THE CHROMOSOMES AND RELATIONSHIP OF
CREPIS SYRIACA (BORNM.)

BY
DONALD ROSS CAMERON

University of California Publications in Agricultural Sciences

Volume 6, No. 10, pp. 257-286, plates 12 and 13, 3 figures in text

Issued October 11, 1934

Price, 50 cents

University of California Press
Berkeley, California

Cambridge University Press
London, England

PRINTED IN THE UNITED STATES OF AMERICA

THE CHROMOSOMES AND RELATIONSHIP OF
CREPIS SYRIACA (BORNM.)*

BY

DONALD BOSS CAMERON

INTRODUCTION

Although it has long been accepted that chromosome numbers are
highly constant within species, a great many occurrences of deviation
have been cited. It is very rarely, however, that a species is found in
which variation is the rule rather than the exception. Such a discovery
was reported by Hollingshead and Babcock (1930) in Crepis. While
examining and comparing the chromosomes of a large number of species
within this genus, individual plants of C. syriaca were found to possess
from ten to thirteen chromosomes in somatic cells. The extra chromo-
somes, where present, were morphologically similar to one another, but
different from any members of the basic group of ten. This condition in
a species in which chromosome morphology is so well marked made it
most favorable material for the investigation of supernumerary chro-
mosomes. Other particularly advantageous characteristics of the species
are the relatively low chromosome number and the fertility of plants
with variant genoms.

The taxonomic relationship of C. syriaca is also of interest, particu-
larly with respect to C. alpina L., of which it was formerly held to be a
variety. Both species are in the same section of the subgenus Barkhausia,
being most closely related to Crepis rubra L. and C. foetida L. of the
same section (Babcock and Lesley, 1926, figs. 1 and 2, and table 2; Bab-
cock and Navashin, 1930) . In general appearance they are similar, yet
there are many constant differences which serve to distinguish them as
species. The more outstanding differences are tabulated as follows :

C. syriaca C. alpina

1. Habit low, spreading. 1. Habit tall, erect.

2. Herbage light green, lacking tomen- 2. Herbage gray, tomentose.

turn. 3. Radical leaves obovate-oblong, ob-

3. Radical leaves oblanceolate, acute, tuse, denticulate, sometimes with
dentate or runcinate-pinnatifid or 3-4 irregular shallow lobes near
pinnately parted, the segments apex.

acute, dentate. 4. Cauline leaves entire or denticulate

4. Cauline leaves laciniate at base. at base.

* Crepis syriaca (Bornm.) Babcock, MS.

[257]

258

University of California Publications in Agricultural Sciences [Vol. 6

C. syriaca — (Continued)

5. Heads nodding before anthesis.

6. Heads fully expanded in anthesis.

7. Ligules deep yellow.

8. Anther-tube appendages 0.7 mm.
long, obtuse.

9. Achenes about 14 mm. long, not al-
ways sharply divided into two types.

10. Pappus 4.5—5.5 mm. long.

11. Plants flower about 108 days after
planting.

C. alpina — (Continued)

5. Heads erect before anthesis.

6. Heads partly expanded in anthesis.

7. Ligules pale yellow.

8. Anther-tube appendages 0.4—0.5 mm.
long, acute.

9. Achenes about 18 mm. long, of two
distinct types, marginal and inner.

10. Pappus 6-7 mm. long.

11. Plants flower about 135 days after
planting.

Although, in respect to their comparative morphology, there appears
little doubt of the correctness of recognizing these plants as separate
species, still their relationship is so close that a cytogenetieal analysis is
warranted. Individuals from the two species were therefore crossed and
the later generations subjected to a cytological and less extensive geneti-
cal investigation.

These problems were suggested by Professor E. B. Babcock, to whom
the writer is indebted for constant assistance and advice. It is also a
pleasure to acknowledge the generous cooperation of Professor R. E.
Clausen throughout the course of the work.

MATERIALS AND METHODS

The majority of the Crepis syriaca plants used were derived from seed
from herbarium material (accession number 1923). The original speci-
mens were collected by M. Chijik, in Galilee, in 1924. Several plants were
grown from this source by the writer in 1929 and from a study of these
and their progeny most of the results here presented were obtained.
Later accessions which were also used are numbers 3106, 3132, 3158,
3159, and 3167 from northern Lebanon near Hasroun, and 3168 from
Antilebanon, near Baalbek. Accessions of C. alpina used for comparison
and for hybridizing are numbers 1640 from the Copenhagen Botanic
Garden, 1641 from Tiflis (indigenous in the Caucasus region), 2769
from the Leningrad Botanic Garden, and 2783 from the Charkow Bo-
tanic Garden.

Some variation in external morphology occurs among these accessions
of C. alpina, especially between individuals of the Tiflis type and those
from localities in Asia Minor, Kurdistan, western Persia, Transcaucasia,
and Crimea. The former plants are sufficiently different to be classed
as a subspecies of C. alpina. In habit they are somewhat more inclined to
branch from the base of the plant, a character which is even more notice-
able in C. syriaca. The radical leaves are acute, rather than obtuse, and
are not so gray as the typical representatives of the species. The invo-
lucral bracts are purple-tipped, a character which is transmitted to all

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 259

Fx plants in crosses with C. syriaca. The corollas are a brighter yellow
color and the flower heads are, in general, smaller with shorter ligules.
The achenes are of about the same length as typical C. alpina, but the
beak is longer, in this respect also resembling C. syriaca. As will be
pointed out later, these characters are of interest in connection with the
possible origin of the latter species.

Root-tips obtained from plants five to ten days after transplanting to
6-inch pots were fixed in chrom-acetic-formalin solutions (cf. Hollings-
head and Babcock, 1930). They were imbedded in paraffin and cut 8/i.
thick. Buds were fixed in acetic alcohol (2 parts absolute alcohol, 1 part
glacial acetic acid) and were examined from 70 per cent alcohol in iron-
aceto-carmine.

Figures were drawn with the aid of a camera lucida at a magnification
of 3750, using a Zeiss 1.8 mm. n.a. 1.3 oil immersion objective and Zeiss
12x compensating ocular.

THE SOMATIC CHROMOSOMES OF CREPIS SYRIACA

Aside from the supernumerary units present in C. syriaca, the chromo-
somes of this species are virtually identical with those of C. alpina
(2n = 10). This fact coupled with the instability of plants with higher
numbers is very good evidence that 10 is the basic diploid complement
of C. syriaca.

