Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

Olga A Filatova

doi:10.1371/journal.pone.0236749

. 2020 Jul 30;15(7):e0236749. doi: 10.1371/journal.pone.0236749

Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

Olga A Filatova ^1,^*

Editor: William David Halliday²

PMCID: PMC7392277 PMID: 32730308

Abstract

Each resident-type (R-type) killer whale pod has a set of stereotyped calls that are culturally transmitted from mother to offspring. The functions of particular call types are not yet clearly understood, but it is believed that calls with two independently modulated frequency components (biphonic calls) play an important role in pod communication and cohesion at long ranges. In this study we examined the possible functions of biphonic calls in R-type killer whales. First, we tested the hypothesis that the additional component enhances the potential of a call to identify the family affiliation. We found that the similarity patterns of the lower- and higher frequency components across the families were largely unrelated. Calls were classified more accurately to their respective family when both lower- and higher-frequency components were considered. Second, we tested the long-range detectability of the lower- and higher-frequency components. After adjusting the received levels by the killer whale hearing sensitivity to different frequency ranges, the sensation level of the higher-frequency component was higher than the amplitude of the lower-frequency component. Our results suggest that the higher-frequency component of killer whale biphonic calls varies independently of the lower-frequency component, which enhances the efficiency of these calls as family markers. The acoustic variation of the higher-frequency component allows the recognition of family identity of a caller even if the shape of the lower-frequency component accidentally becomes similar in unrelated families. The higher-frequency component can also facilitate family recognition when the lower-frequency component is masked by low-frequency noise.

Introduction

Killer whales are among the few mammalian species that possess vocal learning abilities [1]. Each resident-type (R-type) killer whale pod has a unique call repertoire–vocal dialect consisting from a set of stereotyped calls that calves learn from their mothers and other matrilineal relatives [2]. The dialects, being patterns of socially learned behavioral phenotypes, therefore represent a form of animal culture [3].

Cultural traditions have been described in many animal species [4]. Social transmission of behavioral innovations is especially beneficial in situations of rapid environmental change, when instinctive behavior that evolves through natural variation and selection is too slow to follow the arising challenges and fast shifts in behavioral adaptations are required. In killer whales, the vocal culture serves a different function. The repertoires change simultaneously with social divergence, and the cultural inheritance of dialects provides a system for recognition of relatedness of the social units [2]. Each R-type killer whale pod shares the same dialect; new pods form gradually through the split of an ancestral pod, and the dialects of the splitting pods slowly diverge in time due to changes in call structure. This provides the complex system of dialects with varying degrees of similarity to each other that more or less reflect the degree of relatedness between the corresponding pods [5, 6].

The functions of particular call types are not yet clearly understood, but it is believed that calls with complex structure consisting of two independently modulated frequency components play important role in pod communication and cohesion at long ranges [7, 8]. These ‘biphonic’ or ‘two-voiced’ calls are produced more often during encounters of several pods rather than in a single-pod context [9–11]. Additionally, the differences in directionality of the lower- and higher-frequency components allow a listener to infer the direction of movement of a caller [10].

Biphonation is widespread in mammals and occur in many species as diverse as canids [12–14], primates [15–17] and whales [18, 19]. However, functional interpretations of biphonation are scarce. The proposed functions of this phenomenon include the increase of unpredictability and indication of physical condition [20] and enhancement of individual recognition [20–22]. Volodina [22] reported that the correct individual classification was much higher in dhole (Cuon alpinus) calls when using both components of a biphonic call together than when using each of them separately. Among cetaceans, the role of biphonic call in individual recognition has not been tested, though biphonic calls were reported to serve as individual signature sounds in bottlenose dolphins [23, 24].

In R-type killer whales, recognition of a family dialect appears to be at least as important as individual recognition, because a whale stays in its natal family for life, and family affiliation is important for its survival. In contrast to dholes and most other mammals, killer whales possess an ability to change the structure of their calls deliberately in a specific direction [1]. However, the speed of call change is rather slow [25], and optimal complexity of call contours is limited by propagation loss that attenuates fine details of contour structure. In this situation, whales can benefit from ability to change different components of a call independently, which would increase the degree of structural divergence of calls from related families.

Biphonic calls occur in all studied killer whale populations [2, 8, 26–28]. In most populations both in the North Pacific and the North Atlantic, the average frequencies of the higher- and lower-frequency components are more or less similar [29] except for the North Pacific ‘transient-type’ (T-type) mammal-eating killer whales, which have been shown to be the most genetically divergent group among all studied killer whale populations [30]. The widespread occurrence and homogeneity in frequency of both components across populations suggest that biphonic calls are an essential part of the acoustic repertoire of the killer whale and bear specific functions that are similar in different populations.

In this study we examine the possible functions of biphonic calls in R-type killer whales. First, we test the hypothesis that the additional component enhances the potential of a call to identify the pod and family affiliation. We compare the similarity patterns across the lower- and higher-frequency components and test whether either of the components or both of them in combination work better to mark the family identity. Second, we test the long-range detectability of the lower- and higher-frequency components by comparing their received levels and adjusting them by killer whale hearing sensitivity to different frequency ranges.

Material and methods

Ethics statement

This study was part of the research topic of Lomonosov Moscow State University “Principles of high-frequency and ultrasound communication in mammals”. According to the laws of Russian Federation, no permit is required for distant non-invasive research of cetaceans. The field studies did not involve endangered or protected species.

Data collection

The materials and data for this study were collected as part of the Far East Russia Orca Project (FEROP) in Avacha Gulf, Kamchatka, during the summer months of the years 2012–2019. The underwater sound recordings were made from 4–4.5 m inflatable boats while the engines were turned off, at a sampling frequency of 48 or 96 kHz. For the recording we used Offshore Acoustics hydrophone (nominal sensitivity -154 dB re 1 V/μPa, frequency response curve 6 Hz to 14 kHz 1–3 dB according to manufacturer’s specifications) with Zoom H4 and Zoom H6 flash recorders. The photographic identification (photo-ID) method was used to identify individual killer whales and families. To take photographs, we approached the whales to a distance of 20-50m, or moved the boat 200-300m ahead of the animals and off to the side and waited until they passed. Photographs of the left side of individual whales were taken to show the details of dorsal fin and saddle patch, using the technique developed by Bigg et al. [31].

The resident (R-type) killer whales of Eastern Kamchatka, Russia, are known to range along the east coast of Kamchatka peninsula from Avacha Gulf to Karaginsky Gulf and east to the Commander Islands [32]. Whales from this population live in stable social units that include maternal relatives with no dispersal observed [33]. We do not have enough data to reconstruct the full genealogies of these units, and we suspect that in some cases one unit can include more than one matriline, so we use the term "family" to denote these units. Families that share the same vocal dialect are attributed to the same pod, and pods with similar dialects form clans. To date we recognize at least 62 families, which belong to 20 pods in three acoustic clans forming a single community: Avacha clan, K19 clan and K20 clan [33, 34]. Avacha clan, consisting of more than 13 pods and 30 families, is the most common. For this study we used only the families from Avacha clan for which sufficient data were available.

Calls were classified according to the existing catalogue [35] with some additional subtypes identified through our further studies. For this study, we used calls of two most common biphonic types of Avacha clan—K5 and K7 types.

Similarity of contours of the lower and higher frequency components

For the analysis of contour similarities, I used high-quality calls recorded from particular families when they were alone in the area with no other families nearby. In total I used calls produced by 14 families from five different pods, 10 calls from each type per family (except for K5 call type of Ikar family which had two distinctive subtypes, so we used 10 calls of each subtype from this family, 20 calls from this family in total). To cover intra-group variation, I selected calls from as many independent recording sessions from each family as possible. No sample from any family contained calls from fewer than three independent sessions with that family.

Call contours were extracted using a custom-made MATLAB script for manually tracking frequency contours of each syllable (for the detailed description of the algorithm see [36]. After the operator selected enough points to track all modulations of the frequency contour, the algorithm smoothed and interpolated them to produce a vector of frequency measurements with a sampling interval of 0.01 s. The contours of the lower- and higher-frequency components were extracted independently of each other.

The similarity of contours was measured using dynamic time-warping, which allows limited compression and expansion of a signal’s time axis to maximize the frequency overlap with a reference signal. For this study, I used the warping algorithm developed by Deecke and Janik [37]. The percent similarity of contours was calculated by dividing the smaller frequency value by the larger value at each point and multiplying by 100. From the resulting similarity matrix, a cost matrix was constructed that kept a running tab on the similarities of the elements making up the curves while adding up these costs to produce a final number that indicated the percent similarity between the contours. To calculate a distance measure between each pair of calls in our analysis, I subtracted this value from 100%.

In order to estimate the potential ability of killer whales to assess the family identity of calls, I classified the calls using ARTwarp algorithm in MATLAB [37]. The similarity between the extracted contours and the reference contours was calculated using dynamic time-warping. As my aim was to assess the rate of correct classification to existing categories (families), I modified the ARTwarp algorithm used by Deecke and Janik [37] in order to prevent it from creating additional categories and modifying the reference contours. For this, on the first step I calculated a reference contour for each family using the corresponding module of ARTwarp. On the second step, I used model contours as weight matrix and fixed the number of categories, so that all new contours had to be assigned to any of the existing categories, represented by the family reference contours. Then, I calculated the rate of correct classifications for lower- and higher-frequency component contours separately and for the contours that consisted of both components taken together.

Detectability of the lower- and higher-frequency components over distance

I measured the amplitude in the middle of the fundamental frequency contour of the lower- and higher-frequency components using the rectangle cursor tool in AviSoft SASLab Pro (Fig 1). As the aim of the measurements was to access the difference between the amplitude of the higher- and lower-frequency components, and not the absolute source level values, I did not account for the parameters of the recording system.

Fig 1 — Locations of the amplitude measurements taken from the lower- and higher-frequency components of K5 (top) and K7 (bottom) call types.

The calls were collected for the analysis at different locations within Avacha Gulf. Spectral sound propagation loss can vary between locations due to differences in bathymetry, sediment structure and water properties; some variation is also introduced by differences in depths at which killer whales produce their calls. To address this, we measured a large amount of calls produced in various conditions and in different situations to cover as much natural variation as possible. As killer whales also experience all these propagation effects when listening to the conspecific calls, we assume that our sample set is an adequate representative of what killer whale normally hear in the wild.

I measured the total of 689 K5 calls and 372 K7 calls from 2012–2019. To estimate the average received levels of these call components, I measured all observed calls of these types regardless of their quality and signal-to-noise ratio. If the higher-frequency component was not detectable on the sonogram, the call was not used for the further analysis.

