76 



Fishery Bulletin 105(1) 



contaminant analysis, lipid and fatty acid, and stable 

 isotope analysis (see Herman et al., 2005). Genetic 

 analysis involved sequencing the entire mitochondrial 

 control region (see Barrett-Lennard [2000] for details). 

 Acoustic recordings were made when whales were vocal- 

 izing and ambient noise levels permitted using an Off- 

 shore Acoustics^*^ (Offshore Acoustics, Nanaimo, British 

 Columbia) hydrophone with a built-in preamplifier and 

 a Sony WM-D6C (B and H Photo, New York, NY) cas- 

 sette recorder. This system had a frequency response of 

 10 Hz to 8 KHz (±3 dB). 



Single, continuous observation periods with killer 

 whales were termed encounters. During these vessel- 

 based observation periods, the location of the killer 

 whales was plotted at approximately 5-min intervals 

 by using a global positioning system (GPS) linked to a 

 computer with Nobletec^"' (Nobletech, Beaverton, OR) 

 navigational software. Time spent in different behav- 

 ioral states (e.g., feeding, socializing, resting, traveling) 

 was recorded on data sheets. Whales were observed con- 

 tinuously during encounters, and any signs of possible 

 predation were recorded. During behavioral observa- 

 tions, marine mammal kills were confirmed only when 

 marine mammal parts were observed in the mouths of 

 the whales, or when bits of blubber, skin, or viscera, 

 hair were collected, or blood or oil was observed on the 

 surface of the water. Predation on fish was confirmed 

 by observations of fish in the mouths of whales or by 

 collecting and inspecting floating parts. To document 

 the potential marine mammal prey in the region during 

 the period of the study, we recorded the time, location, 

 and number of all marine mammals sighted. 



Analytical methods 



Photo-identification All photographic negatives were 

 examined over a light table with an 8.0 power Peak^^ (B 



and H Photo, New York, NY) magnification loop. Identifi- 

 able individuals were recorded and assigned a unique 

 alphanumeric name in order to be tracked throughout 

 the study. Whales that could not be positively re-identi- 

 fied were not assigned a name. From this photographic 

 database, the actual number and identity of individual 

 killer whales and groups of whales present for each 

 encounter were determined. Because some of the pho- 

 tographs were of poor quality, these photographs were 

 rejected from further analyses; thus not all the whales 

 encountered were identifiable. 



Acoustics We inspected acoustic recordings for the 

 presence of discrete calls by listening to tapes and 

 monitoring real-time spectrograms using Cool Edit 

 2000™ (Syntrillicum Software Corp., Phoenix, AZ) 

 sound manipulation software. Calls were analyzed fol- 

 lowing the protocol of Ford (1991) and Yurk (2005). 

 Killer whales produce a variety of different types of 

 vocalizations that can be described as clicks, whistles, 

 and calls (Ford, 1989). Calls are the most common type 

 of vocalization, occurring in over 90% of all encounters 

 with vocalizing killer whales. These pulsed vocalizations 

 occur as either signals of a frequently repeated acous- 

 tic pattern or as signals of variable acoustic pattern 

 (Ford, 1991). Discrete calls were chosen for this analysis 

 because they retain their recognizable acoustic structure 

 for many years and likely for many generations (Yurk, 

 2005). Recognized calls were digitized at a 44.1-kHz 

 sampling rate with a 16-bit sample size and further 

 analysed spectrographically using Canary 1.2.4 sound 

 analysis software (Cornell Laboratory of Ornithology, 

 Ithaca, NY). The spectrographic analysis was done by 

 using fast-Fourier transformations (FFT) of time series 

 of the recorded sound pressure waves with sizes of 1024 

 points for each analyzed time series. The FFT identifies 

 the composing sine waves in sound pressure waves of 



