UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Echolocation in wild killer whales (Orcinus orca) Barrett-Lennard, Lance 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_barrett-lennard_lance.pdf [ 2.82MB ]
JSON: 831-1.0086068.json
JSON-LD: 831-1.0086068-ld.json
RDF/XML (Pretty): 831-1.0086068-rdf.xml
RDF/JSON: 831-1.0086068-rdf.json
Turtle: 831-1.0086068-turtle.txt
N-Triples: 831-1.0086068-rdf-ntriples.txt
Original Record: 831-1.0086068-source.json
Full Text

Full Text

ECHOLOCATION IN WILD KILLER WHALES (ORCINUS ORCA).byLANCE BARRETT-LENNARDB .Sc., University of Guelph, 1980A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Zoology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADecember 1992© Lance Barrett-Lennard, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study . I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives . It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission .(Signature)Department of 	 ZoologyThe University of British ColumbiaVancouver, CanadaDate	Dec.DE-6 (2/88)ABSTRACTEcholocation by odontocete whales has been demonstrated in captive settings many times,yet little is known about its use and function in the wild . In this thesis, I describe echolocationsounds in pods of killer whales (Orcinus orca) off the coasts of British Columbia and Alaska.I examine the relationships between echolocation and water clarity, ambient noise, andbehavioural activity. I compare echolocation by two populations with different feeding habitsand foraging patterns . The study provides new insight into the functional role of odontoceteecholocation, and into the relative contributions of vision and passive listening to spatialperception and prey detection.Underwater recordings were made during 111 encounters with killer whales . Knownmembers of the resident (fish-eating) population were photographically identified 85 times, andtransients (mammal eaters) 23 times . Click sounds characteristic of echolocation were identifiedin the recordings aurally, and by spectrographic and waveform analysis . Most clicks occurredin series (trains), but isolated clicks were occasionally detected.Echolocation behaviour differed strikingly between residents and transients. Anecholocation index (EI) was defined as the average percentage of time that an individualproduced click trains . The mean EI for residents was 4 .24%, 27 times greater than fortransients. The duration of resident click trains averaged 6 .83 s, compared to 0 .86 s fortransients. Resident click trains were comprised of evenly spaced clicks, whereas transient trainshad uneven click spacing . Transient individuals used isolated clicks once every five minutes,111four times as often as residents . For resident killer whales, EI values were significantly higherduring foraging and travelling than during other behavioural activities . In residents, EI declinedwith group size . This was also true of transients, for the recordings in which click trains weredetected. Transient EI levels were significantly higher when close to shore than when offshore.No relationship between EI and ambient noise level was found for either residents ortransients. Residents increased the amplitude of their clicks in response to increasing ambientnoise levels . No relationship was found between water clarity and EI for either type of killerwhale .I suggest that the differences in echolocation behaviour between residents and transientsare accounted for by their different prey . Fish have little or no aural sensitivity in the frequencyrange of killer whale clicks . Marine mammals are able to detect clicks, and may use them toevade killer whale attacks . The use of isolated clicks and short, irregular, quiet click trainsmakes transient echolocation less detectable by marine mammals than the echolocation used byresidents . Passive listening is probably the principal technique that transients use to locate prey,whereas residents use echolocation in combination with passive listening when foraging. Visionis not a major factor in locating prey, but may be used by either whale type during pursuits.Finally, I suggest that both residents and transients obtain much of their positional andorientational information using passive listening alone .ivTABLE OF CONTENTSPageABSTRACTTABLE OF CONTENTSLIST OF TABLESLIST OF FIGURESACKNOWLEDGEMENTSCHAPTER ONE : General IntroductionCHAPTER TWO : Transient killer whales seldom click: intraspecificdivergence in patterns of echolocation useIntroductionStudy SpeciesMethodsStudy LocationsSampling MethodologyBehaviour CategoriesAcoustic AnalysisPhotographic analysisResults	16Killer Whale EncountersDetectability of Echolocation SoundsCharacteristics of Echolocation PulsesClick Train Characteristicsiiivviviiix155788810111516161718vEcholocation Index ValuesContexts of Echolocation Use1919Behavioural Activity 19Group Size 22Water Clarity 22Ambient Noise 23Discussion 24Resident and Transient Echolocation Levels 24Patterns of Echolocation Clicks 25Behavioural Contexts of Echolocation Use 27Group Size and Echolocation Use 28Echolocation Use and Water Clarity 29Echolocation Use and Ambient Noise 31Conclusions 33Echolocation and Prey Detection Strategies 33CHAPTER THREE : General Conclusions 53REFERENCES 60APPENDIX 69viLIST OF TABLESPageTable I. Study locations, dates and areas surveyed . 35Table II . Characteristics of resident and transient click trains . 36Table III. Echolocation by resident and transient killer whales . 37viiLIST OF FIGURESPageFigure 1 . Map of the northwest coast of North America showing the study areas .	38Figure 2 . Waveform of a killer whale echolocation click .	39Figure 3. Energy distribution of two echolocation clicks .	40Figure 4. Waveform and spectrogram of four seconds of a resident killer whaleclick train .	41Figure 5. Waveform and spectrogram of four seconds of a transient killer whaleclick train.	42Figure 6. Number of analyzed five minute recordings of resident and transientkiller whales, by behavioural activity .	43Figure 7. Mean echolocation index values for resident and transient killer whales,by behaviour category .	44Figure 8. Cumulative frequency distribution of echolocation index levels forresident and transient killer whales .	45Figure 9 . Echolocation index level as a function of group size for residentkiller whales .	46Figure 10. Echolocation level as a function of group size for transient killer whales . 47Figure 11 . Photocell resistance as a function of depth .	48Figure 12. Echolocation index as a function of Secchi depth for resident killerwhales .	49vii'Figure 13. Echolocation index as a function of Secchi depth for transient killerwhales . 50Figure 14. Echolocation index plotted against environmental and vessel noise levelsfor resident killer whales . 51Figure 15. Echolocation index plotted against environmental and vessel noise levelsfor transient killer whales . 52Figure Al . Cumulative frequency histograms of echolocation level by resident killerwhales engaged in four behavioural activities . 70Figure A2 . Cumulative frequency histograms of echolocation level by transient killerwhales engaged in four behavioural activities . 71Figure A3 . Plots of resident and transient killer whale echolocation level againstSecchi depth, by behaviour category . 72Figure A4 . Plots of resident killer whale echolocation level against environmentaland vessel noise levels, by behaviour category . 73Figure A5 . Plots of transient killer whale echolocation level against environmentaland vessel noise levels, by behaviour category . 74ixACKNOWLEDGEMENTSA remarkable number and variety of people contributed ideas, inspiration, equipment, anddirect assistance to this project . I owe a debt of gratitude to each one of them.The late Michael Bigg, friend and mentor, helped plan this project, and blessed it withhis infectious enthusiasm and perpetual good humour . I believe that he would have been tickledat the further unfolding of the resident/transient killer whale story, and I hope I have done hismemory justice . I was privileged to have had two supervisors . One was John Ford, who gotme started in killer whale acoustics, and who encouraged me to return to graduate school aftera long hiatus in the free world . John suggested this project, and shared my excitement as theresults emerged. Jamie Smith co-supervised . Jamie was a wellspring of enthusiasm, advice, andsupport . This thesis was improved immeasurably by the application of his legendary editingskill . I owe him many red pencils and many thanks.Graeme Ellis deserves special mention for doing the photo-identifications, and forknowledge and advice freely offered . Craig Matkin of the North Gulf Oceanic Society madeit possible to carry out research in Alaska . Eva Saulitus, Molly Freeman, and Olga vonZiegesar, my co-researchers in Prince William Sound, were constant sources of support andideas, and the best of friends. The stalwart crew of the "Lucky Star" supplied engine parts andfresh vegetables, and countered the angst of biologists with the practical wisdom of fishermen.In British Columbia I was assisted by the Haida Watchmen, and by Ron Hamilton and DougBuries of the Canadian Parks service . Anna Reid, Julie Kimmel, Leah Saville, Jamie Smith, andmy father Godfrey Barrett-Lennard endured heavy seas, bad food, cramped accommodation, andcranky leadership while serving as crew . Alex Morton and Robin Baird shared details of theirobservations of transient killer whale foraging behaviour. David Bain, Marilyn Dalheim, PaoloDomenici, Christophe Guinet, Gillian Muir, Oskar Painter, Dick Repasky, Tiu Simila, AndrewTrites, David Ward and David Westcott offered ideas and advice . Robin Liley served on mycommittee and supplied feedback and comments on earlier versions of this thesis . Don Ludwigoffered many helpful suggestions throughout, and provided support for a summer assistant todevelop acoustic analysis software . Pete Matthews gave many hours of his time helping mecome to grips with the physics of underwater acoustics.I owe the most to Kathy Heise, who was involved in the project from beginning to end,who was indefatigable in the field, and whose immaculate field notes helped patterns emergefrom chaos. Kathy reminded me on the most difficult days to open my eyes : after all, we werealive and well and working in paradise.I received financial support for research expenses from the Vancouver Aquarium, andfrom an NSERC operating grant awarded to Jamie Smith . I received personal support from aUBC teaching assistantship, and from a BC Science Council G .R.E.A.T. scholarship . Equipmentwas loaned by the North Gulf Oceanic Society, the Vancouver Aquarium, the CanadianDepartment of Fisheries and Oceans, and the University of British Columbia. I thank theCanadian Parks Service for permitting me to conduct research in Gwaii Haanas/South Moresby,and the US National Marine Fisheries Service for allowing me to combine my research with anongoing population study of Prince William Sound killer whales .1CHAPTER ONEGeneral IntroductionSound is well-suited for use as a carrier of information . It takes little energy to produce,it propagates efficiently through air and water, and its bandwidth properties allow it to carrymany signals simultaneously (Norris and Evans 1988) . Numerous animal species have theability to both detect and create sounds . Thus, it is not surprising that many species also havethe ability to actively gather information using the technique of echolocation . Species for whichecholocation has been reported include rodents, insectivores, birds, bats, humans and whales(Fenton 1980 ; Norris & Evans 1988) . Perhaps what is most surprising is that only the bats andtoothed whales are known to have highly-evolved echolocation capabilities.The advantage of echolocation seems evident: it is a technique that can be used toacquire spatial information when visual, olfactory and passive acoustic cues are not sufficientfor a particular activity . The most obvious liability of echolocation is that the user risksrevealing its location and identity to competitors, predators, and prey . Owls and pinnipeds suffersimilar constraints to vision as bats and toothed whales, and have obvious hearing and vocal pre-adaptations for echolocation, yet they use echolocation rarely, if at all (Konishi 1973 ; Watkins& Warztok 1985). Thus, conditions that restrict vision do not necessarily result in net selectionfor echolocation.Researchers of bat sonar have long been aware of both the costs and the benefits ofecholocation (eg Fenton 1980) . Echolocation facilitates the pursuit of insects and the avoidance2of obstacles (Fenton 1980) . However, some species of prey take effective evasive measuresupon exposure to bat echolocation sounds (Roeder 1975), competitors can eavesdrop onecholocators and share prey (Barclay 1982), and predators may track bats by homing in on theirecholocation calls (Fenton 1980) . Many bats are known to hunt part or most of the time usingpassive listening rather than echolocation . In these species the costs of echolocation exceed thebenefits (Andersen & Racey 1991 ; Faure & Barclay 1992) . In general, echolocation is a mixedblessing sensu Fenton (1980), and a cost-and-benefit approach is important to understanding andpredicting its use.The role of echolocation in the behavioural ecology of bats is much betterunderstood than for toothed whales (odontocetes) . Echolocation by bottlenose dolphins (Tursiopstruncatus) was discovered in the 1950's (McBride 1956 ; Kellogg 1958) . Since then, muchresearch effort has been devoted to describing the sonar capabilities of various odontocetespecies in captivity (eg Diercks et al. 1971 ; Au 1980; Moore & Pawloski 1990; Au 1990;Kamminga & Beitsma 1990, Turl 1990; Turl et al . 