134 J- Moore, 



Unless the two last divisions of mammalian spermatogenesis have 

 nothing in common with the reduction divisions of invertebrates, which 

 seems very unlikely, there is nothing equivalent in the higher type to 

 the second division (theoretically the most important) as performed in 

 the lower one. Nor is this all, for, as I shall show later, the longi- 

 tudinal division can, among mammals, apparently be dispensed with. 

 At any rate, I have not been able to satisfy myself that it is diffe- 

 rentiated fi-om the succeeding divisions in the' dog; while in this animal 

 the last division (i. e. that in which the sixteen chromosomes are in 

 the rat reduced to eight) is so much modified as to present characters 

 intermediate between karyokinesis and akinesis, in which there are no 

 distinct chromosomes to count — yet this is the last division, the 

 one most vitally concerned in the equation of the two pro-nuclei. 

 Moreover, akinesis proper appear^s to be reintroduced between this last 

 mitosis and the formation of the spermatozoa! 



In such species as the rat, although the chromosomes of the 

 longitudinal divisions are difficult to count, the difficulty arises from 

 their small size and close crowding. We are quite sure that they 

 exist as distinct chromosomes, and that the chromatin of the parent 

 cell is halved with almost mathematical exactitude; so also in this 

 species with the last and hetrotype division. 



In dogs the first division seems to be either dispensed with, or 

 to have become indistinguishable from the practically akinetic division 

 of the regenerative stock; and, as I have said, the division corre- 

 sponding to the beautiful hetrotype one in the rat is so reduced in 

 its karyokinetic characters that it is quite impossible to count the 

 chromosomes, not because of any crowding or of the small size of 

 these bodies, but simply because they are not differentiated one from 

 another — a fact which seems to me of very considerable theoretical 

 importance, since, in consequence of the chromatin not being cleanly 

 divided, we are not sure that it is accurately distributed between the 

 daughter elements. Is it however possible to imagine that although 

 the chromatin is apparently confused, a real numerical although invisible, 

 identity of the chromosomes remains? 



There is certainly less evidence to justify this view than there 



