52 BULLETIN 1090, U. S. DEPARTMENT OF AGRICTTLTUBE. 



however, a remarkably fortunate combination of genetic factors 

 segregated out in one line, and this line ultimately displaced all other 

 lines. 



Family 2 had a somewhat similar history. There were many lines 

 in the early generations, but after the seventh generation all of the 

 matings traced back to a single pair in the fourth generation. The 

 family began as one of the weakest in nearly every way and was ab- 

 solutely the poorest in rate of gain and in frequency of litters. The 

 successful line continued to produce small litters of undersized ani- 

 mals, and held the record for smallness of gains in the second period, 

 as well as in the first. The percentage born alive, however, increased 

 from very poor to medium, and the percentage raised of those born 

 alive increased from medium to the very best. The most remarkable 

 change was from the lowest in litters per year in the first period to a 

 tie with Family 23 for the highest record in the second. Since 1915 

 small litters and very light weight have remained characteristic, but 

 also frequent litters and great success in raising the young born alive. 

 In spite of its defects it is one of the easiest families to maintain. Its 

 ability to raise the young which are born alive seems to be correlated 

 with high vitality thereafter. 



Study of the pedigrees indicates that Family 24 must probably be 

 considered as next to Family 35 in the extent to which there has been 

 a shifting in the importance of lines of descent. Table 5 shows that 

 it stands third in change of characteristics. In this case, however, 

 the change was for the worse in most respects. Apparently a number 

 of fairly vigorous lines were superseded by an inferior one. This 

 would seem hardly as likely to happen as the expansion of a good 

 line, noted in Families 35 and 2, but it is not impossible. It must be 

 remembered that the characters of a line of descent can be determined 

 only from the average of many individuals. At a given time the 

 genetically inferior line might well happen to be represented by the 

 more vigorous individuals. 



Among the remaining families, Xos. 20, 21, and probably also 34, 

 were ones in which the rather important changes in character may 

 also well have been due to the expansion of particular lines of descent. 

 The changes in 23 and 39 can not be explained so satisfactorily in this 

 way. In certain cases families remained fairly true to type, even 

 though the predominant lines of descent altered considerably. This 

 was true of Families 1, 9, 13, 19, and 38. This, however, is not sur- 

 prising. In the remaining families the original lines of descent run 

 parallel to each other through both periods and there were few im- 

 portant changes in the family characteristics. 



A consideration of these family histories, especially those of Fam- 

 ilies 35, 2, and 24, strengthens the argument for the inheritance of 

 characters in which the coefficient of correlation between the sue- 



