MIGRATION OF SALMON IN COLUMBIA RIVER. 
147 
way. He applied this principle in his studies of the ehinook salmon of the Sacramento 
River. Following the variations in the catch of the fisheries at the different towns 
along the bay and lower Sacramento, he estimated that a school of salmon made its 
way from Vallejo, on the lower bay, to Sacramento, on the river, in 4 days for the spring 
run when the river is relatively high. In the summer and fall they move more slowly. 
This he explains bv the fact that the river is low and the tides in the bay therefore more 
nearly equal in time, thus requiring more time for the salmon to pass through the bay. 
My fish were marked in August, hence are to be compared with the movements of 
fall fish as described by Rutter. I accept Rutter’s hypothesis as partially explaining the 
movements of salmon in tide water. Undoubtedly currents in the rivers are directive 
on the movements of the migratory fishes. In tidal waters this factor is still active. 
In the tidal area at the mouth of a river the relative time of the flood and ebb currents 
rapidly changes toward the upper tidal limits, where the former entirely disappears. If 
salmon were directed by currents alone they would make the journey more and more 
continuously as they come within the brackish area. Figured on the basis of the dif- 
ference of the duration of the flow of the flood and ebb currents as against the observed 
speed of salmon, it is obvious that the fish would pass through the tidal area in a much 
shorter time than these observations indicate. Other factors are operative, for currents 
alone are not sufficient to account for the movements. I believe that a much more 
influential factor is the condition of the water as regards its amount of salt. Salmon 
are sharply responsive to the stimulus that comes from variation in the degree of admix- 
ture of sea water and river water in the tidal area, a stimulus that is doubtless in the 
nature of a negative chemotaxis. Attention has already been called to the changes in 
the osmotic equivalents of the blood in fresh-water salmon as compared with those in sea 
water. These changes, though slight, are due in large measure to the transition from a 
sea-water environment to one of fresh water. Such physiological adaptations require 
a relatively long time. If a salmon entering the mouth of the Columbia should swim 
into an area of water relatively fresh before his gills and other epithelial tissues were 
sufficiently adapted to it, chemotactic reaction would stimulate him to increased activity, 
which, by the law of such reactions, would lead him in the end toward salt water. These 
journeys into areas now relatively fresh, now relatively salt, but in the balance ever 
toward fresh water, will continue until the epithelial tissues of the individual fish have 
become adapted to life in fresh water. The rate at which this adaptive process takes 
place determines the total time required for the passage through the tidal area. The 
observations recorded in this experiment indicate a very much longer time spent in 
tide water by the salmon on the Columbia River than allowed by Rutter for salmon 
on the Sacramento. While not numerous enough and not sufficiently varied to make 
the deductions absolutely conclusive, yet these experiments strongly indicate that 
salmon spend not less than from 30 to 40 days in passing the tidal area of the lower 
Columbia. 
