STUDIES ON CHROMOSOMES 411 



the third case X divides in but one spermatocyte-division (either 

 the first or the second according to the species) and in the other 

 passes undivided to one pole. It is here perfectly clear, as has 

 been urged by McClung ('01, '02) and myself ('05 c), that the 

 failure to divide in one division is due to the absence of a 

 synaptic mate; and it is thus rendered doubly certain that in case 

 of the .XT-pair but one true division (an equational) takes place, 

 the other 'division' (reductional) being merely the separation of 

 the synaptic mates. This is, I think, a conclusive demonstra- 

 tion in the case of these chromosomes of the reality (1) of the con- 

 ception of bivalence, (2) of the reduction-division in its original 

 and unmodified sense. That these conclusions are not limited 

 to the sex-chromosomes is shown by the ra-chromosomes of the 

 Hemiptera, which have no relation to sex as far as known. In this 

 case conjugation (synapsis) is usually delayed until the last pos- 

 sible moment before the first division. Their separation, which 

 immediately ensues, is again a true reduction-division of the m- 

 bivalent ; but what we here call a ' division' is obviously not prop- 

 erly such but only the disjunction of two distinct bodies that 

 have but just come into momentary contact. In this case, as 

 in that of the -XT-pair, the term reduction 'division' is a mis- 

 nomer. But one actual division takes place, the equation-divi- 

 sion. This is fully borne out by the interesting anomaly that I 

 described in a single individual of Metapodius in my sixth 'Study,' 

 consisting in the presence of three m-chromosomes instead of 

 two. Here all three uniformly couple to form a triad element in 

 the first division, which immediately breaks up into its compo- 

 nents, of which one passes to one pole and two to the other. In 

 the second division all three divide equationally, so that half 

 the spermatids receive but one w-chromosome half two. This 

 is, of course, exactly in accordance with expectation; and it is a 

 remarkable fact that the two m-chromosomes that are present in 

 half the secondary spermatocytes do not disjoin but divide equa- 

 tionally, as they should. 



The case of the autosomes is different, owing to the intimate 

 union and possible fusion that follows synapsis; and it seems prob- 

 able that the reduction-division must here be regarded in a differ- 



