494 



Fishery Bulletin 104(4) 



IRI), followed by the fish Chromis crusma (3.54 % IRI) 

 and unidentified Osteichthyes (2.12 9( IRI). Polychaetes 

 were not found in this area, and only one type of prey 

 belonging to the phylum Mollusca was recorded. 



Valparaiso Nineteen prey items were identified; of 

 those, 14 were crustaceans. In this location, three prey 

 items were predominant: the crab C. setosus (34.32% 

 IRI), the shrimp, R. typus (27.38% IRI), and the crab, 

 P. des/narestii (19.09% IRI). The fishes were relatively 

 well represented (unidentified Osteichthyes 8.49% IRI). 

 The Mollusca (2.60% IRI) were represented by the small 

 octopus Robsonella fontaniana. and the snail Turritella 

 cingulata. 



Talcahuano A total of 24 prey items were identified 

 from this location: 18 crustaceans, four fishes, one snail, 

 and one polychaetae. The most important prey item was 

 Mysidacea spp. (70.94% IRI), followed in importance 

 by the crabs P. desmarestii (6.4% IRI) and C. setosus 

 (5.09% IRI). 



Valdivia Twenty four prey items were identified from 

 this area. Mysidacea spp. was the most important in this 

 location (92.13% IRI). However, six of the fish species we 

 identified were represented in this area. 



Aysen Fjord Despite the importance of unidentified 

 prey items in this area, we observed an absence of most 

 of the prey items found in the other locations. However, 

 the trend of the predominance of crustaceans contin- 

 ued. The most important prey item was the megalop 



of P. sicarius (94% IRI) and unidentified Osteichthyes 

 (4.15%). The larval stage of P. sicaiius occurs season- 

 ally; therefore, their importance in the diet of S. capensis 

 from this geographical area can be associated with the 

 sampling period (November). 



Golfo Nuevo In this area, 12 prey items were iden- 

 tified. Lobster krill Miinida subriigosa (32.07% IRI), 

 the brachyuran crab Halicarcinus planatus (23.53% 

 IRI), and Munida gregaria ("15.84% IRI) were the most 

 important prey items. The percentage of unidentified 

 Osteichthyes (22.47% IRI) was higher than in other 

 locations on the Chilean coast. 



Feeding strategy 



Figure 3 shows prey-specific abundance plotted against 

 frequency of occurrence of prey items for S. capensis 

 in the entire study area (Fig. 3A), and in each subarea 

 (termed "location" in this article) (Fig. 3, B-H). This 

 pattern of point distributions was similar for almost all 

 locations. Most of the points (frequency of occurrence) 

 were close to the left side of the .v axis, and they tended 

 to be positioned towards the upper left corner of the 

 graph (specific abundance). There was a high between- 

 phenotype contribution (BPC) to the niche width in 

 Antofagasta (3B), Coquimbo (3C), Valparaiso (3D), Aysen 

 fjord (3G), and Golfo Nuevo(3H) (i.e., individual preda- 

 tors specialized on different prey types and each food 

 category was consumed by only a limited fraction of the 

 predators). A more mixed feeding strategy, with varying 

 degrees of specialization and generalization for different 

 prey types was observed in Talcahuano (3E ) and Valdivia 

 (3F) (i.e., a high BPC and lower within-phenotype con- 

 tribution (WPC) to the niche width). On the other hand, 

 predominant prey items were observed only in Coqui- 

 mbo, Valdivia, and Aysen fjord (R. typus, Mysidacea gen. 

 sp., and the megalops of P. sicarius, respectively); rare 

 or occasional prey items (crustaceans, fishes, mollusks, 

 and polychaetes) were found in all locations. 



Discussion 



To determine the dietary variability of a predator, in 

 order to describe its feeding pattern over its entire 

 geographical range, is a difficult task because of the 

 need for large sample sizes in a number of locations. 

 Additionally, fish, such as S. capensis, whose percent- 

 ages of empty or everted stomachs are high (50% ), make 

 an assessment of diet variability yet more difficult. We 

 managed these difficulties by sampling fish during the 

 same season, sampling over a similar depth range, and 

 by using known information about the geographical 

 distribution of prey items. 



Geographical variation in the diet 



The high percentage of crustaceans in the diet of S. cap- 

 ensis indicates that is primarily a carcinophagous fish, 



