MEIOSIS IN CERTAIN INTERSPECIFIC HYBRIDS 
 
 IN CREPIS AND ITS BEARING ON 
 
 TAXONOMIC RELATIONSHIP 
 
 BY 
 
 E. B. BABCOCK and S. L. EMSWELLER 
 
University or California Publications in Agricultural Sciences 
 
 Volume 6, No. 12, pp. 325-368, plates 14 and 15, 8 figures in text 
 
 Issued November 27, 1936 
 
 Price, 50 cents 
 
 University of California Press 
 Berkeley, California 
 
 Cambridge University Press 
 London, England 
 
 PRINTED IN THE UNITED STATES OF AMERICA 
 
MEIOSIS IN CERTAIN INTERSPECIFIC 
 
 HYBRIDS IN CREPIS AND ITS BEARING ON 
 
 TAXONOMIC RELATIONSHIP 
 
 BY 
 
 E. B. BABCOCK and S. L. EMSWELLER 
 
 CREPIS NICAEENSIS Balb. x C. SETOSA Hall, f . 
 
 Most Cbepis species are excellent subjects for cytological research be- 
 cause of tbeir generally large and morphologically distinct chromo- 
 somes. The genus includes three subgenera, in each of which numerous 
 representative species have been studied with reference to comparative 
 morphology, geographic distribution, and chromosome relations (Bab- 
 cock and Cameron, 1934 ; Babcock, 1936) . Cytogenetic studies on several 
 hybrids between species in the same subgenus and in different subgenera 
 have exhibited phenomena that throw light on problems of taxonomy, 
 phylogeny, and evolution (Babcock and Navashin, 1930). The present 
 investigation was undertaken with the hope of contributing further evi- 
 dence on some of these problems. 
 
 Species and Hybrids Investigated 
 
 These investigations are concerned with hybrids between Crepis nicaeen- 
 sis Balb., of subgenus Eucrepis, and two subspecies of Crepis setosa 
 Hall. f. of subgenus Barkhausia. Some of the F 2 derivatives were also 
 studied. Plants of the two subspecies of C. setosa, namely, typica (ac- 
 cession 2623) and Topaliana (accession 2671) were used as paternal 
 parents, and a single nicaeensis plant (accession 2700) furnished the 
 maternal gametes. The diploid chromosome number of both species is 8. 
 The strain of C. nicaeensis used in this study came from northern 
 Italy. The species occurs sporadically in lower montane regions from 
 the Eastern Pyrenees to Macedonia. Plants found in the wild state are 
 tall, erect, and simple-stemmed with a few branches near the top bear- 
 ing a few medium-sized flower heads. Under cultivation the plants are 
 either annual or biennial ; they retain the same upright habit as in the 
 wild but are very vigorous, with long branches and many heads. As a 
 rule nicaeensis plants have a stronger central axis than do setosa plants 
 and the basal or rosette leaves are very different, although there is no- 
 ticeable variation in both species. In details of the heads, florets, and 
 fruits the differences are even more marked. Differences in the achenes 
 
 [325] 
 
326 University of California Publications in Agricultural Sciences [Vol. 
 
 b 
 
 Fig. 1. Achenes of: a, Crepis setosa typica (2623) ; b, C. setosa Topaliana (2671) ; 
 c, C. nicaeensis (2700) ; d, C. nicaeensis x C. setosa typica Fij e, C. nicaeensis x C. se- 
 tosa Topaliana Fj ; /, an F 2 derivative with 8 chromosomes ; g, h, F 2 derivatives with 
 24 chromosomes ; i, C. biennis. 
 
1936] Babcoch-Emsweller : Meiosis in Interspecific Hybrids in Crepis 327 
 
 are of special interest in this study. In nicaeensis the achenes are golden 
 brown, about 3 mm. long and 0.6 mm. wide, terete, narrowed at base and 
 summit, with no beak or prolongation of the apex, and with ten broad, 
 nearly smooth, longitudinal ribs (fig. lc). 
 
 In C. setosa, considered as a whole, the plants are more slender, the 
 habit more bushy, and the heads more numerous and smaller. Besides 
 other distinguishing characters, the achenes are tawny, 3-5 mm. long 
 and 0.3-0.6 mm. wide, and the body is terete, with ten ribs, narrowed at 
 base and attenuate at summit into a slender beak equal to or shorter than 
 the body (fig. la,b). The strain of C. setosa typica used in these investi- 
 gations was obtained from a wild population in Savoy, France. This sub- 
 species occurs rather commonly at low elevations from the eastern Pyre- 
 nees to Macedonia. Our accession of subsp. Topaliana came from a wild 
 population in eastern Thessaly. This subspecies has a restricted distribu- 
 tion in northern Greece and a few forms which are intermediate between 
 the two subspecies have been found in northwestern Thessaly and 
 Epirus. 
 
 Some of the characters that distinguish the two subspecies are listed 
 in table 1. Although these data were obtained from only two plants of 
 one subspecies and one of the other, yet these plants were fairly typical 
 of their respective entities. Plants of typica are normally larger and 
 more robust than Topaliana plants. Differences in the basal and cauline 
 leaves are shown in figure 2a, b. Certain differences in the achenes are 
 discernible in figure la, b. In the typica strain the beak is equal in length 
 to the body of the achene and all the achenes in a head are closely simi- 
 lar. In Topaliana the marginal achenes are notably different from the 
 rest of the achenes in the head. They are somewhat compressed, paler in 
 color, and merely attenuate, not finely beaked, as is shown by the right- 
 hand achene in figure lb, and the inner achenes have a relatively longer 
 body and shorter beak than do those in the typica strain used. Another 
 difference which, although minute, was consistent in each plant was 
 found in the setae (bristles) borne on the involucral bracts and pedun- 
 cle. In the typica strain these setae are slightly longer and more slender 
 than those of the Topaliana strain. This is especially interesting because 
 of the failure of setae to appear in the F 1 (30.31) of which typica was a 
 parent, whereas setae were abundant in the other F x (30.32) , as is shown 
 in plate Id, e. The first impression given by these facts is that the par- 
 ticular typica plant which was used as male parent of 30.31 must have 
 been heterozygous for setae, whereas the Topaliana plant used in pro- 
 ducing 30.32 was homozygous. It was shown by Collins (1924) that in 
 Crepis capillaris glabrous involucre it inherited as a simple recessive to 
 pubescent involucre. But here the pubescence consists of hairs, not setae. 
 Furthermore, plants with glabrous involucres have never been reported 
 
328 
 
 University of California Publications in Agricultural Sciences [Vol. 6 
 
 in C. setosa; on the contrary, these peculiar setae seem to be a constant 
 specific character in this species. Certainly no plants with glabrous in- 
 volucre were observed among several strains of typica while they were 
 under cultivation. This fact, together with the structural difference in 
 the setae of the two subspecies noted above, may be indicative of a 
 
 TABLE 1 
 
 Differentiating Characters in the Subspecies of Crepis setosa 
 
 2623. Typica 
 Large, robust 
 
 Height (2 plants), 65-75 cm. 
 
 Stem erect with several long branches 
 from base, each with a secondary 
 branch from each node; secondary 
 branches erect, forming a narrow 
 angle with main branch 
 
 Involucre setose, the bristles slightly 
 longer and more slender than in 2671. 
 Bristles of same type also rather plen- 
 tiful on peduncle 
 
 Achenes all similar, or the marginal ones 
 parthenocarpic, but even then usually 
 beaked and colored like the inner ones; 
 3-4 mm. long, deep tawny, the body 
 oblong, subterete, strongly 10-ribbed, 
 beak equal to body, conspicuously ex- 
 panded to form a white inverted cone 
 below the broad pappus disk 
 
 Style branches bright green 
 
 2671. Topaliana 
 Small, slender 
 
 Height (1 plant), 38 cm. 
 
 Stem erect with a branch from each 
 node, but branches shorter than axis 
 and spreading at a wide angle from it 
 
 Involucre setose, the bristles shorter 
 and slightly thicker. Bristles on pe- 
 duncle very few and diminished 
 
 Achenes of two shapes, 3-4 mm. long; 
 the marginal ones very pale or whit- 
 ish on ventral face, long-fusiform, 
 strongly attenuate toward summit or 
 shortly beaked, obscurely ribbed; 
 inner achenes pale tawny, the body 
 fusiform, subterete, delicately 10- 
 ribbed, beak a little shorter than 
 body, not expanded below pappus 
 disk 
 
 Style branches pale greenish yellow 
 
 genetic difference between the subspecies with respect to setae. At- 
 tempts were made to obtain hybrids between these two subspecies of C. 
 setosa, but they failed, so there is no further evidence bearing on this 
 particular question. 
 
 Hybrids between C. nicaeensis and C. setosa were obtained with diffi- 
 culty. Only four F 1 plants were produced, two of each with typica and 
 Topaliana as male parent. For brevity's sake the two crosses will be re- 
 ferred to hereafter as hybrid A and hybrid B. The F x hybrid sibs were 
 very similar, but there were consistent differences between the two F x 
 
1936] BaococTc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 
 
 329 
 
 families. Some of these differences are shown in table 2. The leaf types 
 of the parents and F x hybrids are shown in figure 2. Not much signifi- 
 cance can be attached to these differences because the parents, like those 
 in other species of Crepis, are rather variable in leaf shape ; but, taken 
 along with the other differences in the hybrids, these leaf differences 
 may also indicate genetic diversity between the subspecies. Certain dif- 
 ferences in anthocyanin in the F x plants may have been due to heterozy- 
 gosity in the nicaeensis parent. One plant from each cross was pressed 
 
 TABLE 2 
 Differentiating Characters in Hybrids A and B (2 Plants of Each) 
 
 A. C. nicaeensis 9 X C. setosa typica <? 
 
 Achenes average 4.27 mm. long; ribs as 
 in C. nicaeensis 
 
 Involucres and peduncles glabrous 
 
 Anthocyanin restricted to base of plant 
 
 Basal and cauline leaves laciniate 
 
 B. C. nicaeensis 9 X C. setosa Topaliana cf 
 
 Achenes average 3.47 mm. long; ribs 
 similar but less prominent 
 
 Involucres and peduncles setose 
 
 Anthocyanin present in upper part of 
 plant 
 
 Basal leaves coarsely dentate; cauline 
 leaves nearly entire 
 
 for the herbarium ; the others were used in the present study. The two 
 hybrids differed notably in fertility. Hybrid A produced 545 open- 
 pollinated achenes and hybrid B, only 120. 
 
 From hybrid A one hundred F 2 achenes were sown. They produced 
 ninety-three seedlings. Three of these died at the cotyledon stage and 
 three more after they had developed a small rosette of basal leaves (see 
 plate 2h, i). The chromosomes of the six seedlings that died were not 
 examined. The remaining plants were grown, to the large rosette stage. 
 Chromosome numbers were determined from somatic plates in root tips. 
 Only three had 8 chromosomes ; the remaining eighty-four all had 24 
 chromosomes. 
 
 From hybrid B twenty-five achenes were sown, and nineteen seedlings 
 were obtained. Two of these died in the cotyledon stage. The other seven- 
 teen were grown to maturity ; of these, two had 8 chromosomes and fif- 
 teen had 24 chromosomes. 
 