A preliminary survey of the chromosomes of C. syriaca was presented
by Hollingshead and Babcock (1930). Since this study was made, how-
ever, new chromosome numbers have appeared in the progeny of self-
fertilized lines. Plants derived from accession number 1923, which had
fourteen chromosomes or less, show five pairs of chromosomes con-
sistently (see fig. lb ) resembling those of C. alpina very closely. In addi-
tion, there is a characteristic supernumerary unit which is present one,
two, three, or four times in 11-, 12-, 13-, and 14-chromosome plants
respectively. This extra chromosome appears as a small unit, possessing
a subterminal constriction and bearing a large satellite at the proximal
end (fig. lo-X). In plants with fifteen, sixteen, and eighteen chromo-
somes respectively, the increase seems to involve more than this X chro-
mosome. For example, in fig. 2b, showing a complement of fifteen chro-
mosomes, there appear to be two extra E chromosomes and three X's
(cf. fig. la). This may be due to a nondisjunction or noncon junction in
members of the basic complement and hence may be of rare occurrence.
The chromosome groups containing sixteen and eighteen chromosomes
(fig. 2c and d) do not show all the X chromosomes as pictured in figure
la. There is little doubt, however, that the extra members are of this
same type. Throughout the investigation, it was found that the difficulty
in obtaining clear plates was greater in forms with higher numbers.

260

University of California Publications in Agricultural Sciences [Vol. 6

The drawings shown here were made with the primary purpose of
depicting the variation in number, since only composite representations
could indicate the complete morphology of the various groups.

Jill-

A B C D E

J

3A

Fig. 1. Crepis syriaca. a, somatic chromosomes (haploid set) ; b, 1923 B6-2-8 (ten
chromosomes) ; c, 3106 Kl (ten chromosomes) — note the large satellites on two non-
homologous chromosomes; d, 1923 A 11-2-3 (eleven chromosomes), one supernumer-
ary X chromosome; e, 1923 B7-1-3-1 (twelve chromosomes) ; /, 1923 B7-1-2 (thir-
teen chromosomes).

1934] Cameron : Chromosomes and Relationship of Crepis syriaca 261

>'ff W

w.

Fig. 2. Crepis syriaca. a, 1923 B7-3-1 (fourteen chromosomes) ; o, 1923 A 11-4-7
(fifteen chromosomes) ; c, 1923 A 11—4—6 (sixteen chromosomes) ; d, 1923 A 11—4—2
(eighteen chromosomes) ; e, 1923 A 11—4—6. Chromosomes of tetraploid cell, 32. The
units in the lower right portion were in an adjacent section and some were obviously
cut.

262 University of California Publications in Agricultural Sciences [Vol. 6

In the 16-chromosome plant obtained, a very high frequency of tetra-
ploid cells occurred. The chromosomes of such a cell are shown in figure
2e. Those pictured at the lower right were from another section and some
were obviously only a portion of chromosomes already drawn. A total of
thirty-two units was counted in several cells and many others showed
approximately the same number. Many binucleate cells were noted in
this plant. Another unique circumstance was the presence of a large
number of cells which exhibited profile views of metaphase chromosomes,
with a consequent displacement of the usual regularity of cell layers.
This was the only example of any abnormality in somatic mitosis which
appeared during the study.

A much greater variability appeared in the cultures designated 3106.
Figure lc shows the chromosomes of a 10-chromosome plant from this
accession. Instead of the usual five pairs ordinarily found in other 10-
chromosome plants, a new abnormality appeared, involving the C and
E chromosomes. Two large fragments of chromatin are attached to mem-
bers of these pairs, upsetting the balance of the genom. All the plants
derived from accession number 3106 appeared similar in their external
morphology; yet there was not only a variation in chromosome number
from plant to plant, but also variability in chromosome morphology
within a single individual. So far no occurrences of variation in number
of chromosomes among the cells of a single plant have been distinguished,
nor do the two large pairs, A and B, ever become modified. In agreement
with this observation Carothers (1917) suggests that, so far as she has
observed, smaller chromosomes do not have a very pronounced effect
when behaving irregularly, but that such irregularities associated with
larger ones have a deleterious result. Among the varying chromosome
groups in the cultures from accession number 3106, several cells were
observed which contained a unit very similar in appearance to the X
chromosome described earlier. It is conceivable that forms such as these
may represent an intermediate step between normal 10-chromosome
plants and the types described in the cultures grown from accession 1923.

VARIATION IN CHROMOSOME NUMBER

Nearly all the accessions that have been studied up to the present time
have shown variation, although the amount of material available in some
was quite limited. Plants grown from various accessions have been found
to exhibit chromosome variability as follows: (a) variation in number
between plants — 1923; (b) variation in morphology within plants —
3106, 3132, 3167, 3158; (c) variation in both number and morphology —
3159; (d) a single plant of accession number 3168 showed the normal
ten chromosomes of C. syriaca.

It is of course entirely possible that further variations in chromosome

1934]

Cameron: Chromosomes and Relationship of Crepis syriaca

263

number may be found in accessions other than 1923, since this one was
studied much more intensively than the others.

In 1929 a number of cultures were grown from open-pollinated seed
collected from plants grown from original accession number 1923. Table
1 shows the results of chromosome counts of plants obtained from this
source. Among the cultures classified as derived from 10-chromosome
parents, several were grown from seed collected from plants of which
the chromosome number had not been determined. However, from the
situation in the progeny there was no doubt concerning their chromo-
some complements.

TABLE 1
Frequencies of Plants Having Different Chromosome Numbers

Parental plants

Frequencies of plants obtained with numbers of:

Number of
plants

Number of
chromo-
somes

10

11

12

13

7
2
2
1

10
11
12
13

81
5
1

2
12

6
2
8
1

1
1

4

It is significant that, among the progeny, by far the greater number of
plants had ten chromosomes. Since the parental plants were selected at
random it is obvious that in the parental population 10-chromosome
plants were more numerous than those with higher numbers. In each
progeny group there was a marked tendency to reproduce the parental
chromosome number. Among the progeny of 10-chromosome parents
the variant numbers were 11 and 12; while the progeny of 11- and 12-
chromosome parents varied in both directions from the parental number.
In no plant was a complement of less than ten chromosomes found.

The foregoing results seem to indicate that certain lines have a greater
tendency to vary than others. Thus, referring to table 2 of Hollingshead
and Babcock (1930), the parent plant 1923-2 showed much more varia-
tion in its progeny than the plant or plants giving rise to culture 1923-A.
From the results outlined here, it is altogether likely that plant 1923-2
had twelve chromosomes, while the parents of culture 1923-A had ten.
Further similar evidence will appear later. Additional determinations of
plants from accession number 1923 amplify the last line of this pre-
viously published table as follows :

Chromosome number
10 11 12 13

27. 1923-2 Open-pollinated

264

University of California Publications in Agricultural Sciences [Vol. 6

In order to eliminate the possibility that fluctuations were brought
about by crosses between plants which differ in chromosome number, in-
dividual selections were selfed to produce the later generations. Thus,
except where specifically designated, the remaining data were obtained
from self-fertilized lines.

