CLASSIFICATION OF VARIATIONS 95 



present which is responsible for the production of that character. 

 Where there is no definite dominance and the heterozygote is 

 of an intermediate nature we should be unable to declare on 

 which side the factor concerned was present and from which side 

 it was absent. The degree of dominance becomes thus the 

 deciding criterion by which we distinguish the existence of factors. 

 But it should be clearly realized that in any given case the argu- 

 ment can with perfect logic be inverted. We already recognize 

 cases in which by the presence of an inhibiting factor a character 

 may be suppressed and purely as a matter of symbolical expres- 

 sion we might apply the same conception of inhibition to any 

 example of factorial influence whatever. For instance we say 

 that in as much as two normal persons do not have brachydacty- 

 lous children, there must be some factor in these abnormal persons 

 which causes the modification. Our conclusion is based on the 

 observed fact that the modification is a dominant. But it may 

 be that normal persons are homozygous in respect of some factor 

 N, which prevents the appearance of brachydactyly, and that in 

 any one heterozygous, Nn, for this inhibiting factor, brachy- 

 dactyly can appear. Similarly the round pea we say contains 

 R, a factor which confers this property of roundness, without 

 which its seeds would be wrinkled. But here we know that the 

 wrinkled seed is in reality one having compound starch-grains, 

 and that the heterozygote, though outwardly round enough, is 

 intermediate in that starch-character. If we chose to say that 

 the compoundness of the grains is due to a factor C and that two 

 doses of it are needed to make the seed wrinkled, I know no 

 evidence by which such a thesis could be actually refuted. That 

 such reasoning is seemingly perverse must be conceded; but 

 when we consider the extraordinary difficulties which beset 

 any attempt to conceive the mode of origin of a new dominant 

 factor, we are bound to remember that there is this other line of 

 argument which avoids that difficulty altogether. In the case of 

 the " Alexandra "-eye in Primula, or the red calyx in Gates's 

 Oenothera, inverting the reasoning adopted in the text, we may 

 see that only the Primula homozygous for the yellow eye can 

 develop it and that two doses of the factor for the rubrinervis 



