THE DETERMINATION OF DOMINANCE. 305 



In this preliminary account of the experiments, I have given 

 only the crosses between the female L. signaticollis and the males 

 of L. diversa and the female L. undecimlineata and male L. sig- 

 naticollis. The reciprocal crosses have been made and confirm in 

 every way the results obtained in the crosses given, but in 

 this paper the object is not the investigation of variability of 

 behavior in reciprocal crosses, but an elementary consideration 

 of the effect which incident or surrounding conditions may have 

 upon the behavior of characters in hybridization. A more de- 

 tailed treatmsnt of this work must be postponed for a more 

 extensive publication. 



The results obtained in both sets of experiments are in general 

 very much in harmony with results which have been obtained by 

 certain workers in experimental embryology, with this difference, 

 that in this case the analysis has been carried out through two 

 or more generations, and in the case of experimental embryology 

 the experiments cease with the obtaining of larval forms. 



The whole question of dominance is at the present time a try- 

 ing one. Practically all the criteria which have been proposed 

 from time to time for the detection of the dominant member 

 of a cross have in one way or another been broken down, and 

 Davenport's definition, that the dominant member of a cross is 

 only to be detected, in many organisms at least, by the determina- 

 tion of which character is the more variable in subsequent genera- 

 tions, seems to me not likely to be of general applicability. In the 

 materials which form the basis of this paper it could not possibly 

 be used, because both of the extracted forms are at times equally 

 variable, at times one is more variable than the other, and at 

 other times both are relatively invariable. 



What determines the dominance of one parental type over the 

 other is a question of vital importance in both normal and cross 

 fertilization. Apparently, one of the earliest workers to attack 

 this question was Vernon in 1898. He suggested that the hybrid 

 echinoderm larvae which in the summer were largely maternal in 

 type, and in winter were largely paternal in type were due to 

 the relative differences in the ripeness of the sexual products used 

 in the cross. 



Doncaster, in 1904, concluded that temperature was the chief, 



