Edwards and Perkins: Estimated tuna discard from dolphin, school, and log sets 
215 
76% of observed school sets, and 
75% of observed log sets. The set 
data reported 134, 1,098, and 
9,819 tons of observed discard, 
respectively. A nonzero amount of 
discard was reported in 65% 
(Table 1) of log sets where discard 
was recorded, but in only 8% of 
school sets and only 0.7% of dol- 
phin sets where discard was re- 
corded. Frequency distributions 
for tons of tuna discarded per set 
were clustered near zero for dol- 
phin sets, were somewhat wider 
for school sets, and were much 
wider for log sets (Fig. 3), indi- 
cating the increasing frequency 
and increasing tonnage of tuna 
discard from dolphin, school, and 
log sets, respectively. 
Geographic patterns in discard 
reflected geographic patterns in 
effort devoted to the three differ- 
ent set types. Both effort and 
positive discard occurred most 
frequently in area 1 for dolphin 
sets and in area 3 for school and 
log sets (Table 1). 
Expected discard per set 
Expected discard per set for log 
sets was 10.5 short tons of tuna 
(all tuna species combined; SE= 
0.83, n=998, means range 7.1- 
15.4 short-t depending on geo- 
graphic area; Fig. 4). Expected 
discard per set for school sets was 
1.15 short tons of tuna (SE=0.24, 
rc=960, means range 0.97-1.57 
short-t) depending on geographic 
area). Expected discard per set 
for dolphin sets was 0.06 short 
tons of tuna (SE=0.2, n=2,110 
fishery wide). 
Estimated average annual discard 
Despite the much larger number of sets (greater fish- 
ing effort) on dolphins than on school or logs (Table 
1), estimated average annual discard was much 
greater for log sets than for school or dolphin sets, 
reflecting the enormous difference between set types 
in expected discard per set. Estimated average an- 
nual discard for the entire U.S. fleet during the study 
period was 88 short tons per year (SE=29, n=2,110) 
from dolphin sets, 595 short tons per year (SE=125, 
n=960) from school sets, and 5,400 short tons of tuna 
(all species combined) per year (SE=426, n=998) from 
log sets, based on observed patterns of effort (Fig. 5). 
Redistributing the great number of dolphin sets 
to either of the other set types considerably increased 
predicted annual discard of tuna from school or log 
sets (or both), and from the U.S. fleet as a whole. 
Redistributing all dolphins sets to school sets in- 
creased predicted annual discard of tuna from school 
