CHROMOSOME STUDIES ON THE DIPTERA. 257 



ning out proceeds the threads become entangled and lose much of 

 their staining capacity, making this the most difficult stage to 

 analyze. In fact, it is impossible to trace all of the chromosomes 

 in any one nucleus. It is significant, however, that as soon as the 

 aggregates have elongated two kinds of threads may be observed : 

 single and double ; and occasionally in favorable nuclei the two 

 members of a double one may be seen to diverge in the form of a 

 Y, or may even form a loop essentially like those of later stages. 

 The most convincing evidence as to the nature of events at this 

 time is obtained from an examination of the threads attached to 

 the "dense body." As has been noted above, this body in later 

 stages is attached to the two single threads making the loop in one 

 of the large chromosome pairs. By following this structure 

 through the early stages, then, it should be possible to determine 

 whether or not the loop is present here also. A careful study of 

 this feature has convinced me that the loop is normally present 

 throughout the early stages from the time the chromosomes first 

 elongate. In some cases the entire structure may be seen, as 

 shown in Figs. 4 and 6, and in others it is evident that the threads 

 running out from the dense body are single, not double. The fact 

 that single threads and occasionally loops (Fig. 5) may also be 

 seen in other parts of the nucleus makes it seem almost certain 

 that the same conclusion applies to the other large chromosomes. 

 It is practically impossible to differentiate the smallest chromosome 

 pair from the others in the early stages, hence I have been unable 

 to determine whether or not it possesses a loop at this time. A 

 little later the chromosome pairs move to the periphery of the 

 nucleus and become separated sufficiently to permit of individual 

 analysis which brings them into the stage with which our descrip- 

 tion began (Figs. 7 to 9). 



It is possible, of course, to assume that the chromosome pairs 

 do not behave synchronously in the very early stages, and to 

 imagine them undergoing, one at a time, a complete synapsis fol- 

 lowed immediately by a partial opening out into loops. This 

 would account for the constant presence of both single and double 

 threads in the nucleus. Such an explanation seems improbable, 

 however, from analogy with other forms, and especially in view 

 of the evidence furnished by the chromosome pair attached to the 



