544 



Fishery Bulletin 90(3), 1992 



> 



o 

 o 



-I—' 



c 

 o 



(U 



100 J Station 1. n 



75- 



101 



^ 



_KfS_^^_ 



-EfSL- 



BC 



MP 



MC 



BS CB 



CM 



SC 



S0-- 



Station 4, n = 1 34 



^ 



_s;sL. 



_KJSl_ 



F* p^l ■ 



SB 



BC 



BS 



CB CM 



SC 



75 

 50-- 



Station 6, n = 50 



^ F^ I ^ E^S 



BB 



BC 



BS CB 



Bottom Type 



Figure 2 



Percent cover of the ten most abundant bottom-type combinations at Heceta 

 Bank, Oregon. RR = rocl< ridge; BB = boulder; BC = boulder-cobble; BS 

 = boulder-sand; CB = cobble-boulder; CM = cobble-mud; SC = sand-cobble; 

 MC = mud-cobble; MP = mud-pebble; MM = mud (see text for a descrip- 

 tion of bottom-type codes), n = number of habitat patches sampled per 

 station. 



Among non-schooling species, the estimated total 

 length ranged from 105 cm (dogfish shark Squalus 

 acanthias) to a mean of 12 cm (unidentified small Sebas- 

 tes) (Table 1). Many smaller fishes were seen, but could 

 be identified only to genus {Sebastes juveniles) or family 

 (Gobiidae). Among schooling species, average length 



varied from 42 cm (yellowrtaO rockfish Sebas- 

 tes flavidus) to 11cm (Sebastes juveniles). 



Ontogenetic habitat changes 



Several rockfish species occurred both in 

 schools and singly, including pygmy, yel- 

 lowtail, sharpchin S. zacentrus, redstripe 

 S. proringer, and canary S. pinniger. Pyg- 

 mies were the single-most abundant spe- 

 cies identified in either category. Schools 

 of pygmy rockfish consisted of significant- 

 ly smaller individuals, averaging 16.1cm, 

 whereas non-schooling aggregations and 

 solitary individuals averaged 19.4cmTL 

 (ANOVA, F 18.0, df 1,699, P<0.01). Fur- 

 thermore, schools were usually associated 

 with ridge tops shallower than 100 m, while 

 non-schooling fish were on cobble and 

 boulder bottoms at depths of 100-150 m. 

 A similar analysis of the data for yellow- 

 tail, sharpchin, redstripe, and canary rock- 

 fish showed no significant difference be- 

 tween sizes of individuals in and out of 

 schools (ANOVA, all P>0.05). 



Differences among stations: 

 Fish assemblages 



Results of the PCA indicated striking dif- 

 ferences among stations in both the com- 

 position of fish assemblages and the 

 similarity of individual transects within 

 stations (Table 2, Fig. 3). Bartlett's spher- 

 icity test indicated that the first two axes 

 described significant non-random patterns 

 of variation among species, and accounted 

 for 70% of the total variation. The first 

 axis, which accounted for 49% of the varia- 

 tion, primarily contrasted transects at sta- 

 tion 4 (intermixed mud and rocky ridges) 

 vs. transects at station 5 (mud) (Fig. 3). 

 Transects from stations 1 and 3 (shallow 

 rocky bank tops) formed intermediate 

 homogeneous groups, while transects from 

 stations 2 and 6 (medium-depth boulder 

 and cobble fields) formed intermediate 

 heterogenous groups. Variable loadings in- 

 dicated that this axis primarily contrasted 

 variation in the relative abundance of rosethorn, 

 pygmy, canary, sharpchin, yellowtail, and greenstriped 

 rockfish, which were abundant at stations 1, 3, and 4 

 with the relative abundance of thornyhead rockfish, rex 

 sole, sablefish, poachers, and zoarcids, which were 

 abundant at station 5 (Table 2). 