1991) . By contrast, the published literatureon echolocation in wild cetaceans is limited to a few papers describing the acoustic analysis ofsingle encounters (Ford & Fisher 1978 ; Steiner et al . 1979; Goodson et al . 1988), or listingincidental observations (Watkins 1980).Field studies have lagged behind research on captive animals for several reasons . First,field studies of echolocation lack the precision of behavioural observation possible in a captivesetting. For example, it is seldom possible to link specific individuals in a group with specificsequences of echolocation pulses. Second, acoustic characterisations of sounds recorded in the3field are inexact. For example, the frequency structure of echolocation sounds varies dependingon an animals's orientation with respect to the hydrophone (Au et al . 1978) . Since orientationcannot be controlled or determined in the field, only general descriptions can be made of thefrequency characteristics of echolocation sounds (Watkins 1980) . Third, overlapping sequencesof clicks from large groups may be difficult to differentiate, and sounds from nearby animalsmay mask quieter sounds from those of more distant individuals (Watkins 1980).These problems, however, do not rule out studies of wild populations aimed at describingthe behavioural, social and environmental contexts of echolocation use (Evans & Awbrey 1988).Identification of specific echolocators and precision in the frequency component of acousticanalysis are not essential to answering "when" and "why" questions about echolocation use . Theproblem of overlapping sequences may be . Mressed to some extent by focusing on smallergroups. Studying echolocation in a wild population of odontocetes requires a sacrifice inprecision in order to see general patterns.Populations of killer whales (Orcinus orca) along the north-west coast of North Americaare well-suited for echolocation research . Killer whales in this area are accessible, predictablein distribution, and familiar with the presence of small vessels (Bigg et al . 1987 ; Leatherwoodet al . 1990) . They travel in small groups, and pass through areas where ambient noise levelsare low . Both of these characteristics facilitate acoustic analysis . Their non-sonar vocalbehaviour has been described (Ford 1984, 1989, 1991), the majority of individuals have beenidentified using natural markings (Bigg et al. 1987; Ellis 1987; Heise et al . 1992), and researchhas been carried out on their social structure (Bigg et al . 1990) . Diercks et al . (1971) confirmed4that the species uses echolocation, and described its echolocation sounds . Most significantly,two non-associating populations of killer whale inhabit this region, one of which eats fish andthe other marine mammals (Ford 1984 ; Bigg et al . 1987; Morton 1990) . This makes it possibleto relate echolocation use to foraging strategy, and to assess the flexibility of echolocationuse .In this thesis I examine echolocation in the two forms of coastal killer whale from BritishColumbia and Alaska . In Chapter Two, I compare echolocation use between the two formsin various behavioural and environmental contexts. I consider the question of whetherecholocation substitutes for vision by examining the relationship between water turbidity andecholocation use . I also ask whether echolocation substitutes for passive listening by examiningwhether ambient sounds that potentially restrict passive listening are associated with increasedecholocation use. Finally, I propose strategies of echolocation use for fish-eating and mammal-eating killer whales . In Chapter Three I discuss the "mixed blessing" of echolocation in lightof these findings .5CHAPTER TWOTransient killer whales seldom click:intra-specific divergence in echolocation use.IntroductionSince the discovery in the late 1950's that some toothed whales use echolocation, thesonar abilities of various species have been investigated extensively. This research has focusedon the following topics : 1) the acoustic structure and directionality of echolocation pulses (Au& Hammer 1978 ; Hol & Kamminga 1979); 2) target resolution ability (Au et al . 1978;Kamminga & Beitsma 1990) ; 3) effects on sonar performance of noise and reverberation (Au1980; Au & Turl 1984; Turl et al . 1991); and 4) the effects of target size and distance on sonarpulse structure, amplitude and repetition rate (Morozov et al . 1972; Turl & Penner 1989 ; Moore& Pawloski 1990) . These questions have all been addressed through experiments on captivedolphins and porpoises . In most cases individual animals have been trained to perform taskssuch as detecting silent targets or discriminating between similar objects while blindfolded, priorto the running of trials.This experimental paradigm has provided many answers to "what" and "how" questionsof sonar capability . Unfortunately progress in this area has outstripped progress on "when" and"why" questions of sonar utility . For example, there is certainly overlap in the function of visionand sonar, yet we know very little about how and when individuals make decisions about whichto use. Several authors have suggested that passive listening is a third means of orienting and6locating prey (Norris 1967 ; Wood & Evans 1980; Evans & Awbrey 1988) . This has not beeninvestigated systematically in captive studies, apart from measuring hearing capabilities(Andersen 1970 ; Thompson & Herman 1975 ; Thomas et al . 1990; Bain 1992).Four recent sets of observations indicate that wild odontocetes may use echolocation lessoften or for different functions than was previously suspected. 1) Equipping fishing nets withacoustically-reflective targets generally does not reduce accidental entanglements (Evans &Awbrey 1988; Lien et al . 1990; Dawson 1991). 2) Recently-captured or untrained dolphins areoften unable to use echolocation effectively for simple tasks such as obstacle avoidance (Wood& Evans 1980) . 3) Recordings of wild killer whales (Orcinus orca) indicate that silent travellingis not uncommon, even under conditions of restricted visibility (J . Ford, pers . comm.). 4)Blindfolded captive bottlenose dolphins can pursue live fish without emitting echolocationsounds (Wood & Evans 1980), presumably using passive listening alone . Since the advantagesof echolocation to individuals living in dark or turbid environments seem obvious, theseobservations imply significant costs to echolocation use.Various costs of echolocation use have been reported for bats (Fenton 1980 ; Barclay1982). These include attracting predators and competitors, and stimulating evasive behaviourin prey. Similar costs may apply to odontocetes . In addition, echolocation in social odontocetesis probably affected by information sharing, and by the use of echolocation sounds incommunication . These questions have not been addressed to date, due to the lack of descriptivedata on echolocation use by wild populations .7This study provides baseline data on echolocation in the wild for one species, the killerwhale . I describe echolocation in free-ranging populations on the north west coast of NorthAmerica, and examine the costs and benefits of echolocation use by comparing populations ofkiller whales with different feeding specializations . I investigate the behavioural correlates ofecholocation use, to test whether echolocation functions principally in food-detection andorientation. I test the hypothesis that information acquired using echolocation is shared betweengroup members. Finally, I assess tradeoffs between echolocation, vision, and passive listeningby comparing echolocation use across a range of water clarities and ambient noise levels.Study SpeciesThe coastal killer whale populations of the northern Gulf of Alaska and British Columbiahave been actively studied for 12 and 20 years, respectively. Almost all individuals have beenidentified and catalogued on the basis of natural markings and fin shape (Bigg et al . 1987; Heiseet al. 1992), and their seasonal movements, diet, social structure, and life history parametershave been described. (Stevens et al. 1989; Bigg et al . 1990; Nichol 1990; Olesiuk et al . 1990).Two sympatric non-associating forms, known as residents and transients, inhabit these areas(Bigg et al . 1987) . Residents live in stable matrilineal groups or pods of 5 to 40 individuals andforage principally (perhaps entirely) on fish (Bigg et al . 1987) . Transients live in less stablegroups of 2 to 10 animals and feed principally or entirely on marine mammals (J . Ford and G.Ellis, pers . comm.). Killer whales produce three distinct type of vocalisations: pure tones orwhistles, pulsed tones, and clicks (Ford 1989) . In odontocetes, whistles and pulsed tones arebelieved to be primarily social signals, and clicks are used in echolocation (Popper 1980) . Ford8(1984, 1991) reported pod-specific repertoires of discrete pulsed calls for residents, whereas alltransient groups are believed to share a single repertoire (Ford 1984).MethodsStudy LocationsThe study sites and dates are listed in Table I, and the locations are shown in Figure 1.All three sites were sheltered from heavy seas and had relatively low levels of vessel traffic,making it possible to record under low noise conditions . In Prince William Sound and thecentral coast of British Columbia, the water varied widely in clarity between different locations,due to glacial meltwater and surface run-off. The killer whale populations of British Columbiaand Prince William Sound do not overlap, hence two discrete resident and two discrete transientpopulations were examined.Sampling MethodologyIn Alaska I used an 8 m vessel for field accommodation and for all aspects of theresearch . The vessel was driven by a single-screw stern drive, and had a maximum cruisingspeed of 16 knots . In British Columbia a 13 m vessel was used for accommodation and researchactivities . It was driven by a single screw on a straight shaft, and had a maximum cruisingspeed of 8 knots . A 5 m inflatable boat powered by a 25 hp outboard engine was also used forsearching and observing in British Columbia, when sea conditions permitted . I worked with oneassistant in Alaska, and one or two assistants in British Columbia .9In order to locate killer whales we visually searched areas known to be commonly usedby killer whales . Sighting information received from mariners or aviators by radio was usedwhenever possible to focus searches . During searches we travelled at approximately 8 knots,scanning with binoculars . If local conditions permitted, we periodically dropped an observer offon shore to scan from a high vantage point with 7X50 binoculars or a 20X spotting scope . Atapproximately 1 hour intervals a hydrophone was deployed to listen for vocalising pods out ofvisual range. If distant whales were detected, we attempted to find them using a directionalhydrophone.When killer whales were located we approached slowly to within about 25 m, and thenparalleled their course while taking identification photographs with a 35 mm camera fitted witha 300 mm lens and shoulder brace . We photographed the left side of each individual, using theprotocol described by Bigg et al . (1986) . When all animals present had been photographed, wemoved 500 m ahead of the animals and shut off the vessel engine . A Bruel and Kjaer 8101hydrophone was lowered on a 25 m cable, and recordings were made on Ampex 407 tape usinga custom-built calibrated pre-amplifier and Nagra IV-SJ reel-to-reel recorder running at 37 cm/s(effective system frequency response 25 Hz - 35 kHz +/- 1 dB) . Recordings were made at hourlyintervals . Five minute recording sessions were used, since the animals generally remained withinrange for detecting echolocation for that period . Simultaneous voice recordings of behaviouralobservations were made on a separate track of the same tape . Backup recordings were madeconcurrently using a Sparton 60CX123 hydrophone on a 20 m cable and a Sony WM-D3cassette recorder.10Immediately after each recording session, water clarity was measured using a 20 cmSechii