176 



THE AMERICAN NATURALIST 



[Vol. LI 



Eureka than in the Samoa race, though here again the 

 difference related to averages and did not hold for all 

 individual cases. 



A comparison of the mean figures for absolute tail 

 length in two series of mice is not entirely justifiable, par- 

 ticularly if the two lots of individuals differ somewhat 

 in mean body size. But the relative tail lengths (ex- 

 pressed as percentages of body-length) may be fairly 

 compared, since there is good evidence that these ratios 

 remain nearly constant after the first few months of life. 

 The following table allows of a comparison between the 

 two races, in respect to this character : 



The differences between the Samoa lot (sexes com- 

 bined) and the Eureka males and females are 6.91 per 

 cent, and 6.12 per cent., respectively. These differences 

 are about seven and six times their probable errors, re- 

 spectively. Their significance may therefore be regarded 

 as fairly certain, despite the small numbers comprised 

 in the Samoa series. 



As regards foot-length, the two races do not differ 

 significantly. But the ear, as already stated, is appre- 

 ciably longer in the Samoa mice, this difference being 

 perceptible, even without measurement. Here, as in the 

 case of tail-length, a simple comparison of gross averages 

 for the two groups would be unjustifiable. But in the 

 present instance, the conversion of the absolute values 

 into percentages of body-length would be equally unjus- 

 tifiable, since the growth of the ear is not at all propor- 

 tionate to that of the body as a whole. We must there- 

 fore resort to the method of **size groups," i. e., we must 

 divide each of our two lots of animals into small groups 

 comprising individuals of nearly equal size. 



In the case at hand, we have fifteen groups, or rather 



