'y

CHROMOSOMES AND PHYLOGENY IN CREPIS

BY
LILLIAN HOLLINGSHEAD AND ERNEST B. BABCOCK

inn

University of California Publications in Agricultural Sciences

Volume 6, No. 1, pp. 1-53, 24 figures in text

Issued January 4, 1930

University of California Press
Berkeley, California

Cambridge University Press
London, England

CHROMOSOMES AND PHYLOGENY IN CREPIS

BY

LILLIAN HOLLINGSHEAD and EENEST B. BABCOCK

INTRODUCTION

In connection with genetic and taxonomic studies of Crepis, an
examination of as many species as could be brought into cultivation
has been in progress for about ten years. The earlier work on the
chromosomes was done by Dr. Margaret Mann Lesley, who studied
particularly numbers and sizes (Mann, 1922, 1925; Babcock and
Lesley, 1926). The work of M. Navashin (1925, 1926) and Taylor
(1925, 1926), who described satellites and constrictions for the first
time in this genus, showed that a closer morphological study of the
chromosomes from suitably fixed material would be of value for com-
parative studies of related species.

It is the purpose of this paper to present our knowledge of number
and morphology of the chromosomes in seventy species and to con-
sider this evidence in relation to a system of classification based on
phylogenetic relationship. But the present paper is not intended to
serve as a taxonomic treatise. Therefore no keys or descriptions of
species will appear and there will be no attempt to set forth the
detailed evidence for the phylogenetic groupings proposed, as such
descriptions and data will appear in a taxonomic treatment now in
preparation. The specific names used have been carefully verified as
to identity, priority, and authorship, and are in nearly every case the
same as those which will be used in later publications. In the present
paper it is proposed merely to present the evidence derived from
cytological investigation and to discuss the phylogenetic groupings of
those species of Crepis which have been investigated. These group-
ings, however, have been worked out by combining the data on chro-
mosome number and morphology with the evidence from external
morphology of the plants, at the same time giving consideration to
geographic distribution and to the genetic evidence derived from
experiments on interspecific hybridization.

The point of view held by the writers with regard to the funda-
mental relations between phylogeny and taxonomy is in general

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

agreement with that of Hall and Clements (1923). The concept of
species denned by these authors and accepted by us as a satisfactory
approximation to the truth is stated as follows:

The evolutionary view of the species is that it is a definite phylogenetic stock,
sprung from and related to similar stocks, and itself undergoing modification into
a number of variads. As they have recently come from the same stock, these
variads are more nearly related to each other than they are to those of any other
species, and they represent a definite phylogenetic unit, the species, at the same
time that they mark its further differentiation.

But the view of these authors, that gross morphological difference
between plants, as contrasted with such cytological features as number
and morphology of the chromosomes, is the only definite measure of
progress in evolution, is too limited, as has already been pointed out
by one of us (Babcock, 1924). The importance of the chromosomes,
especially their appearance in somatic cells at mitotic metaphase. as
an index of taxonomic relationship, has become increasingly evident
as the number of species of Crepis examined has increased.

Acknowledgments

It is impossible at this time to mention the many institutions and
individuals who have assisted by providing seeds or roots of the
species herein discussed. These will appear in connection with sources
to be acknowledged later. We are especially indebted, however, to
Dr. M. Navashin for seeds of several species, which he brought from
Russia, and for active interest and frequent help throughout the
investigation. We also gratefully acknowledge the constant cooper-
ation of Mr. C. W. Haney in connection with growing the plants,
assisting in collection of root tips, and many other details.

CYTOLOGICAL MATERIALS AND METHODS

Somatic metaphases of root tips have been used exclusively for
chromosome counts and morphological studies. In most cases the
material was obtained from plants in the rosette stage after they had
been transplanted to five- or six-inch pots. Occasionally the roots
of very young seedlings were used but they were on the whole less
satisfactory. Some of the counts on the American species were pro-
cured from tap roots dug from the wild and transplanted to pots in
the greenhouse, where they produced new roots, giving material well
suited for chromosome studies. In other American species the counts

1930] HoUingshead-Babcock : Chromosomes and Phytogeny in Crepis 3

were obtained from young seedlings. Introduction of most American
species into garden culture has been found very difficult but is grad-
ually being accomplished.

Roots were fixed in Taylor's (1925a) chrom-acetic-osmic-maltose
solution, in weak and strong Flemming, and in chrom-acetic-formalin.
Most of the investigations and all of the drawings have been made
from material fixed in one of the two chrom-acetic-formalin solutions
given below.

1. A. 65 cc. water 2. A. 65 cc. water

10 cc. glacial acetic acid 10 cc. glacial acetic acid

1 gr. chromic acid 1 gr. chromic acid

B. 40 cc. formalin (commercial) B. 10 cc. formalin (commercial)

35 cc. water 65 cc. water

Mix one part A with one part B just Mix one part A with one part B just
before fixing. before fixing.

No constant difference in quality of fixation between material
fixed in these two solutions has been established. The chromosomes
fixed with these fixatives are usually slightly shorter and broader than
those fixed in the osmic fixatives, so they may be somewhat contracted.
Morphological details are quite clear in well fixed material, however,
and in material so fixed the homogeneity of the cytoplasm, against
which the chromosomes stand out very clearly, facilitates greatly the
examination of the sections for suitable figures.

The material was imbedded in paraffin and sections were cut 6 to
12//. thick depending on the size and number of the chromosomes.
Heidenhain's iron haematoxylin was used for staining.

A Bausch and Lomb (90 x) apochromatic oil immersion-objective
and Zeiss compensating oculars have been used in this study. All the
drawings have been made with the help of a camera lucida at a
magnification of approximately 3750 and reduced to 2500 in repro-
duction.

Where the chromosomes of a species were being investigated for
the first time, counts from two or more plants were made whenever
possible. The number of plants investigated is indicated in the table
of chromosome numbers. In those instances where the chromosome
number of a species newly investigated is based on one plant this plant
is the only one which has been grown in our cultures. In the case of
Crepis incana, the plant whose chromosomes were studied did not
bloom ; and as this was the only plant available and there were no
herbarium specimens accompanying the seeds, the identification was
based on achene and leaf characters alone.

4 University of California Publications in Agricultural Scien-ces [Vol. 6

Where the species had 12 or fewer chromosomes an attempt was
made to find a metaphase plate, a drawing of which would depict the
complex so that each individual chromosome might be recognized if
it could be distinguished from the others under the microscope. Lack
of sufficient material or particular difficulties have in some cases
forced us to be satisfied with figures of plates which do not reach this
ideal. In most cases, however, the chromosomes can be distinguished
in the figures and the pairs recognized without difficulty.

Where the number exceeds 12, the problem of distinguishing the
various chromosomes becomes more difficult and in most cases it has
been impossible to distinguish all the chromosome pairs in one plate,
although various types are readily picked out. In those species with
more than twenty chromosomes the plates for drawing were chosen
with the primary aim of depicting clearly the number, and, second-
arily, of showing as many as possible of the chromosomes which could
be distinguished by their particular morphology.

GENERAL MORPHOLOGICAL FEATURES

Marked differences in the frequency of occurrence of good meta-
phase plates were noted between species. The particular species in
question, as well as the condition of the roots, affected the number
of mitotic figures. Some species characteristically showed plates in
which the chromosomes were well spaced in a single plane while in
others the usual arrangement of chromosomes was such that a morpho-
logical study was impossible. The species with higher numbers and
larger chromosomes presented greater difficulties in this respect.

Most species, in material well fixed in chrom-acetic-formalin and
well stained, show black chromosomes, their outlines clear and distinct
against the gray cytoplasm. When differentiated so as to clear the
cytoplasm, some species show chromosome outlines which are less
sharp, especially under high magnification. The extreme phase of this
condition was found in certain species (C. bulbosa and C. japonica)
which have been included under Crepis but which at one time or
another have been assigned to other genera. The same is true of
Ixeris graminea, which for a time was thought to belong in Crepis.
Their chromosomes are small, and it may be that this is the reason
why the outlines under differentiation become less clear so as to make
it difficult to determine the shapes with exactness. The black and
white figures used in this paper are not suitable for showing such
differences.

1930] HoUingshead-Babcoclc : Chromosomes and Phytogeny in Crepis 5

Size difference continues to be one of the most useful features by
which chromosomes may be distinguished. The relative lengths of
the components of a particular chromosome complex are rather con-
stant and although the lengths of the same chromosome may differ
noticeably from cell to cell this variation is relatively small. Differ-
ences in the thickness of chromosomes of different species are obvious
but no consistent difference between the widths of the various chromo-
somes of any one complex has been observed.

Satellites or "trabanten," chromatin balls attached to the chromo-
somes proper by thin threads, first described by S. Navashin (1912),
are now generally recognized as of wide-spread occurrence. The list
of plants and animals in which satellites have been observed, given
by Kuhn (1928), will serve to illustrate in what diverse groups these
peculiar structures occur. Most, if not all, species of Crepis contain
at least one pair of chromosomes marked by the presence of satellites.
They vary in size from a large ball (C. sctosa, fig. 4c) to a tiny sphere
(C. palaestina, fig. 18d). As in most other genera they are usually
attached to the proximal ends (those which lie toward the center of
the plate) of chromosomes which have short arms or heads (cf. below),
but in some cases (C biennis, fig. 12a) they may be attached to chro-
mosomes with arms more nearly equal in length. In only one species
(C. pulchra fig. 19) was there found a pair of chromosomes with
distal satellites. In several species (C lyrata, C. mollis, C. hieros-
olymitana, fig. 13) structures resembling large satellites at times, but
at others appearing to be segments cut off by distal constrictions,
have been seen. Such variations may be due to imperfect fixation as
suggested by Taylor (1924). In this connection it should be noted
that Geitler (1929) figured the chromosome complexes of C. capillaris,
C. dioscoridis, C. rubra, and C. blattarioides, depicting morphological
features similar to those shown in our figures. Considerable variation
in the appearance of the satellited chromosome of C. blattarioides was
described, including the form and size of the satellite and even its
presence or absence. Some if not all of this variation can perhaps be
attributed to poor fixation.

The behavior of the satellites during the mitotic anaphase of C.
capillaris has been described and figured by Taylor (1926) and the
writers' observations add nothing new to this knowledge. Whether or
not the satellites are to be found on the nucleolus at prophase as
stated by M. Navashin (1925) has not been particularly studied but
prophases have been seen which could be interpreted as exhibiting
this phenomenon.

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

The clearness with which the satellites are visible depends on their
size, the quality of the fixation, and the degree of differentiation. The
thread which attaches the satellite to the chromosome proper may be
contracted under poor fixation until a small satellite appears as a
little protuberance (C. incamata, fig. 18c) or a large satellite as a
segment. Excessive or even ordinary differentiation may render
invisible a satellite thread or a small satellite. They are not always
to be seen even in well fixed material but it is believed that in such
cases they are hidden by the chromosomes proper and that they are
constant morphological features of the living chromosome. The
satellite may vary slightly in apparent size and shape from cell to cell
within a species but on the whole it is as constant as any morpho-
logical feature. On occasion a satellite thread may be drawn out to
an unusual length (C. amplexifoJia, fig. 5c) or the spherical shape
may be obliterated (C. parviflora, fig. 11a). These, however, are
exceptional cases and may be due to poor fixation.

One certain case (C. foetida 2048, fig. Id) and several possible
cases of a constant difference in size between satellites on homologous
chromosomes within a species have been found. M. Navashin (1926)
reported such a case in C. dioscoridis and similar cases in other genera
are known (cf. Kuhn, 1928). Detailed studies of more material
would be necessary to decide the questionable cases.

