‘260 PRINCIPLES OF ANIMAL NUTRITION. 
‘ gain) plus the energy expended in producing chemical and physical ° 
changes in the body. 
‘In. actual experimentation it: is witaa impossible to so 
adjust the food that there shall be absolutely no gain or loss of body 
substance, although its amount can be made relatively small. 
Experiments’ on this subject, then, necessarily fall under Cases 3 
or 4,and as a matter of fact, in all the experiments hitherto 
made, the subject has either done no mechanical work or this 
work has been converted into heat inside the calorimeter and 
measured along with that directly given off by the body, so that 
all these experiments fall.wnder Class 3. ‘ 
The quantities to be determined, then, are 
1. Potential energy of food. 
2. Potential energy of excreta (feces, urine, hydrocarbons, etc.). 
3. The heat produced (including that into which any mechani- 
eal work is converted). 
4, The potential energy of the gain or loss of body substance. 
5. The energy expendéd (or evolved) in producing changes in 
the body. 
If we can determine accurately these five factors, and having 
done so find the equality stated under 3 to exist in a large number 
of cases, we shall be justified in the conelusion that the law of the 
conservation of energy applies to the animal organism. 
The methods by which the first four of the above factors may be 
determined formed the subject of the preceding chapter. As re- 
gards the fifth, it has commonly been assumed that in an experi- 
‘ment begun and ended at the same hour of the day and under com- 
parable conditions, which has been preceded by a considerable 
period of uniform feeding and other conditions, and in which the 
subject was in apparent good health, the initial and final states of 
the body are substantially the same. While it seems highly prob- 
able that this is true, an actual demonstration of its truth is not an 
easy matter. With respect-to the body temperature in particular 
it is worthy of note that even a.slight variation would materially 
affect the results. Thus in a 1000-pound ox, assuming an average 
specific heat: of 0.8, a variation of one fifth of a degree Celsius would 
correspond to 160 Cals. The rectal temperature affords the best 
available means of control on this point, and a very ingenious 
