FUTUEE OF THE SCIENCE OF BREEDING 179 



from the standpoint of our discussion, forms four kinds of 

 gamete, viz. (I) those containing all three factors ; (2) those 

 containing any two of them (whether A + A^, A + A^, or 

 A^+A^ does not matter) ; (3) those containing but one ; and, 

 lastly, those containing none. Further, these four types of 

 gamete are produced in the ratio 1:3:3:1. When two such 

 series of gametes are brought together by mating the first- 

 cross animals with one another, it is easy to calculate that seven 

 different kinds of individuals are formed according as they 

 contain six or less doses of any of the three factors. In point 

 of size they will form a continuous curve (Fig. 9) of which 

 the extremes are identical with the weights of the original 

 parents, and the mean is that of the first crosses. 



On the whole the phenomena observed under experimental 

 conditions are similar to those expected on the hypothesis just 

 outlined. In the case of our poultry, for example, Maj or Bailey 

 and I showed that the observed phenomena could be explained 

 on these lines, in terms of four factors. There is of course a 

 further test. Certain of the individuals in the F2 generation 

 must be homozygous for either one or two or three of the 

 postulated factors. The extremes must clearly be in this 

 condition, and among the others homozygous individuals 

 should occur in each of the two classes of 15. In actual 

 experiment, testing for these homozygous intermediates 

 would be a hopeless task to enter upon. Nevertheless, if the 

 hypothesis is to serve as an acceptable basis for future work, 

 we ought to be able to test it by breeding the more easily 

 identified extremes together, and by showing that strains 

 breeding true to intermediate size can be produced. In the 

 case of the Sebright, Major Bailey and I showed that the small 

 F2 birds breed true, and I have now some evidence which 

 seems to make it fairly certain that a constant strain of 

 intermediate size can be formed. 



The h3rpothesis of multiple factors, then, presents a view 

 of these cases of quantitative variation which goes some way 

 towards bringing them into line with cases in which the 

 hereditary characters are of a distinctly qualitative nature. 

 But before we can regard this interpretation as more than 



