Full text of "Contributions from the Laboratory of Plant Genetics"

See other formats

iiMtaMtfjjS'

f - <3 ■

I
9

HARVARD UNIVERSITY
BUSSEY INSTITUTION OF APPLIED BIOLOGY

CONTRIBUTIONS

FROM THE

Laboratory of Genetics

Vol. II.

BOSTON
1913-1919

SCTENCE

CONTENTS

Xenia and the Endosperm of Angiosperms. E. M. East.

A Genetic Analysis -of the Changes Produced by Selection in Ex-
periments with Tobacco. E. M. East and H. K. Hayes.

Studies of Teratological Phenomena in Their Relation to Evolu-
tion and the Problems of Heredity. Orland E. White. Pts. I-II.

Observations on the Relation between Flower Color and Insects.
E. M. East and R. W. Glaser.

Johannsen's Elemente der exakten Erblichkeitslehre. Zweite
Auflge. E. M. East.

An Interpretation of Self -sterility. E. M. East.

The Phenomenon of Self -sterility. E. M. East.

E. S. Carman. E. M. East.

An Interpretation of Sterility in Certain Plants. E. M. East.
The Chromosome View of Heredity and its Meaning to Plant

Breeders. E. M. East.
Further Experiments on Inheritance in Maize. H. K. Hayes and

E. M. East.

Studies on Size Inheritance in Nicotiana. E. M. East.

Significant Accuracy in Recording Genetic Data. E. M. East.

Inheritance in Crosses Between Nicotiana langsdorffii and Nico-
tiana alata. E. M. East.

Hidden Feeblemindedness. E. M. East.

The Explanation of Self -sterility. E. M.* East.

The Bearing of Some General Biological Facts on Bud- Variation.
E. M. East.

Dominance of Linked Factors as a Means of Accounting for Hete-
rosis. Donald F. Jones.

On Reversible Transformability of Allelomorphs. H. Terao.

Maternal Inheritance in the Soy Bean. H. Terao.

The Inheritance of Doubleness in Chelidonium ma jus Linn. Karl
Sax.

The Behavior of the Chromosomes in Fertilization. Karl Sax.
Studies on Self -sterility. I. The Behavior of Self -sterile Plants.

E. M. East and J. B. Park.
Studies on Self -sterility. II. Pollen-tube Growth. E. M. East and

J. B. Park.

Studies on Self -sterility. III. The Relation between Self-fertile
and Self-sterile Plants. E. M. East.

Studies on Self-sterility. IV. Selective Fertilization. E. M. East.
Studies on Self-sterility. V. A Family of Self-sterile Plants Wholly

Cross-sterile Inter se. E. M. East.
The Role of Reproduction in Evolution. E. M. East.
Intercrosses Between Self -sterile Plants. E. M. East.
The Effects of Inbreeding and Crossbreeding Upon Development.

D. F. Jones.

Heredity of Quantitative Characters in Wheat. George F. Free-
man.

XENIA AND THE ENDOSPERM OF ANGIOSPERMS

E. M. EAST

Reprinted for private circulation from
The Botanical Gazette, Vol. LVI, No. 3, September 19 13

XENIA AND THE ENDOSPERM OF ANGIOSPERMS

E. M. East

As is well known, the term "xenia" was proposed by Focke to
describe any effect of pollen of another race upon the tissue of a
seed plant apart from that initiating the formation of an embryo.
As it has been exceedingly questionable whether any such effect
beyond a chemical irritation ever occurs, the word has come to be
applied to the appearance of the Fz hybrid endosperm produced by
the fusion of the second male nucleus with the so-called endosperm
nucleus of the embryo sac, when its characters are different from
those exhibited by the mother plant after self-fertilization.

Since the fact of this fusion was proved cytologically by Guig-
nard (10) and Nawaschin (ii), data on this type of xenia have
interested botanists because of the differences of opinion existing
concerning the phylogenetic significance of the angiosperm endo-
sperm.

The most detailed observations on xenia have been those on
maize, since numerous maize varieties exist with differences in
endosperm characters. The behavior of the following factors in
heredity is known from the researches of DeVries (5), Correns
(2), Webber (12), East and Hayes (7), and Emerson (9). In
addition, East (6) has found good indications of at least three
additional factors that modify the expression of the red and the
purple aleurone colors.

Factor

Action

Causing full development of starch grains

Causing yellow color throughout endosperm

Similar to Yt but not allelomorphic to it

Basic color factor necessary for color in aleurone cells

Present with C gives red color in aleurone cells

Present with R and C gives purple color in aleurone cells

Inhibits aleurone color when present with RC or PRC

Observations on crosses wherein these characters have been
concerned have made it possible to formulate the following law
regarding xenia :

217] [Botanical Gazette, vol. 56

218

BOTANICAL GAZETTE

[SEPTEMBER

When two races differ in a single visible endosperm character in which
dominance is complete, xenia occurs only when the dominant parent is the
male; when they differ in a single visible endosperm character in which domi-
nance is incomplete or in two characters both of which are necessary for the
development of the visible difference, xenia occurs when either is the male.

It is evident that such- a statement can be true only if the two
male nuclei always carry the same hereditary factors and if a male
nucleus always enters into the formation of the endosperm. The
first requirement has been satisfied in every experiment thus far
recorded; the second requirement will now be considered.

In particular cases where xenia has followed the crossing of
races differing in endosperm color, aleurone color, or ability to
mature starch grains, the seeds are not uniform in appearance.
One may be half starchy and half wrinkled; another may be half
yellow and half colorless ; still another may have half of the aleurone
cells red or purple and the other half colorless. Examples of this
kind are rarely found, although it is a common thing to find seeds
with a mottled appearance affecting only the aleurone colors.

Correns and Webber suggested independently that in these
cases the male nucleus may fail to unite with the fusion nucleus and
each divide independently, forming either the half-and-half seeds
or those which are mottled. Webber also suggested, as an alter-
nate hypothesis, the fusion of the male nucleus with one of the polar
nuclei, the other polar nucleus remaining independent and dividing.

East and Hayes have shown that Correns and Webber were
dealing here with two phenomena. The seeds that are mottled
become so only from the development or non-development of color
in the aleurone cells. They merely exhibit irregularity of Mende-
lian dominance, since in some crosses practically all seeds hetero-
zygous for one of the factors producing aleurone color are mottled,
although homozygotes are fully colored. Furthermore, the
mottling does not extend to the color or other character of the
deeper endosperm tissue in case the parental varieties had such differ-
ences, which necessarily would be the condition if the endosperm
had been formed according to either of Webber's independent
development hypotheses. This criticism has also been made inde-
pendently by Emerson (8).

EAST—XENIA AND ENDOSPERM

219

The other cases, where the endosperm is divided more or less
equally into two types, remain to be explained. The hypothesis of
independent development of the male nucleus seems improbable if
one may judge from relevant cytological data on both animals and
plants. The second hypothesis is very plausible. As a third pos-
sibility, East and Hayes have suggested ordinary " endosperm
f ertilization " with subsequent vegetative segregation similar to that
occurring in bud sports. This could be proved, according to them,
if among the Fx seeds of a cross between parents differing in two
allelomorphic pairs, individuals should be found in which the
parental characters were combined differently. No such cases
have been recorded.

The difficulty of deciding between the first and the second
hypothesis of Webber lies in the fact that individuals of this kind
are very rare, and when they have been found the investigator has
not been able to say which particular endosperm character was
carried by the male cell and which by the female cell. This was
because they have occurred in selfed hybrids where both pollen
and egg cells were segregating various Mendelian factors. In the
experiments now to be described, this difficulty has been overcome.

The red color in the aleurone cells of maize is due to the inter-
action of two factors that may be represented by the letters C and
R; this color may be changed to purple by the presence of a third
factor P. Red is RC and purple is PRC, therefore, although it
must be understood both that other factors which have never been
lost in any variety may enter into the combination, and that other
factors which have been lost in certain varieties may affect the
development of color.

Six homozygous white varieties may exist with the following
zygotic formulae: PPRRcc, PPrrcc, PPrrCC, ppRRcc, pprrCC, and
pprrcc. Any cross between these varieties of such a nature that
R and C or P, R, and C are brought together results in the red or
the purple color respectively.

Among the selfed maize ears that had been produced in the
course of the writer's experiments were a number giving red
wrinkled and white wrinkled seeds in the ratio of 3:1. These
white seeds must have either the formula ppRRcc or pprrCC.

220

BOTANICAL GAZETTE

[SEPTEMBER

White seeds from three such ears were planted in isolated plots and
used as male parents on the flowers of plants arising from white
seeds found on selfed ears of 13 other families. A number of these
families had the proper formulae to produce color, and about
60,000 red or purple seeds were produced. There were all-purple
ears and all-red ears in several families. Other combinations gave
purple and white seeds or red and white seeds in the ratio of 1:1.
How this came about is clear if one assumes either of the formulae
given above for the male parent. Suppose the male parent had the
formula ppRRcc: a family with the formula pprrCC gives all-red
ears, while one with the formula pprrCc gives ears with red and
white seeds in the 1:1 ratio; a family with the formula PPrrCC
gives all-purple ears, while one with the formula PprrCC or PPrrCc
gives ears with purple and white seeds in the 1 : 1 ratio.

Considering first only the all-purple and the all-red ears, one
must conclude that the fusion of the " endosperm nucleus" and the
second male nucleus always occurs. If it did not occur, white seeds
would result, because a factor from each parent is essential for the
production of color.

Among these 60,000 seeds, 6 were found that showed the half-
and-half condition; that is, color had developed on one side and
not on the other. They were typical illustrations of the phenome-
non which Webber's two hypotheses were devised to explain.
They occurred in only 0.01 per cent of the fertilizations, but in
spite of their rarity they show that Webber's first hypothesis,
assuming independent development of the male nucleus, is unten-
able, since independent development of the paternal and the
maternal nuclei could produce no color. No decision can be made
between Webber's second hypothesis — fusion of the male nucleus
with one polar nucleus and independent development of the other — ■
and the hypothesis of vegetative segregation after partial develop-
ment. The bilateral symmetry of the halves of the seeds with and
without color favors Webber's idea; at the same time, it must be
pointed out that the frequency of the occurrence is not too great
to compare favorably with the frequency with which "bud sports"
originate. Though it would afford some satisfaction, a precise
explanation of these rare aberrations is not a necessary requisite

EAST—XENIA AND ENDOSPERM

221

to several conclusions indicated by the experiments. It is evident
that in the varieties of maize used, a paternal and a maternal
nucleus carrying the same hereditary factors as are borne by the
true gametes — in the case of the 7 factors investigated — always
fuse in the formation of the endosperm. For this reason geneticists
investigating maize have been correct in treating the endosperm as
if it were an embryo. The endosperm characters have behaved
exactly like plant characters. Two white varieties of sweet peas
may carry factors both of which are necessary for the production
of color. When they are crossed, color develops. Color develops
in maize in a quite similar manner when the two complementary
factors are carried by the " endosperm nucleus" and the second
male nucleus. Nevertheless, one should keep in mind that the
problem is complicated. Collins (i) found a white ear of maize
in a yellow variety that behaved as if its seeds were crossed with
the yellow. He interpreted the phenomenon as a mutation showing
reversal of dominance, although the data on succeeding generations
corroborated those obtained by previous investigators in which
yellow was partially or completely dominant. It is not unlikely,
however, that Collins merely happened upon a plant from white
seed in which the male nucleus did not enter into the formation of
the endosperm, although other interpretations are possible. This
may seem like an odd statement after having shown that the two
nuclei always fuse, but it is made advisedly. In most varieties of
maize the two nuclei do appear always to fuse, but Hayes is now
working out the details in a cross in which a Mexican starchy corn
is one of the parents where the nuclei appear never to fuse. In
other words, it seems that there may be varieties of maize in which
endosperm formation is the opposite of that just described, and
within each category no change to the other has been found. But
may not such a change occur ?

Whether or not the last suggestion ever proves to be true, it
seems to me that from the data now collected one is entitled to
discuss angiosperm endosperm formation from the viewpoint of
experimental genetics.

The endosperm of the gymnosperms is essentially vegetative
tissue of the female gametophyte. It results from continuous cell

222

BOTANICAL GAZETTE

[SEPTEMBER

formation originating with the germination of the megaspore,
although fertilization occurs during the process. From the time
of Hofmeister the morphological character of the endosperm of
angiosperms was considered to be the same as that of the gymno-
sperms until the double fertilization was discovered. This fact
gave rise to the idea that the angiosperm endosperm might be
a sporophytic rather than a gametophytic structure, its nature
being that of a monstrous embryo, or possibly that it is a composite
tissue neither gametophytic nor sporophytic.

Most botanists, however, have held with Strasburger to the
original idea that the endosperm is gametophytic. Strasburger
concluded that the second fusion is not a true act of fertilization
uniting the parental qualities and forming an embryo, but a vege-
tative fusion acting merely as a stimulus to growth. Miss Sargent,
however, believes that it is a degenerate embryo, the monstrous
character being caused by the interference of the antipodal nucleus
having a vegetative character and an indefinite and usually redun-
dant number of chromosomes in the act.

The difficulty in the situation appears to be the obscurity of the
phylogenetic history of the fusion of the two nuclei in the embryo
sac and the subsequent fusion with the second male nucleus. The
problem is further complicated by the irregularity of endosperm
formation in various species. Although triple fusion appears to
occur in the majority of angiosperms, the following important
general variations have been noted. In addition to these general
variations many minor deviations have been found (Coulter and
Chamberlain 4). (1) Vegetative endosperm formation may take
place in a similar manner to that occurring in gymnosperms. This
may occur without fertilization, or before or after fertilization.
Usually the endosperm tissue is formed from the descendants of
the antipodal cells, but the chalazal nucleus may degenerate and
the endosperm be formed from the micropylar polar nucleus.

(2) The polar nuclei may not fuse, but divide independently.

(3) Fusion may include many cells.

Furthermore, endosperm formation may be initiated by free
nuclear division, or the sac may be divided into two parts by a cell
wall after the first division. Even when the latter phenomenon

1913] EAST—XENIA AND ENDOSPERM 223

occurs, endosperm tissue may be formed in both chambers, although
usually division proceeds only in the micropylar chamber.

These general cytological data being given, how do the facts
from pedigree cultures bear upon the problem ?

Just how much weight should be given to data from only one
species when discussing the morphological significance of the endo-
sperm is questionable. But in maize it is evident that Stras-
burger's distinction between vegetative and generative fertiliza-
tion will not hold. Cytological work on other species does not bear
out Miss Sargent's conception, since endosperms form quite
regularly without the interference of the antipodal vegetative ( ?)
nucleus. If the perfectly regular manner in which the above-
mentioned endosperm characters of maize are transmitted is con-
sidered apart from other facts, there appears to be no escape from
the conclusion that the endosperm is sporophytic in character.
But there is another way of looking at the matter that makes the
view of Coulter seem more probable.

Coulter (3) has concluded that conditions in the embryo sac
favor fusions of any number or kind of free nuclei — an indefinite
process without a necessary phylogeny that results in a growth
which is practically gametophytic. It is not dependent upon a
male nucleus, a polar nucleus, or even a reduction division.

The experimental evidence accords perfectly with this view.
The superficial endosperm characters are indeed transmitted regu-
larly when a male nucleus takes part in the fusion, but there is no
reason for believing that the remaining maternal nuclei carry all
the characters borne by the egg because these characters are the
same in the nuclei concerned. The egg must usually have an
organization somewhat different from that of the other maternal
nuclei; although it is recognized that other nuclei sometimes func-
tion as eggs. It is likely that a differentiation has ensued which
makes a particular nucleus an egg, and that it is not wholly a matter
of position. The general belief in the vegetative character of the
antipodal cells of the embryo sac is an admission that they have
not received all the properties retained by other four cells. It is
not very heretical, therefore, to assume that the cell that becomes
the egg is different from its associates. Botanists hesitate to assume

224

BOTANICAL GAZETTE

[SEPTEMBER

the differentiation during ontogeny admitted by zoologists. They
desire to believe that most plant cells can reproduce the whole
plant. But this is a belief and not a fact, and until it becomes a
fact it is well to recognize this plausible alternative in considering
matters such as periclinal and sectorial chimeras as well as endo-
sperms.

Harvard University

LITERATURE CITED

1. Collins, G. N., Heredity of a maize variation. U.S. Dept. Agr. Bur.
Plant Ind. Bull. 272. pp. 7-23. 1913.

2. Correns, C, Bastarde zwischen Maisrassen mit besonderer Beriicksichti-
gung der Xenien. Bibliotheca Botanica 53:1-161. 1901.

3. Coulter, J. M., The endosperm of angiosperms. Bot. Gaz. 51:380-385.

IQII.

4. Coulter, J. M., and Chamberlain, C. J., Morphology of angiosperms.
pp. x+348. New York. 1909.

5. DeVries, H., Sur la fecondation hybride de l'albumen. Compt. Rend.
Acad. Sci. 129:973-975. 1899.

6. East, E. M., The Mendelian notation as a description of physiological
facts. Amer. Nat. 46:633-655. 191 2.

7. East, E. M., and Hayes, H. K., Inheritance in maize. Conn. Agr. Exp.
Sta. Bull. 167. pp. 1-141. 1911.

8. Emerson, R. A., Inheritance of color in the seeds of the common bean,
Phaseolus vulgaris. Ann. Rep. Neb. Agr. Exp. Sta. 22:67-101. 1909.

9. , Aleurone colors in F2 in a cross between non-colored varieties of

maize. Amer. Nat. 46:612-615. 1912.

10. Gutgnard, L., Sur les antherozoides et la double copulation sexuelle chez
les vegetaux angiospermes. Rev. Gen. Botanique 11:129-135. 1899.

11. Nawascrtn, S., Resultate einer Revision des Befruchtungsvorgangs bei
Lilium Martagon und FritUlaria tenella. Bull. Acad. Imp. Sci. St. Peters-
bourg 9: no. 4. 1899.

12. Webber, H. J., Xenia, or the immediate effect of pollen in maize. U.S.
Dept. Agr., Div. Veg. Phys. and Path. 22:1-44. 1900.

A GENETIC ANALYSIS OF THE CHANGES PRO
DUCED BY SELECTION IN EXPERIMENTS
WITH TOBACCO

PROFESSOR E. M. EAST and H. K. HAYES

NEW YORK
1914

[Reprinted without change of paging, from the American Naturalist, 19 14.

[Kepnnted from The American Naturalist, Vol. XLV1II., Jan., 1914. |

A GENETIC ANALYSIS OF THE CHANGES PRO-
DUCED BY SELECTION IN EXPERIMENTS
WITH TOBACCO1

PROFESSOR E. M. EAST and H. K. HAYES
Bussey Institution of Harvaud University

The Pkoblem

In 1903 Johannsen announced that continued selection
of the extreme values of certain quantitative characters
in successive self-fertilized generations of a number of
strains of beans had produced no changes in the mean
values of the characters. He concluded that these par-
ticular strains were homozygous for the gametic factors
whose interaction resulted in the characters investigated,
that these homozygous characters may be properly de-
scribed by one or more gametic factors nonvariable in
transmissible qualities and properties, and that the varia-
tions observed in the characters of any single fraternity
were due entirely to the action of environmental condi-
tions during ontogeny and were not inherited. Funda-
mentally, these conclusions were a recognition of the gen-
eral value of Mendelian description for all forms of in-
heritance through sexual reproduction, combined with an

1 These investigations were conducted with funds furnished by the Con-
necticut Agricultural Experiment Station from their Adams' appropria-
tions, by the Bureau of Plant Industry of the United States Department of
Agriculture, and by the Bussey Institution of Harvard University, and the
writers desire to take this opportunity of expressing their sincere appre-
ciation of this hearty cooperation which made the work possible.

THE AMERICAN NATURALIST [Vol. XLVIII

admission of disbelief in the inheritance of ordinary
adaptive changes. The latter conception was Weismann-
ian in that all inherited variations were held to be changes
in the germ cells. It was not necessary to suppose it im-
possible for the environment to produce such changes and
therefore to have been of no value during the course of
evolution, but merely to suppose that during the compara-
tively short period of experimental investigations no gam-
etic variations have occurred traceable to such a cause.
For his first conclusion to be justified, it was assumed that
the changes which every biologist knows do follow the
continuous selection of extremes under certain conditions
are to be interpreted entirely by the segregation and re-
combination of hypothetical gametic factors which are
constant in their reactions under identical conditions.

Numerous investigators working on "pure lines" with
different material corroborated Johannsen's conclusions,
and, as it was seen to be possible to interpret in the same
manner changes made by selection in experiments where
self-fertilized lines were not used, such as those of the
Vilmorins and others on sugar beets and those of the
Illinois Agricultural Experiment Station on maize, many
biologists accepted them and considered them a great ad-
vance over former conceptions of the mechanism of
heredity. On the other hand, there were those who main-
tained a skeptical attitude, the chief criticism directed
against the conception being that all progress due to
selection must have a limit, which in many of these ex-
periments had already been reached, and that even if re-
sults were being obtained action might be too slow to be
detected.

The Material

These criticisms were reasonable when applied to cer-
tain specific cases, and in 1908 the experiments reported
in this paper were designed with the hope of testing their
validity, using the species ordinarily grown for commer-
cial tobacco, Xicotiana tabacum, as the material. This
plant satisfies the conditions which are requisite for

No. 565]

CHANGES PRODUCED BY SELECTION

material used in pure line studies. It has characters that
can be estimated readily and accurately and which are
affected only slightly by external conditions. It is easily
grown, is naturally self-fertilized, reproduces prolifically,
and is known in many markedly different varieties. In
fact, it is an ideal subject for work of this kind.

The investigations were not patterned after the stand-
ard type set by Johannsen wherein the constancy of suc-
cessive generations of pure lines grown from selected
extremes were tested, since even if it were possible to
gather a quantity of data at all comparable to that col-
lected by Johannsen ( :09) and Jennings ( :08) in their
brilliant investigations, the criticisms mentioned above
might still be made. The plan chosen was that of cross-
ing two varieties of tobacco which differed in a character
complex easily and precisely determined, and of selecting
extremes from a number of families of the F2 generation.
If Johannsen 's views be incorrect, such continued selec-
tion should affect each family in the same degree. If his
conclusions be justified, selection should reach an end-
point in different generations in different families, and
there should be no relation between the number of genera-
tions required to reach this end-point and the progress
that is possible.

There should be no need of a historical summary of the
previous investigations that have been interpreted as cor-
roborating or refuting Johannsen 's conclusions. Such
summaries have been made in other papers. It should be
mentioned, however, that the classical researches of Pearl
( :11) on the inheritance of fecundity in the domestic
fowl have been so planned and executed that certain of
the criticisms directed against Johannsen mentioned above
are not justified, yet Pearl finds himself thoroughly in
accord with the Danish physiologist's position.

Several hundred varieties of Nicotiana tabacum exist
which differ from each other by definite botanical char-
acters, yet only two general characters suitable for our
purpose were found. We desired to confine our observa-
tions to quantitative characters that were influenced but

THE AMERICAN NATURALIST [Vol. XL VIII

little by environment, and number of leaves and size of
corolla were the only ones that satisfied this requirement.
Such character differences as height of plant and size of
leaf, while undoubtedly transmissible, are influenced so
strongly in their development by nutrition that work with
them is exceedingly difficult. For example, if a certain
variety of Nicotiana tabacum is grown under the best of
field conditions, the longest leaves are about 28 inches and
the total height about 6 feet, but a portion of the same
seed fraternity may be grown to maturity in 4-inch pots
without reaching a height of over 16 inches or having
leaves longer than 4 inches. On the other hand, several
experiments conducted in the same manner have shown
no difference between the frequency curves of variation
in number of leaves or of size of corolla, whether starved
in small pots or grown under optimum conditions. The
character complex number of leaves was chosen for this
investigation rather than the size of corolla because vari-
eties that differ greatly in number of leaves are common.

TABLE I

Frequency Distribution of Number of Leaves per Plant when
Starved in Small Pots

(Compare "with frequency distribution under normal field conditions at
Forest Hills, Massachusetts, in Tables VII and XI)

No. of Leaves per Plant

^lant JNo.

23 1 24

(6-1)

(6-D-l

(6-2)

(6-2)-2

(56-1)

(56-2)

...

2
1

3
6

10
8
1

15
15
0
1
12
6

8
16
8
0
6
10

7
12
7
1
7
13

5
14
0

2
8

8
12

3
17

3
16

8
4

Previous Work of the "Havana" X "Sumatra" Cross

Several crosses have been made between varieties of
tobacco that had a mean difference of seven or eight
leaves, bnt the majority of the data reported here were
collected from the descendants of a cross made by A. D.
Shamel between the types known in Connecticut as
" Havana" and " Sumatra." The "Havana" parent was

No. 565] CHANGES PRODUCED BY SELECTION

from a variety that had been grown for a number of years
at Granby, Connecticut. It averages about 20 leaves per
plant although ranging from 16 to 25 leaves. The aver-
age height is about 1.4 m. and the average leaf area about
7 sq. dm. The " Sumatra' ' parent was a type specimen
of a variety that had been introduced into Connecticut to
be grown under cloth shade. It averages between 26 and
27 leaves per plant with a range of from 21 to 32 leaves.
The average height is nearly 2.0 m., but the average leaf
area is only about 3 sq. dm.

According to Shamel, the first hybrid generation of
this cross developed somewhat more vigorously than the
parent types and was uniform in its habit of growth.
The second generation, he thought, was hardly more vari-
able than the first. Several F3 families, the progeny of
inbred F2 individuals, were grown in 1906 and proved to
be a variable lot. One of these plants produced 26 small,
round-pointed leaves with short internodes between them.
This plant was thought by Mr. E. Halladay, upon whose
farm the experiment was conducted, and Mr. J. B. Stewart,
of the U. S. Department of Agriculture, to be worth sav-
ing from its promise of producing a desirable commercial
type.

In 1907 the Department of Agriculture made an agree-
ment with Mr. Halladay to grow two acres of tobacco for
experimental purposes, and on his own initiative Mr.
Halladay grew a number of plants from inbred seed of
the one that bore 26 leaves. This selection, numbered 2
h-29 in accordance with the department nomenclature,
was comparatively uniform in appearance and several
plants were selfed. In Mr. Halladay 's absence, how-
ever, all of the plants were ' 1 topped,' ' except one that
happened to be rather late. This plant was selfed. It
had 26 medium-sized, round leaves and grew to about the
same height as the Connecticut Havana.

In view of Mr. Halladay 's high opinion of the type, the
seed of this plant and the remaining seed of its parent
were planted in 1908. The plants of this generation pre-
sented a uniform appearance and promised a high grade

THE AMEBIC AN NATURALIST [Vol. XL VIII

of wrapper tobacco, but the crop when cured lacked uni-
formity. Some leaves of exceptionally high quality were
produced, but the crop in general lacked that characteris-
tic known as " grain' ' and had too large a proportion of
heavy leaves — the so-called 4 'tops."

From this 1908 generation 100 seed plants were selfed,
their leaves harvested, cured and fermented separately,
and data on quality recorded. The type was also grown
commercially on a large scale. The commercial results,
however, have been reported in another paper. We are to
consider only the results of the selection experiment that
began in 1908, through the cooperation between the U. S.
Department of Agriculture and the Connecticut Agricul-
tural Experiment Station, a joining of forces that in 1909
included the Bussey Institution of Harvard University.
Shamel ( :07) considered the strain produced by this cross
to be the result of a mutation. From a study of the
data from the previous work on the cross it seemed to the
writers that a different interpretation of the results might
be made. While it was not impossible that the many-
leaved type that had been isolated was the result of a
mutation, it appeared much more probable that it had
arisen through a recombination of Mendelian factors.
The type had the habit of growth and size of leaf of the
pure ' 'Havana' ' variety and the number of leaves of the
' ' Sumatra' ' variety, a combination that might reason-
ably be expected to be the result of the Mendelian law.

Results on the Reciprocal Cross, "Sumatra"
X "Havana"
To test the hypothesis that the new tobacco was the
result of such recombination and could be reproduced
whenever desired, the reciprocal of the original cross was
made in 1910. The female parent, "Sumatra," was the
direct descendant of a sister of the plant used as the
male parent of the original cross by Shamel in 1903
through seven generations of selfed plants. The male
parent, "Havana," was from the commercial field of the
Windsor Tobacco Growers' Corporation at Bloomfield,

No. 565] CHANGES PRODUCED BY SELECTION

Connecticut. It was a descendant in a collateral line of
the plant used by Shamel in 1903 as the female parent in
his cross.

Table II, giving the frequency distribution for the num-
ber of leaves of the two parents and the first and the
second hybrid generations, is a complete justification of
our prediction as to how the hybrid type produced by
Shamel originated. The ' ' Sumatra ' ' and the ¥1 genera-
tion were grown at New Haven, Connecticut, in 1911, the
' ' Havana' ' was grown at Bloomfield, Connecticut, in 1911
from commercial seed of the same variety as the plant
used for the male parent, while the F2 generation was
grown at New Haven, Connecticut, in 1912. The F1 gen-
eration, producing an average of 23.3 ± .14 leaves per
plant, is intermediate in leaf number, since the ' 6 Havana ' '
variety shows an average leaf number per plant of 19.8
± .08 and the " Sumatra' ' variety 26.5 ± .11. The varia-
tion as determined by the coefficient of variability is some-
what less for the F1 than for either parent. The value
for the "Sumatra" variety is 6.64 per cent. ± .28 per
cent., for the "Havana" variety 6.98 per cent. ± .27 per
cent, and for the F1 generation 6.24 per cent. ± .41 per
cent. Taking into consideration the probable error in
each case, one may say that the variability of the three
populations is almost the same.

The variability of the F2 generation, however, is greatly
increased. This is shown by the high coefficient of vari-
ability, 10.29 -± .23 per cent., although a glance at the fre-
quency distribution with its range of from 18 to 31 leaves
brings home the point without recourse to biometrical
calculation.

The appearance of the plants in the field corroborated
the data of Table II in other characters. The F1 genera-
tion was intermediate in the various leaf characters, such
as shape, size and texture, that distinguish "Sumatra"
from "Havana" tobacco, and in these characters it seemed
as uniform as either of the parental varieties. On the other
hand, the F2 generation was extremely variable. Some
plants could not be distinguished from the pure "Suma-

THE AMEBIC AX XATURALIST [Vol. XLYIH

tra," others resembled "Havana," although of course the
majority were intermediate in various degrees. Several
plants combined the leaf size and habit of growth of the
"Havana" parent with the leaf number of the "Suma-
tra" parent. In other words, plants were produced in
the F« generation by the recombination of MendeUan fac-
tors that exactly repeated the type which Shamel had ob-
tained in the F3 generation of the reciprocal cross made
in 1903 and which he thought was due to a mutation.
This fulfilled adequately the prediction made by us in
1908.

Results of Selecting for High Number and Low Num-
ber of Leaves ix the "Havana" X "Sumatra"

Cross

In describing the reproduction of Shamel 's hybrid with
numerous large leaves by a reciprocal cross, there has
been a chronological inversion. This was done simply to
show that the original hybrid known commercially as
"The Halladay" was actually a recombination of Men-
delian factors in which the "Havana" and the "Suma-
tra" varieties differed. We will now describe the effects
of selection on the original "Halladay hybrid."

It will be recalled that the selection experiment which
is the principal subject of this paper began with the sell-
ing of 100 seed plants of Shamel 's Halladay hybrid in
1908. These plants were the F4 and F5 generations of the
cross "Havana" X "Sumatra." Plants numbered from
1 to 49 were the F4 generation ; those numbered from 50
to 100 were the F5 generation. They were apparently
breeding true, for the short habit of growth and large-
sized leaf of the "Havana" parent and the goodly num-
ber of leaves of the "Sumatra" parent. The casual ob-
server either would have said with Shamel that here was
a mutation breeding as true as any tobacco variety, or
that a fixed hybrid, a hybrid homozygous in all of its
gametic factors, had been produced. Accurate data
taken on the progeny of those of the F4 and F5 seed plants
which it was possible for us to grow in our limited space,

No. 565]

CHANGES PRODUCED BY SELECTION

however, show that such judgments would have been
superficial. The general type of the plant did appear to
be fixed, but the frequency distribution for number of
leaves of the F5 and F6 populations were not the same.
Strictly speaking, they were not fixed. What would be
the result of selecting (and selfing) extremes from these
different families for a number of years? A tentative
answer to this question is to be obtained by examining
the remainder of our tables.

The tables are arranged roughly in the order of the
effect that selection has had in changing the mean of the
various families that were the starting points of this part
of the experiment. The selections were grown near Bloom-
field, Connecticut, on the light sandy loam of that region,
soil typical of that which produces the famous Connecti-
cut Eiver Valley wrapper tobacco. Duplicate experi-
ments with several of the original families were made at
New Haven, Connecticut, however, on an impoverished
soil not fitted to grow a good quality of tobacco even after
supplying large quantities of tobacco fertilizer, and in
the condition used not fitted to grow good crops of any
kind. Two families were also grown in triplicate, the
third selections being planted at Forest Hills, Massachu-
setts, on a very fine type of rich garden land which brought
out maximum luxuriance of growth, but which did not
produce good tobacco quality. These experiments were
not true repetitions of the experiments at Bloomfield,
Connecticut, since aliquot portions of the seed from the
selfed plant grown there were not sent to the other places
to be grown. But they were duplicates in that each
family came from the same F4 or F5 mother plant,
although, beginning with the F5 or FG population, differ-
ent selfed seed plants furnished the starting point of selec-
tions carried on independently. In this way there were
afforded a greater number of chances to see what selec-
tion could do.

Table III shows the results obtained from family No.
77. This family arose from an F5 plant having 23 leaves,
one below the modal leaf number if we may judge from

THE AMERICAN NATURALIST

[Vol. XLVIII

-H -H -H -H

■H -H -H -H

GO CO »C i— •

CO 1> iO

p< H h cm

■fl -H -H -H

CO iO CO Tfi

O (N LO 00
—i i-i

d c

rt c3
> >
c3 03
WW

X X

c3 c3
s-,

c3 o3 o3 03

> s s s

3 3 3 3

o
c
-

Eh
O

-H -H -H -H -H -H -H

SONOCNNOO
CO CO IN CO CO OS ^

oo co © cd ©' o

1.54 ±.05
1.41 ±.04
2.03 ±.06
1.35 ±.07
2.32 ±.07
1.84 ±.05
1.93 ±.06

18.4 ±.07

21.3 ±.05
21.9 ±.08

22.4 ±.11
24.9 ±.11
26.6 ±.07
25.8 ±.08

CO CO CO iO .-1 o

co o co x cm x x

CM CO CM IN N ^

j— ( i-H

• 1— 1 i-H l-H

. CM Tt*

; ; ; ; io <n co

; ; ; ; o »h cm

.... cN i-i

; ; cm ; io ^ o

. . . iH Tf CM

r--

; : w rH n © m

. . CM iO

Plar

! CM © © © 00 CM
CO CO CO

u
•

h CO -h id OS 00 o

CM CO CO l-O

a
>

HTfOOlOTfrt

h <n h n in Tjt

CM CM b- CO O »0

CO CO CM CO iH CM

_£>

-tf iO !> t« io CM iO
CO IN h

Numl

co t^- co <m i— • ; i-H

CO CO "tf

x o o : : : :

iO CM CM . . . .

cm ua : : : :

co : : : :
*

co : : ; : : !

i-H

;uaaBj JO S9ABarI
jo joqnin^i

O O O CO o o
CM CM CM CM CM CO CO

nMOJto

CM i-t o © O CM

^ _ O rH rH

C! O O O Ci C C!

uonuaanao

en oo r» <o r- oo oo

— — — — — — —

HifH CM CM CN
III III
l>

1> t> l> l>

No,

i.565] CHANGES PRODUCED BY SELECTION

CO O CC O Cl N N

-H -H -H -H -H -H -H

1-1 i-i *-< ^ cm

■H -H -H -H -H -H -H

q n n Tf ifl

6 h ?i n ui to cs

N tN CN (N CN M N

in -h c h tj. x

« ^ X O IN

-I —I <N M — ' -<

5 l-

jo joqranv^

UAYOjr) JB8 ^

rt H S O 2 H H

aoijBjaaaj)

fa fa fa fa fa fa fa

i-h (M N

I I I « I I .1

r~ b- b»

i>. i> t> b- b-

5 £

3§

P
or

^ N H N M
HlOMNtNn

fli till
XI 1 - — b- OC CO

cm q co cm co cm
id cd co co CO CO

1.26 ±.03
1.53 ±.13
1.60 ±.08
1.53 ±.05
1.80 ±.06
1.69 ±.05

23.9 ±.05
25.2 ±.19
24.1 ±.11
24.4 ±.07
26.1 ±.08
26.9 ±.07

Total

C h C) C! o >o
HCOXOMN

CO CM CM CM

Number of Leaves per Plant

• 1—1 1— c

• • • - —1 CM

'. '. ] H 1/J

! i-H

: h o i>

... CM CN

rH i i-h CO ^ rt<

CM lO

Ol CO Tf O lO Ci

CO
CN

h (M iO

24 25

CO 00 O O 00
CO h ic iO CO

»0 CO rl CO
O CM iO CO t-i
1—1

CO
-M

cm co co o a> rt<

b- l-l Tj<

<M
<N

O C5 CO CM i—i
CO rH

O Tt< •

i— 1 l—l i—l

Number
of Leaves
of Parent

CO CM CO b- CO C5
CM CM CM CM CM CM

Year
Grown

i— i C O O i— 1 CM

^ O !-(

O C5 CO OS © ©

Gener-
ation

•

fa fa fa fa fa fa

No.

(76~1)-1
(76-1)
76

(76-2)

(76-2)-l

(76-2)-l-l

l—l

3 -4-1

o >— <

g ^

THE AMERICAN NATURALIST [Vol. XLVIII

the F2 generation of the reciprocal cross where the mode
was at 24 to 25 leaves. The F6 fraternity that it pro-
duced was somewhat smaller than one would wish if
he were to be confident of the calculations made. The
mode is 22 leaves and the mean nearly the same, 22.4
± .11 leaves. From among these plants, a minus variant
having 20 leaves and a plus variant having 27 leaves were
selected to produce the F7 generation. The modes in this
generation are 21 and 25 leaves, respectively, a difference
of 4 leaves ; and the means are 21.9 ± .08 and 24.9 ± .11
leaves, respectively, a difference of 3 leaves. Progress in
both directions continued when a 20-leaved plant was
selected to carry on the minus strain, and a 30-leaved
plant was selected to carry on the plus strain. The modal
classes of the F8 generation are 21 leaves in the minus
selection and 26 leaves in the plus selection, while the
means are 21.3 ± .05 leaves and 26.6 ± .07 leaves, respect-
ively. In the F9 generation the plus selection was lost,
but the minus selection grown from a 20-leaved plant had
the mode dropped to 18 leaves and the mean to 18.4 ± .08
leaves. In order not to lose the plus selection entirely,
however, more of the F8 generation seed was grown in
1912. The mode is the same as in 1911, but the mean
dropped slightly to 25.8 ± .08 leaves.

Here one notices what is very common throughout the
experiment ; the extremes selected for mother plants were
not members of the most extreme classes. This means
simply that vigorous healthy specimens were always
selected as the mother plants, and often the most extreme
variants did not come up to the standard. It is hardly
just to criticize this procedure, however, for with the best
care that it was possible to give, the experiments with
several families were terminated on account of non-
germination of seed or for some similar reason, it being
impossible, on account of the pressure of other work, to
self many plants in each selection. Even where seed
from several mother plants was collected, it did not in-
sure the continuation of that selection. The necessary
space and care involved in growing so many seedlings in

No. 565]

CHANGES PRODUCED BY SELECTION

7.26 ±.25
6.54 ±.21
6.81 ±.27
7.14 ±.26
6.11 ±.20

1.87 ±.07
1.76 ±.06
1.79 ±.07
2.05 ±.07
1.78 ±.06

25.8 ±.09

26.9 ±.08
26.3 ±.10
28.7 ±.10
29.2 ±.08

Total

ONNC5M

• • . 1— 1

• . • 1-1 •

• • • rH CN

rH • . lO IO

CM
CO

rH CM : t> r«

per Plant

r-t CM i CM CO

. rH CM

rH CN »0

ives

t» CO CN CO rH

CM rH rH ■<*

-—

CO
CM

CO CO rH c© OS
rH tH CM CM CO

lO CO rH rH tH

CM iC CO CM CM

3
fc

CO
CM

CO rH OS 00 O

1©
CM

CM tO CO lO CM
rf CM CM

O OS CM CO rH
CO rH rH

CO
CM

CO CO CO .
rH . .

rH • rH • •

Number

of Leaves
of Parent

rH rf CO OS O
CM CM CM CM CO

Year
Grown

rH O OS O rH
rH rH O rH j— I
OS OS OS OS OS
i-{ r~i r-i r-i i-t

Genera-

r- o to <o r»

pH pEH pL< pE< fl<

rH CM

6
3

(19-1)-
(19-1)
19

(19-2)
(19-2)-

>
d

OS CM 00 b- OS 00 tf5

rH TH rH CM rH CM CM

-H -H -H -ti -H -H -H

rH OS GO rH T* rH CO

00 CM CM GO lO © t>
rj? GO lO CO *C CO CO

1.32 ±.05
2.08 ±.11

1.33 ±.05
1.79 ±.07
1.50 ±.05
1.70 ±.08
2.02 ±.08

GO lO CO O rH rH
<Z> r-l O r-i O y-t i-t

-H -H -H -H -H -H -H

CO rH CM CO rH CO O

N iO © N CO O
CM CM CM CM CM CM CO

Total

cm o os o oo o

rf OS OS rH OS O CO
rH rH j— 1 rH i— 1 i—l

Number of Leaves per Plant

: ^

rH CO

co
co

O O

CO CM
rH

CO '. '. i CM t» rH

: . . co

rH CM ! CM CO CO
rH CO

i> rH ; CM CM r» O

rH . rH CM rH CO

00
!M

O CO N ^ CO N OS
CO CM rH CM rH

rH O OS rH l>. rH OS

U5 rH CM CO rH CM

CO
CM

OS 00 lO OS rH rH tH
CM rH rH CN CO rH

•O N N CO lO N H
rH lO CM CM

CO CO CO CM b» ! rH
HHIH

OS lO CO

CM
CM

l> CM rH

r-( y-t r-i

rH • •

Number
of Leaves
of Parent

rH CM rH CO OS OS CM
CM CM CM CM CM CM CO

Year
Grown

CM rH O OS O rH CM
rH rH rH © rH rH rH

Os Os OS OS Os OS OS

Genera-
tion

eo r» <e x> to n. ao

fn fn fn Ph: Pm Pm;

No.

(19-D-l-l
(19-1)-1
(19-1)
19'

(19-2)

(19-2)-l

(19-2)-l-2

THE AMEBIC AN NATURALIST [Vol. XL VIII

isolated seed pans filled with sterilized soil made it im-
possible to start more than two sets of plants for each
plus and each minus selection. Generally both sets grew
perfectly, but occasionally both failed, and in that case it
was usually too late in the season to start a third set even
if it were available.

The second part of Table III shows the results obtained
on the poor soil of New Haven, Connecticut, with the same
family. There was continuous progress in both direc-
tions. The minus selections during the three generations
show a constant reduction of mode, the figures being 23,
22 and 21; the plus selections show an even greater in-
crease in mode, the figures being 25, 27 and 28. The same
decrease and increase occur in the means until in the F9
generation there is a difference of nearly 9 leaves, the cal-
culated means being 20.9 ±.08 leaves and 29.7±.14 leaves,
respectively.

Figs. 1 and 2 show typical plants of the plus and minus
strains of this family as developed by 3 years of selection.
Fig. 3 illustrates an interesting change of phyllotaxy in
some plants of (77-2)-l-l as grown at New Haven in 1912.

Passing to the data on Family No. 76 (Table IV) there
is the same evidence of the effectiveness of selection, ex-
cluding the minus strain in 1910, of which only 31 plants
were healthy. This effect is markedly less than with the
other family. The mode of the minus selection remained
at 24 leaves and the mean was reduced only from 24.1
± .11 leaves to 23.9 ± .05 leaves, — hardly a significant
figure. The mode of the plus selection crept up to 26-27
and the mean to 26.9 ± .07 leaves, there being here one
more generation than in the case of the minus strain.

Table V gives the data on plus and minus selections of
Family No. 19 at Bloomfield for two generations. The
original family stock of the F5 generation has the mode at
27 leaves and the mean at about 26 leaves. A 24-leaved
plant of this generation became the parent of the minus
strain, giving in the F6 generation a population with the
same mode and a slightly higher mean (26.9 ± .08 leaves).
Continuation of the strain through a 24-leaved plant gave

No. 565] CHANGES PRODUCED BY SELECTION

an F7 population with the mode one class lower and the
mean at 25.8 ± .09 leaves. Whether this slight reduction
really means anything we are unable to say. At least, if
it yields at all to selection,
the progress is very slow.
On the other hand, a con-
siderable gain has been
made in the plus selec-
tions. The mode rose im-
mediately to 29 leaves
when the progeny of a 29-
leaved plant were grown,
and went up to 30 leaves
the next generation, the
modal condition being the
same as the number of
leaves of the parent plant.
The means are 26.3 ± -10
leaves, 28.7 ± .10 leaves
and 29.2 ±.08 leaves, the
amount of progress being
— as may be seen — 2.4
leaves and 0.5 leaf in the
two successive genera-
tions. This result appar-
ently indicates a slowing
down of the effect of selec-
tion.

The continuation of the
table gives the results ob-
tained at New Haven on
this same family. Here
there are data from three
generations, and these
data modify the conclu-
sions based on the results

obtained at Bloomfield. Both plus and minus strains
nearly parallel the Bloomfield results for two generations,

Fig. 1. Plant of Halladay Ha-
vana Tobacco (77-2) -1-1, which Av-
erages 29.7 Leaves Per Plant. It
is the Result of Three Years of Se-
lection for High Leaf Number in
Family 77, which Averaged 22.4
Leaves Per Plant in 1909. New
Haven, 1912.

THE AMERICAN NATURALIST [Vol. XLVIII

the F7 generation means being 28.3 ± .11 leaves and 25.1
± .15 leaves, respectively, but in the F8 generations they
differ. Selecting minus extremes for the first two genera-

Fig. 2. Plant of Halladay Havana Tobacco (77-1) -1-1, which Averages
20.9 Leaves Per Plant. It is the Result of Three Years of Selection fob
Low Leaf Number in Family 77. New Haven, 1912.

tions reduced the mean of that line from 26.3 ± .10 leaves
to 25.1 ± .15 leaves, but the third selected generation (F8)
had a higher mean than the original family (27.3 ± .08
leaves) . The parent plant of this F3 generation produced

No. 565] CHANGES PRODUCED BY SELECTION

24 leaves, and as the strain indicated that it was hetero-
zygous for a number of factors by showing a coefficient of
variability of 8.29 ± .42 per cent-, it is possible that the
selected parent plant may have belonged gametically to a
higher class than was indicated somatically; nevertheless,
it can not be denied that three generations of selected
minus extremes have produced no results. This conclu-
sion is not valid for the plus strain. Starting with 26.3 =h
.10 as the mean number of leaves (F5), the succeeding gen-
erations had means of 27.1 ± .07 leaves, 28.3 ± .11 leaves
and 30.0 ± .11 leaves. The differences are 0.8, 1-2 and 1.7
leaves, respectively. Progressive change has certainly f ol-

Fig. 3. Change of Phyllotaxy in Some Plants of (77-2) -1-1 Grown in New

Haven in 1912.

THE AMERICAN NATURALIST [Vol. XL VIII

lowed, and unless one considers that the results of 1912 are
somewhat too high (probably a valid assumption), the
change has increased instead of decreased. Naturally
there must be a decreased momentum in change of mean
time, but this decrease is not yet shown by the figures.

Fig. 4. Plant of Halladay Ha- Fig. 5. Plant of Halladay Ha-
vana Tobacco (19-2)-l-2, which Av- vana Tobacco (19-1)-1-1, which Av-
erages 30 Leaves Per Plant. It erages 27.3 Leaves Per Plant. Three
is the Result of Three Years of Se- Years of Selection for Low Leaf
lection for High Leaf Number is Number Have Proved Unsuccessful.
Family 19, which in 1909 Averaged New Haven, 1912.
26.3 Leaves Per Plant. New Haven,
1912.

No. 565] CHANGES PRODUCED BY SELECTION

Eepresentative plants of the plus and minus strains of
family 19 as obtained by three years of selection at New
Haven are shown in Figs. 4 and 5.

Family No. 5 (Table VI) shows a decrease in mode
from 28 to 26 leaves, and a similar decrease in mean from
28.1 ± .06 leaves to 26.6 ± .09 leaves as a result of the first
minus selection. A second minus selection, however, in-
dicates either that the future progress is to be very slow
or that the entire effect of selection was manifested in the
first selected generation.

With the three parts of Table VII we take up the re-
sults on Family No. 6 at all three stations. The minus
strain was carried on only two generations at Bloomfield,
but with this exception there are data upon three genera-
tions. At Bloomfield the two generations of selected
minus extremes resulted in 0.6 leaf decrease in the mean,
but at New Haven the results were negative, the means
advancing from 25.8 ± .06 leaves to 27.9 ± .12 leaves in
three generations, while at Forest Hill the mean remained
practically the same. Surely selection was unprofitable
here.

The first year of selection from the other end of the
curve, however, resulted in marked progress. The mean
advanced nearly 5 leaves in each case. The original F5
mean is 25.8 ± .06 leaves, but the three F6 means are 30.7
± .09, 29.6 ± .08 and 30.8 ± .12 leaves. This is a remark-
able concurrence of results. The means in the two suc-
ceeding generations were about the same in the Bloomfield
and New Haven experiments, but there was another defi-
nite advance at Forest Hills. Such a result should not
be unexpected. If the F6 generation were almost but not
quite a homozygous lot, and if one assumes that selection
of extremes from homozygous population has no effect
in shifting the mean, it would frequently happen that
some individuals selected to continue the line would be
homozygous in all factors and some heterozygous in one
or more factors.

The cause of the peculiar distribution of the population
(high variability) of the F8 generation grown in Bloom-

THE AMERICAN NATURALIST [Vol.XLVIII

•H -H -H -H -H -H

-H -H -H -H -H -H

-H -H -H -H -H -H
10 co h <n o ©

CO CO CO Oi CO OS

>- a

H H T(* H lO

-H <N JO 00 00

r-4 lO lO Tj<

S3 •©

o os o i-i

1-1 1-1 l-l

HH <N <N

OHNNHPJO
tH <N t-H <N i-H CO

41 -H -H -H -H -H

§ § 3 3 5 S

1.52 ±.05

1.53 ±.05
1.59 ±.04
1.90 ±.06

1 91 ± C\A

3.03 ±.09

25.1 ±.07

25.2 ±.08
25.8 ±.06
30.7 ±.09
oq i -»- n*c.

30.5 ±.13

N H M C) (N N

Number of Leaves per Plant

• <M •

• CM •

i—i

IH O

Tt< l> lO
l> lO

O CO CO

MHH

iH <N

i—i

■ <N

• <N

Num-
ber of
Leaves

of
Parent

Year
Grown

O) C5 C> Ol o o>

U0T*
-■BI9U39

CO CO CO CO co

No. 565]

CHANGES PRODUCED

BY SELECTION

>
d

■H -H -H -H -H -H -H

COOOOHOIO

to o in h io n o

b» Is- CO CO IC CO* lO

00 © rt< © t> ©

q q q q q q p

•h -h' -h -h' -h •«' -h

h N to O CO 00 «5

CM O 00 CO 0C O ©
»— l »— i O O O >-h o

-H -H -H -H -fi -H -H

OS t}J |> CO CO t>- CM

iv" io ui io os oo ©

cm: cm cm cm cm cm co

00 O »C N r-l Tt< id

rt< t)* CS O O »C
rH rH i— 1 CO N H H

Number of Leaves per Plant

00
CO

• 1—1

. . . . o • •

• • • • 1-1 1—1

• • • • O i-l

• • • ■ © CM i-l

[ ; ; ; co co o

CM CM i ! i-t CM CO

CO H H H JO C6

CO J

t> i-l CM rH CM rH tH
i— ( IO CM H^

NHHHOtDiO

i— 1 H H O CO N

CO
CM

rH © H< © CO CM t>
CM H N CO ^ H

.-t <N CO CM tJH CO
CO H CO CO H N

CO
<N

CO CO H N H H H
CM CO I> i—i

co co n o h h ;

i— I CO

(M Tf h o h h ;

H CN lO

icooto ; ; ;
i— 1 1— 1 1— •

CM
CM

co >o H ^ • • •

Numberof
Leaves of
Parent

lO CO lO 00 O O CM
CM CN (N Cm CO CO CO

Year
Grown

CM i-l O OS O i-H CM
i— ( i— i i— i O i— < i— i i— i

OS OS OS OS OS OS OS

Gen-
eration

a s « n is n n

Ph fa fa fa Ph Ph

No.

1-1 1-1
1 1

1— 1 l— 1 rtH

£h £h £h cm'cm'cm'
co ci ci ^ ci ci ci

4.77 ±.18
4.81 ±.19
5.04 ±.25
6.16 ±.17
5.52 ±.27
4.65 ±.18
3.87 ±.14

*o *o CD ^ 00 CO *o
©000©©©

i -h -h' -h -h' -h' m

CM CM CO lO IO CM

l> r— Os co cm os t^-

■w* U "w1 W ^ *w> W

■H 41 -H M -H -H -H

cd co io io o* co cm*

CM CM CM CM CO CO CO

CM CO CO I> OS 00
iO iO OS O OS CO

i-l i-t CO HH

o>
eo

• CM >

t> CM

; ; ; ; cm io rH

'. ; : : ^ i> Tt-

.... CM CO

Plan

eo
eo

: : : : © t- i-i

. . . . --I ^ ©

per

'. '. '. \ © »o co

. . . . rH CM CO

aves

; i-H ; rH io © ©

CM rH rH

: rH i rH rH CO X

o
u

Numl

CO H CX) OS N ; i

rH CO CM . .

00 iO iO CO cm ; ;

(M >0 rH © . .

*£>

t»h © CO iO [ '. '.
CO rH CM IO . . .

^ io © ; : ;

rH HIO . . .

co 1-* cm © ; ; ;

CM
CM

. . . ' . .

Number
of Leaves
of Parent

■HH © IO 00 © iO 00
CM CM CM <M CO CO CO

Year

Grown

CM rH © © O rH CM

1-H 1-H rH O rH 1"H rH

©©©©©©©

Genera-

tion

<*> t- <o to <o r- ao
PH Ph Ph rH Ph rH

rH <M

HH CM CM

d
A

PrH^P C^ M cm"
co © cO © cO © cO

THE AMEBIC AN NATURALIST [Vol. XLYIII

o
3

OOi© wo
<M i-h CM CN <N

-H 41 41 41 41

00 rH rH rH TjH

lO CO CD tO

1.47 ±.05
1.43 ±.05
1.40 ±.07
1.72 ±.05

1.48 ±.05

25.2 ±.07
23.4 ±.06
22.9 ±.09
24.1 ±.08
27.0 ±.08

Total

00 i— < O CO CO
C5 O CM

<M
CO

• • • • CM

• • • i—t

rH • • • 10

'. . C5
i—l

per Plan!

. . CM

CM CM ! 00 rt<
CM . »0

aves

rH CO CO t>- CM
Tj< rH CM rt<

ber of Le

CO^NOO)

Ci l> t> it}

Num

ONH05H

i—l CO CO to

conhn ;

CO N H

; i-h cm oo ;

. CM i—t

O
CM

.' ^ ^ 10 j

• t-H • • •

3 <» eS

Z «M V.

i-l O 05 O i-l
rH rH O i-H rH

05 03 05 05 05

i-h t-i CM CM

II II

CO CO CO CO CO

Z'S

-H -H

•H 41

-H -H

rH l>

■H -H -H -H

•H -H 41 -H

O CM O 05

CO i-i lO 00

rH CM rH rH

41 41 41 41

itS rH 00 O

CO O CO CO its 00 rH

05 rH rH rH GO CO

CM CM rH CM rH CM

t> CO CM rH

C5 Ol CO ^ 00 o o
CM CM CM CM CM CO CO

CM rH O 03 O rH CM

i— I rH i— I O rH rH i— I

05 05 05 0 0)03 05

rH rH rH CM CM CM

CM CM CM CM CM CM CM

No. 565] CHANGES PRODUCED BY SELECTION

field is not clear. It is possible that the plants having
from 18 to 23 leaves were diseased, but no such condition
could be recognized in the field. Again, it is possible
that a few Havana plants were mixed in by mistake,
although as the leaves of the selection are characteris-
tically different from Havana and as the plants with low
leaf numbers resembled the remainder of the row, this
supposition is improbable. The most likely explanation
is that mutation occurred in a few gametes of the mother
plant, a condition that did arise, or that we assume to
have arisen, in Family 41 (see Table X). At any rate,
the change did not follow the path of selection.

In Figs. 6 and 7 are shown typical plants of Family No.
6 obtained by three years of selection in the effort to pro-
duce strains of high and low leaf number, respectively.

Family No. 34 (Table VIII) is peculiar — although this
is not the only time the phenomenon occurred — in that
the F5 population grown from a 24-leaved F4 plant seems
not to have given the true mean. Plants with a low num-
ber of leaves (22 and 20) were selfed to carry on the
minus strain, but both gave means higher than was shown
by the F5 generation. Perhaps further selection will
produce results, but the case is not a hopeful one. The
only evidence for such an assumption is the increased
mean of the F7 plus strain. If it is assumed that 24.0 is
nearer the true mean of the F5 population than the 22.9
actually calculated, then the jump to 27.0 ± .08 leaves in
the F7 generation gives us a basis for expecting results in
F8 in the minus strain.

Nothing can be said as yet about the minus strain of
Family No. 12 (Table IX), for it happened that the first
selection was a complete failure. Six plants were ob-
tained, but the lowest number of leaves was 29. One of
these plants was selfed and gave an F7 population having
a mean of 28.7 ± .09 leaves. Unfortunately the selections
from this fraternity did not germinate and in 1912 we had
to fall back on the reserve seed from which the 1911 crop
came. The crops of 1911 and 1912 are therefore dupli-
cates. The plus strain made an advance from 24.5 ± .10

THE AMEBIC AN NATURALIST [Vol. XL VIII

leaves to either 26.8 ± .07 or 29.0 ± .08 leaves. The first
advance is 1.6, the second 0.7. We can give no explana-

Fig. 6. Plant of Halladay Havana Tobacco (6-2) -1-1, which Averages 30.2
Leaves Per Plant. It is the Result of Three Years of Selection for High
Leaf Number in Family 6, which Averaged 25.8 Leaves Per Plant in 1909.
New Haven, 1912.

No. 565]

CHANGES PRODUCED BY SELECTION

tion of the failure of the results of 1911 and 1912 to dupli-
cate. This is the greatest deviation obtained in the course
of our experiments. The results of 1912 are probably
too high. It is yet too early to say whether or not this

Fig. 7. Plant of Halladay Havana Tobacco (6-1) -1-1, which Averages 27.9
Leaves Per Plant. Theee Years of Selection to Decrease the Leaf Number
of this Type Have Proved Unsuccessful. New Haven, 1912.

THE AMEBIC AX XAIURALIST [Vol. XL VIII

strain is decreasing in the average annual shift of the
mean.

Family No. 41 shown in Table X gave perhaps the most
peculiar results of any of the selections. It may be that
no great shifting of the mean toward the minus end of the
curve should have been expected, because the minus
mothers were each rather high in number of leaves. There
was one with 25 leaves and one with 24 leaves. This was
unfortunate, but was made necessary by the number of
late and diseased (mosaic) plants in the selection. Never-
theless, each of these plants was below the mean of the
previous generation and if a marked change would have
followed the selection of extreme individuals, some change
should have followed the selections of the individuals that
were the actual mothers. But in spite of this fact the
mean persistently rose from 23.9 ± .07 leaves to 26.3 ± .08
leaves, then to 28.1 ± .07 leaves, although the duplicate of
this selection grown in 1912 went down slightly to 27.4
~ .07 leaves. In the plus strain successive generations
of mothers having 2S and 30 leaves caused a small upward
shift of the mean ; it became first 25.7 ± .09 leaves then
25.6 ± .14 leaves, although the 1912 duplicate of the last
population had a mean of 26.9 ± .08 leaves.

The extraordinary phenomenon to which we wish to
call particular attention, however, is not this behavior of
the minus and plus strains in the regular selection ex-
periment, but rather the origin of a few-leaved strain
from a single individual that appeared in the F6 genera-
tion of the plus strain. Eef erring to the table, it will be
seen that in this generation a 12-leaved plant appeared.
This is really a peculiar phenomenon, for we had never
before observed a normal 12-leaved plant among the many
thousands that have come under our observation. They
do not occur. In this population the plant with the next
lowest numbers of leaves had 20 leaves, and in classes 20
and 21 there was only a single plant of each. This 12-
leaved plant was selfed and gave rise to a population
ranging from 8 leaves to 30 leaves, and having a vari-
ability of 23.50 per cent. ± .11 per cent. The mean of the

No. 565] CHANGES PRODUCED BY SELECTION

distribution was 19.8 ± .28 leaves. A 10-leaved plant of
this lot was selfed and gave a progeny with a mean of
17.9 ± .08 leaves and a variability of 11.24 per cent. ± .33
per cent. What interpretation can be given these facts?

We believe a distinct mutation occurred, a mutation
different from those of DeVries. At least DeVries be-
lieves that the mutations that he has observed always
breed true. If the following hypothesis as to the origin
of the 12-leaved plant be true, it is unnecessary to sup-
pose with DeVries that mutations always breed true or
even that they often breed true. Of course DeVries be-
lieves that his Oenothera mutations obey laws different
from those of whose mechanism we know a little. He be-
lieves that species crosses always breed true; that they
do not Mendelize. This belief we hold to be unfounded.
Species crosses have never been shown to breed true.
There have been statements to the effect that crosses be-
tween Eubus species breed true, but no good evidence has
been submitted in their support ; while the data of Tam-
mes ( :11) on Linum species crosses, Davis ( :21) on (Eno~
thera species crosses, and of East ( :13) on Nicotiana
species crosses, concur in showing that species as well as
varieties obey Mendel's Law of segregation and recom-
bination. Furthermore, we think that Heribert-Nilsson's
( :12) beautiful experiments on DeVries 's own material
show that the latter did not collect sufficiently exact data
on his own crosses to find out whether they bred true or
not.

If one is to believe that a mutation in a hermaphroditic
plant breeds true he must suppose that constitutional
changes occur both in the male and the female gam-
etes, or that the change occurs after fertilization. But it
seems more probable that such a change will take place
either in the one or the other gamete and not in both. This
we believe to be the explanation of the appearance of the
12-leaved tobacco plant. A mutation occurred in either
an egg cell or a pollen cell. It does not matter in which
one it is assumed because there is no evidence favoring
either case to the exclusion of the other. This cell with

THE AMERICAN NATURALIST [Vol. XL VIII

U I «

2? ■ ci

g « OS
3rH&H

CD -*»

•H-H-H-H-H-H-H-H-H

lOiflOLOOOOtOO
OOOOOrH(MOO

Hllillll

^HrHrHrH ,-hcNtIHCNCN

OOlNINOMTflNHN

H H CON

CM rt< CO CO

K5U5^tOMO(NOO

NhOO)Ohh(N(N

rH rH i— C O rH r- 1 i— I H H I

CM CM Jh CO rH CO

rH^rHrH CN^cJc^cNcN

5.71 ±.55
6.31 ±.20
6.44 ±.21
6.86 ±.18
6.18 ±.21
5.60 ±.18

t)( ifj ifl lO tfi IC

nqqoqo
M M -H -H* -H

•*NO)tO WO
^ iO iO tO CO iO

«<

25.2 ±.19
24.9 ±.07
24.7 ±.07
24.2 ±.06

26.7 ±.08

26.8 ±.07

Total

lO O CO lO to CM
<N CM CO CN CM

§3

CM •

! CM

00 CO

i CM CM CM t>
CM CM

«
-

h N CO N CO h

CM "Ct*

i per

MOH 05 00 O
CN CN H ^ CO

Leaves

CO CO lO IQ

U5
CM

CO ^ ^ CN W X
CO ^ N CN CN

COKOOOlO
rH *0 l> rH rH

Num

M h CO O (N CN
CO CO ^

CM
CN

HflJOH
rH CO

: o

CN CO CC

\ rH

i rH TJH | j

. . rH • •

Number
of Leaves
of Parent

^ co co «o oo i>

CM CM CM CM CM CM

Year
Grown

CM rH O 05 O r-t
i— 1 rH r-l O r- 1 rH
OS OS Oi OS OJ OS
rH i— I rH rH i—l rH

Gener-

ation

a « h * b it

fn fn pS4 pn Ph

No.

HHHrt CNCN
CO CO CO CO CO CO

<W >»✓

1
!

rfl

(-H

'•+3

«H

l-H

*3
id

CP
fl

insuf

CO
>

00
rH

■n

to*

and i

2
5

■a

r*»

CP 03

> CP
©

S r2

t CP

DO .2

« 3

rfl *2

5 rfl

■ bo

► '§

fl C3

03 CP

cS fl

00 0~

-t-J
O

CP
t-i

rfl rS

E-» <1

00 CS

rfl

No. 565]

CHANGES PRODUCED BY SELECTION 33

5.04 ±.20
6.26 ±.24
6.39 ±.20
6.86 ±.18
5.84 ±.20
7.32 ±.28
7.16 ±.29

1.23 ±.05
1.50 ±.06
1.61 ±.05
1.66 ±.05
1.60 ±.06
1.93 ±.08
1.97 ±.08

N 00 N (O 00 H H
© O O O © rH i-h

-H -H -H -H -H -H -H

^ O CM CM tJ< lO

t)5 tj3 id th" t> co

CM <N <N CM CM CM CM

Total

rt< 00 r- io b- -h cm

\T <-AJ "J ^T"
HHI^OOHHH

rH •

• • • • • o •

'. '. i O rH

• • • j j rH <<* j

. . '. iO (N iO

ant

; ; ; ; © ^ cm

rH rH

I ; CO CM CO O rH
. . CO rH CM

CM CM CO CN CM O O
rH Tt< CM CO

►
S3

CO CO r}< OS CM OS rt«
CM rH Tf CM CM

mber of J

rH -HH CO CO 00 CO

CM rH ^ t}< CM CO CM

IC
CM

^ 00 r>. CM CO CO Tt^

iO CO t> OS CO CM l>
tP^^N CM

O CO O) H IfJ H

CO CO CM <c*

CM
CM

cm oo iq rH ; ; ;

-1 CO . . .

'. *o '. co ; ; ;

• rH T}H j j i

• • . rH • • •

Number
of Leaves
of Parent

CO CO CO 00 00 o
CM CM <N CM CM CM CO

Year

Grown

CM rH O OS © rH CM
rH rH rH O rH rH rH

OS OS OS OS OS OS OS

Genera-

tion

o> oo r» <o r~ oo at

'— pt, pG| pE|

No.

V 1"

rH rH rH rH

II II

- — - - — v ^\ ✓-v - — •> ,• — s

rH rH rH CM CM CM

CO cO CO cb co O co
iO ifl «o U3 iO O iO

>
d

6.14 ±.22
4.50 ±.17
5.01 ±.24
6.86 ±.18
4.26 ±.20
4.91 ±.19
4.42 ±.17

O W5 O iO iO o >o
© © © O © © ©

ji ji ji ji ji ji ji
■H U t) H -H -n -H

OS t> b- CO CO CM 00

IO rH CM CO rH tJ< rH

00 CO OS CO GO 00 CO

o o o o o o c
■H -H i -H -H' -h' -h'

OS O CO CM CM OS l>

id co id t}5 oo co

CM CM CM CM CM CM CM

Total

CO rH H IO CO OS ^
MOOhO"*US

rH rH rH CO rH rH rH

eo
eo

rH •

'. '. '. '. co :

o
eo

io ; ; ; co ^ ^

Number of Leaves per Plan!

OS
<M

"SCO ; CM b- rH CO

00
CM

OOO^NNOOO

rH rH CM CO CM

b- rt< © OS © OS 00

CN CO H H Tf H IO

CD
CM

io os co co os io b-

-*if)Mi*H CO

I©

^HONNHN

CO CM b- rH

H«
CM

b- CM CO OS rH ; b-

io th os ;

"tf . . .

CM
CM

CM rH CM rH \ \ [

CO . . .

! " '. rH '. '. '.

. . . rH . . .

3^3h

CO CO CO CO 00 CM rH
CM CM CM CM CM CO 00

Year
Grown

CM rH © OS O rH CM
rH rH rH O rH rH rH
OS OS OS OS OS OS OS
rH rH rH rH rH rH rH

Genera-
tion

at «o r> e t> oo et
{X, pE, Ph ft,

No.

rH CM
HH CM CM

H H H ^ CNN C?

cO cO CO CO cO CO co
IO KS o >o iQ o

rrj

«g ca

2 ©

o O

« s

§ e

o o

~ u

O O

THE AMERICAN NATURALIST [Vol. XLYIII

a changed gametic constitution, — a loss of gametic fac-
tors,— was fertilized by an unchanged cell. The un-
changed cell may have had any of the gametic possibil-
ities open to the germ cells of the 28-leaved plant of the
F5 family in which the mutation arose, and we know that
certain factors in this plant were heterozygous, for pro-
gressive change followed the selection of a plus extreme
in the next generation. The 12-leaved plant was there-
fore a hybrid. It resulted from the union of a mutating
germ cell of the mother plant that furnished the F6 gen-
eration with an unchanged germ cell. We can even as-
sume that the mutating germ cell, if fertilized by another
of the same kind, would have produced a plant with less
than 12 leaves. The reasons for believing this are simple.
There is experimental evidence (Hayes, 1912) that the
F2 generation of a cross between varieties differing in
their number of leaves is intermediate in character. Our
12-leaved plant is the lone representative of such an ¥x
generation. The F2 generation therefore should give
plants with less than 12 leaves, and in fact such plants
did occur. The distribution marked Fa in the table is
the F2 generation, and this accounts for its extreme vari-
ability. The distribution marked Fb is the F3 generation,
and its variability is less than half that of the preceding
generation.

Family No. 56 was the second family to be grown at all
three of the experimental stations (Table XI). It arose
from a 26-leaved plant of the F5 generation which pro-
duced an F6 progeny with a mean of 24.2 ± .06 leaves and
a mode at 24 leaves. The three generations of the minus
strain grown at Bloomfield remained practically the same.
The last generation did indeed show a mean 1.0 leaf
higher than the original population, but no dependence
can be placed in data from only 25 plants. The data on
the minus selections grown at New Haven are for this
reason a little more dependable. They show a fluctuat-
ing mean, but no progress due to selection, the F9 genera-
tion having a little higher mean than the F6 generations.
The three minus selections grown at Forest Hills also

No. 565] CHANGES PRODUCED BY SELECTION

resulted in higher means, those for F7, F8 and F9 being
25.3 ± .09, 26.0 ± .06 and 25.9 ± .08 leaves, respectively.

This peculiar result implies only that the mean of the
original F6 population which was grown at Bloomfield
was lower than it would have been if grown on the Forest
Hills' soil. This is not a direct effect of environment on
the growing plant. It has been shown conclusively in
our pot experiments, as stated before, that starvation or
optimum feeding has scarcely any effect on the number of
leaves, although it has a marked effect on the develop-
ment of many other characters. On the other hand, en-
vironment does appear to have a marked effect on the
number of leaves that a plant is to develop, if it acts
during the development of the seed. It is well known by
plant physiologists that the environment produces many
of its effects very early in the life history of the indi-
vidual or in the development of the organ concerned. For
example, the so-called light leaves of the beech with two
layers of palisade cells are differentiated from the shade
leaves with only one row of palisade cells by the amount
of light that falls on a branch during the season preceding
the development of the leaves: that is, it is determined
during the laying down of the bud from which the next
season's growth of twig and leaves comes. This period
during which a particular change is possible is called the
critical period for that change by plant physiologists.
Thus a plant may have hundreds of critical periods in its
ontogeny, each marking an end-point of development be-
yond which a certain feature is irrevocably fixed. For
example, the critical period for that cell division that de-
termines leaf size in the beech is much later than that
which determines the number of layers of palisade cells.

Now the critical period for influencing the number of
leaves of the tobacco plant is practically at an end when
the embryo plant goes into the resting stage of the seed.
Before that time the number of leaves may be influenced
by the external and the internal influences that form the
total environment of the mother plant; after that time
environment has little influence on the number of leaves.

THE AMERICAN NATURALIST [Vol. XLVHI

The rise in the mean of the population of the F8 genera-
tion of Family No. 56 is due partially to the effect of en-
vironment, therefore, in that the mother plant was grown
under better conditions, but is probably not to any great
extent due to the conditions under which the plants them-
selves were produced.

The better environment of the mother plants does not
account for all the rise in the means in populations F8
and F9, but it accounts for part of it. It will be noticed
that all of the populations grown at Forest Hills had
higher means than those grown at Bloomfield and New
Haven, although the F6 mother plants were grown at
Bloomfield and not at Forest Hills. The greatest shift
of the mean, however, comes in the F8 and F9 generations,
for the mother plants of both of these populations were
grown on the more fertile soil. There is a simple ex-
planation of these facts, an explanation that is of great
economic importance to practical tobacco growers. A
part of the rise in mean at Forest Hills was due to set-
ting the plants in the field there when they were in an
earlier stage of development than those at Bloomfield and
New Haven. They were not set earlier in the season (at
least, one year they were set early, one year they were set
at the average time and the third year they were set late),
but they were set as small plants. When small plants
(about 4 inches high) are set in the open the root system
is equal to the task of supporting the aerial parts and the
plants start right in to growing normally. There is no
period of passivity. The plants produce leaves spaced
with normal internodes and these leaves develop suffi-
ciently to have a commercial value. But when the plants
reach a height of 8 or 10 inches in the seed pans or seed
beds and are then set in the field, the normal metabolism
is likely to be upset for a time. The plant takes some
time to recover its equilibrium and start a normal growth.
During this period basal leaves begin to develop, but the
internodes are so close together that they do not obtain
their aliquot share of nutriment, hence they grow only to
one quarter or one third their normal size and soon wither

No. 565] CHANGES PRODUCED BY SELECTION

and drop off. The leaf scars are left, but they are so
close together that it is difficult to make a correct count of
the number of leaves. But more important than this,

Fig. 8. Plant of Halladay Ha-
vana Tobacco (56-2) -1-1, which Av-
erages 27.5 Leaves Per Plant. It
is the Result of Three Years of Se-
lection for High Leaf Number in
Family 56, which in 1909 Averaged
24.2 Leaves Per Plant. New Haven,
1912.

Fig. 9. Plant of Halladay Ha-
vana Tobacco (56-1) -1-1, which Av-
erages 24.4 Leaves Per Plant. Three
Years of Selection for Low Leaf
Number Have Proved Unsuccessful,
New Haven, 1912.

THE AMERICAN NATURALIST [Vol. XLVI11

-H -H -H -H -H -H -H

N IN CO N 00 M >0
N CO CO N C N

-H -H -ti -H -H -H -H

■H -H -H -H -H -H -fl

»o lO CO 00 O 00

O) CN 00 H O CO
OO^COCOOW
CN CN h HdlN

tJ* CI O CO O CD CO

CN <N <M CM CM (N <M

CNhOOOhN

_ _ rH O rH —I rH

05 05 Oi Ci C5 C5 05

o
o

_ o

w
p
a*

7.12 ±.23
5.71 ±.19
7.64 ±.25
5.39 ±.15
6.58 ±.22
6.37 ±.22
6.18 ±.34

1.83 ±.06
1.46 ±.05
2.01 ±.07
1.45 ±.04
1.73 ±.06
1.80 ±.06
1.65 ±.09

oo t>. a> co oo C5 co

O O O O O O rH

-H -H -H -H -H -H -H

N CO CO OJ CO (N N

id id co co co oo co

CM CM CM CM CM <M CM

Total

rH l> CO O rH LO CO

CNHHOOOiN
CM CM CM CO CM rH

: : : : :** :

eo
co

• • rH • • CM ■

\ rH rH ' rH '
rH

rH rH 00 Tt< 00 TH CO

Plar

Ci
CN

LO lO CO Ol rH rH GO
rH HNH^I

per

co OS ^ a oo cm

CM rH CM r> CM Tt< rH

aves

CNHtOOO^HH
CO CO 00 tH CO CM

CO
<M

Tt< l> CO CO CO C5 CO
LO CO ^ LO tH rH rH

o
u

£i

IQ
<M

o> O © CO H CO
CO lO CM CM CO rH rH

Sum

O N "O 00 r)t CN tJi

CO CO CM rH rH

Q 00 O CO G '. rH

rH rH

<M
<N

lO CM CM • O ■ *H

CO CM CM • O • •

• rH ' •

Number
of Leaves
of Parent

CO lO 00 O rH (M
(N CM CM CM CM CO
CO

Year

Grown

CM rH O Oi O rH CM
rH rH rH O rH rH rH

OS Oi 05 OS O O OS

Gener-

e> oo r» <o r» oo 9
pH Pm Ph Ph Ph rH Ph

No.

rH CO

1 1
CM CM CO CO

XX^. ^.XX

CO CO CO CO CO CO CO
1> £^£3^ www

No. 565] CHANGES PRODUCED BY SELECTION

the tobacco grower loses an average of from one to two
of his most valuable leaves.

The plus strain of Family No. 56, which we were dis-
cussing when we digressed to speak of the critical periods
of development, did show a considerable shifting of the
mean following the selection of high-leaved mother plants.
In the Bloomfield selections the mean went from 24.2
± .06 to 26.7 ± .08 leaves, then to 26.8 ± .07 leaves ; in the
New Haven experiment the mean shifted to 27.4 ± .08
leaves, — a gain of 3.2 leaves, — and then dropped to 26.4
± .11 leaves, recovering again in the F9 generation to
27.5 ± .11 leaves ; in the Forest Hills experiment the suc-
cessive means were 27.2 ± .08, 28.9 ± .08 and 26.7 ± .06
leaves. Summing up the data from this experiment, it
may be assumed to be reasonably certain that no progress
resulted from the selection of minus extremes, but that
there was a slight effect gradually diminishing in quan-
tity when plus extremes were selected.

Eepresentative plants of Family 56 obtained by three
years of selection in the effort to produce strains of high
and low leaf number, respectively, are shown in Figs.
8 and 9.

Family No. K (Table XII) was grown on a farm near
the Bloomfield experiments, in 1910. The records of the
F5 generation consisted of the number of leaves of only
31 plants. From among these individuals two plants
were selfed to become the mothers of the F6 generation.
Since no dependence can be placed on the F5 distribution
by reason of the few plants and since it is not absolutely
certain that the mother plants of F6 had 20 leaves each,
the selection really began in 1911 with theF7 generation.
There is a difference between the minus strain and the
plus strain in 1911 and 1912, — 0.5 leaves the first year and
1.3 leaves the second year, — however, so that one may
assume the possibility of a slow shifting of the mean in
both directions.

The data on Family No. 73 are shown in Table XIII.
This family came from a 28-leaved plant, one of the
highest of the F5 generation. The F6 progeny of this

THE AMEBIC AN NATURALIST [Vol. XL VIII

2 > »
g » c«

cp O

CO CO CN "HH rH OS 00
|H H M H N rH rH

•h -h -H -h' -h' -h* -h'

tO iO N iO O iO CD

iO rH tO iO iO >fl

o o o o o o o

-h' m -h' i M 4 i

IO N ^ O CO O N
*> CO C3S rJJ CO lO CO
r-5 i-i rH rH rH rH rH

t>. CO 00 CO l> t~ CO

0 o o o o o o

-H -H -H -H -H -H -H

01 co w oo cn q q

lO CO CO t>" co io

CN CN CN CN CN CN CN

Total.

CO t}H Tt< O CN CN lO

rH rH lO O <N CN ©

CO <N CN CO CN CN CO

NHTjlH • • O

cn <N cn os ; ; o

— ■
=

Oi
CM

CO 1> CN O CO CN lO

rH CN CN rH

I> CN OS CO CO t>
CO CN CN rj< rH CN rH

o
p.

as
a>

rH Tf lO OS rH OS OS
CO CO CO O CO CO CN

r of Leavi

«o

OS rH O t> lO 00 lO

CO i> lO CO lO 1>

lO
CM

CO CN rH CO lO t- CO

co co co co io Tt< l>

mbe

O CO rH OS lO rH
rH CN rH CO CO lO

CN TJH CM CN CO CN 00
rH rH CN CO

CM
CN

QOHiOCNNHHt

O • rH • CO H U5

CN CN OS 00 00 00
CN CN CN <N CN CN CN

CN rH O OS O rH CN
i— I i— I i— I O rH rH rH

OS OS OS OS OS OS OS

rH rH CN CN CN

l> !> !> b- b- b- t»

CN <N CN CN CN <N CN

s
p

6.74 ±.21
5.27 ±.17
7.33 ±.25
5.08 ±.14

1.80 ±.06
1.40 ±.05
1.95 ±.07
1.37 ±.04

26.7 ±.08
26.6 ±.06
26.6 ±.09
26.6 ±.05

Total

lO N N

CN rH O OS

CN CN CN CN

Number of Leaves per Plant

CM
CO

CN • ■ •

CO rH CO rH

CO ^J* CO rH
rH

O CN CN CO
CN rH (N rH

00
CN

CO o o o
co co co

O CN CO rH

co co co os

CO
CN

00 CO rH CO
Tt< iO b-

rH lO rH 00

CN CO CO CO

CO rH rH IC
HHWH

CO
CM

CO CM T* b-

lO • CN •

rH '. © •

• rH

Number
of Leaves
of Parent

O O 00 o
CO CO CN CO

Year
Grown

CN rH O OS
O

OS OS Os OS

i-^ i-t t-< y-i

Gener-
ation

££££

No.

(82-2)-l
(82-2)-l
(82-2)
82

No. 565] CHANGES PRODUCED BY SELECTION 41

individual showed a mean of 26.9 ± .06 leaves, and from
among them plants having 25 and 29 leaves, respectively,
were selected to start the minus and the plus lines. These
two mother plants gave F7 populations alike as to mean,
but differing by one class as to mode. The minus line
had the higher mode. The extremes of this generation
used in carrying on the experiment differed by 8 leaves,
and the resulting progenies apparently followed the selec-
tion. The means are 25.6 ± .07 and 28.2 ± .09 leaves.
Whether these shifted means represent a permanent
change or not we are not prepared to say. The minus
mean is probably somewhere near the correct figure for
in the F9 generation it was practically the same, but in
the F9 generation of the plus strain the mean dropped
from 28.2 ±.09 leaves to 26.7 ±.13 leaves. This is a
slightly lower point than that of the original F5 distribu-
tion, but it was calculated from only 76 individuals. A
conservative estimate of the significance of the results
would probably be as follows: the mean of the minus
strain has shifted slightly but permanently and is now
fixed, while the mean of the plus strain has not changed
but has shown evidence of some heterozygosis in one gen-
eration.

We come finally to consider Families No. 27 and No. 82,
the data on which are listed in Tables XIV and XV. Two
generations of both plus and minus selection were re-
corded for Family No. 27, but only plus selections of
Family No. 82 were grown. There is no necessity for
considering either in detail because a simple inspection of
the tables shows that selection has accomplished nothing.

Conclusions

The cumbersome and no doubt dry details of the ex-
periments to the close of the year 1912 having been de-
scribed, let us give a brief resume of the conclusions that
we believe may reasonably be drawn from the data that
have been offered. There can be no doubt that the orig-
inal "Halladay" type of tobacco, isolated and propa-

THE AMERICAN NATURALIST [Vol. XL VIII

gated by Mr. Shamel and Mr. Halladay from the cross
between "Havana" and "Sumatra" tobaccos, arose
through the segregation and recombination of the Men-
delian factorial differences of the two plants, and not as
a mutation. It is simply a union of the factors that stand
for leaf size and height of plant in the "Havana" variety
with the factors that bring about leaf shape and high
number of leaves in the "Sumatra" variety. It hap-
pened that the somatic characters of these varieties ac-
count for all the characters of the hybrid. At the same
time one must remember that strains were obtained by
selection that averaged higher in number of leaves than did
even the ' ' Sumatra ' ' parent. We can only conclude from
this fact that the difference between the "Havana" and
the "Sumatra" varieties in leaf number is greater fac-
torially than somatically. Besides certain factors com-
mon to the two varieties, the factors for leaf number in
"Havana" tobacco might be represented by the letters
AA, and those of "Sumatra" tobacco by the letters BB,
CC, DD, EE. By recombination, this would give plants
with a smaller number of leaves than the "Havana"
variety and plants with a greater number of leaves than
the "Sumatra" variety. Both combinations were ob-
tained; and further, the theory has been shown to be cor-
rect by the results of other crosses where both types ap-
peared (Hayes, '12). It is probably unwise to suggest too
concrete a factorial analysis of the cross, yet the factorial
difference assumed above will account for all of the facts
obtained, by simple recombination. We assume a factor
in the heterozygous condition to account for the produc-
tion of one leaf and a factor in the homozygous condition
to account for the production of two leaves. The mean
of the " Havana" variety is about 20 leaves and the mean
of the "Sumatra" variety about 26 leaves. Somatically
there is a difference of 6 leaves or three factorial pairs
for which to account. But in order to have the theory
coincide with the facts there must be at least one (pos-
sibly two or three) factorial difference that does not show
in the two varieties. The meaning of this statement can

No. 565] CHANGES PRODUCED BY SELECTION

be shown best by an illustration. The 20 leaves of the
"Havana" variety and the first 20 leaves of the "Suma-
tra" variety are represented by 10 pairs of factors, of
which nine are the same and one different in the two
strains. The "Havana" variety is nine leaf factors plus
AA, the first 20 leaves of the "Sumatra" variety are nine
leaf factors (the same as those in the "Havana") plus
BB. The additional leaf factors of the "Sumatra" are
CC, DD. EE. With these assumptions, the recombina-
tions of a tetra-hybrid will represent our facts fairly
accurately. But, as was stated above, it does not seem
wise to take this interpretation of the facts too literally.
That some such factorial combination will represent our
facts superficially there can be no doubt, but in reality if
one could grow hundreds of thousands of individuals and
follow the behavior of each he would likely find himself
constrained to represent his breeding facts by a much
more complex system. There would probably be gametic
couplings and factorial differences whose main effect
would be on some entirely different character or complex
of characters, but which would have some slight jurisdic-
tion over leaf determination. To become diagrammatical,
the unit characters of a house are its cornices, its win-
dows, its floors and what not, but a collection of these
components is not a house. We may even exchange
dormer windows with our neighbor, but we can exchange
them only if they fit. Again, we may put on a coat of
paint, a color unit, but this color unit affects the appear-
ance of many other parts that are just as truly units.

The essential part of our conception of the origin of
this hybrid type is that recombinations of characters
quantitative in their nature can be expected and predicted
in crosses in exactly the same manner as is done with
qualitative characters. On the other hand, it must be
borne in mind that here was a hybrid type that appeared
to be breeding true to the general characters that we have
described, in the F4 generation. That it was not breed-
ing true is clear from the results of the selection experi-
ments, yet out of the small number of F5 and F6 families

THE AMERICAN NATURALIST [Vol. XL VIII

taken under observation at least two were found to be
breeding true for all practical purposes in the F5 and F6
generations. We were able to reproduce the ' ' Havana' '
type by continued selection in Family 77 and were able
to produce strains breeding approximately true to 30
leaves or so by the selection of mother plants in several
families. But can we say that any of our families are
now fixed so that no progress can be made by selection?
We can not. But we can say that some of them are so
constant that it would be a loss of time for selection to be
continued for economic results. It is important to know
whether plant or animal populations can reach such a
state of constancy by inbreeding that no profitable results
can afterwards be obtained by the practical breeder. We
believe it demonstrated by even these few data that such
a state, a homozygous condition, occurs in a definite pro-
portion of F2 offspring, and can be propagated commer-
cially at once if a sufficient number of families are grown
to be relatively certain of including the desired com-
bination.

As to the problem of theoretical importance, the ques-
tion of the true constancy of homozygotes generation
after generation, we believe it to be fair to conclude that
a state so constant is reached, that even for the theoret-
ical purposes of experimental genetics it may be assumed
as actually constant. Further experiment and larger
numbers may show that selection can always cause a shift
in the mean, but will necessarily be a shift so slight that
it can be detected only by a long-continued experiment
and enormous numbers. Assuming for the purpose of
argument that this is the case, the matter would affect
only the question of the trend of evolution. It may come
to be believed, from evidence now unknown, that evolu-
tion may progress slowly in this manner, but if it does,
its course can hardly be demonstrated experimentally be-
yond a reasonable doubt. The problems of experimental
genetics can be attacked, however, from the standpoint
that experimental evidence of the shifting of the mean of
a homozygous population by selection is negligible.

No. 565] CHANGES PRODUCED BY SELECTION 45

Mutations may occur. We have shown the origin of
one family by a very wide mutation. In this particular
case it was not difficult to show that a constitutional
change took place in a single germ cell of the mother
plant. It was only by a lucky chance that this fact could
be demonstrated, for with smaller changes such proof
would be impossible ; but there is no reason to believe that
this phenomenon is unique or even rare. It is much more
reasonable to assume that mutations usually arise in
single gametes than that the same change occurs simul-
taneously in many germ cells. One should expect the
somatic result of a mutation in an hermaphroditic plant
— the sporting plant itself — not to breed true, therefore,
but to behave as an F1 hybrid between a mutating and an
unchanged germ cell. It is true that the mutations ob-
served by DeVries in Oenothera Lamarckiana are sup-
posed to have bred true, but this is sometimes question-
able even from DeVries 's own data. The Lamarckiana
" mutants' ' that did breed true are much more reason-
ably explained as segregates from complex hybrids.
They can be interpreted by Mendelism with no essential
outstanding facts, but if they are to be interpreted as
mutations, several discrepancies between what actually
occurred and what should be expected on DeVries Js own
theory must be explained. It must be shown why the
changes took place in numerous germ cells, — in both the
male and the female gametes, — and why these germ cells
always fused at fertilization; for the changed germ cells
must have fused with each other because many Lamarck-
iana plants were produced by the same mother plants that
produced the mutations, while the mutations are sup-
posed to have bred true. On the only other possible theory
of mutation, that the change occurred in the developing
zygote after fertilization, one would have to explain why
the mutants did not often appear as bud variations, in-
stead of these being much rarer than the supposed muta-
tions, as is actually the case.

We do not deny the theory of mutation as modified to

THE AMERICAN NATURALIST [Vol. XLVIII

assume only that constitutional changes usually occur in
the germ cells, but on this belief the sporting plants must
often be Fx hybrids, and the plant breeder must resort to
selection to isolate his pure mutation. And by the same
reasoning one gametic change may produce many new
creations, for there is a chance to recombine it with all the
known gametic differences in the species.

No one can say how often mutations arise. It is likely
that changes other than the one observed took place in
our tobacco experiments, but it is not likely that they
are sufficiently numerous to base a system of selection
within a pure race on the possibility of their occurrence.
The fact that no changes ensued that could be detected in
several of our selected lines is an argument against it.
The comparatively large jumps are the ones likely to
have the greatest economic importance, and these are
easily detected without refined methods of procedure.
Small jumps can be economically important only if they
are numerous, and, as there are absolutely no data to
show either that they are numerous or that changes can
be produced rapidly within homozygous pure lines through
any other cause, it seems unwise to recommend that the
practical breeder expend time and money to bring about
results that either can not be expected at all or that are
so slow and so trifling that they can not be detected in
carefully planned and accurately executed genetic inves-
tigations. On the other hand, the results of the last de-
cade show that important economic results can be ob-
tained easily and surely by selection from artificial hy-
brids or from the natural hybrids that occur in cross-
fertilized species by the recombination of Mendelian
factors. TVe believe, therefore, that the isolation of ho-
mozygous strains from mixtures that are either mechan-
ical or physiological, that are either made artificially or
are found in nature, offers the only method of procedure
that the practical plant breeder will find financially
profitable.

Finally, we should like to call attention again to the

No. 565] CHANGES PRODUCED BY SELECTION

practical importance of determining the duration of the
period in the course of which particular plant characters
are responsive to the action of environmental influences.
The character complex that has been the basis of this
study is a striking illustration of how results from such
investigations may be applicable to farm practise. One
may plant a portion of the seed from a self-pollinated
tobacco plant on poor soil or on good soil and the average
number of leaves per plant and the general variation of
the plants in number of leaves will remain nearly the
same in both cases.2 But seed selected from mother
plants grown on the good soil will produce plants aver-
aging slightly higher in leaf number than the plants com-
ing from seed on mother plants whose environment is
poor. Consequently, it is better to select seed from well-
developed mother plants — mother plants whose environ-
ment has been good — than from mediocre mother plants.
There is no question here of the inheritance of an acquired
character or of continuing to raise the number of leaves
by cultural treatment. One simply takes advantage of
the fact that during seed formation there is a period of
mobility at which time the potential number of leaves of
the young plant are practically fixed. Pending the end
of this critical period, the number of leaves can be in-
fluenced by external conditions within the limit of fluctu-
ating variability.

In the same connection, the effect of time of planting
on the tobacco plant should again be mentioned, as this
also emanates from environmental change. The actual
number of leaves is, of course, practically fixed at the
time of setting the plants in the field, but this is not true
of the number of leaves that will have a commercial
value. For example, a seedling with 26 potential leaves
is planted. If it is planted when about four inches high,
the general physiological disturbance due to transplanta-
tion is negligible and the plant continues its normal cycle
of development without a pause, bringing to maturity

2 Garner's (:12) results on Maryland Mammoth are an exception to this
statement because this variety is indeterminate in growth.

THE AMERICAN NATURALIST [Vol. XLVIII

about 22 leaves. If planting is delayed until the seedling
is eight or ten inches high, there is a different state of
affairs. Development is arrested, the plant pauses to ad-
just itself to the change. It soon recovers and continues
its normal ontogeny, but the period of reduced growth
has left an ineffaceable record. Several of the leaves —
among them the more valuable leaves — have been so
affected during this readjustment, that they develop to
only a fraction the size that they should attain because
the internodes between them are so short, due to the con-
stricted development that normal metabolism does not
occur. Thus there is a loss of one or two leaves, which
on several acres of tobacco may make the difference be-
tween profit and loss. Hence, the grower should not de-
lay setting his plants in the field until they have become
overgrown in the seed bed.

March, 1913

LITERATURE CITED

Davis, B. M. Genetical Studies in CEnothera, III. Amer. Nat., 46: 377-
427. 1912.

East, E. M. Inheritance of Flower Size in Crosses between Nicotiana

Species. Bot. Gas., 55: 177-188. 1913.
East, E. M., and Hayes, H. K. Inheritance in Maize. Conn. Agr. Exp. Sta.

Bull. 167: 1-142. 1911.
Garner, W. W. Some Observations on Tobacco Breeding. Ann. Rpt. Amer.

Breed. Assoc., 8: 458-468. 1912.
Hayes, H. K. Correlation and Inheritance in Nicotiana Tabacum. Conn.

Agr. Exp. Sta. Bull. 171: 1-45. 1912.
Heribert-Nilsson, N. Die Variability der CEnothera Lamarclciana und das

Problem der Mutation. Ztschr. Abstam. u. Vereb., 8: 89-231. 1912.
Jennings, H. S. Heredity, Variation and Evolution in Protozoa, L Jour.

Exp. Zool., 5: 577-632. 1908.
. Heredity, Variation and Evolution in Protozoa, II. Proc. Amer.

Phil. Soc., 47: 393-546. 1908.
. Assortive Mating, Variability and Inheritance of Size, in the Con-
jugation of Paramecium. Jour. Exp. Zool., 11: 1-133. 1911.
Johannsen, W. Uber Erblichkeit in Populationen und in reinen Linien.

Jena, Gustav Fischer, pp. 1-515. 1903.
Pearl, Raymond. Inheritance of Fecundity in the Domestic Fowl. Amer.

Nat., 45: 321-345. 1911.
Shamel, A. D. New Tobacco Varieties. Yearbook U. S. Dept. Agr., 1906:

387-404. 1907.

Tammes, Tine. Das Verhalten fluktuierend variierender Merkmale bei der
Bastardierung. Bee. Trav. Bot. Neerl., 8: 201-288. 1911.

ORLAND E. WHITE

STUDIES OF TERATOLOGICAL PHENOMENA IN
THEIR RELATION TO EVOLUTION AND THE
PROBLEMS OF HEREDITY

I A STUDY OF CERTAIN FLORAL ABNORMALITIES IN NICOTIANA AND
THEIR BEARING ON THEORIES OF DOMINANCE

Presented in Partial Fulfillment of the Thesis Requirement for the
Degree of Doctor of Science at the Bussey Institution of
Harvard University. 191 3

Reprinted from the AMERICAN JOURNAL OF BOTANY Is 23-36. text figs.

1-4. Ja 1914.

[Reprinted from American Journal of Botany, Vol. I, No. i, January, 1914.]

STUDIES OF TERATOLOGICAL PHENOMENA IN THEIR
RELATION TO EVOLUTION AND THE
PROBLEMS OF HEREDITY

I. A Study of Certain Floral Abnormalities in Nicotiana

AND THEIR BEARING ON THEORIES OF DOMINANCE1
Orland E. White

When Mendel's law was rediscovered, dominance was considered
as essential and as important a principle as segregation. Further
investigation soon demonstrated the phenomenon of "imperfect
dominance," and still later studies led to a substitution of the "pres-
ence and absence" factor hypothesis for Mendel's conception of
contrasted character pairs. De Vries (1902), Bateson (1909), Daven-
port (19 10), Castle and others look upon dominance as an attribute
of the factor or determiner, and according to the last two investigators,
variation in dominance, at least in part, is the result of variable
potency, or variation in the power of a determiner or factor to express
itself in ontogeny. De Vries held the racially older characters to be
dominant over the younger, a conception which the last ten years of
experimental investigation has not upheld. On the other hand, East
(19 1 2) and Emerson (1912) think of dominance as a result of the
activities of one or more specific factors, plus the modifications pro-
duced by the whole factorial organic complex (all the other factors
concerned in the organism's heredity) and by the external environ-
ment (climate, soil, etc.). In other words, under identical genotypical
and external environments, the factor A would always give the same
expression, no matter how often the experiment was repeated.

The chief value of the data which I have to present lies in its
bearing on this important question of dominance. The abnormalities
concerned are three in number, viz., petalody and pistillody of stamens
and that peculiar form of corolla doubling to which de Vries and others

1 Contribution from the Laboratory of Genetics, Bussey Institution of Harvard
University. Brooklyn Botanic Garden Contributions, No. 7. Read at the Annual
Meeting of the Botanical Society of America, Atlanta, Ga., 1913.

ORLAND E. WHITE

apply the term catacorolla. The data on each are given in some detail,
followed by a short discussion and summary.

The work was done in the Laboratory of Genetics, Bussey Institu-
tion of Harvard University, under the direction of Prof. E. M. East, for
whose kindly interest and criticism, I wish to express my appreciation.

The material was obtained from various pure line cultures of
Nicotiana species, which had been under observation for several
years. All pure species used in this study bred comparatively true
and no abnormal variations appeared in them, except in Nicotiana
langsdorffii grandiflora, which was subject to petalody, and gave
evidence of being a hybrid as it was heterozygous for yellow and blue
pollen, the true form according to Comes (1899) having only blue
pollen.

1 . Petalody

This teratological character is an extremely common feature of
garden flowers, and, as usually found, is variable even among the
stamens of the same flower, i. e., one stamen may possess it, or it may
be present in two, three, four or all of them. On one stamen, the
petal-like outgrowth from the filament, which constitutes the char-
acter, may be very small, while another filament in the same flower
may show an anomalous enlargement from three to ten or twelve
times as great. It presents its extreme form in the common double-
flowered races of Dianthus, Rosa, Prunus and Ranunculus. The
majority of gardeners as well as many scientists believe that such
double-flowered races can be created from single-flowered varieties by
selection. A very excellent treatment and historical resume of this
subject is given by de Vries (1906, Chap. 17) in which he produces
historical proof that many of our common double-flowered races arose
suddenly and in full possession of their peculiar character. His
experimental studies led him to assign doubleness because of its
variability, to the category of "ever sporting" characters. In many
of our cultivated races, double-flowered plants quite faithfully repro-
duce themselves if they are fertile at all. The majority of these
races have arisen as mutations, the causal factors of which are largely
unknown. Among horticulturists the belief is prevalent that intense
cultivation is responsible for the anomaly, but there are no data from
controlled experiments to support such a belief. Peyritsch (Goebel
1900, I, p. 195) induced all degrees of doubling ift^the floral organs of

STUDIES IN TERATOLOGICAL PHENOMENA

Cruciferae by artificial parasitization with Phy to plus, and, according
to Hus (1908), Molliard caused the formation of double flowers by
mechanical irritation. From these facts, one may conclude that
double flowers may result from many different causes.

In Nicotiana, petalody arose in at least two dozen plants of four
or five hybrid families on which observations were being made for
other purposes. The pure species from which these hyrbids were
derived, while under observation for five years, never developed petal-
ody. Further, this abnormal condition was never observed in Fi
hybrid generations, although thousands of flowers were examined.

Two of these abnormal plants were self -fertilized, and the progeny,
grown under approximately the same environment as the mother
plant, reproduced the character, showing it to be a hereditary and not
an induced phenomenon. One of the races was derived from an F2
segregate of N. langsdorffii X N. forgetiana. The expression of the
character in the stamens was very variable. Table 1 gives a general

Table i

Number of affected stamens per flower 1 2 3 4 5 Total

Number of flowers 1 14 4 6 25

idea of the extent of this variability among the different flowers of the
mother plant. The progeny, over 100 in number, all possessed the
abnormality. The throats of the corolla tubes in some plants were,
however, almost packed with anomalous stamens; while in others,
perhaps only a single stamen was malformed. An examination of the
progeny plant by plant for differentiating characters showed that
segregation in flower color, habit of plant, leaf shape, etc., had occurred,
indicating that the mother plant was heterozygous for a large number
of factors.

The other race of these anomalous stamen-bearing plants was
derived from selfed seed of a plant which appeared to be N. langsdorffii
grandiflora. The variability of the abnormal character is shown in
Table 2. In 19 12 under the same field conditions, 70 plants were

Table 2

Affected stamens per flower 123 45 Total

Number of flowers 1 20 4 25

grown from selfed seed of this mother plant. The inspection of these
70 plants showed " (HP parent to have been homozygous in all its

ORLAND E. WHITE

grosser morphological features, excepting pollen color. Habit, foliage,
height and floral characters were in all plants practically of the same
type and no evidences of a difference in genotypical constitution were
to be observed except for the case mentioned. The anomaly expressed
itself to about the same degree in all 70 plants, and had I desired to
begin selection work toward securing a double-flowering Nicotiana,
one plant would have been as good a starting point as another.

Summarizing these facts, one finds that where the anomalous race
was heterozygous in many characters, the expression of petalody
was extremely variable; while in the race largely homozygous, prac-
tically no variation in the abnormality was noted.

2. PlSTILLODY

This anomaly consists of the presence of small pistils in connection
with the anthers. Sometimes these little pistils amount to no more
than a style and a stigma; at other times, the anther or pollen-sacs
may be partly changed into carpels and rudimentary ovules produced.
Occasionally such ovules are fertile and produce seeds. An examina-
tion of the literature on the subject shows the character to be neither
common nor rare. Usually it is so small and inconspicuous that it
passes unnoticed, but in the opium poppy, it is showy and character-
izes a distinct horticultural variety. Papaver somniferum var.
monstruosum or var. polycephalum, as it is sometimes called, affords the
material for a very interesting chapter on pistillody in "Species and
Varieties, their Origin by Mutation" (de Vries, 1906, Chap. 13).
The writings of Masters, DeCandolle, and Hofmeister also contain
valuable information on this subject. Masters considered the anomaly
to be an accidental phenomenon, while DeCandolle in his Prodromus
described pistilloid wall flowers as a distinct variety. The pistilloid
poppy is at least a century old, and was grown as a field crop in Europe,
being especially valuable because its anomalous condition did not
allow the capsule to open and scatter the seed. De Vries (1906,
PP- 369-99) found these poppies, in respect to their chief peculiarity,
very sensitive to environment, especially during the first two to five
weeks of their seedling stages. By manipulating the soil conditions
at the proper time, he was able to increase and decrease the anomalous
expression. Plants almost normal and those extremely abnormal
were produced in this manner. Selection had no permanent effect

STUDIES IN TERATOLOGICAL PHENOMENA

on its expression. De Vries classified it as an "eversporting" variety.
Although it was possible almost to destroy the character or inhibit
the expression of its hereditary elements by modifying the environ-
ment, it was never abso'utely eliminated by this treatment. In
addition to the action of the external surroundings, internal factors
must have had some part in making this an extremely sensitive char-
acter, because poppies, like corn, are cross-fertilized, and hence are
more or less heterozygous, and, while the external conditions are no

Fig. 1. Nicotiana flower showing pistillody.

doubt very important for the characteristic development of the
anomaly, the eversporting condition one may ascribe at least partially
to the effect of segregating genes.

The race of pistilloid Nicotianas with which I experimented origi-
nated from the guarded seed of a single anomalous mutant which
was discovered among the segregates of an F2 generation from N.
langsdorffii X N. alata. Two or three hundred of these F2 plants

ORLAND E. WHITE

from the same cross were grown, but no other pistilloid mutant was
found among them. The plant was designated (-2-1A) and in all
subsequent experiments will be known under this number. Over no
of its flowers were examined, all of which showed the character in each
stamen, although there was considerable quantitative variability.
No semblance of an ovary in connection with the pistilloid stamens
was found in these -2-1 A flowers, although this occurred in its de-
scendants. Cuttings of the mutant were made, and selfed seed procured
from which 90 offspring were obtained, 72 of which reproduced the

Fig. 2. Stamens from a single flower showing pistillody in detail.

character in all faithfulness, and were in all apparent respects like the
parent. Eleven of the progeny developed flowers with only two or
thre^ cr at most four pistilloid stamens, and in these, the anomalous
pistils were much smaller than those of the original (-2-1A) or of its
72 offspring. Seven of these offspring entire'y lacked pistilloid sta-
mens. At first, such a state of affairs was very puzzling, as the
possibility of technical error was not taken into consideration. How-
ever, there were sap-colored flowers among the progeny, which was

STUDIES IN TERATOLOGICAL PHENOMENA

very surprising, inasmuch as the hybrid family had contained only
cream and white-flowered plants even to the grandparental generation.
Table 3 shows the ratio of white to colored plants and their stamen
character.

Table 3

Pistillody not

Color Pistillody fully expressed Normal Total

White 71 10 81

Colored 1 1 7 9

72 II 7 90

When I found that some of the progeny with sap-colored (magenta,
etc.) flowers possessed pistilloid stamens, I was more puzzled than
ever, because I had already found it to be completely recessive in the
crosses I had made. When the conception of dominance and reces-
siveness as characteristics, not of the unit "character" or factor alone,
but of the latter plus the effect produced upon it by its internal
(genotypical) and external environments, was brought to bear upon
the problem, the explanation was simple, especially as 90 Fi and'381 F2
progeny of a cross between -2-1 A and 321 (N. alata) had given nothing
but white-flowered plants. During the winter I had been working
with many colored-flowered F2 segregates of N. forgetiana (314) X N.
alata (321) and had not been careful enough about cleaning my pollen-
izing tools before selfing the flowers on the cuttings of the original
(2-1 A) mutant, and, as a result, a few. hybrid seeds were produced.
Pistilloid stamens in the colored-flowered plants were due to dominance,
complete in one case and partial in the others, of the anomalous condi-
tion over that of the normal. In the other 7 progeny with colored
flowers, the expected condition, i. e., the dominance of the normal,
prevailed. Probably all 18 progeny belonging to the normal and inter-
mediate classes were hybrid. Further experiments are in progress
to determine this. The change in dominance is not thought to have
any special connection with the color factors, but is interpreted in the
same manner as the anomalous results secured in some of my un-
published studies on fasciation, viz. : the modifying influence of other
factors. The 18 plants which were causing confusion had, in the
majority of cases, a very different and distinct habit from the original
pistilloid mutant, and this was especially true of the plants with colored
flowers. The 72 or more pure abnormal (2-1 A) progeny were very
similar in habit, flower color and other characters, so much so that I

ORLAND E. WHITE

inferred that the parent plant (2-1 A) had been largely homozygous in
its genotypical constitution.

From the cross referred to above (2-1 A X 321), 90 Fi progeny
were grown, all of which were intermediate in both habit and in size
of floral organs, but absolutely normal as regards pistillody. Two
of these were selfed and F2 progeny grown. The results are tabulated
in Table 4.

Table 4

Pedigree

Normal

Abnormal

Total

(394 X 321) — 2 — iA X 321 — 1

103

185

(394 X 321) — 2 — iA X 321 — 2

152

196

Total

257

124

381

Expected

28575

95-25

381

Deviation

—28.75

+28.75

One family (-2) gave a fair approximation to the 3 :i ratio, but the other
had a large excess of abnormal segregates, which I am at present unable
to account for, because the two families were grown from the same
grandparental stock, and under the same external environment.
Many other characters of a structural nature had segregated in this F2
generation, and the variation in the expression of the anomaly was
large. Many plants were as abnormal, and many much less so than
the grandmother. Other abnormalities appeared, both in pistilloid
and normal segregates. Split corolla tubes and 3- to 4-loculed ovaries
were not infrequent. Some of the segregates, as well as a number of
the pure line (?) progeny, possessed flowers with pistilloid anthers
containing numerous small ovules. Where these occurred, the pollen-
sacs were deformed, sterile, and usually the ovules were exposed, owing
to hypertrophy of the anther-sac walls.

3. Catacorolla

This is not an uncommon anomaly, and hereditary races of it have
long been known, e. g., hose-in-hose primula, and a garden variety
of gloxinia, first described by Prof. E. Morren (see Masters, 1869,
pp. 451-52, figs. 213-14). Catacorolla has been exceedingly well
described by both Morren and Masters, so I shall not take the space
here for a general detailed description, but confine myself to the form
it takes in the particular race with which I worked. This race (4-1 A)
is descended from a single plant which possessed the catacorolla

STUDIES IN TERATOLOGICAL PHENOMENA

peculiarity to a more marked degree than any other one of the 15
anomalous plants which appeared in a family of 50 F2 segregates from
a cross between N. langsdorffii X N. alata. In fact, this hybrid
family was derived from the same grand parental cross as that in
which the pistilloid mutant occurred. Instead of a bud mutation
occurring shortly after fertilization, as was probably the case with
the homozygous pistilloid character (2-1 A), this catacorolla mutant
(-4- 1 A) must have originally arisen as a change in the gametes of
one or the other of the grandparental types or in the cells concerned
in their ontogenesis, if we are to interpret the succeeding experimentally
obtained results in accordance with our general knowledge of heredity.
In the F2 generation grown from guarded Fi seed from a cross between
two normal individuals occurred a segregation of 15 anomalous and
35 normal plants, making a ratio of 7 normal to 3 abnormal or 2.33 : 1.
Supposedly the abnormals would have all bred true, for the one plant
(-4- 1 A) which was selfed produced 20 progeny all of which faithfully
repeated the parental peculiarities in respect to catacorolla, habits of
growth, character of foliage, size and color of flowers, and color of
pollen. It is not supposed that only one "altered" egg cell or pollen
grain was necessarily produced in attempting to explain the place in
ontogeny at which this mutation arose. Possibly many were formed
as the result of a prematuration mutation, but if such were the case,
and if they united with unaltered gametes, the resulting seeds possibly
were not planted, or if planted, only one Fi plant of this sort chanced
to be included in those selfed for further propagative purposes.

Catacorolla in this race is typical of the anomaly as it appears in
other plant species. Petalloid segments are produced outside the
ordinary corolla, and partially adhere to it, these segments having,
colored outer and plain green inner surfaces. In other words, the
normal corolla appears to have been separated at some time during
its ontogeny into five segments. Later when these fused to produce
the normal gamopetalous Nicotiana flower, the union appears not to
have taken place through the careful growing together of the edges of
each segment, but on the contrary, to have been brought about in such
a manner as to leave a seam like that made by a tailor. At the
point of union of two segments, there is a slight waste of material, and
it is this which is reflexed back in the mature flower and gives the
catacorolla effect. The segment then is really a piece of left-over petal.
In some flowers, the petaloid segments are not united with the normal

ORLAND E. WHITE

corolla except at their bases, and, in such cases, other factors have
interfered and effected a distinct separation. The anomalous char-
acter, then, is the result of imperfect fusion of the corolla segments
in ontogeny. This theory is further supported by the relation that
exists between the number of normal corolla lobes and the number of

Fig. 3. " Catacorolla." Nicotiana flowers from the parent plant of the -4-1 A
race, showing variation in the expression of the anomaly.

extra-corolla segments. Table 5 shows the character of this relation-
ship in 28 flowers taken from the original parent (4-1 A).

Table 5

Number of segments per flower 1 2 3 4 5 6

Number of flowers 3 2 6 9 5 3

STUDIES IN TERATOLOGICAL PHENOMENA

A more extended investigation from the standpoint of anatomy and
morphology is necessary before such a theory can be demonstrated
as a truth. The fact that six extra corolla segments are sometimes
present can be explained by supposing that two segments sometimes
result from a single "seam." The size of these segments varies from a
slender, thread-like structure to one as broad as the normal lobe. In
some flowers they fuse and produce a supernumerary corolla. This
variability is characteristic of the race as a whole, i. e.y some plants
are not more variable than others, so that the character may be said

Fig. 4. Flowers from Fx hybrids between catacorolla and normal races, showing
variation in expression ("dominance and recessiveness "). Each flower represents
the typical expression in a single hybrid plant.

to be eversporting only in the sense that a single plant may possess
both very abnormal and slightly abnormal flowers.

Several series of hybridization experiments are in progress, but
they have reached only the Fi stage. The most interesting of these
experiments relates to a study of the dominance and recessiveness of
catacorolla. In addition to the selfed seed produced by the parent

ORLAXD E. WHITE

(4-1 A) plant, a large amount of "open field" seed was gathered from it.
Thousands of hybrid F2 segregates of various crosses such as N.
forgetiana X N. alata and N. alata X N. langsdorffii were grown in the
same field, and in the same year as the 4-1 A parent. These were all
normal in respect to catacorolla, excepting the 15 plants already men-
tioned. Cross-fertilization was more favorable to the production of
seed on this one (-4-1A) selection than self-fertilization. This means
that the open field seed would produce largely hybrid Fi plants.
One hundred and sixty-two plants grown from this seed gave 43
homozygous 4-1 A progeny and 119 hybrid Fi progeny, the latter
representing almost as many different Fi combinations as there were
individuals. As a consequence, they were extremely variable in
almost every taxonomic feature, — in habit, height, foliage; in flower
color, size and shape; in pollen color, and in many other less prominent
characters. Sixty of the 119 were colored, and 59 were white. Some
of the flowers were as small as those of N. forgetiana, while others were
as large as those of N. alata. Fig. 4 is an attempt to show something
of these differences in flower size, as well as in the variability of the
catacorolla character. Each flower represents a single plant. The Fi
variation in the expression of the catacorolla was remarkable. Sup-
posedly each of the 119 plants represented a different genotypical
complex, and hence one would, on the conception of dominance
supported by East, expect a great deal of variability. Table 6 shows
the results of classifying the whole 162 progeny by color and by their
expression of the anomaly.

Table 6

-4- 1 A Hybrids
Color Pure Homozygote Intermediate Normal Total

White 43 33 26 102

Colored _ 11 49 60

43 44 75 162

Those classified as normals showed absolutely no expression of the
character.

Guarded crosses were made between the -4-1A and -2-1A strains.
The genotypical constitutions were very different, as each had a dis-
tinct growth habit, leaf size, etc. About 150 Fi plants were grown in
the same field and under approximately the same conditions as the
other "catacorolla" cultures. In this cross, the Fi expression of
catacorolla was intermediate, with a fluctuation towards complete

STUDIES IN TERATOLOGICAL PHENOMENA

dominance of the normal, although never approaching that state.
The pistillody was absolutely recessive.

4. Discussion and Summary

1. Nicotiana plants showing petalody were selfed and progeny
grown from them. In one race the abnormal character was extremely
variable, some plants showing a large expression, other plants showing
it only to a slight degree. This race varied in many other characters,
proving the mother plant to have been very heterozygous. In another
race, the abnormality was reproduced in all the progeny to the same
degree as in the mother plant. With the exception of pollen color, no
variation in other characters occurred in this race.

2. Pistillody originated as a discontinuous variation and was
inherited in the same manner, crosses with the normal in one case
giving in F2 a progeny closely approximating a simple 3 : 1 ratio.
In two hybrid Fi families, it was completely recessive, while in what
appears to be another hybrid Fi family, it is wholly dominant. The
first two families differ from the last family in a large number of char-
acters, as the ancestry of the latter involves another species.

3. The catacorolla race of Nicotiana originated from a discon-
tinuous variation. When crossed with normal races, the Fi progeny
were either intermediate in character or absolutely normal, though
the individual Fi progeny from each cross showed no variation among
themselves. Great variation existed between the different pollen
parents of many of these Fi individuals.

As a whole, the data secured from hybridizing races of normal
plants with those possessing the three abnormalities discussed above
support the view that dominance and recessiveness are not in any way
attributes of the factor or "character" in itself, but are the result of
the factor expression plus the modifying influence of the environment,
whether genotypical or external (soil, climate, etc.). The variability
in the expression of catacorolla in the 119 Fi plants of -4-1 A crossed
with the 119 different normals is strong supporting evidence that this
conception of dominance is the most tenable of those recently ad-
vanced by geneticists.

LITERATURE CITED

Bateson, W. Mendel's Principles of Heredity, Cambridge Univ. Press, pp. 1-396.
1909.

Castle, W. E. The Origin of a Polydactylous Race of Guinea-pigs. Contrib. from
Zool. Lab. of Mus. Comp. Zool., Harvard Univ., No. 176, pp. 17-29. l9°5-

ORLAND E. WHITE

Castle, W. E. Heredity in Relation to Evolution and Animal Breeding. New York
and London, pp. xii + 1-184. 191 1.
The Inconstancy of Unit Characters. Am. Nat. 46: 352-362. 1912.
Clos, D. Essai de Teratologic Taxinomique. Toulouse, pp. 1-80. 1871.
Comes, O. Monographic du genre Nicotiana, Naples, pp. 1-80. pi. 14. 1899.
Davenport, C. B. Inheritance of Characteristics in Domestic Fowl. Carneg.

Institution Pub. No. 121: 1-100. pi. 1-12. 1910.
DeVries, H. Die Mutationstheorie, Leipzig. 2 Bd. 1901-1903.

Species and Varieties, their Origin by Mutation. Open Court Pub. Co.,

Chicago, pp. 1-847. 1906.
The Mutation Theory. Open Court Pub. Co., Chicago. 2 Vols. 1909-1910.
East, E. M., and H. K. Hayes. Inheritance in Maize. Conn. Agr. Exp. Bull. Noj.
167 and Contrib. Lab. of Genetics, B. I. H. U. No. 9, pp. 1-142. pi. 1-25. 191 1.
The Mendelian Notation as a Description of Physiological Facts. Am. Nat.
46: 633-655. 1912.

Emerson, R. A. Inheritance of Certain "Abnormalities" in Maize. Rpt. Am.

Breed. Assoc. 8: 385-399. 1912a.
Goebel, K. Organography of Plants. Clarendon Press, Oxford. 2 Vols. 1900-

1905.

Hus, H. Fasciations of Known Causation. Am. Nat. 42: 81-97. 1908.
Johannsen, W. The Genotype Conception of Heredity. Am. Nat. 45: 129-159.
1911.

Masters, M. T. Vegetable Teratology. London, pp. xxxviii + 534. 1869.
White, O. E. The Bearing of Teratological Development in Nicotiana on Theories
of Heredity. Am. Nat. 47: 206-228. 1913.

Uberreicht von dent Verf

Sonderabdruck aus der

Zeitschrift fiir induktive Abstammungs-
und Vererbungslehre

1916 Bd. XVI Heft 1/2

Verlag von Gebruder Borntraeger in Berlin W35

ORLAND E. WHITE:

Studies of Teratological Phenomena in their
Relation to Evolution and the Problems of
Heredity.

Studies of Teratological Phenomena
in their Relation to Evolution and the
Problems of Heredity.

II. The Nature, Causes, Distribution, and Inheritance of
Fasciation with Special Reference to its Occurrence in

Nicotiana1).

By Orland E. White,.

Brooklyn Botanic Garden, Brooklyn, N. Y., U. S. A.
(Eingegangen 1. Juli 1914.)

Contents.

A. Introduction p. 50

B. Definition, occurrence, and distribution „ 51

C. Classification „ 56

1. Morphological „ 57

a) Morphological theories „ 62

2. Physiological „ 63

a) General considerations „ 63

b) Inherited (germinal) form of fasciation „ 68

c) Uninherited (somatic) form of fasciation „ 70

D. Discussion and summary of Parts B and C „ 75

E. Review of Mendelian studies „ 78

1. Pisum „ 78

2. Zea mays „ 79

F. Special study of fasciation in Nicotiana „ 82

1. Problems „ 82

2. Materials and methods n 82

a) Description of species and varieties used in the investigation . . . „ 83

b) Number of plants grown „ 94

c) Methods „ 94

*) Contribution from the Laboratory of Genetics, Bussey Institution of Harvard
University. Brooklyn Botanic Garden Contributions, No. 11.

Inductive Abstammungs- und Vererbungslehre. XVI. 4

50 White.

3. Fasciation and environment „ 95

4. Fasciation and selection „ 100

5. Fasciation and hybridization „ 103

a) Fasciation X normal ,,103

b) Fasciation X calycanthemy „ 114

6. Summary and conclusions „ 119

Of. General discussion, showing the bearing of these data on certain

general problems of heredity and evolution „ 128

Table 1—26 ,,135

A. Introduction.

The present paper is the outcome of an extended series of studies
on the phenomenon of fasciation in plants. The first part consists
largely of compiled date on its occurrence and classification, together
with a review of the researches of de Yries on this anomaly. In the
second part, I have described in some detail, a series of hybridization
experiments, in which a mutant variety of Nicotiana tdbacum breeding-
true to fasciation was crossed with several distinct normal varieties of
this same species, as well as with several strains belonging to markedly
distinct species. In the course of this account, I have tried so show
the necessity of dispensing with the latent character conception of the
morphologists and of being more precise in our use of terms. Particular
emphasis has been laid upon the fact that a character always is the
result of both internal and external factors and hence non-existent as
a continuous entity in the germ-plasm of two or more successive gene-
rations of organisms. Characters are either present or absent and
never latent. Characters morphologically indistinguishable, and present
in the same species of organism, may be entirely unrelated when viewed
from the standpoint of cause. Such facts have a very important bearing
on the numerous morphological studies of evolution which the last half
century has brought forth. Many of the morphological studies concerning
the origin and relation of various plant and animal groups must be
reinvestigated from this standpoint before the final word as to their
place in the evolutionary scale can be said, for it is obvious that a
plant with a character caused by a combination of a certain protoplasmic
material with a certain environment is not necessarily even remotely
related to an organism with the same character produced by a different
kind of protoplasmic material in a different or perhaps even the same
environmental medium.

Studies of Teratological Phenomena.

The character of my material has made possible a tabular pre-
sentation of data which will give one an unprejudiced picture of the
actual results from a study of the inheritance of fasciation. The
meristic nature of this character makes such results a I have obtained
more nearly free from the personal element than those involving studies
such characters as colors. Further studies of this particular character
are in progress1).

The major portion of this investigation was carried on under the
direction of Prof. E. M. East, to whom much credit is due, for helpful
criticism and encouragement, I wish also to express my appreciation
for the helpful suggestions and criticisms given me by other members
of the biological faculty of Harvard, and to Director J. H. Maiden, of
of the Sydney Botanic Gardens for the compilation of data on the
occurrence of fasciation in Australia, I have drawn on the papers of
several investigators for text illustrations and for these I have given
proper credit in connection with the legends.

B. Definition, occurrence and distribution.

Fasciation, as commonly defined, is a flattened, strapped-shaped or
ribbon-like expansion of the main axis or axillary organs of a plant; at
the base it is generally cylindrical, at the apex, combed (truncate), or
diffusely branched in "witch-broom" fashion. Its presence in a plant-
may alter the arrangement of both foliar and floral leaves, and increase
their number. Under classification, detailed information concerning both
morphological and physiological aspects of this anomaly are given.

Taxonomically, fasciation is a very widely distributed anomaly,
largely confined it would seem to the vascular plants, as I have been
unable to find records of tj^pical cases occurring in the lower groups.
Hus (1908, p. 83) cites its presence in fungi, but gives no descriptions
or references to where such data may be found. More or less typical
fasciations have been recorded from 102 of the 290 families into which
Engler (1909) divides the living vascular plants. These 102 families
are listed in alphabetical order in Table A. According to this table,
less than two fifths of the total number of families have contained

*) The greater part of this study was presented in 1913, in partial fulfillment
of the thesis requirement for the degree of doctor of science of the Bussey .Institution
of Harvard University.

White.

fasciated individuals. In a consideration of the potential ability of all
plants to become fasciated, this fact is only negative, and hence largely
valueless evidence. Table A is necessarily incomplete. Owing to the

Table A. List of plant families from which fasciated
individuals have been recorded.

Acanthaceae

Combretaceae

Lauraceae

Primulaceae

Aeeraceae

Compositae

Leguminosae

Ranunculaceae

Aizoaceae

Convolvulaceae

Liliaceae

Resedaceae

Amarantaceae

Cornaceae

Linaceae

Rhamnaceae

Amaryllidaceae

Crassulaceae

Loranthaceae

Rosaceae

Anacardiaceae

Cruciferae

Lycopodiaceae

Rutaceae

Anonaceae

Cucurbitaceae

Lythraceae

Salicaceae

Apocynaceae

Cyatheaceae

Malvaceae

Sapindaceae

Aquifoliaeeae

Dioscoreaceae

Meliaceae

Sapotaceae

Araceae

Dipsacaceae

Moraceae

Saxifragaceae

Araliaceae

Droseraceae

Myrtaceae

Scrophulariaceae

Asclepiadaceae

Epacridaceae

Oenotheraceae

Simarubaceae

Berberidaceae

Equisetaceae

Oleaceae

Solanaceae

Betulaceae

Ericaceae

Orchidaceae

Sterculiaceae

Bignoniaceae

Euphorbiaceae

Orobanchaceae

Tamaricaceae

Borraginaceae

Fagaceae

Oxalidaceae

Taxaceae

Bromeliaceae

Gentianaceae

Palmae

Thymelaeaceae

Cactaceae

Geraniaceae

Papaveraceae

Tiliaceae

Campanulaceae

Gesneraceae

Phytolaccaceae

Umbelliferae

Capparidaceae

Goodeniaceae

Pinaceae

Ulmaceae

Caprifoliaceae

Gramineae

Piperaceae

Valerianaceae

Caryophyllaceae

Gutti ferae

Plantaginaceae

Verbenaceae

Casuarinaceae

Haemodoraccae

Plumbaginaceae

Violaceae

Celastraceae

Halorrhagaceae

Polemoniaceae

Vitaceae

Chenopodiaceae

Iridaceae

Polypodiaceae

Clethraceae

Labiatae

Polygonaceae

scattered nature of the literature on the subject, records which would
add other families to the fasciated column probably have been overlooked.
One must also remember that many families are small, inconspicuous
and limited in their distribution, facts which would often make them
inaccessible to observers. From my own studies, I would conclude that
there is no evidence for believing that the individuals of any particular
family may always be exempt from fasciation, but on the contrary, it
would seem reasonable to adopt the view that all vascular plants under
the "right" conditions may become fasciated. Sorauer (1906, p. 334)
practically holds this opinion, while de Vries (1910; 2, p. 502) believes

Studies of Teratological Phenomena.

the character to be entirely absent from some plant "groups", but present
in the individuals of the majority of them in a latent or inactive con-
dition. Moquin-Tandon and Godron (1871 — 72) state it to be more
common in dicotyledons than in monocotyledons, and several present-day
students, including de Vries, have expressed the same opinion.
Differences exist between the various families, genera and species as
to the frequency with which their members become fasciated. According
to Masters (1869, p. 20) the anomalous character is especially common
in certain species of Delphinium, Cheiranthes , Matthiola, Brassica,
Cichorium, Campanula, Euphorbia, Celosia, Fraxinus and Fritillaria.
Records of its occurrence in the Oenotheraceae and Compositae are
extremely numerous. These data, however, as I hope to show later, are
not to be taken as evidence that under the "right" x) conditions, any
one species is more capable of producing fasciations than is any other.

Examples of this anomaly have been recorded from trees, shrubs,
vines, and herbaceous plants. In the first two divisions, the branches
are most frequently fasciated; in the latter, the main stam is usually
altered. Annuals, biennials and perennials are subject to the monstrous
condition. De Vries secured the best developed examples in biennials,
such as Crepis (see Fig. 5), but among annuals, Celosia cristata, when
cultivated properly, produces magnificent specimens.

Halophytic and hydrophytic environments do not seem favorable
to the production of fasciations, as I can find no evidence of such
plants in species characteristically growing under these conditions.
Xerophytic surroundings do not inhibit the creation of this class of
anomaly, as is shown by the frequency of its appearance in the hedge
cactus, Cereus marginatus (Starr 1899) of the dry Mexican plateau,
and in the desert loving genera Cereus and Epiphyllum.

Celosia cristata does not lose its prominently combed inflorescence
under artificial drouth conditions. Sedum is characteristically a xero-
phytic genus, but a well known fasciated variety of Sedum cristata has
existed in Europe for two centuries. Nicotiana tabacum fasciata still
retains its flattened stem when grown in dry, impoverished soil.
Fasciated plants may be produced in xerophytic species and persist
under such conditions, but an environment of this sort is not favorable

*) "Right" conditions may signify many different sorts of environment, as environ-
ment favorable for the production of fasciation in one species may prove altogether
unfavorable in the case of other species.

White.

to their maximum development. Among mesophytes, the anomaly is
common and in this type of plant attains its greatest degree of expression.

Geographically, fasciation may be said to be known from almost
every botanically explored region of the torrid and temperate zones.
The largest number of records, as one would expect, refer to examples
in old settled countries, but these are by no means the most inter-
esting. Phyllocladus glauca often produces fasciated male cones. New
Zealand's especially beautiful tree-fern, Hemitelis smithii has been
recorded as fasciated. A marvellous specimen of Araucaria cunning-
hamii Ait., 79*3 dcm. high, about 18 years old and possessing huge
combed -branches is pictured in "Pines of Australia" ([R. T. Baker &
H. G. Smith 1909] See Fig. 1).

Frequency of fasciation in wild and cultivated plants compared.

The data on the occurrence of fasciation do not appear to support
the commonly accepted notion that teratological variation originate more
frequently under artificial conditions than in nature. This anomaly has
been recorded many times in wild plants.

According to de Vries (1894, 1906) wild fasciated plants of
Crepis biennis, Aster tripolium, Geranium molle, Taraxacum officinale,
Oenothera Lamar chiana, Raphanus raphanistrum and Pedicularis palustris
are common in Holland. Hus (1908) mentions the frequency of fas-
ciation in Erigeron canadensis. Often it takes on the character of a
disease, and sweeps over a locality, affecting only plants of certain
species. Instances of this kind have been recorded in connection with
Rudbeckia hirta, Ranunculus bidbosus, Taraxacum officinale and Lepachys
columnaris. Conrad (1901) in the first case reports the occurrence
of fasciated Rudbeckia plants by the hundreds in a field near Haines-
port, X. J. in 1899. A similar outbreak occurred in Ranunculus in a
meadow near Haddonfield, X. J. in 1893. Mr. C. T. Brues informs me
that several years ago he noticed large numbers of plants of Lepachys
in the fields around Austin, Texas, were affected. Dandelions (T.
officinale) are often fasciated, both in wild and man-made environments.
Sometimes they are numerous, 20 or more plants being noted by
M. Breviere (1881), near the village of Saint-Saulge, France; in other
cases, only two or three individuals have been found.

There are only a few of the large number of observations on fas-
ciation in wild plants that might be cited, but they are believed sufficient
to indicate the especial frequency of its appearance in nature. A com-

Studies of Teratological Phenomena.

plete list would probably contain examples, in some cases by the dozen,
from very nearly all the families listed in Table A.

On the other hand, records of the anomaly in cultivated plants,
though more numerous, afford no proof that it originates more often in

Fig. 1. Fasciated tree of Araucaria eunninghamii, Ait.
(Drawn from a photograph after Baker and Smith.)

White.

domesticated races than in wild forms. This statement is based on the
following facts: First, cultivated plants are much more accessible to
observation than those in nature, and secondly, when the hereditary
form of this variation occurs, it is more likely to be perpetuated, both
as a homoz3Tgote and heterozygote , than were it subject to natural
selection. In view of these facts, the more frequent appearance of the
anomaly in cultivated plants may be erroneously credited to repeated
origin, when it is only a case of hereditary transmission (see
Emerson 1912b).

Among field crops such as sugar beets, corn, peas, sweet potatoes
and pine -apples, fasciations are often extremely numerous. These
anomalous individuals are commonly found growing beside the normal
plants in the same field and under apparently the same environment.
Blodgett (1905) reports a case of a field of peas (Pisum sativum)
where not over 10 per cent of the vines were unfasciated. Conard
(1901) finds the sweet potato so commonly fasciated in all parts of the
United-States where it is an important crop, that he concludes it to be
a hereditary phenomenon. M. T. Cook (1906) mentions a variety of
pine-apple ("Smooth cayenne") which gives over 25 per cent abnormal
fruits of a fasciated nature. Other pine-apple varieties such as ^Puerto
Eico" also produce fasciated fruits, but not in large numbers. Sugar
beet plantations and cornfields both yield a plentiful supply of fasciated
material, the former of the stem, the latter of the female inflorescence
(ear). I doubt if there is any corn-grower who has not observed these
occasional flattened ears in his field. East, Hayes and Emerson
have each isolated pure strains of such plants. Recently I have observed
fifty or sixty extremely fine examples of fasciation among a couple of
hundred hybrid Rosa Wichuraiana plants planted along a Boston parkway.
Other species of cultivated plants in which the anomaly is common are
Lilium speciosum album corymbiflorum, Evonymus japonica, Eubus sp.,
Tetragona expansa, Helianthus annuus, Cucurbita melo (all deVries);
Cotoneaster macrqphyUa (Worsdell, 1905); Pru nus sp. (Maiden, 1913,
White, 1912).

C. Classification.

Variation may be viewed from two angles: the strictly morpho-
logical, which takes into consideration external form, color, anatomical
structure, and other physical features, or the physiological, which

Studies of Teratological Phenomena.

involves a study of the conditions necessary to produce the character,
its transmission from generation to generation through seed, its vege-
tative propagation and the factors favorable to its minimum and optimum
development.

1. Morphological.

Fasciations, on the basis of changes which they bring about in
the external form of the stem, may be divided into linear (the ordinary
form), bifurcated, multi-radiate and ring categories.

The first is the commonest type, the second has been observed by
de Vries to be a variation of the first, and often associated with it.
The third is also a variation of the first, in which the inflorescence or
affected structure separates at the apex into three or more short
branches. De Vries (M. T. 2: 497—8, 1910) figures this type for
Amaranthus speciosus. Quadri- radiate fasciations have been found by
the same investigator in Digitalis httea and Celosia cristata, in the
case of the latter on a branched individual. Tri-radiate fasciated heads
are common in the Compositae.

Ring fasciations are quite distinct morphologically from the other
forms and are not very common. Typical cases are found in the here-
ditary fasciated race of peas. Pisum sativum umbellatum and not un-
commonly in Veronica longi folia and Taraxacum officinale. In Veronica,
according to de Vries, they are less than a centimeter long, while in
Peperomia maculosa (M. T. 2, p. 496, 1910), they are sometimes a deci-
meter in length. Typical ring fasciation differs from the ordinary linear
form in that the main axis becomes distended into a funnel-shaped
structure, with the inner cavity somewhat freely exposed to the atmo-
sphere. De Vries calls these annular fasciations because the vegetative
cone is transformed into annular wall. I have observed a case or two
comparable to ring fasciations in Nicotiana. In Pisum, the character
is strictly hereditary, my statement being based on the observation of
over 300 plants grown in 1912. This type of fasciation appears to be
common in legumes, as it is also the form described by Blodgett
(1905) for Pisum. De Vries is inclined to separate it from the ordinary
type because of its morphological and anatomical peculiarities, but Knox
(1908) says all the various forms are related to each other, the difference
being morphological, not physiological. Possibly the dissimilarities in
anatomical structure in some species and genera bring about the
distinctive character of the ring type, at least in those species where

White.

it occurs as the usual t}^pe form as would appear to be true in certain
legumes.

According to an earlier, but somewhat more detailed system of
classification (Godron 1871 — 72), fasciations were grouped on the basis
of the special part of the plant which they altered. In some fasciated
plants, the inflorescence is often entirely inhibited (e. g. Oenothera
biennis L.), while in others no modification of this structure may take
place. In this manner, Godron sorted out six different morphological
types. Recent physiological studies on this anomaly have shown,

however, that all of these types may occur
within the same species or even on the
same plant.

In order to understand thoroughly
the nature of the changes caused through
fasciation, a somewhat detailed descrip-
tion of the monstrous variations produced
in the different plant organs is ne-
cessary. Those structures which call for
special attention are roots, underground
stems, main stems, branches, leaves, in-
florescence, flowers, fruit, a ad incidently,
seedlings.

Roots. Braun (Worsdell, 1905)
Fig. 2. Fasciated potato (Photo- described fan-shaped aerial roots in the

graphed from an drawing in the CactllS, Epiphyllum hookeri. J. C.
Gardener's Chronicle). CosterilS and J. J. Smith, Jr. (1896)

mention the same anomaly in Saccolabium
blumei (Orchidaceae). Other orchids in which anomalous roots have been
observed, are Aerides crispum and Phalaenopsis schilleriana (Gard. Chron.
1874, p. 703) (Fig. 2). In the former, the roots are usually contracted
into flattened masses, irregularly plaited, and give rise to contorted
ramifications.

Underground stems. Fasciations of these structures occur in
Spiraea sorbi folia (de Vries, M. T. 2, 1910, p. 505), Solarium tuberosum
(Gard. Chron. 1885, pp. 80—81) and Oxalis crcnata "Oka" (Hus, 1906).
In the latter case the tubers transmitted the character.

Main stem. This is the plant structure most commonly altered
by fasciation. The variation in form this takes has been adequately
described in the preceding pages and in a former paper (White 1913).

Studies of Teratological Phenomena.

A more detailed account of the linear type as it occurs in Nicotiana
is given under "materials" in the part devoted to a special study of
Nicotiana tabacum fasciata. Often the main stem is the only organ
altered, and the degree of this alteration in plants of the same pure
line1) may vary from a perfectly normal to an extremely abnormal
condition. De Vries states this variability to be true of all of his
"eversporting" fasciated races. "Ring" and „linear" are the two main
morphological types, but probably
"witch-brooms" should also be
classed as a form. Very generally
(and this is true especially of
fasciation in woody plants) the
surface of the stem is striated by
the prominence of woody fibers
(Fig. 3). Fasciated branches or
main stems, owing to more rapid
growth on one side than on the
other, are not infrequently curved,
presenting an appearance that may
be likened to a shepherd's crook.

Branches. Branches may
remain unfasciated, even though
it is expressed in the main axis.
In Pisum and Rubvs, both are
altered; in Nicotiana, egnerally
only the latter. In cases represen-
ting Godron's fourth class, a
small branch may be the only
part of the plant affected. Examples
of this kind are common in trees
and large shrubs.

Leaves. Leaves are usually modified by the presence of fas-
ciation only through their relation to the main axis. Changes in number
and position are common, changes in size and shape uncommon.
Alterations in number and position are exceedingly irregular in character

(a) G>)

Fig. 3.

(a) Fasciated branch of Acer rubrum.

(b) Fasciated stem of Erigeron canadensis.

*) The term "pure line" as used in this paper refers to the descendants of a
single self-fertilized ancestral plant, and does not refer in any way to the degree of
homozygosity of this descendant population.

White.

as illustrated by Nicotiana tabacum fasciata, in which the phyllotaxy
is distorted and the number of leaves in extreme cases is increased
from the normal 24 to 160, the fluctuating arc for the fasciated race
being from 28 to 160. Occasionally fasciated plants of Nicotana pro-
duce from the same point on the stem, two leaves, which have the
appearance of resulting from a post-genital fusion. J. C. Costerus and
J. J. Smith (1896) describe a fasciated plant of Hymenocallis senegambia
which produced all gradations between single normal leaves, partly fused
leaves, and two independent leaves attached to the same point on the
main axis. The leaves on fasciated stems may be smaller than those
of normal stems, if there has been a very great increase in number.
In clovers and other pinnately-leaved plants, the presence of the anomaly
may increase the number of pinnae (Kajanus 1912).

Inflorescence. Fasciation may express itself in this organ of
the plant by greatly increasing the number of flower -bearing twigs or
by suppressing the production of flowers altogether. In the former case,
if the twigs (pedicles or peduncles) remain unfused, a "witch -broom"
effect is produced, such as occurs in Erigeron, Nicotiana, and some
plumose tj^pes of Celosia cristata. When the floral twigs are shortened,
and the main floral axis is broadened out into a fan -shaped, truncate
structure, a "combed" tj^pe is produced such as is found in the dwarf
races of cockscomb and in Phyllocladus glauca. All gradations exist
between these two main types of inflorescence. In either type,
bifurcated and multi-radiate crowns may occur. Inflorescences have also
been observed in which the anomaly expresses itself very slightly, perhaps
only in the "apparent" fusion of two or more of the terminal pedicles.

Flowers. Flowers borne upon fasciatad stems are usually,
although not necessarily, altered in structure, as is evidenced by the
normal floral organs of Pisum s. umbellatum. Alterations commonly
take the form of a repetition of parts that may extend even to the
locules of the anthers, in linear arrangement of parts, and in hypertrophy
and atrophy. In the first case (polyphylly) repetitions may occur that
include any one or any combination of the four whorls of organs. The
lowest whorl (calyx) is most likely to be modified, and the likelihood
of alteration of the other three follows in the order of their axial
attachment. There is, however, no close correlation between the increase
in number of parts in the separate whorls, as far as I was able to
observe in Nicotiana. The repetition in Geranium molle fasciatum
deVries, is said to represent a series of duplication of whorls, the

Studies of Teratological Phenomena.

normal flower being 5-parted, the abnormal having approximately 10,
15 and 20 lobes. Such a series is no present in JSicotiana, irregularity
in number being characteristic of all increases in the number of parts
to its whorls.

Alterations in flowers expressed in a linear manner or as a single
plane are usually confined to the gynoecium. The other whorls generally
retain their normal shape, unless rendered impossible through changes
in the form of the gynoecium.

Hypertrophy and atrophy are commonly present. The former
expresses itself as an increase in size relations, the latter as abortions

Fig. 4. Fasciated pineapple fruit with 71 crowns.
(After M. T. Cook.)

of organs. Atrophy associated with fasciation frequently takes the
form of abortion of gynoecial and androecial (contabescence of anthers)
structures, producing either partial or complete structural and functional
sterility.

Many minor alterations in the character of each whorl may be
associated with fasciation. Petalody of sepala (calycanthemy) and
stamens, pistillody of stamens, adhesion and cohesion of the different
floral organs, synanthy, syncarpy, and dialysis are somewhat common
attendant phenomena. One may rightly infer from the preceding
account that in the floral organs, the expression of this anomalous
character very often reaches its maximum, and produces its greatest
alterations.

White.

Fruits. The term "fruits" is used in a general sense and not
in its strictly technical meaning. The seed capsules of Nicotiana t.
fasciata are frequently distorted, hypertrophied and atrophied structures,
always having an increased number of locules. Pasciated medlar fruits
(Owen 1885) have been recorded which are curved like a ram's horn
and possess 40 instead of 5 calyx teeth. Ferhaps the most striking
example of the effect of fasciation on fruits, may be found in certain
varieties of the pine-apple, Ananas sativus. The "pine-apple" of course
is a multiple fruit, in reality a mature inflorescence, so that properly,
this example should be described as fasciation of the inflorescence.
Cook (1906) describes the following variations of the anomalous pine-
apple fruits: fruits with 2 separate crowns, with two united crowns,
with. 3, 4, 5, 6, 7, 8, 13 separate and compound crowns; flattened
fanshaped compound fruits enlarged by a more or less continuous series
of crowns. A specimen of the latter character, weighing 18 pounds,
and containing 71 crowns, was observed (Fig. 4). Conard (1901)
mentions the occurrence of large fasciated fruits on the commercial
variety of strawberry "Clyde".

Seedlings. Seedlings of fasciated dicotyledonous plants not
infrequently posses more than the normal pair of cotyledons, but even
in such hereditary races of the anomaly as Nicotiana iabacum fasciata
and Celosia cristata, the great majority of the young plants are normal.

a) Morphological theories.

Two theories regarding the morphological nature of the fasciated
organs have been advocated, each by a famous botanist. Moquin-
T an don holds that fasciation is the result of the flattening (enlargement)
of a single growing point. Linne, on the other hand, held it to be
the result of an increase in number of buds that, owing to their crowded
quarters, subsequently fused. A discussion of the arguments for and
against each theory is given in Masters' Vegetable Teratology (1869),
Masters himself, concludes in favor of the opinion advanced by Linne.
Eecent investigators on the anatomical structure of fasciations are
inclined to agree with Moquin-Tandon, as the internal vascular
structure does not appear to uphold the "concrescence theory" of Linne.
Compton (1911) on the basis of detailed investigation of ring fasciation
in Pisum s. umoellatwn concludes the anomaly to be the enlargement
of a single growing point; although he advances a suggestion which

Studies of Teratological Phenomena.

would explain the manner in which the peculiar anatomical features
could be produced through the fusion of several normal stems.

Knox (1908) from an investigation of the anatomical features of
fasciated Oenotheras also concludes in favor of Moquin-Tandon's
deduction. She finds no evidence of fusion of stems in the growing
region and calls attention to tho fact that ring fasciations may break
on the side and develop the linear type. According to this author all
types are the result of the enlargement of a single growing point.

Church (1905) and Worsdell (1905) favor a very modified form
of the concrescence theory of Linne. Worsdell believes fasciation to
be the result of a compromise between two inherent ancestral tendencies,
and rarely a case of real mechanical fusion in the Linnean sense. Two
opposed forces are operating in the organism, — one inducing integrity,
the other producing plurality of parts. Fasciation in higher plants is
a reversion to the ancestral branching character of the low7er plants,
such as lycopods, ferns and algae. In other words, this anomalous
character in a morphological sense is a case of the congenital fusion
of an unusual number of branches. The reason advanced to account
for the fusion in one plane is the ancestral or primitive branching
character, — algae, ferns, etc. being said to branch primarily in a
single direction1). In this sense, Worsdell' s hypothesis of congenital
fusion and increase in number of parts is a modification of the "con-
crescence" theory.

The morphological aspect is mainly descriptive and gives but little
insight into the real or perhaps more fundamental nature of fasciation.
On the other hand, physiological investigation, though increasing the
complexity of the problems involved in a study of fasciation, gives at
least deeper descriptive knowledge of its nature and the causes which
produce it.

2. Physiological.

a) General considerations.

Variations of any kind, from the standpoint of physiology, may
be placed for all practical purposes into twro main categories, those
inherited (germinal) and those uninherited (somatic). This statement

*) Braun as far back as 1859, advanced a somewhat similar view (see Roy. Soc.
Publ., London).

White.

implies that such a classification is very simple, which is very far from
the case as the following pages will abundantly testify. Characters, as
we have come to use the term, are definite morphological realities.
We divide a plant or animal up into parts, more or less arbitrarily,
using function, origin, form, or some other criterion as a basis, depending
on the special phase of biology in which one is most interested. A horn,
a pattern, an organ, become, through continuous familiarity with this
thought, absolutely definite entities, entirely separated, in our minds,
from the remainder of the organism. In this way, we come to think
of the brown eye in man, not as two or more separate character-
entities, but as one distinctive character, whereas from the standpoint
of genetics, the color and the remainder of the eye must be considered
separately. As systematists , it becomes hard for us to adopt the
physiologist's or perhaps the chemist's basis of classification, founded
as it is upon experimental evidence, the methods and nature of which
we find rather strange. We are prone to think of the inheritance of
characters as though they were actually handed on from cell generation
to cell generation, a conception very foreign to fact, for the character
is the combined expression of a bit of protoplasm and a specific
environment.

In dividing characters into hereditary and non-hereditary classes,
I realise I am adopting an arbitrary classification which represents but
a part of the whole truth. But it represents that phase of the question
in which I am most interested in a clear manner. As I understand
the term, heredity simply implies that a given material under a given
specific condition or environment presents certain physical phenomena
which we describe as characters. We start with a standard material,
and if other material under the same conditions does not present this
character phenomena, wc hold it to be absent, but if under other con-
ditions it can be induced, we call it an environmental effect, and contend
that it is not inherited because under its normal (usual) conditions,
the offspring will not reproduce it.

Looking at the character fasciation from this standpoint, it
becomes comparatively easy in many cases to distinguish between the
inherited and the non-inherited form, but in certain cases, the difficulties
of classification are very greatly increased through our general ignorance
of the nature of the material and the various combinations of material
and environment necessary to produce fasciation. Environmental and
hereditary effects are apparently hopelessly mixed.

Studies of Teratological Phenomena.

Take, for example, the investigations of de Vries upon this
anomaly, the results of which were partly responsible for his formulation
of the conception of "eversporting characters". This investigator defines
an eversporting race of plants as one in which the „character" under
observation is inherited by all the individual progeny, but only expressed
(somatically) in part of the individuals. In other words, some of the

Fig. 5. Fasciated rosette of Crepis biennis.
(After de Vries.)

progeny possess the character fasciation and some do not. This per-
centage could be and was increased by selection in a plus direction,
but the permanency of the increase was always subject to the caprice
of the environment. Even in a homozygous pure line, de Yries looks
upon this anomaly as inconstant, although its transmission to all the
progeny is said to be perfect. In some of de Vries' cultures, this
character was so inconstant that often over half the individuals of one
of these races were normal, although the race was held to bre edtrue

Induktive Abstammungs- und Vererbungslehre. XVI. 5

White.

to fasciation as far as transmission was concerned, for seeds of self-
fertilized normals produced progeny in which the percentage of fasciated
individuals was nearly as high as that from seed of self -fertilized
abnormal plants. Some fasciated races were poor as regards somatic
expression, others were rich, and a race producing few anomalous plants
could not be induced by selection to give a higher percentage of them.
Rubia tinctorum and. Pedicularis palustris are typical examples of the
former, in cultural trials producing the anomaly in very few individuals,
even under what appear to be the most favorable environmental con-
ditions (de Vries, 1906, p. 410). These poor races are technically known
as half races.

On the other hand, the rich races (eversporting or '-middle" races)
often gave as high as 50 per cent fasciated progeny.

Crepis biennis (fasciated) is a typical example of the latter and I
will recount its history in the Amsterdam garden as it is related by
de Vries.

Crepis biennis is commonly fasciated in Holland and de Vries
found two such plants growing among hundreds of normal plants in a
meadow near Hilversum in 1886 (Fig. 5). From the normals, he collected ripe
seed and from this seed grew about 100 plants in 1887 — 88, 12 per cent
of which were fasciated. Seed from fasciated plants of this generation
gave 120 plants, of which 40 per cent showed fasciated rosettes the
first year. The remainder were destroyed. Of the 40 per cent fas-
ciated individuals, three of the finest fasciated plants were selected and
used as seed parents for the next (4 th) generation, which gave 30 per
cent fasciated plants. The fifth generation gave 24 per cent; the sixth
was very rich in the anomaly, although no exact figures are given. The
seventh generation produced only 20 per cent fasciated progeny, only
rosettes of the first year being counted. The eighth generation was
sown on a small scale and no percentage is recorded. The hereditary
constitution of this race is said to be fairly constant under normal
conditions and the average percentage of fasciated individuals fluctuates
between 30 and 40.

Out of 350 plants raised from seed of isolated normal (atavist)
individuals of the Crepis biennis fasciata race, about 20 per cent were
fasciated. Again in 1895, 41 individuals of Crepis biennis fasciata were
abundantly manured with horn meal. Under these conditions, the number
of fasciated individuals rose to 85 per cent, the race under ordinary
conditions at this time, producing 20 to 40 per cent. In the manured

Studies of Teratological Phenomena.

culture, the plants stood close together, or, says de Vries (M. T. 2:
p. 516) "I should probably have succeeded in inducing the anomaly in
every one of them."

No sharp limit, according to de Vries can be drawn between the
normals (atavists) and the fasciated individuals, and he again uses the
Crepis race to support his statement. From an isolated group of
3 fasciated plants of this race, seed was saved from the one most
abnormal and 150 progeny raised under the most favorable conditions.
The following results were obtained:

or, altogether, about 80 per cent were fasciated. The breadths of the
108 fasciated stems were tabulated by classes and a curve plotted,
0 indicating the group of the 33 normals (atavists). Transitions between
normals and fasciated individuals occurred, but are said to have been
relatively rare. Two pure types were produced then, as shown by the
two peaks of the curve. Practically the same results were secured in
all the numerous fasciated races with which de Vries experimented,
and even the cockscomb, Celosia cristata, belongs to this category of
eversporting hereditary varieties, although in its case, "complete atavists
are very rare". (M. T. 2: p. 519.) De Vries (see 1906, p. 401, also
M. T. 2: pp. 525 — 526) summarizes his conclusions regarding fasciated
eversporting varieties as follows: (Nos. 7 and 8 are a free translation).

1. "Races always consist of fasciated individuals and atavists"
(normals).

2. "The proportion of the former varies greatly, often amounting
to only 40°/o or less, but not infrequently to more (Geranium
and Crepis with 65°/o and 85%; Celosia cristata).'"

3. "The fasciated individuals are connected by transitional forms,
which are, however, rare; and the statistical curves representing
them have therefore 2 apices."

4. "These proportions are to a large extent dependent on external
conditions of life, which can transform atavists into fasciated
plants and vice versa. This transformation obviously takes
place during the plastic period in youth, before the character in
question is actually developed."

Stems without fasciation . . .
„ slightly fasciated at top
„ fasciated along whole length

. . 33

. . 9

. . 108

Total 150

5;;:

White.

5. "The atavists, as well as the selected individuals, produce
fasciated offspring, and often in proportions very little lower
than those in which the selected individuals produce them."

6. "Between the broadened specimens and the atavists there is
no essential or fundamental difference, in spite of the great
difference in their external forms."

7. Fasciation is due to some internal hereditary quality, which
though often latent, becomes active in response to external
conditions. Its wide distribution causes one to assume that
it arose in some common ancestor of the forms which now
possess it. Hence, phylogenetically, it must be very old.

8. Poor races and rich races both may be strengthened or weakened
by selection and I treatment, but the limits between races are
never transgressed. A poor race cannot become a rich one
through selection. The external conditions being the same, the
hereditary factor must be variable.

The de Vriesian conception of eversporting hereditary characters has
burdened the science of genetics with an extremely complex inter-
pretation of a set of facts that may be given a much clearer and simpler
explanation, and certainly more in accord with the modern strict use
of the term heredity. De Vries has urged the need of much more
research upon anomalies, and gives his results and conclusions rather
as suggestions than unquestionable facts. Further discussion of ever-
sporting races is given at the conclusion of this review.

b) Inherited (germinal) form of fasciation.

Pisum sativum umbellatum Mill., Celosia cristata, Nicotiana tabacum
fasciata, and some races of Zea mays L. are well known examples of
hereditary fasciated races. With the exception of Nicotiana, nothing
is known concerning the manner in which they originated or the
genetic character of their immediate parents. Numerous observers vouch
for the absolute constancy in the transmission of this character by seed.
Pisum s. umbellatum is figured in Gerarde's Herball as a separate
species. Lynch, (1900) Kornicke and Rimpau have grown this race
for a number of years, and all three regard it as perfectly constant in
its transmission of the fasciated character. My own observations have
led me to adopt the same view. Goebel grew Celosia cristata, and
found, contrary to de Vries that it was absolutely constant as regards
fasciation.

Studies of Teratological Phenomena.

Races which are suspected of being hereditary, but about which
little is known genetically, have been from time to time recorded as
variants of Cirsium (Moq.-Tand.), Reseda and Myosotis (de Vries),
Curcurbita pepo (Mazzani and de Vries), Oxalis crenata (Kuntze
and Hus), Ipomoea batatas Poir. (Conard) and Ananas sativus Schult.
(M. T. Cook).

There is another class of fasciation commonly present in woody
and herbaceous plants, which appears to be transmitted asexually. There
is no experimental evidence that they
are germinal variations, but the fact
that the anomaly reappears in every
season's renewal of growth, is regarded
by some observers as proof that it is a
hereditary phenomenon. In Abies (Fig. 3)
de Vries describes a fasciated condition
that reappeared year after year in every
season's growth of wood. Hus (1906)
gives similar facts regarding a specimen
of Rhus diversiloba. Repeated annual
fasciation is a charactistic of a specimen
of Sophora secundiflora (Vasey, 1887)
described from Texas (Fig. 6). Rheum
mooreroftianum (W or s dell, 1905)
plants at Kew send up a number of
fasciated shoots each year. The sweet
potato1) regarded by Conard as a
constant fasciated race, has been pro-
pagated entirely asexually and in this
manner, the anomalous character is said
to have been so widely distributed,
that in many areas unfasciated plants are difficult to find. Numerous
instances of a similar nature are recorded in connection with other woody
and herbaceous plants.

Unless the situation were considered carefully, one might conclude
prematurely that in these plants the anomalous character is reproduced
through seed, but no one so far as I know has demonstrated this to

*) One may consider fasciation in this case a bud sport and account satisfactorily
for its wide distribution asexually.

Fig. 6. Fasciation in Sophora.
(After Vasey. Photographed from
a drawing in the Bot. Gazette.)

White.

be a fact. On the other hand, there is some circumstantial evidence
that certain of these cases may need the services of a pathologist rather
than those of a geneticist in order to determine the nature of their
ailment. I have no doubt, that in some instances, the presence of
fasciation in wood}' and herbaceous plants is a strictly hereditary
character. In other instances, some of which have come directly under
my own observation, I believe the anomalous condition to be due to
perennial fungi or to bacteria. I have no other proof than circumstantial
evidence for this statement, but several perennial fungous diseases of
economic importance, such as those resulting from the presence of
several species of Exoascus, are known to be capable of bringing about
marked modifications in the character of the host-plant structures. As
a rule, they do not affect every individual plant of a group to the same
extent on their first appearance, even though these plants may all belong
to the same variety. Often only a few branches show the anomaly,
the disease in some years gaining, in others, losing ground. Some
species of Exoascus produce "witch -brooms", which in the matter of
increasing the amount of woody tissue through stimulation, is comparable
with what takes place in the production of a fasciated branch.

c) Uninherited (somatic) form of fasciation.

De Vries and others regard heredity as a matter of degree. A
single fasciated plant appears in a normal culture of a species, and the
next year, seed from this plant produces another large culture with
perhaps a single fasciated plant or perhaps three or four present. The
per cent is small and cannot be increased by selection. It is designated
as a poor race. Rich races produce larger numbers of the anomaly.
In order to show that the value of heredity as a conceptional term will
be decreased if a sharp line cannot be drawn between non- inheritance
and inheritance of characters, it seems to me necessary to emphasize
the importance of this point, and I shall go into greater detail here
than the subject would otherwise warrant, Somatic fasciations may be
classified under several heads on the basis of difference in causal
factors, although these factors, from a physiological standpoint, function
in producing the character in the same manner. Fasciations may be
caused by insects injuring the young embryonic tissue, by mutilation
through the agencies of frosts, higher animals and man by abruptly
increasing the supply of nutriment, either by checking the plant's ability

Studies of Teratological Phenomena. 7 ]

to use it or by an actual increase, perhaps by fungi and bacteria, and
other unkown factors.

Insects. Knox (1908) has shown the moth Mompha to be the cause
of fasciation in several Oenothera species (Fig. 7). Injuries were inflicted

"Fig. 7. Fasciated plant of Oenothera parviflora. (After Knox in Carneg. Publ.)

White.

on the initial meristem of the growing point and could only be detected
microscopically during the earliest growth stages. The greatest number
of fasciations were produced under optimum conditions, showing increase
of nutriment in an abrupt manner to be the real cause, though this
abrupt increase was made possible through crippling the metabolism of
the plant by the insect. Knox's investigations are the most extended,
but other observations on other plants support her conclusions.

A specimen of Hieracium vulgatum with a broad flattened stem
was found in relation to a gall of Aulax hieraeii (de Yries, 1901, p. 291).
Below the gall the stem was normal, above, it was fasciated. Hus
(1908) figures fasciated specimens of Erigeron canadensis, containing
larvae of Cecidomyia erigeroni. Other specimens contained larvae, and
though abnormal, were not fasciated. In all the specimens examined,
the fasciated and abnormal modifications began only above the gall-like
swellings. Molliard (1900) found coleoptera larvae at the base of
fasciations in stems of Raphanus raphanistrum L. and lepidopteran
larvae occupied the same position in fasciated Picris hieracioides indi-
viduals. In the former case, a score or so of affected plants were
examined and the larvae were present without exception. In the latter
example, the relation between the insect and fasciation was not so
clear. More recently, the same investigator has shown that a relation
exists between the presence of fasciation in Senecio jacobea and certain
insect larvae inhabiting their roots or the bases of their stems.

Branching palms are not uncommon in India and are classed by
F. Scott of the Agri.-Hort. Society of India as of the nature of fas-
ciation. A particular case is given on the authority of Dr. Beaumont
(Gard. Chron. 1874), which is unique. This is a specimen of the
„ common date palm with 22 branches, 18 of which rise vertically, and
are so closely packed that it was not possible to give a clear idea of
them in the picture". S. Pulney Andy (1869, p. 661), commenting on
these branched palms, states that the intelligent native farmers, give
insect depredations, particularly beetles which bore into the growing
point, as the cause of this condition, especially as found in bifurcated
trees of Cocos nucifera. Petch (1911) states that the fasciations
frequently present in young trees of Hevea braziliensis are probably
due in some cases, to insect and fungous attacks, although these factors
will not account for the presence of the anomaly in every cases.

Fraxinus excelsior and F. ornus are often affected with "fasciations"
which are sometimes "so abundant that it looks as if the trees had

Studies of Teratological Phenomena.

been sown with them", according to Kerner and Oliver (1902, p. 549;,
These "fasciations" are caused by a gall-mite Phytoptus. Judging from
my own observations, and from pictures, these are not typical fas-
ciations, such as really occur in ash trees at times. The typical linear
fasciation illustrated in Fig. 119 (Fraxinus excelsior, see Kidd 1883)
may occur in connection with these hypertrophied inflorescences, but as
to this I have no information. Cases of fasciation which may be and
have been interpreted as the result of insect mutilation are numerous,
but definite information is absent from these observations in the majority
of cases.

Natural elements, higher animals and man. The examples
of traumatic response to injuries from these sources are numerous, but
not always accompanied by desirable details. Cereus marginatus, under
the name of "Organo" is largely used as a hedge plant in Mexico.
A hedge of these plants (Starr, 1899), (Hus, 1908, Fig. on p. 86)
which were partly injured, probably because cuttings were taken from
them for planting, showed numerous fasciations: Krasan (Klebs,
1903—06, p. 134) observed fasciations induced by loss of foliage through
the action of June beetles or spring frosts. According to Sorauer
(1906, p. 334), a fasciation in Tecoma radicans was brought about
through appression to a wall, the parts above the wall also showing
the anomalous character. Lopriore (Hus 1906) however, did not
succeed in producing this condition in Vicia roots through prolonged
pressure. An asparagus grower (Hus, 1906) in California, claims fas-
ciated shoots of his crop are more common among those plants which
first pierce the ground, especially after a cold winter. Fasciations
(Hus, 1906) were very frequent in wild and cultivated plants at
Berkeley, California, within a week after a heavy rainfall at a most
unusual time of the year (September, 1904). Other teratological phe-
nomena were also common. Dandelions from time to time appear on
lawns, along sidewalks and in fields in a fasciated state. Hus believes
these to be the result of mutilation.

M. T. Cook (Letter, 1912) writes me that Cuban planters
believe mutilated pine-apple stock will produce malformations (fasciation
included) and that his own observations tended to confirm this belief.
Richly nourished, but uninjured individuals of Weigelia (Ooebel, 1900)
sometimes produce fasciated shoots.

Experimental production. Sachs (1859) was one of the first
to produce fasciations experimentally. By cutting off the chief axis

White.

above the cotyledons in such plants as Phaseolus multiflorus and Vicia
faba, the axillary shoots frequently became fasciated. Double leaves
and changes in phyllotaxy appear not infrequently on individuals thus
treated. Lopriore (1904), following Sachs' methods, cut off the root
tips of seedlings of Vicia faba and obtained fasciated roots in a large
number of cases. Goebel (1: 1900, p. 190) calls attention to the
production of "fasciations artificially by causing the 'sap' to flow
rapidly and with great intensity into a lateral bud which otherwise
would only obtain a small part of it". This is his explanation for the
common occurrence of fasciated suckers and stool shoots. An intentional
slight injury of the growing tip produced fasciation in Ibervillea sonorae
(Knox 1907). By crushing young stems of Viola tricolor, var. maxima,
fasciated shoots were produced (Blaringhem, 1904—5). By removing
the main stem of Agroslemma githago (de Vries M. T. 2, p. 501) just
above the cotyledons, the axillary buds, which as a rule do not develop,
grew out under this treatment, and frequently became fasciated.
Mutilation of the main stem and branches of Barkhausia taraxaci folia,
induced more or less fasciation in the branches and inflorescences of
this plant, according to Lamarliere (1899).

Hus (1906) was able to induce fasciation by the use of the
following method: 'Plants, just previous to flowering time were subjected
to the environments described below':

"About the time of the appearance of the first flower, the
plant is kept as dry as possible, only enough water being given
to prevent wilting. As a result, the flowering period will be
comparatively short, and in an indeterminate inflorescence, the
buds near the end of the spike remain undeveloped. If at this
time, the plants are abundantly irrigated daily, occasionally with
manure water, numerous fasciations will make their appearance.
But it must be remembered that this result is usually reached
only with plants which throughout their existence have been
well nourished and well cared for generally. For no apparent
reason, one plant will fasciate; while the next one, belonging
to the same species, remains normal."
During 1905 fasciations were obtained by this method in Antirrhinum
majus, Actinomeris squarrosa, Solanum lycopersicum "Magnus", Lythrum
virgatuma, Oenother Lamar eMana and Collomia grandiflora. Experiments
with Solanum pseudo- capsicum, Capsicum annuum, Solanum nigrum
and Abutilon avicennae yielded no fasciations.

Studies of Teratological Phenomena.

Fasciations (perhaps produced in this manner) of Solarium lyco-
persicum, Antirrhinum majus, Echeveria glauca and others have been
propagated by cuttings at the Missouri Botanic Garden (Hus 1908).

Eeed (1912) induced fasciations in seedlings of Phaseolus multi-
florus through the removal of the plumule when it was about an inch
Jong. Shoots were thus caused to develop from the axillary cotyledon
buds, and many of them showed fasciation. The removal of the apical
buds from these shoots caused still more fasciated and twisted structures
to delelop. Through this treatment, the hypocotyl also often became
fasciated. Yicia faba and Pisum sativum under the same treatment as
P. multiflorus produced only a few slightly fasciated structures.

Epigeal types such a Phaseolus vulgaris, Lupinus douglasii, Ricinus
communis and Cucurbita pepo have fleshy cotyledons, and hence a
large supply of reserve food. When subjected to the same treatment
as P. multiflorus, no fasciations were produced. Epigeal-type seedlings
were given a plentiful supply of nitrogenous manures, and some were
mutilated. Although the checks were vigorous in their growth, the
mutilated individuals did not produce any fasciated structures.

Daniel (1904) induced fasciation in the common European pear
by a method of pruning called "a onglet complet", which consists
of removing all the buds, the terminal included, from a branch, and
allowing this branch to remain on the tree.

D. Discussion and summary of Parts B and C.

All characters, whether somatic or otherwise, may be regarded as
the resulting expressions, in an organism, of stages in the development
of a factor (gene, or germinal unit) plus the modifications of this
expression brought about through the presence in the organism of other
factors, and through the action of external environmental conditions such
as soil, climate, insect depredations, etc.

Fasciation, on the basis of this conception of a character, may be
caused by many diverse and unrelated combinations of internal factors
and external conditions. Hence, to speak of it as latent, when it does
not exist as an observable entity, is an absurdity. If one holds to such
a vague form of interpreting certain facts, the furniture dealer is justified
in speaking of chairs being latent in mahogany trees or of the latency
of office desks in oak trees. The fore-going data show us beyond all
doubt that many factors, both internal and external, are responsible for

White.

fasciation. Mechanical injury, sudden arrest of growth, insect and
fungous depredations, injury due to climatic factors (frosts, increased
humidity), poor seed associated with richly fertilized soils and intensive
cultivation are probably only a very few of these numerous environ-
mental conditions. But the internal factor or factors are just as
important as the external ones, as shown by the results obtained in
attempting to produce this phenomenon experimentally. All plants, even
of the same species or of even closer taxonomic affinity, when subjected
to apparently identical environmental conditions do not respond in the
same degree or in the same manner. This is amply illustrated by the
results obtained by Hus, Reed and others. Cucurbito pepo seedlings
did not become fasciated with Reed's methods, but fasciated plants of
this species are recorded by de Vries. Fasciated races never bred
true with de Vries, although he tested out many fasciated individuals
belonging to numerous diverse species, by growing large numbers of
their progeny. Even the old established horticultural varieties of
cockscomb proved to be inconstant in his cultures, as far as the
character fasciation was concerned. On the other hand. Goebel found
the cockscomb to be absolutely constant in his cultures even when they
were grown in sterile sand. Fasciation is also known to be an absolutely
constant character in several races of plants, as the hundreds of
individuals comprising several generations that have been raised under
controlled conditions, testify. Normal strains of these same fasciated
races are also common, and in the case of peas, have bred true to
absence of fasciation for at least a quarter of a century. Are we to
believe that the character fasciation is latent in these normal strains,
or are the facts more clearly expressed by looking upon it as absent?
From the standpoint of genetics, the latter interpretation is by far the
more preferable, because it more clearly expresses the facts in the case,
as we know them. There are undoubtedly constant hereditary races of
fasciated plants. — races, in which every plant derived from the seed
of a self- fertilized fasciated parent, is fasciated. if both parent and
progeny are grown under identical environments. There are also constant
normal strains of these same species, which breed true to normalness,
when all are grown under the same environment as their fasciated
relatives. There are still other races of plants in which fasciated
individuals are common and the percentage of these abnormals vary
greatly as in the case of the Crepis with which de Tries experimented.
Progeny grown from such fasciated races under the same conditions as

Studies of Teratological Phenomena.

the constant normal and constant fasciated races continue to remain
inconstant, even though they may all be grown from seed of a single
self-fertilized plant. Such races are usually plants which are naturally cross-
fertilized and hence may be heterozygous in many internal factors. The
segregation of these factors, even in the germ-cells of a single selfed plant,
may produce a very diverse progeny. As the development of one factor
may be hindered or helped by the presence or the absence of others, one
may conclude that part of this eversporting condition is due to the
segregation that takes place in each generation. Further, more than
one primary factor may be involved in producing fasciation in some
races of plants. In attempting to explain the eversporting character
of de Vries' fasciated races, one must not forget the prevalence of the
various external factors, which especially in an old settled country, are
always on hand to commit depredations. When these external factors
operate in the form of an insect, they are very hard to trace, as has
been shown by Knox and others. My own experience with Oenotheras
has shown me how hard it is to guard against such factors. I grew
200 seedlings from two very fasciated wild plants of Oenothera biennis,
which were obtained for me through the kindness of S. M. Blake of
the Gray Herbarium. They were grown under conditions generally held
to be favorable to the development of fasciation, and on ground only a
few rods from which in former years, many fasciated Oenotheras had
been observed. Of the 200 seedlings only one was fasciated and that
only to the extent of a small twig. It is obvious that in this case,
fasciation was not hereditary in the sense in which we ordinarily use
the term, yet had I previously believed such characters to be inherited in
this inconstant manner, I would have never considered the true cause
— insect mutilation.

Summary of Parts B and C.

1. The character fasciation is widely distributed in the plant world,
both in wild and cultivated plants. Sufficient data have not been
collected to prove that it is absent from any taxonomic group.

2. Certain ecological conditions are favorable to its development,
but these conditions are not necessarily essential.

3. The character fasciation may occur in almost any part of the
plant. Morphologically, it appears to be an enlargement of a single
growing point, so that considering the fasciated plant as a whole, the
amount of tissue is greatly increased over that of its normal relatives.

White.

4. Very numerous, diverse internal and external factors operate
singly or in combination to develop fasciation. If these factors are
largely internal, and the race is naturally cleistogamous, the character
is generally hereditary and constant. If the factors are largely external
(insect depredation, mutilation, etc.) the character fasciation is neither
hereditary nor constant. If the fasciated race, to begin with, is not
genotypically homozygous, as is not the case when the individuals of a
species are naturally cross-fertilized, the character fasciation may be
hereditary, but present in only part of the progeny, owing to the
segregation of the primary internal factors for fasciation or to the
segregation of numerous other internal factors which may modify in
some manner the expression of those factors especially concerned in the
production of fasciation.

E. Review of previous Mendelian studies.

1. Pisum sativum umbellatum Mill, is the only race of fasciated
plants which has been fully tested as to the discontinuous nature of its
inheritance in crosses with the normal race (Fig. 8). The fasciated character
of this plant was one of the original seven Mendelian character -pairs.
Mendel (Bateson 1909) found that crosses of fasciated X normal in
Fi gave complete dominance of the normal condition. In F2 the ratio
of abnormal to normal was 3*14:1, Bateson and Punnett (Bateson,
1909, p. 25) repeated tins experiment, but secured intermediates in F2.
Fasciated (terminal inflorescences) . . . 207
Normal (axial „ ) ... 651

Total 858

Mendel's experiment was again repeated by Lock and later by
Darbishire. Lock hesitatingly confirms Mendel's results. The hesitancy
is caused by the variation in the fasciated character. Normal X fas-
ciated in F2 gave Lock (1908) approximately a 3 : 1 ratio, but many
of the fasciated F2 segregates expressed their anomalous character in
a much slighter degree than the fasciated grandparent ("Irish Mummy").
Some of this modification in expression was ascribed to crowding and
to other unfavorable growth conditions, since the fasciated F2 segregates
"had to compete with thrice their number of normal sister plants, but
it did not seem likely that this would account for the whole difference".
•Seeds of very slightly fasciated plants were sown and the resulting F3's

Studies of Teratological Phenomena.

grown under optimum conditions. All the plants thus produced were
in every case fully fasciated, almost, if not quite as much as the original
grandparent.' The slight variability of the character was therefore
ascribed by Lock to environmental influences. Fi plants grown at the
Bussey Institution from seed ("Irish
Mummy" X "Chinese Native") fur-
nished by Darbishire, gave ab-
solute dominance of the normal
condition.

2. Zea Mays L. De Vries,
East and Hayes, Emerson and
Hus have all experimented with
races of maize which produce fas-
ciated ears.

De Vries (1894) finds this cha-
racter to belong to the "eversporting"
class. Cultures that were grown by
him contained 40 per cent abnormal
plants. Hus and Murdock (1911)
secured results similar to those ob-
tained by de Vries.

East and Hayes (1911) found
an ear of this fasciated type in a
culture of field corn which had been
selfed for three generations. The
seed was grown aud 34 abnormal
and 12 normal- eared plants were Fig- 8. Pisum sativum umbellatum.
produced. Another fasciated ear (After Gerarde.)

appeared in the F2 generation of a

cross between two normal strains, one of which had been recorded
as throwing abnormal -eared plants. This ear produced 62 abnormals:
23 normals. The normals appeared to breed true, and the abnormal
condition is regarded by them to be dominant. The character itself
fluctuated between very abnormal and (superficially) almost normal
states.

The most extensive investigations on the inheritance of fasciation
in maize have been made by Emerson (1912). In his cultures, the
degree of fasciation varies much even between the different ears of a
single plant, some ears being very broad-tipped, while others are only

3. Pisum umbellatum.
Tufted or Scottish Pease.

White.

Fig. 9. Fasciation in Zea mays, illustrating constancy of the fasciated character in this
strain. The horizontal ear is from the parent plant of the plants producing the verti-
cally placed specimens. (Photograph by Emerson.)

Studies of Teratological Phenomena.

slightly flattened. The opinion is expressed that different degrees of
fasciation may be inherited.

One of Emerson's corn families shows an extreme type of fas-
ciation, not present in any of the others (Fig. 9). 'From a 1910 family
of popcorn that contained both normal and abnormal ears, four fasciated
ears were selected and used as parents of 1911 families.' The results
obtained were as follows:

No.
of Plants
grown

Character of Ears.

1911

Ear 1
Ear 2

Ear 3
Ear 4

18
46

- 28

All strongly fasciated.

All strongly fasciated.
f 12 more or less fasciated.
1 3 perfectly normal.

[Strongly abnormal, normal and all intergrades
t making classification impossible.

From one of the first two families, I examined and collected ears,
and found very little variation in the expression of the anomalous
character. Another family grown from a fasciated ear gave a proportion
of 32 plants with more or less fasciated ears, and 35 plants with
apparently normal ears, though some of the latter may have been
slightly flattened. The parent of the above family was also crossed
with an 8 -rowed dent corn plant. The Fi generation contained 63
plants, all producing perfectly normal ears. In other crosses (1912)
between fasciated and normal races, the Fi was also perfectly normal.
About 25 per cent fasciated plants occurred in some F2 families, while
in other crosses even less than 25 per cent were fasciated. Emerson
believes soil and climatic conditions to have considerable influence on
the expression of fasciation in maize. In some strains, he thinks
perhaps two Mendelian factors are involved. The suggestion is also
made that 'interaction between a single fasciation factor' and the
diverse characters present in the different strains, may provide just as
good an interpretation for the complexities in the results as the postu-
lation of more than one factor.

Induktive Abstammungs- und Vererbungslehre. XVI. 6

White.

F. Special study of fasciation in Nicotiana.

1. Problems.

Chief among the problems involved in this particular study, is the
effect of diverse and unrelated genetic factors in their ontogenetic
expression upon that of the fasciation factor. Owing to the historical
data extant concerning the origin of this fasciated race, one is almost
justified in saying that in this particular case, through comparison with
its normal parent, and through crossing, an isolation of a factor has
been accomplished. If to some, this statement is not justified by the
data which is to follow, I feel that at least, I have found ideal material
to serve as a standard with which to compare the variation in expression
which takes place when this race is crossed with other races, and thus
better appreciate the true nature of the phenomena of dominance. Other
problems more or less associated with this main problem are : the nature
and causes of dominance, the ability of selection to modify a unit factor,
the relation between fasciation and environment, the nature of Mendelian
segregation when two abnormal hereditary characters are combined
through crossing, the appearance of mutations in controlled cultures,
and the fidelity with which pure homozygous F2 segregates breed true
in later generations.

It is realized, however, that the data secured on the problems of
selection and the modification of unit factors, are too few to be of
more than suggestive value.

2. Materials and methods.

The material upon which the study is largely based is a fasciated
race of Nicotiana tabaeum. In connection with the hybridization
experiments, cultivated and wild species and varieties of Nicotiana,
which have been tested in controlled cultures for at least two years
were used. These types were obtained by Dr. East from Prof. W. L.
Setchell, from Mr. J. S. Dewey, from Prof. A. Splendore and Prof.
0. Comes through the kind offices of Mr. D. Fairchild. To all of
these gentlemen the writer wishes to express his grateful thanks.
Nicotiana is an excellent genus on which to conduct investigations, and
of the numerous species it contains, Nicotiana tabaeum is one of the
most favorable from a genetic standpoint, because of its large number
of distinct subspecies and varieties, the majority of which are practically

Studies of Teratological Phenomena.

cleistogamous, hence, yielding races which are at once almost natural
pure lines. Varieties and species in this genus in many cases are fertile
inter se, and the seed produced by one flower furnishes an abundance
of progeny, thus helping to eliminate the arduous technique of making
crosses. The species and varieties which were used in this study are
described by number.

a) Description of material-species and varieties.

300 — 309 Nicoiiana tabacum fasciata (Fig. 10). Mutant derived from
"Cuban" variety of N. tabacum, J. S. Dewey, 1907. This race was obtained
from selfed seed of a mutant found growing in a field of Cuban tobacco
in the district of Partidos, near the town of Alquiza, Cuba in 1907.
J. S. Dewey, who was connected with the company on whose plantation
the discovery was made, describes the original plant as follows: 'Stem
fasciated; leaves 152, not over 8 cm. long when dry, flowers abnormal,
very little seed produced.' Owing to the cleistogamous nature of
N. tabacum, the strain from which the mutant arose was probably a
natural pure line, the characters of which were largely homozygous.
The isolation of pure true-breeding lines ot Cuban tobacco from a mixed
population by Has selb ring (1912) seems to substantiate such a claim.
And if this be true, hybridization had no part whatever in producing
the mutation. As only one fasciated plant occurred in the field, so far
as is known, and as this plant was homozygous and bred true upon
selfing for the abnormal character, the actual place in ontogeny at
which the change from the normal to the abnormal took place must
have been shortly after fertilization. If it had occurred later in ontogeny,
the fasciated character would have appeared first as a bud -sport. If
it had taken place before fertilization as the result of a disruption in
one of the cells involved in the maturation of the egg or sperm, a
double mutation would have been necessary (i. e., a similar single
mutation in both egg and sperm ancestry) in order to account for the
homozygous condition of the original mutant. In the latter case, the
element of chance is so great, that it is very improbable, even had
such mutations occurred, that it would have been possible for them to
unite, and had all the germ-cells of a whole plant changed, more than
one abnormality should have appeared. One may say, of course, that
on its first appearance, it was heterozygous and that the single plant
found was a representative of an F2 or possibly of a backcross. But

White.

it is hardly probable in a crop so closely attended and scrutinized as
is tobacco, and with so prominent a change in character as the original
mutant showed, that additional abnormal plants would have escaped
unnoted. The heterozygote (Aa) produces plenty of seed under even

Fig. 10. Fasciated Nicoliana plants growing under shade
in Connecticut.

unfavorable conditions, so its scarcity would not account for the presence
of only one plant. So many mutations are coupled with maturation
phenomena at the present time that it seems worth while to call
attention to cases that one may feel tolerably certain did not arise as
a consequence of reduction or maturation phenomena disturbances.

