408 EDMUND B. WILSON 



nary chromosomes ('autosomes') the process maybe different in 

 some important respects from that seen in the sex-chromosomes; 

 and we must not lose sight of the wide difference of behavior in 

 other respects that differentiates the latter from the former. It is 

 possible that the sharply marked process of conjugation and 

 disjunction characteristic of the sex-chromosomes and m-chro- 

 mosomes may be correlated with their specific functional relations. 

 The case for the autosomes must therefore rest upon their direct 

 study, and a reduction-division can only be fully established by 

 tracing the bivalents as double bodies or 'gemini' through every 

 stage from the time of their conjugation to that of their disjunc- 

 tion. 



Whatever view of synapsis be adopted, this is a difficult task. 

 If we take the view that the chromosomes are arranged in linear 

 series in a spireme-thread which breaks into bivalent segments 

 each consisting of two chromosomes in telosynaptic union there is 

 no guarantee, as far as I can see, that the latter ultimately sepa- 

 rate at the synaptic point. If we accept parasynaptic conjuga- 

 tion the difficulty is of a different kind, namely, the extremely 

 close union of the conjugants side by side, which as nearly all 

 observers are agreed, follows upon synapsis. The most that has 

 been asserted by these observers has been that evidence of longi- 

 tudinal duality can always be seen in some of the bivalents at 

 every stage. Without reviewing all these cases, I will only recall 

 that in the case of Batracoseps, for example, Janssens says, 

 "Pendant le long stade du bouquet, les anses sont simples et par 

 aucune methode cytologique nous ne parvenons a y reconnaitre 

 la moindre trace de dualite" ('05, p. 401). In case of Tomopteris 

 the Schreiners admit that at a period shortly following synapsis 

 no longitudinal division can be seen in the pachytene-threads; 

 and these authors are compelled to fall back upon indirect evi- 

 dence in support of their conclusion that the duality is not really 

 lost. Again, in Montgomery's recent work on Euschistus ('11) 

 he expresses the conviction that there is no valid evidence of any 

 actual fusion between the conjugants; yet in point of fact states 

 that after the completion of synapsis ''The autosomal loops 

 (bivalents), are in one-half the normal number and, for the most 



