54 F - B - SUMNER AND R. R. HUESTIS. 



case. A conspicuous exception to this general rule is. however, 

 to be noted in the case of foot pigmentation. 



(5) In general, coefficients of fraternal correlation seem to be 

 lower for the F 2 generation than for the F,. 



(6) Pairs of characters which are correlated in the parent 

 races and in hybrids between these show, in general, a higher de- 

 gree of correlation in the F 2 generation than the F,. 



(/) Certain elements which contribute to the general pelage 

 color of these mice (aside from those which are obviously inter- 

 dependent) are found to be correlated with one another, and thus 

 either to have a common genetic basis or at least to depend upon 

 factors which are more or less closely linked. On the other hand, 

 another pair of characters which differ in the same direction in the 

 two races (foot pigmentation and percentage of ''black" in the 

 pelage) are apparently not at all correlated within the single race, 

 although individual differences in these characters are known to 

 be partly genetic. Since these two characters are known to have 

 varied together geographically in more than one species of Pcro- 

 myscus, their coincident modification has probably resulted, in 

 some way, from external agencies and not from correlated vari- 

 ations, due to internal causes. 



Items I to 6, in the foregoing summary, bear upon the theory of 

 "multiple factors," i.e., the view that each of the subspecific dif- 

 ferences here discussed is due to the cumulative effect of a number 

 of independent genetic factors, whose effects, taken singly, are 

 similar in their nature. Without entering into an extended dis- 

 cussion, it may be said that the facts here presented conform, on 

 the whole, to the requirements of that theory, and that no facts nar- 

 rated in the present paper constitute serious objections to it. 

 Items 3, 4 and 6 represent findings which are somewhat at vari- 

 ance with those presented in certain recent papers by the senior 

 author, based upon other hybrid series, and in large degree upon 

 other characters than those here considered. 



After careful consideration of the present series, and a further 

 statistical examination of previous material, the senior author is 

 prepared to admit that the case for a " multiple factor " explana- 

 tion, while far from conclusive, is considerably stronger than he was 