disk divided into black and white quadrants . The disk was lowered on the lee side of thevessel until it disappeared from view. It was then raised until it re-appeared, and the Secchidepth was taken to be the average of the two . The surface of the water was shaded from directsun during this measurement. To assess whether water clarity varied with depth, an ambientlight profile was made using a Archer 276-116A photocell attached by a 30 m cable to aresistance meter . Readings were taken every meter as the cable was lowered . Clear layers wererevealed when depth increments resulted in little or no change in resistance reading : rapidchanges indicated the presence of turbid layers.Behaviour CategoriesDuring the recording sessions the behavioural activities of the whales were noted . Thesewere grouped into the six categories listed below . The first four are based on categories usedby Ford (1989).Foraging : This included all occasions when the whales were known or suspected to be feedingor searching for prey. Direct evidence was available when whales were seen carrying prey, orif fish scales, tissue, blood, or blubber were seen floating at the surface. Kills were confirmedwhenever possible by collecting remains with a fine mesh net for close inspection . Erratic high-speed swimming, sudden lunges, and swimming in tight circles were also considered to beevidence of foraging by residents . Swimming close to shore, entering into shallow bays, andcircling kelp patches and small islands was evidence of foraging by transients (Ford 1984).Transients swimming in dispersed formation across open areas were also considered to be11foraging (Baird & Stacey 1989).Travelling: A group was considered to be travelling when its members formed a single largegroup or several tightly-knit subgroups and moved on the same course at speeds > 6 km/hr.Resting : Groups were categorized as resting when members formed tightly-knit groups, movedat speeds < 4 km/hr, and surfaced in unison in a co-ordinated manner, without travellingconsistently in any direction.Socializing: Whales grouped into clusters and engaged in physical interactions such as chasing,rolling, and thrashing were considered to be socializing . Breaches (when an animal threw mostor all of its body out of the water), spyhops (vertical surfacing with the head lifted clear of thewater), and the slapping of flippers, dorsal fins and tail flukes on the surface were allcharacteristic of social activities.Milling : Groups were considered to be milling when their members grouped loosely and madelittle or no progress through the water . Unlike resting, breathing during milling was neitherregular nor co-ordinated.Slow-travelling: Groups that moved at about 3-6 km/hr, showed no signs of foraging, engagedintermittently in surface activities such as breaches and tail fluke slapping, and organized intomoderately dispersed subgroups of 3-4 animals, were considered to be slow-travelling.Acoustic AnalysisThe recordings were analyzed using a Kay Elemetrics DSP-5500 sound analyzer, capableof performing acoustic analysis in real time. The tapes were played back at reduced speed (1/8to 1/2 original speed, depending on the complexity of the sounds being analyzed) . I displayedthe analysis results on a divided video screen . Half of the screen displayed a colour-enhanced12512 point spectrographic display with a 64 kHz bandwidth (compensated for tape speed) in a4 second time window. This display was useful for detecting fine-scale variations in frequency.Although the system response declined beyond 35 kHz, useful information extended to 50-55kHz. Killer whale echolocation pulses or clicks have been reported to lie well within this range(Diercks et al . 1971) . The second half of the screen displayed the signal waveform, which wasuseful for detecting signal strength fluctuations over time . Input and dynamic range levels wereadjusted periodically to compensate for varying signal-noise ratios . I also monitored theplayback aurally during all phases of the analysis.Sonar pulses usually occurred in regular series, hereafter referred to as click trains.During the analysis, the beginning and end times of each click train were recorded using adigital stopwatch . The individual pulses in any given click train were typically consistent intiming and frequency structure, and were readily-distinguishable from those in other trains. Thismade it possible to distinguish overlapping trains . If the whales were widely-dispersed ortravelling rapidly during a recording session, it was sometimes difficult to determine whethercertain trains had ended or faded into background noise . If this occurred, the analysis of thatsession was discarded. Five minute recordings of grouped pods moving at normal speeds couldusually be analyzed successfully . Recordings of rapidly-travelling or very widely-dispersed podscould not always be analyzed, particularly under high noise conditions.I used a simple echolocation use index (EI) to quantify sonar activity . This was theaverage percentage of time that an individual whale present during a recording session emittedclick trains, and was calculated as follows :13EI=100 dsnwhere d is the sum of the durations of all the click trains, s is the duration of the recordedsession, and n is the number of whales present. Values of the echolocation index are alsoreferred to as echolocation index levels throughout this paper.Echolocation indices for the pooled resident and transient data were compared by t-test.Differences among EI means across all behaviour categories were square-root transformed toequalize variances and were tested using a one-way ANOVA . In the case of transients, a non-parametric analysis of variance was used (Kruskal-Wallis test, Zar 1984, p 176) since thevariances could not be equalized by transformation . When means differed, Scheffe's multiplecontrast test (Zar 1984, p 196) was used to test whether echolocation was more stronglyassociated with travelling and foraging than with other behavioural activities.Click trains were selected using random number tables for measurements of average pulserepetition rate (R), calculated asR= c-1twhere c is the number of clicks in the train, and t the duration of the train . Repetition rateswere compared between residents and transients using t-tests . The variances in repetition ratewere compared using Bartlett's test (Zar 1984, p 181) . Every train detected was also rated forthe evenness of intervals among its component clicks . Trains with irregular timing were flagged14and counted during analysis . A train was defined as irregular if the interval between any pairof adjacent clicks differed by more than 10% . Thus, trains that graded gradually from onerepetition rate to another were considered to be regular. Most click trains could be readilyclassified by listening to a slowed playback of a recorded session and viewing its spectrographicform. When irregular trains were detected, I paused the playback and compared click intervalsusing time cursors on the spectrographic display . The percentages of irregular trains werecalculated session-by-session. These data were normalized using a modified Freeman and Tukeyarcsine transformation (Zar 1984, p 240) . Mean percentages were then compared using t-tests.Isolated short-duration broadband sounds closely resembling sonar pulses but notoccurring in trains were detected periodically . These are referred to as isolated clicks, and werequantified using the following index (IC):IC=1000 C.nswhere c; is the number of isolated clicks, n the number of animals present and s the duration ofthe recorded session in seconds . Isolated click use was compared between transients andresidents using t-tests on the arcsine-transformed data.Finally, the average background levels of both environmental and vessel noise were ratedon discrete scales of 0-5 . Each unit increment represented a change in average sound pressurelevel of 10 dB. This was determined by reference to the input levels of the recorder and thepre-amplifier at the time of recording, and the analyzer input level settings. The absolute sound15pressure level of the baseline was not determined, however it was consistent between allrecording sessions . Resolution beyond 10 dB was not practical since noise levels fluctuatedconsiderably within a 5 minute recording session. Environmental noise was caused by wind,rain, waves, surf, and the fizzing of melting glacial ice ; it was generally broadband and relativelyconstant. Vessel noise was produced by boat engines and propulsion systems, and was alsobroadband although lower in peak energy than environmental noise . Vessel noise varied froma constant roar to an intermittent throbbing . When both environmental and vessel noise werepresent, the relative contribution of each to total ambient noise was estimated . Associationsbetween these noise types and echolocation index levels were sought using simple and multiplelinear regression techniques on the raw and log-transformed data pooled across all behaviourcategories, for each behaviour separately.In many instances more than one recording session was analyzed from the sameencounter. To ensure that recordings were statistically independent, I selected sessions foranalysis which were separated by a distinct break in behaviour . Thus, if an encounter consistedof 3 h of foraging followed by 2 h of resting and then 3 h of foraging, I analyzed at most onesession from each of the three activity periods . On occasion I analyzed concurrent transientforaging sessions that were separated by a kill and feeding session or by a transition betweenforaging modes (for example a move from an inshore area to an offshore area).Photographic AnalysisThe whales present during each encounter were photographed to identify them as residentor transient. Ilford HP-5 black and white film was exposed at 1600 ASA using a shutter speed16of 1000/sec or less and developed as described by Miles (1990). The negatives were examinedunder low-power magnification with a dissecting microscope, and the animals identified usingcatalogues by Bigg et al . (1987), Ellis (1987), and Heise et al . (1992), and the photographcollections of the Pacific Biological Station, Nanaimo, British Columbia, and the NationalMarine Mammal Laboratory, National Marine Fisheries Service, Seattle, Washington. Allanimals photographed were positively identified and were known previously as residents ortransients, except for three encounters with "new" groups . Recordings of these new animalswere excluded from the data set .ResultsKiller Whale EncountersKiller whales were encountered 111 times during this study . Residents were present 85times, transients 23 times, and uncategorized groups the remaining 3 times . Both residents andtransients were observed in a range of behavioural activities and under varying ambient noiseand water clarity levels . An average of 5 h was spent photographing, observing and recordingthe animals during each encounter. Successful acoustic analysis was performed on 162 recordedsessions of approximately 5 min duration each, comprising 74 sessions with residents and 88with transients.Detectability of Echolocation SoundsEcholocation clicks were often detected from approaching killer whales up to or beyond3 km; under most conditions complete trains could be distinguished when the animals were17within 1 km. When a pod passed the boat, clicks dropped in volume quickly, and detectionbecame unreliable within 500 m . This directional effect was most evident in the high frequencypart of the sound spectrum, as has been found in other odontocetes (Diercks et al . 1973; Au1980). Transients produced quieter clicks than residents, but this was offset by their smallergroup sizes, making it possible to be closer to all animals . Choppy sea conditions sometimescreated sufficient slapping on the vessel hull to mask quiet clicks ; however, more than 75% ofrecorded sessions were analyzed successfully.Characteristics of Echolocation PulsesSonar pulses occurring within trains were broad-band, short duration signals characterizedby rapid rise times and exponential decay (Fig. 2). Generally four to six peaks were resolvablebefore the signal became buried in noise, corresponding to an interval of approximately 0 .5 to15 msec. Peak frequencies ranged between 4 and 18 kHz (Fig . 3), however some clicks hadmeasurable energy extending to the upper limit of the recording system (c . 