396 
Vol. XXX, No. 5 
Joumal of Agricultural Research 
production for the 36 periods is higher 
than that observed by 1.0 per cent. 
Grouping the results as before one has: 
In 34 cases, or 94.4 per cent of the total 
number, the difference exceeds 1 per 
cent, of which 23 are plus; in 26 cases, 
or 72.2 per cent of the total number, the 
difference exceeds 2 per cent, of which 
20 are plus; in 14 cases, or 38.9 per cent 
of the total number, the difference ex¬ 
ceeds 3 per cent, of which 9 are plus; in 
9 cases, or 25 per cent of the total num¬ 
ber, the difference exceeds 4 per cent, 
of which 6 are plus; in 7 cases, or 19.4 
per cent of the total number, the dif¬ 
ference exceeds 5 per cent, of which 4 
are plus; in 2 cases, or 5.6 per cent of 
the total number, the difference ex¬ 
ceeds 6 per cent, of which 1 is plus; in 
2 cases, or 5.6 per cent of the total 
number, the difference exceeds 7 per 
cent, of which 1 is plus; and in 1 case, 
or 2.8 per cent of the total number, 
the difference exceeds 8 per cent which 
is minus. 
In considering, then, the results of 
Tables I and II, the following questions 
suggest themselves: First, in compar¬ 
ing the heat production obtained by 
the direct and the indirect method 
which of the two possesses the greater 
accuracy? Second, what are the pos¬ 
sible causes for the wider differences 
between the computed and the ob¬ 
served heat production in the cow ex¬ 
periments than in those made on steers? 
In studying these questions let us con¬ 
sider, briefly, the sources of error and 
the methods used in each. 
SOURCES OF ERROR IN THE IN¬ 
DIRECT METHOD 
The indirect method here considered, 
by which the figures for computed heat 
production in Tables I and II were 
obtained, is the second of the two in¬ 
direct methods referred to above, the 
so-called balance method. The basic 
principle upon which both of these 
indirect methods are based is that the 
oxidation of a given substance in the 
body liberates the same amount of 
energy as does its oxidation outside 
the organism. The difference, then, 
between the gross energy of the feed 
and the energy of the total excreta, as 
determined by analytical methods, 
gives the energy derived from tKe 
feed, and is equal to the heat given off 
by the animal, in case the animal does 
not gain or lose body tissue. A loss of 
body tissue, of course, involves a 
liberation of energy exceeding that 
derived from the feed; therefore the 
energy-equivalent of the body tissue 
lost is added to the energy derived 
from the feed in computing the total 
heat production. On the other hand, 
a gain of body tissue means that not 
all of the energy derived from the 
feed was liberated, a part of it being 
stored in body tissue, the energy- 
equivalent of which is subtracted from 
the energy of the feed in computing the 
heat production. 
Representing the heat production by 
H, the energy of the feed by F, the 
energy of the total excreta (feces, 
urine, methane, brushings, and milk) 
by E, and the energy of the body gain 
by G, the general working formula for 
heat production is H = F —E —G. 
The estimation of the gain or loss 
of body tissue, which is an essential 
feature of this method, is computed 
in accord with the conception of the 
schematic body, which regards the 
organic matter of the animal as 
composed essentially of protein and 
fat, with at most comparatively small 
amounts of carbohydrates (glycogen). 
The supply of glycogen is assumed to 
remain constant during the experiment, 
and the gain or loss of protein and fat 
is estimated from a balance between 
income and outgo of nitrogen and 
carbon. 
The factors used in the computation 
are the following: protein = N X 6; 
fat=CX 1.307; carbon in protein = 
52.54 per cent; heat of combustion of 
protein per gram = 5.7 Calories; heat 
of combustion of fat per gram=9.5 
Calories. The derivation of these 
factors is fully discussed by Armsby 
( 0 ). 
The analytical data needed for 
computing the heat production by the 
balance method are: (1) Dry matter in 
the feed, feces (or feces and urine 
mixture) and brushings; (2) nitrogen 
in the feed, feces, urine, milk, and 
brushings; (3) carbon in feed, feces, 
urine, milk, and brushings; (4) carbon 
in the gaseous excreta (respiration 
apparatus); (5) carbon in methane 
(by respiration apparatus); (6) energy 
in the feed, feces, urine, milk, and 
brushings. 
The accuracy of this indirect method 
obviously depends on the accuracy of 
the factors used and on the accuracy of 
the analytical results. To what extent 
the factors used may be responsible 
for the differences between the com¬ 
puted and the observed heat produc¬ 
tion is difficult to say. Certainly any 
error due to the factors used should be 
least where the gain or loss of body 
tissue is least. Referring, however, to 
the results in Tables I and II, one 
notices in many cases relatively small 
differences between the observed and 
