Rbactions and Pkoducts in Interspecific Ceosses 109 



in type. In Pj the signaticolUs type bred true in all instances, the mid-type 

 gave the customary F^ array in nearly perfect proportions. This obTiously 

 heterozygous line shows nothing of any interest or irregularity, no matter how 

 far it is carried and regardless of the conditions of the medium within limits. 



In the signaticolUs line, however, I learned to recognize differences in the 

 population, delicate and not capable of determination except by measurement. 

 In general appearances the population was uniformly signaticolUs type, bred 

 true, crossed with pure signaticolUs stock, gave only signaticolUs types in the 

 progeny, so that one might easily pass the entire population as extracted signa- 

 ticolUs. I found that with close attention I could separate three form-types 

 from this Fj signaticolUs population. These I first recognized as present in an 

 Pa population that was signaticolUs to all outward appearance, of form-types 

 that had the suggestion of diversa, of signaticolUs, and of an intermediate type 

 with intermediate conditions in body, form, and aspect. 



In 1907 I had measured three random fraternities of this Fa signaticolUs 

 type to obtain the form-index. The result of the determinations showed in the 

 iPj signaticolUs type population a trimodal polygon, one mode of which was 

 upon that of the diversa parent, the other that of signaticolUs, the third variable 

 and intermediate, that of typical heterozygotes. There was not the least doubt 

 of the presence of the three modes in the Pj fraternities measured, which were 

 supposedly homogeneous. In figure 7 I have shown the results of this determi- 

 nation for the fraternities and the determinations for the two parent species. 

 I also made like determinations of the form-index in a normal heterozygous 

 line and which showed that the conditions in the intermediate groups and in 

 the recognized mid-type was essentially the same. The three fraternities, each 

 the progeny of a single pair of parents, were large — 171, 180, and 204— and 

 each showed the same arrangement into a trimodal curve as is shown in the 

 summation of them shown in figure 7. I was thus able also to detect these 

 modal differences of the form-index and so was able to select mates from the 

 two extreme modes that corresponded with the two parent species modes for 

 breeding, and found that the response was immediate to produce in the next 

 generation complete isolation of the two groups, with no more intergrades in 

 form-index than are present in the normal species. These lines were kept 

 breeding on through 1908 and 1909 without change or transgression of the 

 limits of the condition first determined in 1907. 



In the latter part of 1908 it was observed that in fraternities belonging to the 

 lower end of the range there was not uncommonly a development of faint, 

 diffuse, irregular pigmentation along the rows of punetation on the elytra 

 (plate 8, figure 11), in position and relations in which pigment is never found 

 in the pure signaticolUs stock, and has not been made to develop by the extension 

 of the pigment in the pits. It was further found that these conditions in the 

 race were permanent and could by mating likes be intensified to the extent that 

 the elytron presented often a faint brownish tinge extending out on all sides 

 from the system of pits in the surface of the elytron. 



These two experiences lead to the hypothesis that the F.^^ signaticolUs race 

 was not signaticolUs at all, but a masked heterozygous race that had the diversa 

 gametic system present but in combination, such that it was not able to produce 

 the normal reaction in the complex to give the usual heterozygote form. This 