 The 24-chromosome F 2 derivatives must have resulted from natural 
 crossing with Crepis biennis (n = 20). Plants of this species were in 
 flower along with the F 1 hybrids, which were purposely left unprotected. 
 The leaves of the 24-chromosome derivatives were more or less like those 
 
330 
 
 University of California Publications in Agricultural Sciences [Vol. 6 
 
 of C. biennis. This is not necessarily significant, because the leaves of 
 nicaeensis also resemble those of biennis ; but when these leaves are com- 
 pared with those of their 8-chromosome sibs, the difference is striking 
 
 Fig. 2. Basal and cauline leaves from: a, Crepis setosa typica; o, C. setosa Topa- 
 liana; c, C. nicaeensis; d, C. nicaeensis 5 x C. setosa typica (JxF,; e, C. nicaeensis 
 $ x C. setosa Topaliana rf F,. 
 
 because the leaves of the latter strongly resembled those of setosa (see 
 plate 15/, g), whereas the leaves of the 24-chromosome plants were very 
 like those of biennis (plate 15a-e). Fortunately these 24-chromosome 
 derivatives were slightly fertile, and the achenes were like those of Men- 
 
1936] Bab cocTc-Emsw slier : Meiosis in Interspecific Hybrids in Crepis 331 
 
 nis in size and shape (see fig. 1) and in the higher number of ribs which 
 is characteristic of biennis. 
 
 The five 8-chromosome F 2 derivatives were closely similar in general 
 appearance and very precocious in development as compared with the 
 24-chromosome sibs. Two of them are shown in rosette stage in plate 15/, 
 g. The one on the right is three weeks older than the other. Two of the 
 8-chromosome derivatives died from an unknown cause when the flower 
 stalk was a few inches high, and another was attacked by a species of 
 Botrytis just before flowering. The two plants that produced flowers 
 were from hybrid A. Their fertility was low, one producing twelve 
 achenes and the other, seven. 
 
 This paper is chiefly concerned with a study of meiosis in the ¥ x hy- 
 brids and the two 8-chromosome F 2 derivatives that produced flowers. 
 
 Cytological Methods 
 
 Root tips were fixed in chromacetic formalin and stained with Haiden- 
 hain's iron haematoxylin, as described by Hollingshead and Babcock 
 (1930). At first the buds were fixed in the Carnoy-Navashin solution, 
 dehydrated in the alcohols, cleared in xylol, and imbedded in paraffin. 
 Buds treated in this manner were usually cut with difficulty because of 
 hardening in the higher alcohols. To remedy this situation, a butyl alco- 
 hol series was substituted as a clearing agent in place of the regular 
 xylol series. In this method the buds were run up to 80 per cent alcohol, 
 then through the following series of alcohols (Zirkle, 1931) : 
 
 (a) Water 15 ec, absolute ethyl alcohol 50 ee., butyl alcohol 35 cc, 2 to 3 hours. 
 
 (b) Water 5 ec, absolute ethyl alcohol 40 cc, butyl alcohol 55 cc, 2 to 3 hours. 
 
 (c) Absolute ethyl alcohol 25 ec, butyl alcohol 75 cc, 2 to 3 hours. 
 
 (d) Butyl alcohol. At least 2 changes, leaving in the 2d change at least 12 hours. 
 
 Infiltration was accomplished by pouring melted paraffin into a vial, 
 allowing it to cool, and then pouring over it butyl alcohol containing the 
 buds. As few as four changes in the oven over a period of two days were 
 found to be sufficient to remove all traces of butyl alcohol; the buds 
 treated in this manner were soft, easily cut, and stained readily in both 
 gentian violet and Haidenhain's iron-alum haemotoxylin. 
 
 In order to concentrate more material on a slide, imbedded buds were 
 remelted in the oven, and from two to four were mounted on one block. 
 Melted paraffin was slowly dropped over the buds and cooled rapidly, 
 and the block was finally trimmed for cutting. Both cross and longi- 
 tudinal sections were cut. The latter proved to be most useful, because a 
 large number of PMC's could be studied in the same locule, and because 
 irregular stages found in F 2 plants were more easily studied when ad- 
 
332 
 
 University of California Publications in Agricultural Sciences [Vou 6 
 
 jacent to "normal" cells. All slides from each block were kept separate, 
 and a slide containing material from near the center of the buds was 
 stained to determine whether the desired meiotic figures were present ; 
 if so, the remaining slides of the series were stained. 
 
 /y. 
 
 •5, 
 
 Sa 
 
 M/ft 
 
 
 
 Fig. 3. Chromosomes at mitotic metaphase in root-tip cells of: a, Crepis setosa; 
 b, C. nicaeensis; c, C. nicaeensisxC. setosa F,; d, an 8-chromosome F, derivative, 
 x 2500. 
 
 Figure 3 and figure 4c were drawn at a magnification of 2500 diam- 
 eters and were reproduced without reduction. Figures 4a and 4& were 
 drawn at a magnification of 4000 diameters and reproduced without re- 
 duction. Figures 5e and 6c were drawn at a magnification of 1800 diam- 
 eters. All other drawings were made at a magnification of 2500 diameters 
 and were reduced one-fourth in reproduction. 
 
1936] BabcocTc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 333 
 
 Somatic Chromosomes 
 
 Somatic metaphase chromosomes of Crepis setosa are shown in figure 
 3a; those of C. nicaeensis in figure 3b. As the chromosomes of the two se- 
 tosa subspecies were apparently identical, only one plate is shown. Ac- 
 cording to Navashin's scheme of chromosome designation, both species 
 have one pair each of types A, B, C, and D. These designations are not 
 
 Sefosa 
 
 f 
 
 /V/'caeens/s 
 
 )" 
 
 Ae Ph/?f 
 
 3d 
 
 S* 
 
 My 
 
 m 
 
 Fig. 4. a, o, Haploid chromosomes of Crepis setosa and C. nicaeensis x 4000. c, chro- 
 mosomes at mitotic metaphase in a root-tip cell of an 8-chromosome F 2 plant showing 
 three doubtful members, X,, X 2 , X s . x 2500. 
 
334 University of California Publications in Agricultural Sciences [Vol. 6 
 
 intended to imply homology but merely general similarity in size and 
 shape. There is difference enough in chromosome morphology to make 
 identification of each set in F x a relatively simple matter (fig. 3c) . Nava- 
 shin (1928) reported loss of the satellite of the tectorum D chromosome 
 in the F x capillaris-tectorum hybrid. This observation was confirmed by 
 Hollingshead in the same hybrid (1930) . In the F x nicaeensis-setosa hy- 
 brid the nicaeensis D chromosome is similarly modified and careful 
 measurements indicated an increase in size of the proximal arm. This 
 change makes it more difficult to distinguish between the B chromosome 
 of setosa and the modified D chromosome of nicaeensis. The chromosomes 
 of the five F, plants discussed in this paper are shown in figure 3d. In 
 order to check the camera-lueida drawings, photomicrographs were made 
 from normal somatic plates (plate 14/). The chromosomes appear to be 
 as follows : from setosa, one A, two B, one C, and two D ; and from ni- 
 caeensis, two modified D chromosomes. This genom, however, is outside 
 of normal expectations and studies of meiotic behavior in both F t and 
 F, indicate that one of the two chromosomes, which is morphologically 
 similar to the nicaeensis D chromosome, is perhaps a modified setosa B 
 chromosome. The evidence favorable to this hypothesis will be discussed 
 under meiosis. 
 
 It is unusual that all five of these plants should have the same chromo- 
 some complement, and that no other combination of eight chromosomes 
 was found. The eight F, plants that died in the seedling and very early 
 rosette stages were most likely 8-chromosome types that were unable to 
 survive. When the first lot of F 2 seed was planted, some care was exer- 
 cised to eliminate the shrunken achenes. When a second planting was 
 made, the achenes were carefully graded into large and small lots. There 
 was a mortality of 40 per cent in the seedlings from the small achenes 
 and of only 2 per cent in seedlings from large achenes. 
 
 Meiosis in F x 
 
 Hybrid A (C. nicaeensis § x C. setosa typica <$) 
 
 This hybrid exhibited rather high regularity in its meiotic behavior. In 
 figure 5 the drawings are from several slides, all stained with Haiden- 
 hain's, with the exception of b, d, and h which were stained in gentian 
 violet. The number of bivalents ranged from one to four. Figure 5a-d 
 shows associations of 4„, 3„ + 2,, 2„ + 4i, and In + 6i, respectively. The 
 chromosomes in b and c are drawn out of position for the sake of clarity. 
 Figure 5e shows four adjacent PMC's in the same locule, all exhibiting 
 complete pairing. In three a small chromosome is loosely paired with a 
 large one. The small chromosome is undoubtedly the setosa C chromo- 
 some. Figure 5/ shows 3 n + % at IM. The univalents were apparently 
 
1936] Babcoclc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 335 
 
 (O/O r©^ ^v 
 
 Fig. 5. Meiosis in hybrid A (C. nicaeensis x C. setosa typica). a-d, selected diaki- 
 nesis figures showing variable pairing (in ft and c certain chromosomes have been 
 drawn outside the nucleus for clarity), e, adjacent PMC's showing complete pairing. 
 /, IM — 3n +• 2i. g, IA. h, IIM. i, IlA. j, k, m, n, types of irregularities. I, "normal" 
 tetrad, x 1875, except e, 1575. 
 
336 University of California Publications in Agricultural Sciences [Vol. 6 
 
 not arranged on the spindle but were placed to the side of it. Their size 
 indicates that neither is the setosa C chromosome. In ten cells showing 
 3n + 2 I( the C chromosome of setosa was one of the two unpaired chromo- 
 somes in only four cells. 
 
 Figure 5g shows a I A in which the setosa C chromosome is easily rec- 
 ognized. The chromosomes have not yet assumed their characteristic 
 shapes, as is also true in the IIM shown at h, which may have followed 
 such an anaphase. Figure 5i is an excellent IIA, showing the chromo- 
 somes in a nearly normal somatic condition. In the group to the right, 
 the D chromosome of setosa can be recognized as well as its small C 
 chromosome in the group at the left. Such a IIA would probably give 
 rise to a tetrad similar to figure 51. Very few irregularities were observed. 
 Figure 5j shows a IT in which a portion of the chromatin has failed to 
 reach either pole and will probably be left in the cytoplasm as a micro- 
 nucleus. In figure 5k unequal nuclei are seen, with a chromatic body in 
 the cytoplasm at the right. There was no trace of a membrane about this 
 chromatic body. Possibly figure 5m represents a triad with a diploid and 
 two haploid nuclei. It was the only example of this type found in the 
 material examined. A type of tetrad irregularity that occurred very 
 rarely is shown in figure 5n. There has been an unequal division of chro- 
 matin in the IIM of one nucleus of the dyad, whereas the other nucleus 
 appears to have divided more regularly. 
 
 The frequencies of pairing in the two F 1 plants and in the F 2 plants 
 are shown in table 3. 
 
