Testing the acoustic prey debilitation hypothesis - Oregon State

Testing the odontocete acoustic prey debilitation hypothesis: No stunning results Kelly J. Benoit-Bird College of Oceanic and Atmospheric Sciences, Oregon State University, 104 COAS Administration Building, Corvallis, Oregon 97331, and Hawaii Institute of Marine Biology, University of Hawaii, PO Box 1106, Kailua, Hawaii 96734

Whitlow W. L. Au Hawaii Institute of Marine Biology, University of Hawaii, P.O. Box 1106, Kailua, Hawaii 96734

Ronald Kastelein SEAMARCO, Julianalaan 46, 3843 CC Harderwijk, The Netherlands

共Received 19 January 2006; revised 9 May 2006; accepted 12 May 2006兲 The hypothesis that sounds produced by odontocetes can debilitate fish was examined. The effects of simulated odontocete pulsed signals on three species of fish commonly preyed on by odontocetes were examined, exposing three individuals of each species as well as groups of four fish to a high-frequency click of a bottlenose dolphin 关peak frequency 共PF兲 120 kHz, 213-dB peak-to-peak exposure level 共EL兲兴, a midfrequency click modeled after a killer whale’s signal 共PF 55 kHz, 208-dB EL兲, and a low-frequency click 共PF 18 kHz, 193-dB EL兲. Fish were held in a 50-cm diameter net enclosure immediately in front of a transducer where their swimming behavior, orientation, and balance were observed with two video cameras. Clicks were presented at constant rates and in graded sweeps simulating a foraging dolphin’s “terminal buzz.” No measurable change in behavior was observed in any of the fish for any signal type or pulse modulation rate, despite the fact that clicks were at or near the maximum source levels recorded for odontocetes. Based on the results, the hypothesis that acoustic signals of odontocetes alone can disorient or “stun” prey cannot be supported. © 2006 Acoustical Society of America. 关DOI: 10.1121/1.2211508兴 PACS number共s兲: 43.80.Ka, 43.80.Nd 关JAS兴

I. INTRODUCTION

The idea that odontocete cetaceans can use sound to debilitate their prey was first proposed by Bel’kovich and Yablokov 共1963兲 and later reviewed by Norris and Mohl 共1983兲. Odontocetes, or toothed whales, a group of approximately 70 species in six families, are predators that feed on a range of prey including cephalopods, crustaceans, fish, birds, and other marine mammals. As a group, they rely on sound for orientation, navigation, communication, and predator and prey detection 共Reynolds and Rommel, 1999兲. It has been proposed that loud, impulsive sounds made by cetaceans in the vicinity of prey can stun, disorient, and debilitate their prey, rather than just being used to find prey. Support for the prey-stunning hypothesis includes observations of dolphins in laboratory and field settings. Data that sperm whales have intact squid lacking any bite marks in their stomachs led Norris and Mohl 共1983兲 to suggest that the squid, rapid swimmers that are highly maneuverable, must be immobilized by the whale before consumption. A number of observations 共summarized in Marten et al., 2001; summarized in Norris and Mohl, 1983兲 that fish appeared lethargic and fish schools were depolarized by feeding dolphins in both captive and wild circumstances have been considered consistent with prey stunning by dolphins, though the mechanism has not been tested. Evidence to support the prey-stunning hypothesis also comes from results that show acoustic signals can indeed 1118