The morphological differences between the normal Cuban (402) and
the mutant (300 — 309) which constitute the somatic character called

Studies of Teratological Phenomena.

fasciation, will be described in detail, beginning with the grosser structures
such as stem and inflorescence, and concluding with a description of
the changes brought about in the reproductive organs. Special care
will be observed in the description of the leaves and floral parts
because of the variability in number produced by the presence of this
character.

Seedlings. Generally normal; tricotyls rare.

Stem. Cylindrical base, gradually developing the characteristic flat,
ribbon -shaped, fasciated condition. Grooved or ribbed by fine vascular
strands. Linear width, 1*25 — 5*5 cm. Some stems more flattened than
others. Fasciated part of stem not the same throughout its whole
extent. Variable, often slightly curved owing to irregularities of growth.
Usually unbranched except for the cluster of small twigs constituting
the inflorescence. Pith an ellipse in cross section. Anatomical pre-
parations give no support to the "concrescence" theory.

Leaves. Phyllotaxy very irregular. Double leaves and leaves
with broadened apices not uncommon. Great increase in number as
compared with normal parent race. Many normal-sized leaves and many
smaller than normal (402). Variation in number of leaves correlated
with size and character of main axis. Variation in number per plant
ranges between 28 — 152 as taken from records of over 200 plants
grown under five different environments and during five seasons. Normal
as regards form. (See Table 2.)

Inflorescence. Bifurcate, multiradiate, rarely "annular" or
funnel-shaped, often single main axis, abruptly terminated by a number
of small twigs. Flower -bearing twigs small, densely clustered into
'witch-broom'-like bunches or irregularly distributed along the sides and
apex of the stem.

The floral structures were deformed in a large number of ways,
the most prominent being the increase in number of parts to each whorl
per flower. This numerical increase is not of a constant nature, in the
sense of a variation from one definite number to another, a statement
also true of the change in phyllotaxy and leaf number. Nor does it
at all substantiate deVries' notion of a duplication in number such
as is said to occur in Geranium molle fasciatum. In a figurative way,
one may compare the variability of Nicotiana tabacum fasciata to that
of an arc made by a pendulum, provided there was a force behind the
device to change the rate of its swing and the size of its arc when
affected by things external to itself. Tables 3 and 4 give a better idea

White.

than mere description, of the extent and nature of this variability, the
former as regards the variability of the race and the latter as regards
the range per individual plant. By inspection of Table 3, the average
number of parts per whorl of a flower is seen to show a progressive
increase in the expression of the fasciation factor. This may be expressed
in tabular form by subtracting the normal number of parts per whorl
(5 for the first three and 2 for the gynoecium) from the average for
the number of parts per abnormal flower. In order to compare the
first three whorls with the gynoecium, their differences should be divided
by 2' 5* as there are 25 times as many petals, sepals and stamens
to a flower as ovary- locules. The results for families 301 — 1 and
303—1 are thus:

Table C.

301—1
Average No. of
floral parts above
normal

303—1
Average No. of
floral parts above
normal

Sepals ....

1-74 +

0'69 +

2-03 +

0-80 +

Petals ....

2*65 +

1*06 +

3'24-h

1*29 +

Stamens . . .

2-70 +

1-08 +

3'47 +

1-39 +

O.-locules . .

1-93 +

2-86 +

2'86 +

The calyx is thus seen to be the least, and the gynoecium the most
affected. The latter is almost three times as abnormal as the former
in 301 — 1, and more than three times in the case of 303—1. This
progression in the manifestation of abnormal condition is in accordance
with the observations on other parts of the plant. The seedlings appear
to be normal; the first few leaves are not deranged as to phyllotaxy,
and the whole stem remains normal, even in the most fasciated specimens
for a foot above ground. The linear expansion increases in extent and
the leaves in number as the plant approaches maturity. At maturity,
the apex of the stem shows the greatest linear expansion, and sometimes
becomes so abnormal that the whole inflorescence is partially inhibited
in its development. The greatest alteration in phyllotaxy and the
largest increase in number of leaves is characteristic of this terminal
portion of the main axis. Owing to the fluctuation in expression of
the factor, the stem may not always show the linear expansion through-
out its whole length, but in all cases it shows in the inflorescence.

Studies of Teratological Phenomena.

In addition to the increase in number of parts, the floral structures
are subject to many minor abnormalities. The first flowers to bloom
on a plant are much more abnormal than those appearing later. Because
of this, all data on the abnormal flowers of a family have been taken

a b
Fig. 11.

a Nicotiana 300 — 309 (fasciated).
b Nicotiana 402 (normal).

as nearly as possible at the time when all the plants were approximately
in the same blooming stage. These early flowers were often so split
(dialysis) and deformed as to lose all semblance of belonging to any
regular- flowered family such as the Solanaceae. The later flowers,
though generally possessing as many floral leaves as the earlier ones,
were usually as symmetrical as those of the normal 402 (Fig. 11).

White.

A detailed study of the various whorls disclosed many more
anomalous characters. The calyx, in addition to being irregular,
sometimes possessed a sepal attached to the outside of the regular whorl
(Fig. 12). Sepal lobes were irregular in size, and occasionally one would
occur with a slight reddish color on the tip (calycanthemy). Calyx and
corolla whorls not uncommonly were present as one spiral whorl
(„speiranthie"). Two flowers were sometimes fused and enclosed in a
continuous calyx (adhesion). Once or twice, a flower was found con-

Fig. 12. Stems and flowers of the fasciated
(300—309) and normal (402) races of N. tabacum.

sisting of only a corolla and a few stamens, growing on the side of
and fused with the main corolla (synanthy). Instead of an increase in
flower parts through a multiplication of whorls (pleiotaxy) the increase
takes place through a multiplication of the number of parts per whorl
(polyplryUy). All four whorls are increased in the number of their parts
in this manner. Polyphylly of the androecium increases the number of
stamens per single flower, the range of variability being 4 — 25.
Filaments are fused to each other (cohesion) and to the walls of the
corolla (adhesion). The anther-sacs are sometimes split at the end into

Studies of Teratological Phenomena.

two segments; in other cases an actual increase in number occurs and
a single stamen may have as high as six. Petalody and pistillody of
the stamens are rarely present. In the former case, the petals are
very slightly developed; in the latter, three or four rudimentary pistils
occur, developing from almost any point on the anther- sac. In an
examination of thousands of flowers of this race, I have found petalody,
pistillody and calycanthemy only in a dozen or so cases. Abortion of
pollen (contabescence) is common as well as various distorted conditions
of the anthers.

The pistil frequently was wholly or partly incapable of functioning,
owing to various forms of distortion, including proliferation, staminody,
pleiotaxy, and meiophylly of the style and ovary-locules. The increase
in number of locules ranged between 2 (rarely) and 21, the mode being
about 4. The style was often shortened and twisted. Ovary-locules
were so crowded at times, owing to polyphylly that many were abortive,
resulting in a much distorted capsule. From 2 — 4 pistils (pleiotaxy)
were often present in the same flower, sometimes all capable of
functioning; in other cases, all but one abortive. Sterility was present,
but in the majority of cases, examination of a mature capsule demon-
strated fertility to be almost perfect.

Cytology. In a preliminary paper, the normal conditions were
briefly described for both this abnormal strain and the normal (402).
The chromosome number was 48, reduced in the germ -cells to 24.
Cytological variations in the normal (402) were rare, either as to
chromosomes and their number or in other structures. Many anthers
' of the abnormal race when examined cytologically were entirely normal
in all their maturation phases. Others showed evidences of almost total
sterility through premature breaking down of the archesporial tissue,
while still others were only partially sterile. Abnormal variation in
the rate of progress of maturation stages was often characteristic of
the abnormal anthers. Atrophy and disintegration following physiological
abnormalities causes the not infrequent appearance of only a very few
mature pollen grains in the mature anther. This breaking down of the
pollen mother-cells began in the early prophases of the first division,
and persisted as late as the prophases in the second division. Ab-
normalities were not as common in the second reduction division as in
the first. Deformed nuclei were common in the first maturation division.
The nucleoli and chromatin were not infrequently clumped together as
though overheating on the slide had taken place. Nuclear fragmentation

White.

was common. Great variation in the staining properties of the
preparation was noticeable, this probably resulting from physiological
abnormalities. The chromosomes of a mother-cell were either increased
in number by abnormal division or nuclei divided and never separated.
In one case 51 chromosomes were counted in a reduction phase, and
from the manner of their occurrence, it would seem there was no other
interpretation than of increase in number through a division of only
part of the chromosomes of this cell. The cell was in a state of
disintegration. Other cases occurred in which there were 30 where
only 24 should have been counted. Irregular divisions and lagging
chromosomes were not infrequent. The chromosomes of Nicotiana are
small, though very distinct in the maturation stages, but in cases where
abnormal conditions prevailed, one could not always be certain they
were counting heterotypic or homotypic chromosomes or both, owing to
premature division of some of the heterotypic chromosomes. Increases
in number were rare, but many of the other abnormalities mentioned
were common. Supernumerary pollen grains such as occur in Hemerocallis
were never found. Divisions of chromosomes in somatic cells on account
of their size, were studied with difficulty, and gave no data. Cytological
observations on maturation in the gynoecium were only superficial as
compared with those of the anthers. No special abnormalities in the
reduction divisions were noted in the few sections examined, but a
small percentage of the embryo-sacs appeared to be abortive.

From data taken in connection with the floral leaf counts, I
should judge the contabescent anthers to be from 4 — 5 per cent, but
such data probably gives too low an estimate, because one cannot be
always certain that an anther, externally normal, represents the same
internal state.

Summarizing, one may say that although these abnormal cellular
conditions are strictly inherited, their morphological aspects closely
resemble the cytological changes produced in plants through external
stimuli, such as fungi, insects and chemicals, after the manner in
which these have been described by Molliard (1897), and others.
Gregory (1905) described similar phenomena in the case of the abortive
anthers of sweet peas, though in this case as in Nicotiana, the pheno-
menon was of strictly hereditary nature.

Fasciation in Nicotiana is comparatively rare, as somewhat extensive
search through the literature brought to light only a few cases of floral
fasciation in Nicotiana affinis (alata) (deVries).

Studies of Teratological Phenomena.

The relation of this race to different environmental conditions
and the nature of the variability in expression of the fasciation factor
will be further discussed under "selection and environment".

402. Nicotiana tabacum L., var. "Cuban" (Fig. 13). (13—29 U. S.
Dept. of Agr.) Grown in controlled cultures for at least three years.

Fig. 13. Enlarged view of the inflorescences of Nicotiana tabacum, races 300 — 309, 402.
(Photographs from F2 segregates of 304 X 402.)

Remarkably constant in all its characters. As contrasted with 300—309,
it is normal in all the organs altered in that race by the fasciated
character. A possible exception to this generality is found in the rare
occurrence of an extra sepal or petal. Stem cylindrical, number of
leaves per plant 18 — 24 with mode of 20, ave. 19*65, based on counts
made by East and Hayes on 124 plants grown at Bloomfield, Conn,
in 1911; inflorescence branched, flowers pink, three whorls 5-lobed,
gynoecium 2-loculed. Number of parts to androecium and gynoecium
practically constant. Plants in our cultures, uniform in height and
other gross morphological features. Same variety as that from which

White.

300 — 309 was derived, but obtained from a different source. Maturation
phenomena in the anthers normal. No contabescence.

353. N. tabacum, var. fruticosa Hook. fil. U. S. Dept. of Agriculture,
1908. Orig. from Portici, Italy. General description given in Comes'
Mon. (1899, p. 8) Sp. 1, var. 1. Comes' var. fruticosa not that of Hook,
fil., according to Setchell (1912). Inbred for three years, constant
for characters noted in crossing experiments. Low, shrubby, profusely
branched plants, height ave. (1912) from 9 plants, 14*55+ dcm. Lvs.
petioled, non-auriculate, leaf count made on main axis from 9 individuals
(1912), ave. 12*33 4- leaves per plant. Range in variation 10 — 14, but
method was very unsatisfactory and results are only approximate.
Flowers normal, occasionally 6-sepaled or 6-petaled, 100 flowers examined
(1912) were all perfectly normal, slightly darker pink than 300 — 309,
deeply lobed corolla, petals acuminate, slender throat and tube. Sepals
long, acute, and reflexed at tips.

373. N. tabacum, var. havanensis, (Lag.) Comes' Mon. (p. 16),
angustifoliae, vern. Loemodjang. Comes' 1908 (See S. & P. Int. Inv.
No. 14, p. 40) U. S. Dept. of Agr. 22164. When first grown, variable
in leaf and flower characters. Selected by East for constancy in corolla
shape and leaf characters. Race from selfed seed, constant for twTo
years in characters here noted. Hght. 17*54 — 19*06 dcm., ave. 18*3 dcm-
in 1912. Unbranched main axis, lvs. in number 18 — 21, ave. (1912)
19*33+, auriculate, sessile; inflorescence spreading, flowers dark red,
inflated throat, corolla pentagonal, varying toward subrotund lobing,
normal as to number of parts, no 6-sepaled or 6-petaled flowers being
noted. Fertile. See Comes7 Mon. (p. 16) for general description.

396. N. tabacum, var. fruticosa hybridae, fruticosa X macrophylla
purpurea. Hort. N. calyciflora Caille. Comes' Mon. (p. 10). From
Portici, Italy. Same as Nicotiana tabacum var. calycina of Setchell
(1912, p. 6) (Fig. 14). Race from inbred seed, constant for two years or
more, except in characters noted as otherwise (See Setchell on constancy
of this variety). Low, shrubby, profusely branched, 353 -type of plant.
Hght. constant in inbred stock, except for the appearance of a single
nana plant in 1912. Ave. hght. from 24 (396 — 1) individuals,
13*84+ dcm. Lvs. sessile, auriculate, ave. no. per plant 11, with
range of 9 — 12 [count includes 24 plants determined from observations
on main axis (unsatisfactory)]. Inflorescence normal; flowers reddish
pink, subject to splitting of corolla tube and other morphological
irregularities. Sepals, petals and stamens fluctuate between 5 and 6

Studies of Teratological Phenomena.

per flower. Filaments not infrequently fused to corolla tube (adhesion).
Three or four of the sepals are always roseate colored in whole or in
part (calycanthemy). This character is variable and the anomalous
sepals usually are longer than the non- colored. The peculiar crinkled
effect shown in Fig. 6 is also characteristic and constant when even
the calycanthemy does not appear. Gynoecium normal. Fertility
100 per cent, and especially noticeable. Corolla is usually irregular.
The constancy of the anomalous character is perhaps better shown in
Table 5. Table 5 gives an indivi-
dual record of 25 flowers of plant
396—1 (1911), together with six of
its progeny grown from inbred seed.
24 other 396 — 1 progeny were grown
in the same row and their range of
variability was approximately the
same as the six plants on which the
table is based. Masters (p. 384—85)
says teratological coloration of sepals
is especially common in gamose-
palous flowers and cites numerous
cases.

Calycanthemy usually means Fig. 14. N. calyciflora, Caille.

simply a change in color, but may Note the calycanthemous sepals,

not infrequently be accompanied by

structural changes, and this is said to be especially true in cases where
displacement of organs has occurred. In the 396, as grown in our cul-
tures, structural changes were always present in some form and degree.
Splitting of the corolla tube and the formation of and adherence of an
extra petal were especially common. Coloration in at least one sepal
and usually three was characteristic. The dwarf (mutant?) mentioned
differed from the race only in height, not in number of leaves nor in
type of flowers.

Other Species.

324. N. Ugelovii Watson. U. C. Bot. Garden 1909. Answers to
Comes' (p. 43) description , except in floral characters. Sepals , petals
and stamens vary between 5 and 7, per flower, 6 being very common.
The race breeds nearly true to a 3-loculed ovary. Sp. 25. (See Setchell
1912, p. 25).

White.

327. N. glutinosa L. U. C. Bot. Garden 1909, Comes' Mon.
(p. 24) Sp. 3. Constant from selfed seed for 3 years.

331. N. paniculata L. U. C. Bot. Garden 1909, Comes' Mon.
(p. 25), Sp. 4. Constant from selfed seed for 3 years.

332. N. sylvestris Speg. et nob. (Nova, sp.) U. C. Bot. Garden
1909. Comes' Mon. (pp. 34—35). Sp. 19. Constant.

b) Number of plants grown.

This investigation of the inheritance of fasciation in species of
Nicotiana is based on data from studies of about 5,000 plants, grown
over a period of five years, and under several distinct physiological and
geographical environments. A nearly complete list of the different
species, races, families and hybrids, together with the exact number of
each grown, the year and the environment are given in Table 1.

c) Methods.

All plants used in these experiments were grown as nearly as
practicable under the same external environmental conditions. Plants
for comparative study were often grown side by side. After each
operation in making crosses, all instruments were carefully cleansed in
95 per cent alcohol. Pollen was used only from unopened flowers.
The technical work was always personally looked after. The methods
used in the cytological phase of this investigation are described in an
earlier paper (White 1913).

Tabulation. Data were collected on each plant and tabulated
separately. The character on which most of the studies were made
were: — extent of fasciation in the main axis, number of leaves per
plant, number of floral parts per whorl. 25 flowers from each plant
were taken and the number of parts per whorl for each flower was
recorded separately, and in such a manner that all the whorls per
single flower remained identifiable and their correlation could be shown.
Minor abnormalities of all kinds, such as calycanthemy, pistillody of
the stamen, united filaments (cohesion), abnormally distorted anthers,
petalody of the stamen, deformed styles and stigmas, cohesion or fission
in leaves, were recorded for each flower and plant. In presenting in
tabular form, the mass of data thus accumulated, two types of frequency
tables are used, each of which shows the variability of the fasciated
character expressed in terms of numerical-plant-organ alterations, as for
example, one flower may have 8 sepals, 10 petals, 12 stamens, and

Studies of Teratological Phenomena.

5 ovary -locules, while another may be altered from the normal 402
condition by having 10 sepals, 10 petals, 12 stamens and 4 ovary-
locules.

One type, such as Table 4, gives the frequency distribution of
floral parts (sepals, petals, stamens, etc.) per flower, and shows the
fluctuation of the character fasciation per plant in a whole family.
The second type of table such as No. 3 disregards the individual plant
entirely, and gives the frequency distribution of the number of parts
per flower and leaves per plant for the family as a whole. The first
type is especially useful in comparing the individual inconstancy of the
character in two different races. The second is only valuable as a
basis for the comparison of the average variation of different races.
In order to show the variation in the organs of a single plant, when
affected by fasciation, tables of records of typical individuals have been
used (See White 1913, pp. 212—13). The race 402 is at all times
accepted as the normal, and a basis for comparison as to what con-
stitutes an abnormal Nicotiana tabacum deviation.

3. Fasciation and Environment.

In order to study accurately the manner of the inheritance of a
"somatic" character, it is first necessary to thoroughly understand its
reaction toward its particular environment, and this is especially true
of such a character as fasciation, as has been repeatedly emphasized
by de Vries, Knight (1822) and a host of seedsmen and florists.
De Yries found the influence of conditions surrounding the parent seed
plant to be a factor that must always be taken into consideration in
studying anomalies. In his work, plump seeds gave the plant a better
start in life and as nutriment and good care are very essential to
producing large fasciations, such seeds, he says, should be selected.
Seed from branches favorably situated should produce more anomalies
than seed from 'poorer' flowers and weaker branches (considering the
latter to be atavistic or more normal structures), but in his experimental
work, this expectation bore but little fruit.

De Vries has very often called attention to the relation between
plant vigor and the production of anomalous variations, — "the stronger
a branch is, the more liable it is to flatten out". Biennials and
perennials when allowed to fasciate the first year gave but very small
expressions, and in some cases, a heritable race would show no sign
of its fasciated character. On the other hand, high percentages of

White.

individuals with beautiful comb-like expansions were secured the second
year. Time of sowing, according to de Yries (1909 — 10), (M. T. 2,
p. 498) is also a factor in the production of fasciated individuals. In
Holland, sowings of Crepis in April and May gave 30 — 40 per cent
fasciations, sowings at the end of July 20 per cent, and those made in
September, none at all. The same results were obtained with Taraxacum
officinale. Apparently early sowing gives the plant time to make a
strong rosette before winter stops growth. Crowding, pruDing, and
nutrition are important factors. Crowding is said to lessen the percentage
of fasciated individuals, and increase the atavists in a hereditary race.
Pruning diverts the sap rapidly into lateral branches, and this, according
to de Yries and Goebel, arouses the latent character to somatic
expression. Nutriment is very important as evidenced by the remarks
on crowding, selection of seeds, etc. For the reason that fasciated
branches are often somewhat weakened by growth expansion, de Yries
thinks atavistic (normal) branches may perhaps produce the best seed
for continuing the race. By the same method of reasoning, absence
of fasciated branches on a fasciated main stem, which is very common,
are explained. Briefly then, in order to favor the expression of such
characters as fasciation, even though the character is said to be strictly
hereditary in a race, it is necessary, according to de Yries, to take
into consideration, environment of seed parents, character of seed (weight,
plumpness, etc.), climatic conditions, time of sowing, crowding, pruning,
and other methods of diverting sap or increasing the food supply
abruptly, and nutriment (soil, water, fertilizer, etc.).

Growing in sterile sand decreases, while richly manured soil
increases the percentage of fasciations in a hereditary race, and the
finest specimens are produced by pot culture, rich soil and subsequent
transplantation (de Yries, 1899, M. T. 2, p. 501).

Nicotiana tabacum fasciata furnishes especially favorable material
on which to test out some of these ideas and throw further light on
the interpretation of the facts, for in this race, there are so many
structures on which reaction-phenomena can be noted. The experimental
cultures of this race have nearly always been grown beside the normal
402 race, and it is understood that both races (300—309 and 402)
under all the environments tested, constantly show the differences noted
in the preceding descriptions.

Commercial growers of tobacco fan their seed, sowing only the
heaviest. In my work, unselected seed was planted in seed pans,

Studies of Teratological Phenomena.

containing moderately rich soil. Often these pans were crowded with
young plants, but they did not remain in such quarters long enough to
become stunted. In pricking out plants, naturally and unconsciously,
the best seedlings were selected, though this was not always the case,
especially when seed was scarce. Variation between the plants became
greatest after they had been pricked off into flats, and when final
transplantation time arrived, there were some more or less stunted
individuals, but all were usually planted. The normals (402) were
always subjected (in my own work) to the same treatment as the
fasciated race. Plants were grown in a variety of environments, and
in order to show the constancy of the race under these environments,
Table 2 was constructed. The number of leaves per plant is extremely
variable, as evidenced repeatedly in a leaf count of the progeny of a
single selfed plant, when all had been grown under the same conditions.
124 plants from selfed seed of a single 402 plant, grown under shade
in Connecticut in 1911, gave an average of 19 '65 leaves per plant,
with a range of variability between 14 and 24, and a mode of 20. In
1908, under about the same conditions, 99 plants from selfed seed of
the original mutant were grown, and leaf counts made by J. S. Dewey,
gave an average of 69*7 leaves per individual, with a range of
variability between 30 and 133 and a mode of 57. Included in Table 3
are abnormal segregates, but these as far as I can judge, are
indistinguishable from the pure abnormal race. The range of variability
in the number of leaves in different years is well shown in Table 2.
301 — 309 consisted of 148 progeny grown at Bloomfield, Conn., from
selfed seed of nine of Dewey's 1908 plants. 301 — 1 and 303 — 1 were
selections from the 1909 cultures, the progeny of which were grown
under field conditions at the Bussey Institution in 1910. X is the
progeny of a single selfed plant of Dewey sport (genealogy lost) grown
in the Bussey greenhouse, the winter of 1909 — 1910. The 1911
selections were all grown under the same environment at the Bussey
Institution. The same is true of the 1912 cultures, except that the
five 301 — 1 plants were neglected before transplantation, and given
poor soil in field cultures. These data are subject to the criticism that
in a study of environmental effect on plants, seed from the same,
instead of from different plants, should be used. But this criticism is
probably invalid here, because the fasciated race is a pure line upon
which selection (White, 1913) seems to have no effect. So for purposes
of comparison, seeds of different individuals of a pure line have the

Induktive Abstammungs- und Vererbungslehre. XVI. 7

White.

same value whether from the same plant or generation or different
plants in different generations. The tent -grown plants in Connecticut
had a higher average number of leaves than the plants of the 1910
culture at the Bussey Institution. 1911 was an especially favorable
year for fasciations. After transplantations in June, a drouth ensued
through which the plants barely lived. July followed with much rain
and excessive heat. 1912 was a fair year for field cultures. A June
drouth was followed by rain in July, but the change was not so abrupt
nor so extreme as in 1911.

From the table, it is apparent that this race shows very decided
variation in the number of leaves in the different families, and if I am
correct in considering the variability as not due to a mixed population,
the only other alternative is to admit the effect of environment. And
yet environment does not affect the expression of this character in
Nicotiana to the degree claimed by de Yries for his fasciated races.
In all my field cultures and those reported by Mr. Dewey and Dr.
East (aggregating 694 plants grown over a period of five years in
four distinct environments) the fasciated character has bred very true
in the sense that its somatic expression always showed to some extent
in the stem, in leaf number, in the character of the inflorescence and
the floral organs, and that this development never fluctuated toward
the normal sufficiently to make classification even remotely questionable.
No so-called atavists have appeared at any time.

Another experiment was inaugurated in the winter of 1912 to test
further the constancy of the fasciated character under different environ-
ments. Cultures of normal and abnormal plants were started in February
by sowing seed of these strains in ordinary sterilized soil. Germination
was excellent, and the young plants were kept in their seed quarters
for about four months. No additional food was supplied them, and a
struggle for existence ensued. As a consequence, the survivors do not
represent more than one tenth of each original "pot" society. From
these 5 '08 and 10*16 cm. pots (orig. seed quarters) the survivors were
removed, with all of their soil, to 15*2 cm. pots and fed once a week
with a solution made up as follows:

2 pt. superphosphate

1 pt. sulphate of ammonia

1 pt. potash

1 teaspoonful to 7*57 litres of water.

Studies of Teratological Phenomena. 99

Table D.

Effect of environment on the expression of the factor A.

Designation

Survivors |

oomatic
A "nnpar-
ance

Hght. in
dcm.

No. of
leaves

Flowers

TJ.pmfi'rlf s

xvviuai a o

301—1—8

Abnormal

7 63— 92

42-51

Abnormal

*) One plant with no

fasciation.

301—1—32

7-63—9-2

62—77

One plant with ex-

treme fasciation.

303—1—13

1 plant, 80

Stems characteristic-

ally flat.

402—1

Normal

4 plants, 9'2

20—22

Normal

(304X402)— 1-

-28

3-05—9-2

17—22

(304X402)— 1-

-8

Abnormal

3 plants, 9*2

76—80

Abnormal

Stems all fasciated.

(304X402)—!-

-31

Normal

5plants,9'2

19—

Normal

Total

When the text photographs were taken (Fig. 15), the plants were
6 months old, having passed 2 months in the 15*2 cm. pots. In the
course of these two months many died, and the remainder had bloomed
and matured seed. Many of their leaves were yellow and the bottom
four to seven leaves had fallen. Table D gives their pedigree, the
number of survivors to each pot, their height and character. The flowers
were as large and as unaffected as though the cultures had been given
the best care. The normal flowers were in all respects similar to those
of field -grown plants. The leaves were reduced to a fourth of their
normal area, but as shown by the table, their number remained unchanged.
The main plant stems were very small, but flattening was as charac-
teristically expressed, though in 'baby-ribbon' dimensions, as that of
field plants.

While other characters such as leaf size and plant size are
modified in expression by adverse conditions this does not seem to be
true of fasciation as it appears in Nicotiana, except through its relation
to other characters, such as size of stem. Not one single individual
of the whole 61 surviving abnormals but what could very easily be

x) Stem so small in diameter that inflorescence fasciation would be difficult to
determine through casual observation.

. 7*

100

White.

distinguished from the normals grown under the same conditions. So
far as this race is concerned "atavists"' in the de Vriesian sense do
not exist. And when abnormals and normals are grown under certain
specific identical environments, my experience gives me reason to believe

there will always be the same deci-
ded features by which to distin-
guish them.

4. Fasciations and selection.

In a former paper (White, 1913)
the statement was made that the
fasciated character of the 300—309
race did not seem an}' more amenable
to selection than the race of cocks-
combs with which de Tries worked.
In other words, from two generations
of selection for normalness, no effect
seemed to be apparent, so the attempt
for the time, was given up. The exact
data on which this conclusion rests
were by no means conclusive. The
results obtained were clearly enough
interpreted, but the experiment was
not of long enough duration, nor
detailed enough to furnish any but
indicative data. In 1909, East selec-
lo- ted one of the most abnormal and

one of the least abnormal plants of the
fasciata race cultures grown at Bloomfield to be selfed. 301 — 1 is the
pedigree number of the least abnormal, and that of the most abnormal
plant is 303 — 1. Both the parents were shade-grown plants raised by
Dewey. The parent of 301—1 had 64 leaves, that of 303—1.
133 leaves. East took no definite data on the number of leaves of
301 — 1 and 303 — 1, but remarks in notes that 303 was the most
abnormal of all the eight families grown in 1909. The leaf count
varied between 40 and 100. The plant selected as 303 — 1 was one of
the most abnormal of its family. Plants from selfed seed of these two
strains were grown in 1910 side by side at the Bussey Institution.
Table 4 shows the individual variation of each plant in the number of

Studies of Teratological Phenomena.

101

its floral parts while Table 2 gives this same information for the number of
leaves. Table 3 is a summation table showing the frequency distribution
in number of flower parts per flower for the race. Adding together
the total number of flower parts of all the flowers from all the plants
of each race examined, and dividing this sum by the total number of
flowers examined a constant is obtained by which to more accurately
compare the difference between the two races (301 — 1 and 303 — 1).
Thirty three 301—1 plants with a total of 825 flowers, gave a constant
of 6*509+, while 303 — 1 with a total of 850 flowers from 34 plants,
gave 7*152+, the difference in abnormalness being 0*643+ in favor
of the 303 — 1 race. This difference in abnormalness is apparent in the
averages calculated for all the flower parts, and is also true of the
stem-flattening and the number of leaves. The average number of
leaves per plant for 32 of these same 301 — 1 plants is 34*18+, while
that for 36 (303 — 1) plants is 45*52+ leaves per plant. Table 4 gives
the range of variation in flower parts per flower for each plant of the
two races. The range in individual plant variation is about the same
for the two races with a slight advantage in favor of 303 — 1.
The modes for the variation in number of parts per flower are also
the same, with the exception of the ovary -locules, 303 — 1 having
almost one more locule per average flower then 301 — L Selection so
far seemed to be producing results, so four plants were selected from
the progeny of 301—1 and 303—1 and selfed. These were 301—1—2,
303—1—14, 301—1—29 and 303—1—12. The two former were
approximately the least abnormal progeny of their respective families
grown in 1910, while the two latter were approximately the most
abnormal. The extent of their abnormalities in floral structures may
be noted in Table 4, and changes in leaf number in Table 2.

Selfed seed of these four selections was grown in 1911 and the
plants matured under about the same environment as surrounded the
301 — 1 and 303 — 1 cultures of 1910. Time did not allow me to make
an elaborate examination of the 239 plants thus produced, nor of the
303 — 1 and 301 — 1 plants that grew beside them, serving as checks.
However, by going through them at maturity, I was able to classify
them roughly by the extent of their stem-fasciation into slightly abnormal
and abnormal classes (see Table E).

Slightly abnormal simply means that stem fasciation only appeared
in the region of the inflorescence. This was the stem condition of the
parents 301—1—2, 301—1—29 and 303—1—14. Parent 303—1—12
had an exceedingly abnormal, bent flattened stem.

102 White.

Table E.

Jr edigree

Leaves of
Parent

Selected
toward

Slightly
Abnormal

Abnormal

xoiai

301—1—2

Normal

301—1—29

Abnormal

303—1—14

Normal

303-1—12

Abnormal

Total

192

239

No plant in any of the four families was any less normal than
the parents, and there were many more extreme abnormals in all the
selections than were present in either of the 1910 families. Roughly
301 — 1 — 29 appeared to be more abnormal in stem-fasciation then
301 — 1 — 2, but this was not true of the families 303 — 1 — 14 and
303 — 1 — 12, although the two parent plants represented extreme con-
ditions. If one may draw conclusions from such scant data, I should
interpret these results as showing the ever -varying nature of the
character, not ever- varying however, in the sense of de Vries. No plant
ever approached the normal, and were it not for labels, I should have
been unable to have distinguished the two 301 — 1 selections from the
parent strain growing beside them. De Vries, too, after two or three
years, found selection of little value, and this was especially true in
his attempt to produce a normal cockscomb by selection from an
abnormal race. The value of selection for the first two or three years
in de Vries' cultures may be accounted for by the fact that his plants
came from the wild, of whose immediate ancestry he was ignorant.
His fasciated races were also plants which as a rule were cross-fertilized.
Selection work on this Nicotiana race should be continued, and careful
detailed records taken during many years before a dogmatic decision in
regard to selection of an abnormal from a normal and a normal from
an abnormal could be made. The material is ideal for such work, as
it fills the conditions called for by the advocates of Johannsen's pure
line theory admirably, and the objection of bisexual inheritance is not
here applicable as in the case of animals. Starting with what is in
all probability a natural pure line, made homozygous by thousands of
generations of inbreeding, and making use of a character that acts very
clearly as a single unit, it would seem that the question of the power
of selection to modify this particular gene could be irrevocably settled.

Studies of Teratological Phenomena.

103

5. Fasciation and hybridization.

The hybridization work with the fasciated race was the most
important part of the investigation. Crosses between it and various
normal species and varieties have been numerous, and one such hybrid
family has been grown to the F4 generation. Crosses were also made
between it and a calycanthemous- flowered race, N. t. calyciflora. The
Fi progeny will all be described together; the later generations of all
crosses, under their respective headings. All parents used in this
hybridization work are typical plants of the forms described under
materials. Where the individual record of any fasciated parent used is
known, it may be found in Table 9. The number, place and year in
which any cross was grown is given in Table 7. The factor for fasciation
is designated (A), that of calycanthemy (B).

a) Fasciation X Normal.
Fx progeny (Fig. 16 u. 17).
Eight successful species and varietal crosses were made with the
fasciated race. The species hybrids were all sterile, while the varietal
Fi hybrids were all fertile. The
species cross (331 X 301) was not
grown in sufficient numbers to give
valuable data. The others, though
sterile, gave important data regar-
ding the nature of dominance.
Tables 6 — 12 give the frequency
distribution of floral parts (sepals,
petals, etc.) per flower per indivi-
dual plant of each different hybrid
family. Table 13 shows the frequency
distribution of parts per flower for
the family as a whole, and gives

Fig. 16. Expression of calycanthemy (B)
in Fx flowers from 396 X 342. Note
the petaloid sepals.

the mode, number of flowers and
plants examined and the character
of single average flower for each
of the crosses.

Exceptions to uniformity in Fi. All the Fi plants of a given
cross and its reciprocal were alike in the characters under observation
with two exceptions. One of these was probably a reversal of dominance

104

White.

in connection with corolla shape, possibly resulting from internal
environmental changes. It occurred in a family of 36 Fi plants of
301 — 1 — 5 X 373, and all its (Plant 17) flowers were deeply lobed,
similar to those of the mother, rather than to the pentagonal corolla-
shaped flowers of the pollen parent. Cuttings were taken and grown
the next year in field cultures, with the result that both of the parental
types of flowers as well as intermediates were produced on the same
plant. Seed was not saved from these three kinds of flowers to

Fig. 17. Flowers of parents and Ft hybrid (303—1—24 X 332).
From right to left: 332, 301—24 X 332 and 4 flowers of 300—309 race.

determine if somatic segregation had taken place. Selfed seed from
the original plant gave only an ordinary segregating F2 population.

The other exception was in the nature of a mutation. It occurred
in a 304 X 402 Fi family of 39 individuals grown in 1910. Its
resemblance to a plant of the pure fasciated (300 — 309) strain was very
striking (see Table 6, Plant 39) and I would have eliminated it as a
stray contamination, without any hesitation, had I not found after
repeated trials to secure selfed seed, that it was at least self- sterile.
And as I recall, cross -fertilization was also attempted, but resulted in
failure. The latter point is one on which I have no notes, so I am

Studies of Teratological Phenomena.

105

not absolutely certain that it was tried. Dozens of flowers were bagged
and hand pollinated, but no seed ever resulted. One may call it a
reversal of dominance, i. e., a change from the intermediate to the
completely abnormal condition, but some internal disturbance of a
mutative nature must have accompanied this change, in order to account
for the sterility. Unfortunately, I knew so little concerning the nature
of my material and problems at the time that neither cuttings, nor

Fig. 18. Flowers of parents and Y1 hybrid (300—1—5 X 373).
Four center flowers belong to hybrid.

cytological material were secured. In the light of Digby's (1912)
studies on sterility in Primula Iceivensis, a cytological examination of
this case might have proved interesting.

Flower variability on individual Fi plants (Fig. 18).

The first flowers on a plant of the pure fasciated race were
usually the most abnormal. Flowers on plants maturing large numbers
of seed capsules were likely to be the least abnormal. Environmental
disturbances generally affected the variability of these organs in either
a plus or minus direction. And in order to compare the range of
variability of the fasciated expression in different plants and different

106

White.

families, it was necessary to take their flower records during approximately
the same blooming phase.

Table F shows a study of 88 flowers collected from the same plant
at different intervals of time. The study is given here in order to
show the fluctuation there may be in dominance when environmental
factors enter, for the differences in the three lots are largely due to
this cause.

Table F. Effect of environment on the expression of (A)
in an Fi hybrid plant.

) Aug. 10, lyii

No. of

Floral leaf el

ass range

Jj (V)

F(v)

Constant

Flowers

Number

OCUals ....

1 99,

J — o

V 1 9

— w**o

Petals ....

126

5-04

2-016

Stamens ....

>»

127

5 08

2-032

Ovary-locules . .

2*12

2'120

2) Sept. 1911

Sepals ....

199

5'38

2 152

Petals ....

196

5-29

2-116

Stamens ....

197

532

2-129

Ovary-locules . .

2-18

2-189

3) Feb. 1, 1912

Sepals . . . '. ,

137

5*27

2-107

Petals ! . . .

145

5'58

2-230

Stamens ....

7»

144

5*54

2215

Ovary-locules . .

2*38

2-384

Variability of dominance. Table 13 is a comparison between
the average abnormal condition of the Fi flowers of the different crosses,
the data from all, except the first and last resting on a study of
25 flowers from each Fi individual of the respective hybrid families.

324 X 301 appears to be the most abnormal, followed somewhat
closely by 303—1—24 X 332 and 304X402. The least abnormal Fi

x) Field.

2) Same field as Aug. 10.

3) Greenhouse (richly fertilized soil).

Studies of Teratoiogical Phenomena.

107

families were 353 X 301, 301 X 396 and 301—1—5 X 373. The cause
of the high abnormal condition of 324 X 301 Fi flowers is easily
accounted for by the character of the 324 parent, as the flowers of
this race are 5 — 7-petaled and almost constantly in possession of a
three-loculed ovary. In all cases except the one just mentioned and
301 X 396, one parent is normal. My Fi data on this subject would
be more satisfactory if I had used only a single fasciated plant as the
abnormal parent in these crosses. Personally, I believe the race to be
a homozygous pure line, and if the factor A cannot be modified by

Fig. 19. Flowers of parents and Fx hybrid (303—1—13 X 327).
Three center flowers belong to hybrid.

selection, any one of the pure strain individuals would be similar to
every other one, so that the flowers of an Fi family grown from

303 X 373 should not be more nor less abnormal than those from a
301 X 373 culture, providing both were grown under the same environ-
ment. An experiment of this kind is in progress. Two or more crosses
of the same kind, so far as my experiments have gone have always
given similar data on dominance, leading one to infer that internal
environmental differences of the normal races are responsible for the
variability in the expression of A in each of these hybrid families (Fig. 19).

Cytology. The maturation phenomena in the Fi anthers of

304 X 402 was investigated, but as compared with the pure strain, the

108

White.

abnormal expression of A was small. Contabescent anthers occurred
in this cross, as well as in all the others, but their number per flower
or plant depended always on the degree of dominance expressed by
the factor A.

F29 F3 and F4 progeny.

F2 generation from three different varietal crosses have been
grown, these being 301—1—5 X 373, 301—1—1 X 353 and reciprocal,
and 304 X 402.

N. tabacum fasciata (304) X N. tabacum (402) (Fig. 20).

The experimental results and data from this cross have been given
in detail in an earlier paper (White, 1913). The F2 population con-
tained abnormals (AA), intermediates (Aa), and normals (aa) in the
proportion of 1:2:1. All the apparent classes were easily and accu-
rately separated. The abnormal segregates bred asolutely true in F3
and F4, while the intermediates continued to produce AA, Aa and aa
progeny in these later generations. Such a population is represented
by Tables 15 — 16. By glancing at it, one may see how sharply the
three classes are differentiated, and this contrast becomes more marked
when one sees the plants themselves rather than mathematical charac-
terizations of them.

Table 14 shows the ration of AA, Aa und aa segregates among
the progeny of six different selfed heterozygotes. In some cases, where
the families are small, the actual and expected ratio are far from
agreement, but in the case of (304 X402) — 1 — 34), one could not
possibly expect a closer approximation between the two. The total
ratio from all the heterozygous families also closely approaches that
demanded by theory.

In this cross, so far as one could determine by observation, no
segregation in other factors took place, and one would expect none to
occur if the two races differ only by the genetic factor A, as I strongly
believe.

Character of aa and AA segregates.

Table 17 shows the character of an F3 population from a selfed
AA F2 segregate, while Tables 18 — 19 represent two families from
selfed aa F2 segregates. A glance at the latter may cast doubt upon
my statement that pure aa segregates occur in F2 and breed absolutely
true, because these tables show there is a slight variability in the
number of parts per flower. But the normal (402) itself, so far as

Studies of Teratological Phenomena.

109

casual observation can determine, also shows this slight variability, and
in practically the same degree. In an omitted table, the frequency
distribution of these variations of parts per flower and the number of
the abnormal flowers to each plant were calculated. Out of 11*475
flowers collected from 459 normal segregate plants or their progeny,
there were 11*079 normal to 396 abnormal, or over 3*5 per cent of
the latter. These abnormal flowers, representing 240 of the 459 plants

Fig. 20. Young inflorescences of 304 — 1, 304 X 402 and 402. Photographed from

herbarium specimens.

concerned, had 608 extra floral parts (sepals, petals, etc.) or about
1*5 part per flower. Their absence from the remaining 219 plants is
of no significance, as each plant of the whole 459 was represented by
only 25 flowers, and as abnormal ones are more common on branches
just entering the blooming period, I might easily have overlooked them
on some plants. Plants on which two or three such flowers had been
found were selfed and their progeny grown, and from an examination
of these, I concluded this variation was not heritable (Fig. 21).

As to whether the flowers of the aa segregates, in respect to
number of parts, are as normal as those of the 402 grandparent, is an

110

White.

Studies of Teratological Phenomena.

Ill

important question theoretically, on account of its bearing on the theory
of gametic contamination as opposed to that of gametic purity. My
casual observations on the 402 race are not in a sufficiently definite
form as yet to convince others than students of this same material that
they are accurate. This being true, two interpretations can be placed
upon my data. One may say, until proof to the contrary in a tabulated
form is presented, that segregation of fasciation- determining material
was not complete in the formation of the F2 seed, and consequently
slight abnormalities in aa flowers appear, the latter being evidence in
favor of such a supposition. However, one may also say that these
abnormal flowers occur just as frequently in the 402 race, that they
are just as abnormal, and that their presence is the result of accidents1)
in ontogeny, in which case, internal character-materials have undergone
absolute segregation and the F2 aa plants are pure normals.

In order to prove the former contention, it must be shown that
the F2 aa segregates under the same environment as 402 plants,
continuously produce flowers more abnormal than the latter. Further,
if contamination is at all common, repeatedly back - crossing normal
segregates with the pure abnormal (300—309) race, should eventually
give one a pure abnormal race, and each generation of back- crosses
should show some progress. In making such an experiment, one must
first be certain that homozygous lines are being used as material,
otherwise, if favorable, one might look upon its results as due to
unconscious selection, for if one believes as does Castle (1912) in the
creative power of selection, there is no apparent physiological reason
why, through self-fertilization and selection, an abnormal race such as
300—309 could not be produced from these slightly abnormal -flowered
F2 segregates.

If one adopts the "fixed factor" conception of East (1912) the
extra parts of normal flowers on aa and |402 plants are simply and
plausibly explained as ontogenetical accidents. For all practical pur-
poses, the aa segregates are as normal as (402) individuals.

Cytology. The maturation phenomena of the anthers were
investigated. Cytological irregularities in AA segregates were similar
to those of the abnormal grandparent; in Aa anthers, similar to the Fi

*) The term accident is used throughout this paper in a very specific sense. In
using it, 1 intend to convey the idea that is expressed when a very accurate technical
device, such as a rotary microtome, misses cutting a section.

112

White.

conditions: while in aa segregates, normal conditions prevailed as in the
normal grandparental race. The proportion of contabescent anthers and
the cytologieal irregularities which they contain fluctuates as the
grosser alterations produced by the factor A vary. These facts are of
theoretical interest, because of their possible bearing on the chromosome
theory of inheritance (see White. 1913).

X. tabacum fasciata (301 — 1 — 5) X X tdbaeum havanensis (373).

Two F2 families were
grown (Fig. 22). Tables 20 a
andb give the ratio of abnormal
to normal segregates and the
proportion of fasciate to non-
fasciate- stemmed plants. It
was impossible to accurately
classify the heterozygotes,
owing to the numerous modi-
fications the character fas-
ciation had undergone, as
compared with its appearance
in the pure strain and the
304 X 402 A A and Aa segre-
gates. The small proportion
of 1 fasciated stem to 5 normal
one was perplexing, inasmuch
as theory demanded three of
the former to one of the
latter. The individuals of this
population did not vary much
in height, but segregation of
factors governing color, habit
and leaf character had occurred.
The leaf count per plant varied
between 16 and 111: for those classified as normal, it ranged between 16
and 28 and for "possible heterozygotes" as low as 19 and commonly 20
to 25. Table 21 shows the character of an F2 population, in terms of
its flower variability and leaf number per plant.

A", tabacum fruticosa (353) X JV. tabacum fasciata (301).
Only two F2 families of this cross were grown and the ratio of
abnormal to normal plants may be found in Tables 20 a and b. The

Fig. 22. Two F2 segregate plants (AA) from
301—1—5 X 373.

Studies of Teratological Phenomena.

113

difficulties in classifying the three kinds of segregates were greater than
in 301 X 373, heterozygoses very often being classed as normals, until
repeated inspections had taught me to look the whole plant over carefully
for abnormal flowers. Individuals with fasciated stems occurred in a
proportion of about 1 F to 9 N, whereas one would expect 3 to 1.
353 differs very markedly in character from 301, and one might almost

a Fig. 23. b

a) Fx plant with 18 leaves (301—1 X 353—3).

b) F2 AA segregate with 100 leaves.

say dozens of these differences had segregated, so that the F2 generation,
owing to the many new factor combinations, represented an extremely
variable population. Variation in height ranged from 10*7 to 24*4 dcm.,
and the leaf count from 13 to 212 leaves per plant. The relation
between these characters is shown in Table 23. The segregate
bearing 212 leaves was 17 '54 dcm. high and had a very fasciated stem,
while the one with 13 leaves was 16 '8 dcm. and normal -stemmed
(Fig. 23 and 24). Table G gives the frequency distribution of height in deci-
meters and number of leaves per plant for 24 apparently normal segregates.

Induktive Abstammungs- und Vererbungslehre. XVI. g

114

White.

Table G.

Hght. in
decimeters

12-2

12-96

13*72

14-48

1524

16-00

16*76

17-52

18-28

19-04

19-80

20-56

21-32

22-34

Frequency

Lvs. per
Plant

Frequency

Extremely abnormal segregates,
judging from the records, have a
leaf count as low as 24 and a height
of only 12 '2 dcm. A survey of
the data in Tables 22 and G shows
there is some correlation between
height and number of leaves. In
addition to the characters just
mentioned, these segregates differed
as to color, habit of growth (bran-
ching or non-branching), leaf charac-
ters, flower characters and many
others not so easily noted.

b) Fasciation X calycanthemy
and reciprocal.

This cross was especially in-
teresting for two reasons; first,
the segregation of two characters
which were more or less associa-
ted in the same organs, and second
on account of the numerous modi-
fications of fasciation, even in AA
segregates.

In the first case, both charac-
ters had given simple 1:2:1 ratios
in an F2 generation obtained by
crossing each of them separately with the normal 402. (See Table H.)
Both were partial dominants when heterozygous, hence in monohybrid

Fig. 24. Abnormal F2 segregate (pro-
bably AA) from (353 X 301—1). Note
absence of fasciated stem.

Studies of Teratological Phenomena.

115

crosses with 402 , the Aa and Bb classes were always easily dis-
tinguished. In the dihybrid cross under discussion then, an F2 ratio
of 1 AABB : 2 A ABb : 2 AaBB : 4 AaBb : 1 AAbb : 2 Aabb : 1 aaBB : 2 aaBb :

Fig. 25. Typical flowers from F2 segregates of calycanthemy X fasciation, each
flower representing one of the 9 somatic classes. Beginning at the top and left,
they are (a) AABB, AaBB, AABb, AaBb (b), aaBB, aaBb, AAbb, Aabb, aabb.

laabb was to be expected, the 9 gametic classes all being identifiable
(Fig. 25 and 26). The actual results obtained by growing 477
plants from three different Fi individuals bore out these con-

116 White.

elusions, but owing to complications caused by the segregation of
numerous other characters, the classification was extremely difficult.

Table H. (396 X402 F2 generation.) (Calycanthemy X normal.)

Ratio

Classes of F2 segregates

Total

Actual . . .
Expected . .

15
13-25

25
26-50

13
13-25

53
53

Some individuals could be assigned to their particular category
with accuracy and speed, while others were so hopelessly modified that

Fig. 26. Extremely abnormal flowers from F2 AABB segregates
of 396 X 402.

even after repeated trials and extensive study, they could only be
placed by guess. This state of affairs was not a particular class
characteristic, but true of all. In order to be as accurate as possible,
three separate classifications were made: one early one involving only
a part of the population, and two later ones in which nearly all of
the plants were considered. Those not included, were either destroyed
accidentally or were not yet in bloom at the time of the last inspection,
and as they were few in number, the absence of data from them is of
no significance. Table 23 gives the results of the three classifications.
I consider that of the second the most accurate, as most of the plants
were in full bloom when this was made. Count 3 was taken later in
the season, when many of the segregates possessed few flowers, but

Studies of Teratological Phenomena.

117

numerous maturing seed capsules. Under the latter conditions neither
the character fasciation (in floral expression) nor calycanthemy are
expressed typically. This is especially true of the individuals of the
Aa and Bb categories. Table 25 gives the character of the individuals
of this F2 population, and enables one to note the fluctuation as
regards each class. I consider the individuals of the classes A ABB,
AABb, AaBb, A abb and aaBB to be the most accurately identified, and
those of aaBb and aabb as probably the most inaccurately placed, as
some of the former no doubt are AaBB and AaBb and some of the
latter also are AaBb. But all the discrepancies between the actual and
expected ratio cannot be accounted for by assuming preventable error
to be the cause. The classes aaBb and aabb especially are too high
to be explained in this manner, and in all counts the greatest dis-
crepancy occured in the class aabb. But the factor B was not causing
the distortion of the ratio, for in a population of 469 F2 individuals,
351 were abnormal (BB or Bb) and 118 normal (bb)1), a very close
approach to the expected ratio. The factor A was causing the trouble
as shown by the total number of abnormal (AA and Aa) to the total
normal (aa) segregates. An examination of the whole F2 population as
regards stem-fasciation gives an added basis for my statement. Table 25
gives the total fasciate- stemmed plants to those with normal stems,
the proportion being about IF to 8*5 N., whereas theory demands
3F:1N, and at least *119 (*1F:2FN:1N) plants fasciated enough
in this structure to be easily distinguished, and twice that number with
slightly altered stems. In making the observations, all plants showing
the least indication of monstrous alteration in stems were included in
the (F) class.

It is very evident in this hybrid population, that too many aa and
not enough AA and Aa segregates were present to accord with theory.
A glance at the nature of the characters present additional to those
caused by factors A und B disclosed similar conditions to those found
in the 353 X 301 and 301 X 373 populations. The two grandparental
races differed by a large number of factors, and these were segregating.
Height, leaf character, branched and unbranched habit, normal leaf
number and numerous floral characters are a few of the more prominent,
and as a consequence of this great shuffling of factors, the 477
segregates presented striking dissimilarities. Hardly two of the whole

*) Note second count.

118

White.

lot, but were distinguishable by some character difference, and this
population was by far the most variable of all those which I grew. In
Table 25, some idea of the striking differences in height may be gained.
The range is: aa segregates, 12*2 — 20*6 dcm.; AA and Aa segregates,
10*7 — 24 '4 dcm. In leaf count, the aa class varied from 10 to 17;
the Aa between 10 and 33, and the AA, from 14 to 85 leaves per plant.
Fasciate-stemmed segregates are not confined to any particular type,
but are present in a large number of the different character combinations,
among them being dwarf individuals (9*15 dcm.) with only 15 leaves,
and extremely tall plants with either few or many leaves. The
correlation between height and number of leaves per plant was not
marked. Branched fasciated stems appeared for the first time. The
branching type had come from the 396 grandparent and the fasciated
stem character of the 300 — 309 race had spread itself out over the
three or four main branches, expressing itself even in the little twigs.
This was a distinctly new type, and not a bifurcate or multiradiate
fasciation. Other well defined types with fasciated stems occurred, and
one of these, especially distinctive, ganietically AABB or AaBB, had
a slender, unbranched, flattened axis bearing only 10 or 15 leaves.
All types as described in Table 25, were selfed and much more light
will be thrown on the subject by the F3 generation. Pending the
growing of this, the explanation given for changes in dominance of A
in the various Fi crosses, is presented to account for the unexpected
distortion in the F2 ratios.

Discussion.

From the results of these three varietal crosses, together with
the data from 9 Fi hybrid families, it would appear that other factors
must markedly affect the somatic expression of the factor A. And this
may be so much modified that the intermediate expression of dominance
in the Aa segregate may be changed to complete dominance of the normal
(aa) condition, provided certain other unrelated, but interacting factors
were present in the zygote. If this occurred, more segregates somatically
normal would be expected in F2, because heterozygotes of this kind
could not be distinguished from aa plants, except through the breeding
test. The failure of the factor A to alter the normal appearance of
the stem can be accounted for in the same manner. It is very evident
from the F2 results of the cross 304 X 402, that when two homozygous
pure lines differing in a single factor are crossed, the F2 individuals

Studies of Teratological Phenomena.

119

of each apparent class will differ but very little from each other, and
this slight difference may be described as the difference in external
environment. On the other hand, when two homozygous pure lines1)
differing in numerous factors are crossed, F2 individuals of each apparent
class may express the character especially under observation in many
distinct morphological forms, each of these changes in expression being
the result of modification by other factor developments. Calycanthemy
(B) is not so easily altered in its expression by changes in factorial
complexes as is factor A.

6. Summary and conclusions.

The following conclusions may be drawn from this study:

1. According to the evidence already presented, the original plant
which became the progenitor of this fasciated race, must have mutated
from unknown causes, at or shortly after fertilization had taken place
(somatic mutation in the embryo). For the reasons given in the body
of this paper, the mutant is believed to have arisen in a homozygous
condition as regards the factor A and upon self-fertilization to have
bred true.

2. The Cuban variety of tobacco known as 402 in the B. I. of
H. U. cultures is believed to differ from the Nicotiana tabacum fasciata
race (300 — 309) only in the absence or presence of a single genetic
factor (A). In all other respects, the two races in breeding tests
give results that would lead one to believe them to be identical
genotypes.

3. As a generality, N. tabacum varieties and races are held to
be natural pure lines (as reported by Hasselbring) and to give uniform
progeny in breeding tests of self-fertilized seed parents. Owing to
cleistogamy and the consequent rarity of cross -fertilization, such pure
lines, in the majority of cases, have bred true for thousands of
generations. The sexually produced progeny of these N. tabacum
homozygous pure lines, would differ in no way from the progeny that
might be created by asexual methods such as cuttings.

*) In all the crosses of the fasciated race with normal varieties and species of
Nicotiana, all the Fj generation plants of a given cross were practically uniform in
appearance (height, flower-shape, and color, foliage and stem characters, etc.). This
fact indicates these normal parents to have been practically homozygous genotypes.
(See Tables.)

120

White.

4. The race (300 — 309) is such a homozygous pure line because
it has been propagated from seed obtained by selfing a single (original)
plant, which is believed to have arisen or mutated by a single factor A
from such a homozygous pure line.

5. The factor A of this race, in expressing itself somatically,
when surrounded by its original gene complex (402) affects a large
number of important plant organs. Prominent among these may be
mentioned the stem, the leaf number and arrangement (phyllotaxy), the
inflorescence and the flowers. The nature of this expression is held
to be ontogenetically progressive, as the seedlings and the early juvenile
stages do not show any distinguishing fasciata features. As the plants
of this race progress toward maturity, the factor more and more
implants its distinctive morphological characteristics upon the various
organs, so that those which develop last, exhibit the greatest alterations.
Hence, the last flower whorl laid down in ontogenetic development is
the most altered from its normal expression by the factor A.

6. The hereditary nature of the fasciated condition has been
tested by breeding large numbers of progeny from the seeds of a single
selfed plant. In all cases, the character is constant in the sense that
its extreme fluctuations do not approach the normal (402) condition
near enough so as to call forth any question as to which is which when
the two are grown together. In other words, all the progeny of selfed
plants of this race express the character fasciation, as described under
"materials" to some degree. Its fluctuation is largely "inherent" and
not the result of the "external" environments under which the ex-
perimental cultures were grown. As no "atavists" appeared, the character
is not "eversporting" in the de Vriesian sense.

7. The repetition in the number of organs, such as leaves, sepals,
petals, stamens and ovary-locules is not a duplication of whorls
(pleiotaxy) or of practically whole organisms, as the theories of some
anatomists would seem to imply, and as de Vries suggests in explanation
of his data on Geranium molle fasciatum. No evidence of congenital
mechanical fusions is given by cross-sections of the mature stem. The
different whorls in the flower appear to vary somewhat independently
of each other, as the correlation in number of parts between those of
a single flower is far from perfect, though probably exceeding 50 per
cent. The progressive expression of the factor in its ontogenetic
development may entirely account for this.

Studies of Teratological Phenomena.

121

8. Under five distinct geographical and topographical environments,
the characteristic expression of the fasciated race has remained quali-
tatively, but not quantitatively constant. One of these environments
was characterized by impoverished soil, general cultural neglect, and
over -crowding, but the plants, though dwTarfed in the size of their
leaves and stems, had flowers and stems of the prescribed abnormal
type. From comparisons in leaf number, between individuals grown
under different field conditions, it is believed that environment, especially
the factor weather, is able to change their variability in number per
plant. The change in number of floral segments may also be associated
with climatic factors.

9. The 300—309 race is probably not as well ''adapted" to
different environments, as is the parent race from which it mutated.
The few seed set by the original mutant might indicate that had it
arisen under natural conditions, its ability to persist would have been
very problematical. In the Harvard University cultures, the race is
very fertile, though falling short in this respect when compared with 402.
This means, that on a basis of fertility, the latter would be the more
successful competitor in the struggle for existence. The modifications
produced by a combination of the factor A, and certain environmental
conditions would further cause this variety to be a losing competitor in
the struggle with the normal, as under extremely favorable environmental
conditions (for the factor A) very few flowers would develop and mature
seed. As far as roots and general plant vigor are concerned, when
placed under ordinarily favorable tobacco growing conditions, one race
is no better equipped for existence than the other. Under many tropical
environments, I have no doubt that the race, as now grown, would
persist and compete successfully with many other plants for existence,
if selfed seed were sown in such places.

10. The results secured from the selection experiment are only
indicative and possibly may be interpreted as favorable to the idea of
the creative power of selection. I prefer to interpret them as indicating
the inability of selection to modify the fasciated character, so as to
produce eventually a normal. The data are not conclusive, however.

11. The data from crossing a fasciated plant with a normal (402)
plant demonstrated in a clear manner that the two races apparently
differed only in the possession of a single unit factor A. The Fi was
intermediate in character and the F2 gave abnormal (AA), heterozygous
(Aa), and normal (aa) segregates in the ratio of 1 : 2 : 1 or 3 abnormals

122

White.

to 1 normal. The three classes were clearly recognizable and very
distinct, the heterozygote always being an intermediate. No segregation
of other factors, which might have been hypostatic (as in Bateson's
sweet peas) occurred. The total progeny of 304X402 Aa plants gave
an extremely close approximation to the theoretical expectancy.
Actual ... 98 AA:192 Aa : 103 aa; Total 393,
Theoretical . . 98 AA : 197 Aa : 98 aa; „ 393.

12. Four generations of the cross (304 X 402) have been grown
and the F2 segregates have all bred true to their respective character,
AA plants producing always AA, Aa individuals always producing the
original F2 ratio of 1:2:1, and aa plants always giving rise to aa
progeny. F2 normal segregates and their F3 and F4 progeny have
possessed two or three abnormal flowers. These, as pointed out in the
body of the paper, may be interpreted by some biologists as evidence
in favor of a theory of gametic contamination as opposed to that of
gametic purity, because I have nothing better than casual observations
to prove that the normal 402 plants have the same character in the
same degree. This character is not hereditary in the strict sense, and
upon the fixed unit factor conception, these slightly abnormal flowers
may be explained as accidents in ontogeny.

13. The cytological investigations of the maturation phenomena
in the anthers and ovaries of the pure abnormal (300 — 309) race showed
numerous irregularities in the normal processes of reduction. Chromatin
and cytoplasm were alike affected, and many of the pollen mother-
cells were destroyed through the presence of factor A. Only a certain
percentage of anthers and ovules are abortive, and this partial sterility
is made good as far as seed production is concerned by the increase
in the number of ovules and ovary -locules. In crosses with 402, the
Fi plants show the same irregularities, but to a lesser degree and in
a lesser number of anthers. The cytological observations on the F2
generation show that a perfect correlation exists between the gross
morphological and cytological features of this fasciated race. The factor
for fasciation then produces its abnormal effect even in the germ-cells,
and this effect is subject to alteration in both somatic and germinal
structures through changes in dominance. The same factor A expresses
itself in the soma by altering the form of many of the characteristic
organs, and in the germ -cells by destroying a certain percentage of
cell materials, or by producing irregularities in their normal processes
which ultimately cause their own destruction. Because of the F2 ratio

Studies of Teratological Phenomena.

123

of this cross, it is held that only one factor is involved, and if that
factor is present in a chromosome, it is in duplex condition in each 2n
cell and simplex in each n cell. All anthers and other organs are held
to be gametically similar in such a race, and every cell, except for
environmental modifications is identical with every other cell of the
organism. Pollen mother-cells are identical as to factor composition and
can differ from one another only in environmental modifications.
Environmental conditions must be practically the same in the same
anther and in all the anthers of a single bud. The questions then
arise why the factor A is latent in some cells and patent in others,
and why a certain definite percentage of each anther's pollen is not
aborted rather than all of it, etc.1). From these data, one would find
no support for the chromosome hypothesis of inheritance and in theory,
there is directly opposing evidence.

14. Three other crosses were made, the Fi plants of which were
fertile. One of these (301 — 1 — 5 X 373) did not appear to differ much
from 402 or 300 — 309 in factors affecting height, though in other
respects the factorial complexes of the two varieties were very
different. This was even more true of the varieties 353 and 396, as
they were branching, shrubby, dwarf-statured plants. When crossed
with 300 — 309 and the F2 generation observed, it was very evident
that the latter variety differed from the two former in an exceedingly
large number of characters (factors). The crosses of 396 with 301
demonstrated the extreme extent of these differences, as hardly any
two plants were alike among the whole 477 F2 progeny, and many
of their distinguishing features were sharp and clear-cut genetic
characters.

15. The ratio of abnormal to normal plants in these three crosses
was very confusing, as in all cases, there was excess of "somatic"
normals and this divergence from the theoretical expectancy increased
as the apparent character differences between the two grandparents
became more numerous. In 301 — 1 — 5 X 373, the excess of normals
above the expected was only 2*8 per cent; in 353 X 301 — 1, the excess
rose to 4'8 per cent, while in 396 X 301 and reciprocal crosses, the
increase in normals over the theoretical had mounted to 9' 7 per cent

*) Attention is called to the similarity between the disturbances and modifications
produced by the factor A, and that caused by certain parasite mutilations. All the
experimental data are opposed to the theory that the factor A is an internal parasite
foreign to the organism.

124

White.

(second count) or 15*2 per cent on the basis of the third count. In
each succeeding cross, the total number of F>> progeny grown was
larger and the discrepancies between the actual and the theoretical
ratios should have been less in the case of 396 X 301 than greater.
In addition to the discrepancies between the actual and theoretical
ratios, the abnormal character was modified in its somatic expression
to such a degree that it was very difficult to separate the three classes,
AA, Aa and aa. Since going over my records, I think it would be
impossible to classify them accurately except through breeding tests.
Owing to the variable nature of doaiinance and recessiveness in these
crosses, even the selection of the aa (normal) segregates became
somewhat difficult, and no doubt the deficient number of abnormal
plants (AA and Aa) in the actual ratio obtained, may be explained by
my inability to distinguish properly between normal (aa) and Aa
segregates. Variation in the expression of the abnormal character was
extremely noticeable in all three of these crosses. The proportion of
fasciate - stemmed Fa segregates to those with normal stems was
respectively 1 — 5, 1 — 9 and 1 — 8*5. Many abnormal plants (AA) then
were not characterized by fasciated stems. In the case of the cross
304 X 402, all the F2 AA segregates possessed the fasciated stem con-
dition in varying degrees, but in some of the AA segregates of these
other crosses, the stems were as normal as any normal tobacco plant's
main axis well could be (Fig. 27). The explanation for this difference in
the expression of the factor A appears to lie in the different nature of the
gene complexes or genotypes. In 304 X402, both parents appear to
be genotypically alike except for the factor A, while in all the other
crosses, it is evident that this was not the case. Factor A expresses
itself as described under materials when in genotypical environment 304
and 402, but very differently under genotypical environments 373, 353,
396, etc, because it is modified in its somatic expression in all these
environments. Taking 304 or 402 as the standard genotypical environ-
ment by which to compare the remainder and calling it Xi, the other
environments may be referred to as X>, X3, or X4, etc. Under Xi,
factor or gene A always gives a certain typical somatic expression,
while under any other X, that somatic expression may or may not
remain the same. Under Xi environment, no branched fasciations were
produced, the ribbon -like linear expansion being characteristic only of
the main axis, but in the cross 396 X 301, under Xn environment,
segregate plants appeared expressing this anomalous condition in several

Studies of Teratological Phenomena.

125

of their branches. And of these plants, some may be homozygous in
enough of the factors which transmit the branching habit, that a constant
race could be obtained. Again, under X2 genotypical environment, the
factor A may not be able to express itself at all in the stem, and the
abnormal floral and phyllotaxy conditions may be reduced, owing to the

Fig. 27. F2 types of stems resulting from crossing the fasciated
race with normal-stemmed races of N. tabacum.

presence of modifying factors in this gene complex X>, that were absent
in that of Xi. Hence among the F2 progeny of crosses 353 X 301,
396 X 301, etc., plants segregated out which, judged by the expression
of the factor A under Xi conditions, would be recorded as heterozygotes,
yet they may be just as truly AA segregates, as any AA plants could
be. Such a state of affairs causes extreme confusion in a study of

126

White.

characters in heredity, but the phenomena are easily understood and
more simply expressed when one can work with progeny, the male and
female parentage of which is contained within the same plant. Under
the conditions above mentioned, an F2 population such as (396 X 301)
— 12 may contain: —

AA segregates resembling AA plants of Xi standard,

AA „ „ Aa „ £ „ „

AA „ very nearly as normal as those aa plants of Xi

standard which have two or three abnormal flowers.
AA segregates which have numerous new modifications in the
expression of the factor.
Such an F2 may also contain:

Aa segregates resembling AA plants of Xi standard,
Aa „ jj Aa „ „ „ „

Aa „ „ aa „ „ „ „

Aa „ with many new modifications.
One can readily see from this illustration the confusion which would
ensue in attempting to classify such an F2 progeny. The study of
this factor A and its expression in the „soma" under differing geno-
typical conditions is exceedingly complicated because it is capable of
expressing itself as a modification in so many of the important plant
organs. The study of a color character is much simpler, as I shall
soon indicate.

16. Dominance and recessiveness, being regarded as characteristics
of the environment (both genotypical and external) are expected to
show a large degree of variation in a series of crosses such as the
present investigation involves. A study of the Fi generation of four
very distinct species crosses and three fairly distinct varietal crosses
with the abnormal race (300—309) has fulfilled these expectations. All
the species and varieties are normal as regards fasciation, except 324
and perhaps to a slight degree 396. When crossed with the abnormal
(300—309) race, if other factors did not produce modifying effects, the
Fi plants from these different combinations should all show the same
degree of dominance and recessiveness. Such is not the case, however;
Fi plants of 301 X 373, etc. are the most normal, followed by 353 X 301,
301 X 353, 301 X 396, 303 X 327 in the order named, through to
332 X 301 or 332 X 303, 304 X 402, and 324 X 301, which are the most
abnormal. (Consult Tables 6—12.) The species crosses were all sterile.
The F2 generation of all the varietal crosses further substantiated this

Studies of Teratological Phenomena.

127

interpretation of the variableness in dominance shown in the plants of
the different Fi families. It is not improbable that the F2 progeny
of some of these species crosses, if they were obtainable, would represent
plants of the constitution AA, which would promote or allow a very
much greater abnormal expression that that of the standard Xi race
(300 — 309). Theoretically, I should expect such segregates.

17. It is of interest to note the somatic effect when the factor A
is present in simplex condition. Instead of a progressive increase in
expression, as the plant approaches maturity, the factor seems to be
unable to maintain its normal (under Xi) rate of development after the
first one, two, or three floral whorls are laid down in morphogenesis.
As a consequence, the gynoecium does not usually show the most
abnormal expression of the factor A in the eight or nine different Fi
combinations grown. In 303 — 1—13 X327, the greatest alteration
occurs in the corolla, the least in the calyx and the next to the least,
in the gynoecium. Practically the same relationship between the different
floral whorls and the somatic expression of A, is apparent in all the
other Fi hybrid families represented in Table 13. The exceptions to
this generalization are 324 X 301, 304 X 402 and 301 X 396. The first
combination does not follow the floral organs (see materials). The
second case, 304 X 402, is explained on the grounds of relationship,
the gene complex in both races being extremely favorable to the normal
expression of A. 301X396 shows only a slight deviation from the
general rule and is possibly due to error resulting from the manner in
which the data from it were collected.

18. The origin of factor A could be pictured as either a loss or
a gain in actual protoplasmic substance. It was either a sudden dropping
out of something essential to normal development (402) or it may have
been an abrupt change in the germ -plasm. As Morgan and others
before him have suggested, changes in chemical configuration (isomerism)
of protoplasm may account for the origin of new factors and such a
conception is preferable to the idea of an actual protoplasmic loss. The
Nicotiana factor A is both discontinuous in origin and in inheritance.
Whatever be its germinal nature, it arose as a unit and is trans-
mitted intact.

19. Calycanthemy (B) in crosses is transmitted as a single unit
factor, giving in the progeny of an Fi Bb plant, segregates in the
proportions of 1 BB : 2 Bb : 1 bb. When the calycanthemous and
fasciated races are crossed, the resulting AaBb individuals show only

128

White.

partial expression of the A and B factors. In F2, the progeny are
divisible into 9 gametic classes in the proportion of 1 AABB : 2 AaBB :
2 AABb : 4 AaBb : 1 AAbb : 2 Aabb : 1 aaBB : 2 aaBb : 1 aabb. These
classes also represent the apparent or visible classes, because the
heterozygotes are distinguishable from the homozygotes. In practise,
accurate classification of the progeny into these classes was impossible,
owing to the presence of modifying genes in the two grandparental
races, which were also segregating in this F2. The ratio of BB and
Bb to bb individuals in this F2 was as accurate as was to be expected,
but a great excess of apparently normal aa segregates occurred. Some
of these undoubtedly belong gametically to the other classes. Calyc-
anthemy is only slightly altered by the presence of modifying genes or
ordinary changes in external environment. This is to be expected, when
such a character is compared with one much more fundamental (in that
it expresses itself by such numerous and far-reaching alterations) in the
organism's life cycle, such as is true of fasciation.

G. General discussion, showing the bearing of these
data on some of the more general problems of evolution

and heredity.

Under this heading, I wish to present certain general conclusions
which my data, in my opinion, justify, The chief of these concerns the
latent character hypothesis of the morphologists, in so far as it bears
on problems of genetics and evolution. This conception, especially in
the form presented by deVries is not only confusing, but actually
contrary to fact, as I have shown in the body of this paper. A further
example will, I hope, make my previous contention clear. The character
fasciation, as it occurs in Pisum, appears in at least two physiologically
distinct forms. In Pisum umbellatum it is absolutely hereditary under
the ordinary environments in which peas are grown. Lynch, Rimpau,
Koernicke, Mendel, Lock, as well as myself, all vouch for this fact.
Under these same conditions the ordinary varieties of peas do not
possess this character. But in a case reported by Blodgett (Fig. 28),
environmental phenomena induced this character in a variety of Pisum to
the extent that not more than 10 per cent of the plants remained normal.
This effect was distinctly an environmental one and was not hereditary
in the absence of the inducing agent. Blodgett was not able to show

Studies of Teratological Phenomena.

129

this by experiment but after carefully considering the data he gives as
to its extent and origin, I think there is but slight doubt that had
experiments been feasible, the results would have proved the truth of
this assumption. A comparison of the photographs in Blodgett's paper
with those of fasciation in Pisum umbellatum (Fig. 29) show the two forms of
the character to be morphologically indistinguishable. In the presence
of such data, the question arises, are we justified in speaking of fasciation

as being latent in the normal peas, such as those reported byBlodgett?
It is far simpler, it seems to me, to regard it as hereditary in both
forms, for under the same conditions, it is reasonable to suppose that
this particular variety of Pisum sativum would always produce the
phenomena Blodgett observed. The interesting point is that there is a
hereditary difference between Pisum umbellatum and all other peas,
when all are grown under ordinary environmental conditions. In neither
case is a character latent, but rather absent, as a character is just
as much an effect of a specific environmental medium as it is
an effect of a bit of protoplasmic material. With appropriate
material and appropriate environment, this effect can always be produced,

Induktive Abstammungs- und Vererbungslehre. XVI. 9

Fig. 28. Fasciation in Pisum sativum due to environment.
(From photograph after Blodgett.)

^30 "White.

if we are to believe what chemistry and physics teach us. A character
is always the result, of a chemo-physical reaction and not a continuous
entity existing from generation to generation. When a character is
handed on from generation to generation, it is formed anew each time.

On this viewpoint, historical knowledge is the only criterion by
which the newness or the oldness of a character may be established.
Dominance has often been suggested (de Vries, Castle, etc.) as a

Fig. 29. Fasciated plants of P. sativum umbellatum arranged like Fig. 28 in
order to show their close morphological similarity to those described by Blodgett.

means of distinguishing between old and new, progressive and retro-
gressive characters. But since dominance itself is an expression
phenomenon, due to both environment and heredity, it must be
dispensed with as a criterion. So that in the end, the primitiveness
as well as the progressiveness of characters must be determined by
palaeontological evidence and logic.

This brings us to a still more important point, i. e., to the question
of the validity of the work of many morphologists who have drawn
deductions as to which characters in certain groups are primitive and
which are progressive. These deductions are not infrequently made

Studies of Teratological Phenomena.

131

regardless of the very necessary support from evidence based on studies
of fossils, and without paying due consideration to the effects of
environment. For example, the character fasciation in peas, from the
morphologist's viewpoint, is a single kind of character, for how are
they to know, in the absence of experimental data, that there
are, at least, two absolutely distinct forms of this character. Eelatively
speaking, perhaps these two forms of the fasciation character are even
but remotely related. They differ as to cause and in behavior, but
morphologically they are absolutely indistinguishable. If factors have
a definite reality in the sense in which we think of the chemical atom,
then probably a different combination of factors as well as of environ-
mental phenomena are responsible for these two forms, just as the
color, quality or character red, may be found in very distinct and
relatively remotely related forms of mineral matter. And if the
relationships of organisms cannot be judged by characters, how can one
formulate a natural system of classification?

With the elimination of the latency conception from biological
discussion, the various perplexing data on fasciation resolve themselves
into an orderly and simple scheme.

From the standpoint of genetics, all fasciations may be divided
into two classes ; those inherited and those uninherited, the former due
primarily to one or more genetic factors, the latter largely the result
of external environmental conditions. The "half" and "eversporting"
fasciated races of de Vries probably belong largely to the latter class.
No necessity exists for and confusion results from maintaining the con-
ception of "eversporting" fasciated races. Had the studies of de Vries
been made at the present time rather than in the pioneer period
of genetics, it is safe to say that this conception would never have
arisen.

One may further deduce from this study of fasciation, certain
conclusions regarding the phenomenon of dominance. Dominance and
recessiveness in the case of the character fasciation is always dependent
on the three elements held to be the basis of a character, — the primary
factor or factor complex, the complex of all the other genetic factors
constituting the total heritage of the organism, and the external environ-
ment. Changes in any one of these may give rise to a new character
or modify an old one, the distinction between old and new being
largely a matter of convenience, for, when a factor still retains its
usual or common expression, except for slight modifications which do

132

White.

not alter its distinctive features, it may still retain its old name, but
when these changes in expression are so complete as to make their
relation to the same factor unrecognizable, it were better to describe
them as new. This interpretation is opposed to that conception of
heredity which implies dominance to be a possession of the factor or
factors primarily concerned in the inheritance of a character. On the
other hand, it accords with and supports in every way, the fixed factor
hypo thesis of East.

Literature Cited.

Baker, R. T. and Smith, H. G., The Pines of Australia. Techn. Educat. Ser. No. 16,

Dept. of Publ. Instruction, N. S. W., Australia (see illustration p. 333), 1910.
Bateson,W., Mendel's Principles of Heredity. Cambridge Univ. Press, pp. 1 — 396, 1909.
Blaringhem, L., Anomalies hereditaires provoques par les traumatismes. Compt.

Rend. Acad. Sc. 140, 378, 1905.
Blodgett, F. H., Fasciation in field peas. Plant World 8, 170—177, 1905.
Breviere, M., Extrait d'une Lettre de M. Breviere, de Saint -Saulge. Bull, de Soc.

Bot. de France 28, 5, 1881.
Castle, W. E., The inconstancy of unit characters. Amer. Nat. 46, 352 — 362, 1912.
Church, A. H., On the relation of phyllotaxis to mechanical laws, III. Secondary

growth phenomena. Oxford, 1905.
Clos, D., Essai de teratologic taxinomique. Toulouse, pp.1 — 80, 1871.
Comes, O., Monographic du genre Nicotiana. Naples, pp. 1 — 80, PI. 14, 1899.
Compton, R. H. , The anatomy of the mummy pea. New Phytologist 10, 249—255,

Fig. 1—4, 1911.

Conard, H. S., Fasciation in the sweet potato. Univ. of Penn. Bot. Lab. Contrib. 2,
205, 1901.

Cook, M. T. , Teratologia de la Pina. Tornado del Primer Informe Anual de la
Estaction Central Agronomica de Cuba, pp. 244—247, Lam. 43—46. 1906.

Costerus, J. C. and Smith Jr., J. J., Studies in tropical teratology. Ann. Jard. Bot.
Buitenzorg, 13, 97—118, 1896.

Daniel, L., Jardin 18, 268—270, 276—278, 1904.

Darwin, C, Origin of Species, 6th ed. London, pp. 1 — 538, 1872.

— Variation of animals and plants under domestication, 2 vols. 1868.
Darwin, F., The life and letters of Charles Darwin, 2 vols. London, 1887.
Darwin, F. and Seward, A. C, More letters of Charles Darwin, 2 vols. London, 1903.
Davenport, C. B., Inheritance of characteristics in the domestic fowl. Carneg. Inst.

Pub. No. 121, 1—100, PI. 1—12, 1910.
DeVries, H., Over de Erfelijkheid der Fasciatien. Bot. Jaarboek 6, 72—118, 1894.

— Sur la culture des fasciations des especes annuelles ct bisannuelles. Rev. Gen. de

Bot. 11, 136—151, 1899.

Studies of Teratological Phenomena.

133

De Vries, Die Mutationstheorie. Leipzig. 2 Bd., 1901—1903.

— Species and Varieties, their Origin by Mutation. Open Court Pub. Co., Chicago,

pp. 1-847. 1906.

— The Mutation Theory. Open Court Pub. Co., Chicago, 2 vols, 1909—1910.

East, E. M., The Mendelian Notation as a Description of Physiological Facts. Am.
Nat. 46, 633-655, 1912.

— Inheritance of Flower Size in Crosses between Species of Nicotiana. Bot. Gaz. 55,

177—188, PL VI— X, 1913.
East, E. M. and Hayes, H. K., Inheritance in Maize. Conn. Agr. Exp. Sta. Bull.

No. 167 and Contrib. Lab. of Genetics, B. I. H. U. No. 9, pp. 1—142, PI. 1—25, 1911.
Emerson, R. A., Genetic Correlation and Spurious Allelomorphism in Maize. Ann.

Rpt. Nebr. Agr. Exp. Sta. 24, 58—90, 1911.

— Inheritance of Certain "Abnormalities" in Maize. Rpt. Am. Breed. Assoc. 8,

385—399, 1912 a.

— Getting Rid of „Abnormalities" in Corn. Ibid. 8, 400—404, 1912 b.
Engler, A., Syllabus der Pflanzenfamilien. Berlin, pp. I— XXII + 256, 1909.
Gerarde, J., The Herball of Generall Historie of Plantes, London, 1597, 1st ed., see

pp. 323—325. See also pp. 1220 (Fig. 3) —1221, ed. of 1633.
Godron, A., Melanges de Teratologic Vegetale. Soc. Nat. des Sci. Naturelles Cherb.

Mem. 16, 81—127, 1871—72.
Goebel, K., Organography of Plants. Clarendon Press, Oxford, 2 vols, 1900 — 1905.
Gregory, R. P., The abortive development of the pollen in certain sweet -peas

(Lathyrus odoratus). Proc. Cambridge Phil. Soc. 13, 148—157, PI. 1—2, 1905.
Hasselbring, H., Types of Cuban Tobacco. Bot. Gaz. 53, 113—126, PI. IV-X, 1912.
Hus, H., Fasciation in Oxalis crenata and Experimental Production of Fasciation.

Rpt. Mo. Bot. Gard. 17, 147—152, PI. 17—19, 1906.

— Fasciations of Known Causation. Am. Nat. 42, 81 — 97, 1908.

Hus, H. and Murdock, A. W., Inheritance of Fasciation in Zea mays L. Plant

World 14, 88—96, 1911.
Johannsen, W., Elemente der exakten Erblichkeitslehre. G. Fischer, Jena, pp. I — VI.

515, 1909.

— The Genotype Conception of Heredity. Am. Nat. 45, 129—159, 1911.
Kajanus, B., Uber einige vegetative Anomalien bei Trifolium pratense L. Zeitschr.

f. Abst. u. Vererb. 9, 111—133, 1913.
Kerner, A. and Oliver, F. W., The Natural History of Plants. London, 2 vols, 1902.
Kidd, H. W., On fasciation. Science Gossip 19, 196—198, 1883.

Klebs, G., Uber kiinstliche Metamorphosen. Abh. naturf. Gesell. Halle, 25, 134,
1903—1906.

— The Influence of Environment on the Forms of Plants. Darwin and Modern Science,

Cambridge Univ. Press, pp. 223—246, 1909.
Knight, T. A., On cultivation of Cockscomb. Trans. Hort. Soc. of London, 4, 321, 1822.
Knox, A. A., Fasciations in Drosera, Ibervillea and Cecropia, Torreya. 7, 102 — 103, 1907.

— Induction, Development and Heritability of Fasciations. Carneg. Inst. Pub. No. 98,

20 pp., 1 Textfig., 5 pi., 1908.

134

White.

Kraemer, H., Some notes on the modifications of color in plants. Science, n. ser. 29,
828, 1909.

Lamarliere, L. G. de, Sur la production experimentale de tiges et d'inflorescences

fasciees. Comptes Rend. 128, 1601, 1899.
Lock, R. H., Present State of Knowledge of Heredity in Pisum. Ann. Roy. Bot.

Gardens, Peradeniya, Ceylon 4, 92—111, 1908.
Lynch, L, The Evolution of Plants. Journ. Roy. Hort. Soc. 25, 34—37, 1900.
Makousky, A., Verh. Naturf. Ver. in Briinu, 3, 19, 1864.

Masters, M. T., Vegetable Teratology. London, pp. I— XXXVIII + 534, 1869.
Maiden, J. H., Personal letter. 1913.

Mendel, G. J., Versuche iiber Pflanzen-Hybriden. Verh. Naturf. Ver. in Brtinn, Bd. 10,
Abh. p. 1, 1865.

Molliard, M., Hypertrophic pathologique des cellules vegetales. Rev. Gen. de Bot.
9, 33, 1897.

— Cas de virescence et de fasciation d'origine parasitaire. Rev. Gen. de Bot. 12,

323—327, 1900.

— Comptes Rend. Acad. Sci. (Paris) 139, 930—932, 1904.

Moquin-Tan don, A., Elements de Teratologic Vegetale. Paris, pp. I — XII-}-103, 1841.
Nestler, A., Untersuchungen iiber Fasciationen. Oster. Bot. Zeitschr. 44, 343, 1894.
O shorn, H. F., The Continuous Origin of Certain Unit Characters as Observed by a

Paleontologist. Am. Nat. 46, 185—206, 249—278, 1912.
Owen, Fasciated medlars. Gardener's Chronicle 23, 112 — 113, Fig. 21, 1885.
Penzig, O., Pflanzen-Teratologie. Genua, 2 Bd., 1890—94.

Petch, T., The physiology and diseases of Hevea brasiliensis. London, p. 240, 1911.
Pulney-Andy, S., Branching Palms. Trans. Linn. Soc. 26, 661, 1869.
Reed, T., On Morphology and Physiology of Fasciated Seedlings. Ann. Bot. 26.
388—402, 1912.

Sachs, J., Physiologische Versuche iiber die Keimung der Schminkbohne. Sitzungsb.

d. k. k. Akad. d. Wiss. in Wien, 37, 57, 1859.
S etch ell, W. A., Studies in Nicotiana I. Univ. Calif. Publications in Bot. 5, 1 — 86,

Pis. 1—28, 1912.

Shull, G. H., A New Mendelian Ratio and Several Types of Latency. Am. Nat. 42,
433—51, 1908.

Sorauer, P., Handbuch der Pflanzenkrankheiten. 1, 334, 1906.
Starr, F., (See Trans. Acad. Sc., St. Louis 9, XX), 1899.

Vasey, G., Fasciation in Sophora secundiflora Lag. Bot. Gaz. 12, 160, PI. X, 1887.
Worsdell, W. C, Fasciation; Its Meaning and Origin. New Phyt. 4, 55—74, 1905.
White, O. E., The Bearing of Teratological Development in Nicotiana on Theories of
Heredity. Am. Nat. 47, 206—228, 1913.

— Studies of teratological phenomena in their relation to evolution and the problems

of heredity. I. A study of certain floral abnormalities in Nicotiana and their
bearing on theories of dominance. Am. Journ. of Botany, 1, 23 — 36, Figs. 1 — 4, 1914.

— A new cytological staining method. Science, n. ser. 39, 394 — 396, 1914.

Studies of Teratological Phenomena.

135

Tables 1 — 26.

Explanatory note. In the majority of these tables, the various races are
cited by pedigree number rather than by name. The number and name are both given
in the body of the paper under "materials and methods". Tables 21, 22 and 25,
representing the character of certain F2 populations, were tabulated from selected plants
and not taken from plant after plant as they stood in the row. These selected plants,
however, represent practically all the various types that appeared. In all other tables
of the same type, the plant numbers represent the plant populations as they stood in
the row in the field-plots.

These tables are so constructed that a definite unprejudiced conception of the
variableness of the expression of the fasciation factor may be easily gained. Take, for
example, the character of plant number 1 in Table 6. 25 flowers were examined;
15 had 5 sepals, 9 had 6 sepals and one had 7; 8 had 5 petals, 11 had 6 petals and
6 had 7 petals; 10 had 5 stamens, 10 had 6 stamens, and 5 had 7 stamens; 12 flowers
had 2 ovary-locules while the remaining 13 had 3. This plant was 19 '0 dcm. high at
maturity and had 27 leaves. The figures within circles ((14)) are floral leaf class range
numbers placed thus to avoid lengthening the table. In tabulating the inheritance of
calycanthemy, O'O represents absolute absence of this condition, 0*5, its presence on a
portion of 1 sepal, l'O, 2'0, 3'0, its presence on 1, 2 and 3 sepals.

136

White.

Table I. Materials.

Designation

No. of
Indi-
viduals

Year

Grown at

Original Mutant (300)

1907

Alquiza,

Cuba.

Dewey's Nos. 1—99 (E. 300—309)

1908

No. Bloomfield, Ct.

East1** Nos 300 309

148

1909

ii >j

307 V 13 2Q (4.02^

r i

1909—10

Greenhouse, B. I. H. U.

1909—10

Pure strain (X) Fasciated, No.

iOSI

1909 — 10

901 1 f^Jnrmnl eplpptiAn^

Ovl X yXlUIiliCll oclCLLlUi-ly

1910

B. I. H. U. Exper. Plots.

ouo — i ^^xunoinicU .. j

1910

ii ii

324 X 301

1910

>■>

ii »

30J. V 109

1910

ii ii

f90J. V 409^ 1

1910

n n

C301 V 339^

1910

n n

fq9i v 301
^ooi /\ ouiy

1910

ft »

J.09

1910

ii »

399

1910

ii n

373

191U

ii n

9^3
OOO

1910

ii »

901 1

oui — 1

1911

n »

903 1

ouo — 1

1911

11 n

oui — i — z ^oeieciiou exper. j

1911

» n

QA1 1 OQ

1911

» r>

Q AO 1 in

ouo — 1 — i-Z „ „

1911

ii ii

OUO — 1 14 „ „

1911

ii 11

A AO
4U—

124

1911

No. Bloomfield, Ct.

/"9n i v ^^o^ i

\^OU-± /\ 4U^j — 1

1911

B. I. H. U. Exper. Plots.

„ „ — l— o net.

1911

ii »

( » >5 ) — 1 — 7 Norm.

TT
*s

1911

ii n

( „ „ ) — 1—8 Abn.

1911

ii n

( „ „ )— 1— 10 Norm.

1911

n n

( „ „ )-l-12 Abn.

1911

n w

( „ „ ) — 1 — 14 Norm.

108

1911

n ii

( „ „ )— 1— 28 Norm.

109

1911

ii ii

( „ „ ) — 1 — 31 Norm.

104

1911

ii n

( „ „ )— 1— 34 Het.

108

1911

ii »

301—1—1 X 353—3

3ft

1911

ii n

353—3 X 301—1

1911

ii »

301—1—5 X 373

1911

ii ii

303—1—24 X 332

1911

ii ii

303—1—13 X 327

1911

n ii

402

1911

n 11

Transport

1797

Studies of Teratological Phenomena. 137

Table I. Materials, continued.

Designation

No. of
Indi-
viduals

Year

Grown

x i diiopui L

1 7Q7

X 1 V t

OOo

1 Q1 1

lu XX

B. I. H. U. Exper. Plots.

373

1 Q1 1

1311

11 11

VJl CCilllU UoC Mai VCtLlUIX LulLLllCo

1 UULO BUI VIVUIB

1 Q12

B. L H. U. Greenhouse.

1 Q1 2

X <J1^

B. I.H.U. Exper. Plots (Poor soil).

QUI X

1912

9fl3 1

OyJO X

1912

» i?

QflJ. 1

1912

X . ' 1 —

n »

jno i

1 Q1 2

X Jl_

ii ii

379 1

1912

n ii

353 i

1912

» ii

(lf)9 1 V 303 1 .3^ ~F

1912

ii ii

(R04. V 402"> 1 fi 2 Hpf F

1912

ii ii

( ^ 1 7 in Norm F

^ ^ ,j ) X < XU i.1 UI 111. Jl. ^

1912

ii »

\ 11 „ ^ X lu OU xN 0 VJLU. J? ^

1 Q1 2

X X —

ii ii

C ^ 1 1Q 3Q Ahn F

^ 11 ii J X X^ OO AU11. X. ^

1912

ii n

*(304 X 402)— 1—28— 32 Het. F4

1912

ii ii

301—1—1 X 353—3 Fx

1912

ii ii

(353—3 X 301— 1)— 12 F2

227

1912

ii ii

(301—1—1 X 353— 3)— 7 F2

1912

ii ii

301—1—5 X 373 Fx

1912

ii ii

(301—1—5 X 373)— 17 & —5 F2

269

1912

ii ii

Total

3043

Fasciation X calycanthemy

396

1910

B. I. H. U. Exper. Plots.

396 X 402

1910

ii n

396

1911

ii ii

(396 X 402)— 1

1911

ii ii

396—1

1912

ii ii

301—1—2—5 X 396-1

1911—12

B. I. H. U. Greenhouse & Exper. Plots.

301—1—3 X 396

1911—12

ii n

ii 11

301—1—1 X 396

1911—12

ii ii

11 11

396 X 301—1

1911—12

ii n

11 11

396—1 X 303—1—200

1912

ii ii

H 11

(396 X 301— 1)— 10

135

1912

ii ii

11 11

(301—1—2—5 X 396—1)-

11 F2

233

1912

n ii

11 11

(301—1—1 X 396)— 12

109

1912

ii ii

n ii

Total

724

138 White.

Table 2. Frequenc}^ distribution

Designation

Year
Grown

Grown at

No.
Plants

26 27

29 30

32 33 34

uiiginai Jiiutani ^ouuj

iyu /

Alcjuiza, Partidos, Cuba

jjewey s jnos. i — yy

1 QflQ

(Shaded) Bloomsfield, Conn.

i^ast s in os. oui — ouy

iyuy

» » >»

i /ie

140

East's Plant X Progeny

B. I. H. U . Greenhouse, Mass.

o
_

oui — 1

1 Q1 0

JJiXptJl llucULdl X 101S, X>. X. XX. U .

• J —

o
—

ouo — 1

1 Q1 0

iyiu

" " "

1
1

oui — i — _ — o

1 Ql 1

Lull

QAQ 1 19 1 Q1 1

ouo — 1 — X& — iyn

1 Ql 1

I I I 1

•i

oUo — 1 — ZUU

1 Ol 1

uii

(304X402) -1—12

1911

(304 X 402>-l— 8

1911

11 11 n

(304 X 402)-l-12-38

1912

304—1

1912

11 11 11

301—1

1912

Table 2. Frequency distribution

Designation

Year
Grown

Grown at

No.
Plants

70 71

7(5 77

Original Mutant (300)

1907

Alquiza, Partidos, Cuba

Dewey's Nos. 1 — 99

1908

(Shaded) Bloomsfield, Conn.

East's Nos. 301—309

1909

» » a

148

East's Plant X Progeny

1909—10

B. I. H. U. Greenhouse. Mass.

301—1

1910

Experimental Plots, B. I. H. U.

303—1

1910

ii ii ii

301—1—2—5

1911

303—1-12—1911

1911

" " 1

303—1—200

1911

I I I J

(304 X 402)- 1—12

1911

n ii a

(304 X 402)— 1—8

1911

n ii ii

(304 X 402)-l-12-38

1912

ii n }i

304—1

1912

n ii ii

301—1

1912

n ii ii

Studies of Teratological Phenomena.

of number of leaves per plant.

139

36 37 38 39 40 41 42 43'44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68

4 6

of number of leaves per plant (continued 1).

79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

2 4

102 103 104 105

106 107

108

140 White.

Table 2. Frequency distribution

Designation

Year

Grown at

No.

ioy

nil
112

113

114

Grown

Plants

llOjlll

Original Mutant (300)

1907

Alquiza, Partidos, Cuba

Dewey's Nos. 1 — 99

1908

(Shaded) Bloomsfield, Conn.

East's Nos. 301—309

1909

n n n

148

East's Plant X Progeny

1909—10

B. I.H. U. Greenhouse, Mass.

301—1

1910

Experimental Plots, B. I. H. U.

303—1

1910

» n n

301-1—2—5

1911

303—1—12—1911

1911

>5 » » |

303—1—200

1911

n n »

(304 X 402)— 1 — 12

1911

. n » n

(304 X 402)— 1—8

1911

» n n

(304 X 402)-l-12-38

1912

» » »

304—1

1912

n n n

301—1

1912

n n n

Table 2. Frequency distribution

Designation

Year
Grown

Grown at

No.
Plants

139

140 141

142

143

144

Original Mutant (300)

1907

Alquiza, Partidos, Cuba

Dewey's Nos. 1 — 99

1908

(Shaded) Bloomsfield, Conn.

East's Nos. 301—309

1909

» » 5)

148

East's Plant X Progeny

1909—10

B. I. H. U. Greenhouse, Mass.

301—1

1910

Experimental Plots, B.l.H.U.

303—1

1910

301—1—2—5

1911

303—1—12—1911

1911

':':':]

303—1—200

1911

(304 X 402)— 1—12

1911

» >j n

(304 X 402)— 1—8

1911

(304 X 402)-l-12-38

1912

» n n

304—1

1912

» » n

301—1

1912

n » n

Studies of Teratological Phenomena. 141

of number of leaves per plant (continued 2).

115

116

117 118 119

120 121

122

123

124

125

126 127

128 129 13o| 131 132 133

134

135

136

137

138

of number of leaves per plant (continued 3).

145

146147 148

149

150 151

152

153

154 155;156 157 158jl59;160

161 162jl63

164

165

166

167

142

White.

CD
—

CD
>

Average

rr -t x co
O CM — X

t- X X ^

b- t* CO

i> -r

F b

5.19^0^ -ox

0S8

GO CD l> ^
tD H IO i

10 co co oi

EC 3? CO CM

Ci O CM — -
lC l> fc» -T

Floral leaf classes

iO
CM

-t

CO
CM

CM
CM

r— 1

Ol — — 1

Ol Ol

CO
t— I

01 CM

CM CM tH

CM CO

„ CM

Tl*

co 10

T O CO

T— 1

CO — ! —

y— 1 CM CO

CM CO —1 CM
CM CO

T— 1

Ci uO — 1
CM

x x co

O
1— I

CO CM t» »— 1
CM CM

CO lO CM CO
iO iO

X CM CM

1— 1 iO rH iQ
CM "^t X 1—1
1— 1

CO co iO X

— 1 co 10

— CM — 1

~t x co

i—i CM — CO
CM CM CM

—1 co 0 co

l> Ci CO —

co co

-r uo co
cs C!
CO CM CM

ic co Ci -t
C: CO X CO
CM

Ci t* fc» t»

05 t-»
rH

A IO iO >G
CM CM

CO X
CO —
CM

Kind of
Organ

Sepals
Petals
Stamens
Ovary-locules

Desig-
nation

1—108

T — 80S

Studies of Teratological Phenomena.

143

Table 51).

Frequency distribution of floral abnormalities per flower per
plant in two generations of Nicotiana t. calyciflora.

Progeny of 396

— 1

Plant No.

396—1

396—2

Sepals

r>
D

Petals

6
7

Stamens

af class

6
7

Ovary -locules

Floral

O'O

1-0

Calycanthemy

2-0

3'0

4-0

x) Table 4 next page.

144

White.

r-j

s — '

b£

•<s>

«H-H

•<s>

•i— i

+=>

"r-l

=—

«H-H

bJD

CO
CO

m in co

1-1

co co o

co co in

CO i-l

CM t- CM

CO in

t- 00

mo©

-* t-

CM CO O

rH t-

00
CM

OS CO CM

O
CM

CO CO CM rH

CO CM i-l

CO
CM

rH t- in

O CO ^ H

rH as as

■r*

CM iH

m as

as in —

CO
CM

CO O OS i-i

CM CM O

N W « rt

rH O 30

CO r-l

CM lH

H W W

CO CO CO

rH CO CO

rH O rH

to -# m

o t-

CM rH

CM tH CM

m m in

rH rH

rH CO in

Tjl N M rl

in rH

CM CO CS

CO CM

co co m

m o

t- CM CO

i— 1 i— <

OS rH CM

H OS

r-l rH

m co cm

rH rH CM

CO O CO

t- cm m

rH CO

cm m

CM CO HO

rH as

CO CO

CO rH

H H 90

rH rH

CO -*

rH CM CM

CM CO -r rn

OS rH rn

CM rH

co co m

t- O H

CO CM

00 Of.

m c- co

a»

rH m in cm

in m co cm

CM CO
CM

No.

-* m

CO !>• 00 OS

c —

n in

t—

OS O rH

CM CO

m co t~

tj* in co t— oo

la nt

Studies of Teratological Phenomena.

145

t- o n

n o -

CM
CO

o
~

own'

cr.

to
©

cm —

R 0 (D H

«*

cm in

i-H CO ^H

-H<

CO
CO

to
©
co

CO CO CO i-l 1-H rH i-H

OJ
CO

o
b

N 51 (i) rt

O O CO CO i-i

00
CO

CO
CO

an o

CO X

so
eo

eo
o

cm ~*

IS N O CO

10 tO CM CM

Ifl - t- H H

rH CO X CM

at
co
co

o IN x

CO
CO

A ^ M N rt

i-l CO CO kO <M i-l

to
b

CO ^ CO i-l SO

to
eo

^< co a>

o
eo

\a
t-

CO 1-H

CO 00 CM CM

to
eo

9
b

^ fll H H H

— CO

CO « <«

01 N H i-H CO

CO i-l lO CO i-l i-i i-l

SO i-l

CO 00 »-« SO iH
i—(

h 0 n

to
eo

rH OJ

rt N (• H H « i-l

CM i- i-H

Cl
CO

CO —

i-l Ci © rH CM SO

94 rH

i-i to irt CM rH

1ft

00
C".

N H N H iH

eo ^ so ~h

t- ; h h

CO
eo

IO 00 H

0>n««i(ie«t»»QOrtei»>i(

«4rHrHrHvHr*rHr4r*»HeMeMeMCl r,

N « Li (O f 30 C C — "UJS^hh51?

Leaf Count

Height
(dcm.)

suaniBjg s8iTK>oi-£reAO

Induktive Abstammungs- und Vererbungslehre. XVI. 10

Xfl

H-3

*rH

T— 1

1
1

v — '

b£)

• rH

•<s>

«2.

ten

•ri

ber

• <>*

•i— i

•rH

+->

abi

[an

Cr-I

<X>

ph!

<H— 1

0)0(0

<*5 t H< CO rH

eo CD m

H « O H rH

rH O0 00 in rH rH rH

/cm\

eo c-

rl tO (O H rH

VCM J

h co tji in cm 1 — '

CM H" rH rH

rH CO 00 CS rH

cm — © co

CO rH OS rH

CO
CO

H N ^ t" H

rH CM t- CM rH CM

rH O O

"* CO CM rH

rH CO m rH rH rH

rH H< rH in

CO in 00 OS

CM Ot rH CO

co t— in

CO © CS

rH CS © CM rH CM

rH * O CM

CS
CM

t- ej» th

rH UJ CO CM rH CM

rH t- CO

00 CO CM rH »H

CM til Oi CD -H rH rH

/oo\

h* O O CM

t- -T H*

V— .J

rH 0 <N w

in rH OS

CO
CM

CO © CO H< rH

inoco^eM oicMrH

m CO CM CM

h* m co

cm t- m -h

CO CS rH

H<
CM

^ 00 CO ■* i-( CM

H N N rf CN irtrH rH rH rH rH

/ScTy Os\

cm eo cm

CM If) 00 rH

IrH A-H J

00 CO rH rH WW

rH 00 -HH rH

CM
CM

cm CO eo

•* CS I- CM rH CM

eo © cm

CO ® (M

irtCDHIinCMrHrHrH

in co in cm

O
CM

m -f o

co co eo eo

rH rH CO

HI CM CO (M rH

OS CO -* H< rH rH

CM 00 CO

<M t- CO CO

I* O CO H w

•n cm m cm

-h CM rH.rH

00 in rH rH

m oo cs

CM t— m rH

H< O CO CO rH rH

CO (M t-

CM O i-l CM

rH rH rH 00 rH rH

rH rH rH CO in

CO HI t- rH

CM CM CM O CO in rH

CM rH CM

Ol H m H

rH-^COeOeOCOCOrH rH

-H h* CO CM

i-i co in co

» a co

rH CM t— CO

CM t-» CO

OS CO If} m rH rH r-l

t- ■* CO

eo o oo

rH t- © ^H rH rH rH

CO CD CO

CM CO CS CO

rH CS rH CM rH rH

rH (M 00 t- H<

CO CO CO

CO *# eO CM rH CM

rH co eo co rH rH

O hJ( m Hj< rH rH

rH O m CM

lrt

cm co cm in

rH O O t- r- rH

rH 00 00 CM

rH CM CM C5 rH

CO rH h* 03 rH

rH OS © rH

-# H< -h h* CM

rH CM CM 00 CM

(M rH CM 00 -H

CM rH CO CO

CM CO CO CO CM CM

(M rH in rH

CO t- (M CO

rH CO 00 O CM rH

H C. t-

Plant No.

■*lfl(Ot»XeiOHIM

Tfiincot-aocsOrHCMeorjiinco t-

— O cc t- x —

S9SllBI ssup

CO CM •"»< rt

x>
>b

R r4 iH »H

(£T)

05 CM w

■* O ■* « to N

ifS

lO
iH

o
b

■f « o

■£
t-

■* N m CM CM

^ ^ ^

i-i O CM tH

CO
CO

lO tH O)

N * S H H H

H H M —

M C W CO H i-H

CM 1-1

O O CM CM

CM
CM

X 05 CC

iH CO <-l

«4 fH

i W 5) H i-l

N CM CM I-H ~H

«Meo-*eMco«i-imiHiH i-i tm -h

_ _ (g)

<E>

CO CO

H ffl lO N CO i-l

eo
ill

e h it

^ lO rt if J! N rt CM

iH CM CO i-H

t- QO r( i* ^1 H

— — -f -H ,-1

i-( rc cm iO i-H i-i —

— (f)

f » H i-H

iH CM

«* X i-H <N

t- ^

CO
O)

~. « -

-H t- (N CO CM

CM
CM

as ao i-H

coeMcot-cMi-ii-i —

,_,

^ «

n eo to « s ■* « — «

•<*< <0 ^ CM ^ —

* 16 N

w te * n io «q fi —

C H N H

■«* as -f eo —

jH CM

e oo t- n h h

Kan

9 * « <a

00 ^1 t- 93 H H »H

O CM — i-H —

- CM rH

N H »• ^1 H

irt «D

— W CM CO CM i-i

CM « tH

O O CO *H iH

£ — rieo-*iOVOt-=000'-i(McO'*<
— i-i -1 — — — ^HrH^H^HCMMCMeMCM

M«itisehi£noHMn<«io^t>oocio

1-4t-(1-|i-<,H»-<iHtHi-Ii~ICM

Leaf Count

Height
(dcm.)

10*

148

White.

Table 6. Range of variability in height, number of floral leaves

Fi generation of (304X402) and

Plant No.

1 K

o
&

Q
O

rr
/

1 Q

i (\
10

1 a
lb

1 n
1U

rr
(

1 o

1 A

1 Q

Q
O

1 o

15a

-i
1

o
O

1
1

Sep

10
11

af class ranges

9
10
11

Stamer

9
10
11
12

o
o

Ovary-

5
6

Leaf Count

Height
(dcm.)

19-0

19-8

20-6

19-0

190

19-8

20-6

19-8

21-4

21*4

20-6

206

19-8

19-0

198

Studies of Teratological Phenomena.

149

per flower per plant, number of foliage leaves per plant in an
reciprocal (402 — 1 X 303 — 1 — 35).

I1)

21)

4
1

3
1

12
10
1

1
1

12
1

4
(1

?>1

10
6
8

11
1

7
1

5
(1

8
6
8
2

17 19

13
6

. 4
Pi

19-8

21-4

19-8

21-4

20-6

21.4

21-4

221

21-4

22-1

21-4

229

18-3

22-1

190

22-1

21-4

19-8

18-3

214

J) Reciprocal.

150 White.

CD
CO

— —

IQ
Ol

iO
CO

CM
CM

„

Ol
CM

— CI
- —

—

01 Ol

a n (I

DO
CO

—

l> rH

CM iH

—

"+ CM

•

1 «■

-r

CO —

Ol
Ol

1"*

-t

iO
CM

-t

—
—

! O

-t

Ol
01

CM
Ol

tO
Ol

Ol
Ol

X
Ol

tO
CM

—

CO
Ol

i Ol

-f

>
C

CO
Ol

^ —

lO
CM

CO
01

-f

S-

1 eg

CO CM

-f

i—i

b-
Ol

iO
CM

tC
Ol

o
=

CM
CM

lO 01

Ol
Ol

tO
Ol

"-r
CM

Ol
Ol

» Ol

iO
CM

Ol
Ol

i-H

0Q
Ol

i o

§3

iO
Ol

< —

CD
OX

CM
CM

tO
CM

CM DC
CM

CM
Ol

1
1

tO
Ol

DC —

1— t

X rH

-f

tO
Ol

01 —

t- CM

CO
Ol

• -

CM
Ol

iO
Ol

O
Ol

H/
CM

• —

1 s

CO i—i

Ol CO

-t

93
—

—

Ol Ol

-t

1—1

T* iH

~* —

»o

i-H

X —

X i-1

X —

IO
Ol

-t

.m:a Ji

-r

-T

iO
Ol

Ol
01

-r —

CM
Ol

tO
CM

—

X rH

—

uO
Ol

Oi
CM

Ol
Ol

be
=

—

I> H

X
CM

CO
Ol

—

X
01

Ol
01

Ol —

lO
CM

X
Ol

co t*" X

"ic IT

—

He i

■a s

X 2

Studies of Teratological Phenomena.

151

—
—

-r^

—

"—

^ ^

CO
1

—

i—i

1— 1

tt-H

«rH

N — '

"—

i ii

h->

•iH

•i— 1

•—

OJD

CO © CO

io © *h

rH © rH

00 01

Cq 01 rH
1—1 1— 1

o co oq

rH rH

CI CO

rH rH

■H/ rH

CO
CO

oq UO

rH ^

lO)

H rH

CD X rH

-J.

1—1

rH" ^

X t*
rH

-t rH

CO
CO

5q £q _
i—l i—l

t* <qi

i— i

co oq
oq

rf rH

»o
cq

Oq ^

oq co
cq

iO
Ol

co oq

rH i— 1

rH i— I ^

X CD rH
rH

Oi
(M

O i—l

i— i i— i

X CO

oo uo oq

cq co
oq

00
Cq

— - q

h i— i

O rH ^

O ^ rH rH
i— 1

»o
oq

CO l-H

t—

— en m

r~ 1

1—1

io x oq
—

os oq

1—*

i— i

rH lO 00 rH

i—l

oo io oq
i—i

rH © IO

cq co
oq

»o
cq

01
Cq

H 2

rH rH

lO o
i— i i— (

in
cq

i— 1 Tt^

i—l Tf<

co oq
cq

cq *°

5h ^

cq co
cq

— rH

cq co
cq

1— 1

O Tf< t— 1

oq °

CO X rH

rH 2

t>- t>» rH
rH

oq u"

O lO

in
cq

lO CO rH

iO ^

co oq
oq

JO © rH

~t X CO

CD X rH
i— 1

>o
oq

i-H

^ t oq

CO X rH

oq co
oq

t"» co oq

r^ oq oq

CO rH

in
cq

i— (

oq © ^

CO CO CO

o cq co

rH i— 1

— rH

X CO rH

oq 00 1-1

rH CO rH

»o
cq

CD X rH
rH

x io oq

rH rH

m
cq

co co co

t> CO »o

tJ< i— i

fc-

t— co oq
i—i

t"» lO CO

cq co
oq

X t—

t*~ X

in
oq

CD X rH

X © rH

in
Cq

^ rH

co oq

"t i—i
cq

OS CD

co oq
oq

oq co
oq

co oq
cq

co cq
oq

»o
cq

X o
1—1

oq co
oq

Plant No.

rjt *a co co

iO CO t- X

■h/ iC © C- X

cq co

S9|noo^

-JLlTSAQ

152

White.

Table 8b. Reciprocal of the Fi

Plant
No.

oi
-i

1 Q

1 4
14

i k

i e

iSl

i ft

1 Q

1 A
14

1 7
1 /

1 A
14

i ft
lo

1 Q

i y

1 9

9fi

i a

—

7
1

Q
O

1J.
14

1 3

Q
O

1 fi
ID

1 1
11

3
1

Petal

E class range

6
7

Q
D

9
l

o
_

10
i

5
i

4
1

11
i

i
l

i
i

il leal

91
_ 1

i ft

1 7

1 fi
ID

9J.

1 3

1 ^

1 9
1Z

1 Q

9fi

o
-

9ft

Q
O

1
1

1 9

i n

A
4

1 fi
ID

-t-s

7
8

1
1

03
S

o
o

1— 1

Table 8c. Range of variability in height and number of

Plant No.

Leaf Count

Height (dcm.)

122

14-5

13'0

14-5

15-3

14'5

15*3

16-8

17'5

15-3

14-5

Studies of Teratological Phenomena.

153

cross tabulated in Table 8a.

leaves per plant in an Fi generation of (301 — 1 — 1 X 353 — 3).

21 22

16-8

14-5

17-5

16-8

16-0

16*8

17'5

17-5

19-0

16'8

17-5

16-0

16-8

16-0

154

TV hite.

Table 9. Range of variability in height, number of floral

plant in an Fi generation

Plant No.

1 1
ii

1 Q

Q
O

1 o

o
O

Sep;

8
9

1 A
1U

i
l

o
_

1
1

Q
O

o
&

i
i

Q
O

1
1

Q
O

■x>

6
7

o
21

1 1

A
•1

7
16

1 o
10

Q
O

7
/

1 o
1Z

1 A
1U

Floral

Q
O

o
_

i
i

fi
o

8
9

ooules

<— i

>>
u

Ova

Leaf
Count

Height
(dcm.)

30-o|28-2 28-2 27-5

27-5

25-927-5

275 27-5 27'5

27-5|23-625*2

26-727-526-725-027-5

25-926-7

x) Poor soil. — Nos. 19 — 40. Plants tabulated later in the season. —

Nos. 41 — 42. From

Studies of Teratological Phenomena.

155

leaves per flower per plant and number of foliage leaves per
of (303 — 1 — 24 X 332).

— '

2
1

.11

(

2
1

19
1

13
1

14
1

13
1

26*7 26.7 27*5 25*2 27"5 24*4 28*2 25'9 25*9 25'9 28'2 27'5 28'2 28-2I27-5 28*2!27'5 22*9 19'8 22*9

28-2

27-5

a different cross of these same species, grown in 1911.

156

White.

Table 10. Range of variability in calycanthemy and number of

families of (396 X 301 — 1)

Plant No.

ll1)

12s)

Sepals

6
7

3
1

5
1

13
1

2
1
1

3
1
1

2
1

Petals

5
6
7

l
i

4
o

5
1

2
o

5
i

<j
2

31
1 0

a
\j

5
9

K
O

o
—

A
*±

eaf class ranges

6
7

4
1

38
3

22
4

5
1

2)i

5
1

Ovary -locules

Floral L

2
3
4

5
1

3
2

3
3

38
3

19
7

5
1

5
3

4
2

0-5

Calycantln

1- 0

2- 0

3- 0
4*0

1
1

2
1

1
4

4
2

2
4

3
2

4
1

12
1

3
1

3
2
1
1

2
4

4
1

3
1

J) 1 flower selected at random from each of 41 field-grown plants.
2) Field-grown plants.

Studies of Teratological Phenomena.

157

floral leaves per flower per plant in 5 different Fi generation
and reciprocal (301 — 1 X 396).

41")

42*)

432)

Q
O

o
_

1
i

o
-

i
i

o
—

9
-

9
_

i
i

o
O

(

158

White.

Table 11. Range of variability in number of floral leaves per
flower per plant in an Fi generation of (303 — 1 — 13 X 327).

Plant No.

Sepals

CO
CD

6
7
8

1
1

i n

D
1

i
i

fl
d
Fh

a
D

r
0

-i
1

Petals

eaf class

6
7
8

6
6

5
9

7
7

Stamens

Flor*

6
7
8

6
4

4
1

5
2

2
9

9
5

5
8

1
2

Ovary-

locules

Table 12. Range of variability in number of floral leaves per
flower per plant in an Fi generation of (324 X 301).

Plant No.

Sepals

Petals

range

6
7

4
21

2
22

5
20

3
22

4
20

3
22

6
8

8
17

4
20

1
20

■3

Stamens

Ovary-

locules

5
6
7

4
2

3
1

Studies of Teratological Phenomena.

159

Table 13. Frequency distribution of number of floral leaves
per flower in Fi crosses involving the factor A.

Pedigree

Kind of

Floral leaf classes

Total
of flower

54-1 50

o S

33
O

Ave.

No.

Structure

i | 4

5 6

10 11 12

parts

c *

324

Sepals

234

2225

6-85

Petals

275

2253

325

6-93

301

Stamens

251

2267

6- (J7

Ovary-loc.

143

139

1163

3-58

304

Sepals

431

411

158

5934

5-79

Petals

168

458

374

6422

1025

6-26

402

Stamens

196

445

352

6397

6-24

Ovary-loc.

268

717

2869

2-80

303—1—24

Sepals

286

428

276

6203

6-05

332

Petals
Stamens

121
151

325
336

526
488

52
44

6657
6628

1025

7
7

6-49
6-47

Ovary-loc.

592

419

2497

2 44

303—1—13

Sepals

585

5*62

Petals

643

104

6-18

327

Stamens

634

6-09

Ovary-loc.

238

2-29

301

Sepals

197

1596

537

396

Petals

204

1589

297

5-35

Stamens

281

1509

5-08

Ovary-loc.

257

i
i

636

2-14

301—1—1

Sepals

561

256

4560

5-36

X
353-3

Petals

561

244

4584

850

5-39

Stamens
Ovary-loc.

833

611

211

4515
1717

5
2

5 " 31
2-02

353—3

Sepals

700

300

5470

5-34

Petals

739

262

5430

1025

5 30

301—1

Stamens
Ovary-loc.

1009

785

223

5371
2066

5
2

5-24
2-02

301—1—5

Sepals

707

173

4715

5-24

Petals

635

234

4798

900

5-33

373

Stamens

699

179

4720

5-24

Ovary-loc.

868

1832

2-04

Total No. floral leaves 108,723 from 5551 flowers from 292 plants.

Table 14. Classification of progeny of Fi, F2 and F3
heterozygotes of the cross (304 X 402).

Designation

Gen.

A A

Total

Theoretical

(304 X 402)-

-1

50 :

(304 X 402)-

-1

75 :

(304 X 402)-

-1—6

50 :

(304 X 402)-

-1—34

109

25 :

(304 X 402)-

-1—6—2

75 :

(304 X 402)-

-1—28—32

50 :

Totals

192

103

393

25 :

196

160

White.

btudies or ler

t(I CO CO

atological Phenomena.

CM CO

161

rt
<

CM
CM

CO CM
CM

a
a

■# Ol N

T-l

O in

CM
CO

rt
<

t- CO « <N r- N

HOOOOCOrlOHH

CO
CO

CO CM
CM

rt
cs

rt
<

O lO
CM

o
b

CO o>

ta o

CM
~*

CM
CM

rjl tH

CM CO

-*
CO

CO
OS

CO
CM

rt
rt

CO
CO

at
<

i< !S •* H

O tH

t- CO CM

O m

CM
CO

rt
<

1ft

o
as

rt
rt

CO H Ol H H

m m ih

m O

I-H TH

CO
CO

t- t- iH

CO t-

rt
<1

00 T-l CN iH H iH i-H

OJ CM r-t H iH tH

<4
<3

rt
rt

CO
CM

rt
rt

i-H CO CO tH CM CM CM CM CM i-t

-(£)

CO lrt •«* t-i in ?H iH tHiHtHtHi-i

(?)

«!
<J

CO Sl CO '

CM CO ' — '
TH rH

CO
CO

rt
<

o
as

CD CO CO

ift o

CM tH CO CO i-t

CO T-i i-l tH

CM
t-

r—

rt
rt

rt
<

■* CO 10

-# © »H
CM

TH t- t-

00 t—

lO
CO

CO CO O — 1 H tH

t*» CO OS tH tH

<4
<]

rt
<

t-i CM CM

00 t~

CM
CO

co co co

00 CO «-l

CM
CO

o
cb

i*mcot-ooo50T-icMeo-*mcot-coGiOTH

NM^iaCOt-MOlOrlNCO^IlOCfl

OJD S

co
co
rt

Induktive Abstammungs- und Vererbungslehre. XVI. 11

162

White.

—

• f— 1

< —

1! t

pfl

bJD

• i— 1

—

i -»

•r- (
p"

C£

c+h
O

S-l

co
to

X (O rH

x m iH rt

Ifi
o

O IM CO

m co

■*
o

in
CM

i-H eo t-i

cm
co

eo eo

in to

o
CO

(VY) IBUiJouqy

•># i-i

00
IO

ri n h

m co •*

tr-
m

O "* H

CM t- CO

m
lO

!C X n

m t- eo

eo
in

CO CM

O X i-H

IN
in

CO X H

CO CM t-

rH =r. * H

CM CO
CM

i-l CO i-l

t< ad co

^ X CO

O CO CM

CO H X

H N f)

x cm m

eo t- m

^ — - cn i— i

O CO CM

«* to «* cm i-i

to i« e n h h cm h
(?)

s s ~

o n f w

rt « 91 K

CO CM t-

05
CO

H O) O t* rl H

0) CO t- r4 CM

C- t- i-l

f h ^ n h

to o co n H d CM H

CO
CO

|H O t- 1-

in x

i d in

CM X O

■*
eo

CO X a CM tH CM

O (• H N l-H i-i i-i

Plant No.

-*ocot-xCiC;»^eMeo-*incot-

^u-COt-XOSO^CMCO^^CO^CCWOjH

Sp3d9g

Studies of Teratological Phenomena.

163

i-i eo —

<
<

00 t-

b
tH

o
b

ao
b

—
<

<
<

c—

1- X

c;
<

its o

ci
<

2 12

ir.
t-

O iO

cS
<

Si 1-1

o
b

lH 00 CO

1-1 iH

1ft

o cm

o
b

cS
<

-1^-1

*** t-i

CO
00

L — _

»-l i— 1

(g)

<
<

— ^JZ ®

t- co «

CO
CO

c"*

30
-

<
<

CO 0>

-(g)

i-l i-t

t- — CM i-l CO ,H

-M

<
<

rt a «

6)—^—

Tit
CO

W ■* L1) «

X O N

x>
b

CO ^ i— 1 H H N H H H

H CM SO H CO H H

cp
b

<
<

n co cst> asaoHnn^ivjco

be 1

suaniBig ssTiiooi-iLiBAO

11*

164

White

©q
—

H«HO
i—l i— (

iO CM t- tH
i—l

i— i 05 co oq
i—i

i— 1 lO OS CO CO i-H

IO
CM

OS
CO

co oo th
i— i

io os i> oq oq

oq o oq th
i— i i— i

CM rH CM

HODlOH
i— i

co oq oo cm

CO
CO

lO
CM

IO
CM

CO rH t» rH

Tji rH

CO
CO

i-i O CO i—l
1—1 1— 1

CO t}1 t> rH
rH

CM
CO

oq ic t- i-i
i—i

O-. rH
rH CO OS CO Tt*

1—1

os co ©q i—i

O rH
rH rH

c- ^

T— I

HO-^H

©q

co oq o
i— i i—i

CM Oq rH

©q

CM OS CO i-l
i—l

rH CO rH
rH

©q

00 CO
1— 1

CM iO 00

oq o oq th
i—i i—i

CM CO

OS CO CO
1—1

rH CO rH
rH tH

oq
oq

1—1

rH 00 ©J CO rH

i—i
oq

Tt< i-l

t rH

i— 1 iO OS
1— 1

oo oo io

H CD t> H
i—l

rH "TJI IO lO

T— 1

rH 00 00 00

HflOt^H rH

H 05 ^ H

T— |

CO 00 OS CM CO

lO
1—1

CO O H H
1— 1 1—1

CO 00 rH
i—l

lO
CM

i— i

00 JO tH
1— 1

l—f rH i— 1 CM
H rH

>o
oq

O CO CM
i—l i— i

CO 00

femionqy

jeuxionqy

r> 00

l> 00
rH

oq o co
i — i —

rH 00 IO rH
rH

CO O 00 t— 1

oq oq rH

iH rH

CO t> io

rH ^

rH rH

i-l ^ o
1— 1 1—1

t- 00

io co r>
i— i

HO^
rH rH

i—i

Plant No.

iOCOt>XOOHCqcC^iOCOI>CC

T^iococ^ooooiHoqco^iococ^ooosor-

Parent

H CO H
i—l i— (

"H oq os

Studies of Teratological Phenomena.

165

H t» O O H i— 1

pH i-H CM tH
i—i i—l

iO
T

CO ^ l— 1

1—1

~. i ~ —

i-H

»o

CO
01

— :~ i~ — — —
i—i

CM i—l
i— f

(M <M i— i
CN

iO CM CO i—i

H H t> CM

1— 1

Til

tO
CM

IOC5H

CM CO
CM

3Q t*» i— <

tJ< i— l

H N CO 01 i— i h CO 03

i^ X CO CO CM tQ i— i i— i i— t

CM
tO

i-H CM

2 2 1-1

CM
CO

t— O i— I

H CO H

CM CO
CM

1—1

tO
CM

— i— I

CM O CO
CM

tO
CM

DC l~ t*-

CM CM i-H

CM
CO

w 00 H

T— I

ci co c

X t"»
i— i

Ol
CO

i— i I> CO CO i—

i—l r-i CO

T i—l

tO
CM

cm CO cc

i-H CO X CO

i-H

CO CO t> CM
1— 1

i—i CM O CM
CM

uO 05 O- i—i i—( CM

Ot>lOH i-H i-H
i— 1

CO iH rj* lO CM
i—l

CM iO CO ^ h
i—i

lO
CM

CO
CM

1 0
CM

tO
CM

iC
CM

CQOh
i— i

»o

iO
CM

iO CM tO i— i CM

H t* N H

■"t" CM OS

CO H H

. CO

{Bnuouqy

[■BOLiouqy

X
CO

^O — H X
1—1

L~ X) T— 1

— t> O H

CM CM r-i

CO
CO

CO CM
CM

rH O

i-H — H

tO
CM

CM
Ol

»0^t-XOOi-iCMCO"^iOOt>XOO

NCO'<#lOCOt>Q0050HCMCO,*lOCOI>QO

Leal
Coun

Heig
(dcm

S9|n00|-XlBA()

— £ ,o

^ O i-i
CM

Table 17. Range of variability in number of floral leaves,

segregate plant (AA) and its F3 gene-

Parent

Plant No.

—1—12

Sepals

9
10
11

3
1

12
13
14
15
16
17
18

1
1

.">

10
1 ^

2
1

3
3

2
3

5
1

2
6

7
2

1
1

2
1

3
4

Petals

12
13
14

1
1

2
1

2
2
2

1
1

15
16
17

18
19

(21) j
l

&>
c
2

"cS

:'.

Stamens

12
13
14

4
3
1

1
1
1

7
1

5
1

4
2
1

3
2

2
1
1

16
17

18
19

(^} 1

(g)1
(30)

" 1

Ovary-

9
10

locules

16
17

®i

Leaf Count

105

48 |

Height (dcm.)

20 6j 19-8j 21*4j 22-9| 21^! 19 o| 2Vi\ 22'9| 19 o| 19 oj 24'4; 25'2| 19-8; 22"9| 244

in height and number of foliage leaves in an F2 generation
ration progeny (304 X 402)— 1 — 1 2.

•

9A
6i

or,

9ft

y i

V i

{22)

4)
*j 1

9
O

1 n

1 1

Q
0

108

22-9 1 214

20-6

24-4 24'4| 22-9

190

19-0

19-8

24-4

21-4

24-4

19-0

190

21-4

22-9

21-4

22-9| 21-4| 21-4

168

White.

Table 18. Range of variability in height, number of floral

per plant in an F3 homozygous

Plant No.

Sepals

5
6

24
1

23
2

Petals

class n

5
6

24
1

23
2

Stamens

■ —

S3
-

5
6

24
1

23
2

Ovary-
locnles

All

2-loculed

Leaf Count

Height (dcm.)

17-5

16*8

17-5

19-fl

198

183

i9-a

19-8

i -

206

19-8

190

19-8

Table 19. Similar to Table 18, but representing the

parent (aa)

Plant No.

Sepals

m
bit

5
6
7

24
1

23
2

23
1

Petals

05
X

5
6

24
1

2.r>

•':

24
1

23
2

Stamens

—

5
6

24
1

■

24
1

23
2

Ovary-
locules

All 2-loculed unless

Leaf Count

Height (dcm.)

175

168

19-0

16-8

18-3

168 190

18-3

1 v.

190

1S-3

18-3 20 6

19-S

19 0

198

Table 20a. Ratio of abnormal to normal segregates in F2
generations from two distinct normal varieties crossed with

the fasciated race.

Designation

Gen.

Classes

Total

Theoretical

AA Aa

Expectancy

(301—1—5 X 373)— 17 + — 15

194

269

201-75: 67-25

(353—3 X 301— 1)— 12

P.,

44 110

227

56* 75 : 112*5 : 56*75

(301—4—1 X 353— 3)— 7

11 47

18-75: 37*5:18*75

212

302

226*5 : 75*5

Studies of Teratological Phenomena.

169

leaves per flower per plant and number of foliage leaves
population (aa) [(304 X 402)-— 1 — 7].

24 25

24
1

23
2

24
1

23
2

22
3

23
2

24
1

All 2-loculed

23 24

26 25

23 22

23 23 25

23 23

23 24 21

19-8 | 19-8| 19-8| 19-8j 19 8| 19-s! 198 19-8 1 20"6 19*s| 19'0; 19 S> 19"8! 19"8! 19 8j 19*8| 19-8; 19 0| 19 8| — | 19"8 | 20-6

character of the progeny of a different F2 homozygote
[(304 X 402) — 1 — 10).

24
1

23
1
1

24
1

otherwise noted

24
1

23 |

19-8 | 19-8| 20-6| 20-6| 19-8 1 19*8| 19 8| 19 8] 19"8 19 8 19-8 19"8 19'8 j 198 | 19*8 j 19-0 j 19*8 j 19 8 j 198 | 19'8 | 20'6

Table 20b. Ratio of f asciate-stemmed plants to normal-
stemmed plants in the population noted in Table 20a.

Designation

Gen.

Classes

Total

Fasciate- stein

Normal-stem

(301—1—5 X 373)— 17 -J 5

224

269

(353—3 X 301— 1)— 12

201

227

(301—1—1 X 353— 3)— 7

170

White.

Table 21. Range of variability in height, number of floral

per plant in an F2 generation

Parents

Plant No.

1-5

1—17

6
7

13
8

12
1

13
3

2
19

4
17

11
1

GO
QQ

8
9
10

a
\j

■+J
0)

p-i

GO
d>

be
=
-

7
8
9
10
11
12
13

10
13
1

6
6
2
1
2

7
13
3
1

22
3

3
1

Stamens

Floral leaf <

6
7
8
9
10
11
12

13
7
3
1

13
3
3
1

6
6
2
4
4

11
6
6
1

6
13
3
1
1
1

1
20
4

9
2

12
1

1
1

8
2

(15)1

Ovary-locules

3
4
5
6
7
8
9
10
11
12

7
12
4

1
1
1

10
2

4
2

17
4

2
2
1

7
1
1

Leaf Count |

229

Heig

ht (dcm.)|

206

24-4

19-8

19-0 22 9

19-8

16-8

206

19-8

19-0

19-8

21-4

19-0

19-5

21-4

19-8

19-s

19-8

Studies of Teratological Phenomena.

171

leaves per flower per plant and number of foliage leaves
[(301 — 1 — 5 X 373) — 17 and —5].

25-5

4T5

5
12

5
3

5
2

7
11
7

15
10

18
4

24
1

15
8
1
1

24
1

18
6
1

6
2

18
6
1

i 1

13.
7
2
1

1
10
14

4
13
7
1

22
3

13
10
2

3
13
9

22
3

3
14
8

2
11
11

3
8
14

12
12
1

7
17
1

9
16

7
12
6

21
4

8
13
3
1

24
1

17
7
1

9
24
2

13
11
1

9)1

2
4
7
10
1

1
12
10

6
16

19
6

3
8
14

13
10
2

16
9

24
1

9
16

4
7
11
1
2

12
12
1

6
19

9
16

6
13
5
1

22
3

5
16
4

16
8
1

11
13
1

18
7

4)1 (

2
2
9
8

17)1

13
10
1
1

7
11
4
1
2

23
2

2
9
13
1

9
2

4
13
8

24
1

9
12
4

11
14

2
13
7
2
1

21
4

5
20

15
10

23
2

19
6

23
2

17
2

2
2

16
7

1
1

1
21

1
1
1

3
22

10
14

229

244

229

221

19 8

17-5

20-6

21-4

183

20-6!

22 1

206 1 17-5

10-7

20-G

19-0

20-6

19-0

18-3

25-9

206

23-6

22-9

236

172

White.

Table 21 (continued). Range of variability in height, number of

per plant in an F2 generation

Parents

Plant No.

4:;

45'7

55 56

1-5

1-17

22
3

*oS
Ph

5
6
7
8
9
10

22
2
1

9
14
2

14
1

24
1

6
18
1

4
13

9
12

11
11

22
2
1

24
1

9
13

6
16

9
1

1
4
9
V

24
1

b£
P

5
6
7
8
9

10
11
12

9
2

10
10

5
16
4

24
1

12
10

2
2
12
8
1

5
13
7

4
16
4
1

12
12

3
17
2

22
2
1

1
11

5
13
7

9
15
1

1
9
8
1

23
2

OB
S.

C3
~

1
1

5
G
7

10
11
12
13

16
7

o
—

10
i -±
1

12
12

i
i

2
3

7
1

7
16

9
a

a
o

1 1

j. j.

3
1

2
8

4
19

23
1

i
i

2
19
4

5
12
8

14
11

1 9

2
1)1

22
3

^—
0

2
3
4
5
6
7
8
9
10
11
12

19
6

3
18
4

6
2

19
6

9
11
2
1
2

13
10

24
1

23
2

21
4

20
5

6
13
4

1
I

Leaf Count

229

Height (dcm.)

21*4

198

22 9

24*4

20-6

20.6

1S 3 22-1

198

22-1

221

206

24-4

21.4

25-9

19 8

Studies of Teratological Phenomena.

173

floral leaves per flower per plant and number of foliage leaves
[(301 — 1 — 5 X 373) — 17 and —5].

h 1

16 ! 31

121

70 !

111

183

15-3

21-4 23 6] 20-6

18'3

18-3

198

18-3

18'3| 19-8

198

21-4

20-6

19-S

21-4!

18'3

26-7 1 22 -9

160

21-4! 23 6

20-6

22-9

174

White.

Table 22. Range of variability in height, number of floral

per plant in an F2 generation

Plant
No.

25-5

5
6

8
9
10

24
1

8
13
4

6
10
9

24
1

3
17

3
22

12
10

3
22

12
2

3
12
10

1
15
9

17
6
1

9
15
1

1
2
20
2

3
9
12
1

9
12
4

19
6

23
2

2
12
11

3
10
12

11
9
5

2
21
1
1

Petals |

lass ranges

5
6
7
8
9
10
11
12
13
14

23
2

7
14
4

6
10

24
1

2
20
2
1

23
2

8
13
4

2
16
5
1

11
12
2

4
9
8
4

1
12

8
3
1

11
12
1

9
12
4

2
19
3
1

4
10
11

5
14
2
1

17
8

24
1

9
15

2
8
15

8
3

2
19
1
1
1
1

| Stamens |

Floral leaf c

5
6
7
8
9
in

12
13
14

24
1

6
13
6

9
15
1

24
1

9
12
4

1
4

6
10
3
1

12
6
5
1

1
16

2
1

12
11
1

14
9
2

6
16

4
12
8
1

14
8
3

19
6

2
9
13
1

1
9
14
1

2
18

2
1
2

Ovary - locnles

2
3
4
5
6
7
8

14
11

18
7

2
17
5
1

13
10
1
1

23
2

rr
1

14
3
1

24
1

1
12
12

8
13
3
1

23
1

24
1

10
14
1

10
15

18
5
1
1

12
13

11
14

19
6

19
1
1
2

Leaf
Count

Height 1
(dcm.) 1

16-8

160

10-7

175

17-5

160

18-3

21-4

22-1

19-0

16-0

20-6

16-0

21-4

18-3

19'8

15-3

22*9

15-3

18-3|

19-8

23-6

Studies of Teratological Phenomena.

175

leaves per flower per plant and number of foliage leaves
from (353 X 301) and reciprocal.

35-5

3
6

15
1

21
2
2

5
17
3

5
8
1

12
11
1
1

3
20
2

7
15
2
1

1
11

12
2

6
17
2

1
12
12

24
1

2
9
10
3

2
13
10

16
4

8
12
4

24
1

16
7
2

8
13
3

9
13
2
1

4
13
3

6
7
11
1

20
3
2

18
6
1

6
15
4

11
10
4

18
6
1

4
21

9
10
6

19
6

6
1

20
4

10
14
1

24
1

1
2
11
6

13
6
6

19
5
1

3
17
3
2

24
1

16
8
1

1
2
15
7

2
13
8
1
1

2
13
4
1

7
11
7

23
2

18
6
1

7
13
4
1

18
6
1

23
2

3
4
17
1

7
13

22
3

20
5

22
3

13
11
1

2
11
11

9
12
4

20
3
1

1
3
8
5
6
1

20
4
1

1
2
16

3
8
13

6
11

22
3

23
2

6
18
1

24
1

15
9

' 1

23
2

18
6
1

10
13
1
1

2
16
7

10
14
1

4
9
2
4
5
1

6
18
1

4
17

6
11

17-5 J 21-4| 19-0

153

160

137

153

20-6

10-7

214

20-6

21-4

18-3

20-6

21-4 1 24'4

16-8

20-6

22-9 1 16-o| 22-1

19-0

16-0

19-0

176

White.

Table 22 (continued). Range of variability in height, number of

per plant in an F2 generation

Plant
No.

e 0

t>8

fin

/ 1

7Q
/O

8
9
10

6
1

9
1

—

Q
O

1
1

1
±

22
3

8
13

21
2

17
4

17
2

2
22

10
7

[ass ranges

10
11
12
13
14

cS
S3

1
|

10
11
12
13

3
1
1
1

1
1

(16)1

1 Ovai

6
7

2
1

Leaf
Count

Height
(dcm.)

198

21-4

20-6

18-3

14-5

16-0

13-7

19-0

16-8

22-1

16-8

13-7

18-2

221

22-9

160

18-3

20-6

16-0

22-9

168.

Studies of Teratological Phenomena.

177

floral leaves per flower per plant and number of foliage leaves
from (353 X 301) and reciprocal.

101

7
f

-LO

1 1

a
\j

K
o

7)1

5
2

13-0

| 13-0

1 22-9

17-5

18-3

137

| 14-5

153

206

16-0

168

13-0

1 22'9

19-8

17-5

153

114

153

13-7

18-3

160

24-4

17-5

Induktive Abstammungs- und Vererbungslehre. XVI. 12

178

White.

O co

H o

CO t-
00 o o

GO CO CO
H H N

o cs

i—l CM rH

CO 05 GO

rH O lO
r— I CO CO

CM CM CM

GO r— I OS
rH CM CM

CO CO GO

CM CO O

rH rH CM

CM CO CO

t- OS CM
CM rH CM

OS CO CO CO
© t> l> H
^ ^

rH t~ « rH

^ ^ O

CM C5i 7^ CM
OS

O ^ ^ .H

co co ^

ri5 rQ

lO lO
CM CM

PQ PQ

iO uo
CO CO

H O IC
CM CM CO

C5 O CO

CO OS T4 CM

OS CO 10 rH
^ OS

OJ CO Oi CO

t> co t>

rH OS CO
CM rH rH

OS GO rH
CO lO

» OS
CO CO CO

GO ^N

rH S 00

OS ^ H
CM CM rH

rH rH 71 CM

IO CO ^

GO CM CO O
H CO OS

IO rH

CO ^ 00
rH CM CM

CO rt<

o co r1 CM

^ CO rH

^ ^ OS

PQ PP

iO CO
CO CO

CM CO CM CO

8 «

o '43

X
pq

X
a

Studies of Teratological Phenomena.

179

Table 24. Fasciate- to normal-stemmed plants in an F2
population from (396 X 301) and reciprocal.

Designation

Fasciated

Normal

Total

(301—1—1 X 396)— 12

109

-*a

(301—1—2—5 X 396— 1)— 11

207

233

(396 X 301— 1)— 10

127

135

Total 1 st count

431

477

Total 2nd count

427

477

Total 3rd count

427

477

Total Count by alterations in normal ^
phyllotaxy J

422

477

12*

180

White.

1 — 1

£^

T— 1

<h-

f— 1

r-H

TT
CO

, ,

I
1

i — l

1 — 1

1
1

1 1

r-J

r— I

1
1

-4-3

CP
—

d
*—i

"r" '

-r^>

,s J

•i— 1

—

• rH

-4-3

t>.

—

5-1

C£

—

CM rH h
i— ( >—i

-T CO

DO OS

Ol
L-

IO OS —

CM CO CM Tf rH
H

CM Ci 05 rH

r- 1
t-

HQOOCOWH

1—1

HHO CM rH CM rH i— I rH
r—

rH CM H< rH

o
t~

1—1

CM O O H CM rH

H" CO rf SO rH CM

co
co

CO ^ CO lO
rH

CO O CO rH rH
rH

CM CM ^ CM CO CM

CM CO t>
1—1

CM lO rH
r— I

CM t> iO

—

CO O CM

rH rH

r-< 00 fc- IQ

CM IO X CO — Ol

Tf4

CO CO H

1—1

CM t> CO iO CO CM

rH CM IO CO CO CO

OS
CO

io a

—

CO O O rH rH

CO t> rH — 1 rH
rH

X
CO

UO t> CM
1—1

O Tf< CO CO rH

rH IO IQ rH CM

iO
CO

H lO 1> t-

1—1

CM X rH CM rH rH
rH

rH IT— X T — CO

i—l C— CO rH

1—1

rH O X rH
rH

rH X H" CO -t

CM lO O 00
rH

t- t- CO rH

CO X lO CM CM

CO
?!

X O CO —

— ~ -T
rH rH

00C5 1> rH

CO
CM

CM O lO CO

T— 1

CM IO -tf CO rH

rH H^ X IO CO 01

— CO

CM CO

x co co

r- 1

H N CO
rH

CM CO CM CO CM

CO CM CM CM

CM O C5 CO
1— 1

C5 CM CM rH rH

HI> Orf HOI

lO
CO

T— 1 t— CO T— i

O O CO CO H

O lO i0

i—l

CO CD N ^

x lo ^ co io

rt no co ia io co

i— I

X lO CM

CO CO X rH
rH

ABB

rH (M — 1 rH

(3)

CO rH rH rH rH rH

-f

r— 1 IO CO 00 00

lO CO rH IO rH rH

CO CM iO "t

rH CM CM

H(MQ0t|
rH

rH CO 00 IO IO

CO O lO i>
rH

OS -t CO t*4 CM

rH X CO UO

KO
CM

lO t> CM rH

X CO rH

Hj4 OS X CO —

rH H^ CO CO H
rH

CM rH

IO OS CO CO rH rH

OS TJ4 IO rH rH

Class

No.

lOCl-XOO

—

CM CO

rH rH

COi'iOCOt>CCC50HCMCOi'iOCOt>

^lOONXOO

ant

SaSuBI SSB[0

Studies of Teratological Phenomena.

181

rH CO

HH rH CS rH
rH

»o

l'-

Ol GO Ol rH rH
i—l

HH rH

0
10

GO
OS

T— 1

01 rH Ol HH

lO (M H ^ H H i-H
i—l

CO rH rH

GO
CO

i-l r-i

00 CO 03
1—1 1—1

GO rH ©1 HH
rH

CO
rH

HH ,— 1 — 1

HH HH HH HH HH Ol 01

H- HH rH IO rH
rH

OS
01

CO
GO

i— 1

CO OS 01 rH

T—I

O IO rH
tH

O
CO

lO HH tH 01 CO

tH CO rH CO rH
rH

Ol
Ol

GO
1— 1

HH 0* rH rH

rH CTS T—I HH CO 01 tH rH

^ H CO IC (M
rH

1-H

00 CO 01 t—i

T-H

rH HH HH CO CO
1— 1

GO
OS

T— 1

0*S Ol 01 CO Ol T-H T-H T-H

1— 1

CO t>" rH Ol T—I t-H
1— 1

H 1> CO H
1— 1

■HH

IO CO Ol IO
1— 1

rH CO tH

CO
tH

1— 1

1— 1 O HH T-H T-H T-H Ol

Ol 0 O CO

rH rH

cq
id
1— 1

IO t- Ol t-H
-H

CO CO HH 01
1— 1

Ol t> Ol t-H 1— 1 rH rH

I> IO rH Ol
rH

GO
1— 1

rH O t>
rH

01 l_H

CO
CO

t-H

1—1 T— 1

Ol CO CO Ol rH rH
rH

rH Ol GO CO rH
rH

t-
tH

Ol 01

Ol CO CO Ol rH rH
rH

CO rH GO CO
rH

CO
GO

Tj<OH

HHJ OSO rH
rH

HH
rH

H GO Q rH CO rH Ol

tH Ol IO IO CO HH 01

■H/

1— 1

T— 1

VO CO CO rH

i-H Ol O O Ol
1— 1 rH

GO
Ol

rH
rH

rH CO

rH CO rH 1— 1

1— 1 rH

rH CO Ol Ol

CO
t~

HH
rH

HCOHH

IO rH rH HH Ol
rH

H CM t> CO CO

rH CO HH <M -HH rH

rH CO CO

rHOlCiOlOlOllOrH rH

GO t> CO HH

HH
CO

GO
CO

Ol C- 01 Ol rH rH
rH

H CO OS I> (M

CO
Ol

rH Ol

(20)2

CO t- 00 rH Ol rH

rH rH rH

(I)

rH lO C5 CO Ol rH rH

tH
HH

OlCO^iOCOC-GOOOrHOlCO^fiO

0100000000900

boHwm^ibit-ooob

i-H

■+J ^— %

Heig
(dcm

sajnooj-toAQ

a>
hH

182

White.

00 iO Ol

rH CO — 1

Ol CO
Ol

O t> H H

co t> oi

00 CO

co o oq

^ O CO Ol rH

-f i-H rH

CO 01 CO CO

^ Ol Ol LO Ol CO CO rH

LO lO "T rH Ol

CO OS r— I i—l i— I

iflOl» Ol

H CD ^

CO H X CO

O H

rH Ol 00

^ 00 GO CO

CO CO 01

Tt< 00 Ol i-i

Ol 00 ^

r- I CO O LO rH i— I i— I

CO 00 CO LO

Ol CO
Ol

co oi

H ^ (M t> H

t— 1 i— 1 CO LO r— I

01 t- 00 CO CO

Ol Ol rH

i-i Ol Ol
Ol

'—I T*<

Ol CO o

CO Ol
Ol

LO CO 00 O © r- I Ol CO

C0^L0C0t>00CiOi-l01C0'*L0C0t>.

LO CO 00 OS o

Studies of Teratological Phenomena. 183

lO
<M

0
©1

^ rH

1— 1 T— 1

©q

iO
©q

lO
©J

-t

lO
Ol

CO i— 1 i— 1 Qj

©q

CS CO
i— l

©q

oq
1—1

l> 00
1—1

iO
©q

»b

1—1

t- 00
1—1

©q

OS
rH

1— 1 1— 1

Oi^ HH

1—1

©q cC(Mh

lO O5HCQCOC0HH

»"H

rH Tj< rH

iO Tfl ©q CO i-H
i—i

IP
1— 1

o i-H oq oq
oq

1—1

uO
t-

oq HH

CS
1—1

l> iO CO
i—i

i-i co o OS oq
i—i

CO
O

O CO i— 1 i—l
©J

tH rH

CS
1— (

i— I CO i— 1
1— ( 1— 1

CO iO CO GO

t-
1—1

i— 1 t— rH *H/ i— 1 i— 1

1—1

i—l CO i—l

E~
i-H

©q CO

him oq
oq

i-H

IO
(M

oo os oq oq co r-i

t>
rH

O
OS
rH

©q

CO CO HH '
i—l

lO t» CO
1—1

©q

O
CO

-f iH
©q

i© M

oq ^

CO
CO

oq
oq

^ CS t> rf r-(

1—1

GO
i— I

05 1OH
i—l

co
oq

CO ©q

©q

io t~ "HH oq

oq
oq

©q i-h ©q

©q

CO i— I i— I

1— 1 1—1

1—1

CO
rH

CO CS

1—1

co os oq i— i
i— i

O
i— (

rH ©qCO^lOCOt-GCOS

oqcorj<iocoi>G0050i-(oqcoTj<io

909999090900

1—1

Leaf Count

Height (dcm.)

SU9UTB;§

iCm9l{;UBD^BQ

184

White.

r>5

I— 1

*tH

ce
o

*©

r— 1

^— J

«HH

T-H

(TO

resei

be
•i— i

lit

c5
be

<X>

cter

], as

i— i

«HH

i— i

aabb

. ._ .

IO
<M

»o

(M
CO

lO
(M

IO
CM

lO
CM

lO
lO

CO
CM

IQ
(M

lO
CM

IO
CM

CO CM
CM

-+ rH

IO
CM

rH
tH

IO
CM

IQ
CM

IO
CM

iO
CM

CO CM
CM

IO
CM

UO
CM

CO CM
CM

lO
CM

IO
CM

CO
CM

oo i— 1 1— i

CO CM
CM

IO
CM

IO
<M

PQ
PQ

CO
CO

O iC
CM

rH O CO rH
rH rH

CO 00 tJ*

l>
i— 1

CM CO
CM

■HH ^

rH CO rH

lO
i—i

Tt< rH
CM

lO
CM

rH CO rH
rH rH

O lO
CM

rH CM CM
CM

O rH
rH rH

Class

Plant No.

iO®l>00CiOH(MCC

rH rH rH rH

W^iOCOt*Q0050H(MCO^iOCOI>

H/ lO SO 00 05 O

spscfog

suanre;g

Studies of Teratological Phenomena.

185

i—i

CM
i—i

CM »

00
C5
i—l

t-H

00
1— 1

lO
CM

1—1

i-H

CO oq
CM

fr-
ee

lO
CM

CO'

i-H

CO
rH

CO
lO

IO
CM

CM
i— i

IO
CM

1—1

CM .

T-H

Tj< 1-H

p
cc

O
CM

tti

CM t-H

CM
i—l

m h h

t-
i-l

9°

1— 1

lO
CM

CM i—l CM

CO
1— 1

lO
CM

C5 CO CO

lO
1—1

CO
CM

lO
CM

io hi

CM W

1—1

00
OS
i—l

i—l

i— I i— 1 GO lO

i-H

O CM CM

i-H

At*
rH

— i—l

H CO H

i-H

00
1— t

co cm

l>(MiOH

i-H

CMCO^iOCOt-OOCSOi-iCMCO'^iO

1— I 1— 1 1— 1 T— 1 T— 1 1— 1

oopopooppopp

~£

Height (dcm.)

P
O

suaniB^g

SapOO[-^JBAQ

Referate.

187

wohl die alteste Arbeit, welche uns von einer zweifellosen Klonumbildung bei
Bakterien berichtet, die auch von den extremsten Anhangern der Lehre von
den Bakterienmutationen bis vor kurzem als „ Mutation" aufgefafit werden
mufite. Die Sachlage ist die folgende.

Gruber hatte in einer 307 Tage alten Kultur des Vibrio Finkler-Prior
auf Gelatineplatten neben zahlreichen typischen Kolonien anscheinend als
Verunreinigung in geringer Zahl Kolonien einer weniger rasch verflussigen-
den Bakterienart gefunden. Dieseiben erregten seine Aufmerksamkeit da-
durch, dafi sie bei oberflachlicher Betrachtung eine gewisse Ahnlichkeit mit
den Kolonien des Choleravibrio zeigten. Auf Grund dieser Beobachtung
stellte dann Firtsch seine Untersuchungen an.

Die von Gruber gefundene, anfangs als Verunreinigung angesprochene
Form des Finkler-Priorschen Vibrio unterschied sich vom Typus vor allem
durch das Verhalten auf Nahrgelatine; der Verflussigungstrichter im Gelatine-
stich bildete sich viel langsamer und nahm andere Form an als beim Typus ;
zudem war die Beweglichkeit des neuen Vibrio geringer, das Bild der Ober-
flachenkolonie abweichend usw. Bei langerem Suchen wurde dann auf
weiteren alteren Kulturen wiederholt dieselbe abweichende Form gefunden,
aufierdem aber kamen noch zwei andere abweichende Formen zur Beob-
achtung. Dabei stellte es sich heraus, dafi der am wenigsten abweichende
Typ in 33 — 54 Tage alten Gelatin estichkultur en (Vibrio I), der zuerst ge-
fundene, mittlere Typ (Vibrio II) in 48 Tagen bis 1/2 Jahr alten Kulturen
auftrat, wahrend der letzte Typ (Vibrio III) in iiber einem Jahr alten Gela-
tinekulturen vorherrschte. Wahrend nun Vibrio I seine Eigenschaften nicht
dauernd behielt, sondern relativ leicht, wenigstens teilweise, in den ursprung-
lichen Typus ruckfiihrbar war, gelang diese Riickwandlung bei Vibrio II und
III bei regelmafiigem Verimpfen innerhalb einer iiber vier Monate sich er-
streckenden Arbeitszeit nicht. Bemerkenswert ist weiter, dafi aufier den
drei besonders hervortretenden Typen noch relativ konstante Zwischentypen
gefunden wurden.

Als wichtigstes Ergebnis fuhren wir die folgenden Worte Firtschs an:
Aus ein und derselben Bakterienart wurden vier Formen gezogen, die in
ihrer Kolonienform (teilweise auch in der mikroskopischen Wuchsform) durch-
greifende Verschiedenheiten zeigen und von denen wenigstens drei (der ty-
pische Proteus Vibrio II und III) diese unterscheidenden Merkmale mit solcher
Zahigkeit bewahren, dafi sie einzeln fur sich untersucht — nach dem bis-
her geiibten Modus der Artbestimmung — zweifellos als besondere Arten
aufgefafit werden miifiten.

M. a. W.: Firtsch hat in seinem Vibrio II und III eine „Bakterien-
mutation" im Sinne zahlreicher neuerer Bakteriologen im Jahre 1886 fest-
gestellt; er hat aber gleichzeitig in seinem Vibrio I eine Dauermodifikation
im Sinne Jollos usw. (vgl. Lehmann, 1916, S. 297) beschrieben.

Durchaus bemerkenswert und lehrreich fur die Bakteriologen, welche
jeden in einer alteren Kultur beobachteten etwas abweichenden Stamm als
Mutation bezeichnen, ist aber der fo]gende Satz aus Firtschs Arbeit: So
bedeutend die Unterschiede im Aussehen der Kolonien (der verschiedenen
Vibrionenformen) auf Nahrgelatine sind, so lassen sie sich doch mit grofier
Wahrscheinlichkeit auf verschiedene Grade von Abschwachung der Wachs-
tumsenergie uberhaupt, der Fahigkeit, die Gelatine zu verfliissigen und der
Eigenbewegung zuriickf uhren ; Abschwachungsvorgange, die gewifi nicht be-
deutsamer sind, als der Verlust der Fahigkeit, Sporen zu bilden, der Gar-
tatigkeit, der Virulenz.

188

Referate.

Wie recht Firtsch hatte, geht aus den Untersuchungen Fiirsts hervor.

Als wichtigste Vervollkommnung der Firtschschen Versuche wird von
Fiirst das Burrische Tuschepunktverfahren zur Isolierung einzelner Bak-
terien als Ausgangsmaterial der Staimne eingefiihrt. Wenn allerdings Fiirst
hierdurch glaubt, mit reinen Linien zu arbeiten, so verfallt er in denselben
Fehler, welcher von so vielen neuzeitlichen Bakteriologen, die sich mit „Bak-
terienmutationen" beschaftigten, gemacht wurde. Es handelt sich bei solchen
auf ein Individuum zuriickgehenden Bakterienstammen nicht urn reine Linien,
sondern um „Klone", wie ja den Lesern dieser Zeitschrift gelaufig ist.

Bei vier solchen Klonen des Vibrio Finkler-Prior liefien sich nun die von
Firtsch gemachten Angaben hinsichtlich der Zeit des Auftretens der Vi-
brionenformen und ihrer kulturellen Eigenschaften in vollig ubereinstimmen-
der Weise wiedererlangen. Zudem werden die Untersuchungen Firtschs
noch in bedeutsamer Weise erweitert.

Einmal untersuchte Fiirst, ob durch Auslese von Plus- oder Minus-
varianten der in der Grofie sehr variablen Individuen des Typus eine Ver-
schiebung des Gipfels der Variationskurve in der Nachkommenschaft moglich
sei. Es gelang dies nicht. Ebensowenig hatte eine solche Auslese einen
Einflufi auf das hamolytische und peptolysierende oder agglutinative Ver-
mogen der aus ihnen hervorgegangenen Stamme.

Die verse hiedenen Typen aber lassen sich, wie Fiirst zeigte, noch
durch eine Reihe anderer Merkmale, als Firtsch angegeben hatte, differen-
zieren. So konnte festgestellt werden, dafl bei den extremsten Typen mit
dem Verlust der Beweglichkeit der Verlust des Geiflelapparates Hand in
Hand ging. Parallel damit ging weiterhin der Verlust der spezifischen — also
der Gelatineagglutinabilitat. Hingegen blieb das Saurebildungsvermogen bei
alien Typen das gleiche.

Fiir die Frage der „Bakterienmutationen" weitaus am wichtigsten ist
indessen die Feststellung Fiirsts, daft wohl bei regelmafiiger Weiterimpfung
von Gelatine- zu Gelatinekultur die extremen Stamme sich konstant er-
halten, dafi aber bei langerem Stehen der Kultur auch Vibrio II und IH
Firtschs stets von 2 — 21/2 Monaten an nach und nach wieder in den Normal-
typus zuriickschlagen.

Damit aber ist, wie Fiirst zu Ende seiner Arbeit mit Recht sagte,
bewiesen, dafi es sich wenigstens bei den von Firtsch beobachteten, mehr
oder weniger lang vererbbaren Variationen nicht um echte Mutationsvorgange,
die zur Entstehung neuer Arten Anlafi geben, handeln kann.

In einem Nachwort hebt dies Gruber noch weiter hervor, indem er
die beobachteten Abweichungen von der Norm nicht als genotypischer,
sondern phanotypischer Natur aufgefafit haben will: „Das lang dauernde Be-
stehenbleiben der neu aufgetretenen Eigentiimlichkeiten bei den Nachkommen
auch unter veranderter Lebenslage (z. B. bei Umziichtung auf Agar) ist nicht
echte Vererbung, sondern „falsche", auf „Nachwirkung" ( Woltereck) be-
ruhende".

So sind also die urspriinglich nur als „Mutationen" deutbaren Firtsch-
schen neuen Vibrionenformen durch Fiirst als Dauermodifikationen erwiesen.

Es ist das eine wertvolle Lehre fiir die Zukunft. Findet man ab-
weichende, iiber kiirzere oder langere Zeit konstante Bakterienstamme, so
bezeichne man sie nicht sogleich als Mutationen, sondern als Klonumbildungen,
bis ihre Natur als Dauermodifikation durch Riickbildung erwiesen ist, oder
aber es spater einmal moglich wird, die eine oder andere als auf geno-
typischer Grundlage zustandekommend zu erweisen. e. Lehmann.

Referate.

189

Kiefiling, L. Erbanalytische Untersuehungen iiber die Spelzenfarbe des
Weizens. Ein Beitrag zur angewandten Vererbungslehre. Landwirt-
schaftliches Jahrbuch fiir Bayern 1914. Nr. 2.

Vor 10 Jahren wurde vom Verfasser mit der Beobachtung der Folgen
spontan bei Weizen (Triticum vulgare) eingetretener Bastardierungen begonnen.
Eine erste Mitteilung iiber dieselben erfolgte 1908 in „Fuhlings landwirt-
schaftlicher Zeitung", S. 737. Einige Nachkommenschaften zeigten beziiglich
des in erster Linie beobachteten Merkmals, Spelzenfarbe ein Verhalten, das
der Annahme des Pmm-Schemas entspricht: braune Spelzenfarbe : Br X weifie
Spelzenfarbe : br in Fx : Brbr mit Dominanz von Braun, in F2:BrBr; Brbr;
brBr; brbr oder aber ein Verhalten, das mit der Annahme des Zea-Schemas
in Einklang zu bringen ist. Bei anderen ergaben sich aber andere Spaltungs-
verhaltnisse, insbesondere erschienen weniger weifispelzige Pflanzen aus
braunspelzigen und es gaben auch extrahierte weifispelzige Pflanzen, die rein
vererben sollten, einzelne braunspelzige Pflanzen. Die Erklarung, die Nils son-
Ehle fur Falle seiner Hafer- und Weizenbastardierungen anwandte und
die mehrere gleichsinnig wirkende Anlagen fiir eine sichtbare Eigenschaft
annimmt, wurde dann zur Erklarung der selbst beobachteten abweichenden
Falle bemitzt. Sind zwei Anlagen fiir braun vorhanden, von welchen die
eine: Brx gewohnliches Braun, die andere: Br2 helleres Braun und beide zu-
sammen: B^B^ dunkleres Braun bewirken, so wiirde sich in F2 ergeben:

^BT1'BT1Br2BT2 dunkelbraun, rein vererbend,

B^B^brgbro braun, rein vererbend,

br1br1Br2Br2 hellbraun, rein vererbend,

brx br1 br2 br2 weifi, rein vererbend,

B^BrjB^br^ dunkelbraun, nach 1:2:1 spaltend,

B^b^BrgBrg dunkelbraun, nach 1:2:1 spaltend,

Brx brxBr2 br2 braun, nach 1:2:2:4:1:2:1:2:1 spaltend,

Brt b^ br2 br2 hellbraun, nach 1:2:1 spaltend,

b^b^B^b^ sehr hellbraun, fast weifi, nach 1:2:1 spaltend.

Die Spaltungsverhaltnisse aufgefundener Individuen von Winter- und
Sommerweizen wurden nun weiter verfolgt und ihr Ergebnis durch die er-
wahnte Erklarung zu deuten versucht. Die Individualauslesen , innerhalb
welcher jahrlich einige Individuen ausgelesen wurden, standen dabei ohne
kiinstlichen Schutz gegen Fremdbestaubung nebeneinander. Es ergaben sich
verschiedene Abweichungen : wiederholte Abspaltung von braunspelzigen
Individuen in der Nachkommenschaft von weifispelzigen, Spaltungszahlen, die
sich nicht in das Schema einfiigen, unregelmafiig erscheinen, oder die An-
nahme von mehr als zwei Anlagen fiir braune Spelzenfarbe nahelegen, Auf-
tauchen von Abstufungen der Braunfarbung und abweichende Spaltung nach
solchen Individuen, Auftreten von Fleckung der Spelzen und dergleichen.
Weiterhin ergab eine genauere Durchmusterung des Versuchsmateriales und
einer Reihe von Formenkreisen von Sommer- und Winterweizen, dafi es
rein weifie Spelzen nicht gibt, und eine Untersuchung der Farbe der Weizen-
spelzen legt die Annahme nahe, dafi die verschiedene Farbung der Spelzen
durch Abbau oder Umwandlung des Chlorophylls bewirkt wird. Beide Fest-
stellungen fuhrten zu dem Schlufi, dafi es sich bei Braun und Weifi nicht
um die Wirkung einer Anlage und ihres Fehlens handeln konne, sondern
dafi auch weifispelzige Individuen irgend welche Anlage fiir Braunspelzigkeit
besitzen. Der Verfasser nimmt im weiteren Verlauf zur Erklarung der Ver-
haltnisse eine Anzahl von Anlagen an, die alle Braunfarbung bedingen, aber

190

Referate.

verschiedene Wirkungsintensitaten aufweisen. Br ist dabei die Gesamtanlage
fur Braunfarbung; Br1 Br2 Br3 usw. sind Einzelanlagen, die alle eine bedeu-
tendere Intensitat bewirken und z. B. die untereinander nur wenig ver-
schiedenen Werte 4,001, 4,002, 4,003 besitzen, wahrend Br0m Br0n BrQ° Einzel-
anlagen sind, die eine geringere Intensitat bewirken und z. B. die unter-
einander wenig verschiedenen Werte 0,999, 1,000, 1,001 aufweisen; br1? br2,
br0m, br0n deutet dann das Fehlen der betreffenden Anlage an. Die bei der
Berechnung der Gesamtintensitat sich ergebenden Intensitaten 4, 3,99, 3,97
und die nachsten Stufen wiirden braun entsprechen, die um 1 gelegenen
Stufen weifi und die zwischen 1,5 — 2,5 liegenden hellbraun, Eine mono-
hybride Spaltung wiirde demnach in der zweiten Generation geben:
1 mal Brx Brx Br0 Br0; 2mal B^ brt Br0 Br0; 1 mal b^ b^ Br0 B0 und die

4 + 4 + 1 + 1

Berechnung der Gesamtintensitat dann Intensitat: '- = 2,5;

4 + 0 + 1 = 2; ° + ° + 1 + 1 i Die dihybride Spaltung, mit Br0
bei beiden Eltern, wiirde sich dann wie folgt darstellen:

Strukturformeln: Intensitatsrechnung: Farbe:

a) 1 • BriBrjBrgBrgBroBro 4 + 4 + 4 + 4 + 1 + 1 = 18 : 6 = 6 = braun,

b) 2 • BrtBrj Br2 br9 Br0Br0 4 + 4 + 4 + 0 + 1 + 1 == 14: 5 = 2,8 = braun,

c) 2 • Brj b^BroBiCBroBro 4 + 0 + 4 + 4 + 1 + 1 = 14: 5 = 2,8 = braun,

d) 1 • By1 B^ br; br^ Br0 Br0 4 + 4 + 0 + 0 + 1 + 1 = 10 : 4 = 2,5 = hellbraun1),

e) 4 • B^ brx Br2 br2Br0Br0 4 + 0 + 4 0 + 1 + 1 = 10 : 4 = 2,5 = hellbraun1),

f) 1 • brx b^BrgBrgBroBro 0 + 0 + 4 + 4 + 1 + 1 = 10:4 = 2,5 = hellbraun1),

g) 2 • Brx brx br2 br2Br0Br0 4 + 0 + 0 + 0-t-l + l= 6:3 = 2,0 = hellbraun,

h) 2 • br1 brxBr2 br2Br0Br0 0 + 0 + 4 + 0 + 1 + 1= 6:3 = 2,0 = hellbraun,

i) 1 • hr1 bT± br2 br2Br0Br0 0 + 0 + 0 + 0 + 1 + 1= 2:2 = 1,0 = weifi.

Auch wenn Br0 nur von einem Elter kommt, wobei F1 dann z. B.
Brx br1 Br0 br0 ist, lafit sich die dihybride Spaltung und die Intensitatsrechnung
in analoger Weise darstellen.

Zur Erklarung der beobachteten Fleckung: auf fast weifien Spelzen
dunklere Zeichnung entlang den Randern und Nerven, wird ein Fleckungs-
faktor angenommen, der nur wirkt, wenn Farbfaktoren vorhanden sind. (FBr
gefleckt, Fbr nicht.) Heterozygoten fur Br sowohl wie fur F (Brbr und Ff)
wiirden kleinere oder seltenere Flecke aufweisen. Eine weitere Moglichkeit
bei der Erklarung der Fleckung ist gegeben, wenn eine zweite Anlage fur
Braunfarbung angenommen wird, die von F nicht beeinflufit wird. Es lafit
sich aber auch ohne Annahme einer besonderen Anlage fiir Fleckung aus-
kommen nur mit Annahme von bestimmten Braunanlagen , welche neben
Braunfarbung auch die Fleckung veranlassen.

Die vom Verfasser aufgestellte Intensitatshypothese arbeitet mit einer
grofieren Zahl von Anlagen, die sich voneinander nur durch verschieden ab-
gestufte Wirkung unterscheiden. Die in der Arbeit verwendeten bestimmten
Zahlen fiir diese Wirkung sind nur beispielsweise. Die Verteilung der An-
lagen bei der Geschlechtszellenbildung und die Zusammentritte der ver-
schieden veranlagten Geschlechtszellen erfolgt entsprechend Mendel, Vor-
handensein und Fehlen je einer Anlage wird dazu angenommen. Die Be-
griffe Dominanz, Rezessivitat, Pravalenz, Epistasie und Hypostasie sind

x) An der Grenze von braun.

Referate.

191

durch einen Zahlenwert, die Intensitatszahl, ersetzt, welche die gegenseitige
Beeinflussung der einzelnen Anlagen scharf ausdriickt.

In Beziehung auf die Zuchtung lafit die Arbeit auch wieder den Wert
der Fortsetzung der Auslese, auch bei Selbstbefruchtern , erkennen, da sie
nachweist, dafi viele Spaltungen sich der Wahrnehmung entziehen und dafi
erkennbare Folgen von, nicht als solche erkannter, Heterozygotie nach
weiterer Spaltung fruher oder spater auftauchen konnen. Fruwirth.

Stark, Peter. Untersuchuiigen uber die Variabilitat des Laubblattquirls
bei Paris quadrifolia. Zeitscbr. f. Bot. 1915, 7, S. 673—766.

Die statistische Untersuchung findet immer mehr Eingang in die bota-
nische Wissenschaft. Fiir die Abstammungs- und Vererbungslehre ist das,
wie langst erkannt, von besonderer Bedeutung. Denn die auf statistischer
Grundlage gewonnenen Ergebnisse bilden eine sichere Grundlage, auf welcher
Abstammungs- und Vererbungslehre weiterbauen konnen.

Der Verf. der vorliegenden Arbeit hat es nun unternommen, die
Variabilitat des Laubblattquirls bei Paris quadrifolia auf statistischer Grund-
lage nach den verschiedensten Richtungen zu untersuchen und ist dabei auf
sehr interessante Tatsachen und zu weitgehender Klarung der untersuchten
Variabilitatsverhaltnisse gekommen. Betrachten wir die Hauptergebnisse:

Die Blattzahl der Einbeere schwankt in dem untersuchten Gebiet
zwischen 1 und 7; die Variabilitatskurve zeigt einen sehr steilen Gipfel
uber 4. Die Gesamtkurve aller Standorte zeigt infolge der starkeren Aus-
bildung des linken Schenkels eine unverkennbare Asymmetric Dei Mittelwert
liegt dementsprechend etwas unter 4,0. Umfangreiche Messungen fuhrten
dann zu dem Ergebnis, dafi gleichzeitig mit der Gliederzahl auch Stengel-
lange und Blattlange ansteigen.

Das Gesamtmaterial lafit sich in bliihende und nichtbluhende Stengel
einteilen; bei den letzteren iiberwiegen die Minus-, bei den ersteren die
Plusvarianten. Bei den bliihenden Sprossen ist also der rechte, bei den nicht-
bluhenden der linke Gipfel besonders stark ausgepragt.

Junge Pflanzen haben eine geringe Gliederzahl, altere steigen unter
unregelmafiigen Oszillationen bis zum normalen Viererstadium, bei giinstigen
Verhaltnissen aber bis zu hoheren Quirlzahlen empor. Bei Betrachtung
grofierer Genossenschaften auf statistischer Basis findet man im Verlauf der
Generationen bestimmte Gesetzmafiigkeiten in diesen Oszillationen, welche
in einem steten Hinfluten zum Mittelwert gipfeln.

Zwischen Haupt- und Seitensprofi bestehen enge Beziehungen im
Variationsverhalten. Die Seitensprosse wiederholen die Erscheinungen der
Hauptsprosse im allgemeinen. Durch Verfaulung von Rhizomstucken kommt
es zu Einzelindividuen, niederzahligen Zwergexemplaren, die sich auch durch
kunstliche Zerstiickelung erzielen lassen.

Von besonderem Interesse sind weiter die Resultate, welche fiir die
Abhangigkeit der Variabilitat von aufieren Bedingungen vorliegen. Verf.
hat in bezug auf Unterlage, Pflanzengenossenschaft usw. verschiedene Stand-
orte getrennt untersucht. Er findet eine enge Beziehung zur Unterlage.
Kalkboden begunstigen Plus-, Kieselboden Minus varian ten. Auch wirkt Boden-
feuchtigkeit, Belichtung usw. erhohend. Sehr interessant ist es, die Angaben
uber die Variabilitat in Beziehung zu den Bestanden im einzelnen zu ver-
folgen. Mykorrhizenbildung scheint Plusvarianten zu begunstigen.

192

Referate.

Sodann glaubt Verf., dafi sich eine gewisse Erblichkeit der Glieder-
zahlverhaltnisse darin kund tut, dafi die Nachkommen hochzahliger Sprosse
denen niederzahliger in mancher Hinsicht iiberlegen sind. Sie keimen
rascher, wahrscheinlich auch in grofierer Anzahl und steigen in ihrer Ent-
wicklung schneller zu hoheren Gliederzahlen empor. Das durch Aussaat-
versuche gewonnene Material ist zwar nicht sehr umfangreich, die soeben
ausgefiihrte Uberlegenheit der Abkommen von Plusvarianten ist aber nicht
zu bezweifeln. Indessen lafit sich hier wohl sicher nicht von Vererbung,
sondern nur von einem ernahrungs-modifikatorischen Einflufi der Elter-
generation sprechen, etwa wie bei der sogenannten kongenitalen Vererbung,
zieht sich ja doch durch die gesamten hier erorterten Variationsverhaltnisse
der Einflufi der Ernahrung auf die Variabilitat der Gliederzahl. Eine exakte
Vererbungsuntersuchung von Paris stofit aber auf ganz erhebliche technische
Schwierigkeiten, welche hauptsachlich in der langen Dauer bis zum Eintritt
der Bliihreife begriindet sind.

Die Schwankungen der Gliederzahl bei Paris quadrifolia werden dann
auch unter phylogenetischen Gesichtspunkten erortert. Verf. schliefit auf
eine standige Zunahme und gleichzeitig damit eine entsprechende Vermehrung
der Quirlzahlen in der Trillium- Pam-Gruppe.

Es ist hervorzuheben, dafi samtliche Zahlenwerte auf exakt kritischen
Werten beruhen.

Die Untersuchungen sollen noch weiter auf die Blutenglieder ausgedehnt
werden. Ref. wiirde als besonders wiinschenswert erachten, wenn dabei
auch der Korrelationsverhaltnisse der Zahlen in den aufeinanderfolgenden
Wirteln eingehend gedacht wiirde.

Alles in allem bietet die Arbeit eine Fulle interessanter Tatsachen
und es ware nur zu wiinschen, dafi bald ahnlich vielseitige Variationsunter-
suchungen an anderen Pflanzen angestellt wiirden. e. Lehmann.

Zeitschrift fiir induktive Abstammungs- und Vererbungslehre

Inhaltsverzeichnis von Bd. XVI Heft 1/2

Abhandlungen Seite

Iwanow, E. und Philiptschenko,Jur., Beschreibung von Hybriden

zwischen Bison, Wisent nnd Hausrind 1 — 48

White, Orland E., Studies of Teratological Phenomena in their

Relation to Evolution and the Problems of Heredity .... 49 — 185

Referate

Fiirst, Th., Untersuchungen iiber Variationserscheinungen beim Vibrio

Finkler-Prior (Lehmann) 186

Kie filing, L., Erbanalytische Untersuchungen iiber die Spelzenfarbe
des Weizens. Ein Beitrag zur angewandten Vererbungslehre

(Fruwirth) 189

' Lang, Arnold, Die experimentelle Vererbungslehre in der Zoologie

seit 1900 (Baur) 186

Stark, Peter, Untersuchungen iiber die Variability des Laubblatt-

quirls bei Paris quadrifolia (Lehmann) 191

[Reprinted from Psyche, Vol. XXI, No. 1]

OBSERVATIONS ON THE RELATION BETWEEN FLOWER
COLOR AND INSECTS.

By E. M. East and R. W. Glaser.
Bussey Institution, Harvard University.

In 1909 a cross was made between the small red flowered Nico-
tiana forgetiana Hort (Sand) and Nicotiana alata Lk. and Otto yar.
grandiflora Comes, the large white N. affinis of horticulture, for
the purpose of studying certain problems of heredity. About
fourteen thousand plants of the second, third and fourth hybrid gen-
eration have been grown, and it has been established beyond a rea-
sonable doubt that each plant is completely self-sterile though it
crosses easily with any of its neighbors. Several hundred carefully
controlled self-pollinations have not yielded a single seed, while
histological studies have shown self-fertilization to be practically
impossible. On the other hand, hundreds of artificial cross pollina-
tions have yielded capsules full of seed in almost every instance,
showing with what ease cross-fertilization takes place, for artificial
pollination is usually not as successful as natural pollination.
The fact that every capsule formed naturally on these plants must
have resulted from a cross-pollination produced by an insect,,
serves to excuse our adding to the already huge literature on the
relations between insects and plants. The sixteen different color
forms that have segregated from the original cross permit observa-
tions on the percentage of flowers cross-fertilized and the selective-
value, if any, of distinct color varieties.

Our knowledge of the behavior of insects relative to flowers has
been greatly extended during the past few years by the work of
Plateau, Forel, Lovell, Grsenicher and others, but it has resulted in
that obscurity which precedes aggregation and precipitation by
disclosing the marvelous complexity of the relation. The adjust-
ment between certain insect forms and certain types of flowers is.

Psyche

[February

just as obvious now as when pointed out by Sprengel, but few
entomologists or botanists will admit its adequate interpretation
by the simple natural selection idea as believed by Hermann
Miiller and his followers who did not see the obstacles to this view
as plainly as did Darwin.

The attitude of botanists has been affected chiefly by genetic
investigation. Mendelian research and hypotheses regarding
mutational evolution have at least gained a serious reconsideration
of the origin, inheritance, and cause of survival of flower forms.
Investigations on cross- and self-fertilization, by giving a clear
and reasonable interpretation of the vigor of first generation
hybrids and the converse — the apparent deterioration through
inbreeding hybrids — -have caused us to view mechanisms for cross-
pollination at a new angle. Self-pollination gives inherently
stranger races (vigor not masked by heterozygosis) and insures
reproduction, but practically precludes the trial of variations not
of decisive value or of various recombinations of new variations
with old characters. On the other hand, cross pollination, while
permitting the survival of weak types through the vigor of hetero-
zygosis, and while rendering reproduction more dubious, does
assure a trial of all new variations in all the combinations possible
in a mendelian sense.

The appreciation of the intricacy of the behavior of insects to-
ward flowers is due primarily to the knowledge of insect sense or-
gans, to the ingenuity of the experiments of animal psychologists,
and to the passing of the tendency to interpret all the actions of
the lower animals as tropisms.

For these reasons the question as to whether particular flower
colors have a survival value due to the preference of certain insects
for them, upon which we have gathered a few data, would probably
be answered somewhat as follows by the majority of biologists.
Excluding any question of olfactory sense, it may be assumed that
insects perceive color differences from short distances but seldom
if ever exercise a choice. Night flyers, of course, perceive white
much more easily than colors. These conclusions are supported
by the data in the following table :

1914] East and Glaser — Relation Between Flower Color and Insects 29

T^rkfol nninnpr

AlJLtll II 111 1 I'M |

rf»r ppnt of

KJl llLIWClo vJll

vJl 11 W CI 3

LX\J YV KsL O

10 average plants.

fertilized.

White

18,035

7,052

39.10

Yellow

26,686

4,836

18.12

Red

14,165

2,154

15.21

Purple

9,721

1,628

16.74

Ten average plants of each of the four colors — white, yellow,
red and purple — were selected at random. The total number of
flowers produced on each color type during the flowering season
(July 15 to October 15) was determined by counting the places
on the racemes where flowers had been. The number of capsules
present was assumed to be the number of flowers fertilized, al-
though this count is not as accurate as the first by reason of the
accidental loss of capsules. Long experience with Nicotianas,
however, leads us to believe that this error is small.

The first point to be noted is the comparatively small percentage
of cross-pollination by insects. Numerous experiments on artificial
cross-pollination have shown that a very small amount of pollen
causes normal development of the capsules, yet the yellow, red
and purple types had only about 17 per cent, of their blossoms
crossed. According to the table, the percentage of white flowers
fertilized was more than twice as high as any of the colored types.
The reason for this is obvious. From the beginning of the flower-
ing period, about July 15, to the end period of summer heat, about
September 15, the flowers opened at about 4 p. m. and remained
open until about 7.30 a.m. During the last month of flowering,
the weather was so cool that the flowers also were open throughout
the day. Nearly two thirds of the fertilizations occurred diB%tg
the last month as could be determined by the positions of the
flowers on the racemes. Furthermore the percentage of fertiliza-
tions on the white type during the last month was about the same as
on the colored types. Roughly, one might say then that about 6 per
cent, of the pollinations of the colored types were made by night-
flyers (Sphingidse, etc.), while during the same period these insects

Psyche

[February

pollinated from 20 per cent, to 25 per cent, of the white type. In
other words, there was a high rate of selection of white flowers
during the period when the flowers were pollinated at night, but
there was no selection of colors when daylight pollinations were
made by the Hymenoptera and Diptera that frequented the plants.

