72 
Fishery Bulletin 95 ( I ), 1997 
there were local tag releases that resulted in most of 
the tag returns from those locations. These local tag- 
return rates (0.126 and 0.223, respectively) were used 
as the lower bounds for the reporting rate distribu- 
tions for Solomon Islands and Philippines. For the 
other locations, it was not possible to identify sets of 
local releases to calculate local tag-return rates be- 
cause the release locations of the local returns were 
widely distributed throughout the tag release area, 
not just in the vicinity of the unloading ports. In these 
cases, I made the minimal assumption that the “lo- 
cal” releases comprised all tag releases except those 
returned from other locations. Thus, the shapes of 
the reporting-rate probability distributions are de- 
termined by the tag-seeding data and by this notional 
minimum possible return rate. Note that the means 
and medians of such distributions could be quite dif- 
ferent from the tag-seeding sample means p r In 
general, the differences will be greatest where is 
close to 0 or 1 and n is small. 
For returns from locations where no tag-seeding 
data were available or for tag returns that could not 
reasonably be pooled with purse-seine returns be- 
cause the recovery processes were different (14% of 
all returns), values of p' were sampled from uni- 
form distributions [7(0.5, 1.0). While somewhat arbi- 
trary, this procedure is meant only to reflect some 
knowledge of the minimum possible reporting rate 
from these locations. In fact, this assumption prob- 
ably understates the likelihood of tags being reported; 
almost all of these tags were recovered in Indonesia 
and in Pacific Island countries, where widespread 
publicity and the attractiveness of cash rewards are 
likely to have resulted in high reporting rates. 
Results 
Tag shedding 
In all, 4,541 tuna (2,557 skipjack, 1,493 yellowfin, 
and 491 bigeye) were double-tagged during the RTTP. 
Return rates of double-tagged tuna were comparable 
to those of single-tagged tuna. 
Returns from 525 double-tagged tuna were avail- 
able for analysis. Fitting the model specified in Equa- 
tion 2 to the pooled data provided estimates of p and 
L that, according to the model, would result in 89% 
(95% confidence interval of 82%-94%) of the origi- 
nal tags being retained after two years at large (Table 
1) . Both p and L were significantly different from 
zero (P<0.001). 
Fitting the model to the three species separately, 
although yielding somewhat different tag-retention 
rates (Table 1), did not result in an overall statisti- 
cally significant improvement in fit (P=0.334, Table 
2) . Similarly, there were differences in tag-shedding 
estimates for the different taggers (Table 1), but over- 
Table 1 
Double-tagging results (m is the number of returns bearing two tags and n is the number of returns bearing one tag), tag- 
shedding parameter estimates, the estimated proportion of tags retained after two years at liberty (Q 2yr ), an< i likelihood function 
values (ft) for fits of the tag-shedding model to the various data sets. 
Tag-shedding parameters 
95% confidence 
Data set 
m+n 
m 
n 
P 
L (/mo.) 
^2 yr 
interval for Q 2yr 
ft 
Pooled data 
525 
457 
68 
0.05861 
0.002312 
0.89 
0.82-0.94 
201.498 
Skipjack tuna 
241 
211 
30 
0.03485 
0.007179 
0.81 
0.68-0.93 
87.637 
Yellowfin tuna 
204 
176 
28 
0.06574 
0.001532 
0.90 
0.81-0.95 
81.434 
Bigeye tuna 
80 
70 
10 
0.06667 
0.000000 
0.93 
0.74-0.97 
30.142 
Total 
525 
457 
68 
199.213 
Tagger 1 
42 
40 
2 
0.01111 
0.002788 
0.92 
0.76-1.00 
7.909 
2 
45 
36 
9 
0.07348 
0.009636 
0.74 
0.42-0.93 
22.023 
3 
45 
39 
6 
0.07143 
0.003400 
0.86 
0.76-0.94 
17.670 
4 
53 
47 
6 
0.06000 
0.000000 
0.94 
0.68-0.98 
18.718 
5 
106 
94 
12 
0.03604 
0.004257 
0.87 
0.73-0.96 
36.533 
6 
68 
57 
11 
0.08800 
0.000000 
0.91 
0.70-0.95 
30.096 
7 
81 
67 
14 
0.00000 
0.018460 
0.64 
0.50-0.85 
34.184 
8 
60 
53 
7 
0.03566 
0.007403 
0.81 
0.55-0.97 
21.110 
Other taggers 
25 
24 
1 
0.00000 
0.003059 
0.93 
0.78-1.00 
3.600 
Total 
525 
457 
68 
191.843 
