FACTS OF INHERITANCE 145 



any given pedigree, are eliminated by a law that 

 deals only with average contributions, and the 

 varying prepotencies of sex in respect to different 

 qualities are also presumably eliminated." Thus 

 an inheritance is not merely dual, but through the 

 parents it is multiple, and the average contribu- 

 tions made by grand-parents, great-grandparents, 

 etc., are definite, and diminish in a precise ratio 

 according to the remoteness of the ancestors. 



EXPERIMENTAL STUDY. Perhaps we may most 

 profitably illustrate the experimental study of 

 heredity by asking what the possible results are 

 of pairing two hypothetical organisms. Although 

 prediction as to the result of any individual pairing 

 is apt to be falsified (except in clear cases of 

 Mendelian inheritance), there are some well- 

 known alternatives of expectation. 



(1) Pairing of Similar Pure-bred Forms. Let 

 us begin with the offspring of similar pure-bred 

 organisms. When similar forms are bred together 

 for several generations a certain uniformity of type 

 is likely to result. If by selection the most similar 

 are mated together, while the least similar are 

 persistently removed from the stock, and if there 

 is also some measure of inbreeding, then there is 

 likely to be more or less constant uniformity of 

 type. These " pure-bred " organisms produce 

 others like themselves, and we suppose this to mean 

 that the hereditary items in the ovum have not 

 only their counterpart, but their equivalent, among 

 the hereditary items in the spermatozoon. This, 

 then, is one of the modes of inheritance that the off- 

 spring closely resemble the parents and one another. 

 The variability is restricted within a small range. 



(2) Blending. Passing from the mating of 



10 