50 kHz). Similarfrequency peaks and ranges were reported by Awbrey et al. (1982) for wild killer whales . Asthe hearing of killer whales extends to 105 kHz (Bain 1992), I paid careful attention to the upperfrequency portion of the spectrographic display for evidence of the lowest components of veryhigh frequency clicks. I found no indication of such clicks in the 162 sessions analyzed.Isolated clicks were often very difficult to distinguish against background noise . Indeed,early in my analysis these sounds were missed entirely . I eventually realized that single ordouble pulses closely resembling those found in trains were recorded in the presence of killerwhales, and were not detected in recordings made in the same areas when whales were absent18(12 sessions, 60 min total duration) . Isolated clicks bore similar characteristics to clicksoccurring within trains, both in duration and frequency structure, and when paired had inter-clickintervals of 100 msec or less.Click Train CharacteristicsThe characteristics of resident and transient click trains are summarized in Table II.Residents produced click trains that were more than eight times as long, on average, than thoseproduced by transients . The percentage of irregular click trains was approximately 18 timeshigher for transients than residents . During spectrographic analysis I noted that resident clicktrains were more consistent in frequency structure and distinctly louder relative to backgroundnoise than those produced by transients . However, I did not systematically quantify thesefactors . Figures 4 and 5 are examples of resident and transient trains . Note the irregular timingand frequency structure of the latter.Resident click repetition rates were about twice those of transients, but this differencewas not statistically significant (Table II) . However, the variance in repetition rate wassignificantly greater for residents than transients (Bartlett's test, p<.001) . In the bottlenosedolphin the inter-click interval (the inverse of repetition rate) varies with target distance suchthat the interval is greater than the time required for a sound to travel to a target and back(Morozov et al . 1972; Au et al. 1974; Au et al . 1982; Roitblat et al . 1990). The beluga(Delphinapterus leucas) differs in that clicks may be produced in consistently-structured"packets" of several clicks, with some inter-click intervals less than the two-way travel time tothe target (Turl & Penner 1989) . Click packets were not detected in this study and have not19been reported by other authors . Based on the assumption that inter-click intervals exceed thetwo-way travel time to a target, the mean maximum target distances for residents and transientswere 55 m (95% confidence interval : 34-133 m) and 109 m (95% confidence interval : 89-143m) respectively.Echolocation Index ValuesResidents and transients also differed strikingly in their echolocation index values, assummarized in Table III . On average, echolocation trains were produced 4 .24% of the time forevery resident individual in a group, compared to 0 .16% for every transient, a 27-fold difference.In contrast, isolated clicks were used four times more frequently by transients than residents.The indices in Table III are equivalent to intervals between isolated clicks of 21 .4 minutes forresidents and 5 .0 minutes for transients.Contexts of Echolocation UseBehavioural ActivityFigure 6 shows the number of sessions successfully analyzed within each behaviourcategory. This is only an approximate behaviour budget, since some categories of behaviourpresented greater difficulties in acoustic analysis than others and were discarded more frequently.Transients spent a greater percentage of their time foraging than residents (as reported by Ford1984, and Morton 1990), and were never observed socializing. Residents travelled infrequently(as reported by Ford 1989) and rested more often than transients.Strong differences in echolocation index levels between transients and residents were20consistent across all behaviour categories (Fig . 7). For residents, analysis of variance indicatedsignificant differences in EI between behaviours (p= .012). Foraging and travelling combinedhad significantly higher EI's than the remaining four behaviour categories combined (Scheffe'stest, p=.004). Norris (1967) reported that different types of echolocation were used in foragingand orientation by bottlenose dolphins . To test whether this was true of resident killer whales,mean EI levels of foraging and travelling groups were compared . No significant difference wasfound (Scheffe's test, p= .574).For transients, a nonparametric analysis of variance was used, since the variances differedmarkedly between categories despite the transformation . No differences in EI between behaviourcategories were found (Kruskal-Wallis test, p= .924) . However, the pattern of relative levels ofecholocation between the various behaviour categories is similar to that seen in residents.No recordings of socializing transients were obtained during the course of this study,reflecting the rarity of this behaviour (Morton 1990) . Four recordings of socializing transientswere provided by E . Saulitus from four years of research on transient killer whales in PrinceWilliam Sound. These records are of interest in that they show EI values of 1 .11 to 7 .08, witha mean of 3.36, i .e. approximately 20 times higher than the mean values I observed for non-socializing transients . These data are not included in the analyses in this paper.Residents and transients differed in their EI frequency distributions, in addition to meanEI's (Fig. 8). The transient frequency distribution is heavily skewed towards zero, with noecholocation at all during the highest proportion of sessions . The resident frequency distribution21is much less skewed. This may indicate that for transients a cost to echolocation is experiencedeven at very low levels, whereas with residents constraints to echolocation increase moregradually. The general pattern of echolocation distribution is consistent across behaviourcategories (Appendix, Figs . Al and A2).Transients were observed foraging both inshore and offshore. Harbour seal kills wereconfirmed 15 times during inshore foraging and seven probable kills were seen but notconfirmed. During offshore foraging we confirmed three kills of Dall's porpoise and saw threeprobable kills . Echolocation index levels were significantly higher when transients were nearshore than when offshore (inshore: n=57, mean EI= .230, sd=.498 ; offshore: n=26, mean EI=.025, sd= .075; t-test: p<.041) . Transients were observed using two techniques for catching seals.The first was to travel long distances underwater to attack seals that were swimming near theshore. The second was used after seals were already alerted by a kill or attack. Seals near shoreusually hauled out immediately in response to the presence of killer whales, but seals somedistance from shore took refuge in kelp patches, on islets or shoals, or in underwater crevicesor caves. In this case, the transients either moved away a short distance and returned, or milledin the area. Seals were caught when they left underwater refuges to surface for air or when arising tide made them accessable to the whales . During these waiting periods, quiet echolocationclicks were relatively frequent.During offshore foraging, transients swam more than 500 m underwater before attackingDall's porpoises in three observed incidents . In the remaining three cases rapid surface chasesbegan 25 to 50 m from the porpoises, which fled along the surface and frequently changed22direction. Although the porpoise groups contained one to three individuals, no more than onekill per attack was confirmed. I detected no echolocation sounds prior to or during a chase.Group SizeEcholocation index levels by resident killer whales declined significantly with increasinggroup size (Fig. 9; for EI vs log group size, r 2 = .142, p=.001) . To test whether this result simplyreflected sampling error due to large groups being spread out over longer ranges than smallgroups, I repeated the analysis using only recordings made when the animals were in closeproximity to each other . This did not affect the result.A similar but much weaker trend was observed for transients (Fig. 10; r2=0.023,p=0.157) . However, when I removed all sessions with no echolocation at all, the negativeassociation between EI and group size became stronger (Fig . 10; r2=0.143, p=0 .052). Thissuggests that group size does not strongly affect whether or not a transient group usesecholocation, but if it is used, echolocation per individual declines with increasing group size.Groups of three transient individuals were most common (49 of 88 sessions analysed) . Threetransient groups of this size were recorded. One of these accounted for 28 sessions . Thus, itis possible that the observed trend is due to differences between pods rather than an effect ofgroup size.Water ClarityEcholocation index levels appear to be independent of water clarity across the range ofclarities encountered . Secchi depth was taken to be a reliable indicator of water clarity, since23(Fig. 11). No significant associations were found between EI level and Secchi depth for eitherresidents or transients (residents : r2=0.000, Fig. 12; transients : r2=0.007, Fig. 13). Associationswere also sought separately for each behaviour category, without success (Fig . A3). Lines werefitted on the latter plots using a Lowess smoothing technique (Wilkinson 1990) to help identifynon-linear relationships : none were apparent.Ambient NoiseEcholocation activity levels for both residents and transients were independent of bothenvironmental and vessel-produced noise levels. Simple linear regressions were attempted usingthe echolocation data pooled from all behaviour categories on environmental and vessel noiseseparately, and multiple linear regressions were attempted on both noise types . None of thesemodels were statistically significant . Pooled EI's plotted against both noise types are presentedin Figures 14 and 15; the surfaces, which are fitted using distance-weighted least squaresmethods (Wilkinson 1990), show the absence of clear trends.I also searched for relationships between the two types of noise and EI for eachbehaviour category separately (Figs . A4 and A5) . Lowess smoothing was used to explore thepossibility of non-linear trends, as described above ; again, no clear relationships emerged . Itwas noted consistently that resident clicks increased in intensity during periods of high ambientnoise, whereas transient clicks did not do so in an obvious way . Thus it is possible that sometransient echolocation, difficult to discern at the best of times, was not detected during noisyperiods .24DiscussionResident and Transient Echolocation LevelsThe most striking finding of this study is that resident and transient killer whales differedmore than 25-fold in average echolocation index levels . I suggest that this is a consequence ofthe different responses of fish and marine mammals to echolocation sounds . Peak energies ofkiller whale echolocation clicks in this study ranged between 4 and 18 kHz ; Diercks et al.(1971) reported peak energies of 25 kHz for a captive killer whale . Audiograms have not beenperformed on Pacific salmon (Onchorhynchus spp.), the most common prey of resident killerwhales (Bigg et al . 