Aside from size and the occurrence of satellites, a number of
investigators have found that the most useful distinguishing feature
of chromosomes is the occurrence of a fiber attachment constriction
dividing the chromosome into two arms. The importance of the ' ' two-
armed" nature of the chromosomes has been reemphasized by Heitz
(1928), who has found it in many widely separated families and
indeed in different divisions of the plant kingdom, and he holds it to
be a universal characteristic of chromosomes. S. Navashin (1916&)
has stated that absolutely every chromosome possesses an "achromatic
transverse fissure" which divides the chromosome into two arms
(cited from M. Navashin, 1926). Taylor (1925, 1926), who has dis-
cussed fully the kinds of constrictions, has failed to find a case among
any of the medium or large chromosome types studied which had a
truly terminal fiber attachment and he is inclined to the opinion that
an actual constriction zone or related structural differentiation is
always present and can be demonstrated by suitable means. All
investigators agree that one arm may be so short as to constitute a
small head.

1930] Hollingshead-Babcock : Chromosomes and Phylogeny in Crepis 7

In Crepis, fiber attachment constrictions have been found on prac-
tically every chromosome which could be identified (except in C.
japonica and C. bulbosa). Where the chromosome is bent so as to
obscure the constriction, the relative lengths of the arms give a fair
indication of its position, for the chromosome usually bends at the
constricted region. The constrictions are particularly apt to be
observed on chromosomes where they are to be found very near the
ends, as on the satellited chromosomes of C. alpina var. syriaca
Bornm.1 (fig. 2c).

THE CHROMOSOMES OF SEVENTY SPECIES

Mann (1925) summarized the chromosome numbers of the twenty-
seven species of Crepis reported by various investigators up to that
time and Tischler (1927) has a complete bibliography of papers deal-
ing with Crepis chromosomes which includes additional counts given
by M. Navashin (1925), Babcock and Lesley (1926), and some un-
published chromosome counts by M. Navashin. The chromosome
number of C. reuteriana has been given by Babcock and Hollingshead
(1929). In no case have counts obtained by the writers differed from
those reported by Mann, Navashin, or Babcock and Lesley.

With a few minor exceptions, no details of somatic chromosome
morphology other than size differences were described in earlier
works on the genus. De Smet (1913-14) figured satellites on C. virens
(= C. capillaris) but did not describe them. Rosenberg (1918) recog-
nized flexures in C. tectorum chromosomes as constant in position at
somatic anaphase and later (1920) says, speaking of the chromosomes
of C. capillaris (wrongly called C. reuteriana, cf. Babcock, 1924?>),
"ein sog. Trabantenbildung kommt vor die aber meiner ausicht naeh
von ganz anderer Natur ist als die ' Quersegmentierung. ' " It is
likely that the so-called "trabanten" which he saw were merely heads
separated from the chromosomes proper by constrictions. De Litar-
diere (1923) located the point of fiber attachment on one chromosome
of C. virens (= C. capillaris) and noted that the chromosome was
usually bent at that point. Mann (1925) saw the large satellite of
C. setosa and stated later (Babcock and Lesley, 1926) that satellites
were "not always present" in her material. Such results can, in
most cases, be attributed to the use of unsuitable fixatives.

1 This "variety" should be recognized as a distinct species, but the change in
nomenclature will not be made at this time.

s

University of California Publicatioiis in Agricultural Sciences [Vol. 6

M. Navashin (1925, 1926) and Taylor (19256, 1926) established the
constant occurrence of satellites and fiber-attachment constrictions as
features of use in distinguishing the various chromosomes in a number
of species of the genus. Navashin (1925) figured ten species in which
he had determined the morphological features of each chromosome.
These species were C. pulcherrima (= C. pulchra), C. grandiflora
(= C. conyzaefolia), C. dioscoridis, C. tectorum, C. rubra, C. mar-
schaUi, C. virens (t= C. capillaris), C. rhoeadifolia (=C. foetid a),
C. alpina, and C. parviflora. To these he has added C. aspera in
another paper (1927). Taylor (19256) gave figures showing the
morphological features of the various chromosomes of C. setosa and
C. capillaris, and later (1926) he figured C. capillaris in somatic
metaphases and anaphases. The investigations reported here have
corroborated those of Navashin and Taylor in practically every detail,
and drawings of the species they figured are included only for the
sake of uniformity and completeness.

The present investigations add twenty-seven species to the list of
those whose chromosome numbers are known (including four species
of other genera closely related to Crepis). Table 1 gives the somatic

TABLE 1

The Somatic Chromosome Numbers of Sixty-Seven Species of Crepis

(Three other species are listed, with numbers in italics, which are not accepted
in Crepis by the authors.)

OLD WORLD SPECIES

Species

C. aculeata (DC.) Boiss

C. alpina L

*('. alpina var. syriaca Bornm.

C. amplexifolia (Godr.) Willk

C. aspera L.
*C. asturica Lacaita

C. aurea (L.) Cass

C. biennis L.

C. blattarioides (L.) Vill

C. bulbosa (L.) Tausch

*C. bungei Ledeb

*C. burejensis F. Schmidt
C. bureniana Boiss

Somatic

chromosome

number

8

10

10, 11, 12, 13

8

8

10

10

39, 41

8
18

8
16

Number
of plants
examined

9

1

20

5

2

3

2

5
1
3

16t
3

Accession

1602

1499

1923

1019

1135, 1073

2088

2170

1874, and others

2033

1303

1827

2174

2747

1655

•Chromosome number reported for the first time.
t One plant was triploid with 12 chromosomes.

1930] Hollingshead-Babcoclc : Chromosomes and Phylogeny in Crepis

TABLE 1 — (Continued)

Species

C. bursifolia L

C. capillaris (L.) Wallr

C. chondrilloides Jacq

*C. chrysantha Froel.

C. ciliala C. Koch

C. conyzaefolia (Gouan) D. T
C. dioscoridis L

C.foetida L

*C. gymnopus Koidz

*C. hackeli Lange

*C. hierosolyrnilana Boiss

C. hookeriana Ball

*('. incana Sibth. et Sm

C. incarnala Tausch

C.japonica (L.) Benth

C. lacera Tenore

*C. leontodontoides All

*C. lybica Pamp

C. lyrata Froel

C. marschalli C. A. Mey.

C. mollis (Jacq.) Asch

C. monlana Urv

C. multicaulis Ledeb

C. myriocephala Coss. et DR.
*C. nana Richards

C. neglecta L

C. nicaeensis Balb

C. palaeslina (Boiss.) Bornm.

C. paludosa (L.) Moench

C. pannonica (Jacq.) C. Koch

C. parviflora Desf

*C. pontana (L.) D. T

C. pi-aemorsa (L.) Tausch

C. pulchra (L.)

C. reuteriana Boiss

C. rubra L

C. senecioides Delile

C. setosa Hall, f

C. sibirica L

C. taraxacifolia Thuill

Somatic

chromosome

number

Number
of plants
examined

Accession

8

3**

1220

6

20

X, 982 and others

8

2

2180, 1907

8

3

2179

40, 42?

2

2181

8

2

2183

8

3

1742, 972, 1455

|

3

1751, 2188

10

1

2048

1

3

2307

8

2

2746

16

3

1873

12

3

2619

8

I

1458

16

1

1667

8

1

1304

16

2

1045, 2132

8

4

1914

10

1

1807

8

1

1698

12

4

1644

8

3

1532

12

2

2201

12

2

1175

10

2

1480

8

1

1557

14

2

2698

8

1

1753

8

6

2700

8

1

1552

12

1

1825

8

1

1695

8

2

1630

10

2

2204

8

3

2133

8

J

4

1213, 1483

1

1

1894

8

4

2134, 2218

10

]

2

1506

I

1

1176

8

5

1044

8

1

1036

I

1

1510

10

2

1862

8

1

5

1806, 10G4, 1704

1

1895

* Chromosome number reported for the first time.
** One plant was a trisomic with 9 chromosomes.

10

University of California Publications in Agricultural Sciences [Vol. 6

TABLE 1— (Continued)

Species

Somatic

chromosome

number

Number
of plants
examined

Accession

C. tectorum L

*C. tenuifolia Willd.

• !

15
10
8
10
16
10

9

1
8
1
4

2
2

1

1498

1702
1826

C. tinaitana Salz

C. vesicaria L

*Rodigia commutata Spr.

1681
1576
1666, 2219

*Ireris oraminea Nakai

*Pterotheca sancta (L.) K. Koch

2568

2582 or 2583

AMERICAN SPECIES

44?

1

1778 (typical)

33

6

1922 (typical)

33

1

2096 (typical)

*<'. acuminata Nutt. <.

55?

1 or morej

1830 (form)

33

1 or more

1848 (form)

33

4

1919 (form ?)

33

4

1934 (hybrid form)

*('. andersoni Gray (

22
22

3

2

2086 (typical)

2136 (form)

*C. elegans Hook

14

5

2654

(

88?

1 or more

1840 (typical)

*C. barbigcra Leib J

88?

1 or more

1842 (typical)

44

1 or more

1838 (hybrid form)

{

88?

6

1959 (hybrid form)

*C. glauca (Nutt.) T. and G

99

1 or more

2079 (typical)

*C. gracilis (Eat.) Rydb

22

55?

2
6

2572 (typical)
1695 (form)

*C. monticola Coville

55?

1 or more

2771 (typical)

C. nana (see Old World Species)

22

2

2772 (typical)

*C. occidentalis Nutt

22?

1

2220 (form)

44

5

1921 (form)

22

1829 (form)

22

2075 (form)

22

2078 (form)

22

2066 (form)

22

2065 (form)

*('. nincinala (James) T. and G. .A

22

2068 (form)

22

2069 (form)

22

1 or more

2076 (form)

22

2077 (form)

22

2083 (form)

22

2071 (form)

*C. scopulorum Cov

44?

2

2773 (form)

* Chromosome number reported for the first time.

t Indicates that the roots were fixed under the accession number only and no record was kept of the

number of plants fixed together.

1930] Hollingshead-Babcock : Chromosomes and Phylogeny in Crepis 11

chromosome numbers of the species, the chromosomes of which are
described below, the numbers of individual plants, and the accession
numbers of the various plants examined. An accession number is
given to each lot of seeds or roots when it is received and the progeny
of any accession is likewise designated by that number. An asterisk
denotes that the chromosome number is reported for the first time.
Three of the species designated by asterisks (C. lybica, C. runcinata,
and C. glauca) were first counted by Dr. Margaret Mann Lesley and
one (C. leontodontoides) was first examined by Miss Priscilla Avery
(unpublished observations) .

With the discovery of the somatic number 14 in C. nana, the
known series of somatic chromosome numbers in the Old World
species of Crepis becomes 6, 8, 10, 12, 14, 15, 16, and 40±. That
of the American species is 14, 22, 33, 44, 55 ?, and 88 ?.

In the following descriptions the species are taken in groups
corresponding to the phylogenetic groups to be considered later.
There are four major groups or subgenera— Paleya, Barkhausia,
Catonia, and Eucrepis. The arrangement of species within each of
these divisions is shown in the chart on page 30.

Paleya

Crepis asturica (2?i = 10; fig. la), the one representative of the
subgenus Paleya which has been examined cytologically, has a com-
plex which illustrates several of the chromosome types common in
the genus. The two longer pairs, nearly equal in length, can be dis-
tinguished from one another by a small difference in the relative
lengths of the arms which are separated by the fiber attachment con-
strictions. The satellited pair, shorter, has the fiber attachment con-
strictions closely subterminal, forming small heads to which the small
satellites are attached. The two remaining pairs, nearly equal in
length to the satellited pair, have fiber attachment constrictions
median or nearly median and they commonly appear as small Vs.