Reprinted from the Botanical Gazette, 57: No. 3, March 1914

CURRENT LITERATURE

BOOK REVIEWS
Genetics

Some one once said, perhaps more epigrammatically than truthfully, "the
progress of a science is in direct proportion to the mathematics used in its
development." Whether generally true or not, the constant and rapid progress
of genetics since the introduction of Mendel's mathematical notation is a great
argument in favor of the statement. At the same time, the chaos that can
result from the unwarranted use of mathematics without other premise or
analysis is only too familiar to biologists. It has seemed as if those best trained
in mathematics were the first to forget that their science is merely a shorthand
method of stating the facts, that no more can come out than goes into the
mill, though it should come out in a shape more conducive to thorough mental
digestion. The slogan of certain biometricians, "there are no premises, all is
treatment," has brought many biologists to that state of mind in which they
could take seriously Poe's sly dig in the "Purloined Letter." In speaking of
the necessity of putting oneself in the mental attitude of the thief if the hiding
place of the stolen letter were to be discovered, he says: "As poet and mathe-
matician, he (the thief) would reason well; as mere mathematician he could
not have reasoned at all."

It remained for Johannsen to prove that he is poet, biologist, and mathe-
matician, by showing some four years ago the true relation of Karl Pearson's
beautiful developments of mathematical methods to genetic research. The
motto through the whole 25 chapters of his 500-page book was: "Wir mussen
die Erblichkeitslehre mit Mathematik, nicht aber als Mathematik treiben!"
Johannsen's work on the comparative permanence of homozygous types pub-
lished under the title Ueber Erblichkeit in Populationen und in reinen Linien
(1903) had already been enthusiastically received by many investigators, partly
by reason of the author's mastery of a persuasive style and partly because the
conclusions fitted data with which his readers were personally familiar. For
these reasons, this elaboration of his ideas met with a cordial reception that is
not the fate of many textbooks. But one unfavorable criticism of any impor-
tance could be made. The author did not treat adequately the numerous
genetic researches in which the problems of heredity had been attacked by
methods unlike his own. There is no hesitancy, therefore, in saying that th
new edition,1 with its 30 chapters and 722 pages, to which this criticism may not

Johannsen, W., Elemente der exakten Erblichkeitslehre. Zweite Auflage.
8vo. pp. xi+723. figs. 33. Jena: Gustav Fischer. 191 3.

239

240

BOTANICAL GAZETTE

[3IARCH

be applied with justice (if one excepts cytological research) , will be a welcome
addition to genetic literature.

In its present form, the work might very easily be divided into two books
with separate titles that could be used independently. The one is a thorough
introduction to statistical methods as they should be used in the service of
biology; the other is a well balanced discussion of the present status of genetic
conceptions.

As might be expected, it has been the general discussion of heredity that
has received the bulk of the revision; the chapters on biometry were admirably
done in the first edition, and the static nature of their substance was such that
little change has been necessary. Scarcely a word has been altered in the first
five chapters, though Charlier's short method for determining the standard
deviation has been added. In chapter 6 the discussion of mean error has been
revised and a demonstration from the domain of plant physiology has been
added. From this point to chapter 22, only chapters 12 and 13 are new, but
the remainder of the book is entirely as written.

In chapter 12 the more recent investigations concerning the possible effect
of selection on pure lines are described, while in the next chapter the "misunder-
standings" of certain authors who have opposed the theory of permanence of
homozygous types are taken up and disposed of with very clear logic, though
the style of the rejoinder is sometimes a little caustic.

The last seven chapters of the book are so crowded with information that
only a hint as to their contents can be given. They must be read by all who
are interested in genetics. Sixty pages are given up to the influence of the
factors of environment on variation and 160 pages to Mendelism in its various
phases, including heterozygosis, inbreeding, sterility, coupling, and sex determi-
nation. Mutations are considered rather concisely in the next to the last
chapter, the author being rather of the opinion that the peculiar behavior of
Oenothera Lamarckiana will ultimately be shown to be the result of segregation
and recombination, as has been suggested recently by Heribert-Xilsson.
The final chapter is a resume, with observations on eugenics, race hygiene, and
evolution.

With reference to the position taken in his earlier work concerning the
action of selection, the author remains as firm as a rock. He adds further data
of his own to support his position and shows very clearly that the seemingly
opposing conclusions of various investigators either are due to fallacious
reasoning or are based upon material that is not easily divested of complications
that confuse the main issue. To critics who deal only with generalities he
makes the following reply that may well be taken to heart by those who deal
with evolution from an easy chair:

Man hat mich kurzsichtig genannt, in Bezug auf die Selektion. Ich konstatiere
dies mit Vergniigen; die Pramissen einer oft maszlosen spekulativen Fernsichtigkeit
waren ja gerade zu untersuchen und wiirden wertlos gefunden.

CURRENT LITERATURE

241

It will doubtless surprise many that Johannsen maintains a firm
Lamarckian attitude throughout his book, dealing particularly sympathetically
with the work of Semon. He says: "Man hat mich ferner 'reiner Weisman-
nianer' genannt. Jeder solche 'man' hat mein Buch nicht gelesen oder
nicht verstanden." The reviewer must admit, therefore, that he has not
understood the author, for after reading the volume he is still firmly convinced
that in its essentials it is more nearly Weismannian than Lamarckian. Of
course he would not accuse the author of maintaining the morphological
hypotheses of Weismann with the biophores, determinants, and ids all built
into a beautiful structure, but the germ-to-germ inheritance, the dependence
of transmissible qualities upon germinal constitution, the invalidity of any
particular assumption as to breeding power from the appearance of the soma,
and the comparative freedom of the germinal substance from the influence of
ordinary environmental changes, as maintained throughout the work, will be
classed by most biologists as belonging rightly within the scope of Weismann's
conception of heredity.

Very few new terms are introduced by Johannsen in this edition of his
book, but two have appeared that seem justified in spite of the abuse that has
been showered on the roots used. Individuals that belong to the same pheno-
type are "isophenous"; individuals that belong to the same genotype are
"isogenous." In addition he has adopted Webber's term "clone" for a bud
individual.

Taken all in all, one must be very critical to have anything but praise for
the new Erblichkeitslehre, and it is confidently predicted that it will long remain
a classic. — E. M. East.

Reprinted from the Proceedings of the National Academy of Sciences
Vol. I, p. 95, 1915

AN INTERPRETATION OF SELF-STERILITY

By E. M. East

BUSSEY INSTITUTION. HARVARD UNIVERSITY
Piesented to the Academy. December 28. 1914

In certain hermaphroditic animals and plants, self-fertilization is
often impossible. This gametic incompatibility has been called self-
sterility. In the vegetable kingdom it is known to be comparatively
widespread; in the animal kingdom, though it may be found later to be
characteristic of many species, as yet only the Ascidian Ciona intestin-
alis has furnished material for study of the problem. (See Morgan,1
Adkins, in Morgan,2 and Fuchs.s)

Ciona is not perfectly self-sterile. Individuals appear to vary in de-
gree of self- sterility, though no case has yet been found where self-
fertility is equal to cross-fertility. Morgan believes that there is a
great difference in the compatibility of ova to sperm from other indi-
viduals, though Fuchs maintains that 100% of segmenting eggs can be
obtained in every cross with normal ova if a sufficiently concentrated
sperm suspension is used.

Fuchs has shown a chemical basis for the phenomenon by the differ-
ence in ease of cross-fertilization after contact of ova with sperm from
the same animal and by the variation in ease of self-fertilization after
certain artificial changes in the chemical equilibrium of the medium
surrounding the ova, and by this work has brought the matter of self-
sterility in Ciona in line with that in Angiosperms as worked out by
Jost.4

Jost has shown that in the plants with which he worked only short
tubes were formed after pollination with pollen from the same plant,
though the necessary length of pollen-tube was easily developed after
cross-fertilization. He saw as cause of these phenomena a chemotropism
due to the presence of 'individueller Stoffe.' Pollen was indifferent to
Tndividualstoff' from the same plant, but was stimulated by that from
other plants.

To Correns5 such an explanation of self-sterility seemed too general.
He believed that a simple interpretation would account for the results
he had obtained from Cardamine pratensis. Two plants B and G were
crossed reciprocally and sixty of the offspring tested by pollinating from
the parents, on the parents and from sisters. The back crosses of (B X G)
or (G X B) with B and with G seemed to him to indicate four equal-
sized classes with reference to gametic compatibility: (1) plants fertile

GENETICS: E. M. EAST

with both B and G; (2) plants fertile with B but not with G; (3) plants
fertile with G but not with B ; (4) plants fertile with neither B nor G.

These facts were interpreted by assuming the existence of two inde-
pendently inherited factors that inhibit the growth of pollen-tubes.
Representing these factors by the letters B and G, the original plants
must be supposed to have had the formulae Bb and Gg respectively,
since it is clear that type BB and GG could never be formed. When Bb
is crossed with Gg the four types BG, Bg, bG and bg should result, of
which the first three should be self-sterile. Plants BG should be fertile
with plants bg, plants Bg should be fertile with bG and bg, plants bG
should be fertile with Bg and bg, while plants bg should be self-fertile
as well as cross-fertile with the other three classes. Attractive as this
theory is, it is not clearly in accord with the facts. Plants of the type bg —
inherently self -fertile — were not found, and the other classes showed many
discrepancies.

Morgan2 has offered another hypothesis that fits the data from both
plants and animals. If I have not misunderstood the meaning of his
rather general statement of the proposition, my own theory is only
an extension of it, laid down perhaps a little more specifically. He
says:

The failure to self-fertilize, which is the main problem, would seem to be
due to the similarity in the hereditary factors carried by the eggs and sperm;
but in the sperm, at least, reduction division has taken place prior to fertili-
zation, and therefore unless each animal was homozygous (which from the nature
of the case cannot be assumed possible) the failure to fertilize cannot be due
to homozygosity. But both sperm and eggs have developed under the influence
of the total or duplex number of hereditary factors: hence they are alike;
i.e., their protoplasmic substance has been under the same influence. In
this sense, the case is like that of stock that has long been inbred, and has
come to have nearly the same hereditary complex. If this similarity decreases
the chances of combination between sperm and eggs we can interpret the results.

My own work has been done with the descendants of a cross between
Nicotiana forgetiana (Hort) Sand., a small red-flowered species, and
Nicotiana alata Lk. and Otto. var. gr audi flora Comes, the large white-
flowered sort commonly known as Nicotiana affinis. Both parents were
undoubtedly self-sterile as over 500 plants of the Fi, F2, F3, and F<
generations have been found to be self-sterile by careful tests.

Several experiments were made in which selfing, crossing inter se,
and back crossing were done on a large scale, using plants of theF2,F3
and F4 generations which had segregated markedly in size and were of

GENETICS: E. M. EAST

at least eight different shades of color. In the F2 generation, twenty
plants coming from two crosses between Fi plants were selected for
experiment. Each was selfed many times and in addition 131 inter-
crosses were attempted, from four to twelve flowers being used in each
trial. All attempts at selfing failed, while only two attempts at crossing
were unsuccessful. Of the 129 successful inter-crosses, all but 4 pro-
duced full capsules, and it is probable that even this variability in ease
of cross-fertilization was caused by attending conditions. One hundred
and twenty other inter-crosses were made in the F2 generation, with
three failures.

In the F3 generation, about one hundred inter-crosses were made be-
tween twelve plants wThich were the progeny of two sister F2 plants.
Six of these attempts failed

In the F4 generation, fifty-eight inter-crosses were made between ten
plants that were the daughters of two F3 plants. Fifty- three of these cross-
fertilizations were successful.

Back crosses also were made in considerable numbers, though not to
the extent one might desire. Plants A, B, C and D were combined in
four different ways and among the plants resulting from these combi-
nations eighty-rive back crosses were attempted, of which five failed.

These facts will not fit any simple Mendelian formula similar to that
proposed by Correns; furthermore, data from an experiment of a differ-
ent kind appear to support Jost's idea of 'Individuals toff e' rather than
Correns' idea of inhibitors. Pairs of plants were provided to furnish
series of selfed and crossed flowers. The pistils of these flowers were
fixed at regular periods after pollination, stained, sectioned and the
pollen- tubes examined. Since the flowers on each plant had about
the same length of pistils, curves of pollen-tube development for both
crossing and selfing could be constructed. The pollen grains germinated
perfectly on stigmas from the same plant, from 1200 to 2000 tubes having
been counted in sections of single pistils. The difference between the
development of the tubes in the selfed and the crossed styles was wholly
one of rate of growth. The tubes in the selfed pistils developed steadily
at a rate of about 3 mm. per twenty-four hours, with even a slight ac-
celeration of this rate as the tubes progressed. If the flowers were
sufficiently long-lived, one could hardly doubt but that the tubes would
ultimately reach the ovules, though this would not necessarily mean that
fertilization must occur. Since the maximum life of the flower is about
11 days, however, the tubes never traverse over one-half of the distance
to the ovary. On the other hand, the tubes in the crossed pistils, though

GENETICS; E. M. EAST

starting to grow at the same rate as the others, pass down the style
faster and faster, until they reach the ovary in four days or less.

From these facts it seems reasonable to conclude that the secretions
in the style stimulate the pollen-tubes from other plants instead of in-
hibiting the tubes from the same plant.

The whole question, therefore, becomes a mathematical one, that of
satisfying conditions whereby no stimulus is offered to pollen- tubes from
the same plant, but a positive stimulus is offered to tubes from nearly
every other plant.

The nearly constant rate of growth of pollen-tubes in the pistils of
selfed flowers, compared with the regular acceleration of growth of the
tubes from the pollen of other plants, undoubtedly shows the presence
of stimulants of great specificity akin to the ' Individuals toff e' of Jost,
though I believe their action to be indirect. Experiments by several
botanists, which I have been able partially to corroborate, point to a
single sugar, probably of the hexose group, as the direct stimulant. The
specific 'Individuals toff e' I believe to reside in the pollen grains and to
be in the nature of enzymes of slightly different character, all of which
except the one produced by the plant itself for the use of its own pollen,
or by another individual of the same genotype, can call forth secre-
tion of the sugar that gives the direct stimulus. At least this idea
links together logically the fact of the single direct stimulus and the
need of 'Individuals to fie' to account for the results of the crossing and
selfing experiments But whether or not. this be the correct physiologi-
cal inference, the crossing and selfing experiments call for an hypothe-
sis that will account for no stimulation being offered the tubes from the
plant's own pollen while at the same time great stimulation will be given
the tubes from the pollen of nearly every other plant. This is a straight
mathematical problem, and it is hardly necessary- to say that it is in-
soluble by a strict Mendelian notation such as Correns sought to give.
This is obvious to anyone familiar with the basic mathematics of Men-
delism. On the other hand a near Mendelian interpretation satisfies
every fact.

Let us assume that different hereditary complexes stimulate pollen-
tube growth and in all likelihood promote fertilization, and that like
hereditary complexes are without such effect. One may then imagine
any degree of heterozygosis in a mother plant and therefore any degree
of dissimilarity between the gametes it produces, without there being the
possibility of a single gamete having anything in its constitution not
possessed by the somatic tissues of the mother plant. From the chromo-

GENETICS: E. M. EAST

some standpoint of heredity the cells of the mother plant are duplex in
their organization: they contain N pairs. The cells of the gametes
contain N chromosomes, one coming from each pair of the mother cell;
but they are all parts of the mother cell and contain nothing that that
cell did not contain. These gametic cells cannot reach the ovaries of
flowers on the same plant because they cannot provoke the secretion of
the direct stimulant from the somatic cells of that plant.

All gametes having in their hereditary constitution something different
from that of the cells of a mother plant, however, can provoke the
proper secretion to stimulate pollen-tube growth, reach the ovary before
the flower wilts, and produce seeds. Such differences would be very nu-
merous in a self-sterile species where cross-fertilization must take place;
nevertheless like hereditary complexes in different plants should be found,
and this should account for the small percentage of cross-sterility actually
obtained. It must be granted that this hypothesis satisfies the facts,
but that is not all. It is admittedly a perfectly formal interpretation,
but from a mathematical standpoint — granting the generality of Mendel-
ian inheritance — it is the only hypothesis possible that can satisfy the
facts.

In conclusion it should be mentioned that the cross-pollinated pistils
show a considerable variation in the rate of growth of individual pollen-
tubes, though our curves of growth have been made by taking the aver-
age rate of elongation. Is this variation a result of chance altogether or
must one assume a differential rate of growth increasing directly with the
constitutional differences existing between the somatic cells and the vari-
ous gametes? If it is assumed that any constitutional difference between
the two calls forth the secretion of the direct stimulus to growth, chance
fertilization by gametes of every type different from that of the mother
plant will ensue; if there is a differential rate, selective fertilization will
occur. This question cannot be decided definitely at present, but two
different lines of evidence point toward chance fertilization:

1. Flowers from a single plant pollinated by different males show no
decided difference in rate of fertilization.

2. Color differences are transmitted to expected ratios.

Further, it will be recalled that beginning with the F2 generation sister
plants crossed together have given us our F3 and F4 populations, and that
these F3 and F4 populations apparently have given a constantly increas-
ing percentage of cross-sterility. This is what should be expected under
the theory that a small difference in germ plasm constitution is as active
as a great difference in causing the active stimulation to pollen tube

100

GENETICS; E. M. EAST

growth. Breeding sister plants together in succeeding generations
causes an automatic increase of homozygosity as is well known. This
being a fact, cross- sterility should increase. Such an increase in cross-
sterility has been observed in the F3 and the F4 generations, but it would
not be wise to maintain dogmatically that it is significant.

1 Morgan, T. H., Some further experiments on self-fertilization in Ciona. Biol. Bull.,
8, 313-330 (1905).

2 Morgan, T. H., Hertdity and Sex. New York. Columbia University Press, ix + 1-282
(1913). (page cited 217).

3 Fuchs, H. M., On the conditions of self-fertilization in Ciona. Archiv. f. Entwickl. d.
Org., 40, 157-204 (1914); The action of egg-secretions on the fertilizing power of sperm.
Archiv. f. Entwickl. d. Org., 40, 205-252 (1914).

4 Jost, L., Zur Physiologie des Pollens. Ber. d. deul. Bot. Ztg., Heft V and VI (1907).

6 Correns, C, Selbststerilitat und Individualstoffe. Festschr. d. med. nat. Cesell. zur 84.
Versamml. deulsch. Naturforscher u. Arzle. Miinster i. W.f pp. 1-32 (1912).

THE PHENOMENON OF SELF-STERILITY

PROFESSOR E. M. EAST

NEW YORK
1915

[Reprinted without change of paging, from the American Naturalist, 191 5.]

[Reprinted from The American Naturalist, Vol. XLIX., Feb., 1915. |

THE PHENOMENON OF SELF-STERILITY 1

PROFESSOR E. M. EAST
Bussey Institution, Harvard Universitt

In both animals and plants in which the two sexes have
been combined in the same individual, cases have been
found where self-fertilization is practically impossible.
This gametic incompatibility has been called self-sterility,
although the term is hardly proper as applied to normal
functional gametes that may fuse with their complements
in the regular manner, provided each member of a pair
has been matured in a separate individual.

In plants the phenomenon has been known since the
middle of the nineteenth century, in animals a correspond-
ing discovery was made in 1896 by Castle, the species
being one of the Ascidians, Ciona intestinalis. During
the eighteen years that have passed since Castle's dis-
covery, Ciona has been studied on a large scale by Morgan
(1905), Adkins (Morgan, 1913), and Fuchs (1914). The
botanists, however, have lagged somewhat behind ; for, in
spite of having been acquainted with self-sterility in
plants for over half a century, and having found over
thirty species where a greater or less degree of self-
sterility occurs from which to select material, very few
thorough investigations into the physiology of the subject
have appeared.

The main facts regarding fertilization in Ciona intesti-
nalis are about as follows :

1. Under uniform suitable conditions, individuals vary
in degree of self-sterility, it being exceptional to find an
animal that is perfectly self-sterile.

2. Self-fertility has never equaled cross-fertility, though
the possibility remains that some animals may be self-

i Bead by title at tlie thirty-second meeting of the American Society of
Naturalists, December 31, 1914.

THE AMERICAN NATURALIST [Vol. XLIX

fertilized as easily as they may be crossed with certain
particular individuals.

3. The ease with which the ova of any animal "A" may
be fertilized by the sperm of other individuals may vary.

Morgan (1913) concluded from his own work and that
of Adkins that there were wide differences in the compati-
bility of ova to different sperm. Fuchs (1914) maintained
that 100 per cent, of segmenting eggs can be obtained in
every cross if the ova are normal and a sufficiently con-
centrated sperm suspension is used. It is possible that
Fuchs is correct and that varying concentrations of sperm
suspension were the cause of Morgan's and Adkins ?s re-
sults, yet the possibility of differences in this regard in-
herent in the individual is not to be overlooked. It will
be seen later that I regard the matter as of great impor-
tance to the general subject.

4. A chemical basis for self-sterility is shown in Fuch's
experiments by (a) the decrease in ease of cross-fertiliza-
tion after contact of ova with sperm from the same ani-
mal, and by (b) the difference in ease of self-fertilization
after various artificial changes in the chemical equilibrium
of the medium surrounding the ova.

From the botanical side various studies on the physiol-
ogy of self-sterility have appeared since such investiga-
tions were initiated by Hildebrand in 1866. At this time
itisnecessary for us to consider only those of Jost (1907),
Correns (1912), and Compton (1913).

Jost was able to show that in self-sterile plants tubes
formed from their own pollen were so limited in their
development that fertilization did not occur, although the
necessary length of pollen tube was easily developed after
a cross-fertilization. He saw as the cause of these phe-
nomena the presence of 1 ' individueller Stoff e. ' ' Pollen
was indifferent to "Individualstoff " of the same plant,
but was stimulated by that of other plants.

Correns (1912), working with one of the bitter cresses,
Cardamine pratensis, obtained results to which he gave a
simpler interpretation. Starting with two plants, B and

No. 578]

SELF-STERILITY

G, he crossed them reciprocally and tested 60 of the off-
spring by pollinating from the parents, on the parents,
and inter se. The back crosses of (B X G) or (G X B)
with B and with G apparently indicated fonr classes about
equal in size with reference to gametic compatibility:
(1) plants fertile with both B and G; (2) plants fertile
with B but not with G\ (3) plants fertile with G but not
with B\ (4) plants fertile with neither B nor G.

To these facts Correns gave a Mendelian interpretation
by assuming the existence of two factors each of which in-
hibits the growth of pollen tubes from like gametes. Bep-
resenting these factors by the letters B and G, it is clear
that types BB and GG could never be formed. The orig-
inal plants were supposed to be of classes Bb and Gg, re-
spectively. When crossed there resulted the four types
BG, Bg, bG and bg. Plants of types BG, Bg, and bG
should be self -sterile, while plants of the type bg should be
self-fertile. Plants BG should be fertile with plants bg*
plants Bg should be fertile with bG and bg, and plants bG
should be fertile with Bg and bg. As a matter of fact
Correns 's results were not clearly in accord with the
theory. Plants of the type bg were not self-fertile, and
the other classes of matings showed many discrepancies.
It is only fair to say, however, that the author recognized
some of these difficulties, but believed them to be due to
other inhibitors.

In a part of Compton's (1913) work, a still simpler
interpretation of self-sterility is offered, at least for a
particular case, that of Reseda odorata. Darwin's origi-
nal discovery that both self-sterile and self-fertile races
of this plant exist was confirmed and the following results
obtained in crossing experiments. Self-sterile plants
crossed either with self-sterile or with self- fertile plants
gave only self-sterile offspring. .Certain self-fertile

^jfc1!^6:^^^^ self~

^ pollinated^ Other gelf -fertile plants gave ratios of 3 self-
fertile to 1 self-JSSie offspring when self-pollinated, and
ratios of 1:1 when crossed with pollen from self-sterile

THE AMERICAN NATURALIST [Vol. XLIX

plants. For these reasons he regards self-fertility as a
simple Mendelian dominant to self-sterility in the case
studied. I believe Compton would draw no such sharp
line about self-sterility in general. In fact, he follows
Jost in suggesting the presence of a diffusible substance
in the tissues of the style and stigma which retards or
promotes pollen tube growth after self-pollination or
cross-pollination in some manner analogous to the mech-
anism that promotes animal immunity or susceptibility
after infection.

The only alternative general hypothesis has been pro-
posed by Morgan, and this can be discussed more advan-
tageously after the presentation of my own work, of which
only an abstract will be given at this time.

In 1909 I made a cross between a small red-flowered
Nicotiana, Nicotiana forgetiana (Hort.) Sand, and the
large white-flowered Nicotiana of the garden Nicotiana
alata Lk. and Otto. var. grandifiora Comes. All of the
plants of the YY generation appeared to be self-sterile.
Tests of Nicotiana forgetiana2 have shown these plants
also to be self-sterile, but both self-fertile and self-sterile
plants of the other parent have been found. From data
gathered later, there seems to be no doubt that a self-
sterile plant of Nicotiana alata grandifiora, was used in
the actual cross. This conclusion seems reasonable in
view of the fact that of over 500 plants of the Flf F2, F3
and F4 generations tested, not a single self-fertile plant
was found.

The plants of the Fj generation were all vigorous and
healthy, and in spite of the fact that they resulted from a
species cross which Jeffrey claims always produces large
amounts of abnormal pollen, a large number of examina-
tions of pollen from different individuals showed from 90

2 I thought originally that both of these species (East, 1913) were self-
fertile. Seed had been obtained from a carefully bagged inflorescence of
each species in 1909. Either the plant of N. forgetiana which gave this
seed was self-fertile — something that I have never been able to find since
that time — or there was an error in manipulation. At any rate, the plants
resulting from this seed were all self -sterile.

No. 578]

SELF-STERILITY

to 100 per cent, of morphologically perfect pollen grains,
a condition about the same as was found in the pure spe-
cies. To this statement there is one exception. A single
plant was found with only about 2 per cent, of good sound
pollen.

Several experiments were made in which crossing and
selfing was done on a large scale, using plants of the F2,
F3 and F4 generations which had segregated markedly in
size and were of at least 8 different shades of color. In
one of these experiments 20 plants of the F2 generation
coming from 2 crosses of F1 plants were used. It was
planned to make all possible combinations of these plants,
400 in all. This task proved overburdensome, however,
and in addition to the self-pollinations but 131 inter-
crosses were made with the following results.

1. Each plant was absolutely self-sterile.

2. Leaving out of consideration the plant with shrunken
imperfect pollen only two crosses failed. This failure of
1.5 per cent, of the crosses may have been due to im-
proper conditions at the time of the attempts, but as a
number of trials were made the possibility remains that
there is a small percentage of true cross-sterility.

3. Of the 129 successful inter-crosses, 4 produced cap-
sules with less than 50 per cent, of the ovules fertilized.
The remaining crosses produced full capsules. It is
barely possible that this result shows a slight variability
in ease of cross-fertilization, but I am more inclined to
believe that these 4 cases where a low percentage of fer-
tilized ovules were obtained were accidental.

Other crossing experiments of the same kind have cor-
roborated these results. Out of 120 inter-crosses, only 3
failed.

Later, something over 100 inter-crosses were made be-
tween 12 plants of an F3 population resulting from cross-
ing two sister F2 plants. Six of the attempts at cross-
fertilization— 3 to 8 trials per plant being made— were
failures. These plants as well as others tested were com-

THE AMERICAN NATURALIST [Vol. XLIX

pletely self-sterile, and apparently there was cross-steril-
ity in about 6 per cent, of the possible combinations.

In the F4 generation, 10 plants resulting from crossing
two sisters of the F3 generation were selected for experi-
ment. Unfortunately, I was able to make only 58 inter-
crosses, 5 of which, almost 10 per cent., failed.

Back crosses have furnished another line of experiment,
though they have not been carried on as systematically as
were those of Correns. Nearly 85 back-crosses using
plants from the progeny of four combinations which in-
cluded four individuals as parents, have been made. The
plants themselves all proved self-sterile, and in addition
5 of the back crosses failed.

When these experiments were begun I expected to find
that the facts would accord with a simple dihybrid Men-
delian formula similar to that which Correns later pro-
posed as an interpretation of his results, yet only by con-
siderable stretching and a vivid imagination will Cor-
rens 's data fit such an hypothesis, and my own data do
not fit at all. No self-fertile plants have been produced
by any combination, and cross-sterility is a possibility in
only from 1.5 to 10 per cent, of the combinations. Fur-
thermore, Correns 's idea of inhibitors appears unlikely
from some other data I have gathered with the help of
Mr. J. B. Park. Ten plants were involved in this experi-
ment. Paifs of plants were provided to furnish series
of selfed and crossed flowers. The pistils of these flowers
were fixed at regular periods after pollination, stained,
sectioned, and the pollen tubes examined. Fertilization
not later than the fourth day marked the end point of the
crossed series, the dropping of the flowers between the
eighth and the eleventh day ended the selfed series. As
the flowers on each plant had about the same length pistils,
curves of pollen tube development for both crossing and
selfing could be constructed. The pollen grains germi-
nated perfectly on stigmas from the same plant, from
1,200 to 2,000 tubes having been counted in sections of
single pistils. The difference between the development

No. 578]

SELF-STERILITY

of the tubes in the selfed and the crossed styles is wholly
one of rate of growth. The tubes in the selfed pistils de-
velop steadily at a rate of about 3 millimeters per twenty-
four hours. There is even a slight acceleration of this
rate as the tubes progress. If the flowers were of an
everlasting nature one could hardly doubt but that the
tubes would ultimately reach the ovules, though this would
not necessarily mean that fertilization must occur. Since
the maximum life of the flower is about 11 days, however,
the tubes never traverse over one half of the distance to
the ovary. On the other hand, the tubes in the crossed
pistils, though starting to grow at the same rate as the
others, pass down the style faster and faster, until they
reach the ovary in four days or less.

From these facts it seems reasonable to conclude that
the secretions in the style offer a stimulus to pollen tubes
from other plants rather than an impediment to the de-
velopment of tubes from the same plant.

The whole question, therefore, becomes a mathematical
one, that of satisfying conditions whereby no stimulus is
offered to pollen tubes from the same plant, but a positive
stimulus is offered to tubes from nearly every other plant.

Morgan has given an answer to this question in a gen-
eral way. If I understand his position correctly, my own
conclusions are not very different from his, but are some-
what more definite. Morgan (1913) states that the re-
sults of Adkins and himself on dona intestinalis can best
be understood by the following hypothesis :

The failure to self -fertilize, which is the main problem, would seem
to be due to the similarity in the hereditary factors carried by the eggs
and sperm; but in the sperm, at least, reduction division has taken
place prior to fertilization, and therefore unless each animal was
homozygous (which from the nature of the case cannot be assumed
possible) the failure to fertilize can not be due to homozygosity. But
both sperm and eggs have developed under the influence of the total or
duplex number of hereditary factors; hence they are alike, i. e.} their pro-
toplasmic substance has been under the same influences. In this sense,
the case is like that of stock that has long been inbred, and has come
to have nearly the same hereditary complex. If this similarity decreases

THE AMERICAN NATURALIST

[Vol. XLIX

the chances of combination between sperm and eggs we can interpret
the results.

I make this quotation to show Morgan's viewpoint. It is
for him to say whether the following conclusions are ex-
tensions of his own or not.

The tolerably constant rate of growth of pollen tubes in
• the pistils of selfed flowers, compared with the great ac-
celeration of growth of the tubes from the pollen of other
plants as they penetrate nearer and nearer to the ovary,
undoubtedly shows the presence of stimulants of great
specificity akin to the "IndividualstorTe" of Jost. We
are wholly ignorant of the nature of these stimulants, but
I am inclined towards a hypothesis differing somewhat
from his. Experiments by several botanists, which I
have been able partially to corroborate, point to a single
sugar, probably of the hexose group, as the direct stimu-
lant. The specific "Individualstoffe" I believe to reside
in the pollen grains and to be in the nature of enzymes of
slightly different character, all of which, except the one
produced by the plant itself for the use of its own pollen
or by other plants of identical germinal constitutions,
can call forth secretion of the sugar that gives the direct
stimulus. At least, this idea links together logically the
fact of the single direct stimulus and the need of "Indi-
vidualstoffe" to account for the results of the crossing and
selfing experiments. But whether or not this be the cor-
rect physiological inference, the crossing and selfing ex-
periments call for a hypothesis that will account for no
stimulation being offered the tubes from the plant's own
pollen, while at the same time great stimulation is given
the tubes from the pollen of nearly every other plant.

This is a straight mathematical problem, and it is
hardly necessary to say that it is insoluble by a strict
Mendelian notation such as Correns sought to give. This
is obvious to any one familiar with the basic mathematics
of Mendelism. On the other hand, a near Mendelian in-
terpretation satisfies every fact.

Let us assume that different hereditary complexes stim-

No. 578]

SELF-STERILITY

ulate pollen tube growth and in all likelihood promote fer-
tilization, and that like hereditary complexes are without
such effect. One may then imagine any degree of hetero-
zygosis in a mother plant and therefore any degree of
dissimilarity between the gametes it produces, without
there being the possibility of a single gamete having any-
thing in its constitution not possessed by the somatic tis-
sues of the mother plant. From the chromosome stand-
point of heredity the cells of the mother plant are duplex
in their organization; they contain N pairs. The cells
of the gametes contain N chromosomes, one coming from
each pair of the mother cell ; but they are all parts of the
mother cell and contain nothing that that cell did not con-
tain. These gametic cells can not reach the ovaries of
flowers on the same plant because they can not provoke
the secretion of the direct stimulant from the somatic cells
of that plant.

All gametes having in their hereditary constitution
something different from that of the cells of a mother
plant, however, can provoke the proper secretion to stim-
ulate pollen tube growth, reach the ovary before the flower
wilts and produce seeds. Such differences would be very
numerous in a self-sterile species where cross-fertilization
must take place; nevertheless like hereditary complexes
in different plants should be found, and this should ac-
count for the small percentage of cross-sterility actually
obtained. It must be granted that this hypothesis satis-
fies the facts, but that is not all. It is admittedly a per-
fectly formal interpretation, but from a mathematical
standpoint,— granting the generality of Mendelian inheri-
tance,— it is the only hypothesis possible that can satisfy
the facts.

Let us now look into a few of the ramifications of the
subject. Examinations of the pistils that have been sec-
tioned after cross-pollination show a considerable varia-
tion in the rate of growth of individual pollen tubes,
though our curves of growth have been made by taking
the average rate of elongation. Is this variation a result

THE AMERICAN NATURALIST

[Vol. XLIX

of chance altogether or must we assume a differential rate
of growth increasing directly with the constitutional dif-
ferences existing between the somatic cells and the vari-
ous gametes ? If we assume that any constitutional dif-
ference between the two calls forth the secretion of the
direct stimulus to growth, chance fertilization by gametes
of every type different from that of the mother plant will
ensue ; if there is a differential rate, selective fertilization
will occur. This question can not be decided definitely at
present, but two different lines of evidence point toward
chance fertilization.

1. Flowers from a single plant pollinated by different
males show no decided difference in rate of fertilization.

2. Color differences are transmitted in expected ratios.
Further, it will be recalled that beginning with the F2

generation, sister plants crossed together have given us
our F3 and F4 populations, and that these F3 and F4 popu-
lations apparently have given a constantly increasing per-
centage of cross-sterility. This is what should be ex-
pected under the theory that a small difference in germ
plasm constitution is as active as a great difference in
causing the active stimulation to pollen tube growth.
Breeding sister plants together in succeeding generations
causes an automatic increase of homozygosity as is well
known. This being a fact, cross-sterility should increase.
Such an increase in cross-sterility has been observed in
the F3 and the F4 generations, but it would not be wise to
maintain dogmatically that it is significant.

There are various questions, including the important
one of the origin of self-sterility, that can not be discussed
at this time. In conclusion, therefore, let us turn once
more to the phenomenon of self-sterility in Ciona intes-
tinalis. It seems to me that the hypothesis outlined above
has few, if any, drawbacks when applied to self-sterility
in plants. The question there, as far as we have gone, is
one of pollen tube growth, and the theory that the secre-
tion of the direct stimulant can be called forth only by a
gamete that differs in its constitution from the somatic

No. 578]

SELF-STERILITY

cells between which the pollen tube passes, is logical. If
the same theory is to be extended to animals, however, it
follows that the external portions of the membranes of
the animal egg that have been shown by the wonderful in-
vestigations of Loeb and of Lillie to have such important
functions, must be functionally zygotic in character. I
am aware that this suggestion may be considered pretty
radical, but it certainly should be given consideration.
I do not like to draw an analogy between the animal egg
and a pollen grain, but it may be mentioned that in these
structures— surely comparable to the animal egg in the
fineness of their membranes and walls— both color and
shape are inherited as if they were zygotic in nature.

December 5, 1914.

LITERATURE CITED

Castle, W. E. The Early Embryology of Ciona intestinalis Flemming (L.). .

Bull. Mus. Comp. Zool., Harvard University 27, 201-280. 1896.
Compton, E. H. Phenomena and Problems of Self -sterility. New Fhytolo-

gist, 12, 197-206. 1913.
Correns, C. Selbststerilifat und IndividalstofTe. Festschr. d. mat.-nat.

Gesell. zur 84. Versamml. deutsch. Naturforscher u. Arzte, 1912.

Minister i. W., pp. 1-32.
East, E. M. Inheritance of Flower Size in Crosses between Species of

Nicotiana. Bot. Gaz., 55, 177-188. 1913.
Fuchs, H. M. On the Conditions of Self-fertilization in Ciona. Archiv.

f. EntwicTcl. d. Org., 40, 157-204. 1914.
The Action of Egg-secretions on the Fertilizing Power of Sperm.

Archiv. f. Entwickl. d. Org., 40, 205-252. 1914.
Hildebrand, F. Ueber die Nothwendigkeit der Insektenhilfe bei der Be-

fruehtung von Corydalis cava. Jahrb. wiss. Bot., 5, 359-363. 1866.
Jost, L. Zur Physiologie des Pollens. Ber. d. deut. bot. Gesell., 23, 504-

515. 1905.

TJeber die Selbststerilitat einiger Bliiten. Bot. Ztg., Heft V and VI.

1907.

Morgan, T. H. Some Further Experiments on Self-fertilization in Ciona,.

Biol. Bull., 8, 313-330. 1905.
Heredity and Sex. New York. Columbia Univ. Press, pp. ix -f- 1-

282. 1913 (page cited 217).

E. S. Carman

E. M. EAST

Reprinted, without change of paging,
from the Journal of Heredity (Organ of the
American Genetic Association), Vol. VI,
No. 2, Washington, D. C, February. 1915.

E. S. CARMAN

One of the Greatest of American Plant Breeders — His Work Too Little
Appreciated — Success With Potatoes Most Noteworthy — His
Activity as a Journalist.

E. M. East
Bussey Institution, Forest Hills, Massachusetts.

IT IS a delightful epigram but hardly
the actual truth that "If a man
preach a better sermon, write a
better book, or build a better
mouse-trap than his neighbor, though
he hide himself in the wilderness, the
world will make a beaten path to his
door." The world as a whole is likely
to give its applause to some very unim-
portant people. And after all is it not
probable that too general a commenda-
tion encourages superficial rather than
solid work? The anti-socialistic argu-
ment that a more even distribution of
earthly comforts would oppose progress
because it limits ambition is a pure
sophism. Few things worth doing have
been done with either money, power or
fame in view. For this reason there is
no need to feel sorry that E. S. Carman,
great alike as agricultural journalist,
public spirited citizen and creator of
new varieties of plants, never received
the panegyrics of which some others
have been since the recipients. He had
the happiness described by Marcus
Aurelius: "A man's happiness — to do
the things proper to man." Not that
Mr. Carman was unknown — perhaps
the editor of no rural paper was admired
and trusted more — but, even with the
temptation of a private medium for
exploiting his triumphs, he did no more
than describe carefully and impartially
success and failures alike with the
honesty of a true nature-lover and born
investigator.

Mr. Carman would probably have
denied that he was a great plant
breeder. He originated no new methods
and made few contributions to the
study of heredity; but he did discover
many interesting facts during his hy-
bridization experiments and he added

hundreds of millions of dollars to the
wealth of the country, keeping nothing
for himself. He was a national bene-
factor, and who will say he was not a
great man when he placed public
service before private gain? His atti-
tude in the matter is summed up in the
final paragraph of an article on the five
famous potato varieties placed on the
market between 1882 and 1896. "It
will now appear that for our 16 years of
potato work, we have sold five kinds for
precisely $1,000. We dare say that,
had we used our columns for advertising
the three kinds now offered for sale,
retaining the entire control as long as
possible, The Rural New-Yorker might
easily have made a snug little fortune.
But, tell us friends, were we to crack up
the plants that have originated at the
"Rural Grounds" while we sold them to
you either directly or indirectly, do you
think that you would place as much
confidence in the thorough impartiality
of our plant reports, as you do now?"
Ten years ago the writer made a trip
through the great potato regions of
Wisconsin and Minnesota. During it
one of the most successful and best in-
formed growers stated that in the
previous decade 80% of the potatoes of
the country were either Mr. Carman's
productions or seedlings from them.
How much truth there was in this state-
ment it is impossible to say, but dis-
count it as much as one will, can it be
said that there is no such thing as
altruism ?

POTATO CREATIONS.

The famous potatoes from the Rural
Grounds were Rural Blush, Rural New-
Yorker No. 2, Carman No. 1, Carman
No. 3 and Sir Walter Raleigh. They

The Journal of Heredity

were not raised from hand hybridized
seed, though this had been the original
intention. Sixty-two varieties were
grown as prospective parents, but cross-
ing proved impossible; no functional
pollen was formed. A few natural seed
berries were found, however, and from
them after years of testing these five
kinds proved to be the fittest. Even
the records of the maternal parents were
lost, but the goal set at the beginning
was reached. New potatoes better
than the old Early Rose and Peachblow
were produced. Considering the amount
of time and space at command, it was
probably the most successful practical
plant breeding experiment ever tried.

In all of the other hybridization
work, Mr. Carman made careful castra-
tions of the flowers used as female
parents, protected the blossoms from
foreign pollen and made the crosses by
hand. "Guess work in hybridization
or crossing," he says, "is altogether
abominable, because it is impossible
to know whether anything has been
effected or not, while the variations
sure to appear in the seedling plants,
it will be assumed, are evidences of
cross-bred parentage."

One of the most interesting pieces of
work brought to a successful conclusion,
was a cross between the beardless Arm-
stron ,r wheat and rye made in 1882.
Several varieties from this cross were
finally introduced, but whether they
battled successfully with pure wheats
or ryes, I have never heard.1 The im-
portant thing was the variation in a
first hybrid generation which was con-
clusively demonstrated — work which it
would be interesting to repeat even now
as the constancy or comparative homo-
zygosity of the parents was unknown —
and the pioneer work Of showing the
possibility of making crosses between
these two generically different cereals.
Mr. Carman saw the salient point very
clearly as the following quotation shows :
"What do they promise? If the hybrids
give us a grain less valuable than rye or

wheat, nothing will be gained in this
case, except the curious fact that a
cross between two different genera of
grain is possible. This established,
however, the way is opened for further
hybridization the pregnant results of
wh ch can only be guessed at."

Another interesting specific cross made
by Mr. Carman was between the black-
berry and the raspberry. It gave noth-
ing of commercial importance, though
by repeating it Luther Burbank is said
to have produced a valuable berry.
Neither Mr. Burbank nor Mr. Carman,
however, was the first to make this
cross; Mr. Carman, himself, admits
obtaining the idea from William Saun-
ders of London, Ontario, who had pro-
duced similar hybrids some five years
before.

WORK WITH SOLANUMS.

Mr. Carman's taste evidently was
partial to the Solanaceae. He worked
for many years on tomatoes, and suc-
ceeded in isolating from his various
crosses five types that were worthy of
introduction to the trade. They were
the Longkeeper, Lemon Blush, Terra
Cotta, Autocrat and Democrat. Auto-
crat and Lemon Blush were known for
years as the finest of their kind. He
also crossed the common tomato with
both the Currant Tomato L. pimpinel-
lifolium and the nearly related genus
Physalis. Whether any valuable types
were produced from the first cross or
not, I have been unable to find out,
but it was demonstrated that the first
hybrid generation was intermediate in
character and that a few of the indi-
viduals of the latter generations com-
bined a fairly large size of fruit with the
racemic type of inflorescence. The
generic cross was not sufficiently fertile
to be propagated, and died out after
a couple of generations.

Various other crosses of all kinds
kept up the interest of Mr. Carman in
his work, in which he was efficiently and
enthusiastically aided by Mrs. Carman,

1 W. Van Fleet, who was associated with Mr. Carman in his breeding work, states that none
of the real hybrid types survived continued propagation. Segregation occurred to such an extent
that the progeny soon became, to all appearances, either rye or wheat. None of the rye types
proved of particular value, but several of the wheat types are still in use. Farmers Bulletin No.
616 of the U. S. Department of Agriculture, "Winter Wheat Varieties for the Eastern States,"
recommends the soft "Rural New Yorker No. 57," one of Carman's creations. — The Editor.

East: E. S. Carman

although with one exception the rose
hybrids were the only ones that were
extremely valuable. This was the Car-
man Gooseberry. Here was a goose-
berry that might have revolutionized
gooseberry growing since in a limited
test it was mildew proof, but unfor-
tunately the seed firm to which it was
sold was unable to propagate it.

The roses were perhaps the real
attraction of the "Rural Grounds."
The Rosa rugosa of Japan was the
foundation stock, and upon it were
crossed first the Austrian hardy yellow
rose known as Harrison's Yellow, then
Hybrid Perpetuals and afterwards Hy-
brid Teas. From these crosses hundreds
of plants were raised — most of them, of
course, worthless, but some of remark-
able beauty. From the first cross men-
tioned came the Agnes Emily Carman, a
fine, hardy, longlived, though thorny
variety. In color it was like the
Jacqueminot, but many times as profuse
in blossoming. From other crosses
came procumbent roses, hedge roses, tea
roses, etc., etc. They did not attain
pre-eminence as did the potato varieties
but they helped and still help to brighten
many a flower garden.

Elbert S. Carman was born on
November 30, 1836, in Hempstead,
Long Island. He entered Brown Uni-
versity in 1854, rooming with John
Hay. He was obliged to withdraw after
two years of work, however, on account
of illness. In 1873, he married Agnes
E. Brown, by whom he had two chil-
dren. Immediately after his marriage
he moved to River Edge, N. J., where
he began to plant and experiment on
the place that afterward became so
well known as the "Rural Grounds."
While here he became so interested in
Moore's Rural New-Yorker as a con-
tributor, that he purchased the paper
and became its editor in 1876. Through
an absolutely open and honest policy,
he made this journal a power in the
agricultural world. For many years it
has stood out against all frauds and
impostures to the farmer, even though
this went against its monetary interests.
Air. Carman died February 28, 1900,
regretted by the many friends he had
made in his editorial capacity, who
wrote of him like the hero of Leigh
Hunt's ever popular poem, "as one
who loved his fellow men."

AN INTERPRETATION OF STERILITY IN CERTAIN

PLANTS.1

By E. M. EAST.
{Read April 23, 191 5.)

It is obvious that it is impossible to investigate the cause of
sterility in hybrids by the pedigree culture method when such
sterility is complete. Occasionally, however, one finds hybrids
which are not wholly sterile. Such is the case in the historic cross,
Nicotiana rustica L. X Nicotiana paniculata L. This hybrid holds
an enviable position in experimental botany, since it was the first
artificial hybrid to be studied. It was made by Kolreuter in 1760
and was studied by him for several years by means of back crosses
with each parent.

This cross I repeated in 1909, using as the N. rustica parent a
small variety N. rustica humilis Comes obtained from Dr. Comes
through the kindness of Dr. D. G. Fairchild. It has now been
studied through five generations both in the field (general morphol-
ogy) and in the laboratory (histology and cytology). The essen-
tial points noted, as I see them, are as follows :

Two species giving extremely uniform progeny when selfed
have, when crossed, given an intermediate F1 population as uniform
as themselves, and an inordinately variable F2 population.

The germination of F2 seeds varies in different samples from
20 to 60 per cent.

Practically no two F2 plants are alike, and the parental forms
are recovered once in every 100 to 200 F2 plants.

In Fj, from 1 to 6 per cent, of the § gametes are functional. It
is impossible to determine the percentage of viable $ gametes formed
from the pollen mother cells, but from 2 to 6 per cent, of the

1 It is impossible to reproduce the photographs shown by means of lan-
tern slides, but an illustrated paper giving the details of the investigation is
to be published shortly.

Rebrinted from Proceedings American Philosophical Society, Vol. liv , rgij.

EAST— STERILITY IN CERTAIN PLANTS. [April 23,

pollen found is morphologically perfect. The maturation difficulty
in spermatogenesis is largely at the first spermatocyte division.

Fx plants are as fertile inter se as in back crosses with either
parent.

Segregation of determiners for fertility occurs in Fx, so that by
recombination some perfectly fertile plants are obtained in F2.

Nearly all fertile F2 plants selfed give only fertile progeny. Oc-
casionally a fertile F2 plant selfed may give a slightly non-fertile
daughter.

Numerous combinations that should be possible in F2 are omitted
in the population obtained. Combinations approaching N. rustica
seem to be more frequent than those approaching N. paniculata.
Many more homozygous combinations occur in F2 than might be

expected.

Perfectly fertile plants giving perfectly fertile progeny, hetero-
zygous for many -allelomorphs, do occur in F2.

Xo more than a very general formal interpretation of these facts
can be made at present, but assuming that the chromosomes carry
the hereditary character determiners, and that these react with the
cytoplasm under proper environmental conditions to build up the
soma, attention is called to the following possibilities of satisfying
the conditions imposed by the data.

1. There is selective elimination of F2 zygotes.

2. There is no evidence of selective fertilization. (I infer this
from the fact that F1 plants are as fertile inter se as in backcrosses.)

3. The selective elimination of non- functional gametes that must
occur in Fx and the recombinations of functional gametes that give
different grades of fertility in F2 cannot be interpreted by a Mende-
lian factorial notation without subsidiary assumptions, but possibly
may be the result of one of the two following hypotheses :

(A) Through multipolar spindles, mating of non-homologous
chromosome pairs at synapsis, or other mitotic aberrations at the
reduction division, the 24 chromosomes characterizing each of the
two species may be irregularly distributed at gametogenesis. If
some of these irregular gametes may function, the majority of the
experimental data are satisfied, but there are reasons which there is
not time to consider which make this scheme improbable.

I9I5-]

STERILITY IN CERTAIN PLANTS.

(B) On the other hand the facts may be interpreted without
assuming irregularities of chromosome distribution if (i) there is
a group of chromosomes in each parent that cannot be replaced by
chromosomes from the other parent; if (2) there is a group of
chromosomes from each parent, a percentage of which may be re-
placed by chromosomes from the other parent, but where func-
tional perfection of the gametes varies as their constitution ap-
proaches that of the parental forms; if (3) there are other chromo-
somes that have no effect on fertility and therefore can promote
recombinations of characters in the progeny of fertile F2 plants ; if
(4) a naked male nucleus entering the normal cytoplasm of the egg
in the immediate cross can cause changes in the cytoplasm that will
affect future reduction divisions; if (5) this abnormally formed
cytoplasm is not equitably distributed in the dichotomies of gameto-
genesis in the F± generation; if (6) it follows from (4) and (5)
that F2 zygotes may be formed which are less perfect in their
gamete forming mechanism than those of the F1 generation; and
if (7) the heterotypic division of gametogenesis does not necessarily
form two cells alike in their viability.

Bussey Institution,

Harvard University.

THE CHROMOSOME VIEW OF HEREDITY AND
ITS MEANING TO PLANT BREEDERS

E. M. EAST

NEW YORK
1915

[Reprinted without change of paging, from the American Naturalist, 191 5.]

[Reprinted from The AMERICAN NATURALIST, Vol. XLIX, August, 1915.]

THE CHROMOSOME VIEW OF HEREDITY AND
ITS MEANING TO PLANT BREEDERS1

E. M. EAST
Bussey Institution, Harvard University

Definite advice as to practical procedure must be based
on a firm foundation of fact if the leaders in the applied
science are to retain any confidence in those who lay the
first stones in the pure science. At the same time, if it is
clearly understood that science only approximates truth,
that so-called "established laws' ' are only highly prob-
able and never absolute, it can hardly be said to be unwise
if an inventory of fact is taken at any time. The hand-
writing on the wall is never finished ; some words are dim
and the erasures and omissions are many, but that is no
reason why one should not try to read it and to see what
it directs if he has translated aright.

This preliminary justification of the title of this article
is made because our present stock of facts regarding
heredity points clearly to the chromosomes as vital parts
of the mechanism, and I wish to emphasize some impor-
tant practical deductions in case this position continues to
become more firmly established.

A just and complete dissertation upon the role of the

i This paper is based upon two lectures delivered at Harvard University
in 1914. I hope that any cytologists who may have their attention called
to it will overlook the repetition of some well-known facts in the first few
pages, as it is intended to be merely a general statement of a particular
point of view with certain deductions that follow if it be accepted. I wish
to thank Doctors O. E. White, T. H. Morgan and E. Goldschmidt for their
kindness in giving me many suggestions, but in justice to them I should state
that they are not responsible for the conclusions drawn.

457

458

THE AMEBIC AX XATUBALIST [Vol. XLIX

chromosomes in heredity not only would fill many pages,
but would expose numerous gaps in our present knowl-
edge, gaps that leave several important questions in the
balance. We shall assume frankly therefore that the
chromosomes are the bearers of the determiners of prac-
tically all of the hereditary characters that have been in-
vestigated by pedigree culture methods, acknowledging
freely our ignorance on many points, but maintaining that
while no facts have been discovered which offer insur-
mountable arguments against the viewpoint taken, the
following logical sequence of truths discovered at various
times and by different methods of research make a pretty
sound case upon which to base our practical conclusions.

Belative Importance of Nucleus axd Cytoplasm
There are several reasons for believing that of the two
parts of the cell, the nucleus and the cytoplasm, the former
plays the greater role in heredity.

In general it is believed that the two parents contribute
equally in the production of offspring — that the male and
female contribution of potential characters is practically
the same. If there were a difference it would be shown
by divergent results in reciprocal crosses, but the investi-
gations following Mendel's method make it probable that
with the exception of sex and sex-linked characters, the
results of reciprocal crosses are generally alike. This
being true, it would appear that the principal basis of
inheritance must be sought elsewhere than in the cyto-
plasm, for in most observed cases the sperm is very much
smaller than the egg, and this difference is largely a dif-
ference in the amount of cytoplasm each carries. Is one
not to look for some significance in this disparity in size?
Strasburger, as well as other botanists, has even gone so
far as to declare the male generative cell in certain angio-
sjDerms to be simply a naked nucleus that slips out of its
cytoplasmic coat into the embryo sac, leaving the dis-
carded coat behind, and that stimuli proceeding from the
nucleus control the assimilation of food in the cell and
determine even the character of the cytoplasm itself.

No. 5S4]

HEREDITY AXD ITS MEANING

459

This belief may be too radical. The machine must have
all of its parts to do proper work ; and it may be, as Conk-
lin suggests, that such characters as polarity, symmetry
and localization of organ bases in the egg have their chief
seat in the cytoplasm. This is only a possibility and not
a fact, however, for one must admit that etiological inves-
tigation has not disclosed the presence of a material basis
of heredity in the cytoplasm, though he may not be con-
vinced that it is unimportant. Does the same statement
hold for the nucleus ?

The nuclear cavity contains four substances as they are
ordinarily described in connection with morphological in-
vestigations. These are nuclear sap, linin, nucleolar ma-
terial and chromatin.

Xuclear sap probably belongs as much to the cytoplasm
as to the nucleus, and we know nothing as to its possible
significance and importance within the nucleus.

Linin by some investigators is regarded as very similar
to chromatin. Others (Strasburger) consider it to be
the framework of the chromosomes, and the only real sub-
stance within the nuclear cavity that is continuous from
generation to generation. It is a thread-like material
staining lighter than chromatin upon which the chromo-
somes appear to be strung in the early prophases of nu-
clear division.

Nucleolar substance, though it stains in a different
manner from chromatin, is considered by many to be
chromatin-like in its nature. It is the substance of which
the nucleoli are composed; but as these bodies become
vacuolated and finally disappear during nuclear division,
one is led to believe with Strasburger that they are tem-
porary storehouses of some necessary food material.

Chromatin, however, as the material of which the
chromosomes are composed, plays such a peculiar part in
the activities of the cell, that hypotheses as to the mean-
ing of its behavior are certainly more than shrewd guesses,
as will be seen.

The chromosomes may be described as morphological

460

THE AMEBIC AX NATURALIST [Vol. XLIX

elements, of various shapes and sizes that are found
within the nucleus; they are especially demonstrable as
deeply staining bodies, definite in number for each cell at
the period of division. In many cases in both plants and
animals they have been found to be made up of small
particles, the chromomeres, and various investigators
have expressed the belief that these, too. are definite in
number and play an important part in the larger collective
entity, the chromosome.

Almost from their discovery, the chromosomes have
had an especially important part assigned to them in the
drama of heredity because of the previous philosophical
deductions of Weismann. Weismann reasoned that if
there were no reduction of heritable substance in the life
cycle of an organism, it would pile up indefinitely because
of the nuclear fusion at fertilization. He, therefore, pre-
dicted the discovery of some mechanism by which the
character conserving substance would be divided. A few
years later his prediction was verified in its important
details by actual observation of the chromosome reduc-
tion in the formation of germ cells in Asearis. From this
discovery and from the facts that a specific number was
found for the cells of each species, that all the cells of an
individual appeared to possess the same number (except
when they were halved at gainetogenesis), that they were
apparently permanent organs, that they were longitudi-
nally halved in division so as to give each daughter cell
the same number as well as an exact half of each chromo-
some possessed by the mother cell, investigators were
early tempted to place upon chromosomes the whole
burden of inheritance.

Our observations regarding chromosomes and the re-
duction divisions in plants now rest on a basis of cy to-
logical investigation of over 250 species, representing
over 150 genera and divided among the four great groups
of this kingdom. Montgomery's 1906 list of chromosome
numbers in animals represents investigations on 185 spe-
cies, comprised in about 170 genera, distributed among

No. 584]

HEREDITY AXD ITS MEANING

461

nearly all the phyla of the animal kingdom. Sex chro-
mosome studies have undoubtedly increased these figures
for the animal kingdom to date, by hundreds of species.

Variation in chromosome number among the cells of an
individual plant or animal is a recognized fact among
cytologists, but this variation is not regarded as of par-
ticular significance, as commonly it is held to exist only
among old cells, cells highly specialized, or, at any rate,
cells which will never have anything in common with re-
production. To quote from Strasburger,

the number of chromosomes in the nuclei of the somatic cells of both
the sexual and the asexual generations have been found to vary. But so
far as my experience goes, these observations are always to be observed
in the nuclei of cells which are no longer embryonic, like those in an
embryo or growing point, but which, on the contrary, are to some ex-
tent histologically specialized and are not destined eventually to give
rise to reproductive cells. The determinate number is still more fre-
quently departed from in nuclei which are definitely excluded from the
sphere of reproduction.

In the reproductive cells, chromosome division is, on
the other hand, very exact, and the numbers found, almost
invariable, with one exception. This exception is the so-
called accessory chromosome or chromosomes, that ap-
pear to be coupled with sex differentiation. And the
very fact that such accessory chromosomes do exist and
by their presence or absence parallel sex distribution,
forms one of the most unanswerable arguments in favor
of the chromosomes being the chief bearers of character
determinants.

Morphological Individuality of the Chromosomes
The next topic to consider is whether there is sufficient
evidence to support the idea that these bodies— the chro-
mosomes—are morphological entities persisting from one
cell generation to another.

Prochromosomes are deeply staining bodies found in
the resting cell nuclei of plants, which probably corre-
spond in number, but not in size, to the chromosomes
which are found in the dividing nuclei. These bodies are

462

THE AMEBIC AX X ATI' R ALT ST [Vol. XLIX

thought to represent the resting nuclear condition of the
chromosomes. Prochromosomes have been found in at
least sixty species of plants, and various structures com-
parable to them in many others. These investigations
favor the thought that the chromosomes are persistent
morphological entities : nevertheless they are not suffi-
cient to establish the matter if there were no other data
at hand.

There is a series of facts, however, which is more con-
vincing. We are told that in addition to each species of
animal or plant having in the larger part of its cells a spe-
cific number of chromosomes, there is a constant reap-
pearance of the different shapes and sizes of these chro-
mosomes in the same positions relative to one another
during cell division after cell division.

Strasburger says: "The observation of such a series
of stages of nuclear division as can be obtained by the
laying open of embryo sacs in which development of
endosperm tissue is commencing, makes it difficult to re-
sist the impression that it is always the same chromo-
somes which make their appearance over and over again
in the repeated divisions. In the prophase, the chromo-
somes are seen to appear in precisely the same position
that they occupied in the preceding anaphase, and if the
picture of the anaphase were proportionally enlarged, it
would exactly correspond to that of the succeeding pro-
phase/'

The facts from which these general conclusions have
been drawn can not be denied. Baltzer found odd-shaped
chromosomes of similar shape in many maturing eggs of
sea urchins. Boveri, Montgomery and later SchafTner
pointed out a constant difference in the form and the size
relations of the two chromosomes of Ascaris megalo-
cepliala univalens. Sutton thought he could recognize
each individual chromosome in eleven consecutive cell
generations of the maturing germ cells of the lubber
grasshopper Brachystola magna. The so-called sex chro-
mosome which has been found in so many insects and

No. 584]

HEREDITY AND ITS MEANING

463

other animals, is a clear case of constancy in appearance.
In plants the same phenomenon has been observed. Ro-
senberg* investigated the pollen mother cells of Crepis
virens and in certain stages in division invariably found
two long, two intermediate and two very short chromo-
somes. Division figures in the somatic cells showed the
same differentiation, and in an examination of the nuclei
of the pollen grain he found only one chromosome of each
kind present. Such other species of this genus as have
been investigated also show some variation in chromo-
some form, although it is not so striking as in C. virens.
Hieracium venosum, exceptionally good material also in-
vestigated by Rosenberg, has shown the same thing.
Edith Hyde remarks on the fact of the constant reappear-
ance of certain chromosome forms among hundreds of
division figures which she observed in Hyacinthus orien-
talis. Sauer mentions a very long chromosome constantly
present in pollen mother cell preparations of the lily-of-
t he-valley, and Strasburger and Lutz found a large
chromosome among many small ones in Lychnis dioica.
In certain species of Yucca this chromosome differentia-
tion takes on a dimorphic aspect, ten of the chromosomes
being very large and about forty-five very small.

Taking into consideration all of these facts, of which
hardly more than a random sample has been given, one
is clearly justified in concluding that these cell characters
are reproduced generation after generation. Why this
constancy if they are not important !

Physiological, Individuality of the Chromosomes
There is also considerable reason for believing that the
various chromosomes of a cell may have different func-
tions.

Boveri was the first to endeavor to test this hypothesis
by allowing sea-urchin's eggs to be fertilized by two sper-
matozoa. Three nuclei, each with eighteen chromosomes,
were thus present in the same egg, two male and one
female. Although cytoplasmic division seemed to pro-

464

THE AMERICAN NATURALIST

[Vol. XLIX

ceed normally, the chromosomes were usually distributed
irregularly by a three-poled or a four-poled spindle. As
a result three or four cells were produced at the first divi-
sion of the doubly fertilized egg, instead of the two cells
that arise after normal fertilization. Various abnormal
larvae were produced later. In such embryos, Boveri found
the organism to be divided into definite regions, thirds or
fourths, each part traceable to one of the three or four
original cells, and the cells of each part differing from
the cells of the other parts in their combination of chro-
mosomes and usually in their chromosome number. In
rare cases normal embryos were produced, but these were
more commonly developed from a doubly fertilized egg
which in its first division was three-celled, than from one
in which it was four-celled. The thought occurs at once
that three cells have a better chance than four cells in
securing a full set of chromosomes, both as to number and
kind. If the division were normal, each nucleus would
receive a full set in the case of the chromosome distribu-
tion to three cells, but the division is usually irregular,
and because of this irregularity each cell does not usually
secure its normal set of chromosomes. Nevertheless it
is clear that the embryo parts developed from the three-
celled cleavage stand a much greater chance of being
normal than those from the four-celled type, although
through irregularities in division an eighteen-chromo-
some-celled region might be formed even where the first
division was four-celled.

In some cases, the embryo was completely normal as
regards skeleton and pigmentation in one or even two of
its thirds, while the remainder was entirely lacking in
these characters. Nearly normal embryos occurred which
were perfect as to parts and specific characters, but indi-
vidual variations which normally should have appeared
in separate larvae were present among the thirds. Asym-
metrical larvae also were formed.

More important still are the results Boveri obtained by
isolating the three cells of the three-fold type and the

No. 5S4]

HEREDITY AND ITS MEANING

465

four cells of the four-fold type and allowing them to de-
velop into larvae. When the four cells of a four-celled
stage of a normal embryo are separated, each cell pro-
duces a normal dwarf embryo alike in every respect, but
the three- or four-celled embryos from double fertilized
eggs, when treated in the same manner, never produce
normal dwarfs even when the chromosome distribution
has been numerically equal. Large numbers of larvae
brought into existence through this experiment showed all
possible combinations of characters, just as all possible
chromosome combinations were found in their nuclei,
and from these and other data the conclusion is drawn
that "not a certain number, but a certain combination of
chromosomes is necessary to normal development, and
this clearly points out that chromosomes have different
qualities." In other words, the sea urchin has a set of
eighteen chromosomes, each chromosome performing at
least some different functions from its neighbors, making
it necessary for the whole set to be present in order to
insure normal development.

In further investigations, Boveri placed sea-urchin eggs
which had been normally fertilized and were about to di-
vide under pressure. As a result, division of the nucleus
took place, but often no division of the cytoplasm. Such
eggs on again dividing often formed more than two poles,
resulting in inequalities in chromosome distribution and
abnormal larval development. Boveri puts upon these
cases an interpretation similar to that of the preceding
experiments, as the irregular chromosome distribution
seems to be all they have in common.

Morgan comments on Boveri 's experiments as follows:

The evidence makes probable the view that the different chromosomes
may have somewhat different functions and that normal development
depends on the normal interactions of the materials produced by the
entire constellation of chromosomes.

Artificial parthenogenesis and experiments with enu-
cleated eggs have proved that only one set of chromosomes
is necessary to normal development of embryos, but it is

466

THE AMERICAN NATURALIST

[Vol. XLIX

important, in considering these experiments, to note that
two sets of similar chromosomes are present in a normal
sexually produced organism.

Pairs of chromosomes of each shape and size (if they
differ in shape and size) are nearly always found in the
somatic cells — the exception being when the so-called
accessory chromosomes are present. And since but one
of each kind is found in the two gametes that fuse to form
the new organism, it is only natural to suppose that one
set was contributed by the maternal parent and the other
by the paternal parent.

The numerous cases in which this phenomenon has been
demonstrated are to many the most convincing evidence
of some sort of a morphological individuality of the chro-
mosomes. To them the fact implies pairs of freight boats
loaded with the essential materials of life, to others— the
minority— it is no more wonderful than the constant re-
currence of other plant organs. At any rate, it has been
shown that these sets of chromosomes continue an appar-
ently independent existence for some time. Moenkhaus
produced hybrids between the two species of fish, Fundu-
lus heteroclitus with long straight chromosomes and
Menidia noted a with short curved chromosomes, and the
early divisions of the fertilized egg showed clearly com-
plete sets of chromosomes from each parent. Eosenberg
obtained similar results in crosses between the two sun-
dews, Drosera longifolia, which has forty small chromo-
somes, and Drosera rotnndifoUa, which has twenty large
chromosomes. In some cases similar to the latter, where
one parent contributes a greater number of chromosomes,
it should be noted that the organism seems to have regula-
tory powers. The chromosomes unnecessary for a double
set are either thrown out or take no part in the activities
of cell division. For example, in the supposedly hybrid
sundew, Drosera obovata, Eosenberg found that its thirty
chromosomes behaved in the following peculiar manner.
Ten of them paired with another ten, but the other ten
remained unpaired and acted in a very abnormal fashion

No. 584]

HEREDITY AND ITS MEANING

467

in the reduction divisions. The ten pairs separated nor-
mally, one of each pair going to each pole; but the ten
unpaired were irregularly distributed, sometimes nearly
all of them going to one pole, sometimes most of them be-
coming lost in the cytoplasm and forming small nuclei.
Embryos were produced in a very few cases and these
only through back-crossing with pollen of D. longifolia.
Unfortunately these embryos only developed through a
few cell divisions.

These chromosome pairs have been distinguished by
the name homologous chromosomes. For a long time it
was thought that the paternal and the maternal set of
chromosomes separated from each other bodily at the re-
duction division. Now it is believed to be only a matter
of chance which chromosome of a pair passes to a particu-
lar daughter cell. There is some cytological evidence for
this view, but the main argument in its favor is that this
behavior is all that is necessary to fit nearly all the known
facts of heredity, with the chromosomes playing the
part of the active heredity machinery as will be seen
shortly. This statement is true in a broad sense, but the
word nearly is used because there is an exception to it.
Chance apportionment of either member of a homologous
pair of chromosomes to a daughter cell accounts for all
facts of alternative (Mendelian) inheritance except where
there are breaks in the correlation between characters
usually inherited together. Since such breaks in corre-
lation are common, it is clear that there must be a period
when chromosome pairs have such an intimate relation
that material can be exchanged. Many biologists believe
that such a period is found during the maturation of the
sex cells. The particular point at which such a conjuga-
tion or approximation of chromosome pairs takes place is
called synapsis; it occurs as a part of the prophase or
first stage of the reduction division. Some investigators
have been unable to demonstrate any real chromosome
fusion at this time, but all agree that there is an approxi-
mation between the two sets, and a chance for some kind
of an exchange or interaction to take place.

468

THE AMERICAN NATURALIST

[Vol. XLIX

Evidence of the physiological individuality of the chro-
mosomes may be concluded by referring briefly to the so-
called accessory chromosome. This fraction of a chro-
mosome, whole chromosome, or in some cases, group of
chromosomes, possesses no true synaptic mate, and there-
fore at reduction division two types of daughter cells are
found. The presence or absence of the "accessory" is so
closely associated with sex determination that most biolo-
gists now regard it as the morphological expression of a
germinal sex determinant. The essential result of re-
searches on this body may be summed up in the following
words of Wilson.

They have established the existence of a visible difference between the
sexes in respect to these chromosomes, and have shown that it is trace-
able to a corresponding difference in the nuclei of the gametes of one
sex or the other.

The simplest type of accessory chromosome, where the
male possesses an unpaired chromosome which passes
to one pole undivided in one of the spermatocyte divisions
and hence enters but half the spermatozoa, was discovered
by Henking (1891) in Pyrrhocoris. This work was con-
firmed in certain species of Orthoptera in 1902 by Mc-
Clung, who advanced the hypothesis that the odd chromo-
some was a sex-determiner. Shortly afterward this was
made more probable by Wilson and by Stevens who
proved for several species of Hemiptera that the body
cells of the males contain one less chromosome than the
females. Two accessory or X chromosomes are present
in the female, while but one is present in the male.

About the same time, both Wilson and Stevens inde-
pendently discovered another kind of dimorphism in male
germ cells of certain Hemiptera. Here the X chromo-
some of the male has a smaller synaptic mate Y. The
body cells of the female, however, show two of the large
X chromosomes. The sexes, therefore, both contain the
same number of chromosomes, but have the same type of
chromatin difference as was first discovered. The female
is XX and the male XY.

No. 584]

HEREDITY AND ITS MEANING

469

Baltzer claimed in 1909 that in the sea urchins Sphccre-
chinus and Echinus the sex with the dimorphic germ cells
is the female instead of the male, but the work of Tennent
has shown him to be in error and he has retracted the
statement. There is, therefore, no undisputed cytological
evidence demonstrating this type of dimorphic eggs ; but
since breeding results on certain species of birds and of
lepidopters can be interpreted only on such an assump-
tion, it is safe to assume that sooner or later they will be
found.2 Whether or not there are animals of this type,
however, is of no particular importance in the present
discussion. What we desire to emphasize is that a large
number of animals, including man, have been shown to have
a chromatic difference between the sexes, and that this
difference is readily explained by the fact that the eggs
are of a single type and the spermatozoa of two types.

In dioecious plants no such morphological differentia-
tion has been found. But this fact does not negate the
idea that the visible differences found in animals are really
sex-determining differences. We have only to suppose
that the dimorphism is primarily qualitative and second-
arily quantitative. Indeed Wilson has found that the Y
chromosome— the synaptic mate of the X— may vary in
different species from a size equal to that of X until it
disappears entirely, leaving X without a mate.

There is only one criticism in this whole matter. One
may admit these cytological differences between the sexes,
but hold that they are early appearances of secondary sex-
ual characters. Morgan, von Baehr and Stevens have
answered this impeachment. In the phylloxerans and
aphids all the fertilized eggs produce females ; males arise
only by parthenogenesis, though females may arise in this
manner. The cytological facts are as follows: Under
favorable external conditions eggs develop without reduc-
tion and females are formed. Under unfavorable condi-
tions one or two chromosomes (the sex determiners) are
thrown out. If these eggs develop without fertilization

2 Dimorphic eggs in Lepidoptera have recently been demonstrated by
both Doncaster and Seiler.

470

THE AMEBIC AX NATURALIST [Vol. XLIX

males arise. The somatic condition of the females may
therefore be termed XX and that of the males XY. If
both reduced normally at any time, ordinary fertilization
might be expected to give both males and females. But
the spermatocytes without X degenerate, leaving only one
type of functional spermatozoa, which produces females.
Thus actual causal connection between the X chromosome
and sex determination appears to have been demonstrated.

These are the main cytological arguments in favor of
the chromosome view of heredity that seem to me to be
insuperable. There are minor arguments both pro and
con, which, as I said in the beginning, we have not space
to consider. Instead it seems more profitable to show
how Mendelian results interlock with those from cytology
like the parts of a jig-saw puzzle.

Chromosomes axd Mexdellajst Ixheeitaxce
The principal phenomena of Mendelian inheritance are :
(1) characters that breed true; (2) uniformity of the
population of the first hybrid generation in particular
traits in which homozygous parents differed; (3) inde-
pendent segregation of certain character determiners;
(4) recombination of certain characters; (5) perfect
coupling between certain characters; and (6) partial
coupling between certain characters. Let us see how
plausibly one can picture the mechanism through which
such phenomena may result without imputing to the
chromosomes any behavior that is not known to occur.
To do this simply let the imagination portray a plant spe-
cies having four chromosomes, each chromosome having
three character determinants that can be followed through
the breeding results that are obtained.

Our figures represent the immature germ cells of the
plant just previous to the reduction division. Fig. 1
shows the germ mother cell with a duplicate set of heredi-
tary determinants. The mature germ cells are exactly
alike, therefore the plant breeds true to the characters
concerned.

No. 584] HEREDITY AND ITS MEANING 471

Fig. 1

Fig. 2

Suppose, however, that a change in the germ plasm has
occurred (Fig. 2) at some time or other. In one member
of the first pair of chromosomes, determinant "A" has
become "a." The mature germ cells differ from each
other by one factor. For this reason the plant does not
breed true, but gives a mono-hybrid Mendelian result.

Fig. 3

472

TEE AMEBIC AX X ATE BALI ST [Vol. XLIX

Again, if such a change occurs that A becomes A' (Fig.
3), a series of triple allelomorphs giving monohybrid re-
sults with each other, is formed. "A" is allelomorphic
to "A'" or "a."

Fig. 4

But there are other character determinants in the first
pair of chromosomes. What happens if both "A" and
"B" become changed? There are two possibilities, as
shown in the two parts of Fig. 4. If one of the members
of the pair of homologous chromosomes becomes abC
while the other remains ABC. there is a positive corre-
lation between the inheritance of "A" and "B." On the
other hand, if the change is such that the two chromo-
somes are aBC and AbC, there is a negative correlation
between A and B. In other words, the determinants re-
main correlated in the same way they entered the com-
bination. There may be breaks in these correlations,
however, as Morgan has shown in DrosophUa; and these
breaks in correlation occur in a constant ratio. Diagram-
matically, it may be said that A and B are always the same
distance apart in the chromosome structure and that the
determinants " cross over" from one member of a pair
to the other every so often. All of the gametes in the
first case are not ABC and abC, for example. Some of
them will be AbC and aBC. And the same percentages
of these cross overs are found in the second case where
"A" and "B" are correlated negatively. Furthermore,

No. 584] HEREDITY AND ITS MEANING

473

if C should become c, and the chromosome pair take the
form ABC and abc, there are definite relations between the
three determinants. Breaks in correlation occur, and this
ratio is constant, so that if given the percentage of breaks
of correlation between "A" and "C" and "B" and "C,"
the percentage of breaks between "A" and "B" can be
predicted. If there is a break in the correlation between
"A" and "C" 30 times in 100, and a break between "B"
and 10 times in 100, then there will be breaks in the
correlation between "A" and "B" 20 times in 100.

Fig. 5

Likewise, the determinants in the second pair of chro-
mosomes are coupled together in their inheritance. D, E
and F have each their peculiar linkage to the other, a link-
age that remains comparatively constant. Yet the de-
terminants in the second pair of chromosomes are entirely
independent from those in the first pair in their inheri-
tance. For example, if, as shown in Fig. 5, "A" should
become "a" in either member of pair number one, and
"D" should become "d" in either member of pair number
two, Mendelian dihybridism would result. Furthermore,
if "A" and "D" should each have the function of affect-
ing the same general character complex in somewhat the
same manner, there would be an apparent 15 : 1 ratio if
dominance were complete or a series of types ranging
from the type of one grandparent to that of the other, if
dominance is lacking.

These are the main features that have been established

474

THE AMERICAN XATUEALIST

[Vol. XLIX

by recent work on hybrids. AYe have pictured them as
actual chromosome functions, because every part of the
description has been actual fact as far as the breeding
experiments go. Our picture, it is true, is fictitious, for
we do not know absolutely that the heredity mechanism is
of this nature. But the facts do fit perfectly all that is
known of chromosome behavior. It seems impossible,
therefore, that there should be so many coincidences.

There are also two other pieces of evidence that fit in
and round out the case. Bridges has shown that females
occasionally occur in Drosopliila bearing the sex-linked
characters borne by the mother but showing no influence
of those borne in the father. Such exceptional females
were found to inherit directly from their mother the power
of producing like exceptions, and it was proven cytolog-
ically after the prediction had been made from the breed-
ing facts that these females resulted from the non-disjunc-
tion of the X chromosomes at the maturation of the eggs
from which they came, and that one half of their daughters
did in fact contain a Y chromosome in addition to two X
chromosomes. This appears to be definite proof that sex-
linked genes are borne by the X chromosomes.

The other important basis for regarding the chromo-
somes as the material basis for heredity also comes from
Morgan's work on Drosopliila ampelophila, this being the
only species upon which sufficient work has been done to
give a reasonable basis for the conclusion. All of the hun-
dred and thirty or so mutations in this species upon which
Morgan and his students have worked are so linked to-
gether in heredity that they form four groups correspond-
ing to the four pairs of chromosomes found in the species.
If one single character should be found that did not fit into
one of these four groups, the whole theory tcould break
down. But no such character has appeared.

This completes the case for the chromosomes as regards
the main facts, and it seems only proper that a fair-
minded jury of scientists should render verdict for the
plaintiff. Xo case is so bad, however, that a lawyer can

No. 584] HEREDITY AND ITS MEANING 475

find nothing to say for the defense and scientists in this
respect resemble the men of the bar. Certainly there are
some outlying facts, but they are comparatively unimpor-
tant. If a series of important facts should at any time
be found which do not fit, the chromosome mechanism
should be looked into. It is likely that the explanation
will be found in an abnormal chromosome behavior as was
the case in the aphis.

Practical Conclusions and Discussions
If now it be accepted as a reasonable premise that the
chromosomes are the chief if not the sole bearers of he-
reditary determinants of body characters, and that their
behavior is a rough indication of the mechanism of he-
redity ; what cytological facts, if any, can be made useful
at present or in the future to plant and animal breeders?
If such data exist, they should be put to service ; if it is
likely that such facts can be found, investigations should
be undertaken. The broad question may be divided into
three parts which will be discussed in regular sequence :

1. What are the relations of chromosomes to somatic
characters ?

2. "What are the relations of normal chromosome beha-
vior to the transmission of characters 1

3. What are the relations of peculiar or unusual chro-
mosome behavior to the transmission of characters?

Eelations of Chromosomes to Internal Characters
Some very interesting observations have been made on
the relations of internal and external characters to chro-
mosome number.

Farmer and Digby in a comparative study of the cells
of a fern of the genus Atliyrium with similar cells of three
of its varieties, found that the measurements were suc-
cessively larger in the three varieties than in the species,
and that there was a corresponding increase in the number
of chromosomes, the gametic numbers for the species and
its varieties being estimated at 76-80, 84, 90 and 100,

476

THE AMEBIC AX XATURALIST

[Vol. XLIX

respectively. Investigations on another fern, Lastrea, did
not corroborate these results^ however, in one variety the
chromosomes being more numerous and the cells smaller
than in the parent type.

Gates by comparing nuclei and cells of different tissues
of CEnotliera Lamarckiana and similar structures in its
"rnutant" 0. gigas with double the number of chromo-
somes, found that the 0. gigas cells and nuclei were always
larger, varying from a comparative ratio of 1 : 1.5 to 1 : 3.
At the same time, it would hardly be wise to maintain that
this is always the case, for only a few individuals were
investigated.

Primula sinensis has two forms in cultivation, similar
except as to size. The giant form has flowers about one
and one half times the size of those produced by the ordi-
nary form. Gregory investigated these two forms cyto-
logically to determine the cause of this increase. The
nuclei and the chromosomes of the giant form were a little
larger, though the difference was hardly a measurable one.
The chromosome number was the same in both the forms.
In a later investigation he has found that some exceed-
ingly large plants with nuclei distinctly larger than those
of the normal form had double the number of chromo-
somes normal to the species.

Boveri investigated this same relation of cells and nu-
clei to chromosome number in X., 2N and 4X larvae of the
sea urchin. From these studies, he concludes that chro-
matin is non-regulatory, and in the case of decrease, un-
regenerable, the cytoplasm in contrast showing the fullest
regulatory activity. Further, the size of the larval cells
is governed by the chromosome mass and the cell volume
is directly proportional to the chromosome number. On
the other hand, Conklin's investigations on annelids, mol-
lusks and ascidians lead him to take a position opposed
to that of Boveri. He says :

The size of the nucleus, centrosomes and chromosomes is dependent
upon the volume of the cytoplasm is clearly shown in C re pi did a, where
in large and small blastomeres, these structures are invariably propor-
tional in size to the volume of cytoplasm.

No. 584]

HEREDITY AXD ITS MEANING

411

Neither chromosomes nor nucleus control, the size of the
cell in annelids, mollusks or ascidians.

Relations between Chromosomes and External
Characters

Thus there seems to be no constant relationship even
between nuclear or cell size and number of chromosomes,
and bonds of union between external taxonomic charac-
ters and chromosome number seem to be still more tenu-
ous. It is true that certain giant Primulas and (Enotheras
had more chromosomes than were characteristic of the
normal forms, but it is 3'ust as clear that all giant Primulas
(and the same is probably true of (Enotheras, from the
work of Heribert-Xilsson and of Geerts) do not have ab-
normal chromosome numbers.

Results on several species of both animals and plants
are interesting in this connection.

The thread worm, Ascaris megalocephala, has two va-
rieties, bivalens and univalens, the former having as a 2N
number four chromosomes, the latter two chromosomes.
Xothing is known as to the origin of these two forms.
They are found parasitic in the same host individual and
neither form is rare. According to Herla, they hybridize
freely and produce embryos whose cells have three chro-
mosomes, but no mature hybrids have ever been found.
Meyer could distinguish no anatomical differences be-
tween the two varieties.

Rosenberg investigated the reproductive structures of
two species of sundew and found one to have double the
chromosome number of the other. A subsequent com-
parison of anatomical and taxonomic characters failed to
show any sharply marked differences between them ex-
cept in size. The form having the smaller chromosome
number was smaller and less robust. They inhabit the
same territory and produce natural hybrids which are
sterile.

Rosa canina has two varieties which have the same taxo-
nomic characters, but one form has thirty-four while the

478

THE AMEBIC AN NATURALIST [Vol. XLIX

other has only sixteen chromosomes. The form with
thirty-four chromosomes is apogamous and reproduces
without fertilization, but that one must not conclude that
apogamy is necessarily associated with a double or an in-
creased chromosome number, is clear from the case of
Rumex. Rumex was investigated by Eoth; one species,
R. cordifolius, having forty chromosomes as its 2N num-
ber, required fertilization to produce offspring; another
species, with only sixteen chromosomes, was apogamous.

A short list of nearly related species or species with two
varieties varying in their chromosome numbers with their
character differences, if present, is given below.

Name

Date

Characte rs

Investigator

All '11 T7V ^7^7 „„„ '11 —

Alchemilla Eualcnemilla . . .

1904

Apogamous

Strasburger, E.

aphanes

1904

Ascaris megalocephala

1883

Alike externally

Van Beneden

1895

Meyer, O.

" " and others

1887

Boveri, T.

1887

1911

Ishikawa, M.

1911

1909

Rosenberg, O.

1909

More robust, etc.

Echinus microtuberculatus .

1888

Boveri, T.

1902

1903

Alike externally

Ancel, P.

1896

v. Rath, 0.

1908

128

None mentioned

Yamanouchi, S.

1908

132

CEnothera lamarckiana

1911

Gates, R. R.

1909

Large and

coarser

Primula sinensis

1909

Gregory, R. P.

" giant form

1909

More robust

1914

1909

Apogamous

Rosenberg, 0.

1904

Strasburger, E.

1909

Overton, J. B.

" purpurascens. . .

1909

Apogamous

Zea Mays, "White Flint".

1911

Kuwada, Y. "

" , "Sugar"

1911

What conclusions can be drawn from these facts % Cer-
tain botanists have attempted to connect chromosome
doubling with apogamy, as usually the chromosome num-
ber in apogamous species is higher than in the normal
species of the same genus; but there is no evidence of

No. 584]

HEREDITY AND ITS MEANING

479

apogamy in Oenothera gigas, and in Ramex the form with
the low number of chromosomes is apogamous while the
form with the high chromosome number requires fertili-
zation. On account of these exceptions, therefore, it
seems probable that the cause of apogamy is deeper than
a mere doubling of the chromosomes, even though doub-
ling may usually accompany such a change in reproduc-
tive habits.

Variation in chromosome number in the same species
has been proposed as a cause of general variation in so-
matic characters, but the evidence is not clearly in favor
of such a theory. In the fern Nephrodiam molle Yama-
nouchi found spermatid cells to be of two sorts, those with
sixty-six and those with sixty-four chromosomes. This
would mean that Nephrodium has two gametophyte forms
and two sporophyte forms, externally identical, so far as
our present knowledge goes, but differing in their chro-
mosome numbers.

Further, sporophytes developing from the prothallia of
ferns without the intervention of a sexual process have
the X instead of the 2X chromosome number, yet apoga-
mously developed fern sporophytes, except as to chromo-
some number, are indistinguishable from normal sexually
produced individuals of the same species.

Many writers have been tempted to postulate a causal
relation between the numerical variation of chromosomes
among the species of a genus and the genera of a family
and their specific and generic characters. The thirty or
more species of Compositae investigated have shown a
remarkable variation in their chromosome numbers, the
2N numbers ranging between six and sixty, and, as is well
known, the Compositae possess an infinite variety of
sharply contrasting characters. But the lily family also
has an enormous number of characters in its species and
genera, and the genus Lilium, with its great variety of
characters distributed among forty-five species, is typical
of the other genera of the family, as far as present inves-
tigations go, in having the same chromosome number for

480

THE AMEBIC AX XATURALIST

[Vol. XLIX

all of its species. Others suggest that the more chromo-
somes a plant species possesses the greater is its varia-
bility. Thus Spillman3 speaks of the low variability of
rye, suggesting its small chromosome number (six or
eight) as a possible reason; for maize, having probably
from twenty to twenty-four chromosomes, is infinitely
more variable than rye. However, Britton's "Manual"
selects Crepis virens for special mention as an extremely
variable species from among the four or five other species
listed under that genus, and it is known that C. virens
has only six chromosomes, while three other species of
Crepis investigated all have higher numbers. Again, ac-
cording to Wiegand, the Comma has only six chromosomes,
yet every gardener is well acquainted with the infinite
variety in Cannas.

The Chromosomes axd Variability
After a consideration of the above facts, one may well
hesitate to state that there is even a high degree of corre-
lation either between variability in chromosome number
and general variability, or between high numbers of chro-
mosomes and a high degree of variability in specific char-
acters. On the other hand, it is not certain that the data
upon which our discussion is based are relevant to the case
in hand. We have discussed a possible relationship be-
tween chromosome numbers and species complexity and
variability as found in the wild. This is not at all the
same thing as discussing the relationship between chro-
mosome number and true variability. It is true that com-
plexity and specialization of plants and animals seem to
have no connection with chromosome number, and that
within a family a genus or a species profusion of taxo-
nomic characters do not go hand-in-hand with high chro-
mosome numbers. But in these cases our data come from
persistent forms. What the actual inherent variability
of the protoplasm is in most cases we do not know. Dro-
sophila ampelophUa, a species with only four chromo-

s Six according to Westgate 's unpublished data ; eight according to Xakao.

No. 584]

HEREDITY AND ITS MEANING

481

some pairs, is considered to be very constant in its char-
acters from the taxonomist's standpoint, yet by careful
continued observation Morgan has succeeded in detecting
over 130 mutations.

From a strictly mathematical standpoint, it would seem
that if other things are equal, variability would take place
in proportion to the number of chromosome units. The
difficulty is that in no case do we know anything whatever
about the relative complexity of any particular chromo-
some unit. One must infer, however, that the 47-48 chro-
mosomes in man are individually much more complex than
the 128-132 chromosomes in the fern Neplirodium molle.
If this inference be correct there are reasons why altera-
tion in determinants may occur in direct proportion to
the number of chromosomes or rather to the mass of chro-
matin without there being visible somatic variability in
the same ratio. Let us construct an imaginary plan for
preventing visible variation without preventing change
in chromosome determinants. Unquestionably the sim-
plest means is to double the chromosome number. Se-
lecting, for example, a species with four chromosomes, let
us suppose that fertilization occurs without a reduction
division at some time or other. Then instead of a dual
organism with two sets of chromosomes of similar func-
tion, one from the male and one from the female parent,
there would be a quadruple organism with twp sets of
similar chromosomes from each parent. Any germinal
change which would produce a new dominant character
would be apparent immediately, but for a recessive change
to appear— and these are many times as numerous as the
others— it would be necessary to have identical changes
occur in two chromosomes. Following out this line of
reasoning, it is not hard to see what a great possibility
for uniformity there is in further chromosome duplication,
provided the actual fact of duplication makes no great
change in the organism. That chromosome doubling has
no decided visible effect we have seen from the cases
already described ; and since so many nearly related spe-

482

THE AMEBIC AX NATURALIST

[Vol. XLIX

cies and varieties have their chromosome numbers in
series 1:2:3:4, etc., it seems by no means improbable that
what we have imagined above has actually occurred many
times. And if one may believe that the event has the
result supposed, all the worry about relationships between
chromosome number and height of species in the scale of
evolution may be eliminated.

Xormal Chromosome Behavior axd Heredity
The second query, concerning the relation of normal
chromosome behavior to the transmission of characters,
is much more important than the one just examined, but
it can be discussed more briefly. By normal ^chromo-
some behavior' ' is meant a reduction division where ma-
ternal and paternal chromosomes approach each other in
definite pairs (if homologous pairs are present), chance
only governing the passage of either to a particular
daughter cell. This is probably the usual behavior in the
higher plants and animals, and upon this behavior depends
Mendelian heredity in the narrow sense. The thesis to be
submitted and scrutinized is the following: The maximum
possible difficulty in the improvement of animals and
plants by hybridization usually depends directly upon the
chromosome number.

TVhen a mutation in a single determinant takes place in
the germ cells of a plant, such as may cause the loss of
red color in the corolla, crosses between such a form and
the normal give a monohybrid Mendelian result. Two
mutations in non-homologous chromosomes gives in a
similar way a dihybrid result. Such simple conditions,
however, are not met with very frequently. For example,
White found that a fasciated tobacco when crossed with
the type from which it sprang and from which it probably
differed only by this single determinant, gave a mono-
hybrid Mendelian ratio in the F2 generation ; but when the
fasciated type was crossed with other types the result was
a complex F2 population. This population was suscepti-
ble of analysis, nevertheless, and showed that the various

No. 584]

HEREDITY AND ITS MEANING

483

varieties with which the fasciated type was crossed dif-
fered from it by several determinants, each of which was
transmitted independently though they every one affected
the development of fasciation. This illustration is not
one of a rare phenomenon. It is what geneticists find
constantly in their experiments. Presence or absence of
a particular character may depend upon the presence or
absence of a particular essential determinant, but, given
this determinant, sooner or later the investigator finds
several other determinants which modify the expression
of the character. The existence of these modifiers has
been the cause of a great deal of confusion in the analysis
of breeding results, but in reality the inheritance is sim-
ple. The experience that all investigators who have
worked intensively have had with them shows that prac-
tically all somatic characters are due to multiple determi-
nants in the germ cells. It merely depends on the rela-
tive difference between the germ plasms brought together
in crosses, how complex the resulting F2 populations ap-
pear. Since even apparently simple characters are thus
due to complex germinal interactions, that results of
crosses made for the purpose of improving such intangi-
ble things as yield, size, quality, etc., should be complex,
is not astonishing. In the comparatively extensive expe-
rience that the writer has had in breeding tobacco, maize,
peas and beans the wide variability of the F2 population
in crosses between distinct varieties leads him to think
that it is extremely common for such varieties to differ
qualitatively in every chromosome. Furthermore, the
relative complexity of the segregating populations is
much greater in tobacco than in corn and greater in corn
than in peas or beans. What can this mean but that when
varieties are found that differ qualitatively in all of their
chromosomes, the complexity of the result varies directly
with the number of chromosomes present.

In Mendelian inheritance the number of actual types
(both homozygous and heterozygous) present in the F2
population when all are represented is 3n, and the number

484

THE AMERICAN NATURALIST [Vol. XLIX

of individuals that must be present to give an equal chance
for the presence or absence of an individual of every type
is 4% where n represents the number of allelomorphic
pairs. This being true, if differences in all of the chro-
mosomes are predicated in tobacco and in pea crosses, the
maximum number of individuals necessary in the F2 gen-
eration to allow for one reproduction of each of the grand-
parental forms is 424 in the first case and 47 in the second
case. It is clear that there is an absolutely overwhelming
difference in the difficulty of recovering the grandpar-
ental forms in the two examples.

Xow this is about what one wishes to do in many plant-
breeding problems. It is desired to combine one or two
characters from one parent with all of the other qualities
of the second parent. And such has been my experience
that I believe that this maximum possible difficulty in the
operation as predicated by qualitative differences in all of
the chromosomes often occurs. There can be no question
on these grounds of the importance of determining the
number of chromosomes in a species before beginning
a complex plant-breeding problem, and thus being able to
comprehend the maximum possible difficulties that may
be encountered. How greatly these difficulties vary may
be seen in the very incomplete list of chromosome counts
in common plants that is given below.

Common Name

Scientific Name

Date

Investigator

Banana

Musa sapientum, "dole".

"16"

1910

Tischler, G.

Musa sapientum,

"Radjah Siam"

"32"

1910

«« ««

Musa sapientum, "Kladi"

1910

Bean

Phaseolus vulgaris

1904

Wager, H.

Calla lily. . . .

Richardia Africana

1909

Overton, J. B.

Carina

Canna indir.a.

1900

Wiegand, K. M.

more than

1904

Kornicke, M.

Zea Mays, "yellow

starchy" "amber rice

pop," "black starchy,"

"golden broach field,"

"white flint"

"20"

1911

Kuwada, Y.

Zea Mays, "red starchy,"

"red sugar"

9-10

"18-20"

1911

ii ««

Zea Mays, early 8-rowed

sugar

9-12

1911

No. 584]

HEREDITY AND ITS MEANING

485

Common Name

Scientific Name

Date

Investigator

1 9
1Z

1 Q1 1

iy 1 1

"56"

1903

Cannon, W. A.

1910

Balls, W. L.

o
o

1 QAft

iyuo

1 iscnier, kj.

Q
O

"1ft"
ID

1 QAQ

iyuy

Rosenberg, 0.

1 Or\r- 1 Q

nUAn+ Oft

aDOUt ZD

1 qa^
iyuo

Tn«l TT C\

juei, xi. yj.

1 ft

Qft

1 QAQ

iyuy

Lagerberg, T.

1909

Davis, B. M.

1 QA7

iyu/

(jreerts, J. JM.

7
1

1 Q1 1

vJiiltJo, XV. XV.

1 j.

9ft

1 QAQ

iyuy

vjraies, iv. xv.

128

1908

Yamanouchi, S.

132

1900

Strasburger, E.

1 a

1 Q1 1

iy 1 1

TnKaro A T 07.,] A T

lanara, ivi., ana ivi.

Ishikawa.

0
0

1 QAQ

iyuy

Rosenberg, 0.

A
*±

0
0

iyuo

JUci, xx. \j.

O
O

1 ft
ID

1 Q1 A

iy iu

Tolinrn A T

i anara, ivi.

1 9
1 _

1 Sftzt

Guignard, L.

1 ft

"Qft"
OD

1 ftQQ

ioyy

wiegana, xv.. ivi.

1 ft
ID

1 QAQ

iyuy

oauer, i_<. w .

I 7?

I I ;

A(\ ^A

1 qi n
iy iu

x 1 1 iicii a, ivx.

9ft

i qi a
iy iu

i anara, ivi.

Qft
CO

79?

1 QAQ

iyuy

Winkler, Hans

"ft"
0

"ift"
ID

1 ftQft

ocnanner, j. xi.

1 9
1Z

"91"

1 ftQQ

isyo

Overton, E.

1 A
14

1 QAQ

iyuo

Cannon, W . A.

ou or

1911

Hague, Stella M.

1899

Chamberlain, C. J.

244

1910

Kuwada, Y.

1904

Strasburger, E.

1913

White, 0. E.

1909

Winkler, Hans

1901

Ernst, A.

1899

Atkinson, G. F.

1896

Kornicke, M.

"16"

1893

Overton, E.

"16"

1908

Dudley, A. H.

Cotton. . . .

Currant . . .
Dandelion.

Elderberry.
Evening

primrose
Evening

primrose.

Evening

primrose.
Fern

Zea Mays, early sugar. . .
Gossypium," hybrid". . .
" , "Egyptian".

Ribes 2 sp

Taraxacum confer turn. . .

Sambucus sp

(Enothera grandiflora. . . . i

. i

0. lamarckiana,

0. gigas ,

Nephrodium molle

Flag

Hawksbeard

Lily

Lily-of-the-

Valley. . .
Lily-of-the-

Valley. . .
Mulberry. . .

Iris squalens

Crepis lanceolata var.
platy phylum

Nightshade

Onion

Peony ,

Pea

Persimmon .

Pine

Rice

Rose

Tobacco. . . .
Tomato

Tulip

Wake-robin .
Wheat

Crepis virens ....
Crepis tectorum. .
Crepis japonica .
Lilium martagon .

Convallaria majalis . .

Convallaria majalis
Morus alba, "Shirowase'

Morus indica

Solanum nigrum

Allium Cepa

Paeonia spectabilis

Pisum sativum

Diospyros virginiana

Pinus laricio

Oryza sativa

Rosa sp. — 3 species. . .

Nicotiana sp

Solanum lycopersicum.
Tulipa Gesneriana. . . .
Trillium grandiflorum.
Triticum vulgare

Among these figures are found four of our most impor-
tant crops— wheat, tobacco, corn and cotton. They con-
trast strikingly in their chromosome numbers. Wheat
and tobacco, species in which the flowers are naturally
self-pollinated, have 8 and 24 chromosomes, respectively,

4 "But we often find a larger number." Quotation marks refer to in-
ferred numbers rather than actual countings.

486

THE AMEBIC AN NATURALIST [Vol. XLIX

in their gametes. Corn and cotton, species usually cross-
pollinated, have 10-12 and 20-28 chromosomes, respect-
ively, in their germ cells. These species all have been
under cultivation since before there has been recorded his-
tory. Many varieties of each exist. It is not at all im-
probable that with thousands of years of cultivation and
selection under diverse conditions, mutations in most of
their chromosomes have persisted. If, then, improvement
means working on character complexes that involve al-
most all of the plant functions, it does not seem improb-
able that the actual difference in the difficulty of improving
wheat and tobacco is as 4s : 424, or about 1 to 4,295,000,000.
In . like manner corn and cotton compare in the ratio
410 : 428, or 1 to 68,720,000,000. And is it not true that
modern improvement in most of these crops does involve
nearly all the plant functions? Yield in wheat involves
number and size of grain, and number of culms, with all
that these things include in plant economy; yield of to-
bacco involves number, size and thickness of the leaves.
Quality, a mystical word, is perhaps still more complex.
In wheat, it takes in habit of growth of both root and stem
and such other characters as go to make up strength and
hardiness, thickness of pericarp, size of aleurone cells, and
the physical and the chemical character of both endo-
sperm and embryo, as well as their size ratios in regard
to each other. In tobacco, it includes thickness and
strength of leaf, color, texture and all chemical and physi-
cal characters that make for flavor and "burn."

One may say that this is all very well as a theory, but
that it is all theory, and ask what support is given to it by
practise. I have had personal experience with but two
of these four crops. I have worked extensively and in-
tensively with corn and tobacco for some ten years. But
I have followed carefully the published experiments in
breeding wheat and cotton and have seen several of the
more important experiments. And I may say that it teas
my observation of the extreme difficulty in the experi-
ments with cotton and tobacco as compared with com and
wheat that led to this theory of the cause.

No. 5S4]

HEREDITY AND ITS MEANING

487

In proposing this thesis, the chromosomes have been
considered as pairs of freight boats loaded with character
determiners, shifted bodily to the daughter cells by in-
ternal forces of which we are ignorant. Yet this is not
the whole truth. The determiners in particular chromo-
somes seem to be tied together more or less tightly, but
they are not always transferred as one package. They
are coupled in their transmission to the next generation,
but this coupling is not perfect. Breaks in the coupling
occur and there is order and regularity in these breaks.
Our knowledge on these matters rests upon the researches
of Morgan on Drosophila, Bateson on the sweet pea, and
Tanaka on the silkworm, so it is not certain whether these
are common grounds for this regularity or whether each
species has regular laws which control the breaks in cor-
relation. But in either case, these breaks do not inter-
fere with our proposition. They only complicate matters.
In most of the cases in Drosophila, where they are best
known, linkage is comparatively tight, i. e., breaks are
somewhat rare; but they may become so frequent as to
simulate inheritance from separate chromosomes. In
those cases our theory is of no value, but if Drosophila
is any criterion by which to judge, such conditions are
very unusual.

Abnormal Chromosome Behavior and Heredity
The third query concerning the relations of peculiar or
unusual chromosome behavior to the transmission of
characters may be passed over with a word. In certain
insects, for example, bees, wasps, aphids, phylloxerans,
etc., odd sex ratios and attendant complexities have long
been known. These have been cleared up more or less
completely by cytological studies. They depended upon
chromosome behaviors that are not usual in animals or
plants. Similar peculiar chromosome mechanisms may
be present in many other species. Attention is merely
called to the fact that if experiments on any plant species
appear to show that its characters do not obey the laws
that have been demonstrated for so many types, their

488

THE AMERICAN NATURALIST

[Vol. XLIX

cytological eccentricities should be looked into. In them
will probably be found the key to the situation. The
CEnotheras may be mentioned as a case in point. Their
heredity in many cases is not what would be expected by
analogy with other plants. We know that in some ways
the behavior of their chromosomes is irregular. Further
study will probably show that this is the sole cause of their
anomalous heredity.

LITERATURE CITED

Allen, C. E.

1905. Xuclear Division in the Pollen Mother-cells of Lilium canadense.

Ann. Bot., 19: 189-258.
Atkinson, G. F.

1899. Studies on Reduction in Plants. Bot. Gas., 28: 1-26.
Balls. W. L.

1910. The Mechanism of Xuclear Division. Ann, Bot. 24: 653-665.
Baltzer, F.

1910. Ueber die Beziehung zwischen dem Chromatin und der Ent-
wickelung und Vererbungsriehtung bei Echinodermenbastar-
den. Archiv. f. Zellforsch., 5: 497.

Bateson, W.

1909. Mendel's Principles of Heredity. Cambridge, Cambridge Uni-
versity Press. See bibliography reported therein.

Bonnevie, K.

1909. Chromosomenstudien I. Archiv. f. Zellforsch., 1: 450-514,1908.
II. Ibid., 2: 201-278.

Boveri, Th.

1892. Die Entstehung des Gegensatzes zwischen den Geschleehtszellen
und den somatischen Zellen bei Ascaris megalocephala. Sitzber.
Gesell. Morph. u. Physiol. Niinchen., Band 8.

1902. Mehrpolige Mitosen als Mittel zur Analyse des Zellkerns. Verh.
Rhys. u, Med. Gesell. zu Wiirzburg., X. F., 35.
Bridges, C. B.

1914. Direct Proof Through Xon-disjunction that the Six-linked Genes
of Drosophila are Borne by the X-chromosome. Science, N. S.,
40: 107-109.
Cannon, W. A.

1902. A Cytological Basis for the Mendelian laws. Bull. Torr. Bot.

Club, 29: 657-661.

1903. Studies in Plant Hybrids: The Spermatogenesis of Hybrid Cot-

ton. Bull. Torr. Bot. Club, 30: 133-172.
1903. Studies in Plant Hybrids : The Spermatogenesis of Hybrid Peas.
Bull. Torr. Bot. Club, 30: 519-543.

Cardiff, I. D.

1906. A Study of Synapsis and Reduction. Bull. Torr. Bot. Club, 33:

271-306.

No. 584]

HEREDITY AND ITS MEANING

489

Chamberlain, C. J.

1899. Oogenesis in Pinus laricio. Bot. Gas., 27: 268-280.
Conklin, E. G.

1901. The Individuality of the Germ-nuclei During the Cleavage of
the Egg of Crepidula. Biol. Bull, 2: 257-265.

1905. The Mutation Theory from the Standpoint of Cytology. Sci-
ence, N. S., 21: 525-529.

1908. The Mechanism of Heredity. Science, N. S., 27: 89-99.
Coulter, J. M., and Chamberlain, C. J.

1903. Morphology of the Angiosperms. N. Y., Appleton.
Davis, B. M.

1909. Cytological Studies on (Enothera: I. Pollen Development in

(Enothera grandiflora. Ann. Bot., 23: 551-571.

Delage, Yves.

1903. L'heredite et les grands problems de la biologie generate. 2me
ed., Paris, Reinwald.

DeVries, H.

1910. Intracellular Pangenesis: Including a Paper on Fertilization

and Hybridization. C. S. Gager, trans. Chi. Open Court.

Dixon, H. N.

1894. Fertilization of Pinus silvestris. Ann. Bot., 8: 21-34.
Driesch, H.

1899. Die Lokalisation morphogenetischen Yorgange. Ein Beweis
vitalistischen Geschehens. Leipzig, Engelmann.
Farmer, J. B., and Moore, J. E. S.

1905. On the Maiotic Phase (Reduction Division) in Animals and
Plants. Quart. Jour. Micr. Sci., N. S., 48: 489-557.
Parmer, J. B., and Digby, L.

1907. Studies in Apospory and Apogamy in Ferns. Ann. Bot., 21:

161-199.

Fick, R.

1905. Betrachtungen uber die Chromosomen, ihre Individualitat, Re-
duction und Yererbung. Arch. Anat. u. Physiol, Anat. Abt.,
179-228.

1908. Zur Conjugation der Chromosomen. Arch. f. Zellforsch., 1:

604r-611.

Gates, R. R.

1907. Pollen Development in Hybrids of CEnothera lata X O. Lamarlc-

iana, and Its Relation to Mutation. Bot. Gas., 43: 81-115.

1909. The Stature and Chromosomes of (Enothera gigas DeYries.

Arch. f. Zellforsch., 3: 525-552.

1911. The Mode of Chromosome Reduction. Bot. Gas. 51: 321-344.
Geerts, J. M.

1908. Beitrage zur Kenntnis der cytologischen Entwicklung von

(Enothera Lamarcl'iana. Ber. Deutsch. bot. Ges., 26a: 608-
614.

1907. Ueber die Zahl der Chromosomen von (Enothera Lamar chiana.
Ber. deut. bot. Gesell, 25: 191-195.

490

THE AMEBIC AN NATURALIST [Vol. XLIX

Gregoire, V., and Wygaerts, A.

1904. La construction du noyau et la formation des chromosomes dans

les cineses somatiques. I. Bacines de Trillium grandiflorum
et telophase homceotypique dans le Trillium cemuum. La Cel-
lule, 21: 7-76.

Gregoire, V.

1905. Les resultats acquis sur les cineses de maturation dans les deux

regnes. Revue critique de la Litterature. (I Mem.) La
Cellule, 22: 221-376.

1907. Les fondements cytologiques des theories courantes sur l'heredite

Mendelienne. Ann. Soc. Boy. Zool. et Malacol Belgique, 42:
267-320.

1908. Les phenomenes de l'etape synaptique representent-ils une caryo-

cinese? La Cellule, 25: 87-99.

1910. Les cineses de maturation dans les deux regnes. L 'Unite es-

sentielle du processus meiotique. (II Mem.) La Cellule, 26:
223-422.
Gregory, E. P.

1909. Xote on the Giant and Ordinary Forms of Primula sinensis.

Proc. Cambridge Phil. Soc., 15 : 239-246.

1909. The Mode of Pairing of Chromosomes in Meiosis. New Phy-

tologist, 8: 146-153.
1914. On the Genetics of Tetraploid Plants in Primula sinensis. Proc.
Boy. Soc, B. 87.

Guyer, M. F.

1911. Nucleus and Cytoplasm in Heredity. Am. Nat., 45: 284-305.
Haecker, V.

1907. Die Chromosomen als angenommene Vererbungstriiger. Ergebn.
u. Fortschritte d. Zoologie, 1: 1-136.

1910. Ergebnisse und Ausblicke in der Keimzellenforschung. Zeitschr.

Abstammungs. u. Vererbungslehre, 3: 181-200.

Hague, S. M.

1911. A Morphological Study of Diospyros virginiana. Bot. Gaz., 52:

34-42.
Hegner, E. W.

1911. Germ-cell Determinants and their Significance. Am. Nat., 45 :
385-397.

Henking, H.

1891. Ueber Spermatogenese und deren Beziehung zur Eientwicklung
bei Pyrrhocoris. Zeitschr. wiss. Zool., 50: 1891.

Heider, K.

1906. Vererbung und Chromosomen. Versamml. deut. Natur u. Artze.
Jena.

Herbst, C.

1909. Vererbungsstudien VI. Arch. f. Fntw., 27: Heft 2.
Hertwig, O.

1898. Die Zelle und die Gewebe. Jena, Fischer.
Hyde, E.

1909. The Seduction Division in the Anthers of Eyacinthus orientalis.
Ohio Nat., 9: 539-544.

No. 584]

HEREDITY AND ITS MEANING

491

Ishikawa, M.

f 1911. Cytologische Studien von Dahlien. Bot. Mag., Tokyo, 25: 1-8.
Janssens, F. A.

1905. Spermatogenese dans les Batraciens: Evolution des auxcytes
males du Batrahoseps attenuatus. La Cellule, 22: 380-425.

Juel, H. O.

1S97. Die Kernteilungen in den Pollen-Mutterzellen von Hemerocallis
fulva und die bei denselben auftretenden Unregelmassigkeiten.
Jahrb. wiss. Bot., 30: 205-226.
Koernieke, M. Der heutige Stand der pflanzlichen Zellforschung.

1904. Ber. dent. bot. Gesell., 21: 66-134.
Kolliker, A.

1885. Die Bedeutung der Zellkerne fiir die Vorgange der Vererbung.
Zeitschr. wiss. Zool., 52: 1-46.

Kuwada, Y.

1910. A Cytological Study of Oryza sativa. Bot. Mag., Tokyo, 24:

267-281.

1911. Maiosis in the Pollen-mother Cells of Zea Mays, L. Bot. Mag.,

Tokyo, 25: 163-181.

Lagerberg. T.

1909. Studien iiber die Entwickelungsgeschiehte und systematisehe
Stellung von Adoxa moscTiatellina. Kungl. Sv. Vet. Alad.
Handl, 44 : 1-86.
Lawson, A. A.

1911. Xuclear Osmosis as a Factor in Mitosis. Trans. Roy. Soc.

Edinburgh, 48: 137-161.
1911. The Phase of the Xucleus Known as Synapsis. Trans. Soy.

Soc. Edinburgh, 47: Pt. Ill: 591-604.

Lotsy, J. P.

1906. Torlesungen iiber Descendenz-theorien. Two vols. Jena, Fischer.
Lundegard, H.

1909. Ueber Reduktionsteilung in den Pollenmutterzellen einiger diko-

tyler Pflanzen. SvensJc. Bot. Tidshr., 3: 78-124.

1910. Ueber Kernteilung in den Wurzel-spitzen von Allium cepa and

Vicia faba. SvensJc. Bot. Tidslcr., 4: 174-196.
Lutz, Anne M.

1909. Notes on the First Generation of Hybrids of O. lata X O. gigas.

Science, N. S., 29: 263-267.
Moenkhaus, W. J.

1904. The Development of Hybrids between Fiuidulus heteroclitus
and Menidia notata with Especial Reference to Behavior of
Maternal and Paternal Chromosomes. Am. Journ. Anat., 3:
29-64.
Montgomery, Jr., T. H.

1906. Chromosomes in the Spermatogenesis of the Hemiptera heterop-

tera. Trans. Am. Phil. Soc, N. S., 21: 97-173.
1908. On the Morphological Difference of the Chromosomes of Ascaris
megalocephala. Arch. f. Zellforsch., 2: 66-75.
Morgan, T. H.

1910. Chromosomes and Heredity. Am. Xat., 44: 449-496.

492

THE AMEBIC AN NATURALIST [Vol. XLIX

1913. Heredity and Sex. N. Y., Columbia University Press.
Mottier, D. M.

1907. The Development of Heterotypic Chromosomes in Pollen Mother-
cells. Ann. Bot., 21: 309-347.

Nakao, M.

1911. Cytological studies on the nuclear division of the pollen mother
cells of some cereals and other hybrids. Jour. Col. Agr. Sap-
poro (Japan), 4: 173-190.
Nawaschin, Sergius.

1910. Naheres iiber die Bildung der Spermakerne bei Lilium Marta-
gon. Annates d. Buitenzorg, Sleme Supplement, 2e Partie:
871-904.

Nemee, B.

1909. Zur Mikrochemie der Chromosomen. Ber. deutsch. Bot. Gesell.,

27: 43-47.

1910. Das Problem der Befruchtungsvorgange und andere cytologische

Fragen. Berlin, Borntraeger.
Overton, J. B.

1909. On the Organization of Nuclei in the Pollen-mother Cells of
Certain Plants with Special Eeference to the Permanence of
the Chromosomes. Ann. Bot., 23: 19-61.

Eabl, C.

1885. Ueber Zelltheilung. Morpliol. Jalirb., 10: 214-330.
Eemak, E.

1858. Ueber die Theilung der Blutzellen beim Embryo. MiUler's
Arch., S. 178-188.
Eosenberg, O.

1905. Zur Kenntnis der Eeduktionsteilung in Pflanzen. N. Bot. Not.,

S. 1-24.

1906. Cytological Investigations in Plant Hybrids. Ept. 3d Intern.

Conf. on Genetics, pp. 289-291.

1907. Experimental and Cytological Studies on the Apogamy in Hiera-

cium. Svenslc. Bot. Tidslcr., 2: 143-170.
1909. Ueber den Bau des Euhekerns. Svenslc. Bot. Tidslcr., 3: 163-
173, Taf. 5.

1909. Ueber die Chromosomenzahlen bei Taraxacum und Bosa.

Svenslc. Bot. Tidslcr., 3 : 150-162.
1909. Cytologische und morphologische Studien uber Drosera longi-

folia X B. rotundifolia. Kungl. Svenslc. Vetenslc. Handl., 43:

1-64.

1909. Zur Kenntnis von den Tetradenteilungen der Compositen.
Svenslc. Bot. Tidslcr., 3: 64-77.

Sargent, E.

1896. The Formation of Sexual Nuclei in Lilium Martagon. I.

Oogenesis. Ann. Bot., 10: 445-477.

1897. The Formation of Sexual Nuclei in Lilium Martagon. II.

Spermatogenesis. Ann. Bot., 11: 187-224.

Sauer, L. W.

1909. Nuclear Divisions in the Pollen Mother-cells of Convallaria
majalis. Ohio Nat., 9: 497-505.

No. 584]

HEREDITY AND ITS MEANING

493

Schaffner, John H.

1909. Chromosome Difference in Ascaris mcgaJocephaki. Ohio Nat.,
9: 506-508.
Stevens, N. M.

1905. Studies in Spermatogenesis with Especial Reference to the

11 Accessory Chromosome. ' 7 Carnegie Inst., Wash., Pub. 36, I.

1906. Studies in Spermatogenesis, II. A Comparative Study of the

Heterochromosomes in Certain Species of Coleoptera, Hemip-
tera and Lepidoptera, with Especial Reference to Sex Deter-
mination. Carnegie Inst., Wash., Pub. 36, II.

Stomps, T. J.

1911. Kerndeeling en Synapsis by Spinacea oleracea. S. 1-162.
1910. Review Biol. Centralbl., 31: 257-320.
Strasburger, E.

1894. Periodic Reduction of Number of Chromosomes in the Life His-
tory of Living Organisms. Ann. Bot., 8: 281-316.

1904. Ueber Reduktionsteilung. Sitzber. K. Preuss. Al'ad. Wiss., 18:

587-614.

1905. Typische und allotypische Kernteilung. Histologische Beitrage

zur Vererbungsfrage. Jahrb. wiss. Bot., 52: 1—71.

1907. Ueber die Individualist der Chromosomen und die Pfropfhy-

briden Frage. Jahrb. f. wiss. Bot., 44: 482-555.

1907. Die Ontogenie der Zell seit 1875. Progressus rei Botanicce, 1:

1-138.

1908. Chromosomenzahlen, Plasmastrukturen, Vererbungstrager, und

Reduktionsteilung. Jahrb. f. wiss. Bot., 45: 479-570.

1909. Meine Stellungnahme zur Frage der Pfropfbastarde. Ber.

Deutsch. Bot. Gesell., 27: 511-528.
1909. Zeitpunkt der Bestimmung des Geschlechts, Apogamie, Parthe-
nogenesis und Reduktionsteilung. Histologische Beitrage, Heft
7, S. 1-124. Jena.

1909. The Minute Relation of Cells in Relation to Heredity. Darwin

and Modern Science, pp. 102-11. Cambridge, Cambridge Uni-
versity Press.

1910. Chromosomenzahl. Flora, 100: 398-446.
Sutton, W. S.

1903. On the Morphology of the Chromosome Group in Brachystola

magna. Biol. Bui., 4: 24-39.
1903. The Chromosomes in Heredity. Biol. Bui., 4: 231-251.
Tahara, M.

1910. Ueber die Kernteilung bei Morus. Bot. Mag., Tokyo, 24:
281-289.

1910. Ueber die Zahl der Chromosomen von Crepis japonica, Benth.

Bot. Mag., Tokyo, 24: 23-28.
Tahara, M., and Ishikawa, M.

1911. The number of chromosomes of Crepis lanceolata var. platy-

phyllum. Bot. Mag., Tokyo, 25: 119-121.

Tanaka, Y.

1913. A Study of Mendelian Factors in the Silkworm, Bombyx mori.
Journ. Co}. Agr., Sapporo, 5: 91-113.

494

THE AMEBIC AN NATURALIST

[Vol. XLIX

Tennent, D. H.

1908. The Chromosomes in Cross-fertilized Echinoid Eggs. Biol. Bui.,

15: 127-134.

Tischler. G.

1906. Ueber die Entwicklung des Pollens und der Tapetenzellen bei
Ribes-Hybriden. Jalirb. f. wiss. Bot., 42: 558.

1008. Zellstudien an sterilen Bastardpflanzen. Arch. f. Zellforsch.,
1: 3-151.

1910. Untersuchungen iiber die Entwicklung des Banana-Pollens.
Arch, f. Zellforsch., 5: 622-670.

Wager, H.

1904. The Nucleolus and Nuclear Division in the Root-apex of Phaseo-
lus. Ann, Bot., 18: 29-55. (Number of chromosomes deter-
mined by White from a study of Wager's figures.)
Weismann, A.

1892. Das Keimplasma. Jena.
White. O. E.

1913. The Bearing of Teratologieal Development in Nicotiana on
Theories of Heredity. Am. Nat., 52: 206-228.

Wilson. E. B.

1904. The Cell in Development and Inheritance. New York, Mac-
millan.

1901. Experimental Studies in Cytology. A Cytological Study of
Artificial Parthenogenesis in Sea Urchin Eggs. Arch. f.
Entw., 12: 529-596.

1909. Studies on Chromosomes, V. The Chromosomes of Metapodius.

A Contribution to the Hypothesis of the Genetic Continuity of
Chromosomes. Joum. Exp. Zool., 6: 147-205.

1910. The Chromosomes in Relation to the Determination of Sex.

Science Progress, 5: 570-592.

1911. The Sex Chromosome. Arch. milr. Anatomie, 77: 249^271.

(Bibliography on accessory chromosome.)
Winkler, Hans.

1908. Ueber Parthenogenesis und Apogamie im Pflanzenreiche. Pro-

gressus rei Botanicce, 2: 293-454.
1910. Ueber die Nachkonimenschaft der Solanum-Propfbastarde und
die Chromosomenzahlen ihrer Keimzellen. Zeitschr. Bot., 2:
1-38.
Yamanouchi, S.

1908. Sporogenesis in Nephrodium. Bot. Gaz., 45: 1-30.
Ziegler, H. E.

1906. Die Chromosomen-Theorie der Vererbung in ihrer Anwendung
auf den Menschen. Arch. Eassen-Gesellsch. Biologie, 3: 797-
812.

Zoja, R.

1895. Sulla independenza della eromatina paterna et materna nel
nucleo delle cellule embrionali. Anat. Anz., 11: 289-293.

THE CONNECTICUT

AGRICULTURAL EXPERIMENT
STATION

NEW HAVEN, CONN.

BULLETIN 188, SEPTEMBER, 1915

FURTHER EXPERIMENTS ON
INHERITANCE IN MAIZE

H. K. HAYES and E. M. EAST.

CONNECTICUT
AGRICULTURAL EXPERIMENT STATION.

OFFICERS AND STAFF.

BOARD OF COXTROL
His Excellency, Marcus H. Holcomb, ex-officio, President.

Prof. H. W. Conn, Vice President Middletown

George A. Hopson, Secretary Wallingford

E. H. Jenkins, Director and Treasurer New Haven

Joseph W. Alsop Avon

Wilson H. Lee Orange

Frank H. Stadtmueller Elmwood

James H. Webb Hamden

Administration

K. H. Jenkins. Ph.D., Director and Treasurer.
Miss V. E. Cole, Libra/ ian and Stenographer.
Miss L. M. Brautlecht, Bookkeeper and Stenographer.
William Yeitch, In charge of Buildings and Grounds.

Chemistry.

Analytical Laboratory.

John Phillips Street. M.S., Chemist in Charge.
E. Monroe Bailey, Ph.D., C. B. Morison, B.S.
C. E. Shepard. G. L. Davis. Assistants.
Hugo Lange, Laboratory Helper.
V. L. Churchill, Sampling Agent.
Miss E. B. Whittlesey, Stenographer.

Proteid Research.

T. B. Osborne, Ph.D., D.Sc, Chemist in Charge.
Miss E. L. Ferry, M.S., Assistant.

Botany.

G. P. Clinton, Sc.D., Botanist.
E. M. Stoddard, B.S., Assistant Botanist.
Miss E. B. Whittlesey, Herbarium Assistant.
G. E. Graham, General Assistant.

Entomoloi

W. E. Britton, Ph.D., Entomologist ; State Entomologist.

B. H. Walden, B.Agr., First Assistant.

O. S. Lowry, B.Sc, I. W. Davis, B.Sc. I A .

r, _ „ _ ' i" Assistants.

M. P. Zappe, B.S. *

Miss G. A. Foote, Stenogt aphcr.

Foresti

Walter O. Filley, Forester; also State Forester

and State Forest Fire Warden.
A. E. Moss. M.F., Assistant State and Station Forester.
Miss E. L. Avery, Stenographer.

Plant Breeding.

Donald F. Jones, B.S., Plant Breeder
C. D. Hubbell, Assistant.

Vegetable Growing.

Howard F. Huber, B.S.

FURTHER EXPERIMENTS ON INHERITANCE IN

MAIZE.

H. K. Hayes * and E. M. East.

This paper is a report on the inheritance of certain differ-
ences in the endosperm of various maize races that have been
made the basis of a division into the subspecies everta, indurata,
indentata and amylacea. To these investigations, a genetic
study of the shape of seed which characterizes the socalled
rice pop corns is added.

The writers take pleasure in acknowledging the efficient
aid of Air. A. F. Schultze. assistant botanist at the Connecticut
Agricultural College, and Mr. C. D. Hubbell, assistant at the
Connecticut Agricultural Experiment Station, in the consider-
able amount of field work involved.

MATERIAL AND METHODS.

The parental races used in the crosses were self-fertilized
for several years before any hybrids were made, and are be-
lieved to have been homozygous for the characters studied.
The material from which these races originated was described
in a previous publication (See East and Hayes, 1911), but
the following additional points regarding it should be noted :

1. Zea mays everta. The pop corns.
No. 6T White rice pop.

This white pop is one of the lines which has been pro-
duced from No. 23, (East & Hayes. 1911). It breeds true
to the "rice" type of seed, — sharply pointed where the style

* Mr. Hayes resigned January 1, . 1914, to take charge of plant
breeding work in the Experiment Station and College of Agriculture of
the University of Minnesota. The experimental work here reported was
carried on at the Connecticut Station as an Adams Fund Project. The
Minnesota Experiment Station and the Bussey Institution of Harvard,
should be given credit for time spent in the preparation of this paper for
publication.

2 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

(silk) was attached,— although there is some variation in the
degree to which this character is expressed. The seeds con-
tain only very small amounts of soft starch.
No. 65. A white, flint-like pop.
This is a strain produced from No. 26, of our previous
publication. Its seeds resemble those of a typical flint variety
in shape, and contain only very small amounts of soft starch.

2. Zea mays indurata. The flint corns.
No. 5. Watson's white flint.

This variety is a true white flint which developes a red
pericarp in full sunlight. The depth of tint which developes
naturally is therefore inversely proportional to the thickness of
the husk. The seeds contain a larger proportion of corneous
starch than many races of flint corn, though .less than that shown
by the two pop varieties just described. As in all flints, how-
ever, there is a small zone of soft starch in the center of the seed.

3. Zea mays indent at a. The dent corns.
No. 6. Learning dent.

This is a vigorous strain of a famous yellow dent. Like
all varieties of its group, the soft starch extends over the whole
summit of the seeds, yet the layer is thin enough to allow the
race to be classified as a smooth dent (i. e. not beaked).

4. Zea mays amylacea. The flour corns.
No. 10. White flour.

This is a floury race with seeds resembling the average 8
rowed flint in shape. Though the seeds usually contain only
floury starch, sometimes an almost imperceptible layer of corn-
eous starch developes in the exterior of the endosperm. It
seems likely that this variation is an effect of external condi-
tions rather than of gametic impurity.

The plantings have always been made from the original
seed envelope, and pains have been taken to prevent the mis-
placement of seeds.

The different families were marked in the field by heavy
stakes to which wired tree labels were attached, but to prevent
error through their misplacement a planting plan was made each
year showing the exact location and the number of hills of each
strain.

INHERITANCE IN MAIZE.

Classification of seeds was made only from hand pollinated
ears, although the remaining ears of a selection were always
examined, and in the case of those seed characters not immediate-
ly affected by pollination, were used in determining the range
of variation.

The various races were given different numbers as Xo. 10
flour corn and No. 5 flint corn. A cross between 10 and 5 was
then written as 10 x 5 the female parent appearing first. Differ-
ent self-pollinated ears obtained from growing the cross between
(10x5) were labeled (10x5)-l, (10x5)-2, etc. Later genera-
tions were labeled as (10x5)-l-2, (10x5)-l-3, (10x5)-2-4, etc.
If the Fj_ generation was pollinated with pollen from the flint
parent, this ear received the label (10x5)-l x (5-2)-8-3, as the
case might be. This back cross was planted the following year
as (10x5 x5). Thus we had complete records of the parents
and ancestry of our various lines.

The field technique has been described in previous publica-
tions.

For convenience the various crosses will be considered under
special headings.

Family (10x5), Flour x Flint.

A cross between the floury race Xo. 10 and flint race Xo. 5
was made in 1910, the resulting seeds resembling the female
parent. As indicated above, the characteristic difference between
these races is the amount of soft starch in the seeds. The flint
race produces a small quantity of soft starch in the center of the
seed, surrounded by a large layer of corneous starch, while the
flour race produces only an occasional trace of corneous starch
around the exterior of the endosperm. Xo immediate effect of
pollination through double fertilization was expected, as both
our own earlier results and those of other investigators (Correns
and Lock) were thought to imply that these differences in the
starchy character of the endosperm behaved in heredity as if
they pertained to the plant rather than to the endosperm. On
growing this cross in 1910, however, we were much surprised
to find a clear segregation of seeds on each ear. This fact
showed that the physical condition of the starch in these races

4 CONNECTICUT EXPERIMENT STATION. BULLETIN 188.

was not a maternal character, since in that case we should have
expected a uniform population of seeds on the F1 ears, resembling
either the male or female parents or intermediate between them.

A classification of the seeds from the ears of the F1 gen-
eration plants, is given in Table 1. Only two classes could be
made ; corneous seeds like the flint parent, and floury seeds re-
sembling the floury parent. There was no difficulty in dividing
the seeds into these two classes. Of the thirteen ears shown
in Table 1, some contained a greater proportion of flint or of
floury seeds than others, but all gave close approximations to a
1 to 1 ratio. This being a novel F1 ratio, further experiments
were made to find a genetic interpretation of it.

TABLE 1.

Self-pollinated Ears from the F1 Generation of a Cross
Between No. 10 Flour and No. 5 Corneous Flint.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5)-l

145

186

-3

208

142

-4

169

161

-5

156

169

-6

181

166

-7

189

172

-8

175

203

" -9

168

165

-10

213

-11

209

205

-12

238

237

-13

190

197

-14

252

223

Total

2493

2439

The floury seeds of (10 x5)-"i and (10x5)-S were labeled
(10xo)-7S and (10x5)-8S to distinguish them from the cor-
neous (flint-like) seeds 'of the same ears, which were labeled
(10x5)-7C and (10x5)-8C respectively. The data from sev-

1 The word hybrid in these discussions is used in a peculiar sense to
avoid longer descriptions. It means a cob bearing a population of seeds
belonging to more than one phenotype.

INHERITANCE IN MAIZE.

eral self-fertilized ears obtained by growing the floury seeds
are given in Table 2. Of a total of 11 hand-pollinated ears, 8
were hybrid1, and gave 1 to 1 ratios with a total of 748
corneous to 691 floury seeds. The other 3 ears bred true for
the floury habit.

Of the open field or naturally pollinated ears, 28 were hy-
brids and 23 pure floury. This gives a total of 36 hybrids to
26 pure floury, which, considering the number grown, is a rea-
sonable approximation of a 1 to 1 ratio.

TABLE 2.

Self-pollinated Ears Obtained Through Growing Floury
Seeds of Ears (10 x 5) -7 and (10 x 5) -8.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5)-7 S-l

108

125

-7 S-2

" -7 S-4

162

126

-7 S-7

" -8 S-5

100

" -8 S-6

" -8 S-7

" -8 S-8

100

" -8 S-2

Pure Floury

" -8 S-3

u a

-8 S-4

it a

Total in hybrid ears

748

691

Table 3 gives the results of planting the corneous seeds of
ears (10xo)-7 and (10xo)-8. Of a total of 9 self-fertilized
ears, 5 proved to be hybrids and 4 were pure corneous. The
ratio of corneous to floury seeds in these 5 hybrid ears was
461 corneous to 482 floury, a close approximation of 1 to 1.
Of the open field ears 38 were corneous and 34 hybrids. Thus
in this case the hybrid and the pure corneous ears are clearly
in a 1 to 1 ratio.

(5 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

TABLE 3.

Self-pollinated Ears Obtained Through Growing Corneous
Seeds of Ears (10 x 5)-7 and (10 x 5)-8.

Ear X umber

Corneous Seeds

Floury Seeds

(10 x 5)-7C-6

* -7C-9

101

-8C-3

" -8C-8

191

211

-8C-10

-7C-5

Pure corneous

-7C-8

<< tt

-8C-5

<< «

" -8C-6

tt tt

Total in hybrid ears

464

482

Table 4 gives the results of pollinating ears of the Fx plants
with pollen from the parental strains No. 10 flour, and No. 5
flint, respectively. Only 1 ear was obtained from the back
cross between (10x5) and the No. 10 parent. This ear had
156 corneous and 184 floury seeds. Three ears resulted from
crossing plants of (10 x 5) with the flint, or No. 5 parent. These
ears showed various ratios of corneous to floury seeds, but the
deviations from 1 : 1 ratios were not all in the same direction.
Of the total number of seeds in the four ears, 544 were corneous
and 543 floury.

TABLE 4.

Ears of the First Generation Cross of (10 x 5) Pollinated
With Pollen From the Pure Parents, No. 10 Flour
and No. 5 Corneous Flint.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5) -13 x (10-3) -14

156

184

- 3 x ( 5-3) -1

102

- 1 x ( 5-3) -3

107

- 5 x ( -5-3) -7

179

201

Total

544

543

[NllERITANCE IX MAIZE.

Table 5 gives the results obtained from planting floury seeds
of ears (10 x 5)-5 x (5-3)-7 and (10 x 5)-l x (5-3)-3 of Table
4. It was expected that such seeds would be hybrids between
the corneous and floury types and should therefore give hybrid
ratios when grown. The table shows 10 self-pollinated ears which
gave a ratio of 1014 corneous to 850 floury seeds. Seventy-
nine naturally pollinated ears were all hybrids showing a definite
segregation. The corneous seeds of ears (10 x 5)-5 x (5-3)-')
and (10x5)-l x (5-3) -3 were also tested. A total of 13 self-
fertilized and open field ears were pure corneous flints like
the corneous flint parent, No. 5.

TABLE 5.

Self-pollinated Ears Obtained From Planting Floury
Seeds of Ear (10 x 5) -5 x (5-3 )-7 and Ear
(10 x 5)-l x (5-3)-3.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5) x 5-7S-6

102 •

116

-5

125

137

-1

-8

126

110

-2

128

106

-7

(10 x 5) x 5-3S-2

-8

-1

126

-6

Total

1014

850

Table 6 gives the results obtained from planting corneous
seeds of ears (10x5)-13 x (10-3)-14. As these seeds were as-
sumed to be the result of a cross between corneous and floury
types, it was to be expected that all resulting ears would show
segregation. Five self-fertilized ears evidently came from hybrid
seeds as they gave a total ratio of 653 corneous to 620 floury
seeds. Of 5"J open field ears, 56 came from- hybrid seeds. One
ear which was somewhat immature probably was a pure soft
floury ear. This result may be explained by assuming that one
floury seed was planted by mistake.

8 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

Of the 7 self-fertilized ears obtained from planting the floury
seeds of the cross between (10x5) -13 x (10-3) -14, all were
pure floury. Of the open pollinated ears, 11 were unquestion-
ably pure floury while 2 indicated segregation. These ears may
have come from corneous seeds planted by mistake, althougn
it is possible that a few stalks were mislabeled at harvesting
time, as the stalks bearing the open pollinated ears all were
shocked on the same field.

TABLE 6.

Self-pollinated Ears Obtained From Planting Corneous
Seeds of Ear No. (10 x 5)-13 x (10-3)-14.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5) x 10-3-14C-4

127

116

-9

200

172

-6

-10

183

167

Total

653

620

Table 7, gives the results of planting seeds of Ear No.
(5-3) -20, pure corneous flint, which was pollinated with pollen
from Fj generation cross (10 x 5). There was no immediate
effect of the pollen of (10 x 5) -6 upon the pure flint ear (5-3) -20.
Of 5 self-fertilized ears obtained from growing this cross, 4
showed segregation, giving a total of 528 corneous to 508 floury
seeds, and 1 was pure corneous. Of the open field ears 24 were
pure corneous and 34 showed segregation. These results show
that the pollen grains carry the factors for corneous and floury
starch in the ratio of 1 to 1.

Table 8 gives the results of planting seeds of ear (10-3) -13,
which was pollinated with pollen from an F1 ear (10 x 5)-14.
There was no visible effect on the endosperm of (10-3)-13 due
to crossing. Three of the self-fertilized ears obtained from this
cross had a total of 397 corneous to 377 floury seeds; 6 self-
fertilized ears were like the floury parent. Of the open field
ears, 32 were homozygous floury and 30 were hybrids.

INHERITANCE IX MAIZE.

TABLE 7

Self-pollinated Ears Obtained From Planting Corneous
Seeds of Ear No. (5-3) -20 x (10 x 5) -6.

Ear Number

Corneous Seeds

Floury Seeds

5 x (10 x 5) -2

116

113

-3

116

-o

126

120

-8

170

159

-9

Pure corneous

Total in hybrid ears

528

508

TABLE 8.

Self-pollixated Ears Obtained From

Planting Floury

Seeds of Ear No. (10-3)-13 x (10 x 5)-14.

Ear Number

Corneous Seeds

Floury Seeds

10 x (10 x 5)-6

158

156

-7

-4

155

142

-1

Pure floury

" -2

-3

-4

a -

-5

-6

Total in hybrid ears

397

377

Table 9 gives the results of planting the corneous seeds of
(10 x 5)-8C-8 and (10 x 5)-8S-8. This F, generation was grown
to determine whether a constant splitting into a 1 to 1 ratio in
the hybrid ears could be expected. The results show no great
deviations from this ratio. On 9 selfed ears showing segrega-
tion there were 996 corneous and 954 floury seeds.

The total progeny of (10x5)-8C-8 consisted of 12 hybrid
ears and 15 corneous ears, while the progeny of (10x5)-8S-8C
included 17 hybrid and 10 pure corneous ears. Considering
the few individuals grown the data corroborate those of the pre-
vious generation.

10 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

TABLE 9.

Self-pollinated Ears Obtained From Planting Corneous
Seeds of F2 Generation Ears (10 x 5)-8C-8 and
(10 x 5)-8S-8.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5)-SC-SC-l

150

116

-3

116

133

" -2

Pure corneous

-5

-7

-8

(10 x 5)-8S-SC-l

114

132

-2

96'

115

-4

103

-5

142

104

-6

114

it rt
- t

101

-8

-3

Pure corneous

Total in hybrid ears

996

954

Table 10 gives the results of planting floury seeds of ears
(10 x 5)-SC-S and (10 x 5)-8S-8. In 8 self-pollinated ears there
were a total of 966 corneous and 99 T floury seeds. Among the
progeny of (10 x 5)-8C-8S there were IT segregating ears and 16
floury ears, while the progeny of ( 1(> x 5)-8S-8S gave a total of
12 segregating and 10 floury ears. The data in these two tables
show that the progeny of an ear which is a cross between floury
and corneous may be expected to give a ratio in F2 of 1 cor-
neous. 2 segregating to 1 floury ear.

r N H ERIT A N CE IX MAIZE. SUMMARY.

I 1

TABLE 10.

Self- pollinated Ears Obtained From Planting Floury Seeds
of (10 x 5)-8C-8 and (10 x 5)-8S-8.

Ear Number

Corneous Seeds

Floury Seeds

(10 x 5)-8C-8S-l

112

132

-2

157

174

-5

155

150

-6

100

-7

150

(10 5)-8S-8S-2

107

-3

100

-4

-1

Pure Floury

-4

(10 x 5)-8C-SS-S

-3

Total in hybrid ears

966

997

To test the purity of apparently homozygous segregates the
seeds of pure corneous ear (10 x 5)-8C-6 were planted. A total
of 63 ears were all pure for the corneous habit. Pure floury
ear (10x5)-8S-2 gave a progeny of 78 ears. All were of a
similar character and contained seeds which were nearly filled
with soft starch. There were traces of corneous matter in some
seeds, but under Connecticut conditions the floury parent also
produces traces of corneous matter in a few seeds.

Summary and Interpretation of Results.

In general, no matter which variety was used as the female
parent, there was no immediate visible effect of the male parent
in the endosperm of crosses between No. 5 flint and No. 10
floury maize. The F1 generation plants produced ears in which
there was a clear segregation of corneous and floury seeds in
a 1 to 1 ratio. This ratio was unaffected whether the Fx ears
were pollinated with pollen from either the pure flint or the pure
floury parent. The progeny of a cross between Fx and the flint
parent gave a ratio of 1 hybrid ear to 1 pure flint ear. Like-
wise the progeny of a cross between Fx and the floury parent
gave a ratio of 1 floury ear to 1 hybrid ear. Seventy-six

J 2 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

F2 ears produced from a self-fertilized Fx ear of cross (10 x 5),
gave a ratio of 1 pure flint ear, 2 hybrid ears and 1 pure
floury ear. The flint and the floury ears bred true in later
generations.

A total of 69 self-fertilized ears showing segregation gave
a ratio of 8,803 corneous seeds to 8,562 floury seeds. This is
a ratio of 1 to .961 or approximately 1 to 1.

The above results prove that the visible endosperm character
of a seed shows the potentiality of the female gamete which
entered into that particular seed, and that the male gametes
have no immediate effect on the endosperm to determine whether
they be corneous or floury. Data from later generations, how-
ever, show that the pollen grains of plants from hybrid seeds
transmit both the corneous and the floury condition, approxi-
mately J/2 carrying a factor for corneous seeds and the other
half a factor for floury seeds.

Two hypotheses will explain the facts : either there is no
fusion between the female endosperm nucleus and the so-called
second male nucleus of the pollen grain, in which case the en-
dosperm developes wholly from the endosperm nucleus of the
embryo sac and therefore exhibits the gametic character of the
egg cell : or, there is dominance of the condition of the mother.
As ordinarily two female polar nuclei unite with a single male
nucleus to produce the endosperm it might be expected that this
double dose of the female character should predominate over a
single dose of the male character, so that by inspection the seeds
would be classed as of the mother type. Correns (1901) used the
second hypothesis to account for certain results in his study of
the inheritance of color in the aleurone cells, where there ap-
peared to be a dominance of the maternal condition. Although
East and Hayes (1911) were able to show that Correns' assump-
tion was unnecessary in the case of aleurone color, the aberrant
ratios obtained being due to the interaction of several factors,
it does appear to fit the facts in the crosses just described.

A cross between a yellow corneous race and a white floury
race would show the correct explanation of the results of the
floury-flint cross, for if in F2 the ratio of yellow to white was
3 to 1, and of corneous to floury, 1 to 1, it would then be

INHERITANCE IN MAIZE.

established that there was a fusion of the female polar nuclei
with a male generative cell. Emerson suggested that the
same test could be made by pollinating ears which were expect-
ed to give a 1 to 1 ratio with pollen from a yellow corneous flint.
At the time this test was to be made no seeds of the immediate
cross between the corneous and floury races were available, but
a number of seeds of hybrid ears (10 x 5)-8C-8S-6 were planted
and the resulting plants crossed with pollen from a corneous
yellow flint known to breed true. Four ears were obtained of
a cross between (10 x 5)-8C-8S-6C (the corneous seeds) and
the yellow flint. They proved to be yellow corneous flints. Of
the naturally pollinated ears obtained from (10 x 5)-8C-8S-6C,
12 were pure corneous flints and 18 hybrids. Five ears of
(10 x 5)-8C-8S-6S (the floury seeds) were also pollinated with
pollen from the yellow corneous race. All five ears were yellow
and four were yellow floury ears. One ear was a definite hy-
brid, however, and gave a ratio of 55 floury seeds to 59 cor-
neous seeds. Of the open field ears of (10 x 5)-8C-8S-6S, 13
were pure floury and 14 hybrids.

The ear which had all ycllozi' seeds and yet showed a ratio
of 55 floury to 59 corneous, seems sufficient evidence for con-
cluding that the apparent dominance of the condition of the
mother is due to the fact that the endosperm is produced from
a union between two female polar nuclei and one male cell.
Thus two doses of a flour corn factor dominates one dose of the
corneous factor and vice versa. This fact has an important
bearing on the multiple factor hypothesis for interpreting the in-
heritance of quantitative characters, for it shows that a series of
factors may have cumulative somatic effects.

Family (10x6), Flour x Dent.

This cross was made in 1909 between self-fertilized strains
of Learning No. 6 and floury No. 10. An F1 generation was
grown in 1910, and an F2 generation from the seeds of Fx ear
(10 x 6)-l was produced in 1911. There was no appreciable
effect on the physical condition of the starch in the seeds of No.
10 due to the pollen of No. 6. On the Fx ears the seeds were in-

14 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

termediate between No. 10 and No. 6 in size, and were rather
uniformly dented. As regards the appearance of the starch in
the seeds, there was definite segregation, but classification was
difficult due to the fact that all seeds contained soft starch at
the cap and sides and were dented. The seeds of the self-fertiliz-
ed F1 and F2 ears were all examined carefully against a strong
light, however, and were classified as accurately as possible. The
results of this classification are given in Table 11. Considerable
variation in the ratios on the different ears is exhibited, but as a
rule there is an indication of a 1 to 1 ratio.

Although this seed classification may not have been as ac-
curate as might be desired owing to the difficulties involved, the
division of the total population of F2 ears into corneous, hybrid
and floury types as shown in Table 12, is exact and serves as a
complete corroboration of the theory. Thirty-six ears were
classed as pure corneous, eighty as hybrids and thirty-seven as
pure floury. This is certainly a close approximation of a 1:2 :1
ratio.

An examination of the F2 ears showed that there was con-
siderable range of variation between the different ears which
were classed as corneous or floury types. There was a little va-
riation among the seeds of the same ear, but this was not
greater than could be explained by differences in development
due to physiological causes. The pure corneous or pure floury
ears, however, differed from each other by a considerable amount,
and it seemed likely that some of this variation would be in-
herited. Of the ears of Table 11, (10 x 6)-l-13, (10 x 6) -1-3
and (10 x 6) 1-4 bred true to the floury type.

The corneous seeds of ear (10 x 6) -1-5 produced 13 hybrid
and 17 pure corneous ears, while the floury seeds yielded 19 pure
floury and 16 hybrid ears. Corneous ear (10 x G) -1-5-2 was
grown the following year and produced dented ears which bore
seeds containing a fair proportion of corneous starch.

Ears (10 x 6)-l-6, (10 x 6)-l-9, (10 x 6)-l-12 and (10 x 6)
-1-14 of Table 11 were grown in 1912. All produced ears hav-
ing seeds with a considerable proportion of corneous starch, the
progeny of No. (10 x 6)-l-6 and No. (10 x 6)-l-12 having

INHERITANCE IX MAIZE.

about the same proportion, and of No. (10 x 6)-l-9 and
No. (10 x 6) -1-14, having a greater proportion of corneous
starch than the dent parent.

TABLE 11.

Record of Self-Fertilized Ears of F1 and F2 Generation of
Cross Between No. 10 and No. 6.

Ear Xumber

Corneous Seeds

Floury Seeds

/-in -v R\ K

104

(10 X o) -Z

159

226

(10 x 6)-l-l

123

a ti e\
-d

157

152

it a Or*

-63.

199

145

-4a

307

250

U «_g

242

208

" "-6a

149

" "_7

199

182

" "-8

124

" u-9a

259

202

" "-10

226

196

" "-11

212

209

" "-12a

182

106

" "-13a

107

108

-14a

" "-15

» 85 ■

"-16

hybrid ear, immature

"-6

considerable corneous starch in all seeds

" "-9

all seeds very corneous

"-10

some varibility, no seeds as Xo. 10 (varia-

tion probably due to immaturity)

"-11

pure corneous

"-12

pure corneous

"-14

all seeds very corneous

"-15

pure corneous

"-3

as Xo. 10 >

proved pure

"_4

probably as Xo. 10 )'

floury in 1912

"-16

"-17

as Xo. 10

"-18

"-19

"-13

"-20

"-21

Probably as Xo. 10

"-22

"-23

"-24

Total in hybrid ears

2863

2393

16 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

TABLE 12.

F2 Ears Obtained From Growing Ear (10 x 6)-l.

Parent Stock

Pure
corneous

Hybrids

Pure Floury

[Hand pollinated ears]

Dark yellow seeds of (10 x

6)-l

Light yellow seeds of (10 x

6)-l

W hite' seeds of (10 x 6)-l

Total

All ears obtained from ear No. (10 x 6) -1-9 selfed had small
seeds with traces of dent. On some ears there were merely
traces of dent, but other ears showed, the dented condition in all
seeds. Selections were made to determine whether these varia-
tions were inherited. In 1914 a self-fertilized ear which bore
seeds with only a few traces of dent was grown, also an ear with
all seed dented. The progeny of these ears is given in Table 13.

TABLE 13.

The Progeny of Ears Xo. ( 10 x (i ) -1-9-1 and ( 10 x 6)-l-9-2.

Progeny Classes.

Condition of Parent Ear

% seeds

V2 seeds

Few seeds

No seeds

dented

Few seeds dented

% seeds dented

These results show that little progress was made by the se-
lection.

Of the self-fertilized ears obtained from ear Xo. ( 10 x 0' )
-1-14, one showed no trace of dent, all of the seeds containing a
large proportion of corneous starch. This ear was grown and
compared with another self-fertilized ear which showed traces of
dent in nearly all seeds. The results are given in Table 14.

[NHERITANCE IX MAIZE, CONCLUSIONS. LI

TABLE 14.

The Progeny of Ears No. (10 x 6)-l-14-l and (10 x 6)-l-14-2.

Progeny Classes

Condition of Parent Ear

V2 Seeds dented

Few seeds
dented

Xo seeds dented

Xo seeds dented
Half seeds dented

4
11

27
19

In this case there seems to be some effect of selection, al-
though the number of individuals grown is not very large.

Conclusions.

There seems to be a close agreement between the results of
the cross between 10 and 6 and those reported for the cross be-
tween 10 and 5. It was, however, more difficult to classify the
seeds in the (10 x 6) cross as in No. 6 corneous starch is pro-
duced only on the sides of the seed, the cap and the immediate
vicinity of the embryo being filled with soft starch.

The essential difference between No. 10 and No. 6 in type
of starch produced is evidently one factor, yet since different F3
families showed variations in the amount of corneous starch pro-
duced, there must be several minor factors which modify its
development. There is good evidence that at least some of these
minor factors are factors which have a direct effect on totally
different tissues. For example, the size and shape of the seed
which is at least partly controlled by the type of pericarp (a
maternal character) has considerable influence upon the appear-
ance of the starch. To put the matter roughly, in plants which
fundamentally have the same zygotic possibilities as regards the
type of starch in the endosperm, the amount of soft starch ac-
tually developed is directly proportional to the size of the seed.

Family (10 x 64), Floury x Rice Pop (Very Corneous. |

The No. 10 parent had been self-fertilized for three years
and the No. 64 parent had been self-fertilized for two years
prior to 1909 when the cross was made. There was no visible
effect of the pollen of No. 64 on No. 10. F1 ears were grown in

IS CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

1910, but in no case was there a clear segregation among the
seeds like that occurring in the Ft ears of crosses (10 x 5j and
(10 x 6). This may have been due to the fact that the ears were
somewhat immature. The seeds of three F1 ears were separated
into two classes ; first, seeds as floury as Xo. 10 ; second, all re-
maining seeds. These partially corneous seeds showed a range
of variation from very corneous seeds to those which contained
only a little more corneous matter than the Xo. 10 flour parent.
The result of this classification is shown in Table 15.

TABLE 15.
F2 Ears of Cross Between (10 x 64).

Ear

Xumber

Floury Seeds

Corneous Seeds

(10

x 64) -7

216

(10

x 64) -10

349

(10

x 64) -12

168

353

Total

343

918

The seeds of (10 x 64)-7 and (10 x 64)-10 were planted in
1911. Those which had been classed as of the floury type like
Xo. 10 were planted as (10 x 61)-TS and (10 x 64)-10S. The
remainder of the seeds of the same ears were planted as (10 x 61)
-TC and (10 x 61) -10C respectively. The results obtained from
a classification of the progeny of these ears are given in Table 16.

TABLE 16.

Ears Obtained From Planting (10 x 64) -7C and TS and
( 10 x 61) -10C and 10S.

Parent Type

Progeny Classes

□ Ih
g P

Definite
I lybrids

1 ntermediate
Corneous

Pure
Corneous

(10 x 64) -TS

(10 x 64)-10S

(10 x 64) -7C

(10 x 64)-10C

INHERITANCE IN MAIZE.

1!)

There is a similarity in the variability of the populations ob-
tained from the floury seeds of (10 x 64)-7S and (10 x 64 I-10S;
the progeny of the corneous seeds of (10 x 64) -10 and (10 x 64)
-7 also show about the same percentage of ears in the different
classes.

* Two self-fertilized F2 ears (10 x 64)-10S-5 and (10 x 64)
-10C-4 were classed as definite hybrids. The corneous seeds of
these ears gave a range of variation from purely corneous to
definitely hybrid ears, there being 3.3 times as many corneous,
intermediate, and definitely hybrid ears, as there were pure cor-
neous ears. The floury seeds of (10 x 64)-10S-5 and
(10 x 64)710C-4 produced 4.2 as many hybrid and intermediate
ears as pure floury ears. Thus these two F2 ears showed as
variable a progeny in F3 as had been found in F2.

Five self-fertilized F2 ears of the intermediate floury class
from the progeny of (10 x 64)-10S gave a total population of
165 ears ; of which 19 approached pure corneous but contained a
larger percentage of soft starch than the corneous parent, 12 ap-
proached the floury parent, and 134 were intermediate. Many
of these intermediate ears showed some variation among the seeds,
but no clear segregation.

F2 corneous ears, (10 x 64)-10C-9, (10 x 64)-7C-9, and
(10 x 64)-70l bred true for the corneous habit in F3. (10 x 64)
-10C-9 was grown in F4 and again bred true.

Pure floury ear (10 x 64)-7S-13 bred true in F3 and F4 for
the floury habit.

One self-pollinated intermediate F2 ear, (10 x 64)-7C-2
proved to be a hybrid and gave in F3 15 corneous ears, 32 definite-
ly hybrid ears showing clear segregation, and 18 intermediate cor-
neous ears which showed some variation. This is a 1 :2 :1 ratio.

Two F4 ears bred from the intermediate class, (10 x 64) -7C-
2-10 and (10 x 64)-7C-2-l, together produced 14 ears approach-
ing pure corneous, 68 intermediate variable ears and 4 approach-
ing pure floury. These ears are probably all intermediates, the
variation being due to maturity and possibly due to the effect of
other inherited factors. Of 3 other F2 ears classed as inter-
mediate, 2 gave intermediate progeny and 1 proved to be a definite
hybrid. Self-pollinated ears of selections (10 x 64)-7S-l and
(10 x 64)-7S-7 from the intermediate class were grown the fol-
lowing year. These results are given in Table 17.

20 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

TABLE 17.

Progeny of Ears No. (10 x 64)-7S-l and (10 x 64)-7S-7
Which Were Classed As Intermediate Variable

Ears.

Classification of Progeny

Ear Xo.

Parent type

■ "~ IT.

u i:

hi in

u ; —

f 3 1

C_ —

(10 x 6-0-7S-1-2

Most corneous ear

(10 x 64)-7S-l-6

" floury ear

•t

(10 x 64)-7S-7-10

Intermediate ear

* -2

««

-4

-8

$46

$ Of this population. 25 open field ears were very variable and
showed definite segregation. The self-fertilized ears were comparatively
uniform.

The data in Table IT show that intermediate variable ears
tend to give intermediate variable progeny. The ears did not all
become thoroughly mature, and this may be the explanation of
their variable endosperms. There is also the possibility that
other heterozygous factors may have influenced development in
such a way as to produce variation. ( East & Hayes 1911).

F3 ear (10 x 61:)-10C-l-6 produced intermediate and corneous
seeds in a ratio approaching 1 :1. The corneous seeds of this ear
gave a progeny of 28 purely corneous and 24 definitely hybrid
ears, while the intermediate seeds gave a progeny of 2 corneous
ears. IT definite hybrids and 23 intermediate variable ears. This
is a close approximation of a 1 :2 :1 ratio. That only 1 factor de-
termined whether corneous or intermediate seeds were to be pro-
duced in this ear is further indicated by the separation of seeds
from five self-pollinated ears which were classed as definite
hybrids. The results are given in Table 18. The total number
of corneous seeds in these five ears were 514: and of intermediate
seeds 491. This clearly approaches a 1 to 1 ratio.

INHERITANCE IN MAIZE, SUMMARY.

TABLE 18.

Classification of Seeds of Hybrid Ears Obtained From
Planting Intermediate and Corneous Seeds of Ear
(10 x 64)-10C-l-6.

Ear Number

Corneous Seeds

Intermediate Seeds

i 10 x 6-P-1CC-1-6I-2

101

-9

. 92

-8

(10 x 14)-10C-l-6C-8

135

124

-10

120

124

Total in hybrid ears

514

491

Summary and Interpretation of Results.

The pollen of Xo. 64. pop apparently had no effect on the
character of the endosperm of Xo. 10 flour. This is in agree-
ment with the results of the crosses (10 x 5) and (10 x 6). The
Fx ears showed the results of segregation, although in this case
there was a range of Yariation from the floury to the corneous
type. Seeds of this Fx generation (F2 seeds) produced a popu-
lation of ears ranging from the pure corneous to the pure floury
type.

One uniformly floury ear bred true in F3 and F4 for the
floury habit ; three ears with purely corneous seeds also bred true.

Two F2 ears (10 x 64)-10S-5 and (10 x 64)-10C-4: gaYe as
Yariable an F3 progeny as had been found in F2. The ratio in this
case was approximately 1 pure corneous ear to 6.2 intermediates
and definite hybrids to 0.8 pure floury ears.

Other F2 ears gaYe a 1 :2 :1 ratio in F3 as was the case in the
(10 x 5) and (10 x 6) crosses. An example of such a ratio is
that obtained from F2 ear (10 x 64)-7C-2, which produced 15
corneous ears, 32 definitely hybrid ears and 18 intermediate ears.

SeYeral self-fertilized intermediate F2 ears bred comparatiYe-
ly uniformly, giYing a progeny which contained more corneous
starch than the Xo. 10 parent but less than the No. 64 parent.
Thus intermediate ear ( 10 x 64J-TS-1 produced ±1 ears of the
intermediate type none being either purely corneous, definitely
hybrids, or clearly floury. A self-fertilized ear ( 10 x 64)-7S-l-2
which contained more corneous starch than other self-fertilized

22 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

ears, yielded a progeny of 35 variable intermediate ears and 2
ears approaching the corneous condition although they were not
truly corneous ears like No. 64. Self-fertilized ear ( 10 x 64) -TS-
1-6 which approached the floury type, produced 40 intermediate
variable ears and 1 ear with somewhat more floury matter, though
it did not compare with No. 10. Thus in a total of 119 ears from
this intermediate line (10 x 64)-7S-l there were no pure cor-
neous, pure floury or definitely hybrid ears. This variation may
largely be due to differences in the maturity of the seeds and
ears, as the amount of corneous starch is directly dependent on
the maturity of the seeds, although of course the hereditary con-
stitution determines the amount which can be produced under
favorable conditions, but there is also considerable likelihood that
what one may call minor inherited factors modify the expression
of the character. Whether more than one major factor affecting
the endosperm is involved is still a question. The ratio obtained
among the progeny of ears (10 x 64)-10S-5 and (10 x 64)-10C-4,
the facts that certain F2 ears produced an F3 progeny similar to
the 10 x 5 cross, and that others bred approximately true to the
intermediate, the pure floury, or the pure corneous types might
seem to indicate two such factors, but analysis is so difficult that
this is only a reasonable guess, as will be shown by a considera-
tion of all of the facts.

The following conclusions we hold to be justified by the
data at hand.

1. The factors directly responsible for the differences in the
physical condition of the starch exhibited by the so-called starchy
sub-species of maize, the flour, dent, flint and pop corns are as
truly endospermal in their inheritance as endosperm color char-
acters. They partake of the nature of the embryo and not of the
plant on which they are borne.

2. These characters appear superficially to be maternal for
the following reasons. The endosperm nuclei are triploid due
to the fusion of two nuclei from the female gametophyte with
one nucleus from the male gametophyte. In the characters under
discussion, the presence of two factors always dominates the
presence of one factor. Thus corneous female (CC) x floury
male (F) is phenotypically corneous, while floury female (FF) x
corneous male (C) is phenotypically floury. These characters.

INHERITANCE JN MAIZE.

therefore, appear to be inherited in a different manner from
endosperm colors where the presence of one color factor is suf-
ficient to cause perfect development of color. This is the first
proof of a cumulative somatic effect of factors.

3. From the fact that in these crosses, as well as in num-
erous others involving the same subspecies of maize that we have
examined, the F2 reproduces the grandparental and no types more
extreme than the grandparental types (with possibly a rare ex-
ception), it follows that a large series of multiple allelomorphs
affecting the starchy condition of the endosperm exists.

4. From the facts (a) that where no complications such as
differences in shape and size of seed exist (viz. cross 10 x 5)
segregation is simple and definite, (b) that where such differences
in shape and size of seed do exist segregation occurs but is diffi-
cult to demonstrate clearly until these complications have been
eliminated, it follows that although only the presence of factors
in the endosperm affect these characters directly, the maternal
zygotic constitution has an indirect effect. This effect is roughly
a direct correlation of size of seed with floury condition of the
endosperm.

Having these facts in mind, let us see what difficulties ob-
struct analysis if it be assumed that two factor differences may
differentiate the endosperms of certain maize varieties in respect
to starch as seemed possible in the case of cross (10 x 64).

The simplest assumption would be that each of these factors
has a similar effect, and when one sees the difficulties thus in-
volved, and considers that such a simple assumption is less prob-
able than one in which each factor has a different effect, it is
clear why we do not wish to assert dogmatically that two such
factors are involved in the cross between the flour and the pop-
corn.

Let the flour corn be AABB and the pop corn aabb, it being
understood that the phenomenon of dominance is in this case
wholly a quantitative reaction. The Fx generation in the cross
and its reciprocal would be

AAa BBb

and
aaA bbB

"24 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

In each case, the predominant influence of the mother would be
such that any effect of the father would scarcely be noticeable.
Four types of gametes would be formed in the Fx generation as
usual, AB, Ab, aB and ab, — but the appearance and breeding
qualities of the zygotes formed would be peculiar, as is shown in
the following table, due to the fact that the "gametes" of the
embryo sac are the fusion cells AABB, AAbb, aaBB and aabb.

Appear alike breed differently

The grandparental types have appeared of course and will
breed true, but other individuals will look like the grandparents
though they will breed differently and will ultimately give the
whole series if crossed together. Other complications will occur
to any one who takes the trouble to study the table.

Family (65 x 64), White Pearl Pop x White Rice Pop.

In 1910 a cross was made, between white rice pop No. 64 and
pearl pop No. 65 for the dual purpose of determining the probable
value of such a cross for the commercial production of first gen-
eration hybrid pop corn, and to study the inheritance of the
pointed seed characteristic of the rice pop corns.

The Fj plants were considerably more vigorous than either
parent. The seeds produced approached the length of those of
the longer type, the white rice pop, and the width of those of the

1 AAABBB 1

1 AAABBb

1 AAaBBB
1 AAaBBb J

1 AAAbbB 1

1 AAAbbb

1 AAabbB

1 AAabbb J

1 aaABBB 1

1 aaABBb

1 aaaBBB

1 aaaBBb J

1 aaAbbB 1

1 aaAbbb

1 aaabbB

1 aaabbb

[NHERITANCE IX MAIZE

broader parent, the pearl pop. Thus the F2 seeds (those borne on
F1 plants) were considerably larger than those of either parent,
and since the pericarp was weaker rather than stronger than that
of the pure types, they did not pop as well.

TABLE 19.

Inheritance of Seed Shape in a Cross Between Wtiite Rice
Pop Xo. 64 and Pearl Pop Xo. 65.

Condition of

Progeny

Ratio of Pointed
ears to intermedi-
ate and noil point

Ear Number

Parent
Type

Pure
Point

Inter-
mediate
Point

Non
Point

64-4

Pure pt.

147

65-8

Xon pt.

200

(65 x 64) Fi

132

(65 x 64) -1 F2

Int. pt.

1 :10.S

" -3 "

1:10.8

-o

1 :6.5

" -6 "

1 :15

" -1-13 F3

Int. or non
pt.

*21

* Possibly non-point as the point was scarcely perceptible.

The data on the cross are given in Table 19. The F l gen-
eration was of intermediate habit, — there being some projection
of the seeds at the point of attachment of the silk. Four selfed
F1 ears furnished F2 generations. The progeny of these ears was
variable, the seeds of some ears being as completely pointed as
the white rice pop parent, the seeds of others non-pointed like
the pearl pop parent, while the greater number were of various
intermediate types. Of a total progeny of 263 individuals, 2±
ears were classed as pure pointed like the white rice parent.
This is an indication of a 15 :1 ratio, although one can not be cer-
tain that the classification was correct because these ears were
not selfed and could not be tested by the type of progeny produc-
ed. A number of F2 ears were self-pollinated, but none happened
to be obtained which could be classed as typically pointed. One

2() CONNECTICUT EXPERIMENT STATION. BULLETIN 188.

ear having seeds but slightly pointed (possibly non-pointed) was
grown in F3. The twenty-one ears produced were like the parent
ear. showing only slight projections on the seeds at the tip of the
ear.

The difference between the pointed seed characteristic of the
white rice pop corn and the normal shape of seed typical of other
varieties can not be explained by a single factor. If. however, we
assume that there is a difference in two factors, that each factor
is allelomorphic to its own absence and is inherited independently
of the other, that both are necessary for the production of the
pure pointed condition, and that either of them alone may pro-
duce a tendency to a pointed condition (intermediate point), the
data accord fairly well with the theory. But since on this hypo-
thesis it is assumed that a factor in the heterozygous condition,
produces only half as great an effect as when homozygous, one
can appreciate the difficulty of classifying the ears correctly by in-
spection, and since classification must be exact to prove such a
case merely by the ratios obtained it must be admitted that our
evidence is open to some criticism. On the other hand, we be-
lieve that the facts are clear enough to make them of some value
in practical plant breeding, and we do not believe that the case is
sufficiently important to make it worth while overcoming the
difficulties that stand in the way of a more acceptable proof.
Furthermore, the data on the next cross appear to corroborate our
earlier facts.

Family (64 x 6), White Rice Pop x Leaming Dent.

This cross was made in 1909 between self-bred Leaming and
white rice pop strains. The purpose of this cross was a further
study of the mode of inheritance of quantitative differences in
seed size, of the proportion of corneous to soft starch, and of the
pointed habit of the white rice pop.

The results on inheritance of seed shape are given in Table
20. These results again indicate that two factors are involved.
Furthermore, examination of Table 20 and Table 21, shows that
the pointed character is inherited independently of the position of
starch in the seeds.

INHERITANCE IN MAIZE.

TABLE 20.

Inheritance of Seed Shape in a Cross Between No. 6
Leaming Dent and No. 64 White Rice Pop.

Condition of Progeny

Ear Number

Parent
Type

Pure

Inter-

Non

Point

mediate
Point

Point

64-4

Pi i v t

r uic ±> t.

147

6-3-4

I ^ pii t tirin nt

107

6 x 64

112

(6 x 64) -4

Int pt.

-6

-6-6

-6-3

-4-8

Pure ( :>) nt

-4-9

-6-4

Int pt.

-4-6

-4-4

-6-5

Xon pt.

4 (?)

-6-7

-4-3

-4-7

-4-10

-4-5

(6 x 64) -6-6-4 F4

Pure pt.

-6-6-1 "

Pure or int. pt. ( ? )

27a

-4-8-8

Pure pt.

61b

1 (?)

-6-3-6 "

-6-5-4

Int. pt.

1 (?)

-6-5-3

1 (?)

-6-7-8

-4-8-3

-4-3-7

-4-10-5 F4

-4-10-3 F4

Xon or int.

-4-3-5

Non

a — 4 ears with points not as strongly developed as the remaining ears,
b — 1 ear with points not as strongly developed as the remaining ears.

The F1 generation was intermediate as regards the pointed
condition, and there was segregation into pointed, non-pointed and
intermediate ears in F2. Thirteen self-pollinated F2 ears were
grown in F3. Of these, the following F2 ears were classed as
pure pointed, (6 x 64)-6-6, (6 x 64)-6-3, (6 x 64)--T-8, (6 x 64)
-4-9. Two of these ears, (6 x 64)-6-6 and (6 x 64)-6-3, bred
true in F3, while (6 x 64)-4-8 and (6 x 64)-4-9 showed segrega-

28 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

tion in F3 with a total of 85 pointed and 31 intermediate pointed
ears. Two self-fertilized ears, (6 x 64) -4-8-8 and (6 x 64)
-4-8-3, were grown in 1914. One proved to be a pure pointed ear
and the other again gave pure pointed and intermediate pointed
seeds. These results might have been obtained if ear (6 x 64)
-4-8 were homozygous for one factor for point and heterozygous
for a second factor.

Three self-fertilized F2 ears of the intermediate class showed
a range of variation in F3 from pure pointed to non-pointed ears.
Six F2 ears classed as non-pointed were proved to have been
hybrids by the F3 results. One of these, (6 x 64) -4-7, produced
52 intermediate and 13 non-pointed ears. As no typically pointed
ears were obtained it seems fair to conclude that the parent ear
(6 x 64) -4-7 was heterozygous for 1 factor for pointed seeds.

Two self -fertilized F3 ears of line (6 x 64) -6-6 which bred
true for the pointed habit in F3 were grown in F4. Ear (6 x 64)
-6-6-4 gave a progeny of 35 ears, all of which were pure pointed :
while (6 x 64) -6-6-1 had a progeny of 23 pure pointed ears and 4
with points more strongly developed than the intermediate class,
but not so strongly developed as the 23 pure pointed ears. This
may be a physiological variation or it may possibly be due to
chance pollination. As these four were open field ears, it is im-
possible to determine the matter by further breeding.

The results are an excellent illustration of the old Vilmorin
Isolation Principle, — in modern times the genotype hypothesis, —
for they show that the only sure method to determine the breeding
value of an ear is to grow and examine its progeny. A part of
the pure pointed class gave a pure pointed progeny ; other ears
proved to be hybrids. There was also considerable difference in
the progeny of different intermediate ears ; some being apparently
homozygous for one factor for point and heterozygous for
another, while others appeared to be heterozygous for a single
factor.

These results, as did those in the case of the (65 x 64) cross,
indicate that two factors are involved in the production of strong-
ly pointed maize seeds.

Table 21 gives the results of a study of the dented condi-
tion and the proportion of corneous to floury starch in the same

INHERITANCE IX MAIZE.

2{J

cross. The white rice pop parent contains only a small amount
of floury starch, while the dent variety has corneous starch at the
sides of the seed and floury starch at the cap and next the em-
bryo. There was no effect on the development of the amount of
corneous starch in No. 6 dent due to the pollen from No. 64 pop.
The Fj generation cross produced ears with intermediate sized
seeds. These ears would have to be classed as dents.

TABLE 21.

Inheritance of Dented Habit and Proportion of Corneous
to Floury Starch.

Condition

of Progeny

dumber

Parent Type

in
—

(LI

to
-c

-z

Nearly
pure de

Half se
dented

£ c

— T3

J s

pop, non-dent

107

dent

147

112

-4 F2

nearly pure dent

t( « <<

-6 "

-4-4 F3

pure dent

-4-5 "

-4-8 "

a u

-6-6 "

u a

-6-5 "

a <«

-6-7 "

nearly pure dent

<< a tt

-6-4 "

-4-7 "

half seeds dented

-4-9 "

-4-6 "

few seeds slightly dented

« « «< «

-6-3 "

-4-10 "

non-dented

-4-3 "

it *(

-6-3-6 F4

pure dent

-6-5-3 "

-6-5-4 "

« u

-6-6-1 "

-6-6-4 "

-6-7-8 "

half seeds dent

-4-8-8 "

half seeds slightly dent

-4-8-3 "

seeds slightly dent

-4-3-7 "

few seeds dent

-4-3-5 "

no seeds dent

-4-10-5 F4

few traces of dent

-4-10-3 "

non-dent

64-4
6-3-4
6 x 64

30 CONNECTICUT EXPERIMENT STATION, BULLETIN 188.

Two Fj ears (6 x 64) -4 and (6 x 64) -6 were grown in F2.
Both populations showed a wide range of variation. The ears
were classed as pure dent, nearly pure dent, half seeds dent, few
seeds dent and non-dent. Ear (6 x 64) -4 had progeny of each
class, while (6 x 64) -6 produced progeny in all classes except the
non-dent class. Thirteen F2 ears were grown in Fa. Two non-
dented ears gave a progeny of non-dented ears and ears with a
few seeds slightly dented. Xo ears bred true in F, or F4 for the
pure dented condition, although some selections gave a progeny
with a much larger proportion of dented ears than others.

Twelve F3 ears were grown in F4. Ear ( 6 x 64) -6-5-3 pro-
duced the greater proportion of its progeny in the pure dent class.
Ear (6 x 64) -4-10-3 bred true to the non-dented character, and
the corneous non-pointed condition. Ear (6 x 64) -6-5-3 bore
seeds which approached the size of those of the Xo. 6 Learning
parent, although the range of variation was somewhat greater.
Ear (6 x 64) -4-8-3 gave a uniform progeny in 1914. and bred
comparatively true to the seed size of the pop parent.

The seeds of those ears which were classed as non-dents and
those with a few seeds dented, popped perfectly when tested.
The condition of the other families is shown in the table.

Summary and Interpretation of Results.

The data from these two crosses indicate strongly that two
independently inherited factors are necessary for the production
of a strongly pointed seed. The rice pop point can be transferred
from the pop parent to dented seeds by crossing and selection ;
the inheritance of these characters being entirely independent of
each other.

A study of the proportionate amount of corneous and floury
starch in the 6 x 64 cross shows a wide variation in F2. One ear
(6 x 64) -4-10-3 bred true for about the same amount of corneous
starch in F4 as that of the X'o. 64 parent. Other ears were again
as variable as F2, while still others showed a smaller range of
variability. It is impossible to state how many factors are in-
volved in producing these somatic differences, but it is a fact that
the parental types can be recovered easily and will breed true.

INHERITANCE IN MAIZE, CONCLUSION.

Conclusion.

Since a summary of the results obtained for each cross has
been given in its proper place, it seems unnecessary to repeat
them here. If the reader will refer to them he will find an ab-
stract of the paper.

Literature Cited.

CORRENS, C.

1901. Bastarde zwischen Mahrassen mit besonderer Be-
riicksichtigung der Xenien. Bibliotheca Botanica.
53:1-161.

EAST, E. M. and HAYES, H. K.

1911. Inheritance in Maize. Connecticut Expt. Sta. Bull.
167:1-142.

EMERSON, R. A. and EAST, E. M.

1913. The Inheritance of Certain Quantitative Characters
in Maize. Nebraska Station Research Bull. 3 :1-120.

SHUEL, GEORGE HARRISON.

1914. Duplicate Genes for Capsule-form in Bursa bursa-
pastoris. Zeitschrift fur induktive Abstammungs-und
Vererbungslehre XII: 97-149.

PLATE I.

Xo. 5, corneous flint at bottom, Xo. 10 flour at top and Fi at left.
The two lower center ears show the result of planting corneous Fi
seeds and the two upper center ears show the result of planting
floury Fi seed.

PLATE II.

a. Xo. 10 flour at left. Xo. 6, Learning dent at right. The four other ears
represent the F3 generation of cross. They are uniformly very corneous with
slight traces of dent. The seeds are smaller than those of either parent and
of uniform size.

b. Average ears of Xo. 65 pearl pop at left, Xo. 64 rice pop at right with
average Fi in center. The two remaining ears represent the extremes of F2.

PLATE ill.

b. F2 generation of cross (6 x 64). Xote the segregation of characters.

PLATE IV.

(6XtY)-y-/0

Upper row, F3 generation ears, with large amount of corneous
starch. Some ears with slight trace of dent. Middle row, average
progeny of F2 ear which bore good sized dented seeds. Lower row,
average progeny of F2 ear which bore intermediate dented seeds
with a well-developed point.

PLATE V.

£>4

((>*MM-lO-3

(6*64)-(>-$-3

Average ears of parental types Xo. 6, Learning dent and No. 64, white
rice pop above. The ears below represent the variation in 4 F4 families.
(6 x 64) -6-6-4 bred true for the rice point, (6 x 64) -4-10-3 bred true for
corneous, non-dented seeds, (6 x 64) -4-8-8 is a small-seeded selection and
(6 x 64) -6-5-3 is a large-seeded selection. (Photo by Walden.)

PLATE VI.

(/0X6</)-7S-/3

(/0X6V)-/0S-/

(/OX69)-/QC-9

Upper row, F3 generation of cross between No. 10 flour and Xo.
64, rice pop, which bred true for the floury habit.

Middle row, F3 generation of same cross which bore seeds of
intermediate type.

Lower row, F3 generation of same cross which bred true for the
corneous habit.

PLATE VII.

3 ^'

(/OX6VJ-/OS-6

a. Fa generation of cross between Xo. 10 and Xo. 64 which bred true for
the seed size of Xo. 10.

00XM)-7C-2

b. Fz generation of cross between Xo. 10 and X"o. 64 which bred true for
the seed size of Xo. 64. The corneous seeds popped perfectly.

STUDIES OX SIZE INHERITANCE IN NICOTIANA

E. M. East

Harvard University, Btasey Institution, Forest Hilist Massachusetts

Reprinted from GEXETICS 1:164-176, Mar

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board
William E. Castle

Harvard University

Edwin G. Conklin Rollins A. Emerson

Princeton University Cornell University

Charles B. Davenport Herbert S. Jennings

Carnegie Institution of Washington Johns Hopkins University

Bradley M. Davis Thomas H. Morgan

University of Pennsylvania Columbia University

Edward M. East Raymond Pearl

Harvard University Maine Agricultural Experiment Station

George H. Shull, Managing Editor
Princeton University

Genetics is a bi-monthly journal issued in annuai volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum. Canadian subscribers
should add 25 cents for postage. To all other countries 50 cents should be
added for postage. Single copies will be sent to any address postpaid
for $1.25 each.

All subscriptions, notices of change of address, and business corre-
spondence should be sent to the Princeton University Press, Princeton,
New Jersey, and all remittances should be made payable to the Princeton
University Press.

Entered as second-class matter February 23, 191 6 at the post office at
Princeton, N. J., under the act of March 3, 1879.

STUDIES ON SIZE INHERITANCE IN NICOTIANA

E. M. East

Harvard University, Bussey Institution, Fjrest Hills, Massachusetts

STUDIES ON SIZE INHERITANCE IN NICOTIAN A

E. M. EAST

Harvard University, Busscy Institution. Forest Hills, Massachusetts
[Received January 6, 1916]

As various writers have pointed out, all Mendelizing characters prob-
ably are due to the interaction of several genes, and presumably every
gene may exhibit several somatic effects, yet no one doubts that the
Mendelian notation describes the inheritance of such things as color
accurately and concisely. It is strange, therefore, that some geneticists
still refuse to believe that the inheritance of size characters can be de-
scribed in the same way, without further assumptions.

Various reasons are assigned for this disbelief.

It is held that one should not assume the absence of dominance, as
has been done by those who have investigated size characters. But as
a matter of fact, absolute dominance is rare. A heterozygous gene very
seldom produces an effect identical with that of homozygous genes.
Full dominance is the extreme, the limiting condition, not the common
condition. Even with such simple and possibly superficial characters as
colors, careful examination usually shows incomplete dominance.

A further misconception of the phenomenon of dominance is the ob-
jection to the assumption of genes having cumulative effects. As stated
in the first sentence, most Mendelizing characters have been shown to
be due to the interaction of several traceable factors, in addition to an
ever present factorial residue of which nothing is known. This unex-
plored ground may be reduced in its extent by new mutations affecting
the character in question, but proof can never be offered that it has been
entirely eliminated. For the same reasons it follows that one should not
assume that the simplicity of the known facts proves actual simplicity in
the hereditary transmission of any character. This complexity in the
germinal basis of characters, is, of course, general proof of the cumula-
tion effect of genes, but in addition a specific case has recently been found
in maize (Hayes and East, 1015). When reciprocal crosses of "floury"
and "flinty" maize races are made, the maternal erfdosperm character is
dominant. This dominance has been proved to be due to the fact that the

SIZE INHERITANCE IN NICOTIAN A 165

endosperm is produced by the union of two maternal and one paternal
nuclei. Thus two genes of either kind dominate the effect of one.

The only other criticisms worthy of notice are directed against as-
sumptions of gametic purity and of factorial constancy. As criticisms
of the Mendelian interpretation of quantitative characters they come
no nearer the mark, for they apply to Mendelism as a whole.1

The true reason for objecting to the theory, therefore, seems to be —
as is often the case — that those who disapprove of it have not given it
sufficient study to be convinced that any real evidence in its favor can be
cited. For example, Castle (1914) says:

"When races are crossed that differ widely in size, the first filial (Fx)
generation is intermediate between the parents and often not more variable
than one of the parent races. But the second filial (F2) generation, though
still intermediate, commonly shows increased variability, the range of which
may even extend into or include the size range of one or both parent races.
This increased variability of the F2 generation is the only evidence of
Mendelism in size crosses."

With this view I cannot agree. It is true that one may not expect
dimorphic phenotypes in simple ratios in the F2 generation. Somatic
appearance is not so highly correlated with genetic constitution that defi-
nite ratios always appear when characters like color are studied. Even
in such cases one must prove the classification of the phenotypes by
further breeding. By carefully studying what actually occurs in simple
and obvious Mendelian phenomena, however, the mathematical require-
ments where size characters are involved can be worked out. If these
requirements are independent, — i.e., if they are not restatements of the
same conditions, — and if the breeding facts meet them fairly and
squarely, the case is good.