1987; Nichol 1990) . However, most bony fish that have been tested,including Atlantic salmon (Salmo salar), have very low auditory sensitivity above 3 kHz(Hawkins & Johnstone 1978 ; Hawkins 1986; papers in Tavolga 1976) . Thus these species mayperceive little or no energy from killer whale clicks . Marten et al. (1988) tested the reactionsof 13 species of bony fish to high intensity bottlenose dolphin sonar and detected no behaviouralreactions . Pinnipeds and cetaceans, the prey of transients killer whales, have acute hearing tofrequencies beyond 30 kHz (Mohl 1968 ; Andersen 1970 ; Mohl & Andersen 1973 ; Moore &Schusterman 1987 ; Terhune 1988).Pinnipeds that detect approaching killer whales may evade capture by leaving the water,moving into shallow areas, or hiding in kelp patches or amongst rocks on the bottom . Duringmy study harbour seals were seen using these techniques successfully . Frost et al . (1992)reported similar behaviours by walruses . Porpoises were seen swimming away from transientsat high speed; other authors have reported dolphins and grey whales moving into shallow water25when killer whales were nearby (Baldridge 1972 ; Wursig 1989). Thus the risk of alerting preymay be a significant cost of echolocation for transients . The finding that transients echolocatedmore when nearshore than when offshore may have resulted from the fact that during many ofthe nearshore recording sessions, seals in the area were already alerted to the whales . The costof echolocation was presumably reduced at this point, since the element of surprise was lost.Are there alternative explanations for the difference in echolocation use between residentsand transients? Residents and transients differ strikingly in behaviour : residents have large groupsizes, do not follow shorelines closely, make short dives, and spend approximately 50% of thetime foraging; transients live in small groups, often travel in shallow water following shorelinesclosely, have relatively long dive times and spend approximately 77% of the time foraging(Morton 1990 ; Heimlich-Boran 1988). Transients may require more precise orientation skillsthan residents, since near-shore travelling has the attendant risks of accidental stranding andcollision with the bottom . In addition, longer dive times provide reduced opportunities toconfirm position by visual reference to above-surface features . Finally, if positional informationis shared among group members as proposed by Norris (1980), then transients should requiremore information per individual than residents simply by virtue of group size. All thesedifferences in travel patterns, group sizes and dive-times suggest that frequent echolocationwould be more beneficial to transients than residents, the opposite of the observed pattern.Patterns of Echolocation ClicksI found in this study that transient killer whales produced short and irregular click trainscompared to the trains of resident killer whales. I also noted incidentally that transient clicks26were more variable in frequency structure and quieter than resident clicks . These observationsare consistent with the hypothesis that stealth is more important to the hunting success oftransients than residents . Sequences of short duration sounds that are irregular in timing andfrequency structure resemble random noise more closely than regular sequences . Similarly, thelonger and louder a sequence, the greater the opportunity for recognition . Thus, potential preyof killer whales are more likely to detect resident echolocation trains against background noisethan they are transient trains.The use of isolated clicks may also yield echolocation information without alertingpotential prey . My difficulty in recognizing isolated clicks while listening to slowed playbacksmay be indicative of the problems faced by an acoustically-vigilant porpoise or pinniped . Isuggest that isolated clicks are cryptic echolocation "snapshots" . Isolated clicks could onlyprovide crude information compared to that possible with click trains (Norris 1967) . However,this disadvantage may be offset by the stronger echoes returned by pinnipeds and smallcetaceans relative to fish.The pulse repetition rate was significantly more variable in residents than in transients.This suggests that residents use echolocation to investigate wider distance ranges than dotransients. For example, residents may use it for long-range navigation, detecting distant schoolsof fish, identifying and selecting specific fish to hunt, and tracking individual fish during pursuit.Transients may use it primarily to detect prey that are close enough to capture even if alerted.In the latter case, there should be a close match between click intensity and the range underexamination, with the sound being sufficiently intense to return detectable echoes, but not so27intense as to warn prey at greater distances. This agrees with my qualitative observation thattransient clicks were consistently quieter than those of residents.Behavioural Contexts of Echolocation Use.The patterns of echolocation use by resident killer whales resembled the patterns of calluse reported by Ford (1989), except that the high call rates of socializing groups were notmatched by high echolocation rates . In addition, foraging groups in this study used echolocationconsistently, while Ford (1989) reported periods with no calls ranging from 1 min to 1 hr duringforaging. Evans and Awbrey (1988) warned that : "Assignment of sounds to mutually-exclusivecommunication or echolocation categories may or may not match their uses by dolphins".However, the more consistent use of echolocation than calls during foraging implies thatecholocation plays a greater functional role in that activity. Ford (1989) suggested that callswere used as social signals between scattered pod members, since the first call in a seriesseemed to trigger a chorus of responses. Click trains did not appear to elicit responses, and therewas no evidence that they had a similar function to calls. Clicks were louder and pulserepetition rates more variable during foraging than during other activities, also suggesting anactive role of sonar in foraging. Average echolocation index levels were high during travelling,but silent travelling was also observed on occasion. This suggests that residents may experiencesome cost to echolocation . It may be advantageous to move to new foraging areas withoutrevealing this information to other resident pods, as has been described in bats (Barclay 1982).Silent travelling by residents was twice preceded by the arrival of transient killer whales within1 km. In both instances the resident pods abruptly terminated their previous activities, ceasedall vocalizations and grouped up tightly before beginning to travel . This uncharacteristic28behaviour appeared to be an avoidance response, and is evidence that resident killer whaleecholocation may be suppressed in certain social contexts.Transient killer whale echolocation use differed less among behaviour categories than inresidents, but a similar trend was seen . The infrequent use of echolocation in all behaviourcategories, (except, possibly, socializing), might be explained in two ways. Firstly, transientsmay be alert for opportunistic encounters with prey even when not actively searching . Secondly,prey behaviour patterns may be affected by the detection of transients so that foraging successis reduced in the future. These ideas are both supported by my observations . Twice transientgroups engaged in travelling and slow-travelling (not classified as foraging because the whaleswere swimming in close formation) suddenly pursued porpoises . Frequently killer whalespassing close to harbour seals caused them to haul out or move rapidly into shallow water orkelp patches, although no pursuits or aggressive interactions took place . Increased vigilance bythese seals might reduce their risk of being attacked for some time, thus, transients could paya future cost for present detection.Group Size and Echolocation UseThe negative correlation between echolocation use and group size for resident killerwhales suggests information sharing. There would seem to be little advantage to each individualseparately acquiring information about prey or surroundings, and possible interference costs.Information sharing may occur in three ways . First, socially dominant individuals may useecholocation to make decisions and then lead other group members in movements and activities.Second, by monitoring the echolocation activity of distant group members, individuals may be29able to infer the location of foraging "hotspots" . Wilkinson (1992) reported eavesdropping ofthis type in bats . Third, killer whales may be able to interpret the echoes of pulses producedby other group members directly, or "share" clicks. Scronce and Johnson (1976) showed thata blindfolded captive bottlenose dolphin could detect a target using artificially-produced pulsesfrom a sound projector . Click sharing may be more useful to a resident group that is swimmingin formation and scanning ahead with echolocation than one engaged in foraging, since foraginganimals are usually dispersed and often randomly oriented (Ford 1989) . In addition, residentkiller whales foraging on schooling fish such as salmon are likely to be presented with a"clutter" of small acoustic targets, from which they select one or more to pursue . This activityrequires greater intensity and frequency of clicks than coarser-scaled discriminations (Au & Turl1983), and therefore may require greater individual control of click production.Unlike residents, transients usually forage in a co-operative manner, (Heimlich-Boran1988; Baird et al. 1992). This almost certainly means that information exchange occurs.However, the intermittent and low-intensity nature of transient echolocation makes it unlikelythat click-sharing is a principal mechanism for this exchange . In this study the effects of groupsize were weak if all recorded sessions were considered, but much stronger if the sessions withno echolocation were dropped . A possible explanation is that in transient groups only oneindividual echolocates, and that individual does not base its decisions on group size.Echolocation Use and Water Clarity.I found no significant association between echolocation index and water clarity for eitherresidents or transients . This is an important observation since it implies that caution must be30used in extrapolating the results of echolocation experiments in captivity . Echolocation does notalways take over where vision leaves off . Wood and Evans (1980) described a study designedto examine echolocation use by a dolphin during the capture of live fish . Contrary toexpectations, the dolphin repeatedly located and approached live fish without producing anydetectable echolocation sounds . The study demonstrated that delphinid cetaceans can detect,track and catch fish using passive acoustic cues alone . This is highly relevant to understandingthe foraging behaviour of free-ranging populations . Unfortunately, this type of experiment hasnot been repeated on other odontocete species . However, sound production by many species offish is well documented (Myrberg 1981 ; Boyes 1982; Hawkins 1986), and it seems likely thatmany species of cetaceans make use of this source of information.Evans and Awbrey (1988) described wild beluga whales feeding actively on salmon insilt-laden water without producing detectable sounds except for an occasional buzz . Either theanimals were producing pulses with virtually no energy below 60 kHz, the limit of theequipment used (which is unlikely, see Kamminga & Wiersma 1981), or they were relying onpassive acoustic cues . The same authors reported silent feeding under low-visibility conditionsby bottlenose dolphins.Norris et al . (1961) reported the opposite lack of correspondence between water clarityand echolocation. They found that a blindfolded bottlenose dolphin retrieved dead fish usingecholocation, but also used echolocation when performing the same task in clear water withouta blindfold. Evans and Awbrey (1988) also reported that wild bottlenose dolphins eating deadflying fish used intense echolocation despite clear water conditions . They speculated that the31echo characteristics of a fish may reveal more about its condition than does visual appearance.In my study foraging residents used echolocation under clear conditions . However, becauseinter-click intervals can be used to infer maximum but not actual target distances (Au et al.1982), and because maximum distances exceeded the clarity, it is not known whether clicks wereused to investigate fish within visual range.In this study, transient killer whale attacks on Dali's porpoises were not associated withecholocation . Since in every case the prey were detected by the whales well beyond underwatervisual range, and visual scanning above the surface was not seen, it appears that passive listeningwas the most likely detection mechanism . Dall's porpoise echolocation is centred on frequenciesof 135-149 kHz (Au & Jones 1991) and thus may not be detectable by killer whales, which havelittle sensitivity above 105 kHz (Bain 1992) . However, I was able to detect the underwatersounds of porpoises breaking the surface and breathing within 25 m under quiet ambientconditions, and transients may make use of such sound at much greater distances . Jefferson etal . (1992) cite several reports of cetaceans becoming silent and motionless when approached bykiller whales . This supports the idea that transients use swimming and surfacing sounds to findmarine mammal prey.Echolocation Use and Ambient Noise.Several authors have suggested that passive acoustic cues play important roles in thenavigation, foraging and social behaviour of delphinid cetaceans (Norris 1967 ; Wood & Evans1980; Evans & Awbrey 1988 ; Hoelzel 1991) . Passive listening cannot be easily measured inan observational study of free-ranging populations, however, the probability of its use may be32inferred from the relationships between ambient noise levels and echolocation behaviour.Echolocation sounds are generally louder than those providing information about prey, andpotentially louder than those providing orientational cues, such as distant surf . Thus, as ambientnoise levels increase, passive listening should become ineffective as a sensory mode sooner thanecholocation . The masking effects of broadband noise on echolocation by beluga and bottlenosedolphins has been described by Turl et al . (1987) and Au (1980, 1990) . Both species respondedto increased ambient noise by increasing the intensity and frequency structure of their clicks intarget detection experiments . If this is also true of killer whales, then echolocation is probablypreferred over passive listening under noisy