TABLE 2
Frequencies of Plants (from Selfing) with Different Chromosome Numbers

Year

Pedigree

Chromo-
some
number
of parent

Frequencies of plants with numbers of:

grown

10

11

12

13

14

15

16

18

1930

A2-10

10
10
10
10
10
10
10
10

11
11

12
12
12
12
13
13
14

2
5
8
3
9
9
2
5
1

3
3
1

2

1

5

6

1

4

1
3
4

2

1

1
1

2
2

5

1

3
1

1

2

1930

A2-15

1930

B6-2

1931
1931
1931

A2-10-2
A2-15-5
B6-2-5 .

1932
1932
1930

A2-10-1
A2-15-5-2....

All-2

1931
1930

Al 1-2-8
B7-3

1931
1931
1931

Al 1-2-1
Al 1-2-6
B7-8

1931

All-4 ...

1

1932
1932

All-4-3
All-4-4

It is unfortunate that larger numbers of plants could not be examined.
This was often impossible because of the small amount of seed set by cer-
tain plants and the inviability of plants with particular somatic comple-
ments. In other examinations larger numbers would not have contri-
buted any valuable information since the chromosome numbers could be
predicted from previous determinations. In 1932 all the progeny from
an 11-chromosome individual died after reaching the rosette stage. Only
one survived until root-tips could be collected and these were so poor
that an accurate determination was impossible. At the same time an en-
tire culture of plants from a 10-chromosome parent succumbed. Appar-
ently there was a factor influencing the physiology of these plants, mak-
ing them more susceptible to slightly adverse environmental conditions.
Other cultures, alternating with these, remained perfectly healthy.

Table 2 shows the frequencies of plants having different chromosome
numbers in self-fertilized lines. In all the cultures derived from 10-
chromosome parents, only one variant appeared. This 11-chromosome
individual may have been the result of an accidental cross or of a mix-
ture of seed, but at any rate a very high degree of constancy is revealed.

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 265

This result is only to be expected if ten is the "normal" somatic comple-
ment of the species. In the progenies obtained from open-pollinated
plants, the majority of individuals have the same chromosome count as
their parents. In self ed lines, however, this tendency grows progressively
less marked with the increase in chromosome number. It is still quite
noticeable in the 10- and 11 -chromosome series, but in 12- and 13-chromo-
some lines the tendency is toward the production of still higher numbers.
In the culture from an open-pollinated 13-chromosome plant, for ex-
ample, four 13's and one 12 were obtained. When another 13-chromosome
plant was selfed, individuals with thirteen, fourteen, fifteen, sixteen,
and eighteen chromosomes were obtained. This, of course, may have
been owing to differences in the parental plants. In general, there did
not appear to be any tendency on the part of plants with larger numbers
to return to the 10-chromosome type, although, when this condition was
reached, it was virtually always maintained.

RELATION OF CHROMOSOME NUMBER TO EXTERNAL

MORPHOLOGY

In respect to the external characters exhibited by the plant, the extra
chromosomes appear to have small and inconsistent effects. In general,
in plants having from two to eight supernumeraries, the flower heads are
small and asymmetrical, and frequently have split ligules. The 10- and
11-chromosome individuals are usually normal in this respect although
some individuals among the plants possessing eleven chromosomes have
narrower ligules with longer teeth than in the normal forms (plate 12,
figs. 2 and 3) . There is also a tendency for plants with higher numbers
to have broad, coarse leaves with margins which are merely toothed
instead of margins more deeply cut, as in the 10-chromosome types.
This alteration seems to depend more on the parental complement than
on the number of chromosomes actually contained in plants with the
type of foliage described. The difference is brought out clearly in the
1932 cultures. In the progeny of a 13-chromosome plant, two individuals
appeared having ten somatic chromosomes. These were similar to normal
individuals except in their broad radical leaves and generally more ro-
bust habit. There is here an indication that other changes besides gross
morphological ones may be taking place in the chromosomes. As far as
can be seen from a microscopic examination, the chromosomes appear
identical in all 10-chromosome plants, regardless of their origins. On
the whole, the 10-chromosome individuals are of a paler green color
than are the plants containing supernumeraries, but there is sufficient
variability to make a thorough color investigation of little value.

Another occasional abnormality is the development of an irregular
rosette. This occurs in individual plants and is associated with eleven,

266 University of California Publications in Agricultural Sciences [Vol. 6

fourteen, and sixteen somatic chromosomes. In spite of their morpho-
logical irregularity, these plants show uniform reduction divisions and
produce a high percentage of good pollen. No plants of this type occur
in the lines characterized by the possession of ten chromosomes. In sev-
eral of the irregular forms the rosette has two distinct centers of growth,
giving to some of the leaves the appearance of being inverted.

Most of the cultures derived from 10-chromosome plants were con-
sistently uniform in external morphology. However, in 1930 an inter-
esting new type appeared among them. In two cultures, five small plants
appeared, which were approximately 15 cm. high as compared with
45 cm. for their normal sibs. The contrast between these two types can
readily be seen by a comparison of plate 12, figures 1 and 2. Little at-
tention was given these plants at the time, since occasionally dwarfing
occurs as a result of unfavorable growing conditions. In 1931, however,
a culture of ten plants was grown from one of these variants, all of which
were of the same dwarf habit. From another individual, in the same year,
progeny were grown which showed segregation for this character. The
tallest plant in the culture was 50 cm. in height, the smallest about 15 cm.
The nine plants of this culture which reached maturity could readily be
grouped into two size-classes of 7 tall : 2 dwarf. Although these numbers
are far from adequate as statistical data, it seems altogether likely that
the first selected plant referred to must have been homozygous, and the
second, heterozygous for the character. Hence, it is probable that an in-
termediate height condition appears in the heterozygote, as it was a rela-
tively small plant that gave rise to the segregating population. In 1932
a culture was grown from one of the dwarf segregants, of which all the
plants were dwarf. In order to determine the nature of the genes asso-
ciated with the new character, crosses were made between the two types.
The resulting seeds were inviable.

MEIOSIS IN CREPIS SYRIACA

A study of microspore mother-cells in this genus is attended with more
than ordinary difficulty. The chromosomes are very large and irregular
in shape, making an analysis of the meiotic behavior not an easy matter.
The majority of cells in smear preparations present a profile view of the
metaphase plate and the total number of cells obtained from one floret
is small.

