202 I The Process of Evolution 



complexes, which are the despair of the taxonomist and the cyto- 

 geneticist ahke. The genera Rubus (blackberries and raspberries) 

 and Poa (bluegrass) are familiar examples in which relationships 

 are so blurred as to prevent any clear understanding of the group as 

 a whole. The American species of Crepis (false dandelions) form 

 the only agamic complex of size that has been studied in its entirety 

 from both the systematic and cytological points of view. It is interest- 

 ing that, of the 196 species of Crepis ( which have been grouped into 

 27 sections ) , apomixis and polyploidy are known only in five species 

 of the section Pyrimaches in southeastern Asia and in nine species 

 of the section Psilochaena in America. 



The ten North American species of section Psilochaena make up 

 a group more heterogeneous than any other. Their basic chromo- 

 some number, x = 11, is not found in any other section, and it may 

 have originated as a result of interspecific hybridization of Asiatic 

 species with n = 7 and n = 4, or with n = 5 and n= 6. Crepis 

 runcinata is the only member of the section that shows neither 

 apomixis nor polyploidy, having a chromosome number of 2n = 22. 

 Seven of the remaining species each have a 22-chromosome diploid 

 with which are associated numerous polyploid apomictic forms. The 

 other two species include only polyploid apomictic types derived 

 by hybridization between two or more of the above-mentioned 

 seven. 



Thus there are seven primary diploid types involved in polyploidy 

 and apomixis. Five have greatly restricted geographic distribution 

 and very different ecological preferences. No localities are known 

 in which two diploids occur together, and no diploid hybrids have 

 been found. The polyploids, with chromosome numbers of 33, 44, 

 55, 77, and 88, show combinations of characteristics of two or more 

 of the diploids and all are apomictic ( some facultative ) . In general, 

 the apomicts exceed in geographic area their diploid ancestors. 



The analysis of this agamic complex may serve as a case study 

 for such problems in general. The unveiling of relationships and the 

 construction of a useful classification were dependent upon recogni- 

 tion of the primary sexual diploids. Around these could be grouped 

 the derivative autopolyploid, and allopolyploid asexual forms. Al- 

 though the variation at first appears baffling and overwhelming, it 

 can be made sensible; in so doing, interesting facts are revealed. 

 For example, diversity is greatest in the regions where the sexual 

 species are found and is less in the peripheral areas of the range 

 occupied only by apomicts. The latter appear to have radiated, sub- 

 sequent to their formation by hybridization and polyploidy, from 

 the more central area of their sexual ancestors. 



