GENIC INTERACTION 1 79 



parts of the foot and the organism as a whole that most of the genetic system concerned 

 with thumbs, great toes, and small toes is necessarily still present. The loss of these 

 digits is presumably due to some simple, superimposed, inhibitory process that stops 

 formation of the normal number of lobes on the developing limb buds. Any genetic or 

 environmental effect that inhibits this inhibitory process releases the substrate for action 

 of the whole array of genes that shaped these missing digits in the remote ancestors. 

 Homology, whether replicative in the same individual or phylogenetic, must be con- 

 sidered to be a matter of degree, the result of calling into play more or less similar 

 systems of genetic reactions by more or less similar developmental processes. 1434 



The multifactorial Polydactyly may conceivably involve ancestral genes that have 

 been carried at low frequencies throughout the history of the Caviidae or have been 

 brought back by reverse mutation, but this is hardly likely for Px. Yet if Pxpx acts 

 merely by inhibiting a relatively simple inhibition acquired in evolution, the thumb, 

 small toe, and great toe that develop in this genotype under the released ancestral 

 heredity may be considered as essentially homologous to these digits in the mammals 

 that have never lost them, in spite of the monstrous characteristics of the foot in PxPx. 



Thresholds have been shown to play a major role in other morphologic deviants 

 of the guinea pig of which otocephaly has been most intensively studied. 1433, 1455, 1458 

 It may be added that Griineberg 501 reports that similar "quasicontinuous" variation 

 interrupted by thresholds is the commonest situation in the mouse in the case of 

 morphologic deviants. 



TYPES OF INTERACTION 



The effects of factor replacement under varied genetic or environmental conditions 

 may be put into four categories. 



Constant effects. — There is usually constancy of effect in cases in which the replace- 

 ment in question is associated with variations in a character that seems obviously un- 

 related; for example, R, rr (rough and smooth fur respectively) and any pair of colors. 

 This also often holds for widely different aspects of a character that is single only in a 

 broad sense; for example, A, aa {agouti and nonagouti respectively) on a black (B) or 

 brown (bb) background. 



With respect to grades of a single quantitative character, constancy of effect 

 requires somewhat arbitrary definition. One may, for example, choose to treat 

 either a consistent additive effect or a consistent multiplicative effect as a constant effect. 

 Strict examples of either of these types of constancy are, however, probably uncommon. 

 They were not found in the quantitative studies of intensity of the colors of the guinea 



Pig- 



Cumulative but not constant effects. — There are many examples of this among the 

 guinea pig colors. Thus on assigning 100 as the measure of quantity of yellow pigment 

 with CF, that with c d c d F was 38, with Cff 33 and with c d c d ff 5. Any two-factor case 

 such as this could be made additive arbitrarily by a suitable transformation of scale, 