The 10-chromosome plants showed five pairs at I M in all preparations
where a definite determination was possible. In other Crepis species which
have been examined, a small amount of variation has been seen, so that
it is very likely that a similar condition would occur here if the studies
were prolonged. In good preparations a large number of cells were seen
which had the appearance of the one represented in figure 3a. Figure 3b

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 267

shows a drawing of a plate typical of the 11-chromosome forms. Six
units were consistently present involving five bivalents and a univalent.
Usually the bivalents behaved normally, although at times one of these
units was seen to divide precociously. The univalent divided at the first
division and at the second the halves separated to either pole. Thus,
apparently 5- and 6-chromosome gametes were regularly produced by
this type of plant. On this basis an 11-chromosome plant when selfed
should yield 10- 11- and 12-chromosome progeny in the ratio of 1 :2:1
if a similar situation exists in the embryo sac. Table 2 shows, however,
that the distribution is not of this type.

The discrepancy between observed and expected frequencies of 10-,
11-, and 12-chromosome plants derived from 11-chromosome parents
may result from a qualitative difference in the chromosomes involved.
There seems to be a definite tendency for plants with more than five
pairs of chromosomes to yield progeny with more supernumeraries than
they possess. This fact leads to the suggestion that the supernumerary
unit may possess genetic material which is of considerable value in
development. However, 10-chromosome plants show little or no inclina-
tion toward yielding progeny with larger numbers of chromosomes. Let
us assume that the X chromosome arose from a reorganization of the
chromatin of the basic five pairs by fragmentation and attachment. Then
the portion of genetic material contributed would be duplicated in the
genoms of plants with more than ten chromosomes. If this portion were
associated with increased vigor or viability, it might reasonably be ex-
pected that 6-chromosome gametes of an 11-chromosome plant would
function at the expense of 5-chromosome gametes and hence would lead
to a preponderance of higher-numbered forms.

In several preparations there was an aberrant behavior of the divided
univalent at I A in 11-chromosome plants. Instead of the halves going
to opposite poles, they were both seen in close proximity to one end of
the cell even after the other chromosomes had merged and were no longer
distinguishable as separate units (fig. Be). No micronuclei were appar-
ent, however, so these must have been incorporated in the daughter
nuclei at the second division.

A greater amount of irregularity could be seen in 12- and 14-chromo-
some plants. In spite of this, a remarkably high percentage of complete
pairing was evident. In 12-chromosome forms only minor irregularities
were noted and these occurred infrequently. In nearly all cells which
presented clear views, six bivalents could be distinguished (cf. fig. 3c).
Occasionally a precocious division of a bivalent occurs, or lagging of the
univalents after the bivalents have split.

In 14-chromosome plants (fig. Bd) a peculiar condition was revealed.
Some flower heads showed a very high percentage of regularity in
meiosis, seven bivalents being regularly formed. In other florets on the

268

University of California Publications in Agricultural Sciences [Vol. 6

Fig. 3. Crepis syriaca. a, 1923 B6-2-5 (ten chromosomes forming five bivalents) ;

b, 1923 A2-10-1-1 (eleven chromosomes forming five bivalents plus one univalent) ;

c, 1923 A 11-2—1-7 (twelve chromosomes forming sis bivalents) ; d, 1923 A 11-4—1
(fourteen chromosomes forming seven bivalents) ; e, 1923 A 11-4—4—1 (eleven chro-
mosomes; telophase showing both halves of divided univalent at one pole).

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 269

same plant there was great irregularity. Chromatin masses appeared in
place of the usual independent units. These in many preparations were
accompanied by small fragments and division seemed to be brought about
by a simple fission of the general mass. This highly abnormal type of
division would account for the large percentage of poor pollen formed
by such plants.

In maize, Randolph found all types of pairing among the extra chro-
mosomes. An occurrence was also reported of pairing between one of the
supernumeraries and one of the normal bivalents. There was a difference
in the staining reaction of the extra univalents, the latter being some-
what darker than the rest of the complement.

EFFECT OF CHROMOSOME NUMBER ON POLLEN
FORMATION AND FERTILITY

Although there appears to be no clear-cut correlation between chromo-
some number and morphology in C. syriaca, an excess of chromatin does
have a marked effect on pollen formation and fertility. Furthermore, the
amount of excess chromatin has a greater effect apparently than the
chromosome unbalance. This can readily be seen by an examination of
table 3, which shows the percentage of undeveloped pollen grains in
different chromosome-number classes. Since some of the figures were ob-
tained, it has been found that occasionally the first flower formed yields
highly aberrant results. For this reason there have been omitted a few
plants which were obviously of this type and which would consequently
give misleading average values for the different classes.

The plant numbers marked with an asterisk refer either to individual
plants or to an average of two, which occurred in populations derived
from plants with chromosome numbers other than their own. The rest
were derived from plants with the same number, as is indicated in the
second column. Only a few 14-chromosome plants were available for
pollen counts and these showed from 30 per cent to 50 per cent unde-
veloped grains.

These figures are not conclusive, since a comprehensive study of pollen
development and fertility was not intended. The table does show, how-
ever, that there is a marked difference between 10-chromosome forms and
those which contain larger chromosome complements. It might well be
expected that 12-chromosome individuals, having a balanced chromo-
some condition, would show a higher percentage of good pollen than the
ll's and 13's. This does not prove to be the fact. There seems to be a
definite tendency for the higher-numbered individuals to have a larger
percentage of undeveloped pollen, irrespective of whether the chromo-
some number is odd or even.

270

University of California Publications in Agricultural Sciences [Vol. 6

TABLE 3

Cheomosome Numbers and Pollen Development and Fertility

Pedigree number

Chromosome
number

Percentage of
poor pollen

Fertility

A2-10

10
10

10
10
10
10
10
10
10
10
10
10

1.9
1.4
2.9
0.6
2.5
1.3
0.9
4.2
4.5
2.1
0.7
18.7

Fair

A2-15

Poor

B6-2

Fair

A2-15-5

Good

B6-2-5

Fair

A2-15-5-2

Fair

A2-10-1

Good

*All-2-6

Poor

*All-2-l

Fair

*All-2-6

Fair

*B7-8-2

*All-4-3

Poor
Poor

Average Percentage of Poor Pollen for 10-Chromosome Plants, 3.4

All-2

14.5
20.1
10.1
1.9
21.2

Poor

All-2-8

Poor

*All-4-4...

Fair

*All-4-5

*A2-10-1-1

Fair
?

