^^

v^

ANNALS

OF THE

MISSOURI BOTANICAL GARDEN

t

V

- ^

*^

Annals

o

fth

e

Missouri Botanical

Garden

/

Volume XXXIV

1947

tt ft • n ^
• • • •

ft •

With 48

platp^V^i m*ap§ arid 9<

"» t • • * •

99 figures

Published quarteriV*'^;^alcsVurg,\TllInoi^,3b*' -thV pojrfl of Trustees of

the J^isjolixi BofaiiicaJ'tjSfdcii, StNtouis; lAtt

Entered as second-class matter at the post-ofiGce at Galesburg, Illinois,

under the Act of March 3, 1879.

<\*f

U,'

V

LIBRARY

Annals

of the

Missouri Botanical Garden

A Quarterly Journal containing Scientific Contributions from the
Missouri Botanical Garden and the Henry Shaw School of Botany of
Washington University in affiliation with the Missouri Botanical

Garden, ^

Information

i

The Annals of the Missouri Botanical Garden appears four times
during the calendar year: February, May, September, and November. Four
numbers constitute a volume. ^^

« n • «

Subscription*;l^rjce .y./*„^,!.^^*.* $10.09 per volume
Single Numbers *2!^6 each

Contents of previcJiiS iJsW^ df the 3^]jJMAj.y^<5F-7^E MISSOURI Botanical
Garden are listed in the Agncuftural" Index, 'published by the H. W, Wilson

Company.

**-

• ■ t

• • «

« •

I

^

STAFF
OF THE MISSOURI BOTANICAL GARDEN

Director

George T. Moore

Hermann von Schrenk,

Pathologist

Jesse M. Greenman,

Curator of the Herbarium

Carroll W. Dodge,

Mycologist

Edgar Anderson,

Geneticist

Robert E. Woodson, Jr.

Assistant Curator of
the Herbarium

Henry N. Andrews,

Paleobotanist

Robert W. Schery,

Research Associate

Gustav a. L, Mehlquist,

Research Horticulturist

^

George B. Van Schaack,

Honorary Curator of Grasses

Nell C. Horner,

Librarian and Editor
of Publications

Gerald Ulrici,

Business Manager

BOARD OF TRUSTEES
OF THE MISSOURI BOTANICAL GARDEN

President

Richard J. Lockwood

Vice-President
Daniel K. Catlin

Second Vice-President

Eugene Pettus

L, Ray Carter

Dudley French
Henry Hitchcock

John S. Lehmann

George T. Moore
A. Wessel Shapleigh

Ethan A. H. Shepley

EX-OFFICIO MEMBERS

Arthur H. Compton,

Chancellor of Washington
University

A. P. Kaufmann,

Mayor of the City of
St. Louis

Stratford L. Morton,

President of the Academy of
Science of St. Louis

William Scarlett,

Bishop of the Diocese of
Missouri

C. Oscar Johnson

President of the Board of Education of St. Louis

Gerald Ulrici, Secretary

TABLE OF CONTENTS

A Direct Relationship between Pantothenate Con-
centration and the Time Required to Induce
the Production of Pantothenate-synthesizing

««

Mutants'' in Yeasts

Mendehan Inheritance of Genes Affecting Vitamin-
synthesizing Abihty in Saccharomyces :_

American Origin of the Cultivated Cucurbits

Fossil Polypores from Idaho

Monograph o£ the North American Species of

CorydaHs

PAGE

The Northern FHnt Corns

WiUiam L. Brown and Edgar Anderson 1-28

Mycocandida riboflavina Carroll W. Dodge 31-36

Inheritance in the Carnation (Diauthus caryophyl-
lus) . III. Inheritance of Flower Color

Gustav A. L. Mehlquist and T. A. Geissman 39- 74

The Effect of the Medium on Apparent Vitamin-
synthesizing Deficiencies of Microorganisms

Carl C. Lindegren and Caroline Raut 75- 84

Carl C. Lindegren and Caroline Raut 85- 94

Carl C. Lindegren and Gertrude Lindegren 95- 99

Thomas W. Whitaker 101-111

Henry N. Andrews and Lee W. Lenz 113-114

John Henry Britts — Physician and Fossil Hunter

Henry N. Andrews 115-117

The Idaho Tempskyas and Associated Fossil Plants .

Henry N. Andrews and Ellen M. Kern 119-186

Gerald B. Ownbey 187-258

A Study of Hevea (with Its Economic Aspects) in

the Repubhc of Peru R. J. Seibert 261-352

Some Dynamics of Leaf Variation in Asclepias

tuberosa Robert E. Woodson, Jr. 353-432

Field Studies of Guatemalan Maize

Edgar Anderson 433-467

General Index to Volume XXXIV 469-474

Annals

of the

Missouri Botanical Garden

Vol. 34 FEBRUARY, 1947 No, 1

THE NORTHERN FLINT CORNS

WILLIAM L, BROWN

Ccnctichf, Pioneer Laboratory, Fionccr Hi-Bred Corn Co.j Johtntou, Iowa

AND EDGAR ANDERSON

Geneticist to the Missouri Botanical Garden
EngcUnann Professor in the Henry Shaw School of Botany of Washington University

The slender-eared, wide-kernellcd flint corns of New York State and New
England were for centuries (see Table III) the commonest type of maize in eastern
North America. As dent varieties pushed northward and as earlier and earlier
varieties of dents have been developed, these wide-seeded flints have been restricted
to an ever-narrowing fringe along the northern edge of maize cultivation. Today
they are of secondary economic importance but their role in the production of
the very varieties which supplanted them makes their study imperative to the
modern corn-breeder. In addition to their intrinsic interest as a well-marked and
formerly widespread type of Zca Mays, their close identification with the Indians
of the eastern United States renders their history and relationships of compelling
Interest to the American archaeologist.

During 1944-46 a collection of these northern flint varieties was brought to-
gether and grown in the experimental plots of the Pioneer Hi-Bred Corn Company,
at Johnston, Iowa. We are especially indebted to Dr. R. G. Wiggans of Cornell
University for suggesting sources of seed for a number of eastern varieties.

Tables I and II list the varieties by name, in so far as this was known, and
their places of origin. A photographic record was made of one plant of each col-
lection, and herbarium specimens were prepared of two or more tassels (male
inflorescence). Internode diagrams (Anderson and Schregardus, '44) were made
of representative plants, and the following record was made of the tassels: tassel
branch number, condensation index (Anderson, '44), number of tertiary branches,
presence of whorling in the central spike, and number of paired spikelets per
whorl. Open-pollinated ears were obtained from each culture and were scored
for cob and kernel color, kernel width, kernel thickness, amount of denting, and
diameter of the shank below the ear. These various scores and measurements are
presented in Tables I and IL

(1)

[Vol. 34

2 ANNALS OF THE MISSOURI BOTANICAL GARDEN

For cytological study, sporocytes from each of the varieties were killed and
fixed In 3 parts alcohol to 1 part propionic acid. After 24 hours at room tempera-
ture they were stored in a refrigerator until they were smeared in propionic
carmine. Chromosome knob numbers were obtained from each culture. The
results are tabulated in Tables I and II and are discussed in detail below.

MORPIiOLOGY

It was immediately apparent that, in spite of much plant-to-plant variation,
the northern flints were essentially homogeneous at the eastern end of their range
in New York and New England but became increasingly variable as the Great
Plains were approached. This is equally true whether one considers the morphology
of the plants, the appearance of the ears, or the knob numbers of the chromosomes.

true

and which is described later in this report. The following description therefore
applies to the relatively uniform material from the Northeast. As shown in
Table II, similar varieties are also found in the northern Great Plains but there
they are accompanied by other kinds of flint corns (pis. 3 and 4).

+

The cars of the northeastern flints are characteristically long and slender with
8 to 10 rows of wide, crescent-shaped kernels (pis. 1, 2). The cob is strong and
proportionately large, particularly toward the base, and the shank or ear-stalk is
thick and well-developed. Frequently the base is noticeably larger than the rest
of the ear, and even in those varieties which do not exhibit this character in a
prominent fashion, a tendency in this direction may be seen in increased row
numbers, irregular kernels, or irregular rowing at the base of the ear. Tliis in-
creased basal development In the northeastern flints is most conspicuous when
comparisons .ire made with 8 -rowed varieties from western Mexico and with some
of the flints of the southwestern United States. Such varieties taper toward ilic
base instead of becoming increasingly wider.

be

They

have more suckers or tillers than the common dent varieties from the same area,
and these tillers are usually shorter than the main stalk and often bear malformed
ears and tassels. Prop-roots are less common than in United States dent varieties;
there are usually very few above the level of the soil surface. The culms are small
and slender with long Internodes and are lighter green than most dent varieties.

The leaves are narrow and the ears are borne on long shanks. The leaves of the
ear shoot (the husks) have conspicuous blades (fig. 4) which are sometimes re-

r

ferred to as "flag-leaves" by sweet-corn breeders. The combination of slender
culms, irregular tillers, and well-developed flag leaves gives all these flints a
distinctive general aspect.

The tassels of the northeastern flints are wiry and open. Tassel branch num-
bers are mostly from 12 to 16. There is little or no condensation (fig. 2), -and
the spikelct pairs are thinly and evenly spaced along the secondary branches. The
t.isscls have a slender axis with long internodes. The central spike Is thin and

1947]

BROWN & ANDERSON

NORTHERN FLINT CORNS

3

/

V

Fig. 1. Actual diagram showing branching pattern and spikelet arrangement of a typical tassel
of Longfellow Flint. Left: the 11 nodes of the branched portion of the tassel, showing the number

of branches at each node (in this case 1 or 2) and the approximate direction in which they pointed.
Immediately above these 11 nodes are the first 5 nodes of the central spike, each of which bore 2
pairs of spikelets. Right: a diagram to scale of the actual position of all the spikelets on an 8 -cm.
section of the central spike. The technique is adapted from that used by Mangelsdorf (*45) and
represents a portion of the spike as if it had been slit down one side and flattened out from a
cylinder to a rectangle. Scale at left in cm. All the spikelets in this specimen were in pairs, one
member of each pair being sessile and one pedicellate. All were in 4 ranks and, aside from 1 extra
pair at the lower right-hand corner, were arranged 2 pairs to a node, the pairs being at right angles
to the next nodes above or below; in other words, decussate, which means that if there were spike-
let pairs at the north and south ends of a particular node, those at the next node would be on the
east and west sides, those at the second node at the north and south again, etc.

without the conspicuously thickened central portion so characteristic of most dent
corns and certain varieties of popcorn. The arrangement of the tassel branches is
more regular than appears from casual observation. In those varieties which are
mostly 8 -rowed it is as follows: The upper two branches are opposite and below

4

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

20

10

C.I, 1.0
No. Ill's

WILBUR'S FLINT

20

10

C.I. 1.0
No. Ill's

PARKER'S FLINT

:s

PQ
>1

CO

a

20

S 10

C.I. 1.1

No. Ill's 1

SMUT NOSE

20

10

C.I. 1.1
No. Ill's 2

ASSINIBOINB

20

10

C.I. 1.A

No. Ill's 1

DAKOTA WHITE

20

10

C.I. 1.1
No. Ill's 2

LONGFELLOW

SBC014T)ARY BRAJXHES FHOU EASE

Fig. 2. Lcngtlis of tassel branches, condensation index (C. L), and number of tertiary brancln.'s
for typical individuals of six varieties of northern flint. (Sec Anderson, '44 for details of scoring.)

19 47]

BROWN & ANDERSON

NORTHERN FLINT CORNS

5

them IS a series of branch pairs which are opposite or practically so. Toward the
base of the tassel these pairs become increasingly indistinct until finally there is a
single branch at each node. The lower two branches are usually 2 -ranked and on
opposite sides of the axis though well separated. There is a strong tendency for
the tassel branches, as a whole, to be quite regularly 6-ranked but aside from this
we have been able to find no general regularity in the way they are arranged on
the stem, which varies between the clearly opposite pair just below the central
spike and the alternately 2 -ranked pair at the base of the tassel. Detailed records
of a typical tassel are presented in fig. 1.

X

8

Chromosome knob number

Fig. 3. Frequency distribution of chro-
mosome knob numbers in dent corn inbreds

The arrangement of the central spike is surprisingly simple and does not seem
to have been previously described. In the 8 -rowed crescent-seeded flints it is

L

clearly whorled. At each node there are two pairs of spikelets, one of each pair
being pedicellate and one sessile. The pairs at each node are at right angles to
those immediately below and immediately above, so that the spike, as a whole, is
4-ranked and decussate. This simple arrangement is somewhat masked by a slight
twisting of the axis, and in some plants by a low degree of multiplication (Cutler,
'46, p. 269). In 10- and 12-rowed varieties the patterns may be modified in
various ways. If there is enough condensation (Anderson, '44) to telescope some
of the nodes on the secondary branches, there may be additional spikelets at the
nodes, or the clear division into nodes may not be apparent and the spike will seem
to be arranged spirally instead of being whorled. In some of the Plains flints with
12-rowed ears the central spikes may be whorled but the whorls have three spike-
let pairs instead of two as in the 8 -rowed varieties.

6

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

<

X

o

C

Pi

'V>

>

w

O 6

S

o c

rj

rj

fS

O

u

z

o

00

00

00

oo

DO

<^^

rj

00

1

E

u*\

so

^ 00

r^

a\

so

^4

o

4>

c ^

E

o

E

a

.a

o

U

o

C/3

s

C

>

\r\

ON

00

SO

•^

(N

oc

^ -i-

ir«

I

"a

3

J

I

"T3

"a

O

o

o

-a

u

1

o

4> O

o

I

'ft*

-a

V

■u

V

4>

V

u

Q>

u

<u

u

:C J5 ^
^ ^ ^

e -^

^ ^

CO

c

>

4-1
V)

4>

o

.2

o

z

m

E

(J

D
O

Pi

o

(/I

en

u

o
-a

•

c

at
>^

Q

S

Q

nj

c

3

X

-2

u

O
O

3

Q

E
E

o

J

i: P^

00

C

o

s

E

E

V)

CJ

3

o

O

z

3

B

o

6

o

o

I

•
3

H H -

^

c

V

CiO

<s

u

NO

rt

u

4rt

E

o

W

*n

NH

O

o

E

o

1U

u

r£

u

'*n

'C

C

t

O

1

^

^^

o

■ n

a

-C

-^

care

i4

o

N

Kn

c

*-*

n

'-M

tc

O

u

v>

O

a>

u

ibl

3

V)

^h4

>

.S

u

o

«.»

o

60

c

c

o

T3

u

(A

na

l-i

U

-n

c

E

3

o

z

*

*

*

194/1

BROWN & ANDERSON NORTHERN FLINT CORNS

7

<

<

1

O

OS

>

I

a>

o *

2

o

O O rN O

ITS O

fN rs 1^

*-M m

<N

7

<N

r^

oo

CM oo

t

O r^j

E

a

u

VI

I

oo

oo

<N <N CO

(N

oo QO

fM M

*N

(N

a\

<N

— « (N

fo ON On ^ r\ r-i •-*

ir\ <N ON if\ fM

o\

^O

C4

I

C

c

Q- =

O ! O O

<N

E

ON

3N

On

c^corN,or^^oooooNOK^oo^

ON

OS K

OS o

Crt

CJ

E

■^

rj

N3 O -t ^
-1- ^ Tj- ir^

•^ rvj rvj o -t

^ Tf -4- r-~, ^- Tl~ Tj-

ie*

c

(U

^

a —

4*

1-1

3

a

i
O

o

o
i-<

3 3

3

a

1m

o o

o

^ %J

4»

«* >* <U >s >.

^^^^^oT^-^^

a>

o o o o o

-T3

^

c^ t; b ^

3 3

t4

3

-2 -£

o

V

o

V

u

' ^ i~i i-i i~t t-i i~i i~i i^ ._i

(U

o

V V

tU u

^ ^ i S S %$^^ti^ti^$if^ $ t$ it

o

3
O

3

o

2:

Q

O

o

rl

Q

In
O

2

<• ^

V3

Q

^4

vJC

'Z*

4^

«

L

3

O

c?

ty5

2

o

^ rt rt
-M w w

O O O O

^ ^ ^ ^ ^

rt n m -^ C

|Q Q Q Q-g

W5 jC -^ --C --C O

1^^ 4-» -M i-J W O

^ T3 u. 1-^ Ui u i^

? ri o O O O ■ "<

&

ffl

'^ fi ^

> i^ ^

I « o

«>

^

§

(A

rt O

O O
^ -id

Q Q

C/1

HZ S

01

o o

X

v.-

4>

u^

k^

^j

c

>

o

rt

„>

««

w

o

1

h

c

4 ^4

oine

Squ

^

^

rt

d

r^

ft

^

o

!e

o

o

(aC

C/1

pi^

^

M

L-i

'75

«

n

1

■<1-

<

<

Q

Q

OJ o

o

E

rt
rt

< ^

u

u
C C u.

a o X hj S

ti rj rt

w pq UJ

rt 0! rt

Lh Ui Im

o

^ J^ \^

d,-^

X X x;

«>

o ^

O C C C PL,
^ rt rt rt ^ ^

S

eu Q

o

Z

■5 ^

C VI « ;/3 oO
. 2 V> to 1/5

4^

9

C u

rt 3 3 3 2

Pi cii =r:i a; J?

V5

s

wo

n 1

EX (^

OO ^

bO

OJ

o

to

o

V5

o

S

o

O

c

ft Fh

Is

« §

O U 4>

.12 ja "U

> « c

o

O CQ

J I

u

' ^?

be C

=5 t; ^

.-H 3 O

-2 5
f 3 6

r o ^

O V3

Z

* *

* * *

8

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

m

w

in

xn
O

<
>

w

N

<

u

o

H

t/3

o

O

PS

5 o

o

w

O

O

w
o

o

Pi

p

ft

E

W

g

8

tn

o *

tf)

c

P P-i

o o

I

cj o <i>

(A

Wl

v; tn 4_t bn M

CJ rt rt C ^ C
^ ^ OJ ^ <U

6

.2

8

I

d

CJ

M

rj

E

I

a\

oo

K

M)

O

O

u

u

E

^o

■^

<N

..

00

o

§

oo "O

o

rt

<

to

c/^

'^ I S

60 o 2 ,_,

o *-• ^

to

d

•n

C

M

&o

* ^4

ft

c

rt

u

pairi]

-a

C

*

ent see

ong ro

ent see

Fort An

-1700

enting

wasco'

g row
shank

1

o

-T3

Stron
big

Cre:

1-1

U

150

O

V

U

O

a

u
O

c

o

tn

t^

-^

ts

«^ f-«

fA

*-%

r4

(N w^ tr\

\o

ro

(N

*N

<M

r^

r^

CM

■^

0\ Tj- ITN

U^

r^

'^ ^*^ so

CM

^

Pipi cti XO O

00

s

S2

S 2SS<SSS2i;

o

V)

c

^ In

1

o

u

u

u

C

u

C

u

00 60

t- a

d ^

d
o

d
o

pa

o

Qi

o

O

o

■ - rt

U3 U

-T3
C

O

K u^

CM

G d d

*<

4rt

nj

pq

CO

.5 o

IT fc^: Z

V5

rt

O

pq

1/1

o

z

d

O

u

rt

u d

rt

O w

d

rt 3

York

York
York

New

zz

u
O

Z

o

zo

d

rt

60

>^ OJ Q> IJ
1^ CI n> gj

G

-— tri t/i w^ '-rr rt c

O

o o

oo

o o

o

a

o

c

O O O S

_ a; o o 1- to (Ti

«-) d d d f^ u. n

d d d d (J o "^

w-^ U OJ U O V ^^

1947]

BROWN & ANDERSON NORTHERN FLINT CORNS

9

OJ

C

d

w

<

(A

s

t«o

5

o

c

8

ril

airi
ing

M

\ It

*

4>

4>
^ >

kar

denting

Ear elliptica
widest abi

toric Ari

Strong ro
Strong row

widest a

o

Z

His

Ear

o

to

c

u

u

i::^ o ^

h-1

-a

i-l

a

a, w

u

3

W

1/1

o

u

rt

l-i

60 (* O i> '-'

60

G

rt

£

£

^^ 'S «>

a>

o

G

^

E

-a

n

(N

OS

OO

K

^

o

^o

^

£

2:

^

OO

s

(A

s

CO

(y5

r<

(N

rr-,

T3

o

o
c

a>

4-1
O

CO

00

ITS

*N

rr\

OJ -^

<

<

rs|

<N (N

-^ CJ

<S SS

z:z; 1^:2:

o
U

c

d

O

P3

O ^

u

o
O

u

UT)

'5
Q

CO

C

V

CO

o
U

K

C/1

^ CI.

o o

H Q

O O

QQ

JS

IM w u

^ ^ ^

Q> a> ^

1/1

to

c

to
rt

CO

rt
t/>

3

O
trt

IT)

on CO Z Z 2 ^ Ea:^

- "^ o o

E

T|-

*

I

I

3

p

O

(A
Vi

zz

^

<N

s s <

rt

c

<

C
rt

<

22

c c

M V)

U OJ

t> >

rt rt

u u

£ £

o o

Ut 1-1

rt

C
rt

rt
</^

C

rt

S < <:

o

3

cr
o

T
U

w

'5

-0

o

u

rt

o

rt

rt
rt

V5

o
o

rt

U
rt

^

c
P

c

n

o

a

w

E

3

u

(A

3

t »

E

3

(A

3

4 ^

[Vol. 34

10 ANNALS OF THE MISSOURI BOTANICAL GARDEN

CYTOLOGY

The northeastern flints make excellent cytological material; as compared to
most other United States varieties they are easy to smear and yield a high per-
centage of good preparations. This is, in part at least, the result of their being
knobless or essentially so, that being the outstanding cytological feature of these

corns. A summary of chromosome knob numbers for flints from the Northeast
and from the Northern Plains is presented in Tables I and 11. It will be seen that,
aside from the organizer knob on chromosome 6, the northeastern flints arc knob-
less or have onl)' one or two knobs. The organi/cM' knob Is never large as in some
Central American varieties. On the contrary, it is usually extremely small, light-
staining and inconspicuous (fig. 5). The knob positions In the northern flints are
also characteristic, as had previously been reported by Longley ('38) for the
varieties of the northern Indians. When knobs are present in these varieties they
are usually either small and terminal on chromosome 9 or two small knobs on the
long arm of chromosome 6. The form of the terminal knob on chromosome 9 is
a!so characteristic. It may be more or less cleft at the apex, or it may taper to an
acuminate point; it is never large and cylindrical as in certain varieties from

r

western Mexico. In all cases knobs in the northern flints arc small; often they
are only slightly larger than a large chromomcrc.

Since the varieties included in this study were from open-pollinated seed and

were more or less heterozygous one might expect knob numbers to vary con-
siderably within varieties. Actually the variation is slight^ usually not more th:

one knob. Where different knob numbers were observed the number listed rep-
resents the average of a number of counts. Among the varieties examined cyto-
logically, two (Parker's Flint and Twelve-row Dakota) were found to contain, in
addition to the normal chromosome complement, one and two pairs of B chromo-
somes. The B chromosomes in the variety ''Twelve-row Dakota" were unusually
large and the pairs were frequently joined at pachytene of meiosis.

ARCHAEOLOGY

The rapid advances of American archaeology in the last few decades have added
greatly to our knowledge of the history and development of maize. As additional
collections of prehistoric material become available and as cultural sequences are
more accurately determined, the history of corn in North America will no longer
be a matter for argument. It will instead come within the domain of measurement
as various collections are recorded and carefully compared with one another. For

the northern flint corns, while some details remain to be filled in, the outlines of
the story are already clear. They are sho\^ n in map 1 and a detailed report is pre-
sented in Table III. Before going into the minutiae of these collections, the non-
archaeological reader may be helped if we anticipate the chief conclusions;
Eight-rowed flints, similar to those described above, were widespread in pre-
Columbian times in the eastern United States. They go back to approximately
1000 A, D., and over wide areas in the eastern United States no other kind of

\

19471

BROWN & ANDERSON NORTHERN FLINT CORNS

11

._ vs'.-^_-_- ---^^ i>x^r¥'^Xi?^-3?>^ ::5fi^^^^Pet'-':'"»:i: :.tb -r-

Fig. 4. A sini;Ic car of Stevens' Flint showing extensive husk-leaf
development representative of many varieties of northern flint corn.

12

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 3A

mf

V

%

•#»'•"'*»

«

>!■>

>

w^

J^

«'

4^

•

*

Fig. 5. Pachytene cliioMioM)nics ut I,()n^t\.llow I lint showing sm.^ll
organizer knob on chromosome 6. The hirgcH .iriow points directl)
to the ur^ani/er kntib. It will he noted that the knob is so small as to
be scarcely visible against die side of the nucleolus and is only sUghth
larger than the terminal chromomerc of the satellite just below it.
T!io smaller arrow points to the small terminal kni)b on chromosome 9.

1947]

BROWN & ANDERSON

NORTHERN FLINT CORNS

13

Map 1 . Distribution of
large dots, crescent-seeded 8
various other types.

collections
■10 rowed

of prehistoric corn in the eastern United States:
flints (i. c., like northeastern flints) ; small dots,

maize has been reported in archaeological remains. In the Great Plains, by con-
trast, there were widely divergent types of corn as well as northern flints and

apparent mixtures with them. The sequence of these types of flint in the Great
Plains remains to be worked out. The situation in the American Southwest is
equally complex but one fact is certain: The northern flints arrived there relatively
late (about 1300 A. D.), long after other types of maize had been established in
that region. *

The facts on which these conclusions are based are presented in condensed form
in Table 111 and map 1. These summarize the collections of archaeological maize
from the Great Plains and the eastern United States which we have so far ex-
amined, some 3 6 in all. They represent all the collections readily available at the
Rochester Museum, the Peabody Museum of Harvard University, the Ohio State
Museum at Columbus, the Museum of Ethnobotany of the University of Michigan,
the National Museum at Washington, the Library of the Iowa State College at
Ames, and the Chicago Museum of Natural History. To the curators and staffs
of these museums we are greatly indebted. They not only made the material
available for study but supplied us with literature and references in addition to
much general Information on archaeological matters. It will be noted that nothing
from the southwestern states is included In this survey. The situation there is
much too complex for discussion here and is somewhat outside the scope of this
paper. For a detailed report on one prehistoric collection of southwestern maize

14 ANNALS OF 1 HE MISSOURI BOTANICAL GARDEN

[Vol. 34

4

which includes 8-rowcd flints, see Anderson's appendix to Haury's ('45) report on
Painted Cave. For a general discussion of the evidence on types of maize in the
Southwest see the fourth chapter, pp. 39-5 5, of Carter's ('45) "Plant Geography
and Culture History in the American Southwest."

Some of the data in Table III arc summarized in map 1. At least three main
types of corn occur in the Northern Great Plains: (1) Varieties very similar, If
not identical, to the northeastern flints described above. They have wide, crescent-
shaped seeds, thicker at the apex of the kernel than near the germ. They are
straight-rowed with strongly paired rows, and are predominantly 8-rowed. (2) In
the Northern Great Plains, in addition to the above varieties, arc others which are
more or less simihir but have higher row numbers and smaller, squarcr kernels.
(3) From rock shelters and caves from the Ozarks to southern Ohio are found
collections of a very different type of corn. Some of these are well preserved.
From others the evidence is fragmentary. They resemble the so-called prehistoric
Baskctmaker corn of the Southwest in their irregular-shaped kernels, their ears,
which taper to the base as well as to the tip, and their high percentage of cars with
row numbers from 12 to 14. Their presence in this area and their resemblance to
Basketmakcr corn raise questions which are completely outside the scope of this
paper. The point in question is the 8 -rowed crescent-seeded flints. Map 1 and
Table III demonstrate that such varieties have been in the eastern and northern
states for some centuries at least and that they were once very widespread there.

If we catalogue the varieties of corn by their general resemblance to each other
m all characters rather than by the texture of their endosperm (Anderson and
Cutler, '42) it will be seen that a number of sweet corns, a few of the older
varieties of popcorns, and some of the flour corns of the eastern Indians are very
closely related to the northeastern flints. They resemble them in their early
matu:ity, crescent seeds, predominance of 8 rows, tillering, flag leaves on the car,
absence of prop-roots, and structure of the tassel. The few which we have
examined are also very similar cytologically.

DISCUSSION

T/^ fiort/jcrn flifits as ancestors of modern corn-belt varieties.

While the northern flints, as such, are now little more than a curiosity in much
of the region where they were formerly grown, they are indirectly of both practical
and theoretical significance because they are at least one of the ancestral types of
the varieties which replaced them. There Is abundant evidence that the varieties
of the United States corn belt originated by repeated hybridization between the
northern flints and soft-textured southern dents.

Until the early 18 00's nothing like the big cylindrical, yellow maize of the
corn belt, with its keystone-shaped, dented kernel, was known in the United States
or elsewhere. As American agriculture developed and pushed westward th
northern flints were progressively more and more mixed with soft white dents
spreading up from the South. The latter were In many ways similar to some
present day Mexican varieties. Lorain, whose book appeared posthumously In

^ »

1947J

BROWN & ANDERSON NORTHERN FLINT CORNS 15

1825, described the ears of these southern dents as not very long, neither is the
cob so thick as that of the big white and yellow [flint]. But the formation of the
grain makes the ears very thick. They frequently produce from thirty to thirty-
two and sometimes thirty-six rows of very long narrow grains of a soft, open
texture. These grains are almost flat, at their outside ends." He also states that
this dent "ripens later than any other but is by far the most productive." (p. 203).
The commonest name for these soft dents was *'Gourdseed", since the flat kernels
with a collapsed and more or less pointed tip resembled a pumpkin seed or gourd

seed.

Lorain discussed in detail the results to be obtained from mixing Gourdseeds

and flints and went on to say:

"The quantity of the Gourdseed corn mixed with the flinty yellow corns, may be
determined, so as to answer the farmer's purpose. When the proportion of the former
greatly predominates, the grains are pale, very long and narrow, and the outside ends of
them are so flat that but little of the indenture is seen. As the proportion of Gourdseed
decreases in the mixture, the grains shorten and become wider, and their outside ends grow
thicker. The indentures also become larger and rounder, until the harder corns get the
ascendancy. After this the outside ends of the grains become thicker and more circular.
They also grow wider, and the fluted appearance between the rows increases. The indentures
also decrease In size until they disappear, and the yellow flinty varieties are formed. But
as I believe, not so fully but that the latent remains of mixture will forever subject it to
more or less change." (loc. cit., pp. 205—206).

The churning and rechurmng of the Gourdseeds and the flints continued for
several decades. By 1837, P. A. Brown listed seven different varieties known to
him which had originated in that fashion. For the year 1850 we have an unusually
complete picture. Before there was a Federal Department of Agriculture, the
Patent Office published an annual summary of the progress of American agricul-
ture; questionnaires were sent out to leading farmers and the replies were sum-
marized and woven into an essay. For 1850 (U. S. Comm. Patents Rept., '50) the
replies were printed practically as written, not even being sorted according to
states. Since the first question to be answered had been: "What varieties of corn
are most esteemed in your vicinity?", the replies give a detailed picture of the
kinds of corn grown in the United States in the middle of the 19th century. The
corn belt was just then taking shape In Ohio. Three of the replies from that state
describe the mixing of flints and Gourdseeds which was taking place,
vatc several varieties of what is here called gourd-seed. They are all nearly a
hybrid between the rough gourd-seed of the South and the flints of the North."
(p. 371), Another letter asserts that the best varieties arc "obtained by mixing

We

Souther

Another states that

there are "many good varieties, mostly crosses between gourd-seed and the small
flint." (p. 454). Only one reply about corn was received from Illinois which was
then outside the corn belt (p. 245). It reports that in the vicinity of Quincy the
most esteemed variety is "a species obtained by mixing the large yellow corn of
Kentucky with the yellow flint." The white Gourdseed is also said to be planted.
Mixtures of Gourdseed with various southern corns are spcc^ifically mentioned in
reports from North and South Carolina, Virginia, Alabama,v and Mississippi.

16 ANNALS OF THE MISSOURI

lA'oI.. 34

Co

New York, and they were still among the outstanding varieties for Massachusetts,
Ohio, Kentucky, Illinois and Michigan, The expression "dent corn", incidentally,
is used only in the three letters from Michigan (pp. 309, 410, 412).

There can be little doubt then that our corn-belt dents originated during the
first half of the nineteenth century by a manifold mixing of northern 8- and 10-
rowed flints with many-rowed southern dents. In addition to the precise evidence
given by Lorain and the Patent Office report for 18 50 there are numerous ref-
erences and descriptions in other agricuhural writings. For detailed accounts the
reader is referred to Edward Enfield's monograph on ''Indian Corn," published in
1866; Fearing Burr Jr.'s, "Field and Garden Vegetables of America," 1863;
Browne's, "Essay on Indian Corn" in The 'Farmers' Cabinet' for 1838, and the
Transactions of the New York State Agricultural Society' for 1848.

T/jc importance of northern flints in modern corn breeding

The demonstration that our corn-belt dents are derived in part from th
northern flint corns is of more than academic interest. It has been shown (Ander-
son, '39) that in crosses where any considerable number of genes are concerned
the total forces of varietal cohesion are vastly greater than is usually appreciated.
In such crosses all the multiple-factor characters will be partially linked with one
another, and while a bewildering variety of new forms may appear, on the whole,
the combinations of characters which went into the cross together will tend very
strongly to stay together In the hybrids. If the number of segregating genes
exceeds three per average cross-over segment (a not unlikely figure in crosses be-
tween northern flints and southern dents) then the linkages can be broken only by
long generations of controlled breeding. Though approximately a century has
elapsed since the mixture of the southern dents and the northern flints was begun,
we may well expect that enough of the genes contributed by the flint varieties are
still so linked with one another, on the average, to render this linkage worthy of
consideration in any corn-breeding program. In producing hybrid corn, for in-
stance, some of the difi"iculties encountered are due to the fact that we arc not
working with a homogeneous mixture of dent corns as such; we are working with
a mixture containing large blocks of germ-plasm of southern dents and of northern
flints. Hard kernels, a low row number, cylindrical ears, and early maturity were
quahties which went into corn-belt corn from the flints. It is a matter of common
knowledge among experienced present-day corn breeders that these qualities still
tend to stay together. Knowing that these qualities went in together from the
flints, it should not take too long, by experimental breeding, to produce at least a
rough estimate of their distribution in the germ-plasm of corn-belt varieties. Arc
they scattered equally over all ten chromosomes, or are they concentrated on a
few? Are the gene differences to be estimated in the tens, the hundreds, or the
thousands? It should be possible within a reasonable length of time to answer
these questions in at least a provisional way, and data of this nature should be quite

194 7]

BROWN & ANDERSON NORTHERN FLINT CORNS 17

useful to the modern corn breeder in his efforts to improve existing inbred Uncs
and to create new and better ones.

Kelatianship between c/yromosame knob numbers afiJ morphology of inbreds. —

There is as yet little exact evidence as to how completely the gene combinations
introduced from the northern flints have been broken up in modern dent corns.
Our determinations of chromosome knob numbers in 65 inbred lines of corn-belt
maize bear directly on this point (fig. 3).

While the numbers of inbreds investigated is still too small to represent an un-
equivocal demonstration, the general trend in ear and plant morphology from one
extreme to the other as one passes from inbreds with low knob numbers to those
with higher numbers is most suggestive. The inbreds with knob numbers of ap-
proximately two are clearly the most like the northern flints of any of the 65
which have been studied. It would seem that the total effect of the forces holding
the germ-plasm of the northern flints together is so strong in modern maize, even
after a century of mixing, that the coherence can be demonstrated cytologically.
If this be true, it represents racial coherence of a very high order of magnitude,
for the knobs serve as cytological markers for only a portion of the germ-plasm.
Any specific knob can serve only as a marker for the arm or part of arm of the
particular chromosome in which it occurs. Since there are 10 chromosomes and
therefore 20 chromosome arms, the difference between the high knob lines of 8
and the low knob lines of 2 is at most a difference in only 6 out of the 20 arms,

I

or 30 per cent. It seems, therefore, significant that with markers in only 30 per
cent of the chromosome arms, there is still an indication of resemblance to the
northern flints in the low knob inbreds.

It may be that when a larger number of inbreds have been examined the rela-
tionship between low knob numbers and flint-like characters will not be as definite
as these preliminary results have indicated. The low knob number of the northern

flints, however, is definitely established. This fact poses a number of questions
since it seems to be in direct opposition to Mangelsdorf and Cameron^s ('42) pioneer
work on the same subject.

Mangelsdorf and Cameron determined the knob numbers of over 150 varieties
of maize from Guatemala and demonstrated the association of high and of low

nvimbers with various contrasting characters of the ear and plant. Two of the
most definite associations which they established were between high knob number
and cylindrical vs. tapered ears, and with straight rows t's, irregular rowing. On
the basis of their findings we might expect the northern flints to have the highest
knob numbers of any United States varieties of corn. Actually they have the

lowest, as we have shown above. These two facts, however, are not as diametrically
opposed as they might seem. The corn of the United States is not the corn of
Guatemala, nor could all of it have been directly derived therefrom. Much of It,
theoretically the greater part of it, must have spread Into the United States by way
of Mexico. For that country we have only about 50 knob determinations but in

18 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

general they agree with those of Mangelsdorf and Cameron. In wMtern Mexico

F

there is a whole group of varieties with cylindrical cars, high-lodging resistance,
and growing chiefly at low altitudes. They have high knob numbers as Mangels-
dorf and Cameron would predict. In Central Mexico, mostly at very high alti-
tudes, there is a group of tapering-earcd dent corns which lodge badly and are
smut-susceptible. Mangelsdorf and Cameron would predict them to have generally
low knob numbers which they do, being from to 5 in the material we have in-
vestigated. Much of the dent corn of Mexico is intermediate between these two
extreme types (Anderson, '46), and the few examples we have investigated have
intermediate knob numbers as might be expected. It was such varieties as these
which eventually spread northward into the United States. If for the moment we
sidestep the question of where the northern flints came from originally but keep in
mind that they have few or no knobs, then our results come closer to falling into
line with those of Mangelsdorf and Cameron. By a mixture of old, southern dents
with mcdiumly high knob numbers and northern flints with few knobs or none,
then a situation such as we have described would have developed.

Oiigiu of the northern flints, —

The above hypothesis is satisfactory as far as it goes, but it leaves unexplained
the origin of the northern flints and advances no reasons for their having few or
no knobs. Only the most tentative of explanations can be offered at the present
time. As has been pointed out above, the northern flints are characterized by
wide, crescent-shaped seeds on a cylindrical, few-rowed ear with a strong cob more
or less enlarged at the base and borne on a stout shank. This is a distinctive com-
bination of characters. Since somewhat similar varieties are known in the American
Southwest, Mangelsdorf and Reeves originally suggested (*39) that the northern
flints spread into the eastern United States from that direction. In this respect
they are almost certainly wrong. We have specific archaeological evidence that
the northern flints are definitely pre-Iroquoian in eastern North America (see
pi. 6), There is abundant and definite evidence (Carter, '45; Carter and Ander-
son, M5) that varieties like the northern flints did not reacli the American South-
west until after 1200 A. D. There is even some evidence to suggest that they
reached the Southwest as varieties relatively similar to those in the East and that
they then underwent hybridization with some of the varieties already present in
the Southwest to produce the typical long-eared sorts of the modern pueblos
(Haury, '45). Furthermore, the very similar long-eared varieties of northern
Chihuahua most certainly represent relatively late southern extension into Mexico
as had already been determined from cultural evidence (Sayles, '3 6).

If the northern flints could not have spread from the Southwest, whence could
they have come? Varieties with wide kernels, stout cobs, and a more or less en-
larged base are practically unknown from most of Mexico. However, they are
present in Guatemala and adjacent Chiapas. It seems probable that northern flint
corns may be among various cultural traits which spread from south of the

1947]

BROWN & ANDERSON NORTHERN FLINT CORNS 19

tepee

clear record cf the route by which they journeyed.

If the northern flints did come from Guatemala, it is still necessary to explain
their low knob number, since many of their ear and plant characters are essentially
those Mangelsdorf and Cameron ('42) found to be correlated with high knob
number in their Guatemalan survey. They interpreted the high knob number as
due to crossing with teosinte, which is known to have a very high knob number
and which, according to Mangelsdorf and Reeves' hypothesis ('39), was itself
derived from previous hybridization between maize and Trrpsacum. On these
hypotheses, therefore, the high knob numbers and the associated characters came
ultimately from Tripsacumy and Mangelsdorf and Cameron applied the term
"tripsacoid" to these varieties. Since that time, however, Graner and Addison
('44^ have reported that Tripsacum australe of South America, unlike its North
American relatives, is lacking in terminal knobs. Assuming that Graner and
Addison^s observations are typical of the cytological picture in Tripsacum australe^
then we are faced with the possibility that introgression of Tripsacum germ-plasm
into 2ea might have various effects upon knob number, as Cutler ('46) has re-
cently suggested. It is quite possible, therefore, that our results with the northern
flints can be harmonized with the hypotheses put forward by Mangelsdorf and his
collaborators. Before that can be accomplished, however, we shall need to have
a much more detailed understanding than we have at present of the relationships

South

SUMMARY

1, Though no longer of much commercial importance, the northern flints are
of interest to anthropologists as a type of corn once very wide-spread in the eastern
United States. They are also worthy of consideration by modern corn-breeders as
one of the ancestors of modern United States dent corns,

2. A representative collection of northern flint varieties was grown. Its gross
morphology, its pachytene cytology, were systematically investigated. The varie-
ties of flint corn from New York and New England are substantially uniform
morphologically, cytologically and archaeologlcally. Similar varieties are also
grown on the Northern Great Plains, but the collection from that area is more
variable and includes other types, as it has since prehistoric times.

3. The northeastern flints (those from New York and New England) have
slender culms^ irregular tillers, well-developed flag leaves, few visible prop-roots,
and are of early maturity. Their ears are cylindrical, 8- to 10-rowed, with strong
shanks and proportionately large cobs. Their kernels are wide, undented, and not
pointed. The tassels are wiry, with no condensation. The central spike bears its
spikelets in whorls of two pairs: the pairs 4-ranked and decussately arranged on
the spike.

4, The pachytene chromosomes of the northeastern flints show few knobs or
none at all, and the knobs, when present, are usually small.

[Vol. 34, 1947]

20 ANNALS OF THE MISSOURI BOTANICAL GARDEN

5- There is abundant archaeological evidence to show that similar varieties
of corn were common in eastern North America in prehistoric and protohistoric
time. Over wide areas in the eastern states they are the only maize so far obtained
in archaeological excavations,

6. The northeastern flints were widely grown commercially in the United
States in colonial times and afterward. During the first half of the 19th century
they were extensively and repeatedly hybridized with soft dent varieties from the
South, giving rise eventually to the typical cylindrical-cared dent varieties of the
United States corn belt.

7. Though the amalgamation of the northeastern flints and the southern
dents has proceeded for nearly a century, some of the characteristics of the
northern flints are still more or less linked in the germ-plasm of modern United
States commercial corns. A cytological survey of 65 inbred Hnes of dent corn
showed chromosome knob numbers of from 2 to 8. The inbreds with the lowest
knob numbers (i.e. the most flint-like) were most similar to the flints in their
external morphology.

8. The origin of the northeastern flints is briefly discussed. While they are,
in general, unhke Mexican maize but show strong resemblances to certain varieties
from Guatemala, the problem cannot be seriously approached until more detailed
information concerning the morphology and cytology of Centra! and South
American varieties is available.

LITERATURE CITED

Anderson, Edgar (1939). Recombination in species crosses. Genetics 24:668-698.

, (1944). Homologies of the ear and tassel in Xca }Aay$. Ann. Mo. Bot. Gard. 31:325-344.

■, (1946). Mai/c in Mexico. A preliminary survey. Ibid, 33:147-247.

. and Hugh C. Cutler (1942). Races of Zea Mays, I. Their recognition and classification.

Ibid, 29:69-88.

. and Dorothy Schregardus (1944). A method for recording and analyzing variation in

internode patterns. Ibtd, 31:241-247.

Brown, P. A. (1837). Cited in Wallace, H. A., and E. N. Bressnian (1937). Corn and coTn
growing. 436 pp. New York.

Carter, George F. (1945). Plant geography and culture history in the American Southwest. Viking
Fund Publ. in Anthropol. No. 5:1-140. New York.

Cutler, Hugh C. (1946). Races of maize in South America. Leafl. Bot. Mus. Harvard Univ.
12:257-292.

Graner, E. A., and George Addison (1944). Meiose em Tripsacum australc Cutler e Anderson.
Anais da Escola dc Agn'cultura, PIracicaba, Brazil 1944:213-224.

Haury, Emil W. (1945). Painted Cave, northeastern Arizona. Amerind Foundation No. 3:1-87.
Dragoon, Arizona.

Longley, A. E. (1938). Chromosomes of maize from North American Indians. Jour. Agr. Res.
56:177-195.

Lorain, John (1825). Nature and reason harmonized in the practice frf husbandry. 563 pp.
Carey & Lea, Philadelphia.

Mangclsdurf, P. C, and J. W. Cameron (1942). Western Guatemala, a secondary center of origin
of cultivated maize varieties. Leafl. Bot. Mus. Harvard Univ. 10:217-256.

■ , and R. G. Reeves (1939). The origin of Indian corn and its relatives. Texas Agr. Exp.

Sta. BulL 574:1-315.

Sayles, E. B. (193 6). An archaeological survey of Chihuahua, Mexico. Medallion Papers No.
22:1-191. Globe, Arizona.

U. S. Commissioner of Patents. (1850). Report for the year 1850. Part IL Agriculture. Wash-
ington. 1851.

[Vol. 34. 194:;

22

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plates 1-5

In these plates the lance-shaped object at the left-hand margin is a tracing of the first
leaf above the ear. In the photographs of the cars each of the divisions on the scale rep-
resents one centimeter,

i

Plate 1. Stevens' FKnt — typical plant, ears, and kcrnoN.
Plate 2. Parker's Flint — typical plant, ears, and kernels.

Plate 3. Dakota Wliite, a variety of the Great Plains having many characteristics of

eastern flints.

Plate 4. Twelve-row Dakota. This type is quite different morphologically from the
northeastern flints but similar in many ways to certain varieties of the south-
western states.

Plate 5. Spanish Popcorn. Representative plant, ears, and kernels. A flint of very
' early maturity whose morphology suggests relationship to both northeastern and
Great Plains varieties.

L I

\

4

+

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 1

Stevens' Flint
BROWN & ANDERSON— NORTHERN FLINT CORNS

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 2

1

^dNMf

mi^

^m

■mmn:

>

Ti=-J

■^■«ps

^-

- ^ ^^ ^

■ . . y.

i-i^

^i^-.

■f %«

^.-"^

s

?*i*

i^.^-n

.^.

mm^m m

■ y^^^

« ^

■^.

Parker's Flint

BROWN & ANDERSON— NORTHrRN FLINT CORNS

Ann. Mo. Bot, Card., Vol. 34, 1947

Plate 3

Dakota Whiie
BROWN & ANDFRSON— NORTHERN FLINT CORNS

Ann. Mo. Bot. Gard., Vol. 34, 1947

Pi ATE 4

TwELVL-Row Dakota
BRO>X^N & ANDERSON— NORTHFRN FLINT CORNS

Ann. Mo. Bot. Gard., Vol. 34, 1947

P[ ATE 5

fc^Mft ' *■■ '^'^^t^ ,^^-lr4,^»^^

SrxNisH Popcorn

BROWN & ANDERSON— NORTHERN ELINT CORNS

28 ANNALS OF THE MISSO

[Vol. 34, 194

EXPLANATIOX OF Pi ATE

PLATE 6

rhotOKr.iphs of charred maize cobs from Gibraltar Site, Wayne Co., Mich., collected
by Dr. J.merson F. Grecnman, Museum oi AnthropoIoi;y, University of Michi>;an, sum-
mer 1938. ^'Owasco" (probably before 1200 A.D.). Note the S-rowed cobs, tlu' Avide
alveoli, strongly paired rows, and lar>;e shanks. Photoi;raph courtesy of Volncy Jones and
the Museum c»f l.thnt>botnny of rlie University of Michigan.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 6

_-i _'_^j j_'_-

p.. ^PJT

i

■\

1

I

I

i

^.■.

^

BROW^N & ANDERSON— NORTHERN FLINT CORNS

MYCOCANDIDA RIBOFLAVINA

CARROLL W. DODGE

Mycologist to the Missouri Botanical Garden
Professor in the Henry Shaw School of Botuny of Washington University

The following study is based on culture no. 921, resulting from a long series
of selections from stock culture no, 321 in the collection of Anheuser-Busch, Inc,

St. Louis.

J

/

Mycocandida nboflavina Dodge, sp. nov.

Pseudomycelium ex cellulis longe ellipsoideis, 10-11 X 1.6-2.5/x, ramis laterali-
bus 1-2 (-4) ad qucmque nodum, cellulis 1-3. Colonia parva, cremea vel albida,
laevis vel subfoveolata, margine tenui. Colonia rugosa cremea, crateriformls, rugis
radiantibus, subelevatis, margine crassiori. Gelatina tarde liquefacta. Glucosa,
fructosa, mannosa, sucrosaque fermentatae.

Cells in young cultures variable in shape from ellipsoid to long-ellipsoid, or
ovoid, often with both ends rather acute, budding polar, but cells not apiculate as
in Kloeckera (Psendosaccharomyces) , mostly single, a few in short chains.

In old liquid cultures (four months) similar to young cultures but chains
somewhat longer with 2 (rarely up to 4) short branches at the nodes (up to 3
cells long), cells ellipsoidal to subpyriform, terminal cells short-ellipsoid to
nearly spherical; one spherical cell seen with 4 buds at one end and 3 at the other.

On old malt agar cultures, pseudomycelium well developed, cells 10-11 X 1-^"
2.5^, branching lateral, only 1 or 2 branches at a node, cells long-cUipsoid or with
the end bearing the branch slightly enlarged and more rounded; terminal cells

shorter.

No ascospores produced on old cultures (some completely dried out) nor on

gypsum blocks nor on Gorodkova agar.

COLONY CHARACTERS AND SECTORING

On malt extract agar (15° Balling), colony small, margin thin, sloping gently
to the center, surface smooth with very minute pitting and with some very shallow
radial valleys, with a small rugose sector. Transfers from the rugose sector pro-
duced colonies with a shallow central crater, with low broad radial folds and a few
cross folds, margin circular and somewhat elevated, with a smooth sector occupy-
ing about one-sixth the circle. Colonies cream buff with a lighter margin. Trans-
fers from the smooth sector produced colonies with a very low dome in the central
crater, sloping gently to the margin, with 4 or 5 radial valleys, surface smooth
with some very shallow pits (visible under a 9 X hand-lens), margin very smooth
and slightly elevated. Colonies cartridge buff or darker, margins somewhat lighter.
No further sectoring occurred on smooth colonies.

'Mrak, E. M., H. J. Phaff, R. H. Vaughn, and H. N. Hansen. Yeasts cM:currmg In souring figs.
Jour. Bact. 44:441-450. 1942.

Issued March 29, 1947.

(31)

\

[Vol.. 34

32 ANNALS OF THE MISSOURI BOTANICAL GARDEN

On Sabouraud glucose agar, colonics with a low central plateau, broadly crenatc
margins with narrow radial valleys connecting the central plateau to the notches
in the margin. Transfers from the smooth type of colony on malt extract agar
produced colonies with a nearly smooth center (only a very faint suggestion of a
crater), sloping very gently to the margin, surface rather dull from minute pitting
(about the limits of visibility with a 9 X hand-lens), no radial valleys nor ridges,
margin less elevated, with the faintest suggestion of marginal strmtion (under 9 X
lens). Color pale ochraceous salmon. Transfers from the rugose sector on inalt
extract produced colonies with a low central dome in a very shallow cxcentric
crater, sloping gently to the margin, a very few slight ridges and radial valleys
(arranged as the lamellae of a mushroom). Colony light ochraceous salmon, mar-
gin paler. No sectoring was observed on Sabouraud glucose agar.

Yeast decoction agar: colony more elevated, very moist and shining, faint
depression in center with about 4 very shallow, radial valleys, pure white. Trans-
fers from either smooth or rough sectors on malt extract agar produced the same
type of colony on yeast decoction agar.

BIOCHEMICAL ACTIVITY

In general, liquid cultures produced a slight ring on the sides of the tube, no
islets nor pellicle; the liquid remained clear and the sediment finely granular; in
old liquid cultures (about 4 months), the ring is a little better developed and
the sediment becomes slightly more flocculent. In fermentation tubes of glucose,
mannose and lactose, a few islets developed but they were never abundant nor did
they coalesce to form a pellicle. Litmus milk was neither acidified nor coagulated;

a slight ring and abundant sediment developed In the tubes, showing that growth
occurred. Gelatin was very slowly liquefied, complete in 16 weeks, abundant
sediment, but no ring nor pellicle.

Fermentation: Gas is produced with glucose, fructose, mannose and sucrose,
none with maltose nor lactose. Gas is produced much more slowly than with
Saccbaromyces cerevhtae, not showing until the second day after inoculation. No
acid was produced under anaerobic conditions (long arm of fermentation tube) ;
acid with glucose and lactose, none with mannose, maltose nor sucrose under aerobic

conditions (short arm of fermentation tube). Sediment abundant in all tubes.

Since our organism was thought possibly related to Brcitanomyces and si
most species of the latter produce an after-fcrmcntation of beer, 95 per cent ethyl
alcohol was added to 15° Balling malt extract to make final concentrations of
5-12 per cent ethyl alcohol. At 5 per cent after three weeks, there was good
growth with abundant flocculent sediment; at 6 per cent growth was good, but
with less and more granular sediment; at 7 per cent growth was poor with very
fine granular sediment. No growth at 8-12 per cent, hence the limiting concen-
tration of alcohol lies between 7 and 8 per cent ethyl alcohol.

Two hundred ml. portions of Budweiser beer (4 per cent alcohol) were
measured into flasks under aseptic conditions, and one set inoculated with Myro-

1947]

DODGE MYCOCANDIDA RIBOFLAVINA 33

boflaiimi^

The flasks were weighed

daily. In 6 weeks, 2 gm. loss of weight was recorded, with a decrease of alcohol
to 1.1 per cent (volume) and an increase in non-volatile organic acids measured.

Had the organism been a species of Brettanomyces, an increase In ethyl alcohol and
volatile organic acids would have been obtained.

TEMPERATURE RELATIONS

Week-old cultures in malt extract {15° Balling) were placed in a water bath
at the desired temperature for half -hour and hour Intervals. Good growth on
subsequent plating occurred at 48'' C. for a half hour and at all lower temperatures.
No growth occurred at 49° C. nor above to 58° C after half -hour exposures.
About 20 colonies per plate were found on plates from cultures exposed to 48"
for one hour; hence we conclude that the thermal death point is 49° for a half-
hour exposure, and the maximum temperature for growth is 47-48° C. While
no extensive experiments were made to determine optimum temperature for growth,
such experience as we have had Indicates that the optimum Is about 3 C.

RELATIONSHIPS

At first examination of young cultures, our organism might be taken for
Kloeckera (PsenJosaccharomyces) y as the cells are rather elongate with acute ends,
but not truly apiculate. Little or no pseudomycellum has ever been reported in
Kloeckera, while our organism predominantly produces pseudomycellum in old
cultures, placing It In the Eremascaceae Imperfectae. Brettanomyces, when grown
on potato agar, resembles our organism In morphology, but fails to liquefy gelatin,
produces sufficient volatile acid in malt extract to kill the cultures quickly (unless
calcium carbonate Is added to the medium), and ferments sugars very slowly.
Our organism liquefies gelatin slowly, produces no volatile acid under similar con-
ditions, and ferments sugars more rapidly if they be fermented at all.

Both morphologically and physiologically, our organism belongs in Mycocandida
Langeron & Talice\ based on the type species Candida viortifera Redaelli^. Our
organism differs from the type species in its fine granular sediment rather than
flocculent growth with little sediment, liquefying gelatin more slowly and fer-
menting fewer sugars but growing better in those it does not ferment. Our
organism has longer, slenderer ellipsoidal cells in malt and other media, while
M. mortifera has nearly spherical ellipsoid cells.

Our organism resembles Candida Gnilliermondi (Cast.) Langeron & Guerra^ In
its fermentative ability, but the latter forms a pellicle on liquid media, fails to
liquefy gelatin, renders litmus milk alkaline, fails to assimilate lactose, and has
larger cells and pseudomycelium. Although some strains approach the morphology

^Langeron, M., et Talice, R. V. Nouvelles methodcs d'etudc et essal dc classification dcs
champignons levuriformes. Ann. Parasitol. Hum. Comp. 10:1-80. 5 ph. 1932.

^Redaclli, P. I miceti come associazione microbica nella tuberculosi polmonare cavitaria. Osser-

vazioni micopatologiclie e spcrlmcntali. pp. 21-24. Pavia, 1925.

^Langeron, M., et P. Guerra. Nouvelles rechcrches de zymologie medlcale. Ann. Parasitol. Hum.

Comp. 16:429-476. pi. I2-24. 1938.

[Vol. 34, 1947]

34 ANNALS OF THE MISSOURI BOTANICAL GARDEN

of Mycocandida on some media, In general it has the type of pseudomycelium of
Syrifigospora (Mycotorula).

Our organism differs from strain 488 (isolated by L. J. Wickersham from
sour milk and used by P. R. Burkholder'* In patent 2,363,227 under the name
Candida Gtiillicrmondia) in producing as good growth In lactose as in mannose
and sucrose. Candida Guilliermondi (Cast.) Lang. & Guerra produced no growth
on lactose and maltose while C. Gnilliermondia Burkholder produced better growth
on maltose than on glucose and only slight growth on lactose.

Our organism was thought to be an aberrant strain of Saccharomyces fragilis
when first isolated and might be considered an asporogenous variant. It has the
same general morphology when first transferred to fresh media, but is smaller, the
cell walls are not fragile, and it has produced no ascospores by the usual techniques.
It liquefies gelatin much more slowly and fails to ferment lactose. I have found
no report of S. fragilis producing pseudomycelium in old cultures.

In view of the above considerations, our organism is an undescrlbed species of

the Ercmascaceae Imperfectae for which I propose the name Mycocandida ribo-
flavina.

In conclusion, I wish to express appreciation to Anheuser-Busch for a research
grant and permission to study this organism; to Dr. George W. Freiberg for helpful
suggestions and photostats of pertinent literature not available in local libraries;
to Dr. J. E. McClary for determinations of ethyl alcohol and volatile and non-
volatile acids; to Dr. Lilian Nagel for the drawings illustrating morphology; and

ool of Botanv of Wash

University, for kindly interest in this study.

^Burkholder, P. R. Fermentation process for the production of riboflavin (vitamin B2). U. S.
Patent Office 2,363,227:1-3. 1944.

Explanation of Plate

PLATE 7

Fig. 1, Cells from 3 -day agar culture at room temperature.

Fig. 2, Cells from dried-out colony on agar plate, soaked in distilled water and
stained with aceto-orceln. 10 X ocular, 90 X objective.

Ann. Mo. Bot. Gakd., Vol. 34, 1947

Plate 7

e

($

<?>

@

C3>

1

Q>

U --

o

o

o

\

X ^ *

o

C>

a

t

J

10^

DODGE— MYCOCANDIDA RIBOFLAVINA

[Vol. 34. 19 1"

36 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PLATE 8

Fig. 3, Unusual types of cells from old lactose fermentation tube. Free-Kand sketches.

Fig. 4. Pseudomycelium and cells from moist colony on agar slant, stained with aceto-
orcein. 10 X ocular, 90 X objective.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 8

a?

\:f^

L

i iOft

&

t \

L" *

4 4

* «

h-.

* X

I* *

» *

'-.-

I'

F' 1

_ k

*»*.

1^

r* ^*

* -J

* ^1

m

^ »^j

DODGE— MYCOCANDIDA RIBOFLAVINA

INHERITANCE IN THE CARNATION (DIANTHUS CARYOPHYLLUS)

III. Inheritance of Flower Color

GUSTAV A. L. MEHLQUIST

Horticulturist, Missouri Botanical Garden
Associate Professor, Henry Shaw School of Botany of Washbigton University

(Formerly of the University of California, Los Angeles)

AND T. A. GEISSMAN

*

Associate Professor of Chetnistryy University of California, Los Angeles

Introduction

4

This carnation study was begun by the senior author in 1932 at the University
of Connecticut, and although only about 3,500 plants were grown there, that
represented much of the work necessary before larger populations could profitably
be grown. ^ In 193 6 the study was transferred to the University of California
where, between 1936 and 1942, 35,000 to 40,000 plants were grown. Since 1946
the work has been continued at the Missouri Botanical Garden,

The purpose of this study was partly to produce superior carnation varieties;
especially in the yellow group where good commercial varieties have always been
scarce, and partly to learn the genetical basis for some of the characteristics which
contribute to the make-up of a good commercial variety. As the project expanded
it was found necessary to limit the study to one or two major characteristics.
Since a pleasing flower color is one of the primary requirements of any plant grown
for ornamental purposes, this feature was gradually given preference, while otbers
were given attention only as they appeared in the cultures grown for color analysis.
Although the study Is by no means complete, it seems justifiable to report the
data obtained to date, as it may be some time before the work, interrupted by the
war, can be resumed on full scale.

Material and Methods

The carnation material available during the first season consisted of ten com-
mercial varieties, or clones, namely: arctic, betty lou, fairy queen, ivory,

MAINE SUNSHINE, MATCHLESS, PINK ABUNDANCE, SPECTRUM, SURPRISE, and

WOBURN. A few Others were added during the next two years. Since carnation
varieties of this type are ordinarily reproduced by cuttings, they were expected to
be rather highly heterozygous. In order to get an idea of the degree of hetero-
zygosity and at the same time make a start toward the production of relatively
pure lines, self-pollination of these varieties was immediately undertaken. How-
ever, one variety (pink abundance) produced no pollen whatever during the
entire season, and two varieties (spectrum and ivory) failed to set any seed
whether self-pollinated or cross-pollinated. On the whole, selfing proved to be
difficult and produced relatively few viable seeds per capsule. Crosses, on the

I

■^The senior author is indebted to Professors R. H. Patch, G. S. Torrey, A. S. Porter, and S. P.
HolHster for their kind interest in the project while it was carried on at the University of Con-
necticut, It was through their combined efforts that the necessary facilities were provided.

(39)

[Vol. 34

40 ANNALS OF THE MISSOURI BOTANICAL GARDEN

other hand, resulted in fair amounts of good seed and were easily made. Those
varieties which produced Uttle or no pollen proved to be among the best seed-
producers when cross-pollinated. In the generations following these crosses many
plants were eventually obtained that were reasonably self-fertile and could be in-
bred until relatively pure lines were established. Whenever a line of twenty-four
or more seedlings failed to show any segregation for the characteristic being
studied, the line was considered homozygous for the corresponding gene. This
number Is obtained by solving the equation 1 — (%)"=: .999, where n is the
number of self seedlings that must be grown from a plant to indicate with a
probability of ,999 whether a plant that does not segregate is homozygous for
the genes under investigation.

It was later found that varieties which could not ordinarily be selfed during
the winter In the greenhouse, either due to lack of pollen or because of failure to
set seeds, could be selfed with at least a fair amount of success if they were grown
in the field during the summer and in the fall transferred to rather small pots and
placed in the greenhouse. Even varieties which when benched, as is ordinarily
done with this type of carnation, produced no pollen, with pot culture produced
at least a few anthers and set good seeds. If the nitrate level was kept fairly low
and the plants held rather on the dry side, this partial fertility often lasted well
into the winter.

The seed was germinated in the greenhouse, the bulk of it in sterilized sand or
soiL The seeds of the most important lines, and those which for some reason were
poorly developed, were germinated on blotters and transferred to soil shortly after
germination. Regardless of which method was used, most of the seedlings were
transferred to 2-inch pots or 2-inch plant bands and later planted out in the field.
A few were transferred to 4-inch pots and flowered In the greenhouse. Whether
grown In the greenhouse or in the field, the progenies from crosses generally
bloomed in from five to eight months whereas those from selfed lines were decided-
ly more Irregular, requiring from five to fifteen months from planting of seed to
flowering.

The chromosome number was determined from root-tip material on over 100
different plants. The 2n number was 30 except for occasional tetraplold

root

or sectors. Meiosis has been studied only in some 30 plants, all of which showed
15 blvalents at IM. Included in these 30 were 4 female sterile, 4 male sterile and
3 which ordinarily failed to produce seeds because of the prevalence of secondary
ovaries. All underwent regular meiosis, w ^ 15 is the x number for the
genus Diantbus (Darlington, '45). All observations arc based on permanent prep-
arations that were stained according to StockweU's safranln-crystal violet schedule
(Stockwell, '34).

In recording flower color, the names used in commercial carnation culture were
retained; but new colors were given descriptive names.

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 41

Results

To facilitate tlie analysis, the colors of the carnation have been divided into
four main groups, namely:

I. The acyanic group, containing only those colors that are due to antho-

xanthins^. These colors are pale yellow, clear sulphur-yellow and white.

II. The cyanic group, in which the colors are due to anthocyanins on ivory

' base. This group contains two distinct series depending on whether the

anthocyanin involved is pelargonidin or cyanidin. Each of these scries
may again be divided into two sub-series depending on whether the antho-
cyanin occurs as a monoglycoside or as a diglycoside,

a. Pelargonidin monoglycoside colors: salmon (eleanor, charm); red

(spectrum, king cardinal, TOM KNIPE, WM. SIM) .

aa. Pelargonidin diglycoside colors: light pink (Virginia); deep pink

(pink abundance, boston ward, JOHN briry).

b, Cyanidin monoglycoside colors: lavender-pink (no commercial) ;
crimson (woburn, topsy, seth parker).

bb. Cyanidin diglycoside colors: lavender-pink (no commercial); purple
(royal purple, potentate).

IIL The transition group in which the color is due to partial development of
anthocyanin on yellow base. This group contains the salmon-yellow,
orange, salmon-orange and pale maroons. Some of these self colors may
be variegated with anthocyanin, in which case they are specifically dis-
cussed in the next group.

IV. The variegated group containing all those types in which either acyanic
or cyanic colors occur In stripes or zones on lighter background. Five
types of variegations will be discussed as follows:

a. random narrow,

b. random broad.

c. picotee pattern.

d. salmon-red.

The fifth type of variegation, flushed, because of its more natural relation-
ship to the self colors, will be discussed in connection with the acyanic

group

L The Acyanic Group

a. Yellow versus White.

Most of the F^^ progenies from crosses between white and yellow have been
either pure white or white lightly striped with anthocyanin color, but some have
been anthocyanin self-colored. In Table I are summarized the results from those
in which the Fj^ were white or white-variegated. As variegation is discussed sep-
arately, only the self colors are considered here. The results indicate that two
independent genes govern development of the yellow and white colors respectively,

^"Anthoxanthin" is a rather general term applied to sap-pigments other than those of the
anthocyanin type. It refers in most cases to flavone derivatives.

I Vol. 34

42

ANNALS OF THE MISSOURI BOTANICAL GARDEN

TABLE I

>

(

-

'4:

*

O

PROGENY

»

-

RA r

*

TOTAL

RATIO

PARENTAGE

^

1

P

■

r ^

1

White
var. pir

White

Yellow

Pale
Yellow

1

1

*

MAINE SUNSHINE*, yellow

p,

63

15

68

3:1

.20

34006-2, yellow

V,

17

7

24

3:1

34518-1-14*, yellow

Pi

24

24

38192-14, yellow

p,

17

27

\

38ir;8-l, pale yellow

Pi

I

21

21

w

37054-6, white

Pi

*>h

37079-29, white

p.

30

30

37109-1, white

Pi

23

23

■

1

38594 — 34518-1-14x37054-6

F,

30

*

30

Two plants

F,

95

85

41

12

233

12:3:1***

.65

■

38626 — 34006-2 x 37109-1

F,

13

14

17

1:1

Two pi., white var. D. P.

F»

4s :

30

19

6

103

12:3:1

.95

One plant, white

Fa

1

66

20

5

1

91

12:3:1

.75

38628— M. S. X 37109-1

F,

9

5

i

14

One plant, white

F.

61

24

85

3:1

.45

One plant, white

F,

42

11

53

3:1

.45

4

40522 — 38192-14 x 37109-1

F,

27

18

45

1:1

Three pi. white var, D. P.

F,

69

94

38

12

213

1

12:3:1

.85

40576 — 38168-1 x 37079-29

F,

26

26

One plant

F.

17

7

24

3:1

* MAINE SUNSHINE at timcs had occasional broad, white stripes and faint, narrow pink stripes
34518-1-14 had faint narrow reddish stripes.
** 37054-6 was female sterile, hence no Pi population.
***The white and white variegated pink have been added.

and tfiat the gene for white is epistatic to the one for yellow. Because the so-
called whites are really ivory-colored, at least in the bud stage or until bleached
in sunlight, the gene controlling the development of this color has been designated
L The gene for full yellow color has been designated Y. Thus YI and y1 are
white, Y i yellow and yi pale yellow.

The whites used as parents in the crosses summarized in Table T, with one ex-
ception, were pure ivory-white on which no red or pink marks had ever been
observed. The one exception, 37079-29, in the greenhouse during the short days
of winter at times had a faint tinge of pink. Under field conditions it had pure
white petals with tinted anthers. The four yellow parents, on the other hand, reg-
ularly produced a few reddish or pinkish stripes. Some of the yellow F2 plants
also had some reddish or pinkish stripes but in the field they were so indistinct that

no accurate scoring could be made for this feature. The crosses in which the F^
progenies were anthocyanin-colored are summarized in Table IV.

19-171

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

43

TABLE II

RATION

PROGFX^

r

TOTAL

RATIO

1

PARENTAGE

*

*

P

1

GENK

A-color

White*

1

Yellow
md Or£

1

MAINE SUNSHIWE, ycllow

34520-6, red

1

Pi
p,

1
157 red

f 15p.y.
( 63 yel.

1

78
157

3:1

.20

3 5009-5. red

Pi

j 3 8 red
( 12 salmon

50

3:1

.85

34520-6-12, red

Pi

3 5 red

35

37117-37, light pink

Pi

23 1. pink

23

33002-3, deep pink

Pi

j 25 d. pink
I 8 1. pink

33

3:1

.90

37531 — 34520-6xM.S.

F,

27 d. pink

27

1

Three plants, deep pink
38558 35009-5 x M. S.

F.
F,

1

52

J 14 d. pink
I 12 l.pink

28

24

1
1

1

104
26

27:21:16
1:1

.25
.65

One plant, deep pink

F,

57

20

19

96

9:3:4

.40

One plant, light pink

F.

39

20

17

76

27:21:16

.25

38596 — 37117-37 x M. S.

F,

4

13 1. pink

13

Two planes

F,

90

78

62

230

27:21:16

.60

38619 — 33002-3 x M.S.

F,

] 7 d. pink
I 3 1. pink

¥

10

One plant, deep pink

F.

71

1

23

26

' 120

9:3:4

.65

38637— M. S. X 34520-6-12

F,

1

2 3 d. pink

1

23

One plant

F.

164

59

73

296

9:3:4

.85

Two plants

F,

152

115

75

342

27:21:16

.40

* The column for A-coIor includes salmon, red, light pink, and deep pink.
**The column for white includes white variegated red or pink.

aa. Yellow versus Antfjocyanin,

I

In Tables II and III are summarized the data from crosses between yellow and
anthocyanin self-colored plants. The F2 results conform to two different ratios,
the 9:3:4 and the 27:21:16, indicating segregation for two and three genes re-
spectively. It should be noted that in the crosses where segregation occurred
according to the 27:21:16 ratio the yellow parents (maine sunshine and
34006-2) were heterozygous for pale yellow, and that segregation according to
the 9:3:4 ratio occurred in the same crosses. The pale yellow parent, 37039-14
in cross 38 583, was a segregate from selfing 34006-2. Both plants selfed from
this cross gave segregation according to the 27:21:16 ratio. On the other hand,
the orange-yellow, 35003-1 (34518-1-1) and the yellow 34518-1-14 (Table
VIII), were both from lines in which no pale yellow plants have ever been re-
corded. The seven F^ plants that were selfed from these crosses all segregated
according to the 9:3:4 ratio. Furthermore, the composition of the orange and
yellow groups differed according to the nature of the segregation types. When-
ever the segregation ratio was 9:3:4 the orange and yellow group was composed of

[Vol. 34

44

ANNALS OF THE MISSOURI BOTANICAL GARDEN

TABLE III

^

c

IM^OCFNY

PARENTAGE

<

»

1^

c

TOTAL

RATIO

P

1

Up

*

oO

1

1

w

U
1

<

— t3

37039-14, pale ycHow

Pi

1

15 p.y.

15

34006-2, yellow

Pi

f 17 yel.
I 7 p.y.

1

24

3:1

.60

34518-1-14, yellow

Pi

1
1

24 yel.

24

35003-1, orangc-ycllow

Pi

1

1

1

1
1

\ 10 or.
I 3 yel.

13

i

34520-6-12, red

Pi

3 5 red

1

35

34520-6-13, red

Pi

40 red

40

35019-1, light pink

Fi

j 13 I. pink
/ 3 salmon

16

3:1

.60

33002-3, deep pink

Pi

) 25 d. pink
) 8 1. pink

^K ^v ^t

33

3:1

.90

38559 — 35019-1 x 35003-1

F,

3 7 d. pmk

1 4 red

11

1

Two plants, red

F,

] 44 red

I 1 8 salmon

1
1

20

29

111

9:3:4

.90

Three plants, deep pink

Fj

127

49

48

224

9:3:4

.30

4

3 8564 — 34518-1-14 x 34520-6-12

hi

12 red

12

Two plants

F,

152 red

37

67

256

9:3:4

.20

38605 — 34518-1-14 x 3 3002-3

F,

14 d. pink

14

Two plants

F,

149

43

63

255

9:3:4

.70

38550 — 34006-2 x 33002-3

Fi

\ 44 d. pink
(9 1. pink

53

3:1

.18

One plant, deep pink

F,

22

19

7

48

27:21:16

.20

One plant, deep pink

F.

39

n

20

70

9:3:4

.65

One plant

F,

51

28

25

104

1

27:21:16

.30

3 85 83 — 34520-6-13 x 3703 9-14

F,

22 d. pink

1

22

Two plants

F,.

282

204

140

626

27:21:16

.40

* The column for A-color includes salmon, red, light pink and deep pink,
**The column for white includes white variegated red or pink. ■

only two types, namely, orange and yellow In the proportions of 3 orange to 1
yellow. Thus a more complete ratio for this type of segregation may be written:
9 A-colored: 3 white: 3 orange: 1 yellow. When, on the other hand, segregation
occurred according to the 27:21:16 ratio the orange and yellow group consisted
of three different types of individuals, namely, orange, yellow, arid pate yellow in
proportions approximating 9 orange : 3 yellow : 4 pale yellow.

These, results suggest that the gene determining segregation cither according to

4

the 9:3:4 or the 27:21:16 ratios in this case is a member of the Y-y pair. Tliat is,
those Fi plants that segregated according to the 9:3:4 ratio were homozygous for
the y gene, whereas those that segregated according to the 27:21:16 ratio were
heterozygous for this gene. The third gene involved may be assumed to be a basic
anthocyanin gene A, acting with the genes Y and / to produce normal anthocyanln
color. The data on variegations (Section IV) indicate that no anthocyanin color

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 45

whatever is produced in the presence of its allele a^ but that certain variegation
patterns are possible with an intermediate allele a"^^^^.

The interaction of these three gene pairs, all of which are necessary for full
production of anthocyanin color, may be represented thus:

Y 7 A

y i a

17 Y J A

9 Y I a

9 y I A

3 y I a

9 Y i A

5 Y i a

3 y i A

1 y i a

A-coIored
white
white
whit

27 A-coIored

2 1 white

* *

transition group
yellow
pale yellow
pale yellow

16 yellow, orange,

maroon

In Table IV are summarized the data from the crosses between white and
yellow in which the Fj^ progenies were A-coIored. On the basis of the genotypes
suggested these data should conform to the 9:3:4 and 27:21:16 ratios. Although
the progenies from these crosses are rather small, the segregations conform to these
requirements.

TABLE IV

O

k

<

Z:

PROGENY

TOTAL

RATIO

■

PARENl'AGE

A-color*

*

IE

Yellow

und Orange

P

37079-21, white
37079-29, white
37109-1, white
38192-14, yellow
34520-17-35-12, salm. yellow

MAINF. SUNSHINE, yelloW

39578 — 37109-1 x salm. ycl.
One plant

40552 — 37079-21 x salm. ycl.
One plant

40553 —37079-21 x 38192-14
Two plants

40526 38192-14 x 37079-29
Three plants

40584 M.S. X 37079-29
Two plants, white
Two plants, deep pink

Pi
Pi
Pi
Pi
Pi

Pi

F.

F2

Fi

F2

Ft
F^

Fi

F2
F,

7 L p.

3.2 ,

1
1

26 d. p.

50

26 d. p.
65

28 d. p.
89

13 d. p.

38

20 white
3 white
23 white

1
1

i
■1

1

29
19
39

85

7 white
52
29

27 yellow
23 s. yel.
f 63 yellow
( 15 p. yel.

29

20

33
51

14 p. ycl.
23

20
30

23

27

23

78

7
90

26

26

137

28
225

20
66

27:21:16

9:3:4
27:21:16
27:21:16

3:1

27:21:16

.20
.25

.75

.40

.30

.45
.85

*>!■

^ The column for A-color includes salmon, red, light pink, and deep pink.
The column for white includes white and white-variegated. *

b. Wtnte versus AntJx)cyamn,

In Tables V and VI are summarized the results from crossing white with A-

[Vol. 34

46

ANNALS OF THE MISSOURI BOTANICAL GARX)EN

TABLE V

o

PROGENY

PAKKNTACK

H

< !

W
W

Pi

A-color

Group

1

02

White

TOTAL

RATIO

P

^^^^^ ^^M ^^ ^" ^ ^ ^^ ■

Deep

pink

Light
pink 1

1

1

1

Salmon

IVORY, white

1

10

10

37054-6, white

»:■

1

37079-29, white

Pi

30

30

34520-17-16, salmon

P.

31

31

34520-6-13. red

P.

40

1
1

40

1

37117-37, light pink

Pt

23

1
1

23

t

38578 — 34520-6-13 x 37054-6

Ft

18

1
1

1

18

Three plants

F,

239

68

262

569

9:7"'*

1

.25

38579 — 34520-6-13 x IVORY

Fi

13

10

!

23

1

1:1

1

.50

One plant, deep pink
One plant, red

F2

F.

52

10

92

54
63

116
155

9:7

; 9:7

1

.50

.40

58617 — 37117-37 x 37054-6

F,

13

13

1

1

Two plants

F,

102

24

92

218

9:7

.60

40531 = 34520-6-15 x 37079-29

F.

20

20

Two plants

F,

96

24

48

168

3:1 !

1

.25

40569 — 34520-17-16 X 37079-29

Fi

26

26

1

Three plants

F,

70

26

22

11

52

181

1

.20

* 37054-6 was female sterile.
**All the A-color groups have been added together.

TABLE VI

H ■ ^

1

PROGENY

1

PARKN lACiE

<

<U C

c

C/2

White
var. A-col.

1

>

TOTAL

RATTO

1
1

r

34520-6-12, red

34520-6-13, red

34520-17-35-2, salmon
37109-1. white

Pi

Pi

Pi

1

1
1

r

35
40

28

23

3 5

40
28

23

1

38580 = 34520-6-13 x 37109-1
Two plants, red

Two plants, deep pink

F.
F,

1

1

12
93

42

13

' 202

20

5

48
49

94
97

25
344
306

1:1
9:7
9:7*

.80

.35
.20

38624 — 37109-1 x 37117-37
Two plants, li^ht pink

Two plants, deep pink

Ft

F2

Fa

14
68

5

lOl

21

w

28
30

48
43

19

177
162

9:7
9:7

.80
.75

38624 — 37109-1 x 34520-6-12 |

One plant, red

One plant, deep pink

One plant, deep pink

Ft

F2
F.

F.

13

33

27

14

12

41
12
10

11
3

11
11

20

1

18
16
14

25
81
72
88

1:1
9:7
9:7
9:7

.80
.15
.25

.35

* All the A-color have been added together against white and whice-varlegated

19471

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 47

u

colored. In the crosses 38578, 38579, 385'80, 38624, 38625, and 38617, in which
the white parents involved were pure white, never having shown any anthocyanin
color whatever, all the F2 populations grown conform to the 9:7 ratio, indicating
segregation for two independent genes. The results from crosses 40531 and

40569, on the other hand, indicate a single gene difference. The white parent
(37079-29) involved in these two crosses was occasionally slightly flushed with
pink. It is the same plant that was discussed in connection with Table I.

In the process of purifying many of the original A-colored lines by selfing,
numerous small progenies were obtained which segregated for white in the propor-
tions of 3 A-colored to 1 white. Furthermore, many crosses were made between
a number of whites selected from crosses 38578 and 38579 (Table V). These
Fl progenies contained all possible combinations, namely, all white, 3 white to 1
A-colored, 1 white to 1 A-colored, or all A-colored.

In most of the crosses between pure white and full A-color, between yellow
and full A-color, and between yellow and white that resulted in full anthocyanin
color, some of the whites occasionally were tinted pink or red and in some, whose
products indicated segregation for both y and a, a goodly number of the progeny
were strongly flushed pink or red on white background. One plant (37079-29)
that occasionally produced a faint tinge of color in the petals has already been
discussed In connection with Tables I and V. This plant, when crossed to two
different yellow plants, produced colored F^ progenies (Table IV) which in the
next generation (Fm) segregated for white and yellow in 27:21:16 proportions;
that 15, segregation by three genes. On the other hand, when it was crossed to
pale yellow (Table I), the result was a white Fi and segregation only for pale
yellow in the second generation in proportions indicating segregation by only one
gene. Likewise, when crossed to homozygous salmon and red, the F2 results Indi-
cated segregation by one gene (Table V). The only genotype possible that would
account for these results is y J A,

As already stated, many crosses were made among whites selected from the F2
generations from crosses 38578 and 38579 (Table V), Several of these plants,
including some that were lightly tinted, were crossed to 37079-29 and some of
its self-seedlings. In every case tinged selections, wKen mated to 37079-29 or its
self-seed!ings, produced only tinged progeny. On the other hand, the same selec-
tions produced colored progeny when mated to certain pure whites with which
37079-29 also produced colored progenies, suggesting that the tinge or flush of
color was inherent in the ;y-gene or some allele to it. As different whites of
known genotypes gradually became available, numerous crosses were made In
order to test this hypothesis. The results (Table XVI, p. 60) bear out the
hypothesis that the tinged and flushed plants belong to the ^f-whites.

As may be seen In figs. 2-4 of plate 9, the anthocyanin in the flushed indi-
viduals varies not only in amount but also in distribution. In matings between
strongly flushed plants and near- whites the colors of the F^ generations usually
were intermediate, but sometimes they were stronger than in either parent. How-

48

IVOL. 34

MISSOURI

ever, as it has not yet been possible to grow such progenies under controlled con-
ditions In the greenhouse, it Is not known whether this increased color was due to
the genotype or the environment.

In the early stages of this study, when many lines were inbred in order to
provide homozygous plants, numerous lines were obtained whose segregations
indicated that white flushed with anthocyanin is a simple recessive to full self-color
and a simple dominant to that type of whites which produce a slight tint or flush
of color on!y under favorable conditions. The monogenic relationship between
white-flushed anthocyanin and the corresponding self-color is further demon-
strated by the crosses between flushed and variegated individuals (Tables XVII
and XIX) .

On the basis of the results obtained so far, it can be said that the lowest allele
of y that has been obtained to date, may with 7 and A produce a faint tinge or
flush of anthocyanin on the petals of the flower. The anthers and the tips of the
stigmas arc usually faintly colored in this type, even when the petals are white.
The intensity of color varies with the specific genotype and the environmental
conditions. Usually j-whites with R can be distinguished from r plants but
whether a plant has the dominant allele of S or M cannot be determined except by
breeding tests. The gene for flushing has been designated y^ Probably different

mod

expression

The occurrence of white variously striped with A-coIor in many of the crosses

is discussed in Section IV.

II. The Cyanic Group

a. Pelargonidin MonoglycosiJe Colors.

In 193 3 a red seedling, which, because of sparse pollen production, could not
profitably be selfed, was pollinated by spectrum supreme, a commercial red
variety which only rarely sets seeds (due to the prevalence of secondary ovaries)
but usually produces good pollen. All Fi plants were red (Table VII). Three of
the four F^ plants that were selfed segregated in the proportions of 9 red : 3 sal-
mon : 3 salmon-orange : 1 salmon-yellow, while the fourth did not segregate. In
the F3 one red plant again segregated in this manner, another red segregated for
salmon in the proportions of 3 red : 1 salmon, while the third red did not segregate.
Of the three salmon plants selfed in this generation two segregated for salmon-
yellow In the proportions of 3:1, while the third bred true, as did also the only
salmon-yellow plant selfed. In the F4 the salmons either segregated for salmon-
yellow or bred true. The two salmon-yellows that were selfed bred true.

These results, together with those from the crosses 38610, 39525 and 39583
summarized at the bottom of Table VII, clearly demonstrate the difference in one
gene between red and salmon, red and salmon-orange, salmon and salmon-yellow,
salmon-orange and salmon-yellow, but a difference of two genes between red and
yellow. The presence of orange and yellow indicated segregation for the / gene

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 49

-

TABLE VII

L
L

^

J

:neration

PROGKNY

TOTAL

1

RATIO

PARENTAGE

T^

UOUI

1

^

Ho

P

Sal

c/}

00 >.

33511-3, red

No Pi

3

3

SPECTRUM SUPREME, red

No Pi

1

t

i

1
1

38187-10, salmon

Pi

37

1

37

1

\

34520 33511-3 xspectrum supreme

Fi

23

23

34520-3, red

F2

10

8

4

2

24

9:3:3:1

.25

34520-6, red

F.

157

157

34520-10, red

F2

83

24

31

8

146

9:3:3:1

.80

34520-17, red

F2

125

43

28

12

208

9:3:3:1

.25

34520-17-2, salmon

F3

32

12

44

3:1

.70

34520-17-5, red

Fs

40

13

53

3:1

.90

34520-17-6, red

F3

43

19

12

4

78

9:3:3:1

.60

34520-17-8, red

Fa

79

79

34520-17-16, salmon

F3

i

1
1

31

31

n

34520-17-19. salm. yel.

F3

1

1
i

12

12

34520-17-35, salmon

F3

1

26

10

36

3:1

.70

34520-17-35-1, salmon

F4

1
1

25

9

34

3:1

,80

34520-17-35-2, salmon

F4

28

28

34520-17-35-12, salm. yel.

F4

23

23

34520-17-35-31, salm. yel,

F4

27

27

38610 — 34520-6 x 34520-17-35

Fi

24

24

Two plants

F2

193

61

254

3:1

.70

Fix 34520-17-35-2

BC

64

55 ,

119

1:1 '

.45

Fi X 34520-17-35-12

BC

27

31

1

58

1:1

.60

39525— -17-35-1 x 17-35-12

BC

1

29

26

55

1:1

.65

395 83—34520-6 x 38189-10

Fi

12

12

Three plants

F2

162

62

224

3:1

.50

(see under III). Chemical determinations have shown the anthocyanin in both
red and salmon to be a monoglycoside of pelargonidin, but in different concentra-
tions (Geissman and Mehlquist, '47). The genes corresponding to these different
concentrations have been designated S and 5 respectively. Red, or scarlet as this
color often is called in commercial carnation culture, may thus be designated by
the genotype Y I AS while salmon would be Y I A s.

Results from a similar cross are summarized in Table VIII. The F2 segrega-
tions here are in the same proportions as those just discussed, but one of the genes
involved is different. The presence of yellow and orange-yellow again indicates
segregation for the / gene. The presence of white but absence of pale yellow in-
dicates segregation for a gene of the A locus. All the whites from this cross had
from one to many narrow red stripes. The yellows were at first recorded as pure

redd

No such

stripes were ever found in the orange-flowered group. When

tnpe

observed (cross 38 564^ Table IX) ail the whites in the F2 had occasional red

50

[Vol.. 34

ANNALS

MISSOURI BOTANICAL GARDEN

TABLF

VITT

1

ION

PROGENY

1

1

1

pakkxta(;e

<

Red

White
var. red

Orange

Yellow
var. red

TOTAL

RATIt)

r

3 3 506-3. red

NoP,

1

33514-M, red

No Pi

1

ri

34520-6, red

P,

157

157

345IS 3350/>-3 x 33514-1 1

F,

14

14

Two plants^ red

F,

137

57

38

12

244

9:3:3:1

.20

Three plants, red

F,

170

(^(^

236

3:1

,30

34518-1-!, or.iagc

F.

10

3

13

3:1

34518-I-I2, wliite var. red

Fa

1
1

26

8

34

3:1

.80

34518-1-13, white var, rod

F.,

1

IS '

S

26

3:1

.50

34518-1-14, yclh)w var.

Fh

24

1

24

34518-9-2, white var. red

F.

3 8

38

34518-9-2 X 34518-1-1

F4

18

17

35

1:1

34518-1-14 \ 34518-9-2

F^

2*

2

1

^ 1 from above cross

F=

65

1

20

85

3:1

.75

#2 from above cross

F-,

1

69

1

23

93

3:1

,95

^2 red mutant

F,

51

23

74

3:1

.20

3451S-1-12 X 345I8-I-13

F,

7*

2

9

3:1

One plant frnm above

F.

42

22

64

3:1

1

' .07

Red mutant

F.

48

18

. 1

66

3:1

.60

38546 — 34518-1 x 34520-6

F,

55

1

55

3 8546-5, red

F.

J

49

15

13

6

S3

9:3:3:1

.85

38546-6, red

F. 1

65 .

15

80

3:1

.20

One plant of each of these lots produced red-flowered branches which were vegetativcly propagated
and then self-pollinated.

Stripes; most of the yellows had faint reddish stripes; but none of the oranee was
ever recorded as having them. For reasons discussed under section IV this gene
for white with red stripes must be considered an allele in the A~a series.

As in cross 34520 (Table VII) red differs from yellow in two genes whereas
there is a single gene difference between red and white, red and orange-yellow,
orange-yellow and yellow, and white and yellow.

The red of cross 34518 was somewhat duller or more toward the salmon-red
hue than the red from cross 34520. When crosses were made between reds from
these different families the F^ plants were always dull red and in the Fo generations
the deeper red of the 34520 line reappeared. However, adverse w^eather condi-
tions made accurate classification difficult. Somewhat less than one-fourth of the
progeny was classified as deep red, and of the remainder some were distinctly dull
red and many appeared to be intermediate. Lately a still deeper red has appeared
in one line derived from the cross 38579 (Table V). Again, this red totaled
about one-fourth, whereas the remainder was apparently all the kind just discussed
as deep red. For the purpose of reference, the red from cross 34520 has been des-
ignated "standard" red, while the dull red, deep red, and any other red that might
be met with in future work will be measured against this standard.

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

51

■a

^X^hen the salmon-orange from 34520 was crossed to the orange from 34518
the Fi was orange and the salmon-orange reappeared in the F2 to the extent of
about one-fourth of the total. When yellow from 34518 was crossed to salmon-
yellow from 34520, the F^ was orange and the F2 was approximately 9 orange
to 7 yellow. The orange here contained orange, salmon-orange and what appeared
to be intermediate shades. Likewise, the yellow group contained both clear

yellow and salmon-yellow.

The single gene difference between dull red and standard red, between standard
red and deep red, as well as between orange and salmon-orange might be due
either to different alleles of the S gene or to an independent modifying gene
determining the intensity of the anthocyanin. However, when crosses were made
between various derivatives of crosses 34518 and 34520 (Table IX) all the F^
were dull red and the F2 included not only dull red and deep red but also salmon.
From these observations it must be concluded that the varying shades of red are
not due to multiple alleles of the S gene but rather to an independent modifying
gene influencing the concentration of the anthocyanin. Further work is necessary

TABLE IX

rARENTACE

34520
34520
34520
34520

34520
34518
34518
34518
38565

6-12, red
17-16, salmon

17-19, salm. yel.

17-35, salmon
17-35-12, salm. yel.
1-12, white var. red
1-14, yellow var. red
1-17, orange
2, white var. red

38566 = 34518-1-14 x 34520-17-35
Two plants, red
Two plants, orange

38574 = 34520-17-16 x 34518-1-12
Two plants, red
Six plants, red

3 8574::= 34520-17-16 X 34518-1-17
Two plants, red
One plant, red

3 85 64 = 34518-1-14 x 34520-6-12
Two plants, red

39304 = 39520-17-19 x 3 85 65-2

395 54 = 34520-17-3 5 X 3 8 565-2

o

H

<

W

t3

a:

PROGENY

o

B

1>

>

Pi

35

p.

30

p.

Pi

26

■

Pi

Pi

26

Pi

NoP,

Pi

Fi

13

Fj

83

23

F,

Fi

24

F2

86

29

F^

142

50

Fi

15

F«

110

32

F.

40

9

Fi

12

F2

152

F,

32

Fi

39

70

17

29
60

12

37

O

3

30

85

it *

42**

14

(>7

«■ )>

30

33

o

12

10

23

8

24*^

23

22

TOTAL

RAT 1

35

30
12

36
23
34
24

92

16

168

52

24
144
337

15

184

75

12
256

62

72

* These yellows were lightly variegated red.
**The field conditions did not permit accurate separation of yellow from orange.

3:1

3:1
27:9:12:16

9'.7

9:3:4
27:9:12:16

9:3:4
27:9:12:16

9:3:4

1:1

1:1

P

.93

.60
.60

.80

.50
.95

.60
.30

.20

80

.95

52

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

before the gene or genes causing these differences can be properly identified.

The results, summarized in Table IX, in all other respects confirm the con-
clusions based on the data from Tables VII and VIII.

aa, Pelargonidhi Diglycoside Colors.

In Table X arc summarized the results of the crosses made between red and
deep pink, red and light pink, Ught pink and deep pink, and salmon and deep
pink. One of the crosses between red and salmon from Table VII is included for
comparison. It is evident that deep pink and light pink differ frotn red and salmon
respectively In one gene and that deep pink diflfers from salmon in two genes, the
salmon being the double recessive while deep pink is the double dominant.

Chemical determinations have shown that the deep pink and light pink are due
to a pelargonidin which is not a monoglycoside, as was red and salmon, but
diglycoside. Thus the gene that differentiates deep pink from red and light pink

TABLE X

4

GENERATION

■

PRO(

:;eny

TOTAL

RATIO

PARKXTACE

J

4J c

iJ'S. [

05

C

o
S

C/1

P

33002-3, deep pink
37117-37, light pink
34520-6-12, red
34520-6-13, red
34520-17-16, salmon
37010-1-12, salmon

p.
Pi
p.
p,
p,
p.

70

27

35

40

30
24

70
27
35
40
30
24

1

1

38610 34520-6-13 x 34520-17-16
Two plants

Fi X salmon parent

1

1

1

24

193
91

1
1

1

61
88

24
254
179

3:1
1:1

.70

.70

39583 34520-6-13 x 37010-1-12

Three plants

Fi X salmon parent

F,

13

173

81

63

77

13
236
158

3:1
1:1

.50
.75

38609 34520-6-13 x 34002-3
Six plants

Fi X red parent

F,
BC

47

99

87

39
90

47
138
177

3:1
1:1

.35

.80

38621 37117-37 x 34002-3

Three plants

Fi X light pink parent

F:

F.
BC

}7
84
63

29
59

1

37
113
122

3:1
1:1

.85
.70

38620 33002-3 x 34520-17-16
Two plants

Fi X salmon parent

F,

F,
BC

14
88

129

31

115

1

26
110

12
92

14

157
446

9:3:3:1
1 :1 :I :1

.90
.30

1

38597 37n7-37x 34520-6-12
Four plants

Fi x salmon (34520-17-16)

F,

F.

BC

19

70
39

19

44

20
38

7

43

19
116
164

9:3:3:1
1:1:1:1

.80

.85

38622 34520-6-12 x 37117-37

Two plants

Fi X salmon (34520-17-16)

F,
F,

BC

25

100
77

35
57

38

78

11
56

25

184
268

9:3:3:1
1:1:1:1

.90
.08

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

53

from salmon apparently does so by causing the development of a diglycoside in-
stead of a monoglycoside. This gene has been designated M. Then the genotype
of deep pink is Y I ASM, light pink Y I A s M, red Y I ASm and salmon
Y I A s m. The diglycosidic anthocyanin apparently is less stable than the cor-
responding monoglycoside, for in strong sunlight deep pink and light pink bleach
much more than red and salmon. In fact, under California field conditions, the
light pinks often bleach to almost white whereas the salmons retain their color

fairly well.

The same differences in intensity of color noted for the reds and salmons
obtain in the deep pinks and light pinks. In all probability, the same genes are
responsible for the differences in both series of colors,

b. Cyanidin Monoglycoside Colors.

Table XI gives the results of crossing red with crimson. Unfortunately,
neither of the crimson plants used as parents was homozygous for crimson but
the fact that the F2 progenies contain variegated individuals as well as crimson
and red does not obscure the monogenic relationship between these two colors.
Only one cross between salmon and crimson is available so far. The crimson was

roon

salmon-yellow. As shown in Table XII, the F^ consisted of 21 crimson and 4

TABLE XI

GENERATION

PROGENY

TOTAL

1

RATIO

PARENTAGE

Crimson

r

1
1

, Maroon

var. crimson

Orange
var. red

1

P

34520-6-13, red

Pi

40

40

37107-2, crimson

Pi"

1

37107-3, crimson

Pi

17

8

25

3:1

.45

1

37107-3-9, maroon var. crimson

Pi

19

6

25

3:1

.80

37107-3-20, crimson

Pi

39

11

50

3:1

.60

3 7107-3-24, crimson

Pi

29

11

40

3:1

.70

\

38581 —34520-6-13 x 37107-2

Fi

13

11

24

1:1

38581-13, crimson

F2

62

22

29

7

120

9:3:3:1

.45

38581-21, crimson

F.

105

35

140

9:3:3:1

1.00

38581-22, red

Fs

75

75

38581-23, crimson

F2

115

38

50

12

215

9:3:3:1

.50

38581-13 X red parent

BC

26

29

55

1:1

.65

38581-21 X red parent

BC

26

22

48

1:1

.50

38582 — 34520-6-13 x 37107-3

Fi

14

14

38582-2, crimson

F2

72

19

91

3:1

.35

38582-8, crimson

■

¥2

152

39

45

10

246

9:3:3:1

.12

1

39516 34520-6-13 x 37107-3-9

Fi

16

16

Two plants, crimson

F,

64

20

20

6

110

9:3:3:1

.90

39516-1 X var]e£:atcd parent

BC

39

11

37

9

96

3:1:3:1

.75

* Complete Pi segregation for 37107-2 was 35 crimson, 8 red, 8 maroon var. crimson, 2 orange
var. red, 10 lavender, 2 salmon.

54

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

X

-I

<

y.

W

Oh

o

H

K

.J
<

O
(-1

;)>iinMQ

K 00

• *

*^ r^

r^, rn

^O Tt" OC tX l/^
f--. r^ rj fs* fN

O 0\

p-m

uouijut^

^J3

"O ir^ 00
c4 rsi fs*

japUOAT?''!

IIOOJKXV

UOOJEJV'

oc o\

UOSlUji^

K

O O -^ CM ^

^O

* •

■ i

* «

ON

ON

v.-

t *

OS

» •

t ■

ir\ ^ O so

r^ ^ oo oo ir\

— K

00 0\

•^

rj (M

K

K

"«*- so

K IT, K

fv) **S OO ^H

N0IXVHHN3i)

W

a,

ft^ p- c- cT C-.

c* «

M

tL, tu ph u, y

l/N

o

I
K

E -

c

=1 s

E E

>

r4

t4

C/5

1

I
K

t

I

I

CM

I

I

I

c

o

to

E

I

n

E

I

::; 5

s

I

K

r4

m

K

K

00

oo

I

ir\

I

K

oo

oo

il

>

C

o

o

E

I

K

OO

I

I

I

r*-i

I

IN

OO
OO

l/\ lA

B ■

ON

■ *

K
Csl

ON O

K

ON r-i

K

fM

K

<S| NO OO

Csl SO
ON

T1 n ?t

U- P^ u.

I

I

I

lA

I

I

C

o

E

v^

C

o

E

c

n3
>

C

c

o

oo

e

£

o

O

o

CD
>

k-.

c
2

o

V)

G
O

E

H

bO

C

o

V

«>

C

E

2

e

d

G

(«

u
u

VI

g

3

a

«>

H

3

G
ft

I

s

C 3

5 5^
S .Si

-^3

4>

O o

« o

<u

A *

*

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

55

maroon-variegated-cnmson. Only two crimsons and one maroon-variegatcd-
crimson were selfed. As both the maroons and the variegated types are members
of the transition group, only crimson, red and lavender need to be considered
here. Although the proportions of lavender and salmon are somewhat too small,

f

the reasonably good fit to a 9:3:3:1 ratio suggests that two pairs of independent
genes are involved. The back-cross 39552, although small, supports this hypoth-
esis. Since the genotypes for red and salmon are respectively Y 1 AS and Y I As,
the genotype for crimson and lavender may be written Y I ASK and Y I A s R,
the gene for crimson being designated by R, 'When three lavender plants from
this cross were selfed, they segregated for salmon in the proportions of 3 lavender
to 1 salmon, and when lavender was crossed to red the Fj result was crimson.
This is just what would be expected on the basis of the genotype suggested.

The anthocyanin in both the crimsons and the lavenders has proved to be a
monoglycoside of cyanidin. The function of the 21 gene then apparently is the
production of cyanidin to the exclusion of pelargonidln, v/hereas in the presence
of r pelargonidin only is produced.

hb. Cyanidin Di^lycoside Colors.

When a crimson that was heterozygous for maroon-variegated-crimson was
crossed to a homozygous deep pink the F^ generation was magenta-purple (Table
XIII). The anthocyanin present in this magenta-purple proved to be a diglycosidc
of cyanidin. Thus, the gene Ai introduced through the deep pink parent functions
also here as a modifying gene concerned with the development of the correspond-
ing diglycoside. The independence of M with respect to R and S is clearly shown
in Table XIII. The only genotype left in this series which has not been accounted
for is Y I A s RM, This was produced by crossing light pink Y I A s r M with
lavender-pink Y I A s R vi. The F^^ appeared to be slightly paler than the lav-
ender-pink parent but in the Fo generation it was impossible, by inspection, to
separate accurately the plants having M from those having the recessive allele vt,
but chemically they proved quite distinct. All plants with the gene M contained

TABLE XITI

parknta(;f

33002-3, deep pink
34520-6-12. red
37107-3, crimson

38603 = 37107-3 x33002
38603-2, purple

38603-8, purple
38603-9, purple
Total for -8 and -9
38603-8 X 34520-6-12

-3

TION

PROGENY

<

o ^

TO-

X

w

V

C
O

1) '^

TAL

'p.
:3

rim;

13

Maro
var, c

c *-

6^

a

Ph

^

Q a 1

0^ ,

1

70

70

Pi

+

35

35

Pi

17

1

8

25

Fi

9

9

Fa

23

9

6

2

11

3

54

F2

143

46

35

12

236

F2

26

6

13

3

48

F2

169

52

48

15

i

284

BC

70

79

66

85

1

300

RATIO

3:1

27:9:9:3:12:4

9:3:3:1

9:3:3:1

9:3:3:1

1:1:1:1

I

.45

.90
.30
.25
.65
.25

[Vol.. 34

56 ANNALS OF THE MISSOURI BOTANICAL GARDEN

a diglycoside while those with its recessive allele m contained the corresponding
monoglycoside.

III. The Transition Group

The results summarized in Table VII show that salmon-yellow differs from
red in two genes, but only In one gene from either salmon or orange. One of
these genes must be /', as otherwise yellow could not be expressed since / has been
shown to be epistatic to Y, The other gene must be Sy since this yellow could be
obtained as a segregate by selfing salmon heterozygous for /. The genotype of this
yellow then must be Y/ A.sv, and since salmon has already been shown to be Y / A s,
the only genotype possible for salmon-orange is Y lAS, On the basis of these
genotypes all segregations shown In Table VII are possible.

The yellow In Table VIII likewise differs from red in two genes. For the

reasons stated in the preceding paragraph one of these genes must be /". The other
could be an allele of A since segregation also took place for white, or near-white,
but no pale yellow; it might also be a new gene. However, when white-variegated
red plants from this source were crossed to plants known to h(t Y I a or yl a the
Fi were always white-varlegated red or white-varlegated deep pink, but when

they were crossed to yl A plants the Fi progenies were fully anthocyanin-colored
and segregated In the F.j In the proportions of 9 A-colorcd : 3 whitc-variegated :
4 white. This second gene then must be a member of the A-a series. The geno-
type for this yellow might tentatively be represented thus: Y i a^'^^^ S,

When this yellow was crossed to the salmon-yellow from 34520 the resulting
F^ was Intermediate between the salmon-orange of 34520 and the orange-yellow
of 34518 but more like the hitter. The F2 consisted of apparently 9 orange to 7
yellow. The orange group contained orange-yellow, salmon-orange, and what
appeared to be Intermediate shades. The yellow group likewise contained both
yellow and salmon-yellow. Most of the clear yellows had faint reddish stripes but
none were found on any of the salmon-yellows or on any member of the orange

groups.

Chemical determinations made on diffeicnt salmon-oranges and orange-yellows

showed that the color In both groups was due largely to a non-anthocyanin sub-
stance plus a small amount of anthocyanin probably of the pelargonidia groups.
However, it has not yet been possible to determine whether or not the difference
between these groups is due to a difference in concentration of one or both
pigments.

When salmon-yellow and orange were obtained as segregates from crimson and
purple (Table XII), segregation for two other members of the transition group,
maroon and pale maroon, also occurred in proportions suggesting a ratio of 9
maroon : 3 pale maroon : 3 orange : 1 yellow. On subsequent selfing some of the
maroons repeated this segregation, but pale maroon and orange, on selfing, either
bred true or segregated for salmon-yellow only. The genotype of the maroon
must therefore be Y / A S R, and the pale maroon Y i AsK. Chemical determina-
tions have shown these colors to be due to a combination of anthocyanin, probably
of the cyanidin type, and a non-anthocyanic substance.

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

57

Wh

be determined except by genetical tests. The amovint of anthocyanin is evidently

bet

mined by inspection.

IV. The Variegated Group

a. Random Narrow Variegation

The first type of variegation to appear in these studies was that shown in pi. 9,
figs. 5 and 6. We have termed it rando^n narrow because of the narrow, well-
defined stripes which arc more or less randomly distributed, although they some-
times tend to be concentrated toward the distal ends of the petals. Variegated
lines show considerable variation in the amount of striping, from an average of
less than 1 stripe per petal up to as many as 20 or more. Occasionally a whole
petal or even a whole flower is colored. The color of the stripes Is determined by
the genotype of the self-colored normal type from which segregation takes place;
that is, if this type of variegation segregates from a red-flowered line the stripes
are red, from a deep pink line the stripes are deep pink, and so on. Variegation of
this type has been obtained from all of the anthocyanin colors. The background
color IS ordinarily white though It may be yellow. If yellow, the stripes are usually
so faint that they often escape attention unless the flowers are carefully examined.

Whenever this type of variegation has segregated from normal self-color the
proportions have always been such as to indicate a rnonogenlc difference between
variegation and self-color. All individuals variegated on white ground that have
been selfed have either bred true or segregated for pure white, or yellow faintly
striped with the same anthocyanin color or one recessive to it (see Table VIII).
The results from crossing plants with this type of variegation with plants of
known genotypes are shown in Table XIV.

Although the Fo data from the crosses listed in Table XIV are as yet very
meagre, they do support the hypothesis alluded to in sections lb and Ila, namely,
that this type of variegation is due to a gene which is allelic to the A-a pair. That

is:

A

= full color, a r^ pure white; while a^^^^ permits the development of fully
colored narrow stripes of anthocyanin on white background, or, in conjunction
with i, faintly colored stripes on yellow background. The monogenic relationship
between full self-color and white-variegated is definitely demonstrated by the
crosses summarized in Table VIII.

Apparently different alleles of aT^^ exist, or the expression of this gene is
modified by other genes, for through selection it has been possible to select lines of
white-variegated that differ only in the amount of variegation. When such lines
have been intercrossed the Fj generations have generally been intermediate, but in
the Fo generations the variegation range sometimes exceeded that of both parents.
That is, in the F2 from a cross between heavily variegated and lightly variegated
the range was extended from very lightly to very heavily variegated. This increase

mi

[Vol.. 34

58

ANNALS OF THE MISSOURI BOTANICAL GARDEN

i

TABLE XIV

CROSS

RESULTS

Known

Unknown

.

Color

Genotype

White

vartcgaccd

red

3<

: White

ylASrm

Red

»»

•ff

X

ylASRm

Crimson

i

ft

M

a

»*

ylASrM

Deep pink

»f

»

s

1*

ylaSrm

White var. red

>¥

ff»

^

*»

YlaSRm

" " crimson

tt

>l

x

»t

YlaSrM

" deep

pink

»f

• »

n

>>

ylaSrm

" red

• I

*>

X

»»

ylaSrM

" deep

pink

±

t*

X

; Orange

YiASrm

Red

t»

»•

3C

"

YiASrM

Deep pink

9»

4

ff

X

. Salmon-yellow

Y/Asrm

Red

t»

I*

X

: Yellow

YiaSrm

White var. red

»

»

X

»

YiaSrM

" deep

pink

>l

»t

X

: Pale yellow

yJaSrM

" deep

pink

Yellow

♦

f*

X

. White

ylASrm

Red

»f

9 1

ff

X

I*

• YlaSrm

White var. red

•>

f*

X

Orange

YiASrm

Orange

pf

*»

X

Yellow

YiaSrm

Yellow var. red

fft

ff

X

Pale yellow

yiaSrM

" pink

* These yellow-variegated

striped with red. All
segregates from red.

red were only faintly variegated but the Fi with YlaSrm was quite well
the yellow-varicgatcd-rcd plants that were used in these crosses were

4

*

TARLE

XV

CROSS

1

1

Know a

RESUr.lS

Unknown

I

■

Color

Genotype

!

n

Orange

variegated

red

X

Red

YIASrm

1

Red

•>

99

99

X

Orange

YiASrm

Orange var. red

>»

99

99

X

Maroon

YiASRm

Maroon var. crimson

fi

99

99

X

Yellow

YiaSrM

Orange var. deep pink

>»

11

99

X

Pale yellow

yiaSrM

Orange var. deep pink

Maroon

99

crimson

X

Red

YIASrm

Crimson

f>

f >

99

X

Orange

YiASrm

Maroon var. crimson

9f

' 99

99

K

White var.

crim.

Yla'^'^^SRm

Crimson

Yellow

1*

white

f

X

Yellow

YiaSrM .

Yellow var. white

»»

19

it

X

Orange

YiASrm

Oranj^e var. deep pink

>i

99

99

X

Maroon

YiASRm

Maroon var. purple

»»

99

99

X

Crimson

YIASRm

Purple

Orange

99

red

X

White

1

ylASrm

Red

tt

99

99

X

II

ylASRm

Crimson

>f

it

91

X

91

YlaSrm

Red

tv

IV

11

X

11

YlaSrM

Deep pink

tt

tl

11

X

11

yJaSrM

Deep pink

VI

11

11

X

White var.

red

Yla^'^^Srm

Red

Yellow

11

white

X

11

ylASrm

Deep Pink

19

99

11

X

99

YlaSrm

White

19

99

91

X

11

^

yJaSrM

White

19

19

91

X

Yellow var.

red

Yia-^'^^Srm

Yellow vir. white and red

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 59

tion, or it might be the result of other genes modifying the expression of the
variegation gene.

Fi generations between homozygous white-variegated and pure white (a-white)
have always been white-variegated. The Hmited F2 generations that have been
grown so far from such crosses have indicated segregation for one or two genes
although usually there tends to be an excess of whites. This excess is probably
due to the bleaching of the anthocyanin stripes under field conditions. At any
rate, plants that have been classified in the field as pure white sometimes proved
to be variegated when transferred to the greenhouse during the fall and winter.
Two such plants selfed in the greenhouse segregated for pure white, so it must be

assumed that they actually were of the genotype.

The crosses summarized in Table I are of interest in this connection. The
plant 37054-6 was a pure white that had remained so under all conditions in or
out of the greenhouse. When it was crossed to a yellow faintly variegated pink
the Fj^ (30 plants) was white with deep pink stripes. In the F2 generation it was
impossible to separate definitely the variegated and non-variegated in the yellow
group but in the white group considerable care was taken to check the plants from
time to time in order to ascertain the exact proportion of variegated individuals.
The final count of 95 variegated to 8 5 non-variegated indicates segregation for
two genes giving a 9:7 ratio.

The plant 37054-6, on the basis of its behavior in other crosses (see Table V),

must be assumed to be of the -T-^-rrg^riotype. The other plant (3 7075- 14) was

ylaSrM

Y i a^"'' Srm

V^^^^^^ ' Therefore, with respect to variegation we should expect from this

cross the following genotypes: 9 Y iif^^^ : 3 Ytf : 3 ya^^^ : 1 ya,o{ which only the
first should be variegated.

Another pure white plant 37109-1 was crossed to a pure yellow, probably of

the genotype tt"- * The result (40522, Table I) was 13 white-variegated and

Y I a

14 white. As in the previous cross it was impossible to classify the yellows in the
F2 generation into variegated and non-variegated plants, but in the white group
from selfing two variegated plants, 48 were classified as variegated and 30 as non-
variegated, again indicating segregation for two genes. The one non-variegated
plant that was selfed produced non-variegated plants only. From other

crosses it had been established that the most likely genotype for 37109-1 was

y I a^'^^ S r M

— 1; — • ■ : The results obtained from this cross are In agreement with these

y I a s r m

genotypes. When this white was crossed to another yellow which, as far as can
be ascertained, was also of the genotype Y/<?, the Fi contained 27 lightly varie-
gated to 18 non-variegated. In the F2 there was a considerable deficiency in the
variegated group which in all probability was due to bleaching so that some
lightly variegated plants were classified as white. When the same white was

[Vol. 34

60

ANNALS OF THE MISSOURI BOTANICAL GARDEN

crossed to homozygous red (crosses 38580 and 28625, Tabic VI) lightly varie-
gated individuals again occurred in the F2, but, as may be seen from the table,

the variegated proportion is less than expected. The same occurred when maine
SUNSHINE, a commercial yellow variety, was crossed to red (cross 3 8637, Table
II) or to a white of the y-typc resulting in a deep pink F^ (cross 50584, Table
IV). In cither case lightly variegated individuals occurred in the F^ generation
but in somewhat smaller proportions than would be expected on the basis of the
genotype suggested.

b. Random Broad V arte gat ton.

This type of variegation (pi. 10, fig. 4) was first met with in crosses involving
the commercial yellow carnation maine sunshine. This variety, although gen-
erally classified as a self-colored yellow, occasionally produces faint pink stripes

such as described under IVa. It was therefore no surprise to find individuals In

the F2 with narrow stripes of full anthocyanin color on yellow or white ground.
However, in addition many individuals were obtained with randomly distributed
stripes that were much broader and less definitely delimited than in the randoiti
narrow variegation described above. The color In this type of variegation ranges
from yellow striped with white up to maroon striped with purple. Thus this
type of variegation is limited to the transition series. Now, since all the members
of this series are / /, it seemed logical to assume that the gene responsible for this
type of variegation may have been a multiple allele of the J-/ scries.

The results of crosses between members of this variegation series and plants of
known genotypes are shown in Table XV. Although the number of F^ popula-
tions raised to date from the crosses listed in Table XV are few, the results indicate
that each member diflfers In one gene only from the corresponding self-colors.
That is, maroon-var.-crimson behaves In a simple recessive with respect to crimson
but as a simple dominant to maroon; orange-var.-red bears a similar relation-

TABLE XVI

CROSS

1

^

Known

RESULTS

Unknown

Color

Genotype

White

flushed pink* :

1^ White

YlaSrm

Red or deep pink

>»

YlaSrM

Deep pink

>>

YJaSRm

Crimson or purple

»

ylASrm

White flushed pink

»»

ylASrM

» » i«

r ''

ylASRm

" purple

»

yJaSrm

" pink

9*

ylaSrM

n »» >>

+

K YelL

3W

Yia

Red or deep pink

K Pale

yellow

yia

White flushed pink

* It is difficult to distinguish between m and M types in this group. Except in heavily flushed
individuals red and pink flush gives the same appearance. Some of the flushed plants used here
had M, others m.

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

61

ship to red and orange.
so far showed a correspo

been obtained
The cross to

j-white gave full self color. The results obtained (Tables XI, XII, XIII and XV)
are all compatible with the hypothesis that this type of variegation is due to a
gene multiple allelomorphic to the T-i series. That is, 1= full color; i^^^ = broad
random variegation; i := self -color of the transition series.

The most interesting cross in this group is one between yellow broadly varie-
gated white and yellow faintly variegated narrow red. Seven of the 17 F^ plants
were yellow faintly variegated red, but the other 10 had both broad white stripes
and narrow pink stripes. Furthermore, where the two types of variegation over-
lapped (that is, where the narrow stripes overlapped the broad) the narrow stripes
were of a bright deep pink color; but when the narrow stripes were between the
white stripes (that is, on yellow ground) they were as faint as in the parent from
which they were introduced. Thus it is evident that wherever the white stripes
do occur the conditions are the same as if the whole flower had been I instead of

:var

c. Picotee Pattern,

This variegation pattern (pi. 10, figs. 1-3) appeared in an F2 population from
a cross between a commercial crimson (woburn) and a commercial white
(matchless) of the ^/-series. The Fj contained only 12 plants of which 3 were

TABLE XVII

\

TION

PROGENY

PARENTAGE

T\

TOTAL

RATIO

P

GENE

Self-
color

White

var. re<

1

White

flushed

red

White

57050-6, white flushed red

Pi

33

33

37079-18, white*

Pi

1

26

26

3703 0-16, white var. red rand. nar.

Pi

27 \

q

27 ,

b

37078-11, " " " " "

Pi

29

29

3 8201-4, " " " ••

P.

1

23

23

40529— 38201-4 X 37030-6

F,

23

1
!

23

Three planes, red

F,

150

58

70

278

9:3:4

.60

40532 — 37030-6X 37078-11

F,

8 !

8

Two plants, red

Fi

53

21

38

t

112

9:3:4

.05

40536 — 37030-16 x 37030-6

F,

19

19

Three plants, red

F,

' 89

26

33

148

9:3:4

1

.60

40548 — 37079-18 x 3 820N4

Fi

22

22

Three plants, deep pink

F.

119

38

22

29

208

9:3:4**

.65

40550 37079-18 x 37078-11

Fi

25

25

Three plants, deep pink

F2

73

27

9

31

140

9:3:4"^*

.45

* In the field this plane was pure white but under favorable conditions in the greenhouse the petals
would show an occasional flush of anthocyanin.
**The white and white-flushed were added.

[Vol. 34

62

ANNALS OF THE MISSOURI BOTANICAL GARDEN

TABLE XVIII

PARENTAGE

PROGENY

TOTAL

PINK MATCHLESS, dccp pink

WOBURN, crimson

3453 5 =: VOB. x P. MATCH.

34535-2, purple

34535-4, white var. purple

34535-4-6, white var. deep pink

345 3 5-4-1 2, white var. crimson

Pi
P.

F,

F,

Fa

None

163 self-colored, 3 9 whitc-varlcgatcd, 5 8 transition color
7 self-colored, 2 white variegated, 3 transition color

148 self-colored, 44 white-variegated, 49 transition color
1 8 white var.purp. or crim.,4 wh. var. red or pink, 1 1 wh,
47 white var. d. p.. all with picotee pattern, 7 white
72 white var. crimson, all with picotec pattern

260
12

241
33
54
77

purple, 4 deep pink, 3 maroon-broadly-variegated-purple, 1 pale lilac-variegated-
purple, and 1 white-variegatcd-pink. It is from the pale lilac-variegated-purple
(Table XVIII) tliat all the lines with this pattern on whitish ground have been

derived.

This pattern occurs in all variations of intensity from the faintest suggestion
to the deeply colored shown in fig. 1, pi. 10, When the pattern is strong either
in extension or intensity of color, the background also becomes Hghtly colored.
That IS, if the pattern is red or deep pink the otherwise white ground becomes
faintly colored pink, and if the pattern is crimson or purple the ground becomes
pale lilac. Under field conditions this ground color often bleaches to white but in
the greenhouse it usually remains. On clear yellow ground the pattern is very
faint, often limited to a pale edge at the distal ends of the petals. The same pat-
tern occurs in the transition series (fig. 3, plate 10), but here it appears to be
made up of broader stripes and blotches than when it occurs on whitish ground.
Because most of the plants with this pattern have also had stripes typical of either
the r^^^ or a'^^^ variegations, it was thought that perhaps this pattern was only
expressed in i^^*' and a'^'^*' genotypes and that the apparently ''pure" picotee pattern

TABLE XIX

RATION

PROGENY

TOTAL

1

RATIO

PARENTAGE

1

1

P

GENE

Self-
color

White
var.

White
flushed

40534—34535-4-12-1 x 37030-6

F.

26 cr.

26

Two plants, crimson

F,

127

45

80

252

9:3:4

.04

405 35 — 3703 0-6 X 3453 5-4-12-2

Fi

21 cr.

%

21

■

Two plants, crimson

Fa

73

27

41

141

9:3:4

.45

40546 — 37079-18 x 34535-4-12-1

Fi

24 purp.

24

Four plants, purple

Fa

154

57

95

306

9:3:4

.04

40580 — 345 35-4-12-1x37079-29

Fi

26 purp.

M

26

Two plants, purple

Fa

93

37

39

169

9:3:4

.50

40581 —34535-4-12-2 x 37079-29

Fi

29 purp.

29

1

Two plants, purple

F2

97

41

47

185

1

9:3:4

.50

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 63

in reality was due to a relatively "low" allele of F^^ or tf^^^ with a "high" allele
of the gene determining the picotee pattern.

