252 



Fishery Bulletin 98(2) 



into geographical areas based on the spa- 

 tial distribution of rookeries (Fig.l) and 

 the reported animal movements between 

 rookeries in northern Patagonia (Crespo 

 and Pedraza, 1991; Crespo^; Dans et al.^). 

 Sex and standard length (SL, cm) were 

 recorded when possible. Males ranged from 

 114 to 243 cm SL, whereas females ranged 

 between 102 to 196 cm SL (Fig. 2). 



Stomach content analysis 



Stomach contents were preserved in 70% 

 alcohol or frozen at -20°C. Hard pieces were 

 recovered by using sieves of different mesh 

 sizes (from 0.5 to 10 mm) flushed with water 

 and by using decantation trays. Fish oto- 

 liths and bones, cephalopod beaks, crusta- 

 cean exoskeletons and other hard remains 

 were used to quantify and identify the prey 

 species. Identification was made by using 

 local species reference collections at the 

 Marine Mammal Laboratory, Centro Nacio- 

 nal Patagonico, CONICET, and available 

 catalogues (Clarke, 1986;Mennietal., 1984; 

 Roper et al. , 1984; Boschi et al., 1992; Gosztonyi and 

 Kuba''). Complete and undigested elements (com- 

 plete prey, otoliths and beaks) were measured with 

 digital calipers. When digested and broken hard 

 pieces were found in a stomach, the measurements 

 for these elements were assigned from a random 

 sample of undigested and whole parts of the same 

 species obtained within the same stomach (Koen 

 Alonsoetal., 1998). 



Size (total length ([TL]) offish and dorsal mantle 

 length (IDML]) of squid, cm) and wet weight (W, g) 

 of prey were estimated from hard pieces by using 

 allometric regressions (Clarke, 1986; George-Nasci- 

 mento et al., 1985; Koen Alonso et al., 1998) (Table 

 2). In those cases where regressions were not avail- 

 able, regressions of related species were employed 

 (Table 2). When related species regressions did not 

 exist, weight was assigned by direct comparison with 

 measured and weighed individuals of similar size for 

 the same species or by weighing the fragments found 

 in the stomach. The presence of stomach stones and 

 parasites was recorded and all the gastroliths were 

 weighed in each stomach. 



110 130 



150 170 190 210 

 Standard length (cm) 



230 250 



I Males a Females 



Figure 2 



Length-frequency distribution of the total sample of South American 

 sea Hons analyzed in this work. Three males and four females were 

 not included in this figure because their standard lengths could not be 

 obtained. 



•* Gosztonyi, A., and L. Kuba. 1996. Atlas de huesos craneales y 

 de la cintura escapular de peees eosteros patagonicos. Informes 

 Tecnicos del Plan de Mancjo Integrado de la Zona Costera 

 Patagonica. Fundacion Patagonia Natural i Puerto Madryn. 

 Argentina) 4:1-29. 



Data analysis 



The relative importance of prey species was evalu- 

 ated by means of the index of relative importance 

 (IRI) (Pinkas et al., 1971). The IRI was calculated for 

 each prey species as 



IRI = i'7(N -^^WWcFO, 



Where %FO = the percent frequency of occurrence; 

 %N = the percentage by number; and 

 %W = the percentage by regression-esti- 

 mated wet weight. 



This IRI is a modified version of the index where the 

 original term of percentage by volume was replaced 

 by the '/fW term (Koen Alonsoetal., 1998). In order to 

 make easier the interpretation of the IRI, this index 

 was expressed on a percent basis i'JdRI) (Cortes, 

 19971. Graphical representation of the diet was also 

 employed to present some results (Cortes, 1997). 



Two overlap indices, the general overlap index 

 (GO) and the specific overlap index (SO) (Petraitis, 

 1979; Ludwig and Reynolds, 1988), were used to 

 examine dietary differences. These indices were 

 selected because they are based on the same theo- 

 retical framework, have associated statistical tests 

 (Petraitis, 1979), and the GO presents a small bias 

 even when the sample size is small (Smith and Zaret, 

 1982). 