Barkhausia

The chromosomes of C. alpina, C. alpina var. syriaca, C. rubra,
C. foetida, and Rodigia commutata, all with a somatic number of
ten, resemble those of C. asturica but with the exception of C. rubra
they are generally smaller. Those of C. foetida (fig. lb, c, d) re-
semble those of C. asturica most closely and are almost as large.

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

C. foetida 2188 (=C. rhoeadifolia, Navashin, 1925) does not differ
noticeably from typical C. foetida in chromosome morphology, nor
does C. foetida 2307 (=C. gland idosa Guss.), but C. foetida 2048
(fc=C. interrupta S. et S.) is distinguished by larger satellites and
the plant examined showed a marked difference in the size of the two
satellites (fig. Id). That a careful study involving investigation and
measurement of many chromosomes might establish further differ-

Ct asturica c. foetida 1751 C. foetida 2188

b * c

&<£ ik

C. foetida 2048 Rodlgla commutata

d e

Figure 1.

ences between these foetida strains cannot be denied. Rodigia com-
mutata (fig. If) has a chromosome complex very similar to that of
C. foetida.

In C. alpina (fig. 2a) one of the two short pairs of chromosomes
has small heads. Plants of C. alpina var. syriaca (fig. 2b, c, d, e)
which were examined varied in chromosome number from 10 to 13.
Table 2 gives the number of plants counted and their origins. With
one exception only the first root of a seedling was examined and in
many cases it was not possible to determine the shapes of all the
chromosomes. However, it was established that at least some of the
plants with 10 chromosomes had a chromosome garniture which
differed noticeably from that of alpina in the shape of one of the
small pairs. As far as could be determined the variation in number
involved a small chromosome with a head to which was attached a
rather large satellite, the 10-, 11-, 12- and 13-chromosome plants

1930] Hollingshead-Babcock : Chromosomes and Phytogeny in Crepis

13

TABLE 2

The Somatic Chromosome Numbers Found in C. alpina

VAR. SYRIACA

Origin of seed

Culture
number

Number of plants with
chromosome numbers of

10

11

12

13

Univ. Calif. Herbarium sheets 313831 and
313832*

27. 1923
28. 1923a
28. 1923b

1

Univ. Calif. Herbarium sheet 313832

6

1

1

3

27.1923-2 open pollinated

5

3

•Several p'ants collected in Galilee, Palestine, environs of the Menahamiah Company, April 20,
1324, by M. Chijik.

J.I

■>?

Sfc&

C. alpina

C. alpina

var. syrlaca (3)

C. alpina C. alpina

var. syriaca (1) var. syriaca (2)

C. alpina

var. syriaca (4)

Figure 2.

having none, 1, 2, or 3 of these chromosomes respectively. This sit-
uation recalls that of the supernumerary chromosomes in maize
(Randolph 1928a and b) and merits further study.

The chromosomes of C. rubra (fig. 3«, b) are larger than those
of C. asturica and they differ in other respects, the most noticeable
of which is the occurrence of a pair of short chromosomes with large,
often constricted, satellites. No other satellites were seen in C. rubra
1506 but in rubra 1176 one chromosome pair bore very small satellites.
M. Navashin (1925) shows the small satellites in his figure of rubra
but in strains examined later he failed to find it (unpublished obser-
vations).

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

3°$

C. rubra 1506 c. rubra 1176

a b

Figure 3.

Crepis aspera, C. bursif cilia, C. setosa, and C. senecioides (fig. 4).
with 8 chromosomes each, are quite different in details of their chro-
mosome morphology. C. setosa is characterized by the presence of a
pair of large satellites and C. senecioides is outstanding among the
8-chromosome species by the small size of its chromosomes.

C. aspera C. bursifolia

a b

C. setosa C. senecioides

c d

Figure 4.

Crepis bureniana, C. aculeata, and C. amplexifolia (fig. 5) with
8 chromosomes, are quite distinct in chromosome morphology. C.
bureniana and C. aculeata have chromosomes of the same size order;
those of C. amplexifolia. are smaller. The exact position of the fiber
attachment constriction on one chromosome of bureniana has not been
determined, owing probably to inferior fixation, but it is very near
the end.

1930] Hollingshead-Babcock : Chromosomes and Phylogeny in Crepis 15

V

r., bureniana C. aculeata C. amplexifolla

a b c

Figure 5.

The chromosome garnitures of C. myriocephala (2n = 8) and C.
lybiea (2w = 8, fig. 6a, 6) appear to be indistinguishable from each
other and differ only in their slightly larger size from those of C.
vesicaria, C. taraxacifolia, and C. marschalli (fig. 6c, d, f) which are
extremely similar to one another.

C. myriocephala c. lyblca C. vesicaria

a b c

C. taraxacifolia C. hackell c. marschalli

d e f

Figure 6.

Crepis hackeli (2n = 16, fig. 6e) is probably a tetraploid species.
Good plates showed 4 similar satellited chromosomes resembling
closely those seen in vesicaria, etc., and 4 chromosomes resembling the

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

largest of those in the same group of species. The increased number
did not permit an exact comparison of the remaining chromosomes
which are rather similar in morphology. They do, however, resemble
in shape and size the intermediate chromosomes of the vesicaria group.

# H

C. slbirlca C. pontana

a b

Figure 7.

Catonia

Crepis sibirica and C. pontana (2m = 10) have chromosome com-
plexes which differ in details but are similar in size and much larger
than those of any other 10-chromosome species examined (fig. 7). In
the plate of C. sibirica figured, one of the satellite threads has con-
tracted so that the satellite appears as a small protuberance on the
proximal end of the chromosome. This is not a constant condition.

Crepis blattarioides (2w — 8, fig. 8a) has, on the whole, chromo-
somes slightly smaller than those of other species of this phylogenetic
line. One pair bears rather large satellites. C. burejensis, C. chry-
santha, and C . conyzaefolia (fig. 8b, c, d) have similar garnitures of
8 large chromosomes, the first differing noticeably from the other two
in the shape and size of the satellited chromosome. In the plate
from which the figure of C. chrysantha was drawn, one of the satel-
lites, two of which were found in other plates, was apparently hidden
by another chromosome.

Crepis paludosa (2w = 12) is easily distinguished from the other
species of this subgenus which have been examined cytologically by
the number and shape of its chromosomes ('fig. Se).

In C. aurea (2n = 10, fig. 9a) the chromosome complex resembles
that of C. asturica, differing markedly from it in only one chromosome
pair. The chromosomes of C. bungei 1827 (2w = 8, fig. 9c) are larger
than those of C. hookeriana (2« = 8, fig. 9/;). C. bungei 2174
(2n=16, fig. 9<7) has chromosome types resembling those of C.

1930] Hottingshead-Baococlc : Chromosomes and Phylogeny in Crepis 17

bungei 1827 but they appear to be slightly smaller. Although several
of the chromosome types in the 16-chromosome race may be repre-
sented more than twice, only two satellited chromosomes were seen
and M. Navashin, from whom the material was obtained, has not seen

C. blattarioldes c. burejenaia

C. chrysantha C. conyzaefolia

C. paludosa
e

Figure 8.

more in similar material (unpublished observations). Plants of these
two accessions are now in flower. Although 1827, with 8 chromo-
somes, is more nearly typical of the species in shape of leaves and
habit of branching, yet the flower-heads and finer details of the

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

inflorescence are very similar in the two. Apparently 2174, with
16 chromosomes, is a form of this species resulting from some sort
of chromosomal variation, the precise nature of which can be deter-
mined only by further study.

C. aurea
a

C. bungei 1827 C. bungei 2174

If

C. tingitana
e

Figure 9.

C. tingitana (2n= 10, fig. 9e) has chromosomes resembling those
of C asturica but differing markedly from them in one chromosome
pair.

Eucrepis

The chromosomes of C. leontodontoides (2n = 10, fig. 10a) and
C. multicaulis (2n = 10, fig. 10b ) are noticeably different in size,
those of C. multicaulis approaching those of C. asturica, those of C.

1930] Hollingshead-Babcock : Chromosomes and Phytogeny in Crepis 19

leontodontoide s being much smaller. Both complexes differ in other
details from that of C. asturica and from each other.

C. tectorum (2n = 8), C. neglecta (2n = 8), C. parviflora (2n = 8),
C. capillaris (2w = 6), and C. nicaeensis (2w = 8) do not particularly

C. leontodontoide 9 C. multlcaulls

a

Figure 10.

b

C. parviflora C. tectorim C. neglecta

a b c

T M

C. capillaris C. nicaeensis

d e

Figure 11.

resemble each other in chromosome morphology (fig. 11). Those of
neglecta are the smallest, those of tectorum the largest of this series.
In C. biennis the chromosome number has been variously reported
as » = 16 (Marchal, 1920). n = 20 (Rosenberg. 1918), n = 21,
2n = 42 (Rosenberg, 1920), and 2n = 40 (Mann, 1925). Investiga-
tions of several plants (table 1) have shown that there is an actual
variation in number from plant to plant and this is confirmed by

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

unpublished observations of M. Navashin (cf. Collins, Hollingshead,
and Avery, 1929). Counts of 39 and 41 have been obtained and the
plant from which the plate in figure 12o was drawn had 39 chromo-
somes. Satellites have been seen on chromosomes of at least two
different types but those on only one type were visible in this figure.

C. biennis C. ciliata

a b

Figure 12.

A similar situation with respect to variation in number was
found in C. ciliata (fig. 12b) the two plants examined having 40
and 42? chromosomes respectively. The latter count is not well
established but there were more than 40 chromosomes. Again satel-
lites were seen but the large number of chromosomes precluded a

C. mollis C. lyrata

a b

Mi

C. raontana C. hierosolymitana

c d

Figure 13.

1930] Hollingshead-Babcock : Chromosomes and Phylogeny in Crepis 21

decision as to their exact number. The chromosomes of C. biennis
and C. ciliata are of the same size order.

The group of four species with 2n = 12 (fig. 13), C. mollis, C.
Iijrata, C. montana, C. hierosolymitana, gave material on which it
was somewhat difficult to make out exact details of chromosome
morphology. The chromosome outlines were not quite so clear as
in most species and the increased number added to the difficulty.
True satellites were found in montana and distal constricted ends
which often resembled satellites were found on V-shaped chromosomes
of lyrata, mollis, and hierosolymitana. Whether there are other small
true satellites has not been definitely decided.

C. nana C. elegan3

a b

Figure 14.

The material of C. nana and C. elegans (2« — 14, fig. 14«, b) was
limited to a few root tips from germinated seeds and the figures
available were anything but satisfactory for a study of chromosome
morphology though the number was quite clear. The complexes are
similar, containing long and medium V's and probably a satellited
pair of chromosomes with submedian constrictions. In each of the
figures reproduced only one satellite appears, but in C. nana other
plates were seen containing two satellites.

Extensive study of material from six plants of C. tenuifolia, grown
from seed collected from the wild, revealed always a somatic number
of fifteen chromosomes. Two of these plants have flowered with the
help of artificial light and one produced a number of achenes. Two
plants which grew from these achenes were examined and each had
15 chromosomes. Reduction divisions of the plants which flowered
were very irregular showing sometimes 15 univalent chromosomes and
sometimes both bivalents and univalents. An examination of the
somatic chromosomes (fig. 15a) shows no good evidence of hybrid
origin since many of the ohromosom.es have seemingly morphological
mates. In the plate drawn, the upturned chromosome lying beneath

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

C. tenuifolia Pterotheca sancta

a b

Figure 15.

another in the central region is probably similar to the short one in
the middle lower portion of the figure. This was the only plate
observed in which there appeared to be 4 satellited chromosomes
although 3 were seen very commonly. Whatever the origin of this
species the evidence points to some form of parthenogenesis or apog-
amy as the usual means of reproduction. This would reconcile the
irregularity in reduction divisions and low proportion of good pollen

2

*^2

C. dloscoridia C. pannonlca

a u

C. lacera C. chondrllloides

c d

Figure 16.

