The Spermatogenesis in Pentatoma up to the Formation of the Spermatid. 73 



riog becomes therefore divided transversely twice. This, however, 

 cannot be the case if each ring has the value of only 2 chromosomes, 

 as I have tried to show above. On the contrary, the two divisions of 

 the rings must be either, in the first division, a transverse division 

 of the two chromosomes of a ring, and in the second mere separation 

 of the remaining portions ; or the first division must result in separat- 

 ing the two chromosomes of a ring from one another, and the second 

 result in the transverse division of the remaining chromosomes. From 

 the figures given by Wilcox it is practically impossible to determine 

 which of these modes occurs. But in neither case is there a possi- 

 bility of an equation division, and in both cases there must be, if my 

 definition of a chromosome be correct, one transverse division of the 

 chromosomes, and one elimination division (vide infra). Whichever way 

 we regard it, our definition shows that the rings have the value of 

 only 2 chromosomes each. 



In Gryllotalpa (Vom Rath, '91, '92, '95) the continuous spirem 

 thread is described as splitting longitudinally for its entire length, and 

 then dividing transversely into 6 segments; the normal number of 

 chromosomes is 12, and hence our definition shows that there is one 

 half the normal number of chromosomes in the 1st spermatocytes. 

 Each of these 6 chromosomes becomes a ring in a manner which has 

 not been very clearly described by Vom Rath, though his later paper 

 ('95, in which the fig. 6 is of prime importance) would seem to show 

 that each of the 6 rings is produced by the opening up of the longi- 

 tudinal split, the ends of the chromosome then remaining in contact 

 with one another, which is followed by a transverse division of the 

 ring. Each ring is, however, equivalent to one bivalent chromosome, 

 not to 4 chromosomes as assumed by Vom Rath, since it was pro- 

 duced in the first place from one segment of the spirem, and the 4 

 parts of the ring remain up to metakinesis connected by liuin. 

 Vom Rath ('95) has described in other cases (e, g. Euchaeta) that 

 the 4 portions of a tetrad are not connected by linin, which he regards 

 as a proof of his opinion of the value of each portion of a ring as a 

 whole chromosome. But some linin connection must nevertheless be 

 present, even if not demonstrable, since otherwise it could not be ex- 

 plained how the 4 parts of a ring always keep close together. Here 

 may be cited a criticism on the part of Henking ('92) of Hertwig's 

 assumption that each separate chromatin element in the reductions 

 divisions of Ascaris has the value of a chromosome: "Es ist nämlich 

 . . . immer beobachtet worden, dass die beiden Tbeilstückc der anfangs 



