212 
Fishery Bulletin 96(2), 1998 
nual discard by the U.S. fleet during the study pe- 
riod (1989-92) for observed patterns and to predict 
discards that would have resulted if sets on dolphins 
were redirected to school or log sets. We then predict 
annual average discard from the international fleet 
(of which the U.S. was only a small part) under the 
same redistributions of effort and discard rates as- 
sumed for the U.S. fleet. 
Although providing the ultimate impetus for the 
current study, we consider here only briefly the 
change in dolphin bycatch that would result from 
replacing dolphin sets with school and log sets be- 
cause the effect is quite easy to estimate and is cur- 
rently almost irrelevant in terms of dolphin popula- 
tion dynamics. The effect is easy to estimate because 
dolphin mortality is almost nonexistent in school and 
log sets. Thus replacing dolphin sets with school or 
log sets (or both) will virtually eliminate fishery-re- 
lated dolphin deaths. This result is irrelevant in a 
population dynamics sense because annual fishery- 
related dolphin mortality is currently less than 5% 
of estimated annual replacement rates (compared 
with dolphin mortality rates approaching 100% of 
annual replacement rates during most years prior 
to 1993). Current mortality levels are about 2,600 
dolphins per year from a variety of stocks composing 
an overall abundance of about 9.5 million dolphins in 
the ETP. Even for the stock most commonly affected by 
the fishery, the northeastern spotted dolphin ( Stenella 
attenuata), current kill levels are about 800/year from 
a stock of about 700,000 animals with an estimated 
replacement rate of 14,000-28,000 animals/year. 
Thus, replacing all dolphin sets with school or log 
sets would likely have little effect on dolphin stock 
dynamics, given current levels of fishery-related 
mortality. However, it appears very likely from our 
analysis that such a replacement could substantially 
reduce the abundance and biomass of the commer- 
cial tuna stock by substantially increasing the 
bycatch of yellowfin tuna that are too small for mar- 
ket sale. Discarding these small fish affects both cur- 
rent catch and future size composition. Thus we fo- 
cus here on the potentially detrimental effects on the 
commercial stock of ETP yellowfin tuna that could re- 
sult from redirecting dolphin sets to school and log sets, 
rather than on the negligible effects such effort changes 
are likely to have on ETP dolphin stock dynamics. 
Methods 
Data confection 
As part of a bycatch study initiated in 1989 by the 
Inter- American Tropical Tuna Commission (IATTC), 
data were collected by scientific observers from either 
the National Marine Fisheries Service (NMFS) or 
IATTC, placed aboard U.S. tuna purse seiners during 
routine fishing trips to the ETP. Observed purse sein- 
ers included only those with greater than 400 short-t 
holding capacity. Those with a smaller capacity do 
operate in the ETP and are numerous for some coun- 
tries but are not included in the ETP observer pro- 
gram (IATTC 1 ’ 2 ’ 5 ’ 6 ). 
Each agency contributed about 50% of the observ- 
ers. Agencies alternated sending observers on depart- 
ing trips. Observers recorded time, position, and set 
type of all sets made by U.S. vessels during the 31- 
month “study period” from 1 September 1989 through 
30 March 1992. Position data were used to allocate 
each set to one of three geographic areas of fishing 
effort, as explained below. 
Short tons of tuna discarded per set was calculated 
by subtracting the reported “total tons tuna loaded” 
from the reported “ total tons tuna captured.” After 
completing the original analysis of these data 
(Perkins and Edwards, 1996), we were informed that 
the data set we had been provided for analysis did 
not distinguish between true discard and those cases 
in which the calculated discard also included some 
tuna that were transferred to another vessel rather 
than loaded by the capture vessel because the cap- 
ture vessel had reached capacity during that set. 
Thus in our original analysis we missed a small num- 
ber of sets in which some of the captured fish were 
transferred to another vessel and became part of the 
commercial catch rather than discard. We recalcu- 
lated the discard estimates for this paper using the 
corrected data; therefore the estimated average dis- 
card (and associated parameters) reported here dif- 
fer slightly from our earlier paper. Because the prob- 
lem occurred in only a few sets, it led to very little 
change between the estimated model parameters 
presented in Perkins and Edwards ( 1996) and those 
presented here. The largest change (approximately 
10%) occurred in the parameter p for estimating dis- 
card from dolphin sets, but this change has almost 
no effect on subsequent analyses or conclusions because 
so little discard overall occurs from dolphin sets. 
Observers recorded discard data for only 50% of 
the sets made during the period 1 September 1989 
through 30 July 1990 because only IATTC observers 
were collecting discard data during this period. Dur- 
ing the remaining period, 1 August 1990 through 30 
5 IATTC Annual Rep. 1991. Inter- American Tropical Tuna 
Commission, c/o Scripps Institute of Oceanography, La Jolla, 
CA 92038, 271 p. 
6 IATTC Annual Rep. 1992. Inter-American Tropical Tuna 
Commission, c/o Scripps Institute of Oceanography, La Jolla, 
CA 92038, 315 p. 