 The large number of instances in which four bivalents occur is rather 
 remarkable for Crepis hybrids involving different subgenera. The pair- 
 ing is also rather close, and there is apparently a very intimate associa- 
 tion of the chromosomes of most bivalents. The small C chromosome of 
 setosa can frequently be identified, and it is usually very loosely paired 
 with a nicacensis chromosome. The generally close association of these 
 chromosomes from widely separated species is quite different from the 
 looseness of pairing reported in most of the other interspecific hybrids 
 of Crepis. Avery (1930) published meiotic figures of hybrids involving 
 Crepis leontodontoides with five other species — tectorum, capillaris, par- 
 viflora, Marschallii, and aurea. In many cells the bivalents are composed 
 of rather loosely associated chromosomes which typically form elon- 
 gated bivalents. The same situation also occurs generally in the hybrids 
 studied by Babeock and J. Clausen (1929), who also published meiotic 
 figures of Crepis aspera and C. oursifolia. The bivalents formed in these 
 species show an association more nearly resembling the bivalents in our 
 hybrid. Such close association in this hybrid indicates at least that con- 
 ditions were rather favorable for exchange of material between chromo- 
 somes of the two species. This supposition is strengthened by the investi- 
 
1936] Bdbcoch-Emsweller . Meiosis in Interspecific Hybrids in Crepis 
 
 337 
 
 gation of meiosis in the hybrid Crepis divaricata x C. Dioscoridis by 
 Miintzing (1934) . Although the amount of pairing was noticeably lower 
 than in our hybrids, the nature of the association was similar. Cross- 
 shaped bivalents sometimes occurred, which would indicate interstitial 
 chiasmata, and "the frequent occurrence of chromatin bridges and frag- 
 ments at IA is too regular to be due simply to non-homologous associa- 
 tion." 
 
 TABLE 3 
 Bivalent Formation and Achene Production in Fi and F 2 
 
 
 4n 
 
 3ii2i 
 
 2n4i 
 
 1 1 1 6 1 
 
 0n8i 
 
 Per cent 
 irregu- 
 larities 
 
 Per cent 
 "good" 
 pollen 
 
 Number 
 
 of 
 achenes 
 
 Hybrid A. 
 
 122 
 
 10 
 
 1 
 
 1 
 
 
 
 9 
 
 92 
 
 545 
 
 Hybrid B. 
 
 64 
 
 32 
 
 16 
 
 2 
 
 
 
 43 
 
 58 
 
 120 
 
 F 2 no. 1 . . . 
 
 22 
 
 44 
 
 2 
 
 6 
 
 1 
 
 71 
 
 48 
 
 7 
 
 F 2 no.2.. . 
 
 31 
 
 72 
 
 7 
 
 12 
 
 2 
 
 75 
 
 42 
 
 12 
 
 Hybrid B (C. nicaeensis JxC. setosa Topaliana J 1 ) 
 
 In this hybrid the meiotic behavior of the chromosomes is considerably 
 more irregular than in hybrid A. As shown in table 3, these irregulari- 
 ties comprise 43 per cent of the total, as compared with about 9 per cent 
 for the same classes in hybrid A. In hybrid B, as well as in hybrid A, no 
 cells were found with a complete failure of bivalent formation. 
 
 A diakinesis figure with 4n is shown in figure 6a. In figure 6b there 
 are 3 n + 2 r ; and the smallest univalent is very probably the C chromo- 
 some of setosa. Figure 6c represents three adjacent pollen mother cells 
 with 4n, 3n + 2i, and 2 n + 4r and shows the occurrence of loosely paired 
 bivalents and the close association of variable pairing in the same locule. 
 This figure may be compared with figure 5c, to show the looser type of 
 pairing and relatively less frequent formation of 4n in this hybrid than 
 in hybrid A. In figure 6d the association of the small setosa C chromo- 
 some with a large nicaeensis chromosome forms a bivalent somewhat 
 similar to one formed in hybrid A. One of the two cells observed, in 
 which but one bivalent was formed, is shown in figure 6e in which the 
 chromosomes were very loosely arranged near the equatorial plate. 
 Figure 6/, g represent adjacent PMC's, one of which shows 2 n + 4 r and 
 the other 3n + 2 T . In the second cell, two of the bivalents are very loosely 
 associated, in striking contrast to the same situation in hybrid A (fig. 
 5/) . In figure 6h one chromosome is seen lagging at the equator, four are 
 moving to one pole, and three are going to the other. Three of the dis- 
 joined chromosomes already show equational splitting for the homoeo- 
 typic division. In figure 6i one chromosome is at one pole, five are at the 
 
338 University of California Publications in Agricultural Sciences [Vol. 6 
 
 Fig. 6. Meiosis in hybrid B (C. nicaeensisxC. setosa Topaliana). a, b, diakinesis. 
 c, adjacent PMC's showing variable pairing, d-g, metaphase plates showing rather 
 loosely formed bivalents. h, i, IA showing unequal distribution of chromosomes. /, 
 normal tetrad, k, dyad, l-o, variations found in same locules with "normal" types, 
 x 1875, except c, 1575. 
 
1936] Babcock-Emsweller : Meiosis in Interspecific Hybrids in Crepis 339 
 
 other pole, and two are lagging at the equator but appear to have begun 
 migrating toward the pole with the larger chromosome number. An ap- 
 parently normal tetrad is shown in figure 6j. Figure 6k represents a 
 dyad in which the nuclei appear to be equal in size. Very few irregulari- 
 ties of this sort were seen in this plant, and where they did occur they 
 were in locules in which tetrads occurred. Figure 61 shows a dyad in 
 which the nuclei are unequal in size. This condition could have followed 
 an unequal distribution of chromosomes like that shown in h. A normal 
 monad was not found in this plant, but several similar to figure 6m 
 were seen. Most of the tetrads were abnormal, micronuclei being very 
 common. 
 
 Meiosis in F 
 
 The various stages in meiosis of the F 2 plants with 8 chromosomes are 
 shown in figure 7. These drawings were made from smear preparations 
 from the two plants which flowered. Diakinesis figures with 4n and 
 3n + 2i are shown in figure la, b. Some IM figures are shown in figure 
 7c-j depicting 4n, 3 n + 2 r , l n + 6i, and 8i. In figure li the C chromosome 
 of setosa is a member of the only bivalent formed. Two IA groups are 
 seen in figure Ik, I. In figure 11 some of the separated chromosomes are 
 splitting for the homoeotypic division. An irregular assortment of 5 and 
 3 chromosomes is seen in figure 7m. A division of this type could give 
 rise to very small pollen grains like those shown in figure 8i. The chro- 
 mosome group in figure In is interesting because it suggests a possible 
 method of origin for polyploid gametes. The seven units at one pole may 
 have occurred by nonconjunction in three pairs of chromosomes. This 
 situation was observed but once and cannot account for the high fre- 
 quency of large pollen found in this plant. Many of the tetrads were nor- 
 mal (fig. lo), but usually irregular types were found (fig. Ip and fig. 
 8f,g,h). 
 
 The frequency of the irregularities in meiosis in these two plants is in 
 rather close agreement, as is shown in table 3. It is very unusual that in 
 so many cells only one or two bivalents are formed. Unfortunately meio- 
 sis in the two subspecies has not been studied, so it is not known whether 
 either of them shows an appreciable amount of nonconjunction. 
 
 Chromosome morphology would seem to indicate that these F, plants 
 had six setosa chromosomes, one A, two B, one C, and two D, along with 
 modified D chromosomes from nicaeensis. This complement, however, 
 could be produced only if the nicaeensis D chromosome had paired part 
 of the time with setosa C and part of the time with setosa A. This be- 
 havior would necessitate that the nicaeensis D chromosome have sectors, 
 one homologous to a sector in the setosa C chromosome and another to a 
 sector in the setosa A chromosome. If this were true, we should occasion- 
 
340 University of California "Publications in Agricultural Sciences [Vol. 6 
 
 Fig. 7. Meiosis in 8-chromosonie F 2 plants, a, b, diakineais showing 4 n and 3n + 2 t ; 
 in b note the small unpaired chromosome (see text), c-j, IM showing variation in 
 number of bivalents. k-m, IA (note small C chromosome of setosa in I and unequal 
 distribution of chromosomes in m). n, seven chromatin bodies at one pole; possibly a 
 result of attachment, o, p, "normal" and abnormal tetrads, x 1875. 
 
1936] Bdbcoclc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 341 
 
 ally find trivalents formed in these plants. As trivalents are not found 
 this view is apparently not tenable. 
 
 The interpretation which seems most logical is that one of the two 
 chromosomes which appear like nicaeensis D chromosomes is in reality 
 a new chromosome resulting from crossing over in F 1 . Although five F 2 
 plants had eight chromosomes that were apparently of the types de- 
 scribed above, only two flowered normally and both were descendants 
 of hybrid A. The pairing in this F 1 hybrid was exceedingly regular, and 
 bivalent associations were in most cells very intimate. Hence exchange 
 of material probably took place in this hybrid. If pairing occurred be- 
 tween setosa A and nicaeensis C, and crossing over took place at or near 
 the constriction of setosa A, a new chromosome would be produced, the 
 major portion of which would be setosa A, and the presence of the prox- 
 imal arm of the nicaeensis C chromosome would produce an element 
 resembling a nicaeensis D chromosome. On this basis the complement of 
 these F 2 plants may consist of the following setosa chromosomes : one A, 
 two B, one C, two D, and one modified A, together with one modified 
 nicaeensis D chromosome. The haploid sets of setosa and nicaeensis, to- 
 gether with an F 2 somatic plate, are shown in figure 4. In the F 2 plant 
 the chromosomes in question are marked X 1; X 2 , and X 3 . 
 
 The chromosomes of these F 2 plants may perhaps represent recombi- 
 nation products from crossing over in F r . Although the chromosomes of 
 each species are distinctively different in morphology, it would become 
 exceedingly difficult to identify many possible new recombinations. 
 Plants with other assortments of 8 chromosomes were unfortunately not 
 found ; they might have thrown considerable light on the question at 
 issue. Navashin (1926) reported the appearance in Crepis of large atyp- 
 ical chromosomes, a phenomenon to which he gave the name "novation." 
 Their occurrence was thought to result not from end-to-end fusion, but 
 perhaps from translocation. Avery (1930) suggested that the variable 
 pairing observed in the Crepis hybrids which she investigated "is a re- 
 flection of the transformational processes presumably responsible for 
 the differentiation of the specific genoms." In a recent paper Navashin 
 (1934) reports several occurrences of "sporadic amphiplasty" in which 
 morphologically changed chromosomes were observed in the progeny of 
 species hybrids. Poole (1932) has found cytological evidence of new 
 chromosome types in derivatives of his rubra-foetida amphidiploid. In 
 the hybrid, Crepis divaricataxC. Dioscoridis, Miintzing (1934) ob- 
 served the occurrence at IA of chromatin bridges and fragments and 
 concluded that the simultaneous production of long chromatids with 
 two spindle-fiber attachments and fragments without attachment is 
 probably caused by crossing over between homologous segments in chro- 
 mosomes from the two species. He emphasizes the point that "an unlim- 
 
342 
 
 University of California Publications in Agricultural Sciences [Vol. 6 
 
 ited number of new chromosome types may arise by the mechanism at 
 work in this hybrid (or a number only limited by the number of chro- 
 momeres)." 
 