In Table XIX is summarized the data from crosses between picotee pattern
and 3/- whites. It is apparent that, with the exception of the crosses 40534 and
40546, the segregation from full self-color is the same as if the genes in question
were y^^ and ^^^ The results from the crosses 40534 and 40546 do not agree too
well with the hypothesis but it was noted that the self-colored and white-flushed
plants from these crosses were, on the average, much more vigorous than the
variegated plants. The reason for this difference in vigor is not known. Among
the variegated individuals there were some that appeared to be picotee only, others
that were rando^m narrow, while the majority showed both types of variegation.
If one considers those that appeared to have only random narrow variegation
against the remainder, the proportions for the five crosses are: 7:38, 6:21, 10:47,
8:29 and 10:31, or 41:166 for all of them, which is approximately % of the total.

The Fi plants of each of the crosses between picotee pattern and random
narrow variegation showed both types of variegation (Table XX). In the F2
generation there was segregation for both patterns. Since a heavy picotee pattern
might mask the stripes of the other variegation pattern, it is safer to consider
those having only random^ narrow variegation. By so doing it becomes evident
that in four of the six crosses this variegation occurred in about ^4 ^^ ^^^ total
number of plants. On the other hand, in the other crosses (40538 and 40540),
only one plant of the four that were selfed segregated for random variegation. The
plant 37078-11 that was used as one parent in these crosses came from a line in
which weak picotee patterns had been observed and, although this plant had been
classified as having random variegation only, it is possible that it also had a weak
picotee pattern. It was not possible to check on this as the plant was no longer
available when the difference among these crosses became apparent.

A white-flowered plant (39024-26) obtained from a lavender line segregating
for white (40 lavender; 11 white) was crossed to a pure-breeding white whose
genotype had been determined to be Y I a Sr m. The result was 15 F^ plants, all

of which were variegated crimson on white ground, 3 with random narrow varie-
gation, and 12 with both random and picotee patterns. Two F^ plants that
plainly showed both types of variegation were selfed. In the F2 generation of 96
plants, 60 were variegated while 3 6 were white. Of the 60 variegated plants, 12
were classified as having picotee pattern, 18 random variegation, and 30 with both
random and picotee. The proportion of 60 variegated to 36 non-variegated sug-
gests a 9:7 ratio or segregation for two genes probably y and a. If this is correct
the white extracted from the lavender line must have been of the genotype

yla^^^ so that the F^ plants were — ■- . This genotype would account for

the segregation of variegated and white in approximately 9:7 proportions. The
ratio between all the plants showing the picotee and those with the random type
variegation is 42 to 18, suggesting that the F^^ plants were heterozygous for a
dominant gene capable of producing the picotee pattern only in the presence of

64

ANNALS OF THE MISSOURI BOTANICAL GARDEN

I Vol. 34

PARK NT AGE

40537 — 37030-16'- x 3453 5-4-12-2'='^
40537-11

40537-16

40538 = 37030-16 x 3 4 53 5-4-12-3

40538-2

40538-4

40540 = 37078-11 x 34535-4-12-1

40540-6

40540-9

40542 = 37078-11 \ 3 4535-4-12-1

40542-8

40542-10

40577 = 34518-1-14**=' x 34535-4-12-1

40577-9
40577-19

40578 = 345 18-1-14 X 34535-4-12-2

40578-2

40578-8

TABLE XX

*For
**For

**3f

For

Pi data on
Pi data on

P] data on

o

<

on

c

Fi
F.

F.

F,

F,

Fi

F^
F,

Fi
F-.
F.

F,

Fa

F.

F,
F,
F.

PROGENY

White var. crimson

20

18
17

57
10

26
15

7
20

1
2

+

a

25
26
58

10
18

Z2

13
5
8

11
26

12

26

10
10

21

22

27

37030-16 and 37078-11 see Table XVII.

34535
34518

4-12-1, 12 and 13 see Table XVIII.
1-14 see Tabic VIII.

d

u

10
18

12
7

12

5
5

8
6

White var. red

u

1

2

1

13
16

4
(2

8

6

6

14

3

7

5

8

c

6

13

15

5

3

2
2

TOTAL

4

5

25
76

105

10

70
64

13
71

40

11
61

42

26

27

46

21
40
49

another variegation gene, in this case tf^"*".

When a plant having flowers that were orange variegated with red picotec
pattern was crossed to a white-variegated-red of the random narrow type, the Fj^
of 21 plants consisted of 1 1 red self-colored plants and 10 with white flowers
variegated red with both random and picotee patterns. No Fo generation has yet
been grown from this cross.

Much more work is needed before the exact inheritance of the picotee pattern
will be known. The best hypothesis that can be made at this time is that it is
determined by a dominant gene non-allelic with the other variegation genes dis-
cussed and capable of producing its characteristic pattern only in the presence of
either I^'"^ or ^^''^^, For purposes of identification this gene will be designated Pic.

d. Salmon-Red Variegation. —

This type of variegation was first found on salmon ground but has since oc-
curred on every member of the 5 series, that is, salm-on, fight pink, and lavender.
It is the most erratic of the different types of variegation encountered in this study.
Red variegation on salmon ground is the only color that has been studied for the
inheatance of this feature, all the data pertaining to this variegation In other
colors having been derived incidentally from crosses made for other purposes. The

/

19471

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION

65

origin of this type of variegation, as far as this study is concerned, can be traced
to the commercial variety spectrum. This variety has been found in these
studies to be heterozygous for yellow and salmon. Thus the genotype is

— . _^ Jt is of a rather dull red color. During the 20 years that it has been

Y i A s T m

widely grown, it has produced at least one mutation toward a deeper, more at-
tractive red which has largely replaced the parent variety. It is known in the
trade as spectrum supreme. A salmon mutant, also widely grown commercially
and known as salmon spectrum, has occurred several times. This salmon
mutant in turn frequently mutates back to red, but most of these mutations are
limited to a few red stripes or sectors of individual flowers only rarely involving
whole flowers. Other commercial salmon-colored varieties known to be genetically
related to spectrum, such as charm, laddie and surprise, frequently mutate to
red in the same manner (pi. 10, figs. 5, 6).

TABLE XXI

PARENTAGE

NK ABUNDANCE, dcCp plllk

SPECTRUM, red

SURPRISE, salmon
34520-6-13

3 3 503 ^SURPRISE X SPECTRUM

33503-2, salmon

33514^PINK ABUND. XSPECTRUM

33514-20, salmon var. red

34509
34509
34509
34509
34509
34509
34509
34509
34509
34509

33514 X 33503-2
3, salmon var. red
5, salmon var. red

10, salmon var. red
12, salmon var. red

11, red

14, red

10-1, salmon var. red
10-1-1, salmon var. red
10-1-2, salmon var. red

34509-10-1-2
Plant #1
Plant #2
Plant #3
Salmon from
Salmon from

X 34520-6-13

#2
#3

*The ratio is based on
** Less than .01.

O

<

NoPi
NoP,
NoP,

Pi

Pi

Fa

Fi

F,

Fi
Fj
F=
F,

Fo

F,
Fa

Fa

F4
F4

Fi
F2

F2
F2
F3
Fs

V

'J4

self-color versus variegated

PROGENY

(7) >

o

rt

xn

TOTAL

40

40

6

ri

4
23

10

23

13

11

6

30

9

3

12

6

8

14

5

66

37

108

38

11

49

30

12

42

27

24

51

32

14

46

34

9

43

16

8

24

1

27

28

1 •

36

3

40

27

1

17

37

9

46

43

8

1

52

35

16

2

53

37

14

51

1

27

11

39

RATIO

1:1

3:1'='
3:1

1:1

?

3:1
3:1

3:1
3:1
3:1
3:1

?

3
3
3
3
3

1
1
1
1
1

P

.65
.60

.40
.50

.40

.40
.20

.15
.65
.60

[Vol. 34

66

ANNALS OF THE MISSOURI BOTANICAL GARDEN

In Table XXI are shown the crosses of particular interest in connection with
this type of variegation. The cross 34509 indicates that this variegation is a sim-
ple recessive to full self-color. The other crosses show that such is the case. On
the other hand, nearly all the salmon-variegated-red plants that have been selfed
have given more salmon selfs than was expected on the basis of a single gene dif-
ference. However, some salmon plants extracted from such progenies in the next
generation produced again a majority of salmon-variegatcd-rcd plants, as if they
in reality had been salmon-variegated-red. This irregular behavior and the fact
that most of the spontaneous occurrences of this type of variegation have been
limited to a few stripes or sectors involving only one or two petals indicate that
svich variegation is due to some instability of the s gene or to some other gene
capable of causing the s allele to mutate to S. Tliat it is the s gene which mutates
is evident by the variegation being limited to the 5-serIes. In order to identify

Th

ere is no evi-

this gene for further studies it will be designated 5^"^

dence that the gene for picotee pattern, discussed in the preceding section, has any

effect on this gene.

Discussion

As far as we are aware, the only previous published data on the inheritance of
flower color in the carnation, aside from the preliminary report by the senior
author in 1939, is that of Connors ('14). From the results of a cross between a
commercial white and a commercial yellow carnation, he concluded that white
was dominant to yellow and red, and yellow in turn to red. Our results show that
he was right in concluding that white is dominant to yellow (actually epistatic)
but not as to white and yellow being dominant to red or pink. The appearance
of red or pink stripes on white or yellow flowers from selfing what was supposed
to be pure whites, in all probability, was due to mis-classification of the F^ plants.
In fact, Connors himself stated that at the end of the season the yellow parent,
JAMES wiiiTCOMB RILEY, produced somc flowers that were streaked with red.
That places this parent in the variegated class. The white parent, white per-
fection, must have been homozygous for a^ as otherwise the Fj generation would
have been anthocyanin self-colored. One of the parents must have been heterozy-
gous for y^ as otherwise no pale yellow or cream-colored individuals would have
occurred in the F2 generation. If one assumes that the yellow parent was homozy-
gous for d^"^** the results are entirely compatible with the genotypes suggested by
this study. The whites obtained in the F^ were probably lightly variegated but
grown under conditions unfavorable for the production of this variegation. Under
field conditions in California it was found necessary to check the populations sus-
pected of variegations several times during the year to be reasonably certain that
plants classified as whites were actually white.

The genotypes suggested here are in many respects similar to those suggested
for other plants. As Wheldale found in AntirrJjimim ntajtis ('10), Lawrence and
Scott-Moncrleff in Dahlia variabilis ('35), and Buxton in Primula acaulis ('32),
two genes arc concerned with the production of the anthoxanthins in the carna-

19471

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION (J

O rt

M

^

o

O

2 o

c re

OO
ON

NO

D

3

CO
ON

o

n

a.

O n
P

>

o

»-l

ft

^

r»

(A

O

o

o

00 H^

3 3^^

ft o

OS

"H,

09

I

^3

I

X

■

^ ^ -n ^

M M M ►-»

K>

-^ o

N 00

4^ OO

VI NO O

\-n ON VI

<^ On

>-tJ NO ON Wl

000-4

3 3 t3 ^

** a o Q

'^ 2 S =^

'7 3 3

5'

On

NA

II

HO HH

^ 2 ^ ^

o ** o o

^

I

I

^^\ JS

M

V

4^

I

I

-1 ft*

s

CU

3"^

o

3

VI

I

VI

f

5'

VI

I

"T] ^ ^ 'T] "rrt

M M M to t-'

"Tl "Tj "Tl *Tl

H M

NJ NO

0\

VI NO

4^

NJ1

4^

K> VI OO

NO On ^

NO

VI 4*^

K>

VI

VI

h-» V« NO

rs> v^ N^ bj

VI 4^ VI

4^

4».

4^ t^ H- 1

v CO 4h.

VI 4^ ^ VI

OO N> OO 4^

V4 o ^ O VI

Vj«» O '-^^ ^

v^ N> 4^ ^ ".^

Cn OO 4^ ^ OO

On OO 4»- ON ^^

O OO '-M NO 0\

VI V4
I I

I •
>^ ^

6a

NO
I

^

O

3

o

3

3

o

3

OO 0\

>

2

m

Vi

C

X
o

^ "^ t; ^

iNi

Sad
>

a

(JEXERATION

Deep pink

Light pink

var. d. pink

Light l)ink

Red

Sahuoii
var. red

Salmon

White
var. pink

White

O

X
X

Yellow var.
red or pink

ON
VI \^

Pale yellow
:iud yellow

K> Nj^ K> VJ

OO ON W CO

N> Ni NO
• • • • <.^j

^- I— 4^

ON ON *

\0 VX) VO ^

V4 xi xa VJ

O
H

>

>

On ^ On
wi O O

VO 4>- OO VO

O VI N/l NO

NJ

^

68

[Vol.. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

tlon, Y with yellow and I with ivory (white). As in Antirrhinum, I is cpistatic
to y, but plants with the allele y are white only in combination with /. In com-
bination with the recessive allele / they are of a pale yellow which in strong sun-
light may bleach to cream. As in Antirrhinum vm]in, VrimnJa acaulis, Tropacolum
majus (Scott-Moncrieff, '36), and Pharhitis nil (Hagiwara, '32), only one gene A
is concerned with general anthocyanin production. Plants homozygous for a in
the presence of / are pure white, as is also yJ a. Plants with yl A usually have
colored anthers, tips of stigmas, leaf bases and nodes and, under favorable condi-
tions, a trace of anthocyanin in the petals. The gene S determines the concentra-
tion of the anthocyanin, permitting full intensity, while in the presence of its
recessive allele a much smaller amount of anthocyanin is formed, resulting in a
series of pale colors. One, perhaps two, as yet unidentified dominant genes further
suppress the amount of anthocyanin. As in the China Aster (Callistcmwa
chinensts (L.) Skeels) studied by Wit ('37), the gene M controls the glycosldic

type of the anthocyanin. In all genotypes with M the number of sugar molecules
attached to the anthocyanidin molecule is two, in genotypes with in, only one.

TABLE XXIH

NUMMARY OF GENOTYPES AND I'HENOTYPES FOR SELFCOLOKED CARNATIONS

(^icnotypes

PhenotypL'S

Y J A% KM —

Maecnta-purplc

Y I AS R ni —

Crimson

Y 1 A S r U —

Deep pink

Y 1 A s KU —

Lavender

Y I A S r m —

Scarlet^ red

Y I A s R m —

Lavender*^

y / ^ 5 r U =

Light pink

Y I A % r m —

Salmon

y \ A "^"^ —

White petjls, anthocyanln-colored anthers and stigmas**"^

Y I a —

Pure white petals, white anthers and stigmas

y I a - — — —

Pure white petals, white anthers and stigmas

y / A S R M —

Maroon

Y i AS K m "

Maroon

y /■ A S r M —

Orange

y / A £ R M -

Pale maroon

y i A S r /./ —

Orange

Y i A % K m ~

Pale maroon

Y i A % T III —

Salmon-yellow

Y i A s r m =

Sahnun-yellow

Y ; a '^* —

Yellow

y i A - - ~ —

Pale yellow

y i a - — — —

Pale yellow

'^This lavender cannot be distinguished from M lavender except by breeding tests. The same is
true for maroon, pale maroon, and orange.

*** Any allele of S R M may be substituted for without change in appearance.

Under favorable conditions the petals also may be faintly flushed with anthocyanin. The kind
of anthocyanin will depend on the specific genotype but only the pink-red series with r and
the crimson-magenta series with R can be recognized by inspection. Whether the plants have
w or M cannot be determined with certainty by inspection.

* * *

19471

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 69

Also, as in Aster, the gene K determines the kind of anthocyanin. In genotypes
with R the product is cyanin, whereas with r it is pelargonin only.

The inheritance of flower variegation in the carnation needs further study.
The more or less continual outcropping of variegated individuals in crosses made
to study self-colors was at times quite a nuisance, but now that the main genes for
the self -colors are established and the connection between them and the genes for
variegation are at least partly known it will be easier to plan the required critical
crosses necessary to complete the picture. All of the genotypes possible with the
genes identified so far are listed in Tables XXIII and XXIV.

It is of interest that all of the flower color genes identified in this study ap-
parently also are concerned with the general vigor of the plants. The recessive
types have been, on the average, less vigorous than the corresponding dominants
and the multiple recessives definitely weaker than the multiple dominants.

The genes 7 and M are of particular interest in this connection. Plants with
i (that Is, yellows) and members of the transition series are usually quite deficient
in the cuticular waxy material responsible for the bloom or glaucousness of the
leaves and stems. Plants with /^^^ are generally somewhat better in this respect
but still deficient. This deficiency seems to be of relatively little consequence in
the greenhouse but out-of-doors, especially in hot and dry weather, the plants are
much harder to grow. Probably this deficiency in cuticular wax means less
protection against excessive transpiration.

By selection it has been possible to obtain i plants with so much more glaucous-
ness that they are indistinguishable from I plants in the greenhouse and do very
well under most field conditions. However, all these plants also have M. Every
selection made among ^ -m plants has been definitely inferior to the best selections
from the i M group- It would appear therefore that the dominant allele of M^ or
genes associated with it, can in part make up the deficiency in glaucousness caused
by /.

TABLE XXIV
SUMMARY OF GENOTYPES AND PHENOTYPES FOR VARIEGATED CARNATIONS

IVa. Random Narrow VarieKation

y 7 a^fl^ S R M z= White with narrow stripes of purple
y 7 a^ar S R m

Y I a^'^^ S r M

j> t> *> ij i>

y> 9» >* fy f9

crimson
deep pink

Y I a^^^ s R M = " " " " " lavender

y 7 tf^«^ S r m

J3 $f 19 >« »*

red

y 7 a^'ar s R m = " " " " " lavender

J

»> ft >J >f 97

light pink
salmon

Y / fl^ar s r M

Y I a'^'^^ s r m
y I a^(^^ = White

Y i a"^^^ S R M = Yellow with narrow stripes of purple

y i a^'^^ S R m

Y i a-^^r S r M

Y i a'^^*' s R M

Y i tf^^*" S r m

Y i a^<^r s R m = " **

y / j^'fi^ s r M

Y i i/^«^ s r m
j; ;• ^T^rtf _ „ « — Pale yellow

99 9*

I 5> >»

> » »

» »» >f

crimson
deep pink

lavender

red
lavender

light pink
salmon

70 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

IVb. Random Broad Variegation

y i^'^^ A S R M ^

Maroon with broad stripes of purple

Y i^'^r A S R m —

" " crimson

Y /^'«^ A S r M —

Orange with broad stripes of deep pink

Y i^<^r /i , j^ j^ —

Pale maroon with broad stripes of lavender

Y i'^'^^ A S r m z=i

Orange with broad stripes of red

Y ivar A s R m ~

Pale maroon with broad stripes of lavender

Y r^'^i^ A s r M zzz

Salmon-yellow with faint broad stripes of pink

1

Y i^ar ^ J r ^, —

Salmon-yellow with faint broad stripes of pink

y ivar a rr

Yellow with broad stripes of white

y i^(^r A —

Pale yellow

y fVar ^j _ „ _ ^^^

Pale yellow

IVc. Picotce Pattern-

— This pattern can presumably be superimposed on any i'^'^^-

by the gene Pic.

or tf'^f^- genotype

IVd. Salmon-Red Va

ricgation

Y I A s"^'^^ R M zrz Lavender with purple stripes

y / A s'^<^^ R m =1 Lavender with crimson stripes

y / A 5^'*^^ r M =z Light pink with deep pink stripes

y / A s^<^^ r m ^=z Salmon with red stripes

IVe. Flushed Variegation

yn I A S R M=: White flushed magenta-purple
yf^ I A S R m =

yf^ I A S r M

yf^ 1 A 5 R M ~

yf^ I A S r m

yf^ I A s R m

yfi I A s r M

yf^ I A s r m

yfl I a = White

yf^ i A — — - := Pale yellow flushed deep yellow to orange

^/^ i a — - - =: Pale yellow

crimson

deep pink
lavender

red
lavender

light pink
salmon*

Mixed types of Variegation

y s"^'^^ d^'«^ — - - rr Yellow with broad stripes of white and narrow stripes of any antho-

cyanin color depending on specific genotype.
yf^ I tf^'if - - - =: White, or white flushed with anthocyanin. depending upon relative

"strength" of the alleles,
yfl i^'<^r A — Not known.

* The "flushed" phenotypes, lavender, light pink, and salmon, cannot be distinguished except by
breeding tests.

Summary

Six independent genes for self-colors in the carnation Have been identified.
Their functions may be sumtnarized as follows:

Y controls the production of yellow anthoxanthln. It is hypostatic to f. In
the presence of the recessive allele y, only a limited amount of anthoxanthin
is developed, resulting In pale yellow or cream-colored flowers.

I controls the production of ivory-white anthoxanthin. It is epistatic to Y.
The recessive allele / permits the production of yellow anthoxanthin.

A is the basic gene for anthocyanin. It is fully effective only in combination
with Y and 7. In combination with i only a small amount of anthocyanin

1947]

MEHLQUIST & GEISSMAN INHERITANCE IN THE CARNATION 71

h

m

is produced, resulting in a series of pale colors on yellow background (the
transition series). In the presence of the recessive allele a no antliocyanin
IS produced. The interrelationship of these three genes Is shown by the
following genotypes:

27 Y I A ^ full anthocyanln self-color.
9 y I A z= white or near white.
9 Y I a :::^ pure white.
i y I a ^ pure white.

9 y i A = transition colors (small amount of anthocyanln on yellow

background).

3 Y i a ^ yellow,

3 3; i A = pale yellow.

1 y i a =^ pale yellow.

S controls the amount of anthocyanin. In the presence of its recessive allele 5
much less anthocyanin is formed. One, possibly two, as yet unidentified
genes modify the effect of S-5.

R determines the kind of anthocyanin. The dominant allele causes the pro-
duction of cyanin resulting in crimson or dark red flowers, whereas its re-
cessive allele r causes the production of pelargonin only, resulting in bright
red or scarlet flowers.

M determines the number of sugar molecules attached to the anthocyanin
molecule. With the dominant allele there are two sugar molecules at-
tached whereas in the presence of the recessive allele m only one sugar
molecule occurs.

The number of sugar molecules attached to the anthocyanin has a marked
effect on the anthocyanin. For instance, M with r changes the color from bright

■

red or scarlet to deep pink and M with K changes crimson or dark red to magenta-
purple. In general, it may be said that the addition of the second sugar molecule
has a bluing effect on the anthocyanin color. It has no visible effect on the

anthoxanthin.

At least five genes are concerned with the different types of flower variega-
tion in the carnation. Four of these appear to be multiple alleles with genes for
self-color. They are:

yf^ causes limited amounts of anthocyanin to be produced under favorable
conditions. This anthocyanin occurs as a tinge or flush on white back-
ground. This type has been termed fluslyed.

f^^^ with a causes broad, indefinite, randomly distributed stripes of Ivory
anthoxanthin on yellow ground, and with A similar stripes of anthocyanin
on colors of the transition series. This variegation has been termed randoTtt
broad.

[Vol. 34, 19471

71 ANNALS OF THE MISSOURI BOTANICAL GARDEN

gvav causes narrow, definite, randomly distributed stripes on white or yellow
background. This variegation has been termed random narrow,

s"^'^^ causes sporadic, irregular striping on any member of the 5 series (salmon,
light pink, lavender).

Pic causes a definite variegation pattern, plcoteCy in the presence of i^'"'' or
^var^ The recessive allele pic probably has no visible effect.

The results indicate that more multiple alleles of these genes concerned with
flower variegation exist, or that their action is influenced by modifying genes.

All of the genes for flower color appear to be concerned also with the general
vigor of the plants, for the reccssives were, on the average, somewhat less vigorous
than the corresponding dominants, and multiple reccssives were definitely weaker
than the multiple dominants.

The gene 7 seems also to be directly Involved in the development of the cuticu-
lar waxy material responsible for the "bloom" or glaucousness of the leaves and
stems, as plants with / are quite deficient in this respect. The gene M or genes
associated with it appears to be able partly to make up this deficiency caused by /.

Literature Cited

Buxton, B. H. (1932). Genetics of the primrose, Primula acaulis. Jour. Genet. 25:195—205.

Connors, C. H. (1914). Heredity studies with the carnation. Proc. Am. Soc. Hort. Sci. 1914:95-
100.

Darlington, C. D., and E. K. Janaki Animal (1945). Chromosome atlas of cultivated phints.
London,

Geissman, T. A., and G. A. L. Mchlquist (1947). Inheritance in the carnation Dianfhus cary-
opfjyllus, IV. The chemistry of flower color variation I. (In press),

Hagiwara, T. (1932). On the gcnetico-pliyslological studies of the colour development of flowers
in Pharbitis nil, Proc, Imp. Acad. Japan 8:54.

Lawrence, W. J. C, and R. Scott-Moncrieff (1935). The genetics and chemistry of flower colour
in Dahlia: a new theory of specific pigmentation. Jour. Genet. 30:155—226.

Mehlquist, G. A. L. (1939). Inheritance in the carnation, Dianthiis caryophyllus. T. Inheritance
of flower color. Proc, Am. Soc. Hort. Sci. 37:1019-1021.

Scott-Moncrieff, R. (1936), A biochemical survey of some Mendclian factors for flower colors.
Jour, Genet. 32:117-190.

Stockwell, Palmer (1934J. A stain for difficult plant material. Science 80:121-122.

Wheldale, M. (1910), Die Vererbung dcr Bliitcnfarbe bei Antirrhinum wajns, Zeitschr. f. Induk.
Abstam. u. Vcrerb. 3:321-3 33.

Wit, F, (1937). Contributions to the genetics of the China Aster. Genctica 19:1-104.

Explanation oi Plate

' PLATE 9
Diantbus caryophyllus

Fig. 1. Pure white.

Fig. 2. White flushed red toward center

Fig. 3. Flushed red toward edges.

Fig. 4, Evenly flushed.

Figs. 5 & 6. Kartdam narrow variegation

Ann. Mo. Boi . Card., Vol. 3 4, 1947

PiATr 9

MEHLQUIST & GEISSMAN— INHERITANCE IN THE CARNATION

74

[Vol. 34. 19^7

ANNALS OF THE MISSOURI BOTANICAL GARDEN

r

-\

EXPLANATIOX or Pl.A FE

TLATL 10

Diiiiilluis CiiV} f>[fli) Uh^

Fig. 1. SiroHi; crinisttn picofcc pnltcrn i>n white back^u^'uiul.

Fig. 2. Light l^icolcc pattern with some raiufom narrow stripes.

Fig. 3. Strang red pirofee pattern on orange background.

Fig. 4. Rauilom brihul red stripes on orange background.

Fig. 5. Salniofi-rcil variegation In left tliird of salnn)n nt)\\er (charm)

Fig. ^. lndi\'idual petals from tlo\\'er in fig. 5.

A NX. i\lo. BoT. Gakd., Vol. 34, 1947

Plate 10

MEHLQUIST & GFTSSMAN— INHERITANCE IN THE CARNATION

Annals

of the

Missouri Botanical Garden

Vol. 34

MAY, 1947 No. 2

THE EFFECT OF THE MEDIUM ON APPARENT VITAMIN-
SYNTHESIZING DEFICIENCIES OF MICROORGANISMS^

CARL C. LINDEGREN
Research Frofeswr, Henry Shaw School of Botany of Washington University

AND CAROLINE RAUT
Research Assistant, Henry Shaw School of Botany of Washington University

Beadle and Tatum*s work (*45) on biochemical mutants of Ncurospora has
resulted in wide acceptance of the view that it is relatively easy to distinguish
strains genetically capable from those genetically incapable of synthesizing vita-
mins. They discovered many mutants apparently incapable of performing specific
syntheses. The present paper reveals, however, that different synthetic media so
affect the growth of organisms that conclusive demonstration of specific deficiency
requires a much more critical study of the environment than has hitherto been
achieved. Differential growth of cultures in a synthetic medium which deviates

■

in many respects from the natural substrate may prove very useful for genctical
diagnosis but may not give reliable information concerning the synthetic activity
of the organism under normal conditions, "Without critical testing it is difficult
to say whether a given deficiency is absolute. The following data will show that
the standard test of ability to grow In the presence and inability to grow in the
absence of a given vitamin in a synthetic medium is inadequate and that the
criterion of equal increments of growth for equal additions of the vitamin may
give an ambiguous answer.

A COMPARISON OF TWO SYNTHETIC MEDIA

Table I shows the formulas of the synthetic media used by Burkholder ('43),
Hutner (unpublished), Wickerham ('46), and Beadle and Tatum ('45). Many
experiments have shown that Hutner's medium is a very different substrate for
yeast growth from Burkholder's. This was clearly revealed by growing cultures
Nos. 3 and 10 (Table II) on two batches of each medium, one to which no
pantothenate was added and the other containing 50 y of pantothenate per liter
(fig. 1). Culture No. 3 does not begin to grow in Burkholder's medium without

-^This work was supported by grants from Anheuser-Busch, Inc., The American Cancer Society,
and Washington University.

(75)

[Vol. 34

7^

ANNALS OF THE MISSOURI BOTANICAL GARDEN

I

I

9

o

8

O

s

I

Im

0Q

C

C

c
o

O

rs

z

^ d

o '^

^

1

H

o

G
O

O

u

<u

^ 2

>

^»

n 60

6 ^

W

^ 2

rt

"3

c

C
rt

c -a

c

u

D

s

o

-o

o
c

'5

n

*-*

o

o

It

rs
C

c
a.

O

O

u

o •—

6 "

OJ

-a

n

»r\

o

C

o

U

c

c

4>

o

c

o

c
o

E

u

Ml. 5

^2

G

o £

irv

C qj

o

c

C
O

1947]

LINDEGREN & RAUT VITAMIN-SYNTHESIZING DEFICIENCIES 77

added pantothenate until after 200 hours, but growth is completed in Burkholder's
medium containing 50y of pantothenate per liter after 50 hours. Readings taken
at 71 hours would be interpreted to mean that it was a "nonsynthesizer." Culture
No. 10 is capable of more rapid growth than culture No, 3 in Burkholder's with-
out added pantothenate and it grows much more slowly in the absence of
pantothenate than in its presence. Moreover, it synthesizes pantothenate after
a considerable lag in Burkholder's medium; In Hutner's medium the lag is much
shorter. Burkholder's medium is a much better diagnostic medium than Hutner's
since growth occurs more rapidly on it than on Hutner's when pantothenate is
supplied, but when pantothenate is not added, growth occurs less rapidly on
Burkholder's than on Hutner's. That is, Burkholder's is a better medium when
pantothenate is added but a poorer medium without added pantothenate.

THE "VITAMIN" PEDIGREE

Pedigree I (Table II) describes the melibiose-fermenting capacity of the vari-
ous cultures which were subsequently investigated In some detail for their
"vitamin-synthesizing" activity on Burkholder's medium according to his tech-
nique. The original diploid culture of S. cerevhiae (No. 1) was incapable of
fermenting melibiose, and its offspring, cultures Nos. 3, 4, 5, and 6, were similarly
incapable. In this pedigree numbers are used to indicate the different cultures,
and each group of four in a single column includes the cultures produced by the
four spores of a single ascus. Thus cultures 3, 4, 5, and 6 were originally from
one diploid ascus of S. cerevhiae. Culture No. 2 fS. carls her gensi^) fermented
melibiose, as did all of its haploid progeny. Culture No. 7 was the only survivor
of a single ascus, and a mating between it and No. 4 produced a hybrid supposedly
heterozygous for the ability to ferment melibiose.

This pedigree involves a non-Mendelian phenomenon which will be considered
in greater detail in a later paper (Lindegren & Lindegren, '47). The pedigree is
merely presented here for subsequent reference in tracing the descent of the dif-
ferent cultures.

HYBRIDS BETWEEN YEASTS SUPPOSEDLY DIFFERENT IN VITAMIN-

SYNTIIESIZING ABILITY

The members of this pedigree were characterized (Lindegren and Lindegren,
'45) as * Vitamin-synthesizers" and *Vitamin-nonsynthesizers." The distinction
between ability and inability to synthesize was made by Burkholder's method
with his medium. This procedure defines readings taken at 72 hours as diagnostic.
We have since discovered that this method does not give conclusive results, for if
growth were allowed to continue, the supposedly nonsynthesizing yeasts will
eventually begin to grow and will finally, in most cases, attain a level equal to
that of the so-called synthesizers. However, at the end of 72 hours a 10- or 20-
fold difference in growth often exists as indicated by reading the turbidity with a
Klett Photoelectric Colorimeter, Haploids of S. carlsbergensis were characterized

78

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

TABLE I

FORMULAS OF DIFFERENT SYNTHETIC MEDIA

Dextrose

Sucrose

Recrystallizcd

asparagine
K2HPO4
KH0PO4

MgS04.7H20

CaCl2.2Hi:0

(NH4)2S04
KI

NaCl

Sodium citrate
Citric acid

(NH4)2HP04

Ammonium tartrate

Burkholdcr's

Beadle & Tatum's

Grams per liter

(NHi)NOa

Parts per million

Boron

Manganese

Zinc

Copper

Molybdenum

Iron

Micrograms per liter

Thiamin

Pyridoxine

Nicotinic acid

Pantothenate

Biotin

Inositol

Riboflavin

p-amino-

benzoic acid

200
200
200
200

2

10,000

200
200
200

200
2

10,000

pH 5.0
with NaOH

pH 6.0

with citric
acid

20.0

5.0

Wickcrham's

10.0

1.0

0.125

1.0

0.875

0.5

0.5

0,1

0.1

1.0*

0.0001

0.1

0.1

5.0

w

1.0

0.01

0.07
0.01

0.05

400
400
400
400
2
2,000
200
200

pH 5.3

* Either asparagine or (NH4)2S04 was used.

+

+

; haploids of S. cerevisiae were
ridoxine 4-, according to Burk-

characterized as biotin — , pantothenate — , and p
holder's technique.

Table III shows the Klett readings obtained after 72 hours by Burkholder's
technique of these different cultures. It appears that culture No. 4 is a non-
synthesizer of pantothenate, while No. 7 is a synthesizer. A hybrid between No. 4

194 7]

LINDEGREN & RAUT VITAMIN

79

TABLE II

*

PEDIGREE OF A HYBRID BETWEEN 5. CEREVISIAE (LK) AND S. CARLSBERGEI^SIS

(MRAK. 126)

1

3
4
5
6

S. cerevisiae me diploid (Lk)

a me

a me
a me
a me

2

7
76
77
78
79

a

S. carhbergensis Me diploid (Mrak, 126)

Me
Me

Me
Me
Me

74
80

81

82

83

Me

Me
Me
Me

Me

4 X 7 (a me X a Me)

20
21
22
23

a
a
a
a

Me
Me
Me

Me

r

12
13
14
15

Me
Me
Me
Me

16 Me

17 Me

18 Me

19 Me

35
36

37
38

10 diploid

Me
Me

me
Me

A

90 Me

91
92
93

me

Me
Me

94 Me

95 Me

96 Me

110
111
112

Me

me
Me

4 X 20 (a me X

a

Me)

25

26

Me
Me

27 me

28 me

29 Me

3 me

31 Me
32

3 3 Me

34 me

35 Me
3 6 me

39

40
41
42

me

Me

me

5 X 25 (a me X a Me)

45
46
47
48

Me

me
me

49
50
51
52

Me
me

53
54
55
56

me
me
Me

57
58
59
60

Me

me
Me

me

61
62
63

64

me
Me

65
66

67

63

me

Me

69
70
71

72

me
Me
Me
me

86
87
88
89

Me
Me
me
mc

20 X23 (a Me X a Me)

124
125
126

127

Me

me
me

128
129
130
131

Me

Me

Me
Me

136

137

138
139

Me

Me
Me
Me

140
141
142
143

Me

me

Me
Me

144
145

146

Me
Me
Me

150

151

152
153

Me

Me
Me

4 X 49 (a me X a Me)

4 X 55 (a me X a Me)

104
105
106

Me
me
me

117
118
119
120

me
Me

Me

121
122

123

5 X 72 (a me X a me)

7 X 72 (a Me X a me)

98

me

101

me

99

me

102

me

100

me

103

me

Me
Me
Me
Me

me
Me
Me

[Vol. 34

80 ANNALS OF THE MISSOURI BOTANICAL GARDEN

and No. 7 (culture No. 10) appears capable of synthesizing pantothenate, but
the four haploid progeny, Nos. 20, 21, 22, 23, do not reveal any segregation for
this character according to this specific criterion. (The pantothenate character
does segregate according to Mendclian ratios In other pedigrees which do normally
segregate regularly.) It was subsequently discovered that none of these cultures
was a nonsynthesizer and that the data appearing in Table III do not give an ade-
quate picture of the synthetic ability of the organisms. The inadequacy of this
characterization will be dealt with in detail below.

VITAMIN-SYNTHESIZING ABILITY OF SUPPOSEDLY DEFICIENT YEASTS

IN A NATURAL MEDIUM

Dr. F, W. Tanner, Jr. (unpublished) grew various members of pedigree I
(Table II) in a natural medium containing molasses and corn-steep liquor for 71
hours. When the yeast and the medium were assayed it was found that under
these conditions all the cultures synthesized similar amounts of the different vita-
mins. Apparently the cultures were not clearly differentiated by ability and
inability to synthesize the vitamins, but were all capable of synthesis under
favorable conditions.

CONTINUOUS OBSERVATION OF GROWTH

Our present technique, an example of which has been presented in fig. 1, re-
quires a much longer observation period. Many of the same cultures were rein-
vestigated and were sometimes observed for as long as a month, readings being
made over the entire period. This is quite different from the standard practice of
discontinuing thf experiment after 71 hours. If the period of observation was
extended for a longer time, haploid cultures of S. cereihiae (as well as of S-
carhber gcnsh) were found to be able to grow on Burkholder*s medium in the
absence of pantothenate. This was equally true of all the progeny of the hybrid,
many of which had previously been described as "nonsynthesizers," In some cases
there was a delay of more than 600 hours before growth began. We grew the cul-
tures in 6 X ^-inch Kimble tubes and determined the amount of growth by
measuring turbidity on the Klett Photoelectric Colorimeter adapted to take the
larger tubes. This made It possible to make readings over the entire period with-
out discarding them.

THE INADEQUACY OF THE CRITERION, EQUAL INCREMENTS OF GROWTH FOR

EQUAL INCREMENTS OF VITAMIN

Culture No. 7 was planted in a series of media containing different concentra-
tions of pantothenate, and the data produced a family of parallel curves (fig. 3 in
following paper). If the experiment were terminated at an arbitrary time (standard
practice in nearly all assay experiments), curves could be obtained in which equal
increments of vitamin appear to produce equal increments of growth. The curves
in fig. 2 were obtained by readings taken at various times. At 50 hours there
appears to be a straight-line relation between the amount of added vitamin and

1947]

LINDEGREN & RAUT VITAMIN-SYNTHESIZING DEFICIENCIES 81

TABLE III

THE AMOUNT OF GROWTH AFTER 12 HOURS OF CULTURES FROM PEDIGREE 1 ON

BURKHOLDER'S MEDIUM DEFICIENT IN THE INDICATED VITAMINS

j^\ 1 ^ T

Photometer Reading

Culture No.

Photometer Reading

Culture No.

-?y.

—fa.

—Bu

^Py.

-Pa.

—Wu

; (diploid)

2 (diploid)

350
24

11^
280

11
280

1

4
7

350
310

22

15

20
200

12

8

125

4x7

JO (diploid)

315

240

135

20

274

45

50

12

21

312

202 ;

55

13

22

290

200

■

59

14

23

300

1

210

110

15

90

1

318

212

140

91

100

210

140

92

300

235

134

93

345

235

110

4 X 20

25

33

26

310

25

145

34

27

325

28

145

35

h

28

36

%

5 X 23

57

305

1

16

33

4

69

340

1

237

275

58

37

94

14

70

93

15

r

59

38

100

10

71

358

16

230

60

347

16

20

72

86

400

315

15

154

355

200

120

87

380

150

150

155

12

165

30

88

400

300

10

156

425

140

300

89

460

50

200

157

3

20

365

20 X 23

136

318

1

256

160

137

360

222

193

138

340

240

200

■

139

320

170

[Vol. 34

82 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the amount of growth over a considerable portion of the curve. However, this
culture was observed continuously, and eventually nearly as much growth was
attained in the medium without the added vitamin as in the one containing rela-
tively large amounts of added pantothenate (fig. 3 in following paper).

Oioo

c^

CULTURE NO. 7

~T3 W

MlCTOQrams of Pantothenate per Liter

TN>

Fig. 2. The relationship between increments of pantothenate In Burkholder's
medium and the amount of growth of culture No. 7 at different time intervals. In
•ome phases of the curves a straight-line relationsliip exists between the amount of
growth and the amount of added pantothenate but in all the culture tubes the growth
eventually rose to approximately the same level.

THE EFFECT OF MULTIPLE DEFICIENCIES OF THE B VITAMINS

I

A haploid culture of Saccharomyces cereiisiae which grew well on an agar con-
taining Hutner's synthetic medium with the vitamins, biotin, inositol, thiamin,
and pantothenate, but which was unable to grow on the same agar medium lack-
ing all these vitamins, was tested on agar containing various combinations of
vitamins. Pour plates following Lederberg and Tatum's ('46) technique were
made. In this technique a layer of sterile agar is poured in the bottom of the petri
dish, then a layer of agar seeded with 500 cells poured on top, and a third layer of
sterile agar on top of the second. This technique prevents colonies at the top and
bottom of the agar from growing diffusely over or under the agar. A penicillin
assay cup placed in the center of each plate was filled with a solution containing
the combinations of vitamins.

Colonies were counted and a number of methods of scoring were tested, but
simple -j- and — scoring is probably the most informative. The following tabu-
lation gives the results:

1947]

LINDEGREN & RAUT — VITAMIN-SYNTHESIZING DEFICIENCIES 83

VITAMINS PRESENT

B I Th Pa

I Th Pa

B Th Pa
B I Pa
B I Th
Th Pa
I Pa
I Th

B Pa

B Th

B I

Pa

Th

I

B

VITAMINS ABSENT

B
I

Th
Pa
B I

B Th

B Pa ■
I Th
I Pa
Th Pa

B I Th
B I Pa
B Th Pa

I Th Pa
B I Th Pa

SCORE

+ + +

+ + +

+ + +

+ +

+ + +

+ +

+ +

+ + +

+

+

+

+

+ +

+

+ +

The addition of inositol is apparently not very helpful in this concentration
(10,000 7 per liter) when biotin, thiamin, and pantothenate were also present, for
there was more growth when it was absent than when it was present. Actually
there were fewer colonies when it was absent, but those which grew did much
better. However, the culture supplied pantothenate alone grew less well than
that supplied both inositol and pantothenate. There were probably considerable
amounts of biotin in the agar, for removal of biotin did not usually reduce growth
greatly. Removal of biotin and inositol simultaneously was serious. When either
inositol or pantothenate were removed singly no serious effect occurred, but when
both were removed together there was considerably less growth. The cells could
synthesize both inositol and pantothenate easily when only one was absent but
lacking both they synthesized poorly. These facts indicate that inability to grow

in the absence of vitamins may involve simply lack of capacity to grow, or begin
to grow, under the prescribed conditions rather than inability to synthesize the

absent vitamins under all conditions (Williams, *41).

THE RELATION BETWEEN VIABILITY OF CELLS AND INABILITY TO GROW

ON DEFICIENT MEDIUM

The ability or inability to grow in a deficient medium may merely involve
inviability in the new medium rather than absolute inability to synthesize the
vitamin in question. If the cells are inviable in the deficient medium they will be

unable to begin synthesis of the required vitamin. The fact that most yeasts begin
to synthesize when they are permitted to stand a sufficiently long time suggests
that continued examination of cultures is necessary. This points up an important
difference between Net^rospora and yeasts. A conidium of Ncurospora can put
out a germ-tube and begin growth in distilled water, but if it is to continue growth

r

[Vol, 34, 194; !

84 ANNALS OF THE MISSOURI BOTANICAL GARDEN

It must be immediately supplied with the substances which it requires. It is rela-
tively difficult to establish a dormant myccVnim capable of awaiting future devel-
opments. Probably one of the most important advantages of a single-celled form
over a filamentous fungus is that the single-celled form can go through one or two
divisions and settle down to dormancy while a filamentous organism with special-
ized spores must continue to grow once the filament Is produced or it will perish.
This may give the appearance of absolute deficiencies to many mutants which die
shortly after producing a filament in a medium lacking a specific vitamin.

These results have been supported by further work on three other mutants
differentiated by ability to synthesize paraminobenzoic acid, pyridoxine, and
uracil. So-called nonsynthesizers of pyridoxine eventually grew in a medium con-
taining no added pyridoxine. Diploids heterozygous for a gene controlling synthesis
of paraminobenzoic acid produced two classes of offspring, one which grew
rapidly and one which grew slowly in the absence of paraminobenzoic acid. Al-
though these two classes were clearly differentiated, the weak synthesizers were
always able to grow eventually. An even more important bit of confirmatory
evidence was obtained from the stock incapable of synthesizing uracil. The
uracil-deficlent cultures responded quickly to large amounts of uracil added to
Burkholder's medium, but in the medium without added uracil they died. Trans-
fer from uracil-deficient medium to a complete medium a few days after inocula-
tion proved that the cells had died.

CONCLUSIONS

(1) Our results indicate that all so-called ' nonsynthesizing" yeasts which
remain alive when incubated in a deficient medium will eventually grow and
synthesize vitamins in that medium.

(2) We have decided that Burkholdcr's conclusion that certain yeasts arc
"nonsynthesizers" requires qualification because synthesis might have been dis-
covered if observation were made over a longer period.

(3) We suggest that many of Beadle and Tatum's "vitamlnless" mutants
may appear to be "nonsynthesizers" because they die in the deficient medium.
"Reversions" of Neuwspora mutants to wild-type, i, e., slow growth of "vitamin-
less" mutants in the deficient medium have been reported by Bonner, Tatum, and
Beadle ('43) and other workers in this field, but they are usually regarded as
exceptional cases rather than the standard expected behavior. The above results
suggest the possibility that an improvement of the medium, without addition of
the vitamin for which the stock is deficient, might lead to a higher frequency of
reversions.

A DIRECT RELATIONSHIP BETWEEN PANTOTHENATE CONCEN-
TRATION AND THE TIME REQUIRED TO INDUCE THE PRODUCTION
OF PANTOTHENATE-SYNTHESIZING "MUTANTS" IN YEAST^

CARL C. LINDEGREN

Research Professor, Henry Shaw School of Botany of Washington University

AND CAROLINE RAUT

Research Assistant, Henry Shaw School of Botany of Washington University

This paper describes experiments indicating that the concentration of panto-
thenate bears a direct relation to the time required to restore pantothenate synthesis
in a yeast cell. Various members of the pedigree shown in Table II- were grown in
batches of Burkholder*s medium made up with the following amounts of panto-
thenate added per liter: 100, 50, 20, 10, 5, 2, 1, 0.5, and y. Each tube was
inoculated in a uniform manner with a loop. Three hundred colonies grew from
each loopful of cells on plating, but since the haploid cells were typically aggre-
gated the total number of cells was probably less than 1500.

Figure 3 shows the results with S. cerevhiae (culture No- 1), the turbidity
being plotted against time in hours. The graphs are made by plotting the average
of the turbidity produced in duplicate tubes, except in a few cases in which
the tubes were so widely different that averaging did not seem to be a permissible
practice. Usually the readings differed by only a few units and averaging was
obviously acceptable. After 45 hours, growth is practically completed in the
media containing 50 and 100 y of pantothenate, but it is fully 75 hours before
appreciable growth is recorded in the tube without the added pantothenate. This
culture had previously been characterized as a synthesizer of pantothenate- These
data show that diagnosis depends largely on the time at which readings are taken.
Comparison of the 100 y and y tubes at the end of 45 hours would have resulted
in characterizing this particular organism as a "nonsynthesizer" of pantothenate.
The relationship between the amount of added pantothenate and the time at
which growth begins is quite clear, since the curves are all closely parallel
during early and logarithmic growth and overlapping occurs only after the
logarithmic phase of growth has been completed. There is a sharp diflference be-
tween the time at which growth begins in the tubes containing 0.5 and 1 y of
pantothenate per liter as well as between growth in tubes containing 1 and 2 y

of pantothenate per liter.

The culture of S. cerevhiae^ whose reactions are recorded in fig. 3, was induced
to sporulate, and similar tests with the four haplophase cultures are shown in
figs. 4 and 5. Cultures No. 3 and No. 4 are remarkably similar in behavior.
According to previous techniques, these would have been classified as "nonsyn-
thesizers" because growth in the absence of pantothenate did not begin until after
250 hours. The particularly interesting feature of the behavior of these cultures
is the direct relation between the length of the delay before growth begins and
concentration of pantothenate in the medium.

^This work was supported by grants from Anheuser-Busch, Inc., The American Cancer Society,
and Washington University.
*See preceding paper,

(85)

86

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

r<

ac

o
o

o

■o

o

t/1

Q£

o

z

AilOl9»ni

o
o

8

Of

o

X

8

O

o

3
O

o

8

Aiioiauni

o

o

o

8

t ^4

H

" E

S i

rt c

O (t

C C

,2 '3

C '
c -^

o .

(J

ex (A

c c

2 8

E

4J

.3

4>

E

o

3

P3

c

C «-

: 8

o

C3

IS

a

^T3

. s
;2 "

S

3

u

• fk

s

o

4-1

n

3

c

o

Im

C

to

H

4>

c
a.

M

1947]

LINDEGREN & RAUT PANTOTHENATE SYNTHESIS IN YEAST

87

3

o

D. o

O o

Si o

o

D
»

Q-
O

n

3

(A

p

CO

c

a
in

3

rt

5

rt

(A

c
3

o

o 5

3

3

rt

(A

rt

c
o

D

(A

rt

rt

3

o

3

n

rt
3

n 3

O- P

O W

I 3

^ ■ rt

H

rt

•n

O

3

rt

(A

o

TURIlDrTV

o

8

88

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

1

3

m

I

O"*

I

AiiOii«nA

3

;?

h^

\ 1 ^

3

C

G

a,

C

o

rt

w
«^

C

o

u
C

c
o

s

E

I

U

o

3

c

o

C

s

z

(1>

3

o

be

iZ

¥

1947]

LINDEGREN & RAUT PANTOTHENATE SYNTHESIS IN YEAST 89

In culture No. 5 (fig, 5) the different concentrations of pantothenate also
bear a direct relation to the delay before growth begins. A similar picture exists
for culture No. 6 (fig. 5), except that in the concentration of 1 y and y per liter
the duplicate tubes differed so markedly from each other that it was not permissible
to average the results. This is one of the few cases in which growth in ly per
hter in one of the duplicate tubes occurred later than that in the tube containing
0.5 y per liter.

The behavior in culture No. 5 (fig. 5) shows an almost ideal example of
general tendency of the "delayed" cultures. Growth in media containing 100
and 50 y per liter takes place at a very rapid rate. In the medium containing 20 y
the rate is somewhat decreased, and in lOy considerably so. A further decrease

i

occurs In 5 y per liter, so that there is a continual decrease in rate of growth in
the 50, 20, 10 and 5 y media, respectively. The case is quite different in the 2, 1,
0.5 and Oy media, where beginning of growth is delayed more and more as the
concentration decreases but once growth begins the rate is uniform and more rapid
than in the 5 y medium. The decreasing rate in the first five curves indicates that
where there is an excess of pantothenate the growth bears a direct relation to the
concentration of pantothenate, indicating that synthesis is suppressed when more
than 5 y per liter are present. (See also culture No. 4, fig. 4).

The rate of growth in the last four curves is practically identical, but the
delay before growth begins bears a direct relation to the concentration of panto-
thenate. This is interpreted to mean that in each of the last four curves the
growth begins after the induction of a "mutation" which possesses the ability to
synthesize pantothenate and that the rate of growth depends on the synthesis of
the vitamin by the cell. The basic assumption for this interpretation is the view
that de novo mutations from inability to ability to synthesize are extremely in-
frequent and the mutations observed in the laboratory are practically all "loss"
mutations. On this assumption, an agent which produces regular and precise
changes in cells from "nonsynthesizers" to "synthesizers" does not produce a
change of a completely non-existent locus to a synthesizing locus but merely
acts to bring a partially degraded or temporarily inactive gene into function-
al activity. The rate of growth Is independent of the concentration of panto-
thenate originally present in the medium (below 2 y per liter). However, the
time at which the "mutation" is induced (the "delay") depends on the concentra-
tion of pantothenate present; possibly directly on the number of molecules of
pantothenate Impinging on the gene. In a medium containing 1 y per liter, more
molecules would collide with any given surface than in one containing 0.5 y per
liter. The fact that only a small number of cells was used in Inoculating the tubes
and that easily detectable and constant differences exist between the low concen-
trations Indicate that mutation occurs in many of the viable cells in a given tube at
approximately the same time, rather than that one cell mutates and then outgrows
its neighbors. If the latter were the case, the curves would overlap and the pre-
cise differences between the different concentrations would not be detectable. It

y

90

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

i

Aiiaifluni

§

r*

Anoifluni

C

V

o

z

C

a

c

o

e

§

§

c

§

i

AiiaiBuru

1

«

§

AiiaiBuni

u

O

M

w

a

s

rsi

a

*N

8
2

3

s

I

1947]

LINDEGREN & RAUT PANTOTHENATE SYNTHESIS IN YEAST 91

is, of course, difficult to call these organisms "mutants" because the specific test
for mutation is segregation. When transfers from the culture tubes without the
added pantothenate were made to similar tubes likewise without added panto-
thenate (0-0) growth occurred about 150 hours sooner in the second than in the
original test. These curves of growth are shown as dotted lines on the graph. As
the yeast was transferred serially in the medium, the delay before growth started
was further shortened. The fifth transfer (0-0-0-0-0) began to grow sooner In
the medium without the addition of pantothenate than the cultures which had
been originally classified as synthesizers.

The original cultures came from a slant of yeast extract agar. Cells were sus-
pended in 10 cc. sterile distilled water, and transfers to Burkholder's medium were
made with a loop of this diluted suspension to Insure against transfer of vitamins.
A small but uniform number of cells was transferred in each loop. Irrespective of
the concentration of pantothenate in Burkholder's synthetic media none of these
original transfers failed to grow, but each grew after the delay indicated on the
graphs. Many other transfers were subsequently made from one tube of Burk-
holder's synthatic medium to another with the same concentration of pantothenate;
all these resulted in growth. Generally speaking, 0-0 transfers began growth much
sooner than the original transfer, indicating that a "mutation" had occurred In
the first transfer and that growth began due to the "mutation" or that some new
channel of synthesis was established which became more eflFicient with continued

use.

Saccbaromyces carlsbergensis (culture No. 2, fig, 3) is an undelayed synthesizer
of pantothenate, and growth in all concentrations of pantothenate is completed
before 100 hours. The single haploid offspring of S. carlsbergensis (No. 7, fig. 3)
is smiliarly an undelayed synthesizer but is spectacularly capable of utiUzing any
available pantothenate, as is shown by the beautifully parallel curves on the dif-
ferent concentrations. A hybrid between undelayed synthesizer (culture No. 7)
and delayed synthesizer (No. 5) produced the hybrid culture No. 10, which was
an undelayed synthesizer (fig. 3). When four haploid progeny from hybrid
No. 10, cultures Nos. 20, 21, 22, and 23 (fig. 6), were tested, all showed the
ability to use whatever pantothenate was available, as evidenced by the parallel
nature of the curves for different concentrations. However, these four progeny
were all undelayed synthesizers of the vitamin, and no clear-cut Mendellan segre-
gation occurred. This does not necessarily mean that the difference is not one
under gene control, for this pedigree is one in which gene transformation fre-
quently occurs. This matter is being dealt with in an article now in press
(Lindegren and Lindegren, '47). The pantothenate character segregates regularly
In other pedigrees In which Mendellan segregation of other gene-determined
characters normally occurs.

DISCUSSION

Non-Random ^^MutationJ* — Mutations are generally supposed to result from
random changes in the gene which occur independently of substrate with a specific

[Vol. 34

92 ANNALS OF THE MISSOURI BOTANICAL GARDEN

frequency. The probability that spontaneous or induced mutations would produce
adaptive or "progressive" changes in a gene are generally thought to be about as
likely as that the act of throwing a wrench at a motor would result in an im-
provement in the machine. Skoog and Lindegren ('47) have presented evidence
indicating that mutation to glucose utilization was influenced by the nature of the
substrate. The above data suggest that "mutation" which enables the cell to
synthesize pantothenate depends directly on the concentration of pantothenate in
the environment. The "mutations" induced by pantothenate arc quite different
from the ordinary recessive mutations used in genctical Mendelian analysis; they
may merely be the result of the addition to the gene of one of its essential com-
ponents. Such a component might correspond to what I have called the cyfogene.
This presupposes that pantothenate synthesis Is under genetic control. Most
previously described "vitaminlcss" mutants are probably genotypes unable to sur-
vive In the deficient synthetic medium. The genotypes which we described here
synthesize pantothenate when the level of pantothenate In the medium drops below
a certain minimum. However, it is suggested by the data that some pantothenate
(either In the cell or in the medium) must be present before the synthesizing
mechanism can operate.

Organized versus Molecular Genes. — The gene is probably a loosely organized
complex structure rather than a precisely definable chemical compound. The
tendency to conclude from (1) the experiments of Stanley (in which an Isolated
crystalline nucleoprotein was shown to produce the same effect as tobacco mosaic
virus) and from (2) the experiments of Avery, McLeod, and McCarty (in which
a nucleic acid was shown to be capable of transforming one type of pneumococcus
into another) that the gene Is either a crystalline nucleoprotein or a nucleic acid
disregards the possibihty that the nucleoprotein and the nucleic acid may be only
a part of the organized structure making up the gene. The fact that thousands of
molecules of the mosaic virus nucleoprotein are required to produce a single in-
fection has been Interpreted to result from the difficulty of securing Infection
with a single particle. An alternative Interpretation Is possible: It may be that
only one particle in a thousand of the "purified" preparation is so organized that
It is capable of infection. In the pneumococcus experiment the transformation

may have been achieved because the complex which comprised the gene producing
the smooth mutant form was brought into functional activity by the addition of
a single nucleic acid, just as a machine can be made to operate by adding a single
nut. This does not mean that the nucleic acid is the gene, any more than the nut
is the machine. Our experiments with pantothenate show that by adding molecules
of It to a suspension of yeast cells a cell Incapable of synthesizing pantothenate
could be transformed Into one capable of performing the synthesis. The fact that
a gene-controlling synthesis has become functional may not mean that a gene has
been added but merely that one component of the complex which makes up the
gene has been supplied. This component, though essential, may be only a part of
the total organized structure.

1947]

LINDEGREN & RAUT PANTOTHENATE SYNTHESIS IN YEAST 93

CONCLUSIONS

The evidence presented above indicates that in the presence of a large excess
of pantothenate no synthesis of pantothenate occurs although growth of cells by
utilization of the available pantothenate goes on at a very rapid rate. At concen-
trations not in excess of, but greater than the minimum required for growth, the
cells do not synthesize, and the rate at which they grow is determined by the
amount of pantothenate supplied. When the concentration of pantothenate is less
than the minimum required for growth the cells "mutate" so that they are able to
synthesize pantothenate and grow. The time required for this "mutation" to take
place is determined by the small amounts of pantothenate which are present in the
media. The data may not completely exclude the possibility that only a small
fraction of the population has been affected and that selection has been an im-
portant factor in the phenomenon; further tests of this view are in progress.
The present indications support the view that a large fraction of the popula-
tion is involved and if this be true, pantothenate can be regarded as an agent
which acts to repair a partially degraded gene. The data suggest that in the
complete absence of pantothenate neither synthesis nor growth can begin. (The
cells in the medium to which no pantothenate has been added did not necessarily
begin growth in the absence of pantothenate, for each cell probably carried a suf-
ficient amount to initiate growth.) Synthesis occurs in Burkholder's medium
only if enough pantothenate to initiate synthesis but not enough to suppress it is
present-

ACKNOWLEDGMENTS

We are grateful to Dr. J. O. Lampen and Dr. E. L. Tatum for helpful criticism
of the manuscript and to the members of our own research group for assistance
in the work.

BIBLIOGRAPHY

Avery, O. T., C. M. McLeod, and M. J. McCarty (1944). Studies of the chemical nature of the

substances included in transformation of pneumococcal types. Jour. Exp. Med. 79:137—140,
Beadle, G. W., and E. L. Tatum (1945). Neurospora. II. Methods of producing and detecting

mutations concerned with nutritional requirements. Am. Jour, Hot. 32:678—686.
Bonner, D., E. L. Tatum, and G. W. Beadle (1943). A genetic control of biochemical reactions in

Neurospora: A mutant strain requiring isoleucine and valine. Arch. Biochem. 3:71—91.
Burkholder, Paul R. (1943). Vitamin deficiencies in yeasts. Am. Jour. Bot. 30:206-211.
Hutner, S, H. (1946). Organic growth essentials of the aerobic nonsulfur photosynthetic bacteria.

Jour. Bact. 52:213-221.
Lederberg, J,, and E. L. Tatum ( 1 946 ) . Detection of biochemical mutants of microorganisms.

Jour. Biol. Chcm. 165:3 81-3 82.
Lindegren, C. C, and G. Lindegren (1945). Vitamin-synthesizing deficiencies in yeasts supplied

by hybridization. Science 102:33-34.
^ ^ ^ 1 947 ^ ^ Gene to gene transfer of gene-controlled fermentative ability in

yeast. (In press) .
Skoog, F. K., and C. C. Lindegren (1947). Adaptation In a yeast unable to ferment glucose. Jour.

Bact. (In press)
Stanley, W M. (1940). The biochemistry of viruses. Ann. Rev. Biochem. 9:545-570.
Wickerham, L. J, (1946). A critical evaluation of the nitrogen assimilation tests commonly used

in the classification of yeasts. Jour. Bact. 52:293—301.
Williams, R. J. (1941). Growth-promoting nutrilites for yeasts. Biol. Rev. 16:49-80.

MENDELIAN INHERITANCE OF GENES AFFECTING VITAMIN-
SYNTHESIZING ABILITY IN SACCHAROMYCES^

CARL C LINDEGREN

Research Professor, Henry Shaw School of Botany of Washington Unhersity

AND GERTRUDE LINDEGREN

Research Assistant, Henry Shaw School of Botany of Washington University

Pedigrees describing both Mendelian and non-Mendelian inheritance of the
ability to ferment carbohydrates In Saccharomyces have been reported by us
(Lindegren and Lindegren, M7). Genes controlling the fermentation of galactose,
maltose, or melibiose are transmitted in a regular Mendelian manner in some pedi-
grees, and in a non-Mendelian manner in others. Present indications are that this
IS due to the gene-to-gene transfer of some essential gene-component controlling
fermentation. This phenomenon complicated the problem of genetical analysis
of yeasts until regular Mendelian pedigrees were available.

The diagnosis of the fermentative ability of any selected culture is clear-cut,
no difficulty being experienced in distinguishing a fermenter from a non-fermenter.
In the present pedigree the fermentation of sugar is usually complete after 48
hours; the negatives do not ferment when held for three weeks. When regularly
segregating pedigrees became available, the problem of genetical analysis of fer-
mentative ability was capable of an uncomplicated solution.

Burkholder's medium (Lindegren and Raut, '47) is an excellent diagnostic
medium for distinguishing pantothenate "synthesizers" from "nonsynthesizers,"
because a so-called nonsynthesizer grows rapidly in this mcduim containing
pantothenate, but requires weeks or months to produce growth in its absence.
However, genes affecting vitamin synthesis are apparently transmitted in some
pedigrees in a non-Mendelian way similar to that displayed by genes controlling
fermentation. The first pedigree on the inheritance of "vitamin-synthesizing"
ability in Saccharomyces (Lindegren, *45) failed not only to reveal regular
Mendelian inheritance of this ability but also of genes controlling the fermenta-
tion of carbohydrates. In our selected inbred strains, the ability to ferment
galactose and maltose is transmitted in a regular Mendelian manner, and the
present paper shows that genes affecting the synthesis of paraminobenzoic acid,

pantothenate, pyridoxine, and thiamin are transmitted with corresponding regu-
larity. These genes are described as "affecting" rather than "controlling" the
synthesis of vitamins, because we have not discovered any absolute deficiencies in
yeasts. Lindegren and Raut have shown that a so-called nonsynthesizer of panto-
thenate eventually will grow in a medium without the addition of pantothenate,
although some cultures do not begin growth until they have stood in the tubes for

nearly a month,

^This work was supported by grants from Anheuser-Busch, Inc., Washington University, and
the American Cancer Society.

(95)

[Vol. 34

96

ANNALS OF THE MISSOURI BOTANICAL GARDEN

In gcnetical analysis, it is relatively unimportant wKether absolute or partial

deficiencies are dealt with; all that is required is a clear-cut differentiation of the

haplold offspring of a hybrid into two different categories. This is easily effected

in our present yeast pedigrees by using Burkholder's medium with and without

added pantothenate. Genes affecting pantothenate and pyridoxine synthesis are

easily diagnosed; the "nonsynthcsizers" do not begin to grow until a week after

planting while the "synthesizers" attain nearly full growth after 48 hours. The

former may produce a turbidity reading of between 200 and 300, while the latter

still show a reading of between and 5. After the tubes have been held for two

months it is difficult to distinguish them, but at 4 or 5 days the difference is

pronounced. Cultures differing in genes which affect the synthesis of thiamin and

paramlnobcnzoic acid show distinct differences at the end of 48 hours, but by the

fourth day it is difficult to tell them apart. However, any clear-cut segregation

of the progeny into two classes supplies the geneticist with an adequate gene
marker.

Table I describes 8 asci dissected from a hybrid heterozygous for mating type,

TABLE I

ANALYSIS OF ASCI FROM A HYliRID HETEROZYGOUS FOR MATING TYPE, FERMEN-
TATION OF GALACTOSE AND MALTOSE, AND GENES AFFECTING THE SYNTHESIS

OF PANTOTHENATE
(Ascosporcs from Hybrid 1426 X 1428 (a G ma pan X a g MA PAN) )

Culture

Type

G

MA

PAN

Culture
No.

Type

G

MA 1

PAN

No.

1

■

2

2101

a

-

_

274

2105

a

+

5

2102

a

—

1

KO

2106

a

+

190

2103

a

+

3

2107

a

+

2

2104

a

+

+

4

2108

a

1

+

254

t

5

4

2109

a

1

+

6

1

2113

a

1

3

2110

a

+

2

2114

a

—

+

210

2111

a

—

145 +

2115

a

+

220

2112

a

+

+

274

2116

a

+

I

+

5

d

>

2121

a

1

^ .^

1

+

2125

a

1

2

2122

a

+

+

^

200

2126

a

+

+

270

2123

a

+

5

2127

a

_

^_

4

2124

a

■ +

345

2128

a

■■—

'■ ■

200

7

8

2147

a

+

1

+

290

2151

a

+

4

2148

a

+

137 +

2152

a

+

250

2149

a

3

2153

a

—

140 +

2150

a

1

8

2154

a

»

m^^m.

4

I

1

1

1947]

LINDEGREN & LINDEGREN INHERITANCE IN YEASTS

97

galactose fermentation, maltose fermentation, and a pair of genes affecting the
ability of the organism to grow in Burkholder's medium without added panto-
thenate. The — and -[- signs under the columns G and MA indicate whether or
not the organism produced gas In a medium containing galactose or maltose re-
spectively. The figures under the column PAN show the turbidity reading reg-
istered in a Klett Photoelectric Colorimeter, after four days in a culture tube of
Burkholder's medium without added pantothenate.

TABLE II

ANALYSIS OF ASCI FROM A HYBRID HETEROZYGOUS FOR MATING TYPE. FERMEN-
TATION OF GALACTOSE AND MALTOSE, AND GENES AFFECTING THE SYNTHESIS

OF PARAMINOBENZOIC ACID, THIAMIN AND PYRIDOXINE.

(Ascospores from Hybrid 2236 X 2090 (a G MA pab th py X a g ma PAB TH PY) )

Culture
No,

24C9
2410
2411
2412

2423
2424

2425
2426

2431
2432
243 3
2434

2439
2440

2441
2442

Type

G

MA

PAB TH

1

5

7

9

a

a

a

died

+

+

+

11

+
+

2447

a

1

2448

a

+

+

2449

a

—

+

+

2450

a

+

^

+

PY

1

1

a

--

^—

a

^—

+

+

a

+

^_

a

1

+

+

1

+

+

a

1

1
1

.

a

+

1

+

+

a

+

—

—

a

+

+

1

+

a

1

+

+

+

a

+

+

+

a

+

+

a

1

1

k

1

+

+

4-
+

Culture
No.

Type

2427

2423

2429

2430

2445
2444
2445
2446

2451

2452
2453
2434

G

MA

I

PAB

TH

4

1

■

1

2419

a

+

2420

a

—

+

+

2421

a

+

+

2422

1

a

i

6

8

10

a
a
a

died

+
+

+

+
+

•^

-H

+

+

a

—

+ - ^

a

t '
i

+

a

+

+

+

a

+

1

1

.

a

i-

+

a

■

+

a

+

+

+

12

a

a

a

+
+

+
+

+
+

+

+
+

PY

+

+

+
+

+
+

+

+

98

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

TABLE II (Continued)

Culture
No.

2471
2472
2473

2485
2486
2487
2488

2495
2494
2495
2496

Type

a

a
a

G

I

MA PAB

TH

13

15

17

+
+

+

+

21

23

+

+

PY

2455

a

+

1

_

+

2456

a

+

+

+

2457

a

—

+

+

+

^

2458

a

+

+

4

2463

a

+

+

—

**—

+

2464

a

+

+

+

2465

a

—

—

+

+

^—

2466

a

+

—

^^

^^

+

a

+

+

i

+

a

^_

4-

+

—

a

+

+

+

a

— ■'

■ -■

m^im

1 •

1

+

+

a

■ —

^

a

+

+

+

— -

a

+

+

+

+

+

25

2501

a

-, -'

^^m

+

-U

2502

a

—

—

+

+

^

2503

a

+

+

^

+

—

2504

a

+

+

+

Culture
No.

19

2477

a

+

1

+

+

2478

a

—

—

+

2479

a

+

+

+

2480

a

^^

+

2467
2468
2469
2470

2474
2475
2476

2481
2482
2483
2484

2489
2490
2491
2492

2497
2498
2499
2500

Type

a

a
a

a

a

a

a
a

G

MA

PAB

TH

14

16

a

+

.

a

-^

a

+

+

a

+

+
+

20

+

+

+
+

+
+

22

24

+

+

+

+

+
+

26

+

+

+
+

+

+

a

+

+

+

b -.

a

—

^—

^—

+

a

+

+

I

a

+

+

■*

PY

2459

a

+

+

2460

1

a

+

—

—

^

+

2461

a

^—

—

+

+

+

2462

a

+

+

+

+

+

18

a

+

a

+

—

+

+

+

a

+

+

+

+

+

+

+

+

+

+

1947]

LINDEGREN & LINDEGREN INHERITANCE IN YEASTS 99

Cultures 2111, 2148, and 2153 produced the recorded turbidity in the panto-
thenate-free medium after 48 hours and were discarded. They would doubtless
have grown more, this being indicated by the -f" sign after the turbidity reading.
Each ascus produced two cultures with a turbidity reading of less than 8 and two
with more than 160 four days after inoculation. The genes controlling mating
type, galactose fermentation, and maltose fermentation also segregated regularly In

each of the eight asci.

Table II is a pedigree describing the cultures grown from the ascospores dis-
sected from 24 asci. These asci are derived from a hybrid heterozygous for mating
type, galactose fermentation, maltose fermentation, and genes affecting the syn-
thesis of paraminobenzoic acid, thiamin, and pyridoxine. The -}" and — signs
indicate whether or not the cultures ferment galactose or maltose, and whether
they grow in Burkholder's vitamin-free medium. The readings on the paramino-
benzoic- and thiamin-free media were made after 48 hours, while those in the pyrl-
doxine-free medium were made after four days. Two of the cultures from each
ascus produced heavy turbidity in the vitamin-frec media while two produced prac-
tically no growth at the time diagnosis was made. The readings were all recorded
numerically just as were the pantothenate readings shown in Table I, but for the
purposes of clarity were converted into -{- and — signs in the table. The only
exception to the expected Mendelian segregation of 2:2 in each ascus is found in
asci Nos, 5 and 22 in which three fermenters of galactose were discovered, al-
though the mating type, maltose fermentation, and vitamin characters segregated
regularly.

Tests were made for linkage between all possible pairs of genes, and usually

■

free assortment was indicated. In some cases linkage to each other or to different
centromeres was suggested but the evidence was not sufficient to warrant definite
conclusions. These data are presented to establish the fact that genes affecting
vitamin synthesis may segregate In a regular Mendelian manner In selected inbred
pedigrees.

ACKNOWLEDGMENT

We are indebted to Dr. E. L. Tatum for the information that one of the parent
cultures was paraminobenzoic acid-dcficlent.

REFERENCES

Lindegren, Carl C. (1945). Yeasc genetics. Bact. Rev. 9:111-170.

, and Gertrude Lindegren (1947). Gene-to-gene transfer of gene-controlIcd fermentative

ability in yeast. In presi.

, and Caroline Raut (1947). The effect of the medium on apparent vitamin-synthesizing

deficiencies of microorganisms. Ann. Mo. Bot. Card. 33:75-84.

AMERICAN ORIGIN OF THE CULTIVATED CUCURBITS^

I. Evidence from the Herbals
II. Survey of Old and Recent Botanical Evidence

THOMAS W. WHITAKER

Geneticist, Bureau of Plant Industry, Soih^ and Agricultural Engineerings

17. S. Department of Agriculture, La Jolla, California

L

INTRODUCTION

There are four species of Cttcurbita that rank as cultivated plants (C. Vcpo L.;
C. moschata Poir.; C. maxima Duch.; and C. jicifoli<r Bouche), and there is good
archeological evidence that the first three were present in the Americas in pre-
Columbian times (see Carter, '45). However, it has never been decisively demon-
strated that this group may not have been common to both Old and New Worlds
as seems to have been the case with the white-flowered gourd, Lagenaria sicercia

(Molina) Standi,

In the course of his investigation on the association of the cultivated cucurbits
with the various Amerind cultures of the Southwest, the writer had occasion to
examine most of the published work that concerns the origin of this group. The
present report is an attempt to evaluate this evidence, and draw the indicated
conclusions.

With the exception of Cucurbita ficifoliaj the four species with which we are
concerned are annuals. All have 20 pairs of chromosomes. They rarely, if ever,
produce species hybrids, except by means of artificial pollination, and then only
with difficulty. Up to the present, none have been discovered in the indigenous

state.

I. evidence from the herbals

The herbals of the 16th and early part of the 17th centuries are invaluable
sources of information in tracing the origins of the cultivated species of Cucurbita,
Prior to the establishment of contact with the New World in 1492, the herbals
contained no recognizable description or illustration of these plants. Surely plants
as large and distinctive in vine and fruit as squash and pumpkins would not have
escaped the notice of an astute group of observers such as the herbalist-scholars
of the 15th century appeared to be. A century after the discovery of America,
the record as traced through the various herbals indicates that two of the annual

^An investigation carried out while a Fellow of the John Simon Guggenheim Memorial Founda- ~
tion, 1946—1947. The writer Is grateful to the Director, Librarian, and staff of the Missouri
Botanical Garden for their courtesy in making available for study the excellent collection of
pre-Linnean literature found at that institution. Thanks are due to Professor Edgar Anderson
for his customary stimulating advice and criticism.

'^ Cucurbita ficifolia is ordinarily not thought of as a cultivated plant. The work of the Russian
investigators Bukasov, Zhiteneva ('30), Parodi ('34), and more recently the collections of Sauer,
West, and others (personal communications), indicate that it has a long history of cultivation, and
must be regarded as a cultigen. There are no archeological records of its occurrence. It is a per-
renial with 20 pairs of chromosomes.

(101)

[Vol. 34

102 ANNALS OF THE MISSOURI BOTANICAL GARDEN

species of Cucurbita had reached Europe, and one of them (C. Pepo) was rep-
resented by several varieties.

Fuchs (1542) seems to have been the first herbalist to note a cultivated cucur-
bit and produce a recognizable figure of it. His illustration, labelled "Turcklsch
Cucumer,

»> •

evidently some variety of Cucurbit a Pcpo (pi. 11, fig. 1). The
deeply lobed leaves and general appearance of the plant suggest that it may be allied
to our present-day Vegetable Marrows. From the shape of the fruit, there is reason
to believe that the illustration labelled **Meer Cucumer" is a variety of C. Pepo
currently known as *'Small Sugar" (pi. 11, fig. 2). Like the illustrations of most
herbalists, Fuchs' are somewhat conventionalized, in order to accommodate the

wood

Pcpo

Matthiolus (1560) has an illustration of what seems to be a field pumpkin

labelled Cucurbita indica (pi. 12, fig. 1) Dalechamps (1587) has a

Pcpo)

»perly

Al-

Pcp

the remaining morphological characteristics makes it seem certain that the plant
is referable to this species.

cur bit a Pepo. J

pones lati) of what appears

Scallop" (pi. 12, fig. 2).

White Bi

\ plantaru

Pep

Cucurbita verrucosa (pi. 12, fig. 3). Bauhin (1650-51) has a reversed copy of
Dalechamps' figure, and Bailey ('29) is undoubtedly correct in assigning It to

C. Pep

species.

to tl

lis

Lobelius (1591)

Pepo (Pep

Pepo rotundus compressus Melonis effigie, Melo-pepones latiores Clypeiformes,
Melo-pepo teres, and Melo-pepo compressus alter). The fruits pictured under the

pepones latiores Clypeif^

!oD-fruitcd summer saua

12. fig. 4).

?cpo

with any of our present-day varieties. In addition, Lobelius has produced the first

pressus

Pep:

Tabernaemontanus (1591) is particularly rich in the number of varieties of
Cucurbita Pepo which are illustrated. A total of nine forms are figured, some of
which can be recognized as closely allied to our present-day varieties. Melopepo
clypeatus is undoubtedly a form of the "White Bush Scallop" summer squash;

^pitata is much hke the former with a slightly different fruit shape.

drils.

pepo teres and M, compr

Pepo

1947]

WHITAKER ORIGIN OF CULTIVATED CUCURBITS 103

Vegetable Marrow type; the same is true for Pepo Indicus viinor oblongus. The
form designated as "Pepo Indicus minor rottmdus is quite similar in shape to our
modern variety, "Perfect Gem." ?epo Jndicus minor clypeafus and Pepo Indicus
minor an gnJosus (pL 12, fig. 6) are forms whose fruit shape and general appearance
are strongly reminiscent of the modern "Table Queen" or "Acorn" squash.

The results of this survey provide strong evidence that none of the cultivated
species of Cucurhita were known to the botanists of the Western World before
1492^. In the following century at least two species (C. Pepo and C. maxima)
were recognized by the herbalists, and for one of them fC. Pepo) a number of
varieties were known. It seems strange that C. moschata was not introduced into
Europe during this period. There may be several reasons for this: (1) In general,
this species is more subject to range restrictions by low temperatures and short
days than either C. Pepo or C. maxima; (2) recent distribution data indicate
that it is found only in the more inaccessible regions of Mexico, Central America,
and Colombia.

Cucurbita ficifoUa, with its relatively hard shell and rather coarse, stringy
flesh, lacks the edible quaHties of the annual species. This may have been the chief
reason for its neglect by the early explorers. Furthermore, this species requires a
relatively long photoperiod, and it is doubtful whether it would mature fruits in
Europe, except perhaps in the extreme southern portions and under exceptionally
favorable cultural conditions.

II. SURVEY OF OLD AND RECENT BOTANICAL EVIDENCE

far the Old World origin of the cultivated species of Cucurbita

species

World

'Taudin (1856), At the beginning of
his extensive and illuminating memoir, which has laid the experimental foundation

4

for our understanding of the species of this group, he devotes a single paragraph
to their origin. He states that of the six known species (C. moschata, C PepOy
C. maxima, C m^lanosperma, C. perennis, and C digitata) the first three have

been cultivated for a considerable length of time in Europe. The nativity of C.

It is claimed, without documentation, that

fnaxtma

C.Pepo

maxima and C- moschata are more mod

into European gardens ("leur introduction dans nos jardins ne remontant guere
au dela de deux siecles").

^Sturtevant ('19, p. 219) has summarized this line of evidence in a remarkable lucid statement,
**If we consider the stability of types and the record of variations that appear in cultivated plants,
and the additional fact that, so far as determined, the originals of the cultivated types have their
prototypes in nature and are not the products of culture, it seems reasonable to suppose that the
record of the appearance of types will throw light upon the country of their origin. From this
standpoint, we may, hence, conclude that, as the present types have all been recorded in the Old
World since the fifteenth century and were not recorded before the fourteenth, there must
be a connection between the time of discovery of America and the time of appearance of pumpkin
and squashes In Europe/*

104

[Vol. 34

MISSOURI

Naudin, in discussing Cucjirbita ficifolia (C. Jfielanospcrma Caspar.), states
that It was Introduced into Europe about 1800 A. D., probably from southern
Asia as Indicated from Its common name, "Courge de Siam.'* Reports of travelers
led him to bch'evc that at this time it was grown in China on a large scale; thus
confirming his opinion that the species originated in Asia, Naudin thought that
C. ficifolia has important potentialities as an economic plant, for use as human
food if properly prepared in the immature stages, and as cattle food because of its
long-keeping qualities.

In a later paper, Naudin (1859) reports further experimental work with various
genera of the Cucurbitaceae. He does not make any positive statement about the
origin of the cultivated cucurbits, although he infers that C. moschata is an Old
World indigene. He states that seed of several varieties collected in India have
been grown at the Museum. Since the early terminology of cucurbitaceous fruits
was in much confusion, it is highly probable that Naudin has mistaken Pliny's
reference to watermelons, melons, cucumbers, and gourds as including some mem-
bers of the genus Cucurhita. There is no evidence to support the belief that Pliny
was familiar with the latter group.

The widely held conviction that the three commonly cultivated species of
Cncurbita were of other than American origin was continued by De Candolle
('83) on very slender, and for the most part, questionable evidence. Later in-
vestigators (Cogniaux, 1881; Pitticr, *26; Herrera, '41) have propagated De Can-
dolle's views without critical reexamination of their basis. From De Candolle's
discussion of the origin of the four species under consideration it is apparent that
he is positively in favor of an Old World origin only in the case of Cncurbita
maxima^ and there is some reason to doubt that he felt that the record was en-
tirely convincing here. In terminating his discussion of the origin of C. maxima
he makes the statement, "En definitive, sans ajouter une foi implicite a Tindigenat
sur les bords du Niger, fonde sur le dire d'un seul voyageur, je persiste a croire
Tcspece originalre de Tancien monde et introduite en Amerlque par les Europeens."

The best evidence De Candolle could muster for his Old World theory of the
origin of Cncurbita vtaxima was Hooker's (1871) citation of localities for certain
collections: i.e. "Upper Guinea. Nupe on the Niger, apparently indigenous. Bar-
ter!" Wclwitsch's discovery of this species in Angola is also referred to, but there

is no indication as to whether or not it was an indigenous plant. The fact that
Barter's plants were collected along the banks of a large river would lead to the
supposition that It was an introduced species. Welwitsch's collection was made in
or around a village, and it is therefore quite likely that the plants were escapes. At
best, De Candolle's arguments for an Old World origin of C. maxima rest on an
extremely flimsy foundation.

As for Cucurbita Vepo, De Candolle presents the documented evidence for and
against its Old World origin. His position may be summed up by stating that the

historical record does not contradict the opinion that this species may be of

American origin.

1947]

WHITAKER ORIGIN OF CULTIVATED CUCURBITS 105

According to De Candolle^ the origin of Cticurhita moschata presents an un-

^ solved problem. However, he is inclined to attach some weight to the unproven

assertion that this species was more widespread in southern Asia than in any other

region during the seventeenth century. As stated previously, C. moschata was
unknown to the botanists of the fifteenth and sixteenth centuries. The first record
of its occurrence seems to be the excellent illustration published by Van Rhede in
Horfiis Maliharicus (1688). During the 17th and 18th centuries it appeared in
several floras of southern Asia and Africa (Wight, 1843; Clarke, 1879; etc.).
However, in no case was it claimed to be an indigenous plant.

Evidently Cucurbita Tuoschata was introduced into European horticulture from
southern Asia (Naudin, 1856), rather than directly from the Americas. The
common names given to varieties of this species were indicative of Old World
origin, i. e. "Pleine de Naples," "Pleine de Barbaric," "Muscade de Provence," etc.

De Candolle suggests that Cticurhita ficrfolia is of American origin, since up to
the time of his investigations, all the perennial species of the genus were natives to
California or Mexico, whereas the annual species were assumed to be of Old
World origin. This argument has now lost whatever cogency it may have had.
Bailey ('43) has described several species from North America which are un-
doubtedly annuals.

Evidence for the New World origin of the' cultivated species of Cucurbita

In a critical review of certain phases of De Candolle's book, Gray and Trum-
bull ('83) present the evidence for an American origin of the three annual species.
Their report can best be summarized by quoting directly:

"Allusion has already been made (under Lagenaria) to the difficulty of distinguishing the
genera of the Cucurbit aceae, under names by which they are mentioned by voyagers and explorers
of the first century after the discovery of America; and the question of species is particularly diffi-
cult. Yet we find abundant evidence — especially as respects North America — (1) that in various
parts of the country, remote from each other, the cultivation of one or more species of Cucurbits
by the Indians was established before those places are known to have been visited by Europeans;
(2) that these species or varieties were novel to Europeans, and were regarded by botanists of the
16th and 1 7th centuries, as well as by the voyagers and first colonists, as natives or denizens of the
region in which they were found; and (3) that they became known only under American names;
one of these names (Squash) becoming, in popular use, generic, and two others (Macock and
Cushaw) surviving, as names of varieties, into the present century."

Gray and Trumbull then present strong evidence for tbeir conclusions, follow-
ing a chronological scheme as nearly as possible. First, the reports of several early
explorers and historians are cited. Although it is usually difficult or impossible
to determine precisely to what species these writers have reference, it is almost
certain that they are concerned with one of the three annual species of Cucurbita,
probably C, Vepo. The reports of the following explorers are cited: Columbus,
Cuba, 1492; Cabega de Vaca, Florida, 1528; De Soto, Florida and Mississippi,
1539-41; Carrier, Canada, 1535; Sagard, Canada, 1642; Lahontan, Southern Can-
ada, 1703; also the historians who mentioned pumpkins, macocks and squashes
Captain John Smith, 1606-08; Strachey, 1610; Higglnson, 1629; Beverley, 1705;
and others.

106 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

Further support is provided by the works of the 16th century herbalists
Fuchs, Dodoens, Matthiolus, etc. It is clear, as Gray and Trumbull point out, that
the Cucurbitas were considered foreign or novel by these botanists. Furthermore,
the word "Indian" as applied to the area of origin of various species did not neces-
sarily mean that they came from Asia, but rather the West Indies or the Americas.
Much confusion has come about through a misinterpretation of the word "Indian."
De Candolle and others have, for the most part, interpreted it in a narrow sense as
applying only to British India, but the truth seems to be that Cucurbita Pepo and
C. maxima reached Europe from the West Indies or directly from the American
continent-
Gray and Trumbull regard Nuttall's (1818) statement of particular impor-
tance in establishing the American origin of the cultivated Cucurbita. Nuttall
mentions two species, C. Lagenaria and C. verrucosa (Warted Squash), and of the
latter, he observes, "Cultivated also by the Indians of the Missouri to Its sources."
Cucurbita verrucosa is one of the warted varieties of C. ?cpo. Trumbull's work
(1876) in tracing the Indian origin of the words squash, cushaw and macock is
considered by Gray and Trumbull as being especially significant in establishing a
case for the North American origin of Cucurbita Pepo and C. moscbata. Trum-
bull states in summarizing his conclusions, ''As regards North American varieties,
the evidence seems conclusive. These varieties at least bear Indian names, which
date from the first coming of the Europeans, and of these varieties we have no
mention before they were found in North America."

Recent botanical evidence.-

The Russian plant explorers (Bukasov, '30; Zhiteneva, '30) have contributed
an Immense amount of data to our knowledge of the distribution of the cultivated
Cucurbitas. Briefly, they have found that the greatest diversity of the group is

ficifol

Pebo

general area (Mesa Central). It is important to note the omission of C. maxima
from the above list. Apparently this species has never been cultivated by the
natives of Mexico, Central America, or the northern portions of South America.
Cucurbita ficifolia, according to the Russian investigators, is the most widely
distributed species of the group. It is found in all countries from Mexico to
Chile along the Cordillera. There are white-seeded and black-seeded forms; other-
wise, the composition of the species is very stable over the entire range. Cucurbita

almost as widely distributed as C. ficifolia. It is extensively grown in

Ce

and Chile.

pecies of Cucur

The forms found in Mexico and parts of Central America are typically white-
seeded, while those of Panama and Colombia are brown-seeded. The Russians

Mesa Ce

po

collections, and appears only sparsely in their records from Central Mexico.

1947]

WHITAKER ORIGIN OF CULTIVATED CUCURBITS 107

Parodi ('3 5) has made a significant contribution to the subject in his study of
pre-Hispainic agriculture in Argentina, He finds that Cncurhita maxima was one
of the principal species of plants cultivated by the Guarnies in northeast Argen-
tina and Paraguay at the time of the conquest of the Rio de la Plata.

Cardenas ('44), after completing his studies of the cultivated Cucurbitas of
Bolivia, comes to the conclusion that there were several varieties of Cucurbita

r

maxima present in the Andean valleys of South America at the time of the con-
quest. On the other hand, C. Pepo and C. moschata are evidently of recent intro-
duction into the cultivated crop complex of Bolivia, Paraguay and Argentina. He
suggests that a thorough exploration of the temperate and tropical portions of
Bolivia and Peru might uncover wild relatives of the cultivated cucurbits that
would be helpful in deciphering their relationships. The basis for this suggestion
is the discovery of a small, warted gourd (el "joko*') in an isolated region of
Bolivia (upper canyon of the Rio Caine). This gourd is cultivated for food and
is believed to be closely related to C. fepo.

SUMMARY

1. Negative evidence of the presence of Cucurbita Pepo and C. maxima In
Europe prior to 1492, also familiarity of the herbalists of the 16th and 17th cen-
turies with these species, suggest very strongly that they were introduced into
Europe from the Americas.

2. An examination of the evidence in favor of the origin of the cultivated
species of Cucurbita in the Old World indicates that it is very fragmentary, and in
general unacceptable.

3. The botanical record, while not as extensive or decisive as it might be,
clearly favors an American origin of the cultivated cucurbits.

4. Finally, the archaeological and botanical records lead inescapably to the

conclusion that the four cultivated species of Cucurbita, C Pepo, C. moschata,

C. maxima, and C. ficifolia, are New World in origin. The possibility that C.

moschata may have been common to both hemispheres is not ruled out, but it does
seem relatively remote.

LITERATURE CITED

Bailey, L. H. (1929). De CucurbUis domesticatis. I. Gentes Herbarum 2:63-115.

, (1943). Species Cucurbitae. Ibid. 6:267-322,

Bauhin, J. (1651). Hisioria universalis plantarum 2:222.

Bukasov, S. M. (1930), The cultivated plants of Mexico, Guatemala, and Colombia. Bull. Appl.

Bot., Genet, and Pi. Breed. Suppl. 47:1-553.
Cardenas, M. (1944). Las cucurbitas cultivadas de Bolivia. Rev, de Agr, 2:3—12.
Carter, G. F. (1945), Some arcKeologic cucurbit seed from Peru, Acta Americana 3:163—172.
Clarke, C. B. (1879), Cucurbitaceae. In Hooker, J. D. Flora of British India 2:604-635.
Cogniaux, A. (1881). Cucurbitaceae. In de Candolle's Monographiae Vhanerogamarum 3:325—951.
Dalechamps, J. (1587). Historic generaUs plantarum 1:615.

DeCandolle, A, (1883). Origine des plantes cultivees. Paris. 377 pp.

Dodoens, R. (1563). Cruydeboeck. 508 pp.

Fuchs, L. (1542). De historia stirpium. pp. 69%-699.

Gray, A., and J, H. Trumbull (1883). Review of De Candolle's "Origin of Cultivated Plants."
Amer. Jour, Sci. 25:370-379.

[Vol. 34. 19471

108 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Herrera, F. L. (1941). The flora of Cuzco 1:398.

Hooker, W. J. (1871). Order 64. Cucurbitaceae. In Oliver, D. Flora of tropical Africa 2:555-556.

Lobellus. M. {l59l).Icoms stirpium 1:642-643.

Mattliiolus. P. A. (1560). Commentarli secundo auctu 290 pp.

Naudin. C. (1856). Nouvelles rechercHes sur les caract^res spccifiques et les varietes des plantes du

genre Cucurbifa, Ann. Sci. Nat. IV, 6:5-73.

(1859). Revue des Cucurbitacdes cultivees au Museum, en 1859. Ihid. 12:79-164.

Nuttall, f. (1818). The genera of North American plants 2:228.

Parodi, L. R. (1934). La alcayota en la Argentina. Rev. Argent. Agron.^ 1:85-86.

, (1935). Relaciones de la agricultura prehispanlca con la agricultura Argentina actual.

Observaciones generales sobre la domesticacion de las plantas. Ann. Nac. Agron. y Vet. Buenos

Aires 1:115-167.
Pittier, H. (1926). Plantas usuales de Venezuela. 458 pp.
Seringe, M. (1828). In De CandoUe's Prodromus systematls naturalh 3:316.
Sturtevant, E. L. (1919). Notes on edible plants. (Edited by U. P. Hedrick). State of N. Y. Dept.

of Agric. 27th Ann. Rept. 22:1-686.
Tabernaemontanus, I. T. (1591). Neuw Kreuterbuch. 2:178-181.
Trumbull, J. H. (1876). Vegetables cultivated by the American Indians. I. Bull. Torrey Bot.

Club 6:69-71.

Van Rhede, H. (1688), Hortus Malibaricus, pars %xpl 2.

Wight, R. (1843). Iconci plantarum Indiae OrientaUs Z\t, SO?*

Zhiteneva, N- E. (1930). The world's assortment of pumpkins. Bull. Appl. Bot., Genet, and Pi.

Breed. 23:157-207.

, (1930o). Survey of the principal literature on the systematics of pumpkins and squashes.

Ibid, 343-356.

Explanation of Plate

PLATE 11

Fig. 1. "Tiirckisch Cucumer" of Fuchs, apparently a variety of Cucurbit a Pepo.
Fig. 2. "Meer Cucumer" of Fuchs, evidently C. Pepo var. "Small Sugar."

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 11

4

X

>

P^

O

o

n
c

<

>

D

n

c

n

c

^
(j:)

NJ

Ann. Mo. Bot. Card., Vol. 34, 1947

Plate 12

CVCVRBITA INOICA,

Pepbnes Uci.

t»:crD<prporrtm.

1

2

3

Mclo-pffonej UtiomClypcirot

Pepo maxir.mi Indirui comprcf-

5t(«in2\n^^^nif(6(cf<*fi39^>oncn.

Pcpo 1 ndicus minot anguloriu.

4

5

6

\C^HITAKER— ORIGIN OF CULTIVATED CUCURBITS

IVoL. 35, 1946]

WHITAKER ORIGIN OF CULTIVATED CUCURBITS 111

Explanation of Plate

PLATE 12

Fig. 1. Cucurbit a indica from Mattliiolus, probably Cucurbita Pepo,

Fig. 2. Pepones lati from Dodoens — Cucurbita Pepo, possibly van "White Bush

Scallop."

Fig. 3, Cucurbita verrucosa from Dalechamps, evidently a warted variety of C. Pepo,

Fig. 4. Melo'pepones latiores Clypeiformes from Lobelius, probably Cucurbita Pepo
van "Golden Custard."

Fig. 5. Pepo maximus Indicus compressus from Lobelius — the first illustration of
Cucurbita maxima.

Fig. 6. Pepo Indicia minor angulosus of Tabernaemontanus, probably Cucurbita
Pepo van "Table Queen."

V

^

FOSSIL POLYPORES FROM IDAHO

HENRY N. ANDREWS
AND LEE W. LENZ

In June, 1946, we spent a day searching for fossil plant remains in tKe late
Tertiary deposits in southwestern Idaho. Our primary quest in this region was for
petrified evergreen cones that have occasionally turned up, although only in such
quantity as to whet the appetite of collectors. The focal point of that day's col-
lecting was approximately 10,5 miles south of Bruneau and .5 mile east of state

runs

County into Nevada.

J

M

fied polypore. More recently, Mr, S. H. Osgood, of Rupert, Idaho, has sent us a
fragment of another specimen. Although both of our specimens seem to be
clearly referable to the fossil Fames idahoensis Brown (Brown, '40), in view of
the great rarity of fossil polypores a brief record of the specimens seems worth

while.

This part of Idaho Is well known to local mineral collectors for its abundance
of fossil wood, as well as the occasional cones. Most of these fossils are weathering
out of a loosely consolidated, fine white sandstone which is overlain by a brownish-
buff sandstone of a harder texture likewise yielding petrified plant remains. Over-
lying the productive plant beds is a horizon which yields an abundance of well-
preserved fish jaws (Mylocyprinns rohustus Leidy). The only stratigraphical
study of the beds in this region is that of Piper's ('24), and the horizon from
which our fossils were obtained was apparently in his group No. 8 which is de-
scribed as "Lake sediments, semi-consolidated, white, gray, and buff sandstones and
sandy shales, volcanic ash; . - . .". These beds have been considered to be of
Pliocene age although it is possible that they may be of later origin. In a recent
letter (February 20, 1947) Dr. Bobb Schaeffer has informed us of a collection of
Mylocyprinns rohtisftis fossil fish pharyngeals in the American Museum that were
collected "from an area in southwest Idaho between 'Catherine and Sinker Creeks.'
This particular locality is considered to be Pleistocene and as this genus has not
previously been reported from older formations I am wondering if your horizon
might not also be referable to that period.'*

h

POLYPORACEAE

FoTnes idahoensis Brown. — The Dodds specimen is a nearly complete sporo-
phore, only a small portion of one side having been lost. As seen in top view
(fig. 1) it measures 10 x 11 cm. It bears but one layer of pores, which are 12 to
15 mm. long and number approximately 720 per square cm. Judging from the
portion remaining it did not attain a thickness exceeding 2 cm. The rings of
growth, characteristic of the living bracket polypores, are clearly defined on the
upper surface. While it is not sufficiently well preserved to reveal any significant

(113)

[Vol. 34. 1947]

114 ANNALS OF THE MISSOURI BOTANICAL GARDEN

diagnostic characters, a longitudinal ground thin section reveals a fine filamentous
structure suggesting mycelium. The specimen is preserved as No. 5000 in the

collections of the Henry Shaw School of Botany.

The Osgood specimen (No. 5001) is a fragment of an appreciably larger poly-
pore that was probably about 15 cm. in diameter. The pores of this specimen
attain a length of slightly more than 20 mm.

The primary interest in these fossils lies in their evident position in the Poly-
poraccac, and within this family they appear to be most closely related to the
genera Fames and Polyporus. Their general aspect is that of a Fomes^ and because
of the close resemblance to Fames idaJjoensis we have assigned them to that
species. The lack of dependable color preservation in these, as in most fossils,
detracts appreciably from making an entirely dependable comparison with modern
species of Fames and Falyporus, Brown has, however, noted a rather close sim-
ilarity between F. iJa/x)Cf7sis and the living F. pinicola (Sw.) Cooke.

Only two undoubted American fossil polypores have been recorded previously.
Mason ('34) has described a specimen of Fames applanatns (Pers.) Gill, from the
Pleistocene Tomales formation of Tomales Bay, California. Brown*s specimen of
FarTves idalyoemh was collected about 5 miles north of the Bruneau locality from
which our specimens were found.

We wish to thank Mr. J. M. Dodds and Mr. S. H. Osgood for kindly presenting

these fossil fungi for preservation in the Henry Shaw School of Botany collections.

Thanks are also due Dr. Bobb Schaeffer, The American Museum of Natural History,

for identifying the fossil fish pharyngeals as being referable to Mylacyprwtis and

probably the species rohtistus of Lcidy. Through the courtesy of Dr. R. W. Brown

we have been able to compare our specimens with a portion of the type specimen
of Fames idaJjoensis,

REFERENCES CITED

Brown, R. W. (1940). A bracket fungus from the late Tertiary of southwestern Idaho. Jour.

Wash. Acad, Sci. 30:422-424.
Mason, H. L. (1934). Pleistocene flora of the Tomales formation. Carneg. Inst. Wash,, Publ.

415:83-179.
Piper, A. M. (1924). Geology and water resources of the Brimeau River Basin, Owyhee County,

Idaho. Idaho Bur. Min. and Geol., Moscow, Idaho. Pamphlet No. 11:1-55.

Explanation of Plate

PLATE 13

Fames idaJyoensis Brown

Fig, 1. Upper surface of the sporophore. Pores may be noted where portions of the
context have broken away.

Fig. 2, Under surface.

Fig. 3. Side view showing the single layer of pores.

Fig. 4. A portion of the under surface showing the pores enlarged.

Figures 1—3 nearly natural size. Figure 4 magnified 4 times. All photographs of
specimen No. 5000.

Ann. Mo. Bui. Gard., Vol. 34, 1947

Plate 13

ANDRKW'S &: LFNZ— FOMES IDAHOENSIS

Ann. Mo. Bot. Gard.» Vol. 34, 1947

Plati 14

^ . m

r

^

4

t

DR. JOHN HENRY FiRIT'lS

November 1, 1836 — Novfinhfr IS, 1909

JOHN HENRY BRITTS— PHYSICIAN AND FOSSIL HUNTER

HENRY N. ANDREWS

In 1899 the United States Geological Survey published a classical volume on
the Carboniferous fossil plants of the western Missouri region. The author of this
work is the late David White, a noted and able geologist and paleobotanist, while

behind the scenes, contributing the fossil plants that made it possible, worked a

g ■

country physician of Clinton, Missouri, Dr. John Henry Britts.

The daily routine of Dr. Britts* life appears to have been no less crowded than
is usual for men in his profession, although it was somewhat more colorful, judging
from the variety and scope of his undertakings. In addition to his successful medi-
cal practice he found considerable time to serve the state in capacities of very
lasting benefit. One of the most significant of these Is the part he played in re-
vealing the floral splendor of Missouri as It existed some 250 million years ago.
But before mentioning In detail these more purely scientific achievements it may
be well to briefly sketch in the background of his earlier life.

John Henry

Jane Rogers Britts. His great-grand-

country

til

the Revolution. George M. Britts moved from Virginia with his parents apparent-
ly at about the turn of the 19th century and located at Ladoga, Indiana. He
studied medicine and practiced in Parke and Montgomery counties until the year
1842. The son, John Henry Britts, was born November 1, 1836, in Montgomery
County, Indiana. He attended the state schools until the age of 19, when he began
the study of medicine. At that time he went to live with his maternal grand-
father, Dr. Henry Rogers, with whom he remained until the spring of 1857, when
he moved with his family to Clinton, Missouri. He then resumed his study under
the preceptorship of his uncle. Dr. John A. Rogers, and during the college year
1857-58 attended lectures at the St. Louis Medical College. It is thus quite evi-
dent that he hailed from a family in which the medical tradition was well estab-
lished, and In view of the scientific interests that he displayed through life it is
not surprising that he followed this course at the outset. His formal training cer-
tainly was not extensive as compared with modern concepts, yet his learned
relatives imparted to young Britts a sound and comprehensive scientific foundation,
judging from the success and distinction that he later achieved.

Shortly after Dr. Britts began the practice of medicine in Cass County in 1859
Governor Jackson issued a call for troops to repel the Federal invasion of Missouri.
Britts responded and proceeded to raise a company of which he was made Captain.
Six months later he joined General Price's army at Springfield and helped to organ-
ize Waldo P. Johnson's battalion, which became later a part of the Fourth Infantry
Regiment of the Confederate States Army. Dr. Britts was made surgeon of the

(115)

[Vol. 34

1 1 6 ANNALS OF THE MISSOURI BOTANICAL GARDEN

regiment with the rank of Major, later being promoted to Brigade Surgeon. On

June 9, 1863, while on duty at the City Hospital during the siege of Vicksburg,
he was wounded by a shell thrown into the city by Porter's fleet, and it was found
necessary to amputate his right leg. The following August he left Vicksburg, a
paroled prisoner, and after his recovery served as surgeon in Alabama and Georgia
until the end of the War. Upon his return to Clinton in 1865 Dr. Britts formed
a professional partnership with Dr. P. S- Jennings which lasted until the death of
the latter thirty years later.

Dr. Britts' practice, like that of most physicians working in small cities and
towns, extended over a wide territory. It was this frequent local travelling, com-
bined with an intense scientific curiosity, that made possible the accumulation of
his collections of fossil plants. At that time many coal mines were operating in
Henry County and the adjacent territory, and Dr. Britts often visited them.
Along with his medical kit, he always carried in his buggy a bag of tools, including
pick and hammers, to carve out a few choice specimens from the shales that the
coal miners laid aside for his study.

In addition to the numerous specimens that he furnished to private individuals
and public institutions a large collection of Dr. Britts' fossils were placed many
years ago in the Chicago Academy of Science's Museum. Mr. Eliot C. Williams,
Jr., has kindly given me the following information concerning these specimens:

"The accession record shows that on May 12, 1904, the Academy received a collection
of 1124 coal plants collected by Dr. John H. Britts, in Missouri and Pennsylvania. This
gift was made by Mr. Francis Peabody, A notation in the accession record states that the
collection contains many types.

"I have checked on the collection, and it seems to be in good shape, but I did not count
to see whether or not there are still 1124. I would judge that the collection is probably
intact.

*"A label in one of the cases indicates that this collection was the basis for Monograph
#37 of the U. S. Geological Survey on the Fossil Flora of the Lower Coal Measures of
Missouri, by David White." (Letter dated July 18, 1946.)

Dr. Roland W. Brown has informed me that the U. S, National Museum houses
approximately 1,000 of Dr. Britts' specimens from the Missouri Coal Measures.
Many of these are type and figured specimens. I have not had occasion to study
these two Important collections of American Carboniferous plants, although it is
believed that paleobotanists working In this field would be glad to know of their
whereabouts. Apparently the two collections contain all of the types in the above-
mentioned Monograph.

Following the publication of the Missouri Monograph Dr. Britts continued
collecting plants from the coal mines. Most of the specimens composing this
later collection were acquired during or shortly prior to 1904. It was through
the kind offices of Mr. D. K. Greger, formerly a curator of paleontology at Wash-

4

ington University, that the existence of this collection was brought to my atten-
tion some eight years ago. At that time It was in the possession of a grandson of
Dr. Britts, Mr. J. B. Owen of Clinton, from whom it was later purchased for
Washington University In St, Louis. It is not a large collection, but a considerable

1947]

ANDREWS — JOHN HENRY BRITTS 117

portion of the specimens are of exquisite beauty both in their preservation and
scientific and teaching value. It is especially rich in fine examples of Astero-
phyllites and Annularia.

Most of the mines from which these fossils were acquired have long since
closed, and it is questionable whether it will ever again be possible to continue
where Dr. Britts left off. At least we have a reasonably representative selection
of the plant fossils from his old hunting grounds which will always serve as an
invaluable foundation for such paleobotanical studies as may be carried on in that
region. Our knowledge of Carboniferous floras is increasing every year as new
regions are opened up and old ones reworked, and it is largely through the efforts
of such active amateur workers that this knowledge is forthcoming.

Thanks are due Mr. John B. Owen for placing at my disposal biographical
data pertaining to his grandfather.

THE IDAHO TEMPSKYAS AND ASSOCIATED FOSSIL PLANTS^

HENRY N. ANDREWS

AND ELLEN M. KERN

CONTENTS

Historical introduction 1 19

I

The present status of our knowledge of Tempskya 121

The localities 122

Size and form of the trunks 125

Ontogeny of the trunk and the restoration 131

The roots 138

I

Comparisons with other plants, living and fossil 143

Taxonomlc considerations 146

Associated plant remains 151

Notes on methods 153

Summary 155

Acknowledgments 156

Literature cited 156

Plates 158

Appendix — Discovery of the Wayan Tempskyas, by Henry Thomas 184

HISTORICAL INTRODUCTION

Very few fossil plants have ever presented a more distinctive anatomy or more
challenging problems than the Cretaceous Tempsykas. It is perhaps well to point
out in these first few lines that we do not pretend to have arrived at a final ac-
counting of all the existing gaps in our knowledge of these ferns. However, cer-
tain significant facts have been discovered concerning their habit and the plants
associated with them in life which will stand unchallenged, while our interpreta-
tions may raise doubts or be modified by future investigations. It is intended that
this should be taken as simply another chapter in our growing knowledge of the
Tempskyas.

Through a most fortuitous circumstance that has been described elsewhere
(Andrews, '47) I was able to comb certain of the hills in southeastern Idaho^ in

This study was aided by a grant from the Penrose Fund of the American Philosophical Society,

^Collecting activities were carried on by the senior author and certain persons in Idaho whose
names are given in the Acknowledgments and elsewhere in this paper.

(119)

[Vol. 34

120 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the early summer of 1942 under the guidance of Mr. W. A. Peters of Jerome,
Idaho, chiefly in search of petrified trunks referred to the genus Te?npskya, In a
technical paper it may seem out of order to dwell on an introduction to the subject
at hand, yet there are so many details of botanical, historical, as well as general
human interest, attached to this group of plants that we believe they should be
recorded for the benefit of those who may continue with studies of this and other
plant groups in Idaho.

In the summer of 1939 I was presented with a small fragment of a Tempskya
which had been collected in a gravel pile near the Haddenham fossil shop at Fossil,
Wyoming. At that time even a fragment seemed like a treasure — it meant material
for class study, but, of greater importance, it meant that Tempskyas should be
found in much more abundance near by. In later years we traced the probable
origin of that fragment to an Upper Cretaceous horizon running north and south a

few miles to the east of Kemmerer, Wyoming. It was not, however, until an abun-
dance of large, well-preserved specimens were found in the adjoining Idaho hills
that we became fully aware of the importance of these plants in the Cretaceous
vegetation.

Through an intimate knowledge of their countryside, a number of local col-
lectors have enabled us to acquire a considerable quantity of specimens. The vig-
orous collecting activities of Mr. C. Henry Thomas, of Wayan, Idaho, and Mr. E.
Manlon, of Firth, should be noted in particular. Among the numerous westerners
whose acquaintances have enriched my life and laboratory the name of Henry
Thomas should be recorded as a great collector and a Tempskya specialist. In his
assiduous search for these fossils he may be compared only with Wieland, who
collected cycads in the Black Hills, or the early bone hunters such as Sternberg or
Hatcher. It is a comparison on a smaller scale and of a somewhat more specialized
nature, but the pioneering spirit and prodigious productivity differ but little.

My first contact with Mr. Thomas was in 1942, when he still occupied his
former ranch on the Williamsburg bench area. I was not prepared to lunch in these
rather remote though beautiful hills with a rancher whose cabin was lined with
hundreds of books. While lacking the literary capacities of a Thorcau it was soon
evident that here was a man who understood and appreciated the world he lived

in. At the rear of Mr. Thomas's cabin a wooden platform already displayed scores
of fine Tempskyas collected mostly within a radius of a few hundred yards. En-
couraged by Dr. Roland W. Brown and myself, Mr. Thomas set to work scouring
the near-by hills with increased interest and enthusiasm, with the result that the
collection has been increased manyfold, consisting now of some few tons of fine
specimens. In all probability far more than in all other collections combined
(fig. 2). Believing that the Tempskyas are destined to occupy an important niche
in Cretaceous floras, I requested Mr. Thomas to write in his own words a few lines
pertaining to his discovery of the fossils in this region. This has been included in
the Appendix for such historical interest as it may have for future paleobotanlsts.

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 121

THE PRESENT STATUS OF OUR KNOWLEDGE OF TEMPSKYA

Although petrified trunks belonging to the genus Tcmpskya were discovered
well over a century ago in Europe, the first really informative accounts of these
plants were those given by Kidston and Gwynne-Vaughan in 1911, and Seward in
1924, More recently Read and Brown ('37) and Read ('39) have given us much
more comprehensive treatments. A detailed review of previous contributions is
included in their account and will be repeated here only in so far as is necessary
to orient the reader and to compare our findings and concepts with those of
previous workers.

It is especially important to note that prior to Read and Brown's important
work all descriptions had been based on either very poorly preserved specimens or
a few fragmentary ones. The only possible exception to this statement is Seward's
description of Tcmpskya Knowltofti from the Colorado shale of Montana. A de-
tailed consideration of that specimen will be taken up later. This historical aspect
of the subject is particularly significant in the case of Tcmpskya for its anatomy
is so peculiar as to lead to highly erroneous conclusions concerning the habit of the
plant unless adequate material is available for study.

Six species of Tcmpskya have now been described fiom North America. These
include a specimen from Maryland described by Berry in 1911. According to
more recent workers this was very poorly preserved and is of little interest or
importance other than as a geographical record for the genus. Later Seward de-
scribed his T. Knowltoni from Montana, and in 1937 Read and Brown described
two more species and recorded specimens from a considerable number of localities
in Wyoming, Idaho, Utah, Montana, and Oregon, Most recently Arnold ('45)
has described two more species, Tcmpskya WcsscUi and T. wyamingcnsis^ from
Montana and Oregon, and Wyoming, respectively. Combined with the previous
European reports which record specimens from Russia, Bohemia, and England, the
wide distribution of these plants in Upper Cretaceous times is well established.

We should like to emphasize that our own studies have not been primarily of
a taxonomic nature. We are inclined to doubt that certain of the better-known

American species are sufficiently distinct to warrant the specific names that they
bear but with this phase of the Tcmpskya story we have no quarrel or primary in-
terest. When dealing with anatomical characters it is not always possible to
arrive at entirely satisfactory criteria for the segregation of species. In the rather
large quantity of material that we have had available for study there is consider-
able variation In the gross form of the trunks, but with the exception of a very
few specimens it has seemed most feasible to assign all of these to one species. Our
efforts have been directed primarily toward arriving at a clearer concept of the
general habit of the plants, their ontogenetic development and physiology. In
pursuing this course we have perhaps tended to put less emphasis than previous
workers on the segregation of species. However, In view of the undeniably close
relationship of the species of Tcmpskya we do not feel that our approach has
materially slighted a sound taxonomic treatment. While we have sectioned many

[Vol. 34

122 ANNALS OF THE MISSOURI BOTANICAL GARDEN

specimens through a wide variety of size and shape, a great many others, par-
ticularly the larger ones in the Thomas collection, have not been available for this
purpose. We believe that we have studied In cellular detail an adequate number of
representative specimens, and from scores of others we have drawn information
concerning the habit of the plants. There Is obviously some practical limit to the
number of specimens that can be handled, and with some field experience with the
Tempskyas one may select representative material with a minimum danger of
missing essential details.

A typical transverse section of a Tcmpskya trunk reveals a most unique
anatomy. It consists of numerous, small, siphonostelic stems held firmly together
in a dense matrix of diarch roots (figs. 24, 28). Taken individually, the anatomy
of a single stem is not unlike that of a modern maidenhair fern (figs. 20, 21, 22),
such as Adiantum pedatumy while leaf traces are given off in two rows toward the
nearest external point of the trunk (text-fig. 2). This unit aggregation of many
branching stems with their petioles and roots has been called a "false-stem" by

previous writers. It is, we feel, a superfluous term as well as somewhat misleading.
If a special term must be used it would be more appropriate to call it a "super-
stem," and while an adequately descriptive phrase would be cumbersome, we have
preferred to use the term tnink as one that involves no new creation and can
hardly be misinterpreted.

In order to define clearly the objectives in our own study It may be most ex-
pedient to note the chief gaps in our knowledge of these fossils. Sufficiently large
collections had not been available for study to settle many of the concepts con-
cerning the habit of the trunk — whether It was creeping, ascending, or upright.
Read ('39) has discussed this In some detail In an Interesting and critical paper.
The Ontogeny of the trunks, their unique physiological set-ups, and the manner
in which the foliage was borne present problems that have been but partially
explained. It is to these categories that we have been especially drawn. Further-
more, previous work on Tcmpskya has offered but little evidence of the kinds of
plants that were associated with them In life. We have been fortunate In finding
in the Wayan, Idaho, district the fossil wood of a conifer, a dicotyledon, as well
as a cycad specimen with the silicified fern trunks.

THE LOCALITIES. AGE AND AREA

The greater part of the Tcmpskya specimens In our collection and all of those
in the Thomas collection were obtained from an area of a few square miles im-
mediately east of the Wayan post-office. In order to show this area precisely we
have reproduced in text-fig. 1 the northeast corner of the United States Geological
Survey's topographic map of the Lanes Creek quadrangle. Although the silicified
trunks have been gathered over the greater part of this territory the most produc-
tive areas are shown within the dotted lines. Most of the collecting that has been
done has been simply a matter of exploring the surface of the hillsides and small
stream beds. Excellent material is obtained In this way, and in most cases the
specimens show no evidence of long transport either before or after fossllization.

1947J

ANDREWS & KERN

IDAHO TEMPSKYAS

123

•H

H

3

o

5 =>

o

ri-
ft

o

"-1
o

fa- r^

o

?

o

w

3

n
•-I

rb

C
w

-< i^

o' ^

3

O

ft

3

ft O

OS.

I"

ft

3

C/5

rt

O

3
4

S

8

g"

1

Y

[Vol. 34

124 ANNALS OF THE MISSOURI BOTANICAL GARDEN

During a week's collecting in the autumn of 1945 two days were spent ex-
cavating in section 27 in the bank of a draw from which Mr. Thomas had previ-
ously obtained some exceptionally fine material. Many of his largest and most
complete trunks were obtained from a pit at this location. In the course of about

r

three hours' digging one may expect to take out as many hundred pounds of speci-
mens. However, with the exception of one other pit, very little digging has been
attempted in this area. It seems likely that large quantities of the fossils remain .
underground.

This topographical area falls within the bounds of the Wayan formation al-
though the exact position of the latter within the Cretaceous is still uncertain.
The most detailed stratigraphical account of the region is that of Mansfield ('27)
in which it is noted that "Definite correlation of the Wayan formation is impos-
sible at present." In a chart showing the geographical distribution and strati-
graphical correlation of Tempskya deposits in the United States, Read and Brown
have tentatively placed the Wayan formation near the base of the Upper
Cretaceous.

In 1942 Mr. E. Manion, of Firth, Idaho, kindly guided a small party of us to
a hillside approximately 25 miles east of Ammon, The exact position of this
locality Is: NW'A sec. 5, T. 2N, R. 4lE, Hell Creek quadrangle, Idaho. Speci-
mens have been found here in some abundance although the area is limited to a few

acres in extent. The locality was revisited in 1946 and a dozen small specimens
obtained. On that occasion we continued our search in the surrounding hills with-
in a radius of two or three miles but found no other fossils. It is to be wondered
that any one should have ever happened on this small outcrop, yet we feel that
there must be many more in the vast extent of the Cretaceous beds that go far to
the south. This locality lies approximately 3 5 miles northwest of Wayan and in
all probability represents an extension of the same Tern pskya-hc:inng beds. Speci-
mens obtained here are generally well preserved although they do not differ
anatomically from those collected near Wayan. This will be referred to as the
"Ammon locality" in future references in this paper.

In Wyoming, shortly to the east, Tempskya is found in the Aspen shale and
the Thermopolis shale, both of lower Upper Cretaceous age. Tempskya Kuowltoni
from Montana was found in the Colorado shale which extends into the middle
Upper Cretaceous, while Berry's T. Whitei was derived from the Patapsco forma-
tion in Maryland, of upper Lower Cretaceous age. T. Wessel/i (Arnold, '45) is
reported from the Kootenai formation northwest of Great Falls, Montana (as well
as a placer outwash at Greenhorn, Oregon). There is a possibility that the Mon-
tana specimens may have weathered out of the overlying Colorado group (Arnold,
'45, p. 26). T. wyomingcnsis was found twenty miles northeast of Grcybull,
Wyoming. Arnold notes that "Fragments of Tempskya are widely scattered
within the valley of Beaver Creek and its tributaries, but they are nowhere
abundant. They have been found only where the Morrison formation is exposed
and arc associated with enormous numbers of dinosaur bone fragments and

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 125

gastroliths;".

It is thus clear that the Tempskyas ranged from the upper part of the Lower
Cretaceous through middle Upper Cretaceous times.

SIZE AND FORM OF THE TRUNKS

With only one or two exceptions all the Tempskya specimens that we have ex-
amined from the Wayan and Ammon localities compare most closely with
Tempskya Wcssclii Arnold. Although the following discussion is based on this
species unless otherwise noted, the views that are expressed concerning its habit,
ontogeny, and physiology are probably generally applicable to the genus as a whole.

Some concept of the size that the Tempskyas attained was known as early as
1836 from Fitton's description of a trunk 9 feet long and 12 x 4 inches in
diameter, found in the southeast of England (Stopes, '15, p. 14). Much more
recently Read has reported trunks up to 10 inches in diameter. The largest speci-
men in the Thomas collection, and so far as we are aware the largest yet reported,
is 16 inches in diameter. Other fragmentary specimens, which do not constitute
complete transverse sections, indicate trunks of even larger size, so that a maximum
of 18 or 20 inches in diameter seems very likely. It is thus clear that these were
plants of no mean magnitude, although the evidence indicates that they did not
attain a great height.

As a matter of convenience we shall consider the hundreds of specimens in

r

the Thomas collection in three categories: basal specimens, tips, and intermediate
portions which will be called discs. This latter term will apply to any specimen
that is complete in transverse section but may be quite variable in length.

The bases, — One of the chief objectives of the 1945 trip to Wayan was to
obtain basal portions of the trunks in order to shed further light on their general
habit; that is, whether they were ascending or upright. Fortunately some fine
specimens were collected in the field and others located in the Thomas collection.
In all cases the specimens which we have interpreted as being the basal portions
flare outward slightly at the very bottom (figs. 3, 5, 7), present a characteristic
knobby lower surface, and are composed exclusively of roots. It is apparent, as
Read and Brown have pointed out, that the stems in the older, basal portions of
the trunks decayed completely, their place being taken by roots. A more detailed
discussion of the anatomical details will be given later.