1930] Hollingshead-Babcoek : Chromosomes and Phytogeny in Crepis

23

with the apparently constant odd number of chromosomes and the
fairly high proportion of achenes obtained.

Pterotheca sancta (2n = 10, fig. 15b) has chromosomes of the
C. foetida type. We are indebted to Dr. M. Navashin for kindly
giving us the opportunity of reproducing the figure from his slide.
In this, the only plant examined, apparently one of the short chromo-
some pairs was heteromorphic, the spindle-fiber attachment of one
chromosome being median, and of its mate being submedian.

C. dioscoridis, C. pannonica, C. lacera, and C. chond rill aides (fig.
16) have eight chromosomes in garnitures which resemble each other
and those of C. burcjensis, C. chrysantha, and C. cony zae folia of the

C. incana
Figure 17.

subgenus Catonia (above). The chromosomes of dioscoridis seem to
be slightly smaller and those of pannonica slightly larger than those
of the other species in this group. C. incana (2« = 16, fig. 17) is
probably a tetraploid species since it has 4 similar chromosomes in
each set. The chromosomes resemble those of the group just described.
The complexes of C. reuteriana, C. praemorsa, C. incarnata and
C. palaestina (fig. 18). and of C. pulchra and C. gymnopus (fig. 19),
with 8 large chromosomes, are similar, those of C. incarnata and
C. praemorsa being probably indistinguishable. Each of these species
is characterized by the presence of one pair of long V-shaped chromo-
somes. Each has a pair of very small satellites. 'On the shortest
chromosome pair of C. pulchra in well fixed, darkly stained material
there could be found very small distal satellites. These have not been
seen in other species. This pair of chromosomes differed in shape

C. reuteriana C. praemorsa

C. incarnata c< paiae3tlna
c d

Figure 18.

C. pulchra (1) c# pulchra (2)

C. pulchra (3) C. gymnopus

c d

Figure 19.

1930] Bollingsliead-Babcock : Chromosomes and Phylogeny in Crepis 25

in the strains of pulchra examined. In one strain (1213, fig. 19c) a
plant occurred in which the attachment constriction of one of the
members of the pair was nearly median, in the other member it was
distinctly nearer the proximal end. In another plant of the same
strain (fig. 19b) in both members of the pair the constriction was
median. In plants of two other strains (1483 and 1894) each mem-
ber of the pair had the constriction nearer one end of the chromosome
(fig. 19a).

fflk

C. occidentalis 2772 C. occldentalia 1921

a

C. montlcola
c

Figure 20.

The chromosomes of the American species (except C. nana and
C. elegans) (figs. 20 to 22) are intermediate in size and the various
species show similar types of chromosomes. Satellites have been seen
in nearly every species but only in some of the 22-chromosome forms
was it possible to determine their number, 4 satellited chromosomes
having been established in several cases.

26

University of California Publications in Agricultural Sciences [Vol. 6

The C. occidentals, monticola, scopulorum and the C. gracilis, acu-
minata, barbigera groups (figs. 20 and 21) show polyploid series with
a base number of 11 (table 1). In some cases, indicated by ( ?) after
the chromosome number, the large number of chromosomes or paucity
of material has prevented an exact determination but the number

C. acuminata 1778
a

C. gracilis 2572
c

C. acuminata 1830

C. barbigera 1842

Figure 21.

given was the most likely one. In C. Occident alls, C. gracilis, and
C. acuminata, variation in number between different forms has been
found. The occurrence of the odd numbers 33 and 55 point to natural
crossing and there is abundant morphological evidence that natural
crossing, even between species, is no uncommon occurrence. The
occurrence of the numbers 33 and 44 in typical C. acuminata suggests

1930] Hollingshead-Babcock : Chromosomes and Phytogeny in Crepi.s 27

the possibility that the somatic number 22 exists in this species, too,
as it does in occidentalis and gracilis.

The C. andersoni, glauca, runcinata group (fig. 22) of which
several plants representing different accessions (table 1) have been

C. andersoni 2086 C. andersoni 2136

m

C. runcinata 2065
e

Figure 22.

examined, has given uniformly complexes of 22 chromosomes contain-
ing similar types. Whether it would be possible by a prolonged study
to reveal differences in chromosome morphology between these species
cannot be said.

28 University of California Publications in Agricultural « ■

^yncuituial Sciences [Vol.6

in certain .^gT L STCf £ """ '^ -*""'
PMogy has much in eommon ^Iwhirt .i^ 7°"°'? "k0" mor-
<Wibed so noticeably that fte"™^ toge!^ T ^
P-POses (% 23). The general ^u.e ^ h hev JST™

and Sharp„ess „, outli ^ b » ""* »M differ, s,ze

shapes of most of the eh™m„ ' " each sPecles 'he

attachments A, , a , „, 7 T ?"* ^^ °r S"bm«,ian ^
or structures rt^Tthem * eh"°mOSOmeS » "* "— -«*«

«q;, o<W^ /*^5r

Ixeris grarninea n k„-iv. ~~

a . c- Japonica

Figure 23.

c

THE PHYLOGENY OP SXXTT-SEVEN SPECIES OP CBEPXS

divisions o, the ehart, flgl X ZlZZ^ ">' "" *" Maj°"
Barkhamia the upper ri„ht Z t ^ °Ml",ies the center.
E«crepis the lo^T^^T^ T* ** """ ^ "«<
subgenus are indicated by ft * bet",!en the SPMi"s »> each

leading ,„ or towarTfte ~J„! "7*F?* "' groups and bv iines
-e number of .^ S" ^ * ^ ? *-
chromosome morphology it will be necZrv tn , , """'"""

presented in foregoing pages. ' C0"s"" the ^"'^

Paleya
confab;; Clnrlyflvtehe m°St "?M" °f ^ *» ■%«« I.

d-stribntion b::;.:,i' r;;:?;:0^,;?^^18 »f -«««'

western Europe, Abyssinia * ^ ™ddy °™ areas (south-

unlike other Creni* th» , i Himalaya) . In these species,

i^tepis, tiie involucre is not cIpavK- /I,-*
outer and inner series of h™ . * , " differentiated into

series of bracts, at least as regards length, although

1930] Eollingsliead-Babcoclc : Chromosomes and Phytogeny in Crepis 29

in one species the outer bracts differ in color from the inner ones.
The leaves are large and nearly entire and the heads are few and
large. The achenes are large, elongate, and in three or four of these
species they are more like Barkhausia achenes while in the other they
are more like those of Catonia. C. asturica is one of those with Bark-
hausia-like achenes and the marked similarity between its chromo-
somes and those of foetida and its close relatives has been noted. It
•seems reasonable to assume that Paleya was formerly a widespread
group containing more numerous species than at present and that
among these species there existed primitive forms of all three of the
other subgenera. As most of these primitive forms are now extinct,
evidence is lacking for direct connection between Paleya and some
Catonia species and between Paleya and all Eucrepis species.

Barkhausia

Of the other three subgenera, Barkhausia exhibits its relationship
to Paleya most obviously both in external morphology and in chromo-
somes, but it must be remembered that as yet the chromosomes of
only one species of Paleya have been seen. Barkhausia contains about
one-fourth of the species under discussion ; these occur in portions
of Europe, Asia, and Africa bordering on the Mediterranean. A few
of these species are of wide distribution within the area defined, but
most of them are of rather restricted range. The species of Bark-
hausia are all characterized by having definitely beaked achenes in
which the beak is usually equal to or longer than the body of the
fruit. This specialization for seed dispersal indicates considerable
advancement beyond Catonia and Eucrepis, but, as was noted above,
the line of development is well advanced in most of the present species
of Paleya. There is also present in Barkhausia a strong tendency
to have the marginal achenes different from the inner ones either in
shape or in color and texture of the pericarp or in both respects, and
in some species, like alpina, aspera, and aculeata, these differences in
outer and inner achenes are very striking. Although a few species
of Eucrepis have strikingly modified marginal achenes and it is not
uncommon in both Eucrepis and Catonia to find the marginal achenes
slightly different in shape from the inner ones, yet in neither of these
subgenera is found as great specialization of the marginal achenes as
occurs in Barkhausia. Nearly all species of Barkhausia are annuals
which may also be taken as evidence of more recent development than
in Catonia and most Eucrepis. There is much variation among these

C A T 0 N I A

BARKHAUSIA

senacioides 8

aetosa 8

capillaris
parvif lora

nicaeensis

biennis 40±/
ciliata 40±

palaestina 8
pulchra 8

tenuifolia 15

praemorsa 8
incarnata 8

gymnopus 8

occidentalis 22,44
monticola 55?
scopulorum 44?

gracilis 22,55?
acuminata 33,44,55?
barbigera 88?

glauca 22
andersoni 22
runcinata 22

Figure 24.

EUCREPIS

1930] Eollingshead-Babcocl : Chromosomes and Phylogeny in Crepis 31

species and within some of them, in leaf characters, number and size
of heads, and size of achenes.

Anisoderis, a subgroup of Barkhausia, contains the species most
resembling Paleya. This subgroup originally contained C. aljrina,
rubra, and foetida, and to these may be added two species now known
as C. alpina var. syriaca and Rodigia commutata. The close resem-
blance between the chromosomes of these species and those of G. astu-
rica has been noted. All the other species of Barkhausia have 8 (or
16) chromosomes. Their arrangement in the chart, figure 24, is
based primarily on external morphology, particularly on achene char-
acters, and the achenes of some of them have been illustrated in an
earlier paper (Babcock and Lesley, 1926). Although aspera and
aculeata are rather closely related, their very different achenes show
their relationship to different species. Thus aspera most nearly re-
sembles dioscoridis (of Eucrepis) in its marginal achenes, while its
inner ones are more like those of bursifolia, setosa, and senecioides,
but these five species are all widely separated from one another in
other characters. It may be that the aspera line and the aculeata
line arose from a common stock of distinct origin from the alpina line,
the supposed progenitor being some (extinct?) species of Paleya
having 8 chromosomes. For present purposes, however, it is more
convenient to indicate these two lines as arising from the alpina line.

The four species, bureniana, aculeata, ample xifolia, and myrio-
cephala differ from the remaining five species, grouped near the top,
in having the marginal achenes distinctly different from the inner
ones. C. bureniana most nearly resembles foetida in external mor-
phology, while the divergence is progressive until in myriocephala one
finds a highly specialized species with very numerous, small heads and
many small fruits. The other eight chromosome species of this line
are all closely related to myriocephala but are not so highly special-
ized through reduction in size of heads, flowers, and fruits. C. lybica,
indeed, has the largest heads and fruits of the six species, although
its chromosomes are the most like those of myriocephala. This may
be explained, however, as the result of genie mutations unaccompanied
by marked changes in chromosome morphology. In C. hackcli the
evidence from external morphology is in fair agreement with the
cytological evidence that this is a tetraploid or amphidiploid species.

Fig. 24. Phylogenetic chart of sixty-seven species of Crepis. The subgenus
Paleya, center, is the most primitive. In the other three subgenera, Barkhausia,
upper right; Catonia, upper left; and Eucrepis, lower half, the degree of relation-
ship is roughly indicated by connections or absence of connections with Paleya.

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

Ixeris graminca probably should not be included in Crepis.
Although its achenes are more like those of Barkhausia than they are
like those of the typical Ixeris of Cassini, yet in other details of
external morphology it is certainly closer to Ixeris and Lactuca than
to Crepis. The evidence regarding the chromosomes provides assist-
ance at this point. Although the number happens to be 16, the
chromosomes are very much smaller than in any species of Barkhausia
and in size and appearance they closely resemble those of C. bulbosa
and C. japonica which are mentioned under Catonia and Eucrepis
respectively.

