102 Scientific Proceedings, Royal Dublin Society. 



For our present purpose there is no need to collect more figures. In the 

 succeeding volumes, similar phenomena can be seen. Most often there is no 

 serious difference between dam and daughter ; but frequently they are 

 widely apart, and the daughter is sometimes tlie better, sometimes the worse. 

 If we look at the figures we see that they run in grades : the lowest being 

 from 550 to 600, and the highest from 1,000 to 1,200; and we see also that 

 a cow of either extreme grade never has a daughter belonging to the other. 

 The daughters of extremes are either extremes of tlie same kind or are of a 

 medium grade of from about 750 to 950, Again, the medium grade 

 cows have daugliters not only like themselves, but also of either extreme 

 grade. 



These phenomena suggest a Mendelian explanation, namely, that the 

 extreme grades are the parent strains, and the intermediate the hybrid. If 

 this suggestion be correct, then we ought to find three corresponding grades 

 of bulls liaving daughters falling into similar grades according to the 

 Mendelian formula. 



The data founded upon this point are fewer than might be desired. They are 

 sufficient, liowever, for the business in hand. Denmark is the place in which 

 most data should have been found ; but tlie Danish herd-books contain only 

 a selection from their stock ; consequently, it is difficult to find many bulls 

 having more than a few daughters in the Herd-book. At home, again, few 

 breeders liave kept milk records long enough for our purpose, and still fewer 

 have retained many daughters of the same bull till they were old enough to 

 have several records kept. We have no other data meantime, however, and 

 the choice lies between using such as we possess and waiting for more. The 

 present inquiry may help to produce more. We shall give the Danish data 

 first : — 