J. Acoust. Soc. Am. 120 共2兲, August 2006

Pages: 1118–1123

cause a loss of buoyancy control, abdominal hemorrhage, and even death in fish. Explosives including TNT and dynamite with exposure levels between 229 and 234 dB and black powder with an exposure level between 234 and 244 dB have been shown to cause debilitation 共Hubbs and Rechnitzer, 1952兲. Spark discharges with exposure levels between 230 and 242 dB can cause similar effects on fish 共Zageski, 1987兲, though negative results have been reported for some cephalopods 共Mackay and Pegg, 1988兲. An airgun exposing fish to an average of 216 dB peak-to-peak has been shown to cause death and debilitation in fish after multiple exposures over several hours 共Marten et al., 2001兲. Acoustic stunning of some fish has also been shown from highamplitude pure tones 共Hastings, 1995兲. Most of the acoustic signals that have experimentally shown to cause prey debilitation have rapid rise times 共several tens of microseconds or less兲 and high pressures 共⬎210 dB peak-to-peak兲, as do dolphin clicks. However, most of the received threshold levels shown to cause responses to these sounds are 10 to 25 dB above the maximum intensity recorded for dolphin clicks 共Au, 1993兲. In other words, they contain about 300 times more energy than a single echolocation click. Only repeated exposure to the airgun every 15 s for 4 h caused debilitation at received sound levels reported for dolphin sounds. Perhaps the most interesting evidence towards potential prey stunning by odontocetes comes from a very different animal taxa. Snapping shrimp produce broadband, impulsive sounds that reach source levels of 220 dB peak-to-peak

0001-4966/2006/120共2兲/1118/6/$22.50

© 2006 Acoustical Society of America

共Schmitz et al., 2000兲. Snapping shrimp have been shown to stun their prey using a single snap 共Herberholz and Schmitz, 1999兲. Although it is not known if the mechanism for stunning by snapping shrimp is the sound itself or the mechanical stress caused by the cavitating bubble or the jet of water, these are some of the first data to suggest that a biologically created sound, rather than anthropogenic sounds, could cause debilitating effects on prey and serve a predatory function. While the hypothesis that odontocete signals can directly cause debilitation in their prey has not been directly tested, the hypothesis has been attractive to both scientists and the popular media. In the last decade, Norris and Mohl’s 1983 synthesis of the hypothesis has been cited 35 times. Some of these papers invoke acoustic prey stunning as a mechanism to explain their results despite the lack of direct support for the mechanism. Media coverage of papers that have discussed the prey-stunning hypothesis 共Cranford, 1999; Marten et al., 2001兲 has included dramatic headlines such as “Killer clicks” 共New Scientist, 31 January 2001兲 and “Porpoise stun gun” 共Science Frontiers No. 29: Sep–Oct 1993兲. The Santa Cruz County Sentinel 共26 April 1998兲 summarized one study saying, “mysterious monsters are killing fish with murderous death beams.” The general acceptance of the hypothesis by both professional and lay people led us to attempt a direct test of the hypothesis that odontocete acoustic signals can debilitate or disorient prey. II. METHODS A. Study area

The experiment was conducted in an outdoor tank at the field station of the Netherlands National Institute for Coastal and Marine Management 共RIKZ兲 at Jacobahaven, Zeeland, The Netherlands in a freestanding rectangular tank 7.0 m long and 4.0 m wide with a water depth of 2.0 m. The tank rested on a 3-cm-thick layer of hard-pressed Styrofoam to reduce contact noise from the environment and was covered by a canopy to improve the clarity of video recordings. The water in the tank was pumped directly from the nearby Oosterschelde 共a lagoon of the North Sea兲. The salinity was 30%– 33%, the pH 7.9–8.1, nitrogen concentration ⬍5 mg/ l, and nitrate concentration ⬍0.1 mg/ l. The water was circulated via a sand, UV light, and carbon filter. To make the environment inside the tank as quiet as possible, the filter unit had a low-noise “whisper” pump. To reduce contact noise entering the pool, the pump and filter unit were also placed on hardpressed rubber. Water parameters 共temperature, salinity, oxygen, and nitrate兲 were measured daily and remained well within the boundaries suitable for normal activity of the fish. B. Subjects

We tested the effects of simulated odontocete clicks on three species of fish: Atlantic herring 共Clupea harengus兲, sea bass 共Dicentrarchus labrax兲, and Atlantic cod 共Gadus morhua兲. The Animal Welfare Commission of the Netherlands government stipulated that the fish used must feed readily in captivity, so they had to come from aquaria, though all were originally wild-caught. Sea bass and cod were fed to satiation each day after the sessions on a diet of J. Acoust. Soc. Am., Vol. 120, No. 2, August 2006

FIG. 1. Experimental setup. The fish is enclosed in monofilament enclosure next to an acoustic transducer. While in the cage, the subjects were videotaped from above and from the side.

