WENRICH: spermatogenesis of PHRYNOTETTIX MAGNUS. 71 



while the other has none. It will be noticed (fig. 65, b, c) that the 

 details of the two components are quite homologous up to the large 

 distal granule, and that the point of attachment of this large granule 

 seems to correspond to the distal end of the smaller component. These 

 considerations would lead us to suppose that here, too, as in the un- 

 equal type of B, the difference between the members of the pair may be 

 due to the loss by one of them of a definite part possessed by the other. 

 In this case, however, no such equal pair has been found as occurs in 

 B when both members possess the part in question. The side-by-side 

 association of the members of this pair is as evident as it was for A 

 and B and the relations of the two longitudinal splits are the same. 



In regard to the mode of distribution of tetrad parts in the first 

 maturation division, however, we meet in this case a curious exception 

 to the general rule. This pair divides equationally, as shown in 

 figure 65, h-j; but it sometimes divides reductionally as shown at 

 k-m, same figure. From casual inspection it appeared that the divi- 

 sion occurred as frequently in the one manner as in the other. But 

 in order to test the relative frequencies of the two methods, 928 cases 

 chosen at random were counted and it was found that of this number 

 472, or 50.8%, were in process of reductional division, while 456, or 

 49.2%, were dividing equationally. It would seem from these counts 

 that the method by which the tetrad divides is a matter of mere 

 chance. This is the more apparent when we take into consideration 

 the fact, brought out by extended observations, that the two methods 

 occurred side-by-side in the same cysts. It may be that the shape or 

 position of the tetrad when it is first brought under the influence of 

 the mitotic spindle determines the mode of division. 



The fact that this unequal pair divided in the first division reduc- 

 tionally a part of the time made it possible to study the distribution 

 of the two conjugants with reference to the accessory chromosome, 

 which goes to one pole undivided. It was soon found that either 

 member of the pair could accompany the accessory into the second- 

 ary spermatocytes. Consequently counts were made to determine 

 whether the two kinds of distribution occurred with anything like 

 equal frequency. Out of 421 cases counted at random 216, or 51.3%, 

 were found to show the larger member going to the same pole as the 

 accessory (Plate 10, fig. 121, C), while in 205, or 48.7%, of the cases 

 the srnaller member was going with the accessory to the same pole (fig. 

 120, C) . These results seem to furnish a good example of chance dis- 

 tribution of clu-omosomes at maturation. 



The behavior of these three selected clu-omosome-pairs, as described 



