298 
Fishery Bulletin 116(3-4) 
Total length (cm) 
Figure 4 
Length-frequency distributions of 4 species of rockfish (Sebastes spp.) on trawlable low- 
relief habitat from submersible surveys (ST, dark gray bars with diagonal hatches) and 
the Northwest Fisheries Science Center West Coast Bottom Trawl Survey (T, checkered 
bars) conducted off central California (36-37°N latitude) during 2003-2009. The curved 
line represents the bottom-trawl selectivity function from the most recent stock assess¬ 
ment for each species (not available for vermilion rockfish, Sebastes miniatus). The solid 
and dashed vertical lines indicate the mean lengths from submersible and bottom-trawl 
surveys, respectively. 
of vermilion rockfish on the 2 habitats were not sig¬ 
nificantly different, fish <35 cm TL occurred in greater 
proportion on untrawlable habitat. 
The third comparison, that of length distributions of 
fish from submersible surveys on low-relief trawlable 
habitat with those from the trawl survey (Fig. 4) was 
significantly different for all 4 species (Pearson’s chi- 
square two-sample test, P<0.001; Table 4). In these 
comparisons, the dissimilarity in length distributions 
between the two surveys was even more apparent for 
greenspotted, greenstriped, and canary rockfishes than 
when all submersible survey data for these species, re¬ 
gardless of habitat, were compared with trawl survey 
data (Fig. 2, Table 2). The difference in the proportion 
of small fish present in the two surveys was particular¬ 
ly pronounced for greenstriped and canary rockfishes, 
with 10% length quartiles of 5 and 17 cm TL (green¬ 
striped) and 15 and 37 cm TL (canary) in submersible 
and trawl survey data, respectively. In contrast, length 
distributions for vermilion rockfish were more similar 
between surveys than length distributions for the other 
species, owing to a greater proportion of large vermil¬ 
ion rockfish (>45 cm TL) on trawlable habitat in the 
submersible survey data. 
Discussion 
Our study provides a useful comparison of length 
data collected by the West Coast Bottom Trawl Sur¬ 
vey and nearby submersible surveys in untrawlable 
areas, for some deepwater rockfishes off central Cali¬ 
fornia. Although the trawl survey samples areas of 
soft and low-relief habitats where relatively low den¬ 
sities of many rockfishes occur, there were enough 
length data for comparisons of some species that 
associate mostly with mixed and high-relief rocky 
habitats. The broader length and depth distributions 
present in the submersible survey data allowed infor¬ 
mative comparisons with the trawl survey data. We 
could not directly address whether the low proportion 
of small sizes in the trawl data was due to habitat 
(i.e., small fish not available on trawlable habitat) or 
gear selectivity. However, the greater proportion of 
small sizes present on trawlable habitats in the sub¬ 
mersible survey data (in particular for greenstriped 
rockfish), and the similar maximum lengths present 
in both surveys (Fig. 4) suggest that gear selectiv¬ 
ity was the cause. Similarly, commercial trawl gear 
selects for larger sizes and would not be expected 