Then, I calculated the difference between the amplitude of the lower- and higher-frequency components as perceived by killer whales. For this, I used killer whale audiogram [38, 39] to adjust for the hearing threshold differences between the frequencies of the lower- and higher-frequency components. The audiograms of individual killer whales differed from each other, so I used the ‘model’ audiogram (Fig 3 in [39]) for the calculations. From the audiogram, I estimated the threshold at the mean frequency of each component of each call type. For K5 type, the mean frequency of the lower-frequency component was 1.1 kHz and of the higher-frequency component– 9.7 kHz, which corresponded to the thresholds of 93 and 60 dB re 1 μPa, respectively. Therefore, I assumed that the difference in the sensitivity to the lower- and higher-frequency components of K5 type should be about 33 dB. For K7 type, the mean frequency of the lower-frequency component was 1.9 kHz and of the higher-frequency component– 6.5 kHz, which corresponded to the thresholds of 83 and 66 dB re 1 μPa, respectively. Therefore, I assumed that the difference in the sensitivity to the lower- and higher-frequency components of K7 type should be about 17 dB.

Results

Similarity of contours of the lower and higher frequency components

The similarity patterns of the lower and higher frequency components of both K5 and K7 calls across the families were largely unrelated (Fig 2 and S1 Table). The distance matrices of the lower and higher-frequency components of K5 and K7 calls had very weak correlation which was significant only for K5 (Mantel test, K5: r = 0.089, p = 0.01; K7: r = 0.046, p = 0.099).

Both for K5 and K7 call types distances between lower- and higher-frequency components from the same families and the same pods were smaller on average than between different pods (Table 1, Fig 2). Distances between the lower-frequency component contours of each call type were higher than the distances between the high-frequency contours for the same family/pod category of the same type, except for K7 call contours of the same families, where the relationship was reversed (Table 1).

Table 1. Distances (mean ± SD) between pairs of calls calculated through dynamic time warping for lower- and higher-frequency components of K5 and K7 call types between the same families, between different families from the same pods, and between different pods.

	K5		K7
	lower-frequency component	higher-frequency component	lower-frequency component	higher-frequency component
Same families	4.4 ± 2.8	3.4 ± 1.9	2.8 ± 1.9	3.8 ± 2.1
Different families, same pods	8.4 ± 4.2	5.4 ± 3.4	5.5 ± 3.7	5.1 ± 2.8
Different pods	12.3 ± 5.8	8.9 ± 4.9	8.2 ± 3.7	6.3 ± 2.9

Open in a new tab

The scatterplot of K5 call type distances clearly demonstrates the discrepancy between the similarity patterns of the lower- and higher-frequency components: there were many pairs of calls with highly different lower-frequency components and relatively similar higher-frequency components, and also many pairs of calls with dissimilar higher-frequency components and relatively similar lower-frequency components, but few calls that were highly dissimilar both in low- and higher-frequency components. This pattern did not occur in K7 call type.

Both K5 and K7 calls were classified more accurately to their respective family when both lower- and higher-frequency components were considered. For K5 calls, the correct classification rate was 70% when only the lower-frequency component was involved, 56% with only the higher-frequency component, and 81% when both components were used. For K7 calls, the correct classification rate for the lower-frequency component was 66%, for the higher-frequency component 51%, and for both components together 68%.

Detectability of the lower and higher frequency components over distance

The higher-frequency component was detectable in most calls: only 16% of K5 calls and 5% of K7 calls had no visible higher-frequency component on the sonograms (S2 Table). All of the calls without detectable high-frequency components were distant and the amplitude of the lower-frequency component was also low, confirming that the higher-frequency component was not detectable due to transmission loss, and not because it was absent in the original call. All good-quality calls of these types had pronounced higher-frequency component. The frequencies of the higher-frequency component absorb roughly 1 dB/km more than the frequencies of the lower-frequency component; due to this effect, at 10 km distance, the higher-frequency component loose approximately 10 dB more than the lower-frequency component [7].

The received amplitude of the higher-frequency component was substantially lower than that of the lower-frequency component in K5 call type, but not in K7 call type (Fig 3). In K5 call type, the mean amplitude of the higher-frequency component was 17.5 dB lower than the mean amplitude of the lower-frequency component. In K7 call type, the mean amplitude of the higher-frequency component was only 0.1 dB lower than the mean amplitude of the lower-frequency component.

Fig 3 — **Measured amplitude in the middle of the lower-frequency component (LFC, red) and in the middle of the higher-frequency component (HFC, blue) of K5 (top) and K7 (bottom) calls.** Green identify measurements of the higher-frequency component adjusted by killer whale audiogram.

The higher-frequency component of K5 call type had higher frequency than that of K7 call type (mean±SD, K5: 9715±785 Hz; K7: 6505±305 Hz). Consequently, due to the higher hearing sensitivity of killer whales to the higher frequencies, more pronounced adjustment by the hearing threshold was applied for the higher-frequency component of K5 call type (33 dB) than for that of K7 call type (17 dB). The perceived amplitude of the higher-frequency component of K5 call type after adjusting it by the hearing threshold was 15.5 dB higher, than the amplitude of the lower-frequency component. The perceived amplitude of the higher-frequency component of K7 call type after adjusting it by the hearing threshold was 16.9 dB higher, than the amplitude of the lower-frequency component.

Discussion

In killer whales, the main function of biphonation has been hypothesized to be the better discrimination of a caller’s orientation by listeners. Miller [40] has shown that when a calling killer whale is oriented towards a listener, the overlapping high frequency component bears substantially more energy, than when a caller is oriented away from a listener. It happens because high-frequency sounds are more directional in toothed whales, likely due to specific adaptations of their sound producing and enhancing structures. The difference in the relative energy of the lower and higher frequency components allows a listener to infer the direction of a caller’s movement, which is crucial to the coordination of individuals on a distance. Underwater visibility is low compared to air; for example, in the North Pacific in summer it is hardly possible to see anything further than several meters. Biphonic sounds give the whales clues to identify the direction the other animals are moving. Moreover, the directionality of the higher-frequency component can provide additional benefits, such as allowing a caller to emit family/clan membership information towards specific whales or groups within an aggregation selectively.

However, if the orientation marking was the only function of the higher-frequency component, there would be no need to vary its shape across families: the same contour for all families would work equally well. Besides, the higher-frequency component is not the only element of acoustic repertoire that can indicate the orientation of the caller. Other sounds, such as echolocation clicks, burst-pulse sounds and even whistles also have directional properties [41].

In several terrestrial species, the overlapping higher-frequency component has been suggested to increase individual recognition. For example, it was shown that joining of two independent call components into a common vocalization may function to enhance individual recognition in the dhole [22]. In king and emperor penguins, biphonation enhances the accuracy of parent-chick and mate-mate recognition [21].

Additional hypotheses for the function of biphonation in other species have included: increasing unpredictability and providing an indication of physical condition [20]. However, these are likely not appropriate considerations as possible functions of biphonation in killer whales. These hypotheses suggest that the caller can choose when to include or not include a biphonic component in each call it makes. However, calls that comprise killer whale dialects are stereotyped, and if a call is biphonic, the presence of the higher-frequency component is obligatory, i.e. the same call type is not normally produced alternately with and without the higher-frequency component. Most likely, the higher-frequency component of killer whale biphonic calls functions as an alternative contour that bears information on the caller’s identity (probably on the family level, in contrast to individual level in dholes and penguins). The presence of two independent components makes the recognition system twice more error-proof both in space and time. In space, if one of the components is not recognizable due to distance or masked by noise, the whales can still use the other component to identify the caller’s family affiliation. In time, if in some families the contour shape of the higher- or lower-frequency component randomly converges through the process of cultural evolution [42], the other component still remains different between the families and allows listeners to discriminate between them.

The marine environment favors the acoustic channel for information transfer, because visibility underwater is very poor, while sound travels much further than in air. However, acoustic transmission has its limitations which are different for the low- and high-frequency sounds. Low-frequency sounds attenuate less and travel further than high-frequency sounds [43], therefore, in quiet conditions the lower-frequency component has better potential for long-range communication. On the other hand, underwater noise (both natural and man-made) is normally more pronounced in lower frequencies, often masking the lower-frequency component of killer whale calls.

Ship noise is a serious issue that can substantially reduce the detection distance of killer whale calls [44]. Killer whales can increase the amplitude of their calls in response to ship noise [45], but no studies have examined the relative noise immunity of the higher- and lower-frequency components of killer whale calls so far. Although at close ranges the energy of ship noise can extend above the higher-frequency component [46], normally it is mostly concentrated on low frequencies [47]. Therefore, it is likely that the higher-frequency component which lies above noise may function to facilitate pod and family recognition in noisy conditions. The directionality of the higher-frequency component can also help increase the signal-to-noise ratio in the direction of a receiver in noisy conditions. The amplitude comparisons in our study suggest that both higher- and lower-frequency components are equally important: the lower-frequency component is normally slightly louder, but this is compensated by the better hearing sensitivity of killer whales to higher frequencies [38, 39]. The higher-frequency component has higher sensation level, i.e. the whales normally hear the higher-frequency component slightly better than the lower-frequency component, which highlights its potential importance for their underwater communication.

Most biphonic calls of killer whales have heterodyne frequencies below and above the higher-frequency component (Fig 4). Heterodyne frequencies arise from the interaction of the two components, when the lower-frequency component is amplitude modulating the higher-frequency component [8, 12, 48]. This leads to the appearance of sidebands below and above the higher-frequency component that are equal to the difference or sum of the fundamental frequency of the higher-frequency component and the fundamental frequency and harmonics of the lower-frequency component [48]. Therefore, from the shape of the higher-frequency component and the heterodyne frequencies it is possible to infer the shape of the lower-frequency component when it is masked by low-frequency noise. It is unknown whether killer whales employ this capacity of biphonic calls, but it emphasizes the potential of the independently modulated higher-frequency component for underwater acoustic communication.

When both call components are detectable, biphonation can still be useful to duplicate the recognition potential of stereotyped calls, because the shape of call contours can randomly converge in unrelated social units [42]. Killer whale calls are not inherited genetically, but are rather learned from mother and other maternal relatives. Complex repertoires of stereotyped calls–vocal dialects–represent a form of animal culture [3]. Killer whale dialects slowly change in time through learning errors and innovations [2, 25, 49–51]; this process of cultural change is called cultural evolution [52, 53]. Since the variability of sound contours is limited due to the natural physical constrains, call contours of unrelated killer whale social units can sometimes become more similar due to the random convergence [42]. This would impede the discrimination between these two pods on a distance, when some call features are masked. The presence of an independently modulated and separately evolving call component would enhance recognition, because the probability that both components would converge in two pods is very low. Indeed, our results demonstrate that call discrimination by dynamic time warping was better when both components were used, compared to each component separately.

