Perkins and Edwards: Capture rate as a function of scfiool size for Stenella attenuata 



553 



mixed schools in our analysis. Thus, we took as our 

 population of schools not just those composed purely 

 of NE offshore spotted dolphins, but all schools con- 

 taining them. School sizes were taken as the total 

 number of animals in each school. This approach 

 would not have been appropriate if we had been es- 

 timating a stock-specific abundance (e.g. Wade and 

 Gerrodette, 1993). However, as long as there is no 

 bias in the species composition of schools that are 

 set on, our approach is valid. An exploratory data 

 analysis indicated that the distribution of species 

 proportions was very similar for both research and 

 tuna vessel data. 



There was some indication that pure spotted 

 schools tended to be smaller on average than mixed 

 spotted-spinner schools. We did not pursue this be- 

 cause it did not affect our results. 



Encounter rate for very large schools 



Inspection of Figure 2 raises the question of why so 

 few very large schools { 1000 animals or gi'eater) were 

 sighted from the research vessels when so many were 

 set upon by tuna vessels. Only five schools ( f/r of 

 sightings) in that range were reported by research 

 vessel obsei-vers, and the largest was estimated to 

 be 2617 animals. In that range, 896 schools (26"^ of 

 observed sets) were reported set upon by tuna vessel 

 observers, and 97 were estimated to be larger than 

 2617 animals. These largest schools from the set data 

 did tend to include slightly higher percentages of 

 species other than spotted dolphins. However, they 

 were still primarily made up of spotted dolphins (just 

 over an estimated 7Q'7( on average), and it was not 

 the case that they were due to an association with 

 large groups of, for example, common dolphins (Del- 

 phinus delphis). which are known to form very large 

 schools (e.g. Edwards and Perrin, 1993). 



At least four explanations for this apparent dis- 

 crepancy are possible. First, this may simply reflect 

 the tuna vessel captains" preference for setting on 

 large schools. Second, the difference may be due to 

 relative bias in size estimation between the two types 

 of observers, as discussed earlier. However, to explain 

 all of the difference, the two sets of obsei'vers would 

 have to differ on average by a factor of five in their 

 estimates. Third, the research vessels may have 

 missed a relatively rare segment of the population 

 of schools, which the tuna vessels are able to seek 

 out with a much greater search effort and a nonran- 

 dom search strategy. Fourth, some of these large ob- 

 servations in the set data may have been from in- 

 tentionally repeated sets on the same schools. There 

 is evidence in the tuna vessel obsei^ver data for both 

 of these last two explanations, i.e., that localized ar- 



eas of high density or school size (or both) may exist 

 and that repeated sets on a single school may occur. 



Conclusions 



The results of this study indicate that tuna purse- 

 seiners in the ETP fishing on NE offshore spotted 

 dolphins have a strong preference for setting on 

 larger than average dolphin schools, and that such 

 schools were subject to being set on at a much higher 

 rate than were smaller schools. Specifically, the larg- 

 est schools considered, those of 1000 or more ani- 

 mals, were estimated to be set on approximately once 

 every week, whereas the smallest schools considered, 

 those of 100 animals, were estimated set on less than 

 once a year. Our estimated capture rates should be 

 taken as averages for a given school size and do not 

 account for variation due to other factors, such as 

 geographic location or the amount of associated tuna. 

 Also, although we estimated rates in terms of sets 

 per year, we do not assert that the short-term cap- 

 ture rate for a given school is constant, i.e. that sets 

 occur at evenly spaced intervals throughout the year. 

 For example, relatively few sets are made "on dol- 

 phin" in the NE offshore stock range during June 

 and July (e.g. Edwards and Perkins, 1998). 



These results do not account for any errors in esti- 

 mation of dolphin school size. Although potential er- 

 rors in school-size estimates made by research ves- 

 sel observers were investigated, no corresponding 

 study of potential errors for tuna vessel observer es- 

 timates was possible. 



To draw conclusions about capture frequency for 

 an individual dolphin, we must consider the size 

 range of the schools with which a given individual 

 tends to associate. Our results imply that dolphins 

 associating primarily with large schools will be sub- 

 jected to capture much more often than individuals 

 associating primarily with small schools. However, 

 we also estimated that the largest schools are rela- 

 tively rare and account for a minority of the total 

 number of individual dolphins at any given time. 

 These results may imply that a fixed but relatively 

 small percentage of the dolphin population was con- 

 sistently subjected to a high rate of capture in purse- 

 seine nets but that a majority of dolphins occur in 

 schools smaller than those apparently preferred by 

 purse-seiners, and experience relatively few captures 

 per year. 



However, little is known about the spatial and tem- 

 poral dynamics of dolphin schools and their sizes, 

 and other conclusions are possible. If dolphins asso- 

 ciate with a wide range of school sizes, then the cap- 

 ture rates for individual dolphins would tend to "av- 



