Murie et al. : Distribution and activities of Sebastes spp. 



315 



widest. Specifics of the oceanographic characteris- 

 tics of Saanich Inlet are detailed elsewhere 

 (Herlinveaux, 1962; Anderson and Devol, 1973). Of 

 relevance is that the basin of the inlet reaches a 

 maximum depth of 234 m, but a submerged sill at 

 75 m at its mouth in Satellite Channel restricts the 

 renewal of deep-water into the inlet. The resulting 

 oxygen deficiency, anoxia, and production of hydro- 

 gen sulfide is offset in some years when well-oxy- 

 genated, dense bottom-water intrudes over the sill 

 (Herlinveaux, 1962; Anderson and Devol, 1973). We 

 therefore did a hydrocast at depths of 0-225 m to 

 determine the depth of the oxycline in the inlet and 

 whether hydrogen sulfide was present at the time 

 of the surveys. The sampling station was located at 

 lat. 48°37.80'N and long. 123°30.00'W (Liu, 1989). 



Three sites within Saanich Inlet were surveyed: 

 five transects were traversed at Elbow Point, eight 

 in an area north of McKenzie Bight, and three in 

 an area north of Sheppard Point (Fig. 1). These ar- 

 eas were known from preliminary SCUBA dive sur- 

 veys to have rockfish present in >30 m water depth 

 (Murie et al., pers. obs.). All surveys were conducted 

 during daylight hours (09:30 hours to 16:00 hours) 

 and were also restricted to a depth range between 

 20 and 150 m because the buoyancy of the Pisces PV 

 is not finely controlled above 20 m, and bottom time 

 restrictions precluded the transects starting at the 

 basin floor (-200 m). Depth at which each transect 

 started therefore varied among sites owing to slope 

 and positioning of the submersible, with starting 

 depths of 95-109 m at Elbow Point, 92-154 m at 

 McKenzie Bight, and 67-74 m at Sheppard Point. 



At the start of each transect, the Pisces PV sub- 

 merged in open water and on reaching depth the 

 external floodlights were lit and the submersible 

 was manoeuvred horizontally, slowly, toward the 

 cliff face. Once the bottom substrate (cliff) was lo- 

 cated, the submersible began a slow vertical ascent 

 (~5 m-min -1 ), keeping the viewing ports (port, pilot, 

 and starboard) directed perpendicular to and ap- 

 proximately 3 m from the substrate. Underwater 

 visibility at the time of the surveys was -5—6 m with 

 external illumination. 



On ascent, an audio-record was made of the spe- 

 cies, time, depth, estimated size (whenever possible), 

 activity, and habitat for each rockfish observed. Each 

 observer (port and starboard) recorded all rockfish 

 encountered within a plane bisecting the pilot's 

 viewport and extending outward at an angle of ap- 

 proximately 45°, corresponding to approximately 3 

 m of horizontal distance across the substrate (i.e. 

 viewing width). To avoid counting the same fish 

 twice, any rockfish swimming across the path of the 



submersible or positioned close to the pilot's view- 

 ing area was pointed out to the other observer. Size 

 was visually estimated (±5 cm) by comparing the 

 fish with an externally mounted graduated rod. 

 Rockfish were designated as small (<20 cm total 

 length [TL]) and large (>20 cm TL); large referring 

 to the size at which they enter recreational and com- 

 mercial fisheries (Richards, 1986). Activity of each 

 fish was scored according to whether the fish was 

 perched in the open, positioned in a crevice, occu- 

 pying a shelter hole, hovering off the substrate, or 

 swimming. Habitat was categorized as vertical wall 

 (may have cracks, small crevices, or ledges; score=l), 

 complex (comprising broken rock and boulder fields; 

 score=2), or sand-mud (score=3). Any change in the 

 habitat or slope of the substrate (±10°) was recorded 

 and the depth noted. 



Rockfish density was estimated for each habitat 

 type over 20-m depth intervals. The total number 

 of fish recorded by both observers within each habi- 

 tat type over a 20-m depth interval was divided by 

 the total area of that habitat type viewed over the 

 depth interval. The area viewed was calculated by 

 multiplying the viewing width of both observers (i.e. 

 6 m) by the ratio of the change in depth to the sine 

 of the slope. 



Median densities of small and large fish were 

 calculated for each habitat type and 20-m depth 

 interval, with transects pooled for increased sample 

 size. Density distributions for rockfish were skewed 

 so densities were calculated as medians with 25% 

 and 75% quartiles. Densities of quillback rockfish, 

 S. maliger, among depth intervals and habitats were 

 analyzed by using Kruskal-Wallis tests (SAS, 1985), 

 as this species had an adequate median density (>1 

 fish- 100m 2 ). Statistical significance was indicated 

 by P < 0.05. Analyses for the other rockfish species 

 (median densities <1 fishlOOm 2 ) were limited to 

 qualitative comparisons of their depth distributions 

 and numerical abundances. 



Activity of each species of rockfish was analyzed 

 using percent occurrence, which was calculated by 

 dividing the sum of all individuals observed in each 

 activity by the total number of individuals of the 

 species for which activities were recorded, and mul- 

 tiplying by 100%. Species associations were deter- 

 mined for individual fish of each species by scoring 

 the presence of a conspecific or a heterospecific 

 within 3 m. The sum of the number of individuals 

 which were observed in the presence of a con- or 

 hetero-specific was then expressed as a percentage 

 of the total number of individuals of the species. 

 Individual rockfish with no other rockfish within 3 

 m were considered to be 'alone' (solitary). 



