372 bulletin: museum of compaeative zoology. 



possible explanation, though a rather artificial one ; and that is, that 

 the chromosomes both of somatic and of the earlier generations of germ 

 cells are bivalent, and that it is to the " chromomeres " that we should 

 turn for the univalent chromosomes. Evidence in favor of such a view 

 is afforded by the apparently double nature and outline of the chromo- 

 somes, but I put it forward merely as a suggestion, which I am by no 

 means prepared to support. 



There can no longer be any doubt that in many animals — e. g., worms 

 (Calkins, '95), crustaceans (Nichols, : 02), Peripatus (Montgomery, :00,) 

 chilopods (Blackman, : 05), insects (Montgomery, '98 a ; Sutton, : 02 ; 

 Baumgartner, : 02) — numerical reduction is the result of an end-to-end 

 union of individual chromosomes. This view has been urged so forcibly, 

 and it has fitted in with current theories of heredity so well, that the 

 earlier view — namely, that reduction is caused by a segmentation of 

 the spireme into half the somatic number of chromosomes — has been 

 largely abandoned. Thus Blackman (: 05, p. 95) writes that "it [end- 

 to-end union] may well be common to all sperm cells." But to apply 

 this rule without reserve to all Metazoa is, in the present state of our 

 knowledge, going too far. Although such a simple process of reduction 

 as a mere breaking of the spireme into half the somatic number of seg- 

 ments may never occur, yet it is evident that the course of events in 

 coelenterates and in Ascaris is very different, at least superficially, from 

 that seen among the arthropods. After studying the careful and con- 

 scientious figures of Brauer ('93), Sabaschnikoff ('97), and Tretjakoff 

 (: 04), it is hard to avoid the conclusion that in Ascaris, at least, the 

 numerical reduction is the expression of a re-grouping of the chromatin 

 granules, rather than a pairing of pre-existing chromosomes. Unfortu- 

 nately, however, the process in Ascaris is obscured by the occurrence of 

 a " contraction stage " at the critical period. In Gonionemus, likewise, 

 the conditions are most readily explained on the assumption that there is 

 a re-grouping of the chromatin granules. For it seems to me unreasonable 

 to give to the pairing of chromatin structures in the spermatocyte a sig- 

 nificance which cannot be attributed to the very similar process in the 

 spermatogonia. 



There is one very important point which appears from the evidence 

 now at hand : in every case in which an early pairing of individual chro- 

 mosomes has been demonstrated beyond doubt, — such cases are, with 

 one possible exception, limited to the arthropods, — there is good reason 

 to believe that the chromosomes preserve their morphologic individuality 

 from the last spermatogonia! until the first maturation division. This 



