324 Universifij of California Publications in AuricuUuniJ Sciences [Vol. 2 



grouped under Section II all having 5 pairs of chromosomes of similar 

 size, while those under Sections III and IV have 4 pairs but differ 

 somewhat in individuality. It is worthy of note that one character 

 commonly used in distinguishing between these species, viz., the posi- 

 tion assumed bj' the young flower heads before anthesis, whether erect 

 or nodding, has been found to be too variable in the case of foetida to 

 make it of diagnostic value. 



In its dimorphous fruits, the inner ones beaked and the outer ones 

 winged, C. aspera exhibits relationship with Barkhausia on one side 

 and the Dioscoridk group on the other (cf. fig. 4. k, k'). Its chromo- 

 some group resembles those of the three Barkhausia species in having 

 chromosomes of medium size, and it has been crossed with two of these 

 species. But these hybrids exhibit very abnormal reduction phe- 

 nomena, whereas hybrids between certain Barkhausia species (t^es-i- 

 caria, MdrschaUii and taraxaeifolici) show normal pairing and reduc- 

 tion. Thus all the evidence indicates that aspera belongs in a class by 

 itself. Furthermore, amploxifolia, which closely resembles aspera 

 morphologically, also has 4 pairs of medium-sized chromosomes 

 (p. 331). 



Subgenus {!)) is a heterogeneous group which is scarcely capable 

 of satisfactory classification on the basis of fruit characters alone. 

 Thus in the case of sections V, VI, and VII there is much stronger 

 affinity, as indicated by comparative morphology, than would appear 

 from Hoffmann's synopsis. In all three of the species concerned the 

 inner involucral bracts of fructiferous heads are conspicuously thick- 

 ened or much hardened. Then, too, palaestina has a combination of 

 some of the distinguishing characters of the other two species, and 

 yet it is in no sense an intermediate form such as might arise from 

 hybridization. The flower heads in pahtestina are large and showy, 

 and the marginal fruits are enclosed within the inner involucral bracts, 

 in these respects resembling Dioscoridis, while the inner fruits bear a 

 strong resemblance to those of puJchra. Furthermore, the fruits in 

 pidchra, contrary to Hoffmann, are sometimes of two distinct shapes, 

 the marginal ones being flattened as in palaestina (cf. fig. 4). Without 

 going into further details at this time, we may suggest that these three 

 sections might Avell be combined into one. The chromosome groups 

 of pulchra (N = 4), palaestina' (N = 4), and Dioscoridis (N = 4) 

 are indistinguishable, and the F^ of pidchra X palaestina is highly 

 fertile. 



Section VIII, Aetheorrhiza, must stand alone, at least for the 

 present. While the inflorescence of hidhosa suggests strong relation- 



