336 
Fishery Bulletin 115(3) 
LD 
< 
E 
Average 
Figure 4 
Comparison for fixed allocation and proportional allocation of the within-length interval variance, (A) 
for fixed allocation and (B) for proportional allocation, and of the between-length interval variance 
for the species investigated (estimated with aging error [AE1] by using data from the NOAA Alaska 
Fisheries Science Center bottom trawl surveys for the Gulf of Alaska [GOA, 1984-2011], Aleutian 
Islands [AI, 1980-2010], and Bering Sea [BS, 1982-2011]). Barplots indicate the averages of (C) the 
between-length variance and within-length interval variance, (D) for fixed allocation and (E) for pro- 
portional allocation for species types. 
tal biomass in the final year was correlated with the 
ratio of recruitment variability divided to the survey 
index uncertainty, which was 0.96 (Fig. 7C). 
Discussion 
When thinking about the most efficient and appropri- 
ate manner to distribute sample sizes for age across 
multiple species, whether in a fishery-independent sur- 
vey or from a commercial or recreational fishery, there 
are a number of factors to consider. All of these can be 
reduced to answering a single question: What is the 
relative value of each additional otolith to stock assess- 
ment? Although this can be posed as a simple ques- 
tion, it is not simple to answer. It is an optimization 
problem that requires balance among biology and life- 
history, statistics and stock assessment, and the com- 
mercial value and importance of the fisheries to user 
groups. Although this study is the first attempt in the 
fisheries literature to directly address this question, 
we recognize that many more factors than those con- 
