Sperm whale predator-prey interactions involve chasing and buzzing, but no acoustic stunning

Abstract

The sperm whale carries a hypertrophied nose that generates powerful clicks for long-range echolocation. However, it remains a conundrum how this bizarrely shaped apex predator catches its prey. Several hypotheses have been advanced to propose both active and passive means to acquire prey, including acoustic debilitation of prey with very powerful clicks. Here we test these hypotheses by using sound and movement recording tags in a fine-scale study of buzz sequences to relate the acoustic behaviour of sperm whales with changes in acceleration in their head region during prey capture attempts. We show that in the terminal buzz phase, sperm whales reduce inter-click intervals and estimated source levels by 1-2 orders of magnitude. As a result, received levels at the prey are more than an order of magnitude below levels required for debilitation, precluding acoustic stunning to facilitate prey capture. Rather, buzzing involves high-frequency, low amplitude clicks well suited to provide high-resolution biosonar updates during the last stages of capture. The high temporal resolution helps to guide motor patterns during occasionally prolonged chases in which prey are eventually subdued with the aid of fast jaw movements and/or buccal suction as indicated by acceleration transients (jerks) near the end of buzzes.

Sperm whale predator-prey interactions involve chasing and buzzing, but no acoustic stunning

OPEN

R

A P

A. Fais,, M. Johnson, M.Wilson,, N. Aguilar Soto, & P.T. Madsen,

The sperm whale carries a hypertrophied nose that generates powerful clicks for long-range echolocation. However, it remains a conundrum how this bizarrely shaped apex predator catches its prey. Several hypotheses have been advanced to propose both active and passive means to acquire prey, including acoustic debilitation of prey with very powerful clicks. Here we test these hypotheses the acoustic behaviour of sperm whales with changes in acceleration in their head region duringprey capture attempts. We show that in the terminal buzz phase, sperm whales reduce inter-click prey are more than an order of magnitude below levels required for debilitation, precluding acoustic stunning to facilitate prey capture. Rather, buzzing involves high-frequency, low amplitude clicks well suited to provide high-resolution biosonar updates during the last stages of capture. The high temporal resolution helps to guide motor patterns during occasionally prolonged chases in which prey are eventually subdued with the aid of fast jaw movements and/or buccal suction as indicated by acceleration transients (jerks) near the end of buzzes.

It remains, then, to be inquired in what way the sperm whale (Physeter macrocephalus) usually does supply his enormous frame with sufficient food wrote the whaling surgeon Beale in 18401. Today, almost 200 years later, that question is still relevant for the largest tooth-bearing predator on the planet, a major nutrient recycler in the worlds oceans2 responsible for an annual biomass turnover that has been compared to the combined catches of human sheries3. Sperm whales are mainly teuthophagous predators that spend more than half of their life below 500 meters depth where they target meso- and benthopelagic prey. The main food source for sperm whales are medium sized squid (0.5 m and 13 kg), of which 7080% are slow moving, ammoniacal species4,5, while the remainder are faster moving squid and sh targeted at meso- and epipelagic depths68. Occasionally, much larger squid such as jumbo, giant and colossal squid4,9,10 have been found in sperm whale stomachs, as well as rarer prey items ranging from sardines to seals6.

The body plan of sperm whales is dominated by a blu-fronted head that, in old males, may constitute up to 1/3 of the body length11. Despite possessing a long, slender and underslung jaw with large conical teeth, there are in general no tooth marks on ingested prey items, and several examples of healthy sperm whales with severely deformed lower jaws led Berzin6 to conclude that there is reduction not only in the function of the teeth, but also that of the entire lower jaw in the digestive process of the sperm whale. In addition, there is a distance of several meters between the tip of the nose and the oropharyngeal opening, meaning that sperm whales must pass well over their prey before they can be ingested, leaving ample opportunity for escape. Thus sperm whales push a large nasal complex ahead of them while trying to capture small, oen agile prey items with a narrow underslung jaw.

The apparent conict between the gigantic nose of sperm whales and their foraging success12 has, in combination with their cryptic deep diving life style, led to a plethora of more or less informed speculations as to how sperm whales may go about subduing between 100 and 500 prey items per day4. Beale1 aired the common belief of whalers that sperm whales use the light coloration of their mouth region to attract prey. That theory was elaborated upon by Clark13 who proposed that foraging sperm whales hang motionless at depth by regulating their

BIOECOMAC. Dept. of Animal Biology, La Laguna University, Spain. Zoophysiology, Department of Bioscience, Aarhus University, Denmark. Scottish Ocean Institute, University of St. Andrews, Scotland. Institute of Biology, University of Southern Denmark, Denmark. CREEM, University of St. Andrews, Scotland. Murdoch University Correspondence and requests for materials should be addressed to A.F. (email: [email protected])

SCIENTIFIC REPORTS

1

www.nature.com/scientificreports/

buoyancy via selective heating and cooling of the spermaceti organ. However, recently Amano and Yoshioko14

have shown that sperm whales are active predators that at times accelerate to speeds of more than 3 m/s during rapid changes in body orientation and sharp turns at depth15. Aoki et al.16 have further shown that such speed bursts are associated with squid ink and body parts passing cameras placed on the back of tagged sperm whales. These results show that sperm whales actively pursue prey during foraging dives. Hence, sperm whales are active predators, but the knowledge of how sperm whales nd and subdue their prey is still sparse. Fristrup and Harbison17 suggested that sperm whales forage actively using vision to identify and locate prey. This visual predation hypothesis predicts that sperm whales may detect their prey by its bioluminescence or should consistently turn their ventral side towards the surface to facilitate stereo vision of silhouetted prey6.

In 1972, Norris and Harvey18 proposed that sperm whales use echolocation to nd and discriminate prey by producing powerful biosonar clicks with their hypertrophied nasal complex. Recent experimental evidence has supported this theory by showing that the nasal complex is indeed a sound generator1922 that can produce highly directional clicks with source levels (SLs) exceeding 235dB re 1Pa (peak-to-peak; pp) (equivalent to an energy ux density (EFD) of approximately 182dB re 1Pa2s), the highest instantaneous sound levels from any known biological sound source23. These source properties and the fact that sperm whales produce clicks throughout their foraging dives strongly suggest that sperm whales use long-range echolocation to nd prey7,2427. Gordon24 also

proposed that echolocation is used during the last seconds of prey acquisition and compared the fast bursts of clicks (also called creaks) recorded from diving sperm whales with buzzes produced by bats in the terminal phase of prey capture28. Miller et al.29 went on to show that sperm whales display an increase in manoeuvring between some 12s before and aer the end of buzzes, with the peak of activity coinciding with the end of the buzzes. This coupling between buzzing and movements lends support to the hypothesis that sperm whale buzzes play a role in active prey acquisition as is the case for beaked whales30, porpoises31 and delphinids32. Furthermore, it was shown that buzzes occurred both while moving up and down in the water column29, suggesting that visual imaging of prey against down-welling light is not essential for sperm whale hunting as also evidenced by the ndings of healthy, but blind sperm whales6.

