STUDIES ON CHROMGSOMES 99 
but this question is clearly not yet ready for discussion. How- 
ever such associations have arisen, the result is equally appli- 
cable to the explanation of sex-limited heredity. 
(d) Secondary sexual characters. Castle (09) has offered the 
interesting suggestion that the free Y-chromosome may be re- 
sponsible for the determination of secondary sexual characters 
in the male. Though I have criticized this view (’09c) I now 
believe it may be true for certain cases. It is obviously excluded 
when the Y-chromosome is missing; and since nearly related 
species—in Metapodius even different individuals of the same spe- 
cies—show the same or similar secondary male characters whether 
this chromosome be present or absent, it seems probable that 
these characters are in general determined in some other way. 
But if, as I have suggested, sex-limited heredity may arise through 
a modification of the Y-constituents of the X-element, it follows 
that the YY-pair thereby becomes heterozygous. In such case, 
the free Y-chromosome, being confined to the male line, should 
continue to represent characters that are no longer present in 
the female, and hence would be indistinguishable from secondary 
male characters otherwise determined. It has further become 
evident (as is indicated below) that the chromosome-groups are 
so plastic that their specific composition may vary widely from 
species to species. It may very well be, therefore, that Castle’s 
suggestion may apply to some forms. 
6G. Modes in which the chromosome-number may change 
The constant and characteristic duality of the ‘d-chromosome’ 
in the second division suggests a series of questions regarding the 
mode in which the chromosome-number may change that have 
an important bearing on those already considered. The appear- 
ance of this chromosome must suggest to any observer that it is 
a compound body, consisting of two closely united components 
that are invariably associated in a definite way; but it is especially 
noteworthy that its duality does not certainly appear before the 
last division. This case must be added to the steadily increasing 
evidence that chromosomes which appear single and homoge- 
