A, P. Armsby—Digestion Experiments. 357 
Sampling.—Since the value of a digestion experiment is 
largely dependent upon the correct sampling of fodders and 
dung, particular attention was paid to this point. 
ach day’s dung was sampled by spreading it out in a shal- 
low tray and taking small portions from different parts of the 
tray until the desired amount was obtained. 
_ The malt-sprouts and cotton-seed meal were sampled at the 
time of weighing out for feeding. The amount required for the 
last four days of the preliminary feeding and the six days of the 
actual experiment was weighed ovt at one time, each day’s feed 
separately. As each one of these twenty portions was weighed 
out, a small amount was laid aside, and the mixture of these 
small portions constituted the analysis sample. 
Substantially the same method of sampling was applied to 
the hay, but as it was desired to ascertain by these trials the 
average composition and digestibility of a considerable quantity 
of hay, the process was somewhat more elaborate. 
The clover hay, to the amount of about three and one-half 
tons, was run through a power feed cutter, using nominally a 
half inch cut, although most of the hay was not actually cut as 
fine as that, The cut hay was spread out on a tight floor and 
thoroughly mixed, being handled entirely with shovels to pre- 
vent, as far as possible, the sifting to the bottom of the finer 
Portions. It was then spread out about two feet deep in an 
analysis sample was taken in the way above described for the 
other fodders. 
? In the experiments with hay alone the fodder for the pre- 
liminary feeding and the actual experiment was weighed out at 
different times, so that we have two analyses of each of the 500 
Pound samples mentioned above. In the subjoined table these 
are numbered 1, 2, 8 and 4, and their close agreement attests 
the accuracy of the sampling. 
