Local fish species were used to feed a captive blue-eyed shag Phalacrocorax atriceps bransfieldensis during 45 days of the austral summer at Jubany Station, King George Island, South Shetland Islands. The otoliths identified in pellets were affected by the digestive process and consequently the fish species were differentialy underrepresented in number and size (length/mass). Except Gobionotothen gibberifrons and Nototheniops nudifrons all the species were underestimated in number (N. rossii fully). G. gibberifrons was also the less affected is size, being Pagothenia bernachii, N. nudifrons, Trematomus newnesi and Notothenia coriiceps largely underestimated. Thus, preliminary correction factors were obtained to improve the accuracy of weight estimations of fish ingested, calculated y means of equations based on otolith-lengths. The shag produced a total of 16 pellets, with a frequency of 1 every 2.5 days. It ingested willingly a mean ration equivalent to 31% of its mass, which is a higher energy requirement than that experimentally observed in other not antarctic shag species. Algae and ploychaetes were found in the casts and came from the fish stomachs. Therefore, their importance in the diet of the blue-eyed shag could have been overestimated in previous studies.