However, while echolocation can explain how sperm whales nd and track their prey, it does not address how the largest toothed whale is able to catch highly mobile nektonic prey which make up a portion of its diet. Norris and Mhl12 advanced the so-called biological big bang hypothesis suggesting that the powerful usual clicks of sperm whales serve to acoustically debilitate prey and thereby facilitate capture. This hypothesis presents a coherent scenario potentially explaining both the apparent hunting success of these large predators7 and the evolutionary driving force behind the hypertrophied nasal complex. A prerequisite for acoustic debilitation is that the prey is negatively aected by sound pressure levels within the capabilities of the sperm whale sound generator. Zagaeski33 exposed guppies to intense transients from a spark generator and found that 50% of the sh were stunned at received levels (RLs) of 235dB re 1Pa (pp). In an unpublished study from 1982, Norris and Mhl (B. Mhl pers comm) exposed the cephalopods Loligo and Sepia to exploding blasting caps and found that it took RLs of 253 to 266 dB re 1 Pa (pp) to debilitate these species. However, transients from both a spark generator and blasting caps have signicant energy at low frequencies coupled with very fast rise times and therefore little spectral and temporal resemblance to toothed whale echolocation clicks. Indeed, when using ultrasonic transients closely mimicking those of toothed whale echolocation clicks on squid34 and on sh with swim-bladders35,36, RLs up to 226dB re 1Pa (pp) did not impair the swimming behaviour of either species. Hence, acoustic debilitation seems unlikely if sperm whales do not expose their prey to more than 235 dB re 1 Pa (pp) during prey capture. In fact source level estimates of buzz clicks from sperm whales have been estimated to be less than 210dB re Pa (pp)20, so if sperm whale prey are indeed caught at the end of buzzes, their capture cannot be explained by acoustic debilitation.

A detailed understanding of how sperm whales catch prey and a specic test of the dierent and in some cases opposing hypotheses require ne scale data on the acoustic behaviour of foraging sperm whales and simultaneous information about the spatial relationship with their prey. Here we use high resolution sound and movement recording tags on free-ranging sperm whales to quantify their behaviour during prey captures. Our analysis supports previous reports that sperm whales forage actively and conrms that sperm whales do not consistently orient themselves as predicted to optimise the use of visual cues provided by the downwelling light. We also show that sperm whales do not debilitate prey acoustically, but rather employ a fast repetition rate buzz to provide high resolution echolocation sampling during active prey chases. We propose that prey are subdued near the end of the buzz with a rapid acceleration consistent with fast movements of the mandibles and the tongue to grasp and suck in prey.

Results

Foraging behaviour. During approximately 66 h of combined acoustic and movement data, the 6 tagged sperm whales performed a total of 82 foraging dives (Table1). Echolocation periods, evidenced by production of usual clicks, lasted on average 38 (7.6) minutes during deep foraging dives performed by the three whales tagged at low latitudes of the US East coast, amounting to 86% of dive duration. The three sperm whales o northern Norway echolocated on average 26 (7.9) and 38 (7.1) minutes during shallow and deep foraging dives, respectively, corresponding to 93% and 90% of dive duration. Buzzes were produced mostly during the bottom phase of foraging dives (Fig.1, Table1). Sperm whales foraging in the low-latitude habitat made on average 24 (4) feeding buzzes per dive at depths ranging from 922 to 1197 meters. Whales foraging in the high latitude habitat switched between shallow (48217 m) and deeper (2531862 m) foraging layers. They performed fewer buzzes during shallow dives (4 3) compared to deeper dives (15 12). In total the whales produced 674 buzzes that lasted a median of 9.1s (minimum: 2s; lower quartile: 6.2s; upper quartile: 14s; maximum: 114s). Of these, 567 buzzes matched the criteria used here to select independent buzzes (see Methods) and thus were included for further analysis (Table1).

SCIENTIFIC REPORTS

2

www.nature.com/scientificreports/

whale

code year location

time

recording

# foraging

dives

jerk [m/s3] #/buzzes

speed [m/s]*

#/buzzes

201b 2003 NA 2h 2 56/48 880 (2221193) 8 (441) 41 (5569)/48 1.3 (1.31.4)/3 202a 2003 NA 38 1 17/16 787 (596920) 5 (242) 59 (7214)/16 1.8 (1.81.8)/1 206a 2003 NA 3h 47 5 103/89 852 (5211024) 5 (258) 82 (7945)/85 1.9 (1.32.2)/8 196a 2005 NN 24h 17 29 154/122 125 (87492) 14 (7114) 19 (6174)/100 0.6 (0.60.6)/1 199a 2005 NN 20h 42 28 143/109 347 (301600) 12 (494) 47 (51275)/99 2 (1.82.2)/13 199b 2005 NN 16h 46 17 201/182 1343 (401854) 7 (425) 14 (5196)/179 2.2 (1.22.7)/26

Table 1. Tagging information. Whale codes are formed by the Julian day and the deployment order of the tag in that day. Foraging dives are dened as echolocation dives with a maximum depth greater than 25m. Information is given as median (range). NA= North Atlantic; NN=northern Norway. *Calculated using the orientation-corrected depth rate method for buzzes with absolute pitch >45.

# total/analysed

buzzes buzz depth [m]

buzz

duration [s]

Figure 1. (A) Image of sperm whale 199b tagged in 2005 o northern Norway. The gure shows the tag position at the moment of deployment. (B) Section of a dive prole recorded from a male sperm whale o northern Norway. The colour indicates approximate swim speed estimated using a Kalman lter matching pitch angle to depth rate. This estimate loses accuracy at low pitch angles. Each grey circle marks a prey capture attempt (buzz). (C) Roll orientation of the diving sperm whale during the foraging dive. Distribution of the absolute roll angle during the foraging phase of pooled (D) shallow dives, depth <220m (n= 31, 3 whales), and (E) deeper dives (n= 64, 6 whales). Coloured bars denote dierent animals.

Speed during buzzing. As exemplied in Fig.1, tagged sperm whales moved forward at estimated speeds between 1 and 2.5m/s. By design, the Kalman lter used to estimate speed in this gure tends to de-emphasize short bursts of high speeds associated with prey chases. Also, swimming speed is poorly estimated at low pitch angles. Therefore, the more robust method of orientation-corrected vertical speed was applied to 52 buzzes (n= 1 to 26 per whale), all with an absolute pitch angle greater than 45 throughout the buzz. This analysis shows that the whales

SCIENTIFIC REPORTS

3

www.nature.com/scientificreports/

moved forward during prey capture attempts at an average speed of 1.9 (0.4) m/s (representing the mean over all buzzes of the average speed in each buzz), attaining minimum and maximum speeds of 0.6m/s and 2.7m/s (Table1).

Testing for multimodal sensing. Fristrup and Harbinson17 proposed that vision is essential for sperm whale foraging and hypothesized that whales should roll upside down to facilitate stereo vision6 of prey against down-welling light. To test this we computed the absolute roll angle of sperm whales throughout the foraging phase of dives, that is, between the rst and last buzz within each dive7. Sections with high absolute pitch angles (>60) were excluded from this analysis to avoid erroneous roll estimates due to gimbal lock37. The roll analysis was done separately for shallow dives (<220 m, n = 64, 3 whales) and deep dives (n = 31, 6 whales) to test for a preferential use of vision near the surface. The histograms of roll angles showed a bimodal distribution, with peaks around 10 and 110 degrees, for both dive types (Fig.1C,D).