conditions.I found no correspondence between echolocation index and ambient noise level forresidents. This agrees with the above conjecture that for residents echolocation is not costly andis normally used. Transients also showed no response in echolocation index to ambient noiseof either type, which suggests that they do not trade off echolocation and passive listening . Anyattempt by transients to overcome the masking effect of noise by increasing their echolocationindex levels or the loudness of their clicks is likely to result in their detection by theiracoustically-sophisticated prey . Anecdotal reports indicate that transients avoid vessels moreactively than residents (E . Saulitus, pers . comm.), perhaps as a response to interference withpassive listening. Twice during this study transient groups suddenly began pursuit of Dall'sporpoises immediately after my vessel engine shut down, and similar observations have beenmade by G. Ellis and J. Ford (pers . comm.). These observations support the conjecture thatvessel noise affects the ability of transients to find prey using passive listening .33ConclusionEcholocation and Prey Detection StrategiesBased on the foregoing discussion, I propose that resident and transient killer whales useecholocation strategies as outlined below . I do not imply that these strategies are proved, butrather that they fit well with my findings and those of others, and form a useful basis for futureresearch.Resident Killer Whales : Residents detect fish using passive listening, echolocation and vision.Echolocation is normally low in cost . Thus, it is used even when conditions for vision are good,both to probe beyond the visual range, and to obtain additional information such as the size,species and condition of prey . When vision is blocked by turbidity, click rates do not risegreatly, because the echolocation channel is already in use . Passive listening is also used todetect schools or aggregations of fish . Passive listening is an important mechanism inorientation and navigation, thus echolocation index levels are reduced when resident whales arenot searching for prey . Echolocation use may be curtailed altogether if costs to echolocationarise. The risk of attracting competitors may be one such cost.Transient Killer Whales : Echolocation is used sparingly in the form of isolated clicks or short,quiet, and irregular trains to reduce the risk of alerting potential prey (marine mammals) . Sincethe prey of transients are large and sparsely distributed, random underwater visual searching isseldom effective . Transients run a "trap-line" between areas of known local abundance such asharbour seal haulouts, and search other areas using passive listening. As such, their behaviour34is not affected by water clarity, except perhaps during the final phases of a pursuit . Sincetransients depend to a greater extent on passive listening than residents, particularly in openwater, their hunting efficiency may be reduced by high ambient noise levels .35Table I . Study locations, dates and areas surveyedLocation Lat./Long. Area Surveyed(km2) DatesPrince WilliamSound, Alaska60°05'N., 147°30'W . 1600May 25-Aug. 30,1990July 8- Sept . 10/91July 25-Sept . 6/92Central MainlandCoast, BritishColumbia52°0'N., 127°50'W. 900 May 27-June 18/92May 20-June 2/92South-EasternQueen CharlotteIslands52°30'N., 131°0'W. 1200 June 3-July 9/9236Table II. Characteristics of resident and transient click trainsMean ClickTrain Duration+/- SE (s)[numbersessions]Mean Per-centage ofIrregular Trains+/- SE[numbersessions]MeanRepetition Rate+/- SE (/s)[number oftrains]Resident 6.83 +/- 0.54 3.46 +/-1 .08 13.75+/-3.98[74] [70] [32]Transient 0.86 +/- 0.20 60.84 +/- 8 .16 6.83+/-0.77[88] [28] [23]t-test p<.001 p<.001 p=.15137Table III. Echolocation by resident and transient killer whales, based on 74 recorded sessionsfor residents, 88 for transientsMean Echolocation Index(EI) +/- SEMean Index of IsolatedClick Use (IC) +/- SEResident 4.24 +/- 0.51 0.78 +/- 0.23Transient 0.16 +/- 0.04 3.29 +/- 0.99t-test p<.000 p=.02438away sareaCentral British Columbia CoasteStudy AreaFigure 1 . Map of the northwest coast of North America showing the study areas .390	7 .4Time (cosec)Figure 2. Waveform of a killer whale echolocation click. The duration of the click isapproximately 2.2 msec .'' ,	i 11111 .4I(a)400	16Frequency (kHz)32	 6	 I	 +	 1	 b	 1	 0	I'	ni	 ~(b)0	16	32Frequency (kHz)Figure 3. Energy distribution of a click from a click train (a), and of an isolated click (b) .41("_.r '•,,' 1; N, ,, .r e000'ti ISM,t`•41001, ,vy : .i„ h LPtiH 1'„x ''1 ,~y 4 ,. itOyu' ' i"'fit,	, 1.	:~` Y,~~ t ,•ryy~t~	~•	" ~" ,''r1'	1	'r	•r i.~ '	~.' '~~tYi .	a . .ri r,•L .Y r.'',l ..Iii1~ _	•., ..	1 , u r. .~l~f,2	4Time (s)Figure 4. Waveform and spectrogram of four seconds of a resident killer whale click train.The dotted line on the waveform plot is a measure of relative amplitude . Clicks are darkvertical lines on the spectrogram plot ; the diffuse shaded areas from 20 kHz up are click echoes.(a)VaEa32(b)042(a)a 0EQ11Ir ~~ .1 r.0	2	4Time (s)Figure 5 . Waveform and spectrogram of four seconds of a transient killer whale echolocationclick train . The dotted line on the waveform plot is a measure of relative amplitude .5040302010q TransientsResidents43f	ml	r	s	St	tBehaviour categoryFigure 6 . Number of analyzed five minute recordings of resident and transient killer whales,by behavioural activity. Abbreviations: f=foraging, ml=milling, r=resting, s=socializing,st=slow-travelling, t=travelling .3.53 .02.52 .01 .51 .00.50.0f44f	ml	r	S	St	tBehaviour categoryFigure 7. Mean echolocation index values for resident and transient killer whales, by behaviourcategory. Resident means are stars, transient means are circles. The echolocation index is square-root transformed as follows: x'=(x+.05).5. Standard error bars are shown.45(a)(b)1 .	3	6	9	12Echolocation index1 .	1	2	3	4Echolocation indexFigure 8 . Cumulative frequency distribution of echolocation index levels for resident (a) andtransient (b) killer whales . Note the differing axes .46Group sizeFigure 9 . Echolocation index level as a function of group size for resident killer whales . Thedata points have been "jittered" (shifted by small random amounts) to reveal overlapping values .47Group sizeFigure 10. Echolocation level as a function of group size for transient killer whales . Thecontinuous line is a logarithmic regression of echolocation on group size, for all the data shown.The broken line is a logarithmic regression of echolocation on the non-zero values shown . Thedata points have been jittered to reveal overlapping values .20(a)000 15XEmSept 1- - - - Aug 11	 Aug 8- - - July 28	 July 25485	10	15	20	25Depth (m)(b)302010June 15	 June 10- - - May31	 May 3010	15	20	25Depth (m)Figure 11 . Photocell resistance as a function of depth. Readings were taken in Prince WilliamSound, Alaska (a), and on the central mainland coast of British Columbia (b), in the summer of1991 .493 .5 ►03 .0°2 .50°0 00 °o °2 .0°°°q °°•°1 .5°° 0°°° 0 °0°1 .0	0	° °	° 00.50.00°q °00°0°0o 00000-0.50 2 4	6	8►10 12Secchi depth (m)Figure 12 . Echolocation index as a function of Secchi depth for resident killer whales . Thedata points have been jittered to reveal overlapping values .2 .00000 000 000000o 0	00°°	°q 0 0 0 0 o W ~O•oO1 .51 .00.50.012104	6	82-0.5050Secchi depth (m)Figure 13 . Echolocation index as a function of Secchi depth for transient killer whales . Thedata points have been jittered to reveal overlapping values .51Figure 14. Echolocation index plotted against environmental and vessel noise levels for residentkiller whales . Noise is in 10 dB increments from an arbitrary baseline . The surface was fittedusing least distance weighted squares techniques .21052Figure 15. Echolocation index plotted against environmental and vessel noise levels fortransient killer whales . Noise is in 10 dB increments from an arbitrary baseline . The surfacewas fitted using least distance weighted squares techniques .53CHAPTER THREEGeneral ConclusionsMy work has shown that resident and transient populations of killer whales varymarkedly in their use of echolocation. The extent of this difference suggests that generalisationsabout the echolocation behaviour of any species of odontocete may be risky at best . Populationsmay develop echolocation strategies that are dependent on available prey, predators, and otherlocal ecological factors . Such adaptive strategies would maximize the functional aspects ofecholocation, such as prey detection, and minimize the incidental costs, such as predatorattraction. I assume in this thesis that the echolocation strategies of both resident and transientkiller whales are indeed well-suited to their particular ecological niches . This is not anadaptationist approach sensu Gould and Lewontin (1979) : I have not assumed that becauseparticular traits have evolved, they are necessarily adaptive in their present context . Based onevidence that echolocation skills are learned (Wood & Evans 1980 ; Evans & Awbrey 1988), Isuggest instead that odontocetes adapt their innate echolocation capabilities to suit their specificecological circumstances.Transient killer whales used isolated clicks and short, quiet, irregular click trains, whichwere difficult to discern against background noise . I suggested in Chapter Two that this use ofecholocation is part of a hunting strategy involving stealth. This does not necessarily mean thattransients are proficient echolocators that derive large amounts of information from their crypticclicks . On the contrary, there are two reasons to suspect that transient echolocation may berelatively undeveloped . First, transient foraging behaviour probably affords few opportunities54for learning sophisticated echolocation techniques, since echolocation is used relativelyinfrequently. Second, the task of hunting marine mammals with sonar may be relatively simplecompared to the task of finding fish, since the lungs of marine mammals are good reflectors ofunderwater sound.I reported in Chapter Two that resident killer whales used echolocation much more oftenthan transients . They used longer, louder, and more regular click trains and used isolated clicksless frequently than transients . Resident echolocation was particularly strongly associated withforaging and travelling. During this study resident killer whales fed principally on salmon(Onchorhynchus spp.) that were aggregated into schools . A difficulty for predators feeding onschooling fish is that unless they are able to track and chase down specific individuals, they facean array of prey that never tires . I suggest that a function of echolocation for residents is thatit enables them to "lock on" to specific fish and chase them through visually-confusing schools.This mechanism would depend on the fish being individually-recognizable by echolocation.Since tissue in water is relatively transparent to sound, the strongest echo from a fish comesfrom its air bladder . The air bladder forms a resonant chamber that rings on exposure to soundwith a frequency that varies with its volume . Thus, it is likely that individual fish do indeedreturn unique echoes . There have been no tests of the ability of an odontocete to recognizeindividual fish . However, Thompson and Herman (1975) showed that bottlenose dolphins arecapable of discriminating very slight differences in frequency, which is a requirement for thistechnique. A similar mechanism may enable residents to discriminate desirable from undesirableprey from a distance. On several occasions in Prince William Sound, I saw killer whales rapidlytravel 25-50 m, bypassing schools of pink salmon, (O. gorbuscha) to catch larger fish, probably55chum or coho salmon (O. keta or O. kisutch).I suggested in Chapter Two that passive listening is probably used in important ways byforaging killer whales, particularly transients. Transients use echolocation sparingly, and oftentravel silently . They do not appear to locate prey using underwater vision, since they initiateattacks from well beyond the range of visibility. Transients were seen raising their heads abovethe surface ("spyhopping") near seal haulouts during this study, but were not seen scanningabove the surface in open water . As noted in the previous chapter, transients have beenobserved to locate and pursue prey immediately after vessel noise was reduced, which isevidence of passive listening . In addition, Guinet (1991,1992) in a study of predation by killerwhales on elephant seals, was able to attract killer whales close to shore by striking the surfaceof the water to simulate seals entering the water . Hoelzel (1991) reported similar evidence ofthe use of passive listening by killer whales hunting sea lions near shore.It is harder to draw inferences about the role of passive listening in foraging residentkiller whales . In Chapter Two, I suggested that fish sounds