Comparing the similarity patterns, I have found that they were different for the lower- and higher-frequency components. Many calls from different families had highly different higher-frequency component and relatively similar lower-frequency component and vice versa, but there were relatively few calls that were dissimilar in both components. This can indicate that dissimilarity in killer whale calls is costly and/or difficult to achieve because of the structural constrains [42], and for this reason it rarely occurs in both components.

Two examined call types had different patterns of similarity of the lower- and higher-frequency components: in K5 call type, adding the higher-frequency component to the contour substantially increased the rate of correct classification to family, while in K7 call type it increased the correct classification rate only slightly, suggesting that the higher-frequency component of K7 call type differs little across families. Different patterns of similarity of the lower- and higher-frequency components in different call types have been demonstrated previously for the eastern North Pacific killer whales: in some call types, adding the higher-frequency component increased the discrimination error, and in other call types, the effect was the opposite [54]. It is possible that the higher-frequency component of K7 serves the function of clan rather than family recognition. In Kamchatka, biphonic calls with higher-frequency component on 6–7 kHz, like K7, are found only in Avacha clan [34], which makes this clan easily acoustically recognizable on a distance. Yurk [55] also found significant differences across killer whale clans of Alaska and British Columbia in frequencies of the higher-frequency component, but not the lower-frequency component. This suggests that in some populations the two components may serve for recognition on different levels–the lower-frequency component on the family level and the higher-frequency component on the clan level. This can happen if the components evolve with different speed. Variation in the rates of cultural evolution of different elements of killer whale vocal repertoire has been reported previously: Deecke et al. [25] showed that different call types can change with different speed, and Filatova et al. [5] found that different syllables within calls diverge with different speed, including syllables of both the lower- and higher-frequency components. Yurk [55] performed McDonald-Kreitman test for positive selection and found different selective pressures for the lower- and higher-frequency components. Overall, these findings suggest that independent cultural change of both components of biphonic calls may facilitate acoustic recognition of different levels of social structure (clans, pods or families) in killer whales. However, despite all of this evidence is consistent with independent evolution of the two components, it does not demonstrate it directly and rigorously. In future, analysis of the component variation over time is necessary to exclude the alternative drivers of the observed patterns.

In conclusion, I suggest that a likely function of the higher-frequency component is to duplicate and/or complement the social identity marking when the lower-frequency component is masked by noise or accidentally appears similar in unrelated social units. The combination of both components provides a redundancy of information that is beneficial for these animals to maintain contact over distance in the noisy underwater environment. The cultural evolution of two components of biphonic calls is happening independently of each other and can occur with different speed, which ensures the lack of correlation in their similarity patterns, providing the whales with an error-proof back-up mechanism to recognize the social affiliation of their conspecifics within distance of acoustic contact.

Supporting information

S1 Table. Similarity matrices for the lower- and higher-frequency components of K5 and K7 calls.

(XLSX)

Click here for additional data file.^{(945.9KB, xlsx)}

S2 Table. Measurements of the amplitude of the lower- and higher-frequency components of K5 and K7 calls.

The table also shows the time stamp of the measurement within a file, and the frequency at which the amplitude was measured.

(XLSX)

Click here for additional data file.^{(62.3KB, xlsx)}

Acknowledgments

I am grateful to all members of FEROP expeditions, especially to Mikhail Guzeev and Anastasya Danishevskaya who were responsible for sound recordings, and to Tatiana Ivkovich for photo-identification of individual killer whales.

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

This work was supported by Russian Science Foundation (http://www.rscf.ru/) funding to OAF, grant number 19-14-00037. This research received no additional funding from any public, commercial or not-for-profit sectors. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Abramson JZ, Hernández-Lloreda MV, García L, Colmenares F, Aboitiz F, Call J. Imitation of novel conspecific and human speech sounds in the killer whale (Orcinus orca). Proc Royal Soc B Biol Sci. 2018; 285(1871): 20172171 10.1098/rspb.2017.2171 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Ford JKB. Vocal traditions among resident killer whales (Orcinus orca) in coastal waters of British Columbia. Can J Zool. 1991; 69: 1454–1483. 10.1139/z91-206 [DOI] [Google Scholar]
3.Rendell L, Whitehead H. Culture in whales and dolphins. Behav Brain Sci. 2001; 24(2): 309–324. 10.1017/s0140525x0100396x [DOI] [PubMed] [Google Scholar]
4.Laland KN, Hoppitt W. Do animals have culture? Evol Anthropol. 2003; 12(3): 150–159. 10.1002/evan.10111 [DOI] [Google Scholar]
5.Filatova OA, Ivkovich TV, Guzeev MA, Burdin AM, Hoyt E. Social complexity and cultural transmission of dialects in killer whales. Behav. 2017; 154(2): 171–194. 10.1163/1568539X-00003417 [DOI] [Google Scholar]
6.Deecke VB, Barrett-Lennard LG, Spong P, Ford JKB. (2010). The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca). Naturwissenschaften. 2010; 97:513–518. 10.1007/s00114-010-0657-z [DOI] [PubMed] [Google Scholar]
7.Miller PJO. Diversity in sound pressure levels and estimated active space of resident killer whale vocalizations. J Comp Physiol A. 2006; 192: 449–459. 10.1007/s00359-005-0085-2 [DOI] [PubMed] [Google Scholar]
8.Miller PJO, Samarra FI, Perthuison AD. Caller sex and orientation influence spectral characteristics of “two-voice” stereotyped calls produced by free-ranging killer whales. J Acoust Soc Amer. 2007; 121: 3932–3937. 10.1121/1.2722056 [DOI] [PubMed] [Google Scholar]
9.Filatova OA, Fedutin ID, Nagaylik MM, Burdin AM, Hoyt E. Usage of monophonic and biphonic calls by free-ranging resident killer whales (Orcinus orca) in Kamchatka, Russian Far East. Acta Ethol. 2009; 12:37–44. 10.1007/s10211-009-0056-7 [DOI] [Google Scholar]
10.Foote AD, Osborne RW, Hoelzel AR. Temporal and contextual patterns of killer whale (Orcinus orca) call type production. Ethol. 2008; 114:599–606. 10.1111/j.1439-0310.2008.01496.x [DOI] [Google Scholar]
11.Weiß BM, Symonds H, Spong P, Ladich F. Intra-and intergroup vocal behavior in resident killer whales, Orcinus orca. J Acoust Soc Am. 2007; 122: 3710–3716. 10.1121/1.2799907 [DOI] [PubMed] [Google Scholar]
12.Wilden I, Herzel H, Peters G, Tembrock G. Subharmonics, biphonation and deterministic chaos in mammal vocalization. Bioacoustics. 1998; 9:1 71–196. 10.1080/09524622.1998.9753394 [DOI] [Google Scholar]
13.Riede T, Herzel H, Mehwald D, Seidner W, Trumler E, Tembrock G, et al. Nonlinear phenomena and their anatomical basis in the natural howling of a female dog-wolf breed. J Acoust Soc Am. 2000; 108: 1435–1442. 10.1121/1.1289208 [DOI] [PubMed] [Google Scholar]
14.Volodin IA, Volodina EV. Biphonation as a prominent feature of dhole Cuon alpinus sound. Bioacoustics 2002; 13: 105–120. 10.1080/09524622.2002.9753490 [DOI] [Google Scholar]
15.Fischer J, Hammerschmidt K, Cheney DL, Seyfarth RM. Acoustic features of female chacma baboon barks. Ethol 2001; 107: 33–54. 10.1111/j.1439-0310.2001.00630.x [DOI] [Google Scholar]
16.Brown CH, Alipour F, Berry DA, Montequin D. Laryngeal biomechanics and vocal communication in the squirrel monkey (Saimiri boliviensis). J Acoust Soc Am. 2003; 113: 2114–2126. 10.1121/1.1528930 [DOI] [PubMed] [Google Scholar]
17.Riede T, Owren MJ, Arcadi AC. Nonlinear acoustics in pant hoots of common chimpanzees (Pan troglodytes): frequency jumps, subharmonics, biphonation, and deterministic chaos. Am J Primatol 2004; 64:277–291. 10.1002/ajp.20078 [DOI] [PubMed] [Google Scholar]
18.Tyson RB, Nowacek DP, Miller PJO. Nonlinear phenomena in the vocalizations of North Atlantic right whales (Eubalaena glacialis) and killer whales (Orcinus orca). J Acoust Soc Am. 2007; 122(3): 1365–1373. 10.1121/1.2756263 [DOI] [PubMed] [Google Scholar]
19.Quick N, Callahan H, Read AJ. Two‐component calls in short‐finned pilot whales (Globicephala macrorhynchus). Mar Mam Sci, 2018; 34(1): 155–168. 10.1111/mms.12452 [DOI] [Google Scholar]
20.Fitch WT, Neubauer J, Herzel H. Calls out of chaos: the adaptive significance of nonlinear phenomena in mammalian vocal production. Anim Behav 2002; 63: 407–418. 10.1006/anbe.2001.1912 [DOI] [Google Scholar]
21.Aubin T, Jouventin P, Hildebrand C. Penguins use the twovoice system to recognize each other. Proc R Soc Lond B Biol Sci. 2000; 267: 1081–1087. 10.1098/rspb.2000.1112 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Volodina EV, Volodin IA, Isaeva IV, Unck C. Biphonation may function to enhance individual recognition in the dhole, Cuon alpinus. Ethol. 2006; 112(8): 815–825. 10.1111/j.1439-0310.2006.01231.x [DOI] [Google Scholar]
23.Kriesell HJ, Elwen SH, Nastasi A, Gridley T. Identification and characteristics of signature whistles in wild bottlenose dolphins (Tursiops truncatus) from Namibia. PLoS One. 2014; 9(9): e106317 10.1371/journal.pone.0106317 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Papale E, Buffa G, Filiciotto F, Maccarrone V, Mazzola S, Ceraulo M, et al. Biphonic calls as signature whistles in a free-ranging bottlenose dolphin. Bioacoustics. 2015; 24(3): 223–231. 10.1080/09524622.2015.1041158 [DOI] [Google Scholar]
25.Deecke VB, Ford JKB, Spong P. Dialect change in resident killer whales: implications for vocal learning and cultural transmission. Anim Behav. 2000; 60: 629–638. 10.1006/anbe.2000.1454 [DOI] [PubMed] [Google Scholar]
26.Yurk H, Barrett-Lennard L, Ford JKB, Matkin CO. Cultural transmission within maternal lineages: vocal clans in resident killer whales in southern Alaska. Anim Behav. 2002; 63: 1103–1119. 10.1006/anbe.2002.3012 [DOI] [Google Scholar]
27.Strager H. Pod-specific call repertoires and compound calls of killer whales, Orcinus orca Linnaeus, 1758, in the waters of northern Norway. Can J Zool. 1995; 73: 1037–1047. 10.1139/z95-124 [DOI] [Google Scholar]
28.Wellard R, Pitman RL, Durban J, Erbe C. Cold call: the acoustic repertoire of Ross Sea killer whales (Orcinus orca, Type C) in McMurdo Sound, Antarctica. Royal Soc Open Sci. 2020; 7(2): 191228 10.1098/rsos.191228 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Filatova OA, Miller PJO, Yurk H, Samarra FIP, Hoyt E, Ford JKB, et al. Killer whale call frequency is similar across the oceans, but varies across sympatric ecotypes. J Acoust Soc Amer. 2015; 138: 251–257. 10.1121/1.4922704 [DOI] [PubMed] [Google Scholar]
30.Morin PA, Archer FI, Foote AD, Vilstrup J, Allen EE, Wade P, et al. Complete mitchondrial genome phylogeographic analysis of killer whales (Orcinus orca) indicates multiple species. Genome Res. 2010; 20: 908–916. 10.1101/gr.102954.109 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Bigg MA, MacAskie I, Ellis G. Photo-identification of individual killer whales. Whalewatcher. 1983; 17(1): 3–5. [Google Scholar]
32.Burdin AM, Hoyt E, Sato H, Filatova OA. 2006. The Killer Whales of Eastern Kamchatka. Alaska SeaLife Center; 2006. [Google Scholar]
33.Ivkovich TV, Filatova OA, Burdin AM, Sato H, Hoyt E. The social organization of resident-type killer whales (Orcinus orca) in Avacha Gulf, Northwest Pacific, as revealed through association patterns and acoustic similarity. Mamm Biol. 2010; 75: 198–210. 10.1016/j.mambio.2009.03.006 [DOI] [Google Scholar]
34.Filatova OA, Fedutin ID, Burdin AM, Hoyt E. The structure of the discrete call repertoire of killer whales Orcinus orca from Southeast Kamchatka. Bioacoustics. 2007; 16: 261–280. 10.1080/09524622.2007.9753581 [DOI] [Google Scholar]
35.Filatova OA, Burdin AM, Hoyt E, Sato H. A catalogue of discrete calls of resident killer whales (Orcinus orca) from the Avacha Gulf of Kamchatka Peninsula. Zoologicheskii Journal. 2004; 83: 1169–1180 [Google Scholar]
36.Filatova OA, Deecke VB, Ford JKB, Matkin CO, Barrett-Lennard LG, Guzeev MA, et al. Call diversity in the North Pacific killer whale populations: implications for dialect evolution and population history. Anim Behav. 2012. 83: 595–603. 10.1016/j.anbehav.2011.12.013 [DOI] [Google Scholar]
37.Deecke VB, Janik VM. Automated categorization of bioacoustic signals: avoiding perceptual pitfalls. J Acoust Soc Amer. 2006; 119(1): 645–653. 10.1121/1.2139067 [DOI] [PubMed] [Google Scholar]
38.Szymanski MD, Bain DE, Kiehl K, Pennington S, Wong S., Henry KR. Killer whale (Orcinus orca) hearing: Auditory brainstem response and behavioral audiograms. J Acoust Soc Am. 1999; 106: 1134–1141. 10.1121/1.427121 [DOI] [PubMed] [Google Scholar]
39.Branstetter BK, Leger JSt, Acton D, Stewart J, Houser D, Finneran JJ, et al. Killer whale (Orcinus orca) behavioral audiograms. J Acoust Soc Amer. 2017; 141(4): 2387–2398. 10.1121/1.4979116 [DOI] [PubMed] [Google Scholar]
40.Miller PJO. Mixed-directionality of killer whale stereotyped calls: A direction of movement cue? Behav Ecol Sociobiol. 2002; 52(3): 262–270. 10.1007/s00265-002-0508-9 [DOI] [Google Scholar]
41.Branstetter BK, Moore PW, Finneran JJ, Tormey MN, Aihara H. Directional properties of bottlenose dolphin (Tursiops truncatus) clicks, burst-pulse, and whistle sounds. J Acoust Soc Amer. 2012; 131: 1613–1621. 10.1121/1.3676694 [DOI] [PubMed] [Google Scholar]
42.Filatova OA, Samarra FIP, Barrett-Lennard LG, Miller PJO, Ford JKB, Yurk H, et al. Physical constraints of cultural evolution of dialects in killer whales. J Acoust Soc Amer. 2016; 140(5): 3755–3764. 10.1121/1.4967369 [DOI] [PubMed] [Google Scholar]
43.Hodges RP. Underwater acoustics: Analysis, design and performance of sonar. John Wiley & Sons; 2011. [Google Scholar]
44.Erbe C. Underwater noise of whale‐watching boats and potential effects on killer whales (Orcinus orca), based on an acoustic impact model. Mar Mam Sci. 2002; 18(2): 394–418. 10.1111/j.1748-7692.2002.tb01045.x [DOI] [Google Scholar]
45.Holt MM, Noren DP, Veirs V, Emmons CK, Veirs S. Speaking up: killer whales (Orcinus orca) increase their call amplitude in response to vessel noise. J Acoust Soc Amer. 2009; 125(1): EL27–EL32. 10.1121/1.3040028 [DOI] [PubMed] [Google Scholar]
46.Veirs S, Veirs V, Wood JD. Ship noise extends to frequencies used for echolocation by endangered killer whales. PeerJ. 2016; 4; e1657 10.7717/peerj.1657 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.McKenna MF, Ross D, Wiggins SM, Hildebrand JA. Underwater radiated noise from modern commercial ships. J Acoust Soc Amer. 2012; 131(1): 92–103. 10.1121/1.3664100 [DOI] [PubMed] [Google Scholar]
48.Brown JC. Mathematics of pulsed vocalizations with application to killer whale biphonation, J Acoust Soc Amer. 2008; 123(5): 2875–2883. 10.1121/1.2890745 [DOI] [PubMed] [Google Scholar]
49.Foote AD, Griffin RM, Howitt D, Larsson L, Miller PJO, Hoelzel AR. Killer whales are capable of vocal learning. Biol Lett. 2006; 2: 509–512. 10.1098/rsbl.2006.0525 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Filatova OA, Samarra FIP, Deecke VB, Ford JKB, Miller PJO, Yurk H. Cultural evolution of killer whale calls: background, mechanisms and consequences. Behav. 2015; 152(15): 2001–2038. 10.1163/1568539X-00003317 [DOI] [Google Scholar]
51.Filatova OA, Burdin AM, Hoyt E. Is killer whale dialect evolution random? Behav Proc. 2013; 99: 34–41. 10.1016/j.beproc.2013.06.008 [DOI] [PubMed] [Google Scholar]
52.Mundinger PC. Animal cultures and a general theory of cultural evolution. Ethol Sociobiol. 1980; 1: 183–223. 10.1016/0162-3095(80)90008-4. [DOI] [Google Scholar]
53.Lumsden CJ, Wilson EO. The relation between biological and cultural evolution. J Soc Biol Struct. 1985; 8: 343–359. 10.1016/0140-1750(85)90042-9. [DOI] [Google Scholar]
54.Nousek AE, Slater PJ, Wang C, Miller PJO. The influence of social affiliation on individual vocal signatures of northern resident killer whales (Orcinus orca). Biol Lett. 2006; 2: 481–484. 10.1098/rsbl.2006.0517 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Yurk H. Vocal culture and social stability in resident killer whales (Orcinus orca). Dissertation, University of British Columbia; 2005. [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0236749.r001

