October 1986, at Bamberton (lat. 48°35'N, long. 

 123°3rW), located half-way down Saanich Inlet, a 

 24 km long fjord on the southeast coast of Van- 

 couver Island. Because of its close proximity to 

 regional oceanographic laboratories, this inlet has 

 been well studied oceanographically (Carter 1934; 

 Herlinveaux 1962; Richards 1965; Anderson and 

 Devol 1973; Pickard 1975; Emerson et al. 1979; 

 Thomson 1981) and biologically (Tunnicliffe 1981; 

 Burd and Brinkhurst 1984, 1985). At its mouth, it 

 has a submerged (75 m) sill behind which there is 

 a deep (maximum depth = 234 m) basin, and water 

 below the sill depth is typically isolated, oxygen- 

 deficient, and may contain hydrogen sulphide. The 

 basin is flushed only when water above the sill is 

 sufficiently dense to cascade into it. Throughout the 

 year, there typically remains one or more oxyclines 

 in the water column structure. The study area is 

 generally characterized by a 20-30° slope to about 

 70 m, and a 30-60° slope from this depth to the 

 bottom of the inlet (209 m). 



Observations were made by two people from the 

 submersible Pisces IV, which has been previously 

 described by Mackie and Mills (1983). Two 3-h dives 

 were made each day— one typically starting at 1300 

 and the other at 1900, about one hour after sunset. 

 Dives to the bottom of the inlet were conducted at 

 random locations over a 3 km length of shoreline 

 and began with a direct descent to the bottom of 

 the inlet. The submersible was then moved upwards, 

 horizontal and about 1 m from the cliff, from the 

 bottom of the inlet to a depth of 20 m. After sur- 

 facing, the submersible was then towed to another 

 dive location, where the process was repeated. A 

 total of 17 ascents were completed during the 

 8 dives. 



Ascent speed varied according to habitat complex- 

 ity, slope, and crustacean abundance, but was slow 

 enough to permit recording of the species observed 

 except in areas where extremely dense concentra- 

 tions of animals were found. Observers were sta- 

 tioned on opposite sides of the submersible, with no 

 overlap in their visual field. Each observer was able 

 to scan approximately 90° on one side of the sub- 

 mersible 's path. 



As discussed by Richards and Schnute (1986), a 

 general problem with use of submersibles has been 

 the quantification of species abundance. Both the 

 presence of lights and the submersible itself might 

 affect animal behavior and hence bias observations. 

 Prawns tend to be cryptic and were frequently found 

 in association with bottom debris, and occasionally 

 in holes. After extensive observations, including 

 moving at specific locales with lights on or off and 



moving at variable speeds, it was concluded that 

 prawns and other demersal crustaceans were not 

 apparently affected by Pisces's presence, allowing 

 them to be effectively studied. On the approach of 

 Pisces, demersal crustaceans would occasionally 

 swim a short distance with rapid flicks of their 

 abdomens, but in most cases, they would simply 

 assume an aggressive stance towards the submers- 

 ible and continuously face it as it passed by. They 

 did not retreat under cover. 



Visibility of benthic animals varied somewhat 

 because of changes in bottom topography and its 

 effect on distance of the submersible from the sub- 

 strate. Water clarity was generally good (>7 m), but 

 resolution was poor at the perimeter of the illum- 

 inated area (about 5 m maximum). 



Data were initially recorded into handheld tape 

 recorders, and tapes were transcribed shortly after 

 each dive. Prawns were individually counted and 

 their depths of occurrence noted over a vertical 

 transect of about 3 m. However, this procedure 

 could not be used when live prawn abundance ex- 

 ceeded approximately 6 m"^, because of their high 

 density and because the prawns stirred up the sub- 

 strate by their movements. This situation only oc- 

 curred over a narrow depth range (between 70 and 

 85 m depth, depending on date of observation). Dead 

 prawns, which often only consisted of exoskeleton 

 fragments, were also difficult to count at a density 

 greater than about 2 m"^. In both situations, num- 

 ber per meter of depth was conservatively estimated 

 by multiplying the respective minimum average den- 

 sity per square meter by 3 m, the transect width 

 over which live and dead prawn abundances were 

 being assessed. 



Qualitative notes on abundance of munids. Muni- 

 da quadraspina Benedict, 1902, the dominant ben- 

 thic crustacean present, were recorded by depth 

 interval. Observers noted other invertebrate and 

 fish species present in each transect. 



Movement of observed crustaceans was sufficient- 

 ly slow, relative to the submersible' s movement, to 

 prevent their crossing the submersible's path and 

 possible double counting by the observers. Body size 

 of some individuals was estimated by comparing 

 them to a 30 cm rod, marked in 10 cm intervals, that 

 hung in front of the left viewport. Sizes were later 

 confirmed by measurement of carapace lengths of 

 specimens collected using the extendable arm of the 

 submersible. 



Water samples, and on some occasions crusta- 

 ceans, were collected at selected depths by pump- 

 ing water through jars attached to the exterior of 

 the submersible until they had been thoroughly 



602 



