GERM CELLS OF GRYLLOTALPA 295 



these chromosomes on the spindle is a matter of chance. This 

 method of distribution causes four kinds of secondary spermato-. 

 cytes. These according to Voinov are as follows: 



1. 4 dyads plus m-chromosome plus bivalent plus Y; 



2. 4 dyads plus m-chromosome plus accessory plus X; 



3. 4 dyads plus m-chromosome plus bivalent plus X; ; 



4. 4 dyads plus m-chromosome plus accessory plus Y. 

 Voinov does not discuss the relation of these four kinds ofi 



spermatocytes to sex production. Neither does he discuss 

 whether the XY pair is related to sex production or whether it 

 is an unequal pair of autosomes, as described by Carothers and 

 Robertson. 



c. A comparison of my observations made on material collected 

 at Naples with those of Voinov. Unfortunately I do not have 

 enough good material to work out the chromosomal history in 

 either the Freiburg or the Naples form. I wish, however, to 

 compare the observations of Voinov with those which I have 

 made on the material collected at Naples. It seems probable 

 that our observations have been made on the same form. 



As mentioned before, Voinov figures in the spermatogonia! 

 group one small chromosome which he calls an m-chromosome. 

 In the first spermatocyte division he figures the m-chromosome 

 as a bivalent. This would indicate that it is made up of two 

 spermatogonial chromosomes, and I think such is the case. In 

 my figures of the spermatogonial group (fig. 4, A, B, C) there 

 are clearly 15 chromosomes and two of these are small. Hence 

 it seems probable that Voinov has overlooked one of these small 

 chromosomes. My material showing the first spermatocyte divi- 

 sion is inadequate and hence I cannot draw definite conclusions. 

 It indicates, however, that there are eight chromosomes present 

 instead of seven. This is what we would expect if there are 15 

 chromosomes in the spermatogonial cells and if the accessory 

 is not linked to another pair. If it is linked as Voinov describes, 

 this would reduce the number to seven. I wish to point out at 

 this place a possibility of error in Voinov's interpretation. Cer- 

 tainly his figures are not convincing. The possibility of an 

 error seems all the more probable when we compare his results 