The basal periphery and under surface of these stumps are distinctive. The
former is characterized by slightly buttressed, rounded projections and the latter
by shghtly raised knobs and cavities, or in a few instances by one large shallow
cavity. The general uniformity of these specimens would seem to support the
view that they represent the original stump portion of the trunk and not simply a
broken sector taken from some higher level. It is pertinent to add, therefore, that
they are all upright, indicating perfectly erect trunks.

One cannot be certain whether the roots extended out uniformly in all direc-
tions or whether they tended to aggregate into more massive "compound roots."
The knobby character of the extreme base may point to the latter explanation

[Vol. 34

126

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Text-fig. 2. A diagrammatic drawing of a transverse section of specimen T3 8 (peel 14)
showing the distribution of the stems and the position of the xylem of stele and petiole traces. It
may be noted that stem branchings are numerous and the trunk is radially symmetrical with respect
to departure of the traces. Natural size.

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 127

although this is oflFered only as a suggestion. An objection may be raised that
these small almost microscopic roots (pis. 20, 21, 24) could not have adequately
anchored such a massive trunk. When it is considered, however, that they probably
radiated out by the tens of thousands, that at some points they may have been ag-
gregated to form compound roots, and that the individual roots possess an extremely
stout sclerotic cortex (fig. 19, etc.), there can be no doubt that their supporting
capacity was very great. Another objection may be raised that this knobby sur-
face is an erosion artifact caused either before or after fossilization. However, the
sides and upper surface of these specimens show no such effect, and the stumps
always flare outward slightly at the extreme bottom, as might be expected.

The disc specimens, — The specimens that we have interpreted as being basal
^nd terminal portions of trunks are considerably in the minority, which indicates
that the plants did attain a height of at least some few feet.

If the cross-sectional form of these disc specimens could be depended upon as
a specific taxonomic character the number of species represented would be very
nearly endless. Different specimens vary from circular, to broadly elliptical, to
strongly flattened In transverse section (figs. 10, 24, 25); and one specimen was
found in which the trunk is crescent-shaped (fig, 9). Since we have not been
able to observe any correlation between these variations in form and the internal
structure of the trunks, such variations would seem to be of no taxonomic sig-
nificance. It is possible that the variously flattened specimens have resulted from
crushing caused by overlying sediments prior to silicification. A more detailed
anatomical consideration bearing on this problem will be given on a later page.

The size and form variation of some specimens in the Thomas collection is
recorded in Table I, this information having been compiled chiefly with the view
of arriving at some concept of the height that these plants attained. In compiling
these data a representative selection of specimens has been taken, all of which were
complete in transverse section. Many more might have been added to the list, but
in general they would have affected only the quantitative aspect of the table.

Since none of the specimens represents a complete trunk we cannot arrive at an
exact figure for the height of a plant in life; however, from many observations of
their diameters and the rate of tapering we may calculate a reasonably dependable
minimum. In the cases of the base and disc specimens we have recorded the
diameters at both ends in order to indicate the rate of taper. Of the terminal
trunk specimens only the basal diameter can, of course, be given.

An examination of these figures for the trunk (disc) specimens will show that
none of them taper abruptly from one end to the other. Such tapering is found
only In the undoubted terminal specimens. The longest one that we have ob-
served is in the Manion collection (specimen A), from the Ammon locality, and
through its length of 21 inches it displays no tapering. The same holds true in a
general way for the basal specimens, which flare slightly at the extreme base but
otherwise eive no evidence that the trunks were very short (figs. 3 and 7).

[Vol. 34

128

ANNALS OF THE MISSOURI BOTANICAL GARDEN

TABLE 1

MEASUREMENTS OF A RET^KESKNTATIVE COrLECTION OF TEMPSKYA SPF.CIMENS

Specimen number

9

Nature of speclmenf

11

19

21

26

Tr

Tr

Tr

T

Tr

31

T

32

Tr

33

38

39

40

Tr

Tr

T

Tr

42

Tr

45

4rt

47

48

Tr

Tr

Tr

T

51

A (Manion coll.)
W. U. (T216)

Tr. (near base)

T

Tr

6

10

B

13

22

25

34

fi

B

B

B

B

41

43

44

B

B

B

49

23

24

W. U. (T47)

W. U. (T230)

B

Ti

Ti

Ti

Ti

Diameter (s) In inches

9. X 6.5:|:
11. x7.

9.3 X 6

10. X 7.

n. X 104:1;

12.3 X 10.5

4.7 X 2.5
6. X 4.

14. X4.7

8.5 X 6.5
9. X 7.

7. X 5.

8. X 5.5

13. X 7.3

8.75

6.5 X 6

5. X 3.3
5.7 X 3.7

9.3 X 5.3
10. X 5.3

14
16

9.

6. X 4.5

7. X 5.5

7.
7.5

8. X 5.5

9. x7.

11. X 7
12.5 x7

6. x3.7

6. x3.5

10.5

X 5.

5.5

X 3.7

3.

4.3

X 1.7

X 2.

16.7

X 13.

8.7
9.

X 9.7
X 10.

9.5
9.

X 4.
X 6.

9J
8.7

X 7.3
X 7.3

9.5

11.

13.

X 12.

7.

7.5

X 4.3
X 5.5

8.5

^ 13.35

10.

X 4.5

Length in inches

12.3

il

11.

6.

11

11.

10

13.

6.

7.

7.

6,

11.

10.

7.

9.

10.

21.5

16.

8.

9.

9.

8.

7.

15.

10

12.

9.

12.

7.

12.

7,

8.5

* Specimen numbers refer to the autlior's field notes, all measurements being taken from specimens

in the Thomas collection unless otherwise indicated. jTr — trunk (disc) ; B — base; Ti tip.

JA slightly elliptical specimen, the figures In the upper and lower lines being for the top and
bottom diameters respectively. :|::l:No appreciable taper of the specimen, the top and bottom
diameters being the same. JNo appreciable taper, and the specimen is cylindrical or nearly so.
55In t-he case of tip specimens this refers to the diameter at the base.

1947]

ANDREWS & KERN — IDAHO TEMPSKYAS 129

In this respect it is necessary to consider the internal structure of the trunks.
The densely compacted stems and roots that make up the trunks must have had in
life the consistency of a very tough strand of rope. In texture it probably was not
unlike an Osmunda rhizome with its vast coat of petioles and roots. In the
Tempskya trunks, however, there were many stems and the whole aggregation was
bound together very tightly. Thus while the trunks must have been very tough
they were probably not extremely rigid. The thousands of roots created a closely
interwoven unit; yet, lacking an interspersed ground tissue, such a height as is
attained by a coconut or royal palm would seem to be improbable. One other
pertinent point may, however, be mentioned here. Unlike other unbranched,
columnar trees such as living palms, tree ferns, and cycads, Tempskya bore very
small fronds, as Is evidenced by the relatively minute size of the petioles, and we
shall offer evidence that these were borne not merely in a crown at the top but for
a considerable distance down the trunk as well. Thus, in all probability the
Tempskyas did not have the mechanical problem of a large weight of foliage con-
centrated at the top, a feature which allows for a considerably taller trunk than
might otherwise be expected.

It may be noted that some of the longest specimens from the Wayan region
(Table I, Nos. 33, 34, T216) display but little tapering from one end to the
other. The nineteen disc specimens considered in the table all have an average
taper of approximately .6 inches per foot. Thus, assuming a uniform taper
throughout, a trunk 10 inches in diameter at the base would taper to a point at a
height of about 16 feet. This is probably in excess of the height actually attained
since, among other factors, the apex of the trunk tapers abruptly to a blunt point
(figs. 26, 27). From the terminal specimens at hand we may suppose, then, that
at a diameter of 3 or 4 inches our trunk terminated, giving a height of about 12
feet. In view of the relative proportion of basal and terminal trunk specimens
along with the discs this figure seems quite reasonable.

The specimen of maximum diameter (#51 in the table) measures 14 and 16
inches at the upper and lower ends, respectively, and is 10 inches long. This had
been exposed for some time prior to collection and is somewhat lichen-encrusted.
No evidence of stems could be observed in the transverse sections, indicating that
the specimen came from near the base of a trunk, although it does not represent
the basal-most portion as both end surfaces are Irregularly broken. On the basis
of the estimate given above for' a trunk 10 inches in diameter it is possible that
this specimen may represent a plant that attained a height of about 19 feet.

Prior to the publication of Read and Brown's work, accounts of Tempskya
were based on so few, as well as fragmentary, specimens that a satisfactory concept
of the habit of the trunks was not possible. Read has more recently presented a
more detailed discussion ('39) dealing with "The evolution of habit in Tempskya,*'
While certain of his concepts are clear-cut and sound, we are not able to agree in
all respects with his conclusions. Read states, "In short, the writer's concept of
the growth form of the dorsiventral false stemmed species of Tempskya is an as-

130 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

ccndlng, climbing type of fern with numerous liana adaptations. It is of course
obvious that the very basal portions of the stem were horizontal or oblique. How-
ever it is doubtful if these subterranean portions developed the dense mass of
parallel roots characteristic of the false stem. Rather they must have been
markedly divergent." (p. 70).

In the hundreds of specimens that we have collected or studied from the Idaho
localities the evidence points toward an upright, self-supporting trunk without
liana adaptations, and we cannot agree that It Is "obvious" that the basal portions
were horizontal or oblique. Our evidence as gained from a study of the external
form of the trunks may be considered first.

It seems significant that we have encountered only two specimens (one shown
In fig. 8) which show any noticeable longitudinal curvature. Basal specimens give
no evidence of other than upright habit from the start.^ Two concepts that seem
to have become indelibly impressed in the minds of those paleobotanlsts who have
seriously studied the Tempskyas deserve analysis at this point. The first of these is
based on the specimen of Tempskya Knowltoni described from Montana by Seward
in 1924. That specimen Is described and figured as being 3 3.5 cm. long and obcon-
ical in form, the supposed basal end being 1.5 cm. In diameter and the enlarged
apical end 6.5 in diameter. It is certainly apparent that a Tempskya trunk of such
an obconical form would have been mechanically incapable of attaining any appre-
ciable height, and even if it could in such a small specimen as this It must have
been dangerously top-heavy. A very likely flaw, however, in Seward's interpreta-
tion of this as a complete trunk lies In the anatomy of the specimen. He has indi-
cated (text-fig. 2, page 490) that the trunk is anatomically dorslventral, that is,
the petioles for the most part pass out toward one side. Read and Brown likewise
figure Tempskya minor as showing predominantly dorsiventral orientation of the
stems composing the trunk. We feel that it is very likely that these authors have
been dealing, in such cases, with portions of much larger trunks in reaching these
conclusions. By sectioning some 70 specimens and preparing peel preparations of
the entire transverse surface we observed the course of the stems and petioles.
These specimens have ranged from 5 to 30 cm. in diameter, and in nearly every
case, whether the trunks were circular or oval In transverse section, the orientation

of the stem-petiole organization with respect to the trunk as a whole is strictly

radial, — that is, the petioles depart toward the nearest outer 'point of the trunk.
This evidence of radial arrangement is based, furthermore, on trunk specimens
that show no indication of appreciable weathering or fracturing. There can be no
doubt that they represent complete transverse sections, with the exception of the
outermost projecting stem tips and petiole bases. Seward Indicates, moreover, that
with his specimen "the surface appears to be waterworn." Evidence from a few
specimens could thus be very misleading, and fragmentary ones must be expected

In the earliest stages of the sporeling the first formed stem may have been creeping or ascending,
but concerning this no information is available.

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 131

to display this apparent dorsiventrality. Specimens in our own collection, as well
as many in the Thomas collection, would, individually, give this impression if
complete transverse sections were not available for comparison. We do not wish
to criticize Seward's very excellent anatomical study of the single specimen he had
available, but rather we wish to point out the errors of interpretation that
may readily result from conclusions based on sucK limited material. The over-all
shape of his specimen also argues most strongly in favor of our view that it is
but a fragment, for in all of the Idaho specimens the basal portions are clearly the
largest in diameter and taper toward the apex. A more detailed consideration of
chis feature may best be saved for the following section dealing with the ontogeny
of the trunk. (Also see discussion of the living Dicksonia fibrosa on page 145).

Bower has shown ('35, fig. 296, 1930, etc.) that in the ferns the young sporo-
phytes are obconical in form, but it must be remembered that this is most apparent
during the very early stages. Generally, the stelar system soon attains a maximum
diameter as in most of the rhizomatous forms, or increase in diameter is rendered
possible by polystely of one sort or another. The ferns have been remarkably adept
at modifying their primary stelar tissues to make up for a lack of mastery of the
cambium. Such rather divergent structural types as are represented by Psaronius
and Tempskya illustrate the high state of organization that has been made possible.
In the case of Tempskya we do not know what the wery earliest stages in the de-
velopment were like but there can be little doubt that maximum individual stelar
size was soon attained and that stelar divisions started very soon after the sporeling
stage. Perhaps during the first two or three feet of vertical growth the trunks
were obconical, although it is most likely that root development soon counter-
acted this to produce a trunk that generally tapered from the base toward the apex.

The tips. — We have in our own collections three specimens of the terminal
portions of trunks, all of which (figs. 26, 27) taper rather abruptly to a blunt

apex

THE ONTOGENY OF THE TRUNK AND THE RESTORATION

With the exception of the basal ones, characterized by their anatomical compo-
sition of roots and distinctive external features, all the specimens that we have
examined display, in transverse section, stems scattered quite uniformly through
them, from the extreme periphery to the center. Some specimens show consider-
ably more stems per unit area, which is due, in part at least, to the position of the
section, whether nearer the base or apex of the trunk. One of the most striking
anatomical features is the short life span of the leaves. It is not possible to indicate
precisely how long a frond persisted, but judging from modern ferns, cycads, and
palms it probably was not more than one year. The evidence for this lies in the
fact that petioles are found only in close proximity to the stem from which they
were derived. Thus is would seem that any single stem must have been at the out-
side of the trunk, terminally or laterally, at the time it was bearing active petioles.

Previous workers have assumed that the Tempskya trunks bore a crown of
fronds at the top in a fashion generally comparable with that of a modern tree

132 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

fern or cycad. The only previously figured restoration is that of Kidston and
G Wynne- Vaughan*s which appeared in Stopes ('15) Lower Greensand flora. In
that restoration a ring of stems is shown composing a sort of compound crown
at the top. They are shown branching two or three times, each with a rather
dense cluster of leaves. From our observations, such a supposed habit is quite
contrary to the evidence afforded by the internal structure of the trunks, as well
as the external form of the tip specimens.

The tendency to compare these fossils with modern plants such as tree ferns
and cycads is understandable. However, in Tcmpskya we are dealing with a dif-
ferent mode of increase in the diameter of the trunk, as well as a highly unique
physiological set-up with regard to photosynthesis and transport of fluids. While
most tree ferns, palms, and cycads bear a crown of relatively few, large leaves the
Tempskyas bore a great many small ones, as is evinced by the minute size of the
petiole (figs. 20, 28), It is evident from the dimensions of the petioles, by com-
parison with those of living ferns, that the fronds must have been very small,
probably little more than a foot long. The physiological problem of Ught rela-
tionship where numerous relatively very small fronds are aggregated at the top
of the trunk has not previously been given the consideration that we believe
is necessary for a reasonable concept of the habit. A quantitative comparison with
plants of supposed similar habit will clearly point out the difficulty.

We have measured the trunk and petiole diameters, as well as the number of
leaves in a crown, for some cycads, palms, and a low-growing tree fern that are
cultivated in the Missouri Botanical Garden greenhouses. The data are presented
in Table II, together with those for a few representative specimens of Tcmpskya
Wesselii. A comparison of these living columnar-trunked plants, with their
crowns of leaves at the top, with the trunks of Tcmpskya reveals certain significant
structural divergences. A wide range of trunk types has been purposely included,
and of them we may immediately eliminate from close comparison those with tall
and uniformly slender trunks and a few large leaves, such as the palms Hexopction
mcxjcanum and Cbamacdorea Tepejiloie. These plants are In no way comparable to
the more massive trunks of Tcmpskya, The stouter-trunked forms such as Pbocnix
rcclinafa, and more especially P. dactilifera, present a closer structural comparison.
The frond/trunk relationship Is, however, worth careful consideration. Two
specimens of P. dactlJifera in our greenhouse measure about 12 and 14 feet high,
respectively (up to the crown of leaves), and these have diameters of 14 and 16
inches, respectively, including the very heavy armor of leaf bases, the latter ac-
counting for at least one-third of the trunk diameter. Of particular interest is
the base of the petiole which tapers from 4 inches (in its wide diameter) close to
the trunk to lYz inches through a distance of 12 inches. A generally similar or-
ganization prevails in the larger-leaved cycads, the basal portion of the petioles
being stoutly bulbous to support the weight of the leaf. In all cases we have
prepared the petiole/trunk ratio from measurements of the petiole out beyond this

1947J

ANDREWS & KERN

IDAHO TEMPSKYAS

133

TABLE II

A COMPARISON OF CERTAIX MORPHOLOGICAL DATA IN LIVING PLANTS WITH

SPECIMENS OF T, WESSELII

Living plants

Cycas niicholtzii
C. cirdnalis
C siamensh

C. rcvohtta
Dfoon spinulosum
Enccphalartos
aUcnsteinii

Ciboiiutn sp.
Phoenix reclinata

P. dactilifcra

Hexopction mcxicannm
Chamacdorca Tcpejilote
Thritjax parti flora

Caryofa nrcns

Ratio of petiole

diameter to trunk

diameter

1:13

1:11
1:9

1:24

1:9

1:10

Fossils

Tetnpskya Wcsselii

specimens
T18
T3 3
T51
T90

1:11
1:4

1:10

1:4

1:3

1:2.5

1:2.5

Number

of leaves

in a crown

15

22
40
'16
40
12

27

28

10

4

15

5

Remarks

Very little taper to trunk, which is en-
closed in dense armor of leaf bases.

Trunk encased in very dense armor of
leaf bases.

No appreciable change in diameter through
Its 10 feet of height.

Very slender clean trunk with little
change in diameter through its 8 feet.

Gently tapering trunk, clean below with
bulbous base 5 in diameter which tapers
to about 3" at departure of first leaf.

Trunk ensheathed with closely appresscd
leaf bajes, uniformly tapering from 5

at ground to 3" where first leaf is given
off at height of 6'.

great basal swelling. Turning to the Tempskyas, in specimen T3 3 the mean ti-unk
diameter is 170 mm. and the petiole diameter (taken immediately after the de-
parture of the petiole from a stem) Is 2.7 mm., the petiole/trunk ratio being 1:63.
In comparison with a stout columnar cycad such as Cycas sia?nensis with a trunk

J

10 inches in diameter (including the leaf bases) and a height of 9 feet, with a
petiole 2 inches in diameter close to the base we have a corresponding ratio of 1:5,
or when the petiole diameter Is taken out beyond the swollen base, a ratio of 1:9
as shown in the table.

The larger-leafed cycads and palms have, as might be expected^ an exaggera-
tion of this enlargement of the petiole base to support the weight of the fronds.
The relative size of this bulbous base is noticeably smaller In the smaller-leafed
species.

Since there Is no evidence to indicate that the petioles of a Tempskya Increased
appreciably after their departure from the stem, they must have borne relatively
very small fronds. The actual diameters of the petioles in a number of trunks of

[Vol. 34

134 ANNALS OF THE MISSOURI BOTANICAL GARDEN

T. Wessclii immediately after their departure from the stem varies from 1.5 to
2.75 mm. Dimensions of the petioles of cycads (taken immediately beyond the
swollen base) range between 8 and 25 mm., those of Cibotium sp., 13 mm., and
of the palms from 7 to 40 or more mm. Since these figures for the living plants
are taken beyond the bulbous base, and the trunk diameters include the leaf base
armor, the recorded divergence between their ratios and those of the Tempskyas
IS an extremely conservative one.

The Stopes restoration is ingenious In that it allows for a considerable prolif-
eration of the leaf-bearing area. However, the actual terminal trunk specimens
do not suggest any such appearance. All the specimens of this nature that we
have observed, representative ones being shown in figs. 26 and 27, indicate a
rather bluntly tapered apex like that of a living cycad, fern, or palm. In view of
the generally good preservation of the Wayan Tempskyas, most of which show no
great wear due to transport either before or after fossilization, we should expect
to find some evidence of the stems or at least the rather massive stem aggregates
as shown in the Stopes restoration. Such evidence is quite lacking,

A point that we wish to make is that the apex of a Tcmpskya probably could

not have borne sufficient photosynthetic surface to have satisfied the requirements
of the plant.

From the evidence afforded by internal structure, one of the most striking
features of all Tcmpskya trunks (excluding the basal portions) lies in the fact
that in any transverse section free petioles are rarely found more than a few milli-
meters beyond the stem that bore them. In other words (as previous authors have
pointed out), the leaves were not persistent for any great length of time.

In order to arrive at a clearer understanding of the mode of lateral growth in
Tempskya we have: first, observed the stcm-pctiole-root organization in transverse
section In many trunk specimens, varying from approximately 2 to 10 inches in
diameter; and, second, followed the course of individual stems in single specimens
by means of serial sections. This latter procedure is considered in some detail in
the following paragraph.

Two specimens, each approximately 4 inches in diameter/ were cut into a
series of thin slices in order to determine the extent of branching in the individual
stems, their destination, and the number of petioles that depart through a given
length. The branching of the stems proved to be so frequent that the slices had to
be taken between ^4 ^nd Yz inch apart in order to follow them with certainty.
With reference to stem branching, Read and Brown noted: "The writers at-
tempted to determine the distance between these successive bifurcations by cutting
a block several inches long into serial sections, but they found that this character
is so variable that it has little value either for morphologic or for taxonomic con-
siderations." (p. 110). While agreeing that the taxonomic value of stem branching

*^Small specimens were selected for this purpose because ^ the extreme diiTiculty of cutting the
larger trunks. From a comparison of many specimens varying in size from 2 to 12 inches in
diameter there can be no doubt that the branching as described (based on specimens T5 1 and T90)
is representative.

1947]

ANDREWS & KERN — IDAHO TEMPSKYAS

135

may be negligible we shall try to point out
that It is of the utmost significance in an
interpretation of the general habit and phys-
iology of the plant as a whole.

It is evident, even without making serial
sections, that branching of the stems is
very profuse, for in almost any single trans-
verse section of a complete trunk a consid-

F

erable number of stems may be seen to be
dividing. Taking more or less at random
complete transverse sections from ten dif-
ferent specimens an average of 45 per cent
of the stems was observed to be branching.
Thus the serial sections, upon which text-
fig. 3 is based, serve to confirm a three-
dimensional picture that might have been
prepared In a somewhat less exact fashion
from a single transverse section. The worth
of the peel method has proven an invaluable
aid in anatomical studies of this sort. We

have not relied upon it exclusively, but it

Is the only feasible way In which one can
prepare complete sections, and often ex-
cellent ones, of trunks up to 1 and 1 2
inches in diameter.

Previous accounts of the stem branch-
ing in Tempskya have indicated it as being
dichotomous, and while this is predominant-
ly the case it is not always so. In some
specimens there may be appreciable variation
in stem diameters as is shown in figs. 20,
21, and 22. .

In order to present a three-dimensional

aspect of stem branching we have selected

r

several stems from specimen T90 which
have been followed through a distance of
4.5 cm. (text-fig. 4). The average distance
between successive divisions is approxi-
mately 1.5 cm. This abundant branching
activity must result, through any appre-
ciable distance. In either a great congestion of stems, or one of the two divisions
soon ceases to grow. The latter Is observed to be the case.

Tcxt-fig. 3. A diagrammatic drawing
of a transverse section of specimen T2
(peel 13) showing radial symmetry in the
departure of the leaf traces In a strongly
flattened specimen. Natural size.

[Vol. 34

136

ANNALS OF THE MISSOURI BOTANICAL GARDEN

This frequent cessation of growth of many of the stems and the continued
growth in the trunk as a whole, by the branches, present a distinctly different
type of organization from that of other living or fossil plants with which
Tempskya may be compared. In Tempskya the trunks Increase in diameter as
well as in height by this same process. While we cannot follow the growth stages
of a single specimen in plants that lived 100 million years ago, we can arrive at an
explanation of this developmental anatomy by observing trunks of varying sizes,
and in all of them it seems clear that increase in diameter has taken place by the
continued division of stems at the periphery of the trunk. In virtually every
cross-section of a disc specimen stems may be observed (text-figs. 2, 3, 5) ac-

l//

Hit

Gu ^

r

II

Cm --

Da

C,

3

tf

n ,

^^^fa^

Text-fig. 4. A three-dimensional aspect of stem branching and the production of petioles as
obtained by following the branches by means of serial sections, through a distance of 4.5 cm., of
specimen T90, Each small division on the horizontal and vertical scales represents 1 mm. In their
proper positions on the scale, letters to the left indicate the section, and numbers Indicate the peel

of that section used In constructing the illustration.

tually departing from the trunk. J

proper

sport

It is difficult to observe these stems on the

outside of the trunk because they departed obliquely and are not readily dis-

roots

It is

probable that they did not extend more than a centimeter or two at the most, as
a greater length would have resulted in excessive crowding of the foliage.

peno

We

1947]

ANDREWS & KERN IDAHO TEMPSKYAS

137

Text-fig. 5. A diagrammatic drawing of a transverse section of specimen T226 (peel 3) show-
ing radial symmetry in the departure of the leaf traces in a cylindrical trunk. Natural size.

grams given in text-fig. 6 this apparent mode of growth. As stated above, we have
no knowledge of the earliest development of the sporeling but it is evident from
the smaller specimens that profuse branching of the stems was initiated very early
in the development of the trunk. Just how long the original stems retained direct
continuity with the ground likewise cannot be determined. The decay of the
stems in the lowest portion was, of, course, gradual and the apical growth con-
tinued at a considerably faster pace. However, taking as an example a plant with
a basal diameter of 10 inches and a height of 12 feet, it is probable that the lower
2 to 2.5 feet of the trunk was composed of roots, (see extreme right diagram in
text-fig. 6).

[Vol. 34

138

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Text-fig. 6, A hypothetical series of longitudinal sections through Tempskya trunks showing
the supposed mode of growth. The solid lines represent stems, and the dotted lines stems in various
stages of decay.

In summary, tKese seemingly important points may be emphasized: The
trunks had a generally tapering form from base to apex; branching of the stems
was profuse and apparently uniform throughout the life of a plant, producing
lateral as well as apical growth; leaves were not long-persistent and their small
size would not have afforded sufficient photosynthetic surface as an apical crown
alone.

Thus we feel that tbe Tempskya plants appeared in life as indicated in the
accompanying restoration (text-fig. 7). While, as noted at the outset, this is based
on specimens that we have assigned to Tempskya Wcsseliij the close anatomical sim-

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 139

ilarity of all of the described species of the genus suggests a generally comparable
habit. We wish to note that the kind of foliage borne by these plants is not known.
Although the leaves shown in the restoration are of the Anemia type this does not
necessarily imply such a relationship. They have been used to indicate rather the
approximate size and dhtrihnition of the foliar organs.

THE ROOTS

The roots of the Tempskyas have been described by previous investigators in-
cluding Kidston and Gwynne-Vaughan (*11), Seward ('24), Read and Brown
('37), and Arnold ('45). A rather close similarity of the anatomy of these organs
has been noted in the descriptions of the various species, and we are in agreement
with these previous workers in that the roots seem to offer no recognizable specific
characters. There are, however, certain points pertaining to their physiology that
seem deserving of further consideration.

If it is kept in mind that the basal portion of the trunk of the mature plant
is composed entirely of roots, and the trunk at any point is composed largely of
them, it is evident that they played a more than ordinary role in the absorbing,
conducting, and supporting functions of the plant. A detailed study of these
roots presents some rather challenging problems, and the literature of comparable
physiological set-ups in living plants is by no means a copious one.

The roots, like the stems, branch profusely and present in any section con-
siderable variation in size, degree of maturity, and preservation. The stele is
small and, like the stem, consists only of primary wood. In well-preserved speci-
mens (fig. 13) the phloem and endodermis are clearly defined. It is not often
possible to identify positively any tissue that may be called the pericycle. In fig.
13 a thin crushed row of cells, apparently the pericycle, may be noted between

the endodermis and the large metaxylem tracheid in the upper-right portion of the
figure. Very early in the maturation of the extra-stelar tissues a conspicuous
fibrous middle cortex is developed. This tissue is extremely variable, at times ex-
tending to the endodermis (fig. 13), and its development is accomplished to a
considerable extent by abundant radial cell divisions which result in a distinctive
tangential alignment of the cells (fig, 19). Taken by the thousands it is evident
that such structures would develop a trunk of great strength — certainly not a
brittle one — yet the close organization of roots suggests one of considerable
rigidity. Without this sclerotic cortex is a rather broad, thin-walled outer cortex
(figs, 11, 17), in most cases largely decayed, the outer remnants of it forming a
collapsed loop which encloses the tissues.

Although the roots in general average a little less than 1 mm. in diameter
there Is considerable range in size. Within an area of a square centimeter roots
may be found that are less than .5 mm. and others nearly 2 mm. in diameter. The
smallest of these (mature) roots may have a middle cortex consisting of only two
rows of the very thick-walled fibrous cells, a row or two of large, thinner-walled
outer cortex, and epidermis.

The nature of the outer cortex may have some bearing on an interpretation of

[Vol. 34

140

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Text-fig. 7. A restoration showing the probable habit
of Tempskya, bascJ on specimens of T. Wcssclti, Foliage
has not been found attached to the trunk. For further
explanation see text.

1947J

ANDREWS & KERN IDAHO TEMPSKYAS 141

the compressed forms in which the trunks often occur. It is evident that many of
the roots in a normal trunk, prior to removal of the plant from its place of growth
or fossilization, were largely decayed. Just how much decay did take place im-
mediately before siHcification cannot be determined and certainly varied with
different specimens. Certain parts of specimen T53 are exceptionally well pre-
served, and within a small area (fig. 14) some of the roots still retain the most
delicate tissues while intimately associated with them are others with nothing but

the sclerotic cortex intact. Some of the former are young roots in which little
thickening of the sclerotic cells had occurred and they show root hairs as well;
the latter may be interpreted as older roots that had been dead for some time.

In contrast to fig. 14, the majority of sections show very few roots in which
the outer, large, thin-walled parenchymatous tissue is preserved. Here and there
a root may be in almost perfect preservation, and less occasionally a considerable
group will be well preserved. Others have the inner cortex and stele intact and in
still others the stele Is missing, as well as the remnants of the outer cortical loop.
Many of them reached this stage through death and decay during the normal life
of the plant. If, however, any appreciable number of roots were alive and active
at a given time (as must have been the case) a considerable percentage of the area
of the transverse section was composed of this large, thin-walled, readily decayable
tissue. It would seem, then, that the general decay of this tissue immediately prior
to fossilization would have allowed even a relatively slight lateral pressure of
overlying sediments to have compressed the trunks.

A comparison of circular and variously flattened specimens has been made in
order to determine the mechanism of flattening. However, no differences were
ever observed that might point positively to mechanical crushing. It may also be
noted that in specimen T53, where the preservation of immature roots is so per-
fect, there is no marked indication of distortion. If we should assume that a
large percentage of the roots was actively functioning and with their outer cortex
intact the roots would have been so closely compacted as to have been strongly

angular in shape. Yet where an appreciable number of roots happen to be well

preserved in a small area (fig. 14) this is not the case. There is some compaction
due to crowded growth but it is not excessive. It may also be noted that in most
areas the roots, or remains of roots, are so crowded that there could not have been
room for them all to have existed with their outer cortical tissues at one time.
The evidence therefore indicates that only a portion of the roots composing a
trunk was active at a given time. In this connection it may be noted that Schoute
found very few live roots of the many composing the dense matrix of the trunk of
Hetnitclia crenulata (see page 144).

In his treatment of Tempskya Knowltoniy Seward ('24, p. 494) makes the fol-
lowing pertinent remark: *'The contrast between the large number of roots with-
out any visible connexion with their mother-organs and the small number of
which the origin is demonstrated is remarkable. Most of them must have come
from stems or leaf-bases that are unrepresented in the specimen,"

[Vol. 34

142 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Unfortunately, this evidence does not indicate positively whetlier the flattening
of the trunks was natural or the result of the pressure of overlying sediments prior
to fossIli?:ation. In summary we are inclined to believe, however, that the flattened
form is not natural for the following reasons:

1. In any trunk there was, in life, a mixture of live, active roots as well as

others in various stages of decay.

2. This organization allowed sufficient "inter-root-stem" space for the re-
organization of these elements when crushed by overlying sediments without
presenting a recognizable microscopic effect.

3. No other observable characters correlate with the wide range in transverse

shape.

4. Circular and flattened trunks alike in our collection display a symmetry

that is radial with reference to the departure of leaf traces.

5.

spec

rms

Root hairs. — In many of our specimens we have been able to observe well-
preserved root hairs, some of which are very long, as shown in fig. 12. These root
hairs are often found on the larger roots which, judging from their size and gen-
eral maturity, would seem to indicate at some appreciable distance back from the
apical meristcm. That these hairs occur on the older portions of the
moreover, adequately attested by the fact that they may be observed on numerous

roots

roots

be

root hairs produced at an earlier growth stage of the root or that they are simply
developed from the epidermal cells throughout the life of the root. There is
evidence to indicate both modes of origin. The root shown in fig. 11 represents
one of average size and certainly mature. Numerous short root hairs may be
noted. These are complete hairs, as evidenced by the uniformly rounded tips, and

not simply broken remnants.

In describing living specimens of Dicksonia fibrosa in New Zealand (see
page 145) Field ('90) refers to the absorbing capacity of the aerial roots. In
Tempskyas this function must have been of considerably greater importance than
in most of the living ferns where the stem stelar system extends down to ground
level. In the larger Tempskya trunks (10 to 12 inches in diameter) the stems
had died away from the lower two feet, and possibly more, of the trunk. That
portion of the trunk above this "root-stump" depended upon its water supply,
then, either through the long slender roots reaching down from the stems or, more
directly, through rain water absorbed from the apical and lateral surface of the

:. There would seem to be little doubt that most of the minerals were taken
up from the soil through the length of the trunk. However, the external surface
of the trunk throughout must have been very absorbent and the trunk itself
capable of retaining considerable moisture. The fact that deep within the trunk,
roots are found with root hairs intact would seem to indicate that they functioned
thus in drawing off this water reservoir.

trun

1947]

ANDREWS & KERN

IDAHO TEMPSKYAS

143

COMPARISONS WITH OTHER PLANTS, LIVING AND FOSSIL

The morphology of the trunk of Tempskya represents a peak of structural
evolution along a line that is manifest in a generally comparable fashion in a
number of ferns, both fossil and living. The ferns have developed some very
remarkable and ingenious devices in the organization of their primary tissues to
bring about increased size. Why they have never been successful in the use of a
cambium is a mystery that we may never know, yet their other modes of develop-
ment have been successful to a considerable degree and are none the less interesting.
We have no reason to believe that any of the other plants mentioned below are
closely related to Tempskya, yet they seem significant in offering clues to the
racial origin of the Tempskya trunk (or *'false-stem").

Scattered rather sparsely through the literature there are references to a num-
ber of living ferns having upright trunks composed of branching stems which are
held together to a greater or less degree by a mass of adventitious roots. Since
some of these references are extremely interesting, yet obscure and mentioned
only briefly, if at all, by previous writers, a few of the most pertinent ones will be
considered in some detail. In his "Ferns of New Zealand" H. C. Field ('90)
presents a number of interesting accounts of ferns with stems of upright habit:

"'Rhizomes are of various kinds. The simplest form is that which grows 'erect' and
produces its fronds in a crown or tuft at the top; in which case the plant is called a
'crowned* or 'tufted* one. Tn many ferns, this erect rhizome is prolonged above ground to
a great height, as in tree ferns, and it is then called a *caudex.' This caudex , • , is always
clothed with fibrous rootlets by means of which moisture is imbibed from the atmosphere
and helps the upward growth of the plant. The number of fronds which form the crown
of the plant depends very much on the number of these aerial root fibres, our Dicksonia
fibrosa, for instance, in which the actual caudex is only about two inches thick, while the
fibres form a felted or interwoven mass, sometimes two feet In diameter, having often as
many as forty fronds in its crown. Some caudices have a large conical base of root fibres;
and in others this cone extends to the very top of the caudex, which is then of no great
height, the whole mass being called a 'rootstock.' Sometimes an erect rhizome or caudex
becomes flattened at the top and produces a great number of fronds, while at others it
becomes divided into two or more branches, each of which produces a separate crown; and
occasionally fresh crowns burst out of the side of a caudex. It is not quite certain how
these originate. In some cases it appears to be the nature of the fern to divide itself in this

fashion; in others, it seems as If young plants had grown on the face of a caudex; while
in others it occurs by accident. I watched a case where a falling tree strained a supplejack
tightly across the crown of a tree fern; with the result that the next spring the plant pro-
duced two crowns, one on each side of the supplejack, and thenceforth was "forked. I have
seen a cyathea dealbata with five branched caudices and crowns, and an aspidium aculeatum
with seven, varying from three feet to fiye feet high," [pp. 11—12.]

From the point of view of comparison with Tempskya perhaps the most in-
teresting of these extant New Zealand ferns is HemiteUa Smithii. This tree fern,
which is known to attain a height of 20 to 30 feet, often divides at the tip into
two or three branches. In 18 86 Buchanan described a remarkable specimen from
the slopes of Mount Carglll, near Dunedin. Buchanan's sketch of this tree-fern
is reproduced in pi. 26. It is reported as being 16 feet high and with 16
branches and several buds. "The budding and branching may proceed from any
part of the stem, and the specimen has several branches diverging in various direc-
tions, which again divide, as in dicotyledonous trees." According to the author*s

[Vol. 34

144

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Text-fig. 8. Spores and fragments of annuli found in a
ground thin section made from specimen Tl6 (slide 1400).
X 220.

brief description and figures, the branching is not dichotomous but rath

er "a

oody

inner or true stem of the tree, and also the close fibrous outer covering." Shortly
after the division of the single stele in the main trunk of this plant the resultant
branches become separate as shown in the sketch. It thus differs from Tempskya
in which the branches (except the extreme terminal portions) are permanently
held together by the dense mass of roots. An interesting point of comparison,
however, is that in Hemitelia Smitbii the crowns of leaves on the respective branch-
es are borne at different levels, thus differing from other tree ferns with "divided
crowns," but not entirely unlike our concept of the mode of habit in Tcfnpskya.

Another curious living fern is Hemitelia crcnvlata Mett., from the forests of

Kandang Badak, Java, which was described by Schoute in 190^. It is a "tree fern

f «

1947J

ANDREWS & KERN IDAHO TEMPSKYAS 145

(pi. 25) of considerable size, attaining a circumference of 201 cm. at 30 cm.
above the ground. The basal two or three feet of the trunk consist of numerous

branches enclosed in a dense matrix of roots, while above this the branches are
free. Ji^dging from Schoute's illustration of a specimen with a man standing
beside it the plant attained a height of at least 12 to 15 feet. At ground level
the trunk of HemitcVta crcnulata contains but one stem while about 19 cm. above
this it branches into three, while 28 cm. higher up seven branches are found. In
the largest specimen reported, 3 3 branches are displayed which bear leaves in a

r

crown as well as laterally. As may be noted, the branches grow horizontally out
for a short distance and then ascend sharply.

The stem of Todca barbara Hook. f. seemingly presents an organization that is
comparable to the above two ferns although we have been able to find but little
information concerning the gross morphology of the trunk, most of the accounts
dealing only with the cellular structure of individual stems, petioles, etc. Seward
and Ford ('03) give the following account although it is not as pertinent to the
present discussion as one might wish:

*'The seem of ToJra barbara may reach considerable dimensions, forming a short and
thick mass covered with a dense felt of brown roots, which completely hide the main bifur-
cated axis. One of the numerous plants of T, barbara sent to Europe by the late Sir Ferd.
von Mueller has been figured, in which the stem reaches a breadth of 2.5 metres, a height

of 1.76 m., and a thickness of 1 m. J. Smith also described a specimen from the Victorian
Alps of Australia measuring 5 ft. 8 in. in height, with a diameter of 7 ft. 9 in., and weigh-
ing I ton 3 cwt.; he adds that a plant was received at Kew in 1869 weighing 15 cwt. and
bearing 3 crowns and 160 fronds. The stem of a Todca barbara in the Cambridge Botanic
Garden measures 8 ft. in circumference and 3 ft. in height, with 14 distinct 'crowns'; at
the present time the crowns bear 23 fronds, with an average length of 7 ft. 6 in.'* [p- 239.]

From the large number of crowns that these Todea plants bear there must be
rather profuse branching of the stems composing a trunk. Unfortunately, we
know very little about this, for, as Sahni ('28) notes: "The mode of formation of
the false stem still needs elucidation."

Field's ('90) account of Dicksonra fibrosa is also worth quoting. He writes:

"The caudcx seldom, if ever, attains the height of 25 ft., but is extremely stout in ap-
pearance owing to the mass of matted fibrous aerial roots which envelops the actual caudex
and which is often 15 in. to 18 in._ in diameter, and occasionally even more. Curiously
enough, ic is often larger In diameter above than below, particularly in plants not exceeding
6 it. or so in height; which shows to how great an extent this fern absorbs nourishment
from the atmosphere by means of its aerial roots."

This is the only case of obconical trunk shape, such as Seward reported for his
single specimen of Tempskya Knowltorii^ that we have come across in living ferns
(other than in the very young stages). And, judging from Field's statement, it
is not found in a completely mature plant. In view, then, of this data on living
ferns and the fact that none of the hundreds of specimens of Tempskya from Idaho
suggest such a trunk form it seems reasonably certain that Seward's specimen is
either a very exceptional one or that it represents a fragment of a larger trunk.

The only fossil plant that seems to merit comparison with Tempskya is the
Carboniferous zygopterid tree-fern from New South Wales described by Sahni

146 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

/

I

('28) as Clepsydropsis aus trails. Later, he ('32) gave reasons for its exclusion
from that genus and proposed the name Ausfroclcpsis. So far as we are aware its
present valid name, then, is Austroclepsis aiisfralis (E. M. Osborn) Sahni.

It should be noted that there is no close genetic relationship between A.
ausfrafh and the genus Tempskya for their individual stems and petioles have a
strikingly divergent anatomy. Such resemblance as may be observed is only In the
general habit of the trunks. In A. atistralis the petioles arc large, almost equalling
the stems in diameter; many petioles appear in a transverse section, indicating that
they were long and the fronds persistent for some time; furthermore, the divisions
of the stems are not nearly as profuse as in Tempskya. Thus the general appear-
ance of Austroclepsis with its single crown of rather large fronds (Sahni, *32, text-
iig. 14) must have been quite different from that of Tempskya. These two genera,
widely separated in time and space as they were, possibly present in the gross
organization of their trunks a similar evolutionary trend such as is exhibited in
the above-mentioned living ferns, and it is a trend that probably took place in-
dependently in a number of different groups of ferns at different geologic times.

TAXONOMIC CONSIDERATIONS

■ s

As stated at the outset, our primary Interest in this Investigation was with the
general habit, mode of growth, and physiology of the Tcmpskyas. A few points
may be worth mentioning, however, In order to clarify our own .taxonomic treat-
ment as well as to offer suggestions that may be of use to future Investigators. A
careful identification of the specimens has, of course, been basic, and while we
have assigned the name Tempskya Wessdii to most of our specimens other ob-
servers might find cause to split these Into more than one species. Such "species"
would be of very doubtful significance, and it seems certain that concepts pertain-
ing to the life form and functions of the plants would stand unaltered.

I

Since the roots have presented no recognized taxonomic characters the identi-
fication of stump specimens can be made only to the genus. However, In view
of their constant and uniform association with the trunk (disc) and tip specimens
in the Wayan region there can be no doubt of their identity.

In view of the lack of evidence of natural affinities of the genus, Read and
Brown ('37) created the family tempskyaceae, which constitutes the most ex-
pedient treatment. The only disappointment in our own investigation has been
that such spores and sporanglal fragments as were found in the trunks offer no
positive help.

Prior to the work of Read and Brown eight species of Tempskya had been de-
scribed: from Sussex, England; the basin of the Karaganda River In Russia;
Bohemia; Maryland; and Montana. Many of these, because of their fragmentary
nature and poor preservation, are certainly not worth further consideration and
In their Syfwpsis Read and Brown have dealt with only two of them — Tempskya
rossica Kidston & Gwynne-Vaughan from Russia, and T. Knowltoni Seward from
Montana, and they have added two species, Tempskya grandis from Wyoming and

194 7]

ANDREWS & KERN IDAHO TEMPSKYAS 147

T. minor from Wyoming and Idaho.

More recently Arnold (M5) has described two more species: T. wyomingcnsis
from "about twenty miles northeast of Greybull, Bighorn County, Wyoming, and
r. Wesselii from Greenhorn, Oregon, and Great Falls, Montana.

Since it is perhaps most expedient to the present discussion we present the
Synopsis of Read and Brown, to which we have added Arnold's two species in
accordance with his concepts of their relationships.

SYNOPSIS'^

1. Individual stems of false stem large, with very short internodes as
indicated by the numerous leaf bases present in transverse sections.
Xylem exarch or possibly slightly immersed in some specimens. False
stem chiefly radially symmetrical. Xylem ring containing much

parenchyma,

A. Inner cortex a broad zone of parenchyma containing near its inner

margin an Irregular but continuous tract of sclerenchyma. Outer
layer of "pith" a similar zone of parenchyma, containing scleren-
chyma, especially in the vicinity of the nodes Tempskya grandis

B. Inner cortex a narrow zone of large-celled parenchyma. Presence

of an Inner sclerotic layer not recorded Tempskya rossica

C. Inner cortex with two bands of stone cells. Smaller stems and long-
er Internodes than the above two.... Tempskya Wesselii

D. Individual stems large (6-8 mm. in diameter), very close to T.
grandis, differing chiefly In having a double layer of stone cells;

internodes and stems larger than in T. Wesselii Tempskya wyomingcnsis

2, Individual stems of false stem small, internodes of such length as to
permit only a little overlapping (2-3) ,of leaf bases. Xylem exarch.
False stem dorsiventral. Xylem ring containing little, if any, paren-
chyma.

A. Xylem very compact; protoxylem commonly segregated into definite

groups. Inner cortex broad, parenchymatous. Petioles common in
false stems, indicating persistence of leaves; xylem arch, fairly
flat. Stems averaging larger than those in the next group Tempskya minor r

B. Xylem compact but with parenchyma In places interspersed with
the trachelds. ' Inner cortex usually narrow, parenchymatous. Pet-
ioles rare in false stem; xylem arch, rounded : Tempskya Knowl tain

*From: Read, C. B. and R. W. Brown: American Cretaceous ferns of the genus Tempskya.
U. S. Geol. Survey Prof. Paper 186-F, p. 119. 1937, except parts C and D under I, which have
been abstracted from: Arnold, C. A.: Sillcified .plant remains from the Mesozoic and Tertiary of
western North America. Mich. Acad. Sci., Arts, and Letters 30:24-3 3.

Read and Brown have considered the gross organization of the trunks, that is
whether radially or dorsiventrally symmetrical, of importance, while Arnold
places considerable weight on the structure of the cortex in the deUmitation of
species.

In almost all cases where it is certain that we are dealing with complete trans-
verse sections the symmetry of the trunks is essentially radial. In order to present
some quantitative evidence of this we have taken a representative sampling of
specimens displaying different shapes and noted the direction of departure of the
petioles (Table III). Tliese show quite clearly that the departure of the petioles
is usually toward the nearest periphery of the trunk, as shown in text-figs. 2 and 5.
It is not surprising that an occasional stem should bear its leaves toward the center
rather than the periphery of the trunk but this does not necessarily point toward

148

[Vol. 34

MISSOURI

TABLE III

COMPILATION. FROM A KKPRESENTATIVE SELECTION OF SPECIMENS, OF STEMS.

CONTRIIIUTING TO RADIAL OR DORSIVENTRAL SYMMETRY OF THE TRUNKS.

THE FORMER BEING STEMS IN WHICH THE PETIOLES PASS TOWARD THE

NEAREST PERIPHERY OF THE TRUNK. THE LATTER THOSE WHICH

ARE IRREGULAR IN THIS RESPECT

Specimen and
peel number

Shape and

dimensions in

transverse section

(inches) *

Total

number of

stems

Ratio of dorsi-
vcntral to radially
symmetrical stems

Percentage of

stems contribut-
ing to radial
1 symmetry

T205, Bl

E 3 x2

57

1:50

— •

98

T2, T5

E 53/4 X VA

50

3:38

95

T47, TI9

E 53/ix 3/2

134

0:114

100

T3. T3

E 3 ^2 X 2 1/2

63

5:58

92

T90, AT8

E 3/2 X2J/2

67

3:54

95

T51. GBTI

T 4x 3

74

5:56

92

T20U Al

E 23/4 X 2

53

3:47

94

T202, Bl

T 3 X 1 J/2

27

2:23

92

T5, T2

c 2V2

20

2:18

90

T4, Tl

E 3/2x23/4

70

1:55

98

Tl7. T3

c 23/4

52

7:47

87

T53, T3

E 6/4x2/2

96

1:81

99

T45. T5

E 33/4 x3

: 60

6:48

89

T42. T7

E 4x3/2

68

5:58

92

T33, T7

E 73^x4

158

3:122

98

*E, elliptical in transverse section; T, triangular; C, circular. The stems in the central third
(diameter) of the trunks have been omitted since these are more variable in the direction of
departure of the petioles and represent for the most part the terminal stems of the trunk.

asymmetry of the trunk as a whole. Since the stems probably projected out a
short distance beyond the trunk proper it is quite conceivable that they could have
given off leaves on the Inner as well as the outer side.

In only one of our specimens is there a tendency for the petioles to pass pre-
dominantly in one direction. This Is a small specimen, or quite possibly a portion
of a larger one. It must be admitted that while our own observations are based
on a large number and variety of specimens their geographical distribution Is not
great. We feel inclined to predict, however, that when comparable collections are
gathered from other regions and when it is certain that the individual specimens
are complete in transverse section, they will reveal radially symmetrical trunks.

With respect to the anatomy of the cortex, we have noted under the "Tech-
niques" section differences that may be encountered in a study of identical stems

of a trunk using peel preparations vs. ground thin sections. Also important is
th

iatlon in this character (particularly the apparent presence or absence of
the stone cell band on the inside and outside of the inner cortex) that may show up
in consecutive peels taken from the same specimen. With the proper etching
time the sclerotic bands stand out in striking contrast to the parenchymatous
tissue; but where the time Is too little or too great the contrast may be much less
apparent, or even barely discernible.

These differences are pointed out as neither an indictment against the methods
themselves or the results of previous workers but to show that there are two points
of the utmost importance in studying the genus Tempskya:

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 149

1. Specimens displaying complete or nearly complete transverse sections of
the trunk are absolutely essential to taxonomic and morpKologic studies.

2. Great caution must be exercised in delimiting species with reference to the
way in which sections are prepared.

We feel that our own investigation has shown the probable lack of taxonomic
significance in the gross symmetry of the trunks as used in the preceding synopsis.
Furthermore, in view of the rather close gradation of the other characters that
have been used to delimit species, as well as the fact that some have been based on
rather scanty material, there is considerable doubt whether Tempskya minor^ T.
KnowUonij T. Wesselii, and T. wyomingensis represent distinct species. We do
not wish to carry this taxonomic problem further but as our knowledge of
Tempskya grows it is likely that rather extensive revisions will be necessary.

^

pore

pskya identified as T, Schhnpe

Wight. Th

tetrahcdral," with a spore wall conspicuously sculptured with long bars. He
noted a close comparison of these spores with those of the living Anemia efcgans

Ceratopteris thalictroides (Park-

eriaceae) has spores with a similar sculpturing.

Knowlton

the annuli of a Schizaeaceous fern, and probably belong to the genus Anemia. It
is largely on account of these reproductive organs that Tempskya has been thought
to be of Schizaeaceous affinities.

The only contribution that we have been able to make in this direction per-
haps confuses the picture more than it helps to clarify it. Ground thin sections of
two of our specimens (Tl6 and T3) contain considerable numbers of spores and
fragments of the annuli of fern sporangia- TTie spores occur as isolated individuals,
as well as aggregations of about 100 to 200, and in all cases the exine appears quite
smooth. They are mostly collapsed, but a few appear triangular to slightly
elongate (text-fig. 10 and fig. 23), these averaging about 50 x 40 /jt. The spore
shown in fig. 23 measures 51 x 3 6 /x.

The lack of any sculpturing of the spore wall and the occurrence of spores In
large masses might lead one to infer that they are simply immature. However,
since the exine is smooth in all of them, and an appreciable number is scattered
about as individuals, it Is probable that this is a mature character- For whatever
the comparisons are worth it may be pointed out that there is a general lack of
surface ornamentation throughout the genus Gleichcnia (Gleicheniaceae) , while
Thyrsopterh elegans (Dicksoniaceae) has spores that are likewise smooth-walled
(Knox, '39) and compare closely with our fossils.

Associated with these spores are some annulus fragments (text-fig. 10).
Speculations on these fragments can lead to no definite conclusions although it
seems clear that they are not Schizaeaceous, nor Is it likely that they represent the

[Vol. 34

150 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Osmundaceae or Gleicheniaceae. A much closer comparison seems to be afforded
by the modern Polypodiaceae or possibly the Cyatheaceae-Dicksoniaceae.

Wq have chosen to refer our specimens to T. Wesselii Arnold because they
compare most closely with the published description of that species. For those
who may be especially interested in Tempskyaj as well as for the purposes of
record, we have compiled short descriptions of a representative selection of speci-
mens of T. WcsseJii in our own collections, pointing out especially distinctive
characters that the respective specimens present:

Spechncn TjJ, — A disc specimen measuring nearly 8 x 4J/2 inches in transverse
section and € inches long. This seems to be representative of the larger, flattened
trunks. Approximately 160 stems are found in a single transverse section, of
which about 74 are shown in various stages of branching. The individual stems
average 3.5 mm. In diameter, most of them being fairly close to this figure, al-
though a maximum variation of from 1.5 mm. to 5 mm. In diameter may be found.
The steles in the peripheral inch of the trunk are, for the most part, appreciably
better preserved than those deeper within. Such differentiation of preservation is
not as noticeable in the smaller specimens.

specimen T2. — A much-flattened tip specimen (fig, 26 and text-fig. 3) ap-
proximately 6^x1^ inches in diameter and 2^ inches long. Of the 57 stems
shown in transverse section 18 are branching. The stems average 3.5 mm. in
diameter, with a maximum variation of 2 to 4 mm. The preservation is uniform
throughout the specimen, as might be expected in a small trunk tip.

Specimen T53. — A flattened, nearly oblong-shaped disc measuring 6^2 x 2^
Inches in transverse section and 2 J/2 inches long. This is unusual only in the mode
of preservation, the smoothed surface (and peels) presenting a characteristic
blotched appearance due to the irregular quality of preservation.

Specimen T4y, — A somewhat flattened tip specimen 9 inches long, in which
a transverse section 5 inches from the top measures 6x3^/2 inches. Of the 134
stems in one transverse section, 45 are shown In various stages of dichotomizing.
Each stem is giving off petioles toward the nearest point on the periphery of the
trunk, thus producing perfect radial symmetry. Stems In the central part of the
trunk are shown in perfect transverse section in any peel preparation, whereas
those toward the periphery are shown, In the same peel, In somewhat oblique sec-
tion, indicating that the stems at the edge of the root mass bend slightly outward
while those In the center of the trunk keep a more or less vertical course.

Specimen T216. — A disc specimen 13 Inches long which Is somewhat triangular
in transverse section. The transverse section at the bottom of the specimen
measures 6^ x 5 J/4 inches, while at the top it measures 6 x 4J/2 inches. Several
hundred stems are present In a single transverse section, the radial arrangement of
which is apparent.

Specimen T2y. — A very much flattened, small disc specimen from Mr-
Thomas's collection, about 4 J/2 inches long and measuring about 3 /2 x ^g inches

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 151

in transverse section at the top. At that point only eight stems are present in the
transverse section, and all but one of these are in some stage of branching. The
stems average lYi mm. in diameter, and are well preserved.

Specimen Tgo. — A small, very well-preserved disc specimen, approximately
35/2x25/2 inches in transverse section and about 2^ inches long. Serial sections
were made of this specimen, and the branching of three of the stems was followed
for a distance of 4 J/2 cm., as illustrated In text-fig. 4. Approximately 55 stems
are found in a single transverse section, of which 22 are in various stages of
dichotomizing. Individual stems vary from 1 5/2 to 4 mm. in diameter, with most
stems measuring 3 mm. in diameter. Preservation of the stems varies somewhat,
although in no apparent set pattern.

Specimen T22g. — 'This is a rather large disc specimen (fig. 9) and is especially
unique in that It Is crescent-shaped In transverse section, the latter measuring
nearly 13x6 inches. Two discs, totaling about 12 inches long, were found, one
of which is in Mr. Thomas's collection and one in ours under the above number.
Most of the stems are In a rather advanced state of decay, due primarily to the
fact that the specimen came from toward the lower part of a trunk as Is Indicated

roots

ASSOCIATED PLANT REMAINS

In his description of Tempskya tvyomingensis Arnold (*45) has mentioned
"occasional fragments of the trunks of Cycadeoidea, which resemble those from
the Freezeout Mountain locality north of Medicine Bow" associated with the fern
material in the valley of Beaver Creek, Bighorn County, Wyoming. Aside from
that reference, nothing has been reported so far as we are aware concerning the
plants with which Tempskya may have been associated in life.

We have been fortunate in finding in section 27 (Lanes Creek quadrangle, see
text-fig. 1) rather abundant fragments of a dicotyledonous wood, a coniferous
wood, and a portion of the trunk of a Cycadeoidea. Although these plant remains
were found only in the one locality they serve to give us some concept of the
ecology of the Tempskyas.

Judging from the very faintly defined annual rings of the fossil woods seasonal
climatic fluctuations probably were not great, and the presence of the cycad, as
well as the cycad and dinosaur fragments reported by Arnold, would suggest a
generally warm climate. If the growth requirements of Tempskya were at all
comparable with those of modern tree ferns the climate must have been a tropical

one.

In trying to arrive at a tentative comparison with modern floras and climates
we have drawn on the extensive field experience of Mr. Paul H. Allen, the Garden's
tropical plant collector in Central America. The following is, we feel, of some
k:omparative significance:

"The only modern conifers associated with tree ferns in the American tropics would be
species of Pocfocarpus, usually found in the highlands between 3000 and 7000 feet. Slender

152

[Vol. 34

MISSOURI BOTANICAL GARDEN

species of tree ferns occur as isolated specimens in heavy rain forest from sea level to 6000
or 7000 feet, being replaced by stouter, handsome species at higher elevations. Greatest
concentrations of individuals, however, are found in open, unshaded locations having ample
moisture, such as banks of small streams in pastures (vicinity of Villavicencio, Mcta,
Colombia), abandoned fields growing up to second growth (highlands of Chiriqui — 4,000
to 6,000 ft.), moist sunny canyons in dry grassland (badlands of lowland Code), or sunny
moist roadside banks in forested areas (vicinity of Puerto Pilon, Canal Zone), or National
Highway near Remedios, Chiriqui Province, Panama. No true cycads occur in the Americas
in close association with tree ferns in modern times, but I have seen species of Zaniia grow-
ing in the same area with them about M.idden Lake In the Zone, and in patches of forest
along the Rio Arlari in Colombia.

"Summarizing, tree ferns grow in greatest concentration in relatively open, sunny situa-
tions, and there is, so far as I know, nothing that could be described as a typical tree-fern
association of plants. Thus, while tree ferns might by pure coincidence be found with
PoJocarf)U5 or Zamia, they might just as often, or rather more often, be found wnth other
things/'^

In a consideration of the climate of this region during middle Upper Cretaceous
times it is pertinent to recall the presence of Anemia Fremont! and Gleichenitcs
coloradensh in the Frontier formation of southwestern Wyoming. These are fern
species with undoubted relationships to the modern genera Anemia and Glciche^tia,
both of which are tropical to warm sub-tropical in their present distribution.
Although the fossils were found south of Kcmmcrcr, Wyoming, the Frontier for-
mation extends north to a point less than 2 5 miles from Wayan, and the actual
distance to the Kcnimerer locality is less than 100 miles.

We do not know what the exact correlation is between the Frontier and Wayan
formations; the latter may lie slightly below the former (Read and Brown, *37,
pi. 27). Yet, it is safe to assume that the tw^o are not far apart. Thus since
these two ferns of tropical affinities were contemporaneous and inhabited the same

Wayan Temp.

Wy

tropical forests covering undulating hills of altitudes up to possibly 4000-7000
feet, and favored by a climate that was uniformly moist and w^arm throughout
the year. The more exact floristics of the ''Tempskya forests" must remain in
doubt, although the Colombia and Canal Zone Habitats suggested by Mr. Allen
seem to present a very likely comparison.

Tempskya as animal food. — It does not seem entirely improbable that the
Tempskyas constituted an important dietary item for certain larger animals of
the time, such as the Cretaceous herbivorous dinosaurs. Diversified as the dinosaurs
were in form and environmental adaptations, some of them almost certainly must
have occupied the habitat of these ferns. The association of their bone fragments
(see p. 124) with Tempskyas adds support to this belief.

ing plant. Wi

Temp

considerable portion of the trunk,

rather than in merely a crown at the top, the leaves were available to herbivores
both small and large. Each plant bore a considerable quantity of foliage even
though the individual leaves were small. From the abundance and wide distribu-
tion of the Tempskyas in certain regions it would SL'cm possible that they may have
been of considerable importance as animal food.

^From a letter received from Paul H. Allen, Gamboa, C.ui.il Zone, May 7, 1946.

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 153

BENNETTITALES

Cycadeoidea sp. — In the summer of 1946 Mr. Thomas found two cycad speci-
mens in section 27 (see text-fig. 1). One of these is a very small fragment while
the other, described here briefly, is a portion of what was apparently a columnar
trunk and measures 7 inches long and 9 x 7.5 inches in diameter. Its owner has
preferred that the specimen remain intact, and since it is rather poorly preserved
and with no strobili in evidence it is doubtful whether sectioning would reveal
significant evidence. In the transverse section the leaf base zone is about 1.5
inches wide, while the entire central portion of the trunk is largely replaced by
silica. At one point a small fragment of the \
less than 2 cm. thick. Like the Tempskya specimens, this cycad trunk fragment
shows no evidence of prolonged water transport; thus its chief interest lies in the
probable association of these plants with the ferns in the Cretaceous landscape of

ood

Waya

CONIFERALES

Cupressinoxylon sp. — In transverse section (fig. 31) this wood presents two
conspicuous features: (1), the annual rings are not sharply defined, there being
very few "late summer" tracheids marking that year's growth from the first cells

ttered

wood

In tangential view (fig. 32) the rays may be seen to vary from 2 to 40 cells
high and predominantly uniseriate with an occasional biseriate one. There is no
evidence of pitting in the tangential walls of the tracheids. The preservation is
not sufficiently good to allow observation of the finer diagnostic details in radial

section.

type

part uniseriate, while the rays appear to be uniformly parenchymatous, there being
no evidence of ray tracheids.

DICOTYLEDONEAE*
SOME NOTES ON METHODS

trun

16

inches in diameter need hardly be emphasized. Furthermore, although the abun-
dance in which the fossils were found during the second and third trips to Wayan
was most encouraging, one could not help but wonder whether a sufficiently
comprehensive study could be completed within a reasonable period. In general,
we have selected only the more complete and apparently better-preserved specimens
for microscopic study. Some sort of critical concentration of material was ob-
viously necessary at the outset, and while specimens of importance, possibly rep-
resenting new species of significant morphological features, may have escaped our

number

bee

peel methods. Specimens up to about 5 inches in diameter were cut in our
. laboratory, while the larcer ones have been cut by the Pickel Stone Company

* As it has not been possible to prepare a discussion of the associated dicotyledonous wood without
seriously delaying publication this will appear at a later date.

[Vol. 34

154 ANNALS OF THE MISSOURI BOTANICAL GARDEN

of St. Louis. Such firms, having equipment for cutting and polishing large monu-
ment stones, may prove of considerable aid in paleobotanical work, and the cost is
not excessive.

In all cases we have first made peel preparations of the complete transverse

There is almost no limit to the area of a section made with this method

section.

provided adequate cutting and grinding equipment is available. These peel prep-
arations proved Invaluable in studying the gross organization of the trunks as
well as detailed cellular anatomy In the better-preserved material. A few other
points with reference to the peel technique seem worth recording. The quality of
preservation in different portions of a peel taken from a transverse section often
varies considerably. Generally, In the larger trunks, the central portion is not as
well preserved as the more peripheral parts. This would be expected In accordance
with our concept of the mode of growth of the trunks. In some, mineralization
apparently was not uniform throughout the trunk.

Especially important Is the fact that the degree of etching prior to pouring
the peel solution had to be especially precise in this material, far more so than in
any other petrifactions that we have studied. As typical of this we may point
out the differences observed in the parenchymatous Inner cortex of the stems where
that tissue Is bounded on the Inside by sclerotic nests and often on the outside by
a thin sclerotic band. With the proper etching time these sclerotic tissues stand
out in striking fashion. However, when the time was too long, very dark peel
resulted, or when the time was too short, the peels were so light that such tissues
could hardly be distinguished at all. In this respect Arnold ('45, pp. 27-28) re-
ports that in T. Wcsselii, *'A peculiarity in the chemical make-up of these stone
cells is that in sections prepared by the 'peeV method they are not recognizable,
which indicates that they are soluble in the hydrofluoric acid used in the etching
process." While this may be true in some specimens It has been our experience
that the peels can be made to show the sclerotic tissues just as well as ground sec-
tions although some experimenting must be done with the etching time to have it
perform to the best advantage.

Ground sections have been prepared from the better-preserved specimens.
These have revealed certain of fhe more minute cellular details, especially in the
root structure, In a more satisfactory manner than peels. Most of the best
Tcmpskya specimens are black, indicating the presence of a large percentage of
the original organic matter. These necessarily must be ground very thin to render
sufficient transparency. We have found that the thermoplastic cement known as
^'Lakeside No. 70," prepared by the Lakeside Chemical Company in Chicago, is far •
superior to balsam as an adhesive. It Is convenient and economical to use, sticks
very tenaciously to glass, and does not, as Is often the case with balsam, present
the diflficulty of being too soft or too brittle.

19471

ANDREWS & KERN IDAHO TEMPSKYAS 155

SUMMARY

1. Fossil plants referable to the genus Tefjipskya have been known for over
a century from European localities including England, Bohemia, and Russia. More
recently a number of species have been described from northwestern United States.

2. The Tcnipskya trunk (false-stem) is composed of many branching slphon-
ostelic stems held together by a dense mass of small, diarch, branching, sclerotic
roots. Over 200 stems have been found composing the trunk of some specimens.

3. The specimens described in this report were collected In southeastern
Idaho chiefly In the vicinity of Wayan, and have weathered out of the Upper

Cretaceous Wayan formation. Other specimens have been obtained from a locality
east of Ammon, Idaho.

4. The largest specimen measures 16 inches in diameter although fragments
of others suggest a somewhat greater maximum size. It is calculated that a trunk
10 inches in diameter attained a height of approximately 12 feet, while the largest
ones may have reached heights of 19 or 20 feet. The trunks were erect or very
nearly so.

5. In transverse section the trunks vary from circular to strongly flattened,
although in nearly all cases the departure of the petioles indicates radial symmetry.
It seems very likely that the flattening has been caused by crushing prior to
silicification.

6. In larger trunks the basal portion consists of roots only. As the stems in

this region decayed their place was taken by roots.

7. Lateral and longitudinal growth took place by frequent branching of the
individual stems, one of the two usually soon ceasing to grow. The leaves were
small judging from the relatively minute size of the petioles, although they were
very numerous on the trunk as a whole. The leaves were given off in two rows
from each stem and probably were borne over two-thirds or more of the length of
the trunk instead of only in a crown as in modern tree ferns and cycads.

8. A detailed consideration Is given of the organization and apparent physi-
ology of the roots.

9- Comparisons are drawn between certain living species of Hi'niUdici and

Toi/ca as well as a Carboniferous Clcpsydropm, The highly peculiar anatomical
organization of Tcmpskya does not compare closely with that of any other fossil
or living plant although those mentioned above, among others, seem to present a
similar "growth tendency" which apparently originated independently in a number

of fern groups.

10. A synopsis (taken from the works of Read and Brown, and Arnold) is
given for the better-known American species, and all of the specimens on which
this paper is based are referred to Tcmpskya Wcsselii Arnold. It seems clear that
all of the American species were closely related and very similar in general habit.

11. Spores and sporangial annull have been found in two specimens.

12. Associated with the Tcmpskya trunks in the Wayan district arc fragments
of coniferous and dicot woods, and a specimen of a cycadeoid.

156 ANNALS OF THE MISSOURI BOTANICAL GARDEN

|\\)i.. ^4

ACKNOWLEDGMENTS

I am sure that my friends in Idaho have wondered, during the years that have

elapsed, about the little there was to show in the way of pubhshed resuhs, and this

paper is brought out with a feehng of deep gratitude for their cooperation and

patience and the hope that some measure of worth and satisfaction has rewarded
their efforts.

For his generosity and ever-ready willingness to guide me along the little-
known paths of his state I am very grateful to William A. Peters of Jerome, Idaho.
Sincere thanks are also due Ralph and Blanche Peters. A very large share of the
credit for whatever contributions may have been made to the story of the
Tempskyas is due Henry Thomas, of whom appreciation has been recorded else-
where in the text. For the specimens obtained from the Ammon locality thanks
are due Mr. E. Manion, of Firth. Last but not least it is a pleasure to acknowledge
the kind hospitality of the ranchers of Wayan whose cooperation in many ways has
added immeasurably to the success of the field work.

We are also indebted to Dr. C. A. Arnold for kindly presenting specimens of
r. Westell} and T. wyomingcnsis from his own collections.

LITIKATURE CITED

AnJreWf, H. N. (1943). Notes on the genus Timinkya. Am. Mid. Nat. 29:153-136.

, (1947). Ancient plants and the world they lived in. Comstock Publishln>; Co. (In

press).

Arnold, C. A. (1945). Silicified plant remains from the Mesu'/oic and Tertiary of western North

America. Mich. Acad. Sci., Arts, and Letters. Papers 30:3-34.
Berry, E. W. (1911). Maryland Geological Survey, Lower Cretaceous, pp. 295-299.
Boodle, L. A. (1S95). Spores in a specimen of Tf/nlfskya ( FnJogcfiffrs) . Ann. Bot. 9:137-141.
Bower, F. O. (1930). Si/e and form in plants. London.

, (1935). Primitive land plants. London.

Buchanan, J. (1886). On a renurkable branchin^^ specimen of Ucmitclij .smit/jii. New Zealand
Inst., Trans. & Pnx:. 19:217.

Field, H. C. (1890). The ferns of New Zealand and its immediate dependencies. London.
Kidsion, R. and D. T. Gwynne-Vaughan (1911). On a new species of Tcm[nkya from Russia.
Russ. Kais. Min. Ges. 48:1-20.

Knox, Elizabeth M. (1939). The spores of Pteridophyta, with observations on microspores in
coals of Carboniferous ai;c. Bot. Soc. Edinb., Trans. &: Proc. 32:43 8-466.

Mansfield, G. R. (1927). Geography, geology, and mineral resources of part of southeastern Idaho.
U. S. Geol. Surv., Prof. Paper 152.

Read, C. B. (1939). The evolution of habit in Tcmpskya. l.loydia 2:63-72.

y and K. W. Brown (1937). American Cretaceous ferns of the genus Tvmpskya. U. S.

Geol. Surv., Prof. Paper 186-F.

Sahni, B. (1928). On ClcpiyJropds, australh, a zygt)pterid tree-fern with a Tempskya-likc false
stem, from the Carboniferous rocks of Australia. Roy. Soc. London, Phil. Trans. B217:l-37.

, (1932). On the genera Cicpsyifropsis and Cludo.xylon of Unger, and on a new genus

Ausfroilrpsh, New Phytol. 31:270-27S.

Schoute, J. C. (1906). Fine neue Art der Stammesbildun^ !m Pflan/enreich. (Ufmitclia cretndata
Mett.). Jard. Bot. Huiten/org, Ann. 20:198-207.

Seward, A. C. (1924). On a new species of Tempskya from Montana: Tcmpskya Kinmltoni,
sp. nov. Ann. Bot. 38:48 5-507.

, and S. O. Ford (1903). The anatomy of ToJt^j, with notes on the geological history
and atlinities of the Osmundaceac. Linn. Soc. London, Trans. 11 Ser. Bot. 6:237-260.
Stopes, M, C. (1915). Catalogue of the Mesozoic plants in the British Museum (Natural History).
The Cretaceous Flora. Part IT. — Lower Greensand (Aptian) plants of Britain, pp. 9-21.

[Vol. 34, 1947]

158 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or Plate

rLATF 15

Fi^. I. A rcprcscnrntivc view of the hills southeast of Wnyan, Td.iho. Numerous
specimens were fouiul alon^; the slope in the center, and on the left slope of the liill in
the background.

Fii;. 2. Mr. Henry Thomas and a portion of his Tcmp\kya collection. Wayan, Idaho.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 15

-I . . . ^

^-^■-

^■^ ^>»rw?S5?=w<i^^

^^^

■^i

-1

I

[^i^

"4

ANDREWS & KERN— IDAHO TEMPSKYAS

[Vol, 34. 1947]

1 60 ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXPI-ANATTON OF PlATE

rLATK 16

Fig. 3. A specimen of the basal end of a trunk showing roots only; basal circum-
ference of specimen 36 inches, upper circumference 32 inches. This specimen differs from
most In that the under side of the base Is uniformly hollowed, forming a large single
cavity with a maximum depth of about 1 inch in the center. Thomas collection.

Fig. 4. A portion of a trunk showing somewhat more rapid tapering than is usual in
one of this length; basal circumference 29 inches, upper circumference 24 inches, length
13 inches. Thomas collection.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 16

Y

^

>

z

2

D
>

X

o

>

;/5

c^>

?

[Vol. 34. 19a7 \

162

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PLATE 17

Fii;. 5. A basal portion of a trunk nearly circular at the bottom :ind somewhat
tlattcneJ above.

Fig. 6. A view of the under side of the base of the same. Henry Shaw School of
Bot;iny collection, T83.

Ann. Mo. Box. Garh., Vol. 34» 1947

Plate 17

ANDREWS 5c KERN— IDAHO TEMPSKYAS

[Vol. 34, 1947J

164 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation oi Plate

PLATE 18

Fig. 7, A typical specimen of the base of a trunk, composed only of roots; circum-
ference at upper end 23 inches. Thomas collection.

Fig. 8. One of the two specimens in the entire W'ayan collections which shows any
appreciable curvature; lower circumference 23 inches, upper circumference 11 inclies. In
transverse view this specimen is somewhat flattened, the diameters of the upper end being
4^ X 8!4 inches respectively. Thomas collection.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 18

]

-^

.^V'** ■■f*

LA

'■^'

-'-. "n

^ ■

y. -L^'

■^

"^ >

ANDREWS & KERN— IDAHO TEMPSKYAS

[Vol. 34, 1947|

166 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation oi Pi. ate

PLAIT- 19

Trunk (disc) spcclniens of Tcnil^kya sliowini; tlic flattened form in which many are
found.

Fig, 9. A rather large and unique specimen in that it is crescent-shaped in transverse
view. This is a short disc, the major and minor diameters measuring 13 x 6.5 inches.
Henry Shaw School of Botany collection, T229.

Fig. 10. End view of three disc specimens from the Thomas collection showing
varying degrees of flatness in transverse section. The specimen at the left is not a complete
disc, the side at the bottom of the photo representing a broken surface. Dimensions of
the (transverse) ends shown In the photo arc: left, circumference 17 inches, diameters
7x2 inches; middle, circumference 17*4 inches, diameters 6^ x 3 ^4 inches; right, cir-
cumference 17 inches, diameters 7x3 inclies.

Ann. Mo. But. Card., Vol. 34, 1947

Plate 19

ANDRFVS & KERN— IDAHO TEMPSKYAS

[Vol. 34. 194/

168 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation' or Plate

PLATE 20

Fig. 11. An older root showing youni; root li:ilrs produced on the "loop** of the
outer cortex. From slide 1406, x 70.

Fit;. 12. Root-hair development of smaller root. From .slide 1406, x 80.

Fii;. 13. Stelar structure of a root in which xylem, pliloem, pericycle and endodcrmis
are dlstin.^u;shabIe. From slide 1411, x 175. Detailed description in text.

Fig. 14. A portion of the root mass showing various stages of preservation in ad-
joining r(x>ts, some perfectly preserved, others with merely the circle of sclerotic middle
cortex remaining, thus indicating that roots were not all alive but in various stages of
decay at the time of fossilization. From peel T53-4, x 17.

Ann. Mo. Bot. Card., Vol, 34, 1947

Plate 20

,..a

AiNDRE^S & KI-RN— IDAHO TEMPSKYAS

[Vol. 34, 1947]

170 ANNALS OF THE MISSOURI BOTANICAL GARDEN

I

Explanation o\ Pi ate

PLATE 2!

Fig. 15. A small root which has penetrated a stem and is hearing a root hair, indi
catcd by arrow. Fri>m slide 1409, x 48.

Fig. 16. Enlarged view of the rout hair in fi^;. 15, x 2R0.

T\^. 17. A young root with just the hcginnini^s of a sclerotic middle cortex, and the
delicate outer cortex complete. From peel T53-2, \ 50.

■

Fi>;. 18. A very small, but older root, as indicated by the sclerotic middle cortex
From peel T17-11, x 80.

Fifi. 19. An older root showing; the tangential rows of sclerotic cortical cells and

the "loop*' produced by the disappearance of the dehcate outer cortical cells. From peel
T17-8, X 53. ^

Ann. Mo. Bot. Garu., Vol 54, 1947

Platf 21

■-■ t-

1

f

''^^■*--- '-9 ,-*-flt -' ' -r

ANDREWS & KERN— IDAHO TEMPSKYAS

L\'oi.. 34, 194/i

172 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or PiATr

PLATF 22

Figs. 20 and 2L Two stems found in peel Tl 7-7, illustrating the variation in stelar
size in a single trunk erection. The large stem shown in fig. 20 hears two leaf traces, a and
h. A third one (c) may be seen to be nearly ready to depart from the stem stele. Both

figures x 20.

Fig. 22. A stem showing unequal dichotomy, from peel T3 3-2, x 20.

Fig. 23. One of the spores found down among leaves and stems in specimen Tl6.
From slide 1400, x 500.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 22

V

1 n I 1^1 I I 13 nn^^^Ti l^---^iL -* MiJ I L^iLk-ririAJ4_^L ^^-^-^-^ ^

1

r ^rVJ, --^f^^f-- -\ -x^i^. -T r.^-L-rt L> 44i^ rO-.^-j ^j ! ■ ^ ■ ^ -| i-

^C.. K^j ^ . \j^.^V . ■-. jvi-n ^^ . T\^■^ i>^ 4* ^^V--?

^n\ -^ V^'-^riV^^ ' " PaiV-^a ■ ^L3 r^T-r V ^^J-J -f-/ -f ^ J*^ *-' ^ '- ' ^ "il^ili nh ii*^"- "^b'-^J ^i--4.i F-- ■ l.">^- ^ ■ ^ iji-lxi- j^riV

ANDREWS & KERN— IDAHO TEMPSKYAS

[Vol. 34, 19471

174 ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXX^LANAIION OF PLATE

ri.ATE 23

o

Fii;. 24. A specimen from tlic Thomas collection, wIili 192 sterns present in Its aren
£5^2 ^ -^ Inches. The specimen was from % to lYz inches thick.

Fig. 25. A much flattened disc specimen (T27) which measures 3 J/2 x % inches in
transverse section at the top, and 4]/2 inches in length.

Fig. 26, A flattened tip specimen (T2) wliich measures GYz x 1 J/2 inches in trans-
verse section at the base and 2% inches in length.

Fig. 27. A tip specimen (T230) which measures 6 x 3 J/2 inches at the base and 12
inches in length.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 23

I-

ANDREWS & KERN— IDAHO TEMPSKYAS

IVoL. 34. 19471

176 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ExpLANAiioN OF Plate

I'l ATE 24

FI>i. 28. Transverse view of paVt of a Tcml^kya trunk showing atout 26 stems
From pcci T51, C, Tl3, x nearly 5.

I
^ -.

■' *

*

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 24

28

ANDRFWS & KFRN— IDAHO TEMPSKYAS

[Vol. 34, 1047]

178 ANNALS OF THE MISSOURI BOTANICAL GARDEN

/

/

\
\

Explanation of Plate

PLATE 25

rig. 29. A partially dissected specimen of the modern Ucfnifclia crcmilata. From:

Sclioutc, J. C, Fine neuc Art dor Stammesbildung im Pilanzenrcich {Hcn/itt'Iia crcimlata
Mett.). Ann. Jard. Rot. Ruitenzorg, pi. iQj fi^. a. 1906.

Ann. Mo. Bot. Gakd., Vol. 34, 1947

Plate 25

t

I

Eh

5=

I

^* ii

' >K-

m

<i.- .-

'^^-

V' ^, ■»^.->.

\'i

-t

•M^'-

:m

\

' I-

1 9^i ^A. ^ '

^.

*^"'..-

^^

^1

■> F^

^ A/

:3*

-TO

:»-?;

ll>.

t ^

*

K,: M

s^-^ *

*AV

?

>?

Ai,*V'

.MW-

^ :k

i^^-

^ .. ■■

' *

-ifci

i

29

ANDRF.WS & KERN— IDAHO TEMPSKYAS

[Wn. 34, 1947]

180 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ExPLANAiioN oi Plate

PLATE 26

Fig. 30. A profusely branchini; specimen of the modern HemitcVta Swithii from
New Zealand. From: Buchanan, J., On a remarkable branching specimen of Hfmifclia
Sntif/tii. Trans and Proc. N. Z. Inst., pL 22. 1886.

Ann. Mo. Bot, Card., Vol. 34, 1947

Platl 26

'^^.

K

^'F

■>>-

^'v* *.'

^ T

"m^ ^

Jf

t K^

-■fi

- #

-k .

■'■^■J

>^> '

■■tH.A:

sK

^__ *

'stV:!

.^;

_ifti

^

-^

^r

?

V^L

, J

/ *.

7-*

t

I

^ll. . X .^

il H

>

*//■)' 1-^^

ts\ \^ .^ Y' "'^^^'i/

'n

"<-

J

■.^

1-

IV

/I

i-S

■ ;*■

^■^

#??^

//

^ ^

/.

I

X

■>*■'

J

ft

.-.^

:V

i*

V.^i

t»,

^:

-"1^ .

X>v.^

■/

t^

*^i

n

1 T

•i

i ?

lI.

/ .*'

'*-

^-'
-=*^>^

V-x *v

^

T >^

^

<v

4

^*

4-fj

I

V

\

^1

\

*tf*r*

/

-■^i

.--!'.

»^«:

V t J*»,

\'H

'^^^^■'

* ti

*^

■^

l*^

Sf-

"t -^

^^^

V

.*

'.t «_

:^

A

>y

s I i^.i.

■c*y^

,-^

^ ^ *

*^

-ri-^

»^"

V

■*-

■J

i

f.

i

i

'♦

''■^-

1

V

J <

^* 4

I

-^

\' *

H

L

.-*

^.

f^

fS^'V

o »v

C'>

>^

$

jW

i'^ -

K.'

1 *■'

^»i* I-* a;

-f,vi;

■ f^

. >.

><^

/

30

ANDREWS & KERN— IDAHO TEMPSKYAS

[Vol. 34, 1947]

182 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or Pi ate

riATF. 17

Figs. 31, 32. Transverse and tangcMuial sections respectively of a specimen of coni-
ferous wood {Ciiprcssiftoxylon sp.) found associated with Tcmpskya cast of Wayan, IJaho.

Fig. 31 from slide No. 1474, specimen T8, x 62; fig. 32 from slide No. 1475, specimen
T8, X 62.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 77

ANDREWS & KERN— IDAHO TEMPSKYAS

1947]

ANDREWS & KERN IDAHO TEMPSKYAS 185

Appendix

SOME COMMENTS ON THE DISCOVERY OF TEMPSKYA IN THE VICINITY OF

WAYAN, IDAHO

r>y C. IIEXRY THOMAS

In answer to your request I have included in the following lines a brief histor-
ical sketch of my Te77ipskya~co\\ecting activities In the Wayan district.

I located in that part of the state in May, 1915, having migrated from the
Scottsbluff country in Nebraska where I was raised on the frontier. Mine has
been pretty much of an outdoor life. I was born in a one-room sod house in a
still untamed country and have herded sheep or otherwise worked with livestock
ever since the age of seven. As a bo) T roughed and toughed it on the wind-swept
prairies of western Nebraska, which were then treeless and monotonous with not

even a shrub in sight. But since then they have been put under irrigation and
become most productive and desirable.

After locating in the Grays Lake country of Caribou County I became occu-
pied as a sheep herder and was naturally attracted to the odd and out-of-the-
ordinary petrified fossil remains which are frequently seen on the range. At the
time I did not know whether they were plant or bone, or in fact what they were;
It was evident only that they were fossil remains. If bone It seemed most likely
that they were saurians, or lizard-like reptile remains, and if plants, I had no idea
of their origin.

Most of us arc born with a sense of curiosity. We like to pry into the un-
known. There are charms in new ventures, and thrills in first discoveries. More-
over, most of us are pack rats. We like to accumulate, although not many become
enthralled with rocks.

I started gathering specimens in a small way almost simultaneously with my
arrival on the Williamsburg bench. Later, as my collection had grown to a size-
able extent this matter became noised abroad, bringing a number of mineral
collectors and rock-hounds from far and near. But no one knew what they were.

Myths and tales precede scientific knowledge. The human mind wants an ex-
planation. And such settings gave birth to wild and fantastic tales. In one In-
stance a certain oil-stock promoter, on visiting Mr. Sam Sibbctt*s ranch, claimed
to be able to trace the outline of some huge bird which was of such gigantic
proportions that It could seize an elephant by the nape of its neck! Such were the
earlier local concepts of these interesting fossils that we later learned were petrified
Tree fern trunks.

In the fall of 193 6 I read in the Pathfinder of Mr. Roland W. Brown being
associated with the Smithsonian Institution and doing paleontological work in
Idaho and other far western states, so I sent some specimens to him for identifica-
tion. Mr. Brown informed me that they were the so-called Tempskyas, or the
petrified remains of a peculiar fern of the Cretaceous period.

Mr. Brown contacted Mr. W. W. Rubey (also of the U. S. Geological Survey)
and as a consequence Mr. Rubey, who was doing field work on the Wyoming side

[Vol. 34, 1947

186 ANNALS OF THE MISSOURI BOTANICAL GARDEN

during the .summer of 1937, came over to Investigate. He expressed surprise at
the abundance of Tempskyas in this region, and also took a number of leaf imprints
of semi-tropical plants found in sedimentary rocks adjacent to grounds where

Tempskyas weather out.

In August, 193 8, Mr. Brown and Mr. Carl Mumm came to study the Temp-
skyas and the stratification of the beds out of which they weather.

Mr. W. A. Peters of Jerome, Idaho, who, of all rock-hounds, undoubtedly has
the largest and most diversified collection In the state of Idaho, paid a visit In the
spring of 1942 and brought with him Mr. Henry N. Andrews of the Missouri
Botanical Garden, St. Louis. Mr. Andrews came at a favorable time of the year,
when the snow was gone and there w^asn*t much vegetation to hide rocks, and the
ground was soft so that they could readily be pried out with a wrecking bar,
Paleobotanists know best how to appreciate plant remains of past geological ages.
All those who have visited here have appeared to be highly interested and have
commented on the excellent state of preservation of our Tempskyas.

During the summer of 1943 I herded sheep for Mr. Emil Stoor on ground
adjoining my former holdings to the north and east, and that is when I really
found most of my larger and better specimens.

This is an ever-changing u^orld. Fossils are the evidence of the existence of
former forms of plant and animal life. Scientists by tracing these clues endeavor
to read the history of the earth's past geological transformations. Nothing just
happens. Everything is the result of preceding forces. A rock is the product of
nature's workings In the past.

Annals

of the

Missouri Botanical Garden

Vol. 34 SEPTEMBER, 1947 No. 3

MONOGRAPH OF THE NORTH AMERICAN SPECIES OF CORYDALIS^

GERALD BRUCE OWNBEY^

Introduction

My attention was attracted to the genus CorydaUs of the family Fumariaceae
some years ago, since it seemed to offer many unsolved problems in the systematic
Interpretation of various species. The genus had received no special attention
from any American botanist since the time of Engelmann and Gray, and the pro-
posal of nearly forty new names for members of the genus in America by Fedde
during the early years of the present century had made it imperative that their
proper status be determined so that scientific literature no longer would be encum-
bered with superfluous nomenclatonal terms.

A great volume of herbarium material has been available, in the study of which
I have attempted to make full use of classical methods. In addition, it was felt
that field studies would help in the understanding of some of the more difficult
species. With this in view a six-wceks field trip was made through the western
United States during the summer of 1946.

I have attempted to view all species in the light of modern concepts of specia-
tion. Population studies of one species have confirmed the presence of minor
measurable differences between inbreeding colonies of this species, even those not
widely separated geographically. I also have had opportunity to grow several
species under greenhouse conditions and to test the stability of minor variants.
Further work along these lines doubtless would do much in clearing up obscure
problems not amenable to standard methods of the herbarium taxonomist.

■^An investigation carried out at the Missouri Botanical Garden and submitted as a thesis in partial
fulfillment of the degree of doctor of philosophy in the Henry Shaw School of Botany of Washing-
ton University.

^Instructor in Botany, University of Minnesota, Minneapolis,
Issued October 31, 1947.

(187)

[Vol. 34

188 ANNALS OF THE MISSOURI BOTANICAL GARDEN

History of the Genus

Linnaeus^ included all of tlie species of fumariaceous plants known to liim in
tlie single polymorphous genus Vnmaria. The subsequent subdivision of this
heterogeneous group of plants was left to later authors who attempted to revive
names of pre-Linnacan botanists in something resembling their original sense.

The nomcnclatorial complications of the generic name Corydalis have been in-
vestigated by Spraguc^, from whose account the following discussion largely is
abstracted. The name has been used in two distinct senses as follows:

(1) Corydalis [Knaut, Meth. PL 153. 1716; Dill. Cat. PL App. 129. /. 7.

ticab

1763 (Cystica I

1791), and is based

upon Tumaria vcsicaria L. This monotypic genus sometimes is united with Cory-
dalis Vent., but in the opinion of students of the family, such as Hutchinson and
Feddc, it should be retained separately. As the fruit of Cysticapuos is inflated and
bladder-like, it seems probable that sufficient grounds exist for segregating the
species gencrically from Corydalis.

(2) Corydalis Vent. Choix de PL /. ig, 1803. [Capnoidcs Tourn. Inst. Rei
Herb, 423. /, 237, 1719]; Capnoides Adans. Fam. PL 2:431. 1763, Ventenat
treated only a single species, Corydalis fitngosa (Adlnmia fungosa Greene), which
is now universally recognized as a separate genus. The generic name, however,
must be credited to Ventenat, even though the single species is referable to
Adliiwia, since the author states in a footnote that his generic concept is founded
upon that of Tourncfort, who described and figured Corydalis sempcrvircns (as
Capnoides sempcri'irens).

Because of the widespread acceptance of the name Corydalis in its modern sense,
perhaps occasioned by its adoption by de Candollc in his monumental works^-*,
the International Botanical Congress of Vienna conserved it over Capftoidcs Adans.,
Cisticapfios Adans., Ncckeria Scop., and Vseudojumaria Mcdik. The conservation
of Corydalis has insured its permanent use, and has precluded the revival of any
other generic name which otherwise might supersede it.

General Morphology

The aerial parts of Corydalis arc succulent, and annual or biennial in all of our
species. The glaucous foliage and finely dissected leaves give a characteristic
aspect to the plants.

In distinguishing sections and species, greatest Importance Is attached to the
morphology of the outer petals, stigma, fruits, seeds, and underground parts. An
account of the peculiarities of structure and the special terms used in referring to
them is given, therefore, In the following discussion.

^Linnaeus, Sp. Pi. 2:700. 1753.

^Spraguc, In Kcw Bull. Misc. Inf. 1928:351. 1928.
^DeCandolle, Reg. Vcg. Syst. Nat. 2:113. 1821.
*DcCandolle, Prod. Syst. Nat. \:{26. 1824.

1947J

OWNBEY AlONOGRAPH OF CORYDALIS 189

Root: The nature of the root Is an important diagnostic character In delimit-
ing sections. In annual, winter annual, and biennial species of Section Eucory-
DALis^ an ordinary tap root Is present. This often Is quite succulent and may be
somewhat lignified when the plant reaches maturity. The roots of the perennial
species must also be classified as tap roots, although they become fleshy at a very
early seedling stage. In C pane/flora of Section Pes-gallinaceus the mature root
is tuberous and ordinarily bifurcate. In Section Ramoso-sibiricae, the seedlings
develop a tuberous sweHing the first year. This grows to large proportions during
succeeding years and often is crushed and flattened by pressure of the soil.

Stem: A rhizome Is present only In perennial species of Section Ramoso-
sibiricae. This gives rise apically to the annual stems.

The hollow, annual stems are succulent in all of our species, although some-
times semi-ligneous at the base. They are monopodlal In growth except In Section
EucoRYDALis where they predominately are sympodlaL The nodes are somewhat
abbreviated toward the base.

Leaves: In our species of Section Ramoso-sibiricae only a single leaf is pro-
duced annually until the plants reach flowering age. The number produced in
Section Pes-gallinaceus is unknown, but presumably is low. In all members of
Section Eucorydalis a basal rosette of leaves Is developed prior to the development
of the flowering stems. Leaves are produced alternately, the later ones often being
progressively reduced in size and intcrgrading Imperceptibly Into the floral bracts.
The larger stem leaves are petlolate and pinnate except in C. paticiflora where they
are simple and ternately divided. The primary segments are themselves once or
twice pinnatifid or incised. The petioles are somewhat expanded at the base, es-
pecially those of the larger cauline leaves. A few sheath-like cataphylls some-
times are present at the base of the stem.

Inflorescence: The Inflorescence is a terminal raceme or panicle, the flowers
being crowded at first but becoming more distant during and after anthesis
through elongation of the floral axis. The floral bracts offer very little In the way
of diagnostic characters. For the most part they are successively smaller from the
base to the apex of the floral axis. The uppermost are often extremely minute.

Flowers: Corydalis flowers are bilaterally symmetrical. They are dimerous,
having two inconspicuous sepals, two laterally placed outer petals, one of which
is spurred, and two inner, dorsl-ventrally placed petals opposite the sepals. There
is some cohesion but no true fusion of the petal margins at the base. The stamens
are arranged in two phalanges of three each, which are opposite the outer, lateral
petals. Morphology of all the parts presumably Is conditioned to some extent by
compression in the bud. Floral structure Is quite uniform throughout the genus.

Tlie very much reduced, rudimentary sepals are scarlous and fugacious, and

function as protective organs only In the early bud stage. Although they are of

little diagnostic value, they are described In detail for each species treated in this
paper.

[Vol. 34

190 ANNALS OF THE MISSOURI BOTANICAL GARDEN

The presence of a single spurred petal and the polyspermous fruit are para-
mount characters in distinguishing the genus CoryJalis from closely allied genera.
The relative size and shape of the spur vary ui diflfcrent species. In measuring the
length for this study the distance from the point of attachment of the pedicel to
the tip of the spur is taken. The free end of the petal is carinate. This carina is
referred to in the following descriptions as the hood. It is often provided with a
medial exterior fold, the crest, and an expanded border of greater or lesser width,

the wing margin.

The unspurred outer petal is similar to the spurred one with respect to the
hood, crest and wing margin. When the flowers are fully developed, the petal
sometimes is geniculate immediately posterior to the hood. Tn Section Ramoso-
SIBIRICAE there is a distinct basal gibbosity which probably represents a rudi-

*

mentary spur.

The two asymmetric inner petals are connate at their apices and enclose the
anthers and stigma at anthcsis. They consist of an outer broader portion, the

(IW

is a longitudinal fold which lies between the margins of the outer petals in the
bud. In addition, on the exterior basal half of the blade adjacent to the spurred
petal, there Is another simple longitudinal fold or fleshy protuberance. Morphology
of the inner petals ordinarily is not of diagnostic importance below the sectional
level.

Each stamen phalange consists of three stamens whose filaments are united
laterally. The anther of the central stamen Is di thecal; those of the outer stamens
are monothecal. The phalange opposite the spurred petal is provided with a
nectiferous stamen spur which extends into the petal spur and is adnate to it for
the greater portion of its length. No morphological characters of value in dis-
tinguishing species are to be found In the androecium.

Both the stigma and style are persistent. The style is slender, short, and not
distinctive in character. The flattened stigma, however, often is quite character-
istic. Stigmatic surfaces are located on papillary projections numbering from four
to eight in our species. As the projections are somewhat delicate, an unopened
flower or large bud should be selected for examination. Tn these, germinating
pollen does not obscure details of structure as it does in older flowers. The nature
of the stigma is of considerable diagnostic value, especially in defining sectional
li

ines.

Teratological flowers of the type termed "peloric" have been reported from

time to time. I myself have observed two instances of this phenomenon in which
both outer petals were spurred.

Cleistogamy in CoryJalis has been recognized for many years. In C. micrantha
and C. pseudoviicrantlja the potentiality for cleistogamy is present at all times.
In other species it is rare, non-existent, or unrecognized. The problem of self-
fertility is well worth investigation. The flowers of all species studied show evi-
dence of germination of the pollen which is clustered around the stigma. Under

1947]

OWNBEY MONOGRAPH OF CORYDALIS 191

such conditions it is difficult to determine just how much self-pollination actually
is occurring. Opportunity for cross-pollination is not lacking as witnessed by the
large number of insects, both as to individuals and species, that visit Corydalh
having brightly colored flowers.

Fruit: The young fruit is enclosed by the stamen phalanges at anthesis, and is
oriented so that the placentae are opposite the Inner petals in the dorsiventral
plane and the flattened stigma in the transverse plane. When mature, it is a few-
to many-seeded, bicarpellate capsule varying in shape from narrowly linear to
broadly elliptical, oblongoid or obovoid, and sometimes flattened at the base.
Dehiscence is accomplished by separation of the two valvate portions from the two
placentae. In Section Ramoso-stbtricae the valves roll up elastically from the
base and the seeds are scattered to a considerable distance from the parent plant.

All American species of the genus fall into two well-defined groups with re-
spect to disposition of the fruit on the pedicel. In Sections Ramoso-sibiricae and
Pes-gallinaceus the body of the fruit is geniculate at the base and is reflexed to
a marked degree upon erect or spreading pedicels. In Section Eucorydalis the
fruit is not geniculate and is not reflexed except by actual curvature of the pedicel.

The ovules are campylotropous. At a very early stage a comb-Uke or sheath-
ing caruncle appears near the point of attachment of the funicle. At maturity
the caruncle covers a greater or lesser portion of the seed. The testa is seen to be
essentially smooth to variously reticulate or muricately decorate'd when viewed
under magnification. The nature of these decorations, as well as gross size of the
mature seed, is of importance in specific diagnosis.

A pathological condition in which the fruits become swollen, spongy and sterile

is not uncommon In C. Caseana and C. aurea. Upon examination such abnormal
fruits arc found to contain an insect larva. The t^^ of the adult insect apparently
is lodged in the young fruit at flowering time; the resultant larva passes through
the early stages of its existence enclosed within the tissues of its host. Faulty
interpretation of the peculiar fruit developed under these conditions has led to
some confusion in terminology.

Generic Relationships

It is not within the scope of this paper to discuss the relative merits of the
■ many genera of the family Fumariaceae. Hutchinson'"* lists eighteen genera from
Europe, Asia, Africa and North America, and Fedde^ proposes one additional genus
from Asia. In North America only four genera are represented, namely, Fumaria,
Dicentra, Adlnmia, and Corydalh. All of our species of Fumaria are weeds intro-
duced from Europe.

The genus Corydalis, the largest of the family, includes a heterogeneous aggre-
gate of species- all of which, however, have certain fund'amental characters in
common. The petals are free or essentially so, and the corolla is zygomorphic,

"Hutchinson, in Kcw Bull. Misc. Inf. 1921:97. 1921.

*^Fedde, in Engler & Prantl, Nat. Pflanzenf. ed. 2. 17b:121. 1936.

[Vol. 34

192

ANNALS OF THE MISSOURI BOTANICAL GARDEN

having 1 single spurred petal These two characters distiaguish the genus clearly
from Adlumia^ which has petals united below, both of the outer ones barely saccate
at the base, and, furthermore, it is of climbing habit. The genus Dkcntra has both
outer petals equally spurred, but does not differ fundamentally in any other respect
from Corycialis. The fruit of CorydaVn is a 2- to many-seeded, dehiscent, bivalvate
capsule, while Fumaria has an indchiscent, one-seeded fruit.

The distinctions between extra-American genera of the Fumarlaceae and Cory-
dalis are of like magnitude to those of the American genera. An important cle-
ment in the family is its naturalness and the equal systematic value of the features
characteristic of its included genera.

Evolutionary Tendencies and Interspecific Relationships

A discussion of evolutionary tendencies often is speculative. The initial as-
sumption generally made is that distant or close genetic relationships are indicated

curvlalllqua

aeaD«rTlr«na

C* pauolf lora

C« aurea

C, mlcrantha

C. Scoulerl

C- Ca&eana

C* paeudonlcrantha

C* crystal Ilna

ANNUAL on ,s,

BISNKIAI /»/ fERlNHIAL

Fig. 1. Suggested interspecific relationships.

4

by greater or lesser morphological similarities. This assumption ordinarily is justi-
fiable providing due consideration is given to such modifying factors as parallel
evolution. Discussions of this nature also arc of value in bringing into perspective
the probable direction of evolution within the group, and serve to emphasize not
only inter-rclationships of species, but the characters which are undergoing
basic change. This makes possible a prediction of the type of subsequent change

to be expected.

Annual species usually arc assumed to have arisen from perennial species by

compression of the life cycle into a single year. Many genera have both perennial

and annual species, Corydalis being such a genus. Presumably, one or more lines

of perennial species have given rise to the annuals which now predominate in

America.

Among perennial species of Corydalis are some with rhizomes and some with
tuberous roots, one of which type may have given rise to the other at some time in
the past. Also, among perennials arc species with pinnate leaves and those with sim-

1947]

OWNBEY MONOGRAPH OF CORYDALIS 193

pie, ternately divided leaves. The upper stem leaves of the pinnate-leaved species
are much reduced, and, strictly speaking, are simple. Historically, a species with
simple leaves could then be derived from one with compound leaves by a fore-
shortening of the axis or by a reduction of compound leaves to their terminal
segments. It is not improbable that this has occurred in Corydalis,

Among both perennial and annual species are those with paniculate and those
with racemose inflorescences. The racemose type could be derived from a panicu-
late type by reduction of branching.

In some annual species, such as C. mirea^ both sympodial and monopodial
growth is found. Racemes in both types are terminal. After flowering of the
primary raceme in the sympodial type, growth in length is taken over by the
uppermost axillary shoot. This in turn gives rise to a terminal raceme. The evo-
lutionary significance of the sympode is not clear. I believe, however, that it

4

should be looked upon as derived from the monopodial type. It may represent
only an adaptation for continued growth provided conditions remain favorable.

Concept of Species and Subspecies

I have attempted to portray the species as natural, biological units, the mem-
bers of which are more closely related genetically to one another than to members
of other species. Closeness of genetic relationship Is manifested by relatively close
morphological similarities. Comparative morphology, then, remains the most Im-
mediately usable criterion of genetic relationship. The individuals of each species
are potentially interfertile, or at least their historical progenitors were Interfertile.
Further, each species occupies a "natural" distribution determined by factors in-
herent within it.

Great morphological variability In some species of Corydalis makes over-all
statements of distinguishing characteristics of species and subspecies misleading If
unquahfied. There apparently is a corresponding amount of genetic diversity in
some of these plastic or polymorphic species. C. atireay for example. Is relatively
uniform throughout the northern part of its range. In southwestern United States

it breaks up Into innumerable forms or ecotypes. According to recent concepts,

these may be looked upon as expressions of genetic differences due to isolation of
small segments of the whole. Under such circumstances there is said to be a '
potential loss of heterozygosity and its accompanying morphological variability,
together with potential evolution of the isolated segments along lines divergent
from that of the species proper. This appears to be a satisfactory explanation of
the condition in C. aurea, but experimental data to support this view are insuf-
ficient. Although colony-to-colony variation can be demonstrated statistically in
C. flavtJa, this does not carry over Into any recognizable regional pattern of
variability. I hope to have opportunity to discuss this question more fully in a
future study.

Another type of variability common to many plant species is most striking In
C. sempervirens and C micrantha. The conditions under which the particular

[Vol. 34

194 ANNALS OF THE MISSOURI BOTANICAL GARDEN

plant grows affect gross size and form to a marked degree. For example, plants
subjected to abnormally dry conditions often are dwarfed; those growing in shade
have fewer branches, more delicate-textured leaves and more slender stems. When
plants of C. inicrantha grow closely crowded together they are less branched and
a higher percentage of normal flowers is developed.

The probability of interspecific hybridization in nature is limited to C aurea,
C, murantha and C curvisiliqua, • These species seem to be closely related, and the
possibility or even probability of hybridization among them must be considered.
Evidence for this is at present inferential and is based upon plants of intermediate
character collected in southern Missouri, Oklahoma, and Texas, Controlled
crosses between these species are necessary to supply affirmative or negative evi-
dence of potential hybridization in nature.

4

In adopting the category of subspecies, I have attempted to maintain its usage
In the strictest sense. According to this usage, each subspecies has a discrete or
nearly discrete distribution of its own within that of the species as a whole. One
may have potential overlap in situations where the habitats favored by the sub-
species themselves overlap or intergradc. In Instances where two or more elements
of the species have been isolated historically by some barrier subsequently removed,
the elements may again intermingle along their zone of contact providing they
are still interfertile.

The second attribute of a subspecies in Corydalis is minor but perceptible
morphological differentiation. The subspecies are not always mutually exclusive
on morphological grounds, but each has a norm of variability which differs from
that of other subspecies. It sometimes becomes a matter of judgment as to
whether to describe two closely related elements as species or subspecies. There
is no hard-and-fast rule which will be universally applicable due to the funda-
mental nature of speciation.

Chromosomal Complements

The basic chromosome number In Corydalis is 8 in species, all European, so far
reported, with a single exception having a probable basic number of 7- There is
evidence of the occurrence of polyploidy in two species, but the data are too scanty
to justify generalization.

Species Reported by n 2n

C bulbosa . Maude"^

C. cava Tischler^ 8

C. lutea Kellct^

C. pumila Nemec^^

24

56?
16

/

■^Maudc, in New Phyt. 39:18. 1940.

^Tischlcr, In BIoI. Ccntralbl. 48:343. 1928; Planta 8:696. 1939.
®Kcllct. ace. tb Darlington & Janaki, Chromos. Atlas Cult. Pi. p. 69. 1945.
^^Ncmcc. ace. to Tischlcr, in Planta 8:695. 1929.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 195

In two American species which I examinedj the diploid number has been tenta-
tively established as 16. I hope that a further report on American species will be
possible when data from material now under study are compiled.

Alkaloidal Properties

The alkaloidal properties of a large number of fumarlaceous and papaveraceous
species have been reported by Manske^^. American species of Corydalh investi-
gated have been C. Caseana (ssp. Caseana) , C. Sconleriy C. aurea (ssp. aurea) ^
C, aurea ssp. occidentalh (as C. montana), C. micraiitha (ssp. australis 7), C.
crystallina and C. sempervircns. It is of interest to taxonomists that each species
was found to contain a particular set of alkaloids, some of which are common to
other species but not in the same combinations.

Manske has drawn certain conclusions about interspecific relationships which
are substantiated on morphological grounds. For example, from a chemical stand-
point, he agrees that C. micrantha and C. crystallina are species distinct from C
aurea. However, he treats C. aurea ssp. occidentalis (C. montaua) as a distinct
species, a view that I am not able to support on the basis of comparative mor-
phology. Manske's work has gone a long way In confirming for this group an
assumption that perhaps is sometimes unwarranted, that is, that physiological dif-
ferentiation may accompany morphological differentiation, even In lower system-
atic categories. It would be of great interest to the taxonomlst to know whether
changes In alkaloidal properties are present in widely separated geographical seg-
ments of a species In wliich little or no morphological differentiation is present.

Economic Importance

I

In so far as is known, the species of Corydalis are not of great economic im-
portance. According to collectors' notes, the plants are utilized by the Zuni
Indians of the Southwest, but to "what extent or purpose is unknown to me. As a

forage for livestock, they are of no importance both because of their relative

scarcity and the apparent unpalatibility of the foliage. On account of their high
alkaloidal content^ It is probable that they are distasteful to livestock as well as
toxic if eaten In quantity.

Many species of Corydalis have been grown in gardens as much for their value
as curiosities as for their Intrinsic decoratlveness. Of the American species, C.
Caseana ssp. Brandegei and ssp. Cnsickii are especially recommended for trial. Both
are handsome plants In nature, but greatly restricted in habitat. C Scmilcri^ C.

.h

the past.

■*^^Series of papers in Can. Jour. Res. Ser. B, beginning with 7:258-264. 1932, and still continuing.

196 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

Geographical Distribution

The genus Corydalis is confined almost exclusively to the northern hemisphere.
Its center of diversity Is In Eurasia, there being several times the number of species
there as are found in North America. In North America, the greatest number of
species are concentrated in eastern Oklahoma and adjacent Texas, Arkansas, and
Missouri, all species found there being placed in Section Eucorydalis. The closest
affinities of each of the sections represented in North America are, however, with
their Asiatic and European counterparts, and not with each other. Problems of the
limits of distribution of the American species are discussed individually in the
taxonomic section which follows.

Acknowledgments

I am greatly indebted to the Missouri Botanical Garden and to its director,
Dr. G. T. Moore, for the use of its library and herbarium facilities during tbe

course of this study.
E. Wood

J

I wish to extend my thanks to the curators of the several herbaria who have

permit

However, in

order to conserve space in this paper it has been necessary to omit detailed
lists of specimens examined with the exception of the material from Mexico,
Canada, Alaska, and that of C. micrantha ssp. texensis first described in this paper.
The disposition of all numbered and many unnumbered collections viewed in the
course of this study may be ascertained by reference to the Index to Exsiccatae.
For the information of curators of herbaria and other Interested persons, citation
of authenticated specimens for each county hsted may be found in the original
manuscript which is deposited in the library of Washington University, St. Louis,
Missouri. Following Is a list of the herbaria, together with the abbreviations
adopted. Material at each of these herbaria has been viewed and annotated by me.

CA — Herbarium, Colorado Agriculture and Mechanics College.
CAS — Herbarium, California Academy of Sciences.
CIUC — Clokey Herbarium, University of California.
D — -Dudley Herbarium, Leland Stanford University.

DU — Herbarium, Duke University.

G — Gray Herbarium, Harvard University.

IH — Intcrmountain Herbarium," Utah State Agricultural College.

M — Herbarium, Missouri Botanical Garden.

NMA— Herbarium, New Mexico College of Agriculture and Mechanic Arts.

NY — Herbarium, New York Botanical Garden.

RM — Rocky Mountain Herbarium, University of Wyoming.

T — Herbarium, Tulane University.

UA — Herbarium, University of Arizona.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 197

UC — Herbarium, University of California.
UM — Herbarium, University of Minnesota,
UO — Herbarium, University of Oklahoma.
US — United States National Herbarium.
UT — Herbarium, University of Texas.
WS — Herbarium, State College of Washington.

Taxonomy

CoRYDALis Vent. Choix de PI. /. /p. 18 03, nam. conserv.y exclusive of Corydalis
fiingosa Vent.; not Medik.

[Capnoides Tourn. Inst. Rel Herb. 423, t. 23/. 1719].

Fumaria L. Sp. PL 2:700. 1753, In part.

Capnoides Adans. Fam. Pi. 2:431. 1763, 7iam, rejic,

Neckeria Scop. Tntrod. Hist. 313. 1777, nom. rejic,

Pseudo-Fiunarta Medik. Phil. Bot. 1:110. 1789, nom. rcjic,

Pistolochia Bernh. Syst. Verz. Pfl. 57. 1800; not Raf.

Borckhatisenia Gaertn. ex Mey. & Scherb. Oekon.-Tcch. Fl, Wett. 3:4. 1801.

Odoptera Raf. Cat. 15. 1824.

Capnites Dumort. Fl. Belg. 117. 1827.

Bulbocapnos Bernh. in Linnaea 8:469. 1833.

Sopborocapnos Turcz. in Bull. See. Nat. Mosc. 211:570. 1848.

Cryptoceras Schott & Kotschy, in Ocster. Bot. Wochenbl. 4:121. 1854.

CorydaUis Aschers. Fl. Prov. Brand. 2:9. 1864.

Capnodes Ktze. Rev. Gen, 1:13. 1891.

Annual, biennial or perennial herbs from a tap root, tuberous root or rhizome;
stems monopodia! or sympodial; leaves simple or pinnate, the pinnae deeply once or
twice divided and incised; inflorescence a panicle or raceme, terminal, bractcate;
flowers bilaterally symmetrical; sepals 2, scarious, often fugacious; petals 4, free
or somewhat coherent at the base, in two whorls of two petals each; outer petals
dissimilar, one spurred, the other sometimes gibbous at the base, both more or less
distinctly keeled or hooded at the apex; Inner petals similar, connate at the apices,
clawed; stamens In two groups or phalanges opposite the outer petals, each phalange
with three anthers, the outer two of which arc monothecal, the central dithecal;
phalange opposite the spurred petal having a distinct glandular spur which is ad-
herent to the Inner surface of the petal spur except at the tip; stigma persistent,
flattened, sometimes 2-lobed, with 4-8 papillary stigmatic surfaces; style distinct,
slender; fruit a bicarpellate, many-seeded capsule, with two sterile valves and two
persistent placentae; seeds having a distinct chalazal appendage or caruncle, smooth
or variouslv decorated under magnification.

Standard Species; C sempervirens (L.) Pers. Syn. Pi. 2:269. 1R07.

KEY TO THE SECTIONS

A. Perennial from a rhizome or tuberous root; leaf blades pinnate or
simple; flowers never yellow; fruits obloneold to obovoid, reflexed
upon erect or spreading, straigliC pedicels; stiema approximately rec-
tanf;u?ar or triangular, as long as broad or longer.

B. Rank-growing, bydrophllous species of western United States;
rh'V.ome and roots large and fleshy; leaf blades pinnate, the pinnae

198 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

once or twice plnnacifid or incised; flowers pink or white; stigma

rectangular, or if triangular, narrowest at the apex Section I. RAMOSO-siBfRiCAE (p. 198)

BB. Low-growing species of the far north; roots small, tuberous, usual-
ly bifurcate; leaf blades simple, once tcrnately divided, the segments
incised; flowers blue or purplish-blue; stigma triangular, broadest
at the apex Section II. Pes-gallinaceus (p. 207)

AA. Annual or biennial, with a somewhat succulent root ; leaf blades
pinnate, the pinnae once or twice pinnatifid or incised; flowers yellow
(except in sp. 4); fruits narrowly to broadly linear, never reflexed
upon the straight or curved pedicels; stigma approximately rectangu-
lar, broader than long (except in sp. 4) Section III. Eucorydalis (p. 209)

Section 1. Ramoso-sibiricae (Fedde) G. B. Ownbey, stat. nov.

Ramoso-sibirlcae Fedde, in Engler & Prantl, Nat. Pflanzcnf. ed. 2. 17b: 1)1. 1936, as

subsection.

KEY TO TI IE SPECIES AND SUBSPECIES

A. Stem leaves about 3; primary axis of inflorescence with about 25
flowers; flowers pink, the inner petals not tipped with deep red or
purple; outer petals having no wing margin, the hood generally hav-
ing a very high crest; stigma approximately triangular, narrowest at
the apex, about as broad as long; fruit obovoid; seeds about 3.5 mm.
in diameter, distinctly papillose under magnification; coastal "Wash-
ington, Oregon, and Vancouver Island 1. C, Scolder i

AA. Stem leaves 3-5; primary axis of inflorescence often with 50 or more
flowers; flowers pink to white, the inner petals always tipped with
deep red or purple; outer petals usually having a well-developed wing
margin, the hood with a low or obsolescent crest; stigma approxi-
mately rectangular, longer than broad; fruit oblongoid, elliptical,
rarely obovoid; seeds about 2.5 mm. in diameter, obscurely papillose
under magnification.

B. Plants about 10 dm. tall; wing margin of the outer petals lacking

or narrow, the unspurred outer petal acute; California 2. C. Cascana

ssp. Cascana

BB. Plants 8-20 dm. tall (except ssp. 2b); wing margin of the outer
petals moderately to very highly developed, the unspurred outer
petal not acute,

C. Outer petals rounded, sometimes mucronulate, the wing margin
scarcely folded back upon the hood.

D. Plants mostly 10-15 dm. tall; flowcn pink or white; outer

petals mucronulate; Colorado .2a. C. Cascana

ssp. BramJcgci

DD. Plants mostly 4-10 dm. tall; flowers white; outer petals oc-
casionally barely retuse, not mucronulate; Utah 2b. C. Cascana

ssp. brachycarpa
CC Outer petals cmarginate, tbe wing margin folded back upon
the hood.

D. Inflorescence not profusely branched; spurred petal 18-24
mm. long; wing margin very highly developed, not erose;

northeastern Oregon and southern Idaho 2c. C. Cascana

ssp. Cusickli
DD. Inflorescence profusely branched; spurred petal 16-20 mm.
long; wing margin moderately developed, minutely erose;
northern Idaho 2d. C. Casca7ia

ssp. hasfafa

As here understood' this section includes only two species, C. Scoideri of
coastnl Oregon, Washington, and Vancouver Island, and C. Cascana of widespread
but sporadic occurrence in many mountainous districts of western United States.

1947]

OWNBEY- — MONOGRAPH OF CORYDALIS 199

Under C. Caseana are included several variants which hitherto have been regarded
as distinct species. These variants are essentially identical with respect to colora-
tion of the flowers and detailed morphology of the inner petals, stigmas, fruits, and
seeds. They differ appreciably in what are better considered as minor characters,
such as the development of a wing margin on the outer petals, length of spur, and
length of pedicel. The leaves and gross size of the plants vary to some extent
among the different elements of the species, but the taxonomic value of these must
be discounted as about the same type and degree of variability are found in other
species of the genus.

Corydalis Caseana is an excellent example of the type of morphological
divergence commonly met with when component parts of a species are isolated
geographically. The subspecies might be thought of as incipient species whose
modified genetic make-up and consequent morphological divergence have not yet
reached the species level. In another sense they might be thought of as remnants
of a species which through isolation have lost a large portion of the genetic vari-
ability present in the ancient stock.

The members of this section have well-defined habitat requirements, any
deviation from which is sufficient to prevent survival. The plants grow in or
near a continuous source of fresh, running water, in springs, along small creeks,
and in the case of ssp. Brandegei also in wet, open, subalpine forests. All require
considerable sunlight for best development, but at the same time will tolerate some
shade. Plants growing in the sun tend to have smaller, more firmly textured
leaves. C. Scouleri grows at elevations of sea level to about 2500 feet. The sub-
species of C. Caseana grow at elevations of 3000—11,000 feet.

The time required for these plants to reach flowering size is not known. At
one locality I have seen seedlings of at least three size classes. These classes very
likely correspond to age intervals of one year, yet the largest of the seedlings was
still relatively small. It therefore seems probable that these plants do not attain
flowering size until they are four years or more old.

1. C. Scouleri Hook. Fl. Bor. Am. 1:36. t. 14. 1829.

Corydalis macrophylla Nutt. apud Torr. & Gray, Fl. N. Am. 1:69. 1838; Torr. & Gray,

1. c. 665. 1 840, as syn.
Capnodcs Scouleri Ktze. Rev. Gen. 1:15. 1891.
Corydalis Allenii Feddc, Rep. Spec. Nov. 10:478. 1912.
Capnoides Scouleri Thornber ex Fedde, in Engl. & Prantl, Nat. Pflanzenf. ed, 2. 17b: 13 3.

1936, nom. nud. in synon.

Perennial from a rhizome; stems 1 or more, usually 5—10 dm. tall, branched
above; stem leaves widely divergent, about 3, very long-petiolate; blades pinnately,
more or less ternately, compound, 1 dm, or more long and broad, the primary seg-
ments once or twice pinnatifid or incised, the ultimate segments variable, some-
times broadly elliptical or less commonly ovate or obovate with rounded apices, or
sometimes narrowly elliptical with acute apices, but more often intermediate, 1—8

[Vol. 34

200

ANNALS OF THE MISSOURI BOTANICAL GARDEN

I

cm. long, 0,5—4.0 cm. broad, minutely apiculate; inflorescence not strongly mono-
podia!, consisting of 1 or more simple racemes or sparingly branched panicles
arising from the axils of the stem leaves, each raceme usually with fewer than 25
flowers; bracts obscure, the lowermost narrowly elliptical, the upper much re-
duced, linear; pedicels erect, 2-5 mm, long; sepals ovate or broadly lanceolate,
laciniatc or toothed, 1—2 mm. long, deciduous at anthesis; flowers light to deep
pink, the Inner petals not tipped with purple; spurred petal usually somewhat
arcuate, 20-25 mm. long, the lanceolate spur 14-20 mm. long, the regular crest
moderately to very highly developed, extending to and beyond the acute tip of
the hood, wing margin absent; unspurred outer petal 12-15 mm. long, naviculate,
the crest similar to that of the spurred petal; Inner petals usually 9-11 mm. long,
the blade much broader at the apex, the slender claw equalling the blade in length;
stamen spur one-half to two-thirds the length of the petal spur, bent or hooked at
the apex; stigma roughly triangular, the width at the lower lobes about the same
as the length along the medial line; style slender, about 3 mm. long; fruit obovoid,
10-15 mm. long, 4-5 mm. broad; seeds black, about 3.5 mm. in diameter, con-
spicuously papillose under magnification.

