interruptedly, in a higher temperature than 95° and in very 

 humid air, they will produce many weakly and not well de- 

 veloped butterflies, or will die; when, however, chrysalides 

 exposed to as high as 115° temperature for about two hours 

 daily, the butterflies will emerge, then, sometimes about two 

 days later than the regular time of five days from the chrys- 

 alides bred in 90° temperature, and will then produce mostly 

 some rare aberration in color or markings. 



Lot (a) in the second generation produced from 100 

 caterpillars, 32 females and 35 males; the wilt disease got hold 

 of a good many caterpillars ; however, we saved enough to 

 give us a fairly good bunch of butterfles to continue this ex- 

 periment and show the percentage of change in color and 

 markings. The males had the normal color of our local form 

 except dark in ground color; from the 32 females, three of 

 them had the large ocelli of the hindwings purple in color, and 

 these three females had also the most purplish color on the 

 forewing; with this butterfly the female is always progres- 

 sive in development of new color and markings, only in later 

 generations when the percentage of the change becomes 

 higher, then the males will get these changes also. The re- 

 duced size of both ocelli of the hindwing was, in this genera- 

 tion, still more reduced in many specimens, in males as well 

 as females. 



Lot (b), also the second generation, developed 18 females 

 and 26 males and the percentage of change in color and size of 

 the ocelli was about the same as in lot (a). We retained some 

 of the purple females for breeding, alone in a cage, and suc- 

 ceeded in obtaining enough eggs in both separate lots of but- 

 terflies, with purple color, to secure our experiments; how- 

 ever, we had, at that time, not enough space and time to keep 

 the more yellow ones, which were colored like our local form, 

 and we were more interested in this new purple color, and 

 wished to improve this by selection. 



In our third generation we obtained in lot (a) 42 females 

 and 35 males, and in this generation we had about 50 per 

 cent with purple ocelli on the hindwings, and also the males 

 had about the same percentage of purple ocelli. The lot (b) 

 developed 58 females and 55 miles; here the percentage of 

 butterfles with purple ocelli was also about 50 per cent. Many 

 more had the innerfield of these ocelli not quite full purple, 

 there was a small space of lighter color left in the upper circle ; 

 it is for this reason very difficult to give the exact per centage 

 of the change of color. In this lot (b) we got, for the first 

 time, three females with a very small spot directly under the 

 ocelli of the forewings. just the same way that the newly 

 developed appendix of these ocelli started to develop in our 

 old line of experiments described in Pomona College Journal 

 of Entomology. However, this small start, in later bred gen- 



32 



