306 . CHESTER A. STEWART 
(13 rats; 6 M, 7 F), the average net body weight at each age being 
approximately 10, 138 and 15 grams respectively. 
In addition sixteen rats, starting at three weeks of age, were 
underfed for much longer periods (to 139 to 412 days of age; see 
table 1). They were then killed and autopsied. Tor one litter 
of the group fasting for a long period (M 29), the underfeeding ° 
started at the age of five weeks. Four rats of litters S 26 and 
M 29 (one male and one female of each litter), and one male of 
litter St 44 were accidentally asphyxiated by illuminating gas. 
These rats are all included in the two groups of test rats weighing 
approximately 70 grams. The two asphyxiated individuals of 
litter M 29 had been refed about thirty-six hours, which has 
resulted chiefly in a great increase in the contents of the alimen- 
tary canal. 
In general the plan was to kill the test rats at the same body 
weight as the corresponding controls, but this was not always 
possible. Nevertheless, the average body weight of the different 
groups of test rats differs only slightly from that for the cor- 
responding controls, as is shown in table 1. In comparing the 
data from the test rats with those from the controls throughout 
the paper, this difference has generally been disregarded, although 
strictly speaking, there should be a slight correction in every case 
for differences in body weight. Such a correction should be based 
upon the net body weight (excluding content of stomach and in- 
testines), rather than upon the gross body weight, however. As 
the differences in body weight are in all cases small, it seems 
justified to ignore them in making comparisons. 
The control rats remained constantly with the mother through- 
out the nursing period (three weeks). After this time they were 
fed an abundant diet of whole wheat (Graham) bread soaked 
in whole milk. Water was also supplied. 
The test rats (in the experiments upon very young animals) 
were removed from the mother at frequent intervals for periods 
of usually 15 to 24 hours (occasionally longer, the maximum 
being 43 hours), and were permitted to nurse during the inter- 
vening time. For convenience, these test rats are frequently re- 
ferred to throughout this paper as the rats fasting from birth 
