The other-loss rate for both tagged and untagged fish 

 includes natural mortality and outmigration. In addi- 

 tion, the other-loss rate for tagged fish is augmented 

 by tag dropoff and any additional natural mortality 

 caused by the tag. Although no information was 

 available on the relative importance of the different 

 contributors to the other-loss rate, it is believed that 

 escapement to spawning streams was the dominant 

 factor. Continuing tag dropoff may have been a 

 significant factor in these experiments. The type of 

 tag used was a strap tag attached to the dorsal edge of 

 the caudal fin. As mentioned above, the fishing mor- 

 tality rate would be underestimated if there were 

 either immediate tagging mortality or incomplete 

 recovery and reporting of recaptured tagged fish. 

 Original correspondence which accompanied some of 

 the tag returns indicated that there was substantial 

 nonreporting of tags during these experiments. Unfor- 

 tunately there is not sufficient information to es- 

 timate the magnitude of this type of error. In the opin- 

 ion of the authors the estimated fishing mortality 

 rates given in Table 16 are too low. 



The estimates of F and X given in Table 16 were 

 calculated from the values of ^ and s under the 

 assumption that the fishery during the years of these 

 experiments was periodic; i.e., the fishing season was 

 open for 5.5 days and closed for 1.5 days of each week. 

 The method of calculation will be outlined briefly: 



Let F = weekly exponential fishing mortality rate, 

 X = weekly exponential other-loss rate, 

 Ai = length of first part of weekly open period, 

 A2 = length of weekly closed period, 

 A3 = length of second part of weekly open period, 

 and 



A, + A, -1- A3 = 1. 



Then 



_ -(f(Ai+d2)+X) 



F ., -(f + X)A, 



F + X 



+ e 



(Fa,+X (Ai+i2)) F .r, ^ 



£ (1p-(f+X)A3) 



F + X^ 



Since the length of the weekly closed period was 1 '2 

 days, A2 = 1.5/7 and A, + A3 = 5.5/7. It was assumed 

 that A3 = for releases on Sunday and that Ai = for 

 those on Saturday. Values of A, and A, for midweek 

 experiments were assigned on the assumption that 

 releases were made exactly at midnight of the tagging 

 day. The two nonlinear equations shown above were 

 solved for the unknowns F and X by an approximation 

 technique using nomographs. 



The estimates for F and X for each of the 30 experi- 

 ments together with their averages by area and year 

 are listed in Table 16. The F values varied from 0.207 

 to 1.208 and averaged 0.514. The X values ranged 

 from 0.234 to 1.287 and averaged 0.595. Thus, for 

 these experiments the estimated rate of removal of 

 tagged fish by the fishery was slightly less than the 

 rate of removal by all other causes. 



Estimates of F and X computed from the results of 

 the tagging experiments conducted in 1950 in Icy 

 Strait by Elling and Macy (1955) were reported by 

 Paulik (1963). The opening date for the commercial 

 fishing season in 1950 (15 August) was later than the 

 closing dates of the seasons during the years 1938 

 through 1945. In 1938-45 the season in Icy Strait was 

 usually opened about 20 June and closed the first 

 week in August. Elling and Macy (1955) tagged fish 

 on consecutive days from 25 July through 14 August. 

 Using their data Paulik (1963) obtained an in- 

 traseasonal weekly X value of 0.641 and a weekly F 

 value of 1.230. Although the value for X in 1950 agrees 

 fairly well with the overall average of the X values for 

 the tagging experiments in 1938-42 and 1945, the 

 1.230 value for F in 1950 is considerably higher than 

 the average F of 0.514 for 1938-42 and 1945. We 

 suggest that the discrepancy between these F values is 

 due in a large degree to nonreporting of recaptured 

 tags in the earlier experiments. 



In 1941 the X rate increased as the release site was 

 moved from Icy Strait to middle Chatham Strait 

 (Table 16) . The F values appear to be highest for mid- 

 dle Chatham Strait. Although the intraseasonal 

 variability in F and X values was extremely high, 

 there appeared to be an increasing trend in the values 

 of F and X during the season. 



To estimate the fraction removed by the fishery of 

 the total run entering the northern part of 

 southeastern Alaska during a given year, it is 

 necessary to take into account the pattern of recruit- 

 ment of pink salmon into the fishing areas during the 

 fishing season. A computer program which calculates 

 the rate of exploitation as a function of the fishing 

 mortality rate and the temporal pattern of fishing 

 (i.e., opening and closing dates and the number of 

 days of fishing per week) when there is intraseasonal 

 recruitment is decribed by Paulik and Greenough 

 (1966). To use this program it is necessary to assume 

 that the recruitment into the fishing area can be ap- 

 proximated by a triangular entry pattern. Data input 

 to this program includes the times of the beginning, 

 the peak, and the end of the entry pattern; the begin- 

 ning and the end of the fishing season; and seasonal 

 variations in the fishing mortality rate and the other- 

 loss rate. 



The cumulative catch at any time during the season 

 is calculated by means of two simple algorithms: 



34 



