POLACHECK: DISTRIBUTION OF SEARCHING EFFORT 



Does Searching Behavior Bias 



the Overall Estimate of 



Dolphin Encounter Rates? 



When the two different methods of estimating 

 the overall encounter rates within a cruise are 

 compared, the adjusted encounter rates tend to be 

 smaller (Table 7). However, the differences are 

 not large, not because the biases are necessarily 

 small, but because positive and negative effect of 

 concentrating searching effort tend to cancel each 

 other. Positive biases would be expected in the 

 unadjusted rate due to concentrating of searching 

 effort in clusters dominated by sets made in asso- 

 ciation with dolphins, since these appear to be 

 areas of high dolphin densities (Table 5). Simi- 

 larly, negative biases would be expected due to 

 searching in non-dolphin clusters. For the data 

 considered here, substantial and roughly equal 

 amounts of searching occurred in both types of 

 clusters. The effects of concentrating searching 

 effort within the two types of clusters tended to 

 cancel each other. This suggests that the major 

 effect of nonrandom searching on measures of 

 relative abundance of dolphins will change in re- 

 lationship to the importance of dolphin and non- 

 dolphin fishing. Such changes tend to occur with 

 changes in the relative abundance of large yel- 

 lowfin tuna compared to skipjack and small yel- 

 lowfin tuna. 



DISCUSSION 



The fundamental question in interpreting the 

 results of this paper is whether the clusters that 

 have been defined bear any relationship either to 

 the searching strategy of the vessels or to the 

 underlying distribution of dolphin and tuna 

 schools. Two factors hamper answering this ques- 

 tion: The first is the large variability among ves- 



sels; the second is the lack of appropriate null 

 hjqjotheses by which to test the results. The large 

 variability is to be expected and is inescapable. 

 Not only is there a large stochastic element in the 

 catch and encounter process, but large differences 

 can be expected in searching strategies among 

 vessels. Thus, a large range exist in skill and 

 experience among fisherman. In addition, the 

 amount of information, which is shared among 

 vessels varies and some vessels may be acting as 

 scouts for other vessels (Orbach 1975). 



The specification of null hypotheses is difficult 

 because an infinite number of searching models 

 are compatible with the definition of random 

 search (i.e., defining a random search as one in 

 which the search path is independent of the dis- 

 tribution of the objects being sought). In order to 

 actually model a random searching process, the 

 probability of changing the direction of the search 

 path needs to be specified. A random search could 

 encompass an)^hing fi-om Brownian motion to 

 random straight line crossings of an area. With a 

 finite amount of searching, these will not neces- 

 sarily yield the same results. 



Most of the results from this paper are compat- 

 ible with a model of clustered searching for clus- 

 tered prey, and some of them seem hard to explain 

 unless the searching and the schools of dolphins 

 are nonrandomly distributed. That the distance 

 travelled relative to the actual distances between 

 sets tends to be greater in areas where the density 

 of sets is high, as is the proportion of this distance 

 which is spent searching, are unlikely results un- 

 less searching is concentrated in these areas. 

 Also, the higher encounter rates in clusters dom- 

 inated by dolphins, the comparisons of the dis- 

 tances and frequency of the nearest set with the 

 preceding or next set, and the high percentage of 

 sets which fall into clusters are results that would 

 be expected if searching, tuna, and dolphins were 



Table 7. — Comparison between the unadjusted overall encounter rate (nunnber of 

 schools of Stenella attenuata per 1 00 miles of searching) for a cruise and the encounter 

 rate adjusted for possible bias due to the concentration of searching within clusters. 



'One cruise had no recorded sighting of spotted dolphin schools. 



363 