This species is limited In distribution. It Is found In wet, cool habitats of
northwestern Oregon, northward to Vancouver Island. Morphologically, it is most
easily distinguished from C. Caseana by its generally highly developed crest and
absence of a wing margin on the outer petals. The obovoid fruits most typical of

Map 1. Distribution of CoryJalh Scouleri Hook.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 201

this species are rarely approached in C. Caseana, and its approximately triangular
stigma may be contrasted with the nearly rectangular stigma found in that species.
Finally, the seeds of C. Scouleri are considerably larger and more distinctly

papillose.

Within the species there is considerable morphological variability, especially
with respect to the leaves. The very small, narrowly elliptical ultimate leaf seg-
ments found on some specimens are in part the basis for Fedde^s proposed segregate,
C. Allenii, which I cannot maintain on valid grounds. C. SconJcri also is variable
with regards to its flowers, particularly in length of spur, amount of curvature of
the spurred petal, development of the crest, and gross size. When considered
against the background of the species as a whole these variants lose their systematic
Significance. As possible examples of population variability potentialities of the
species from locality to locality they are of considerable interest.

Moist, shady woods, especially along water courses; Vancouver Island, western Wash-
ington and northwestern Oregon at elevations of sea-level to about 3 500 feet. Flowers
about April 15 to June 15; fruits about May 15 to July 30.

British Columbia: Vancouver Island.

Washington; Clallam, Clark, Grays Harbor, Jefferson, King, Mason, Pacific, Pierce,
Thurston, and Wahkiakum counties.

Oregon: Benton, Clackamas, Clatsop, Columbia, Hood River, Marion, Multnomah,
Tillamook, and Washington counties.

2. C. Caseana Gray ssp. Caseana G. B. Ownbey, stat. nov.

CorydaJis Caseana Gray, in Proc. Am. Acad. 10:69. 1874.

W

1880.

Capnodcs Bidiuelliannm Greene, Fl. Fran. 280. 1891.
Capnodes Caseanum Greene, I. c. 1891,
Capnodes Caseanum Ktze. Rev. Gen. 1:14. 1891.

F

Glaucous perennial; stems 1— several, generally 10 dm. or less in height; stem
leaves about 5, pinnate, the primary segments again once or twice pinnatifid or
deeply incised; ultimate segments narrowly to broadly elliptical, mostly 1—2 cm.
long, apiculate; inflorescence paniculate, consisting of a strong numerous-flowered
central axis and 1 to several shorter, fewer-flowered lateral axes, these often again
branched at the base; bracts inconspicuous, usually narrowly elliptical, rarely
broader in outline, the lowermost about 10 mm. long, greatly reduced upward;
pedicels semi-erect, 3-5 mm. long; sepals variable, sometimes with a broad base
and a very long-attenuated central lobe, sometimes orbicular and denticulate at
the margin, sometimes otherwise, 2—4 mm. long, rarely persisting through anthesis;
flowers light pink or probably also white, the inner petal tips reddish-purple;
spurred petal often curved, usually 16-22 mm. long, rarely longer, the spur grad-
ually to rapidly tapered to the blunt apex, 12-16 mm. long, the hood crested, the
crest regular or denticulate, rarely obsolete, extending to and beyond the acute
tip of the petals, the wing margin, if present, narrow, regular or more or less
denticulate and not folded back toward the hood; unspurred outer petal 10-12
mm. long, the crest and margin similar to that of the spurred petal; inner petals

202

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

usually 9-10 mm. long, the claw 3-5 mm. long; stamen spur one-half to three-
fourths the length of the petal spur; stigma approximately rectangular, with 8
papillary stigmatic surfaces; style about 3 mm. long; fruit oblong, 10-15 mm.
long, 3-4 mm. broad; seeds dark brown to black, minutely papillose under magnl-
fication, about 2.5 mm. in diameten

Morphologically, the subspecies is best distinguished by its narrow or obsolete
wing margin, its generally narrower, curved spurred petal, and acute apices of
the outer petals. It Is probably most closely similar to ssp. Brandegci from which
it differs appreciably In the smaller gross size of the plants, smaller flowers, and
narrower petal margins. The variant named C. B'ulwelliac by Watson is of no

/

J M^^l ' ^k^P^*

i'-..

/

/

i

f?-

I' 4

t ■

*

■ '•fc

*«■■-*

* M ■

.■*v4 4 ■, a/^ ^^4 t^

t, \

_.■%

t 4

I

f \

■

>

:*->

ioV

^vrr:

\

X

'A

^

">

\

'J

— . •

M. >
/\?.> ^

■ ■^

I

>

'-v^

lI'V

\

f^yP

1/

i

t

^^ ^:n I

h. U

I ^.

'» I,

'*

.^

V

« *.

. J

i

e

/

^51

^

V ^.- \/»

A

i

i\

-It

I

I

^'^,^

r>

/

V /

I

_h r

/ 1

Pv

*

^

/

\»

^.

'"it

.-' V*

t

Ir^

/?

;

^

Caseana Ownbey
IP ssp. Brandegci (Wats.) Ownbey
ssp. brachycarpa (Rydb.) Ownbey

ssp. CusickJi (Wats.) Ownbev

ssp. hastata (Rydb.) Ownbey

11/2^^ ^

(^.

V

7

^^

3V\

Map 2. Distribution of CoryJalis Caseana Gray,

1947]

OWNBEY MONOGRAPH OF CORYDALIS 203

systematic Importance. Leaflet size, petal margin, crest and slenderness of the

spur upon which Watson*s proposed species was based vary to some extent even at
a single locality.

This plant has been collected as far south in California as Truckcc, by Sonne,
but there are no recent collections from this area.

Very moist, often shady situations, in springs and on gravel bars, in and along streams;
southeastern Shasta County southward and eastward to Placer County, California, at ele-
vatlons of about 4000-6000 feet. Flowers from about June 1 to July 30; fruits from
about July 1 to September 1.

California: Butte, Lassen, Placer, Plumas, Sierra, and Tehama counties.

^

2a. C. Caseana Gray ssp. Brandegei (Wats,) G. B. Ownbey, stat. nov.

Corydalis Brandegei Watson, Bot. Calif, 2:430. 1880,
Capnoides Brandegei Heller, Cat. N. Am. Pi, 55. 1898,

Glaucous perennial; stems 1— several, 5-15, mostly 10—15, dm. or more in

height; stem leaves about 5, the lowermost sometimes 10 dm. long, pinnate, the
pinnae once or twice pinnatifid or deeply incised, the ultimate segments mostly
elliptical, 1—5 cm. or more long, apiculate; inflorescence paniculate, consisting of
a central numerous-flowered axis and often 1— several fewer-flowered secondary
axes, these sometimes again branched; bracts inconspicuous, narrowly elliptical to
linear, much reduced and minute upward; pedicels semi-erect, about 5 mm. in
length, up to 10 mm. in fruit; sepals 2-3 mm, long, ovate or orbicular, the margin
irregularly toothed; flowers light pink to white, the inner petals tipped with deep
red or purple, inverted, the spur often nearly upright along the raceme; spurred
petal 18—25 mm, long, the spur 12—16 mm. long, the hood crested, the crest low
and regular, extending to and beyond the rounded apex of the petal to form a
short beak, the wing margin broad, scarcely folded back toward the hood; un-
spurred outer petal about 12 mm. long, the crest and wing margin similar to
that of the spurred petal; inner petals 10—12 mm. long, the claw 4—5 mm, long;
stamen spur about two-thirds the length of the petal spur; stigma approximately
rectangular, with 8 papillary stigmatic surfaces; style about 3 mm. long; fruit
oblong, 15—18 mm. long, about 5 mm. broad; seeds dark brown to black, about
2.5 mm. in diameter, minutely papillose under magnification.

This subspecies is distinguished most easily on the basis of the wing margin,
which is broad, regular, not rctuse at the apex, and not appreciably folded back
upon the hood as in ssp. Ciisickii and ssp. ha%tata. It sometimes appears so when
the flowers are distorted in pressing. The manner in which the low, regular crest
extends beyond the rounded apex Is characteristic. Occasionally the spur is nearly

erect and the fruits, when mature, reflexed nearly to the pedicels. This situation
is found especially in plants from the southern portion of the range.

Subspecies Brandegei is very abundant in Colorado from the summit of Wolf
Creek Pass, Mineral County, for approximately 4 miles down the west side, at
elevations of about 10,000-10,800 feet. Only a few plants are present on the

[Vol. 34

204 ANNALS OF THE MISSOURI BOTANICAL GARDEN

east side of the pass. At this site the flowers arc uniformly pinkish-hivcndcr in
color. At similar elevations in Kcbler Pass, Gunnison County, the plant also is
abundant, especially on the west slope for at least two miles from the summit.
The plants arc essentially identical to those at "Wolf Creek Pass except with respect
to flower color. Here there Is a preponderance of very light, nearly white-flowered
individuals.

Very moist, suhalpine situations, especially along water courses, at elevations of about
8,000-11,000 feet; Gunnison and Delta counties, Colorado, southward to northern Rio
Arriba County, New Mexico. Flowers from about June 10 to August 10; fruits from
about July 10 to September 10,

Colorado: Archuleta, Conejos, Delta, Gunniion, tlinsJale, and Mineral counties.

New Mexico: Rio Arriba County.

2b, C, Caseana Gray ssp. brachycarpa (Rydb.) G. B. Ownbey, stat. nov.

Capnoidcs brachycarpnm Rydb., in Bull. Torr. But. Club 34:426, 1907.
Corydalis brachycarpa Fedde, Rep. Spec. Nov. 10:3 15. 1912.

Glaucous perennial; stems 1-6, 4—10 dm. tall; leaves 3—5, the lower ones long-
petiolate, pinnate, tbe pinnae once or twice pinnatifid or incised; ultimate leaf
segments elliptical, acute at both ends, usually 1—3 cm. long and 0.4-1.0 cm,
broad, minutely aplculate; inflorescence paniculate, consisting of a stout central
axis and often one or more sccondarj^ axes; bracts linear, the low^er ones about 15
mm. long, much reduced above; pedicels stout, spreading to semi-erect, about 5

I

mm. long at flowering time, up to 10 mm. or more long at fruiting time; sepals
ovate or broadly lanceolate, more or less undulate or toothed at the margin, 3-5
mm. long, sometimes persisting through anthesis; flowers white, the inner petals
tipped with deep red or purple; spurred petal 18—22 mm. long, the spur straight,
gradually narrowed to the blunt apex, 9—12 mm. long, the wing margin broad,
undulate, stiff, not appreciably folded back toward the hood, rounded at the apex,
occasionally barely retuse, the crest obsolescent or lacking; unspurrcd outer petal
12-14 mm. long, the margin and crest similar to that of the spurred petal; inner

petals 9—11 mm. long, the claw about 4 mm. long; stamen spur one-half to two-
thirds as long as the petal spur; stigma approximately rectangular, with 8 papillary
stigmatic surfaces; style about 3 mm. long; fruit oblong in outline, about 12 mm.

long and 4 mm. broad; mature seeds not seen.

Subspecies brachycarpa Is a well-marked unit. It is best distinguished morpho-
logically on the basis of the broad, spreading wing margin of the outer petals
which are commonly neither acute-tipped nor emarglnate, but rounded, at the
apex. The broad margin is very w^ell developed even In the bud. The plant pos-
sibly is closest to ssp. Cn^ickJi, but in addition to the above-mentioned differences,
It IS only about one-half as large. The leaves are very similar to those of ssp.
Cusickii as found in Oregon,

The name brachycarpa is something of a misnomer if it was intended to call
attention to a fruit difference between this and other members of the complex.

1947J

OWNBEY MONOGRAPH OF CORYDALIS 205

The normal fruits, althougli perhaps smaller than those ordinarily found in other
subspecies, are in no way significantly different. It is probable that the name was
applied because of a misinterpretation of the swollen fruits very commonly found
on living plants. All such fruits examined were found to contain an insect larva,
the adult counterpart of which has not been identified. The stimulation of
growth of pathogenic tissue results in a globose, spongy, abnormal, sterile fruit.
A similar situation is not uncommon in C. aiirca.

This subspecies is of very limited distribution, and it is possible that the adult
population numbers no more than a few hundred individuals.

On gravel bars along stream courses at elevations of about 8500-10,000 feet; "Wasatch
Mountains, Salt Lake and adjacent Utah counties, Utah. Flowers from about July 1 to
July 30; fruits from about July 30 to August 30.

Utah: Salt Lake and Utah counties.

2c. C. Caseana Gray ssp. Gusickii (Wats.) G. B. Ownbcy, stat. nov.

4

Corydalh Cusickii Watson, Bot. Calif. 2:430. 1880.

Capnoides Cusickii Heller, Cat. N. Am. Pi. 55. 1898.

Corydalh Hendersonii Fedde, Rep. Spec. Nov. 12:278. 1913; not Hcmsl.

Corydalh idahoensh Fedde, 1. c. 16:19 5. 1919.

Glaucous or green perennial; stems 1— several, 8—15 dm. tall; leaves 4-6, pin-
nate, the pinnae once or twice pinnatifid or deeply incised; ultimate leaf segments
usually narrowly, sometimes broadly, elliptical, apiculate, 1—5 cm. long, 0.5—1.5
cm. broad; inflorescence paniculate, consisting of a stout central axis bearing
numerous flowers and 1— several shorter, fewer-flowered secondary axes; lowermost
bracts verj^ narrowly to broadly elliptical or obovate, often 15 mm. long, much
reduced and usually linear above; pedicels spreading, 5-10 mm. long at flowering
time, often up to 15 mm, long at fruiting time; sepals ovate to lunate, often

toothed or laclnlate, 2—4 mm. long; flowers white or tinged with pink, the Inner
petals tipped with deep red or purple, the apices of the outer petals widely diver-
gent; spurred petal 18-24 mm, long, the spur generally straight, 10—14 mm. long
and not rapidly tapering to the blunt apex, the crest, when present, low and in-
conspicuous, the wing margin extremely broad, deeply notched at the apex and

folded back upon the hood; unspurred outer petal 12-15 mm. long, the margin
similar to that of the spurred petal; inner petals 9—11 mm, long, the slender claw
about 4 mm. long; stamen spur straight, one-half to three-fourths the length of
the petal spur; stigma approximately rectangular, with 8 papillary stigmatic

surfaces; style 3—4 mm. long; fruit oblong-elliptical, 10—15 mm. long, 4-5 mm.
broad; seeds dark brown, minutely papillose under magnification, about 2.5 nim.
in diameter.

The highest development of the wing margins of the outer petals found any-
where within the species is present In ssp, Cusickii, This and the long pedicels are
Its most distinctive features. The wing margin and emarglnate apex Is emulated
on a much lesser scale by ssp. hastata. Both ssp. brachycarpa and ssp, Brandcgei
have broad margins, but they never are greatly emarglnate at the apex and the

/

[Vol. 34

206 ANNALS OF THE MISSOURI BOTANICAL GARDEN

margins are stiflfer and never appreciably folded back upon the liood when tlie

flowers are fresh.

This plant was described by Watson from material collected by Cusick in
the Wallowa and Blue mountains of northeastern Oregon. It still is collected

occasionally in the Wallowa Mts. and is present in some abundance above Cornu-
copia. It has been collected a few miles above Sumpter in the Blue Mts. and if it
is at all abundant in this area it is at this point or further south. In Idaho, a
variant with broader bracts, generally broader leaflets, and less-branched inflores-
cence is common in some localities. The Idaho form was described as C. Hcndcr-
sovit by Fedde. It is found along the tributaries of the South, Middle and North
Forks of the Boise River and the South Fork of the Payette River at known eleva-
tions of 5000-7500 feet. It does not often occur along the larger streams. On a
hillside about 6 miles northeast of Rocky Bar, Elmore County, there is d pure
stand covering approximately one-half acre. The plant here reaches its maximum
development,

L

Growing in and along springs and small streams at elevations of about 5000-7500 feet;
mountains of southwestern Idaho and northeastern Oregon. Flowers from about June 15
to July 30; fruits from about July 1 to August 15.

Idaho: Boise, Camas, Elmore, and Valley counties.
Oregon: Baker and Union counties.

2d. C. Caseana Gray ssp. hastata (Rydb.) G. B, Ownbey, stat. nov.

CapnoiJes hasfafvm Rydb.. in Bull. Torn Bot. Club 34:426. 1907.
CorydaJis hastata Fedde, Rep. Spec. Nov. 10:315. 1912.
CorydaVn Cusickii var. hastata Fedde, 1. c. 12:279. 1913.

Glaucous or green perennial; stems 1-several, 10—18 dm. tall; main stem leaves
3-5, the lower ones widely divergent from the stem, the blade deltoid, pinnate,
the pinnae once or twice pinnatifid or deeply incised; ultimate leaf segments
broadly elliptical, ovate- or obovate-elliptical, rounded or acute at the ends, usually

1.5-4 cm. long and 0.5-1.5 cm. broad, minutely apiculate, of very thin, tissue-
like texture when dry; inflorescence paniculate, delicately and profusely branched,
consisting of a main axis and several more or less branched secondary axes; bracts
often foliose, ovate to obovate, somewhat reduced above; pedicels about 5 (5-10)
mm. long, semi-erect or spreading; sepals about 2 mm. long, usually with an
elongate, lanceolate medial lobe and two basal auricles which arc often somewhat
toothed at the margins; flowers pale pink to white, the inner petals tipped with
deep red or purple; spurred petal 16-20 mm. long, the spur 10-12 mm. long,
straight or incurved, gradually narrowed toward the bfoad, blunt apex, the wing
margin moderately well developed, reflexcd toward the hood, commonly arose,
retuse at the apex, the low crest extending over the apex of the hood into a short
beak; unspurred outer petal 10-12 mm. long, the margin and crest similar to that
of the spurred petal; inner petals 7-9 mm. long, the stout claw about one-third
the total length; stamen spur two-thirds to three-fourths as long as the petal spur;
stigma approximately rectangular, with 8 papillary stigmatic surfaces; style 2-3

1947]

OWNBEY MONOGRAPH OF CORYDALIS 207

mm. long; fruit oblong, 12-16 mm. long, about 4 mm. broad; mature seeds dark
brown, minutely papillose under magnification, about 2,5 mm. in diameter.

Although characterized by numerous morphological differences of greater or
lesser value, this plant must be included with C. Caseana in the broad sense. It is
best distinguished morphologically on the basis of the broadly spreading deltoid
leaf blades, the broadly elliptical ultimate leaf divisions, and the profusely branched
inflorescence. In floral characters it most closely resembles ssp. Cusickii^ but
differs from it in having considerably smaller flowers with shorter inner petals and
outer petals with a much narrower, usually erose wing margin which, as in that
subspecies, is reflexed upon the hood and usually is emarginate at the apex. Only
in ssp. Caseana is the wing margin narrower. The sepals are, indeed, as noted by
Rydberg in his original diagnosis, somewhat characteristic. Sepals in Corydalis
are, however, a very much reduced organ, and there is everywhere considerable
variability in outline. I believe that the sepals of ssp. hastata cannot justifiably be
given much weight as a distinguishing feature.

This subspecies is of limited and as yet not definitely circumscribed distribu-
tion. It is found in Idaho from southwestern Shoshone County, southward and
eastward to northern Idaho County, probably only at medium elevations. It is
especially abundant along Orograndc Creek, Clearwater County. It has been re-
ported from the upper reaches of the Selway River (Moose Creek Trail), but its
presence there should be confirmed.

Very wet situations, in and along streams at elevations of about 3000-4000 feet;
mountains of northern Idaho, Flowers from about June 15 to July 30; fruits from about
July 15 to August 30.

Idaho: Clearwater, Idaho, and Shoshone counties.

Section II. Pes-gallinaceus
Sect. Pes-gallinaceus Trmisch, in Abh. Nat. Ges, Halle 6:273, 1862.

Corydalis §. II. Capnites DC. Reg. Veg. Syst. Nat. 2:115. 1821.
Pistolochia Bernh. Syst. Verz. Pfl. 57. 1800.
Bulbocapnos Bernh. in Linnaea 8:469. 183 3.

KEY TO THE SPECIES

A single representative in North America .3. C. pauciflora

3. C. pauciflora (Steph.) Pers. Syn. PL 2:269. 1807; Cham. & Schlecht. in
Linnaea 1:560. 1826, not Edgew.

Fumaria pa7iciflora Steph. ex Willd. Sp. PI. 32:861, 1803.

Corydalis pauciflora 7 parvi flora Regel, in Bull. Soc. Mosc. 343:136, i861.

Capnodes pauciflorum Ktze. Rev, Gen. 1;14. 1891.

Capnoides pauciflorum Gov. in Brooks et al. Rec. Cape Nome & Norton Bay Reg. 170.
190L

^//,

1919.

Perennial; root deep-set, tuberous, usually bifurcate, having a central channel,
the fibrous rootlets mostly at the base, accessory buds at the summit 1-several, or
lacking; stems usually 1-3, unbranched, erect, mostly 8-20 cm. tall, often with
1-2 basal cataphylls; basal leaves none; stem leaves 2-5, long-petioled, simple, the

[Vol. 34

208

ANNALS OF THE MISSOURI BOTANICAL GARDEN

blades ternately divided, the segments again incised into 2-4 (usually 3) lobes,
the ultimate lobes elliptical; peduncles stout, terminal, with 3-5 inverted flowers
crowded at the summit; bracts ovate to obovate, 4-10 mm. long, 3-5 mm. broad,
the lowermost larger; pedicels stout, erect, 4-10 mm, long at flowering time, up
to 20 mm. long at fruiting time; sepals scarious, fugacious, 1-2 mm. long and
broad, variously toothed; flowers blue, often tinged with purple; spurred petal
17-20 mm. long, the hood short and inconspicuous, the low, regular crest extend-
ing to or nearly to the obtuse apex of the petal, the wing margin narrow, the spur

7

near

12

mm. long, nearly as broad basally as apically, the apex 1-2 mm. longer than that
of the other petals, the crest similar to that of the spurred petal, the margin re-
flexed; clawed inner petals 8-10 mm. long, the slender claw occupying one-half or
more of the total length, the blade obovate; stamen spur clavate, two-thirds to
three-fourths the length of the petal spur; stigma triangular, broadest at the 4-
lobed apex; fruits reflexed," about 12 mm. long and 5 mm. broad, elliptical to
obovate; seeds turgid, black, shiny, essentially smooth under magnification.

This distinctive species is well known to students of boreal floras. It is the
only Asiatic species of Corydalh whose distribution extends across the Bering
Straits into America. It is beyond the scope of this paper to discuss names pro-
posed for variant forms of the species found in Asia. An excellent discussion of
the application of the name to the American plant is to be found in Hultcn's

t work (Flora of Alaska and Yukon 5:810. 1945).

#

0"?

Map 3. Distribution of Corydalh pauci flora (Steph.) Pers.

1947]

OWNBEY— MONOGRAPH OF CORYDALIS 209

In tundra; islands of the Bering Sea and Straits eastward throughout Alaska to the
Yukon and northern British Columbia at elevations of sea level to about 3 500 feet; also
widely distributed in Asia. Flowers from about June 1 to July 15; fruits from about
July 1 to August 1.

Alaska: Ft. St. Michaels, Norton Sound, 1865-66, Bannister (G, US); Nome, July,
1890, Blaisdell 6j (UC) ; Seal Islands, 1875, Bryant (US); Anvik, near the Mission, Lower
Yukon River, June 11, 1924, Cha]^man I (NY); Mission premises, Anvik, without date,
Chapman 28 (G) ; near Chinik, Seward Peninsula, July 3, 1900, Collier (US); St. Paul
I., Bering Sea, July 9, 1899, Coville 6 Kearney l8lO (US) ; Port Clarence, July 12, 1899,
Coville & Kearney 1966 (US); Hall I., Bering Sea, July 14, 1899, CoviUc 6 Kcarytey
2033 (G, US); McKinley Park Sta., Mt. McKinley Nat. Park, June 4, 1932, Dixon 17
(UC); roadside, Igloo Creek, same locality, June 13, 1932, Dixon 25 (UC) ; White River
Valley, near the boundary, 1909, Eaton (US); St. Matthew I., July 8-13, 1916, Hanna

(US); on hillside, Goodnews Bay, July 14, 1919, Harrington 57 (US); St. Paul I., June
30-Aug. 20, 1910, Heath (D) ; Nome, 1914, Hill 65 (US) ;^ wet brook banks, Karluk,
June 14, 1901, Home (NY); St. Paul L, Aug. 1, 1897, Kincaid (UC) ; St. Paul I., with-
out date, Maclntyre (G); St. Paul I., July, 1892, Macoun (NY); St. Paul I., July, 1891,
Macovn (M, G, US); St. Paul I., June 29, 1914, Maconn (NY, US); Cape Lisburne, Aug.
13, 1931, Mason (M, UC, NY, G) ; Iviktook Lagoon, St. Lawrence L, July 10, 1931,
Mason (UC) ; Old Man Creek, a branch of the Kovukuk, 4 mi. above camp, near Caribou
Mt., July 6, 1901, Mcndenhall (US); between Yukon River, Nation River, and Inter-
national Boundary, 1930, Mertie 60 (US); damp moss in small gulch, open land near
Teklanika River, Mt. McKinley Nat. Park, 3600 ft. alt., June 24, 1928, Mexia 2040
(UC); Golovin Bay, 1881, Mnrr 168 (G) ; moist thicket near headquarters, Mt. McKinley
Nat. Park, May 31, 1939, Nelson ^ Nelson W-2206 (RM) ; Nelson L, July 6, 1921,
Palmer 194 (US); St. Paul L, June 10, 1890, Palmer 178 (US); near Karluk, Kodiak L,
June 1, 1897, Rutter (D) ; same locality. May 23, 1897, Rutter g2 (D) ; same locality,
June 13, 1903, Rutter lyg (US) ; same locality, June 14, 1903, Rutter 2o6 (M, US) ; Mt.
McKinley Nat. Park, June 13-22, 1937, Scamvian 620 (G) ; Camp Retreat, June 28,
1886, Stoncy (US); Anvil Mt., vicinity of Nome, June 29, 1918, Thornton 319 (US);
damp hillside near creek, Tanana, June 14, 1914, Thouscn 6 (DU) ; St. Paul L, July 9,
1899, Trelease 6 Saunders 3872 (M) ; Hall L, July 4, 1899, Trelease 6 Saunders 3873
(M); St. Matthew L, July 15, 1899, Trelease & Saunders 3874 (M) ; St. Paul L, July
28, 1895, Trtie & Prentiss 12 (NY, G, US); Noatak, July, 1929, Wagner (US); vicinity
of Port Clarence, July 16, 1901, Walpole 1457 (US); same locality, July 18, 1901, Wal-
pole 1467 (US); St. Paul L, July, Aug., 1879, White (G).

British Columbia: Mountains near head Iskut River, Cassiar Dist., July 30, 1910,
Preble ^ Mrxter 6ig (US).

Yukon: Across Bonanza Creek, Dawson, June 19, 1914, Eastwood 307 (G, US); 24-
mile house, Dawson, June 25, 1914, Eastwood 380 (CIUC, G, US).

Section III. Eucorydalis

Sect. Eucorydalis Prantl, in Englcr & Prantl, Nat. Pflanzenf. 3^:144. 18 89

Corydalis §. III. Capnoides DC. Reg. Veg. Syst! Nat. 2:122. 1821.

F

KEY TO THE SPECIES AND SUBSPECIES

A. Flowers pink, the petals tipped with yellow, the hood not crested,
the claw of the inner petals much longer than the blade; stigm.i not
distinctly 2-lobed, with 4 papillary stigniatic surfaces; fruits erect,
very slender, usually 3 0—3 5 mm. long; seeds about 1 mm. in diam-
eter; Georgia to Newfoundland, British Columbia, and Alaska.. 4. C. sent per vir ens

A A. Flowers pale to bright yellow throughout, the claw of the inner
petals equalling or shorter than the blade; stigma 2-Iobcd, each lobe
having 3 papillary stigmatic surfaces; seeds 1.5—2.0 mm. in diameter.

I Vol. 34

210 ANNALS OF THE MISSOURI BOTANICAL GARDEN

B. Spurred petal 7—9 mm. long, the hood having a high, undulate or
toothed crest, the spur incurved, about 2 mm. long; fruits broadly
liticar, usually straight, pendent on very long pedicels; central to
eastern United States 5. C. flavula

BB. Spurred petal 10-22 mm. long (in normal flowers), the spur not
appreciably incurved, usually 4-8 mm. long; fruits erect, on rela-
tively short pedicels (except in sp. 9).

L

C, Spurred petal 16-22 mm. long, the hood with a very high crest,
the wing margin very broad; fruits densely beset with trans-
parent, clavate pustules; southwestern Missouri to central Texas.. <?. C. crysfaUina

CC. Spurred petal 1 0—1 8 mm. long, the wing margin moderately
broad to narrow; fruits essentially glabrous, although sometimes
obscurely granulose along the sutures.

D. Plants often bearing clcistogamous flowers; spurred petal of
normal flowers 10-15 mm. long, the hood with a low, regular,
undulate or obsolescent crest; seeds about 1.5 mm. in diam-
eter, nearly smooth under magnification.

E. Normal-flowered racemes not greatly exceeding the leaves,
often short; spur usually somewhat globose at the tip;
fruits often stout» commonly 10-15 mm. long; central

United States 7. C. micrantba

ssp. micranfha

EE. Normal-flowered racemes often greatly exceeding the leaves,
elongated; spur not globose at the tip; fruits slender,
15—3 mm. long.

F. Stems usually weak and not strongly striate when dry;
foliage green to glaucous; fruits 15-20 mm. long; south-
central to southern United States 7a. C. ntjcra7}iha

ssp. ausfralis

FF. Stems usually stout and strongly striate when dry; foliage

glaucous; fruits 25—30 mm. long; coastal south Texas 7b. C. inicrantha

ssp. tcxcnsh

DD, plants seldom bearing clcistogamous flowers (except in sp.
10); spurred petal mostly 14—18 mm. long; seeds about 2
mm. in diameter, essentially smooth to variously decorated
under magnification.

E. Seeds distinctly murlcate or murlculatc under magnification;
central Texas to southern Kansas.

I

F. Hood crcstless or with a moderately well-developed crest;
fruits 26—34 mm. long, usually abruptly acute; seeds dis-
tinctly muricate under magnification; central and western

Texas 8. C. ntrvisiliqna

ssp, cnrvisiliqna

FF. Hood with a well-developed crest; fruits 20-25 mm.
long, gradually tapered; seeds muriculate under magni-
fication; north Texas to southern Kansas 8a. C curvhiliqiia

ssp. grandihracteata
EE. Seeds never muricate though sometimes muriculate at the
margin under magnification,

F. Racemes usually surpassed by the leaves; hood sometimes
crested; fruits spreading or pendent, usually 18-24 mm.
long; seeds with no ring margin; northern United States
to Alaska and southw^ard in the Rocky Mountains to
Mexico 9. C. anrea ssp. aurca

FF. Racemes usually surpassing the leaves (except in sp. 10);
hood usually not crested; fruits erect, often incurved;
seeds usually having a ring margin.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 211

G. Cleistogamous flowers not present; fruits stout, in-
curved, usually 12-20 mm. long; southwestern United

States and adjacent Mexico 9a, C. aurea

ssp. occldenfalh
GG. Cleistogamous flowers present; fruits slender, not
strongly incurved, 25-3 mm. long; mountains of
eastern Mexico 1 0. C. pscudomicrantha

P

fi

somewliat apart. C. scmpcrvircnsy althougK properly placed in this section, is quite
distinctive in several ways, and conceivably could provide the basis for a subsection.
Fedde (in Engler & Prantl, Nat. Pflanzenf. 17b: 129. 1936.) includes all of these
species in his subsection Eucapnotdes of Eucorydalis.

4. C. SEMPERVIRENS (L.) Pers. Syn. PL 2:269. 1807.

Finnaria sempervircris L. Sp. Pi. 2:700. 1753.

Neckcria sempervirens Neck. Elem. Bot. 3:60. 1790.

Fumaria glauca Curt. Bot. Mag. 5: /. lyg. 1792.

Capnoides glauca Moench, Meth. Pi, 52. 1794.

Capnoides sempervirens Borkh. in Roem. Arch. f. Bot. 1^:44. 1797.

Corydalh glauca Pursh, Fl. Am. Sept. 2:463. 1816.

CoryJalis rosea Eaton, Man. Bot. 79. 1817.

Corydalis annua Hoffmeg. ex Steudcl, Nomen. Bot. ed. 2. 1:423. 1841, as syn.

Neckeria glauca Millsp. Fl. W. Va. 327. 1892. (W. Va. Agr. Exp. Sta. Bull. 2),

Very glaucous biennial; stems usually 1, approximately 30—8 cm. tall, much
branched, erect; earlier cauline leaves long-petioled, crowded; later cauline leaves
nearly sessile, much reduced; leaf blades pinnate, the basal with 5 main segments,
the upper with 3 main segments, the segments ternately divided, then once or
twice incised, the ultimate lobes oblong-elliptical, obtuse, apiculate; Inflorescence
a raceme or panicle, terminal, each axis 1- to 8 -flowered; bracts narrowly elliptical,
minute, 2—5 mm. long, 0.5—1.0 mm. broad; pedicels slender, erect, 5-20 mm.
long at maturity, successively shorter upward; sepals 3 mm. or less long, ovate,
short-attenuate, white or tinged with pink; flowers pink, the petals tipped with
yellow; spurred petal 10-15 mm, long, the saccate spur 3—4 mm. long, blunt, the
very short hood not crested, the wing margin minute but relatively broad, folded
back upon the hood; spurlcss outer petal 10—13 mm. long; inner petals 9—12 mm.
long, the slender claw 6—8 mm. long, occupying about two-thirds of the total
length, the blade obovate, much broader near the apex, having a very high, angular,

r

medial fold; stamen spur blunt, 1 mm. or less long, one-third the length of the
petal spur; stigma slightly broader aplcally, with 4 papillary stigmatic surfaces;

fruits erect, the slender body 25-50 (usually 30-3 5) mm. long, straight or some-
what curved, many-seeded; seeds about 1 mm, in diameter, black, shiny, turgid,
distinctly decorated under magnification, the margin obtuse.

C. sempervirens Is recognized easily by its erect habit, its divaricately branched
stems, its pink and yellow flowers, its very slender, erect or spreading fruits, and
numerous minute, decorated seeds. It has no close relatives In America,

This widely distributed species has no discernible geographical variants. Two

[Vol. 34

212

ANNALS OF THE MISSOURI BOTANICAL GARDEN

local variants, however, have been observed. Plants collected by Bartlctt at Med-
ford, Mass., have exceptionally stout, incurved fruits, and a collection made by
Ehlcrs at Prentis Bay, Michigan, has unusually small flowers. It Is doubtful if
either of these variants is of nomenclatorial consequence.

In shallow, often dry soil, rock ledges, crevices, and talus, and on burned or otherwise
disturbed areas, at elevations of about 500-5000 feet; northeastern Georgia to Maine and
Newfoundland, thence westward to Montana and British Columbia and northwestward
to Alaska. Flowers throughout the summer months from about May 15 to September
15; fruits from about June 1 to September 30. '

Map 4. Disirlbution of CorydaVn scmpcrilrcns (L.) Pcrs.

Gi£OKGIA; Rabun County.

South Carolina: Pickens County.

North Carolina: Alexander, Buncombe, Burke, Caldwell, Forsythe, Haywood,
Henderson, Jackson, Macon, Mitchell, Transylvania, Watauga, and Wilkes counties.

Tennessee: Carter County.

Kentucky: Bel! and Harland counties.

Virginia: Augusta, Bedford, Carroll, Giles, Lee, Madison, Loudoun, Page, Pulaski,
Rappahannock, Rockingham, Slienandoah, and Smythe counties; Shenandoah National
Park.

West Virginia: Grant, Mineral, Monroe, Pocahontas, Preston, Raleigh, and Webster
counties,

Maryland: Allegany, Frederick, and Garrett counties.

Pennsylvania: Adams, Bedford, Bucks, Elk, Fayette, Indiana, Lancaster, Luzerne,
Lycoming, Monroe, Montgomery, Northampton, Perry, Philadelphia, Union, Westmore-
land, and York counties.

1947]

OWNBEY— MONOGRAPH OF CORYDALIS 213

New Jtrsey: Bergen, Essex, Hunterdon, Ocean, Passaic, Sussex, and Warren counties.

New York: Albany, Dutchess, Essex, Franklin, Greene, Herkimer, Jefferson, Nassau,
Orange, Putnam, Rensselaer, St. Lawrence, Saratoga, Tompkins, Warren, Washington, and
Westchester counties,

Connecticut: Fairfield, Hartford, Middlesex, New Haven, and New London counties.

Rhode Island; Providence County.

Massachusetts: Berkshire, Essex, Franklin, Hampden, Hampshire, Middlesex, Nor-
folk, Suffolk, and Worcester counties.

Vermont: Addison, Caledonia, Chittenden, Orange, Rutland, and Windham counties.

New Hampshire; Carroll, Cheshire, Coos, Grafton, and Hlllsboro counties.

Maine: Aroostook, Cumberland, Franklin, Hancock, Knox, Lincoln, Oxford, Penob-
scot, Sagadahoc, Somerset, Washington, and York counties.

Ohio: Lake and Portage counties.

Indiana: Lake and Starke counties.

Illinois: Cook, LaSalle, and Ogle counties,

Michigan: Cheboygan, Crawford, Emmet, Ingham, Keweenaw, Leelanau, Mackinac,
Marquette, Muskegon, and St. Clair counties.

Wisconsin: Adams, Ashland, Bayfield, Brown, Oneida, Polk, Sauk, Sawyer, Shawano,
and Vilas counties.

Minnesota: Aitkin, Beltrami, Carlton, Cass, Chisago, Clearwater, Cook, Itasca, Lake,
Millelacs, and St. Louis counties.

Montana: Flathead County; Glacier National Park.

Newfoundland: gravelly railroad embankments, Grand Falls, July 4, 19n, Fentald
& Wiegand 5455 (NY, G) ; sandy terraces, n. bank of river above the falls, Bishop*s Falls,
valley of Exploits River, July 28 & 29, 1911, Fernald & Wiegand 5456 (G); dry woods,
Buchan (?) Junction, July 13, 1930, Jaussa?i (G) ; railway, Gambo, July 14, 1893, Wag-
horne 21 (UA).

Nova Scotia: digby co. — clearing at border of deciduous woods, Wentworth Lake,
Sept. 4, 1921, Fernald & Long 23866 (G). Inverness co. — dripping cliffs, Big Intervale,
July 17, 1941, Roland 41416 (G). lunenburg co. — recently burned clearing w. of
Bridgewater, Aug. 18, 1921, Fernald ^ Long 23865 (G).

Prince Edward Island: prince co. — dry clearings, Alberton, July 11, 1912, Fernald

d St. John /502 (WS, NY, G, US).

New Brunswick: kent co. — Bass River, June 10, 1869, Fowler (G).

Quebec: brome co. — Mt. Elcphantis, Brome, July 30. 1902, Pease 606 (G) ; dry
mountain ledge, Bolton, July 25, 1926, Knotdton (G). chambly co.— St. Hubert:
Tourbieres, environs de Montreal, June, 1915, Victorin 2o6 (M, WS, US), charlevoix
CO. — vicinity of Cap a L'Aigle, July 27, 1905, Macotin (G). gaspe co. — alluvial woods,
York River, July 29, 1905, Williams, Collins G? Fernald (G). kamourasra CO.- — -dry,
quartzite hills, Ste. Anne, July 14, 1922, Fernald d Pease 25088 (G). lake st. john
DisT. — carriere de granlt, Roberval, July 20, 1921, Victorin 15757 (G). megantic co.
dry, serpentine slopes and crests of Caribou Hill, Black Lake, Aug. 26, 1915, Fernald G?
Jackson I20g8 (G). nissisquoi co. — rocky places, Philipsburg, June 22, 1910, Edmond-
sott 4995 (G), montmagny co. — Grosse-Ile, Testualre du Saint-Laurent, July 31, 1935,
Victorin, Kolland -Germain, Rousseau ^ Meilleiir 40082 (G). pontiac co. — Ile-des-
Soeurs, Lake Timiscaming, June 26, 1918, Victorin 8365 (US). Richmond co.— dry
ledge, Cleveland, July 30, 1923, Chamberlain 6 Knowlton (M, G). rimouski co. — dans
un champ prcs du chemin du cap a TOrignal, Aug. 19, 1927, Rousseau 26971 (CIUC).
TERREBONNE CO. — sur Ics gneIss laurentiens, St. Jerome, July 4, 1930, Victorin ^ Rolland-
Germain 33122 (RM, G). two mts. co. — La Trappe, Oct. 9, 1926, Louis-Marie (G).
CO. UNCERTAIN — on rocks along Matamck River, n. shore, July 26, 1927, Bowman 247
(G); Lac Kamatose, sur le ballast de la route, 101 milles au nord de Mont-Laurier, Aug.
23-25, 1941, Victorin, Rolland -Germain ^ Dominique 260 (G).

Ontario: algona dist. — thin soil, e. ridge, Havilland Bay, 47'' 00' N., 84° 45' W.,
Aug. 12, 193 5, Taylor, Hosie, Fitzpatrick, Losee ^ Leslie 1310 (US), carleton co.
Cascades, vicinity of Ottawa, July 31, 1920, Victorin IOO55 (WS). frontenac co.

214 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

TIMISKAMING DIST.

Battersea, June 13, 1898, Eilmondson lJo6 (NY); Kinsston, May 30, 1901, Fowler (G,
US); Barricficld, June, 1897, Boyd (M). kt-nora dist. — Minakl, July 25, 1915, Thomp^
son 30 (M). LEEDS CO.— Jones Falls, June 7. 1895, Vou'ler (US), muskoka dist.
crevices of rocks, Lake Joseph, Muskoka, Aug. 20, 1881, Burgess (M) ; Lake Muskoka,
Aug. 16-18, 1898, Topping (US), nipissing dist. — common on sunny rocks, Twin
Islands, Timagami region, July 24, 1926, Anderson G? Anderson 26ojg (M, NY, G) ;
Cache Lake, Algonquin Park, July 4, 1900, Macoun (US); open, grassy woods, Sturgeon
Falls near Lake Nipissing, Aug. 14, 1937, Nelson ^ Nelson 239^ (RM). parry sound
DIST.— Island 74 in French River, July 5, 1939, Drircy I (US); in soil pockets on granite
ridge, s. side of French River Harbor, n. w. part of Parry Sound, Sept. 6, 1932, Grassl 3766
(NY). RENFREW CO.— bluffs, Bonne Chere Mts., July 20, 1899, Umbach (US), thun-
der bay dist.— Munj^o Park Point, Nipigon Lake, I9I2, VuUing (G) ; crevice of altered
lava, rocky knoll, flat e. of Schrelber, Aug. 21, 1937, Hosic^ Losee ^ Bannan 1412 (G);
dry ledges, S. Slate Island, July 6, 1933, Pease & Bean 23,540 (G); black loam along
Amadis River, 1912, Pulling (G); diabase crevices, Shangolna Island, Sibley Tp., 48° 20'
N., 88° 50' W., July 6, 1936, Taylor, Losee 6 Bannan 505 (D).
Moose River Basin, 1903, Bell (G).

Manitoba: Selkirk dist.— Elk Island, Lake Winnipeg, July 20, 1887, Macoun (NY).
DIST. uncertain— Piguitonay (mile 214), route of Hudson Bay Railway, July 8, 1917,
Emerton (G) ; Lake Winnipeg Valley, 1857, Bourgeau (G).

Saskatchewan: north shore, Athabnska Lake, July 26, 1920, Laing 1/4 (US); ex-
posed rocky slopes, Chariot Pt., Lake Athabaska, about 59° 36' N., 109*' 13' W., June 12,
1935, Raup 6088 (G); in clay and sandy soil. Sulphide Lake (Lac la Ronge), Oct. J,
1941, Studer 4-16 (CIUC).

Alberta: athabaska dist.— Egg Lake, Athabaska Delta, July 18, 1920, Harper 53
(US); Smith Landing, June 13, 1903, Preble d Gary 13 (US); Granite Hill, Gov. Hay
Camp district, Slave River, about 59° 31' N., 11° 28' W., Wood Buffalo Park, Mackenzie
Basin, Aug. 14, 15, 1928, Raup 2443-a (G); short distance e. of Sand Point, n. shore of
Lake Athabaska, about 58° 57' N., IIO' 42' W., 700 ft. alt., Sept. 1, 1932, Raup & Abbe
4531, PEACE RIVER DIST.— Notikewin, Peace River region, roadside in poplar woods, 57°
N., 118" W„ July 12, 1931, Moss 2255 (WS).

British Columbia: cariboo dist, — Hargreaves Ranch, 2900 ft. alt., Mt. Robson,
Aug. 19-26, 1943, Scamman 3272 (G, US); In brush on rocky slope, Campbell Island in
Summit Lake, 31 mi. n. of Prince George, Aug. 1, 1941, Weber 2600 (WS, M, NY, G,
US). COAST dist. — Bute Inlet, without date, Anderson (WS). kootenay dist. — Revel-
stoke, May 27, 1890, Macoun (US); deserted log road, Revelstoke, July 6, 1905, Shaw
830 (NY, G, US), new WESTMINSTER DIST. — Chcak Kamis, June 25, 1920, ex herb.
Anderson (WS) ; Mons, P. G. E. Railway, June 20, 1916, Macoun (NY, G).

Yukon: n. side of Moose Creek near Clark's Peak, 3 500 ft. alt., Mayo District be-
tween Stewart and MacMillan Rivers, Aug. 7, 1939, Bostock 60 (G) ; Klondyke, Aug. 23,
1898-1901, Maclean (UC, G) ; Moosehide Mt., Dawson, July 14, 1902, Macoun (NY);

Bonanza Creek, Aug. 11, 1899, Tarlefon i;8a (NY); Dawson, July 17, 1898, Williams
(NY). ' J 7 .

Alaska: edge of airfield, Franklin, Fortymile dist., July 16, 1941, Anderson ^ Gasser
7320 (RM, G); dry ground. The Birches, 55 mi. below Tanana, on the Yukon River,
July 8, 1902, Brooks (G) ; roadside near Knik, Oct., 1913, Ghaney 151 (M) ; Eagle to
Valdes trail, June 30, 1902, Gollier 72 (US); headquarters, Mt. McKinley Nat. Park,
June 28, 1932, Dixon 45 (UC) ; Hot Springs on the Tanana River, July 28, 29, 1909,
Hstchcock (US); Rampart, July 26, 1901. Jones 67 (US); Mt. McKinlcy Nat. Park,
summer. 1932, Kaye 1501 (UC) ; Dall River Trail, 3 mi, above Dall City, Ft. Hamlin,
Yukon River, to Bergman, Koyukuk River, June 29, 1901, Mendenhall (US); n. of super-
intendent's office, 3000 ft. alt.. Mt. McKinley Nat. Park, July 23, 1928. Mexia 2106 (M,
UC, D, NY, G, US) ; exposed hillsides above spruce woods, e. side of Wonder Lake, near
center of n. boundary of Mt. McKinley Nat. Park, Aug. 14. 1928, Mexia 2240 (M, UC,
D, NY, G, US); McKinley Park Station, July 31, 1922, Mnrie (US); rocky soil near
Park Headquarters, Mt. McKinley Nat. Park, June 22, 1939, Nelson ^ Nelson W-2151

N

1947]

OWNBEY MONOGRAPH OF CORYDALIS

215

(RM) ; open woods just below Park Headquarters, Mt. McKinley Nat. Park, July 3, 1939,
Nehan ^ Nelson ^622 (RM, G); roadside near Park Headquarters, Mt. McKinley Nat.
Park, July 16, 1939, Nelson Gf Nelson 3833 (M, RM, IH, NY, G) ; Fairbanks, June, 1927,
Palmer i/'Sg (US); recent clearings and open woods, Goldstream Creek and Pedro Dome,
51 mi. n. of Fairbanks, 65° N., 147" 30' W., 800-2000 ft. alt., June 13, 1926, Porsrld &
Porsild 13^ (G) ; open, forested bottom lands, Kokrines Mts., n. side of divide, towards
Melozitna River, 65° 20' N., 154'' 30' W., 800-4000 ft. alt., June 23-July 5, 1926,
Porsild d Porsild y3g (G); hill s. of Mitchell Creek, 3000 ft. alt., Copper River region,
Aug. 6, 1902, Poto 121 (US); Anchorage area. May 23, 1943, York 4 (M) ; sandy hill-
side near Palmer, July 5, 1943, York Pa2o8 (M) ; Mt. McKinley Nat. Park, 63*" 43' N.,

W., J

(G) ; Livengood, about 80 mi. n.-n.w, of

149^ 15'

Fairbanks, June 19-21, 1940, Scamman I/3S (CIUC, G) ; Gens de Large (Chandler

River) & Koyukuk rivers, 1899, Schradcr (US),

5. C. FLAVULA (Raf.) DC. Prod. Syst. Nat. 1:129. 1824.

Flint aria flavula Raf. in Dcsv. Jour. Bot. 1:224. 1808,

Corydalis aurea a flavula Wood, Am. Bot. & Fl. 34. 1870.

Corydalis flavidtda Chapm. Fl. S. U. S. Suppl. 1:604. 1883, sphalm.

Capfiodes flavidum Ktze. Rev. Gen. 1:14. 1891.

Neckeria flavula Millsp. Fl. W. Va. 327. 1892 (W. Va. Agr. Exp. Sta. Bull. 2).

Corydalis Geyeri Fedde, Rep. Spec. Nov. 10:311. 1912.

Green or glaucous winter annual; stems 1— several, sympodial, commonly 15—30
cm. tall, erect while young, often prostrate or ascending when older; basal leaves
long-petioled; cauline leaves sliort-petioled to almost sessile, bardly reduced in
size upward; leaf blades pinnate, with 5-7 segments which are again plnnatifid into
about 5 lobes, these again incised; ultimate leaf segments narrowly to broadly
elliptical, subapiculate, varying greatly in size; racemes equalling or barely ex-
ceeding the leaves, commonly 6- to 10- or more flowered, sometimes poorly de-
veloped; cleistogamous-flowercd racemes, when present, inconspicuous, 1- to 5-
flowered; floral bracts broadly to narrowly elliptical, 6-12 mm. long, 3-7 mm.
broad, the lowermost often foliose or variously incised, becoming entire and re-
duced upward; pedicels slender, erect at anthesis, reflexed In fruit, 6—15 mm. or
more long; sepals scarious, fugacious, about 1 mm. long, lanceolate; flowers pale
yellow, somewhat crowded; spurred petal 7—9 mm. long, the hood crested, the
crest high, undulate or toothed, the wing margin well developed, also un-
dulate or toothed, the incurved spur about 2 mm. long; spurless outer petal
6—8 mm. long, the crest and wing margin as in the spurred petal; clawed

Inner petals 5-7 mm. long, the claw 2-3 mm. long, the blade approximately

twice as broad near the apex as at the distinctly lobed base; stamen spur less than
1 mm. long, less than one-half the length of the petal spur; stigma broader than
high; fruits reflexed or variously disposed, 14—22 (often 18—20) mm. long,
straight, essentially glabrous; seeds about 2 mm. in diameter, black, shiny, on
magnification seen to be concentrically, submuricatcly decorated on the narrow,
acute ring margin.

This species is easily distinguished by the small, crested flowers, the very short,
incurved spur, and long, reflexed pedicels. Clelstogamous-flowered plants are col-

I

I

216

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

lectcd occasionally, and these often have much broader and larger ultimate leaf
segments and weaker, more diffusely branched stems. Such plants can be deter-
mined accurately by fruit characters alone.

Moist, loose soil, wooded slopes and bottom lands, at elevations up to about 2000 feet;
Connecticut and New York to North Carolina westward to northern Louisiana, eastern
Oklahoma, Kansas and Nebraska. Flowers in early spring from about March 15 to May
15; fruits from about April 1 to June 1.

Map 5. Distribution of CoryJalis flavula (Raf.) DC.

Connecticut: Middlesex County.

New York: Nassau, Onondaga, Rockland, Ulster, and Yates counties.

New Jersey: Camden, Hunterdon, Mercer, and Somerset counties.

Pennsylvania: Allegheny, Franklin, Huntingdon, Lancaster, Montgomery, Perry,
Philadelphia, Snyder, Washington, and York counties.

Delaware: Newcastle County.

Maryland: Allegany, Baltimore, Carroll, Cecil, Hartford, Howard, Montgomery, and
Prince Georges counties.

Virginia: Albemarle, Alexandria, Bedford, Botetourt, Buckingham, Dinwiddie, Fair-

W

Warren, and W

West

North Carolina: Durham, Forsythe, Halifax, and Madison counties.

Woodf

Tennessee: Blount, Davidson, Knox, and Obion counties.

1947]

OWNBEY ^MONOGRAPH OF CORYDALIS 217

Alabama: Tuscaloosa County.

Mississippi: Sharkey County.

Louisiana: Natchitoches and Rapides parishes.

Ontario: Essex County, Point Pelcc and Pelee Island.

Michigan: Kalamazoo County.

Ohio: Clermont, Franklin, Hamilton, Ottawa, Ross, Scioto, and Warren counties.

Indiana: Floyd, Lawrence, Marion, Montgomery, Orange, and Perry counties.

Ilunois: Hancock, Jackson, Mason, Pike, St. Clair, Union, Wabash, and Will counties.

Missouri: Barry, Boone, Butler, Callaway, Camden, Clay, Cooper, Franklin, Howell,
Jackson, Jasper, Jefferson, Lawrence, McDonald, Madison, Maries, Marion, Montgomery,
Morgan, Oregon, Ozark, Perry, Pettis, Phelps, Pulaski, St. Clair, St. Francois, Ste. Gene-
vieve, St. Louis, Shannon, Texas, Warren, and Washington counties.

Iowa: Pottawattomie County.

Nebraska: Sarpy County.

Kansas: Atchinson and Wyandotte counties.

Oklahoma: Adair, Cherokee, LeFlore, McCurtain, and Muskogee counties.

Arkansas: Carroll, Cross, Garland, Jackson, Searcy, Van Buren, and Washington

counties. •

6, C. crystallina Engelm. apud Gray, Man. Dot. ed. 5. 62. 1867; Bot. Gaz.
11:189. 1886,

Corydalis aurea /S. ? crystallina Torn & Gray, Fl. N. Am. 1:665. 1840.
Corydalis crystallina Engelm. ex Torr. & Gray, 1. c. 1840, as syn.
Capnodes crystallinum Ktze. Rev. Gen. 1:14. 1891.

Capnoides Halei Small, in Bull. Torr. Bot. Club 25:157. 1898, as to most of Hale's col-
lection from Louisiana.

Corydallis crystallina var. strictissima Fedde, Rep. Spec. Nov. 10:479. 1912.

Glaucous winter annual; stems l-several, often sympodial, 20—40 cm. tall,
erect or ascending; basal leaves long-petioled; cauline leaves sliort-petioled to

4

sessile, somewhat reduced upward; leaf blades pinnate, the segments pinnatifid and
once again incised, the ultimate lobes broadly lanceolate to linear-lanceolate, sub-
apiculate; primary racemes surpassing the leaves, 8- to 18- (ordinarily 12- to 15-)
flowered, the later secondary racemes fewer-flowered; bracts ovate to ovate-acum-
inate, 5-12 mm. long, 3—6 mm. broad, usually much reduced upward; pedicels
stout, erect, about 1 mm. long; sepals scarious, fugacious, 2 mm. or less long,

broadly ovate to cordate, somewhat attenuate, the margin sometimes incised, es-
pecially at the base; flowers bright yellow, crowded at first, becoming more distant
at anthesis; spurred petal 16-22 mm. long, the hood always crested, the crest very
high, undulate or toothed, the wing margin very broad, reflcxed upon the hood, the
spur 6—8 mm. long, the blunt tip distinctly globose; spurless outer petal 12—14
mm. long, about 3 mm. longer than the inner petals, the wing margin wide, not
reflexed upon the hood, enclosing the margins of the spurred petal in the bud, the
crest as in the spurred petal; inner petals oblanceolatc, 9—11 mm. long, the narrow

claw 4—5 mm. long, the blade about twice as wide at the tip as at the base, the

f

basal lobes small; stamen spur 3.5-5.0 mm. long, clavate, curved or bent near the
apex; stigma about twice as broad as high; style long and slender; fruits erect,
14—18 mm. long, stout, straight or moderately incurved toward the floral axis,

densely beset with transparent, clavate pustules which often break open at

218

[Vol, 34

ANNALS OF THE MISSOURI

GARDEN

maturity or, rarely, glabrate; seeds black, about 2 mm. in diameter, distinctly
submuricatcly decorated under magnification, having no ring margin.

This species is distinguished from all other species of Corydalis by the peculiar
type of pubescence of the fruit. The pustules sometimes appear ligulatc when
desiccated or, as is often the case, when they rupture at maturity. The crest and
margins of the hood of the outer petals are more highly developed than in any
other yellow-flowered species.

Map 6. Distribution of Corydalis crystallhia Engelm

1947]

OWNBEY MONOGRAPH OF CORYDALIS 219

C. crystalUna van strictissima Is a habitat variant and of no systematic value.
The type was collected by F. L. Flarvey in "Orchards, grain fields, etc., Northwest
Arkansas," and was distributed as Curtiss's North American Plants I2^''\ Fedde
cites the collection but gives the number erroneously as 12 Ja, Too, he previously
had described C. micrantha var. diffusa, a synonym of C. micrantha ssp, aiistraUs,
on the basis of the true Curtiss's North American Plants distribution #l2^a, col-
lected by Curtiss himself in Duval County, Florida.

Prairies, fields, open woods, and wasteland; southwestern Missouri to central Texas.

Flowers in early spring from about April 1 to May 15; fruits from about April 1 5 to
June 1.

Missouri: Bates, Benton, Cass, Greene, Henry, Jasper, Lawrence, McDonald, Newton,
St. Clair, and Vernon counties.

Arkansas: Ashley, Benton, Carroll, Drew, Franklin, Nevada, Pope, Sebastian, and
Washington counties.

Kansas: Cherokee and Montgomery counties.

Oklahoma; Atoka, Cleveland, Craig, Haskell, Latimer, LeFlorc, McCurtain, Mayes,
Muskogee, Okmulgee, Osage, Pittsburg, Pontotoc, Pushmataha, Rogers, and Tulsa counties.

Texas: Brazos, Colorado, Denton, Fannin, Grayson, Kaufman, Lamar, Navarro,
Tarrant, Van Zandt, and Victoria counties.

7. C. micrantha (Engelm.) Gray ssp. micrantha G. B. Ownbey, stat. nov.

Corydalh aurca var. micrantha Engelm, apud Gray, Man. Bot. ed. 5. 62. 1867.
Corydalh 7nicrantha Gray, in Bot. Gaz. 11:189. 18 86, in part.
Neckeria micrantha MacMillan, Metasp. Minn. Valley. 2 5 5. 1892.
Cafmoides mrcranthiim Britton, in Mem. Torr. Bot. Club 5:166. 1894.
Corydalis micrantha var. pachysiliquosa Fedde, Rep. Spec. Nov. 10:380. 1912.
Corydalis monilifera var. ferruginifcra Fedde, 1. c. 11:498. 1913.

r

w

Glaucous or nearly green winter annual; stems 1-several, usually 15—25 cm.
tall, erect or ascending, sparingly branched; basal leaves crowded, long-pctioled;
cauline leaves short-petiolcd to nearly sessile, gradually reduced upward; leaf blades
pinnate, the 5-7 primary segments pinnatifid and again incised, the ultimate lobes
oblong-elliptical or obovate, subapiculate; normal-flowered racemes usually present,
slightly exceeding the leaves, 6- to 16-flowercd, not surpassed by the fcwer-flow-
d secondary racemes; cleistogamous-flowcred racemes, when present, incon-
spicuous, 1- to 5-flowcred; bracts elliptical, the lowermost 5-8 mm. long and 2-4
mm. broad, the upper much reduced, often minute on cleistogamous-flowered
racemes; pedicels erect, the lower usually 2-4 mm. long, gradually decreasing in
length upward; sepals scarious, fugacious, 1,5 mm. or less long, ovate, often un-
dulate or toothed at the margin; flowers pale yellow, often somewhat crowded
throughout anthcsis; spurred petal 11-15, usually 12-14 mm. long, the hood
crested, the crest low, undulate or rarely obsolescent, the wing margin well devcl-
oped, the spur 4.5-6.0 mm. long, the apex distinctly globose; spurless outer petal
9—11 mm. long, semi-geniculate, the crest low; inner petals 7—9 mm. long,,
oblanceolatc, the claw 3-4 mm. long, the blade twice as broad at the apex as at the
obscurely lobed base; stamen spur 3-4 mm. long, about three-fifths the length of

-^

y

[Vol. 34

220 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the petal spur, straight or curved, sometimes clavate; stigma 2-lobed, rectangular,
twice as wide as high; fruits erect, commonly 10-15 mm. long, rarely longer, often
shorter in clelstogamous-flowered racemes, straight or moderately incurved; seeds
about 1.5 mm. in diameter, black, shiny, turgid, concentrically but moderately
decorated under magnification, obtuse at the border, with no ring margin.

The subspecies of C. viicrantha are all characterized by very small seeds and
can be distinguished from all other yellow-flowered species by them alone. Sub-
species viicrantha usually can be distinguished from ssp. ausfralis by its less
elongated racemes, generally smaller flowers, globose tipped spur and generally
shorter, stouter fruits. The two subspecies intcrgrade in all of these characters,
especially in southern Missouri and Oklahoma, but in most cases the disposition of a
given specimen is not difficult. In southern Missouri a form is also found which
is characterized by larger, more showy flowers, longer pedicels, and relatively short
fruits. This large-flowered form has been confused by various authors with C.
aurea. As flower size is not considered a good criterion for separation of the segre-
gates of C. micranfha, this form seems better left with the typical subspecies.

A dwarf form of ssp. viicrantha was collected by Reverchon at Columbia,
Brazoria Co., Texas. This locality is far removed from the expected range of the
subspecies and it seems likely that data on the label are mixed. Its presence there
should be verified.

Although encountered occasionally in other species of CorydaVs the clcisto-
gamous condition reaches its highest development In C. micrantha. It occurs at
random throughout the range of the species, A single plant may have only normal
flowers, only cleistogamous flowers, or both cleistogamous and normal flowers. In
the last instance, the cleistogamous flowers arc produced only on the smaller, less
well-developed secondary branches. Plants havhig only cleistogamous flowers are
quite different in aspect, usually being much more profusely and delicately
branched. The racemes are short, weak, and ordinarily have 1-5 small, undevel-
oped, self-fertilized flowers and ultimately the same number of fruits crowded near
the apex. The disposition of such specimens Is likely to prove difficult for one
who is unfamiliar with this type of variation.

The conditions which institute clclstogamy are not entirely understood. It is
notable, however, that plants growing in the shade or those which are crowded
together so that they shade each other are predominantly cleistogamous. Also, age
of the plant is of some significance, as the later racemes are often entirely cleisto-
gamous on plants which at first produced only normal flowers.

Along bluffs, rocky hills, open woods, and river banks, often in disturbed soil; southern
Minnesota to Illinois, Kansas, and northern Texas. Flowers in early spring from about
April 1 to May 15; fruits from about April 15 to June 1.

Minnesota: Fillmore, Goodhue, Hennepin, Murray, Nobles, Olmsted, Pipestone,

Steele, and Winona counties.

Wisconsin: Pepin and Rock counties.

Iowa: Dickinson, Emmet, Floyd, Henry, Jackson, Muscatine, Page, Palo Alto, Polk,
Van Buren, Wapello, and Wright counties.

1947]

OWNBEY MONOGRAPH OF CORYDALIS

221

Map 7. Distribution of Corydalis mlcrantha (Engelm,) Gray ssp. mkrantha Ownbey.

[Vol. 34

222 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Illinois: Henderson, LaSallc, Menard, Peoria, and St, Clair counties.

Missouri: Adair, Boone, Camden, Cass, Cedar, Christian, Dade, Daviess, Gasconade,
Greene, Henry, Iron, Jackson, Jasper, Jefferson, Lawrence, Livingston, McDonald, Marion,
Miller, Moniteau, Phelps, Polk, Pulaski, Ralls, Randolph, St. Charles, St. Clair, Stc. Gene-
vieve, St. Louis, Shannon, Stone, Texas, Wayne, and Webster counties.

Arkansas: Benton, Carroll, Logan, and Washington counties.

South Dakota: Clay County.

Nebraska: Cass, Cedar, Gage, Lancaster, Nance, Otoe, Richardson, and Sarpy counties.

Kansas: Bourbon, Chase, Cowley, Geary, Harvey, Lyon, Miami, Osborne, Riley,
Sedgwick, Shawnee, and Wyandotte counties.

Oklahoma: Carter, Comanche, Kay, Kingfisher, LcFlore, Logan, ATcClain, Murray,
Oklahoma, Payne, Rogers, and Tulsa counties.

Texas: Brazoria, Dallas, and Upshur counties.

7a. C. MiCRANTHA (Engclm.) Gray ssp. australis (Chapm.) G. B. Ownbcy,
Stat. nov.

Corydal'n aurca var. ausfraJis Chapm. Fl. S. U. S. Suppl. 1:604. 1883.

Corydalh micrauiha Gray, In Bot. Gaz. 11:189. 1886, in part.

CapnoiJcs llalei Small, in Bull. Torr. Bot. Club 25:137. 189 8, as to the Curtiss collections

from Jacksonville, Florida, but not as to most of Hale's specimens from Louisiana.
Capfiohhs rampcsfrc Britton, Man. cd. 2. 1065. 1905,
Corydcilis nin'lsiliqua var. fencrior Feddc, Rep. Spec. Nov, 10:365. 1912.
Corydalh micranfha var. diffusa Fcdde, 1. c. 3 80. 1912.
Corydalis micranfha var. lepfosiliqua Fcdde, !. c. 11:497. 1913.

Corydalh camljcstris Buchholz & Palmer, in Trans. Acad. Sci. St. Louis 25:115. 1926.
Corydalis Ilalci Fcrnald & Schubert, In Rhodora 48:207. 1946.

Green or somewhat glaucous annual; stems 1-scvcral, usually 20-40, occasion-
ally up to 60 cm. tall, the earlier usually stouter, semi-erect, the later ascending;
basal leaves crowded, long-pctiolcJ; cauline leaves short-pctiolcd or nearly sessile,

reduced upward; leaf blades pinnate, the 5-7 primary segments pinnatifid and
again incised, the ultimate lobes longer than broad, approximately ovate, sub-
apiculatc; normal-flowered racemes usually present, much surpassing the leaves,
10- to 20-flowcred, not surpassed by secondary racemes; clcistogamous-flowered
racemes, when present, inconspicuous, 1- to 5-flowcrcd; bracts elliptical, usually
less than 8 mm. long and 4 mm. broad, the upper mucK reduced; pedicels erect,
the lower 3—6 mm. long, decreasing in lengtli upward; sepals scarious, fufl^clous,
1.5 mm. or less long, broadly ovate, the margin undulate or toothed especially at
the base; flowers pale yellow, becoming distant during anthcsis; spurred petal 12-14
mm. long, the hood nearly always crested, the crest low, regular or undulate, the
wing margin well developed, the spur 4-6 mm., long, the tip blunt, never distinct-
ly globose; spurlcss outer petal 9-11 mm. long, geniculate, the crest low; inner
petals 8-10 mm. long, oblanceolatCj the claw 3-4 mm. long, the blade twice as
broad near the apex as at the obscurely lobed base; stamen spur 2.5-3.5 mm. long,
about tbree-fifths the length of the petal spur, usually straight, sometimes bent
near the tip, clavate; stigma 2-lobed, rectangular, twice as wide as high; fruits
erect, 15-20, rarely 25 mm. long, slender, straight or moderately incurved; seeds

1947]

OWNBEY MONOGRAPH OF CORYDALIS . 223

about 1.5 mm. In diameter, black, sKiny, concentrically but moderately decorated
under magnification, obtuse at the border, with no ring margin.

Subspecies atisf rails is best distinguished by Its elongate normal-flowered
racemes, its short, saccate spur which is never clearly globose at the tip, Its slender,
erect fruits, and its minute, nearly smooth seeds.

I

The peculiarities of this plant were first recognized by Chapman who in 1883

r

published a short and accurate description of it in the first supplement to his
Tlora' under the name Corydalis atirea var. australis. In 18 86 Gray (Bot. Gaz.
11:189), in his study of C. aiirea and its allies^ concluded that Chapman's variety
belonged with C. viicrantha and reduced it to synonymy under that species. In
conformance with Gray*s treatment Chapman, then, also treated his variety as a
synonym of C. micraniha in the third edition of his Tlora' issued in 1897. The
following year Small rcdescribed Chapman's plant as Capnoides Halei...

Small's description of Capnoides Halei was drawn on the joint basis of Hale's
collection from Louisiana and Curtiss's collections from Florida. Hale's plants
(s. n. in Herb, N. Y. Bot. Gard.), are all C. crystallina except for one small speci-
men which is referable to the subspecies described above. It is evident from the
general aspect of the plants of C. crystallina that they are those referred to by
Small in comparing the "new" species with Capnoides cnrvisiliqmim (Corydalis
ctirvisiliqiia) when he distinguished it from that species by ". . , • its more slender
habit, and especially by the more coarsely dissected leaf-blades."

Also cited by Small was Ciirtiss 45^5 from Jacksonville, Florida, and with the
exception of the sentence quoted above, it is from these Florida specimens that the
description is drawn. They, therefore, should be designated the authentic type if
the species were maintained.

This entity has been the subject of a recent paper by Fernald and Schubert
(Rhodora 48:207. 1946) who, recognizing its distinctness, have revived Small's
name, and transferred it to Corydalis. I do not believe, however, that the dif-
ferences between this and ssp. ttitcrantha are of specific level, and am therefore
taking up Chapman's earlier varietal name australis in its new rank of subspecies.
In support of this view it may be mentioned that seeds of ssp. 7nicrantha and ssp.
australis are identical in size and decoration of the testa, that the flowers are similar

in that both have a low crest, and that the geographical distrubution of the two
subspecies taken as a whole is not unnatural. Even greater variability in habit,
together with similar minor morphological variability of the floral organs and
fruits, is found in other species of Corydalis such as C. aitrea and C. Caseana,

Plants of ssp, anstralis from coastal North Carolina, South Carolina, and
Georgia are appreciably smaller and more strict in habit than those from other
parts of the range of the subspecies. The flowers also are noticeably smaller, the
hood is not crested, and the fruits are very slender and often moniliform.

The center of diversity of this subspecies is eastern Oklahoma and southeastern
Missouri, and plants intermediate between this subspecies and ssp. micrantha and

[Vol. 34

224

ANNALS OF THE MISSOURI BOTANICAL GARDEN

C. curi'isiliqua are not uncommonly collected in this area. The disposition of a
given specimen, liowever, ordinarily is not difficult.

In disturbed, often s:indy soil, abandoned fields and waste areas, along roadsides, and in
open woods; from soutbcrn Missouri and eastern Kansas to Texas, Florida, and North
Carolina. Flowers In early spring, about February 15 to April 30; fruits from about
March 1 to May 15.

Map 8. Distribution of CoryJalis micrantha (Engclm.) Gray ssp. australis (Chapm.) Ownbcy.

North Carolina: Bladen, Brunswick, Craven, Jones, Lenoir, and New Hanover
counties.

South Carolina: Beaufort, Charleston, Georgetown, and Horry counties.

Georgia: Camden, Glynn, and Pulaski counties.

Flork^a: Alaclniaj Duval, Franklin, Hernando, Leon, Marion, Nassau, Putnam, and
St. John counties.

Alabama: Mobile County.

Mississippi: Harrison County.

Louisiana: Jefferson, Natchitoches, Orleans, and Rapides parishes.

Texas: Anderson, Austin, Bastrop, Bell, Brazos, Burleson, Caldwell, Clennan, Dallas,
DeWitt, Frio, Galveston, Gonzales, Grayson, Grcg.s;, Harris, Henderson, Jackson, Kauf-
man, McLennan, Nueces, Rusk, San Augustine, Smith, Tarrant, Travis, Upshur, Victoria,
Waller, "Washington, and Wharton counties.

Oklahoma: Carter, Cleveland, Creek, Logan, Murray, Muskogee, Oklahoma, Payne,
and Pottawatomie counties.

Arkansas; Benton, Crawford, Hempstead, Hot Spring, Jackson, and Pulaski counties.

Missouri; Barry, Carter, Cedar, Dunklin, Jackson, Jefferson, Madison, Mississippi, St.
Clair, Scott, and Vernon counties. ^

Kansas: Miami County.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 225

7b, CoRYDALis MiCRANTiiA (Engelm.) Gray ssp. texensis G. B. Ownbey, ssp

nov.

Herbae annuae glaucae; caulibus 20-45 cm. longis saepe crassis post exsic-
cationem valde stnatis; foliorum laminis pinnatis, segmcntis pnmariis pinnatifidis
incisis ultimis oblongo-acutis subapiculatis; racemis saepe crassis folia superantibus,
bracteis ovato-attcnuatis marglne denticulatis infimis ca. 5 mm. longis 2 mm. latis
supcrioribus aliquid minoribus; floribus flavis primo congcstis in anthcsim
remotioribus, pedicellis erectis patulisve 2-4 mm. longis, 'sepalis fugaceis ca. 1.5
mm. longis ovato-attenuatis, petalo calcarato valde arcuato 12-15 mm. longo
carinae cristo satis humili undulato margine bene manifesto supra cristum inflexo
calcare obtuso Kaud globoso, petalo ecalcarato exteriorl ca. 10 mm. longo margine
baud reflexo duobus interioribus 8-10 mm. longis oblanceolatis ungui 3-4 mm.
longo, lamina apicc quam basi multo latiorl, calcare staminali ca. 2 mm. longo
clavato, stigmate 2-lobato ca. bis longiori latiorl; fructibus erectis vel incurvatis
gracilibus 25-30 mm. longis; scminibus ca. 1.5 mm. diam. sub lente aliquid
ornatis margine obtuso.

Glaucous annual; stems 20-45 cm. long, often stout and strongly striate when
dry, prostrate-ascending; leaf blades pinnate, the primary segments pinnatifid and
again incised, the ultimate lobes oblong-acute, subapiculate; racemes often stout,
surpassing the leaves; bracts ovate-attenuate, denticulate at the margin, the lower-
most about 5 mm. long and 2 mm. broad, the upper somewhat reduced; pedicels
erect or spreading, 2-4 mm. long; sepals fugacious, about 1.5 mm. long, ovate-
attenuate; flowers yellow, crowded at first, becoming more distant during anthesis;
spurred petal strongly arcuate, 12-15 mm. long, the hood with a low undulate
crest, the wing margin well developed, reflexed upon the hood, the blunt spur 5-7
mm. long, not globose; spurless outer petal about 10 mm. long, the margin not
reflexed; inner petals 8-10 mm. long, oblanceolate, the claw 3-4 mm. long, the
blade much broader at the apex than at the base; stamen spur about 2 mm. long,
clavate; stigma 2-lobed, twice as broad as high; fruits erect or incurved, slender,
25-30 mm. long; seeds about 1.5 mm. in diameter, moderately decorated under
magnification, obtuse at the border, with no ring margin.

This well-defined subspecies is endemic to the coastal plain of southern Texas.
It is most closely comparable to ssp. atistralis, but Is easily distinguished by its
longer fruits and more strongly arcuate spurred petal. In habit and foliage It Is
very similar to C. cnrvisiliqna ssp. cnrvhiliqtia with which It is often confused. It
can be distinguished from the latter by Its non-muricate seeds and shorter spur
which is not globose at the tip.

Moist, often ^sandy soil, open ground of alluvial plains and uplands; south coastal
Texas. Flowers in early spring from about February 20 to March 20; fruits from about
March 1 to April 10.

Texas: atascosa co. — moist, alluvial ground, Campbclton, March 10, 1917, Valmer
II2jg (M, G, US). BEE CO. — Bceville, March 30, 1932, Jones 2gj6j (M, type), cal-
HOUN CO. — Bahia del Espiritu-Santo, ex herb. BerlanJler 548, l^gg, igjS (G). Cameron

[Vol. 34

226

ANNALS OF THE MISSOURI BOTANICAL GARDEN

CO.

-Palm Grove, March 3, 1940, Parks 142Q (M). goliad co. — Goliad, Feb., 1927,
Williams II (UT). jim wells co. — sandy loam, about 600 ft. alt., Romarsid Rancb,
March 18, 1943, Freeborn 338 (UT). kendall co. — Edge Falls, March 26, 1938, Parks
2gjOO (G). KLEBERG CO. — Rivicra, Feb. 22, 1930, Harrison (US), live oak co. —
sandy upland, 41 ml. n. of Ahce, March 1, 1944, Painter ^ Barklcy 14461 (UT). NUECES
CO.— Corpus Chrlstl, May, 1913, Orcuti 5S29 (M); Robstown, March 26, 1920, High gi
(M), vicToarA co. — Victoria, April 6, 1900, Eggert (M) ; sandy, open ground, Victoria,
March 4. 1916, Palmer ^064 (M, D, US) ; Victoria-Gohad, March 29, 1930, Tharp (UT).

WILLACY CO.

in open ground, sandy situations, March 21, 1937, Kunyon 161S,

Map 9. Distribution of CoryJalis mtcruutha (Fngclm.) Gray ssp. tcxcnsis Ownbcy.

8. C. curvisiliqua Engclm. ssp. curvisiliqua G. B. Ownbey, stat. nov.

Corydalis aurea var. curvisiliqua Gray, in Proc. Acad. Phila. 1863:57. 1864, nom. nud.
Corydatis curvisiliqua Eni^clm. ex Gray, I. c. 1864, nom. nud. in synon.; apud Gray, Man.

Bot. ed. 5: 62. 1867; Bot. Gaz. 11:188. 1886.
Capnoides curvisiliqua Ktze. Rev. Gen. 1:14. 1891.
Neckeria curvisiliqua Rydb. in Univ. Nebr. Bot. Surv. Nebr. 3:24. 1894.

Glaucous winter annual or perhaps biennial; stems 1-sevcral, the primary often
erect, the 1-sevcral secondary ascending, 10-40 cm. long, often somewhat
branched; basal leaves long-pctiolcd; caulinc leaves short-petiolcd, reduced in size;
leaf blades pinnate, the pinnae twice pinnatifid, rarely again incised, the ultimate
segments oblong, obtuse or rounded; peduncles usually surpassing the leaves, the
primary 6- to 18-, usually about 12-flowercd; the secondary fewer-flowered;
bracts ovate, 10 mm. or less long, 6 mm. or less wide, the lowest sometimes foliose,
much reduced upward; pedicels stout, spreading, 2-3 mm. long; sepals scarious,

' 1947J

OWNBEY MONOGRAPH OF CORYDALIS 227

broadly ovate to ovate-attenuate, often more or less toothed or undulate at the
margin, about 1 mm. long; flowers bright yellow, often strongly arcuate, crowded
on the raceme at first, becoming more distant during anthesis; spurred petal 16-18
mm. long, with a very broad wing margin, the crest absent to well developed and
undulate or toothed, the spur 7-^ mm. long, often somewhat globose at the blunt
tip; spurlcss outer petal 12-15 mm. long, geniculate, about 3 mm. longer than the
inner petals, the crest similar to that of the spurred petal; inner petals oblanceolate,
9-11 mm. long, the slender claws nearly half the total length; stamen spur clavate,
bent near the apex, 4-6 mm, long; stigma 2-lobed, twice as broad as high; style
slender; fruits slender, erect, moderately to strongly arcuate or incurved toward
the floral axis, usually 26-34 mm. long; seeds about 2 mm. in diameter, black,
muricate, with essentially no ring margin at maturity.

Subspecies curvhiliqua is most easily recognized by its extremely long, erect,
incurved fruits, and its seeds which are distinctly muricate under magnification.
The latter character is approached nowhere else in the genus, and, indeed, is the
strongest character upon which the species is based. The tetragonal character of
the fruits mentioned by Gray (Bot. Gaz. 11:189. 1886), although perhaps more
pronounced here, especially in fresh material, is by no means unique.

Floral characters which arc of value In recognizing this subspecies are the well-
developed wing margins of the outer petals, the much-reduced, claw-like basal
portion of the unspurred outer petal, and the well-developed spur which is about
one-half the total length of the spurred petal The degree to which the crest is
developed Is extremely variable; material from the type locality usually has no
apparent crest. Some plants, however, have a moderately well developed crest.
This diversity is general throughout the range of the subspecies.

Disturbed soil, sandy bottoms, abandoned fields, open woods, hillsides, and valleys;
central to western Texas. Flowers in early spring from about March 1 to May 1; fruits
from about March 15 to May 15.

Texas: Bexar, Brewster, Caldwell, Comal, Crockett, Culberson, Edwards, Frio,
Gillespie, Hays, Irion, Jeff Davis, Karnes, Kerr, Kinney, Llano, Maverick, Medina, Presidio,
Taylor, Terrell, Tom Green, Travis, and Uvalde counties.

8a. C. cuRVisiLiQUA Engclm. ssp. grandibracteata (Fedde) G. B. Ownbey,
Stat. nov.

Corydalh niruisHiqua var. grandibractcafa Fedde, Rep. Spec. Nov. 11:291. 1912.

Glaucous winter annual; stems 1-scveral, stout, ascending, commonly 20-30
cm. long; basal leaves numerous, moderately long-pctiolcd; cauline leaves somewhat
reduced, shortcr-petioled; leaf blades pinnate, the primary segments pinnatifid and
usually again incised, the ultimate segments elliptical to obovate; peduncles stout,
surpassing the leaves; bracts conspicuous, usually ovate-acuminate, the lowermost
usually 10-15 mm. long and 4-6 mm. wide, somewhat reduced upward; pedicels
spreading, usually 2-3 mm. long; sepals ovate, variously

too

yellow; spurred petal 15-18 mm. long, having a well-developed wing margin, the
hood crested, the crest conspicuous, regular or undulate, the stout spur 7-9 mm.

228

rVoL. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Map 10. Distribution of CoryJalh curtlulhiuj Engclm.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 229

long, somewliat globose at the tip; spurless outer petal 12-15 mm. long, geniculate,
the basal portion slender, claw-like, the crest similar to that of the spurred petalj
inner petals oblanceolate, 9-11 mm. long, the claw slender, 4-5 mm. long; stamen
spur about two-thirds the length of the petal spur; stigma twice as broad as high;
style slender; fruits slender, erect, incurved toward the floral axis, 20-25 mm. long,
gradually tapered apically; seeds about 2 mm. in diameter, black, having a narrow
ring margin, distinctly muriculate under magnification.

This subspecies is best distinguished by its slender, lanceolate, erect, incurved
fruits, its relatively large flowers, the usually highly developed crest and wing
margin, and the large ovate floral bracts. The muriculate character of the seeds,
so striking in ssp. nirvhiliqua, is here reduced nearly to the condition found in C.
aurea ssp. occidentajis. Hybridization between the two and with C. micrantha
ssp. ansfralis may account for the anomalous nature of many specimens. However,
because of the greatest agreement In floral morphology with C. nirvhiViqtia and
because its range Is a northward extension of that species T believe that it is

properly placed here.

The isolated occurrence of ssp. gravdibracteata in Muscatine Co., Iowa, perhaps
is best explained by a chance introduction of seeds.

Usually in sandy soil, open ground, alluvial plains, roadsides, prairies, and slopes;
southern Kansas to northern Texas; eastern Iowa. Flowers from about April 15 to May
15; fruits from about May 1 to May 3 0.

Iowa: Muscatine County.

Kansas: Chautauqua, Stafford, and Sumner counties.

Oklahoma: Alfalfa, Caddo, Canadian, Cleveland, Comanche, Grady, Kingfisher,
Kiowa, Logan, McClain, Oklahoma, and Stephens counties.

Texas: Archer, Clay, Collin, Dallas, and Navarro counties.

9. C. AUREA "Willd. ssp. aurea G. B. Ownbey, stat. nov.

F

Corydalh anrca Willd. Enum. Hort. Berol. 2:740. 1809.
Fumciria aurea Muhl. ex Willd. 1. c. 1809, as syn.
Ftimaria aurea Ker, Bot. Reg. 1:/. 66, 1815.

Odoptcra aurea Raf. Cat. 15. 1824.

Corydalis montana Engelm. ex Gray, in Mem. Am. Acad. 4:6. 1849, nom. nud. in synon.

Corydalls aurea var. typica Regcl, in Mem. Acad, St, Petersb. 4'*: 19. 1861 (Tent. Fl.

Ussuri. 19. 1861); Bull. Soc. Mosc. 343:145. 1861.
Corydalis aurea var. parviflora Regel, in Bull. Soc. Mosc. 34*^:146. 1861.
Corydalis aurea p. macrantha Wood, Am. Bot. & Fl. 34. 1870.
Capnodes aureiim Ktze. Rev. Gen. 1:14. 1891.

Neckeria aurea Millsp. Fl. W. Va. 327. 1892 (W. Va. Agr. Exp. Sta. Bull. 2).
Corydalis Wetherillii Eastw. in Bull. Torn Bot. Club 29:524. 1902.
Corydalis wyomingensis Fedde, Rep. Spec. Nov. 10:312. 1912,
Corydalis tortisiliqua Fedde, 1. c. 313. 1912.
Corydalis Gooddingii Fedde, 1. c. 1912.
Corydalis hypecoiformis Fedde, 1. c. 314. 1912.
Corydalis Engelmannii Fedde, 1. c. 365. 1912.
Corydalis aurea var. robusta Fedde, 1, c. 379. 1912.
Corydalis monilifera Fedde, 1. c. 417. 1912.
Corydalis washingtoniana Fedde, 1. c. 419. 1912.
Corydalis macrorrhiza Fedde, 1. c. 479. 1912.
Corydalis Albertae Fedde, 1. c. 11:196. 1912.
Corydalis Jonesii Fedde, 1. c, 1912.

. [Vol. 34

230 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Coryilalis orcgana Feddc, 1. c. 290. 1912.

Corydalh densicoma Fcddc, 1. c. 291. 1912.

Capuoides WethcriUii Heller, in Aluhlenbergia 7:123. 1912.

Capiioida enchlamydeum Woot. & Standi, in Contr. U. S. Nat. Herb. 16:122. 1913.

Corydalh forthiliqua var. htigihracteata Fcddc, I. c. 11:497. 1913.

Coryddl'n Engcfmaiuiii var. exalt ata Fcddc, !. c. 1913.

Corydalh hopyroides Feddc, 1. c. 498. 1913.

Corydalh hopyroides var. Mcarnsii Feddc, 1, c. 12:37. 1913.

Corydalh wyomhigcnsh var. lafivaghiafa Fcddc, 1. c. 3 8. 1913.

Capnoides Engelniannii Cockcrell, in Univ. Colo. Stud. 11:216. 1915.

Capnoides macrorrhha Cockerell, 1. c. 1915.

Corydalh eucHamydea Fcdde, 1. c. 18:32. 1922.

Glaucous winter annual or biennial from a more or less branched caudex;
stems sympodial, prostrate-ascending, 10-50, usually 20-3 5 cm. long; basal leaves
long-petiolcd; cauline leaves barely reduced in size upward, also usually long-
petioled; leaf blades pinnate, with 5-7 pinnae, these pinnatifid into about 5 seg-
ments which are again incised; ultimate leaf segments broadly to narrowly ellipti-
cal, 1.5-severaI rimes as long as broad, greatly variable in gross size, subapiculate;
peduncles short, terminal; racemes shorter than to barely exceeding the leaves, the
primary 10- to 30-, usually 10- to 20-flowercd, the secondary 4- to 12-flowered;
bracts elliptical to linear, the lowest 4-10 mm. long and 1-2 mm. broad, rarely
larger, often denticulate at the apex, much reduced upward; pedicels erect when
young, generally reflexed or recurved in fioiit, the lowermost 5-10 mm. long;
sepals scarious, fugacious, broadly ovate or ovate-attenuate, irregularly toothed,
1-3 mm. long; flowers pale to bright yellow; spurred petal 13-16 mm. long, the
hood usually not crested, the crest when present low and incised, the wing margin
moderately to well developed, the spur straight or slightly incurved, 4-5
long, the tip somewhat globose; spurless outer petal 9-11 mm. long, the hood and
crest as in the spurred petal; inner petals 8-10 mm. long, the claw 3.5-4.5 mm.
long, the blade somewhat broader and more distinctly winged distally; stamen
spur 2-3 mm. long; stigma about twice as broad as high; fruits commonly 18-24,
rarely up to 3 mm. long, usually slender, often erect when young, generally
pendent at maturity, straight to moderately arcuate, often moniliform, the valves
often torulose when dry; seeds nearly 2 mm. in diameter, black, shiny, turgid,
obscurely decorated to nearly smooth under magnification, broadly acute at the
edge, with no ring margin.

This subspecies is best distinguished on the basis of the generally weak racemes
and slender, pendent or spreading fruits. Tlie racemes ordinarily do not exceed
the leaves except in early stages of growth. It intcrgrades at times with ssp.
occiJcitalis, but in general can be distinguished without difficulty when the plant

Is in fruiting condition.

Contingent upon the broad view of the subspecies adopted here it has been
necessary to reduce to synonymy a large number of specific and varietal epithets
proposed by Fedde. For the most part, they arc founded upon minor variant forms
which are by no means mutually exclusive. A brief discussion of the proposed

mm.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 231

biological basis for the type of variability found in C. aiirea is given in tbe intro-
ductory material to this paper.

Corydalis aiirea ssp. aarea is of north temperate and subarctic distribution. In
the northern part of its range it is found at low elevations, but in the southern
part it is confined largely to mountainous districts, and may grow at elevations of
11,000 feet or more. Consequent to its wide range and adaptation to a diversity
of habitats, this subspecies has become quite polymorphic. It seems probable that
a good deal of minor genetic differentiation has taken place in each of the isolated
mountain ranges of the Southwest. Ultimately the many forms thus produced
may be of nomcnclatorial rank. At present their nomenclatorial recognition can
add nothing to an understanding of the group.

Among the many recognizable variants which in my opinion are not nomen-
clatorially important the following are mentioned briefly:

Plants from Rimouski County, Quebec, described as C. aurea var. rohiista by
Feddc, are of Interest primarily because of their very follose stems. Plants with
similar foliage are found in southeastern Canada and northern United States as far
west as the shores of Lake Superior and Lake Michigan. Fruit characters empha-
sized by Fcdde do not set these plants apart from the subspecies proper.

At low elevations in Washington, Oregon, Idaho, Nevada, and Wyoming
plants sometimes are found with stout stems, semi-ercct fruits, and short, obtuse
ultimate segments of the basal leaves. These arc all better placed with ssp. avrca
because of the unreliability and probable superficiality of the distinguishing

characters.

In the mountains of Otero and Lincoln counties, New Mexico, is a variant
having very large, foHose bracts. In fruit characters it is intermediate between
ssp. aurca and ssp. occidentaUs, This variant was described as Capnoides cuchla-
viydcum by Wooton and Standley, but I cannot see that the differences are in
any way essential, and am therefore reducing the name to synonymy. ^

Particularly striking variant forms having weak, diffusely branched, leafy
stems, weak, 1- to 4-flowered racemes, and very broad, incompletely divided
ultimate leaf segments occur sporadically throughout the Southwest. Among
many localities where plants of this description are found may be mentioned es-
pecially Crandall Canyon, Carbon County, Utah, the Charleston Mts., Clark
County, Nevada, the Grand Canyon National Park, Arizona, the Santa Catalina
and Chiricahua Mts., Arizona, the Black Range, Grant and Sierra counties. New
Mexico, the Guadalupe Mts., Culberson County, Texas, and the Davis Mts., Jeff
Davis County, Texas. The tendency for plants of this sort to be produced, there-
fore, is widespread. Viewed separately they often appear significantly different,
but against the background of the subspecies proper their significance fades.

Although the outer petals of ssp. aurea ordinarily are not crested, an occasional

eptlon to this generalization is met with in northern and western United States.
The fact that the presence of a crest apparently is not tied up with any constant
morphological difference, together with Its erratic occurrence, leads me to believe
that it does not warrant serious consideration.

*

232

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Map 11. Distribution of CorydaVn aurea Willd. ssp. anrca Ownbcy.

Loose, open, often gravelly soil, lake shores, talus slopes, ledges, rocky hillsides and
creek bottoms, gravel pits, road cuts, and burned-over areas; northeastern United States,
northward and westward to Quebec, the Dakotas, Mackenzie, and Alaska, southward in
the Rocky Mountains to Arizona, New Mexico and western Texas, at elevations of a few
hundred feet In northern United States and Canada to over 11,000 feet in the Colorado
Rockies, Flowers throughout the summer months, from about May 1 to August 30;
fruits from about May 15 to September 15.

Pennsylvania: Snyder County.

New York: Essex, Jefferson, and Tompkins counties.

Vermont: Addinson, Chittenden, Rutland, and Windsor' counties.

New Hampshire: Grafton County.

Illinois: Cook and Winnebago counties,

Michigan : Alpena, Keewenaw, Mackinac,
counties.

Montmorency, Oscoda, and Schoolcraft

Wisconsin: Brown and Door counties.

r

Minnesota: Aitkin, Becker, Beltrami, Carlton, Cass, Chisago, Clearwater, Cook,

1947]

OWNBEY MONOGRAPH OF CORYDALIS 233

Crow Wing, Dakota, Goodhue, Hennepin, Hubbard, Lnke, Meeker, Ottcrtail, Polk, Pope,
Ramsey, Renville, St. Louis, Todd, Wabasha, and Winona counties.

South Dakota: Brookings, Custer, Fall River, Harding, Lawrence,. Meade, and

Pennington counties.

North Dakota: Benson, McLean, Morton, Pembina, and Rolette counties.

Montana: Carbon, Cascade, Chouteau, Deerlodge, Flathead, Gallatin, Jefferson,
Lewis & Clark, Meagher, Missoula, Park, Powell, and Ravalli counties; Glacier National

Park.

Wyoming: Albany, Big Horn, Crook, Fremont, Johnson, Lincoln, Park, Sheridan,

Sublette, Teton, and Uinta counties; Yellowstone National Park.

Colorado: Boulder, Chaffee, Clear Creek, Conejos, Custer, El Paso, Fremont, Gilpin,
Grand, Gunnison, Hinsdale, Jefferson, Lake, La Plata, Larimer, Mineral, Montezuma,
Montrose, Ouray, Park, Pueblo, Rio Grande, Saguache, San Juan. Summit, Teller, and
Weld counties; Rocky Mountain National Park.

New Mexico: Bernalillo, Catron, Colfax, Dona Ana, Eddy, Grant, Lincoln, Luna,
Mora, Otero, Rio Arriba, Sandoval, San Miguel, Santa Fc, Sierra, Socorro, and Taos
counties.

Texas: Brewster, Culberson, and Jeff Davis counties,

Arizona: Apache, Cochise, Coconino, Gila, Graham, Mohave, Pima, and Yavapai

counties; Grand Canyon National Park.

Utah: Beaver, Cache, Carbon, Daggett, Duchesne, Garfield, Iron, Salt Lake, San
Juan, San Pete, Summit, Uintah, Utah, and Wasatch counties.

Nevada: Clark, Elko, Esmeralda, Humboldt, Lincoln, Nye, and White Pine counties.

California: Modoc County.

Idaho: Bannock, Blaine, Bonner, Clark, Custer, Fremont, Kootenai, and Owyhee

counties.

Oregon: Crook, Grant, Harney, Lake, and Wallowa counties.

Washington: Chelan, Douglas, Ferry, Kittitas, Okanogan, Pend Oreille, Spokane,

Stevens, and Whitman counties.

Quebec: bonaventure co. — Restlgouchc River, Matapedia, Aug. 1, 1936, Victorviy
Germain G? Domhnquc 48gg6 (UO, UC, CIUC, G). rimouski co. — Massacre Island, on
conglomerate covered with moss, coniferous woods, Bic, Aug. 12, 1927, Rousseau 26S/I
(US); same locality, June 30, 1927, Rousseau 26401 (M, WS, G); humus in crevices of
calcareous rock, July 8, 1905, Collins 6 Fernald 85 (G, UC, NY, US), temiscaming

DIST. — Polnt-au-vent (Lake Temiscaming) , June 25, 1918, Yictorin 8358 (M, G, US).
>x^RiGHT co.-^Aylmer, May 26, 1901, Fowler (US).

Ontario: algoma dtst. — waste ground by Algoma Central Railway, Gray (Mile
229), June 23, 1921, Pease l8o2Q (G) ; ballast near Coppermine Point, Lake Superior,
July 7, 193 5, Pease ^ Ogden 2jl6l (G). BRUCE co. — Lion's Head, on damp calcareous
rocks, June 11, 1932, Yictorin ^ Prat 45945 (RM, G). carleton co. — vicinity of Otta-
wa, May 28, 1921, Rolland 15761 (WS, NY, US), frontenac co. — Gardiner's Farm,
near Kingston, June 10, 1897, Langford (M). lambton co. — on sides of sand hills, near
Port Franks, May 24, 1906, Dodge I (US), manitoulin co. — dry cliffs, Gore Bay,
Manitoulin Island, July 5, 193 5, Pease d Ogden 25190 (G, US), thunder bay dist.—
rich shore of Lake Superior, about Lat. 48° 45' N., Long. 87° 15' W., 1 mi. n.e. of
Schreiber, Aug. 16, 1937, Hosier Losee ^ Bannan 1413 (G) ; damp diabase ledge. Norma
Creek, Thunder Cape, June 26, 1936, Taylor, Losee & Bannan 504 (CIUC).

Manitoba: marquette dist. — ^Fort Ellice, along the line of the Grand Trunk Pacific
Railv/ay, June 27, 1906, Macoun fjS Herriot (G). portage la prairie dist. — Portage la
Prairie, along the line of the Grand Trunk Pacific Railway, May 28, 1906, Herriot (G);
Carberry, 1898, Thompson (M). dist. uncertain — Piguitonay, Mile 214, route of Hud-
son Bay Railway, July 8, 1917, Emerton (G) ; Charleswood, June 5, 1915, Thompson

97 (M).

Saskatchewan: moose jaw dist. — newly burnt woods, Cypress Hills, June 15, 1884,
Macoun (G). qu'appelle dist. — moist woods, Qu'Appelle Valley, June 26, 1938,
Shevkenek II 5 (G), dist. uncertain — in rich, moist ground, usually in burnt-over
ground, McKague, June 21, 1940, Breitung 577 (M, IH, UT, NY).

/

234 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

Alberta: calgary dist. — gravel banks and rocky hills, Shaganappi, vicinity of Cal-
gary. 3400-3600 ft. alt., May 30, 1913, MooJie 137 (NY. US). Edmonton dist.—
burned area in woods, Edmonton, May 21, 1931, Moss 21 4.0 (WS). jasper nat. park —
Jasper, 3472 ft. alt., Aug. 31-Sept. 2, 1943, Scamwau Jj6q (G, US), medicine hat
DIST. — moist, rocky woods, vicinity of Roscdalc, 2200-2500 ft. alt., May 27, 1915, MooJie
QTT (M, CIUC, D, UT, NY, G, US), red deer dist.— n. e. of Buffalo Lake, May 23,
1926, Brinkman 201 j (US); Sarcce Reserve, June 15 to Aug. 15, 1905, Goddard 48Q
(UC). rocky mts. NAT. park — Bow River Valley, 4500 ft. alt., Banff, June 9-18, 1906,
Broii'fi 62 (M, NY, G, US); roadside near the village, vicinity of Banff, 4500 ft. alt.,
June 19, 1899, McCalla 2124 (NY, US), victoria dist. — grain field, Fort Saskatchewan,
June 10, 1930, Turner (G). dist. uncertain (probably athabaska) — Athabaska
Landing, July 28, 1914, Hitchcock 12064 (US); Fort Chlpewyan, Athabaska, June 5.
1903, Preble d Cary 5 (US) ; muddy river bank along lower Fircbag River near its mouth,
June 3, 1935, Ra7ip 60J3 (G) ; base of eastern slope of Caribou Mts., about 58'' 57' N.,
113° 55' W., and 58° 51' N., 113° 57' W., Wood Buffalo Park, Mackenzie Basin, July 17,
1930, Raup 243g-a (NY, US),

British Columbia: cariboo dist. — Alexis Creek, June, 1914, Newcowbe IQ (G),
cassiar dist. — above Discovery on road to Surprise Lake, July 10, 1930, Sefchcll ^ Parks
(UC); near head of Ingenika River, Sept. 8, 1910, Preble ^ Mixtcr 68gb (US); near
bead Iskut River, July 29, 1910, Preble & Mixter 601 (US), kootenay dist. — near Goat
Creek, 27 mi. n. of Natal, July 4, 1941, Weber 22g6 (M, RM, WS, NY, G); Kicking
Horse Valley, vicinity of Field, 4000 ft. alt., June 20, 1906, Brown 21 4 (M, NY, G, US).
YALE dist. — near Gulchon Creek, 13 mi. s. of Savona, 50° 32' N., 120'' 52' W., about
3500 ft. alt., June 23, 1941, Hitchcock ^ Martin 7412 (M, RM, WS, UC, IH, NY, G);
along Bolean Creek, about 1 ml. n.w. of Falkland, 2400 ft. alt., June 30, 1941, Hitchcock

ef Martin ^48$ (M, RM, WS, UC, IH, NY, G).

Mackenzie: Fort Resolution, no date, Onion, Kennicott ^ Hardisty (NY).

Yukon: Fifty-Mile River, Aug. 4, 1899, Bolton (US); Dawson, June 3, 1914, £^5/-
wood /J? (WS, CIUC, G, US); recent burns, Fort Selkirk, June 13, 1899, Gorman IO23
(NY, US); Klondyke, 1900, Maclean (US); Bonanza Creek, Aug. 11, 1899, Tarleton
49b (NY, US); Walker Gulch, July 16, 1899, Williams (NY); Lake Lebarge, June 23,
1899, Tarleton 4Qa (NY, US).

Alaska: Eagle to Valdes trail, June 30, 1902, Collier 7J (US); vicinity of Copper

Center, 1908, Heideman 66 (US); Hot Springs on the Tanana River, July 28, 29, 1909,
Hitchcock (US); Yukon River country, no date, Ketchum (G) ; banks of railroad cut,
Mt. McKinley Nat. Park, Aug. 2, 1939, Klelson fj Nelson 40TO (M, RM, NY, G) ; Fair-
banks, June, 1927, Palmer ly^O (US); Gopher Center, Copper River region, June 1, 1902,
Poto 14 (US). ' *

9a. C. aurea Willd. ssp. occidentalis (Engelm.) G. B. Ownbey, stat. nov.

Corydalis aurea var.. Gray In Smiths. Contr. Knowl. 5:10. 1853 (Pi, Wright. 2:10).

Cprydalis m&fitana Engelm. apud Gray, Man. Bot. ed. 5. 62. 1867.

Corydalis aurea var. occidentalis Engelm. apud Gray, 1. c. 1867; Bot. Gaz. n:188. 1886,

Capnoides montanum Britton, in Mem. Torr. Bot. Club 5:166. 1894.

Neckeria aurea occidentalis Rydb. in Univ. Nebr. Bot. Surv. Nebr. 3:24. 1894.

Corydalis crassipedicellata Fedde, Rep. Spec. Nov. 10:364. 1912.

Corydalis bilimbata Fedde, 1. c. 379. 1912.

Corydalis chihuahuana Fedde, 1. c. 418. 1912.

Corydalis curvisiliquaeformis Fedde, 1. c. 11:289. 1912.

Corydalis fonesii var. stenophylla Fedde, 1. c. 497. 1913.

Corydalis pseudomicranfha var. Griff ithsii Fedde, I. c. 12;37. 1913.

Corydalis pachyloba Fedde, 1. c. 38. 1913.

Capnoides pachylobum Greene ex Fedde, 1. c. 1913, nom. nud. in synon.

^

Glaucous winter annual or biennial; stems often erect while young, usually
10-25 cm. or more long; basal leaves long-pctioled; cauline leaves few, often some-

1947]

OWNBEY MONOGRAPH OF CORYDALIS 23 5

what reduced in size; leaf blades pinnate, having 5—7 pinnae, these pinnatifid and
again incised; ultimate leaf segments usually oblong, 2—5 times longer than broad,
subapiculate; peduncles usually stout; racemes surpassing the leaves at least in the
early stages of growth, 5- to 20-, usually 8- to 12-flowered; bracts elliptical to
linear, 10 mm. or less long, much reduced upward; pedicels erect, 1-5 mm. long;
sepals scarious, fugacious, ovate, often toothed at the margin, 2 mm. or less long;
flowers mostly bright yellow; spurred petal 14-18 mm. long, the hood usually not
crested, the wing margin well developed, the blunt spur 5-9 mm. long, often
somewhat globose at the tip; spurless outer petal 8—13 mm, long, geniculate, the
hood and margin as in the spurred petal; inner petals 8—11 mm. long, the claw
about one-half of the total length; stamen spur 3-6 mm. long; stigma about twice
as broad as high; fruits 12-30, commonly 16-18 mm. long, erect, stout, curved
upward and inward or obliquely along the floral axis, not moniliform; seeds about
2 mm. in diameter, black, acute at the edge, usually having a narrow marginal
ring which is distinctly reticulate under magnification.

This subspecies is most often confused with ssp. aicrca. The two are best dis-
tinguished by the more strongly monopodial growth form, stouter racemes, gen-
erally larger flowers and longer spurs, and, most important, the stouter, more
strongly curved, erect or semi-erect fruits of ssp. occidentalis. In southwestern
United States ssp. occidentalis is found at lower elevations as a general rule, but
since the seeds of ssp. aitrca often are washed down from the mountains the latter
also sometimes is found at low elevations. The difference between the two rep-
resents a summation of several divergent tendencies which together form a rather
reliable index to the proper disposition of any given specimen. At the same time
the two are segments of a fundamentally heterogeneous species and true inter-
mediates do exist.

Gray referred to this entity in the fifth edition of his 'Manual* but made no
clear choice between the two pames suggested by Engelmann, C, aiirca var. occi-
dentalis and C. vtonfana, as he was undecided whether the plant represented a new
species or variety. In 1866 Engelmann, in a letter to Gray (still preserved at the
Gray Herb.), made the following statement: "If you retain montana as a species
you must keep the name, I suppose, but as a variety of anrea the name of occi-
dentalis is preferable . . .", In 188 6, concluding that the entity was truly a
variety of C. aurea, Gray accepted the name occidentalis and published C. vtontana
as a synonym. In accordance with this point of view, there seems to be no doubt
that the name occidentalis should be retained in its modified rank of subspecies.

The specimens cited by Gray In 18 86 are of historical interest. Fendler's 1847
collections from Santa Fe, New Mexico, are cited first. These specimens appear
to me to be typical but depauperate C. aiirea ssp. aitrea. The fact that they were
first cited has led to their general acceptance as the historical type of C. aurea var.
occidentalis. The second collection cited is Wright IJOQ from El Paso, Texas,
which Gray said is better representative of the entity. This is quite true, and this
collection is typical of the subspecies as understood today. Next cited is Pringle

236

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

ig8 (later taken as the type of C. chihtiabnana Fedde) which is again typical ssp.
occiden talis as are Palmer*s 1865 collections from Arizona (at least as to Cones
& Palmer 2Q4), and Rnsby Q from the Burrow Mountains. The portion of Hall
& Harbour ji, cited last, deposited in the Gray Herbarium, is true ssp. occideutalis;
materia! bearing identical labels deposited in the Missouri Botanical Garden Her-
barium and in the United States National Herbarium is ssp. aiirca.

Loose, often sandy, dry soil, bottom-lands, prairies, plains, foothills and mesas, and

Map 12. Distribution of Corydalis aurea W\\\d, ssp. occidcniaUs (Engelm.) Ownbey.

1947]

OWNBEY MONOGRAPH OF CORYDALIS

237

/

along ditches, railroad embankments and washes, at elevations of about 1000-6500 feet;
southwestern South Dakota and eastern Wyoming to western Oklahoma, Texas, northern
Mexico, and Nevada. Flowers in spring at lower elevations, in summer at higher eleva-
tions, from about March 15 to July 30; fruits from about April 1 to August 15.

South Dakota: Fall River County.

Nebraska: Banner and Dawes counties.

Kansas: Stafford County.

Oklahoma; Beckham, Blaine, Caddo, Canadian, Cimarron, Custer, Grady, Greer,

Jackson, Jefferson, Kingfisher, Kiowa, Texas, and Tillman counties,

Texas: Childress, Comanche, Crosby, Culberson, Dickens, Fisher, Hall, Hudspeth,
Jeff Davis, Kent, Lubbock, Nolan, Reeves, Scurry, and Sutton counties,
Wyoming: Albany and Platte counties.

Colorado: Arapahoe, Archuleta, Baca, Boulder, Denver, Fremont, Garfield, Gunni-
son, Huerfano, La Plata, Larimer, Las Animas, Moffat, Montrose, Ouray, Pueblo, Rio

Grande, and "Weld counties.

Utah: Duchesne, Emery, Garfield, Grand, Millard, Piute, Salt Lake, San Juan, Sevier,

Uintah, Utah, and Washington counties.

Nevada: Lincoln County.

Arizona: Apache, Cochise, Coconino, Gila, Graham, Maricopa, Mohave, Navajo,

Pima, Pinal, Santa Cruz, and Yavapai counties.

Ne>)7 Mexico: Catron, Colfax, Dona Ana, Eddy, Grant, Hidalgo, Luna, McKinley,
Quay, Rio Arriba, Sandoval, San Juan, San Miguel, Socorro, Taos, Torrance, and Valencia

counties.

Chihuahua: Sept., 1934, Dobie 13 (UT) ; Casas Grandes, June 2, 1899, Goldman
433 (G, US); St. Diego, 6000 ft. alt., April 18, 1891, Ilartman 600 (NY, G, US);
Chihuahua, spring, 1936, LeSiieur Mex-^ld (UT) ; Majalca, June 24, 1936, LeSuenr 1207
(M, G); vicinity of Chihuahua, about 4250 ft. alt., April 8-27, 1908, Valmcr 4 (M, NY,
G, US); valley near Chihuahua, March 22, 1885, ?rtngle 19S (UC, NY, G, US— Tsotypcs
of C. chihuahuana Fedde) ; 14 mi. s. e. of Minaca, 6500 ft. alt., July 25, 1937, Shreve
8012 (UA); near Colonia Garcia In the Sierra Madres, 7300 ft. alt., July 25, 1889, Town-
send e? Barber 163 (M, UC, NMA, NY, G, US— Isotypes of C. crassipedicellata Fedde);

Santa Eulalia plains, 18 85, Wilkinson (D).

DuRANGo: San Ramon, April 21 to May 18, 1906, Vahner 72 (M, UC, NY, G, US);
Otinapa, July 25 to Aug. 5, 1906, Valmer ^gg (M, NY, G, US).

Sinaloa: By spring water In shady canyon, near Platano, Sierra Monterey, March 9,
1940, Gentry 5869 (M, UA, NY, G).

Sonora: Bablspe, 5330 ft. alt., Dec. 24, 1890, Hartman 358 (G) ; no definite locality,

1890, Lloyd jdg (G).

State Uncertain: Mexico, no date, Coulter 664 (M, NY, G).

10. C. PSEUDOMicRANTHA Fcdde, Rep. Spec. Nov. 11:499. 1913

i

Glaucous or green biennial (or annual?); stems 1-several, sympodial, usually
20-40 cm. long, prostrate-ascending; basal leaves crowded, long-petioled; cauline
leaves sKort-petioled, hardly reduced upward; leaf blades pinnate, the primary
segments pinnatifid and again Incised, the ultimate lobes elliptical, subapiculatc;
normal-flowered racemes, when present, 6- to 12-flowered; cleistogamous-flowered
racemes abundant, 1- to 5 -flowered; bracts elliptical to obovate, 2-8 mm. long,
1-5 mm. broad, often minute on cleistogamous-flowered racemes; pedicels erect,
the lower 1-3 mm. long, shorter upward; sepals about 1 mm. long and 0,5 mm.
broad, ovate-attenuate; flowers pale yellow, inconspicuous, crowded at anthesis;
spurred petal 10-12 mm. long, the hood crestless, the wing margin narrow, the
spur 3—4 mm. long, not globose at the tip; spurless outer petal 8-9 mm. long,
slender, usually straight; inner petals 7-8 mm. long, narrowly oblanceolate, the

\

[Vol. 34

238

ANNALS OF THE MISSOURI BOTANICAL GARDEN

claw about two-fifths tlic total length; stigma 2-lobcd, rectangular, twice as broad
as high; fruits erect, commonly 2 5-3 mm. long, slender, straight or moderately
curved; seeds about 2 mm. in diameter, black, submuricately decorated under
magnification especially at the often distinct ring margin.

This subspecies is best distinguished by its slender, erect and usually straight
fruits, in contrast to those of C. iiurca ssp. aurca which are mostly pendent and
curved. The presence of cleistogamous flowers suggests an affinity with C-
wicrantha but size and decoration of the seeds indicate that it is more properly
maintained as a distinct species.

Mountains of southern Coahuila to Vera Cruz, Mexico, at elevations of about 7000—
9500 feet. ^Flowers and fruits throut;hout the sprini; and summer months.

Coahuila: Saltillo, Sept., 1898, Palmer J56 (G, US); Sierra Je Parras, 8000-9000
ft. alt., July, 1910, rurpin 4602 (M, UC, G, US, type).

Map 13. Distribution of Corydalis pseudoynicrantha Fedde.

1947]

OWNBEY MONOGRAPH OF CORYDALIS 239

NuEVO Leon: Sierra Madre Oriental; lower San Francisco Canyon, about 15 mi. s. w.
of Galena, 7500-8000 ft, alt., June 12, 1934, Mueller ^Mueller J73 (UT, G).
Tamaulipas: Canyon de Garrapata, April, 1926, Runyon I02I (G, US).
Vera Cruz: Boco del Monte, Aug., 1908, Purpus 30/J (M, UC, NY, G, US).

Introduced Species

C fnfea DC, a European species, was collected at Elk Rock, Multnomah
County, Oregon, M. W. Gormon 4076, June 2, 1917 (WS, D) ; /. C. Nehon 6 M.
W, Gormon I2^Q, same date (G). Gormon made the following comment on the
label: "Com. Probably esc. from cultivation. Native of S. Eur. where it runs
wild as a weed." This apparently is the only recorded instance of a native Euro-
pean or Asiatic species having escaped from cultivation in the United States.
Doubtless It has happened other times, as many Eurasian species are attractive
horticultural curiosities and have been grown in this country. Sporadic occurrence
of such species or of weedy species accidentally introduced is to be expected.

C. Intca has the following characteristics: Leaves thrice tcrnately compound
or incised, the ultimate segments elliptical; flowers yellow, the spur about one-
fourth the total length of the spurred petal; fruit about 10 mm. in length, long-
pedicellate.

Excluded Species

Corydalis hiaurlta Hornem. Hort. Hafn. 2:668. 1815 = Dicentra sp.

C. bracteosa Spreng, Syst. Veg. ed. 16. 3:162. 1826 = Dicentra sp.

C. cafiadensis Goldie, in Edinb. Phil. Jour. 6:329. 1822 = Dicentra cana-
densis (Goldie) Walp.

C. Cucidlaria Pers. Syn. Pi. 2:269. 1807 = Dicentra Cucullaria (L.)
Bernh.

C. eximia Link, Enum. Hort. BeroL 2:218. 1822 = Dicentra eximia (Ker)
Torr.

C. formosa Pursh, Fl. Am. Sept. 2:462. 1816 = Dicentra Formosa (Andr.)

DC.

C. fungosa Vent. Choix de Pi. /. jp. 1803 = Adlumia fungosa (Ait.)

Greene.

Mosc

doe

C paeoniae folia Pers. Syn. Pi. 2:269, 1807. Listed as a questionable synonym
of C. Scouleri Hook., Torrey and Gray, Fl. N. Am. 1:69. 183 8. This Asiatic
species actually is not found in America.

C. tenuifolia Pursh, Fl. Am. Sept. 2:462. 1816 = Dicentra sp.

Index to Exsiccatae

The collector's numbers are printed in italics^ or if the collection Is unnumbered, it is
indicated by a dash following the collector's name. The numbers m parentheses are those
assigned to the species and subspecies In this revision.

Abrams, L. R. g222 (1); 7264 (9). Alexander, A. M. 549a, 54gby 54gc (9).

Adams, J. W., & E. T. Wherry. 4698, 4748 Allard, H. A. 832, 7895 (4) ; 232, 2562,
(5). 6612, 7620, 7620a, 7630 (5).

(

[Vol. 34

240

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Allen, O. D. Il8, 311a (1).

Ames, Mrs. M. E. P.

(2).

Benson, L. I42g (1); 2202 (2).
Benson, S. B. 72 (9).

Anderson, D., Rhinchartj & Nelson. 84^ Bcrcman, S. D. 704 (9),

(9a).

Bergman, H. F.

(9).

Anderson, E., & D. M. Anderson. 26ojp Bcrlandier, J. L. I^QQ (7a); 548, I/'gg,

(4).

W

Anderson, J. R.

(1);

(4); 804 (9). Bethel, E.

1933 (7b); 216, 1476 (8).

Berry, R. F. 82 (9a).
Bertaud, Bro. '55 (9).

(2a).

W

Anect, Bro. 51, 16 J (9).

Bethel, E., F. S. Willey, & I. W. Clokcy.

4128 (2a).

Applcgate, E. I. 8517, S610 (9); 8450 Bidwell, Mrs. J.

(2).

(9a).
Armstrong, M. ^1$ (1).
Arsene, Bro. G., & Bro. A. Benedict. 15144

(9).

W

Biltmore Herb., I2gi (4); 2082a (5);
5453a ( 7a ) ; 2079a ( 9 ) .

Arthur, J. C.

(7).

Bissell, C. H.

(4).

Artz, L. 523 (4).

Ashe. W. W. — (4).

Austin, Mrs. R. M. — , 557, 1 393 (2);

(9).
Averill, H. —

Blaisdell, F. E. 67 (3).
Blake, S. F. 9320 (5).
Blanchard, F. —

(4).

W

(4).

■, 661 (9).
— (8a).

Bacigalupi, R. 162S (2); 2146 (4); 978
(9).

Bailey, L. H. Jr.
Bailey, W. W.

(9)

Blankinship, Laura A.

Blomquist, H. L. 37H (4); 37^^ 7^95

(5); 10228 (7a).
Blomquist, H. L., & D. Corrcll. 4710 (4).
Bogusch, E. R. — , S9^y 601 (7a) ; 600 (8).

(4).

Bolton, A. L.

(9).

Baker, C. F. 258, 339, 716 (2a); — (4); B^kcr, J.

(9a).
Bostock, H. 60 (4).

349. 517 (9) ; 183, 338, 340 (9a).

Baker, C. F., F. S. Earle, & S. M. Tracy. Bourgcau, E.

304, 910 (9).

Baker, M. S. 4S32G, 9374 (9).

Bannister, H. M.

(3).

Barber, H. S. US (9a).
Barber, M. A. 167, 257 (9).
Barkley, E. D. g6 (7a).

Boyd, A. A.

Brackett, E.
Bradv. A. W

— (4).
W. 247 (4).

Brainerd, E.
Brandcgce, T, S,

(4).
(7a).

(4).
(9).

1097, 4263, 13233

Barkley, F. A.

(7a).

Barkley, Mrs. M. W.

— (4).

(7).

(2a); 620 (9); 36,284 (9a).
Brass, L. J. 14175 (9a}.

Barlow, B. —
Barndell, E. M.

(1).

Braun, E. L.
Brav. W. L.

Barnhart, J. H. J/, 443, 793, IO18 (4);

474 (9).
Bartlett, Mrs. F. 55 (9).

J

(5).

(7a); 75 (8).

,577 (9).

Briggs, F. P. 1458 (4).
Brigham, Mrs. R. H. 13730 (9).

Bartlett, H. H.

251 (4).

Bartlett, Mrs. W. H. 292 (9).

Bartram, E. B.
Bates, J. M. —
Beal, W. J. —

Bebb, M. S. -

(9).

(9a).

(4).
(9).

Bright, J. 14931
Brinkman, A. H.

5253 (9).
Britton, N. L.

(4); 1 428 1, 14283 (5).
4436 (4); 2015, 4114a,

(5).

Britton, N. L., E. G. Britton, & A. M. Vail.

(4).

Bebb, R. 2736, 3763, 3916, 5093 (6) ; 1446 Broadhead, G. C.

(6);

(7).

(7a); 2362 (8).

Brooks, H. E.

(4).

Bell, J. M.

(4).

Bell, W. B. T460 (9).

Benke, H. C. 5492 (6); 5141 (7); 3338,
5824 (7a)

Benner, W. M. —
Bennett, F. L. 364 (9).

Brown, A. H. 37 (6).
Brown, S. 62, 214 (9).

Bruce, Mrs. C. C.
Bruhin. T. A.

— , 1192 (2); 2206 (9).
(7).

(4); 5013 (9).

Bryan, W. C
Bryant, Mrs.

74 (9).

(3).

1947]

OWNBEY MONOGRAPH OF CORYDALIS

241

Buflfum, B. C. 5/ (9).

Bull, R.

(9a).

Burgess, T. L W.
Burnham, S. H.

(4).
(4).

Bush, B. F. — , I, 22, 63, 1386, 4375, 7445,

Clements, F. E., & E. S. Clements. IIO,

257 (9).
Cleveland, D. —

Clokey, I. W. 27^

2771 (9a).

(2).

7914, 13237, 1 3237 A, 132 j9, 13259A, Clover, E. U. 6354 (9a).

13280A, 14400, 14403, 14408, 14460, Cockerell, T. D. A.

(9).

14480, 1452s, 14543 (5); 523, 1295, Cockerell, T. D. A., & M. D. Cockerell.

1610, 5615, 5615A (6); 8, 22, 23, 71,

32 (9a).

617, 887, 1081, 1328, 1649, 4369, 4925, Cocks, R. S.

