484 



Fishery Bulletin 91(3). 1993 



1.1 



Bosed on 19SB-ag — Nunericol Rbundance 



B 



8.9 



5TRTION 



Based on 198B-Q9 Blonass 



n, 



Figure 5 



Results of cluster analysis of the 1988-89 AMLR survey data. 

 Station labels are read vertically under each branch of the 

 dendrogram. The dashed line represents the level at which 

 significantly different clusters begin. (A) Dendrogram of 

 clustering based on numerical abundance; (B) dendrogram of 

 clustering based on biomass. 



of large, epibenthic invertebrates: branching sponges, 

 corals, tunicates, crinoides, etc.). 



The C. aceratus-P. georgianus association is most 

 likely a result of the similar feeding behavior and habi- 

 tat preferences of these species. Both are relatively 

 sedentary, benthic fish which feed heavily on krill and 

 other fish. Their diets strongly overlap (64%, McKenna, 

 1991) and their distribution may well reflect the dis- 

 tribution of their prey. 



Significant associations suggest some biological in- 

 teractions between the species involved. However, little 

 can be said about the cause of these associations. Two 

 species may display a negative association owing to 

 distinct habitat requirements; they may have similar 

 habitat requirements but experience a niche shift be- 

 cause of competition ( McNaughton and Wolf, 1979b); 

 or they may not coexist because of predation by one on 

 the other. Most of the significant negative associations 

 in this study involved one piscivorous member 



SPECIES GROUPS 



STATION 

 GROUPS 



•O.irj O.10 0.2g| 0.3^ 0.4H 0.5H 0.6|g 0.7| 



Nodal Constancy Diagram (1986-87 biomass) 

 SPECIES GROUPS 



STATION 

 GROUPS 



<'□ '0 =^ m >m 



Nodal Fidelity Diagram (1986-87 biomass) 



Figure 6 



Results of nodal analysis of the 1986-87 AMLR survey data, 

 based on biomass. The width of each band is proportional to 

 the number of entities in the associated group. 



(McKenna, 1991), which suggests predator-prey rela- 

 tionships. The strongest of these negative associations 

 were between M. microps and two of the smallest spe- 

 cies in the region, N. nudifrons and A. mirus (Table 3). 

 These three species are found close to the bottom and 

 much of their diet consists of benthic organisms (37% 

 overlap, McKenna, 1991). However, M. microps grows 

 to over twice the size of the other two (Fischer and 

 Hureau, 1985) and will feed on fish (McKenna, 1991) 

 (though its small mouth probably limits any predation 

 on the other two species to juveniles). Information about 

 available habitats and the diets of the individuals 

 within each habitat are necessary before these rela- 

 tionships may be more clearly defined for the South 

 Georgia system. 



The classification of the species into groups (by in- 

 verse clustering) and their distributions were not 