raw fish. Herring were similarly fed a diet of Trouvit pellets size no. 00 共Nutreco Aquaculture兲. Fish were returned to the aquaria following the experiment to comply with animal care regulations. Fish species were chosen to represent a variety of swimbladder types because of the hypothesis that the mechanism of fish debilitation is the interaction of the acoustic signal with air-filled cavities in the fish. The herring species has a modified form of the primitive swimbladder as a physostome, with air bladder extensions to the lateral line and labyrinth 共Blaxter et al., 1979; 1981兲. Species in this family have been shown to be especially sensitive to sound including high frequencies 共Mann et al., 1997兲 and the Pacific herring 共Clupea pallassi兲 and American shad 共Alosa sapidissima兲 have even been shown to respond to echolocation signals 共Mann et al., 1998; Wilson and Dill, 2002兲. We hypothesized that the herring was the most likely candidate for stunning because of these anatomical adaptations. Sea bass is a euphysoclistous fish; they are physostomous as juveniles but physoclistous as adults 共Chatain, 1986兲. Cod have the most derived swimbladder form, a completely closed physoclist bladder. The swimbladder has also been shown to play a role in the hearing abilities of cod 共Sand and Enger, 1973兲. If debilitation occurred, the variety of swimbladder types used could help elucidate the mechanism of acoustic interaction with the fish. C. Experimental setup

For testing, fish were placed inside a 0.5-m-diameter, 0.3-m-high cylindrical monofilament enclosure supported by a plastic ring on the top and bottom 共Fig. 1兲. This allowed the fish to swim freely while keeping the subjects positioned in front of the transducer producing the stimulus. The enclosure was suspended 1 m below the water’s surface with the acoustic transducer placed immediately adjacent to the enclosure, making the subject fish, on average, 0.25 m away from the source of the stimuli. Two underwater video cameras were focused on the net enclosure, one looking directly down on the enclosure from above, and one looking from the side, directly across from the transducer. These cameras alBenoit-Bird et al.: Acoustic prey stunning by odontocetes

1119

FIG. 2. Acoustic stimuli used in the experiment were three synthetic odontocete echolocation signals recorded using the calibration hydrophone. The waveforms 共time-amplitude plots兲 of each signal are shown on the left, and the spectra 共frequency-amplitude兲 plots on the right.

lowed us to observe the behavior of the fish without causing disturbance, and permitted us to make sure the fish was centered in front of the acoustic transducer when the stimulus was presented. The size of the enclosure was a compromise between allowing fish movement and controlling exposure level. Previous studies of stunning have observed loss of buoyancy control and changes in small-scale swimming behavior which this size enclosure should not limit. Individual fish or fish in groups of four were placed inside the enclosure and allowed to acclimate to their new conditions for at least 30 min. The behavior of the fish was recorded using the two video cameras for 5 min prior to the presentation of the acoustic stimuli and 5 min postexposure for comparison. The hypothesis was that fish would be dramatically affected by the presentation of clicks so we looked for changes in activity level that would indicate avoidance or loss of swimming ability as well as changes in tilt or roll angle of greater than 30 deg, indicating loss of buoyancy control. We also looked for fish survival for 1 week following the experiments as an indication of long-term internal damage. We were prepared to immediately euthanize any fish that appeared to be in extreme distress immediately following acoustic exposure. D. Stimuli

A variety of odontocete species have been hypothesized to debilitate their prey with echolocation clicks, so we simulated clicks of three different species in our experiment. The clicks as recorded by a calibration hydrophone are shown in Fig. 2. Clicks were produced using a function generator computer plug-in board with a sampling rate of 1 MHz. We measured the source level and the exposure level of each signal using a calibrated B&K 8103 hydrophone placed at a variety of ranges from the transmit transducer; the results are shown in Fig. 3. The first signal we tested, a bottlenose dolphin 1120

J. Acoust. Soc. Am., Vol. 120, No. 2, August 2006

FIG. 3. The measured sound-pressure level of each of the three signals as a function of range from the source transducer. The levels the fish were exposed to at the center of the cage, indicated by the paired lines, were calculated based on these curves.