Decision Letter 0

William David Halliday

22 Apr 2020

PONE-D-20-05337

Independent cultural change of higher-frequency component can facilitate call recognition in killer whales

PLOS ONE

Dear Dr Filatova,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jun 06 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

William David Halliday, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Additional Editor Comments (if provided):

Four reviewers have assessed this manuscript, and all agree that it has merit, but requires significant revision before it can be published. I agree with this assessment. Please address all comments from each reviewer when revising your manuscript. I look forward to seeing the revised manuscript. Reviewer 3 uploaded their comments in a file - please remember to download this file and address those comments.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

Reviewer #3: Yes

Reviewer #4: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript describes an important bio-acoustical subject, the use of bi-phony (two-voiced calls) by killer whales and suggest possible functions.

The topic has been subject of interest not just within the killer whale or cetacean research community but is of interest to the much broader research community interested in the evolution of communication signals and is specifically addressing adaptive function of signal structure used not just by killer whales.

The manuscript is well written and understandable for an audience with some background in bioacoustics and also provides explanations of some of the more complex bio-acoustic phenomena such as heterodyne frequencies in two component pulsed vocalizations. As such the manuscript provides understandable information for a wider audience of PlosOne.

recognition in killer whales” and support some of their main conclusions, such as the complementary function of both the lower and higher frequency call components in family recognition. The authors provide support for that conclusion in their investigation of family classification accuracy of two calls of resident type killer whales in the Russian Far East by considering classification means of the two call components separately. The argument that both components can be used independently to identify family membership of the caller is well laid out. The author also provide good support from their data that calls that have both components visible in the spectrograms have higher classification means than those where only have on of the two components is visible thereby validating one of the main conclusions of the study that calls with more than one components, especially those with heterodyne frequencies, have a higher ability to be used as family group identifiers in a variety of acoustic environments.

The authors argument, however, that the two-component structure of calls provides better propagation capabilities is only weakly supported by the data provided. The reasons for my assessment is based on the physics of underwater acoustics that make it difficult to interpret the data analysis in the way the authors did.

The calls collected for the analysis likely were collected on a number of occasions taking place at different locations and at different times of the year with some consistency in the latter due to field work seasonality. Spectral sound propagation loss varies tremendously between locations, i.e. different frequencies attenuate very differently at different locations due to things like bathymetry, sediment structure and water properties, the latter varies not only spatially but also temporally (e.g. vertical sound speed profiles can vary within hours at some locations depending on tidal, wind, and water mixing conditions). We can assume that the calls were not produced all at the same depths although based on tagging studies resident killer whales may produce most of their calls in the upper water column (< 20 m water depths). This upper layer of water is often varying in sound propagation conditions mostly due to temperature fluctuations.

All of the above introduces variation in the propagation of the different components that cannot be confounded by using received levels at one location with on hydrophone. Although the authors tried to minimize this variation by only looking at the sound pressure differential between the frequency components, the reality is that this variation can be have a much higher magnitude between locations and seasons than is considered by the authors. What this means that the same call components produced by the same animal may propagate very differently relative to each other in different locations and at different times in the same location.