 Origin of Polyploid Gametes 
 
 When tetrad stages and pollen were studied in these F, plants, it was 
 found that a comparatively large number of monads and dyads were 
 formed and that the pollen grains could be sorted into four rather defi- 
 nite size groups, as shown in figure 8j-l. The group listed as "small" 
 (table 4) and shown in figure 8j is probably of the "normal" size. On the 
 basis of measurements of pollen grains in this group these grains were 
 equivalent to "good" pollen in the parent species. 
 
 TABLE 4 
 Size of Pollen in F 2 
 
 Giant 
 
 Large 
 
 Small 
 
 Very small 
 
 35 
 
 78 
 
 115 
 
 16 
 
 In meiosis in F 1} and normally in F 2 , whole chromosomes that were 
 paired at IM separated and moved to the poles ; and univalents were as- 
 sorted at random without splitting. In the F, plants, IA figures were 
 finally found in which splitting of both disjoined and univalent chro- 
 mosomes occurred. Figures 8a and 8b indicate that the splitting has 
 taken place slightly earlier in the disjoined chromosomes. In both chro- 
 mosomes there is no doubt that the split had been completed in early 
 anaphase and that the resulting nuclei would have had 8 chromosomes 
 each. This type of behavior will probably result in a binucleate condi- 
 tion like that shown in figure 8/. The complexity of the chromosome 
 situation in this F„ plant makes it extremely probable that the resulting 
 nuclei are very rarely alike qualitatively, even though each has 8 chro- 
 mosomes. As the equational split has already occurred, it is unlikely 
 that a second division of the chromosomes will occur ; and various types 
 of "diploid" gametes should result. Figure 8k is probably a "diploid" 
 pollen grain resulting from such a division. 
 
 A second irregularity in meiosis was found which probably gives rise 
 directly to the giant pollen grains that occur in these F, plants. Here, 
 spindle formation at IM is very weak or entirely lacking. The chromo- 
 somes arrange themselves on the equatorial plate and some bivalents are 
 probably formed, although this point is not certain and is not essential 
 to what follows. Each of the 8 chromosomes splits equationally, but 
 there is no separation of the halves ; a nucleus of 16 chromosomes is the 
 result. The 16 chromosomes can be counted readily in figure 8c. Various 
 
1936] BabcocTc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 343 
 
 Fig. 8. Origin of polyploid gametes in an 8-chromosome F 2 plant, a, b, first ana- 
 phase showing precocious splitting of chromosomes which will probably result in 
 "diploid" nuclei, c-e, origin of 16-chromosome gametes through splitting of the 8 
 univalents in equatorial region with no succeeding division. /, g, h, "dyad," monad, 
 and "triad." ir-l, pollen grains of four size classes, all x 1875. 
 
344 University of California Publications in Agricultural Sciences [Vol. 6 
 
 stages of this type of division are shown in figure 8c, d, e. This condition 
 would result directly in a monad (fig. 8g) which would develop into a 
 giant pollen grain (fig. 81). 
 
 Investigators working with widely scattered genera have reported the 
 occurrence of giant pollen grains in many plants. De Mol (1923) found 
 in Hyacinth, which had been subjected to abnormal conditions, large 
 pollen grains in the nucleus of which was the diploid number of chromo- 
 somes. He interpreted their origin as due to an ineffective karyokinesis 
 in the microspore rather than to any irregularities in the first or second 
 division of the pollen mother cell. 
 
 Sakamura and Stow (1926) observed marked variation in the size of 
 pollen grains in Gagea lutea from plants exposed to high temperatures. 
 Cytological studies revealed several irregularities which they thought 
 caused the abnormal pollen grains. In the first division they occasionally 
 found an irregular distribution of chromosomes that resulted in pollen 
 grains with varying chromosome numbers. Before the second division 
 in some cells they observed that all chromosomes of the primary sporo- 
 cyte gathered at the center of the cell and reconstructed a new nucleus 
 without a cell division. They suggested that nuclei of this type probably 
 gave rise to diploid pollen grains if a regular second division took place, 
 or that, if this division were omitted, giant tetrad pollen grains would 
 be formed. Second divisions were not observed, 
 
 Rosenberg (1926-27) investigated meiosis in a number of partheno- 
 genetic species of Hieracium. He found a graded series of degeneration 
 of meiosis ranging from the boreale type, in which a variable number of 
 doubles and singles occur, to the laevigatum type, in which there is no 
 conjugation whatever, namely, "semiheterotypic division." The latter 
 type is often interrupted by a premature homoeotypic division whereby 
 a "restitution nucleus" is formed. A new nuclear wall encloses the entire 
 spindle figure and the chromosomes of this single large nucleus divide 
 normally, producing pollen grains with the diploid number of chromo- 
 somes. 
 
 Karpechenko (1927), in tracing the origin of polyploid F 2 plants 
 from his generic hybrid BrassicaxRaphanus, found diploid pollen 
 grains formed in the F 1 in the following manner. At IM pairing between 
 the chromosomes of each haploid set did not occur and spindle forma- 
 tion was very weak with no distribution of univalents. The complete 
 omission of the heterotypic division was followed by a normal homoeo- 
 typic division, each chromosome splitting equationally and separating 
 to the poles. This resulted in diploid dyads. A second splitting of some 
 univalents was occasionally observed. Tetraploid gametes were found to 
 originate as follows : "In the cells of the archesporium, at the division 
 preceding reduction division, a division of the nucleus sometimes occurs 
 
1936] Babcock-Emsweller : Meiosis in Interspecific Hybrids in Crepis 345 
 
 without formation of a new cell wall. As a result there are produced 
 pollen mother cells with two nuclei, each of them containing 18 chro- 
 mosomes. The chromosomes in both nuclei remain nonconjugating ; at 
 the first division both spindles fuse so that the anaphase shows all the 
 univalents which later on form one nucleus containing all the chromo- 
 somes. A splitting of all the chromosomes then ensues, and in the ma- 
 jority of cases dyads arise with about 36 chromosomes in each cell." 
 
 The two F 2 plants in which the polyploid gametes were observed were 
 grown under separate environmental conditions. After the chromosome 
 complements were determined, one plant was removed from Davis to 
 Palo Alto and isolated. The second plant was isolated at Davis. Although 
 Palo Alto and Davis are less than one hundred miles apart, there is 
 quite a difference in the average temperatures of the two places. At 
 Davis the temperature frequently reaches 100° F, while at Palo Alto it 
 is usually not more than 80° to 85° F. The data on metaphase pairing 
 in these two plants are given in table 3 ; number 1 was at Palo Alto, and 
 number 2 at Davis. There is very little difference in the percentage of 
 each type of pairing observed. This fact suggests that the occurrence of 
 meiotic irregularities in these similar F 2 plants was not affected by such 
 differences in temperature as were involved. It seems probable, there- 
 fore, that this temperature range (80°-100° F) may be included within 
 a certain optimum range for meiotic regularity in these hybrids. 
 
 Later, the twelve achenes from F 2 plant number 2 were planted, and 
 eleven seedlings were produced. At the same time, larger numbers of 
 other Crepis seeds were planted, and all populations were transplanted 
 into 4-inch pots in the same type of soil. The eleven seedlings from plant 
 number 2 grew very rapidly and soon produced fine rosettes. They did 
 not produce many roots, however, and finally the entire population be- 
 came chlorotic and died. An examination of the plants a few days before 
 death showed a very meager root system as compared to top growth. The 
 other Crepis plants on the same bench grew vigorously. Only a few root 
 tips were obtained from three of these F 3 plants and they were in poor 
 condition. They showed that these three plants had 8 chromosomes ; but 
 the figures were so poor that identification was impossible. Definite con- 
 clusions concerning viability of the polyploid pollen cannot be drawn 
 from such meager data, but Matsuda (1927) has shown that large pollen 
 grains of Petunia are much less viable and slower to germinate than are 
 normal ones. 
 
 Intraspecipic Relations and Meiotic Behavior 
 
 The differences in amount of pairing at metaphase in hybrid A and hy- 
 brid B (table 3) show that the chromosomes of C. setosa Topaliana fail 
 to form bivalents at metaphase with the nicaeensis chromosomes as 
 
346 University of California Publications in Agricultural Sciences [Vol. 6 
 
 regularly as do the chromosomes of C. setosa typica. Also, when bival- 
 ents occur, the association of the paired chromosomes is generally less 
 intimate in hybrid B than in hybrid A. That is, there are more bivalents 
 in hybrid B showing merely terminal association. Such differences in 
 meiotic behavior, however, do not necessarily indicate wide genie diver- 
 sity between these two forms of C. setosa, because comparable differences 
 have been found among closely similar individuals of a single species 
 and in hybrids involving less diverse forms of a species. 
 
 Plants of a certain race of C. capillaris known as the X strain are mor- 
 phologically typical of the species and closely similar to one another 
 except for minor variations. Hollingshead (1930) found univalents at 
 metaphase in from 12 to 44 per cent of PMC's in different plants of this 
 strain growing under the same favorable conditions. More recently it 
 has been discovered (Richardson, 1935) that, in progeny from a diploid 
 branch of a haploid individual of this same strain of C. capillaris, the 
 occurrence of univalents at IM is caused by failure of chiasma forma- 
 tion between homologous chromosomes which are found to be regularly 
 associated at pachytene. She reports that such univalent homologous 
 chromosomes frequently lie in juxtaposition at diakinesis and this is 
 thought to be a result of their earlier association. Whatever the cause of 
 failure in chiasma formation may be, the fact remains that marked 
 irregularities in metaphase pairing may occur in similar individuals of 
 the same species. It has been found by Beadle (1930) and others in 
 several species of plants that such differences between individuals may 
 be caused by a single gene difference. 
 
 Furthermore, Hollingshead (1930) studied the occurrence of irregu- 
 larities in metaphase pairing in hybrids between C. capillaris (n = 3) 
 and C. tectorum (n = 4) . In these experiments the X strain of C. capil- 
 laris was crossed with two strains of C. tectorum (1498 and 1648) . These 
 two strains of tectorum were similar in most respects, but it was noted 
 by Hollingshead that the 1498 strain was more uniform and had de- 
 scended from one plant, whereas the 1648 strain was grown from wild 
 seed and varied in leaf shape and time of maturity ; also that 1498 came 
 from the Copenhagen Botanic Garden, but that 1648 was collected near 
 Tomsk, Siberia. Although the original source of the Copenhagen strain 
 is not known, it is probable that the two strains represent widely sep- 
 arated geographic races. Furthermore, the achenes of these two strains 
 are significantly different in width and possibly in length. At any rate, 
 the Siberian strain has relatively narrow achenes. Such a difference is 
 sufficient to warrant the recognition of these strains as distinct forms or 
 perhaps as varieties, but they are certainly not sufficiently distinct in 
 morphology to admit of their being recognized as subspecies. Yet the 
 difference in meiotic irregularities in the hybrids between these two 
 
1936] BabcocTc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 
 
 347 
 
 strains of tectorum and the same strain of capillaris are even greater 
 than the corresponding differences between the hybrids involving the 
 two subspecies of setosa. This is shown in table 5, in which the original 
 data are made more easily comparable by stating the percentages of 
 PMC's showing the various numbers of pairs at metaphase. From these 
 data it appears that hybrids 1A and IB show less regular pairing at 
 metaphase than do hybrids 2 A and 2B. But this may be entirely due to 
 the known tendency to meiotic irregularity in the capillaris X strain and 
 not to a difference in the two tectorum strains. This may also account 
 for the apparent tendency to greater difference in amount of meiotic 
 
 TABLE 5 
 
 Comparison of Meiotic Irregularities in Capillaris-Tectorum Hybrids and in 
 Nicaeensis-Setosa Hybrids 
 
 Hybrids 
 
 Per cent of PMC's with 
 
 
 4n 
 
 3n 
 
 2n 
 
 In 
 
 On 
 
 1A. capillaris X tectorum 1498 
 
 
 55 
 
 36 
 
 8 
 
 1 
 
 
 
 IB. capillaris X tectorum 1648 
 
 
 18 
 
 30 
 
 25 
 
 27 
 
 
 
 2A. nicaeensis X setosa Topaliana 
 
 91 
 
 7 
 
 1 
 
 1 
 
 
 
 
 
 2B. nicaeensis X setosa typica 
 
 57 
 
 27 
 
 14 
 
 2 
 
 
 
 
 
 irregularity between 1A and IB than between 2A and 2B. From these 
 data, therefore, no conclusion can be drawn concerning the bearing of 
 the amount of pairing at metaphase on the degree of genetic relationship 
 between two subspecific entities. 
 
