FISHERY BULLETIN: VOL. 69, NO. 4 



is based on the data for the years 1946-1957, 

 when the escapement estimates were made 

 largely from ground and aerial surveys. Pre- 

 sumably this estimate is less reliable -than the 

 second estimate based on the years 1958-1966, 

 when the escapement estimates were based 

 largely on very reliable tower counts. The two 

 ratios are 1.4438 and 1.4433. 



Thus the average Wood River escapement has 

 formed a remarkably constant proportion of the 

 total Nushagak escapement regardless of which 

 years are used for calculation. Consequently, 

 the ratio 1.4436, based on data for all the years 

 1946-1966, has been used to enlarge the early 

 Wood River weir counts from 1908 to 1919 to 

 reflect the Nushagak escapement for the same 

 years. 



Estimated in this manner, the annual Nusha- 

 gak escapement for the period 1908-1919 aver- 

 aged 1,126,113 spawners, and the total Nushagak 

 sockeye run averaged 5,771,526 salmon, or a rate 

 of exploitation of 80.5%. 



THE PERIOD 1925-1945 



No weir counts exist for these years, and quan- 

 titative stream surveys were not conducted. 

 However, escapement estimates can be made 

 from available data on the catch, size distribution 

 of the fish, sex ratio, and expended effort. 



From 1926 on, the legal minimum mesh size 

 was 51/2 inches, and from 1927 the maximum 

 length of the drift gill nets i-emained unaltered 

 at 150 fm. An exception must be made for 1937, 

 when the maximum length was reduced to 100 fm 

 for this one year in the Nushagak fishery. There- 

 fore, given a measure of the catch and the in- 

 stantaneous rate of fishing for each centimeter 

 group by the 5i/.>-inch gill nets and an estimate 

 of expended fishing effort, the escapement can 

 be calculated by centimeter groups from the 

 formula for competitive fishing units and 

 summed over the size range observed in a year 

 to give the total escapement: 



E 



h 



e '"i 



1 



where E = the unknown total escapement, 



Cj = the known catch for size group /, 



Qi = the coefficient of catchability for 



size group ./, 

 / = the number of standardized fish- 

 ing units, and 

 a and b = lower and ui)per bounds of the 

 size range. 



No natural moi'tality has been assumed during 

 the fishing season. 



Because of the different selection curves for 

 males and females by 5V2-inch mesh size, these 

 calculations must be done separately for each 

 sex. The necessary data for this calculation 

 follow. 



Sex Ratios 



It has been assumed that no selection for sex 

 was exerted in the collection of samples for size 

 and age composition. Consequently, the num- 

 bers of males and females measured in a day 

 provide an estimate of the sex ratio in the catch 

 for that particular day. This procedure was 

 necessitated by the absence of specific sex ratio 

 samples. 



Size Composition of the Catch 



During the years considered here, the Bureau 

 of Fisheries stationed biologists at selected can- 

 neries for collection of scale samples and length 

 measurements. At other times, resident people 

 were hired for the same purpose and paid a fixed 

 amount for each scale book collected. 



Generally the type of length measurements 

 made is not indicated in the records; but it 

 has been assumed that the procedure was to 

 measure length from the tip of the snout to the 

 fork of the tail. This assumption was verified 

 by a comparison of the resulting length-fre- 

 quency curves with the mean lengths of 2- and 3- 

 ocean fish in postwar years. 



Since 1946, the common procedure has been 

 to measure the length of the sockeye salmon 

 taken in the fishery from the middle of the eye 

 to the fork of the tail. The Fisheries Research 

 Institute took a series of double measurements 

 in 1946 to provide a basis for constructing a i-e- 

 gression line between the two types of measure- 

 ments, and a conversion can be made from one 



754 