(7a).

5519, 7109, 7440, 7568, 7568 A, 10409, Coghill, G. E. 95 (9).
13280, 13459, 14461, 14481, 14498 (7); Cole, L. A. 103 (4).

4, 19, 317, 519, 1 01 4, 1120, 1345, 1377, Collier, A. J.

(3); 72 (4); 73 (9).

14578, 14606, 14629 (7a); 1209 (8); Collins, J. P., & M. L. Fernald. 85 (9).

575 (8a).
Butler, G. D. 12, 41, 10975 (6);

Collom, Mrs. R. E. 140, 143 (9).

(7).

Commons, A.

(5).

Butters, F. K., .& M. F. Buell. 349 (4); Conard, H. S. 150 (1).

4^9 (9).
Butters, F. K., & C. O. Rosendahl. J342

(9).

Cain, S. A. 56 (9).

Cameron, C. 53 (9a)-

Camp, W. H. 1 28 5 (4); 1 333 (5).

ice, L., & F. W. Pe
W. B. 1 4041 (5).

Coombes, Mrs. A. L.

— (1).

(2).

Cooper, Dr. —
Cooper, W. S. .
Copeland, E. B.

(2).

Canby, W. M.

(4);

(5).
(2).

Cantelow, Mrs. H. C. —

Carberry, C. 43 (^)'

Carleton, M. A. 35 (6); 4g (7a).

Carr, L, G. 911 (4). -

Carr, W. P. 125 (9).

Copeland, H. F. 20 ^^ 758 (2).

Core, E. L. — . 2886 (4).

Correll, D. 107 (4).

Cory, V. L. 291, 38687 (8); 5447, 13662

(9a).

W

4773> 59^ (9).

Carrasco, L.

(8).

Carter, M. R. 23 (9).

Coues, E., & E. Palmer. 294 (9a).
Coulter, T. 664 (9a).

Carter, W. R.

(1).

Coville, F. V.

(4);

(9).

Case, E. L., & J. G. Lemmon.

(2).

Coville. F. v., & T. H. Kearney. 1810,
1966, 2033 ( 3 ) .

Castetter, E. F. II42, 1403 (9); 1416, Cowen, J. H. 39, 629 (2a); 30 (9); 38,

1436, 1482 (9a).

. 476'i (9a).

Chamberlain, E. B., & C. H. Knowlton. Cowles, H. C. 629 (1).

(4).
Chamberlin, Myrtle. 12 (7).

Chandler, A. 2127 (7).

Chancy, R. W. 151 (4).

Crandall, C. S. 12 44 (4); 28, 238, 470

(9); 236, 238, 473 (9a).
Crandall, C. S., & J. H. Cowen. 35 (9).

Chapman, A. W.

(4);

,65 (7a).

Cratty, R. I.

(7).

Chapman, J. W. 28 (3).

Chapman, Mrs. J. W. / •(3).

Charette, L. A. 254 (4).

Chase, A. 2 1 73 (5).

Chase, V. H. 5159 (5); 3822 (7).

Child, M. 562 (9a).

Churchill, J. R. —

Cronquist, A. 2854 (2c) ; 1414, 1 41 5,
2724, 2834, 3158, 3323 (9); 895 (9a).

Curtis, L. B. — (7).

Curtiss, A. H. — , // (5); — , N. Am. Pi.
125 & 125a, 4208, 4515, 4516 (7a).

— (4);
(7a); — (9).

Clark, H. S. — (5),

Clark, O. M. 8412 (9).

Clark & Devitt. 10 (6).

(5);

2431a (2c).

, 190, 354, 1728, 2431,

. 41S Cutler, H. C. 771, 4638 (9a).

Dale, E. E. Jr. 43 (9).
Damon, W. E. 67 (4).

Daniels, F. 82 (9a).

Clausen, R. T., & E. R. Clausen. 5702 (5). Darlington, W.

(5).

Clausen, R. T., & H. Trapido. 2813 (9).
Clemens, Mrs. J. 11589 (7).

Davis, J. — , 73, 1435, 3833, 4415, 4415a,
6363, 6387 (5); 121 1 (7).

Clemens, J., & Mrs. J. Clemens. 797 (8). Davis, R. J. 2746 (2c); 275-37 (2d).

[Vol. 34

242

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Dcam, C C. 31641 (4); — I2J/4 (5), EpIIng, C. 63^1 (1); 5<?^-? (7).

Degencr, O. 4S41 {9a),

Degener, C, & L. Peiler. l6/'go (9).

Epiing, C, & J. M. Houck. gjjS (2d).

Fpling, C, & W. Roblson,

(2).

Dcmaree, D. 4732, IO394, 12059 (5);
3039, I2III, 15095, 20884, 20885,
20920, 22023 (6); 14516, 2o8jO (7a);
11898, 12351, 12547 (8a); 7470, 12193

(9a).
Dewey, L. H. / (4); 225 (5).
Dixon, J. If, 25 (i); 45 (4).
Dobie, Mrs. B. McK. 13 (9a).
Dodge, C. K. 43 (5); / (9).

Eskew, C. T. T557 (7); 1585 (8a); 203,

243 (9).
Evans, W. J. 43 (6).

Eyles, D. E. 7725 (5).

Fassett, N. C. 3853 (4) ; 20937 (5) ; 201 1 3

(7).
Fassett, N. C, & W. C. Meyer. 5717 (7).

Faxon, C. E.

(7a).

Dougan, L. M.

(7).

Downey, J. 181S (9).

■ (7).

(6); jf^-/ (7).

Doyel, Mrs. -
Drouet, F. —

Drummond, T. 16 (7a).
Drushel, J. A. 1387 (4);

(5); 1S19 (9).
Duran, V. 3069 (9).
Dutton, D. L. — (4); —
D wight, N. E. 9 (9a).

Ealy, R. 20 (6).

Fames, A. J., & L. H. MacDaniels. 2366,

2367 (9).
Earle, F. S. 503, 503a, 643 (9) ;

Faxon, E., & C. E. Faxon.
Fendler, A. 1 7 (9).
Fendlcr, F. S. 1 445 (4).

(9).

Fernald, M. L.

(4).

(9).

- (9a).
Earle, F. S., & S. M. Tracy. 1 35, 392 (8).

Earle, R. E.

(4).

Eastwood, A. 1758 (2);

3S0 (3); 133 (9); 6093, 8204 (9a).

Eaton, D. C.
Eaton, D. W.
Eby, A. F. -

(7a).
(3).

— (4);
Eby, J. H. — (4);

Edie, M. C. 2 (9).

Edmonds, H. W. —

(5);

(9a).

(4).

Edmondson, T. W. II06, 4995 (4).

Fcrnald, M. L., & J. F. Collins. 1050 (9).
, 4059, 4094 Fcrnald, M. L., & H. B. Jackson. I2098

(4).

Fernald, M. L., & J. B. Lewis. 1 4536 (5).
Fernald, M. L., & B. Long. 23865, 23866

(4); 7041, 1 1838, 12083 (5).
Fernald, M. L., B. Long, & E. C. Abbe.

Fernald, M. L., B. Long, & A. S. Pease.

I16S5, 1 1686, 11687 (5).
Fcrnald, M. L., & A. S. Pease. 25088 (4).
Fernald, M. L., & H. St. John. 7502 (5).
Fernald, M. L., & W. C. Strong. 429 (4).
Fernald, M. L., & K. M. Wicgand. 5455,

545^ (4).
Fcrnald. M. L., K. M. Wiegand, & A. J.
Eames. T4295 (4).

Ferris, R. S., & R. Duthie. 1002 (2c);
106S (9).

Fiero, K. 2J (9).
Fiker, C. B: 1674 (9).

(2a); 307,

(9).

Edwards, O. T.
Eggert, H. —

(5);

(1).

(7);

(7a);

(7b); 106 (8a).

Fink, B. —
Fisher, G. L.
Fisher, H. L.

(7).

(7a).
(4).

Eggleston, W. W. 5656 (2a); 3015 (4);

10439, ^^955, ^S674, 20251 (9); 6402,

Fitzpatrick, T. J., & M. F. L. Fitzpatrick.

/ (7).

(.5).

6427, 6447, 14322, 19825, 19982 (9a). Flint, M. B.
Eggleston, W. W., & M. L. Fernald. — (4). Flodman, J. H. 486 (9).
Ehlers, J. H. 252, 6208 (4) ; — , 7558 (9). Fogg, J. M. Jr. 4984, T3459, 14552, 15115,

(4); 1579 (5).
Ford wood, W. H, 139. 365 (9).

Foster, R. C, & J. F. Arnold. 199 (9a).

Ellis, C. C. 6 (9).
Elmer, A. D. E. — , TO18 (9).
Elmore, F. H. 8 (9a).
Elrod, M. J. 51 (9).

Fowler, J.

(4);

(9).

Emerton, J. H.

(4);

(9).

Freeborn, R. 338 (7b).

Emig, W. H. 520 (7).

Freeman, O. M.

Engelmann, G.

^32 (6);

Engelmann, H.

(1);
(7);

(9a).

(4);

(9).

(5);

(4);

(7a).

French, G. H. — (5).

Friesner, R. C. 2726, 8528, 16649 (5).

Engleman, J. 1569, 1 571 (7); 213, 1570,

Frye, W. M.
Fuller. T. O.

(4).
(7a).

1572 (8a); 1566, T567, 1568, 1695 (9a). Fulton, H. J. 7165 (9a).

1947]

OWNBEY MONOGRAPH OF CORYDALIS

243

Galway, D. 8415 (9a).

Hall, E.

(7); II (7a).

Garrett, A. O. 54g8, 8453 (9) ; 2555 (9a). Hall, E., & J. P. Harbour. 31 (9, 9a).

Garvin.

(9a).

Gary, L. B. 575 (5).

Gates, F. C. 12453 (4); IO205 (5).

Gates, F. C, & M. T. Gates. IO36S (4).

Hall, H. M. g^oS, gyg6 (2).

Hall, H. M., & E, B. Babcock, 42gi (2).

Hamilton, L, P.
Hanna, G. D.

(9a).

Gayle, E. E.

(7).

Gentry, H. S. 586g (9a).

Geyer, A.

(5).

Gilbert, F. A. 43 (5).
Gilbert, F, A., L. O. Will

son. 404 (5).
Gilbert, G. — (4).
Gilbert, J., & T. Gilbert.

Gilman, M. F. 47 (9a).

— (3).
Hanson, H. C. — , 364 (8); A108, C363

(9a).

Hanson, H. C, & E. E. Hanson. Aliyp,

A1180 (9a).

Hardin, E. —

(9).

Harper, D. C.

(2a);

(9).

(4).

Harper, F. 53 (4).

Harper, R. A.

(4).

Girard.

Harper, R. M. 32()6 (5); 10 (6).

(8).

Glatfelter, N. M.

(4);

,61 (5).
(5).

(7a).

Harrison, J. J.

Glcason, H. A. p72J (4); —
Glismann, D. 42 (9a).
Goddard, P. E. 48g (9).

Godfrey, R. K., & R. N. White. 705
Goldman, E. A. 433 (9a).
Goodale, A. S. 6gg62 (7a).
Goodding, C. O. 642g (9).
Goodding, L. N. dp, 6ga, 1027^ Il8g

^3^S> 1501. 2071 (9); pj/ (9a).
Goodman, G. J. 2533 (6).
Goodman, G. J., & C. L. Hitchcock.

1484 (9).

Goodman, G. J., & L. B. Payson. 2S40 Havard, V.

Harper, R, M., & H. Kurz. —

Harrington, G. L. 57 (3),

Harrington, H. D. 1621 (2a); 3 (9).

Harris, M. 3605 (9).

Harrison, B. F, //J, 6/5<?, 747^^ gi63,

IO165 (9); djpp, g630, 1013s (9a).
Harrison, G. J. 7Sg4 (9).

(7b).

Hartman, C V. 338, 600 (9a),
Harvey, F. L. — , /, 4, j6, 123'' (6);

(7).
Harvey, F. L., & L. H. Harvey. —

(4).

72(95, Hasse, H. E. —

Hastings, G. T.

(7a).

(4).

(7a);

(8);

(9a).

W

1023

(9).

(7a).

(9a).

Gorman, A
Gould, L.
Grace, M. C. 132 (8a).

Graham, E. H. 3771, 8635, g287 (9) ; 7S82,
8014 (9a).

Grant, A. L. 1643 (5).
Grant, G. B, 132, 2726 (9).

Hayden, A. 5045, 5047,5036, 1 1246, II247

(7).
Hayden, F. V.

— (9a).

(5);

(7); 56 (9);

Hawk, H. A.

(6).

Grant, J. M.

7936 (1).

Grant, M. L. 2845 (4); 2846 (9).
Grass!, C. O. 3766 (4).

W

Gray, A.

(9);

1757 (7).

— (9a).

Hayward, H. E. 638, 645, 1335 (9).
Heald & Wolf. 7// (8).
Heath, H. — (3).
Hedrick, D. C. 252 (8).

Heideman, C. W. H. 66 (9).

Heidenrcich, V. T. ig3{9).

Heller, A. A. 1 2047, 14685, 16284 (2);

,9 (4);

(5); 1433 (7a).

Gray, F. W. 5134 (4).

Greenman, J. M. 235, 236, 238, 372 (4);
3958, 4212 (5); 2184, 3392 (9).

Heller, A. A., & E. G. Halbncb. 1209 (4).
Heller, A. A., & E. G. Heller. 3871 (1);
3500. 3529 (9); 3515 (9a).

Greenman, J. M., O. E. Lansing Jr., & R. Henderson, J. B. 33, 5689 (6).

A. Dixon. 132 (4).

Griffith, H. R. 35 (5).

Griffiths, D. — - (9); 2280, 2379, 2648,

3553, 384^, 4176 (9a).

Griscom, L.
Grout, A. J.

(4).
(4).

Henderson, L. F. 49, 546 of 1924 (1);
3265 (2c); 3716 (9).

Hermann, F. J. 7290, 7529, 7748, 8729

(4); 10087 (7a); 5247, 7749 (9).
Herrick, C. L. 2$9, 524, IO06 (9).

Herrlot, W.

(9).

Haberer, J. V. 55 (4).

Hale, J.

(7a).

Hall, D. 8415, 8506 (9a).

Hershey, A. L. 3971, 3972, 3974, 3978,

3980 (9) ; 3973, 3973, 3976, 3977, 3979

(9a).
Hexamcr, F. M. —

(4).

[Vol. 34

244

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Hicks, G. H,

(4).

High, M. M. gi (7b).

W

Hill, G. A. 65 (5).

2 1 00 J (7a).

Hills, F. G.

(4).

Hinckley, L. C. 28g^ (8); 21 y (9).
Hitchcock, A. E. IIIQ, 1287 (9).

Johnston, E. L., & G. G. Hedi^ecock. 820
(9a).

Jones, G. N. — , 8Yg4 (l); 506S, 6g6i,
6g62, 6g6j (2c); J40S (9).

Jones, I. 6/' (4).
Jones, M. E. —

, Iigr (2b);

Hitchcock, A. S.

(6); 32, 1041 (7); IO38
12042, 12064 (9).

(4); 70^0 (5); IO]Q

(2); -

2g365 (7b) ; 85, 50S5, 5486a, 5684] (9) ;

, Si^S^^ 5^96, 3^9^c, 324ga, 5315a,

(8a);

535^, 6002b, 25760 (9a).

Joor, J. r.

Hitchcock, C. H.

(7a).

(9).

Hitchcock, C. L. 2311 (9).

Kaye, D. 1501 (4).

Hitchcock, C. L., & J. S. Martin. 7412, Kearney, T. H^ 123,^01^4)^.

74^5 (9).

Hitchcock, C. L, & C. V. Muhlick. (?//?,

g747, 10216 (2c); g37o, 11568, 12138

(9).

Hogin, L.
Holman, P,

(8).
(4).

Kearney, T. H., & R. H. Peebles. g237

II251 (9).
Keck, D. D. 641 (9).
Kecfe, A. M. — (4).
Kccscckcr, B. P. i2 (6).

Keever, C. 412 (4).

Holmgren, A. H. 875, II13, ig73 (9); ^'"''f'^'' J- ]\

3279 (9a).

(5);

(6);

Holmgren, A. H., & S. Hansen. 350g (9). Kennedy, G. G.

Holzingcr, J. M.

(7);

Hooker, J. D., & A. Gray.
Hope, C. g340 (9a).

,33 (9).

- (2b).

- (4);

— (7).

— (4).
Kennedy, P. B. 44S8 (9).

76, 281, gi6

Kennicott, R.
Ketchum, F. E

(4).

(9).

Killian, O. L. — , 6g54 (7a).

H„pki„, M ,66, (.,; ,4,6. 30S, (,,.,; ^^^V:- 3^07^ -^4^9 (^a).

1374 (9a).

Kincaid, T.

(1);

(3).

Hopkins M., & G. L. Cross. 1549 (5); Rirkwo^d, J. E. 77/2 (2d); 2^5-/ (9).

1774 (6).
Hopkins, M., E. R. MacDowcll, & M. P.

Copeland. 63g4 (7).
Hopkins, M., A. Nelson, & R. A. Nelson.

206, 233 (8a); 267 (9a).

W

(3).

Kirkwood, J. E., & J. W. Severy. 7/72

(2d).
Knight, W. C. 50 (7).
Knowlton, C. PI. — (4).

— (5); 6 (9).

Knowlton, F. H.
Kriebcl, R. M. 3105 (5).

Hosic, R. C., S. T. Losce, & M. W. Bannan. Krotkov, P. V. 74'jg, go6^ (9).

7^77, 7^72 (4); 7.^7J (9). '^'"^ '

House, H. D. 3651, 5153, 16214, 24625, Laing, H. M. 174 (4).

2S184 (4); 3228 (7a).

Howell, J., & T. J. Howell.
Howell, T. J. — , J7, (I);
Hudson, Bonnie L. 72^ (5).

(1).

(9).

Lakcla, Olga. TJOO, 2461, 2485 (4); 2403,

2g26 (9).

Lamb, F. H. 7odj (1).

W

(7).

Hughes, E, L. 44 (9).
Ifft, J. 34 (1).

Inncrs, R. R. d/^ (7rt); 340 (8),

Lane, W. C.

Lanvjford, T. E.
Lani;Iois, A. B.

(4).

(9).
(7a).

Lappin, A. F. 84 (9); (?5 (9a).

Large, J. W., & K. T. Clausen. 1264 (4).

Lea, M. C,

(4).

Inncs, R. R., & B. H. Warnock. 423, 472 Lcding, Mrs. 777p (9).

(8).

Jaussan, K. P.
Jeff & Watts. ■
Jennings, O. E.

Lee, P. S.

(7a).

Leggett, W. H.

(5).

(4).
(9a).

(5).
Jennison, H. M. 7/^ (5).

Jcrmy, G. —

(7a);

, 20s (8).

Lclbcrg. J. B. 5/c?^ (2); ligo (2d); 473^

5H (9); 5503 (9a),

Lcmmon, J. G. — , p, j6 (2).

Leonard, F. E. — (2b).

LcSucur, H. yiex-Sl6, I2oy (9a).

Joegcr, H. F.

(9).

Lcttcrman, G. W.

(5);

(7);

(9).

Johnson, P, 276 (9).

Johnston, E. L. 2/" (9); II2I (9a).

Lcwton, P. L, 45 (7a).

Lighthipe, L. H. /p5, (55i^ (7a).

1947J

OWNBEY MONOGRAPH OF CORYDALIS

245

(4).

Lighthipe, M. L.
Lindhcimcr, F. L.

Lindsay, A. A. 4556 ( 1 ) .

Little, C. S. / (9).

Little, E. L. Jr. 54, 1023 (5); 123, 651,

1022, 1024 (6); 534 (9a).
Little, Mrs. E. L. I02I (5); 1020 (7a).
Little, E. L., & Mrs. E. L. Little. 1 19 (5).
Livingston, G. A. 47 (2a).
Livingston, R. B. 17$, 301 (9),
Lloyd, C. E. 369 (9a).

W

433, 546, 663 (8). Maguire, B. 86la, 3460 (9).

(9).

Maguire, B., & H. L. Blood. J364, 1863
(9a).

Maguire, B., & G. Piranian. 124J8 (2a).

J

W

(9).

1862 (9).

Marcclline, Sister M. 2063 (9).

Lloyd, C. G.
Lloyd, F. E.

(5).

(1);

(4).

Marie, P. Louis.
MarsKall, W. F.

Martin, B.

Mason, H. L.

(4).
475 (9a).

(7a).

(3); 3423, 3473 (9).

Lodewyks, M. C. 87 (7).

Mason, R. F. 61, 85, 87 (6);

(7).

Loesc, E.

(8a).

Long, B. 50047 (4); 4294S, 56265 (5).

Loomis, H. F. g26 (9a).

Lucas, W. D. 242, 245 (9).

Lucas, W. D., J. T. Painter, & F. A. Bark-
ley. J 42 36 (8).

Luckhardt, R. 185 (7).

Masscy, A. B. 849 (4); 1749, 3799 (5).
Mathias, M. E. 143 (4); 386, 593 (9).

Maxon, W. R.
Maxon, W. R.,
Maxon, W. R.,
Mearns, E. A,

75 (5).

1265, 37go, 4g68 ( 9 ) ;

obinson. J (4).

indlcy. 25J (5).
(7); 27c?, 56^5, 1045,

303 (9a).

Lunell, J.

(9).

(4).

(3);

McAtee, W. L.
McBryde, J. B.
McCalla, W. C
McCart, W. L.
McCauley, Lt.

(4).

- (7a); M1007 (8a).
2124 (9).

(2a).

McCIellan, B. 10 (7).
McClellan, J. T. 12 (7).
McCree, J.
McCubbin.

633 (5).
(5).

McDonald, F. E.

(7).

Mearns, E. W, 57, 58 ■
Memmmger, E. R, —
Mcndenhall, W. C. —
Mericle, L. 449 (9a).
Merrill, E. D., & E. N. Wilcox. 646 (9).
Mertie, J. B. 6c> (}).
Mertz, H. N. — (5).

Metcalfe, O. B., 877, 980 (9); 7p, 1050
(9a).

Metz, Sister M. C. 21 51 (8).

Mexia, Ynes. 2040 (3); 2/od, 22^0 (4).

Meyer, F. G. ^dc?, 2/59 ( 1 ) .

Miller, A.

(2c).

McFarland, F. T. 22 (4).

Miller, G. S. Jr. 443 (7a).

W

Mcintosh, A, C. /Jc? (9a).

McKelvey, S. D. 1686 (7a); 2000 (8); Milligan, Mrs. J. M.

<?<5 (4).

2402, 4802 (9); ^JJO, 4438 (9a).

McLean, D. 27 (6).

(6).

Millspaugh, C. F, 446 (4); /^ (5).

McMullen, E.

(6).

McMurry, F. B. 655 (8a).
McVaugh, R. 4950 (4); -^pO(5 (5).

Mohr, C. — (5);
Moldenke, H. N.
1 98 1, 8204 (9).

• 4^ (9).
353, 4^3 (4).

Macbride, J. F. S95 (2c).

Macbride, J. F., & E. B. Payson. 3287, Moodie, M. E. 137,011 (9).

- (7a).

8768 (4); 243s (5);

3457 (9).

MacDougal, D. T. 230 (9); 83, 103, I14 Moore, E. J.

Moore, A. H. 1667 (4); 3092 (5).

(9).

(9a).
Macfadden, F. A. 13314 (2c).

Moore, F. L., & C. E. Moore.

(4);

(7);

(9).

Maclntyre, Mrs.
Mackaness, F. -

(3).

— (4).
Mackenzie, K. K. 677, 2612 (4);

(5);

MacLean, J.
Macoun.

— , 3756

Moore, H. E. Jr. 623 (7a).
Moore, J. A., & J. A. Stcycrmark. 3063,
3664 (9).

17, 23, 33, 30 (7); 38 (9).

(4);

(9).

W..,

(9).

99 (4); 7, 7/2 (5);

700

(9).

^333^

Macoun, J. M.
Macoun, John.

(3);
(4);

(4);
(5);

(9).
(9).

Moritmcr, M. F. 83 (5).
Morton, M, D. 200 (6).

Moseley, E. L.

(5).

[Vol. 34

246

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Moss, E. H. 2255 (4); 214.0 (9).

Osterhout, G. E.

v 7^2, 1748^ 22S0,

Moyer, Dr.

(4).

S98r, 2448, 3702, 6984 (9); 1950, 2614
(9a).

Mueller, C. H., & M. T. Mueller. 773 (10).

Muenscher, W. C. 42 (4); 13312 (5); Otis, I. C. 1420, J312 (I) .

3138 (7).

Over, W. H. 5139 (7); 1753 (9a).

Muenscher, W. C, & B. Maguirc. 2280 (4) ; Overholts, L. O. — , 10096 (9).

2326 (9).

Ownbey, M. 7 10 (9).

Muenscher, W. C, W. E. Manning, & B. Ownbey, M., & G. B. Ownbey. 2935, 2938

Maguire. 394 (4).
Muenscher, W. C, & M. W. Muenscher.

14279 (7a).
Muir, J. 168 (3).
Mulford, A. I. — (2c). •
Murdock, John Jr. 3022, 3092, 4640 (9). Oyster, J. H.

(2); 3034, 3034^, 3053 (2a); 3057

(2b); 3062, 3064, 3066, 3067, 307T
(2c); 3077, 3089, 3089a (2d); 7077,
2815, 3016, 3018, 3032, 3035, 3041,
3048, 3050, 3055 (9); 3007 (9a).

(7, 7a).

Murie, O. J.

(4).

Murphey, E. V. A. 65 1/2 (9).
Myers, I. J. 75 (9a).

Myers, J.

(5).

Nash, G. V. 289 (4).

Nease, F. 23 (8a).

Nelson, & D. Anderson. 978 (9).

Nelson, J. Smith, & Merkle. 709 (8a); 70^

(9a).
Nelson, A. 10886 (5); I0858 (6); I18,

124, 1241, 1679, 1921, 2395, 3164, 3926,

4027, 4676, 7181, IO50I, IT 495 (9);

369 (9a).
Nelson, A., & J. F. Macbrlde. 1833, 21 54

(9).
Nelson, A., and E. Nelson. 3684, 6221 (9).

Nelson, A., & R. A. Nelson. W-2206 (3);

■, W-2151, 2395, 3622, 3833 (4);

5006, 5019 (8) ; 2103X, 4010 (9) ; 1 395,

1 491, 2005, 2005a, 2786 (9a).

Nelson, E. STO (9).
Nelson, E. W. 4841 (9).
Nelson, J. C. 2569 (1).
Nelson, M. L. 43 (8a).
Nelson, N. L. T. 27 (2a).
Nelson', P. M. 45 (6).
Newcomhc, W. A. 19 (9).

Nichol, A. A.
Noll, H. R. -

(9).

(4).

Normand, J. F.

(6).

Norton, J. B. 174 (4); 43T (7a); T4, T4a

(7).
Nuttall, L. W. —

(5).

Ohlweiler, W. W.

(5).

Olguin, M. — (8).

Onion, I. S., R. Kennicott, & W. L. Hard-

isty.

(9).

Oosting, H. J. 1787 (4).

Orcutt, C. R. 5829 (7b).

Osborn, B. 1487K (5); 461K (8a).

Osborn, W. J., & H. J. Fulton. 7152 (9a).

Pace, L. 61 (7a); 295 (9).

Painter, J. H. 2 (5).

Painter, J. T., & F. A. Barkley. 14461

(7b).

W

ley. 14235 (7a).
Palmer, Edward.

(8a); 77 (9); ^, ^2,

jpp, 9902 (9a); J56 (10).
Palmer, E. J. 289, 1616, 1676, 1676A,
5428, 9414, 14758, 24778, 29748, 33202,
33209, 39156, 39996 (5); — , 37. 288,

888, 1738, 1738A, 1792, 3412, 4921,

J3428, 24726 (6); 1625, 1645, 1674,
1678, 1702, 1710, 5514, 24594, 24769,
24847 (7); 4827, 7106, 13126, I3146,
1343^, 39166 (7a); 9064, I 1239 (7b);
1 1 396, 34424 (8) ; 3118S, 31346, 34296,

34335, 34543, 37394, 37 39^, 37 44^ (9) ;
33966 (9a).

Palmer, L. J. 194 (3); 1769 (4); 7750

(9).

W. 178

W.. & W

(2).

Pammel, L. H. — (7); 307 (9a).
Pammel, L. H., & E. M. Stanton. 46 (9).

Parish, W. F. 5 (9a).

Parks, H. B. 2202 (7a); 1429, 29500

(7b).
Parry, C. C.
Patterson, H. N. — (7); 7/7 (9).
Paxton, B. 77/ (6).

Payson, E. B. 1664 (5) ; ^7 (9);2^p (9a).
Payson, E. B., 8t L. B. Payson. 2799, 4722

(9); 383S (9a).
Pease, A. S. 606, IO773 (4); 7jp6, 79076

(5); 7^070, 18029 (9).
Pease, A. S., & R. C. Bean. 23549 (4);

2J5J7, 26^^<?, 26393 (9).

Pease, A. S., & E. C. Ogden. 25777, 257^7,
23190 (9).

Peck, M. E. 4T33, 4136, 9616, 1533^,
20946, 21442 (9).

19471

OWNBEY MONOGRAPH OF CORYDALIS

247

Peebles, R. H., & E. G. Smith. 11^ jj (9). Reynolds, H. C. I4ig, 2g68, 2g8i, so^g.

Penard, E. jp (9).
Penfound. W

T.— (5);-
W. 5j6l (9a).

(7);

(7a).

3327, 3846 (7).

W

Perrin, L.
Perry, R. C.

Phelps, O. P, 4go (4).
Pickett, R. 64 (9).

, 77 (6).
- (6).

Rice, E. 37 (8a).

(4).

Ricker, P. L.

Riedcl, M.
Riehl, N.

W

(4).

(7a).
(7).

Piper, C. V. — , ig, 4gj8 (1); 4037 (2d); Rinehart, F. 210 (7a); 2J^ (8a).

. 3233 (9).

Piper, R. H. 112, 1323 (4).
Plank, E. N.

W. W

(6);

Robinson, B. L. — , 80 (4); 25/ (7a).

Pollard, C. L. 48, 1 36 (5).

(7a).

W

Pollock, W. M.
Poison, M. ig (6).
Poole, S. F. 121 (7).

(4).

Rodgers, L. 343 (4).
Roland, A. E. 41416 (4).
Rolland, Fr. 7130, 13761 (9).
Rollins, R. C. 2003 (2a); 7p(5j (9).

Porsild, A. E., & R. T. Porsild. 133, 73g Rose, J. N., & J. H. Painter. 8128 (4).

(4).

Porter, C. L. 3043 (9); 2838 (9a).
Poto, W. L. 121 (4) ; 7^ (9) .
Preble, A. E., & M. Gary. 13 (4); 5 (9).
Preble, E. A., & G. Mixter. 61 g (3); 6oij

68gb (9).

Preble, E. A., & A. E. Preble, gi (4).
Pringle, C. G. — , ig8 (9a).

Pulling, H. E.

(4).

Rose, J. N., W. R. Fitch, & T. H. Park-
hurst. 17710 (9).
Rose-Innes, R. 41022 (7a).
Rosendahl, C. O., & F. K. Butters. 1363

(1); 3ig3 (7).

Rossback, G. B., & R. P. Rossback. 476

(1).
Rothrock, J. T. 22, 28 (9).

Rousseau, J. 23264, 26g7I (4); 26401,

Hastings.

(4).

Purpus, C. A. 6550 (9); 705S (9^); 3073, ^^59^. 26657, 26871 (9).

4602 (10). Rowlee, W. W., K. M. Wieg

Pusonett, K. 59 (7a).

Quarterman, E. /Op, I0g4 (5).
Quick, C. R. 1072 (2d).

Radford & Stewart, 32 (7a).
Ramaley, F. 1165, 1413 (9).

Runyon, E. 47 (8a, 9a),

Runyon, R, t6i8 (7b); I02I (10).

Rupp.

(5).

Rusby, H. H. 282, 2gi (4); p (9a).
Russell, C.

(7);
Rust, H. J. 621 (9).

(7a).

Ramaley, F., & K. R. Johnson. 14731 (9). ^^th, A. J<57 (4); ig4, 228, 231, 233,

Ramaley, F., & Richards. 13060 (9a).

W. W

Rand, E. L., & B. L. Robinson. 121 (4).

W. 21 (7).

J(5j, 1674, 1 841 (5); 7775 (6); 1033
(7a); J (9a).

Ruttcr, C. — , g2, I7g, 206 (3).
Rydberg, P. A. 6848 (2b); 8igi, g23g.

Randolph, L. F., & F. R. Randolph. 77^J 9^33 (4); 725, 312, 313 (9); lo, 313

(4).

(9a).

Raup, H. M. 2443-a, 6088, 6113, 7624 Rydberg, P. A., & A. O. Garrett. 88/3

(4); 2438 -a, 243g-a, 2440-a, 2443-a,
6033, 6032 (9).

(9); g2o8 (9a).
Rydberg, P. A., & R. Imler. 403 (6).

Raup, H. M., & E. C. Abbe. 4473, 4331, Rydberg, P. A., & R. K. Vreeland. 6262,

4595 (4).
Redfield, J. H.

280,281 (5); 283 (9).
Reed, E. L. 1823 (8).

287, 288 (4); 27g,

6263 (9); 6261 (9a).

Safford, H. T.

Safford, P. —

(5).

(2).

Reed, J. F.

(4).

Reed, J. F., & M. S. Reed.

(7a).

Rcppert, F.

(8a).

Reverchon, J. — 2736, 371 1, 433J (6); , ...... „. ^_^.

I8g6 (7); — . i8g6, 2733, 2g63, 3710 Sandberg, J. H. — 163 (4);

Safford, W. E. 776 (4).

St. John, H. 8662 (1); 6361, 762g (9).

St. John, H., & B. Long. IO30 (5).

St. John, H., & D. White. 22 (4).
St. John, Mrs. O.

(9a).

(7a);

. 3707, 3709 ( 8 ) ;

(8a).

1023 (9).

> 18, 434,

[Vol. 34

248

ANNALS OF 1 HE MISSOURI BOTANICAL GARDEN

Somes. M. P. 23 (9).

Sonne, C. F.

12 (2).

S:indbcrg, J. H., & J. B. Lclbcrg. 432 (9).
Sandbcrg, J. H., D. T. MacDougal, & A. A.

Hcllcr. 139, 765 (9).

Sartwcll, H. P., — (5).

Sarvis, J. T. 46 (9).

620 (3); 213, 1735, 3272 Spraguc, R. 255 (9).

Soth, Mrs. M. E. C-I0() (9).
Spcrry, O. E. T171, T601 (8).

Spiej^clbcrg, C. H. 202, 203 (9).

Scamman, E.

{^); 2402, 3369 (9). Sprcadborogh, W. 70320 (9).

Schallcrt, P. O. S135 (4); 5/i^, iP'^J (5); StanJley, P. C. 6706 (2a); 3342, J2I26,

(7a);

(9a).

Scbedin, L. M., & N. T. Schedin. 6o2,

16360, 18313 (4); 11291 (5); 4147'
5108, 6593, 1 5321, 17006 {9) ; —

40615

603 (9).

W. 2c? (9).

(9a).

Standley, P. C, & H. O. Bollman. IO964

Schmoll, H. M. 1050, 1268 (9a).

(9).

Schneck, J.
Schradcr, F. C.

(5).

W

(8).

(4).

Schrcibcr, Beryl O. 2527 (2).

Scbrenk, H. von.

(9).

149

Schrcnk, J.

Schuette, J. H.

(4).

(4); 10 (9).
Schulz, E. D. 21 (8).
Scribncr, F. L. 8c (9).
Sellon, G. I. 85 (9).
Scnn, H. A. 1334 (9).

.Serchell. W. A.

Stearns, F. 348 (9).
Steele, E. S. 98 (5).
Steele, E. S., & Mrs. E. S. Steele.

(4).
Stephenson, B. C.

Stephenson, M. R. 171 (7a).

Stevens, G. W. 15 (7, 7a); 73.I (7); 94.I

(7a); 192 (8a); 409, S07V2 (9a).

Stevenson, E^ 16 (9).

(5).

(4).

Stewart, L. M.

(4).

w

(9).

Seymour, F. C. 378, S37S (4).
Shafer, J. A. 38,632 (5). ,
Shaw, C. H. 830 (4).

Stcycrmark, J. A. 237, 331, 420, 43T, 333,

4380, 4619, 4693, 4930, 8029,

574,

10078, loogo, 10189, 10235, 10439,
18302,18600, 18613, 18633, 18805, 18908,

Shear, C. L. — (4); 47 (7); 4226, 4430, 19051, 21265 (5); 18636, 18757, 18760

(6); 554, 769, 451^, 4528, 4530, 454J,

8038,

4718, 4834, 5090 (9).

Sheldon, C. S. 340 (9).
Sheldon, E. P. — (4);

W. 115 (9).

(9).

Shimck, B.

(7);

(8a).

455^, 4564, 4575, 4740, 5729,
18519, 1863T, 18646, 1S668, 18712,

18719, 187 51, 18766, 18797,
18821, 18826, 18833, r886i.

18817,

Shinncrs, L. H. 364S (5).

W

Shocklcy, ^X^ H.

(9).

Sholly, G.

w

(9).

18873,

19209, 19215 (7); 813, 10210, 10249,
10260 (7a).
Stillingcr, C. R. 29 (2d).

SLilllnger, R. C. 5/ (9).

(5).

Stokes, S. G.

(93).

Shreve, F. 5227
8012 (9a).

Shreve, F., & T. H. B.

(9); 5426, 6285, 7332, Stone, Mrs. F. M. 266, 532 (9); 20, 78,

(4).

333, 444 (9a).

Storey, G. M.

Shriver, H.

(4);

(5).

W

Skehan, J.
Small, J. K.

(7a); 87 (9a).
I29rh (4);

(5).

9 (4).

— (3).
Stratton, R. 676a (7); 676b (7a).

Studcr, A. 4-26 (4).

Studhalter, R. A. ///J (9a).

Studhalter, R. A., & J. Marr. S-JO15 (9).

Small, J. K., & A. A. Heller.

Smart, Dr. 1S2 (9a).

Smith, A. D. 16 (9).

Smith, C., & F. Rindhart. IT3 (9a).

Smith, C. C. 385, 505 (7); 505 (8a).

Smith, C. P. '3621 (2c) ■,1361, 2332 (9).

(9a).
(4).

Smith, E, C.
Smith, G. —

(1).

(7a).

Smith, J. D.

(5);

(7a).

Smith, P. — (2a); // (8a).
Snyder, Mrs. M. — (7).

Sturgis, W. C.

Sturtevant, F. L.

Sudworth, G. B. 88 (4).

Suksdorf, W. N. 1948, 6666 (1); J (9).

Svenson, H. K. 4455 (5).

Sylvester, C. H. 30 (5).

Tarlcton, J. B. 178a, 178b (4); 49a, 49!?

(9).
Tatnall, E. — (5).
Taubcnhaus, J. J. 2787 (7a).

1947]

OWNBEY MONOGRAPH OF CORYDALIS

249

Taylor, B. 2^01 (9).

Taylor, B. C.

UmbacK, L. M

(4).

, 706 (4);

(5).

Taylor, T. M. C, R. C. HosJe, R. E. Fitz-
patrick, S. T. Loscc, & A. Leslie. 1310,
1311 (4); J314 (9).

Bannan. JOj, jo6, 508 (4); $02, ^04 ^j^'^^Y' ^- R-

Underwood, L. M., & A. D. Sclby. 208
(9).

Van Eseltinc, G. P. 2jT, 282 (5).
Van Valkcnburgh, A. N. p (6).

(9).
Taylor, K. A.

(5).

Texas, University of.

(9); -

(9a).

(7a);

(8);

— (8a);

Thames, L.
Tharp, B. C.

(9a).

(6);

(7a);

(7b);

35000, 37000

(4).
Vestal, A. G. 262 (9a).

Victorin, Marie. 206, 836^, lOOj^, II328,

^5757, J 57(^0 (4); 8358 (9).

Victorin, Marie. & H. Prat. 45945 (9).
Victorin, Marie, & Roland-Germain. 3312
49743 (4).

Thompson, E.
Thompson, E. S.

— , 36000, 37001 (8).

(!)■
(9).

erm

Thompson, J. W. 630, 4110, Q40Q (1);

13406 (2c); 7000, 831 1, I1447, 11905, vlT'c c

inique 260 (4); 48()g6 (9).
Victorin, Marie, Rolland-Germain, J. Rous-
seau, & R. Meilleur. 40082 (4).

(5).

13502, 14162 (9).

Visher, S. S. 4021 (7); 152 (9).

W

Volk, E.

(5).

2og (9).

Thompson, S. L. 30 (4); 97 (9).

(7);

Thornber, J. J.

2822, 40 S3 (9a).
Thornber, J. J., & Brown.

— 5699 (9);

^^* Vreeland, F. K. gi6 (4).

Waghorne, A. 21 (4).

Wagner, R.

(3).

(9a).

Thornber, J. J., & F. Shreve. 7809 (9).

Wahl, H. A. 676 (4); 563 (5).
Waite, M. B. —

W

Waldron, C. B.

(4).

Thouscn, Mrs. O. T. 6 (3).

Thurber, G. 146 (9a).

Tidestrom, I. 3500, 3776 (2a); 333, 1879,

9468, 10935 (9).
Timmerman, M.
Tinsley, J. D.
Tolstcad, W. L,
Topping, D. L.

Walker,

(9).

(5).

(4).

(5).

(4).

Tosh, J. P. 564 (4); 97 (5).
Toulouse, B. — (9a).
Toumey, J. W. 49 (9a).

Townsend, C. H. T., & C. M. Barber. 16
(9a).

Townsend, E. C. —

Walker, E. H. 2280, 2464 (4).
Walker, E. P. 254 (9); 1 46 (9a).

Walpofe, F. A. 1457, 1467 (3).

Ward, C. G. 16 (7a).

Ward, L. F. 64, 520 (5); 5 (9a).

Warnock, B. H. 57, 46022 (7a) ; 63, 423,

472, 20621, 21397, 46021 (8); T135

(8, 9).

Warren, E. R. 1846 (2a); 1794 (9).

Washburn, E. W.

(7).

3

Tracy, S. M. 9212 (7a); 37 (9a).
Tracy, S. M., & F. S. Earle. 392 (8).

Train, P. 2176, 2616, 3053 (9); 2536

Waterfall, U. T. 621 (6); 2615 (7a); 410,

2591 (8a); 3277 (9); 2030 (9a).
Watson, S. 27 (2d) ;

52 (9,92).
Watt, D. A.

(4); 26, 53 (9);

(4).

(9a).

Trelease, W.
Trelcase, W

(4);
& D.

— ^07 (5);

A. Saunders.

3^73, 3874 (3).

- (9).

3872,

(4).

J2 (3).

True, F. W., & D. W. Prentiss Jr.

Tuftc, E. T. 1 01 (9).

Turner. 167 (9).

Turner, G. H. 4 (9).

Tweedy, F. 429 (5); 136 (8); 1 23, 3534,

4949, 4951 (9); 4950, 5536 (9a).

Weathcrby, C. A. 6088 (7a).

Weatherby, C. A., & L. Griscom. 16532
(7a).

Webb, R. J.

Weber, W. A. 2600 (4); 2296 (9).

Wehmeyer, L. E., F. N. Martin Jr., & H.

F. Lovcland. 5003a (9).
Wells, Mrs. E. M. 67 (6).
Werkenthin, F. C.

Wctherill, A.
Wheeler, C. F.

(9a).

(9).
(9).

Tyler, A. A.
Tyler, E. E.

(4).
(7).

Wheeler, H. N. 517, 2622 (9a).
Whetzcl, H. H, —

White, M.

(5).

(7).

250

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

White, R. — (3).

Whitcd, K. 347, 3146 (9).

Whitehead, L. C.
Whitehouse. E.

Williams, L. O., & R. WilHams. 3122,

(9a).

3714 (9).
WilHams, R. S.

8282 (9a).

684,818 (8); 817 (9); Williams, T. A.

(4);

(7);

, 26 (9).
(9).

Wh
Wh
Wic
Wieg

75(^/823, 1053/

2897 (2).

3297,3298 (9).

Wicgand, K. M., & M. C. Wicgand. 65?,

^39 (7a); 634, 6^6 (9); 644, 8682A Woodbury, A. M. 41 (9).

Willits, V. lOI (9).
(9a). Wilson, C. B. 107, 251 (2d).

• Wislizcnus, F. W. 447 (4); 705 (7).

Wolf, C. B. 3002 (2a).
$294, Wolf, J., & Rothrock, J. T. 776 (93).

Wolff, S. E. 371 (7a); 2046 (9a).

Wood, F. E., & F. J. Wood. 123 (4).

E.

(9a).
Wilcox, E. N.
Wilcox, E. V.
Wilcox, T. E.
Wilkcns, H. ^
Wilkinson,
Wilkinson,
Williams, '
Williams, :
Williams,
Williams,

Fcrnald.
Williatns.

(9).

6 (9aj.

Woods, C. N. 358 (9).

Woods, C. N., & I. Tidcstrom. 257$ (2c).

Woodson, R. E. Jr. 280 (5).

Woolson, G. C.
Wooton, E. O.

(1).

(9);

— , 3810 (9a).

(9a).

II (7b).

Wooton, E. O., & P. C. Standley. 3374

(9a).
Worthlcy, I. T. —
Wright, C. n09 (9a).

(9).

E. F. — (4).

E. F., J. F. Collins, & M. L.

— (4).

York, C. L. 4'
York, W. II.
Young, M. S.

(7); 578, 2176, 2202a (9).

487 (4); 1436 (5); 1505 Zcllcr, S. M.

Va2o8 (4);

395 (7a).
(7a).

(1).

(7a)

Zundcl, G. L. 167 (9).

Index to Scientii ic Names

r

Previously published names accepted here are in Roman type; new names, combina-
tions, and the page numbers on which the taxonomic treatment appears arc In bold face;
synonyms are In italics.

Adlumla ^ 191

fungosa 188, 239

Borckhauscn'ia 197

BuWocapfios 197, 207

Cal)fiifes

197, 207

CatyfioJes '— .. ..__J 97

aurciim „.__-,-. 229

BidwcUianum 201

Cascafjum

201

crystaUhintn

jlai'ulum 2 1 5

panel florum _„_..-207

Scolder} . 199

Calwohfcs 188, 197, 209

hrachycarpum 204

BrafiJegei 203

Catfipcsfre _— — — _ 222

.._ . 22 3, 226

-205

Engelmarnjii 2 3

ciichlaffiyJcuni ___.230, 231

glauca -„-„__.. _-_.-., 2 1 1

Ilalc! . _ _ 217, 222, 223

curvhiliqunm
CusfrkJi

230
219

hcf$fat7(77i -_- _.- 206

niacrorrhiza

micranihum .

vionfanuni 234

pachylobiitn -— ,2 3 4

pauciflornm 207

Scouhri ..___.199

scmpcrvircfis 188, 211

Welbcrillii • 230

217 Cht'icapnos 188

Corydalls - : 197

Albcrfac 229

AUniti 199, 201

an n ua 2 1 1

aurea 191, 193, 194, 195. 205, 220,

223, 229
ssp. aurea._„195, 210, 229, 235, 236, 238

australis 222, 223

21 7

cnrvhiliqua - 226

fUvida 2 1 5

var. vjacrantha

var.

var. crystallhia

var.

var.

var. mtcrattiKya „

th

._ 229
_.219

1947]

OWNBEY MONOGRAPH OF CORYDALIS

251

ssp. occidentalis 195, 211, 229. 230,

231,234

van occidentalis . 234, 23 5

var. parviflora 229

var. robiata 229, 231

var. typica 229

biaurifa 239

BhJwcUiae - 201, 203

bilimbata 234

brachycarpa 204

bracteosa 239

Brand egei .— T- 203

bulbosa 1 94

campestrh 222

canadensis _•_ 239

Caseana 191, 198, 199, 200, 201, 223

ssp. brachycarpa 198, 204, 205

ssp. Brandegei „

195, 198, 199, 202, 203, 205

ssp.
ssp.

ssp

cava

chihiiahtiana

Caseana 195, 198, 201, 207

Cusickii 195, 198, 203, 204,

205, 207

hastata 198, 203, 205, 206

194

Jonesii 229

var. stenophylla 234

lutea 194, 239

macrophylla 199

macrorrhiza 229

micrantha 190, 193, 194, 195, 219,

222, 238
ssp. australls_„_:_ 210, 219, 220, 222,

225, 229

var. diffusa 219, 222

var. leptosiliqua 222

ssp. micrantha 210, 219, 223

var. pachysiliquosa „„... 2 1 9

ssp. texensis __196, 210, 225

monilifera 229

var. ferruginifera 219

monfana 195, 229, 234, 23 5

ochotensis ...—2 3 9

orcgana 2 3

234

239

235

234,

crassipedicellata 2 34

crystallina 195, 210, 211, 217, 223

var. strictissima 217, 219

CuciiUaria 23 9

curvisiliqua 194, 223, 224, 226

ssp. curvisiliqua 210, 225, 226, 227

ssp. grandibracteata 210, 227

var. grandibracteata 228

var. tenerior 222

pachyloba

paeoniaefolia

pauciflora 189, 207

var. Chamlssonis -— 207

var. parviflora 207

pseudomicrantha 190, 211, 237

var. Griff it bsii 234

pumila -_- ,„.-. 194

rosea : - 211

Scoulcri 195, 198, 199, 239

sempcrvirens--_.188, 193, 195, 197, 209, 211

tenui folia 239

tortisiliqua — — — 229

230

washingtoniana 229

Wetherillii 229

var. longibracteata

curvisiliquaeformls -—11 A wyomingensis 229

Cmtckii 2 5

hastata 206

var.

191, 192, 239
239

densicoma 230

Engehnanntl 2 3

var. exalt at a ... . 230

euchlamydca 230

eximta „...— _ „. — _____.„_„„239

flavidula -2 1 5

flavula 193, 210, 211, 215

formosa 239 Eucapnoides -. — - 211

fungosa 188, 197, 239 Eucorydalis 189, 191, 196, 198,209

var. lativaginata 230

Corydallis 197

Cryptoceras 197

Cysticapnos 188

Diccntra

canadensis

Cucullaria 2 3 9

eximla — 2 39

formosa 239

Geyeri . -.__.. -2 1 5

glauca 2 1 1

Gooddingii 229

Halei 222, 225

hastata 206

Render sonii ..„.......„._„_._..205, 206

hypecoiformis 229

idahoensis 2 5

isopyroides 2 3

var. Mearnsii - 2 3

Fumaria 188, 191, 192, 197

aiirea 229

flavula 2 1 5

glauca — ^— — — 2 1 1

pauciflora _ 207

sempervirens 21 1

vesicaria 188

Neckeria 188, 197

_.229

'dentalis 234

a urea

var. oca

252 ANNALS OF THE MISSOURI

[Vol. 34, 1947]

curvhiliqna „.. ._.___.._.„.226 Pes-gallinaceus 189, 191, 198, 207

flat'nJa 2 1 5 Pisfolochia 1 97, 207

^laura _„_.. 211 rscuilofnmarla __. 188

mlcraniha ..___ 2 1 9 Pscucfo-Fumnria 197

sempcnircns 21I Ramoso-sibiricae 189, 190, 191, 198

OJoptcra _-_- __^_197 Sopborocapnos __._ [ ll97

ai( rca 229

Explanation of Plate

PLATE 2 8

Generalized floral morphology of CoryJalh drawn from C. Caseana Gray ssp
Branclcgei (Wats.) Ownbey.

Fig. 1. Dorsal view of gynoecium.

Fig. 2. Lateral view of gynoecium.

Fig. 3. Unspurrcd stamen phalange.

Fig. 4. Exterior view of spurred petal.

Fig. 5. Interior view of spurred petal.

Fig. 6. Internal structure of flower showing arrangement of parts.
Fig. 7. Sepal.

Fig. 8. E"xterIor view of clawed Inner petal.

Fig. 9. Interior view of clawed Inner petal.

Fig. 10. Interior view of unspurred outer petal.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 28

10

OWNBEY— MONOGRAPH OF CORYDALIS

[Vol. 34, 1947 i

254 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PLATE 29

Figs. 1
show mor
species; X about 8,

l-ll. Gynoccia of species of Corydalh at flowering time; drawn especially
phology of the stigma. Each drawing is representative of all subspecies of

to

species of the

Fig, 1. C. Caseatia Gray.

Fig. 2. C. Cascatia Gray; side view of stigma.

Fig. 3. C Scoulcri Hook.

Fig. 4. C, pauci flora (Steph.) Pers.

Fig. 5. C. micrantba (Engclm.) Gray.

Fig. 6. C. nirvhillqua Engelm.

Fig. 7. C. anrea Willd.

Fig. 8. C. flai'ula (Raf.) DC.

Fig. 9. C. crystallhia Engelm.

Fig. 10. C. pscudomicrautba Fedde.

Fig. 11, C. sempcrvircns (L.) Pers.

Figs. 12-15. C. Cascana Gray; drawings representative of Sections RAMoso-sinmicAE
and Pes-gallinaceus.

Fig. 12. Raceme in fruit; X about 1.

Fig. 13. Seed; X 4.

Fig. 14. Seed; X 2.

Fig, 15. Fruit, showing manner of dehiscence: X 4.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 29

I

3

5

2

7

9

14

OWNBEY— MONOGRAPH OF CORYDALIS

256

[Vol. 34, 1947]

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PLATE 3

Flowers of the subspecies of Corydalh Caseana Gray and of Corydalh Scoulcri Hook.
An interior view of the unspurrcd outer petal is shown in each case to illustrate difler-
ences in structure; X 1^.

Figs. 1-3, C Caseana Gray ssp. Caseana Ownbcy.

Figs, 4-5.

(W

Figs. 6-7, C. Caseana Gray ssp. brachycarpa (Rydb.) Ownbey

Figs. 8-9. C. Caseana Gray ssp. Cusickii (Wats.) Ownbey.

Figs. 10-1 L C. Caseana Gray ssp. hastata (Rydb.) Ownbey,

Figs. 12-13. C. Scoulcri Hook.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 30

12

OWNBEY— MONOGRAPH OF CORYDALIS

258

[Vol. 34. 19471

ANNALS OF THE MISSOURI

GARDEN

Explanation of Plate

■ PLATE 3 1

CoryJaJh panci flora (Stcph.) Pers.

Fig. 1. Raceme In fruit; X about 1^
Fig. 2, Habit of plant; X about IJ/2.
Fig. 3. Stigma; X 8.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 31

OWNBEY— MONOGRAPH OF CORYDALIS

A STUDY OF HEVEA (WITH ITS ECONOMIC ASPECTS) IN THE

REPUBLIC OF PERU!

R. J. SEIBERT-

Intkoduction

Natural rubber holds, and will continue to hold, a leading position among the
commodities of the world. It is relatively a newcomer among the necessities of
our advancing civilization, yet its absence would change the conveniences of
modern life to drudgery.

Various articles made from Hevea were perhaps first described from the
Amazon valley in 153 5 by the historian Ovicdo y Valdes. Two hundred years
elapsed before La Condamine, during 1734-1744, brought out samples of rubber
from the Amazon valley, introduced the strange material to European nobility,
and later published his reports, which Included a crude drawing of the Hcvca tree.
The Hcvca rubbertree received its formal botanical treatment in 1775 by the

French botanist Aubletj who described it as Hevea gnianeftus Irom material col-
lected in French Guiana.

Colonial policies of rigid forign trade barriers prevented rubber from reaching
the open market until after the Napoleonic invasion of Portugal, In 1823, how-
ever, the first commercial shipment of rubber reached the United States in the
form of several hundred pairs of rubber shoes manufactured by the Para Indians.
Stimulated by the invention of the vulcanization process by Charles Goodyear in
18 39 and the great advances of the automobile industry of this century, rubber
became the "gold" of the Amazon. An ever-increasing demand of a decreasing
supply resulted in the decline of the Amazon jungle exploitation. Its complete
collapse was brought about by the development of Hevea plantations In the British
and Dutch East Indies, which, with their inevitable large-scale industry and con-
sequent lower prices, rapidly took over world production during the second decade
of this century. Since rubber was no longer supplied to the United States from
South America, the penalty was paid during the recent war for virtual East Indian
monopoly. We can appreciate, now, the efforts of our government in attempting
to stimulate small farm rubber production throughout Latin America (Blandin,
1941; Kllppert, 1942).

Since Aublet's first description of the genus Hevea in 1775, about 100 species,
varieties, subspecies and forms have been described under various name combina-
tions. Although Hcvca brasilicnsiSy under cultivation, has been studied with con-

^ An investigation carried out at the Missouri Botanical Garden in tlie Graduate Laboratory of
the Henry Shaw School of Botany of Washington University and submitted as a thesis in partial
fulfillment of the requirements for the degree of Doctor of Philosophy. Field observations and data
were obtained in the course of the writer's oflficial surveys and jungle exploration work during 1940
to 1946 as Botanist for the Division of Rubber Plant Investigations, Bureau of Plant Industry,
Soils and Agricultural Engineering, Agricultural Research Administration, United States Depart-
ment of Agriculture.

^Botanist, Division of Rubber Plant Investigations, Bureau of Plant Industry, Soils and Agricul-
tural Engineering, Agricultural Research Administration, U. S. Departn^icnt of Agriculture.
Issued October 31, 1947.

(261)

[Vol. 34

262 ANNALS OF THE MISSOURI BOTANICAL GARDEN

slderablc detail, little is actually known about it, and less about the otber species
as tlicy occur naturally in the Amazon valley. The majority of collections and
studies have been confined to the navigable rivers, a narrow margin away from
these streams, and around centers of habitation. What Hevea forms exist, and
how they exist, between the major streams and their headwaters In the eastern
Andean foothill arc that skirts the range limit of the genus, will remain a question
,as long as intcrstream areas are undeveloped. This question will furthermore re-
main as long as transportation is largely confined to the main waterways and until
more than an occasional Individual with scientific interest studies the genus, its
components, and its ecology.

Through the studies of Dr. Adolfo Ducke, the great authority on Amazonian
botany, and those of Drs. Richard Evans Schultes and J. T. Baldwin, Jr., we are
coming to realize that the specific entities of the genus are limited to less than a
dozen. These entities can be rather clearly delimited morphologically and
ecologically In spite of considerable intraspeclfic variation. It is becoming evident
that both geographic and morphological forms arc being established within the
species, that interspecific hybridization frequently occurs in nature, and that ap-
parently few genetic barriers exist between the species.

With the recent war-time stimulation to wild-rubber tapping in the Amazon
valley, modern advances in transportation methods, and the ever-increasing need
for plant improvement programs, the Division of Rubber Plant Investigations,
U. S. Department of Agriculture^ has pursued basic studies of Hciea in Its natural
habitats. In addition, it has undertaken the selection of superior strains of
Hevea brasilienm^ as well as of other species and varieties from the jungles, and
established them in Tropical American experiment stations where a planned system

of Hevea plantation improvement is under way (Brandcs, 1941, 1943; Rands,
1942),

Material and data derived since 1940 from the efforts of the Department of
Agriculture, in cooperation with the Latin American Republics surrounding the
Amazon valley, have augmented substantially our previous knowledge of Hevea
as a whole. With such data we can begin to visualize In an over-all manner the
morphology, taxonomy, distribution and genetics of the genus preliminary to a
comprehensive monographic work, though many gaps remain to be filled in all
countries concerned. Moreover, we can begin to make sound progress, through
selection and breeding, toward the quality of planting material, disease resistance
and adaptability, as well as increasing production wlille lowering costs.

Scope of the Work

This study considers the genus Hevea as it is known to occur in the Republic
of Peru. Although I devoted three years to its study there, and others before me
have collected and written of the Peruvian species, vast areas exist between the
relatively few stations studied for w^hich hear-say and assumption still must take
the place of accurate information.

1947]

SEIBERT HEVEA IN PERU 263

It has, of course, been impossible to consider the Peruvian Hevcas without first
taking cognizance of speciation In the adjoining Amazonian countries, particularly
in Brasil. The genus, besides being of great economic importance, has the distinc-
tion of being sharply confined within the hylaca, or Amazon River drainage basin,
except in its northeastern distribution where the range extends to the watersheds
of coastal drainage basins in the Guianas and to some extent In southern Venezuela.

Speciation in Hevea has been considered a difficult problem and has been
treated in various fashions. Because the genus is composed mostly of large trees It
not only, has been difficult to collect but seldom has yielded complete study ma-
terial. The result has been that species often have been described solely on sterile
material, only to be described later as another species on the basis of fruiting or
flowering material. Comprehensive studies of comparative morphology are lacking,
although there has been some study of floral morphology, particularly by Hcmsley
(1898). Unfortunately, natural intraspecific variation has not been given due
consideration; and an amazing number of prominent and reliable specific characters
have been entirely overlooked.

Much of the following discussion will deal with several newly proposed floral
and vegetative characters which appear to have constant specific value and are of
great practical use to both the herbarium and the field worker. That such sig-
nificant characters have been overlooked previously is further evidence that con-
stant and repeated observation has no substitute for revealing new plant features.
Characters present but not previously recognized for their value, or inspirations
from that subconscious feeling of "indescribable differences," under repeated ob-
servation of the unit as a whole, may loom to the conscious as significant features
in the key to solution.

This paper is of necessity provisional It will try to bear out new evidence for
speciation in Hevca along with that known and used in the past. An attempt is
made to devise a practical key to the species which may be of equal use to the
herbarium and field worker. For the first time it will bring together the taxonomy
of the Peruvian Heveas as a unit, as well as those of the neighboring country,
Bohvia. The paper will solicit trial by those who encounter this group of plants,
and, It is hoped, will stimulate further comparative morphological observations by
others, so that monographic treatment of the genus may be eventually in order.
It will attempt to add its bit to the promotion of cooperation between the taxono-
mist and the geneticist, both of whom will have a long and fertile field of research
In this highly important genus of trees.

Acknowledgments

To the Division of Rubber Plant Investigations, Bureau of Plant Industry,
Soils and Agricultural Engineering, Agricultural Research Administration, U. S.
Department of Agriculture, with which I have been associated as Cooperative
Agent since July 1940, I am indebted for facilities and experience gained through
their Plantation Hevea Improvement Program carried on In cooperation with

[Vol. 34

264 ANNALS OF THE MISSOURI

thirteen Latin American countries. The Division has kindly extended to me a
year's leave of absence during which time It has been possible to continue my
studies at the Missouri Botanical Garden.

Material for much of this paper was gathered from the Republic of Peru. I
am deeply grateful for aid rendered by government officials concerned with the
rubber program in that country. The Estacion Experimental Agricola de Tingo
Maria, Peru, a cooperative agricultural experiment station maintained by the Office
of Foreign Agricultural Relations, U. S. Department of Agriculture, and
the Peruvian Ministry of Agriculture through its Peruvian and United States
employees, was of constant aid in propagating and maintaining living material of
jungle selections. To the many rubber tappers, dnr'mgcros (Seibert, 1947), with
whom I worked and Uvcd, must go much credit. They are the men who help one
in the "bush" when help is most needed. Tlicy are the men who know rubbcrtrccs
Instinctively, and from them came many basic facts which we in return can put

into scientific language.

Preparation of this paper has meant the amassing of most of the Hcvca speci-
mens from the major herbaria of the United States, namely, U. S. National Her-
barium, New York Botanical Garden, Chicago Natural History Museum, Missouri
Botanical Garden^ Gray Herbarium, Arnold Arboretum, and the National Arbo-
retum, to the curators of which I am indebted for loaned material.

Dr. John T. Baldwin, Jr., and Dr. Richard Evans Schultes, both of whom are
actively concerned with somewhat similar studies in other Amazonian countries,
have made available to mc their collections of Hevea, They have been most con-
siderate and helpful in discussions, both verbally and through correspondence, and
have tested for mc a number of morphological characters here proposed. Mr. Hans
Sorensen, Agent, Rubber Plant Investigations, stationed at the Instltuto Agron-
omico del Norte, Belcm, Brasil, has kindly tested and confirmed my observations
regarding the question of short-shoots. Dr. John B. Carpenter, Agent, Rubber
Plant Investigations, stationed at the Estacion Experimental Agricola de Tingo
Maria, Peru, very kindly has forwarded to mc herbarium speceimens of critical
material from my living jungle selections as they have come into flower since my
departure from Peru. To these Institutions and individuals and to many others
who have been of aid in this phase of the Hevea Improvement program, I wish to

coo

so much needed in coordinating large-scale programs of national and international

extent.

Morphology of the Genus

Habit .

Without exception the genus is woody. For the most part It is composed of
medium-sized to large trees, which in Hevea giiiaucnsis var. httea and especially
H. brasilienshy frequently may reach 45 meters in height under most favorable
growing conditions. Largest trunk diameters are found within H. hrasiliensiSy and
in Madre de Dios, Peru, it Is not uncommon to find trees 1 meter, occasionally 1.5

^To Dr. Robert E. Woodson. Jr., I am indebted for valuable guidance of the research and presen-
tation of the morphologic and taxonomic aspects of this paper.

i

1947]

SEIBERT HEVEA IN PERU 265

meters, in diameter at 1 meter above the ground.

In contrast to the large trees of the genus, two entities have been reported in
which the habit is low and shrubby. The case of H, camporum Ducke (1925),
collected by R. Monteiro da Costa in the campinas^ between the headwaters of
the rivers Manicore and Marmellos, southern tributaries of the lower Madeira, is
very poorly understood because of sparse herbarium material collected. The region
from which it comes apparently is hilly, semi-open, grassy and scrub-forest land,
unfavorable to good tree growth. According to Ducke (Schultes, 1945) the
species perhaps is only a dwarf form of H. patwiflora var, coriacca.

The other interesting case of dwarfed or shrubby habit in the genus is Hevea
nitida var. toxicodendroides (Schultes, 1947), discussed by Schultes (1944) and

*

described as H, viridis var. toxicodendroides Schultes & Vinton, This variety was
discovered in Colombia from the upper Apapores Basin, growing on and around
apparently old sandstone outcrops of at least Triassic age, on which scmi-
xerophytic conditions exist. These plants are about 12 feet tall and are quite bushy
or shrubby in aspect. Otherwise they resemble the species in morphology and size
of the leaves, flowers, fruit and seed. Thus it appears that variation may be more
ecological than morphological. Experimental growing only can determine this.

From apparent intcrgradations between the normal type of tree branching and
the low branching of the shrubby types one may observe and interpret conditions
frequently encountered in plantations of the commercially grown H. brasiliensis.
Pruning is a common practice and certain "clones"^ under normal conditions tend
to form a low branching habit. It is necessary, through pruning, to prevent the
formation of such low branches as it would interfere with a good tapping panel.
Under conditions of undue drought and where soil Is not suitable for growth of
H, brasiliensis almost all individuals of the species will tend to have profuse and
low-branching habits, not at all typical of the same plants growing under natural
conditions.

Trunk.

Growing under flooded conditions, where the trees stand in various depths of
water for nearly the entire year as do H, Spm^ceana, Benfba?niana and micro phylla"^^

the trunks are distinctly swollen toward the ^base. This "bellying'^ is ^oTntedly
referred to by the Brasilian name for H. Spritceana, seringucira barriguda^. The
amount of bellying and the height to which it extends may possibly be some indi-
cation as to how high flood waters reach up the trunk. Above the uniformly swollen
portion, the trunk suddenly tapers upward. A graphic example of this bellying

effect due to flooding appears to be well illustrated by contrasting the H, hrasil-
iensis of the periodically inundated land, ficrra baja^ of the upper Amazon proper,

'^Campinas in Brasil refer to grassy, scrub-forest hilltop land.

^A clone in Hevea plantation terminology refers to an individual tree which is vegctatively
propagated through successive generations by means of bud-grafting.

'^'Schultes (*47) has shown from type studies at Kew that H. Tnifjor, as at present known, should
be referred to H, microphyUa.

^Seringucira barriguda in Portuguese means big-bellied rubbertrce.
Tierra baja in Peru refers to low inundated land along rivers; and ficrra aliura to land above
the level of river flooding.

[Vol. 34

266 ANNALS OF THE MISSOURI BOTANICAL GARDEN

with that growing on well-drained ficrra altura of Pando in Bolivia and Madre de
Dios in Peru. Wherever the trees are subjected to periodic inundation the swollen
bases are conspicuous and rather suddenly tapering several feet above the ground.
On tkrra altura the trunks are definitely cylindrical with no more than the normal
gradual tapering In girth.

The normal trunk for the species growing on well-drained land is cylindrical
to the ground level with a very slight tapering or girth decrease upward. Under
usual forest conditions the first branches depart from the upper third of the tree.
It is not uncommon to find forest giants with 9b to 100 feet of cylindrical trunk

to the first branches.

There appears to be no reference to, nor have I seen any cases of buttressing.
However, Dr. Baldwin informs me that he found a buttressed tree on the Rio
Negro, which he is inclined to feel resulted through intcrgeneric hybridization with
Cuuurla, It would appear that buttressing Is not a characteristic of Hcvea.

Interesting evidence concerning the age of H. brasiliensh trees from Bolivia
and the Acre Territory of Brasil is given by La Rue (1926). These areas and that

of adjoining Madre de Dios, Peru, arc characterized by having one distinct rainy
and one distinct dry season per year, conducive to the formation of annual tree
rings. Although no Increment borer was available, I was able to observe the trunks
of a number of trees at Iberia, Peru, which were felled to make way for an air
strip. Annual rings were not noticeable until nearly a year after felling, when
partial decomposition of the wood had set in. A radial cross-section of one of
these trees, 84 cm. In diameter, showed 211 annual rings. There Is indeed Uttle
doubt that some of the largest forest giants were already growing before the
discovery of America.

B-ark. — True outer bark color and other characteristics in Hevca are often dis-
tinctively overshadowed by the predominance of crustaceous lichens which makes
the trees easily spotted in the forest. As yet there has been insufficient investiga-
tion and description of the bark as applied to specific delimitation. In general, It
Is quite smooth with some scaling and color range from light gray to dark brown.
There appears to be much Intraspccific variation in bark characters, ranging

through very smooth, pustuled, flaky, shaggy, to definitely corky within H. hrasil-

icnsis itself. An extreme has been described by Bartlett (1927) as H. hrasilicusis
mut. Grantham}, Bark variants of this and many other types occur not only in
plantation material but in the wild as well, together with the intcrgrading forms.
Schultes (1945) reports evidence from Colombia that bark variations are of Im-
portance In distinguishing subspecific variants In Hevca. These bark variants are
important factors in a selection of trees for plantation use, in that ease of tapping

may bo considerably hampered by such rough strains. Outer bark variations would
appear to be of considerable use in clone distinction (Frey-Wyssling, 1933).

The inner portion of the bark, or phloem, is, from the economic standpoint,

the important part of the tree anatomy, since the latex vessels located here furnish
the natural rubber of commerce when cut in the process known as "tapping."

1947]

SEIBERT HEVEA IN PERU 267

Here again there may be as much or seemingly more intraspccific as inter-
specific variation. Within H. brasilicnsis the phloem of mature trees may
vary in thickness from about 0.5 cm. to about 2 cm. Apparently, due to the
cambial development of concentric rings of phloem, the latex vessels occur in con-
centric rows, the number of which is highly variable, ranging from about 8 to 3 5.
There would seem to be no correlation between number of latex rows, their indi-
vidual size or productivity, and age or size of the mature tree. Perhaps one of the
most striking variations in phloem is the color. This variation, or series of varia-
tions, for H. brasilicnsis appears to reach its maximum toward the southwestern
and western part of the species range. Observations of thousands of trees seem-
ingly have shown all range of variation through tan, brick, purphsh red, reddish
purple and blackish purple. La Rue (1926) has a considerable discussion of these
color variations. There is a feeling among the rubber tappers that the trees with
purplish phloem, and particularly those with the Irreta^ or blackish purple color,
give the best yield and the superior quality of rubber. Although there may be
such a tendency, this color is by no means a constant criterion as evidenced from
jungle selection work carried on in the Madre de Dios area.

Texture of the phloem likewise is variable, ranging from very hard, with high
number of stone cells, and difficult tapping, to soft, easy-cutting, almost "cheesey"
texture. There seemingly is a tendency for trees with purplish phloem to have a
softer texture than those with more tannish phloem. It is interesting to note that
though a tree may have virgin bark of a tannish or light-colored phloem, the re-
newed phloem from previously tapped portions of the same tree has a reddish or
purplish color. Trees with a reddish or purplish virgin phloem, as well as those
with renewed secondary reddish or purplish phloem which is tan in virgin condi-
tion, often exude a reddish or purplish dye-like fluid independent of the latex flow
when freshly cut.

With the exception of the seemingly complete absence of so-called [yrcta trees
from the lower Amazon region, there would seem to be no geographic, ecologic or
edaphic range limitations of these various bark variations. Any randomly selected
estraJa^ from the Madre de Dios may contain a fairly complete range of color and
other bark variations within close proximity. There furthermore is no conclusive
evidence from the Madre de Dios region that any of the numerous bark variations
are constantly associated with such intraspecific morphological variations as leaf
size and shape.

Although little is known of bark variations in other species than H, brasilicnsis,
it appears that similar variations exist in H. gutancnsis var. lutea. Here the outer
bark also shows variations in type and amount of scaling. The phloem color ap-

'Prcta, meaning black in Portuguese, Is used to describe trees of H, hrasiliensis in which the
phloem is distinctly blackish purole in color.

^An estraJa consists of about 100 to 150 rubber trees, joined one to another by a forest trail
running from and returning to the shirivgcro's house in somewhat loop fashion. A map of an
estrada has been excellently figured by Preusse-Sperber (1916). The rubber tapper is usually
assigned two estradas, the trees of each being tapped on alternate days.

[Vol. 34

268 ANNALS OF THE MISSOURI BOTANICAL GARDEN

pears to vary from ligKt tan to a somewhat purplish red. The texture and thick-
ness also are quite variable.

Ijitcx, — Slight chemical differences distinguish the rubber from the latex of the
different species. Great differences, however, do occur when the latex as a whole.
Including Its serum and non-rubber content, is taken into consideration (Parkin,
1900). It is well known that H. hrasiJicfisis as a species gives the most abundant
yield and the best quality of rubber from the manufacturing standpoint, while H.
BenthamJana probably ranks second. Tt is also true, but less well known, that
In the species H. brasiliensiSy within areas of its distribution as well as between
Its regions of distribution, there are great variations in these qualities. On the
basis of these facts, large-scale selection and later breeding programs in the Far
East have been able to step up yields per acre from less than 450 pounds annually
to recent experimental yields as high as 2000 pounds. Such has been done from
the original stock of less than 100 trees of H. brasilionis^ finally surviving from
those Introduced by Henry Wickham into Ceylon and Singapore from seed he col-
lected near Boim, Rio Tapajoz, Brasil. From this might be visualized ultimate
yield possibilities by choosing stock from such areas In the southwcstcrnmost
limits of H. hrasiliensis, as In the Mad re de Dios, where average tree yields are
proving to be some three times as great as the region from which the Far Eastern
stock was obtained.

Although Zf. gtmncnsis and Its varieties may be characterized partially as
having yellow latex and considered as having low yields of weak rubber, there is a
great deal of tree to tree and area to area variation. It Is usually noted within
this group that the color of the latex may become nearly white after successive
tapping over periods of time. Latex color appears to be no criterion of quality
inasmuch as both H. brasiUonis and //. Spniccaria nre characterized by white
latex. The former species produces excellent quality and the latter very inferior
quality shunned by all rubber tappers. Cases of latex reaching sulphur-yellow
tones are known to exist in //. brasiJicnsis, both among plantation-developed
clones and among jungle trees. There is no evidence that such individuals produce
inferior rubbers. Trees have frequently been found In the jungle selection work
where the latex of the young branches Is yellow, while that from the trunk is

perfectly white.

Cases arc known where the latex is very his^h in resin content, producing a very
"tacky"^^ rubber. Such occurrences are frequent in areas below Tquitos where
apparently introgression of H. [yuicijlora Into the H. giiiancnsis complex has re-
sulted In the weakening of the rubber to such an extent that freshly coagulated
rubber fails to keep its shape even over night. H. [yauciflora itself has a very poor
resiny rubber and never Is tapped commercially. Allen (No. 3^9^) reports a

*^ According to Mann (1940), "It seemed fairly clear, however,^hat the whole of the rubber in
Malaya originally came from 27 seedlings of the original Wickham collection that went to Singapore
in 1876."

^° "Tacky" rubber is that which in coagulated form remains rcsiny and sticky, refusing to
maintain its original shape. It is usually very weak with little elasticity.

1947]

SEIBERT HEVEA IN PERU 269

collection of H. nitida from tlie Vaupcs region of Colombiaj in which adulteration
of the latex with that of good latex from other species resulted in the ruining of
all the latex, even preventing proper coagulation.

Latex yield, and often its color and quality, cannot be gauged from the first
incision made into the tree. Both plantation and jungle tappers have found that
the normal tree is stimulated by what is termed "wound response"^^ under system-
atic tapping. Yields from the individual tree Increase as much as 20 per cent
after several tappings or even several weeks of tapping, and so long as regular
systematic tapping Is carried on this optimum yield is obtained. On the other
hand, rare cases have been seen in which the tree gradually "drys"^^ by the same

system.

Latex flow from tapped trees is, of course, affected to some extent by the water
content of the latex. Individual trees in the same jungle area will show consider-
able normal deviation from the accepted average of about 30-3 3 per cent Dry
Rubber Content (DRC) and appear to range from between 20 per cent to as high
as 45 per cent DRC, Seasonal variations also occur and lag somewhat in direct
proportion to the amount of rainfall. A climate with a more or less uniform
rainfall throughout the year and with no long pronounced dry season would be
necessary if year-around tapping were desired. A further factor affecting latex

flow, and one which is poorly understood, seems to be related to temperature and
time of day at which tapping is done. Plantation experiments appear to have
proved that flow and yield are substantially larger if tapping is done very early In
the morning, before the sun reaches its full effect, or before 10:00 a. m. Further-
more, on cloudy days the time of maximum flow is extended considerably. Similar
effects have been observed by the jungle tappers and apparently is in part directly
responsible for the custom of many tappers to rise as early as an hour or two
before sun-up and start their work with the aid of a lantern. During abnormal
periods of cold weather in the upper Amazon reglons^^ it also has been observed
that tree yields considerably increase above normal.

Branch system, —

The normal branch system of the genus is composed of the prominent erect
main axis from which arises a symmetrical system of secondary branches. Under
usual forest conditions branching is found on the upper one-half to one-third of

^^ "Wound response'* (Royal Botanic Garden, Ceylon, 1899) is the plantation term given to the
normal increase in a tree's yield induced by systematic regular tapping over a period of days or
weeks, from the comparatively low first-tapping yield to the normal yield of the same tree after it
is in regular tapping.

^"It is frequently convenient, wlien speaking of Hcica trees, to use the terminology of the dairy
industry. To go "dry" is one of those terms. .

^^A rather broad area of the upper Amazon valley, roughly between Porto Vclho and the Andean
foothills, is subject to yearly cold periods lasting about 2—3 days known in Brasil as friagcin and
in Peru variously as friaje or varuza. One or more of these periods usually occurs around the
months of July or August during which time strong southeast winds, recorded at Maldonado as
high as 100 mph., precede a driving rain of short duration. This is followed by lighter winds,
cloudy weather and temperatures as low as 10 C.

270 ANNALS OF THE MISSOURI

[Vol. 34

tlic tree. Radical exceptions to this form occur in H. caniporuii} and H. nitida
var. tox'icoilcuJroiiJcs which are characterized by low shrubby forms. Wind
damage to forest trees, as well as artificial topping or pollarding occasionally
practiced* in plantations, appears to have similar effects in stimulating strong lateral
branch development. There also would seem to be inherent branch differences
among individuals or strains of the same species.

Two major habital forms within ff. brafiltcns'n may be observed: (1) possibly
the typical form in whicli there is a prominent main axis with small lateral
branches; and (2) the form with lateral branches as prominent as the central axis
and frequently even replacing it. Although plantation material appears to b
largely of the latter type, representatives of the former also occur. Seed progeny
from the low-river areas, as Belcm and Rio Tapajos, as well as from such up-river
areas as Acre, Madre de Dios and Iquitos, all growing at the Estacion Experimental
Agricola dc Tingo Maria, show both types of branching and intermediates to exist
from all areas. There is a tendency, however, for the latter type to predominate

in progeny from low-river areas, while the former appears to be predominant in
up-river areas. '

Anotlier branch variation, discussed with reference to clonal differentiation by
Frey-Wyssling (1933), concerns the angle and form at which the lateral branches
arise from the main axis, being from nearly vertical to nearly horizontal. Until we
have more information from both natural and cultivated habitats for all of the
species, branching habits would seem to be primarily intraspecific variations and
of more use in clonal than in specific delimitation.

Roots.

W

«

A

long, prominent tap root seems to be the rule, and also that rather prominent
laterals radiate out from just below the ground surface to form a surface-feeding
network. Under certain cultivated conditions, such as are found at Bayeux, Haiti,
where a permanent high water table lies one and a half to three feet from the sur-
face, the long tap root of H. brasilicnsis either branches profusely at the water
table or suddenly turns and follows parallel with it. In the Madre de Dios region
where the species grow naturally on well-drained land, the root system is as found
under normal plantation conditions. In the Tquitos area, where the species nat-
urally grow part of the year In several feet of water and the rest of the year with
a comparatively low-water table, the roots of several trees, uprooted by high winds,
also have been found to be of the normal type. I know of no reference to the root
systems of such species as H. Bcntbawlana, nihropbylla and Spruceafia, which
naturally grow permanently in water. //. Spruccana^ however, does react nor-
mally In its root development when planted on well-drained land.

Long-sboots and sborf-sboots (Flusbes and hitcrflushes): The genus Hevea is
characterized by periodic (annual except In seedling stage), rapid stem growth
or elongation alternating with a long period during which stem clonc;ation

1947]

SEIBERT HEVEA IN PERU

271

- — ^^ • • • • • A

4

F

c

%^^^^^

X)

1 1'^Aa

E

r

A

B

C

D'

B

.. E

Text-fig. 1. Leafy shoots (flushes) alter-
nating with relatively elongate caducous-
scaly short-shoots (interflush short-shoots) :

A — conspicuous interflush short-shoot; B —
long-shoot or flush; C — lateral branch de-
velopment from a short-shoot or spur; D —
leaf-scar; E — leaf scales on lower part of
flush adjoining the short-shoot; F — lateral
spur development.

Text-fig. 2. Leafy shoots (flushes) alter-
nating with narrow rings of bud -scale
scars (interflush rings) : A — inconspicuous

interflush short-shoot; B — long-shoot or
flush; C — lateral branch development from
an inconspicuous short-shoot; D — leaf scar;
E — leaf scales on lower part of flush ad-
joining the interflush ring; lateral spur
development inconspicuous.

[Vol. 34

272 ANNALS OF THE MISSOURI BOTANICAL GARDEN

and leaf formation are arrested or reduced. These intervals of rapid stem
development or long-shoots have been referred to In plantation use as "stories"
(Frey-Wyssling, Heusser & Ostcndorf, 1932) and perhaps more popularly as
"flushes" (Assoc. Cent. Exp. Sta., 1939). With the exception of noting the

presence of terminal bud scars, and of scale buds at the area between "flushes," the
center of the "corona"^"* region in budding terminology, little attention has been
paid to this character in specific determination. It appears that the interval be-
tween flushes, or terminal bud scar region (text-fig. 1), might best be known as
the "interflush" which, perhaps, is referable morphologically to a **short-shoot"
(pi. 32, fig. 2).

No literature has yet come to my attention making any reference to one of the
most striking morphological characters in the genus Hevca — the '*short-shoot."
The short-shoot, or spur, is well known In Ciuk^^o (Collins, 1903) and a number
of conifers (Chamberlain, 193 5) . Short-shoots occur in the rosaceous fruit

trees. They would appear to occur on the underground rhizomes of Voa and per-
haps other grasses Including Bamboo (note figs. 23 & 123 in Arbcr, 1934) and

may be found in some of the higher plants, both tropical and temperate. It would
not be hard to presume that a comprehensive study of short-shoots throughout the
plant kingdom not only would furnish an Interesting field of research, but also
might lead to important taxonomic and evolutionary data.

Short-shoots generally have been given attention only when they develop
prominent lateral spurs. They may continue for several years before suddenly
changing their growth nature by transformation into a long-shoot. However,
close examination shows that they may, and frequently do, occur terminally on
the mnin axis and later are seen to be alternating between two long-shoots. Due
to much compression of the intcrnodes on these shoots, the normal long-shoot leaf
phyllotaxy superficially appears to be changed. Furthermore, the leaves arising
from these short-shoots are not entirely normal (Chamberlain, 193 5).

In collecting Hevca for the herbarium, branching specimens are seldom chosen
for the press. Thus little evidence of the presence of spurs is ever presented to the
herbarium worker. In Hevca the short-shoots are not prone to linger for several
years before developing into a long-shoot, though such has been seen on rare
occasions. Ducke (1935) apparently came close to recognizing their presence,
when, in his description of //. guiafionis^ he remarked: "Old branchlets with per-
sistent scales at the terminal bud," and with regard to H. Spmccana pointed out
"the numerous pointed scales which cover the vegetative huds."

Since, in the plantation, only H. hrasiUensis has been under critical observation,
and since short-shoots are so condensed (text-fig. 2) in this species as to be rep-
resented by little more than a narrow ring of a few scale bud scars between flushes
(pi. 32, fig. 1), the character for the genus as a whole has not come Into recog-
nition. Even the classic work of Parkin (1904) on nectlferous and non-ncctiferous

^^"Corona" is a term being adopted by the U. S. Department of Agriculture to the area of the
stem on which the buds are most crowded together.

1947]

SEIBERT HEVEA IN PERU 273

bud scales fails to recognize the presence of the short-shoot. Furthermore, in the
nursery where close observation might well bring out such a feature, we find that
young Hevea seedlings and buddings have somewhat different vegetative habits
from older trees. The young plant sends out a new vegetative flush barely after
the previous one has matured. All the leaves from several flushes are persistent,
whereas after the tree reaches about three years of age it will normally send out
only one flush per year and that only after defoUation of the previous one, as in
the case of H. brasiliensh. It is only after growth-habit maturity has been at-
tained that the species showing distinct short-shoots will develop the character.
We have a rather parallel condition in Ginkgo where the young plant develops only
long-shoots until after it has reached some size.

The length of the flushes appears to be quite variable from year to year, prob-
ably depending both upon climatic conditions and upon the amount of shading the
branch has received from other branches. From this standpoint the short-shoots
of any one species appear to be less variable in length by comparison. As yet no
instance has been seen where the short-shoot grades imperceptibly into the long-
shoot.

The short-shoot In Hevea makes its terminal (and axillar>0 appearance during
the maturation period of the flush and reaches conspicuous proportions just before
the appearance of the inflorescence (pi. 33, fig. 1), In the species where
short-shoot is pronounced (H. gnianemh, Spruceana, paiccijlora, righJijolia and
nitida) it consists of a segment about 1 cm. long (pi. 32, fig. 2; pi. 3 3, figs. 2-3)
covered with caducous scales (pi. 3 3, fig. 1-2) which morphologically are reduced n

th

Wy

The nodes of this portion

of the stem are much compacted and number as many as 100, with Internodes
naturally being obsolete. The first branches of the inflorescence seem to arise
from the axils of the uppermost scales (pi. 33, fig, 2) and shortly may be fol-
lowed by the appearance of the young long-shoot or flush carrying the leaves.
Frequently inflorescence branches also arise from the axils of the lowermost leaves
of the long-shoot. On the older portions of the stem these short-shoots alternate
with the long-shoots and are conspicuous at the Interflush regions (pi. 3 3, fig. 3).

In those species where the short-shoot Is much condensed or inconspicuous (H.
brasilknsis, microphylla and Bcnthamiam — pi. 32, fig. 1), one not familiar with
the genus as a whole might have difficulty In recognizing it as such. Neverthe-
less, the short-shoot is morphologically the same as In the previous group except
that the scales or reduced leaves are very few and can be recognized at the inter-
flush areas of the older stems as only a narrow ring of few bud-scale scars.

Cases of both natural and artificial hybridization between species of contrasting
short-shoot condition, as H. brasilicnsis X Sprnc^ana, H. Bcnthainiana X g^iianen^
sis, H. hrasiliensis X pancijlora, etc., appear to show the conspicuous short-shoot,
indicating its probable dominance as a factor in contrast to its opposite extreme in
which the short-shoot is represented bv an inconspicuous narrow ring of bud-scale

274 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

scars. Hybrids frequently appear which closely resemble the species except for
having the conspicuous short-shoot. Such cases must be borne in mind when oc-
casional specimens of H. Benthamiana varieties, for example, refuse proper place-
ment in the proposed taxonomic key.

This conspicuously contrasting character is considered Here as being of key
significance in dividing the genus as a whole into two groups. Whether or not
these two groups are entirely natural has not yet been determined; in fact, in com-
bined consideration with many floral characters tlicy do not appear to be entirely
natural. Nevertheless, these divisions seem to be of more practical importance
than previous ones based on anther number and anther whorls. It is hoped that
further detailed study may be given to the question of short-shoots by others
engaged in Hevea field work.

Leaves, —

The general structure of the leaf is uniform throughout the genus, being digi-
tately 3-foliolate with the leaflets joined to the relatively long petiole by noticeable
petlolules. A small, lateral, early caducous stipule is found on the stem at each
side of the petiole base. Although 3 leaflets arc the rule, 1, 2, 4 or 5 leaflets, or two
of the leaflets grown together in varying degrees, may be found on rare occasions
(Frcy-Wyssling, 1931). Such abnormalities are found in both the plantation and
jungle where either an occasional leaf may be abnormal or many of the leaves
from the same tree show such conditions. Seedling leaves, so far as I have been
able to observe, likewise normally are 3-foliolate. When leaf abnormalities do
occur there seems to be a stronger tendency for the production of more than 3,
rather than fewer leaflets. Conspicuous glands, or extra floral nectaries, which
normally occur on the upper surface of the petiole just below the junction of the
3 petlolules, are apt to be extremely variable even on the same tree. It is from
these pctlolar glands that extra leaflets appear to be derived, and they may or may
not be of stlpular origin.

Phyllotaxy has been determined for H. brasilicnsis (Ostcndorf and Ramaer,
1931). Leaves are spirally disposed on the flushes and diverge at an angle usually
of 138° (2/5) or occasionally at 103° (2/7). This appears to be the extent of
variability throughout the genus. Some deviation from this might be expected In
the bud scales or on the short-shoots of other species. A comparative study from
fresh material will be undertaken at a later date in connection with a thorough
consideration of the short-shoot w^ithin the genus,

isling, Heusser, and Ostendorf (1932) have discussed in detail the.

Wv

servations

tremendous variations of all parts of the leaf in //. braulicnsis, Ol
both field and herbarium specimens of other species indicate considerable intra-
spccific leaf variation to be characteristic throughout the genus. With some
striking exceptions, which wall be pointed out in the followmg paragraphs, leaves
hold greater significance and consistency in distinguishing individuals or clones
than for specific distinction.

^

1947]

SEIBERT HEVEA IN PERU 275

Leaf persistence: Those who have observed Hevea growing, both In the planta-
tion and in the Amazon valley, are familiar with the deciduous wintering habit of
//. brasiliensis. During or near the end of the dry season it ''winters" by losing all
the leaves of the most recent flush and going through a dormant period of

several weeks. The dormant period is followed by the appearance of inflorescences
and, immediately after, the rapid growth of the new vegetative flush (pL 34,

fig. 1).

Under jungle conditions in any limited region, the trees of H, brasiliensis

usually go through these stages at the same time, any one tree not lagging or being
advanced by more than a few days. Under plantation conditions, particularly
where there is no pronounced dry season or in abnormal latitudes or altitudes, the
leaf-shed and flowering seasons may vary several months from tree to tree. In
some extreme cases the same tree may be In several stages at the same time. Some
branches may be defoliating and resting, others sending out inflorescences and
young flushes, while still others may have mature vegetative flushes and already
maturing fruit. Yet, so far as I am aware, within H, brasiliensis^ the previous
flush always defoliates completely before the appearance of the inflorescence on
that particular branch. It Is said that a few oriental cloncs"^^ tend to hold their
leaves while flowering, but the details are not clear to me. I am of the opinion that

■

the actual branch which is flowering has defoUated; or, it may be that the clone
is not genetically pure for the species. Such instances have been observed at Tingo
Maria and Iqultos where individuals of Iquitos origin (referable to H. brasiliensis
but very likely carrying genes of another species) hold a few leaves on the previous
flush while flowering.

It should be explained that plants up to about three years of age do not show

this seasonal wintering. They normally send out new flushes regularly every

month to six weeks. During this stage of growth, leaves of about the top three

flushes persist while those on the lower ones gradually absciss. The transition from

this habit to the mature growth habit is sharp. At the proper age and size the

young plant undergoes complete defoHation and continues thereafter to send

out yearly flushes which completely defoliate before the appearance of the next
flush.

There is good reason to believe that the H. brasiliensis type of defoliation and
wintering does not occur In all species of the genus. No information on the sub-
ject appears to be recorded, and detailed field study of the flush and inflorescence
habits of the genus as a whole is wanting.

The contrasting condition seems to exist in which leaves of mature tree flushes
are persistent until well after the appearance of the inflorescence, its maturation
and the presence of the new flush (pi. 33, figs. 2-3; pi. 34, fig. 2). It Is not clear
what, if any relationship exists between interflush rings and interflush short-shoots
and the degree of dormancy. Unfortunately, insufficient field evidence Is available

'Oriental clones are those developed and selected from cultivated Hevea brasiliensis growing on
Far Eastern plantations.

LVoL. 34

276 ANNALS OF THE MISSOURI BOTANICAL GARDEN

to draw final conclusions, particularly from cultivated conditions wKere all tlie

L

species are growing together in the same climate and soil. Mostly from evidence
of herbarium specimens, the following categories appear to be evident with refer-
ence to the various species:

1. Leaves persistent on the previous year's flush until well after the appearance of the inflores-
cence or its maturation.

H, Spruccaifa — From the specimens it seems that the new flush is very slow in appearing
after the Inflorescence and that the fruit may be well along in its development before
the new flush appears. Meanwiiile, the previous flush leaves are very persistent. All
inflorescence branches arise from the axils of the scalei on the short-shoot (pi. 3 3^
fig. 2). ^

H. rlgidifolia — The previous flush leaves are very persistent until well after inflorescence
maturation. From the few specimens at hand it is not possible to determine how long
after inflorescence maturation the new flusli makes its appearance. As above, all in-
florescence branches arise from the short-shoot scale axils.

J1. pauciflora — ^The previous flush leaves show a very strong tendency to persist until after
the appearance and maturation of the inflorescence, as well as the appearance of the
new flush which follows almost Immediately after that of the inflorescence. Most of
the inflorescence branches arise from the short-shoot scale axils but some may arise
from the axils of the lower leaves on the new flush. A few specimens show a tendency
for defoliation at the time the new flush makes its appearance. This may be due
partly to removal of excess leaves by the collector (a frequent bad practice to
''improve'* tlie herbarium specimen) or it may be partly due to frequent hybridization

with other species. The study of these characters in living material is especially
critical in the H. pauciflora complex.

H. guhincnsis and varieties — A stronger tendency is shown for defoliation or partial
defoliation preceding the appearance of the Inflorescence than in H. pauciflora. Never-
theless, the majority of the specimens show at least a few leaves, particularly upper
ones, remaining at the time the inflorescence appears and until the appearance of the
new flush which shortly follows. Most inflorescence branches arise from the short-
shoot scale axils. However, as in II. pauciflora^ the upper inflorescence branches tend
to arise from the axils of the lower leaves of the new flush.
H, mihia — This species is so very poorly represented in flowering material that little can be
said except that some previous year's flush leaves (upper) tend to persist until after
the appearance of the inflorescence. Although the basal inflorescence branches arise
from the upper short-shoot scale axils, a greater number of the upper inflorescence
branches arise from the leaf axils on the lower half of the flush.
2. T. caves not persistent on the previous year's flush, /. c. they absciss before the appearance of

the inflorescence on that branch.

H. Bcntbamrana — Leaves are completely absclssed before the appearance of the inflorescdnc^.
The flush immediately follows the appearance of the inflorescence. Tlie basal in-
florescence branches arise from the uppermost bud-scale axils but mostly from the
axils of the leaves on the lower two-thirds of tlic flush. A number of specimens,
tiipcrficially resembling H. Bcuthamlana^ show some persistent leaves, but close exam-
ination of the leaf, shoot and flower gives ample evidence that the specimens
represent hybrids between H. Bcnthamhna and //. pauciflora^ ^ulancnsh var, lufca,

or Sprnccana^ from which the persi?;tcnc leaf character probably Is derived.

H. microphylld — Leaves arc completely absclssed before the appearance of the inflorescence.
Lower inflorescence branches arise from the upper bud-scale axils while the rest arise
from the leaf axils of the lower third of the flush. However, on account of lack of
sufficient material, the above observation may be somewhat inaccurate.

H. brasilicnsh — Leaves are completely abscissed before inflorescence appearance. Lower-
most inflorescence branches arise from the lowermost scale leaves of the flush and
continue to arise from the axils of nearly all leaves of the flush except from a few of
the uppermost ones^^.

The fact that most axillary buds of the flush leaves have developed inflorescence branches may
be of importance when it is necessary to bud graft material from the branches of mature trees. If
such budding be necessary, as it is in selecting Jungle trees for experimental and plantation use, one
should choose so-called "corona" buds, those which occur in a rather crowded position near the
terminus of the flush and from the inconspicuous short-shoot where no inflorescence branches likely
have arisen.

1947]

SEIBERT HEVEA IN PERU 277

17

It perKaps is of importance to note that the species falling in the first category
showing persistent leaves are also those considered to have conspicuous short-shoots.
Those of the second category with non-persistent leaves coincide with those species
without conspicuous short-shoots. Although the character is used to a limited
extent in the proposed key, I cannot as yet place full emphasis on it until further
study has been made from living plants.

Leaflets: In a consideration of leaf characters as an aid to distinction of the
various species one is forced to choose carefully and even then some overlapping
can be expected. Although Hcvea specimens usually are very large and difficult
to collect, some leaflets may be found on the ground beneath the trees at any season.
It is important, then, that we be able to derive maximum specific use from even a
fallen, disintegrating leaflet and the aid of a pair of field glasses in determining an
uncoUectable forest giant from some area where Hevea has perhaps previously not
been collected.

Position: Position of the mature leaflets with reference to the petiole axis is of
prime importance in associating a particular plant within certain groups of species.
Two distinct contrasts may be found, /'. r.: (1) Leaflets erect to slightly horizontal,
in which case they stand up above the axis of the petiole to which they are at-
tached. This type of position is found in the H. giiianensis complex. (2) Leaflets
distinctly horizontal to reclinate, the leaflets standing out in a plane parallel to
that of the petiole or hanging down nearly perpendicular to the petiole axis. Hevea
brasilieusis and H. rigidifolia are typified by reclinate leaflets while the remaining
species are rather intermediate between horizontal and reclinate.

Lower lepidote surfaces: H, fiitida is the only species in which scales are lacking
or, at most, extremely sparse and sufficiently microscopic to produce a strikingly
concolorous leaflet. All other species are characterized by having varying densities
of minute, whitish, epidermal scales distinguishable with a strong hand lens. The
presence of these scales accounts for the characteristic dull lustre and occasional
whitish color of the lower surface. The scales are most densely disposed on the
mature leaflets of H. rigidifolia and H. patici flora, where they lie one against the
other in tile-like regularity and are noticeably angular (hexagonal) in outline. In

■

H. Bentbamiaiia they are quite densely disposed but appear to be rather lens-shaped.
In all other species they approach a circular outline, are more or less separated from
one another, and vary in density from tree to tree. This specifically insignificant
variation of density accounts for the description of a variety siibcojicolor under H,
brasiJicfisis. This species shows a tendency for many individuals, throughout its
entire range, to have relatively fewer, more widely spaced scales which in extreme

i

cases would appear to produce the "subconcolor" condition.

■^^Since this character is frequently hard to distinguish in herbarium material, it is advisable that
collectors note it In their observations.

[Vol. 34

278 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Pubescence: When present, leaflet pubescence is confined to the lower surface.
In H. Bcnthamiana and H, Spraceana pubescence normally is found over the entire
lower surface, associated with the veins and veinlets. In the former species it is
usually reddish (occasionally mixed with whitish), while in the latter it is whitish.

In H. guian4^n$h and H. gnlanemh var. liitca varying amounts of whitish or
reddish-tinged pubescence normally are found associated with the midvein. In
the latter the tendency is towards reddish or mixed reddish-tinged and whitish
pubescence. In //. gnianevsis var. marg/nafa^ a slight whitish pubescence generally
is associated with the midvein but may be harder to distinguish than in the
previously mentioned members of the complex^ ^. Type specimens of H. pauciflora,
as well as those of IL confusa and H. pauciflora var. coriacea, all show very slight
whitish pubescence associated with some portion of the midvein, usually within
' the upper half.

All the other species normally are without pubescence. Frequently, at least In
H. hrasilicnshy the very young leaflets may show some pubescence during the first
days of their appearance. These minute hairs, however, soon are entirely caducous.
Their presence or absence varies from plant to plant. There are specimens referable
to H. brasilicnsh in which pubescence can be noted along the midvein and even
the lateral veins. Close observation, however, has given additional morphological
evidence that they probably are introgressive hybrids.

Shape and Size: In many respects these features, especially the size, should be
given very little significance. In the past, they have been responsible for perhaps
more confusion and promiscuous varietal description than anything else. Size
varies with the age of the plant; it varies from year to year depending on climatic
conditions; and most noticeably It varies with the portion of the flush from which
the leaves are taken. As a general rule, //. viicrophylla has the smallest leaflets
within the genus, while H. brasilknsis produces some of the largest. It is not un-
common on any one flush of H. hrasilicnsis to find leaflets ranging in length from

5 cm. to 20 cm. or longer.

Shape at least is more constant for the individual than size. In general, H.
viicrophylla has a narrow lanceolate leaflet, while H. gniaucnshy particularly

var. marg/fiafa, tends toward an obovate outline. H. pauciflora^ brasilicnsh and
rigiJifolia usually have broadly lanceolate leaflets.

Leaflet Tips: Leaflet tips are a good character to use in distinguishing H. rigid-
ifolia and H, pauciflora, which from sterile material alone frequently may be hard
to tell apart. Rather voluminous material of H, pauciflora has failed to reveal a
case In which the midvein extends all the way to the end of the blade tip. Further-
more, the end of the midvein always appears to be excised or calloused, and on
mature leaflets produces a "socket" or glandular efi^ect (text-fig. 3). This char-

^®It should be stated tliat care must be exercised, particularly in the study of herbarium speci-
mens, in distinguishing correctly between pubescence and ilmulated pubescence caused by hyphae
and small fruiting bodies of certain minute fungi which protrude from the midveins. Apparently
the veins form an excellent medium of fungus growth during the drying stage of the specimens.

1947]

SEIBERT HEVEA IN PERU

279

Text-fig. 3. Leaflet tip on which the
midvein terminates short of the blade tip
and is glandular-calloused. Typified by H.
pauci flora.

Text-fig. 4. Leaflet tip on which the
midvein extends to the end of the blade

tip and is not calloused.
rigidifoUa,

Typified by H,

acteristic is unique in the genus. It is In strict contrast to H. rigidifolia where the
midvein extends to the end of the long, narrowly acuminate blade tip, producing a
cuspidate eflfect (text-fig. 4). In other species the midvein also extends to the end
of the blade tip or even slightly beyond, as occasionally noted in H. micropJyylla
and H. hrasiliensis.

Texture: Hard coriaceous leaflets are found in H. rigidifolia. In H. paticiflora
the leaflets appear to mature so slowly that a coriaceous condition can be expected
only in the fully mature flush of perhaps several months of age. A coriaceous to
subcoriaceous texture is the general rule for H, guianensis var. marginata^ and to
lesser extents in H. guianensis^ guianensis var. Ititea, nitiday and Benthamianaj es-
pecially when growing under poor soil and climatic conditions.

H, brasilieusisy microphylla and Spruceana are typified by membranaceous leaf-
lets with considerable individual differences that are distinguishable as good clonal

variations.

Margins: Revolute margins are the rule in ff. rigidifolia. They are noticeable
on the mature leaflets of H. paiiciflora and form the main character in segregating
the variety viarginafa from H. guianensis.

Inflorescence,

The inflorescence arises from the terminal (occasionally axillary) bud regions
of the young stems. It is composed of numerous panicles arising from the axils
of the upper scales of the short-shoot. In a majority of the species panicles also
arise from the axils of the lower scale leaves of the young flush and frequently
from the axils of many of the flush leaves themselves. Under ideal growing condi-

[Vol. 34

280 ANNALS OF THE MISSOURI BOTANICAL GARDEN

tions trees will commence flowering when between three to five years old. Under
forest conditions slow, competitive growth may prevent flowering until the tree
IS twenty to twenty-five years old.

The flowers of the inflorescence always are monoecious. The terminal flower of
the primary and stronger secondary axes of the panicles is pistillate, all others being
staminate. Flowering occurs normally once a year, following a dormant or winter-
ing season during which some of the species defoliate completely, H. hrasHicnsh
being an example. Other species, however, tend to flower without defoliation.
Still others (inly partially defoliate or gradually defoliate during flowering or after
subsequent flush development.

Thus far I have been able to make field studies on only two species with ref-
erence to inflorescence habits and their interesting relation to wintering (defolia-
tion or non-defoliation), short-shoots, and the development of the flush. This

inter-relationship between the various species is very strongly suggested by meager
and incomplete herbarium collections from the Amazon valley, the assumption
previously having been that these features are the same in all species. It does appear
that a complete comprehension of the genus w^ill be aided materially by a full
understanding of the inflorescence and its habits for the different species; a sum-

mary of which is suggested under the discussion of leaf persistence.

Buds and Flowers: Sketches of both male and female buds and flowers of each
important taxonomic entity of the genus are presented in pis. 35, 36, 37, 38,
and 39.

The flowers, though small and difficult to dissect, are relatively easy to study
after one has mastered the fundamentals of their morphology. With the key here
proposed, microscopic dissection of the flower rarely seems necessary, and then only

for critical study. Usually sufficient characters may be seen from the superficial
bud and flower aspects to make more technical observations unnecessary.

Tlie staminate flow^er has a short pedicel subtended by a small caducous bract,
in the axil of which usually an abortive bud is found. After anthesis the flower
abscisses from the peduncle. The perianth Is composed of a short calyx tube and
five valvatc lobes. Petals are absent but a disk is present which may be represented
merely as a slight swelling or enlargement of the base of the staminal column, by

the presence of five gland-like protuberances, or by five relatively large petal-like
acute lobes which seldom reach farther than to the lowest anthers. The anthers
range from five to ten in number. They are sessile and attached in one regular
or irregular, two irregular, or two regular whorls on a staminal column. The
staminal column represents the fused filaments in conjunction with the rudi-
mentary pistil, the tip of which extends beyond the anther whorls and may or may
not be slightly lobed or divided giving a stigma-like impression.

The pistillate flower has a somewhat longer pedicel than that of the staminate
and likewise abscisses from the peduncle unless it be successfully fertilized. The
pedicel dilates into a torus which is conspicuous only in the case of H. niicrophylla.

1947]

SEIBERT HEVEA IN PERU 281

Here again the perianth is composed of calyx tube and five valvate lobes. Contrary
to the habit of the staminate flower, the tube abscisses after anthcsis at its point of
junction with the torus. Neither petals nor stamens are present but a conspicuous
or inconspicuous disk is found. In some cases It is possible to distinguish between
petal-like and stamen-like disk lobes. The pistil is present in the form of a tri-
locular,' tricarpcllary ovary, each cell of which has but one ovule. The three
stigma lobes may be entire or somewhat lobed and are sessile, though in rare cases
they may be very shortly stipitate.

Color: With the exception of H. Spruceaua, which normally has a dark reddish
purple to purplish brown calyx tube, all the species have flowers which are uni-
formly cream-yellow to brownish yellow in color.

Pubescence: Variation in floral pubescence appears to be of some significance
in speciation. In general, the species of Hevea have their buds and flowers covered
with very short whitish hairs. In H. Bcntbamiana these hairs not only are longer
than in other species but have a distinctly reddish color. Within H. gniancnsis and
its variety lufca there frequently is a reddish tan cast to the short floral pubescence,
while the color of the hairs on the peduncles and pedicels tends to be more reddish
in tinge. At the point of contact between the peduncle and pedicel there appears
to be a narrow band of longer, more dense hairs which is noticeable in H. Bentham-
iana, guiaiiensis and varieties, and rigidifoUa.

The staminate buds and flowers in all species appear to be rather uniformly
pubescent without. The pistillate flowers of four species, however, show remark-
able variation. H. pauciflora, having pubescent lobes without, is distinctly
glabrous below the lobes on the tube and well on the pedicel. H. micropbylla be-
comes conspicuously less pubescent on the tube below the lobes and has a glabrous
torus and pedicel. The lower-center portion of the lobes of H. nitida become
glabrous as well as the tube. H. brasilicnsis is slightly less pubescent on the tube

than on the lobes. All other species appear to have uniformly pubescent lobes and
tube.

The inner surface of the calyx lobes apparently is somewhat pubescent in all
the species, the pubescence being conspicuously longer in H. pmuciflora.

The ovary, although characteristically short-pubescent or somewhat silky

■fl

-'f

Calyx Lobe Acumination and Lobe Tips: Bud shape, particularly as to whether
it be obtuse or acuminate, has been frequently used as a taxonomic character with
special reference to the staminate flower (Ruber, 1906). The illustrations (plates
40 and 41) will show the importance of the degree of acumination with ref-
erence to the calyx lobes of the mature flower as well as the bud. It should be
noted that the degree of bud acumination is not always indicative of the degree of
mature calyx lobe acumination; furthermore, that the degree of bud and lobe
acumination is not always the same for both the male and female flowers.

282

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Two other characters, neither of which appears to

have been noted previously, are of importance when
taken with the consideration of acumination: (1) Shape
of the lobe tips (pL 43, figs, 1-2), i, e., whether they
be acute and pubescent as shown in text-fig. 5, or
whether they be blunt and glabrous (calloused) as
shown in text-fig. 6. Calloused calyx lobe tips of both

P'

//'

brasil

5

?nsis, (2) The absence or degree of contortion
of the bud tip and resulting mature lobes. Contortion
In its most pronounced degree is found in both the
staminate and pistillate flowers of H. rigidifolia (pi. 43,
fig. 3) and to a lesser degree in H. hrasilknsh and H.
micropbyUa (pis. 38-39),

Disk: The disk, both in staminate and pistillate
flowers, is a rather conspicuous feature of the genus
Hevea. It Is of taxonomic significance when Its dcvel-
opment in the various species Is compared (sec pi. 42).
The disk of the staminate flowers appears to represent
petals. When lobes can be distinguished their position
alternates both with that of the calyx lobes and the
anthers of the lower whorl. In the pistillate flower a
study of the disk lobes, where conspicuously present,
Indicates the disk to represent both petals and stamens.
Frequently the petal-like lobes alternate with structures
which appear to be much-reduced stamens, both fila-
ments and anthers being distinguished. Most frequently
the disk lobes are attached to the base of the ovary, but

cases have been noted in dissections where they are slightly adherent to the calyx

at the point of its abscission.

Like other characters, the disk developments are Inclined to show a great deal
of transition from one extreme to the other from one to the next species. As in
the case of acumination, the same degree of development does not always carry
through in a parallel fashion between the staminate and pistillate disk of the same
species.

Torus: Morphologically all of the species of Hcira can be considered to have
a torus in the pistillate flower (pis. 41-42). It is only in H. nticrophylla
that it is so pronounced (pi. 43, fig. 2) that it may be used immediately
In distinguishing that species from all others. Though not a conspicuous character
In the genus, It Is rather well developed also in H. pauciflora.

6

Text-fig. 5. The normal
calyx lobe wlilch is acute
and pubescent to the tip.

*

Text-fig. 6. Calyx lobe
which is bluntly acute, cal-
loused and glabrous at the
tip. Found conspicuously in
H. p and flora and hL ntiida,
and less well developed in
rr. hrasilicnsis.

\

h

1947]

SEIBERT HEVEA IN PERU 283

Anthers: Each stamen consists of a bilocular, longitudinally dehiscent anther,
sessile or nearly sessile and attached directly to the staminal column (see upper
figures, pi. 42). The column is composed of fused filaments as well as the rudi-
mentary pistil. This accounts for the columnar portion extending above the
anther whorl or whorls and explains why this stamina! column tip frequently is
found to be so variable within the species, i, e,y it may be acute, blunt, long, short,
entire, bi-lobed or even tri-lobed.

In taxonomic studies of Hevea considerable stress has been paid to anther num-
ber, on the basis of which two sections, Euhevca Muell.-Arg. and Bhiphonia
(Baill.) MuelL-Arg., were erected by Mueller-Argoviensis (1865). Through a
study of more complete material I have arrived at the opinion that exact number
of anthers is of little taxonomic significance within the limits of certain tendencies.
These tendencies are towards 5 anthers in one whorl as typified by H, guiancnsis
and its variety marginata; towards 6-8 or 9 anthers in two irregular whorls in
which the upper Is the more irregular, as found in H. gniancnsis var. hitcUy
Bentbamianay and occasionally in Spruceatiay pauciflora, and rigidifolia; and towards
10 anthers in two regular whorls of which H. hrasiliensis, nitida and micro pbylla
might be considered typical. However, the number of anthers may vary within
the species and between flowers on the same tree. As is to be expected, if there be
two whorls, the anthers of the upper alternate in position with those of the lower.

An anther character of perhaps considerable significance, and one not previously
stressed, is the size. Anthers within the genus appear to fall into two definite size
groups, those which approach 1 mm. in length and those of about 0.5 mm, in
length, with no definite recognizable gradation. The nearest approach is in
H. gniancnsis var. lutca^ where both sizes frequently are present in the same
staminatc flower, but with Kttle size gradation one into the other. So far
as I am able to determine, the large anthers are associated only with the H.
gniancnsis complex. They appear quite uniformly in the predominant number
of 5 in one whorl in both H. gniancnsis and its variety marginata. In all other
species the anther size approaches 0.5 mm. with the exception of H. gniancnsis var.
In^ca where both sizes may sometimes be present (usually only one or two of the
large size). However, it must be remembered that this Is an extremely variable
and transitional variety in which frequently none of the larger anthers are present
on the irregular whorls, but if present are on the lower whorl.

Frnit.

The fruits arc disposed terminally on the inflorescence branches but the ratio
of developed fruits to the total number of female flowers is perhaps only one to
ten or twenty. So far as I know, the fruit of all species tends to hang down due
to its weight on the peduncle. Because of the comparatively long inflorescence in
both H. brasilicnsis and H. pattcifloray and because the terminal pistillate flower
tends to be the one which most frequently develops fruit, the peduncles of
these two species seem to be longer than those of the other species, and of
course droop more. The departure of the peduncle from the branches of H.

t

t

[Vol. 34

284 ANNALS OF THE MISSOURI BOTANICAL GARDEN

panciflora is conspicuous in that it is at a right angle, after which rather sudden
drooping occurs. In H. byasilicfish and other species, the peduncle departs from
the branch at considerably less than a right angle and droops in a more gradual arc.
The fruit is quite uniform throughout the genus, normally being a 3-carpellary
dehiscent capsule of relatively large size. The carpels are bivalvcd, composed of a
coriaceous pericarp and a woody cndocarp which varies in thickness with the
species. Each carpel normally contains one seed. Tn some trees, particularly
noted in plantings of //, hrasilicnsis, the capsules are occasionally 4-, or even 5-,
carpelled, all carpels containing a normal, viable seed. Perhaps through abnormal
or imperfect fertilization^ capsules sometimes mature normally in every respect
except for the seeds, which may be rudimentary or abortive in one or more of the

carpels.

Three distinct variations in the capsule of Ilcvca occur: (1) The most prev-
alent type is subglobosc in outline v/ith more or less cmarginate, mucronate tip.
In cross-section, it is distinctly 3-lobcd. Tn dehiscence this type is violently ex-
plosive, both the seeds and capsule parts being thrown as much as 15—20 meters.
All that remains on the tree is the peduncle and placenta. There appears to be
considerable variation in the thickness of the woody capsule walls or endocarp
between species with this type of fruit. //. giiiancmh and its varieties, H, brasil-
icfisis, fiifiJa^ and panciflora tend to have a thick, woody endocarp in which the
valves retain their original shape with little noticeable contortion. On the other

T

hand, II. BoithawiUia and II. rlgidifolia have a relatively thin, woody endocarp,
the valves of which show very noticeable contortion at dehiscence, especially H.
rigrdiflora. (2) The fruit of H. microphyUa is unique in that its shape is pyra-
midal, tending toward an acute apex, and is noticeably keeled. The carpel walls
are thin and leathery, being composed of a coriaceous pericarp and an almost paper-
thin, woody endocarp. Dehiscence is not explosive; rather, the valves appear to
open slowly, greatly contorting and allowing the seeds merely to fall. The valves
appear to be persistent to the receptacle for a considerable time after dehiscence;
the torus of this species Is very conspicuous at the base of the fruit. (3) The fruit
of H. Spruccaua likewise Is unique for the genus in that it is much larger, ellipsoid
to subovoid, obtuse at the tip, and round in cross-section. The carpel walls arc
composed of the coriaceous pericarp and a very thick, woody endocarp. Although
dehiscence Is somewhat explosive, there Is little contortion of the valves, and they
are persistent to the receptacle for some time. The seeds are not thrown far.

J

SccJ^: Hevea seeds^*^ are similar in shape and color patterns to those of the
castor bean, Kicinusy but they are, with few exceptions, very much larger and they
always lack the persistent caruncle in mature condition. Reserve material has a
very high percentage of oil. Length of viability is normally only a few weeks.
Exposure to full sun and drought will shorten this period, while packing in a cool,

^"'XVilJ pigs, Peccary, of which two common types arc known locally in Peru as huangana and
sajirto, arc extremely fond of the seeds. Deer, vevado^ appear to relish the young seedlings coming
up under tlio trees. Wild game is particularly abundant during the Jlcvca seed season, augmented
by many members of the Cat family whicli prey on botli the wild pigs and deer.

1947]

SEIBERT HEVEA IN PERU 285

slightly moist medium can extend it to two or three months. Although abundant
moisture and deep shade is essential for seed germination, the young seedlings soon
perish if not exposed to rather full sunlight. Various types of seed twinning have

been reported but this seems to be rare.

Much consideration has been given to the color patterns of the outer seed sur-
face of H. brasilicnsis (La Rue, 1919) as a means of identifying individual trees
within the species. Color patterns are relatively stable for the Individual. It
would appear that coloration is of little use in speciation, with the possible excep-
tion of H, Bcnthaniiana, where the background is light in comparison with
the more tannish background of other species. The brown mottling, too,
appears to be somewhat clearer and more brilliant than in other species.

Over-all seed size is apt not to be constant, except in H. Spmceana^ which is
longer than any other species. In this connection the length/thickness ratio is the
important factor, seeds of H. Spruceana being at least twice as long as thick while
in all other species the length is less than twice the thickness.

Shape of the seed, especially in cross-section, tends to be relatively constant for

4

the species and can give considerable aid in speclation. The following generaliza-
tions may be made regarding shape and other seed tendencies as an aid to species

determination:

H. Spruceana. — Several unique characters are present in the seeds. They are at least twice as
long as thick, in contrast to those of all other species whose length is less than twice their
f thickness. The longitudinal-section through the dorslvcntral plane shows a slight curved
(approaching rcniform) condition, suggesting a similarity in both shape and size to the
kernel of a Brazil-nut. In cross-section, the seed is sharply angled in that the compressed
flattened ventral and the two lateral angles formed at the junction between dorsal and
ventral surface arc very prominent. The dorsal surface tends to be uniformly rounded

with a rather faint dorsal angle.

H. inicrophylla. — The seed is unique. It is triangular-ovate in outline, with the smaller end
the micropylar end.

H. brasilicnsis. — Seeds normally are ellipsoid in outline, somewh:it compressed on the ventral
surface. In cross-section, the dorsal surface tends to be uniformly rounded, as does the
more flattened ventral surface. If any angulation is present it is seen at the junction of
the two surfaces. It appears that the presence of germ-plasm from other species will
produce noticeable effects on the angulation or generally rounded condition.

H. Bent ha mi an a. — In general, this species has smaller seeds than H. hrasiliensis but, like In that

species, they tend to be quite rounded. In color, they have the noticeably whitish back-
ground and brilliant brown spots in contrast to the tannish background of other species,
yere again hybridization, or at least presence of germ-plasm from other species, seems to
produce pronounced effects on the shape.
H. gulancnsh and its varieties. — The seeds tend to show a distinctive 4-angled cross-section In
which the dorsal sides are longer than the ventral sides, producing a kite-shaped effect.
Each side of both the dorsal and ventral surfaces shows a shallow concavity running

throughout most of the seed length.

H. pauclflora and //. rig'nllfolia. — The seeds show a decided hexagonal cross-section In which
two sides are on the ventral surface and four sides make up the dorsal surface. Each of
the dorsal sides tends to have a shallow concavity running nearly the entire length of the
seed. On the ventral surface the two sides form prominent lateral concavities on the
upper two-thirds of the seed but tlils Is replaced by a central concavity on the lower third.

H. nifida. — Seeds of this species appear to have the upper half of the ventral surface composed
of two lateral concavities, replaced by one larger, central concavity on the lower half.
On the dorsal surface there are two shallow concavities. In the cross-section of the
micropilar end a 4-anglcd, kite-shaped effect is produced which is not noted In. the basal
cross-section. Not enough seeds of this species have been seen on which to base an

accurate description.

