
DIGESTIBILITY OF FISH AND POULTRY. NaS; 
_ the digestibility of these materials in the ordinary mixed diet, 
than would experiments in which the diet was limited to a 
single material. 
Description of food materials.—In these experiments, well 
known brands of canned salmon and canned chicken were pur- 
chased at a local grocery. No preparation was necessary, other 
than the removal of a small amount of bone and other refuse 
from the chicken, and the materials were then eaten as pur- 
chased. 
As a type of lean fish, fresh cod was selected. Several 
fresh cod steaks were purchased each day, cut into small 
pieces and prepared by frying. The pan was slightly greased 
with butter, and the fish thoroughly mixed with a knife during 
the cooking. 
The duck used as a type of fat poultry was purchased un- 
cooked, and prepared by roasting. The fat which cooked out 
in the process was carefully saved, and constituted the ‘‘duck- 
fat’’ hereafter referred to. 
The bread was ordinary white bread purchased of a local 
baker, and was left at the laboratory each day. Immediately 
after purchase, it was stored in large glass jars in a cool place. 
The milk used was from a mixture of the total product of 
several cows of a local herd. Milk from the same cows was 
included each day, as previous experience had indicated little 
variation in composition of such milk from day to day. 
The butter was purchased at a local grocery, a fresh supply 
being obtained for each series of experiments. 
The sugar was granulated cane sugar and was obtained at a 
local grocery. The rock-candy was a portion of a supply kept 
on hand for combustion work with boinb calorimeter. 
- Sampling of food.—As previously stated, the chicken and 
salmon were purchased incans. At the beginning of the series 
of experiments, each subject was given one of these cans. At 
the first meal, each subject opened his can, transferred the 
contents to a large weighed bowl, and secured as thorough 
mixing as possible by stirring with a large spoon. Portions 
as nearly equal as could be easily estimated were then taken 
from each bowl for a sample for analysis. This procedure was 
