PAIRING OF CHROMOMERES 177 



centromeres into one central undifferentiated mass or magma, 

 with which the separate arms of the large autosomes of D. melano- 

 gaster for example, appear to be separately connected. Certain 

 regions of the X and Y chromosomes that are known to be inert 

 seem to take part in this magma, so that the small active part of the 

 Y has only recently been discovered (Prokofieva-Belgovskaya, 1935 ; 

 Frolowa, 1936). The inert parts are either less hypertrophied or 

 less closely paired than the active parts. 



In different species the behaviour of the chromosomes differs. 

 In D. simulans the chromosomes do not always pair, and the 



"■^ 



'/m 



W\ (St 



Fig. 60. — Two chromosomes before pairing is complete in D. psendo- 

 obscura, showing left-hand relational and relic coiling in the 

 unpaired segment. Only a few of the chromomere-bands are 

 shown. X 1000 (Koller, i935)- 



centromeres never fuse (Geitler, 1934 ; Bauer, 1935, a and b). It 

 is particularly significant that pairing fails most often in the shortest 

 chromosome to which a nucleolus is attached, since this indicates 

 that the difference is due to delay in coming together. Delay in 

 pairing means inhibition of pairing as we have seen in similar 

 circumstances in meiosis, and it is possible, but unlikely, that this 

 applies to the salivary gland nuclei. 



While the thicker striations often appear as uniform bands, 

 close comparison of a number of species has shown that each of the 

 finer bands consists of a group of granules which resemble the 

 chromomeres of meiotic chromosomes and may be properly so 

 described (Koltzoff, 1934). A plate of thirty-two or sixty -four of 