Catonia

Catonia has less obvious connection with Paleya than has Bark-
hausia, yet the evidence from external morphology as well as from
chromosomes strongly indicates such an origin. The existence of one
species of Paleya (C. oligocephaJa Sch. Bip.) with Catonia-like char-
acters is especially significant. Unfortunately the chromosomes of
this species have not been seen. Catonia includes slightly less than
one-fourth of the species thus far investigated cytologically and these
are all perennials of the Old World. They are distributed in re-
stricted areas scattered throughout Europe, Asia, and northern
Africa ; the subgenus is of much wider distribution than Barkhausia.
All of these species have achenes resembling those of C. oligocephaJa
more or less, but some of them are sufficiently different in their
achenes and in other characters to necessitate the assumption of more
than one prototype. Two principal lines are indicated in the chart,
therefore, also two minor groups in each of which only one species
has been examined cytologically.

Crepis sibiriea exhibits closest resemblance to C. oligocephala in
external morphology and the other species of this line are progres-
sively distinct from the supposed prototype. The second group, aurea,
hookeriana, and bungei, are so distinct from C. oligocephala as to
necessitate the assumption that they arose from some other species.
C. tingitana, on the other hand, resembles aurea in certain respects
although very distinct in others; it does not seem improbable that the
two arose from a common prototype. C. paluclosa resembles sibiriea
and its allies in habit and leaves but the much smaller heads, and
especially the very distinct achenes, make it necessary to assume a
separate origin for paludosa.

Crepis bulbosa has long been recognized as one of the most out-
standing species of the genus. In fact it has been referred to five

1930] Hollingshead-Babcock : Chromosomes and Phytogeny in Crepis 33

other genera, including Aetheorhiza of Cassini, who treated it as a
monotypic genus. It is the only species of Crepis thus far examined
having 18 chromosomes, and, as in Ixeris graminea, the chromosomes
are very small. These facts would appear to warrant the reinstate-
ment of Cassini 's genus.

Eucrepis

These species not only show the least morphological resemblance
to those of Paleya but they also comprise the most heterogeneous of
the four subgenera. There is general resemblance within the sub-
genus, however, in the more numerous and smaller heads, the marked
reduction of the outer involucral bracts, and the mostly smaller, beak-
less achenes. Of the thirty-eight species thus far studied cytologically,
about one-third are annuals or biennials and two-thirds, perennials.
This subgenus is the most widespread in geographic distribution,
being represented in North America as well as throughout the Old
World north of the equator. As is indicated in the chart, there are
eight groups of obviously related species and some half-dozen species
which are distinct from each other and from the several groups.

Crepis leontodontoid.es and C. multicaulis are sufficiently similar
to the Catonia species, tingitana and aurea, to suggest a common
origin. The geographic distribution of three of these species is not
inconsistent with such a view. C. multicaulis, however, is widely
separated from the other three geographically. At the same time it
is among the most outstanding species in the genus because of extreme
reduction in size of inflorescence although it is perennial like the other
three. It seems probable, therefore, that multicaulis sprang from the
same ancestral stock as the other three but is a more recent species
and that there were several connecting forms which have disappeared.

The two annuals, parviflora and capillaris, are obviously close in
external morphology and occupy different adjacent geographic areas.
On the basis of specialization through reduction in size of flowers and
fruits parviflora seems to be more recent and this view would be con-
sistent with the chromosomal evidence if fragmentation of one of the
capillaris chromosomes and marked changes in the others could be
assumed. Fairly close to these are two other annuals, neglecta and
tectorum, yet both are so different in their chromosomes as to require
the hypothesis of a separate origin from each other and from the
capillaris group. Both tectorum and neglecta exhibit a marked
tendency to the production of a beak on the achene, but this is true
also in certain forms of leontoclontoid.es. It would appear that these

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

are really intermediate between Eucrepis and Barkhausia and that
in course of time they may develop definitely beaked achenes.

Another group includes nicaeensis, which is found to be either
annual or biennial, and biennis, which is usually biennial but some-
times blooms the first year. The close resemblance of the two species
has been emphasized by Bischoff (1851), yet one has 8 chromosomes
of medium size and the other about 40 rather smaller ones. It was
deemed necessary to assume a common ancestral form with 10 chromo-
somes because of the evidence from interspecific hybrids (Collins and
Mann, 1923; Collins, Hollingshead, and Avery, 1929) that C. biennis
is an octoploid species. Crepis ciiiata may also have derived its large
number of chromosomes through polyploidy, but morphologically it is
too distinct from biennis and nicaeensis to warrant the assumption of
a very recent connection.

Crepis tenuifolia is a very outstanding species of Eucrepis not
only because of its peculiarities of external morphology but also on
account of its odd number of chromosomes discussed above. The
fact that it appears to produce seed apomictically suggests the possi-
bility that it originated as a hybrid between two species with 14 and
16 chromosomes respectively, although, as has been pointed out,- the
shapes of the various chromosomes offer no good evidence for this
method of origin.

The four species with 12 chromosomes are very distinct from one
another, but have quite similar heads, flowers, and fruits, and they
are all perennials. C. mollis, with its nearly entire leaves and few,
rather large heads, appears to be the least specialized ; the other three
are about equally advanced and are correspondingly less widely dis-
tributed.

C. nana and elegans form a unique group in external morphology,
in geographic distribution, and in the number of their chromosomes.
C. nana is the most widely distributed species in the genus. It occurs
sporadically from Turkestan to northeastern Siberia and in North
America from western Alaska to northwestern Canada, in Labrador,
the Rocky Mountains, and the Sierra Nevada. C. elegans occurs in
Alaska and the Rocky Mountains. There is good morphological evi-
dence that elegans has been derived from nana and the extensive boreal
and alpine distribution of nana indicates that it is either a much older
species or a more successful species, perhaps both. There are no close
relatives among the other species which have been examined cyto-
logically.

1930] Hollingshead-Babcoclc : Chromosomes and Phylogeny in Crepis 35

Of the six species, reuteriana, praemorsa, incarnata, gymnopus,
pulchra, and palaestina, the first four are perennials and the other
two annuals. C. praemorsa and C. incarnata are very closely related
although fairly distinct from each other morphologically, and still
more distinct from their near relative, gymnopus. Similarly pulchra
and palaestina are obviously related although still more distinct from
each other. C. praemorsa, incarnata, and gymnopus bear less resem-
blance to reuteriana, however, than do palaestina and pulchra; they
probably diverged from the reuteriana line considerably earlier than
did palaestina and pulchra. Of the last two, pulchra is more highly
specialized through reduction in size of flowers and its high self-
fertility.

Crepis pannonica, laccra, and chondrilloides are obviously closely
related perennial species and they probably stand in the above order
in their phylogenetic relations to one another. This is indicated by
differences in degree of dissection of the leaves. In pannonica the
leaves are nearly entire ; in lacera they are deeply dissected ; while in
chondrilloides they are extremely finely dissected. The geographic
distribution is in agreement with this relation, since pannonica occurs
from Hungary to Turkestan, lacera occurs only in central and south-
ern Italy, and chondrilloides is restricted to a small region in Istria.

Crepis incana represents a group of perennial species which is
closely related to the foregoing although very distinct. It was prob-
ably derived from the same ancestral stock.

Crepis dioscoridis bears the evidence of its relationship to C.
aspera (of Barkhausia) in its peculiar marginal achenes and annual
habit. If it arose from the same stock as aspera it must have diverged
at a rather remote time. It is possible, of course, that this similarity
in the marginal achenes is merely a case of parallel variation but if
such is the case the ancestry of dioscoridis is still more obscure.
Although it bears some resemblance to the pannonica group, especially
in appearance of the involucre and in the morphology of its chromo-
somes yet in most respects it is very distinct from those species.

Pterotheca sancta seems to belong in Crepis, not only on the basis
of its somatic chromosomes, but also on morphological grounds, the
details of which cannot be discussed here. The origin of this species,
however, is quite as uncertain as that of dioscoridis or tenuifolia. Its
possible derivation from the Barkhausia stock is suggested by the
resemblance of its chromosomes, but morphologically it seems quite as
probable that it had a common progenitor with the reuteriana group.

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

The American species which have been examined cytologically,
excepting elegans and nana, comprise a unique group with a different
basic chromosome number (n = ll) from any other Crepis. None of
them occurs east of the Mississippi River and the center of distribu-
tion is in southwestern Canada and northwestern United States. This
suggests the probability of an Asiatic origin, but thus far no oriental
species has been found with n = ll chromosomes. Neither have
any oriental species been found which closely resemble any of these
American species. A few Asiatic species are known, however, which
resemble certain American species in some ways and it may be shown
eventually that they arose from a common prototype.

Although the problem of the phylogenetic relation of these
American species to other Eucrepis is unsolved, certain assumptions
may be warranted on the basis of chromosome numbers, external
morphology, and ecology. All nine species are perennials. Three of
them, runcinata, glauca, and andersoni, occur in moist meadows and
alkali bogs and have fleshy rootstocks. The other six occur mostly
in dry, rocky or gravelly places and have very long, slender, woody
rootstocks. The first three are equally diverse from the other six in
leaves, heads, and achenes. Thus there are two very different sub-
groups of American species and the situations with regard to chromo-
some numbers are equally diverse. Although fourteen different forms
of runcinata, glauca, and andersoni have been examined, they all have
2w=22 chromosomes. But among the other six species, both tetra-
ploidy and octoploidy occur as well as unbalanced polyploidy (33
and 55) accompanied by seed-bearing, which suggests reproduction
by some form of apomixis. The existence of two such diverse groups
in the same general region and with a common basic chromosome
number suggests a common origin but divergence must have occurred
at a comparatively remote period. Yet as compared with the Old
World species of Eucrepis, these nine are probably of relatively
recent origin. This would necessarily be the case if they were derived
from Asiatic ancestors and it is strongly indicated by the higher
chromosome numbers of this group. The probable method of origin
of these remarkable groups of species is that first suggested by Winge
(1917) and later confirmed experimentally by Clausen and Good-
speed (1925). This method consists of interspecific hybridization
followed by delayed mitosis in the fertilized egg cell or in some initial
cell of a shoot, during which the chromosomes all divide without pass-
ing to the poles. Thus a complete diploid set of chromosomes from

1930] Hollingshead-BabeocJc: Chromosomes and Phytogeny in Crepis 37

both parents is provided after which mitosis proceeds regularly.
Some such amphidiploid hybrids are known to be vigorous and fully
fertile. It is only necessary to assume natural crossing between two
species with n = l and » = 4 chromosomes, followed by amphi-
diploidy, to explain the origin of species with 22 chromosomes.
Because of the diversity between the two groups of American species,
however, it seems necessary to assume that at least three different
species were involved in the original crosses. Some of the other
numbers which occur in the Eucrepis series may also have originated
through amphidiploidy.

Crepis japonica (Youngia lyrata Cass.) is the type species of
Cassini's genus. As in the case of Ixeris graminea and Aetheorhiza
bulbosa, the chromosomes are much smaller than Crepis chromosomes.
This evidence, together with that from external morphology and geo-
graphic distribution, will probably be considered sufficient to warrant
the reinstatement of Cassini's Youngia as that genus was originally
defined.

