478 



Fishery Bulletin 90(3). 1992 



gatus could be attributed to different dive sites or 

 bottom types and to determine whether there was 

 observer bias. Independent variables for the ANOVA 

 were dive site, substrate material, substrate particle 

 size, and observer. 



Comparisons of trap landings 

 and density estimates 



Trap catch rates were obtained in March 1988 from one 

 pyramid shrimp trap set at each of the five dive sites 

 and allowed to soak overnight (see Ralston and Tagami 

 (1992) for details). Trap catches ofH. laevigatus were 

 regressed against mean densities obtained from visual 

 counts for the five study sites, fitting a linear model 

 with a zero intercept. The slope of this regression is 

 an estimate of q, which was compared with that 

 reported by Ralston and Tagami (1992). 



Results 



A total of 923 shrimp were captured in the 5 pyramid 

 traps set at the study sites. Of these, 705 (76%) were 

 H. laevigatus (Table 2), 217 wereH. ensifer, and 1 was 

 Acanthephyra eximia. During the scheduled observa- 

 tion periods on the submersible dives, a total of 494 

 shrimp were observed (194 total quadrant observa- 

 tions). Of these, only 95 (19%) were if. laevigatus, and 

 the remainder consisted primarily of Plesionika sp., 

 tentatively identified as P. ensis, and a few individuals 

 of P. alcocki, H. ensifer, A. eximia, and Gnathophausia 

 longispina. All H. laevigatus observed from the sub- 

 mersible appeared well within the size range of those 

 captured in the traps, indicating that both stock assess- 

 ment methods sample the same population. 



Visual censuses 



During our dives, the shrimp showed little reaction to 

 the presence of the submersible or its lights. When the 

 submersible came within a few inches of the shrimp, 

 they swam a short distance avoiding collision. When 

 the photoflash was used, the shrimp within a few feet 

 of the submersible started, darting a distance of 1-4 

 cm. No other reactions to the submersible or its lights 

 were observed. Several behavioral differences were 

 noted between the various species observed. 



A total of 94% (89 of 95) of the H. laevigatus ob- 

 served during census periods were seen between 550 

 and 675 m, the depth range herein defined as the zone 

 of maximum abundance. Individuals of H. laevigatus 

 were observed at each dive site, but not necessarily dur- 

 ing the scheduled census periods conducted within the 

 zone of maximum abundance. Only counts taken within 



this depth range were used in the analysis. 



Heterocarpus laevigatus were observed as solitary 

 individuals on the bottom, usually stationary but oc- 

 casionally walking, and rarely swimming near the 

 bottom. They showed little activity in the presence of 

 a baited container and were not observed crawling 

 over or entering it. Conversely, H. ensifer were found 

 in groups near relief features (e.g., large sea anemones) 

 at shallower (450-550 m) depths, either stationary on 

 the bottom or swimming about 1 m above the bottom. 

 They were very active in the presence of a baited con- 

 tainer, aggregating quickly and crawling over and 

 entering the container through the mesh and other 

 holes. Plesionika alcocki usually were seen on the 

 bottom, whereas P. ensis generally were seen hang- 

 ing motionless in the water column ~l-2m off the 

 bottom. Each showed some activity around the baited 

 container. Acanthephyra eximia and G. longispina 

 were observed swimming 1-2 m off the bottom, but 

 were not seen at the baited container. 



Bottom temperature varied during the dives from a 

 low of 3.9°C at 920 m to a high of 6.0°C at 480 m. The 

 temperature range within the zone of maximum H. 

 laevigatus abundance was 4.8-5.9°C. 



Bottom type varied considerably among the sites. 

 The bottom at the two Oahu sites was classified as 

 coralline sand making up an even, featureless plain 

 with a gradual (<20°) slope. The bottom at the three 

 Kona coast sites was much steeper, generally about a 

 35-45° slope, but with some sections near vertical or 

 even slightly undercut at the Kona site 3. At Kona site 

 1, the bottom was nearly uniformly composed of small 

 (5-10 cm diameter), sharp-edged volcanic rocks and 

 very little coralline material. Kona site 2 differed from 

 site 1 in that the small volcanic rocks were more 

 weathered and the substrate had a greater coralline 

 component. Kona site 3 had many sandy areas, at 



