61 
TAGGING SALMON IN ALASKA, 1923 
TO BRISTOL BAY 
The preceding discussion of the rate of travel has not included those fish that 
went to Bristol Bay. As noted above, consideration of this migration has been 
reserved for separate treatment, since the results are seriously affected by the closing 
of the season in Bristol Bay on July 25. Early in the season the average time 
required is about 20 days; later it is about 10 days; but this fact alone is not 
conclusive evidence that the fish are actually traveling faster, since the slower fish 
from the later experiments are eliminated from the records by the closing of the 
season in Bristol Bay. The result is that the average time required later in the 
season is certain to be less, even though the fish were not traveling faster. It has 
been possible, however, by other methods, to show that there is actually an accelera- 
tion of the rate of travel toward the latter part of the season. Tables 36, 37, and 38 
show the time required for the fish to go to Bristol Bay from, respectively, Unga 
Island, Morzhovoi Bay, and Ikatan Bay. The results from various experiments 
are given in separate columns, and it is easily seen that the upper parts of the dis- 
tributions were cut off by the closing of the season on July 25. The heavy line at 
the lower end of the columns indicates this date. Very few returns were reported 
after this date, and the few that were reported later than July 25 were probably 
tags that had been retained by fishermen and were not turned in previously. 
In studying the comparative rate of travel during different parts of the season 
it has been necessary, in order to eliminate the error resulting from the closing of the 
fishing season on July 25, to consider only the first few fish to be reported in Bristol 
Bay from each tagging. In order that the number actually caught might not affect 
the results, the numbers which we have taken from each distribution have been based 
on the number tagged, not on the number retaken. Several methods of analysis 
were tried, but with practically identical results. The simplest and most obvious 
thing to do was to take the time required for the trip by the first arrival. In addition 
to this we took the time that elapsed before 1 per cent of the total number tagged 
were taken in Bristol Bay and the same for 2 per cent and 3 per cent; then we took 
the average time required before 1, 2, or 3 per cent of the total number tagged had 
reached Bristol Bay. We thus had 7 different series of data from each of 3 localities 
in which tagging was carried on, or 21 in all. 
These data all told the same story, showing conclusively that the fish were 
actually traveling at a faster rate later in the season. Figures 10 to 12 present 
graphically the data on the time required by the first arrival and the time required 
for 2 per cent of those tagged to arrive. The time of first arrival is naturally the most 
variable, but is important, since it is subject to no possible error due to the closure 
of the season. The time required for 2 per cent of those tagged to arrive was selected 
from among the other series of data on account of the fact that it appeared to be the 
most stable. The curves shown on the graphs have been drawn in “by eye,” as the 
character of the data did not seem to warrant more refined methods of curve fitting. 
The data on which these graphs are based are given in Table 39, which also shows 
the median time required for the journey from the various points of tagging to 
Bristol Bay and the rate of travel in miles per day. 
44699—27 5 