Summarizing briefly with respect to subgeneric relationship,
Paleya is a small group of species which appear to be more primitive
than all other Crepis at least in certain characters. There are two
distinct subgroups of Paleya, viz., species with Barkhausia-like
achenes and one species with Catonia-like achenes. Barkhausia con-
tains species which show strong resemblance to certain Paleya -species,
although certainly more highly specialized. At the same time, within
Barkhausia are found the highest degree of specialized structure in
the beaked achenes and the largest amount of adaptation through
development of the annual habit. The combination of these two lines
of specialization, however, while advantageous in particular environ-
ments, has not resulted in wide distribution of the subgenus as a
whole. This restriction of adaptation accompanying marked special-
ization must be considered along with the evidence that many species
of Barkhausia appear to be of recent development. In Catonia is
found a group of perennial species some of which exhibit evidence
of relationship with an existing species of Paleya while for others
it is necessary to assume other prototypes. The species of Catonia
are less highly specialized than those of Barkhausia and Eucrepis,
and most of the species are of comparatively restricted distribution.
Eucrepis, the largest, most widely distributed, and most heterogeneous
division, exhibits connections with both Barkhausia and Catonia in
certain of its species but there is no evidence, among the species dis-

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

cussed in this paper, of close connection with Paleya. It seems most
probable that the prototypes of Eucrepis and the present or earlier
species of Paleya arose from a common stock subsequent to its differ-
entiation from closely related genera such as Hieracium and Lactuca.

CHROMOSOMES AND PHYLOGENY

The cytologist or geneticist cannot fail to have noticed in the
literature of recent years an increasingly close relationship between
the study of the chromosomes and of external morphology with a view
to classification. The impression is rather strong that in most cases
it is the cytologist or the geneticist who has realized how useful or in
some cases how fundamental his findings are to the elucidation of
phylogenetic problems, and so has been led by a combined study of
chromosomes and external morphology to formulate classification
systems and phylogenetic hypotheses or to criticize those based on
external morphology alone. It is to be hoped that taxonomists will
come to the same realization and that in the future it will be con-
sidered as desirable to know the nature of the chromosomes of any
plant as it is to be familiar with the details of its external morphology,
particularly in cases where critical decisions are necessary.

Among the most outstanding cases where a knowledge of chromo-
some numbers has been useful in classification is that of the genus
Rosa. Hurst (1925) utilized the cytological findings of Tackholm
(1920, 1922) and Blackburn and Harrison (1921) in the classification
of this genus, which in spite of the life-long labors of several botanists
was still in an admittedly unsatisfactory state. These investigations
had shown that there was in the genus a multiple series of chromo-
some numbers with a base number of seven. With this in mind a
workable system was formulated, the most outstanding feature of
which was that the whole Caninae section was characterized at reduc-
tion division by a condition often seen in hybrids, viz., the presence
of univalent as well as bivalent chromosomes.

Quite different but equally interesting results were obtained by
Heilbom (1924) in Carex. When the chromosome numbers were
arranged according to the taxonomic relationship within the genus
groups of adjacent numbers were found, nearly related species having
numbers of about the same magnitude. In the genus Triticum a
remarkable agreement in classifications based on morphology, resist-
ance to disease, serological reactions, and chromosome numbers is

1930] Hollingshead-Babcoclc: Chromosomes and Phytogeny in Crepis 39

evident (Sax, 1921, 1923). Clausen (1927) found that Viola species
of the same systematic subgroup belonged as a rule to the same series
of chromosome numbers and in the Melanium section of the genus he
made a new subdivision on the basis of chromosome numbers. He
believes his investigations afford further proof of the great importance
of cytology as an aid to taxonomic research.

It is the intention of the writers to do no more than point out a
few of the many instances in which a knowledge of chromosome num-
bers has either aided in the making of a classification system or has
added weight to one already made. It is true that in some instances
a study of chromosome numbers has not helped greatly in classifica-
tion, for species widely different morphologically may have the same
chromosome number, as in Crepis. When such is the case this study
has shown that a knowledge of the morphology of the chromosomes
may be very illuminating. The chromosome number 8 occurs in
thirty-two out of the seventy species discussed and in three of the
four subgenera. The possibility is good that it occurs also in Paleya.
But a comparative study of chromosome morphology has shown that
these species with the same chromosome number are usually char-
acterized by very different chromosomes.

A study of chromosome number and morphology in relation to
classification was carried out by Sveshnikova (1927) in thirty species
of Vicia. Three main groups — Ervum with 14 chromosomes in all
species, Cracca with 12, 14, 24, and 28, and Euvicia with 12 and 14 —
were represented in the study. The chromosomes were divided into
four groups according to the relative lengths of the arms and the
presence of satellites and each species was described on this basis.
When the species were arranged in groups according to the numbers
of chromosomes of different classes, the classification was in general
agreement with that usually applied by systematists. Indeed this
investigator was able to make a key based on chromosome number and
morphology which corresponded very nearly with one worked out by
Ascherson on external morphology.

Wexelsen (1928), however, found no such parallelism in the
differentiation of chromosome complexes and external morphology
in his studies on chromosome numbers and morphology in eighteen
species of Trifolium. Species which were far removed taxonomically
and very different in morphology had very similar chromosome com-
plexes, and very nearly related species had very different complexes.
This he believes presents a very clear demonstration of parallel vari-

40 University of California Publications in Agricultural Sciences [Vol. C

ation, for example the presence of one pair of satellited chromosomes
in many species would be due to independent parallel mutations and
not to the fact that they have been derived from a common source.

In most instances in Crepis similarity of chromosomes is most
certainly associated with a common phylogenetic origin, although it
may be necessary in one outstanding case at least to assume parallel
variation to account for the similarity of the chromosomes of species
far removed phylogenetically. We refer to the resemblance of the
chromosomes of the conyzae folia, chrysantha, burejensis group of
Catonia to those of the pannonica, lacera, chondrilloides group of
Eucrepis.

It is interesting to note what kinds of changes with respect to
chromosome number and morphology the phylogenetic chart in figure
24 involves. The first difficulty we are faced with is the fact that only
one species of the subgenus Paleya, which is believed to contain or to
have contained the progenitors of the species of the other subgenera,
has been examined cytologically. It seems probable that, since each
of the other subgenera shows variation between species in the number
of chromosomes, the same would be found to occur here. For the two
lines (one in Barkhausia, one in Catonia), however, where the con-
nection with Paleya is clearest ten is apparently the most primitive
number. In each line one or more reductions to eight have occurred.
This is true also of the second main line of Catonia.

The 8-chromosome species of Eucrepis could be assumed to have
arisen from 8-chromosome Paleya species not examined cytologically
or now extinct. If so it is necessary to assume that within Paleya a
reduction from 10 to 8 or an increase from 8 to 10 chromosomes has
occurred during the evolution of the species of this subgenus. Within
Eucrepis the relations depicted in the chart (fig. 24) involve a reduc-
tion to 8 from the supposed 10-chromosome progenitor of nicaeensis
and possibly a reduction to 6 from an 8-chromosome progenitor of
capillaris.

The two known methods by which a reduction in number of chro-
mosomes may take place are (1) the elimination of a pair of chromo-
somes following irregularities in meiosis and (2) the fusion of non-
homologous chromosomes. Examples are known where plants pos-
sess one pair of chromosomes less than the normal diploid set as in
Triticum (Kihara, 1924, 1925), Avcna (Iluskins, 1927), and Primula
kewensis ■ (Newton and Pellew, 1929), but the species in question are
admittedly of polyploid origin and the plants are partly or wholly

1930] Hollingshead-Babcock : Chromosomes and Phylogeny in Crepis 41

sterile. In other genera where monosomic plants are known (for
example, in Datura and Nicoliana), which would be expected to give
progeny lacking one chromosome pair, such progeny have not been
reported. Moreover, M. Navashin's extensive study (1926, 3929)
which included several thousand plants 'of three species of Crepis,
failed to find a plant with even one chromosome missing. In view of
these facts the explanation of reduction in chromosome number by
elimination of a pair is, to say the least, questionable.

Similarly, evidence is lacking in this genus that a decrease in
number could have resulted from end to end fusion of non-homologous
chromosomes. M. Navashin found no instance of it in his investiga-
tions and neither does a study of the chromosomes of the various
species in which a reduction is supposed to have occurred show any
evidence of such a process. To take an extreme case, the 8 chromo-
somes of senecioides, presumably derived from a 10-chromosome stock,
are on the whole smaller than those of the supposed stock from which
they have sprung and certainly no one of them is long enough to
represent simply an end to end fusion of two chromosomes of the
parental stock.

It has been suggested in the animal kingdom by Wilson (1925)
and Painter (1925) and in the plant kingdom by Delaunay (1926)
and favorably taken up by Jaretsky (1928) that the occurrence of
small chromosomes in a complex may represent an intermediate step
in the process of gradual diminution in the size of a chromosome
which terminates with its disappearance. Rather small chromosomes
are not rare in Crepis although none have been found as small as
those which first prompted this hypothesis.

It is also theoretically possible, although not very probable in
species having such low chromosome numbers, that such a reduction
in number might result from hybridization between two 10-chromo-
some species. This necessitates the assumption that such hybrids
occasionally produce functional gametes having only 4 instead of the
normal 5 chromosomes. Self-fertilization in such a hybrid might
rarely produce new stable forms having 8 chromosomes. Thus far the
experimental evidence from interspecific hybrids is against such a
hypothesis (Navashin, 1927; Hollingshead, 1930).

The phylogenetic chart involves as well the assumption of numer-
ous instances of increase in chromosome number. In some of these
cases, such as haekelei, bungei 2174, and incana, all with 16 chromo-
somes, each species is morphologically closely related to certain 8-

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

chromosome species. Such cases are readily explained by chromosome
doubling. Perhaps in the case of hackeli (from external morphologi-
cal evidence), and of bungei 2174 (from the fact that no more than
two satellited chromosomes were seen) the doubling followed hybrid-
ization between two 8-chromosome species. It has been pointed out
earlier that a 'similar explanation may account for the chromosome
numbers of biennis, ciliata, and the American species with a base
number of 11, and the same explanation might conceivably be applied
to the origin of the species with 12 and 14 chromosomes. Such an
explanation would involve the assumption that species with 6 chromo-
somes featured in the ancestry of each of the last two groups. This
seems rather unlikely since the one species known which has 6 chromo-
somes is, judging by its external morphology, of comparatively recent
origin. It seems more probable that they have arisen from species
with lower chromosome numbers by some method other than one
involving a doubling of all the chromosomes.

How to account for an increase in number of one or more pairs
of chromosomes is another problem. In many instances the evidence
from chromosome morphology, as pointed out by Navashin (1925),
is not in accord with an explanation which supposes the simple dupli-
cation of a pair of chromosomes already present as a result of irregu-
lar meiotic behavior, as suggested by Rosenberg (1918, 1920), for
many species in which an increase is supposed to have occurred show
no two pairs morphologically alike. Moreover, although M. Navashin
(1926) found trisomies rather frequently in two Crepis species he
has never found tetrasomics in their progeny (unpublished data).
It must be borne in mind, too, that experimental evidence has shown
that the tetrasomics investigated up to the present, as in Datura
(Blakeslee and Belling, 1924), Avena (Huskins, 1927), and Triticum
(Huskins, 1928) are usually much less viable than normal diploid
plants and even when equally viable as in Nicotiana (Clausen and
Goodspeed, 1924) they do not breed true. In this connection, however,
the occurrence of the numbers 10, 11, 12, and 13, with the variation
probably involving only a single chromosome type, in apparently typical
and quite viable plants of C. alpina var. syriaca is of great interest.
One may assume that it will be possible to isolate constant 10-. 12-, and
possibly even 14-chromosome races from this species. Whether the
supernumerary chromosome is a relic of some hybrid ancestry, or
whether it represents the phenomenon designated by M. Navashin as
"novation," or whether it represents part of a normally 12-ehromosome

1930] Hollingshead-Babcoclc : Chromosomes and Phytogeny in Crepis 43

complex are as yet matters of speculation only. Instances are known
in the genus where a chromosome from one species has been added
to the normal complex of another as a result of hybridization
(Hollingshead, 1930) but the plants are weak and sterile. Similarly
M. Navashin's plants which contained apparently new chromosomes
of unknown origin were inviable and it has been pointed out above
that plants lacking one pair of chromosomes (except in polyploid
species) have not been found. Of the three suggested hypotheses to
account for this situation in syriaca, the first (hybrid origin) appears
to be the least improbable.