共Tursiops truncatus兲 click, had a peak frequency of 120 kHz and a source level of 203 dB re : 1 ␮Pa peak-to-peak, exposing fish in the middle of the cage to levels of 213 dB re : 1 ␮Pa peak-to-peak. This signal was produced using a custom-built planar transducer consisting of a 1–3 composite piezoelectric element of 6.4 cm diameter. This transducer has a 12.5-deg beamwidth in the far field, and though its beamwidth in the near field is unknown, subject fish were often directly in front of the transducer during the experiment, exposing them to the highest possible sound levels. The second signal, a killer whale 共Orcinus orca兲 click 共Au et al., 2004兲, had a peak frequency of 55 kHz and had a source level of 200 dB re : 1 ␮Pa peak-to-peak, exposing the fish to a level of 208 dB re : 1 ␮Pa peak-to-peak. The final, low-frequency click had a peak frequency of 18 kHz and a source level of 187 dB re : 1 ␮Pa peak-to-peak, exposing the fish to 193 dB re : 1 ␮Pa peak-to-peak 共Fig. 3兲. Both of these lower-frequency signals were produced using a custom-built transducer consisting of three piezoelectric ceramic spheres having diameters of 5.0, 3.8, and 1.3 cm and stacked in a column. Each sphere has a different resonance frequency and the three were driven in parallel so that the transducer has a low directivity index. The frequency range, duration, amplitude, and energy flux density of each of the three simulated clicks were comparable to the clicks produced by a range of odontocete species, providing comparable acoustic exposure to natural clicks, and thus should affect the subjects in the same way as natural clicks. Benoit-Bird et al.: Acoustic prey stunning by odontocetes

Clicks were presented to three individual sea bass and three individual cod as well as groups of four individuals of each species in two ways. The first was at static, regular pulse rates of 100, 200, 300, 400, 500, 600, and 700 clicks/ second for exposure times ranging from a minimum of 7 s up to 1 min. This type of click presentation is most similar to regular echolocation by odontocetes at close ranges 共less than 0.4 m, Evans and Powell, 1967兲. We also presented clicks in modulated pulse “sweeps.” Clicks repetition rate increased from 100 to 700 clicks/ s over either 1.1, 2.2, or 3.2 s. This click pattern is similar to the “terminal buzz” produced by echolocating animals as they make the final approach to a target 共Au, 1993兲. We presented this signal type to three individuals cod and three individual herring as well as to a group of four cod and a group of four herring. E. Behavioral parameters

To observe effects on buoyancy control, the video data were used to determine the tilt and roll angle of each fish in both individual and group tests every minute for the 15 min pre- and postsound exposure. To determine if fish avoided the sound’s source, the percent of time each fish spent in each quadrant of the cage was recorded for the pre- and postexposure observation periods. To measure fish movement, the number of times each fish crossed between the quadrants of the cage was compared from pre- and postexposure periods.

FIG. 4. The distribution of fish roll 共top兲 and tilt 共bottom兲 angles before and after exposure to the acoustic stimulus. An angle of zero represents a vertical orientation. There were no significant differences between pre- and postexposure periods and no significant effects of species, number of animals in the cage, or signal type, so all stimuli and fish species results are pooled in this graph.

F. Statistics

Tilt and roll angle of each fish in both individual and group tests every minute for the 15-min pre- and postsound exposure was compared using a repeated-measures analysis of variance 共ANOVA兲. A series of paired t-tests was used to compare the pre- and postexposure quadrant residence time for each fish to determine if fish avoided the transducer. To measure fish movement, the number of times each fish crossed between the quadrants of the cage was compared from pre- and post-exposure periods using a paired t-test. III. RESULTS

No clear changes in behavior were observed during the relatively short ensonification period. To observe effects on buoyancy control, the video data were used to determine the tilt and roll angle of each fish in both individual and group

tests every minute for the 15-min pre- and postsound exposure 共Fig. 4兲. A repeated-measures ANOVA revealed no significant differences between pre- and postexposure angles and showed no interaction of this with signal type, the number of fish in the cage, or species 共Tables I and II, p ⬎ 0.05兲. To determine if fish avoided the sound’s source, the percent of time each fish spent in each quadrant of the cage was recorded for the pre- and postexposure observation periods. A series of paired t-tests was used to compare the pre- and postexposure quadrant residence time for each fish 共Table III兲. No significant differences were observed in any of the four quadrants 共p ⬎ 0.05兲. This suggests that there was no avoidance of the sound’s source. To measure fish movement, the number of times each fish crossed between the quadrants of the cage was compared from pre- and postexposure periods using a paired t-test. No