Fast movements near the end of buzzing. The acceleration rate (jerk) was computed in the nominal approach phase and during the buzz to test for transients consistent with prey capture38. Jerk was normalized to the mean jerk during steady swimming (see Methods), and a normalized jerk threshold of 5, derived by the log-frequency distribution of the normalized jerk in approach and buzz phases, was used to detect strong transients. Normalized jerk transients above this threshold were found in 527 buzzes (93% of all buzzes analysed) and occurred in the last ve seconds of buzzes in 70% of cases (Fig.2). Peaks in normalized jerk during buzzes had a mean of 73 (interquartile range, 12 to 72), some 20 times higher than the normalized jerk during steady swimming (Table1). Buzzes with normalized jerk transients and with high absolute pitch angle, for which accurate vertical speed estimation was possible, were included for further analysis (n= 27). For these 27 buzzes, the prey range at the start of each buzz was estimated by back-calculation, using the time delay between the normalized jerk peak to the start of the buzz to estimate the hand-o distance, i.e., the prey range at which whales radically alter the ICI at the onset of buzzes. The mean estimated hand-o distance for these buzzes was 9m (interquartile range, 5 to 14m).

Acoustic output adjustment as a function of range to prey interception. Using the normalized jerk peaks as an indicator of when prey are intercepted during a buzz, we quantied the ICI and the apparent output level (AOL) as a function of time, and therefore range, prior to target interception (Fig.3BD) for steep buzzes in which speed estimation was reliable. Although ICI and AOL changed continuously through the approach and buzz, there was a major transition at the switchover from usual to buzz clicking. This was marked by an order of magnitude reduction in ICI from about 0.5s to 20ms (Fig.3B). The impact of this on the potential acoustic eld of view39 can be visualized by comparing the ICI with the estimated two-way travel time (TWT) from whale to prey (Fig.3B). While the 0.5s ICIs before the buzz leave a long lag time between the reception of an echo and the emission of the following click, the fast ICIs in the buzz are much closer to the TWT of the target.

During the usual click approach phase the tagged whales showed a bimodal pattern in ICI: the low-latitude whales produced stable ICIs up until the start of the buzz (Fig.4A), whereas the sperm whales tagged o northern Norway tended to continuously decrease ICIs in this nominal approach phase (Fig.4D). During buzzing all tagged whales continuously decreased ICI (Fig.4B,E) with the exception of occasional uctuations (e.g., Fig.2D) which presumably track prey movements. As most usual clicks were clipped in the recording, an accurate measure of the AOL reduction from usual clicking to buzzing is not possible but a reduction of 12 orders of magnitude in peak pressure is conservative (i.e., a drop in level of 2040dB compared to the minimum outputs during the approach phase) (Figs3B and 4). Within the buzz, where signal levels are well below clipping, there is a gradual decrease of AOL with time amounting to a roughly additional 10 dB level reduction over the course of the buzz (Fig.4C,F).

Sound exposures of prey. The sound exposure of prey to regular clicks was estimated by combining the median hand-o distance of 9m with the maximum 235dB re 1Pa (pp) source level estimates of sperm whales (EFD: 182dB re 1Pa2s)23. Subtracting the approximately 20dB of transmission loss over the 9meters, the maximal sound exposure of prey just before the buzz will be no higher than some 215dB re 1Pa (pp) (EFD: 161dB re 1Pa2s). When sperm whales switch to a buzz they increase click rates and lower source levels by some 2 orders of magnitude20. The 2040dB reductions in source levels from the maximum of 235dB re 1Pa (pp) mean that prey exposures during the close encounters of the buzz are likely to be well below 200dB re 1 Pa (pp) (EFD: 146dB re 1Pa2s).

That estimate is supported by a fortuitous situation in one tag deployments where a Dtag aer detachment from a whale at 630 meters depth was approached by the whale while buzzing. This ensonication provides a unique example of the sound eld that may be received by a prey as it is approached by an echolocating sperm whale at depth (Fig.5). The RLs of the buzz clicks in the tag varied as the whale scanned its directional sound beam across it, but never exceeded 187dB re 1Pa (pp) (EFD: 133 dB re 1Pa2s) (Fig.5). The received buzz clicks, which were not clipped due to the lower SL during buzzes, showed the distinctive monopulse nature of sperm whale on-axis clicks (Fig.5B), with a duration of ca. 100s (Fig.5C), a broad bandwidth and a centroid frequency >25kHz (Fig.5D). The 46kHz recording bandwidth was clearly insufficient to sample the full frequency range of buzz clicks. The buzz followed the characteristic drop in ICI to 1020ms observed in other buzzes (Fig.5A). The whale bumped into the tag 5seconds before the end of the 20s long buzz (Fig.5A), consistent with the timing of jerk transients during buzzes recorded by tags on whales attempting to capture prey. In order to investigate the inuence of the limited bandwidth of our recordings, we simulated a signal with a bandwidth of 96 kHz (sampling rate 192kHz), using a spectrum that was close to identical to the spectrum of the most broadband signal in our recordings, but with the gentle high frequency roll o extrapolated so as to continue up to 96kHz. We then

SCIENTIFIC REPORTS

4

www.nature.com/scientificreports/

Figure 2. (A) Normalized jerk levels recorded in the 30seconds prior to the end of each buzz (n= 154) of whale sw05_196a. Jerk was normalized to steady swimming. Values during steady swimming are represented in white. Values above the threshold set at 5, as indicated by the red line in (C), represent abnormally fast movements.(B) Normalized jerk relative to the end of buzz, pooling all buzzes from all whales with duration <21s and time between consecutive buzzes (IBI) >16 s (n = 567) (see Methods for the denition of these thresholds).(C) Individual-weighted average of normalized jerk (black line) for buzzes shown in B. The upper and lower limits of the grey coloured area represent the maximum and minimum values of the individual averages, respectively. The red line denotes the 5 threshold used to detect a jerk transient. (D) Inter-click interval and normalized jerk signature of an echolocating sperm whale during a prey pursuit. The panel shows data from 10seconds before the onset of the buzz and until 2seconds aer the end. Clicking alternates between faster and slower rates, accompanied by high jerk transients. The sperm whale may be tracking acoustically the moving prey by increasing its ICIs when prey escapes, and occasionally attempts to capture prey (as indicated by the high normalized jerk) when close enough.

evaluated the eect on the p-p amplitude measurements of this simulated signal by reducing the bandwidth again to 48kHz, showing that the eect on the RL was below 1dB.