Furthermore, the authors weigh the received levels of these components by assigning different components different sensitivities based on the hearing curve of the animals. While frequency based hearing sensitivity definitely plays a big role in the detection of the call components it is unlikely that killer whales can detect sound pressure differences linearly but their hearing is dependent on the auditory filter bandwidth that applies to the specific auditory system. Usually, that would be an octave fractal, such 1/3, 1/5, 1/6 or even 1/12 octaves auditory filter bandwidth over which the animal integrate sound pressure. So, while it is definitely true that higher frequency components are detected according to the higher hearing sensitivity of killer whales for those frequencies, we cannot assign a numerical value ( the authors chose 33 dB) as a weighting pressure when comparing perception of low and high frequency components. Since fractal filters are proportional filters, they become wider with increasing frequencies, which means the animals may be able to detect smaller pressure differences in lower frequencies while pressure differences for higher frequencies need to be more pronounced for the animal to perceive.

So, I don’t think the different propagation leads to the described effects in detection differential described by the authors. So, this section on propagation difference would need be revised to include uncertainty in the weighting assumption and the conclusions based on that section should also be revised.

Reviewer #2: Comments PONE-D-20-05337

General

Multiple publications have hypothesized the functions of biphonic calls in resident killer whales through contextual use of these calls and call features using source levels: group identity, contact over large distances, and inferring direction of the caller. This study uses recordings from Russian resident killer whales to test two of these hypotheses: group identity and contact over longer distances.

Throughout the manuscript the author refers to ‘killer whales’ when what is more correct is resident killer whales or fish eating killer whales

The author uses assignment to family group based on the low frequency component (LFC), high frequency component (HFC), and the LFC and HFC together to test the hypothesis that biphonic calls provide information on group identity. Wouldn’t one expect more information usually provide better classification? The LFC appears to account for the majority of correct classification when the two are combined, and the addition of the HFC only minimally improves classification success. Wouldn’t comparing the success rate of classification of biphonic and monophonic calls be more appropriate to test this hypothesis?

The author uses a large set of calls to compare the received levels of the LFC and HFC to test long range detectability each. As mentioned in the discussion, there are many things that can impact the detectability of low and high frequency sounds: direction of the caller, distance to the caller, spreading loss, and background noise. These can impact the LFC and HFC in different directions, therefore having an impact on the relative amplitude of the LFC and HFC. It is unclear how the authors account for this. The large sample size, and recording calls across a variety of scenarios, may be adequate, but the author need to address this.

From the methods presented it is unclear how the adjustment for the hearing threshold was done. My understanding is that the relative amplitude differences of the LFC and HFC were adjusted based at the hearing sensitivity at those frequencies. But this would require absolute received levels of the calls, which we do not have for this study (see above). This section of the methods needs more details.

Specific:

Introduction

Par1,line 2: pod and or community?. Maybe say every group of killer whales or say “In resident killer whales, each killer whale pod….”

Par2,line 3: evolve should be evolves

Par2, line 6: instead of the “- “ start a new sentence

Par3, line 6: ‘Besides’ is awkward wording here. ‘Additionally’ or ‘furthermore’

Par5, line 1-2: again this is the case only for resident killer whales. Others, like Bigg’s and offshores disperse.

Materials and Methods

Similarity of contours of the lower and higher frequency components section:

You mention the number of families, but how many different pods does this represent?

Were the 10 calls quality graded as in Deecke and Janik? Were the calls of highest quality? Adequate signal to noise ratio?

Paragraph 3- is this final number what is referred to as distances in the results? This should be clarified.

Detectability of the lower- and higher-frequency components over distance section

Wouldn’t the direction of the caller impact the amplitude of the HFC of calls more than the LFC?

Recording quality and signal to noise ration can impact the ability to make reliable measurements. If the dataset does not have high levels of background noise that would impact these measurements, some clarification/quality grading/ analysis of impact should be done.

Calls with no detectable HFC… it should be clarified that these weren’t used in the analysis

Results

Detectability of the lower and higher frequency components over distance section:

The mean difference in amplitude between the HFC and LFC is reported but the range or SD should also be reported.

Reviewer #3: I have uploaded my review comments as a separate document because it exceeds 20000 characters (all comments are geared towards providing constructive feedback to help improve the manuscript and arguments). Overall, the manuscript is well written, but would benefit from copy editing to improve syntax and grammar. Please refer to my attachment for further comments and details.

Reviewer #4: Overall this is a valuable study that advances our understanding of this acoustic phenomenon in an important way. The conclusion that “the main function of the higher-frequency component is to duplicate and/or complement the social identity marking when the lower-frequency component is masked by noise or accidentally appears similar in unrelated social units.” Is well supported by the analyses and so I recommend the paper be accepted subject to some changes needed in interpretation and presentation.

My biggest concern is that the Title and second primary conclusion is not adequately justified as ‘independent cultural change’ has not been shown- that can only be done using a temporal study showing how the features have changed over time. The pattern found could just be due to natural variability within otherwise fixed contours – or contours that slowly evolve but not independently. New title could be: Independent acoustic variation of higher-frequency components can facilitate call recognition in killer whales

The manuscript is important without this unsupported claim. You should propose specific future work by which the question of whether or not these call types truly evolve independently is tested more directly and robustly.

A second overall concern is that the write-up doesn’t cite all of the most relevant work in the field. It is important to cite and discuss those published studies that most closely relate to your current work.

Minor suggestions:

Ln 39 - suggest you delete ‘excellent’. The cited study shows rather poor copies of sounds made by one captive killer whale.

Ln 40 add ‘call repertoire’ before dialect to be more precise.

Ln 46-‘evolves’

Ln 49 – no ‘the’ before ‘social’

Ln 55- A paper by Deecke The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca) - seems important to cite here.

Ln 59 or 62. A paper by Miller Caller sex and orientation influence spectral characteristics of “two-voice” stereotyped calls produced by free-ranging killer whales -seems important to cite here.

Ln 149-it is strange to read ‘we’ for a single authored paper.

Ln 158 and Discussion: One highly relevant paper that isn’t cited here looked at error rates of classification for LFC versus HFC: “The influence of social affiliation on individual vocal signatures of northern resident killer whales (Orcinus orca)” It appears that very similar methods and results were found, though on a slightly different social scale.

Methods: consider the possibility that the ‘same family’ results may have been influenced by some of the calls being produced by the same individual.

Ln 225 and 226 – add ‘pairs of’ before ‘calls’

Table1and Figure 2 – what are the units here? This value seems to come from a black box, so make an effort to convince the reader that it is a valid indicator of the similarity.

Table 1- this is a mean of the distances averaged across a lot of pairs of calls, right? If so, state that this is a mean value, and add the standard deviation and sample size to the table.

Table 1 title is confusing- State that is the distance between pairs of calls.

Figure 2-consider to use smaller symbols and open face symbols to better show the data underneath other data points.

Figure 2 – add an x-axis label. Correct top panel y-axis label.

Ln 240-246: Did you ever see cases when only the higher component was visible?

Discussion:

The difference between a signal and the audiogram sensitivity at that frequency is commonly known as the ‘sensation level’. This is a dictionary definition, and could be useful for your paper.

Ln 309 – how much more quickly does the higher frequency component attenuate than the lower frequency component-in dB/km? Is that difference enough to overcome the higher sensation level of the higher component with distance? (personally I think the difference is very small, but it is worth to show you looked at this).

LN 331-333: Again here it’s relevant to cite Miller et al Caller sex and orientation influence spectral characteristics of “two-voice” stereotyped calls produced by free-ranging killer whales -- as it supports the presence of heterodyne frequencies in the two component calls from a totally different population of killer whales.

Ln 336-338. I’d suggest caution here regarding your point on evolution of call types. It is clear that we don’t know the actual mechanism by which these contours are produced, which limits our ability to be certain that each component can be modified independent of the other. There may be aspects of sound production that fundamentally limit the flexibility of one component to change without affecting the other.

Ln 399- all of this evidence is consistent with independent evolution of the two components, but doesn’t demonstrate it directly and rigorously. Alternative drivers of the patterns you found remain possible. Stronger to here to indicate future temporal analyses that would be able to tackle the question more directly.

Acknowledgements reads odd with ‘We’ if this is a single author work.

Overall, very good work.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Harald Yurk

Reviewer #2: No

Reviewer #3: No

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: LineReview_PONE-D-20-05337.docx

Click here for additional data file.^{(24.7KB, docx)}

PLoS One. 2020 Jul 30;15(7):e0236749. doi: 10.1371/journal.pone.0236749.r002

Author response to Decision Letter 0

28 May 2020

Dear Reviewers,

thank you very much for your time and the thorough assessment of my manuscript. I have substantially revised it following your corrections and suggestions. I have changed the title and added the clarification about the ecotype of the study population. I have added the relevant references and added more details to the methods regarding the call recording, selection and analysis. I have expanded the discussion as suggested by the reviewers. Overall, I am very grateful to all of you for your detailed corrections, which I believe have substantially improved the manuscript.

Reviewer #1: The manuscript describes an important bio-acoustical subject, the use of bi-phony (two-voiced calls) by killer whales and suggest possible functions.

The authors provide a number of intriguing suggestions in support of the main topic of the manuscript as stated in the title “Independent cultural change of higher-frequency component can facilitate call recognition in killer whales” and support some of their main conclusions, such as the complementary function of both the lower and higher frequency call components in family recognition. The authors provide support for that conclusion in their investigation of family classification accuracy of two calls of resident type killer whales in the Russian Far East by considering classification means of the two call components separately. The argument that both components can be used independently to identify family membership of the caller is well laid out. The author also provide good support from their data that calls that have both components visible in the spectrograms have higher classification means than those where only have on of the two components is visible thereby validating one of the main conclusions of the study that calls with more than one components, especially those with heterodyne frequencies, have a higher ability to be used as family group identifiers in a variety of acoustic environments.

Response: I agree that our recordings varied greatly in sound propagation properties, but I think that given the large sample size and time span, this is actually a good thing: this means that we likely captured most of the variation that killer whales themselves experience while listening to conspecific sounds, at least in the coastal waters in summer. I have included this explanation into the methods.

Response: This comment would be relevant if I was assessing the ability of whales to detect pressure differences across some frequency range. However, this was not the case – I was looking for the hearing sensitivity of killer whales at some particular frequencies separately. I looked at the audiogram (model audiogram at Fig. 3 in Branstetter et al. 2017) and saw that the threshold at the frequency of the lower component (1.1 kHz) was about 93 dB re 1 µPa, and at the frequency of the higher component (9.7 kHz) it was about 60 dB re 1 µPa. Therefore, the difference in the sensitivity to the lower and higher components should be about 33 dB. I have added more details of this process to the methods.

Reviewer #2: Comments PONE-D-20-05337

General

Throughout the manuscript the author refers to ‘killer whales’ when what is more correct is resident killer whales or fish eating killer whales

Response: I have changed “killer whales” to “resident-type (R-type) killer whales” when appropriate (we try not to use terms ‘resident’ and ‘transient’ when referring to Russian killer whales because it is confusing for our officials, so we adopted the terms ‘R-type’ and ‘T-type’ instead).

