STUDIES ON CHROMOSOMES 407 



fait tres etonnant, mais absolument certain;" but he very justly 

 adds, '' De ce que nous ne pouvons pas expliquer par quel mechan- 

 isme la soudure se realise dans de tels cas, il ne s'en suit rien 

 centre son existence indubitable" ('08, p. 167). However diffi- 

 cult such a mode of union may seem a priori, the preparations of 

 the Schreiners actually demonstrate double loops that are united 

 at both ends while widely separate along their middle portions — 

 shown for example in figs. 16, 17 and 18 of their paper of 1908, 

 which accurately represent the facts, as I am able to confirm from 

 examination of these identical nuclei in the original preparations. 

 We must seek to discover by observation how the conjugating 

 loops disentangle themselves from the apparent chaos of the lep- 

 totene-spireme. The chaos may however be apparent rather than 

 real. The interesting facts worked out by Janssens in regard to 

 the persistent orientation of the loops in the pre-synaptic stages 

 of Batracoseps indicate that their polarity is not lost at any time 

 between the final spermatogonial anaphases and the amphitene 

 stage, and that their free ends always converge towards the cen- 

 trosome. It seems quite possible that the way for synapsis may 

 be prepared already in a very early pre-synaptic stage, by a defi- 

 nite regrouping of the chromosomes that may take place before 

 the leptotene loops are -formed as such. It is evident that the 

 central portions of the loops are constantly shortening as the 

 peripheral portions come together (possibly as a result of the 

 progressive torsion of the latter) . It seems therefore by no means 

 a hopeless task to undertake a definite solution of the puzzle by 

 observation. 



2. The question of the reduction-division 



The history of the sex-chromosomes in Oncopeltus affords, I 

 believe, complete demonstration of the occurrence both of synap- 

 sis and of a true reduction-division in the original sense^ — i.e., 

 of the disjunction of two entire chromosomes that have previously 

 conjugated in synapsis. But, although this creates a certain 

 presumption in favor of the occurrence of a similar process in case 

 of the other chromosome-pairs, this argument must not be pushed 

 too far — indeed, there is reason to believe that in case of the ordi- 



