437] COLORS OF TIGER BEETLES—SHELFORD 43 



nines had been cut off the pine belt in Gary wliere my specimens were 

 collected. It is about as open as the cottonwood belt where evaporation 

 from the porous cup atmometer is about twice that of the oak dunes 

 in which the ]Miller specimens were collected. The soil temperature 

 goes very high in the Gary locality. 



Distance below Temperature in degrees G 



surface Air 36 C 



1 1-4 cm. 47 



3-4 cm. 38 



8-9 em. 35 



10-11 cm. 33 



12-13 cm. 32 



17-18 cm. 30 



These forms pupate at a depth of 15 cm. and thvis at a tempera- 

 ture of 31°C. on the warmest days. The temperatures in the shade in 

 oak covered sand dunes are much lower being about 27°C. under the 

 same conditions. 



Plate XXV shows the division of the various subspecies of C. pifr- 

 purca into classes. Here the primary division of the group, shown in 

 the immaculate form in the center of the group which is ver^v rare, is 

 an habitudinal one — those at tlie left are the patterns of a series of races 

 which inhabit level ground usually among scattered vegetation. To tlie 

 right are tliose that occupy steep banks, particularly clay banks. Classes 

 a and b, cimarrona, and t, 10 guttata, do not appear to be so differentiated 

 and accordingly the graph perhaps should have been revei-sed with the 

 generalized patterns in the center, though further investigation would 

 be necessary to determine this. The present arrangement is based on 

 resemblances between the two, cimarrona and those at the left, and C. 

 10 guttata and those at the right. The distribution of the two gi-oups 

 shown at the right and the left of the center are shown in figures 471(7 

 and 472. 



If one notes the localities represented by the graplis showing the 

 distribution of classes, it is evident that there is no striking difference 

 in tlie distribution of classes in Puget Sound, Massachusetts, and Color- 

 ado. The modal ela.ss for Manitoba, Topeka, and Chicago, is tlie same. 

 This goes to indicate that the main line of separation is habitudinal 

 rather than geographic. 



Similar relation could be shown for other species. The main differ- 

 ences in patterns are primarily associated either with different local- 

 ities usually separated geograpliically, or witli differences in habitat 

 preference. 



The figures on plate XXVI II (Figs. 473 to 536) are arranged 