Response: Not necessarily. If all HFCs were identical, adding them to the analysis would not improve the classification success. If they had different similarity pattern unrelated to family affiliation (for example if some HFCs in family 1 were more similar to family 2, and other to family 3), adding them would in fact decrease the correct classification rate.

The LFC appears to account for the majority of correct classification when the two are combined, and the addition of the HFC only minimally improves classification success. Wouldn’t comparing the success rate of classification of biphonic and monophonic calls be more appropriate to test this hypothesis?

Response: No, it wouldn’t. The problem is that monophonic and biphonic calls in killer whales differ not just by the presence of the higher-frequency component; they have different source levels, different diversity, and their usage varies depending on the behavioral context, suggesting different functions. Therefore, we can expect that their group-specificity can be also different, and therefore we cannot extrapolate results from monophonic calls to the lower-frequency component of biphonic calls.

Response: I have added a paragraph on this issue to the Methods: “The calls were collected for the analysis at different locations within Avacha Gulf. Spectral sound propagation loss can vary between locations due to differences in bathymetry, sediment structure and water properties; some variation is also introduced by differences in depths at which killer whales produce their calls. To address this, we measured a large amount of calls produced in various conditions and in different situations to cover as much natural variation as possible. As killer whales also experience all these propagation effects when listening to the conspecific calls, we assume that our sample set is an adequate representative of what killer whale normally hear in the wild.”

Response: I have added more details on this to the methods. In short, I estimated the difference in thresholds at the frequency of the lower and higher components, and then added this difference to the received level of the higher component of each call.

Specific:

Introduction

Par1,line 2: pod and or community?. Maybe say every group of killer whales or say “In resident killer whales, each killer whale pod….”

Response: I changed it to “Each resident-type (R-type) killer whale pod”.

Par2,line 3: evolve should be evolves

Response: Changed, thank you!

Par2, line 6: instead of the “- “ start a new sentence

Response: Done.

Par3, line 6: ‘Besides’ is awkward wording here. ‘Additionally’ or ‘furthermore’

Response: Changed to ‘Additionally’.

Par5, line 1-2: again this is the case only for resident killer whales. Others, like Bigg’s and offshores disperse.

Response: I added ‘R-type’ here to clarify this.

Materials and Methods

Similarity of contours of the lower and higher frequency components section:

You mention the number of families, but how many different pods does this represent?

Response: I added this information to the text.

Were the 10 calls quality graded as in Deecke and Janik? Were the calls of highest quality? Adequate signal to noise ratio?

Response: The calls were not quality graded, but we used only the highest-quality calls with all syllables clearly visible.

Paragraph 3- is this final number what is referred to as distances in the results? This should be clarified.

Response: Sorry, I should have clarified this: the distances are the opposite of the similarities, i.e. 100 minus the final number obtained through dynamic time warping. I have added this information to the text.

Detectability of the lower- and higher-frequency components over distance section

Wouldn’t the direction of the caller impact the amplitude of the HFC of calls more than the LFC?

Response: It would, but I assume that given the large sample size, I will get the relatively uniform distribution of caller’s orientation towards the hydrophone, which will be a good approximation of what killer whales normally hear.

Response: Actually, it doesn’t. I took very simple measurements – amplitude in the middle of the lower and higher-frequency components. They are pretty obvious and easily obtained even from calls with low SNR. However, most calls were of reasonable quality, because we normally do not do recordings in presence of loud noise.

Calls with no detectable HFC… it should be clarified that these weren’t used in the analysis

Response: I added this clarification.

Results

Detectability of the lower and higher frequency components over distance section:

The mean difference in amplitude between the HFC and LFC is reported but the range or SD should also be reported.

Response: I have added SDs to the table.

Overall Review Summary:

Overall, the study meets PLOSone’s Publication Requirements. I believe the study provides important insights into the potential functionality of biphonation in resident killer whale vocalizations and will be an important addition to the available literature on the subject – I recommend it be published, after the comments in this review are taken into account. The study is well executed. The field efforts to collect this acoustic data for the study for so many different family groupings is impressive and worthy of recognition. The author does an adequate job of citing and reviewing relevant literature on the subject from multiple taxa throughout the manuscript. The study was performed using established methods that have been published previously, with some modification. I would have liked to see this study include more than one population of resident killer whale to increase scope, such as also including recordings from the Northern Resident population, but this is not essential and the study still contributes in a meaningful way.

Overall, the manuscript is well written, but would benefit from copy-editing to address some issues in grammar and syntax. There are some sections that require re-wording for clarification and interpretability (see line-by-line comments below).

I may use the abbreviations ‘HFC’ and ‘LFC’ in some portions of my review, these just refer to ‘higher frequency component’ and ‘lower frequency component’, respectively.

General comment:

The author makes many references to ‘killer whales’, in general, throughout the manuscript, but the statements really only apply to a given ecotype (most often, ‘resident’ killer whales). And because the author only assesses vocalizations for resident-type killer whales in this study, the conclusions cannot be generalized over all killer whales/types due to the different acoustic behaviour, acoustic population sub-structuring, and vocalizations that other ecotypes exhibit. I suggest the author clarify their use of ‘killer whale’ in the paper to limit it to just the resident ecotype where appropriate.

Response: To clarify this, we have added ‘resident-type’ or ‘R-type’ to the text when appropriate. We are trying to avoid using the term ‘resident’ when referring to Russian residents because this provides too much confusion when discussing the ecotypes with our officials. Therefore, over the last 2-3 years we have been using mostly the terms ‘R-type’ and ‘T-type’ instead of ‘resident’ and ‘transient’ in our papers.

The author also makes many references to ‘pod’ or ‘pods’ throughout the manuscript, but this can be misleading and ambiguous without providing a definition of the term to relate it to the previously published definitions of the term. See comments below for discussion on this.

Response: See the response about the ‘pod’ definition below.

Line-referenced Comments:

Title: I think the title can more accurately reflect the messages of the manuscript. First, it only speaks to the high frequency component, but the paper argues that the presence of both the high- and low-frequency component together provide killer whales with a tool that will ensure transmission of information in many more conditions/situations than either component individually. Second, the title only speaks to ‘call recognition’ but the paper’s main arguments revolve around the information that the calls carry – ie. Social group affiliation at various resolutions. Third, the title generalizes across killer whales, whereas I do not think these results can be interpolated beyond the ‘resident’ ecotype. I may recommend the following title (or something similar) to more accurately reflect the paper’s outcomes/arguments and reach: “Independent culturally-induced change of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in resident killer whales”. This title more accurately represents the author’s conclusions in the discussion, and the fact that this really only relates to the resident killer whale ecotype (until similar research is conducted on the other ecotypes the use of the biphonation cannot be assumed to be the same without further study because different ecotypes seem to have different acoustic population sub-structuring, and thus the information communicated with the biphonation and the contexts it is used could be quite different).

Response: I changed the title to “Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales”. I removed the reference to cultural change from the title according to the comment of reviewer #4.

Line 12: the author’s use of the term ‘pod’ needs to be defined. See further comments on this topic below.

Response: See the response about the ‘pod’ definition below.

Lines 28 and 29: Change “…component allows to recognize the family identity…” to “…component allows the recognition of family identity of a caller…”.

Response: Changed.

Line 40: The author uses the term ‘pod’ throughout the manuscript, however, the meaning of ‘pod’ is ambiguous. The author’s intended meaning of the term ‘pod’ needs to be defined. “Pod” has several meanings in the context of killer whales. “Pod” can be used in the traditional sense to describe a group of cetaceans (like a ‘herd’ of cattle, or ‘flock’ of birds). However, in killer whales, ‘pod’ was also defined by Bigg et al. 1990 as a functional social unit within the nested social structure of resident killer whales in British Columbia, as “A subpod made of groups who spend more than 50% of the time together“. Although this definition has been adopted by many researchers monitoring other killer whales around the world to describe social groupings, Ford and Ellis (2002, Reassessing the social organization of resident killer whales in British Columbia) determined that the ‘pods’ within Northern Resident Killer whales that were grouped base on the definition in Bigg 1990 no longer met this definition of ‘pod’ after ~30 years of monitoring the population (ie. a “pod” according to the social definition in Bigg et al 1990 is not a stable social grouping, and thus the use of the term ‘pod’ in this context should no longer be used). The author should define their intended meaning of ‘pod’ throughout the paper, or use a different descriptor. Or if the social groupings in their study population still adhere to Bigg’s definition of ‘pod’, then describe this (although, as mentioned, even though the population may still adhere to the definition, it likely won’t always, and a different term should be used).

Response: Well, it is a good point. Actually, we had found that we cannot use Bigg’s 50% definition of pod for our population, because the associations between matrilineal units are more fluid: some units can spend 90% of time together during one summer and only 10% of time together next summer. Therefore, we rather adopted Ford’s dialect-based definition of pod as a set of matrilines that share the same dialect. I referred to this definition in the Methods: “Families that share the same vocal dialect are attributed to the same pod, and pods with similar dialects form clans.” To clarify the term earlier in the text, I added the dialect-based definition to the Introduction: “Each R-type killer whale pod shares the same dialect”

Line 52: “New pods form gradually through the split of an ancestral pod after a matriarch dies…”. First, the use of ‘pod’ needs to be defined or not used (use ‘group’ instead), as described in comment above. Second, this is not the only mechanism that drives the creation of new groups in resident killer whales. New groups also form through group-splitting that takes place while the maternal ancestor is still alive (intra-geneological splitting), as described in Stredulinsky 2016: Determinants of Group Splitting in a Threatened Population of Fish-eating Killer Whales. Group splitting is a more complex process than this sentence suggests, and as such, so is dialect evolution.

Response: I have deleted “after a matriarch dies” statement. Now it reads “New pods form gradually through the split of an ancestral pod, and the dialects of the splitting pods slowly diverge in time due to changes in call structure”. Of course it is a simplistic description, but it is accurate enough to give an idea of the process to a reader who is new to the field, and provides enough background to comprehend the further inferences in the text.

Line 73-75: Strongly suggest a change in wording. Natal philopatry is not present in all types of killer whales (therefore the author should only relate this sentence to resident killer whales, not killer whales in general). Also, I would argue that family life is not ‘critical’ for survival in killer whales. Family cohesion has some very notable benefits that make maintaining family cohesion more advantageous than not in many killer whale societies, but it is not critical. There are some killer whale societies where emigration out of one’s family is important (such as in the West Coast Transient “Bigg’s” killer whale population). The author should refrain from generalizing across killer whales generally, and try to stick within the study population (resident killer whales).

Response: I have added ‘R-type’ here to clarify that it applies only to resident-type killer whales, and replaced ‘critical’ with ‘important’.

Line 77: ‘Intentionally’ and ‘deliberately’ are synonyms.

Response: I have deleted “intentionally”.

