STUDIES ON CHROMOSOMES 379 



one is much longer than any of the others and evidently corre- 

 sponds to the large pair of chromosomes that are a constant feature 

 of the diploid groups in this species (photos. 33, 34) . From these 

 facts it is clear that each of the double threads is a bivalent, which 

 corresponds to a pair of spermatogonial chromosomes, and that 

 synapsis must have taken place during the period of contraction or 

 synizesis, as many other observers have concluded in both ani- 

 mals and plants."* In what manner synapsis takes place, and 

 whether the longitudinal halves of the diplotene-threads represent 

 the original conjugants in side by side union are questions that 

 here present very great practical difficulties to direct observation. 



Stage g. The diffuse or confused period. The diplotene-nuclei 

 now undergo a remarkable transformation, characteristic of many 

 Hemiptera, in the course of which the double threads as such 

 completely disappear from view, giving rise to a diffuse, lightly 

 staining net-like stage in which the boundaries of the individual 

 bivalents are indistinguishable (figs. 66, 67, 97, photos. 11, 14, 

 16). In Oncopeltus and Lygaeus I have found it impossible to 

 arrive at any clear notion as to the exact nature of this transfor- 

 mation. In Hermann preparations of Largus all the transitional 

 stages are shown with great apparent clearness, yet even here it is 

 difficult to reach a certain result. This question — one of the most 

 important involved in the maturation-process — will, I believe, 

 repay careful study in smear-preparations, which I hope to under- 

 take hereafter with more adequate material than I have at 

 present. 



As the process begins, the threads become less regular and at 

 the same time longer and thinner, while the longitudinal cleft is 

 still more evident. A little later the two halves of the double 

 threads become more or less contorted, more granular and irregu- 

 lar in structure, and at the same time are often seen to be separat- 

 ing in an irregular way (figs. 101-103). By the continuation of 

 this change the double threads as such disappear from view, 

 and the whole nucleus is traversed by rather thin, irregular, con- 



1" In Syromastes Gross ('04) believed that the somatic or diploid number of 

 double threads could be counted in the post-synizesis s-tages, and that synapsis 

 took place at a later period. 



