48 BULLETIN 1090, U. S. DEPARTMENT OF AGRICULTURE. 



periods. There are evidently genetic factors which act ahke on the 

 rate of growth before and after birth. It would, indeed, be surprising 

 if it were otherwise. The correlation between the family means 

 for weight at birth and weight at a year, of males born in litters of 

 three, was given as +0.63 in an earlier paper. (Wright, 1917) ^ 

 The difference between the undersized guinea pigs of such a family 

 as No. 2 and the large ones of such a family as No. 13, is in fact 

 even more marked among the adults than among the young. The 

 demonstration of factors which affect the rate of growth at all ages 

 has, however, no bearing on differentiation in general vigor. 



The other two significant correlations connect size of litter with 

 birth weight and gain, respectively, in the period from 1911 to 1915. 

 Each of these correlations was +0.62 ±0.08. The correlations, in 

 the period from 1906 to 1910, were not certainly significant, although 

 relatively high ( + 0.26 and +0.37 respectively). Because of the' 

 close connection between birth weight and gain, it is evident that 

 we are really dealing with a single correlation, one between size of 

 litter and rate of growth (for a given size of litter). There is an in- 

 dication of an interrelation here, although, as will be shown later, 

 there is another possible explanation which must be considered. 



In opposition to the hypothesis of a differentiation in general vigor, 

 we have the evidence of the remaining 30 correlations, only 1 of which 

 is as high as +0.30, and 11 of which are negative. The low correla- 

 tions in certain cases are especially noteworthy. The percentage 

 born aUve, and the percentage of those raised to weaning, show 

 virtually no correlation in the first period ( + 0.03) , and only a doubtful 

 one in the second ( + 0.30) . It would appear that these depend, to 

 a large extent, on independent hereditary factors. There is evidence 

 which is brought out in Bulletin 1121 that the percentage born 

 alive depends more on the vigor of the dam, while the percentage 

 raised of those born alive depends more on the vigor of the young 

 themselves. There must, of course, be important factors in common, 

 but the effects of these seem to be neutraUzed by the tendency toward 

 negative correlation which independent factors would have in this 

 case. Factors which tend to remove the less vigorous young at or 

 before birth tend thereby to increase the percentage raised of those 

 born alive, if they do not also prejudice the chances of the young 

 after birth. 



But even though there may be factors in common, the differences 

 among the families in vigor in these respects is in marked contrast 

 to the differences between stocks of guinea pigs raised under different 

 environmental conditions. Under unfavorable conditions both the 

 percentage born alive and the percentage raised of those born alive 

 tend to decline. 



5 See footnote 2. 



