In using the above formula, there was presented a misleading interpre- 

 tation of the number of fishermen. The calcialated number was not truly- 

 representative of the number of fishermen, but of the number of "stops" 

 along the shoreline comparable to the duration of the "stops" of fishermen 

 within the complete census areas. The calculated number of fishermen 

 was thus considerably high. By interviewing a sufficient number of 

 anglers to gather data on their average fishing time, it will be possible 

 to adjust this estimate so as to be more representative of the true number 

 of fishermen. 



With the formulae above, estimates for the shoreline fishery for the 

 19^1 season were calculated. Results of the revised 1950 and 19^1 season 

 are shown in table 12, The rates of effort and catch-per-unit-effort for 

 the two years are mean values of rates calculated for all 2-week periods 

 for each year. 



Table 12 

 Besults of shoreline census, 1950 and 1951 



Catch-per-fish- 

 Year Wo, fishermen No. hours No. fish Effort erman hour 



1950 50,260 85,281 55,370 1.79 0.6o 



1951 llli,7i;3 120,651 i;8,965 1.30 O.U 



As there is no method of determining the comparability of the results 

 of the two seasons of shoreline fishing because each was derived by a 

 different method, any conclusions must necessarily be based on the above 

 figures. It is ^parent that a considerable increase in effort was ex- 

 pended in 1951 to catch less fish than in 1950. The rates of catch-per- 

 unit-effort verify this and indicate less fish being available to the 

 shoreline fishermen in 1951 (fig. 10), 



Table 13 is a summary of the 1951 estimates of hours of effort and 

 of catch by 2-week periods. The standard deviations as shown are not 

 derived from actual variances, but rather from variance estir.ates of N. 

 and Nq based on the internal variation within the complete census. These 

 variances fail to include the sampling distribution of the Kj_i, and are 

 therefore, strictly speaking, not actual variances. The random variables 

 used to compute the "variance estimates" of Nh and Nc are, respectively, 

 the length of time fished by each fisherman and the catch by each fisher- 

 man. There is no corresponding random variable for the count of fisher- 

 men, and the calculated number is simply a straight ratio estimate of Nf, 



22 



