THE DETERMINATION OF DOMINANCE. 3O3 



In this experiment in the Fj generation three distinct types 

 appeared, one like the male parental type, one like the female 

 parental type, and a mid type. Each of the derived types was 

 inbred for four consecutive generations and continued to 

 breed true, neither one giving any indication of the other 

 stock. Of the mid types several pairs were mated and gave 

 the same results as those obtained from the mid types in 

 Exp. No. H 700, H 701, etc. Thus, in Pair A, the larvae in the 

 first stage were all alike; in the second stage they split into 

 white and yellow in a ratio of wh. 55 : yl. 45. The white split 

 into white without spots (Whs 27), and white with spots (WhS 23) ; 

 and the yellow split into yellow without spots (Yls 14), and yellow 

 with spots (YIS 25). The four classes of larvae gave adults of the 

 Fj generation, each producing three classes sharply demarked 

 one from the other, as in the previous experiments. Plate 

 VIII. shows the behavior and results obtained in this experi- 

 ment. 



In this series of experiments the behavior in the first and second 

 hybrid generations indicates a variability in the reactions which 

 take place between the combinations brought about in the germ 

 plasm at fertilization. As stated in the introduction, I have 

 not been able to show the existence of additional characters 

 latent or recessive, which might bring about the variability in 

 behavior observed, and the problem is to account for this vari- 

 ability. 



In the absence of evidence to the contrary, the characters seen 

 may safely be assumed to represent the characters involved, and 

 it is perfectly clear, in crosses between L. signaticollis and L. 

 diver sa, and between L. signaticollis and L. undecimlineata, 

 that when identical materials were placed under dissimilar con- 

 ditions, dissimilar types of behavior resulted without the pro- 

 duction of anything new or unusual in the way of attributes in 

 the resulting progeny. In the case of Exp. No. H 409, the be- 

 havior there strongly indicated that either one or the other of 

 the two parents was heterozygous in character, and such an inter- 