 Other differences were found, however, between the two subspecies of 
 C. setosa, besides the difference in behavior of their chromosomes when 
 associated with nicaeensis chromosomes in hybrids. The morphological 
 differences already noted (table 1) justify their rank as subspecies, and 
 the behavior of their chromosomes in the hybrids discussed above is con- 
 sistent with this classification. Babcock and J. Clausen (1929) have 
 shown that the closer the phylogenetic relationship between the species 
 used in a cross, the higher the percentage of good pollen grains and the 
 greater the fertility of the hybrid. The pollen counts of the hybrids, 
 given in table 3, show a very striking difference. The percentage of meio- 
 tic irregularities observed in hybrid A, as compared with those in hybrid 
 B, is of nearly the same ratio as the numbers of seed produced by each 
 hybrid. Although their chromosomes correspond closely in size and 
 shape, these subspecies of setosa may have had their chromosomal or- 
 
348 University of California Publications in Agricultural Sciences [Vol. 6 
 
 ganization greatly altered. The fact that attempts to cross them were 
 unsuccessful might indicate such a situation, but such negative results 
 are of slight value in judging of relationships. The use of other forms of 
 the same subspecies might possibly result in hybrids. If they do seldom 
 cross in the regions where their ranges overlap, this would be an impor- 
 tant factor in the divergence of the two entities. The point of impor- 
 tance to taxonomy is the fact that two entities exist which appear to 
 have diverged from one species and which differ in several morpholog- 
 ical details, but which are still so similar that practical considerations 
 oppose their being recognized as distinct species. "When such a situation 
 is found, together with intergrading forms in the region of overlapping, 
 the most satisfactory systematic disposition of these forms is their treat- 
 ment as subspecies. 
 
A REVIEW OF THE EVIDENCE ON MEIOSIS IN 
 
 INTERSPECIFIC HYBRIDS OF CREPIS WITH 
 
 REFERENCE TO INTRAGENERIC 
 
 RELATIONS 
 
 The three major subdivisions of the genus, namely, Catonia, Eucrepis, 
 and Barkhausia, may be considered as natural in the sense that, as a 
 general rule, the species within each subgenus are more closely related 
 to each other than they are to species in the other two subgenera. But, as 
 has already been pointed out (Babcock, 1936), the very close relation- 
 ship between the subgenera, as indicated by comparative morphology, 
 makes it difficult to separate them by hard and fast lines. Thus certain 
 species of Eucrepis resemble Catonia in their relatively unspecialized 
 involucre, while several others approach Barkhausia in the structure of 
 their fruits. Also, at least one species of Eucrepis shows certain affinity 
 with species in both the other subgenera. This particular species, C. leon- 
 todontoides, was involved in the five interspecific hybrids studied by 
 Avery (1930). Its relationships to the other species involved in these 
 hybrids are considered below. The point to be emphasized here is that 
 C. leontodontoides is a bridging species between Catonia, the most primi- 
 tive, and Barkhausia, the most advanced, of the three subgenera. Con- 
 cerning the genus as a whole, the evidence from chromosome number 
 and morphology (Babcock and Cameron, 1934) and from geographic 
 distribution (Babcock, 1936) also indicates the close interrelationship 
 between the three subgenera and suggests a monophyletic origin for the 
 genus, or at least that it is a natural group which arose from a common 
 progenial stock. 
 
 The study of interspecific hybrids in Crepis has provided additional 
 evidence bearing on phyletic relations within the genus. In the early 
 work of Collins and Mann (1923) the first occurrence of autosyndesis 
 was discovered in the hybrid C. setosa x C. biennis. As the 20 chromo- 
 somes of biennis formed 10 pairs at IM the 4 setosa chromosomes were 
 distributed at random. Thus metaphase pairing furnished only negative 
 evidence on the relationship between these species. Subsequent work on 
 this hybrid led to the discovery that C. biennis is an octoploid species 
 with a base number of 5, which was later confirmed by direct study of 
 the somatic chromosomes (Babcock and Swezy, 1935). In another hy- 
 brid, C. capillaris x C. setosa (Collins and Mann, 1923), it was first re- 
 ported that no pairing of the chromosomes was observed and that the 
 fertility was extremely low. But it was reported later by Hollingshead 
 (1930 ; based on unpublished data of Mann Lesley on a hybrid involving 
 different strains of capillaris and setosa) that variable pairing was ob- 
 served. It is possible that a small amount of metaphase pairing occurred 
 
 [349] 
 
350 University of California Publications in Agricultural Sciences [Vol. 6 
 
 in the original hybrid of Collins and Mann. Yet another hybrid, C. capil- 
 lars x C. aspera, has been reported (Navashin, 1927) as very irregular 
 in meiosis, with random distribution of unsplit or split univalents ; but 
 possibly a small amount of pairing also occurs in this hybrid. The occur- 
 rence of pairing in all other interspecific Crepis hybrids examined cer- 
 tainly warrants the recognition of this possibility. 
 
 Most of the other studies of meiosis in Crepis hybrids were made 
 before the relations between pachytene chiasmata and pairing at meta- 
 phase had been worked out (Darlington, 1932) and the few attempts to 
 study the premeiotic prophases in Crepis hybrids had proved unsuccess- 
 ful. The recent contributions of Richardson (1935a, b) on meiosis in 
 Crepis have shown that the striking individuality of the chromosomes 
 and their meiotic behavior in species with low numbers offsets the tech- 
 nical difficulties involved in the small size of the florets and the paucity 
 of pollen mother cells. Roller (1935) investigated meiosis in Crepis 
 aurea (to = 5) of subgenus Catonia and C. rubra (to = 5) of subgenus 
 Barkhausia and found that there is a definite difference with respect to 
 the initial chiasma frequency and the degree of terminalization. It was 
 concluded that the chromosome complements of two taxonomically well- 
 differentiated species of the same genus may be similar in external 
 structure, but dissimilar in behavior in meiosis, thus showing the impor- 
 tant internal differences between these species. 
 
 Hitherto, the study of meiosis in species hybrids in Crepis has been 
 restricted to the counting of bivalents and univalents appearing at meta- 
 phase. In these studies the proportion of bivalents observed in the va- 
 rious hybrids has been regarded as an indication of the degree of genetic 
 affinity between the parent species. According to Darlington (op. cit., 
 pp. 159-160), amount of pairing between hybrids of various species 
 cannot be compared as an indication of genetic similarity without a 
 knowledge of chiasma frequencies in the parents; and differences in 
 amount of pairing will have a very indirect relation to the differentia- 
 tion in genetic properties of the parent species. Accurate data on chiasma 
 frequency and behavior in parents and hybrids are no doubt essential 
 for the elucidation of metaphase association in hybrids characterized by 
 a mediocre or low amount of pairing. But if persistent chiasmata are 
 essential for bivalent association at metaphase, then the amount of meta- 
 phase pairing must directly result from the number of such chiasmata. 
 Also the abundance of evidence that pairing of chromosomes at pachy- 
 tene is usually determined by the conjugation of homologous chromo- 
 meres justifies consideration of the occurrence of a number of bivalents 
 at metaphase as direct evidence of genetic homology between the parent 
 species. Moreover, when high and low amounts of pairing at metaphase 
 are positively correlated with other evidence on genetic relationship, 
 
1936] BabcocTc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 351 
 
 these differences in pairing may be considered as consistent and cogent 
 evidence on relationship of the parent species. With these considerations 
 in mind, it seems worth while to reexamine the data on chromosome pair- 
 ing in all the interspecific hybrids of Crepis in which such studies have 
 been made. 
 
 In order to compare the evidence thus far available from the study of 
 these hybrids with the other categories of evidence already mentioned, 
 comparable data on the meiotic behavior of the chromosomes in ¥ t plants 
 from eleven different interspecific crosses have been assembled in table 6. 
 In connection therewith, references are given to the original papers. In 
 this table the proportions of PMC's with various numbers of bivalents 
 are expressed as percentages of the whole number of PMC's in which it 
 was possible to make counts and, in order further to facilitate compari- 
 son of the hybrids, the mean numbers of bivalents per PMC are given. 
 The classification of the species in subgenera is indicated typograph- 
 ically as follows : Catonia, capitals ; Eucrepis, bold face ; Barkhausia, 
 italics. The phylogenetic relations of the species are roughly indicated 
 by four categories in ascending order, viz. : primitive, less primitive, less 
 advanced, advanced. It is unnecessary to go into detail here concerning 
 the basis of classification into these categories. The features most used 
 are life cycle, size of plant, shape of leaves, size of heads, size of outer 
 involucral bracts, modification of inner bracts, size of florets, size and 
 shape of fruits, number and size of chromosomes. It is believed that the 
 classification here given is a fair representation of the relative phylo- 
 genetic position of the several species. Phyletic relations as based on 
 morphology alone, however, are only approximate and the classification 
 of the species into the four categories is more or less arbitrary. 
 
 Three of the hybrids, numbers 1, 9, and 11, are preeminent in the 
 large amount of pairing exhibited. In each hybrid one of the parent 
 species is more or less primitive and the other parent is in the next 
 higher class. These facts are consistent with the idea that in more primi- 
 tive species there have been relatively fewer genetic changes than in 
 more advanced species. But hybrid 10 is like hybrids 9 and 11 with re- 
 spect to phyletic classification of the parent species, and hybrid 10 shows 
 a comparatively low amount of pairing. A brief consideration of the spe- 
 cific relations between the parents of each hybrid will throw light on 
 this situation. 
 
