168 PROCEEDINGS OF THE ACADEMY OF [Feb., 



sometimes the opposite ends also are joined, elongated rings resulting 

 instead of other forms. The space between the two arms of such a 

 bivalent chromosome does not represent a longitudinal split, but the 

 area between two entire bivalent chromosomes. The true longitudinal 

 split becomes apparent as a cleavage of the small chromatin masses 

 forming each univalent chromosome, and is a line of chromatin sepa- 

 ration within each univalent chromosome; sometimes it cannot be 

 seen, which is due to the chromosome being viewed from the edge. 

 This is the first and only longitudinal split of the chromosomes from the 

 time of first formation of the spermatocytes up to the stage of the 

 spermatid. No trace of a longitudinal split can longer be seen in the 

 heterochromosome, which now has in most cases the form of two nearly 

 parallel rods, produced by the bending at the middle of the original 

 straight one. This synapsis stage corresponds to the similarly named 

 stage of the Hemiptera in the close massing of the chromosomes near 

 one pole of the nucleus, but we have seen that the conjugation of the 

 chromosomes becomes affected at an earlier period, that of the figs. 17 

 and 18. There is evidence that in Syrbula, as I have shown to be the 

 case in Peripatus, there is a continuous linin spirem during the synapsis 

 stage ; but at no period of the first spermatocyte is there a continuous 

 ^chromatin spirem. The splitting of the chromatin globules does not 

 occur simultaneously for all composing a chromosome, but rather suc- 

 cessively; and each globule or granule is a mass of demonstrable 

 smaller microsomes. Hence there is no proof that each smallest visible 

 microsome divides into two during the longitudinal splitting of a 

 chromosome. 



Then comes a post-synapsis stage in which the chromosomes are no 

 longer densely grouped, and when the longitudinal split is very clear. 

 In figs. 23 and 24, illustrating this stage, only three and four respectively 

 of the nine bivalent chromosomes are drawn ; and the point x on each 

 marks the linin band connecting every two univalent chromosomes. 

 Very rarely does the longitudinal split become wider than shown in 

 these figures, but sometimes it wddens as much as is shown in the largest 

 chromosome of fig. 25. This was the maximum extent of separation 

 seen of the halves of a split univalent chromosome, and from this stage 

 through the following this split narrows gradually. 



Unlike most of the Hemiptera no rest stage follows, but the sperma- 

 tocytes enter immediately upon the prophases of the first maturation 

 mitosis; successive steps of this process are shown in figs. 27-31. The 

 nucleus enlarges, the chromosomes lie close to its wall, the delicate 

 linin fibres change their character and break each into a row of minute 



