306 
Fishery Bulletin 111(4) 
With the hypothesis that the levels of bottomfish 
prey and current speed are greater over hard-high 
habitats than over other environments (Ralston et al., 
1986; Haight et al., 1993a; Kelley et al., 2006), it could 
be inferred that Opakapaka, Kalekale, and Onaga 
move into this habitat type upon reaching sexual ma- 
turity to increase their foraging rates and maximize re- 
productive output and gamete dispersal. On coral reefs 
in Hawaii, the Yellow Tang ( Zebrasoma flauescens) has 
been found to shift into habitats with increased food 
resources when it reaches reproductive size to possibly 
improve its reproductive ability (Claisse et al., 2009). 
No actual bottomfish spawning events were recorded 
during our study. Opakapaka and Onaga are known to 
spawn at night (C. Kelley, unpubl. data), and camera 
deployments were restricted to daytime hours. Other 
than seasonality, habitat and environmental parame- 
ters of bottomfish spawning have yet to be determined. 
It remains possible, however, that the observed onto- 
genetic habitat shifts occurred as a result of a repro- 
ductive cue — given that the change in habitat roughly 
coincided with sexual maturity. 
Another factor that may influence ontogenetic habi- 
tat shifts is habitat complexity. Laidig et al. (2009) 
found that juvenile roekfishes on the continental shelf 
off central California were associated with boulder and 
cobble habitats before they moved into the slope habi- 
tats used by adults. It is plausible that juveniles and 
smaller species of bottomfishes use more complex habi- 
tats in a similar manner for protection and predator 
avoidance. However, because habitats were classified at 
a 200-m scale, our study did not take into account hab- 
itat heterogeneity within grid cells and smaller-scale 
habitat characteristics, such as complexity or rugosity. 
Structural complexity and the combination of habitat 
types in a given area are likely to influence fish distri- 
butions at their respective scales. Future work is need- 
ed to investigate the role of habitat complexity and 
heterogeneity on size distributions of bottomfishes and 
to look more closely into how specific habitat types are 
used. Such an approach could provide more informa- 
tion about the cause of the ontogenetic habitat shifts 
observed in this study. 
The regional variations in relative abundance and 
mean length could be related to differential fishing 
pressure or large-scale habitat features. It can be ex- 
pected that remote locations, such as Niihau, would 
have less fishing pressure than locations closer to 
major ports and, thereby, would have greater relative 
abundances and lengths of target species. Contrary to 
this expectation, the highest levels of relative abun- 
dance were found at Hilo for Opakapaka and in Pailolo 
Channel for Ehu. Both areas are easily accessible to 
fishing; therefore, other factors may have driven the 
observed distributions. Protection did not have an in- 
fluence on the relative abundance of any of the 4 spe- 
cies studied, a finding that is consistent with the re- 
sults of Moore et al. (2013). In terms of mean length, 
the largest Opakapaka may have been found at Niihau 
because of the remote location and longevity of the 
protection of this small island. The Niihau BRFA has 
been closed to fishing since 1998. The opposite may 
be true for Hilo, where the smallest Opakapaka were 
observed. Before the implementation of the revised 
system of BRFAs, fishing in the depth range of Opak- 
apaka was permitted because the BRFA boundary be- 
gan at 200 m. How protection and fishing pressure af- 
fect abundance and size distributions of bottomfishes 
should be investigated further because these factors 
may confound any trends attributed to habitat or oth- 
er environmental variables. 
Mega-scale habitat features (scale from Greene et 
al., 1999: macro=l-10 m; meso=10-1000 m; mega=:l-10 
km), such as pinnacles, banks, terraces, and even fea- 
tureless carbonate flats, also could be influencing bot- 
tomfish distributions. In this study, juvenile Opakapa- 
ka and Onaga were found to associate with hard-low 
habitats. There is a large terrace at Hilo, where most 
juvenile Opakapaka were observed, and flat, hardbot- 
tom habitats predominate in Pailolo Channel, where 
most Onaga juveniles were present. These large-scale 
features predominantly have low slopes and hard bot- 
toms and match the observed habitat preference of 
these species at the meso-scale. However, because of 
the difference between the habitat classification scale 
(200x200 m) used in our study and the size of mega- 
scale features, further investigation is required to es- 
tablish a conclusive connection between the bottomfish 
distributions observed in this study and mega-scale 
features. In the case of Pailolo Channel, for example, 
with its large, flat areas of hardbottom habitat, our re- 
sults agreed with a finding of another survey effort. 
Previous fishing surveys have indicated that this area 
possibly was a nursery ground for Onaga (C. Kelley, 
unpubl. data). Because the smallest mean length (42.80 
cm FL) and about 75% of all juveniles of this species 
measured (<61 cm FL) in this study came from Pailolo 
Channel, it is highly likely that a nursery ground for 
Onaga exists in this area. 
Conclusions 
This study has improved our understanding of the 
species-specific ecology of 4 bottomfish species in the 
MHI. Analyses of habitat preferences on the basis of 
relative abundance and length-frequency distributions 
showed that habitat types other than hard-high envi- 
ronments are important to each of the species studied, 
often as a result of ontogenetic shifts in habitat use. 
Given that these bottomfishes are found throughout 
the Indo-Pacific region, these findings may provide the 
framework for the prediction of species distributions 
outside of Hawaii. Because juveniles of Opakapaka and 
Onaga were associated mostly with hard-low habitats, 
it is imperative that future definitions of the bottom- 
fish EFH take into account habitat associations by life 
stage. Although some species share similar preferences, 
