68 
Abstract .—Estimates of tag-shed- 
ding and tag-reporting rates are re- 
quired for an estimation of fishing and 
natural mortality rates from tagging 
data. For this purpose, double-tagging 
and tag-seeding experiments were un- 
dertaken by the South Pacific Commis- 
sion, in conjunction with a large-scale 
tuna tagging program, in the western 
tropical Pacific Ocean during 1989- 
1992. Estimates of tag-shedding rates 
indicated that 89% (95% confidence in- 
terval of 82%-94%) of tagged tuna still 
retained their tags after two years at 
liberty. Differences in shedding rates 
among skipjack, yellowfin, and bigeye 
tuna, and differences in shedding rates 
among taggers were found not to be sta- 
tistically significant. Tag seeding car- 
ried out on board purse seiners by ob- 
servers resulted in 342 returns of the 
532 tags seeded, for a return rate of 64% 
(60%-68%). The return rate of seeded 
tags varied significantly by unloading 
location (most tags were recovered dur- 
ing unloading), but not by species. The 
highest return rates of seeded tags oc- 
curred from American Samoa, Philip- 
pines, and Solomon Islands, whereas 
Korea and Thailand had the lowest re- 
turn rates. The overall average report- 
ing rate, weighted by the estimated 
numbers of tags recovered at each lo- 
cation, was 0.59. A bootstrap procedure 
was used to estimate a 95% confidence 
interval of 0.49-0.67. These results 
implied that, of the 146,581 tags re- 
leased during the large-scale tagging 
program, 31,166 (27,208-37,264) were 
recaptured, of which 18,266 were re- 
turned to the South Pacific Commission. 
Manuscript accepted 17 July 1996. 
Fishery Bulletin 95:68-79 (1997). 
Estimates of tag-reporting and 
tag-shedding rates in a large-scale 
tuna tagging experiment in the 
western tropical Pacific Ocean 
John Hampton 
South Pacific Commission 
B.R D5 Noumea Cedex, New Caledonia 
E-mail address: wjh@spc.org. nc 
Tag release-recapture experiments 
are commonly used to estimate pa- 
rameters, such as growth, mortal- 
ity, and population size, of exploited 
fish stocks (Beverton and Holt, 
1957; Seber, 1973). One method 
used to estimate mortality rates is 
to fit a tag-attrition model to a time 
series of tag-return data (Seber, 
1973; Kleiber et al., 1987). In its 
simplest form, the tag attrition 
model can be expressed as 
<j>j =(1 -a)T 
exp [-(F + M + X)(j- D] — ~ — - 
F + M + A 
[l- exp(-F - M - A)], 
( 1 ) 
where (f), is the predicted number 
of tag returns in time period j, a rep- 
resents all type-1 tag losses, T is the 
number of tag releases, F is the in- 
stantaneous rate of fishing mortal- 
ity (assumed constant), M is the in- 
stantaneous rate of natural mortal- 
ity (assumed constant), and A rep- 
resents all continuous type-2 tag 
losses. Type-1 tag losses include im- 
mediate tag shedding, immediate 
tagging-induced mortality, and fail- 
ure to report recovered tags. Type- 
2 tag losses include continuous tag 
shedding, continuous mortality di- 
rectly attributable to the tag, and 
emigration of tagged fish away from 
the area of the fishery. For unbiased 
estimates of F and M to be obtained, 
it is clear from Equation 1 that 
these tag losses must be estimated 
and included in the tag-attrition 
model. 
In general, type-1 and type-2 loss 
rates cannot be estimated directly 
from tag-return data, although es- 
timation of type-1 losses may be 
possible under circumstances where 
fishing intensity is highly variable 
(Beverton and Holt, 1957). More 
commonly, loss rates are estimated 
from independent experiments car- 
ried out in conjunction with a tag- 
ging program. Tag-shedding rates 
may be estimated from double-tag- 
ging (two tags per fish) experiments 
(Wetherall, 1982) or from direct ob- 
servation of tagged fish held in cap- 
tivity. Tag-reporting rates may be 
estimated from tag-seeding experi- 
ments (Youngs, 1974; Green et al., 
1983; Campbell et al., 1992), from 
sequential observations of recover- 
ies at different stages of catch han- 
dling and processing (Hilborn, 
1988), and by comparing tag return 
rates from the fishery with those 
from a control group (such as ves- 
sels carrying fisheries observers) 
assumed a priori to report all tag 
recoveries (Paulik, 1961; Seber, 
1973). Type-1 and type-2 tagging 
mortality rates may, for some spe- 
cies, be estimated from observations 
of tagged and untagged fish held in 
captivity. 
The South Pacific Commission 
(SPC) recently conducted a large- 
