248 



Fishery Bulletin 93(2), 1995 



H. jordani - 



H. splnosus - 



Bathymaster spp. 

 A. fimbria — 



C. aleutensls 

 H. hemilepidotus 

 Mallotus villosus 



Stlchaeidae 



Affinity Level 

 0.40 



0.30 



Myoxocephalus spp. 



A. hexapterus 



H. decagrammus 



H. stelleri 



P. monopteryglus 

 Hemilepidotus spp. 



Figure 7 



Results of recurrent group analysis on neuston data (larvae) for all cruises ( 1981-86). 

 Boxes enclose members of recurrent groups that have affinity levels of 0.4 or higher 

 with each other. Lines connect taxa with affinities outside their groups. 



i l l i ' ' ''..'' i i i i i 



of P. monopterygius larvae with the outer shelf and 

 slope particularly off Kodiak. Hemilepidotus spp. had 

 a similar pattern of distribution. 



Two species which were in- 

 cluded in the analysis, but did 

 not display significant affinities 

 with any of the other taxa, were 

 Theragra chalcogramma and 

 Zaprora silenus. It seemed that 

 at least in the neuston, T. 

 chalcogramma, which is the 

 dominant larval taxon in this 

 region, had a unique pattern of 

 occurrence, largely dissimilar to 

 the other neustonic larvae. The 

 lack of affinity of Z. silenus with 

 other species was probably due 

 to its infrequent occurrence in 

 these samples. 



Kendall and Dunn ( 1985 ) and 

 Rugen 3 found a variety of recur- 

 rent groups and inter-species 

 linkages among the neustonic 

 fish larvae in the western Gulf 

 of Alaska over four seasons. The 

 species groups and affinities 

 changed seasonally and were 

 inconsistent among two-week 



random (Fig. 8) but di 

 trends which reflected 



sampling periods. Similar to 

 the results presented here, T. 

 chalcogramma did not occur 

 among the recurrent groups 

 or associated linkages of spe- 

 cies identified in these studies. 



Five species groups and 

 eight sector groups (sectors= 

 sampling sectors in Fig. 2) 

 were identified from the 

 agglomerative hierarchical 

 classification of data com- 

 bined for all the cruises 

 (Table 5). The Bray-Curtis 

 dissimilarity coefficient val- 

 ues at which these groups 

 were formed were high ( mini- 

 mum value of 0.63), particu- 

 larly among the sector groups. 

 These indicated that the 

 groupings were weak and 

 that species were only loosely 

 affiliated with each other in 

 terms of density and distri- 

 bution patterns. 



The distribution of the 

 eight sector groups seemed 

 splayed certain geographical 

 a variety of distribution pat- 



i i i iii .i. i 



I I ,'.U' ' A+AJ i 



T tV*R 



4 2 5 8 3 1 

 192274722 



I 7 2 

 2 8 12 



I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 



Figure 8 



Distribution of sector groups resulting from numerical classification of density data 

 for the dominant taxa offish larvae in the neuston, based on all cruises ( 1981-86). A 

 plus sign ( + ) indicates that no fish larvae were caught in that sector. 