Discussion

In the last decades, the development of new tools to study the sounds and movements of free-ranging echolocating whales has extended our knowledge about sound production in sperm whales21,22,40,41, the acoustic properties of their clicks23,42,43, their echolocation behaviour4448 and movements during foraging1416,29,49. However, information about how the largest tooth-bearing predator in the world tracks and captures small and agile prey is still scarce. A number of hypotheses have been advanced, but they remain largely untested due to the difficulties in sampling ne-scale predator-prey interactions in the deep sea. To address this data gap we used multi-sensor tags

SCIENTIFIC REPORTS

5

www.nature.com/scientificreports/

Figure 3. (A) Sperm whale shallow foraging dive prole with four prey capture attempts, i.e. buzzes, indicated by red circles. The black line indicates the echolocation phase of the dive. (BD) represent dierent parameters from the start of the buzz to the peak in normalized jerk, taken as a proxy for prey interception. (B) A detailed view of the click rate during the nal buzz (encircled in black in (A)). ICIs, from 10s before the buzz to theend of the buzz, are colour coded by their relative AOL (dBpp re max pp). The red line marks the estimated two-way-travel time (TWT) from the sound source to the presumed prey capture location calculated from the forward speed of the whale. Note the log scale on the TWT axis. (C) Normalized jerk signature during the same period and (D) distance between the front of the whale and the presumed prey interception location, derived from the forward swimming speed of the whale, calculated using the orientation-corrected depth rate method.

to obtain a high-resolution picture of the movement and acoustic behaviour of free-ranging sperm whales during prey capture. This unique data set allows us to address the three hypotheses for how sperm whales hunt and capture their prey: the visual predation hypothesis17, the biological big bang hypothesis12 and the hypothesis of suction feeding50. To do so, we rely on a well-established indication of when whales are attempting to capture prey in the form of echolocation buzzes. Strong evidence for the role of buzzes in close prey approaches has been found in beaked whales where echoes from prey can be monitored30 and the consistent connection between buzzes and movement in sperm whales29, beaked whales51, porpoises31 and dolphins32 suggest that they serve the same role in other echolocating species. Using the end of the buzz as a rst proxy for when whales have completed or abandoned the capture, we examine how tagged whales move in the preceding seconds. In the following we will use these new data to critically evaluate the three hypotheses for how sperm whales hunt and capture their prey.

The visual predation hypothesis. Animals will likely make use of all sensory cues available to improve the efficiency of their activities52. Fristrup and Harbinson17 proposed that vision is central to sperm whale foraging and predicted a ventral-upward posture of sperm whales in order for them to use stereo-vision to search for silhouetted prey against a bright background. Such visual predation should be more prominent in well-lit shallower waters12. Our data show that tagged whales did not consistently swim upside down to spot and track prey silhouetted against the lit surface waters using stereo vision (Fig.1BD) augmenting the study by Miller et al.29 of downward and upward buzz directions. Instead, they spent a large portion of their foraging time rolled to one side which would enable only monocular vision of prey in down-welling light. Although sperm whales may be complementing echolocation with monocular vision, the fact that the whales use similar rolling behaviour in shallow and deep dives despite the very large dierence in light levels, suggests that this behaviour has

SCIENTIFIC REPORTS

6

www.nature.com/scientificreports/

Figure 4. ICIs of analysed buzzes as a function of time to the end of buzz for (A) sperm whales in theNorth Atlantic (NA; n= 154 buzzes), and (D) sperm whales in northern Norway (NN; n= 413 buzzes). The histograms on the right show the ICI distribution in the approach phase. Expanded view of the buzz phases in (B) North Atlantic and (E) northern Norway. Relative AOL (dBpp re max pp) as a function of time to the end of buzz for (C) North Atlantic sperm whales, and (F) sperm whales o northern Norway. Each colour represents a dierent animal.

little to do with vision. Rather this nding may indicate that sperm whales have preferred approach angles for echo-guided prey capture. Specic body positions may favour the orientation of the whales sonar beam towards prey when approaching it, or may be linked to feeding mechanisms and/or to prey behaviour. Similar behaviour has been observed in some beaked whales, where specic approach angles and body orientations are adopted to capture particular prey53.

Fast movements indicate strikes at prey. Many toothed whales use suction feeding to acquire prey54,55, and it has recently been shown that porpoises31, dolphins32 and beaked whales30 all subdue their prey by suction and/or raptorial feeding in the last part of the buzz. These feeding events require rapid movements in head and neck musculature which can be detected by accelerometers mounted near the head. As done in studies with non-echolocating marine mammals37,56, we used the dierential of acceleration or jerk as a processing step to emphasise these sudden movements. Here we show that strong jerk transients are recorded consistently during buzzes by tags attached to the crest of the skull of sperm whales, and that they are seldom recorded outside of buzzes.

Caldwell et al.50 suggested that sperm whales ingest prey via a sucking motion of the tongue and gular region. That hypothesis has gained anatomical support by Werth57, who concluded that the morphology and relative positions of the tongue and the very large and exible hyoid are indicative of suction feeding via an orice formed by the oropharyngeal isthmus. In the current study, strong transients of jerk were found around 5seconds before the end of the majority of buzzes (Fig.2). As the dierentiation used in calculating jerk de-emphasises relatively slow manoeuvres and swimming motions37,58, jerk transients are most likely an indication of rapid movements in the cranial and mouth region. Although acceleration signals associated with sound production have been shown recently in n whales59, it is unlikely that the jerks found here are caused by clicks because jerk peaks occur late in the buzz, when acoustic output levels recorded by the tag are the lowest. Rather, the consistent presence of jerk transients near the end of buzzing, irrespective of buzz duration (Fig.2A), strongly supports the interpretation that these are caused by rapid movements in the gular region during strikes at prey. We cannot infer from the jerk whether the strikes involve suction generation or raptorial jaw movements or both, although evidence of intact prey in stomach contents6 and the functional anatomy of the gular apparatus57 lend support to the former. Either way, if our interpretation is correct, the jerk transients do provide a strong indication of when prey are intercepted during the buzz phase

SCIENTIFIC REPORTS

7

www.nature.com/scientificreports/

Figure 5. Data from a free-oating Dtag ensonied by a buzzing sperm whale that eventually bumps into the tag. (A) Absolute sound pressure level recorded by the tag. (B) The tag is ensonied at normal buzz rate clicking, and the clicks show the distinct monopulse nature of sperm whale on-axis clicks21, but the RLs never exceed 187dB re 1Pa (pp). (C) The waveform and (D) power spectral density (SPD) of the 5 most powerful clicks received on the tag (grey lines) and the mean overlaid (black line).

Acoustic output adjustments to prey range. The transition from usual clicking to buzzing in the tagged sperm whales marks a dramatic gaze adjustment (Fig.3B, refs 25 and 26). The tagged sperm whales reduced their ICIs by more than an order of magnitude when switching from usual to buzz clicking (Fig.4A,B). Simultaneously they reduced the output level of their clicks by 2040dB (Fig.4C,F). This echolocation behaviour is very similar to that described for smaller toothed whales, which increase click rates and reduce output levels in the terminal phase of prey capture31,39,60. A high click rate enhances the information ow on prey location to guide motor-patterns during close range manoeuvres, but at the potential expense of range ambiguity and spatial aliasing from distant targets61. A way to reduce the distance over which echoes can cause range ambiguity at high click rates, and thus to simplify the acoustic scene for perceptual organization, is to reduce the SLs. It is also possible that reduced SLs when clicking rapidly could relate to pneumatic restrictions of the sound generator of toothed whales62,63, but this does not alter the potential value of low SLs for close tracking of prey.

