Reliability of the Estimates 



For the estimates obtained directly from a 

 regression equation it is easy to compute 

 confidence interval estimates to suggest their 

 reliability. Final estimates, however, are ob- 

 tained by additional adjustment which add addi- 

 tional variability. On the other hand, the 

 agreement between estimates obtained in dif- 

 ferent ways yields additional confidence. 



Some confidence intervals are: 



For the estimate of the 4-year-old male 

 kill to 31 July by method 1, Y lies between 

 16,900 + 10,500, that is, between 6,400 and 

 27,400 with 95 percent confidence. 



For the estimate of the adjusted kill of 

 the 1961 year class by method 2, the 95 per- 

 cent confidence interval for Y is 33,800 + 

 23,000, that is, from 10,800 to 56,800. 



The width of these confidence intervals 

 should indicate the caution that must still be 

 exercised in regard to forecasts of the male 

 harvest. 



Final Predictions by Ages and by Islands 



The estimates given above are averaged to 

 yield St. Paul Island figures. Further, the 

 long-term relation between St. Paul and St. 

 George Island in regard to the male kill is 

 the basis of the extrapolation of the estimate 

 to St. George Island. It should be pointed out, 



however, that in 1964 as in 1963 the fraction 

 of the kill on St. George Island was higher 

 than the long-term average. The average of 

 the St. George Island kill to the total for both 

 islands for the period 1950-64 is exactly 20 

 percent. The results are shown in appendix 

 table 4. 



Appendix table 4. --Prediction of the kill in 1965, by age, 

 Pribilof Islands, Alaska 



Literature Cited 



CHAPMAN, DOUGLAS G. 



1964. A critical study of Pribilof fur seal 

 population estimates. U.S. Fish and 

 Wildlife Service, Fishery Bulletin, vol. 

 63, no. 3, p. 657-669. 



ROPPEL, ALTON Y., ANCEL M. JOHNSON, 

 and DOUGLAS G. CHAPMAN. 



1965. Fur seal investigations, Pribilof Is- 

 lands, Alaska, 1963. U.S. Fish and Wild- 

 life Service, Special Scientific Report- - 

 Fisheries No. 497, 60 p. 



29 



