28o KELLOGG AND BELL 



deterininate vai'iation, the condition being the result of a ten- 

 dency in the species, produced by certain to us unknown ex- 

 trinsic or intrinsic influences, not selective or purposeful? To 

 our mind this color pattern variation, any more than that in 

 Hippodainm^ is not to be looked on as sufficient basis for a life 

 and death selection. Where it takes close examination by our 

 trained eyes, often aided by the simple lens, to distinguish the 

 exact character of the variation, is a beetle to lose or save its 

 life by the relation of this variation to the eyes of swooping 

 birds or darting predaceous insects (which, with toads and 

 lizards, are almost the only external selective agencies with 

 which the color pattern has to do)? Predaceous insects do not 

 see slight differences in pattern, at least at any range beyond a 

 few centimeters. They discover their prey by the visual per- 

 ception of its movements, and by their keen sense of smell. 

 Birds do not distinguish protectively colored insects ; but shall 

 soro?' individuals with two spots each a millimeter in diameter, 

 fused, held to be more safely colored than soror individuals 

 with these two spots separated by a hair's breadth? But that, it 

 seems to us, is the necessary admission if natural selection is to 

 be looked on as the only factor, outside of purely fortuitous 

 variation, in the gradually changing character of the color pat- 

 tern of soror. 



It will be of interest to note the condition of variation in cer- 

 tain lots of soror all collected in October, 1902, at various 

 localities in California other than the Stanford campus. We 

 wished to discover if soror throughout its range shows this 

 same tendency toward the production of a dominant variety of 

 the type of class D (in the Stanford series). We were able to 

 obtain several (usually small) series from a few distant locali- 

 ties, covering, however, only a small part of the range of the 

 species. We hope to add gradually to this series of geograph- 

 ically scattered lots. 



A lot of 405 individuals from Santa Rosa, California (about 

 sixty miles from Standford campus), shows variation frequencies 

 as follows : 



Class A: 228 individuals, with all spots free. Class B: 105 

 individuals with both middle pairs fused. Class C: 12 individ- 

 uals with left middle pair fused. Class D : 18 individuals with 



