'931 The secondary association of chromosomes 355 



and three bivalents respectively at first metaphase (Figs. Ill and V), 

 though the association is not so pronounced as in the second meta- 

 phases figured, and varies in degree. Ishikawa makes no mention of 

 chromosome association in Dahlia coronata (2n=32) which he also ex- 

 amined, but his figures of that species show a similar grouping of 

 bivalents at first metaphase (Fig. II) and especially at second metaphase 

 (Fig. VI). The association is not so strong as in the same stages in 

 D. variabilis. Ishikawa does not say whether the bivalents are asso- 

 ciated at diakinesis in these species. He concluded that the association 

 of pairs of bivalents was a sign of their homology, and further that 

 D. variabilis and D. coronata were tetraploid and diploid species respec- 

 tively. 



The associations which Marchal observed in Amblystegium were 

 apparently all, or nearly all, quadrivalent rings, and therefore are 

 not relevant to this discussion, not being due to secondary asso- 

 ciation. 



No further mention is made of the association of chromosomes until 

 Frost (1925) reported with regard to some tetraploid Citrus seedlings 

 that ' ' part or all of the bivalents are often more or less closely associ- 

 ated in quadrivalent groups." Cra.ne and Darlington (1917) also refer, 

 to frequent " secondary association " in Rubus species and hybrids as 

 an indication of the homology of the chromosomes. Hagerup (1927) 

 found association of bivalents in Empetrum hermaphroditum (2n=52), 

 and this association, unnoticed by Hagerup, is also seen in his drawings 

 of first metaphase in E. nigrum (2n=26). Darlington (1928) working 

 on Prunus, reported that " a careful study of most plates seen in polar 

 view shows that the bivalents are to a certain extent lying in pairs." 

 He further stated that ' ' this secondary pairing, begun at diakinesis, 

 is perhaps most strongly developed at metaphase of the first division. 

 It is not therefore a development of or continuation of a prophase rela- 

 tionship—but must be a secondary association." Secondary associa- 

 tion was also observed and figured by Buxton and Newton (1928) in 

 Digitalis species and hybrids. 



In each of the species and hybrids studied by the above workers 

 primary and secondary chromosome associations occur together, but at 

 this time no certain distinction could be made between the two types 

 of association. As the true nature of prophase pairing became under- 

 stood it was possible to differentiate between the types of association, 

 by examination of the critical stage of diakinesis, and from other 

 evidences to be discussed later. 



