PROGRESS IN BIOLOGICAL INQUIRIES, 1926 639 
other destructive factors, would vary with the species, with the stream, and 
with the season. It could confidently be expected that certain seasons, in a 
given stream, would be more favorable than others, and it might well happen 
that a smaller spawning reserve in a favorable year would produce a larger 
ultimate run than would be produced by a much larger spawning reserve in a 
year less propitious. These fluctuations, due to natural causes, are inde- 
pendent of the size of the spawning colony and form one of the greatest 
obstacles to a scientific administration of the fisheries. Through them, even the 
fact of depletion becomes difficult of detection, and the trend of the fishery can 
be determined only through observations extending over a considerable period 
of time. They make difficult, also, the solution of our special problem here 
considered, and will produce such a variation in our results, from year to year, 
as will yield us nothing more than an average expectation on which to base our 
procedure. Even this, however, will be invaluable. We shall have established 
what is a normal or average expectation of yield from a given number of 
spawning fish in the stream in question, and we shall have a basis for exact 
determination of the full swing of the pendulum of natural fluctuations in the 
yield of these colonies, a basis that heretofore has been lacking. 
We selected as our species the red salmon of Alaska (Onoorhynchus nerka)— 
the most valuable species, commercially, and the one concerning the life history 
of which the mest is known. It develops spawning runs in such streams only 
as have one or more available lakes in their course, and spawns almost ex- 
clusively along the gravelly shores of the lakes or in their tributary streams. 
The young, after hatching, seek the deeper waters of the lake and live there 
for varying periods before they enter on their downward migration to the sea. 
In selecting streams for our experiments, it was necessary to find such as 
admit annually of a complete census of the run, including that portion that 
is taken commercially, as well as the portion that escapes the fishermen and 
ascends to the spawning grounds. Such streams are difficult to find, for usually 
at some portion of their spawning migration the mature fish mingle with those 
bound for other rivers and are there the subject of an important fishery. It is 
impossible to segregate the different races involved in such a fishery, and the 
commercial part of the yield is therefore left in doubt. Fortunately, in the 
Karluk River, on Kodiak Island, and in the Chignik River, on the southern slope 
of the Alaska Peninsula, we have two important red-salmon streams, the runs 
to which are substantially free from capture until they reach the vicinity of 
their respective mouths, where the fisheries are located. These two streams 
have been selected for our experiments. Agents of the bureau collect daily 
statistics of the commercial catch, and the spawning escapement is ascertained 
at counting weirs, one of which is operated in each of these rivers. The Karluk 
weir has been maintained yearly since 1921, and the Chignik weir since 1922. 
Such a weir consists essentially of a picket fence extending entirely across 
the river, with a number of adjustable gates, through which the ascending: 
salmon are permitted to pass, virtually in single file. Employees of the 
bureau are stationed at these gates day after day throughout the season and 
count the fish as they pass through. The run in the Karluk River begins in, 
May and continues uninterruptedly into October. During this long period 
the watchmen are constantly on duty, shutting the gates during the darker 
hours of the night or for limited periods in the daytime when the run is. 
slack; but otherwise, in wind and rain-and the searching cold of the Alaska 
summer, they stand on the weir and count the salmon streaming through the. 
gates. Although the run is constantly in progress during the summer months, 
the numbers that pass on different days varies widely, from a few hundred 
on some days to over 100,CO00 on others. The magnitude of the task becomes 
apparent when it is considered that in the season of 1926 over 2,500,000 red 
salmon were counted through the Karluk weir. 
If all the progeny of a given year matured at the same age and returned 
in the spawning run of a single year, our problem would be a relatively simple. 
one. Unfortunately, this is far from being the case. They may mature at 
any age from 3 to 7 years and return mingled in widely varying proportions. 
with salmon of different age derived from other brood years. To ascertain 
the yield of any single year’s spawning, therefore, it becomes necessary to. 
analyze the runs of all the different years in which the progeny may appear. 
The first year the Karluk weir was in operation was 1921, when about 1,590,000: 
spawners ascended the river to the lake. Part of these returned as 5-year fish, 
in 1924, as 4-year fish in 1925, and as 5-year fish in 1926. Others may be. 
expected as 6-year fish in 1927 and as 7-year fish in 1928. 
