212 
THE BOOK OF FISHES 
great expense and labor, placed across 
Wood River an intercepting rack, which 
compelled all migrating salmon to pass 
through narrow tunnels or gates provided 
for the purpose and so arranged that 
the fish would be readily visible to persons 
on watch. 
Men provided with an automatic count¬ 
ing and registering device were stationed 
on the rack night and day and kept a 
tally of the salmon as they passed up¬ 
stream. The run continued during all of 
July and part of August, and on one day 
over 324,000 fish were recorded, and on 
another more than 402,000. The total 
tally was 2,603,655 salmon, all of the red 
species. 
These were fish that had escaped the 
very active fishing in Nushagak Bay, and 
in addition to them several million other 
fish are known to have gone up other 
tributaries of the bay to their spawning 
grounds, the data available indicating 
that the total run of red salmon in the 
Nushagak basin in 1908 was as many as 
13,600,000, with 10,100,000 as the mini¬ 
mum, of which 6,400,000 were caught 
and utilized at the local canneries. 
Therefore, under the most favorable 
conditions for reproduction, nearly 53 
per cent of the run escaped, and under 
the most unfavorable 37 per cent. 
TAKING A SALMON CENSUS 
During each of the three following 
years the rack was reconstructed at the 
same place, and the census of the run 
was taken in the same way, with the fol¬ 
lowing results: 1909, 893,244 fish; 1910, 
670,104 fish; 1911, 354,299 fish. 
In 1912, the fourth year of the rack 
test, a large run corresponding to the 
run of 1908 was expected. Only 325,264 
fish were counted, however, but for some 
unaccountable reason a big run occurred 
in 1913 totalling 735,109 fish. No work 
was performed during 1914. In the 
following two years the run was low 
according to expectations and then came 
in large in 1917, the fourth year after 
the large 1913 run. Again the count fell 
off In 1918 and 1919. 
It was thought that this experiment 
and similar trials in other streams would 
afford accurate data relative to the 
natural increment of the fish, so that 
the approximate size of the run being 
known, the minimum number of fish neces¬ 
sary to maintain the supply may be 
allowed to pass to the spawning grounds 
each year and the remainder of the run 
placed at the disposal of the fisherman. 
Except for the off year of 1912 the test 
has provided suitable data for such a 
determination. 
