Barry et al.: Capture efficiency and size selectivity of four pot types in the fishery for Enteroctopus dofleini 
41 
Figure 2 
The four pot types we evaluated in Kachemak Bay, Alaska, in 2006 to determine their effi- 
ciency in capturing North Pacific giant octopus (Enteroctopus dofleini) for use in a directed 
fishery: (A) Lair pots; (B) Korean hair crab pots; (C) sablefish pots; and (D) shrimp pots. 
sessed by comparing mean weight of octopuses caught 
in each pot type. Because of heteroscedasticity in the 
data we used a nonparametric Welch analysis of vari- 
ance (ANOVA) to test for variation in the weight of oc- 
topuses by pot type, followed by Tukey’s post hoc tests. 
In order to diminish the potential confounding effects 
of differences in depths at which pots were set, we re- 
analyzed weight by pot-type data using only giant oc- 
topuses caught within Eldred Passage where depth did 
not differ significantly by pot type, using an ANOVA 
and Fisher’s least significant difference (LSD) test. 
Individual chi-square tests were used to determine 
whether there was a disproportionate number of males 
or females caught in any one trap type. 
No standardized method exists to measure octopus 
CPUE. In octopus fisheries around the world CPUE 
is tailored to the type of fishery (Defeo and Castilla, 
1998; Hernandez-Garcia et al., 1998; Sanchez et al., 
2004). There are markets for both food and bait giant 
octopuses and in Alaska the fishery would not be size 
selective. In a full-time directed fishery where the 
ability to turn over gear is important, “the total catch 
in kilograms (kg) per days soaked” may be a most use- 
ful expression of effort. In contrast, kilograms per pot 
may be a more useful indicator of effort in a part-time 
fishery. The number of pots soaking may be more im- 
portant if a vessel is employed to simultaneously take 
part in another fishery while its pots for giant octo- 
pus are soaking. Irrespective of the fishery structure, 
whether full-time or concurrent, both CPUE estimates 
provide valuable information for resource manage- 
ment. Both CPUE estimates were calculated for each 
individual survey and then averaged over surveys. No 
statistical tests were performed on CPUE data because 
the averages calculated over the two surveys are based 
on two estimates. Catch-per-unit-of-effort values for 
shrimp pots represent point estimates because shrimp 
pots were fished only in the first survey. We compared 
the average soak time of lair pots that caught giant 
octopuses to those that did not by using a post hoc 
t-test to investigate anecdotal evidence that would 
indicate that longer soak times increase the likelihood 
of capture. 
Results 
We captured a combined total of 254 giant octopuses, 
four of which were dead and omitted from further analy- 
sis. The number of individuals captured was consistent 
