451] COLORS OF TIGER BEETLES— SHELFORD 57 



111 general such correlation is closest in relation to rainfall, but this 

 correlation is not so good as one would expect (Fig. 470 a, PI. XXIV). 

 This is perhaps to be expected in the ease of species which belong to 

 local conditions which is true of most of the species of Cicindela. This 

 subject has been discussed in some detail ( Shelf ord, 1911). Here it 

 was shown that species which were distributed in a major climatic 

 habitat had a distribution correlated with the distribution of vegeta- 

 tion, which in turn is correlated with the distribution of climatic con- 

 ditions. I showed further that species such as C. tranqucharica trav- 

 ersed almost the entire continent without much variation by virtue of 

 living in moist soil, due either to climatic moisture or to local stream 

 moisture or lake-shore moisture. C. scutcllaris, C. purpurea and most 

 of the other species noted are found in some special kind of soil such 

 as sand containing a little humus (Shelford, 1911, 1913b) or steep clay- 

 banks or some other restricted situation. Taking C. scutellaris for 

 example, this species being found in well drained or dry sand contain- 

 ing a little humus and bound by scattered vegetation throughout its 

 range, it is to be expected that the distribution of the species will be 

 correlated with some sort of measured soil conditions such as soil tem- 

 perature, soil wilting coefficient, or the like ; but no such conditions 

 have been recorded or mapped. There is some evidence of soil effects 

 in this species (see Fig. 558, PL XXIX). Some specimens from the 

 verj' coarse sands resiilting from the weathering of St. Peter's sand 

 stone, near Utica (Starved Rock), Illinois, are purple. No purple 

 forms have been taken elsewhere. Two specimens from sandy clay 

 (Suman, Indiana) had an unusual silky appearance. When soil tem- 

 perature work under way is published, I shall attempt to make use of 

 the extensive records which have been accumulated for the purpose of 

 working out correlation between conditions and color and pattern va- 

 rieties. Conditions associated with altitude influence color in some 

 cases, but there is no unity of conditions or colors. 



GEOGBitPHlC CENTER OP THE GROUP ON THE BASIS OP PATTERNS 



The usual criteria for the center of distribution (Adams, 1902) 

 indicate that the Oriental region or at most the Oriental and Ethio- 

 pian regions (shores of the Indian Ocean) are the geographic center 

 or center of distribution of the group. The first evidence presented 

 which indicates this is found in table I, in which eleven groups of 

 species are shown to occur in the Oriental region and in other regions, 

 while not more than six occur in any one other region and at the same 

 time in still others. 



Patterns are divisible into three great gro\ips: first those without 

 the spots at the base and along the inner border of the elytron shown 

 to the left of the bottom of figure 580; these patterns represent the 



