334 
Francis B. Sumner 
To determine the extent of these ditferences, we shall consider, not 
the dift'erences between the gross averages, for reasons already stated, 
but the average difference between the »warm« and the >cold« figure 
within each size-group. Tlie mean difference in tail length, as thus 
computed, shown by the parent generation at the age of six weeks, 
was 8.195 mm, that shown by the offspring being 1.264. Thus the 
latter figure is a little over 15 per cent of the former. 
6 7 S 9 10 11 12 13 14 15 16 17 18 19 
Showing foot and ear length of same mice as those represented in Fig. 6. Note that vertical scale is 
five times as great as in preceding figure. For further explanation, see the latter. 
As regards foot and ear length, we may profitably compare the 
generation with which we are dealing bere with the first offspring 
of the same parents. These first offspring, it will be recalled, were 
subjected to temperature differences similar to (though not as great 
as) those to which the parents had been exposed. An instructive com- 
parison may therefore be made between the curves (Figs. 3 and 4) 
for these mice which had themselves been directly modified and 
