268 



E. H. HARPER. 



been collected which bear upon this point, and the same data 

 have been tabulated with reference to the two other matters men- 

 tioned at the outset, the prepotency of sex and the relation of 

 age to prepotency. 



THE SELECTION OF DATA TO DETERMINE PREPOTENCY. 

 With the view of determining certain of the factors of prepo- 

 tency, two thousand individuals of pure black or gray color have 

 been selected at random, consisting of one thousand colts and 

 one thousand fillies, the parents likewise being of pure color, 

 one black the other gray. This method of selection gives a 

 positive resemblance of the offspring to the one or the other 

 parent. It seemed to the writer that the pure colors were 

 the only ones from which inferences could be drawn, since the 

 descriptions of color are too meager to warrant a safe infer- 

 ence in cases of parents or offspring of mixed color. It 

 would be impossible in the case of mixed colors to decide which 

 was the preponderating color or to infer the resemblance of the 

 offspring to one or the other parent. This defect is apparently 

 inherent in data not originally collected for any scientific purpose. 

 Any consideration of the Mendelian hypothesis is precluded. 

 And while that fascinating theory would naturally be the first to 

 turn one's attention to investigations of this sort, it has seemed 

 that the limited use of data as above outlined might prove to be 

 warrantable as a method for certain purposes proposed in this 

 this paper. All of the data presented in this paper outside of 

 Table I. consist of that just referred to, namely, the 2,000 indi- 

 viduals selected at random according to the method described, 

 and including 1,000 colts and 1,000 fillies. 



TABLE II. 



PREPOTENCY IN RELATION TO SEX. 



In Table II. the data are arranged with reference to sex pre- 

 potency. It is seen that of 1,000 colts 543*01- 54.3 per cent. 



