THE DETERMINATION OF DOMINANCE. 



301 



These have been carried on into subsequent generations, and 

 it has been found that the twelve extracted types resulting in the 

 second hybrid generation are different from each other, even 

 though they may appear superficially alike. The general be- 

 havior and results obtained in this experiment are shown in 

 Plate VI. 



L. undecimlineata 9 X d" L. signaticollis. 

 Exp. No. H 700. 



A virgin female of L. undecimlineata, from Exp. No. 722, g. 

 I, was crossed with a male L. signaticollis, from Exp. No. 419, 

 g. XI (DC), under the following conditions: 



Max. ± 81° F. 

 Min. ± 70° F. 

 Av. 75-6° F. 



Food : Normal — uniform. 



R. H. 



Max. zb 87 per cent. 



Min. ± 69 per cent. 



Av. 77-11 per cent. 



This cross gave rise in the F^ generation to larvae all alike in 

 the first stage, which in the second stage separated into two 

 classes, yellow and white, and in the third stage remained as 

 two classes, white without spote (Whs) and yellow without spots 

 (Yls) (Whs 37 : Yls 45). The white larvae (Whs) gave rise to 

 twenty-one individuals exactly like the female type (126", 9 9), 

 and the yellow without spots (Yls) gave rise to twenty-four 

 individuals of the mid type (iic? 13 ?), a ratio of 7:8. The 

 individuals like the female parental type, when inbred, gave a 

 typical repetition of the ontogeny of the female parental stock 

 with adults like themselves, and continued to do so for four 

 consecutive generations without any indication of the male parent 

 type appearing. The mid type, when inbred, gave in the Fg gener- 

 ation typical results. The larvae of the first stage were all alike; 

 they divided into yellow and white larvae in the second stage in the 



