FISHERY BILLETIN: \( )L. Ki. NO. -.i 



Dragovich (1969) reviewed existing information on 

 the food habits of Atlantic skipjack tuna. Since that 

 time food habits have also been reported in studies 

 by Dragovich (1970) and Dragovich and Potthoff 

 (1972) for skipjack from the East and West Atlantic 

 and by Batts (1972) for skipjack in North Carolina 

 waters. Zavala-Camin (1981) examined predator- 

 prey interactions of fishes, including skipjack cap- 

 tured north of the area in this study. 



The primary objective of this study was to discover 

 if skipjack tuna feed upon their young. The presence 

 of juveniles in bait-caught skipjack stomachs would 

 verify the study area as a spawning-rearing ground. 

 Knowledge of the prey and their relative importance 

 also contributes to the understanding of prey- 

 predator interactions, which affect population 

 distributions and fluctuations. 



MATERIALS AND METHODS 



Stomach samples for this study were collected on a 

 monthly basis from October 1981 to December 1982 

 from skipjack tuna caught off Rio de Janeiro (Fig. 1). 

 National Marine Fisheries Service (NMFS) person- 

 nel collected stomachs from frozen fish transhipped 

 to Puerto Rico, and Superintendencia do Desenvol- 

 viemento da Pesca (SUPEDE) personnel sampled 

 fish landed locally in Rio de Janeiro. Fish from the 

 Puerto Rican source were caught within 1 mo prior 

 to sampling; fish from the Brazilian source were sam- 

 pled 3 to 5 d after the recorded catch date. The sam- 

 pling design required collecting about 15 stomachs 

 from each 10 cm length group, measured to the 

 nearest cm per month. However, the number of 

 stomachs collected was dependent on the catch-size 

 distribution. Once the stomach was removed from 

 the fish, it was preserved in 10% buffered Formalin- 

 and shipped to the Southwest Fisheries Center 

 (SWFC) for analysis. 



Stomachs were examined from 1,041 fish between 

 44 and 81 cm fork length. In the laboratory each 

 stomach was opened. The volume of the food bolus 

 was measured, and the contents were identified to 

 the lowest possible taxon. The taxonomic groupings 

 were then measured by volumetric displacement. 

 and the individuals counted. Whole undigested fish 

 were identified by comparing external characters 

 with those described in published keys or with iden- 

 tified museum specimens from Scripps Institution of 

 Oceanography, La Jolla, CA. Digested animals, par- 

 ticularly juvenile scombrids, were identified by verte- 



bral, gill raker, and fin ray counts, as well as other 

 skeletal characteristics, described by Potthoff and 

 Richards (1970), Miller and Jorgenson (1973), and 

 other published keys. Cephalopods were identified by 

 comparing beak characters with published illustra- 

 tions, descriptions, and keys (see Wolff 1981). Crus- 

 taceans and other invertebrates were identified by 

 specialists from Scripps Institution of Oceanography 

 and SWFC. 



The occurrence of bait in the stomachs may have 

 biased the relative importance of fish in the diet. The 

 bait primarily consisted of Sardinella brasiliensis, 

 Harengula jaguana, and Engraulis anchoita; 

 however, other fish families may have been included 

 in the captured bait. The sardines were readily iden- 

 tifiable from their external characters and usually 

 were undigested. The anchovies, in contrast, were 

 often quite digested, creating difficulties in identifi- 

 cation. Gary Nelson^ nevertheless was able to verify 

 these fish as Engraulis anchoita. Although the least 

 digested item in the stomach was usually the last 

 meal (bait), stomachs were removed from a few days 

 to 1 mo after capture, and presumably postmortem 

 digestion occurred. As a result, the degree of diges- 

 tion was not a reliable indication of distinguishing 

 bait from natural prey. The time required for com- 

 plete gastric evacuation of smelt fed to skipjack tuna 

 is estimated to be 12 h (Magnuson 1969). Although 

 the bait was captured in nets from bays and estuaries 

 (Rinaldo4), Matsuura et al. (1978, 1981) have con- 

 firmed that a spawning stock of £■. anchoita does ex- 

 ist in waters inhabited by skipjack tuna. It is unlikely 

 that the sardines served as prey for skipjack. 

 However, I could not distinguish between E. an- 

 choita consumed as natural food and as bait. There- 

 fore, although these species were considered bait, 

 some may have been ingested as natural food. Bait 

 was not considered prey, and stomachs containing 

 only these species were treated as empty. 



Two methods of analysis were employed to rank 

 the food items in terms of availability and impor- 

 tance to the skipjack tuna: 



1) An index of relative importance (IRI) was calcu- 

 lated for each prey type in terms of numbers, 

 volumes, and fre(|uencies (Pinkas et al. 1971): 



IRI = {N + V)F 



''Reference to trade names does not innply endorsement by the Na- 

 tional Marine Fisheries Service, NOAA. 



'G. J. Nelson, Department of Ichthyolojjx-. American Museum of 

 Natural History, New York, NY 10024. pers. commun.. May 1982. 



■•R. R. Rinakid, Southwest Fisheries Center La .lolla Laboratory, 

 National Marine Fisheries Service, NOAA. I'.O. Box 271, La Jolla, 

 CA 92038, pers, commun., .June 1982, 



380 



