POSTNATAL GROWTH IN THE ALBINO RAT 



19 



is never safe to draw conclusions from a single litter, and there 

 is some risk even with several litters. These conclusions are of 

 practical importance in selecting animals for experimental work, 

 and in interpreting the results of such work. Although fraternal 

 variability was not directly calculated for the individual organs, 

 the same principle doubtless holds good with them; for, as will 

 be shown later, all the organs are more or less closely correlated 

 with the bod}^ weigh J;. 



In the human species, fraternal variation appears to be rela- 

 tively greater than in the rat. For example, Galton ('94) found 

 the mid-stature for the human adult (male) population to be 68.2 

 inches. The quartile (or probable) deviation for the whole adult 

 population he found to be about 1.70 inches, and that for brothers 

 (average of four methods) about 1.06 inches. Thus for human 

 stature by this method the fraternal variability appears to be 

 only about 62 per cent of the racial variability. On the basis 



TABLE F 

 Coefficient of variation in individual litters at different ages 



^ Two males killed in litter M 9 at 7 days, and 2 females in A 33 at 20 days; one 

 (a 'dwarf') in litter M 7 died between 20 and 30 days. 



2 The large variation here was due to a 'dwarf which caught up with the others 

 in body weight before the age of 14 days. 



' At 10 weeks the coefficient of variation in this litter was 17.0, the increase 

 being due chiefly to sexual differentiation in body weights. 



