Watters and Dick: Length distributions of Sebastes spp. off central California 
295 
Table 1 
Characteristics of samples from submersible and bottom-trawl surveys used to compare length distributions of 4 species of 
rockfish (Sebastes spp.) off central California (36-37°N latitude) during 2003-2009. Habitats categorized as trawlable within 
the submersible surveys were considered rough proxies for the trawl survey because they were too small to be trawled. Com¬ 
mon depth ranges for species are from Love et al. (2002). 
Species 
Survey 
No. of 
transects 
or hauls 
Total area 
sampled 
(m 2 ) 
Area of 
trawlable 
habitat 
surveyed (%) 
No. 
of fish 
measured 
Avg. depth 
and depth 
range (m) of 
measured fish 
Common 
depth range 
(m) of 
species 
Greenspotted rockfish 
Submersible 
503 
250,970 
40 
3282 
132(55-307) 
30-363 
Trawl 
20 
318,699 
100 
292 
109(85-239) 
Greenstriped rockfish 
Submersible 
369 
189,596 
47 
2157 
125(80-299) 
100-250 
Trawl 
28 
461,814 
100 
601 
126(94-264) 
Vermilion rockfish 
Submersible 
139 
65,846 
22 
695 
111(55-203) 
50-150 
Trawl 
6 
90,812 
100 
80 
111(81-121) 
Canary rockfish 
Submersible 
120 
53,313 
27 
667 
108(55-233) 
80-200 
Trawl 
9 
144,062 
100 
90 
98 (84-161) 
different (Fig. 1). A total of 609 submersible transects 
were clustered in canyons and areas of relief, whereas 
35 trawl hauls were dispersed outside of these areas. 
Length data from the trawl and submersible surveys 
were collected from overlapping depths throughout the 
common depth range of all 4 species (Table 1). Range 
of depth was broader for the submersible data. Traw¬ 
lable habitat represented less than 50% of the total 
habitat sampled along submersible transects and var¬ 
ied by species. Transects with vermilion and canary 
rockfishes present contained the least amount of traw¬ 
lable habitat, while those with greenspotted and green- 
striped rockfishes contained greater amounts of habi¬ 
tat categorized as trawlable. In general, the amount 
of seafloor area sampled in relation to the number of 
fish measured was considerably greater for trawl hauls 
than for submersible transects. 
The first comparison, that of length distributions of 
all individuals of each of the 4 species, revealed signifi¬ 
cantly different distributions for the 2 surveys (Pear¬ 
son’s chi-square two-sample test, P<0.001), and there 
was a broader range of lengths and greater proportion 
of small fish in the submersible data (Fig. 2, Table 2). 
Greater proportions of greenspotted and greenstriped 
rockfishes <30 cm TL and <20 cm TL respectively, and 
canary and vermilion rockfishes <40 cm TL and <45 
cm TL respectively, were present in the submersible 
data than in the trawl data. Binned maximum lengths 
from the 2 surveys were the same for greenstriped (40 
cm TL) and vermilion (60 cm TL) rockfishes, 5 cm TL 
larger in the submersible survey for greenspotted rock- 
fish (50 cm TL), and 15 cm TL larger in the submers¬ 
ible survey for canary rockfish (70 cm TL). Greenspot¬ 
ted rockfish length distributions were most similar be¬ 
tween the surveys, although two peaks were present in 
the trawl survey data; all but the smallest (5 cm TL) 
and largest (50 cm TL) length bins were represented in 
trawl survey data for this species. Length distributions 
of vermilion and canary rockfishes were most dissimi¬ 
lar between the surveys; several length bins with data 
from the submersible survey were missing data from 
the trawl survey, and the 10% length quartiles differed 
by ca. 10-cm-TL. 
Trawl survey selectivity curves for greenspotted and 
greenstriped rockfishes are consistent with a reduced 
proportion of small fish compared with the proportion 
from the submersible survey (Fig. 2). The dispropor¬ 
tionate number of small canary rockfish in the sub¬ 
mersible survey, compared with that in the trawl sur¬ 
vey, is not consistent with the estimated trawl selectiv¬ 
ity curve for that species. If canary rockfish larger than 
15 cm TL are 100% vulnerable to the trawl survey, as 
implied by the selectivity curve, the expected propor¬ 
tions of small sizes would be at least as large as those 
in the submersible survey. 
The second comparison, that of length distributions 
of fish from the submersible survey on untrawlable 
and trawlable habitats (Fig. 3), revealed significantly 
different distributions for greenspotted, greenstriped, 
and canary rockfishes (Pearson’s chi-square two-sample 
test, P<0.001), whereas those of vermilion rockfish were 
not (Table 3). Mean lengths for all, except vermilion 
rockfish, were smaller on trawlable than on untraw¬ 
lable habitat. All sizes of greenspotted rockfish were 
present on both habitat types; however, small (<20 cm 
TL) fish occurred in greater proportion on trawlable 
than on untrawlable habitat. Greenstriped rockfish, 
with almost equal numbers of lengths from the 2 habi¬ 
tats, had a greater proportion of small (<20 cm TL) 
fish on trawlable habitat. The small number of canary 
rockfish that were surveyed on trawlable habitat (35 
individuals) also had a greater proportion of small (<20 
cm TL) fish than the proportion of small fish present 
on untrawlable habitat. Although length distributions 
