18 C. M. JACKSON 



Since the coefficient of variation for each Utter and also the mean 

 of the total population remains unchanged by this procedure, the 

 products may be considered as forming a single large litter, whose 

 coefficient of variation may be calculated in the usual way. The 

 results by this method are seen to correspond closely to those by 

 the preceding methods. In fact all three methods agree much 

 more closely than might have been anticipated, considering the 

 relatively small number of litters and of individuals. 



Comparing the coefficients of fraternal variation, calculated on 

 the litter basis, with those of the total population calculated by 

 the ordinary method at corresponding ages, it is evident that 

 (excepting only the unusually large fraternal variation for the 

 female at ten weeks) the litter variation is always very much 

 smaller, being usually only from one-third to one-half as large as 

 the total variation. It also appears that the increased varia- 

 bility in the body weight of the total population from twenty 

 days onward is due to an increased variability between litters more 

 than to increased variability among the individuals within the 

 litter. Dunn ('12) states (p. 137) that if conditions of growth 

 are favorable the rats of unlike initial weight within a litter 

 tend to approach each other in weight as growth goes on. If 

 this be true, the fraternal (intra-litter) variation should decrease 

 with age. 



Table F includes my observations upon the variation in indi- 

 vidual litters at successive ages. The sexes are here taken 

 together. It is evident that there is considerable variation, both 

 among different litters at the same age and in the same litter at 

 different ages. The data suggest that there is a tendency for 

 intra-litter variation to decrease up to the age of about twenty 

 days, followed by a tendency to increase. This is supported to 

 certain extent by table E, showing the coefficients of variation 

 on the litter basis. 



It may be concluded from all the evidence available that in 

 general the variation in body weight within a given litter of 

 albino rats is probably less than half that of the general popula- 

 tion of the same age under similar environment. Nevertheless, 

 even within the litter, variations are sometimes so great that it 