At a mean estimated distance of 9 meters to the prey, the tagged sperm whales switched from the approach to the buzz phase. This hand-o distance is about a body length of a female sperm whale and signies the range from the prey to the sound source at the anterior part of the nasal complex, which is an additional 12 meters anterior of the oropharyngeal opening. A similar hand-o distance of 0.5 to 1 body lengths has been found in beaked whales60 and porpoises31. Hence, hand-o distances between the approach and buzz phase in echolocating toothed whales seem to be scaled with the size and possibly manoeuvrability of the predator53. Longer hand-o distances imply longer initial ICIs in buzzes, which is consistent with the overall click rate during buzzing that is about one order of magnitude slower in sperm whales compared to smaller toothed whales (Figs3B and 4AE; refs 31 and 61). This in turn implies less frequent updates of prey location in buzzing sperm whales. Given the similar forward speeds of dierent species of toothed whales during foraging (Fig.3D; refs 14, 64 and 65), the clicking rate in buzzes may thus be more related to the absolute manoeuvrability of the echolocating whale66,67

than its speed53.

SCIENTIFIC REPORTS

8

www.nature.com/scientificreports/

A hand-o distance of about 9meters corresponds to a TWT to the targeted prey of 12ms which is fairly close to the initial ICIs in buzzes (Fig.3B). Our hand-o distance estimates are approximations as they assume that prey do not move when the buzz is initiated. When this assumption is not realized, it may result in an estimated TWT that is greater than the ICI as in the example of the 20-second long buzz in Fig.3. It is more likely that sperm whales, as with beaked whales and bats, indeed use buzz ICIs that are greater than the TWT, but that the prey of the buzz in Fig.3 starts to move away from the buzzing whales at some point during the buzz, prolonging the time from buzz start to terminal jerk at the end of the buzz. Pursuits of moving prey result in prolonged buzzes in both bats68 and toothed whales8. This could help explain the occasional long buzzes observed here, which can last up to two minutes being some 10 times longer than the normal short buzzes of sperm whales foraging in deeper waters (Fig.2). In some of these longer buzzes, the clicking rate cycled several times between faster and slower modes, accompanied by strong jerk transients, suggesting that the whale is tracking the changing range to an elusive prey which it attempts to capture several times (Fig.2D). Hence, sperm whales, like smaller toothed whales, target evasive prey that at times require long-distance pursuits (Fig.2D8) and, occasionally, an increase in speed1416 to be captured.

The ICIs of usual clicks during the approach phase were much longer than the estimated TWT at the start of buzzes (Fig.4A,D), suggesting that sperm whales, like beaked whales, maintain a large sensory volume even aer they have detected and selected a target. This likely facilitates tracking of several prey items for sequential capture53,69,

supporting recent suggestions that sperm whales are able to organize perceptually a multi-target auditory scene48.

The tagged sperm whales showed a bimodal clicking behaviour during the approach phase, where ICIs either remained stable (Fig.4A) or displayed a gradual reduction (Fig.4D) as prey were approached. Free-ranging dolphins have been shown to gradually decrease ICI as they close in on a target during the approach phase70.

Conversely, beaked whales studied in the wild typically approach individual prey without such range locking in the approach phase53,69. The fact that sperm whales do not consistently conform to either of these modes suggests that clicking rate in the approach phase is possibly not species-specic, but could be dynamically adapted to dierent prey types or prey distribution found in specic habitats, which in turn may aect the echoic scene that must be processed by the echolocating toothed whales. Stable ICIs in the approach phase may be a way to deal with a complex auditory scene, calling for auditory stream segregation via slow sampling rates69. Conversely, if the whale targets a larger or more active prey in a simpler auditory scene, it may be able to apply range locking as it approaches it.

Sound exposures of prey. The so-called biological big bang hypothesis predicts the use of acoustic debilitation to facilitate prey capture12. As found by Aoki et al.15 for sperm whales in the north Pacic, whales in our study occasionally swam at relatively high speeds during the foraging phase of dives. These bursts of speed were inferred in15 to be chases aer active prey although the lack of acoustic data in that study meant that the connection between movement and echolocation phase could not be established. Acoustic data from the tags used in our study conrm that the fast swimming occurs just before and during buzzes and co-occurs with highly dynamic ICI patterns in buzzes. This strongly supports the interpretation that whales are chasing prey over distances of 10s of meters which is inconsistent with prey debilitation that should involve eortless interception of immobile prey. Acoustic debilitation also implies that sperm whales should seek to maximize sound exposure of their prey. With a median hand-o distance of 9meters from the approach to the buzz phase at which time the click outputs are reduced by two orders of magnitude, we show that sperm whales do not do that. Rather we estimate that prey items during the approach phase are unlikely to be exposed to more than 215dB Re 1Pa (pp), and that the low source levels in the buzz phase results in prey sound exposures well below 200dB re Pa (pp) (Fig.5). Those observations are also supported by a maximum received level of 187dB re 1Pa (pp) on a free-oating tag in front of a buzzing sperm whale (Fig.5). Clearly this observation has to be evaluated with caution, given that it is derived from a single buzz from a single animal. Nevertheless, the back-calculated SLs from Fig.5 and the 187dB re 1Pa (pp) exposure are consistent with independent buzz source level estimates from array recordings20. Furthermore, sperm whales have been reported to ingest a variety of non-food items71, and porpoises have been shown to use the same echolocation behaviour when targeting prey31 as when homing in on articial targets39. These observations in concert lend weight to the notion that prey targeted by sperm whales are likely not exposed to more than 190dB re 1Pa (pp) during buzzing. These estimated RLs are at or below levels shown to have no eect on tested cephalopod or sh prey species with gas-bladders, which would likely be more prone to sound induced damage3436.

Hence, it appears that sperm whales do not attempt to maximize the sound levels imping on their prey items. Instead they dramatically reduce source levels when they are within a body length of their prey resulting in exposure levels that are 1 to 2 orders of magnitude below those required to debilitate prey. The short, low-energy, broad-band and high-frequency clicks (Fig.5) at high repetition rates in a buzz are therefore quite unsuited for prey debilitation, but rather well suited to provide both fast updates on prey location and to provide clutter reduction during the last few meters of the terminal prey capture phase as would be required to eectively track and capture evasive prey. In bats, buzz calls are of short duration and lower frequencies than search and approach calls, resulting in broad sound beams providing a wide eld of view during close range target interception72.

Interestingly, the higher centroid frequencies of sperm whale buzz clicks as compared to usual clicks would suggest an increase rather than decrease in directionality during buzzing, which is puzzling given the already high directionality of sperm whale usual clicks23. It may be that conformation changes in the sound producing nasal structures can oset the frequency eect by increasing the functional aperture, but clarication of this issue must await beam measurements of buzzing sperm whales.

SCIENTIFIC REPORTS

9

www.nature.com/scientificreports/

Conclusion

Here we have addressed several foraging hypotheses for sperm whales using high resolution on-animal measurements to uncover the movement and echolocation behaviour before and during prey capture. Sperm whales produce the most powerful biological sounds in the animal kingdom in order to echolocate prey at long ranges, but reduce acoustic outputs by several orders of magnitude when they are about their own body length from their prey. This behaviour is inconsistent with the big bang hypothesis indicating that the extreme nasal complex of sperm whales did not evolve to debilitate prey with sound, but rather to produce powerful clicks for long range echolocation of prey. There is likewise no indication that visual predation plays a signicant role in sperm whale foraging. Instead, prey capture attempts by sperm whales are remarkably similar to those of other echolocating toothed whales comprising a sudden change in click rate and output level when the whales are about a body length from their prey. High repetition rate buzzes at low output levels provide high temporal and spatial resolution tracking to inform motor patterns in the last few meters before capture. While some prey targeted by sperm whales appear to be easy to subdue, others, oen at shallower depths, require lengthy chases, leading to prolonged buzzes and higher than average swimming speeds. Strong and sudden changes in acceleration near the end of buzzes are supportive of the notion that sperm whales employ suction feeding to engulf their prey57 and continue buzzing during post-acquisition prey handling. Nonetheless, it still remains a conundrum how the sperm whale can keep track of a prey item during the nal meters from when the prey drops below the sonar beam and until it reaches the buccal cavity, and we still do not understand the details of how the prey is engulfed. Hopefully, advances in animal-attached cameras will help answer these questions in the future.