Average Percentage of Poor Pollen for 11-Chromosome Plants, 13.5

B7-3

12
12
12
12
12
12
12
12

34.3
22.1
17.0
13.2
21.7
50.2
2.3
10.9

Poor

All-2-1

Fair

B7-8

Fair

*All-2

Poor

*B7-l-5

Poor

*All-2-8-7

Poor

*B7-l-3-l

?

All-4-3-3

?

Average Percentage of Poor Pollen for 12-Chromosome Plants, 21.4

B7-1

13
13
13
13

27.2

9.1

48.1

50.0

Poor

All-4

Fair

*B7-3-9 ...

Poor

♦All-2-1-1

Poor

Average Percentage of Poor Pollen for 13-Chromosome Plants, 33.6

* Either an individual plant or an average of two plants.

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 271

If the pollen grains containing supernumeraries are the nonfunctional
ones, it is hard to see why the higher-numbered forms do not yield chiefly
10-chromosome plants. The explanation of the results obtained may lie
in the possible occurrence of functional maternal gametes with more
than five chromosomes or of nonfunctioning pollen grains having five
chromosomes which lack something contained in the extra chromosome.

Little evidence of correlation between pollen development and fer-
tility has been shown in the results obtained up to the present time. To
be sure, the 10- and 11-chromosome plants show a higher degree of fer-
tility and a lower percentage of deficient pollen in general. Among indi-
vidual plants, however, those with the best pollen are not necessarily the
most fertile. The 11-chromosome plants are virtually as fertile as those
with ten, while 12's, 13's and 14's are nearly sterile and the rest com-
pletely so; and these do not show a proportional increase in nonfunc-
tional pollen grains.

RESULTS OF CROSSES WITH CREPIS ALPINA

Following the method outlined by Hollingshead (1930), several crosses
were made between individuals of C. syriaca from accession number
1923, and C. alpina plants obtained from accessions 2783, 1641, and 2769,
in order to determine, as far as possible, cytogenetical relationships be-
tween the species.

When C. syriaca is used as the female parent there is a much higher
percentage of seed set than in the reciprocal cross. This may be due to
the fact that usually syriaca plants were chosen which had a chromosome
number higher than ten; and it has been shown in wheat hybrids involv-
ing plants which differ in chromosome numbers, that it is preferable to
use the plant with the higher number as female (Watkins, 1927). Only
eleven of the achenes obtained from the cross alpina § x syriaca <§ pro-
duced plants and of these all but two died before reaching maturity.
A glance at plate 13, figure 3, shows the total dissimilarity of these to
the other hybrids and to the parents. The cause of this dissimilarity is as
yet unknown. It is possible that environmental conditions may provide
the true explanation, although occurrences of unlike reciprocal hybrids
have been reported. Further crosses are planned in order to clear up
this question.

From the reciprocal cross, twenty-six hybrids were obtained, being
derived from syriaca plants having ten, eleven, twelve, and thirteen
somatic chromosomes. In all of them hybrid vigor was marked. With
respect to leaf type and general habit they were more like the alpina
parent, but they had the flower-head type and color and the nodding
buds of syriaca. Nearly all of them were virtually as early maturing as
syriaca and were consistently highly fertile. The latter characteristic,

272

University of California Publications in Agricultural Sciences [Vol. 6

along with the compatibility of their chromosomes, is evidence of the
close relationship of the two species.

In crosses where alpina was used as a female parent, the resulting
hybrids all possessed ten chromosomes. This, of course, may only be
because of the small number obtained. However, in view of the amount
of variation ordinarily met with and the fact that only plants with
supernumeraries were used as pollen parents, it is much more likely that
there is a preferential functioning of pollen grains in favor of 5-chro-
mosome grains.

In the syriaca $ x alpina rf crosses, the hybrids revealed varying num-
bers of chromosomes, the situation being similar to that in the female
parent. The resulting chromosome numbers are shown in table 4. Since
it may be assumed that the alpina parent always contributes five chromo-
somes, the number present in functioning egg-cells is readily calculated.

TABLE 4

Frequencies of Hybrids Having Different Chromosome Numbers

Parental

chromosome

numbers

10

n

12

13

Chromosome
number in func-
tioning egg-cells

10x10
11x10

12 x 10

13 x 10

13
1
4

2
1

2

2

1

5
5,6

5,6,7, 8
6

As with C. syriaca, the F1 progeny of 10-chromosome parents all pos-
sessed ten chromosomes. Where a higher number was involved, variation
again occurred and to a considerable degree. The somatic complements of
the hybrids resemble closely the ones which have been described in con-
nection with different number-groups of syriaca. The basic five pairs are
consistently present with one, two, or three of the small supernumeraries,
the appearance of which has already been described. Assuming a regu-
lar reduction division, we should expect that the progeny of a 12- and a
10-chromosome plant would contain eleven chromosomes. Such, however,
does not prove to be true. Not only are both the extra chromosomes pres-
ent in the offspring, but even an additional one may be added to yield
a 13-chromosome form. As has already been indicated by the production
of an 18-chromosome syriaca individual from a 13-chromosome parent,
there must be a further multiplication of extra chromosomes in the divi-
sions prior to, or during, meiosis. If feasible, a study of embryo-sac
mother-cells should be of great assistance in clearing up this point. It
has already been pointed out that syriaca pollen probably does not trans-
mit the extra chromosome in crosses with C. alpina. Furthermore, the
entire lack of resemblance of these individuals to their maternal parents

1934]

Cameron: Chromosomes and Relationship of Crepis syriaca

273

and their similarity to one another precludes the assumption of a devel-
opment initiated by a doubling of the maternal genom. It is possible that
syriaca pollen may transmit the extra chromosomes in self-fertilization,
while showing quite a different behavior in alpina pistils.

MEIOSIS IN THE F1

An extensive investigation of meiotic behavior was not attempted since
a preliminary survey indicated that this was not recognizably different
from the parental type. This is only to be expected, considering the simi-
larity between the chromosome groups of the two species involved and
the high degree of fertility exhibited by the Fx plants. In all the 10-
chromosome Fx's examined, complete pairing was consistently the rule.

POLLEN STUDIES IN THE F,

Pollen counts of Px plants revealed some rather unexpected results. It is
significant, for example, that in 10-chromosome plants the percentage of
poor pollen was much greater than in C. syriaca itself while the per-
centage of fertility was as high, if not higher. This indicates that there
is not a very close correlation between percentage of good pollen and
degree of fertility. However, there is always considerable variation even
within an individual plant from time to time, so that these results are
not conclusive evidence on this point. The percentages of poor pollen
and the parental chromosome numbers of the F1 cultures are shown in
table 5.

