PI.ACI: VARIx\TlON. 103 



ation. We must use care and discretion, for biometric expressions are only 

 a little more precise, only a little less liable to error, than qualitative expres- 

 sions of biological facts. 



ShuU (1904) well remarks: "To some these results may seem to preclude 

 the possibility of deriving anything of further value from quantitative studies 

 of variation, while to others many new problems of great interest and impor- 

 tance will be suggested." 



As yet we know very little of the laws which govern this phenomenon, or, 

 indeed, of the part which it may play in evolution. In my own studies, data 

 have been gathered which point to the importance of this phenomenon, and 

 may indicate some of the laws by which it is governed. 



The data in the foregoing tables show conclusively that place variation 

 affects every part of the animal, and that not all species are alike in their 

 variability. For example, decemlincata and midtitcsniata are highly variable 

 in this respect, while signaticollis, undecimUneata, and dilecta are among the 

 least variable of all the species. If we compare the place variability of these 

 species with the individual variation of the entire population it is at once 

 apparent that the species with a high individual variability also present a high 

 place variability. Moreover, the highly variable species decemlineata and 

 multitceniata are also widely distributed, while those of low variability are 

 more restricted in their habitat, as signaticollis and dilecta. Hence it follows 

 that wide-ranging species, which are more variable than those of restricted 

 range, are also more variable in one place than are those of narrower distri- 

 bution. This wide distribution, high individual, and high place variation all 

 indicate in the species a high degree of sensibility to stimuli, and an adapta- 

 bility and lack of conservative tendencies in the population. 



This difference between species we may examine to advantage in the speci- 

 mens of vudtitccniata obtained at Guadalupe, Federal District, Mexico, and 

 those of signaticollis from Cuernavaca, Mexico. One lies on the east of the 

 Sierra de las Cruces, the other on the west. One lives in a habitat where the 

 yearly rainfall is from 16 to 20 inches, the other where it is from 50 to 60 

 inches. In the first generation of midtitccniata in 1903 the polygon of place 

 condition was fairly normal, while that of the second generation was strongly 

 skewed prophetically. During the growth of the first generation conditions 

 were average ; during that of the second the rainfall was far in excess of the 

 average. We may attribute this place variation to the increased rainfall, 

 although it is possible that it was only the indirect cause. During this same 

 season signaticollis, although subjected to the same variation in precipitation, 

 showed no increased place variation. In the following year (1904), again, 

 under average conditions, the polygon of the first generation in miiltitceniata 

 showed a historic skewness and that of the second a normal condition, while 

 signaticollis remained invariable. Of the two species one is immensely more 

 susceptible to changes in environment, undergoing modification very readily. 



8-T 