Materials and Methods

Tagging. Field work was performed o the Eastern US seaboard (38N, 70W) where three sperm whales were tagged in July 2003, and in the Bleik canyon (6925N, 1545E), approximately 30km W of Andya, northern Norway, where three sperm whales were tagged in July 2005. Tagging was conducted under permits 369-1440-01, 981-1578 and 981-1707 from National Marine Fisheries Service to P. Tyack, and permit # 2005/7720-1 from the Norwegian Forsksdyrsutvalg to P. T. Madsen. The study was performed in accordance with Woods Hole Oceanographic Institution, and as such, tagging methodology was approved by the Woods Hole Oceanographic Institution animal use and care committee.

Sperm whales tagged in the North Atlantic may have been adult females or immature males, both similar in size, whereas all three whales tagged o northern Norway were adult male sperm whales. Whales were tracked acoustically to predict their surfacing locations; once at the surface, whales were approached slowly from behind, and a Dtag (version 2) was placed using a 15m cantilevered pole mounted on a rigid-hulled inatable boat. All tags were placed in the neck region to ensure clear recordings of clicks and to maximize the chances of detecting acceleration from muscle movements associated with feeding (Fig.1A). Tags were attached with a square array of 4 suction cups to minimise relative movement between the tag and the body. The Dtag is an archival device that samples 3-axis accelerometers, 3-axis magnetometers and pressure at 50Hz37. Sound was recorded simultaneously with sensor sampling from one (2003) or two hydrophones (2005) at 96kHz (16-bit). Aer the release of the tag from the whale, a VHF radio beacon in the Dtag aided tracking and recovery of the tag.

Data analysis. Tag data were analysed using custom soware in Matlab 7.5 (Mathworks). To exclude shallow submersions from further analysis, a foraging dive was dened as being deeper than 25m and containing echo-location clicks. Foraging dives less than 220m deep were considered shallow dives following Fais et al.48. Sound recordings were displayed as scrolling spectrograms (512 point FFT, Hann window, 50% overlap) to identify the start and end of echolocation clicking in each dive, along with the approximate time of buzzes. Clicks from tagged whales were separated from those of nearby clicking animals based on the consistently high RL of the former. Following Teloni et al.8, the precise start time of a buzz was dened as the rst click in which the inter-click interval (ICI) dropped below 0.22 s. Conversely, the rst click interval longer than 0.22 s was taken as indicating the end of a buzz. Within each dive, the time between the end of one buzz or buzz event and the start of the next was dened as the inter-buzz interval (IBI). Sequences of buzzes in close succession have been interpreted as capture attempts on the same target or a tight school of targets in other toothed whales73. Buzzes interrupted by air recycling pauses or by a brief increase in ICI (below buzz threshold) were, therefore, assigned as capture attempts of the same target and thus dened as a single buzz event, reaching durations up to two minutes (Fig.2D). To avoid treating related buzz sequences as independent events, long buzz events (>21 seconds) comprised of dierent buzz sequences and buzzes spaced closely (<16seconds apart) were excluded from further analyses. These criteria were chosen from log-frequency distribution plots of buzz length and IBI.

The emission time of each individual click was determined by a supervised click detector adjusted to the frequencies and ICIs previously reported for sperm whales23,26. The RL of the clicks at the tag, a proxy for the relative SL termed the apparent output level (AOL)69, was computed from the peak-to-peak amplitude expressed as dB re maximum pp26. In absence of information about when individual prey were detected and selected for capture, the approach phase to the prey28,61 was conservatively dened as the ten second time window preceding the transition into the buzz53.

Sensor data were decimated to 5 Hz and corrected for the tag orientation on the whale37. The orientation of the whale, represented by the Euler angles, pitch, roll and compass heading, was then estimated using standard procedures37. Swim speeds prior to, and during, buzzes were estimated from the depth rate and pitch angle of the whale sensu Miller et al.74). The accuracy of this orientation-corrected depth rate method decreases with declining absolute pitch angle and so the method was only used during buzzes in which the absolute pitch was consistently >45. To visualize the approximate forward speed throughout the dive (Fig.1), a two-state Kalman lter matching pitch angle to depth rate was used.

SCIENTIFIC REPORTS

10

www.nature.com/scientificreports/

In order to detect changes in acceleration associated with prey capture and deglutition, the three accelerometer axes were dierentiated and the norm of the resulting triaxial jerk was computed at each sample38,51,58. The

maximum range of the accelerometer was 2 g, and for this analysis sensor data were used at the full sampling rate of 50 Hz. In each dive the norm-jerk during buzzing was normalized by dividing by the RMS norm-jerk measured during seconds 10 to 40 of the descent phase, an interval that contained only steady swimming in all animals (normalized jerk coefficient [m/s3*(m/s3)1]). The rst 10s of the descents were excluded to avoid strong uke strokes. This normalization procedure compensates to some extent for dierent tag locations and was used for all jerk analyses. The normalized jerk signature within each buzz was inspected visually and the maximum peak in each buzz was identied. Assuming that the normalized jerk peak represents the moment of prey interception and so the time at which the whale-prey distance is zero, the target range as a function of time before the jerk peak was estimated for buzzes performed at high absolute pitch angles (>45 degrees) by integrating the estimated speed. This analysis assumes that the prey are stationary or move little in comparison with the whale during the approach. This may over-estimate the distance to prey that attempt to escape early in the buzz.

References

1. Beale, T. The Natural History of Sperm Whales. (John Van Voorst, 1839).2. Lavery, T. J. et al. Iron defecation by sperm whales stimulates carbon export in the Southern Ocean. Proc R Soc Lon 277, 35273531 (2010).

3. Clarke, M. R. Beaks, nets and numbers. Symp Zool Soc Lon 38, 89126 (1977).4. Clarke, M. R. Cephalopoda in the diet of sperm whales of the Southern Hemisphere and their bearing on sperm whale biology. Discovery Reports 37, 1324 (1980).

5. Kawakami, T. A review of sperm whale food. Sci Rep Whales Res Inst 32, 199218 (1980).6. Berzin, A. A. The Sperm Whale (Israel Program for Scientic Translations 1972).7. Watwood, S., Miller, P. O. J., Johnson, M. P., Madsen, P. T. & Tyack, P. L. Deep-diving foraging behaviour of sperm whales (Physeter macrocephalus). J Anim Ecol 75, 826835 (2006).