At least eight such requirements, most of which are independent
mathematically, should be met by the pedigree-culture data when all
populations succeeding the original cross are obtained by self-fertili-
zation.

1. Crosses between individuals belonging to races which from long-
continued self-fertilization or other close inbreeding approach a homozy-
gous condition, should give F1 populations comparable to the parental
races in uniformity.

2. In all cases where the parent individuals may reasonably be pre-
sumed to approach complete homozygosis, F2 frequency distributions
arising from extreme variants of the F1 population, should be practically

1 The question of the validity of these criticisms when directed against the entire
Mendelian theory, is not under discussion.

Genetics 1: Mh 1916

E. M. EAST

identical, since in this case all Fx variation should be due to external
conditions.

3. The variability of the F2 population from such crosses should be
much greater than that of the Fx population.

4. When a sufficient number of F2 individuals are available, the
grandparental types should be recovered.

5. In certain cases individuals should be produced in F2 that show
a more extreme deviation than is found in the frequency distribution of

Figure i. At left, young plant of Nicotiana longiflora var. (383) ; at right, young
plant of N. longiflora (330).

either grandparent. This phenomenon was predicted by the writer
(East, 19 10) as an expected result of Mendelian recombination before
actual cases had been discovered.

6. Individuals from various points on the frequency curve of an F2
population, should give F3 populations differing markedly in their modes
and means.

SIZE INHERITANCE IX NICOTIAN A ^7

7. Individuals either from the same or from different points on the
frequency curve of an F2 population should give F3 populations of
diverse variabilities extending from that of the original parents to that
of the F2 generation.

8. In generations succeeding the F2, the variability of any family
may be less but never greater than the variability of the population from
which it came.

Not all of these eight conditions are met by the data to be presented

Figure 2. Average flowers of parents with an average flower of the F2 generation
(383 X 33o) in the center.

in this paper, but all of them have been met many times in the course of
other experiments, and not one fact has been discovered directly op-
posed to them.

Genetics 1: Mh 1916

E. M. EAST

The data to be considered here were obtained by measuring the.
length of the corolla in a cross between two varieties of Nicotiana
longiflora Cav. The seed of Xo. 330, which is probably the type, was
obtained from Prof. W. A. Setchell (see Setchell 1912, pp. 21-22).
The seed of Xo. 383 was received from the Instituto Sperimextale
per le Cultivazioxi dei Tabacchi at Scafati, Italy, through the kind-
ness of Dr. A. Splexdore. It was known there as N. plumbaginifolia
Viv., but seems to be merely a small variety of N. longiflora.

Each corolla length recorded is expected to represent the phenotype
of a single plant. The method of recording them and the accuracy that

Figure 3. Average flowers of parents at A (330) and D (383) ; extreme segregates of
the F2 generation at B and C.

8 111

SIZE INHERITANCE IX NICOTIANA

I - I I \ [ I I 1 I I I M

169

I I I I I I I

O 00 <N

00 <n co

I M I

I I I I I

s 1

IMS

1 1

s 1

III?

1 1

l-H

a l

IMS

1 1

CO 00

§ l

1 1

10 1
00 1

1 1 1 1

1 M

O
to

i-h

£ 1

II II

1 M

I>-

R 1

1 II 1

1 *

CN
CO

O
"3-

£ 1

1 "S

I -

Ov
M

Ov
CO

£ 1

II II

1 S1

Ov
CO

-J

rf vo

[

a 1

II II

CO co

O
IO

co vO

CO M

'-O

1/1

£ I

II II

O <N

-1-

S 1

1 1 II

to 00
m

a 1

vO 1

! 1 I 1
1 1 1 1

HI CO

-1-

-f

1
1

00 1
10 1

1 1 1 1
1 1 1 1

■

O vO

CN
CN

Tj"

1 -

1
1

—

LO I

10 1

^t"

r^.
CO

to
to

0* 1

LO 1

1/-.

LO
LO

s 1

II II

tO
10

LO
CN

-r

•# 1

CN
—

? 1

<N O i-i 1

CO HH |

CO
cn

I I

I I I

I I

II II

I I

GC
t

W l-H CN O

vo O r-^

O O

LO o

r>- 1-1 O O cn

00 00 to 00

co to o.
Tt GO 00

II I I I

m CN com CN com CN CN cococococococococo^"^-^J"^J-toto
Ov 0s Ov Ov 0s O Ov Ov 0 Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov O' Ov Ov Ov

e
.2

CN CO^M CO LO O

I I I I I I I I

MMMCNCNCNCNCN

I I

HI hH CN

I I I

CO vO vo co vO

I I I I I

M CN CN M CN

^OOOOOOOOOOOOOOOOO

O cococococococococococococococococo
cOcococococococococococococococococo

^XXXXXXXXXXXXXXXXX

_ „ _ cococococococococococococococococo

cococpO O O coooooooooooooooooooooooooooooooocoo

OC 00 GO co CO co GO cococococococococococococococococo

CO CO co co co co CO ^— ' ^— ' v~-x w w w -w w w v^-

OOOOOOOOOOOOOOOOOOOOOOOO

Genetics 1: Mh 1916

I/O

E. M. EAST

may be expected of the method have been discussed in another paper
(East 1916).

Both of the varieties used as parents may reasonably be supposed to
be homozygous in most of their characters for they are generally self-
pollinated naturally. Whether either or both of them had been self-
pollinated artificially before I obtained them is not known, but they had
been self-pollinated for two generations after I received them before
the cross was made. The crosses and the succeeding selfings gave full
capsules, and the germination of the seeds was almost perfect. .

As shown by the tables, if the frequency distributions of the pure
varieties for 191 3 are excluded on account of the small number of plants
grown, the average mean of No. 383 is 40.54 mm and of No. 330, 93.30
mm. The species has corollas over twice the length of the variety. The
average of both parents is 66.91 mm, and since the mean length of the
F-l generation is 63.53 mm, it is clear that the vigor induced by heterozy-
gosity, which was fairly well marked in the vegetative characters, had no
effect on the flowers. This fact is in keeping with previous observations,
since it has been shown that corolla length is very slightly influenced by
external conditions, and that heterozygosity effects a result comparable to
favorable external conditions.

The variability of the Fx population appears to be exactly the mean
of variety No. 383 (the more variable parent) for the two years 191 1
and 1 91 2, though considerably higher than the variability of variety No.
330 for the same period. Theoretically one ought to expect this only
when both parental varieties are completely homozygous, therefore the
data might be supposed to show such a condition. But, while the two
varieties used here probably approach a homozygous condition, the
similarity of the two constants obtained is possibly more apparent than
real. The measurements have been thrown into three-millimeter classes
for convenience, but these classes are manifestly too large for small flow-
ers like those of No. 383. If one-millimeter classes are used there is less
distortion of the figures and the percentage variability is smaller. For
this reason I believe that it is fair to conclude that the variability of the
- Fx population is slightly larger than that of either parent. The proponents
of the Mendelian theory may maintain that this merely shows a slight
degree of heterozygosity in the parents, therefore, while its opponents
may see in the results indication of a slight increase in variability due to
the cross itself. No one can object to this view when considered apart
from other facts, but it should be pointed out that the difference to be
accounted for is very small in either case.

SIZE INHERITANCE IN NICOTIANA

171

Table 2

Statistical constants of the frequency distributions shown in table 1.

Designation

No.
Ind.

Mean

S. D.
in mm

C. V.
in percent

JNo.

383-

-I9II

125

46 d=

•75±

4-33±

_ 0
18

XT~

JNo.

383-

-I9I2

6i±

OO db

4-92db

XT~

No.

383-

-I9I3

76±

09 db

2.74d=

JNo.

330— IQII

22d=

29 =b

I 2

2 .40d=

XT~

JNo.

330—1912

37±

23 ±

2-39±

JNO.

33O—I913

I2±

70db

2-93±

JNo.

(330

x 383)

1?
r 1

173

53 ±

92rb

4-OOdz

XT~

JNo.

(330

x 383) -1

T7
*2

211

Si±

91 ±

0 — - 1

8-75±

XT~

JNo.

(330

X 383) -2

XT'

233

78±.3o

79=b

9-73±

XT~

JNo.

(330

x 383) 1-1

170

I4±

82 ±

5. 22±

No.

(330

X 383) 1-2

143

47 ±

74 ±

6.99=b

No.

(330

X 383) i-3

147

20±

I7±

6-3i±

No.

(330

X 383) i-4

175

34±

07 dr

7. 22±

No.

(330

X 383) 2-1

IS9

04 ±

OOdr

6.85d=

No.

(330

x 383) 2-3

i43

34 ±

06 dz

6.63zb

No.

(330

X 383) 2-4

166

01 ±

85±

6-55±

No.

(330

x 383) 2-5

160

97±

04 dr

5-74±

No.

(330

x 383) 2-6

162

20dr

76±

5-93±

No.

(330

x 383) 1-2-1

184

7i±

37±

5.l8dr

No.

(330

x 383) 1-3-1

189

25±

87 dr. 06

4-04dr

No.

(330

X 383) 2-6-1

i95

25±

30dz

4.01 d=

No.

(330

X 383) 2-6-2

164

86 db

83 ±

7-°4d=

No.

(330

X 383) i-3-i-i

161

98 db

30 ±

5-49±

No.

(330

X 383) 2-6-2-1

125

88 ±

52db

6.28d=

Examination of the F2 frequencies shows that only one individual
reaches the lower size limit of No. 330 and that no individual comes
within two classes of the upper size limit of No. 383. Viewed from this
standpoint the results are less in accord with Mendelian theory than any
of those obtained in the numerous size studies I have made. At the
same time, one may say that this is because the numbers are too small
to expect an exact duplication of the grandparents in a species which
in all probability has 24 chromosomes in its germ-cells and in which
grandparental duplication should be expected only once in 265 million
millions of F2 individuals. The difficulty here would be not to account
for the non-appearance of the grandparental sizes in F2 populations of
about 200 individuals, but to conceive how extremes differing by 36
millimeters had arisen. One appears to have but a single alternative:
either the differences between types that give fertile F2 generations are
due to relatively few factors, the remaining germ-plasm being identical,

Genetics 1: Mh 1916

1J2

E. M. EAST

or the extremes recovered are not like the grandparents but merely re-
semble them.

From another standpoint the variability of the F1 and the F2 gener-
ations is very different. Theoretically if recombination is possible, the
number of classes between the extremes varies directly with the square
root of the number of individuals involved. The coefficients of vari-
ability of the two populations, however, should not change with larger
numbers except as regards the confidence to be placed in the calculated
constants. The coefficient of variability of the Fx generation is 4.60
±.17 percent and the coefficients of variability of the two F2 populations
grown are 8.75 ±.29 percent and 9.73 ±1.30 percent, respectively. Thus
jthe average variability in F._, is just double that of the Ft generation.

Figure 4. Average flower of Xo. 383 (A) compared with modal condition (C), and
with an extreme (B), of Fa family (383 X 330) 1-3; and, ditto Xo. 330 (F)
compared with modal condition (D), and with an extreme (E), of F3 family

SIZE INHERITANCE IX NICOTIANA I73

One can scarcely appreciate the significance of this immense difference
until he recalls that the difference between the means of the pure varieties
and the mean of the F2 generation is only about four times the standard
deviation of the latter, while the difference between the means of the
varieties and the mean of the F± generation is about nine times its
standard deviation.

Let us now examine the means of the populations that have resulted
from selling selected individuals of the F2, F3 and F4 generations. Whether
one can isolate rapidly lines with markedly different mean values after
a blend such as occurred in the F1 generation is a question of consider-
able practicable importance. Nine F3 populations were grown. The ex-
treme parental types from which they were grown were 46 mm and
82 mm, a difference of 36 mm. The greatest mean difference between
the F3 populations was 30 mm, although it happened that the difference
between the parents that produced these two lines was only 32 mm.

From the F3 generation two "short corolla" and two "long corolla"
lines were grown. In this generation it wTas possible to select extremes
a little farther apart, 43 mm to be exact. The difference between the
means of the mpst extreme resulting populations was 37 mm.

Only two F5 populations were grown, one from a plant with flowers
41 mm long and the other from a plant with flowers 90 mm long. The
families to which they gave rise had means of 42 and 88 mm in round
numbers. Thus a relatively small number of selections has given a type
averaging but two millimeters longer than the smaller parent, and a
type averaging less than five millimeters under that of the larger parent.
To attain these ends only twelve families from extreme parents were
grown. It is impossible to say just how many selections of F2 indi-
viduals would have had to have been made to reach the same goal on
the recombination theory, but one can estimate the probability of the
occurrence of individuals of the desired size in F2 from which to select.
Consider the F2 generation in which the standard deviation is 6.79 mm.
Assuming this distribution to be normal the expected frequency beyond
the distance from its mean represented by one-half the mean of Xo. 330
minus mean of Xo. 383 is .0619 percent. In other words, one might
expect an F2 individual with the size of the modal class of either grand-
parental variety about once in every 1600 plants.

These facts indicate clearly the proper procedure of the plant breeders
in such cases, as has already been brought out by Emersox and East
(1913). If it is technically possible to grow an F2 large enough to be
reasonably certain of obtaining several individuals with the desired com-

Genetics 1: Mh 1916

*74

E. M. EAST

bination, the breeder is tolerably sure of success. But the numbers are
often prohibitive in practice, and at the best the work involved is great.
On the other hand, though success is not so certain because the plants
with the gametic possibilities desired may be dropped out at any point,
selection continued for several generations gives a high probability of
success with comparatively little work.

A study of these means with reference to their bearing on Galton's
Law of Regression is also interesting.

In thirteen out of the fifteen fraternities descended from the two F2
populations there was regression towards the mean of the fraternity
from which the parent came; these two individuals, however, produced
populations with means further removed from the means of the parental
population than were the parent individuals themselves. Further, the
deviations of the parents from the mean of their fraternity show no
correlation with the deviation of the mean of the progeny from the
parental value. In other words, in selfed lines of this kind an ex-
treme variant is almost as likely to produce a type like itself as is a
slight variant. This is to be expected with the hypothesis of plural
segregating factors but not with the old Galtonian hypothesis in which
somatic resemblance is the sole measure of heredity. Our observation
is not new since Galtonian regression in the original sense is now en-
tirely discredited, but our data illustrate the point.

The remaining arguments are based upon the variabilities of the
fifteen fraternities whose means have just been considered.

In the first place, it is essential that one should know whether he may
expect to obtain fraternities that breed as true as the parental varieties
at once, after long continued selection, or not at all. These data do not
show fraternities comparable to either parent variety in variability among
the nine F3 families, but out of the four F4 families two show as narrow
a variability as Xo. 383.

A more important question, however, is that of continuous reduction
of the coefficient of variability due to the automatic tendency toward com-
plete homozygosis produced by continued self-fertilization. Theoreti-
cally, a fraternity produced by self-fertilization may be as variable as
the fraternity from which its parent came, but it can never be more
variable, provided breaks in any linkage between characters are equally
probable in both cases. Of course when dealing with small populations
one should not place too much confidence in the probable error calcu-
lated for any particular biometrical constant. If one could be certain
that the calculated coefficient of variation represented the true values in

SIZE INHERITANCE OF NICOTIAN A
Table 3

The pedigrees of the families and their coefficients of variation.

175

8.75^.29

9-73±-3o

5.22±.I9

6. 99 ±.28

6.3i±.25
7 . 22 ± . 26

6.85±.26
6.63^.26

6.55^.24
5-74±-22

5.93^.22

5.i8±.i8
4. 04 ±.14

5-49 ±.27

/ 4.oi±.i4
\ 7.04^.26

6.28±.27

a series of populations of this kind a single coefficient of variation higher
than that of the preceding generation would be a critical failure of the
theory of plural Mendelian determiners to meet the breeding facts. In
small populations from one hundred to three hundred, however, the
matter can only be tested by induction from a large number of experi-
ments. Table 3 is a contribution toward this end. Among the fifteen
families reported there are two exceptions to the rule which are noted
by bold-faced type. The remainder of the families all show lower vari-
abilities than the families from which they came.

Considering these data apart from other known facts, one may say
that the evidence tends to justify the use of plural segregating factors
in interpretating size inheritance, nevertheless the writer believes that
dogmatic conclusions on such a broad question should not be drawn
from a single set of experiments. Only when the numerous size studies
of such investigators as Belling, Castle, Davenport, East, Emerson,
Hayes, Heribert-Nilsson, Kajanus, MacDowell, Nilsson-Ehle,
Pearl, Phillips, Punnett, Shull, Tammes, and Tschermak are
considered together, is it possible to make a reasonable judgment of the
mechanism by which such characters are transmitted. The volume of
this work is large and the data reported, without exception, can be in-
terpreted as Mendelian. Furthermore, such an interpretation is not
merely formal, as some writers have stated, but is as genuinely helpful
to the breeder as is any Mendelian data.

In view of these facts many biologists may question the desirability

Genetics 1: Mh 1916

i76

E. M. EAST

of increasing the literature by papers of the same type. They may hold
with considerable justice that the case has been proven. At the same
time, though one may not question the value of any of these investiga-
tions, it must be admitted that the material used in most of them is
undesirable for a critical test of the theory involved. In all of the
zoological researches, bisexuality introduces a constant error into the
results. Many of the races of plants involved were markedly heterozy-
gous. The difficulty of drawing just conclusions from the botanical in-
vestigations was also increased by the use of characters affected strongly
by environmental differences. For these reasons, I hope to report the
results of several other studies of this kind in which the constant errors
are reduced to a minimum, believing that the theory must be proven or
disproven under such critical conditions. If with such material the Men-
delian notation is justified — as I believe is true when one considers the
work of Belling, Emerson, Hayes and myself on plants naturally self-
fertilized — then it will be impossible to criticize its use in those experi-
ments where some allowance must be made on account of the peculiarities
of the material involved.

LITERATURE CITED

Castle, W. E., 1914 Multiple factors in heredity. Science, N. S. 39 : 686-689.
East, E. M., 1916 Significant accuracy in recording genetic data. Amer. Jour. Bot.
(in press).

Emerson, R. A,, and East. E. M., 1913 The inheritance of quantitative characters in
maize. Neb. Agr. Exp. Sta., Research Bull. No. 2, pp. 1-120.

Hayes, H. K., and East, E. M., 1915 Further experiments on inheritance in maize.
Conn. Agr. Exp. Sta., Bull. 188, pp. 1-31.

Setchell, W. A., 1912 Studies in Nicotiana I. Univ. of Cal. Pub. Botany 5: 1-86.

Information for Contributors

The length of contributions to Genetics will not be arbitrarily limited, but articles
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitations as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication.

Contributors are requested to observe the usual care in the preparation of manu-
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published in a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer to
them, and should be separated from the text by lines. All footnotes should be num-
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds for
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is aV% X lV* inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be i2l/2 X ?H inches. Legends for all figures should be typewritten on a
separate sheet.

Authors will receive galley-proofs, and, if specially requested, page-proofs will be
sent also, it being understood that the latter must be returned very promptly, and that
no extensive change may be made in page-proofs, which is not compensated for within
the same paragraph, or in an adjacent paragraph on the same page.

Contributors will receive 100 reprints of their papers, with cover, free of charge,
and may secure additional copies at actual cost of manufacture, provided these are
ordered when the corrected galley-proofs are returned.

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

TABLE OF CONTENTS

PAGE

Bridges, Calvin B„ Non-disjunction as proof of the chromosome

theory of heredity, (concluded) 107

East, E. M., Studies on size inheritance in Nicotiana 164

Tupper, W. W., and Bartlett, H. H., A comparison of the wood
structure of Oenothera stenomeres and its
tetraploid mutation gigas 177

Harris, J. Arthur, Studies on the correlation of morphological and
physiological characters: The development of
the primordial leaves in teratological bean
seedlings 185

A Co-operative Effort

Genetics has been founded through the cooperative effort of many
persons interested in the discovery of the principles of heredity and in
their application. The plans for this journal do not contemplate any
financial gain from its publication. As soon as the subscriptions pay
the cost of manufacture and distribution on the basis of plans already
announced, the income from additional subscriptions will be devoted to
increasing the size or the frequency of issue, and to improving the quality
of the journal. Subscribers are urged therefore to assist in extending
the subscription list as rapidly as possible, in order that Genetics may be
enabled to publish an increasingly large share of the work of American
investigators in this field.

SIGNIFICANT ACCURACY IN RECORDING GENETIC

&r*J:£U • J- : / data ' ,•-

E. M. East

Reprinted from the AMERICAN JOURNAL OF BOTANY, 5: 211-222,

May, 1916.

(Reprinted from the American Journal of Botany, 5: 211-222, May, 1916.]

SIGNIFICANT ACCURACY IN RECORDING GENETIC

DATA

E. M. East

In 1913, I contributed a paper to the Botanical Gazette (55: 177-
188) on the inheritance of flower size in a cross between Nicotiana
alata grandiflora Comes, and a type thought to be Nicotiana for getiana
Hort. Sand. Corolla size had been selected for study because in this
genus it is "so comparatively constant under all conditions attending
development" — something which could not be said of any other size
character that had been under observation. Since other investiga-
tions of the same kind were under way, and a larger amount of data
might be reported later, the "liberty of asserting the truth of this
statement" with only the following data in its support was requested.
This paragraph followed.

"During the past four years, I have grown about 20 species of
Nicotiana in considerable numbers. They have been grown under
very diverse conditions. Some have been starved in four-inch pots,
others have had the best of greenhouse treatment; some have had
poor field conditions, others have had all field conditions practically
at their best. The height of the plants, the size of the leaves, and
similar size complexes have varied enormously, but the size of the
corollas has scarcely varied at all. For example, plants of Nicotiana
sylvestris Speg. & Comes, grown to maturity in four-inch pots, pro-
duced no leaves longer than 7 inches. On the other hand, sister plants
of the same pure line produced leaves 30 inches long in the field.
Both series, however, produced flowers with the same length and
spread of corolla. Furthermore, cuttings from 20 of the field plants
reported in this study were rooted and grown in small pots (6 inch)
in the greenhouse. Their blossoms were the same size as those of the
field grown plants from which they came."

[The Journal for April (3: 135-210) was issued April 18, 191 6.]

211

212

E. M. EAST

Recently Goodspeed and Clausen have published in this Journal
(2: 332-374. 1915) an immense amount of data on the influence
that certain environmental factors have on flower size in Nicotiana.
The conclusions they draw are eight in number based upon 25,000
measurements of the length and spread of the corollas of Nicotiana
tabacum var. macrophylla and three hybrids between N. tabacum
varieties and N. sylvestris, and run somewhat as follows:

1 . Both length and spread of corolla decrease during the flowering
season to such an extent that at the end of six weeks the average
spread may drop 6 mm. and the average length 4.5 mm.

2. The Fi N. tabacum X N. sylvestris hybrids are shoit-lived
perennials, and the flowers of the second season are of approximately
the same size as those of the first season.

3. Removal of open flowers during the normal flowering season
prevents nearly all decrease in size.

4. Flowers apparently fully opened are smaller before than they
are after anthesis, even though the anthers are partially sterile.

5. Flowers on pot-grown cuttings are smaller than those borne on
the field plants from which they were taken.

6. Under favorable and unfavorable greenhouse conditions, flower
size varies distinctly and in the same direction as the vegetative
characters.

7. Length of corolla is more stable than spread of corolla under
environmental stimuli.

8. "The only true distribution representing the flower size of a
population must be based upon measurements which, for each plant,
extend over the greater part of the period of flowering normal for the
given species or hybrid group, or cover an identical portion of the
flowering period of each plant."

The data were collected and these conclusions drawn, the authors
say, "to establish tentative criteria in keeping with which flower
size investigations, in Nicotiana at least, should be carried on and
interpreted."

The statements of Goodspeed and Clausen and those quoted from
my own paper seem at first sight to be irreconcilable. Indeed, the
authors have done me the honor of devoting a considerable portion
of their paper to criticizing my views and methods. For example,
because it was maintained that flowers are constant under different
environments compared with the changes exhibited by vegetative

SIGNIFICANT ACCURACY IN RECORDING GENETIC DATA 213

organs, they have assumed that no precautions whatever were taken
to eliminate environmental differences. Since the statement was
made that plants were grown under diverse conditions, a fact men-
tioned merely in connection with the question of the effect of stimuli
on corolla size, they seem to have concluded unjustly and unreasonably
that the data from these experiments were used in the paper under
consideration.

On the other hand, Goodspeed and Clausen are perfectly justified
in asking for a description of the way in which my data were taken.
I wish to make such a statement, therefore, in order to support my
former paper and some other studies on the inheritance of flower size
which are to be published in the near future, and because of the op-
portunity presented to illustrate a question of considerable general
interest. This question, which as a teacher of genetics I have found
neglected by research students more than any other, is: What is sig-
nificant accuracy in recording data?

The seemingly opposed statements of Goodspeed and Clausen
and of myself serve to illustrate the thought in mind. The two
allegations are not wholly discordant. Although I do not wish to
withdraw or to modify my own statements, at the same time I am
willing, in a broad sense, to accept most of their conclusions. Ex-
cluding certain differences in our data that are undoubtedly due to
dissimilar conditions at Berkeley, California, and at Boston, Massa-
chusetts, my own results are similar to theirs except as to the magni-
tude of the changes caused by environmental differences. The point
upon which we differ decidedly is the significance of the results in
relation to the problem at hand — the inheritance of differences in
corolla size in Nicotiana.

One of my college instructors once said to me: "It is seldom
necessary, in the interests of scientific accuracy, to weigh a ton of
hay on an analytical balance." That statement might be made the
basis of a course on Precision of Measurements. One is hardly ever
required to impress mechanical accuracy upon really earnest students.
They will weigh and measure material with the utmost pains (in spirit
at least). What is difficult is to impress an idea of true precision.
It is not uncommon to see measurements recorded to tenth milli-
meters after the random use of two scales having a one percent differ-
ence, or material for analysis weighed to the fourth decimal place
with weights that have never visited the Bureau of Standards, on a

214

E. M. EAST

balance with very unequal arms. It is rare to find students who
think of these errors and endeavor to correct them, although such
correction is as necessary in biology as in physics. Let us see how
our biological problem fits the rules for the treatment of errors in use
in experimental physics.

It was desired to record, in such a manner that they would be
comparable, numerics that represented the phenotypes of series of
plants of species of Xicotiana in regard to corolla length and spread,
sufficiently accurately that genetic analysis of the results might be
made.

The investigation was initiated by a series of preliminary measure-
ments designed to show the practical physical limits to the precision
of the direct measurements. Repeated measurements of the same
flowers showed that there were residual errors beyond one millimeter
in the case of length and two millimeters in the case of spread of
corolla. Measurement to millimeters was adopted, therefore, although
these measurements were afterwards thrown into larger classes for
reasons that can be justified biometrically.

Then came a study of ontogenetic variation in order that the
factors affecting such variation might be detected. The factors that
would naturally occur to anyone who had had experience in growing
plants were time of planting, physical and chemical condition of the
soil, moisture, age of plant, flowering period, age of flower, position of
inflorescence on plant and position of flower in the inflorescence. To
determine the effect of each of these factors, it was necessary of course
to eliminate the influence of all the others as far as possible. Since
the cultures to be compared were nearly always planted at the same
time, and since this variable is somewhat dependent upon others that
were under consideration, it was neglected. My cultures have also been
grown in well-drained soil very uniform in its fertility, but it was
thought wise to determine how much effect extreme soil conditions
might have. Several species growing outside in soil of good tilth were
compared with greenhouse pot cultures. Three-inch, four-inch, five-
inch and six-inch pots were used in various species, but the treatment
was uniform for each species. The species were N. tabacum (several
varieties), N. rustica (several varieties), N. longifiora (two varieties),
N. sylvestris, N. paniculate, N. acuminata, N. forgetiana and N. alata
grandiflora. Since only from ten to twenty plants could be grown
in the greenhouse in most cases, statistical constants were not calcu-

SIGNIFICANT ACCURACY IN RECORDING GENETIC DATA 215

lated, for I have not the faith of Goodspeed and Clausen in probable
errors based on nine or ten observations (see their tables II a, b and
III a, b). Averages of five flowers per plant taken when first in full
flower, however, indicated means within a millimeter of each other
for length and within two millimeters of each other for spread of
corolla for over half of the species, when compared with the sister
plants in the field. The greatest difference was in a N. alata grandi-
flora test where the starved plants showed an average of about 5 mm.
shorter and 7 mm. narrower flowers. Hybrids were also tested.
As I do not consider it necessary to cite figures endlessly where they
serve so little purpose, however, only a table of results on a cross
between two varieties of N. longiflora is given, the field records and
the pot records being made by different observers. The general

Table I

Frequency Distribution for Length of Corolla in Cross between N. longiflora Varieties

Designation

No. 383 field. .
Xo. 383 pots. ,
No. 330 field .
Xo. 330 pot. .

(383 x 330)

F3A field...
Ditto, pot

(383 x 330)

F3B field...
Ditto, pot

(383 x 330)

F3C field...
Ditto, pot

(383 x 330)

F3D field . .
Ditto, pot

(383 x 330)

F3E field...
Ditto, pot

Class Centers in Millimeters

37 40 43! 46 49 52 55 58 I 61 I 64 67 70 73 76 7Q 82 1 85 88 91 94 97

2 6

3 9

70 19 10
. . . 2

1 1

59 41 19
4 1

32
2

20*
2

10 1

effect of starvation can be seen even without having the means calcu-
lated. A comparatively small number of observations were made
on each population, but they serve as samples of the frequencies found.
Certainly no marked decrease in size is apparent, and since the vegeta-
tive organs of the pot-grown plants varied from one half to one fifth
the size of those in the field (linear dimensions), it seems that one

216

E. M. EAST

should be justified in stating that comparatively starvation had no
effect on the flowers.

Both sets of these plants had a sufficient supply of moisture to
keep them healthy. When this is not the case there is some difference
in flower size. For example, some N. rustica plants each showing a
mean flower length of 20 mm. with extremes of 18 mm. and 22 mm.
at the first of the season, decreased in their mean flower length to
18.8 mm. after being in flower for four weeks during which very little
rain fell. Then came four inches of rain within forty-eight hours.
After this, stout vigorous laterals arose from the lower part of the
main stems bearing flowers with a mean length of 21. 1 mm. (extremes
were 19 mm. and 23 mm.). Thus a marked difference in activity of
cell division shows its effect on the flower.

This factor is probably the cause of the greater size shown by
flowers on lateral branches when compared with those on terminal
branches in Goodspeed's and Clausen's work (Tables XIII, XIV, XV).
These authors also found that the flowers on new vigorous branches
after "cutting back" were increased in the same way.

These facts should be taken into consideration when examining
the conclusion of the California botanists that flower size decreases
markedly as the length of the flowering season increases. Their data,
as well as my own, proving that flower size may keep up to that of
the first of the season and even increase if the weather conditions
remain favorable for the production of vigorous new lateral branches,
show that it is questionable whether a significant decrease in flower
size occurs during the time that data would be likely to be taken.
Their data showing marked decreases from the first of the season to
1 mid-season are from populations of 9 and 10. During similar periods
I have found no measurable decrease in flower length in N. tabacum,
N. longiflora, N. paniculata and N. rustica. I have found a mean
decrease of 1.0 mm. to 1.5 mm. which possibly is due to this factor in
certain cultures of N. langsdorffii, N. acuminata, N. forgetiana and
N. alata grandiflora, but I think the true occasion of the decrease was
lack of moisture. On the other hand, there seems to be evidence in
Goodspeed's and Clausen's data that toward the end of the season
there is likely to be a decrease in flower size. My own data have
shown a drop of from 4 mm. to 8 mm. in both corolla length and
spread in various species in the last dozen or two flowers produced.
This shows as a sudden change which is evidently due to physiological

SIGNIFICANT ACCURACY IN RECORDING GENETIC DATA 21 7

reasons. The true state of affairs is masked, therefore, when this
decrease is treated as a gradual drop in flower size during the season.
If measurements on greenhouse cultures grown in proper sized pots
are taken daily over a long period, they simply show comparative
uniformity in flower size until about the end of the flowering season.
Then a decrease which produces a sharp bend in the curve occurs.

As to variation in size owing to age of the flower, I have found that
this is largely a mechanical difficulty. There is no difference in length
between flowers before and after anthesis, for anthesis takes place
normally either before or within 10 hours after the flower opens in all
species of Nicotiana under Boston conditions. A flower if unpolli-
nated may open for as many as 5 successive days, and there is a slight
increase in both length and spread of the corolla. But a pollinated
flower seldom opens on more than two successive days. The flower
becomes less firm however and the spread of the corolla may appear
to increase.

Flowers of the same relative position on vigorous branches are the
same size whether they be on the main stalk or on laterals in species
like N. forgetiana and N. alata grandiflora which are characterized by
vigorous lateral branches from the base of the stem. Flowers on
lateral branches in species like N. tabacutn where the main stem is so
much more vigorous, average (in my counts) slightly less (under 1 mm.)
than those on the main stem.

After about the sixth flower on the species having racemes, and on
the flowers coming out after the first full glory of the panicled species,
there is also a slight decrease in size owing to decrease in the conducting
channels of the fibro-vascular system.

What information do these observations, which are the preliminary
" qualitative" tests made in every investigation, give us? They show
that to record the phenotypes of flower size of a series of Xicotiana
plants, the seeds should be sown at the same time in uniform soil, the
plants should be pricked out uniformly and set at the same time in a
plot of uniform fertility. The flower records should be made within
two or three weeks of each other at the first of the season, allowing no
marked climatic change to intervene if possible. The flowers recorded
should be the vigorous flowers (as stated in the last paragraph) of
vigorous branches, and should be measured on the same day that they
open.

This procedure should be followed where it is physically possible,

218

E. M. EAST

and any departure noted in order that a correction for any constant
error due to it may be calculated, if it be advisable. But, one might
ask, would not any trained geneticist have taken these precautions
anyway? What has been gained?

The advantages are real. Unsuspected constant errors often come
to light through such preliminary investigations. The good fortune
that none appeared here certainly makes it no less satisfactory. It
showed that control of conditions in such a manner that constant
errors will be negligible in the end result is technically possible. It
gave a definite idea of the magnitude of the error produced when
various environmental factors do vary, and this is very necessary in
determining the probable limits of error.

There is a way of testing the conclusion that with the conditions
controlled as suggested the constant error is negligible. If the same
plants are measured during similar portions of successive periods of
flowering activity, there is but one other obvious variable — total age
of plant. If the latter has no measurable effect the two frequency
distributions should duplicate. On this point I have no data, but
Goodspeed and Clausen have corroborated the expectation in their
conclusion number two. I do have some data on random samples
of the same pure line grown in different years. This will be taken up
later, however, as another point is involved.

Now the question arises: If records are made in this uniform
manner, how many records from each plant are needed to obtain a
measure of that plant with the precision necessary for a genetic investi-
gation? Goodspeed and Clausen say that twenty-five flowers is the
minimum. At the beginning of my Nicotiana investigations (1908),
I used the same number, curiously enough. But I soon found that
this was "accuracy with no significance," and the number was reduced
to five. I now use but one measurement per plant. This is done
because the precision is so nearly that of using twenty-five flowers,
that it would be a waste of labor to try to attain the other. Further-
more the precision obtained by measuring twenty-five flowers is only
appreciably greater when it can be done in a short time, otherwise
constant errors may become very much greater.

The precision attained by measuring one flower per plant is all
that is required for the use to which the data are to be put, and it is
a rule of experimental physics not to strive for greater accuracy.

This matter can and has been tested in two ways. The first is to

SIGNIFICANT ACCURACY IN RECORDING GENETIC DATA

219

compare random frequency distributions of the corolla size of single
plants with frequency distributions of the flowers when selected from
vigorous branches and measured on the same day they have opened.
This procedure gives a measure of the accuracy of single flower selec-
tions. To illustrate this, data from two species with very different
sized flowers are submitted.

Table II

Comparison of Random Samples of Corolla Length on Single Plants and Samples in
which Constant Errors have been Largely Eliminated

Name

Class Centers

in Millimeters

20 | 21

N. paniculata, Random

3
4
14
17

16
18

4
3
4
3
4

3
1

Selected

Ran

4
5

Sel

Ran

16
20
1

5
2

Sel

Ran

Sel

Name

Class Centers in Millimeters

IOC

N. alata gr., Ran

3
2
1

16
22
6

3
2

4
1

3
2

" " " Sel

" " " Ran

3
4
17

' Sel

" " " Ran

" " " Sel

These plants are among the most uniform and the most variable
respectively, and give an idea of the range of variability involved.

The other test made was to select fifteen flowers on a plant at
random, and determine the mean to the nearest millimeter; then to
find the deviation from this mean when single flowers were selected.
In 100 tests of flowers shorter than thirty millimeters 88 selections were
made within the 3 millimeter class to which the mean belonged.
The remainder were in contiguous classes. On flowers between 70
and 100 millimeters long 82 out of 100 selections were within the 6
millimeter class to which the mean belonged. The remainder with
2 exceptions were in contiguous classes.

From these tests it will be seen that the probable error of the
selection (equal chances) is not over plus or minus 2 percent. If this

220

E. M. EAST

were a constant error it would be considerable. But it must be
remembered that it belongs to the class of accidental errors and that
in the long run the minus errors are compensated by the plus errors.

Such compensation can be clearly seen and the accuracy of the
method perhaps most clearly demonstrated by comparing frequency
distributions of the same pure line, daughters of the same plant, during
successive seasons. In a number of cases populations of sister plants
were grown for two and three years. The seed in each case came from
single 1909 or 1910 plants, and since the percentage germination
remained practically constant, the different populations are in the
nature of duplicate and triplicate determinations. If then the fre-
quency distributions are sufficiently alike that they may be presumed
to be random samples of one population, the method is accurate enough
for genetic purposes. A sample of the result is shown in Table III.

Table III

Random Samples of the Same Population Grown in Different Seasons

Name

Class Centers in Millimeters

Means

37 40

100

N. longiflora, var. A, 191 1 .
" " " 1912.

" " 1913.
N. longiflora, var. B, 191 1.

80
28
32

32
16
I

40.46 zb. 1 1
40.61 rfc. 19
39.76dr.i2
93.22dr.l6

93.37i.20
92.12dr.37

6
2
5

22
16

49
32
10

11 11 11 11 I9I2

" " I9I3-

When one takes into consideration' the difference in size of corolla
among Nicotiana species and varieties that will cross and give fertile
hybrids — L e. N. langsdorffii 21 mm. and N. alata grandi flora 85 mm.,
it is scarcely necessary to enter into a biometrical argument on the pre-
cision of the method. Here are two small samples of the same popu-
lation of N. langsdorffii grown in 191 1 and 1914:

Designation

Class Centers in Millimeters

191 1 plants

3
9

1
7

2
1

1914 plants

Can it be doubted that the phenotype for corolla length to which
N. langsdorffii belongs is shown here with an accuracy much greater

SIGNIFICANT ACCURACY IN RECORDING GENETIC DATA 221

than is necessary when an analysis of the hybrid progeny of it and
N. alata grandiflora is contemplated? Biometrical methods are
much too imperfect to demand more. There is no intention to dis-
cuss here the reasons why the biometrical methods in general use in
genetics are imperfect. But it must be emphasized that they are
merely used in default of better, since many of them cannot be de-
fended either mathematically or biologically. For example, common
sense tells us that equal-sized classes should not be used for the two
very different species shown in Table III, where the corolla of one is
three times that of the other, yet no satisfactory method has been
proposed which does away with the difficulties involved. Since it is
necessary to use such poor methods in calculating our end results in
genetic studies of size, however, one should remember that labor to
record data far more precisely than these methods require is labor
wasted.

At the same time, though one may believe that biometrical methods
are imperfect for certain purposes, they are founded on the theory of
probability and when used should be used with this in mind. Having
recorded his data with the precision desired, one should not try to
analyze them until he has collected a sufficient number of observations
to make calculations of residual errors have meaning. Just what the
minimum number should be varies with the problem and cannot be
discussed in this paper. There are several textbooks on the Theory
of Measurements in which the matter is treated in detail. All I wish
to point out here is that in every problem capable of biometrical
analysis there is such a minimum, and if the data to be analyzed are
far under this required minimum, no over precision (in cases where
this is possible) in making the records will give them value.

An excellent illustration of this is found in Goodspeed's third
article on Quantitative Studies of Inheritance in Nicotiana Hybrids.1
The author used his method of recording measurements of flowers
through a considerable portion of the flowering season in order to
determine the phenotypes to which the plants belong, and yet has
made analyses of frequency distributions having such a small number
of entries that they possess no meaning whatever. Among 44 fre-
quency distributions, 29 have less than 12 plants recorded. He
recognizes the fact that the number of plants involved is too small,
but feels that this deficiency is balanced by the accuracy of his records.

1 Univ. Cal. Pub. Bot. 5: 223-231. 1915.

222

E. M. EAST

He says: "Data which have been submitted, however, leave no room
for doubt in my own mind that investigations on the inheritance of
flower-size demand the recognition of certain definite criteria and that
the results of such investigations are vitally influenced by inherent
as well as externally induced physiological states peculiar to the plant.
Thus it remains to be seen if as many as 800 plants are necessary to
establish the validity of an expanded Mendelian notation in F2 of
a flower-size hybrid, whether the 40,000 to 80,000 measurements,
seemingly essential to a fair expression of results, can be accumulated.
In other words, the experiment with which this paper deals has been
a partially successful effort to measure many flowers on a few plants
with the thought that the conception of flower-size would thus be
approximately perfect for a few, rather than certainly imperfect for
many plants. It is undeniably true that the number of plants is
smaller than it should be, and it is perfectly evident that if the flowers
on a larger number of plants cannot be correctly measured the attempt
is not worth making."

One could hardly find a better illustration of "accuracy without
significance." These views are absolutely indefensible mathemati-
cally. It has been shown that the method used by Goodspeed in
making his records has only a fallacious claim to great precision; but,
granting that the method is extremely accurate, it is an accuracy
unnecessary to the end result. On the other hand, it should be clear
that records in sufficient number to make probable errors significant
is positively essential for a biometrical analysis. This end can only
be attained by recording larger numbers of plants and not by over-
refinement in the plant records. The plant records should have the
precision required by the end result, but greater precision does not
influence this result.

Harvard University

INHERITANCE IN CROSSES BETWEEN NICOTIAN A LANGS-
DORFFII AND NICOTIAN A ALATA

E. M. East

Harvard University, Bussey Institution, Forest Hills, Massachusetts

Reprinted from GENET I OS' 1 : 311-333, July, 1916

INHERITANCE IN CROSSES BETWEEN
NICOTIAN A LANGSDORFFII AND NICOTIAN A ALATA

E. M. EAST

Harvard University, Bussey Institution, Forest Hills, Massachusetts
[Received March 14, 1916]

TABLE OF CONTENTS

PAGE

Introduction 311

Early work 3J3

Inheritance of pollen color and of flower color 3l7

Fertility of the hybrids 3*9

Height 319

Rapidity of growth 32°

Leaves 320

Corolla length 322

The reciprocal cross 329

Conclusion 332

Literature cited 333

INTRODUCTION

In a rather intensive genetic study of the genus Nicotiana including
some sixty inter-specific crosses, the writer has found very few fertile
crosses between species whose status would not be questioned by taxono-
mists. Of these, the one showing the most perfect fertility is to be
described in this paper.

Nicotiana Langsdorffii Weinm. and Nicotiana alata Lk. and O. are so
different from each other in their characters that they were placed by
George Don in the different sections of the genus, that he called Rustica
and Petuniodes, and have been kept there by Comes, the most recent
monographer of the Nicotianas. The writer agrees with the suggestion
of Lock (1909) that N. Langsdorffii should be removed from the
Rustica section to the Petuniodes section on the basis of its genetic be-
havior when crossed with N. alata, but the very fact that taxonomists
without access to genetic data have seen fit thus to separate them is an
indication of a specific distinction not to be questioned except by those
who would fuse all types giving fertile hybrids.

Nicotiana Langsdorffii was described by Weinmann (Roem. &

Genetics 1: 311, Jy 1916

312

E. M. EAST

Schult. Syst. iv. p. 323) from Brazil. It probably lias a wide distribu-
tion in South America as it has been found in Chile (Comes 1899). The
immediate sources of my plants were Setchell (1912, his If) and
A. Splendore, Scafati, Italy. I do not know where Setchell obtained

Figuki: 1 Figure 2

Figure i. A young flowering plant of Nicotiana alata Link and Otto, var. grandi-
fJora Comes.

Figure 2. A young flowering plant of Nicotiana Langsdorffii Weinm.

his plants, and the two strains may be from the same stock. At any rate
they are practically identical, both corresponding with the plate in the
Botanical Magazine (1825 pi. 2555).

The plants are from 120 — 145 cm in height, vigorous, profusely
branched, the branches erect. The basal leaves are 20 — 30 cm long,

CROSSES BETWEEN NICOTIAN A LANGSDORFFII AND .V. ALATA 313

obtuse, ovate, sessile, narrowed and decurrent at the base. Upper
leaves are lanceolate, and all are extremely rugose above. Inflorescence
racemo-paniculate. Flowers are about 20 mm long, very uniform in
size; corollas funnel-shaped, a gibbous ring above, the limb concave,
spreading, and very slightly notched; greenish yellow, pendulous. The
pollen is blue.

Nicotiana alata as described by Link and Otto (see Ic. PL Rar. I,
63, t. 32. DC. Prodr. XIII. I. p. 567. Garten flora tab. 1010. Comes
1899, p. 35) from Brazil (found in Uruguay and Paraguay according to
Comes), I have never seen. The type used in these experiments is the
common N. affinis Moore (Gardn. Chron., 1881, p. 141) referred by
Comes to the variety grandiflora. The variety seems to have no points
by which it can be distinguished from the species. It is described as
having larger flowers with more perfume, more zygomorphism and less
gibbosity than the species, but these are. very indefinite and inconstant
qualities.

The strain with which our crosses were made has plants 1 10-130 cm
high, appearing shorter because of the loosely spreading habit. Basal
leaves are acute, ovate, quickly narrowed to a slightly decurrent base,
slightly rugose ; upper leaves lanceolate to linear. Inflorescence is a
raceme. Flowers are 75 — 95 mm long, tube gradually enlarging toward
the limb and slightly gibbous at the top, light greenish yellow faintly
lined with purple; limb broadly expanded into obtuse, ovate lobes the
lower two being distinctly smaller than the other three and giving the
flower a decidedly zygomorphic form. The corolla limb is pure white on
the inside and cream with sometimes a tinge of purple on the outside.
One anther is usually somewhat shorter than the others. The pollen
is white or yellowish. Some plants are self-fertile, others are completely
self-sterile.

EARLY WORK

These two species were crossed and studied by at least three of the
earlier hybridizers, Naudin, Godron and Focke. Concerning their re-
sults, I quote Focke (1881) :

"N. alata Lk. X Langsdorffii Weinm. Gartner found no foreign species
with which he was able to fertilize N. Langsdorffii. Reciprocal crosses be-
tween N. alata and N. Langsdorffii are not difficult, however ; Naudin ob-
tained especially good, well filled capsules by fertilizing N. Langsdorffii
with pollen from Ar. alata, and although only one pollination of A\ alata
with Ar. Langsdorffii pollen was successful, in this case also a large capsule
full of seeds matured. I found no difficulty with either cross. Of Ar".

Genetics 1 : Jy 1916

3i4

E. M. EAST

Langsdorffii $ X N, alata 6 (X '. Persico-Langsdorffii Naud. 1. c. p. 74)
Naudin produced 118, and of A', alata $ X N. Langsdorffii <2 (N. Langs-
dorffii-Persica Xaud.) 53 examples; all of which were exactly like one
another. They were 130 — 160 cm high (Ar. Langsd. ca. 100 ; Ar. alata 60 cm)
and because of their spreading branches more nearly resembled N. Langs-
dorffii. The blossoms were medium large, greenish white, with the limbs
distinctly rounded. Pollen bluish gray. Fruitfulness perfect. I have made
the same crosses with like results. Pollen was plentiful and the grains
well-formed. The capsules contained in the neighborhood of 500 seeds.

"Naudin's hybrids were to be distinguished from Ar. commutata by their
higher stature, their larger and more greenish flowers, and their darker
leaves.

"Later generations. Through continued self-fertilization, Naudix's hy-
brid plants gradually returned toward the condition of the parent species,
although this was never fully reached. Godrox received from Alex.
Braun of Berlin, seed of A\ alata-Langsdorffii (as well as of N. Langs-
dorffii 2 X N. alata c?) and raised many forms from it; among others were
varieties with yellow, with cream, and with pure white flowers. The leaves
were variable, the decurrence at the stem being sometimes very pronounced,
sometimes just traceable and sometimes lacking.

"Two varieties in which crossing had been prevented by gauze produced
fruits whose seeds reproduced the mother form exactly."

From this extract, it is clear that Focke was familiar with the facts
that in the cross under consideration — as well as in other crosses — the
Fx generation is more vigorous than either of the parent species, that the
population is uniform and the individual plants fertile, and finally that
the F2 and following generations are variable and may produce plants
having a striking resemblance to the original parents. In this he was
merely copying Naudin. Both Naudin and Godron perceived the es-
sential facts of inheritance in hybrids much more clearly than other
contemporary hybridizers, and we may be assured that had Naudin
had an opportunity of reading Mendel's paper, as did Nageli, he would
have appreciated its significance. He came very close to an enunciation
of what we now know as the Mendelian laws, but either he lacked the
ability for mathematical analysis that characterized Mendel because of
the latter's training in physics, or was prevented from making such an
analysis by the greater complexity of the hybrids he studied.

Naudin (1865) says of his cross, Nicotiana Persico-Langsdorffii:

"The two plants here united, although very different at first sight, have
distinct analogies in their habit of growth, the form of their leaves, their
general aspect, and up to a certain point, in their long, tubular, pendent
flowers. One feels these analogies more strongly if he remembers that
there exists a form exactly intermediate between the two (N. commutata,
Fisch.), of which I shall speak later. As well as I can judge by the descrip-
tions, N. Per ska, of which there is a pretty good figure in the Botanical
Register, pi. 1592, appears to be identical with the N. alata of Dunal;

CROSSES BETWEEN NICOTIAN A LANGSDORFFII AND .N. ALATA 315

not being sure, however, I have preserved the name that it carried at the
Museum."

Naudin wished to see whether Ar. commutata was a natural hybrid
and if he could reproduce it by this cross. He obtained 118 plants from
his cross, "all of the most uniform appearance" and from 130-160 cm
high. He concludes that N. commutata is not the hybrid N. Persico-
Langsdorffii, "at least of the first generation." Since he made this con-
clusion solely on account of the greater height and vigor of his artificial
hybrids, it must be that he suspected that he might have duplicated
N. commutata when by inbreeding his plants had lost their hybrid vigor.
I have duplicated plants of N. commutata grown from seed received
from Dr. Splendore of Scafati, Italy, several times in my own crosses
and have obtained F3 families that bred as true to the form (intermediate
between N. alata grand, and N. Langsdorffii) of the so-called N. com-
mutata as did the species (?) itself.

Naudin found that the reciprocal cross was so nearly like the other
that "without the labels the two lots would have been taken the one for
the other." Unfortunately, however, although these crosses were per-
fectly fertile, Naudin did not self them and continue his observations.
On the other hand, he did obtain some information regarding later
generations by a consideration of the volunteer seedlings that appeared
during the next few years on the plat that had borne the original cross.
He says :

"Without having given these hybrids of the second and the third gener-
ations the attention they merited, I have noticed that their forms became
more and more divergent, some approaching N. Per sic a and others dis-
tinctly tending toward N . Langsdorffii."

Some of these plants he potted, and obtaining seed from one that re-
sembled N. Langsdorffii he grew a population that bred true to a type
that could scarcely be distinguished from N. Langsdorffii. These experi-
ments were continued, and from seed of this generation, he raised in
1863, fifty plants nearly all of which "had returned to the type well
known as Ar. Langsdorffii.,,

Thus it is seen that Naudin observed nearly all the essential facts of
Mendelian heredity in this one cross, — a uniform Ft generation, a segre-
gating F2 generation, and a later generation which showed that certain
of the extreme segregates bred true. But the observations on this par-
ticular cross are not so important as the general conclusions to which
Naudin was led by his broad experience as a hybridizer. Under the
heading "Physionomie des hybrides," he says :

Genetics 1: Jy 1916

316

E. M. EAST

''In order to have a correct idea concerning the phenomena presented by
hybrids, it is essential to distinguish between the first generation and those
that follow.

"I have always found, in the hybrids I have made myself or of whose
origin I was certain, a great uniformity of aspect between individuals of
the first generation and originating from the same cross no matter what
their number. This fact we have seen exemplified in Petunia violaceo-
nyctaginiflora, Datura Tatulo-Stramonium and D. Stramonio-Tatula,
D. Meteloido-Metel, D. Stramonio-laszis, etc., Xicotiana Texano-rustica and
rustico-Tcxana, X. Persico-Langsdorffii, etc. ; having already emphasized
these resemblances it is useless for me to dwell upon them."

"In fact, one may say that hybrids of the first generation resemble each
other as much or nearly as much as the individuals that come from a single
legitimate species."

(It is well to note that the hybrids with which Naudin supports his
thesis here are all between solanaceous species that are generally self-
pollinated naturally, and may be presumed to approach homozygosis.)

"Beginning with the second generation, the aspect of hybrids is changed
in a remarkable manner. Ordinarily, the perfect uniformity of the first
generation is succeeded by a regular medley of forms, some approaching the
specific type of the father, others that of the mother, a few returning
suddenly and entirely to the one or the other form. At other times, this
progress toward the original types is by degrees and slowly, and sometimes
one sees a whole collection of hybrids incline toward the same side. The
important fact, however, is, that it is the second generation, in the great
majority of cases (and perhaps in all), that starts this dissolution of the
hybrid forms, a phenomenon recognized by many investigators, doubted by
others, but which appears to me to-day to be established beyond argument.
We shall explain the cause in the following paragraph."

"All of the hybrids of which I have studied the second generation with
some care, have shown these changes in appearance and have manifested
this tendency to return to the forms of the original species, and this when
conditions have been such that the pollen of the species themselves could
not have been the cause. We have seen striking examples of it in Primula
o ffi c inali-g r and i flora, in all of the hybrids of Datura Stramonium, in D.
Meteloido-Metel, in the reciprocal hybrids of Nicotiana angustifolia and
macrophylla, X. Persica and Langsdorffii, Petunia violacea and nxctagini-
flora, in Luffa aeutangulo-cylindrica, and further in Linaria purpureq-
vulgaris. In the second generation of several of these hybrids there has
been a complete return to one or the other or to both of the parent species
together with individuals approaching each species in varying degrees; in
other cases also we have seen intermediate forms continued at the same
time that other specimens from the same family have effected the return
of which I speak. Moreover, we have observed cases (Linaria purpureo-
vulgaris oi the third or fourth generation) of actual retrogression toward
the hybrid form, sometimes a plant that had apparently returned entirely
to one of the two species, has even given rise to individuals that very nearly
resembled the other species. All of these facts are explained naturally by
the segregation (disjunction) of the two specific essences in the pollen and
the ovules of the hybrid."

CROSSES BETWEEN NICOTIAN A LANGSDORFFII AND .V. ALATA 317

Space may not be claimed to show just how Naudin's views differed
from those of Mendel, except the bare statement that he did not grasp
the idea of a unit-character inheritance. Our quotations are already
somewhat lengthy. They may well be pardoned, however, since they are
taken from a paper not readily available to most geneticists, and have
a considerable theoretical and historical interest. But it was not for this
alone that I have used them, nor because they contain observations upon
the particular cross that is the subject of this paper. And in passing let
me say that there is scarcely a doubt but that Naudin's Nicotiana Per ska
and N. Langsdorffii are the same as our own N. alata grandi flora and
N. Langsdorffii. The particular reason for the citations is this: While
it is to be hoped that with the fruitful hypotheses of modern biology as
guides, contemporary genetic research is to be more productive than that
of the early nineteenth century hybridizers, it must not be forgotten that
very often we are merely repeating more carefully, more quantitatively
and with a better idea of relative values, the experiments of these
pioneers. The observations of such men as Naudin have been confirmed
and as far as they go are usually correct. For this reason I think that
we may accept their facts until the same experiments have been repeated
more carefully and have given us more precise data. This being true,
there is no question but that these numerous observations on hybrids be-
tween species belonging to so many different groups, showing as they
do all the essential phenomena of Mendelian inheritance, go far toward
proving Mendelian heredity in quantitative characters.

The only recent work upon N. Langsdorffii-alata hybrids is that* of
Lock. Lock made a number of crosses between species of Xicotiana in
the years 1906-8, but published only one paper (1909) on the subject.
He crossed A7, alata and N. Langsdorffii reciprocally, made several back-
crosses, and studied the selfed progeny of the F1 generation. He noted
the uniformity of the Ft generation and the variability of the F2 gener-
ation and reported a few measurements of the flowers. He established
the dominance of blue pollen over yellow and of yellow corollas over
white. He also believed that the facts indicated the dominance of
gibbous over funnel-shaped corollas.

INHERITANCE OF POLLEN COLOR AND OF FLOWER COLOR

In a cross such as this between two distinct species it is important to
know whether any distinct qualitative difference shows a Mendelian be-
havior. I found only two such differences, pollen color and flower color,

Genetics 1: Jy 1916

3Ig E. M. EAST

and have corroborated the results of Lock in regard to them. I was not
able to corroborate his conclusions in regard to corolla shape, as F2
plants all showed some development of the gibbous condition. And it
seems to me that this was to be expected for it is characteristic of
both species.

N. alata has yellow pollen and N. Langsdorffii blue pollen. No matter
which way the cross is made the pollen is blue. The pollen of the
heterozygotes is often lighter than that of the pure N. Langsdorffii, but
not invariably so. Microscopical examination showed no distinctly yellow
grains on the F1 plants so that, like pollen shape in Lathyrus, the color is
a sporophytic character. This is less astonishing than the phenomenon
in the sweet pea, for it is well known that pollen color is a tapetal deposit.
One F2 consisted of 342 plants with blue pollen and 100 plants with
yellow pollen. Counts of smaller segregating populations corroborated
these results, though there was an excess of blues in all but one case,
an F3 family consisting of 39 blue and 22 yellow. Just what this excess
of blue-pollened plants means, I am unable to say. It may be only a
technical difficulty, as the anthers of both species are blue. At any rate,
there seems to be no possibility of other factors being concerned directly.
Yellow-pollened plants have never given blues.

The flower of N. Langsdorffii is greenish yellow both outside and in-
side the corolla. N. alata, on the other hand, though slightly greenish
yellow with sometimes a faint tinge of purple on the outside of the
corolla, is pure white on the inside of the limb. Apparently the cells
just beneath the epidermis on the inside of the limb of these flowers
contain no colored chromatophores and very few plastids of any sort.
The flowers of the Fx hybrids (made either way) are cream-colored, but
appear to be variable because the old flowers are so light as to be some-
times mistaken for whites. Further the smaller-flowered plants appear
to be a darker yellow owing to a concentration of chromoplasts which
show through the upper two layers of cells. There is no question but
that the inheritance of these differences is Mendelian, but it is not
certain that only one factor is involved. Three F2 populations gave
ratios of 196 yellows to 61 whites, 50 yellows to 15 whites, and
57 yellows to 15 whites respectively. A heterozygous F3 family also
gave a ratio of 112 yellows to 29 whites, but one of our F2 families pro-
duced 70 yellows to 6 whites. This constant excess of yellows leads one
to suspect complications, but it can be said that no white ever produced
yellows after self-pollination, though a number of such families were
grown.

CROSSES BETWEEN NICOTIAN A LANGSDORFFII AXD N. ALATA 319

It is possible that there is a correlation between small flowers and
yellow color though this could not be established.

FERTILITY OF THE HYBRIDS

These slightly distorted ratios give some cause for the surmise that
differential fertility exists among the gametes and the zygotes. That all
the possible gametic constitutions mature at spermatogenesis cannot be
asserted without a cytological study of the early stages. The capsules on
the F1 plants were well filled, however, and the germination of the seeds
was between 90 percent and 100 percent.1 This seems a fair proof that
the ovules were all functional and that there was no selective elimination
of zygotes.

On the other hand, all of the pollen produced by the F± plants and of
the plants of later generations was not well formed. An examination of
the pollen of 20 Fx plants after having simply shaken it out on slides
showed both when dry and in glycerin or in sugar solutions that ap-
parently functional pollen grains existed in percentages varying from
70 to 96. Of course one cannot say that all of these seemingly well-
formed pollen grains are functional, as Dorsey (191 5) has shown that
in certain Vitis species they sometimes contain no generative nuclei ; but
since in nearly all the plants there are around 85-90 percent perfectly
formed pollen grains one may be fairly certain that if much selective
elimination of gametes occurs it occurs before the pollen grains are
formed, for the parent species themselves show only from 80-90 percent
of well formed grains.

HEIGHT

Nicotiana Langsdorffii (328) and Ar. alata (321) are nearly the same
height,^ — about 132 cm and 120 cm respectively, — but they are very diffi-
cult to measure owing to their becoming so profusely branched during

1 Goodspeed (1913) has criticized a table published by East and Hayes (1912, p. 28)
entitled "Condition of hybrids in crosses between species of Xicotiana" because a
number of Nicotiana hybrids were tabled as showing 100 percent germination. This
table was published to indicate the general type of certain hybrids with regard to
vigor, and I think served its purpose. It was distinctly stated, however, (p. 29) that
"the voluminous data that have been collected on these hybrids have been condensed
and approximated so that they include only facts germane to the matter in hand." It
would seem that it might have been clear to Goodspeed from this statement that
these germinations were only classes. Possibly it would have been better to have
said germination "high, medium, low and failing," but it does not seem to me that
the readers were led far astray. As a matter of fact the germinations tabled as 100
percent, included all hybrids that tested over 90 percent.

Genetics 1: Jy 1916

320 E. M. EAST

the latter part of the season. The Fj generation was as uniform as
either parent and showed distinct evidence of hybrid vigor. The plants
varied from 140 cm to 160 cm. In the second hybrid generation there
was no evidence of segregation into distinct types, either as regards
height or general habits of growth but plants varied from below the
height of Xo. 321 (extremes about 100 cm) to that of the Fx generation.

RAPIDITY OF GROWTH

Though both of these species continue flowering until frost. A'. Langs-
dor ffii commences flowering earlier than A', alata. when planted at the
same time, and owing to the multitude of ripe capsules formed, takes
on a more mature appearance in September. The variation in time of
flowering within each species is very slight. In seasons with normal
rainfall, sunlight and heat, plants of A'. Langsdorffii planted in the
greenhouse at the same time and set in the field on the same dav, come
into blossom within three days of each other. Plants of A', alata treated
in the same manner, show greater variation, sometimes a week elapsing
between the time that extremes begin blossoming. The Fx plants are as
uniform in this respect as A'. Langsdorffii and are slightly earlier. The
F2 plants, on the other hand, are more variable than those of A', alata,
and this variability is not wholly an effect of environment as is beauti-
fully demonstrated by the F3 cultures. Ten progeny rows from different
F2 plants showed a difference of 25 days in the time the plants began to
flower. Four of the families were variable like the F.2 population, but
the remainder were very uniform within the family. On the fifteenth of
July two families were in full blossom without an exception, one family
had just begun to bloom, one family had the central stalks well ad-
vanced and two families were in the rosette stage.

LEAVES

In general the shape of the leaves of both of these species is the
same. The basal leaves of A', alata, however, are acute, with redundant,
folded margins, while those of A'. Langsdorffii are obtuse and not re-
dundant at the margins. A'. Langsdorffii is much more rugose than
A', alata. Both species are decurrent. The tips of the leaves of the Fx
plants are intermediate, but in other qualities the leaves are like those of
A', alata. The F.2 plants run the whole gamut of these variations. There
are plants, the rugosity of whose leaves is like A\ Langsdorffii, that are
like A', alata in other respects (except that the flowers are smaller).

~c-~c--cv:-t^.v. ceo
o -r -r cc ^ t s c n .c <-^.
ci in —

- 5

o «n

o -

o M m

vO ci CI m

tj- m t>-
ci O

— ci ci

h no
— >o — c* o

O «o

- Z

- - ~~. — — c

S I I I I I I I I

CI CI CI CI

~ - - _ _ _

~ ~ w ci ri

— — — XXXXXXXXXXX

- - " r x x x x x r. v: v:

JZ — M ri M M d CI C* CI CI CI CI

CI C« Ncocococococncocorocnco

c*} co — • — — ■ — - — — — — - — — —

322 E. M. EAST

Those plants that have returned to the N. Langsdorffii habit and size of
flower, however, always have rugose leaves though they may be acute.

In all of these respects the reciprocal crosses are so nearly alike that
it is impossible to distinguish between them.

COROLLA LENGTH

As I have explained in other papers (East 191 6 a, b), corolla length
is an excellent character upon which to make genetic studies because of
the very slight effect produced by environmental differences. Corolla
measurements of single flowers when taken with due precautions as to
uniformity of age of plant, age of flower, position of flower, etc., well
represent the phenotypes of the plants concerned.

Table 1 gives the frequency distributions of such measurements upon
populations of the pure species, the cross when Ar. Langsdorffii was used
as the female, a single F2 population, the total of several F2 populations
and eight F3 families. Only three of these distributions contain as many
individual measurements as I should like, and one — (328 X 32I) I_5 —
must be discarded entirely on account of the small number of plants.

The statistical constants for these distributions are shown in table 2.

Table 2

Statistical constants for the frequency distributions of corolla length shown in table I.

Pedigree
X umber

Size in
parent

Mean

No.

328, (1914)

21.41

.11

1. 19

.08

5.56

•37

328, (191 1 + 1914)

21-43

.10

1. 17

.07

546

•31

321, (1911)

81.76

•49

5.08

•35

6.21

.42

(328 X 321) Ft

40.78

.22

2.20

•15

5-39

.38

(328 X 321) — 1 F2

3777

-h

•24

5.63

•17

14.91

•45

256

(328 X 321) total F2

38.30

-h

.17

5.99

.12

15-64

.32

58i

(328 X 321) 1 — 1 F3

22.65

.12

I.24

.08

5-47

•37

(328 X 321) 1 — 2 F3

35-44

.15

I.62

-h

.11

4-57

•31

(328 X 321) 1 — 3 F3

39.31

•25

2-54

H-

.17

6.46

•44

(328 X 321) 1 — 4 Fa

52.04

•44

5.52

•31

10.61

.61

(328 x 321) 1 — 41 F3

51.02

•51

6.l6

.36

12.07

•71

(328 X 321) 1 — 5 F3

' 54

49.24

1.09

8.05

■+-

•77

16.35

H-

1.60

(328 X 321) 1 — 6 F3

52.79

•35

6.79

.25

12.86

.48

168

(328 X 321) 1-7F3

21.34

±2

.12

1.24

.08

5.8i

-+-

•39

As may be seen, N. Langsdorffii (No. 328) has a very low variability.
This is to be expected, for N. Langsdorffii is practically always self-fertil-

CROSSES BETWEEN XICOTIANA LANGSDORFFII AXD N. ALATA 323

ized naturally. N. alata is often self-fertilized, but evidence of consider-
able cross-fertilization has been found by observing the actions of
Sphingidae in the evening, by isolating plants, and by self-sterility studies.
Though the coefficient of variability (6.21 ± .42 percent) is almost as

Figure 3 Figure 4

Figure 3. A, N. alata grandiflora; B, Fx of X. Langsdorffii X -V. alata grandiflora;
C, N. Langsdorffii (1911) X V2.
Figure 4. D and E, extremes of the F2 generation (191^) X V2.

low as that of Xo. 328, therefore, it is probable that No. 321 is not so
nearly homozygous. Furthermore, the number of individuals measured
is small. On the other hand, since a single plant of No. 321 was used in
the cross, it is possible that the true variability of this "blood" intro-
duced, is somewhat smaller than that represented by the frequency
distribution.

Curiously enough the mean of the Fx population is smaller than the
average of the two parents. Thus there apparently is no effect of
heterosis on the flowers. The square root of the Fx mean is more nearly

Genetics 1: Jy 1916

324 E- M- EAST

that of the average of the square roots of the means of the two parents,
but I do not feel justified in attaching any significance to the fact.

The coefficient of variation of the F2 generation is nearly three times
that of the F1 generation. Though extremes like each parent were not
produced, it is hardly possible to see any other cause for this great
difference in variability than segregation and recombination of Mendel-
ian factors. From the theory of probability one might expect to recover

Figure 5. F, extreme of the F:; generation, and G. pure A*, alata grandiflora (1913)

X V2.

both parents with a comparatively small number of F2 plants, but the
variability of F1 is so small that even the plants obtained in F2 could
not be expected in the F1 if the whole of Xew England were planted
with them.

In the F3 generation there was regression toward the mean of the F2
population in six out of seven cases ( excluding Xo. 1-5 on account of the
small number of plants), but the greatest extremes gave the least regres-
sion. The coefficients of variability were lower than that of F2 in every
family, and three of them bred as true as the parental species. F3 family
Xo. 1-7 reproduced N. Langsdorffii exactly.

The Mendelian theory calls for the production of the same type of
F2 population no matter what F1 parent is selected, when the original
individuals entering the cross are homozygous. Critics of the use of
the Mendelian terminology in crosses involving size characters have
maintained, however, that small F1 individuals will give F2 populations

Figure 6 Figure 7

Figure 6. Individual plant produced by an F2 segregate that was like Ar. Langs-
dor ffii in every detail (E of figure 4).

Figure 7. Progeny row showing uniformity of F3 family to which the plant shown
in figure 6 belonged.

with lower means than will large individuals. As there are many indi-
cations that the plants entering into this cross were very nearly true
homozygotes I have endeavored to test this proposition. Of course, as
might be expected by pro-Mendelians, in such a cross the variability of
the F1 population is so low that the extremes selected differed by only
8 mm. Nevertheless five F2 frequency distributions from different
F1 parents are presented in table 3. The statistical constants shown in
table 4 emphasize the fact that the means and the standard deviations

Genetics 1: Jy 1916

0
OS

3 c

- -1-

N -

i/3 —

X it. X X

vl N X CO O

in r o nx

•O i- 01 ^

X X X X X

CROSSES BETWEEN NICOTIAN A LANGSDORFFII AND N. ALATA 327

Table 4

Statistical constants of the corolla length in the five F2 families reported in table 3

Number

Size in
parent

Mean

No.

(328 X 321) — 1

37.77 ± .24

5.63

± .17

14.91

-+-

•45

256

(328 X 321) — A

37-55 ± -36

5.65

± .26

15.05

109

(328 X 321) — 2

39-73 ± .52

6.32

± -37

15.91

•95

(328 X 321) — 3

38.21 ± .40

5.21

± .28

13.63

•75

(328 X 321) — 4

40.08 ± .56

7.11

± .40

17.24

± 1

.00

very nearly overlap. In other words the curves are very nearly identical,
. and it can be shown mathematically that the probability is very high that
they are all samples of the same population. The similarity of the
curves is shown graphically in figure 9. The points of the theoretical
curves of these five F2 populations were calculated and are shown in

Figure 8. A, a random sample of N. Langsdorffii flowers from six different plants ;
B, a random sample of flowers from twelve different plants from the progeny row
shown in figure 7 (X JA).

comparison with the theoretical curve of the total distribution of all F2
observations. The extreme classes are so nearly identical that curves
could not be distinguished when drawn super-imposed, so that only the
points are indicated. Where no points for a particular curve are given it
is understood that they lie on the single curve of total observations
which is drawn.

Genetics 1: Jy 1916

328

E. M. EAST

Certainly no one can well maintain that these curves show any de-
cided difference when the probable errors are taken into consideration.
There is no dissimilarity in variability like that shown by F3 populations
from different points on the F2 curve. The mean of the population from
the 36-mm parent is higher than that from the 44-mm parent. This
fact is not to be taken as significant; it is merely a coincidence. It is
very evident that the only just conclusion is that selection has no effect.

Figure 9. The points of the theoretical curves for flower length, calculated from
the data obtained from growing F2 populations from iFx individuals of various sizes in
cross 328 X 321. The curve drawn in full is that calculated from the combined
observations.

In tables 1 and 2 one may notice an F3 family, No. (328 X 321)1 — 7,
that seems to have repeated the small parent, N. Langsdorffii. This
was indeed the case. Several F2 plants duplicated N. Langsdorffii in
every feature and two of them were selfed and their progeny grown.
Family No. (328 X 321)1 — 1 from the larger of these two plants as

CROSSES BETWEEN NICOTIANA LANGSDORFFII AND N. ALATA 329

regards flower size was very uniform, but the flowers were slightly
larger than those of the N. Langsdorffii strain used, and the plants dif-
fered from it slightly in other ways. In short, it could hardly be main-
tained that the great-grandparent had been duplicated. On the other
hand family (328 X 321)1 — 7 was exactly like a family from a selfed
plant of No. 328. I could not find a distinguishing trait by the most
minute examination. Figures 6, 7 and 8 show this fact plainly, but since
photographs of the other small family would appear to show the same
thing, so small are the differences between it and our strain of No. 328,
a table of corolla measurements has been introduced in order to demon-
strate the matter quantitatively.

Since it is obvious that the use of 3-mm classes in table 1 — classes as
small as can be treated conveniently in connection with such great size
difference — obscures somewhat the true distribution of the corolla length
of No. 328, table 5 gives the distribution of the corolla lengths of the

Table 5

Frequency distribution of No. 328 and of an F3 family that bred true to the
characters of this species.

Pedigree
Number

Class centers in
millimeters

Mean

S. D.

C. V.

No.

19 20 21 22 23

328, (191 1)

3 12 1 2

328, (1914)

1 9 33 7 1

20.96 ± .06

0.69 ± .05

3.27 ± .21

328, total

1 12 45 8 3

21.00 ± .06

0.72 ± .04

3.43 ± .20

(328 X 321) 1 — 7 F3

11 33 6

20.00 ± .05

0.57 ± .04

2.75 ± .19

same plants in i-mm classes. This brings out the wonderful uniformity
of the populations of both No. 328 and No. (328 X 321)1 — 7 and the
marvelous similarity between the two families. Furthermore, it shows
how similar are two populations of No. 328 grown from the same seed
but in different years.

THE RECIPROCAL CROSS

The cross in which No. 321 was used as the mother, was not a true
reciprocal of the other in that the same individuals were not used. In
fact a different strain of N. Langsdorffii known as No. 328 — 1 was used,
which had flowers slightly smaller than No. 328. For this reason as well
as that each generation of this cross was grown a year later than the

Genetics 1: Jy 1916

33°

E. M. EAST

MO1-!

N O 00

00 - y-»
h n

o o

N 00 00

N ~ - ~

N m if

1) &>

W> "2

Ls. -i.

ta-fe" III

S-fl— I I I I I I

O - 00 00 M 00 00 00
' — ^ M M M CI C) C)

f i ^xxxxxx

ro >I2 >I2 «£2 >I2 >I2

332 E. M. EAST

other, the slight differences between the two can not be attributed to
the different way of making the cross. The cross with No. 321 as the
mother was more difficult to make but this is probably due to the greater
length of the style of No. 321.

Cross 321 X 328 is notable for the extreme uniformity of the first
hybrid generation and the great increase in variability in the second hy-
brid generation, as is shown in tables 6 and 7. N. Langsdorffii was again

Table 7

Statistical constants for the frequency distributions of corolla length shown in table 6.

Pedigree
Number

Size in
parent

•M

can

328 — 1, (1910)

19.40

•13

1.02

-+-

.09

5.26

328 — 1, (1912)

19.23

.19

1.42

.13

7.38

.69

328 — 1, total

19.32

.11

1.23

.08

6.37

•41

321, (1911)

81.76

•49

5.08

•35

6.21

.42

(321 X 328—1) F1

42.42

•19

1.60

.14

3-77

•32

(321 X 328—1) F2

3779

.28

5.36

.20

14.18

•54

163

(321 X 328— 1) 1 —

1 F8

19.30

.IO

I.50

.07

7-77

•37

101

(321 X 328 — 1 ) 1 —

2F3

43.63

.28

3-73

.20

8.55

•45

(321 x 328 — 1 ) 1 —

3 Fa

45-34

•32

4.80

.22

10.59

105

(321 X 328—1) 1 —

4F3

44-52

•32

3.85

.22

8.65

.50

reproduced in F2 and plant (321 X 328) 1 — 1 bred true to its characters.
There was no nearer approach to No. 321, however, than there was in
cross 328 X 321. The cross appeared to be fully fertile and the seeds
germinated well though in general not so perfectly as those of the reverse
cross. It does not seem as if the slight infertility shown, however, could
be the explanation of the failure to reproduce the larger parent.

Again the coefficients of variability of the four F3 families grown are
below that of the F2 generation. Considering them together with the
other later generations previously reported it would seem as if the case
for Mendelian inheritance were pretty clearly proven.

CONCLUSION

A fertile cross between two distinct species, Nicotiana Langsdorffii
and Nicotiana data grandiftora, each uniform in its characters, has been
reported here with the following results, no matter which way the cross
was made.

(CROSSES BETWEEN NICOTIAN A LANGSDORFFII AND N. ALATA 333

1. The F1 populations are as uniform as the parents.

2. The F2 generations are nearly three times as variable as the Fx
generations.

3. Individuals reproducing -the smaller species were found in the F2
generation.

4. Certain of these F2 individuals reproduced N. Langsdorffii popula-
tions in the F3 generation.

5. No F2 individuals reproducing N. alata grandi flora were found,
but F3 plants approaching such a type were produced.

6. Galtonian regression occurred, but selected extremes regressed no
more than those deviating moderately from the parental mean.

7. Individuals from the same point on the F2 curve showed different
variabilities in F3.

8. The variabilities of F3 families were invariably smaller than those
of F2 families.

9. The above conclusions are based upon corolla length measurements
but apparently are true for other characters, except that in other char-
acters, N. alata grandiflora types were reproduced.

10. Mendelian inheritance of corolla color and pollen color is shown.

11. Mendelian inheritance seems to be the only logical interpretation
of the other phenomena.

LITERATURE CITED

Comes, O., 1899 Monographic du genre Nicotiana. Naples : Typographic Cooperative.
Dorsey, M. J., 1915 Pollen sterility in grapes. Jour. Heredity 6 : 243-249.
East, E. M., 1916 a Significant accuracy in recording genetic data. Amer. Jour. Bot.
5: 211-222.

1916 b Studies on size inheritance in Nicotiana. Genetics 1 : 164-176.
East, E. M., and Hayes, H. K., 1912 Heterozygosis in evolution and in plant breeding.

U. S. D. A. Bur. Plant Ind. Bull. 243, pp. 58.
Focke, W. O., 1881 Die Pf lanzenmischlinge. 569 pp. Berlin : Borntraeger.
Godron, D. A., 1863 Des hybrides vegetaux considered au point de vue de leur fecondite

et de la perpetuite ou non-perpetuite de leurs caracteres. Ann. sci. nat, Bot.

Ser. IV, 19 : 135-179.

Goodspeed, T. H., 1913 On the partial sterility of Nicotiana hybrids made with N.

sylvestris as a parent. Univ. Cal. Pub. Bot. 5 : 189-198.
Lock, R. H., 1909 A preliminary survey of species crosses in the genus Nicotiana from

the Mendelian standpoint. Ann. Roy. Bot. Gar. Peradeniya 4: 195-227. PI.

18-29.

Naudin, Ch., 1865 Nouvelles recherches sur l'hyforidite dans les vegetaux. Nouv. Arch.
iMus. Paris 1 : 25-176. PI. 9.

Genetics 1: Jy 1916

Hidden
Feeblemindedness

E. M. EAST

Reprinted, without change of paging,
from the Journal of Heredity (Organ of the
American Genetic Association), Vol. VIII,
No. 5, Washington, D. C. May, 1917.

HIDDEN FEEBLEMINDEDNESS

One Person in Fourteen of the American Population Probably Carries the Trait in
a Recessive Form, Although Normal to all Appearances — One-Fourth
of Offspring will be Feebleminded if Mating is Made
with Another Carrier

E. M. East
Bussey Institution, Forest Hills, Mass.,

THE increase in the number of
feebleminded in the United States
during the past few years has
been such that the heredity of
the trait, and the classification and
treatment of those so afflicted, have
been the subject of much careful study.
The result of this activity has been
very creditable. Thanks to the re-
searches of Goddard, the method of
inheritance of feeblemindedness is as
clear as that of any other heritable
variation in the human race. Owing
to the ingenious psychological methods
of Binet and Simon, the grade of
mentality can be determined reasonably
well. Even our slowly moving legisla-
tive bodies have been somewhat dis-
turbed by the facts and have passed a
considerable number of laws designed
to cut off this defective germplasm.
either through segregation of the sexes
during the reproductive period or by
sterilization.

One can have only words of com-
mendation for the serious efforts to face
the problem; nevertheless, in the nu-
merous papers on feeblemindedness that
have been published during the last
decade, not a single author appears to
have appreciated the real menace. Our
modern Red Cross Knights have
glimpsed but the face of the dragon.

Goddard has shown that feeble-
mindedness is transmitted as a Mendel-
ian recessive. In other words feeble-
minded individuals may be produced
in three ways. If feebleminded mates
with feebleminded all of the offspring
will be feebleminded. If a feeble-
minded individual mates with one carry-
ing the trait in his or her germcells, on
the average one-half of the offspring

will be feebleminded. It is these two
types that segregation or sterilization
will affect. But these are not the only
sources of feeblemindedness, and per-
haps they are not the most dangerous.
If two carriers of feeblemindedness
mate, one-quarter of their offspring will
exhibit the trait and one-half of them
will transmit it. Let us endeavor to
see what this means.

THE NUMBER AFFECTED

It appears that in our present popula-
tion of 100,000,000 or thereabouts,
there are 300,000 persons who are
feebleminded through an hereditary
defect, a ratio of 3 per 1,000. This is
an estimate to be sure, but it is s< i
conservative that it probably veils the
true state of affairs.

Now how many of these defectives
have been the result of a mating
wherein at least one of the parents was
feebleminded? This question is a dif-
ficult one and can only be answered
with a rough approximation. The best
estimate that I can make from a careful
examination of the meagre statistics at
present available is 100,000. The dose
must not be too bitter, however, so let
us double this estimate. This leaves
100,000 feebleminded persons that must
have been produced by the mating of
two transmitters of feeblemindedness
who did not show defective mentality
themselves, unless an unprecedented
percentage of origin de novo be assumed.

These 100,000 defectives were pro-
duced during a period in which there
were rather less than 20,000,000 married
couples of reproductive age. They
were produced by parents both of which
carried feeblemindedness. .But only

215

216

The Journal of Heredity

one-fourth of the progeny of such
matings show feeblemindedness. There-
fore, at least 100,000 couples of this
type . were reproducing during this
generation. This would presuppose the
survival of four children per couple
long enough to have their mental status
determined, an assumption that would
probably require a total reproductivity
of seven children per married pair.
Among the children from these matings
would be some 200,000 carriers of de-
fective germ-cells, but we will omit
them from our considerations. The
important point is that out of 20,000,000
pairs of married persons, if we treat the
problem as static, 100,000 were trans-
mitting feeblemindedness. What then
is the number of such persons in the
population?

Let us state the question in another
way. A certain number of persons
out of a population of 40,000,000 of a
marriageable age have defective germ-
cells. If two of them marry, one-quarter
of their children will be feebleminded.
If 100,000 of such marriages did occur,
what is the ratio of carriers of feeble-
mindedness to normals in the general
population? The correct answer will
depend of course upon how much selec-
tive mating takes place. There is un-
questionably a general tendency for
carriers of feeblemindedness to be
brought together and a marriage to
result. But this cannot be taken into
account very accurately and had best
be left out of our calculations.

Pairing among carriers of feeble-
mindedness has occurred in the ratio
of 1 to 200 marriages; then, if no selec-
tive mating has taken place, carriers
of feeblemindedness must occur in the
general population in the ratio of 1
to 14.

One-fourteenth is approximately the
square root of 1/200. If 1/14 of the
population carry feeblemindedness and
13/14 are normal, then the probability
of normal mating with normal is 13/14X
13/14= 169/196, the probability of nor-
mal mating with carriers of feeblemind-
edness is 1/14X13/14+13/14X1/14
= 26/196, and the probability of two
carriers of feeblemindedness mating is
1/14X1/14=1/196.

Possibly this figure is somewhat too
high for the single trait feebleminded-
ness. We have not corrected for
changes in the population during the
length of the period considered or for
selective mating. But, to balance this
we have used a low estimate of the
number of feebleminded, a high esti-
mate of the number of defectives pro-
duced by parents of which at least one
exhibited defects, and a high birth-rate
in families of those transmitting the
defect. Further, no mention has been
made of epilepsy and of certain types of
insanity, which are inherited in the
same way, and to which the same line of
reasoning applies. In view of these
facts it is probable that the conclusion
that 1 person out of every 14 carries
the basis of serious mental defective-
ness in one-half of his or her reproduc-
tive cells understates rather than over-
states the facts.

The problem of cutting off defective
germ-plasm, therefore, is not the com-
paratively simple one of preventing the
multiplication of those so affected.
This task, though sufficiently difficult in
practice, is possible: the way has been
pointed out; something has been ac-
complished. It is rather the almost
hopeless task of reducing the birth-
rate among transmitters of serious
defects.

NEED FOR RESEARCH

A stupendous task necessitates pro-
digious efforts. Already there is a
tremendous selective birth-rate in favor
of lesser civic worth, and it is extremely
doubtful whether, under our present
economic system, much can be accom-
plished by recommending early marri-
ages and large families among those
whose accomplishments have proved
their social value. Whether family
limitation among those carrying defec-
tive germ-plasms can be effected must
be decided in the future. It will be a
distant future if a stupid government
persists in refusing to countenance
rational parenthood among those least
fitted to reproduce the race, the while
shutting one eye and winking the other
at what has become a national practice
among those best fitted to build a

East: Hidden Feeblemindedness

217

greater America. There is one sugges-
tion, however, at which no one will
cavil. We have assumed that a normal
mentality is completely dominant over
a defective one. Is this true? Com-
plete dominance is rare among those
characters commonly studied by animal

and plant geneticists. Is it not likely
that the Binet-Simon or other proper
tests would show that carriers of mental
defects exhibit a lower mentality than
pure normals ? Would it not be wise to
start some investigations along this
line?

The Explanation of
Self-sterility

E. M. EAST

Reprinted, without change of paging,
from the Journal of Heredity (Organ of the
American Genetic Association), Vol. VIII,
No. 8, Washington, D. C, August, 1917.

THE EXPLANATION OF SELF-STERILITY

E. M. East, Bussey Institution, Forest Hills, Mass.

IN a recent paper by C. W. Moore1 on
the subject of self-sterility, several
ill-advised statements were made to
which attention should be called.
The paper begins with the sentence :

Several who have made a study of the
problem of the inheritance of self-sterility of
plants have obtained results which did not
point to any one definite manner in which
flowers act when self -pollinated.

One might read into the meaning of
this statement either that there was
great difference of opinion regarding the
behavior of self-sterile plants or that
little was known regarding self -sterility
before the appearance of the paper under
discussion. As a matter of fact a great
many details regarding self-sterile plants
are known. Darwin dealt with the
matter at some length, and more
recently extended researches by Jost,
Correns, Compton and Stout have
appeared. The present writer has also
investigated the subject rather minutely
although only preliminary reports of
the work have been published. As to
the gross facts, there is not a great
difference of opinion among the later
writers. Each has found that pollen
grains germinate after self-pollination
as readily as they do after cross-pollina-
tion, but that they grow more slowly,
and the present writer has determined
that the growth curves of self-pollen
tubes are approximately straight lines,
while growth curves of cross-pollen
tubes are similar to those of auto-
catalytic reactions. Each has found
that there is cross-sterility of the same
nature as self -sterility. In other words,
the plants of a self-sterile race are not
only self -incompatible, but some com-
binations are cross-incompatible. The
differences of opinion come in inter-
pretation of these results, and these
differences are due largely, we believe,
to the fragmentary character of the
evidence.

Moore founds an hypothesis by which
to explain self -sterility on the supposed
fact that self-tubes are greater in diam-
eter than cross-tubes. In fact this
seems to be the main thesis of his paper.
He says :

. . . the greater width of the self-pollin-
ated pollen tubes of Tradescantia is due to
the fact that the food supply is more favorable
to the nourishment of a self-pollen tube than
it is to a cross-pollen tube. On account of
the abundant food supply the pollen tubes
did not lengthen, but grew wider since they
were in a very favorable medium. By this
hypothesis it is possible to explain most of
the data here presented. . . .

What Moore did was to measure short
self-pollen tubes and long cross-pollen
tubes as he distinctly states on page 204.
Now if he had measured self-pollen
tubes and cross-pollen tubes of the
same length, as he should have done,
it is almost certain that he would have
found them to be of the same width.
At least this is the observation of the
writer on numerous pistils of three
different self -sterile species of Nicotiana.
Moore's main thesis, therefore, seems
to be based upon an improper observa-
tion.

The second point made in the paper,
involving a criticism of the present
writer, is similarly without foundation
He says:

He [East] states that "all gametes having in
their hereditary constitution something differ-
ent from that of the cells of the mother plant,
however, can provoke the proper secretion to
stimulate the pollen tube growth, reach the
ovary before the flower wilts, and produce
seeds." From this it may be inferred that
there may be an enzyme in the pollen grain
that in a cross-pollination is able to induce the
stigma to excrete a stimulating substance so
that the pollen tube is able to grow. In a
self-pollination this enzyme is not able to act.
However, if this were the case, when a few
cross-pollen grains were placed on a self-
pollinated stigma, they would be expected to
germinate and cause the stigma to produce the
stimulating substance. Thus the pollen tubes
from the self-pollination would also benefit
by the stimulating influence and should be

1 Journal of Heredity, viii, 203-207, 1917.
382

[Reprinted from The American Naturalist, Vol. LI., March, 1917.]

THE BEAEIXG OF SOME GEXEEAL BIOLOGICAL
FACTS OX BUD-VAEIATIOX1

PROFESSOR E. M. EAST
Bussey Institution. Harvard University

I take it no one denies that in the Angiosperms vari-
ations may be produced in connection with reproduction
by means of buds and that these variations may be per-
petuated by the same method. Practically, as horticul-
turists and plant breeders, we care little about the occur-
rence of budrvariations elsewhere in the organic world.
Xevertheless, it may help in the orientation of our ideas
if we remember that budding is not a rare or unconven-
tional method of reproduction. In a generalized form,
the earliest method, it has persisted throughout the plant
kingdom from the most primitive to the highest and most
specialized types. Sexual reproduction has not replaced
it, but has been added to it. Even in the animal kingdom,
though eliminated among the higher forms, it still exists
as an occasional alternate method in three fourths of the
phyla. Such being the case, it would seem logically to
follow that variation must have been within its possi-
bilities.

The cause, the frequency, the type, the constancy, the
mechanism, of these variations are more debatable, how-
ever, and on these questions many biological facts which
superficially seem unconnected, have a direct bearing. In

iRead before the meeting of the Society for Horticultural Science, De-
cember 28, 1916.

129

130

THE AMERICAN NATURALIST

[Vol. LI

fact, on certain phases circumstantial evidence is the only
evidence at hand.

The exact nature of the cause or causes of bud- variation
can hardly be discussed profitably. We may imagine
irregularities of cell division directed by combinations of
unknown factors, but to describe these factors in concrete
terms is at present impossible. At the same time, cause
can not be neglected entirely even at present, for cause in
a generalized sense is intimately connected with frequency
in that vigorous perennial the question of the inheritance
of acquired characters. The data on this subject are so
voluminous that each for himself must give them careful
conscientious consideration. Here no more can be done
than to point out some of the conclusions to which I, per-
sonally, have been driven, and their connection with the
subject in hand. These conclusions are:

1. Broad and varied circumstantial evidence indicates
unmistakably that the inheritance of acquired characters
has played an extremely important role in evolution.

2. Numerous experimental investigations designed to
test the possibility of such inheritance directly have either
failed utterly or have been open to serious destructive
criticism. Direct proof of the inheritance of acquired
characters is therefore lacking.

3. If conclusions 1 and 2 are to be harmonized, either
modifications are fully inherited so rarely that proof that
they do not belong to the general category of chance
changes in constitution of the germ-plasm is impossible,
or the imprint of the environment is so weak that ex-
tremely long periods of time— perhaps geological epochs
— are necessary for its manifestation.

Diametrically opposed views on the inheritance of ac-
quired characters are held tenaciously and unequivocally
by equally eminent biologists. Those who concur with
the Lamarckian position are nearly always the students
of evolution who approach the subject from the historical
or the philosophical side and who rely almost entirely
on circumstantial evidence ; those who adhere to the side

No. G03]

BUD-VARIATION

131

of Weismann are usually experimentalists whose evi-
dence is indeed direct, but often questionable, usually
capable of various interpretations, and always fragmen-
tary. I have been bold enough to grasp both horns of
the dilemma, and to plead that each is right from his point
of view. My confession of faith is, the environment has
been an immense factor in organic evolution, but its
effects are shown either so infrequently or after the elapse
of so great a time, that for the practical purposes of
plant breeding we can neglect it as we would neglect an
infinitesimal in a calculation. As Bergson, I think,, said :

We have been trying to prove that the hour hand moves, in a second,
of time.

A few words will make clear the general arguments in
favor of this position, although adequate support to the
thesis would require considerable time.

In the first place, it seems to me the possibility of the in-
heritance of acquirements must be admitted. Weismann 's
general contention that the chromatin of the germ-cells
is the actual hereditary substance, and that the germ-cells
themselves may be regarded as one-celled organisms re-
producing by fission and conjugating at certain times,
while the body must be considered simply an appendage
thrown off from and independent of the germ-cells, is not
supported merely by the embryological researches of
Boveri, Kahle and Hegner on two or three animal forms,
or by the ingenious ovarian transplantations made by
Castle and Phillips on guinea pigs, but by all of the recent
pedigree culture and cytological genetic work, botanical
as well as zoological. Nevertheless it has not been and
logically can not be proven that there is no way for en-
vironmental forces to produce germ-plasmic changes.
Memory is just as strange a phenomenon and Semon has
done biology a service by pointing out the analogy be-
tween the mechanical requirements for memory and for
the inheritance of somatic modifications.

This possibility being admitted, one may well concede
the plausibility of the arguments of the numerous pale-

132

HIE AMEBIC AX XATURALIST

[Vol. LI

ontologists, taxonomists and ecologists in favor of La-
marckian principles, in spite of the fact that their evi-
dence is circumstantial. They take a comprehensive view
of the actual conditions that exist among organisms,
which is impossible to the experimentalist. It will not do
simply to say that the manifest convergence of analogous
organs in all parts of the organic world, or the wonderful
adaptations of the social insects may be explained in some
other way. Of course there may be other explanations
for these phenomena; but until more satisfactory ex-
planations are forthcoming it is rightfully a custom in
science that the adequate interpretation at hand should
be accepted.

On the other hand it is equally wrong for the ardent
devotees of Lamarckism to clutch at every isolated case,
every inadequate and abortive experiment, when judicial
consideration shows not a single unassailable instance of
the inheritance of a somatic modification. Many of these
experiments have a direct bearing on bud-variation, and
I shall attempt to show where they lead us.

1. Inheritance of Mutilations. — The most radical La-
marckians of the present day only go so far as to sup-
pose that mutilations are inherited on very rare occa-
sions—and they are always zoologists. Ethnology has
furnished us with so many histories of mutilations of
ears, of lips, of feet, of reproductive organs, long con-
tinued in the folkways of a people, that new laboratory
experiments have been deserving of the ridicule they
have received. Botanists have seldom had any delusions
on the subject. Plants are so continually mutilated in
the buffetings they receive during life, with no result in
the next generation, that the non-inheritance of the effects
of such injuries is taken as a matter of course. Yet there
is occasionally one whose reason fails at the critical
moment, and who holds that cuttings from the chrys-
anthemum with the large flower resulting from the re-
moval of lateral branches, will produce larger flowers in
the next generation than will an untreated sister plant.
If not this, some equally indefensible doctrine.

No. 603]

BUD-VARIATION

133

2. Effects of Changed Food Supply.— This last ex-
ample was really one of changed food supply induced by
mutilation. Change of food supply by other methods has
been the basis of scores of experiments, particularly on
insects. Many insects are so very whimsical about what
they eat that it seems possible their selective appetite
may be an inherited instinct impressed by the environ-
ment of countless generations. But the total result of all
experiments on them is merely to prove that a second
generation may be influenced in the start they get in life
by the nutrition of the mother.

The same thing is true in plants. We fertilize a pop
corn to get a bumper crop of good plump healthy seeds,
but we don't expect a dent corn as the next year's result.
We very properly endeavor to give our potatoes a bal-
anced ration, in expectancy of a larger yield of well-
matured, healthy tubers, but we should not expect these
tubers to affect our next season's supply other than by
their health. Similarly we take scions from well-lighted
parts of the tree where growth has been good. In such
twigs the graft union heals easily and properly, and a fit
channel for conveying nutrients is established. In doing
these things we are practising sanitation or preventive
medicine, as it were, a laudable proceeding. But the hor-
ticulturist who promises a different variety by such
means is illogical and misleading.

Yet we find Bailey so imbued with the idea of making
out a perfect case for Lamarckism that he lends the
weight of his authority to the following statement among
others :2

Whilst these ' 1 sports ' ' are well known to horticulturists they are generally
considered to be rare, but nothing can be farther from the truth. As a
matter of fact, every branch of a tree is different from every other branch,
and when the difference is sufficient to attract attention, or to have com-
mercial value, it is propagated and called a " sport. "

We may admit the differences between the branches of
a tree without cavil. What is more serious is the impli-

2 " Survival of the Unlike,' ' p. 72.

134

THE AMERICAN NATURALIST [Vol. LI

cation to the reader that all variations have the same co-
efficients of heredity, that a bud-variation is simply a
wide fluctuation imposed by external conditions. If this
were true the whole organic world would be chaos. But
species and varieties do exist. They may be " judgments ' '
in one sense, but in another they are concrete things. In
fact we learn this further on in this volume when it suits
Bailey's purpose to have asexually propagated varieties
very constant. He says (p. 353) :

At first thought this fact — that varieties may be self-sterile — looks
strange, but it is after all what we should expect, because any variety of
tree fruits, being propagated by buds, is really but a multiplication of one
original plant, and all the trees which spring from this original are ex-
pected to reproduce its characters.

3. The Effects of Disease.— The influence of disease is
in many ways like that of malnutrition, in that it is wholly
an effect on the physiological efficiency of the reproducing
cells. This fact is fairly clear when dealing with diseases
with outstanding symptoms. In many instances, how-
ever, diseases are not easily diagnosed. There may even
be no suspicion that disease is present. In such cases it
is rather hard to believe that selection is not accomplish-
ing a positive and radical improvement. A good ex-
ample of this is the selection of potato tubers. No one
consciously selects a seed potato infected with blight. In-
dependent of the probability of reinfection, there is the
likelihood that the diseased tuber will not be able to pro-
duce a normal plant because of the effect the fungus has
had on its own cells. One doesn't usually believe, how-
ever, that rejection of this tuber and selection of the
healthy sister is going to lead to the formation of a new
race. Yet numerous experiments on potatoes in which it
is shown that successive selections have raised the
average yield over that of the unselected tubers, are prob-
ably of just this type. The race is kept up by the re-
jection of diseased tubers, but there is no evidence what-
ever that it is unproved. I am not going to argue that
desirable asexual variations may not occur during this
time, and be retained. I say only that any improvement

No. 603]

BUD-VARIATION

135

indicated by the raw data must be discounted by the
amount of deterioration shown by the unselected variety
under similar conditions. Such deterioration is very
common, and is due to disease, I believe, rather than to
any supposed disadvantage of asexual reproduction
per se.

This category of facts has been cited under the discus-
sion of the inheritance of acquired characters, because
such phenomena have perplexed other than botanists.
Belief in the transmission of disease, or the effects of dis-
ease, by sexual reproduction was current for many years.
It is only since the possibility of infection in the egg
itself was demonstrated for various diseases, that the
true state of affairs has been known.

Many other types of experiments designed to demon-
strate Lamarckism might be cited, but they have no direct
bearing on bud-variation except in so far as a positive
case would affect our general attitude on the frequency
of their occurrence. They are all similarly negative or
questionable, however, so that we must conclude with
Weismann that no case of inheritance of acquirements
has been proved beyond a reasonable doubt. In other
words we grant such a possibility but believe it to be so
rare or so gradual that practically it may be disregarded.

In reality one could hardly have expected any other
conclusion from the type of experiment by which the
question has been attacked. Generalized they are some-
thing like this. Species X having been grown under en-
vironment A for numerous generations is removed to en-
vironment B. An adaptive change occurs which persists
during several generations. Later the descendants of
the original plants are returned to environment A and
the change is reversed. When the reverse change occurs
more slowly than the original change, it is argued that
Lamarckian inheritance is shown. The logic used to draw
such a conclusion is indefensible, even if the difficulty of
correcting properly for changes due to normal heredity
is left out of consideration.

If acquired characters are inherited and the changes

136

THE AMERICAN NATURALIST

[Vol. LI

induced are reversible, the long period under environ-
ment A should have produced a deep impression on
species X. Change under environment B should be slow.
Reversal should be rapid, however, because of the slight
impression environment B must be supposed to have
made during the very few generations in which its influ-
ence was possible.

If acquired characters are not inherited, precisely the
same changes should occur, owing to somatic adaptation,
the only differences being that the total amount of change
in each case would be reached in the second generation
after the environment had acted during the earliest stages
of the life history.

If, on the other hand, the changes induced by environ-
ment B are not reversible, judgment must be based on
the percentage of individuals changed by B and not re-
changed by A. One can readily see how a just judgment
would be clouded by probable reversible somatic effects
in such cases. Instances of the inheritance of acquire-
ments, unless they were very frequent, which from our
general evidence is unthinkable, would be indistinguish-
able from ordinary chance variations.

Such methods of attack on the subject being almost
predestined to failure from the inherent difficulties of
the problem, it would seem wiser to seek for a more hope-
ful methodology, and in the meantime to accept the only
conclusion justified by the data at hand; namely, the
inheritance of acquired characters is either so rare an
occurrence or so slow a process, that by plant-breeders
it may be assumed to be non-existent. One realizes of
course that the problem of sexual transmission of somatic
acquirements is not necessarily the same as that of asex-
ual transmission, but the experimental results have been
the same in both cases. Let us, admit, therefore, that
one can not hope to obtain real improvement in asexually
propagated varieties merely by selecting buds from
plants or parts of plants which have developed under
especially favorable conditions.

This does not mean that radical environmental changes

No. 603]

BUD-VARIATION

137

may not be the direct cause of such a modification. Dr.
H. J. Webber once informed the writer that immediately
after the great Florida freeze of the early nineties bud-
variations in the citrus fruits of that region were greatly
increased. Such variations may have been induced by
the freezing, but they were not adaptive variations.

The conclusions reached thus far have not involved a
point of theory which practically is difficult to separate
from the one just discussed. It is this. If we disregard
adaptive variations, is there not still a reason for select-
ing fluctuations? Are there not internal factors which
so act that there is a narrow but appreciable variability
in an asexually produced population which may offer a
basis for selection? In other words, how constant is an
asexually propagated race?

We can make an effort to compute the frequency of
marked bud-variations. But have we any right to assume
that these represent the sum total of all bud-variations?
Are not bud-variations and perhaps all inherited vari-
ations like residual errors, the small ones frequent, the
large ones rare? This may be the case, but I should like
to emphasize the fact that we have no true criterion for
determining the size of a variation. A variation that ap-
pears large by visual criteria may be an extremely small
change in the constitution of the plant, and vice versa.
In view of this fact together with the practical consid-
eration that commercially valuable variations must be
measurable within a reasonable duration of time— say a
lifetime— it is by no means certain that we are going far
astray in calculating the frequency of bud-variations by
the so-called marked jumps or mutations.

Furthermore the range of the fluctuations of asexually
propagated varieties of most species is very small even
when broadened— as it always is— by the addition of the
effects of variable external conditions. It is not hard to
recognize a Winesap apple, a Clapp's Favorite pear or a
Concord grape, even though these varieties have been
grown extensively for a considerable number of years.
Certain local subvarieties of the pome fruits are said to

138

THE AMERICAN NATURALIST [Vol. LI

exist, but they are so extremely rare that one may admit
all cases of disputed origin and still have very little
asexual variation to account for.

I have never seen a published calculation of the fre-
quency of bud-variation, and presume it would be of little
value anyway, since the general evidence indicates a dif-
ferent frequency for different species and even for the
same species at different times. It may be mentioned,
however, that in personal examination of over 100,000
hills of potatoes belonging to several hundred varieties,
12 definite bud-variations have been seen, a frequency of
1 in 10,000; while just as careful a scrutiny of about
200,000 plants belonging to the genus Nicotiana has
brought to light but 1 case.

Probably a more practical and just as satisfactory an
estimate of the frequency of bud-variations in economic
plants is the record of varieties that have been produced
in this manner. Naturally such a record contributes little
to theory because only a portion of the variations arising
are observed, and only a fraction of those observed are
propagated. Further, the origin of comparatively few
commercial varieties is known. Yet we may get some idea
of what to expect in the future, by noting what has oc-
curred in the past.

Data, gathered in this manner will appear to give us
different values depending on how we approach the
matter. For example, in Cramer's wonderful monograph
on bud-variation, the grape is cited as one of the species
that often varies in this manner. He cites some 25 or
more sneh varieties. Yet in the lar^e list of American
grapes in Hedrick's " Grapes of Xew York" only one
doubtful case of bud-origin is reported. When one re-
members that hundreds of varieties of grapes are grown
and millions of vines are examined each year, improve-
ment by this method seems rather hopeless. And ex-
amination of the list of present-day apples, pears, plums
and cherries, of the bush-fruits, or of potatoes— all groups
of considerable horticultural importance— is still more
disappointing, for I venture to say that the varieties of

No. 603]

BUD-VARIATION

139

these types in cultivation which have originated as bud-
variations can be counted on the fingers of one hand.

At the same time it would be wrong not to attribute
any importance to bud-variation as a plant breeding ad-
junct. Cramer lists several hundred chrysanthemums
and over a hundred roses as of bud-origin, as well as a
smaller number of varieties in species where bud-varia-
tion appears to be less prevalent. Further, Shamel is
said to have found bud-variation in the citrus-fruits to
be sufficiently common to be worthy of an extended inves-
tigation.

These species, however, with perhaps the banana and
the pineapple— the origin of whose varieties is little
known— are the outstanding examples of comparatively
frequent bud-variation, picked from our whole long list
of cultivated plants. The first two examples, moreover,
are species belonging to the domain of floriculture, where
rather superficial characters such as color are valuable.
In very few other species have bud- variations been re-
corded in sufficient numbers to justify us in employing
any other adjective than ' i rare ' ' in describing them. And
of the sum total of these varieties only an extremely small
percentage are of such a nature that agriculture would
suffer a material loss if they were eliminated.

Perhaps these last statements appear to imply a very
limited type of bud-variations. This is not true. Bud-
variations are wholly comparable to seed-variations in
their nature, but they are handicapped because recom-
binations of variant characters are possible only in sexual
reproduction. N bud-variations in a species are simply
N variations, but A7 seed-variations may become 2n seed-
variations provided they are not linked together in hered-
ity. An immense advantage thus accrues in favor of
seminal reproduction because by far the greater number
of commercially valuable characters are complex in their
heredity, i. e., they are represented in the germ-plasm by
several factors independently inherited.

Cramer divides bud-variations into the same classes
that de Vries has used for sexual mutations : progressive.

140

THE AMERICAN NATURALIST [Vol. LI

where new characters arise; retrogressive, where a char-
acter becomes latent or lost ; and degressive, where latent
characters become active. In this important monograph
practically all recorded bud-variations to the date of pub-
lication, 1907, are discussed. Yet not a single case of
progressive variation is listed. They are all catalogued
as retrogressive or degressive. Their classification is
correct, however, only when a progressive variation is
defined as the addition of a character wholly unknown in
the previous history of the species.

As examples of what bud-variation does produce we
may well study Cramer's painstaking work. There are
losses of thorns, hairs and other epidermal characters,
together with an occasional degressive change of the
same kind. There are changes in color in vegetative
parts. Green becomes red or "aurea" yellow, or a loss
of anthocyan occurs. Sometimes the changes are such
that the plants remain striped or otherwise variegated.
Flowers and fruits exhibit the same types of color varia-
tions in considerable numbers. They are mostly losses,
with the appearance of what in Mendelian terminology
is called hypostatic colors, but once in a great while
epistatic colors recur anew.

Monstrosities appear. Other parts of the flower take
on the appearance and form of petals or of sepals. Dou-
bling occurs in several different ways. Fasciations arise.
Changes in the character of the reproductive apparatus
are not uncommon, sometimes giving us seedless fruits.

Plants change their habit of growth. They become
dwarf. They retain juvenile characters. They become
laciniate, or develop the trait known as 1 1 weeping.' '

Thus we see that bud-variation is not limited in its
manifestations; and what is more important, we realize
that bud-variations are very comparable to seminal varia-
tions, there being hardly a type of change known in
sexually reproduced plants that has not been duplicated
asexually. What then is the difference, if any, between
true somatic changes and true germinal changes in con-
stitution? We can get clues which indicate a fairly satis-

No. 603]

BUD-VARIATION

141

factory solution of this problem from three different lines
of research, pedigree cultures, graft-hybrids and cell-
studies.

It is a noteworthy fact that the character of the progeny
produced sexually by bud-variations has been studied in
a comparatively few cases, and in most of these instances
self-pollinations were not made. Nevertheless Cramer
believes the following conclusions are justified:

1. In a vegetative Mendelization, of the progeny of a
branch with the positive character 75 per cent, have the
character and 25 per cent, are without it, while the prog-
eny of a branch without the character all lack it.

2. In a vegetative "Zwischenrasse" by which he gen-
erally means a variegated race, of the progeny of each
type (original and variant), a part retain and a part lack
the character, the percentage being variable.

3. In a vegetative mutation, by which he means any
change not a "Zwischenrasse" and which did not appear
to him to be Mendelian in type, of the progeny of a branch
retaining the positive character, either all possessed it or
a part were with and a part without it, while the progeny
of a branch without the character were all of the same
type.

If we allow for some deviation due to cross-pollination,
I believe that Cramer's records support this view, and
that modern genetic research suggests the interpretation.

In the first place, the "Zwischenrasse" are evidently
of the type studied principally byrCorrens and by Baur
in sexually reproducing races. They are due to chro-
matophore changes, and in many cases at least are not
the result of nuclear activity. This being true, one would
expect in neither asexual nor sexual reproduction the
same type of inheritance for variegated races that obtains
for other types of variation. Inheritance will parallel
cytoplasmic rather than nuclear distribution ; an expecta-
tion apparently realized for both types of reproduction.

t Omitting the<<Zwischenrassen,,, therefore, we have two
phenomena to explain, both of which are similar to cases
of inheritance in sexual reproduction where chromatin

142 THE AMERICAN NATURALIST [Vol. LI

distribution parallels the facts. In each instance the
negative variant— may we call it the recessive— breeds
true. In one case the positive variant breeds true, in the
other case it gives a simple Mendelian ratio.

The mechanism necessary for such phenomena is not
difficult to picture. Bud-variations are many times more
frequent in hybrids, that is, in plants heterozygous for
one or more characters, than they are in pure species.
This is the view of Cramer, this was the view of Masters,
the eminent English student of bud-variations and tera-
tological phenomena, this was the conclusion drawn by
the present writer in several articles published some
years ago. Such results would be obtained either when
the proper germinal change occurs in the chromosome
whose mate lacks a character for which the plant is hetero-
zygous ; or, when there is a dichotomy in which the chro-
mosomes of such a pair are not halved but pass the
material basis necessary for the production of the posi-
tive character to one daughter cell and not to the other,
provided the daughter cell lacking the character gives
rise to a branch.

A bud-variation in a character for which the plant was
homozygous would be obtained only when simultaneous
like changes occur in both chromosomes of a homologous
pair, or when the material basis necessary for the pro
duction of the positive character all passes to one daugh-
ter cell, as described above.

This hypothesis would account for the fact that hetero-
zygotes give rise to bud-variations more frequently than
homozygotes, since a germinal change seldom gives rise
to a new positive character, and a change in one chromo-
some of an identical pair tending toward the production
of a recessive, would not show in the latter case.

I am not certain that this hypothesis may not with
reason be applied to variations that are usually consid-
ered seminal. There is no particular ground for assum-
ing that such variations occur only at the maturation of
the grerm-cells. We know that progressive variations of
whatever origin are extremely rare. Why then may not

No. 603]

BUD-VARIATION

143

most variations be produced in cell divisions previous to
the formation of the germ-cells? When recessive we
should not note them as bud-variations unless the plant
is heterozygous and the mutating cell gives rise to a
branch ; when dominant we should only note them in the
latter eventuality. But if these mutating cells should
later give rise to germ-cells, the change would become
apparent in the progeny.

We have still one other hypothetical case to consider.
It is said that some bud-variations are not transmitted
by seed. I have not been able to trace an authentic case,
but such is the general belief, fathered, I think, by Dar-
win. The usual citation is the nectarine, which sometimes
is said to give nectarines but at other times gives only
peaches. Whether trichome characters only behave thus
I do not know. But if that be true, we can understand
why if we refer to Winkler's work on the so-called graft-
hybrids.

Winkler found that the most interesting of these pecul-
iar phenomena are caused by the tissue of one species
growing around the tissue of the other. He therefore
gave them the euphonious name of periclinal chimeras.
Cytological examination showed that the epidermal tis-
sues only are from one race, the remaining tissues being
from the other. It is really a symbiosis and not a union.
Xow as the germ-cells are formed wholly from subepi-
dermal and never from epidermal tissues, the seeds of
these plants always produced seedlings like the type
forming the inner cell-layers.

It seems probable that the production of the nectarine
may be analogous. If the change producing the nec-
tarine occurs after the epidermal tissue has been segre-
gated from other tissues, the cells which are ancestors
of the germ-cells should not be affected and the nectarine
seedlings would give peaches. If, on the other hand, the
change producing the nectarine, has occurred before any
such segregation, the progeny sexually produced should
in part be nectarines.

DOMINANCE OF LINKED FACTORS AS A
MEANS OF ACCOUNTING FOR
HETEROSIS

DONALD F. JONES

Connecticut Agricultural Experiment Station, Neuu Ha<ven, Connecticut

Reprinted from Genetics 2: 466-479

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle
Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Pennsylvania

Edward M. East

Harvard University

Rollins A. Emerson

Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan

Columbia Unirersity

Raymond Pearl

Maine Agricultural Experiment Station

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum. Canadian subscribers
should add 25 cents for postage. To ail other countries 50 cents should
be added for postage. Single copies will be sent to any address postpaid
for $1.25 each.

All subscriptions, notices of change of address, and business corre-
spondence should, be sent to the Princeton University Press, Princeton,
New Jersey, and all remittances should be made payable to the Princeton
University Press.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J.", under the act of March 3, 1879.

DOMINANCE OF LINKED FACTORS AS A
MEANS OF ACCOUNTING FOR
HETEROSIS

DONALD F. JONES

Connecticut Agricultural Experiment Station, Nov Haven, Connecticut

DOMINANCE OF LINKED FACTORS AS A MEANS OF AC-
COUNTING FOR HETEROSIS1

DONALD F. JONES

Connecticut Agricultural Experiment Station, New Haven, Connecticut
s [Received March 1, 1917]

A stimulation resulting from hybridization in both plants and animals
has long been recognized. The increased growth as the result of crossing
is so common an occurrence that it is probably familiar to everyone who
has made any hybridization experiments.

This stimulation, variously spoken of as "hybrid vigor," stimulus due
to heterozygosis, heterosis, etc., was clearly established as an organic
phenomenon by the abundant cases cited by early investigators such as
Kolreuter (1766), Gartner (1849), Darwin (1877) and Focke
(1881), as well as a large number of other investigators at that time
and an increasingly large number since then. The important investiga-
tions in recent times (East 1908, 1909; Shull 1908, 1909, 1910, 191 1 ;
East and Hayes 191 2) are so familiar that it is not necessary to do
more than mention them.

Concrete explanations as to the cause of these results have not ac-
companied the accumulation of facts. Various hypotheses have at-
tempted to account for the results, but they have been little more than
outlines of the problem.

The valuable contributions of East (1908, 1909) and of Shull
(1908, 1909, 1 910, 191 1 ) established the fact that continued inbreeding
is not a process of continuous degeneration but that the reduction in the
amount of growth is due to the isolation of unlike biotypes differing in
the amount of growth attained at normal maturity. Together with this
isolation of biotypes there was a loss of a stimulation which was assumed
to be derived in some way from crossing. This decrease of vigor be-
comes less after continued inbreeding and to all appearances ceases as
complete homozygosis is approached. This stimulation has been shown
to be correlated more or less closely with the degree of heterozygosity.
The whole subject has been ably presented and discussed by East and

1 Contribution from the Connecticut Agricultural Experiment Station and from
the Bussey Institution of Harvard University.

Genetics 2: 466 S 1917

DOMINANCE OF LINKED FACTORS AND HETEROSIS 467

Hayes (1912). A quotation from this paper (pp. 36 and 37) presents
the matter as it stands at present :

"The hypotheses in regard to the way by which the act of fertilization
initiates development are numerous, but since they are entirely speculative
it is not necessary to discuss them here. The only conclusion that seems
justified is that they are not immediately psychological or vitalistic in na-
ture. Loeb's remarkable researches prove this. But whatever may be the
explanation of the means by which the process is carried out, the statement
can be made unreservedly that the heterozygous condition carries with it
the function of increasing this stimulus to development. It may be me-
chanical, chemical, or electrical. One can say that greater developmental
energy is evolved when the mate to an allelomorphic pair is lacking than
when both are present in the zygote. In other words, developmental stimu-
lus is less when like genes are received from both parents. But it is clearly
recognized that this is a statement and not an explanation. The explana-
tion is awaited.''

Keeble and Pellew (19 10) first suggested a concrete explanation to
account for the results of this nature which they obtained with peas.
Two varieties of garden peas, as grown by them, each averaged from 5
to 6 feet in height. The Fx grown from this cross averaged from 7 to 8
feet in height, 2 feet taller than either parent. A result of this kind is
comparable to heterosis. The F2 was put into four classes : one class con-
taining plants as tall as the Fi, two classes of semi-tall plants similar in
height to the two parents, and one class of dwarfs shorter than either
parent. The two classes of semi-tall plants, similar in height, were dif-
ferentiated in the same manner as the two parents ; one had thick stems
and short internodes, the other had thin stems and long internodes.
Other differences helped to distinguish the two classes of equal height.
The number of plants falling into these four classes agreed closely with
the expectation from a di-hybrid ratio where two factors showing domi-
nance were concerned, giving a 9 : 3 : 3 : 1 ratio.

The writers assumed two factors to be concerned : one producing thick
stems, the other long internodes. These factors they designated T and
L. One of the parental varieties was medium in height because it pos-
sessed one of these factors, e.g., that for thick stems, but lacked the
other. Such a plant had the formula TTll. The other variety was of
medium height because it lacked this T factor but possessed the factor
for long internodes, and was given the formula ttLL. Both of these
factors showed dominance over the allelomorphic condition ; hence the Ft
was taller than either parent because both factors were present together.
Whether or not later investigations have justified the interpretation that
Keeble and Pellew have placed on the data as explaining height of

Genetics 2: S 1917

468 DONALD F. JONES

their peas makes no material difference to the discussion here. Taken as
it stands, it is a beautiful illustration of the way in which dominance
may increase a character in Fx over the condition of either parent.

Curiously enough, this explanation has never been considered an ade-
quate one or in any way essentially related to the universal phenomenon
of heterosis. This hypothesis of dominance accounting for heterosis, as
outlined by Keeble and Pellew, has two objections which have up to
the present been considered insurmountable.

The chief objection has been that, if heterosis were due to the domi-
nance of a greater or less number of factors governing the amount of
development, it would be possible in generations subsequent to the F2
to recombine in one homozygous race all of the factors resulting in large
growth and, conversely, the negative condition in another homozygous
race. In other words, it would be possible to obtain one strain having all
of the dominant factors, and another with all of these dominant factors
lacking. Both of these races should be homozygous, hence self-fertiliza-
tion should not result in less vigorous progeny. The completely recessive
race should be below the parents in its power for development, as the Fi
and the complete dominant were above the parents. That all of these
supposedly necessary corollaries are not supported by the facts is well
known.

Both Shull (1911) and East and Hayes (1912) have considered
this objection to be valid. A quotation (p. 39) from the latter makes
their position on this point clear.

"Keeble and Pellew (1910) have recently suggested that 'the greater
height and vigor which the F1 generation of hybrids commonly exhibit may
be due to the meeting in the zygote of dominant growth factors of more
than one allelomorphic pair, one (or more) provided by the gametes of one
parent, the other (or others) by the gametes of the other parent.' We do
not believe this theory is correct. The 'tallness' and 'dwarfness' in peas
which Keeble was investigating is a phenomenon apparently quite differ-
ent from the ordinary transmissible size differences among plant varieties.
Dwarf varieties exist among many cultivated plants, and in many known
cases dwarfness is recessive to tallness. It acts as a monohybrid or pos-
sibly a dihybrid in inheritance, and tallness is fully dominant. Varietal
size differences generally show no dominance, however, and are caused by
several factors. Transmissible size differences are undoubtedly caused by
certain genetic combinations (East 1911), but this has. nothing to do with
the increase of vigor which we are discussing. The latter is too universal
a phenomenon among crosses to have any such explanation. Furthermore,
such interpretation would not fitly explain the fact that all maize varieties
lose vigor when inbred. "

Another objection to the hypothesis of dominance has been raised by

DOMINANCE OF LINKED FACTORS AND HETEROSIS 469

Emerson and East (1913). In this publication it is said that, if the
effect of heterosis were due to dominance, the distribution of the F2 in-
dividuals would be unsymmetrical in respect to characters in which hete-
rosis was shown in Fi. This follows from the familiar Mendelian ex-
pectations where there is dominance and any number of factors is con-
cerned. For the purpose of illustrating this point let us take the case of
height of peas already cited. In the F2 population a distribution of the
individuals in respect to height is, theoretically, 9 tall plants (with both
factors present), 6 medium-tall plants (3 with one factor -\- 3 with the
other), and one short plant (with both factors lacking).

Similar asymmetrical distributions in F2 would occur with any number
of factors (if there were no other facts to be taken into consideration),
as seen from the figures given in table 1 modified somewhat from those
given by Baur (1911, p. 63).

In any case of a size character similar to height of peas with any num-
ber of factors, the plotting of the number of individuals in F2 occurring
in the classes given in row B in table 1 would give an asymmetrical dis-
tribution. This is on the assumption that the individual having the
greatest number of dominant factors present (whether in the simplex or
duplex state) would attain the greatest development of the size character.

In the vast amount of data accumulated upon the inheritance of quan-
titative characters no such tendencies toward an, asymmetrical distribu-
tion is evident in the majority of cases recorded. In Emerson and
East's paper, referred to, dealing with quantitative characters in maize,
and in Hayes's publication (191 2) dealing with the same type of char-
acters in tobacco, the distributions in F2, where heterosis is shown in Fi,
are all considered to be of the type of normal frequency distributions.
If any skewness is shown by any of these it is too slight to suggest the
types of curves obtained by plotting the figures in table i, B.

It is perfectly evident that the two objections raised against the hy-
pothesis of dominance as a means of accounting for heterosis, as out-
lined by Keeble and Pellew, and as it has been considered up to the
present, are valid. But both these objections to dominance as an inter-
pretation of heterosis have failed to take into consideration the fact of
linkage.

Abundant evidence is fast being accumulated2 to show that characters
are inherited in groups. The different theories accounting for this link-

2 It is unnecessary to give references to the convincing results obtained by Morgan,
Bateson, and their collaborators, as well as to those obtained by many others whose
work is of great importance if not so extensive.

Genetics 2: S 1917

4/0

DONALD F. JOXES

Table i

Distribution of Fn individuals when each character shows complete dominance and

each has a visible- effect.

»\ U II 1 Uc I Ul

1 ciL iUl S 111
\\ 1 1 IV. 11 L11C X ^
13 I1CLC1 UiJ

gous

Distribution of the individuals

X Uldl 1 1 Ll 1 1 1

npr 1 tt flip
L/^l 111 III C

t~\ i~\ 1 1 1 ifi r\Y\
pUjJLllclLlUIl

3: 1

3:1

1 : 0

1 : 1

9:3:3:1

9:6 : 1

2:1 : 0

1:2 : 1

27 : 27 : 9 : 1

3: 2 : 1 :o

1 : 3 : 3 : 1

A £

17:27: 27

: 27 19:9:9:9:9:9:3:

3:3:

B I

\i:

108

: 54 :

: 2 :

256

i :

: 6 :

A 3»

:Sn

-1 •

. 3„_i . 3„_, . 3„_2 :3„-

2 •

etc.

.... : 1

B i(3») :

D( 3""1

) : D(3»-2) : ....

etc.

. . : 1

C n

: 7/ — 1

: 11 — 2 :

etc.

. . . ?/ — n

(2»)2

D i

... etc

. = coefficients of the

expanded

binomial (a-\-a) >l .

: i

A, The distribution into the visibly different categories. B, The distribution into
categories with different numbers of dominant factors present (either in a homozy-
gous or heterozygous condition). C, The number of dominant factors in which the
categories differ. D, The number of visibly different categories with the same num-
ber of dominant factors present.

age of characters make no essential difference in the use to which these
facts will be put here. It is only necessary to accept as an established
fact that characters are inherited in groups and that it is these groups of
factors which Mendelize. The chromosome view of heredity, as de-
veloped by Morgan and others (1915), will be used because it gives a
means of representation in a simple, graphical manner.

The increasing complexity of Mendelism points very strongly to the
probability that the important characters of an organism are determined
by factors represented in all or most of the chromosomes or linkage
groups. This idea has been proposed by East (191 5) and seems to be
in accord with the facts. If this view is approximately correct, and if it

DOMINANCE OF LINKED FACTORS AND HETEROSIS

may also be assumed that, in addition to the factors which differentiate
varieties, many different factors may bring about the same visible effect,
then it is possible to meet the two objections raised against dominance as
a means of accounting for heterosis.

As an illustration of what is meant by different factors bringing about
the same visible effect, an example may be taken in which one variety of
plants grows to an average height of six feet because of one set of fac-
tors, and another variety grows to approximately the same average height
but attains this height through the operation of a different set of factors.
This is comprehensible when it is remembered that height is only an ex-
pression of a plant's power to develop. Hereditary factors which affect
any part of the plant may indirectly determine height. Direct proof as
to the essential correctness of this assumption, i.e., of different factors
producing the same somatic effect, is at hand in the cases of duplicate
genes producing the same morphological result in Avena saliva (Nils-
son-Ehle 1909) and Bursa bursa-pastoris (Shull 1914), as well as the
other cases of duplicate genes reported by Nilsson-Ehle (1908) and
East (1910).

The widespread occurrence of abnormalities and other characters
detrimental to the organism's best development is well known in both the
plant and animal kingdoms. This is especially true in naturally cross-
pollinated species of plants. It may be taken for granted that no one
variety has all of these unfavorable characters nor, on the other hand, has
it all the favorable characters. For the most part each variety possesses
a random sample of the favorable and unfavorable characters. There
are differences between varieties in their power for development, how-
ever, just as there are differences in superficial characters. Some varie-
ties of plants grow taller than others; some grow faster; some produce
more seed. But, on the average, most of the varieties of a species tend
to grow to about the same extent, however much they may differ in
superficial characters.

If, for the most part, these favorable characters are dominant over the
unfavorable (if normalities are dominant over abnormalities) it is not
necessary to assume complete dominance in order to have a reasonable
explanation of the increased development in Fx over the average of the
parents or any subsequent generation. It is in Fi, and in ¥x only, that
the maximum number of different factors can be accumulated in any one
individual.

Because of linkage it is impossible to recombine in any one individual
in later generations any greater number of characters in the homozygous

Genetics 2: S 1917

472

DOXALD F. JONES

condition than were present in the parents if the factors were distributed
uniformly in all of the chromosome pairs. Possible exceptions to this
statement will be discussed later. This view of the situation explains
why the effects of heterozygosis result in a greater development in Fi
than in the parents, and not less. Why should crossing not have re-
sulted in a depressing or indifferent effect instead of a stimulating one,
according to previous views?3 It also makes it seem probable that the
effects of heterozygosis remain throughout the life of the sporophyte,
even through innumerable asexual generations. Furthermore, it will be
shown that no skewness in the distribution of F2 is expected.

Let me submit in the form of a concrete illustration the abstract view
that I have tried to present in the preceding paragraphs. A purely hy-
pothetical case will be assumed, in which two homozygous varieties of
plants, having three pairs of chromosomes, both attain approximately the
same development as represented by any measurable character. This de-
velopment will be considered to amount to 6 units, 2 of which are con-
tributed by each chromosome pair. One of these varieties, which will be
called "X," attains this development because of factors distributed in the
three pairs of chromosomes. Any number of factors may be chosen,
but, for the sake of simplicity, only three in each chromosome will be
employed. These are numbered 1, 3. 5; 7, 9, 11 ; and 13, 15, 17; in the
following diagram, each different in its contribution to the plant's de-
velopment. The other variety, "Y", develops to an equal extent in the
character measured, and this development will also be considered to
amount to 6 units. It attains this same development, however, by a dif-
ferent set of factors distributed in the three chromosomes, numbered
2, 4, 6; 8, 10. 12 ; and 14, 16, 18. It is also assumed that these 6 factors
are fully as effective in the in condition as in the 2n condition, i.e., show
perfect dominance. It will be seen from the diagram that the Fx develops
to twice the extent of either parent, because there are present here 18
different factors (in the in condition), whereas the parents have only 9
(in the 2n condition). In the diagram, any other factorial complex
common to both varieties is ignored. The development of the parents of
6 units and of the Fi of 12 units is additional to that afforded by this
common factorial complex.

Following this hypothetical case into the F2 generation by selfing or

3 Crosses between plants not closely related do result in no greater development than
the parents and in many cases much less than the parents. This is because characters
which are widely dissimilar are unfavorable to the organism's best development when
acting together.

DOMINANCE OF LINKED FACTORS AND HETEROSIS

473

breeding together these Fi plants, the theoretical results given in table 2
are obtained.

Summing up the results of this tabulation, it will be found that eight
plants are homozygous and have the same development as either parent,
i.e., of six units. Eight plants are heterozygous in all three chromosome

P. X : 6 Y : 6

X x Y : 12

Diagram i. — To show how factors contributed by each parent may enable the first
generation of a cross to obtain a greater development than either parent.

pairs and have the same amount of growth as Fi, i.e., of twelve units.
The remaining 48 plants fall into two equal-sized groups developing to
eight and ten units respectively. In other words, the distribution is sym-
metrical, and this symmetry remains, however many chromosomes are
concerned.

Furthermore, it should be noted that the mean development of F2 is
nine units, which is an excess above the parents of just half of the ex-
cess of the Fi over the parents. In other words, the extra growth de-
rived by crossing the two varieties has diminished 50 percent. In F3 from
a random sample ©f F2, it can be shown that this excess again diminishes
50 percent, so that the effect is only 25 percent as great in F3 as in Fly

Genetics 2: S 1917

474

DONALD F. JOXES

Table 2

Composition of a tri-hybrid in F2 according to Mendelism, and the development which
each individual attains depending upon the number of heterozygous chromo-
somes contained and thereby the total number of different factors present.

Number of indi-

Contribution o£

viduals in each

category

pair

A A B B C C

2 + 2 + 2

A A' B B C C

4 + 2 + 2

A A BB'CC

2+4 + 2

A A B B C O

2 + 2 + 4

AA'BB'CC

4 + 4+2

A A BB'CC

2 + 4 + 4

A A' B B C C

4 + 2 + 4

AA'BB'CC

4 + 4 + 4

A A BB CC

2 + 2 + 2

A A B B' CC

2+4 + 2

A A' B B C'C

4 + 2 + 2

A A' B B' C'C

4 + 4+2

A A B'B' C C

2 + 2 + 2

A A B'B' C C

2 + 2 + 4

A A' B'B' C C

4 + 2 + 2

A A' B'B' C C

4 + 2 + 4

A' A' B B C C

2 + 2 + 2

A' A' BB'CC

2 + 4 + 2

A'A' B B C C

2 + 2 + 4

A' A' BB'CC

2 + 4 + 4

A'A' B'B' C C

2 + 2 + 2

A'A' B'B' C C

2 + 2 + 4

A'A' B B C'C

2 + 2 + 2

A'A' B B' C'C

2 + 4 + 2

A A B'B' C'C

2 + 2 + 2

A A' B'B' C'C

4 + 2+2

A'A' B'B' C'C

2 + 2 + 2

64 Total

Distribution of the F9 individuals according to the development attained.

Classes

= 4

Number of classes

Frequency

=64 ■

Total population

and so on in subsequent generations. This is in accord with the mathe-
matical prediction made by East and H«\yes (1912), to which actual
data obtained from maize roughly approximate, as shown by Jones

(1916).

The development attained by any individual in table 2 is correlated
with the number of heterozygous factors present. This has been main-

DOMINANCE OF LINKED FACTORS AND HETEROSIS

tained by all recent writers on the subject as a rough description of the
facts as obtained in actual experiments.

When different numbers of chromosomes are concerned, according to
this scheme, the number of individuals in the different classes making
up the whole F2 population is given in table 3.

In any F2 distribution there are as many individuals heterozygous for
all factors (duplicating Fx individuals) as there are individuals homozy-
gous for all factors concerned in the original cross (two duplicating the
parents; the remaining forming new homozygous combinations). The
remaining individuals fall into a symmetrical distribution between these
two end classes. The theoretical figures for any F2 distribution in which
n Mendelizing units are concerned can be obtained by taking the coeffi-

Table 3

Distribution of the individuals in F„ according to the number of heterozygous
chromsomes pairs they contain.

Number of
chromosome
pairs in
which the
F is het-
erozygous

I Total num-

Classes with different number of heterozygous chromosome ber of indi-
pairs and the number and ratio of individuals j viduals in

in these classes 1 the popula-

tion

5 i

256

160

320

160

1024

etc

2«

(2n)2

etc

coefficients of the

expanded binomial (a + a

)•

cients of the expanded binomial (a -f- a)n and multiplying these by 2n, as
shown in table 3. Since the expanded binomial is used to illustrate a
normal frequency distribution, there can be no question as to the sym-
metry of the F2 distributions if the diagrammatic scheme outlined is, in
this respect, a description of the actual facts.

In the preceding purely diagrammatic representation of the way in

Genetics 2: S 1917

476

DONALD F. JOXES

which dominance may account for the effects of heterozygosis, perfect
dominance was assumed. Such an assumption is neither justified nor
desirable. Many theoretical explanations of the inheritance of quantita-
tive characters are based on exactly the converse assumption, i.e., that
factors in the in condition have just half the effect that they have in the
2n condition.

In the development of an organism, however, all types of factors are
concerned, both qualitative and quantitative. Partial dominance in quali-
tative characters is a normal occurrence. The concensus of opinion at
the present time is that there may be, in reality, no cases of perfect domi-
nance. In those cases in which the heterozygote cannot be distinguished
from the pure dominant, it is assumed that the similarity is only ap-
parent and not real. The heterozygote merely approaches the condition
of the dominant type more or less closely. However much it may be
true that perfect dominance rarely or never occurs, the fact and univer-
sality of partial dominance can hardly be denied.

In this connection it should be realized that the difference between the
heterozygote and the recessive type in many cases is one of kind, while
the difference between the heterozygote and the dominant type is one of
degree. A good illustration of this point is found in the case of albinism
in maize. Plants heterozygous for the factor (or factors) determining
the production of chlorophyll cannot be distinguished from normal green
plants — a case of apparently complete dominance. If there is in reality
a difference between these heterozygous and homozygous normal green
plants, although not apparent, that difference is very slight as compared
with the difference between the heterozygote and the abnormal recessive.
In the former case the difference, if there is any, is quantitative. The
heterozygote may not have as much chlorophyll as the normal homozy-
gote. In the second case the difference is qualitative. The heterozygote
has chlorophyll; the recessive has none. This is a difference which de-
termines the life or death of the organism.

All the evidence at hand leads to a seemingly logical conclusion, one
necessary to the conception of dominance as an explanation of heterosis,
which is, that many factors in the in condition have more than one-half
the effect that they have in the 211 condition. Whether or not this is a
logical conclusion and one that is justified by the facts remains to be
seen. It certainly has the advantage of being more definite and compre-
hensible than the assumptions previously made (Shull 191 i ; East and
Hayes 1912), that factors in the heterozygous condition stimulate de-
velopment by virtue of their being in that condition, without showing in
any way why this should be so.

DOMINANCE OF LINKED FACTORS AND HETEROSIS

There is abundant evidence to show that many abnormal characters
exist in a naturally cross-pollinated species and that they are recessive to
the normal condition. In maize innumerable examples can be cited. In
addition to the complete lack of chlorophyll already mentioned, there are
also other chlorophyll factors which distinguish yellowish-green plants
from normal green plants, just as there are cases of both conditions in
other plants, e.g., Pelargonium (Baur 191 1). By inbreeding, strains of
maize are isolated which are dwarf ; some are sterile ; some have con-
torted stems; some fasciated ears. Some are more susceptible to the
bacterial wilt disease, and still others have brace roots so poorly de-
veloped that they cannot stand upright when the plants become heavy.
It is unnecessary to mention more examples, because their occurrence in
many kinds of material is familiar to everyone. All the characters cited
are recessive, either completely or to a large degree, to the normal con-
dition. More than one of these unfavorable characters may be present
together in one inbred strain. Xo one strain so far known has them all.

Crossing many of these strains of maize together produces perfectly
normal Fx plants. They are normal because the factors which one strain
lacks are supplied by the other, and conversely. Because more of the
favorable characters are present when the strains are united in Fi than
in either parent, the Ft is naturally able to attain a greater development.
This effect is heterosis.

In the preceding diagrammatic illustration of the way in which hetero-
sis may be brought about it was assumed that all factors had equal ef-
fects, that they were evenly distributed in the chromosomes, and that there
were no crossovers. This is probably far from describing all the actual
conditions. All deviations from this uniformity add to the complexity
of the problem. It remains to be seen whether or not the assumption of
dominance as an explanation of heterosis will not meet all or most of the
requirements raised by all these complicating factors. It is only neces-
sary to consider that a large number of factors is concerned, and that
those factors are in most cases fairly evenly distributed among all the
chromosomes, and that, in the main, crossovers in some places are bal-
anced by crossovers in others.

Crossing over also provides a means of understanding why certain
homozygous individuals (and varieties) may possess a greater number
of desirable characters than others. Exceptionally good individuals
might be formed by crossing over in heterozygotes occurring in such a
manner that all, or a large number of, desirable characters would be
combined together eventually in one individual. Such a condition, ac-

Genetics 2: S 1917

478

DONALD F. JOXES

cording to the laws of chance, would be exceedingly rare, which is well in
accord with the facts.

Without going into all the possibilities which this viewpoint opens up,
it is onlv necessary to say that a way is offered to meet the objections
which have been raised against the conception of dominance as a means
of accounting for the facts of heterosis as so far known.

There is still the possibility that there may be a stimulus derived from
crossing quite apart from hereditary factors. The view presented here
simply coordinates the existing knowledge of heredity so as to give a
comprehensible view of the way in which heterosis may be brought about.

SUMMARY

1. The phenomenon of increased growth derived from crossing both
plants and animals has long been known but never accounted for in a
comprehensible manner by any hypothesis free from serious objections.

2. The conception of dominance, as outlined by Keeble and Pellew
in 1 910 and illustrated by them in height of peas, has had two objections
which were: a. If heterosis were due to dominance of factors it was
thought possible to recombine in generations subsequent to the F2 all of
the dominant characters in some individuals and all of the recessive char-
acters in others in a homozygous condition. These individuals could not
be changed by inbreeding, b. If dominance were concerned it was con-
sidered that the F2 population would show an asymmetrical distribution.

3. All hypotheses attempting to account for heterosis have failed to
take into consideration the fact of linkage.

4. It is shown that, on account of linked factors, the complete domi-
nant or complete recessive can never or rarely be obtained, and why
the distributions in F2 are symmetrical.

5. From the fact that partial dominance of qualitative characters is a
universal phenomenon and that abnormalities are nearly always recessive
to the normal conditions, it is possible to account for the increased growth
in Fi because the greatest number of different factors are combined at
that time.

6. It is not necessary to assume perfect dominance. It is only neces-
sary to accept the conclusion that many factors in the in condition have
more than one-half the effect that they have in the 211 condition.

7. This view of dominance of linked factors as a means of accounting
for heterosis makes it easier to understand: a, why heterozygosis should
have a stimulating rather than a depressing or neutral effect ; and b, why

DOMINANCE OF LINKED FACTORS AND HETEROSIS

479

the effects of heterozygosis should operate throughout the lifetime of the
individual, even through many generations of asexual propagation.

LITERATURE CITED

Baur, E., 191 1 Einfiihrung in die experimentelle Vererbungslehre. pp. vi -f- 293.
Berlin : Borntraeger.

Darwix, C, 1877 The effects of cross and self-fertilisation in the vegetable kingdom.

pp. viii + 482. London : D. Appleton & Co.
East, E. M., 1908 Inbreeding in corn. Connecticut Agr. Exp. Sta. Report for 1907,

pp. 419-428.

1909 The distinction between development and heredity in inbreeding. Amer.
Nat. 43: 173-181.

1910 A Mendelian interpretation of variation that is apparently continuous.
Amer. Nat. 44 : 65-82.

1915 The chromosome view of heredity and its meaning to plant breeders.
Amer. Nat. 49 : 457-494.

East, E. M., and Hayes, H. K., 1912 Heterozygosis in evolution and in plant breed-
ing. U. S. Dept. of Agric, Bureau of Plant Industry Bull. 243. pp. 58.

Emersox, R. A., and East, E. M.f 1913 The inheritance of quantitative characters in
maize. Nebraska Agr. Exp. Sta. Research Bull. 2. pp. 120.

Focke, W. O., 1881 Die Pflanzen-Mischlinge. pp. 569. Berlin : Borntraeger.

Gartxer, C. F., 1849 Versuche und Beobachtungen iiber die Bastarderzeugung im
Pflanzenreich. pp. xvi + 791. Stuttgart: C. F. Gartner.

Hayes, H. K., 1912 Correlation and inheritance in Xicotiana tabacum. Connecticut
Agr. Exp. Sta. Bull. 171. pp. 45.

Joxes, D. F., 1916 Inbreeding in maize. Paper read before the Annual Meeting of the
Botanical Society of America, at Columbus, Ohio, Dec. 1915. Abstract in
Science N. S. 63: 290, 25 F 1916.

Keeble, F., and Pellew, C, 1910 The mode of inheritance of stature and of time of
flowering in peas (Pisum sativum). Jour. Genetics 1:47-56.

Kolreuter, T. G., 1 766 Dritte Fortsetzung der Vorlaufigen Nachricht von einigen das
Geschlecht der Pflanzen betreffenden Versuchen und Beobachtungen. Leipzig:
Gleditschen Handlung. Reprinted 1893 m Ostwald's Klassiker der exakten
Wissenschaften, No. 41. Leipzig.

M org ax, T. H., Sturtevaxt, A. H., Muller, H. J., and Bridges, C. B., 1915 Mechan-
ism of Mendelian heredity, pp. xiii -}- 262. New York : Henry Holt & Co.

Nilssox-Ehle, H., 1908 Einige Ergebnisse von Kreuzungen bei Hafer und Weizen.
Botaniska Notiser pp. 257-294.
1909 Kreuzungsuntersuchungen an Hafer und Weizen. Lunds Lniversitets
Arsskrift, N. F., Afd. 2, Bd. 5, Nr. 2, 122 pp.

Shull, G. H., 1908 The composition of a field of maize. Rep. Amer. Breeders' Ass.
4 : 296-301.

1909 A pure line method of corn breeding. Rep. Amer. Breeders' Ass. 5 : 51-59.

1910 Hybridization methods in corn breeding. Amer. Breeders' Mag. 1 : 98-107.

191 1 The genotypes of maize. Amer. Nat. 45: 234-252.

1914 Duplicate genes for capsule form in Bursa bursa-pastoris. Zeitschr. f. ind
Abst. u. Vererb. 12:97-149.

Information for Contributors

The length of contributions to Genetics will not be arbitrarily limited, but article
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitations, as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication.

Contributors are requested to observe the usual care in the preparation of manu*
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published in a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer t*
them, and should be separated from the text by lines. All footnotes should be num
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fcf
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4$i X 7JA inches' and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be 12^2 X lV* inches. Legends for ai! figures should be typewritten on a
separate sheet.

Contributor will receive 100 reprints of their papers, with covers, free of charge,
and may secure additional copies at actual cost of manufacture, provided the*e are
ordered when the corrected galley-proofs are returned.

Manuscripts and ail editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull. 60 Jefferson Road, Princeton, N. J.

STATEMENT OF THE OWNERSHIP, MANAGEMENT, CIRCULATION, ETC., REQUIRED
BY THE ACT OF CONGRESS OF AUGUST 24, 1912,
Of Genetics, published bi-monthly at Princeton, N. J., for April 1, 1917.
State of New Jersey, County of Mercer, ss.

Before me, a notary in and for the Srate and county aforesaid, personally appeared
C. Whitney Darrow, who, having been duly sworn according to law, deposes and says that he is the
Business Manager of Genetics, and the following is, to the best of his knowledge and belief, a true
statement of the ownership, management, etc., of the aforesaid publication for the date shown
in the above caption, required by the Act of August 24, 1912, embodied in section 443, Postal
Laws and Regulations, to wit:

L That the names and addresses of the publisher, editor, managing editor, and business
managers are: Publisher, Princeton University Press, Princeton, N. J* Managing Editor,
George H. Shull, Princeton, N. J.; Business Managers, Princeton University Press, Princeton. N. J.

2. That the owners are: Princeton University Press, Princeton, New Jersey.

3. That the known bondholders, mortgagees, and other security holders owning or holding
1 percent or more of total amount of bonds, mortgages, or other securities are: None,
C. VV. Darrow, Mgr. Princeton University Press. Sworn to and subscribed before me this
4th day of Jane, 1917.

[seal] Elmep. C. Nelson.

(My commission expires Feb. 8th, 1922.)

GENETICS, SEPTEMBER 1917

TABLE OF CONTENTS

Pearl, Raymond, Studies on the physiology of reproduction in the
domestic fowl XVII. The influence of age upon re-
productive ability, with a description of a new reproduc-
tive index 417

Little, C. C, The relation of yellow coat color and black-eyed

white spotting of mice in inheritance 433

Bridges, Calvin B., Deficiency 445

Jones, Donald F., Dominance of linked factors as a means of ac-
counting for heterosis 466

Shull, A. Franklin, Sex determination in Anthothrips vcrbasci. . 480

Robbixs, Rainard B., Some applications of mathematics to breed-
ing problems 489

[Reprinted from The American Naturalist, Vol. LI., November, 1917. J

ON REVERSIBLE TRANSFORMABILITY OF
ALLELOMORPHS

H. TERAO

The Imperial Agricultural Experiment Station, Tokyo, Japan

In genetical studies of variegation in plants, the fact
has been observed occasionally that with a certain fre-
quency a dominant allelomorph occurs in the correspond-
ing recessive homozygote (De Vries,1 Correns,2 and
Emerson3). In this paper the author presents a new in-
stance of a similar phenomenon, which it is hoped may
throw additional light on the subject.

In certain pedigree cultures of the rice plant, Oryza
sativa L., there happened to occur in 1912 families con-
taining besides ordinary fertile plants a number of sterile
plants. These sterile plants were normal in their growth,
but showed a considerable barrenness at the ripening
season. Some of them yielded no seed whatsoever, others
bore a small number of normal seeds, and very few were
mosaic forms with higher fertility. These families, two
in number, each belonging to a different variety, were
derived from single plants of the former generation, and
were very uniform in other characters. From them the
experiment was started.

The rice plant, being a self-pollinated species, is con-
venient material for breeding experiments. Although
the experiments in this investigation were made largely
from open-pollinations, the results obtained were always
similar to those from experiments in which plants were
artificially protected against accidental natural crossing.

The observations of 1912 and 1913 are shown in sum-
marized form in Tables I and II, a and ~b, and point to the
following conclusions. Sterility behaves as a simple re-

iDe Vries, H., "Die Mutationstheorie, ' ' Bd. I, 1901, pp. 489-511;
" Species and Varieties, their Origin by Mutation/ ' 1905, pp. 309-339.

2 Correns, C, Berichte der Veutschen Botanischen Gesellschaft, Bd. 28,
1910, pp. 418-434.

3 Emerson, R. A., American Naturalist, Vol. 48, 1914, pp. 87-115; ■
Genetics, Vol. 2, 1917, pp. 1-35.

690

601

THE AMERICAN NATURALIST [Vol. LI

cessive to fertility, and the seeds resulting from partial
fertility of sterile plants again give segregating families.
In Family A, which shows an exceedingly slight fertility
of sterile plants, the segregation ratio in the offspring
derived from fertile individuals is quite close to expecta-
tion, but in Family B which shows a considerably higher
grade of partial fertility of sterile plants, the progeny of
fertile individuals exhibit considerable deviations from
the expected segregation ratio.

TABLE I

The Segregating Families, A and B, in 1912

Fam.

Segregation

Partial Fertility of
Sterile Plants

Fertile
Plants

Sterile
Plants

Total
No. of
Ind.

Steriles Je

Ratio per 4

Total
No. of
Spike-
lets

Fertile Spikelets

No.

ca.

. 13

26.53

2.94

1.06

9,000

0.02

B ....

105

130

19.23

3.23

0.77

14,941

434

2.90

TABLE II

The Families Derived from Families A and B
(a) The Progeny of the Fertile Plants

No. of Families

Ratio per 3

Segregating Families

Family
In 1912

Uniformly
Fertile

Segregating

Total

Uni-
formly
Fertile
Families

Segre-
gating
Families

Fertile
Plants

Sterile
Plants

Total
Number

Indi-
viduals

Steriles

a:...

0.94

2.06

1,068

346

1,414

24.46

B....

105

1.17

1.83

4,874

1,301

6,175

21.06

(b) The Progeny of the Seeds on the Sterile Plants

Family in

Number of

Total Number

1912

Families

Fertile Plants

Sterile Plants

of Individuals

Steriles %

0.00

401

115

516

22.29

These facts may be interpreted by the following hy-
pothesis. The dominant and the recessive types con-
cerned are assumed to be transformed by certain un-
known causes into the other allelomorph. The recessive
allelomorph which has made its appearance in Families
A and B is assumed to have originated in the preceding

No. 611]

ALLELOMORPHS

692

generation by the transformation of the dominant allelo-
morph. This recessive state of the hereditary substance,
however, has a tendency to revert into the original domi-
nant state. Such reversion is especially likely to occur in
vegetative cells, where each recessive allelomorph seems
to be able to revert independently. Consequently, in reces-
sive homozygotes the reversion generally will produce
heterozygotic cells, either one of the two recessive alle-
lomorphs being changed into the dominant. The hetero-
zygotic cells thus formed will give rise to partial fertility
in otherwise sterile plants. Again, the recessive allelo-
morph in heterozygotic cells may be subject to similar
reversion, and such reversion may occur both in the
heterozygotic cells of sterile plants and in normal hetero-
zygotes. Here, however, heterozygotic cells will be
transformed into dominant homozygotic cells without
visible effect on the plant concerned. The consequence
of this reversion in the next generation will be that the
proportion of the dominant segregates may exceed the
theoretically expected figure. Finally, it may be assumed
that between Families A and B there exists a difference
in the reverting tendency of the recessive allelomorph,
which necessarily will effect the differences in both the
intensity of partial fertility of sterile plants and the devi-
ations in the segregation ratio.

In Table III the segregating families derived from the
fertile plants of Family B are classified according to the
magnitudes of the deviations in terms of probable errors.
The true percentage for the recessive is assumed, in the
one case as 25 per cent, (the Mendelian ratio), and in the
other case as 21 per cent, (an arbitrary number), In
comparing the two different frequency distributions made
in this manner with the theoretical frequency distribu-
tion, it is observed that while the frequency distribution
of the deviations from 25 per cent, shows a considerable
discrepancy from the theoretical, the latter fits the fre-
quency distribution of the deviations from 21 per cent,
rather closely, the goodness of fit being P = 0.915. Con-
sequently, the ca, 4 per cent, deficiency of recessive segre-

693

THE AMERICAN NATURALIST [Vol. LI

gates is a normal expectation and not an experimental
error.

TABLE III

The Frequency Distribution of the Deviations in the Segregation
Ratios in the Group of 64 Segregating Families Descended
from Family B of the Year 1912

Dev. /P.E.

-5 —4 — ;

I -2

L 0-1

2 +3 +

4 +

5 Total

Experimental frequency
Experimental froquency
Expectation

(I) -...

(II) ...

11 7
1

0.2: 1.2

2
4.3

17
11

10.3

15 6|
19 14
16.0 16.0

4
10
10.3

5 2
4.3 1.2

0.2

64
64
64.0

Note: In the experimental frequency (I) the true percentage for re-
cessives is taken as 25 per cent., and in (II) as 21 per cent.

Such an aberrant segregation ratio seems to be a con-
stant tendency all through the generations descended from
Family B. This is shown in Table IV in which the ex-
periments in the years from 1912 to 1915 are summarized.

TABLE IV

The Aberrant Segregation-Ratios Obtained in the Years 1912-1915

No. of

Parent-

No. of

Ster. <i

D./P.E.

Years

Fams.

plants

Inds.

Fertiles

Steriles

Dev. i

p.e. i

1912

Fertile

130

105

19.23

5.77

2.55

2.3

1913

6,175

4,874

1,301

21.06

3.94

0.37

10.6

1914

1,560

1,207

353

22.63

2.43

0.74

3.3

1915

4,696

3,732

964

20.52

4.48

0.47

9.5

Total. . . .

128

12,561

9,918

2,643

21.04

3.96

0.26

15.2

1913

Sterile

516

401

115

22.29

2.71

1.21

2.2

1914

994

779

215

21.63

3.37

0.93

3.6

1915

684

522

162

23.68

1.32

1.12

1.2

Total

2,194| 1,702

492

22.43

2.57

0.62

4.1

Again, in regard to the intensity of partial fertility of
sterile plants, the descendants of Families A and B ex-
hibited respectively relations similar to those seen in 1912.
(Family A was not traced after 1913.) A count of fertile
spikelets on sterile plants descending from Family B was
made in 1914 on 281 plants bearing a total of 101,412
spikelets. In this count the number of fertile spikelets
was 3,857, corresponding to 3.78 per cent, of the total
number of spikelets. The latter figure may be regarded
as the average fertility of sterile plants in the progeny
of Family B.

No. 611]

ALLELOMORPHS

694

The fertile spikelets of sterile plants are generally scat-
tered at random over the panicle, and each fertile spikelet
may be regarded as representing a separate case of re-
version ; but in mosaic forms which show higher fertility
and are of rarer occurrence, the reversion may have
taken place in earlier stages of plant development, result-
ing in larger fertile sections. Consequently, when the
count of fertile spikelets is made with only the first type
of sterile plants, a more correct value for the frequency
of reversion may be obtained. The result of such a count
on 902 panicles containing 93,635 spikelets is 1,858 fertile
spikelets, i. e., 1.98 per cent, of the total number of
spikelets.

The mosaic forms appear in several different grades of
partial fertility. In a panicle either one or more branches
or even one half of the panicle can be highly or entirely
fertile, the remaining part being absolutely or nearly ab-
solutely sterile. Similarly, in a single plant some whole
panicles can be entirely or highly fertile while others are
of the ordinary grade of partial fertility. Furthermore,
similar mosaic conditions were also observed in single
flowers of sterile spikelets. While all six anthers of a
sterile spikelet generally bear none or but few pollen
grains, occasionally flowers appear in which certain
anthers contain a considerable number of pollen grains
of normal appearance and others show the ordinary state
of sterility. Hence it may be assumed that the reversion
can take place at any sta'ge of plant development.

The partial homozygosity of heterozygotes, correspond-
ing to the partial fertility of sterile plants, may be esti-
mated in the following way. Assuming that the possi-
bility of reversion at any stage of a plant's life, similar
to that observed above, may also occur in heterozygotic
cells, then we may distinguish for convenience two differ-
ent types of reversions; there is the reversion which will
cause partial homozygosity within a single flower, and the
reversion which will produce an entirely homozygotic
spikelet or larger homozygotic sectant. Suppose then
that the latter reversion will give to the heterozygote

695

THE AMERICAN NATURALIST

[Vol. LI

liomozygotic (AA) spikelets in any part "x" of the total
number of spikelets which is taken as a unit, and again
that in the remaining (1 — x) part of the total number of
spikelets, the other type of reversion will occur, turning
some part "yn of the whole generative tissue taken as a
unit from the Aa state to the AA state. For simplicity,
however, we may substitute "x" for "y" in the above re-
lation, because it seems presumable that a similar prob-
ability of reversion may exist constantly all through the
plant life. Such a plant will have the following consti-
tution in regard to the generative tissue:

x(AA) + (1 -x)[x(AA) + (l-x)(Aa)].

As the result of self-pollination, the progeny of such a
parent plant will show the constitution :

x(AA) + (1 - x) [J(l + x)2(AA) + id - x2) (Aa)
+ i(l-x)2(aa)].

Applying arbitrary values to "x" in this formula, we
shall get numerical relations among segregates. In
Table V the results of such calculation are compared with
results obtained by the experiments in 1913-1915. Thus
we may find the average partial homozygosity of hetero-
zygotes around 4 to 6 per cent., the average partial fer-
tility of sterile plants being, as was already shown, ca.
4 per cent.

TABLE V

Calculations on Data of Table IV

(AA+Aa)

4 %

77.88 %

22.12%

38.74%

61.53%

5 %

78.57

21.43

39.69

60.31

6 %

79.24

20.76

40.89

59.11

9,918
78.96%

2,643

21.04%

; 94

41.05%

135
58.95%

It has also been noticed that the sterility concerned is
associated with an abnormality represented by the be-
havior of chlorophyll at the ripening of seeds. While, at
the ripening season, the chlorophyll in the fertile sections
of the mosaic forms turns to yellow just as in ordinary
fertile plants, the chlorophyll in the sterile sections still

No. 611]

ALLELOMORPHS

696

remains green. The fertile spikelets occurring in a small
number on the otherwise sterile panicle appear on rip-
ening as yellow spots scattered among green spikelets;
the plants with both sterile and fertile panicles appear in
the fall also as mosaic forms with green and yellow
leaves. This feature of the sterile plants is in direct
contrast to the behavior of the mosaic plants with the
variegated and the entirely green leaves studied by De
Vries and Correns.

The observations in the foregoing pages seem to paral-
lel those made by the authors cited at the beginning of
this paper. In the present investigation, however, there
was observed also the transformation of allelomorphs in
the opposite direction, that is, the transformation of the
dominant allelomorph into the recessive allelomorph,
something scarcely mentioned in the investigations re-
ferred to above. The observations in this regard were in
brief as follows.

In the first place, the spontaneous occurrence of segre-
gating families was observed again among the descend-
ants of the families which had proved in the experiments
already described to be constantly fertile. This suggests,
just as did the occurrence in Family A and Family B in
1912, the probability of the A A cell changing into the Aa
cell.

In the second place, a constant tendency of the dom-
inant allelomorph to be transformed into the recessive
allelomorph was observed in certain strains. In 1913,
special attention was paid to such segregating families in
which the excess of recessive segregates over the theo-
retical expectation was particularly high. Although, as
already noted, the variation among the segregating fam-
ilies in 1913 with regard to the deviations from the reces-
sive proportion might possibly have arisen from experi-
mental errors associated with a certain probability of alle-
lomorphic reversion from recessive to dominant, yet it
was deemed not impossible that the very considerable
excess of recessives exhibited by certain families might be
caused by other reasons. This point was seemingly de-

697

THE AMEBIC AN NATURALIST

[Vol. LI

cided by the experiment made with Family Z?80 in 1913
(Table VI), since in this family there was noticed a con-
stant tendency toward the allelomorphic transformation
under consideration.

TABLE VI

The Segregation of Family B/80 and Its Descendants

Year

No. of
Families

Parent-
plants

No. of
Indi-
viduals

Fertile
Plants

Sterile
Plants

Recessives

Deviation of
Recessives

P. E.

1913

Fertile

30.30%

+ 5.30%

2.95%

1914

1,020

727

293

28.73

+ 3.73

0.91

1915

435

309

126

28.89

+ 3.89

1.40

1916

11,013

7,832

3,181

28.88

+ 3.88

0.28

Total. .

114

Fertile

12,567

8,937

3,630

28.89%

+ 3.89%

0.26%

1914 (a) .

16i

Sterile

199

147

23.62%

- 1.38%

2.04%

1914(6) .

131

Sterile

100

95.00%

+70.00%

2.92%

1915

592

548

516

94.16

+69.16

1.25

1916

1202

1,436

1,337

93.11

+68.11

0.77

Total. .

1923

Sterile

2,084

136

1,948

93.47%

+68.47%

0.64%

1 Derived from the family in 1913, i. e., Family B/80.

2 Derived from the group (6) in 1914.

3 Excluding the group (a) in 1914.

In Table VI there is beside the ca. 4 per cent, ex-
cess of recessives in the families derived from fertile
parents, a remarkable excess of recessives in the families
descended from the sterile parents in the group (b) in
1914. The sterile plant of this type could not be distin-
guished from those which, as was shown in Table IV,
gave segregating families with an excess of dominants in
the intensity of the partial fertility as well as in the be-
havior of chlorophyll at the ripening of the seeds. Con-
sequently, it may be presumed that although these two
types of sterile plants have the same genetical constitu-
tion originally, the dominant allelomorphs resulting from
the reversion of their recessive allelomorphs are of dif-
ferent stabilities in the dominant state ; that is, in the first
type of sterile plants such dominant allelomorphs are
very easily re-transformed into the recessive state, while
in the second type the corresponding dominant allelo-
morphs tend to remain in the reverted condition.

No. 611]

ALLELOMORPHS

(>98

Corresponding to the excess of recessive segregates, a
deficiency of dominant homozygotes among dominant seg-
gregates was also noticed. Among 153 families derived
from fertile plants in the experiment above mentioned, 40
families, were uniformly fertile, the remaining 113 fami-
lies showing segregation. The former, therefore, is 26.14
per cent, of the total number of families, and shows
7.19 per cent deficiency from the theoretically expected
percentage, 33.33 per cent., the probable error being
± 2.68 per cent.

In conclusion it may be stated that the allelomorphs
concerned in this investigation are probably subject to
reversible transformations, and that the probable fre-
quency of the allelomorphic transformation may be prac-
tically constant in a certain strain, and possibly may be
different in different strains. As to the conditions under
which such allelomorphic transformations take place,
nothing is yet certain except that these conditions are of
a hereditary nature. The manner in which different in-
tensities of allelomorphic transformations are inherited
will be the subject of further investigation.

A word may be added here regarding the conception
of dominance and recessiveness. Bateson's theory of
"presence and absence of f actors' ' is sometimes under-
stood in the sense that the dominant allelomorph is re-
garded as due to the real presence of an hereditary mate-
rial unit which is absent in the recessive allelomorph.
Such a conception is not in full accordance with the idea
of the reversible transformability of allelomorphs as de-
scribed in this investigation. There is another possibility
of the nature of allelomorphs. The dominant and the re-
cessive allelomorphs may be supposed to represent two
alternative conditions or phases of a single hereditary
substance, somewhat resembling the chemical conception
of polymerization. Consequently, the interchangeability
between the dominant and recessive allelomorphs is not
improbable theoretically.

Busset Institution,
August 26, 1917

[Reprinted from The American Naturalist, Vol. LII„ January, 1918.]

MATERNAL INHERITANCE IN THE SOY BEAN

H. TERAO

The Imperial Agricultural Experiment Station, Tokyo, Japan

The soy bean, Glycine hispida Maxim., shows as differ-
ent types two cotyledon colors, yellow and green. The
beans with yellow cotyledons have two types of seed-coat
colors, namely, green and yellow, while the beans with
green cotyledons have always green seed-coats.1 The in-
heritance of these types of cotyledons and of seed-coats
has been proved by the author's experiments to be ma-
ternal. A brief notice of the experiments will be given
in the following.

The green and yellow colors of cotyledons and seed-
coats are obviously attributed to chlorophyll, which, •
on the ripening of the beans, is either changed from green
into yellow or remains green. Further, according to the
author's observations, the chlorophyll in the vegetative
parts of the plant shows the same behavior as the chloro-
phyll of the cotyledons; in other words, the leaves and
stems of the varieties with yellow cotyledons turn to a
yellow color when they are gradually dying coincident
with the ripening of the beans, while those of the varieties
with green cotyledons remain green sometime after the
dying of the whole plant. These facts suggest that the
two types of cotyledon colors may represent two kinds of
chlorophyll, one which changes into yellow under certain
physiological conditions and one which is not so affected.
The chlorophyll of the seed-coats, however, seems to be-
have somewhat differently from the chlorophyll in all

i Black and brown pigments also appear in the seed-coats of certain
varieties. These pigments are entirely independent of the green and yellow
colors here referred to in their inheritance, but they make the latter colors
invisible or at least indistinct. By proper crosses, however, one can test
whether a seed-coat covered by the black or brown pigment belongs to the
green or the yellow category.

THE

AMERICAN NATURALIST

[Vol. LII

£ *0

§§

>»

»o /
bo \

8 ^
©ON

b£

>> O
a

=3 CM ^

g oo ^

bjO N*

o o

o J
83 co

a c
>> o

lO CO

8 a

0> CD-

S-l Sh

bO bl)

8 co

M O

M V

Z to

o ^

-83 »

>> O

Is.

CD Z

bD >

83 00 81
CP CM tO
>> CM

b£ 1>

CO i-H

tH CO

O .83
■£ °

O "rt

k CO

o 3

i §

.2; co

>> a

8 13
CU CM

8 *
CO 83 -
bfi >>

S S3

o ra
«- 8
bfi O

00 i-i
rf< O
CM lO

8 O

No. 613] INHERITANCE IN SOY BEAN

other parts of the plant, since, as was already noted, yel-
low cotyledons are accompanied by green seed-coats in
certain varieties.

The crossing experiments which have been made by the
author since 1910 with these different types of beans have
produced the results shown in Table I, the main facts
being summarized as follows.

I. The F2 cotyledons of the crosses reciprocal to each
other are of the same character as the female parents.
In respect to the cotyledon colors, the F2 and following
generations show the characters of the ¥\ generation ex-
clusively, instead of a Mendelian segregation between the
yellow and green colors. Hence we are probably dealing
with characters which can be inherited only through the
female parents.

II. The inheritance of the seed-coat colors is a more
complicated phenomenon. In the cross 1 ' green cotyle-
dons, green seed-coat" (?) X 4 'yellow cotyledons, yellow
seed-coat" (c?), the green seed-coat is inherited through
the female parent exclusively, just as in the case of the
cotyledon colors; but in the reciprocal cross the green
and yellow seed-coats show Mendelian segregation, the
former being dominant.

The maternal inheritance observed above was not due
to self-fertilization succeeding failures in artificial cross-
ing, because several other characters showed inheritance
through the male parents.

An interpretation of the inheritance phenomena under
consideration is suggested as follows. In the first place,
let us refer again to the two different kinds of chlorophyll
assumed to be concerned in producing the green and yel-
low cotyledons; namely, the chlorophyll which can be
changed into yellow and the chlorophyll which remains
green. (These will be denoted respectively as "(Y)"
and "(G)" in the later descriptions.) These character-
istics of chlorophyll may be due to heritable traits of the
chromatophores or of the cytoplasm, and not to hered-
itary elements in the nucleus. As, on the fertilization of

THE AMERICAN NATURALIST

[Vol. LII

the egg-cell, the chroinatophores and the cytoplasm of the
female gamete will probably remain as such without
being supplemented by those from the male gamete, their
characteristics would naturally be inherited only through
the female parent. In the second place we may assume
that a pair of Mendelian factors is concerned in the inher-
itance of the colors of the seed-coats. The factor "Hn
inhibits the chlorophyll "00" in the seed-coat of the
beans with yellow cotyledons from changing to yellow,
producing beans with yellow cotyledons and green seed-
coat; the absence of the factor "H," expressed by "h,"
allows the seed-coat of the bean with yellow cotyledons to
remain yellow. The seed-coat of the bean with green
cotyledons remains green no matter whether the factor
"H" is present or absent, because the beans of this kind
have the chlorophyll " (G) " which is incapable of chang-
ing the color.

The justice of the contention regarding the bean with
green cotyledons, moreover, is supported by the following
observations. The F2 families of the crosses "green
cotyledons, green seed-coat" (?)X "yellow cotyledons,
yellow seed-coat" (c?) were actually composed of two
kinds of individuals which were distinguishable from each
other by a slight difference of the intensity of green color
in the seed-coats, and the numerical relation between these
two kinds of individuals was approximately the Mendelian
mono-hybridal segregation ratio, the darker seed-coat
being dominant to the lighter one. Again, in the F3 gen-
eration of these crosses, there were obtained three types
of families, two which were uniformly of the darker and
of the lighter seed-coats respectively and one which was
a mixture of both. By comparing the green seed-coats of
the female parents in these crosses with those of the prog-
eny, the former was found to belong to the darker class
mentioned above. These variations in the green color of
the seed-coats may be regarded as being due to the in-
fluence of the Mendelian factors "H" and "h" respec-
tively on the chlorophyll "(G)"; from which it follows

No. 613]

INHERITANCE IN SOY BEAN

that the method of inheritance in the beans with yellow
cotyledons obtains also in the beans with green cotyle-
dons.

Keeping these statements in mind the cases in Table I
may be illustrated as follows:

Parents
Cotyledons
Seed-coat

Cotyledons
Seed-coat

Crossing No. I

(G)HH (9)X(Y)hh (tf)
green yellow
green vellow

\ /

(G)Hh
green
green

Crossing No. II
(G)HH (9)X(Y)HH (<?)
green yellow
green green

\ /

(G)HH
green
green

Cotyledons
Seed-coat

Parents
Cotyledons
Seed-coat

Cotyledons
Seed-coat

(G)HH (G)Hh (G)hh

25% 50% 25%
green
green

Crossing No. Ill

Y(hh) (9)X(G)HH (<?)
yellow green
vellow green

\ /

Y(Hh)
yellow
green

(G)HH
100%
green
green

Crossing No. IV

Y(HH) (9)X(G)HH (cf)
yellow green
green green

\ /

Y(HH)
yellow
green

Cotyledons
Seed-coat

(Y)HH (Y)Hh (Y)hh

25% 75% 25%

yellow yellow

green yellow

(Y)HH
100%
yellow
green

If the foregoing interpretation really represents the
facts in this investigation, we may consider also crosses
in which forms such as (G)Hh, (G-)hh, and (Y)Hh were
used as the parents, since in these crossings phenomena
different from those in Table I would be expected. These
expectations have been fulfilled in further experiments in
which individuals from the previous experiments repre-
senting different intensities of seed-coat color were used
as the parent plants. The results of these crosses, accom-
panied by interpretations, are shown in Table II.

THE AMERICAN NATURALIST [Vol. LIT

Crosses

TABLE II

MADE AMONG THE PROGENY OF THE HYBRIDS SHOWN IN TABLE I

Parents

Female Male

No. of
Indi-
viduals

Crossing
VII .. .

No.

Crossing
VIII .

No.

Crossing
IX . ..

No.

Cotyledons

Seed-coat

Interpret.

yellow
yellow
(Y) hh

green
green
(G) hh

Cotyledons

yellow

green

Seed-coat

green

Interpret.

(Y) Hh

(G) hh

Cotyledons

yellow

green

Seed-coat

green

Interpret.

(Y) Hh

(G) Hh

Char-
acter

yellow

vellow

(Y) hh

100%

yellow

f green

\ yellow

J (Y) Hh

50%

l(Y)hh

50%

yellow

j green

\ yellow

f (Y) HH

25%

\ (Y) Hh

50%

I (Y) hh

25%

NO. Of

Indi-
viduals

2,381

1,963

1,108

The maternal inheritance described in this paper seems
to be essentially the same phenomenon as the inheritance
of the character " albo-niaculata" which was studied by
Correns2 in Mirabilis Jalapa and also by Baur3 in Antir-
rhinum majus. In each case one is dealing with chroma-
tophore characters.

Harvard University, Bussey Institution,

2 Correns, C, Zeitschr. f. ind. Abst. u. Vererbungslehre, Bd. I, 1909,
pp. 291-^329; Ibid., Bd. II, 1909, pp. 331-340.

3 Baur, E., Zeitschr. f. ind. Abst. u. Vererbungslehre, Bd. IV, 1910, pp.
81-102.

THE INHERITANCE OF DOUBLENESS IN CHELID ONI UM

MAJUS LINN.

KARL SAX

Harvard University, Bussey Institution, Forest Hilts, Massachusetts

Reprinted from Genetics 3 : 300-307, My 1918

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle Edward M. East

Harvard University Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport
Carnegie Institution of Washington

Bradley M. Davis

University of Pennsylvania

Rollins A. Emerson
Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan

Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 191 6, at the post office at
Princeton, N. J., under the act of March 3, 1879.

THE INHERITANCE OF DOUBLENESS IN CHELIDONI UM

MAJUS LINN.

KARL SAX

Harvard University, Bnsscy Institution, Forest Hills, Massachusetts

Reprinted from Genetics 3 : 300-307, My 1918

THE INHERITANCE OF DOUBLENESS IN CHELID ONIUM

MAJUS LINN.

KARL SAX

Harvard University, Bussey Institution, Forest Hills, Massachusetts
[Received September 3, 1917]

The most thorough investigation of the inheritance of doubleness in
flowers has been made by Miss Saunders (191 o, 191 1, 191 7). In the
Petunia the double flowers set no seed and all crosses must be made with
the single plant as the female parent. In the F± of such a cross there
is either a segregation into 3 single : 1 double, or into 9 single : 7
double. All single plants selfed or crossed inter se produce only sin-
gles. Sauxders explains these results on the assumption that pollen
from the single flower used carries only factors for singleness, while
ovaries of the single flowers lack the factor for singleness in some cases
and possibly in all. The pollen of double flowers is further assumed
to be heterozygous. In the stock the inheritance of double and single
flowers is explained on the assumption that two factors are involved
which are linked in the pure singles, but net linked in the eversporting
singles. It is also assumed that ''single" factors are distributed only in
the female gametes according to a system of partial linkage.

In the Welsh poppy (Meconopsis cambrica) Saunders (1917) found
that a simple 3 : 1 ratio was obtained in F2 by crossing single and double
plants. Doubleness was found to be dominant. In the hollyhock (Al-
thaea rosea, A. fici folia) the offspring of single X full double plants
were intermediate in F1 and in the F2 produced a 1 : 2 : 1 ratio. In
the carnation (Dianthus caryophyllus) a cross of double X single pro-
duced an Fo ratio of 3 double : 1 single, while a cross of double X sin-
gle in Dianthus barbatus (sweet william) gave the same F2 ratio, but
singleness was found to be dominant.

Norton (1907) and Batchelor (1912) have found that doubleness
in the carnation is of two types; the ordinary or standard double and
the type known as "bullhead" or "buster". By crossing a full double
or buster with the single the resulting F1 is standard double. The Ft

Genetics 3: 300 My 1918

INHERITANCE OF DOUBLENESS IN CHELIDONIUM

selfed or crossed inter se produces an F2 ratio of i single : 2 standard
doubles : 1 full double or buster.

The doubling of flowers is usually due to petalody of the stamens,
according to de Vries. This is true of the plants investigated by Saun-
ders, and has been recorded by many writers, Goebel (1913), Masters
(1869), de Vries (1906), and others; but no statistical data have been
presented to show the degree of correlation between stamen number and
petal number.

The inheritance of doubleness in Chelidonium majus and a statistical
study of the relation of petals and stamens will be considered in the pres-
ent paper. A double- and a single-flowered plant growing near the
Bussey Institution were transplanted and reciprocal crosses made.
In the Fx about sixty plants of each cross were grown and in each case
produced practically an equal number of single and double plants. Seed
from the F-l singles and doubles were planted separately. Of 133 plants
raised from seed of single F1 plants, 109 were single and 24 double. Of
in plants raised from seed of the double Fx plants, 6 were single and
105 double. If we assume that the high number of singles resulting
from Fx singles, and the 6 singles among the doubles, were due to con-
tamination by crossing, or possibly due to volunteers from wild plants
growing near by, then it appears that singleness is dominant, and that
the original cross was made with a heterozygous single. We would then
expect a ratio of 1 : 1 in the F1 irrespective of which plant is used as
the female parent. In the F2 the Fx single segregates should give a
ratio of 3 single : 1 double, while the Fx double segregates should breed
true, when selfed or crossed inter se. In the F2 of this particular cross
all of the doubles should be homozygous, while two-thirds of the singles
should be heterozygous.

Of the 244 plants raised in the F2 the petal number and stamen num-
ber of the flowers of 147 plants were found. An average of 20.6 flowers
per plant were counted.

The average numbers of petals and stamens of the F2 plants are shown
in table 1. The mean petal number is 10.54 ± 0.31, with a standard
deviation of 5.60 ± 0.22. The mean stamen number is 18.30 ± 0.26,
with the standard deviation of 4.72 ± 0.19. The coefficient of correla-
tion between stamen and petal number is — .90 ± .01. In table 2 all
of the flowers of the 147 F2 plants are plotted in respect to stamen num-
ber and petal number. The mean petal number is 10.89 ± °-°7 with a
standard deviation of 5.84 ± 0.05, while the mean stamen number is

Genetics 3: My 1918

302

KARL SAX

Table i

Average stamen and petal number of 147 F2 plants. Chelidonium, single X double.

Stamens

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

5
6

7
8

9
10
^•11

■g 12
^ 13
14
15
16

17
18

19
20
21
22

6 23 10 4 I

56
0
0
0

I
I

6
6

15
22
6
6
7
8
4
4
0
1

6 13 14 11 13 12 10 5

6 6 23 10

1 147

Petals

M = 10.54 — 0.31

cr = 5.60 ± 0.22

r = — .00

* Not weighted for deviation in number of flowers counted.

Stamens

M = 18.30 ± 0.26
cr = 4-72 ± O.19

.01

18.14 ± 0.06 with a standard deviation of 5.15 ± 0.04. The coefficient
of correlation is — .863 ± .003. The variation of the doubles is espe-
cially striking in table 3, and there is apparently a continuous gradation
from single to full double.

There is no significant variation of the petal number (four) in the
singles, while the mean number of stamens is 23.68 ± 0.14 with a
standard deviation of 1.56 ± 0.09. Although the doubles are recessive,
the variation of petal number and of stamen number is much greater
than in the singles. The F2 doubles are plotted in table 3. The mean
number of petals is 14.56 ± 0.20 with a standard deviation of 2.81 ±
0.14, and the mean stamen number is 14.99 ±0.17 with a standard de-
viation of 2.39 ± 0.12. The coefficient of correlation is — .58 ± .05.
It is apparent that the singles, even though two-thirds of them are heter-
ozygous, are much less variable than the recessive doubles.

The sum of the petal number and stamen number is about the same in
all individuals, whether single or double. The mean sum of the petals

■ — >

•si

CO O C\
vg CM

hH 00

04 co lo On

CO Tf O vovo 00 OVO O CM O CO Tf vo hh
O co O 00 m p-H 0 On vo CM

On
CM

On CM

00
CM

hH co

CM °*

hH CM

vo
CO

co cm

hh o CM
CM

co CM,

co
cm

CM ^ VO

hH hH

COOO
On

tJ-00

^VO i-i CM

CO l~l hH

vo cm
On

00 H M Tj-

CO CM CO hh

On CM
vo

tNVO VO Tj"

NLOl^M ION HH

CM
CO

VO IO00 tN. VO CM HH hH HH
CM HH

conO O VO

ONl^O'OvorooOhH h m
tJ- CM CM hh

hH covo vo

O "<frO "+ O CO iO vo W hh ro
CO CM hh _

HH CM 00 On

O CO HH HH IJ-i CO TfNO HH

vo Tf co CM CM

hH vo On

00 VO vo £J tJ- OVO ^tVO tJ- CM
co CM CO CM CM

-a-

CM Tj" Tj"

CM 00 vo On txvo hh io txvO vo co
cOCO^tOJCMHHCMHH

00
CM

00 hh hVO KN vovo 00 00 CO hh

HH hH hH CM CM HH hH

co f^00 CO Tt co CM cot^O ^ CM

HHhHCM^CMhHCMCMHHhH

CM co CM CO Ov <"0 covO O00 co CM

00 NOm M OnvO O co CM CM

22
oo

o»

CM hh hH IO00 O CO CM hh

M H TfTf N ^OCf CM HH

VO
CO

HH CO

o d

+1 +1

HH hH

00 io

175

f> vo

q o
o d

+1 +1

00 00

d vo

co vovO txOO O O hh

CM co tJ- vovo tN.00 O O hh CM co vovo t>
MHHhHHHhHhHhHhHCMCMCMCMCMCMCMCM

304

KARL SAX

and stamens of the doubles is 29.49 ±0.15 with a standard deviation ot
2.24 ± 0.12, while in the singles the mean sum of petal number and
stamen number is 27.68 db 0.14 with a standard deviation of 1.56 ±
0.10. The sum of the petal number and stamen number in the doubles is
significantly larger than in the singles, but with a much greater variation

Table 3

Average stamen number and petal number of F2 double plants. From table 1

Stamens

9 10 11 12 13 14 15 16 17 18 19 20 21

9
10
11
12
13

% 15
* 16
17
18

20
21
22

1 2 4 0 13 14

Petals

M = 14.56 ± 0.20

cr = 2.81 ± O.14

13 12 10

6
6

22
6

6
7
8

4
4
o
1

Stamens

M = 14.09 ± 0.17
a = 2.39 ± 0.12

.58 ± .05

If the pedigree culture results did not show the singles to be dominant,
one might expect the larger number of petals and stamens and the greater
variability of the doubles to be due to heterozygosis. The behavior of
the doubles may, however, be due to splitting of the stamens in some
cases, in addition to petalody.

