STUDIES ON CHROMOSOMES 353 



ing fibers are much less conspicuous, and often appear single. 

 Up to the middle anaphases the sex-chromosomes remain always 

 separate (figs. 18, 19). In the later anaphases all the chromo- 

 somes draw more closely together, and often come more or less 

 into contact, though without losing their original grouping; 

 but in case of the autosomes the contact is but casual and tem- 

 porary, while the sex-chromosomes become definitely attached 

 to each other to form a dumb-bell. shaped body at the center of 

 the group (figs. 19, 21). By this process the total number of sepa- 

 rate chromatin-elements is reduced from nine to eight (the hap- 

 loid number). In Lygaeus the process takes place in exactly the 

 same way and may be seen with equal clearness, both in polar 

 views and in side-views. In both species polar views of the final 

 anaphases show that the chromosomes, save for their more 

 crowded condition, have retained the same grouping as in the 

 metaphase (figs. 24 to 29), the XT-bivalent being at the center, 

 surrounded by the other chromosomes in the form of a ring. 

 These facts are seen so clearly and in so great a number of cases 

 as to remove every doubt that in these species the conjugation 

 between X and Y regularly takes place at the poles of the spindle 

 before the first maturation-division has been completed. 4 ' 



In Lygaeus the XF-bivalent thus formed is readily distinguish- 

 able by the inequality of its two components (figs. 28, 29, 46). 

 In Oncopeltus it is often not thus marked but its identity is no 

 less certainly revealed in another way. As the figures show, the 

 autosomes still show but slight indication of a transverse constric- 

 tion, and can hardly be described as dumb-bell shaped until a 

 later period. The XF-bivalent, on the other hand, is invariably 

 deeply constricted, so as to have a conspicuously dumb-bell shape, 

 and it often still appears like two chromosomes that are merely in 

 contact. This characteristic difference persists throughout the 

 entire interkinesis, and is still perfectly obvious in the ensuing 

 metaphase of the second divi ion. 



4 I described and figured this process in the case of Coenus in my first 'Study' 

 ('05 b) but did not recognize its constancy. I now incline to think that it will 

 be found to occur in the same way in many other forms. 



