n] Imperfect Dominance 53 



roan is a heterozygous character, caused by the meeting of 

 the factors for red and white*. The blue-roan so often 

 seen in the cross between black Aberdeen-Angus cattle and 

 white shorthorns is presumably the corresponding hetero- 

 zygote form for black and white. 



With further knowledge of the details and closer exami- 

 nation of material probably many such cases will be found. 

 Darbishire, for example, has lately shown that though it is 

 usually impossible in the case of peas to tell pure round 

 seed from the heterozygous rounds, by external appearances, 

 yet on microscopical examination the two classes can be 

 distinguished at once by the different structure of the starch 

 grains (94). 



Many such cases where dominance is imperfect are now 

 known. This phenomenon has no bearing on the more 

 important question of the degree of perfection with which 

 segregation is accomplished. The supposition that domi- 

 nance was an essential phenomenon of Mendelism was of 

 course a delusion. Imperfection of dominance does not 

 even obscure the application of Mendelian analysis. The 

 cases in which difficulty does arise are those in which 

 dominance is irregular and the recessive class cannot be 

 distinguished with certainty. In the fowl, for instance, the 

 extra toe is usually a dominant, but in some strains there is 

 irregularity, and birds without the extra toe may neverthe- 

 less transmit it. So also the blue colour of maize seeds, 

 though usually a dominant, may sometimes be carried on 

 by seeds which appear white (Lock, 174). Even in these 

 examples, however, there is no reason to think that such 

 irregularities are indications of imperfect segregation. 

 It is not impossible that they may be ascribed to inter- 

 ference caused by the presence of other factors in various 

 combinations, and sometimes, no doubt, to disturbance by 

 external conditions. 



All observations point to a conclusion of great import- 

 ance, namely that a dominant character is the condition 



* The evidence brought forward by Mr Wilson suggests that there is a 

 further complication in this case, and that probably red x white sometimes 

 gives dominant red, not roan. The irregularities in his table of results 

 distinctly point to some such conclusion, which would be quite consistent 

 with appearances seen elsewhere. 