Line 113-114: Were the hydrophones calibrated across frequencies? What method did the author use to determine the reportedly flat response (manufacturers documentation/technical specifications of the equipment? End-to-end system calibration including recording equipment? Hydrophone-only calibration?). If the hydrophones were not calibrated, this needs to be stated, with an indication as to how the flat response was determined.

Response: For the Offshore Acoustic hydrophone, we rely on the manufacturer’s specifications, and the CetaPhone hydrophone was calibrated by us in combination with Zoom H4 flash-recorder. However, I checked specifically the recordings I used for this study and found that they were all made with the Offshore Acoustic hydrophone. I have added its manufactuter’s specifications to the Methods.

Line 108-135: Were any observations about group behavioural state collected during the recordings? As noted in previous publications (e.g. Ford 1990), behavioural state can affect how whales produce vocalizations, with more aberrant versions of stereotyped calls being produced during periods of socializing. Since behaviour affects how calls are produced, it is possible that call similarity would be influenced by behavioural state. If some families included in the study were only recorded during bouts of intense socializing, the recorded calls may not accurately represent the structure of the stereotyped calls made by those families in other behavioural states more conducive to very stereotyped call production (such as foraging or consistent travel, etc). To truly compare similarity of call structure amongst family groups, shouldn’t behavioural state also be incorporated into the analysis as a variable?

Response: Group behavioral state was noted during the acoustic recordings, but it was not accounted for during this analysis, because it was not necessary. Aberrant calls reported by Ford (1990) are easily discernable from the normal versions of stereotyped calls, and I did not use them for this study.

Line 120: Did the author use a catalog to identify the individuals? Is the catalog citable? It should be cited if so.

Response: I added reference to our catalog ‘The Killer Whales of Eastern Kamchatka’.

Line 139-141: Why 10 calls per call type per group? Is 10 calls enough to capture the variability within the calls adequately? How were the 10 calls per type per group chosen to make up the dataset (were selection criteria based on quality or SNR (or other – like the first ten calls of each type in each recording of each family? The time-warping contour selection is too laborious for a larger sample size?))? Please state reasoning.

Response: Ten call per group were selected as a reasonable number of calls I can obtain from a reasonable number of groups. If I decide to use 20 calls per group, I will be able to get this sample size from only four or five families. It is surprisingly difficult to obtain a sample of even just 10 good-quality calls reliably assigned to a particular family. Killer whales are more vocal in presence of other families – a context that makes call assignment to particular family rather difficult. For example, we have many encounters of two-three families of the same pod with hundreds of great good-quality calls, but we cannot use them, because we do not know which of the families produced them (and as the families are from the same pod, their dialects look identical to me, so I can’t just reveal it by the call structure). We also have a lot of single-family encounters during which the whales were almost totally silent – they usually are not very vocal when only their own family is present. For this reason, I had to limit my sample size to 10 calls per family.

Did any examinations take place that looked at the overall quality of the final selected calls by family group? It seems to me that better quality calls may lead to better measurements in the dynamic time-warping analysis. If some groups had a higher proportion of poor quality calls, this may affect the comparisons.

Response: All the calls used for the contour-tracking were extremely good quality calls (which is also why I could get so few of them). These calls were initially selected for the comparison of different syllables for my previous paper, and this required very good quality calls with all syllables clearly visible.

The manuscript and interpretability of the results would be improved by including a few more details in the methods about the data set selection process.

Response: I have added more details on the call selection process.

Line 141: remove ‘each’ from ‘per each family’.

Response: Removed.

Line 142-146: Is the custom MATLAB contour extraction tool previously published? If so, please cite, or provide a few more details on the method: how did the operator determine they’ve selected enough points to adequately describe the contour? If the operator has to do this for each call, this may introduce subjectivity. Could this subjectivity bias the contour shape slightly of each call? If you repeated the point selection and extraction process 10 times from the exact same call clip would the resulting contours be the exactly the same? If not, how much variability exists among these contours?

Response: The MATLAB script was published for the first time in Filatova et al. 2012. That paper includes more detailed description of the contour extraction process, so I added this reference to the text.

Regarding the subjective variability of contours, I did not assess this, but I expect it to be very low, because the operator can see the resulting contour on the background of the spectrogram and make sure that the contour follows the call precisely. Therefore, the variation does not exceed several Hz, which is a normal error level even when you measure call parameters with AviSoft SASLab or another similar software.

Also, these methods need to be clearer on how the high-frequency and low-frequency contours were treated. Were they selected and extracted independently of each other (so there would be a HFC contour and a LFC contour for each call clip?)? In Line 168-169 the author states: “..and for the contours that consisted of both the components taken together.” Does this mean that the analyst selected points to extract the HFC and LFC as one contour containing features of both? More clarity on the contour extraction process and how the contours for both components were treated is needed to help interpretability and reproducibility.

Response: Yes, the LFC and the HFC contours were extracted independently. When I say “contours that consisted of both the components taken together“ I mean that for this analysis the LFC and HFC contours were concatenated horizontally to form a single contour that included both the LFC and HFC contours – this was necessary to assess the rate of correct classifications when both components were used.

Line 149-150: The author should briefly state why adopting a modified approach was necessary.

Response: The modification was made for our previous analyses and it was described in detail in Filatova et al. 2012: “Because the algorithm of Deecke & Janik (2006) only allows expansion or compression of the time axis by a factor of three, the algorithm cannot be used to compare calls that differ in length by more than a factor of three. In this case, their similarity is considered 0%. This constraint biased the results in comparisons where many short or long contours were present in the repertoire of one population but not the other. To avoid this, we developed an additional algorithm that stretched the shorter contour through interpolation to make it one point longer than a third of the longer contour.” For the current study this modification in fact was not necessary because we compared the calls of the same type, and none of them differed in length by more than a factor of three. We used the modified algorithm only because it was already modified and there was no point to change it back because it does not change anything when the calls have similar length. For this reason, to avoid confusion I have deleted ‘modified’ from the text, because the results did not differ from those that would be obtained by the original algorithm.

Line 150-155: This description needs to be clearer. It could be made clearer if supplemented by a figure of a contour and an example output from the time-warping algorithm to illustrate what they mean by “smaller frequency value by the larger frequency value at each point” and “a running tab was kept on the similarities of elements making up the curves while adding up these costs…”. More detail into how the cost matrix works would be great. A figure with a flow diagram of how the values from the time-warping get turned into percent similarity, then combined into the cost matrix resulting in a single value.

Response: I did not develop this algorithm and it was described in detail by its authors in Deecke and Janik 2006, which I refer to. The algorithm itself is available online (for example, in the appendix of of Volker Deecke’s thesis). Also, there are millions of illustrations of dynamic time warping process easily available to anyone who can use Google search. Therefore, I believe that describing the details of the algorithm with figures in this particular manuscript would be definitely an overkill.

Line 158-160: This sentence is an exact replica of a sentence in the methods of Deecke and Janik 2006, citation [34] in this manuscript. This sentence must be re-worded to avoid plagiarism.

Response: Changed.

Line 161-162: the original ARTwarp algorithm was not developed by Deecke and Janik [34], but as Deecke and Janik [34] describe, it was originally developed by Carpenter and Grossberg (1987), and modified for use by Deecke and Janik.

Response: I changed the wording to “ARTwarp algorithm used by Deecke and Janik”.

Line 176-177: This part of the paper is very interesting and has the potential to provide great insight into how these two call components may be perceived and used by killer whales. However, using relative amplitude of the two components is only valid if the recording system’s frequency response is truly flat. The author previously mentions in the methods that the hydrophones used have a flat response, but provide no information for how that was determined. The detection range comparison compares the relative amplitude across frequencies, and thus, if the recording system does not actually have flat response, the relative contribution of the high and low frequency components will not reflect reality. For example, if the recording system is actually more sensitive in the 5-8kHz band than in the ~1kHz band, the amplitude of the high-frequency component would have been over-represented by assuming a flat response. The author should provide more information on how their recording system’s flat response was determined/verified (end-to-end calibration? hydrophone-only calibration? Manufacturers technical documentation? , etc).

Response: As stated above, I used the manufacturer’s specifications provided with the hydrophone. The sensitivity of the hydrophone was flat within 3 dB range up to 14 kHz, which is well above the highest of the high frequency component in this study. Three dBs is an acceptable error level given the approximate nature of other parameters, including the averaged killer whale audiogram.

Line 188: The author cites two papers as the source for the audiogram information used. However, the two papers present many different audiograms from many different individual killer whales (and even the composite audiograms presented in the two papers are slightly different). For reproducibility, the author must state which specific audiogram was used.

Response: I used the ‘model’ audiogram from Branstetter et al (2017) for the calculations. I have added this information to the text.

Line 217-221: Figure 2: Figure 2’s axes need to be labelled. Only the y-axis is labelled. Also, the x-axis label has a spelling error.

Response: I have fixed the axes labels.

Line 242: “All these calls…” This should be more specific, such as: “All of the calls without detectable high-frequency components…”

Response: Changed.

Line 271: Discrimination of a caller’s orientation, not ‘direction’. The direction to a caller is not related to the high-frequency component’s directionality (their hearing is directional and thus can infer direction to a caller inherently). The high-frequency component gives information about the orientation of the caller relative to the receiver.

Response: Changed to ‘orientation’.

Line 276-278: Direction of movement does not necessarily equate to the direction of calling. The whale could temporarily face a certain direction to emit a call, then continue on it’s intended path. Also, the high-frequency component is not necessary for determining direction of travel. The caller only needs to emit two omnidirectional sounds of any type while travelling and the listener can discern direction of travel. In this context, the higher frequency component’s directionality can only provide the caller’s orientation. And the orientation information is quite ambiguous, especially considering the 4D environment the whales live in and the number of possible orientations that are possible for the same amplitude ratio of the high to low frequency component. And on top of this, the whales make other types of directional sounds, so the HFC isn’t the only vocalization that can elude to orientation of the caller (I think echolocation is likely a better indication of orientation because it is made very frequently and consistently and the returning echoes from features can also be used…).

This is all to say that the author can strengthen their argument that the HFC is not solely used for orientation/group coordination using the above points. Also, the discussion can be improved by speaking to the other potential reasons for the directionality in the high-frequency component. The directionality of the HFC can be more useful than for just providing orientation information: the directionality of the component would allow a caller to emit family/clan membership information towards specific whales or groups within an aggregation selectively (just as a human in a noisy room would yell (everyone in the room would hear), but ALSO make eye contact (directional component) with a person on the other side of the room if they were the intended target of such communication). The directionality can also help increase the SNR in the direction of a receiver in noisy conditions to be heard better in noise – not just the simple presence of the high frequency component, because that can be masked quite easily in some situations, but the ability to create a higher SNR by directing the call towards other members of the group in noisy conditions).