 In hybrid 1, although aurea is of Catonia and leontodontoides of 
 Eucrepis, both species are perennials, and they are rather similar in 
 certain morphological features, particularly in size and shape of the 
 fruits. Also they both have 5 pairs of chromosomes which are closely 
 similar in size and shape. 1 Furthermore, they are indigenous in the same 
 
 1 A11 comparisons of karyotypes are based on Babcock and Cameron (1934). 
 
352 
 
 University of California Publications in Agricultural Sciences [Vol. 6 
 
 ►J 
 pq 
 
 < 
 
 « hh 
 
 W 
 
 "^ O 
 
 r-H © 
 
 
 
 .1 
 
 
 
 
 
 
 
 
 
 
 
 "030 
 
 O 
 
 CO 
 
 t~ 
 
 IO 
 
 CM 
 
 OS lO 
 
 t^ 
 
 c» 
 
 00 
 
 CO 
 
 ^HOO 
 
 
 CM 
 
 ■* 
 
 CM 
 
 CM 
 
 OO 
 
 CM 
 
 CM 
 
 
 CM 
 
 CM r-H 
 
 Mean 
 imber 
 ivalen 
 erPM 
 
 41 
 
 41 
 
 41 
 
 41 
 
 41 
 
 41 -H 
 
 M 
 
 -H 
 
 4i 
 
 41 
 
 -U-H 
 
 t~ 
 
 
 
 "5 
 
 10 
 
 CO 
 
 >fflTt< 
 
 00 
 
 lO 
 
 "5 
 
 00 
 
 o*t< 
 
 a-° a 
 
 ■* 
 
 CM 
 
 CM 
 
 »— < 
 
 - 
 
 CMrt 
 
 CM 
 
 CM 
 
 CO 
 
 ** 
 
 CO CO 
 
 So 
 
 
 Tf 
 
 t^ 
 
 Ui 
 
 CO 
 
 CM CO 
 
 O 
 
 O 
 
 CO 
 
 lr^ 
 
 ■*•* 
 
 f2 s 
 
 CO 
 
 IO 
 
 CM 
 
 CM 
 
 >c 
 
 3*- 
 
 co 
 
 >c 
 
 CO 
 
 «5 
 
 COrt 
 
 
 _, 
 
 
 
 O 
 
 CM 
 
 •* 
 
 
 
 HN 
 
 
 
 00 
 
 
 
 o> 
 
 00 
 
 
 
 
 
 CM 
 
 CM 
 
 •>* 
 
 co 
 
 (M 
 
 
 
 
 
 
 
 _ 
 
 
 
 05 
 
 Tfl 
 
 CD 
 
 CO 
 
 OOlO 
 
 CO 
 
 ■«=)< 
 
 
 
 t^ 
 
 ■H CM 
 
 
 
 
 T— I 
 
 O 
 
 i-H 
 
 !— 1 
 
 CM 
 
 
 
 
 CO 
 
 
 jd 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 P 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 GO 
 
 b 
 
 s 
 
 111 
 
 M 
 
 
 
 CO 
 
 O 
 
 O 
 
 co 
 
 CO O 
 
 t~ 
 
 a> 
 
 00 
 
 
 
 ,-ITl< 
 
 CN 
 
 
 CM 
 
 CM 
 
 CM 
 
 CM 
 
 coco 
 
 CO 
 
 »— 1 
 
 
 CO 
 
 , ~ l 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 *o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ■»* 
 
 
 eo 
 
 lO 
 
 ia 
 
 O 
 
 CM 
 
 1000 
 
 l^ 
 
 
 
 
 «# 
 
 t^t^ 
 
 a 
 
 
 co 
 
 
 1-1 
 
 1-1 
 
 CM 
 
 CM 
 
 Ifl 1-H 
 
 <M 
 
 •* 
 
 CO 
 
 y ~ i 
 
 CM 
 
 
 
 ,_, 
 
 CO 
 
 O 
 
 CM 
 
 
 as 
 
 <n 
 
 O 
 
 
 
 ^ 
 
 
 
 ^H^ 
 
 
 
 CM 
 
 CM 
 
 ■<s« 
 
 
 
 CO 
 
 
 CO 
 
 " 
 
 C32 id 
 
 
 10 
 
 •* 
 t^ 
 
 
 
 
 
 
 
 
 
 
 <5« 
 
 
 
 
 
 
 
 
 •<*< 
 
 
 
 
 
 
 
 
 
 
 •* 
 
 
 
 
 
 
 
 
 
 
 
 
 ■<*> 
 
 
 .C3 
 
 "* 
 
 
 
 
 5 
 
 a 
 
 a 
 
 > 
 
 
 
 
 
 "a 
 
 
 s 
 
 
 
 53 
 O 
 
 
 
 
 T3 
 
 
 
 
 
 CO 
 
 00 
 1. 
 
 
 53 
 
 H 
 (3 
 
 00 
 
 
 
 
 
 
 
 
 
 B 
 
 
 K 
 
 
 
 
 
 
 
 
 
 -s 
 
 
 53 
 
 3 
 
 
 
 
 T) 
 
 
 
 
 
 
 
 
 
 
 ■* 
 
 
 a 
 
 a 
 
 •d 
 
 B 
 
 
 ■*f 
 
 •* 
 
 CO 
 
 
 CO 1 * 
 
 ■<*< 
 
 
 
 CO 
 
 
 
 O 
 
 eo 
 
 O 
 
 CO 
 "|H 
 
 
 .2 
 
 CO fcH 
 
 H 
 
 O 
 
 53 
 
 ■* 
 e 
 
 ■3 
 
 *G 
 
 
 u 
 
 to 
 
 ■* 
 
 C3 
 
 8 
 
 
 -t-» 
 
 > 
 
 
 
 •-H -VJ 
 
 •*» 
 
 u 
 
 CO 
 
 
 
 C5 
 
 
 O 
 
 hi 
 
 9" 
 
 
 O. O 
 
 O 
 
 A 
 
 s. 
 
 
 CD 
 
 CO 
 
 eg 
 
 
 eg cd 
 
 CD 
 
 00 
 
 09 
 
 5 
 
 V 
 
 
 
 
 O. 
 
 <j 
 
 
 *» 
 
 ** 
 
 53 
 
 C3 
 
 00 
 
 
 iC 
 
 »o 
 
 "5 
 
 iC 
 
 10 
 
 
 
 
 
 
 
 O 
 
 to 
 
 CO 
 
 ca 
 
 CO 
 
 CO 
 
 
 
 
 
 
 
 > 
 
 0) 
 
 CO 
 
 CD 
 
 CD 
 
 CD 
 
 
 
 
 
 
 
 
 !2 
 
 T3 
 
 ■O 
 
 •0 
 
 "O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 J 
 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 Tf< 
 
 +* 
 
 ■** 
 
 ■** 
 
 ■** 
 
 •*» 
 
 
 
 
 
 •* 
 
 CO 
 
 a 
 
 a 
 
 a 
 
 a 
 
 d 
 
 d 
 
 
 
 
 «* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C3 
 
 CO 
 
 d 
 
 $ 
 
 T3 
 
 •a 
 
 ■a 
 
 T3 
 
 •0 
 
 
 
 
 2 
 
 'e 
 
 J 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ca 
 
 .« 
 
 CD 
 CD 
 
 
 d 
 
 a 
 
 a 
 
 d 
 
 d 
 
 
 
 
 C3 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 a 
 
 
 
 
 « 
 
 CD 
 
 CD 
 
 CD 
 
 CD 
 
 
 
 
 
 
 53 
 
 •** 
 
 13 
 
 '5 
 
 
 "5 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 < 
 
 
 
 
 
 
 
 
 
 
 
 .■£ 
 
 w 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 tf 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .. 
 
 
 CM 
 
 CO 
 
 ■* 
 
 IO 
 
 CO 
 
 1> 
 
 00 
 
 OJ 
 
 
 
 I— ( 
 
 
 
 
 
 
 
 
 
 
 
 
 
 z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "2^ 
 
 2.S 
 
 CdO 
 Si 
 
 o 2 
 
 So 
 
 t-T) o 
 
 pq S 73 
 
1936] Bdbcotfc-Emsweller : Meiosis in Interspecific Hybrids in Crepis 353 
 
 geographic area, although aurea is a montane species and leontodon- 
 toides occurs at low altitudes. This evidence, together with the fact of 
 high degree of pairing in the hybrid, warrants the inference that al- 
 though leontodontoides has developed the more specialized involucre of 
 Eucrepis and in other respects is a less primitive species than aurea, yet 
 there is a high degree of homology between the chromosomes of the two 
 species. 
 
 The parents of hybrid 9, aculeata and aspera, are in the same section 
 of Barkhausia and are obviously closely related ; but aculeata is some- 
 what more primitive, having larger heads, florets, and fruits, less dif- 
 ferentiated marginal achenes, and similar but much larger chromo- 
 somes. Therefore a high degree of genie homology may be assumed, and 
 an even larger amount of pairing at metaphase than was observed might 
 be expected. As 61 per cent of the PMC's had 4n and 31 per cent 3n, it is 
 possible that the difference in size or some other structural difference 
 prevented chiasma formation or terminalization in one particular chro- 
 mosome type. A study of pachytene and later stages in this hybrid would 
 be of interest. The two species occur in the same geographic area. 
 
 Hybrid 11 is the subject of the first part of the present paper, and 
 the differences between HA and 11B in number of pairs at metaphase 
 have been discussed. At this point only hybrid A will be considered. 
 Although nicaeensis is of Eucrepis and setosa of Barkhausia, and al- 
 though nicaeensis is somewhat more primitive, being biennial, larger 
 throughout, and with thicker, beakless fruits, nevertheless the two spe- 
 cies are generally similar. This is true especially when setosa typica is 
 compared with nicaeensis, and when habit, type of involucre, and size 
 of heads, florets, and fruits are considered. The chromosomes are of the 
 same four types, and in only one of the types do the representatives of 
 the two species differ greatly in size. Furthermore, setosa typica is in- 
 digenous in low altitudes of the same geographic area as nicaeensis, 
 which is montane. Hence the nearly complete metaphase pairing and 
 the closeness of the association in the bivalents can be interpreted as 
 indicating a high degree of genie homology in these species. 
 
 In hybrid 10 a quite different situation exists. The phyletic relation 
 is the reverse of that in number 11, the Barkhausia species, divaricata, 
 being somewhat more primitive than the Eucrepis species, Dioscoridis. 
 Both species have 4 pairs of chromosomes. But on the basis of compara- 
 tive morphology they are widely separated ; for they differ throughout 
 and notably in that divaricata has less reduced outer involucral bracts 
 and monomorphic, shortly beaked achenes ; whereas Dioscoridis has a 
 rather highly specialized involucre and dimorphic, unbeaked achenes. 
 The two species are widely separated geographically, divaricata being 
 endemic in Madeira, and Dioscoridis in Greece. The amount of chromo- 
 
354 University of California Publications in Agricultural Sciences [Vol. 6 
 
 some pairing in this hybrid is practically the same as in hybrid 5. In the 
 latter the parents were leontodontoides, a relatively primitive species of 
 Eucrepis with 5 pairs of chromosomes, and Marschallii, an advanced 
 species of Barkhausia with 4 pairs. These species are indigenous in 
 widely separated regions, leontodontoides in Italy and southwestern 
 France, Marschallii in the Caucasus and Caspian region. All this evi- 
 dence indicates that in both of these hybrids the parents represent 
 widely divergent phyletic lines within the genus and is consistent with 
 the notion that the low amount of metaphase pairing in these two hy- 
 brids indicates relatively low amounts of genie homology. 
 
