MIGRATION OF SALMON IN COLUMBIA RIVER. 1 39 
will be ignored by the fish within a very short time — a time probably measured by 
minutes. 
The chief objection one can raise here is to assume that the button when once 
inserted acts as a continuous source of stimulation to the individual fish, thus driving 
it into panic. One may assume that the button is not where the fish can see it 
and that it makes no sound which the salmon can hear, granting the questionable fact 
that the fish recognizes unusual sounds. The only other possibility is that the button 
is a continuous source of cutaneous sensory stimulation. This last seems plausible, 
but the fact is that either the wound will heal and adapt the surface to contact with 
the button or the injured surface will begin to degenerate, in which process the local 
nerve endings will soon lose their function and become insensitive. 
Those conditions which lead to the migration of the salmon are the chief directive 
stimuli for the salmon at this phase of its existence. They overshadow all others. In 
comparison with this series of reactions, the so-called migratory instinct, small physical 
injuries are as nothing. If it were not so, the numerous fish that are injured by seals 
or sea lions, that are torn by hooks and the rocks, that are even more profoundly injured 
in the escape from the gill nets, would not appear in such vast numbers on the upper 
fishing grounds of the river. By my own count on different occasions net-injured fish in 
the catch of some of the wheels during the summer of 1908 amounted to from 25 to 60 
per cent of the total, and I am reliably informed that at certain times the per cent may 
run to 80 or 90. My observations indicate that some of the salmon recover from these 
bruises received from the gill nets, though what per cent of recovery occurs I can not 
say. Salmon are, however, frequently taken on the Celilo fishing grounds with injuries 
so profound that one wonders how they could have survived so long, yet these severely 
injured fish are forging ahead toward the spawning grounds. The migratory stimuli 
overshadow even these most profound injuries and continue to do so until death ends 
the struggle, and death must inevitably end the struggle of these unfortunates long 
before the spawning act is consummated. 
DETAILED RESULTS OF EXPERIMENT. 
The location chosen for the marking of the salmon of this experiment is the Wash- 
ington state fish trap, a few hundred yards above the head of Sand Island. The point 
is some 7 or 8 miles within the mouth of the Columbia, on the Washington side, and 
10 or 12 miles below Astoria. The border of the channel above the island is bounded 
by a line which represents the legal limits regulating the setting of fish traps by the 
fishermen. The state trap is located just outside these limits, permission having been 
secured for the location by the Washington fisheries authorities from the United States 
engineers in order to catch fish for the Chinook hatchery. The point also marks the 
limits on the north to the area over which gill-net fishermen drift their nets. In fact, 
gill netters occasionally have their nets caught by the cross currents and thrown on 
this trap. Standing, as it does, just on the border of the north channel on the line that 
separates the gill netters’ field on the one hand from the set traps on the other, this 
trap is especially well located for this experiment. It is in the area of brackish water, 
