1092 
Journal of Agricultural Research 
Vol. XXXI, No. 11 
Accordingly, we have excluded Period III of experiment 179 from 
the computation of the average heat-increment value of corn meal, 
because of the small quantity of meal represented in this period, and 
we have also excluded from the average the heat-increment value of 
clover hay and corn meal as obtained by comparing Periods III and 
V in experiment 220, because of the small difference in feed. 
Among the sources of error in the heat production, and con¬ 
sequently in the heat increments, in some of the experiments here 
considered, two require especial mention. These are (1) the refusal 
of feed in some of the periods of heavy consumption, and (2) the 
probability that the temperature of the calorimeter was below the 
critical temperature for the animal in some of the periods in which 
the subjects on submaintenance rations suffered extensive losses of 
energy from the body. 
The amounts of feed refused have been reported in the publication 
of the details of these experiments (7, 2, 8 , 6 , 7, 8, 9) and will not 
be repeated here. In experiments 209 and 210, which have been 
published only in condensed form (i 6 ) and not in full detail, there were 
considerable amounts of feed refused during some of the periods. 
During Period IV, experiment 209, the animal was offered daily 
6.8037 kg. of dry matter of alfalfa hay, but the average daily amount 
eaten was only 6.1737 kg. (See Table II.) During Period I of the 
same experiment the animal rejected 147.6 grams of dry matter of 
hay and 9.3 grams of grain. In experiment 210 the animal was offered 
daily during Period I 7.6303 kg. of dry matter of corn stover, and dur¬ 
ing Period II 4.2040 kg., but the average daily amounts eaten during 
these periods were 4.3353 kg. and 3.5475 kg., respectively. 
In comparing the heat production and the rations of the different 
periods, for the derivation of the heat increments, the amounts of 
feed refused were in each case subtracted from the total offered. It 
is, however, readily understandable that such a procedure may not 
always be effective in accomplishing the desired correction for 
refusal of feed, when it is considered that the balance of matter 
represents an average of 10 days’ feeding, while the heat production 
is derived from only two days. It is clear that the success or failure 
of such an attempt at correction of the heat production must depend 
to a large extent on whether the refusal of feed was uniform through¬ 
out the entire digestion period, including the calorimeter days. 
In connection with the effect of refusal of feed on the heat produc¬ 
tion, its effect on the determination of metabolizable energy, which 
also is involved in the computation of net-energy values, must ilso 
be mentioned. 
A refusal of feed occurring early enough in the collection period to 
be wholly reflected in the fecal outgo of this period would be corrected 
for by subtraction of the amount of the refused feed from the amount 
of the feed offered, but if the decrease in fecal outgo consequent upon 
the refusal of feed does not all occur prior to the termination of the 
collection period, the attempt at correction of the balance of matter by 
subtraction of the amount of feed refused from the amount of feed 
offered would at best be incomplete, and, under some conditions, 
might actually introduce into the balance a new error of the extent of 
the entire amount of the refused feed thus subtracted from the amount 
of feed offered. In the experimental practice of this institute the 
