7 76" PHYSIOLOGIC GENETICS 



particular in kinds of interaction, that is to be found in this case in even a rather limited 

 genetic system. 



POLYDACTYLY 



Attention is now directed to a different sort of morphologic character, the occur- 

 rence of atavistic digits. The guinea pig, like all species oiCaviidae, lacks the thumb, 

 great toe, and small toe. A fourth toe, exactly like the small toe of related rodents, is 

 not, however, uncommon. Among 22 inbred strains there were 1 1 that were invariably 

 3-toed on the hind feet (including strains 13 and 32), the small toe appeared sporadic- 

 ally in five (including strain 2, the major branch of which, however, was entirely free 

 of the trait) and in six the incidence was fairly high (including strain 35). 1445, 1446 ' 1447 

 A strain, D, with 100 per cent occurrence of a well-developed small toe was produced by 

 Castle 181 by selection. Crosses were made between strain D (4-toed) and the 3-toed 

 strains 2, 13, and 32 and between D and strain 35 with 31 per cent in the branch des- 

 cended from a single mating in the 12th generation. 1451 



The results from 2 x D and 32 x D gave complete dominance of 3-toed in F l5 

 except for one individual in 26 in the latter, and passable 3 : 1 and 1 : 1 ratios in 

 F 2 and the backcross to strain D respectively. These results suggest segregation of a 

 single, essential, recessive factor for the small toe but tests of the supposed segregants 

 from (2 x D) x D gave results that completely vitiated this hypothesis. The 3-toed 

 segregants produced only 23 per cent 3-toed (in 186), whereas the 4-toed segregants 

 gave nearly the same result (16 per cent 3-toed in 119). The most plausible hypothesis 

 seemed to be cumulative action of multiple factors on the underlying physiology and 

 two thresholds: one for any development of the small toe, below which there is homeo- 

 static control of the normal 3-toed foot and, slightly higher, a threshold (or better, a 

 ceiling) above which development of the small toe is controlled homeostatically, or 

 canalized in Waddington's 1352 terminology. The cross 13 x D gave 67 per cent 3-toed 

 and 33 per cent 4-toed in F x . Tests of the two F x types gave results that did not differ 

 significantly either in F 2 or in the backcross to D. There was no simulation of one- 

 factor heredity, but the results fit well with the interpretation of multiple factors and 

 two thresholds. Most of the young from 35 x D were 4-toed. The results in F 2 

 and the backcross to 35 again fit the above hypothesis. 



The results from these crosses in F 1 and F 2 are represented in figure 28 according 

 to this threshold hypothesis. There was considerable variability in strain 35 in relation 

 to age of mother (high percentage of 4-toed from immature mothers) and season 

 (excess of 4-toed in winter and early spring) , 1414 The standard deviation on the physio- 

 logic scale can be determined as that of the normal curve that yields the observed 

 trichotomy, taking the interval between the thresholds as the unit. It is approximately 

 0.80. The same standard deviation was assumed to apply to the other inbred strains 

 and to the three Fj/s. The mean of D was put at 2.5a (2 units) above the upper 

 threshold as a value which might be attained by selection but at which overlap of the 