[Vol. 34

286 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Pollination

I

Very little is known regarding the pollination of Hcvca In nature. Since the
pollen grains arc rather sticky, wind can be eliminated as a factor. Ramacr (1935),
in certain hand-pollination experiments, found that cross-pollinations were con-
siderably more successful than self-pollinations. Workers in Hevea consider that
cross-pollination Is the rule, but cases are known where a tree isolated from other
trees by many miles repeatedly set fertile fruit year after year. There are other
cases, particularly in isolated mono-clonal plantings, where certain clones will not
set fruit unless planted in close proximity to other clones. The fact that Hevea
has set fruit wherever it is planted and comes into flower might indicate that no
highly specialized insect adaptation Is necessary for its pollination. The terminal
position of the pistillate flowers on the inflorescence and its major branch axes
could indicate that cross-pollination by flying insects might be favored.

One never can find a jungle Hevea tree in bloom that is not covered with ants,
usually of varied species, Including Icaf-cuttcrs, occasionally carrying the flowers to
their nests. Though ants cannot account for cross-pollination, they may be an
important factor in selfing. Maas (1919) states that members of the Nitidulidac,
Phlacridae, small Curculionidae, fly species and small bees, have been seen on Hevea
flowers in the Far East. I have observed the common Honeybee to frequent the
flowers of Hevea growing at the Plant Introduction Station, Coconut Grove,
Florida, but little, if, any, reference has been made regarding the possible pollinators
In the natural habitats.

Since we find that natural hybridization is taking place among the species in
the Amazon valley, it is imperative to know something of the conditions of cross-
pollination and what insects might account for it. H. brauliensis, at least, tends
to open its flowers during the latter half of the afternoon but pollen is said to be
fertile only for perhaps a day, losing its viability rather rapidly in dry sunny
weather (Maas, 1919). The pistillate flower apparently is receptive for two to
three days. In view of this, both day- and night-flying insects must be taken into
consideration. It would appear that conditions for successful cross-pollination,
where species arc separated by distances of a mile or less, would be best at night.
Within the Peruvian range of Hevea, species of stingless bees, both Melipotia and
Trigona, have been seen in abundance around and on flowers of Hevea. Meliponaj
especially, remains on the flowers even after the felling of a branch. Little seems
to be known regarding the distances traveled by members of these genera, but
Michcncr (1946) has made such observations In Panama. Tt would appear that
these stingless bees tend to concentrate on a few trees and work within compara-
tively limited ranges, accounting mostly for selfing. Dr. Herbert F. Schwartz, of
the American Museum of Natural History (correspondence Feb. 5, 1947), has
suggested that members of the Megalopta bees (Halictidae) might account in part
for the night-flying insects visiting Hevea. He has identified a specimen of this
group taken from a Hevea flower as a species of Augochlora. Besides the bees,
^wasps also arc frequent visitors of Hevea trees In flower, the nests often being
encountered.

1947]

SEIBERT HEVEA IN PERU 287

At the present time, there appears to be more positive evidence that self-
ing may be the rule under jungle conditions. Here trees, even of the same
species, may be separated by a relatively few meters or up to a kilometer or so.
Nevertheless, it must be pointed out that we scarcely know anything of the upper-
story fauna in tall forest trees even during the day, much less at night.

Cytological Summary

According to Baldwin (1947), all the species of Hevea thus far studied cyto-
logically have a normal 2n chromosome count of 36, confirming the work of
Ramacr (1935), Paddock (1943) and Perry (1943). Baldwin found one indi-
vidual of H. paticiflora to have 18, and another of H. guianensis var. marginata
with 54. Perry (1942) states that '*all (Hevea) species studied are tetraploid."
Ramaer (1935) considered 18 chromosomes as the basic 2n number for Hevea,
Baldwin believes the normal 36 2n chromosome number likely to be tetraploid, and
that the individual with a 54 2n number probably is hexaploid,

Intraspecific Variation

Some excellent research has been carried on at Far Eastern experiment stations
concerning morphological variations of trees of H. hrasiliensh. No part of the
plant's anatomy seems to be without its range of intraspecific variation (Frey-
Wyssling, 1931, 1933; Frey-WyssKng, Heusser and Ostendorf, 1932; Assoc. Cent.
Exp. Sta., 1939). Furthermore, Bobiloff (1931) has brought out that there are
physiological differences between clones of the same species with regard to latex
color reactions through the addition of calcium chloride. More recent work is deal-
ing with variations in individual susceptibifity to diseases, being carried on by

Langford (1945).

Casual observations sometimes lead one to feel that there are more differences

between two trees of the same species than between two species. This is especially

true when the observer is highly familiar with the details of many clones of H.

brasilknsis but not with the other species.

My own preliminary observations, as well as those of previous workers, indicate
that there is a tendency towards opposing extremes in any one morphological
variation; but since the extremes are probably based genetically on multiple factor
differences, there are always intermediates present. It also has been observed that
though one or many characters may be intermediate in nature, there are other
characters which more closely approach one or the other extreme of the total ex-
pected variation. A list of variable intraspecific characters can be of material aid
in tabulating ranges of variation within species. Large-scale scoring of intra-
specific variations, in comparing species, would show amounts of parallel variation
ranges between species; and may eventually be used, when more material is avail-
able, in showing that a number of species grade one Into the other.

Speaking from a more practical standpoint, intraspecific variations are of great
value in distinguishing Hevea clones. The proper identification of young, sterile.

288 ANNALS OF THE MISSOURI

[Vol. 34

bud-grafted material or clones in the field planting, budwood garden, and experi-
mental plot is of utmost importance since mixtures due to lack of careful super-
vision and other unavoidable factors frequently occur and, if not detected, can
lead to high production losses.

For the relatively few clones in commercial use in the Far East, the written
descriptions amply have served their purpose, even without the actual devising of
keys for their identification. The cooperative Hcvea Plantation Improvement
Program, being carried on by the U. S. Department of Agriculture in cooperation
with many of the Latin American countries and commercial rubber companies, has
resulted in the amassing and distribution of hundreds of Hciea clones selected
from superior jungle trees. It also has resulted in the selection of thousands of
nursery seedlings from various seed progenies collected throughout the Amazon
valley, and is resulting in thousands more of hand-pollinated crosses, all of which
arc undergoing experimental tests. Once these individuals are proved resistant to
strains of the South American Leaf BHght, Bothidclla UUi, both under natural
and artificial innoculatlon, they must be distributed to various experiment stations
and cooperators for field trials. With each move and each distribution, the chances
for an error in labelling increases many-fold.

A comprehensive study of individual morphological variations and their con-
sistency for use in the accurate identification of many clones will In Itself be a
long-time work. It will be one In which frequent revisions and changes will have
to be made to keep abreast of the advancing development and introduction of
more and more proved clones into commercial use. The work at first will be
largely devoted to the use of vegetative characters to be found on the young
budded plants. From this, it will advance to include characters of the mature
tree, as bark, latex, trunk and branching. Finally, it must include variations in
inflorescence, flower, fruit and seed, which, though not so important for the
planter, are necessary for the geneticist and plant breeder.

In making up preliminary lists of intraspeclfic variable characters, I have, of
course, drawn heavily from the previous work on clone characters; but in working
with and selecting jungle material and subsequently studying it as young buddings
m nurseries, many additional variations have come to my attention. Such would
be expected since the new material is coming from widely separated localities and

from the jungle instead of from the original stock on which the Far Eastern
industry is based. Certainly many other contrasting characters are yet to be
observed, and one can readily see that all possible combinations of the many char-
acter variations would lead to astronomical figures. There is little reason to doubt
that keys could be devised to take care of any number of clones desired to be

identified.

The most practical means of large-scale clone identification would not neces-
sarily have to be in the form of a key. The punch-card system might better be
adopted In which only combinations of strongly contrasting characters be used,
disregarding all Intermediates.

1947]

SEIBERT HEVEA IN PERU 289

The following intraspecific characters are listed with their contrasting condi-
tions which have been found to exist within H, brasllicn^is. Since the greatest
practical need for intraspecific variations lies in the differences on young budded
clones, the characters found on young, sterile plants are stressed. I have had little
opportunity for practical observation of characters of this sort from large progeny
numbers of species other than H. hrasilienstSy or, for that matter, of many of the
known interspecific hybrids. Yet from some such observations made on H.
guianensis var. Jutea there is reason to believe that parallel conditions exist in other
species. With little revision, these lists of intraspecific variations might well apply
to the hybrids as well as the species, after the material in question has been given
its proper specific rank.

Trunk (Stem)

growth: strong or weak

form: erect or leaning

BASAL cross-section: round or fluted

BARK

green: with or without bloom

BRO^'N-GREEN

lenticels: conspicuous or inconspicuous

color: whitish or as cork
first cork: on flush or on interflush

in streaks or in spots

BROWN

color: grayish, tannish, reddish or brownish
lenticels: conspicuous or inconspicuous
size: large or small

color: whitish or as cork
cork: smooth or rough

GROW^th crack interval: fine or coarse

LATEX

color: white, cream or yellow
consistency: watery or thick

BUDS""

terminal

BUD SCALES: few or many
shape: linear or deltoid
axillary: depressed or exscrted; early-sprouting or late-sproutlng
LEAF SCAR: protruding or not protruding
margin: protruding or not protruding

stipules: conspicuous or inconspicuous

early caducous or late caducous
SHAPE: linear or deltoid
typical flush

SHAPE: asymmetrical or symmetrical

half-globular or globular segment
conical or truncate-conical
dimensions: large or small

broad or narrow

tall or short
density: sparse or dense
stories: continuous or separated

Leaves —

petioles: longer or shorter than blades

direction: downward, horizontal, or upward
form: straight, arcuate, inverse-arcuate, or sigmoid

Bud characters could be substantially augmented when considering other species in addition to
H. brasiliensis.

[Vol. 34

290 ANNALS OF THE MISSOURI BOTANICAL GARDEN

base: normal or much swollen

subcrized or not suterizcd
dimensions: long or shore

thick or thin

ANGLE BETWEEN EACH: bfoad (above 90 ) or narrow (below 70 )
direction: downward, horizontal, or upward
form: clawed or not clawed
SIZE: long or short

LEAF BLADES

color: yellowish green, light green, or dark green

PUBESCENCE"^: absent or present on veins

UPPER SURFACE LUSTRE: glossy or duU

LOWER SURFACE LUSTRE! subconcolorous or not subconcolorous

texture: membranaceous or coriaceous

SHAPE: lanceolate or oblanceolate

ovate or obovate

diamond or rhombic

suborbicular (orbicular leaflets have not yet been noted)

: large or small

margin: plane, wavy or crisped

revolute or not rcvolute
MIDVEIN: terminating short of blade tip, extending to end of blade tip, or

extending beyond blade tip

lateral veins: continuous, forked or branched

tip: obtuse: attenuate (acuminate) or short (not acuminate)

acute: acuminate or not acuminate

base: acuminate or not acuminate

obtuse or acute

long-section profile: flat or convex

CROSi^-SECTloN PROFILE: flat, V-shapcd or boat-shaped

position to plane OF petiole: erect, semi-erect, parallel, declined or horizontal

POSITION TO EACH OTHER: apart, touching or overlapping

Miscellaneous Abnormalities

leaves: more than 3 leaflets

less than 3 leaflets
leaflets concrescent

The above intraspecific variations are but a few wlien considering that these
arc found in the young plant. An over-all consideration of the mature trees
would include not only the above but would be augmented by many more, some
of which have been mentioned under the morphological discussions. A few of
the more pronounced intraspecific variations of mature trees may be found in the

nature of the corky bark surface, Its color, and its method of exfoliation. The

color of the phloem Is highly cKaracteristic, and five color divisions already have
been mentioned. The nature of the branching also can be placed in various classes.
Besides the study of variations in seed-color pattern, very little has been done re-
garding intraspecific variations in the inflorescence, flowers, floral pubescence, fruit
and seed shapes, all of which show Innumerable variable contrasting characters.
However, as yet they are not sufficiently well studied to be presented in this paper.

^^In considerations including other than //, brasiJicfJsh, pubescence characters can be greatly
augmented and highly significant.

/

1947]

SEIBERT HEVEA IN PERU 291

THE

A. Leafy shoots (flushes) alternating with relatively elongate, caducous-
scaly short-shoots*^ (interflush short-shoots"^).

1, Leaflets distinctly erect to slightly horizontal"^; staminate flowers
without disk lobes; seeds with 4-angled (kite-shaped) cross-section.

a, Staminate buds broadly obtuse to rounded (may be somewhat
acute in var. marginata) ; anthers about 1 mm. long, normally

5 in one whorl.

a. Leaflets membranaceous to subcoriaceous, not revolute 1. H. guianensis^^

/3. Leaflets coriaceous, revolute X H. guianensis

var. Tnarginata

b. Staminate buds acute to acuminate; anthers about 0.5 mm. long
(one or more may approach 1 mm.), normally 5—7 in two ir-
regular whorls .........la. H. guianensis

var. LUTEA

2. Leaflets distinctly horizontal to rccllnate; staminate flowers with
disk lobes; seeds variously shaped.

a. Leaflets concolorous, not lepidote on the lower surface; leaf
flush tending to defoliate before appearance of inflorescence, at
least before appearance of new flush; staminate disk lobes very
conspicuous, attaining lower whorl of anthers 2, H. nitida

b. Leaflets not concolorous, densely whitish-Iepidote on the lower
surface; leaf flush usually persistently leafy until after inflores-
cence maturation and appearance of new flush; staminate disk
lobes inconspicuous, short.

a. Mature leaflets coriaceous, glabrous below; flowers yellowish;
diameter of staminate flowers normal for the genus (about
2.5-3 mm.) ; pistillate disk lobes conspicuous, long-acute;
fruit with 3-lobed cross-section; seeds hexagonal in cross-
section, length less than twice the thickness.
L Leaflets conspicuously revolute, hard-coriaceous, cuspidate,
midvcin extending to end of blade tip, not callosc-tipped;
staminate buds long-acuminate, conspicuously contorted;
calyx lobes acute-tipped, pubescent to tip, not calloused.... //. rlgidifolia

IL Leaflets not conspicuously revolute, membranaceous when
young, becoming coriaceous with slow maturation, not
cuspidate, with midvein terminating before reaching blade
tip, having a glandular, calloused tip; staminate buds
obtuse, never contorted; calyx lobes blunt-tipped, the tips
calloused, glabrous.. 3. H. pauciflora

/3. Mature leaflets membranaceous, usually noticeably whitish-
pilose over entire lower surface; flowers reddish to brownish
purple, at least the calyx tube; staminate flowers largest in
diameter for the genus (about 4.5 mm.); pistillate disk lobes
inconspicuous; fruit with round cross-section; seeds vcntrally
compressed-angular, otherwise, rounded in cross-section, length
at least twice the thickness.... H. Spruceana

B, Leafy shoots (flushes) alternating with narrow rings of bud-scale
scars"^ (interflush rings) ^^.

^^Refer to tcxt-fig. 1 and pi. 32, fig. 2.

"■'* Attention Is called to cases where hybridization between species of this group and those in
which the short-shoot is not conspicuous has produced rather rare individuals, though otherwise
characteristic for the most part of one of the short-shoored species which may not have this feature
In evidence.

"^Position of leaflet frequently of little use in the herbarium.

^^Specles printed in capitals are those represented In Peru; those printed In lower case have not

been found to occur in Peru.

^^Rcfer to tcxt-fig. 2, and pi. 32, fig. 1.

^^ Attention Is called to cases where hybridization has occurred with conspicuously short-shooted
species, the short-shoot being usually conspicuous in the hybrid. As is particularly evident in H.
Bcntha7nlana hybrids, the specimen may superficially resemble that species in most characters except
a conspicuous short-shoot.

[Vol. 3-4

292 ANNALS OF THE MISSOURI BOTANICAL GARDEN

1. Leaflets usually rcddish-pilosc over entire lower surface; antlicrs
normally less than 10 in two irregular whorls; seeds showing dis-i

tinct whitish background under brilliant, clear brown spots 4. H. Benthamiana

2. Leaflets glabrous below; anthers normally 10, in two regular whorls;
seeds showing distinctly tan background under usually dull brown

Spots.

a. Leaflets relatively small (5—12 cm. long); flowers very long

(staminatc about 7 mm.; pistillate about 11 mm.) for the
genus, the pistillate with an enlarged torus; calyx lobes acutely
acuminate, pubescent to tip, not calloused; both staminate and
pistillate disk lobes small but conspicuous; ovary glabrous;
capsule pyramidal, acute; valves thin, leathery; seeds triangular-
ovate in longitudinal section H. mlcro^hyUa

b. Leaflets relatively large (5-30 cm. or more long); flowers of
median length (staminatc 5 mm.; pistillate 8 mm.) for the genus,
the pistillate without an enlarged torus; calyx lobes bluntly
acuminate, with small calloused, glabrous tips; both staminatc
and pistillate disk lobes very small and inconspicuous; ovary
silky-pubescent; capsule subglobose, emarglnate; valves thick,

woody; seeds oval in longitudinal section 5. H. brasiliensis

The Peruvian Species

1. Hevea guianensis Aubl. Hist. PL Guiin:i Fr. 2:871. (/?/. 335 as H. pcnniam,
sphalm). 1775.

Jafrop/ja (?) elastica L. Sp. Pi". Suppl. 422. 1781.
Caoutchoua elastica (L.) H. F. Gmel. Syst. 1007. 1791.
Slpbonia Calmcbu Rich, ex Willd. Sp. PI. 4:567. 1805.
Sipbonia clasfica (L.) Pcrs. Syn. Pl. 2:588. 1807.

Sip/jofih ^luancfuis (Aubl.) Juss. Euphorb. Gen. 40 {pl 12, fig. 38a as Siphonta clasfica,

sphalm). 1824.
Hcica nigra Ule, in Engl. Bot. Jahrb. 35:667. 1905.
Hcvca caiicho Posada, Estudlos Cient. 212. 1909, nom. nud.
Uevca coUina Hubcr, in Bol. Mus. Gocldi 5:249. 1909.

Hcvca guravcfisis var. collina (Iluber) Duckc, in Archlv. Jard. Bot. Rio de Janeiro 4:109.
1925.

Heica guianensis var. cuneafa (Ruber) Duckc, 1. c. 6:51. 1933, in part,
Hevea giUanensis ssp. occidcnfaJis Duckc, in Archiv. Inst. Biol. Veg. Rio de Janeiro
2:229. 1935.

Hevea guianensis var. occidcnfaJis Ducke, 1. c. 193 5.
Hevea guianensis ssp. fypica Duckc, I. c. 227. 193 5.

Medium-sized to large tree to 40 m. tall; trunk cylindrical; brancbcs somewbat
rcddisb; sbort-sboots very conspicuous, of somewhat greater diameter than the
long-shoots; bud scales very numerous, linear, about 6 mm. long, early caducous.
Leaves partly persistent until after appear;ince of inflorescence; mature leaflets
erect, membranaceous to thinly coriaceous, usually obovate with short acuminate
tip, the pubescence of sparse hairs on lower surface along midvein, tannish or
somewhat reddish, the scales of the lower surface roundish, the midvein continuous
to end of blade tip, not calloused. Flowers yellowish, staminatc buds rounded to
obtuse, not contorted, the short pubescence whitish-tan to reddish, rather uni-
formly distributed except at abscission region of pedicel where the hairs are more
dense and longer; staminatc flowers about 3.5 mm. long and 3 mm. broad, the calyx
lobes acute and acute-tipped, not calloused and not acuminate, the disk inconspicu-

\

T

a

1947]

SEIBERT — HEVEA IN PERU 293

ous, represented by a slight flaring of the staminal column base, the anthers normal-
ly 5, approximately 1 mm. In length, in one whorl; pistillate buds obtuse to acutish,
not contorted, the pubescence short, uniformly whitish tan to reddish; pistillate
flowers about 6 mm. long and 3 mm. broad, the calyx lobes acute and acute-
tipped, not calloused, scarcely acuminate, the disk inconspicuous, the ovary silky-
pubescent. Fruit maturing green in color, subglobose, emarginate-apiculate with
3-lobed cross-section; capsules ligneous, explosive, the valves thick, showing no
contortion at dehiscence; seeds quadrangular-kite shaped in cross-section, to about
20 mm. long and 18 mm. thick, latex sulphur to cream-yellow; rubber and yield

rather inferior.

Vernacular Names: shhniga dcbil, jebe dcbil (Peru).

Known Natural Distribution: Upland forests of the Guianas, Venezuela, Colom-
bia, Brasil and eastern north-central Peru.

Peru: dept. loreto: Rio Napo, Clotilde, fl.^^ Sept. 1940, Skutch 4q8j.

Heiea giiianens'n in pure strain appears to have been collected rarely in Peru.
Ducke has collected it (No. 1433, Feb. 2, 1942) from near the mouth of the Rio
Yavari on the Peru border, and various collections are reported from the Colom-
bian border on the Rio Putumayo. It appears to reach its western limits on the
low tierra altura hills on the Rio Napo, extending into the vicinity of Iquitos

I

where collections indicate it to exist in forms contaminated with H. paticiflora.
Other collections, as the type of H. nigra (Ule sSgS), from the upper Rio Jurua,
Acre Territory, would indicate H. gii'ianensis possibly to exist in Peru on the range
of hills east of Contamana which center on the Peru-Acre border between 7^ and

L

about 8.5'' S. However, there is morphological evidence that the type of H. nigra
is not pure H. guiancnsis but shows some contamination with H. pauciflora. The
most notable evidence of such contamination appears on the lower leaflet surfaces,
where the scales are angular and quite densely crowded. Specimens which match
Ducke's H. guianensis var. occidentalis likewise frequently show this H. pmiciflora
type of lepidote lower leaflet surface. There is a definite tendency for such trees
to have considerably weaker rubber than is ordinarily found in true H. guianensis.
H. nigra and what has been known as H. guianensis var. occidentalis may represent
H, guiancnsis with somewhat stabilized admixture of H. pauciflora germplasm.
A discussion of Peruvian specimens representing hybrids between H. guianensis
and H. pauciflora Is given In the succeeding discussion of Putative Hybrids.

la. Hevea guianensis Aubl var. lutea (Spruce ex Benth.) Ducke & Schultes,
In Caldasia 3:249. 1945.

Siphoftia lutea Soruce, ex Benth. In Hook. Kew Jour. 6:370. 1854.
Siphonia brevijolia Spruce, 1. c. 7:194. 185 5, nom. nud.
Siphonia apicnlata Spruce, ex Bail!, in Adansonia 4:2 8 5. 1864.
Hevea lutea (Spruce ex Benth.) Muell.-Arg. in Linnaea 34:204. 1865.
Hevea peruviana Lechl., ex Benth. & Hook. Gen. Ph 3:290. 1S80.

^^The following abbreviations will refer to the condition of tlie examined collection, i.e.:
fl. := in flower, fr. = in fruit, st. ^ sterile.

[Vol. 34

294 ANNALS OF THE MISSOURI BOTANICAL GARDEN

llevca lutca var. ctincata Hubcr, In Bol. Mus, Gocldi 3:3 57. 1902.

Hevea cuneata Ilubcr, 1. c. 4:626. 1906.

Hcvca hraulicnsh var. ciincafa (Huber) Pax, in Engl. Pflanzcnrcich 4:123. 1910.

Ilevca gutanensis var. cuneata (Huber) Ducke, in Arcliiv. JarJ. Bot. Rio de Janeiro 6:51.

1933, in part.
Uci'ca hitca var. pllosula Ducke, I. c. 6:53. 193 3.
Hevca lufca typica Ducke, I. c. 193 3.

Hevea lutca f. pilosula Ducke, in Archiv. Inst, Biol. Vcg. Rio dc Janeiro 2:224, 231. 1935,
Hevca guianctnh var. htfca i. peruviana (Lechl. ex Bench. & Hook.) Ducke, in Inst.

Agr. do Norte, Bol. Tec. 10:24. 1946.

In general, as //. guiaiicftsh, but somcwbat larger trees with leaflets erect to
somcwbat liorizontal, tending toward broadly lanceolate and having more distinct
pubescence along the midvcin. Male buds and calyx lobes slightly but distinctly
acuminate; anthers normally 5-7, in two irregular whorls, occasionally with one

or more anthers of the lower whorl approaching 1 mm. in length, all other
anthers about 0.5 mm. Female buds and calyx lobes somewhat acuminate. Seeds

very distinctly quadrangular-kite shaped.

From sterile material alone this variety frequently Is hard to distinguish from
H. guiancnsh. The floral morphology, which In the two extremes Is quite distinct,
frequently shows Intergrading characters.

Vernacular Names: jcbe dchil, jcbc Jchil Jc alfura, jcbc awapa, dyiringa dcbil,
shhinga ifc alfura, shirtnga dc ccrro, shirhii^^a aniariUo (Peru).

Known Natural Distribution; Upland rain forests of Colombia, Brasil, Bolivia
and Peru.

Pi Ku: HEPT. huanuco: Rio Huallaga: Tingo Maria, upland forest, alt. 675-765 m.,
fl. Dec. 1942, BaUivin 3S24, fl, Aug. 1940, Skutcb 4063, 4066, old fr. March 1946,
Scibcrf 2404, 2406, 240"/. Rio Pachitea: Pto. Tnca, H. and old fr. Oct. 1945, Scibcrt 21S4,
old fr., 2lSfi. DEPT. SAN MARTIN: Rio Huallaga: Rio Azul, alt. 800 m., fl. and old fr.
July 1945, Scibcrf 2o8j; Rio Pucarte, st. Aug. 1945, Langamack 5. ?/., old fr. Dec. 1945,
Scibcrt ^ Langamack 226t^ 2262, dept. loreto: Rio Huallaga: Yurlmaguas, upland
forest, St. Dec. 1942, Raid win 2826, 2S27, fl. Aug. -Sept. 1929, Killip ^ Smith 28706;
Rio Paranapura, Chambira Brook, st. May 1943, Fletcher s, n,; Rio Shishinagua, st. May
1943, Fletcher s. n. Rio Maranon: Pongo de Manscrlche, 1923, La Rue s. n.; near Borja,
fl. Sept. 1940, Skutch 4080, st. 498 1 y 4Q82. Rio Ucayali drainage: Rio Yurac Yacu,
Boqueron, alt. 1000 ni., old fr. July 1945, Scibcrt 20^8; Pampa del Sacramento: Ccrro
dc Chancliahuayo, st. Oct. 1898, Huber 1^77 (type of H. cuneata Huber); Rio Yurac
Yacu, old fr. July 1943, Seibert 2074; Rio Aguaitia, fr. Dec. 1944, Scibcrt IQ78, st. Nov.
1945, Seibert 22J4, fr. 22^6, Rio Ucayali, Requcna, st. Jan. 1947, Carpenter ^ Lescann
s, n, (P-151 grown from seed at Estacion Experimental, Tingo Maria). Rio Amnzon:
Pinto Cocha, Rio Nanay, st. June 1929, Llewelyn Williams 8t8; near mouth of Rio Nanay,

St. Dec. 1942, Baldwin 2822, 2823; Rio Napo, Singapor, fl. Oct. 1943, Scibcrt 1848, st.
i8so; Rio Ampiyacu, old fn Feb. 10, 1943, Russell s. n.; Huanta, fl. Oct. 1943, Scibcrt

^^^55y fl- ^'^S^ (intermediate between H. guianensis and its variety lutca), st. t8J)7; Oro
Negro, Rio Moto Huayo, old fr. Oct. 1943, Seibert l8j4, nFPT. puno: San Gavan, fl.
Collector? (H. peruviana Lechler ex herb.), di-pt. cuzco: Quince Mil, 1000 m., old fr.
May 1946, Seibert 2426, dtpt junin: Satipo, SOO m., old fr. Jan. 1946, Seibert 2370,
2371; fl. Sept. 1940 Skutch 4974,

Bolivia: Colonia, Rio Negro, fr. July 1943, Baldwin 2961,

TTcvca guianvfisis var. lufca is perhaps the most widely distributed entity of
the genus. It is a characteristic tree of the Peruvian monfaria~'\ It is found on

^"TIic term moutafia in Peru refers to al! of the heavily forested Liiid cast of the Andes. It in-
cludes the eastern Andean foothills, as well as the low expanse of tlie upper Amazonian basin.

1947]

SEIBERT HEVEA IN PERU 295

much of the Peruvian tierra altura and hilly land of the Peruvian Amazon basin,
above the flood levels of the UcayaU, Huallaga, Maranon, Napo and the Amazon.
It also IS found on the eastern Andean foothills, occasionally as high as 5000 feet.
As an entity the variety is extremely variable and frequently appears to grade
into H. gniancnsis, with which it certainly is most closely allied.

On the basis of floral morphology and pubescence, both on the flowers and
lower leaflet surfaces, H. gn'ianemh var. hitea frequently shows strong tendencies
toward H, Bcnthamlana. Morphological evidence might well suggest its having
been derived through the hybridization of H. gviancnsis and H. BejithamianUy
from which was established a true-breeding, relatively stable, but highly variable
entity. An occasional plant, although growing far from the known distribution
range of H, Bcnfhamiana, appears among typical Jf. guianemh var. lutea as a
probable recombination or throw-back in certain of its aspects to be more refer-
able to H. Bcntbamiana,

In habit and habitat the variety Jutca is similar to H. gnianensis, being a rather
large tree and, so far as I know, always found on the tierra aJfura, Though
frequently seen close to periodically inundated areas, it appears to be confined
to the land above inundation level. The position of the leaflets is erect to semi-
erect, like that of H. gttiancnsis or at most intermediate between that and the
horizontal position characteristic of H, Bcntbamiana. Leaflet size and shape,
being variable, show a tendency towards an intermediate condition.

The pubescence on the lower leaflet surface is interesting. It will be recalled

gnianenm there may be slight, sparse whitish pubescence alon^ the
midvein, while in H. Bentham'iaua there is typically a rather dense reddish pubes-
cence over the entire lower surface, associated with the veins and vcinlcts. The
indument of H. guianensh var. hitca is extremely variable. It ranges from sparse
whitish to mixed whitish and reddish, or reddish on the midvein and frequently
extending to the secondary or branch veins. One case has been found {Ray
Russell 5. //., May 14, 1943, Santa Rosa, near Pinglo, Rio Maranon, Loreto) where

dense whitish pubescence occurs over the entire surface associated with the
veinlets.

It appears that this variable pubescence character has at least twice accounted
for the description of separate entities, namely, H. lufca var. pilosiila Ducke and
H. Foxii Huber. Such variations, where hitea and Bentham'iana grow close to-
gether, may result from further hybridization between the two entities. In other
Instances these variations may be natural tendencies toward recombination. Th
range of variation and distribution as known from specimens at hand appear to be
too intergrading for any decisive subspecific naming at the present time.

The f)resence of a distinct interflush short-shoot Is strongly estabUshed and
allied to that of H. gjiiancjish, though here, too, there may on occasion exist a

H

[Vol. 34

296 ANNALS OF THE MISSOURI BOTANICAL GARDEN

condition that strongly suggests H. Bcntbamhina influence. The inflorescence
habit anJ the tendency for some leaves of the previous flush to remain until after
the appearance of the inflorescence are characteristic of H. gnmftemis, but with
modifying influence seemingly suggestive of the H. Bcntbafniatia deciduousness.
The shape of the male bud and lobe acumination of the flower in anthesis suggest
none other than an intermediate condition between guiancnsh and Bcnthamiana.
An intermediate condition also is exceptionally well shown In the anthers and in
the whorl irregularities. The disk, on the other hand, is strictly that of //.
^iuancNsIs, being represented only by a slight flaring of the staminal column base
in the staminate flower. Fruit, carpel thickness and seed shape are those of H.
guiancush, while seed coloration on occasion appears to show weak H. Bcnthamiana
tendencies. Latex flow, quality, and color are highly variable from tree to tree
and from region to region. In general, the variety Jutca produces the weak rubber
and relatively poor yield of IL guiancnsh, though individual trees may have rela-
tively high flow, relatively good quality, or frequently a whitish latex; all of
which may or may not be due to the probable //. Bcnthamiana background in its

phylogenctic history.

Where H. gulaucnsh var. hitca is found within close proximity to other species
It appears that hybridization readily takes place under proper conditions. In Peru,
specimens indicate that this variety hybridizes with H. brasilicnsh, pauciflora and
Bcnthamiana, Furthermore, evidence of hybridization Is not necessarily restricted
to the region of close proximity with the other species. Evidence of introgression
may be found over considerable distances into the IL gntancnsis var. hitca distri-
bution and away from the species with which hybridization must bave taken

lace.

P

Where

Bolivia, Madre de Dios, and in various places along the Ucayali River, infiltration of
IL brasilicnsis germ-plasm into the hifca distribution may be noted from morpho-
logical characters. In the Madre de Dios this is perhaps most striking. At Maldonado,
IL guiancnsrs var, Jiitca comes in from the soutb and west, stopping rather ab-
ruptly at the Rio Madre de Dios. To the north there Is at present a gap of some 50
kilometers in which practically no Hcvca is found. Then suddenly one encounters
H. brasilicnsis from the north and cast. It is interesting to note tbat the H.
gutancnsis van lutca from the Maldonado area has considerable infiltration of H.
brasilicnsis germ-plasm, morphologically recognizable. Furthermore, the H. brasiJ-

Ttn

a specimen collected by Hodge (No. 6013) from the upper Inambari is super-
ficially no different from specimens around Maldonado which show this H, brasil-

icnsis Introgression,

Q

Mil, one is unable to find morphological indication of H. brasilicnsis germ-plasm
in the specimens of H. guiancnsis var. lutca.

1947]

SEIBERT HEVEA IN PERU 297

1

The presence of this H. brasiliensh germ-plasm is also indicated in the quality
of rubber from that area south of the Madre de Dios and cast of the Inambari.
This area was tapped in the past boom and also during World War II. The rubber
is considered as a superior quaUty lutca rubber. Though H. brasilknsis itself docs
not ascend the Rio Pachitea or Rio Pichis, a specimen from Pto. Inca showed
strong morphological evidence of its presence within lufca and may well account
for the superior quahties of weak rubber coming from these areas in general.

From the immediate Iquitos area there are sufficient specimens to indicate
strongly the presence of an hybrid swarm on the repeatedly cleared land, some of
which is in pasture (pi. 44). Other areas consist of second growth at the present
time. From specimens at hand, part of this complex hybrid swarm has resulted
from Interspecific hybridization of H. giiknensh var. lufea and pauciflora in which
the majority of the collections show natural segregation to simulate most closely

-'fi

bra

and pauciflora. Here there is likewise a tendency for natural segregation to simu-
late H. paiiciflora most closely. Occasionally slight morphological tendencies
indicate certain specimens to contain germ-plasm of all three species. However,
they are not sufficiently clear-cut or shown in sufficient number of collections to
be convincingly measured.

Citation of specimens showing hybridization of H. gnJanenm var. lufca with
other Peruvian species may be found under the section Putative Hybrids.

2. Hevea nitida Mart, ex MuelL-Arg. in Mart. Fl. Bras. 112:301. 1874.

Sipboma uitiJa Mart, ex Muell.-Arg., 1. c., 1874.

Hevca virhlh Huber, in Bull. Soc. Bot. France 49:48. 1902; emend. Huber, in Bol.
Mus. (Gocldi) 7:235-236. 1910.

\

Small to medium-sized tree to 3 m. tall; trunk cylindrical; branches reddish;
short-shoots rather conspicuous; bud scales numerous, Hnear, about 2 mm. long,
very early caducous. Leaves partly persistent until after appearance of inflores-
cence; mature leaflets horizontal to somewhat recUnate, membranaceous but
gradually becoming at least subcorlaceous in late maturity, drj^ing reddish, lance-
olate to oblanccolate, acuminate, glabrous, concolorous, the scales of the lower
surface lacking or so sparse and minute as not to alter its color or lustre, the mid-
vcin continuous to the end of the blade tip, not calloused. Flowers whitish-yellow;
stamlnate buds obtuse, becoming somewhat acuminate, not contorted, the short
pubescence white, rather uniformly distributed; stamlnate flowers about 5 mm.
long and 3 mm. broad, the calyx lobes slightly acuminate, blunt-tipped, con-

r

spicuously calloused, the disk very conspicuous, stellate with 5 acute lobes reaching
the lowest anthers, the anthers normally 10, approximating 0.5 mm. in length, in
two regular whorls; pistillate buds acute, becoming long^acumlnate, not contorted,
the pubescence short, white, becoming very sparse towards the lower center of

[Vol. 34

298

MISSOURI

the lobes and on the tube; pistillate flowers about 9 mm. long and 3.5 mm. broad,
the calyx lobes long-acuminate, deeply incised, blunt-tipped, calloused, the disk

mm

what glabrescent. Fruit maturing purplish in color, subglobosc, emarginate-
apicuiate, with 3-lobed cross-section; capsules ligneous, explosive, the valves thick,
showing little contortion after dehiscence; seeds angular, in cross-section quad-

rangular kite-shaped toward the micropilar end, becoming hexagonal tow^ard

opposite end, to about 21 mm. long and 13 mm. thick. Latex white to buff, not
abundant; rubber very inferior.

Vernacular Namf: pnra shiringa. shiringa mapa^ jebe debil muerto (Peru).

Known Natural Distribution: Apparently both on rocky hillsides and period-
ically inundated land but closely associated with old sandstone or granitic outcrops
apparently of Cretaceous, Triassic, and Precambrlan origin. Colombia, Brasil and Peru.

Peru: dept. loreto: Rio Huallaga: Rio Yanayacu, between Rio Huallaga and
Rio Ucayali, swampy land, st. Dec. 189 8, Ilubcr I^J4 (type of H. viridis Huber).
Rio Amazonas: Rio Nanay, Iquitos area, st. June, 1929, Llewelyn Williams 8Sg, Rio
Putumayo: Occidcntc on the Peru-Colombia border, fl. 1910, Fox 2 (type emend.
U. viriJis Huber) — not examined. cultivated material: a living plant brought
from the Rio Yanayacu (Peru) by J. Huber, growing at the Jardln Botanico do Musco
Gocldi, Belem, Brasil, fl. Sept. 1942, Archer 7582, st, Sept. 1931, Kmkofj i6jS, st.
Feb. 1924, La Rue 5. ;;.

The cultivated specimens cited are of considerable interest since we have so
little good material of the species collected from Peru. They appear to represent
topotypical cultivated material from Huber's type locality of H. viridis. The one
flowering collection made by Archer is quite referable to H. nitida in floral mor-
phology, the short-shoots and, in general, the leaflets. The lower leaflet surfaces
of this and other specimens of the cultivated plant, however, tend to show a
minute lepidote condition slightly atypical of H. nitida. The scales, notwith-
standing, are neither of sufficient size nor density to affect the concolorous aspect.
There remains some question, since the leaflets do show a slight H. brasiliensis
aspect, whether or not Huber's H. viridis had some admixture of H. hrasilicttsis

gcrm-plasm.

Having seen only one leaflet from the type of H. tiifida, Ducke (1935)

questioned its affinity with H. viridis. Schultcs (1945) felt that it should belong
with H. hrasiliensis var. subconcofor. Througb the excellent photograph, made
by the Chicago Natural History Museum, of the entire type specimen of Martius*
collection deposited in the Herbarium at Munich, it has been possible to identify
LL fiitida as H. viridis with some degree of certainty. Tlie presence of interflush
short-shoots, as well as the glossy under-surface of tlic leaflets, leaves little doubt
that H. viridis sbould henceforth be rcfercd to H. nitida.

The species appears to be associated with inundablc areas along streams but
closely associated with, as well as being found on, rocky outcrops or hills of the
cafifiga^^ ^ype. H. nitida var. foxlcodrndrordes (Schultes & Vinton) Scbultes is

"^A llglu forested, rocky hill is known as a catinga in Brasil.

1947]

SEIBERT HEVEA IN PERU 299

apparently confined to areas of ecological extremes found on tKe tops of hills
jutting out of the Amazon valley floor, in Colombia, as described by Schultes
(1944), These hills arc flat-topped and may be composed of sandstone, quartzite
or granite. They are undoubtedly very old geologically, probably of Cretaceous,
Triassic, and in some instances Precambrian age.

Collections of H. nitiJa from Peru are extremely few, and the habitats are not
at all well known. Since in Colombia and Brasil the species commonly is asso-
ciated with the old geologic formations one is incited to look for some indication
of similar outcrops in the Peruvian collection areas. Reference to the Geological
Society of Americans map of South America (1946) indicates such outcrops exist
between the Huallaga and the UcayaH, southeast of Yurimaguas. It shows much
of the area north of Iquitos, through which the Putumayo passes, to be of Pre-
cambrian origin. Although no indication of this is given on Iquitos itself, it must
be said that the immediate area on which Iquitos is built is considerably higher
than the surrounding country. The soil of this area is not typical of Tertiary
deposits of the surrounding lower area, being a yellow, much compacted, clayey
sand. It is quite possible that the immediate Iquitos area is itself a relic area. If

specimen can be satisfactorily explained.

Will

appe

H. nit'nia has ycvy widely scattered distributions of rather confined and small areas.
This may indicate it to be a survival or relic species. Undoubtedly many more
localities of H, mtrda still exist to be discovered. The same may be said for
probable outcrops of these ancient formations jutting out above the Tertiary de-
posits of the Amazon valley. These probable discoveries will undoubtedly picture
more fully the chain of rehc areas running southwest from the Guianas and more
or less skirting the eastern edge of the northern Andes.

In a number of characters, Including floral and fruit structure, habitat asso-
ciation with old geological formations, and the very poor-quality rubber, there
would appear to be close relationship of H. nitida to H. pauciflora. H. mtida,
however, is easily distinguished by its concolorous leaflets, the exceedingly well-
developed staminate disk, and the pronounced acumlnation of the female bud and
mature calyx lobes.

The Peruvian specimens referable to H. nitida show no distinct morphological
evidence of hybridization with other species. However, it should be stated that
the presence of some minute scales on certain specimens could indicate the presence
of H, braulicusis germ-plasm. A selection of cultivated

from seed collected at Culparl on the Rio Huallaga and growing at TIngo Maria,
on the other hand, shows some floral and lepidote conditions that might indicate
slight contamination of H. nitida. Unfortunately, the number of specimens is
insufficient for making accurate measurements upon which to bear out Baldwin's

H. bra

[Vol. 34

300 ANNALS OF THE MISSOURI BOTANICAL GARDEN

suggestion (1947) that H, brasilienus var. suhcoticofor possibly resulted from
introgrcssion of //. nifhh genes into H. Inasilrmsis,

3, Hevea pauctflora (Spruce ex Benth.) Muell.-Arg. In Linnaea 34:203. 1865.

Siphonh pauciflora Spruce, ex Bentli. in Hook. Kew Jour. 6:370. 1854.

llevca memhranacea Muell.-Arg. in Mart. Fl. Bras. 11":299. 1874.

Ilevea confusa Hemsl. In Hook. Tc. Pi. 6:2, sub ]>/. 2^/0, pi 2S75, figs, 7-J, pi 2575,

figs. 12 6 13. 1898.
llevca tnembrauacca var. Ichgyne Duckc, in Archiv. Jard. Bot. Rio de Janeiro 6:57.

1933.
llevca pauciflora ssp. typica Duckc. in Arcliiv. Inst. Biol. Veg. Rio de Janeiro 2:239.

1935.
llevca pauciflora ssp. coriacea Ducke, 1. c. 193 5.
llevca pauciflora var. coriacea Ducke, I. c. 1935.
Hevea mcmhranacca f. Iciogync Duckc, 1. c, 193 5.

Small to large tree to 30 m. tall; branches brownish; trunk cylindrical; short-
shoots conspicuous; bud-scalcs very numerous, deltoid-acuminate, about 3 mm.
long, very early caducous. Leaves persistent to partially persistent until after
inflorescence maturation and appearance of new flush; mature leaflets horizontal
to slightly reclinate, at first membranaceous but slowly becoming quite coriaceous
and even revolute in late maturity, usually broadly lanceolate with short obtusely
acuminate tip, the pubescence none or of few short, white hairs along part of the
midvein, the scales of the lower surface very dense, angular in tile-Uke compact-
ness, producing a conspicuous whitish surface, the midvein terminating short of
the blade tip, calloused or gland-like. Flowers pale yellow; staminate buds ob-
tuse, not contorted, the pubescence white, tomcntosc, usually dense and rather
uniformly distributed; staminate flowxrs about 3-4 mm. long and 3.5 mm. broad,
the calyx lobes acute, blunt-tipped, conspicuously calloused, scarcely acuminate,
the disk inconspicuous, but of 5 small gland-like lobes, the anthers normally 10,
approximating 0.5 mm. in length, in two regular whorls; pistillate buds obtuse to
acutish, not contorted, the pubescence white, conspicuous on the lobes, but
scarcely present on the tube and pedicel; pistillate flowers about 5-6 mm. long,
3-4 mm. broad, the calyx lobes acute, blunt-tipped, conspicuously calloused, be-
coming slightly acuminate, the disk very conspicuous, of acute lobes about 1.5
mm. long, the ovary glabrate to short-pubescent. Fruit maturing purplish red-
(or green?) in color, subglobose, cmarglnate, short-apiculate with 3-lobcd cross-
section; capsules ligneous, explosive, the valves thick, showing slight contortion
at dehiscence; seeds hexagonal in cross-section, quite variable in size, about 13-25
mm. long and 10-18 mm. thick. Latex whitish to tan, oxidizing black; rubber
resinous, sticky, very weak and with little elasticity, the yield very poor.

Vr^RNACULAR Names: No vernacular nnnies referable to this species have been

encountered in Peru.

Known Natural Distridution: Apparently associated with geologically old

(Precambrlan, Triassic, and Cretaceous) outcrops on rocky or sandy slopes, frequently

1947]

SEIBERT^ — HEVEA IN PERU 301

swampy, in light forest, British Guiana, Venezuela, Brasil, Colombia, and Peru.

Peru: dept. loreto: Iquitos, Estrada Morona, marshy second growth, st. Nov. 1942,
Baldwin 2800, fr. 280 1, 2S02, fl. 28ojy 2805, 2806, fr. 2808, fl. 28og, fl. and fr. Dec.
1942, 281 S, 28 JO, Vic. Iquitos: wet, grassy area, £1. Oct. 1940, Skutch 4gpO, fl< and fr.,
4gpi, St. 4992; Mishuyacu, fl. Sept. 1929, KilUp & Smith 29919; Punchana, fl. Dec. 1942,
Baldwin 2816, fl. and fr. 2818; San Juan, fr., Dec. 1942, Baldwin 2820.

Although the above specimens are referred here to H. pane/flora, it is question-
able whether they represent this species In pure strain. As to their composite
morphological characters as a whole, they are taxonomically referable to H. pauci-
jlora. In general aspect, certain morphological details of the flowers and seeds,
habit, habitat, variable flowering and fruiting time, there are perhaps more than
faint indications of germ-plasm infiltrations from //, hrasiliensis and H. gntaucmis

var. liitca.

The H. pauciflora complex appears to be one of the older entities of the genus.
It has a wide range of distribution, but is limited within that range to what ap-
pears now to be a belt of small isolated habitats extending from Iquitos in a
general northeasterly direction to British Guiana. These habitats become progres-
sively more numerous, somewhat merging into a large area of distribution In
southeastern Venezuela and British Guiana. This complex also appears to limit
itself to areas which represent geologically old outcrops jutting through the more
recent Tertiary Amazon valley deposits. In this distributional respect H. pauci-
flora is similar to H. nifida with which it appears closely related morphologically.

The complex has gone through various phases of taxonomic splitting, in which
the Guiana material is usually considered as H, confiisa. The Brasilian material
from the Rio Negro and Sollmoes Is referred to H, pauciflora with coriaceous-
leaved specimens as H, paiuiflora var, corlacea. Although Hemsley described H,
co7tfusa as distinct from H, pauciflora, he later (1901) came to the conclusion
that it was synonymous with the latter. More recently the feeling has been that
//. confusa is synonymous w^lth H. pauciflora var. coriacca.

In my comparative morphological studies of material from the entire known
range of the complex, it has become apparent that the differences in leaflet texture
is not inherent, but a condition due to maturity. Unlike most other species
(possible exceptions are H. uifida and H. rigidifolia) , the leaflets of H. pauciflora
appear to take a relatively long time to reach their full texture maturity. Although
the leaflets reach mature size very sliortly after their appearance, they are at first
quite membranaceous and for several months gradually become coriaceous. This
species tends to hold its leaves until after the appearance of the new flush and the
new leaves have reached mature size. Through actual specimens, It has been possible
to see the previous year's coriaceous, revolute leaflets, and at the same time see the
current year's mature, membranaceous leaflets. This leads me to believe that no
valid varietal difference can be made on the basis of leaflet texture. A search has
been made, with little success, to find other morphological characters sufficiently
stable to base taxonomic segregation of these supposed entities. The specimens at

302

[Vol. 34

MISSOURI

hand show some striking variations in fruit and seed size, though perhaps no
greater than is the range in //. hrasiliensis. The seed shape would appear, however,
to be of a rather uniform pattern, having an hexagonal cross-section. There are
so few specimens of the [niudflora complex having seed that I am unable to de-
termine whether or not the seed differences are of varietal or subspecific value.

Spruce (Bentham, 18 54) has mentioned that the seeds of Hevca on the Rio
Negro are prepared and eaten by the Indians; and Dr. Baldwin informs me that
//. paudjlora is frequently grown by the Indians in their yards for the seeds which
they cat (Baldwin, 1947). It could be possible that in tlie hundreds of years
Indians have been along the Rio Negro, they have not only distributed the species
outside of its natural habitats, but also subconsciously selected for seed size.
Perhaps some of the exceptionally large-seeded specimens arc coming from old

trees

I

A further character which may or may not be of taxonomic use in the complex
lies in the color of the maturing fruit pod. Apparently it usually matures
purplish red, but sometimes the color appears to be green. Until further evidence
can be shown that valid morphological differences of taxonomic significance exist
within the complex, it seems best to consider the complex as one entity under H.
pa?iriflora. It is recognized that there appear to be few genetic boundaries to
prevent natural hybridization with other species when natural or man-made con-
ditions are favorable for it. This may be .\n important factor in the seeming
confusion within the H, panciflora complex.

Both //. Immilhr and H. pahiJosa have been described from the immediate
vicmity of Iquitos. The type material, in both species, is thought to represent
segregating material from an hybrid swarm derived through interspecific hybrid-
ization of H. panciflora and IL gulam'nsis var. lufea. In both H. paludosa and

paucifl

paiicifl

The presence of H.

guiauaish var. httca is morphologically more difficult to distinguish in the types
of //. huiuilior than in H. paludosa; but, at least through bud acumination and

pubescence characters, there can be Httle doubt of its presence.

If H. panciflora and H. guiancus'is var, lutca were the only species concerned
in producing the Iquitos hybrid swarm the problem would be relatively simple.
The swarm is complicated in that segregates of //. panciflora X hrasilicusis also
appear. Morphological evidence from the leaves, short-shoots, bud acumination
and contortion, calyx-lobe acumination, and seed characters can leave little doubt
of such a condition existing.

Although specimens show natural segregation most closely towards forms of
the //. panciflora parent, there is evidence from cultivated trees planted at Hac.
Chanticlair that, given ideal and uniform growing conditions, gradation takes
place in the direction of both parents. A discussion of this will be found under

1947]

SEIBERT HEVEA IN PERU 303

H. brasilicns'n X pauclflora in the section Putative Hybrids. The Iquitos hybrid
swarm complex is still insufficiently known and collected to give more than faint,
mconclusive evidence that certain specimens show influence of all three species
within the same plant. Undoubtedly future collections will show this to occur.
Discussion and citation of specimens from the Iquitos hybrid swarms may be
found under the section Putative Hybrids.

It seems significant that frequently the same tree is both in fruit and flower
at the same time. This is borne out by the range of flowering dates when grouping
together all specimens from Iquitos cited as H, patuiflora, H, giiianensts var. hitea
X pdHciflora, and H. brasilknsls X p^uciflora. Flowering appears to occur be-
tween July and March; furthermore, it must frequently occur twice a year to
account for flowers and mature fruit on the same tree. Segregation along morpho-
logical lines not only is taking place, but it appears that the normal flowering time
regulator has been upset, or at least modified to fit in with the extremely variable,
seasonal conditions from year to year and within the year found around Iquitos.
Frequent definite wet and dry periods alternate throughout the year. This could
give rise to a set of conditions which, when correlated and emphasized by such
man-made conditions as deforestation of the jungle, with resultant grazing and
second-growth, could well fit in with the unstable pattern of both frequent and
sporadic flowering.

The large-scale man-made changes in the immediate area of Iquitos (pi. 44),
and the natural conditions within that area, in which all three parent species exist,
together with the few genetic boundaries, would all seem to constitute an ideal
set of conditions for the development of such hybrid swarms. Since most of the
members of the swarms appear most closely to simulate H, pauciflora^ it must be
assumed that the man-made changes have simulated most closely the habitat of
that species.

The immediate area around Iquitos Is somewhat higher in elevation than the
surrounding country, and the rather compacted, sandy, clay soil is not typical of
the surrounding lower areas. It appears that Iquitos might well represent an
isolated Triassic or Cretaceous sedimentary outcrop, especially since we do find
H. paucifloray a species which In other regions seems to be confined to such old
outcrops. Practically all of this Iquitos area has been cut over In the past, not
only once but perhaps many times. It was presumably inhabited by Indians long
before the coming of white man. Here, then, rises the question suggested imme-
diately following the citation of the specimens under H. pauciflora: are these
specimens pure strain H. paiictflora? It might be possible that H. paiicijlora in

pure strain at Iquitos long since has been destroyed by man, but still persists in
the form of an hybrid swarm which is attempting to maintain its identity as H.

pauciflora.

[Vol. 34

304 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Even though H. pauciflora were not originally indigenous to Iquitos, we know
that the seeds of this species were eaten by the Indians of the Rio Negro and that
they transported seeds for planting in their yards. It would not be too hard to
presume that H, punriflora may have been Introduced into Iquitos before white
man arrived. Successive introductions of this species along the rivers at various
points would have led to a man-induced, natural selection for adaptation to a

wide range of habitats.

Food for man is scarce in the Amazonian jungle and the Indian augments his
jungle harvests with planted root and seed crops, not only in his garden but
frequently at scattered points along his hunting trails. One evidence of this is
the presence of several Brasihnut trees, Bcrthollct'ia excelsa, near the river between
Iquitos and its suburb, Punchana. The Brasll-nut tree apparently is not Indigenous
as far up the Amazon as Iquitos. Its size, as well as the opinion of the older in-
habitants, Indicates It to have been planted by Indians long before Iquitos was a
modern town. The "peach palm," pijjiayo or InfuayOy Giiilichini sp., a native of
the Andean slopes, is found frequently in Isolated stands of a few trees on well-
drained, ideal camping spots near waterways, apparently all through the Amazon
valley. These Instances are able further to substantiate a theory that //. pauci^
flora, too, may have been Introduced by the Indians.

4

4. Hevea Benthamiana Muell.-Arg. in Linnaca 34:204. 1865.

Small to medium-sized tree to 25 m. tall; trunk conspicuously swollen toward
base; branches reddish gray; short-shoots inconspicuous, of narrow ring of bud-
scale scars; bud scales few, thin, linear-acuminate, about 3 mm. long, very early
caducous. Leaves deciduous before the appearance of the Inflorescence; mature
leaflets horizontal to slightly reclinate, firmly membranaceous, very broadly
lanceolate to oblanceolate, shortly acuminate, drying reddish, usually reddish-
pubescent below, the scales of the lower surface rather dense, whitish, more or less
lens-shaped in outline, the pubescence usually dense, typically reddish over the
entire surface, and confined to the veins and vcinlets, the midvein continuous to
the blade tip, not calloused. Flowers yellowish; staminate buds acuminate, not
contorted, uniformly dense, long and reddish-pubescent, the pubescence longer
and more dense at the point of abscission; staminate flowers about 3—4 mm. long
and 2 mm. broad, the calyx lobes acuminate, not contorted and not callose-tippcd,

the disk of 5 small but conspicuous lobes, the anthers normally 8-10, about 0.5
mm. long, in two irregular to regular whorls; pistillate buds acuminate, not con-
torted, densely reddish and longish-pubescent, the pubescence becoming less dense
towards the base of the tube; pistillate flowers about 6 mm. long and 2.5 mm.
broad, the calyx lobes acuminate, not contorted and not callose-tippcd, the disk
inconspicuous, of very short lobes, the ovary densely short-pubescent. Fruit
maturing green, subglobose, emarginate-apiculate, with 3-lobed cross-section;
capsules ligneous, explosively dehiscent, the valves rather thin, not noticeably
contorting at dehiscence; seeds ellipsoidal in cross-section, ventrally compressed,

1947]

SEIBERT HEVEA IN PERU 305

but otherwise rounded with scarcely any indication of angUng, about 19 mm.
long and 14 mm. thick, the brilliant, clear, brownish mottUng having a whitish
background. Latex white, abundant, the rubber and yield considered second only
to that of H, brasiliensis.

Vernacular Names: Thus far, no vernacular names have been encountered in Peru.

Known Natural Distribution: Apparently confined to the deeply inundated areas
and igapos^^ near the major streams along and north of the Amazon, southern Venezuela,
Colombia, Brasil, and apparently along the lower Peruvian portion of the Rio Putumayo.

Of this species, I have seen no material of apparent pure strain collected from
Peruvian soil. Schultes (1945) states that H. Benthamiana occurs along the Rio Putu-
mayo of Colombia below Arica, so it is quite possible it may exist on the Peruvian side as
well. I have seen specimens, apparently representing hybridized forms of H. guiancnm
y:iT.Jutca X Benthamianay from the Peruvian-Colombian Putumayo, which have been
referred to H. Foxii and H. glabrescens. These are discussed under the section Putative
Hybrids.

Since no specimens from Peru yet coming to my attention apparently have
represented pure strain H. Bcnthamiana I am at present giving no synonymy for
the species. However, It has been necessary to revise the descriptive terms for the
species in keeping with the morphological revisions of the other Peruvian species.
This has been done largely from an isotype specimen of H. Benfhavira?ia, Spruce
2560, a very fine specimen collected from the Rio Negro region of Brasil, near

Panurc on the Rio Vaupes.

Many varieties and forms of H. Benthavtiana have been described, and it
appears that most of them have conspicuous short-shoots, a character which is not
conspicuously present in typical Bcnthamiana material. Ducke (1943), Schultes
(1945) and Baldwin (1947) recognize that H. Bcnthamiana hybridizes rather
readily with the H. gniancnsis complex, H. pauciflora and H, Spntccana, Exami-
nation has shown that many specimens considered as varieties of H. Benthaviiana
simulate It in general, except for having the conspicuous short-shoots. This, as
w^cll as the presence of other morphological characters, has convinced me that
many of these named varieties are actually the result of very frequent hybridiza-
tion between H. Bcnthamiana and these other species.

5. Hevea brasiliensts (HBK.) MuclL-Arg. in Linnaea 34:204. 18 65.

Slphonia brasiliensis Willd., ex Juss. Euphorb. Gen. 40, 113, pi 12, jig. 38b. 1824, nom.

nud.
Sipbonia brasiUcrnis HBK. Nov. Gen. ct Sp. 7:171. 1825.
}Sipboma Kuuthiana Baill. Etud. Gen. Euphorb. 326. 1858.
Hevea jancircnsis Muell.-Arg. in Mart. Fl. Bras. 11^:706. 1874.
Hevea Siebcri Warb. Kautschukf. 32-33, fig. 1900,
?Hev'ea Kimthiana (Baill.) Hubcr, in Bol, Mus. Gocldi 3:349. 1902.
Hevea brasiliensis var. angustifolht Ule, in Tropcnpflanzer, Bcihcft 6:8. 190 5.
Hevea brasiliensis var. lati folia Ule, I. c. 1905,

Hevea brasiliensis var. stylosa Huber, in Bol. Mus. Goeldi 4:640. 1906.
Hevea Randiana Huber, 1. c. 63 6. 1906.

^^Jgapo is a Brasilian word for areas subject to very heavy yearly inundation and which arc
rather permanently swampy.

306

[Vol. 34

MISSOURI

Ilevea hrasiUcnsis var. Randiana (Huber) Pax, in PflanzcnrelcK 4:123. 1910.

Hcica hrasiVionis vnr. ]aneirenm (Mucll.-Arg.) Pax, 1. c. 12L 1910.

Uet^ea hrasiUctnis var. acreana Ulc, in Entjl. Bot. J.ilirb. 50:14. 1914,

llcvca brasilicfish f. fypica Ducke, In Arch. Jard. Bot. Rio dc Janeiro 6:55. 1933.

Ilcrca brasiVicnsh var. subconcolor Ducke, I. c. 1933.

lli'ica hrasilicfisis f. subconcoJor Ducke, In Archlv. Inst. Biol. Vcg. Rio de Janeiro 2:224.
1935.

Hcvca brasilicusis £. KanJhvia (Huber) Ducke 1. c. 193 5.
Sipbofiia riiUcyana Cook, in Jour. Wash. Acad. Sci. 31:46. 1941.

Large tree to 50 m. tall; branches grayish brown; trunk cylindricalj but
noticeably swollen towards base when growing In periodically inundated land;
short-shoots inconspicuous, of narrow ring of bud-scale scars; bud scales few to
about 10, linear-deltoid, about 3-4 mm. long, early-caducous. Leaves deciduous
before appearance of inflorescence; mature leaflets reclinate, membranaceous, us-
ually lanceolate to broadly lanceolate with rather long-acuminate tip, glabrous,
the scales of the lower surface whitish and roundish in outline, the midvcin con-
tinuous to end of blade tip or extending slightly beyond, not calloused. Flowers
creamish yellow; stamlnate buds noticeably acuminate, slightly contorted, the
pubescence short, white, uniformly distributed; staminate flowers about 5 mm.
long and 2.5 mm. broad, the calyx lobes acuminate, blunt-tipped, calloused, and
contorted, the disk inconspicuous, of 5 rudimentary lobes or swellings, the anthers
10, approximating 0.5 mm. in length, In two regular whorls; pistillate buds
noticeably acuminate, slightly contorted, the short pubescence white, becoming
sparsely distributed below the lobes on the tube; pistillate flowers about 7 mm. long
and 3 mm. broad, the calyx lobes long-acuminate, blunt-tipped, calloused and
contorted, the disk Inconspicuous, the ovary silky-pubescent. Fruit maturing
green in color, subglobose, cmarginatc-aplculate, with 3-lobed cross-section; cap-
sules ligneous, explosive, the valves thick, showing no contortion at dehiscence;
seeds ellipsoidal in outline, ventrally compressed but usually without noticeable
angling in the ventrally compressed ellipsoidal cross-section, variable In size,
16-3 8 mm. long, 14-24 mm. thick. Latex white or rarely cream to yellowish;
rubber and yield superior for the genus.

Vernacular Names: jcbc fino, sbiriuga fino, shiringa legifimo, sbirhiga or scrrnga
(Peru), and scringucira (near the Pcru-Brasil border).

Known Natural Distribution: Periodically inundated land along the Amazon and
the lower courses of its larger tributaries In Venezuela, Colombia, Brasil and Peru. Also
on well-drained inter-river plateaus or gently rolling land in Parana, southeastern
Amazonas, Acre and northern Matto Grosso, Brasil; Pando, Bcni and northern La Paz,
Bolivia; and Madre de Dios, Peru.

Peru: dept. san martin: Rio Hualiaga nbove Yurimaguas, periodically inundated
land along Rio Cuiparl (seed >;rown at Estacion F.xpcrimental Agricola de Tingo Maria),
fl. Aug, 21, 1946, Carpcufcr rj Lcsnnio s. //. (P-142), fl. Sept. 1946, Carpenter 6 Lescarto
s,v, (P-143), fl. Aug. 1945, Seibcrt JJ/j. dept. lorfto: periodically inundated land: ^
Rio Maranon, confluence with Rio Ucayali, st. Dec. 1942, Bahlwiu 282S. Rio Pacaya,
aflfkicnt of Rio Ucayali, st. Mar. 1943, Russell s. n. Rio Tapiche, affluent of Rio Ucayali,
Uscar, St. Nov. 1943, Selberf iSoj; Callao, st. Mar. 3, 1943, Russell s. n., st. Mar. 4,
1943, Russell j. ?;. Rio Ucayali, Lago Curuhuaitl, above Requena, st. Nov. 1943, Seibert

1947]

SEIBERT HEVEA IN PERU 307

1882, 188 J, 1884. Rio Amazon: Rio Itaya, st. May 1929, Lleivelyn Williams 2o6; Iquitos:
Punchana, st. Dec. 1942, Baldwin 2821, Pro, fl. Aug. 1929, Llewelyn Williams 200J;
Yana Mono Island, mouth of Rio Napo, st. Oct, 1943, Seibcrt 18'/ J; Oran, below mouth
of Rio Napo, St. Oct. 1943, Seibcrt J8^J; Firmeza, across from Pebas, old fr. Oct. 1943,
Seiberf 1861; Quebrada Yanayacillo, st. Oct. 1943, Seibcrt 1864, fl. 18/6; Fortaleza, Rio
Peruate, fl. Oct. 1943, Seibcrt 1872, t8/j, st. /<?7^; La Victoria, fl. Aug.-Sept. 1929,
Llewelyn Williams 2gjl; Caballo Cocha, st. Aug. 1929, Llewelyn Williams 21^6; (From
Caballo Cocha seed progeny grown at Tingo Maria) fl. Aug. 1945, Seibcrt 22y§, 22 jy^ fl.
Aug. 1946, Carpenter 6 Lcscano s,n. (P-146), fl. Sept. 1946, 5.;;. (P-146), fl. Aug,
1946, s,n. (P-147); Islandia, mouth of Rio Yavari, fl. and fr, Oct, 1940, Sluitch 4()8y;
Rio Ataquari, Peru-Colombia border, st. Oct. 1943, Seibcrt l86g. dept. madre de dios:
Usually on well-drained land between streams, soil yellowish to reddish, sandy, clay loam:
Rio Acre drainage: Inapari, Centro Viejo, st. June 1945, Seibcrt 20 jj, Rio Tahuamanu
drainage: Iberia: fl. July 1944, Seibcrt igjQ; Centro Alianza, st. July 1944, Seibcrt 1932;
Centro Arrozal, st. May 1945, Seiberf 2021, st, June 1945, 2023, 2024, 202j, 2026, old
fr. 2027, 2028, 20J0, 20JI, st, 20J2, old fr. 20JJ; Centro Brusscllas, st. July 1944,
Seibcrt igjj, 1 936, 1937, ^93^1 Centro Miraflorcs, st. Apr, 1944, Seibcrt 1904, old fr.
igoj, St. July 1944, Seibcrt ig33, 1934, fl. Aug. 1945, Seibcrt 2130, fl, and old fr, 2141,
2142, 2143, 2144; Centro Portillo, fl. Oct. 1944, Seibcrt igSS, st. ig56, fl. ig57, 1958,
igS9> i960, igSl, 1962, 1963, St. 1964, rgdj; Centro Primavera, st. June 1945, Seibcrt
2060, old fr. 2062, 2o6j; Centro Urquilla, st. June 1944, Seibcrt ig2j; Centro Villa
Nueva, st. June 1944, Seibcrt ig26.

Bolivia: dept. pando: On well-drained land between streams, soil yellowish loam to
sand-clay loam: Rio Acre drainage: Cobija, st. Dec. 18, 1923, La Rue s. n, (three collec-
tions) ; "one hour cast" of Nazaret, fl. Aug. 1945, Seibcrt 211 j; Nazaret to Nauruedino,
between Acre and Tahuamanu drainage, fl. and old fr. Aug. 1945, Seibcrt 2ll6; Ulti-
matum to Peru border, Rio TahuamLinu drainage, fl. and old fr. Aug. 1945, Seibcrt 2I20.
Rio Tahuamanu drainage: Porvcnir, st. Dec. 23, 1923, La Rue s^n, Rio Abuna drainage:
Rio Pacahuaris, Santo Domingo, old fr. June 1943, Baldwin 2gj^, ^957' dept. beni:
well-drained ticrra altura: Ribcralta: fl. Sept. 28, 1923, Wicr s. n,; Hac. El Prado, fl. Aug.
1^45, Seibcrt 2102; junction of Beni and Madre de Dios rivers, fl, Aug. 1886, Kiisby 885,
Ivon, Rio Ivon, st. Feb. 1922, White 237S, Rio Guapore drainage: Lago Guachi, fl.
Sept, 1943, Baldwin 2gg8,

*

Within the Department of Loreto, along the Amazon (upper Solimocs)

- and its larger tributaries, as the lower Huallnga, Maranon and Ucayali, H. brasil-

iensis is almost uniformly associated with the periodically Inundated areas. Very

exceptionally, it is associated with the slightly higher, non-inundable land adjoin-

mg the periodically flooded areas, tliough the distance and tlie altitude between the
two habitats may be but a few meters.

As to the origin of H. brasiliensis as a species, I hesitate at this time to draw
any definite conclusions. Morphologically, it appears to be a complex made up of
characters both simulating and distinct from other species. Its leaflets are more
reclinate than in any other species except possibly H. rigidifolja, which, however,
are not very well known from this standpoint. The sbort-shoots of H, brasiliensis
arc less pronounced than in any other species. Typically it lacks any pubescence
on the lower mature leaflet surfaces. The staminate and pistillate calyx lobes arc
interesting in that they show small calloused tips. Calloused lobe tips are found in
no other species except H, panciflora and H. nitida^ where they are quite pro-
nounced. The calyx lobe tips of H. brasiliensis show some variation, but, In general,
are less conspicuous, suggesting an Intermediate condition between callosity and

tVoL. 34

308 ANNALS OF THE MISSOURI BOTANICAL GARDEN

n 4

the normal acute tip. Both the disk and number of anthers (10, in two regular
whorls) appear to he quite constant. The fruit, the valves, and, In a way, the
seeds perhaps most closely resemble those of H, pauciflora. The lower leaflet sur-
face IS strikingly similar to that of H. fuicrophylla. As a species, H. hrasilinnis
forms the largest trees in the genus. Its most outstanding feature is its superior

latex yield and rubber quality.

In contrast to the preferred habitat of H. hrasriiensis along the Amazon, its

habitat in the northeastern portion of the Department of Madre dc Dios and the

adjoining regions of Brasil (Acre) and Bolivia consists of well-drained, rolling

land. The soil here is a friable, reddish to yellowish, sandy, clay loam of excellent

quality. Although very little of this area, even along the major streams, is subject

to long-period flooding, the species is not at all common where considerable flooding

occurs.

Although throughout much of Its range //. brasiliensis is associated with
periodically Inundated conditions, such is not the case in its southwestcrnmost
limits of distribution. Usually it is not found on inundable land in Bolivia, Matto
Grosso or In the state of Parana. //. hrasil'icusis of the Tapajos and regions between
the Tapajos and the Xingu apparently is found on the well-drained plateau areas
above the rivers. Due to lack of sufficient collections it is not clear to mc where,
or If, there is a distinct zone of transition between upland and lowland H. brasiU
icnsis. The critical area for such determination will lie within the southern half
of the Brasihan State of Amazonas. Men in Peru who have been in the upper
Yavarl, the Rio Blanco, and the Rio Ahiquh, all rivers which have their source
from the range of hills along the Peruvian border adjoining northwestern Acre
Territory, have reported that Hcvca hrasilicfisis gradually takes to the higher, well-
drained land.

There are several interesting cases, from both the Peruvian and Bolivian border
areas, in which specimens referable to 11. hrasilicnsis show various gradations of
H. gniancnsis var. hitca influence in their characters, and vice versa. As will be •
discussed under Putative Hybrids between these two entities, there is evidence that
the ficrra altnra H, bras/licvsis is a geographic race resulting from introgression
and ecotypic selection and carries a sliglit contamination of H, gniajiensis var,
lutea germ -plasm.

In southern Madre de Dios and the adjoining region of Puno of Peru, and in
the Departments of La Paz and Beni of Bolivia, both the limits of //. hrasiJicftsis
entering from the northeast and H, gii'Mnensis var. lutea descending along the
Andean foothills are quite sharp and distinct, a comparatively narrow belt exist-
ing between the two (seemingly of varying width) in which very little H^vea is
found. Typified by the area between Maldonado and the Rio Manurlpc in Madre
de Dios, and between Rurrenabaque and Mapiri in Bolivia, it is along the border
areas of H. guiancnsh var. Uitea that there appears to be further hybridization
between the two entities. This has resulted in a zone of intcrgradatlon along the

1947]

SEIBERT HEVEA IN PERU 309

zone of intcrgrading habitat between the relatively flat, rolling area of tierra alt^ira
H. hrasilicnsh and the steep foothill slopes of H, guianensis var. hifea.

Hybridization between the tierra ba]a H. hrasilicnsh and H. guianensis var,
hifca is indicated in the following specimens: from tierra altura areas on the Rio
Pachitea; near the junction of the Maranon and Ucayah, at Nauta; on the lower
Napo; and at the mouth of the Yavari. The fact that specimens are found which
are intermediate or approach one or the other of the species indicates that natural
hybridization occurs between the two entities at any place where conditions are
such that the two species come close together. There, may be a resulting tendency
for simultaneous and parallel development of the upland race of H. brasiliensis at
many places throughout the Amazon valley.

Citation of hybrid specimens and further discussion may be found under the
section Putative Hybrids, where reference Ukcwise Is made to Interspecific hybridi-
zation involving H. panciflora.

Natural Occurrence and Species Range

The accompanying map shows the distribution of the Ilevca species in Peru
and bordering areas of neighboring countries. Collections and collecting stations
of Heiea In Peru are far too few upon which to base a complete and accurate dis-
tribution map. Little is known of large areas between many collecting stations
cited with the species descriptions. It has been necessary to fill in a large pro-
portion of the map with questionable data gained from hearsay, transient rubber
tappers, reports of rubber coming from various isolated areas, and from Impres-
sions gained through having flown over much of the lowland area of Peru east of
the Andes.

Ule (1905) has presented a map showing the region of the entire Amazon
valley In which species of Hevea are found. It presents a rather good picture of
the distribution of H. brasiliciisis and part of H. discolor scnsu H. Bcritbamianaj
but makes no attempt to show distributions of other species. The picture pre-
sented for Peru is very sketchy and incomplete. The map given In Schurz et aL
(1925), also showing the entire Amazon valley, is not much more complete in that
it deals only with H. brasiliensis and H, Bcnthamiana. The distribution of H.
Bcnthaniiana on the Peruvian Amazon and Napo appears to be entirely inaccurate.
H. brasiliensis distribution in Peru Is shown to extend into regions known to have
H. guianensis var. hifca, a very abundant species in that country but not shown

on the map. Neither map gives the over-all picture of Hevea species distribution
In Peru.

It appears that the genus reaches its southernmost distribution of about 16° S.
in the Department of La Paz, Bolivia, and Its westernmost limits around the Pongo
de Manseriche in the vicinity of the Rio Maranon In the Department of Amazonas,
Peru. In both these extremes the genus Is represented by H. guianensis var. lufea,
as it Is for the altitudinal extremes of the genus where it occasionally reaches 5000

[Vol. 34

310 ANNALS OF THE MISSOURI BOTANICAL GARDEN

feet on the Andean foothills.

Specific distributional features have been discussed under each species, but
little h:is been said concerning plant associations found with the Peruvian Hevea
species. In many cases these features are not sufficiently known for present dis-
cussion. In northwestern Madre de DIos, where Peru joins Brasil and Bolivia, it
is significant that H. brasilicusis is associated with the Brasil-nut tree, BcrthoUctia
ci. excclsay which in this region has somewhat smaller fruit and apparently better-
flavored kernels than trees from lower on the Amazon. Here both trees prefer
and arc usually confined to the well-drained fierra altura. Up to as many as eight
trees per hectare of IL hrasiliefisis, and slightly less of Bcrthollclia, have been
found on average areas of more dense distribution. Neither Hevca nor BcrthoUctia
is uniformly distributed throughout the entire area. Both appear to be in local
"pockets" of from about 10 to 100 sq. kilometers, separated by several
kilometers in which the distribution of Hevca especially may be quite sparse.
These localized areas of dense Hevca distribution are known in that region as
ccvtros; these being subdivided into colocasioves in which several shiringcros or
rubber tappers live more or less together, working their individual cstradas.

As a w^hole, the area could be classed as a JIcvea-BertholJetia forest in which
these two species are most frequently encountered and form the largest of the
forest giants. This area of Madre de Dios r^nd that adjoining in Bolivia are
frequently characterized by having a thick undergrowth of semi-climbing spiny
bamboo, paca, Gnadiia tO))ic>ifosa Hack ^ Lindn.'^^ (Scibcrt 2o6S), These mats
of GuaJua undergrowth, pacales^^, are frequently many square kilometers in ex-
tent in which little other tree growth than IL brauVicnsis and BcrthoUctia may
be found. It appears that the Guadua is slowly enveloping the forest, replacing
the old trees as they die out and preventing any other trees from gaining a foot-
hold. It is perhaps because both Hevca and BcrthoUctia form such old trees that
they still persist in these paralcs. Many estradas are at least partially located in
these bamboo areas, in which trail maintenance is a difficult and time-consuming
task because of the rapid growth and viciously spiny nature of the undergrowth.
Ule (1914) has mentioned this bamboo association with H. brasilicnsis consider-
ably further to the northeast In Brasil, referring it to Gimdua Wcbcrbauri Pilgcr.

Unlike the Hevca Inasiliensis of the Madre de Dios, that of the Amazon
and lower Ucayali grows in the periodically inundated land adjoining the
rivers where BcrthoUctia is non-existent and GnaJna is seldom found. The distri-
bution of H, brasiUensis in Peru and adjoining Brasil and Bolivia is apparently not
too dissimilar to that shown by Record & Hess (1943) for Sivictcnia wacropbyUa,
Coaba, the Peruvian mahogany. Although Sivictcuia macro phyUa extends
into regions where H. guiancnsis var. hi tea Is found, in the Iquitos area
it occurs on the periodically inundated areas skirting the river with H, brasiUensis,

^^Provisionally determined by Dr. F. A. McClure.

^ Pacales are areas in which the dominant vegetation is paca, Guaaua sp.

DISTRIBUTION OF SPECIES OF HEVEA IN THE REPUBLIC OF PERU AND BORDER REGIONS OF ADJOINING COUNTRIES

Base map copied from the Map of the Americas, South America — sheet North 1:5.000,000, American Geographical Society of New York, 1942. % original size

194 7]

SEIBERT HEVEA IN PERU 311

but appears to make its best growth on the tierra alfura. In the Madre de Dios
area it is found with H, brasiliens/s,

Ccilrela odorata in Peru includes in its hillside forest form cedro virgin, grow-
ing in areas which are typical for H. gniancnsis var. hitea, while its fierra ba]a
form, found on periodically inundated land in the Iquitos area, is frequently asso-
ciated with H. brasilicnsis.

It is not known to what extent the distribution of H. gniancnsis overlaps that
of H. gnianensis var. Iv^tea in Peru and adjoining areas. Specimens which are
referable to H, gnianensis appear to exist along with -those of H, gnianensis var.
hitca in a number of regions. On account of the sparsity of good H. gnianensis
material, no accurate lead can be given as to whether or not the two entities are
separated ecologically. It appears that the areas where H. gnianensis is recogniz-
able taxonomically also contain recognizable hybrids between either H. gnianensis
and panciflora or H. gnianensis var. hitea and panciflora. The distribution of H.
gnianensis as represented on the map is of that which Ducke interprets as H.
gnianensis var. occidenfalis. This variety has been suggested in the section, Puta-
tive Hybrids, as being of introgressive origin. Its distribution might suggest for
the most part, that, though in morphological aspect it is referable to H, gnianensis
in that there is a predominance of 5 large anthers in one irregular whorl, it might
even have resulted from H. gnianensis var. Infea X H. panciflora. The conflicting
evidence at hand strongly suggests that these areas of overlapping distribution
are badly in need of further collection and field study.

This type of distribution mapping, though probably inaccurate in man)

r re-

spects, serves as a much-needed guide in pointing out areas badly in need of further
field study and collection. It should be noted that the distribution areas showing
//. fiitida and H. pancfflora also are points at which further field study is much
needed. Although not plotted on the map, R. de Lemmos Froes recently found
H. Sprnceana to extend up the Amazon to the Rio Jutai, above Fonte Boa. It is
very doubtful, however, that this species will ever be found to occur in Peru in

pure strain.

Historical Ecology

In considering the origin and distribution of Hevea it will be necessary to have
a clear picture of the geological history of the Amazon valley and its sur-
roundings. Though many details are lacking, the Geological Society map for
South America (1946) gives the compiled knowledge from existing published
works on the subject.

The Andean uplift, which skirts the western reaches of the Amazon valley,
peters out to the north and west of the Orinoco River in Venezuela. Previous to
the late Mezozoic folding of the Andes, or at least previous to their Pliocene uplift,
it is thought that much of the drainage of the present Amazonian region passed
towards the Pacific. At that time, it would appear that an older range, still very

I

(Vol. 34

312 ANNALS OF THE MISSOURI BOTANICAL GARDEN

prominent In the Guianas (Maguire, 1945) and typified by Cerro Duida (Tate &
Hitchcock, 1930) and Mt, Roraima (Tate, 1930) In Venezuela and British Guiana,
extended towards the southwest as well as southeastward through the lower,
present Amazon region. This range, presumably, was continuous with the ranges
still extant which skirt the southeastern reaches of the Amazon valley in Goyaz

and Matto Grosso.

The Andean uplift must have stopped the westward flow of water to

the Pacific, resulting in a huge, inland lake now evidenced by the Tertiary deposits
of the Central Amazonian basin. This lake, in an effort to find an outlet, had to
push through the ancient, eastern range. Its rising waters at the same time
isolated many higher areas as islands within it. As evidenced by the extremely
low divide existing between the Amazon and Orinoco drainage at an actual Junc-
tion of the upper Casiquiare and the Orinoco in Venezuela, the Amazonian lake
may possibly have broken through northward to the Atlantic previous to its
present course. Further study of existing maps might indicate that a break once
existed at the present low divide between the upper Guapore and Paraguay rivers,
flowing out southward through the Parana basin also previous to its final break-
through and the formation of its present course toward the east.

In light of such complicated geological history as major changes in water-flow
to the four directions (referred to as "sloshing" by Baldwin, 1947) and the pres-
ence of a huge inland lake in which isolated peaks of an old land mass existed with-
out being flooded, one can begin to picture the genus as having had a complicated
genetic history in its adaptation to major ecological changes. Presumably, th
genus had its origin on the Triassic land mass or even old Precambrlan outcrops at
that time and subsequently exposed permanently. If we can base any faith on the

pre«ump

■'fi'

types. From these, and possibly through intcrgencrlc hybridization in some cases
with Cnnuria or other closely allied genera (Baldwin, 1947; Baldwin and Schultes,
1947), other species were derived and ecotypically selected. Such natural selections
were for adaptation to a succession of changing water-flow directions, new
habitats caused by the draining of the huge lake, and development of the present
Amazon drainage system.

Presumably the genus had been evolved previously and was encroaching on the
newly formed land as the lake subsided. It is possible that, due to lake conditions
and flood waters, seed dispersal at the time could have been in all directions and
very widespread through water currents to many portions of the subsiding lake
shores. It probably would have been necessary for the genus to evoK^e types
which were adapted to very wet and more or less permanently inundated condi-
tions. Such types are still extant in the form of H. Sprnccana^ Bcnthamiana and
tnirroplyylla, as are types which prefer only periodical inundations as H. hrasiliemh
and nitida. Such adaptations have been evolved not only through interspecific

1947]

SEIBERT — HEVEA IN PERU 313

hybridization, chromosome aberrations, and natrual selection, but also possibly
through intergeneric hybridization. This may be substantiated by H. Spriiceana
which suggests strong Cwniria influence in the structure of its fruit and in habitat
preference.

As the lake subsided, leaving the shores high and no longer subject to inunda-
tion except along the forming stream channels, there was further need for the
evolution of types which could again persist on well-drained land. This
habitat was to typify the greater part of the Amazon valley. In the light of this,
we find at present that H. guianensis var. hitca fits into such a habitat pattern
and has the largest distribution of any of the species. Also, we find that H. brasiU
ensis has a rather large distribution south and west of the Amazon. It exists not
only as a species in areas of comparatively light periodic inundation but In much
larger areas of an intermediate zone on the rather low but well-drained lands be-
low the relatively higher zonal distribution of H. guiancnsis var. Infea. As has
been pointed out in this paper, there is some morphological evidence that H.
hrasiliensis, as It exists on the tierra altura, is ecotyplcally and genetically different
from that which grows on periodically Inundated areas, probably having been
evolved through Introgression of H. guianemis var. lutea germ-plasm Into H.
brasiliensis.

Putative Hybrids in Peru

I

The occurrence in nature of interspecific hybridization in the genus Hcvea^
as evidenced from both wild and cultivated trees, Is a fact which can neither be
ignored nor questioned. Though Intraspeclfic hybridization within H. brasiliensis
has long been practiced as a means of Hevea improvement in the Far Eastern
plantations (s'Jacob, 1931), It is not known accurately when the first artificial
Interspecific hybrids were produced.

One of the older recognized artificial hybrids Is represented in the Herbarium
of the Arnold Arboretum by a fine flowering specimen. It was collected at the
Singapore Botanic Garden, December 17, 1923, by Btirkill s, n.. Is labeled H.
brasilicus'is X confiisa, and includes notes concerning the two parents. A study
of morphological characters shown by this specimen leaves little doubt but that
this hybrid arose from a cross between H. hrasiliensis and panclflora (confitsd) .
It is of significance that this specimen matches remarkably well two specimens
(Scibcrt 1840 and Kublman 1/2/), collected at Iqultos, Peru, which are of un-
doubted natural hybrid origin between the same species.

Ramaer (1935) has proved that artificial interspecific hybridization Is possible
between H. brasiliensis and Sprnceana, and Schmole (1938, 1941) has reported
hybrids from this cross to be superior root stocks for budded clones of H. brasil-
iensis as tested at the A.V.R.O.S. General Experiment Station in Sumatra.

Pearson (1912) and Huber (1913) recognized that hybrid swarms existed be-
tween H, brasiliensis and confusa in the Botanic Garden and plantations in Trini-

[Vol. 34

314 ANNALS OF THE MISSOURI BOTANICAL GARDEN

dad. Hubcr tKen became quite conscious of certain floral and seed variations in
trees around tlie low lands of the Amazon mouth region which appeared to him
intermediate between H. bvasilicnsis and H. Sprnceana. He recognized without
doubt that hybridization was occurring between these two species and through
comprehensive seed measurements suggested a method of selecting H, brasilicnsis
from //. Spruccana and resulting hybrids on the basis of seed size.

As Huber's successor, Ducke (193 5, 1943) has devoted sections In his taxo-
nomic works to the recognition and Interpretation of natural hybrids occurring In
the Amazon valley of Brasil. Furthermore, he has recognized that a number of
named species and varieties are of hybrid origin. With additional opinions and
evidence presented by La Rue (1926), Cook (1941), Schultes (1945), Baldwin
(1946, 1947), and others, it has become necessary to include sections dealing with
recognized h)'brids in taxonomic works on Ilci'du

Recognizable natural hybrids are morphologically distinguishable between
many of the species and appear in general collections of Heica from many parts
of the Amazon valley. It appears that there are no strong genetic barriers between
the species. If two or more species come together In their natural distribution
and if natural conditions such as flowering time be favorable, natural hybridization
may be expected to take place. If ecological conditions be naturally favorable or
made favorable by man through clearing, planting, or pasturing, In such a way as
to eliminate part or much of the strong natural selection found under normal
jungle conditions, the hybridized seed may be expected to, and does, attain the
status of mature, flowering and fruiting trees. In the jungle perhaps all but one
out of a million seeds meet the fate of being eaten by animals; or, if seeds germi-
nate, the seedlings are eaten by other animals; or the seedlings die within one to
four months after germination due to lack of sufficient light on the forest floor.
It is an extremely rare seedling which, due to a fallen jungle tree or being along a
trail or clearing, ever attains maturity.

HeVEA BRASILIENSIS X GUIANENSIS var. LUTEA

Vernacular Names: ]ehc dcbil-fuio, shiringa dcbil-fino, jchc dcbil-fi)w dc alfura,

sljiriuga Uauha (Peru).

Kno^"N Naturae Distributfon: Uusually on or near borders where the two species

come together,

Tlie specimens at hand rarely appear to be truly intermediate. In all cases it
has been possible to note them as simulating more closely one or the other of the
species. The following specimens appear most closely to simulate H. guiaucnsh
var. lufea:

Peru: dept. loreto: Rio Napo, Progreso, st. Oct. 1943, Seihert i84g. dept,
PUNO: Upper Rio In;imbari, valley of the Rio HiKiri-T luari, alt. 1000 m., st. May 1943,
Hod^e 6oi^, DEPT. MADRE DE Dios: Vic. Maldonado, south of Rio Tambopata (Seed
progeny grown at Estacion Experimental Agricola dc Tlngo Maria), st. Jan. 1947, Car-
pctitcr & Lcscano S.1U (P-I27), fl. Sept. 1946, 5.;;. (P-128), st. Jan. 1947, s.n. (P-129).
St. Jan. 1947, s. n. (P-130).

1947]

SEIBERT HEVEA IN PERU 315

These specimens, through the presence of erect leaflets and pronounced short-
shoots, are referable to H. guianensh var. liitea. Since most specimens represent
selections from seed progeny obtained south of Maidonado, it is unfortunate that
only one of these specimens is in flower, that of P-128. Here the flowers appear
intermediate but lack the small calloused tips of H. brasHiensis. They have 6-8
anthers in two irregular whorls. The leaflets appear to be intermediate, but have
the vein pubescence of lutea. As far as is known, the Hodge specimen was taken
from a distance of about 175 kilometers from the nearest H. brasHiensis. If this
be true, here is some interesting evidence of the distance to which introgression
can penetrate into species distributions.

Hevea giiianensis var. lufea is characteristic of the eastern slopes of the Peruvian
Andean foothills where it occasionally reaches an altitude of 5000 feet. In the
southern half of Peru it extends eastward into Bolivia south of the Rio Madre de
Dios. Studies made from material originating near Maidonado show it to be

H

H

Along the trail from Maidonado north to Iberia, a distance of 211 kilometers,
there is a gap of some 50 kilometers between Maidonado and the Rio Manuripe in

Hei

anuripe

occurs. From reports, this gap narrows at the BoUvian border east of Maidonado.

H

established germ-plasm of H. brasHiensis. It appears to have been derived through

H

brasHiensis characteristic of the adjoining northeastern Madre de Dios and north-
eastern Bolivia. In yield and rubber quality the Maidonado material is character-
istically intermediate. The short-shoot character frequently is more or less
intermediate; and the leaflets are often more horizontal than erect.

brasiliensis:

H

Peru: dept. huanuco: RIo Paclikea, Pto. Inca, old fr. Oct. 1945, Seibert 2 1 86
DEPT. LORETO: Nauta, St. Nov. 1943, Seibert & RusseJl l88l; Iqultos, Estrada Morona,
fl. Dec. 1942, Baldwin 2829; Rio Napo, Progrcso, st. Oct. 1943, Seibert 1847; Rio
Yavari, Islandia, st. Oct. 1940, Skutch 4g88. dept. madre de digs: vie. Maidonado,
south of Rio Tambopata (from seed progeny grown at Estacion Experimental Agricola!
Tingo Maria), st. Jan. 1947, Carpenter & Lescano s. n. (P-131).

Bolivia: dept. pando: Rio Tahuamanu, Porvenir, st. Dec. 23, 1923, La Rue s.n.
DEPT. LA PAZ: Isapuri, St. Oct. 1901, R. S. Williams 1658, 165Q; Mapiri, fl. Sept. 1939,
Krukoff 10824, St. Sept. 1907, Buchticn 1622; Palmos, st. June 1902, 71. S. Williams 1657.

These specimens appear to show the reclinate leaflets and inconspicuous short-
shoots of H. brasiliensis, but the leaflets have the vein pubescence of liitea. Where
flowers are present, the anthers are less than 10 in two irregular whorls. Th
flowers have intermediate acumination and the calyx lobes occasionally are without
small calloused tips. All specimens appear to have been collected from areas above

y

t

[Vol. 34

316 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Wh

equen

color, IS superior to that normally found in lutea.

The La Rue specimen from Porvenir, Bolivia, comes from a region In which
H. gidancmis var. htea is not known. The nearest known occurrence of Uitea is
some 225 kilometers to the south and southwest. Here Is a specimen from a
center of supposedly pure H. brasiliensis, yet It shows //. guianensis var. lutea
characters with Uttle question, even to the Inferior rubber and yellow latex.
Though this specimen more clearly shows IL guianensis var. hfea influence than
any other I have seen from the Pando-Madre de Dios area, it should be pointed out
that H. hrasilicnsh trees with yellow latex and even inferior rubber arc not too

trees occasionally show slight intergrading char-
acters In the flowers and leaves, discussed under H. brasiUensis, The indication Is
that the so-called Acre-Beni H. brasiliensis carries some slight H. guianensis var.

lutea germ-plasm.

It is known that the seed which Sir Henry Wickham took from Brasil to in-
augurate the development of the plantation rubber Industry of the Far East came
from the Tapajos River area near Boim, where the trees grow naturally on well-
drained plateau land above any periodic Inundation. Early impressions that H.
brasiliensis grew on Inundated land led to test plantings in the Far East on flooded
land. It has long since been proved that the plantations did much better on land
not subject to flooding, though the trees could live and grow if flooding were not

over too long a period of time.

Is there any difference between H, brasiliensis from the tierra ba]a and that
from tierra alfura? There seems to be no difference sufficiently distinct to base
taxonomic delimitation. Yet, there must be a genetic difference!

Repeated observations of trees and specimens from trees over the entire range
of //. brasiliensis, and from the two types of habitat, have given some rather good
morphological evidence that the tierra altura-loving H. brasiliensis has been derived
through Its Incorporation of H. guianensis (or probably var, lutea) germ-plasm.
It has been only through the detailed study of a relatively large number of speci-
mens from the Tapajos, Matto Grosso, Acre, Bolivia, and Madre de Dios that cer-
tain otherwise unaccountable features sometimes crop up. An occasional specimen
will show a tendency toward having a short-shoot. Rather frequently the mid-
veins of the lower leaflet surface will show some pubescence. There may be a
strong tendency here and there for the leaflets to approach a horizontal or semi-
Position of the leaflets, one to another, is an interesting feature
varying from apart to touching and to overlapping. The overlapping condition
is a rather constant feature of the H. gimnensis complex. Much has been made
of this variable character in clone identification work (Frey-Wyssling et. a!.,

193 2). Frequently, at least in seeds from the Madre de Dios, there is rather
strong angularity, approaching the kite-shaped cross-section. Furthermore, there
are occasional trees producing cream-colored and, more rarely, sulphur-yellow

pos

t •

1947]

SEIBERT HEVEA IN PERU 317

latex. Again reference is made to a striking specimen (La Ktu 5. «., Dec. 23,
1923, ^^itajiba'\ from Porvenir, Bolivia), which combines several pronounced
features of H. gtdancnsis van liitca. These features are rather conspicuous short-
shoots, erect leaflets, and vein pubescence, all in the same specimen. Another speci-
men {Seibert 2120, Ultimatum, Peru-Bolivia border, near Iberia), has the male
buds with only slight acumination plus pubescent leaflet veins. These two cases
occur in the center of large areas in which only H. brasiliensh is represented.
There is, it seems, no chance for direct hybridization to have taken place.

It would seem that the establishment of this "ecotype*', which resulted from
past hybridization, became relatively stabilized superficially as H. brasilietnh
except for its habitat preference. However, it is still not sufficiently well estab-
lished to prevent certain morphological throw-backs or recombinations resembling
its minor constituent. It may well be referred to as a geographic race derived
through ecotypic selection.

If in this discussion the true situation is approached, one can conclude that the

bra

H

H. gn/afiensis van lutea germ-plasm.

HeVEA BRASILIENSIS X PAUCIFLORA

Vernacular Names: None typical of this group has been encountered in Peru.

Known Natural Distribution: Thus far it has been collected from the vicinity
of Iquitos and Caballo Cocha.

P^RU: DEPT. LORETO: Iquitos! Marshy second growth on outskirts of city, fl. and fr.
Dec. 1942, Baldwin 280/", fl., fr. 2810, 281 1; between Iquitos and Morona Cocha in old
clearing along swampy stream, fl. Sept. 1944, Seibcrt 1940, fl. Oct. 1940, Skutch 4993;
Punchana, marshy land, fl. Dec. 1942, Baldwin 2S17; Mishuyacu, fl. Jan. 1930, King 8l2.
Caballo Cocha: fl. Aug. 1929, Llewelyn Williams 2jOJ.

The above Iqviitos specimens are from an hybrid swarm, being most closely

referable to H, pauciflora through the presence of the conspicuous calloused calyx

lobe tips, dense, angular scales of the lower leaflet surface, peduncle departure,

and seed characters. The varying degrees of bud and calyx lobe acumination,

somewhat rcclinate leaflets, and the continuous midvcin to the end of the blade
tips all show H. brasilicnsis influence.

In these wild specimens there is not very good evidence of intergrading segre-
gation from one species to the other, since they tend most closely to simulate H.
pauciflora. The following cultivated specimens taken from a progeny growing
at Hacienda Chanticlair on the edge of Iquitos give some experimental evidence
that, where natural selection has to some extent been eliminated, there tends to be
segregation of intergrading types from one to the other species.

r-

DEPT. LORETo: Hacienda Chanticlair, cultivated in garden, trees 12-20 m. tall, fl.
and fr. Sept. 1943, Seibcrt 1840, fl. 1841, 1842, 1843^ fl. 1844, fl. Mar. 1924, Kuhlman

On the outskirts of the city of Iquitos some 3 trees were planted about

[Vol. 34

318

ANNALS OF THE MISSOURI BOTANICAL GARDEN

twenty-five years ago from local seed in the garden of Hacienda Chanticlair. The
trees show a beautiful series of scgregational intermediates between H. brasilicfnis
and //. panciflora, indicating them to be of hybrid origin. At the time of collec-
tion some of the trees were partially defoliating, others were in full flower, some

~ +

with botli flowers and maturing fruit, some with mature fruit, while still others

4

were in sterile condition. In size they ranged from 10 to 20 m. tall. The trees
had been tapped for a short time at the beginning of the war but tappuig was
abandoned because of the uniformly low yield of poor quality rubber, high in
resin and turning blackish as is characteristic of //. panciflora. The following
table will attempt to demonstrate the morphological variations between the two
species involved. It is unfortunate that more trees were not in flower at the time
of collection to show the complete range of variation which was demonstrated by
the group of trees as a whole.

HYRRTD SWARM 11. BRASIUENSJS X PAUClFLOrxA, HACIENDA CHANTICLAIR, IQUITOS

^1

^

ilyx
ion

-d

-T3

M

•T3

Ln

:3

to

'^ t:

3

^

^

-T3

XI c

OJ

cj

oj E

o

CJ *-*

U O

4->

*-< *-•

+-. o

- -*

w C

^ "H

rt

n C 1

rt -n

rt

rt C

rt t:

C

C O

C t-

7^ vi

^ o

■g

Stami
conto

'e s

^ (J

Sta

n 3

C/1 rj

ex

>

01

tj

C

u

«tt

^j

w

sis

8

u

, f^

V

(/)

«n

a

1

OJ

•-♦-y

rt

ni

o

CJ

01

-C

I-)

to

H. brasilicnsh

Scibcrt 1840

Scibcrt 1841

Scibcrt 1842

Scibcrt 1843

Scibcrt 1844

Kuhlnuui 1727

//. panciflora

5

2

5

5

5

3

5

2

3

1

3

5

?

5

?

1

2

2

1

5

?

?

?

2

2

5

2

2

?

3

4

5

Continuous,
not calloused

Round

3

?

3

2

2

5

?

2

4

5

Continuous,

calloused
Continuous.

not calloused

Continuous,
calloused
Continuous,
not calloused
Continuous,

calloused
Continuous,

callou.sed

Inter-
mediate
Round

Round

Round

Round

Angular

3

5

Short of tip,

calloused

Angular

5

* The numbers represent relative degrees to which the ch;iracter is pronounced

The group of trees is not only characterized by having very poor rubber, but
IS relatively free of South American Leaf Blight, Dofhidclla UUi, a character
which appears to be, in the Iquitos region, more inherent in H. pancijloru as a
whole than in H. hrasilicusis. This meager evidence would indicate some link
between poor quality and low latex yield.

HeVCA GUIANENSIS X PAUCIFLORA

4
5

4

5

VrRNAcuLAR NA^tKs: sh'ir'Diga banartcrciy jchr dchil mucrto (Peru)

1947]

SEIBERT HEVEA IN PERU 319

Known Natural Distribution: Soutli western Colombia, western Brasll, eastern

nortK-central Peru.

Peru: dept. loreto: Tqultos: Estrada Morona, outskirts of city, fl. and fr. Dec.

1942, Baldwin 282g-A; Punchana, hillside in old second growth near stream, st. Feb.

1944, Seibcrt 226g, Rio Napo: Curaray, st. Oct. 1940, Sluttch 4g86; 40 kilometers

above mouth of Rio Napo, low hills away from river, fl. and old fr. Oct. 1943, Sclbert

185I.

Although morphologically referable to H. gnianeus'n in nearly all respects, the
lepidote condition of the lower leaflet surface approaches that found in H. panel-
flora. The rubber from these trees is extremely poor. It docs not retain its shape
in ball form after smoking, but rather flattens out, even overnight. The yellowish
tan, resiny latex soon oxidizes black to appear as stains on clothing characteristic
of banana juice stains — thus the vernacular name, banancra. Too little material

h

from Peru is available for thorough study, cither from the standpoint of distribu-
tion or segregational variability.

Hevea guiancinis var. viarg'niata is questionably known from Peru, by a collec-
tion made by Skutch (No. 49S6) from Curaray, Rio Napo. Many specimens of this
variety from Manaos have been examined and found to have the lepidote condition
and, on rare occasions, the disk development of H. pauriflora. The somewhat
revolute leaflet edges appear superimposed on what otherwise seems to be H.
guianensis with obovate leaflets, rounded at the tip. In tlicsc characters (disk is
not known) and in superficial aspect the Peruvian collection agrees. It would
appear that the development of H. gniaucnsis var. marghiata, through introgres-
sion of H, pauciflora into H. guianensis, should be taken into consideration. The
Manaos material indicates that the more or less stable entity has been derived
through hybrid origin.

Hevea guianensis var. lutea X Benthamiana

H. Foxii Hubcr, in Bol. Mus. Goeldi 7:228. 1913.
//. glabresccns Huber, 1. c, p. 230, in part.

Vernacular Names: ituri (Peru).

Known Natural DiSTRiBUTiONt Rio Putumayo, Peru-Colombia border and Rio
Maranon.

Peru: dept. loreto: Rio Putumayo: Liberia, fl. and fr. Feb. 1911, Fox 5. ?;. (sYNTYPE
of H. glabresccns Huber) ; Ultimo Retiro, fl. and fr. Oct. 1910, Fox 7 (type of H. Foxii
Huber). Rio Maranon: Santa Rosa near Pinglo, st. May 14, 1943, Russell s.n.

These specimens morphologically are most closely allied to H, guianensis var.
hitca, with its distinct short-shoot, lack of disk lobes, and very irregular anther
whorls. In Fox's Liberia specimen, the anthers are of two sizes, two approaching
1 mm. in length, the others about 0.5 mm. However, these characters are compli-
cated by leaflet pubescence, long reddish floral pubescence, and the bud and calyx
lobe acumination, which are definitely characters of H. Benthamiana.

Two specimens of Huber's H. glabresccns were cited in the original descrip-
tion. Both these Fox collections from Sombra and from Pcbas appear to me to be
better placed in H. guianensis var, lutea X pauciflora.

\

320

[Vol. 34

MISSOURI

It is unfortunate that the RusscU specimen from the Rio Maranon Is sterile,
preventing accurate placement. It is such an unusual specimen that superficially
some might think it to be a new species. The presence of short-shoots and other
characters indicates to me that it Is allied to //. giiiancnsh var. lutea, but the
pubescence of the lower leaflet surfaces is extremely dense, as dense as in the most
typical of H. Berttbaniiana specimens. Tlie pubescence, however, is white, a char-
acter not too often found in good //. Benthamiuna. With the exception of H.
Bcntbamianay reported found on the Peru border on the lower Putumayo, H.
Benthamiaua appears actually never to have been collected in Peru, much less as
far up on the Maranon as Pinglo. The question arises as to how H. Bcutbamiana
characters can arise in H. guiancnsis var. hifca specimens so far away from JFf.
Bcuthamiaua range. Either H. Bcntharuiana exists along the Peruvian Amazon
and Maranon and has been overlooked, or //. guhficnsis var. lutca is carrying H.
Bcnthamiana germ-plasm which occasionally recombines in certain specimens to
show Itself rather strongly, as in the Russell specimen. This latter theory at
present appears plausible to n^. Through comparative morphological studies of
the flowers, leaves, stems and short-shoots, and through the great amount of
segrcgatlonal-like variablHty of //. guiaucnsh var. luteal T feel that It may

have been derived through interspecific hybridization of H. guiancnsh and H.
Bcnthamiafia,

According to Fox's notes published with Ruber's description, 11. Foxii pro-
duced 75 per cent of the rubber of the Putumayo. This region was again in
production during the past war, from which c^me the "Putumayo Block" and
''rubos dc Putumayo^'^'' grades of rubber slightly superior to normal H. gniancnsh
var. lutca. This better quahty rubber, as well as morphological characters,
would indicate that the H. guiaucmis var. hitca of the Putumayo may
have incorporated some of the H. Bcutbamiana rubber quality and more of the
other H. Bcufbawiaua tendencies than is generally found in other regions.

Hevea guianensis var. lutea X i'aucifiora

//. gldbrcsccns Hubcr in Bol. Mus. Goeldi 7:230. 1913, in part.

In morphological aspect the following specimens are most closely referable to
H. guiaucusis var. lutca:

Vfrnacular Names: ]cbc dcb'd, ]cbc debit dcbil, jebe debit banancra (Peru).

Known Natural Distribution: T terra aliura, probably In secondary growth, scat-
tered alon^ the Rio Putumayo and the Rio Ama7anas, from Tquitos to Caballo Cocha.

Peru: dipt. LORFTOr Rio Putumayo: Sombra, fr. Dec. 24, 1910, Vox s. n. (syntype
of IL glabresccns Hubcr). Rio Amazon: Rio Nanay, Tierra Doble, st. June 1929,

Literally tran^latoJ from llic Spanish as "Putumayo tails," used during the past rubber boom
as a term describing tail-like appearing masses In the ciassificatioo of a rubber from that
region of Peru. These so-called *Tutumayo tails" are made up of scrap rubber, taken from the
tapped trees, and is wrapped inro shapes simulating tails. Before the use of the more modern
tapp'ng knives came into cfTect, trees were tapped with a small hatchet known as the inachadino.
In the Putumayo area, coagulated rubber from these wounds was placed in a crude press to form
huge masses or "blocks" of rubber, known as "Putumayo blocks."

1947]

SEIBERT HEVEA IN PERU 321

Llewelyn Williams gi8, Pebas, st. Mar. 4, 1911, Fox s.n, (paratype of k. glabrescens
Huber); Caballo Cocha, Quebrada Mazamore Cana, st. Oct. 1943, Seiberi l86j, fl. l868.

These specimens all show remarkable similarity to the one syntype of H.
glahrescens collected by Fox at Sombra on the Rio Putumayo. They arc referable
to H, guiancnm var. lutea. However, the way in which the peduncle departs
horizontally from the short-shoot and hangs down, the short-shoots, leaflet per-

/>'

/^'

tnensis X panciflora.

The following specimens are most

pan erf I

indications that they are members of an hybrid swarm in which segregation and
natural selection arc such as to produce types most closely simulating H, pauciflora.

H. paliidosa Ule, In Engl. Bot. Jahrb. 3S:666. 1905.

H. humilior Ducke in Archiv. Jard. Bot. Rio de Janeiro 5:154, pi 20, fig, 50. 1930.

Known Natural Distribution: On tierra altara, but on the edges and near local
marshy areas of pasture land containing second growth and in second growth in the

vicinity skirting Iquitos (pL 44). - • n

Peru: dept. loreto: Iquitos: Estrada Morona, marshy land, outskirts of city, fl.
Nov. 1942, Baldwin 28 1 2, 281 3, fl. and fr. 28 1 4; road to Morona Cocha, fr. Nov. 1945,
Ducke 1774, fl. and fr. Oct. 1927, Ducke 20602 (syntype-2 of H. humilior Ducke), fl.
March 1924, Kublman 1526 (Jard. Bot. Rio No. 2411), (syntype-1 of «. humilior
Ducke), fl. July 1902, Vie 6260 (type of H. paludosa Ule); vie. Punchana, fl. and fr.
Sept. 1943, Scibert 1838, fl. iSjQ; San Juan, st. Dec. 1942, Baldwin 2819; Mishuyacu,
fl. Oct.-Nov, 1929, King 128.

These trees, up to 20 m. in height, appear to have a variable range of flowering
time. From the presence of both maturing fruit and flowers on the same tree, it
would seem that the trees may flower more than once a year. There is some bud
acumination, usually an irregularity of the anther whorls, less than 10 anthers,
and a reddish tinge to the floral pubescence — all indications of H. gntancnsis var.
hitea. The calyx lobes always show calloused tips, the peduncle departs at right
angles from the stem and droops, and the vein tips are usually calloused and stop

short of the blade tip —

pauci flora, Th

intergradc between the type of the two entities. The rubber, where known, is
always very poor and resinous, turning black when drying.

As in the case of Kuhlman's syntype of H. humilior, it is not always too easy
to distinguish morphological characters of H. gniancnsis var. lutea except in a
rather vague way through pubescence characters. Furthermore, the leaflets tend
to be more horizontal than erect, so strongly do these members of the hybrid
swarm tend towards H, paueiflora.

Hevea guianensis var. lutea X pauciflora X brasiliensis

Peru: dept. loreto: Rio Ampiyacu, Puca Orquillo, fl. and young fr. Oct. 1943,
Seiberf 1862,

This specimen has caused me considerable concern because it shows morpho-

^

[Vol. 34

322 ANNALS OF THE MISSOURI BOTANICAL GARDEN

logical influence of all three species, but, in general, it perhaps most closely simu-
lates H. gnianonh var. hifca on the basis of the erect leaflets, the short-shoots, 8
or less anthers in two whorls, and the vein pubescence of the lower leaflet surfaces.
The inflorescence is intermediate between H. pauci flora and H. hrasilknsis while
the way in which the peduncle of the young infructcscence departs from the stem
is that of H. pauciflora. The leaves, which persist until after inflorescence
maturation and the appearance of the new flush, appear as those found in H.
pauciflora. The flowers, through acumination and pubescence, arc superficially
those of IL hrasilicmh, even to the slight bud contortion. The calyx lobes lack
the calloused Lips of either H. hrasilh'nsis or pauciflora. The presence of disk
lobes In the pistillate flower simulates those of //. pauciflora. The leaflets are
quite intermediate between H. giiiauensi^ var. hitca and H. hrasilicnsis. It is
interesting to note that the male buds always absciss before anthesis. Perhaps this
is a type of male stcrlhtv.

The tree was rather small, 15 m. tall, growing on a low hillside well above
inundation level. It came from an area where H. guianensis var. fnfea predomi-
nates and is known as jehe dchil fjuo dc altwa, yet the rubber from this tree
appeared to be quite inferior. Tlic cream-colored latex stains the hands red
before rapidly oxidizing to black. This blackish oxidation of the latex is a
character whicli in my experience always shows up where admixture of H. pauci-
flora is suspected.

Economic Aspects oi Current Investigations

Langford (1945) has shown that within the native habitat of the commercially
grown H. hrasilicusis there arc strains which naturally resist the virulent South
American Leaf Blight, Dothidclla Ulci P. Henn. Notably resistant strains have
been found to exist in the Acre territory of Brasil and from the Leticia region of
Colombia on the Peruvian border. More recent studies have included the region
of northeastern Madre de Uios, Peru, within the ranize of resistant H. hrasilieus

IS

St rams.

Living material from these areas is now being grown and tested at various
Latin American cooperative experiment stations, on the basis of which it is now
possible to develop a sound industry of commercially grown rubber in this hemis-
phere. Since the South American Leaf Bhght exists in many of the Latin
American countries and threatens to spread to those in which it has not yet been
reported, the developmcftt of this natural rubber industry is being based on
material naturally resistant to the disease.

Although clones of //. hrasilicusis have been developed in the Far East which

have proved to be of superior yield, none of them have withstood resistance tests

. against this disease. As a temporary means of utilizing the high-producing, but

susceptible, Oriental Clones in the trial plantings within this hemisphere it is

necessary to top- or crown-bud these plants with indigenous clones of proven

1947]

SEIBERT HEVEA IN PERU 323

resistance to the Leaf Blight (Sorensen, 1942). Clones for top budding may in-
clude resistant strains of other species than H. hrasilicnsis.

Thus far, time has been a limiting factor In testing and proving yield poten-
tialities of more than a comparatively few of the earlier-found resistant clones.
It would seem likely that through large-scale testing of many jungle-selected
clones, particularly from superior yielding trees from such promising areas as
Madre de Dios and adjoining Acre and Pando, a number of clones will prove to
have naturally inherent characteristics of combined superior yield and high re-
sistance. Seedling progeny from seed collected in such areas are growing to
maturity at various stations. From these progeny further desired selections may
be made and may prove even better than actual individual jungle selections. In
addition to these methods of obtaining desirable planting material for Latin
America, a breeding program Is under way. By artificial breeding it Is possible
to utilize and combine desired characters of many clones. It Is necessary that
these clones meet specified requirements of many localities throughout Latin
America where the growing of Hcica as an additional small farm cash crop would
be of benefit to the community.

Hevea brasilicnsh^ as we are beginning to understand it throughout its very
large range In the Amazon Valley, is an extremely variable species. It is variable
not only in its morphology, but In Its habitat preferences, altitudlnal range, dry-
season tolerance, disease resistance, latex yield, rubber quality and many other
specialized features. It becomes apparent along modern lines of genetic thought
that within this species Itself arc the basic ingredients for breeding artificial
clonal material suitable to many of the varied conditions found throughout Latin
America. For example, the fact that dense human populations and small-farm
communities are more or less confined to elevations above the hot, insect-Infested
lowlands in which Hevca is considered to grow best, need not mean that human
populations must be moved to Hcvea-grov^mg areas. It should be the aim, since
high-elevation stock Is available, to develop Hevea so It can be taken to the
populations existing between 2000 and 4000 feet, or even higher.

Though natural hybrids between many of the species arc recognized, we do
not know much more about them than that there appear to be few genetic barriers
between the various species.

No species has yet been found to have superior yield of rubber to that of H.
hrasilieush. Where known, hybrids between it and any other species appear to
result in a considerable lowering of qualities which are of prime commercial im-
portance. Indiscriminate or promiscuous interspecific hybridization, if not con-
trolled, could well lead to expenditures of huge sums of money and disastrous
results. There are, however, a number of features of some other species such as
exceptional disease resistance In strains of H. paticifJora and certain other species,
the xerophytic nature of some forms of H. nitida, and many others which hold an
interesting problem of interspecific hybridization for the plant breeder In the
improvement of plantation Hevea,

[Vol. 34

324 ANNALS OF THE MISSOURI BOTANICAL GARDEN

List of Peruvian and Bolivian Specimens Studied

Archer, W. A. 7582 (H, nstida); 7jSj (H, guianrnsis var. lufea).

Baldwin, J. T., Jr. 2800, 2S0T, 2S02, 2803, 2805, 2806 (\H, paucifhra) ; 2807 (IL brastU

icnsis X pauciflora); 2S08, 2S0Q (IL pauciflora) ; 28 10, 28 1 1 (H. brasilicmis X

pauciflora); 2812, 2S1J, 281 4 fff. f^uiaficfish var, lutca X pauciflora); 28lSy 2816

(11. pauciflora); 2817 (H. hranVtcnm X pauciflora); 2S18 (H. pauciflora); 28TQ (H.

guiaucusrs var. lufca X pauciflora) ; 2820 (H, pauciflora) ; 2821 (H. hrasilicf^sis) ;

2822, 2823, 2824, 2826, 2827 (U. guianeu$is> var. lutea); 2828 (H, hrasiljcnsis) ;

282Q fW. hrasiliensis X guianensis var. lufca); 282Q-A (H. guianctjsis X peruciflora) ;

28 JO (H. pauciflora) ; 2g^^, ^957 fW. hrasilietjsis) ; 2g6l (H. guiaueusis var. lutca);

2Qq8 (H. brasilienm) .
Buchticn, Otto. 1622 (IL brasilicnsis X guiancnfn var. lufca).
Carpenter, J. B., & Manuel Lcscano. .?. n. P-A'^'\ P-B, P-1, P-24, P-25, P-26, P-36, P-39,

P-52, P-5i5, P-58, P-65, P-66, P-67, P-69, P-73, P-76, P-77, P-78, P-80, P-81,

P-83, P-84, P-85, P-86, P-87, P-88, P-90, P-91,P-99fH. brasilicusis) ; P-I27, P-128,

P-129, P-1 30, P-1 31 (H, braulicush X guiannnis var. lufca); P-142 (H, brasilicftsis) ;

P-143 (H. brasilicusis X ?nifida); P-1 45, P-1 46, P-147 (H, brasilicrtsis) ; P-151,

P-1 5 3 (H. guianensis var. lutca).
Ducke, Adolfo. 1774 CH. guiaucusis var. lutca X pauciflora) ; 20jq8 (H, guianensis var.

lutea); 2o6o2 (H. guiancnsh var. lufca X pauciflora).
Fletcher, Claude, s. n, (H. brasilicnsis) ; 5.;;. (H. giuancnsk var. lufca).
Fox, W. 7, Herb. Rio I184S (H. guiancnsh var. lufca X Bcnfbamiana) ; Herb. Rio T1847

(H. guianensis var. lufca X pauciflora) ; Herb. Rio 1 1848 (H, guianensis var. /wAra

X Bcnfbamiana) ; Herb. Rio Tl84g (H. guianensis var. lutca X pauciflora).
Hodge, W. H. (5o/j (^H". brasilicnsis X guianensis var. lufca).
Huber, J. /J77 TW. guianensis var. lufca); 1^34 (II. nitiJa).
Killip, E. P., & A. C. Smith. 2^406 (II. brasilicnsis X H. guianensis var. lufca); 28706

(H, gula?iensis var. lufca); 2QgiQ (IL pauciflora).
Klug, G. 128 (H. guianensis var. lufca X pauciflora); 8 12 (II, brasilicnsis X pauciflora).
Krukoff, B. A. 1628 (H. nifida); 10824 (II. brasilicnsis X guianensis var. lufca).
Kuhlnian, J. G. Tj26 (II, guianensis var. lutea X pauciflora); 1 527 (II. brasilicnsis X

pauciflora) .
Langemack, Victor. 5.;/. (H. guianensis var. lutea).
La Rue, C. D. s. n, (II, brasilicnsis); s. n. (II. brasilicnsis X guianensis var. lutea);

9*n, (H, guianensis var. lufca).
Rusby, H. H. 88s (H- brasilicnsis).
Russell, Raymond, s. n. (IL brasilicnsis) ; s,n. (IL guianensis var. lutea); (H. guianensis

var. lufca X Bcnfbamiana) ,
Seibert, R. J. /5?(S*, /.^JO fff. guianensis var. ////r*/ X pauciflora); 1840, 1841, T842,

1843, 1844 OH. brasilicnsis X pauciflora) ; T847 (H. brasilicnsis X guianensis var.

lufca) ; 1848 (IL guianensis var. lutca) ; f84g (IL brasilicnsis X guianensis var.

lufca); 1S50 (H. guianensis var. lufca); iSjI (IL guiancnsh X pauciflora); 18^3

(H, brasilicnsis) ; 1854 (H, guiancnsh var. lufca X pauciflora) ; 18 jj, ^856, 7857

(H, guianensis var. lutca); 1861 (I I. brasilicnsis) ; 1862 (IL guianensis var. lufca X
paucifUna X brai^liensis) ; 1864 (II, brasilicnsis) ; T867, 1 868 (IL guianensis var.
futca X pauciflora); l86g, 1872, 1873, 1874, 1876, 1877, 1882, 1883, J 884, l8g3,

IQ04, igos, ig23, ig2S, IQ26, 1032] 1033, ig34. ig35, ig36, ig37, 103S, 1039

CH, brasilicnsis) ; ig40 (IL brasilicnsis X pauciflora) ; ig43 (IL guianensis var.

lutea); 1944, igs5> W5^> W57. ^0^8, 1030, ig6o, ig6i, ig62, ig63, 1964; 1963,

TQ74, ig73 (H. brasilicnsh) ; ig78 (IL guiancnsh var. lutca); 2021, 2023, 2024, 2023,
2026, 2027, 2028, 2030, 2031, 2032, 2033, 2033, 2060, 2063, 2063 (II. brasilicnsis) ;
2074, 2078 (H. guianensis var. lufca); 2082 (H, brasilicnsh) ; 2087 (H. guianensis

var. lufca); 20gi, 2102, 2IO3, 2Jo6^ 21 T3, 21 t6, 2T20, 21 30, 2141, 2142, 2143,

Collectors' number lacking, these being references to the clone number of the plant collected.
The clones represent jungle selections from Peru cultivated at the Estacion Experimental Agricola
de Tingo Maria, Peru, where they are undergoing experimental testing.

1947]

SEIBERT HEVEA IN PERU 325

2144 (H. brasilicnm); 2184, 218 j (H. gjtlanensh var. lutca); 2l86 (H. hrasilic?ish
X guianensis var. hifea) ; 2234, 2236, 226r, 2262 (H. gtdanensh var. lutea) ; 226q
(H. gidanensh X pauciflora) ; 23^0, 237T, 2404, 2406, 2407, 2426 (H. guianemh
var. Jiifea).

Seibert, R. J., & Manuel Lescano. 2270, 2271, 2272, 2273, 2274, 2275, 2276, 2277,

2278, 2279, 2280, 2281, 2282, 2283, 2284, 2285, 2286, 2287, 2288, 228g,

22go, 22QI (H. brasiliensis).

