SEGREGATION OF HOMOLOGOUS CHROMOSOMES 485 



conceivable that this weakness might result merely in a con- 

 striction such as marks certain homologues. Figure 38 a (plate 

 11), where one of the larger tetrads in a member of form B is 

 dividing at such a point, favors this interpretation. This is the 

 only instance of this sort found in animal number 38. 



So far, in all of our work, we have never found one of the larger 

 tetrads with unequal homologues. On the other hand such 

 peculiarities occur rather frequently among the smaller ones. 

 This may indicate that the smaller chromosomes are less im- 

 portant. That they may be in part dispensed with was shown 

 by Wenrich ('16) who finds that his tetrad 'C lacks both terminal 

 granules in certain individuals; and it seems possible that 

 chromosome number 2 has been entirely eliminated from Cir- 

 cotettix. The homologues of the larger tetrads may be so 

 important that a zygote lacking in even a part of one cannot 

 develop. 



The further suggestion presents itself that irregularities in 

 mitoses such as have just been described, may be responsible 

 for the occurrence of supernumerary chromosomes. Gametes 

 containing a complete dyad plus a portion of its homologue 

 would probably be viable and one may readily imagine that the 

 extra portion might become detached in a succeeding division 

 and might then form a supernumerary. Such an element would 

 probably contain merely a reduplication of factor loci already 

 present in the two dyads from whose homologue it was derived 

 and might therefore have no influence in heredity. 3 On the other 



3 Dr. P. W. Whiting has suggested that it may be of interest to geneticists 

 to point out that chromosomes such as numbers 7 and 8 (plate 1), which are in- 

 distinguishable morphologically, behave as duplicates, corresponding to Shull's 

 theory of duplicate genes, and not as tetraploid homologues, as Muller has sug- 

 gested for Gregory's tetraploid Primulas, e.g. If A A' equals the atelomitic 

 dyads and a a' their telomitic homologues on Shull's theory A would always 

 segregate from a, and A' from a', or telomitic from atelomitic, giving a 1-2-1 

 gametic ratio. On the other hand, if they were tetraploid homologues in Mul- 

 ler's sense, A might segregate from A' and a from a' just as frequently, giving 

 a 1-4-1 gametic ratio. In other words we would sometimes find the atelomitic 

 dyad of chromosome number 7 mated with the corresponding one of number 8 

 and the same for the rod-shaped homologues. Since this condition does not 

 occur, we must conclude that their morphological resemblance does not in- 

 volve their behaving as tetraploids. 