The greater variation in number of parts in the double flowers loses
its significance when the doubles of the F2 are grouped by themselves
and compared with the double flowers on individual double plants of the
F2. The mean petal number of all of the F2 double plants is 14.56 ±
0.20 with a standard deviation of 2.81 ± 0.14 and the mean stamen
number is 14.99 ± O-1? with a standard deviation of 2.39 ± 0.12
(table 3). In tables 4, 5 and 6, three F2 double-flowered plants are
plotted in respect to petal number and stamen number. In table 4 the

INHERITANCE OF DOUBLENESS IN CHELIDONIUM

305

Table 4

Stamen number and petal number of flowers of a single F2 'double' plant.

Stamens

Petals

M = 17.23 ± 0.20

<7 = 2.77 ± 0.14

Stamens

M = 15.64 ± 0.18
a = 2.43 ± 0.13

r — — .46 ± .04

Table 5

Stamen number and petal number of flowers of an F2
'double' plant.

Stamens

14 15

jjn 16

2 I

& 17

£ 18

3 3

Petals

Stamens

M = 17.15 ±

0.16

M =

11.36

± 0.18

a —

1.71 ±

0.1 1

<7 =

1.08

± 0.13

- .64

•05

mean petal number is 17.23 ± 0.20 with a standard deviation of 2.77 ±
0.14 and the mean stamen number is 15.64 ± 0.18 with a standard devia-
tion of 2.43 ± 0.13. In table 5 the mean petal number is 17.15 ± 0.16

Genetics 3: My 1918

306

KARL SAX

Table 6

Stamen number and petal number of flowers of an F2 'double' plant.

Stamens

9 10 ii 12 13 14 15 16 17 18 19 20 21 22

10
11

• 13

^ 15

* 16
17
18

19
20

I 2

I I

II I
2 I 2 I

12 2

2 12
I I

I I

I I I I
I

Petals

M = 13.29 ± 0.22
a = 2.11 ± 0.16'

Stamens

M = 16.58 ± 0.30
a- = 2.88 ± 0.21

.66 ± .06

with a standard deviation of 1.17 ± 0.11 and the mean stamen number is
11.36 ± 0.18 with a standard deviation of 1.98 ± 0.1.3. ^n table 6 the
mean petal number is 13.29 ± 0.22 with a standard deviation of 2. 11 ±
0.16 and the mean stamen number is 16.58 ± 0.30 with a standard
deviation of 2.88 ± 0.21. The coefficient of correlation of all the F2
doubles is — .58 ± .05 while the coefficients of correlation of the three
F2 double plants are — .46 ± .04. — .64 ± .05, and — . 66 ± .06, re-
spectively. The variability of the petal number and stamen number of
all of the F2 doubles is, in general, not greater than the variability in
individual F2 double plants.

CONCLUSIONS

Doubleness appears to be a simple recessive character in Chclidonium
ma jus.

There is apparently a continuous series in degree of doubling from
singles to full doubles in the F2.

There is much greater variation in the doubles, which are recessive,
than in the singles, of which two-thirds are heterozygous. The F2
doubles are however no more variable than individual double plants of
the F2.

There is a high degree of negative correlation between petal number
and stamen number in the F2, due to petalody.

INHERITANCE OF DOUBLENESS IN CHELIDOXIUM

307

LITERATURE CITED

Batchelor, L. D., 1912 Carnation breeding. Ann. Rept. Amer. Breeders' Ass. 7 :
199-205.

De Vries, H., 1906 Species and varieties ; their origin by mutation, pp. xvii -f 874.

Chicago : Open Court Pub. Co.
Goebel, K., 1913 Organographie der Pflanzen. Erster Teil. pp. — .
Masters, M. T., 1869 Vegetable teratology, pp. xxxviii -f- 534. London : Ray

Society.

Norton, J. B., 1007 Heredity in carnation seedlings. Ann. Rept. Amer. Breeders'
Ass. 3: 81-82.

Saunders, E. R., 1910 Studies in the inheritance of doubleness in flowers. 1. Pe-
tunia. Jour. Genetics 1 : 57-69.
191 1 Further experiments on the inheritance of "doubleness" and other charac-
ters in stocks. Jour. Genetics 1 : 303-379.
1917 Studies in the inheritance of doubleness in flowers. II. Meconopsis. Al-
thaea, and Dianthus. Jour. Genetics 6 : 165-184.

■Genetics 3: My 1918

Information for Contributors

■4

The length of contributions to Genetics will not be arbitrarily limited, but article
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitation*, as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication.

Contributors are requested to observe the usual care in the preparation ef manu-
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published in a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer tij
them, and should be separated from the text by lines. All footnotes should be nun*
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fof
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4^iX7lA inches and all illustrations intended for use as single plates should not
exceed 5 X 7f4 inches, or should allow of reduction to these dimensions. Double
plates may be 12Y1 X Wa inches. Legends for all figures should be typewritten on a
separate sheet.

Contributors will receive 100 reprints of their papers, with covers, free of charge,
and may secure additional copies at actual cost of manufacture, provided these are
ordered when the corrected galley-proofs are returned.

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. Geof ?e H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, MAY 1918

TABLE OF CONTENTS

PAGE

La Rue, Carl D., and Bartlett, H. H., An analysis of the

changes involved in a case of progressive mutation 207

Hance, Robert T., Variations in the number of somatic chromo-
somes in Oenothera scitttillans De Vries 225

Goodale, Hubert D., Feminized male birds 276

Sax, Karl, The inheritance of doubleness in Chelidonium majus

Linn 300

THE BEHAVIOR OF THE CHROMOSOMES IN
FERTILIZATION

KARL SAX

Harvard University, Bussey Institution, Forest Hills, Massachusetts

Reprinted from Genetics 3 : 309-327, July 1918

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle
Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis
University of Pennsylvania

Edward M. East

Harvard University

Rollins A. Emerson
Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan
Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J., under the act of March 3, 1879.

THE BEHAVIOR OF THE CHROMOSOMES
IN FERTILIZATION

KARL SAX

Harvard University, Bussey Institution, Forest Hills, Massachusetts
[Received October 23, 1917]

The problems of heredity are intimately related to the two critical
stages in the life cycle of the higher organisms, the reduction divisions
of sporogenesis in plants and of gametogenesis in animals, and the cyto-
logical processes of fertilization. The first of these has received much
attention and is generally considered to be closely correlated with the
segregation of the factors which determine the characters of the next
generation. The union of the gametes at the time of fertilization is
equally important, although we know little of the behavior of the chro-
mosomes at this period. Recent research has made it desirable from
both a cytological and a genetical standpoint to examine critically their
behavior in the higher plants.

The behavior of the chromosomes in animals at the time of fertiliza-
tion is comparatively well known. In some cases the sexual nuclei unite
while in the resting condition so that the male and female contribution
of chromosomes can not be distinguished. It has been shown, however,
that in many species the male and female chromosomes are formed be-
fore the fusion of the sexual nuclei and that they may maintain the two
distinct groups during the first division of the zygote. In some species
the independence of male and female chromosomes has been traced
through several divisions and even to later cleavage stages. Indeed there
are reasons for believing that the male and female chromosomes may
maintain their independence until gametogenesis. In all cases the first
division of the fertilized egg appears to be essentially similar to any
other normal somatic division.

In plants there is a great deal of variation in the behavior of the sexual
nuclei in fertilization. In the algae it has been shown that in some
cases the first division of the zygote is a reduction division. In Spiro-
gyra (Karsten 1908) it has been found that the first division of the zy-
gospore reduces the chromosome number from twenty-eight to fourteen.

Genetics 3: 309 Jl 1918

KARL SAX

Allen (1905 ) has found that the first division of the oospore nucleus of
Coleochaete is also a reduction division. In all cases described the
gametes unite completely and male and female chromosomes are not
found in separate groups.

The fungi afford an interesting variety of phenomena in connection
with fertilization. As Harper ( 1910) points out. they have served to
enlarge our conception of the sexual process elsewhere. In the rusts
a fusion of two cells may occur independently of nuclear fusion as was
shown by the work of Blackmax 1.1904") and Christmax ( 1905).
The nuclei remain separate throughout the sporophytic phase until they
fuse in the teleutospore just before reduction. The long period in which
the nuclei exist side by side in the sporophyte without fusing and their
final fusion prior to reduction is significant. There is no question but
that the male and female chromosomes are independent in the sporo-
phytic generation. Xo less striking are the conditions found in the
Ascomycetes. Harper's ( 1905 ) work on the mildews has shown the
existence of two nuclear fusions in the life cycle, one at the origin of
the ascocarp where there is a fusion of differentiated gametes, and the
other in the ascus. The nuclear fusion in the ascus is followed by three
successive divisions which are thought to be correlated with the occur-
rence of the double fusion.

A most remarkable fusion which takes the place of the normal ferti-
lization has been described in Aspidium falcatum, an apogamous fern,
by Miss Allex" (1914). In this case the sporophyte develops through
a vegetative bud from the prothallus and has presumably the i.r number
of chromosomes. Sixteen spore mother cells are present as in other
ferns but these fuse in pairs to give eight cells from which a maximum
of thirty-two spores may be developed. This fusion sometimes occurs
while the spore mother cells are in the spireme stage so that two com-
plete spiremes are present in the fusion nucleus. The fusion is followed
immediately by the reduction division characteristic of spore formation.

The behavior of the chromosomes in fertilization has been described
for many species of Gymnosperms. It was found by Blackmax
(1898), Fergusox ( 1904 ). and Chamberlain ( 1899), that in Pinus the
male and female chromosomes are formed independently for the first
division of the fertilized egg. The independent formation of male and
female chromosomes has also been described in Tsuga (Murrill 1900),
Tuniperus TXorex 1907), and Abies (Hutchixsox 1915). All of
these accounts are in general accord with the conditions found in some
other plants and in many animals.

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 311

A most unusual behavior of the chromosomes in fertilization has
been described in Abies balsamea by Hutchinson ( 1915). The chromo-
somes of the male and female gametes are formed independently as is
the case in most conifers. As the two groups unite the chromosomes
are said to become paired side by side and to twist about one another
in a manner similar to their behavior in the reduction divisions. The
number of chromosome pairs is haploid. According to Hutchinson
there is then a transverse segmentation of the chromosome pairs. The
resulting pairs of chromosomes are diploid in number, and of about
half the length of the pairs before segmentation. The chromosomes are
described as then separating to form 4X chromosomes, half of the num-
ber passing to each pole in the first division. The description is not
supported by convincing figures.

Chamberlain (1916) in Stangeria paradoxa finds a similar pairing
of chromosomes at fertilization. Although he apparently upholds
Hutchinson's conclusion he does not state that in this form there is a
transverse segmentation of the chromosome pairs in the first division
of the egg.

In the angiosperms there is no detailed account of the behavior of
the chromosomes during the first division of the fertilized egg. In most
cases the sexual nuclei fuse while in the resting condition. Cases have
been described in Lilium (Guignard 1891), Cypripedium (Pace 1907),
and Fritillaria (Sax 191 6), where the gamete nuclei were rarely in the
spireme stage before fusion. The first division of the fertilized egg in
angiosperms has been described by Guignard (1891), Ernst (1902),
Goldschmidt (1916), Renner (1914) and the writer (Sax 1916),
but in no case have the descriptions been complete and rarely has more
than a single division figure been shown in any of the papers.

Atkinson (1917) in a recent genetical paper on Oenothera has pre-
sented some results which he maintains can be explained only on the as-
sumption that there is a segregation of factors in the Fx zygote. He
also states (p. 257) :

"The germ plasm is peculiarly sensitive to shock from the meeting of
sperm and egg, particularly when there is a genotypic difference between
the two germ plasms. This results more or less in interchange, crossing
over, dominance, as well as blending, of factors in the zygote, often ac-
companied by selection of factors into different associations in different
zygotes giving rise to more than one hybrid type in the F1 generation of
crosses."

Atkinson's statements are largely theoretical, and, as Davis (i9J7)

Genetics 3: Jl 1918

312

KARL SAX

has indicated, have little experimental and no cytological basis, yet we
must consider such possibilities in the present study.

The purpose of this study is to consider not only the behavior of the
chromosomes of the gametic nuclei, but also in the "triple fusion." Al-
though it is questionable if we may consider the triple fusion as a real
fertilization, yet so far as the inheritance of endosperm characters is
involved it is quite comparable to the union of the gametic nuclei. For
this reason and also because of the great importance of endosperm
characters in cereal breeding, we will give some attention to the chromo-
somes in the triple fusion in both Fritillaria and Triticum. It is also
necessary to describe briefly the development of the embryo sac in Fritil-
laria in order to understand the chromosome number found in the first
division of the endosperm nucleus.

MATERIAL AND METHODS

The material for this paper was secured from about four thousand
cases of fertilization in Fritillaria pudica Spreng. and not less than two
hundred cases of fertilization in Triticum durum hordciformc Hort. var.
Kubanka. Flemming's stronger solution and chrom-acetic acid were
used as fixatives for the Fritillaria ovaries. At the suggestion of Dr.
Osterhout about 0.5 percent of urea salts were added to the above
fixatives and gave excellent results with the wheat ovaries. Besides as-
sisting as a fixative, the urea salts reduce the surface tension of the so-
lution and aid in the penetration of the fixative. Sections were cut from
10 n to 20 fi thick. Modifications of Flemming's triple stain, and Heid-
enhain's iron alum haematoxylin were used in staining with good results.

THE CHROMOSOMES IX FRITILLARIA

A brief consideration of the chromosome number in the development
of the embryo sac in Fritillaria is necessary before describing the later
stages. Xumerous counts of the chromosome number in the various
stages of embryo sac development were made. In the heterotypic divi-
sion of the megaspore mother cell twelve chromosomes pass to each
pole. The second division usually appears to be normal and the four
resulting nuclei each receive twelve chromosomes. These nuclei pass
into the resting stage. The nuclei in the third division present consider-
able variation in respect to chromosome number. The two nuclei at the
micropylar end of the embryo sac have the usual ix number of chromo-
somes and when they divide twelve chromosomes pass to each pole. One
of these resulting nuclei becomes the upper polar nucleus. The other

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 313

three constitute the nuclei of the cells of the egg apparatus, one of
which functions as the egg. One of the two nuclei nearest the chalazal
end of the sac in the four-nucleate stage disintegrates. The other,
which has received twelve chromosomes from the previous division, has
at the time of division not twelve chromosomes, but about twenty-four,
and in the division about twenty-four chromosomes pass to each pole.
One of the resulting nuclei becomes the lower polar nucleus. As a re-
sult of this development of the embryo sac the egg and upper polar
nuclei each contain ix or twelve chromosomes while the lower polar
nucleus possesses approximately 2x chromosomes. The lower polar nu-
cleus and the normal antipodal nucleus are considerably larger than the
other nuclei of the embryo sac, presumably because of the 2x number
of chromosomes which they contain.

The general behavior of the sexual nuclei in Fritillaria has been de-
scribed in an earlier paper (Sax 19 16). The contents of the male
nuclei while free in the embryo sac were found to be in an irregular
dark-staining network. The male nucleus soon after coming in contact
with the egg nucleus loses this netlike structure of its chromatin and
passes into the usual resting condition. The union of the sexual nuclei
occurs while they are in the resting condition as shown in figure I,
plate 1. The male nucleus can be recognized by its slightly smaller size.
In both nuclei there are nucleoli, and the chromatin is in a reticular net-
work. In the cases figured the nuclei are flattened at the point of con-
tact but in some cases the egg nucleus is somewhat indented. The cyto-
plasm is distributed uniformly throughout the lower part of the egg
cell, while the upper part is almost entirely occupied by a large vacuole.
The gametic nuclei remain distinct for a comparatively long time, often
until the endosperm has reached the four- or eight-nucleate stage.

The disappearance of the adjacent nuclear membranes of the sexual
nuclei results in a fusion nucleus which is also in the resting stage. At
first there is a trace of the outlines of the two nuclei but later the fusion
nucleus shows no indication of its binucleate origin. At this stage all
of the chromatin stains alike and it is impossible to distinguish the male
and female chromatin as separate contributions. Such a case is shown
in figure 2. The chromatin is in the resting condition and more nucle-
oli are present than in the nucleus of the unfertilized egg.

The fusion nucleus remains in the resting stage but a short time.
The chromatin forms a fine threadlike structure and the nucleoli begin
to disappear. Figure 3 shows an early spireme stage where the spireme
is not completely formed. Delicate threads can be seen among the

Genetics 3: Jl 1918

3i4

KARL SAX

heavier chromatin threads. The spireme thread thickens until it ap-
pears as illustrated in figure 4. Here the nucleoli have not entirely dis-
appeared and traces of the nuclear membrane may be seen. The broken
appearance of the spireme is partly due to the passage of the microtome
knife through the edge of the nucleus. But in many other spiremes ex-
amined it was also impossible to demonstrate the presence of a single
continuous spireme.

After the formation of the spireme the nuclear membrane disappears
entirely. The spireme contracts and becomes somewhat thicker just be-
fore the segmentation into chromosomes occurs. The segmentation of
the spireme is shown in figure 5. The nuclear membrane has completely
disappeared and the segmenting spireme has contracted. Surrounding
the nuclear cavity is a dense area of cytoplasm from which delicate
threads pass to the newly formed chromosomes.

A significant variation in spireme formation of the zygote has been
described for Fritillaria in an earlier paper (Sax 19 16, fig. 21). In
this case the sexual nuclei were found in the spireme stage before
fusion. This condition is very rare in the fusion of the egg and male
nuclei, but is not uncommon in the nuclei of the triple fusion.

The chromosomes in the metaphase of the first division are long and
often cover much of the spindle. The spindle is bipolar and the chromo-
somes are in a single group and not paired. The number of chromo-
somes at this time is twenty-four. A longitudinal section of the meta-
phase of the zygote is presented in figure 6. All of the chromosomes
are not shown in this figure, but in this and the two following figures
the proportion of chromosomes omitted, due to their presence in an
adjacent section, is about the same. The cytoplasm at this time is more
uniform than in the preceding stages.

As the chromosomes divide twenty-four pass to each pole. There is
no evidence that the male and female chromosomes are separated in two
groups. The spindle is typically bipolar. The anaphase of the first
division of the zygote is shown in figures 7 and 8. In figure 7 the
chromosomes are not so large as in the metaphase. In figure 8 the
chromosomes are nearer the poles and they are larger than in the pre-
ceding figure. The spindle at the lower end appears to be three-parted,
possibly due to the fact that it is cut somewhat obliquely. The other
pole has no appearance of being divided. In neither case figured nor
in many others examined is there any evidence that the chromosomes
pair.

In the telophase the chromosomes form a single compact group at

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 315

either pole. The fibers between the poles thicken and the cell plate is
formed. A clear telophase is shown in figure 9. Although the chro-
matin groups are very compact the outlines of the chromosomes are still
apparent.

The second division of the embryo appears to be the same as any other
normal somatic division. Often one of the cells of the two-celled em-
bryo divides before the other, but usually they divide at the same time.

The second male nucleus and the two polar nuclei usually fuse in the
resting condition. Often, however, the nuclei are found in the spireme
stage before they fuse and occasionally before the upper polar nucleus
and male nucleus have joined the lower polar nucleus. A striking case
of the latter condition is shown in figure 10. The nuclei are in early
spireme stages. The difference in the size of the nuclei is evident. In
this triple fusion the large lower polar nucleus contributes about 2x
chromosomes while the upper polar nucleus and the male nucleus each
contribute ix chromosomes. Thus $x chromosomes are contributed by
the maternal parent and only ix by the male parent.

The number of chromosomes in the metaphase of the first division
of the "endosperm nucleus" is approximately 4X or forty-eight. From
figure 11 it is evident that this division appears to be normal. There
is no segregation of chromosomes into groups, no pairing of chromo-
somes, and the spindle is not multipolar.

The chromosomes of the metaphase split longitudinally and about
forty-eight daughter chromosomes pass to each pole. In the anaphase
shown in figure 12 it was impossible to show all of the chromosomes.
Apparently an equal number of chromosomes pass to each pole and the
division in every way appears to be regular.

THE CHROMOSOMES IN TRITICUM

The mature embryo sac in wheat consists of the egg cell, two syner-
gids, two polar nuclei in the primordial endosperm, and a large number of
antipodal cells. The cytoplasm of the egg cell is more or less vacuolate
near the periphery, but a denser layer is found around the nucleus.
The egg nucleus contains one large nucleolus and often several small
ones. The chromatin is arranged in a loose irregular network. The
polar nuclei come in contact with each other, but do not fuse, before
fertilization. They are very large and are about equal in size. The
cytoplasm of the "Endospermanlage" extends from the polar nuclei to
the egg apparatus.

The male nucleus as it enters the egg cell is small, elongated and its

Genetics 3: Jl 1918

316

KARL SAX

contents are so dense that it appears almost homogeneous in structure.
It takes a brilliant red stain with the safranin and is easily recognized.
A male nucleus free in the cytoplasm of the egg cell is shown in figure
13. It is long and curved at the ends. The structure of the male
nucleus appears almost homogeneous except for small vacuoles and
darker-staining particles. There is no trace of a separate cytoplasmic
sheath around the male nucleus.

The male nucleus when in contact with the egg nucleus is often more
or less coiled, but has the same structure as when free in the egg cell.
Such a case is presented in figure 14. Immediately around the male
nucleus the cytoplasm is less dense and a few delicate threads cross this
vacuolate area. This condition was also found in other preparations.
The chromatin of the egg nucleus at this time is still in the resting con-
dition.

The male nucleus apparently penetrates the egg nucleus before in-
creasing in size. Many stages were found where the compact spireme
of the male nucleus could be seen inside of the membrane of the egg
nucleus. Its appearance is striking and unmistakable because of its den-
sity and darker-staining properties. In figure 15 the male nucleus is
shown inside of the egg nucleus near the membrane. It has increased
in size and the irregular compact spireme thread can be seen. There are
several small nucleoli in the male contribution of chromatin. The chro-
matin of the egg nucleus is in the early spireme stage and portions of
the spireme thread are often more or less parallel. In figure 16 the
early spireme of the male chromatin is especially distinct. Its dark-
staining closely coiled spireme is quite distinct from the loose spireme
of the egg nucleus. The nuclear membrane is still apparent at this stage.
In both cases figured the large nucleolus of the egg nucleus is shown
quite separate from the male chromatin.

The segmentation of the two spiremes of the zygote results in 2.x
chromosomes which are unpaired. The metaphase of the first division
of the fertilized egg is illustrated in figure 17. There are approximate-
ly twenty-eight chromosomes to be seen in this section. Several pieces
of chromosomes, or possibly whole chromosomes, were found in an ad-
jacent section. There is no indication that the chromosomes contributed
by the male and female gametes lie in separate groups and they are not
in pairs. In several chromosomes shown in this figure a longitudinal di-
vision may be seen. This longitudinal splitting of the chromosomes
continues until all of the chromosomes have divided, as illustrated in
figure 18. There are about twenty-one pairs shown in this figure and

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 317

there are six or seven pairs in an adjacent section. These split chromo-
somes are comparatively long and are often curved. The daughter
halves are most widely separated at the center and in some cases they
appear to be slightly twisted about one another. After the completion
cf the longitudinal fission 2,r daughter chromosomes pass to each pole.
An anaphase of the first division is shown in figure 19. The daughter
chromosomes are about as long as the mother chromosomes of the meta-
phase and are often curved. The spindle is clearly bipolar. The dense
layer of cytoplasm so conspicuous about the egg nucleus in the earlier
stages is apparent around the spindle.

The chromosomes pass to each pole in a single group. In figure 20
approximately twenty-eight chromosomes may be counted at each pole
in the late anaphase. The chromosomes have shortened a little and are
in compact groups. It will be noted that the dense area of cytoplasm in
the egg cell is for the most part around only the upper pole, while the
chromosomes of the lower pole lie in the more vacuolate cytoplasm.
The chromatin of the telophase passes into the resting stage, a cell plate
is laid down and the two-celled embryo is formed.

The male nucleus which unites with the polar nuclei appears to be
quite similar to the one which unites with the egg nucleus. In figure 21
the triple fusion is shown. The dark-staining male nucleus is coiled at
the side of one of the polar nuclei. The polar nuclei are large and of
about equal size. Each contains a large nucleolus atad the chromatin is
in granules in an open network.

At the time of spireme formation the chromatin contributed by each
of the three nuclei may be distinguished. Such a case is shown in fig-
ure 22. Here the spireme of the male nucleus is in a comparatively
compact ball at the upper side of the figure. The spireme of each polar
nucleus is rather open, very irregular, and quite distinct. The nucleolus
and outlines of each polar nucleus may be seen, but the nuclear mem-
branes have almost entirely disappeared.

The chromatin of the male nucleus and polar nuclei maintain their
independence not only in the spireme stage, but apparently also in the
early division stages. Figures 23 A and 23 B are drawn from adjacent
sections of the first division of the triple fusion nucleus. This is the
early metaphase when the chromosomes are on the nuclear plate. There
are approximately forty chromosomes to be counted. The chromo-
somes appear to be in three groups, each group containing approximately
an equal number of chromosomes. The spindle in figure 23 A is clearly
two-parted at the upper end and more or less so at the lower end. The

Genetics 3: Jl 1918

3i8

KARL SAX

spindle in figure 23 B appears to be distinct from the spindles in figure
23 A. Later stages in the first division of the triple fusion nucleus were
not observed, but judging from many counts of later endosperm di-
visions, the first division is regular and an equal number of chromo-
somes pass to each pole. The fact that there are about forty chromo-
somes in the endosperm divisions indicates that the polar nuclei each
contribute ix chromosomes. After the first division the endosperm
nuclei divide rapidly and cell walls are not formed in the early develop-
ment of the endosperm.

DISCUSSION

In many animals and in most gymnosperms the independence of the
formation of the male and female chromosomes following fertilization
is unquestioned. There are a few cases in the angiosperms where the
independent formation in the zygote of the spiremes from the gametes
has been demonstrated. In an earlier paper (Sax 19 16) the writer de-
scribed the gamete nuclei of Fritillaria, in rare cases, as in the spireme
stage before fusion. In wheat the spireme of the male and female
nuclei are formed separately even though the male nucleus is within the
egg nucleus at the time of spireme formation (figures 15 and 16). In
both Fritillaria and Triticum the nuclei of the triple fusion may be in
the spireme stage before fusion. In Triticum the- chromosomes con-
tributed by each nucleus appear to be in more or less separate groups
even in the metaphase of the first division. Although there is con-
siderable variation in the condition of the sexual nuclei at the time of
fusion, I believe that we are justified in assuming that the male and
female chromosomes are formed independently in the zygote, even in
cases where the nuclei fuse in the resting condition.

The behavior of the chromosomes during the first division of the egg
in Fritillaria and Triticum is essentially not different from any other
normal somatic division. There is no indication that the chromosomes
of the male and female gametes are in separate groups, or that they
pair in the first division of the zygote. In Triticum the chromosomes
in the metaphase, when first split longitudinally, appear much like certain
stages in Abies, which Hutchinson (191 5) interprets as a pairing of
chromosomes. But in Triticum the chromosomes of the metaphase be-
fore splitting are distinctly 2.r in number and after splitting there are
2x pairs. If a pairing of male and female chromosomes occurred we
would expect to find only ix pairs of chromosomes.

In neither Triticum nor Fritillaria is there anything comparable to

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 319

a reduction division in the fertilized egg nor are there any irregularities
which might account for a zygotic segregation of factors. It is pos-
sible, as Atkinson (1917) suggests, that there is a "shock" as the re-
sult of the meeting of genotypically different germ plasms, which may
cause an irregular behavior of the first division of the zygote. The
only cytological work which might support this theory is that of Gold-
schmidt (1916). He explains the occurrence of patrocliny in Oeno-
thera crosses as a result of merogony. There is no good cytological
evidence of any behavior which would explain a segregation of factors
during the first division of the zygote. The segregation of factors in
the Fj as described by Atkinson, might be accounted for on the as-
sumption that in Oenothera the chromosomes of the zygote behave as
described in Abies. But this assumption would lead into other problems
more difficult to explain. It is much more probable, as Davis (19 17)
has suggested, that Atkinson is dealing with genotypically impure
parent species, and that his assumptions concerning the segregation of
factors in the zygote are entirely unnecessary.

The behavior of the chromosomes of the zygote in Abies, as described
by Hutchinson (1915), if verified, would be of considerable genetical
significance to those who consider the chromosomes the bearers of the
hereditary factors. According to Hutchinson the chromosomes of the
male and female nuclei pair in the zygote so that there are ix pairs.
Each pair segments transversely, forming 2x pairs of chromosomes.
The chromosome segments separate and 2.x chromosomes pass to each
pole. There is no longitudinal division of the chromosomes as is the
case in the usual type of somatic division. If we assume that the he-
reditary factors are located in the chromosomes it is obvious that if
homologous chromosome segments pass to opposite poles the heredi-
tary factors will be segregated according to the usual Mendelian be-
havior. We would then expect a segregation of characters in Fx indi-
viduals in crosses when the parents differ in Mendelizing characters. If
we assume that not only are the hereditary factors located in the chro-
mosomes, but also in a definite linear arrangement in the chromosomes,
it is evident that the chromosomes can not divide longitudinally at one
time and transversely at another time, as described in Abies, without
causing chaos in the distribution of the hereditary factors. Further-
more if this process continued through several generations each chromo-
some would finally contain but a single factor and most of the factors
would be lost. Certainly the behavior of the chromosomes in Abies, as

Genetics 3: Jl 1918

320

KARL SAX

described by Hutchinson, can not be brought into harmony with the
theory of a linear arrangement of the factors in the chromosomes. To
be sure the latter theory is only a working hypothesis, but the research
of Morgan and his students, and other experimental work, can best be
explained by it.

If in Abies there is a double "reduction,'' one at the time of sporo-
genesis, the other during the first division of the zygote, we might ex-
pect a compensating fusion at some period in the life cycle. A compar-
able case is found in the mildews where the double fusion is thought to
be followed by a double reduction in the ascus. In view of the present
study and the theoretical questions raised by Hutchinson's results, it
would be well to examine more critically the behavior of the chromo-
somes in the first division of the zygote in Abies.

The pairing of the chromosomes in somatic cells has been described
in Galtonia, Funkia, Oenothera, Thalictrum and Yucca. It appears
that the pairing of the chromosomes may take place at the time of
fertilization, as Hutchinson maintains is the case in Abies and as
Chamberlain (1916) has described in Stangeria, or that the pairing
may not occur until synapsis as is true in the rusts. There may be in-
termediate examples where the chromosomes pair at various times in
the sporophytic life cycle. It is possible as Miss Fraser (1912) sug-
gests, that "the clearest cases of Mendelian inheritance will perhaps be
those correlated with a late association of the chromosomes in pairs."
Experimental work, however, makes it appear probable that the chromo-
somes are quite independent, and that if there is any influence or inter-
change of factors between chromosomes it does not occur until synap-
sis. According to Harper (1910) the behavior of the chromosomes in
the rusts "is certainly strongly suggestive that synapsis and its accom-
panying phases represent a stage of mutual influence if not of inter-
change of physical material between the chromosomes much more inti-
mate than any which has preceded it in the life of the sporophyte."

In Fritillaria and in Triticum there is no evidence that the chromo-
somes pair in fertilization. The male and female spiremes are formed
independently. At synapsis there is a pairing of male and female
chromosomes. We might expect that here two spiremes are formed,
one male and one female, as in fertilization, and that these spiremes or
portions of these spiremes, pair side by side (parasynapsis), rather
than expect that homologous chromosomes become arranged alternately
end to end and then pair as in telosynapsis. In Aspidium where the

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 321

spore mother cells fuse while in the spireme stage there is a reduction
division immediately following. In this case there is probably a pairing
of homologous chromosomes as is generally thought to occur in reduc-
tion divisions. It is more difficult to assume that here the chromosomes
should follow the telosynaptic method of reduction than to assume that
the spiremes conjugate side by side as in parasynapsis. In either case,
however, the physical difficulties seem to be very great. If there is a
somatic pairing of chromosomes in cases where the telosynaptic method
of reduction occurs we might expect the male and female chromosomes
to become arranged alternately end to end at or soon after fertilization.

In the development of the embryo sac in Fritillaria twenty-four
chromosomes are found in the metaphase of the third division at the
chalazal end of the embryo sac, and as a result of this division the lowTer
polar nucleus receives about 2x or twenty-four chromosomes. A simi-
lar doubling of the chromosome number at the chalazal end of the
embryo sac has been described in Lilium martagon by Guignard
(1891), Mottier (1897), Sargant ( i 896), and Strasburger (1908).
According to Strasburger this doubling is due to a premature longi-
tudinal splitting of the chromosomes and is dependent on the food sup-
ply of the plant. In Fritillaria the lower polar nucleus which contains
about 2x chromosomes is considerably larger than the upper polar nu-
cleus which contains ix chromosomes. Since both nuclei are surrounded
by a common cytoplasm, the difference in size is probably due to the dif-
ference in chromosome number. In Triticum the polar nuclei are of
equal size and so probably have the same chromosome number, presum-
ably ix each. The difference in size between the polar nuclei and the
egg nucleus is apparently due to the nuclear cytoplasmic relations.

In the triple fusion in Fritillaria there are about $x chromosomes
contributed by the female parent and ix chromosomes contributed by the
male parent. If we assume that the hereditary factors are borne by the
chromosomes, then the female parent contributes three sets of factors
and the male parent contributes only one set. This phenomenon is of
considerable importance in relation to the multiple factor hypothesis.
For instance Hayes and East (1915) have found in a particular cross
in corn that the endosperm character borne by the mother was always
dominant. One of the parents used had a starchy endosperm, the other
a corneous endosperm. Either way the cross was made the F1 endo-
sperm was always of the same kind as the endosperm of the mother
parent. This type of inheritance is explained on the assumption that the

Genetics 3: Jl 1918

322

KARL SAX

two factors contributed by the mother, as a result of polar fusion, are
always dominant over the single factor contributed by the male parent.
The effect of multiple factors is clearly demonstrated here, and if the
female parent contributes not two, but three factors, the maternal con-
tribution should be still more potent.

It is unnecessary to consider the morphological character of the endo-
sperm because this subject has been dealt with fully by East (1913). It
may be added, however, that the development of the endosperm is inde-
pendent of the development of the embryo. While working with Mr.
E. F. Gaines at Washington State College, I found about a dozen
cases in wheat where the endosperm had developed normally and yet no
embryo was present. One such case has been found in corn. Such a
condition, although very rare, is of theoretical interest.

The significance of double fertilization has been a puzzle to many
botanists. The suggestion of Nemec (1910) seem to be the best expla-
nation and has some experimental proof. Nemec maintains that an
endosperm whose characters are contributed by both parents is better
suited to the nourishment of the embryo than an endosperm of wholly
maternal origin. Stingl's (1907) work supports this theory. He
has transferred embryos of various cereals to their own and other endo-
sperms. In most cases the embryo grew better on its own endosperm
than on that of another genus. Preliminary experiments of my own in
transferring hybrid corn embryos to the parent endosperms and vice
versa, have not as yet given definite results. In view of Stingl's re-
sults and my own work it appears that the experimental error is too
large to permit small differences in adaptability of embryo and endo-
sperm to be noted. If we assume that a hybrid endosperm is better able
to nourish the hybrid embryo than an endosperm derived from the
mother only, we might expect that "wider" crosses could be made among
plants where the endosperm is derived from both parents.

CONCLUSIONS

A complete series of stages in the first division of the fertilized egg
has been found in both Fritillaria pudica Spreng. and in Triticum durum
hordeiforme Host. var. Kubanka.

In Fritillaria the male and female nuclei are of about equal size be-
fore fusion. The sexual nuclei usually unite while in the resting con-
dition, although occasionally they are in the spireme stage at the time
of fusion. The presence of a single continuous spireme in the zygote

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 323

could not be demonstrated at any time. In the first division of the
fertilized egg each chromosome contributed by the male and female
gametes splits longitudinally and twenty-four chromosomes pass to each
pole.

In Triticum the male nucleus is small and almost homogeneous in
structure even when in contact with the egg nucleus. The male nucleus
enters the egg nucleus and forms a separate compact spireme. At the
same time the spireme of the egg nucleus is formed. In the first division
approximately twenty-eight chromosomes are found in the metaphase.
These chromosomes split longitudinally and the 2x number pass to each
pole.

In both Fritillaria and Triticum the first division of the zygote ap-
pears to be essentially like any other normal somatic mitosis. The male
and female chromosomes are formed independently; they are not found
in separate groups, nor do they pair. There is no chromosome behavior
at this time which might account for a zygotic segregation of factors.

In the triple fusion in Fritillaria the nuclei are often found in the
spireme stage before fusion. The first division of the endosperm nu-
cleus appears to be very regular. The chromosomes are not paired and
there is no segregation of the chromosomes contributed by each nu-
cleus. In this division about 4X or forty-eight chromosomes pass to each
pole. This number is due to the fact that there is a doubling of the
chromosomes at the chalazal end of the embryo sac in the third di-
vision and as a result the lower polar nucleus receives about twenty-four
chromosomes. Thus in the triple fusion the female parent contributes
3«r chromosomes and the male parent contributes only ix chromosomes.
This phenomenon is of interest in relation to the multiple factor hy-
pothesis.

In Triticum the nuclei involved in the triple fusion form separate
spiremes. There is evidence that the chromosomes contributed by each
nucleus may remain more or less separate even in the metaphase of the
first division. Each nucleus contributes ix or about fourteen chromo-
somes.

In both Fritillaria and Triticum the chromosomes of the triple fusion
are formed independently, there is no pairing of chromosomes and the
first and following divisions appear to be regular.

This work was done under the direction of Dr. E. M. East to whom
I am indebted for suggestions and criticisms. I am also indebted to
my wife for much assistance.

Genetics 3: Jl 1918

324

KARL SAX

EXPLANATION OF PLATES

All drawings were made with the aid of the camera lucida. The
figures were drawn from single sections with the exception of figure 22.
The magnification is as follows: . Figure 3, 11 20 diameters; figure 21,
900 diameters; all other figures, 1200 diameters. No reduction.

Plate i. — Fritillaria

Figure i. — Male nucleus and egg nucleus in contact.
Figure 2. — Complete fusion of male and egg nuclei.
Figure 3. — Early spireme of fertilized egg.
Figure 4. — Spireme of fertilized egg.

Figure 5. — Segmentation of spireme. Formation of chromosomes.
Figure 6. — Metaphase of the first division of the fertilized egg.
Figures 7, 8. — Anaphase of the first division of the fertilized egg.
Figure 9. — Telophase of the first division of the fertilized egg.
Figure 10. — Early triple fusion.

Figure ii. — Metaphase of the first division of the endosperm nucleus.
Figure 12. — Anaphase of the first division of the endosperm nucleus.

Sax, Karl, Behavior of the chromosomes in fertilization

Plate i

Genetics 3: Ji- 1918

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 325

LITERATURE CITED

Allen, C. E., 1905 Die Keimung der Zygote bei Coleochaete. Ber. d. d. bot. Gesell.
23 : 285-292, Taj. 13.

Allen, R. F., 1914 Studies in spermatogenesis and apogamy in ferns. Trans. Wis-
consin Acad. Sci. 17: 1-56, pis. 1-6.

Atkinson, G. F., 1917 Quadruple hybrids is Oenothera. Genetics 2:213-260.

Blackman, V. H., 1898 On the cytological features of fertilization and related
phenomena in Pinus silvestris L. Phil. Trans. Roy. Soc. 190:395-426, pis.
12-14.

1904 On the fertilization, alternation of generations and general cytology of the
Uredineae. Ann. Bot. 18 : 323-373, pis. 21-24.
Chamberlain, C. J., 1899 Oogenesis in Pinus laricio. Bot. Gaz. 27:268-280, pis. 4-6.

1916 Stangcria paradoxa. Bot. Gaz. 61 : 353-372, pis. 24-26.
Christman, A. H., 1905 Sexual reproduction in the rusts. Bot. Gaz. 39 : 267-275,
pi. 8.

Davis, B. M., 1917 The test of a pure species of Oenothera. Proc. Amer. Phil. Soc.
54 : 226-245.

East, E. M., 1913 Xenia and the endosperm of angiosperms. Bot. Gaz. 56:217-224.
Ernst, A., 1902 Chromosomenreduction, Entwicklung des Embryosacks und Be-

fruchtung bei Paris quadrifolia L. und Trillium grandiflorum Salisb. Flora

91 : 1-46, Taj. 1-6.

Ferguson, M. C, 1904 Contributions to the life history of Pinus. Proc. Washing-
ton Acad. Sci. 6: 1-202, pis. 1-24.

Fraser, H. I. C, 1912 The pairing of the chromosomes. New Phyt. 11:58-60.

Gold Schmidt, R., 1916 Nochmals iiber die Merogonie der Oenotherabastarde. Ge-
netics 1 : 348-353, pl- 4-

Guignard, L., 1891 Nouvelles etudes sur la fecondation. Ann. Sci. Nat. Bot. 14:
163-296, pis. 9-18.

Harper, R. A., 1905 Sexual reproduction and the organization of the nucleus in
certain mildews. Carnegie Inst. Washington. Publ. 37. 92 pp., 7 pis.
1910 Nuclear phenomena of sexual reproduction in fungi. Amer. Nat. 44:595-
603.

Hayes, H. K., and East, E. M., 191 5 Further experiments on inheritance in maize.

Conn. Agr. Exp. Sta. Bull. 188. pp. 1-31, pis. 1-7.
Hutchinson, A. H., 1915 Fertilization in Abies balsamca. Bot. Gaz. 60:457-472,

pis. 16-23.

Karsten, G., 1908 Die Entwicklung der Zygoten von Spirogyra jugalis Ktzg. Flora
99: i-ii, Taj. 1.

Mottier, D. M., 1897 Ueber das Verhalten der Kerne bei der Entwicklung des
Embryosacs. Jahrb. f. wiss. Bot. 31 : 125-158, Taj. 2, 3.

Murrill, W. A., 1900 The development of the archegonium and fertilization in the
hemlock spruce (Tsuga canadensis) . Ann. Bot. 14:583-607, pis. 31, 32.

Nemec, B., 1910 Das Problem der Befruchtungsvorgange und andere zytologische
Fragen. 532 pp. 5 Taj. Berlin : Gebriider Borntraeger.

Noren, C. O., 1907 Zur Entwicklungsgeschichte des Juniperus communis. Upsala
o

Univ. Arsskr. 64 pp., 4 pis.
Overton, E., 1893 tiber die Reduktion der Chromosomen in den Kernen der Pflanzen.

Vierteljahrsschr. Naturf. Gesell. Zurich 38:169-186.
Pace, L., 1907 Fertilization in Cypripedium. Bot. Gaz. 44 : 353-374, pis. 24-27.

Genetics 3: Jl 1918

^26

KARL SAX

Plate 2. — Triticum

Figure 13. — Male nucleus in the egg cell.

Figure 14. — Male nucleus in contact with the egg nucleus.

Figures 15, 16. — Male nucleus within the egg nucleus.

Figure 17. — Metaphase of the first division of the fertilized egg.

Figure 18. — Late metaphase of the first division of the zygote.

Figure 19. — Early anaphase of the first division of the zygote.

Figure 20. — Late anaphase of the first division of the zygote.

Figure 21. — Early triple fusion.

Figure 22. — Spireme stage in the triple fusion.

Figures 23A, 23B. — Adjacent sections of the early metaphase of the first division
of the triple fusion nucleus.

Genetics 3: Jl 1918

BEHAVIOR OF THE CHROMOSOMES IN FERTILIZATION 327

Renner, O., 1914 Befruchtung und Embryobildung bei Oenothera Lamarckiana und
einigen verwandten Arten. Flora 107:115-150, Taj. 12, 13.

Sargaxt, E., 1896 The formation of the sexual nuclei in Lilium martagon. Ann.
Bot. 10:443-477, pis. 32, 33-

Sax, K., 1916 Fertilization in Fritillaria pudica. Bull. Torrey Bot. Club 43 : 505-
522, pis. 27-29.

Strasburger, E., 1908 Chromosomenzahlen, Plasmastrukturen, Vererbungstrager
und Reducktionsteilung. Jahrb. f . wiss. Bot. 45 : 479-570, Taj. 1-3.

SfriNGL, G., 1907 Experimented Studie iiber die Ernahrung von Pflanzlichen Embry-
onen. Flora 97 : 308-331.

Genetics 3: Jl 1918

Information for Contributors

The length of contributions to Genetics will not be arbitrarily limited, but articlu
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitations as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication.

Contributors are requested to observe the usual care in the preparation of manu-
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published v? a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer to
them, and should be separated from the text by lines. All footnotes should be num
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fot
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4H X 7V2 inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be 1254 X 7Va inches. Legends for all figures should be typewritten on a
separate sheet

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, JULY 1918

TABLE OF CONTENTS

PAGE

Sax, Karl, The behavior of the chromosomes in fertilization 309

Harris, J. Arthur, Further illustrations of the applicability of
a coefficient measuring the correlation between a variable
and the deviation of a dependent variable from its prob-
able value 328

East, E. M., and Park, J. B., Studies on self-sterility. II. Pollen-
tube growth 353

Wright, Sewall, On the nature of size factors 367

Robbins, Rainard B., Some applications of mathematics to breed-
ing problems* III 375

Robbins, Rainard B., Random mating with the exception of sister

by brother mating 390

STUDIES ON SELF-STERILITY I. THE BEH/^IOR OF
SELF-STERILE PLANTS

E. M. EAST and J. B. PARK
Harvard University, Bussey Institution, Forest Hills, Massachusetts

Reprinted from Genetics 2:505-609, November 1917

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board
George H. Shull, Managing Editor

Princeton University

William E. Castle

Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Pennsylvania

Edward M. East

Harvard University

Rollins A. Emerson

Cornell University
Herbert S. Jennings

Johns Hopkins University

Thomas H. Morgan

Columbia University

Raymond Pearl

Maine Agricultural Experiment Station

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 191 6, at the post office at
Princeton, X. J., under the act of March 3, 1879.

STUDIES ON SELF-STERILITY I. THE BEHAVIOR OF
SELF-STERILE PLANTS

E. M. EAST and J. B. PARK

Harvard University, Bussey Institution, Forest Hills, Massachusetts
[Received March 19, 1917]
TABLE OF CONTENTS

PAGE

Introduction 506

The occurrence of self-sterility 506

Early work on self-sterility 507

Recent work on self-sterility 518

The material used and the general plan of the present investigations 528

The effect of the environment on self-sterility 530

Inter-specific pedigree culture experiments 538

Cross No. 1. Nicotiana Forgetiana X N. alata (self-sterile X self-sterile) . . 539

The F2 generation 539

The F, generation 539

The F3 generation 542

The F4 generation 545

The F5 generation 545

Argument on cross No. 1 549

Cross No. 2. N. alata X AT. Forgetiana (pseudo self-fertile X self-sterile) 555

Cross No. 3. N. Forgetiana X N. alata (self-sterile X pseudo self-fertile) 555

The Fa generation 557

Family D, N. alata plant 53 X plant 44 of class C, cross No. 3 561

Family E, N. alata plant 58 X plant 44 of class C, cross No. 3 565

Family F, plant 34 of class A X plant AA of N. Forgetiana 567

Family G, plant 44 of class C, cross No. 3 X plant AA of A7*. Forgetiana. . 571

Family H, plant 44 of class C, cross No. 3 X plant 10 of class A, cross No. 2 575
Family I, plant 44 of class C, cross No. 3 X plant 34 of class A, cross No. 2 583
Family J, plant 52 of class B, cross No. 3 X plant 23 of class A, cross No. 2 588
Family K, plant 52 of class B, cross No. 3 X plant 44 of class C, cross No. 3 592

Argument on cross No. 2 and cross No. 3 593

Cross No. 4. N. commutata X N. Forgetiana (self-sterile X self-sterile) 596

Argument on cross No. 4 598

Intra-specific pedigree culture experiments 598

Summary and interpretation of the results 601

Literature cited 607

Genetics 2: 505 X 1917

5o6

E. M. EAST AND J. B. PARK

INTRODUCTION

The occurrence of self -sterility

Among both hermaphroditic animals and plants forms are known in
which fertilization of the eggs by sperm or by pollen of the same indi-
vidual is difficult or even impossible. This condition is known as self-
sterility,1 although the term is not a happy one, since both the male and
the female gametes are morphologically perfect and are functional with
the complemental gametes of other individuals.

Self -sterility is probably a widespread phenomenon though its pres-
ence has been proved experimentally in comparatively few plants and in
only one animal. The result, one might even say the aim, of self-sterility,
however, is cross-fertilization. Regarded from this standpoint, it is to be
classed with the various other specializations of animals and plants, such
as morphological differences in the accessory sexual organs, dichogamy,
moncecism, dicecism, etc., which tend toward the same end; and since
these obvious contrivances for cross-fertilization are so numerous and so
dispread, it is difficult to believe that the less easily detected self-sterility
is rare, particularly as it has arisen independently in widely separated
groups.

The important role played by cross-fertilization in the evolution of
animals and plants may be attributed in some degree, therefore, to the
phenomenon of self-sterility; hence, any light thrown upon its meaning is
a contribution toward an explanation of the significance of cross-fertiliza-
tion in general.

Among animals only Ciona intestinalis has been proved to be self-
sterile (Castle 1896), though the condition is suspected in several
other forms.

Among Angiosperms self-sterility is rather generally distributed.
Knuth (1898, Vol. I, pp. 42-45) gives a list of 134 self-sterile species
representing 46 families and including both monocotyledons and dicoty-
ledons. This list is the best compilation of recorded cases and may be
considered fairly complete to-day as very few additional records have

1 The words self-incompatibility and self-impotence have been substituted for self-
sterility by various writers. These terms seem to us to be neither more nor less
objectionable than self -sterility, since neither takes into consideration the fact that
the same type of infertility may exist between different individuals. The important
point in the matter is that one should not confuse the phenomenon with any of those
types of true sterility where there is either complete or partial incapacity for the
production of gametes functional per se. For a discussion of the differences between
self-sterility and true sterility see Kraus (1915) and Stout (1916).

THE BEHAVIOR OF SELF-STERILE PLANTS

507

appeared in subsequent publications. It is naturally somewhat inaccu-
rate, inasmuch as several cases are recorded in which cross-pollination was
merely prevented by bagging the inflorescence or by isolating the plants
and self-pollination not insured. Nevertheless, at least 70 percent of the
records are properly proved cases of a self-sterility that is something
more than an ephemeral condition due to environmental changes or to a
fleeting period of reproductive inactivity that is normal in the life
history of so many plants.

There remain, then, in the neighborhood of 100 well endorsed in-
stances of self-sterility scattered over some 35 families. These families
are so different in their modes of reproduction that no general conclusion
can be drawn regarding the development of self -sterility. There are
legumes which are usually self-fertilized, and orchids that have developed
quite wonderful floral mechanisms favoring cross-fertilization; there are
showy flowers, and flowers peculiarly inconspicuous; there are flowers
with perfume, and flowers without it; there are anemophilous plants, and
plants that would be classed as strictly entomophilous. In certain genera,
such as Passiflora, there is a general tendency toward self -sterility; in
other genera, for example Verbascum and Nicotiana, closely related spe-
cies behave very differently.

In other words self-sterility has arisen many times, and often in
groups where there was apparently no need for it if the necessity is
assumed to be that of cross-fertilization. Not only is this an irresistible
argument in favor of the idea already expressed that only a small fraction
of the cases of self-sterility have been discovered and that self-sterility
has been a much more important factor in plant evolution than has pre-
viously been suspected, but it also indicates that certain of the mechanical
devices that have received great credit for promoting cross-fertilization
were inadequate for the needs of many plants.

EARLY WORK ON SELF-STERILITY

The discovery of self -sterility in plants probably should be credited to
Kolreuter, the first2 real student of hybridization, although his case is
somewhat doubtful. Kolreuter (1764) found that during two years
three plants of Verbascum phceniceam set no seed with their own appar-
ently good pollen, although they seeded readily with pollen of V. Blat-

2Thos. Fairchild crossed Dianthus caryophyllus with D. barbatus in 1719, and
Linneus brought his hybrid between Tragopogon pratensis and T. porrifolius into
flower in 1759, but neither of them contributed to the world any important facts
regarding hybridization.

Genetics 2: N 1917

508

E. M. EAST AND J. B. PARK

taria, V. nigrum, V. phlomoides and V. Lychnitis. Later these plants
showed sporadic fertility alternating with sterility of pollen or of eggs
or of both sex-cells, so that this instance may be only one of induced
true sterility due to conditions. It seems to deserve priority as an in-
stance of self-sterility, however, for Darwin (1872, p. 341) found
V. phceniceum and V. nigrum to be self-sterile, although the related
species V. Thapsus and V. Lychnitis were self-fertile.

Sprengel (1793), the other important hybridist of the 18th century
does not mention the subject.

Several true instances of self-sterility were discovered by Herbert
(1837) in his experiments with the Amaryllidacese. He says :

"Nine very fine crosses of Hippeastrum were flowering [there] at the
same time; one a natural seedling from Johnsoni or Regio-vittatiim, two,
J ohnsoni-pulveralentum, one J ohnsoni-vittatum, one psittacino- Johnsoni
crossed again by vittato-J ohnsoni, one from Johnsoni by solandriflorum,
and two from vittato-J ohnsoni by the same. Being desirous of blending
again these plants which were all cross-bred, different flowers were touched
with pollen from their several neighbors and ticketed, and other flowers were
touched with their own pollen. Almost every flower that was touched with
pollen from another cross produced seed abundantly, and those which were
touched with their own either failed entirely or formed slowly a pod of in-
ferior size with fewer seeds, the cross impregnation decidedly taking the
lead."

"It is only from the superior efficacy of the pollen of another plant that
we can account for the circumstances of some hybrid plants, which breed
freely with plants of either parental stock and fecundate them, not pro-
ducing seed readily when left to themselves ; for if their pollen is able to
fertilize and their ovary to be fertilized, there can be no positive sterility in
the plant, though there may be a want of sufficient energy under certain, or
perhaps under ordinary, circumstances.,,

These observations of Herbert referred to hybrids, though he also
found self-sterility in the species Zephyranthes carinata, and Darwin
in discussing them very properly sets them apart from the cases of self-
sterility in pure species. We shall show later, however, that absolute
self-sterility exists both in pure species and in hybrids, and is one and the
same phenomenon. In fact Herbert himself very nearly demonstrated
this. In a letter to Darwin (1875) written in 1839, he states that after
a duplication of these experiments with like results, he was led to make
similar trials on a pure species. He selected a plant of Hippeastrum auli-
cum which he had recently imported from Brazil. Three of its flowers he
selfed without result; a fourth flower he crossed with pollen of a triple
cross between H. bulbulosum* regince and vittatum and obtained good
seed.

3 Probably H. rutilum Herb.

THE BEHAVIOR OF SELF-STERILE PLANTS

509

Later work cited by Darwin (1875) ^so supports this idea. Bidwell
in New South Wales found Amaryllis belladonna to be partially self-
sterile, though fertile to the pollen of other species. E. Bernet, of
Antibes, a man having a wide experience in crossing species of Cistus,
found that their hybrids when fertile (he does not mention the pure
species) were completely self-impotent. His statement is that, quoting
Darwin, "the flowers are always sterile when the pistil is fertilised by
pollen taken from the same flower or from flowers on the same plant."
"But," he says — without the italics — "they are often fertile if pollen be
employed from a distinct individual of the same hybrid nature, or from a
hybrid made by a reciprocal cross." A. Rawson, a well known English
horticulturist, found the same absolute self-sterility in various named
varieties of Gladiolus that were said to have descended from Gandavensis,
an old race produced by crossing G. natalensis by G. oppositiflorus. The
interesting point in Rawson's work was that none of the plants of tb
same variety would set seed when interpollinated. As each variety haa
been propagated asexually by bulbs, he was of course actually dealing
with plants of the same germinal constitution, though under somewhat
different environmental conditions. For this reason it is extremely im-
probable that these were cases of induced true sterility.

"Altogether, Mr. Rawson, in the year 1861 fertilised twenty-six flowers
borne by four varieties with pollen taken from other varieties, and every
single flower produced a fine seed capsule; whereas fifty-two flowers on the
same plants, fertilised at the same time with their own pollen [which had
been proved to be good by the crosses], did not yield a single seed capsule."

Returning to the phenomenon as exhibited in pure species, Wm.
Mowbray, gardener of the Earl of Mountn6rris, in a letter to the Secre-
tary of the Horticultural Society (England), dated October 29, 1830,
states that he could get fruit only from Passi flora alata and P. racemosa
by reciprocal fertilization.

Observations on self-sterility in this genus continued to be made later
by a number of observers. The most important work was done by
Robertson Munro (1868). Munro found P. alata, P. racemosa, P.
cccrulea, P. Bellottii, P. kermesina, P. holosericea and P. fulgens to be
self -sterile, while Darwin obtained evidence that P. laurifolia and
P. quadrangular is were in the same condition. The evidence of perfect
self -sterility in the first three species is incontrovertible, in the remaining
species it is highly probable.

Some of the details from Munro's work are exceedingly interesting.

Genetics 2: N 1917

E. M. EAST AND J. B. PARK

In the first place he found plants of P. alata to be highly fertile with their
own progeny as the following quotation shows.

"I impregnated a considerable number of these flowers with their own
pollen, everyone of which proved abortive. But on impregnating eighteen
flowers on the mother plant with pollen from her own self-impotent seed-
lings, I got eighteen fine plump ovaries full of seed."

Again, Munro found that self-sterile plants were sometimes cross-
fertile and sometimes cross-sterile with plants of the same species and
presumably of the same generation. For example, three self-sterile
plants of P. cocrulea all produced seeds with pollen from one other plant.
The same experiment on P. alata showed cross-sterility in two instances
and cross-fertility in one instance.

A curious case of a return to self-fertility in P. alata through grafting
was also reported by Munro. Mr. Donaldson, gardener at Keith Hall,
grafted a self-sterile plant upon stock of an unknown species. Though its
pollen still refused to fertilize certain other plants of the same species, it
was markedly self-fertile and fertile with at least one other plant. Seed-
lings from this plant were all self-sterile but were fertile with the mother
plant.4

Gartner (1849), wno was among the most reliable of the early
hybridizers, found a number of self-sterile species. Dianthas japonicus
was. sterile both with its own pollen and with the pollen of D. barbatus.
Two plants of Lobelia f idgens likewise proved self-sterile. Their pollen
was good on L. cardinalis and L. syphilitica, their ovules could be fer-
tilized by the pollen of these species, but self-pollination yielded nothing.
A plant of Verbascum nigrum was also completely self -impotent though
fertile as a male with V. Lychnitis and V. anstriacam and fertile as a
female with V. Thapsus.

Similar conditions in certain exotic orchids were reported by Scott
(abstract 1863, complete paper 1865). A duplicate of a table in his
paper and a summary of his conclusions follow.

Scott and Munro (Darwin 1875) eacn independently found
Oncidium sphacelatnm also to be wholly self-sterile after some three
hundred attempts at self-pollination, though the species was fertile
reciprocally with other Oncidiums. Munro in addition confirmed
Scott's observations on O. divaricatum and added O. flexuosum to the
list of self-sterile plants.

4 It is likely that this phenomenon is similar to the pseudo self-fertility due to
conditions, which is discussed later in this paper.

THE BEHAVIOR OF SELF-STERILE PLANTS

Unions between Oncidium microchilum, 0. divaricatum var. cupreum and
0. ornithorhynchum.

)er of flowers
tilized

number of
ules produced

)er of good
sules

ated number

seeds

Estimated number
of good seeds

By calculation

Total

caps'

Numl
cap

Total
seeds

Good
seeds

0. ornt//t. X
0. micro. (No. 2)

20200

4242

or as 1000 to 210

0. micro. (No. 2) X
0. ornith.

1 2

0. ornith. X
0. micro. (No. 1)

23360

3737

or as 1000 to 160

0. micro. (No. 1) X
0. ornith.

0. divar. cup. X
0. micro. (No. 2)

22050

7938

or as 1000 to 360

0. micro. (No. 2) X
0. divar. cup.

I >

/•

0. divar. cup. X
0. micro. (No. 1)

26240

8922

or as 1000 to 340

0. micro. (No. 1) X
0. divar. cup.

17700

1434

or as 1000 to 420

0. micro. (No. 1) X
0. micro. (No. 2)

45800

34350

or as 1000 to 750

10.

(j. micro. \\1\0. 4) /\
0. micro. (No. 1)

ii.

0. micro. (No. 1) X
own pollen

12.

0. micro. (No. 2) X
own pollen

"By a summary comparison of these results we have the following highly
interesting facts disclosed. First, we see that the male element of 0. micro-
chilum (No. 1) will fertilise the female element of the two distinct species
O. ornithorhynchum and O. divaricatum var. cupreum and yet be completely
impotent upon its own female element ; nevertheless the susceptibility of the
latter (female element) to fertilisation is shown by its fertile unions with
another individual of the same species, and likewise by a fertile union with
an individual of a distinct species, namely 0. divaricatum var. cupreum.
Secondly, the male element of O. microchilum (No. 2) will fertilise the
female element of O. ornithorhynchum and O. divaricatum var. cupreum,
and likewise another individual of its own species, though on its own female
element it is utterly ineffective."

Genetics 2: X 1917

512

E. M. EAST AND J. B. PARK

These observations, together with similar ones on 0. Cavendishianum
recorded by Lecoq (1862) from the experiences of Riviere were made
on hot-house plants and Darwin originally attributed their self-sterility
to the peculiar conditions under which they were grown. He was
forced to modify his conclusions, however, through information received
from Fritz Muller. The latter self-fertilized over one hundred flowers
of Oncidiuni flcxuosiim at Desterro, Brazil, where it is native, without
obtaining a single seed, but he did discover the important fact5 that each
plant was fertile with the pollen from any other plant.

Scott and Muller each independently made the further discovery
that the tissue of the style of the self-sterile plants was penetrated freely
by the pollen tubes after selfing, though fertilization did not subsequently
occur.

As Darwin noted :
"Another observation made by Fritz Muller is highly remarkable, namely
that with various orchids the plant's own pollen not only fails to impregnate
the flower, but acts on the stigma, and is acted on, in an injurious or poison-
ous manner."

We have not been able to find any confirmation of these results, and it
seems entirely probable that the apparently poisonous action of the pollen
after an "illegitimate" pollination, might have been due to the action of
bacteria or fungi, since the work was done under tropical conditions.
But the facts are so exceptional that we give Darwin's (1875, vol. 2. p.
112) account.

"Fritz Muller observed the poisonous action of the plant's own pollen
in the above mentioned Oncidium flexu-osum, 0. unicorne, pubes (?), and
in two unnamed species. Also in two species of Rodriguezia, in two of
Notylia, in one of Burlingtonia, and of a fourth genus in the same group.
In all these cases, except the last, it was proved that the flowers were, as
might have been expected, fertile with the pollen from a distinct plant of
the same species. Numerous flowers of one species of Notylia were fer-
tilised with pollen from the same raceme ; in two day's time they all withered,
the germens began to shrink, the pollen masses became dark brown, and not
one pollen grain emitted a tube. So that in this orchid the injurious action
of the plant's own pollen is more rapid than with Oncidium flexuosum
Eight other flowers on the same raceme were fertilised with pollen from a
distinct plant of the same species ; two of these were dissected and their
stigmas were found to be penetrated with numberless pollen tubes ; and the
germens of the other six flowers became well developed. On a subsequent
occasion many other flowers were fertilised with their own pollen, and all
fell off dead in a few days ; whilst some flowers on the same raceme which
had been left simply unfertilised adhered and long remained fresh. We

5 It is probable that cross-sterility existed, but was not discovered.

THE BEHAVIOR OF SELF-STERILE PLANTS

513

have seen that in cross unions between extremely distinct orchids the pollen
long remains undecayed ; but Notylia behaved in this respect differently ; for
when its pollen was placed on the stigma of Oncidium flexuosum, both the
stigma and pollen quickly became dark brown, in the same manner as if the
plant's own pollen had been applied."

Muller suggests an explanation of this phenomenon which must be
pleasing to the minds of strict Natural Selectionists. He believes it to be
an advantage to the species to have its pollen positively deleterious rather
than simply neutral, because the flowers would then soon drop off, and
the energies of plants no longer be directed toward nourishing a part
which would not finally function.

Another quotation from Darwin (ibid., p. 113) is interesting both for
the facts contained and for the deductions of Muller.

"The same naturalist found in Brazil three plants of a Bignonia growing
near together. He fertilised twenty-nine flowerets on one of them with
their own pollen, and they did not set a single capsule. Thirty flowers were
then fertilised with pollen from a distinct plant, one of the three, and they
yielded only two capsules. Lastly, five flowers were fertilised with pollen
from a fourth plant growing at a distance, and all five produced capsules.
Fritz Muller thinks that the three plants which grew near one another
were probably seedlings from the same parent and that from being so
closely related, they acted very feebly on one another. This view is ex-
tremely probable for he has since shown in a remarkable paper (Muller
1873) that in some Brazilian species of Abutilon, which are self-sterile, and
between which he has raised some complex hybrids, that these, if near
relatives were much less fertile inter se, than when not closely related."

This work of Muller (1873) consisted in noting the fertility of
various matings of 8 species of Abutilon that he denotes by the letters
A, C, E, F, M, P, S and V, the individual plants being distinguished by
subscripts. Thus the plants EF.FX and EF.F2 are similar combinations
formed by crossing species E with species F and crossing the first gen-
eration hybrids thus formed with Fi and F2. The principal results were
as follows:

Number of flowers

Mother plant

Source of pollen

Number
of fruits

Average
No. of seeds

F.EF,

Others of same stock

F.EF1

F.EF, EF.FX and EF.F2

i-3

F.EF

FE and FE2

4-5

F.EF

EF2 and EF3

4.6

F.EF

4.6

F.EF

F.CF1 and F.CF,

4.5

F.EF

FS,

4-7

He says that the results following the intercrossing of sister plants

Genetics 2: N 1917

E. M. EAST AND J. B. PARK

were not due to bad pollen, as on other plants it was completely potent;
the pollen of F.EF2 producing fruit full of seeds on FSlf that of EF.F
on FE2, that of EF.F2 on F, and that of F.EFX on F, F.CF2, FSX
and FS2. In explaining the phenomenon he follows Darwin in suppos-
ing inbreeding to be the cause.

Most of these observations and investigations were known to Darwin
who not only published historical accounts in the "Origin of species" and
''Variation of plants and animals under domestication," but between i860
and 1880 carried out numerous experiments on the subject which were
reported in a series of papers in the Journal of the Linnean Society
and other places and were brought together in the three classics, "On the
various contrivances by which British and foreign orchids are fertilised
by insects" (1862), "The effects of cross- and self-fertilisation in the
vegetable kingdom" (1876), and the "Different forms of flowers on
plants of the same species" (1877).

Darwin's investigations on fertilization in the orchids are only re-
motely related to the subject in hand, but his experiments on heterostyled
dimorphic forms are, we believe, concerned with an analogous phenome-
non. The "illegitimate" unions according to Darwin include certain
matings other than self-pollination, but the greatly decreased fertility
after self-pollination in practically all of these species as well as the
absolute self-sterility of so many forms indicate that the condition is one
like ordinary self -sterility though complicated by a linkage with style
length and with pollen size. The work of Bateson and Gregory ( 1905)
on the inheritance of heterostylism in Primula has done something to-
ward clearing up these relationships, but much remains for the future. As
these investigations of Darwin are readily available and cannot, at
present, add materially to our discussion of self-sterility on account of
moot points, they will not be described further; but we shall abstract
from the experiments on those plants usually considered to be genuinely
self-sterile.

Darwin (1876) investigated rather thoroughly the conditions in five
self-sterile species, Eschscholtzia calif ornica, Abutilon Darwinii, Senecio
cruentus, Reseda odorata and R. lutea.

A plant of Eschscholtzia calif ornica had been accidentally found to be
self-sterile by Fritz Muller (1868, 1869) while working in southern
Brazil. This induced him to investigate its behavior through six gener-
ations, during which time he found all of the plants to be completely
self -sterile though fertile between themselves. As Darwin had found

THE BEHAVIOR OF SELF-STERILE PLANTS

515

English plants comparatively self-fertile and as Hildebrand had dis-
covered no complete self-sterility in plants grown in Germany, he ob-
tained from Muller seed of the Brazilian plants known to be self-
sterile and from them raised seedlings. These while not wholly self-
fertile, tended toward fertility, which fact Darwin attributed to the
lower English temperature. A second generation of seedlings proved to
be still more self-fertile. Conversely, seed of English stock sent to Brazil
proved to be more self-fertile than the native race, though one plant thus
exposed to the climate of Brazil for two seasons, was wholly self-sterile.

These results were paralleled by the behavior of AbiUilon Darwinii
which is self-sterile in its native Brazil, but became moderately self-
fertile late in the first flowering season in Darwin's greenhouse.6

Darwin made no extensive experiments on self-sterility with Bra-
zilian plants in collaboration, so to speak, with Fritz Muller; but this
was not for the lack of material, for in a letter to Focke (1893),
Muller says the number of self-sterile species of plants in Brazil is
very large, and that different species of the same genus often behave
differently in regard to self-pollination. He observes that self-sterility is
often associated with unusual vegetative vigor and that species of Oxalis
having trimorphic flowers which are all self-sterile make unusually
vigorous growths. This condition observed by Muller is doubtless
merely another example of the hybrid vigor or heterosis so common
among both plants and animals, and shows the reason, of course, why
self-sterility has been maintained by natural selection.

Darwin's experiments on Senecio cruentus are noteworthy only be-
cause the varieties used were descendants of garden hybrids.

Two plants of a purple-flowered and one plant of a red-flowered
variety were found to be self-sterile and cross-fertile.

The experiments with Reseda odorata were more detailed. Those of
1868 are shown in tabular form, the letters representing individuals and
the subscripts pollinations. As may be seen, the seven plants used were
absolutely self-sterile. The number of pollinations made allow us no
doubts about the matter, F and G being selfed many times as well as the
others, though in these two cases no figures were reported. Sixteen
cross-matings, on the other hand, were all fertile.

In the spring of 1869, four other plants were raised from fresh seed
and isolated under nets. Three of these proved to be wholly self-fertile,
while the fourth was not completely self-sterile.

6 Cf. our results on flowers late in the season.

Genetics 2: X 1917

Si6

E. M. EAST AND J. B. PARK

Darwin's experiments on Reseda odorata in 1868.
Male parents

s16

^19

stt

Much surprised at these divergent results Darwin raised six more
plants in 1870. Of these, two were almost self-sterile and four were
completely self-fertile. The former produced altogether five seeds, which
were grown the following year. These plants made a luxuriant growth,
but were almost completely self-sterile like their parents [an indication of
pseudo-fertility]. The progeny of the self-fertile plants was not
followed.

These varying results were attributed by Darwin to a difference in
inherited sexual constitution. He says in his general conclusions (1876,
P- 346) :

"Finally, the most interesting point in regard to self-sterile plants is the
evidence which they afford of the advantage, or rather of the necessity, of
some degree or kind of differentiation in the sexual elements, in order that
they should unite and give birth to a new being. It was ascertained that
the five plants of Reseda odorata which were selected by chance, could be
perfectly fertilised by pollen taken from any one of them, but not by their
own pollen ; and a few additional trials were made with some other in-
dividuals, which I have not thought worth recording. So again, Hildebrand
and Fritz Muller frequently speak of self-sterile plants being fertile with
the pollen of any other individual ; and if there had been any exceptions to
the rule, these could hardly have escaped their observation and my own.
We may therefore confidently assert that a self-sterile plant can be fertilised
by the pollen of any one out of a thousand or ten thousand individuals of the
same species, but not by its own. Now it is obviously impossible that the

THE BEHAVIOR OF SELF-STERILE PLANTS

517

sexual organs and elements of every individual can have been specialised
with respect to every other individual. But there is no difficulty in be-
lieving that the sexual elements of each differ slightly in the same diversified
manner as do their external characters ; and it has often been remarked
that no two individuals are absolutely alike. Therefore we can hardly avoid
the conclusion, that differences of an analogous and indefinite nature in
the reproductive system are sufficient to excite the mutual action of the
sexual elements and that unless there be such differentiation fertility fails."

These inductions are cleverly drawn and clearly expressed, but they
are not all justified by the data in Darwin's possession. The matings
between self-sterile plants made by Hildebrand, Muller and Darwin
were neither individually nor collectively sufficient to establish the point
that "a self-sterile plant can be fertilized by the pollen of any one out of
a thousand or ten thousand individuals of the same species," and it is
upon this supposition that the generalization is based. Further, Munro,
whose work was known to Darwin, had found cross-sterility in
Passiflora.

As it is not proposed to make this review a check list of species which
are, as a whole or in part, self-sterile, but rather to set forth the known
facts concerning the behavior of self-sterile plants and to outline the
various theories that have been suggested to interpret the phenomenon,
we shall pass Darwin's conclusions without further comment. His work
properly stands as the outpost of advance in the subject until the re-
discovery of Mendel's Law in 1900. The method of analysis of pedigree
cultures foreshadowed by Vilmorin but really initiated by Mendel has
made a methodological revolution. It seems fitting, however, to close this
part of our paper with the work of a botanist who, though making no
outstanding contributions to the subject, was a contemporary of and an
aid to Darwin, and who from the chronological standpoint links the
work of Darwin to that of the present day.

Hildebrand worked and wrote indefatigably upon questions of fecun-
dation in plants from 1863 until 1908. His first paper (1863), on di-
morphism in Primula sinensis appeared almost simultaneously with that
of Darwin, and since that time in the neighborhood of seventy contribu-
tions on similar subjects have appeared under his name.

Hildebrand (1866) published some rather extensive experiments
with Corydalis cava in which he showed that the plants were absolutely
self-sterile although both pollen and ovules were functional. But his
investigations were noteworthy with respect to the large number of spe-
cies in which he established a high probability of self-sterility, rather

Genetics 2: N 1917

5i8

E. M. EAST AND J. B. PARK

than for any fundamental researches on the genetic problem concerned.
We will mention only one other paper, therefore, merely to show the
large numbers of self-sterile plants that are sometimes (possibly often)
to be found in a single family when said family is even partially
investigated.

In 1896 he published on the Cruciferse and found Hesperis tristis,
Lobularia maritima (=Alyssum maritimumham.) , Cardamine pratensis,
Rapistrum rugosum, Iberis pinnata and Sobolewskia clavata fully self-
sterile, Aethionema grandiflorum and Hugueninia tanacetifolia (=Nas-
turtium tanaceti folium Hook.) nearly self-sterile, and only Draba verna
and Brassica rapa fully self-fertile.

RECENT WORK ON SELF-STERILITY

The work of the last decade on self-sterility has been less concerned
with the discovery of new cases than with an interpretation of the phe-
nomenon in keeping with modern biological thought. Several note-
worthy investigations on both plants and animals have appeared.

Jost (1907) repeated Hildebrand's experiments on Corydalis cava,
and unlike the latter, observed a small percentage of self-fertility. In
his experiments 93 selfed plants yielded 6 capsules, whereas 42 crossed
plants produced 30 capsules. Self-sterility was also noted in Secale
cereale (a variety montanum) and Lilium bidbiferum. The immediate
cause of the different behavior of these plants after self-pollination and
after cross-pollination was found to be the difference in rate of pollen-
tube growth. In Secale, pollen tubes were found to have penetrated the
micropyle in about eight hours after cross-pollination, although after
self-pollination the tubes had merely reached the base of the pistil after
twenty-four hours. Pollen tubes also appeared to grow somewhat faster
than after self-pollination when crosses (?) were made between flowers
on the same plant, but in view of the fact that asexually propagated
plants from a single seed appear to behave very similarly this observa-
tion may not be correct. In this connection it should be mentioned that
Focke (1890 and 1893) found that Lilium bidbiferum plants of the
same clonal variety were completely cross-sterile, although sister seedlings
were cross-fertile. Similar observations on asexually propagated pome
fruits have been made by Waite (1895) and Lewis and Vincent
(1909), but in these cases "fruitfulness" rather than "fertility" was
noted.

To explain his results Jost had recourse to the old concept of 'Tndi-
vidualstoffe." He believes that individuals not only of the same species

THE BEHAVIOR OF SELF-STERILE PLANTS

519

but of the same family differ qualitatively in their chemical composition,
that the gametes of any plant possess the "Individualstoff" of that plant,
and that pollen tubes grow well only in tissues having a different
'Individualstoff/'

In 19 12 a very important paper by Correns appeared in which a
Mendelian interpretation of results was proposed. His experimental
work began with a hybrid between Petunia nyctaginiflora and Petunia
violacea that had been produced in 1901, and of which 11 individuals
had passed through the winter. Six of these plants were found to be
self-fertile, three completely self-sterile and two nearly self-sterile.
Among the self-sterile plants certain combinations proved easy to make,
while others' were impossible. It was sometimes impossible even to cross
the self-sterile with the self-fertile plants [probably pseudo-fertile]. For
several reasons, however, Correns found Petunia unsatisfactory and the
work was dropped until 19 10; it was then recommenced with Cardamine
pratensis, a Crucifer that had been shown to be wholly self-sterile by

HlLDEBRAND (1896).

Concerning the "cause" of self-sterility, borrowing the term from the
author, he gives the following facts : The pollen grains germinated on
the stigma of the self-pollinated flowers, but produced only short tubes
that did not penetrate the tissues of the stigmas, while after cross-
pollination the pollen tubes were found in the upper part of the ovaries
after only 48 hours.

The pedigree culture investigations began with two plants, B having
very light lilac flowers, and G having flowers of a more intense lilac.
These plants were crossed reciprocally, the combination B2 X being
designated No. 1 and the other No. 2. From each of these matings, 30
plants were raised, and formed the basis of the remaining experiments.
They were numbered ia, ib, ic, 2a, 2b, 2c, etc.

These plants were first tested for their fertility when used as females
by crossing each individual with the pollen of two unrelated plants from
Lake Zurich and Schwabia respectively. These pollinations were suc-
cessful without an exception, proving that pollen from a single plant
could fertilize each of the 60 Fx sibs.

From 3 to 15 pollinations were then made upon every ¥t plant with
the pollen of each parent B and G. About half of these pollinations
were uniformly fruitful or uniformly unfruitful as the case might be,
but the other half showed variations in behavior that made classification
of the results difficult. For example out of ten pollinations of plant 10

Genetics 2: N 1917

520

E. M. EAST AND J. B. PARK

with the pollen of B, 6 were successful and 4 unsuccessful. This plant
was classed as fertile with B. Again, plant ik pollinated 7 times with
the pollen of G yielded 3 good capsules, 2 poor capsules and 2 failures
Correns classes this plant as sterile with G with a question mark. These
results seem at first sight to indicate a definitely graduated fertility in
Cardamine. This is not impossible; but, arguing from our own experi-
ence (Nicotiana alata), it appears to be more probable that the plant is
in a rather unstable condition physiologically and can be influenced easily
by external conditions.

Correns did endeavor to test the question of the influence of age of
plant on fertility by (1) making 17 duplicate pollinations the next year
with pollen from a plant raised from a cutting of B, and by (2) making
18 reciprocal pollinations from the Fx plants upon B and G. The pollina-
tions with pollen from the cutting of B made in 19 12 checked with those
made in 191 1 with pollen from the original plant B in a remarkable
manner. Of the reciprocals, 7 were successful both ways, 5 failed both
ways, 4 were rather indefinite but similar, while only one showed a con-
flicting result (2 failures one way and 3 successes the other).

In spite of these facts, however, it is apparent from Correns's account
that the plants were at all times kept in as fine condition as possible so
that the behavior under a poor environment or during different phases
of the flowering period was really not determined. What these experi-
ments did do was to prove beyond a reasonable doubt the physiological
similarity of cuttings with respect to cross-fertility and cross-sterility,
and to indicate that reciprocal crosses always behave in the same manner.
Unfortunately for the latter thesis, however, there are a few conflicting
results in his table 8, though this he does not mention. Of the 53 recipro-
cals recorded there, 31 give the same results, 17 give different results,
while 5 are questionable.

Correns concluded that the behavior of the F1 individuals with the
pollen of the parents was such as to indicate equal-sized classes of defi-
nitely fertile or definitely infertile plants, the behavior of the reciprocals
being the same. His classification gave the following groups :- — fertile
with B, 32 ; sterile with B, 28; fertile with G, 30; and sterile with G, 30.

He further concluded that the action of an F1 individual toward one
parent was wholly independent of its action toward the other, and that
the population could be divided into 4 classes with reference to the
behavior of the individuals toward both parents, as follows :

THE BEHAVIOR OF SELF-STERILE PLANTS

521

Fertile with both B and G, type bg,

16 plants

14 plants

Fertile with B, sterile with G, type bG,
Fertile with G, sterile with B, type Bg,
Sterile with both B and G, type BG,

An explanation of these facts was sought by assuming that each parent
B and G carried at least one transmissible factor, B and G respectively,
which actively inhibited pollen-tube growth, besides at least one inactive
factor, b and g respectively. The formulae for these plants would then
be Bb and Gg, and when they are crossed four equal-sized classes of
zygotes will be formed BG, Bg, bG and bg, because B and b, and G and g
segregate at reduction. These four Fx classes should behave when back-
crossed with each parent in the manner shown above.

There seems to be no reason in his hypothesis why plants of the type
bg should not be self-fertile though this is not the case. In fact all of the
60 Fx plants are assumed to be self-sterile although two cases showing
some self-fertility (probably pseudo-fertility) are shown in table 8c.
But this discrepancy is probably due to an imperfect description of the
hypothesis by the author, as the relation between self-fertile and self-
sterile plants is evidently meant to be left out of consideration.

The intra-class and inter-class pollinations between the Ft plants of
which he made about 700 (tables 8a-8d), hardly come up to expectations,
but there is a regularity that cannot be overlooked.

Compton (1913 a) confirmed Darwin's report that both self-fertile
and self-sterile plants occur in the mignonette, Reseda odorata. From
experiments on crossing these two races he obtained the following facts :

( 1 ) Self-sterile plants when bred inter se threw self-sterile offspring
only. This was thought to indicate that self-sterility is a Mendelian re-
cessive. (2) Certain self-fertile plants, when self-fertilized gave self-
fertile offspring only. When crossed with self-sterile plants the same
result was obtained. These plants Compton regarded as homozygous
dominants. (3) Other self-fertile plants, when self-fertilized, gave ap-
proximately 3 self-fertile to 1 self-sterile offspring. The same plants
crossed with self-sterile individuals produced about one-half self-fertile
and one-half self-sterile progeny. These he regarded as heterozygous.
All of these facts are satisfactorily interpreted by the hypothesis that
self-fertility is a simple dominant to self-sterility.

In a later paper Compton (1912) suggests, as Jost had previously
done, the presence in the pistil of diffusible substances which stimulate or
retard pollen-tube growth after cross- or self-pollination respectively.

Genetics 2: N 1917

522

E. M. EAST AND J. B. PARK

The growth of pollen tubes in the style and the growth of fungus hyphae
in a host appealed to Compton as analogous, and he suggests that self-
sterility may be due to agents similar to those which govern immunity or
susceptibility in animal or plant.

These results confirm a Mendelian hypothesis already suggested by
Baur (1911) without reporting detailed results. He crossed the self-
sterile Antirrhinum molle with the self-fertile A. ma jus and obtained only
self-fertile offspring. The F2 generation consisted of both self-fertile
and self-sterile plants, the former being in the majority. Baur gave
these hybrids to Lotsy (191 3) who raised a large F2 generation with
similar results although he was inclined to believe that the plants showed
variable degrees of self-fertility and self-sterility. Neither Compton,
Baur nor Lotsy touched the question of the behavior of self-sterile
plants among themselves.

Since self -sterility was discovered in the Ascidian Ciona intestinalis
by Castle ( 1896) , its reproductive behavior has been studied by Morgan
(1905, 1910), Morgan and Adkins (Morgan 1913), and Fuchs
( 1914 a) . Morgan and Adkins showed that these animals vary in degree
of self-sterility. Perfectly self-sterile individuals were the exception, but
self-fertility never equaled cross-fertility. Individuals also varied in the
ease with which their eggs might be fertilized by the sperm of other
individuals. The following matings were made with the results noted in
percentage of eggs fertilized :

THE BEHAVIOR OF SELF-STERILE PLANTS

523

Fuchs (1914 a), however, has criticized Morgan's work, maintaining
that 100 percent of segmenting eggs can be obtained in every cross with
normal ova if sufficiently concentrated sperm suspension be used. He
showed, among other things; that (1) an increased concentration of
sperm suspension caused an increase in the number of eggs self-fertilized,
(2) a greater concentration of sperm was usually necessary to bring
about any self-fertilization than would cross-fertilize 100 percent of
foreign eggs, and (3) contact with suspension of own sperm decreased
the ease of later cross-fertilization.

The work of Fuchs suggests a physico-chemical basis for self-sterility,
since contact of eggs with their own sperm appears to cause changes in
the egg membranes which inhibit entrance of own sperm and to some
extent of foreign sperm, yet his criticism of Morgan's statements is not
to the point for by the submission of the eggs to different sperm concen-
trations he has increased the number of variants under investigation.

Morgan (1913, p. 217) explained his facts by means of this
hypothesis :

"This failure to self-fertilize, which is the main problem, would seem
to be due to the similarity in the hereditary factors carried by the eggs and
sperm; but in the sperm, at least, reduction division has taken place prior
to fertilization, and therefore unless each animal was homozygous (which
from the nature of the case cannot be assumed possible) the failure to
fertilize cannot be due to homozygosity. But both sperm and eggs have
developed under the influence of the total or duplex number of hereditary
factors; hence they are alike, i.e., their protoplasmic substance has been
under the same influences. In this sense, the case is like that of stock that
has long been inbred, and has come to have nearly the same hereditary
complex. If this similarity decreases the chances of combination between
sperm and eggs, we can interpret the results."

This interpretation of self-sterility endeavors to give a modern render-
ing of Darwin's idea that the condition is analogous to the decreased
fertility often resulting from other modes of inbreeding. From his other
numerous observations on cross- and self-fertilization, Darwin felt in-
stinctively that such an analogy should exist, even though self -sterile
plants were continually cross-pollinated and must of necessity have a
mixed ancestry. Morgan's contribution was to show in a general way
how such a similarity might come about. His suggestion is unquestion-
ably stimulating and we have been glad to acknowledge our indebtedness
to it (East 1915).

One should not ascribe more breadth to the hypothesis than the author
really intended, however; for certain coordinate problems that may or

Genetics 2: X 1917

524

E. M. EAST AND J. B. PARK

may not have the same underlying cause, were not included in its scope.
For example, it assumes nothing regarding the origin of self-sterility or
the difference between self-sterility and self-fertility. At first sight one
feels that there is a great weakness in its failure to account for self-
fertility, since the eggs and sperms of self-fertile races also develop
under the influence of the total or duplex number of hereditary factors,
and it is difficult to see why this should decrease the attraction between
eggs and sperm in some cases and not in others. But the difference be-
tween self-fertile and self-sterile organisms is not of necessity the same
problem as the behavior of self-sterile organisms.7 This distinction is
manifest if one refers to Compton's work. In his material the difference
between self-fertility and self-sterility is that of a single Mendelian
factor, — self-sterility being recessive. But Compton does not attempt
to account for the behavior of his self-sterile plants.

Darwin, on the other hand, made no serious attempt to interpret the
behavior of self-sterile plants, or to describe the fundamental difference
between self-fertile and self-sterile races. He was concerned chiefly with
the origin of self-sterility. The basic reason for the evolution of self-
sterility, he thought, lay in a necessity for cross-fertilization. In this
we believe he was unwise. The benefits of cross-fertilization, no one
doubts. With the vigor of heterozygosis as the immediate advantage for
natural selection to grasp, with the immense ultimate advantage of
multiplicity of forms brought about by Mendelian recombination, one can
see reason in all the host of devices for producing cross-fertilization in
animals and plants, — including even bisexuality itself. But this does
not mean that cross-fertilization is an inevitable need, as Darwin be-
lieved was so clearly demonstrated by his observations on the deleterious
effects of inbreeding. It is rather merely an asset in the struggle for
existence, as recent experiments have shown.8 Consequently emphasis
should be placed on the assured benefits of cross-breeding and not on the
doubtful evils of inbreeding. One can understand therefore why self-
sterility might be desirable, and why it should be retained by natural
selection after coming into existence, but the cause of its origin must still
be denoted by that useful word chance, the veil of ignorance.

In view of these facts — and all of the important facts regarding self-
sterility have been cited — the fundamental questions involved are almost
as obscure now as they were when Darwin left them. But the work of

7 Stout (1916) continually confuses these two problems.

8 See East and Hayes (1912) and the papers there cited.

THE BEHAVIOR OF SELF-STERILE PLANTS

525

Morgan, Correns and Compton encourages the hope that their solution,
if one may use that term for scientific description, will be accomplished.
An interpretation in harmony with modern biological conceptions which
will in its turn be helpful, ought at least to be possible when all of the
facts are at hand.

Since the historical part of this paper was written, Stout (1916) has
published a bulky memoir on self-sterility in Cichorhim intybus. A large
portion of this paper is devoted to destructive criticsm. Darwin and his
contemporaries, Baur, Compton, Correns, East, Jost, Lotsy, Morgan
and Shull are "placed upon the carpet" and dealt with severely. One
wonders whether all of these writers can be wholly wrong in the views
that have been assailed, and if not, just wherein the differences of opinion
lie. We cannot help but feel that they are due largely to his miscon-
ceptions of the views of the various writers concerned.

As examples of what is meant by this statement, let us mention two of
the points on which Stout lays great stress. He feels strongly that self-
sterility is a markedly variable character, and that this has not been
recognized by previous writers. But since the existence of variability in
the somatic expression of self-sterility has been admitted unanimously
by the writers with whom we are acquainted, the true point at issue is not
this, but rather the question whether any considerable part of the varia-
tion in this character is the result of genetic differences. This question
has been investigated in Nicotiana, and there the variation seems to be
almost wholly due to environmental changes, as is shown later in this
paper. Considered with this point in mind, a reasonable and constructive
interpretation of our own and many other self-sterility data can be given.
Where before there was chaos a certain order appears. Stout's failure
to recognize these truths is probably the reason why he has been unable
to make any constructive analysis of his own numerous data for the fact
that some of his families arising from selfed seed behaved exactly as the
families arising from crossed seed shows that he is often (at least)
dealing with a pseudo self-fertility (see p. 531).

Xow this argument of Stout's, we gather, is meant to be only a par-
ticular instance advanced in favor of his general view that characters
are (always?) too variable genetically to be represented properly by
fixed Mendelian factors. The justice or injustice of such a contention
cannot be discussed here, but we should like to point out that in assum-
ing— as is so often done — that geneticists commonly believe in an ele-

Genetics 2: N 1917

526

E. M. EAST AND J. B. PARK

mental stability of characters, the attitude of the great majority of such
workers is misconstrued. If we have interpreted Mendelian investigators'
views correctly, they believe that characters are variable, but in different
degrees in different species ; and that there is adequate evidence to show
that most characters in most species are so constant throughout the num-
ber of successive generations ordinarily available for experimental pur-
poses when viewed under the conditions most likely to eliminate variables
other than heredity, that the abstract idea of fixed germinal factors can
be used properly and helpfully in genetic analysis.

As a second case where we believe Stout has not represented fairly the
views of the writers criticized, the section of his paper entitled "Relation
of vegetative vigor and fertility to inbreeding and cross-breeding" may
be cited. Stout criticizes in particular the views of Darwin, Shull,
and East and Hayes on this subject. He rests his case on a paper by
Burck (1908) in which the writer holds, that (quoting Stout) :

"(1) plants that are regularly self-fertilized show no benefits from cross-
ing, (2) that nowhere in wild species is there evidence of an injurious effect
from self-fertilization, and that there is abundant evidence of continued
vigor and high fertility resulting from long-continued self-fertilization, and
(3) that the advantage derived from crossing within or between garden
varieties appears when there is doubtful purity ; and is due to the fact that
both vigor and fertility have already been decreased by hybridization, and
that when crosses do give increased vigor and fertility the cross has re-
stored in increased measure the original nuclear organization of the parent
species.,,

The logic of the third statement is too delightful for comment, being
worthy indeed of Mother Eddy. Vigor is decreased by hybridization.
Vigor is increased by hybridization. It is increased by restoring "nuclear
organization." Not only is nuclear organization restored, but it is
restored in "increased measure."

The second statement has never been denied by modern writers, to our
knowledge. It was emphasized by East and Hayes (1912), who
pointed out why the advantage of cross-fertilization in plants should
be stressed rather than the disadvantage of self-fertilization. This ad-
vantage, if one may recall it, lies in the fact that n inherited variations
can produce but n forms under self-fertilization, and may produce 2n
forms under cross-fertilization by Mendelian recombination.

The first statement is simply not in accord with the facts. We are
astonished that one who has the acquaintance with the literature that

THE BEHAVIOR OF SELF-STERILE PLANTS

527

Stout has shown, should quote it with approval. Every hybridist of
experience from Kolreuter (1760) to the present day has cited so
many data diametrically opposed to it that the matter is no more worthy
of discussion than is a denial that the earth is round.

Of course as to the interpretation of the facts one may hold a differ-
ence of opinion. The hypothesis of heterosis advanced independently
by Shull and East has, we think, served a useful purpose. The last
word has not been said, however, and data accumulated by H. K. Hayes
and D. F. Jones in their continuation of the experiments reported by
East and Hayes (1912) have led the senior author to modify his views
on several of the points there discussed, though not on the main conclu-
sions. But in the meantime it is disconcerting to have our published
statements misunderstood and misinterpreted. For example Stout
says (p. 419) "East and Hayes believe that heterozygosity gives an
increase of both vigor and fertility in proportion to the number of
heterozygous factors in the organism." There are two errors in this
statement. Neither Shull nor East has maintained that crossing in-
creases fertility. The number of flowxrs and fruit is often increased,
but no data have appeared which indicate a decreased percentage of non-
functional gametes. Second, East and Hayes used the words "roughly
proportional to the number of heterozygous factors." Leaving out the
word "roughly" and taking the statement from its context, conveys a
very wrong impression for it was not assumed that every germinal factor
affected vigor and it was expressly stated that one could not assume equal
effects for different factors. Again Stout achieves a remarkable misin-
terpretation of the results reported in table 5 of this same paper. Here
42 inter-specific crosses are reported, of which 14 show decreased vigor
(this figure should be 13 instead of 14 owing to a typographical error
in reporting the first cross, N. alata X Forgetiana, which was 125 percent
of the parental average in height, instead of 25 percent). Stout leads
his readers to infer that this table is the sole basis of the conclusions
regarding heterozygosis, and that the conclusions are incorrect because,
as he states : "There was increased vigor in only 17 cases, but there is no
apparent reason why, if it is simply heterozygosity that increases vigor,
more of the combinations should not show increased vigor."

Now what are the facts. The statements on the previous page (p. 27)
of the paper make it clear that many varietal crosses were made (over
100 in Nicotiana alone to that date), which showed vigor equal to, or
greater than the parental average. While not expressly stated, it may

Genetics 2: N 1917

5^8

E. M. EAST AND J. B. PARK

be inferred that none was found with decreased vigor. If it had been
otherwise it would have been stated. Multiplication of such data was
thought unnecessary in view of the exceedingly numerous results of
Kolreuter, Knight, Gartner, Naudin, Focke, Darwin and others,
on the increased vigor of such hybrids. This table then, as is shown on
pages 29 and 30, was submitted for the particular purpose of trying to
establish a wholly different thesis, viz., that as germ plasms become more
and more unlike, there comes a time when hybrids show ( 1 ) an inability
to form germ cells (sterility), and (2) difficulty in somatic cell division.
Our typographical error was unfortunate, but in view of the text given
the statement made by Stout is an inexcusable perversion of our work.

We have mentioned but two out of a goodly number of misconstruc-
tions of work with which we have been concerned. We have done this
because we believe that they are paralleled in the author's criticism of
most of the writers mentioned above, and because we realize that if we
undertook to point out these misunderstandings in the case of other
writers, the answer would be that it was merely a difference of opinion.

On the other hand, Stout has given us a classification of types of
sterility, and has reported a really immense amount of data. We hope
that he will give a more constructive analysis of them later.

THE MATERIAL USED AND THE GENERAL PLAN OF THE PRESENT

INVESTIGATIONS

The investigations described in this paper may be said to have been
begun in 19 10, when, in connection with some genetic studies on size in
the genus Nicotiana, the two species Nicotiana Forgetiana (Hort.) Sand,
and Nicotiana alata Lk. and Otto var. grand i flora9 Comes were found to
be self-sterile. These two species have been made the basis of our experi-
ments, though later some work was done upon Nicotiana angusti folia R.
and P. var. crispa? Cav., N. commutata Fisch. & Meyer, and N. glutinosa
L., in which self-sterility had been discovered.

The characters of these species and of Nicotiana Langsdorffii L., a
self-fertile species used, are described in Comes (1899), Setchell
(1912), and East (1913, 1916).

From the technical standpoint the material has been ideal. Any com-
bination of the three species N. Forgetiana, N. alata, and N. Langsdorffii
can be made, the Fx hybrids being completely fertile (in proper cross-
fertile combinations). N. glutinosa and N. angustifolia, however, can

9 Hereafter N. alata grandiflora will be known as A7, alata and N. angustifolia crispa
as N. angustifolia.

THE BEHAVIOR OF SELF-STERILE PLANTS

529

neither be crossed together nor with the other species. The plants of
each race grow rapidly and vigorously, and are not easily affected ad-
versely by sudden changes in environmental conditions. They are not
subject to serious parasites. Cuttings root well, and with care old roots
will live through a second and occasionally even a third season. Emas-
culation and pollination are easy to perform, and seed production in
fertile crosses is high.

N. Forgetiana, N. alata and N. angustifolia belong to the subgenus
Petunioides, a fact worthy of note because nearly all of the species of
this section have both showy flowers and abundant nectar which attract
insects and thus promote cross-pollination. Even N. glutinosa has
rather conspicuous blossoms, though belonging to the subgenus Rustica
in which most of the species have small and unattractive flowers that are
self-pollinated naturally. In other words all four of these species prob-
ably had evolved structural modifications which aided cross-fertilization
long before the development of their self-sterility. We are dealing, there-
fore, with plants desirable both from the viewpoint of the experimentalist
and of the student of evolution, a most unusual combination.

The general problem presented by this material obviously was to dis-
cover the facts regarding self-sterility, and to determine whether these
facts might be fitted by a simple mathematical or chemico-mathematical
description. It has been attacked along three distinct lines : ( 1 ) pedigree
cultures; (2) histological studies of pollen tubes in crossed and in selfed
pistils, and in inter-specific and inter-generic crosses; and (3) physi-
ological studies of pollen tubes cultivated on artificial media.

Work along this general plan has been carried on at the Bussey Institu-
tion of Harvard University continuously since 19 10, though it has not
been our sole interest. It was our good fortune to have the very efficient
aid of Dr. O. E. W^hite, then a graduate student and assistant at Har-
vard University, during the winter of 1911-12. The junior author's
connection with the work began in February 1914, and has continued
until the present time. In addition, Miss Grace Sheerin and Miss
Bertha Kaplan have assisted in the pollination work for limited periods
of time.

It being impracticable to present and to examine these various data
within the limits of a single article, we propose to take up only a portion
of the pedigree culture work in this paper, leaving the remaining ques-
tions to be treated later. The pedigree culture investigations have thus
far involved four studies: (a) the effect of environment on self-sterility;

Genetics 2: X 1917

530

E. M. EAST AND J. B. PARK

(b) the relations existing between self-sterile plants in intra-specific and
inter-specific crosses; (c) the relations between self-sterile and self-fertile
plants; (d) selective fertilization. The first two studies will be discussed
here.

The usual precautions used by plant geneticists have been carefully
observed, including castration of all flowers on self-sterile plants used as
pistillate parents. This safeguard would not be worthy of especial
mention except for the fact that it is wholly disregarded in Stout's re-
cent paper (1916). We shall show in a later paper that effective pollen
mixed with "own" pollen causes scarcely any acceleration of "own"
pollen tubes in Xicotiana. But we cannot find that Stout determined
this for chicory, and to take for granted that there is no such effect seems
to us a laxity in a scientific work.

Every important fact described has been confirmed independently by
each of us, and certain of the data that have been remarkably orderly (for
example, table 11) have been collected by several persons in such a man-
ner that personal equations were largely eliminated.

It may be noted here that a preliminary report of some of the work
which we now report in detail was published in 1915 (see East 191 5 ) .
With more data in hand more definite ideas on the subject have been
possible, hence several differences will be noted between the statements
made then and now. It is scarcely necessary, however, to point out
every difference in the interpretations, as we shall endeavor to give in
full our reasons for the present conclusions.

THE EFFECT OF THE ENVIRONMENT ON SELF-STERILITY

In beginning the description of our experiments with a section on the
effect of environmental changes on self-sterility a chronological inversion
is made which needs explanation, particularly as carefully planned ex-
periments designed to show the effect of individual environmental factors
when all others are controlled have not been carried out. Work on the
relation between self -sterile plants was started with the idea, that even
though Darwin were correct in supposing that self-sterility is seriously
affected by changes in the environment, conditions might be kept so con-
stant that no difficulties would be encountered. Indeed, this is probably
the case, since no particular difficulties were experienced during several
years in spite of certain environmental factors being constantly varied.
There came a time, however, when troubles arose which were puzzling
for a considerable period. Our inquiries regarding the effect of the

THE BEHAVIOR OF SELF-STERILE PLANTS

531

environment on self-sterility have finally removed the stumbling-block
and have made a clear and reasonable analysis of the pedigree culture
work possible.

In brief these conclusions are as follows :

1. Self-sterility is a condition determined by the inheritance received,
but can develop to its full perfection only under a favorable environment.
This is not a strange conclusion, for perhaps particular environmental
combinations are necessary for the full development of all positive
somatic characters. But certain characters are much more seriously
affected than others by the environmental variations likely to be met under
ordinary conditions. For example, Baur (1911) showed that Primula
sinensis rubra produces red flowers when grown at a temperature of 20°
C. and white flowers at a temperature of 300 C. ; East and Hayes
(1911) found that the red pericarp characteristic of a certain maize
variety developed in sunlight but not in shade; Miss Hoge (Morgan
et al., 191 5) discovered that in a Drosophila mutant with supernumerary
legs the character was only called out when the animals were kept at
io° C. Self-sterility is such a character. It develops fully only under
conditions which promote a normal healthy vegetative growth, and dur-
ing the active part of a flowering period.

2. At the end of a flowering period and under conditions adverse to
vegetative growth, self-sterility declines until a few seeds may sometimes
be obtained after self-pollination. Occasionally even a full capsule is
produced. The immediate cause of this partial return to a pseudo-
fertility is the acceleration of pollen-tube growth that obtains under these
conditions. Since we have reason to believe that the difference between a
fertile and a sterile combination in these plants is the ability of the pollen
grain through something inherent in its constitution to call forth in the
tissue of the style in the former and not in the latter case a secretion
which accelerates pollen-tube growth, it follows that in weakened style
tissue some change has occurred that renders this secretion more easily
produced.

3 Self-sterility can be restored in weakened plants by allowing them
to go through a period of rest and then, by proper treatment, bringing
them into flower anew as vigorous plants. Truly self-fertile plants can-
not be forced into self-sterility by any treatment. This last conclusion
is of course largely a conclusion by analogy and is not subject to rigorous
proof.

4. Self-sterile races differ in their norms for self-sterility. Thus in

Genetics 2: N 1917

532

E. M. EAST AND J. B. PARK

N. Forgetiana and in N. angustifolia the character is much more stable
than in N. alata and N. glutinosa. In many ways this behavior indicates
the existence of multiple allelomorphs for self-sterility.

The basis for these conclusions is the whole of our experience with
self-sterile plants, which, it is scarcely necessary to say, cannot be cited
statistically in this place. But the following facts will show, we hope,
that they are well founded.

Cross No. i between TV. Forgetiana and N. alata was made in 1909
using N. Forgetiana as the female. At that time both of the parents
were thought to be self-fertile because a carefully bagged inflorescence of
each species had yielded seed ; but when the plants of the F± generation
turned out to be self-sterile, the status of the parents was investigated
more carefully. Over two hundred plants of N. Forgetiana have been
tested under various conditions. Plants growing out of doors both on
good soil and on poor soil have been tested throughout the growing
season. Greenhouse-grown plants have been tested not only throughout
a normal flowering period (about 3 months), but have been forced
through an abnormally long flowering period during the test. Plants
well nourished have been compared with plants poorly nourished, and
plants well watered with plants under conditions of drouth. Both old
roots and cuttings brought into a second flowering period in fine condi-
tion have been compared with much pruned old roots and cuttings in
poor condition.

Only 3 cases of seed production have been observed. 2 plants at the
end of their flowering period under conditions adverse to vegetative
growth produced 1 and 2 capsules respectively having about 50 seed each
(the normal is ca. 300) out of 14 tests. The third plant was not tested
until near the end of its flowering period. At that time it was noted that
it seemed to be self-fertile. Under test it did indeed produce several
fine seed capsules after self-pollination and would undoubtedly be called
a self-fertile plant were there not the following reasons for considering it
an unstable self-sterile (see description of N. alata).

1. The plant when first tested was in a late flowering stage, yet pro-
duced capsules only in about half the tests.

2. After pruning and resting for a time the plant was brought into
vigorous flower a second time. The tests during the first two weeks of
this period (about 20 flowers) were all negative. The plant seemed
to be perfectly self-sterile. Gradually, however, self-fertility returned
as the flowering period waned.

THE BEHAVIOR OF SELF-STERILE PLANTS

533

3. Twenty-four plants grown from selfed seed of this individual,
tested during the height of their flowering period, all proved self-sterile.

We are therefore forced to concede the probability that an error of
manipulation or of record was made in 1909, although we may have
happened upon a plant like the one just described since the original
selling was done at the end of the flowering season. Be that as it may,
the conclusion is inevitable that N. Forgetiana> (and N. angastifolia has

Table i

Progeny of pseudo self-fertile N. alata plant used in cross No. 2. Grand-progeny of
original pseudo self -fertile plant. Subscripts show number of pollinations made.

Ped. No.
$

No.
selfings
sterile

No. selfings giving capsules

Plants with
which cross-
fertile
$

Plants with
which cross-
sterile
$

1-10 seeds

10-50 seeds

250-300 seeds

532, 572, 582

53, 57, 58, 59

53, 59

314,53* 54, 56;, 57

58, 79

58, 62, 71, 79

314, 58

58, 66

1 pollination 53 X 54 and 2 pollinations 59 X 53 produced 1-10 seeds each.
314 = N. Forgetiana.

yielded similar results) is a self-sterile species of remarkable stability,
which only occasionally (1 in 300?) produces a plant that shows some
self-fertility under adverse conditions.

N. alata, on the other hand, has proved to be more unstable10 in its
self-sterility; or better, it has proved to have a norm more nearly inter-

10 N. glutinosa appears to behave like N. alata, but has not been tested very
thoroughly.

Genetics 2: • N 1917

534

E. M. EAST AND J. B. PARK

mediate between the extremes complete self-sterility and perfect self-
fertility. But fundamentally it is a self-sterile species like N. Forgetiana.

Numerous N. alata plants have been tested for self-sterility under the
same conditions as described above for N. Forgetiana. The results have
been similar in that the plants were practically always completely self-
sterile during the early part of a vigorous flowering season. But under
adverse conditions during the latter part of the flowering period, rather
a high percentage of the plants produced capsules with from i to 50
seeds each. Only 2 plants have been found, however, that appeared to
be almost completely fertile from the middle of the flowering period
onward under normal conditions. Of these plants more is to be said.

Assuming that no mistake was made in 1909 and that selfed seed was
actually obtained from a field-grown plant of N. alata, we have records
of its progeny for three generations.

Twenty-five seedlings from this seed were grown in 19 14. These
plants were tested for self-sterility as field-grown plants, though not as
thoroughly as might be desired. 2 plants showed some self-fertility, —
no tests having been made until the latter part of August. From 1 of
them selfed seed was obtained and a second generation grown. 23 of
these plants were tested in the greenhouse with the results shown in
table 1.

Fourteen of these plants produced no seed when selfed; 9 showed
some degree of self-fertility. This fertility apparently occurred only

Table 2

Progeny of pseudo self -fertile N. alata plant No. 56.
Great-grand-progeny of original pseudo
self-fertile plant.

Ped. No.

No. selfings
sterile

No. selfings giving capsules

1-10 seeds

10-50 seeds

250-300 seeds

THE BEHAVIOR OF SELF-STERILE PLANTS

535

when the flowering season was waning and the plants were under adverse
conditions, as was stated before; but it cannot be proved that this was
always the case, for one cannot draw a definite line between vigorous
and weakened plants. 3 plants, excluding No. 56, produced some full
capsules, but in these cases the remaining self-pollinations and sterile
cross-pollinations show that the plants were not truly self-fertile. Plant
No. 56, however, showed no direct indications of self-sterility in con-
nection with the 3 self-pollinations tried. More pollinations should
have been made on this plant at the beginning of a second flowering
period. Unfortunately, it was discarded. The evidence of self-sterility,
therefore, is wholly circumstantial. It is, that though having functional
ovules No. 56 was sterile to the pollen of plants 53, 57, 58 and 59, and
though having functional pollen it was sterile crossed on plant 59.

A small population was grown from the selfed seed of this plant. It
is shown in table 2.

Although 5 of the plants produced some seed, if one considers the
date of manipulation and the state of the plants, the evidence is all in
favor of the idea that this was an effect of external conditions. There
is no reason whatever for believing that any of the plants were truly
self-fertile.

All told then, we have three generations of AT. alata plants, each gener-
ation being grown from selfed seed produced from plants apparently
weakened at the time of seed production, without the occurrence of a
single plant which behaved in every way like a truly self-fertile individual.
It seems to us, therefore, that this selfed seed might be thought of as
having been produced artificially.

If this be the correct view of the matter, it is clear that there is no
reason why fusion between gametes produced by a self-sterile plant
cannot occur provided the male generative nucleus enters the embryo
sac. Such unions may take place without affecting the self-sterility of the
progeny. Even by the selection of apparently self-fertile plants for three
generations no tendency toward the formation of a self-fertile race is
indicated. Just how broadly one may generalize from these data is still
problematical, but the two following conclusions are certainly more than
guesses.

(1) Unless a male gamete complementary to every female gamete is
formed, there is no selective fertilization, for full capsules have been
found on plants that in the early part of the season and in crosses showed
they were really self-sterile. Other evidence militating against selective

Genetics 2: N 1917

536 E. M. EAST AND J. B. PARK

Table 3

Progeny of X. alata Xo. 38 X N. ahta Xo. 56.

Ped. Xo.

sterile

Xo. sellings giving capsules

1-10 seeds

10-50 seeds

250-300 seeds

1 1

101

lUi

8
8

I06

108

109

110

T T T
111

tt6

1 1U

TT7
A A/

Il8

T TO
1 iy

120

121

122

T2 1

127

1 2b

133

135

136

137

139

140

141

144

146

fertilization which will be presented in a later paper has also been ob-
tained by a different method of attack.

(2) It follows therefore that self-sterility behaves as a sporophytic
character and is not the result of incompatibility between gametes.

One other bit of evidence regarding A", alata should be presented here.
It is the behavior of the progeny of a cross between the self-sterile plant
Xo. 58 and the apparently self-fertile plant Xo. 56. These data are
reported in table 3.

Here again we find a considerable percentage of plants, a third to be
exact, giving a few capsules having from 1 to 50 seeds each. Here again

THE BEHAVIOR OF SELF-STERILE PLANTS

537

it was the plants near the end of their flowering season, the plants that
had been cut back strongly, the plants that were producing flowers on
one or two weakened branches, that gave the seeds. To be sure, as in
other families, one or two plants apparently vigorous behaved in the
same way near the end of the flowering season. But the correlation
between weakened failing branches at the end of their flowering period
and tendency toward self-fertility was very high even when judged only
by external appearances.

The remaining data on this subject cannot be discussed in this place
without repetition, since they include nearly all our pedigree culture
work. And at any rate they are important only as corroborative evi-
dence, for in our regular experiments extremely weak and old flowering
branches were seldom used. For this reason we rarely had to contend
with any approach toward self-fertility in self-sterile plants. But the
phenomenon when met lent support to our hypothesis. Furthermore,
cross-sterile combinations behaved in the same way.

These conclusions have been a great aid to us in analyzing our pedi-
gree culture facts. Without them the data from two or three of our
populations, where pollinations were carried on up to the end of the
flowering season, would have been somewhat chaotic. They reveal, for
example, that N. alata is just as much of a self-sterile species as Ar. For-
getiana- though the expression of the character is affected more easily
by external conditions. They show clearly why selection for three years
accomplished nothing. The selected extreme was a non-inherited fluctu-
ation. It is clear also why crosses between these apparently self-fertile
plants and plants unquestionably self -sterile, yielded no truly self-fertile
offspring in either the F± or F2 generations. The plants were really self-
sterile; they were pseudo-fertile, and will be so called.

In this connection it may be recalled that Darwin (1876) found that
self-sterile plants of Abutilon Darwinii became partly self-fertile at the
end of their flozcering season.

Keeping these things in mind, one is able to classify the pedigree cul-
ture results with great accuracy, though there are five possibilities of
error.

1. There may be error of record. This we believe to be slight, owning
to our various methods of checking results.

2. A true sterility either partial or complete may exist. This usually
can be discovered by a microscopical examination of the pollen, and may
be tested by reciprocal crosses. The reciprocal cross test has never

Genetics 2: X 1917

538

E. M. EAST AXD T. B. PARK

brought to light a case of ovule sterility and pollen fertility, but the
converse is sometimes true.

3. Combinations made but once and failing must be reported as
sterile ; but this is an error about 4 times per hundred, since this is the
ratio of failure found in combinations known to be fertile, by reason of
an imperfect technique or other unknown causes. We cannot correct
accurately for this error, but it must be considered when discussing ex-
ceptions to a general scheme which other data fit.

4. Combinations may fail once and succeed once in two trials, or in
very rare cases fail twice and succeed twice in four trials. Experience
has shown that if the capsules are normal in size and full of seed, the
combination is fertile. Fertile combinations always give full capsules.
There is no partial fertility in fertile combinations except as true sterility-
exists in some degree ( see error 2 ). Conversely, it is possible of course
to meet with a pseudo self-fertile plant like A', alata plant Xo. 56. which
under adverse conditions might give full capsules of normal size after a
''sterile" combination had been made. But under the environmental
conditions that usually obtained during our work, this would be ex-
tremely rare, — to the best of our knowledge and belief not over 1 per
200 plants.

5. Combinations may give capsules with from 1 to 50 seeds as well as
failures. These are sterile combinations. They probably occurred in only
three families, because only in these families were the plants utilized
during the zcJiole of their flowering period. Unfortunately it must be
admitted that a few errors of record may have been made with these
cases. A small number of apparently successful matings were not
recorded until the capsules had opened. Since the capsules were of
normal size and each had contained a number of seeds, these combina-
tions were recorded fertile, but the matter is not certain.

It is not believed that these errors are serious even when taken to-
gether but some allowance must be made for them in considering the
few exceptions noted in the analyses we have made of the tables that
follow.

IXTER-SPECIFIC PEDIGREE CULTURE EXPERIMENTS

All of the crosses reported in this paper are between species or varie-
ties believed to be self -sterile for the reasons set forth in the foregoing
section. But because certain plants were used which under the peculiar
conditions at the time of the test for self-sterilitv vielded some selfed

THE BEHAVIOR OF SELF-STERILE PLANTS

539

seed, these plants are distinguished by the term "pseudo self-fertile."
Their behavior in these crosses is further evidence that the term is
justified.

Cross No. i. N. Forgctiana X N. alata (self -sterile X self -sterile)
The cross to be described first is that mentioned previously in connec-
tion with the- discovery of self-sterility in the genus Nicotiana. It was
made in 1909, using N. Forgctiana as the female and Ar. alata as the
male.

The Fx generation
The F1 population consisted of vigorous plants twenty-five11 percent
taller than the average of the two parents and was extremely uniform in
size and in color of flowers, though the latter were not so dark a red as
those of the male parent. A few individuals tested for fertility in 19 10
and others from the same original cross again tested in 1912, all proved
to be self -sterile. The actual tests made, some 20 plants altogether, were
too few to claim self-sterility for every individual, but careful observa-
tion of about 50 other plants in the field indicated this to be the case.
These observations were made by estimating the number of capsules
which developed naturally on each plant, it having been determined that
on self-fertile plants of an allied species, N. Langsdorffii, from 10 to 20
times as many capsules develop as on self-sterile plants of N. alata,
though the ratio of flowers formed on the two species is only about
3 to 1.

No extended experiments were carried out to test the fertility of these
plants in intercrosses. 6 intercrosses between sister plants were made
and each was successful, but whether some cross-sterility existed or not
is unknown. The pollen, however, was good in every plant examined
(about 30).

The Fo generation
From these 6 intercrosses between pairs of F1 plants almost a thou-
sand individuals were grown. They showed a most remarkable varia-
tion in all their characters, the range including the modal values of both
grandparents. The frequency distributions for length and for breadth
of corolla have been discussed in another paper (East 1913), and it
will suffice to note here that while the coefficient of variation for length
of corolla in the Fx generation was 8.28 ± .38 percent, in the F2
generation it was 22.57 ± .39 percent.

11 By a typographical error the height of this cross is made 25 percent instead of
125 percent in table V, East and Hayes 1912.

Genetics 2: N 1917

540

E. M. EAST AND J. B. PARK

There was also a great range in color of corolla, which even with the
considerable number of subsequent generations grown, has not been
analyzed to our complete satisfaction. 4 Mendelian factors appear to
describe the breeding results best, giving the 7 forms, red, magenta,
light red, light magenta, light red on exterior of corolla only, light
magenta on exterior of corolla only, and white. Red is epistatic to
magenta, and the darker colors are epistatic to the lighter ones.

These details are given in order to emphasize the fact that here we
have two races sufficiently distinct from each other to be designated as
separate species, which cross easily and give a fertile F1 generation and
a wide range of forms in the F2 generation. The fertility of the F±

Table 4

Result of mating s on 20 plants of the F2 generation of cross No. 1
N. Forgetiana X A7, alata.

Plants used as males

THE BEHAVIOR OF SELF-STERILE PLANTS

541

generation indicates absence of any selective elimination of gametes or
zygotes in its daughters, and the variation exhibited by these daughters
shows conclusively that the original parents really did differ by a
considerable number of hereditary factors. These matters are impor-
tant in connection with the inbreeding experiment that followed.

About 40 plants from the F2 generation were crossed and selfed on a
rather large scale. One of these experiments in which 20 plants coming
from 2 intercrosses between F± plants were used, is shown in table 4.
The vertical columns give the number of the plants when used as males;
the horizontal rows are the same plants when used as females. The
result of each mating made is denoted by the letters F for fertile and S
for sterile.

It was planned to make all possible combinations of these plants ; but
this proved to be impracticable, and only 1 5412 were accomplished. The
pollinations on the plants of this generation as well as those on the
succeeding generations included in this experiment were made under
various conditions of sunshine, temperature, moisture, food supply and
age, but these variables appeared to have no influence on fertility. The
results always checked. A small number of matings were made in the
open field in August and September, 191 1. The remainder were per-
formed in the greenhouse. A part of these were made upon some of the
old plants that had been transplanted during the late fall, and the others
upon cuttings from the plants in the field which were again ready for
operation in April, 19 12. But in all the work on the 20 plants tabled
it should be noted that pollinations were made during the height of the
flowering period when the plants zvere in good condition. Nevertheless,
there may have been errors. If such did occur, cross-fertility would
have been favored; since at the time the work zvas done upon the F2, F3,
and F4 generations of this cross, pseudo self-fertility zvas not suspected.

The plants were each selfed from 2 to 10 times, an average of 4 times
per plant, without a single seed being obtained.

Of intercrosses, 132 were made. 3 of these are indicated by question
marks on the table. This is because plant 5 had defective pollen, it being
the only one of the twenty in which the pollen did not show from 90 to
100 percent of morphologically perfect grains. None of the crosses
where this plant was used as the male gave capsules over half-filled with

12 A few of the figures given here differ from those given in the preliminary report
on this work (East 1915). This is due to rechecking the results and to the addition
of a few more data. There have been no essential changes and the present figures are
believed to be correct.

Genetics 2: X 1917

542

E. M. EAST AND J. B. PARK

seeds, but since 7 matings had from 30 to 100 seeds per capsule, and
since the reciprocal matings were all successful we have classed them as
fertile. The matings questioned, 11 X 5, 14 X 5, and 20 X 5, ought
also to be classed as fertile, since the reciprocals were fertile, but as they
yielded only 2 to 10 seeds per capsule, they have been omitted from these
next calculations.

Of the remaining 129 intercrosses, 126 were successful; 4 of them
produced capsules having less than 50 percent of the ovules fertilized
(2 pollinations each being made), the remainder produced full capsules.
There were few failures among these intercrosses, though from 2 to 12
repetitions of the matings were made in almost every case. It seemed
as though an intercross possible at one time could be made at any other
time at the first attempt. In other words, there seemed to be no varia-
bility in ease of cross-fertilization. The failures in the fertile inter-
crosses were less than 4 percent, and these were complete failures which
may be attributed to the technique used.

Twenty-eight intercrosses between these plants and other plants of
the F2 generation were also made with 28 successes. In addition, 92
other combinations were made between plants not shown in the table.
They are not reported in detail because only a fewT matings per plant
were made ; but the gross results were 89 successes and 3 failures.

Altogether among these matings there were 54 pairs of reciprocals
each of which gave the same result.

The failures in the intercrosses remain to be considered. The table
shows 3 cases; of which 10 X 13 was tried 2 times; 11 X 12, 12 times;
and 12 X 11, 6 times. The last pair are reciprocals, but we shall treat
reciprocals separately for the present. Of the other 3 cases, 2 of them
were tried 3 times, but the third was made only once, which of course
does not settle the matter. Thus there were 4 definite cases, 1 probable
case, and 1 questionable case of cross-sterility, a matter of 2.4 percent
(6 out of 249).

The F3 generation

Out of the many fruitful combinations of F2 plants, 29 F3 families
were grown, — 50 to 150 individuals of each being transplanted from the
greenhouse to the field with due care that random samples were obtained.
Field examinations as described above, indicated a total absence of
self-fertile plants, and from 3 to 6 attempts to self individuals of each
family resulted in failures.

The progeny of 2 red-flowered plants of the F2 generation furnished

THE BEHAVIOR OF SELF-STERILE PLANTS

543

Table 5

Result of matings on 12 plants of the F3 generation of cross No. 1,
N. Forgetiana X N. alata.

Plants used as males
1 2 3 4 5 6 7 8 9 10 11 12

S F F F F F F

S FF F FFFF
F S F F F F

F F S S F F F

F S F F F F

F FSFS FFS F
FFF FFSFF FS
FFF F S FFF
FFFF S F F

FFSSFF SF
F FF FF FSF

F F S

the material for the continuation of our intercrossing experiment. Most
of the work was done on 12 plants as set forth in table 5. Fruitless self-
pollinations averaging over 3 per plant proved they were self-sterile.
102 cross-pollinations were made: 75 are shown in the table; 27 were
made in a less systematic manner with 11 other plants of the same
family. These resulted in 95 successes and 7 failures. Again the
"possible" combinations were almost always successful. The unsuccess-
ful matings were 4 X 6, 6 X 4, 6 X 10, 10 X 6, 7 X 12, and 10 X 5.
Combination 6X4 was made twice and combination 7 X 12 once, the
remainder were made three or more times. The first 4 matings consist
of 2 pairs of reciprocals. The reciprocal of 7 X 12 was also made, but
proved to be fertile. This is evidence that with further trials 7 X 12
would also have been successful, for we have invariably found recipro-
rocals to behave alike when a number of pollinations sufficient to deter-
mine definitely the status of the cross has been made. In fact 26
reciprocals gave the same result in this population. The remaining
combination showing cross-sterility was between plants 21 and 27.
Eliminating combination 7 X 12, therefore, 6 percent gross of cross-
sterility is shown in the F3 generation.

Genetics 2: N 1917

544

E. M. EAST AXD T. B. PARK

Table 6

Result of matings on 10 plants of the F4 generation of
cross Xo. i, X. Forgetlana X -V. alata.

Plants used as males

Table 7

Result of matings on 20 plants of the F5 generation on cross
Xo. 1, X. Forgetiana X N. alata.

Plants used as males

19 2C

S F

s s

F S

F F

S F

U
-.

i —

■s.

F S

S S

THE BEHAVIOR OF SELF-STERILE PLANTS

545

The F4 generation

Only 2 of the F3 combinations were grown during the next season and
the pressure of investigations along other lines was such that but little
work was done upon them. Field examination and tests on 21 plants,
however, showed us no self -fertility. 10 of the progeny of 2 red-flowered
F3 plants, had 52 matings made upon them, 15 being reciprocals giving
duplicate results. In addition 6 random matings with other plants of
the family were tried with 1 failure. There were 48 successful and 4
unsuccessful matings on the 10 plants shown in table 6. The fertile
matings yielded good capsules as usual with 3 exceptions, there being
but 7 complete failures out of over 200 pollinations. Of the unsuccess-
ful combinations, pollinations were made as follows : 2X8, 4 trials ;
5 X 2, 6 trials; 5 X 8, 4 trials; 8 X 5, 5 trials; and 8 X 12, (not shown
in the table) 4 trials. Each of these cases is fairly certain, therefore,
and gives us a gross cross-sterility ratio of nearly 9 percent.

The F5 generation
Only 1 F- family was studied, but as it was planned to discontinue this
particular experiment, considerable attention was given to it. As was
also true of the F3 and F4 generations, the work was carried on under
field conditions. Similarly again, it was produced by mating two red-
flowered sibs.

A random sample of 20 plants was marked for work, and 439 pollina-
tions made (table 7). Of these pollinations 92 were wholly unsuccess-
ful attempts to secure selfed seed made on 17 plants, an average of 5.5
pollinations per plant. Thus there is no question about the self-sterility
of each plant tested. Plants 4, 5 and 20 were not tested. Plant 4 had
such bad pollen that results with it are valueless, and plants 5 and 20
were somewhat sickly. Plant 9 also had such poor pollen that the seed
capsules were not full, but a classification of the matings where it was
used could be made without any serious chance of error. 274 pollina-
tions were made on the 119 intercrosses that proved fertile. Only 12 of
these attempts failed, and 5 of them were on crosses between Xo. 9
and No. 3. Thus only 4 attempts per hundred failed in the intercrosses
that were classed as fertile from records of other pollinations, showing
conclusively, we think, that inbreeding had produced no quantitative
diminution in fertility among "possible" combinations, the percentage of
failures in fertile crosses in the F2 generation being about the same.

The remaining 73 pollinations were unsuccessful attempts to obtain
seed in 33 intercrosses. The details are shown in table 8.

Genetics 2: X 1917

546

E. M. EAST AND J. B. PARK

Table 8

Record of unsuccessful cross-pollinations made on the F5 generation
of cross No. I, Nicotiona Forgetiana X N. alata.

Pollina-

Mating

tions

Mating

tions

Mating

tions

i X 4

7 X 9

16 X 20

2X4

7 X io

17 X 8

2 X S

7 X ii

17 X ii

2 X 16

9X5

17 X 12

2 X 17

io X 5

18 X 3

2 X 20

12 X 8

18 X 7

3 X io

13 X 3

18 X io

io X 3

14 X 6

18 X ii

3 X ii

15 X 6

18 X 13

7 X 3

15 X 14

19 X 5

7X4

i6 X 9

19 X 20

It will be seen that only i reciprocal cross was made on these plants
and this was by accident. A large number of reciprocals had been made
on other crosses ahvays with the same results when tried a sufficient
number of times to make classification conclusive. It was decided
therefore, to make as many distinct matings as possible in order to make
a thorough test of the mating proclivities of the plants under observa-
tion. The result is that the percentage of cross-sterility found in the F5
generation is not strictly comparable with the percentages found for the
earlier generations where matings were made at random and each
mating counted. To be sure a few reciprocal matings13 were made in
F5, but the percentage is very much less than in the preceding genera-
tions. The gross cross-sterility found in F- was 21.7 percent, if the 8
crosses where only 1 pollination was made be counted. By the theory
of error 1 of these cases might be excluded, wrhile for certain other
reasons (see table 9) error is suspected in another case, but since this
correction would reduce the cross-sterility percentage by only 1.2, the
figures 21.7 will be allowed to stand.

Eight other intercrosses between other plants of this same population
were also made. We have not thought it necesssary to include them in
the table because the attempts at crossing were so sporadic, but the per-
centage of cross-sterility would scarcely be changed, for 7 out of 8
intercrosses were fertile.

A number of other facts appear in the data shown in tables 7 and Sr
which are not apparent without careful study. In accordance wTith their
behavior in intercrosses, the plants may be grouped into 5 classes in

13 Seven reciprocals were made altogether in this family with like results.

THE BEHAVIOR OF SELF-STERILE PLANTS

547

which there is intra-class sterility and inter-class fertility, with very few
exceptions. This grouping is shown in table 9. The two columns at

Table 9

Plants of Fs (feneration of cross No. 1, N. Forgctiana X -V. alata, grouped in
accordance with their behavior in intercrosses.

Fed.

Number cases fertile
within group

Number cases sterile
within group

JNO.

! E

10
11

0
0

3
4

3
3

3
1

3
3

1?
1

0
0

i §

the left show the division into groups, and the pedigree numbers of the
plants within each group. The next 5 columns show the number of indi-
vidual cases of cross-fertility within each group. For example, plant
Xo. 3 was fertile with 5 plants of group B, with 2 plants of group C,
with 3 plants of group D, and with the single plant comprising group E.
The last 5 columns show the number of individual cases of cross-

Ge.netics 2: X 1917

54«

E. If. EAST AXD J. B. PARK

sterility within each group. The exceptional cases where there is inter-
class sterility or intra-class fertility are printed in bold-face type. By
utilizing the mating record of a plant either when used as a male or
female in making the classification, all of the plants could be grouped
excepting number 4 which had very bad pollen. It is excluded on this
account.

The number of exceptions appear at first sight to be rather large but
it must be remembered that one exceptional mating makes two irregulari-
ties appear in the table. If 7 is sterile with 9. 9 is sterile with 7, for
example, and both exceptions are noted.

Xumber 2 and number 17 are anomalous plants; the remainder
behave very regularly. 2 is sterile with 17 where one would expect to
find fertility: this is also true of the mating 17 X II- Both of these
matings were made twice, which establishes the sterility rather definitely.
In addition 2 is fertile with both 9 (thrice) and 19 (twice) of the same
group, though it properly belongs in group B from its sterility with 5,
16 and 20. and its fertility with at least 1 plant of each of the other
groups. The mating between 2 and 9 was fertile only 3 times in 6 trials,
however, and may indicate a pseudo-fertility due to external conditions.
The cross was made reciprocally; 2X9 was fertile in both trials, but
9X2 was fertile but once out of 4 trials.

Eliminating plants 2 and 17 from consideration, there are left only 4
unconformable matings. There are 2 cases of inter-class sterility. 9
with 7 and 10 with 5. Each of these matings was made but once,
however, and their sterility is questionable because 4 times per 100 one
obtains no seed in matings that otherwise prove fertile. The exceptional
fertile matings, 5 with 16 (thrice ) and 16 with 19 (twice), on the other
hand, appear to be definitely established.

If one admits the possible fertility of combinations 9X7 and 10 X 5.
then, 16 plants allow themselves to be grouped into five classes A, B, C.
D, and E, with no anomalous behavior whatever. Each is cross-sterile
with every plant of its own class and cross-fertile with every plant of
every other class with which it is tested. True, 3 anomalies remain,
plants 2. 16 and 17. Xumber 17 of class D shows a perfectly regular
behavior except with plant 11 of class A and plant 2 of class B. Plants
2 and 16 show their irregularities only within their own class except in
the cross between 2 and 17, which leads us to suspect pseudo-fertility.

The conclusion seems just, therefore, that this grouping is real and
significant, since the great majority of these plants ( in this sample of the
population. 84 percent ) shows an absolutely regular behavior and the
small minority of exceptional plants presents but a few irregularities.

THE BEHAVIOR OF SELF-STERILE PLANTS

549

If one admits the justice of this classification there comes the ques-
tion of the number and composition of such groups in the F5 generation
of this cross. 19 plants form a very small sample of such a population.
What is the composition of the whole population? The first thing to be
noted is the varied size of the groups. The number of individuals in
each class is 6, 6, 3, 3, and I, respectively. Even with a due allowance
for the smallness of the sample, it is clear that there is little probability
of the plants being distributed in equal-sized classes. It is hardly more
probable that the distribution will fit a Mendelian + YaY expansion.
It is reminiscent, however, of a normal binomial expansion (J/2 + l/i)n.
The resemblance is possibly illusory, but o, 3, 6, 6, 3, 1 is too much
like 1, 5, 10, 10, 5, 1 to escape notice, particularly as on the theory of
random sampling it is possible for the whole population to contain from
1 to 3 more classes. Be that as it may, we can certainly conclude that
the F5 generation of this particular cross contains no more than from
6 'to 8 groups — the chances are practically negligible that there might be
10 — which are intra-class sterile and inter-class fertile, and within which
the distribution of individuals bears some similarity to that of a normal
frequency distribution.

Let us now consider whether a possible meaning can be attached to
the results obtained in this experiment.

Argument on cross No. 1

We early assumed a working hypothesis in part similar to and in part
different from that of Morgan, viz.; first, self-sterility is heritable;
second, as regards that part of the constitution of pollen grains which
affects the behavior of self-sterile plants all pollen grains produced by
each plant are alike, i.e., with reference to self-sterility pollen grains
behave as if they were sporophytic; third, under normal conditions the
pollen tubes produced by pollen from any self-sterile plant will not grow
in styles of that plant with a rapidity sufficient to reach the ovules during
the "life" of the flower, on account of this "likeness" of constitution;
fourth, pollen tubes will grow with a rapidity sufficient to allow fertiliza-
tion to occur if the constitutions of the two proposed parental plants
differ from each other in any of these essential factors, by reason of a
stimulus possibly analogous to that which makes growth more vigorous
in first generation hybrids.

The first assumption has been demonstrated more or less clearly by
all who have worked upon self -sterile plants. It is proved for self-
sterile Nicotiana species both by the experiments reported here, and by

Genetics 2: N 1917

.550

E. M. EAST AND J. B. PARK

those to be published later on the relation between self-fertile and self-
sterile plants.

The second assumption is proved circumstantially by the fact that
reciprocal matings always duplicate each other. Direct experiments
showing that selective fertilization does not occur have also been made,
and will be the subject matter of another paper.

If there be any justification for the third and fourth assumptions, a
cross between two self -sterile species differing by a large number of
hereditary factors (expecting some of the differences to be effective)
should show a high degree of cross-fertility in the Fx and F2 generations,
followed by an increasing percentage of cross-sterility in later genera-
tions produced by the closest possible inbreeding. The reason for such
a belief is, of course, the well-known fact that inbreeding increases
homozygosis. Such being the case, plants ought to appear with "like"
constitution as far as the factors affecting cross-fertility are concerned,
and these should be cross-sterile to each other. If the factors affecting
cross-fertility are relatively few in number, a small number of intra-
sterile, inter-fertile groups should be found after a comparatively limited
amount of inbreeding. This, broadly speaking, we believe to be a plaus-
ible interpretation of the facts found. A detailed interpretation is given
later.

In general, the F2 generation of such a cross — between species —
might be expected to show an approach to the maximum limit of cross-
fertility, since the F2 generation usually shows greater variability than
succeeding generations. But in the case of self-sterility where the self-
sterile plants must be supposed to differ in constitution among them-
selves, this is probably not strictly true. If one could test a large series
•of F2 populations from various original and F± matings, he ought to
find a variable degree of cross-fertility, with the maximum reached only
in certain cases.

In this instance, no claim can be made that we are dealing with the
maximum. We can only report the results for this case, pointing out
that in crosses No. 2 and No. 3, the cross-fertility is much lower.

One of the best systems of inbreeding in the case of self-sterile plants
is to mate sister plants in successive generations, for such crossing, after
an original mating Aa X Act, by Mendelian recombination ultimately
gives a population in which A A and aa each approach y2 and Aa ap-
proaches o. Expectation of homozygosis in successive matings is ]/>,
H> if - - - - 1 (Jennings 1916). This system seemed to suit

THE BEHAVIOR OF SELF-STERILE PLANTS

551

our purpose better than any scheme of mating parent with offspring,
because of the difficulty of keeping plants alive for several years.

It is regretted that so little is known about the cross-fertility of the
F1 generation, but this bit of ignorance does not affect our test seriously.
This really begins with the inter-cross of two self-sterile Fj plants,
which were similar in appearance, but were producing numerous differ-
ent types of gametes, as is proved by the extremely variable re-
generation.

The cross-sterility14 of the FL> generation was 2.4 percent, if the sixth
case of sterility where only one pollination was made, be included. The
result on the 20 plants tested rather thoroughly was 3 cases of cross-
sterility out of 131 matings. The result on the other twenty-odd plants
tested less thoroughly was 3 cases of cross-sterility out of 120 matings.
And this percentage of cross-sterility may have been too low, as was
mentioned before, because of our failing to suspect pseudo cross-sterility
at this time. But taking this low estimate of cross-sterility at its face
value, it is clear that no hypothesis of Individualstofife (Cf. Jost 1905) is
necessary to account for the results. The presence of even 6 cases of
■cross-sterility in 251 matings eliminates this requirement definitely.

The number of classes which would be necessary to give such an
amount of cross-sterility, on the assumption of inter-class fertility and
intra-class sterility, depends upon what is presupposed as to the fre-
quencies within the classes.

As we shall have a number of such estimations to make, let us con-
sider the matter here. It is always dangerous to calculate a posteriori
probabilities. But because this danger is realized, and the probabilities
calculated must be used with caution, it seems best to use as simple an
approximation as possible. Therefore, we have assumed that if ^ and F
represent the total number of sterile and of fertile matings found, the

obtaining this figure, self X self is added to the cross-sterility of course,

then gives us a measure of the probable number of classes for

14 In our preliminary paper (East 1915) judgment was withheld as to the validity of
the apparent cases of cross-sterility in the F, generation of this cross. The recovery
of a misplaced data card with records of duplicate cross-pollinations made on the com-
binations that had shown apparent cross-sterility, by Dr. White, gives us the grounds
for our present conclusions.

probable error of the determination

5 representing the total of sterile combinations. The fraction

S + F

Genetics 2: X 1917

552

E. M. EAST AND J. B. PARK

(cv + cy + Cr2)

, where r is the number of

s + f i(Ca + cb+ cry

classes, and Ca, Cb, etc., are the number of individuals within each class.
If the classes are of equal size, the ratio of sterility to total number

of combinations is

where n is the number of classes; for if there

nx2 i Tr
= . It on the

are x individuals in each class the sterility is

W x* n

other hand, the distribution of individuals within the classes is that of
the coefficients of the point binomial, these coefficients must be
substituted.

With these two assumptions as to distribution, the following percent-
ages of sterility to total number of matings is found :

Number

Point

Number

Point

of classes

Equal size

binomial

of classes

Equal size

binomial

25.0

31-3

6.6

15.0

20.0

273

6.2

14-5

16.7

24.6

5-9

14.0

14-3

22.6

5-6

13.6

12.5

20.9

5-3

13.2

11. 1

19.6

5-o

12.8

10. 0

18.5

4-8

12.5

9.0

17.6

4.5

12.2

8.3

16.8

4-3

1 1.9

7-7

16.1

4.2

1 1.7

7,1

15-5

2-5

9.0

Should one wish to make the calculation from cross-sterility only on
account of the self-sterility determinations being selected values the
formula becomes

sc _2(C7 + cv+ Cr2)-2(ca + c6+ cr)

Cr)

Sc + F 2(Ca + C6+ Cry — 2(Ca + c6 +

but this correction is unnecessary under most circumstances.

Taking now the gross returns on the F2 generation at their face value,
2.4 percent cross-sterility, or 15.8 percent total sterility on the 40 plants
used, the number of classes of approximately equal size necessary to
account for the results is between 8 and 14. But the groups which were
afterward found in the F5 generation, were not of equal size. Their
frequencies resembled rather those of a point binomial. Assuming such
a distribution within the classes of F2, the number of classes wTould lie
between 12 and 25.

THE BEHAVIOR OF SELF-STERILE PLANTS

553

These class number determinations have been made roughly on pur-
pose. There are three reasons for doing this.

In the first place, there is reason to believe that the proper percentage
of cross-sterility was not obtained. Our calculations were made by
including with the matings listed in table 4, 28 matings of plants shown
in table 4 with other plants, and 92 matings made rather unsystematic-
ally between about 20 plants not shown in that table. Data as to the
age, condition, flowering period, etc., of these plants were not recorded.
Furthermore, fertility and sterility were usually recorded merely as
F and S without data as regards the percentage of seeds in the capsules.
For our present purpose, therefore, they have not the value of the data
recorded in table 4.

Probably the correct way to treat the data of table 4 would be as
follows. Consider every mating as if it were made reciprocally whether
actually accomplished or not. For example, 5 X 1 is fertile; then
assume 1 X 5 to have been fertile even though that mating was not
attempted, since reciprocals always have given the same results. If this
be done the records show 184 cases of cross-fertility, 4 cases of cross-
sterility and 20 cases of self-sterility. Another question then arises.
These plants supposedly were all in good condition and in general were
mated only at the height of the flowering season. But we did not at that
time suspect pseudo cross-fertility, and made no particular attempt to
clear up doubtful cases, as was done later on crosses No. 2 and 3. Now
crosses 3X17, 5X15, 6X8 and 19 X 14, although made twice each,
showed less than 50 percent of the ovules fertilized. The reciprocal of
6X8 was clearly fertile, so this mating remains in the "fertile" column.
But there is good reason from analogous results in the other families for
considering the other 3 matings as sterile. The mating 5X15 may be
questionable, but as 5 had such bad pollen we cannot be certain of the
placing of mating 15 X 5, as was stated earlier. If then we remove
these matings from the fertiles to the steriles, which seems the logical
thing to do, there are 178 cases of cross-fertility, 10 cases of cross-
sterility and 20 cases of self-sterility. The total percentage of sterility
is 14.4 (30:208), with very little selective advantage to sterility on
account of self-fertilizations.

With these facts in view, we believe it reasonable to assume that
between 8 and 14 approximately equal-sized intra- sterile classes or be-
tween 12 and 25 intra-sterile classes with the individuals distributed
according to the point binomial coefficients, are represented in the F2
generation, — these being taken as distributional extremes.

Genetics 2: N 1917

554

E. M. EAST AND J. B. PARK

The second reason for approximating the number of classes is because
the number of individuals investigated is comparatively small, and the
probability that they are not a fair sample of the population corre-
spondingly large.

The third reason is that the probable upper limit of the number of
classes is all that is essential to our purpose. The point is, that should
the anszver lie between 2? and 81 classes, the difference could be ac-
counted for by 1 additional Mendelian factor pair. The number of
actual classes in the F2 generation of a Mendelian population is 3" where
n represents the number of allelomorphic pairs; and 33 is 27, while 34
is 81.

Thus it is clear that with the assumptions made previously regarding
the cause of self -sterility, our probable maximum cross-fertility can be
interpreted by 3 (possibly 4) effective allelomorphic pairs.

For the same reasons for which it was thought best to correct the
gross percentage of cross-sterility found in the F2 generation, the later
generations of this cross ought to be revised.

Considering then only the matings of the F3 generation shown in
table 5, if one counts reciprocals fertile or sterile as the case may be with
the mating made, there are 98 fertile combinations and 6 sterile combina-
tions. But mating 1X5, made twice, yielded capsules only 30 and 35
percent full, respectively; and mating 9X3, made thrice, yielded cap-
sules only from 20 to 30 percent full. If, as seems probable, these are
really sterile matings, the ratio of cross-sterility to the total number of
cross-combinations becomes 10 to 104 or 10 percent, and the ratio of
total sterility to total number of combinations becomes 22 to 116 or

19. 1 percent.

Similarly correcting the results listed in table 6 for the F4 generation,
we find 16.2 percent of cross-sterility in the cross-combinations and 26.2
percent of total sterility in all combinations, with indications that plants
2, 5 and 8 belong in one class, plants 6 and 7 in a second class, and
plants 9 and 10 in a third class. This result is obtained thus : there are
listed 68 fertile and 6 sterile combinations, but matings 10 X 9 (made
twice), 6X7 and 7X6 are now classed as sterile because they uni-
formly gave capsules less than 40 percent full.

The cross-sterility of the F- generation has already been analyzed
sufficiently carefully in explaining table 9. Measured as above it is

22.2 percent.

Unquestionably the samples of the populations from which these re-
sults were obtained were so small and the number of matings so few,

THE BEHAVIOR OF SELF-STERILE PLANTS

555

that the probable errors are large; but rough as the determinations may
be, zee think that no one can question the general conclusion that in these
three generations from repeated sib matings cross-sterility has increased
immensely.

The cross-fertility of F2 in this cross, as compared with the cross-
fertility in those to be described next, is high. It may not be the maxi-
mum cross-fertility possible in a population from one original mating,
but it is the highest found in 16 families that we have studied rather
thoroughly. High as it is, nevertheless, the probable maximum number
of inter-fertile, intra-sterile classes which it contains is less than 25,
and this number may be interpreted by the permutations of 3 Mendelian
allelomorphic pairs. Further the probable number of these classes in
the F5 generation can hardly be more than 8, a figure which may be
interpreted by only 2 effective allelomorphic pairs. We were decidedly
in error, therefore, when in 191 5 we said (East 191 5) : "This is a
straight mathematical problem and it is hardly necessary to say that it is
insoluble by a strict Mendelian notation such as Correns sought to give."
In justice it should be said, however, that at that time, the existence of
cross-sterility in the F2 generation was uncertain through a supposed
lack of confirmatory data which was really in our possession and had
been overlooked.

Cross 2. N. (data X Ar. Forgetiana- (pseud 0 self -fertile X self -sterile)
and cross J. N. Forgetiana X N. alata (self -sterile X pscudo self -fertile )

The two crosses to be described next are reciprocals made with the
same two individuals. It was our intention to repeat the cross just
described together with its reciprocal, and to make a more thorough
study of the first hybrid generation. At the same time we intended to
study the relation between self-sterility and self-fertility by crossing
N. Forgetiana with a fertile plant of N. alata, since N. alata was then
supposed to be a mixed population consisting of self-sterile and self-
fertile plants. Both of these crosses were made. In crosses No. 2 and
No. 3 the "self-fertile" daughter of the original supposedly self-fertile
plant described on page 534 was used as the N. alata parent. Soon after
work was started on these plants, our evidence was so conclusive that
N. alata was always self-sterile and that this particular individual
showed only pseudo-fertility caused by external conditions, that we
decided to use N. Langsdorffii as the self-fertile strain in a series of
crosses and to continue this work as a repetition of cross No. 1.

Genetics 2: N 1917

556

E. 11 EAST AND J. B. PARK

Table io

Result of matings on Fx plants o to 39 of cross So. 2, N. alata X N. Forgetiana and
on plants 40 to 32 of cross Xo. 3, X. Forgetiana X N. alata. Xumber
of pollinations shown by subscripts.

9
10
11
12

13
14
15
16
17
18

19
20
21
22
2^
24
25
26
27
28
29
30
31
32
33
34
35
36

37
38
39
40

42
43
44
45
46
47
48
49
50

44, 463

2, 3, 4, 6, 41

4, 18, 41, 44, 52
2, 9, 14, 23, 29
22, 9, 10, 44=

2, 3, 6, 9, io., 182, 46

5, 10., 43, 44

2, 13, 22, 44

6, 9, 10, 39, 40, 463

3, 18, 44, S%

4, 6, 18, 40, 44

2, 8, 12, 15, 34, 44, 46
9, 16, 22, 43

3, 8, 182, 44, 46
18.., 20. 43

1. 3, 16,, 17, 18, 20

13, 14, 18, 25, 433, 46

14, 18, 19, 20, 22. 30

22, 2I2, 23, 28, 34, 36. 44

17, 22, 28, 34, 44

2, 83, 9, l6, l8. 20;, 21, 22:, 26, 36, 40, 44

4, 12, 16, 18, 46
12, 42, 44

32, 6, 20, 26, 28, 44

8, 33, 442, 463

9, 18, 22, 23:, 25, 40, 48
35, 18, 32, 44, 46

2* 3, 23., 27, 39, 4^1

2, 143, i&, 22, 23, 24, 25, 30, 34, 37, 41, 46
83, 29, 33, 44., 45, 46

22, 32, 52

9, 21, 23, 29, 30, 34, 43, 44

8, 16, 23, 31, 46
28, 41, 44, 46

3, 9. 18, 21, 27, 30, 34, 37, 42
8„ 33, 442, 46

39, 42„ 43. 44, 46.
28, 35, 39, 42, 43, 462

9, 44

22, 43, 44, 47, 49

10, 37, 44, 48
20, 44

5, 27. 33, 38, 39. 40,. 42, 44. 46, 51
10, 14, 23, 34, 45

18, 44, 48

10, 22, 37, 44, 51

20, 42, 44, 45, 46, 51, 52

40, 41, 43, 46
42, 44, 45

18. 39. 5*i 52

9, 18, 23, 39, 45, 46, 50

10, 23, 29, 37, 51

222, 34, 383, 49
&

9, 22, 23,

4., 63, i83, 4i3, 46
18

8=, 442

3c 43, iS,, 40=
183, 46
5z, 44*

2*, 10, 232, 37, 48

22, 23. 242, 27:, 343, 482
63, 183, 463, 52

2, 92, 153, 212, 34t

10, 34i

93, .I32, 14, 23., 44
I7:, 293
16. 26-, 44,

3, 463
182

432

2, 94, 222, 25, 27a 37
14, 23,, 24, 36, 483
92. io-, 37, 482

102, 222, 23, 30, 37
22, 9. 233, 27

28, 293, 443

2, 92. 303, 343. 48

8, 262, 293, 443
5:, 26, 28, 31,, 443

9, 2I4, 223, 27

82, 29, 36. 44
18, 33, 462
185, 322

io,, 232, 24, 372
83

103. 23

92, IO3, 22, 234, 343, 38

34?, 3~3. 47
18, 40,, 423
6, 333. 462
332, 'O, 46,
393, 41. 45*

462, 52

52,

382

10, 232, 242, 272, 34
o, 9, 27. 343, 47
9, 273, 372
8, 29

32, 42, 6. 182. 4 1 2, 45s, 462

THE BEHAVIOR OF SELF-STERILE PLANTS

557

It is reasonable to consider these crosses in a sense to be repetitions of
cross No. i, but one must not assume that they are duplicates of cross
No. i. Both N. alata and Ar. Forgetiana must consist of plants which
differ among themselves in the factors that affect self -sterility, hence
only by following through a number of F1 generations where these spe-
cies are involved could one expect to find results duplicating those of cross
No. i. The data are none the less interesting, however, because the
crosses are only similar and not identical.

The F1 generation
All of the individuals resulting from this cross were grown in a green-
house as potted plants. The F1 generation came into blossom during
the latter part of the winter. Conditions were extraordinarily favorable
for growth and the pollinations were all made while the plants were
vigorous, hence scarcely any trouble arose over classification of the
results.

Our study was made on a population of 53 plants. Pedigree num-
bers from o to 39 inclusive represent cross Xo. 2, N. alata X N. Forge-
tiana; pedigree numbers 40 to 52 inclusive represent cross Xo. 3,
N. Forgetiana X AT. alata.

Each plant was selfed one or more times, and all proved absolutely
self-sterile. Further each plant was back-crossed with pollen from a
single plant of each of the parent species with complete success in every
instance. The plants used in this work were not the individuals that
entered into the cross under discussion, however, for unfortunately
these were not available.

The numerous cross-pollinations made are shown in table 10. There
were 103 reciprocal matings. Of these 100 gave duplicate results, 39
pairs being fertile and 61 pairs sterile. The three which did not check
are :

2 X v sterile, 1 pollination ) . . - ...

0 _ r' . [ classed as lertile

3X2, lertile, 1 pollination )

6 X S2< fertile, 1 pollination 1 . ,
v„> . . „. . > classed as sterile

52 X o, sterile, 1 pollination J

classed as sterile

37 X 21, fertile, 1 pollination
21 X 37? sterile, 1 pollination

558

E. M. EAST AND J. B. PARK

shows conclusively that 3 should be fertile with 2, 6 sterile with 52, and
21 sterile with 37. They have been classed accordingly. That this
grouping is correct is further shown by the fact that the mating 3X2
(classed as fertile) was made at the height of the flowering season,
while the matings 6 X 52 and 37 X 21 (classed as sterile) were respec-
tively the last and next to the last matings made on those plants.

In spite of the fact that plants 0-39 are from cross No. 2, N. alata X
N. Forgctiana, and plants 40-52 are from cross No. 3, N. Forgetiana X
N. alata, they behave as one family in inter-crosses. The entire popula-
tion can be grouped into 6 classes in which there is inter-class fertility
and intra-class sterility (table 11). The following explanation may be
necessary to make it clear just how table n was obtained from table 10.
Table 10 shows all of the matings; but in the form given it is not easy to
see at a glance every combination in which a particular plant was used,
both as male and as female. It was necessary, therefore, to make a new
table in which the pedigree numbers in the column at the left were
tabled as males, and the pedigree numbers in the columns headed "Fertile
matings" and "Sterile matings" were tabled as females. Thus plant 2
used as a female was fertile with pollen from plants 4, 18, 41, 44, and
52, and sterile with plants 9, 22 and 23; but pollen from plant 2 was
fertile on plants 1, 3, 4, 5, 7, 11, 18, 20, 28 and 29, and sterile on
plants 9, 10, 13, 21, 25 and 27. It is clear, therefore, that instead of
the 8 matings on plant 2 that table 10 appears to show, there are really
21, the 3 reciprocals of course being counted but once.

These tables were combined for analysis. In the interest of economy
only one is shown, however, since the second can easily be made from
the first.

The four exceptions in this huge set of matings are in reality negligi-
ble though they are emphasized in the table by bold-faced type. Matings
15 X 44 and 31 X 36 were sterile, though they do not belong to the
same class. Plant 15 was sterile to 4 plants of class A and fertile to
2 plants of class B, 3 plants of class C, and to the isolated individuals
forming classes D and F. It is unquestionably a member of class A.
Plant 44 was sterile to 7 individuals in class C and fertile to 17 plants
of class A, 12 plants of class B and to the singletons forming classes
D, E and F. This evidence places it unmistakably as a member of class
C. Plant 31 is also a member of class C as evidenced by 3 sterile matings
within that class and by fertile matings with 1 plant of class A and 3
plants of class B. Plant 36 is like plant 15 thrown into class A by its
sterility with 3 others of that class, and by its fertility with 3 individuals

THE BEHAVIOR OF SELF-STERILE PLANTS

559

Table ii

Plants of F1 generation of rcciproeal cross between N. Forgetiana and X. alaia,
grouped in accordance with their behavior in inter-crosses. Plants O-jo
are products of cross No. 2; plants 40-52 arc products of
reciprocal cross So. 3.

-j -
|

| Ped.

Number cases fertile within group

Number cases sterile within group

Group

| No.

c |

D |

E |

5 1

Q
O

1 9

0 1

1 10

0 j

7 1

3 1

1 14

2 1

4 I

I 15

1
1

1 21

2 |

0
0

! 22

i» 1

r\
U

1 1

4 1

r»
U

-->

1 "JO

5 1

e
»)

1 34

0 1

5 1

4 I

I j

1 «5U

2 1

2
0

->
3

T 1
1

•5°

r»
\J

-j
»5

AT
1 4/

7
/

ACk

I !

■7
0

0
0

1 4

1 0

r\
*J

1 7

T
I

rv
yj

9 1

J J

1 39

r»
\J

1
0

A T
41

r»

-3
0

r»

L»

1
0

T 1

1 —

—

1 0

1 —

—

-1
0

1 0

1 8

1 7

1 0

1 16

! 0

| 26

1 0

I 6

! 0

1 7

1 3

1 0

1 3

1 0

1 44

! 12

! 1

1 0

1 7

1 0

1 51

1 5

1 0

1 2

1 0

| 20

1 3

1 °

1 0

1 43

1 8

1 4

| 1

1 0

1 °

1 11

1 -
1

560

E. M. EAST AXD T. B. PARK

of class B, with 2 of class C, and with the lone plant of class D. In
view of this evidence and the fact that in these two matings but one
pollination was made in each case, they are much more likely to be
errors of record or of technique than true exceptions to our classification.

The other two exceptions, matings 45 X 18 and 33 X 46, were fertile
where from the evidence of numerous other matings they should have
been sterile. Here again but one pollination was made in each case;
and, coincidence though it may be, each pollination was the last mating
made on that particular plant. What is more probable than that this is
a pseudo fertility appearing during the wane of the flowering season of
the two mother plants. Xo. 45 and No. 33 ?

Six groups appear in table II, but there is proof of the existence of
only five. Groups A, B, C, D and E are definitely established. Plant 11,
on the other hand, is an isolated individual rather than a class. It does
not belong to groups A, B, or C; but unfortunately it was not crossed
either with class D (plant 20) or with class E (plant 43), hence one
cannot say that it does not fall into one or the other of these two
classes.

In the three large groups the distribution of individuals is 22, 16 and
12. About all that can be said about the type of this distribution is that
the classes appear not to be of equal size. On the other hand, it is inter-
esting to note that the plants of both cross Xo. 2 and cross No. 3 fell
into the three groups as if they were samples of the same population.
There were 40 plants of cross No. 1, and 12 plants of the reciprocal
cross No. 2. In the classes A, B and C the proportions were 18, 10,
10 and 4, 6, 2, respectively. This similar behavior of the progeny of
reciprocals seems to us strong corroboratory evidence in favor of the
conclusion that reciprocal crosses always behave in like manner as re-
gards self-sterility.

It is interesting here to check our a posteriori probabilities with the
facts. There were 278 fertile matings made in this family, of which 39
were reciprocals, making 478 (278 X 2 — 78) fertile combinations alto-
gether. There were 167 sterile matings, of which 61 were reciprocals,
making a total of 212 (167 X 2 — 122) cross-sterile combinations. If to
the cross-sterile combinations, the 53 self-sterile combinations be added,
there is a total of 265 sterile combinations out of 743, — a percentage of
35.6 db 1.2. Assuming a point binomial distribution of individuals we
should expect 4 intra-sterile classes for this percentage of sterility; but
since we must discount the selection of self-combinations a little, per-
haps 5 classes may be taken as the probable expectancy.

THE BEHAVIOR OF SELF-STERILE PLANTS

It was planned to continue the study of this family — considering it
as a single cross — on populations obtained by back-crossing a representa-
tive of each of the large classes A, B, and C with both parents, and by
intercrossing the same three individuals among themselves. This rather
Herculean task has not been finished. The progeny of a part of these
matings was investigated as thoroughly as time permitted in 191 5-16,
but much remains to be done. These families came from the following
combinations :

Family D, N. alata plant 53 X plant 44 of class C
Family E, iV. alata plant 58 X plant 44 of class C
Family F, plant 34 of class A X Ar. Forgetiana
Family G, plant 44 of class C X Ar. Forgetiana .
Family H, plant 44 of class C X plant 10 of class A
Family I, plant 44 of class C X plant 34 of class A
Family J, plant 52 of class B X plant 23 of class A
Family K, plant 52 of class B X plant 44 of class C
In families D and E we have two N. alata plants 53 and 58 crossed
with the same plant of cross 3 (table 11), No. 44 a member of class C.
Families F and G were produced by crossing individuals of classes A
(34) and C (44) with the same plant of the other parent species A".
Forgetiana. The four remaining families are true F2 generations formed
by mating two F1 plants. There is a duplicate test of plant 44 (class C)
with two plants of class A, 10 and 34. Then there is a test of plant 52
(class B) with plant 23 of class A and plant 44 of class C. Thus plant
44 of class C enters into two back-crosses with N. alata, one back-cross
with N. Forgetiana, and matings with two individuals belonging to class
A and one individual belonging to class B.

Family D, — Ar. alata plant 53 X plant 44 of class C, cross No. 3

The first of the eight F2 populations of crosses No. 2 and No. 3 was
produced by back-crossing. Plant 53 of N. alata (table 1), a plant
apparently15 fertile with sister plants 57 and 58, and sterile with sister
plants 54, 56 and 59, was crossed with the pollen of plant 44 of class C,
cross 3. In a manner of speaking, it may be called Pt X Fx, if it be
remembered that plant 53 is not the same plant of N. alata used in making
cross No. 3.

Table 12 shows the self-pollinations made on 39 plants. They behaved
in much the same manner as the Ar. alata plants recorded in tables 1-3.
One-third of them produced some seed, though from 1 to 10 failures

15 See page 533.

Genetics 2: X 1917

562 E. M. EAST AND J. B. PARK

Table 12

Family D. — Record of self-pollinations on progeny of X. alata plant 53 X plant 44

of Fx of cross No. 3

Ped
No.

Xo. of selfings giving capsules with

Xo. selfings -
sterile

1-10 seeds

10-50 seeds 1 250-300 seeds

151

153

154

155

156

157

158

159

160

l6l

l62

163

I64

165

166

167

168

169

170

171

1/2

173

174

1/5

176

177

178

179

180

l8l

182

183

184

I8.S

186

187

188

189

were also recorded for the same plants. The remaining plants produced
no capsules. There was an extremely high correlation between this par-
tial fertility which we have regarded as false, and the close of the repro-
ductive period. Yet one cannot say that every plant can be made to pro-
duce seeds at this phase of the life cycle, even under adverse conditions.
This may be the case, but we have been unable to demonstrate it. 4
plants in this family, however, gave a very nice demonstration of the fact

THE BEHAVIOR OF SELF-STERILE PLANTS 563

that complete self -sterility returns with the return of a new flowering
season. A number of these plants were selfed at various times during
two flowering periods, and plants 156, 166, 177 and 178, though giving a

Table 13

Family D. — Record of cross-pollinations on progeny of N. alata plant 53 X plant
44 of .Fi of cross No. 3 outside of family D.

Ped. No.
9

Fertile with
Ped. No.

Sterile with
Ped. No.

152

204 Family E

153

204 Family E

167

201 Family E

171

201 Family E

174

58 N. alata

Table 14

Family D. — Record of cross-pollinations on progeny of X. alata plant 33 X plant

44 of F1 of cross Xo. 3.

Fertile with

Sterile with

Ped. Xo.

Ped. No.

151

159

152

160

185

153

152

154

151

153

155

154

157

154

159

160

174

161

162

163,, 168

163

185

168

173

170

168

172

175

173

I592

174

175

185

177

168, 182,

183

179

177

180

177

181

183

182

160, 183

185

183

177

184

185

188

185

160, 174

186

185

160?

187

185

188

185

Genetics 2: N 1917

5^4

E. M. EAST AND J. B. PARK

few poor capsules at the end of the first flowering season, showed com-
plete self-sterility from the beginning to the height of the second flower-
ing period. Then, in two cases, the slight degree of fertility shown at
the end of the first flowering season returned. 3 plants produced full
capsules. Xo. 160 and No. 175 yielded 1 each, both according to the late

Table 15

Family D. — Progeny of N. alata plant 33 X plant 44 of F, of cross No. 3 grouped
in accordance with their behavior in inter-crosses.

Group | Xo.

! Ped.

Xo. cases fertile within group

No. cases sterile within gr

oup
Ind.

1 1 1
A | B | C

1 1 1

D 1 E 1 Ind.

1 !

1 1 1 1 1
A I B 1 C J D E
1 I I 1 1

153
154
155
175
182

185
187

0 | 1

0 | —

0 | —
0 j 1
0 | 3

2 —

— ! 2
1 - 1 - '

2
1

2 | 0 j — | — | — | —
2 1 - 1 - I - 1 - I -

2 1 — 1 — — 1 — 0

i j j j

,i|-|-|-|-jo
11 Jo] — 1 0 1 — 1 —
I 4 j 0 I 0 1 — 1 0 | —
ill— 1— 1— 1— I —

| 160
B I 174
186

3 1 0

1 i 0

1 | 0

-. i -
- 1 -

0I2T— 1— 1— 1—
' 0 | 1 i - I - 1 - | -

161

r 162
L 1 163
! 168

1 -

0
0

- 1 —

- 1 -

1 - 1 -

1 ! —

- 1 - 1 1 1 - 1 - I -
!- 1 - 1 3 1 - 1 - 1 -

— i — i 1 0 I — 1 —

1 177
D ! 181
I 183

1 1 —
1 —

1 1 0 1 —

— 1 0 1 —

2
-

0 1 — | 0 | 1 | —

* 1 1 . 1 % 1
0 — ! — 1 2 I —

E 1 184
^ 1 188

1 I —
1 | —

— 1 — 1 0

— I — ! 0

I 1

1 ,

0 1 — I — I — | 1

0 1 — f — 1 r— I I

1 1 !

Table 16

Family E. — Record of cross-pollinations on progeny of N. alata plant 58 X plant 44

of Ft of cross Xo. 3.

Ped. No.
9

Fertile with Ped. Xo.

Sterile with Ped. No.

191

197

192

191

193

44* 583

194

58,

195

193, ^04

197

199

195

200

204

202

197

204

152 Family D

205

204

THE BEHAVIOR OF SELF-STERILE PLANTS

565

season expectation. No. 165, on the other hand, was somewhat of an
exception to the usual rule, in that it produced 6 full capsules out of 11
pollinations. There were 2 failures and 2 small capsules with from 1-10
seeds each from pollinations made during the height of the flowering
season. Toward the end of the flowering period the plant was tested
again and yielded 6 good capsules out of 7 flowers selfed.

Five plants of family D were crossed with individuals outside of that
group, as is shown in table 13. 2 plants, 167 and 171, were fertile with
the pollen of plant 201 of family E, while 2 other plants, 152 and 153,
were sterile to the pollen of plant 204 of family E. Plant 174 was fertile
with N. alata plant 58.

Only 36 cross-matings were made between plants of this family (table
14). Of these, 16 were failures. In spite of this small number of inter-
crosses, 20 out of 28 plants can be shown to belong to not over 5 classes
wherein the plants are intra-class sterile and inter-class fertile (table 15).
The other 8 plants show only 1 or 2 cases of cross-fertility and no cross-
sterility, and may or may not belong to separate groups. Their fertility
with the other classes is shown in the column marked "Indeterminate."

There are no exceptions in table 15. Each plant in every group is
wholly intra-class sterile and inter-class fertile as far as it was tested.
But these five groups are not necessarily independent. A is not B, C, Dt
E, 151, 157, or 172 ; B is not A; C is not A, D, 170 or 173 ; and D is not
A, C, 179 or 180. Therefore B may be C, etc., and the existence of only
three groups is demonstrated.

An estimation of the number of classes by formula is hardly desirable
on account of the small number of combinations made per plant, though
the total number of combinations is larger than appears at first sight
because only 1 reciprocal (sterile) was made. There are really 70 com-
binations of which 30 are sterile, a cross-sterility percentage of 42.8.

Family E. — N. data plant 58 X plant 44 of class C of cross No. 3

Family E resulted from a cross between N. alata plant 58 and plant 44
of class C, cross No. 3. The interesting thing about the family is its lack
of fertility not only when selfed but also in crosses. 10 plants were
mated together in such a manner that the chain of evidence was not
broken, as can be seen by studying table 16, with no evidence whatever of
any fertility between them. They all belong to one class showing perfect
intra-class sterility. In addition, if one may assume that all of the
individuals would have behaved as plants 193 and 194, the group was
sterile to the 2 parents. Plant 204 was also sterile reciprocally with plant

Genetics 2: X 1917

E. M. EAST AND T. B. PARK

152 of family D. and as a male with plant 153 of family D. The onlv
sign of cross- fertility shown was when pollen from plant 201 i which
also belonged to family E) was used on plants 167 and 171 of familv D.
yet in appearance the pollen of these plants was perfectly good.

It is unfortunate that the behavior of more plants of this family was
not investigated, but a good many plants needed attention at the same
time during the period these were in flower, and the importance of
establishing definitely whether the entire family belonged to one class
was overlooked until too late. It is dear, however, that if other classes
existed, they must have contained relatively fewer individuals than the
one found.

Judged by its parents family E appears to be a duplicate of family D.
A", alata plant 58 was apparently fertile to its sister plants 53 and 59.
and sterile to its sister plants 54. 56. 57. 62. 64. 66. 71 and 79 : plant 53.
the female parent of family D. was apparently fertile to plants 57 and 58.

T.V3LE 17

Family F. — Record of self-polKnations on progeny of plant 34 of F, of cross
Xo. 2 X plant A A of X. Forgetiama.

Xo. of settings giving capsules with

Xo. selfings

Pe<L Xo. sterile 1-10 seeds 10-50 seeds 250-300 seeds

211

212

—

214

21 =

216

217

218

219

-225

2."

228

229

230

231

232

234

-}-

236

237

2}±

2J0

--5

^44

THE BEHAVIOR OF SELF-STERILE PLANTS

and sterile to plants 54, 56 and 59 of the same family. But considering
the behavior of Ar. alata plants 53-79 of table 1 as a whole there is good
reason to believe that they all belong to 1 intra-sterile class and that the
fertility of matings 53 X 57, 58 X 53 and 58 X 59 is pseudo-fertility.
For this reason one might expect family D and family E to behave simi-
larly; but unless one assumes the existence of other classes of low fre-
quency in family E, their behavior was different.

Family F. — Plant 34 of class A X plant AA of .V. Forgctiana

Family F resulted from crossing plant 34 of class A, cross Xo. 2, with
a plant of N. Forgeticuia; but, as in families D and E, it was not a true
back-cross, since the plant of N. Forgctiana used was not the individual
that participated in the original mating.

Selfings were made on 27 hothouse-grown plants with the results
shown in table 17. It will be noticed that only 3 individuals produced
any seeds at all. Xo. 225 yielded 1 capsule containing 8 seeds in 12
tests; No. 236 produced 1 capsule containing 7 seeds in 7 trials; and Xo.
241 finally produced a single capsule having about 30 seeds after 12
pollinations. This is a considerably smaller seed production than was
recorded for family D, and we believe it to be due to the fact that family
F came into blossom somewhat later than family D, thus making it
practicable to conclude the pollinations during the height of the flower-
ing season.

A few pollinations were made between plants of this family and plants
of family G, the results of which are set forth in table 18. They will
be discussed when describing that family.

We were able to make 151 cross-matings on this family, with the re-
sults shown in table 19. Some of these matings, unlike the self-pollina-
tions were made rather late in the flowering season. These made trouble
in some cases, and had to be repeated several times before a proper
decision as to fertility or sterility could be made. In all there were 17
matings that gave seeds in some tests and none in other trials. If the
capsules were full and the majority of pollinations succeeded, the mating
was called fertile; if the capsules were small and poorly filled, and the
majority of the pollinations failed completely, the mating was called
sterile.

These 17 matings, we believe, are listed correctly, but there are a
few matings made but once during the latter part of the season which
may be recorded erroneously.

In addition, plant No. 225 had poor pollen and decision as to the

Genetics 2: N 1917

568

E. M. EAST AND J. B. PARK

Table 18

Family F. — Record of cross-pollinations on progeny of plant 34 of Fx of
cross Xo. 2 X plant AA of N. Forgctiana outside of family F.

Pod. No.
9

Fertile with Ped. No.

Sterile with Ped. No.

216 .

278 fam. G.

247 fam. G.

219

250 fam. G.

239

247 fam. G.

241

250 fam. G.

243

247 fam. G.

244

247 fam. G.

Table 19

Family F. — Record of cross-pollinations on progeny of plant 34 of F1
of cross No. 2 X plant AA of N. Forgctiana.

Ped. No.

Fertile with Ped. No.

Sterile with Ped. No.

207

211, 216, 225

209

212, 2l6, 231

211

2093, 214, 2l6

2I22, 215

212

214., 2i64, 231

214

209, 211, 2l62, 217, 219, 228

215

214, 217,, 219, 222

216

212, 217,, 2I92, 223

239

217

2l62, 2l8, 2I92

212

218

217, 219

219

2140, 215, 216, 2I74, 2180, 227

228

221

227

222

2I72, 2l8, 2I92

223

2l6, 2253, 227, 228, 230, 236

224

217, 219, 223, 2252

225

2l6, 2I72„ 219, 221, 223, 228, 2302, 234,, 235

226

227,, 228, 230, 2343

223

227

2190, 224, 2253

228

222, 2232, 2252, 227, 230

219

229

209, 214, 2l6, 219, 23I4
223, 2252, 227, 234, 236

230

231

212, 214, 219, 229-

232

236, 239,, 243

219, 234

233

223, 234, 239

234

225, 226, 230,, 239,

219, 228, 2322

235

236., 239

232

236

232, 234., 239., 243

233

237

2352

238

239, 243. 244,

239

219, 232, 235, 236, 240

2432

240

234, 236, 239, 243

241

2340, 236, 2443

239, 2433

242

239, 2430

243

234, 2442

2394, 241, 242

244

236, 2380, 239,, 241,, 243,

245

238., 24 1 2, 243, 244

THE BEHAVIOR OF SELF-STERILE PLANTS

569

character of three matings (with 219, 227 and 230) was made on the
basis of the successes obtained when No. 225 was used as female.

There were 23 unsuccessful and 128 successful cross-matings in this
family. Of these combinations, 55 were reciprocals fertile in both mat-
ings and 10 were reciprocals sterile both ways.

Eighteen of the plants can be grouped into 6 inter-class fertile, intra-
class sterile groups of 2 or more plants each (table 20), but these groups
are not necessarily independent. A is not B, C, D or F; B is not A, C, D,
E or F ; C is not A or B ; B is not A, B or F ; E is not B or F ; and F is not
A, B, D or E. Therefore, C may be D, E or F ; D may be C or E ; E may
be C or D; and F may be C. But since 2 of these alternatives are
mutually exclusive, it is definitely established that at least 4 of these
groups are independent of each other.

This matter is shown more clearly in table 21, where the 17 other
plants which exhibited no cross-sterility are also listed. From this table
by the process of elimination cited above it can be shown that 5 separate
inter-class fertile, intra-class sterile groups must exist. Since there are
16 plants unplaced because they have had only a few cross-matings made
upon them, however, it may be well to compare the number of classes
proved with the number to be expected from the percentage of sterility

Table 20

Family F. — Progeny of plant 34 of F1 of cross No. 2 X plant AA of N. Forgetiana
grouped in accordance with their behavior in inter-crosses.

No. cases fertile within group No. cases sterile within group

JTCU.

Group

No.

219

228

232

234

216

239

241

242

243

211

212

215

223

226

235

237

233

236

570

E. M. EAST AND T. B. PARK

THE BEHAVIOR OF SELF-STERILE PLANTS

571

found, on the theory of a distribution of individuals corresponding to the
frequencies of the coefficients of the binomial expansion. In family F
there are 128 fertile matings, of which 55 are reciprocals, a total of 146
(128X2 — no) fertile combinations. Likewise there are 23 sterile mat-
ings, of which 10 are reciprocals, a total of 26 (23 X 2 — 20) sterile com-
binations. This amounts to a cross-sterility of 15.1 percent. Adding
the 35 self-combinations to the steriles, gives 61 cases of sterility out of
207 combinations, — a percentage of 29.4. We should expect only about
5 intra-sterile classes in this population, therefore, unless a very broad
allowance is made for selection of matings that were sterile.

Table 22

Family G. — Record of cross-pollinations on progeny of plant 44 of Fx of
cross No. 3 X plant AA of N. Forgetiana outside of family G.

Sterile with

Ped. No.

Fertile with Ped. No.

Ped. No.

247

44 $ parent, Fx plant

249

44 $ parent, 35 13, fam. H, 467,., fam. I

258

34, Fx plant

278

219 fam. F, 374 fam. H, 467 fam. I

281

4050 fam. I

293

44 $ parent, Fx plant

308

34. F, plant

Family G. — Plant 44 of class C, cross No. 3 X plant A A of

N. Forgetiana

Family G was produced by mating plant 44 of class C, cross No. 3,
with the same plant of N. Forgetiana used in producing family F. In
all, 53 hothouse-grown plants had some work done upon them, although
in a few cases only one mating was made. These plants were studied dur-
ing a complete flowering season, but nearly all of the work was completed
before the period of decline in reproductive vigor so that only a few
cases of pseudo-fertility were found. 31 of the plants were selfed from
1 to 19 times with the production of a few seeds in one attempt at selfing
only (308). In 12 other matings there was some conflict in the results.
These were classified, as before, by recording as fertile those that gave
full capsules in two or more trials even though one trial failed, or by
recording as sterile those in which a majority of the trials failed even
though a portion of the pollinations did produce a few seeds (less than
15 percent of normal).

Table 22 records the crosses made when plants outside of family G

Genetics 2: N 1917

572

E. M. EAST AND J. B. PARK

Table 23

Family G. — Record of cross-pollinations on progeny of plant 44 of F1 of cross
Xo. 3 X plant AA of N. Forgetiana.

Ped. No.

Fertile with Ped. No.

Sterile with Ped. No.

247

248,, 250,, 253,, 256, 263, 276

248

•?47

24Q

247-, 2^0.. 2^6

2473, 249

253

251

270

2^2

247o, 240, 2S6, 2^7

1/ -J ^ > %J/

250, 255

2473, 256, 262

255

249, 262

252

256

253

257

255, 256.

258

247, 257, 269

2702

2^0

2583, 262

200

25I21 255, 262
266, 271

262

263

250, 253

258,

265

263, 266, 269, 276., 278

266

26^ 270. 28l

267

2^1

268

284

289

26Q

270, 28l

260

- / 1

274

9'c o-fi o7n c?Rt

-^O^j -Ou> ^/u>

273

275, 276, 28l

274.

269, 275,

270.

27^

2^8 270o 274. 278

276

270, 275, 28l

28 1 0, 284.

278

269, 274, 285, 289

27Q
—/ y

289'

281*

281

2763, 284

275, 278

283

270

276

2S4

28l, 285, 289

28;

?8(") 20 ^06 "}OQ

— °y> ^yo> o<JVJ> ouy

3043
289

286

278, 284

288

289

280

26^ 260 284 20^.,

286, 306

20O

^8 278 ^8q

■^/«J, _<jy

284

20 1

20^
yo

^74 28^.. 280 200

284

295

310

297

289

298

284c

303

2852, 304, 3062

304

293, 306o, 309, 310

307

305

312

311

306

284, 293, 304,, 309, 310

289

307

310

304

308

312

308

309

3044, 3103, 311

510

304, 308, 3090, 311, 312

311

3o84, 3092, 3io2

312

309, 3io,

3"i

THE BEHAVIOR OF SELF-STERILE PLANTS

573

were used as pollen parents. The n niatings tried were all successful.
3 back-crosses were made with plant 44, 2 with plant 34 of Fl9 2 with
plants of family H and 3 with plants of family I. It should be noted,
however, that of 7 crosses of plants of family F with pollen from indi-
viduals of family G, 4 were failures. On the other hand, G family pollen
was fertile on 3 plants of family H (table 27) and on 1 plant of family
I (table 30).

Table 23 shows the cross-matings made within family G. There were
126 successful matings, — 19 being pairs of reciprocals, — making 214
successful combinations. 29 matings were sterile, including 5 pairs of
reciprocals, — 48 combinations in all. 314 combinations have been made,
therefore, 100 being sterile (52 selfs + 48 crosses) and 214 fertile.
The probable sterility is thus 31.2 percent ±1.8 percent.

Table 24 shows 27 plants of this family grouped in accordance with

Table 24

Family G. — Progeny of plant 44 of cross No. 3 X plant AA of N. Forgetiana
grouped in accordance with their behavior in inter-crosses.

Ped.
No.

No. cases fertile within group

D I E

No. cases sterile within group

B I C
I

250
252
253
255

o I
o
0
o

258

260

263

270

274

- z

1 ! —
3 ! -

275
278

279
281
284

291

293_

268

286

288

289

306

3
1

- I

1 I -

1 I 2

0 I I

1 I —

1 I

o I
0

0 I —

285

304
307

305
311
312

Genetics 2: N 1917

574

E. M. EAST AND J. B. PARK

1*3

oi
oi

-J

C
"1

1/5
01

01
Is

•I N

01 01
01

On r*5

T 01 N M H M

ro ro 01 01 01

»-c 01 01
01 01 01

N5C - C

^ t if) IT, ir, ir. £

0J 01 01 01 01 01 01

\o^g n nkkx 5 o <5 c o o

01 N 01

oi oi "oi oi oi oi oi rr? co ?r;

THE BEHAVIOR OF SELF-STERILE PLANTS

575

their behavior in inter-crosses. There are 6 classes as tabled with a fre-
quency of 3, 5, 7, 5, 4, 3. There are 3 exceptions among the fertile
matings, 275 X 278, 281 X 284 and 305 X 312. Only one pollination
each was made on the first and third of these combinations, but the second
was made reciprocally — the last of the flowering season — one pollination
each way. There zvere no sterile exceptions.

Though 6 intra-sterile groups are tabled, there is definite proof of the
existence of only 3 classes. This is easily seen by referring to the table.
Classes C, D and E must be different, but the other 3 groups might have
proved to fall in with them had the proper crosses been made. Nor can
the existence of more than 4 intra-class sterile groups be proved even by
the complete table of inter-class fertility shown as table 25. By our
probability formula also the presumption is that there are but 4 or 5
classes, whether the distribution of individuals be according to the
coefficients of the binomial expansion or into classes of equal size.

Family H. — Plant 44 of class C, cross No. 3 X plant 10 of class A,

cross No. 2.

Family H was one of the 30 true F2 populations possible from com-
binations of the 6 different F± classes. It was produced by crossing plant
44 of class C, cross No. 3 with pollen from plant 10 of class A, cross
No. 2. 70 plants were grown in the greenhouse. Self-pollinations were
made on 33 of these individuals with the results listed in table 26. In
view of previous results it seemed hardly necessary to self every member
of the population. If this had been done a truly self-fertile plant might
have been discovered, of course, but it is exceedingly improbable. Of
those selfed, 5 did produce some seed, — the amounts being shown in the
table. These capsules were all produced at the very end of the flowering
season, except 1 with 8 seeds in it on plant 316. There is a chance that
these seeds were produced by foreign pollen, though it is hardly necessary
to "explain" such a rare exception to the general rule.

This family was studied through a long flowering season. Many mat-
ings were made, and the work completed before we were certain of the
effects of environment on self-sterility. For this reason some of the
matings made toward the end of the season were not tested as thoroughly
as should have been done. Further, no records of the number of seeds
were taken in the case of several capsules that were not full. Thus it is
altogether likely that several matings marked fertile were in reality
sterile. The maximum number of such errors, we should judge from a
careful examination of our records ought not to be over 10.