Seibert, R. J., & Raymond Russell. 1881 (H. brasiliensh X guianensh var. hiiea),

Skutch, A. F. 4963, 4966, 4074 (H, gniancnm var. hitea) ; 4976 (H. brasiliensh X
guianemh var. hdea) ; 4980, 4981, 4983 (H. gtiianensh var. lutea) ; 4984 (H, brasiU
iensis); 498 S (H. gjcianensis) ; 4986 (H. guianensh X pauciflora); 4987 (H. brasil-
iensh); 4988 (H. brasiliensh X guianensh var. lutea); 4990, 4991, 4992 (H, pauci-
flora); 4993 (H, brasiliensis X pauciflora).

Ule, E. 6260 (H. guianensh var. lutea X pauciflora).

White, O. E. 2378 (H. brasiliensh).

Wier, J. R. s. n. (H, brasiliensis).

Williams, Llewelyn. 2o6 (H. brasiliensh); Sf8 (H. guianensh var. hifea); 88q (H. nitida) ;
918 (H. guianensh var. htea X pauciflora); 2003, 2176 (H, brasiliensh); 2503 (H.
brasiliensh X pauciflora); 293 1 (H. brasiliensh).

Williams, R. S. 7657, l6j8y 1659 (H. brasiliensh X guianensh var. lufea).

Bibliography

Association of Central Experiment Stations (Ccntrale Proefstations Verecniging) (1939). Tdentifi-

catiekcnmcrkcn van de Voornaamste in de Praktijk Aangeplante Hevea-clooncn. Java.
Arbcr, Agnes (1934). The Gramincae. The Macmlllian Co., New York.
Aublct, M. F. (1775). Histoire dcs plantes de la Guiane Fran^oise. 2:871-873, pi 33$.
Baldwin, J. T., Jr. (1946). Am. Jour. Bot. Suppl. 33:1s.

, (1947). Hevca: a first interpretation. Jour. Heredity 38:54-64.

, (1947). Hevca rl^idifolia. Am. Jour. Bot. 34:261-265.

■, and R. E. Schultcs (1947). A conspectus of the genus Cunuria. Leafl. Bot. Mus. Harvard

Univ. 12:325-331.
Bartlett, H. H. (1927). A corky-barked mutation of Hcvea brasiliensis. Bot. Gaz. 84:200-207,

ill lis.
Bcntham, G. (1854). On the north Brazilian Euphorbiaceae in the collections of Mr. Spruce.

Hook. Jour. Bot. 6:368-371.
Blandin, J. J. (1941). Why rubber is coming home. Agriculture in the Americas 1^:1-10, illus.
Boblloff, W. (1931). Color-reactions of latex as a mark of identification of Hcvea clones. Archicf

V. d. Rubbcrc. 15:302-308.
Brandes, E. W. (1941). Rubber on the rebound — East to W'est. Agriculture in the Americas

13:1-11, illns.
, (1943). Progress in hemisphere rubber plantation development. India Rubber World

108:143-145, ilbis.
Chamberlain, C. J. (1935). Gymnosperms: structure and evolution. University of Chicago Press,

Chicago.
Collins, G. N. (1903). Dimorphism in the shoots of Ginkgo. Plant World 6:9-11.

Cook, O. F. (1941). Jour. Wash. Acad. Sci. 31:46-65.
Ducke, A. (1925) Archiv. Bot. Rio de Janeiro 4:111.

— , (193 5). Revision of the genus Hevea, mainly the Brazilian species. Archiv. Inst. Biol.

Veg. Rio de Janeiro 2:217-346. (Reprinted 1939).

(1943). Novas contribu<;oes para o conhecimento das Scringueiras ("Hevea") da

Amazonia Brasileira. Agr. Serv. Florestal Rio de Janeiro 2:25-43.
Frey-Wyssling, A. (1931). Abnormal leaves of Hcvea brasiliensis as a clonal characteristic. Archief

V. d. Rubberc. 15:114-124.
, (1933). Characteristics of tappable buddings of the AVROS clones. Ibid, 17:7-12,

pi. 1-2.

, C. Heusscr, and F. W. Ostendorf (1932). Identification of young buddings of Hevca.

Ibid. 16:51-99, pJ. I-50, 2 tables.
Geological Society of America (1946). Geological map of South America. New York.

Hemsley, W. B. (1898). Hooker's Ic. Pi. 6:pl. 2 $70-2577-
3 (1901). Jour. Bot. 39:189.

326 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

#

Hubcr, J. (1906). Ensaio d'uma synopse das espccics do gcncro Hcvea. Bol. Mus. Gocldi 4:620-651.

— -^ , (1913). Novas contribu;,^ocs para o conhecimcnto do >;enero Hevca. Ibid. 7:199-281.

Klippcrt, W. E. (1942). Smill farm rubber production. Agriculture In the Americas 2:48-53,

ill as.
La Condaminc, C, M. dc (1755). Sur unc resine clastiquc, nouvcllemcnt dccouvcrtc a Cayenne par

M. Fresfifau: et sur Tusage dc divers sues InitcuN d*arbres de la Guianc ou France Eqiiinoctialc.

Mem. Acad. Scl. Paris 1751:319-33 3, illns,
Lanj;ford, M. 11. (1945). Soutli American Leaf Bliglit of Hcvea rubbertrecs. U. S. Dcpt. Agr.

Tech. Bull. 882.
La Rue, C. D. (1919). Variation in fruits and seeds of Hevca. Vcrslag van de achtste bijeenkomst

van hot tcchnisch personccl der proefstatlons en ambtenaren van het department van hindbouw.

Mcdan, Svmiatra.

, (1926). The Ilevca rubber tree in the Amazon Valhv. U. S. Dept. Agr. Bull. 1942.

Maas, J. G. J. A. (1919). Die blocmbiologie van Ucica brastVicftsh. Archlef v, d. Rubberc.

3:280-312,

M.i,^uirc, Bassctt (1945). Notes on the geology and geography of Tafelberg, Surinam. Geogr.
Rev. 35:563-579.

Mann, C. E. T. (1940). Improvement in the quality of rubber planting material. The Planter
1910:332-342.

Michener, C. D, (1946). Notes on the habits of some Panamanian stingless bees (Hymenoptera,

Apidae). Jour. N. Y. Ent. Soc. 54:179-197.
Mueller-Argovicnsis, J. (1865). Euphorbiaceae. Linnaca 34:203-204.
Onendorf, F. W., and H. Ramaer (1931). On phyllolaxis in Hevca (English abstract). Archief

V. d. Rubberc. 15:437-4-10.
Oviedo y Valdes, G. F. dc (1535). Historia general y natural de las Tndias. Madrid.
Paddock, E. F. (1943). On the number of chromosomes in Hevea. Chron. Bot. 7:412-413.
Parkin, John (1900). Observations on latex and its functions. Ann. Bot. 14:193-214» pi 12.
, (1904). The extra-floral nectaries of Ucrra hrasilicfjsis Mucll.-Arg. (The Para Rubber

Tree), an example of bud scales serving as nectaries. JbiJ. 18:217-226.
Pearson. TI. (I9I2). Trinidad and its rubber. India Rubber World 46:471-474.
Perry, B. A. (1942). Cytological relationships in the Euphorbiaceae. Va. Jour. Sci. 3:140-144.
■ , (1943). Chromosome number and phylogcnetic relationships in the Euphorbiaceae. Am.

Jour. Bot. 30:527-543.
Preusse-Spcrbcr, O. (1916). Die Kautschuk^onen Amerikas. Tropcnpflan7cr 19:201. 1916.
Ramaer, H. (1935). Cytology in Hcvea. Genetica 17:193-236.
Rands, R. D. (1942). Hevea rubber culture In Latin America. India Rubber World 106:239-243,

350-356, 461-465, illns.

Record, S. J., and R. W. Hess (1943). Timbers of the new world. Yale University Press, New
Haven.

Royal Botanic Gardens, Ceylon (1899). Caoutchouc or India Rubber: Its origin, collection, and

preparation for the market, etc. Circular 105-168, June.
Schmolc, J. F. (1938). Het'ea brasilictjsis. en Hcvea Spruccufiu — hybride als onderbtam voor

oculatics. Archief v. d. Rubberc. 22:178-181.

, (1941). Heiea brauUcnsis and Hcrea Sprnccana hybrids as stock for budgrafts II,

Jhiil. 25:159-165.
Schulies, R. E. (1944). Caldasia 3:25-32, illns.
^ (1944). Ibid, 3:249.

— , (1945). Estudio prcllmmar del gencro Hevea en Colombia. Rev. Acad. Colomb. de

Cien. Exact. 6:331-338, illus.

, (1945). The genus Hcvea in Colombia. Lcafl, Bot. Mus. Hnrv.ird Univ. 12:1-19, illus,

— , (1947). Studies in the genus Hevca, I. Ibid. 13:1-15.

Schurz, \Vm. L., O. D. HargJs, C. F. Marbut. and C. B. Manifold (1925). Rubber production in

the Amazon Valley. U. S. Dcpt. Commerce, Trade Promotion Series No. 23.
Seibcrt, R. J. (1947). The shiringero — Upper Amazon rubber tapper. Agriculture in the Americas

7:43-45, illus.

s'Jacob, J. C. (1931). Procvcn over kunstmatige kruls-en zelfbcstuiving bij Hcvea braulicmh.

Archief v. d. Rubberc. 15:261.

Sorcnscn, H. G. (1942). Crown budding for healthy Hevea. Agriculture in the Americas
2:191-193.

Tate, G. H. H. (1930). Notes on the Mount Roralma region. Geogr. Rev. 20:53-68.

. and C. B. Hitchcock. (1930). Tlie Cerro Duida region of Venezuela. Ibid, 31-52.

Ule, E. (1905). Kautschukgewinnung und Kautschukhandel am Amazonenstrome. Bciheft, Tropen-
pflanzcr 6:1-71. 1905.

, (1914). Hevca brasilietjsis im uberschwcmmungsfrcicn Gcblet des Amazonenstromes.
Bot. Jahrb. 50; Belblatt 114:13-18.

194 7]

SEIBERT HEVEA IN PERU

327

General Index

Ecology 265, 270, 294, 298, 303, 309-

313, 317, 323

Ethnobotany 303, 304

Geology 299, 301, 303, 311-313

Hybridization 286, 291, 293, 305, 311

Intergeneric 266, 312

Interspecific-262, 273, 296, 302, 313-323

camporum 265

Intraspccific

286, 323

Introgresslve 293, 296, 299, 308,

315, 320

Morphology 2 64

Buds and flowers

280

Acumination 281

Anthers 2 8 3

Calyx lobes 281

Color .-.- 2 8 1

Disk

282

Pubescence 2 8 1

Torus 2 8 2

Leajflccs 277

Lower Icpldote surfaces 277

Margins 279

Position

.„_.277

Pubescence 278

Shape and size 278

Texture

Tips

279

278

Short-shoots 270

Native names 265, 267, 293, 294, 298,

300, 304, 306, 310, 311, 314, 317, 318,

319, 320, 322

Rubber qualIty„...-267, 268, 316, 318, 319,

320, 322, 323

Taxonomy*^ 292

Caotitchotia elastic a 292

Hcvea 292

Benthamiana 276, 304

brasihensis 261, 276, 305

var. acreana, 306; var. angitstifoliay
305; var. cuneata, 294; mut. Gran-
thaml, 266; var. janc'nensiSy 306; X
guianensis var. lutea, 314; var. lati-
folia, 305; X pauciflora, 317; f ,
Kandiana, 306; var. Rand tana, 306;
var. stylosa, 305; f. suhconcolor,

ca

iicho

292

colUna 292

con f us a 300

cuneata

Foxii

294

319

glabrescens 319, 320

guianensis 292

var. collina, 292; var. cuneata, 292;
var. marginata, il9 ; ssp, occidcntalis,
292; \ar. occidentalism 292; X pauci-
flora, 318; ssp. typica, 292

. lutea 293

guianensis var

X Benthamiana, 319; X pauciflora,

4

320; X pauciflora X brasHiensis,
321; f, peruviana, 294

humilior 32 1

janeirensis 3 5

Kuntbiana _„ 305

lutea

293

var. cuneata, 29 A; f. pilostila, 294;
var. pilosula, 294; var. typica, 294

membranacea

300

f. leiogyne, 300; var. leiogyne, 300

microphylla 276, 285

minor 265

nigra . 292

nitida 276, 297

var. toxicodendroides 265

321

sa

paludo

pauciflora -^276, 300

ssp. coriacea, 300; yzr. coriacea, 300;
ssp. typica, 300

peruviana 293

Kandiana 3 05

rigidifoHa 276

Sieberi 3 5

Spruceana — 276, 285

viridis 297

var. toxicodendroides 265

Jatropha elastica 292

Siphonia apiculata, 293; brasihensis, 305;
brevifolia, 293; Cahicc/m, 292; elastica,
292; guianensis, 292; Ktinthiana, 305;
lutea, 29^; nitida, 297; paiiciflora, ?>Q0;

Ridley ana, 306

}Q6i Y:ir, subconcolor, 306; f. typica, Variation 263, 266, 270, 274, 276, 281,

306 284, 287, 295, 301, 302, 307, 314, 322

* Valid names are in Roman type, synonyms in italics.

["OL. 34, 1947]

328 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or Plate

PLATE 32

Fig. L llevea hrasiliensis. Terminal and lateral branch growth intervals or "flushes"
separated by a narrow ring of bud-scale scars.

}

Fig. 2. Hevca paiiciflora. Terminal and lateral branch growth intervals or "flushes",
separated by a conspicuous "interflush" short-shoot region. Note the conspicuous lateral
spur development preceding long-shoot or "flush" development.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 3 2

.oc->^«t .V. 1^"

CO

m

«

>

a

* *:

N>

tu--

[Vol. 34, 1947

330 ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXPLANAIION OF PlaTF

PLATE 3 3

Fig. I. llcica gi{Uiin'}ish. Terminal bud scales ^nA conspicuous dc\'clopmcnt of \}\q
sIiorL-slioot before the appearance of ihe inflorescence.

Fii;. 2. llcica Spnicaiua, Terminal Inid scales and the short-shoot from which the
ianorescence arises.

Fig. 3. lici'Ca ^liiiiuoish. Three "interflush" short-shoots, from the youngest of
which the inflorescence arises. Note the two "flush'' regions, upon the upper of wliicli
the leaves remain persistent even after the appearance of the new inflorescence.

•

Ann. Alo. Bor. Gard., Vol. 34, 1947

Plate 3 3

^j ^. 1.' ■ ■ 1 L,i

x^'

in

X

<
tn

>

C

K>

y

■>

LVoL. 34, 19-47]

332 ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXPI AXATION OF Pla I b

PLATF H

Fig. 1. Iliii'ii hrcisilirjisis. Complete Ic.if dcfi)Ii.uion of tlic previous "flush'* before
the .ippoar.incc of the inflorcseenee ami new "flush".

lig. 2. lUii'd r'!^!ilif(^li(h Complete Icnf persisteiiee of the previous "flush" even
nfter the appearance and maturailon of the inflorescence.

AxN. Mo. BoT. Gard., Vol. 34, 1947

Plate 3 4

r

CO

>

5;

c

K)

->■- - --X^-S^i-tIj-

[Vol. 34, 1947]

334 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Pi, ate

PLATE 3 5

Fig. I. Ilncii ^liicincfisis. Staniinatc Inul, open and dissected flowers; pistillntc bud,
open and dissected flowers.

Fig. 2. lined i^nijfU}isis var. hifi'a, Staniinaic bud, open and dissected flowers;
pistillate bud, open and dissected flowers.

Ann. Mo. Bot. Gard., Vol, 34, 1947

Plati; 3 5

03

C

Ki

C^

rO

^

->a

^

-K)

^

LO.

'^:r^:-^--^- .':■

V .

• V

*-•,

.■ • S

* + ^ 4

*« -«

^ «.- .'

• ' t

^ B ^

V *-

#• *

V «

I »

■fc , ■

* ^

i #

> > 4

\ %

N _ ■

i i

• :-■

I *

+ V

J k

s- *

« #

»■■-■•>..

■ *'

_ f

t*^" *

1-:

[Vol. 34. 1047]

336 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or Plaie

PI. ATI" 3 6

Fii;. L llciCii Bciillhuniii/iii. St.iniin.uc Inul, tipcn and dissected flowers; pistilKue
bud, open and dissected flowers.

Fie. 2. llci'Cii SpnirciVhi. St;imiii.ite Inid, open and dissected flowers; pistillate bud,

oju-n .ind dissected flowers.

Ann. Mo. Bot. Card., Vot,. 34, 1947

Plate }6

C/5

L * J

5

X

<

>

Z

t * ■*

c

to

o.

a

rO

■yu

•iJk

3

-*L

■K)

-o.

-iN

-K5

[VuL. 34, 1947]

338 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PLATE \7

Fig. L llcvca [Hiucijlora. Staminatc hud, open ■;\r\d. dissected flowers; pistillate bud,
open and dissected flowers.

Fig. 2. Ilcvea fiifiila. Staminatc bud, open and dissected flowers; pistillate bud, open
and dissected flowers.

Ann. Mo. Box. Gard., Vol., 34, 1947

Plate 37

/ 1

^' Vi

*^^

L . A ■

■ffe

\J

» fr

■ 4

/

1. ^'^

V

• w

,-- . J'

^t

' t ^^ ■ »

^ ji

.- M-*J:;.;'Ar :^

*i

•_• :.^ >-::>y'

• r?

* «

* ^ • ' * -J

:.-.v;--';v;Xii.ii3

>.i

^^

-A^:

/£/^

bo

>

m
G

Q.

-+0

■y:.:

'V

-*■ •

4 d *

-.-.*-■■■; -v.;^:-: k-;"-*-: y-^j^^^^i^iz---

Vt

1.I

■ ' >

#■'.

/.-

■ *

* -

'. ^^

. ' » .-

"' * ■•_

« —

«l^

. .--^

^-

rO

r?

Q,

-Jv^

-u

-p

-ti^

-a

• - i

h-

•.-^■■* V« t

* ^

^ ■

' -

'^-.i

4 ^ %

"x:"

• '-'^

n^-t _*

.-■1 ■ .-

I .J

- - '• -

I

*"*

-^

« J

.* »

^ ^-

\«

^*_ ^

* * t ^

1 ^ .

# - > , ^ '

k\':.

x-'c <'-:v'

J

[Vol. 34. 194/]

340 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation or Pi.atl.

n ATF ^8

Fig. 1. Ucxca hrasilicnsis. Staminatc Inul, open mm\ dissected flowers; pistill.ue bud,
open and dissected flowers.

Fii;. 2. Ilcrca ri^idifalia, Stamin:ue Inid, open and dissected flnwers; pistillate Inid,
open and dissected flowers.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 3 8

:y5

<

>

12
C

[Vol, 34. 1947]

342 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation oi Plan:

PLATE 3 9

llctca tfiicrophyUa, StamliKitc bud, open and dissected flowers; pistill.uc bud, open
and dissected flowers.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 39

O

-Jo

01

-^

4 -

' * ^-"

- <

B

-t>,

-o

Ox

*SJ

<:

2:

C

+0

i* •

-0)

-Vo

- ■ •

•*. ' -

.1.'-

A _ f

'.■ \

- ^

• • -^ * ^

« 1 '

* -•

J ri

* _ ^

[Vol. 34, 1947J

344

ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXPI ANATTON Ol Pi. ATE

PL ATI: -10

Stamiii.uc buds above open flowers of ihc same species

Fig

Fig

Fig

Fig
Fig
Fig
Fig
Fig
Fig

. 1. i

rr

. 2. J

1 ■

. 3. J

. 4. i

r r

i %

. 5. i

r/

. 6. i

. 7. J

rr

. 8. i

A m

. 9. i

n

1 m

Spnici'ijfut.
Ihiucijlovit.

Bcull

hnnuuui.

jjiicrophylhi

ri^ijifoliii.

Ann. Mo. Bot. Card., Vol. 34, 1947

Plate 40

i^^

o

t-n

to

K3 -

Oo -

4:--

a

ui -

Cv-

I -

Oc

^ -

g

5

-*^- * . Y-

■'-::':>■ ^

4^

i^

,j"j

cyi

[Vol. 34. 1947J

346 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Pi ate

PLATE 41

Pistillate buds above open flowers of the same species

Fig. I. 11, ^uiaucnsis.

Fig, 2. //. Sprureafht.

Fig. 3. //. pauci flora

Fig. 4. //. fj/tiila.

Fig. 5. H. guiafH'Nsh van Infra

Fig. 6. //. Bcuthumiana,

Fig. 7. H. brauUcJish,

Fig. 8. //. microphy Ua.

Fig. 9. //. ri^hlijolia.

t

Ann. Mo. Bot. Card., Vol. 34, 1947

Plate 41

NJ

Ui

■ I

. r

> i

i .. f

^ _ ^ "■

_+_*

• J-;:'.'^-'

<■•-,-".

+ .* 1-

.1'

C/5

D3

I

>

t^-

c^j-

4^-

cy»-

o>-

en
C

00

^

3

3

■ _-D -

*■■■

■ t«

S^

4^

^ »

.- . ^

■- ^ 1

^L

^ ^ ^?

^1, -■ ^,

i-_^f-^^

1-- A

'i-i

vc ';

[Vol. 34. 19»ri

348

ANNALS OF THE MISSOURI BOTANICAL GARDEN

EXPI ANAIION or Pi ATE

PI ATI- 42

Male reproductive organs above female ()r>;ans of the same species

Fig. 1. H

IV 2.

Fig. 3.

Fig. 4.

Fig. 5.

rig. 6.

Fig. 8.

Fig. 9.

Sprucciifht.
piUiiijtora

fjifiJii.

^uianonh var. lufca

lii'fil/jannafJii.

hriisilictisjs.
fnicrophyllij,
ri^i ill folia.

Ann. Mo. Box. Gard., Vol. 34, 1947

Plate 42

Cn

CA

m

>

2;

to'-

C>J-

^-

Cn-

QNJ

B
B

(^

NJ

ON

O

Oq

^

[VuL. 34, 10471

350 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation of Plate

PL ATI- 4S

Fii;. L llci'ca nif'nhi. Terminal portion of two p;inicles, showing the terminal
pisillhue flower witli calyx lubes antl after calyx-lobe abscission. Note disk lubes at base
of ovary. Staminate flowers conspicuousl) show the calloused calyx lobe tips. Scale in

nnlhmeters.

rig. 2. Ilriru nihyophylUi. Terminal portion of two panicles, showing the terminal
plstilhite (lower with calyx lobes and after calyx-lobe abscission. Note the conspicuous
torus development in this species. Roth the staminate and pistillate flowers show the
acutely acuminate calyx lobes which are not calloused. Scale in millimeters.

Fig. 3. llcica rii^iJifaliit. Portions of panicles, showing contortion of the bud tips
Scale in millimeters.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Pi All; 43

K - I X -

m

>

m

a

tsi

» M

I I I I I I I I I ! M I I M M ! I I » I M M f ' ' ' ' • ' * ' ' • ' '

w

If

^

i

\

h H -.M ^

I

[VoT.. 34. 1947]

3 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Explanation oi Pi.ai

PI.ATi: 44

C li.ini;cs of natural li.ihirat on outskirts of Iquitos, Peru, m;ide by ninn iIiroui;]i clcar-
In^ ^""^^ pastLirini;. Hybrid swarms of 11. hrasilicfisis X fnuiciflora and //. si^uianonh van
Intra X ihuiciflora occur in this typo of liabilat, representatives of which are shown. —
Photos by Dr, Richard livans Schultcs.

Ann. Mo. Box. Gard., Vol. 34, 1947

Plate 44

. - V- -

rr >4T1 ^r

■ ' - ^*'.'

i

'inif

}

SEIBERT— HEVEA IN PERU

Ann. Mo. lioi. (.Aim., Voi . M, \')47

Pi atf 45

U«1_ . . L J

■-^v^-W -i^

>^

Asclcpius tuhcrosa cl. fuhcroui-itjfrrior

Annals

of the

Missouri Botanical Garden

Vol. 34 NOVEMBER, 1947 No. 4

SOME DYNAMICS OF LEAF VARIATION IN ASCLEPIAS TUBEROSA

ROBERT E. WOODSON, JR.

I. Introduction

One day many years ago, as a young student, I paid my first visit to the
Smithsonian Institution in Washington. When noontime arrived, I was taken in
tow by my new friend, E. P. Killip, to the Smithsonian's unofficial "Lunch Mess"
in the kitchen of an old house around the corner. There I was awed by my in-
clusion within a jovial group of biologists previously known to me only by their
eminence. Killip announced me as a budding authority on the Asclepiadales, an
order of Flowering Plants including the milkweed genus, Asdepias. Instantly
the late Dr. F. V. Coville fixed me with a baleful glare and thundered: "All right,

■p,

Struck. I

knew A, tnhcrosa in the field about my home in St. Louis, but it had never occurred
to me that anything was "wrong" with it; and besides, T was having troubles of
my own with my dissertation topic, the exasperating genus Apocynutn.

At any rate, when my taxonomic studies finally brought me to Asdepias
shortly before the outbreak of the recent World War, I already was prepared to
find something "wrong" with A. tiiberosa. And I did. The species, as is usual in

the genus, is beautifully distinguished by sharply defined floral and vegetative
characters; it is easily keyed from its congeners. But within the species extra-
ordinary variation is rampant, particularly in the leaves. By overworking my
taxonomic intuition, at length I was able to distinguish three subspecies, which
went far toward resolving the difficulty. But I could not escape the knowledge
that something still was "wrong," particularly at the peripheries and commissures
of the subspecific distributions. Nevertheless, but for the outbreak of the war I
probably would have been content to let well enough sufFice.

Every taxonomist Is all too familiar with the professional handicaps imposed
by a world conflict. Even for those who fortunately are left at home, special
duties demand attention. Furthermore, necessary facilities for research are cur-

(353)

[Vol. 34

354 ANNALS OF THE MISSOURI BOTANICAL GARDEN

tailed, such as the exchange of authentic or type spcchncns. Nor can one over-
look the diflficulties of publication itself, thanks to wartime industrial disturbances.
Anticipating this prospect, I decided to take advantage of International catastrophe
by familiarizing myself with some of the more recent tools of biological systemat-
ics to the end of applying them to the special problems presented by Asclepias .

t liber osa.

II. Biology of the Species

Asclepias tuhcrosa is familiar to practically any one who is interested in wild
flowers from Sonora to Massachusetts and from Minnesota to Florida. Through-
out approximately the eastern half of the United States it is a common roadside
plant, conspicuous to any passer-by because of its clusters of stems about knee-
high, each surmouted in midsummer by showy trusses of smallish but intensely
brilliant orange, scarlet, or yellow flowers. The plants are long-lived perennials of
easy culture, and are prized by many horticulturists because of their dcpendabiUty,
long season of bloom, and dramatic dashes of color.

Not the least interesting feature of the flowers is their apparently irresistible
appeal to insects, particularly Hymcnoptcra, which are their chief pollinating
agents. Butterflies, as well, arc almost constantly hovering above blooming plants,
and are responsible for their most familiar popular name of butterflyweed. Coral-
weed also Is an appropriate name for them. Fortunately less familiar are the
names plcurisyroot and chiggcrweed.

VEGETATIVE HABIT

A fully developed butterflywc^^d usually is a rather massive plant. The
perennial portion consists of a w^oody tap root as much as three feet long and eight
inches In circumference, surmounted just below ground level by a tightly branch-
ing crown from which a few to as many as a hundred herbaceous flowering stems
may arise each year. Plants attain blooming age two to three years after germina-
tion of the seed and may persist for as long as twenty years or more. It would be
difficult to estimate the age of a large plant upon a single examination because of
the numerous stems produced each season and their crowding at the crown.

Some botanists would refer to the perennial plant body which has just been
described as a "clon," and the term has been applied to the essentially similar
structure in the Viorna section of Clematis (Erickson, 1945). It should be borne
in mind that in buttcrflyweeds, as in other similar plants, the communities of
clustered stems have no greater degree of Individuality than have the separate
twigs of a tree. They are connected organically to the same tap root, and are
ramifications of a single embryonic plumule. In buttcrflyweeds there are no
vegctativcly reproducing stolons, rhizomes, gemmiferous roots, nor other special
propagulae. It is conceivable that accidental or purposeful operation might result
in the successful division of the crown into two or more parts, although I have not
been successful in the attempt. But such division in nature, if It occurs at all.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 355

must be contrary to the habit of the species, since I have observed no instance
amongst the several thousand hving plants which I have examined.

It so happens that in certain other species of Asclepias, as in the common A.
syr'iaca^ adventitious buds upon special gemmiferous roots habitually succeed in
multiplying single plants. These compose true clons such as those of Hepiero-
callis, Iris, RohiiiiUy and other spontaneously vegetative-propagating plants. To
call the plant of butterflyweed a clon in my opinion not only is misleading but
destroys the contrast of the vegetative propagation of such a species as A. syriaca.
It is difficult to imagine a plant which under some unusual circumstance might
not become divided into two or more. If a term is allowed to become a quibble,
its significance Is lost.

The herbaceous stems of butterflyweed range approximately from one to three
feet in height. They normally are unbranched save at the terminal region of in-
florescence, although occasional axillary branches may be encountered most
frequently as the traumatic result of early decapitation. It must be emphasized
at this point that the herbaceous stems are produced at one time at the beginning
of each season, and are of essentially identical age. They are remarkably similar
in height and rate of development, as well as in number and relative size of parts.

^

These factors arc of obvious advantage in the random collection of leaves for
statistical analysis.

The stems of butterflyweed are determinate, ending in an umbclliform cyme
of approximately one to two dozen pedicellate flowers. In most fully developed
plants the terminal cyme is subtended by two or more leafy-bracted, scorploid
branches studded at the nodes with umbclliform cymules developing in acropetal
succession. These branches obviously are the homologues of the branches of a
dichasium. The determinate Inflorescence character of A. tuberosa is an anomaly
amongst the American species of Asclcplas, It Is an extremely fortunate one from
the standpoint of these Investigations, since it further facilitates the collection of
leaves of nearly identical physiological age, which would be a precarious operation
upon an indeterminate axis.

The leaves are simple, entire, and are Irregularly alternate or spirally arranged.
The number, shape, and size of leaves are extremely important characters In the
distinction of the subspecies, and will be discussed in later paragraphs of this
section. Obviously, the leaf variability Is such that It forms the subject matter
for these Investigations.

REPRODUCTIVE HABIT

i

The flowers of Asclepias are classical examples of entomophily and are equalled
in complexity only by those of the orchids. In the present connection It will

m

probably suffice to recall only those features Immediately concerning pollination and
the production of seed. In the center of a butterflyweed flower, as In other milk-
weeds, there arise five cornucopia-shaped bodies which are petalaceous, and are in
fact often mistaken for the corolla. These are the booils, the nectar-secreting

[Vol. 34

356 ANNALS OF THE MISSOURI BOTANICAL GARDEN

bodies which arc the goal of the insect visitors. The hoods actually are outgrowths
of the staminal filaments, the smaller anthers of which they virtually conceal.

The anthers themselves are closely connivcnt in the form of a cylinder or
truncated cone about the stigma. Each anther contains two pollen cavities. A
remarkable feature of milkweed pollen is that it is borne within small, flat bags,
or pollhiia, the confining membrane being derived from the tapetum. Still more
remarkable is the fact that the pollinia of adjacent anthers are joined together in
pairs by means of a delicate yoke-apparatus (tran^latot) surmounted by a pad-
lock-shaped body known as the corpu$cnlnm.

The corpusculum bears upon its outer face a more or less conspicuous cleft. It

is well known that pollination is initiated when a strong insect, such as a wasp,

accidentally thrusts its barbed legs between the anthers while scrambling about the

center of the flower in search of nectar. If a barb of the insect's leg wedges into
the cleft of the corpusculum, a stout pull of the member usually succeeds in

withdrawing the pair of pollinia from the anthers. It is a common sight to see

wasps or bees flying about a blooming buttcrflywecd, their legs laden with pollinia.

Robertson (1892) has enumerated 15 species of Lcpidoptera, Hymenoptera,
and Diptera collected while bearing on their bodies pollinia of Asclcpias tuhcrosa
in the neighborhood of Carlinville, Illinois. Amongst these is Apis nielli f era ^ the
common honey bee, which js known to have a flight radius of one-half rarely to
live miles. It is diflficuk to secure data concerning the radius of flight of the
other possible pollinators, although the phenomenal migrations, some hundreds of
miles, of certain Lepidoptcra are well known.

I do not wish to over-emphasize the efficacy' of the floral mechanism of
Asclcpias with regard to insect visits, since it does not appear to be very high.
Certainly less than 1 per cent of the flowers normally set fruit, except in A.
Incartiataj the swamp milkweed, in which sets may amount to 25 per cent. An
additional factor to recall^ in connection with insects as pollinators of Asclcpias^
is that the pollinia appear to be highly irritating to the carrier. On several oc-
casions I have trapped wasps of the genus Chlorion in transparent bags, when
they invariably appear to be more anxious to divest themselves of the pollinia than

to escape.

The stigma of the milkweed flower is surrounded by the connivcnt anthers

and is a rather complex structure. The receptive surface is not at the flat top, but
actually is divided into five concave surfaces which alternate with the anthers and
arc closely protected by them. In order that a stigmatic surface be pollinated, it
is necessary to introduce a pollinium between the flanged, cartilaginouSj marginal
Wings of the anthers, an extremely delicate and nerve-wracking operation for a
human experimenter. That the feat is accomplished at all by the chance move-
ments of an insect seems nothing short of marvellous, and that so few fruits

usually are observed upon a single plant is quite understandable.

Ann. Mo. Bot. Gard., Vol. 34, 1947

Plate 46

yA.i,/M/C''' ,

fn^

7.

1.

2.

C.-^fjAMx^^ 'fZoe-><Ji'nyi^'''^

Reproductive mechanism of Asclepias tuberosa: 1, flower; 2, gynostcgium with hoods removed,
showing connlvcnt anthers with one pollinium protruding; 3, two anthers as seen from within,
showing pollinia; 4, carpels with stigma head, showing two stigmatic surfaces; 5, Chlorion sp.
bearing pollinia on feet; 6, pollinia enlarged, attached to barb of insect's foot; 7, follicles with
comose seeds.

[Vol. 34

358

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Size and shape of polliiiia, corpuscuLi, anther wings, and stigmatic chambers
vary greatly amongst the species of Asclepias, Hcnse it is not surprising that

Aft

er over a

interspecific hybridization within the genus is very infrequent.

decade spent in studying It, I doubt whether I have observed many more than a

dozen instances of putative hybrids amongst the approximately 80 species.

The only record of a successful experimental cross between distinct species is
that of A. spcciosa X syr/aca^ first performed by Stevens (1945). In crosses in-
volving other species, Moore (1946a) occasionally observed preliminary swelling
of the ovary, followed by abortion. This he interprets as due to somatoplastic
steriHty. Although successful self-pollinations have been reported by several
authors for certain species of Asclepias (notably Stevens, 1945), similar experi-
ments by Moore (1946b) were unsuccessful. My friend F, K. Sparrow has told
me that his extensive pollinations with /I. syriaca reveal clonal self-sterility. My

I

2

3

4

1

Fig. 1. Somatic meiapKases from younj.; leaves of Asclcp/as inbcrusa In JiiTcrcnt
parts of its distribution: 1, Stillwater, Okla.; 2, Glcnd.ilc, Mo.; 3, Wcstport, Conn.;
4, Snail Lake, Minn. — All figures X 1500.

own limited experiments at self- and cross-pollinations involving A, tiibcrosa have
been notably futile. I suspect that special mechanical factors may account for the
difficulty in milkweed pollinations, such as the degree of desiccation of the tapetal
membrane, which must split to allow the emission of the pollen tubes.

Tlie fruit of butterfly weed consists of solitary, or infrequently paired, narrow,
fusiform follicles up to 15 cm. in length. Upon dehiscence as many as 100 com-
pressed, oval seeds approximately 0.4 cm. long are released, each provided with the
micropylar parachute of silky hairs so diligently studied and collected as a substi-
tute for kapok during the recent war. The silky parachute, technically called the
comay is extremely bouyant, and is doubtless capable of conveying the attached
seeds for long distances, even approximate estimates of which, however, are un-
available.

I have found during the course of progeny tests that viability of the seeds is
fairly high as a rule, but that germination is very irregular, proceeding in some
samples for well over a month. Although I am not yet ready to publish con-

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 359

elusions upon these tests, they indicate at the present writing that the various leaf
modifications encountered in different parts of the species' distribution have a
genetic basis and are maintained within sufficient Umits under cultivation in my
test plots.

CYTOGENETICS

Chromosome counts for Asclep/as ftiherosa made by various workers, notably
Moore (1946), agree in the figure n = \\, This is also the base number for all
other Asclepiadaceae investigated. Polyploidy has not been reported in Asclcpias,
although tetraploids occur in certain other genera, particularly in the Orient.
Somatic metaphases from young leaves of A, fuberosa in various parts of its range
are reproduced in fig. 1. The chromosomes are small (about 2 fx long) and rela-
tively uniform, and are poor subjects for configuration or structural studies.

ECOLOGY

In its geographical distribution from southern peninsular Florida to northern
Sonora and from the Ottawa River to the Black Hills of South Dakota, Asclepias
tiibcrosa demonstrates its wide climatic adaptablKty (Map I). It is clear from the
map, as well as substantiated by my field observations, that the species is best
adjusted to its environment in the mesothermal and southern microthermal
cUmatic regions of North America, Where it is found in the western steppe
climates it is probably as a scattered relic of former mesophytic times.

Something of the climatic preferences of butterflyweed can be deduced from
a comparison of the distribution map of the species with standard climatic charts

of North America.

J

temperature is critical in relation to density of population of the plants, a normal
mean surface temperature of 68° to 86° F. indicating roughly the optimal limits.
Precipitation appears to be more critical than temperature, however, suggesting
general requirements of an annual mean of over 20 inches. It is very striking,
when traveling eastward across Kansas in midsummer, to find the butterflyweed
suddenly emerge from more sheltered positions as a common roadside plant on the
outskirts of Topeka, near the boundary of the tall-grass prairies and the short-
grass plains.

In the opinion of some, Asclep/as tnberosa should be considered as a prairie type,
or even as an emigrant from the dry plains or deserts of northern Mexico. To my
mind it accords more closely with the facts of distribution to regard it as indig-
enous to the glades and open woodlands of the southern hardwood forests, from
which it has spread to the southeastern longleaf, loblolly, and slash pine forests,
and to the southwestern pinyon-juniper-yellow pine woodlands. When one con-
siders that approximately the northeastern third of the species' present distribution
has been available for plant colonization only since the Pleistocene, it is clear that
estabUshment of populations is taking place to the northeast much more rapidly
than to the southwest.

[Vol. 34

360 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Butterflywccd is found at elevations from near sca-level to about 6000 feet
altitude. In the western states progressive desiccation of tKe plains appears vir-
tually upon the point of eliminating the species except upon the well-watered
highlands. In the plains the plant must seek the protection of ravines and canyons,
or larger neighboring plants.

The wide distribution of A. tnhcrosa bespeaks its tolerance for a broad range
of edaphic conditions. Horticulturists are in the habit of prescribing for It a
sandy soil, upon which it undoubtedly flourishes, but scarcely more so than upon
the exceptionally tight clays known as "crawfish soil.'* Although I have made no
attempt to determine its pH preferences, they probably are circumneutral since I
have found fine colonics of plants growing impartially upon dolomitic limestone,
argillaceous shale, and granitic dctritis. An unusually large colony was found on
an alkaline flat in western Texas, but the plants appeared depauperate. Good soil
drainage appears essential for optimal growth, and a "poor" soil is generally
preferable to a "rich" one. I do not remember ever having found it growing on

alluvium.

Butterflywccd is found occasionally as single plants, but more frequently in

col

Distance between

colonies varies notably with respect to local climatic and vcgetational conditions.
Roughly east of the 96th meridian the species is a common roadside plant, as has
been discussed before. Through eastern Kansas and Missouri the colonics may be
encountered at intervals averaging about ten miles apart.

Through the prairies of south-central Illinois, on the other hand, I was able to
find only three colonies along the route of about 150 miles between St. Louis and
the Wabash River. From southern Indiana to West Virginia the colonies become
somewhat more frequent, and reach their greatest frequency, somewhat more than

t Virginia to the coast of Virginia. The brief account of
this transect summarizes observations made during a collecting trip in the summer
of 1946. Statistical details will be presented in a subsequent chapter.

Roughly west of the 96th meridian butterflyweed is scarcely ever encountered
along roadsides through the Great Plains. The plants are found, if at all, only

We

requent

The size

of colonics IS considerably smaller than Is customary in the East, and the distances
between them Is far greater. It is only In the pinyon-juniper-yellow pine high-
lands of the southwestern states that butterflyweed again becomes a fairly frequent?

plant.

'"I believe that colonics usually develop from seedlings of a single parent plant.
This is Indicated not only by their considerable degree of isolation, but by the
centering of leaf variation amongst individual plants about more or less distinctive

colonial means for the various characters measured.

3-

O

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

361

2 §

o
o

"I

< o'

2.^

x«

^

i/>

&i

I

o
n

CM

o

3

rt

V)

^ CTQ

o
d

o

a-

3
5

3

• •

>^

i

362 ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

THE THRHE SUBSPECIES

Ordinary lierbarium metliods disclosed at tlic outset of these investigations the
presence of three subspecies of Asclepias tuhcrosa (Woodson, 1944): A, t.
tnherosa, with leaves typically obovate to linear-obLmccoLate, the base usually
cuneate or rounded; A. t, interior, with leaves typically ovate to ovate-lanceolate,
usually with cordate or truncate base; and A. /. Rolfsii, with leaves essentially as
in ssp. tubcrosa but predominantly with more or less conspicuous hastate or cordate
dilation toward the base and with the margins more or less crisped. Map I shows
the known distribution of the three subspecies and indicates the probability that
the centers of modern dispersal, if not of actual origin, may be regarded as the
Paleozoic land masses Appalachia and O/arkia, and the early Me/ozoic "Orange
Island," now north-central Florida, respectively.

It is difficult to indicate with such a map the variation and intergradation of
taxonomic units, since in this case only three types of symbol are employed. In
addition to the symbols for the separate subspecies, however, a fourth, the hollow
circle, has been introduced to indicate intergradations of A. /. tubcrosa and A. /.

yinb

//

Rolf

As a matter of fact, such is far from being true. The distribution of hollow
circles suggests that in southern Alabama and southern Georgia one might expect
to find intergradations of all three subspecies. This actually appears to be the
case, and the practical limits of such a map stand revealed. A fifth symbol is not
used for intergradations of Rolfsii simply because I cannot distinguish the separate
roles of the three subspecies satisfactorily.

Almost the first glance at Map I will show that although ssp. tubcrosa^ indi-
cated by small dots, is distributed roughly from the western Appalachian foreland
to the coast, with ifiterior, indicated by large dots, to the west and north, the
hollow circles, which indicate subspeclfic intergradcs, extend from the commissure
of the subspecies completely through the distribution of ssp. tubcrosa. It is clear

even from routine examination of herbarium specimens that tubcrosa, situated
unstratcgically between interior and the sea, is in the process of genetic dissolution.

This situation appeared so interesting that a biometric study of natural and
artificial populations of A. tubcrosa was begun early In 1942. Begun as a side-line
to my more orthodox systematic studies, my hobby soon grew to occupy the
majority of my research. So many additional topics of interest arise as time passes,
and so many new lines of attack suggest themselves, that the study might possibly
be continued as long, if scarcely so profitably, as the classic investigations of
Sumner on feromyscns.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 363

III. Exploratory Methods

The disadvantages of the type of distribution map which has just been pre-
sented in the previous chapter are the simple consequence of the use of a discon-
tinuous scale with too few intervals. Nevertheless I do not wish to minimize its
use. In dealing with a genus of many species, such as Asclepias, it is ordinarily
the only type practicable. It at least states the known range of one or more
taxonomic units and perhaps suggests the region of any intergradation. To an
imaginative mind one of its virtues may be that it asks more questions than it

answers.

When a continuous scale is available, however, measurements of a large series
of specimens may allow the accumulation of equally spaced means and their ac-
companying measures of variability, and the result is a "phcnocontour map" such
as that advocated by Huxley (193 8). As yet few phenocontour maps have been
published. Possibly the most familiar examples are those of Alpatov (1929) for
Apis mcllifcray the common honey bee, in European Russia. These, in my opinion,
suffer chiefly because of the relatively few and irregularly distributed locaHties In-
to which the many samples fall. A better instance Is provided by Pearson's (193 8)
geographic study of melanism in the Tasmanian bush opossum, TricJjosiirus
viilpecula fiiliginosusy In which definite contour lines ("isophenes") were obtained.
In a class by itself is the world chart of human blood groups presented by Boyd

(1939).

From a phenocontour map benefits may accrue from several directions. The

biogeographcr may plot with a degree of mathematical precision the migrations
and environmental adjustments of which he now speaks in more general terms.
The cytogeneticist may be given wholesale data of population dynamics upon
which to apply a gamut of attractive theory much of which still requires ex-
emplificatlon. Last, but not least, opportunity finally is given the customarily
Inarticulate systematist to prove the detail of his observations to a disbcUeving
world. The inexorable accumulation of specimens to catalogue for the benefit of
others may allow but one such opportunity, and it should be taken.

USE OF THE HERBARIUM FOR POPULATION STUDIES

There can be Uttle doubt that systematic botany has contributed far less to
recent advances In the study of evolution than has systematic zoology, and one
reason for this has been its neglect of modern statistical methods. Yet for ready
collection of data, ease of manipulation, and wealth of museum material, plants In
general are much more favorable subjects for study than are animals. No one has
appreciated these advantages more than did Charles Darwin, who bequeathed a
part of his estate for the founding of the Tndex Kewensis,' the herbarium tax-
onomist's most indispensable single tool.

Herbarium specimens are not a perfect substitute for living plants, but they
offer incalculable advantages for the interpretation of field studies. Not all
plant materials are adequately preserved by the usual herbarium methods of press-

[Vol. 34

364 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ing and drying, but an astonishing percentage is. In the herbarium of the Missouri
Botanical Garden is a collection of several hundred dried plants assembled by one
Gcorg Rudolph Boehmer as material for his 'Florae Lipsiae Indigena/ published in
1750. The appearance of some of these plants is almost as though they had been
pressed and dried but a few months ago. All arc perfectly recognizable, and we

w

have In them a detailed record of the distribution of plants around Leipsic two
centuries ago.

In the early days of botany It was customary to have represented in the
herbarium only one or two specimens of each species. Nowadays any of the
several major herbaria of the world may include a thousand specimens of a single
widespread species from all parts of its range, gathered in all stages of its growth
at all times of year by faithful collectors, living and dead, for at least a century.
When several such collections of a group of plants are united for one's study, the
very mass of it is most Impressive. The accumulation of material so representative
of variation in time and space obviously is beyond the powers of an Individual. It
is a unique evolutionary heritage.

A false impression widely current among non-taxonomlsts is to the effect that
herbarium specimens usually are collected because of some abnormality which
attracts the fancy of the collector. The accusation reveals such prejudice that one

4

IS baffled for an effective retort. Perhaps a denial that falls far short of revealing
the sincerity of a plant collector but may Impress the critical outsider is the fact
that plant collections habitually are made In multlplicate sets bearing identical
serial numbers for purposes of sale or exchange amongst the numerous botanical
institutions of the world. It would be difficult Indeed for even an unusually
perverse individual to pursue his passion under such adverse circumstances. The
chief danger in plant collecting actually is that of choosing too many "normal"

1

specimens. The statistical errors from such likelihood, however, should be in-
effective.

"Mass collections," which I prefer to call "local population samples," have
been advocated recently by Anderson (1941) and others as an aid to the solution
of certain systematic and cytogenetic problems. This method of sampling local
variation surely is a very useful one, and a tool which I have used in part in my
own work. Certain attendant disadvantages should be discussed, however.

It is known, for example, that the phcnotypic responses to the fluctuations of
climate may vary from year to year in a given place, as Lewis (1947) recently has
shown with respect to Delphinhim, Population statistics obtained from a given
locality for a given year may not be compared safely with samples from other
localities at another time, perhaps even during the same year. Employing only
local population samples, the task of effectively covering the entire range of a
single widespread species, in the case of Asclcpias tubcrosa about 1,500,000 square
miles, assumes fantastic proportions.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 365

The most reliable statistics concerning plant populations over a wide area must
be made not during a single season but over a span of years, and the samples must
be randomly selected and as widely distributed as only generations of differently
tempered naturaUsts can accumulate them. It is possible to meet these require-
ments only by the use of herbarium collections. It is not a new convenience:
plant taxonomists have been enjoying it since long before the birth of Linnaeus.

The advocate of ''mass collections" may retort that even though herbarium
samples may cover distribution in time and space more effectively, the samples
are smaller at best than those specially made by his methods. The statistical fallacy
of this argument is obvious. The reader should not infer that I am condemning
the use of local population samples. I am merely attempting to point out what I
consider to be their limitations and to defend my use of herbarium specimens. I
have obtained quite interesting results through using both methods coordinately.

Perhaps this will be an appropriate place to caution firmly against the incon-
siderate use of herbarium material for statistical work. Being an herbarium cus-
todian myself I can anticipate the angry protests which will arise from my
colleagues at the prospect of their precious charges being plucked, petal by petal,
or for that matter, leaf by leaf, by increasing numbers of ''biosystcmatists." We
may not treat herbarium specimens as we would living plants having the power of
regeneration of lost parts. Type or other authentic specimens must remain sacro-
sanct for the use of posterity. Only abundant parts of herbarium specimens should
be sampled in studies such as these, and then only by an experienced student upon
express permission of the proper authority.

Being an herbarium man, my first impulse in beginning my study of leaf vari-
ation in Asclepias tnberosa was to turn to the herbarium of the Missouri Botanical
Garden where I am employed. There I found several hundred sheets of specimens
representative of the entire range of the species. Nearly all the sheets I found to
bear at least one entire stem of the plant. The leaves of each, habitually numerous,
I found to be well preserved, being somewhat leathery in texture. Since they were
so abundant, I discovered that at least some had escaped being glued to the paper
and could be removed without appreciably damaging the specimen.

After experiment I adopted the procedure of selecting a "random" leaf from
about the middle of a single flowering stem of each specimen, if it included more
than one stem, and recording the locahty (state and county for reasons which will
develop) from whence it came. I soon found also that it is desirable to keep a
record of the various collectors' numbers in order to avoid statistical bias from
measurement of duplicated specimens so prevalent in large herbaria. This entails
no unusual inconvenience, being a common monographic practice. The leaves
were boiled in water, separately, until completely exhausted of air. Their outlines
were then traced upon paper with a sharp 2-H lead drawing pencil, using an
illuminated tracing table.

366 ANNALS OF THE MISSOURI

[Vol. 34

For tracing, T have found best adapted to my needs the large sheets of milli-
meter grid paper printed by KeuflFcl & Esscr of New York. The 5- and 10-milli-
mcter lines are specially accentuated on this paper, which facilitates both
orientation of the tracings and subsequent measurement. The grid paper may be

protected from the moistened leaf by placing between them a small piece of wax
paper.

To some, I might appear to have been more scientific had I used some photo-
graphic means of transferring the leaf outUncs. There are several advantages of
the tracings, the first of which is the millimeter grid itself, special advantages of
which will be Indicated. A second advantage is that the tracings can be added to
the large grid sheets consecutively as they accumulate, assigning a special sheet to
each local population sample, or particular geographic division. In short, the
tracings appeared to be more convenient and photographs or blueprints unneces-
sarily complicated and time-consuming. In measuring smaller objects, greater
accuracy doubtless would be obtained by the latter methods. The degree of ac-
curacy obtained through tracing will be discussed presently; I believe that it will
be found sufficient for my needs.

QUANTITATIVE MEASURES

It has been explained in previous paragraplis that the chief, if not the only,
criteria distinguishing the three subspecies of A. fuberosa are found in differences
of leaf shape. Size, as measured by median length and width, is of minor taxo-
nomic Importance except perhaps in providing abstract universe values. This is
usually true because size, although it has a genetic basis, Is more directly influenced
by environment and age of organism than is shape.

It is easily seen In fig. 2, which consists of very small random samples of leaves,
that leaves of ssp. tnbcrosa may be said to be of greatest average width and those
of ssp. inferior of greatest average length, while those of ssp. Rolfsii average least
in both regards. But an experienced plant systcmatist probably would prefer to
point out that the leaves of tubcrosa tend to be broadest above the middle (obovate
to oblanceolate) with rounded or tapered base, those of interior broadest below the
middle (lanceolate to ovate-lanceolate), usually with 2-lobcd (cordate) base, and

If

and with crisped margins.

Determined to find a suitable continuous scale with which to measure leaf shape
Asclepias tuberosa, I finally recoi^nlzed that the differences in shape, at least

nor

differences In width at two points, at approximately one-quarter the median length
from the apex and the same distance from the base, respectively. It was obvious
that in tubcrosa the greater width typically is at the upper quarter and in interior
at the lower quarter. The intergrades usually are found to bear essentially oblong
or elliptic leaves, their widths at the upper and lower quarters being approximately
equal.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

367

For some time thereafter I measured leaf widths at exactly tlie upper and the
lower quarters of median length, hoping that by the ratios T might be able to dis-
tinguish quantitatively the leaves of the two subspecies with respect to unity as
represented by the oblong or elliptic leaves of intergradcs. This activity ceased
when I realized that ratios such as those I was employing are functions of length.
Another disillusioning discovery was that both ratios and their reciprocals afford
warped scales, and furthermore the warping of the two scales Is unequal. I mention
these points for the consideration of any who may contemplate using similar ratios
in biometric studies.

V;

A. i. -tuberosa

A. t. interior

A. t. Roifsii

Fig. 2. Representative leaf types of three subspecies of Asclepias tuberosa.

The methods which I finally adopted for measuring shape are illustrated in
fig. 3. In the diagram to the left of the figure, AOB represents the frame by
which leaf outlines are oriented at equal distances upon my large grid sheets. The
base of the leaf blade, where it joins the petiole, is placed at O and the tip verti-
cally above in the position designated as B. After tracing the outline about the
frame AOB, it is a very easy matter, since 1-, 5-, and 10-mm. distances are indi-
cated on the grid itself, to measure median length to the nearest millimeter as the
distance OB, rounding to the nearest even figure. Median width, MM\ is measured
also to the nearest millimeter, at the point midway between O and Bj rounding as

[Vol. 34

368

ANNALS OF THE MISSOURI BOTANICAL GARDEN

before. It appears unnecessary to read these distances to fractions of a millimeter
in the light of what we shall learn concerning the experimental error involved.

It will be recalled that leaf shape differences of the two subspecies are found
not only in the relative width of the blade at the upper and lower quarter lengths
but also In the fact that the blade predominantly is cuneate in A. t. tnbcrosa (fig,
3, center figure) and cordate In A. /. interior (fig. 3, right figure). To measure
cither character it Is found desirable to draw a horizontal line at right angles to
the median line OB at one quarter the distance of its length from the apex fXX'J
and base (YY')^ respectively.

p-'r I ■

I '' V . '.A ■-■Mi-}

T"

i:i'l::J

Fig. 3. Methods for

lit"

uring leaf shape in buttcrflywccds. — Explanation in the text

In order to measure the direction and extent of apical taper of the leaves two
chords now are drawn, XY and X'Y'. Using a standard protractor the two angles
XYY* andX'Y'Y are measured to the nearest degree, rounding to the nearest even
figure, and their mean is entered as the statistic hereafter to be known as /A.

In fig. 3 it will be seen that measurements of Z A are 94.5" and 85° (treated
as 8 5.0° in computations) for A. /. fubcrosa and A, t. inferior respectively. By
use of this type of measure, a plant taxonomist will recognize that a leaf of ovate
type will always have an angular reading of less than 90*", a leaf of obovate type
one of more than 90°, and a leaf exactly of oblong type a reading of exactly 90'',

An angular measure to distinguish different types of leaf base In butterflywced
was obtained after It was recognized that in cordate leaves the tip of the basal lobe
usually occurs at a point about midway between O and Y on the one side, and
between O and Y' on the other. Consequently, T now drop two perpendiculars
PZ and P'Z' at points on YY^ midway between OB and the margin of the leaf as
indicated in the diagrams to the left of fig. 3. The angles ZOB and Z'OB are now

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 369

read by protractor as before, and the mean rendered as ZB, the measure of the
leaf base. In the diagrams to the center and right of fig. 3 it is seen that the ex-
ample for the cuneate base of ssp. tuberosa has an Z B value of 40° while the
equivalent value for the cordate base of ssp. interior is 110". A truncate base
would have an Z B of 90°. In computations these figures are recorded as though
significant to one decimal.

It will occur to the reader that the values Z A and ZB actually refer to hypo-
thetical leaf halves and thus are not real variables. Such is doubtless the case, but
their other virtues will probably save them from being condemned by the practical
statistician. In addition, my recording of the mean angles to one decimal in excess
of significant digits is open to criticism although I believe It will be allowed since
subsequent computations are limited to one decimal. In practice the process of

averaging, besides halving the scale intervals, has the effect of halving the experi-
mental error.

Although Z A and Z B have been very satisfactory as measures for shape in
A. /. tuberosa and A. t. interior^ I must confess that they are wholly inoperative
with respect to A. /, Rolfsii, This has been a great disappointment, although to
be anticipated since one could scarcely expect to differentiate three value ranges
and all their intermediates upon a linear scale. This is an unfortunate consequence
of the biological reality of Rolf sir. It might be possible to exclude Rolf sit from
the absolute comparison of tuberosa and interior and yet to contrast it with both
by means of an additional system of arbitrary scores, say for the crisped margin
of the leaf. This would mix a discrete with four continuous scales, however, and
the result probably would not be very helpful.

There can be little doubt that Rolfsii interbreeds with both tuberosa and
interior upon the southeastern coastal plain, and it would be interesting to be able
to measure the phenomenon accurately. It is possible to do so only indirectly ac-
cording to my methods. In further studies we shall not forget the role that Rolfsii
undoubtedly plays, particularly with regard to introgrcssion with the other two
subspecies, but we will be able to deal with it only through inference.

ESTIMATES OF EXPERIMENTAL ERROR

The methods just described at one time appeared so crude to me, compared to
the technical refinements of others, that I spent some effort in obtaining estimates
of the experimental error involved. Despite the risk of being tedious, I am report-
ing the results in some detail because I believe that exercises of this sort should be
published more frequently. As a matter of fact, I know of only one other pub-
lished estimate of experimental error of measurements used in population studies,
that of Sumner (1927), in which the author was chiefly interested in variation of
measurement amongst different observers.

In the summer of 1945 ten leaves of Asclepias tuberosa interior were collected
along a roadside near Valley Park, St. Louis County, Missouri. They were placed
in numbered envelopes and subsequently measured with respect to length, width,

[Vol. 34

370

ANNALS OF THE MISSOURI BOTANICAL GARDEN

^A and Z B, wliile in tlie fresh condition.^ After measurement, each leaf was
returned to its proper envelope and dried under pressure for a week. For ten con-
secutive days thereafter the leaves were drawn and measured, being returned to
their proper envelopes each time, but the order of the envelopes changed by shuf-
fling. After the tenth dry measurement the leaves were boiled separately and
measured a last time. In this experiment extra precaution was taken against un-
conscious bias in that I personally traced the outlines on each occasion, while my
friend Richard W. Holm performed the actual measurements independently.

Table I contrasts the measurements of the ten leaves in the fresh condition and
after having been boiled after drying. Table II presents the results of measuring
the ten leaves upon ten different occasions. The first is designed to show, as far
as this case Is concerned, how comparable statistics obtained from fresh leaves and
those from soaked herbarium specimens may be. The second is a gage of accuracy
in the tracing and measuring process Itself, and also provides something of a guide
to the statistics of Table I.

TABLE I

'MEASUREMENTS OF ASCLEPIAS TUBEROSA INTERIOR LEAVES WHEN FRESH AND AFTER

SOAKING IN BOILING WATER

(Means, standard deviations, and coeflficlcnts of variation; angles in degrees, length and width In millimeters)

N

ZA

i

ZB

Length

■

X

s

V

X

s

V

X

■

s

Fresh

Soaked

10

10

85.0
85.2

0.5

1

0.4

0.6
0.5

120.0
120.6

1.6
2.1

L3

1.7

77.3
76.1

1

1.4
1.3

Width

Since I have no similar exercises with which to compare, it is hard to evaluate
the results recorded in Table II. I was surprised indeed, however, when the error
appeared to be so small, in view of my rather crude instruments, ranging from
0.4° or 0.4 per cent for Z A to only 1.5° or 1.3 per cent for Z B. Tlie metric
error likewise appears to be small. In comparing Table I with the discussion of
Table II it is seen to be immaterial, as far as the two angles are concerned, whether
the leaves are measured fresh, dry, or soaked, since all three means for both lie
within the experimental error estimated in Table 11. In length and width, on the
other hand, the three means lie at distances greater than that provided for by the
estimate of error, particularly in width. This is disquieting, but there is no re-
course since it would be impossible to measure all leaves while fresh. Statistics of

I

length and width, however, will play a role subordinate to those for the two angles
in the studies which follow, since they are not important systematically.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

371

MEASUREMENTS OF TEN

TABLE II

DRIED LEAVES OF ASCLEPIAS TUBEROSA
ON TEN CONSECUTIVE DAYS

INTERIOR MEASURED

(Means, standard deviations, coefficients of variation, means of means, means of standard deviations, and

means of coefficients of variation; angles in degrees, length and width in millimeters)

ZA

ZB

Length

Width

Leaf

No.

X

s

1

V

X

5

V

X

s

V

X

s

V

1

85.2

0.2

0.2

118.9

1.2

1.0

76.0

0.0

0.0

20.0

0.0

0.0

2

85.2

0.4

0.5

117.2

2.0

1.7

75.2

0.4

0.5

18.0

0.0

0.0

3

85.1

0.3

0.4

119.1

1.9

1.6

75.9

0.4

0.5

18.6

0.5

2.7

4

85.3

0.3

0.4

121.3

1.4

1.2

75.4

0.5

0.7

19.0

0.0

0.0

5

85.4

0.4

0.5

119.6

2.7

2.2

75.6

0.5

0.7

18.7

0.5

2.7

6

85.2

0.6

0.7

114.0

1.2

1.0

73.2

0.4

0.5

19.0

0.0

0.0

7

85.3

0.4

0.5

117.2

1.1

0.9

74.0

0.0

0.0

18.1

0.3

1.7

8

85.1

0.2

0.2

120.6

1.3

1.1

77.1

0.3

0.4

19.1

0.3

1

1.6

9

85.1

0.5

0.6

113.9

1.6

1.4

75.3

0.5

0.7

18.1

1

0.3

1.7

10

85.1

0.3

0.4

121.8

1.0

0.8

72.6

0.8

1.1

18.0

0.0

0.0

1

85.2

04

0.4

I18.4

+

^-5

1-3

75.0

V

0.4

0-5

1

18.7

0.2

J.O

IV. The Phenocontours

TECHNIQUE OF MAPPING

Phenocontour mapping is such a recent biological technique that it may be
worth while to combine the account of my own practice with some general com-
ments. The subject may be divided into several considerations which impinge
upon one another so closely that they form a sort of continuum. In beginning an
investigation of this kind, the first thing to be done is to become famiUar with the
systematic morphology of the organism chosen for study. Without a clear under-
standing of the critical characters of the species, for example, much time may be
spent measuring size which might be spent more profitably measuring shape. In
plant subjects, recourse should be had to a large, well-ordered, general herbarium
where the problem in all likelihood can be viewed in perspective and plans made
for the most promising direction of attack.

Having selected a problem and noting the most advantageous direction of
attack, suitable biometric measures must be devised as the sine qua nan of all that
Is to follow. A measure must be found which expresses numerically the phenomena
judged as biologically most significant. Any given measure will only infrequently
be found effective for more than the organism for which it was invented. My
method of measuring leaf base in Asclepias fiiherosaj for example, may be quite
useless in measuring that of another species.

A good measure should be duplicable, sensitive to organic variation, and should
provide an unwarped scale. These are rather complex attributes to discuss briefly.

[Vol. 34

372 ANNALS OF THE MISSOURI BOTANICAL GARDEN

They discriminate, in my opinion, against various types of discontinuous scales
encountered in arbitrary scoring. In some cases, as in presence-or-absence criteria,
scoring is the only sensible procedure. In others, the varying characters may be so
complex that scores at first would appear as the only recourse. But scores almost
Inevitably are the product of the personal equation and should be used only when
standard scales, such as the linear and the angular, arc unavailing. The statistical
advantages of fixed continuous scales are expressed most succinctly by Miss Walker
(1943) : "In order to know how much of a trait an individual has or to say that
one thing Is twice another, it is necessary not only to have equality of units but
also to establish a zero point. Only when these two conditions are met can scores
properly be spoken of as measures."

The area chosen for mapping preferably should include the entire range of the
species or other taxonomic unit under Investigation. This sometimes will be very
large and impose considerable handicap in the gathering of data. But as a general
rule the biological interest will be proportional to the area because of the relative
number of factors allowed to operate. When I commenced this study of butter-
flywced several years ago, I seriously considered following a suggestion of Huxley
and confining my efforts to a single profile across the distribution of the species.
Luckily my taxonomic training insisted that the distribution be treated as a
whole; several unsuspected topics of interest emerged as a result.

Equal to the Importance of adequate measures is that of adequate sampling.
The prime requisite of sampling Is that it be random. In biometric work of this
kind, one cannot use the word "random" in quite the same sense in which it is
employed in ordinary statistics; one cannot make use, for example, of the pub-
lished lists of random numbers commonly employed in sampling. It is necessary

to avoid tlie selection of cases which exhibit injury or manifest growth abnormal-
ity. It is necessary also to select cases at equivalent stages of development; there-
fore I have selected leaves from about the middle of a flowering stem rather than
leaves at a given node from the top or bottom of the stem, for some stems produce
a larger number of nodes during the course of their development than do others.
Above all, one must not allow himself to select what he chooses to call "typical'*
cases in the hope of deriving therefrom the benefits of random sampling.

It does not seem quite possible to deal with questions of bias here in the ordinary
way. In my own studies I think of bias as being intellectual, accidental, or bio-
logical. The Intellectual bias Is understood easily as the more or less subconscious
desire to vindicate a predisposition. Perhaps its best antidote Is to remember that
truth may be stranger than fiction. Accidental bias may be occasioned by the
paucity of specimens available for study In a given population. If herbarium
specimens are in use It may be occasioned by duplicated specimens frequently en-
countered. Protection Is taken against this by keeping a list of collectors* num-
bers as in ordinary monographic work. A variety of other sources of accidental
bias come to mind.

1947]

WOODSON LEAF VARIATION IT# ASCLEPIAS TUBEROSA 373

Biological bias is a phenomenon which is less easy for a mathematical statis-
tician to anticipate. In plants, if a species forms true clons it may be difficult if
not impossible to tell whether one actually is sampling a number of genetic indi-
viduals or offshoots from a single plant. Fluctuations of climate are known to
have a pronounced effect upon phenotypic expression (Lewis, 1947), and a
sampling made during any given season or year may be biased as a result. Sampling
made for convenience along roadsides or in occupied areas may present a very
special bias (Wicgand, 193 5). On first consideration It might be thought possible
to escape physiological bias if sampling of an organism were made at random stages
of its seasonal growth. But if any sample is allowed to consist chiefly of cases
collected during a given stage of development, it may be biased with respect to
others made during another stage. I have limited my sampling of butterflyweed
to plants in full anthesis.

The number of samples and number of cases Included is secondary In im-
portance to the degree of randomness obtained. However, it may be worth while
to emphasize that an adequate sampling is dependent more upon the number and
randomness of samples than upon the number of included cases. Of course I do
not overlook the fact that reliability of means and their derived statistical measures
of variabihty increases as a rule with sample size; but size alone Is not an indica-
tion of randomness and hence of reHability. Adequacy of sample size is determined
by the unique degree of variability of each organism and can be determined In
h case only after special observation.

Adequate sampling of the vast distribution of such a species as Asclepias
tuberosa, an area of approximately 1,500,000 square miles, is clearly beyond the
power of a single individual since we require randomness of climate, time, and en-
vironment. In an earlier section of this paper I have explained how herbarium
collections would appear to satisfy these conditions. The argument may be ad-
vanced, however, that the relatively small numbers of specimens available in
herbaria is insufficient for an adequate sampling distribution. Superficially it
appears small indeed. Its adequacy with regard to butterflyweed may be judged
by fig. 7 of this paper, in which statistics of two sets of samples are compared:
one from the herbarium, consisting of 117 cases distributed along an approximately
1200-mile profile from Topcka, Kansas, to Norfolk, Virginia, and the other of
994 cases collected personally by myself and two friends in June, 1946, along
roadsides between those two cities. The close correspondence and consistency of
the two samplings are striking.

If I may extend my comments on sampling a bit further, I should like to call
attention again to the area inhabited by A. tuherosa: approximately 1,500,000
square miles. First impulse might be to obtain statistics from whatever source
and to combine them for the supposed benefits of larger samples. I am obliged
to confess that for this vast area I have been able to collect and measure only
approximately 12,000 cases in the time at my disposal, or about 1 per 115 square

{

[Vol. 34

374 ANNALS OF THE MISSOURI BOTANICAL GARDEN

miles. At first glance tKis appears inadequate indeed, and it minimizes the true
situation since the cases could not, for practical reasons, be distributed uniformly.
Actually my samples fall into three rough categories: cases obtained from hcrba-

miss

fashion throughout the whole species distribution, largely through the kindness of

interested friends.

Cases obtained from these three sources surely cannot be combined, since they
have been accumulated under different conditions, and their discrepant numbers
would constitute a serious bias. Consequently I have kept them separate although
using all for comparison according to the special values accrued from each. At
the risk of over-emphasis, I should like to repeat that of these three categories of
samples, I consider that obtained from herbarium material by far the most rep-
resentative biologically although they are also the smallest numerically, numbering
only about 3,000 cases, or somewhat less than 1 per 470 square miles of the specific
distribution. As shocking as this rati9 will appear, T believe the derived statistics,
on the whole, to be biologically reliable, and I have used them in constructing my
phcnocontour maps to the exclusion of other data.

The projection of data upon a phenocontour map is a rather complex matter
which depends upon such factors as amount of available statistics, size and char- .
actcr of the area involved, and nature of the information which it is desired to

convey. In his study of pelage melanism in the Tasmanlan bush opposum,
Trichosurm vtdpcniU fnlighwsiis, Pearson (1938) employed relative percentages
of two class scores, black and gray, although he mentions unmeasured variation in
both. By using commercial pelt records of approximately 105,000 cases distributed
amongst 48 more or less equally spaced stations, a ratio of about 5 cases per square
mile, he was able to plot a series of four contours ("isophenes'* of Huxley) ap-
proximately separating areas Including 0-25 per cent, 25-50 per cent, 50-75 per
cent, and 75-100 per cent of gray pelts. His conclusions are chiefly historical.

Few biologists will be able to equal the volume of Pearson's data. To be as
representative of distribution for Asclepias fiibcrosa, my records would have to

embrace over 7,000,000 cases, instead of the approximately 3,000 which I have!

Another advantage of Pearson's data Is the distribution of cases amongst 48 rather
equally spaced stations.

Readers famUiar with the composition of a general herbarium already are aware
that the geographical distribution of exsiccatae, even in the United States, is far
from uniform. Greatest concentrations of specimens occur as a rule about wxll-
cstabUshcd cities where botanists long since have resided. Next come states which
have undergone systematic botanical surveys, and there is a gratifying number of
these. Thirdly, there are regions of peculiar scenic or biological interests, such as
our national parks, which attract appreciable numbers of plant collectors. But we
cannot disregard areas, sometimes of considerable extent, where a lamentable hiatus

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 375

of herbarium records Is encountered. There is no fixed pattern to this mosaic,
and It presents the major obstacle In phenocontour mapping from herbarium
collections.

The projection of data upon a phenocontour map presumably should require
the imposition of equidistant statistics from equal areas. I have accomplished this
by dividing the species distribution Into quadrats of equal areas for which I have
combined the data secured from the various localities included within each. This,
of course, creates a system of artificial (in contrast to natural) populations from
which the desired statistics are computed.

The quadrat area chosen clearly depends upon the geographic plasticity of the
species and the nature of the information desired. If the species Is very responsive
to altitude or ecology, the size of the quadrat will need to be much smaller than
if the organism Is not so sensitive. This may be a very serious obstacle in mapping
a large area. Fortunately, Asclepias tuberosa appears to be rather indifferent in
these respects, and the size of the quadrat depends chiefly upon the necessity of
obtaining artificial populations equally distributed and yet of sufficient numbers
for statistical analysis. This amounts to a certain guise of "gerrymandering," but
is legitimate since the same quadrat area is employed throughout.

In this study of butterflywecd I have manipulated quadrat area to the end of
obtaining populations of at least five cases In critical but poorly collected regions.
The quadrats In this instance arc approximately 120 miles square, and there are 136
within the range of A, t. tuberosa and A. t, interior — many fewer than I would
wish. Even with these large areas, it will be seen that certain of them have failed

I

to yield as many as five cases. After recording the quadrat data, isophenes may
be drawn, if desired or possible, either connecting equal statistics or according to
arbitrary ranges, as practiced by Pearson.

The nature of statistics projected upon the map will vary with individual
problems and with the inclination of investigators. If scores are used, Pearson's
system, already explained, would appear admirable. If a continuous scale has been
employed in measurement, it will be natural to compute the mean; and from it
may be derived any of the familiar measures of dispersion such as the standard
deviation and the coefficient of variation. These may be entered on the map to-
gether with the mean. A measure of variability frequently will be necessary In
order to interpret the geographical distribution of means.

Of course there is a vast number of statistical formulae which may be used for
the analysis of biological measurement, and if one has a taste for mathematics the
possibilities are endless. In my own study, after some dalliance, I have limited
myself, as a rule, to the simple calculation of means, standard errors, standard
deviations, and coefficients of variation, the second and fourth largely for the sake
of convention. An outstanding example of another point of view is afforded by

[Vol. 34

376 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Czcczottowa's (1933) remarkably painstaking study of variation in beech leaves,
to which Dr. M. K. EKas kindly has called my attention.

In the phcnocontour maps which follow, quadrats containing five or more
cases arc indicated with the mean in large bold face type and the associated
standard deviation in large italics. Means of quadrats containing less than five
cases are printed in small Roman type. The standard deviation is used instead of
the coefficient of variation because there appears to be no need for percentage
comparison; also the former adapts itself more readily to my procedure, and the
latter appears too sensitive to the relative magnitudes of the means, in addition to
other disadvantages (cf. Kesteven, 1946).

INTERPRETATION OF TME MAPS

JNTRODUCriON

Before turning to the phcnocontour maps, it may be well to consider very
hastily certain details of the paleogeography of eastern North America which will
have a bearing on our interpretations of the population dynamics of Asclepias
fuherosa. The discussion is illustrated by Map II.

It is gcnci'ally recognized that many of the principal families of Flowering
Plants were established by the close of the Mcsozoic era, possibly before the Lower
Cretaceous. Although I know of no indubitable fossil remains of Asclepias,
numerous records of Late Cretaceous and Mcsozoic imprints, such as the form
genus Apocyfiopbyllum, are known which may well represent, at least in part, an-
cestors of our modern milkweeds, if not records of extant species. At any rate,
present distributions of many species of Asclepias correspond so closely to what is
known of Cretaceous geography that I feel wc may hypothesize rather safely much
the same spcciation in those times as that with which we are familiar at present

(cf. Woodson, 1947).

The Cretaceous has been called "the age of greatest submergence of the con-
tinents and the most extensive epeiric seas the Earth has known (Schuchert &
Dunbar, 1933)." The complex submergences and resulting isolation of floras
probably were of the utmost importance to the meteoric evolutionary diversifica-
tion of Angiosperms during this time, and are reflected in present speclation.

Late Cretaceous saw the climax of the dissection of North America with the
submergence of the Rocky Mountain trough from the Caribbean ta the Arctic.
This was accompanied by submergence of the southeastern coastal plain, partic-
ularly north of the Gulf of Mexico, a deep embay ment extending up the present
Mississippi valley as far as southern Illinois. This embayment separated the ancient
Appalachian and Ozark plateaus, including the extension of the latter to the Llano
uplift in central Texas, and is recalled in our present vegetation by numerous
vicarious species and subspecies. Amongst these may be mentioned Asclepias t.
luhcrosa and A. /. interior, respectively.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

377

/■K'

^n-^l

-^

r\

'J

.- s

■vto

\

» ' -.

W'^

lA"

:^-;^:

A

r -

Ti

.4,'

J

■^\

,-r.

•\

■^

T

J

.nr-

_.«

,r^-

-1

■S

J

^

-1

X"

^

I Tin

^

"^'^ i\ i

, V-r^*'^ Ij^

^Z^

vLA y

)"i\

H/

^

^

"N

-r

y

\

\

^'

^"

h T ■*■ k

1 r

A

-^

,i^

\

\^

Ul

&*

* *

' r

f^ '

^j

[/

/

-4

: x^iLiicHii,^

y

WviM.

1 ORANGE ISLAND

A^ V

>

.^

r -

1

.1

. \ .

Map II. Ozarkia, Appalachia, and Orange Island, with reference to the Cretaceous
and early Meso?,oic seas, and Pleistocene glaciation. Explanation in the text.

Of equal interest in this connection is the probably fluctuating emergence of
low islands in what is now northern Florida, culminating in the early Cenozoic in
the appearance of the more sizable "Orange Island," separated from the Georgian
coast by the Suwancc strait (Schuchcrt, 1935). On these islands possibly de-
veloped many or most of the Floridian endemics of Appalachian affinity, including
A. t, Rolfsii and numerous other milkweeds.

Withdrawal of the Cretaceous seas during the early Cenozoic effected the re-

union of the Appalachian and Ozarkian lands In Oligocene, and of emergent

Orange Island to the continent in Pliocene. By Pliocene, therefore, there appar-
ently were no geographic barriers to reunion of the disjunct distributions of the
three subspecies of Asclepias fiiherosa, if indeed they existed at that time.

Pleistocene brought the continental ice sheets approximately to the present
valleys of the Missouri and Ohio rivers, virtually to the head of the old Mississippi
cmbayment of Cretaceous and early Cenozoic times. This undoubtedly provided
a partial secondary separation of the putative ranges of A. t. hcberosa and A, t,
interior. It is well known that four glacial periods occurred during Pleistocene,
interspersed by warm intcrglacials longer, indeed, than the present day is removed
from the last withdrawal of the ice. The time since the retreat of the Wisconsin
ice sheet usually is reckoned at about 25,000 years.

378

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

K

4>

4>

c
o

n

C

X

>

NJ

E

t-l

3
O

a

o

d

a

■ *

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 379

These well-known facts arc outlined to give perspective for wKat we are to
find in the population dynamics of Asclepias tuberosa in recent times. They
underline the possibility that the phenomena which we will discuss are not unique
to the present, but perhaps the repetition of past and similar phenomena. It is dif-
ficult indeed to estimate the effect of such repetitions upon population genetics,
but at least they should not be disregarded.

The phenocontours for Z A which are reproduced in Map III are projected
according to quadrats 120 miles square, as has been explained previously, with the
means of quadrat populations including five or more cases printed in large bold-
face numerals and the associated standard deviations immediately below in large
Italics. Means of quadrat populations including less than five cases appear in small
Roman-type numerals. Three isophcnes have been drawn approximately sep-
arating bold-face means included within intervals of 2° from the Atlantic coast.
Statistical sources for the data used in compiling this map, as well as for the three
which follow, are presented in Table III.

It may be considered that the 90°-91° contour represents the range of "true**

O ^ ^

A. /. ticberosUy while that of "true" A. /. interior Is indicated by the 84 -8 5 con-
tour. The SS^-S?"" and the 86''-87° contours represent subspecific intermediates
except the populations in the far southwest and for the most part in the northern
Mississippi valley which, for reasons to be developed later, are regarded as affected
by intra-subspecific differentiation. Data for A, t. Rolfsii regrettably are absent
for reasons already explained.

After the preceding discussion of paleogeography, It is easy to identify the cen-
tral portions of the *'true" subspecies distributions as centering In general on the
Appalachian and Ozarkian-Llanoan headlands of Late Cretaceous time. Between
these centers a broad cline stretches from the Great Lakes to the Gulf of Mexico,
broader in the northern and abruptly constricted in the southern halves.

The most obvious explanation of this cline Is the hybridization of Individuals
of either subspecies at the commissure of their distributions with continuous back-
crossing to obtain the remarkably broad phenotypic gradient presented by the map.
This process has been termed "introgressive hybridization" by Anderson and
Hubricht (1938), shortened to "introgression" by some recent writers, and the
result is a genocline (Huxley, 1942).

The middle isophene apparently represents the mid-current of gene flow be-
tween the two subspecies. Its virtually straight course is striking, and possibly
would suggest nearly equal introgresslve pressure from east and west. The nearly
rectilinear nature of this isophene Is the only evidence that will be presented at
this time to support my assertion that the peculiar course of the western Isophene
Is due to Intra-subspecific forces operating within A, /. interior, and not to Intro-

380

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

<

<

I

<

a
o

H

w >

•-3
<

fti

og

^ O

f ^

O

CI.

o

E

-T3

C

C

a

C

o

c

a

o

8

a

u
ft

a

<4

2

u
ft

s

4-1

a

^

^

c

Nl

>

QO

X

>

i*rt

X

>

V^

-

fx

c/i

•X

Z

oo

CM

<N^^^l^r^l^^^'--^s,C0rxO^/^O00-^*^fv4(s^c3^T(■(N

'"fr^j'^O^ — *rs|ooOr^ooiryir>

r^ r^

f— '

K

a\

o

\D

ITS

irv

O

<^

f^

rj

|H^

(N

(N

<N

-^^

N

o\

Ooor^ocoo-Tf^Doot-Hc^r^ooi-Hr^cjnKxoO^-Hu^in

-^

FMOO^oor^rjoo-f-— -ir>r^aNOOKOM\OKK^£ooK

K

ooK>/^v^Oooro^^O|\OQ^KO\oo

l^ T^ r-^ l/> O N.

ON oo 00O0\^O0\'^'^«^w^KrN.^-H00<^\0a^i/%^mo

f^ ON ON-^^^O^SD^n^^^^CM^r^KOOr^fN-^KK^D

+1 II +1 +1 +1 +1 +1 +1 +1 II II II II II II II II II II II II II II II

oo ir\ ^ ^ 'y^ rr^r-^ r^ ^ t^ r^ t. oc rA

cc^O\^f^ir4\c^<z>oot^

rv

i^rS"^^*/^'^'i-^'^rn'i-ooi^r^-'^Ti-oo--nTj-r^*^rs|

II II II II II II II II II II II II II II II II II +1 II II II II II II

^1 '-'(N<N'-'^(N"'-^'-'-^(NOfs|'-^r4r^(NrM'— '(N

K

oooc>o'ooooo'oc>—-oooooo

II II II II II II II II II II II II II II II II II II II II II II II II

rvj

Cv*Ar^rf-oOfMoo^D<NrJO

'i'Oi^\cOOCTNCrsOO\£>SD\OSO^Cir>^OOi— iOOOOK^^i^^
OOQ\OOoQC7\(^OOOOOOOOOOOOOOOOOOOCa\0\a\QCOOoOOOOOOOOO

^r^ oo -^

o \o o

ON "O i-« 0\ ^O

f^ p*^ rr% ■— r\4

1^

rs

iiitiiii|iiiiiiil*lii|iitt

II II nil II II II II nil III! II II II II II II II II II II II II .

OooOK^OOOO^Tt-r-^OOOOiAOOSCO^^O'^r-^O'^f^OSirs

*

OC

*

IN

oo

4

OO

«

On
oo

OO

•
oo

*

K

OO

K

K
K

*

t ' L

oc

•

oc

*

*

*

00

107.

NO
00

■
OO

P

oo

P

OC

00

*

o

oo

15.4

OO

■

•

o

so

A

OO

OS

■

o

p

K

P

OG

ON

«

OO

rsi

P

o

ON

NO

#

o

*

o

14.7

•

00

4

ON

p*^

<N

p

00

1

<N

ON

#

00

m

OS

•

SO

•

K

oo

4

in

On

*

NO

4

o

oo

•

O

o

*

K

•

o

r4

^

NO

*

00

P

I

4

NO

•

-1-

«

95.2

40.9

84.7

100.0

48.4

4

00

I
NO

64.0

74.7

83.6

«

OO

84.9

91.3

97.3

NO

On

OO

m

3\

56.9

NO

no'

58.0

66.0

77.2

82.6

96.4

95.0

93.3

105.5

*

2.2

1.9

r4

2.3

1.9

2.0

2.8

<N

1.2

P

2.7

0.7

2.2

1.8

P

2.5

: 2.3

1

1.7

1.6

0.8

*

K

P

*

#

OO

P

OG

P

■

o

■

NO

o

P

NO

P

OO

4

«

o

P

*

O

P

P

P

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

381

o

w

<

^ ^"^

>

s

P5
Nl

<

>

t/5

IX

>

trt

X

>

trt

X

>

'»

X

2

i-i
rt

T|- '^OcJ\^H^^^r^Oooa^O(NO^D

O^r^oci-^OOooK

l/\

aNOOTtK^o<5oo--H(N|irrir\oooooo<Ni-^

CN

rJC^r^rvir^rsir4rsiH^rar^r^(Mf^CNJ<Nr^

0\ OO Tf Cn< N.

rsi r J pvj fM r^

r^

r^

IN

\D ■^'^Dl^rJh^^Olr^^v,OOOU^OOCT^OOOOOOOOCT^|Xl^N.

4

tN ^0

+1 +1 +1 +1 II +1 +1 +1 +1 +1 +1 +1 +1 II II II II II II II II II II II II

nil

-H(^lr^Kr^oOlr^u^r^|TJ-oa^^^l\DKT^^/^^r^^^,^£>^T^^^^r^J1-HT^^^

r4

^0<N'^f^^'— 'oo-^^u^i/^^ mr-HtJv^N'-^^-*!/^!— <0\rj

rs

ITS

(N

4 •

\r\

r^OoorN'^OooKKKooONTt-

rr^ (N fNi ri tnI rsi ^— - (N CN

o\

■ ■

nil

II nil II II II II II II II II II II II II II II II II II II II II II II

OO 4r^ f^ 00 »— (N

Oi^K-— 'cr\f^aN^oGoo^o^'*-o\

0\ OO rvi cc ^ -^ OO

f^ r^ <N <N <N f-H

<N *^ fN rs| i-H

<N

1— CO
K ON

00

\0 ir\

cj

Csl

SO

rsi

II II II II II II II II II II II II II II II II II II II II II II II II II

nil

l^'^0^^01/^O^^Ou^^v^^^^^^■^^\^^\^r^O'^•*^oOTt■ooc^^^^ll^f^^/^l/^ocT-^

r4

^ r^ r^ r^ — ^

!J\ o o K

^ I— l/^ O "^ O

c^i r4 IN .— i CM '— ' fs

(N ^

PnI

<N

r^ f^r>Jr^lli^oo1---r^l^f<^c^c^^l/^t^r^•^^'^'^f^t^*^^^■^ \r\

r^ rj

n n II II II n ii n ii ii n ii n n n ii n n n ii ii ii n ii ii

nn

r^^^O<»r\|/^c^'i■O^A^c^lrxoorxrN.u^oo^Olr^rn<N<Na^o^*^l^^

CCCOaNCN0O0000CCO00O0OOOCX:CT\0\C<)0O0O00O0D0O000OCCOoocXlclOC>Qa\0^

fN'^bv^OO<^000'^»-H

H-.

t*-i O ir^

rs»

\o

O 1— • rj f^

'y-\lii<» lli|IJIIIiltiitl<>

382

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

c3

<

-TO

e

u

03
Nl

<
Nl

>

IX

>

IX

>

iX

>

v»

l><

z

3

fNr^mr4roi-io4<N-^

ITS

^

^

oo

OO

4

o

4

K

■i-

lA

rO

*o

ro

■

OO

■

•

*

•

O

*

25.0

44.1

28.9

29.4

32.7

25.5

29.0

27.1

31.0

31.4

24.2

On

18.6

25.3

1

2.1

6.0

1

•

CO

4.5

5.4

.1 4.1

1

*

4.6

4.4

4.3

P

rO

1.3

r-i

OO

■

try

•

OO

o

■

CM

t

m

*

OO

*

■

rv

m

+111 +1+1 +111 +1+1+1

-^0-^r^0\a\rs|00irs0

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

+1 +1 +1

iriCMrxOooa\K*N

IN "C CO r^ OC \0 <S|

-<f so Cv od ^0 -^

On OO i\ T^

O K *n |\

O ^ ^

* m

O OO 90

CN —

CS

rO

|/^

rO

O

<N

rj

^

o

4

o

W

^0

■

CO

v^

o

r^i

(N

'"'

^^

*""

^^

f-H

■"^

""^

On

«

4

OO

o

*

OO

■

DC

*

^ fN K '^ •^ OO u^
« t « t * * ■

lA •^ ro ^o rs rx "^

^O ir^ 1-. O
* * ■ *

K rj o ^

rn

K ^-^

+1 +1+1 +1+1 +1+1 +1+1

roCTvXClTVTj-.— (N-fJ-

^ OO

1— 'fN-^^ ^—r^i^rsi

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

rs 00 "^

* * *

rj ir\ fo

+1 +1 +1

00

^ O O <N

\0 CO CTs OO

OO hv ^

N. OC r^

^ ^r^ ^

SO ^

ir\rr\— 'O'^^O "^(Ni/^-^ootsOsi^i^oo

K a^ •-^ 'A <N K "^O
<5 ^O K K OO K vo

^ K

<N

K

^

A

ITS

■

O

•

'^

■<*-

V

OO

ir>

U^

^B

CM

Csl

r^

fNl

■^

^^

OO

r^

^

fO

00

i/>

so

rO

o

#o

<N

4

a\

SD

^

■o

r^

c^i -^ o -"i- -^ o OO

NO

■^ OO K

P » #

CO to

r^ On 5s

OS

1-* CN (N

^ CS ^— • ^ ^ 06 *A

(N " ^

C^ O ^
— * rsi r^

u^ VO 1-^

— «N CN

1— I OO 00 -^
c^ ^ OS "-^

•M K -*

« • *

hv — rs

— r^ ^

+1 +1 +1 +1 +1 +1 +1 +1 +1

KlA^fNCsi^^ ^<vii-^CN

+1 +1 +1 +1 +1 +1 +1 +1+1 +1+1

-^ K ir*

4 • B

^ a\ ir\

+1 +1 +1

ooors'-<CTsiA'--rN\coirsOOSO'<f'-^^so(Nror^Ov^rNmo*rNO »-*ir.ooir^

r4 1— « OS IN

l^^ lA '^ OO On

(NfNSOOO#oK^<NONOOKOO O00^/^00f^^ro^£)
OOoo^Oooi/%ir\iAThioKhx OOOOS'-^oooo

hs CO OO

oo OC OS

rvKKOSiTNO'i-^C^

0\ ^ OO o O

f-H p-^ p-. rO CO

^ O -H rH

(N CNl (N C4

yy o\ G
rN o (N

irv "^ fN| "^ r^ o\ •— '

_i f*-i -i^ _( fsj O fN

rr\ K rrs Se K \C
',-;_; ^ (VJ <sl

K OO 00
(N '-' >— I "

^ 00 rx

# ^ »

rN

■i-l/NlNlOl/VrOSO^^

^ CO "^ CN 'O rs

-^ fo m

OO ^ ITN

+1 +1 +1 +1 +1 +1 +1 +1 +1

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

+1 +1 +1

OOi— ^ir\|\CT\K*^cNOiAO^'^^^'^"^^u^^'^*^<^iAOO^^— 'OO
OOO^oohNW^v^'^*ASOTt- fsl^OOOOOvtVoO */^^^^'^^^— iv^fs, i^^sDSD^

ffsOSOOOOOOOOOOOOOOoOoO OOOOOSC\C>.a\OOOGoO OOOOOOOOoOOOCpO OOOOOOoo

OO \0

1— « r^ CN fN On OO

F^ C^ r^ ,-1 r-.

rs4O0^v^*^^o^^^<N

1— . CN ^^ fN f— « »-^

rx so OO XO <N

rO «> "^ --^ -^

ly-N so rx OO Ov

rj -o "^ OO ^ o

rs| ^ trs ^ K OC OS

'— '<Nco'1"*/^sorN.ooOsO

66 6 666 66 6 66 6 6XXXXXXXXXXXXXXXXXX

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

383

1

<

^■^H

^

c

<

>

t/>

IX

>

«1

jx

>

n

IX

lO

iX

2

n

0Nlr^ooONt^^ocfsONhs. Ooo

oo oo

CM

K K ^ O I\ ^O oo
r^ rM f^ f*~i CM f*^ Csl

l/N 1^

o -^

NO ^

O oK0N^^^O\lr^#*^»/^f-^

f^ (N <^

^

CM

r\ON«^'ft-l/NrM'^ONON
CslrHr^rMrMr^r^fNCM

■•l-

^

4

^0

■

-1-

On

o

■

■^

A

rj

^

(M

K

^

CM

o

o

^0

o

ON

\0

^0

^

t^

m

ir\

•^

^

^

V

ir\

^

rA

f^

H^

(N|

(N

lA

<M

r^

<N

*

*

1

oo

•

0\

4

«

•

oo

•^

ON

*

p

OO

O

*

\0

*

^

^

*

OO

*

■

P

oo

1^

4

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1+1 +1+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

^ m r^ fsj ^s.O

K K OO OO

O oo O

<M^OTl-KOir\oorslo
fcr\Oc4C^(NOoooorsi

r^ x}- oo xj-

r4 fn ^ ^

CN

i^OOt™iO\»— **^r*%i— iO\

■-:^OooO\ooKK-^r4^V^

00KO\0l/>r-^^O

r4

^ ^ oo ON

00 *^ r^ rj

0\

•H ^ hs. fH-^ o '^ CT\

O ^D O ^ On ir> ir\

^O

0r^00\00\^00mrvja\ '^^MOOfn

^0

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +i+i+i+i +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

4

CM

*

^

rx

00

00

OO

K

f^

lA

ON

lA

NO

o

o

o

^

NO

■<t

O

T^

in

O

o

tX

<N

^

vn

m

NO

OO

V

OO

4

OJ

OO

K

m

l/N

lA

4

^

OO

4

■

^

rM

CM

NO

lA

<N

ON

m

SO

NO

\D

NO

NO

NO

K

K

^

NO

"^

tA

iTv

NO

^

NO

\0

\D

■i-

ir\

oo lA

NO

NO

NO

NO

lA

NO

K

NO

K

4

■

On

ON

4

4

O

4

4

4

O

■

ON

4

•

OO

4

O

ITS

■

CM

4

O

4

O

4

O

OO

4

CM

4

*

NO

4

ON

CO ITS \0 ^

0\

ITv

O^OOOOOOK'^^O

lAOhH'^OO'^^'— 'oCf^

^^ (S ,— ^ T-« (sj (N

O OO 0\ K

^o o r^ ri

ON

ON

rg-^OOKON^^^C
\0^oo^KrMO\<N^D

CM

r4 '^ •-'

^^oolr^**^N,f*^^^l

H^0O^^^^U^OOl/^mr^

<M

1— < 1-H f^ r^j

SO ^ iTv 1/^

lA

NO lA

Ov oo ^ ''J-
Tf- lA O "^

^ ^ r4

f^ ON so

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1+1 +1+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

47.7

47.6

50.6

49.5

63.0

83.9

92.4

99.5

99.4

108.4

105.9

102.7

84.9

71.0

82.3

79.6

80.7

49.6

65.8

47.6

82.8

OO

4

o

OO

86.4

73.6

NO
OO

94.1

84.1

83.9

87.7

79.4

64.0

*

1.6

1.5

OO

*

2.2

2.7

2.0

ON

4

2.0

2.4

4

m

4

^

2.0

2.5

1.3

2.4

2.2

C4

2.0

4

NO

rJ

u-s

1.4

(N

1.6

On

4

2.4

O

1.7 ;

1

vo

4

K

4

2.1

1.0

r4

m

OO

*

2.3

*

2.1

On

4

r.

O

\0

K

4

4

P

4

4

■

4

4

4

\o

•

4

•

4

4

OO

4

t

4

m

NO

4

m

4

4

rx

4

4

m

m

4

4

4

+1 +1 +1 +1 II +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

0\0\<NSD(NOOONlAr-iN£jf^| 00<SJu^NOO'^00'-Hoo '^OnoO'— 'iTviAtX^O*^

i--oo--o\oor\^DsoiA'^ii-\ u^rN^NosoKi— «o-H,s^ a^ooK^sv^■^^r^Nor\rx

0\0\ONONOOOOOOOOOOOOQCOO QOOOOOOOoOON 0\ ^ OO OOOOOCOOOOOOOOOOOOoO

K K oo NO O

On — r\ <N Tf fs|

00 OO r\ -^ o

rsl *--- r^ r^ 0\ 00 ^
rsl

K t^

NO
CN

■^ianoKooOnt-hi-hi-h,-hi-h,— ji— «,— i,-Hr^fsiiA\DKoo 5^,^,— i^,—.,— (,-i^i^rsi
I I I t r 1 I 1 I I 4 t I t 1 F t ><'>■<*>**>"* '

384

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

>3

U

U4
<

C

>

IX

tn

X

>

«)

IX

>

M

IX

7^

O ON ON

OO

K

C^ -H

cc

•— - w^

00 \D OO

• » •

r4 c^i

rs ON o

» * ■

rr\ (VJ Tf

O ON ^c ir\

r^ ^ NO

-"f

4 •

+1 II +1 +1 +1

+1 +1 +1 +1

O O 'O (N O

K

O -^ O O

Tt-rvif^rv-, oosOoor^

o rv o -^ o

\D <^ ^ a^ ^ '^ fs

r^ IS. */^

• • *

f^ — o

■^ oo -i-

n <^ cni

<N

fN

oo

K i\ rx

t ft 4

X^ r^\ \r\

IN 'n K
» ■ «

*

\r\

■ * «

rj 1^ —

^^ <N ^

« ft

ITS

+1 +1 +1 +1 +1

+1 +1 +1 +1

^O^^'^oKoK^^'^'^O'^

oo o ^/^ i^ "^
^ \D oo ^ K

fx^ c/:. oo
K ^ ^

'I" u^ oo oo ^ '^

"^ rx IN •t" ^'^ ^

Tf Tf- O i-H oo

l/^ u^ i/> 1^ \0

<N IN l/^

» » «

\0

ON

rsj ^C On

-^ ^ oo
i-H rj -^

l/N

<N

ir\ NO NC

ON K ON

N3

0\ -^ oo

■ ■ ■

■— r*^ -^

oo

K

M rv^ In

1 ft «

SX3 ^ ir>

*

00 sD —

ft ft ft

f^ '^ NO

II II II II II

mill II

l/> O O rn

OO iTv lA O '^ O

-^ f^ f*% ir\ ^
r\ l/^ ir\ \o \o

1— « LT^

IS.

K

OO

NO ON
O K

o ^

1.6

2.9

ft

2.4

OO

O "^ O^ '^ O OO

r^ O K oo \0 OO

O ON 00 hs, ^ \o

^O K K vo

r- 1-^ r^

rN

NO H%

\0

rr IT^ fN r-^

« »

^ (N

<N

r^x

rf-

OO f^ \D

*r\

rO r^ r^

II II II II II

ir\ir\OlSvOol^^^i^^
rj ^ (vj d O i^; oo r\ hv no" ^

ONCT\as3^°^00000°'=«0000

III! II II

ft ft ft

\o ^ ^^ In ^ K sO

oo Of) oo oo QO cjO OC

K K ^ \D

ir\ \0 K oo On

ir\N£3fv^ooO\-— — ^-— .-^,^-^rJ^^^^^^^^■^^-
l I i I > I I I I I I I I I I I I * '

1947J

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 385

open

gression with the eastern subspecies. I consider the northeastern limit of "true"
ssp. interior to approximate the converse of the northwestern limit of "true" ssp.
tuberosa.

Attention has been called to the abrupt change of gradient of the genocline at
approximately the middle. A palcogcographic explanation for this is seen in the
fact that this location also approximates the southern terminus of the Pleistocene
ice. North of this point, upon retreat of the ice, unusual opportunity for rapid
migration must have been offered herbaceous plants with effective means of dis-
semination, such as milkweed with its comose seeds. Surely an unparalleled
community" must have been offered them. Furthermore, the southern half of this
cline traverses low, generally alluvial land unsuited to the species, where colonies
are unusually sparse.

To this we must add obvious geographic features such as the more favorable
climatic and edaphic conditions roughly north of this point. These factors ap-
parently combine to explain why Asclcpias fubcrosa colonies of both subspecies are
larger and more frequent roughly in the northern half of their ranges, as I have
observed repeatedly. It is reasonable to expect more rapid gene flow, and con-
sequently a more gradual dine, under such conditions than under those toward the
south where the colonies are smaller and more widely separated.

Although the isophenes follow satisfactorily consistent courses, it is clear that
both the individual quadrat means and their standard deviations would represent
more or less distinctive population parameters within the broad intervals indicated.
This also will be found true, but with different patterns, with regard to the other
characters measured, as succeeding maps will show- Each quadrat population has
its unique cell-like facets of mean tendencies and variabilities; all mutually de-
pendent upon the constitution of their neighbor populations, but reflecting dif-
ferent internal forces. This heterogeneity within continuity will be found upon a
lower level, but even more strikingly, in the account of natural populations which
wjU follow.

The effects of introgression, therefore, arc superimposed upon kaleidoscopic

population patterns previously established by intra-subspecific gene flow in addition

to diverse degrees of genetic drift; the latter, we may surmise roughly in inverse

proportion to the former. I am inclined to discount any direct environmental

effect on the expression of the two angular measures; but climatic and edaphic

influence surely is exerted indirectly In moulding population tendencies In so far

as it affects population size and frequency with regard to efficacy of gene flow,

Introgressive and otherwise.

The roles of population size and frequency In moulding population tendencies
win be examined in more detail In another section of this paper. A preliminary

L

intimation of their importance will be received upon examination in more detail
of the Quadrat statistics within the distribution of ''true" A, /, interior, as indicated

[Vol. 34

38(3 ANNALS OF THE MISSOURI BOTANICAL GARDEN

by the map. In tKe Ozark region, where the subspecies Is sufFiciently frequent to
be a roadside plant, the quadrat means for Z A approximate a declination of 85.5°.
The diminishing magnitude of the means as they progress toward the WTst suggests
some slight, but diminishing, introgressive effect inadequately conveyed by the
arbitrary isophene.

Near the boundary of the short-grass plains, population frequency drops
abruptly, and the species ceases to be encountered commonly along roadsides. From
this point some miles westward, one encounters a north-south distribution of
plants with somewhat more sharply tapered leaves. As it happens, this tendency
IS associated with somewhat reduced length and somewhat increased width, so that
a fairly recognizable race results.

In the high plains, the subspecies is absent or so rare that adequate quadrat
samples are not available. In the better-watered highlands of New Mexico and
Arizona it reemerges, however, but with such increased quadrat means that an
effective mimicry of the introgressive populations of the Mississippi valley is ob-
tained. Here also associated changes In other characters, especially greatly reduced
width, produce something of a distinct racial fades.

ZB

The phcnocontour chart for Z B, reproduced as Map IV, presents quadrat
statistics in the same manner as did that for Z A, but in this case isophcnes have
been drawn at intervals of 20*^ from the eastern seaboard. This interval was
chosen after a hasty examination showed it to be the smallest which would produce
continuous contours; hence It was surprising as well as gratifying when three
Isophcnes emerged, separating four contours as in the preceding map. Here also
the ranges of "true" A. /, tuberosa and "true" A. /. interior have much the previ-
ous relation to the Appalachian and Ozark plateaus, respectively, and the middle
isophene denoting the midcurrent of introgression has somewhat the same course.

Dissimilarities to the ZA map are striking, nevertheless. The isophcnes them-
selves arc more irregular and somewhat closer together, denoting a steeper cline.
The eastern isophene, limiting the range of "true" ssp. fuhcrosuy has a more abrupt
course to the Gulf coast, the reason for which will become apparent.

The phcnocontours of ssp, inferior arc rather more complex than for Z A, and
show Intra-subspecific differentiation in Z B to be greater than in the former
character. A far southwestern isophene Is discovered here as well. Fortunately,*
confusion with the introgressive middle isophene is avoided here, since the two are
not continuous. The most striking feature, however, is the irregularly elliptic
100°— 110° contour eccentric in a northeasterly direction from the Ozark plateau.

The middle, introgressive, isophene Is regular in the northern, glaciated states,
and follows a course precisely similar to that seen in the map for Z A* The
southern half, however, is very irregular, showing three extensive and opposed
embayments. Before interpreting these irregularities, we may observe that they

1947]

WOODSON— LEAF VARIATION IN ASCLEPIAS TUBEROSA

387

S

»

T?

<

0|

s

t3

tr

o
o

O
en

V)

3

r6

tn

o'

n

[Vol. 34

388 ANNALS OF THE MISSOURI BOTANICAL GARDEN

are tKe cause of the peculiarities of the ranges of the "true" subspecies already
noted: the eccentricity of the 100°-110° contour of ssp. inferior^ and the abrupt
course to the Gulf of the delimiting isophcne of "true" ssp. hiherosay or possibly
the reverse is a more defensible position. At any rate, the evidence is mutually
corroborative.

+

I would interpret the anomalous "bays" of the middle isophcne as possibly
connected with ecological, or at least floristic, selection. The Gulf coast embay-

ment of the 60 —70 contour is a striking duplicate of typical southeastern coastal
plain distributional extensions with which every plant geographer is thoroughly
familiar. I believe that A. /. Rolfsii is implicated rather than A. f. fubcrosa in
this instance, because of the different characters of their distribution. Having
been forced to omit the typical Floridian population from statistical analysis, this
and another phenomenon to be presented shortly constitute our only evidence of
the role of Rolfsii in introgrcssion with the other subspecies.

The eastward embayment of the 80"— 90° contour is typical, if less familiar, of
the western floristic affinities of the isolated prairies of the middle-lower Mississippi
valley (cf. Anderson & Woodson, 193 5, in regard to Tradcscantia occiJcufalis in
eastern Arkansas). The third and northernmost embayment westward probably
testifies to the Appalachian affinity of the area about the Nashville basin. The
conclusion to be drawn from the embayments possibly is that in some way these
floristic regions are peculiarly conducive to either subspecies as the case may be,
and that in some way the factors for leaf base are correlated with others governing
ecological adaptability of the subspecies in a manner not affecting leaf taper. With
data as fragmentary as it is, the differential gene flow for the characters Z A
and ZB in this region nonetheless appears obvious.

In discussing the map for /_ A, attention was called to the distinctive statisti-
cal facies of the quadrat populations. The same heterogeneity within continuity
is found to obtain with regard to Z B, although here it assumes more significant

proportions. Attention was directed in the earlier discussion to the southwestern
diversification of A. /. interior. Something of the same phenomenon Is seen to
Involve Z B. Here, intra-subspccific differentiation produces an Irregularly
centrifugal lowering of the quadrat ZB means, i. e., a tendency for the leaves to
become progressively less cordate. This phenomenon, together with others asso-
ciated with It, will be examined more closely In another section of this study.

Perhaps the most striking feature of Map IV is the two convergent series of
broken lines. Tliese lines were drawn connecting quadrats having the greatest
standard deviations when It was discovered that the deviations for ZB means, un-
like those for Z A means, appear to have a more or less defined contour system of
their own. The series of broken lines from central New York to Texas was drawn
first, connecting the highest deviations In each horizontal rank of statistics. After
this was completed, I observed that a similar, almost parallel, contour to the south-
east could be obtained by connecting the deviations next In size.

\

)

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROS/^ 389

As far as I am aware, this type of introgressive contour has not been discussed
previously, and has not been provided with a name. I shall call it merely the
"crest of variability" until I am inspired by appropriate classical roots. At any
rate, involved with physiographic allegory as we are, it is natural to think of these
crests of variability of Z B as being due to a kind of genetic orogeny produced by
introgrcssion pressure from the three subspecies. This interpretation is supported
by the obvious fact that the longer crest follows the 60' -70° contour between
A. t. tuberosa and A. t. interior quite closely, and that the shorter crest to the
southeast approximates the commissure of A. t. tuberosa and A. t. Kolfsii in almost
exactly the position that any plant geographer would predict.

It would appear remarkable indeed if the longer crest of variabihty were to
parallel more exactly the middle isophene throughout its devious southern course.
Two explanations of its rather minor unconformity suggest themselves: one statis-
tical, the other biological. The first concerns the strategic population in central
Alabama which unfortunately contains too few cases for the computation of
standard deviation. Were this statistic sufficiently large, the crest of variability
would pass through it and successfully negotiate the sharp turn into coastal Texas
in perfect harmony with the sinuous contour. The second explanation is less
triflmg: the fact that any increased variability due to heterozygosis must be super-
imposed upon the pre-existing parameter variablUty of each quadrat, the range of
which may be seen by observing the heterogeneity of deviations in areas not pos-
sibly affected by Introgresslon. The Intriguing fact, therefore, is that the crest
is as consistent as it is.

The relativelv close rnnfnrmiVv of tht^ \nr^rror- ^^^c*- ,.f ,,^_.- U:i:*.. _-^L ^L-

60 -70° contour leads me to conclude that introgrcssion is progressing more
rapidly from west to east, and threatens to engulf the Appalachian population.

Rolf

Caught

between the advances of Its sister subspecies upon either side, the ultimate reduc-
tion of ssp. tuberosa to a hybrid swarm would appear unavoidable. Some indica-
tions of the rate and manner in which this is being accomplished will follow
shortly.

\

Lt no crest of variability was produced on the map for ZA, such as that

Th

iport

to me.

veiy

sible for the expression of the two characters. According to the theory of
oligogenes and polygenes (Mather, 1942), one might expect such a crest of
variability In Introgresslons Involving the former and not the latter. The great
difficulty In applying this theory to our phenocontour data is the lack of knowledge
concerning the hereditary mechanism governing the characters measured. Thi

IS IS

scarcely surprising since the plants do not lend themselves easily to breeding
experiments, and further because both characters of leaf shape may be presumed
to be controlled by multiple factors. A consequent difficulty for me Is uncer-

- *

390

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

o

.3

a

o

a

g

c

s

6

U

o

8

c

Ck,

g

t^

>

1947J

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 391

talnty wlietKer one should regard oligogcnes and polygenes as allowing an ample
continuous scale of graded expression; in other words, how many genes may
constitute an oligogene complex, and how few a polygene, or is the distinction
a relative one?

LENGTH AND WIDTH

The phcnocontours of median length (Map V) and median width (Map VI)
may be discussed together since they show much the same characteristics. In both
maps the same tendency of quadrat populations to drift about individual means
and variabilities Is seen, as has been noted in connection with the other characters.
In the maps for length and width, however, it is not practicable to draw isophenes,
since these characters are not involved in the differentiation of the subspecies. In
Map V an irregular tendency may be noted for leaves of both subspecies to be
somewhat longer in the north than in the south. In Map VI the leaves of A. /.
interior are found to be narrowest in the extreme southwest of the range. It seems
most probable that these tendencies are produced in response to climate, from
what we have learned of the relative frequency of the subspecies in different parts
of its range; the result is a series of rather poorly defined ecoclincs.

The most prominent feature of both maps is the oblique scries of broken lines
extending from New York state to coastal Texas. These lines, in both instances,
connect the largest mean in each horizontal rank of quadrats, with the exclusion of
the far southwestern quadrats of Map V. The two courses of broken lines are
almost identical, diverging in respect to only two pairs of quadrats. It is apparent
also that they approximate very closely the crest of variability observed with
regard to Z B, even to the pecuKar switch-back on the Gulf coast.

I cannot conceive, upon the evidence of the three maps, that the courses of
broken lines for length and width can indicate other than hybrid vigor. Environ-
mental effect can be discounted apparently because the hncs run contrary to the
ecocline contours for length. The combination of the two effects may be seen in
both maps In that there is a general tendency for the maximum means to decrease
from north to south. Of further interest is the indication that heterosis, if such
it be, has httlc effect upon the variabihty of the populations with regard to length
and width. These two properties would indicate that heterosis in natural popula-
tions, hke the crest of variability previously discussed, is a quality superimposed
upon previously established quadrate parameters. Heterosis cannot be regarded as
greater in the northern states, and therefore that those populations are more
heterozygous, simply because the northern maxima arc greater than are those in
the south. Northern leaves, in general, are larger than those in the south. This
explains why the crests of length and width maxima do not correspond exactly,
and, I am confident, why they do not correspond exactly with the crest of
variabihty.

Several genetical problems of importance arise In reply to the interpretations
which have just been made. Since a crest of variabihty was observed between the

392

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

4>

O

C

c
.2

J

"a

>

s

s

d
o

C

o

a.

o

>

s

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 393

ranges of ssp. tiiherosa and ssp. Rolf sir, why is there no parallel heterosis similar to
that involving tuberosa and interior? One might suppose that the differences dis-
tinguishing the former two were insufficient to produce the necessary relational
genie disbalance. From the purely systematic point of view, there can be little
doubt that tuberosa and RoJfsii are more closely related than is cither to interior.
The difference between the two former, however, still must be sufficient to produce

the crest of variability.

It is conceivable that tuberosa and interior have been in direct contact only
since colonial times, due to' the clearing of land, the building of roads, etc.,
although we can scarcely be sure of this. Even so, how has heterosis been main-
tained to the present day? Only by continuous gene flow from inbred populations,
we may expect. Of importance in this regard, probably, are the air-borne seeds
and peculiar adaptations to insect poUination described in the earlier portion of
this paper.

Heterosis, according to Mather (1943) and others, results from the bringing
together of polygenic combinations which have not been selected for good rela-
tional balance, and because of this, the phenotype of the hybrid will be likely to
show a greater departure from the optimum than does either parent. It follows
that heterosis is a sure sign of poor adaptation, and must be selectively disad-
vantageous and less ''fit" than the parental types. In this way, it is believed,
isolating mechanisms and hybrid sterility originate; Indeed, ''the avoidance of
heterosis is the most widespread stimulant of Isolating devices."

Certainly It would be unbecoming of me to enter into a controversy of
genctical theory. In my opinion, however, there Is a tendency to ascribe too uni-
form selective value to indiscriminate types of biological variations. Until a
correlated character Involving some aspect of viabiHty or physiological efficiency
is discovered, I can attach Uttle selective advantage to any of the types of leaves
which I have measured. I am convinced that plants of the inclusive species
Asclcpias tuberosa are nowhere as frequent, as fertile, nor as happily adapted to
their environment as In the zig-zag commissure between the subspecies tuberosa
and interior, where they apparently are most heterozygous.

With gradual discontinuance of out-breeding to "pure" types, I would expect
relative subsidence of both excessive variability and heterosis, and a genetic
peneplanation to a poorly differentiated hybrid swarm — the raw material of future
speclation. In the light of our introductory discussion of paleogcography, T can
see no reason why this has not occurred, perhaps repeatedly, In the past. A
primary function of introgression may be the imparting of greater variability to
specialized parental stocks. My experience In the field with the apparently suc-
cessful establishment of introgressive populations of A. tuberosa prompts me to
disbelieve that heterosis alone will lead to the erection of isolating devices to re-
separate the subspecies. As a matter of fact, the following section of this study
will produce evidence that introgression between A. t. tuberosa^ A. /. interior^ and
A. t. Rolfsii Is continuing apace.

[Vol. 34

394 ANNALS OF THE MISSOURI BOTANICAL GARDEN

V. Analysis of a Roadside Profile

In June, 1946, I mack an automobile trip from Topeka, Kansas, to Norfolk,
Virginia, accompanied by two of my friends, Richard W. Holm and George K.
Richardson, The purpose of the journey was to collect leaves of buttcrflyweed
from every colony of five or more plants encountered between those two points,
for comparison with the herbarium data from a similar profile. The route chosen
followed U. S, highway 40 eastward to St. Louis, U. S. 50 and 150 to Louisville,
and U. S. 60 to Norfolk. The final collection, as a matter of fact, was made at
Chuckatuck, Virginia, a village on U. S. highway 17, a few miles south of Norfolk,
The whole transect totals a distance of about 1200 miles.

The massive plants and broad trusses of brilliant flowers render Asclcpias
tiihcrosa a conspicuous object from a distance of many yards, even when sur-
rounded by undergrowth equalling or exceeding its height. Cruising at a speed
averaging approximately 40 m. p. h., my companions and I had no difficulty what-
soever in detecting the species, I should guess, from a distance averaging at least
one-tenth mile along cither side of the highway. With three pairs of eyes vying
for the first welcome sight of orange, I fancy that few plants escaped us.

With the cry of "Buttcrflyweed!", brakes would scream and out we we would
spring to see whether the find consisted of a single plant or two, which we would
neglect, or a colony of at least five. If the colony proved to be a large one, all
hands would set to work, quickly selecting a portion for exclusive attention, and
nimbly divesting one flowering stem of each plant of a single "random" median
leaf. Back in the car, the leaves would be united into a paper envelope inscribed
with appropriate data, Inserted into a press, and probably sat upon by Dick,
George, or myself. Perhaps more important biometric material has been secured
within four days, and from a wider area, but never more happily.

The transect which has just been described yielded a total of 994 leaves for
measurement; these were distributed amongst 53 colonies of five plants or more.
The largest colony numbered 77 plants; the median colony size for the whole
transect was 13. Unfortunately, no record was kept of single plants encountered,
nor of small colonies of less than five plants, but these certainly were met with far
more seldom than the "larger" colonies; perhaps a total of 105 plants were en-
countered in all.

The distance between colonies for which records were kept varied from 3 to
107 miles, with a median distance of 12 miles. Between these, of course, were
interspersed the single plants and smaller colonies. The mean distance between
colonics varied in a manner to which some degree of significance may be attached.

the greatest concentration occurring in Kansas and Missouri (16-22 miles) on the
western end, and Virginia (14 miles) on the eastern. The most sparse colonies
were encountered In Illinois, where the mean distance was 46 miles. From such a
small sampling, only a general inference may be made to the effect that the colonies

1947]

WOODSON— LEAF VARIATION IN ASCLEPIAS TUBEROSA 395

were more frequent the closer to the putative centers of origin of the two sub-
species, an observation which I take to have more paleogeographic than gcnetical

bearing.

As rcHcf from more practical considerations, I have indulged in some con-
jecture concerning the total population of Asclepias tuberosa in the United States,
From the details of the journey which have been given, it may be estimated
crudely that a total area approximating 243.6 square miles was searched for butter-
flyweed, with an estimated total of 1050 plants, yielding a density approximately
of 4.3 per square mile. Estimating the total area of the species as 1,434,000 square
miles, a total population in excess of 6,000,000 plants is suggested. The unreli-
ability of this figure is emphasized by the realization that the area actually sampled
equals less than one-five thousandths of the estimated total area!

Similarly facile computations may be made concerning the potential fecundity
of the species for comparison with the estimated frequency which has been ob-
served. For our crude purposes, we may assume the potential fecundity of an
"average*' plant of butterflyweed to be the product of approximately 100 ovules
distributed amongst 2 carpels for each of 18 flowers per cyme, with 15 cymes per
stem. Assuming an average of 15 stems per plant, a potential seed crop of 405,000
seeds per plant is indicated, a figure which is no more disproportionate to actual
population than is usual in such computations.

In the earlier discussion of the biology of the species, certain mechanical factors
influencing the failure to realize this potential fecundity have been indicated,
i. e., difficulties of pollen transference and the division of the stigmatic surface into
5 separate areas. To these must be added the observed somatoplastic sterility and
relatively low germination in nature, as well as the more enigmatic animal preda-
tors and rigors of ecological competition. Having by no means exhausted the
possibilities, it is clear why plants are such satisfactory subjects for systematic and
geographical studies: why they stay "close to home" as a rule; why they are
conservative to evolutionary change. It also is clear why more precise genetic
concepts are not forthcoming from studies such as these.

Table IV presents the individual statistics for the 53 colonics of butterflyweed
encountered along the 1218-mile profile from Topeka, Kansas, to Chuckatuck,
Virginia. In fig. 4 they are presented in the form of a composite graph, including
the means and standard deviations for the four characters of Z A, Z B, median
length, and median width. The abscissa of the graph indicates the proportionate
distances of the colonies, the numbers corresponding to those employed in Table
IV. The principal physiographic features encountered in the transect are indi-
cated on the face of the graph between the data for length and width. Drawn to
such a scale, the clines and other features noted in the phenocontour maps are
scarcely recognizable in fig. 4, which to this degree approaches reality rather dis-
concertingly. Nevertheless, with a bit of scrutiny, the graphs may be found to
corroborate satisfactorily the clines, as well as the crests of variability and hybrid
vigor to which attention was called by the maps.

396

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

>

nj

u

H

P

en

Q

<
O
a;

O
O

O

o

O

< u

< c/}

to

o
>

.J

O

Uh

tn

w

^

2

J

a
ft

-JC3

c

OJ

C

C

o

c

W5

c

C
n

M

C
O

">

OJ

rt
"0

C

nl

V3

O

V)

d

^

>

ix

C

>

IX

'

>

NI

IX

>

<

cn

ix

2:

s

a

I

o\

NO

rv)

vo

^0

fM

rs

K

+1 +1 +1 +1 +1 II +1 +1 +1 II

^ (^ ^ ^ CM ^

(N Os rN 04 00 0\

• •••»■••

0\

■

4

00

so

00

CM

CM

■"^

<N

CM

•

(M

-^

A

^

w>

^

NO

^

K

tf%

CM

•

•

1

1

+1

+

^

*-H

00

K

•

so

H%

00

CM

^4

^"^

*^

K

lA

m

*

«

■

*

\o

C\

0\

CM

00

v^

CM ^ ^

CM

CT\

00

K —

<M

SO

K

IN

lf\

<7\

IX

CM
00

K

tA

0\

ir»

l\

00

^

^o

>«

lA

ir\

_JJ_

pA

■^

K

CM

9s

CM

CM ^

»<^ i-H

CM

■^

CM

(N

lA

c^

o

lA

%f\

ITS

K ^

00

ITS

00

^«

hv

*/\

00

0\

00

■

00

0\

lA

r«-i

K

K

lA

K

>.

