520 



Fishery Bulletin 105(4) 



Dall's porpoise Abundance estimation for Ball's por- 

 poise is difficult because of their attraction to vessels 

 (Turnock and Quinn, 1991). To obtain unbiased esti- 

 mates, these animals must be detected before they react 

 to the survey vessel. Our data indicate that the behavior 

 of the vast majority of Ball's porpoise seen at low sea 

 states is "slow rolling." This contrasts with the "rooster- 

 tailing" or fast swimming behavior seen by animals that 

 are approaching the ship. However, when effort is limited 

 to calm conditions (Beaufort states 0-2), the amount of 



search effort is greatly reduced (Fig. 1). As a result, the 

 coefficients of variation for Ball's porpoise abundance 

 are greater than would be expected for such a common 

 species. Off California, the temporal pattern shows 

 higher Ball's porpoise abundance in 1996 (Table 7), 

 mirroring the higher abundance that year of cold-tem- 

 perate delphinids. Forney (2000) found that sea surface 

 temperature was a very good predictor of Ball's porpoise 

 distribution. In their 12 year time series of surveys off 

 central California, Keiper et al. (2005) also found that 



