McKenna. Spatial structure and temporal continuity of South Georgian fish community 487 



■o 

 o 



E 



0.05 0.15 25 35 0.45 0.55 65 0.75 0.85 0.95 



Class of |Tau| 



B 



1986-87 AMLR Survey Data 



random 

 st#.#'s prop. 

 sl# bio prop 

 Sl# #S.# 



k. 



005 015 0.25 



35 45 55 0.65 0.75 85 95 



Class of )Tau| 



Figure 10 



Results of skewer analysis of the 1986-87 AMLR survey data. 

 LEGEND: random indicates the bar representing the results 

 from the 20 random tables generated by each analysis, long. 

 = longitude; st# = station number; # = analysis based on 

 numerical abundance; bio. = analysis based on biomass; prop. 

 = use of proportion. (A) Histogram indicating the frequency 

 of the location of the primary mode while examining the trend 

 in longitude; (B) histogram indicating the frequency of the 

 location of the primary mode while examining the trend in 

 date (as indicated by station number). 



sible explanation for changes in size structure of 

 some of the species within a survey. However, those 

 trends were all weak and inconsistent. In general the 

 results indicate that there was little or no structure to 

 the demersal fish community of South Georgia Island. 

 No areas that could be considered nurseries were 

 identified. Nor were there areas of particularly high 

 abundance of only the most valuable commercial spe- 

 cies. Variability in the abundance of each species was 

 evident. However, essentially the same assemblage 



with similar dominant species was found at nearly 

 every sampled location on the continental shelf. 

 This consistency in the species assemblage has been 

 noted in the South Georgia region before (Targett, 

 1981). 



The absence of spatial structure is unusual in natu- 

 ral communities and it is unclear why this situation 

 exists in the South Georgia Island system. The species 

 within this community can be classified into three tro- 

 phic groups: krill-eaters (e.g., C. gunnari and N. 

 larseni), piscivores (e.g., C. aceratus and P. georgianus), 

 and benthic invertebrate feeders (e.g., N. gibberfrons 

 and N. squamifrons) (McKenna, 1991). One would ex- 

 pect these animals to be distributed according to the 

 availability of their prey or appropriate habitat (or 

 both). Although the trophic groups are clear, they are 

 closely linked and the distribution of all these species 

 may be dependent on the abundance and distribution 

 cf krill (Euphausia superba ). If krill was superabun- 

 dant then, even if it were patchy, it might have more 

 than met the demands of the krill-eaters, explaining 

 their uniform distribution. The piscivores will follow 

 their fish prey, which fed either on krill or some benthic 

 resource. The benthic invertebrate feeders were also 

 uniformly distributed. Little is known about the benthic 

 community around South Georgia Island. However, it 

 has been shown that most of the organisms of that 

 habitat are inedible to fish (Belyaev and Ushakov, 

 1957). It might be expected then that the edible benthic 

 resources are distributed in patches within this habi- 

 tat. That was not apparent from the distribution of 

 the fish and may have been concealed by the fact that 

 even the benthic feeders ate krill (McKenna, 1991), 

 including N. gibberifrons and N. nudifrons. These two 

 species are highly adapted to feeding on benthic or- 

 ganisms and yet their diets included krill. Targett 

 (1981) also observed krill in the diet of these benthic- 

 feeding species and suggested that a shoal of krill had 

 moved into shallow water and was forced close to, or 

 in actual contact with, the bottom. This would seem to 

 be an indication of the great abundance of krill in the 

 vacinity of South Georgia Island. 



The results of this research imply that during the 

 austral summer there was, in general, a uniform dis- 

 tribution of the necessary resources throughout the 

 sampled region, at scales of 10 km to 100's of km. It 

 must be emphasized that these surveys took place only 

 in the summer and little is known about this commu- 

 nity at other times of the year. The strong seasonality 

 of this environment may present the inhabitants with 

 an annual 'boom' and 'bust' cycle, having abundant 

 resources in the summer and a 'bottleneck' period 

 (when resources are less available than at other time 

 of year) during the winter (Wiens, 1977; DuBowy, 1988). 

 The abundance of available resources during the sum- 

 mer may allow different species to feed on the same 



