Loher and Hobden: Length and sex effects on the spatial structure of Hippoglossus stenolepis 
49 
desired statistical significance level (here, P=0.05). NND 
less than the lower critical value indicates significant 
aggregation and NND greater than the upper critical 
value indicates significant overdispersion; random dis- 
tribution is indicated by NND between or equal to the 
critical values. 
Sequential segregation analysis Runs test (RT; Barton 
and David, 1958) was performed as described by Upton 
and Fingleton (1985) to test for significant segregation 
of 1) sublegal-size relative to legal-size halibut; 2) known 
males relative to females and individuals of unknown 
sex; and 3) known females relative to males and indi- 
viduals of unknown sex. RT is a bivariate analysis in 
which a significant result indicates “departure from 
complete randomness” but does not indicate the direc- 
tion of that nonrandomness; i.e., whether the events 
are segregated or overdispersed. Results of tests for 
segregation-by-length will be reported with respect 
to sublegal-size halibut but not for legal-size halibut, 
because reporting both would be redundant. Results 
are reported independently for male and female halibut 
because a third alternative (unknown sex) was possible. 
Results 
Spatial aggregation 
On the majority of full hook-inventory sets, halibut were 
randomly dispersed with little clustering relative to 
empty hooks and bycatch: significant aggregation of hali- 
but was found on 22% of stations (NNA; Table 1). The 
tendency to aggregate appeared to be size-dependent, 
with 44% of the sets showing significant aggregation 
within the legal-size segment of the halibut catch and 
only 11% showing aggregation of sublegal-size fish. 
Known male and female aggregations were significantly 
detected on 22% and 11% of sets, respectively. For the 
sets on which the fish were not aggregated, distributions 
were always spatially random as opposed to overdis- 
persed (Table 1). 
Sequential segregation 
Sixteen percent of the sets displayed significant (RT; 
P<0.05) halibut segregation by length a rate somewhat 
lower than that observed for aggregation of legal-size 
halibut and higher than aggregation of sublegal-size fish 
(Table 1). Segregation of known males was observed on 
8% of sets and female segregation on 20%. Comparison 
of NNA and RT results for stations upon which both 
tests were conducted rarely demonstrated simultaneous 
spatial aggregation and sequential segregation for the 
same population segment. Significant segregation within 
a given population segment was typically observed at 
different stations from those at which significant aggre- 
gation was detected (Table 1). Overall, some form of 
significant spatial structure was detected at the majority 
(60%) of stations. 
Discussion 
Our observation of significant spatial structure within 
the majority of longline catches examined is consistent 
with Clark’s (2004) concern that the demographics of 
commercial catch may vary from survey data in unde- 
tectable ways. In particular, aggregation may allow the 
commercial fleet to selectively target females and the 
fastest-growing members of their cohorts, yielding sex 
ratios that differ from sex ratios encountered during 
surveys at any given combination of size and age. In the 
current analysis, aggregation based on length was more 
commonly detected than aggregation by sex similar to 
prior wild-capture (Lpkkeborg and Bjordal, 1992) and 
laboratory (Stoner and Ottmar, 2004) results. Patches of 
larger fish may occur because larger size translates into 
greater swimming speed, feeding range, and a likelihood 
of encountering bait, and because of competitive domi- 
nance. Patches of sublegal-size halibut may form owing 
to higher feeding motivation and more effective search 
patterns (Stoner and Ottmar, 2004); alternatively, some 
areas may simply represent size-specific habitat. Pacific 
halibut undergo ontogenic shifts in habitat use, settling 
in shallow water as juveniles (Norcross et ah, 1995; 
Abookire et al., 2001) and moving deeper with age (Best 
and Hardman, 1982; Hoag et al., 1997). Age-specific 
distribution and commercial catch rates have also been 
documented in U.S. west coast Dover sole ( Microstomus 
pacificus [Jacobson et al., 2001]), and Piet et al. (1998) 
have suggested that flatfish partition with respect to 
gape size. Importantly, segregation by size could cause 
faster-growing individuals within each cohort to aggre- 
gate separately from slower-growing individuals, poten- 
tially generating skewed mean demographics depending 
on the relative distribution of capture effort to patch 
distribution. 
Our ability to draw specific conclusions regarding 
sex-specific aggregation was limited by our inability 
to dissect all individuals, but sex-based structure was 
still detected. Pacific halibut exhibit sexually dimor- 
phic growth (Gorchinsky, 1998; Clark et al., 1999) and 
even in the absence of true sex-specific differences in 
behavior, females likely predominate in legal-size ag- 
gregations owing to larger size-at-age (Clark, 2004); 
males, conversely, should be more abundant in patches 
of sublegal-size halibut. Sex-specific seasonal aggre- 
gation is also likely, given the species’ documented 
seasonal redistribution and concentration at winter 
spawning grounds (St. Pierre, 1984; Loher and Blood, 
2009; Seitz et al., 2011). Likewise, petrale sole ( Eop - 
setta jordani; Hannah et al., 2002) and Kobe Hounder 
( Crossorhombus kobensis [Moyer et al., 1985]) have 
been shown to exhibit seasonal aggregation for mating. 
In the case of Pacific halibut, analysis of behavioral 
data indicates that recent commercial fishing seasons 
have intersected the species’ seasonal migratory period 
to a considerable degree (Loher, 2011), potentially al- 
lowing catch demography to differ during autumn and 
spring, relative to when summer survey data are col- 
lected. Additionally, there is evidence that Greenland 
