304 
BULLETIN OF THE BUREAU OF FISHERIES 
amount of shrinkage undergone in the kippering of the herring prior to canning, 
but the total shrinkage here adopted is considered to be a very close estimate by 
one of the packers who has canned about two-tliirds of all that has been prepared 
in this manner in Alaska. Any errors due to inexactness of the conversion factor, 
will affect only a few years, as the total amount canned was never large except in 
1919. 
All of the oil and fertilizer produced in southeastern Alaska from 1882 to 1918 
were made by one plant at Killisnoo, and the records of the raw herring used are 
available up to and including 1911. From 1912 to 1920, when this company failed, 
44.9 pounds of raw herring were allowed for each gallon of oil, the factor being calcu- 
lated for the period from 1909 to 1911, inclusive (a new cooker was installed in 1909, 
United States Senate, 1912). For the rest of southeastern Alaska from 1919 to 1928 
(and for the new company that organized to run the Killisnoo plant since 1923) 
the raw herring used for reduction has been calculated by allowing 50 pounds of raw 
herring per gallon of oil and 65 pounds of raw herring per pound of fish meal. The 
two results were then added and divided by two as the meal and oil came from the 
same fish. These two factors were calculated from careful records kept by the plant 
at Red Bluff Bay from 1922 to 1927, inclusive. In Prince William Sound the con- 
version factors used were 61.2 pounds of raw herring per gallon of oil, and 8.25 pounds 
of raw herring per pound of meal, as calculated from careful records kept from 1923 to 
1927, inclusive, by a plant in Evans Bay. The factors vary some from year to 
year, but it was found that the factors for oil and meal hardly differed in their relia- 
bility, the coefficient of variation ( a ) for the oil factor being 0.213; for the 
J V Mean/ 
meal factor 0.162. 
For the pickled products, 20 per cent shrinkage of the raw herring has been 
allowed in all districts; that is, the finished product must be increased 25 per cent 
to represent the raw herring actually used in its preparation. (Figs. 41 and 42.) 
In addition, certain allowances must be made for waste of small herring. No waste 
is allowed for herring pickled by companies operating reduction units, as this waste 
is included with the raw herring for the oil and meal. For the other companies 
varying allowances have been made in the different districts. 
The waste allowed in southeastern Alaska previous to 1918 was only 20 per cent 
as practically all of the herring were cured by the Norwegian method in which quite 
small herring are used. To allow for this, the amount of raw herring actually used 
in the finished product was increased another 25 per cent. From 1918 to 1928 the 
records differentiate between the Norwegian and Scotch cured products. During 
this period the waste on the Norwegian cure is the same, but 40 per cent waste was 
allowed on the Scotch cure, which utilizes only the larger fish. 
In Prince William Sound accurate records are available for the amounts of raw 
herring wasted by a company without a reduction plant in 1922 and 1923. In 1922 
they wasted 18 per cent; in 1923, 26 per cent; which means that the raw herring 
actually used for pickling were increased by 22 per cent and 33% per cent, respec- 
tively, to allow for waste. All of the packers are agreed that the waste in Prince 
William Sound was less in 1922 than in any other year, so the 1923 factor was applied 
to all of the other years. In Prince William Sound, in the later years especially, a 
portion of the waste of the companies without reduction plants (of which no accurate 
