BLAYLOCK: DISTRIBUTION OF THE BOTTLENOSE DOLPHIN 



76°|00' 



76° 00 



Figure 1.— Coastal survey areas and bottlenose dolphin sightings (dosed circles) 

 and Chesapeake Bay mouth survey area (enclosed by dark lines) with dolphin sight- 

 ings (open circles). Dashed line represents transit during coastal surveys. Sightings 

 near Fishermen's Island and Cape Henry which occurred during coastal surveys 

 were included only in the coastal survey analyses. 



where n is the number of herds detected and L is 

 the transect length in kilometers. The variance of 

 D was estimated as (Burnham et al. 1980, p. 51, eq. 

 1.17): 



SHD) = Dnicvin)f + (cv(f(0))2]. 



Herd size was not significantly different between 

 study areas (Kruskal-Wallis test (K-W test), x" = 

 0.9953, df = 2, P = 0.61, Sokal and Rohlf 1981) and 

 sightings were pooled to determine the overall mean 

 herd size. Herd sizes were not normally distributed 

 (Fig. 2) and therefore were normalized by log trans- 



formation to calculate the geometric mean (Sokal 

 and Rohlf 1981) and its variance. Bottlenose dolphin 

 density (P) is the_product of herd density 0) and 

 mean herd size (H). The variance of P, following 

 Goodman (1960, p. 710, eq. 7) is 



,2^ _ H's^jD) DH^{H) sHD)sHH) 



S {r) — , + — — , — 



niD) 



n{H) 



n{Dyn{H) 



with 710) equal to the number of herd sightings in 

 the survey area and n{H) equal to the number of 

 herds used in the estimation of H. This assumes 



799 