 Hybrids 2, 3, and 4 may be considered together since the phyletic 
 relations are similar. Unfortunately the data on numbers 3 and 4 are 
 very limited and in number 3 the distribution of PMC's according to 
 number of bivalents present is irregular. In all three hybrids both par- 
 ents are Eucrepis species, and the more primitive one is leontodontoides, 
 which is indigenous in Italy and southwestern France. The natural dis- 
 tribution of the other three species is as follows : tectorum, middle and 
 northern Eurasia; parviflora, eastern Mediterranean to Caucasus; 
 capillaris, southern and middle Europe east to Crimea. The mean num- 
 bers of bivalents per PMC are : hybrid 2, 2.0 ; hybrid 3, 2.5 ; hybrid 4, 
 1.5. The amount of pairing in number 2 is consistent with the view that, 
 in Crepis, species which are widely separated in geographic distribution 
 have less genetic homology than species occupying the same region. The 
 slightly larger amount of pairing in hybrid 3, of which the parents are 
 more widely separated geographically, is in apparent disagreement with 
 this conception, but the small number of PMC's counted and their irreg- 
 ular distribution according to number of bivalents present, make this 
 apparent exception less significant. Hybrid 4 also appears to be an ex- 
 ception to the hypothesis because the geographic areas of the parents 
 are the closest in this hybrid, yet the amount of pairing is the lowest. 
 But such an inference is not warranted for this hybrid, because the capil- 
 laris plants used as parents were all of the X strain in which wide varia- 
 tions in pairing occur between individuals 
 
 Hybrids 6, 7, and 8 may also be considered together. Although in num- 
 ber 6 two forms of tectorum were used, the differences in pairing ob- 
 served were not comparable to those obtained in hybrid 11 because of the 
 known variability of pairing in the capillaris plants used in this cross. 
 For the same reason hybrid 6 cannot be compared with hybrids 7 and 8 
 with any degree of assurance. It may be noted, however, that the amount 
 of pairing in number 6A is more nearly what might be expected from 
 the morphological similarity of the two species and their geographic 
 distribution. Hybrids 7 and 8 can be compared with more assurance. All 
 four parental species have 4 pairs of cln-omosomes. Although number 7 
 
1936] Babcock-Emsweller : Meiosis in Interspecific Hybrids in Crepis 355 
 
 was a hybrid between a Eucrepis species and a Barkhausia species and 
 number 8 involved two Barkhausia species, the phyletic relations be- 
 tween the two pairs of species are similar in that one species is less 
 advanced than the other. The two hybrids differ, however, in the degree 
 of difference in advancement of the parents. In other words, the degree 
 of advancement of taraxacifolia over tectorum is greater than that of 
 bursifolia over aspera. But the amount of pairing in the two hybrids is 
 equal, the slight difference between them not being significant. This ap- 
 parent inconsistency is offset by the evidence from geographic distri- 
 bution. In number 7 the areas occupied by the two species overlap, but in 
 number 8 one parent (aspera) occurs in Asia Minor and the other (bur- 
 sifolia) is endemic in Italy. Therefore the geographic distribution of the 
 two pairs of species seems to correspond with their degree of genie 
 homology, as indicated by metaphase pairing in their respective hybrids, 
 if allowance be made for the difference in phyletic relations. 
 
 In the foregoing review the degree of genetic homology between pairs 
 of species, as indicated by the number of bivalents at IM in their respec- 
 tive hybrids, has been considered with reference to the relative phyletic 
 status of the species and their geographic distribution. The hybrids in- 
 vestigated are too few to warrant any conclusions, and the crossings were 
 not planned with reference to a study of the relation of metaphase pair- 
 ing to either phyletic status or geographic distribution considered alone. 
 But, if both these variables are taken into account when the metaphase 
 pairing in two hybrids is compared, two inferences appear to be indi- 
 cated, as follows: (1) hybrids between more primitive species have 
 larger amounts of metaphase pairing than hybrids between more ad- 
 vanced species; (2) hybrids between species occupying the same geo- 
 graphic region also exhibit greater meiotic regularity than do hybrids 
 between species from widely separated regions. It is to be emphasized 
 that these inferences apply only to the Crepis hybrids herein considered. 
 
 That hybrids between more primitive species of a monophyletic genus 
 will exhibit greater meiotic regularity than do hybrids between more 
 advanced species, is a reasonable a priori assumption, because the more 
 primitive species would be less differentiated genetically in a truly nat- 
 ural group. From a study of meiotic behavior in Nicotiana interspecific 
 hybrids, Goodspeed (1934) reached the conclusion that formation of 
 bivalents in a hybrid is positive evidence that the pairing chromosomes 
 possess many similar or equivalent genie elements ; and that decrease in 
 amount of pairing "is to be related to an accumulation of genie and 
 chromosomal distinctions between species evolved from a common 
 stock." The data on Crepis hybrids are not adequate for any generaliza- 
 tion about phyletic status and meiotic regularity. There should be sev- 
 eral series like hybrids 1-5, all of which involve a single species for one 
 
356 University of California Publications in Agricultural Sciences [Vol. 6 
 
 parent, and both parents should occur in the same region. Even in this 
 series numbers 3 and 4 must be disregarded because of unreliability of 
 the data, and in number 5 the factor of wide geographic separation is 
 introduced. This leaves only the first two hybrids, in which there is a 
 significant difference in amount of pairing and a great difference in 
 phyletic status of aurea and tectorum. The more primitive aurea pro- 
 duced a hybrid with high meiotic regularity and high fertility, and the 
 more advanced tectorum gave a hybrid with low metaphase pairing and 
 extremely low fertility. 
 
 With reference to the second inference, it seems probable that, in a 
 recently evolved genus with many closely related but widely separated 
 species, there is positive correlation between wide geographic distribu- 
 tion and a low degree of genetic homology. This also is a reasonable 
 assumption because groups of most closely related species would be dis- 
 tributed from the center or origin during the course of evolution. In the 
 present study hybrids 8 and 9 support this inference because all four 
 species are of subgenus Barkhausia, the parents of hybrid 8 being widely 
 separated geographically and less similar morphologically, whereas the 
 parents of hybrid 9 occur in the same region and are generally similar. 
 Hybrid 8 has a mean of 2.5 ± .28 bivalents per PMC, and hybrid 9 has 
 a mean of 3.5 ±.18. The difference, 1.0 ± .33, is three times its standard 
 error and hence may be considered significant. A similar situation holds 
 in hybrids 10 and 11 and in these there is a much greater difference in 
 the number of bivalents formed in the hybrids. It should be emphasized, 
 however, that this positive correlation between wide geographic separa- 
 tion and relatively low genetic homology is suggested only for the genus 
 Crepis and, perhaps, for other large, recently evolved, but widely dis- 
 tributed genera. 
 
 As a matter of fact, exactly the opposite relation has been reported 
 in certain genera, such as Philadelphus (Bangham, 1929; Sax, 1931) 
 and Platanus (Sax, 1933). Both of these are old genera. According to 
 Seward (1931), Platanus is one of the oldest of the broad-leaved trees, 
 and we are informed by Chaney (in litt.) that there is no question of its 
 presence in early Tertiary rocks in many parts of the world, including 
 western North America. With respect to Philadelphus, Chaney (1927) 
 has recognized it in middle Tertiary deposits of eastern Oregon, and 
 Cockerell (1908) has recognized it in the Upper Miocene of Colorado. 
 At all events Platanus and Philadelphus are certainly more primitive 
 and certainly much older than Crepis. In both those genera hybrids have 
 been obtained between certain Old World and New World species hav- 
 ing the same chromosome number, and some of these hybrids exhibit 
 complete or nearly complete pairing of the chromosomes, although there 
 is much variation among them with respect to fertility. The perfect or 
 
1936] Babcoch-Emsweller : Meiosis in Interspecific Hybrids in Crepis 357 
 
 nearly perfect pairing at diakinesis or IM in these hybrids indicates a 
 high degree of genetic homology in the parental species. Apparently the 
 chromosomes of these widely separated species have undergone insuffi- 
 cient change to prevent pairing although the species have been isolated 
 for millions of years. In the present discussion it is unnecessary to con- 
 sider the problem of stability and plasticity in species and genera. It is 
 sufficient to note that such differences exist and that relations between 
 geographic distribution and genetic homology of species will vary ac- 
 cordingly. 
 
 Summarizing the evidence on intrageneric relations in Crepis leads to 
 certain generalizations about this genus. Crepis is a large group of 
 closely related species. Although natural major subgeneric groups are 
 clearly indicated and these groups provide a convenient basis for syste- 
 matic classification, yet the species thus classified under different sub- 
 genera are still more or less closely related. That is, their genie com- 
 plements are more or less homologous. This is certainly indicated by 
 the evidence on metaphase pairing in the interspecific hybrids reviewed 
 above. It is supported also by two other categories of evidence. (1) It 
 has already been shown (Babcock and Cameron, 1934; Babcock, 1936) 
 that in several groups of morphologically similar species there is a 
 single karyotype even though the species occupy different geographic 
 areas. Evidently differentiation and distribution has not been accom- 
 panied in these groups by visible marked changes in the chromosomes. 
 Thus the inference seems warranted that speciation has been made pos- 
 sible by a limited number of gene mutations or minute structural 
 changes, leaving the main part of the residual genotype unchanged. (2) 
 This inference is shown to be valid by the results of the investigations 
 on three such groups of species made by Jenkins, Cave, and Smith 
 independently (unpublished data). In each group the interspecific hy- 
 brids studied exhibited almost complete pairing at metaphase ; and 
 genetic experiments with alternative characters in different species 
 proved that these characters are inherited in Mendelian fashion. In 
 general, therefore, the evidence on metaphase pairing in interspecific 
 hybrids in Crepis is consistent with all the other evidence on phylogeny 
 in this genus. 
 
SUMMARY AND CONCLUSIONS 
 
 1. Two subspecies of Crepis setosa, namely, typica from western Eu- 
 rope and Topaliana from Greece, were used in this study. They had 
 been ranked as subspecies because of the differences they exhibited in 
 external morphology. Each subspecies was used as the paternal parent 
 in a cross made with the same plant of C. nicaeensis. The diploid chro- 
 mosome number in both species is 8. The morphology of the chromo- 
 somes of the setosa subspecies is apparently the same, and they differ 
 sufficiently from the members of the nicaeensis genom to make identifi- 
 cation in Pj relatively easy. 
 
 2. The interspecific hybrids were produced with difficulty, and only 
 one plant from each cross was available for cytological study. These are 
 designated as hybrid A (nicaeensis $ x setosa typica $) and hybrid B 
 {nicaeensis § x setosa Topaliana <$) . 
 
 3. The satellite of the nicaeensis D chromosome was not found in 
 either F x plant, and in each plant there was an increase in the size of 
 the head of this chromosome. 
 
 4. Meiosis in the ¥ 1 plants showed that the chromosomes of setosa 
 typica paired very regularly with the nicaeensis chromosomes and ex- 
 hibited a very close association of chromosomes in bivalent formation, 
 whereas those of setosa Topaliana exhibited less regular pairing and 
 looser association. Percentage of good pollen grains was determined and 
 showed a high correlation with meiotic behavior, being much greater in 
 hybrid A than in hybrid B. Comparative fertility under open pollina- 
 tion was consistent with the foregoing evidence. 
 
