THE UTILIZATION OF ENERGY. 545 
mean only one of two things, viz., that the figures for the work done 
are too high or that the deduction on account of the crude fiber is 
too great. 
That a determination of the equivalence of food and work by 
Wolff's method is subject to considerable uncertainty in an indi- 
vidual case is obvious, but there seems to be no apparent reason 
why it should be uniformly overestimated. The measurement of 
the work was made with great care, and while the work of locomo- 
tion is an estimate, its close agreement with the results of Zuntz & 
Hagemann (p. 539) renders it unlikely that it is seriously in error. 
It would appear, then, that with the rations used in these ex- 
periments the energy required for digestion and assimilation was 
less than the energy of the digested crude fiber. How much less it 
was, however, unfortunately does not appear, and we are obliged 
to content ourselves for the present with this negative conclusion. 
Zuntz & HaGEMANN’s CompuTaTions.—These investigators * 
have recalculated Wolfi’s results in a still different manner. In- 
stead of taking for the amount of work equivalent to the ration the 
figures given by Wolff, which, as already explained, are to a certain 
extent estimates, they take the amount of work actually performed 
in each case and correct for the observed gain or loss of live weight. 
This method is in conception more scientific than Wolfi’s, pro- 
vided the requisite correction can be accurately estimated. As the 
basis for such an estimate, Zuntz & Hagemann take an early experi- 
ment by Wolff,t from which they compute that one gram loss of 
live weight is equivalent to one half revolution of the dynamometer 
(at 76 kgs. draft). From the same experiment they compute the 
mechanical equivalent of one revolution as 2694 kgm. This, how- 
ever, aside from the fact that it is the result of a single series of 
experiments, was obtained with the old form of dynamometer, 
whose indications, as we have seen, were too high, but the later 
experiments unfortunately are not reported in a way to permit of 
an estimate of the difference. 
Taking the correction, then, as estimated, Zuntz & Hagemann 
divide Wolff’s experiments into two groups, viz., those in which the 
work was 400 or less revolutions and those in which it was more 
* Loc, cit., pp. 412-422. 
+ Grundlagen, etc., p. 80. 
