110 



INHERITANCE IN GUINEA-PIGS. 



The offspring of male 4003 rough E are of special interest. He was 

 undoubtedly homozygous rough; the chance that he was heterozygous 

 is (|) u (f) 6 = 0.0001. As he was the lowest grade of rough, he very 

 emphatically disproves any necessary relation between homozygosis 

 and high development of the rough character, or between heterozygosis 

 and partial roughness. 



TABLE 54. 



(10) Occasionally a smooth from a cross which produces rough E 

 will transmit the rough factor, breeding like a rough E. Rough E 

 grades into smooth. From a cross which can produce rough E, 8 smooth 

 animals were tested by crossing with 4-toe smooths to determine 

 whether a smooth can ever be like rough E genetically. One such 

 animal, female R201, was found. 



TABLE 55. 



The case is not quite as clear as could be desired, since R201 seemed 

 to show a trace of irregularity on one hind toe when first graded. As 

 an adult she is indistinguishable from a smooth. 



These experiments are sufficient, it is believed, to establish the mode 

 of inheritance of the major variations of the rough character. 



First, it is clear that partial-roughs do not owe their roughness to an 

 allelomorph of the rough factor or to an independent duplicate rough 

 factor, but to the same factor found in full-roughs. Reasons were 

 given under (2). 



Next, any hypothesis is untenable according to which partial-roughs 

 are due to imperfect dominance and hence are necessarily heterozygous 

 either with the ordinary smooth factor of guinea-pigs or with a more 

 potent allelomorph of the latter present in wild cavies and special stocks 

 of tame guinea-pigs. The latter hypothesis was suggested by Det- 

 lefsen (1914), in the case of the partial-roughs among the rufescens 

 hybrids. He represented the rough factor by Rf, the ordinary smooth 

 factor by rf , and the smooth factor of Cavia rufescens by rf. He sup- 

 posed that Rf is completely dominant over rf, but incompletely domi- 

 nant over rf '. Thus Rfrf would be a full-rough, but Rfrf' a partia- 



