21] INHERITANCE IN PEROMYSCUS 



451 



more extensive series, consisting of animals which are 

 free from any pathological modifications. 



We must urge in passing, however, that evidence of 

 segregation, even if valid, is not necessarily to be accepted 

 as evidence of complete segregation. There is no reason 

 why we might not have partial segregation, combined with 

 partial gametic blending, as Castle maintains. 



In two recent illuminating articles (1917, 1917a), Jen- 

 nings points out how Mendelian-mutationists of the most 

 extreme school have been driven by their own researches 

 into a position that does not differ, according to any prag- 

 matic test, from the one which they so long have vehe- 

 mently opposed. More and ever more minute hereditary 

 differences in the manifestation of a given character are 

 recognized, until the limit of distinguishability is ap- 

 proached. This state of affairs has been attributed to two 

 causes: (1) hereditary modifications in the constitution 

 of single factors, resulting in the formation of series of 

 gradations, allelomorphic to one another; and (2) the 

 existence of series of independent modifying factors, 

 cumulative in their effects. 



As remarked earlier in this paper, the contest has lat- 

 terly come to resemble that allegorical one of the two 

 knights, fighting upon the opposite sides of the same bi- 

 colored shield. And yet there would seem to be a differ- 

 ence. The two knights in the legend were both equally 

 right. In the present case, if we may judge by every 

 pragmatic consideration, the larger measure of right be- 

 longs to those who have contended for the frequent per- 

 manent blending of hereditary characters in crossing and 

 the continuous modifiability of these characters through 

 selection. The finely spun theories of their opponents 

 may help us to symbolize the machinery underlying these 

 phenomena, but the phenomena themselves, and not the 

 theories, are the indubitable realities in the case. 



IX. Summary 

 1. In the preceding pages, the differences, structural 

 and pigmental, which distinguish four geographical races 

 of deer-mice are discussed in some detail. The pigmental 



