FISHERY BULLETIN; VOL. 86, NO. 2 



Figure l. — A hypothetical cruise track for a vessel searching 

 for clustered prey. The location of catches are indicated by 

 xs. 



underlying distribution of the fish population 

 (Koopman 1980). The question of optimal search- 

 ing strategies for fishermen has been receiving 

 increased attention (Pazynich 1966; Salia and 

 Flowers 1979; Clark and Mangel 1983). These 

 studies are primarily theoretical at this time and 

 their application to actual fisheries requires 

 knowledge of the spatial distribution of the fish 

 population. 



Locating schools of tuna (e.g., yellowfin, Thun- 

 nus albacares, and skipjack, Katsuwonus 

 pelamis) in the purse seine fishery in the ETP 

 depends on visual cues. Fishermen use a variety 

 of cues including birds which feed on the same 

 prey as tuna, disturbances on the surface of the 

 water, floating debris which frequently have as- 

 sociated tuna, and schools of dolphins which are 

 often associated with tuna (primarily yellowfin). 

 Fishermen have names for the different types of 

 sets depending upon what is associated with the 

 tuna school. They refer to sets associated with 

 floating debris as log sets, sets associated with 

 dolphins as porpoise sets, and sets not associated 

 with other animals (except possibly birds) as 

 school sets. Fishermen when not engaged in a set 



usually spend their day actively searching for 

 signs of tuna. They use 25 x binoculars to scan the 

 water while the boat cruises at speeds generally 

 between 10 and 12 knots. 



The distribution of schools of tuna, as well as 

 schools of the most commonly associated dolphins 

 (i.e., the spotted dolphin, Stenella attenuata), ap- 

 pears to be spatially and temporally clustered 

 within the ETP considered as a whole (Calkins 

 and Chatwin 1967, 1971; Blackburn and Wil- 

 liams 1975; Suzuki et al. 1978; Au et al. 1979^; 

 Polacheck 1983). At finer geographic scales, there 

 is little available information although Au et al. 

 (fn. 2) suggested that schools of spotted dolphin 

 tend to be locally concentrated in areas of conver- 

 gences and fronts. 



Given the above observation, it is not surpris- 

 ing that the detection or encounter process for 

 tuna or dolphins does not, in general conform to a 

 Poisson process when the distribution of search- 

 ing times, searching distances, or physical dis- 

 tances between nearest encounters are analyzed 

 (Polacheck 1983; Allen and Punsely 1984). In 

 such analyses it is impossible to separate or dis- 

 tinguish the effects of nonrandom search from 

 nonrandom distributions of tuna or dolphins. 



In the harvesting of tuna schools associated 

 with dolphins, fishermen chase and capture the 

 associated dolphins (Perrin 1968, 1969), and some 

 dolphins may be incidently killed. NMFS, as part 

 of its responsibility under the Marine Mammal 

 Protection Act of 1972 for managing and monitor- 

 ing the status of dolphin populations, placed 

 trained observers aboard tuna purse seiners. 

 From the data collected by these observers, ap- 

 proximate cruise tracks can be drawn by connect- 

 ing all positions that were recorded. Many of 

 these approximate cruise tracks (e.g.. Figure 2) 

 have superficially a strong similarity to the hypo- 

 thetical one depicted in Figure 1. It was this sim- 

 ilarity that provided the impetus for the analyses 

 presented below. 



METHODS 



NMFS observers aboard tuna purse seiners col- 

 lected a wide variety of information, both on the 

 sightings of marine mammals and fishing opera- 



2Au, D. W. K., W. L. Peryman, and W. F. Per- 

 rin. 1979. Dolphin distribution and the relationship to envi- 

 ronmental features in the eastern tropical Pacific. Natl. Mar. 

 Fish. Serv., Southwest Fish. Cent., Adm. Rep. LJ-79-43, 59 p. 



352 



