LAURS ET AL.: TAG SHEDDING BY ALBACORE 



Table 2.-Estimates of rates of tag shedding, L (on an annual 

 basis), retention, p, from 1972 North Pacific albacore double-tag 

 study- 



Table 3.-Tag releases and returns from North Pacific albacore 

 single-tag studies. 



Item 



Undated returns excluded 

 Undated returns included 



0.098 

 0.086 



0.88 

 0.88 



could not be assigned to seven double-tag and two 

 single-tag returns in 1972 and one double-tag 

 return in 1973. We assumed that 4 was the same 

 for the returns shown in Table 1 and the returns 

 with unspecified recovery dates and included the 

 10 additional returns in a recalculation of p and L 

 The results of the recalculations are similar to the 

 original (Table 2). We estimated p to be about 0.88 

 and L on an annual basis to be between 0.086 and 

 0.098. This means that if no mortality occurs, 8.2 to 

 9.3% of all unrecovered tags are expected to be lost 

 through shedding annually. 



Our estimate of p is similar to the results 

 obtained for yellowfin tuna (p = 0.913) by Bayliff 

 and Mobrand (1972) and bluefin tuna (p = 0.973) by 

 Lenarz et al. (1973). However, our estimate of L is 

 considerably lower than that obtained for yel- 

 lowfin tuna (L = 0.278) and bluefin tuna (L = 

 0.310). 



Methodology for estimation of the variance of L 

 and p when only two periods of recovery are 

 available has not been published. However, we 

 believe that the number of tag returns available 

 for this study is too low for accurate estimates of p 

 and L. We made the following calculations to 

 illustrate the relative level of accuracy. If we 

 arbitrarily assume that the returns of double- and 

 single-tagged fish in 1973 were from a binomial 

 distribution with the probability of a returned fish 

 having only one tag being 0.5, the probability of 

 having 8 or fewer fish returned with only one tag 

 out of a sample of 22 fish from such a population is 

 about 0.14. If 11 fish were returned with single 

 tags (the expected value from the assumed dis- 

 tribution) instead of the 8 observed, our estimates 

 of p would be 0.895 and our estimate of L would be 

 0.172. Thus it appears that there is a reasonable 

 chance that our estimate of L (about 0.09) could be 

 considerably lower than the true value. 



We are not aware of any other data available 

 from double-tag studies on albacore. However, 

 there is a considerable amount of data available 

 from single-tag studies conducted in recent years 

 on albacore in the eastern North Pacific (Table 3). 

 Return rates in the year after release were 0.018 



for the 1971 releases, 0.036 for the 1972 releases, 

 and 0.033 for the 1973 releases of single-tagged 

 fish, for an average of 0.029. If the return rates are 

 divided by 0.88 to account for Type-I tag shedding, 

 the average becomes 0.033. The return rate in the 

 year after release for the double-tag study was 

 0.027. If the rate is divided by 0.99 (1 - (I - p)'^) to 

 account for Type-I shedding of both tags, the 

 return rate is 0.027. Thus the return rates from the 

 single-tag studies give further evidence that 

 Type-II shedding is insignificant, because if it 

 were not, return rates adjusted for Type-I shed- 

 ding from the single-tag releases should be lower 

 than return rates from the double-tag releases, 

 provided mortality rates were similar for these 

 years. 



The above estimates are based on the assump- 

 tion that all double-tag recoveries are reported as 

 double-tag recoveries. A possible source of error is 

 that some fishers may return only one tag from a 

 double-tag recovery. These fishers might return 

 only one tag because of their interest in albacore 

 migrations, but retain the second tag as a souve- 

 nir. This would result in our underestimating the 

 value of p. To illustrate the extreme case assume 

 that p is actually 1.0, but we estimate it to be 0.88 

 because of incomplete reporting. Then assuming p 

 = 1, Equations (1) and (2) become 



(4) 



UMk = FTA^ovr5e-(^+^^ + 2LK. 



and 



na^K - 2F T iVo TT (1 - e-^'k)e-(^ + ^ + ^'^ + (1 - 5) 



FrNo-rre-^^^^^^^^''- (5) 



where B = minimum proportion of double-tag 

 recoveries that are reported as double- 

 tag recoveries. 



Manipulation of Equations (4) and (5) results in 

 («dd2 + ^^ds2) (nddi) 2e^'2 - 1 



{riddi + Wdsl) i'Hdd2) 



2e^'i-l 



(6) 



and 



677 