In some few instances the hypothesis of transverse segmentation
can be advanced to account for increased chromosome number in this
genus, as suggested by Mann (1925). This hypothesis is believed to
account for an increased number of chromosomes in certain other
genera, for example in Secale (Gotoh, 1924, and Belling, 1925) and
in other cases (cf. Kuhn, 1928). M. Navashin (1926) found a C.
tectorum plant in which one chromosome had fragmented and each
part behaved as a new chromosome in mitosis. This process has been
suggested (above) as having played a part in the evolution of C.
parviflora from capillaris. Such a hypothesis, however, while ex-
plaining adequately chromosome lengths, fails to take into account
differences in the shapes of the various chromosomes and must even
in this instance have been accompanied or followed by changes in
position of spindle-fiber attachment constrictions, which will be dis-
cussed later.

In addition to changes involving increase and decrease in number
of chromosomes, the chart necessitates the assumption that changes
in chromosome size have occurred during the evolution of the various
species. In Muscari, Delaunay (1926) came to the conclusion that
diminution in chromosome length by a slow process called by him
"historiation" had occurred in the evolution of the chromosomes of
the species he studied. Jaretsky (1928) favors the same process as
a mode of chromosome evolution from his studies on Polygonaceae.
Heitz (1928) recognizes changes in length of chromosomes as of
fundamental importance in the evolution of chromosomes. Sveshni-
kova (1927) concluded that in Vicia, among the processes which had
occurred during the evolution of the species she studied, there had
been a reduction of chromatin connected with variation in external
morphological features.

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

This study would seem to show that both increase and decrease
in chromosome size have been relatively frequent occurrences in the
evolution of the various species of the genus Crepis. Morphologically
similar species usually have similar chromosomes and it is only within
such closely related groups that it is safe to assume which chromo-
somes are descended from a common ancestor. In the vesica ria,
lybica, etc., group, which have very similar chromosome complexes,
it is possible to pick out lybica by the uniformly slightly larger size
of its chromosomes. In this connection it is worth noting that Wexel-
sen found two varieties of Tri folium repens to have chromosomes of
different size, one having uniformly larger chromosomes than the
other. Though it is true that, generally speaking, the various Crepis
species have "large" or "small" chromosomes, it does not follow
that increase or decrease in length has similarly affected all chromo-
somes in a complex. This is obvious from a glance at such a complex
as that of setosa which contains a very short pair along with compara-
tively long pairs of chromosomes. The evidence from bursifolia-
aspera hybrids (Babcock and Clausen, 1929) indicates that the sat-
ellited chromosomes of these species are homologous. Most of the
chromosomes in the two species are of the same size order but in this
particular chromosome pair the species differ very markedly, aspera
having long, and bursifolia unusually short, satellited chromosomes.

As has been noted earlier, cross-division of chromosomes is not
unknown in the genus and as M. Navashin (1926) has pointed out,
such a process might be followed by elimination of part of the frag-
mented chromosome. Such an occurrence could be supposed to
account for the difference in size between apparently homologous
chromosomes of different species were it not that no evidence is, avail-
able to show that such a loss can take place and the individual still
survive. Further, it has been shown that several authors favor a
gradual change in chromosome size as the way in which chromosomes
in related species come to differ from each other. Navashin (1926)
pointed out that differences in satellite size between races of a single
species may illustrate this kind of change and stated that the origin
of new distinct species could probably be explained by the continual
addition of such small changes. The fact that chromosomes of all sizes
intermediate between largest and smallest are to be found in this
genus would seem to offer good evidence for this theory.

Lastly, any evolutionary hypothesis must take into account
changes in chromosome shape which are largely determined by the

1930] Hollingshead-Babcocl- : Chromosomes and Phylogeny in Crepis 4.3

presence or absence of satellites and the position of the spindle-fiber
attachment. In this connection cases in which homologous chromo-
somes are different (heteromorphic pairs) are instructive. They
would seem to offer evidence that rare changes in chromosome shape
may occur without markedly affecting the external morphology.

Size differences in satellites on homologous chromosomes have been
reported by a number of investigators (cf. Kuhn, 1928). In Crepis,
M. Navashin found races of C. dioscoridis with two large satellites,
with two small satellites, and with one large and one small satellite.
In Rumex scut at us, Jaretsky found some plants in which one of the
chromosomes of a pair have satellites and others in which there were
no satellites at all. Such a condition could be expected to lead to a
race with both members of the pair bearing satellites. The sig-
nificance of such findings for evolution of chromosomes of a different
type is obvious.

Heteromorphism, which involves a difference in position of spindle-
fiber attachment, is well known in the animal kingdom. In plants,
in addition to the instances of sex chromosomes which may differ both
in size and shape, a few cases of heteromorphic autosomes have been
found. In different samples of both Vicia angustifolia and V. sativa,
Sveshnikova (1927) found the same chromosomes showing slightly
different ratios in arm length. Two cases of heteromorphic pairs are
reported in this paper, viz., Pterotheca sancta and Crepis pulchra.
The case of C. pulchra is particularly interesting for here plants with
two chromosomes of each kind and one with one of each were found.
The first two of these races could, conceivably, if isolated, develop in
the course of time into different species each characterized by its
particular chromosomes.

Heitz (1928) holds that the primitive chromosome form is the
equi-armed one and that asymmetrical shortening (or lengthening)
gives rise to chromosomes with unequal arms. Equi-armed chromo-
somes occur in each subgenus and in many species of Crepis. Usually
they are the smaller chromosomes of the complex, although they may
be the largest as in the pulchra, palaestina, etc., group. There is no
consistent evidence in Crepis to support Heitz 's theory that such
chromosomes are more primitive, for although they do occur perhaps
more frequently in some of the supposedly older species (C asturica,
C. foetida) with low numbers, yet they also occur in the highly
specialized C. senecioidcs.

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

Heitz's theory would also imply that a change in the form of a
chromosome results only from a change in length of one of the arms.
Sveshnikova does not state whether the chromosomes which showed
different ratios of arm length in her material were the same length
or not. In the cases observed by the writers, however, the hetero-
morphic pairs were nearly or quite the same length and it seems more
probable that the change in form has resulted from a shifting of the
attachment constriction rather than from a lengthening of one arm
and a shortening of the other.

Summarizing, the following changes in chromosome numbers must
be assumed to have occurred during the evolution of these sixty-seven
species of Crepis, if the phylogenetic grouping proposed here be
accepted. (1) Reduction in number from 10 to 8 and from 8 to 6,
changes which it is difficult to explain in the light of our present
knowledge. It is possible, however, that such changes might have come
about through a process of gradual diminution of a particular pair of
chromosomes or that interspecific hybridization or fusion of non-
homologous chromosomes might have given rise to a new species with
the reduced number, although there is no experimental evidence for
such hypotheses. (2) Increase in number from 8 to 16, a change
which may have occurred in either of two ways, viz., doubling of the
2n group resulting in true tetraploidy ; and doubling in a hybrid
between two 8-chromosome species, resulting in amphidiploidy. (3)
Increase from 10 to 40 or thereabouts (biennis, cilia ta?) which accord-
ing to experimental evidence must have resulted from once repeated
doubling of the 2» group producing octoploidy. (4) Increase from
22 to 33, 44, 55, and 88. The even numbers probably arose through
chromosome doubling as described above and the odd ones may have
arisen through hybridization between even numbered forms or by the
fusion of somatic gametes with normal ones. (5) Increase from lower
numbers to 12, 14, 15, and 22 chromosomes, which in some cases may
have resulted from transverse segmentation. The higher even num-
bers could be explained by the assumption of amphidiploidy and the
odd one could conceivably have arisen as a result of hybridization,
though again experimental evidence for such an origin is lacking.
In addition to changes in number many changes in chromosome size
apparently have occurred and to the present writers it seems probable
that most of these changes have taken place gradually during the
phylogenetic differentiation of the species. Lastly there have been
many changes in chromosome shape, and the occurrence of races

1930] Hollingshead-Babcoclc : Chromosomes and Phytogeny in Crepis 47

within a species which differ in the shape of one chromosome pair and
of plants with heteromorphic chromosome pairs indicates that this is
not a very uncommon occurrence in phylogeny.

On the whole it seems probable that a number of different mechan-
ical processes affecting chromosome organization have played a part
in the evolution of chromosomes in this genus. Processes such as
fragmentation, union, inversion, deletion, translocation, and dupli-
cation have occurred in Drosophila. Painter and Muller (1929) and
Muller and Painter (1929) have recently described changes in chro-
mosome shape resulting from such processes actually induced by
irradiation with X-rays. These studies are extremely significant in
connection with problems of chromosomes and phylogeny, for such
processes can account for changes in size and shape and increase and
decrease in number of chromosomes. In particular they offer a pos-
sible explanation for the 6, 8, 12 portion of the Crepis chromosome
series whose manner of origin from a supposed 10-chromosome ances-
tor presents the most difficult problem in the study of the evolution
of chromosomes within the genus.

SUMMARY AND CONCLUSIONS

The number and morphology of the somatic chromosomes of
seventy species (including three which have been considered Crepis
but will probably be assigned to other genera) were investigated in
connection with a taxonomic study of the genus. The study of
chromosome morphology has increased materially the value of the
cytological findings in connection with classification.

Size differences, the occurrence of satellites, and shape as deter-
mined by spindle-fiber attachment were used to distinguish the various
chromosomes.

The species discussed include twenty-seven whose chromosome
numbers have not previously been counted and a number which have
not been figured previously. A drawing showing a somatic metaphase
which depicts as far as possible details of morphology is included for
each species. The known series of chromosome numbers of Old World
species is 6, 8, 10, 12, 14, 15, 16, and 40±. That of the American
species is 14, 22, 33, 44, 55?, and 88?. Without doubt these latter
numbers are members of a polyploid series, as are some of the num-
bers in the Old World group. Others probably originated through
interspecific hybridization. There is considerable evidence pointing

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

to 10 as the basic number. An alternative hypothesis is the assump-
tion that 8 is the more primitive number in Crepis. No evidence exists
at present to support this assumption.

In regard to size of chromosomes, there is a range of sizes, the
extremes of which may be roughly expressed by the ratio 1:2. It
should be noted, however, that the species having the smallest chromo-
somes (senecioides, nana, elegans, and leontodontoides) comprise only
a small fraction of the entire number of Crepis species thus far
studied. In comparison with these the chromosomes of the remain-
ing species may be designated as medium, medium-large, and large,
but within each of these categories there is considerable variation so
that the entire list of species could be arranged in a nearly continuous
series on the basis of chromosome size. Species representing the
extremes in size differences occur in Eucrepis, while Barkhausia con-
tains one species having chromosomes of the smallest size, and Catonia
has several species with very large chromosomes. The variation in
chromosome sizes, therefore, is not peculiar to any one subgenus ; on
the contrary it is distributed throughout the genus.

With regard to shape .of the chromosomes, while there are many
minor differences between individual species, there is a general simi-
larity in all the species. Occurrence of satellites and position of
constrictions are the most useful differences in shape. The occurrence
of at least one pair' of satellited chromosomes is practically constant
throughout all the species. About one-third of the species have at
least one pair of chromosomes with approximately median constric-
tions. In all other chromosomes the constrictions are located at some
point nearer the proximal end of the chromosome.