Genetics 2: N 1917

5 7 -

E. V ZAST AXD J. B. PARK

It is also probable that the usual experimental error of 4 failures per
hundred in actually fertile ma rings obtains in cases where a mating was
made but once and proved sterile. There were 63 such mating? in the
intra- family crosses, thereby making 3 such errors probable. The re-
maining combinations were judged by several mating? and by reciprocal
crosses, and are likely to be correct-
It is clear that the errors mentioned above are largely compensatory
when figuring :he 7 ercertages : i fertility : r sterility in :he mating? mile,
but they will stand revealed when endeavoring to group the individuals
in intra-sterile classes.

The record of back-crosses and crosses made with plants outside of

doss C) X plant 10 (F„ cross Xo. 2, doss A).

Ped. Xo.

x - « S. __

; .~-

Xo. of settings giving capsules with

1 3

c7~

" ~ *•

-X T

THE BEHAVIOR OF SELF-STERILE PLANTS

577

Table 27

Family H. — Record of cross-pollinations on progeny of plant 44 (F„ cross No. 2,
class C) X plant 10 (Fu cross Xo. 2, class A) outside family H.

'Fertile witJ Sterile with

r ea. ~\o.

parents

Fertile with Ped. Xo.

Sterile with Ped. No.

44. 10,

315

44=

317

44.

10,

311 fam. G

34s

3Jo

319

10,

311 iam. G

320

10.

321

IOt

311 tarn. G

322

10-

324

3-7
320

44;

442

10.

329

330

10,

331

44.

334

335

44=, 10

337

477 fam. J, 524 fam K

339

18 Fj

340

342

44-, 10 poor

I03

467 fam. I

349

350

10,.

351

354

44.

IOt

467 lam. I

34, F„ 401 fam. I

302

10, 8 seeds

10.

34, Fa

363

10,

365

366

367

368

lOj

467 fam. I

371

467 fam. I

373

44*. 10

374 j

378 ;

44;

401 fam. I, 467^ fam. I

379

44^

1 a.

381 44. JO

405 fam. I, 415 fam. I

382 44.-, iO..

384 44

385 kfe

10 1

family H, are shown in table 27, but they can be discussed best after
dealing with the intra-family matings.

Excluding selfings, 312 intra-family matings were made on 56 plants.
If we take all of these plants to be self -sterile — a reasonable assumption
even though a few of them were not self ed— 448 combinations out of a
possible 3136 were attempted. The figure 448 is the sum 153 X 2 =
306 fertile matings, minus 100, the number of fertile reciprocals,
plus 159 X 2 = 318 sterile matings. minus 132, the number of

Genetics 2: N 1917

5/8

E. M. EAST AND J. B. PARK

Table 28

Family H. — Record of cross-pollinations on progeny of plant 44 (F„ cross No. 2,

aass L- y /\ piani 10 cross

No. 2, class A).

1 C O . -NO.

-Tciine \\ lin nu. r\o.

oicriic \\ lin x eu. ->o.

*
0

315

316, 317, 3182

316

320, 324

3I72, 3l8, 321, 331

317

320

3152, 318, 321, 3272, 3283

3l8

320,., 328

3153, 316, 3175, 321,, 324
315, 3l 6, 3 1 75, 3 542, 381

319

347

320

317, 318, 321, 322, 3240, 3280, 354,, 381

342, 351

321

3^o„ 3350, 381

3152, 3162, 322,, 328,

322

320,, 325, 381,

321, 324, 328, 329

324

320.,, 325, 342, 3672, 3792

322, 327, 328, 331, 3542

325

317, 3222, 324, 329

327

351

330, 3363, 3372, 3402, 345

328

335, 337

317, 324, 327, 329, 342

329

325, 347

3243, 327, 328

330

335

327, 334, 374

331

3353

3i63, 327, 3282, 329, 330, 336

333

33C>2

334

335

33i2, 337, 374

335

321, 324, 327, 328, 329, 331

320, 336, 38 12

330

328, 35I0

3272, 33i, 337,, 342, 345
327, 336, 3403

337

339

338

327, 337, 341, 342

339

318, 327, 336, 3372, 340, 342..

338

340

327, 3372, 3423, 345, 346

341

3272, 331, 3372, 340„ 3424

342

3472, 351, 3732, 381

336, 340, 345, 3542, 37i, 374

345

327, 3372, 342

347

3373, 340, 342, 3492, 354

348

342, 347

35i

349

35i2

3422

350

38i2

334, 337, 340, 349, 3542, 359, 363

351

3492, 350, 353, 3542, 362

320, 368, 381,

352

327, 342, 3493

348

353

35i2

3542, 3623

354

35i2, 371

317., 337, 350, 3632, 3743

355

342,, 35i, 3542, 381

358

3544, 362

359

347, 351, 3552, 366,

342, 354., 362, 37ia

360

362, 363.-,

362

368, 381,

340, 354, 358, 363, 365

363

35i, 366, 3682

350,, 354, 365

365

355, 368

354, 359, 362, 363, 374

366

351, 3542, 360, 363, 365, 368,

378

367

3542, 3703, 37i

368

354, 363, 37i, 374

320,, 367, 381,

370

3672, 368s, 373, 378
366, 368, 38i2

37i, 372, 374
3652, 374t

371

372

367.,, 368, 381

37i2, 374

373

3542, 367, 370,, 3743, 385, 371,

320., 368, 381

37-1

373, 378, 381

37U

378

373, 38i;, 383

3792

379

3744, 3812, 383

3542, 3733

351, 367, 368,, 373

367, 368, 3732, 381

381

3i7s, 340, 341, 3422, 3544, 3742, 378

382

3542, 374, 379

383

3784, 384

367

384

381

3782

385

3812

378

Table 29

Family H.*— Progeny of plant 44 (FJ} cross No. 2, class C) X plant 10 (Fu cross
No. 2, class A) grouped in accordance with their behavior in inter-crosses.

IN O.

cases

fertile within

group

IN 0.

cases

sterile within

group

| Ped.

1 1

I 1

Group

1 No.

1 B

1 c

1 D 1

1 I

Ind.

1 B

1 c

1 D 1

1 L

Ind.

1 315

1 —

j —

| |

1 — 1

—

1 —

j —

1 |
j —

—

1 316

| —

I — 1

—

1 0

j —

j — i

—

1 317

1 2

| —

1 — 1

1 0

j —

1 — l

1 3i8

| —

1 1 1

—

1 0

1 —

1 0 j

—

1 319

| —

I — 1

| —

— 1

1 321

| —

1 — 1

—

1 0

1 —

1 — !

—

1 322

| —

1 — 1

| —

j — |

324

| 1

l — 1

1 — 1

1 327

I —

1 2 |

1 —

0 1

1 328

| —

1 — 1

—

| —

1 — 1

—

329

I —

1 — 1

1 —

1 — 1

1 330

| —

1 — 1

—

I —

—

1 33i

1 —

1 — 1

j —

I — 1

334

I —

j — |

—

1 —

— 1

—

1 336

1 —

1 |

I —

0 |

337

—

1 —

2 l

—

I —

1 0 1

340

L —

1 1 1

i —

0 1

342

I —

2 1

I —

0 |

345

—

1 —

I — 1

—

| —

— 1

—

346

349

1 —

1 — 1

1 —

— j

350

1 —

— 1

—

| —

—

353

i —

1 — 1

—

j —

I — j

—

354

1 0

— I

| 1

— |

358

— |

359

1 —

—

1 —

— I

360

—

1 —

—

I —

— 1

362

1 —

— 1

—

I —

— 1

— '

363

1 —

—

I —

—

365

1 —

—

I —

— I

370

— I

—

i 0

— I

—

371

1 —

— I

1 —

— 1

372

I —

— 1

—

I — .

— j

—

374

— 1

—

1 0

— !

—

320

1 —

— I

—

I —

—

335

—

— I

—

j —

— 1

—

348

! —

— |

I —

— 1

351

1 —

— 1

I —

— 1

352

j —

— j

—

I —

— !

—

367

1 0

— 1

—

| 1

— I

—

368

! —

— 1

1 —

— 1

373

! 2

—

— 1

—

381

! 4

— 1

2 j

—

382

| 1

1 °

...

383

I 3

—

I 0

378

—

1 3

— 1

—

379*

1 0

— I

—

| 1

— 1

—

384

—

1 0

—

— 1

—

385

—

1 °

—

j 1

—

338

—

0 I

—

1 |

339

0 1

1 j

325

333

. 1

Ind.

341

347

355 1

366 1

Probably not really a member of group C.

E. M. EAST AND J. B. PARK

sterile reciprocal, plus the 56 self-combinations. The probable total
sterility in the population is 54.0 percent — 1.4 percent, therefore, which
makes it unlikely that more than 3 or 4 intra-sterile classes are present.
These matings are shown in table 28.

The individuals are grouped with reference to their behavior in inter-
crosses in table 29. This table appears to reveal 4 classes containing 34.
11. 4 and 2 plants, respectively, in addition to 6 indeterminate individuals.
Let us see what it really shows.

In the first place, there are 8 exceptions — fertility where there should
be sterility — in the fertility columns. They are as follows, each mating
being made but once.

Class A 316 X 324

" " 318X328

" " 324 X 342

" " 328X337

" " 336X3^

" " 354 X 37i
Class B 320 X 381

373 X 367

There are also 6 exceptions — sterility where there should be fertility —
in the sterile columns, and here one mating ( Xo. 4) was made twice and
one mating (No. 6) three times. These exceptions are as follows:

1. B X A 319 X 381

2. B X A 320 X 342

3. B X A 335 X 336

4. C X A 379 X 354

5. B X C 367 X 378

6. C X B 379 X 373

These exceptions are no more than were to have been expected from
the predictions made above from a priori calculations. Of the fertile
exceptions, at least 5 were made at the last of the season. Xo data re-
garding percentage of seed obtained to seed expected in full capsules
were recorded, unfortunately, but it is probable from our other experi-
ences that the majority of them produced only partly rilled capsules, and
would have proved sterile had they been made earlier. The sterile ex-
ceptions 379 X 354 and 379 X 373- made twice and thrice respectively
are of little consequence because 379 falls into class C only through the
single sterile mating 378 X 379 (made twice 1. Thus we could just as

THE BEHAVIOR OF SELF-STERILE PLANTS

58l

reasonably call 379 an indeterminate, — that is a plant fertile in all
combinations tried, — and have but the sterile exception 378 X 379 for
which to account. It could not go into groups A or B, though sterile
with one plant of each of those groups because it also was fertile with 2
members of group A and with 3 members of group B.

This interpretation maybe made either way without affecting the chief
point the table was designed to show. Xo indeterminate individual and
neither plant of the very uncertain class D, which was based on the
single case of sterility 339 X 338, were crossed with plants of class C.
Therefore the 3 classes A, B and C are the only ones for which we can
claim independence.

A meaning can now be given to the results of the back-crosses which .
were listed in table 27. 38 plants were crossed with pollen from one or
both parents. Out of the 23 plants crossed with Xo. 44 just 1 was
sterile, — a single pollination of 366 X 44- It is possible that this mat-
ing also might have shown fertility if tested further, but it may show
that 366 is the only plant among those tested that belongs to the same
intra-sterile class as 44.

Plant 10 was used as pollen parent with 29 plants, of which 10
produced some seed. Plant 342 produced a few seeds which seemed
to be parthenocarpic out of 4 tests, and plant 362 yielded 8 seeds in 1
of the 4 tests made. Therefore we have no hesitancy in classifying them
as sterile. Plant 314, which was fertile to plant 10 pollen, was dis-
carded early and is not classified in table 29. For this reason it may
be left out of consideration. Plants 335 and 367 were fertile in one
pollination each, and sterile in one pollination each. Since they gave
full capsules in each of the successful pollinations, however, let
us record them as fertile. Now what is the result? Out of 20
sterile matings 18 arc with plants belonging to class A. The first
exception is with the plant 379 which behaved so irregularly —
as shown by table 29 — that it is just as likely to be a member
of class A as class C. The second exception is a single pollination with
plant 385 of class C. Fertility is shown in 7 cases, all of which are with
class B. Furthermore, the 3 sterile matings made with pollen from plant
34, a member of the same F1 class as plant 10, are with plants of class A
of family H. And the 1 sterile mating made with plant 18, a plant of F-,^
class B, is with plant 339, a member of class D of family H. Therefore,
it seems unquestionable that Plant 44 (and thus class A of F±) belongs
to the class A of family H.

Genetics 2: X 1917

E. M. EAST AND J. B. PARK

Sterile with Ped. No.

X *

i 5j |

J J 5J j j I I

Fertile with Ped. No.

351 fam. H

337 fam. H, 477 fam. J, 524 fam. K

374 fam. H
490 fam. J

377s fam. H
381 fam. H

474 fam. J, 475 fam. J
320 fam. H, 381 fam H

278,, fam. G, 320 fam. H, 381 fam. H, 489 fam. J

Sterile with
parents

Fertile with
parents

$ i i

Ped. No.
9

THE BEHAVIOR OF SELF-STERILE PLANTS

583

Table 31

Family I. — Record of cross-pollinations on progeny of plant 44 (F„ cross No. 3,
class C) X plant 34 (Fu cross No. 2, class A).

Ped. No.
9

Fertile with Ped. No.
3

Sterile with Ped. No.

451
455
456
457
458
460
463
464
465
467
468
470.

395, 396

390, 3962, 398, 400, 405
! 413

I 396, 400
401

398„ 413=
413

420

396, 401, 405s, 408, 415, 418, 420
42S2

413, 425
43i

431, 425

4253, 433
4262, 433s

401,, 426,, 433, 439, 455
43 13

405, 426, 431, 439, 4403, 451, 458

433

439, 446
431, 456, 458,

442

444

463,
413

456,, 457^, 465
456, 463

396,
396,

395, 4443, 468,
400, 405.

400,, 401 4, 405,, 415
413

396,

4052, 4i5a, 426, 467

40 1 3, 408, 415, 467,

4052, 409, 415, 426
4083

4I5«

401., 405., 412, 414, 420., 426, 458
412,

40i2, 405, 415., 425u, 4262, 458

467

43i

405, 420,, 440,, 445, 458, 464,
43i, 458
430, 440

440, 444

415, 426., 4392, 4442, 45i2, 457
4442, 451
4403, 451

4302, 431
4212

442, 4443, 455,, 467

415, 426., 440, 45 1*, 45°\>

4552, 457s, 4582, 4672

455, 4564, 458, 467

405, 420, 451, 455*. 4562, 467

392,, 4683

4572

455, 456,
4654

Family I. — Plant 44 of class C, cross No. 3 X plant 34 of class A, cross

No. 2

Family I was produced from seed obtained by pollinating plant 44 of
class C, cross No. 3, with pollen of plant 34 of class A, cross No. 2. It
is therefore a test of the similarity of constitution of plants of class A

Genetics 2: N 1917

5^4

E. M. EAST AND J. B. PARK

of Flf since plant 44 was crossed first with plant 10 of class A to produce
family H and then with plant 34 of class A to produce family I.

83 greenhouse plants were grown; but the task of manipulating that
number proved too great and very nearly one-half of them were dis-
carded after several weeks of work, permitting our efforts to be more
concentrated. We have not thought it necessary to report any of the
pollinations made on the rejects.

Of the plants remaining, 25 were selfed from 1 to 6 times between the
first and the middle of the reproductive period without obtaining a single
seed. Somewhat contrary to what might have been expected, 6 of these
same plants were again selfed several times during the latter part of the
season with the same result. This does not prove that no seed could have
been obtained at that time if further pollinations had been made, how-
ever, as a few seeds were produced in a part of the pollinations of 22
cross-matings made during the waning of the flowering period, where
continued pollinations made before had left no doubt as to the sterility
of the combination. In 9 other matings, 1 pollination each produced no
capsule, but in each case other matings — usually several — giving full
capsules, proved them to be fertile. They were therefore so recorded.

Table 30 shows the record of back-crosses with pollen of the parents,
and also the crosses made with plants outside of the family. It will be
discussed after making the usual classification.

The inter-crosses in this family are shown in table 31. About one-
sixth of the 2025 different combinations possible with 45 plants were
accomplished. The table shows 61 fertile and 97 sterile matings, in-
cluding 13 pairs of fertile reciprocals and 20 pairs of sterile reciprocals.
The total number of different cross-combinations, therefore, is 250, made
up of 96 fertile and 154 sterile combinations. Adding the 45 self-com-
binations, we have 199 steriles out of a total of 295 combinations. The
probable sterility in the population is thus 67.5 percent ±1.8 percent,
and we should scarcely expect more than 3 or at most 4 intra-sterile
classes even if a Mendelian dominant type (3 + 1) of distribution in
the classes be assumed.

The grouping actually obtained is set forth in table 32. Three classes
containing 34, 4 and 2 individuals, respectively, and 5 unplaced plants,
appear. There are 6 fertile exceptions :

400 X 401
412 X 420
442 X 439
444 X 456
444 X 458
465 X 456

THE BEHAVIOR OF SELF-STERILE PLANTS

Table 32

Family I.— Progeny of plant 44 (F1} cross No. 3, class C) X plant 34 C^u cross No- 2>
class A) grouped in accordance with their behavior in inter-crosses.

No. cases

fertile

within

VV 1 Li i 111

No. cases sterile

within

group

Perl

Oroup

IN O.

Ind.

TnH
.lull.

390

— 1

391

392

—

395

396

400

401

405

4O0

409

412

—

— |

—

414

415

4IO

420

421

—

420

—

1 2

—

439

440

! 7

1 1

442

—

444

—

445

448

j 1

451

455

456

457

458

460

464

465

1 Af\-r

2
8

—

I 468

0
0

-
—

—

I 470

—

1 425

1 -

' 430

1 431

446

1 398

1 413

1 387

Ind.

394

1 I

432

1 -

433

1 -

1 0

1 463

1 -

Genetics 2: N 1917

586

E. M. EAST AND J. B. PARK

S
o

eg £
ro rr;

<N in rt Tf

^ CO co

ro co

OS
S

C M_| r-'

o S

00 00 SB

CO CO co

m" m j-t Q cf

u> 00 00 M <N

CO CO CO -O co

04 VO «
On O O O C

co co ^t- Tf

THE BEHAVIOR OF SELF-STERILE PLANTS

587

Four of these matings were made but once, 1 was made twice and 1
was made reciprocally. The last 2 and 1 other were end-season matings,
the others were mid-season matings. There are 2 sterile exceptions,
431 X 440 and 430 X 458, each tried but once. The number of combi-
nations that form the basis of our grouping is so large, that there is
little danger in accepting the classification as given, however, since these
errors might have crept in in various other ways, as has been shown
before. But it should be mentioned that plant 430 falls just as readily
into group A as it does into group B.

The evidence in this table does not support the idea of more than 3
classes. A and B are well established. But C may be B, since neither
members of the class were crossed with any B individuals. Of the
indeterminates, 387, 394 and 463 may be B and 432 may be A. The sole
positive evidence of a third class, therefore, rests upon plant 433, which
is not A (6 matings in evidence) nor B (2 matings in evidence).

Let us now consider the back-crosses shown in table 30. Every
cross made with the pollen of plant 44, 29 in number, was fertile. On the
other hand 15 back-crosses with pollen from plant 34 were sterile,
though an average of over 3 pollinations per plant was made. Seed was
obtained in only 1 instance: 4 pollinations were made on plant 412,
and 2 made late in the season gave some seeds. The interesting feature
in these 15 sterile matings is that 14 of them were made on plants of
class A, and the fifteenth on plant 430, which, though tabled in class B
may just as readily be placed in class A.

But 3 plants were fertile to pollen of plant 34, — plants 425 and 431
of class B and plant 398 of class C.

A single mating of plant 10 on plant 401 of class A was sterile. Since
plant 10 and plant 34 belong to the same class of the Fx generation, this
mating may be compared with the 3 sterile matings of class A plants of
family H with pollen from plant 34.

Note then the similarity between families H and I. Each has 3
independent inter-fertile, intra-sterile groups with almost the same dis-
tribution of individuals within the classes ; each behaves similarly in back-
crosses. With the exception of a single unclassified plant of family H,
all of the plants tested of both families were fertile with plant 44 of class
C of the F± generation, the female parent of both. With regard to
plants 10 and 34, the male parents of families H and I respectively, both
of which belonged to class A of the F1 generation, each was sterile with
class A plants of both families and each fertile with other plants of their
respective families. The conclusion is unavoidable, therefore, that class

Genetics 2: N 1917

588

E. M. EAST AND J. B. PARK

A of the F1 generation, class A of family H, and class A of family I,
are identical.

This is not the only evidence that can be brought forward in favor
of the similarity of these two families. A sufficient number of crosses
(table 33) was made between the two populations to prove that class A
of family H and class A of family I are the same. Ten members of
class A of family I were crossed with plants from family H. Three pairs
of reciprocals were made with like results for each pair. Counting these
pairs as but 1 mating each, members of class A of family I were crossed
14 different ways with members of class A of family H. Of these mat-
ings 11 were sterile, and 3 fertile. But of the fertile matings, 2 were
with 337 and did not give full capsules. These same class A plants
of family I were also mated 9 times with members of class B of family
H, and all matings were fertile. Bearing these results in mind, the single
sterile mating of 460, — family I, class A, — with 380, — unplaced member
of family H, — is pretty good evidence for placing 380 in class A of
family H. Likewise, the sterility between 431 and 377 is evidence that
377 of family H is not a member of that family's class A, a conclusion
supported by its fertility with unplaced 432 of family I. The remaining
cross, plant 413 of class C of family I with plant 374 of class A of family
H, was fertile.

We do not believe it rash to assert that this makes a complete case.
There can be no doubt that families H and I are practically duplicates of
each other. In this instance, then, two plants belonging to a single class
in which all of the individuals were cross-sterile with each other, zvhen
crossed with the same individual have produced populations as similar
to each other in their behavior in crossing as if they were samples of the
same population.

This does not prove that all members of an intra-sterile class crossed
with the same individual would produce identical populations. No such
claim is made. It does indicate very strongly, however, that in this
particular case, these 2 plants of the Fx class A (10 and 34) are identical
in that part of their constitution which affects self- and cross-sterility.
The criticism may be offered that these results show merely a kind of
dominance exhibited by plant 44, but if this be true, it is a dominance
of a strikingly perfect kind.

Family J.. — Plant 52 of class B, cross No. 3 X plant 23 of class A, cross

No. 2

As has just been shown, F± plants of class C when crossed with their

THE BEHAVIOR OF SELF-STERILE PLANTS

589

Table 34

Family J. — Record of cross-pollinations on progeny of plant 52 (F„ cross No. 3,
class B) X plant 23 (F1} cross No. 2, class A) outside family J.

OLCIllC Willi

Ped. No.

Fertile with Ped. No.

Ped. No.

475

524 fam. K

477

377 fam. H, 467 fam. I

487

512 fam. K

489

467, fam. I

490

421 fam. I

495

18 F1

499

i82 F1

502

18 F1( 512 fam. K

sisters of class A give populations having a high percentage of cross-
sterility and by the same token a small number — 2 or 3 — of intra-sterile
groups. Family J tests the behavior of an Fx plant of class B with a
class A sister.

30 plants of this family were grown in the greenhouse, 6 dying or being
discarded. They were all selfed from 1 to 12 times with no production
of seed except on plants 473 and 489. These 2 individuals produced
seed the latter part of the flowering season. No. 473 was selfed 7 times
at various periods. The first 2 pollinations yielded no seed, the third
and fourth a few seeds, and the last 3 half-filled capsules. No. 489 was
selfed 9 times. The first 3 were failures; the remainder induced cap-
sules, the last 3 pollinations producing a full quota of seed. *

Only 1 back-cross was made. No. 474 was fertile with No. 52.

The few other crosses made with plants outside the family are recorded
in table 34. All were successful. It should be noted that 3 of these
successes were with plant 18, another member of class B of the F1
generation.

As usual only a comparatively few of the 576 combinations possible
between 24 plants were made. The record of cross-pollinations listed
in table 35 are sufficient, however, to show the striking difference in
percentage of cross-sterility between this family and the 2 families just
described. There are 65 fertile matings including 14 pairs of fertile
reciprocals, making 102 fertile combinations in all. Since there are no
sterile reciprocals, the 13 sterile matings are equivalent to 26 sterile
combinations. Adding the 24 self-combinations, gives a ratio of sterility
to total combinations of 50 : 152. The probable sterility in this family

Genetics 2: N 1917

590 E. M. EAST AND J. B. PARK

is therefore 32.9 percent zb 2.6 percent, which leads us to expect about
5 intra-sterile groups.

The grouping made possible by the sterile matings is shown in table
36. There are no exceptions. Each individual in every group shows
perfect inter-class fertility and intra-class sterility as far as they were

Table 35

Family J. — Record of cross-pollinations on progeny of plant 52 (Fu cross No. 3,
class B) X plant 23 (F1} cross No. 2, class A).

Stprile with

1 cQ. IN O.

jrcriiic wiiii r cu. i\u.

Ped. No.

473

474, 475e, 4853

40O

474

475s, 48o2, 482, 485
474, 477, 480, 482, 4854

475

477

473, 4753, 482, 485c

478

484, 485

480

474, 4753, 482, 486, 4872, 491

482

474, 484, 4852

484

474, 480, 482, 4870

4853

485

474, 475, 4825, 492

484

486

485, 4923

474, 495

487

474, 482, 484, 486, 492, 499

488

482, 487,

484

489

477, 492

486

490

489

491

480

484

492

484, 4872, 493, 4952

493

502

494

486, 502

495

499

496

492

499

5024, 503

500

486, 493, 499, 502, 503

502

499

495

503

4992, 500, 502,

tested. Apparently there are 4 classes containing 7, 4, 2 and 2 indi-
viduals, respectively, together with 9 plants which showed no cross-
sterility and are unplaced.

Table 37 shows the evidence for independence between these groups
more clearly. A, B and C or D must be independent, but C and D may
belong to one class since they were not crossed together. In addition 475,
477 and 482 are independent of each other and of A, B and C. Thus there
are apparently 6 independent classes with frequencies of 7, 4, 2, 1, 1 and
1, these frequencies being subject to change of course given the data
necessary to fit the remaining individuals into their proper niches.
Before accepting this classification at its face value, however, we ought

THE BEHAVIOR OF SELF-STERILE PLANTS

59 1

Table 36

Family J. — Progeny of plant 52 (Fu cross No. 3, class B) X plant 23 {V \, cross No. 2,
class A) grouped in accordance with their behavior in inter-crosses.

Group

Ind.

Ped.
No.

474
486
489
493
494
495
I 502

484
485

No. cases fertile within group

491
473
480
492
496

475 ! 1
477
478
482
487
490
499
500
503

I I
C I D I Ind.

No. cases sterile within group

Ind.

Table 37

Family J— Progeny of plant 52 (Flf cross No. 3, class B) X plant 23 (Fu cross No. 2,
class A) grouped to show inter-class fertility.

A B C D 475 477 478 482 487 490 499 500 503

4 1

2 1

3 1

2 1

2 3

475

477

478

482

487

400

499

1 1

500

503

Genetics 2: N 1917

592

E. M. EAST AND J. B. PARK

to see whether the independence of any of the 3 single plants is based
upon a single pollination. Plants 475 and 477 were fertile reciprocally,
4 pollinations being made in all, but plants 475 and 482, and plants 477
and 482 were crossed but once. This is also true of the basis of inde-
pendence between 477 and A, 477 and C, and 482 and C. It depends on
1 pollination in each case. . ,

For these reasons it is hardly likely that more than 6 independent
classes exist in this population, and the chances are perhaps even that
there are only 5. Nevertheless, family J unquestionably contains 2 or 3
more intra-sterile classes than family H or family I.

Table 38

Family K. — Record of cross-pollinations on progeny of plant 52 (F1} cross No. 3,
class B) X plant 44 {F1} cross No. 3, class C).

Fertile with

Sterile with

Ped. No.

Fertile with

outside

within

Sterile with

outside

within

Ped. No.

parents

family

parents

family

505

508

507

515

508

505

509

5ii

44, 52,

508, 509

5i23

512

520

515

524

5i7

524

520

443

SI22

521

44, 52

512

524

44, 52

58 N. alata

525

5202

527

5052, 509

528

523

58, N. alata

Family K. — Plant 52 of class B, cross No. 3 X plant 44 of class C,

cross No. 3

Very little was done upon family K, as table 38 shows, though this
family resulting from crossing a plant of class B (52) with our much
used plant 44 of class C, might have proved very interesting. The plants
would possibly all have shown fertility in back-crosses with 52, while
only a part would have proved fertile with the other parent. This is the
indication of the few matings made. There were 6 cases of fertility and
none of sterility with No. 52, and 3 cases of fertility and 4 of sterility
with 44.

2 plants were crossed with N. alata plant No. 58; both were successful.
These were the only crosses made outside of the family with K plants

THE BEHAVIOR OF SELF-STERILE PLANTS

593

used as females. But K pollen was fertile on several plants of other
families; viz., 524 on 337 of family H, on 408 of family I, and on 475
of family J; 512 on 487 and 502 of family J.

The 14 matings made within the family, including as they do 2 pairs
of sterile reciprocals, are hardly a sufficient basis for even a guess as to
the amount of cross-sterility present potentially. We can only say that
the number of intra-sterile classes would not have been large, the per-
centage of sterility probably lying between 35 and 50.

Argument on cross No. 2 and cross No. 5
If further evidence of the beautiful regularity with which plants be-
longing to the same intra-sterile class behave in crosses be desired, it is
found in the crosses between families cited in tables 13, 18, 22, 27, 30,
33 and 34.

Plants 152 and 153 of class A, family D, were both sterile with family
E pollen which is presumably of one kind. The mating 152 D X 204 E
was even made reciprocally. Plants 167 and 171 of family D, which
were discarded after a few matings had been made and were therefore
undetermined as to class, were fertile to pollen of family E.

In family F, plants 216, 239 and 243, all of class B were each sterile
with the pollen from the unplaced plant 247 of family G. Plant 244,
an unplaced plant of family F was fertile with the pollen of 247,
however. On the other hand, plants 216 and 241 of family F, class B
were fertile with the pollen of plants 278 of class C, family G and 250
of class A, family G, respectively. Plant 278 of class C, family G, was
also fertile with the pollen of plant 219 of class A, family F, although
plant 219 was sterile with the pollen of plant 250 of class A, family G.

If we may say that sterility shows likeness of constitution and fertility
unlikeness of constitution, these results show: (1) that class A of family
F and class A of family G are alike; (2) that class A of family F and
class C of family G are unlike; (3) that class B of family F and classes
A and C of family G are unlike, as they should be since classes A of both
families are alike; and (4) that the unplaced plant 247 of family G
belongs in with class B of family F, as might very well be the case.

In the remaining matings between plants belonging to different families
there was no sterility, except among those matings between families H
and I already discussed. They are none the less interesting, however,
because they show that once fertility has been found between classes
belonging to different families, all matings between plants belonging to
these classes will prove fertile barring experimental error.

Genetics 2: N 1917

594

E. M. EAST AXD J. B. PARK

In family G. unplaced plant 249 was fertile with plant 351 of class B.
family H and with plant 467 of class A. family I. Plant 278 of class C
was fertile with pollen from plant 374 of class A. family H. Plants 278
and 281. both members of class A. were also fertile with plants 467 and
405 of class A. family I. respectively. Thus 2 combinations between the
classes A of families H and I proved to be fertile.

Likewise. 3 plants of class A. family H. 317. 319 and 321, proved to be
fertile with the pollen of plant 311 of class F. family G. Another plant
of class A. 337, also proved to be fertile with the unplaced plants 477 of
family J and 524 of family K.

Fertile matings were made as follows between 4 plants of class A.
family I. and plants of families G and J: 408 with 477, of family J
unplaced: 421 with 490, of family J unplaced: 448 with 474 of family
J. class A. and with 475 of family J unplaced; 467 with 278 of family
G. class C. and with 489 family T. class A.

Fertile matings were also made with the pollen of 3 family I. class A
plants on plants of family J. Pollen of 467 was fertile on 477 unplaced
and on 489. class A of family T. and pollen of 421 was fertile on 490
unplaced of family T.

Thus plants of class A of family I were fertile once with a plant of
class C. family G. 4 times including a reciprocal with unplaced plants of
family T. and 3 times including a reciprocal with plants of class A.
family I.

In these matings between families, then, not a single one militates
against our conception of inter-fertile, intra-sterile groups. Wc believe,
therefore, that the fundamental basis of this grouping is established be-
yond doubt, and that the actual groups as submitted in the foregoing
pages are sufficiently exact to be made the foundation of a theoretical
interpretation of the behavior of self-sterile plants among themselves.

Undoubtedly there will come the critic who will say we have been at
some pains to make out a case for the presence of inter-fertile, intra-
sterile classes in this family. He will point out that some of the excep-
tions among the matings may not have been due to experimental errors
and hence must have subtle meanings other than those given, that our
phrase "pseudo- fertility due to environment" veils the real facts. Let
us forestall him.

Of course some of the matings which form exceptions to the rule of
inter- fertile, intra-sterile classes may be the effect of an unknown bio-
logical cause; certainly factors other than environmental may be the

THE BEHAVIOR OF SELF-STERILE PLANTS

595

basis of a portion of the change from sterility to partial fertility in
certain matings as the flowering season wanes.

The first thing to establish, however, was a broad general rule for
the behavior of self-sterile populations. This has been done by the
work on these 2 crosses. The members of any population of the self-
sterile species under consideration fall naturally into a relatively small
number of groups, each individual being cross-sterile reciprocally with
every member of the same group and cross-fertile reciprocally with every
other individual. The sum total of the exceptions to this rule is well
within the limits of experimental error, even though the question is one
in which every bit of evidence, like pieces of a jig-saw puzzle, must fit,
if a solution is to be obtained. The exceptions to the rule, in fact are of
another order of magnitude than the confirmations. If, therefore, true
exceptions do occur, they are so rare that the usefulness of the rule is
not in the least impaired. Other general matters must be settled before
it is even desirable to endeavor to inquire into them.

Lest there be some difficulty in carrying in mind the essential facts
regarding the grouping of the plants of this series, let us summarize them
here.

The two self-sterile species N. Forgetiana and N. alata were crossed
reciprocally. The progeny of these two crosses behaved so similarly that
collectively the 53 individuals studied could be placed in 6 intra-sterile
classes 5 of which were proved to be independent. The remaining
questionable group consisted of one plant.

From this population 8 families were raised which were character-
ized as follows :

D = iV. alata plant 53 X plant 44, class C; probably consisted of 4-6

classes, 3 being established.
K — N. alata plant 58 X plant 44, class C; probably consisted of 1 class.
F = plant 34, class A X plant AA, N. Forgetiana; probably consisted of

5-6 classes, 4 being established.
G — plant 44, class C X plant AA, N. Forgetiana; probably consisted

of 4-6 classes, 3 being established.
H = plant 44, class C X plant 10, class A; probably consisted of 3

classes, 3 being established.
I = plant 44, class C X plant 34, class A; probably consisted of 3

classes, 3 being established.
J = plant 52, class B X plant 23, class A; probably consisted of 5-6

classes, 5-6 being established.

Genetics 2: N 1917

596

E. M. EAST AND J. B. PARK

K = plant 52, class B X plant 44, class A; probably consisted of 4-6
classes.

It was also determined that class A of the F1 generation, class A of
family H, and class A of family I are identical.

Cross No. 4. N. commiitata X Ar. Forgetiana {self -sterile X self -sterile)

The race used here with the pollen of N. Forgetiana was received
from Italy under the name N. commiitata Fisch. and Meyer. It is the
plant called N. Langsdorffii Weinm. variety grandiflora by Comes
(1899). Of it he says: "Elle est connue depuis 1835 dans les jardins
europeens, mais on en ignorait la patrie." It has been duplicated in our
experiments by crosses between N. alata and N. Langsdorffii. It is an
additional argument in favor of such an origin, that it is self-sterile,
since N. Langsdorffii is always self-fertile. When crossed with N.
Langsdorffii the F± plants are self-fertile. The behavior of this race
when crossed with N. Forgetiana is interesting, therefore, whether it be
a true wild species or was produced by hybridization. In the first case,
a new species cross is reported, in the second case, a self-sterile race
extracted from a cross between a truly self-fertile species and a self-
sterile species, is crossed again with a different self-sterile species.

The.Fj plants were highly fertile, in the sense that 90-100 percent of
the pollen was normal in nearly every plant, and that "proper" combina-
tions yielded full capsules.

A rather small number, 12, field-grown F1 plants were used in our
experiments. These were selfed from 3-10 times, an average of over
4 pollinations per plant. 1 1 were completely self-sterile, yielding not a
single seed. Plant No. 3, however, produced 4 good capsules out of 4
pollinations. This plant behaved like a real self-fertile. Crossed as a
female with each of the other 11 individuals it was fertile; crossed as a
male with all but plants 5 and 11, it was also fertile. Further, it was
fertile as a female with N. Forgetiana. The meaning of this behavior
has not been determined conclusively. Two interpretations are possible.
Owing either to its hybrid origin (self-fertile X self-sterile) or to a
recent introduction of N. Langsdorffii "blood," the race is a mixture of
self-fertile and self-sterile plants; or, by reason of its having been grown
near N. Langsdorffii the preceding generation, the seed from which this
plant came was produced by a stray pollen grain of that species. The
second interpretation seems more probable, since we have corroborated

THE BEHAVIOR OF SELF-STERILE PLANTS

Table 39

Result of mating s on F, plants of cross No. 4, N. commutata X N. Forgetiana.

Ped. No.
$

Fertile with

Ped. No.

Sterile

with Ped. No.
S

2o,

4r„

i3, 25, 43, 5s, 63, 73, J

13,

i5> Q4, I04, II,, I23

3a 52, 63

82, I04

64,

14, 3 , 44, 84, io4, I24

74, 94, Hi

34, 4o, I24

III

1 4, 24, 35, 4s

94,

10.,

14, 22> 3, 44, 124

72,

1 3, 24, 3s, 4s, I24

I4, 24, 85, I06

5s,

33, 73, ii3

i3,

22, 4s

Table 40

Plants of F1 generation, cross No. 4, grouped in accordance with
their behavior in inter-crosses.

No. cases fertile

No. cases sterile

with

in group

within group

Ped.

Group

No.

Compton's conclusion that true self-fertility is completely dominant over
self-sterility.16

In this family 70 cross-matings were made, of which 48 were fertile
and 22 sterile. These matings were each made more than once, as is
shown by the subscripts in table 39. There were 22 pairs of fertile
reciprocals and 4 pairs of sterile reciprocals. By multiplying the sterile
and the fertile matings each by 2 and subtracting in each case the proper

16 The relation between self-fertile and self -sterile plants is to be made the subject
of a later paper.

Genetics 2: N 1917

598 E. M. EAST AND J. B. PARK

Table 41

Intercrosses between progeny of pseudo self-fertile N. alata plant used in cross No. 2.

Compare with table I.

T~> _ J XT ~

Jreu. JNo.

Plants with which fertile

± lantb

wiin wnicn sterile

as 6*

as 9

as 6*

as 9

54=

542, 56, 59*

532 572, 582

532, 59

53, 57, 58, 59

59a

542, 56, 59s

53, 59

542, 56, 57, 62, 64, 66, 71, 79

534, 54, 563, 57a

58, 79

58, 62, 71, 79

76, 78, 79

58, 66

62, 65, 66

number to allow for the reciprocals, we find that there were 52 fertile
combinations and 36 sterile combinations.

If the self-fertile plant is omitted, there are 66 cross-combinations,
each well established by more than 1 pollination through which one
may group the remaining 11 individuals in intra-sterile classes. This
grouping is shown in table 40. The 1 1 plants fall into 3 classes consist-
ing of 5, 4 and 2 individuals. There is not a single case of intra-class
fertility and but 2 instances of inter-class sterility. Matings 10X7 an<^
9X8 show sterility where fertility is to be expected.

Argument on cross No. 4

Outside of the fact that a plant which seems to be a true self-fertile
appeared in this family and was tested with 11 self-sterile plants, no
new phenomena are found in cross No. 4. The same cross-sterility, the
same small number of inter-fertile, intra-sterile classes is found here
that is found in crosses No. 2 and No. 3. Cross No. 4 merely furnishes
corroboratory evidence of facts discussed earlier in the paper. It does
show, however, that the facts discovered in crosses 1, 2 and 3, are not
peculiar to a single hybrid.

INTRA-SPECIFIC PEDIGREE CULTURE EXPERIMENTS

Our experiments within each of these species can be described very
briefly for they have been confined largely to self-sterility tests. Not a

THE BEHAVIOR OF SELF-STERILE PLANTS

599

single thorough inquiry into the cross-mating proclivities of the plants
of a pure ( ?) species has been made. This may seem very odd when so
much time has been spent on inter-specific crosses. But our resolution to
favor the wider crosses is not without reason. We have satisfied our-
selves that the crosses within a species behave in a manner similar to that
of the crosses already described. It seems probable, therefore, that intra-
specific crosses would provide no data that could not be obtained from
inter-specific crosses, although the converse might not be true.

N. Forgetiana. Between 200 and 300 plants of N. Forgetiana have
been selfed under various environmental conditions, with pseudo-fertility
in only 3 instances, as has already been described. N. Forgetiana is
therefore a species on which environmental variations have little effect.
It is a species in which, if one could measure accurately the intensity
of the particular environmental factors that affect the full production of
self-sterility, either the norm for a standard average environment would
stand markedly toward the sterile end of the scale, or the dispersion
coefficient would be small. The environmental complex that tends to-
wards the greatest amount of pseudo self-fertility is necessary for any
visible effect on the plants.

A small number of intra-sterile classes has been shown to exist in
N. Forgetiana. Judging from cross-sterility percentages, the probable
maximum is between 5 and 8 groups, but no accurate classification has
been made.

N. angnstifolia. Between 80 and 100 plants of N. angustifolia have
been tested for self-sterility without the production of a single seed. This
work was done during three summer seasons on field-grown plants. A
certain environmental variation obtained of course, but since no pollina-
tions were made at the extreme end of a flowering season, one cannot
maintain that no pseudo-fertility exists. We are only justified in stating
that N. angustifolia is similar to N. Forgetiana in being difficult to in-
fluence by environmental changes.

Intra-sterile groups have also been demonstrated in this species. Their
number has not been determined but is probably no greater than in
N. Forgetiana.

N. alata. We have shown earlier that N. data is a self-sterile species
in which a considerable amount of pseudo self-fertility appears at the
end of the flowering season under adverse conditions. In other words if
the- environmental factors affecting self-sterility could be measured as
suggested in the case of Ar. Forgetiana, either the norm for a standard
average environment would be further toward the fertile end of the

Genetics 2: N 1917

6oo

E. M. EAST AND J. B. PARK

scale than in the latter species, or the dispersion coefficient would be
larger.

As in the other two species, intra-sterile classes have been proved to
exist, the maximum number probably being smaller than in N. Forge-
tiana or N. angustifolia.

The most important new fact discovered in N. alata is the probability
that a population may exist consisting of only one intra-sterile class
(compare family E). Recall that self-sterility is a sporophytic charac-
ter, that inbreeding decreases the number of intra-sterile classes, and
that there is no physiological or morphological obstacle to the fusion of
any two complemental gametes provided they meet. All of these facts
favor the idea that the behavior of self-sterile plants among themselves, —
given the presence of the character self-sterility through the presence
of a homozygous factor X, — is due to underlying causes which may be
pictured as follows. A certain number of factors which affect self-
sterility exist. The action of these factors is not cumulative. Mating
is possible normally only to plants which differ in at least one of these
factors.

If these premises be correct, after a very few generations of self-
sterile plants raised from selfed seed by taking advantage of the phe-
nomenon of pseudo self-fertility, one should find a population resulting
from a single capsule which is homozygous for these effective factors and
which is therefore wholly cross-sterile under normal conditions.

These conditions are very nearly met by the behavior of the grand-
progeny of the original pseudo self-fertile N. alata plant that is recorded
in table i. Table 41 is made up from table 1 by tabling the cross-matings
both ways when only made one way because of our belief that reciprocal
crosses are always identical. By this table it appears that the 3 matings
53 X 57, 58 X 53, and 58 X 59 are fertile. Tabled both ways there
are 6 fertile combinations. But let it be recalled that these matings were
made during a long flowering season, and that during its wane several
of the self-pollinations produced seed. What is more likely than that
some sterile cross-matings should show pseudo-fertility at the same time?
Our evidence is this. Of these matings 1 was made the middle of the
season and did not give a full capsule, the other two were made at the
end of the season. But this is not all. Our demonstration that every
member of an intra-sterile class should be sterile with every other mem-
ber is the result of an experience with nearly 10,000 cross-pollinations.
The exceptions which have been met are very infrequent and are well
within the expected experimental error. Now if table 41 be examined

THE BEHAVIOR OF SELF-STERILE PLANTS 6oi

carefully, it is seen that there is every indication that all of the 14 plants
listed belong to one class and that the 3 apparently fertile matings are
due to pseudo cross-fertility.

N. glutinosa. Not over a dozen plants of N. glutinosa have been tested
for self-sterility. It appears to behave like N. alata. Cross-fertility has
been demonstrated, but the number of cross-matings made is not suffi-
cient to prove the existence of intra-sterile groups. The above statement
also holds for the race described as N. commntata.

SUMMARY AND INTERPRETATION OF THE RESULTS

The experiments on the self-sterile species Nicotiana Forgetiana,
N. data, N. glutinosa and N. angustifolia described in the foregoing
pages, concern only the behavior of self-sterile plants when bred inter se.
All questions connected with the relation between true self-fertility and
self-sterility have been omitted designedly as pertaining to a distinct
problem. The inquiry thus limited is believed to have established the
following points :

1. Self-sterility is inherited.

2. The four species N. Forgetiana, N. alata, N. glutinosa and N.
angustifolia breed true to the tendency toward self-sterility.

3. Self-sterility is fully expressed in these species from the beginning
to the middle of the flowering season. Toward the close of the flower-
ing season, especially in plants exhibiting the effect of adverse environ-
mental conditions, some self-fertility may be shown. That this phe-
nomenon is simply a non-inherited fluctuation is confirmed in four ways :
(a) the graduated character of the increased fertility as the flowering
season wanes, (b) the return to complete self-sterility at the beginning
of a second flowering season, (c) the sterility of all progeny raised from
selfed seed, and (d) the failure to obtain an increased tendency toward
self fertility after three successive generations had been raised from
selfed seed of the most extreme variants. It has been called pseudo self-
fertility.

This fact naturally shows that self-sterility, whatever its nature, is
only a physiological impediment to self-fertilization.

4. Other environmental factors appear to have little or no influence on
self-fertility.

5. The waning of the reproductive period affects N. alata and N.
glutinosa more markedly than it does N. Forgetiana or N. angustifolia.
This indicates multiple allelomorphism in a fundamental factor the
presence of which is necessary for the development of .Lelf-sterility.

Genetics 2: N 1917

602

E. M. EAST AND J. B. PARK

(N,B. This factor should not be confused with any of those assumed
in the interpretation of the behavior of self-sterile plants among
themselves).

6. Cross-sterility in its nature identical with self-sterility was found
in every population of self-sterile plants tested. The percentage of
cross-sterility in different populations, based in each case on numerous
cross-matings, varied from 2.4 percent to 100 percent.

7. Omitting fluctuations toward self-fertility correlated with a wan-
ing flowering period and a few cases of true sterility as indicated by
microscopical examinations of the pollen, no variability in fruitfulness
was noticed in "fertile" combinations. Fertile matings always resulted
in full capsules.

8. Self-sterility behaves as a sporophytic character. This is demon-
strated by the behavior of reciprocal matings, — pairs of reciprocals
always giving like results either when fertile or sterile. It follows from
this fact that no selective fertilization occurs.

9. The F2 generation of a cross between N. Forgetiana and N. alata
showed a low percentage of cross-sterility, 2.4 percent. This cross was
followed to the F5 generation by means of successive sib matings. The
F5 generation showed 21.6 percent cross-sterility.

In a repetition of this cross made with different plants, several F2
populations studied each showed much higher percentages of cross-
sterility.

10. All of the individuals of a family arising from one mating may
be fertile with both parents, but a part of the individuals may be sterile
with one or with both parents.

11. Cross-sterility exhibits a regularity of behavior such that if A
is sterile with B and with C, it may be predicted that B will be sterile
with C. On the basis of this cross-sterility the plants in each family may
be divided into a relatively small number of groups in which each member
of a class is sterile with every other member of that class and fertile
with every member of every other class.

12. The distribution of the individuals within each class in several
of the families studied was such that the classes may not be assumed to
be of the same size. In certain cases this distribution rather resembled
that of the coefficients of a point binomial.

13 Assuming a point binomial distribution of individuals within the
classes as a limiting type, the number of intra-sterile classes necessary to
account for the highest percentage of cross-fertility found is estimated
to U: less than 25. In most of the families tested the number of intra-

THE BEHAVIOR OF SELF-STERILE PLANTS

603

sterile classes varied from 1 to 6. In a cross between N. alata and
N. Forgetiana in which 53 Fx plants were tested rather thoroughly,
5 (or 6) such classes were found.

14. In those instances where a part of the individuals of a family
were sterile to one or to both parents, only the members of a single class
behaved in that manner.

15. Individuals belonging to different families as well as to different
generations may belong to a single intra-sterile class.

16. Individuals belonging to different intra-sterile classes of the F1
generation when mated with the same individual, produced populations
varying in the number of intra-sterile classes.

17. Individuals belonging to a single intra-sterile F1 class when mated
with the same individual, sometimes produced populations having the
same number of intra-sterile classes, a similar distribution of individuals
within the classes, and possibly the same classes (see families H and I).
It is not established that this behavior is universal, however. In the
one case where the status of both the parents and the progeny as regards
cross-sterility was established very definitely (families H and I), the
two populations behaved in this manner; but in a case where the status of
neither the parents nor their progenies (families D and E) was quite so
clear, the two populations appeared to behave differently.

This rather varied series of facts can be given a very simple interpreta-
tion in keeping with recent interpretations of other inheritance phe-
nomena provided judgment be suspended on one or two obscure points.

Let us assume first that a self-sterile species is self-sterile because it is
homozygous for a fundamental self-sterility factor. Second, let us
assume that a series of partially coupled factors affect the behavior of
self- sterile plants among themselves. The action of these factors is on
the sporophyte, and the nature of this action is such that two plants are
not fertile together unless they differ by at least one of these factors.

It is not necessary to define the action of these factors more specific-
ally, although this will be attempted in a subsequent publication. It may
make matters somewhat clearer, however, to state that the immediate dif-
ference between a fertile and a sterile combination is in the rate of
pollen-tube growth. If at the height of the season a series of self-
pollinations and a series of cross-pollinations are made on a single plant
and the pistils fixed, sectioned and stained at intervals of 12 hours, it is
found by plotting the average length of the pollen tubes in each pistil
against time in 12 hour periods that the growth curve of selfed pollen
tubes is a straight line which reaches less than half the distance to the

Genetics 2: N 1917

604

E. M. EAST AND J. B. PARK

ovary during the life of the flower, while the curve of crossed pollen
tubes resembles that of an autocatalysis and reaches the ovary in less
than 96 hours. Further, it is unnecessary to know why gametes, which
themselves bear various factors effective on the behavior of self-sterile
plants, should act during the process preliminary to fertilization as if
each bore the factors characteristic of the plant on which they were
produced. Attention is called, however, to the fact that modern dis-
coveries tend more and more to showT that the sole function of the game-
tophytes of the Angiosperms is to produce sporophytes. The characters
which they possess appear to be wholly sporophytic, the factors which
they carry functioning only after fertilization. In other words, the
hereditary genes carried by pollen grains — and probably by eggcells —
may be thought of as being dormant until the appropriate time comes for
them to play their proper parts.

It may be helpful to draw a picture of what may be expected to happen
under the assumptions which have been made and to see how closely the
actual facts are paralleled. First, it should be stated that no interpreta-
tion of the fact that within a family the intra-sterile classes are often of
unequal size can be made without assuming linkage except by a number
of awkward subsidiary assumptions. Second, our picture is as simple
as possible in view of the facts at hand, but it may be extended ad libitum
as far as number of factors is concerned. Third, since all of the facts
of Mendelism are merely those to be expected from the known behavior
of the chromosomes as carrying bodies for our hypothetical genes,
chromosome diagrams are used without apology.

Assume first then that a plant of N. Forgetiana is heterozygous for
3 linked factors effective on the behavior of self -sterile plants, and that
the homologous chromosomes of an N. alata plant are heterozygous for
different multiple allelomorphs of the same factors. The two plants may
be represented thus.

N. Forgetiayia Ar. alata

A'"

B'"

THE BEHAVIOR OF SELF-STERILE PLANTS 605

These plants cannot be self-fertilized because all of their gametes are
influenced by the> sporophytic constitution ABC.A'B'C and A"B"C".
A^B^C", respectively, nor can either be fertilized by gametes borne
on a plant of like constitution.

Now each of these plants of N. Forgetiana and of N. alata produces 8
types of gametes. N. Forgetiana, for example, produces great numbers
of ABC and A'B'C, medium numbers of A'BC, AB'C, ABC and A'B'C
by one crossover or linkage break, and small numbers of AB'C and
A'BC by double crossing over. N. alata behaves in a similar manner.
Thus the progeny of this cross will consist of 82 =64 intra-sterile, inter-
fertile groups of individuals, the groups being of various sizes. Fur-
ther, since no individuals with constitutions ABC.A'B'C' or A"B"C".
A"'B"'C" are produced in the Fx generation, every F1 class will be
fertile with both of its parents.

Since by hypothesis two plants need differ by but one effective factor
in order to be fertile in inter-crosses, it is clear that matings may occur
in which certain of these factors are homozygous. To illustrate, it is
possible to obtain two plants of constitutions ABC.A'B'C and A"B"C.
A"'B'"C among the grandchildren of this generation. The factor C is
homozygous and can be left out of consideration since the two plants
form only 4 different types of gametes each. The first forms gametes
AB and A'B' in large numbers, and A'B and AB'in small numbers;
likewise the second forms gametes A"B" and A"'B'" in large numbers,
and A"B"' and A'"B" in small numbers. Even with the elimination of
the C allelomorphs as effective differences, therefore, it is possible to
obtain a family having 16 intra-sterile classes by crossing two such plants.
Of these classes 4 will be large, 8 medium and 4 small.

It is not unlikely that 16 classes is the maximum that need be con-
sidered, but what of the smaller number of groups usually found ? The
answer is that simplification can go on and on until very few intra-sterile
classes are formed.

Suppose, for example, that AB.AB' is crossed with AB.A'B; 4 classes
will be formed AB.AB, AB.A'B, AB'.AB and AB\A'B, of which the
third class will be sterile with the female parent and the second class
sterile with the male parent. Or, suppose that AA' is crossed with AA".
Again 4 classes will be formed, A A, A A', A A" and A' A". A A may
then be crossed with AA', and only 2 intra-sterile classes formed.

This may be assumed to be the simplest form in which a natural
population of self-sterile plants may exist, but theoretically it is possible
by taking advantage of the phenomenon of pseudo self-fertility or pseudo

Genetics 2: N 1917

6o6

E. M. EAST AND J. B. PARK

cross-fertility to obtain a family consisting of but i group. In such a
family every plant would be sterile with every other plant. It is possible
that the two families met in the course of our experiments in which cross-
sterility appeared to be universal, were of this kind.

This hypothesis fits perfectly what to us seem the important experimen-
tal facts. One may have F± generations of various types of complexity,
with an increasing simplicity in succeeding generations through inbreed-
ing; or, the F1 generation may be less complex than the F2 generation, —
the effect of inbreeding first becoming apparent in the F3 generation.
Cross-sterility with resultant intra-sterile classes in single or in different
families is explained. Both sterility and fertility in back-crosses is clear.
The similar behavior of reciprocal crosses is reasonable. Perfect intra-
sterility in the asexual progeny of a self-sterile plant is what is to be
expected. The facts established by Darwin and by Correns when
viewed with due consideration for pseudo-fertility become orderly. And
yet this is but hypothesis, to be modified, extended, restricted or super-
seded as becomes necessary. If it proves useful for a time it will have
served its purpose. Even now there are points upon which other heredity
phenomena throw no light. We will devote a concluding paragraph to
their discussion.

In our experimental wTork the number of intra-sterile classes and the
number of individuals within each class were determined as definitely as
possible. But these experiments have been too much of the pioneer type
not to be rough in many ways. With our present experience the same
facts could be determined more accurately and on much larger popula-
tions with less work than the original determinations demanded, and this
appears to be a requisite for further advance. According to our hypoth-
esis, accepting it without subsidiary refinements, the number of classes
should always be even, and the classes should be equal in size when only
2 or 4 make up the population. Furthermore there should always be
pairs of classes containing the same number of individuals. Now in mak-
ing some of our calculations we have assumed that the individuals are
distributed within the classes in numbers corresponding to the fre-
quencies of the point binomial. Such a distribution was assumed only
as a limiting type of unequal grouping, however, there being scarcely
any evidence that such a distribution is characteristic. As a matter of
fact only in the Fx of cross No. 2 and No. 3 and its descendants, families
H and I, is it possible to say that the number of individuals within the
various classes may not be approximately equal. But in these cases we

THE BEHAVIOR OF SELF-STERILE PLANTS

607

st.imble upon an obstacle that cannot be cleared away with our present
knowledge. The distributions found in these families are such that larger
samples of the populations could not give us classes of equal size. For
the present we must accept the conception of a small number of intra-
sterile groups in certain families with all that this involves. We might
explain them by subsidiary hypotheses of differential vitality or by redu-
plication in the sense of Bateson, but since there is no other good reason
for such assumptions we prefer to leave these matters in abeyance.

Bateson, W., and Gregory, R. P., 1905 On the inheritance of heterostylism in

Primula. Proc. Roy. Soc. B 76:581-586.
Baur, E., 191 1 Einfuhrung in die experimentelle Vererbungslehre. Ed. 2, 1914, pp.

viii + 401. Berlin: Borntraeger.
Beaton, D., 1850 On Amaryllids. Jour. Hort. Soc. 5 : 132-136.

Burck, W., 1908 Darwin's Kreuzungsgesetz und die Grundlagen der Bliitenbiologie.

'Rec. Trav. Bot. Neerl. 4:17-118.
Castle, W. E., 1896 The early embryology of Ciona intestinalis Flemming (L.).

Bull. Mus. Comp. Zool., Harvard University 27 : 201-280.
Comes, O., 1899 Monographic du genre Nicotiana. pp. 1-80. Naples: Typographic

Cooperative.

Compton, R. H., 1912 Preliminary note on the inheritance of sterility in Reseda
odorata. Proc. Phil. Soc. Cambridge 17: 7.
1913 Phenomena and problems of self-sterility. New Phytologist 12: 197-206.
Correns, C, 1912 Selbststerilitat und Individualstoffe. Festschr. d. mat.-nat. Gesell.

zur 84. Versamml. deutsch. Naturfoscher u. Arzte, Miinster i. W. pp. 1-32.
Darwin, Chas., 1859 Origin of species. Ed. 6, 1882, 2 vols. New York : D. Appleton.
1862 The various contrivances by which Orchids are fertilised by insects. Ed.

2, 1877. New York: D. Appleton.
1868 Animals and plants under domestication. Ed. 2, 1875, 2 vols. New York:
D. Appleton.

1876 Effects of cross and self fertilisation in the vegetable kingdom. Ed. 2, 1878.
New York: D. Appleton.

1877 Different forms of flowers on plants of the same species. Ed. 2, 1880.
New York : D. Appleton.

1)13 Inheritance of flower size in crosses between species of Nicotiana.
raz. 55 : 177-188.

phenomenon of self-sterility. Amer. Nat. 49 : 77-87.
iritance in crosses between Nicotiana Langsdorffii and Nicotiana alata.
cs 1 : 3H-333-

ind Hayes, H. K., 1912 Heterozygosis in evolution and in plant breed-
U. S. Dept. Agr., Bur. Plant Ind. Bull. 243, 58 pp.
1881 Die Pflanzen-Mischlinge. pp. iv -f- 569. Berlin : Borntraeger.
•suche und Beobachtungen iiber Kreuzung und Fruchtansatz bei Bliiten-
:en. Abh. nat. Ver. Bremen 11 : 413-421.

er Unfruchtbarkeit bei Bestaubung mit eigenem Pollen. Abh. nat. Ver.
Bremen 12:409-416, 495, 496.

Ginetics 2: N 1917

LITERATURE CITED

6o8

E. M. EAST AND J. B. PARK

Fuchs, H. M., 1914 a On the conditions of self-fertilization in Ciona. Arch. f. Ent-

wicklungsmech. d. Organ. 40 : 157-204.
1914 b The action of egg-secretions on the fertilizing power of sperm. Arch. f.

Entwicklungsmech. d. Organ. 40 : 205-252.
Gartner, K. F., 1844 Versuche und Beobactungen iiber die Befruchtungsorgane der

vollkommeneren Gewachse und iiber die natiirliche und kiinstliche Befruchtung

durch den eigenen Pollen, pp. x -f- [2] -J- 644. Stuttgart E. Schweizerbart'sche

Verlagshandlung.

1849 Versuche und Beobachtungen iiber die Bastarderzeugung im Pflanzenreich.
pp. xiv -f- 790. Stuttgart : Hering.
Gregory, R. P., 191 1 Experiments with Primula sinensis. Jour. Genetics 1:73-132.
Herbert, Wm, 1837 Amaryllidaceae. pp. vi + London : Ridgway & Sons.

Hildebrand, F., 1865 Bastardirungsversuche an Orchideen. Bot. Ztg. 23 : 245-249.
1866 Uber die Nothwendigkeit der Insektenhilfe bei der Befruchtung von

Corydalis cava. Jahrb. wiss. Bot. 5 : 359-363.
1869 tlber die Bestaubungsvorrichtungen bei den Fumariaceen. Jahrb. wiss. Bot.
7:423.

1896 tiber Selbststerilitat bei einigen Cruciferen. Ber. d. deutsch. bot. Gesell.
14: 324-331.

Jennings, H. S., 1916 The numerical results of diverse systems of breeding. Genetics
1 : 53-89.

Jost, L., 1905 Zur Physiologie des Pollens. Ber. d. deutsch. bot. Gesell. 23 : 504-515.

1907 tlber die Selbststerilitat einiger Bliiten. Bot. Ztg. 65:77-117.
Kirchner, O., 1905 liber die Wirkung der Selbstbestaubung bei den Papilionaceen.

Nat. Zeitschr. f. Land.- u. Forstwirtsch. 3: 1-16, 49-64, 47-111. A long list of

legumes investigated.

Knuth, P., 1808 Handbuch der Bliitenbiologie. Band I, pp. 400. Leipzig : Engelmann.
Kolreuter, J. G., 1761-6 Vorlaiifige Nachricht von einigen das Geschlecht der Pflan-

zen betreffenden Versuchen und Beobachtungen, nebst Fortsetzungen i, 2 u.

3. pp. 266. Ostwald's Klassiker, Nr. 41. Leipzig: Engelmann.
Kraus, E. J., 1915 The self-sterility problem. Jour. Heredity 6 : 549-557.
Lecoq, H., 1862 De la fecondation naturelle et artificielle des vegetaux et de l'hybri-

dation. pp. 17-425. Paris: Maison rustique. Quotes M. Riviere on Oncidium

Cavendishianum.

Lewis, C. I., and Vincent, C. C., 1909 Pollination of the apple. Ore. Agr. Exp. Sta.
Bull. 104. 40 pp.

Lotsy, J. P., 1913 Hybrides entre especes d'Antirrhinum. IVme Conference Interna-
tionale de Genetique. pp. 416-428.

Morgan, T. H., 1904 Some further experiments on self fertilization in Ciona. Biol.
Bull. 8:313-330.

1910 Cross and self fertilization in Ciona intestinalis. Arch. f. Entwicklungsmech.

d. Organ. 30 : 206-234.
1913 Heredity and sex. pp. ix -f- 282. New York: Columbia University Press.
Mowbray, Wm., 1830 Trans. Hort. Soc. A letter to the editor.
Muller, Fritz, 1868 Notizen iiber die Geschlechtsverhaltnisse brasilianischer Pflan-
zen. Bot. Ztg. 26: 113-116. A letter to Hildebrand.
1873 Bestaubungsversuche an Abutilon-Arten. Jenaische Zeitschr. f. Naturwiss. 7.

22-45, 441-450. Partial translation in Amer. Nat. 8 : 223-227.
1893 See Focke, W. O.
Munro, R., 1868 On the reproduction and cross-fertilization of Passifloras. Bot.

Soc. Edinburgh 9 : 309-402.
Rimpau, W., 1877 Die Selbststerilitat des Roggens. Landw. Jahrb. 6:1073-1076

THE BEHAVIOR OF SELF-STERILE PLANTS

609

Schacht, H., 1858 Uber Pflanzen-Bcfruchtung. Jahrb. wiss. Bot. 1 : 193-232.

Scott, John, 1865 On the individual sterility and cross-impregnation of certain species

of Oncidium. Jour. Linn. Soc. London 8 : 163-167.
Setchell, W. A., 1912 Studies in Nicotiana. Univ. California Pub. Bot. 5 : 1-28.
Sprengel, C. K., 1793 Das entdeckte Geheimniss der Natur im Bau und in der Be-

fruchtung. Ostwald's Klassiker, Nr. 48, 49, 50, 51. Leipzig: Engelmann.
Stout, A. B., 1916 Self- and cross-pollinations in Cichorium intybus with reference

to sterility. Mem. N. Y. Bot. Gard. 6 : 333-354-
Waite, M. B., 1895 The pollination of pear flowers. U. S. Dept. Agr. Div. Veg.

Path. Bull. 5: 1-86.

Ienetics 2: N 1917

Information for Contributor*

The length of contributions to Genetics will not be arbitrarily limited, but articiu.
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitations, as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication.

Footnotes should be inserted immediately after the reference marks which refer t?
them, and should be separated from the text by lines. All footnotes should be num
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fee
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4H X 7lA inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be I2}4 X 7$4 inches. Legends for all figures should be typewritten on a
separate sheet

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, NOVEMBER 1917

TABLE OF CONTENTS

East, E. M., and Park, J. B., Studies on self-sterility I, The be-
havior of self-sterile plants 505

Index 610

STUDIES ON SELF-STERILITY. II. POLLEN-TUBE GROWTH

E. M. EAST and J. B. PARK
Harvard University, Bussey Institution, Forest Hills, Massachusetts

Reprinted from Genetics 3 : 353-366, July 1918

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

vVilliam E. Castle

Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Pennsylvania

Edward M. East
Harvard University

Rollins A. Emerson
Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan
Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J., under the act of March 3, 1879.

STUDIES ON SELF-STERILITY II. POLLEX-TUBE GROWTH

E. M. EAST and J. B. PARK

Harvard University, Bitsscy Institution, Forest Hills, Massachusetts
[Received November 23, 1917]

TABLE OF CONTEXTS

PAGE

Introduction 353

Historical 354

Pollen-tube growth in artificial media 35b

Attempts to self-pollinate mutilated pistols 358

Pollen-tube growth in the pistil 359

Summary and discussion 362

Literature cited 365

INTRODUCTION

In the first paper of this series (East and Park 191 7), various pedi-
gree culture eNperiments involving the behavior of self-sterile Xicoti-
anas when crossed or selfed were described in considerable detail.
Among the points there established are the following:

1. The four species N. Forgctiana, X. alata, N. glutinosa and
N. angustifolia are truly self-sterile, but the character is affected by at-
tendant conditions in rather a peculiar way. The plants are completely
self-sterile during the active part of the flowering season, but toward
the close of this period, especially under adverse conditions, some self-
fertility may be shown. A7, alata and N. glutinosa are influenced thus
more markedly than N. Forgctiana and N. angustifolia. As a direct
corollary of this fact it follows that self-sterility is only a physiological
impediment to self-fertilization.

2. Cross-sterility, apparently of the same nature as self-sterility, exists.

3. Cross-sterility exhibits a regularity in its behavior such that if
plant A is sterile with plant B and with plant C, it may be predicted that
plant B will be sterile with plant C. On the basis of this cross-sterility
the plants within a family or even within a series of families may be
divided into a comparatively small number of groups in which each
member of a group is sterile with every other member of that group and
fertile with every member of every other group.

Genetics 3: Jl 1918

354

E. M. EAST AND J. B. PARK

4. Excluding the pseudo-fertility sometimes manifested during the
wane of the flowering season and true sterility which is due to non-
functional gametes, there is no fluctuation in fertility in compatible com-
binations. Incompatible combinations produce no seed, compatible com-
binations are fully fertile.

5. Reciprocal crosses always behave in the same manner; from which
it follows that the effective hereditary factors controlling compatibility
partake of the nature of the sporophyte rather than the gametophyte
generation.

The immediate cause of these peculiar manifestations is the varied rate
of pollen-tube growth characteristic of compatible and incompatible com-
binations. It is with these phenomena the present paper is concerned.

HISTORICAL

Numerous researches on pollen-tube growth have been made during
the last three or four decades, but owing to conflicting results, com-
paratively few contributions to knowledge have resulted.

Van Tieghem (1869) appears to have been the first investigator to
demonstrate pollen germination. He also showed that certain pollen
grains contain enzymes which invert cane sugar. This conclusion was
corroborated by Strasburger in 1886, who found also that starch could
be transformed into sugar, an indication of diastase. Green (1894)
succeeded in isolating both of these substances, but could not demon-
strate the presence of cytolase, an enzyme he believed must be con-
cerned in pollen-tube growth from the fact that growth is intercellular.
Green concluded that the reserve foods of the pollen grain are starch,
dextrin, cane sugar, maltose and glucose, the style containing the same
substances with the exception of dextrin. These conclusions, however,
can hardly be accepted without question.

Several authors have assumed that the pollen grains of certain spe-
cies have developed specific chemical requirements for germination, thus
accounting for their failure to obtain growth on artificial media; but
this can hardly be true since pollen grains will so often germinate on
stigmas of plants belonging to a different species or even to another
genus, to say nothing of the successes that have been obtained on arti-
ficial media varying from comparatively pure water to the pure (?)
agar agar of the trade. (Cf. Strasburger 1886: Rittinghaus 1887;

MOLISCH 1893; LlDFORSS 1896, 1899; JOST I905, I907; TOKUGAWA

1914). Questioning this conclusion, however, does not imply a denial
that in certain cases a preference may have been developed for particu-

POLLEN-TUBE GROWTH AND SELF-STERILITY

355

lar substances. In other words, pollen grains may germinate and grow-
on a variety of media, but the rate of growth may be much greater in
some cases than in others. No other conclusion seems to interpret ap-
propriately the results of Molisch (1893), Miyoshi (1894), Lidforss
(1896), Jost (1907) and ourselves, even though we agree with Stout
(1916) that experiments on artificial media are rather unsatisfactory.

Molisch (1893) believed he had demonstrated both anaerotropism
and chemotropism for acids and for secretions of the gynaecium, espe-
cially those of the stigma. He did not assume chemotropism to be a
general phenomenon, for there are pollen tubes (e.g. Orobus vernus)
which neither shun the air nor are attracted by the stigmas ; nevertheless
he did feel that chemotropism plays an important role in the passage of
the pollen tube to the egg cell. Molisch unquestionably made a serious
attack on the problem, for he investigated over 100 species.

Lidforss (1899) confirmed Molisch's observation that the pollen
tubes of Narcissus tazetta grow toward their own stigmas in a gelatin
medium, but his endeavors to imitate the effect of the stigmas with
various sugars, organic acids, amides and tannins were without success.
On the other hand, he did succeed in attracting the pollen tubes by the
use of pieces of onion bulb and by granules of a commercial preparation
of diastase. He believed that this success was due to certain proteins,
since the diastase still attracted the tubes after the ferment was killed
by heat, and since egg albumen washed free of mineral salts exerted
the same effect. Casein and "taka" diastase were without effect, but this
was thought to be dependent upon the presence in them of mineral salts.

He states that pollen tubes of Fritillaria hnperialis are more sensitive
to salts than Narcissus. The former were killed by the same diastase
preparation that had attracted the Narcissus tubes. On the other hand,
dialyzed egg albumen exerted a strong attraction for the Fritillaria
pollen tubes. Numerous other experiments were carried out in which
the effect of proteins on various species of Choripetalse was tested. No
tendency to attract was discovered, but this he believed to be due to
their great sensitiveness to small amounts of salts.

Miyoshi (1894) found that the stigma and style of many angio-
sperms contained reducing sugars. Chemotropic effects on their pollen
tubes were obtained by the use of several different sugars and dextrins
in a gelatin medium. Meat extract, asparagin, glycerine and gum
arabic had no effect, and alcohol and certain salts excited more or less
repulsion.

Perhaps the most general conclusion of Miyoshi was that pollen

Genetics 3: Jl 1918

356

E. M. EAST AND J. B. PARK

tubes could be turned from one solution to another if the concentration
of the second be increased as demanded by Weber's law.

These experiments, as well as the later ones along the same lines made
by Martin (1913), Tokugawa (1914), Andronescu (1915), Adams
(191 6) and others, must be accepted writh some reservation. There
is certainly a probability that pollen tubes show chemotropism, but it
must be admitted, as Stout (1916) maintains, that the amount of
pollen-tube growth observed in artificial media is small, probably never
over 1.5 mm. This being the case, one is likely to be over-influenced
by working hypotheses, and to conclude in favor of chemotropism with-
out due evidence. At the same time, these investigators must be thanked
for having given us a general idea, though perhaps somewhat superfi-
cial, of the physiology of the pollen tube.

The only data on pollen-tube growth bearing directly on the problem
of self-sterility are those of Jost (1907) and of Correns (1912).
They found that when a self -sterile plant is pollinated with its own pol-
len, the tubes are emitted freely but grow extremely slowly. Since a
cross-pollination on the same plant results in rapidly growing tubes, the
hypothesis was advanced, somewhat differently by each, that special
substances in each plant inhibit the growth of pollen tubes from pollen
of that plant.

POLLEN-TUBE GROWTH IN ARTIFICIAL MEDIA

Our own experiments on pollen physiology through the use of artifi-
cial media were made on N. angustifolia and N. Forgetiana. The me-
dium usually employed was 2 percent agar agar to which various nutri-
ents were added, although pure agar was used successfully. As nutrients
different percentages (1-20) of cane sugar, glucose, levulose and sodium
malate were used. The tubes grew well on all of these media, the best
development being obtained on 2 percent agar plus 20 percent cane
sugar. The maximum growth in this medium was about .6 mm. This
length of pollen-tube is almost negligible when compared with the 3 cm
to 7 cm necessary for the sperm nucleus to reach the ovule. Neverthe-
less, there seems to be no question but that there is a true germination
and a real growth on artificial media. If the pollen grains are perfect
morphologically, that is if no true pollen sterility is present, pollen
tubes are formed in nearly every case. Owing to the comparatively
short length to which they grow, one is hardly justified in plotting a
growth curve, but there is no doubt but that the rate of growth from
germination onward is either progressively slower, or that it starts

POLLEN-TUBE GROWTH AND SELF-STERILITY

357

slowly, reaches a maximum in from 12 to 24 hours and then falls off.
This fact should not be overlooked, as it is not what occurs when a
natural compatible mating is made, and shows clearly the great differ-
ence between pollen-tube growth in a natural and in an artificial medium.

In over 100 experiments of this type a high percentage of pollen
germination (70-80 percent) was obtained, and the pollen tubes grew
well no matter what medium was used. The tubes were emitted in all
directions, but in general the growth of an individual tube was in one
direction, though there was considerable variation from a straight line.

Since the experiments of Molisch, Lidforss, Miyoshi and others
all indicated that pollen tubes are attracted toward certain substances, it
was thought that possibly the same phenomenon might occur if portions
of the gynsecium of flowers of the same species were placed in the
media. Many experiments were tried therefore in which parts of stig-
mas, styles and ovaries (both crushed and uncrushed) or of their ex-
tracts were placed in the media and pollen scattered near them at vari-
ious distances (.5 to 3 cm). In some cases gynaecium parts from the
same self-sterile plant which furnished the pollen were used, in other
instances the gynaecium parts came from one plant and the pollen from
a plant cross-sterile with the first. The tendency of the pollen tubes in
these tests was compared with that of pollen from the same plants when
placed near gynaecium parts of plants with which the pollen was known
to be compatible. We were not able to discover any difference in the
behavior of the pollen tubes in these trials. Occasionally, perhaps in
10 percent of the cases, the pollen tubes seemed to be attracted by the
gynsecium parts, but the percentage was about the same in all cases. If
there was really any attraction at all, which is doubtful, it was no
greater between plant parts from plants known to be capable of effecting
mutual cross-fertilization than it was from plant parts taken from plants
which were cross-sterile together, or even when taken from the same
plant. Notwithstanding the fact that there was no decided turning of
the tubes toward any object or substance placed in the medium, there
was some evidence that the presence of gynaecium parts promoted
growth. On a number of occasions data were secured such as are shown
in table 1.

The evidence of stimulation from the presence of ovules and more
particularly of stigmas is unmistakable but whether the presence of
"compatible'' stigmas or ovules shows an additional stimulation over
that due to "incompatible'' stigmas and ovules is doubtful.

Genetics 3: Jl 1918

E. M. EAST AXD J. B. PARK

Table i

Growth of pollen tubes in a 2 percent agar medium in Van Tieghem cells.
Plants A and B are both self-sterile, but arc cross-fertile.

Exp.

Materials used

Ave. length tubes in ^

Pollen A -)- ovary B crushed

Pollen A -f stigma B crushed

Pollen B -{- ovary A crushed

Pollen B -f ovary B crushed

Pollen B -j- stigma B crushed

Pollen A only

Pollen B only

Pollen A -{- stigma A

120

Pollen A + stigma B

170

Pollen A

Pollen B -j- stigma B

100

Pollen B + ovary B

Pollen B + ovary A

100

Pollen B

Pollen B + stigma B

Pollen B -f stigma A

100

Pollen B + ovary B

Pollen B -f ovary A

Pollen B only

Pollen A only

Note: Data taken after 24 hours in exp. 1, after 48 hours in exp. 2, and after 24
hours in exp. 3.

These results do not corroborate the work of the earlier writers men-
tioned above, but neither do they prove there is no such thing as pollen-
tube chemotropism. They indicate some sort of nutritive value or
stimulative effect of substances contained in gynaecium parts, but there
is no evidence that gynaecium parts are more nutritive than other plant
parts nor that "compatible" plant parts are better than "incompatible"
plant parts. Experiments of this kind are unsatisfactory. They may
not be useless, but it seems improbable that any notable increase in
knowledge will be obtained by their use until the technique is so im-
proved that the growth curve in artificial media compares favorably with
the natural growth curve.

ATTEMPTS TO SELF-POLLIX ATE MUTILATED PISTILS

In the early part of our work numerous attempts were made to ob-
tain selfed seed on self -sterile plants by endeavoring to force pollen

POLLEN-TUBE GROWTH AND SELF-STERILITY

359

tubes to grow in shortened styles. The experiments were of two types.
Various methods of obtaining temporary unions between stigma and
style after excising portions of the latter were tried by means of wax
and glass envelopes. All of these attempts were unsuccessful. Pollina-
tions of decapitated pistils were also made using stigmatic fluid and
various sugars as germination media. In two cases, seed was obtained
where stigmatic fluid was used. The matter is merely mentioned to
show the possibility of developing a successful technique. The experi-
ments were discontinued as soon as it had been proved by end-season
self-pollinations that self-fertilization of self-sterile plants is possible,
and that for this reason self-sterility is no true impediment to the fusion
of an egg with a sperm nucleus which is the product of the same plant.

POLLEN-TUBE GROWTH IN THE PISTIL

The most gratifying experiments along these lines were those con-
cerned with the rate of growth of pollen tubes in the pistils of self-
pollinated and of cross-pollinated plants. Studies were made on pistils
of N. Forgetiana, N. alata and N. angitstifolia. The technique con-
sisted of making series of self-pollinations and of cross-pollinations,
both compatible and incompatible, on a single plant, collecting the pistils
thus treated on successive days, and fixing, dehydrating, imbedding and
sectioning them in the usual manner. Longitudinal sections about 10^
thick were used. Triple staining with safranin, gentian violet and
orange G gave the best results, although safranin alone was almost as
satisfactory.

About 400 slides each containing from 10 to 30 serial sections were
prepared in this manner.

The germination of the pollen was found to be just as high in the
selfed as in the crossed pistils, and the number of pollen tubes in a
single pistil was very large. By actual count it was determined that
single pistils may contain from 1200 to 2000 tubes, a number sufficient
to fertilize from 4 to 6 times the number of ovules in the ovary.

The distribution of pollen tubes in the selfed pistils was always minus
skew as is shown by the following sample frequencies.

Distance from stigma mm

1 2.5 1 3-5 1 4-5 1 5-5 1 6.5 1 7-5 1

8.5 19-5 1 10.5 1 1 1-5 1 12-5 1 13.5 1 14-5

1 section after 6 days

| i| 43 1 30 1 30 1 20 1 I3|

7| 7| 5| 2| i| i|

1 section after 7 days

1 1 1 1 4|20|24|

i8| 9| 7| 2| 2] i|

The large number of short pollen tubes and the few greatly in the
lead raise the important question : Do some pollen tubes grow faster

Genetics 3: Jl 1918

3.6o

E. M. EAST AXD J. B. PARK

than others because they have different genetic constitutions? Should
this be so, selective fertilization would result. We have good evidence,
however, that this is not the case.1 The difference in length of the vari-
ous pollen tubes is probably due largely to variation in the time of
germination. Those pollen grains which are in intimate contact with
the moist stigma absorb its secretions and put forth their tubes more
quickly than do those less favorably situated. If, therefore, a flower
be pollinated at a particular time, one may accept the mode of the fre-
quency distribution of the pollen tubes within the pistil as a proper mea-
sure of pollen-tube growth for a given period.

A considerable amount of data of this kind has been collected for the
purpose of comparing the rate of pollen-tube growth in selfings, and in
cross-fertile and cross-sterile matings. Figures i, 2 and 3 are graphs
made from a random sample of the results secured. Each point repre-
sents the estimated modal length of the pollen tubes in a pistil collected at
the indicated length of time after pollination. Ordinates to the various
broken lines at the top of the figure represent the total length of the
pistil of the respective plants. Figures 1 and 2 are from studies on
Nicotiana Fowgetiana. The rate of growth of selfed pistils in figure 1
and of the composite of selfed pistils in figure 2 are typical of selfed
plants during the active part of the flowering season. The growth
curve is practically a straight line. The pollen tubes grow well; as far
as one may judge visually there is no difference between them and
tubes from cross-pollinations. But they grow so slowly that even after
the extraordinary flower life attained on plant 3 where a 14-day point is
recorded, the pollen tubes are only half way to the ovary.

In crossed pistils, on the other hand, if the mating is compatible,
growth curves are produced which are of a very different character.
The pollen tubes start to grow at about the same rate as in selfed pistils,
but the speed continually increases until fertilization ensues, — usually
after from 3 to 5 days. The curve simulates that of an autocatalytic
reaction.

The points plotted for selfings on plants 6 and 7, shown in figure 2
produce curves intermediate between the true "self" curves and the
"cross" curves. This is the type of change that is brought about late in
the flowering season and which when carried to an extreme produces
pseudo self-fertility.

Figure 3 represents curves similar to those just described, but com-

1 For a preliminary report on this subject see Jour. Heredity 8:382, 1917. Further-
details will be published later.

POLLEX-TUBE GROWTH AND SELF-STERILITY

361

Lenffa of Prst'f oie^f JZ/nm

•>

J-

2 3 -4 S 6

Time in Hoys

Figure 2. — A composite growth curve of pollen tubes from normal selfed pistils
of two plants of Nicotiana Forgetiana. End-season growth curves from selfed
pistils of plants of the same species.

piled from data upon N. angustifolia. The sellings produce straight
lines and the compatible crosses produce curves in which pollen-tube

Genetics 3: Jl 1918

362

E. M. EAST AND J. B. PARK

7/rrre /n /Yours

Figure 3. — Growth curves of pollen tubes from pistils of plants of Nicotiana angus-
tifolia. Plants 5 and 2 are cross-sterile with each other.

growth becomes faster and faster as they approach the ovary. Since no
points were noted on the crossed pistils during the early part of the prog-
ress of the tubes, the curves as drawn are not exactly like those of figure
1, but this difference has no real basis.

One of the curves of figure 3 is from an incompatible cross. It is
practically the same as the self upon the same plant.

Many other curves have been made, but these are fair samples. All
show the same type of growth. Selfs made during the active part of
the flowering season always produce straight-line growth curves. The
growth of the tubes is at a steady rate throughout the. length of the
style traversed. The same is true of incompatible crosses. Toward the
wane of the flowering season, however, the type of curve always changes
until it resembles that of a compatible cross, though there remains a
slight difference. The growth is more rapid than that from mid-season
selfings, but the velocity in these cases is almost constant while in com-
patible crosses there is an acceleration.

SUMMARY AND DISCUSSION

The following general conclusions may be drawn from the data re-
ported in this paper.

1. Pollen grains of the four species of Nicotiana used in these in-
vestigations germinate on many artificial media as well as on the stig-
mas of the plants themselves, but the length of pollen tube obtained is
never over .6 mm and usually is from .1 mm to .2 mm.

POLLEN-TUBE GROWTH AND SELF-STERILITY

2. Though pollen tubes are emitted freely on artificial media, their
growth is unlike that obtaining under natural conditions. There the
growth either remains constant or becomes more rapid, but on artificial
media the latter part of the growth is markedly retarded.

3. Pollen-tube chemotropism was not observed, but since the pres-
ence of gynsecium parts in the medium caused a longer tube to be pro-
duced, this negative result may have been due to rapid diffusion of the
chemical stimulants.

4. It is possible to obtain pollen-tube penetration followed by ferti-
lization in a decapitated style by the use of a proper germinating me-
dium.

5. Pollen grains germinate as well on the stigmas of flowers of the
same plant as they do on the stigmas of flowers of other plants with
which they are compatible.

6. From 5 to 10 times as many pollen tubes are produced as are nec-
essary to fecundate the ovules.

7. The pollen tubes produced after a selfing are indistinguishable in
size from the pollen tubes produced after a cross, when pollen tubes of
the same length are measured.

8. Though there is variation in the length of the pollen tubes both
after a self-pollination and after a cross-pollination, this variation seems
to be due wholly to differences in the rate of germination of the pollen
grains or to other causes which are environmental in nature and not to
any differences in gametic constitution between the pollen grains them-
selves.

9. Pollen tubes produced after self-pollinations grow steadily and ap-
parently normally, but do not reach the ovary before the flower decays
because this growth is slow. Length plotted against time is a straight
line.

10. Pollen tubes produced after a compatible cross start their growth
at about the same rate as the pollen tubes produced after selfing ; but the
growth becomes constantly more and more rapid. Length plotted
against time produces a curve that resembles that of an autocatalytic re-
action.

11. Cross-sterile combinations resemble selfings in the rate of pollen-
tube growth.

12. At the wane of the flowering season rate of growth of self pollen
tubes becomes more rapid, though there is little evidence of acceleration
during their passage down the style.

These results appear to us to show that the pollen tubes in a selted

Genetics 3: Jl 1918

3^4

E. M. EAST AND J. B. PARK

pistil are not inhibited in their growth by substances secreted in that
pistil, but' rather that a substance or substances are secreted in the pistil
after a compatible cross which accelerate growth, and that the direct
cause of this secretion is a catalvser which the pollen-tube nucleus is
able to produce because the zygotic constitution of the plant producing
it is different in certain particular hereditary factors from that of the
plant on which it is placed. Since pollen-tube growth is intercellular,
it may be that some cytolysis occurs, but in the main there appears to be
some local reaction between the pollen-tube nucleus and the contiguous
cells of the style which produces or makes serviceable the nutrients neces-
sary for tube growth. The action must be local because the presence of
compatible pollen tubes does not accelerate the growth of self pollen
tubes. The reaction must be mutual because one cannot account for the
peculiar behavior of self -sterile plants in crosses if one interprets pollen-
tube growth as a simple parasitism.

The action in the case of self pollen tubes or where incompatible
crosses are made, is accounted for by the likeness of the parents in the
effective hereditary factors postulated. Some action must take place in
these cases because growth occurs but this is the action on which the
other phenomenon is built. It does not seem to be in the nature of an
inhibition because the growth of self pollen tubes is constant from the
beginning to the end of the growth. If any change in rate of growth
occurs it is a slight acceleration. Nor does this growth appear at all
analogous to immunity phenomena as has been suggested by Compton
(1913) and Stout (1916). At least if there is similarity, the current
theories of immunity do not serve to make the matter any clearer.

The change in rate of tube growth in selfed pistils toward the end of
the flowering season we hold to be a phenomenon apart from those just
discussed. Here instead of a mutual reaction between active cells, there
seems to be more of a parasitism. The pollen tube is active, but the
style cells are inactive. The active pollen tube, then, feeds on the
broken down cells of the style. Our evidence of this is not conclusive,
but it is very suggestive that the pollen-tube growth in this pseudo-
fertility, as we have called it, seems to be merely faster throughout the
whole period of its growth rather than increasingly rapid. It would
seem that if this were the case, then where there is pseudo cross-fertility
between a plant in active flower and a plant at the wane of the flowering
period, reciprocal crosses would not give the same result as was main-
tained in our earlier paper. But our contention there was that plants in
the height of their flowering period give like results in reciprocal crosses.

POLLEN-TUBE GROWTH AXD SELF-STERILITY

365

Moreover in cases where pseudo cross-fertility was noled, with perhaps
a very few doubtful ones, the plants used in reciprocal crosses were at
about the same flowering stage. We have since made a few incompati-
ble crosses (10), however, between plants at the most active part of
their flowering season and plants at the extreme end of their flowering
season. There was, and we may admit we expected it, a difference be-
tween the reciprocals. Some fertility was shown when the old plants
were used as pistillate parents, there was none in the reciprocals.

LITERATURE CITED

Adams, J., 1916 On the germination of the pollen grains of apple and other fruit

trees. Bot. Gaz. 61 : I3I-I47-
Amici, G. B., 1830 Note sur la mode d'action du pollen sur la stigmate. Ann. sci.

nat. 21 : 3^9-33^-

Andronescu, D. I., 191 5 The physiology of the pollen of Zea mays with special
reference to vitality. Thesis: Univ. of Illinois, pp. 1-36. Champaign: Pub.
by author.

Correns, C, 1889 Culturversuche mit dem Pollen von Primula acaulis Lam. Ber.

d. d. bot. Gesell. 7 : 265-272.
1912 Selbststerilitat und Individualstoffe. Festschr. d. math. -nat. Gesell. zur 84.

Versamml. deutsch. Naturforscher u. Arzte Minister i. W. pp. 1-32.
Coupix, Hexri, 1906 Sur Taction de quelques alcalo'ides a l'egard des tubes pollin-

iques. Compt. rend. Acad. Sci. Paris 142 : 841-843.
Degagxy, Charles, 1886 Sur le tube pollinique, son role physiologique. Compt.

rend. Acad. Sci. Paris 102:230-231.
East, E. M., and Park, J. B., 1917 Studies on self-sterility I. The behavior of

self-sterile plants. Genetics 2 : 505-609.
Greex, J. R., 1894 Researches on the germination of the pollen grain and the nutri-
tion of the pollen tube. Phil. Trans. Roy. Soc. London B 185: 385-409.
Haxsgirg, A., 1897 Beitrage zur Biologie und Morphologie des Pollens. Sitzber.

kgl. Ges. d. Wiss. Prague 23: 1-76.
Jost, L., 1905 Zur Physiologie des Pollens. Ber. d.d. bot. Gesell. 23:504-515.

1907 tiber die Selbststerilitat einiger Bluten. Bot. Ztg. 65:77-117.
Lidforss, Bexgt, 1896 Zur Biologie des Pollens. Tahrb. f . wiss. Bot. 29 : 1-38.

1899 tiber den Chemotropismus der Pollenschlauche. Ber. d. d. bot. Gesell.

1 7 : 236-242.

Maxgix, Louis, 1890 Sur la callose, nouvelle substance fondamentale existant dans
la membrane. Compt. rend. Acad. Sci. Paris 110:644-647.

Martix, J. N., 1913 The physiology of the pollen of Trifolium prat ens e. Bot. Gaz.
56 : 112-156.

Miyoshi, Maxabu, 1894 tiber Reizbewegungen der Pollenschlauche. Flora 78 :
76-93.

Molisch, Haxs, 1893 Zur Physiologie des Pollens. Sitzungsber. Weiner Akad.

Wiss., math.-naturw. Kl. 102: 423-448.
Rittixghaus, P., 1887 Einige Beobachtungen uber das Eindringen der Pollenschlauche

ins Leitgewebe. Verh. nat. Vereines Rheinland 43: 105-122.
Stout, A. B., 1916 Self- and cross-pollinations in Cichorium intybus with reference

to sterility. Mem. N. Y. Bot. Gard. 6 : 333-454.

Genetics 3: Jl 1918

366

E. M. EAST AND J. B. PARK

Strasburger, Em 1886 Uber f remdartige Bestaubung. Jahrb. f. wiss. Bot. 17: 50-98.
van Tieghem, Ph., 1869 Recherches sur la vegetation libre du pollen et de l'ovule.

Ann. sc. nat. V, 12 : 312.
Tokugawa, Y., 1914 Zur Physologie des Pollens. Jour. Coll. Sci. Tokyo 35:1-35.
Tomaschek, Anton, 1882 Das Bewegungsvermogen der Pollenschlauche und Pol-

lenpflanzen. Sitzber. Wien. Akad. Wiss. 84 (1881) : 612-615.

Information for Contributors

The length of contributions to Genetics will not be arbitrarily limited, but article
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitation* as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to bt of
primary interest to the geneticist, will find here an appropriate place of publication.

Contributors are requested to observe the usual care in the preparation of manu-
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published h a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer
them, and should be separated from the text by lines. All footnotes should be nura
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fot
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4$iX7lA inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be I2$4 X 7% inches. Legends for all figures should be typewritten on a
separate sheet

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. Georje H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, JULY 1918

TABLE OF CONTENTS

PAGE

Sax, Karl, The behavior of the chromosomes in fertilization. . . . 309

Harris, J. Arthur, Further illustrations of the applicability of
a coefficient measuring the correlation between a variable
and the deviation of a dependent variable from its prob-
able value 328

East, E. M., and Park, J. B., Studies on self -sterility. II. Pollen-
tube growth 353

Wright, Sewall, On the nature of size factors 367

Robbins, Rainard B., Some applications of mathematics to breed-
ing problems. Ill 375

Robbins, Rainard B., Random mating with the exception of sister

by brother mating 390

STUDIES ON SELF-STERILITY. III. THE RELATION
BETWEEN SELF-FERTILE AND SELF-STERILE PLANTS

E M. EAST

Bussey Institution, Harvard University, Forest Hills, Massachusetts

Reprinted from Genetics : 4 341-345* July, 19*9

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle

Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Michigan

Edward M. East
Harvard University

Rollins A. Emerson
Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan

Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J., under the act of March 3, 1879.

STUDIES ON SELF-STERILITY. III. THE RELATION BE-
TWEEN SELF-FERTILE AND SELF- STERILE PLANTS

E. M. EAST

Harvard University, Bussey Institution, Forest Hills, Massachusetts

[Received March 14, 1919]

TABLE OF CONTENTS

PAGE

Introduction

Compton's work on Reseda

Corroboration of Compton's results by experiments on Nicotiana

Summary

Literature cited

34i
34i

342
345
345

INTRODUCTION

In the first paper of this series (East and Park 191 7), where the be-
havior of self-sterile plants was described in some detail, it was pointed
out that the difference between self-fertile and self-sterile plants might
prove to be a wholly different problem. This statement has been over-
looked by several reviewers who criticized the interpretation of the be-
havior of self-sterile plants proposed because it failed to take into ac-
count the phenomenon of self -fertility. The prediction was not made at
random, however. Even at that time various data had been gathered in-
dicating a simple one-factor difference between self-fertile and self-
sterile plants in keeping with Compton's previous work (1912) on
Reseda odorata. It can now be stated unequivocally that the position
then taken is correct. In the material investigated self-fertile plants
differ from self-sterile plants by a single essential Mendelian factor.
Self-fertility is dominant. Adopting a presence-and-absence mode of
expression, a plant is self-fertile because of the presence of a determiner
for self-fertility; when this determiner is absent, the individual is self-
sterile.

The only investigation in which crosses between self-fertile and self-
sterile plants have been studied is that of Compton (1912, 191 3) on the
mignonette, Reseda odorata. Having had his attention directed to the
species by the observations of Darwin, a number of experiments were
made with the following results :

compton's work on reseda

Genetics 4: 341 Jl 1919

342

E. M. EAST

(i) Self-sterile plants intercrossed produced only self-sterile off-
spring. (2) Certain self-fertile plants when self-fertilized threw ap-
proximately 3 self-fertile to 1 self -sterile offspring. (3) These same
plants when crossed with self-sterile individuals, produced self-fertiles
and self-steriles in the ratio one to one. (4) Other self-fertile plants
yielded none but self-fertile offspring from selfed seed.

These facts are satisfactorily interpreted by assuming a single factor
difference with complete dominance. The recessives produced only re-
cessives. The dominants in part produced only dominants and in part
produced both types in the usual ratio of 3 to 1. He was dealing, there-
fore, in part with homozygous and in part with heterozygous plants,
and the behavior of the heterozygous individuals was checked by the
back cross with the recessive.

CORROBORATION OF COMPTOX's RESULTS BY EXPERIMENTS OX XICOTIANA

These experiments of Comptox have been corroborated by crossing
two of the self-sterile species used in our previous work, Xicotiana For-
getiuna and Xicotiana. data, with a third species Xicotiana Langsdorffii,
which is consistently self-fertile.

Xicotiana Forgctiana and Xicotiana Langsdorffii were crossed recip-
rocally. In each case the plants were very vigorous, exceeding both
parents somewhat in height. They grew quickly, matured rapidly, and
produced a profusion of fertile flowers. The flowers were somewhat
intermediate in size but resembled the larger-flowered parent, Xicotiana
Forgctiana, in form. Xo difference ^ould be discerned in the reciprocals
either in the first or second hybrid generation in appearance or behavior.
The two experiments may therefore be considered as one.

About 400 plants were grown and selfed by hand with the usual pre-
cautions against cross-pollination. In each case, from 6 to 20 blossoms
were operated on. Every plant zcas self-fertile. Seed set in abundance,
filling the capsules. Xot every flowe ' pollinated produced seed, of course,
but the percentage was practically the same as that obtained in check ex-
periments on pure Xicotiana Langsdorffii, 85 percent. The work was
completed as early in the season as possible in order not to be disturbed
by the pseudo self-fertility which is sometimes present in self-sterile
plants at the close of the flowering season.

From selfed seed of the cross N. Forgetiana X N. Langsdorffii, 89
plants were grown and tested for self-fertility by guarded hand-pollina-

SELF-FERTILE AND SELF-STERILE PLANTS

343

tions such as were made in the first hybrid generation. Of them 70
proved to be self-fertile and 19 self-sterile.

From selfed seed of the reciprocal cross, 92 plants were tested. Of
this lot 74 showed self-fertility and 18 self-sterility. There was a sum
total, therefore, of 144 self-fertile and 37 self-sterile plants in F2, a
ratio 3.8 to 1.

If the hypothesis of a one-factor difference is correct the deficiency of
recessives is somewhat greater than is to be expected in a population of
this size. Nevertheless this failure to measure up to expectation need not
disturb us. About one-fourth of the bags used in protecting the flowers
were torn by wind, and the plants had to be tested a second time. This
unfortunate occurrence prolonged the experiment until well into Sep-
tember when the plants were past their prime. It is not unexpected
therefore that some truly self -sterile plants should have been listed as
self-fertile because of "end-season" pseudo-fertility. In fact a slight
fertility was shown by about 30 percent of the plants classed as self-
sterile; i.e., they produced partially filled capsules in about 15 percent of
the pollinations.

These plants were tested further by taking them into the greenhouse
and bringing them into a second season of flowering. Pollinations were
then made at the beginning of the season, and the plants proved to be
fully self-sterile.

If this be not sufficient evidence to prove the case, there is the be-
havior of the third hybrid generation to be relied upon. All progeny of
the recessive {self-sterile) segregates of F2 were again self -sterile.
About 200 were tested.

The cross between N. Langsdor jfii and N. alata yield results similar
to those just described. The plants of the first hybrid generation were
all self-fertile; those of the second hybrid generation were partly self-
fertile and partly self-sterile. About 200 F2 plants were tested, of which
38 were self-sterile. Again there was a deficiency of recessives. The
progeny of the self-steriles were a1 1 self-sterile, but no investigation of
the amount of pseudo self -fertility was made. The matter of particular
interest in this cross was the cross-fertility of F2 plants having flowers
of very different corolla lengths. Flowers were obtained as short as 2.0
cm and as long as 6.0 cm, yet reciprocal crosses were very easy to make.

It will be remembered that Kolreuter was unable to fertilize Mirab-
ilis longiflora with pollen from Mirabilis Jalapa although the reverse
cross could be carried out without difficulty. In interpreting these facts

Genetics 4: Jl 1919

344

E. M. EAST

it has been customary to assume that M. Jalapa pollen tubes are short
and thus unable to reach the micropyles of the ovaries of M. longiflora.
From work on pollen-tube growth (East and Park 191 8) and observa-
tions on the F2 individuals of the cross between N. Langsdorffii and
N. alata, we believe this assumption to be incorrect. Pollen tubes of all
species observed by us have continued to grow as long as the flowers
remained unwithered even in many generic crosses. The real cause of
the occasional lack of success when a long-flowered plant is pollinated
with pollen from a short-flowered plant, therefore, is in the "death" of
the flower before the pollen tube has had time to reach the micropyle.

Though we may conclude that lack of a particular factor F results in
self -sterility, there are some other factors to be considered in the be-
havior of crosses between self-fertile and self-sterile plants. When the
self -sterile segregates of the cross between N. Forgetiana and N. Langs-
dorffii were examined carefully throughout the second flowering season,
the type of self-sterility present did not seem to be the same in all cases.
A majority of the plants exhibited a much greater amount of pseudo
self-fertility than had ever been found in N. Forgetiana. In that species
only an occasional plant produced a few selfed seeds and then only at the
extreme end of the flowering season. Among the F2 individuals of the
cross, however, pseudo-fertility set in about the middle of the season and
from then on it was very easy to get capsules which on casual examina-
tion would be said to be full of seed. As a matter of record only about
30 percent of such pollinations were successful and the capsules on the
average had only about 70 percent of the normal complement of seed.
Nevertheless, some 60 to 75 percent of the F2 segregates classified as
self-sterile showed at least 100 times the pseudo-fertility of the parent
species, N. Forgetiana. The remaining plants were comparable to the
latter in self-sterility.

It was also noticeable that the progeny of the most self-sterile of the
F2 plants were similar to them, while the progeny of the others were in
part like their mother plants and in part like N. Forgetiana.

The simplest explanation of this state of affairs is that there is really
a two-factor difference as regards self-sterility and self-fertility between
N. Forgetiana and N. Langsdorffii. N. Langsdorffii is homozygous for
a factor F ; when this factor is absent the plants are self-sterile. It is also
homozygous for a dilution factor D. The constitution of N. Forgetiana
is dd ff. The Fx individuals, having the constitution Ff Dd, are all self-
fertile. In the F2 generation a ratio of 9 FD : 3 Fd : 3/Z} : 1 fd is

SELF-FERTILE AND SELF-STERILE PLANTS

345

obtained. There are 3 self-fertile to 1 self-sterile because of the distri-
bution of the allelomorphic pair F and /. But of the self-steriles, those
having the constitution fD show a great deal more pseudo self-fertility
than those having the constitution fd. Only the fd plants are wholly
comparable to N. Forgetiana.

In describing the behavior of self-sterile plants this statement was
made (East and Park 191 7) :

"The waning of the reproductive period affects N. alata and N.
glutinosa more markedly than it does N. Forgetiana or N. angastifolia.
This indicates multiple allelomorphism in a fundamental factor the pres-
ence [or absence] of which is necessary for the development of self-
sterility. This factor should not be confused with any of those assumed
in the interpretation of the behavior of self-sterile plants among them-
selves."

The peculiarities of the cross between N. Forgetiana and N. Langs-
dorffii show that subsidiary factors affecting the manifestation of self-
sterility, given homozygosity in ff, are as likely to be the interpretation
of the differences shown in these four species as is multiple allelomorph-
ism.

SUMMARY

Data are reported showing that in Nicotiana self -sterility is due to
the presence of the allelomorph of a dominant fertility factor, F. When
a population is homozygous for this factor, ff, it is self-sterile.

The factors which control the peculiar and systematic behavior of self-
sterile plants when intercrossed among themselves are wholly independ-
ent of this factor and the latter does not need to be considered in an
interpretation of their expression.

The manifestation of self-sterility as evinced by the degree to which
pseudo-fertility shows, is due to a subsidiary inherited factor (or fac-
tors), but without the presence of the principal factor ff there is no evi-
dence that it functions.

LITERATURE CITED

Compton, R. H., 1912 Preliminary note on the inheritance of sterility in Reseda
odorata. Proc. Phil. Soc. Cambridge 17:7.
1913 Phenomena and problems of self -sterility. New Phytologist 12:197-206.
East, E. M., and Park, J. B., 1917 Studies on self-sterility. I. The behavior of self-
sterile plants. Genetics 2 : 505-609.
1918 Studies on self-sterility. II. Pollen-tube growth. Genetics 3:353-366.

Genetics 4: Jl 1919

Information for Contributors

Contributors are requested to observe the usual care in the preparation of manu-
scripts. All references to literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published in a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

No limitation will be placed on the kind of illustrations to be used, but as funds for
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4H X 7lA inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be 12H X 7H inches. Legends for all figures should be typewritten on a
separate sheet.

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N, J.

GENETICS, JULY 1919

TABLE OF CONTENTS

PAGE

Johnson, James, The inheritance of branching habit in tobacco. . . . 307

East, E. M., Studies on self-sterility. III. The relation between self-
fertile and self-sterile plants 341

East, E. M., Studies on self-sterility. IV. Selective fertilization. . 346

East, E. M., Studies on self -sterility. V. A family of self-sterile

plants wholly cross-sterile inter se 356

Jones, D. F., Selection of pseudo-starchy endosperm in maize 364

STUDIES ON SELF-STERILITY. IV. SELECTIVE
FERTILIZATION

E. M. EAST

Bussey Institution, Harvard University, Forest Hills, Massachusetts

Reprinted from Genetics 4 : 34^-355> July, 1919

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle

Harvard University

Edwin G. Conklin

Princeton University

Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Michigan

Edward M. East

Harvard University

Rollins A. Emerson
Cornell University

Herbert S. Jennings

Johns Hopkins University

Thomas H. Morgan

Columbia University

Raymond Pearl
Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 1916, at the post office at
Princeton, N. J., under the act of March 3, 1879.

STUDIES ON SELF-STERILITY. IV. SELECTIVE
FERTILIZATION

E. M. EAST

Bussey Institution, Harvard University, Forest Hills, Massachusetts

Reprinted from Genetics 4 : 346-355> July, r9i9

STUDIES OX SELF-STERILITY. IV. SELECTIVE FERTILIZA-
TION

[Received II irch 31, 1919]

TABLE OF CONTEXTS

PAGE

Introduction

Discussion of the problem

Pollen-tube frequency distributions

346
343

349

Influence of growth of compatible pollen tubes upon incompatible pollen tubes 352

Selective fertilization has been evoked many times as a means of ac-
counting for peculiar or unusual breeding results. Castle's (1903)
original theory of sex-determination, and Cuenot's (1908) interpreta-
tion of the non-appearance of homozygous yellow mice, are examples.
Fortunately, it has always been possible to explain matters without re-
taining the hypothesis; in many cases, in fact, direct proof has been avail-
able that selective fertilization does net occur. Nevertheless, selective
fertilization as a contingency has remained a sort of nightmare to in-
vestigators in genetics. Such antipathy is not unnatural, but one must
have in mind the changes which have taken place in the subject during
the last decade, to understand clearly the reason.

Mendel's discoveries, the Laws of Segregation and of Recombina-
tion, made heredity enticingly simple. All extensions, additions and ex-
ceptions have tended toward complexity. In this genetics has but re-
peated the history of chemistry and physics, yet it is to be expected, per-
haps, that any suggested change in genetic conceptions savoring of in-
creased complexity should find favor slowly. And selective fertilization
is a tenet which would increase the difficulties of the subject a hundred-

The outgrowth of Mendelism has been a theory of inheritance founded
on the conception of specific character determiners, genes, located in the
chromatin. In the sense that the central problem of heredity is clearly

Gexettcs 4: 6161 if 9t£

Summary and discussion
Literature cited

354
355

INTRODUCTION

fold.

SELF-STERILITY AND SELECTIVE FERTILIZATION

347

one of chromosome, or at least of chromatin, distribution, the modern
generalization has a simple grandeur not found in early Mendelism; but
this simplicity is quite delusive, as a short consideration shows.

The conception of the gene is unquestionably the foundation of ge-
netics. Students of heredity have submitted good evidence that characters
are the product of many relatively stable genes which have a real basis in
the germplasm, and that each of these genes may be the cause of various
effects in different parts of the organism. They have shown that while
the effects of a particular gene may not be wholly the same under dif-
ferent environmental conditions, nevertheless neither changes in the fac-
tors of environment nor association in particular combinations in the
germplasm serves to change their individuality or constitution with a
significant frequency.

Heredity, then, is the distribution of genes, and the genes have been
located definitely in the chromosomes. Fortunately, chromosome distri-
bution has been standardized in a remarkable manner in the majority of
plants and animals; hence, the greater part of the phenomena found in
breeding experiments may be described by a comparatively few simple
mathematical formulae. It is to this orderly chromosome distribution
that one must impute the utility of the Mendelian nomenclature, for to
it in large measure is due the regularity with which certain ratios recur.
There are irregularities in chromosome distribution, it is true. They
have even furnished some of the critical tests of the modern theory of
heredity taken as a whole. But because they curtail the practical value
of the theory through limiting the possibilities of prediction, it is well
that they are rare.

The standard chromosome mechanism for distribution of genes is that
in which homologous chromosomes mate at synapsis, and komologous
genes, one from either parent, pass by chance to either pole of the mitotic
figure, in the formation of the mature gametes. The chromosomes may
separate without having exchanged genes, presumably ; or, genes may be
exchanged. Just how this interchange occurs is not wholly clear. Mor-
gan has assumed that the genes have a linear arrangement, and that
there must be transverse breaks in the chromosomes. Castle (1918)
believes the arrangement is not linear, and that breaks may occur in
many ways. It is possible that neither assumption is correct. The writer
has felt for some time that possibly the genes are arranged spatially in a
manner somewhat analogous to that assumed by chemists for organic
molecules, though perhaps it might be better to say in a manner an-

Genetics 4: Jl 1919

348

E. M. EAST

alogous to certain crystals, for there certainly is no evidence that the
genes are radicles belonging to single molecules. But the point is that
with a spatial arrangement similar to that assumed for the radicles of
molecules, with the homologous chromosomes mirror images of each
other, with homologous genes interchanging by a definite mechanism,
a more delicate system of action is possible than with mere chromosome
breaks.

However this may be, the hinge on which the usefulness of this whole
scheme turns is that the genes pass to either daughter cell by chance, and
that the gametes thus formed mate by chance.

Even when such inheritance obtains, selective elimination of both
gametes and zygotes is somewhat common, and causes rather chaotic con-
ditions wherever it occurs. For example, the difficulties which charac-
terize all endeavor to analyze inheritance in the Oenotheras are probably
due in large measure to this cause. The additional difficulties which
would arise should it be found that there is selection of genes in gamete
formation, and selection of gametes at fertilization are so great as to be
hardly imaginable.

DISCUSSION OF THE PROBLEM

Particularly suitable material with which to test the second possibil-
ity is found in those plants which are self -sterile. Since the direct cause
of self -sterility is the slowness of growth of self pollen tubes as com-
pared with cross pollen tubes, it would seem as if selective fertilization
would have a better opportunity to manifest itself under such circum-
stances than under those which obtain in self-fertile plants and in animals.

Experiments with the self -sterile species Nicotiana Forgetiana, N.
alata and N. angusti folia have shown that in self-pollinations and in in-
compatible cross-pollinations the pollen grains germinate as well as in
compatible cross-pollinations. No differences are to be found between
the two types either as to the percentage of grains germinating, the
length of time required for germination, or the size of the tubes after
germination, provided pollen tubes of the same length are measured.
Pollen tubes produced after self-pollination or after incompatible cross-
pollination grow so steadily that length plotted against time is a straight
line; but pollen tubes produced after a compatible cross grow at such a
constantly increasing rate that the growth curve resembles that of an auto-
catalytic reaction. As the flowering season is about to come to an end,
more rapid pollen-tube growth occurs after a self-pollination or an in-

SELF-STERILITY AND SELECTIVE FERTILIZATION

compatible cross-pollination, though there is little evidence of the acceler-
ated growth characteristic of compatible combinations. The pollen tubes
grow more rapidly, but the curve by its constant velocity still resembles
the curve of a "normal" self-pollination.

These facts are the basis of our problem, and naturally they suggest
the possibility of selective fertilization. Part of the work reported in
the first of these studies (East and Park 191 7) was done upon a cross
between N. Forgetiana and N. alata. The segregating generations nat-
urally contained numerous individuals heterozygous for a large number
of hereditary factors. There were differences in height of plant, size
of leaf, color of flower, and size of flower, differences which could hardly
be interpreted as the result of less than twenty or thirty determiners un-
less a great many of the variations shown in different organs were due
to the activity of a single gene. Similar hereditary differences were
marked even in the so-called pure species. This being true, it is im-
portant to know whether pollen tubes whose nuclei carry certain de-
terminers grow faster than those which carry other determiners.

POLLEN-TUBE FREQUENCY DISTRIBUTIONS

One method which throws some light on the probability of selective
fertilization is that of studying the frequency distribution of the pollen
tubes after pollination. When applications of pollen are made, and the
pistils prepared, sectioned and stained at varying periods of time after
pollination, similar results are obtained no matter what the type of com-
bination has been. In table 1, for example, a few frequency distributions
of pollen-tubes after self-pollinating self-sterile plants during the height
of the flowering season, are given. In general they are minus skew, and
show that the greater number of pollen tubes are grouped at points from
3.5 mm to 7.5 mm from the end of the stigma at the expiration of from
5 to 7 days after pollination. A number of tubes have pushed out ahead
of the majority, and a great many polien grains — from 5 to 25 percent
— have not germinated at all.

Distribution of pollen tubes in sections of pistils from self-sterile
plants which had been pollinated at the end of the flowering season, show
practically the same thing (table 2). The pollen tubes have reached dis-
tances comparable to those shown in table 1 in a shorter period of time,
but otherwise no marked difference can be seen.

These two tables are presented merely for comparison with tables 3
and 4.

Genetics 4: Jl 1919

350

E. M. EAST

Table i

Frequency distribution of pollen tubes after self -pollinating self-sterile plants during
tl\£ height of the flowering season.

, Distance from the stigma in millimeters

rce ot

data

■ 'j

4. ^

6.5

8.5

0 "

10 ^

IIJ5

12 ^

1 section after
5 davs

I section after

i section after
7 davs

I section after
7 days

Table 2

Frequency distribution of pollen tubes after self -pollinating self -sterile plants at the

end of tlie flowering season

data

1-5

2.5

3-5

4-5

5-5 16.5

7-5

8.5
1

95
1

10.5

12.5

i3o

14.5

1 section after
3 days

3
6

1 section after
3 days

1 section after

1 section after
4 davs

1 section after

In table 3 some distributions of pollen tubes from the F2 generation of
a cross between N. Forgetiana and A', alata are given. The cross is com-
patible, and since the individuals are unquestionably heterozygous in a
large number of factors, they should show a marked tendency to vary if
there is selective fertilization. The frequency distributions shown in
table 4, on the other hand, where sib matings for three generations ought

SELF-STERILITY AND SELECTIVE FERTILIZATION 35]
Table 3

Frequency distribution of pollen tubes after cross-pollinating compatible plants of F,
generation N. Forgetiana X N. alata.

Distance from tht stigma in millimeters

data

10.5

II-5

12.5

13.5

14.5

15.5

16.5

17.5

18.5

19.5

20.5

21.5

1 section after

1 section after
3 days

Table 4

Frequency distribution of pollen tubes after cross-pollinating compatible plants of F5
generation (sib matings) N. Forgetiana X alata

Source of

Distance from the stigma in

millimeters

data

10.5

12.5

13-5

14.5

15-5

16.5

17.5

18.5

19.5

20.5

21.5

1 section after

1 section after
3 days

1 section after

to have brought about a considerable degree of homozygosis, should be
less variable. As a matter of fact, however, there seems to be no signifi-
cant difference in the two cases.

There is no evidence that variability in gametic constitution is the
cause of variability in rate of pollen-tube growth. In fact, there is no
positive proof that there is a measurable variability in pollen-tube growth.

Genetics 4: Jl 1919

352

E. M. EAST

In both of these types of pollination and in all similar cases examined,
percentages of ungerminated pollen grains comparable to those deter-
mined for incompatible matings were found. The actual percentages
have little meaning, for ungerminated pollen grains are loosely held by
the stigmas and the correct number of grains which do not germinate is
not likely to be obtained. But the fact that a considerable percentage
of grains which contain protoplasm and in every respect seem to be nor-
mal, remain as long as 6 days without germinating, leads one to believe
that difference in the rate of germination is largely responsible for the
varied length of the pollen tubes measured. The pollen grains may
differ among themselves in the thickness of their walls or the composi-
tion of the protoplasm outside the nuclei, thus accounting in some
measure for rapidity of germination, without it being necessary to
assume gametic differentiation as a cause. Furthermore the entire
series of results on the behavior of self-sterile plants reported in the first
paper of this series (East and Park 191 7), makes it unlikely that differ-
ences in gametic composition show themselves in any way before fertili-
zation. The factorial composition of the mother plant controls the be-
havior of self-sterile plants, and all the pollen grains of a single plant
may be taken to have the same factorial composition as far as any func-
tions to be performed before fertilization are concerned.

It is not to be supposed that the variability in length of pollen tube
shown in tables 3 and 4 really represents the difference of time at fertili-
zation. In compatible matings the pollen tubes grow faster and faster so
that the variability shown in a frequency distribution of pollen tubes de-
termined at 1 day or 2 days after pollination may be quite different at a
later date. It has not been found possible to obtain satisfactory measure-
ments of pollen tubes as they approach the micropyles, but it may be as-
sumed that at this time the rate of growth is so fast that practically all
of the ovules are fertilized within a few hours. Selective fertilization is
hardly probable therefore for this additional reason.

INFLUENCE OF THE GROWTH OF COMPATIBLE POLLEN TUBES UPON IN-
COMPATIBLE POLLEN TUBES

In interpreting the results of our experiments on pollen-tube growth
(East and Park 1918), it was assumed that after a compatible cross
substances are secreted in the pistil which accelerate the elongation of the
tube, and that the immediate cause of this secretion is a catalyser which
the pollen-tube nucleus is able to produce because the hereditary consti-

SELF-STERILITY AND SELECTIVE FERTILIZATION

tution of the plant producing it is different from that of the plant on
which it is placed. Superficial consideration might lead one to suppose
that if this were true, incompatible pollen tubes would be accelerated by
the growth of compatible pollen tubes if a mixture of the two kinds of
pollen were placed on the stigma. Second thought, however, shows that
this is probably not the case. Plant enzymes are colloids having large
molecules, hence they do not pass freely through cell membranes. Their
actions are largely local ; where they do not seem to be local, the direct
cause of the reaction is more likely to be a crystalloid produced by ac-
tion of the colloid.

The writer has been able to devise no experiment to measure absolutely
such possible stimulation, but two experiments have shown that when
mixtures of compatible and incompatible pollen are applied to a single
stigma, only the compatible pollen produces seed.

In the first experiment a number of pistils were pollinated with a defi-
nite number of compatible pollen grains. The work was done under a
binocular, and the count is thought to be accurate within an experimental
error of ±2 grains. The pistils were then carefully covered with in-
compatible pollen. Eight capsules matured with the results shown in
table 5.

Table 5

The effect of compatible pollen on the growth of
incompatible pollen tubes.

Pistil Number of compatible

No. pollen grains

1 5i

2 48

3 50

4 62

5 32

6 67

7 61

8 46

Number of seeds
produced

42
4i
49
23
58
54
40

The indications from this experiment are that no incompatible pollen
tubes contributed to the production of the seeds obtained; but of course
it is impossible to maintain that these tubes were not accelerated in their
growth to some degree.

In the second experiment, a more critical test of the matter was made.
Three pistils of a white-flowered self-sterile plant coming from a line of

Genetics 4: Jl 1919

354

E. M. EAST

plants homozygous for this color were selfed. Five or six hours after
these plants were covered with pollen from a self-sterile family bearing
red flowers. Capsules full of seed were obtained. If these seeds were
produced by the compatible pollen only., the resulting progeny should be
red-flowered for red is dominant; if incompatible pollen has functioned,
white-flowered plants should be obtained. Three hundred plants have
been grown with not a single white-flowered individual.

SUMMARY AND DISCUSSION

The experiments described in this paper were designed to test the
possibility of selective fertilization occurring in self-sterile Xicotianas,
it being assumed that from the nature of the material the phenomenon
might here be possible. ( i ) Comparisons were made between the pollen-
tube frequency distributions of highly heterozygous and of comparatively
homozygous plants. (2) The influence of compatible matings on in-
compatible matings was investigated. In neither case was there any
indication of selective fertilization.

Though it is impossible to prove a negative, there is so much circum-
stantial evidence against selection both in the formation of gametes and
zygotes, the probability that it ever occurs is very remote. In the first
place gametes are formed in many animals and plants, particularly in
species crosses, which can never function. If the mechanism of gamete
formation were such as to make it necessary to assume a selection of
genes, a low frequency of non-functional gametes would be expected.
Similarly zygotes are produced in the numbers to be expected by chance
mating of gametes, even though these zygotes have no possibility- of
passing through a complete life cycle. There are two cases in mice, eight
in Drosophila, and four in plants where the evidence of lethal factors
is too complete to be disregarded. In reality there are probably hun-
dreds of such instances in plants and animals which have been investi-
gated during recent years that ought to be interpreted in the same man-
ner.

Again, pollen grains show no tendency to behave as if the genes which
they carry function before fertilization. It will be recalled that Batesox
(1909) found pollen shape and color in the sweet pea to be inherited as
a maternal character. The writer (East 191 6) has corroborated this
discovery for color of Xicotiana pollen. It may be claimed, however,
that these facts are just what is to be expected because of the morpho-
genesis of the outer characters of the pollen grain. This is true; but

SELF-STERILITY AND SELECTIVE FERTILIZATION'

355

the criticism does not apply to the phenomena found in the behavior of
self-sterile plants in cross matings where cross-sterility of groups of
plants exists presumably because of genes possessed by the mother plants.
In fact the only activity shown by a male gametophyte which seems to
be due to the factors it is carrying over into the next generation, is a lack
of any activity. In Belling's (1914) work on the velvet bean, he found
50 percent of the Fx pollen was abortive in a certain cross. It appear^
then that in this instance the presence or absence of a gene of the gener-
ation which would ordinarily function after fertilization, has caused the
pollen grain to abort. This lack of ability to function does not neces-
sarily mean the actual activity of the genes of this generation however;
the machine has simply remained uncompleted, so to speak. For this
reason, there seems to be no wisdom in even suspecting selective fertili-
zation; unless mixtures of pollen (or spermatozoa even) from different
individuals should be used. If pollen grains from a single plant are
alike as far as their activities before fertilization are concerned, there
is no basis for selection.

May wre not extend this conception to animals for the present and ac-
cept as a fundamental genetic hypothesis the tenet of chance segregation
in the germ cells and chance mating of these germ cells?

LITERATURE CITED

Batesox, W., 1909 Mendel's principles of heredity, pp. 396. Cambridge, England:
Cambridge University Press.

Belling, J., 1914 The mode of inheritance of semi-sterility in the offspring of certain
hybrid plants. Zeitschr. f. ind. Abstamm. u. Vererb. 12:303-342.

Castle, W. E., 1903 The heredity of sex. Mus. Comp. Zool. Bull. 40 : 189-218.

Castle, W. E., 1919 Is the arrangement of the genes in the chromosome linear?
Proc. Xat. Acad. Sci. 5 : 25-32.

Cuexot, L., 1908 Sur quelques anomalies apparentes des proportions Mendeliennes
(6e note). Arch. Zool. Exp. et Gen. 9:7-15.

East, E. M., 1916 Inheritance in crosses between Nicotiana Langsdorffii and Nico-
tiana alata. Genetics 1:311-333.

East, E. M., and Park, J. B., 1917 Studies on self-sterility. I. The behavior of self-
sterile plants. Genetics 2 : 505-609.
1918 Studies on sterility. II. Pollen-tube growth. Genetics 3 : 353-366.

Genetics 4: Jl 1919

. Information for Contributors

No limitation will be placed on the kind of illustrations to be used, but as funds for
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4H X inches and all illustrations intended for use as single plates should not
exceed 5 X 7H inches, or should allow of reduction to these dimensions. Double
plates may be 12H X 7Va inches. Legends for all figures should be typewritten on a
separate sheet.

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, JULY 1919

TABLE OF CONTENTS

PAGE

Johnson, James, The inheritance of branching habit in tobacco. . . . 307

East, E. M., Studies on self-sterility. III. The relation between self-
fertile and self-sterile plants 341

East, E. M., Studies on self-sterility. IV. Selective fertilization. . 346

East, E. M., Studies on self -sterility. V. A family of self -sterile

plants wholly cross-sterile inter se 356

Jones, D.*F., Selection of pseudo-starchy endosperm in maize. . . . 364

STUDIES ON SELF-STERILITY. V. A FAMILY OF
SELF-STERILE PLANTS WHOLLY
CROSS-STERILE INTER SE

E. M. EAST

Bussey Institution, Harvard University, Forest Hills, Massachusetts

Reprinted from Genetics 4:356-363, 1919

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

William E. Castle
Harvard University

Edwin G. Conkun

Princeton University
Charles B. Davenport

Carnegie Institution of Washington

Bradley M. Davis

University of Michigan

Edward M. East
Harvard University

• Rollins A. Emerson
Cornell University

Herbert S. Jennings
Johns Hopkins University

Thomas H. Morgan
Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J., under the act of March 3, 1879.

STUDIES ON SELF-STERILITY. V. A FAMILY OF
SELF-STERILE PLANTS WHOLLY
CROSS-STERILE INTER SE

E. M. EAST

Bussey Institution, Harvard University, Forest Hills, Massachusetts

Reprinted from Genetics 4:356-363, 1919

STUDIES OX SELF-STERILITY V. A FAMILY OF SELF-
STERILE PLAXTS WHOLLY CROSS-STERILE INTER SE

E. M. EAST

Harvard University, Bussey Institution, Forest Hiils, Massachusetts
[Received June 26, 1919]

In the first paper of this series (East and Park 191 7), 1 the behavior
of a number of families of self -sterile plants under various schemes of
mating was described. In one cross between the two species Nicotiana
Forgctiam and Xicotmna alata, fifty-three plants of the Ft generation
could be separated into not less than four groups in which each member
of every group was cross-sterile with every other member of that group,
although showing cross- fertility with every member of every other
group. The interpretation given of these and other similar facts was in
brief as follows: (1) a self-sterile species exhibits this peculiarity be-
cause homozygous for a basic self-sterility factor; (2) a series of par-
tially coupled factors affects the behavior of sterile plants among them-
selves; (3) these secondary factors act as if sporophytic in nature, so
that all gametes produced by a single individual are identical in this re-
gard with the plant on which they originated; (4) the nature of the ac-
tion of these secondary factors is such that two plants are not fertile to-
gether unless they differ by at least one of these factors.

Though the self-sterility and the cross-sterility existent in these plants
is fully expressed at the beginning and height of the flowering season,
toward the close of the flowering season, particularly in plants exhibiting
the effect of adverse environmental conditions, occasionally some fertil-
ity is shown. It is possible therefore to obtain seed from truly self-
sterile plants and from combinations that are fundamentally incompati-
ble. By taking advantage of this pseudo-fertility it should be possible to
obtain families of plants wholly cross-sterile inter sc. Such a family,
apparently, is family E described on pages 565 to 567 of the paper we
are discussing. Unfortunately very little work had been done on this
family when that paper was written. Since a reserve supply of the seed

1 East, JE. M., and Park, J. B. Studies on self-sterility I. The behavior of self-
sterile plants. Genetics 2 : 505-609, 1917.

Genetics 4: 356 Jl 1919

SELF-STERILE PLANTS CROSS-STERILE INTER SE

357

from which the family came had been preserved, however, it has been
possible to make a more extended study of the case.

In table i are recorded the infertile crosses made upon 54 plants of

Table i

Record of infertile crosses made on 54 plants of family E(2). Number of pollinations

is shown by subscripts.

Ped. No.!

Sterile with Ped. No. S

Sterile with Ped. No. $

33, 53, io3, 5O3, 75

444,573, 593, 6i3, 724, 733, 743

55, 83, I46, 232, 503

*«■ 53. 69,

33, 94, 10., I43, i63, 233, 503, 742

*3, 35, 543, 553, 5<V 573, 594, 6o3, 61, 747

93, io3, I43, 163, I73, 506

io„, I43, 16 , 17,, 23 , 50

0 0 0 0 0 0

54> 83, 6i3

i4o, i7o, 443, 503

0 0 0 0

J3' 55, 83, 93, 453, 583, 6i3, 673, 686

i6_, i7„, 19 , 26 , 43 , 50o, 75

36> 53> 83, 93, io3, 502, 734

I72, 19,, 4iv 42s, 44o, 508

53i 83, 93, I43

22., 23., 43„, 44„, 50,, 55„

83, 93> io3, I43, i62

23^, 26., 29„, 50,, 61,, 62,, 69., 75,
^4 4 ^3 3 3 3 4 0

i4o, 16,

26., 27., 31^ 50„, 61,, 62„, 69.

3 3 4 3' 3 0 0

17,
' 3

264, 4i6, 426, 442, 503, 563

32, 53, 93. i73, I94, 533> 544, 568, 633, 643, 666,
72o» 73 0

' 3' ' "3

3J^» 35„> 38Q, 50Q, 6i„, 62,

«^ 4> ^ 3 0 3 3 3

I4Q, 19^ 22„ 23^, 42,, 44„ 49,, 55,

~3 4 3 "4 ^ 3 ■'3 ' -^3 ^^3

^y3, jo3» i><->3, uA4. "^3» /o3

313. 333, 383, 503 6i4, 623, 683

I93, 273, 754

333> 354, 383, 5O3, 6i4, 62,, 733

224, 264, 293, 423, 433, 753

353, 386, 397, 407, 4i7, 503

293, 3i3, 384, 753

383, 393, 404, 423, 503, 6i3, 693, 75:

264, 273, 3i4, 333, 683, 7o3, 7i3, 733

334, 393, 404, 423, 443, 5<>4, 6i,8
624, 734

,263, 29s, 3i3, 333, 336, 353, 523, 753

404, 423, 433, 507, 6o3, 6i3, 623, 75,

, 337, 353, 383, 443, 493, 503, 523, 63,, 64,

423, 444, 453, 503, 524, 6i4. 623, 63

,337i 354, 384 , 394

Genetics 4: Jl 1919

43
44

45
46
49
50

5i
52

55
56

57
58
59
6o
6i

E. M. EAST

Table i (continued)
>f infertile crosses made on 54 plants of family E (<?)• Number of pollina-
tions is shown by subscripts.

Sterile with Ped. No. $

443, 463, 523, 533, 543, 74,

263, 313, 443, 503, 75,

3i,. 4ia, 44,, SO., 59-

i4, 263, 393, 463, 493, 506, 52., 53,
544

I03> 443, 463, 503, 6i4

263, 393, 453, 504, 5i3, 62s

I42, 393, 4413, 523, 534 543, 56,

Sterile with Ped No. 9

i63, 236, 337,

433

163, 236, 353,

383,

393, 403, 434

I43, I73, 393,

I(V I(V *7„
5°i3' 523

232,

383, 4o4, 4i3, 423, 433, 453,

403, 493

4I3, 443, 453

443

i3, 33, 53, 86, 93, io3, I49, i68, I73, io8, 223,

233, 263, 273, 293, 3I3, 333, 353, 38^, 2>9V 403,

423, 433, 446, 453, 494, 523, 533, 546, 553, 563,

574, 583, 593, 6o3, 6i3, 63s, 643, 65,, 663, 67,,
683, 69s, 703, 7i3, 723, 743, 753

49,

383, 393, 443, 503, 534, 545, 554,
573, 583, 593, 684

4o4, 4I3, 445, 503

233, 503, 544, 553, 563, 573, 584, 623,
64,

53, 234, 506, 553, 564, 573, 583, 64.

53> 263, 503, 563, 573, 58t, 593, 624

57, 238, 5O3, 573, 583, 593, 6o3, 6i3,
634, 642

i3, 53, 504, 583, 594, 6o3, 6i3, 69,

io3, 5o3, 593, 6o5, 6i8, 65,, 733, 744

413' 443, 504, 524

413, 444, 503, 525, 534

i73, 524, 533, 54c

233, 503, 523, 533, 544, 553

523, 533, 543, 553, 563

523, 534, 543, 554, 563, 57,

i3. 54, 503, 6o3, 6i8, 7i3, 733, 75,

436, 523, 553, 563, 574, 58,

53, 503, 6i3, 693, 7i3, 754

393, 563, 573, 585, 59s

l3, 5, 93, io3, 434, 503, 693, 7i4 J93, 223, 263, 27,, 294, 3i4, 353, 383, 393, 404,

|454, 563, 573, 583, 593, 6o3, 633, 643, 653, 663,
743, 754

I0s» 223> 263' 273> ^ Sis* 384, 393, 403, 49;j,
533, 554, 633, 643, 693, 7i4, 723> 733, 754

SELF-STERILE PLANTS CROSS-STERILE INTER SE

Table i (continued)

Record of infertile crosses made on 54 plants of family E (-?). Number of pollina-
tions is shown by subscripts.

Ped. No

Sterile with Ped. No. $

Sterile with Ped No. 9

233, 393> 5O3, 6i3, 623, 643, 653, 743>
753

404, 564, 733, 744

233> 393> 5<>3, 6i3, 623, 673, 683, 734,
753

533, 545, 562, 63 653, 66 , 69,, 70,, 7i4, 733

i3, 503, 6i3, 643, 683, 6q3, 702

584, 633

236, 5O3, 61 3, 643, 7I31 733, 744

672, 683

io3, 503, 672, 693, 7o3, 7i3

643

io6, 353, 503, 663, 693, 703

293, 52^, 643, 653, 693

33, 503, 623, 643, 683, 733, 743> 753

i94, 223, 353> 574, 6o3, 6i3, 653, 673, 683, 70,,
7i6, 753

35 50 , 64 , 69., 71,, 723, 734, 743

652> 673, 683

353, 503, 624, 644, 696, 723, 733, 743

593, 6o3, 6i4, 663, 673, 703, 724

*4« 233' 503, 623, 7i4, 733, 746

703, 7i3, 734

. 73

i3, I44, 233, 353, 623, 63s, 64s, 724

3*3' 384, 583, 593, 644, 663, 693, 7<>4, 7i3, 723,
744

i3, 57, 503, 6i3, 634, 734, 753

52, 4i3, 584, 633, 664, 69,, 7<>3, 7i3, 72c

294, 3i3, 333, 383, 503, 6i4, 624, 69,

1, I43, I93, 273, 353, 394, 423, 593, 6o4, 63.,
64,, 693, 743

this family, the subscripts showing the number of attempts made for
each combination. The table was constructed by assuming that if a com-
bination had been made one way, the reciprocal had also been made, as
explained in our former study (East and Park 191 7). Thus it can be
seen that while only a fraction of the possible combinations were made,
nevertheless the plants were linked together in an unbroken chain. In
other words, if it be true that when A is sterile with B and with C, B is
sterile with C, then each of these 54 plants is sterile with the other.

It is not true however that no seed at all was obtained in the nu-
merous attempts to combine plants of this family. Table 2 shows that
13 combinations produced capsules. From the number of sterile pollin-
ations made with the same plants and from the fact that nearly all of the
fertility appeared at the end of the flowering season, it would seem that

Genetics 4: Jl 1919

360

E. M. EAST

Table 2

Record of fertile crosses made on 54 plants of family E (<?), —
presumably pseudo-fertility. First number is female.

Number of

Combination

fertile

sterile

sterile reciprocal

pollinations

3X5

5 X 10

5 X 74

See 74 X 5

16 x 50

17 X 22

23 X 44

33 X 38

39 X 50

44 X 50

See 50 X 44

44 X 52

50 X 44

See 44 X 50

56 X 23

74 X 5

See 5 X 74

these apparent exceptions are all illustrations of fluctuating pseudo-fer-
tility. There is the whole of our experience with this type of fertility
back of such an assertion, but there is also some specific evidence on the
case in point.

The number of seeds produced by these plants when crossed with
compatible pollen is in general from 300 to 600 per capsule (table 3),
while the number of seeds in the presumably pseudo-fertile combinations
is usually much less. At the same time 4 of the latter combinations pro-
duced what seemed to be full capsules. Combination 16X50 produced a
full capsule at the seventh attempt, although eight out of nine attempts
were failures, and combination 50 X 44 produced a full capsule at the
twelfth attempt although thirteen out of fourteen attempts were failures.
On the other hand plant 23 gave a full capsule with pollen of plant 44
on the first attempt, plant 33 gave two capsules out of eight attempts with
pollen of plant 38, and plant 39 gave four capsules out of eleven attempts
with pollen of plant 50. Now combination 33 X 38 was about 50 per-
cent fertile, and combination 39 X 50 became progressively more fertile
as shown by the number of seeds produced. These three plants, 23, 39,
and 50 were crossed with a large number of other plants, nevertheless,
and showed cross-sterility. Further, at the beginning of another flower-
ing season crosses 23 X 44 and 39 X 50 were impossible. At the same
time it is not without the bounds of probability that combination 39 X 50

SELF-STERILE PLANTS CROSS-STERILE I.XTER SE

361

Table 3

Comparison of the number of seeds in capsules of tiie presumably
pseudo-fertile combinations in family E{2) with the number
of seeds in the capsules of the same plants when pollin-
ated with pollen from the plant of the F2 generation
of the cross between N. Forgetiana and
N. Langsdorffii.

Xumber of

Number of

Combination

seeds in

seeds in capsules when pollin-

capsules

ated with (lid y 128)

3 X 5

5 X 10

115

365, 382, 421

5 X 74

127

365, 382, 421

16 X 50

418

17 X 22

3/0

23 X 44

436

33 X 38

124, 131

39 X 50

151, 255, 289, 352

44 X 50

128

44 X 52

191

50 X 44

427

56 X 23

74 X 5

185

was for some unknown reason more easy to make than other combina-
tions in this family. We have no theory to offer at present as to why
this may be true. It may stand as an open question. The general con-
clusion from all the evidence is that family E (2) may be considered to
consist of plants wholly cross-sterile inter se.

The question then arises: Is the origin of family E (2) compatible
with our previous conclusions as to the behavior of self-sterile plants
when crossed inter se. First, it must be emphasized that the cross-
sterility found has nothing to do with true sterility. A random sample
of 25 plants was used in a test with the pollen a single plant coming from
the F2 generation of a cross between Xicotiana Forgetiana and Xicotiana
Langsdorffii. Out of 64 pollinations there were only 2 failures (table
4). Again, out of 51 attempts to use the pollen of these plants in crosses
thought to be compatible, there was only 1 failure. The sterility found,
therefore, is wholly of the nature termed "self-sterility," or "incompati-
bility," and must be interpreted as such.

The origin of a family consisting of one class of plants cross-sterile
with each other was to have been predicted on the basis of the interpre-

Gexetics 4: Jl 1919

362

E. M. EAST

Table 4

Record of pollinations made on a random sample of 25 plants of family
E{2) with pollen from a single self -sterile plant of the F2
generation of a cross between Nicotiana Forgetiana
and Nicotiana Langsdorffii (814 X 328).

Successful

Unsuccessful

Number of seeds in

Plan* "NTn

1 lani in 0.

pollinations

each mature capsule

ir\m OXO i/JT

3°o» 3o2» 421

300, 330, 5i0

T
1

3yu, ^UU, 4UO, J/0

t r»

T
1

03 1

T *7

T
1

3/°

3^o, 4O3, 230

0 -

T
X

'2*77

37/

°30> 420

481, 404, 470

4o3> 3:A 40^, 372, 340

432, 192

271, 195, 3*7

105

225, 252, 382, 384

157

420, 177

295, 650, 250, 462

700, 618, 902

327

330

635, 588, 580, 468, 678

338, 230, 376

230, 176

480, 240

291, 358

Total

tations we have used, by taking advantage of the phenomenon of pseudo
self-fertility. Continued self-fertilization is possible by persistent ef-
forts at self-pollination carried to the very end of the flowering season.
And continued self-fertilization should bring about homozygosis in the
secondary factors affecting the behavior of self-sterile plants among
themselves. When such a point is reached, the resulting population
should not only be self-sterile but should belong to a single class all mem-
bers of which are cross-sterile with each other.

Family E (2) was not the result of continued self-pollination and
pseudo self-fertility, however. It was produced as follows : The fe-

SELF-STERILE PLANTS CROSS-STERILE INTER SE 363

male parent was No. 58, a plant of N. alata, the result of three generations
of selling a self-sterile strain at the end of the season. The behavior of
No. 58 and of its sister plants when crossed with each other leads one
to believe they were all members of one intra-sterile class, but the evi-
dence is hardly sufficient to establish the point. The male parent was
a member of the Fx population (plant 44, loc. ext., p. 559) produced by
crossing a self-sterile plant of N. Forgetiana with a sister plant of the
mother of the N. alata plant just described (No. 58).

Now it is obvious that the female parent of this family may have come
from a fraternity homozygous for the secondary factors effective on
compatibility inter se. They may have been, for example, plants with
the composition AABB. It is possible also that the male parent, though
originating from a cross, might have had the formula AABB, since its
parents might have been AABB and AaBB. But a whole population
having a single formula could not have arisen through a cross except
through an illegitimate mating (pseudo-fertility). If then the two par-
ents of the population had the same constitution and produced seed
through pseudo-fertility, then family D {loc. ext., p. 563), coming from
the same male crossed on a sister of plant 58, ought to be a duplicate of
family E (2). This however does not appear to be the case. Family D
consisted of at least two intra-sterile classes, unless a good deal of un-
recognized pseudo-sterility was present. On the other hand both of the
parents, in the few tests made on family E (2), were sterile with their
progeny, — a result to be expected on the theory of homozygosis. It
seems, then, that the unsettled question, a question which must await
further investigation, is, why family D and family E (2) are not similar
in composition and behavior.

Genetics 4: Jl 1919

Information for Contributors

No limitation will be placed on the kind of illustrations to be used, but as funds for
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4$i X 7JA inches and all illustrations intended for use as single plates should not
exceed 5 X 7Va inches, or should allow of reduction to these dimensions. Double
plates may be 12^ X 7$i inches. Legends for all figures should be typewritten on a
separate sheet.

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

.GENETICS, JULY 1919

TABLE OF CONTENTS

PAGE

Johnson, James, The inheritance of branching habit in tobacco 307

East, E. M., Studies on self-sterility. III. The relation between self-
fertile and self -sterile plants 341

East, E. M., Studies on self -sterility. IV. Selective fertilization. . 346

East, E. M., Studies on self -sterility. V. A family of self-sterile

plants wholly cross-sterile inter se 356

Jones, D. F., Selection of pseudo-starchy endosperm in maize 364

[Reprinted from The American Naturalist, Vol. LIL, June-July, 1918. J

THE ROLE OF REPRODUCTION IN EVOLUTION1

PROFESSOR E. M. EAST

BUSSEY INSTITUTION, HARVARD UNIVERSITY

The establishment of methods of reproduction which
maintain variation and inheritance mechanisms on a high
plane of efficiency is naturally a fundamental requirement
in organic evolution. Since, however, inheritance mech-
anisms presumably equivalent are common to every method
of reproduction, one should be able to interpret the evolu-
tionary tendencies in the matter by comparing their
effectiveness in offering selective agencies their raw ma-
terial. Some will hold this statement to be a self-evident
truth; others may maintain as strongly either that the
premises are wrong or that the conclusion is not justified
even if the premises be granted. Perhaps it is safer to
ply the middle course ; if the case is not so obvious as a
Euclidian axiom, as a compensation rigorous proof may
be less difficult.

As a basis for argument, let us sketch the general trend
of reproductive evolution in plants and animals.

Ordinarily, one speaks of two types of reproduction
among organisms, asexual and sexual. This is a conven-
tion that has taken on the dignity of a " folkway " among
biologists. Its employment should imply assent to the
4 proposition that the varied forms in which each of these
classes presents itself are inherently equivalent, and that

i Read by title at the Symposium of the American Society of Naturalists
on the subject 11 Factors of Organic Evolution/' Jan. 5, 1918.

273

274

THE AMERICAN NATURALIST [Vol. LII

the groups considered as units are fundamentally distinct,
but it is doubtful whether any such implication would be
admitted by the majority of its users. In fact one could
hardly maintain that simple division, sporification, the,,
production of gemmules, true budding, fragmentation
with regeneration of parts, and the various kinds of
apogamy and parthenogenesis on the one hand, and all
nuclear fusions on the other, can be grouped together as
if they are of the same evolutionary value, if this term be
used in any narrow or special sense ; but from a broader
viewpoint, the conventional classification has a real and
deep meaning which perhaps the biologist has grasped
instinctively.

There are both asexual and sexual methods of repro-
duction in nearly all groups of animals and plants ; among
animals the second has almost supplanted the first, among
plants the two have continued side by side. In neither
kingdom was sex developed as a more rapid means of
multiplication, since, as Maupas showed, a single infuso-
rian may become the progenitor of some 50,000 individuals
during the time necessary for one pair to conjugate.
Some other requirement was fulfilled; and fulfilled ade-
quately if we may judge by the number of times sexual
differentiation arose and the tenacity with which it was
retained.

Just when sexual reproduction first originated in the
vegetable kingdom is still a question. Among the lower
forms only the schizophytes, flagellates and myxomycetes
have passed it by. Perhaps it is for this reason that
these forms have remained the submerged tenth of the
plant world. It is tempting, as Coulter (1914) says, to
see sex origin in the Green Algse. There, in certain
species, of which Ulothrix is a good example, spores of
different sizes are produced. Those largest in size germi-
nate immediately under favorable conditions and produce
new individuals. Those smaller in size also germinate
and produce new individuals, but these are small and
their growth slow. Only the smallest are incapable of

Nos. 618-619] THE ROLE OF REPRODUCTION

275

carrying on their vegetative functions. These come to-
gether in pairs. Two individuals become one as a pre-
requisite to renewed vigor. Vegetative spores become
gametes. Something valuable — speed of multiplication
— is given up for a time that something more valuable in
the general scheme of evolution may be attained.

This is indeed an alluring genesis of sex. Let us use
the indefinite article, however; no doubt it is a genesis of
ssx, but it can hardly be the genesis of sex. Various mani-
festations of sex are present in other widely separated
groups of unicellular plants, the Peridineae, the Conjugate
and the Diotomea? — the Conjugate being indeed the only
great group of plants in which there is no asexual repro-
duction. In these forms one can not make out such a good
case of actual gametic origin, but the circumstantial evi-
dence of sex development in parallel lines is witness of its
paramount importance.

After the origin of sex, many changes in reproductive
mechanisms occurred in plants, but almost all of them
resulted merely in greater protection of the gametes, in
increased assurance of fertilization, or in provision for
better distribution. First there was a visible morpho-
logical differentiation of gametes, the one becoming a
large inactive cell stored with food, the other becoming
small and mobile. Then came the evolution of various sex
organs, and finally the alternation of generations. In the
higher plants a long line of changes have occurred con-
nected with the alternation of generations ; the spore-pro-
ducing type has developed from a form of little impor-
tance to that which dominates the vegetable world, the
garnet e-prQducing type has degenerated until it consists
of but two or three cell divisions. In these variations
there is reproductive insurance, something which also may
be said of those manifold adaptations which provide
zygotic protection either in the seed or the adult plant,
but they are no more direct changes in reproductive
mechanism than are the diverse means which arose to
secure dispersal. In fact in all of these changes no new

276 THE AMEBIC AN NATURALIST [Yol.LII

process of fundamental evolutionary significance oc-
curred, unless it be the various mechanisms devised to
promote or to insure cross-fertilization, and which may be
interpreted as variations tending to perfect sexuality.

Coincident with the general trend of plant evolution
just mentioned, two important changes in the nature of
retrogressions occurred, which have persisted in many
species. A new type of asexual propagation arose,
apogamy, which though it appeared under several guises,
apogamy in the narrow sense, parthenogenesis and poly-
embryony, is none the less asexual reproduction returned
under another name and apparently with no particular
advantages over the older types. Further, hermaphrodit-
ism was developed and has persisted in numerous lines.
AVe may be wrong in calling hermaphroditism a retro-
gression, for it has the great advantage of a certain
economy of effort in the production of gametes, but never-
theless it is certainly a change which per se is in the
opposite direction from that established when sex was
first evolved. A moment of consideration not only makes
this clear, but gives us a pretty satisfactory proof that
the gain made when continuous multiplication was halted
for a time by the intervention of a fusion at the genesis
of sexual reproduction was in some way connected with
the mixture of dissimilar gernrplasms. This conclusion
is hardly avoidable from the fact that although herma-
phroditism retained the cell fusion mechanism of gono-
chorism it was still necessary for Xature to evolve means
for cross-fertilization. And the multitude of ways in
which she solved this problem must mean that an im-
mense advantage was secured.

In spite of the great morphological differences between
animals and plants, the essential evolutionary changes
affecting reproduction in the two kingdoms have been
so similar as to be almost uncanny. Accepting the divi-
sion of animals into twelve phyla as recognized by many
modern zoologists (Parker and Haswell), one finds the
following facts regarding reproduction. Asexual repro-

Nos. 618-619] THE ROLE OF REPRODUCTION

277

duction in the narrow sense is common in Protozoa, Porif-
era, Ccelenterata and Platyhelminthes, and is sporadic in
Molluscoida, Annulata, Arthropoda and Chordata. If
fragmentation and regeneration be included, Echinoder-
mata and possibly Nemathelminthes are added. If
parthenogenesis is included, Trochelminthes is admitted.
Thus only the Mollusca have no form of asexual reproduc-
tion, and zoologists would hardly feel safe in maintaining
its absence there since the life history of so many forms is
unknown. This being the case, one must admit that
asexual reproduction has been found satisfactory for
most of the great groups of animals as far as actual
multiplication is concerned. For other reasons, however,
it evidently did not fulfill all requirements, since sexual
reproduction is established in every phylum. Further,
omitting the Protozoa in which it is difficult to decide such
sexual differences, gonochorism is present everywhere
except in the Porifera, and hermaphroditism everywhere
except in the Trochelminthes, although in Nemathel-
minthes, Echinodermata and Arthropoda it is rare.

Now if our conclusions regarding the true role played
by sex in evolution are correct, hermaphroditism is a
secondary and not a primitive phenomenon. In this we
follow Delage, Montgomery and Caullery rather than the
majority of zoologists. We believe it to be the only
logical view in spite of the fact that the Porifera, usually
considered so unspecialized, are all hermaphroditic.
Perhaps the Porifera are farther along in specialization
than is admitted, for to find the substance nearest chemi-
cally to the so-called skeleton of the sponges one must
turn to the arthropods (the product of the spinning glands
of certain insects). Hermaphroditism, therefore, as in
plants, is from this viewpoint a regression. And as in
plants it was not found adequate. In giving up diecism
for monecism, something was lost, and this something had
to be regained by further specialization. Hence, even as
in the vegetable kingdom one finds the essential feature of
bisexuality, mechanisms providing for mixtures of dif-

278

THE AMERICAN NATURALIST

[Vol. LII

ferent germplasms, restored by means of protandry,
protogyny or self-sterility.

In even such a brief consideration of the more im-
portant changes which have occurred in the reproductive
mechanisms of animals and plants, one thing stands out
impressively. Both animals and plants have adopted as
the most acceptable and satisfactory modes of reproduc-
tion, methods which are identical in what we deem to be
the essential features, something that can be said of no
other life process. These significant features are the
preparation of cells which in general contain but half of
the nuclear material possessed by the cells from which
they arise, which are differentiated into two general
classes that show attraction toward each other, and which
will fuse together in pairs to form the starting point of a
new organism. This parallel evolution is of itself valid
evidence of the importance of the process. Let us return
to our original proposition for its interpretation.

First, is there any evidence that sexual reproduction
differs from asexual reproduction in what may be called
the heredity coefficient? In other words, does one method
hold any advantage over the other as an actual means for
the transmission of characters? I have answered this
question in the negative, but it must be confessed that the
basis for this answer is a long and intimate experience in
handling pedigree cultures of plants rather than the study
of a large amount of quantitative data bearing directly
on the problem. Quantitative data are to be found, of
course, and plants furnish the best material because of the
ease in handling large numbers of both clons and seedlings
side by side; but even with the best of plant material,
several undesired variables are present. Practically the
inquiry must take the form of a comparison between the
variability of a homozygous race when propagated by
seeds and when propagated by some asexual method.
The first difficulty is that of obtaining a homozygous race
and thus eliminating Mendelian recombination. The
traditionally greater variability of seed-propagated

Nos. 618-619] THE ROLE OF REPRODUCTION

279

strains is due wholly to this difficulty, I believe. It may
be impossible to obtain a race homozygous in all factors.
There may be a physiological limit to homozygosis even
in hermaphroditic plants. The best one can do is to use
a species which is naturally self-fertilized, relying on con-
tinued self-fertilization for the elimination of all the
heterozygous characters possible. I have examined mam7
populations of this character in the genus Nicotiana and
have been astounded at the extremely narrow variability
they exhibit. Even though one can not grow each mem-
ber of such a population under identical conditions as
to nutrition, the plants impress one as if each had been
cut out with the same die. Qualitative characters such as
color show no greater variation, as far as human vision
may determine, than descendants of the same mother
plant propagated by cuttings. Further, in certain char-
acters affected but slightly by external conditions, such
as flower size, the sexually produced population not only
shows no greater variability than the asexually produced
population, but it shows no more than is displayed by a
single plant. Yet one must remember that in such a test
the seeds necessarily contain but a small quantity of
nutrients, and for this reason the individual plants are
produced under somewhat more varied conditions than
those resulting from cuttings, hence it would not have
been unreasonable to have predicted a slightly greater
variability for the sexually produced population even
though the coefficient of heredity of both were the same.

I have made similar though less systematic observa-
tions on wheat— an autogamous plant almost as satis-
factory for such a test as Nicotiana— with practically iden-
tical results. I do not know of any published data on the
subject, however, taken either from these or any other
plants. In fact, there are few other plants from which
data could be obtained with so little likelihood of experi-
mental error.

On the other hand, zoology has furnished a consider-
able amount of such evidence (cf. Casteel and Phillips,

280 THE AMEBIC AX XATUEALIS1 [Vol.LII

1903: Kellogg, 1906; Wright, Lee and Pearson, 1907).
One need only mention Kellogg 's work on bees as a type.
Kellogg assumed that if amphimixis were the principal
cause of the continuous variations postulated by Darwin
and Weismann as the most important source of material
for the use of the natural selection,2 parthenogenetically
produced individuals should be less variable than those
produced sexually. A statistical investigation showed,
however, that the characters of drones probably are more
variable than those of worker bees of the same race.
Since Kellogg believes Darwin's judgment that " males
vary more than females" to have been disapproved, he
concludes that " amphimixis is not only not necessary in
order to insure Darwinian variation, but there is no evi-
dence (that I am aware of) to show that it increases
variation. ' '

It is hardly necessary to point out here the numerous
mathematical and biological pitfalls which should be con-
sidered before one could accept as valid the statistical
differences that appear to exist when coefficients of varia-
tion based on such data are examined. It should suffice to
note that the researches of Wright* Lee and Pearson
(1907) on wasps of the species Vespa vulgaris showed
just as great a difference in variability between workers
and drones in favor of the former. Apparently, the sta-
tistics in these two nearly related groups lead to opposite
conclusions ; in reality probably neither statistical differ-
ence is significant as far as the question we are discussing
is concerned. The only conclusion justified by such data
would seem to be that the coefficient of heredity is as high
in the production of asexual as it is in the production of
sexual forms.

Moreover, one can not expect anything more definite
from this method of attack. Biologists may differ as to

- It should be noted here that all parthenogenetic eggs are not mere
spores. Some preparation often occurs through the emission of one polar
bodv. This may be merely a kind of recapitulation, a vestigial process no
longer having any significance whatever, but since we are not certain it
seems to the writer that the evidence from plants at present must be re-
garded as stronger.

Nos. 618-619] THE ROLE OF REPRODUCTION

281

the definition of fluctuation, mutation, etc., but they are
generally agreed that germinal variations, be they great
or small, are inmost species so rare they can not be gauged
by the use of ordinary statistical methods. For this rea-
son, a comparison between the variability of the drones
and of the workers of a pure race of bees is not likely to
show any difference between these two modes of repro-
duction in the matter of the frequency or the type of the
germinal variation produced, and can not answer the ques-
tion as to whether sexual reproduction contributes more
material for the use of natural selection than asexual re-
production. A study of variability in crossed races,
where the effect of Mendelian recombination can be con-
sidered, would be a more logical attack upon the second
problem, but is hardly necessary in view of the other
evidence available.

One is then justified in claiming there is no experimental
evidence to show that sexual reproduction in itself is not
an exact equivalent of asexual reproduction in the matter
of a heredity coefficient, but is this also true for germinal
variation! We believe it is. Variations there are in
both asexual and sexual reproduction, but it can not be
maintained that they occur more frequently in the latter.
There are insects in Oligocene amber apparently identical
with those of to-day, proving that constancy of type is
possible through long periods of time under sexual repro-
duction; yet germinal variations occur to-day in some-
what noteworthy numbers, as Morgan's work on Dro-
sophila shows, although the proportion of these varia-
tions which show possibilities of having an evolutionary
value, as evidenced by persistence in natural types, is
probably small. On the other hand, the number of varia-
tions produced under the dominance of asexual repro-
duction can not be said to be less numerous, even among
organisms of a relatively high specialization. If there
are those who doubt this statement, let them refer to the
immense list of bud-variations in the higher plants com-
piled by Cramer (1907).

There would be little reason in pushing the claims

282

THE AMEBIC AN NATURALIST

[Vol. LI1

further, since even though there does not seem to be a
sufficient difference between sexual and asexual reproduc-
tion in the matter of variation frequency to make it a
subject of experimental proof, certain theoretical points
raise the suspicion that there is such a difference. All we
would maintain is that to account for the general persist-
ence of sexual reproduction by such a cause, the differ-
ence in its favor should be so great that it could easily be
determined experimentally. Since this is not true, we
believe the hypothesis should be discarded.

The points of theory referred to are these. It will be
allowed by all that there is some considerable evidence of
the chromosomes being the most important conservators
of hereditary factors— the physical bases of heredity in
whatever form they may be. If it is assumed then that
changes in constitution in these cell organoids are fol-
lowed by changes in type, and that such changes in con-
stitution are equally probable in all chromosomes, it
follows that parthenogenetic individuals having the hap-
loid number of chromosomes should show a larger propor-
tion of germinal variations than members of the same
species having the diploid number of chromosomes, be-
cause variations of all kinds should be recognizable in the
former case, while in the latter, recessive variations could
not be detected until the first or second filial generation,
and then only when the proper mating was made. There
is some evidence that this reasoning is not wholly improb-
able. But variations occur much more frequently in
heterozygotes than in homozygotes. To me this simply
means that bud-variations are detected more frequently
in heterozygotes than in homozygotes : and an interpreta-
tion is not hard to find. Eetrogressive variations are
much more frequent than progressive variations, and a
retrogressive variation in a particular character shows
only when the organism is heterozygous for that character.
If a retrogressive bud-variation arises in a homozygote
and gametes are afterwards developed from the sporting
branch it is not at all unlikely that the variation may show
in the next generation, but it will be attributed then to

NOS.61S-619] THE ROLE OF REPRODUCTION

283

gametic mutation. If one compares asexual and sexual
reproduction from the standpoint of frequency of varia-
tion only, then sexual reproduction may seem to hold the
advantage over asexual reproduction in the usual sense;
but parthenogenesis, which is certainly a form of asexual
reproduction, is in theory better adapted than sexual re-
production for giving large numbers of variations.

If, therefore, one is constrained to agree that the bulk
of the evidence points to a practically identical coefficient
of heredity for both forms of reproduction, and that varia-
tion in the sense of actual changes in germinal constitu-
tion may occur with greater frequency in asexual repro-
duction, if there is any difference at all between the two
forms, he is driven either to the conclusion of Maupas
that continued asexual reproduction is impossible through
some protoplasmic limitation or to the conclusion of Weis-
mann that a mixture of germplasms offers sufficient ad-
vantages to account for everything. This is the dilemma3
unless one wishes to maintain that efficient mechanisms
for nutrition, adaptation, protection and distribution
could not be evolved or maintained under asexual re-
production.

The contention of Maupas can not be dealt with experi-
mentally any more successfully than the question as to
the inheritance of acquired characters since experimental
time and evolutionary time are not of the same order of
magnitude. The long-continued experiments of Wood-
ruff in which vigorous strains of Paramecium have been
kept dividing asexually for several thousand generations,
however, as well as the botanical evidence that numerous
species having no sexual means of multiplication have
continued to exist during long periods of time, weight the
balance against him. One need not hesitate to concede
that all of these organisms are rather low unspecialized
types; the modern development of genetics has built up
such a solid structure in favor of Weismann's view that
there is little need of argument along the older line.

3 Naturally another hypothesis wholly new to biology may be submitted
at any time.

284 THE AMEBIC AX XA TURALIS T [Vol. LII

The main argument in favor of Weisinann's viewpoint
does not take long to state. It is this : Mendelian heredity
is a manifestation of sexual reproduction. Wherever
sexual reproduction occurs, there Mendelian heredity will
be found. The very fact that it describes the sexual
heredity of both animals and plants is sufficient proof of
its generality in this regard. Xow if A" variations occur
in the germplasm of an asexually reproducing organism,
only N types can be formed to offer raw material to selec-
tive agencies. But if N variations occur in the germ-
plasm of a sexually reproducing organism 2n types can be
formed. The advantage is almost incalculable. Ten
variations in an asexual species mean simply 10 types, 10
variations in a sexual species mean the possibility of 1.024
types. Twenty variations in the one case is again only 20
types to survive or perish in the struggle for existence ; 20
variations, in the other case, may present 1,032,576 types
to compete in the struggle. It is necessary to hedge the
argument by pointing out that these figures are the maxi-
mum possibilities in favor of sexual reproduction. It is
improbable that they ever actually occur in nature, for 220
types really to be found in the wild competing for place
after only 20 germinal variations would mean an enor-
mous number of individuals even if the 20 changes had
taken place in different chromosomes, and if the varia-
tions were linked at all closely in inheritance the number
required would be staggering. But there are breaks in
linked inheritance, and the possibility is as stated.

These advantages remain even though it should be shown
later that the more fundamental and generalized char-
acters of an organism are not distributed by Mendelian
heredity. Loeb (1916) believes that the cytoplasm of the
egg is roughly the potential embryo and that the chromo-
somes, distributed as required by the breeding facts of
Mendelian heredity, are the machinery for impressing the
finer details. There is something to be said for this point
of view, though at present it is but a working hypothesis.
But granting its truth it does not detract from the ad-
vantages gained by sexual reproduction. Even the most

Nos. 618-619] THE ROLE OF REPRODUCTION

285

strict mutationist would hardly maintain that evolution in
general has come about through tremendous changes in-
volving sterility between the mutant and the parent types.
It seems unnecessary to deny such possibilities; but the
weight of evidence is in favor of the majority of varia-
tions being comparatively small, changes in detail, the
very kind which are known to be Mendelian in their in-
heritance.

Yet sexual reproduction in itself does not assure these
advantages, though they are based upon it. There must
be means for the mixture of germplasms. This oppor-
tunity was furnished originally by bisexuality. Then
came hermaphroditism, manifestly an economic gain, yet
on the whole unsuccessful except as functional bisexuality
was restored by self-sterility, protandry, protogyny or
mechanical devices which promoted cross-fertilization.

The prime reason for the success of sexual reproduc-
tion then, as Weismann maintained, is the opportunity it
gives for mingling germplasms of different constitution
and thereby furnishing many times the raw material to
selective agencies that could possibly be produced through
asexual reproduction. Further, there are three minor ad
vantages which rest upon the same mechanism. They are
minor advantages only when compared to the major, and
should not be passed by.

Let us first consider heterosis, the vigor which accom-
panies hybridization. This phenomenon has long been
known. It is characteristic of first generation hybrids
both in the animal and vegetable kingdoms. It affects the
characters of organisms in much the same manner as do
the best environmental conditions. In other words, the
majority of characters seem to reach the highest de-
velopment in the first hybrid generation. The hybrid in-
dividual therefore holds some considerable superiority
over the individuals of the pure races which entered into
it, and is thereby the better enabled to survive and to
produce the multiplicity of forms which its heterozygous
factors make possible. The frequence of this phe-
nomenon, for it is almost universal, together with the fact

2S6

THE AMERICAS XATURALIST

"Vol. LII

that it seems impossible to fix the condition, led Shull and
the writer independently to the conclusion that certain
factors in addition to their functions as transmitters of
hereditary characters also had the faculty of carrying
some sort of a developmental stimulus when in the hetero-
zygous condition. The recent work of Morgan on linked
characters, however, makes it possible to give another
interpretation, as Jones (1917) has demonstrated. If it
be assumed that several variations have occurred in each
of one or more chromosomes, then it can be shown that the
first-generation hybrid between such a variant and the
race from which it arose will bring together all dominant
or partially dominant characters. In the second hybrid
generation, on the other hand. Mendelian recombination
steps in and makes it improbable that many individuals
shall have such a zygotic composition. And only in the
rare cases where the proper breaks in linkage have oc-
curred can a homozygous individual of this type be
produced.

The latter hypothesis holds the advantage that it
furnishes hope for a homozygous combination as valuable
as that of the first hybrid generation no matter how
rarely it may be assumed to occur, but whether it holds
for the majority of organisms or not may depend on a
future decision as to the frequency of side-by- side
synapsis as compared to end-to-end synapsis. Our knowl-
edge of linkage rests almost entirely on Morgan \s work
on Drosophila where side-by-side synapsis occurs at the
maturation of the germ cells. If the break in linkage be-
tween groups of characters apparently carried by a single
chromosome, which Morgan finds to be so exact in Dro-
sophUa, should actually depend on Jannsen's theory of
chromosome, twisting at synapsis, then some other type
of inheritance may be found in species having end-to-end
synapsis. Perhaps this is the reason why the CEnotheras
have such a peculiar heredity, for in them Davis (1909)
thinks end-to-end synapsis prevails. But. be this as it
may. the vigor of first generation hybrids is a fact and
not a theory, and the advantage it brings to the hetero-

Nos. 618-619] THE ROLE OF REPRODUCTION

2*7

zygotic individual in competition with its fellows can not
be gainsaid.

The investigations of Shull and of the writer on the
effects of cross- and self-fertilization have brought to
light another series of facts with a bearing on the problem
under discussion. It has been shown that the apparent
deterioration of cross-bred species when self-fertilized is
in large measure and perhaps wholly due to the loss of
hybrid vigor4 through the formation of homozygotic
Mendelian recombinations and not an effect of inbreeding
per se because of the union of like germplasms. This is
a plausible argument against Darwin's idea that con-
tinued inbreeding is abhorrent to Nature. It may even
be said to be a valid reason for declining to accept
Maupas's belief in the impossibility of continued asexual
reproduction, for there is no very good reason for dis-
tinguishing between continued asexual propagation and
continued self-fertilization. Inbreeding simply brings
about the opposite effect from crossing, and we can see
no reason for the comparative failure of naturally inbred
types in the wild other than the lack of chances for
progress. The one is the conservative manufacturer who
continues the original type of his article, the other is the
progressive who makes changes here and there without
discouragement until the acceptable improvement is
found. In fact, if this argument be overlooked, the in-
bred types which have persisted hold some advantages
over the cross-bred types. The self -fertilized species are
inherently strong and vigorous, witness tobacco and
wheat. They stand or fall on their own merits. They
are unable, as are cross-bred species, to cover up in-
herent weakness by the vigor of heterozygosis. Cross-
fertilized maize has become the king of cultivated plants
because of its variability, but many of our best varieties
carry recessive characters very disadvantageous to the
species.

The next secondary advantage of sexual reproduction is

4 Accepting the view that the vigor of the first hybrid generation is due
to dominant characters meeting makes this argument even more forcible.

288

THE AMERICAN NATURALIST

[Vol. LII

the division of labor made possible by secondary sexual
characters, using the term very generally and including
even such differences as those which separate the egg
and the sperm. It is not known just how these differences
arose or by what mechanism they are transmitted. The
greatest hope of reading the riddle lies in an investiga-
tion of hermaphroditic plants, for there are technical
difficulties which seem to preclude their solution in ani-
mals. For example, breaks in the linkage between sex-
linked characters occur only in the female in Drosophila,
and as the sex chromosome is double in the female, it
can not be determined whether the differentiation be-
tween male and female is due to the whole chromosome or
not. But this ignorance does not give reason for a denial
of the great advantage which sexes bearing different
characters hold over sexes alike in all characters except
the primary sex organs.

The only glimpse of the truth we have on these matters
comes from recent work on the effect of secretions of the
sex organs on secondary sexual characters. The effect of
removing the sex organs and the result of transplanting
them to abnormal positions in the body have shown that
in vertebrates the secretions of these organs themselves
activate the production of the secondary sexual char-
acters. This does not seem to be the case in arthropods,
however, so one can not say that primary sexual differ-
entiation and secondary sexual differentiation is one and
the same thing.

Finally there is a presumable advantage in gonocho-
ristic reproduction in having sex-linked characters. We
say presumable advantages, for all of the relationships
between sex and sex-linked characters are not clear. The
facts are these: One sex is always heterozygous for the
sex determiner and the factors linked with it. Now it
may very well be that there is an actual advantage in the
heterozygous condition, as we have seen above. But
should the so-called vigor of heterozygosis prove to be
only an expression of the meeting of dominant characters,
still a possible advantage accrues to this phenomenon be-

Nos. 618-619] THE ROLE OF REPRODUCTION

28(J

cause the mechanism contributes toward mixing of germ-
plasms. As an example, let us take the Drosophila type
of sex determination. There the sperm is of two kinds:
the one containing the sex chromosome and its sex-linked
factors, the other lacking it. The eggs are all alike, each
bearing the sex chromosome. It follows then that the
male always receives this chromosome from his mother,
who may have received it from either her father or
mother. Moreover, further variability may be derived
from the linkage breaks which occur always in the female.
This last phenomenon is hardly worthy of special men-
tion, however, until it is shown to be typical of such
reproduction.

This short reconnaissance presents the pertinent facts
in the situation as they appear to the writer. A very
great number of interesting things connected with repro-
duction during the course of evolution have not been men-
tioned. This is because it is felt that the essential feature
of the role of reproduction in evolution is the persistence
of mechanisms in both the animal and plant kingdoms
which offer selective agencies the greatest amount of raw
material. Other phenomena are wholly secondary.

LITERATURE CITED

Casteel, D. B., and Phillips, E. F. 1903. Comparative Variability of
Drones and Workers of the Honey Bee. Biol. Bull., 6: 18-37.

Coulter, J. M. 1914. The Evolution of Sex in Plants. Chicago, Uni-
versity of Chicago Press. Pp. 1-140.

Cramer, P. J. S. 1907. Kritische Ubersicht der bekannten Falle von
Knospenvariation. Natuurkundige Verhandelingen van de Hollandsche
Maatschappij der Wetenschappen. Derde Verzameling, Deel VI, Derde
Stuk. Haarlem, De Erven Loosjes, pp. iii-xviii + 474.

Davis, B. M. 1909. Cytological Studies on (Enothera. I. Ann. Bot. 23:
551-571.

Jones, D. F. 1917. Dominance of Linked Factors as a Means of Account-
ing for Heterosis. Genetics, 2: 466-479.

Kellogg, V. L. 1906. Variation in Parthenogenetic Insects. Science,
N. S., 24: 695-699.

Loeb, J. 1916. The Organism as a Whole. N. Y., Putnam. Pp. v-x +
379.

Wright, A., Lee, A., and Pearson, K. 1907. A Cooperative Study of
Queens, Drones and Workers in Vespa vulgaris. Biometrila, 5: 407-
422.

[Reprinted from Brooklyn Botanic Garden Memoirs, 1: 141-153. 1918.]

INTERCROSSES BETWEEN SELF-STERILE PLANTS

E. M. EAST
Bussey Institution, Harvard University

The fact that self-fertilization is practically impossible in certain
hermaphroditic plants, although both the pollen and the ovules are
functional in crosses, has been known since the time of Kolreuter
( 1 760-1 765). The oddity of the phenomenon has been a lure for al-
most every hybridist from that time forward. As in the case of most
other genetical problems, however, our knowledge of its cause and
meaning remained in status quo from the time of Darwin until Men-
delian days. Indeed when the writer began his investigations on the
subject in 1910, the only considerable post-Darwinian work had been
done by a zoologist (Morgan, 1904) on the self-sterile ascidian, Ciona
intestinalis . Since 19 10 botanical papers have appeared by Correns
(1912), Compton (1913) and Stout (1916), but these investigations
will not be discussed here, as it is proposed to treat in this paper only
certain phases of the work carried on by the author and his associates1
during the past seven years, leaving critical review for another place.
For our purpose it seems essential only to present a hasty sketch of
the subject as left by Darwin.

In addition to the utilization of most of the previous and the con-
temporaneous work, Darwin (1876) carried out several investigations
of his own on the five self-sterile species, Eschscholtzia calif ornica t
Abatilon darwinii, Senecio cruentus, Reseda odorata and Reseda lutea.

Darwin's first important result was that the expression of self-
sterility in Eschscholtzia calif ornica and Abutilon darwinii was influ-
enced by changes in external conditions. Six generations of Esch-
scholtzia calif ornica had been found to be completely sterile in southern
Brazil by Fritz Miiller (1868, 1873). As English plants were self-
fertile, Darwin obtained from Miiller seed of Brazilian plants of known
self-sterility. The plants which they produced in England, while
not wholly self-fertile, tended toward self-fertility, which fact Darwin
attributed to the lower English temperature. A second generation of
seedlings proved to be still more self-fertile. Conversely, seed of
English stock was somewhat self-sterile the first season and one plant

1 The author desires to make grateful acknowledgment to Dr. O. E. White and
Dr. J. B. Park for their painstaking aid in this work. Without it, the numerous
experiments undertaken could not have been completed.

141

142 BROOKLYN BOTANIC GARDEN MEMOIRS

wholly self-sterile the second season, when grown in Brazil. One
may assume, I think, arguing from data of similar character, that this
progressive result was not due to actual inheritance of an acquired
character but rather to the fact that the first generation in each case
passed a portion of its life cycle in the original environment.

Similar results were obtained in the case of Abutilon darwinii,
which though self -sterile in its native Brazil, became moderately self-
fertile late in the first flowering season in Darwin's greenhouse.

Darwin made more detailed experiments on Senecio cruentus,
Reseda odorata and Reseda lutea and found, as he believed, that each
plant though self-sterile was cross-fertile with every other plant.
His pollination experiments with Senecio cruentus and Reseda lutea
were so inadequate that they may be omitted from consideration; it
was really his experiments on Reseda odorata that were thought to
establish the fact of complete cross-fertility.

Darwin's Experiments on Reseda odorata in 1868

Male Parents

Only sixteen cross matings were made, however, and this is not
sufficient to prove the point, as is shown by one of our own experi-
ments, where 131 cross-ma tings were made with only 4 cases of cross-
sterility. From the fertile cross-pollinations Darwin raised four
plants in 1869. Three of these proved to be self- fertile and one self-
sterile. Six more plants were grown in 1870. Of these, two were
almost self-sterile and four were almost completely self-fertile. The
former produced altogether five seeds from self-pollinations, and the
resulting plants proved to be self-sterile like their parents. These
varied results Darwin attributed to a difference in inherited sexual
constitution, but it seems to me that this conclusion should be ques-
tioned. Our own results have proved conclusively that toward the

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 143

very last of the flowering season2 self-sterile plants may sometimes
become somewhat self-fertile.

Darwin's (1876, p. 346) general conclusions are as follows:

'* Finally, the most interesting point in regard to self-sterile plants
is the evidence which they afford of the advantage, or rather the
necessity, of some degree or kind of differentiation in the sexual
elements, in order that they should unite and give birth to a new being.
It was ascertained that the five plants of Reseda odorata which were
selected by chance could be perfectly fertilised by pollen taken from
any one of them, but not by their own pollen; and a few additional
trials were made with some other individuals, which I have not
thought worth recording. So again, Hildebrand and Fritz M tiller
frequently speak of self-sterile plants being fertile with the pollen of
any other individual; and if there had been any exception to the
rule, these could hardly have escaped their observation and my own.
We may therefore confidently assert that a self-sterile plant can be
fertilised by the pollen of any one out of a thousand or ten thousand
individuals of the same species, but not by its own. Now it is obvi-
ously impossible that the sexual organs and elements of every indi-
vidual can have been specialised with respect to every other indi-
vidual. But there is no difficulty in believing that the sexual elements
of each differ slightly in the same diversified manner as do their
external characters; and it has often been remarked that no two
individuals are absolutely alike. Therefore we can hardly avoid the
conclusion that differences of an analogous and indefinite nature in
the reproductive system are sufficient to excite the mutual action of
the sexual elements, and that unless there be such differentiation
fertility fails."

One cannot but admire these inductions Darwin has so cleverly
drawn from such meager data, nevertheless one cannot accept them
today just as they stand. The reasons for this statement will be
seen more clearly when our own data have been presented, but a brief
can be submitted with only the support of the work known to Darwin.

In the first place, the seemingly contradictory results that were
obtained in the experiments on Reseda odorata are not necessarily con-
fusing. As reported, self-sterile plants produced varying ratios of
self-sterile and self-fertile plants. Unfortunately, the progeny of the
self-fertile plants was not followed. If it has been, the problem might
have been more easily solved, for, in all probability, the daughter
plants would have been self-sterile. It is my own belief, however,
that the answer can be read in the casual remarks dropped by Darwin
in the midst of his careful descriptions, remarks to which he paid little
attention. Darwin found that both Eschscholtzia California and
Abutilon darwinii, though self-sterile in Brazil tended to become self-
fertile in England, — especially late in the flowering season. Now

2 Cf. Darwin's observation on Abutilon darwinu.
11

144

BROOKLYN BOTANIC GARDEN MEMOIRS

these facts together with that mentioned above regarding the in-
constancy of the results obtained from planting the seed of self-sterile
plants, may be interpreted by the assumption that he was dealing
entirely with fluctuations in all of the five species investigated. These
species genetically were wholly self-sterile. The tendency toward
self-fertility was due to conditions. In other words, these plants
genetically self-sterile needed conditions conducive to a fine healthy
growth to bring out their self-sterility. In the lower temperature of
England, at a time of decline (the last of the flowering season), they
became phenotypically somewhat self-fertile. In the light of my own
experiences, I believe we can reconstruct a picture of Darwin's experi-
ments on Reseda odorata with considerable confidence. He isolated
the plants that he desired to test under nets; then came pressure of
other work, and the data were not collected until the plants had ceased
flowering. At that time capsules were found beneath the nets, and
this seemed to prove at least a partial self-fertility. But instead of
this procedure, suppose that successive self-pollinations had been
made throughout the season. The presumption is that the plants
would have been declared to be self-sterile with the same remark
added which he jotted down in the case of Abutilon darunnii, viz.,
they "became moderately self-fertile late in their flowering season."

Again, Darwin found no cross-sterility in the plants tested, and
concluded that a self-sterile plant can be fertilized with the pollen
of any one of a thousand or ten thousand individuals of the same spe-
cies. Such a conclusion was less cautious than was Darwin's wont for
it was made from a total personal experience of some twenty-odd cross-
matings only, unless his records are extremely incomplete. Indeed
this conclusion must have been somewhat of a surprise to himself
since he states that "it is obvious impossible that the sexual organs
and elements of every individual can have been specialized with respect
to every other individual." He surmounted this difficulty by assum-
ing that the sexual elements of each plant differ slightly in the same
manner as their external characteristics, and that this slight difference
is sufficient to excite the mutual action of the sex elements necessary
in order to have fertilization ensue. The kernel in this conclusion,
that differences in the reproductive systems of two self-sterile plants
are necessary in order to promote cross-fertilization, is so similar to
that to which the writer has been forced after seven years of rather
intensive work as to be uncanny, for it seems to have been reached
in spite of rather than because of the data at hand. This feeling of
surprise at Darwin's clairvoyancy may seem affected, since he was
usually in advance of his time, but it is a fact perhaps worth men-
tioning as a confession of omission that the writer reached his con-

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 145

elusions as the outgrowth of work on heterozygosis and did not refer
to Darwin's view until recently. Be this as it may, a short com-
parison of Darwin's main induction with the facts from which it came
will, I think, show a real reason for wonderment. He believed in
universal cross-fertility of self-sterile plants, his basis being the small
number of cross-fertilizations made by Hildebrand, M tiller and him-
self; although Robertson Munro (1868), with whose work he was
familiar, had found cross-sterility in Passifiora alata, and even the
works of Hildebrand and M tiller as published leave the matter in
doubt. Now how much more reasonable the general induction
mentioned above seems if one assumes (1) that self-sterile plants breed
true for self-sterility but may show a slight degree of self-fertility as a
fluctuation under certain conditions, (2) that a variable but limited
number of germinal "factors" influence the success of matings, cross-
fertilization being possible only when two plants differ in these effective
factors, and (3) that when two plants have the same effective factorial
composition, cross-sterility of the same type as self-sterility exists.
This is what we believe our own work has shown, as we shall try to
demonstrate.

