that the highly mobile animals were the random element. We know that we 

 violate this premise, since these animals are not randomly distributed 

 but as social animals are found in aggregations, and this very patchiness 

 of sightings leads to additional problems with statistical analyses. 



Another assumption that must be made in a survey of this nature is 

 that the estimate of distance from transect line to target is accurate. 

 Estimation of distances over open water with no landmarks for guidance is 

 extremely difficult and the precision of estimations of distance must be 

 in doubt. However, whatever error or bias that might exist in our dis- 

 tance estimation should be consistent, since the method of the distance 

 estimation was the same from observer to observer. 



For an enumerator to be effective and consistent, the probability of 

 detection of each target must be the same and the detection of one 

 target should not lead to the detection of additional targets. For 

 cetacea this is not the case; small schools may be detected by the behavior 

 (e.g. aerial behavior) of a single animals, while large schools may be 

 located from the actions of only a small proportion of the group. 



Other variables which may aid or hinder the "see-ability" of cetacea 

 are the animal's size, color, or type of movement. In addition, environ- 

 mental factors such as sun angle, light levels, glare, sea state, and 

 visibility all contribute to the probability of detection. And we have 

 yet to consider observer bias. Does he see more when fresh at the start 

 of a day's survey; does the sighting of one school "perk" him up so that 

 he is more likely to see subsequent groups? 



We know that on many, if not all, occasions we have violated from 

 three to five of the assumptions necessary to maintain the statistical 

 accuracy for any line transect theory, and realize we probably grossly 

 underestimate or "undersee" the number of cetacea on any survey. In 

 spite of these errors in methodology, common to all current survey tech- 

 niques, we present five common methods of computation for determining the 

 relative abundance of cetacea. Animal densities/nm from each of the 

 five formulae are then extrapolated to arrive at Bight-wide estimtes of 

 populations. Each of the formulae is outline in turn below, followed by 

 consideration of the most useful and reasonable method to assess realis- 

 tic population numbers. 



Formula #1 . Index of Abundance: 



N = Index 



transect length x path width 



where N = number of animals observed; 



where L (transect length) = linear distance flown in nm; 



Where W (path width) = right angle distance off flight line which 



observer scanned. If each side of aircraft was 

 manned by observers, right angle distance is 

 doubled. In this formula right angle distance 

 was considered to be 1 nm and, since both sides 



B-70 