probably enable killer whales todetect schools under quiet conditions using passive listening. However, residents usuallyecholocate during foraging, arguably for the reasons described above . Residents commonlyproduce little or no echolocation in other behavioural contexts . Passive listening may be theprincipal mechanism maintaining orientation at those times (Norris 1967) . Schevill & Watkins(1966) reported that a killer whale (subsequently identified as a resident, Ford 1984) held in anet pen collided with an obstacle that was quietly lowered into its path at night . After a singlecollision, the whale clicked briefly, then continued to swim silently, avoiding subsequent56collisions unless the object was moved . This experiment was repeated many times with thesame result . Thus the whale was orienting reliably without echolocation or vision, presumablyby using passive acoustic cues.Echolocation in wild bats has been investigated much more thoroughly than in wildodontocetes . Insectivorous bats echolocate more than carnivorous species that forage for smallterrestrial mammals (Fiedler 1979) . Bats hunting flying insects may experience the same kindof advantage of echolocation that I have proposed for resident killer whales : it may enable themto pick out and track individual prey and follow them through a swarm of other insects.Carnivorous bats, like transients, are probably silent to avoid alerting their prey (Fiedler 1979).Perhaps the most obvious point of comparison between resident killer whales andinsectivorous bats is that they both exploit food supplies that are seasonally superabundant.Resident killer whales run little risk of depleting incoming schools of salmon in the summer andfall (Nichol 1990) . Similarly, bat predation probably does not appreciably affect the density ofswarming insects (Wilkinson, 1992) . Thus, neither suffers an appreciable cost by attractingcompetitors with echolocation . During seasons when prey are not abundant, the cost ofecholocation in attracting competitors may be much higher . At the same time, a proposedadvantage to echolocation use (increased efficiency in hunting aggregated prey) would decrease.Thus, I predict that both whales and bats decrease their use of echolocation when food is scarceor dispersed. The risk of attracting predators does not apply to killer whales, which have noknown predators, and is probably also low for bats (Fenton 1980) .57Like many field studies of the behaviour of social mammals, this research has raisedmore questions than it has provided answers . Some of these questions are listed below, withsuggestions of how they might be addressed in future projects.1) Role of passive listening: effect of masking noise . I attempted to assess the role of passivelistening by examining the relationship between ambient noise levels and echolocation behaviour.The only effect I detected was increasing click amplitude by residents in response to increasingnoise levels. However, these levels were not manipulated directly . If noise levels wereartificially elevated by projecting broadband random noise, changes in echolocation behaviourcould be monitored more systematically, in order to test the model of echolocation strategiespresented in Chapter Two . This model predicts that residents, but not transients, should increasetheir click amplitude in response to increasing broadband noise, and that transient foragingsuccess should decline under conditions of elevated noise.2) Role of passive listening : response to simulated prey sounds . Guinet (1991,1992) reportedthat killer whales responded to simulated sounds of elephant seals entering the water . It wouldbe worthwhile to conduct similar trials simulating the sounds of offshore prey, such as Dall'sporpoise . This would test whether passive listening is used in offshore foraging, and if so,whether swimming or vocal cues are most important in prey detection.3) Seasonal effects on echolocation use . One limitation of this study is that it was carried outduring a time of year when resident killer whales were feeding principally on a single abundantprey type, namely salmon (Nichol 1990) . Based on scattered observations and the stomach58contents of several beach-cast carcasses, it appears that residents in the winter may have a varieddiet of non-schooling midwater and bottom fish (G . Ellis, pers . comm .). A study similar to thisone but conducted throughout all seasons could evaluate whether residents echolocate less whentheir prey are dispersed. One difficulty with such a study is that the winter ranges of residentkiller whales are poorly known.4) Reactions of fish to killer whale echolocation . In order to explain differential costs toecholocation, I have argued that fish are unlikely to detect echolocation sounds . This has notbeen empirically tested. It is possible that killer whales voluntarily adjust the bandwidth of theirclicks to control their audibility to fish . At times it may be desirable to produce audible soundsto herd fish (Simila & Ugarte 1991 ; Osborne 1986) . At other times it is probably advantageousfor residents to remain undetected, particularly when chasing individual fish . Unpublishedexperiments that used recorded killer whale sounds have been attempted (J. Ford, pers.comm.).These were inconclusive, partly as a result of the difficulty involved in accurately reproducinghigh intensity, short duration sounds . Tests of the reaction of fish to real killer whaleecholocation could be carried out by placing fish in visually-opaque but acoustically-transparentpens, in tanks with captive killer whales.5) Role of learning in the development of echolocation strategies . I have suggested thatodontocetes adapt echolocation strategies to fit local ecological conditions . Do individualsdevelop their own strategies by trial and error, or are they learned from other pod members?In the latter is true, how long does it take to add new techniques to a repertoire of echolocationskills? The answers to these questions may help to explain the separation of resident and59transient populations . The echolocation skills that residents use for hunting fish are poorly-suited for transients hunting marine mammals, and vice versa . Are the differences inecholocation simply consequences of the feeding specializations, or are they also factorsmaintaining those specializations? To determine whether a commitment to a given echolocationstrategy reflects an inability to alter techniques, it is necessary to know how much innovationoccurs naturally . The best approach to these questions might involve a combination of work oncaptive and wild killer whales . First, the ability of individuals to develop new echolocationstrategies could be tested directly. This could be done by presenting novel echolocation tasksto captive individuals and rewarding innovative forms of echolocation . Categories of innovationmight include click repetition rate, click interval regularity, click amplitude, click frequencystructure, the use of isolated clicks, and/or the suppression of echolocation in favour of passivelistening . A second experiment could assess the ability of individuals to learn by example . Thiscould be carried out by testing whether captive killer whales learn echolocation skills for newtasks from individuals already possessing those skills . A third step would be to look atvariation in echolocation behaviour in the wild between groups of residents or between groupsof transients . If all groups of the same killer whale type use similar echolocation strategies, itwould suggest that innovation occurs only on a long time scale . If groups are highly variable,it would suggest that innovative changes arise frequently .60REFERENCESAndersen, M.E. & Racey, P.A. 1991 . Feeding behaviour of captive brown long-eared bats,Plecotus auritus. Anim. Behay., 42, 489-493.Andersen, S . 1970. Auditory sensitivity of the harbour porpoise, Phocoena phocoena. In:Investigations on Cetacea, Vol . II (Ed. by G. Pilleri), pp . 255-259 . Switzerland : BentiliAG Berne.Au, W.W.L. 1980. Echolocation signals of the Atlantic bottlenose dolphin (Tursiops truncatus)in open waters . In: Animal Sonar Systems (Ed. by R.-G. Busnel & J .F. Fish), pp . 251-282. New York : Plenum Press.Au, W.W.L. 1990. Target detection in noise by echolocating dolphins . In: Sensory Abilities ofCetaceans (Ed. by J. Thomas & R . Kastelein), pp . 203-216 . New York : Plenum Press.Au, W.W.L., Floyd, R .W. & Haun, J.E. 1978. Propogation of Atlantic bottlenose dolphinecholocation signals . J. Acoust. Soc Am., 64, 411-422.Au, W.W.L., Floyd, R.W., Penner, R.H. & Murchison, A.E. 1974. Measurement of echolocationsignals of the Atlantic bottlenose dolphin (Tursiops truncatus Montagu), in open waters.J. Acoust. Soc. Am., 56, 1280-1290.Au, W.W.L. & Hammer, C. 1978. Analysis of target recognition via echolocation by an Atlanticbottlenose porpoise (Tursiops truncatus) . J. Acoust. Soc. Am., 64, suppl. 1 : 587.Au, W.W.L. & Jones, L . 1991 . Acoustic reflectivity of nets : implications concerning incidentaltake of dolphins . Mar. Mamm. Sci., 7, 258-273.Au, W.W.L., Penner, R.H. & Kadane, J. 1982. Acoustic behaviour of echolocating Atlanticbottlenose dolphins . J. Acoust. Soc. Am., 71, 1269-1275.Au, W.W.L. & Turl, C .W. 1983. Target detection in reverberation by an echolocating Atlanticbottlenose dolphin . J. Acoust. Soc . Am ., 73, 1676-1681 .61Au, W.W.L. & Turl, C .W. 1984. Dolphin biosonar detection in clutter : Variation in the payoffmatrix . J. Acoust . Soc. Am., 76, 955-957.Awbrey, F.T., Thomas, J .A., Evans, W.E., & Leatherwood, S . 1982. Ross Sea killer whalevocalizations : preliminary description and comparison with those of some northernhemisphere killer whales . Rep . Int . Whal. Commn., 32, 667-670.Bain, D. 1992. Hearing abilities of killer whales . Unpubl . manusc. for National Marine FisheriesService, Seattle Washington. Contract Nos. 43ABNF002499 and 43ABNF002500.Baird, R.W., Abrams, P .A. & Dill, L.M. 1992. Possible indirect interactions between transientand resident killer whales: implications for the evolution of foraging specializations inthe genus Orcinus. Oecologia, 89, 125-132.Baird, R.W. & Stacey, P.J. 1989. Observations on the reactions of sea lions, Zalophuscalifornianus and Eumetopias jubatus to killer whales, Orcinus orca : evidence of "prey"having a "search image" for predators . Can. Fld.-Nat., 103, 426-428.Baldridge, A. 1972. Killer whales attack and eat a gray whale . J. Mammal . 53:898-900.Barclay, R .M.R. 1982. Interindividual use of echolocation calls : eavesdropping by bats . Behay.Ecol . Sociobiol ., 10, 271-275.Bigg, M.A., Ellis, G.M. & Balcomb, K.C. III. 1986. The photographic identification ofindividual cetaceans . Whalewatcher, 20, 10-12.Bigg, M.A., Ellis, G.M., Ford, J.K.B . & Balcomb, K.C. III. 1987 . Killer Whales : A Study oftheir Identification, Genealogy and Natural History in British Columbia and WashingtonState . Nanaimo, British Columbia: Phantom Press.Bigg, M.A., Olesiuk, P .F., Ellis, G.M., Ford, J .K.B. & Balcomb, K.C. III. 1990. Socialorganization and genealogy of resident killer whales (Orcinus orca) in the coastal watersof British Columbia and Washington State . Rep. Int. Whal . Commn., (Special Issue 12),383-405 .62Boyes, D.R. 1982. The role of sound in the British Columbia troll salmon fishery . M.Sc. Thesis.University of British Columbia, Vancouver.Dawson, S .M. 1991 . Modifying gillnets to reduce entanglement of cetaceans . Mar. Mamm. Sci .,7, 274-282.Diercks, K.J., Trochta, R.T. & Evans, W.E. 1973. Delphinid Sonar: measurement and analysis.J. Acoust. Soc. Am ., 54, 200-204.Diercks, K.J., Trochta, R.T., Greenlaw, C.F. & Evans, W .E. 1971 . Recording and analysis ofdolphin echolocation signals . J. Acoust. Soc . Am., 49, 1729-1732.Ellis, G. 1987 . Killer Whales of Prince William Sound and Southeast Alaska . A Catalogue ofIndividuals Photoidentified, 1976-1986 . Hubbs Marine Research Center Technical ReportNo. 87-200.Evans, W.E. & Awbrey, F .T. 1988 . Natural history aspects of marine mammal echolcation:feeding strategies and habitat . In: Animal Sonar (Ed. by P.E. Nachtigall & Moore,P.W.B .), pp. 521-534 . New York : Plenum Press.Faure, P.A. & Barclay, R.M.R. 1992. The sensory basis of prey detection by the long-eared bat,Myotis evotis, and the consequences for prey selection . Anim. Behay., 44, 31-39.Fenton, M. B . 1980. Adaptiveness and ecology of echolocation in terrestrial (aerial) systems . In:Animal Sonar Systems (Ed. by: R.-G. Busnel & J .F. Fish), pp . 427-446. New York:Plenum Press.Fiedler, J. 1979. Prey catching with and without echolocation in the Indian false vampire(Megaderma lyra) . Behay. Ecol. Sociobiol ., 6, 155-160.Ford, J.K.B . 1984. Call traditions and dialects of killer whales (Orcinus orca) in BritishColumbia. Ph.D. Thesis . University of British Columbia, Vancouver.Ford, J .K.B . 