576 CLARENCE E. McCLUNG 



bination it is apparent that there could be no union between a 

 spermatozoon of class 2 and an egg of classes 3 or 4, for instance, 

 without synapsis of non-homologuos chromosomes. Still using 

 size as a test for homology, we find no such indication of non- 

 homologous chromosome unions in this species. 



This difficulty entirely disappears, however, if we assume 

 that the formation of multiples takes place at fertilization. 

 Such a view is in strict accord with our knowledge of the im- 

 portance of this step in the production of differences between 

 individuals and of constancy of character exhibited by each of 

 these variants throughout its subsequent history. Also there 

 is offered a ready explanation of the formation of multiples, 

 which would, at the same time, explain the variety of elements 

 sometimes involved, as in this case; for if there be a period at 

 which chromosomes are brought into intimate relations in a 

 linear series, multiples would be merely persistent cases of such 

 associations. The contrary might also be true, the multiples 

 showing a tendency to overcome forces keeping the chromosomes 

 apart, but, in view of the variety of elements involved in H. 

 viridis, this would seem less reasonable than the first explana- 

 tion. There are also other considerations, into which I have not 

 the time to enter here, which incline me to believe that mul- 

 tiples indicate the operation of a common principle of chromo- 

 some association. The alternative for this explanation of the 

 conditions in H. viridis, by reason of associations established at 

 the time of the union of the spermatozoon and egg complexes, 

 is the assumption that, at some time, each of the classes repre- 

 sented in plate 2 became established in some way and has since 

 maintained itself by the segregation of its elements and their 

 chance recombinations, with the elimination of all the unrepre- 

 sented conditions. By this we are still left without an expla- 

 nation of the first instance of each class and must confront a 

 selective mortality of considerable magnitude, for which there is 

 no evidence. 



Such an explanation, as that to which I incline, for the condi- 

 tions of the multiple chromosome in H. viridis does not make 

 impossible another for the case of Trimerotropis, and it may 



