344 



Fishery Bulletin 89(3). 1991 



In general, polyspecific associations seem to form 

 when social species of similar foraging ecology join to 

 form larger groups to increase feeding success and to 

 better avoid predators. The former may be ascribed to 

 facilitation of feeding or to greater access to resource 

 information (Clark and Mangel 1984). An additional 

 benefit might be a lessening of intragroup competition 

 (relative to monospecific groups of the same size). The 

 enhanced safety could result from more effective 

 vigilance (Pulliam 1973) or to the "convoy" effect 

 (Olson 1964). It seems that the behaviorally more ver- 

 satile species often exploit other species' behaviors, 

 although there appear to be mutual benefits as well. 

 For the most part, these polyspecific associations seem 

 loosely coupled; the overt interspecies behaviors are 

 circumstantial rather than obligatory, and generally 

 agonistic. Species appear to join and leave the groups 

 mainly in response to foraging situations; according- 

 ly, they may be together for hours or days, sometimes 

 months or even years. 



In this paper I describe species found with tuna in 

 the eastern tropical Pacific (ETP) and propose that 

 tuna schools have a polyspecific nature. What this im- 

 plies about tuna biology, especially that of the yellow- 

 fin Thunnus albacares, is discussed. 



Data and methods 



I used data collected by Southwest Fisheries Science 

 Center (SWFSC) observers placed aboard U.S. tuna 

 purse seiners. Such data have been available since 

 1966, but the records have not always been consistent 

 or extensive with respect to other species caught with 

 tuna (except for dolphins). The reasons include the 

 pressure of observers' duties and the proprietary 

 nature of fishing operations. After 1975, some kinds 

 of bycatch information were no longer collected at all, 

 and only abbreviated codes for certain other associated 

 species were recorded. 



The earlier, pre-1976 observer records, however, 

 contain many candid notes on bycatches and other tuna 

 associates, although even then the logs required only 

 generic, not comprehensive, descriptions. Fortunate- 

 ly, some observers clearly were interested in tuna- 

 associated fauna, and they frequently recorded detailed 

 descriptions of what they saw. 



I therefore selected records from the more infor- 

 mation-rich 1974-75 database on purse seine sets (a 

 "set" is a launch and retrieval of a purse seine net) for 

 information on species that occur with tuna. I compiled 

 a list of 22 observers personally known by myself or 

 other SWFSC workers to have been reliable or to have 

 had keen interest in pelagic fauna, and obtained perti- 

 nent information from their original logsheets. These 



logsheets constituted the set records of each of their 

 33 fishing trips. The trips occurred during all months 

 except December, with 90% of the data collected 

 between January and mid- June. 



I examined the data pooled over all fishing trips and 

 also by individual trip, since the quality of the records 

 varied among observers. The data from one T.M. Duffy 

 (TMD) is presented separately in the analysis below 

 (TMD sets cf. the regular fleet sets) to illustrate how 

 the actual observations of a particular observer could 

 differ from the overall, pooled results. Duffy recorded 

 information copiously and was an experienced and 

 interested observer. 



The data were divided into two areas and according 

 to how the tuna were caught. The areas, delimited by 

 latitudes 15°S and 25°N, were from the Central Ameri- 

 can coasts westward to 100°W ("off Central America") 

 and from 100°W out to 135°W ("off Mexico") (see 

 Figure 1). The first area contained sets mainly off the 

 Central American coasts and southwestward to the 

 Galapagos Islands; this is an important area for fishing 

 both free-swimming and log-associated tuna, and also 

 dolphin-associated tuna (explained below). The second 

 area had sets mainly between southern Mexico and 

 Clipperton Island, where much dolphin-associated tuna 

 is taken. The tuna sets were further partitioned accord- 

 ing to whether they were made around floating logs 

 ("log-fish" sets), on free-swimming tuna ("school-fish" 

 sets), or on dolphin-associated tuna ("porpoise-fish" 

 sets). These are the terms used by the fishermen for 

 the variously caught tuna, i.e., for different set types. 

 The porpoise-fish sets refer to tuna caught usually with 

 spotted and spinner dolphins, Stenella attenuata and 

 <S. longirostris (called "porpoise"). Table 1 gives the 

 numerical breakdown of the 1762 purse seine sets I 

 examined according to the above categories. 



The purse seine sets in each area and set-type cate- 

 gory were examined for frequency of different associ- 

 ated species to obtain a profile of the multispecies 

 composition of the tuna schools. For each of certain 

 species— actually grouped species because identifica- 

 tions were inexact— this frequency was expressed as 

 the fraction (p) of the pertinent sets in which the species 

 was caught. For comparison, I calculated both overall 

 fractions considering all sets from all trips and arith- 

 metic means of the fractions for individual trips. The 

 latter fractions were computed only where fishing trips 

 had more than ten sets in the category of interest. 

 Binomial 95% confidence limits of the overall fractions 

 were determined in accordance with the number of sets 

 involved (n). Ranges were determined for the trip frac- 

 tions. These latter fractions were statistically hetero- 

 geneous among trips, at least for log-fish and por- 

 poise-fish sets. Thus the overall fractions probably best 

 reflect the polyspecific composition of schools. 