TABLE 5
Parental Chromosome Number and Pollen Development in Fj

Culture
number

Parental chromosome
number

Average percentage of
poor pollen

X

10x10

58.7

X3

10 x 10

34.6

X8

10 x 10

59.4

X2

11x10

41.1

X4

12x10

64.7

X6

13 x 10

90.1

The fact that in hybrids the percentage of undeveloped pollen was
higher than in the parental species, irrespective of chromosome number,
was of assistance in determining whether a hybrid had actually been
obtained. Thus, a supposed hybrid which appeared very like C. syriaca
in external morphology, was found to have only 1 per cent poor pollen
and was for this reason classified as resulting from accidental self-
fertilization.

274

University of California Publications in Agricultural Sciences [Vol.6

The number of plants examined was of necessity very small, but, even
so, it seems fair to draw some tentative conclusions from the examina-
tions. The individual results are not tabulated because of the variation
just referred to, but they indicate that the parental chromosome num-
ber had a greater influence on pollen productioh than the number actu-
ally contained in the individual. For example, two 11-chromosome plants
in the progeny of a 13-chromosome syriaca parent had values of 89.1
per cent and 91.2 per cent poor pollen, while similar plants in the prog-
eny of an 11-chromosome progenitor had only about 40 per cent of
undeveloped grains.

As before, there is an increase in proportion of nonfunctioning grains
with an increase in number of extra chromosomes. In spite of the un-
balance in the 11-chromosome plants, they produce progeny which ex-
hibit considerably less poor pollen than the 12's or 13's. This is especially
surprising in view of the fact that six pairs are consistently found in
12-chromosome plants.

F2 CHROMOSOME NUMBERS

From the results discussed so far it might reasonably be expected that
in the F2 progeny a variability in chromosome number similar to that in
the parental and F1 types would be expressed. In respect to the 10-
chromosome lines this proved to be true, all individuals determined hav-

TABLE 6
Frequencies of F2 Plants with Different Chromosome Numbers

Culture

Parental
number

Frequencies of plants with:

number

10

11

12

13

14

15

X3-2

10

10

X4-1

10

9

X2-2

11

1

9

X6-1

11

1

8

X7-1

12

1

1

2

1

ing ten somatic chromosomes. However, in the progeny of 11-chromo-
some Fj individuals the majority of plants had twelve, while the self-
pollination of a 12-chromosome type gave rise to considerable variation
again, the tendency being toward still higher numbers. The obtaining of
such a large proportion of 12-chromosome plants from an 11-chromo-
some parent is very difficult to explain. Referring to the 11-chromosome
groups of tables 1 and 2, it can be seen that eighteen out of a total of
twenty-eight had eleven chromosomes, while only three had twelve. In
these F2 cultures seventeen out of a total of nineteen had twelve, while
only one had eleven (cf. table 6). This may indicate a tendency to pro-

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 275

duce a balanced number of chromosomes and suggests again the theory
that the supernumeraries contain some material which is beneficial in
development. When achenes from 12-chromosome F1 plants were sown,
however, the resulting plants, which are obtained only with difficulty,
showed greater variation in chromosome numbers than the progeny of
11-chromosome individuals, in spite of the unbalance in the cells of the
latter type. F3 results should prove of great interest in connection with
the stability of different chromosome groups. In table 6 are shown the
frequencies of F2 plants having various chromosome counts. These
plants were all quite uniform and, as in Fx, they resembled their alpina
grandparents in habit, leaf type, and fertility, and showed segregation
with respect to flower-head type. They did not mature quite so early as
the F1; but much earlier than pure alpina.

DISCUSSION

Supernumerary chromosomes have been reported in Zea (Randolph,
1928), Secale (Gotoh, 1924; Belling, 1925), Ranunculus (Langlet, 1927;
Sorokin, 1927), Oenothera (Lutz, 1916), Rosa (Blackburn and Harrison,
1921), Crepis (Navashin, 1926), Fritillaria (Darlington, 1930), Viola
(Clausen, 1931), and in a number of insects (Stevens, 1908; Carothers,
1917; Wilson, 1909). In most of these the extra chromosomes arise by
(1) duplication, (2) fragmentation, or (3) hybridization.

Duplication of chromosomes by means of nondisjunction or noncon-
junction was found by Navashin to be the cause of a small amount of
spontaneous chromosomal alteration in Crepis tectorum. The supernum-
eraries in this species had a greater effect on the viability of the plants
than those found in C. syriaca. The morphology of the extra unit in
the latter species precludes the possibility of its origin by duplication
although this may be a factor in the multiplication of it in higher-num-
bered plants.

Fragmentation of chromosomes of the normal complement is assumed
to be the cause of the extra units in Secale, Fritillaria, and Oenothera.
According to Geerts (1911) this phenomenon is likely to occur in extra
chromosomes as a result of hybridization. Measurements of chromosome
lengths in C. syriaca revealed that there was a definite linear increase of
the chromosomes as a group where supernumeraries were present. Thus
it is very unlikely that they arose by a simple fission of a member of the
basic genom. It is possible that fragmentation following an attachment
of two chromosomes may have led to the production of the X chromo-
somes of C. syriaca. Carothers (1917) explains in this way the origin of
extra units in Trimerotropis and Circotettix.

Hybridization is the factor to which is ascribed the origin of extra
chromosomes in Ranunculus and Viola. The fact that in C. syriaca varia-

276 University of California Publications in Agricultural Sciences [Vol. 6

bility is so widespread suggests that the species has arisen by hybridiza-
tion (involving C. alpina, in all probability). So far, no material has
been examined which did not show a fairly high proportion of individ-
uals with extra chromosomes, although in all the material the chromo-
some number was constant for the individual. Furthermore, there seems
to be no tendency toward reversion to the 10-chromosome type. Plants
of this type are regularly found to be in the majority and, when selfed,
yield only 10-chromosome progeny, yet accessions from the wild which
were studied showed many extra-chromosome individuals.

The only related species which possesses a chromosome similar to the
extra units found in C. syriaca is C. rubra. This species is indigenous in
Greece and is not known from Syria, the home of syriaca, nor even from
Asia Minor. If natural hybridization with C. rubra has brought about
the increase in chromosome number, it must be assumed that the cross
was effected a very long time ago and that the unbalanced condition has
been maintained. Furthermore, a careful examination of external mor-
phology fails to reveal the presence of any distinctive rubra characters
in syriaca. For these reasons it is necessary to look further for the origin
of the supernumeraries.

