232 



X. M. STEVENS. 



ing table shows the per cent, of supernumeraries in the two 

 species different years and in different localities. 



As I had never seen any signs of degeneration of the super- 

 numeraries, the natural interpretation of their infrequency at 

 La Jolla would seem to be either that they had originated here 

 more recently, or that they had originally appeared in fewer 

 individuals in this locality. 



The behavior of supernumeraries in all cases where they have 

 been shown to occur at once classes them with the hetero- 

 chromosomes, and in Metapodius Wilson ('09) has shown that 

 they have probably originated in an irregular second maturation 

 mitosis in which both idiochromosomes went to one pole of the 

 spindle instead of separating. He therefore regards the super- 

 numeraries in Metapodius as duplicates of the smaller idiochro- 

 mosome. In 1908 I suggested that there might be two varieties of 

 Diabrotica soror and also of D. 12-punctala, one having only the 

 odd heterochromosome and the other an unequal pair, and that 

 hybridization might have given rise to the supernumeraries 

 with their peculiar behavior, dividing sometimes in one some- 

 times in the other maturation mitosis. I have, however, been 

 able to find no evidence in favor of this view. In 1910 I studied 

 carefully the testesof many individuals where no supernumeraries 

 were present, seeking some clue to the origin of these chromo- 

 somes. 



As a rule the odd chromosome X appears near one pole of the 

 spindle in the metaphase of the first maturation mitosis, but I 

 had always noticed that occasionally X is in or near the equa- 

 torial plate, and in some individuals this is quite common. At 

 La Jolla I found two spindles in which X was between the 

 daughter plates in the anaphase, and stretched out lengthwise 

 (Figs, i and 2). In one of these cases (Fig. 2) X was split so 



