206 



The Mechanism of Evolution in Leptinotaesa 



so on. They can be either somatic or germinal, but no trace of knowledge of 

 this either one way or the other is found from the biometric analysis. From any 

 such body of data any one of many " conclusions " can be drawn, but in all the 

 conclusion is an assumption made to interpret the array of mathematical values 

 which the biometric method has obtained. The method shows a condition only, 



Fig. 24. — Showing directions of variation found in spots c, x, d, 

 and e, in L. panamensis. 



and none too accurately, because it can not distinguish germinal and ontogenetic 

 differences, or any other kind of heterogeneity present. It is, however, a fairly 

 good crude method in obtaining preliminary data of conditions in a population. 

 Examination of the directions of diversity found in L. panamensis (fig. 24) 

 shows differences in the areas, but the general result is clearly a delimited vari- 

 ation. In restricted localities the population shows remarkable delimited move- 

 ment of the population as a whole in certain directions. To what agencies these 

 movements and responses are due is not determined by the methods of the census ; 

 an endless array of possible causes for the observed effects could be suggested and 

 plausibilities put together to no purpose whatever. 



Fig. 25. — Sliowing directions of varia- 

 tion, especially in spots c, h, and e, in 

 L. signaticollis, and some results which 

 these directions of variation may produce 

 in the production of pattern differences. 



In L. signaticollis the same spot (c) occurs and shows some of the same direc- 

 tions of variation as in L. undedmlineata, but in area it much exceeds the first 

 species. In figure 25 are shown the variations thus far discovered, together with 

 the results produced in the pattern. In this species spot c is most commonly 

 present as a round area of variable size, is never absent, with the observed 