Although it appears unlikely that C. rubra was involved with C. al-
pina in the origin of C. syriaca, because of their wide geographic separa-
tion and the absence of any distinctive rubra characteristics in syriaca
plants, there are other possible explanations involving natural hybridi-
zation. It has already been pointed out that two distinct strains of C.
alpina are known which grow in fairly close proximity to each other,
the typical form being distributed from Asia Minor to Transcaucasia,
the other in the Caucasus region. Furthermore, the latter type possesses
some characters which are more like those of syriaca than those of the
typical form. It is quite possible that, after the former became distinct,
natural crossing was effected between them, which, with later genie and
chromosomal modifications, produced individuals of the syriaca type.
These may have become isolated and gradually moved in a southerly
direction to occupy the Lebanon region, while the typical form of C.
alpina was restricted to the present more northern range and the other
to the Caucasus and neighboring localities. This would explain the great
similarity of the basic chromosome groups of all three races, as well as
the origin of certain taxonomic characters of C. syriaca. For example,
the flowers of the latter are brighter yellow than those of typical C.
alpina, but are very similar in color to those of alpina plants from the
Caucasus region. In habit and achene characters, also, syriaca and the
Caucasian form of alpina show a tendency toward mutual resemblance.
Such a hybridization might lead to an alteration in the normal chromo-
some behavior, giving rise to the variation characteristic of C. syriaca.

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 277

Another explanation of the origin of these forms is that C. syriaca
and C. alpina (Caucasian type) are both the result of hybridization with
some form now unknown. This is suggested by the fact that the Cauca-
sian form is more or less intermediate between C. syriaca and typical C.
alpina. This interpretation, however, is less satisfactory since it assumes
the existence of an unknown form. Moreover, the present situation in re-
spect to both comparative morphology and the chromosomes in the two
forms of C. alpina and syriaca can be reasonably explained without re-
course to this alternative hypothesis.

Navashin (1931a) has shown how chromosomes may become altered
in nature. Plants of C. tectorum which were abnormal in appearance
showed, upon cytological examination, that chromosomes quite unlike
those among the normal four pairs were present. These had arisen by a
fragmentation of the normal members, followed by attachments of the
fragments so produced. He points out that Dobzhansky (1930) has found
such alterations occurring in homozygous condition. The individuals
containing them were viable and even capable of reproducing them-
selves.

A preliminary study of plants grown from accession number 3106 in-
dicates that a similar situation exists in C. syriaca. In these individuals
many types of chromosomal abnormality are to be found, including such
units, appearing as supernumeraries, as are found in the cultures de-
rived from accession number 1923. Let us assume that hybridization of
the two forms of C. alpina has produced syriaca-like progeny. Either as
a result of this or through some environmental agency (temperature,
chemical condition of the soil, natural radiation, etc.), an alteration in
chromosome affinity is brought about. Two nonhomologous chromosomes
become attached in the prophase stages of cell division and, upon sepa-
ration, do not retain their original form. For example, a part of one of
the other chromosomes might remain attached to the satellite of the
normal D chromosome. A similar effect would be produced if fragmenta-
tion preceded the attachment. Where fragments are produced which do
not possess a spindle-fiber attachment, they are subsequently lost unless
they become attached to a unit which possesses one. Such an attachment
of a fragment would produce a chromosome with an abnormally large
satellite, as in the X chromosome described earlier. If, in reality, the
supernumerary has arisen by means of a modification of the normal
D chromosome, this alteration would also necessarily involve a shorten-
ing of the distal arm by fragmentation or some other means. It is alto-
gether likely that an alteration of this kind would upset the normal
chromosome balance to such an extent that sterility would result. In
syriaca, however, the X chromosome seems to have only a weak effect
when present one, two, or even three times. Thus, this particular chromo-
some may have been introduced into otherwise normal complements and

278 University of California Publications in Agricultural Sciences [Vol. 6

been able to reproduce itself even in competition witb other more nor-
mal types.

The fact that the extra chromosomes are reproduced in such a con-
stant manner lends strength to the theory of chromosome individuality.
Other similar examples have been used as reason for questioning its
validity.

Little work has been done, to the writer's knowledge, with respect to
studies of progeny from extra-chromosome plants of this type. The pres-
ent study shows the tendency of certain of the higher-numbered indi-
viduals to maintain these complements. It is probable that, by continuing
the work, stable races could be established.

The investigation of cytogenetic relationships of C. alpina and C.
syriaca is of interest in the light of modern theories in taxonomical re-
search. From a purely morphological consideration there is little doubt
that they should be classified as distinct species. Constant differences are
apparent wherever material is found, and although they do exist in the
same general geographical region, their distribution is not known to
overlap. Certain intermediate characters indicate that syriaca has arisen
from alpina originally, but changes have occurred which make its
segregation warranted.

The criteria which, cytogenetically speaking, have so far been of as-
sistance in taxonomic treatments, are : (1) number of chromosomes, (2)
chromosome size and shape, (3) compatibility of chromosomes when
brought together in a hybrid, and (4) interfertility or sterility upon
hybridization. As has been pointed out, the chromosomes of the two spe-
cies under discussion are alike in both number and morphology with the
exception of the supernumeraries present in C. syriaca. When crossed, a
hybrid is obtained which is more or less intermediate in appearance.
It shows a high degree of chromosome affinity and is as fertile as the
parental species, or nearly so. Thus, with reference to chromosome affi-
nity in the basic genom and the fertility of Fx hybrids, the former classi-
fication, that is, treating syriaca as a subspecies of C. alpina, is probably
justified. There remain, however, the supernumerary chromosomes which
are unique in syriaca and the numerous differences in external mor-
phology which serve to set it apart as a species.

SUMMARY

1. Crepis syriaca was originally described as a variety of C. alpina L.,
and unquestionably it is very closely related, but numerous constant
morphological differences exist.

2. Accessions from the wild show considerable chromosomal variation
involving number and morphology. Chromosome numbers are constant
for the individual, but in selfed lines plants with ten, eleven, twelve,

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 279

thirteen, fourteen, fifteen, sixteeen, and eighteen somatic units are found,
the variation involving a small chromosome of distinctive morphology.
Selfed cultures show a much greater stability of 10-chromosome lines
and greater variability of lines derived from plants containing super-
numeraries. This variability does not lead to the production of a ma-
jority of 10-chromosome individuals.

3. The supernumeraries have small and inconsistent effects on the ex-
ternal morphology of plants containing them. Abnormal flower-head de-
velopment is regularly associated with the presence of two or more extra
units. Ten- and 11-chromosome plants were very similar in appearance,
but higher-numbered individuals varied considerably from the normal.
An interesting new "dwarf" form, not associated with chromosome varia-
tion, made its appearance in 10-chromosome cultures.