A chart was drawn up to show the four major subdivisions of the
genus and the chromosome numbers of the species discussed. This
chart depicts the phylogenetic groups which were worked out by com-
bining data on chromosome number and morphology with evidence
from external morphology, consideration having been given to geo-
graphic distribution and evidence from interspecific hybridization.
The chromosomes of the species in the various phylogenetic groups
were discussed and their resemblances and differences noted.

In Paleya, considered the most primitive subgenus and supposed
to contain or to have contained the progenitors of the other subgenera,
only one species has been examined and it has 10 chromosomes. Prob-
ably further study would discover other chromosome numbers in this
section. The subgenus Barkhausia contains species witli 8. 10. and Hi

1930] Hollingshead-Babcock : Chromosovies and Phylogeny in Crepis 49

chromosomes, and these same numbers and one 12-chromosome species
are found in Catonia. Eucrepis, the largest subgenus, contains the
greatest number of species and those whose connection with Paleya
is least obvious. The numbers 6, 8, 10, 12, 14, 16, and 40 are found
in the Old World species of the subgenus. With the exception of
two closely related species, one of which also occurs in the Old World,
the American species form a polyploid series with a base number of
11. These facts, together with the heterogeneity of Eucrepis, lead to
the inference that this subgenus consists of a number of related phylo-
genetic lines and that the earlier connecting forms and the more
primitive ancestors have all disappeared. Several of these diverse
subgroups under Eucrepis are represented among the species thus far
studied by individual species while others contain from two to six
species which are obviously closely related both from their external
morphology and their very similar chromosomes.

Two instances of heteromorphic chromosome pairs involving dif-
ferences in point of spindle-fiber attachment were found (C. pulchra
and Pterothcca sancta) and in the former species, in addition to the
plant showing this condition, other plants with a pair of each of the
different chromosomes were found. A difference in size of satellites
on homologous chromosomes was noted in C. foetida.

In one case (C. alpina var. syriaca) variation in chromosome
number from 10 to 13 was found and the variation appeared to
involve one particular chromosome type.

The phylogenetic system proposed involves the assumption of
several different kinds of chromosome changes. They include increase
and decrease in chromosome number and increase and decrease in
size and change in shape. Ways in which these changes may have
come about are discussed and the most likely ones are pointed out.

In general, in each section of the genus, morphologically similar
species have similar chromosomes and the writers are more firmly
convinced by this investigation of the value of such a study of
chromosomes in relation to taxonomy. Certainly there is a fairly
close parallelism in Crepis between number and morphology of the
chromosomes and phylogenetic relationship.

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

LITERATURE CITED

Babcock, E. B.

1924a. Genetics and plant taxonomy. Science, 59:327-28.
1924&. Species of Crepis. Science, 60:175-76.

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

1926. Chromosome number and individuality in the genus Crepis. II. The

chromosomes and taxonomie relationships. Univ. Calif. Publ. Agr.
Sci., 2:315-41.

Babcock, E. B., and Clausen, J.

1929. Meiosis in two species and three hybrids of Crepis and its bearing on
taxonomie relationship. Univ. Calif. Publ. Agr. Sci., 2:401-32.

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

1929. Crepis reuteriana and its chromosomes. Science, 69:356.

Belling, J.

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

Bischoff, G. W.

1851. Beitrage zur Flora Deutschlands und der Schweiz (Heidelberg), pp.
iii-xv and 244-329.

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

1921. The status of British rose forms as determined by their cytological be-
haviour. Ann. Bot., 35:159-88.
Clausen, J.

1927. Chromosome number and the relationship of species in the genus Viola.

Ann. Bot., 164:677-714.

Clausen, B. E., and Goodspeed, T. H.

1925. Interspecific hybridization in Nicotiana. II. A tetraploid glutinosa-

tabacum hybrid — an experimental verification of Winge's hypothesis.
Genetics, 10:278-84.

Collins, J. L., and Mann, M.

1923. Interspecific hybrids in Crepis. II. A preliminary report on the re-

sults of hybridizing Crepis setosa Hall, and C. capillaris (L.) Wallr.
and with C. biennis L. Genetics, 8:212-32.

Collins, J. L., Hollingshead, L., and Avery, P.

1929. Interspecific hybrids in Crepis. III. Constant fertile forms contain-
ing chromosomes derived from two species. Genetics, 14:305-20.

Delaunay, L. N.

1926. Phylogenetische chromosomenverkiirzung. Zeit. f. Zellforseh. u. mikros.

Anat., 4:338-64.
Gotoh, K.

1924. tiber die Chromosomenzahl von Secale cereale L. Bot. Mag., Tokyo,

38:135-52.

Hall, H. M., and Clements, F. C.

1923. The phylogenetic method in taxonomy. Carnegie Inst., Wash., Publ.,
326:3-30.

1930] Hollingshead-Babcoclc: Chromosomes and Pltylogeny in Crepis 51

Heilborn, O.

1924. Chromosome numbers and dimensions, species formation and phylogeny

in the genus Carex. Hereditas, 5:129-216.

Heitz, E.

1928. Der bilaterale Bau der Geschlechtschromosomen und Autosomomen bei
Pellia Faooroniatia, P. epiphylla und einigen anderen Jungcrmanni-
aceen. Planta (Arch. f. wissensch. Bot.), 5:725-68.

HOLLINGSHEAD, L.

1930. Cytological investigations of hybrids and hybrid derivatives of Crepis
capill-aris and C. tectorum. Univ. Calif. Publ. Agr. Sci., 6, 2 (in
press).

Hurst, C. C.

1925. Chromosomes and characters in Rosa and their significance in the origin

of species. Experiments in Genetics (Cambridge University Press),
38:534-50.

Huskins, C. L.

1927. On the genetics and cytology of fatuoid or false wild oats. Jour.

Gen., 18:315-64.

1928. On the cytology of speltoid wheats in relation to their origin and

genetic behaviour. Jour. Gen., 20:103-22.

Jaretsky, B.

1928. Histologische und karyologische Studien an Polygonaceen. Jahrb. f.
wiss. Bot., 66:357-490.
Kihara, H.

1924. Cytologische und genetische Studien bei wiehtigen Getreidearten mit

besonderer Riicksicht auf das Verhalten der Chromosomen und die
Sterilitat in den Bastarden. Mem. Coll. Sci. Koyoto Imp. Univ.,
Ser. B, 1:1-200.

1925. Weitere Untersuchungen iiber die pentaploiden rn7i>«?>i-Bastarde. I.

Jap. Jour. Bot., 2:299-304.
Kuhn, E.

1928. Zur Zytologie von Thalictrum. Jahrb. f. wiss. Bot., 68:382-430.

DE LlTARDIERE, R.

1923. Sur 1 'insertion fusoriale des chromosomes somatiques. Bui. Soc. Bot.
France, 70:193-97.

Mann, M. C.

1922. Report of the College of Agriculture and Agricultural Experiment
Station of the University of California, 1921-22:96-102.

1925. Chromosome number and individuality in the genus Crepis. I. A
comparative study of the chromosome number and dimensions of
nineteen species. Univ. Calif. Publ. Agr. Sci., 2:297-314.

Marchal, E.

1920. Recherches sur les variations numeriques des chromosomes dans la
serie vegetale. Mem. de l'Acad. Roy. de Belgique, ser. 2, 4:1-108.

Muller, H. J., and Paintfr, T. S.

1929. The cytological expression of changes in gene alignment produced by

X-rays in Drosophila. Am. Nat., 63:193-213.

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

Navashin (Nawaschin), M.

1925. Morphologische Kernstudien der Crepis-nrten in Bezug auf die Artbild-

ung. Zeit. f. Zellforseh. u. mikrosk. Anat., 2:98-111.

1926. Variabilitat des Zellkerns bei Crepis-axten in Bezug auf die Artbild-

ung. Ibid., 4:171-215.

1927. trber die Veranderung von Zahl und Form der Chromosomen infolge

der Hybridisation. Ibid., 6:195-233.
1929. Studies on polyploidy. I. Cytological investigations on triploidy in
Crepis. Univ. Calif. Publ. Agr. Sei., 2:377-400.

Navashin (Nawaschin), S.

1912. On nuclear dimorphism in the somatic cells of Gultonia candicans.

Bui. Imp. Acad. Sci., St. Petersburg, 22:373-85 (Russian).
1916a. Sur 1 'alteration brusque comme facteur dans les variations considerees

comme mutation. Moniteur du Jardin Botan. de Tiflis, annee 12,

livre 3. (Russian with French summary.)
1916b. Sur quelques indices de 1 'organization interieur des chromosomes.

Receuil d 'articles scientiliques dedie an Cement Timiriazeff, Moscow

(Russian).

Newton, W. C. F., and Pellew, C.

1929. Primula kewensis and its derivatives. Jour. Gen., 20:405-67.

Painter, T. S.

1925. A comparative study of the chromosomes of mammals. Am. Nat., 59:
385-409.

Painter, T. S., and Muller, H. J.

1929. Parallel cytology and genetics of induced translocations and deletions
in Drosophila. Jour. Hered., 20:287-98.

Randolph, L. F.

1928a. Chromosome numbers in Zea Mays L. Mem. Cornell Univ. Agr. Exp.

Station, 117:1-44.
1928b. Types of supernumerary chromosomes in Maize. Anat. Rec, 41:102.

Rosenberg, O.

1918. Chromosomenzahlen und Chromosomendimensionen in der Gattung
Crepis. Arch. f. Bot., 15:1-16.

1920. Weitere Untersuehungen iiber die Chromosomenverhaltnisse in Crepis.

Svensk. Bot. Tidskr., 14:319-26.
Sax, K.

1921. Sterility in wheat hybrids. I. Sterility relationships and endosperm

development. Genetics, 6:399-416.
1923. The relation between chromosome number, morphological characters
and rust resistance in segregates of partially sterile wheat hybrids.
Genetics, 8:301-21.

DE SMET, E.

1913-14. Chromosomes, prochromosomes et nucleole dans quelques Dicotylecs.
Cellule, 29:335-77.

Sveshnikova, D. N.

1927. Karyological studies in Vicia. Bui. Applied Bot., Genetics, and Plant
Breeding, 17:37-72. (Russian with English summary.)

1930] Eollingshead-Babcoc'k : Chromosomes and Phylogeny in Crepis 53

Tackholm, G.

1920. On the cytology of the genus Rosa. Svensk. Bot. Tidskr., 14:300-11.
1922.' Zytologisehc Studien iiber die Gattung Bosa. Acta Horti Bergiani,
7:97-381.

Taylor, W. E.

1924. Cytological studies on Gasteria. I. Chromosome shape and individual-

ity. Am. Jour. Bot., 11:51-9.
1925«. The chromosome morphology of Velthemia, Allium and Cyrtanthus.

Ibid., 12:104-15.
1925b. Chromosome constrictions as distinguishing characteristics in plants.

Ibid., 12:238-44.

1926. Chromosome morphology in Fritillaria, Alstroemeria, Silphium, and

other genera. Ibid., 13:179-93.

TlSC'HLER, G.

1927. Pflanzliche Chromosoinenzahlen. Tabulae Biologicae, 4:1-83.

W EX el sen, H.

1928. Chromosome numbers and morphology in Trifolium. Univ. Calif. Publ.

Agri. Sci., 2:355-70.

WlNGE, O.

1917. The chromosomes — their numbers and general importance. Compt.
Eend. Trav. Lab. Carlsberg, 13:131-275.

Wilson, E. B.

1925. The cell in development and heredity (ed. 2; New York).