 5. An F, population of 87 plants was grown from hybrid A, and one 
 of 17 plants from hybrid B. In the first group, eighty-four plants were 
 found to contain 24 chromosomes and three plants had 8 chromosomes ; 
 in the second group, fifteen plants had 24 chromosomes and the remain- 
 ing two plants had 8 chromosomes. The 24-chromosome plants were 
 found to have arisen by spontaneous crossing with Crepis biennis. "With 
 the exception of but six plants, they were all biennial. Each of the six 
 that bloomed the first year was slightly fertile under isolated conditions. 
 
 6. The F 2 plants with 8 chromosomes were studied carefully because 
 of their unusual chromosome complement. The morphology of these 8 
 chromosomes indicated that 6 were from setosa — one A, two B, one C, 
 and two D ; and the two remaining were apparently nicaeensis D chro- 
 mosomes. This supposition, however, was not supported by meiotic be- 
 havior, either in F x or in F 2 ; and it was suggested that an exchange of 
 material occurred in F x and that the F 2 chromosomes represent recombi- 
 nation products of this exchange. 
 
 [358 ] 
 
1936] Bdbcock-Emsweller : Meiosis in Interspecific Hybrids in Crepis 359 
 
 7. The various stages in the formation of giant pollen grains in F 2 
 plants with 8 chromosomes were observed. But polyploids were not ob- 
 tained in F 3 , primarily because of high sterility of the F 2 plants and low 
 viability of the F 3 progeny. Slower germination of the polyploid pollen 
 may have been a secondary factor. 
 
 8. The two subspecies of Crepis setosa are so similar morphologically 
 that their classification as different species does not seem to be justified. 
 Although their chromosomes are exactly similar in appearance, yet they 
 must differ somewhat in genetic composition. This is indicated both by 
 the morphological differences between the subspecies, and by the differ- 
 ences in metaphase pairing in hybrids A and B. Classification as subspe- 
 cies, at least for the present, has the advantage of emphasizing their 
 very close affinity. 
 
 9. The genus Crepis is a group of closely related species. The sub- 
 genera, Catonia, Eucrepis, and Barkhausia, are not widely differenti- 
 ated, but overlap more or less, and several hybrids have been produced 
 between species of different subgenera. Eleven interspecific hybrids 
 between diploid species have been studied by various investigators with 
 particular reference to the amount of pairing at metaphase. The evi- 
 dence indicates that the genie complements of the fourteen species in- 
 volved are all more or less homologous. This inference is consistent with 
 the evidence on chromosome morphology in the genus, on geographic 
 distribution, and on other results (as yet unpublished) from genetic 
 investigations on groups of closely related species. The evidence on 
 metaphase pairing in interspecific hybrids, therefore, supports the con- 
 ception that the subgenera of Crepis had a common origin and are still 
 more or less similar in genetic composition. 
 
LITERATURE CITED 
 
 Avery, Priscit.t.a 
 
 1930. Cytological studies of five interspecific hybrids of Crepis leontodontoides. 
 Univ. Calif. Publ. Agr. Sci., 6:135-167. 
 
 Babcock, E. B. and Clausen, J. 
 
 1929. Meiosis in two species and three hybrids of Crepis and its bearing on taxo- 
 nomic relationship. Univ. Calif. Publ. Agr. Sci., 2:401-432. 
 
 Babcock, E. B. and Navashin, M. 
 
 1930. The genus Crepis. Bibliographia Genetiea, 6:1-90. 
 
 Babcock, E. B. and Cameron, D. E. 
 
 1934. Chromosomes and phylogeny in Crepis II. The relationships of one hundred 
 eight species. Univ. Calif. Publ. Agr. Sci., 6:287-324. 
 
 Babcock, E. B. and Swezt, O. 
 
 1935. The chromosomes of Crepis biennis L. and C. ciliata C. Koch. Cytologia, 6 : 
 256-265. 
 
 Babcock, E. B. 
 
 1936. The origin of Crepis and related genera with particular reference to distri- 
 bution and chromosome relationships. Univ. Calif. Publ. Bot. Setchell 
 Festschrift Volume: 9-53. 
 
 Bangham, W. 
 
 1929. The chromosomes of some species of the genus Philadelphus. Jour. Arnold 
 Arboretum, 10:167-169. 
 
 Beadle, G. W. 
 
 1930. Genetical and cytological studies of Mendelian asynapsis in Zea mays. Cor- 
 nell Univ. Agr. Exp. Sta. Mem., 129:1-23. 
 
 Chaney, B. W. 
 
 1927. Geology and Paleontology of the Crooked Eiver Basin. Carnegie Inst. Wash. 
 Publ., 346:45-138. 
 
 COCKERELL, T. D. A. 
 
 1908. Fossil Flora of Florissant, Colorado. Amer. Mus. Nat. Hist. Bull., 24: 
 71-110. 
 
 Collins, J. L. and Mann, M. 
 
 1923. Interspecific hybrids in Crepis II. A preliminary report on the results of 
 hybridizing Crepis setosa Hall. f. with C. capillaris (L.) Wallr. and with C. 
 biennis L. Genetics, 8:212-232. 
 
 Collins, J. L. 
 
 1924. Inheritance in Crepis capillaris (L.) Wallr. III. Nineteen morphological 
 and three physiological characters. Univ. Calif. Publ. Agr. Sci., 2:249-296. 
 
 Collins, J. L., Hollingshead, L., and Avery, P. 
 
 1929. Interspecific hybrids in Crepis III. Constant fertile forms containing chro- 
 mosomes derived from two species. Genetics, 14:305-320. 
 
 Darlington, C. D. 
 
 1932. Eecent Advances in Cytology, xviii + 559 pp. Philadelphia. 
 
 [ 360] 
 
1936] Babcock-Emsweller : Meiosis in Interspecific Hybrids in Crepis 361 
 
 De Mol, W. E. 
 
 1932. Die Veredelung des hollandischen Varietaten von Hyacinthus orientalis L., 
 und damit in Zusammenhang : einige Ergebnisse iiber Selbstbestaubung 
 und Kreuzbestiiubung bei diploiden mit heteroploiden Formen dieser Pflan- 
 zenart. Studia Mendeliana, Briinn. 
 
 Goodspeed. T. H. 
 
 1934. Nicotiana phylesis in the light of chromosome number, morphology, and be- 
 havior. Univ. Calif. Publ. Bot., 17:369-398. 
 
 HOLLINGSHEAD, E. L. 
 
 1930. Cytological investigations of hybrids and hybrid derivatives of Crepis capil- 
 lars and Crepis tectorum. Univ. Calif. Publ. Agr. Sei., 6:55-94. 
 
 Hollingshead, E. L. and Babcock, E. B. 
 
 1930. Chromosomes and phylogeny in Crepis. Univ. Calif. Publ. Agr. Sci., 6:1-53. 
 
 Karpechenko, G. D. 
 
 1927. The production of polyploid gametes in hybrids. Hereditas, 9:349-368. 
 
 Matsuda, H. 
 
 1927. On the origin of big pollen grains with an abnormal number of chromo- 
 somes. La Cellule, 38:215-242. 
 
 Muntzing, A. 
 
 1934. Chromosome fragmentation in a Crepis hybrid. Hereditas, 19:284-302. 
 
 Navashin, M. 
 
 1926. Variabilitat des Zellkerns bei Crepis-Arten in Bezug auf die Artbildung. 
 Zeitsch. f. Zellforseh. mikr. Anat., 4:171-215. 
 
 Navashin, M. 
 
 1927. Tiber die Veranderung von Zahl und Form der Chromosomen infolge der 
 Hybridization. Zeitsch. f. Zellforseh. mikr. Anat., 6:195-233. 
 
 Navashin, M. 
 
 1928. "Amphiplastie" — eine neue karyologisehe Erscheinung. Zeits. f. Indukt. 
 Abst. Vererb. Suppl., 2:1148-1152 
 
 Navashin, M. 
 
 1933. The dislocation hypothesis of evolution of chromosome numbers. Zeits. ind. 
 Abst. Vererb., 63:224-231. 
 
 Navashin, M. 
 
 1934. Chromosome alterations caused by hybridization and their bearing upon 
 certain general genetic problems. Cytologia, 5:169-203. 
 
 Poole, C. F. 
 
 1932. The interspecific hybrid Crepis rubra x C. foetida and some of its deriva- 
 tives II. Two selfed generations from an amphidiploid hybrid. Univ. Calif. 
 Publ. Agr. Sci., 6:231-255. 
 
 Richardson, M. M. 
 
 1935a. Meiosis in Crepis I. Pachytene association and chiasma behaviour in Crepis 
 
 capillaris (L.) Wallr. and C. tectorum L. Jour. Genetics, 31:101-117. 
 1935b. Meiosis in Crepis II. Failure of pairing in Crepis capillaris (L.) "Wallr. 
 
 Jour. Genetics, 31:119-143. 
 
362 University of California Publications in Agricultural Sciences [Vol. 6 
 
 BOSENBERG, O. 
 
 1927. Die semiheterotypische Teilung und ihre Bedeuting fiir die Entstehung 
 verdoppelter Chromosomenzahlen. Hereditas, 8:305—338. 
 
 Sakamura, T. and Stow, I. 
 
 1926. Ueber die experimentell veranlaste Entstehung von Keimfahigen Pollen- 
 kornern mit abweiehenden Chromosomenzahlen. Jap. Jour. Bot., 3:111- 
 137. 
 
 Sax, K. 
 
 1931. Chromosome numbers in the ligneous Saxifragaceae. Jour. Arnold Arbore- 
 tum, 12:198-206. 
 1933. Species hybrids in Platanus and Campsis. Jour. Arnold Arboretum, 14: 
 274-278. 
 
 Seward, A. C. 
 
 1931. Plant Life through the Ages. 601 pp. Cambridge, England. 
 
 Zirkle, C. 
 
 1931. The use of n-butyl alcohol in dehydrating woody tissue for paraffin em- 
 bedding. Sci., 71:103-104. 
 
EXPLANATION OF PLATES 
 
PLATE 14 
 Fruiting heads 
 
 a. Cii pis st tosa h/pica (2623). 
 
 b. C. setosa Topaliana (2671). 
 
 c. C. nicaeensis (2700). 
 
 d. C. nicaeensis JxC. setosa typica £ F t (30.31). 
 
 e. C. nicaeensis § x C. setosa Topaliana £ F : (30.32). 
 
 Somatic chromosomes 
 
 /. Photomicrograph of mitotic metaphase in a root-tip 
 cell of an 8-ehromosome F 2 plant. 
 
 | 364 | 
 
UNIV. CALIF. PUBL. AGR. SCI. VOL. 6 [ B ABCOCK-EMSWELLER] PLATE 14 
 
 Y 
 
 / 
 
 iVV' 
 
 / 
 
 [365 ] 
 
PLATE 15 
 (Crepis nicaeensis x C. setosa typica) J x C. biennis J 1 
 
 a-e. Basal leaves from five 24-chromosome plants de- 
 rived from open-pollinated Fj. 
 
 Crepis nicaeensis x C. setosa typica, F, progeny. 
 
 /, g. Two of the 8-ehromosome plants. 
 
 h, i. Two of the plants that died in early rosette stage, 
 before the chromosomes were examined. 
 
 | 366 ] 
 
UNIV. CALIF. PUBL. AGR. SCI. VOL. 6 f B ABCOCK-EM SW ELLER ] PLATE 15 
 
 [367]