Emphasis must first be laid upon the fact that the behavior of
self-sterile plants among themselves and the relation between self-
fertile and self-sterile plants are distinct problems. Compton (191 3)
found the relation between self-fertile and self-sterile plants of Reseda
odorata to be that of a simple Mendelian monohybrid with self-fertility
dominant. The same relation appears to hold in crosses between the
self-fertile species Nicotiana langsdorffii and the two self-sterile species
with which our work has been done, Nicotiana forgetiana and Nicotiana
alata. There is some single differential between self-fertility and self-
sterility. Given the proper composition a plant breeds true for self-
sterility. The behavior of self-sterile plants among themselves
therefore must be considered separately.

Our work, as stated before, has been done with the two self-sterile
species, Nicotiana forgetiana and Nicotiana alata, and largely with
crosses between these species. Both of these species are affected in
their manifestation of self-sterility by certain environmental changes,
Nicotiana alata much more than Nicotiana forgetiana. Self-sterility
is determined by the inheritance received, but it can develop fully
only under environmental conditions which promote a normal healthy
growth, and during the period of intense flowering. Toward the
end of the flowering period, especially under conditions adverse to
vegetative growth, self -sterility sometimes shows a marked and rather
sudden decline. A few seeds, or even a well-developed seed capsule
may then be obtained. This is not a common occurrence; indeed, it

146

BROOKLYN BOTANIC GARDEN MEMOIRS

is rare, but it is a possibility. Three cases of seed production out of
over three hundred plants tested have been observed in Nicotiana
forgetiana. A considerably higher percentage of fertility has been
marked in Nicotiana alata. Self-sterility can be restored in such
plants, however, if they are allowed to go through a period of rest and
are then, by proper treatment, brought into vigorous flower again.

This is not the whole evidence that this occasional end-season
fertility is a pseudo-fertility brought about by external conditions —
a fluctuation. Three generations of Nicotiana alata plants have been
grown from selfed seed produced by end-season fertility without the
occurrence of a single plant which behaved in every way like a truly
self-fertile individual. This phenomenon, therefore, while teaching
us to test self-sterility only during the main part of the flowering
season, has shown that there is no reason why fusion between gametes
produced by a self-sterile plant may not occur provided the male
generative nucleus enters the embryo sac. Such unions may take
place without affecting the self-sterility of the progeny.

What is then the difference in behavior that makes a cross-pollina-
tion effect fertilization while a self-pollination produces nothing?
What occurs is this: After a self-pollination the pollen grains germinate
and the tubes pass down the style at such a slow even rate that they
reach only about half way to the ovary before the flower wilts and
falls off; while the pollen tubes after a cross-pollination, though
starting at the same rate as the others, grow faster and faster until
fertilization is effected in four days or less. The curve of distance
traversed plotted against time is in the case of the self-pollination
nearly a straight line, while in the case of the cross-pollination it
simulates that of an autocatalytic reaction.

From these facts it seems reasonable to suppose that the secre-
tions in the style offer a stimulus to pollen tubes from other plants
rather than an impediment to the development of tubes from pollen
of the same plant. And we believe that this stimulus is in some way
caused by certain effective differences in the factorial composition
characterizing two compatible plants and that if two plants do not
have these effective differences in factorial composition they are by
the same token cross-sterile with each other. It is clear that this
assumption presumes that the pollen grains matured by a given plant
behave as if they are sporophytic as regards that part of their con-
stitution that affects self-sterility and cross-sterility. The pollen
grains of any plant may carry many different hereditary factors, they
may even carry several different factors which function in controlling
the success or failure of particular cross-matings in the next generation,
but in their own action on the stigmas of other plants they behave

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 147

as if each carried the composition of the mother plant from which it
came. In other words, as far as its action in fertilization is concerned,
a pollen grain partakes of the character of its mother plant and is
like its sisters; as far as the hereditary characters carried on to the
next generation are concerned, sister pollen grains may differ both
from their mother and from each other.

A part of our evidence on these points we shall present. For
further details the reader is referred to a forthcoming paper in Genetics.3

The first experiment to which attention is called is an inbreeding
experiment performed on a cross between Nicotiana forgetiana and
Nicotiana alata. If sister plants are mated in successive generations
after an original mating Aa X Aa, by Mendelian recombination there
results a gradual approach to 1/2 AA, 1/2 aa and o Aa. Expectation
of homozygosis in successive matings is 1/2, 5/8, 11/16, 24/32 ••• 1
(Jennings, 1916). If, therefore, plants of like constitution as far as
effective factors are concerned are cross-sterile with each other,
cross-sterility should become more and more apparent in generations
succeeding F2. To test this possibility, a comparatively small number
of cross-matings was made on the F2, F3, F4 and F5 generations. In
the F2 generation, out of 131 intercrosses on 20 plants only 4 were
unsuccessful. The percentage of unsuccessful matings increased from
this time on, until in the F5 generation about 21 percent of the cross-
matings tried on 20 plants were impossible to make.

In this experiment as well as in all others, results showed that
reciprocal crosses were alike in their compatibility. If two plants
were fertile together, they were fertile reciprocally; if two plants
were incompatible, they were incompatible reciprocally. This is
proof of the sporophytic behavior of the factors affecting the behavior
of self-sterile plants.

The two crosses to be described next are reciprocals made with the
same two individuals. Made with Nicotiana alata and Nicotiana
forgetiana as parents, they are in a sense repetitions of the cross just
described, but it is hardly probable that they duplicate it. Both of
these species must consist of plants which differ among themselves in
the factors which affect self-sterility, hence any crosses in which
different individuals are used may show different results.

All of the individuals resulting from this cross were grown in a
greenhouse as potted plants. The Fi generation came into blossom
during the latter part of the winter. Conditions were extraordinarily
favorable for growth and the pollinations were all made while the
plants were vigorous, hence scarcely any trouble arose over classi-
fication of the results through end-season pseudo-fertility.

3 This paper has since appeared. See "Studies on Self-sterility I. The
Behavior of Self-sterile Plants." Genetics 2: 505-609. 1917.

148

BROOKLYN BOTANIC GARDEN MEMOIRS

Our study was made on a population of 53 plants. Pedigree
numbers from o to 39 inclusive represent the cross N. alata X N. for-
getiana; pedigree numbers 40 to 52 inclusive represent cross A7, for-
ge tiana X N. alata.

Each plant was selfed one or more times, and all proved abso-
lutely self-sterile. Further each plant was back-crossed with pollen
from a single plant of each of the parent species with complete success
in every case. The plants used in this case were not the individuals
that entered into the cross, however, for unfortunately these were
not available.

TABLE I

Result of Matings on Fi Plants o to 39

N. alata X A7, forgetiana and on Plants 41 to 52 N. forgehana X N. alata
Ped. Xo. Fertile with Ped. No. Sterile with Ped. No.

0 44, 46 22, 34, 38, 49

1 2, 3, 4, 6, 41 . . . 8

2 4, 18, 41, 44, 52 9, 22, 23

3 2, 9, 14, 23, 29 4, 6, 18, 41, 46

4 2, 9, 10, 44 18

5 2, 3, 6, 9, 10, 18, 46 8, 44

6 5, 10, 43, 44 3, 4, 18, 40

7 2, 13, 22, 44 1 8, 46

8 6, 9, 10, 39, 40, 46 5, 44

9 3, 18, 44, 52 2, 10, 23, 37, 48

10 4, 6, 18, 40, 44 2, 23, 24, 27, 34, 48

11 2, 8, 12, 15, 34, 44, 46

12 9, 16, 22, 43 6, 18, 46, 52

13 3, 8, 18, 44, 46 2, 9, 15, 21, 34

14 18, 20, 43 10, 34

15 1, 3, 16, 17, 18, 20 9, 13, 14, 23, 44

16 13, 14, 18, 25, 43, 46 17, 29

17 14, 18, 19, 20, 22, 30 16, 26, 44

18 2, 9, 21, 23, 28, 34, 36, 44 3, 46

19 17, 22, 28, 34, 44 18

20 2, 8, 9, 16, 18, 21, 22, 26, 36, 40, 44 43

21 4, 12, 16, 18, 46 2, 9, 22, 25, 27, 37

22 12, 42, 44 14, 23, 24, 36, 48

23 41 9, i°. 37, 48

24 3, 6, 20, 26, 28, 44 10, 22, 23, 30, 37

25 8, 33, 44, 46 2,9,23,27

26 9, 18, 22, 23, 25, 40, 48 28, 29, 44

27 3, 18, 32, 44, 46 2, 9, 30, 34, 48

28 2, 3, 23, 27, 39, 46 8, 26, 29, 44

29 2, 14, 18, 22, 23, 24, 25, 30, 34, 37, 41, 46 5, 26, 28, 31, 44

30 8, 29, 33, 44, 45, 46 9, 21, 22, 27

3i 22, 32, 52 8, 29, 36, 44

32 9, 21, 23, 29, 30, 34, 43, 44 18, 33, 46

33 8, 16, 23, 31, 46 18, 32

34 28, 41, 44, 46 10, 23, 24, 37

35 3, 9, 18, 2i, 27, 30, 34, 37, 42 8

36 8, 33, 44, 46 10, 23

37 39, 42, 43, 44, 46 9, 10, 22, 23, 34, 38

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 149

38 28, 35, 39, 42, 43, 46 34, 37, 47

39 9, 44 18, 40, 42

40 22,43,44,47,49 6, 33, 46

41 10, 37, 44, 48 33, 40, 46

42 20, 44 39, 4i, 45

43 5, 27, 33, 38, 39, 40, 42, 44, 46, 51

44 10, H, 23, 34, 45

45 18, 44, 48 46, 52

46 10,22,37,44,51 52

47 20, 42, 44, 45, 46, 51, 52 38

48 40, 4i, 43, 46 10, 23, 24, 27, 34

49 42, 44, 45 o, 9, 27, 34, 47

50 18, 39, 51, 52 9, 27, 37

51 9, 18, 23, 39, 45, 46, 50 8, 29

52 10, 23, 29, 37, 51 3, 4, 6, 18, 41, 45, 46

The numerous cross-matings made are shown in Table I. There
were 103 reciprocal matings. Of these 100 gave duplicate results, 39
pairs being fertile and 61 sterile. The three which did not check are:

2 X 3, sterile, I pollination 1

3X2, iertile, 1 pollination J

6 X 52, fertile, 1 pollination 1 , ,

w , ., „. . r classed as sterile,

52 X 6, sterile, I pollination J

37 X 21, fertile, I pollination 1 ^ ^

21 X 37, sterile-, I pollination J

Since but one pollination was made in each of these cases we have
made our decision as to fertility or sterility by a consideration of the
circumstantial evidence. The behavior of these plants in other crosses
shows conclusively that 3 should be fertile with 2, 6 sterile with 52,
and 21 sterile with 37. They have been classed accordingly. That
this grouping is correct is further shown by the fact that the mating
3X2 (classed fertile) was made at the height of the flowering season,
while the matings 6 X 52 and 37 X 21 (classed sterile) were re-
spectively the last and next to the last matings made on those plants.

In spite of the fact that plants 0-39 are from cross N. alata X N.
forgetiana, and plants 40-52 are from cross AT. forgetiana X N. alata,
they behave as one family in intercrosses. The entire population
can be grouped into 6 classes in which there is interclass fertility and
intraclass sterility. The following explanation may be necessary to
make it clear just how Table II was obtained from Table I. Table I
shows all of the matings, but in the form given it is not easy to see at a
glance every combination in which a particular plant was used, both
as male and as female. It was necessary, therefore, to make a new
table, in which the pedigree numbers in the column at the left were
tabled as males, and the pedigree numbers in the columns headed
"Fertile matings" and "Sterile matings" were tabled as females.

150

BROOKLYN BOTANIC GARDEN MEMOIRS

Thus plant 2, used as a female, was fertile with pollen from plants 4,
18, 41, 44 and 52, and sterile with plants 9, 22 and 23; but pollen
from plant 2 was fertile on plants 1, 3, 4, 5, 7, n, 18, 20, 28 and 29,
and sterile on plants 9, 10, 13, 25 and 27. It is clear, therefore, that
instead of the 8 matings on "plant 2 that Table I appears to show,
there are really 21, the 3 reciprocals of course being counted but once.

These tables were combined for analysis. In the interest of
economy of space only one is shown, however, since the second can
easily be made from the first.

The four exceptions in this huge set of matings are in reality
negligible. Matings 15 X 44 and 31 X 36 were sterile, though they
do not belong to the same class. Plant 15 was sterile to 4 plants of
Class A and fertile to 2 plants of Class B, 3 plants of Class C, and to
the isolated individuals forming classes D and F. It is unquestionably
a member of Class A. Plant 44 was sterile to 7 individuals in Class C
and fertile to 17 plants of Class A, 12 plants of Class B and to the
singletons forming classes D, E and F. This evidence places it un-
mistakably as a member of Class C. Plant 31 is also a member of
Class C as evidenced by 3 sterile matings within that class and by
fertile matings with 1 plant of Class A and 3 plants of Class B. Plant
36 is like plant 15 thrown into Class A by its sterility with 3 others of
that class, and by its fertility with 3 individuals of Class B, with 2 of
Class C, and with the lone plant of Class D. In view of this evidence
and the fact that in these two matings but one pollination was made
in each case, they are much more likely to be errors of record or of
technique than true exceptions to our classification.

The other two exceptions, matings 45 X 18 and 33 X 46, were
fertile where from the evidence of numerous other matings they should
have been sterile. Here again but one pollination was made in each
case; and, coincidence though it may be, each pollination was the last
mating made on that particular plant. What is more probable than
that this is a pseudo-fertility appearing during the wane of the flower-
ing season of the two mother plants, No. 45 and No. 33?

Six groups appear in Table II, but there is proof of the existence
of only five. Groups A, B, C, D and E are definitely established.
Plant 11, on the other hand, is an isolated individual rather than a
class. It does not belong to groups A, B or C; but unfortunately it
was not crossed either with Class D (plant 20) or with Class E (plant
43), hence one cannot say that it does not fall into one or the other of
these two classes.

In the three large groups the distribution of individuals is 22, 16
and 12. About all that can be said about the type of this distribution
is that the classes are not of equal size. On the other hand, it is

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 151

interesting to note that the plants of both cross No. 2 and cross No. 3
fell into the three groups as if they were samples of the same popula-
tion. There were 40 plants of Cross No. 1, and 13 plants of the

TABLE II

Plants of F] Generation of Reciprocal Cross between N. forgetiana and N.
alata, Grouped in Accordance with their Behavior in Intercrosses

Plants 0-39 are products of the cross; plants 40-52 are products of its reciprocal

152 BROOKLYN BOTANIC GARDEN MEMOIRS

TABLE II— Continued

Group

Ped. No.

Cases Fertile in Group

Cases Sterile in Group

reciprocal, Cross No. 2. In the classes A, B and C the proportions
were 18, 10, 10 and 4, 6, 2 respectively. This similar behavior of the
progeny of reciprocals seems to us strong corroboratory evidence in
favor of the conclusion that reciprocal crosses always behave in like
manner as regards self-sterility.

The study on this family is but one of several that have been
made but we believe that the data on it alone show unmistakably that
the behavior of self-sterile plants in intercrosses is governed by a
relatively small number of factors which act through pollen as if the
pollen grain possessed the characters of the sporophyte from which it
came, and that the gametes of plants having like constitutions as re-
gards effective factors are incompatible in the sense that they do not
make a normal pollen-tube growth and hence do not reach the ovary
in time for fusion to occur. This interpretation shows both why
plants are self-sterile and cross-sterile. It accords completely with
the fact that a population of plants may be divided into groups on the
basis of their mating proclivities and that each member of any group is
cross-sterile with every other individual of that group although it is
fertile with every individual of every other group.

These assumptions being true, it ought to be possible by con-
tinuous self-fertilization, utilizing end-season pseudo-fertility, to
obtain ultimately a population in which every individual possesses
the same effective self-sterility factors. In such a population all of
the plants will not only be self-sterile, but will be cross-sterile. Such a
population has been obtained.

EAST: INTERCROSSES BETWEEN SELF-STERILE PLANTS 153

REFERENCES CITED

Compton, R. H. Phenomena and Problems of Self-sterility. New Phytologist 12:
197-206. 1913.

Correns, C. Selbststerilitat und Individualstoffe. Festschr. d. mat.-nat. Gesell.
zur 84. Versamml. deutsch. Naturforscher u. Arzte Munster i.W. pp. 1-32.
1912.

Darwin, Chas. Effects of Cross- and Self-fertilisation in the Vegetable Kingdom.

Ed. 2, 1878. N. Y. D. Appleton. 1876.
Hildebrand, F. Bastardierungs Versuche an Orchideen. Bot. Ztg. 23: 245-249.

1865.

Ueber die Nothwendigkeit der Insektenhilfe bei der Befruchtung von Cory-

dalis cava. Jahrb. wiss. Bot. 5: 359-363. 1866.
Ueber die Bestaubungsvorrichtungen bei den Fumariaceen. Jahrb. wiss.

Bot. 7: 423. 1869.

Jennings, H. S. The Numerical Results of Diverse Systems of Breeding. Genetics
1: 53-89. 1916.

Kolreuter, J. G. Vorlaufige Nachricht von einigen das Geschlecht der Pflanzen

betreffenden Versuchen und Beobachtungen, nebst Fortsetzungen I, 2 u. 3.

Pp. 1-266. Ostwald's Klassiker, Nr. 41. Leipzig: Engelmann. 1 761-6.
Morgan, T. H. Some Further Experiments on Self-fertilization in Ciona. Biol.

Bull. 8. 3I3-330- 1904-
Muller, Fritz. Notizen tiber die Geschlechtsverhaltnisse brasilianischer Pflanzen.

Bot. Ztg. 26: 113-116. 1868.
Bestaubungsversuche an Abutilon-Arten. Jen. Ztschr. f. Naturwiss. 7:

22-45, 441-450. 1873-
Munro, Robertson. On the Reproduction and Cross-fertilization of Passifloras.

Bot. Soc. Edin. 9: 399-402. 1868.
Stout, A. B. Self- and Cross-pollinations in Cichorium intybus with Reference to

Sterility. Mem. N. Y. Bot. Gard. 6: 333~454-

CONNECTICUT

AGRICULTURAL EXPERIMENT
STATION

NEW HAVEN, CONN.

BULLETIN 207 SEPTEMBER, 1918

THE EFFECTS OF INBREEDING
AND CROSSBREEDING UPON
DEVELOPMENT

D. F. JONES

The Bulletins of this Station are mailed free to citizens of Connecti-
cut who apply for them, and to others as far as the editions permit.

CONNECTICUT

AGRICULTURAL EXPERIMENT
STATION

NEW HAVEN, CONN.

BULLETIN 207 SEPTEMBER, 1918

THE EFFECTS OF INBREEDING
AND CROSSBREEDING UPON
DEVELOPMENT

D. F. JONES

The Bulletins of this Station are mailed free to citizens of Connecti-
cut who apply ior them, and to others as far as the editions permit.

CONNECTICUT AGRICULTURAL EXPERIMENT STATION

OFFICERS AND STAFF

BOARD OF CONTROL.
His Excellency, Marcus H. Holcomb, ex-officio, President.

James H. Webb, Vice President Hamden

George A. Hopson, Secretary Wallingford

E. H. Jenkins, Director and Treasurer New Haven

Joseph W. Alsop Avon

Wilson H. Lee .Orange

Elijah Rogers Southington

William H. Hall South Willington

Administration.

Chemistry,

Analytical Laboratory,

Protein Research.

Botany.

Entomology.

Forestry.

Plant Breeding.

Vegetable Growing.

E. H. Jenkins, Ph.D., Director and Treasurer.

Miss V. E. Cole, Librarian and Stenographer.

Miss L. M. Bhautlecht, Bookkeeper and Stenographer.

William Veitch, In charge of Buildings and Grounds.

*John Phillips Street, M.S.

E. Monroe Bailey, Ph.D., Chemist in charge.

*C. B. Morison, B.S., C. E. Shepard,

M. d'Esopo, Ph.B. ■ Assistants.

H. D. Edmond, B.S.

Miss A. H. Moss, Clerk.

V. L. Churchill, Sampling Agent.

T. B. Osborne, Ph.D., D.Sc, Chemist in Charge.
Miss E. L. Ferry, M.S., Assistant.

G. P. Clinton, Sc.D., Botanist.

E. M. Stoddard, B.S., Assistant Botanist.

Florence A. McCormick, Ph.D., Scientific Assistant.

G. E. Graham, General Assistant.

W. E. Britton, Ph.D., Entomologist; State Entomologist.

B. H. Walden, B.Agr., First Assistant.

*l. W. Davis, B.Sc, M. P. Zappe, B.S., Assistants.
MiSS Martha de Bussy, B.A., Stenographer.

Walter O. Filley, Forester- also State Forester

and State Forest Fire Warden.
A. E. Moss, M.F., Assistant State and Station Forester.
Miss E. L. Avery, Stenographer.

Donald F. Jones, S.D., Plant Breeder.

C. D. Hubbell, Assistant.

W. C. Pelton, B.S.

* Absent on leave. In service of the United States.

CONTENTS

Page

Introduction 5

Definitions 8

Ea"rly investigations with plants 9

The observations of Darwin upon plants 12

Recent investigations with plants 14

Investigations with animals 18

Universality of heterosis 21

A theoretical consideration of inbreeding 22

The results of inbreeding the naturally cross-pollinated maize plant 27

The approach to complete homozygosity 44

The effect of heterozygosis on vegetative luxuriance 47

The value of inbreeding in plant and animal improvement 59

The effect of heterozygosis upon endosperm development and

selective fertilization 61

The effect of heterozygosis upon longevity, hardiness and viability. 69

The effect of heterozygosis upon the time of flowering and maturing. 76

The relation of the effects of heterozygosis and of the environment. 78

Summary of the effects of inbreeding and crossbreeding 81

A Mendelian interpretation of heterosis 82

The part that heterosis has played in the establishment of sex 93

Literature cited 96

The Effects of Inbreeding and Crossbreed
ing Upon Development*

INTRODUCTION.

Among the higher seed plants certain groups are characterized
by almost universal and continuous self-fertilization. On the
other hand certain other groups are completely, or to a large
extent, cross-fertilized in every generation. Between these two
extremes every gradation in the degree of self- and cross-fertiliza-
tion can be illustrated. The structure and function of the floral
organs have become more or less clearly adapted to the customary
mode of sexual reproduction characteristic of each species. In
the thallophytes, bryophytes and pteridophytes much the same
situation exists whereby the gametes which enter into a sexual
fusion may arise either from the same or from different organisms.

In the lower animals the same variation in the mode of sexual
reproduction exists as in plants. Among the higher animals,
however, hermaphroditism is replaced entirely by bisexuality; and
sexual reproduction, except when parthenogenesis takes place,
results only from the union of gametes originating in different
organisms.

This array of facts has naturally led to searching inquiries as
to the purpose of sexual reproduction as compared to other
methods of propagation as well as to the effects of artificial in-
breeding in bisexual animals and in naturally cross-fertilized
plants. Bound up with this latter problem is that which is con-
cerned with the effects of cross-fertilization in all types of animals
and plants of different degrees of relationship.

The development of the Mendelian theory of heredity, carrying
with it the conception of definable, hereditary units which are
sufficiently stable in their transmission from generation to genera-
tion to be recognized and their somatic expression to be described,

* Submitted to the Faculty of the Bussey Institution of Harvard
University in partial fulfillment of the requirements for the degree of
Doctor of Science, December, 1917.

6 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

has made possible an attack upon these problems which has
opened a way towards their solution.

From the knowledge of alternate inheritance it is possible to
ascribe, very definitely and surely, certain of the results of in-
breeding to the segregation and isolation of hereditary factors
which results were formerly thought to be due solely to inbreeding
as a cause in itself. Certain pathological, abnormal or otherwise
undesirable conditions occurring more frequently in animals
and plants produced by matings between nearly related individuals
were formerly attributed to inbreeding as the cause, and it was
thought that inbreeding must always show such undesirable
results. It is now known that many of these pathological and
abnormal conditions resulting from inbreeding do not owe their
origin to that process, but are due solely to the segregation, into
a pure state of the hereditary factors causing the anomalies
which factors were present in the organisms previous to their
being inbred. Inbreeding, then, has nothing to do with the
origin of the undesirable characters under consideration but
merely brings them into visible expression, and whether or not
they appear depends upon their presence originally in the stock
before inbreeding takes place. There still remains a conviction,
however, that all the manifestations attending inbreeding and
the converse effects of cross breeding cannot be accounted for
solely on the basis of the operation of definable, hereditary factors,
but that there is a stimulating effect resulting from crossing,
which is lost by inbreeding, and that this stimulation differs
somewhat from the expression of hereditary factors which can
be transferred and fixed in different organisms. This stimulation
is supposed to be of a physiological nature appearing when dis-
similar germ-plasms are united, and disappearing as the germinal
heterogeneity disappears in subsequent recombinations.

Since this physiological stimulation has always been purely
hypothetical, having never been definitely proven, and since it
has been used to account for certain facts heretofore inexplicable
in any other way, the existence of such a stimulation may fairly
be questioned, in so far as the facts can be logically accounted
for in other ways. Recent advances in the knowledge of the
methods of inheritance have made it possible to meet certain
objections previously held against the view that the effects of
inbreeding and crossbreeding can be attributed solely to the

INTRODUCTION.

operation of hereditary factors without assuming an additional
hypothetical stimulation.

Some of the previous work bearing upon the effects of inbreed-
ing and crossbreeding is reviewed here and with this are given
original data obtained from the naturally cross-fertilized corn
plant, Zea mays L. The facts at hand co-ordinate with the exist-
ing knowledge of heredity in such a way that it seems to the
writer unnecessary any longer to make the fundamental dis-
tinction between the effects of inbreeding and crossbreeding and
of heredity in development.

No attempt is made to canvas the extensive literature on
hybridization (a bibliography of which alone would fill a volume)
in order to list all the cases in which crossing does or does not
result in increased development and inbreeding in a reduction.
It does not take one long in reading over the many published
results of crossing in animals and plants to become convinced
that an increase in development following a cross is a frequent
occurrence. It is hoped that sufficient references are given to
show something as to the universality and nature of the phenom-
enon and a review of the more important contributions is made
in order to sketch briefly the development of the ideas concerning
the cause of the stimulation and the part it has played in evo-
lution and in breeding practice.

The experiments on inbreeding, which have resulted in the
material from which the data given here have been gathered,
were started by Professor E. M. East at the Connecticut Agri-
cultural Experiment Station and carried on by him and subse-
quently by Professor H. K. Hayes and later by the writer. From
time to time reports on these experiments have been made and
conclusions drawn from the facts as observed. These include
various publications under the titles " Inbreeding in Corn,"
" The Distinction between Development and Heredity in In-
breeding " published by Professor East in the Report of- the
Connecticut Experiment Station and in the American Naturalist
and " Heterozygosis in Evolution and in Plant Breeding " by
Professors East and Hayes in a Bureau of Plant Industry bulletin.
Under the title of " Dominance of Linked Factors as a Means
of Accounting for Heterosis " the writer had proposed a different
view as to the cause of hybrid vigor. This was published in
Genetics and its application is discussed here in more detail.

8 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Further publications are planned which will discuss more ade-
quately much of the data which are scantily treated here.

The significance which these investigations may have for the
practical improvement of plants and animals has only been briefly
alluded to here. This phase of the subject has been reserved
for another time when the methods which have suggested them-
selves as the result of these investigations have been more thor-
oughly tested. Finally this collection of facts and theories should
be viewed as a report of progress rather than a well rounded
presentation of the subject of inbreeding and crossbreeding.

The writer is especially indebted to his predecessors whose
work has made these experiments possible. Grateful acknowl-
edgement is due Dr. E. M. East for his careful supervision of the
work and for his kindly advice and helpful criticism as to the
presentation of the results obtained. The writer alone, however,
must assume the responsibility for the opinions expressed. Much
credit is due Mr. C. D. Hubbell, Dr.- Charles Drechsler and Mr.
G. A. Adsit for their careful assistance in the collection and
preparation of the data.

Definitions.

The knowledge of a stimulating effect resulting from a cross
between different animals and between different plants which
gives progeny which may excel their parents in general vigor,
size or other visible characteristics has naturally led to the use
of terms to describe this effect. This stimulation is variously
spoken of as " vigor due to crossing" or "hybrid vigor." Since
hybrid vigor occurs only in crosses of which the parents are dis-
similar in hereditary constitution more exact and comprehensive
terms were needed. The zygote resulting from a union of unlike
gametes is spoken of as a heterozygote (following the usage of
Bateson), hence the term heterozygosis (used by Spillman, '09)
refers to that germinal heterogeneity which results from the union
of unlike gametes, and the stimulation to development which accom-
panies such a condition is spoken of as a "stimulus of heterozy-
gosis," or "heterozygotic stimulation," meaning the stimulating
effects of hybridity or the stimulation due to differences in uniting
gametes. The converse fact of a reduction in vigor accompanying
a return to a homozygous condition is therefore said to be due to,

EARLY INVESTIGATIONS WITH PLANTS.

or result from, homozygosis. Shull ('14) has proposed the term
■ 'heterosis'' to designate this increase in development which may-
result from a heterozygous condition; hence, heterosis, as used
here, will be considered synonymous with "hybrid vigor" or
"stimulus accompanying heterozygosis," in whatever form this
may be manifested or whatever cause or causes it may be due to.
Shull proposed this term, as he says, "... .to avoid the implication
that all the genotypic differences which stimulate cell-division,
growth, and other physiological activities of an organism, are
Mendelian in their inheritance and also to gain in brevity of ex-
pression. ..." Hence the term heterosis is not meant as a mere
contraction of heterozygosis and is not synonymous with it. The
adjective "heterotic" has also been proposed and such an ex-
pression as "heterotic stimulation" is synonymous with heterosis.

Early Investigations with Plants.
Certain evidence remains from the carvings of the ancient
Egyptians to show that they had some conception of a sexuality
in plants. However, it was not until the last of the 17th century,
when Camerarius first demonstrated such condition, that interest
in the production of artificial hybrids began. Tt is significant
that the first artificial hybrids to be systematically studied,
those of Kolreuter (1776), furnished some of the best examples of
heterosis. Kolreuter made many interspecific crosses in Nicotiana,
Dianthus, Verbascum, Mirabilis, Datura and others, many of
which astonished their producer by their greater size, increased
number of flowers and general vegetative vigor, as compared to
the parental species entering into the cross. Concerning one of
the tobacco crosses he says: (pp. 57-58) "Hybrids obtained
from the cross of Nicotiana maj. 9 and glut, o71 produced a far
greater number of flowers and grew to an uncommonly greater
height and a much greater circumference than the pure species
under the same conditions; the height of the plants which were
kept in the hot bed or were set out in the field after they had ob-
tained full growth, amounted to eight feet and 1 to 10 inches;
the whole circumference of the branches to 24 feet; the largest
diameter of the stalks from 2 inches to 2 inches and 3 lines; and
the largest leaves were 2 feet, 2 inches and 9 lines long and 1 foot
and 4 inches wide. Never has anyone seen more magnificent
tobacco plants than these were."

10 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Thomas Andrew Knight (1799) was one among several at that
time who experimented with hybrids with the view of producing
more desirable varieties of vegetables, flowers and fruits. Knight
observed many instances of high vigor resulting from crossing;
among these we note the following remarks about a cross between
two varieties of peas.

(P. 200) "By introducing the farina of the largest and most luxuriant
kinds into the blossoms of the most diminutive and by reversing this
process, I found that the powers of the male and female in their effects
on the offspring, are exactly equal. The vigor of the growth, the size of
the seeds produced, and the season of maturity, were the same, though
the one was a very early, and the other a very late variety. I had, in
this experiment, a striking instance of the stimulative effects of crossing
the breeds; for the smallest variety, whose height rarely exceeded two
feet, was increased to six feet; whilst the height of the large and luxuriant
kind was very little diminished."

It is evident that, in these crosses, Knight was dealing with
dwarf and standard peas and the dominance of standardness is
expected. A sufficient number of cases, however, were observed
in which the crosses were more vigorous than an average of the
parents to convince him that "nature intended that a sexual
intercourse should take place between neighboring plants of the
same species." It was this principle which Darwin elaborated
50 years later.

Sageret ('26) reports vigorous hybrids in Nicotiana and also
between different types of the Cucurbitaceae. Among other
things he notes that in human crosses between one individual
which shows a hereditary pathological condition and a normal
individual, that the disease disappeared in the first generation
but reappeared in the second and following generations. Wiegmann
('28) gives instances of hybrids in the Cruciferae which showed
distinct evidences of heterosis.

Probably the most extensive series of experiments on hybridiza-
tion were those of Gartner ('49) and of Focke ('81). According
to Lindley ('52) Gartner made 10,000 crosses between 700 different
species and produced 250 different hybrids. Many of these hy-
brids showed distinct evidences of heterosis, and this phenomenon
was manifested in many different ways. Gartner speaks especially
of their general vegetative luxuriance, increase in root develop-
ment, in height, in number of flowers and their hardiness and early

EARLY INVESTIGATIONS WITH PLANTS. 11

and prolonged blooming. Focke made equally extensive; observa-
tions and catalogues his own experiments with many of those
made previously. His valuable book shows clearly that the phe-
nomenon of heterosis is widespread and may be expected in the
gymnosperms and pteridophytes as well as in the angiosperms.
Both the works of Gartner and of Focke have been so thoroughly
reviewed in recent times (East and Hayes '12) in connection with
the problem in hand that it would be a needless repetition to say
more about their results here. Special points in their observations,
as they supplement the experiments recorded here, will be referred
to later.

While the work of Gartner and Focke must always rank high as
contributions to our knowledge of genetics one cannot refrain
from remarking that they both missed by their extensive studies
of many species the point which Mendel discovered by his inten-
sive and careful study in one species.

Naudin ('65) next to Mendel will always be remembered, no
doubt, as the first to conceive of a method in the uniformity of
the first generation and the variability of the second. His con-
ception of the segregation of parental qualities as a whole leads
up naturally to Mendel's law whereby the characters of the
parents segregate as units and when finally appreciated the
chaotic observations of Gartner, Focke and their contemporaries
began to be understood as orderly facts. In Naudin's classical
experiments there are many excellent examples of heterosis.
Out of 36 interspecific crosses which he made in Papaier,
Mirabilis, Primula, Datura, Nicotiana, Petunia, Digitalis, Linaria,
Luffa, Coccinea and Cucumis, 24 show positive evidence of het-
erosis. Among the most notable crosses in this respect was that
of Datura Stramonium with D. Tatula in which both reciprocal
hybrids were twice as tall as either parent. Concerning the
Datura crosses Naudin says :

UA shap$ very much taller than the two parental types, and the pre-
mature falling off of the flowers in the first dichotomies, which leads to
tardy fructification are the principal characteristics of this hybrid of
which all the plants in the collection present the greatest uniformity.
We shall see that these different characteristics appear in all the hybrids
of this section of the genus Datura."

Mendel ('65) also records instances of heterosis in his pea
hybrids as is shown in the following passage:

12 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

"The longer of the two parental stems is usually exceeded by the
hybrid, a fact which is possibly only attributable to the greater luxuriance
which appears in all parts of plants when stems of very different lengths
are crossed. Thus, for instance, in repeated experiments, stems of 1 foot
and 6 feet in length yileded without exception hybrids which varied in
length between 6 feet and 73^ feet."

The Observations of Darwin upon Plants.

Of all the contributors to our knowledge of the effects of in-
breeding and crossbreeding no one has collected as many facts
as Darwin ('75, '77). Although undoubtedly much confusion
and misunderstanding have resulted from Darwin's conclusions
on this problem, one cannot but admire his painstaking efforts
to accumulate facts from the behavior of many species of plants
through many generations of crossing and selfing before advancing
his conclusions. No one was more frank to acknowledge the
discrepancies between the facts as he found them and the con-
clusions he drew from them. Those parts of his results which
were not clear to Darwin are clearer to us through our knowledge
of Mendelism of which he was not permitted to know. Since
his method of experimentation, and the results obtained are
familiar to all interested in the problem at hand no extensive
review of his work is necessary. Only a brief summary of the
results obtained and the conclusions which he drew from them
will be given here, reserving a more detailed review of special
parts for a later part of this paper.

Among animal breeders in Darwin's time it was a common
belief that whatever evil effects resulted from more or less close
inbreeding were due to the accumulation of abnormal, diseased,
or morbid tendencies in the offspring of parents which possessed
such tendencies. Darwin refused to ascribe any large part of the
effects of inbreeding to this cause because he knew so many cases
were weakened and reduced types of both plants and animals
which gave vigorous progeny when crossed among themselves.
Instead of an accumulation of the undesirable traits of both
parents the very reverse seemed to be true. Had Darwin known
of the way by which recessive characters may exist for many
generations without making their appearance, doubtless his views
on this point would have differed materially.

Darwin clearly thought that the evil effects of inbreeding kept
on accumulating until eventually a plant or animal propagated

THE OBSERVATIONS OF DARWIN UPON PLANTS.

in that manner was doomed to extinction. His own results came
far short of proving such an assumption. The two wild plants
with which inbreeding was practiced the longest — Ipomea and
Mimulus — showed .very little further loss of vigor after the first
generation. What these experiments did show, most clearly,
was that there was segregation of the inbred stock into diverse
types which differed in minor, visible, heriditary characters and
which also differed in their ability to grow. In both species
plants appeared which were superior to other plants derived
from the same source and some were even equal or superior in
vigor to the original cross-pollinated stock. They differed from
this race, however, most noticeably in the uniformity of all
visible characteristics.

After several generations of inbreeding Darwin found that it
made no difference in the resulting vigor, whether the plants in
an inbred lot were selfed or were crossed among themselves.
This he correctly attributed to the fact that the members of
such an inbred strain had become germinally alike. From his
views on the effect of the environment on organisms, it is easy
to see why he attributed this approach to similarity in inherited
qualities to the fact that the plants were grown for several
generations under the same conditions. This view he thought
was supported by the fact that crosses of his selfed lines with
the intercrossed lines (also inbred, but to a less degree) did not
give as great increase in vigor as the crosses of either lines with
a fresh stock from distant regions. The crosses between two
inbred lines did give a noticeable increase in vigor, in many
cases, equaling the original variety. This is illustrated in the
Dianthus crosses in which the selfed line was crossed with the
intercrossed line and with a fresh stock. The ratio of both
crosses to the selfed plants in height, number of capsules and
weight of seed produced is as follows:

Selfed
X

Inter-crossed

Selfed
X

Fresh stock

Height, compared to selfed

No. Capsules, compared to selfed. .
Weight of seed, compared to selfed

100:95
100:67
100:73

100:81
100:39
100:33

Like Darwin we now attribute the greater increase of vigor
in a cross with distinct stocks to a greater germinal diversity
although we may differ in our ideas as to the way in which that

14 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

diversity was brought about. Whatever may be the explanation
of that, credit is due Darwin for being the first to see that it was
not the mere act of crossing which induced vigor but the union
of different germinal complexes. This he states clearly in the
following sentences (Cross and Self Fert., p. 270):

" These several cases taken together show us in the clearest manner
that it is not the mere crossing of any two individuals which is beneficial
to the offspring. The benefit thus derived depends on the plants which
are united differing in some manner, and there can hardly be a doubt
that it is in the constitution or nature of the sexual elements. Anyhow,
it is certain that the differences are not of an external nature, for two
plants which resemble each other as closely as the individuals of the same
species ever do, profit in the plainest manner when intercrossed, if their
progenitors have been exposed during the several generations to different
conditions."

Recent Investigations with Plants.

Although Darwin was the first to attack the problem from the
standpoint of determining the effects of inbreeding, it is doubtful
if he clearly recognized that the same phenomenon was concerned
in both inbreeding and crossbreeding. It remained for Shull ('08,
'09, '10, '11 and '14), East ('08, '09) and East and Hayes ('12)
to bring out clearly the fundamental similarity of both processes
and to put the matter in such a light that a far clearer under-
standing of the nature of the effects of inbreeding has resulted.

Their conclusions in regard to the causes of the effects of
inbreeding and crossing were for the most part entirely new and
dependent for their support upon the Mendelian principle of the
segregation and recombination of inherited qualities as inde-
pendent units and upon Johannsen's genotype conception of
heredity. Stated briefly their main tenets, based upon their
own careful experiments and a survey of previous results bearing
upon the problem, are as follows:

1. Inbreeding automatically sorts out into homozygous, pure
breeding lines, the diverse and variating complex of hereditary
characters found in a naturally cross-pollinated species.

2. Although complete homozygosity is difficult to attain in
practice, after several generations of selfing, members of the
resulting inbred lines are uniform among themselves but the
respective lines may differ greatly among each other in visible

RECENT INVESTIGATIONS WITH PLANTS.

hereditary characters. The strains may also differ in their power
of development, some being larger, stronger and more productive
than others at normal maturity. Some individuals are often
isolated which are so lacking in necessary characters that they
perish because of inability to reproduce themselves.

3. Those inbred strains which are able to survive finally be-
come constant; no further reduction in vigor or change in visible
characters is to be expected by continued inbreeding. These
constant types are thus quite comparable to naturally self-
fertilized species and may exist indefinitely.

4. When these pure breeding types are crossed there is com-
monly an immediate and striking increase in general size and
vigor to be expected in the resulting first hybrid generation.

To account for this increase in development, following a cross,
a physiological stimulation was postulated which accompanied
heterozygosity of hereditary factors and disappeared as the
organisms approached homozygosity. As an illustration the
union of factor "A" with it allelomorph "a" was considered to
evolve developmental energy which was lacking when either
" A" or "a" were united with themselves. This stimulus to devel-
opment was considered to be due to the union of unlike factors
alone and to have an effect quite different from whatever part
each factor had by itself in the development of the organism.
Stated in their own words the main conclusions of East and
Hayes ('12) are as follows (p. 8):

" 1. Mendel's law — that is, the segregation of character factors in
the germ cells of hybrids and their chance recombinations in sexual
fusions — is a general law.

2. Stimulus to development is greater when certain, or possibly all,
characters are in the heterozygous condition than when they are in a
homozygous condition.

3. This stimulus to development is cumulative up to a limiting point
and varies directly with the number of heterozygous factors in the
organism although it is recognized that some of the factors may have a
more powerful action than others."

It was clearly apparent to recent investigators that many of
the unfavorable characters which appear on inbreeding a naturally
cross-pollinated species are recessive characters which are segre-
gated out of the original complex. In a naturally crossed species,
these are hidden from sight on account of being continually

16 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

crossed with dominant characters. That dominance of factors
could in any way be an essential factor in the vigor and excellence
of hybrids, an idea first proposed by Keeble and Pellew ('10)
and also by Bruce ('10), has not been accepted by most writers
on this subject. They considered dominance to be totally in-
adequate to account for the widespread and almost universal
occurrence of heterosis in plants and animals and the fact that
nearly all naturally cross-fertilized domesticated species are
reduced by inbreeding.

Collins ('10) has shown clearly that many crosses between
varieties of Indian corn already widely crossed among themselves
and grown in the same regions may not give any increase in
productiveness, but when these same varieties are crossed with
varieties from distinct geographical regions great increases in
productiveness are obtained. Further evidence as to the occurrence
of heterosis is seen in the many publications which have appeared
from time to time urging the commercial utilization of this hybrid
vigor as a method of increasing production in many plants.
Among these are Beal (76- 82), McCleur ('92), Morrow and
Gardner ('93-'94), Swingle and Webber ('97), Hayes and East
('11), Hartley ('12), Wellington ('12), Hayes ('13), Hayes and
Jones ('16).

In view of the innumerable cases in which an increase in devel-
opment, in some character, results from crossing and the converse
fact of reduction following subsequent inbreeding, of which the
preceding paragraphs refer to only a small fraction, it is surpris-
ing to note such radically diverse opinions as are held by Burck
('08) and championed by Stout ('16).

Stout attributes the following statements to Burck: (p. 418)

"That (1) plants that are regularly self-fertilized show no benefits
from crossing and that (2) nowhere in wild species is there evidence of
an injurious effect from self-fertilization, and that there is abundant
evidence of continued vigor and high fertility resulting from long con-
tinued self-fertilization."

If by the first statement is meant that crossing between members
of the same variety or between individuals of a uniform species
does not give an increase in development such a result would be
expected because of the germinal similarity brought about by
long continued selfmg and elimination by selection, either natural

RECENT INVESTIGATIONS WITH PLANTS.

or artificial, of all but one type. On the other hand, there is
abundant evidence to show that crossing between different vari-
eties or between different wild species of self-pollinated plants
often results in striking increases in size and vigor. It is only
necessary to refer to the work of Kolreuter, Knight, Gartner,
Naudin and Mendel where many crosses between different species
or between distinct types of Nicotiana, Pisum, and Lathyrus —
plants which are naturally self-fertilized — give unmistakable
evidence of heterosis.

Turning to the effects of inbreeding, almost no long-continued
experiments have been carried out with strictly wild cross-polli-
nated species of plants. Collins ('18) in a brief note states that
teosinte, a semi-wild relative of maize, is not affected by in-
breeding to the extent that maize is. That there is ' 'abundant
evidence of continued vigor and high fertility resulting from
long continued self-fertilization' ' no one longer doubts. There is,
however, hardly enough evidence from plants, so far on record,
to justify the sweeping statement, which the quotation implies,
that cross-fertilized wild species are never reduced by inbreeding.

What evidence there is indicates that naturally crossed wild
species are not reduced by inbreeding to anything like the extent
that domesticated races are. More will be said about this differ-
ence between wild and domesticated races later. There is some
evidence, however, to show that strictly wild species are affected
by inbreeding. Darwin compared the progeny of artificially self-
fertilized plants with the progeny of artificially intercrossed
plants of many wild species. Many of these species were such as
were for one cause or another almost completely cross-fertilized
in their natural state at all times. Although the difference may
be slightly exaggerated there can be no question but that the
difference in the first generation which Darwin obtained between
the selfed plants and the intercrossed plants represents in many
cases the effect which inbreeding has upon these plants. As
examples of widely crossed wild species in which a reduction in
the first generation of inbreeding was obtained by Darwin, one
can, therefore, cite: Digitalis purpurea, Linaria vulgaris, Saro-
thamnus scoparius and Reseda lutea.

Moreover, no matter how much domestication may change
plants from the wild, one cannot cast aside, as of no consequence,
the results obtained from cultivated plants.
1*

18 connecticut experiment station bulletin 207.

Investigations with Animals.

According to Darwin, the mule, that classic example of hybrid
vigor, was known in the time of Moses, when its hardihood and
general good qualities doubtless endeared this animal to the Jews
no less than to the Southern cotton planters of to-day. A similar
cross of the ass with the wild zebra according to Riley ('10) gives
a first generation hybrid animal of considerable merit.

In the early history of the establishment and fixation of breeds
of livestock we note in Darwin's " Animals and Plants under
Domestication" that certain crosses between different breeds often
resulted in progeny excelling individuals of either parent breed;
just as to-day it is not an uncommon practice for livestock raisers
to cross certain well-established breeds to produce crossed animals
to feed for market.

In looking over the reports of experiments designed to test the
effects of crossing in both wild and domesticated animals there is
little disagreement as to the results usually obtained. All are
practically in accord that crossing diverse breeds or races of
animals, if not too distantly related, may frequently result in
vigorous, large and fertile offspring, excelling either parent in
one or more respects. For example, Castle et al ('06) find that
crossing diverse stocks of Drosophila results in an increase in
fertility and that matings between different inbred lines give
progeny with increased fertility up to or beyond that of the more
fertile parental race. In Meriones Bonhote ('15) states that
fertility and size are increased by crossing. Castle ('16) has
crossed domesticated races of guinea-pigs with the wild species
from Peru with the result that there is a noticeable increase in
body weight over either pure parent. Gerschler ('14) crossed
different genera of fishes and obtained large increases in size in
the first hybrid generation. Xiphophorus strigatus, of which the
males were 43.0 cm. long and the females 52.0 cm., when crossed
with Platypoecilius maculatus, of which the males were 26.0 and
the females 31.0 cm. in length, gave hybrid males 54.0 cm. and
females 57.5 cm. He speaks of their ''gigantic size."

Fischer ('13) in his study of the Rehoboth hybrids, a race in
South Africa resulting from a mixture of Hottentots and Boers,
states that their average height is somewhat greater than either
the Hottentots or the Hollanders and South Germans of whom

INVESTIGATIONS WITH ANIMALS.

statistics are available. All the members of this new race are not
first generation crosses by any means, but they are not many
generations removed and crossing with the pure Hottentots, the
shorter parental race, is frequent.

When, however, the literature on the effects of inbreeding in
animals is examined one finds the greatest diversity of facts and
opinions. We find the extreme views of KTaemer ('13) who states
that "continued inbreeding always must result in weakened con-
stitution, through its own influence" together with the equally
extreme and biased opinion of Huth (75) that in mankind there
is no injurious effect resulting from consanguineous marriages
which cannot be accounted for on other grounds.

Crampe ('83), Ritzema-Bos ('94), Guaita ('98), Fabre-Domengue
('98) and Weismann ('04) by inbreeding mammals and birds
found that the process was accompanied by decreased fertility,
attended more or less commonly by lack of vigor, diminution in
size, and pathological malformations. Castle, Carpenter et al ('06)
inbreeding extensively the fruit fly, Drosophila, maintained fer-
tility by selection, so that at the end of 59 generations of brother
and sister matings in one line the fecundity was no less at the end
of the experiment than it was at the start. There was some indi-
cation of reduction in size of inbred flies when compared to nor-
mally crossed stock flies reared under the same conditions. Fur-
thermore, fertility was increased by crosses between certain
inbred lines and between the inbred lines and stock flies. From
this fact and from the fact that their experiments show that the
number of flies in a brood fluctuates greatly, due to temperature
and food conditions, it is not positive that inbreeding was wholly
without injurious effects. It is evident that their experiments do
show clearly:

1. That inbreeding results in strains of unequal fertility.

2. That the occurrence of absolute sterility was pronounced in
the first part of the experiment with the "A" line but almost
entirely disappeared in the later part of the experiment. The
figures as I have calculated them from their table I, p. 736, are
as follows:

Percent of matings
totally sterile

Generations

6 to 24
25 to 42
43 to 59

17.80
18.47
3.37

20 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

This result is to be expected on the view that inbreeding isolates
homozygous individuals and these whenever sterile are, of course,
eliminated.

Moenkhaus ('11) and Hyde ('14) by similar inbreeding experi-
ments with Drosophila have also found that sterility is increased
in the first stages of inbreeding but tends to be eliminated after
this process is long continued. Hyde found definite evidence that
inbreeding caused reduction in size, vigor, rate of growth, longevity
and fecundity and that there was a return to the normal condition
on crossing. As in the other experiments Hyde found that selec-
tion was an effective agent in controlling sterility.

Both Whitney (12a) and A. F. Shull (12a) have shown that
inbreeding and crossbreeding have considerable effect upon the
rotifer, Hydatina Senta, in the size of family, number of eggs laid
per day, rate of growth and in the difficulty of rearing the animals.

King ('16) has obtained results with albino rats which are
quite in agreement with those of Castle. By growing about one
thousand rats in each inbred generation, and selecting the best
individuals for mating, animals have been carried through 22
generations of brother and sister matings without loss of size,
fertility, longevity, resistance to disease and with constitutional
vigor unimpaired. This writer states:

. "The results so far obtained with these rats indicate that close inbreed-
ing does not necessarily lead to a loss of size or constitutional vigor or of
fertility, if the animals so mated came from sound stock in the beginning
and sufficient care is taken to breed only from the best individuals."

Here, as in Drosophila, inbreeding isolates diverse types of
different degrees of excellence. In this case individuals are ob-
tained which surpass the original stock before inbreeding. Thus
we have "Goliaths" among inbred rats as Darwin found "Heroes"
in morning-glories.

Castle ('16) has found that in inbred rats "races of fair vigor and
fecundity can be maintained under these conditions, but that when two
of these inbred races are crossed with each other, even though they have
their origin in a small common stock many generations earlier, an imme-
diate and striking increase of fecundity occurs."

The evidence from relationship marriages in human stocks is
even more conflicting and conclusions still more difficult to draw.
Huth ('75) has certainly done a service in showing that consan-

UNIVERSALITY OF HETEROSIS.

guineous marriages seldom result in the disastrous effects usually
attributed to them. He has shown that incest was not a rare
custom and that races which have undergone such practices are
many of them far from weak. Certainly, races have practiced
close intermarriage for many generations with no marked deterio-
ration. The Persians, Spartans, the ruling classes among the
Egyptians and Polynesians are cited by Huth in support of this
assertion. The data from human matings, however, are of little
value since the close unions are seldom continued many genera-
tions in succession, and the results from isolated communities mean
little, since often the original stock is exceedingly diverse so as to
make the resulting races extremely heterogeneous in hereditary
constituents. This is particularly true of the Rehoboths and the
Pitcairn Islanders which are cited as instances of close inter-
marrying without loss of racial vigor.

Looking over the experiments upon animals it seems as unwise
to expect that inbreeding may not have some deleterious effects,
which, in some cases at least, cannot be overcome by the most
rigid selection, as it is to hold that inbreeding must always result
injuriously. It is to be expected that all breeds of domestic
animals and wild species will not be equally affected by inbreeding.
Domesticated animals in many cases are more widely crossed and
diversified than wild species, and those characters affected by
inbreeding are more accentuated. Certain wild species, which,
by their mode of life, are forced to endure long periods of isolation,
and consequently more or less close inbreeding, would be expected
to show less change under artificial inbreeding. Finally, as I
shall attempt to show that there is no longer a question as to
whether or not inbreeding, in itself, is injurious, the effect which
inbreeding will have on any organism depends solely on the
hereditary constitution of that organism at the time the inbreed-
ing process is commenced.

Universality of Heterosis.

From the literature on the subject of crossbreeding it is to be
observed, therefore, that the occurrence of an incentive to in-
creased development accompanying germinal heterogeneity is
widespread, as it has been noted in plants in the angiosperms,
gymnosperms and pteridophytes, and according to Britton ('98)

22 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

there is even some slight evidence that heterosis occurs in the
sporophyte of the bryophytes.

In animals the mammals, birds, fishes, insects and rotifers show
the phenomenon of heterosis although in some of the unicellular
animals, as we shall see later, the evidence is not so clear.

I shall now take up, in some detail, experiments on inbreeding
and crossbreeding in cultivated plants, principally in maize.

A Theoretical Consideration of Inbreeding.

Up to the present time it has been maintained that the effects
of inbreeding were of two kinds, an isolation of homozygous
biotypes together with a loss of a physiological stimulation which
was considered to be roughly proportional to the number of heter-
ozygous allelomorphs present in the organism at any time. The
reduction of the number of heterozygous allelomorphs in an inbred
population is automatic and varies with the closeness of inbreeding.

Pearl ('15) on the basis of the number of ancestors which make
up the pedigree of any individual has worked out a coefficient
of inbreeding which is an indication of the degree to which that
individual has been inbred. The fewer the number of ancestors
the greater the degree of inbreeding which may vary from no
inbreeding, in which no one ancestor appears more than once
in the pedigree of an individual, to the closest kind of inbreeding
in which no more than one ancestor is concerned in any one
generation in the production of an individual (self-fertilization).
The latter degree is only approached by hermaphroditic plants
and animals, which are capable of self-fertilization and in function-
ally bisexual animals and plants by brother and sister matings.
This statement of inbreeding must, of course, leave out of con-
sideration any germinal change which might take place by means
other than hybridization and as Castle ('16) has pointed out is
modified by the differences in heteroz3rgosity of the ancestors
making up the pedigree.

The automatic reduction in the number of heterozygous allelo-
morphic pairs in an inbred population, by self-fertilization,
follows the well known Mendelian formula by which any hetero-
zygous pair forms in the next generation 50 percent homozygotes
and 50 percent heterozygotes in respect to that pair. Since the
homozygotes must always remain homozygous and the hetero-
zygotes are halved each time and one half added to the homo-

A THEORETICAL CONSIDERATION OF INBREEDING. 23

zygotes the reduction in the number of heterozygous elements
proceeds as a variable approaching a limit by one half the differ-
ence in each generation. The curve illustrating this condition
is shown as No. 1 in Fig. I. Various formulae dealing with

0 1234 56 789 10

Segregating Generations

Figure I. The percent of heterozygous individuals and the percent of
heterozygous allelomorphic pairs in the whole population in each
generation of self-fertilization.

inbreeding have been discussed by East and Hayes ('12), Jennings
('12, '16), Pearl ('15) and Bruce ('17).

It should be remembered that this reduction applies only to
the whole population in which every member is inbred and all

24 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

the progeny grown in every generation. In practice, in an
inbreeding experiment, only one individual in self-fertilization or
two individuals in brother and sister matings are used to produce
the next generation. Thus the rate at which complete homo-
zygosis is approached depends on the heterozygosity of the
individuals chosen. Theoretically in any inbred generation the
progenitors of the next generation may either be completely
homozygous or completely heterozygous or any degrees in between
depending upon chance. The only condition which must follow
in self-fertilization is that no individual can ever be more hetero-
zygous than its parent but may be the same or less. Thus it is
seen that inbreeding, as it is practiced, may theoretically never
cause any reduction in heterozygosity, or it may bring about
complete homozygosity in the first inbred generation. In other
words the rate at which homozygosity is approached may vary
greatly in different lines. However, as the number of heter-
ozygous factors at the commencement of inbreeding increases the
more nearly will the reduction to homozygosity follow the curve
shown because the chance of choosing a completely homozygous
or completely heterozygous individual in the first generations
will become less.

In Table 1 is shown the theoretical classification of the progeny
of a self-fertilized organism which was heterozygous with respect
to 15 independent mendelizing units. It will be seen that the
bulk of the individuals lie between classes 6 and 11 where none of
the members are heterozygous for more than 10 factors nor less
than 5. In other words any individual selected for the progenitor
of the next generation would probably come from the middle
classes and therefore it would be heterozygous for about half
the factors that its parent was. The chance that this individual
would not come from the mid-classes between 6 and 11 would be
about 1 out of 10. The chance that it would be completely homo-
zygous or completely heterozygous would be 1 out of 32,768. If
20 instead of 15 factors were concerned the chances would be
1 out of 1,048,576.

This condition by which the progenitor of each generation tends
to be half as heteroz3^gous as its parent holds true for any number
of factors and in every generation. Also in Table 1 it can be seen
that the progen}' as a whole has an equal number of heterozygous
factor pairs as homozygous factor pairs in respect to those

A THEORETICAL CONSIDERATION OF INBREEDING.

S 5

£ °
o .c

S.s

OiCOiOOiOOiOOiOOiCOiOOiO

i-ti-HCOCO'-ICOTt<00^COi-ICOCO'-l'-H

(NMTfOOOT)*OOOr(*M(N

HiOiOO'OHiOO»0»CH
»-* CO *C Tfi CO »-<

iOOiCO>00»flOiOOiOO"OOLOO
INMi*OOO^OOO^n01

HiOiOO»CHiCOiOiOH

H M lO "<t M H

.5 ic

"2 J3

cu oj e3

I 5 +

o 1 5

£ 13
^ S w

5 "9 o
3 c o

C c3 w

1-hO^OOOOCOCCOOOiOOi-h
HCOiOCOCOiOMH

00 =
CO <M

ei
co

5 O =3

6 =

p 5

e3 T3
'•3 * »

> £ .£ .5

OOOOOOt^OOOOtJhOOOOOO

C*fi-T 0~cT 00~ <N CO (NCM CO <N 00 C& © rH <M
COOI^CNOOOCDOONO^OIM
■^"^05l>Tt"O0000OTtH^.C3^-^

co TjTTjT oo r^cTo^ oo rri rtH co

ht}(OC0hhc0 05^h

i-H <M <N rH

00 CO

i-lCMCO^iOcDl>00050'-i<MCOT^iOCD

2 +

26 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

characters in which the parent was heterozygous. So it is that
in practice the reduction in growth accompanying inbreeding
(on the assumption that heterosis is correlated with heterozygosity)
is greatest at first, rapidly becomes less and finally ceases for all
practical purposes.

If there were no deviating factors the curve of reduction should,
in the majority of cases, approximate curve 1 in Fig. I. However,
it has never been assumed that the amount of heterosis was
perfectly correlated with the number of heterozygous factors.
Moreover, since the heterozygous individuals are more vigorous
than the homozygous, selection, either unconscious or purposeful,
would favor the more heterozygous so that the tendency might
be that the actual approach to homozygosity would not proceed
at as fast a rate as the theoretical curve would indicate.

Self-fertilization is the quickest and surest means of obtaining
complete homozygosis for the reason that whenever any pair of
allelomorphs becomes homozygous it must always remain so long
as self-fertilization takes place, whereas in brother and sister
mating a homozygote may be mated to a heterozygote. Thus
we see from Jennings' ('16) tables that 6 generations of self-
fertilization are more effective than 17 generations of brother
and sister matings in bringing about homozygosis. The reduction
in heterozygous allelomorphs in a population as a whole follows
curve 1 in Fig I irrespective of the number of factors concerned,
provided, as stated before, that a random sample of all the different
classes of individuals are selfed and used as progenitors for the
next generation and that there is equal productiveness and equal
viability. If the heterozygotes are more productive, as in many
cases they are, the reduction to complete homozygosity will be
delayed.

The number of completely homozygous individuals in any gener-
ation, inbred by self-fertilization, differs according to the number
of heterozygous factors concerned at the time that the inbreeding
process is commenced. The curves showing the reduction in the
number of individuals heterozygous in any factors, where, 1, 5,
10 and 15 factors are concerned at the start are given in Fig. I
calculated from the formula given by East and Hayes ('12).
The curve for the reduction in heterozygous individuals, where
one factor only is concerned, is identical with the curve showing
the reduction in heterozygous factors in an inbred population

RESULTS OF INBREEDING.

where any number of factors are concerned. In any case almost
complete homozygosity is reached in about the tenth generation
on the average, although theoretically it may be reached in the
first generation, or may never be reached when a single individual
is used in each generation to perpetuate the line.

Assuming, then, that the loss of the stimulation, accompanying
heterozygosity, is correlated with the reduction in the number
of heterozygous allelomorphs we should expect to find the decrease
of heterosis greatest in the first generations, rapidly becoming
less until no further loss is noticeable in any number of subsequent
generations of inbreeding, and that, on the average, the loss will
become negligible at about the eighth generation and from then
on no further marked change will take place. Some cases are
to be expected in which stability is reached before this generation
and some cases in which it is not reached until later or may even
theoretically never be reached. With these assumptions in mind
let us see what are the actual results of long continued inbreeding
in maize.

The Results of Inbreeding the Naturally Cross-pollinated

Maize Plant.

The behavior of maize during six generations of inbreeding by
self-fertilization has already been reported by East and Hayes
('12). The same inbred strains have been continued and in some
cases the results up to the eleventh generation are given here.

In the previous publication it was stated that a loss of vege-
tative vigor has followed every case of inbreeding in maize.
Some plants had been obtained which were unable to reproduce
themselves. Those strains which were maintained became uni-
form but differed considerably from each other. It was con-
sidered at the end of the period of inbreeding that some strains
were appreciably better than others in their ability to yield.
Six additional years of inbreeding with this material has confirmed,
in the main, these conclusions. A further appreciable reduction
in productiveness, however, has taken place in all lines together
with certain changes in various parts of the plants.

The original experiment began with four individual plants
obtained from seed of a commercial variety of Learning dent
corn grown in Illinois. This variety was given the number 1

28 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

and the four plants which were self-pollinated and selected for
continuation of the inbreeding experiment were numbered 1-6,
1-7, 1-9 and 1-12. These four strains were continued each year
by self-pollination. In the second inbred generation two self-
pollinated plants in the 1-7 line were saved for seed and from them
two inbred lines were split off which therefore came originally
from one line inbred two generations. These are numbered
1-7-1-1 etc. and 1-7-1-2 etc. In a similar way these, and the other
inbred lines, were further split up in subsequent generations.
After the experiment was started with the dent corn inbreeding
was commenced with other material. Two inbred strains of
floury corn, Nos. 10-3 and 10-4, originally from the same variety,
have been maintained and also two strains of flint, Nos. 5 and 29,
and two strains of popcorn Nos. 64 and 65. Chief attention has
been paid to the inbred strains of Learning corn (the longest
inbred) and most of the data presented here have resulted from
this material. Many other varieties besides these have been
inbred for many generations in connection with other investi-
gations and while they are not specifically mentioned the observa-
tions as a whole include these.

In Tables 2 and 3 the yield and height of some of these inbred
strains are given. In 1916 seed of the original Learning variety
was obtained which had been grown in the meantime in the same
locality whence it was originally secured and was grown for
comparison with the inbred strains. This variety in Illinois in
1905 yielded at the rate of 88 bushels per acre, and in Connecticut
in 1916 at the rate of 74.7 bushels. While there is no proof that
any change has not taken place in the original variety there is no
reason to suppose that it has changed to any great extent. Grown
under the same conditions in 1916 the four inbred Learning
strains yielded from one-third to one-half as much as the original
non-inbred variety.

With regard to rate of reduction in yield or the constancy of
the varieties during the later generations it is difficult to draw
conclusions from these figures owing to the fluctuation in yield
from year to year due to seasonal conditions and to the difficulty
of accurate testing in field plot work, which is recognized by all
who have made such tests. As was stated in the first report the
yields for 1909 were too low and in 1911 much too low on account
of poor seasons. No yields were taken on any of the strains in

RESULTS OF INBREEDING.

Table 2. The effect of inbreeding on the yield and height of maize.

Year
grown

No. of
genera-
tions
selfed

Four inbred strains derived from a variety of Learning dent corn.

1-6-1-3-etc.

1-7-1-1-etc.

1-7-1-2-etc.

1-9-1-2-etc.

Yield
bu. per
acre

Height
inches

Yield
bu. per
acre

Height
inches

Yield
bu. per
acre

1 1 1 > i if 1 1 1

inches

Yield
bu. per
acre

rieigou
inches

1916
1905
1906
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917

0
0

2
3
4
5
6
7
8
9

10
11

74.7
88.0
59.1
95.2
57.9
80.0
27.7

117.3

74.7
88.0
60.9
190759 . 3

190846 . 0

63.2
25.4

111.3

74.7
88.0
60.9
190759 . 3
190859 . 7

68.1
41.3

117.3

74.7

88.0

42.3

51.7

35.4

47.7

26.0

191338.9
191445 4

191521.6

191630.6
191731.8

117.3

86.7

81.1

90.5

76.5

41.8
78.8
25.5
32.8
46.2

39.4
47.2
24.8
32.7
42.3

85.0

78.7
82.4

96.0

83.5

58.5

88.0

97.7
103.7

84.9
78.6

19.2
37.6

86.9
83.8

Table 3. The effect of inbreeding on the yield of maize.

Year
grown

No. of
genera-
tions
selfed

Two inbred strains of
floury corn

One inbred strain of
flint corn

10-3-7-etc.
Yield
bu. per
acre

10-4-8-etc.
Yield
bu. per
acre

5-8-6-etc.
Yield
bu. per
acre

1908

70.5

75.7

1909

56.0

43.0

47.5

1910

67.0

48.7

36.1

1911

39.1

29.3

11.5

1912

1913

32^2

49^5

30!i

1914

52.6

38.1

1915

1916

13.9

1Q.Q

18^3

1917

26.6

24.0

30 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

1912. The yields in 1914 are too high and in 1915 too low for the
same reasons. Also in 1915 the yields are unreliable because
only a few plants were available to calculate yields from as most
of them were used for hand pollination. During the last three
years of the test samples of corn have been dried to a uniform
moisture basis and the yields calculated to bushels of shelled
corn per acre with 12 per cent, moisture. This has probably
had a tendency to reduce the yields somewhat as these inbred
strains are very late in maturing and consequently contain large
amounts of water.

With these points in mind an examination of the table shows
that from the beginning of the experiment to the ninth generation
there has been a tremendous drop in productiveness, so that in
that generation the strains are approximately only one-third as
productive as the variety before inbreeding. From the ninth
to the eleventh generation there has been at least no reduction
in productiveness, and practically no change in visible plant or
ear characters.

In the previous publication it was stated (U. S. Dept. of Agric,
B. P. I. Bull. 243, pp. 23-24) that

. . . . " strain No. 6, is a remarkably good variety of corn even
after five generations of inbreeding. It yielded eighty bushels per acre
in 1910. The yield was low in 1911, but since all yields were low that
year it can hardly be doubted that this strain will continue to produce

good normal yields of grain The poorest strain, No. 12, is partially

sterile, never fills out at the tip of the ear and can hardly exist alone.
In 1911 it yielded scarcely any corn but will no doubt continue its exist-
ence as a partly sterile variety."

These statements will have to be modified somewhat. Although
No. 6 is, in the eleventh generation, still the most vigorous inbred
strain, as a producer of grain, however, it can hardly be considered
to give " good normal yields." The plants, nevertheless, are
perfectly healthy and functionally normal in every way except
for an extreme reduction in the amount of pollen which they
produce. The strain No. 12 was lost. Since the difficulty of
carrying along any inbred strain is very great owing to failure
to pollinate at the right time, attacks of fungus on the ear enclosed
in a paper bag, and poor germination in the cold, wet weather
common in New England at corn planting time, the loss of this
strain might be easily accounted for without supposing that it

RESULTS OF INBREEDING.

simply ran out. It may be that this strain could have been
perpetuated if sufficient effort had been put forth to do so. In
view of the further reduction in the other strains, however, the
maintaining of this strain would have been extremely difficult.

Complete records on the height of plant are wanting for many
of the generations, and, unfortunately, in the first part of the
inbreeding period. What figures are available certainly show
that very little change in height has taken place in all four strains
during the last seven generations. Strain No. 6 has increased
in height, if anything. Height is less affected by environmental
factors than is yield and in that respect is a more reliable indicator.
However, great changes in the structure, size and productiveness
may take place without height of plant being greatly altered.

From the figures given in Table 2 there is some evidence that
these strains have reached about the limit of reduction in pro-
ductiveness and that there has been very little change in the last
three years. This, however, is not proven. The continuation
of inbreeding is necessary for conclusive evidence on this point .
As the crosses between individual plants within these inbred
strains have given very little increase over the selfed strains,
as will be shown later, and from the fact that almost no visible
change has taken place in these four strains during the past three
years that I have had them under observation, it seems apparent
to me that the reduction in vegetative vigor and productiveness
is very nearly at an end.

In Tables 4, 5, 6 and 7 are given the frequency distributions of
height, length of ear, number of nodes and the number of rows
of grain on the cob of the original, non-inbred Learning variety
and several inbred strains derived from this variety after nine
or ten generations of selfing. All the plants from which the data
were taken were grown on the same field in the same year. Four
different plots of the variety were grown in different parts of the
field and the data on these plots are given separately and totaled
in the tables. It can be seen from these that no great variations
in range, mean, standard deviation or coefficient of variability
were caused by environmental factors. The pedigree numbers
show the relationship of the several inbred strains to each other.

From these tables it can be seen that both height of plant
and length of ear have been reduced, but in different degrees
in different lines. In some strains reduction in height amounts
to 40 inches and in length of ear to 3.5 inches. The reduction in

-H -H -H -H -H

CM OO CO CO CO
00 GO CO 00 00

CO CO

-H -H -H -H

* O ^ H

■* tC iO

CO lO

-H-H-H-H

lO OS
N CD (N

O O >0 h

-H -H

o o

00 t>

-H -H

CO rj<

t» o

-H -H -H -H -H -H -H -H -H -H-H-H-H -H-H

CO CM

■H -H

o o io o co

1< KJ O M CC

«fl «5 K5 «5
O CM l> <M

lO iO ifl lO
CO OS

CO CM

OS O

C5 o o o o

■* IO KJ

t» 00 CO

CO b»

»o

iH .-H O CO 00

CD N N N
CO M lO

— < l> CD O

b- l>

^ 00

co co

CM iO
IO CO

-H -H -H -H -H

-H-H-H-H

00 N CO N

-H-H-H-H

lO IN N S

-H -H

CO CM

■W -H

^ tj O N

N CO CO O)
0)0 0)0

00 CM © 00
N » O) 00

CM i-H

00 OJ

t> o

CM iO CO CO CO

CO CO N 00 H

>-H CD O CO

>o tj< Tt* r~-

CM i-H O t>»

CM »0 T-t i-H

N H H if)

O CM CO t— 00 00

u 3

_ v

.5 >>

CO (N H H

^ H U3 W

rj< Ttf IT2 IO

CM CM CM CM

^ Tt^ ^t* ^

TH ^ ^
lilt

CO CO CO CO

co co co co

H H (N (N H CO

CO H N U3 Tf< rH

CM CM Tt< T}< H H

»C »0 "2 »C H H

t> t>- CM CM

Tf< T*< O O

^ rH rH CM CM

t»- t- r~-

Tj< co

co co

iO "0

tf tf

CO CO

CM CM

RESULTS OF INBREEDING.

333-H-H 4l -H -H -H MMMM MM M M

tO ©

CO Tt<

CO O

-HNrHrHrH rH rH CM i-H CM (N -H

CO to
CM rH

r>- cm

t> CO

I 00 —■ 1 OC 00 'O MCHN

I © -H o o o

-H -H -H -H -H

I MhhOW

O O o

-H-H-H-H

O 00 tQ 1— 1

-U-H-H-H MM MM

rH CO

CO 00 © 00

"r^ to
00 »o

CM tO i-h O CO

rH 1-4 H ^ O

CD 00 00 OS CMOS 00 to
OOOO 1-1O OO

-H -H-H-H-H -H-H-H-H -H -H -H -H -H -H -H -H

iO tO CD tO CM
CM

CM CO t^- CO
CD to to CO

CO CM

to o
to to

©i-l rH -CM

i-H CO CO CO ©

CM 1—1 CM

00 l> rH to
1— I

to CM 1> CO

rH -CM

00 CM 00 "tf

CM CM rH rH

I> to CM
CM CM CM rH

CM © O

00 to

rH rH CM ^

rH rH CM CM

H CO

1 1 1 1

rH tO CO

1 1 1 1

4< rH

"tf CO

-4< t|h to to

1 1 1 1

CM CM Tf

1 1
rH rH

co co

CM CM CM CM

1 1 1 1

to to «o to

1 1
rH t— 1

tO to

1 1 1 1

T^H T^l T^l ^

1> !>
1 1 1 1

1 1

CM CM

1 1 1 1

TjH Tt^ ^t1

CD ©

1 1 1 1

Tj"1 ^ T^H

1 1 1 1

CM CM

CO CO CO CO
1 1 1 1

1 1 1 1
r-t r-t i—t r-i
till

CM CM
1 1

1 1 1 1
1 1 1 1

1 1
t 1

^ 5
Ph a

34 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

hcONhh (NOONN OOClO O tO hh

lOOiCNCO CO CM -HH CO "^COCO^O to HH HH CO

-H -H -H -H -H -H-H-H-H -H -H -H -H -H -H -H

cd-^cocon cm i— i hh oo r^- co co oo co t>- h i

Tt* H CO O O HCOC5C5 CM CO CM HH Cll> CO H

00 O 00 CM O to hh CO to CO OS CO t> 1> O CO to

O O O h o

o o o o

to to 00
oooo

CO O

CO "HH

o o

-H-H-H-H-H

-H-H-H-H

-H-H

t-h O O CO CM
CM hh CM CO hh

l-H 1— 1 1— 1 1— 1 1— 1

IO CO CO CO
CO tO 00 t>

hh CM CO tO

I> i-lt> OS
i-H

CO 00
O 00

00 CO

O CM O hh CO

l— 1 i— 1 rH l— 1 O

CO tO 00. CO

oooo

NONO

O H O I— 1

C5 00

o o

i> co
o o

•H -H-H-H-H

-H-H-H-H

-H-H

■H-H

CO 00 hh 00 H

I> to hh CM

00 CO CM

H O

Oi H

HH CO HH CO HH

CM CO

cm oo o o to

CO to tO CO

00 i-l O O
tO CO HH IO

Tj<OSO
to to tO TfH

CO CM
to to

o o

tO CO

i— I i— I i— I t— I HH

CO O "HH iO CM

CM H to CM O

CM i-H i-H H CO

C3 10C0005 CO

HHINHN

CM to Oi CO Oi 00

HH ,-H tH CM

H • O CM I> H t-H i-H i— I

CO to H to CO CO

CO 00 IO HH GO H CO t>- O
H i-H i-H i-H H CM CM CM CO

HOlOOM to O 'HH "HH to H HH CO 00 CM to CO
h CM CMCOCMCO CO CM CO h ,-h i-h i-h

CM h CM tO CM O O CO i— I to to CM hh t-i CO CM

CO i-H i-H l— t

M Si

3 a

Ph C

O o

Ah1 >^

HtHthH

CO CM h i— t

H i-h CM CM

H CO

iiii

"HH t-H tO CO
IIII

iiii

CO h !>. to

i i

■HH i-H

HH CO

"HH HH IO IO
IIII

CM CM hh t}h

1 1

O CO

CM CM CM CM
iiii

iiii

to to to to

1 1
l-H T-H

to to

H^ ^rJH ^H tH
llll

1> l> 1> J>

1 1

CM CM

1 1

^H 'HH ^H ^H
IIII

IIII

HH -HH t)H tH

i i

O OS

1 1

CO co

tHH "HH ^jH "HH

IIII

l-H i-H i-H i-H

CM CM

1 1

HH "HH

CO CO CO CO

IIII

IIII
l-H T-H l-H l-H
IIII

CM CM
i i

1 1

CM CM
i 1

IIII

co co co co
iiii

IIII

iiii

i i
I>l>
i i

CD OS
1 i

RESULTS OF INBREEDING. 35

i-i ^ co co ©

i-l O i-H ©

CO rfi to to

tO © l>.

GO ©

-H -H-H -H-H -H-H-H-H

!N M N (© M

co qo co ©

00 rt< CM ©

-H-H -H-H

00 © 00

-H-H
© ©
^ ©

-H-H

i-i oo

CM ©

>o t^- r>- oo

o eo r- i>

,-H t-H Tt<

1-H ©

T-H © © ©

T-H l— 1 l— 1 i— t

1— 1 i-H

-H -H -fi -H -H

^-H^-H

-H -H -H -H

-H-H

Tfl © © © CM

rt< CM CO ©

CO TJH

00 CM © 00

^ TJH

"tf ©

(NMMMM

hNhh

T-H i-H

i-H

CO i—i CO (N

to to

co oo

M M (M (N h

i— i O i— i i— •

t-h CM i-h CM

i— i i—i

T-H ©

-H -H-H -H-H

-H-H -H-H

■H-H -H-H

■H-H

-H-H

(M ^ CM ^

00 © 00 i-H

00 00 00 00 00

to »o

to to

1— 1 1— 1 1— 1 1— I 1— I

i— I l-H i— I i— 1

T-H T-H

]> T-H © C5 I-

CM © CM ©

CO CM

to ©
tO to

CM TtH

tO CO 1> © i-H

T*00 1>CO<M

T-H T-H T-H tO

T-H T-H l-H tO

(M i-H © to

(N © T-H T-H

CM t*— CO

CM i-H

1-h co co

to ©

CM CM © ©

HH CM tH

* 2
- a

, cu

•Sb-S g-g

O O q> c3

T-H l-H T-H i-H

CO CM t-h t-h

i-H t-h CM CM

T-H CO

i 1 1 l

Tf T-H tO CO

i i 1 i

CO t-h 1> to

Tt^ T-H

Tt< CO

4< ^ to to

CM CM 4< "tf

1 1

i i

1 1 1 1

CM CM CM CM

i i i i

to to to to

t 1

1 1

to to

i i i i

1 1 1 1

l> 1>

I I r i

CM CM

1 1

CO ©

■ i i i

1 1 1 1

CM CM

1 1

CO CO CO CO
1 1 1 1

1 1 1 1

1-H T-H 1-H T—t

1 1 1 1

CM CM

1 1

CM CM
i i

1 1 1 1
1 1 1 1

i i

36 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

length of ear is even more than it seems from this table because
the variety contained plants which produced two ears of which
the second is usually smaller than the first; whereas the inbred
strains almost never produce more than one ear to a plant.

The number of nodes per plant is reduced but as compared to
height and length of ear this reduction is very much less. In
the number of rows of grain on the cob there is a reduction in
some lines and an increase in others. These tables show in the
clearest manner that inbreeding has a greater effect on some
characters than on others, and that segregation of characters
has occurred. Perhaps the most noticeable effect of inbreeding
as shown by these tables is the reduction in variability as brought
out by the range and statistical constants. This reduction in
variability is most apparent in the characters which are the least
reduced by inbreeding — number of nodes and number of rows
of grain on the ear — although the low variability is also apparent
in height and length of ear. In variability, also, there is a difference
between different lines.

The variability in height and length of ear of the inbred strains
is higher than it should be, owing to the fact that it was difficult
to obtain a perfect stand of plants, on account of poor germina-
tion of the seeds of the inbred strains. The aim was to have
three plants in a hill. From four to eight seeds were planted as
far as a limited supply of seed would permit, and later, thinned
to three plants. In spite of this precaution it was extremely
difficult to get ami:hing like a perfect stand, so missing plants
were replanted as soon as possible. These replants, owing to
their late start, never entirely caught up with the other plants
and are shorter in height and have smaller ears in consequence.
It is unfortunate that this practice was followed because it is
believed that much more reliable results would have been obtained
otherwise. On the other hand missing plants introduce another
source of error — that of unequal opportunity to grow. Because
there was abundant seed of the variety, and it germinated well,
practically complete stands of these plants were obtained.

The reduction in variability is more apparent in the details
of the structure of the plants and ears which cannot be expressed
statistically. The beautiful uniformity of these plants in all
characteristics at the present time is one of their most striking
features. This can be seen fairly well in the accompanying
photographs. (Plates I to V).

RESULTS OF INBREEDING.

In view of this fact of great uniformity and constancy as a
result of inbreeding one is astonished at the statement made
recently by Stout ('16) in a discussion of the results obtained
from inbreeding in maize by East and Hayes. Stout says (pp.
420-421) :

"strains similar in homozygosity show widest variation indicative of
spontaneous variation in natural vigor which is suggested that in such
highly cultivated varieties such as corn extreme sporadic variations may
be constantly occurring, a condition which is well sho'^n by the numerous
and well-known results of the ear to row test."

Several curious misconceptions are to be noted in this statement.
In the first place, it has never been maintained by anyone to my
knowledge that an equal number of generations of inbreeding produce
an equal amount of homozgosity in different lines. Secondly, it has
never been proposed that the degree of heterozygosity determined
the form or structure of any organism, but that such a condition
was accompanied by a stimulus to development which merely
increased the expression of many hereditary factors. This stimulus
is considered to be without any great effect in itself on variability.
Granted that the inbred strains were equal in homozygosity at
that time, that was no reason why they should be similar in
vigor or in any other respect — in fact the expectation is exactly
the reverse of this. With regard to "spontaneous" and "sporadic"
variation these inbred strains show unmistakably that there is
practically no sporadic or spontaneous variation, that the indi-
viduals making up an inbred strain are remarkably constant and
uniform after some degree of homozygosity is obtained and that
the diversity between different lines can be perfectly accounted
for on the basis of segregation of characters. Also, in the following
paragraphs in his paper Stout fails to see the distinction between
crosses of diverse inbred lines and between crosses of non-inbred
commercial varieties. Because Collins ('14) and Hayes ('14)
failed to obtain increases in all crosses between commercial vari-
eties of similar type Stout would question whether crossing in
maize was ever beneficial. It is quite to be expected that there
are many varieties already so widely crossed that further crossing
does not result in greater heterozygosity, but may even reduce it.
It is only in crosses between somewhat different varieties, like
flint and dent (Jones and Hayes '17) or between varieties from

38 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

different geographical regions (Collins '10) that any great amount
of heterosis in naturally widely crossed varieties is to be expected.

Although there has been a striking reduction in size of plant,
general vegetative vigor and productiveness in these inbred
strains of maize, and in comparison with non-inbred varieties the
inbred plants are more difficult to grow, emphasis must be put on
the fact that the plants are normal and healthy. The monstrosities
which are common in every field of maize, such as the occurrence
of seeds in the tassels, anthers in the ears, dwarf plants, completely
sterile plants, mosaic and albino plants and other similar anomalies
never appear in these inbred strains. Furthermore, in the details
of the size, shape, structure and position of the tassels, leaves,
stalks and ears, these inbred strains show the most striking uni-
formity. These minor details which characterize each of these
groups of plants are difficult to describe but are perhaps the most
noticeable feature about them. The stalks, the tassels or the ears
of all of these four Learning strains if mixed together could be
separated without the slightest difficulty by anyone familiar with
them. Some of the differences which characterize the ears of
these four strains are shown in Plate lb. It is to be noticed in
this photograph that Nos. 1-7-1-2 and 1-7-1-1, which were
originally from the same line, both have flat cobs. In one of them,
however, it is colored, in the other uncolored. Other differences
are to be seen in shape and color of seeds.

The segregation of row number accompanied by a reduction in
variability in these two strains is shown in Table 8 and Fig II.
Data previous to the third generation are not available but since
then a noticeable change in average row number has taken place
without any selection one way or the other. The variability of
each line has decreased at the same time. Whether the increase
in variability, after the eighth generation, has any significance
is not known. It is possibly due to the fact that both lines have
become irregular in row number so that the correct determination
of the row number has been rendered more difficult in the later
generations. Also the number of plants grown in the generations
from the 7th to the 10th are much too few to base accurate con-
clusions upon. The sharp increase in average row number and
decrease in variability in the 8th generation are probably due to
the unusually favorable growing conditions of that year.

RESULTS OF INBREEDING.

00 CD

O 1>-

1— 1

1> 00

00 ^

CM O0

T* lO

»o

-H

-H -H

■H-H

-H-H

•H-H

-H-H

■H-H

<N CO

O CO

<hh o

£2

CO CM

O CO

CO rH

O CO

rH CO

CM rH

©oo

00 t>

CO 00

oo o

O O

i—i

rH rH

CM »0

lO CO

tH 00

o o

CO rH

CM CM

CM rH

CO rH

rH rH

<N rH

CM rH

rH rH

-H

■H-H

■U-H

■H-H

-H-H

■H-H

CM iO

O CO

CM 00

t> rH

rH 05

CM Ci

O 00

O tO

rH I>

O CD

O iO

CM CM

CM rH

rH rH

CM rH

rH Oi

oo »o

O 00

CM CO

rH OS

CD CO

kft iO

co oo

CO CO

rH rH

CO tH

CM CM

CO tH

O OJ

CD to tH

OS O CD rH

1> TH O rH
rH rH rH CM

rH 00 ^Oi
rH CM

iO CM CO tH

CO CM rH O rH ^

rH CM
rH

CO rH
rH

O 00

rH CM

CO CM
tH

O CM
CM CM

CO »o
rH

tH OS

rH CM rH CM

1> CM

tJ< o
i i

rH CM
rH CM

1> CM

rH CM
rH CM

rh
»o

t^. CM

rH CM
I I
rH CM

IO TH
I I

TjH rH
I I

IO rH

l> CM
i i

r« O

rHCM
I I
rH CM

CM CO
i i

I I
TH rH

I I
IO rH

|> CM

^ O
I I

rH CM
I I

rH CM

<^co

^CM

rH CM
■ I
rH CM

a) fl ©

COCO lOtO COCO l>b-

O O O O i-ir-\

c3 is

0) o

oooo oo ^-(th

O O rH rH rH rH

0 05 o o o o

CM CM

rH rH

O O

COCO 10»0 COCO

HH t—( i—t y—tr-\

OO oo oo oo

rH rH

o o

40 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

345 6789 10 11

Generations Inbred

Figure II. The reduction in variability and segregation of number of
rows of grain on the ear in selfed strains of maize.

East and Hayes ('12) have noted many characters which are
isolated from maize by inbreeding. In addition to these, several
other characters have been isolated in this and in other material.
One of these characters is a constant difference in shade of color
of the foliage — some are dark green, others are light, yellowish
green. Some strains are lacking in root development and never
stand upright throughout the season. Some have a single-stalked
unbranched tassel, while others are profusely branched. Some
strains have peculiarly wrinkled or wavy leaves, particularly
noticeable in the first leaves. Some strains produce a small pro-
portion of connate seeds similar to those observed by Kempon
('13) in nearly every ear, while their occurrence has never been
observed on other inbred lines derived from the same source.
There are also marked differences in susceptibility to disease as
will be shown later.

These illustrations are sufficient to demonstrate beyond doubt
that by far the greatest amount of the fluctuating variability
found among ordinary cross-fertilized plants is due to the segrega-
tion and re-combination of definite and constant hereditary
factors. Many of these characters are seldom seen in continually
cross-pollinated plants, and never are so many combined together.

RESULTS OF INBREEDING.

This is due to the fact that they are recessive in nature and com-
plex in mode of inheritance. The most significant feature about
the characters which make their appearance in inbred strains is
that none of them can be directly attributed to a loss of a physio-
logical stimulation, although undoubtedly many of them may be
modified by the vigor of the plant upon which they are borne.
There is no one specific character common to all inbred strains
but simply a general loss of vigor, a general loss of size and of
productiveness accompanied by the appearance of specific char-
acters more or less unfavorable to the plants' best development
but these unfavorable characters are never all found in one inbred
strain, nor is any one character common to all inbred strains.

Probably the most common result of inbreeding in maize is a
reduction in the amount of pollen produced. This becomes appar-
ent in a smaller size of all parts of the tassel, in shrunken and
abortive anthers which are often never released by the glumes,
with a consequent reduction in the amount of pollen available for
fertilization. A normal corn plant should produce, on the average,
anywhere from lcc. to lOcc. or, in some cases, very much more
pollen. I have made no actual measurements of the amounts pro-
duced. Many of the inbred strains, however, now produce only
a small fraction of a cubic centimeter of pollen, and the production
of this small amount is much affected by weather conditions, so
that many strains, otherwise well developed and productive, are
maintained with the utmost difficulty.

It has been my experience that self-sterility in corn is due to
ovule or pollen abortion. Whenever pollen is obtained it seems
to be able to function. Failures to obtain seed after pollen is ap-
plied are common, but are usually attributed to external factors.
At least I know of no clear case where pollen is produced in which
it fails to fertilize the ovules of plants which were capable of being
fertilized by other pollen. Many cases of complete abortion of
the pistillate part of the plant must occur, as many plants are
lost through failure to set seed when good pollen has been applied.
Just where the trouble lies is not always possible to detect. Un-
doubtedly, many cases of complete abortion of either staminate
or pistillate functions, or both, occur during inbreeding, and the
plants are eliminated for that reason.

Reduction in the amount of pollen produced is less serious than
a reduction in the number of ovules, as a very small amount of

42 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

pollen suffices for fertilization when conditions are right. For that
reason unconscious selection for good ovule production has been
much more rigid than for pollen production. That is the reason,
I believe, that more inbred strains now show a greater reduction
in the staminate function than in the pistillate.

A significant feature of the effect of inbreeding upon sterility is
that some inbred strains are perfectly normal in their production
of pollen, and the amount of pollen produced is only a little less
than non-inbred plants, owing to the reduced vigor and size of
the plants which produce the tassels. Out of about twenty-five
inbred strains carried through at least seven generations, three of
them are perfectly normal in the structure and function of their
staminate parts. One of the Learning strains (No. 1-9) produces
more pollen than many non-inbred varieties growing nearby. In
every case, however, those plants which produce the best devel-
oped ears are the poorest producers of pollen, and those strains
which produce abundant pollen have ears which are poorly de-
veloped. In other words, inbreeding is bringing about a tendency
for maize to change from a functionally monoecious plant to a
functionally dioecious plant although, morphologically, both
staminate and pistillate parts are still present. This is illustrated
in Plates VI, a and b, where tassels and ears of four of the inbred
strains are shown.

Although no systematic selection has been practiced throughout
the inbreeding experiment a great deal of selection upon many
characters has been unavoidable as it is unavoidable in any in-
breeding experiment. In maize, the difficulties of hand pollina-
tion result in the selection of plants whose staminate and pistillate
parts are matured synchronously. Any great differences in this
respect, particularly towards proterandry, would render self-
fertilization difficult or impossible, as pollen, according to Andro-
nescu ('15) has very short viability, which fact my own experience
confirms. Of course, all plants which are weak, sterile, diseased
or in any way abnormal tend to become eliminated wherever
these causes reduce the chance of obtaining seed. This uncon-
scious selection becomes more rigid as reduction in vigor and pro-
ductiveness increases in the later generations of inbreeding. The
small amount of seed produced by hand pollination, under the
most favorable circumstances necessitates the using of the best

RESULTS OF INBREEDING.

ears obtained for planting in order to have enough plants upon
which to make any fair observations.

In every case inbreeding in maize has so far resulted in a reduc-
tion in size, vigor and productiveness. Some thirty or forty
inbred strains have been observed, many 6f which are additional
to the ones reported previously.

From the preceding statements in regard to the effect of in-
breeding it can be said that this process produces types which
differ in their power of development as follows:

1. Plants which cannot be perpetuated.

2. Plants which fail to complete normal development and can
be propagated only with the greatest difficulty.

3. Plants which are perfectly normal but varying in the amount
of growth they attain at maturity.

These normal inbred plants, so far obtained in maize, are not
as a rule as large, vigorous or productive as the original cross-
fertilized plants. It is theoretically possible to obtain such plants,
which cannot be reduced in vigor in a homozygous condition as
will be explained later. There is some evidence from the experi-
ments of Darwin, that such plants have been obtained by in-
breeding in other material, for example, in Ipomea and Mimulus.
Selection will help to obtain these vigorous, unreduceable indi-
viduals but may not be fully effective in doing so. More or less
unconscious selection is unavoidable in any inbreeding experiment.

These homozygous, normal, inbred strains, after the reduction
in growth has ceased, are quite comparable to plants of a naturally
self-fertilized species. Darwin found that self-pollination caused
no reduction in vigor in Nicotiana, Pisum, Lathyrus, Phaseolus
and other genera which are naturally self -fertilized to a large
extent. Hayes and Jones ('17) have found similar results with
the tomato. The only effect that inbreeding may have on such
plants is merely to isolate pure lines, which are quite uniform
among themselves, but may be diverse from one another, as
shown by soy beans (Jones and Hayes ('17), but which show no
reduction in vigor on continued artificial inbreeding. These
results are perfectly in accord with Johannsen's genotype con-
ception.

44 connecticut experiment station bulletin 207.

The Approach to Complete Homozygosity.

It now remains to be seen whether or not these inbred strains
are reaching the limit of reduction. There are two ways of de-
termining this, one is by growing two successive inbred generations
side by side in the same year, the other is by crossing different
plants within the same inbred strain.

In Table 9 the results from two successive generations grown
side by side in the same year are compared. On the whole, an
additional year of inbreeding after the sixth produces very little
change. In Table 10 are given the height, yield and length of ear
of selfed and sib-crossed plants which were grown in 1917. In
1916, in each of the strains of which figures are given in the table,
some plants were selfed and some were crossed by another plant
within the same strain. Since all the plants grown that year in
any one strain came from one individual of the preceding genera-
tion, that generation is the significant one. In other words if
the plant in that generation was homozygous, no increase of the
sib-crossed plants over the selfed plants would be expected. The
figures show that there is, on the whole, a slight increase in all
the characters studied. The increase, however, is no greater in
the cases where the common ancestor was inbred for seven genera-
tions than in the cases where it was inbred nine generations.

Shull ('11) compared sib-crosses with selfed plants in which
the significant generation, as I understand it, was the fourth, and
found that the crossed plants slightly excelled the selfed plants in
height, number of rows on the ear and yield of grain. Similarly
the Fi X Sibs exceeded Fi X self in yield, showing that in the
fourth generation complete homozygosis had not been attained.

Whether or not complete homozygosis has been attained by
some or all of the strains shown in Table 10 cannot be stated
positively from the data given. In most cases the increase of the
sib-crosses over the selfs is slight and probably of no significance
as there are about an equal number of cases in which the reverse
condition is shown. A few of the sib-crosses are, however, con-
siderably greater than the selfs in all three characters and it may
very well be that these strains have not attained the degree of
homozygosity that the other strains have. More data are needed
to establish this point with certainty as environmental factors
which favored a certain plot in one character would also favor
the other character as well.

THE APPROACH TO COMPLETE HOMOZYGOSITY.

O ' CO

tc " —

C I CP

CD ' t>

151

-H -H -H -H -H -H -H

N N w m N H

+ I I f I I +

-H -H -H 41 -H -H -H

41 41 41 41 41 41 41

Tl* H Tj< N ^ (N ^

lO !D CO

41 41 41 41 41 41 41

O CO CO CO <N 00 IQ

H i-i N CN rjl

4+| | 4 | 4

41 41 41 41 -H 41 -H

« M C IN N 00 N

41 41 41 41 41 41 41

M O OS 00 « tO (N

lO N M H

N H © Tt<

+ + 1 T

Oi 00 CO lO

CO b-
CO OS

• Tf< 1-H lO CO

] CO H N ^ lO

n ii h ii ii ei n

lO N N N N ^ ^

O (O O) ^ * rf

O) M -H H Tf

00 rH rH '

lO O! N N N

CO CO

CONNECTICUT EXPERIMENT STATION BULLETIN 207.

r J —

-H-H-H-H-H-H-H-H-H-H-H-H-H-H-H

+ I I + I + + I + I + I + I +

to r~- _

X -

I 4-+ I I I +

a h

- *

z -

c 2

d 5 a

- t -

z. >

= — a

— — — n — — x — — — — L-t-

n — d — — ri — r-: m m c - :i —

-H-H-H-H-H-H-H-H-H-H-H-H-H-H-H

— m — — n — ~ — ti ~ l- ri n —

R x

- B

c c c c n x

- C C C C N C

x x

2
-

<
-

— ?i ?J M :C

- =

ti. x
" —

m rt m m r: m — — —

i i i i i i i i i

i|- -r C C ~ r- 4 ?•! CI

— — — — — — — — — —

« oi ei « rp « w ci «

heterozygosis and vegetative luxuriance. 47

The Effect of Heterozygosis on Vegetative Luxuriance.

The most noticeable manifestation of heterosis in plants is a
general increase in vegetative luxuriance. In maize this is par-
ticularly noticeable in increased height of plant, diameter of stalk,
'root development, length of ear and productiveness of grain (see
Plates III, V, VII, VIII, IX, X and XII). In crosses between
inbred strains of maize the amount of heterosis shown is inversely
proportional to the degree of relationship as shown in Table 11."
Montgomery ('12) has obtained similar results.

Some characters are much more affected by heterozygosis than
others. In comparing Tables 12, 13 and 14 with Tables 15 and 16
it will be noticed that the yield of the crosses is increased 180
per cent., height is increased 27 per cent, and length of ear 29
per cent, over the average of their parental lines. On the other
hand, the number of nodes per plant and number of rows of grain
on the ear is increased only 6 and 5 per cent, respectively. In
other words, heterozygosis does not increase the number of parts
to anything like the extent that it increases the size of those parts.
Those parts of the plants which are more or less indeterminate in
size, like internodes, ears and seeds are augmented by crossing as
the result of an increase in the rapidity and rate of cell division.
The increase in size of parts is probably brought about by an
increase in size of cells as well as an enormous increase in number
of cells. Tupper and Bartlett ('16) have shown that gigas mutants
in Oenothera have larger cells than the non-mutant type, so that
a change in cell size may accompany a germinal change.

From Table 11 it will also be seen that some first generation
hybrids may even surpass the original variety in yield, height or
length of ear, although the comparison is rather unfair as the
Learning variety was not acclimatized as were the inbred strains.
The return of vigor realized in the first generation crosses is often
enormous, and the same is true of crossing inbred strains derived
from totally different types of maize as is shown in Table 17.

Although there is an immediate and striking return to the
vigorous condition of the non-inbred stock there is not a return
in variability as shown in Tables 18, 19, 20, 21 and 22. The first
generation crosses are no more variable than the inbred strains
by which they are produced, in many cases less variable, and show
striking differences when compared to the original stock. The
coefficient of variability is entirely inadequate in bringing out

CONNECTICUT EXPERIMENT STATION BULLETIN 207.

5 © « c

3 > C c;

E s S '?.

= -

I + I +

x t - - x — cc
cm ci —

+ + + + + +

-H -H -H -H
— o — cr.

C: ~. l- — CM CT. X CT; CO CT.

o — — — o — o — o" o — — o

-H-H-H

-H-H-H-H

•-i t» l> l>

-H-H-H

r~ re <e

-H-H -H -H

CC — IQ lo IG — ,~ — cc if; — cc vrc

> w

11+ I ++

PC CN CN

+ + + + + +

CM -r- re x N D -
— - CO — — —

— — IO — I- S§ —

-H-H -H-H fl-M -ti-H-H-H -H-H-H -H-H-H -H

Z ?i rc — ei O GC . — — O ~ Lt ut re

NOMC 30 oa --c
r. rr. x x x t- x

X X X
O 03 X

cm co co — oa o
C: O O CN c: —

B 2 S

CC^ CO CO X

0>COO*C CM — 03
DO CO cc — X CN

l> O IO QC

1— I t- >e 1~

— CM — i O
O X N X

— g

4-4)

0 1 1
■Z-l )

•1-4)

! «o^< —
i i i

i- — CM

^ co lo
i i i

— CJ —
1 1 1

i ' f i

CM Lt —

' — i i i

CO CM CM LO

— LO IQ
l l l

— 1Q —
i i i

4 1>- cm

i i i i

CMt>N
i i i

CM N CM
i i i

i i i
- — ~.

iiii
t- — — —

1 i i

l ■ i

i i ■

— — CM

iiii

CM ~
i i i

CM TJ- 4

i l i

CO — CM

— X X —

i i i

CM CM 4"
i i i

CN ' ' '

i T t

N NN
i i i

re CM — '
III*

rj- 1> — X
ii-ii
— X

XXX

i T i
— CJ —

3 XXX

X^£~ n
i i i

re — <e —
iiii

XXX
tZ — —

1 1 1

— .

xxi
i i

i i i

CM — LO

1 1 1

>>
—

' —

i i i i

M C r- 1-

1 1 1

CM »e —

«-0 CM CM IO

JL 1 1

r i i

1 1 1 1

rr n CM n

1 i i

i i i i

CM CM
1 1 1

TTCJN
1 1 1

i i i i
— — T.

■ i i

i i i i

vr — — —
■ i i i

— cr. cr:
i i i

i | T

i i i i
CM —

i i i
— — CJ

i- CM CM
i i •

• i i 1
X — CM 05
i i i i

i i i
CO — CM
i i i

i i i i

CM CO CO *-H

i— CM CM
i i i

CO CM CN
i i i

cot> ti.ee
i i i i

l l l

CNN
1 1 1

IIII

CC C N
iiii

i i i

NNN
i i i

i i i

ONCfi
i i i

HETEROZYGOSIS AND VEGETATIVE LUXURIANCE. 49

Table 12. The effect of crossing inbred strains of maize
AS shown by the increase in the yield of grain.

Yield of bushels per acre

Pedigree number

Pedigree uu'inbcr

of strain — A

axb

BXA

of strain — B

1.6-1-3-4-4-4-2-4-4

99. 1

99.9

1_7_ i_ i_ i_4_7-5-4_7

1-6-1-3-4-4-4-2-4-4

112.9

1-7- i_i_ i_4-7_5_ 2-6

1-6-1-3-4-4-4-2-4-4

82.4

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1

101.0

1-7-1-2-2-9-2-1-1-1

1-6-1-3-4-4-4-2-5-5

•88. 1

84.4

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

103.2

106.7

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

91.0

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

94.8

1-7- 1_1_ 1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

63.9

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

71.5

• 1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

58.0

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

52.5

100.5

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

59.6

82.1

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

66.3

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

84.9

1-6-1-3-4-4-4-2-4-4

1-7*1-1-1-4-7-5-2-6

40.5

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

59.4

1-6-1-3-4-4-4-2-4-1

78.4

Increase

50.4

Percent increase. . . .

180.00

Table 13. The effect of crossing inbred strains of maize
as shown by the increase in the height of plant.

Pedigree number
of strain — A

Height of plant in inches

AXB

BXA

Pedigree number
of strain- — B

1-6-1
1-6-1
1-6-1
1-6-1
1-6-1
1-6-1
1-6-1
1-6-1
1-9-1
1-9-1
1-9-1
1-9-1
1-9-1
1-9-1
1-7-1
1-7-1
1-7-1

3-4-4-4-
3-4-4-4-
3-4-4-4-
3-4-4-4-
3-4-4-4-
3-4-4-4-
3-4-4-4-
3-4-4-4-
2-4-6-7-
2-4-6-7-
2-4-6-7-
•2-4-6-7-
2-4-6-7-
•2-4-6-7-
2-2-9-2-
1-1-4-7-
•1-1-4-7-

2-4-4

2-4-1

2-5-5

2-5-3

5-3

5-6

1-1-4

5-2-6

5-2-1

Average

Increase

Percent increase. .

97. 8±
97. 8±
97.8zfc
96.7=b
93. 6±
93.6=b
93. 6±
102. 7 ±
80. 3±
80. 3±
80. 3 ±
80. 3±
80. 3±
77. 0±
82. 6±
78.5±
82.2±

117.
117.
115.
121.
112.
116.
113.
116.
111.
110
109.
110
108.
Ill
114.
98.
105

3± .
6± .
4± .
9± .
9±1.
1± .
8± .
0±
1±
5± .
2± ,
9± ,
1±
1±
9±
7±

2-4-

117.2±.44

109. 9± .77
113. 4± .51

94. 0±1. 36
114. 1± .55
109. 5± .76

2±

1-7-1-1-1-4-7-5-4-7

5±

1-7-1-1-1-4-7-5-2-6

0 =

1-9-1-2-4-6-7-5-6

2±

1-7-1-2-2-9-2-1-1-1

2±

1-7-1-1-1-4-7-5-4-7

6±

1-7-1-2-2-9-2-1-1-4

0±

1-9-1-2-4-6-7-5-63*.

2±

1-7-1-1-1-4-7-5-2-1

102

7±

1-6-1-3-4-4-4-2-5-3

7±

1-6-1-3-4-4-4-2-4-1

2±

1-7-1-1-1-4-7-5-2-1

7±

1-7-1-1-1-4-7-5-4-5

6±

1-7-1-2-2-9-2-1-1-4

2±

1-7-1-2-2-9-2-1-1-1

8±

1-6-1-3-4-4-4-2-4-4

6±

1-6-1-3-4-4-4-2-5-5

1-6-1-3-4-4-4-2-4-1

88.0

112.4
24.2
27.44

88.3

50 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Table 14. The effect of crossing inbred strains of maize
as shown by the increase in the length of ear.

Length of ear in inches

Pedigree number

of strain — A

AXB

BXA

of strain — B

1-6-1-3-4-4-4-2-4-4

1±

7.3±.

0±

1-7-1-1-1-4-7-5-4-7

1-6- 1-3-4-4-4-2-4-4

1±

5±

3±

1-7-1-1-1-4-7-5-2-6

1-6-1-3-4-4-4-2-4-4

1±

8±

l=t

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1

2±

9±

3±

1-7-1-2-2-9-2-1-1-1

1 «_1 1 A A A 1 K_K
1-D-1-0-4-4-4-Z-0-0

5±

6.8±.

0±

1 - / - 1- 1-1-4- / -0-4- /

1-6-1-3-4-4-4-2-5-5

0±

2±

8.0±.

1±

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

0±

8±

1±

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

9±

6±

2±

1-7-1-1-1-4-7-5-2-1

1-9.1-2-4-6-7-5-3

9±

7-fc

9±

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

9±

6±

2±

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

9±

5±

5.5±.

11 ,

2±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

9±

5±

7.6±

09 *

2±

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

9±

1±

7.6±

1±

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

1±

3±

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

1±

6±

ldb

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

3±

5±

o±

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

2±

0±

2±

1-6-1-3-4-4-4-2-4-1

Average

Percent increase. . . .

Table 15. The effect of crossing inbred strains of maize
as shown by the increase in the number of nodes.

Number of nodes

Pedigree number

of strain— A

AXB

BXA

of strain — B

1-6-1-3-4-4-4-2-4-4

7±

6±

2±

2=fc

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-4-4

7±

3±

8±

1-7-1-1-1-4-7-5-2-6

1-6-1-3-4-4-4-2-4-4

7±

0±

9±

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1

5±

0±

1-7-1-2-2-9-2-1-1-1

1-6-1-3-4-4-4-2-5-5

4±

9=b

2±

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

4±

5±

1±

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

4±

9±

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

2±

3db

6±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

1±

8±

2±

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

1±

9±

5±

1-6-1-3-4-4-4-2-4-3

1-9-1-2-4-6-7-5-3

1±

2±

5±

6±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

1±

3±

0±

7±

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

Irfc

3±

7±

1±

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

9±

7±

o±

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

1±

6±

7d=

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

8±

3±

4±

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

6±

•

5±

1-6-1-3-4-4-4-2-4-1

Increase

Percent increase. . . .

HETEROZYGOSIS AND VEGETATIVE LUXURIANCE. 51

Table 16. The effect of crossing inbred strains of maize as shown
by the increase in the number of rows of grain on the ear.

Number of rows of grain on the ear

x GQigrGG number

Pedigree number

of strain — A

AXB

B X A '

of strain — B

A O 1. o x x a _» x i

Q -u

y ±

5±

8±

171 1 1 A T B A n

1_fi_1-9_4_4_4-9_4_4

5±

8±

1 A
lO

171 1 147^9^

1^13444944

Q _u

.2±

5±

1 Q

1Q19/<«7rA

1-6-1-3-4-4-4-2-4-1

7±

4±

9db

1-7-1-2-2-9-2-1-1-1

1-6-1-3-4-4-4-2-5-5

1±

4±

3db

8±

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

1 -±-

9±

2±

9±

l-7_l_2-2-9-2-l-l-4

1 1 _ 3-4-4-4- 9- t

i i

■l ±

0±

5±

10 1 9 4 (\ 7 r>-('i

1-6-1-3-4-4-4-2-5-3

3±

•19

4±

0±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

4±

7±

3±

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

4±

7±

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

4±

9±

7±

o±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

4±

8±

0±

1±

1_7_ 1-1_ I-4-7-5-4-5

1-9-1-2-4-6-7-5-3

4db

8±

2±.

9±

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

5±

2±

9±

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

9±

3±

9±

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

8±

6±

1±

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

0±

8±

7±

1-6-1-3-4-4-4-2-4-1

Increase

9 '

Percent increase. . . .

Table 17. The effect of crossing inbred strains derived
from different types of maize.

Yield

Increase

Length

Increase

Type

Pedigree number

bu.

above

Height

above

of ear

above

per

ave. of

inches

ave. of

inches

ave. of

acre

parents

Dent

1-6-1-3-4-4-4-2-4-4-3

Floury

10-3-7-3-9-7-5-4-3

Flint

29-5-2-3-8

Pop

65-8-2-2-6-5-2-4

Dent X Floury. .

(1-6-1-3) X (10-3-7-3). .

4-67.7

122

+36.1

+3.0

Dent X Flint. . .

(1-6-1-3) X (29-5-2-3). .

+72.5

117

+24.2

+2.9J

Floury X Dent. .

(10-3-7-3) X (1-6-1-3). .

+20.6

108

+21.5

+1.3

Floury X Flint. .

(10-3-7-3) X (29-11-4-4)

+51.3

104

+22.4

+3.7

Flint X Dent . . .

(29-5-2-3) X (1-6-1-3). .

+58.6

115

+22.4

+3.6

Flint X Floury. .

(29-5-2-3) X (10-3-7-3).

+63.2

112

+30.8

+4.0

Pop X Dent

(65-8-2-2) X (1-6-1-3). .

+41.4

+ 12.8

+ 1.3

Pop X Flint ....

(65-8-2-2) X (5-8-6-3) . .

CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Table 18. The effect of crossing upon variability as shown
by the height of plant.

Coefficient of variability of height

Pedigree number

of strain — A

A X B

B X A

of strain — B

1 -fi_ 1 _Q_4_ 1 _4_9_i_l

X \J -L • J T — ± I _ I — I

l-± =

10 =

4.01 =

49 =

17 11 1 17 5-17

l-l-l-l- 1 — 1 1 -O— _- i

1-6-1-3-4-4-4-2-4-4

14 =

74 =

75 =

1-7-1-1-1-4-7-5-2-6

1-6-1-3-4-4-4-2-4-4

14 ±

DO =

73 =

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1

40 =

22 =

95 =

1-7- 1-2-2-9-2- 1-1-1

1-6-1-3-4-4-4-2-5-5

14 =

92 =

7.32 =

49 =

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

14±

09 =

5.20 =

05 =

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

14 ±

00 =

73 =

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

11 =

05 =

64 =

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

04 =

66 =

11 =

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

04 =

15 =

40 =

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

04 =

78 =

12.61=1

64 =

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

04 =

18 =

5.48 =

27 =

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

04 =

27 =

7. 94 = . 49

05 =

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

73 =

26 =

95 =

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

05 =

27 =

14 =

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

75 =

81 =

14 =

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

64 =

41 =

40 =

1-6-1-3-4-4-4-2-4-1

Average

j7.11

Table 19. The effect of crossing upon variability as shown
by the length of ear.

Coefficient of variability of length of ear

Pedigree number

of strain — A

A X B

B X A

of strain — B

1-6-1-3-4-4-4-2-4-4

17 = 1

86 = 1

99 ±1

75 ±1

1-7- 1_1- 1-4-7-5^-7

1-6-1-3-4-4-4-2-4-4

17 = 1

33 = 1

58 =

1-7-1-1-1-4-7-5-2-6

1-6-1-3-4-4-4-2-4-4

17 = 1

92 =

77 =

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1

06 =

94 =

04=1

1-7-1-2-2-9-2-1-1-1

1-6-1-3-4-4-4-2-5-5

17 = 1

73 = 1

88 = 1

75 = 1

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

17 = 1

71 =

37 =

33=1

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

17 = 1

44 =

77 ±

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

23=1

55 =

00 = 1

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

32 =

16 =

23=1

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

32 =

00
H-

08 =

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

32 =

46 = 1

45 = 1

00 = 1

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

32 =

69 = 1

55 =

76 = 1

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

32 =

87=1

26=1

33 = 1

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

77 =

73 ± 1

04 = 1

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

33 = 1

87 = 1

17 = 1

1-6-1-3-4-4-4-2-4-4

l-7-l-l.l_4.7_5.2-6

58 =

91=1

17 = 1

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

00 = 1

83 = 1

06 =

1-6-1-3-4-4-4-2-4-1

HETEROZYGOSIS AND VEGETATIVE LUXURIANCE. 53

Table 20. The effect of crossing upon variability as shown
by the number of nodes.

Coefficient of variability of number of nodes

Pedigree number

of strain — A

AXB

BXA

of strain — B

1_6- 1-3-4-4-4-2-4-4

12 ±

88 ±

91±

23 ±

171 1 1 j( 7 s i 7

l-i-l-l- 1-4- /-0-4- /

1-6- 1-3-4-4-4-2-4-4

12 ±

68 ±

27 db

171 1 1_ A— T K O d

1-6-1-3-4-4-4-2-4-4

12±

50 zb

67 ±

1Q1 04A7^R

1-6-1-3-4-4-4-2-4-1

61±

00 ±

77 ±

17199Q9111

1-6-1-3-4-4-4-2-5-5

QA _i_

y-i ±

A.7
1 1

82 d=

09 ±

23 ±

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

94 dz

48 ±

34 ±

7.93±

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

94 ±

67 dz

67 ±

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

98 ±

19 d=

66 ±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

llzb

47 ±

98 db

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

lldb

66 ±

61±

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

53 d=

72 =b

66 ±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

11±

41=b

93 =b

48 ±

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

llzb

86 ±

20 dz

93 dz

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

67 ±

62 ±

77 ±

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

93 ±

15 ±

12 ±

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

27 ±

61=b

94 ±

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

66 ±

67 ±

61±

1-6-1-3-4-4-4-2-4-1

Average

•5

6.80

♦

Table 21. The effect of crossing upon variability as shown
by the number of rows of grain on the ear.

Coefficient of variability of number of rows

Pedigree number

of strain — A

AXB

BXA

of strain — B

1-6-1-3-4-4-4-2-4-4

83 ±

73 ±

68 =b

48 ±

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-4-4

83 ±

50 ±

43 ±

1-7-1-1-1-4-7-5-2-6

1-6-1-3-4-4-4-2-4-4

83 ±

79 d=

21d=

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-4-1 '

48 ±

96 ±

06 ±

1-7-1-2-2-9-2-1-1-1

1-6-1-3-4-4-4-2-5-5

23 ±

31±

24 ±

48 ±

1-7-1-1-1-4-7-5-4-7

1-6-1-3-4-4-4-2-5-5

23 ±

62 ±

11±

40 d=

1-7-1-2-2-9-2-1-1-4

1-6-1-3-4-4-4-2-5-5

23 ±

60±

21 d=

1-9-1-2-4-6-7-5-6

1-6-1-3-4-4-4-2-5-3

96±

92 ±

10 ±

1-7-1-1-1-4-7-5-2-1

1-9-1-2-4-6-7-5-3

08 ±

00 ±

96 d=

1-6-1-3-4-4-4-2-5-3

1-9-1-2-4-6-7-5-3

08 ±

36 ±

48 db

1-6-1-3-4-4-4-2-4-1

1-9-1-2-4-6-7-5-3

08 it

11±

32 dz

lOzb

1-7-1-1-1-4-7-5-2-1

1-9- ] -2-4-6-7-5-3

08 ±

92±

62 ±

94 =b

1-7-1-1-1-4-7-5-4-5

1-9-1-2-4-6-7-5-3

08 ±

35 ±

02 d=

40 d=

1-7-1-2-2-9-2-1-1-4

1-9-1-2-4-6-7-5-6

21±

42 ±

06 dz

1-7-1-2-2-9-2-1-1-1

1-7-1-2-2-9-2-1-1-4

40 ±

68 ±

83 d=

1-6-1-3-4-4-4-2-4-4

1-7-1-1-1-4-7-5-2-6

43 d=

75 db

23 d=

1-6-1-3-4-4-4-2-5-5

1-7-1-1-1-4-7-5-2-1

10±

84 ±

48 dz

1-6-1-3-4-4-4-2-4-1

8.58

N CO U5 M nNMK5MT)i^m'*M

-H 41 4! -H -H-H-H-H-H-H-H-H-H-H

-r 05 '~ O <n u> <-< x -<r o oo t>

r-HTfOO rt N ic 00 f N C C N N

Tf U! 00 K5

IN O O
lO » UJ (D S (O

41 -H -H -H 41

C5 00 N o n

41-H-H-H-H-H-H41-H-H

N O O • CO • • i-H

— 'ONOeNOUiCN

00 « « • • O M • ■ O M N N

« N r

Tj< t>- Tj< CO

r}< TJH IH CO

N IQ H 1(3

TJ< <N t-l

4 ci o

f H M *

ci >-i N N

O N N O

^HCOiOfOrfnCOHTji
lOHioaLlcIjioiONH

NNOOQtoO)HHH
I I I I I | lilt

xxxxxxxiii

MNNiljiOrtHioiOiO
eocoeorj<.^<^evic^<Neji

OtitONNNNOOCl

HETEROZYGOSIS AND VEGETATIVE LUXURIANCE. 55

the beautiful uniformity of these crosses between inbred strains.
In every respect each plant is a replica of the other. A collection
of such vigorous and uniform maize plants in the field is a novel
sight (see Plates 1Kb and Vb).

Shull ('14) has pointed out that vigorous plants may be less
susceptible to the effect of the environment than weaker types
and that first generation hybrids, between uniform strains, may
even show a reduction in variability.

The results obtained show this quite noticeably. Particularly
was this true of several Fi's grown between their parental strains
in a demonstration plot on rich low ground. During both seasons
(19 16-' 17) when they were grown on this piece of ground, the
weather was especially unfavorable when the plants were just
starting, the ground being saturated "with water most of the time.
The germination in the selfed lines was extremely poor and many
plants which did grow were stunted, and remained so throughout
the season and never attained full height nor did they produce
either tassels or ears. The variability of height, in these plants,
was far greater than in many non-inbred varieties. Several
plants, when killed by frost in the fall, were not over 30 inches tall
while the average height of this strain is from 80 to 85 inches.
The hybrids also had a poorer start than non-inbred varieties
grown on the same ground on account of the small seed, but were
able to overcome their handicap and in a few weeks were quite
uniform. At the end of the season the difference in variability
between the Fi on the one .hand and the inbred strains and the
varieties on the other was striking. These plants were not used
in the statistical work given here. The crosses and parents which
were used and which were apparently quite uniform show a slight
reduction in variability, in the number of nodes and in height in
the F^s as compared with their parents as can be seen in Tables
18 and 20. As Shull also pointed out, the variability of some
characters may be increased by heterosis. This is shown in number
of rows on the ear. The inbred strains rarely or never produce a
second ear. The vigorous hybrids almost always do, and as the
data have been obtained by counting all the ears gathered from a
plot, the variability of the crosses, as shown in Tables 19 and 21,
consequently seems greater than it really is as the second ear on
nearly every plant is smaller and contains a fewer number of rows.

56 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Although reciprocal crosses are on the whole nearly equal in
respect to the degree in which heterosis is shown, there is some evi-
dence, from Table 12, that this is not always so. Observations
from the crosses in the field showed clearly that those in which
strain Number 1-6 was used as the female, were usually more
vigorous and productive than the others. In Table 23 the yields
of all the crosses and reciprocal crosses (from 1 to 4 of each) having
the same parental races are averaged. An average of all those

Table 23. Yield of reciprocal crosses among inbred strains of

MAIZE.

(All crosses grown 1916. Yield given as bushels per acre.)

Average 9

Yield selfed:
(1917)
(1916)

Ave. weight
of seed-cg.

1-9-1-2

1-7-1-2

1-7-1-1

62.0

31.8
30.6

16.6

78.4

37.6
19.2

27.9

87.9

42.3
32.7

19.9

1-6-1-3

1-9-1-2

82.1

100.5

86.7

1-7-1-2

63.0

70.9

103.6

1-7-1-1

55.3

57.2

98.7

1-6-1-3

67.7

95.8

92.2

96.3

46.2
32.8

34.1

Average cf
89.8 •

79.2

70.4

85.2

crosses in which each strain was used as the male and in which
each was used as the female parent shows some striking results.
Those crosses on the whole in which strain Number 1-9 was used
as the female gave the lowest yield. Those crosses in which strain
Number 1-6 was used as the female are clearly the most pro-
ductive. Strain Number 1-6 is the one which has the largest
seeds and in which the pistillate inflorescence is the best developed
of the four strains and at the expense of the staminate inflorescence.

HETEROZYGOSIS AND VEGETATIVE LUXURIANCE. 57

Strain Number 1-9 is just the reverse of this. It is tfie best de-
veloped of all the inbred strains in its staminate inflorescence,
always producing abundant pollen, but has the smallest seeds,
and is one of the poorest in the development of its pistillate in-
florescence. Approximately a uniform stand of plants was ob-
tained in all these crossed plots. They were all grown side by side
in the same field in the same year. * There seems, therefore, to be
a marked correlation in the development of the pistillate infloresc-
ence between the mother and her hybrid progeny. The high yield
of the crosses in which Number 1-9 was used as the male is due
to the fact that its average yield was not pulled down by the low
yielding crosses in which it was used as a female. The crosses in
which 1-7-1-1 and 1-7-1-2 were used cannot be compared fairly
with the other two because these two strains are more closely
related. This correlation bears a close relationship with the size
and development of the seed which produces first generation
hybrid plant. The seeds of strain 1-9 are the poorest developed,
those of Number 1-6 are the best. Hence, the plants of crosses
(1-6) x (1-9) have a better start than the plants of the reciprocal
cross. This assumption is borne out by the fact that the second
generation starting from large fully developed seeds grown on
vigorous Fi plants are larger at the start than the Fi plants grown
from small, poorly developed seeds produced on inbred plants.
This is shown in Fig. Ill and Plate IX. The second generation,
however, is surpassed by the first before the end of the season, as
shown in Fig. Ill and Plate X. Somewhat similar results have
been obtained by Castle ('16) in guinea-pigs. F2 animals, out of
vigorous Fi females, are larger at the start than either parent
but do not surpass the Fi individuals as in this case. It will be
seen from this that in plants or animals which are reduced by
inbreeding, the Fi is handicapped in comparison with the F2 and
the immediately following generations.

It is not certain that the differences between reciprocal crosses
can be accounted for on a purely nutritional basis. There is the
possibility of unequal germinal reactions with different cytoplasms.

*The crossed strains were not grown between their inbred parental strains
as was the case in the yields reported in U. S. Dept. of Agric, B. P. I.
Bull. 243. This accounts in part for the extraordinarily large yields
obtained at that time.

CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Figure'III. Growth curves of two inbred strains of maize and their
first and second generation hybrids.

inbreeding in plant and animal improvement. 59

The Value of Inbreeding in Plant and Animal Improvement.

These inbreeding and crossbreeding experiments on corn have
considerable theoretical importance in the improvement of culti-
vated plants and domesticated animals. We have seen that in-
breeding results in the elimination of abnormal, pathological and
undesirable characters in general. This result has been obtained
with a loss of size, vigor and productiveness. When these inbred
strains are crossed, however, vigor and productiveness are re-
turned in increased amount due to the uniform excellence of the
individuals freed from undesirable characters. In this way a new
variety or breed can be synthesized from the purified inbred
strains of an old stock. A great sacrifice is thus made to attain a
great good. Of course such a variety would have to be fixed by
selection during a number of generations. The common practice
of crossing in animals and plants already extremely heterozygous
in order to obtain further improvement is like trying to solve a
picture puzzle in the dark. It is only by resolving a naturally
crossed species into homozygous types by inbreeding that it can
be best analyzed and its desirable characters most surely selected
for the recreation of an improved type.

The practical value of inbreeding has long been recognized by
the breeders of domesticated animals. To gain uniformity and
the highest expression of certain desirable characters they often
practice inbreeding until the vigor of the breed is frequently im-
paired. From the results obtained with maize it seems that they
stop just before the greatest good is to be accomplished. What if
vigor is lost? It can always be regained immediately by crossing.
There is no surer way of eliminating undesirable characters and dis-
covering the best that there is in a stock than by a process of rigid
inbreeding followed by subsequent testing in different crosses. This
is not offered as a practical plan of procedure for the improvement
of animals. It is merely intended to call attention to a principle
which has probably not been used to its fullest extent. It may be
that many domesticated breeds of animals cannot endure in-
breeding to the extent that maize can. The cost of obtaining such
pure types might very easily be prohibitive. The writer believes,
however, that the splitting up of a breed of animals or a naturally
crossed variety of plants by long continued inbreeding of the
closest kind possible followed by the recombination of the most
desirable inbred types, obtained in sufficient numbers to insure

60 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

that nothing of value present in the stock at the start, is lost, is a
valuable, practicable method of improvement in many animals and
plants. According to this method a variety or breed would be re-
created and then continued in a naturally crossed condition just as
it was before. The value of this procedure as a method for plant
improvement is now being tested. It is, of course, a long time
proposition and one that must be carried on extensively to promise
results.

With a few plants which are easily crossed it is possible to
utilize hybrid vigor to the fullest extent by growing only first
generation plants. Attention has been repeatedly called to this
method of increasing the productiveness, particularly of maize
and tomatoes. The greatest amount of hybrid vigor is shown in
maize when the plants have been previously inbred. Unfortu-
nately, when the inbreeding is carried on for several generations
the reduction in the vigor of the resulting plants is so great that
the small size and low vitality of the seeds borne on inbred plants
seriously handicaps the hybrid plants grown from these seeds as
just shown. So what is gained by an increased amount of heterosis
may be partly lost by the poor start which the plants have. This
handicap, in comparison with normally-crossed varieties, the Fi
may not be able to overcome entirely even though it is far more
uniform and free from barren, mal-formed and otherwise unde-
sirable plants — factors which count heavily in maximum pro-
duction.

A way to overcome this handicap suggests itself which is to
cross two vigorous first generation hybrids whose composition is
such that the resulting cross will not be less heterozygous than
either parent and, therefore, theoretically no less vigorous and
productive. This is easily accomplished by taking four distinct
inbred strains which are of such a composition that a cross between
any two of them gives a vigorous product. Now by crossing two
of these strains to make one first generation hybrid, and at the
same time crossing the other two to make another, and then by
combining the two first generation hybrids there should be no
reduction in heterozygosity. These doubly crossed plants, how-
ever, starting from large seeds produced on large, vigorous hybrid
plants would be freed from the handicap which their parents had
and although somewhat less uniform should be more productive.
While it may be out of place to say anything about this method

HETEROZYGOSIS AND SELECTIVE FERTILIZATION. 61

until it has been thoroughly tested it is a method which is more
promising than the plan originally advocated because by this
method crossed seed for general field planting is produced much
more abundantly than when non-vigorous inbred strains are
crossed.

The Effect of Heterozygosis upon Endosperm Develop-
ment and Selective Fertilization.

Together with the increase in size of other parts of the plant
there is also an appreciable increase in the size and weight of seeds
of maize immediately resulting from cross-pollination. This has
been shown clearly by Collins and Kempton ('13) by pollinating
several ears of maize with a mixture of the plant's own pollen and
that of a different variety. Roberts ('12), Carrier ('13) and Wolfe
('15) have also shown that in maize the endosperm is increased by
crossing. The writer ('18) has shown that this increase in endo-
sperm development appears even more strikingly in reciprocal
crosses between different inbred strains of maize. At that time
reciprocal crosses had not been obtained between different indi-
vidual plants. In Table 24 are given the distributions of the
weights of the seeds shown in Plate XIa. Two plants were pol-
linated with a mixture of pollen obtained from these same two
plants. One of the plants had white seeds and the other yellow
and the selfed and crossed seeds on each ear could be easily
distinguished. The same pollen mixture was also applied to a
third plant of an inbred strain different from either of the other
two but more nearly related to one than to the other. The
average difference in weights between the selfed and crossed seeds
on each ear are large. The two out-crossed lots of seeds on the
third ear do not differ as greatly but the heavier seeds resulted
from the wider cross.

Table 25 gives a number of averages of the weights of seeds
from similar pairs of ears each having selfed and reciprocally
crossed seeds. In every case there is a noticeable increase in
weight as the result of crossing. In Table 26 the weights of the
out-crossed seeds resulting from some of the same pollen mixtures
are given. Here again the heavier seeds are those which have
resulted from the wider cross. A and C are two inbred strains
derived from one variety at the start while B is derived from a

CONNECTICUT

EXPERIMENT STATION BULLETIN 207.

O 5 *

-H

-H -H

00 iO

lO 00 o

>o o o

O iO OS CO

^ O CO

»0 O f~

O CM o

i-i CM CM

g J3 M

Q £ 3

-< pa
x x
u o

HETEROZYGOSIS AND SELECTIVE FERTILIZATION.

Table 25. The immediate effect of pollination upon the weight of seeds of
maize. (Selfed and reciprocally grossed seeds from the same ears.)

Pollen

A X B

B X A

Pedigree number of

mixture

Selfed

Crossed

Selfed

parent plant — A

parent plant — B

number

White

Light

yellow

Yellow

91 3 1 ^ Q 7-^7-1

1 4 1 O °.0 4 ^ 7 1 1 -4

1 4~ I U->_>U 4-9 1 -1 I t:

o
Z

91 4 14_Q_7_^7_9

on
Zv

Q
O

0 1

O 1

14 1 O '30 4. 4-7-1 1 -4

o
o

Lj X— O X O f 4 O 4 O

Q 1

oo
ZZ

11 in x 0-4.-4-7-1 1-10

1 4- JL U-oU-4-O- I -X 1 1 U

21-3-13-9-7-57-5

14-10-30-4-3-7-11-2

21-3-13-9-7-57-7

14-10-30-4-4-2-7-6

21-3-13-9-7-57-10

14-10-30-4-3-7-1 1-1

21-3-13-9-7-57-14

A
4

14 1 0-40-4-4-9-7-4

21-3-13-9-7-57-20

1 4_1 0-40-fi-1 1 -3-1 1 -3

VJ— O U U 1 X~0 A 1 J

21-3-13-9-7-57-24

14-10-4-6-4-7-8-5

21-3-13-9-7-57-25

11 1 O 4 A 1 A 9 19ft
14-1 U-4-0- 1 1 Z-o

21-3-13-9-7-57-29

14-10-30-4-3-7-11-7

21-3-13-9-7-57-31

14-10-30-4-3-7-11-8

21-3-13-9-7-57-33

14-10-30-4-3-7-11-9

21-3-13-9-7-57-35

14-10-30-4-3-7-11-18

21-3-13-9-7-57-36

14-10-30-4-4-2-7-14

21-3-13-9-7-57-43

14-10-30-4-4-2-7-2

Average

Increase of crossed

above selfed

Table 26. The immediate effect of pollination upon the weight of seeds
of maize. (Out-crossed seeds resulting from some of the same pollen mixtures
used in Table 25.)

Average weight of seeds in centi

grams

Pedigree number

Pollen mixture

Cross

Cross .

of parent plant — C

number

CXA

CXB

20A-8-5-35-8

20.5

20A-8-5-35-3

19.7

20A-8-5-35-4

25.4

20A-8-5-35-11

20.3

20A-8-5-35-24

27.3

20A-8-5-35-26

25.9

20A-8-5-35-6

20.2

20A-8-5-35-13

25.8

20A-8-5-35-15

27.5

20A-8-5-35-18

20.9

20A-8-5-35-21

18.9

20A-8-5-35-30

21.2

20A-8-5-35-37

21.0

Ave. 21

.5 21.5

22.6

Average

22.7

Increase of (CxB) over (CxA)... .

Percent increase

64 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

different variety. All the data taken together clearly show that
an increase in endosperm development in maize is one of the
common manifestations of heterosis.

Since the crossed seeds receive a noticeable impetus in develop-
ment it seemed quite likely that the foreign pollen might be more
efficient in fertilizing than the self pollen and hence a greater
number of crossed seed than selfed would be produced. Such is
not the case, however, as an examination of a large amount of
data has shown.

In performing the mixed pollinations no attempt was made to
have more than approximately equal quantities of pollen. It is
impossible to get a mixture of equal quantities of functional
pollen because it varies so in respect to viability. Since the same
mixture of pollen was applied to both plants the ratio of the seeds
resulting from "yellow" pollen to the seeds produced by the
"white" pollen should be the same on both ears. Thus if there
were no selective fertilization the percent of white seeds on one
ear plus the percent of dark yellow seeds on the other, selfed seeds
in both cases, should equal the sum of the percents of the crossed
seeds on each ear. An excess of crossed seeds would then indicate
a selective fertilization in favor of the crossed pollen. As a small
excess of selfed seeds was obtained any Selective fertilization in
favor of the foreign pollen certainly did not take place.

The numbers of the crossed and selfed seeds, of which the
weights are given in Tables 25 and 26, together with a large
amount of similar data are not given here for fear of unduly
burdening this publication with tables but they show, on the
whole, a small excess of selfed seeds instead of crossed seeds.
The results of an experiment designed to test this point in a some-
what different way are given in Table 27. Here instead of taking
a mixture of pollen from two plants of two different strains a
large amount of pollen was collected from an approximately
equal number of plants of two long inbred and exceedingly uni-
form strains of maize. The two lots of pollen were sifted to
obtain pure pollen and equal quantities of each were carefully
measured out, thoroughly mixed together and applied to a number
of ears of each of the two strains which furnished the pollen —
A and B — and to a third strain — C — distinct from either.
Although the tassels were bagged on the same day and the pollen
collected two days later and equal quantities of each taken there

HETEROZYGOSIS AND SELECTIVE FERTILIZATION.

OS (M
CO

iO iO

ON I

I J5

iS ***

CO OO OO rt<

CO TjH Tt^

i-H !M CO Tt<

uuou

to iO

•O OS
CO

P * c

U <R * a

Tt< OS

CO CO <M_|_
iO I

os os oo o

i— ' 0<I CO rt» CO
CO OQ CO CO CO

io ic

CO >OC5

C CO

oj os

O CO CO
0C 00

PL, 3

i— i oo co rt< to

eq a s ;_q »

ill

0 x -C

61*

P-, 3
0

to m

1— I tO Tf

to Tti os

NOOOINO
^ 00 t> GO CO CO
CO i-t CM CO CO CM

i—i CM CO >o CO

«««

O LO

OS rt< Tf<
Tti iO CS

t>- oo

iO OS

00 TfH O

i-i CO

o o

& O-

a 2

O Oh

5-i O

s 2 I- d
o a^c o

5? O 02 a

" o a£
Q

O uJS

cent

seed

per

Ave. wt.

grams
Increase.

I —

<M CD

— O

- '3

1 0

5<1

"f O

X CO X

07 c

xs ll -3

-SO X

1 z

c £
_o a d

« 5.
Jig"

c c -r;

£ >>o

O &

66 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

was not equal quantities of functional pollen as the number of
seeds given in Table 27 show. The great inequality of functional
pollen may have been due to the fact that the pollen of the B
strain was more moist and tended to aggregate into a flocculent
mass while the pollen of the other was perfectly dry and each grain
remained separated from the others. For this reason it was dif-
ficult to measure the two lots of pollen equally and the dry pollen
clustered about the fine lumps of moist pollen when the two
kinds were mixed and was probably first to gain access to the
stigmas. The difference between the two kinds of pollen was not
due to any external differences, as far as could be seen, and
indicate differences in the rate of maturing after shedding.

Whatever may be the cause of the great difference in fertilizing
power this does not effect the point under investigation. How-
ever different the pollen may be, the seeds resulting from " yellow "
pollen should be in the same ratio to the seeds resulting from the
" white " pollen on one ear as the ratio of the same two kinds of
seeds on the other ear within the limits of the error of random
sampling if there is no selective fertilization one way or the other.
And both these ratios should be the same as the third ratio
obtained when this same mixture of pollen is used to produce
seeds on a plant of a different variety of maize. Let us see what
the figures given in Table 27 show. Of the reciprocal crosses and
selfs the proportion, expressed as percent, is as follows:

Seed color carried by pollen Yellow White Yellow White

Type of Seeds Selfed Crossed Crossed Selfed

Actual proportion obtained 98.490 : 1.510 :: 96.600 : 3.400

Closest perfect proportion 97.545 : 2.455 :: 97.545 : 2.455

Deviation -|-.945 -.945 -.945 ' +.945

The deviation from the closest perfect proportion is in favor
of the selfed seeds. This theoretical ratio agrees very closely
with the actual ratio obtained from the out-crossed seeds as
shown in Table 27 although there is considerable difference
in the results from the different ears. Letting S stand for selfed
and C for crossed the probable error of the determination
S . .6745 ./(SKC) m a . S

18 ± s^Tc 1 s^c • The fractlon s~Tc

gives the percent of selfed seeds and the probable error is stated

as percent. Likewise the fraction 5 ^ gives the percent of

HETEROZYGOSIS AND SELECTIVE FERTILIZATION. 67

crossed seeds and the probable error is the same as for the percent
of selfed seeds.

This same experiment was repeated with about the same
number of plants with the result of a similar excess of selfed
seeds greater than would be expected from the probable error
on the assumption that there is no selective fertilization. Does
this mean that there is a selective fertilization in favor of a plant's
own pollen and that the plant discriminates against foreign pollen
even though the seeds resulting from that foreign pollen are
greatly increased in size, weight, viability and the rate of growth
of the ensuing plants? Unless there has been a constant error in
classifying the seeds this seems to be the necessary conclusion to
be drawn from the results so far given by maize. A sufficient
number of plants will be grown from this seed to determine
definately whether or not there has been any error in the separa-
tion of the seeds so that this question can be answered with a
high degree of certainty.

In the meantime there is little doubt but that there is no great
selective' fertilization in favor of cross-pollination, if any, however
much that cross-pollination may .benefit the resulting seeds and
the plants grown from them. If this is true crossing is without
effect until the zygote is formed at the time of the union of the
male and female nuclei.

In a consideration of selective fertilization it should be remem-
bered that there are two different conditions which may be included
in the term selective fertilization. One may be said to be the
selection of different germ-plasms; the other the selection of
different cytoplasms. For example a heterozygous plant produces
pollen grains with different germinal compositions but all enclosed
in the same cytoplasm. On the other hand pollen from different
plants may differ in the nature of the cytoplasm as well as in
hereditary factors carried in the nuclear material. East and
Park ('18) have demonstrated that in tobacco there is no selective
fertilization between gametes coming from one plant although
the pollen grains differ in factors which determine fertility or
sterility of the ensuing plants. The case is quite similar to that
of the shape of pollen grains in peas which may be either all
round or all cylindrical according to the germinal composition
of the sporophyte which produced them and not according to
the factors which they carry. Where pollen grains differ both

68 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

in the factors which they carry and in the plants from which
they come, as is the case with these experiments with maize,
the conditions are quite different. It would not be surprising
that there should be selective fertilization in one case and not
in the other. East and Park have shown that a tobacco plant
which was self-sterile, pollinated with a mixture of its own and
pollen from another plant with which it was fertile, gave all
crossed seeds — a maximum of selective fertilization.

Darwin (" Cross and Self Fertilization ") found that there was a
selective fertilization in favor of foreign pollen in different plants.
Many of Darwin's experiments, however, were made in such a
way as to be open to doubt whether or not he really did obtain
such an effect. His experiments, in applying foreign pollen
sometime after self-pollination had taken place, in which he
obtained in some cases many or all apparently crossed progeny,
are open to other interpretations. The purity of the plants
pollinated was not known. External conditions influencing
fertilization were not guarded against. Taken as they stand,
however, his experiments with Mimulus, Iberis, Brassica, Raphi-
nus, Allium and Primula do indicate that in these plants there
may be a selective fertilization in favor of foreign pollen. It is
to be expected that plants which show partial self-incompati-
bility would show selective fertilization when a mixture of self
and foreign pollen was applied. In maize, however, as mentioned
before, the sterility shown is in the nature of pollen and ovule
abortion, and whenever well formed pollen is produced it seems
to be able to fertilize equally any plants if not too distinct in type.
A distinction should be made, then, between self-fertile plants and
self -sterile plants when dealing with selective fertilization.

Hyde ('14) has shown clearly that in Drosophila both males
and females of inbred lines are more productive of offspring
when mated to an individual of a different line than when mated
to one of their own. Both males and females, therefore, produce
more functional gametes than are utilized when individuals of
the same inbred lines are paired. Hence a female, impregnated
with a mixture of two kinds of spermatozoa from the same and
from different lines would produce more hybrid progeny than
inbred progeny even if equal quantities of both types of sperma-
tozoa were available for fertilization. In other words there would
be selective fertilization in favor of cross-fertilization.

LONGEVITY, HARDINESS AND VIABILITY.

Whether or not there may be a similar condition in other animals
I do not know. Even in Drosophila, fertilization by the two types
of sperm may take place equally, and a greater proportion of
close-fertilized eggs, than cross-fertilized, fail to hatch, due to
lesser vigor or lethal factors. In Hyde's experiments the type
of fertilization had no marked effect on the number of eggs laid,
only on the percentage which hatched.

In maize, and possibly all plants which show no self-incompati-
bility, 'the fact seems clear that crossing is wholly without effect
until the fertilization process is completed.

Although there is apparently no effect of crossing in maize
until the zygote is formed, such an effect is apparent immediately
afterwards. In addition to the increase in endosperm development
there is also an increase in the vigor of the embryo. Whether or
not the size of the embryo in the seed is increased has not been
actually determined,* other than by inspection, but it undoubtedly
is, along with the endosperm. When crossed and selfed seeds
from the same ear, grown on a plant which has been inbred
previously for several generations, are planted a striking difference
is soon apparent. The crossed seedlings appear from one to two
days before the selfed seedlings and may be two or three inches
above ground before any of the selfed plants begin to appear.
(See Plate Xlb). From then on the superiority of the crossed
over the selfed plants increases rapidly as shown by the curves
in Figure III.

The Effect of Heterozygosis upon Longevity, Hardiness

and Viability.

An increased longevity, viability and endurance against un-
favorable climatic conditions have been frequently noted in
hybrids. Kolreuter and Wiegmann both mention this fact.
Gartner in his book "Bastarderzeugung im Pflanzenreich" devotes
considerable attention to this feature. Under the heading
" Ausdauer und Lebenstenacitat der Bastardpflanzen" he makco
the following statements.

" There is certainly no essential difference between annual and biennial
plants and between these and perennials in regard to their longevity;
for it is not seldom that different individuals of the same species have a
longer life at times as, for example, Draba verna, which has annual and

70 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

biennial forms; the longevity of a plant furnishes thereby no specific
diiferences but signifies at most a variability as Prof. W. D. I. Koch
has shown. However, in hybrids this difference deserves special con-
sideration. In most hybrids an increased longevity and greater endur-
ance can be observed as compared with their parental races even if they
come into bloom a year earlier. The union of a annual, herbaceous
female plant with a perennial, shrubby species through hybridization
does not shorten the life cycle of the forthcoming hybrid as the union
of Kyoscyamus agrestis with niger, Nicotianz rustica with perennis,
Calceolaria plantaginea with rugosa shows, and so also in reciprocal
crosses when the perennial species furnishes the seed and the annual
species supplies the pollen, as Nicotiana glauca with Langsdorfii, Dianthus
caryophyllus with chinensis, Malva sylvestris with Mauritiana or biennials
with perennials and reciprocally as Digitalis purpurea with Ochroleuci or
lutea and lutea with purpurea or ochrohuca with purpurea. From the
union of two races of different longevity comes usually a hybrid into
which the longer life of one or the other of its parent races is carried
whether it comes from the male or female parent species."

Many more instances are given by Gartner from his own ob-
servations and those, of others to enable him to reach the following

conclusion :

" These examples support the statement of Kolreuter's that the longer
life of hybrid plants is to be counted among their usual properties."

With regard to the resistance of hybrids to unfavorable weather
conditions he goes on to say:

" With their longevity stands, in the closest relation, the fairly
common property of hybrids to withstand lower temperatures than their
parental races without injury to their growth and vegetative life. Kol-
reuter first observed that Lycium barbara-afrum in south Germany
withstood the winter in the open field; although Lycium afrum must be
wintered over, at least, in a cold frame. The cross of Nicotiana Tabaco-
undulata, according to Sageret in France had an increased life, although
in a protected place, in *>pen field. W. Herbert reports that Rhododen-
dron altaclarae, which is a hybrid union of R. pontica-cantawbiense 9 with
the very sensitive Nepalense arboreum coccineum d\ has been grown in
the open in England; also Robert Sweet confirms the same result by
a hybrid crinum and many other hybrids of bulbous plants grown in
open field whose parental species must be grown in the hothouse.

" Lobelia syphilitica-cardinalis wintered over with a light covering
in the winter of 1832-1833 with 5°F in open field. . Lychnicucubalus
albus and ruber lasted three years in open field although cucubalus
viscosus in south Germany did not survive in open field. All hybrids
of genus coccineum stood over the winter of 1842-1843 with 5°F. in the
open, although the pure species seldom lives through our usual winters
of 43° to 9.5° F. Prof. Wiegmann reports similar results.

LONGEVITY, HARDINESS AND VIABILITY.

" Very frost sensitive species of Nicotiana and their hybrids did not
withstand, under the same conditions, such low temperatures as the
afore-mentioned plants; but we have flowered and carried over part
of them wherever they were well covered with snow, for example, N.
quadri-valvis glutinosa, rustica-quadrivalvis, these withstood 25° F. and
yet have continued blooming although N. glutinosa, quadrivalvis, panicu-
lata, T abacum and rustica were already frozen by 32° F. Moreover other
crosses of very sensitive and tender species of this genus as paniculata-
Langsdorfii, vincaeflora-Langsdorfii, vincae-flora-quadrivalvis have been
carried over in an active growing condition two to three years, and
glauca-Langsdorfii three years in a cold house with 39° to 42°. The
hybrid N. paniculatarustica-paniculata was kept over in a cold house in
the cold winter of 1839-40 but its leaves were yellow. Among all the
species of this genus the cross of N. suaveolenti-macrophijlla showed
itself to be the most hardy. On the 16th of October of its first year
(1828) its top was frozen but it did not suffer from this, and 12 days later
put out a new shoot from the root and its leaves lasted through the winter
in a cold house in a fresh, green condition although the other species
were yellow and this plant was the first to start into growth in the spring.
The same endurance Sageret observed in Nicotiana suaveolenti-virginica.
All these plants in the last year of their vegetative life seemed to die
off more as the result of the unfavorableness of the weather than of old
age."

Exceptions are noted by Gartner in that some species which
were not resistant to cold did not give resistant hybrids. In
many cases the hybrids were weak because of the distant re-
lationship of the parental races.

Sargent ('94) reports a remarkably vigorous and hardy hybrid
tree supposed to be a cross of the tender English walnut, Juglans
regia and the common butternut Juglans cinerea. He says:
p. 434

"My attention was first called to the fact by observing that a tree which
I had supposed was the so-called English walnut — Juglans regia, in the
grounds connected with the Episcopal School of Harvard College at
Cambridge, was not injured by the cold of the severest winters, although
Juglans regia generally suffers from cold here — and rarely grows to a
large size. This individual is really a noble tree; the trunk forks ab u
five feet above the surface of the ground into limbs and girths, at the
point where its diameter is smallest, fifteen feet and two inches. The
divisions of the trunk spread slightly and form a wide, round-topped
head of pendulous branches and unusual symmetry and beauty, and
probably sixty to seventy feet high."

Heterosis is also shown in a resistance to bacterial and fungus
diseases. Some of the inbred strains of maize are very susceptible

72 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

to the bacterial leaf-wilt and in some years at the end of the
season all the plants of these strains appear as if they had been
scorched by fire while other strains in adjoining rows are un-
touched. Other strains have quite a large percentage of plants
attacked by smut. Crosses, however, of these susceptible strains
with those which are not affected by these parasitic organisms
are only slightly or not at all affected.

Table 28. Susceptibility to smut ( Ustilago zeae) of a non-inbred

VARIETY OF MAIZE, SEVERAL INBRED STRAINS DERIVED FROM THIS
VARIETY AND THE FIRST AND SECOND GENERATION CROSSES
BETWEEN THE MOST SUSCEPTIBLE AND THE LEAST SUSCEPTIBLE
STRAINS.

Percent of plants affected

Total

Pedigree number

number

percent

of plants

Plot I

Plot II

Plot III

grown

affected

1.75

114

1.75

1-9-1-2-4-6-7-5

0 '

' '.27

.56

596

.34

1-7-1-2-2-9-2-1

2.17

.35

408

.49

1_7_1_1_1_4_7_5

8.79

10. 16

5.77

950

9.79

l_6-l-3-4-4-4-2

992

(1-6-1-3) X (1-7-1-DF,.. .

2.48

439

2.28

(1-6-1-3) X d-7-l-l)F2.. .

5.15

In Table 28 are given the per cent, of plants affected by smut
( Ustilago zea, Beck. Ung.) of the original, non-inbred Learning
variety of maize previously spoken of and four inbred strains
derived from this variety by ten or eleven generations of self-
pollination. Seed of the four inbred strains was planted in three
rather widely separated plots in the same field' in 1917. Two of
the strains showed only a small infection by this parasite; one
showed about 10 per cent infection and one had not a single plant
affected in all three plots in a total of nearly one thousand plants.
Since the differences which these four strains show are fairly con-
sistent in the different places grown it can hardly be doubted but
that segregation of susceptibility to parasitism has occurred in
the inbreeding process. The first generation hybrid between the
most resistant and the most susceptible strain was free from smut
in one plot and but slightly affected in another. The second
generation hybrid grown side by side with first generation showed

LONGEVITY, HARDINESS AND VIABILITY.

considerably more infection although the number of plants grown
was small. This is fairly good evidence that resistance to smut
in maize tends to dominate in crosses between plants which differ
in this respect.

Tisdale, according to L. R. Jones ('18) also finds that in flax
disease resistance tends to be dominant although . the hybrids
are more or less intermediate in this respect and the method of
inheritance is rather complex. Biffen ('12), on the other hand,
concluded that the resistance to rust in wheat was recessive.
Likewise, Weston ('18) states that maize and teosinte-maize hy-
brids are extremely susceptible to a downy mildew (Peronospora
Maydis, Rac.) in Java and other places, although teosinte (Euch-
laena mexicana, Schrad.) is immune.

Data from another source have been obtained from the garden
radish (Raphanus sativus, L.). A white-rooted variety of radish
was allowed to go to seed alongside a red-rooted radish. Seed
collected from the white-rooted plants was sown thickly in a flat
and when they came up it was seen that a number of the seedlings
were crossed from their red colored stems. The seedlings were
quite badly attacked by the "damping-ofF' fungus and large
numbers of them were killed, but a far less number of the crossed
seedlings were affected as shown by the decay of the tissues at
the base of the stem. The figures obtained are given in Table 29.

Table 29. Comparative susceptibility to " damping-off " disease
of selfed and crossed radish seedlings.

Variety of Radish

White Seedlings, Selfed

Red Seedlings. Crossed

Number
grown

Number
affected

Percent
affected

Number
grown

Number
affected

Percent
affected

Short, white. . .
Long, white. . .

349
76

142

40.7
36.8

30
7

4
0

13.3

. 0

Gernert ('17) reports a case of immunity to aphis attack of
teosinte-maize hybrids in which the maize parent was badly
infested whereas the teosinte parent and the hybrid entirely
escaped injury.

Together with these manifestations of heterosis in its influence
on hardiness there is an increase in the viability of crossed seeds
as compared to selfed seeds from the same ears as shown in Table

74 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Table 30. The effect of heterozygosis upon germination — a compar-
ison OF CROSSED AND SELFED SEEDS FROM THE SAME EARS OF MAIZE

Pedigree number
of female parent

Pedigree number
of male parent

Percent increase in weight
of crossed seeds over
selfed

Number of seeds planted
of each

Number of selfed seeds
germinated

Number of crossed seeds
germinated

Percent selfed seeds
germinated.

Percent crossed seeds
germinated.

Excess of percent crossed
seeds germinated over
selfed

21-3-13-9-7-57-13

14-10-30-6-2-13-5-13

8.2

121

19.8

38.0

18.2

21-3-13-9-7-57-17

14-10-30-6-11-3-11-17

18.8

66.7

94.9

28.2

21-3-13-9-7-57-21

14-10-30-6-11-3-11-4

7.1

87.5

90.6

3.1

21-3-13-9-7-57-38

14-10-30-4-4-2-7-38

68.0

63.6

100.0

36.4

21-3-13-9-7-57-39

14-10-30-4-4-2-7-7

18.3

48.5

78.8

30.3

21-3-13-9-7-57-54

14-10-30-4-4-2-7-7

3.9

19.6

27.8

8.2

21-3-13-9-7-57-58

14-10-4-6-4-7-8-15

8.8

100

43.0

68.0

25.0

21-3-13-9-7-57-59

14-10-4-6-4-7-8-10

10 0

75.0

100.0

25.0

21-3-13-9-7-57-63

14-10-4-6-4-7-8-29

17.5

83.9

96.8

12.9

21-3-13-9-7-57-64

14-10-4-6-4-7-8-29

13.3

64.3

100.0

35.7

21-3-13-9-7-57-65

13.5

87.2

95.7

8.5

14-10-30-4-4-2-7-12

21-3-13x9-7-57-38

8.3

96.6

98.9

2.3

Total

16.3

635

339

442

53.4

69.6

16.2

* Seeds crossed but number of parent unknown.

30. Seeds which were secured from some of the mixed pollinations,
reported previously, were sown in flats. Without exception the
crossed seeds showed a higher percentage of germination than
the selfed seeds from the same ears as can be seen in Plate Xlb.
These seeds were planted two months after ripening. Whether
or not an increase in age would show greater differences in viability
is not known but it is quite likely that the difference might be-
come even greater with age up to a certain point. The low germi-
nation of both crossed and selfed seeds in some of the ears was
due to the fact that they were moldy on account of late ripening
and damp weather.

The increased vegetative vigor as manifested by an increased
facility of vegetative propagation in hybrids has been repeatedly
spoken of. Kolreuter, Wiegmann, Sageret and Focke make a
special mention of this phenomenon.

Moreover there is no positive evidence that plants which are
propagated vegetatively lose any of their hybrid vigor which

LONGEVITY, HARDINESS AND VIABILITY.

they may have, no matter how many generations of asexual re-
productions take place. Undoubtedly most varieties of culti-
vated fruits, flowers, ornamental plants and field crops which are
commonly propagated vegetatively, owe their excellence in part
to heterosis.

From time to time the supposed degeneration of plants in long-
continued vegetative propagation has been much disputed.
Knight ('99) and Van Mons ('36) contended that they did degen-
erate, but Lindley ('52) reviewing Knight's work thought that
the evidence did not support such a view. Gartner states that
the characteristics of a hybrid do not change throughout the
whole life cycle of the individual, even when it is propagated and
disseminated by buds, cuttings or^ layers.

Darwin believed that a degeneration took place largely for the
same reason that he thought long continued seif-fertilizaticn was
injurious. Asa Gray ('76), in reviewing Darwin's opinions on
this matter, says (p. 347) :

"The conclusion of the matter, from the scientific point of view is, that
sexually propagated varieties of races, although liable to disappear through
change, need not be expected to wear out and there is no proof that they
do, but that non-sexually propagated varieties, though not especially
liable to change, may theoretically be expected to wear out, but to be a
very long time about it."

Gray, however, cites cases of horticultural varieties propagated
since the time of the Romans with no apparent loss of vigor.
Whitney ('12a, b, c) and A. F. Shull ('12b) believe that an actual
degeneration takes place in parthenogenetic reproduction in the
rotifiers. The work of Enriques ('07), Woodruff ('11) and Jennings
('12) on Paramecium proves almost beyond doubt that there is
no degeneration in this organism although reproduction by
fision in the infusoria may be considerably different from vegetative
propagation in the higher plants. Hedrick ('13), from the evi-
dences of long-continued varieties of fruits, and East ('08) working
with potatoes and reviewing extensively the whole question be-
lieve that there is no evidence that a real degeneration takes place
which cannot be accounted for on the basis of the accumulation
of disease or other external effects. East ('10), however, suggested
that such a degeneration, if ever proven, might be accounted for
on the .basis of a decreasing effect of the physiological stimulation
assumed to be derived from heterozygosity. A. F. Shull (12a)

76 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

holds a similar opinion. From the nature of the problem it can
hardly be settled satisfactorily one way or the other by experi-
mental means. As it stands at present there is no clear evidence
that there is a degeneration in long continued asexually propagated
plants. The burden of proof rests with the positive side.

The Effect of Heterozygosis upon the Time of Flowering

and Maturing.

Many investigations have indicated that there is a hastening
of the time of maturity due to heterozygosis. That there is an
acceleration in the rate of growth is, of course, evident from the
great increase in size shown by hybrids grown in the same season
with their parents. There is, moreover, considerable evidence
from previous work and from the data to be given here to show
that hybrids not only grow to a larger size but complete their
growth in a shorter time than the parents take to complete a
smaller amount of growth. In other words, heterozygosis tends
to hasten the time of maturity as well as to increase size.

The investigations of Kolreuter, Gartner, Focke and Darwin
show a large number of specie- and variety-crosses wherein the
hybrid flowers before either of the parents. Both Kolreuter and
Gartner give instances of perennials which commonly bloom in
the second or third year whose hybrids bloom in the first year.

The most extensive observations bearing on this relation of
heterosis to time of flowering are those given by Darwin in his
"Cross and Self Fertilization in the Vegetable Kingdom. " He
gives the time of flowering of 28 crosses between different strains
within many different species — which show positive evidence of
hybrid vigor. Of these 28 crosses 81 per cent, flower before the
parents. Four cases are given where the crosses are less vigorous
than the parents and in each of these the parents flowered first.

Recent experiments in hybridization show, almost without
exception, that crosses which result in an increase in vigor also
result in a hastening of the time of flowering. One exception to
this statement must be noted in the cross between a large dent
and a small pop variety of corn repoited by Emerson and East
('13). This cross showed distinct evidence of hybrid vigor in an
increase in internode length over that of both parents. The
parents differed in time of flowering by 25 days. The first genera-
tion of the cross grown the same year as the parents was "distinctly

TIME OF FLOWERING AND MATURING.

intermediate' ' in time of flowering. There was an increase in the
rate of growth necessarily as the plants were larger than the av-
erage of the parents.

Data bearing upon the relation of heterozygosis to the time of
maturing has been secured from two different plants, tomatoes
and corn. A large part of the data on tomatoes was collected by
Prof. H. K. Hayes, now at the Minnesota College and Station.

Four commercial varieties of tomatoes were successively self-
pollinated for feur years. Two first generation crosses between
these varieties were grown in each of the four years and compared
as to yield of fruit and time of production with the two selfed
parents. In every case the same plants which were used to pro-
duce the selfed seed for the next generation were also used to
make the crosses. For this reason and because tomatoes are
naturally self-pollinated and are hence in a homozygous condition
the first generation crosses can be compared strictly with their
parents.

From thirty to fifty plants of each variety and cross were grown
each year. The fruit was picked as it ripened at intervals of from
3 to 5 days and the average production per plant was determined.
One of the crosses was between varieties which had approximately
the same time of ripening. This first generation cross did not ex-
ceed, in total yield, the average of the two parents and did not
differ from them in respect to time of production.

The other cross, however, yielded, each year, an average of 16
percent above the better parent. The two varieties used in
making this cross differed in time of production by an average
of five days. The first generation cross while yielding 16 percent
more than the late parent was each year fully as early as the
early parent. Although the difference in time of production
between these varieties is small the consistent results obtained in
four successive years are certainly significant.

Similar results were secured with sweet corn. A first generation
cross between an early variety of sweet com, Golden Bantam and
a late variety, Evergreen, was grown in 1916 together with the
two parental varieties and compared in time of flowering, number
of ears per plant and in height. They were all planted at the
same time but rather late in the season so that the early and late
varieties bloomed at more nearly the same time than is usually
the case. About half of the plants of the early variety were

78 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

showing silks before the late variety commenced to silk ou t
The first generation cross was slightly earlier than the early
parent in producing silks. The cross was noticeably affected by
vigor of crossing in that it was fully as tall as the taller parent
and averaged more ears per plant than either parent although
the ears were not as large as those of the Evergreen variety.

Much more extensive and authoritative data have been secured
from a comparison of inbred strains of corn with their first genera-
tion crosses. Forty-two strains of corn which had been continu-
ously selfed for from 5 to 11 generations and 100 first generation
crosses representing different combinations between these selfed
strains were grown under the same conditions as to time of
planting and culture. Both the inbred strains and their crosses
were exceedingly uniform in time of flowering and maturing.
All the plants in any selection flowered and matured within a few
days. About 60 plants of each were grown. At intervals of one
week during the flowering season the number of selections of the
selfs and crosses which had flowered by that time were noted.
Similarly at the end of the season the selections which were mature
were noted at intervals. Although the time of maturity can not
be so definitely determined as the time of flowering all the plants
in a selection were uniform in this respect. For the flint varieties
the glazing of the ears and for the dent varieties the denting of
the kernels were taken as indications of maturity. The crosses
yielded, on the average, 180 per cent more than their parente.

Together with this increase in the amount of growth there
was a noticeable hastening of both the time of flowering and
maturing. In time of flowering the crosses were four days and
in maturing eight days earlier than the average of their parents.
Since the crosses gave a large increase in the .total amount of
growth and produced this growth in a somewhat shorter time than
their inbred parents it is all the more evident that heterozygosis
increases the rapidity of growth. See Plates VII a and b.

The Relation of the Effects of Heterozygosis and of
the Environment.

East ('16) has stated that heterozygosis " affects a result
comparable to favorable external conditions." In a cross between
two varieties of Nicotiana he found that the first generation

EFFECTS OF HETEROZYGOSIS AND THE ENVIRONMENT. 79

gave a noticeable increase in the amount of growth as shown by
the height and general size of the plant as the result of hetero-
zygosis. The corolla length of the flowers, which is very little
affected by environmental factors, was not increased above th(
average of the two parents.

The similarity of the effects of heterozygosis to the environ
mental effects is also shown in the affect of crossing on the numbei
of nodes and internode lengths of corn. As was noted fron
Tables 15 and 13 the number of nodes is increased only 6 percen
while the height of plant is increased 27 percent. This is exactl}
the effect that nutritional factors have. The height of plant k
reduced under poor conditions by a reduction in internode length
without reducing appreciably the number of nodes.

In general it is evidently true that heterozygosis affects many
characters in the same way as the environment, but it should be
noted that in time of maturity these two factors have directly
opposite effects. It is generally recognized, I believe, that favor-
able external conditions such as increased moisture or fertility,
where these are limiting factors, which result in a greater total
amount of growth tend to prolong both the time of flowering and
the completion of growth. Conversely unfavorable external
conditions which stunt the plants and limit their growth tend tc
hasten their period of flowering and maturity. There are, of
course, certain exceptions to this statement.

Whether or not the effect of heterozygosis in hastening maturity
can manifest itself independent of any increase in vegetative
luxuriance or other manifestations of hybrid vigor is not known.
The results given here would indicate that the vigor derived from
crossing enables the plant to carry on its life processes more
easily and more efficiently and thus to accomplish its task in a
shorter time.

With regard to the effects of heterozygosis in animals much
the same relation is shown with the external environmental
effects as in plants although the rate of growth and size obtained
at maturity may be more definitely fixed in animals than in
plants. According to Castle ('16) there is an increase in the rate
of growth as well as the attainment of a larger size at maturity
in hybrid guinea-pigs. Hyde ('14) also finds an increase in rate
of growth and hastening of sexual maturity on crossing in Droso-
phila. These effects in animals are probably greater than could

80 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

be obtained by any amount of feeding or care just as it is the case
in plants

It may be stated briefly that the effects of heterozygosis in
both animals and plants, not too distantly related, all together
contribute towards an increased reproductive ability and this
effect has probably been of fundamental importance in evolution
in establishing sex.

In the foregoing account of some of the most noticeable effects
of crossbreeding upon development we have been dealing only
with crosses among closely related organisms. It is of course,
well known that in crosses between distantly related forms the
beneficial effects of crossing may disappear and the effects become
increasingly more injurious as the degree of dissimilarity becomes
greater. The most frequent, pronouncedly injurious effect is
the reduction or complete loss of fertility. This may or may not
be accompanied by a great acceleration of growth. This is
shown in many plants, notably by Gravatt's Radish-Cabbage
hybrid and by Wichura's Willow hybrids as well as by many
good illustrations given by Gartner and Focke. It is perhaps
not surprising that the reproductive ability should be the first
to suffer since reproduction is the most difficult task the organism
has to perform. The failure of the reproductive mechanism
might divert the energies into bodily growth and thus in part
account for the large size and great vigor of some sterile hybrids
but, as all are agreed, this can not entirely account for the great
increases in size nor obviously does it apply to the more common
cases where both size and productiveness are increased at the
same time.

To sum up one can therefore say that, in plants, crossing may
have a great range of effects, according to the degree of relation-
ship of the parents, from a condition in which : the cross is not
possible and no seed produced; seed may be produced but fail
to germinate; plants may be produced which are either very weak,
normal or very vigorous without being able to reproduce them-
selves ; plants which are both more vigorous and more productive
than their parents; to a condition in which they are so closely
related that the crossed plants do not differ appreciably from
selfed plants. A similar series can be arranged with animals.

SUMMAKY OF THE EFFECTS.

Summary of the Effects of Inbreeding and Crossbreeding.

Before taking up a theoretical consideration of the cause of
hybrid vigor and its importance in the establishment of sex it is
well to summarize briefly some of the main conclusions, with
regard to the effects of inbreeding and crossbreeding on develop-
ment, to be arrived at from a study of the investigations discussed.

EFFECTS OF INBREEDING.

1. Continued inbreeding results in the segregation of a variable
complex into a number of diverse types which are uniform
within themselves.

2. The segregates which differ in visible, qualitative characters
also differ in quantitative characters; types with abnormalities
appear which cannot reproduce themselves; others appear which
are perpetuated with difficulty; others are obtained which are
perfectly normal in structure and function. These latter are
usually less well developed, but may be as well or better developed
than the original stock from which they are derived.

3. The change in size, structure, or function and reduction in
variability is most noticeable in the earlier generations of in-
breeding, rapidly becomes less and the surviving inbred strains
are uniform and constant.

4. The rate of approach to uniformity and constancy differs
in different lines.

5. These uniform and constant inbred strains are quite com-
parable to naturally self -fertilized species.

6. No single effect can be attributed to inbreeding other than
the reduction in variability.

7. All these results are in conformity with Mendel's law and
Johannsen's genotype conception.

THE EFFECTS OF CROSSBREEDING.

1. Heterosis accompanies heterogeneity in germinal constitu-
tion whether or not the organisms crossed are from the same or
diverse stocks.

2. Heterosis is widespread in its occurrence throughout the
plant and animal kingdoms.

82 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

3. Heterosis is shown as an increase in the size of parts rather
than an increase in the number of parts.

4. Cross-fertilization is without effect until the zygote is formed;
from that time on heterosis may be apparent throughout the life
of the individual.

5. Heterozygosis has an undiminished effect on plants propa-
gated vegetatively.

6. Heterozygosis ma}r have a stimulating effect on some char-
acters and a depressing effect on others in the same organism.

A Mendelian Interpretation of Heterosis.

It is due to the work of G. H. Shull ('08, '09, '10, '11) and of
East ('08, '09) and East and Hayes ('12), supplemented and
confirmed by the results given here, that we no longer believe
that inbreeding is a process of continuous degeneration. Also
these investigators first demonstrated clearly that the same
principle was involved in the loss of vigor accompanying in-
breeding and the increase in vigor resulting from crossing.

To account for this w^ell nigh universal loss of vigor when do-
mesticated races of plants and animals are inbred, they thought
it necessary to assume a physiological stimulation which was
present when unlike germplasms were united and which disap-
peared as homozygosis was brought about automatically by
inbreeding. Part of the effects of inbreeding were due, according
to their views, to the segregations into pure lines of different
hereditary complexes and the appearance of previously hidden
recessive characters, and part were due to the loss of this stimu-
lation.

G. H. ShulFs ('14) opinion as to the way germinal heterogeneity
induces vigor is stated briefly as follows (p. 126) :

"The essential features of the hypothesis may be stated in more general
terms" as follows: The physiological vigor of an organism, as manifested
in its rapidity of growth, its height and general robustness, is positively
correlated with the degree of dissimilarity in the gametes by whose union
the organism has been formed. In other words, the resultant hetero-
geneity and lack of balance produced by such differences in the reacting
and interacting elements of the germ-cells act as a stimulus to increased
cell-division, growth, etc. The more numerous the differences between
the uniting gametes — at least within certain limits — the greater, on the

A MENDELIAN INTERPRETATION OF HETEROSIS. 83

whole, is the amount of stimulation. These differences need not be
Mendelian in their inheritance, although in most organisms they prob-
ably are Mendelian to a prevailing extent."

Both the view stated above and that of East and Hayes assume
that the increase in development is due to a reaction between
different elements in the nucleus and that this stimulus disappears
when homozygosity is reached. A. F. Shull ('12a) has proposed
a slightly different idea in that he assumes the stimulus to be due
to the reaction of new elements in the nucleus, brought in by
cross-fertilization, to the maternal cytoplasm. According to his
view there might still be a stimulation even after complete homo-
zygosity is attained. Also in asexual propagation he supposes
that the cytoplasm might become gradually accustomed to a
heterozygous nucleus, hence long continued asexual reproduction
might lead to a gradual reduction in vigor which this writer finds
does occur in the rotifer, Hydatina senta. ('12b).

It should be remembered, however, that both these hypotheses,
as to the effect of germinal differences, postulate a stimulation to
account for an increase in development as the facts demand. It
would have been even more plausible to postulate a depressing
effect had the facts been otherwise. The only basis for a stimu-
lation of this kind is in the fact that fertilization initiates the
development of the egg. Heterozygosis, however, is not con-
cerned with the starting of the development of the egg, but only
with the rate of development after growth is commenced. Is it
not more plausible that "a lack of balance" occasioned by the
union of unlike germplasms would retard development rather
than stimulate it?

Keeble and Pellew ('10) first suggested that dominance of
characters contributed by both parents might be a factor in the
increased vigor of hybrids. They illustrated this conception by
a cross between two varieties of peas which possessed features of
both parents, and were taller than either.

Bruce ('10) has shown that the total number of dominant
factors is greater in a hybrid population than in either parental
population and that there is consequently a correlation between
the number of dominant factors and hybrid vigor. As far as I
know, Bruce has never followed up this suggestion. He did not
show why it was that the presence of a greater number of domi-
nant factors brought about an increase in growth, nor did he
13*

84 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

show why it was that all the dominant features could rarely or
never be accumulated in certain individuals and races which would
therefore show no reduction in vigor when inbred.

East and Hayes ('12) attempting to distinguish between domi-
nance and the effects of heterozygosis make the following state-
ment (p. 31) :

"The term vigor has hitherto been used with the general meaning
which the biologist readily understands. We will now endeavor to show
in what plant characters this vigor finds expression. It is not an easy
task because of the possibility of confusing the phenomenon of Mendelian
dominance with the physiological effect due to heterozygosis. The con-
fusion is due to a superficial resemblance only. Dominance is the ex-
pressed potency of a character in a cross and affects the character as a
whole. A morphological character, like the pods of individual maize
seeds, or the product of some physiological reaction like the red color of
the seed pericarp in maize, may be perfectly dominant, that is, it may be
developed completely when obtained from only one parent. Size char-
acters, on the other hand, usually lack dominance or at least show in-
complete dominance. The vigor of the first hybrid generation theoreti-
cally has nothing to do with these facts. This is easily demonstrated if
one remembers that the increased vigor manifested as height in the Fi
generation cannot be obtained as a pure homozygous Mendelian segregate,
which would be possible if due to dominance. Furthermore, the univer-
sality with which vigor of heterozygosis is expressed as height shows the
distinction between the two phenomena. If the greater height were the
expression of the meeting of two factors (T2, t2, x ti, T2) both of which were
necessary to produce the character, one could not account for the frequency
of the occurrence. Nevertheless, in practice the confusion exists, and
while we have considerable confidence in the conclusions drawn from our
experiments, we have no intention of expressing them dogmatically."

G. H. Shull's statements of the way in which crossing brings
about increased development, and the relation that this stimula-
tion of growth has to dominance of Mendelian characters is fairly
stated, I believe, in the following passage ('11, pp. 244-245):

"In 1908 I suggested a hypothesis to explain the apparent deterioration
attendant upon self-fertilization, by pointing out that in plants, such as
maize, which show superiority as a result of cross-fertilization, this
superiority is of the same nature as that so generally met with in Fi
hybrids. I assumed that the vigor in such cases is due to the presence
of heterozygous elements in the hybrids, and that the degree of vigor is
correlated with the number of characters in respect to which the hybrids
are heterozygous. I do not believe that this correlation is perfect, of
course, but approximate, as it is readily conceivable that even though
the general principle should be correct, heterozygosis in some elements

A MENDELIAN INTERPRETATION OF HETEROSIS.

may be without effect upon vigor, or even depressing. The presence of
unpaired genes, or the presence of unlike or unequal paired genes, was
assumed to produce the greater functional activity upon which larger
size and greater efficiency depend. This idea has been elaborated by
Dr. East and shown to agree with his own extensive experiments in self-
fertilizing and crossing maize. He suggests that this stimulation due to
hybridity may be analogous to that of ionization.

Mr. A. B. Bruce proposes a slightly different hypothesis in which the
degree of vigor is assumed to depend upon the number of dominant
elements present, rather than the number of heterozygous elements.
While all of my data thus far are in perfect accord with my own hypothe-
sis, and I know of no instance in which self-fertilization of a corn-plant
of maximum vigor has not resulted in a less vigorous progeny, it is quite
possible that I have still insufficient data from which to distinguish
between the results expected under these two hypotheses. However,
for the purpose of the present discussion, it is not necessary to decide
which of these two hypotheses (if either) is correct. Both of them are
based upon the view that the germ-cells produced by any plant whose
vigor has been increased by crossing are not uniform, some possessing
positive elements or genes not possessed by others."

A. F. Shull does not consider dominance as an adequate means
of accounting for heterosis, agreeing with East and Hayes and
G. H. Shull, as the following quotation shows: ('12a, p. 10)

"The view that vigor depends upon heterozygosis of the individual
seems to me inherently more probable than that it is due to the presence
of certain dominant genes. The former view admits of a plausible foun-
dation in cell physiology, and the essence of it may be extended to cases
of decrease of vigor in which there is no change in genotypic constitution,
and which are therefore without the pale of either theory."

Castle is also in accord with the general belief that heterosis is
not due to dominance of factors and draws a distinction between
the effects of inherited characters and the stimulus resulting from
crossing. In speaking of the increase in size in crosses between
diverse races of guinea-pigs he says: ('16, p. 212.)

"So far as heredity is concerned, the inheritance is blending, but Fx
shows an increase in size due to hybridization. This increased size,
however, does not persist into F2. It seems to be due not to heredity at all."

(And again on pages 223 and 224.)

"Cross breeding has, then, the same advantage over close breeding that
fertilization has over parthenogenesis. It brings together differentiating
gametes, which, reacting on each other, produce greater metabolic activity.
Whether or not the uniting gametes differ by Mendelian unit-characters

86 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

is probably of no consequence. That they differ chemically is doubtless
the essential thing in producing added vigor. Heterozygosis is mentioned
merely as an evidence of such chemical difference."

These quotations suffice to show that a distinction is held by
biologists at the present time between the effects of inbreeding
and cross breeding and of heredity in development, and they
believe that dominance of hereditary factors is inadequate to
account for the widespread, if not universal phenomenon of
heterosis. The reason why biologists in general have refused to
believe that dominance was in any way responsible for the in-
creased vigor of hybrids has been due to two objections which
have seemed to make this hypothesis untenable. They thought
that if hybrid vigor was due to the dominance of definitely in-
herited characters that all these favorable characters which bring
about heterosis could be easily recombined into a homozygous
individual which would show no reduction on subsequent inbreed-
ing. Since no clear case was known in maize where a plant did
not lose vigor on inbreeding this seemed to be a convincing argu-
ment. Another objection to dominance as a means of accounting
for heterosis was raised by Emerson and East ('13) in that the
distribution in F2 should be unsymmetrical in respect to those
characters in which heterosis was shown in Fi. Since the usual
frequency distributions in cases of this kind are symmetrical, this
objection appeared to be valid.

How both of these objections do not hold when linkage of hered-
ity factors is taken into consideration, the writer has attempted
to show in a recent publication ('17). Because of linkage, char-
acters tend to pass from one generation to the next in groups and
are not easily recombined. Furthermore, on account of linkage
skewness is not expected in the second hybrid generation. All of
the recently acquired knowledge of heredity makes it seem highly
probable that heterosis may be largely, if not entirely, accountable
on the basis of dominance of linked factors.

In considering these two hypotheses, both attempting to ac-
count for heterosis, the following facts about dominance should
be kept in mind:

1. Partial dominance of characters is a widespread occurrence
in plants and animals.

2. Dominance, of course, does not appear until after the zygote
is formed.

A -MENDELIAN INTERPRETATION OF HETEROSIS. 87

3. In most cases dominance does not change throughout the
life of the individual and remains the same through innumerable
clonal generations.

While none of these features of dominance offers any definite
means of proving the truth of the hypothesis advanced, is it only
a coincidence that they fit in exactly with what the facts of het-
erosis demand? It remains to show that those characters which
enable a plant or animal to obtain the best development are, for
the most part at least, partially dominant over those characters
which retard or prevent maximum growth.

The essential difference between the two hypotheses may be
stated briefly. According to the previous view the hybrid combi-
nation of factors Aa carried the ability to stimulate development
because of the union of unlike elements. This stimulation was
absent in either of the homozygous combinations AA and aa,
and this stimulation had no direct relation to the part that either
A or a had in development as hereditary entities. According to
the conception of dominance, first proposed by Keeble and
Pellew and carried out more fully by the writer, the hybrid union
of AAbb with aaBB, resulting in the heterozygous combination
of Aa Bb, increases development because two dominant characters
are present here together, whereas each parent has only one
dominant character. A similar factorial arrangement has been
proposed by Hyde ('14) to account for the increased fertility of
his crosses among partially sterile strains of Drosophila.

In crosses between different types of domesticated animals and
of cultivated plants it has frequently been noted that there is a
tendency towards a return to the characters of the wild species
from which they were derived. Sageret ('26) makes particular
note of this point. It is well known that crosses between different
breeds of pigeons is quite apt to bring back the wild-type of
plumage. The hybrid between radish and cabbage described by
Gravatt ('14) illustrates this point strikingly. The hybrid pro-
duced had neither a succulent "head" like its cultivated male
parent nor a fleshy root like its female parent. In other respects,
as well, it showed this return to wild-type characters. It was
also exceedingly vigorous, but sterile, like so many hybrids between
diverse stocks.

Drosophila furnishes the best illustration of the appearance of
wild type characters in the first hybrid generation. Of the more

88 CONNECTICUT EXPERIMENT STATION BULLETIN "207.

than one hundred mutations found in DroFophila by far the
largest number of these are recessive. Almost all of them are
characters which are less favorable to development. It is stated
that any attempt to collect a large number of the recessive char-
acters into one race is rendered difficult by the weakened consti-
tution of the flies possessing any great accumulation of recessive
characters (Muller, '16). Whenever crosses are made between
diverse types the first generation fly is in many of its characters
more like the wild stock and hence more vigorous than its parents.
All lethal factors, well illustrated in Drosophila, furnish additional
support to the hypothesis of dominance as a means of accounting
for heterosis. Muller ('17) has shown that a condition of "bal-
anced lethals" ma}' be brought about in which only the hetero-
zygotes can live. As dominant lethal factors are always eliminated
as soon as they occur, so, also, is there always a strong tendency
for selection to eliminate any dominant character which is at all
unfavorable to the organism's best development. Unfavorable
recessive factors also tend to be eliminated, but much more slowly.

If the results obtained in Drosophila are applicable to other
animals and to plants we must infer that recessive mutations
occur the most commonly. Hence recessive mutations make up
the characters, to a large degree, that man has selected in the
production of domesticated animals and plants. Just as in Dro-
sophila, crosses between diverse domesticated types tend to
result in the reappearance of wild-type characters which are more
useful to the plant or animal whose chief aim in life is, apparently,
to reproduce itself.