8. Teloni, V., Johnson, M. P., Miller, P. J. O. & Madsen, P. T. Shallow food for deep divers: Dynamic foraging behavior of male sperm whales in a high latitude habitat. J Exp Mar Biol Ecol 354, 119131 (2008).

9. Clarke, M., Martins, H. & Pascoe, P. The diet of sperm whales (Physeter macrocephalus Linnaeus 1758) o the Azores. Proc R Soc Lon 339, 6782 (1993).

10. Best, P. Food and feeding of sperm whales Physeter macrocephalus o the west coast of South Africa. S Afr Jour Mar Sci 21, 393413 (1999).

11. Nishiwaki, M., Ohsumi, S. & Maeda, Y. Change of form in the sperm whale accompanied with growth. Sci Rep Whales Res Inst 17, l17 (1963).

12. Norris, K. S. & Mhl, B. Can odontocetes debilitate prey with sound? Am Nat 122, 85103 (1983).13. Clarke, M. R. The function of the spermaceti organ of the sperm whale. Nature 228, 873874 (1970).14. Amano, M. & Yoshioka, M. Sperm whale diving behavior monitored using a suction-cup-attached TDR tag. Mar Ecol Prog Ser 258,

291295 (2003).

15. Aoki, K. et al. Active hunting by deep-diving sperm whales: 3D dive proles and maneuvers during bursts of speed. Mar Ecol Prog Ser 444, 289301 (2012).

16. Aoki, K. et al. Visual and behavioural evidence indicates active hunting by sperm whales. Mar Ecol Prog Ser 523, 233241 (2015).17. Fristrup, K. M. & Harbison, G. R. How do sperm whales catch squids? Mar Mam Sci 18, 4254 (2002).18. Norris, K. S. & Harvey, G. W. A theory for the function of the spermaceti organ of the sperm whale. NASA SP 262, 397416 (1972).19. Mhl, B. Sound transmission in the nose of the sperm whale Physeter catodon. A post mortem study. J Comp Physiol 187, 335340 (2001).

20. Madsen, P. T., Wahlberg, M. & Mhl, B. Male sperm whale (Physeter macrocephalus) acoustics in a high-latitude habitat: implications for echolocation and communication. Behav Ecol Sociobiol 53, 3141 (2002b).

21. Zimmer, W. M. X., Madsen, P. T., Teloni, V., Johnson, M. & Tyack, P. L. O-axis eects on the multipulse structure of sperm whale usual clicks with implications for sound production. J Acoust Soc Am 118, 33373345 (2005a).

22. Zimmer, W. M. X., Tyack, P. L., Johnson, M. & Madsen, P. T. Three-dimensional beam pattern of regular sperm whale clicks conrms bent-horn hypothesis. J Acoust Soc Am 117, 14731485 (2005b).

23. Mhl, B., Wahlberg, M., Madsen, P. T., Heerfordt, A. & Lund, A. The monopulsed nature of sperm whale clicks. J Acoust Soc Am 114, 11431154 (2003a).

24. Gordon, J. C. The behaviour and ecology of sperm whales o Sri Lanka. PhD thesis, University of Cambridge, UK (1987).25. Jaquet, N., Dawson, S. & Douglas, L. Vocal behavior of male sperm whales: why do they click? J Acoust Soc Am 109, 22542259 (2001).

26. Madsen, P. T. P. R., Kristiansen, N. U., Wahlberg, M., Kerr, I. & Mhl, B. Sperm whale sound production studied with ultrasound time/depth-recording tags. J Exp Biol 205, 18991906 (2002a).

27. Madsen, P. T. et al. Clicking for calamari: toothed whales can echolocate squid Loligo pealeii. Aquatic Biol. 1, 141150 (2007).28. Griffin, D. R. Listening in the Dark: The Acoustic Orientation of Bats and Men (Cornell University Press 1958).29. Miller, P. J. O., Johnson, M. P. & Tyack, P. L. Sperm whale behavior indicates the use of rapid echolocation click buzzes creaks in prey capture. Proc R Soc Lond 271, 22392247 (2004b).

30. Johnson, M., Madsen, P. T., Zimmer, W. M. X., Aguilar Soto, N. & Tyack, P. Beaked whales echolocate on prey. Biol Letters 271, S383S386 (2004).

31. DeRuiter, S. L. et al. Acoustic behaviour of echolocating porpoises during prey capture. J Exp Biol 212, 31007 (2009).32. Wisniewska, D. M., Johnson, M., Nachtigall, P. E. & Madsen, P. T. Buzzing during biosonar-based interception of prey in the delphinids Tursiops truncatus and Pseudorca crassidens. J Exp Biol 217, 42794282 (2014).

33. Zagaeski, M. Some observations on the prey stunning hypothesis. Mar Mam Sci 3, 275279 (1987).34. Wilson, M., Hanlon, R. T., Tyack, P. L. & Madsen, P. T. Intense ultrasonic clicks from echolocating toothed whales do not elicit anti-predator responses or debilitate the squid Loligo pealeii. Biol Letters 3, 225227 (2007).

35. Benoit-Bird, K. J., Au, W. W. L. & Kastelein, R. Testing the odontocete acoustic prey debilitation hypothesis: No stunning results. J Acoust Soc Am 120, 11181123 (2006).

36. Schack, H. B., Malte, H. & Madsen, P. T. The response of Atlantic cod (Gadus morhua) to ultrasound-emitting predators: stress, behavioural changes or debilitation? J Exp Biol 211, 20792086 (2008).

37. Johnson, M. P. & Tyack, P. L. A digital acoustic recording tag for measuring the response of wild marine mammals to sound. IEEE J Oceanic Eng 28, 312 (2003).

38. Ydesen, K. S. et al. What a jerk: prey engulfment revealed by high-rate, super-cranial accelerometry on a harbour seal (Phoca vitulina). J Exp Biol 217, 22392243 (2014).

SCIENTIFIC REPORTS

11

www.nature.com/scientificreports/

39. Wisniewska, D. M., Johnson, M., Beedholm, K., Wahlberg, M. & Madsen, P. T. Acoustic gaze adjustments during active target selection in echolocating porpoises. J Exp Biol 215, 43584373 (2012).

40. Mhl, B. et al. Sound transmission in the spermaceti complex of a recently expired sperm whale calf. Acoust Res Lett Online 4, 1924 (2003b).

41. Wahlberg, M., Frantzis, A., Alexiadou, P., Madsen, P. T. & Mhl, B. Click production during breathing in a sperm whale (Physeter macrocephalus). J Acoust Soc Am 118, 34043407 (2005).

42. Thode, A., Melling, D. K., Stienessen, S., Martinez, A. & Mullin, K. Depth-dependent acoustic features of diving sperm whales (Physeter macrocephalus) in the Gulf of Mexico. J Acoust Soc Am 112, 308321 (2002).

43. Nosal, E. M. & Frazer, L. N. Sperm whale three-dimensional track, swim orientation, beam pattern, and click levels observed on bottom-mounted hydrophonesa. J Acoust Soc Am 122, 19691978 (2007).

44. Wahlberg, M. The acoustic behaviour of diving sperm whales observed with a hydrophone array. J Exp Mar Biol Ecol 281, 5362 (2002).

