364 H. P. Armsby—Digestion Fuperiments. 
PropasBLE Errors OF THE RESULTS. 
and dung. The hay was weighed out upon a balance sensitive 
n. 
favorable to extreme care in weighiny we will assume 0'L 
gram as the probable error of a single weighing. Then, since 
the weight of the hay was obtained by subtracting the weight 
of the empty from that of the filled bag, the probable error 1m 
the weight of hay for one day is + 0-1? +0-1'=014 grins. 
In like manner, the weight of fresh dung daily excreted 1s 
subject to a probable error of + 0°14 grams. As, however, 
some degree of accuracy for the determinations of dry matter 
nitrogen and crude fiber. For the organic matter and fat the 
data are less complete, and owing to the less importance which 
attaches to these determinations, the computations have not 
been attempted. Nitrogen-free extract has also been excluded 
rom the computation, because, being determined by differences 
all the errors of the other determinations are included in It. 
The amount of dry matter fed in Periods I and II is deduced 
aking this small number as the basis of a calculation of 
probable error we obtain the following results : 