This is well shown in an illustration from maize. As stated
before, inbred strains have been obtained which are markedly
deficient in root development. On these plants the large brace
roots which commonly appear when the plants begin to need extra
support, are almost completely lacking. Consequently, the plants
are blown over when they become heavy at the time of ear for-
mation. I have observed these strains three years and each time
they have fallen down. This character is not determined by soil
conditions or insect damage or any external conditions as far as
can be seen. Other plants on either side are perfectly upright.
When these strains are crossed with other strains, inbred for an
equal or longer period, which have well developed brace roots,
the first hybrid generation has remarkably well developed brace

A MEN DELI AN INTERPRETATION OF HETEROSIS. 89

roots, and usually does not show the slightest tendency to go down,
as shown in Plates Xlla and b. Emerson ('12) describes similar
plants in which the root deficiency is also recessive. Another
striking feature is shown in this illustration. The inbred strain
which lacks brace roots is derived from a floury variety of corn
and shows a decided tendency to branch at the base of the stalk.
These branches form stalks with tassels and ears and many of
them are fully as well developed as the main stalk. In this way
two or three stalks may be developed from one seed. The other
parent of the cross shown never branches in this way and never
even develops small branches or "suckers." The first hybrid
generation shows this tendency to branch even more strongly
developed than the branching parent. The plants shown are from
three hills grown side by side and each hill is the product of three
seeds. Thus it will be seen that both parents have contributed
characters to the hybrid.' Both these characters are such as to
enable the plants to attain a greater development in general
vegetative luxuriance than would be possible if either were lacking.
Emerson ('12) gives an even better illustration of two extremely
unproductive types of maize which give a vigorous hybrid, one
of the parents contributing tall stature, the other green chlorophyll.

Many more illustrations of a similar operation of hereditary
factors favoring a hybrid in its development might be cited. I
believe that enough have been given to clear the way towards the
acceptance of the doctrine that hybrid vigor is due largely to the
normal functioning of definable, hereditary factors.

It is recognized that the characters used as illustrations here are
superficial in nature. The characters which are really concerned
in heterosis are those deep-seated, fundamental, physiological
processes which govern metabolism and cell-division. As to the
mode of inheritance of these characters we, as yet, know little.
There is no reason to believe, however, but that many or all of
them are Mendelian in mode of inheritance and that many of
them operate in the same way to enable hybrid progeny to
attain a more complete development than their parents. If this
hypothesis, as to the way in which heterosis is brought about, is
in its essential features correct, it points the way towards a more
fundamental application of Mendelism to the physiological
processes of growth than is generally acceded.

90 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

There now remains to be discussed the part that crossing-over
or breaks in the linkage of hereditary factors would playaccording
to this hypothesis. If any large number of characters are con-
cerned and the dominant and recessive genes are equally appor-
tioned between the two parents, and distributed at random on
the chromosomes, the chance of crossovers occurring in such a
way as to bring all the dominant factors in one individual at one
time would be almost inconceivably small, especially when the
phenomenon of interference is taken into consideration. How-
ever, when crossing-over does occur in such a way as to bring
about more fortunate combinations in certain individuals, those
individuals would be the ones selected by man in domesticated
races, or by nature in the wild. Partial linkage does not prevent
recombination but merely adds to the complexity of the process.
The chance of fortunate recombinations would be greater in the
more widely crossed animals and plants but such combinations
would be again broken up by further crossing. The tendency
would be, however, for the best combination of characters to
survive and gradually supplant the others in time. In naturally
selfed plants, most of which are crossed at more or less infrequent
intervals, a fortunate homozygous combination would be fixed
and the plants possessing such combinations would in time sup-
plant their less fortunate relations.

Thus there would always be the tendency for all the more
favorable characters to be gathered together and the others
eliminated. In time all the individuals of a locality would tend
to become equal in their hereditary characters and crossing
between individuals in a given locality would not accumulate
any greater number of favorable characters than the parents
possessed and hence would not show any evidences of heterosis.

That this is the condition which is brought about Darwin has
shown. Individuals from the same locality derive little or no
benefit from crossing while crosses between individuals from
different geographical regions show a greater effect of crossing.
The work of Collins ('10) and the results obtained at the Con-
necticut Station (Jones and Hayes '17) show this also — varieties
of maize of similar characters and from the same region give less
increase when crossed than do varieties of diverse type or from
widely separated geographical regions.

A MENDELIAN INTERPRETATION OF HETEROSIS. 91

If by crossing-over and subsequent recombination the charac-
ters which bring about the great development in Fi can all be
accumulated in a homozygous condition in an individual, that
individual should show a greater development even than the Fi
as A. F. Shull ('12) has pointed out. This is on the assumption
that most characters which play a part in heterosis are not fully
dominant. That a factor in the diploid condition has a greater
effect than when in the haploid condition is indicated by the
work of Hayes and East ('15) on endosperm characters in maize.
Their results show that in reciprocal crosses a double dose of
one allelomorph in the maternal endosperm fusion nucleus over-
comes a single dose in the paternal endosperm nucleus. In other
words factors have an accumulative effect.

The evidence that such superior individuals have been obtained
by inbreeding is not very convincing it must be admitted. Dar-
win, however, in Ipomea obtained plants — " Hero " and its
descendants — which were certainly no less vigorous than any
plants at the beginning of the inbreeding period and the same
thing occurred in Mimulus. These are the two species which
were the most extensively inbred. Miss King, as mentioned
before, has obtained inbred rats which are larger and more
vigorous than individuals present in the original stock. Nothing
of this kind has occurred in maize and on account of the small
chance of recombining many of the most desirable characters
in one plant, it is not at all surprising that such individuals have
not as yet been produced.

The production of individuals by inbreeding which excel any
of the original crossbred stock offers some means of deciding
between the two hypothesis attempting to account for heterosis.
According to the 'hypothesis of a physiological stimulation it
would be difficult to see how individuals more vigorous than
the parents could be produced by inbreeding.

The hypothesis of dominance also, possibly, makes it easier to
understand why naturally crossed wild species, which have not
been outcrossed with fresh stocks for long periods of time, may
not show any markedly injurious effects from artificial inbreeding.
According to the former view different characters of equal value
to the organism which might persist indefinitely in a species,
would supply a stimulation when united in a heterozygous com-
bination. This stimulation would be lost whenever individuals

92 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

were reduced to homozygosity by artificial self-fertilization.
According to the view of dominance if the allelomorphs were all
equal in their contributions to development there might be
differences in a species and still no loss of vigor would result
from inbreeding. It is assumed that the less favorable characters
have been eliminated by selection. On either hypothesis there
would be no reduction from inbreeding if all the members of
species were exactly alike whether they are naturally crossed
or naturally self-fertilized.

The hypothesis of physiological stimulation also implies the
assumption that naturally crossed species of cultivated plants
are inherently more efficient as producers than naturally selfed
plants. This is hardly justified when we recall such vigorous
and productive plants as wheat, oats, barley, rice, peas, beans,
tobacco, tomatoes and many others which are usually self pol-
linated. It is, however, difficult to make a fair comparison on
this basis.

To sum up, it may then be stated briefly that dominance of
characters as opposed to the former idea of an indefinable physio-
logical stimulation makes more understandable the facts that:

1. Heterozygosis produces a stimulating, and not an indifferent
or depressing effect in crosses between related stocks and that
the reverse is true in widely diverse stocks.

2. Heterozygosis operates throughout the lifetime of the
individual even through many generations of vegetative propa-
gation.

3. Inbreeding may result in individuals more vigorous than the
original cross-bred stock.

4. Inbreeding may not bring about a reduction in some naturally
crossed wild species.

Whether or not dominance of factors is wholly adequate to
account for all of the immediate effects of exogamy remains to
be seen. The former view that dominance was not concerned
at all has been maintained so insistently that I have taken the
extremely opposite view in order to show that dominance at least
can be held responsible for a large part of the increased develop-
ment shown by hybrids. The treatment of the subject in this
light has been dogmatic. That cross-fertilization may produce
some effect which can never be attained in self-fertilization or a*-
sexual reproduction is still possible. The view of the problem

HETEROSIS AND THE ESTABLISHMENT OF SEX. 93

which is presented here makes certain heretofore indefinite effects
more intelligible. It is not meant to preclude entirely any bene-
ficial physiological stimulation resulting from germinal diversity,
if such an effect can be demonstrated.

The difference between the two hypotheses are not as great as
might seem at first sight. The older hypothesis is general in its
application and does not commit itself to the interpretation of
specific effects. The view presented here is specific in its applica-
tion and may be shown to be inadequate for the interpretation
of all phases of the problem of increased development following
cross-fertilization.

The greatest progress in our knowledge of inbreeding and cross-
breeding was made when their effects were linked with Mendelian
phenomena. This was the big step forward. The two ways of
interpreting these effects discussed here, differ only in minor
features and it is not putting the matter fairly to hold them up as
two rival hypotheses, one to be chosen from the other. Placing
the effects of inbreeding and cross-breeding entirely on a Mendelian
basis is merely the logical outgrowth of the older view as knowledge
of the methods of inheritance increased.

THE PART THAT HETEROSIS HAS PLAYED EN THE ESTABLISHMENT

OF SEX.

Since heterosis is widespread in its manifestation it can hardly
be doubted that it has played some part in the initiation and
maintenance of sexual differentiation in organisms. Jennings
('13), however, has shown that conjugation in Paramecium does
not result, immediately, to the advantage of the organism. The
rate of reproduction is actually diminished and many of the
organisms perish. The advantage which is derived from con-
jugation, he considers with Weismann, is due to the fact that
biparental inheritance makes possible a greater .variability and
consequently a greater chance of recombinations, some of which
are better able to persist. Hence, while many offspring from
conjugating paramecia die, some may be able to survive.
Conjugation therefore makes possible a greater elasticity in
adaptiveness to new and varied surroundings. '

If this immediately depressing effect found in Paramecium is
general in the lower animals, heterosis would probably have

94 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

played no part in the inauguration of sex. Both A. F. Shull
('12a) and Whitney ('12a) have shown, however, that heterosis
occurs in the rotifer, Hydatina senta.

In the lower plants heterosis would have significance only in
spore formation, as the main -life of the plant is carried on in the
haploid condition where heterozygosis could not, of course,
operate. As organisms became more differentiated and specialized
the accumulation of factors in the zygote from two somewhat
different parents would have increasing significance. If, for
example, an organism should vary in a character A by one new
dominant mutation A', the heterozygote AA', according to the
hypothesis of dominance, would be superior to the combination
A A but not to the combination A' A'. According to the former
view of a physiological stimulation the heterozygous combination
A A' might be superior to either homozygous combination. The
matter is not so simple as this, however. The breeding facts
show that recessive unfavorable variations are far more common
than dominant favorable ones. The chances would be that
those individuals which varied by dominant mutations would
also vary from the parental stock, sooner or later, by recessive
mutations as well, so that any hybrid union would tend to accu-
mulate more favorable factors than either parental individual
possessed and hence show heterosis. Heterosis would be an
immedate factor for natural selection to work upon.

Moreover it seems possible that heterosis has had considerable
to do with the rise of the sporophyte and the decline of the game-
tophyte in plants. Recombination of characters can take place
as well when the dominant generation is the haploid as well as
when it is diploid in respect to the chromosome arrangement.
From the standpoint of adaptiveness through recombination
of characters it might even be to the organism's advantage to
retain the haploid generation as the one in which the principal
life processes were carried on, since the different combinations
would then be more surely tested and the best more easily
selected in the simplex than in the duplex condition. Heterosis
can only operate in the sporophyte. The union of different
hereditary complexes gives to the sporophyte an advantage over
the gametophyte in that all new favorable variations work to-
gether whereas segregation in the formation of the gametophyte
reduces the efficiency of this generation. On the basis of the

HETEROSIS AND THE ESTABLISHMENT OF SEX. 95

complimentary action of factors according to the dominance
hypothesis of heterosis the gametophyte would practically always
be at a disadvantage as compared to the sporophyte as long as
variations were occurring so that heterosis must have played
some part in these important changes.

Either on the basis of inducing variability or stimulating
development, sex would be a creation of no value to organisms
which are never cross-fertilized. It may be questioned if many
such exist. In either case the sexual mechanism is so complex
and deep-seated in the life of the organism that it is not to be
discarded easily. Whenever the best possible combination of
factors for a given environment is produced, it is to the advantage
of the organism possessing that combination to give up cross-
fertilization and resort to either self-fertilization or some form
of^sexual reproduction, for the reason that these are more efficient
means of propagation. When the environment changes, those
organisms which are not cross-fertilized may either be doomed
to extinction or handicapped in becoming adapted to new con-
ditions and the perpetuation of the sexual mechanism thereby
accounted for.

Whatever may be the value or significance of heterosis, to
account for this phenomenon it is, for the most part, unnecessary
to assume that there is an indefinite stimulating effect of hybrid-
ization along with the expression of definable hereditary factors.
Hence the distinction is no longer needed between the effects of
self-ferti ization and cross-fertilization and of heredity in develop-
ment. The heretofore indefinite physiological stimulation re-
sulting from heterozygosis and the related effects accompanying
the loss of this stimulation following inbreeding can therefore
be given a strictly Mendelian interpretation.

This being so there is no longer a question as to whether or not
inbreeding per se is injurious. Whether good or bad results from
inbreeding depends solely on the constitution of the organisms
before inbreeding is commenced. Inbreeding is concerned only
with the manifestation of conditions pre-existing. As a means of
analyzing and of purifying a cross-bred stock by the elimination
of undesirable qualities, inbreeding is therefore a method of first
importance in plant and animal improvement.

96 CONNECTICUT EXPERIMENT STATION BULLETIN 207.

LITERATURE CITED.

Andronescu, D. I., 1915. The physiology of the pollen of Zea Mays with
special regard to vitality. Thesis, Univer. of Illinois.

Beal, W. J., 1876-1882. Reports, Michigan Board of Agriculture, 876,
1877, 1881 and 1882.

Biffen, R. H., 1912. Studies in the inheritance of disease resistance.
Jour. Agr. Science 4:421-429.

Bonhote, J. L., 1915. Vigour and heredity, pp. xii+263. London:
West, Newman and Co.

Britton, E. G., 1898. A hybrid moss. Plant World 1: 138.

Bruce, A. B., 1910. A Mendelian theory of heredity and the augmenta-
tion of vigor. Science, N. S. 32: 627-628.

1917. In-breeding. Jour. Gen. 6: 195-200.

Burck, W., 1908. Darwin's Kreuzungsgesetz und die Grundlagen der
Bliitenbiologie. Rec. Trav. Bot. Neerl. 4: 17-118.

Carrier, L., 1913. The immediate effect on yield of crossing strains of
corn. Virginia Agr. Exp. Sta. Bull. 202, pp. 11.

Castle, W. E., 1916. Genetics and eugenics, pp. vi+353. Cambridge:
Harvard Univer. Press.

Castle, W. E., Carpenter, F. W.. Clark, A. H., Mast, S: O., and Barrows,
W. M., 1906. The effects of inbreeding, cross-breeding and selec-
tion upon the fertility and variability of Drosophila. Proc. Am.
Acad, of Arts and Sciences 41: 731-786.

Collins, G. N., 1910. The value of first generation hybrids in corn. U. S.
Dept. of Agr., Bur. of P. I. Bull. 191.
1910. Increased yields of corn from hybrid seed. Yearbook, U. S.

Dept. of Agr., 1910, pp. 319-328.
1914. A more accurate method of comparing first generation hybrids
with their parents. Jour. Agric. Research 3: 85-91.

1918. Maize: its origin and relationships. Notes of the 123d
regular meeting of the Botanical Society of Washington by H. N.
Vinall. Jour. Wash. Acad, of Sciences 8: 42-43.

Collins, G. N., and Kempton, J. H., 1913. Effects of cross-pollination

on the size of seed in maize. U. S. Dept. of Agr., Cir. 124.
Crampe, H., 1883. Zucht-Versuche mit zahmen Wanderratten. Land-

wirtsch. Jahrbiicher 12: 389-458.
Darwin, C, 1875. The variation of animals and plants under domestica-
tion. 2 vols. 1898, New York: Appleton and Co.
1877. The effects of cross- and self-fertilization in the vegetable
kingdom, pp. viii+482. London: Appleton and Co.
East, E. M., 1908a. Inbreeding in corn. Connecticut Agr. Exp. Sta.
Report for 1907. pp. 419-428.
1908b A study of the factors influencing the improvement of the

potato. Illinois Agr. Exp. Sta. Bull. 127. pp. 375-456.
1909. The distinction between development and heredity in in-
breeding. Am. Nat. 43: 173-181.
1910a. A Mendelian interpretation of variation that is apparently
continuous. Am. Nat. 44: 65-82.

LITERATURE CITED.

East, E. M., 1910b. Inheritance in potatoes. Am. Nat. 44: 424-430.

1915. The chromosome view of heredity and its meaning to plant
breeders. Am. Nat. 49: 457-494.

1916. Studies on size inheritance in Nicotiana. Genetics 1: 164-176.
East, E. M., and Hayes, H. K., 1912. Heterozygosis in evolution and in

plant breeding. U. S. Dept. of Agr., Bur. of P. I. Bull. 243, pp 58.

East, E. M., and Park, J. B., 1917. Studies on self-sterility I. The be-
havior of self-sterile plants. Genetics 2: 505-609.

Emerson, R. A., 1912. The inheritance of certain "abnormalities" in
maize. Report Am. Breeders' Assoc. 8: 385-399.

Emerson, R. A., and East, E. M., 1913. The inheritance of quantitative
characters in maize. Nebraska Agr. Exp. Sta. Research Bull. 2,
pp. 120.

Enriques, P., 1907. La coniugazione e il differenziamento negli Infusoria.

Archiv. f. Protistenkiinde 9:195-296.
Fabre-Domengue, P., 1898. Unions consanguines chez les Colombins.

L'Intermediarie des Biologistes 1: 203.
Fischer, E., 1913. Die Rehobother Bastards und das Bastardierungs-

problem beim Menschen. Jena. Reviewed in Jour. Her., 1914,

5: 465-468.

Focke, W. O., 1881. Die Pflanzen-Mischlinge. pp. iv+569. Berlin:
Borntraeger.

Gartner, C. F., 1849. Versuche und Beobachtungen liber die Bastarder-
zeugung im Pflanzenreich, pp. xvi+791." Stuttgart: Hering &
Comp.

Gernert, W. B., 1917. Aphis immunity of teosinte-corn hybrids. Science,

N. S. 46: 390-392.
Gerschler, M. W., 1914. tiber alternative Vererbung bei Kreuzung von

Cyprinodontiden-Gattungen. Zeitschr. f. ind. Abst. u. Vererb

12: 73-96.

Gravatt, F., 1914. A radish-cabbage hybrid. Jour. Her. 5: 269-272.

Gray, A., 1876. Darwiniana: essays and reviews pertaining to Dar-
winism. xii+396. New York: Appleton and Co.

Guaita, G. von, 1898. Versuche mit Kreuzungen von verschiedenen
Rassen der Hausmaus. Ber. naturf. Gesellsch. zu Freiburg 10:
317-332.

Hartley, C. P., Brown, E. B., Kyle, C. H., and Zook, L. L., 1912. Cross-
breeding corn. U. S. Dept. of Agr., Bur. of P. I. Bull. 218.

Hayes, H. K., 1914. Corn improvement in Connecticut. Connecticut
Agr. Exp. Sta., Part vi of the Annual Report for 1913. pp. 353-384.

Hayes, H. K., and East, E. M., 1911. Improvement in corn. Connecticut
Agr. Exp. Sta. Bull. 168, pp. 21.
1915. Further experiments on inheritance in maize. Connecticut
Agr. Exp. Sta. Bull. 188, pp. 31.

Hayes, H. K., and Jones, D. F., 1917. The effects of cross- and self-
fertilization in tomatoes. Connecticut Agr. Exp. Sta., Part v of
the Annual Report for 1916, pp. 305-318.

CONNECTICUT EXPERIMENT STATION BULLETIN 207.

Hedrick, U. P., and Howe, G. H., 1913. Apples: old and new. New

York Agr. Exp. Sta. Bull. 361, pp. 79-135.
Huth, A. H., 1875, The marriage of near kin, pp. lxvii-f-359. London:

Churchill.

Hyde, R. H., 1914. Fertility and sterility in Drosophila ampelophila.

i-iv, Jour. Exper. Zool., Vol. 17.
Jennings, H. S., 1912a. Age, death and conjugation in the light of work
on lower organisms. Pop. Science Monthly 80: 563-577.
1912b. The production of pure homozygotic organisms from hetero-

zygotes by self-fertilization. Am. Nat. 46: 487-491.
1913. The effect of conjugation in Paramecium. Jour. Exper.
Zool. 14: 279-391.

1916. The numerical results of diverse systems of breeding. Genetics
1: 53-89.

Jones, D. F., 1917. Dominence of linked factors as a means of accounting
for heterosis. Proc. Nat. Acad, of Sciences 3: 310-312 and also
Genetics 2: 466-479.
1918. Bearing of heterosis upon double fertilization. Bot. Gaz.
65: 324-333.

Jones, D. F., and Hayes, H. K., 1917. Increasing the yield of corn by
crossing. Connecticut Agr. Exp. Sta., Part v of the Annual
Report for 1916, pp. 323-347.

1917. The purification of soy bean varieties. Same, pp. 348-353.
Jones, L. R., 1918. Disease resistance in cabbage. Proc. of the Nat.

Acad, of Sciences 4: 42-46.
Keeble, F., and Pellew, C, 1910. The mode of inheritance of stature
and of time of flowering in peas (Pisum sativum). Jour. Gen.
1: 47-56.

Kempton, J. H., 1913. Flower abnormalities in maize. U. S. Dept. of

Agr., Bur. of P. I. Bull. 278, pp. 16.
King, H. D., 1916. (Experiments of Miss King on inbreeding rats for

22 generations discussed in Jour. Her. 7: 70-76.)
Knight. T. A., 1799. An account of some experiments on the fecundation

of vegetables. Philosophical Transactions of the Royal Society

of London 89: 195-204.
1841 A selection from the physiological and horticultural papers

published in the Transactions of the Royal and Horticultural

Societies.

Kraemer, H., 1913. liber die ungiinstigen wirkungen naher Insucht.
Mitteilung der Deutsches Landwirtschafts Gesellschaft, 6 and 13
Sept., 1913. Translated in Jour. Her. 5: 226-234.

Kolreuter, J. G., 1766. Dritte Fortsetzung der Vorlaufigen Nachricht
von einigen das Geschlecht der Pflanzen betreffenden Versuchen
und Beobachtungen. Gleditschen Handlung. Reprinted 1893 in
Ostwald's Klassiker der Exakten Wissenschaften, No. 41. Leipzig:
Engelmann.

Lindley, J., 1852. * The theory of horticulture, pp. 20-364. 2d Amer.
Edition by A. J. Downing, New York: Wiley.

LITERATURE CITED.

Mendel, G., 1865. Experiments in plant-hybridisation. Translated in
Castle's Genetics and Eugenics, 1916, pp. vi+353. Cambridge:
Harvard Univ. Press.

McCluer, G. W., 1892 Corn crossing. Illinois Agr. Exp. Sta. Bull. 21,
pp. 82-101.

Moenkhaus, W. J., 1911. The effects of inbreeding and selection on the

fertility, vigor and sex-ratio of Drosophila ampelophila. Jour.

Morph. 22: 123-154.
Montgomery, E. G., 1912. Preliminary report on effect of close and

broad breeding on productiveness in maize. Nebraska Agr. Exp.

Sta., 25th Annual Report, pp. 181-192.
Morrow, G. E., and Gardener, F. D., 1893. Field experiments with corn.

Illinois Agr. Exp. Sta. Bull. 25, pp. 173-203.
1894. Experiments with corn. Illinois Agr. Exp. Sta. Bull. 31, pp.

359-360.

Muller, H. J., 1916. The mechanism of crossing-over III. Am. Nat.
50: 350-366.

1917. An Oenothera-like case in Drosophila. Proc. Am. Acad, of
Sciences 3: 619-626.
Naudin, C, 1865. Nouvelles recherches sur l'hybridite dans les vegetaux.
Nouvelles Archives du Museum d'Histoire Naturelle de Paris.
1: 25-174.

Pearl, R., 1915. Modes of research in genetics, pp. vi + 182. New York:
Macmillan.

Roberts, H. F., 1912. First generation hybrids of American and Chinese
corn. Report Am. Breeder's Assoc. 8: 367-384.

Riley, E. H., 1910. A new zebra hybrid. Am. Breeder's Magazine 1:
107-110.

Ritzema Bos, (J.), 1894. Unterschungen iiber die Folgen der Zucht in
engster Blutsverwandtschaft. Die Biologisches Centrallblatt
14: 75-81

Sageret, A., 1826. Considerations sur la production des hybrides, des
variantes et des varietes en general, et sur celles de la famille des
Cucurbitacees en particulier. Annales des Sciences Naturelles
8: 294-314.

(Sargent), C. S., 1894. A hybrid walnut tree. Garden and Forest 7:
434-436.

Shull, G. H., 1908. The composition of a field of maize. Report Am.
Breeder's Assoc. 4: 296-301.

1909. A pure line method of corn breeding. Same 5: 51-59.

1910. Hybridization methods in corn breeding. Am. Breeder's
Magazine 1:98-107.

1911. The genotypes of maize. Am. Nat. 45: 234-252.

1914. Duplicate genes for capsule form in Bursa bursa-pastoris.

Zeitschr. f. ind. Abst. u. Vererb. 12:97-149.
Shull, A. F., 1912a. The influence of inbreeding on vigor in Hydatina

Senta. Biol. Bull. 24: 1-13.
1912b. Studies in the life cycle of Hydatina Senta III. Jour. Exp.

Zool. 12: 283-317.

100 CONNECTICUT EXPEEIMENT STATION BULLETIN 207.

Spillman, W. J., 1909. Application of some of the principles of heredity

to plant breeding. U. S. Dept. of Agr., Bur. of P. I. Bull. 165, pp. 74.
Stout, A. B., 1916. Self- and cross-pollinations in Chichorium intybus

with reference to sterility. Mem. N. Y. Bot. Gard. 6: 333-454.
Swingle, W. T., and Webber, H. J., 1897. Hybrids and their utilization

in plant breeding. Yearbook, U. S. Dept. of Agr. for 1897, pp.

383-420.

Tupper, W. W., and Bartlett, H. H., 1916. A comparison of the wood
structure of Oenothera stenemeres and its tetraploid mutation gigas.
Genetics 1: 177-184.

Van Mons, J. B., 1836. Arbres fruitiers ou pomologie Beige Experi-
mental et Raisonnee. Vol ii. Louvain.

Weismann, A., 1904. The evolution theory. 2 Vols. (Translated by
J. A. Thomson and M. R. Thomson) London: Arnold.

Wellington, R., 1912. Influence of crossing in increasing the yield of the
tomato. New York Agr. Exp. Sta. Bull. 346, pp. 57-76.

Weston, W. H., 1918. The downy mildews of maize, their origin and
distribution. Notes of the 123d regular meeting of the Botanical
Society of Washington by H. N. Vinall. Jour. Wash. Acad, of
Sciences 8: 43.

Whitney, D. D., 1912a. Reinvigoration produced by cross-fertilization

in Hydatina Senta. Jour. Exp. Zool. 12: 337-362.
1912b. "Strains" in Hydatina. Biol. Bull. 22: 205-218.
1912c. Weak parthenogenetic races of Hydatina Senta subjected to

a varied environment. Biol. Bull. 23:321-330.
Wichura, M., 1865. Die Bastardbefruchtung im Pflanzenreich erlautert

an den Bastarden der Weiden. pp. iv+95. Breslau: Morganstern.
W^iegmann, A. F., 1828. Ueber die Bastarderzeugung im Pflanzenreich.

pp. 40. Braunschweig.
Woodruff, L. L., 1911. Two thousand generations of Paramoecia.

Archiv. f . Protistenkunde 21 : 263-266.
Wolfe, T. K., 1915. Further evidence of the immediate effect of crossing

varieties of corn on the size of seed produced. Jour. Am. Soc. of

Agr. 7: 265-272.

PLATE I.

a. A non-inbred variety of Learning dent corn.

Four inbred strains derived from the Learning variety after nine gen-
erations of self-fertilization showing an ear, a cob and
a cross-section of a cob of each.

PLATE II.

PLATE III.

a Representative ears of inbred strain No. 1-7-1-2, etc.

b. The first generation cross of inbred strain No. 1-6-1-3 by 1-7-1-2.

(Plates I to III inclusive, with the exception of lb, are on the same scale). The plants
which produced these ears were all grown on the same field and the non-inbred variety and
the first generation cross were grown in adjoining rows. The ears of these latter two represent
the best ears produced by 60 plants of each.

14*

PLATE IV.

b. Representative plants of the inbred strain No. 1-6-1-3, etc.

PLATE V.

a. Representative plants of the inbred strain No. 1-7-1-2, etc.

b. Representative plants of the first generation cross of inbred strain No.

1-6-1-3 by 1-7-1-2.

(Plates IV and V are on the same scale.)

PLATE VI.

a. Two fully developed tassels on the left and two partially sterile tassels on
the right characteristic of four different inbred strains of maize. From left to
right they are, 20A-8-5-10; 1-9-1-2; 1-6-1-3; 21-3-13-9.

b. Representative ears from the corresponding strains shown in the illus-
tration above. The first strain on the left produces fully developed tassels
and moderately developed ears. The second produces the best ) developed
tassels and the poorest ears. The other two have poorly developed tassels
and moderately well developed ears.

PLATE VII.

a. Two inbred strains of dent corn, Xo. 1-6-1-3 at the right and No.
1-7-1-1 at the left and the first generation cross in the center. The three
ears were grown under equal conditions and gathered on the same day to
show differences in maturity.

» ' *a *^

ggssisfl

b. Two inbred strains of dent corn, Xo. 1-7-1-2 at the right and No.
1-6-1-3 at the left and the first generation cross in the center showing the
differences in maturity.

PLATE VIII.

PLATE IX.

a. Seeds of two inbred strains of corn and the seeds produced upon the
first generation hybrid plant in the center. The second generation plants
grown from these large seeds have an advantpge over either the parents or
the first generation hybrid.

b. Two inbred strains and their first and second generation hybrids.
From right to left they are: inbred strain Xo. 1-9-1-2, Xo. 1-7-1-1, (1-9x1-7)
F2 and Fi.

PLATE X.

a. The same two inbred strains and their first and second generation hybrids
as in IX b. From right to left they are: inbred strain No. 1-9-1-2, No. 1-7-1-1,
(1-9 x 1-7) F2 and Fi.

M \ ! 1 I I * 'i i i 1 v 1 I M i i l i f I i. i L]

b. Same as above — ten plants of each.

PLATE XI.

a. Selfed, reciprocally crossed and out-crossed seeds obtained by pollinating
plants of three different strains with a mixture of yellow and white-carrying
pollen from the plants which bore the two ears shown below, showing the
ratio and distribution of the two different kinds of seeds produced on each
ear.

(The seeds resulting from the "yellow" pollen were colored by hand on all three ears.)

b. Seedlings showing the rate of growth and the amount of germination
of selfed and crossed seeds from the same ears from five different plants.

PLATE XII.

a. The first generation cross of an inbred strain which lacks brace roots
but has the habit of branching freely from the base of the plant (shown
at the right) with an inbred strain (shown at the left) which has well de-
veloped brace roots but does not branch at the base. The three lots of
plantsjfhave resulted from three seeds each.

A closer view of_the roots of the plants shown in the above illustration.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

GEO. F. FREEMAN
Socictc Sultanienne d' Agriculture, Cairo, Egypt

Submitted to the Faculty of the Bussey Institution of
Harvard University, in partial fulfillment of the require-
ments for the degree of Doctor of Science, April 30, 1917.

Reprinted from Genetics 4 : 1-93, January, 1919

GENETICS

A Periodical Record of Investigations Bearing on
Heredity and Variation

Editorial Board

George H. Shull, Managing Editor
Princeton University

vVilliam E. Castle
Harvard University

Edward M. East

Harvard University

Edwin G. Conklin

Princeton University

Rollins A. Emerson
Cornell University

Charles B. Davenport

Carnegie Institution of Washington

Herbert S. Jennings
Johns Hopkins University

Bradley M. Davis

University of Pennsylvania

Thomas H. Morgan

Columbia University

Raymond Pearl

Johns Hopkins University

Genetics is a bi-monthly journal issued in annual volumes of about
600 pages each. It will be sent to subscribers in the United States, the
Philippines, Porto Rico, etc., for $6 per annum for the current volume,
and $7 per volume for completed volumes until the edition is exhausted.
Canadian subscribers should add 25 cents for postage. To all other
countries 50 cents should be added for postage. Single copies will be
sent to any address postpaid for $1.25 each.

Entered as second-class matter February 23, 19 16, at the post office at
Princeton, N. J., under the act of March 3, 1879.

THE HEREDITY OF QUANTITATIVE CHARACTERS

IN WHEAT

GEO. F. FREEMAN
Societe Sultanienne d' Agriculture, Cairo, Egypt

[Received May 15, 1918]
TABLE OF CONTENTS

PAGE

Introduction i

Material and methods 8

Algerian macaroni (No. i) 8

Algerian red bread (No. 3) 8

Early Baart (No. 34) 8

Sonora (No. 35) 8

Date of first head 14

Macaroni X bread wheat crosses. Algerian macaroni (Xo. 1) X Sonora

(No. 35) 14

Bread wheat crosses. Red Algerian bread (No. 3) X early Baart (No.

34) 22

Summary; date of first head 27

Height 31

Macaroni — bread wheat crosses. Algerian macaroni (No. 1) X Sonora

(No. 35) 3i

Algerian macaroni (No. 1) X Algerian red bread (No. 3) JJB

Height in bread wheat crosses, 3 X 35 43

Red Algerian bread (No. 3) X early Baart (No. 34) 46

Summary; height 52

Width of leaf 57

Macaroni (No. 1) X Sonora (No. 35) 57

Algerian macaroni (No. 1) X Algerian red bread (No. 3) 68

Inheritance of leaf width in bread wheat crosses, Sonora (No. 35) X

red Algerian bread wheat (No. 3) 77

Algerian red bread (No. 3) X early Baart (No. 34) 81

Summary; width of leaf 87

General summary 92

Literature cited 93

INTRODUCTION

This paper forms a report on certain phases of a series of investiga-
tions in wheat breeding under the supervision of the writer, in the De-
partment of Plant Breeding of the Arizona Agricultural Experi-
ment Station. The work was initiated by the making of a number of

Genetics 4 : 1 Ja 1919

GEO. F. FREEMAN

reciprocal crosses between an Algerian white macaroni wheat, an Alge-
rian red bread wheat and two local white bread wheats, Early Baart and
Sonora. The original hybridizations were made at Yuma, Arizona, in
the spring of 191 3, the Fx was grown at Tucson in 191 3-' 14 and the F2
and F3 on the experimental farm at Yuma in 1915, 191 6, respectively.
The data concerning time relations, width of leaf, height, rust resistance,
etc., were, of course, taken in the field. At the time of ripening, the
heads of each plant were harvested and placed together in a paper bag,
care being taken to label each bag so that it could be completely identi-
fied. All other data were taken in the laboratory of the Department of
Plant Breeding at the University of Arizona at Tucson. The summa-
tion and analysis of this data begun some months earlier, has been con-
tinued throughout the present year by the writer while on sabbatical
leave from the University of Arizona. The writer here wishes ex-
pressly to thank the officers and management of the Bussey Institution
for laboratory and library facilities throughout the year and especially
Dr. E. M. East for many valuable criticisms and suggestions. He also
wishes to recall with appreciation the assistance rendered by Mr. Don-
ald F. Jones who made the original crosses, by Mr. Leonhardt
Swingle to whose careful and accurate work may be credited a large
proportion of the field and laboratory notes of the second generation,
and finally, by Mr. W. E. Bryan in his efficient assistance with the field
and laboratory notes for the third generation.

Since the re-discovery and publication of Mendel's original papers,
the question of paramount interest among geneticists and plant and
animal breeders has been that as to whether or not the principles in-
volved in the discoveries of Mendel are of limited or universal appli-
cation. Practically all real progress in the study of heredity has arisen
through experiments and observations designed to test the validity and
universality of Mendel's laws.

At the present time, the inheritance of a large number of characters,
including those both of a qualitative and quantitative nature, in a wide
series of both plants and animals, are almost universally considered to
be best explained by the Mendelian hypothesis. These include all char-
acters which in the F2 and subsequent generations, show definite, discon-
tinuous segregation. Most of the cases of peculiar and unusual ratios
have been satisfactorily explained as due to multiple factors, lethal fac-
tors, gametic coupling, gametic selection, partial sterility, etc.

There are cases, however, which admit of explanation by hypotheses
other than those based upon Mendelian principles. Examples may be

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 3

cited among characters which may be expressed quantitatively. In many
such cases the Fj is more or less intermediate between the parents, and
the F2 and subsequent generations show segregation, but such segrega-
tion as does occur is perfectly continuous. Where a sufficiently large
number of variants are grown, there is found every degree of size from
the lowest to the highest extreme of the hybrid distribution. The ex-
tremes of this distribution may or may not reach or extend beyond the
extremes of the parental races.

There are some geneticists who believe that such a type of inheritance
is not Mendelian. They advocate the application of the Mendelian prin-
ciples in many cases, but maintain that we have no proof that Mendelism
is universal and that cases such as those described above may be just as
easily explained by assumptions other than those of gametic purity and
unchanged segregation.

The literature on the subject of the inheritance of quantitative char-
acters has been collected by Shull (1914) and MacDowell (1914).
and has been summarized with excellent clearness by these writers. It
is therefore not necessary to re-summarize these earlier papers. The
results of original research bearing upon the inheritance of quantitative
characters which have appeared since Shull's and MacDowell's sum-
maries may now be reviewed briefly.

Nilsson-Ehle (1914) shows a genetic linkage between a factor for
yellow glume color and an inhibitor which shortens beard length in oats.

Phillips (1914) crossed Rouen and Mallard ducks which differ
greatly in size. The F1 was intermediate in size between the parents and
not more variable than the most variable parent. The F2, while still in-
termediate in average size, was markedly more variable than either the
F1 or the parents.

Punnett and Bailey (1914) in crosses of bantam with larger breeds
of fowl found the Fx intermediate and the F2 highly variable, transgress-
ing the extremes of both parents. Small F2 fowl bred together gave
an F3 all of small size; large F2 individuals bred inter se produced alto-
gether large offspring. The F3 obtained by mating intermediate F2 in-
dividuals was highly variable. They interpret the results as being due
to the segregation of Mendelian unit factors and give a factorial scheme
to account for the phenomena observed.

Hayes and East (191 5) crossed flour corn with a flint variety and
found that the endosperm character was determined by the mother only,
although it was proved that endosperm character, first visible in the next
generation could be inherited through the pollen. The authors conclude

Genetics 4: Ja 1919

GEO. F. FREEMAN

that this behavior is due to the fact that the endosperm is produced from
a union between two female polar nuclei and one male cell and that the
presence of two factors dominates one in either the direction of starchy
or flinty endosperm. In other flint-starchy crosses, the ratios were not
so definite, due possibly to the difficulty of classifying the seed. It was
thought, however, that the same principles were involved as in the pre-
vious crosses. Crosses involving grains of different shape were made
between rice pop corn, pearl pop corn and a dent corn. The results of
these experiments indicated that several factors were involved which
segregated in a Mendelian fashion in the F2 and F3. Parental types
when once recovered bred true.

East (1916 a) records the crossing of Nicotiana Langsdorffii and
N. alata which differ markedly in corolla length. The Fx was inter-
mediate and no more variable than the more variable parent. The F2
also had an intermediate average but the variability was much higher
than in the F±. There was a wide range in the variability of the dif-
ferent F3 races but they were all lower than in F2. He showed by
F3 pedigrees that segregation had occurred in F2 but did not attempt to
determine the number of factors.

East (1916 b) in a second paper reports the results of crossing a
variety of Nicotiana longi flora having the corolla about 93 mm long
with another variety of the same species having a corolla length of
about 40 mm. He carried the study through the first, second, third, and
in a few races as far as the fourth generation, with sufficient numbers
to calculate the coefficients of variation in the separate races. The
author lays down eight conditions which he assumes the data must ful-
fill in order to be interpreted as complying with the conditions of Men-
delian inheritance. Tables and distributions with the calculated con-
stants are given in detail and the conclusions are that no single phenome-
non has occurred which cannot be interpreted as Mendelian.

Phillips (191 5) after a study of the results of color inheritance in
various duck crosses and pheasant crosses says that "it is almost certain
that the ordinary subspecies of the ornithologist is very far from being
a unit variation."

Since the work of Johannsen on the effect of selection in beans, there
has been no similar work with plants which can compare in volume and
significance with that of Fruwirth (1915). Fruwirth followed the
system of pure line selection as practiced by Johannsen. Choosing a
variety of Lens esculent a with flecked seed, he endeavored through se-
lection to bring about greater flecking on the one hand and the diminu-

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

tion of the flecks on the other. After 13 generations he had made no
progress in either direction. Chevrier beans (Phaseolus vulgaris) pro-
duce seeds which, for the most part, have seed coats of a slightly green-
ish color rather than creamy white but a few seeds are white on one or
both sides. It was attempted, through selection within a pure line, to
secure complete inheritance of the green type. Though carried out for
14 generations no change was produced. In a race of vetch which pro-
duced both green and cream-colored seeds on the same plant, he tried
for 10 generations to fix the green coloration by selection but made no
progress. Likewise two years selection of yellow seed made no progress
in the direction of fixing the type. In a Victoria pea variety with yellow-
ish green and yellow seed three years of selection was without effect.
In a variety of Soja bean having lighter and darker brown seed, three
years of selection could make no progress in either direction of darker
or lighter seed coats. In a certain variety of Pisum arvense the seeds
are variable in color. They may be pure yellowish green, or yellowish
green with violet flecks or bands, or the violet color may be so extended
as to leave the yellowish green appearing only as flecks, or finally the
violet color may prevail altogether. Fruwirth endeavored by selec-
tion to increase the amount of violet color in the seeds on the one hand
and to reduce it on the other. In the selection for more violet color in
the seed coats, 10 generations produced no results. The results of the
selection in the opposite direction can best be given in Fruwirth's own
words as follows (Fruwirth 191 5, p. 200) :

"In beiden Johannsen'schen Linien 1 und A ist die Anlage zur
Ausbildung violette Farbe der Samenschale vorhanden, die Anlage ist
aber stark modifikabel und ausserdem sind beide Linien geneigt spontan
Zweige abzuspalten, in welchen diese Anlage ihre Wirksamkeit ganz (in
1 die Zweige II von Ernte 1909, und IV von 1910 Ernte) oder fast
ganz (in 1 der Zweig III der von Ernte 1908 abgeht und die Auslese A)
eingebusst hat. Eine Neigung rein violettsamige Zweige abzuspalten,
* besteht mcht."

"In beiden Johannsen'schen Linien ist die Anlage zur Ausbildung
violette Farbe in der Hiilsenschale vorhanden, und zwar ist die Anlage —
sowie jene violetter Farbe der Samenschale — stark modifikabel. In
beiden Linien ist die Neigung vorhanden, spontan Zweige abzuspalten,
in welchen die Wirkung der Anlage durchschlagend, ohne Modification
auftritt, so dass dann nur violette Hiilsen gebildet werden. Violette
Farbung der Samenschale ist ganz unabhangig von violetter Farbung
der Hiilsenschale,"

"Auslese nach griiner Farbe der unreifen Hiilse ist wirkungslos, Aus-
lese nach violetter Farbe derselben nur dann — und dann sofort — von einer

Genetics 4: Ja 1919

GEO. F. FREEMAN

Wirkung begleitet, wenn spontan ein violetthiilsiger Zweig abgespaltet
worden ist."

In a selection carried out upon a variety of lupine (Lens esculenta)
having mottled seed, Fruwirth sought by selection to produce both
dark- and light-seeded strains. Six years selection in one direction and
eight years in the other produced some divergence in the selected lines
but was not effective in producing either self-colored dark- or light-
seeded races.

In a variety of vetch which normally produced either greenish or
cream-colored seed (see selection experiment described above) after five
generations of self-fertilization, there appeared in the harvest of 1910,
2 plants having mottled seeds. In 191 2 after 7 generations of self-
fertilization and selection the same line produced 4 plants having mottled
seeds. Finally, "trat diese Variation auch als Variation einer ganzen
Pflanze bei 5 Individuen der Ernte 1910 auf, nach 9 Generationen aus
Selbstbefruchtung, fiinf in der Linie, vier wahrend der vorangegangen-
en Massenauslese." All mottled seed bred true.

In selection work with Soja beans one or two spontaneous variations
were observed. All effects of selection (from a mass lot), however,
were produced in the first year. The spontaneous origin of a white-
flowered vetch is also noted.

White mustard (Snmpis alba) with which Fruwirth worked, pro-
duces both yellow and brown seed. After eight years of selection of
close-fertilized seed, he was unable to fix the type or even materially to
diverge the tendency in one direction or the other.

In extensive selection experiments with oats which for some charac-
ters were carried through ten generations he decides that selection within
pure lines is without effect.

Fruwirth (1915, p. 450) finally sums up by saying:

"Bei einer Reihe von ausseren Eigenschaften zeigte sich durchweg,
dass in einer Johannsen'schen Linie bestimmt gerichtete Auslese auch
bei Fortsetzung durch eine grossere Zahl von Generationen keine Ande-
rung des Liniencharakters mit sich bringt.''

Macdowell (1915) has reported the results of selection experiments
upon a race of Drosophila which possessed more than the normal 4
bristles on the thorax. The average number of bristles increased for
6 generations of selection. The same selection was carried on for 5
more generations without additional effect. The author concluded that
there were several accessory factors limiting extra bristles which were
gradually eliminated by selection. MacDowell has also shown a very

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

strong correlation of extra bristles with body size. The present writer
strongly suspects that the real factors here concerned were size factors
and that MacDowell's extra bristle selection was merely an indirect
means of selecting for larger size.

The paper by Yuzo Hoshino (191 5) on the flowering time of peas
and rice has been the subject of much interesting recent comment.
Hoshino crossed early- and late-blooming varieties of peas. He found
that the variation behaviors of the Flf F2, F3 and F4 races (detailed
distributions of which are given) could for the most part be interpreted
by assuming the Mendelian segregation of two allelomorphic pairs, A
and a, which determined early- and late-blooming respectively and two
modifiers B and b. Those variation behaviors which could not be ex-
plained by these factors, he supposed to have been caused by a "con-
tamination" of genes. What he means by contamination of genes is
not clear for he distinctly states that he does not refer to such a con-
tamination as is assumed by Castle in rodent crosses. He suggests
"secondary factors." This is the same as assuming additional factors
of secondary importance such as are assumed by Nilsson-Ehle in the
report of his com/w^m-squarehead-Landweizen wheat crosses.

Hoshino has also shown a gametic coupling of early-blooming with
white flowers and late-blooming with red flowers. This coupling is
broken (by physiological interference or crossing over) approximately
1 time in 7.

In crossing early- with late-shooting rice varieties he finds the Fx in-
termediate, the F2 showing strong segregation. The behavior of the
F3 and F4 races were such as would be normally expected of segregat-
ing Mendelian factors.

Castle (19 17) has re-stated certain data and conclusions previously
published (Castle 1912, pp. 163-168). In crossing + variants of hood-
ed rats with wild rats he found that "wild'' was dominant in F2 and that
the hooded extractives of the F2 were often higher in hood grade than
were their hooded grandparents. In crossing "mutant" hooded rats (a
race which suddenly appeared with a very high + hooded condition) with
wild rats, the Fx was of the wild type but the hooded extractives of the
F2 did not drop lower than the range of the original "mutant" race.
Castle concludes that these facts cannot be interpreted as Mendelian
and must be explained as the results of changes in a single unit factor.

The present paper is offered as the first in a series of further con-
tributions to the knowledge of the inheritance of quantitative characters.
Wheat has proved an especially favorable subject for such an experi-

Genetics 4: Ja 1919

8 GEO. F. FREEMAN

ment inasmuch as its small size renders feasible the production of large
numbers without prohibitive expense and the fact that it is close-pollin-
ated greatly simplifies the genetic analysis of the F2 and subsequent gen-
erations.

The characters here studied are the date of the appearance of the first
head on each plant, the total height of the plants measured in centimeters
from the ground to the top of the tallest head (not including beards)
and the width of the broadest leaf.

MATERIAL AND METHODS

A brief description of the four varieties of wheat used may be given
as follows :

Algerian macaroni {No. i)
Late, tall; stems large, stiff; leaves broad, dark green, medium width;
heads large, cylindrical, flattened, long; glumes bearded, pubescent,
light straw yellow; grain large, mostly translucent light amber,
and very hard, but with some grains having spots of opaque starch
in the endosperm. Originally obtained from R. Marie, Algiers,
Algeria.

Algerian red bread (No. 5)
Late, tall; stem medium in size; leaves medium in width and color;
heads medium size, square; glumes bearded, smooth, light straw
yellow; grain red, medium soft, opaque. Originally obtained from
R. Marie, Algiers, Algeria.

Early Baart (No. 34)
Early, low ; stem medium in size ; leaves medium width, medium green ;
heads medium size, square; glumes bearded, smooth, light straw
yellow; grain white, medium soft, medium size, opaque. Origin-
ally obtained locally.

So nor a (No. 35)

Early, low ; stem medium in size ; leaves broad, light green ; heads cylin-
drical, square, medium size; glumes beardless, pubescent, reddish
brown; grain white, opaque. Soft. Originally obtained locally.

All planting was done with a nursery row machine by which each
grain was covered 2 inches deep and spaced 3 inches in rows 10 inches
apart: There were fifty hills in each row. Strips of barley were planted
on either side of the plot in order that the end plants should not have

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 9

more space than those within the plots. All plants of the pure varieties
grown in 1914 were from mother plants which were selected from the
191 3 general mass cultures as true to the types of their respective varie-
ties. Of these selected 1913 plants there were 14 of macaroni (No. 1),
3 Algerian red bread wheat (No. 3), and 5 early Baart. The head rec-
ords for Sonora (No. 35) in 1914 came from 12 typical heads of this
variety selected from a mass culture. In 1915, of the 9 nursery rows
of pure macaroni (No. 1), 6 were plant rows from the previous year's
culture and 3 were from a mixture of seeds resulting from threshing
together a number of typical heads of this variety selected from a field
culture. The 3 nursery rows of No. 35, 1 of No. 3 and 1 of No. 34 were
plant rows from the previous year's harvest. In 1916, 5 of the nursery
rows of No. 1 came from a single mother plant in 1915 (No. 52-4-1-4)
and the remaining 2 from a single other 191 5 mother plant (No. 3-12-
1-5). The 5 nursery rows of each of the other varieties originated from
single plants in 191 5 as follows: No. 35 from No. 35-11 -1-4; No. 3
from No. 32-2-38; No. 34 from No. 1-13-3-1-24. In all of the discus-
sions, the word culture is used in the sense of a group of plants, grown
in a single nursery row and originating from a single mother plant of
the previous season. This applies alike to the pure varieties and hy-
brids. The exception in the case of the 3 nursery rows of mass-selected
macaroni, grown in 191 5, has been noted. The expression "pure race'' is
often used to distinguish plants belonging to one of the parental varie-
ties from those of hybrid origin.

The statistical methods used in these investigations were those com-
monly employed by biometricians. The constants used were the arith-
metical mean, standard deviation and coefficient of variation. The
means were calculated to the nearest unit employed in the taking of
the original data. The standard deviations were calculated from the
mean class as a mean, i.e., with the middle of the mean class as the
assumed mean, no correction being made for the true mean. This was
considered sufficiently accurate in view of the fact that different plant
rows of the same pure race (pure line originating from a single mother
plant) often showed more difference in standard deviation in the same
season than could possibly arise from failure to correct for the true
mean. An example will suffice. All of the plantings of pure No. 3
(Algerian red bread) arose from the seeds of a single plant in 1914. In
191 6 there were 5 plant rows of this culture grown in different parts
of the experimental plots for comparison with the various hybrids into
which this culture entered. The data for height and the statistical con-

Genetics 4: Ja 1919

io GEO. F. FREEMAN

stants calculated therefrom by various methods are given below. The
original measurements were made to the nearest centimeter and in the
summation of the data the classes were made to include 5 cm with the
middle points at 2.5 and 7.5, thus 62.5, 67.5, etc.

Table i a

Height of pure No. 3, 1916, in centimeters.

Row
No.

100

105

115

120

125

130

135

140

145

150

•to

104

109

114

119

124

129

134

139

144

149

154

105A . . .

105B ...

105C . . .

105D ...

105E ...

* Not used in calculation of constants given in table 1 b.

Table ib
Statistical constants.

Row
No.

Number of
variants

True
mean
(A)

Mean used
in the calcu-
lation of rr
used in the
discussions
(B)

Approxi-
mate mean
given in the
tables and
discussions

Standard
deviation
calculated on
(A)

Standard
deviation
calculated on
(B)

105A

122.85

122.5

123

6.4

105B

135.00

137.5

138

7.0

N 105 C

137-40

137.5

138

5-8

105D

129.50

127.5

128

5-2

5-5

105E

138.50

137-5

138

7-7

Averages
and totals

243

132.65

132.5

133

8-5

8.6

Now the greatest difference in standard deviation arising from dif-
ferent methods of calculating was .3 or about 3.5 percent of the average
standard deviation, whereas the greatest difference between the different
lines was 2.5 (that between 105D and 105E) or 29.4 percent, a little
over eight times the error introduced by the different methods of calcu-
lation. In view of such facts it was not considered worth while to waste
time in accuracy of calculation which could not possibly add any sig-
nificant value to the constants so obtained.

Although the probable errors of a large proportion of the constants
here given have been calculated they are not given in the text on ac-
count of lack of space and the difficulty of placing them in compli-

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT u

cated tables of distribution, etc. In nearly every case, however, in which
the reader is interested, the probable errors can readily be calculated
from the data given. In the F2 hybrids most of the cultures had from 85
to 95 individuals and in the F3, from 40 to 48.

It has been necessary to devise some means of comparing the varia-
bility of a series of hybrid races with their pure line parents, each of
which may perhaps be grown in several different plant rows in different
parts of the experimental plots. Moreover, if we accept high variability
as a measure or indication of heterozygosity, it will be of interest to
compare the variability of second generation hybrids with the third
generation (F3). In close-pollinated plants like wheat, as the average
of heterozygosity certainly decreases from generation to generation, the
average variability of plant populations (populations arising from single
mother plants) should also decrease. This average increase in homozy-
gosity with respect to any one character is, however, not uniform in all
lines. The recombinations may be such that an F2 plant is just as hetero-
zygous with respect to the factors governing height, for instance, as
was its F-l parent and the same may be said of certain individuals in the
comparison of the F3 plants with their F2 parents. We will therefore
have some F2 plants just as heterozygous as their F: parents that will
give rise to cultures of F3 which are just as variable as were the F2 cul-
tures, but the majority of the F2 plants will be less heterozygous than
their Fx parents and will therefore give rise to F3 cultures less variable
than were the F2 cultures. Xow since the quantitative characters con-
cerned, as well as the variability of the same, are subject to environic
modification (see behavior of pure lines in table 1) there must be some
means of comparing statistically the variability of the F3 cultures with
the F2 cultures in order to demonstrate this general decrease of variabil-
ity in the succeeding hybrid generations.

Three methods are available as follows :

(a) Throw all the cultures of a given generation into a single popu-

lation and calculate the standard deviation of the same.

(b) Superimpose the means of the several hybrid cultures, sum the

equal deviations on each side of this mean and calculate
therefrom a standard deviation for the whole series.

of each hybrid culture separately and show the average and
distribution of these constants.
These methods and the value of the constants so obtained will now
be discussed in order:

Genetics 4: Ja 1919

GEO. F. FREEMAN

(a) The standard deviation calculated by this method from a popu-
lation consisting of several plant rows of a single pure line is always
greater than the average of their standard deviations taken separately.
This is caused not necessarily by differences in the standard deviations
of the plant rows entering into the total population (these may be all
identical) but by differences (environic) in the means of the several rows
whereby the distribution of the population as a whole is much broad-
ened. The distribution of this total population and the standard devi-
ation derived from it are therefore measures of the total effects of the
given different environments in modifying the character concerned. If
now we are dealing with an F2 generation all of which originated from
genetically equivalent Fx plants, part of the differences in the F2 plants
would be due to environic effects and part to the effects of genetic re-
combination. The distribution and standard deviation of a hybrid popu-
lation calculated by method (a) would therefore give the total combined
effect of environment and recombination in producing variability. When
now we come to consider an F3 population arising from genetically
unequal F2 plants we simply re-measure (if we plant all the seeds of
all of the F2 plants or a sufficiently large random sample) the influence
of the same factors as were measured in the F2, i.e., the sum of the
effects of environment and all of the factors entering the cross from the
original parents. We cover up the possibility of discovering any de-
crease in the heterozygosity of the F2 plants since differences in the
means of the F3 cultures, due to the genetically different parents, will
have the same effect in broadening the distribution of the total popula-
tion, as differences in the individuals of a single highly variable culture.

(b) The method of superimposing the means introduces a small but
unavoidable mathematical error where the standard deviation is used as
a measure of the average variability of a number of separate cultures.
It is well known, however, that where the means differ, the standard de-
viation is not a good measure of comparative variability. In order to
overcome this difficulty and obtain abstract numbers which may be com-
pared, the coefficient of variation has been devised. This is the per-
centage which the standard deviation is of the mean. It is therefore
apparent that a given deviation from the mean has more weight in the
determination of the coefficient of variation when it is a deviation from
a small mean than when it is a deviation from a large .mean. When
now we superimpose small means and large means we give equal values
to deviations which are of unequal value in determining the coefficient
of variation. Hence if our data have to do with cultures differing widely

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

in their means, where the coefficients of variability rather than the stan-
dard deviation must be used in the comparison of variabilities, we are
not justified mathematically either in averaging standard deviations or
superimposing means. As a matter of fact, however, it may be said
that the error introduced by this means is not large. Taken alone, how-
ever, the method of superimposing the means has one serious fault. It
covers up wide differences in the variability of different individual F3
cultures. For the purposes of genetic analysis it is necessary to know
whether all of the F3 cultures have decreased in variability or whether this
decrease is confined to the offspring of certain only of the F2 plants.
It is therefore necessary to calculate the standard deviations and coeffi-
cients of variation of each of the cultures separately.

(c) Since, as just stated, a knowledge of the distribution of the coef-
ficients of variation of a series of hybrid cultures is probably even more
important than a single general expression of the average variability
as a whole, method (c) which gives all of these details is usually to be
preferred.

In general the coefficient of variation was used as a measure of vari-
ability. In time relations, however, this is difficult on account of the
necessity of selecting arbitrarily some point from which to estimate the
means. In the case of the date of first heading, if some date in March,
say the first or fifteenth were chosen, it was feared that the differences
in means would be so great as to unduly distort the coefficients of vari-
ation. One may readily see that the later such a basal date be chosen
the greater will be the distortion on this account. On the other hand,
if the chosen date be moved backward, the various means, in compari-
son with each other, approach unity, and the coefficient of variation be-
comes then more and more dependent upon the size of the standard de-
viation. Although all of the plots were planted within a period of
seven days in the fall and all came up at approximately the same time,
it would be questionable whether the total vegetative period would be
the best basis of a determination of the variability of date of first head-
ing on account of the fact that some strains were more active in winter
than others and were therefore given unequal starts in the rapid vegeta-
tive period of spring. In view of these difficulties it was decided to use
the standard deviation (expressed in days) alone as the measure of
variability in all time relations.

In the studies on size relations, the coefficients of variation only are
given.

Where averages of a series of standard deviations are given, or

Genetics 4: Ja 1919

GEO. F. FREEMAN

standard deviations are calculated from artificial populations produced
by superimposing the means of different races, such fact has been ex-
pressly stated, but it must not be understood that the writer would infer
that these are strictly comparable mathematically to an average of a
series of coefficients of variability, for reasons already given. Rather
than true arithmetical averages, such means should be considered as
foci around which the distribution of the given series of constants (here
standard deviations) cluster, and therefore form, as it were, a locus
for thinking specifically.

DATE OF FIRST HEAD

The dates of the first head of the parents and the F1 plants in 1914
were not taken.

Macaroni X bread wheat crosses. Algerian macaroni (No. 1)
X Sonora (No. 55)

In 191 5, 3 pure races of No. 35, 9 pure races of No. I, and 37 cultures
of (1 X 35) F2, were grown at Yuma. The following results were
obtained :

Table 2 a

Date of first head in F2 of cross 1 X 35 and »» the parent strains, 1915.

Number of

Average dates

cr of

Average cr

cultures

individuals

of first head

population

of cultures

Pure No. 35. . .

168

March 17

2.14

1.66

(1 X 35) F2...

2546

" 27

4.00

3-56

Pure No. I...

650

" 31

330

1.87

Table 2 b

Distribution of standard deviation of cultures.

.75 1.25 1.75 2.25 2.75 3-25 3-75 4-25 4-75 5-25 5-75

Pure No. 35.
(1 X 35) F2
Pure No. 1.

12 5
1

The 37 hybrid cultures were from the seed of the 37 Fx plants secured
in 1 91 4 which were sown in plant rows in 191 5. It should here be noted
that the standard deviation of the whole population is markedly higher
than the average standard deviation of the plant rows taken separately.
This was also true of the pure races and can be attributed in part to the
place variation of the different plant rows. Part of this difference may

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

also be due to slight differences in the genetic composition of the indi-
viduals of the parental varieties used in the original cross. However,
these individuals, although not all belonging to one pure line, in their
respective varieties, were carefully selected as belonging to the type of
the variety which they were to represent. The differences between the
average standard deviation of the pure lines taken separately and of
their respective populations is therefore an approximation of the error
introduced by place varation (modification) and whatever genetic dif-
ferences there might have been in the several individuals of the parental
cultures.

The greater variability of the hybrid cultures as compared with the
parental varieties is in accordance with what would be expected from
the recombination of genetic factors in the F2 generation. The mean of
the hybrid cultures was 3 days later than the mean of the parents and
4 days earlier than the late parent. The heading dates of both parents
and of the F2 cultures may be summarized as follows :

Table 3

Date of first head in (1 X 35) F 0, 1915.

March April

Cultures

15
16

17; 19

l8|20

21
22

23 1 25
24 1 26

27 29

28 30

31
I

4
5

6
7

8
9

10 12

11 13

14I16
IS 17

18
19

20
21

Pure No. 35

(1 X 35) F2

Pure No. 1

25
4

85
18

47
74

7
21

4
403

796
11

3061403
78 | 153

266
132

98
134

86
81

42
54

8
2

Means of cultures.

Pure No. 35

(1 X 35) F: |

Pure No. 1 |

17 12

i| 4

From the 2546 F2 plants, 230 were selected and planted in plant rows
at Yuma in the fall of 191 5. These selections were, for the most part,
based upon economic characters. However, the dates of first heading
of the plants in the spring of 191 5 varied from March 15 to April 9
and thus furnished material for the study of the segregation of the
factors relating to time of heading.

For comparison of the parental varieties with these F3 hybrids, 7
pure cultures of Xo. 1 and 5 pure cultures of No. 35 from plants selected
as types from these same varieties of the previous year, were grown. The
results may first be summarized as follows:

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 4

Date of first head in (/ X 35) Fz, 1916.

Culture

X"umber of
cultures or
plant rows

X'umber of
individuals

Average date
of first head

0- of total
population

Average cr
of culture

Pure Xo. 35. . .

247

March 25

1-34

1.27

(1 X 35) F3...

230

9772

April 11

6.24

3-14

Pure Xo. 1

343

April 15

1.99

.91

Distribution of standard deviation.

Culture

•25

•75

1-25 1-75

2.25

2-75

3-25

3-75

4-25

4-75

5-25

5-75

6.25

6-75

7.25

Pure Xo.

(1 X 35)

F3)

Pure X~o.

The increase in the variability of the F3 population of hybrids over
the F2 population is striking and surprising. Knowing that only se-
lected individuals of the F2 were planted, one, at first thought, might be
inclined to attribute this to the selection of extremes from both ends of
F2 as parents, but observation of the column showing number of cultures
in table 4 will show that the distribution of F2 parents forms practically
a normal curve. One can therefore only attribute this increase to cli-
matic differences in the two seasons which emphasized the effects of
extreme combinations more in 191 6 than in 191 5, or else to the fol-
lowing, which probably accounts for the greater part of the increase.
It will be noted that the standard deviations of both the populations and
cultures, averaged separately, of the parental varieties, was less in 1916
than in 191 5, and also that the same was true of the average standard
deviation of the separate cultures of F3 as compared with that of the
separate cultures of F2. These facts indicate that the season of 1916
did not emphasize the extremes either in the pure cultures of that year
or in the F3 cultures taken separately, or at least that in the latter case
the increasing homozygosity of the F3 over the F2 was a little more than
able to offset this effect and thereby reduce the variability of the F3
cultures as compared with the F2 cultures taken separately. Xow in
this increase in homozygosity of the F3 cultures probably lies the in-
crease in variability of the population as a whole. We have already
seen that the heterozygotes here tend to take an intermediate position.
Hence as the percentage of heterozygous forms decreases with the ap-
proach toward homozygosity, the percentage of intermediate types will
grow less, i.e., the curve will be flattened, and the standard deviation of
the population, thereby slightly increased.

o ^

£ >

£ .s

VO 01 Tf

Tf -4

01 1> CO
On

01 to

co VO
O

O0 -3-
^ 00
to

v£

VO
lO
CO

O On vO
co vo co

CVJ CV1 vO VO

2 =

8 ^

iO O N

CO
01

lO fo"
CO

O ^ O

o X <u

l-H

Tf 01

io o

CO
01

to b

a X «->

t-i i-

P W S

GEO. F. FREEMAN

A summary of the distribution of the dates of first head in the plants
of the parental cultures and the F3 hybrids is shown in table 5.

It should now be noted that, considering individual plants, there were
among the hybrids, 29 plants earlier than the earliest of No. 35 and 293
plants later than the latest of pure No. 1. Moreover, considered as cul-
tures, there were three cultures whose average date of first head was
earlier than the earliest average of any of the cultures of pure No. 35
and that there were 19 cultures averaging later than the latest pure
culture average of No. 1. There were in fact three cultures whose
average date of first head was later than the latest individual of pure
No. 1. Does this indicate that by recombination we may be able to
isolate races which are earlier than the early parent and later than the
late parent?

Table 6 shows the distribution of the F3 individuals and cultures ar-
ranged according to the date of first heading of the parent F2 plants.
+ =j the date of the first head on the selected F2 parent. 0 = the
average date of the population arising from such parents (reading hori-
zontally). In the same grouping of cultures there are also shown the
distribution of the means of the F3 cultures taken separately and the
distribution of the standard deviations of these cultures. The first
vertical column at the left shows the number of F2 plants (hence F3
cultures) in each category. In a vertical column are also shown the
average of the standard deviations of the cultures taken separately in
that category.

Table 7 shows the distribution of the F3 individuals and cultures ar-
ranged according to the means of the F3 cultures. 0 = the average
date of first head of the cultures going to make up the population in
that group (horizontal). This table also shows the distribution of the
selected F2 plants which were the parents of the several cultures making
up the corresponding culture groups. The distribution of the standard
deviations of the several races taken separately which make up its cor-
responding category is given. The vertical columns are the same as
in table 6.

Table 6 shows us that the differences observed in the date of first
heading of the individual plants of F2 were largely genetic, since their
offspring (F3) exhibits but little regression toward the general mean.
Again the same thing is perhaps better shown in table 7 where the F3
cultures are grouped and arranged in accordance with their own means.
We then have the distribution of the parents of these groups of F3 cul-
tures. It will be observed that in no case does the distribution of the
parents, for any group of F3 means extend beyond the normal limits of

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 6

Date of first head in (1 X 35) F3, tyib- Distribution based upon date of first head of

the selected F2 parents.

F3 individuals

March

April

May

Number of

17)19

25|27

I0|

24(26128

cultures

22|24

26|28|30

1 1 3

9 1
1

25|2/|29

81 125

+ 1

163

225

ia8

o4°

210

226

363

102

144

235

268

336

867

504

116

3
+

362

233
O

630

580

153

1
+

333

214

128
O

113

105

7i
O

13
O

^ = 1

+ = Selected F2 parents.
O = Mean of F3 group.

Means of

cultures

March

April

Number

Average

'31

of cultures

(T of F, i

5.21

3-88

3.83

3-39

300

3.21

3.00

2.81

3-34

2.78

351

2.58

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 6 (continued)

Date of first head in (i X 35) F3, 1916. Distribution based upon date of first head of

the selected F2 parents.

Standard deviation of F3 cultures

cultures

•/ b

1 2^

^ 7=;

A 2=>

1*/ 0

0-/ 0

6 2=;

6.75

7 2=;

T
1

variation of the most variable parental culture. If the differences in
the means of the F3 cultures in tables 6 and 7 are due to genetic causes,
one would expect the intermediate cultures to be more variable than the
extremes, thus assuming that the extreme cultures are more nearly
homozygous than those which are intermediate.

Xow noting the distribution of standard deviations in the F3 cultures
as given in tables 6 and 7 and the average of the standard deviations
for separate cultures as shown in the vertical columns, we are unable to
discover such a decrease in variability toward the extremes. In the
present material, however, this is not surprising for the following reason :
Xo. 1 and No. 35 differ in so many genetic factors that there is an ex-
tremely wide range in the products of their recombination. As a mat-
ter of fact many of these recombinations are so radical and unbalanced
that they are no longer automatic (i.e., are unable to give rise to a liv-
ing organism). Hence there is a large percentage of sterility in the F2
and later generations. Xow the recombination of factors which govern
(by their interaction) the time of heading in this particular cross are
likely so many and so widely different that all of the possible recombina-
tions would give a range of heading time far beyond (both toward the
early and late extremes) the limit of physiological possibilities of a nor-
mal wheat plant. Hence in the range of variation observed in the F2
or F3 of this cross we have only a small section taken from some part
of the larger theoretical curve. It would therefore appear much flatter
than the corresponding curve of a pure race and there would be but
little difference in the heterozygosity, hence, variability, i.e., standard

Table 7

Date of first head in (1 X 35) Ft, 1916. Distribution based upon the means

of the Ft culture.

F8 individuals

March April May

Number of

I7|i9|2i

23|25

2426

cultures

18 1 20 1 22

24|26

2527

A
•

101

117

187

128

229

150

135

161

159

209

288

167

269

618

200

166

1 175

788

137

545

768

230

153

128

112

122

O = Mean of group.

Distribution of F2 parents

March

April

Number

Average

of cultures

o- of F3

4.84

307

2.70

3-77

2.80

3-34

343

363

33
56

3-3i
2.85

5
3

14
15

8
16

6
14

2-59

3-47

3.56

4.14

3-24

Genetics 4:

21 Ja 1919

GEO. F. FREEMAN

Table 7 (continued)
Date of first head in (1 X 35) F3, 1916. Distribution based upon the means

of the Fz culture.
Standard deviation of F3 cultures.

Number of
cultures

125

1.75

2.25

2.75

3-25

3-75

4.25

475

5.25

5-75

6.25

6.75

7.25

deviation, of the cultures arising from individuals selected from either
the middle or extremes.

Bread wheat crosses. Red Algerian bread {No. 5) X early Baart

(No. 34)

In 1915, 1 culture of pure No. 3, 1 culture of pure No. 34 and 6 plant
rows of the F2 of 3 X 34 were grown. These hybrid rows were from
the 6 F-l plants of this cross obtained in 191 4. As noted above, dates of
first heading were not taken in the Fx plants. A summary of the results
in 191 5 is given in table 8:

Table 8

Date of first head in (3 X 34) F2, 191 5.

Number of

Average date

o- of popu-

Average a

cultures

individuals

of first head

lation

of cultures

Pure No. 3

March 28

1.60

(3 X 34) F2. . .

538

March 23

3-93

2.95

Pure No. 34. . .

March 16

i-75

1.75

Distribution of cr of separate cultures.

•75

125

1-75

2.25

2.75

3-25

3-75

4.25

4-75

(3 X 34) F2

Pure No. 34

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

As previously, it may be noted again that the standard deviation of
the hybrids both as a population and as separate cultures was higher
than that of the parental varieties. The mean of the F2 hybrid popu-
lation was only 1 day later than the mean of the parents. The head-
ing dates of the populations of parental cultures and F2 hybrids may be
given in table 9.

Table 9

Date of first head in (3 X 34) F2, 191 5.

Marrh

April

1 9

|xo

(3 X 34) F2

1 '

33
33

61
6

136

47
1

130

Means of cultures.

Pure No. 3. .
(3 X 34) F2.
Pure No. 34.

From these 538 F2 plants 112 were selected, for economic reasons,
for planting in the fall of 1915. For comparison 5 cultures of each of
the parental varieties were also grown. These were selected from typi-
cal plants of the parental varieties of the previous season. The range
of dates of first heading of the selected F2 plants extended from March
10th to the 29th, thus covering 19 of the 23 days of total variation of
the F2. The first summary of results are given in table 10.

Table 10

Date of first head (3 X 34) F3, 19 16.

Number of

cultures or

Number of

Average date

o- of total

Average G

plant rows

individuals

of first head

population

of cultures

Pure No. 3

242

April 13

1.52

.82

(3 X 34) Fg. .

112

5321

April 5

6.43

2-95

Pure No. 34. . .

244

March 25

3-IO

2.17

Distribution of standard deviations.

•25

•75

1-25

i-75

2.25

2.75

3-25

3-75

4-25

4-75

5.25

575

6.25

(3 X 34) F3

Pure No. 34

The general features of this table are the same as those for the other
crosses, namely, that the average standard deviations for the cultures are

Genetics 4: Ja 1919

GEO. F. FREEMAN

less than those of their respective populations and that the hybrid cultures
are much more variable than the pure lines. Moreover, as in the com-
parison of tables i and 4 we here note also an increase in the variability
of the F3 population of hybrids over that of the F2. (Compare tables
8 and 10.) The failure of the average standard deviation of the hybrid
cultures to decline from 1915 to 1916 should be noted. Does this indi-
cate a lack of progress toward homozygosity?

Such an inference would be natural were it not for the peculiar be-
havior of the parental pure race No. 34.

It will be observed that the variability of this race was strongly in-
creased in 1 91 6 over 191 5, although all of the 5 cultures belong to one
and the same pure line, i.e., the single pure line grown the previous
year, which had originated from a single plant in 1914. Perhaps the
same factors which caused this increase in the variability of the pure
line No. 34 were also able to increase the variability of the hybrid cul-
tures which arose from No. 34 as one parent and that this influence
upon the variability was sufficient to offset that of increasing homozy-
gosity and thus maintain the variability for the two seasons at approxi-
mately the same figure.

The distribution of the dates of first head in the parental races and
in the F3 hybrids for 1916 is shown in the following table:

Table ii

Date of first head in (j X 34) F3, 1916.

March April

21I23I 25

22|24| 26

Pure Xo. 3

1 1

138

(3 X 34) F3

I2|i7|43|i39

415

76i

675

597

842

39i

195

157

1 103

Pure No. 34

30 41 56 23

Means of cultures.

Pure No. 3 |

(3 X 34) F3 |

| 1

Pure No. 34 |

■1 «

21 1 9

6| 10

It is interesting to note here that no hybrid plant was earlier than the
earliest individual of the early culture and that there were only 19 later
than the latest of the late parent. Again considered as cultures, the
means of the hybrid cultures all fall within the limits set by the extreme
means of the parental variety cultures. Here recombination does not
seem to have extended the variability definitely beyond the limits of the
parents.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Tables 12 and 13 show the segregation of the F8 to be just as marked
in this cross as in the cross already discussed. The greater variability
of the intermediate classes is also quite evident. This fact taken in
connection with the fact that there was no indication of partial sterility
among the hybrids seems significant. It is exactly what should be ex-
pected if the segregation of the F2 plants and F3 cultures were due to
recombination. This should be contrasted with the absence of greater
variability of intermediates in the semi-sterile hybrids of the bread
wheat — macaroni wheat crosses.

Table 12

Date of first head in (3 X 34) F* Distribution based upon dates of first head
of the selected F2 parents.

Number

F3 individuals

March

April

cultures

| 9|n

Ji7

| 29

1 2

1 6

jio

1 12

172

158

153

103

104

481

396

248

214

101

121

170

4- 1

. 37

195

151

124

134

619

125 [ 1

+ = Selected
O = Mean of

F2 parents,
group.

Means of F3

cultures

March

April

Number

Average & of

of cultures

| F3 cultures

6.30

1.47

2.01

3-79

346

4.02

382

2.46

2.03

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 12 (continued)

Date of first head in (3 X 34) F3, 1916. Distribution based upon dates of first head

of the selected F2 parents.
Standard deviations of F3 cultures.

Number
of cultures

•25

125

175

2.25

275

3.25

375

4.25

475

5-25

575

6.25

675

•i

Table 13

Date of first head in (3 X 34) F3, 1916. Distribution based upon means of F3 cultures.

F3 individuals

Number | March April

of 1

9|ii|

I5|i7

21 23 25

29 31

8 1

I0|

I2|

I4|l6|l8|

20|22

cultures

I0|l2

16] 18

22|24|26

30 1 I |

2l|23

1 1

4 24

9 [ 16

2H\ 38

1
1

200

72|i95|io8

! ! O

i| 2

| 40 1 157 1 222

! 1 1

101

1
1

I 1

1 1
1 1

55|n6|i36|io8

111

221

1 18

\ *

i 1

| 1

| 20

1 78)136

1 1

192

291

58 1 24

1 2

IOI

1 1

i i

1 '3

1 II 27

1 150

i9| 5

| 79

1 1 1 1

1 1

1 ll I

1 9

I52

| 84 | 45|26| 7

1 52

1 5

1 1 1 1

1 1

1 4

j 27

1 112

1 58

146

|I3|2I3

I 6

1 Ti

'III
(III

1 1

1 2

j 19I 40

ki"

I248

1 1 1 1

Mil

I
1

1 1

3 8

1 2

|352

i ^

1 1

1 1 1

O = mean of group.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 27
Table 13 (continued)

Date of first head in (3 X 34) F3, 1916. Distribution based upon means of Ft cultures.

Selected F, parents

March April

Number

Average a of

of cultures

F3 cultures

2-55

2.05

3-35

4.18

432

4.18

3-74

3-63

2-53

1.21

Standard deviations of F3 cultures.

Number

of cultures

.25

•75

125

1-75

2.25

2-75

3.25

3-75

4.25

4-75

5-25

5-75

6.25

6-75

7
8

1
I

2
1

Summary; date of first head

In both crosses the parents had wide differences in heading dates and
the averages of the F2 and F3 were in every case intermediate and nearer
to the late parent. The range of the individual hybrid plants in no case
extended significantly beyond the range of the early parent toward ex-
treme precocity of heading. Toward the late extreme, however, in the
macaroni — bread wheat crosses, there was a long extension of the range,
much beyond that of the late parent. As a matter of fact many plants
never headed, but remained as dark green, grass-like tufts until they were
killed by the heat and dryness of the summer. Among the bread wheat
crosses the extension of the range of date of first head beyond the ex-

Genetics 4: Ja 1919

GEO. F. FREEMAN

treme of the late parent was never marked and could, in fact, be ac-
counted for by the normal extension of the curve due to greater num-
bers.

The same observations made above with regard to the relation of the
means of the hybrid populations to their parental means, apply also to the
distribution of the means of the hybrid cultures, as compared with their
parents, in the F2. In the F3, however, the matter was somewhat dif-
ferent. In the macaroni — bread wheat cross there were 3 cultures whose
average dates of first head were earlier than the earliest parental aver-
age and there were altogether 19 cultures averaging later than the latest
parental average. Since there were 230 cultures concerned, 8.2 percent
are thus seen to lie outside of the parental range. In the bread wheat
cross, on the other hand, there was no case where the average of a hy-
brid culture was outside the range of averages for the parental varieties.
As regards individuals in the F2 the parental types were abundantly re-
covered in every case. As regards means of F3 cultures (a better cri-
terion of the genetic constitution of the F2 plants) the parental types
were also recovered in all cases.

In all cases where more than one culture was involved the standard
deviations of the population were greater than the average of the standard
deviations of the cultures taken separately and in all cases the standard
deviations of the hybrids1 were greater than those of either parent both
as regards that of the populations and the averages of the cultures taken
separately.

In comparing the standard deviations of the hybrid F3 populations
with their respective F2 parental populations we may note the following
observations: (1) the standard deviation of F3 populations are so de-
pendent upon the range of F2 parents chosen, that conclusions drawn
from the calculation of this constant should be carefully guarded. The
standard deviation of the F3 population of both crosses was greater
than that of the F2 population. Since heading time appears to be im-
perfectly dominant in these hybrids, the number of intermediate types
will tend to be reduced as the population approaches homozygosity. If
therefore we assume a Mendelian inheritance, whenever the selected F2
parents practically cover the range of distribution of the F2 population
and form a random sample thereof, we would expect the F3 population
to have a higher standard deviation than the F2 population.

When we come to compare the average variability (here measured by
standard deviation) of the F3 cultures taken separately with the average
1 It should be remembered that the Fj is not here included.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 29

variability of the F2 cultures we are not hampered in our conclusions, to
so large an extent as mentioned above in comparing the variability of
the F2 and F3 populations. With a Mendelian interpretation there is no
genetic reason why any F3 culture should be significantly more variable
than the most variable F2 culture. Moreover, the average variability of
the F3 should be equal to or less than that of the F2, whatever the mode
of selection. We may now observe as follows : ( 1 ) In the macaroni —
bread wheat cross, 1 X 35. the average variability of the F3 cultures
was significantly below that of the F2 cultures. (2) In the bread wheat
cross some complications arose. The average standard deviations of
the F2 and F3 cultures, of the 3 X 34 were the same (2.95). This, how-
ever, cannot be assumed as evidence of a lack of progress toward hom-
ozygosity, for the following reasons : It will be observed that the vari-
ability of pure race No. 34 was strongly increased in 191 6 over 191 5
(2.17 and 1.75, respectively) although all 5 of the cultures grown in
1 91 6 came from the 1 culture grown in 191 5, which in turn came from
a single plant in 1914. Perhaps the same factors which caused this in-
crease in the variability of the pure line No. 34 were also able to in-
crease the variability of the hybrid cultures which were grown from
No. 34 as one parent and that this influence upon the variability was
sufficient to offset that of increasing homozygosity and thus maintain
the variability for the two seasons at the same figure.

The strongly fluctuating nature of the variability of date of first head
is shown by a study of the distribution of the standard deviations of the
F2. In every case the range of distribution of the standard deviations
of the F2 overlapped the range for one or both parents. This could be
explained by assuming a partial-blending inheritance and assuming that
in some F1 plants the blend was more complete than in others. If this
were true the F3 cultures grown from these low-variable F2 cultures
should also show a low variability. The results are given in table 14.

Table 14

Number of F2

Number of

Average &

Number of F2

Number of

cultures as little

F3 cultures

of these

cultures more

F3 cultures

Average a of

variable as

arising from

F3 cultures

variable than

arising from

these cultures

one parent

these

either parent

these

148

3.12

194

305

It is thus seen that the low-variable F2 cultures gave rise to the higher-
variable F3 cultures. This is what would be expected upon a Mendelian

Genetics 4: Ja 1919

GEO. F. FREEMAN

interpretation if we assumed that the low variability of the F2 cultures
in question were so because but few of the extreme combinations chanced
to occur. It must be admitted however that the difference shown is not
large enough to be significant. We may therefore safely conclude that
the differences in standard deviations of the F2 cultures were wholly
fortuitous and without genetic significance.

In the F3 generation, in all cases, cultures occurred with as lowr vari-
ability as that of the parents, i.e., there were cultures which, insofar as
variability is concerned, appeared as nearly homozygous as the pure lines.

With a Mendelian interpretation we are accustomed to expect those F2
plants which take a position relative to the parents similar to that oc-
cupied by the mode of the Flf to give rise to F3 cultures which are more
variable than the F2 plants otherwise located. In the macaroni — bread
wheat crosses we are not able to observe any relation of this kind. This
fact, however, does not argue the absence of Mendelian segregation for
the following reasons : The macaroni and bread wheats here crossed,
differ in so many genetic factors that there is an extremely wide range
in the products of their recombination. Many of these recombinations
are so radical and unbalanced that they are no longer automatic. Hence
there is a high percentage of sterility in the F2 and later generations.
Such sterility may have the effect of flattening the distribution curve of
the F2 or perhaps even limiting it to one end or the middle or even the
extremes of a curve which would be formed by all of the recombination
possibilities. As already pointed out many of the F2 plants never got
beyond the rosette stage and many plants which made a robust vegeta-
tive growth were completely sterile. The study of sterility in these
crosses will be reserved for a future paper. In circumstances such as
these it is apparent that there may occur very little difference in the
heterozygosity, hence variability, of the cultures from individuals se-
lected from either the middle or extremes of the fertile F2 of such a
population. In the 3 X 34 cross there is a very apparent greater vari-
ability of the cultures arising from the modal F2 plants (see tables 12 and
13). It should be noted that here there was complete fertility and the
F2 selections covered nearly the whole of the range of the F2 population.
A glance at tables 6 and 12, where the F3 individuals are grouped with
reference to the heading date of the F2 parents, yields abundant evi-
dence that some sort of segregation has occurred. The F2 plants were
not alike genetically. All of the phenomena observed can be explained
by assuming that heading date is governed by three or more Mendeliz-
ing unit factors. No attempt has been made to determine the number

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 3I

of factors in any case but the fact that many of the intermediate groups
(see tables 6 and 13) show cultures with low variability would indicate
that the number of factors concerned was rather large, thus providing
the possibility of securing several genetically different but still homozy-
gous types.

HEIGHT

Macaroni — bread wheat crosses. Algerian macaroni (Xo. i) X

Sonora (No. J5)

In this study all height measurements were made from the ground to
the top of the highest head (not including the awns). Lengths were
taken to the nearest centimeter and expressed in the summaries to the
nearest five centimeters. Xo pure Xo. 35 was grown in 19 14 which
was comparable with the pure Xo. 1 and the (1 X 35)Fi- The Xo. 1
grown in 191 4 was not a single pure line but was from seed of several
different mother plants of this variety. A summary of the results for
1 91 4 is shown in table 15.

Table 15

Heights in centimeters in (1 X 35) Fu 1914.

Number | 70 | 80
of plants | 79 | 89

100
109

no
119

120
129

130] 140
139 149

150 | 160
159I169

Aver-
age

C.V.

(1 X 35) F,

*5i | i|
39

3
1

26
1

43 1 49
4|_8

18
21

134
147

1 0.0

8.0

The FA was taller but no more variable than the parent given. Thirty-
eight of these hybrid plants gave rise to hybrid cultures in 191 5. The re-
sults are summarized in table 16.

Table 16
Heights in (1 X 35) F2, 19 15.

Coefficient of variation

Number of

Average

of separate

cultures

individuals

height

population

cultures

648

147

6-7

(1 X 35) F2

2535

122

19.6

19.0

Pure No. 35

166

128

11. 1

6.4

Distribution of the coefficients of variation of cultures.

7 | 8 | 9 io|n|i2|i3|i4|i5

i6| 17

I9|20|2l|22[23|24

» '1 Ml

2 8

1 1 1 1 1

9 8 4| A A 1

(1 X 35) F2

Pure No. 35

1 «

1 M 1 1 1

Mill

Genetics 4: Ja 1919

GEO. F. FREEMAN

It should be noted here that, whereas the F1 was taller than Xo. i, the
tall parent, the average of F2 (where all of the Fx was planted) was
lower than either parent. The high sterility of the F2 plants has al-
ready been noted. As usual the hybrids were more variable than either
parent. It should also be noted that the F2 hybrids were much more
variable than the Fx.

Table 17 gives the distribution of the populations and means of both
parents and the F2 hybrids as regards height.

Table 17

Heights in centimeters in (1 X 35) F2, 1915.

Distribution of individual heights

Distribution of
means of cultures

50 60 70

100

no 120 130

140

150

160

170

180

120

1301140

150

109

119 129 139

149

159

169

179

189

119

129

139

149

159

Pure Xo. 1

4 25 89

:55

217

139

(1 X 35)F2

104

178

226

311 409 447

399

248

4
1

Pure Xo. 35

25 1 29 52

Only three of the hybrid plants were taller than the tallest individuals
of the tall parent, but there were 95 lower than the lowest individual of
either parent. Xo hybrid culture averaged as tall as the highest average
for the low parent, but 4 cultures averaged lower than the lowest aver-
age of either parent. All recombinations so far obtained appear there-
fore to be less vigorous than the parental races. Since the Fx plants
showed considerable range in height, it would be interesting to know
whether this was inherited to any degree in F2, i.e., was the range in Fx
due solely to modification or were these differences partly genetic ? Table
18 shows the F2 cultures grouped according to the parental height. The
class in which the parental height fell is marked +, and the mean of the
population arising from such parents is marked O.

While the last class is 8 cm higher than the first class, considering
the small number of races in each, this difference is not above the prob-
able error. YVe may therefore safely conclude that for all practical pur-
poses the F1 plants were uniform genetically.

Two hundred and thirty of the F2 plants were selected for planting in
the fall of 191 5 and gave rise to hybrid cultures which were measured
just before ripening in 19 16. For comparison 7 pure cultures of No. 1
and 5 pure cultures of Xo. 35 were grown. The first summary of re-
sults follow.

< s

1 1 §

O ON

i i— i >— i

I/".

O Ov
LO ID

+§

-r

O On

.00

CO
CN
M

IX
to

O On
co co

»o

1 O On

* o ^

LT.

O On

8 |

CO
<*

8 S

l-H

On
CO

eg £

O On

vS s

co"

O On

o o

ro co

Number
of

individuals

XT)

&
IN

Average
height of
off spring

l-H

Height of
parent

O On
CO CO

O O
in th

Number of
cultures

GEO. F. FREEMAN

Table 19

Height in centimeters in (1 X 3d) F3, 1916.

X'umber of
cultures

Number of
individuals

Average
height

Coefficient
of variation
of the
population

Average C.V.
of separate
cultures

Pure Xo. 1

344

137

6.6

(1 X 35) F...

230

10084

118

20.3

15.4

Pure Xo. 35...

246

123

7-1

6.3

Distribution of coefficients of variation in (1 X 35) Fg, 1916.

5| 7 I 9

33135

6| 8 |io

Pure Xo. 1 ... |

4* 1

(1 X 35) F3..|

6I15I24

Pure X"o. 35. . |i

2\ I 1 I

As usual it may be observed that the pure races are less variable than
the hybrids and that the average coefficient of variation of the cultures
is smaller than those of the populations. It should be further noted
that the average coefficient of variation of the F3 hybrid cultures is
smaller than that of the F2. This is to be expected in the case of in-
creasing homozygosity.

Table 20 shows the distribution of the populations in 191 6.

Table 20

Heights in centimeters in (; X J5) Fz, igi6.

100

120

130

140

150

160

170

180

190

109

119

129

139

149

159

169

179

189

199

Pure Xo. 1

114

(1 X 35) F...

127

217

404

496

862

1335

1723

1757

1435

1077

47i

Pure Xo. 35. . .

141

Distribution c

>f means.

Pure Xo. 1

(1 X 35) F3

Pure X'o. 35. . .

Only 15 hybrid plants were taller than the tallest individuals of the
tall culture. Considering the large number of hybrids in comparison
with the number of Xo. 1, these few taller plants are without signifi-
cance. At the other end of the scale, however, we find 474 plants lower
than the lowest of the lower parent. Considering means we also note
with interest that there were 86 hybrid cultures averaging lower than the
lowest average for the low parent and one hybrid culture averaging lower
than the lowest individual of the low parent.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Tahle 21

Heights in centimeters in (i X 35) Fs, igi6.

Arrangement of F8 individuals grouped according to F2 parents

cultures

100

120

130

140

150

160

170

180

190

109

119

129

139

149

159

169

179

189

199

+ 0

0 +

123

3-'

160

226

240

128

1 1

12I24I33

106

144

250

326

304

168

20 31

107.

174

310

473

507

328

223

124

186

271

428

423

351

138

• +

141

178

282

385

354

174

Number
of cultures

Distribution of means of F3 cultures

100
109

no
119

120
129

130
139

140
149

150
159

Average coefficient
of variation

3
8
12
24
35
55
48
40
4
1

2
4
5
7
11
1 1
5
3
1

3
8
10
17
7
8
1

5
11

14
10

2
6
14
13

16
18
16
17
15
16

14
14

Number
of cultures

24
35
55
48
40

4
1

Distribution of coefficients of variation

I3|i5
I4|i6

36 38

46 48

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 22

Heights in centimeters in (1 X 35) F3, 1916-

Distribution of F3 individuals

isumDer
of cultures

10 20 30

50 60

100

120

|l30

140

150

jioo.i/O^So

19 29 39

109

119

129

139

149

159

169 179 189

176

I25

207

134

526

118

175

292

534

292

144

3/3

2S0

131

269

481

484

264

234 470

493

183

180

172

Number — ,

of cultures 1 80 90

Distribution of F- parent:

13
21

49
53
42
36
12
1

89 99

100 no I 120 130 ■ 140 150 160 170
109 119 129 139 149 159 169 179

3
4
11
10

5
2

3
4
11
17
11
6
2

1
4
5
7
14
14
3

3
8
10
13
5

Average coemcient
of variation

24
4i
23
21

17
14
14
11
11
5

Distribution of coefficients of variation

of cultures

3
4

5
6

9
10

11
12

13
14

15
1 16

17 19 21

18 20 22

26 28

29
30

|32

33 35 37 39 41 43

34 36" 38 40,42,44

45I4"

|46|48

49
50

13
21

49
53
42
36
12
1

3
1
1

6
2

5
2

3
7

3
9
2

9
7
3

1
7
8
11
8
3

5
9
6

4
2
1

J * 7

'1 S| 1
5L6' 6
6' 4' 4
8' 4' 1

'1 1
1 1
1 1

3
3
2

2
1

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 21 shows the height of the F3 plants grouped according to their
F2 parents, the means of the F3 cultures and the coefficients of variation
of these cultures, respectively, making up each population group. Table
22 shows the height of the F3 plants grouped according to the means of
the F3 cultures, the heights of the parents 'giving rise to these groups and
the standard deviations and coefficients of variation of the F3 cultures,
respectively. It should be noted in table 21 that, while there was con-
siderable regression toward the mean, there was a nearly uniform corre-
lation between the height of the F2 parent and the F3 offspring. By
comparing table 21 with table 20 it will be observed that the distribution
of the means in any group of hybrids is no wider than the range of
variation of the individuals in either of the parental varieties. Observ-
ing the averages and distribution of the coefficients of variation we note
an irregular but yet fairly definite lessening of variability in the taller
groups.

Again comparing table 22 with table 20 we note that for any F3
group (in table 22) the distribution of the parents was not wider than
the distribution of the individuals of the parental varieties. The differ-
ences in the heights of the individuals of these parental groups (which
gave rise to cultures having the same mean) could therefore be assumed
to be environmental modifications of plants of the same or equivalent
heredity so far as height is concerned.

The column showing the average coefficient of variation and the dis-
tribution of these constants in table 22 shows a very decided decrease in
variability of those cultures which have high means.

One conclusion stands out prominently from these tables. The fac-
tors for height were not uniform in the F2 plants. Recombination had
occurred so that on the average (i.e., excluding environmental modifica-
tions), tall parents gave rise to tall offspring and the grading of the par-
ents into a series of ascending heights resulted in a slightly less marked
but still regularly ascending series of offspring groups. The complete-
ness of this series indicates that the number of factors was large.

Algerian macaroni (No. 1) X Algerian red bread (No. 3)

In 1914, 151 plants of pure No. 1 and six plants of pure No. 3 to-
gether with 5 plants of (1 X 3) Fi were measured for height.

The following table shows the distribution of the heights of these
plants and their means. Except for the pure No. 1, the numbers were
too small for the calculation of the standard deviations with any degree
of accuracy.

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 23

Heights in centimeters in the (1X3) F„ 1914.

Number
of plants

70
79

80
89

90
99

ioo| no
I09[ii9

120
129

130
139

140 | 150
149; 159

160
169

Mean

Pure Xo. 1

151

134

(1 X 3) F,

124

118

The numbers are too small to give results of any particular signifi-
cance, but it may be noted that the range of the Fx hybrids lies within the
range of the most variable parent and that the mean of the hybrids
lies between the means of the two parent cultures.

The 5 Fj. hybrid plants gave rise to 5 hybrid F2 cultures in 191 5. For
comparison in the same year 9 cultures of Xo. 1 and 1 culture of No.
3 were available. Table 24 gives a summary of the results.

Table 24
Heights in the (1 X 3) F2, 1915.

Coefficient

Average C.V.

Culture

Number

Number of

Average

of variation

of cultures

individuals

height

of the

of the separate

population

cultures

Pure No. 1

648

147

8-5

6.7

(1 X3)F,...

406

118

21. 1

20.4

Pure No. 3

146

4.2

Distribution of coefficients of variation.

Pure No. 1 . . .

(1 X 3) F2..

Pure No. 3. . .

Whereas the Fx hybrids were intermediate between the parent races,
the F2 averaged lower than either, the two parent races being of prac-
tically equal height. The variability of the hybrids was strikingly higher
than that of the parental cultures.

Table 25 gives the distribution of the populations and the means of
both parents and the F2 hybrids as regards height.

None of the hybrid plants was taller than the tallest individual of the
parental cultures but there were 29 lower than the lowest individual of
the parents. It is striking that all of the means of the hybrid cultures
save one were lower than the lowest parental mean. All recombinations,
therefore, appear to be less vigorous than the parental cultures.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 25

Heights in centimeters in the (/Xj) Fa

29/5«

40
49

50
59

60
69

70
79

80
89

90
99

100
109

no
119

120
129

130
139

I40
149

ISO
159

160
169

170
179

Pure No. i

(i X 3) F2....
Pure No. 3

1
17

4
27

4
39

4
57

25
61

89
87
3

155

34
18

217
27
21

139
8

Distribution of means of cultures.

Pure No. 1

(1 X 3) F2....
Pure No. 3

2
I

3
1

Table 26 shows the F2 cultures grouped according to the height of
their respective F1 parents. The class in which the parental height fell
is marked + and the means of the population arising from such parents
are marked °.

Table 26

Heights in centimeters in the (l X 35) P«, 1915.

Number of
cultures

Height of
parent

Average
height of
offspring

Number |

of in- |30
dividuals|39

50
59

60
69

70
79

80
89

90
99

100
109

no
119

120
129

130
139

140
149

150
159

160
169

100
109

105

82 |

0 +

130
139

54 |

140
149

126

270 |

1
1

•I

-1- 1

45 1 23

Although the range of each of these groups is practically the same,
the distinct correlation between the height of parent and height of off-
spring cannot be disregarded. This would indicate that one or the other
of the parental stocks was not pure as regards the factors influencing
height and that the Fr plants were, therefore, not all equivalent genetic-
ally in this respect. In or,der, therefore, to avoid complications, the sub-
sequent discussion of this cross will be based upon the product of a single
Fx plant (145 cm high) in 1914 from which a culture (No. 32-1) was
grown in 1915, of which the following data may be given:

Table 27

Heights in centimeters in the (1X35) F., 10/5.

Culture

Height of
parent

Average height
of offspring

Number of
individuals

| Distribution of heights of individuals
^6y &C |7o|8o|9o|ioo|iio|i2o|i3o|i4o|i5oli6o
1 79 1 89 1 99 1 109 1 1 19 1 129 1 139 1 149| I59| 169

32-1

145

130

15 | i| i| 4| io| 7| ii| 9| I3j io| 5

Genetics 4:

Ja 1919

4o GEO. F. FREEMAN

From this culture 40 plants were selected as parents in 191 5-' 16. A
first summary of the results may be given as follows :

Table 28

Heights in centimeters in (1 X 3)F3, 1916.

Cultures

Number of
cultures

Number of
individuals

Average
height

Coefficient
of variation
of the
population

Average C.V.
of separate
cultures

Pure JCo. 1

342

137

8-5

6.6

(1 X 3) F3....

1758

123

20.6

14.2

Pure No. 3

243

133

8.0

6.6

Distribution of coefficients of variation

5(7

1 1

Cultures

6|8

4|2

(1 X 3) F,

Pure No. 3

III

Again we perceive that the averages of the coefficients of variation of
the cultures are less than the coefficients of variation of their respective
populations, and that the pure lines are less variable than the hybrids.
The average variability of the F3 is markedly less than that of the
cultures in F2.

Table 29 gives the distribution of the populations and means of both
the hybrid and parental cultures.

Table 29

Heights in centimeters in (1 X 3)F3, 1916.

[ 30

100

120

130

140

150

160

170

| 39

109

119

129

139

149

159

169

179

123

(1 X 3) F3

1 8

100

157

244

274

252

320

I/O

108

Distribution of means.

Pure No. i |

1 1

1 1 *

31 1

(1 X 3) F3 | | 1

1 1 ^

5 9\ 6

Pure No. 3 | j |

1 1

1 1 3

1 1

Observing tables 28 and 29 it is evident that on the average, height-
vigor in the F3 hybrids was again less than for the two parental cul-
tures but that there were two hybrid cultures taller than the tallest aver-

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

age for the taller parent. On the other hand 20 hybrid cultures were
lower than the lowest average of the low parent.

Table 30 shows a fairly uniform correlation between the height of the
selected F2 parent and the average height of its F3 offspring. Table 31
exhibits rather strikingly the fact that the taller F3 cul'tures are much
less variable than those which averaged lower. Now if one will com-
pare the distribution of the selected F2 parents (table 30) with the total
F2 population as shown in table 25, it will be observed that the selections
just cover the upper half of the range. As regards the variability of
the F3, therefore, table 31 and the accompanying column of average co-
efficients of variation might be assumed to represent only a half curve.
The low selections were therefore really intermediate F2 individuals.
The higher variability of these lower F3 cultures, and the very evident
decline in variability as we approach the taller, real, extreme, can be in-
terpreted as being in accord with the idea of hybrid recombination of
height factors with the intermediate forms most heterozygous and hence
more variable.

Table 30

Heights in centimeters in (1 X 3)F3, 1916.

Number of

100

120

130

140

150

160

170

cultures

109

119

129

139

149

159

169

179

13
O

13
+

20
O

89
+

80
O

147

72
O

1
+

+, Selected F2 parents.
O, Means of F3 groups.

Number of
cultures

Distribution of means

of F,

cultures

Average
coefficient
of variation

70
79

80
89

90
99

100
109

no
119

120
129

130
139

140
149

150
159

■Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 30 (continued)
Heights in centimeters in (1 X j) ^3, 1916.

Distribution of coefficients of variation in F, cultures

of cultures

Table 31

Heights in centimeters in (1 X 3)FS, 1916.

Distribution of F3 Individuals

cultures

3o| 40 | 50

100

120

130

140

150

160

I/O

109

119

129

139

149

159

169

179

114

109
O

108
O

194

O, Means of F3 groups.

Number of

Distribution of F2 parents

Average

cultures

100

120

130

140

I50

160

coefficient

| I09

119

129

139

149

159

169

of variatioi

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 31 (continued)
Heights in centimeters in (1 X j) F3, 19 16.

Distribution of coefficients of variation in F3 cultures

of cultures

7 | 9

3«

Height in bread wheat crosses, 3 X 35

No pure No. 35 was grown in 191 4 for comparison with the pure
No. 3 and the Fx hybrids of 3 X 35- The following table summarizes
the data for the pure No. 3 (6 plants, not a pedigree line) and the
(3 X 35) Fi hybrids.

Table 32

Number of
plants

Average
height

Distribution c

f heights of indi\

iduals

Culture

100 | no
109 119

120
129

130
139

140
149

150
159

160
169

170
179

Pure No. 3

118

2 | 1

(3 X 35) F,

142

■

The hybrids are thus seen to be taller than the pure No. 3 and the
range is slightly greater, but not more than would be expected with the
larger number of individuals grown, i.e., one could not infer that the
hybrids were more variable than the pure race.

Each of the 18 Fx plants gave rise to an F2 culture in 1915. For
comparison 3 cultures of No. 35 and one of No. 3 are available. Table
33 summarizes the results for 191 5.

Table 33

Heights in centimeters in (3 X 35) 191 5-

Number of

Number

Average

Coefficient
of variation

Average C.V.
of separate
cultures

Distribution
of C.V.

Culture

cultures

plants

height

of the
population

3
4

5
6

r\9

8 10

Pure No. 3
(3 X 35) F2
Pure No. 35

18
3

l6l I
166

146
148
128

4.2

74
11. 1

4-2
6.0
6.4

I
I

15
1

, 2
1 «

Genetics 4: Ja 1919

GEO. F. FREEMAN

It is here interesting to note that the hybrids are somewhat taller than
the tall parent.

Table 34 gives the distribution within the populations of F2 hybrids
and parental races. In the hybrids, the cultures are arranged in groups
with regard to the height of their F1 parents.

Table 34

Heights in centimeters in (3 X 35) F2, 1916.

Xumber of
cultures

Parental
height

60
69

70
79

80
89

100
109

no
119

120
129

130
139

140
149

150
159

160
169

I/O

179

Pure Xo. 3

(3 X 35) F2

120
129

130
139

144

169

140
149

236

286

<«

150
159

105

««

160
169

(3 X 35) Fa
Totals

1 1

|4| 4

I 5

154

508

663

235

Pure Xo. 35

1 8|25

52 1 38

I3| 1

Distribution of means of cultures.

Pure Xo. 3
(3 X 35) F:
Pure No. 35

j j 1 |t —

1
1

1
6
1

+, Selected F1 parent.
O, Mean of group.

No appreciable correlation between the height of the Fx parent and
the average of the F2 offspring is apparent. We may therefore con-
sider that so far as the height factors are concerned, the Fj. plants were
all equivalent. The range of distribution of the hybrid population
slightly exceeded that of the most variable parent in both directions but
no more than would be expected considering the larger number of plants
grown.

From the above F2 hybrids 80 selections were made for growing in
1915-16. These ranged from 118 to 173 cm high, thus covering all of
the upper but not quite all of the lower end of the range of the F2. For
comparison with these, 5 cultures of each of Nos. 3 and 35 were grown.
A first summary of the results are shown in table 35.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 35

Heights in centimeters in (3 X 35) Fa, 1916.

Culture

Number

of
cultures

Number
of

individuals

Average
height

Coefficient
of variation

of the
population

Average C.V.
of separate
cultures

Pure No. 3

243

133

8.0

6.6

(3 X 35) FB...

3849

143

8.4

6.3

Pure No. 35. . .

246

123

7.2

6-3

Distribution of coefficients of variation.

(3 X 35) F3

Pure No. 35

It should here be noted that the average height of the hybrids is again
greater than that of the taller parent and that there is no diminution
in the variability of the F3 from the F2. Moreover, the hybrids are no
more variable than the pure races.

Table 36 gives the distribution of the populations of the hybrids and
their parental races as well as the distributions of the means of the cul-
tures of each.

Table 36

Heights in centimeters in (3 X 35) F3, 1916.

Distribution of individuals

Distribution of
means of cultures

40
49

50
59

60
69

70
79

80
89

90
99

100
109

no
119

120
129

130
139

140
149

150
159

i6o| 170
169(179

no
119

120
129

130
139

140
149

150
159

Pure No. 3. .
(3 X 35) F3.
Pure No. 35.

1
I
1

3
3

1
I

I
37
10

164
72

519
141

108
1045
17

48
1350
1

6
611

104 | 14

3
12

2
26

That we should here have 42 hybrid cultures (slightly more than
half) whose average heights were higher than the highest average for
the tall parent is somewhat surprising. Especially is this so when we
reflect that the variability of the hybrids is no greater than that of the
pure lines.

From table 37 we observe that the regression of the offspring of ex-
treme selections is quite strong, but it is not complete. The difference
between the means of the offspring of selected extremes is greater than

Genetics 4: Ja 1919

GEO. F. FREEMAN

between the means of the parental races (compare table 35). Compar-
ing the distribution of selected F2 parents forming the groups in table
38 with the distribution of the individuals of their parental varieties in
table 34, we will note that they are not more widely distributed. They
can therefore be assumed to be environic modifications of individuals
representing equivalent genetic combinations so far as height is con-
cerned. There was a fairly well marked decrease in the variability of
the taller cultures.

Table 37

Heights in centimeters in (3 X 35) Fz, 1916.

Arrangement of F3 individuals grouped according to F2 parents

Number of

109

119

129

139

149

159

169

cultures

108

118

128

138

148

158

168

178

" +~~

10
+

5
O

0 +

188

265

314
O

132
+

148

484

669
O

273

43
+

176

212
O

Distribution of means

Distribution of coefficients

Number of
cultures

f F3 cultures

Average C.V. of

of variation of F3

cultures

119
128

129
138

139
148

149
158

F3 cultures

3
4

5
6

7 I

9
10

13
14

12.0

8.0

8.2

6.1

5-9

5
1

5-9
50

4
1

Red Algerian bread (No. 3) X early Baart (No. 34)

In 1 91 4 there were grown 6 plants of pure No. 3, 12 plants of pure
No. 34 and 6 plants of (3 X 34) Fx. These numbers are too small to
warrant the calculation of coefficients of variation but the distribution
and averages may well be given.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 47

Taiile 38

Heights in centimeters in (3 X 35) Fa, 19 16.

Number of
cultures

34
8

F8 individuals arranged in accordance with the means of the
F3 cultures

109

119

129

139

149

159

169

108

118

128

138

148

158

168

178

108

242

150

•

213

542

402

231

797

390

105

170

Distribution of coefficients of variation

Number of

Distribution of F2 parents

of F3 cultures

cultures

109

119

129

139

149

159

169

Average C.V.

118

128

138

148

158

168

178

of F3 cultures

8.1

6.3

5-8

Table 39

Heights in centimeters in (3 X 34) Fu 1914.

Cultures

Number
of plants

Average
height

100
109

no
119

120
129

130
139

140
149

150
159

160
169

Pure No. 3

118

(3 X 34) Fx...

123

Pure No. 34. . .

150

The F1 is here seen to be intermediate in height between the parents
and with a smaller range of variation than either.

Each of the 6 Fx plants gave rise to an F2 culture in 1915. For com-
parison, one culture of No. 3 and one of No. 34 were available. Table
40 gives first summary of the results.

Table 40
Heights in (3 X 34) F2, 1915.

Average

Coefficient

Average

Distribution

Number

of variation

of C. V.

Culture

of cultures

of plants

height in

of the

C.V. of the

centimeters

population

cultures

Pure No. 3. .

146

4-2

(3 X 34) F2.

537

150

7-1

5-0

Pure No. 34.

137

4.1

4.8

-Genetics 4: Ja 1919

GEO. F. FREEMAN

As in the last bread wheat cross (No. 3 X 35) and unlike either of
the bread wheat X macaroni wheat crosses (i X 35 and 1X3) the
average height of the F2 is greater than the mean of the parents, in fact
greater than either of the parents. As usual the coefficient of variation
of the F2 taken as a population was greater than the average of this con-
stant for the separate cultures and the average coefficient of variation
of the hybrid cultures was greater than that of the pure parent cultures.

Table 41 gives the distribution of height in the parental races and the
F2 hybrids of this cross.

Table 41

Heights in centimeters in (3 X 34) F2, 1915.

Distribution of

means

Culture

Distribution of individuals

cultures

100

120

130

140

150

160

170

130

140

ISO

109

119

1291

139

149

159

169

179

139

149

159

Pure

No. 3

(3 X

34) F2

151

232

Pure

No. 34

That we should have 4 hybrid cultures averaging taller than the tall
parent is interesting, but may be ascribed to hybrid vigor.

The following table (table 42) gives the distribution of the F2 popu-
lation grouped according to the height of the Fi parents, + being the

height of Fj parent,
such parents :

and O the mean of F2 individuals arising from

Table 42

Heights in centimeters in (3 X 34) F2, 191 5.

Number
of cultures

Parental
height

80
89

90
99

100
109

no
119

120
129

130
139

140
149

150
159

160
169

170
179

Average
height

no
119

147

120
129

152

130
139

100

155

There is thus seen to be a slight correlation between the height of the
Fj parents and the height of the F2, indicating a possibility of some
genetic differences in the Fx in respect to height. In all further discus-
sion of this cross, as regards height, it will be necessary to segregate the
data into groups so as to consider at one time only plants originating
from a single Fx parent. Since nearly all of the F3 population arose

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

from one or the other of the original Fx plants, Xos. 25-1 and 44-2,
all F3 cultures except such as originated from these two will be excluded
from this study, and these will be kept separate. The distribution of the
F2 of these two cultures were as follows :

Table 43

Heights in centimeters in (3 X 34) F2, 1915.

Culture

(44-2) F2, 1915
(25-1) F2, 1915

Parental

X umber

Average

height

individuals

height

120

152

135

155

Distribution of
individuals

120
129

130

140

150

160

139

149

159

169

170
179

Average
C.V.

4-5
4-9

The selections for the F3 covered the full range of both of these par-
ents. Table 44 gives a summary of the results in F3.

Table 44

Heights in centimeters in (3 X 34) F3, 1916.

Culture

Number of

Average

cultures

individuals

height

243

133

(3 X 34 F3 (44-2)..

2408

133

(3 X 34) F3 (25-1)

2396

131

Pure No. 34

243

121

Coefficient of variation

Average of

Distribution of C.V.

Culture

Population

separate

cultures

8.0

6.6

(3 X 34) F3 (44-2)

9-8

6-5

(3 X 34) F3 (25-1)

5-9

Pure No. 34

6.2

In 1 91 6, it will be observed that the average height of the F3 is prac-
tically the same as the taller parents. The coefficient of variation of the
hybrid population is greater than that of the populations of either parent
but the average coefficient of variation of the hybrid cultures taken
separately was not significantly below that of the pure cultures.

The distribution of the heights of the individuals of the F3 popula-
tion and the parental cultures and also of the means of the separate cul-
tures are given in table 45.

Whereas the ranges of the hybrid populations extend beyond the limits
of the parents, this is here not surprising considering the much larger

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 45

Heights in centimeters in (3 X 34) F3, 1916.

Culture

Pure No. 3

(3 X 34) F3 Total
(3 X 34) F3 (44-2)
(3 X 34) F3 (25-1)
Pure No. 34

Distribution of individuals

100
109

no
119

453
215
184
85

120
129

68
1570
652
770
119

130
139

140
149

108
1819
769
934
14

48
io58

493
428

150
159

6
292
179
29

160
169

170
179

Distribution of means
of cultures

100
109

110
119

120
129

130
139
2
50
23
24

140
149

150
159

3
1

numbers used. It is interesting, however, to note that 17 hybrid cul-
tures had average heights higher than the highest average for the par-
ental cultures.

Table 46 shows the distribution of the F3 grouped according to the
selected F2 parents. In table 47 the F3 is grouped according to the
means of the F3 cultures. Table 46 shows a definite correlation between
the height of the selected F2 parent and the mean of the F3 classes, but
there is a strong regression, especially in the higher groups. The F2 se-
lections, it may be noted, covered practically the entire range of the
F2 population. The distribution of the parents in the F3 groups of
cultures having equal means, was not greater than the normal distribu-
tion of individuals in a pure culture. They could therefore be assumed
to be modifications (environic) of genetically equivalent individuals.

Table 46

Heights in centimeters in {3 X 34) Fz, 1916.

Number
of cultures

Arrangement of F3 individuals grouped
according to F2 parents

1 60 1 70 1 80 1 90 1 100 1 1 10 1 120 1 130 1 140 1 150 1 160 1 170]

109

119

129

139

149

159

169

179

109

119

129

139

149

159

(44-2)

°!

171

218

233

180

218

281

410

124

(25-1)

382

410

129

13 J

279

423

254

+ i

Distribution of means
of F3 cultures, 1916

1 ool no 120 130 140 150

+, Selected F, parent.
O, Mean of F3 group.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 46 (continued)
Heights in centimeters in (3 X 34) Ft, 1916.

Average

C. V.
per cent

Distribution of coefficients of

Number

Mean

variation of F

1 cultures

of cultures

of F3

3
4

5
6

9
10

13
14

15
16

(44-2)

107

16.0

125

5-8

130

6.3

137

6.4

132

7-0

(25-1)

122

6-3

130

6.3

133

5.6

136

4.0

Cultures

Table 47

Heights in centimeters in (3 X 54) -F3, 1916.

Arrangement of F3 individuals grouped
according to means of F3 culture

1 69

79 1

99 | 109

119

129

139

149

159

169

[129

139

149

159

169 179

1 60

7o|

90 100

120

130

140

150

l60

\l2Q

130

140

150

160)170

Distribution of
selected F2 parents

(44-2)

5
O

129

340

207
O

22
8

9
1

257
6

504
52

235
206

130
O

4
4

14
4

(25-1)

23
O

129

517

296

287

225

629

113

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 47 (continued)
Heights in centimeters in ( j X 34) fit> *9*&

Distribution

of coefficients of

Cultures

Mean of

Average

of variation of

F3 cultures

F,
* 3

C. V.

(44-2)

107

16.0

118

6.0

123

7-6

134

6.0

147

4-5

153

5-0

(25-1)

120

4.0

126

6.2

135

5-8 .

142

5-3

Summary; JicigJit

The number of F1 plants grown were too small to give significant re-
sults except in the case of the 1 X 35 and 3 X 35 crosses. In both of
these cases the Fj averaged taller than the tall parent. In the other two
cases the F1 was intermediate. In the two macaroni — bread wheat
crosses (1 X 35 and 1X3) the F2 and F3 averaged below both pa-
rental races. In the two bread wheat crosses (3 X 34 and 3 X 35) the
F2 averaged taller than either parent and the F3 of the 3 X 35 cross
was taller than either parent, but in the 3 X 34 cross the average of the
F3 was 1 cm shorter than the taller parent. The distribution of heights
in F1 did not go significantly beyond the limits of the parental cultures
in any case except that of 3 X 35 in which the whole distribution was
pushed upward about 24 cm. The range of distribution of the indi-
vidual heights of the F2 and F3 in neither case of the macaroni — bread
wheat crosses extended significantly above that of the parents, but in
both cases extended markedly below the parental range. On the other
hand in the bread wheat crosses the range in both cases extended dis-
tinctly above, but not significantly below, the parental ranges in F2 of
both crosses and the F3 of the 3 X 35 cross, but in the F3 of the 3 X 34
cross it did not extend significantly either above or below the parental
range. The same observations made with reference to the distribution
of the individual heights of the F2 and F3 of both kinds of crosses also
apply with perhaps greater emphasis to the distribution of the means
of the F2 and F3 cultures taken separately.

Xow, referring to the appropriate tables, note that the average height

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 53

of Fj in one of the species crosses (macaroni — bread wheat; was above
the tall parent and in the other intermediate between the parents. We
must therefore assume that the maximum heterozygosity of these crosses
will give plants at least taller than the low parent. In both the Fa and
F3 of these crosses, however, the average F2 and F8 height was below the
parent. We are therefore compelled to conclude that recombination and
not antagonistic heterozygosis is the cause of the low averages of the
F2 and F3. A complete double set of macaroni factors, a complete
double set of bread wheat factors, or the combination of one complete
set of factors from each species, was able to produce a plant of standard
vigor, but a large majority of the recombinations of these factors where
a complete set from one of the species was lacking, resulted, through
failure of coordination, in the production of plants of reduced vigor.

Now it should be noted that no F2 plant, tall because it was com-
pletely heterozygous, could give rise to an Fa culture which had a high
average height, for the reasons above given. Hence the majority of
tall F3 cultures must have arisen from F2 plants, tall because they were
genetically completely, or nearly completely, like one of the parents. Now
this is in harmony with the fact (see tables 22 and 31) that the taller
F3 cultures were markedly less variable than were those with a less aver-
age height. Now let us remember that the completely heterozygous F1
plants of the 1 X 35 cross were tall plants with wrinkled seeds. If we
examine the F2 plants selected and pick out all of those which were
taller than the average of the low parent and which also had wrinkled
seed, thus again resembling the Fx plants we find that the average height
of the F3 cultures arising from these were no cm with an average co-
efficient of variation of 19.5 percent, whereas the average height of the
offspring of all of the remaining selected F2 plants taller than the aver-
age of the low parent was 123 cm with an average coefficient of varia-
tion of 1 4. 1 percent. Again, if we pick out all of the selected F2 plants
which were taller than the average of the low parent and which also had
smooth seeds, thus resembling one or the other of the parents, we find
that the average height of the F3 cultures arising from these was 126
cm with an average coefficient of variation of 12.6 percent.

A similar study in the 1 X 3 cross gave for the FT-like F2 plants F3
cultures with an average height of 131 cm and an average coefficient of
variation of 12.9 percent, whereas the parent-like F2 plants gave F3
cultures with an average height of 143 cm and an average coefficient of
variation of 6.6 percent.

While these facts coincide completely with the assumptions above

Genetics 4: Ja 1919

GEO. F. FREEMAN

made, the story does not end here. Returning to the i X 35 cross we
found that there were 30 tall Fi-like F2 plants and 73 tall parent-like
F2 plants. If now we cast the F3 cultures arising from these two groups
respectively into subgroups arranged according to the average heights
of the F3 cultures and find the average coefficients of variation of each
subgroup we may tabulate the results as in table 48.

Table 48

Average heights of F3 cultures in centimeters.

70 | 80 90
79 | 89 | 99

100
109

no
119

120
129

130
139

140
149

150
159

30 F3 cultures from tall F2 plants hav-
the wrinkled seed (Fj-like F£
plants)

Distribution of
heights

2 | 3

Average coeffi-
cients of variation

30.0

26.0 1 23.7

20.5

18.5

16.2

16.0

73 F3 cultures from tall F, plants hav-
ing smooth seed (parent-like F2

Distribution of
heights

1 1 2

20
10.0

9.4

Average coeffi-
cients of variation

23.0 1 16.0

15.6

13.7

13.8

With these results we must conclude that we have not yet succeeded
in separating out genetically equivalent groups and that those F3 plants
which gave rise to tall F3 cultures are genetically more nearly homozy-
gous or else we must postulate some other cause for the suppression of
variability in the taller F3 cultures. This last analysis in no way inter-
feres with the conclusions already drawn, for it clearly shows that in
F3 subgroups of equal height, those cultures arising from F-L-like plants
were always more variable than those which came from parent-like
plants.

Now turning to the bread wheat crosses we note that the average
coefficients of variation of the F2 and F3 generations were in no case sig-
nificantly higher than that of the most variable parental culture (see
tables 33, 35, 40, 44). If, however, we consult tables 38 and 47 we
shall observe a distinct lowering of the variability of the taller cultures.
Let us also remember that the Fly F2 and F3 of the 3 X 35 cross all
averaged taller than the tall parent and note (table 38) that the reduc-
tion of the variability of the taller F3 cultures was uniform, whereas the
F1 of the 3 X 34 cross was intermediate, the F2 taller and the F3 again
intermediate, and while the reduction in variability of the F3 cultures
(table 47) was still apparent (with the exception of 1 erratic extreme)
there was some indication that the intermediate F3 classes (Fx-like)
had a tendency to be a little more variable. There appears, therefore, to

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

be two conflicting forces at work, one (heterozygosis) tending to make
the cultures arising from the Fj-like F2 plants more variable, and another
which tends to suppress variability in the taller cultures.

A means of testing for the presence of a factor suppressing variabil-
ity, which is independent of heterozygosity, is found in the F2 cultures
which came from supposedly genetically equivalent I7! plants. In the F2,
the means and variabilities of the several cultures from any given cross
should be the same. Where slight differences occur, they are in all prob-
ability environic. Nevertheless if the cultures be grouped according to
these slight differences in the F2 means, and the average coefficients of
variation of these groups calculated, if there be a factor suppressing vari-
ability in the taller groups it should become apparent, provided there is
a sufficient number of F2 cultures to give valid averages. Such an analy-
sis of the F2 hybrid cultures for 191 5 is given in table 49.

Table 49

Correlation between average height and coefficient of variation in F2 hybrids.

Total
number

Average heights, 1915

ico| no
I09| 119

120
129

130
139

140
149

150
159

(1 X 35) F;

Number of cultures
Average C. V.

! 4
1 192

30
19.0

18.9

(1 X 3) F2

Number of cultures
Average C. V.

28.5

■

20.2

10.4

14-5

(3 X 34) F2

Number of cultures
Average C. V.

5-5

4-8

(3 X 35) F2

Number of cultures
Average C. V.

7.0

5-9

6.0

The differences, while not large, are as uniform as could be expected
from such small numbers and indicate the presence of a suppression
factor of some sort which slightly reduces the variability of the taller
cultures.

The presence of this suppression factor for variability in the taller
cultures is even more strikingly shown in the pure races. Grouping the
cultures according to their means (without regard to year in which they
ore grown) and calculating the average coefficient of variability for each
group we have the result shown in table 50.

Having now shown that there is a factor which, independent of
heterozygosity, may suppress the variability of the taller cultures, we may
conclude as follows :

(1) Some factor for suppressing variability has been able to com-

Genetics 4: Ja 1919

GEO. F. FREEMAN

Table 50

Correlation between average height and coefficient of variation in pure races.

Total
number

j Average height

no
119

120
129

130
139

140 | 150
149 | 159

Pure No. 1

Number of cultures
Average C. V.

2
7-5

7-5

6 | 4
6-5 1 5-5

Pure No. 35

Number of cultures
Average C. V.

3
6.7

6.4

3-9

Pure No. 3

Number of cultures
Average C. V.

2
7-5

3
5-6

4.2

Pure No. 34

Number of cultures
Average C. V.

6.9

4
6.1

4-8

pletely mask the effect of heterozygosity in a cross where the F2 and
F3 cultures averaged taller than the tall parent (3 X 34).

(2) This same factor has largely suppressed, but not entirely masked,
the variability due to heterozygosity in a cross where the F2 and F3
cultures were approximately as tall as the taller parent (3 X 35)-

(3) The factor for the suppression of variability in tall cultures is
apparent in crosses where the averages of the F2 and F3 cultures are
below those of the low parent, but was in no case able to obliterate the
effect of heterozygosity (see 1 X 35 and 1 X 3)-

The question as to the nature of this suppression factor will be re-
served for future discussion. The fact that the average variability of
the F2 and F3 cultures was not significantly higher than that of the pure-
line parents in the bread wheat crosses might be cited as showing that
a blending inheritance has occurred with the production of a single new
race no more variable than the most variable of the parental races,
were it not for the fact that tables 37 and 46 show a definite positive
correlation between the height of the F2 parents and the means of the
F3 cultures derived therefrom. A distinct segregation occurred in the
formation of the gametes of the Fx plants whereby the F2 plants were
different genetically and exhibited these differences in the means of
their offspring, thus giving rise, not to one race, but to a number of
distinct races. The theoretically expected greater variability of the F2
and F3 cultures are simply here suppressed, but in the macaroni — bread
wheat crosses where this suppression factor was ineffective in masking
the variability due to heterozygosis the variability of the F2 and F3
cultures in all cases averaged markedly above that of the pure-line par-
ents.

In the F3 of all crosses, cultures were secured having the parental

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 57

types both as regards average height and variability. In the bread wheat
crosses the average variability of the F3 cultures was slightly larger than
1 that of the F2 cultures in both cases. This is in accordance with the
circumstance that in both, the average height of the F2 cultures was
markedly greater than that of the F3 cultures and thus called into more
active effect the variability-suppressing factor already shown to influ-
ence the taller cultures. In the macaroni — bread wheat crosses, on the
other hand, the average height of the F2 was greater than that of the
F3 in one case and less in the other, but still the average variability of
the F2 cultures was markedly above that of the F3 cultures in both cases.
This is in harmony with the fact pointed out above that the variability-
suppressing factor visible in all of the crosses was not sufficient to mask
the influence of heterozygosity in macaroni — bread wheat hybrids.

Finally we may conclude that all of the facts observed in the study of
the inheritance of height in the wheat crosses here considered are in
harmony with the hypothesis of the segregation of a number of simple
Mendelian unit characters and that there is present some factor (as yet
unknown) which suppresses variability in the taller cultures of both
pure lines and hybrids and that this factor is sometimes able to com-
pletely mask the variability which would normally be produced by
heterozygosity.

WIDTH OF LEAF

In the following study of the inheritance of width of leaf in wheat
hybrids, all measurements are given in millimeters. Averages are there-
fore given to the nearest millimeter.

Macaroni (No. 1) X Sonora (No. 55)
No pure No. 35 was available for comparison in 1914. The data
with reference to the pure No. 1 and the Fx hybrid plants are given in
table 51.

Table 51

Width of leaf in millimeters (1 X 35) Ft, 19 14.

Number

Distribution

of individuals

Aver-

Coefficient

of plants

I3|i4

I5| i6| 17

20|2I

22 23 24|25|26|2/|

age

of variation

Pure No. 1

151

aj 1

i| 3|»

25I24

32 | 5|io| 3| 3| i|

(1 X 35) F,

1 1 2

4| 4

8|9|9| i| i| |

We will here pause only to notice that both the range and variability
of the pure No. 1 were greater than for the hybrid. The average leaf

Genetics 4: Ja 1919

GEO. F. FREEMAN

width for the hybrid was greater than for the pure No. i, but since the
No. i is here the more narrow-leafed parent we have as yet no indica-
tion as to whether or not we are dealing with imperfect dominance or
hybrid vigor.

In 191 5 there were available for comparison 4 cultures of No. 35, 9
cultures of pure No. 1 and 37 cultures of the (1 X 35) F2. A summary
of these data is presented in table 52.

Table 52
Width of leaf in (1 X 35) F2, 191 5.

Number of
head rows

Total num-
ber of plants

Average
width of
leaf

Coefficient
of variation

of the
population

Average
C. V. of
cultures

Pure No. 1

651

13-0

10.3

(1 X 35) F2...

2537

30.2

29-3

Pure No. 35 . . .

169

13-5

13.0

Distribution of coefficients of variation

Pure No. 1

(1 X 35) F2

Pure No. 35

The average of the hybrids is below that of either parent. The stan-
dard deviations of the populations are greater than the averages of the
standard deviations of the separate cultures making them up, and the
variability of the hybrids is much greater than that of the pure cultures.
All hybrid cultures were more variable than the most variable pure
culture.

Table 53 gives the distribution of the several populations and the
distribution of the means of the cultures.

Studying these distributions we note that there were 16 hybrid plants
having leaves wider than the widest individual of the widest-leaved
parent, but there was no hybrid culture averaging as wide as the most
narrow average for Sonora, the wider-leaved parent. On the other
hand more than half of the hybrid cultures averaged lower than the
lowest average of any macaroni head-row and there were 121 hybrid
plants having more-narrow leaves than the narrowest-leaved individual
of the macaroni parent.

<
W

<
H
o

eans

1-1

o |

o 3

tributi
of (

t— i 1— 1

IT)

01
N

—

00 M

U~;
N

rt
N

On VO
CO

CO O <0>

Tj- t-f

01
oi

h a rt-

m \o M

VO to to
MOP)

11 a 1 c
Uals

00 h n
K N CO

ivid

k\o a

O rf w

O O to

r-1

H N M

of widtl

to On vo
IX <N
CV)

LT,

01 Q 01

uoi}

bin

VO O

01 og m

£J

01 On
1— 1

01 fO
VO

VO
to

M ft CO

Pure
(i X
Pure

PO M
lO _ ii

01 01
to CO

"tO VO ^'

oi oT~ oi ^
On 0\ ^ O
On ^ O On

~~ to vo oT

01 01 00 ^ >-

O 01 i 01 tF
T n _ o>

co io vo o

tOl't ~<0 lO

T 0J_ HH_

oi o i

0 ^o

W ro ?

HH _ 00 1^

>ocT

w ^OfrOcpCgO^ %Cx C

tOOO On to On On ^ fO

GEO. R. FREEMAN

Referring to table 51 it will be observed that there was considerable
variation in the width of leaf of the Fx plants. Table 54 groups the
191 5 F2 plants in accordance with the leaf width of their Fx parents
in 1914.

A glance at this table is sufficient to show that there is no correla-
tion whatever between the parental leaf width in 19 14 and the average
leaf width of the offspring in 191 5. We may therefore conclude that
all of the variation observed in the F1 plants was nutritional and that
they were all equivalent genetically so far as the factors governing width
of leaf were concerned.

From these F2 hybrids 230 selections were made which gave rise to a
like number of F3 hybrid cultures in 191 6. For comparison with these
there were available seven head-rows of No. 1 and five head-rows of
No. 35. The selected F2 plants used as parents ranged in width of leaf
from 10 to 35 mm. The very wide-leaved individual was very striking
in appearance and was nearly sterile. Table 55 gives a first summary of
the results in 19 16.

Table 55

Width of leaf in millimeters in (1 X 35) F3, 19 16.

Class

Number
of cultures

Number of
individuals

Average
width of leaf

Coefficient
of variation

in the
population

Average coeffi-
cient of vari-
ation of sepa-
rate cultures

344

12.0

IO.I

(1 X 35) F3...

230

10123

24.9

20.9

Pure No. 35. . .

246

, s

15-2

14.0

Distribution of coefficients of variation

1 1

I5|i7

2/|29|3i|33[35[37|39

45|47|49|5l|S3

Class

'14

i6|i8

28 3° 32|34|36 38|40

46|48|50 52|S4

Pure No. 1

III II

1 1 1 1

(1 X 35) F3

3I24

35|3i

5|w Si 7\ I 3 3

1 1 1 1 1

Pure No. 35

MINI

1 1 1 1

The average for the hybrids is less than either of the parents ; in
every case the coefficient of variation of the population is greater than
the average for the pure cultures of the same class and the coefficient of
variation for the hybrids is greater than for either parent. The coeffi-
cent of variation both for population and average of cultures among the
hybrids was lower in 1916 than in 1915. This was also true of the pure

— ^

In,

Tj"

C
-I

u-)
~I

—

CM
CM

fN
^1

PO
X

tN,

LT,

lO
In

Is
"I

irj

U~,

—

IN.

-t

-1-

"o~

1— IN.

tN.

l-H

irj

Xn.

po z

Pure

Z CO £
PL.

GEO. F. FREEMAN

cultures, and therefore may be in part environic. One thing, however,
remains to indicate progressive increase in homozygosity among the
hybrids. This is the much greater difference in the coefficient of varia-
tion of population and average of cultures, which was apparent in 191 6.

Table 56 shows the distribution of the populations of pure cultures
and hybrids of this cross in 1916.

The hybrid population shows a distribution far beyond both extremes
of the parents. This is also true of the means of cultures. Part of this
greater distribution is of course due to the normal extension of the
curve from the much larger number of hybrids grown. That the curve
of variation is more flat, however, is shown by differences in the shapes
of the curves of variation which are rendered comparable by reducing
each group class to a percentage of the total number in the population
and disregarding all percentages less than one-half of one percent and
expressing all percentages to the nearest integer (see table 57).

Table 57

Ji'idth of leaf in millimeters in (1 X 3) Fz, 1916.

1 5

Pure No. 1 |

(1 X 35) F3 | 1

Pure No. 35 |

When reduced to equal areas the polygon of the F3 hybrid distribution
is thus seen to be limited by a curve much more flat and with more ex-
tended limits than either of the parent races. This indicates that the
extension of the range of variations of the F3 hybrids over the parental
races is genetic. This is further shown in table 58 where F3 cultures
are thrown into groups or populations in accordance with the leaf width
of the selected F2 parental plants.

Though somewhat erratic at the extremes, these results show a very
definite genetic segregation of leaf width in the F2 as exhibited by the
means of their offspring. The distribution of the means of the cultures
in each of these groups is shown in table 59.

Geo. F. Freeman, The heredity of quantitative characters in wheat

Table 58

Width of leaf in millimeters in (1 X 35) J9l6. Distribution of F3 individuals grouped according to the leaf width of the F2 parents.

Number

Leaf width

of cultures

of parent

2Q>

1
T

■7
0

j j

/~V I

7
/

1
0

"T"

°4

T 7

t e
1 0

T -?

7"?

AA
44

/-\
W

T /I

-5

j j

C T

I J

•5
O

Il8

-1.
T

t e

e
3

-^4

7 >\
o4

T T7

12s

104

I JA
X14

■^4

_I_

48
40

■7
0

r 1

77
00

TOS

107

101

1A
o4

-j
0

118

"T"

24.

1 1

y—

127

130

I Izl
1 14

•

TS7

119

172

\AA
144

135

A1
4o

' J

■2
O

■

"T"

e
0

r
0

1 2

91
z6

/in
4U

/I T

1
O

176

r\
\ 1

3°

II9

101

t8t
101

A3^

3°

t e

8
0

t;8

t e

j j

65 107

112

T27

t;7

7fV

3°

t e

t r

1
O

°6

+, leaf width of parent ; O, mean of F3 groups.

Genetics 4: Ja 1919

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Table 59

Width of leaf in millimeters in (1 X 35) Fz, 1916. Distribution of means of I:3
cultures grouped according to the leaf width of the F2 parents.

Number

Parental

leaf

Mean of group

of cultures

width in

1915

in 1916

. I2

This table exhibits even more plainly than the preceding the correla-
tion between the parental leaf width and the mean leaf width of the
offspring.

In order to determine whether the offspring of narrow-, medium-,
and wide-leaved F2 mother plants exhibited any definite difference in
their variability table 60 was constructed.

There is shown here an irregular but still evident diminution of vari-
ability among the offspring of the wider-leaved parents.

It may be suggested, moreover, that since width of leaf is highly in-
fluenced by the environment and there is therefore a strong regression
of the mean of the offspring of extreme variants toward the general
mean of the population, we may get a better idea of the segregation of
leaf-width factors, by grouping the F3 cultures according to their own
means and then calculating the variability of these groups and observing
the distribution of the parents which gave rise to them. We thus
measure backward, determining the range of environic modification of
individuals which are able to give rise to genetically equivalent prog-
enies.

Genetics 4:

Ja 1919

3> 1- rs
< «

LO VO
LO »0

CO t)-
LO VO

M 0)
LO LO

on o

IX 00

M 04

On O
CO

t>> 00
co CO

co co

co rt-

co co

M 04
ro CO

ON O

04 co

W 01 HH M CO M fS|

tN. 00
04 01

MM M M M

LO O

04 01

HH __|HHTfM04'^-'-" M

co "<t
04 04

MM <M co 04 p-i

M OI
04 04

On O
m 04

i- MlXvOM00 04 rot>»04MO|M M

IX 00

04 ^ co 04 04fOOi-i04 mm

>-ir004TfTj-0|MONT1-i-'— mm m

- MHt^t'NNt^H M

w 04

On O

Average
C. V. for
group

o o omdlocovo c Lo w coonlooo o*-. q q lo o
vd ro ^- lo od d d oi m c oc d> m K. oi oi io o; od O

04 04 04 0) m 04 04 04 04 04 m — 04 M 01 CM M 04 m 0*

Parental
leaf width

O 04 co^tiOVO N CO OnO >-< 04 rorJ-iOVC ixOOiO

Number
of cultures

CO 0) 1- O CO On 00 04 04 04 fO rf 00 K h m 04 m
MM04M0404't-'co04M

h
o

H
§

w
SJ
H

<
w

CT5

NO * S

w s g

I J II

•fc o

£ 8

____

^i*

01
01

01*

01
CO

l-H

«t

ok"

vo*

01
vo

vo 0^

l~'

js,

n§

Tt-
CO

vo
CO

Of)

NO
l-H

°5

l-H

-3-

On
01

VO
01

°oT

l-H

«fr

1— 1

Ol ...
00 O N

W CO

NO
01

— oo~

~~ vo

On
01

l-H

o ^

NO
CO

1— 1

CO
01

0nO

vo
vo
01

On
vo

CN|

VO
VO

NO
CO

00
01

CO
CO

l-H

VO
01

l-H

01
l-H

idth

Mean \v

rj-

l-H

Number
of cultures

01
01

l-H

10 1 >-i

CO
CO

CM
CO

O
CO

On
CM

00
CM

VO
CM

N N h CM

rt
CM

1— I HH (— 1 hH

t— 1 t— ( CM

CM
CM

i-h LT) LO CO 10 CO

►H CM COCM 0\ K fO hh

CM tJ" CO co

CO CO tJ" CO 1—1 CO 1—1

Tf- hh lo rt CM hh

CM CM CO CM m tJ- hi

COCOCOVO'^CM'-H hh

h N Tt ro

M M M >p « « W

01 H H

Mean of
F3 cultures

OOOh-'CMCOTj-mvOt^COOO'-'

Number
of cultures

CM CM 10 CM O 1O00 N00 N vO m CM

HNCOtTtOlMH

1 1

01 |
lO LO 1

C\ o
n- m

l_t 01

CO Tj-

co CO

CO CO

CO r-t-

co co

— _ _ — !— 1

M 01
CO CO

01 i-i 01

0 C

01 co

i-i 01 CO 01 CO

OC
01 01

Ol _ _ „

01 01

01 01 CO 01

*o -

01 N

1-1 01 01 01 01 -i M

M 01
01 01

to in o 1- co 01 oi

01 01 Tj-OOCX fO m

oi mmcox oi >-i m

i_r. O

co ^1"

M CO O ^ N N M M

On O
i— i

Average 1
C.V. of F8

up IOC O 01 DC \q c »n
co O in oi oi i-i o 00 ^* oo to m

tCO0|C|C|0|P|MHMMHH

Means of

Fa cultures

00 2 m ^?2"iCjV£ ~ ^ * oi

Number
of cultures

68 GEO. F. FREEMAN

It is interesting to note in table 61 that practically all of the curves
of group distribution are skew, i.e., they slope more abruptly toward
the upper limit.

While the parental groups in table 62 exhibit considerable range, a
comparison of tables 62 and 56 will show that this is not wider than
occurs in the nutritional variations of a pure line.

Algerian macaroni (Arc. 1) X Algerian red bread {No. 5)

For this cross the Flf grown in 19 14, had too few individuals to give
significant results. As a matter of record, however, the results ob-
tained are given in table 64.

Table 64

Width of leaf in millimeters in (1X3) F1} 19 14.

Number

Class

of plants

Average

Pure No. 1 . . . .

151

(1 X 3)

Pure No. 3

From this material there were grown in 191 5, 9 plant rows of No. 1,
six plant rows (two being taken from one of the mother plants) of
1X3 and one plant row of pure No. 3.

Table 65 summarizes the results obtained.

Table 65

Width of leaf in millimeters in (1 X 3) F,, 1915.

Class

Number of
cultures

Number of
individuals

Average
width

Coefficient of
variation of the
population

Average C.V.
of separate
cultures

Pure No. 1 . . . .

651

13.0

10.3

(1 X 3) F2....

406

27.6

25.8

Pure No. 3 . . . .

11.2

Distribution of coefficients of variation.

Class 1 7 1 9 1 11
Uass ! 8 1 10 j 12

13 15

14 ! 16

23
24

25
26

29
30

(1 X 3) F2 |

Pure No. 3 | | 1

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 69

The average leaf width of the hybrids is below that of either parent.
The coefficient of variation of the populations are greater than the aver-
ages of the separate cultures and the variation of the hybrids is greater
than that of the most variable pure culture.

Table 66 gives the distribution of the individuals of the several popu-
lations and the distribution of the means of the separate cultures.

Table 66

Width of leaf in millimeters in (1 X 3) F2, 191 5.

Class

11 1 12

I 2

117

130

107

(1 X 3) F2

3o|30

Pure No. 3

1 1

Distribution of means of cultures.

Class

(1 X 3) F2

Pure No. 3

We first note that, notwithstanding the fact that there were nearly 200
more individuals in the population of No. 1 than in the hybrid popula-
tion, still the range of leaf width among the hybrids extended markedly
beyond the range of pure No. 1 in both directions, and this in spite of the
fact that no single hybrid culture averaged greater than the narrowest-
leaved culture of pure No. 1.

Now analyzing the relation of the F2 hybrid cultures to their (Fx)
parents we find that there is a possibility that there were some differ-
ences in the genetic constitution of the Ft plants inasmuch as the nar-
row-leaved parents produced offspring with a lower average leaf width
than did the wider-leaved parents. This is shown in table 67.

Genetics 4: Ja 1919

1/5

—

l-l

•x.

— oo~

—

— —
-

+ -

0°

02?

ft.

.£
t_

> — '

1/1

1— 1

—

«C

i_r.

—

o u

° £

5 J

2 ,,

PI M O N

IT — LT,

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

Now grouping these cultures according to their mean in 191 5, table
68 gives the average and distribution of the coefficients of variation of
these groups.

Table 68

Width of leaf in millimeters in (1 X 3) F2, 191 5.

Number
of cultures

Average
leaf width
of culture

in 1915

Average
coefficient
of variation

Distribution of
C.V. of
cultures

23
24

25
26

29
30

26.7

26.5

22.0

The coefficients of variation here show a strong decline in variability
in the wider-leaved cultures.

In 1916 there were available for comparison 7 cultures of pure No.
1, 5 of pure No. 3 and 57 cultures of the F3 hybrid 1 X 3- Table 69
summarizes the results obtained.

Table 69

Width of leaf in millimeters in (1 X 3) Fz, 1916.

Coefficient

Class

Number

Total

Average

of variation

Average C. V.

of cultures

number

leaf width

of the

of separate

of plants

population

cultures

Pure No. 1

344

12.0

10. 1

(1 X 3) F3 (33-1)

406

21.3

18.1

(1 X 3) F3 (49-7)

365

24.1

21.4

(1 X 3) F3 (32-1)

1763

26.5

20.9

(1 X 3) F3 (Total)

2534

253

20.5

Pure No. 3

243

12.2

11.4

Genetics 4: Ja 1919

—

O
M

04 |

N
M

0\ M

—

O oc

CO ^

—

1 >

X
: >

O w

PC C

—

CO i-i
m M

—

LT.

l-l

CO
CO

LT.

hs
—

irj
O
M

l-l o

X
CO

vO
r>.

CM
X

CO
CM

t-H

—

C\ o

O H

—

Tj-

—

Class

^ ^< rt

to 0\ 0) ^, •
CO ^J- CO •

. - . to ^

to to to .

/— s O
/-v ^-v co >7

CO co co ^

X X X X «

01 01

Hi MM

£J

t-H CO ^J"

LP.
M

M l-l M CO »-i

M H M 00 O

CO O HI

CO M CO co

CO W 0C

HH M M T*

irj

—

CO O M tf
CO CO 2

^ to to to fn

o' ^ ^ ^ ^o'

2 CO CO CO co 2

g X X X X cj

3 1— » »— 1 1— 1 1— t 3

P-| <W <w W Vw/ £h

10 0

CO CO

CO co

W l-l

co co

M 1— 1 M

0 c

M co

l-l l-l

t>s oc
M M

10 VQ
M M

M M

M CM

MM rj-

H M
M M

O C
i-k M

11 M CO

►H l-l

1-1 CO co In

1*1

1— 1 1— 1

CO Tt
1-1 1-1

11 11 M CO

M M

l-l M O HH M

0, 0

IT. hh CM CO CO

I I I 113
CO On M "q
CO co g

to to to to ^

^ ^ ^ ^ o

co CO co co 2

.« XXXX g

|S HH M 3

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

A study of tables 69 and 70 will show that it is not worth while to
treat separately the 1 X 3 hybrids originating from the different origi-
nal pollinations, since their means and distributions were practically
equal. They will therefore be treated together hereafter.

In table 69 we observed that the average leaf width of the hybrids was
below both the parents. The coefficient of variation was, however, as
usual, markedly higher for the hybrids. From table 70 we note that the
hybrid range in leaf wTidth extends from a single case markedly above
both parents to plants with almost filiform leaves. The different hybrid
groups show practically the same behavior. Whereas 3 hybrid cultures
showed as little variability (coefficients of variation) as the least vari-
able parental culture, more than half were more variable than the most
variable parental culture.

There were 8 hybrid cultures whose mean leaf widths were as great
or greater than the mean for the wider-leaved parent. It is, moreover,
interesting to note that from the hybrids of parents differing, on the
average, only 2 mm in leaf width, there have segregated out races whose
average leaf width differs by 9 mm. The fact that a large part of the
differences in leaf width observed in the F2 generation were genetic, is
shown in table 70 which exhibits the F3 cultures grouped according to
their parental leaf widths.

There is a distinct correlation between parental leaf width and the
mean of the offspring. Whereas the means show a marked range of
distribution in each of the parental groups, this range is never wider

Genetics 4:

Ja 1915

GEO. F. FREEMAN

Table 71

Width of leaf in millimeters in (1 X 3) F&, 1916.
F3 individual plants grouped according to the heights of F2 parents

Number

Leaf width

nf piiltiirpc

<J1 [Jci 1 Cll I

*7
/

j j

1 2

T 1

t r

T 7

T O

9 1

11
+

8
O

/r
0

c
0

4
+

4
O

59
O

39
+

4. "3
to

1 c

19
O

5
+

24
O

15
O

4
O

Means of cultures, 1916

Number

Alean of group

of cultures

1916

leaf width of parent

average leaf width of offspring

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 75

than the fluctuations of the individuals of a pure line. The coefficients
of variation (see table 72) show a distinct though irregular decline to-
ward the wider-leaved parental groups.

Table 72

Width of leaf in millimeters in (l X 3) W*. Coefficients of variation of F3 cul-
tures grouped according to the leaf width of the F2 parents.

Number
of cultures

2
1
1

4
5
11
12
6
6
2

3
2
1
1

Parental
leaf width

9
11

13
14
15
16

17
18

20
22
23
25

Average C. V.
of group

35-5
24.0
14.0
18.8
20.0

23-5
19.0
19.8

19-7
12.0
23-3
18.5
19.0
10.0

This study of variation is made much more distinct by regrouping
the F3 cultures according to their own means in 1916, as m table 73-

Table 73 ,. . , . . 1

, , ^ • -iv „wo ;„ (t V F, 1Q16 Distribution of F3 individuals grouped
Width of leaf in millimeters in (1 X 3) ^3, ,
WX<XX 1 J according to the means of the Fz cultures.

Number
of cultures

Mean
of culture

13
10
10
3
4
2
2

10
11
12
13
14
15
16

17
18

I4!5
2

2
15

4|30
18

29145'

40
O

130

7I19I20I 61

1 1 1 1 j
it

1 1 1 1 1 1

4I 6|'io| 41

I |2I 1
12 %

\ \ 3|
I il I

37
O
100

13 14

3
11

34|3

93 40
O

57 53
Ol

63|84
25|28|33|i2

ioj 4 2
48|i8| 3| 5
19 61 :

3|2o|2o|46|35|27|i5| 7 8
2| 4I 3 7 13 UsH" 6

I to'

3| «
I I

111)

4| 6|i3| 9|23|i5|i8|4M

Genetics 4: Ja 1919

76 GEO. F. FREEMAN

Table 74

Width of leaf in millimeters in (1 X j) F3, 19 16.

Number

Mean of

Distribution

parents

of cultures

F3 cultures

16 17 iS

19! 20

Number

Mean of

Coefficients of variation of F3 cultures
grouped according to the means of the
F, cultures

Average
C V. of

cultures

F3 cultures

1 1

23 25 27

group

24 26 28

33-0

4
8

10
11

1
1

29.8
26.1

21. 1

21.2

19.9

11.7

130

i7o

10.0

A study of table 74 shows very plainly that there is a distinct and
marked segregation of leaf-width factors in the F2 which gives rise to
F3 cultures whose averages reach or exceed the parental means in both
directions. As measured by the coefficient of variation, the variability
of the hybrid cultures clearly decreased as the average leaf width in-
creased. Does this mean that the wide-leaved cultures are more nearly
homozygous (on the average) than the narrow-leaved segregates? If
this were true it would follow that the factors tending to increase leaf
width are recessive and that the genetically narrow-leaved plants were
so on account of dominant inhibitors. This idea is, however, not sup-
ported by the fact that the leaf width of the F1 plants (see tables 51
and 54) which had the maximum of heterozygosity, has leaf widths

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 77

averaging as high or higher than either parent. If leaf-width inhibiting
factors are dominant the maximum narrowness should occur in the F«
plants. If on the other hand these factors exhibited imperfect domi-
nance one would expect the medium races to have a higher variability
than those approaching the extremes. Such, however, is not the case.
We must therefore seek elsewhere for the explanation of this decrease
in variability as the average leaf width of the cultures increases.

Inheritance of leaf width in bread wheat crosses, Sonora (No. 35) X
red Algerian bread wheat (No. 3)
As previously mentioned no pure No. 35 was available for compari-
son with the Ft generation in 1914. A comparison of the leaf width of
pure No. 3 with the (3 X 35) Fi hybrid plants is given in table 75.

Table 75

Width of leaf in millimeters in (3 X 35) Flt 1914.

Number

Average

of plants

leaf width

Pure No. 3

(3 X 35) F,

While the numbers here given are too small to form the basis of defi-
nite conclusions, they at least indicate that the F± hybrids have leaves as
wide as, or wider than, the parents.

These 18 Fx plants gave rise to 18 plant rows of F2 hybrids in 1915
and there were available for comparison with them 1 pure culture of
No. 3, and 4 pure cultures of No. 35. The results may be summarized
as in table 76.

Table 76

Width of leaf in millimeters in (3 X 35) F2, 1915

Class

Number
of cultures

Number of
individuals

Average
leaf width

Coefficient
of variation
of the
population

Average
C. V. of
cultures

Distribution
of C.V.

7
8

9I11

I0|l2

I3|i5
I4|i5

Pure No. 3

1 1.2

1 '

(3 X 35) Fa

1620

13-9

13-4

I2| 2

Pure No. 35

169

13.6

130

i| 2

The mean leaf width of the hybrids is intermediate between the par-
ents. The average variability of the hybrids is only slightly above that
of the pure cultures.

Genetics 4: Ja 1919

GEO. F. FREEMAN

The distribution of the populations and means for this generation are
given in table 77.

Table 77

Width of leaf in millimeters in (3 X 35) F2, 1915.

Distribution of individuals

Pure No. 3
(3 X 35) Fj
Pure No. 35

Distribution of
means of culture

15 1 16

I7|i8

I9|20|21

7| II

1 1

■

II3|lQI

256

259

225

213

127

2|l4

A 1

2| 6

I| 2\ I

It is interesting here to note that the distribution of the means of the
hybrids did not reach the extremes of the parents and that although the
number of hybrids was many times that of Xo. 35, the range of varia-
tion of the hybrids toward wide leaves did not exceed that of its broad-
leaved parent.

For the F3 of this cross there were available 5 pure cultures of each
of Nos. 3 and 35, and 80 plant rows of (3 X 35) F3. The hybrid F2
plants chosen for planting in the fall of 191 5 included 11 of the 19
classes through which the population of F2 was distributed. A first
tabulation of the results follows :

Table 78
Width of leaf in (3 X 35) F3, 19 16.

Class

Xumber
of cultures

Xumber of
individuals

Average
leaf width

Coefficient
of variation
of the
population

Average
coefficient of
variation

Distributio
C. V.

9|n|i3|i5
io|i2|i4|i6

n of

17
18

Pure Xo. 3

243

12.2

11.4

3| I| 1 I

(3 X 35) F3

3852

15-5

12.9

6 1 29 1 28 1 1 1

Pure Xo. 35

246

15.2

13-8

\ \3\2

One is surprised to find here the mean of the F3 hybrids as high as
the broader-leaved parent and the average coefficient of variation of the
separate cultures of hybrids lower than that for the Sonora (No. 35).

The distribution of the individuals in the populations of hybrids and
pure cultures is shown in table 79.

Table 79

Width of leaf in millimeters in (3 X 35) Fs, 1916.

Distribution of individuals

Distribution of means
of cultures

6|7|.8|9

I7| 18

I5|i6

Pure Xo. 3

1 1 |3

7| 4

(3 X 35) F3

3l3|4|l

223

303

583

554

589)392

392

324

207

Pure Xo. 35

1 1 1

27 1 30

" 2

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

It is particularly interesting to note here that there were ten cultures
with means higher than the highest mean for the wide-leaved parent.
We have here a suggestion that if there be some force limiting variabil-
ity in the wider-leaved races it would more strongly affect these wide-
leaved hybrid cultures and thus aid in reducing the average variability of
the group. In this connection it may be remarked that the average co-
efficient of variability of these ten cultures is 11.4 percent, a figure well
below the average coefficient of variability for pure No. 35, which is 13.8
percent.

It is also interesting to note that whereas in the macaroni — bread wheat
crosses many cultures were grown, the average leaf widths of which
were below that of the narrow-leaved parent, here we have no cultures
lower, but there are eight above the wider-leaved parent.

The segregation and recombination of characters by which these
markedly different races were isolated is shown in table 82 where the
F3 individuals are grouped according to the mean leaf width of the F3
cultures.

Table 80

Width of leaf in millimeters in (3 X 35) Fs, 1916. Population grouped according to the leaf width of

the F2 parents.

Number

Parental

Distribution of F3 grouped according to leaf width

Df F2 parents

Average
in 1916

of cultures

width in 1915

14.6

+ 0

14.8

15.6

16.1

124

i6.e

65.

17.2

16.9

102

104

18.1

184

18.5

16.1

This table shows a regular and nearly uniform correlation between
the parental leaf width and the average leaf width of the offspring. The

Genetics 4: J a 1919

GEO. F. FREEMAN

one exception at the wide extreme came from plant No. 21 -5-2-1, a
plant which stood at the end of the row and was very likely an extreme
variant of about the 18 class (see range of this class in table 80).

Table 81

Width of leaf in millimeters in (3 X 35) F3, 19 16.

Xumber
of cultures

Leaf width
of F2
parents

Distribution of means of
F3 cultures grouped ac-
cording to leaf width of
F2 parents

Average
coefficient of
variation of

Distribution of
coefficients of
variation

cultures

15-5

13-3

12.5

13-4

12.6

13-7

12.0

13.2

12.0

130

There is an indication of some decline in the coefficient of variation
in the wider-leaved groups, but it is too much broken up by irregulari-
ties to be of any particular significance.

The study of variability of the F3 is better made, however, by re-
grouping the F3 cultures in accordance with their own means. This is
done in table 82.

Table 82

Width of leaf in millimeters in (3 X 35) Fz, 1916. Population grouped according to
the average leaf width of the F3 cultures.

Xumber
of cultures

Average
leaf width of
F3 cultures

Distribution

leaf

widths

individuals

146

100

209

170

164

126

104

114

150

118

136

129

O = means of F3 groups.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 8l

Comparing tables 78 and 82 we note that, starting with cultures which
differed on an average by 3 mm in leaf width, we have obtained cultures
whose means differ by 6 mm.

Table 83

Width of leaf in millimeters in (3 X 35) 19*6. F2 parents of F3 cultures grouped
according to the means of the F3 cultures.

Number
of cultures

Mean of F3
cultures

Distribution

of F

2 parents

Average
C. V. of F3
cultures

Distribution
of C.V. of
F3 cultures

9
10

13
14

17
iS

13-8

13-4

13.6

I3.I

12.7

11.2

13-5

From table 83 we observe that the range of parents which may give
rise to an offspring with a given mean is not greater than that of a pure
culture.

When the coefficients of variation are calculated we find an irregular
but still quite definite decline toward the wider-leaved cultures as usual
(see table 83).

Algerian red bread (No. 3) X early Baart (No. 34)

This cross will be of special interest for comparison with the other
crosses inasmuch as the two parents had practically the same width of
leaf. The number of plants grown in 1914 are too small to furnish
trustworthy averages but as a matter of record they may be given as
follows :

Table 84

Width of leaf in millimeters in (3 X 34) Fn J9^4-

Class

Number
of plants

Average
leaf width

Distribution of leaf widths

14 | 15 | 16

i8|i9

21 | 22 | 23

Pure No. 3

1 1

" 1

! 1 1

(3 X 34) F,

1 1

3 1 1

Pure No. 34

I 1

* 1 «

3 1 ^ 1

Genetics 4: Ja 1919

- -

K - —

< « ^

01 HI

Number
of

plants

O 00

01
co

duals

01 co 00

—

ca ^

H-l lO

■O

£ indivi

CO 00

01
01

00 M

hh 01

01
CO

—

c to

— -T

c r -c

O to

CO OC

hi fx

ro o
'j —

00 01

-t oc

00 M

—

str

X 01

- ^ PS

> 1*4

O mvO O0 O M M t\f^

to to to" to — co

o ^ C _ _ - _ c

jgr cocococococoZ

"XXXXXX?

CO co co

-7. 5

Distribution of C. V.

C\ o

M d

IN 00

■LT, \£

H N H

t-C 1— 1

CO "<*■

w N

hi \£ i- lO

M <N

O, O | co co

IN OC | M W

Average i
C. V. of
cultures

^ O OC

M 01 M M
— — — -

C. V. of
population

01 co Cn O

pi co oi oi

Average
width of
leaf

rt vC io

_ H* HH

Number
of

individuals

CO IN VD CO
tj- 00 CO —
01 CO co 01
01 01

Number

of
cuhures

IO O O lO
lO

Class

/"N / N

h* 01
i i

CO lO Tj- Tj-
01 TJ- CO

c£ x XfS

CO co

* — - S — '

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 83

These 6 Ft plants gave rise to 6 plant rows of F2 hybrids in 191 5
and there were available for comparison 1 pure culture of each of Nos.
3 and 34. Since the F2 cultures differed somewhat in accordance with
the leaf width of the Fx plants, the records will be given in full rather
than being summarized as usual (table 85).

Here we have the average of the hybrids less variable than either
parent. It should be observed that the one hybrid culture (No. 44-1 |
which was more variable than either parent had a mean lower than
either parent and that the three cultures having means higher than either
parent all had coefficients of variation well below either parent. The
mean of all of the F2 was equal to the wider-leaved parent and the total
range of the F2 was practically confined to the limits of the parental
range. The means of the F2 cultures varied on either side of the par-
ental means but in such cases kept their total range inside of the parental
range by narrowing their own variability.

In view of these rather marked discrepancies in the means of the F2
cultures subsequent study is confined to the progenies of but two Fx
plants (44-2 and 25-1) and these are kept separate.

In 19 16 there were available for study 5 plant rows of each of the
parental cultures, pure N0.3 and pure No. 34, selected from these strains
of the previous season and for the hybrids 50 selections from the F2 of
25-1 and 49 selections from the F2 of 44-2.

Here the means of the hybrids are above the means of either parent
but unlike the F2 the coefficients of variation are slightly above that of
the parental cultures. In table 87 we note that some of the hybrid cul-
tures were more and some were less variable than certain of the pure

Table 87

Width of leaf in millimeters in (3 X 34) Fz, 1916. Distribution of the populations and means

of cultures of hybrids and parents.

Distribution of population

18 1 19

Pure No. 3

30 1 66

4I 2

(3 X 34) F3 (25-1)..

119 327

405

555

297

321 | 109

1 »

(3 X 34) F3 (44-2) . .

188

249(447

514

45i

178

154 48

Pure No. 34

Distribution of

means

of cultures

Pure No. 3

(3 X 34) F3 (25-1)..

(3 X 34) F3 (44-2) . .

Pure No. 34

Gexetics 4: Ja 1919

GEO. F. FREEMAN

cultures. The differences obtained are, however, not large enough to
have any especial significance.

In table 87, the most interesting feature is the distribution of the
means. Here we have 46, approximately half, of the hybrid cultures
with means higher than either of the parents. The same was true in
the F2 cultures (see table 85), As regards height, it will be recalled
that the hybrids of this class also averaged as high or higher than the
taller parent. The fact that so many races had average leaf widths so
strikingly above either parent would suggest recombination with the
production of races beyond the extremes of the parent. This, however,
is made very doubtful by a study of table 88. There the F3 cultures are
grouped according to the leaf width of the F2 parents. Moreover, seeds
were planted from each of the plants of the F2 of the populations of the
cultures concerned (25-1 and 44-2). If therefore the variations in leaf
width of the F2 plants were partially genetic and partially nutritional
(environic) the averages in the F3 groups should show a correlation with
their F2 parents.

We do not seem to have any correlation whatsoever between the leaf
width of the parent and offspring. We may therefore conclude that so
far as this character is concerned the F2 plants were all genetically
equivalent and that all differences such as did arise were modifications.

A study of the distribution of the means of the F3 cultures grouped
according to their F2 parents also confirms the conclusions already
drawn that the F2 plants were all equivalent genetically so far as leaf

2 + ~

« O <n + ^tOo\ Oo O m 0 0

.J. 01 Q CO O

be -O ~

as On

u £ >-.

< 2

e o

i ^

+ 2

0»

w
m

«t

lO
01

+ c>

ir,

On
Q01

or-

+ co

01
01

—

l-H

t/i

C3
CJ

6
II
O

rt
C

GEO. F. FREEMAN

width was concerned. However, both tables indicate that the strain
originating from the original hybrid plant 25-1 had slightly broader
leaves than that originating from the original hybrid plant 44-2.

Table 89

Width of leaf in millimeters in (3 X 34) Fz, igi6.
Population grouped according to the average leaf width of the F3 cultures.

Average

Number

leaf width

Distribution 0

E individuals

of cultures

F3 cultures

24! 25

(25-1)

148

133

132

IOI

163

227

134

144

122

-73

(44-2)

156

135

187

115

239

225

104

O = means of F3 groups.

In order better to study the variability of the F3 generation of this
cross, the plants were regrouped according to the means of the F3 cul-
tures in table 89, and table 90 gives the distribution of the F2 parents
and the coefficients of variation of the F;{ cultures in the same grouping.

The distribution of the F2 parents in this arrangement appears en-
tirely fortuitous without any correlation whatsoever with the means of
the progenies to which they gave rise. These facts therefore form ad-
ditional evidence that the F2 plants were all equivalent genetically and
that all variations of individuals in the F2 or of means of cultures in the
F3 were due to non-genetic factors.

We are unable to detect any significant difference in the coefficients of

HEREDITY OF QUANTITATIVE CHARACTERS EN WHEAT 87

Table 90

h of leaf in millimeters in (3 X 34) F* 1916. F2 parents and coefficient of variation
of F3 cultures grouped according to the means of the F, cultures.

Number
of cultures

Means of
F3 cultures

Distribution

parents' Average
1 r v ~(

Distribution of the
coefficients of variation

| group

7
8

9
10

13
14

15
16

17
18

19
20

(25-1)

13-5

"•5

12. 1

9-4

12.3

(44-2)

12.0

11.8

12.0

10.8

130

variation of the several groups, whether they be observed from the
standpoint of averages or distribution. If, however, the two groups be
combined and the columns be made to include 2 mm range in leaf width
as is done in table 91 (see row for (3 X 34) Fa), we see a slight but
definite decline in variability toward the wider-leaved groups.

Summary; width of leaf

In the 3 X 34 cross, the parents had essentially the same leaf width.
The average of the ¥x was a little below either parent, the F2 exhibited
quite marked differences in the means of the different F2 cultures but
the average of the whole F2 population was the same as that of the
wider-leaved parent. In the F3 the leaves of the hybrids averaged
wider than those of either parent and there were again considerable
differences in the means of the different hybrid cultures (see table 89).
The differences observed, however, are not genetic differences, as is
shown by the fact that there was no correlation whatsoever between the
leaf width of the F2 selected parents and the mean leaf width of their
offspring (see table 88). In other words, the progeny of the different
variants of the F2 gave results such as would come from the fluctnants
of a pure race. We may therefore justly conclude that so far as leaf
width was concerned, the 3 X 34 hybrids formed a pure race. This,

Genetics 4: Ja 1919

GEO. F. FREEMAN

however, does not mean that these hybrids really formed a pure race in
all characters for we have already seen that they segregated in both
height and date of heading. A plant may easily be homozygous for one
character and heterozygous for a number of others. We may assume
therefore that the 3 X 34 hybrids received the same set of leaf-width
factors from both parents. In the subsequent discussions of leaf width
this group will be considered as a single pure variety.

Before proceeding with the summary and discussion of the other
crosses we may first seek to discover whether or not a cause such as we
found to suppress variability in the tall cultures of wheats was also opera-
tive in reducing variability in the wider-leaved cultures. Table 91 brings
together all available data bearing on this point. The horizontal rows
contain the data from plants or groups which were supposed to be ge-
netically equivalent so far as leaf width is concerned.

The results obtained in table 92 are remarkably uniform and exhibit
without doubt some general cause suppressing variability in the broader-
leaved cultures. The nature of this suppression factor is not yet deter-
mined. Three possible explanations are suggested as follows :

(1) Can it be that the coefficient of variation is not a proper measure
of the variability of quantitative characters in biology?

(2) Is it possible that even pure lines of wheat are still somewhat
heterozygous and that the taller cultures are more homozygous than the
others ?

(3) Can there be some physiological limitation of growth in the
higher classes which restricts the full development or expression of the
plus combinations of factors?

The writer is inclined to attribute this suppression factor to a com-
bination of suggestions (1) and (3). If a car be moving at rate A
and we apply an additional force, say F+ffi, which gives an additional
speed say A-\-nt it will require more force than F+2w to give it a speed

Of A— 2)1,

The effect of a factor, environic or genetic, for increasing size, is
probably much less in a combination which tends to produce a variant
above the racial mean than in combinations, the product of which falls
below the mean. We should have, as it were, a telescoping of variabil-
ity in cultures with higher means. It is possible therefore that a better
measure of the variability of quantitative characters would be a coeffi-
cient derived by dividing the standard deviation by some fractional

power of the mean, thus d = — — — where x is a quantity less than 1.
Returning to the macaroni — bread wheat crosses we remember that

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT 89

Table 91

Correlation of average leaf width of culture and the coefficient of variation of the same
in pure lines and genetically equivalent groups.

Leaf width in millimeters

Culture

Total
number

13
14

15
16

17
18

19
20

Pure No. 1

Xo. of cultures
Average C. V.

11.0

8
10.6

4
99

9.0

Pure Xo. 3

1 Average C. V.

II.8

10.5

Pure No. 34

Average C. V.

12.0

113

11.0

Xo. of cultures
Average C. V.

Average C. V.

14-5

137

14.7

8.0

(1 X 35) F2

30.6

3i
29.1

27.0

Average C. V.

26.7

25.0

(3 X 34) F2

Average C. V.

10.7

3
9.0

(3 X 35) F2

Xo. of cultures
Average C. V.

16
13-6

12.5

(3 X 34) F3

Xo. of cultures
Average C. V.

18
12.0

63
11.8

18
10.3

F3 cultures from tall F2 plants
having smooth seeds (parent-
like) (1 X 35) F3

Xo. of cultures
Average C. V.

8
19. 1

12
17.9

14
15.6

14-5

F3 cultures from tall F2 plants
having wrinkled seeds (Fx-
like) (1 X 35) F3

Xo. of cultures
Average C. V.

• I

30.0

26.5

16
22.8

20.5

130

F3 cultures from tall F, plants
having smooth seeds (parent-
like) (1 X 3) F.

Xo. of cultures
Average C. V.

9 1

19.0

16.5

"•3

1 0.0

F3 cultures from tall F2 plants
having wrinkled seeds (Fx-
like plants) (1 X 3) F3

Xo. of cultures
Average C. V.

9 1 ■

| 20.0

27.8 1 26.7

25-0 1

the F1 had wide leaves and wrinkled grains. The average leaf width
of the F2 was markedly below that of either parent but there were some
F2 plants having leaf widths as great or greater than the parental means.
These wide-leaved F3 plants were of three types, viz., (1) some had
wide leaves and smooth grains (parent-like), (2) some had wide leaves
and wrinkled grains (Fj-like^ and a few had wide leaves and partially
wrinkled grains (of uncertain classification). Now since the average

Genetics 4: Ja 1919

GEO. F. FREEMAN

of the F2 was below that of the parents and the variability was much
above the parental variability, we should expect the F-L-like F2 plants
to give F2 cultures low in mean leaf width and high in variability, where-
as the parent-like F2 plants should give F3 cultures high in mean leaf
width and low in variability. Xow disregarding the wide-leaved F2
plants with partially wrinkled seed (on account of difficulty of classifi-
cation) we find the results shown in table 92.

Table 92

(1 X 35) F3

(i X 3) F3

Number
of cultures

Mean
leaf width

Average
C. V.

Number
of cultures

Mean
leaf width

Average
C. V.

F3 cultures from wide-
leaved smooth-seed-
ed F-j plants (par-
ent-like)

16. 1

17.1

154

12.9

F3 cultures from wide-
leaved wrinkled-
seeded F2 plants
(Fj-like)

14.9

23.S

12.4

26.2

Xo better agreement of the facts with the theoretical assumptions
made, could well be expected. It is, of course, not here assumed that the
parent-like F2 plants were constituted genetically exactly like one or the
other of the parents or that the Fi-like F2 plants were completely hete-
rozygous in every particular in which the Fx plants were heterozygous,
but it is assumed that the genetic agreement is close enough to give
marked similarity in form and hereditary behavior. Where a number
of factors are involved, as there probably are here, it would be extremely
difficult, probably impossible, to pick out plants from the F2 by inspec-
tion, which were exactly like either the parents or the Fu genetically.
This could only be done by judging the F2 plants by the genetic be-
havior of their offspring. The facts developed seem to show that the
wide-leaved F2 plants fell into two groups, the one having a complete
(or nearly complete) set of the factors from one or the other of the
parental races, and that the other group contained plants which were
heterozygous for all (or nearly all) of the characters in which the par-
ents differed. Again therefore we have a situation where a complete
double set of one or the other of the parental races or a complete (or
nearly complete) single set from each of the two parents were able to

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

produce wide-leaved plants, but that the large majority of the new re-
combinations of parental characters resulted in less vegetative develop-
ment.

Now referring to table 92 we find that even in the offspring of these
two groups of wide-leaved F2 plants the factor for suppressing vari-
ability was apparent, but it was not sufficient to mask the effect of dif-
ferences in heterozygosity because in the one c>ase (those of the Ivlike
parents) the means tended to be below that of the standard (pure line
parents). Now when we turn to the 3 X 35 cross where the Fl9 F2
and F3 all had average leaf widths larger than the more narrow-leaved
parent, the suppression factor was able entirely to offset the theoretically
expected increased variability of the heterozygous cultures. If in ac-
cordance with the Fj, the wide-leaved F3 cultures were the more hete-
rozygous and the more narrow-leaved the more homozygous we can
easily see how the suppression factor might reduce the average vari-
ability of all of the F3 cultures to a figure equal to or below that of the
most variable parent especially in a case where the average of the leaf
width of the F3 cultures was equal to that of the wider-leaved parent.

One cannot here assume the formation of a single new blended race,
for table 75 shows segregation in the F2 with the formation of many
distinct races in F3, and moreover, in spite of the suppression factor and
the fact that the F2 had a larger mean than the F3, the average vari-
ability of the F3 was less than that of the F2 (compare tables 76 and 78).

According to Mendelian expectation, the parental types of individuals
in F2 and culture means in F3 were recovered in all cases. In 1 X 35>
recombination formed individuals in F2 and a number of cultures in F3
whose means were significantly beyond, both above and below, the range
of either parent. In 1 X 3 the range of individuals in Fx and of means
of cultures in F3 were significantly below, but not above, the parental
ranges. In 3 X 35 the range of individuals in F2 and means of cultures
in F3 were not significantly above or below the parental ranges.

In the macaroni— bread wheat crosses the average variability of the
F2 and F3 generations were markedly above that of the parents but in
the F3 many cultures were secured which were as little variable as either
parent. In no case was there a single F2 culture, however, which had
as low a variability as the most variable parental culture.

The variability of the bread wheat crosses has already been discussed
with sufficient fullness.

The segregation of simple Mendelian unit factors appears to suffice to

Genetics 4:

Ja 1919

GEO. F. FREEMAN

explain all of the facts so far observed in the inheritance of leaf width
in the wheat hybrids here discussed. No attempt has been made to de-
termine the number of factors but the supposition is that there are
several.

GENERAL SUMMARY

Detailed summaries of the three characters, date of first head, height,
and width of leaf, may be found on pages 27, 52 and 87, respectively.

The F1 of the macaroni — bread wheat crosses developed normally and
were in every case equal or superior to the mean of the parents in vege-
tative vigor and they were no more variable in size characters or time
of maturity than were the pure races. We may therefore conclude that
a single complete set of macaroni wheat characters with a complete single
set of bread wheat characters (the maximum of heterozygosis between
the two varieties) will produce a perfectly normal plant.

In the second generation, on the other hand, many of the seeds would
not germinate and those germinating produced plants differing in vege-
tative growth from those which were more vigorous than either parent
to such as never got beyond the rosette stage. Moreover those which
made a normal vegetative development exhibited every degree of sterility
from completely sterile plants to those entirely normal in seed produc-
tion. It would appear, therefore, that these facts alone refute any idea
of blending inheritance, for if blending had taken place in the F1? sterile
or vegetatively deficient plants would be no more likely to occur in the
Fo than in the Fx. Hence we are compelled to predicate segregation and
recombination in these quantitative characters. There is nothing to in-
dicate even partial blending in any of the factors concerned.

In the use of the coefficient of variation as an indication of heterozy-
gosity in hybrids involving quantitative characters, care should be exer-
cised to make due allowance for the fact that races with high means re-
sulting from increased vegetative growth, have their variability limited
or reduced by the apparent law that size factors are more effective in
producing variability in combinations tending to produce a result below
the mean of the hybrid population than in combinations which tend to
exceed this mean.

The suppression of variability in cultures with high means applies to
pure as well as hybrid cultures. It appears to be a telescoping of vari-
ability as the mean approaches the upper physiological limit of growth
rate for the species concerned.

HEREDITY OF QUANTITATIVE CHARACTERS IN WHEAT

LITERATURE CITED

Castle, W. E., 1912 The inconstancy of unit-characters. Amer. Xat. 46 : 352-362.

1917 Piebald rats and multiple factors. Amer. Nat. 51:102-114.
East, E. M., 1916 a Studies on size inheritance in Nicotiana. Genetics 1: 164-176.

1916 b Inheritance in crosses between Nicotiana Langsdorffi and Nicotiana alata.
Genetics 1 : 311-333.

Fruwirth, C, 1915 Versuche zur Wirkung der Auslese. Zeitschr. f. Pflanzcnzuch-
tung 3 : 173 and 395.

Hayes, H. K., and East, E. M., 1915 Further experiments on inheritance in maize.

Conn. Agric. Exp. Sta. Bull. 188, 31 pp.
Hoshino, Yuzo, 1915 On the inheritance of flowering time in peas and rice. Jour.

College Agric. Tohoku Imp. Univ., Sapporo 6 : 229-228.
Mac Dowell, E. C, 1914 Size inheritance in rabbits with a prefatory note and

appendix by W. E. Castle. Carnegie Inst, of Washington, Pub. No. 196, 55 pp.
1915 Bristle inheritance in Drosophila. Jour. Exp. Zool. 19:61-98.
Nilsson-Ehle, H., 1914 t)ber einen als Hemmungsfaktor der Begrannung auftreten-

den Farbenfaktor beim Hafer. Zeitschr. f. indukt. Abstamm. u. Vererb.

12:36-55'.

Phillips, J. C, 1914 A further study of size inheritance in ducks with observations

on the sex ratio of rrybrid birds. Jour. Exp. Zool., 16:131-148.
1915 Experimental studies of hybridization among ducks and pheasants. Jour.

Exp. Zool. 18:69-112.
Punnett, R. C, and Bailey, P. G., 1914 On the inheritance of weight in poultry.

Jour. Genetics 4 : 23-39.
Shull, G. H., 1914 Duplicate genes for capsule-form in Bursa bursa-pastoris. Zeitschr.

f. indukt. Abstamm. u. Vererb. 12:97-149.

Genetics 4: Ja 1919

Information for Contributors

The length of contribution! to Genetics will not be arbitrarily limited, but articc
of less than fifty pages may be expected to appear more promptly than those of greater
length. The limitation* as to subject-matter will be construed broadly, so that any
paper of sufficient importance, whether in the field of genetics proper, of cytology,
embryology, physiology, biometry, or mathematics, if of such character as to be of
primary interest to the geneticist, will find here an appropriate place of publication,

Contributors are requested to observe the usual care in the preparation of manu-
scripts. All references to the literature should cite the name of the author, followed
by the year of publication, the papers so referred to being collected into a list of
"Literature cited" at the end of the article. In this list great care should be taken to
give titles in full, and to indicate accurately in Arabic numerals, the volume number,
the first and last pages and the date of publication of each paper, if published ii a
periodical, and the number of pages, place and date of publication and name of
publisher, of each independent publication. The arrangement of this list should be
alphabetical by authors, and chronological under each author.

Footnotes should be inserted immediately after the reference marks which refer tt
them, and should be separated from the text by lines. All footnotes should be nu»
bered consecutively in a single series.

No limitation will be placed on the kind of illustrations to be used, but as funds fot
illustration are still very limited, authors are requested to practice restraint, and to
limit themselves to such amount and grade of illustration as seem positively essential
to an adequate understanding of their work. The size of the type-forms of Genetics
is 4H X 7V2 inches and all illustrations intended for use as single plates should not
exceed 5 X lY\ inches, or should allow of reduction to these dimensions. Double
plates may be I2# X Wa inches. Legends for all figures should be typewritten on a
separate sheet

Manuscripts and all editorial correspondence should be addressed to the Managing
Editor, Dr. George H. Shull, 60 Jefferson Road, Princeton, N. J.

GENETICS, JANUARY 1919

TABLE OF CONTENTS
Portrait of Hugo de Vries Frontispiece

Freeman, Geo. F., Heredity of quantitative characters in wheat. ... I

Our Frontispiece

On February 16, 191 8, Hugo de Vries celebrated his seventieth
birthday, and on June 13, 1918, he delivered his last lecture as Pro-
fessor of Botany in the University of Amsterdam. In recognition
of his long and distinguished career his birthday was marked by the
presentation of an album containing the autographed photographs of
his botanical friends and admirers the world over, and arrangements
were made for the collection and reprinting in six large volumes, of
all his contributions to scientific journals. It is particularly fitting
therefore to present this portrait of Hugo de Vries at this time.
The photograph used here is one of several taken by Elliott & Fry, Ltd.,
London, on the occasion of the Darwin Centennial Celebration
in 1909. It is copyrighted by the photographers, from whom the right
to publish here has been purchased. The editor is not aware that this
photograph has been previously engraved; but one of the other two
photographs taken at the same time, is used as frontispiece in the col-
lected papers whose publication is mentioned above.

The reproduction of this portrait of Professor de Vries is made
possible by a gift from Dr. Liberty Hyde Bailey, who was for many
years Director of the N. Y. State College of Agriculture at Cor-
nell University, and who is Author and Editor of many important
works on horticulture and agriculture, including such works of special
interest to geneticists, as "Plant breeding", "Survival of the unlike,"
"Evolution of our native fruits" etc. In the title of the first of these
books, first published in 1895, tne expression "plant breeding" was used
probably for the first time.

Date Due

' i

■

Library Bureau Cat No. 1137

QH433.H3:2

3 5002 00107 7903

Harvard University.

Contributions from the Laboratory of Pla

SCIENCE

02 2
433

H3 188915

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Contributions from the Laboratory of Plant Genetics"

See other formats