TABLE I. Repeated-measures ANOVA for fish roll angle pre- and postexposure to acoustic stimuli. Source Pre vs post *Signal *Group size *Species *Signal *Group size *Signal *Species *Group size *Species *Signal *Group size *Species Error 共pre vs post兲


SS

df

MS

F

p

31.61 1095.51 57.13 30.19 886.39 1676.07 1.10 635.66 9664.05

1 18 1 2 13 36 1 12 168

31.61 60.86 57.13 15.09 68.18 46.56 1.10 52.97 57.52

0.55 1.06 0.99 0.26 1.19 0.81 0.02 0.92

0.46 0.40 0.32 0.77 0.29 0.77 0.89 0.53

Benoit-Bird et al.: Acoustic prey stunning by odontocetes

1121

TABLE II. Repeated-measures ANOVA for fish tilt angle pre- and postexposure to acoustic stimuli. Source Pre vs post *Signal *Group size *Species *Signal *Group size *Signal *Species *Group size *Species *Signal *Group size *Species Error 共pre vs post兲

SS

df

MS

F

p

196.23 5893.71 213.63 1085.74 5231.02 10284.76 183.00 5910.46 54943.68

1 18 1 2 13 36 1 12 168

196.23 327.43 213.63 542.87 402.39 285.69 183.00 492.54 327.05

0.60 1.00 0.65 1.66 1.23 0.87 0.56 1.51

0.44 0.46 0.42 0.19 0.26 0.68 0.46 0.13

significant difference was observed 共mean difference =0.0044± 0.0016 95% CI, t = 0.62, df= 170, p = 0.43兲. We did not have any mortality or even slight abnormal behavior of our subjects, indicating no severe internal injuries from the testing. IV. DISCUSSION

We did not observe any measurable changes in the behavior of any of the species of fish in any of the conditions we tested either during or following exposure to odontocete clicks. There were no noticeable changes in swimming activity, no apparent loss of buoyancy control that would be indicated by tilting or rolling of the fish, and no movement away from the transducer. In order for prey stunning by regular or terminal buzz clicks to be a viable mechanism for fish capture by odontocetes, the fish would need to be dramatically and noticeably affected by the acoustic signals, which was not apparent in our results. In addition, we did not have any mortality of the tested fish, which might indicate internal damage from the acoustic signals. We tested a large number of variables including the frequency of the signal, pulse rate, and pulse sweeping. Our choice of the three signal types was to ensure that a wide frequency range was covered in the event that stunning might have a frequency dependency. These signals also covered the full range of frequencies employed by odontocetes from the relatively low frequency in the range used by the sperm whale to the high-frequency bottlenose dolphin. Fish in our experiment were subjected to very high levels of sounds for exposure times that are long relative to odontocete acoustic behavior during foraging. Norris and Mohl 共1983兲 proposed that the final click sweep, or terminal buzz, was the most likely signal to induce fish stunning. These non-“usual” clicks have very short in-

terclick intervals 共Madsen et al., 2005; 2002兲 with between 50 and 200 clicks per second. The clicks in this terminal phase as the odontocete closes in on the prey are typically lower in amplitude than “regular” echolocation clicks, by about 15 dB. Our terminal buzz signals reached 700 pulses per second for 7 s to 1 min, with a maximum amplitude of 213 dB re : 1 ␮Pa. We can describe a pulsed odontocete signal, p共t兲, using the normalized waveform method of Au 共1993兲 as A pp sN共t兲, 2

p共t兲 =

共1兲

where A pp is the peak-to-peak amplitude in ␮Pa and sN is the normalized waveform. The energy flux density of a single click will be e=

A2pp 4␳c

冕

T

sN共t兲2dt =

0

A2pp 1 u. , 4 ␳c

共2兲

where u=the integral of the square of the normalized waveform and T is the duration of the waveform. The total energy flux density for N clicks will be simply eT = N

