FISHERY BULLETIN: VOL. 86, NO. 4 



C/) 

 Q 

 OC 



lU 



I 



DC 

 LU 

 CD 



25 



20- 



15 



10- 



5- 



22 



r~i . r 



1 1 



Figure 2.— Distribution of bottlenose dolphin herd sizes. Numbers above 

 bars denote the number of herds in that size class. 



independence between herd size and perpendicular 

 sighting distance. 



Bottlenose dolphin abundance is the product of P 

 and the area surveyed. In the CBM area this may 

 be extrapolated to the total area if the transects are 

 distributed randomly with respect to dolphin sight- 

 ings. 



The recorder did not distinguish between observ- 

 er's sightings when recording them, thus observer 

 bias was not investigated. The effects of sea state 

 and sun glare on detectability were not investigated. 

 Surveys were not conducted when sea states were 

 above two on the Beaufort scale, and it is unlikely 

 that sea state influenced the results. However, the 

 effect of glare reduced the observers' field of view, 

 which decreased the number of animals detected and 

 resulted in an underestimation of P. 



I conducted photographic surveys from a 7 m boat 

 on five occasions in 1980 and six in 1981 for the 

 purpose of identifying individual bottlenose dolphins 

 by the shape of, or markings on, their dorsal fins. 

 Contact prints of the 35 mm photographs were 

 examined under a dissecting microscope at 40 x 

 magnification. 



RESULTS 



Six aerial surveys in the CBM averaged 119.4 km 

 per survey, covered an area of 762 km^, and re- 

 sulted in five herd sightings of bottlenose dolphins. 

 Ten surveys along the southern Virginia coast re- 

 sulted in 49 herd sightings. Each coastal survey was 

 32.3 km in length and covered an area of 65 km^. 



An additional survey along the northern Virginia 

 coast was 108 km in length, covering an area of 216 

 km^, and resulted in two herd sightings. 



In line transect the distance at which a bottlenose 

 dolphin herd is sighted is assumed to be indepen- 

 dent of its size (Burnham et al. 1980; Seber 1986). 

 Although it seems reasonable that larger herds 

 would be detected at greater distances, analysis of 

 herd size and sighting distance using the method of 

 Drummer and McDonald (1987) showed no signifi- 

 cant size-bias (P > 0.05). As a check, I also regressed 

 herd size against perpendicular sighting distance. 

 There was no apparent association between herd 

 size and distance from the transect (r^ = 0.001) 

 (Fig. 3). The geometric mean herd size was 14.4 

 bottlenose dolphins/herd (SE = 4.0, n = 56). 



Truncation of bottlenose dolphin sightings at 1 km 

 resulted in the discarding of one herd sighting in 

 the southern coastal area, none in the CBM, and one 

 in the northern coastal area. The truncated sighting 

 in the northern coastal area was at approximately 

 1,200 m from the transect and the herd was ap- 

 parently feeding in the wake of a trawler, thus the 

 sighting was atypical of other sightings during this 

 study and probably influenced by the presence of 

 the trawler (see Leatherwood 1975). Both sightings 

 in this area occurred farther offshore than sightings 

 in the other study areas. 



Several parametric and nonparametric models 

 were investigated for fit to the pooled perpendicular 

 sighting distances (Table 1). None of the models dif- 

 fered significantly from the observed distance dis- 

 tributions (chi-square test, P > 0.05). The coefficient 



800 



