114 BULLETIN 908, U. S. DEPARTMENT OF AGRICULTURE. 
Lot 5 was made from waste obtained from small fish from 3 to 5 
inches long (snippers), which would pack from 20 to 25 to a can. 
This waste material contained from 60 to 70 per cent of whole fish 
which had been discarded from the packing table because it was 
decomposed by feed, broken, and damaged. 
The waste material used for lot 6 was composed of small fish which 
would pack 8 to 10 to a can. These fish were very feedy, and had 
been badly broken during transportation to the cannery. The waste 
in these fish as actually determined amounted to from 80 to 85 per 
cent, the largest encountered during the entire investigation. Whole 
flakes full were discarded at a time. 
The results of the analyses made show that a large percentage of oil 
was removed during pressing, while the percentage amount of water 
was not appreciably changed. Where the material was steam cooked, 
the percentage of water increased, and pressing in these instances 
brought the percentage of water down to practically the same amount 
as that present in the raw waste as it came from the packing tables. 
The sum of the constituents—water, oil, protein, and ash—add up to 
practically 100 per cent in all cases in the results on the raw mate- 
rial, but after drying there is a shortage of undetermined matter of 
from 24 to nearly 4 per cent. This difference is in large part due to 
the formation of nonnitrogenous material during cooking and dry- 
ing, and in part to the fact that the factor 6.25, used for calculating 
the total protem from the nitrogen determined, is not exact for fish 
protein. : , 
The average yield of meal, ranging from 5 to 10 per cent of water, 
as determined in these six experiments, amounts to 29 per cent. 
The average would be higher than this by eliminating the last lot, 
in which the waste was steam cooked an excessive length of time, so 
that undoubtedly a great part of the protein material was lost in 
the water which was formed. It is safe to place the yield of mate- 
rial that can be obtained from this waste at 25 per cent under aver- 
age commercial conditions. 
In the case of the oil, cold pressing gave as good a yield as pressing 
the steam-cooked material. Steam cooking or heating before press- 
ing reduces the time required for pressmg. Hxcessive cooking should 
be avoided. The yield of dry meal is apparently decreased by mate- 
rial lost during cooking in the excess water formed by condensing 
steam. The small amount of oil yielded and the comparatively 
small amount retained in the dried meal made from small fish with 
a low fat content might suggest to the prospective producer of fish 
meal that at times in the season, when these fish are plentiful and 
when they are as poor in fat as those composing lots 5 and 6, it 
would not be necessary to press them at all before drying the waste 
for feeding purposes. As much of the oil as possible should be re- 
~ 
