The Method of Evolution 55 



dose in causing the development of the character. In size 

 inheritance, however, the single and double doses probably 

 produce xtxy different effects. 



A further apparent difficulty encountered in interpreting 

 blending inheritance as unit-character inheritance lies in 

 the multiplicity of the units involved, so that segregations 

 do not occur into a few discontinuous size classes, but into 

 classes so numerous and differing so little from each other 

 that it is very difficult to distinguish them. 



Suppose that in crosses of black with white guinea-pigs, 

 black were represented by two unit-characters B and B' , 

 instead of by one, residing perhaps in different chromosomes, 

 and that either one of these could by itself produce black 

 color, then a larger proportion than three-fourths of the 

 second generation offspring would be black, namely, \ |. 

 If, further, the presence of a larger number of factors for 

 black produced more black pigment in the fur than a smaller 

 number produced, then we should have gradations of black- 

 ness among the second generation offspring as follows: 

 4, 3, 2, T, o. 



Add a third factor for black in the supposed cross, 

 located perhaps in a third chromosome, and the pure whites 

 would be reduced to i to 64 of the second generation off- 

 spring, while the diff'erent gradations or intensities of black- 

 ness would become 6, 5, 4, 3, 2, i, o. The occasional white 

 individual would now difl'er so little from the lightest black 

 one that the two might often be confused, and there would 

 seem to exist all intermediate stages between pure white 

 and pure black, without entire segregation into either. By 

 selecting for the lightest or the darkest condition Avithin a 

 mixed race of such second generation offspring one would 

 obtain with each selection a larger proportion of extreme!}' 