1989. Acoustic behaviour of resident killer whales (Orcinus orca) off VancouverIsland, British Columbia. Can. J. Zool., 67, 727-745 .63Ford, J .K.B. 1991 . Vocal traditions among resident killer whales (Orcinus orca) in coastalwaters of British Columbia . Can. J. Zool ., 69, 1454-1483.Ford, J .K.B . & Fisher, H.D. 1978. Underwater acoustic signals of the narwhal (Monodonmonocerous). Can. J. Zool., 56, 552-560.Frost, K.J., Russell, R.B ., & Llowry, L .F. 1992. Killer whales, Orcinus orca, in the southeasternBering Sea: recent sightings and predation on other marine mammals . Mar. Mamm. Sci .,8, 110-119.Goodson, A .D., Klinowska, M. & Morris, R. 1988. Interpreting the acoustic pulse emissions ofa wild bottlenose dolphin (Tursiops truncatus) . Aquat. Mamm., 14, 7-12.Gould, S . . & Lewontin, R .C. 1979. The spandrels of San Marco and the Panglossian paradigm.Proc. R. Soc. Lond. B ., 205, 581-598.Guinet, C. 1991. Intentional stranding apprenticeship and social play in killer whales (Orcinusorca) . Can. J. Zool ., 69, 2712-2716.Guinet, C. 1992. Comportement de chasse des orques (Orcinus orca) autour des lies Crozet.Can. J. Zool ., 70, 1656-1667.Hawkins, A.D. 1986. Underwater sound and fish behaviour. In : The Behaviour of Teleost Fishes(Ed. by T.J. Pitcher), pp . 114-151 . London: Croom Helm.Hawkins, A.D. & Johnstone, A.D.F. 1978. The hearing of the Atlantic Salmon, Salmo salar. J.Fish Biol., 13, 655-673.Heimlich-Boran J .R. 1988. Behavioural ecology of killer whales (Orcinus orca) in the PacificNorthwest . Can. J. Zool ., 66, 565-578.Heise, K., Ellis, G. & Matkin, C. 1992. A Catalogue of the Killer Whales of Prince WilliamSound. Homer, Alaska: North Gulf Oceanic Society .64Hoelzel, A .R. 1991 . Killer whale predation on marine mammals at Punta Norte, Argentina ; foodsharing, provisioning and foraging strategy . Behay. Ecol. Sociobiol ., 29, 197-204.Hol, W.A. & Kamminga, C. 1979. Investigations of cetacean sonar I . Some results on thethreshold detection of hollow and solid spheres performed by the Atlantic bottlenosedolphin (Tursiops truncatus) . Aquat . Mamm., 7, 41-63.Jefferson, T.A., Stacey, P .J. and Baird, R.W. 1991. A review of killer whale interactions withother marine mammals ; predation to co-existence . Mammal. Rev . 21(4) :151-180.Kamminga, C. & Beitsma, G.R. 1990. Investigations on cetacean sonar IX : Remarks ondominant sonar frequencies from Tursiops truncatus. Aquat . Mamm ., 16, 14-20.Kamminga, C. & Wiersma, H . 1981 . Investigations on cetacean sonar II. Acoustical similaritiesand differences in Odontocete sonar signals . Aquat. Mamm ., 8, 41-62.Kellogg, W.N. 1958. Echo ranging in the porpoise . Science, 128, 982-988.Konishi, M. 1973. How the owl tracks it prey . Am. Sci ., 61, 414-424.Leatherwood, S ., Thomas, J.A. & Awbrey, F .T. 1981 . Minke whales off northwestern RossIsland . Antarctic Journal Review, 17, 154-156.Lien, J., Todd, S ., & Guigne, J . 1990. Inferences about perception in large cetaceans, especiallyhumpback whales, from incidental catches in fixed fishing gear, enhancement of nets by"alarm" devices, and the acoustics of fishing gear. In : Sensory Abilities of Cetaceans (Ed.by J. Thomas & R. Kastelein) pp . 347-362. New York: Plenum Press.Marten, K., Norris, K.S., Moore, P .W.B. & Englund, K.A. 1988 . Loud impulse sounds inodontocete sound production. In : Animal Sonar : Processes and Performance (Ed. byP.E. Nachtigall & Moore, P.W.B.), pp. 567-579 . New York : Plenum Press.McBride, A.F. 1956. Evidence for echolocation in cetaceans . Deep Sea Res ., 3, 153-154.65Miles, E. 1990. Procedures for push processing Kodak and Ilford HP5 film . Rep. Int. Whal.Commn., (Special Issue 12, Annex G), 33-34.Mohl, B. 1968. Auditory sensitivity of the common seal in air and water . J. Aud. Res ., 8, 27-38.Mohl, B . & Andersen, S . 1973. Echolocation : high-frequency components in the click of theharbour porpoise (Phocoena ph. L.) . J.Acoust . Soc. Am., 54, 1368-1372.Moore, P .W.B. & Pawloski, D.A. 1990. Investigations on the control of echolocation pulses inthe dolphin (Tursiops truncatus)In: Sensory Abilities of Cetaceans (Ed. by J.A. Thomas& R. A. Kastelein) pp . 305-316 . New York : Plenum Press.Moore, P.W.B. & Schusterman, R .J. 1987. Audiometric assessment of northern fur seals,Callorhinus ursinus . Mar. Mamm. Sci ., 3, 31-53.Morozov, V.P., Akopian, A.I. & Bardin, V .I . 1972. Repetition rate of ranging signals of dolphinsas a function of distance to target. Biofizika, 17, 139-144.Morton, A.B. 1990. A quantitative comparison of the behaviour of resident and transient formsof killer whales off the central BC coast. Rep. Int. Whal. Commn, (Special Issue 12),245-248.Myrberg, A.A. 1991. Sound communication and interception in fishes . In : Hearing and SoundCommunication in Fishes (Ed. by W. Tavolga, A .N. Popper & R.R. Fay), pp. 395-452.New York: Springer-Verlag.Nichol, L.M. 1990. Seasonal movements and foraging behaviour of resident killer whales(Orcinus orca) in relation to the inshore distribution of salmon (Oncorhynchus sp.).M.Sc. Thesis . University of British Columbia, Vancouver.Norris, K.S. 1967. Some observations on the migration and orientation of marine mammals . In:Evolution and Environment (Ed. by. E. Drake), pp.101-125. New Haven : Yale UniversityPress .66Norris, K.S . 1980. The structure and functions of cetacean schools . In : Cetacean Behaviour:Mechanisms and Functions (Ed. by L.M. Herman), pp . 211-261 . New York: Wiley andSons.Norris, K.S. & Evans, E.C. III . 1988. On the evolution of acoustic communication systems invertebrates . Part 1 : Historical Aspects . In : Animal Sonar (Ed. by P.E. Nachtigall &P.W.B . Moore), pp . 655-669 . New York : Plenum Press.Norris, K.S., Prescott, J.H., Asa-Dorian, P.V. & Perkins, P. 1961 . An experimentaldemonstration of echolocation behaviour in the porpoise, Tursiops truncatus (Montagu).Biol. Bull., 120, 163.Olesiuk, P.F., Bigg, M.A. & Ellis, G .M. 1990. Life history and population dynamics of residentkiller whales (Orcinus orca) in the coastal waters of British Columbia and WashingtonState . Rep. Int . Whal . Commn ., (Special Issue 12), 209-243.Osborne, R.W. 1986. A behavioral budget of puget sound killer whales . In: BehaviouralBiology of Killer Whales (Ed. by B .C. Kirkvold & J .S. Lockard), pp . 211-249 . NewYork: Allan R. Liss, Inc.Popper, A.N. 1980. Sound emission and detection by delphinids . In : Cetacean Behaviour:Mechanisms and Functions (Ed. by L.M. Herman), pp. 1-52. New York: John Wiley andSons.Roeder, K.D. 1975. Neural factors and evitability in insect behaviour . J . Exptl. Zool., 194, 75-88.Roitblat, H.L., Penner, R.H. & Nachtigall, P.E. 1990. Matching-to-sample by an echolocatingdolphin (Tursiops truncatus) . J. Exper . Psychol . Anim. Behay. Processes ., 16, 85-95.Schevill, W.E. & Watkins, W.A. 1966. Sound structure and directionality in Orcinus (killerwhale) . Zoologica, 51, 71-78.Scronce, B .L. & Johnson, C.S. 1976. Bistatic target detection by a bottle-nosed porpoise. J.Acoust. Soc. Amer ., 59, 1001-1002.67Simila, T. & Ugarte, F. 1991. Behavioural ecology of killer whales Orcinus orca in northernNorway. In: Abstracts of the Ninth Biennial Conference on the Biology of MarineMammals, p . 64 . Chicago, Illinois.Steiner, W.W., Hain, J.H., Winn, H.E. & Perkins, P.J. 1979. Vocalizations and feeding behaviorof the killer whale (Orcinus orca) . J. Mammal., 60, 823-827.Stevens, T.A., Duffield, D .A., Asper, E .D., Hewlett, K.G., Bolz, A., Gage, L .J. & Bossart, G.D.1989. Preliminary findings of restriction fragment differences in mitochondrial DNAamong killer whales (Orcinus orca) . Can. J. Zool ., 67, 2592-2595.Tavolga, W.N. (ed.) 1976 . Sound Reception in Fishes . Penn: Dowden, Hutchinson and Ross.Terhune, J .M. 1988. Detection thresholds of a harbour seal to repeated underwater high-frequency, short duration sinusiodal pulses . Can. J. Zool ., 66, 1578-1582.Thomas, J .A., Pawloski, J .L. & Au, W.W.L. 1990. Masked hearing abilities in a false killerwhale (Pseudorca crassidens) . In : Sensory Abilities of Cetaceans (Ed. by J. Thomas &R. Kastelein), pp . 395-404 . New York : Plenum Press.Thompson, R.K.R. & Herman, L .M. 1975. Underwater frequency discrimination in thebottlenosed dolphin (1-140 kHz) and the human (1-8 kHz) . J. Acoust. Soc. Am., 57, 943-948.Turl, C.W. 1990. Echolocation abilities of the beluga, Delphinapterus leucas: A review andcomparison with the bottlenose dolphin Tursiops truncatus . In : Advances in research onthe beluga whale, Delphinapterus leucas . Can. Bull . Fish. Aquat. Sci . 224, 119-128.Turl, C.W. & Penner, T.H. 1989. Differences in echolocation click patterns of the beluga(Delphinapterus leucas) and the bottlenose dolphin (Tursiops truncatus) . J. Acoust. Soc.Am., 86, 497-502.Turl, C.W., Penner, R.H. & Au, W.W.L. 1987. Comparison of target detection capabilities ofthe beluga and bottlenose dolphin . J. Acoust. Soc . Am., 82, 1487-1491 .68Turl, C. W., Skaar, D.J. & Au, W.W. 1991 . The echolocation ability of the beluga(Delphinopterus leucas) to detect targets in clutter . J. Acoust. Soc. Am., 89, 896-901.Watkins, W.A. 1980. Click sounds from animals at sea . In : Animal Sonar Systems (Ed. by R.-G.Busnel & J .F. Fish), pp . 291-297. New York: Plenum Publishing.Watkins, W.A. & Wartzok, D . 1985. Sensory biophysics of marine mammals . Mar. Mamm. Sci .,1, 219-260.Wilkinson, G.S. 1992. Information transfer at evening bat colonies . Anim. Behay., 44, 501-518.Wilkinson, L . 1990 . SYGRAPH: The System for Graphics . Evanston, Illinois : Systat, Inc.Wood, F .G. & Evans, W.E. 1980. Adaptiveness and ecology of echolocation in toothed whales.In : Animal Sonar Systems (Ed. by R.-G. Busnel, & J .F. Fish), pp . 381-425 . New York:Plenum Press.Wursig, B. 1989. Cetaceans . Science, 244, 1550-1557.Zar, J .H. 1984 . Biostatistical Analysis . Second Edition . New Jersey : Prentice Hall Inc .69APPENDIXThis appendix contains supplementary figures .70- 15(b).0(a)1 .02520cc-,000z-550	4	8	12	2	4	6	8Echolocation index'- 15(c)Echolocation index'4(d)10-	1 .o -	3-1054	8	12	16	3	5	7	9	11Echolocation index'	Echolocation index"Figure Al. Cumulative frequency histograms of echolocation level by resident killer whalesengaged in four behavioural activities. The behaviour categories are as follows: (a) foraging,(b) resting, (c) socializing and (d) slow-travelling .7140t.025t.030m020 2z1000z0	t	2	3Echolocation index1	2	3	4Echolocation index4 0(c)	15	(d)20151053020101000cz5ID0	t	2	3Echolocation index1 .004	0.8	1.2	1.6Echolocation index4 0.002zFigure A2 . Cumulative frequency histograms of echolocation level by transient killer whalesengaged in four behavioural activities . The behaviour categories are as follows : (a) foraging, (b)milling, (c) travelling and (d) slow-travelling .2	4	8	8	10Secchi Depth (m)2	4	8	8	10Secchi Depth (m)-030OA2A1412 127212 122	4	8	8	10Secchi Depth (m)2	4	8	8	10Secchi Depth (m)(e)	,ato0o2	4	8	8	10	12	0	2	4	8	8	10	12Secchi Depth (m)	Secchi Depth (m)(h)	'1A0.8OAoo0 02	4	8	8	10	12Secchi Depth (m)2	4	8	8	10	12Secchi Depth (m)Figure A3 . Plots of resident and transient killer whale echolocation level against Secchi depth,by behaviour category . The top four plots are residents engaged in : (a) foraging, (b) resting,(c) slow travelling and (d) socializing. The bottom four plots are transients engaged in : (e)foraging, (f) milling, (g) travelling and (h) slow-travelling . The lines are fitted using Lowesssmoothing techniques (see methods) .35251606-0a-08	06	16	25	36	45Relative Environmental Noise06	15	25	36Relative Vessel Noise46(a) 73251a06V _-06	-05 06	16	25	36	4bRelative Environmental Noise06	15	25	3.5Relative Vessel Noise3.506-06-05 45251606-oa-06	05	16	25	36	45Relative Environmental Noise08	15	25	36Relative Vessel Noise3a(C)25to0a45(d)	3a251a0a06	16	26	36Relative Vessel Noise-0b	-06-0 .6	06	16	25	36	46	-06Relative Environmental Noisetoas352546Figure A4. Plots of resident killer whale echolocation level against environmental and vesselnoise levels, by behaviour category . Behavioural activities (by rows) : (a) foraging, (b) resting,(c) socializing and (d) slow-travelling. Lines fitted using Lowess techniques .(a) 2.016IA06Q0742A	.15 -	-t .0 -	-06 -	--05-05	Ob	1 .5	25	36	46Relative Environmental Noise06	16	26	3.6	45Relative Vessel Noise(b) 2A 2.0tb 1 .5to toO6 05Q0 00-0b -05( C)-06	06	1 .6	26	36	46 -06 05	16	2b	36	4.5Relative Environmental Noise Relative Vessel Noise2A 2At5 -	-x15 - -to -cc IA- -05 - 0b - -OA-0LUWce ao--	--0b0J-06-05	06	15	25	35	45 -06 0.6	15	2b	35	48Relative Environmental Noise	Relative Vessel Noise2.016IA050.0151 .0O60.0-05	-0b 1--06 as	1 .b	2b	36	46	-0bRelative Environmental Noise06	16	26	36Relative Vessel Noise46Figure AS. Plots of transient killer whale echolocation level against environmental and vesselnoise levels, by behaviour category . Behavioural activities (by rows) : (a) foraging, (b) milling,(c) travelling and (d) slow-travelling. Lines fitted using Lowess techniques.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items