45. Zimmer, W. M. X., Johnson, M. P., DAmico, A. & Tyack, P. L. Combining data from a multisensor tag and passive sonar to determine the diving behavior of a sperm whale (Physeter macrocephalus). IEEE J Oceanic Eng 28, 1328 (2003).

46. Mathias, D. et al. Acoustic and diving behavior of sperm whales (Physeter macrocephalus) during natural and depredation foraging in the Gulf of Alaska. J Acoust Soc Am 132, 518532 (2012).

47. Oliveira, C., Wahlberg, M., Johnson, M., Miller, P. J. & Madsen, P. T. The function of male sperm whale slow clicks in a high latitude habitat: Communication, echolocation, or prey debilitation? J Acoust Soc Am 133, 313315 (2013).

48. Fais, A. et al. Sperm whale echolocation behaviour reveals a directed, prior-based search strategy informed by prey distribution. Behav Ecol Sociobiol 69, 663674 (2015).

49. Thode, A. Three-dimensional passive acoustic tracking of sperm whales (Physeter macrocephalus) in ray-refracting environments. J Acoust Soc Am 118, 35753584 (2005).

50. Caldwell, D. K., Caldwell, M. C. & Rice, D. W. Behavior of the sperm whale Physeter catodon L. In Whales, Dolphins and Porpoises (ed. Norris) 677717 (University of California Press, Berkeley 1966).

51. Aguilar Soto, N. et al. No shallow talk: Cryptic strategy in the vocal communication of Blainvilles beaked whales. Mar Mam Sci 28, E75E92 (2012).

52. Dangles, O., Irschick, D., Chittka, L. & Casas, J. Variability in sensory ecology: expanding the bridge between physiology and evolutionary biology. Q Rev Biology 84, 5174 (2009).

53. Madsen, P. T., Aguilar Soto, N., Arranz, P. & Johnson, M. Echolocation in Blainvilles beaked whales (Mesoplodon densirostris). J Comp Physiol A 199, 451469 (2013).

54. Bloodworth, B. & Marshall, C. D. Feeding kinematics of Kogia and Tursiops (Odontoceti: Cetacea): characterization of suction and ram feeding. J Exp Biol 208, 37213730 (2005).

55. Werth, A. J. Mandibular and dental variation and the evolution of suction feeding in odontoceti. J Mammal 87, 579588 (2006).56. Skinner, J. P., Norberg, S. E. & Andrews, R. D. Head striking during sh capture attempts by Steller sea lions and the potential for using head surge acceleration to predict feeding behavior. Endanger Species Res 10, 6169 (2009).

57. Werth, A. J. Functional morphology of the sperm whale (Physeter macrocephalus) tongue, with reference to suction feeding. Aquatic Mammals 30, 405418 (2004).

58. Simon, M., Johnson, M. & Madsen, P. T. Keeping momentum with a mouthful of water: behavior and kinematics of humpback whale lunge feeding. J Exp Biol 215, 37863798 (2012).

59. Goldbogen, J. A. et al. Using accelerometers to determine the calling behavior of tagged baleen whales. J Expl Biol 217, 24492455 (2014).

60. Johnson, M., Madsen, P. T., Zimmer, W. M. X., Aguilar Soto, N. & Tyack, P. L. Foraging Blainvilles beaked whales (Mesoplodon densirostris) produce distinct click types matched to dierent phases of echolocation. J Exp Biol 209, 50385050 (2006).

61. Madsen, P. T. & Surlykke, A. Functional convergence in bat and toothed whale. Physiology 28, 276283 (2013).62. Au, W. W. L. & Benoit-Bird, K. J. Automatic gain control in the echolocation system of dolphins. Nature 423, 861863 (2003).63. Beedholm, K. & Miller, L. A. Automatic gain control in harbor porpoises (Phocoena phocoena)? Central versus peripheral mechanisms. Aquatic Mammals 33, 69 (2007).

64. Sato, K. et al. Stroke frequency, but not swimming speed, is related to body size in free-ranging seabirds, pinnipeds and cetaceans. Proc R Soc Lon 274, 471477 (2007).

65. Akamatsu, T., Wang, D., Wang, K., Li, S. & Dong, S. Scanning sonar of rolling porpoises during prey capture dives. J Exp Biol 213, 146152 (2010).

66. Peters, R. H. The Ecological Implications of Body Size. Cambridge University Press (1986).67. Dial, K. P., Greene, E. & Irschick, D. J. Allometry of behavior. Trends Ecol Evol 23, 394401 (2008).68. Kalko, E. K. Insect pursuit, prey capture and echolocation in pipistrelle bats. Anim Behav 50, 861880 (1995).69. Madsen, P. T., Johnson, M., Aguilar Soto, N., Zimmer, W. M. X. & Tyack, P. L. Biosonar performance of foraging beaked whales (Mesoplodon densirostris). J Exp Biol 208, 181194 (2005).

70. Jensen, F. H., Bejder, L., Wahlberg, M., Aguilar Soto, N. & Madsen, P. Vessel noise eects on delphinid communication. Mar Ecol Prog Ser 395, 161175 (2009).

71. Nishiwaki, M. General biology in Mammls of the sea: biology and medicine (ed. S. H. Ridgeway) 3204 (Thomas 1972).72. Jakobsen, L., Brinklv, S. & Surlykke, A. Intensity and directionality of bat echolocation signals. Front physiol 4 (2013).73. Johnson, M., Hickmott, L. S., Aguilar Soto, N. & Madsen, P. T. Echolocation behaviour adapted to prey in foraging Blainvilles beaked whale (Mesoplodon densirostris). Proc Roy Soc Lon 275, 133139 (2008).

74. Miller, P. J. O., Johnson, M. P., Tyack, P. L. & Terray, E. A. Swimming gaits, passive drag and buoyancy of diving sperm whales Physeter macrocephalus. J Exp Biol 207, 19531967 (2004a).

Acknowledgements

We thank the eld crews during the Andenes study and the NMFS North Atlantic sperm whale study (D. Palka, chief scientist) for their dedicated help. K.A. Shorter kindly brought the floating tag data to our attention. W. Hoggart, M. Fage and Anders Svenson are thanked for expert boat handling. We also thank K. Beedholm, D. Wisniewska, M. Wahlberg and M. de Freitas. Field work in Norway was funded by the Carlsberg Foundation and the National Danish Research Council to PTM. The NMFS study was funded by the U.S. Mineral Management Service. MJ is funded by the Marine Alliance for Science and Technology, Scotland, and by a Marie Curie Career Integration Grant. MW was funded by the Danish Council for Independent Research, Natural Science and NAS is currently funded by a EU Horizon 2020 MSC Fellowship.

SCIENTIFIC REPORTS

12

www.nature.com/scientificreports/

Author Contributions

Conceived and designed the experiments: P.T.M. and M.J. Performed the experiments: P.T.M. and M.J. Analyzed the data: A.F. Contributed analysis tools: M.J., A.F. and P.T.M. Wrote the paper: A.F., P.T.M., M.J., N.A.S. and M.W. Performed statistical analyses: A.F.

Additional Information

Competing nancial interests: The authors declare no competing nancial interests.

How to cite this article: Fais, A. et al. Sperm whale predator-prey interactions involve chasing and buzzing, but no acoustic stunning. Sci. Rep. 6, 28562; doi: 10.1038/srep28562 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS

13