A2pp 1 u. . 4 ␳c

共3兲

In the extreme experimental condition of 700 clicks per second for 7 s to 1 min time period, we exposed fish subjects to over 8 to 70 关from Eq. 共3兲兴 times more acoustic energy than would an odontocete emitting terminal buzz signals with a peak-to-peak SPL of 213 dB 共4.5⫻ 1010 ␮Pa兲 at a rate of 200 clicks per second for 3 s 共Madsen et al., 2005兲. Even if we assume an odontocete emitting terminal buzz signals at twice the amplitude 共9 ⫻ 1010 ␮Pa兲 at the same repetition rate and duration, our extreme case would represent 2 to 17

TABLE III. Paired sample t-tests comparing pre- and postexposure time spent in each quadrant of the cage by each fish. All species were pooled in this analysis. 95% Confidence interval of the difference Quadrant A B C D

1122

Mean paired differences

Lower

Upper

t

df

p

−0.0034 −0.0060 0.0007 0.0087

−0.0118 −0.0127 −0.0073 −0.0049

0.0050 0.0007 0.0087 0.0223

−0.80 −1.78 0.17 1.27

170 170 170 170

0.43 0.08 0.86 0.21



times more acoustic energy. In most of our experimental conditions, the fish subjects were exposed to more acoustic energy than would be expected for animals in the wild based on the observations of wild odontocete foraging. We tested these signals on multiple individuals of different sizes representing three species commonly consumed by odontocetes. These fish were selected based on their potential as prey species, their swimbladder morphology, and their availability in captivity. It has been proposed that acoustic signals can interact with the air-filled swimbladder to result in stunning. We chose a physoclistous, euphysoclistous, and physostomous fish species to be able to test for this effect, though no differences in response were exposed. We also tested any group effects by exposing groups of four individuals of a single species together. We did not observe any changes in coordination or coherence of individual fish within the groups. If stunning does occur in the wild, its induction may require a stress other than or in addition to the acoustic signal. Perhaps additional sensory cues are required. For example, perhaps the fish are stressed or stunned by seeing a large predator or by the pressure wave created by the dolphins as they approach their prey. Another possible explanation is that the odontocete acoustic behavior that induces stunning is different than has been described by previous studies. For example, the “bang” sound reported by 共Marten et al., 2001兲. These relatively long, high-amplitude sounds were recorded in the wild and in the laboratory. However, their source, the phonation system or a body movement like a tail slap, is not known. Despite the use of signals near the maximum source levels recorded for odontocete clicks, we could not induce stunning or even disorientation in the fish tested. Our results do not support the hypothesis that odontocetes use their clicks alone to induce stunning in prey. ACKNOWLEDGMENTS

We thank Sander van de Huel for assistance conducting the experiments. Jan van der Veen, Sea aquarium “het Arsenaal,” The Netherlands, lent us most of the study animals, and Gerard Visser, Michaël Laterveer, and Peter van Putten 共all of the Oceanium Department of Blijdorp Zoo兲, provided the herring and transported them to the study site. Gijs Rutjes 共Coppens International兲 provided some of the sea bass. We thank Mardi Hastings for valuable discussions about this work and Brigitte Kastelein and the volunteers of SEAMARCO for logistical support. The facilities of the research station were made available thanks to Dick Vethaak 共RIKZ兲, Roeland Allewijn 共RIKZ兲, and Wanda Zevenboom 共North Sea Directorate兲. This work was supported by the U.S. Office of Naval Research and the Netherlands Ministry for Agriculture, Nature, and Food Quality 共DKW-program 418: North Sea and Coast. This project complied to the Dutch standards for animal experiments 共Chris Pool, head of the Committee


