478 E. ELEANOR CAROTHERS 



rather extensive review of the latter paper on account of some 

 questionable conclusions and on account of some points of 

 general similarity, as I believe, in our material. 



Voinov is dealing with Gryllotalpa vulgaris. This is the form 

 for which vom Rath ('92) reported the somatic number of chro- 

 mosomes as twelve. He described a doubling of this number 

 just before the first maturation division which reestablished the 

 normal number. The second division brought about reduc- 

 tion to one-half the normal. Vom Rath extended to all animals 

 his conclusion that there is a doubling of the number of chromo- 

 somes before the maturation divisions, both of which are re- 

 ductional and result in halving the normal number of chromo- 

 somes. As this erroneous idea appealed to many as a support 

 of Weismann's hypothesis it was widely accepted for a time. 

 Hence work on this form is of more than usual interest. 



Voinov figures fifteen chromosomes in the spermatogonia. 

 The smallest of these is clearly without a morphological mate 

 and he considers it to be in reality a pair of microchromosomes 

 synapsed in the spermatogonia. This assumption leaves him 

 without a Y for what he considers an XY pair which occurs in the 

 first spermatocytes. He assumes the Y exists in the spermato- 

 gonia but has escaped detection. Therefore, he concludes 

 that seventeen is really the somatic number, while he finds seven 

 in both maturation divisions. This is accounted for by further 

 assuming that two of the seven must be multiples; one an octad, 

 the other a hexad. The hexad he believes to be identical in 

 composition with the chromosome 'en L' of de Sinety ('01) and 

 the hexad of McClung ('05). A comparison of Vomov's chro- 

 mosome marked 'a' and 'bi' (text fig. 8c, see my text fig. 3, 3) 

 with de Sinety's chromosome c-s (fig. 93, plate 3, see my text 

 fig. 3, 1) and with McClung's chromosome 'ac' (text fig. 3, see 

 my text fig. 3, 2) will show a critical difference. In both the 

 latter, if we may assume that de Sinety's chromosome 'en L' is 

 a hexad, the tetrad part divides, resulting in a greater quantity 

 of chromatin passing to one pole than the other, while Voinov's 

 'hexad' divides in such a manner that, though the parts are 

 morphologically different, the quantity of chromatin in each 



