purple colored males and females— because we did not need 

 to divide them and all came out more at the same time. 



As we had many good purple colored females and males 

 we selected again the very best ones, and in the fourth gen- 

 eration we got 64 females, all of these had the purple ocellH 

 and in manv the forewings purple in color; and 77 males, 35 

 of these had the ocelli piirple colored — the number of those 

 which had a purple color on the forewings is difficult to say, 

 as that color, there, is very varied in extent. This increasing 

 of the percentage of purple colored, and the extension of this 

 new color almost over the whole forewing, was without doubt 

 the result of our selections, and we continued this line hi the 

 same way; however, the butterfles for a few months in the 

 spring of 1912 had to be bred in a cool tent without an in- 

 cubator, before the glass house was finished, so the time of 

 breeding a generation became extended ; this had, however, no 

 bad effect on our purple color improvement, and in the ninth 

 generation we had the females, as well as the males, all with 

 purple colored ocelli on the hindwings. 



In the eleventh generation we had trouble with some very 

 small parasites which infested eighteen of our chrysalides", 

 and before that time the ants carried off many eggs over night 

 before their presence was discovered, so that in this genera- 

 tion we got only a very few females, and these emerged so 

 tion we got only a very few females, and these emerged so 

 weak and died before pairing. The only course left open to 

 us to get some good eggs of the purple colored females was 

 to pair these females with males of our other stock from 

 which we had now the thirty-first generation of butterflies. 

 We selected for this purpose six of the best developed males 

 and these males had the ocelli of the forewings greatly en- 

 larged and had the black appendage thereon as developed in 

 the" line of expriments bred in humid air as described in the 

 Pomona College Journal ; already on the next day we saw a 

 pair in copula and put these carefully separate in a wire cage 

 with a pot-plant of Linaria cymbalaria whereon the females 

 deposited their eggs. 



Here we come to the most interesting part of our ex- 

 periments from a scientific point of view ; we could find here 

 what the law of heredity would produce if any such law 

 exists. We had two main separate characters to observe; in 

 the male we had the large black ocelli with an appendix on 

 the forewing, this appendix is developed by breeding in 90° 

 temperature" and humidity and by inbreeding, without cross- 

 ing, and selecting the specimens with the largest appendices 

 v;e developed these appendices to a large size; and there 

 formed in many, mostly females, a little light-colored center, 

 we had then double ocelli (connected together), the surround- 

 ing white field remained in extent the same as in our local 



34 