for Animal Experiments of RIKZ兲 and was conducted under University of Hawaii Animal Care Protocol 04-019. This is HIMB contribution 1234. Au, W. W. L. 共1993兲. The Sonar of Dolphins 共Springer, New York兲. Au, W. W. L., Ford, J. K. B., Horne, J. K., and Newman Allman, K. A. 共2004兲. “Echolocation signals of free-ranging killer whales 共Orcinus orca兲 and modeling of foraging for chinook salmon 共Oncorhynchus tshawytscha兲,” J. Acoust. Soc. Am. 115, 901–909. Bel’kovich, V. M., and Yablokov, A. V. 共1963兲. “Marine animals share experience with designers,” Nauka Zhizn’ 30, 61–64. Blaxter, J. H. S., Denton, E. J., and Gray, J. A. B. 共1979兲. “The herring swimbladder as a gas reservoir for the acoustico-lateralis system,” J. Mar. Biol. Assoc. U.K. 59, 1–10. Blaxter, J. H. S., Denton, E. J., and Gray, J. A. B. 共1981兲. “The auditory bullae-swimbladder system in late stage herring larvae,” J. Mar. Biol. Assoc. U.K. 61, 315–326. Chatain, B. 共1986兲. “The swim bladder in Dicentrarchus labrax and Sparus auratus. 1. Morphological aspects of development,” Aquaculture 53, 303– 311. Cranford, T. 共1999兲. “The sperm whale’s nose: Sexual selection on a grand scale?” Marine Mammal Sci. 15, 1133–1157. Evans, W. W., and Powell, B. A. 共1967兲. “Discrimination of different metallic plates by an echolocating delphinid,” in Animal Sonar Systems: Biology and Bionics, edited by R. J. Busnel 共Laboratoire de Physiologie Acoustique, Jouy-en-Josas, France兲, pp. 363–382. Hastings, M. C. 共1995兲. “Acoustic stunning of marine animals,” J. Acoust. Soc. Am. 98, 2940–2941. Herberholz, J., and Schmitz, B. 共1999兲. “Flow visualisation and high speed video analysis of water jets in the snapping shrimp 共Alpheus heterochaelis兲,” J. Comp. Physiol., A 185, 41–49. Hubbs, C. L., and Rechnitzer, H. B. 共1952兲. “Report on experiments designed to determine effects of underwater explosions on fish life,” California Fish and Game Bulletin 38, 333–366. Mackay, R. S., and Pegg, J. 共1988兲. “Debilitation of prey by intense sounds,” Marine Mammal Sci. 4, 356–359. Madsen, P. T., Wahlberg, M., and Mohl, B. 共2002兲. “Male sperm whale 共Physeter macrocephalus兲 acoustics in a high-latitude habitat: Implications for echolocation and communication,” Behav. Ecol. Sociobiol. 53, 31–41. Madsen, P. T., Johnson, M. P., Aguilar de Soto, N., Zimmer, W. M. X., and Tyack, P. L. 共2005兲. “Biosonar performance of foraging beaked whales 共Mesoplodon densirostris兲,” J. Exp. Biol. 208, 181–194. Mann, D. A., Hastings, M. C., and Popper, A. N. 共1998兲. “Detection of ultrasonic tones and simulated dolphin echolocation clicks by a teleost fish, the American shad 共Alosa sapidissima兲,” J. Acoust. Soc. Am. 104, 562–568. Mann, D. A., Lu, Z., and Popper, A. N. 共1997兲. “A clupeid fish can detect ultrasound,” Nature 共London兲 389, 341. Marten, K., Herzing, D. L., Poole, M., and Allman, K. N. 共2001兲. “The acoustic predation hypothesis: Linking underwater observations and recording during odontocete predation and observing the effects of loud impulsive sounds on fish,” Aquat. Mamm. 27, 56–66. Norris, K. S., and Mohl, B. 共1983兲. “Can odontocetes debilitate prey with sound?” Am. Nat. 122, 85–104. Reynolds, J. E. III, and Rommel, S. A. 共1999兲. Biology of Marine Mammals 共Smithsonian Institution Press, Washington兲. Sand, O., and Enger, P. S. 共1973兲. Evidence for an auditory function of the swimbladder in the cod,” J. Exp. Biol. 59, 405–414. Schmitz, B., Versluis, M., von der Heydt, A., and Lose, D. 共2000兲. “A unique way of sound production in the snapping shrimp 共Alpheus heterochaelis兲,” J. Acoust. Soc. Am. 108, 2542. Wilson, B., and Dill, L. M. 共2002兲. “Pacific herring respond to simulated odontocete echolocation sounds,” Can. J. Fish. Aquat. Sci. 59, 542–553. Zageski, M. 共1987兲. “Some observations on the prey stunning hypothesis,” Marine Mammal Sci. 3, 275–279.


1123