X

>^

c

rt

-^d

S5.

.a

rt rt

rt

o

rt

u

C

1-1

rt

VI CO

rt

Ec/5

00

rt

o
U

rt ^

o t

rt

rt

:U

s S « 6

rt

V)

v»
1>

rt

o

u

. c

o

U

S

6

Vi

P
O

Q

^ rt *rt
rt ■— ■

OJ O u

1-2

. j» d «

o -• ^^

CO

f

00

1

ft

•

•

00

*

•

4

A

CN

CM

r^

CM

CM

lA

<^

rA

K

CM

lA

^

f«%

+1

1

1

1

II

1

II

II

1

1

II

1

1

+

•

<M

*

•

m

•

•

CM

4

C4

ft

CM

00

4

m

t^

K

CM

H^

On

■— >

0\

'^

■—4

K

^

tA

«^

^

^O

rx

r^

■0

rx

K

00

00

00

rx

00

K

K

00

m

00

^

CM

lA

00

CM

ir\

lA

00

00

^0

vo

K

CM

4

4

ON

*

CM

(MirvC\^Or*%|s^fr^r--iOOVOr^oOir\

.• • ■ * • • • • • • 4

^OoooosoK'^'KOK^ooooK

CM

00
00

On

II II +1 II II II +1 II II II II ,11 II II

ITS

#

00

0\

•

0\

■^

CM

<M ^

»-. ^

II II II II I! II II II II +1 II +1 +1 +1

ON

K

lA

„ SO

rt O

fc n c " M

O O rt

-H <N W

rt ?

^ S

rt ^ C

rt .^

-J E

** «-> rt 4_,

«— J 3 CO «i

O rt

O

a

C

e d s d

"T3 ^ "T3 '^

C X c X

Ui U t"^ U

rt
rt

§2

V5

rt

i2 S

o

OJ

C

: ? 2

o

I

c

X *e

6^

c

3

O

rt u

r4

^

£

c

rt

CM

1A

NO

K

ON

CM

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

397

I

>

<

13

C
.-J

P3

Nl

c

>

to

JX

>

CA

X

>

m

IX!

>

ffl

l><1

2

.2

n

O

04

m

vo

00

o

rt

a>

N ^

c ^

K

n

OS

oo

CM

o

o
U

o O
rt 1—1

rj

OS

oo

OS

so 1-1

00

c*%

CS i-H

vc irsO rj -^ 00 k Os

CM

K 9y

■^

n^

K SO

OO

^O

'f

OO

^

K

so

l/>

(S

OS

(N

SO

^

%r\

(N

■^

(N

0\

^O

^^ in

* •

fA

%fS

00 ^O

^ CM

K

Tf

r^

(N|

9\

ir\

Cn|

^

On

r^

r^

o\

00

OS

OS

<7\

OS

OS

1-H oo

K

OS

K '^ lA

CM

lA

lA

00

00OlA0\^0\000vK
CM

lA r^

^o |sl

rs CM r^

i-i <N rj

^ o

c

- o

> o
:;^ o

tin

rt n

o

.2 O

u

a

-a

PQ

u

6

c

o

■t-1

C

rt

O

c ::::
o

rt

•>W K >s

rt

^P t?Q

^

rt

§

C/5

G '-S r9 "1:? j::

^ e rt

O ^

o

</> C

rt rt
O

6

^ o

.. rt

o

- 9 -"^ ^

C « C o

d 5^

<u Id O

Os

f^

a\ fn rj

ir\ r^ *^ ITv

fs

+1 +1 +1 +1 +i II +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

tf\

^ ^

o

OS

^

A

K

^

OO

o

M

c^

4

cA

*

K

lA

K

o

■

O

CO

w

oo

V

V

to

so

o

^

o

^

A

K

rs

CM

4

oo

o

^

fM

T

00

o

ir\

<N

rj

O

00

CM

so

rO

K

CM

^o

^_^

^

o

lA

0\

K

^

OS

K

so

^

^

<N

lA

Cs|

♦

<S

•

4

V

fH

■

^

^

^

■^

+1 +1 +1 +1 +1 +1 +1 II +1 +1 +1 +1 +1 +1 +1 II

<s

OS

SCSDr^0\0NKS0'O^«AS0<Sli-HlAr^CM

O\rA0\N.»-^lAOO'«0s00lAa\0\^0lA

CM (S CM •-' Csl Cn| CM

CM

^nrsiAOOOSiATj-cMOTj-ISjiA

cMr^i-'C^'«i--^iAcM'^'0

+1 +1 II II +1 II II II II II II II +1 II II II

rO

SO

^H

ITS

<M

oo

o

m

CO

o

so

■

so

K

r\

OO

o

CM

OS

OS

lA

^

oo

OS

o

oo

OS

ON

O

o

ON

On

oo

K

K

'•O

ITS

^0

^

K

VO

-«*■

•

•

•

00

•

2.8

■^

m

IN

4

•

9

fo

•

0\

•

CM

•

*

CM

*

Os

o

OS

^

^

ir\

CM

o

AJ

CM

oo

^

4

^

o

V

o

*

CM

s

V

J

■

s

o

H

^B

CM

'O

4

04

fo

•

00

*

CM

*

CM

*

•

fN

*

t

«

*

^

*

■

II II II II II II II II II +1 II II II II II II

*^

NO

SO

•«<

o

rj ■

*A

»-H

^

oo

CM

o

O

■

ft

K

K

b

K

ON

N^

OO

o\

oo

OS

On

OS

*

O

O

o

oo

oo

oo

00

QO

oo

00

00

oo

oo

00

oo

OS

oo

On

o\

f^J

SO

O

oo

lA

lA

•o

K

so

K

'•O

CM

^

ITS

oo

oo

fA

^A

(M

^^^

K

■^

"

rt
>

m

o

2

-T3 K rt (J 1^3 hJ

■a

rt

n O rt

a

rt

s

lA

NO

hv

00

OS

CM

C4

CM
CM

CM

CM

»A

SO
CM

Is

CM

C^

OS
CM

398

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

>

<

13

>

ifi

ix

8

>

tn

IX

<

Nl

>

c/a

Ix

>

wi

IX

z

8

o

c

.2

o

2

OS

>

C

o :

^ o

^ U

SI-

K

OS

fM

r^

r^

OS

^

^^

OS

r^

»A

N

\r-

N

\c

«

•

»

*

»

*

■

«

«

9

•

00

^H

oo

-'t

^

SO

OO

O

so

•*

rn

CM

f^

"

<N

<N

""

f*%

fM

KV

r*

<s

Is, —

so

K

sc

00

*N

*r\

SO

O

ir\

ft
> "-.

>

u

lyi

6 ^

o

c

6

c

O

^

•c

I

C
C ra

u

Cs»

n
>

< s

8

«

•

o

fcn

O ^

c
o

tic
c

■-J

o

JO ^

>

o
U

w ^

w ^

>

1^

< c

P5

> s

O 4^

^ C

< £

CQ

>

e

C

nl

E ^ c

■31

>

u

Vi

o

^

&0

-i! :3 ._•

£

c

n

E

3

u

u

c .

oorJoo<Nr4^QOoor^rsio^<N
KOSOSOr^^Kr^Osoooo

+1 +1 II +1 +1 +1 +1 +1 +1 II +1 +1 +1 +1 +1

4

^H^oorjoot^(Nr<%ooK'0^r^sovo
*^ir\«-"OS•^lr^^^^^O^^«^^so^^oooo

OOOlXKOOSO-^-^T-^CSSDr^r^-^OO

so^Noo'^^'^Koor^os^-ooosrNi

+1 +1 +1 +1 +1 W +1 +1 +1 +1 II +1 11 +1 II

KKK(N»^O\^OsK«00K^K^

r^OS'^'—'i'siri^Of^ooN.fNiossom

OOOKKOOK^lOOOOOOSDKlN^^

os**^f^<T\<NOs^O'— "K'^^oommoo
^r\^^^^lrs»^l/^«-'^/^<y^'^oo**^oooso

OOf-r^l/SOOKSO— "ISO\^O0l/>r^r^f\|

II II II II II II II II II II II II II II II

^N.*^^^^'*^O^ooir\irsosoo^
»Xrx0^^ooOfnosW*'^'<t- — vo

hvinsoKix^^r^^o^^^O'^Krx

SOOOKOO\C»O^^SOOOO\f^OOOO

r^so^ooK^Ooo'^i-'^soK^soso
K'^^rifS'trsrvi^sqiJ;^

dooo*"'

I! II II II II II II II II II II II II II II

r\0\ooooCTs^OOsO\oo^Orx^O
ccooooooooooasoooooooo<j\ooooos

4^

>

a

C
rt

*-* o

u

— (N

ir\

^£)

^

o

CM

■t

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

399

^

c

>

CQ

'E

o

>

in

IX

ITi

\x

>

VI

X

>

l/l

iX

;^

o
o

6

(N

^

^

o\

^-^

21.1

20.2

29.3

00

24.3

On

30.3

24.5

o

*

in

K

o

F

5.6

so

m

OS

•

oo

•

OO

OO

o

O

rs.

Csl

r^

SO

NO

K

K-i

\D

NO

NO

ON

NO

a\

ON

fN

^

N.

NO

r*

K' I-

2

c

o

(N

*^

ON

+1 II +1 +1 +1 +1 +1 +1

14.2

\D

19.1

14.9

NO*

NO
NO

OO

14.7

tx

F

11.8

16.4

1
1

16.9

10.1

10.8

*

rx

rx

o

7.7

NO

*

*

7.9

7.4

12.2

11.4

*

1.8

2.8

1

3.5

K

*

1
2.2

3.2

1

+1 +1 +1 +1 +1 II +1 +1

NO

try

IT

70.6

66.8

63.2

69.2

69.1

75.3

1
1

23.1

1

25.4

18,6

1

1

8*61

19.1

32.9

1

22.0

16.5

': 14.2

17.7

OO

o\

ITS

19.0

12.4

' 9.0

«

CM

«

•

*

<s

+1 +1 +1 +1 +1 +1 +1 +1

l/^

oo

r"*

r^

oo

■^

*

9s

B

^

4

o

o

K

4

NO

SO

»x

K

K

vs

in

ir\

OS

1^

^

NO

K

r4

T

(N

V

^

w

W

rJ

•^

+1 II II I! II 11 II II

<M

On

r3

o

O Csl

rt

c > I-

u § u

c o "

<1>

CO

O w

E

>

U
U

g

>

r 3

o w

60

U

o
(y5

>

o
U

C
O

E

CO

c

vO

rx

00

o\

vs

IT' lA

l/N

[Vol. 34

400 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Aside from area, tlie first feature of the grapli to catch the eye may be the

relative smoothness of the curves for width, and particularly for /A, in contrast
to those for ZB and length. This Is chiefly a statistical effect occasioned by the
small total range of Z A means, on the one hand, and the relatively small size of
width meanSj on the other. If V had been used as a standard measure of variability
Instead of s, for example, the curve for width would have been the most Irregular
by far.

A more significant feature Is seen in the Increased Irregularity of all curves
from the Allegheny Mountains eastward to the coast. Since the profile traverses
the central lowlands westward to the Allcghenies, an accounting factor for the
conspicuous variability eastward might be advanced as the more varied topography
per sc. Considering the roadside origin of my samples, I am not impressed by this
explanation. Another might be the assumption of greater intrinsic variability In
colonies of A. /. tuberosa than in those of A. /. 'ulterior. This is not likely since

certain colonies of the former on the Atlantic seaboard have quite as low vari-
ability as colonics of the latter on the midwestern plains.

I am disposed to view the regularity west of the Alleghcnies and the Irregu-
larity to the east as being evidence of greater migration pressure of the western
subspecies, producing greater heterozygosity, and hence variability, in the eastern
population through introgression. I shall explore this possibility more fully In
succeeding sections.

It is apparent that the colonics behave as microgcographic races, those immedi-
ately adjacent to one another often being more different, in a given character, than
others a hundred or more miles away (cf. Anderson, 1936; Dice, 1940). Never-
theless, for all their haphazard courses, there are definite clines, both for Z A and
for Z B. This composite effect presumably Is the result of random fixation and
loss of genes in colonies with limited effective population size (Dobzhansky, 1941),
combined with the occasional exchange of genes between adjacent colonies. The
latter, i)f course, Is effected through insect agency In the final analysis, but doubt-
less is facilitated by the long-distance transport of seeds from either direction,
particularly from west to east — the course of the prevailing winds.

In view of the relatively great distances between colonies, their relatively few
individuals, and their distinctive parameters, I am of the opinion that such colonies
arc the progeny of a single plant germinating from a single wind-borne seed, as a
rule; which in turn largely accounts for their distinctive facies. In consideration
of the apparently facultative cross- or self -poUi nation, I assume the effective
population size of butterflyweed to be low indeed.

A pair of little scatter diagrams which interest me specially are reproduced :

figs. 5-6, since they do much to explain the genetic mechanism of introgression
with regard to both Z A and Z B. In fig. 5, the dots represent the coincidence of
the means of Z A and ZB obtained from each of the 53 colonies of butterflyweed

i

r

t L

LA

%

s

05

90

86

80

2

1

i-UL

5

f

+

+

10

lA

^

1£

+

20

1

4

J.

♦

90

B&

80

2

1

O

+

+

£5

.J L

X.

i

+

;

so

i—L

J_^

4

35

40

'

45

-J—

j-t-i.

50

t A

95
90
85
80
2
1

X

s

Z.A

X

^B

s

+

110

-100

90

60

eo

so-

lo

^

+

Vs

1 ►

110

iOO

• 90

■eo X

70

60

50

r 20

■10 S

Z.B

X

s

%

W

s

I \ \

+

+

+

4+

+

«

P4

er>

4

i

OS

OS

-110

-100

- 90

- 80

70

- 60

- 20

- 10

- 30

20

- 10

- 10

-

♦

k-

on

a

4

+— ►

+

H

+

o

t3

4

+

B

04

4

4

•+

^

EH
CO

rllO

•100

. 90

■ 60

70

X

-k

60

20

10 S

30

20

10

10

X

w

s

ft

T

T

t 1 1

10

T— T

15

T

T

1 — r

20

~T — r
2&

T

30

T — r

I I 1
3&

T

40

I I

— I"
45

111 1

— r-
60

T r

Fig. 4. Means and standard deviations of 53 colonics of butterflyweed along a roadside profile from eastern Kansas to Tidewater Virginia. Explanation In the text.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

401

L^

92-1

91

90-

69

CHUCKAXUCK, VA

t •

70

PAOLl, IND.

80

UWRENCE. KAN

4.

90

100

110 /b

Fig. 5. Coincidence of colonial means for /A and /B along a roadside profile from eastern Kansas
to the coast of Virginia. Explanation in the text.

Za

94

93

92

91

90

69

B8

67

86

65

84

83

8E

81

CHUCKAlu'CK, VA.

CO o

-K5

o

T

*oo o

X X

ox

XX

X XXX

-IJ

»<x

PAOLI, IND.

LAWRENCE, KAH

*•

. . .

80

20

30

-I—
40

50

60

70

60

— r-
90

loo

110

120

L^

Fig. 6. Relation of /A and ZB values in three colonies of butterflyweed. Explanation in the text.

[Vol, 34

402 ANNALS OF THE MISSOURI BOTANICAL GARDEN

extending from Topeka, Kansas, to CHuckatuck, Virginia. The close relation of
intcrgrading pTnenotypes to geographical distribution appears striking, particularly
with reference to the intergradation of high Z A and low Z B values at the eastern
end, and low Z A and high Z B values in the west. The three colonies specially
identified were chosen not only because they occur at both ends and middle of the
distribution, but because their combined means represent satisfactorily the typical
parameters of both subspecies and their chne, including sufficient cases for further

analysis as well.

In fig. 6, the individual cases for Chuckatuck, Paoli, and Lawrence arc super-
imposed upon the same grid, the three cross-bars indicating the respective coinci-

dences of the means for ZA and ZB. The almost perfect confluence of the in-

dividual phcnotypes of the three colonies, located almost equidistantly along the
profile of nearly 1200 miles, is most satisfying.

Q

found due to

multiple factors exhibiting Mendelian segregation. That such Is the case with
respect to both Z A and Z B in butterflyweed leaves is suggested most forcibly by
fig. 6, In which the Paoli colony assumes the properties of an experimentally
produced F2 between the Chuckatuck and the Lawrence populations. The range
of Z B of the former clearly more than equals the combined ranges of the latter
two colonics. With respect to ZA the situation is not so clear, since, although
the range of Paoli is greater than that for either Chuckatuck or Lawrence, it is
not equal to the combined ranges of the two latter. This, following Muller
(1936), would indicate not the absence of multiple genes, but cancellation of
variability in the To by the Intra-raclal variation of the parent strains.

An additional inference to be gleaned from fig. 6 is the apparent absence at
least of strong linkage between the genotypes of ZA and ZB for both Chucka-
tuck and Lawrence. In the Paoli colony, however, there is a slight tendency for
the association of high Z A and low Z B expression, and vice versa, I assume this
may indicate that the multiple gene complex governing either character is of a rela-
tively high order, and that the genes may be distributed upon several chromosomes,
or at least at several loci.

In view of our interpretations of figs. 5-6, it seems clear to me that the
constellation of means presented in fig. 5 most probably is the result of the early
hybridization of plants of A. /. tuherosa and A. t, interior at the initial juncture
of their ranges, with successive back-crosses to other hetcrozygotes and the an-
cestral types producing the remarkably even dines with which we have become
familiar. From the inferences which I have made concerning the complexity and
distribution of the genotypes, I would expect a high chiasma frequency and
abundant crossing-over to catalyze the process which we have been discussing.
Equilibrium of the genic diffusion may conceivably be upset by such factors as
the erection of sterility barriers, as Mather has suggested, but more particularly by

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 403

preponderant selection and migration pressure from one or tlie other of the intro-
gressing populations, an example of which will be provided in a following section
of this study.

COMPARISON OF THE HERBARIUM AND ROADSIDE PROFILES

It happens that the roadside profile just discussed falls for virtually its entirety
within the horizontal rank of herbarium quadrats from Fj to Fl2 of the pheno-
contour maps. A comparison of the two sources of data presents some rather
interesting topics for conjecture, and is documented graphically in fig. 7. Herba-
rium data have been borrowed from Table III; in Table V the roadside data from
Table IV have been partitioned according to geographical provenience into a
complementary scries of quadrat populations which are numbered from J to 12
and preceded by the initial Q. It is essential, in the discussion which follows, for
the reader to understand what I consider to be the important properties of these
two samplings.

In the first place, I consider the herbarium data to be the most "normal'* in-
formation available for the population of Asclepias tuherosa as it exists through-
out its entire range, for reasons concerning the conditions of their collection which
I have discussed previously. I believe that my arguments are substantiated suf-
ficiently by the relatively consistent contours of the maps. The data, as far as
the element of place is concerned, represent not one special type of habitat, but
the sum-total of all types of environments which have been visited by the in-
numerable participating plant collectors. As far as the element of time is con-
cerned, the properties of the herbarium data are more abstruse. They all have
been collected during the blooming seasons of the plants, but during years as long
ago as 1820 and as recent as 1946. A little research upon our collections in the
herbarium of the Missouri Botanical Garden reveals that the majority of our speci-
mens were prepared between the years 1890 and 1918. The plants, therefore, can
only be said to represent, on the whole, the "recent past."

On the other hand, the roadside data are biased in two very important respects:
they have been collected only along main-travelled highways, and only during the
early summer of the year 1946. The essence of our contrast, therefore, consists
of these two particulars: a "normal" collection, with regard to place, versus a
roadside sample; and a collection made in the "recent past," versus a collection
made during the year 1946. The latter, of course, Is unsatisfactorily vague, but
it is the best that can be done at the moment, and may be of some use. Should a
collection of butterflyweeds be made for comparison twenty-five or fifty years
from now along the route of my roadside profile, measured and treated in the same
way, I expect that some Interesting facts concerning evolutionary topics may be
forthcoming.

Considering the small size of the two samplings, only 117 cases for the herba-
rium profile of over 1200 miles and 994 cases for the roadside, the first impression
of fig. 7 may be the relative agreement of the two sets of curves. A second feature,

[Vol. 34

404

ANNALS OF THE MISSOURI BOTANICAL GARDEN

>
<

W

O

c/3
C

C

Pi

a:
<

I

.J
<

a

c

r.i

>

«

E

£

C

W1

V)

"3)
c

o

C ;2

o

i

c

t/1

c
o

C

s e

w

rt

a

4>

^

C

CQ

NI

>

'X

>

X

>

IX

>

29.2

28.8

27.6

oo

ON

*

21.4

14.7

ft

OO

p

CM

22.5

5.7

5.0

oc

*

00

•

2.2

4.2

4.2

3%

■

+

m

NC

•

■

NC

*

•

IX

:?:

3

+1+1 +1+1 +111 +111 +1+1

17.1

19.8

OO

t

oo

18.5

OC

■

K

o

16.5

oo

ft

15.9

20.0 j

17.5

0\

no"

15.9

14.1

K

NO

15.3

12.6

10.6

o

ft

NO

ft

ft

NO*

15.0

ft

13.2

12.8

10.5

7.5

ON

■

*

4

o

NO

*

*

OO

ft

4

ft

OO

4

o

^

o

r*^

'^

-^

■^

— ■

^

o

++++++++++

p

o

oc

4

rs|

NO

^

ir\

On

OS

00

■^

ON

4

ft

4

o

NO

IN

In.

ON

On

ON

fv

oo

oo

K

O

''O

Csl

ft

0\

ITv

■

NO

4

IN

ON

ON

(N

o

^0

CM

(N

rj

C^

<^

""*

^

■^

oo

K

OO

K

oo

^0

■

in

In

ft

In

0\

4

<N

d

IN

ft

4

CM

4

4

*

o

ft

ft

'I-

4

00

«

+1+1 +1+1 +1+1 +1+1 +1+1

64.7

68.1

68.1

64.1

tN

*

NO
NO

81.9

97.3

98.4

102.1

0*801

2.2

2.0

<N

•

IN

K

4

4

OS

4

ft

2.1

2.0

oo

■

0\

ft

IT'

4

■

4

NO

4

4

4

*

m

ft

4

4

4

4

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1

^ ON

<N| r^ oo OO

o o\

ooNa\ONa\ooKKNo^

^OOODOOOCOOOOOCOOOO

r^OKrNiKirsro'1-0G\D

ir\rs|f-^H%ooo%»-HOir\o\

r^

*^T^■m\OhxOOO^O'—

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 405

which I may not have to point out, is that where there are disagreements of con-
siderable degree, the discrepancies are consistent. These two properties are con-
comitant, and justify the acceptance of the two sources of data in contrasting
their two properties of place and time.

m

Herbarium and roadside curves for both the means and standard deviations are
very similar, but the curve for the roadside means is much smoother than is that
for the herbarium. The consistency of the stepped aspect of the herbarium curve, in
my opinion, argues against its interpretation as a statistical artifact, and that intro-
gressive diffusion is unquestionably farther advanced along the roadside. Although
the interval range of the s curves is almost too close for critical comparison, it
will be seen that roughly the eastern half of the roadside profile is more variable
than the western half. There is no crest of variability in cither the roadside or
the herbarium profile.

IB

The Z B curves for the means and the standard deviations for herbarium and
roadside are more contrasting because of the wider interval ranges. It is seen
that west of the Allegheny Mountains (quadrats 12 to /) the roadside populations
have become somewhat more like the eastern subspecies; that of quadrat S (south-
ern Indiana) strikingly so. From the Alleghenies to the Atlantic coast, however,
the leaves of the roadside populations have shifted to a conspicuously higher ZB,
indicative of a leaf base more nearly approaching that of the western subspecies.
The s curves are rather similar, but it probably is significant that variability of
the roadside populations west of the Alleghenies has decreased, as a rule, whereas
that to the east has been maintained or Increased. This is consistent with our
interpretation of the curves for Z A.

LENGTH AND WIDTH

In interpreting the roadside and herbarium profiles, it should be remembered
that it was found impossible, in dealing with the phenocontour maps, to plot
isophenes for length and width. This apparently was due to the greater genetic
drift of those characters In microgcographic races, but chiefly to the fact that the
subspecies themselves are not differentiated by size factors.

Consequently, neither are our herbarium and roadside curves for length and
width as easy of comparison as are those for Z A and Z B. In this case, the dif-
ficulty is accentuated because the data for the roadside are composed of relatively
few specific colonies (or microgeographic races) which are particularly haphazard
in the characters under consideration. Nevertheless, It will be seen that heterosis
is suggested by the consistently larger leaves roughly from eastern Kentucky to
the Atlantic seaboard, having apparently moved eastward together with the crest
of variability of Z B, as we might expect. The varlablHty in length and width,
rather oddly, has not increased notably; rather, the roadside curves have smoothed.

[Vol. 34

406

ANNALS OF THE MISSOURI BOTANICAL GARDEN

8 3 3 2 3

e

♦ *j

5 • «

z s. t % -z

8 r

; r ? J

s s s

^ tj pH

s

9

\

\

/

/

^«

;!?£S33S2

,9

8

JO

J

s

s

I

c

00

§

4i

-o

Of

3

C

o

rt

£

o

K

o

1947J

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

407

W

id

HEH3ARIUM

11

ti sd

ROA-DSIDS

Fig. 8. Reconstruction of buttcrflywecd leaves from equivalent herbarium and roadside quadrats
Explanation in the text.

I conclude from our examination of the herbarium and roadside profiles that
the details of the phenocontour maps obtained from herbarium specimens are re-
liable under the broad conditions of their compilation. More important, the ob-
servations of Wiegand (1935) and others, that hybridization is accelerated along
roadsides and in disturbed areas, is affirmed and given quantitative expression.

Figure 8 reproduces models of leaves reconstructed according to the specifica-
tions of the ten herbarium and roadside quadrats, to illustrate our discussion
pictorially as well as a commentary on our method of measurement. The basal
lines by which Z B is measured arc included for the sake of easier comparison.

Figure 9 consists of a series of frequency histograms for Z A and Z B along
the roadside profile, reassembled according to standard quadrats. The vertical
curves connect the means of each histogram, and provide new representations of
the clines we have been discussing. The distributions of the /_A quadrats, con-
sidering the size of their samples, are nearly normal throughout. In the histograms
for ZB, however, a consistent positive skewing is noticeable in quadrats 7 to J.
To the west, In quadrats 8 to /o, a negative skewness is found, while the distribu-
tions of quadrats // to 12 are normal or essentially so. In the Z B, therefore, the

[Vol. 34

408

ANNALS OF THE MISSOURI BOTANICAL GARDEN

9!?

Lk

N%d6

/B

X=84,9

xHoe.o

S- 1.4

S= 7*5

/.

9 LL

Lk

N= 58

ZB

X--86,0

X=I02.I

S- 1.6

S= 10.8

Q i?

Lk

N=I04

Z.B

X=87.l

Xs 98.4

S= 1.3

S= 12.7

Q 9

Lk

g»87.8
S= 1.5

Nr 13

Z.B

X» 97.3
S= 13.8

Q 8

Lk

N* 95

L%

X:88.8

h 81.9

S= 1.5

S= 15.7

Q I

Lk
^89.3
S= 1.3

N- 87

ze

X- 66.7
S= 19.5

25

96

Lk

N- 32

Z.B

X-89.2

X- 64.1

S- 1.9

S= 17.7

Lk

)r=89. 1

S= 1.8

9 §

NS237

9*

Lk

N-120

ZB

X=89.9

X- 68.1

S= 2.0

S- 15.7.

Lk

X--90.4

S= 2.1

93
14:152

Figure

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 409

quadrat distributions are skewed in the direction of introgression. The absence
of consistent skewncss in the Z A histograms may be ascribed to the relatively less
sensitivity of this measure; actually, however, I interpret it to be additional evi-
dence of the quantitative nature of the Z A genotype in contrast to the qualitative
differences of the Z B genotypes. This question will be considered in detail in a
future series of studies on butterflyweed. Another aspect of the skewing of dis-
tributions In relation to the direction of gene flow will be presented shortly.

r

VI. Intra-subspecific Differentiation of Asdepias tuberosa 'mterior

In the preceding discussions of the phenocontours of Asdepias tuberosa^ we
found that although minor microgcographic races may be observed in the eastern
A. t, ttiherosa, larger internal patterns, if they exist, are obscured by introgression
with neighboring subspecies, the Floridian A. /. Kolfsn and the mid-western A. t.
interior. The last named, however, has a range sufficiently large for us to extract
for analysis the vast western portion of its distribution, which is not affected by
introgressive gene flow with the eastern subspecies.

In this region, as we have learned from the phenocontours, it is possible to
discern certain more or less definite trends in the median length, median width,
apical taper (ZA),' and nature of the base ( Z B) of the leaves. These internal
patterns may be rather vaguely distinguished by the method of quadrats through
which the phenocontour maps were compiled, but they apparently are of such
greater magnitude that they require a different method of analysis. The situation
suggests an allegoric parallel with a picture printed by the half-tone process, in
which the screened Image must be viewed at a certain distance for best compre-
hensi

ion.

METHODS OF ANALYSIS

Methods for the analysis of the internal differentiation of A. /. interior are
suggested by the phenocontour maps, in which the primary direction of modi-
fication usually is centrifugal from the Ozark plateau, the putative center of
origin of the subspecies. This tendency is seen most clearly in the character of
the leaf base, designated as ZB; the leaves in the central Ozarks {i.e. quadrats
Gil and G12) being deeply cordate as a rule, but becoming successively less
cordate in quadrats progressively outward in all directions from this center. At
the periphery of the distribution, from southern Canada to northern Mexico, the
leaves no longer are cordate, but more or less rounded or cuneate at the base, as in
the eastern subspecies.

In Map VII, an arbitrary, solid line is drawn roughly from Toronto,
Ontario, through San Antonio, Texas, and produced considerably to the southwest
of the latter city; it is approximately midway between the central Ozarks and the
crest of variability of the Z B phenocontour map. To the right of this line, popu-

Fig- ^' ZA and /B frequency histograms of butterflyweed leaves along a roadside profile,
assembled Into equivalent profiles. Explanation in the text.

410

[Vol. 34

ANNALS OF THE MISSOURI BOTANICAL GARDEN

Map VII. Method of study of intra-subspecific differentiation in Asclcpias iubcrosa
interior by means of equidistant concentric circles.

lations of A. /. interior are increasingly affected by introgression with the eastern
subspecies; to the left, evidence of introgression is absent or negligible. Any in-
ternal modification of Z B which we will find in ''true" ssp. inferior surely must
be taken into account in its effect on introgression with ssp. tnbcrosa, but we shall
defer that aspect of the problem for the moment, confining our attention to
strictly intra-subspecific differentiation to the left of our arbitrary boundary.

Since modification within A. /. inferior is believed to be centrifugal from the
central Ozarks, it Is legitimate to test the hypothesis by projecting upon a map a
series of equidistant concentric circles emanating from that point; this is ac-
complished in Map VII by using a center exactly midway between quadrats Gil
and Gl2. Our procedure obviously must be made to agree as far as possible with

the system of 120-mile quadrats previously employed in the phenocontour maps;

consequently, tlic first circle, labelled C/, is inscribed with a radius of 120 miles

about quadrats GiT and Gl2, The compass is then widened to increase the radius
by 120 miles and a second circle is drawn outside Cl and enclosing the circular

area labelled as C2, In the map, this circle actually is drawn only to the left of
the arbitrary solid line, but its course to the right must be visualized for the bene-
fit of discussion which will follow. For our immediate purpose, C2 is an arc
rather than a circle. In the other circles, to C6, the same procedure is followed.
Because A. t, interior extends farther to the southwest than to the northeast, the

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 411

outermost "circles" are lesser arcs touching the arbitrary solid line at only one
point, but they, too, may be visualized as extending to the southeast of the intro-
gression boundary, and thus are designated as C/, CSy and Cp.

In reassigning statistics from the quadrats of the previous maps to the new-
circular populations, the arbitrary procedure is adopted of transfering to a given
circle the data of only those quadrats whose exact center lies within its boundaries.
Thus, in Cly although the periphery of the area intersects four other quadrats,
only the centers of quadrats Gil and Gl2 satisfy our conditions, and the data of
the other four are assigned, together with others similarly chosen, to C2. Similar
assignment is practiced In the remaining circles.

An additional feature of Map VII which requires explanation are the tw^o
broken lines which intersect in the center of C/. These lines arc drawn in order
to separate from the concentric circles two equivalent series of concentric arcs
lying in opposed directions, one to the northeast and labelled from NE2 to ]V£o,
the other to the southwest and labelled from SW2 to SWQ; the purpose of these
will be explained shortly. Rather confusingly, I am afraid, the three southwestern-
most arcs bear two sets of symbols, C/ to Cq and SWj to SWq for designation in
different phases of our discussion, but the populations are the same,

THE EVIDENCE OF INTERNAL DIFFERENTIATION

IB

Since our circle and arc populations wall be provided by the quadrats of the
phenocontour maps, two ways of using the quadrats in their new groupings sug-
gest themselves, which we will illustrate with regard to the data for /B. In
Table VI, the herbarium quadrats assigned to each circle have been broken down
and the constituent cases thrown together in computing new circle means and the
usual measures of variance. The centrifugal low^erlng of / B values w^hich was
intimated in an irregular fashion by the phenocontour map is seen to be a con-
sistent tendency. In Table VII, on the other hand, the means of the means and
the means of the standard deviations of the included quadrats of each circle are
computed for comparison. Only slight deviations from the statistics of Table VI
are found, and the centrifugal tendency is confirmed by either procedure.

It will be appreciated, how^cver, that a curve drawn to the data of Table VII
would be somewhat smoother than one drawn to the data of Table VI, I assume
this to be due to unequal w^eighting caused by the numerous cases in well-
collected quadrats and the fewer cases in quadrats less frequented by plant col-
lectors. This Inequality appears to be compensated in the calculation of the means
of means and means of standard deviations, since a quadrat containing few cases
is weighted equally to a quadrat containing many cases. In addition, of course,
the method of Table VII is more convenient than that of Table VI; consequently
I have used the former In subsequent computations.

The data of Table VII arc presented graphically in fig. 10, In which the
ordinate is in degrees of / B and the abscissa represents the equidistant circles

\

[Vol. 34

412

ANNALS OF THE MISSOURI BOTANICAL GARDEN

110

105-

100-

95-

fo-

85-

eo
zo

X5-

10-

-r
cz

/

/

/

T"
C3

-r

C4

C6

C6

^x

X

s

T
C7

CS

C9

Fig. 10. ^B means of means and means of standard deviations of
Asclcplas tuln'rosa Inferior in tlie central Ozarks and in equidistant concentric
circles to the periphery of the subspecies range. Explanation in the text.

from the Ozark plateau. The means curve is seen to be remarkably even, with
the exception of the interesting "hump" at circles J to 4* The curve for the
standard deviation means is still more interesting, since it rises from the lowest
value, in circle 7 to a crest at circle S> then subsides to a somewhat lower plateau

for the outermost circles.

The validity of the curves presented in fig. 10 may be tested by dividing the
concentric circles into two equal, but opposed, series of concentric arcs to the
southwest and to the northeast of the Ozark plateau. The primary purpose of
these arcs is to withdraw for comparison two equivalent samples from the original
populations; to further this end, the opposed arcs arc isolated by the scries of arcs
to the northwest, upon which the labels of the circles appear. An additional ad-
vantage of our two series of arcs is that they traverse climatic regions which differ
to a very marked degree: from the deserts of the southwestern United States and
northern Mexico to the Ozarks, and from the Ozarks to the continental forests of

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

413

X\ XI 2;

Hi

N>

C7^

*

OO

■

Nl

^O

>

o

n

n

Ni

»--

00

n

CO

00

-1^

^>*-*

^o

_^^b_

sc

K)

%jj

*

■

00

1

00

^^^^

XJ

00

00

n

kt

4^

C\

'^

00

00

n

bj

^

N

\yi

00

•t>^

n

*

so

C^

00

1
I

o\

00

N|

n

*

#

vo

\^

4^

N

to

5j^

h-1

w

>

^

O)

^

W

W

^

to

i

>

ft>

HH

2

'■^

M-v

3

d

>

£3

(A

a

P

^

P

>

a-

■■

H

3

(>^

>

n

"pi

to

9

>

m

<

CA

W

""^

W

^

a

H^

a.

to

a-

r

G-

u

r J

9'

3

iy>

n

n

n
n

n

<i

in

X

Z

>

^^■^tf L

^

K^

5.8

1

6,4

1+

^_B 4F^j_L U.^

n

^

■

N>

t— k

hJ

jp-^

ON

0.0

+

XI

\^

n

1

■

*

NJ

H-t

1— I

4^

Ni

•

K)

K ■

<^

ON

^- 0.8

V

=>

r^

H-h

H^

1

NJ

4.8

1

4.8 !

1+

^

9

1

so

N

00

N>

18.4

18.0

1+

VI

N>

1

00

-^

1

^

Wl

1

22.4

18.9

±. 2.1

00

■

00

s>

OS

•

\J1

1+

N>

4^

4*.

n

00

i-w

ON

20.

*

+

n

00

CO

-^

t

00

N>

Ki

•— '

ctn

*

*

1+

^

N|

OS

1 1

on

D
u

O

CA

r»

s

o

n

-«1

n

o

IN.

to

^

O

<

>

n

O

to

d

t^

C

n
O

n

H
5=0

n
n

n

>

r

a

to

r

a

\

[Vol. 34

414 ANNALS OF THE MISSOURI BOTANICAL GARDEN

the Great Lakes. Thus we may expect to observe any influence of climate upon

the phenotypic gradient.

Two final properties of the arc populations are worthy of attention. The
northeastern arcs lie almost wholly in territory covered by the last glaciation,
whereas the southwestern arcs apparently have been available for plant occupancy
since the beginning of the Mesozoic or before. Furthermore, we have observed
that colonics of butterflyweed are appreciably larger and more frequent in the
northeastern arc, whereas in the southwestern arc the colonics are both smaller and
more isolated. We may expect the three factors of climate, geological history, and
relative population structure to have characteristic effects upon the internal dif-

ferentiation of butterflyweed, as in other organisms. The method of analysis by
two opposed arcs has so many advantages that it has been used not only with regard
to Z B, but to Z A, median length, and median width as well. For the moment,
however, we shall confine our attention to the first named.

Data for Z B, analyzed by means of the two opposed arcs, are provided in
Table VIIL In the uppermost panel of fig. 11, the data are presented graphically
in a manner designed to simulate the geographic relationship of Cl and the two
sets of arcs, the ordinate scales in degrees being erected at the center, and the two
sets of curves diverging to left (southwest) and right (northeast) respectively.
The two sets of curves are strikingly similar, and both obviously confirm those
previously obtained from the circle populations. CUmate is seen to have nothing
to do with expression of the phenotype.

The chief disagreement between the circle and the arc data is seen to be the
accentuation of the "hump" of the means curves previously noted, occasioned by
the relatively low values of arcs 2 and J in both the southwestern and the north-
eastern series. I assume this effect to be due to the failure of the diagonal, solid
line noted In connection with Map VII entirely to exclude the influence of Intro-
gression with the eastern subspecies, and that the actual curves, if introgression .
were quite eliminated, would be appreciably smoother even than as depicted in
fig. 10. I have not attempted to move the diagonal line farther to the left, how-
ever, because such action would result in reducing our populations sufficiently to
impede our calculations. "We are justified, however, in regarding the smooth
means gradient of fig. 10 as being an understatement rather than an exaggeration.

It is important, finally, to emphasize the confirmation by the opposed arcs not
only of the means of means presented in fig. 10, but also of the peculiar curve for
the means of standard deviations, with its interesting crest of variability at circle
5. It is important, also, to call attention to the similarity of this curve to the
gradient of the standard deviations involved in the introgression of subspecies
inferior and tuherosa which were presented In fig. 7.

The basic similarity of the gradients representing the introgression of sub-
species ifjfcrior and fuherosa (fig. 7) and those involved in the internal differentia-
tion of Z B in interior strongly suggests that somewhat similar processes are in
play. This is indicated most particularly by the confirmation of the crest of
variability midway in the arc and circle gradients.

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

415

\

\

SW9 SW8 SW7 SW6 SW5 SW4 SW3 SW2

\

-10

1^

CI

/

NE2 NE3 NE4 h£5 HEB

h

11^

X

LENGTH

X

s

-67

s

SW9 SW8 SW7 SW6 SW5 SW4 sS3 SW2

-83

-3

-1

CI

NE2 NE3 NE4 NE5 NE6

/

N

/

\

/

y

\

SW9 SW8 SW7 SW6 SW5 SWA SW3 SW2

-i&

-10

CI

NE2

NE3 NE4 NE5 NE6

WIDTH

•6

Sl«9 SW8 Wf SW6 SW5 SW4 SW3 SW2

■I

CI

NE2 io Wl WE

NE6

Fig. 11. Means of means and means of standard deviations of ZB, /A, median length, and
median width of Asclepias tubcrosa interior in the central Ozarks and in equidistant series of arcs
to the southwest and northeast. Explanation in the text.

[Vol. 34

416

ANNALS OF THE MISSOURI BOTANICAL GARDEN

VD

o\

^6

2

■^

m

K

OO

4

rs

in

J

2

K

*

•

00

u

t

\o

o

UJ

•

«

CENT

•*f

o

1^

;<?;

O

u

r^

o
o

rsi

o

UJ

Z

•^

■

< £

o

Q 'H

W .2

.-H

*

■

^ S

o

CS

^

^

o -S

*"«

f^ —

e^ ^

o «

T3

(M

-i-

<N

O rt

^

■^

•

■

s

CO

o

f— 1

>

r

P3

'^ o

Q c

"*

rj

IN

<

pq c

t/:)

O

w S

Cfl

■

o

b^

£*

r«%

ft

r4

t>D

o

r^

^^

1

(/3 o

1

^

4

OS

P^ s

VD

ON

rv

a

\0

rj

1

o

t/5

OG

^

b

p^

<

^

«

OO

Pej

t/3

00

f-i

w

W

Ex<

oo

^o

_

O

t/5

-i-

*

oo

v^

C/3

<

t/5

K

•

oo

^0

PQ

^

u

u

_J

z

,1k

<

1

I

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA 417

Since the gradients of tKe opposed arcs in glaciated and In unglaclated terri-
tory are essentially similar, I believe that we may visualize the leaves of A. /.
interior In post-Pleistocene times, throughout its range, as being essentially like
those with rounded to cuneatc bases predominant to-day at the periphery of its
distribution. In relatively recent times, probably only a few thousand years ago,
a major genotypic change producing a deeply cordate leaf base would appear to
have become established In the central Ozarks, and subsequently to have been
diffused to outlying populations in the centrifugal manner which we have been
discussing.

In the absence of detailed breeding and cytologlcal evidence, we may only
conjecture concerning the nature of this genie change. It is apparently multi-
factorial, as evidenced by the properties of its distributional frequencies (cf. fig.
12). One would expect a single mutation, even if having sufficient survival

r

value, to be inoperative, and the simultaneous production of a sufficiently large
number of concomitant mutations effecting the same character to be extremely
unlikely. Since either translocation or inversion would be expected to reduce the
frequency of crossing-over, It is difficult to see how the smooth gradient involved
in the centrifugal modification of Z B could be obtained through chromosomal rc-
angement, as understood at present. I have in progress further genetical studies,

which I hope will throw some light upon the nature of the supposed genotypic
renovation of A. /. interior.

In this connection, however, It is worthy of note that I have found plants of
the cordate-leaved populations of A. /. interior consistently to have more stems
per plant, as well as more leaf nodes per stem, than in plants of the peripheral,
narrow-based populations. It Is natural to ascribe selective advantage to the
former, because of their presumably greater reproductive potential, and we shall
encounter confirmative statistical evidence of this inference presently.

Figure 12 consists of standard frequency histograms of /B in the nine circles

extending from the central Ozarks to the periphery of the range of A, t. interior

(exclusive of the zone of Introgression with A. /. tubcrosa). The vertical curve
connects the means of the several histograms. Although surely distorted by the
discrepant sample sizes, the symmetry of the distributions is remarkable in passing,
centrifugally, from essential normality through Increasing degrees of negative
skewness to a less skewed, platykurtic form near the center of the gradient; the
remaining figures become increasingly bimodal. An intimation of this trend was
seen previously in the histograms of the roadside profile (fig. 9), in which the
figure for Oil (partly within Cl) is normal, whereas that for Ql2 (wholly with-
in C2) is negatively skewed.

In connection with fig. 9, we remarked that skewing of the distributions is
rather consistently In the direction of the introgresslng subspecies parameters, and
thus that It points toward the direction of gene flow. A more likely explanation
of the skewing, however. Is provided in fig. 12, which happens to be away from
the direction of gene flow. In other words, skewing in these figures does not
indicate the direction of gene flow, but probably merely the distribution of domi-

[Vol. 34

418

ANNALS OF THE MISSOURI BOTANICAL GARDEN

X*S0.2 S-16.6

X-S3.6 S-17.4

CI
Jt-ai.2 SM6.5

X-84.5 S-18.9

05

X'97.8 SM8.0

X-IOOJ S«I4.8

st>

?*I03.4 S-12.0

60

X-I04.2 S'ltJ.O

tpU

TMIM S-6.4

no

iiu

lou

uo

I. (4)

I2d

IM

JW

nance, as sui^gcstcd by Fisher, Immcr, and
Tcdin (1932), Consequently, in the
gradual development of bimodality to-
ward the periphery of the range, we
might infer the gradual loss of dominance
in both types. I assume that if our range

were sufficiently great, we might there-
after witness the emergence of increasing
dominance of the ancestral genotype.

Although it may be rather rash to con-
tinue speculation, I am inclined toward
the opinion that in the centrifugal dif-
ferentiation of ZB within A. t. interior,
as in the introgression of that subspecies
with A, /. fuherosa, genic amalgamation
is effected by migration, cross- fertiliza-
tion, and redistribution of the genotypes
through crossing-over. I would assume

■

the bimodality of our peripheral popula-
tions to indicate relatively early steps in
this process. The eventual emergence of
the Ozark strain as dominant I would
predict by virtue of its supposed selective
advantages.

At this point I would wish to insert an
analysis of a roadside profile from the
central Ozarks through the circles, sim-
ilar to that by which the introgression of
subspecies hiterior and tuber os^a was
studied. Since circumstances have pre-
vented me from gathering the necessary
data, Table IX is substituted, in which
will be found statistics of a random as-
sortment of natural butterfly weed col-
onies collected in the central Ozarks and
in the southwestern and northeastern
arcs. As incomplete as the evidence may
be, ^ve may observe additional confirma-
tion of the centrifugal effect; we may
see that here, also, as in the introgressing
subspecies, redistribution of the genotypes
is not an orderly diffusion, but proceeds
from colony to colony with different

Fig. 12. Distribution of ZB in 9 equidistaiic
circles from the central Ozarks to the periplicry
of the range of A. f. hifcn'or. Explanation in
the text.

/

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

419

PRINCETON. MINN.

N-65
X=94.2 S-II.O

ABILENE. TEX.

N=78
X=89.8 S=|2.3

ROLLA. MO.
N-23
X=II3.0 S^9.6

110

leo

CI

NE5

Fig. 13. Distribution of /B in 3 colonies of A. t. inferior. Explanation in the text.

velocity rates, depending on the genie constitution of the reproducing zygotes.

Figure 13 consists of frequency histograms of three natural colonies of A. t.
interior: one drawn from the central Ozarks, the other two from arcs SWj and
^^5^ f^fom left to right, respectively. The figures confirm the properties of the
artificial populations which were discussed with reference to fig. 12. Bimodal dis-
tributions from the far southwest are not available because of the small population
sizes in that region.

If we return, in closing, to the complementary curves for Z B in the southwest
and northeast arcs, reproduced in the uppermost panel of fig. 11, we may remember
that the first evidence of the genetic basis of the centrifugal effect was received
in view of the very great climatic difference of the territories included. Whether

F

toward the Sonoran deserts or the Canadian forests, the phenotypic gradients ap-
parently are not directly affected. That is to say that the concentric lowering of
/B values, per sCy Is not environmental.

In the relative rates of transmission of the new gene arrangement, however,
there is evidence that climate plays a positive role. For in the more equable
climates of the northeastern arcs, the gradient for the means of means is more
smooth than is that in the hotter, more arid southwestern arcs, and the means of
the corresponding arcs are consistently somewhat lower. The curve for the means
of standard deviations Is more even in the northeastern arcs as well, and the crest
of variability is somewhat lower. Genie diffusion obviously is more rapid in the
northeastern arcs because the favorable climate allows the Incidence of both larger
and more frequent colonies of the plant.

In the remaining paragraphs of this section, the data for Z A, median length,
and median width are treated similarly by the comparison of the southwestern
and northeastern arc populations. Data are provided in Tables X, XT, and XII,
respectively, and presented graphically in fig. 11. In all three characters evidence
of genetic drift, with or without the directing influence of natural selection,
appears to be the chief phenomenon of Interest.

Before continuing, it may be well to emphasize again what I take to be the
most important properties of the two series of arcs in this connection, viz., the
rather uniformly equable climate of the northeastern arcs, with their relatively

420

ANNALS OF THE MISSOURI BOTANICAL GARDEN

[Vol. 34

X

M

hJ

H

o

H

(Means, standard errors, standard deviations, and coefficients of variation; angles in degrees, length and width in millimeters)

Width

1

>

28.8
24.0
17.3
18.3

1
1

38.3
15.0
29.4

21.7
22.8
18.6
25.9

20.0

so

*
NO

14.3

14.4
8.5

ROM NORTHEAS

m

2.2

1.0

IS

*

OO f-H

IX

^^f^^i-H"^\0lAr*^S0O

A

•^

*

00 r-i

+ + +1 + + +1 + + + +1 +

1 -+- C

o

+1

o

+

OO

o o

+ +

U,

(N04EJsr^r*%i^O'---Or^O

^

ITS

OV f^

m

DNIF

Length

>

ooiTNi-Hoocar^'-'^c^rx'— '

o

NO

o

OO 1— I

lr^f^r<^r'^•^r*^01— "f^rNti— <

r^

o

o

i-t 00

COL(

n

'-"tr^^^O'^'^'i-KO^

lA

*

o o

J

^

r^^-HO-'— «'—''— *4/^o^'--'-<CT\

ON

OO

'- IS

CORDING TO NATU]
3 RANGE

IX

1

<^^O^^<t|-^Q000^00L\D

1 +1 +1 + + +1 1 + + +1 +

oooor%ooooooKoooocooo

1

NO

+

o

NO
OO

89.5 -^ 3.0

83.7 ± 1.7

93.5 H- 3.1
86.4 ± 1.6

1

CO

1

1
1

>

15.0
11.7
11.9
11.0
16.6
15.1
20.4
14.2
17.6

7.7
9.0

NO

0\

8.2

K O

b^ NO

U

m

5.2

8.4

7.5

•-H OO

• •
NO NO

INTERIOR ASSEMBL
HVVEST OF THE SUB!

IX

r\^OOOfNC\OH%i— (OO-^

+ +1 11 + + + ti + + + 1

x^niONirsNO^O-^Tt-Kr^"^

ooo^o^o^ooQoooooooo^o

79.1 -t- 2.6

91.4 -•- 2.7

91.5 ±: 2.3

107.1 -t- 1.7
112.8 H- 1,5

<

NI

>

1

o

■

4

IN K

t-Hf-^O-HCslf-H^— I^^HO

V

c^

.-. o

UBEROSA
SOUT

M

0.9

0.9

00

4

O 00
•

t- o

IX

iAr^i-<rNi'^r4rsifN<Nr-^m

^

d

NO

MEASUREMENTS OF LEAVES OF ASCI FPIAS T

OOOOOOOOOOO

+ + +1 + + + +1 + + + +1

O\T-H0Nr^0\mrsir^o0NO^

OOOOOOOOOOOOOOOOOOOOOO

87.8 -+■ 0,

o

+

NC

ft

OO

o

+1

OO

■

OO

o o

+ +

OO OO

1
1

z

'T

-

«^ o

1

Colonies

Location

Long Siding,

Mille Lacs Co., Minn.
Princeton ^ 1,

Slierburne Co., Minn.
Princeton -^2,

Sherburne Co., Minn.
Elk River #1,

Sherburne Co., Minn.
Elk River #2,

Sherburne Co., Minn.
Elk River #3,

Sherburne Co., Minn,
Snail Lake ^1,

Ramsey Co., Minn.
Snail Lake ^2,

Ramsey Co., Minn.
Snail Lake #3,

Ramsey Co., Minn.
Cannon Falls ^1,

Dakota Co., Minn.
Cannon Falls ^2,

Dakota Co., Minn.

Hampton,

Dakota Co., Minn.

River Falls,

Pierce Co., Wis.

Hertel,

Burnet Co., Wis.

Kalamazoo,

Schoolcraft Co., Mich.
Silver Lake,

Palo Alto Co., Iowa

1

1
1

<

<

z

tk

*

0t

*^

Z

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

421

1^

s

X!

pa

^

^

C

NJ

<

'5

o

u

u

<

>

CA

X

>

to

X

w

>

tf>

X

^

C
O

<N

Ol

ITS

C4

fN

m

On

K

r4

K

CO

00

ft

oo w^ •-«
* • . .

01 r*-. <S| r^

o\

^

00

SO

a\ OS so

■ * *

NO

oo

<N

i/\

SO

ON

ON

%r\

K

oo

-' \0 1-

u^

^

00

00

(N

P

rx

o\

lA

o\

f^

in

•

u-\

as

CM

K

■ «

K O

CO

■

K

«

m

so

K

•

K

(N

H-i

oo

t

l/>

r^

■

-1-

IX

CM
IS.

1

K

oo

CM

OS

rs|

rs|

00

r.|

OS

-1-

OS

K

CM

4

K

OS

0\

00

^

rx

oo

so

■

K

oo

K
K

K

00
ON

so

*

rx

OS

OS

K

*

OS

C\

r^

rx

as

•
oo

<N

h.

i/S

■

4rs

00

ON

so

GO

so

OS

OS

c^<

00

in

ITS

SO

*

4

4

04

^

f^

oo

in

rM

oo
rsi

(M

m

<N

x

+

+1

+

--

OS

OO

OS

ir\

o

f-^

rn

-

O

O

o

K

'*•

0\

in

OS

OS
OS

(N

in

in

On

K

00

OS
00

K

c^

OS

m

■

fM

K

^

VO

o\

o\

ON

.-. (S

OS

OS

m

m NO

r^

^

oo

in

00

«s

CN

m

f^

^

(N

t*^

^

(N

in

m

CO

so

so

SO

oo

so
oo

in
so

oo

^ K

so

1-

oo

On

■

00

K

CO

o\

so

00

(N

r^i

ON

CM

00

oo
rM

K

so

so

o

O

^ S Q

0.2

CD

c

3

O 3

;« ni o

rt

<-»

G
O

o
U

c

o

n

8

cu

O

o

f<

o

o

U

c

o

O

■

o

*

o

_ o 5

i-i

3

rt

t^

en

"o

O

m

o

O

o
U

CI

*

6

VI

6

k2

c o

■^

K

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1

K

so

so

+1 +1 +1 +1 +1 +1 +1 +1 II +1 +1 +1 +1 +1 +1

irs

ON

ft

so

+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 II

rx

OS

so

+1 II II II II 11 11 II +1 II II II II II +1

\D

r^

in

s

o ^

VI

u

3

X

O

u

o ^uj^u gw^u

C ur, O

1^

K <

3

s

2

u

in

oo

NO K

^ ^ ^ ^

<-o cyn t/:i Lo

422

ANNALS OF THE MISSOURI BOTANICAL GARDEN

fVoL. 34

W

u

Pi

X

O

o

u

o

yf*

V)

1

C

c

4-t

w

rt

c/^

">

o

E

•TD

o

Lh

rt

c

PJ

03

C

01

<

P3

E

H

C/2

and

<

c

W

rt

Pi

o

<

V)

P<

c

P

S

<

&

s

p

Pi

PC

W

c

W

2

2

2

2

U

C/5

C/5

c/5

ON

u

<

1

'«^

On

VO .-«

rN

NO

K

•"t

oo

en

•

90

t*^

SO

oo

r4

IT*
00

00

00

»A

^

NO

tS

NO

NO

ON
OO

■^

NO
OO

K

oo

2

Ih

ir\

^

K

00

<N

NO

X

X

t/5

.J

*>o
W

2

'*

OO

p

oo

*

o

u

K

*-•

a;

HH

u

u

On

o

s

2

K

o

oo

r*\

.

u

■^

O

SO

o

2

^

*

OO

<N

^

PH

o

o

C/3

2

lA

O

f4

u

OO

PK^

<

c

<s

o

o

w

pL]

\

+

ty)

f«^

o

oo

<M

O

Ch

Ph

o

.2

^ ^

o

rt

^H

*

#

o

u

C^4

o

H^

2

-T3

HH

-T3

G

<^

On

K

^3

^

■*

*

o

r«

C/1

1

NO

1—1

^

K%

M-(

1

O

if)

V)

r^

\r\

r»%

ITI

c

^

-^

•

Q>

oo

vo

^^

E

-a

t/3

c

-fc

DRAT

V!

G

8
g

(/5

ir*

NO

>6

OO

ON

<

*-

1

1

^

00

^^

tn

NO

1— ■

K^

n

P

u

s

s

^-^

NO

<^

a\

^

<N

4

OO

oo

tT)

ITS

P^

ffi

O

NO

o

NO

K

*

r4

c/3

<

W

OO

ir*

K

Ki

■*

4

NO

H

O

2

O

^

K

o

NO

SO

4

U

» --

go

NO

<

s

(J

H

tt

w

<;

Z

X

IX

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

423

H

C/3

U

o
o

U

<

Q

O

Ah

o
o

.-J

Pi

Q
<

p:;

Pi!

o

<^

^

o

u^

w

V

?:

-i-

Oi

^

oo

ON

UJ

#

*

2

K

(N

m

■^

^

UJ

B

^:

■^

CM

r^

K

OO

UJ

w ^

»

4

Z

ii^

r*^

rr\

(N

*

O

o

w

*

■

r 1

1-^

K

^

to

■

c
o

* -w^

^ ^

O

■

o

■

*>

u

(M

%r\

^

a>

'"^

'H

T3

Ih

1

nda

^

■

rt

CO

T^

4»t

O

G

^

O

(>D

1^-1

£

-o

c

«

-^

^

«

4

CO

i/\

-^

en

a

C/^

rt

v

1^

*A

s

^

0\

■

IJO

f— "

m

s

s

4

CM

oo

fn

J

K

^

^0

*

t>r!

' oo

hv

CN

>t

*

■

w^^^^

^

K

<N

CO

CO

K

m

oo

O

u

IM

«

Ui

<

^

X

IX

larger and more frequent colonies of butterfly-
weed; and the increasingly hot, arid climate of the
southwestern arcs, in which the colonics are ob-
served to become smaller and less frequent. Of
less obvious bearing to our discussion is the prob-
ability that the southwestern populations are of
greater antiquity than are those of the northeast.

ZA

In view of the relatively narrow interval range
of this measure, the monotonous regularity of the
curve for /A (fig. 11, middle panel) may be
somewhat deceptive. Even so, it is apparent that
in arc SW4 (western Great Plains) there is a
definite tendency for more abruptly tapered leaves,
while in the far southwest the leaves predominant-
ly are more gradually tapered than the normal.
We have observed these tendencies previously in
the phenocontour maps. Since both extremes occur
in hot, arid or semi-arid land, where colonies are
relatively small and infrequent, and apparently
are not parts of a definite geocline, I conclude
that we are observing a rather poor example of
genetic drift, in which the influence of natural
selection is absent or negligible.

LENGTH AND WIDTH

In the curves for median length and median
width (fig. 11, two lower panels), the interaction
of genetic drift and selective pressure toward the
establishment of microphyllous types, so char-
acteristic of xerophytic vegetation, appears In the
southwestern arcs with almost schematic clarity.
In general, length is correlated with width; but
in the northeastern arcs we see that the two
characters may behave independently, the length
regularly being considerably greater than in th
Ozarks, while the width is generally somewhat
less. The various curves presented in fig. 11 ap-
pear to testify to the individuality of the genetic
mechanisms responsible for all four phenotypes,

[Vol. 34

424 ANNALS OF THE MISSOURI BOTANICAL GARDEN

and their freedom of action under the pressures of both natural selection and the
breeding structure of the populations.

RELATION OF THE CENTRIFUGAL EFFECT IN /B TO INTROGRESSION

Having demonstrated the existence of centrifugal modification of the leaf
base in genetically more or less "pure" Asclcpias tubcrosa interior, d.n attempt
should be made to apply what we have just learned to the problem of introgression
of that subspecies with A. /. tubcrosa.

In the lower panel of fig. 14, appears a map of the east-central United States
from the Ozark plateau to the Atlantic coast. Three features of our previous
discussions are superimposed upon this map: the concentric circles of Map VII as
they would appear if projected to the east of the boundary of introgression, the
horizontal rank of quadrats from Gj to Gil of the phenocontour maps (which is
chosen because it coincides with the diameter of the concentric circles), and the
position of the crests of variability between subspecies interior, fuherosa, and
Rolfsii as determined in Map IV. Since the width of the circular areas corresponds
to the sides of the quadrats, It is possible to effect a fairly satisfactory comparison

of the two sets of data.

In the upper panel of fig. 14, the ordinate is in degrees of ZB and the abscissa
records the positions of the coinciding circles and quadrats as they appear in the
map immediately below. The uppermost curve, borrowed from fig. 10, represents
the centrifugal modification of Z B which we might expect to obtain In this region
if there were no introgression with the eastern A. t. tubcrosa. The lower con-
tinuous curve represents the actual gradient of the quadrat means, taken from
Table III, indicative of the effect of introgression of A, /. interior with the eastern

subspecies.

If we refer back a moment to the Z B curves in fig. 7, which were obtained
along the horizontal rank of quadrats from Fj to Fl2, their basic similarity to the
quadrats curve in fig. 14 is apparent; in all three curves the influence of the
"hump'* noted In the circles and arcs gradients may be observed. We shall attempt
to provide possible explanations of this feature in paragraphs which will follow,
but for the moment it Is sufficient to note the participation of the centrifugal
effect in introgression of the two subspecies.

necessary

Rolf

the abrupt rise of Z B values at the eastern end of the quadrats curve in fig. 14.^
It may readily be appreciated that the more or less hastate or cordate leaf base
characteristic of the latter subspecies would produce a relatively high ZB value;

4

and Its proximity to the crest of variability between the two subspecies leaves
little doubt concerning the change of gradient of which we speak. The Floridian
subspecies, in fact, most probably also is concerned in the similar gradient changes

(

1947]

WOODSON LEAF VARIATION IN ASCLEPIAS TUBEROSA

425

CIRCLES

QUADRATS

CIRCLES

QUADRATS

Fig. 1 4. Comparison of / B quadrat means as expected with simple gene diffusion and as
actually obtained in the introgression of A. t. interior and A. t. tiiberosa. Explanation in text.

[Vol. 34

426 ANNALS OF THE MISSOURI BOTANICAL GARDEN

seen at the eastern end of the ZB curves in fig. 7. Since our biometric measures
have been designed for use only with the other subspecies, the recurrent influence
of Rolfsii is rather annoying, however Instructive.

Hemmed in by A. t. interior on the west and northwest, and by A. t. Rolfsii
on the south and southeast, Httlc remains of A. /. tiibcrosa which may be called
genetically pure. Under such conditions, it is impossible to distinguish internal
population patterns except upon the lowest level, and very hazardous to speculate
concerning the original subspecies parameters. In the absence of any evidence to
the contrary, I assume that there is no concentric modification of the leaf base in
A. t. tubcrosa such as we have just demonstrated in A. /. interior. Since the
quadrat populations upon the Atlantic coast farthest from the probable influence
of introgresslon vary roughly between 47° and 55° with respect to ZB, I am esti-
mating a leaf base of 50" as a convenient approximation of the characteristic of
"pure" A. /. tuherosa.

An interesting aspect of the quadrats, or introgression, gradient between sub-
species inferior and fubcrosa is shown by use of the broken line in fig. 14. This
line, purely a hypothesis, is drawn from quadrat Gil in the central Ozarks to
quadrat Gj in the eastern Appalachians, where the herbarium sample approximates
the estimated ZB parameter of "pure" ssp. tubcrosa.

The hypothesis of the broken line is that there is a perfectly balanced diffusion
of genes between subspecies infcrioty emanating from G//, and tuberosa^ emanating
from G^, in much the same manner as we might visualize the diffusion of the
molecules of two equally miscible inorganic compounds. Since we have good evi-
dence of the centrifugal lowering of Z B values in interior^ and are assuming the
stability of a value of 50° in tuberosa^ the broken line is plotted at equally graded
intervals between the "circles" curve and the 50° abscissa, to show 100 per cent
interior and per cent tuberosa influence at quadrat Gil, per cent inferior and
100 per cent tuberosa influence at Cj, 50 per cent influence of both interior and
tuberosa at G8, and so forth, for the intervening quadrats. The points along the
broken line, rounded for convenience, as well as the percentage values of the cor-
responding points along the quadrats curve, are provided in Table XIII.

A comparison of the broken Hne with the quadrats curve in fig. 14 will show
the consistent way in which they differ, for we see that the inferior influence is
greater toward the western half of the cline, and that of tuberosa greater toward
the eastern half than we would expect if genes were exchanged equally between
the two subspecies, other factors being removed. The surplus interior influence
is seen to increase toward the east, progressively from Gil to GS^ whereas that of
tuberosa would appear to decrease toward the west. Finally, in Gj, possibly be-
cause we have set the estimated parameter of "pure" tuberosa too low, the value
obtained from the quadrat data appears to show the influence of Rolfsii. ,Thc
latter feature may be neglected as inconsequential for our immediate purpose; but
the centrifugal Increase of interior Influence and decrease of tuberosa influence is
of more interest.

1947]

WOODSON— LEAF VARIATION IN ASCLEPIAS TUBEROSA

427

TABLE XIII

COMPARISON OF /B QUADRAT PERCENTAGES AND MEANS AS EXPECTED WITH
SIMPLE GENE DIFFUSION AND AS ACTUALLY OBTAINED IN THE INTROGRESSION

OF A, T, INTERIOR AND A. T. TUBEROSA

(Explanation in the text)

Quadrat

Gil

Gio

1

G9

% intj'tub.
expected

loo/o

83/17

67/33

% hitjtuh.
obtained

loo/o

94/6

8l/l9

X expected

111.2

94,6

85.3

X obtained

i

in.i

100.5

1

92.9

d/<T

0.0

1

3.5

2.1

G8

50/50

76/24

75.9

89.1

2.2

G7

G6

33/67

27/73

17/8 3

5/95

63.7

61.7

0.3

56.5

52.0

0.8

G5

0/100

2/9S

50.0

50.8

0.2

Table XIII also presents the quadrat means expected with simple gene flow of
the sort we hypothesize, together with the means actually obtained. Finally, the
coefficient of abmodality, d/n, is calculated. A frequent biometric practice is to
regard d/a values >2 as significant and values <2 as probably not significant.
If we apply this arbitrary rule to our calculations, we find that the centrifugal
increase of interior influence probably is significant, while that of tuherosa prob-
ably is not. The means obtained from the quadrat samples in G/, Gd, and G^,
therefore, may be regarded as coinciding sufi'iciently with the hypothesis, and the
apparently surplus tuberosa influence as dubious.

Several explanations present themselves to account for the interesting prop-
erties of the introgression gradient; they are not mutually exclusive nor do they
exhaust all possibilities:

7. Natural Selection, — In the previous discussion of the centrifugal effect
proper, we have seen that the larger number of flowering stems per plant, as well
as the increased general vigor, characteristic of the cordate-leaved race of sub-
species interior, almost certainly would tend to increase the reproductive potential
of the plants, and thus might well be of importance in natural selection. Hence,
the selective superiority of inferior might enable it to extend more rapidly and
perhaps into a greater variety of habitats than subspecies tuberosa. If this explana-
tion is worthy of consideration in connection with the aggressive role of interior
in introgression with tuberosa, it probably is worthy of consideration as a causative
agent in the change of gradient ("hump") which we noted originally in fig, 11;
/. ^., the supersedure of the ancestral race of interior with cuneate- or round-based
leaves by the younger, cordate-leaved race from the central Ozarks may be due to
selective superiority associated with the latter.

[Vol. 34

428

ANNALS OF THE MISSOURI BOTANICAL GARDEN

2. Migration Pressure. — The quadrats gradient of fig. 14 recalls the obvious
fact that plants tend to migrate outward from their centers of dispersal through
the agency of dissemination and other methods of transport. In this instance, the
migratory powers of interior may be taken as greater than those of tubcrosa. Since
the seeds of Asclepias are classical examples of wind-dissemination, we may con-
strue the direction of the prevailing winds in this section of the United States,
from west to east, as favoring this argument.

5. Dominance, — The gradient may be indicative of the relative dominance of
the inferior genotype and the recessiveness of that of tnbcrosa. According to
Fisher's theory of dominance, I take it that this possible "explanation" is merely

I

a corollary of the first.

4. Genetic Structure* — The different degrees of aggressiveness of interior and
tubcrosa^ on the other hand, may be due to a difference in the structural natures of
the respective genotypes which makes the diffusion of the ^^interior influence"
somewhat more rapid than that of fnberosa. For example, a large gene complex,
rather diffuse with respect to loci, might be more easily redistributed in crossing-
over than a smaller, more compact one. This suggestion Is wholly conjectural
and cannot be supported by our present technique, although it deserves theoretic
consideration.

If the surplus influence of interior which we have just demonstrated with re-
gard to Z B is due to migration pressure, we would expect to find a similar surplus
with regard to ZA. Accordingly, fig. 15 compares the /_ A quadrat means as
expected in simple diffusion of genes and as actually obtained in the Introgression
of A. /. interior and A. /. ttiberosa. The method is precisely the same as that em-
ployed in the construction of fig. 14, the estimated parameter of "pure" interior

TABLE XIV

COMPARISON OF /A QUADRAT PERCENTAGES AND MEANS AS EXPECTED IN
SIMPLE GENE DIFFUSION AND AS ACTUALLY OBTAINED IN THE INTROGRESSION

OF A, T. INTERIOR AND A, T. TU DEROSA

(Explanation In tKe text)

Quadrat

% hit. /tub.
expected

% hit J tub,

obtained

Gil

100/0

96/4

X expected

85.5

X obtained

85.7

d/tr

0.3

GIO

83/17

93/7

86.5

85.9

1.3

G9

67/3 3

6O/4O

87.4

87.7

0.6

G8

50/50

46/54

88.2

88.5

0.6

G7

33/67

30/70

89.2

89.4

0.2

G6

17/83

16/84

90.0

90.1

O.I

G5

0/100

4/96

91.0

91.1

0.5

1947]

WOODSON^ LEAF VARIATION IN ASCLEPIAS TUBEROSA

429

Gil

GlO

09

Gd

G7

Q6

ae

Fig. 15. Comparison of /_K quadrat means as ex-
pected with simple gene diffusion and as actually obtained
in the introgression of A. t. interior and A. t. fnbero^a.
Explanation in text.

in this case being 85.5° and that of "pure" tiiberosa 91'^. The percentages of
"influence" and means, expected and obtained, as well as d/n values for each
quadrat, are provided in Table XIV.

In spite of the fact that the ordinate scale of fig. 15 is made ten times greater
than that of fig. 14, to compensate for the correspondingly discrepant range scales
of Z B and Z A, the close correspondence of the broken line representing the
hypothesis and the continuous curve of the actual gcnoclinc is apparent. The
small d/(j values for each quadrat support the assumption that the genocline,
obtained by use of the herbarium specimens, is essentially what one would expect
to result from the equal diffusion of genes, and that other factors apparently are
of negligible importance in introgrcsslon as far as these genotypes arc concerned.
In fig. 14, then, I believe that wc may assume the strong influence of subspecies

[Vol. 34

430 ANNALS OF THE MISSOURI BOTANICAL GARDEN

interior In tlie western quadrats to be due to the selective superiority of the
corclatc-lcaved Ozark population. The same factors may likewise be assumed re-
sponsible for the concentric modification of Z B observed within A, /. inferior.

VII. Discussion and Summary

Asclcfrias ful/crosa^ popularly known as "buttcrflyweed," is a species of herba-
ceous perennials distributed from Ontario to Sonora, and from Minnesota to
Florida. The plants occur chiefly in colonics of from few to over a hundred
individuals in a wide variety of habitat from near sea-level to about 6000 feet
elevation. They are facultatively self- or cross-fertih/ed through Insect agency,
and disperse their comose seed by air currents.

Three subspecies comprise A. tuhcrosa: A. /. itttcriory centering in the Ozark
plateau and extending to southern Ontario, the Rocky Mountains, and northern
Mexico, A, t. tuhcrosay centering In the southern Appalachian Mountains and ex-
tending to the Atlantic coast, and A. t. Rolfsii^ which centers in Florida and
extends onto the coastal plain of Alabama, Georgia, and the Carolinas. Since the
subspecies apparently are quite panmictic and freely hybridize at the present time,
it is assumed that they have had their origin in isolation on the Paleozoic and Early
Mesozolc land masses, Ozarkia, Appalachia, and Orange Island, respectively.

The purpose of this study is to Investigate the population patterns of the three
subspecies, particularly with regard to tlielr apparent introgrcssive hybridization.
This is accomplished by the measurement of leaf characters, which provide the
principal systematic criteria of the subspecies. Unfortunately, it is found im-
practicable to distinguish leaves of all three subspecies by means of a single
continuous scale, so only those of interior and tubcrosa actually are measured;
information concerning the role of Rolfsii is obtained only indirectly in so far as
it influences the other data. Measurement of median length and width is in milli-
meters, and of apical taper ( Z A) and shape of base ( ZB) in standard degrees of
declination, according to a rather elaborate procedure. Leaves arc measured from
herbarium specimens and from natural colonies growing in the field. The former
are considered to yield the best estimate of the natural parameters of the subspecies
for various reasons.

From the herbarium data separate phenocontour maps arc constructed for /A,
Z B, median length, and median width. Isophene systems are plotted to indicate
gene flow between subspecies inferior and fuherosa with respect to ZA and ZB,
the two characters which best differentiate the populations from the systematic
standpoint. In the map for Z B, "crests of variability," indicative of maximum
heterozygosity, occur midway of the gcnocline of inferior and fuhcrosuy and also
about midway between the centers of dispersal of tubcrosa and Rolfsii.

In the maps for length and width, more or less conspicuous increase In leaf
size is seen to be associated with the former "crest of variability," and this is in-
terpreted as an indication of hybrid vigor. No heterosis is found to be associated

1947]

WOODSON— LEAF VARIATION IN ASCLEPIAS TUBEROSA 431

with the second crest, and neither crest of variability nor heterosis is associated
with the gcnocline of Z A. It is not practicable to draw isophenes for length and
width, but both geoclines and ecoclines are observed, the latter apparently related
to restriction of population size under the influence of selection pressure, particu-
larly in the southwest. In all characters, the individual colonies behave as micro-
geographic races, which are combined variously to form the major population
patterns.

One of the most interesting phenomena to be demonstrated is the concentric
diffusion from the Ozark plateau of a special modification of / B which, because
of certain associated selective advantages, appears to be supplanting the ancestral
race. A crest of variability is associated with the diffusion of this character, which
probably dates since the Pleistocene, but there is no apparent heterosis. For
various reasons, it is interpreted as possibly due to a major change of genotype.

The close correspondence of the centrifugal modification of Z B to Matthew's

(1915) hypothesis of the centrifugal migration of primitive organisms is obvious.

Although our data confirm the general thesis of Matthew, viz,^ the peripheral

distribution of the supposedly primitive forms, climate clearly Is not an active

agent In this case.

Analysis of natural populations suggests that Z A and ZB are due to multiple
factor complexes between which there is slight linkage. Segregation apparently is
Mendelian.

Introgresslon In the subspecies of A. tiibcrosa is believed to proceed from
initial hybridization through back-crossing to other heterozygotes and both an-
cestral types, respectively, to produce a more or less perfect gradation of genotype.
Gene flow from either ancestral type is thought to be facilitated by crossing-over
to effect redistribution, and the velocity possibly may be dependent upon the dis-
tribution of genes, particularly with respect to loci. Velocity of gene flow is
observed to be directly proportional to population density. Although gone flow
may be equal In either direction, as in Z A, associated selective advantages of one
gene complex may Inequallze the balance, as In Z B. Gene flow may be inequalized
also by factors mechanically influencing migration pressure, such as the direction

of prevailing winds; this, however, probably Is of minor importance in Asclcpias
tuberosa.

The most obvious evolutionary role of introgressive hybridization might appear
superficially to be merely the negative one of the obliteration of previous specific
or subspecific distinction. However, the resultant increase of potential variability
may play an important part In subsequent systematic differentiation.

Introgresslon of A. /. interior and A. t, tuberosa appears to proceed more
rapidly along roadsides than in undisturbed areas.

During my studies of Asckpias tuberosa I have been the grateful recipient of
favors from a multitude of friends, unfortunately too numerous to mention Indi-
vidually. These unselfish people, widely distributed throughout this country,
enabled me to continue the investigation of geographical variation during the war

[Vol. 34

432 ANNALS OF THE MISSOURI BOTANICAL GARDEN

years when extensive travel was impossible. To my colleagues Edgar Anderson,
Richard Holm, and Harrison Stalker, I am particularly indebted for their generous
attention, encouragement, and suggestions. I am quite sure that I would not have
been able to bring my work to conclusion without the Instructive companionship
and loyal assistance of Mr. Holm.

References

Alpatov, W. W. (1929). Btonictrlcal studies on variation and races of the honey bee {Aprs mclff-

fera L.) . Quart. Rev. Biol. 4:1-58.
Anderson, E. {19}6). The species problem in Iris. Ann. Missouri Bot. Gard. 23:457-509.
, (1941). The technique and use of mass collections in plant taxonomy. Ibid. 28:287—292.

, and L. Hubricht (1938), Hybridiyarion In Tradcscatifia: The evidence for introgressive

hybridization. Amer. Jour. Bot. 25:396—402.

, and R. E. >X^oodson, Jr. (1935). The species of Tradescatjfia indigenous to the United

States. Contr. Arnold Arb. 9:1-132.
Boyd, W. C. (1939). Bbod groups. Tab, Biol. 17:113-240.
Czcczottowa. H. (1933). A study of the variability of the leaves of beeches: F. oricniaJls Lipsky,

T, sslvafica L., and intermediate forms. Part T. Ann. Soc. Dendr. Pol. 5:45-121 (English

summary: pp. 110—121).
Dice, L. R. (1940). Ecologic and genetic variability within species of Pcrofnyscus. Amer. Nat.
74:212-221.

Dob/hansky, T. (1941). Genetics and the origin of species, cd. 2. Columbia Univ. Press. New

York.
Erickson, R. O. (1945). The Clematis Vrcmontii var. Richlii population in the Ozarks. Ann.

Missouri Bot. Gard. 32:413-460.
Fisher, R. A., F. R. Immcr. and O. Tedin (1932). The genetical interpretation of statistics of the

third degree in the study of quantitative inheritance. Jour. Gen. 17:107—124.
Huxley, J. S. (1938). Clincs: an auxiliary taxonomic principle. Nature (London) 142:219-220.

, (1942). Evolution: the modern synthesis. Harper & Bros. New York.

Kestevcn, G. L. (1946). The coefficient of variation. Nature 158:520-521.

Lewis, H. F. (1947). Leaf variation in Dclphinivm varic^atnm. Bull. Torrey Bot. Club 74:57-59.

Mather, K. (1942), The balance of polygenic combinations. Jour. Gen. 43:309-336.

, (1943). Polygenic inheritance and natural selection. Biol. Rev. 18:32-64.

Matthew, W. D. (1915). Climate and evolution. Ann. N. Y. Acad. Sci. 24:1.71-318.

Moore, R. J. (1946a). Investigations on rubber-bearing plants. ITL Development of normal and
aborting seeds in Asclcpias syriaca L. Can. Jour. Res. C. 24:55-65.

, (1946b). Ibid, IV. Cytogenetic studies in Asclcpias [Tourn.] L. Ibid, 24:66-73.

Muller, H. J. (1936). On the variability of mixed races. Amer. Nat. 70:409-442.

Pearson, T. (193 8). The Tasmanian bush opposum: Its distribution and colour variations. Papers

& Proc. Roy. Soc. Tasmania 1937:21-39.
Robertson, C. (1892). Flowers and insects. Trans. Acad. Sci. St. Louis 5:569-598,
Schuchert. C. (1935). Historical geology of the Antlllcan-Caribbean region. John Wiley & Sons.

New York.
, and C. O, Dunbar (1933). A textbook of geology. Part II — Historical geology. John

Wiley & Sons. New York.
Stephens. S. G. (1944). The genetic organization of leaf-shape development in the genus Gossypium.

Jour. Gen. 46:28-51.

Stevens, O. A. (1945). Cultivation of milkweeds. North Dakota Agr. Exp. Sta. Bull. 333:1-19.

Sumner, F. B. (1927). Linear and colorlmctric measurements of small mammals. Jour. Mam-
malogy 8:177-206.

Walker, H. M. (1943). Elementary statistical methods. Henry Holt & Co. New York.

Wicgand, K. M. (1935). A taxonomist's experience with hybrids in the wild. Science, N. S.
81:161-166.

Woodson, R. E., Jr. (1944). Notes on some North American Ascleplads. Ann. Missouri Bot.
Gard. 31:363-370.

, (1947). Notes on the "historical factor" in plant geography. Contr. Gray Herb.

Harvard Univ. 165:12-25.

FIELD STUDIES OF GUATEMALAN MAIZE^

EDGAR ANDERSON

The variation pattern of T^ca yiay^ is surprisingly like that of man It is made
up of a number of poorly defined geographical races and sub-races, some of which
characterize wide areas while others are of restricted distribution. The members
of any one population vary greatly one from another, and ordinarily it Is only by
statistical methods that one can demonstrate regional differences.

In maize, as in man, there are centers of variation in which strikingly different
forms are found in a comparatively small area. For the maize plant one of these
centers is western Guatemala where, according to Mangclsdorf and Cameron
(1942) '*in an area less than half the size of the state of Iowa arc found probably
more distinct types of corn than occur in the entire United States." This great
variability of Guatemalan maize has attracted numerous collectors and is one of
the reasons why Iowa State College recently established a Tropical Research Center
in Antigua, Guatemala. However, judgments with regard to the comparative
variability of Latin American maize need to be made with greater caution than
they have been in the past. Most of the corn in the United States corn belt is
uniform in color. Much Latin American corn has not been so rigidly selected for
that feature and to our eyes looks more conspicuously variable than it really is.
As every geneticist knows, a few segregating color genes can give the impression
of great variability to a population which is relatively uniform morphologically.
As we shall demonstrate below, Guatemalan fields are, morphologically, among
the most uniform which have yet been studied, though there is indeed a great
variation between different varieties.

Unfortunately, most of the collections of maize from Guatemala are of separate
ears bought in the market or obtained at agricultural exhibitions, or bought from
farmers. Maize, however, is a cross-poHinatcd plant and single ears are therefore
not as significant as in some other crops. In wheat, which is almost continuously
self-pollinated, a single spike, if well chosen, may be an efficient representative of
that variety. In maize (extremely heterozygous and nearly always cross-pollinated
under natural conditions) a single individual is somewhat of an accident. Out of
all the millions of gene combinations which might have occurred in a particular
field any single ear is one of the relatively few gene combinations which did come
into existence. Unless carefully chosen it is not an efficient reflection of the gene
frequencies in the field where it was grown. Were It selected to be representative
it might have more significance but, as Cutler (1946) has recently shown, the very

^Thls work was made possible by a special grant from the Pioneer Hi-Bred Corn Company and
Dr. I. E. Melhus, Director of the Iowa State College Tropical Research Center, who kindly made
available the facilities of his laboratory during my stay in Guatemala. I am also indebted to the
Office of Foreign Agricultural Relations and to Director Antonio Goubaud Carrera, of the Institute
Indigenista Nacional of Guatemala, for many favors.

(433)

[Vol. 34

434 ANNALS OF THE MISSOURI BOTANICAL GARDEN

>.

reverse usually takes place: it is the unrepresentative car which is odd-looking and
gets the attention of the collector. Cutler has actual statistics on this point and
they arc impressive (loc. clt., p. 261) :

In the case of maize, colored or freak cars frequently receive more attention tlian normal
ones. For example, in a harvest of 8000 ears at Santiago de Chiquitos, Bolivia, only four ears
differed from the predominating type, yet in a collection representing this lot, three of the
atypical cars were included and only four of the major type.

There is a further reason why collections mnde in the market place may not be

at all representative. Some of the most distinctive and significant types of maize
do not come to the market on the car and have been missed by collectors. Through
most of Latin America there arc old-fashioned types of corn which seldom or never
appear in the markets but which are used for particular purposes, most of them
survivals of pre-0)lumbian practices. Three examples of such specialty corns are
popcorns, brewing corns, and sweet corns.

Popcorns arc very widespread in Latin America. A glance at the ethnological
and agronomic literature will show that they have been obtained by very few col-
lectors. This is largely because, if marketed at all, it is the popped corn which is
sold, usually in various sweetmeats, while the ears do not generally appear. Brew-
ing corns are widely distributed; how widely wc cannot say for certain until
careful collecting has been done. Since in many places brewing, though common,
is illegal, it requires tact and experience and persistence to obtain ears of such
varieties. Sweet corns (i.e., maize with one or more of the recessive genes for
sugary endosperm) are widespread In Latin America. They are seldom used for
green corn as in the United States. They were apparently pre-Columbian sugar
sources and have survived in the manufacture of certain distinctive beverages and
sweetmeats. They are known in Mexico, Guatemala, Peru, Ecuador, Bolivia, and
from the Hopl Indians in the United States. In most of these places the cars are
not commonly sold in the market, and it is only through careful field studies that
wc shall be able to map their present extent and trace their probable history.

For the above reasons the maize samples reported on in this paper w^cre taken
from fields, from drying floors immediately after harvesting, or from the cribs
(trocbas)^ where the maize was being stored. In so far as possible they are ran-
dom samples of 25 cars from each field. When for any reason they do not represent
such a sample the fact is so stated in the Appendix. In making selections from a
field the two or three outer rows of plants were avoided as being unrepresentative.
An ear was taken from a randomly chosen plant in any one row, from the fourth
plant in the next row, from the eighth In the next, and so on diagonally across the
field until 25 were secured. In making crib or drying-floor selections the calipers
were thrown out onto the surface of the ears and the ear nearest their tip w^as
chosen for measurement. However, nubbins and poorly filled ears were of neces-
sity rejected. It may be well to discuss the reasons behind this decision. In the
part of Guatemala where these studies were made, as in most of the United States,
each normal plant bears one perfect ear, though all the nodes below the one bearing

Guatemalan spelling of "troche.'*

1947]

ANDERSON GUATEMALAN MAIZE 435

the ear are also potentially fertile. In a small percentage of cases one or more of
these lower nodes bears an imperfect ear, as do so also occasionally the axillary
shoots from the base of the plant (the '^tillers" or "suckers") . Most of these small,
imperfect ears can be recognized at a glance by an experienced person. In the
United States they are called "nubbins"; in Guatemala, "mulco" is the commonest
name in the Antigua region. To have included measurements of them with
measurements of the upper ear would almost be like including a few leg measure-
ments with arm measurements when studying a human population. The im-
perfectly pollinated ears are rejected because their kernels do not develop normally
and measurements, such as width of kernel, would be almost meaningless.

It may also be well to discuss the actual way in which the collections were
made and the measurements obtained. During the long dry harvest season in up-
land Mexico and Guatemala it is a simple matter in almost any town to find a
field which is being harvested or a patio or drying floor where maize is spread out
to dry. If the town is Spanish-speaking a request at the gate is almost never
denied, particularly if one explains that he does not wish to buy the corn but
merely wants to study it. Once permission is obtained one sits down by the
corn (or actually on it in the corn crib) and measures his sample of 25 ears. This
takes from one to two hours depending upon interruptions. (A pocket full of
hard candies to pass out to the children of the household is almost as indispensable
as a sliding micrometer.) After the 25 ears have been measured a few of them
are photographed, and then one courteously takes his leave. This process is so
simple that one can scarcely dignify it with the name of a technique and yet it is
of real importance in taking cflFicient samples of Latin American maize and in-
terpreting the results of such sampling. It is time-consuming but the time is well
spent. For the first few moments most of the family stands around watching, and
then the spectacle of a strange foreigner carefully measuring ears of the corn be-
comes dull, even for a Latin American family, and the normal life of the household
begins to go on its usual way. Life in such homes is centered about the patio and
as one sits there busy with the corn, he learns, incidentally and in a painless kind of
way, a great deal about the family who owns the corn. When the men have gone
back to work the old grandmother of the family will enjoy discussing the ways in
which corn is made for food in her family, and frequently she can supply informa-
tion about brewing corns or popcorns that cannot be obtained from local agrono-
mists or corn merchants.

Maize is a sensitive mirror of the people who grow it. It is so highly heterozy-
gous that good or bad management and careful or careless selection leave their
imprints upon the character of the population. There are so many kinds of corn
and they are so different and yet cross so readily that the introduction of alien
sorts leaves a permanent witness of the mixture. One cannot interpret popula-
tion samples of maize efficiently without understanding as much as possible about
the people who grew that maize. The long, dull hours spent in measuring the
samples of maize yield a priceless harvest of understanding.

[Vol. 34

436 ANNALS OF THE MISSOURI BOTANICAL GARDEN

The actual results from the Guatemalan studies arc presented In Table I and
also in the Appendix. The characters measured have been described and discussed
in full in **Maizc in Mexico" (Anderson, 1946). This discussion need not be re-
peated here, other than to point out that the characters were chosen after prelim-
inary studies In the field and in the experimental plot. The scatter diagrams used
in the Appendix attempt to present a picture of the population sample in one
simple, easily grasped diagram. As explained in the introduction to the Appendix,
the scoring of kernel texture^ has been made more objective since "Maize in
Mexico" was published. It is now scored in 6 grades as follows:

0. No soft starch at apex of kernel

L Soft starch but no denting

2. Soft starch and a small dent

3. Soft starch and a deep dent but no wrinkling of the pericarp.

4. Soft starch and wrinkling pericarp

5. Soft starch and the apex of the kernel collapsed

In Guatemala, even more than in this country, the kernels at the tip and the
butt of the ear are often different from those in the middle portioii. The scores
for kernel characters attempt to reflect the average condition of the middle third

of the ear.

Through the courtesy of the O.F.A.R. it was possible to make an experimental
test of the reliability of the methods of sampling and measurement used in these
studies. At the experimental plot at Quezaltenango a common yellow variety
from Salcaja had been used in a series of fertilizer test plots. Numbers 1 and 2
in the Appendix show the result of sampling the untreated plot and that to which
phosphorus and nitrogen had been added. A random sample was taken from each
as It was drying after harvesting, and photographs were made of the corn as it
lay out in the sun. As shown in the Appendix, there are only very slight differ-
ences between the two samples by the methods used In this study. Maize is so
variable and is so visibly affected by differences in soil fertility that I have fre-
quently been asked by agronomists as to how much confidence could be placed in
my 25 ear samples. Since the standards by which these sampling methods had been
developed are essentially those of most taxonomic work it has been difficult to
give an intelligible answer to those unacquainted with taxonomic practice. Though
they seldom put it in words, taxonomists learn to choose characters which are
relatively stable under environmental variation.

My general approach has been to work out methods of sampling and to choose
characters for study which would give consistent results for repeated sampling of
the same field, or for different samples of the same variety, and yet were efficient
in distinguishing between varieties which were manifestly different. The methods
have been used with increasing confidence when they demonstrated the regional

*'See Anderson & Cutler (1942) for a discussion of the reasons for abandoning "flour," "flint,"
and "dent" as special texture categories in Central America.

1947J

ANDERSON GUATEMALAN MAIZE

437

>

o

t3

5

trrcTQ

n

41

O

n
o

n

o

s I?

O

O ^« -

o

O

3

^

3

(A

-U

ft

D

c
3

ft

o

o

^

d

O

D cr
o

D ft

5 <

1-1 o

n

3

o

o

o

^ s

»3

n

o

3

*<;

Q

ft

09

p

c

a

w

O

o o
e d

z^>>^^^

c
ft

no

O

n)

0^"^*^

fl» <^ 63 ^^~ftj'

J== ^ 5 C:'P

2: 3". S 2^3

^ _^ N :

^ ?

3-

ft ;;

" i

ft

I I I I

NJ '-w K) V'

I I I I

Uj 4^ \.^ 4;^

I

I

I I I

I

I

I

I

i

I

ON

ii

i

4x

I I

O K» ^0

I I

4:-

T

^^

ON NJ

I

I

I

i ^ i -t.

^ 00

^ 00 \j\ \/\ \n
I I I I I

I

K* 00 \0

I I I

H-> ^>* 4n. \«n VI

I I I I i

I— I I— » to Ni N)

(^ \D ^-*J t— I VI

I

00

I

I 1

\0 vo

OOOCONJOOOOOOO
I I [ I t I I I I

to

I

ON -ti.

I

OS 4^
I 1

hJ CO

K> hJ 00 00

I I I I

4i^ 4^ K) N)

ooonooOnogcoooo
V- XT ^"^ XT v- N- vr

KJ^ N*^ hJ^ N;^ hj^ Isj^ h>v ^

I I I I I I I I

NJ

COOOOGOOOOK*N>

^s^^ hj^ Ki^ kj^ hjN 1^ ^

I I 1 I I I I

1 I

O00K*OlsJN>4i.4i.4^

t^^ SJ^ KJ^ Ki^ Isj^ |s>- hj^ ^jV jg

K) K> N) ro ro fo
\^ ^yi wi VI ^yi \-/i

2

Q/5 ^

P w

3 T

>* 3

O

c

ft

p *■ ^

P

3

-^ 3

P ;ii

3

P

o

3

O org
■<

ft

I I
I I

1

1

hJ

K>

1

t

\^

\^

CO

I

i

I

I

I

00
I

I

I

I I

ON H-

i I

\0 N

o \o

I I

4^

I

\J\

ISJ

00 ro o o
I 1 I I

O 4^
I I

K> O

K> (sj
I I

00 OS

O 00 00 c\

hJ^ Ki^ K)^ Kj
I I I I

K*^ Kj^ Kj^ ro

I I I I

N> 4i. K> ro

K»^ K)^ KJ^ K)

^- N> N> ro

VrfJ ^^ \^ \^

p
n
p

ri-
ft

c

ft

o

o

>

n
p

lyi c/5 tyi CA) c/i c/i c/3

&j p p lu p p p

3333333

r-t rt «-t

P P P

r-" r r

c d d

n n o

p p p

d 3 3 C

o n P P

p p 3 3

*^ ^ r% ^

_ d- O-

3 w

e: 3

o

p

3

I I

I I I I I I I I

■..M \.^ ^» ^

^^ ro \o ^
4^ \^ 4^ \j\

K> S£> ^-^ '-M

I

ON

4^ 4^

\^

i

I

I

I
VI

CN
I

VI

00 ON

i

-t*' Vt

NO

t 1

v^ 4*.

I i

I

I I
ro N)

4^ \j}

I I

fO ^j^

4*- '^H

ro ►—

I
ro

4^
I

ro

I

to
I

I

I

to ro ro

00 »,yi \Ji

CD 00

I I I I I I

N>ONJNJrVN)00
I I I I I I

oots>c^sON4*4^4^QN

OOCCCjOOCCOOOOO\

rJ^ t^ t^^ KT* hj^ Nj^ Kt

I I I

ISJ

K)^ kT^ KJ^ M^ K)^ kt- n»

I I I I I I

h>

I I

^ro4^rororororo

N>^ K)^ hJ^ Ki^ N^ KT^ Ki^ KJ

rorororororororo

>

3

rr

d
p

31

^D

&^

^^ # ^"^ »

C/i

^ ^

4

4 "4

^

>

3 ^

r-f h-i

a

a> h2.

erg

3»

n

GO

p

C fD S

(JO

p

3
>

QA

o

3

O
3

O
>

d
p

O

ft"

090

I I I

I

I t i

I

^

1

\^

T

1

T

',^

-t-

uj

4^ 4^

1 I

4»- ^0

I I

ON VI

VJ 4^ N

1

I

CN

4^ 4^ -U
O OJ 4^

iii

\0 \| ^

I I I

VI V» Vt

\\> \j\ \o

1

I

4^

I

10

I

ON

ro

I
ro

ro
I

4v

u^ 4i.

I I

ro

ro
I

I

ro

CN

NO '^

I I

ro ro

ON O

I

ro
ro

I

OJ

1

I

ro

ro 4*. o ro ho ro o
I I I I f I I

4^ 4^ 4* 4^ 4^ ON 4^
I I t t I I I

On 00 o\ 00 00 00 oc

ft

o'

3

p
ft

ft

3

3

3

3

CL. P

D-

3
3

p

* CA)

3 p

•^ d

3 3

O 00 C^ CpO cc

K)^ hj^ "^ K/^ K)^ Ki^ hj'
I I t I I t I

o o o O 00 00 o

NP- N^ K)^ Nj

till

hJ^ hj^ Ki

I I

ro ro 10 o o o

[
ro

rJ^ KiN rjN (jN (^ jjN K,

3
3

3
ft

ro ro ro ro to 10 ro

vyi s^ wi '^n 01 >syi <*/!

3

c:

ft

c

3

cr

ft

^

r
n

td

n
o
r
r

n

O

>

H

r
>

LVOL. 34

438

ANNALS OF THE MISSOURI

differences of Mexican maize, and the gradual transition from one region to an-

Howev

vincing to most agronomists. It is gratifying, therefore, to present experimental
data on this point. As can be seen from the photographs, the Salcaja maize differs
markedly in yield and vigor and in percentages of imperfect cars under the two
treatments. Random selections of well-filled ears, however, yielded two similar
samples from the two plots. This is a demonstration of the fact that the characters
we have chosen for measurement have a strong germinal basis and that under con-
ditions which will produce approximately normal plants they are not greatly

affected by soil differences.

The 30 collections presented in Tables I and II and in the Appendix are from
three nearby regions, all in the highlands of Guatemala. Seven are from the
Antigua basin at elevations of about 5000 feet. Twelve are from plateaus near
Antigua but from 500 to 1000 feet above it. Of these 12, 8 are on the San Lucas

region. Nine sam-
ples arc from Quezaltenango, a little over 50 miles to the northwest of Ant
and 1000 feet higher.

tenango

TABLE II

COMPARATIVE VARIAT^ILITY OF MATZK COLLECTIONS FROM GUATEMALA AND

F7?OM WKSTIIRN MEXICO

Maximum number of cars (per
2 5 car sample) -with same
row number and kernel
width.

5

Guatemala (29 samples)
Western Mexico (29 samples)

2

5

6

7

8

9

4

7

3

3

4

7

10

2

2

3

3

II

2

1

14

15

16

1

2

17

1

Although the numbers of concctions are small, two generalizations can be
established from the facts summarizcj in Table I. Both of them find further
confirmation in the Appendix: (1) For the characters measured there is a general
trend with altitude for most of the characters. From Antigua (5000 feet) to the

plateaus above it

Q

(6000 feet), the ears become

generally smaller, with smaller shanks and fewer row numbers. (2) Differences
between varieties are greater in the Quezaltenango region.

One quite unexpected fact is demonstrated over and over again in the Ap-
pendix. Much of the common maize in Guatemala is highly uniform. As will be
shown below, there Is Indeed great variation In type in certain parts of Guatemala,
as Mangclsdorf and Cameron pointed out. It is all the more surprising, therefore,
to report that in many fields and even within certain regions, the plant-to-plant
variation of 7.ea May^ is less than in any other region we have studied, including
even the highly selected open-pollinated varieties of the United States corn belt!
This is particularly striking In the Antigua region, but the same general tendency

1947]

ANDERSON

<;UATEMALAN MAIZE

439

Text-fig. 1. Popcorn balls (some of them wrapped In corn luisks) bein_i; offered
for ^aie by an Indian woman (upper left) in the market at Patztin, December, 1946.

Text-fig. 2. Corn just brouglit in from the field, dr)ing in tlie )ard o( an Indian
home in San Antonio Aguascaliemes. Note its uniformity to type.

440

V.)i-. 34

ANNALS Ol THE MISSOURI BOTANICAL GARDEN

3

4

Text -figs. 3 and 4. Corn in tlu' dtyini; ynrtl of a iion Spanish -speaking; Indian
abo\c Zunil. Superior .mJ uniform cars laid out to llic ri^Iit (close up in tlio lower
picture) arc apparently beinj^ saved fur seed ears. Note that a few cars wiili red
pericarp are being deliberately included \s'itli the white ears, a very widespread practice
in the New World among primitive peoples.

1947]

ANDERSON GUATEMALAN MAIZE 441

will be found to run throughout the collections. It is particularly striking when
the scatter diagrams of the Appendix to this paper are compared with those in the
report on Mexican maize.

The maximum number of specimens for one of the cells on the scatter diagram
is presented in Table II for the 29 Guatemalan collections which had 25 ears each.
It is contrasted with a similar summary of the first 29 such collections in the Ap-
pendix of "Maize in Mexico." It will be seen that while the Mexican maximum
varied from 5 to 12, the Guatemalan maximum went up to 17. The average
(median) value of the cell maximum was 7 for western Mexico and 9 for Guate-
mala. Yet the Mexican collections were made in an area where there are strikingly
fewer types of corn per 100 square miles than in the Guatemalan area.

Four of the collections from the Antigua region have from a half to two-
thirds of the sample falling in the same cell of the scatter diagram, which means
that they have the same row number and do not vary more than one millimeter in
kernel width. Somewhat of the same stability is shown when we make comparisons
between varieties in the Antigua region. Of the 25 ears each, of the 7 samples,
68 (or just over one-third) fall in the same cell (14-rowed, 10-11 mm. width of
kernel) on the scatter diagram.

The reasons for this stability of Guatemalan maize require further study. In
part, at least, they rest upon a rigid selection for physical type in picking seed ears.
Among pure-blooded Indian farmers maize is often very carefully selected for type.
Figure 3 shows the maize dr>^lng floor of an Indian family living on the slopes of a
volcano above Zunil. The family was not Spanish-speaking, and those remaining
at home fled when we approached. However, as shown In the photograph, some of
the best ears had been segregated at one side of the pile. They are obviously seed
ears for next year. The rigid selection for type is about on a par with that prac-
ticed by farmers of the United States corn belt in the days before "hybrid corn"
when most farmers selected their own seed ears.

Of the 1 1 samples of white corn 6, as shown in fig. 4, had their averages in the

same cell of the scatter diagram, with row numbers of 14 and kernels 10-11 mm.

wide. It represents a common, well-marked type In the Antigua region where it

was seen repeatedly in markets and being packed to market, on drying floors and

in the fields. It has (pL 47, left) an ear of about the same size as United States

cornbclt varieties, slightly enlarged at the butt, tapering gently to the tip, usually

straight-rowed except for the enlarged basal portion, and very commonly 14-
rowed.

. As shown in fig. 5, the yellow corns of the same region are fewer- rowed on the
average. They too represent a common, widespread type, particularly at higher
elevations. On the highlands between Solola and Totonicapan they represent nearly
all the field corns of that area. They are most commonly 8- to 12-rowcd, with an
even more strongly marked enlargement at the base of the ear than in the white

varieties.

442

[Vol. 34

MISSOURI

18

O

16

14

12

c

c

10

(U

o

o

o

i

i

10-11

12-13

i

14-15

Kernel width in mm.

Text-fig. 5. Averages of tlic field and crib samples of yellow and white corn made in Guate-
ala. Each circle represents the average value for number of rows and for width of kernel of a
whole sample of corn from one field. Open circles denote white varieties, solid circles, yellow
varieties. It will be noted that the white varieties on the whole have narrower kernels and higher
row numbers. In five instances the same farmer was growing both a yellow corn and a white one.
A narrow ruled line connects the two varieties in each of these cases. It can be seen that in each
of these the white variety had a higher row number than the yellow variety with which it wat

associated.

In four of the examples in the Appendix (San Lucas, Santa Lucia, Dona Laura,

Nueva Cuartel)

J

yellow and white varieties grown by the same farmer, or grown in an adjacent
field. These five cases are diagrammed in fig. 5. It will be noted that in each
instance the yellow variety has a lower row number than the white variety being
grown with it. Though some of the yellows have as many rows as some of the
whites, the white varieties being grown in those localities had even more.

1947]

ANDERSON GUATEMALAN MAIZE

443

These facts suggest very strongly that two of the basic elements in the maize
of the Guatemalan highlands are a many-rowed white corn and an 8 -rowed yellow
corn. Though a great deal of mixing has gone on between them, and still con-
tinues, the white varieties, on the average, are larger-eared and with more num-
erous rows, the yellow varieties smaller-eared and fewer-rowed.

Even within the small area sampled by these field studies It is possible to dem-
onstrate Mangelsdorf and Cameron's center of variability in northwest Guatemala.
The 8 collections made in the Quezaltenango area are distinctly more variable
among themselves than those from In and around Antigua. This Is equally true
whether one considers single characters (Table I) or the general over-all impression

of the ear (Appendix).

Our data are not extensive enough cither to prove or disprove Mangelsdorf and
Cameron*s thesis that this Guatemalan center is connected with the presence of
Tripsantm and tcosinte In the same general area. "We have new Information on
only one point. The commingling of types Is not necessarily the resultant of
isolation Into mountain valleys as they suggest.

Three of our collections were made within sight of each other, on one mountain
slope, yet they are very different types of corn. Two of them were from fields
belonging to the same family. It may be significant that these 3 collections were
made on volcanic slopes above the Samala River, which has been since pre-Colum-
bian times one of the easiest approaches^ from the coast Into the highlands. Yet
It Is In this same area that varieties most like those of central Mexico are en-
countered. They have the sharply tapering ears, and the pointed kernels which
characterize the common varieties of the Mesa Central.

It may be, as Mangelsdorf and Cameron sviggest, that the conspicuous vari-
ability of the maize of northwest Guatemala is due to the actual commingling there
of Tripsacnm, Enchlaena, and Xca, On the other hand, from the available facts
one could argue quite as well that Guatemala is a center where diverse strains of
maize, which were dififcrentiated elsewhere, met and hybridized. Nor are these two
hypotheses mutually exclusive; it may be that the extreme variability of the maize
of Guatemala is in part due to very different varieties from South America and
from Mexico having met and hybridized at that point, and in part due to dis-
tlnctive qualities acquired there by introgresslon from Tripsacum. These are
questions which cannot yet be answered until we have reasonably complete popu-
lation samples of Guatemalan maize.

If teoslnte originated in the highlands of Guatemala as Mangelsdorf and Cam-
eron suggested, and has there introgressed most extensively with maize, we might
expect to find the maize of that area strongly tripsacoid in character. From the
published accounts of Corn X Teoslnte back-crossed with corn we might expect
to find a high percentage of varieties more or less long-eared and few-rowed, with
tapering, appressed ears. They would be borne on tough, narrow-leaved, slender-

^Sec McBrydc (1947), footnote page 10.

[Vot. 34

444

MISSOURI

stalked plants with strong root systems. Such varieties seem to be absent in the
highlands of Guatemala but they do characterize wide areas in western Mexico.
In the deep barrancas of western Mexico Tripsacuifi grows in variety and in pro-
fusion. It is there that the Tarahumare Indians are known (Lumholtz, 1902) to
interplant maize and teosinte to introduce drought resistance and flavor to the
former. It is there that waiz chapolote, a coffee-brown popcorn, one of the
most tripsacoid of maize varieties, is and was commonly grown as a staple.

The varieties of the (juatcmalan highlands often possess the low row numbers
to be expected from teosinte but they combine the character with wide seeds, a
large and differentiated butt, and a thick shank. The origin of these two latter
characters Is difficult to explain on any hypothesis. They reach greater extremes
in Guatemala than in any other area known to me and characterize most of the
maize of the highlands. This is particularly apparent when Guatemalan collections
are compared with those from Mexico. This can be demonstrated when ear
base outlines traced from photographs are compared for Guatemala, western
Mexico, and central Mexico. The slight increase in diameter demonstrated in such
photographs tends to be accompanied by changes in kernel rowing and even in
kernel shape, as can be seen in the Appendix to this paper.

It is difficult even to suggest how such enlarged ear bases might have
originated. Perhaps they came from crosses between cylindrical-eared varieties
and short-eared, globular types like those of the Andes. They are altogether lack-
ing in the extensive collections of preltistoric maize from western South America.
They are developed only to a minor extent in present-day Andean varieties. They
are not found over wide areas in Mexico, and in the few cases where they are highly
developed (Chiapas, Mountain Yellow) it is fairly obvious that they have spread
from Guatemala.

Their behavior in crosses indicates a multigenic basis, and it is difficult to see
how they could have originated out of pre-existing maize varieties even by strong
selection. There is no transparent reason why they might have originated by
hybridization with Tr'ipsacum or teosinte. It may be that they are in some way a
recombination of genes from South American and Central American maize and
that they are most strongly developed in Guatemala since it was there that mix-
tures of these two diverse stocks took place on a wider scale than elsewhere.

However these enlarged bases may have originated, their occurrence outside of
Guatemala is an almost certain indication of a greater connection with the maize
of that region. They are well developed In eastern North American flints (Brown
and Anderson, 1947) and are only one of several characters which those varieties
have in common with Guatemalan varieties.

More precise cytologlcal and histological tests are under way. The "Mountain
Yellow" varieties reported from Mexico also show obvious Guatemalan relation-

1947]

ANDERSON GUATEMALAN MAIZE 445

ships. In the American Southwest, where the prehistoric record has been most
completely analyzed, enlarged bases appear suddenly in Pueblo III (Carter and
Anderson, 1945) and have characterized the region ever since. Carter and
Anderson referred to these Pueblo III long-eared corns as ^'Eastern" because of
their close resemblance to Eastern American flints. There is as yet no evidence
concerning the Immediate center from whence they came. Whether it was from
Mexico, or from eastern North America, or from the Great Plains, it is probable
that ultimately, by some route, they trace back to Guatemala.

POPPING AND BREWING VARIETIES

These have been almost universally neglected by collectors. Mangelsdorf and
Cameron list one popcorn and no sweet corns. McBryde (1947) mentions it only
from Patziln. The Russian Expedition Hsts none of either type. Stadelman (1940)
lists one sweet corn, and none of these authorities make any mention of varieties
used for brewing. Though I was not able to study field samples of any of these
special kinds of maize, I did collect a few ears. It seems probable that such varieties
are fairly common in Guatemala. Since this statement will be received with almost
equal skepticism in the United States and in Guatemala, it may be well to discuss
the probability in detail. The explanation is of greater importance because the
situation is not limited to Guatemala but is quite general in many parts of Latin
America.

Certain distinctive types of maize continue to remain unknown because they
are grown by non Spanish-speaking Indians ("indigenas") . These distinctive
varieties arc cither used exclusively by the "indigcnas" or arc manufactured into
products which do not readily betray their origin when they appear upon the

market.

The gulf between the "indigena" and the "ladino"^ is much greater in Guate-
mala than one might suspect from casual observation, Tlic **indigcnas" make up
the bulk of the population. They are common everywhere, even in the capitol
city, and many of the "ladinos" show unmistakable evidence of considerable Indian
ancestry. It is not until one begins to investigate customs or products which are
peculiarly Indian that he finds out how completely these two kinds of people go
their separate ways, and how Httle one knows about the other. A woman from an
Indian town may come into market regularly for most of her adult life but beyond
the few words used in buying and selling she will have no knowledge of Spanish.
A ladino family may live in a largely Indian village for generations and deal in
agricultural produce, yet have little knowledge of those Indian food crops which
are not brought to market. My understanding of this phenomenon is the result of
efforts to obtain Guatemalan varieties of popcorn for study. It may be worth
while to describe these in some detail, since they illustrate the difficulties cn-

^For a complete discussion of the terms *'Iadino'* and "indigena" see McBryde (1947).

[Vol. 34

446 ANNALS OF THE MISSOURI BOTANICAL GARDEN

countered in getting a complete understanding of maize In Latin America, and
explain why certain very important types of food plants arc still almost unknown
to science.

I went to Guatemala knowing that popcorn had been collected there at least
once, and with a general picture of Its distribution and importance in Latin
America which convinced me that it must occur in Guatemala, at least In the back
country. From Erwin (1934) I knew that a popping Amaranth was also
being grown in southern Mexico. I had been successful, under the tutelage of
Dr. Isabel Kelly, in finding popcorn In various parts of Mexico from which it had
previously not been reported, and my command of Spanish was sufficient to discu
the matter with all the Spanish-speaking people T encountered.

During my first three weeks in Guatemala I got almost exclusively negative
results, though I now know that I was sometimes within sight of mountain fields
where popcorn was being grown. Various visiting American collectors, most of
whom had been on the outlook for strange varieties of maize, knew nothing about
it* The staff of the National School of Agriculture was similarly uninformed,
though they had an excellent and detailed understanding of the field corns of
Guatemala. American residents of Guatemala and Guatemalan farmers, merchants,
housewives, and landed proprietors gave equally negative replies with the excep-
tion of Mrs. Mildred Palmer, a specialist in Guatemalan textiles, who has direct
business connections with various Indian villages. She assured me that popcorn
balls were very commonly made and sold in various parts of Guatemala, though
she knew nothing about the varieties of maize from which they were made. The
ethnologists of the Institute de las Tndigenlstas knew Utcle about the matter but
were most cooperative In gathering further Information. They were soon able to
supply me with a single car of popcorn from the Quezaltenango region and the
advice to try making collections in the town of Patzun.

In Patzun I got in touch with the local corn merchant, a most intelligent man,
who had a wide and accurate knowledge of the field corns of the region and a
lively interest in varieties of commercial importance. He supplied the Information
that popcorn balls were sold on market days In Patzun and that they were made
exclusively by the Indians from special kinds of maize. He knew little, however,
about these varieties. He thought there were two different kinds but was not
certain. He thought one had pointed kernels and the other not, but could give no
further information. I then hurried over to the town market and found popcorn
balls being sold by a number of Indian women, none of whom could speak more
than a few words of Spanish. Through an interpreter I attempted to buy ears of
popcorn or at least seeds of that variety. The women attempted to pass off seeds
of ordinary field corn as the source of their popcorn balls. As can be seen from
the photographs in fig. 1, the kernels in the popcorn balls are fully exploded and
could not have come from any such variety of corn. In the short time before my
bus departed, the best I could do was to purchase a small amount of a rather
mongrelizcd popcorn and to arrange to have more authentic specimens purchased

J947]

ANDERSON GUATEMALAN MAIZE 447

and mailed to me. They proved to be a most interesting variety with phenomen-
ally large kernels for a popcorn, but they pop uniformly well. The kernels are
wide and quite thick, with no indications of a point, and have a translucent white
endosperm. The ears are slender, with 10-12 rows of kernels. They are quite un-
like any native or commercial varieties of popcorn known to me.

The popcorn balls from this market led to further information. They were
immediately recognized by every one In the servant class (i. e.j people of predomi-
nantly Indian ancestry) to whom I showed them in the town of Antigua. They
are known as ajborotos and are very commonly^ brought into town during Lent

by Indians from Patcicia, a town near Patzun. (See also McBrydc, page 10), The
landed proprietors of Antigua to whom I showed them had either never noticed
them before or did not know that they were made of maize. One well-to-do
'*finquero" who makes journeys to Chicago and San Francisco every few years
told me that these popcorn balls were made from a plant closely related to "nihau"
(AmarautJms)\ I have not yet been able to collect any considerable amount of
popcorn from Guatemala but on two occasions (above Zunil and above Solola) I
have seen fields of an extremely small-caned maize growing in good land next to
fields of large maize. It must have been either a popcorn or a special variety used
in brewing.

The difficulties encountered in collecting popcorn are magnified in getting ex-
amples of varieties used in brewing or information about their use. More than one
kind of home-made alcoholic beverage is made from maize in Guatemala, but such
manufacture is illegal and one has to have the confidence of his informant if much
is to be learned. In the short time at my disposal I was able to determine that
varieties with a blue alcurone were preferred for this purpose, since they were
sweeter and smoother. I was also assured by a most intelligent "ladino" woman,
who lived in a town largely composed of Indians, that they had certain highly
prized varieties used in brewing their ancient types of beverages. It it highly
probable that ancient varieties with sugary endosperm are still in existence in
Guatemala and that they are probably used in brewing there, as in South America.
Stadelman lists a single ear, and his description reminds one of the luatz dulce
discovered by Kelly in western Mexico (1943).

For the above reasons it is evident that we still know little or nothing about
some of the most interesting varieties of maize in Guatemala, and that it will
require patience and very special skills to obtain a full understanding of these types.
The effort is worth while, not only because such Information will illuminate the
history of maize, but because, singularly enough, it Is among such specialty corns
that useful genes for modern corn-breeding are quite likely to be found (see
below) .

^ThougK similar confections made from a popping Amaranth and a popping sorghum are even

more common.

[Vol. 34

448 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Sdpor.

Unfortunately none of the fields included in this survey was planted to the
higlily developed ''Salpor" or Flour Corn. This is a large-kernclled variety of
white flour corn which is commonly grown in parts of the Guatemalan highlands.
Judging from samples displayed in several markets, a good deal of the maize sold
under that name in Guatemala is extensively contaminated with other kinds of
corn. It very closely resembles the ^^Cacahnazintle^^ flour corns of Mexico which
were probably derived from it, and is quite similar to the highly developed flour
corns of Andean South America from which it may in turn have been derived.

Multiplication.

Cutler has recently described under this name a bifurcation of the spikclet
pedicel which increases the kernel number in South American maize. In its lowest
grades it is responsible for the extra kernels pushed in between the regular rows of
8- and 10-rowed varieties. With a higher degree of expression it turns 8-rowed
varieties Into 4 quadrants, within each of which the rowing is irregular and ob-
scure. In its extreme manifestation it produces an car in which the regularity of
the rowing, as in the "Country Gentleman" sweet corn, is no longer apparent.
Though these ears look superficially similar to "Country Gentleman," they owe
their increase In kernel number not to the development of the aborted floret (as in
Country Gentleman"), but to a doubling of the entire spikclet.

Multiplication is a common phenomenon in Guatemala and from cursory ob-
servation is more frequent and more extreme at higher altitudes. Actual percent-
age frequencies were obtained for several of the collections and arc presented in
the Appendix (nos. 6, 9, etc.).

Practical Considerations,

A number of different agencies in Guatemala are already concerned about the
yields of Guatemalan varieties, and breeding programs are already under way to

ft

improve them* American agronomists, or Guatemalan agronomists trained in the
United States, are prone to begin any improvement program along the lines which
have proved so conspicuously successful in the United States. In my opinion, this
IS ill-advised. In the first place, the variation pattern of 2ca Mays is wholly dif-
ferent in Guatemala from what it Is In the United States. In the second place, the
conditions under which it is now grown and under which it is likely to be grown
in the near future are dIflFerent there and here. Hybrid corn owes its superiority
In the American agricultural picture as much to Its uniformity as to Its superior
yield. Except on a few large plantations there Is Httlc prospect of growing
Guatemalan corn with power machinery by mass-production methods. Under
Guatemalan conditions, therefore, the uniformity of hybrid corn would be of no
particular advantage. Tlie other advantage of hybrid corn, extreme heterozygosity,
might well be achieved in Guatemala by much simpler methods. Tlie fields in the
Quezaltenango area suggest that Mexican varieties with many-rowed, more or less

1947]

ANDERSON GUATEMALAN MAIZE 449

pointed kernels combine well with Guatemalan varieties. It would be a compara-
tively simple matter to select open-pollinated varieties of white Guatemalan maize
and of white Mexican maize which combine well with each other. They could
be grown and improved by mass-selection methods. If the maximum improvement
was worth the time and expense they could then be carefully selected every five
or six generations for their combining ability with one another, using a modifica-
tion of the plan originally selected by Hull (1945) of Florida. These two elite
white varieties, each increased as an open-pollinated crop, could then be inter-
planted and detasseled as is hybrid corn in the United States, producing first-
generation hybrid seed for sale and distribution.

The probable usefulness of Gnafemalan maize in the United States.

Since maize is of even greater importance in the United States than in Guate-
mala, the extreme over-all variability of Guatemalan maize is of great potential
importance to our agriculture. This does not mean, however, that Guatemalan
maize varieties, as such, can immediately be used in the production of better corn
for the United States. As agronomy advances and it becomes increasingly practi-
cal to breed for particular characters in maize, Guatemalan varieties should prove
increasingly useful. Resistance to particular diseases, high percentages of unusual
amino acids, kernel texture, insect resistance, sugar content, etc. are characters
which might well be expected in one Guatemalan variety or another. Once located,
it would be a comparatively simple matter to transfer any particular one to a
commercial inbred line. Once tliey are so incorporated they may be used effective-
ly in the production of commercial hybrids for the United States.

It should be pointed out that in any such discriminating corn-breeding pro-
gram as that just outlined all Guatemalan maize is of potential importance. As a
source of disease-resistant genes or of increased quantities of useful amino acids,
some small-eared, small-kernelled variety from the mountains may be quite as
useful to United States agriculture as the large-eared sorts of spectacular produc-
tivity. It is even possible that some of the out-of-the-way varieties may be more
useful than the ordinary run of Guatemalan field corns. To be specific, the pop-
corns and brewing corns are extreme types morphologically; the chances are good
that they may also be extreme in their chemical composition and their disease and
insect resistance. For purely practical reasons, therefore, Guatemalan corns should
be systematically and comprehensively surveyed without reference to their immedi-
ate usefulness as field corns. We need, in the first place, a general survey of what
kinds of maize there are in the country, what their general morphological char-

L

acters are (row number, kernel texture, etc.). Then we shall be ready to make
a systematic survey of their chemical compositions, disease resistance, etc., and
will know where to turn for the characters we need in United States corn breeding.

[Vol. 34

450 ANNALS OF THE MISSOURI BOTANICAL GARDEN

SufNffiary,

L Field sampling of maize is contrasted with sampling at markets and fairs.
The latter is shown to give an erroneous and incomplete picture of Latin American

maize.

2. The advantages of personal contact with the families which grew the maize
sample are described and discussed.

3. Experimental d