12 C. M. JACKSON 



evident that in general there is no such thing as a single 'normal' 

 course of growth and variation for a given species. We may, 

 however, determine the process under a given set of conditions, 

 and also learn the effect produced by varying the individual fac- 

 tors. It is probable that some variation in the intrinsic growth 

 factors is largely responsible for the differences noted between 

 Donaldson's data and mine. The possibility of unrecognized 

 pathological factors must also be kept in mind. 



We may now consider the variability in body weight found 

 in the present series.- The variability is best measured by the 

 coefficient of variation, which expresses the percentage of the 

 mean formed by the standard deviation. For the net body weight 

 (table 2), it is seen that the coefficient of variation, taking both 

 sexes together, increases from 12.3 at birth to 15.6 at seven days 

 and to a maximum of 28.4 at twenty days. Thereafter it de- 

 creases, being 21.3 at six weeks, 19.9 at ten weeks, and 19.1 at 

 five months. On account of the small nmnber of observations 

 and probable abnormalities at one year, no calculations were 

 made for the total at this age; but the coefficient of variation in 

 body weight for the 20 females at this age was slightly greater 

 than for those at five months. Even when allowance is made 

 for accidental variations due to the small number of observations 

 (as indicated by the relatively large probable errors), the approx- 

 imate extent of variation, and the general trend according to age 

 are clearly evident. Hatai ('08) in mature albino rats (over 

 150 days old) found the coefficient of variation for the gross 

 body weight in 53 males, 25.076 ±2.675 and in 51 females, 

 12.235 ±0.974. 



The difference in variability between the sexes of the rats used 

 in the present study is indicated in table 2. It is seen that the 

 coefficient of variation appears greater in the males at every 

 age, except at twenty days. No great stress can be laid upon 



^ The variability in body- length was not calculated. According to Donaldson 

 ('09), however, the body length of the albino rat is less variable than the body 

 weight; but the two are closely correlated, the coefficient of correlation being 

 0.90. In the human species, the body length (height) is only about one-third as 

 variable as the body weight (measured by the coefficient of variation). 