4. The presence of extra chromosomes causes a marked increase in the
percentage of undeveloped pollen grains produced, this increase being
more or less proportional to the number of extra units. The number of
supernumeraries present apparently has a greater effect in this connec-
tion than has tbe chromosome unbalance. There seems to be a preferen-
tial functioning in favor of gametes with more than five univalents in
plants with more than ten chromosomes. Although individuals with five
or more supernumeraries are entirely sterile and those with fewer are
partly so, there does not seem to be a direct correlation between pollen
development and fertility in these cultures.

5. At meiosis the basic genom forms five bivalents. In plants contain-
ing two or four supernumeraries complete pairing is consistently found.
In 11-chromosome plants, the univalent divides at the first division. At
the second, the halves go to one or the other pole and are incorporated in
the daughter nuclei. Occasional nondisjunction of supernumerary units
at these divisions may account for the increase in chromosome number.

6. In Fa hybrids from C. syriacax C. alpina the chromosome situation
is closely similar to that in the pure species. So far, the extra chromo-
somes have been shown to be transmitted only by maternal gametes. The
hybrids show complete pairing of chromosomes in 10-chromosome crosses
and are highly fertile.

7. Although high fertility is shown by Fx individuals, the percentage
of undeveloped pollen increases considerably, even in 10-chromosome
plants. The parental genom apparently has a greater effect on the pollen
development than has that of the progeny plant. The proportion of un-
developed pollen increases with the number of extra chromosomes in-
volved in the cross.

8. C. rubra, a related species, has a chromosome resembling the super-
numerary of C. syriaca; however, the possibility that it has been involved
in the origin of the latter species by hybridization is rejected on grounds
of evidence from comparative morphology and geographic distribution.

280 University of California Publications in Agricultural Sciences [Vol. 6

9. A more acceptable hypothesis assumes the origin of C. syriaca
through hybridization between typical C. alpina and a Caucasian form
of alpina followed by chromosomal alterations in some of the hybrids,
giving rise to the supernumerary chromosomes which, in this species,
are unique.

BIBLIOGRAPHY

Babcock, E. B., and Lesley, M. M.

1926. Chromosome number and individuality in the genus Crepis. II. The chromo-
somes and taxonomic relationships. Univ. Calif. Publ. Agr. Sci., 2:315-341.

Babcock, E. B., and Navashin, M.

1930. The genus Crepis. Bibliog. Genet., 6:1-90.

Belling, J.

1925. Fracture of chromosomes in rye. Jour. Hered., 16:360.

Blackburn, K. B., and Harrison, J. W. H.

1921. The status of the British rose forms as determined by their cytological be-
havior. Ann. Bot., 35:159-188.

Carothers, E. E.

1917. The segregation and recombination of homologous chromosomes as found
in two genera of Acrididae (Orthoptera). Jour. Morph., 28:445-521.

Clausen, J.

1931. Viola canina L., a cytologically irregular species. Hereditas, 15:67-88.

Darlington, C. D.

1930. Chromosome studies in Fritillaria. III. Chiasma formation and chromosome
pairing in Fritillaria imperialis. Cytologia, 2:37-55.

DOBZHANSKY, T.

1930. Translocations involving the third and fourth chromosomes of Drosophila
melanogaster. Genetics, 15:347-399.

Geerts, J. M.

1911. Cytologische Untersuchungen einiger Bastarde von Oenothera gigas. Ber.
Deut. Bot. Gesell., 29:160-166.
Gotoh, K.

1924. tiber die Chromosomenzahl von Secale cereale L. Bot. Mag. Tokyo, 38: 135—
152.

Hollingshead, L.

1930. A lethal factor in Crepis effective only in an interspecific hybrid. Genetics,
15:114-140.

Hollingshead, L., and Babcock, E. B.

1930. Chromosomes and phylogeny in Crepis. Univ. Calif. Publ. Agr. Sci., 6:1-53.

Langlet, O. F. I.

1927. Beitrage zur Zytologie der Ranunculazeen. Svensk. Bot. Tidskr., 21:1-17.

Lutz, A. M.

1916. Oenothera mutants with diminutive chromosomes. Am. Jour. Bot., 3:502-
526.

1934] Cameron: Chromosomes and Relationship of Crepis syriaca 281

Navashin, M.

1926. Variabilitat des Zellkerns bei Crepis-Arten in Bezug auf die Artbildung.
Zeitschr. f. Zellforsch. u. mikrosk. Anat., 4:171-215.

1931a. Spontaneous chromosome alterations in Crepis tectorum L. Univ. Calif.

Publ. Agr. Sci., 6:201-206.
1931&. A preliminary report on some chromosome alterations by X-rays in Crepis.

Am. Nat., 65:243-252.

Eandolph, L. F.

1928. Chromosome numbers in Zea mays L. Cornell Univ. Agr. Exp. Sta. Mem.,
117:3-44.

SOROKIN, H.

1927. The chromosomes of Ranunculus acris. Am. Nat., 61:571-574.

Stevens, N. M.

1908. The chromosomes in Diabrotioa vittata, Diabrotica soror, and Biabrotica
12-punctata. Jour. Exp. Zool., 5:453-470.

Watkins, A. E.

1927. Genetic and cytological studies in wheat. III. Jour. Gen., 18:375-396.

Wilson, E. B.

1909. Studies on chromosomes. V. The chromosomes of Metapodius. A contribu-
tion to the hypothesis of the genetic continuity of chromosomes. Jour. Exp.
Zool., 6:147-205.

Plate 12

Crepis syriaca

Fig. 1. Dwarf strain (ten chromosomes).

Fig. 2. Normal plant (ten chromosomes).

Fig. 3. 11-chromosome plant.

Fig. 4. 12-chromosome plant.

[282]

UNIV. CALIF. PUBL. AGR. SCI. VOL. 6

ICAMERONl PLATE 12

Fig. 3

Fig. 4

[283]

Plate 13

Cii pis syriaca

Fig. 1. 14-ehromosome plant.

(r> pis alpina and F, hybrids with C. syriaca

Fig. 2. C. alpina (ten chromosomes).

Fig. 3. 10-chromosome hybrid from cross ('. alpina Jxf. syriaca J.

Fig. 4. 10-chromosome hybrid from cross C. syriaca $ x C. alpina g.

I 284 |

UNIV. CALIF. PUBL AGR. SCI. VOL 6

[CAMERON) PLATE 13

Fig. 1

**&* ■

r

Fig. J

Fig. 4

[285]