BLENDING INHERITANCE 195 



Since the average litter among rabbits is about 

 five, the chances that these five rabbits will breed 

 true to their hybrid parents and form a perfect 

 blend between their grandparents is 20 out of 64, 

 while the chance of their being like either grand- 

 parent is only one out of 64. 



It should be noted further that 50 out of 64, or 77 

 per cent, of these hybrids of the second filial gen- 

 eration would have an ear-length between 140 and 

 180, thus approximating a "blend" closely enough 

 to be so classified upon a casual inspection. 



Moreover, if it should be found that excessive 

 ear-length in rabbits is due to more than three dupli- 

 cate determiners, the possibilities of getting anything 

 but an apparent blend would be much decreased. 



The fact, furthermore, that the fractional ear- 

 lengths of the hybrid rabbits in Castle's experiments 

 bred approximately true in the second and subse- 

 quent filial generations, may also be explained by 

 the Nilsson-Ehle hypothesis. 



For example, half lop lengths, according to this 

 explanation, are those with three doses of the deter- 

 miner for excess ear-length. It follows that the 

 progeny of two rabbits each carrying three doses of 

 a determiner will likewise, after the reduction during 

 the maturation of the germ-cells, have three doses 



/3 _i_ 3 \ 

 of the determiner ( = 3 J. 



It would be interesting to breed rabbits having 

 ears of one eighth lop length in which, according to 

 the foregoing hypothesis, there would presumably be 