I think the directionality aspect of the HFC is a very important functionality of the HFC and should be discussed more in the discussion (because it ultimately helps out the author’s arguments and final conclusions). Being able to provide family membership more reliably is fine, but the ability to provide family membership in certain directions is a very useful tool for a very social species (including in noisy conditions).

Response: Thank you! I have added more discussion of the directionality aspects, as well as the argument that other sounds are directional too.

Line 282-284: I don’t think this is necessarily true. The structure of a tone (whether it is frequency modulated or not) can also affect its detectability in different noise conditions. There are types of background noise where a non-frequency modulated tone would be more detectable than a modulated one, but there are also situations where a modulated tone would be more detectable than a non-modulated one. This argument needs to be strengthened.

Response: I have shifted the focus of this statement to address the reviewer’s comment, now it reads “there would be no need to vary its shape across families: the same contour for all families would work equally well”.

Line 290-292: I think this sentence seems a little abrupt, as these uses of biphonation haven't been brought up in the paper before, and also not in relation to killer whales - the abruptness hides the intended purpose of the sentence. I suggest changing the sentence to make it less abrupt, and capture the attempted purpose of the sentence: "Additional hypotheses for the function of biphonation in other species have included: increasing unpredictability and providing an indication of physical condition [19]. However, these are likely not appropriate considerations as possible functions of biphonation in killer whales. These hypotheses suggest that the caller can choose when to include or not include a biphonic component in each call it makes. However, calls that comprise killer whale dialects are stereotyped and ..." Or something like that.

Response: Thank you! I have included this in the text.

Line 294: The use of the term ‘mandatory’ is slightly misleading. Yes, in the specific call types that contain HFC’s, the HFC component is mandatory (the whale does not choose to whether include or remove the component), but the way the sentence is written it could be interpreted that the HFC is mandatory in any call the whale makes.

Response: To clarify this, I added “if a call is biphonic” to the text.

Line 297: “probably on the family level…” Why only at the family level? In this manuscipt, the author only examined the frequency-time component of the call components, but more information is provided in a vocalization, such as it’s timbre? Could this not carry individual information in the HFC/LFC? There needs to be a citation for why the author believes only family level is possible in killer whales – there is arguably strong pressure for individual recognition in resident killer whales – such as in kin-directed prey-sharing, Wright et al. 2017.

Response: I strongly agree that killer whales can recognise their family members individually by their calls, but this sentence only speaks about the high-frequency component “Most likely, the higher-frequency component of killer whale biphonic calls functions as an alternative contour that bears information on the caller’s identity (probably on the family level, in contrast to individual level in dholes and penguins).” It does not imply that killer whales can’t recognize each other on the family level.

Line 312: ‘shifted’ may not be the appropriate word. “concentrated” or “is more pronounced in lower frequencies”, etc. The noise isn’t shifted to lower frequencies, it is just the nature of most anthropogenic ambient noise.

Response: Changed to “more pronounced in lower frequencies”, thank you!

Line 319-320: The author should also consider addressing this argument in relation to Viers et al 2016: “Ship noise extends to frequencies used for echolocation by endangered killer whales”. It should be noted that, at times, even the HFC can be masked by ship/anthropogenic noise. But the author can also add another advantage of the directionality of the HFC in this context. If noise levels are high, even to the point of starting to mask the HFC, the caller can point in the direction of the receiver while calling to increase the HFC’s SNR. This all adds to the potential function of this component.

Response: I have added the reference to Viers et al 2016 and the possible additional function of directionality.

Line 324-325: “…better hearing sensitivity of killer whales to higher frequencies.” Cite [35, 36] here.

Response: Done.

Line 338: This is a very cool possibility, but I wonder if it is meaningful biologically? The heterodyne frequencies only show up when call is good quality (high SNR or close), and in this sense, it may not be that effective when in noise because to resolve it the whales need to be relative close to one another, and if they are relatively close, they will already be able to discern the low frequency component?

Response: This is probably true in most situations, but still I can imagine situation when there is very intensive noise on low frequencies that masks the lower-frequency component but the higher-frequency component and the heterodynes are audible far enough on the caller’s axis. Anyway, it is just a speculation, but I think an interesting one.

Line 357: “pods” should be replaced by “groups” or “social units” for clarity.

Response: Changed to ‘social units’.

Line 362: “This process of cultural change is called cultural evolution.” The author is not defining what cultural evolution is for the first time in this paper. Please cite.

Response: I added references to Mundinger (1980) and Lumsden & Wilson (1985).

Line 363: use of ‘pods’ needs to be defined or changed to ‘groups’.

Response: Changed to ‘social units’.

Line 368: “…probability that both components would converge in two pods is very low.” The probability is likely relative to whether the two ‘pods’ are members of the same or different populations. And if in part of the same population, relative to the overall population size and the frequency that different groups in that population encounter one another.

Response: No, I mean random convergence here – the one that occurs just because the number of contour shapes is limited, and therefore there is always a probability that in the process of cultural change two contours can randomly become more similar.

Line 371-377: This paragraph would benefit from re-wording to provide better clarity.

Response: I have re-worded this paragraph.

Line 402-405: I think these conclusions can benefit from a more thorough discussion of the advantages of directionality (as mentioned previously in these comments).

Response: I have added the statement about the directionality here.

Supplemental Information:

S1: I noticed that the matrices for the K5 and K7 call types are different sizes. Why are the K5 datasets 150x150 whereas the K7 datasets are 140x140? In the methods the author describes data collection as: 14 families, 10 calls of each type per family, so shouldn’t they both be 140x140? I may have misunderstood something? If not, then the author should address this in the methods.

Response: Sorry, I should have clarified it in the methods. One of the families had two distinctive subtypes of K5 calls, and I used them both for the analysis, so from this family I selected 20 calls (10 from one subtype and 10 from another). I have added these details to the methods.

Response: I have changed ‘cultural change’ to ‘acoustic variation’ in the title and the abstract. In the Discussion, the independent cultural change is mentioned in the context of other papers that have demonstrated the change of call features over time, so I believe it is justified there.

Response: I have added the suggested citations.

Minor suggestions:

Ln 39 - suggest you delete ‘excellent’. The cited study shows rather poor copies of sounds made by one captive killer whale.

Response: Deleted.

Ln 40 add ‘call repertoire’ before dialect to be more precise.

Response: Added.

Ln 46-‘evolves’

Response: Changed.

Ln 49 – no ‘the’ before ‘social’

Response: Deleted.

Ln 55- A paper by Deecke The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca) - seems important to cite here.

Response: Added.

Ln 59 or 62. A paper by Miller Caller sex and orientation influence spectral characteristics of “two-voice” stereotyped calls produced by free-ranging killer whales -seems important to cite here.

Response: Added.

Ln 149-it is strange to read ‘we’ for a single authored paper.

Response: I have changed ‘we’ to ‘I’ where appropriate.

Response: Thank you for drawing my attention to this paper! I’ve read it before, but I have forgotten that it had compared the LFC and the HFC in the similar way to what I did. Unfortunately, that paper is very brief and does not provide enough details to compare our results. Nevertheless, I mentioned it in the Discussion.

Methods: consider the possibility that the ‘same family’ results may have been influenced by some of the calls being produced by the same individual.

Response: It is definitely possible, but there is no way around it because many of the families had less than 10 whales, so there is no way to select 10 calls from such family without some degree of pseudo-replication.

Ln 225 and 226 – add ‘pairs of’ before ‘calls’

Response: Added.

Table1and Figure 2 – what are the units here? This value seems to come from a black box, so make an effort to convince the reader that it is a valid indicator of the similarity.

Response: Technically, these are percent. The dynamic time warping gives the result as % similarity, and to calculate distances I subtracted the similarity values from 100%. I added these details to the Methods.

Table 1- this is a mean of the distances averaged across a lot of pairs of calls, right? If so, state that this is a mean value, and add the standard deviation and sample size to the table.

I have added the SD and stated that the table has mean +-SDs in the title. However, I do not think it makes sense to add sample size. The sample size here is the number of pairs, which is much larger than the number of calls. Adding the number of pairs looks like inflating the real sample size, while adding the number of calls would be confusing because the values were calculated from the number of pairs.

Table 1 title is confusing- State that is the distance between pairs of calls.

Response: Done.

Figure 2-consider to use smaller symbols and open face symbols to better show the data underneath other data points.

Response: I have made the symbols twice smaller. Open face symbols are ugly; I think the semi-transparency is good enough to show the points underneath.

Figure 2 – add an x-axis label. Correct top panel y-axis label.

Response: Done.

Ln 240-246: Did you ever see cases when only the higher component was visible?

Response: No. I would expect this to happen when the lower component is masked by intensive low-frequency ship noise, but we usually don’t do recordings when there are loud ships nearby.

Discussion:

The difference between a signal and the audiogram sensitivity at that frequency is commonly known as the ‘sensation level’. This is a dictionary definition, and could be useful for your paper.

Response: Thank you very much! I have used it where appropriate.

Response: I checked this before and frequency-dependent absorption differences were indeed rather low compared to killer whale sensitivity differences to higher frequencies. For example, the difference between 1 and 10 kHz is about 0.7 dB per km. Given that normally we record killer whales within few kilometers, the difference would be only 1-2 dBs. I am not sure I should mention this in the text, because anyone who knows about the frequency-dependent absorption would also probably realize that its influence is low, like you do.

Response: Reference added.

Response: Not sure which lines this comment referred to, because lines 336-338 were actually about heterodynes, not about evolution. Anyway, I agree that we don’t know the actual mechanism, but the fact that their variation is independent directly follows from the results. If the flexibility of one component was limited, then the distances between the pair of its contours would be much lower than of the other non-limited component, but the results show that the distance variation of both components is at about the same scale.

Response: I have added these two sentences here: “However, despite all of this evidence is consistent with independent evolution of the two components, it does not demonstrate it directly and rigorously. In future, analysis of the component variation over time is necessary to exclude the alternative drivers of the observed patterns.”

Acknowledgements reads odd with ‘We’ if this is a single author work.

Response: Changed to “I”.

Overall, very good work.

Response: Thank you!

PLoS One. doi: 10.1371/journal.pone.0236749.r003

Decision Letter 1

William David Halliday

24 Jun 2020

PONE-D-20-05337R1

Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

PLOS ONE

Dear Dr. Filatova,

The authors have done a very good job of addressing the original reviews. Two of the original reviewers (2 and 4) reviewed the revised manuscript, and Reviewer 4 has a few minor suggestions that could further improve the manuscript. Please respond to these additional comments. Following this, the manuscript should be acceptable for publication.

Please submit your revised manuscript by Aug 08 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

William David Halliday, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

Reviewer #4: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #2: Yes

Reviewer #4: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #2: Yes

Reviewer #4: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #2: Yes

Reviewer #4: Yes

**********

6. Review Comments to the Author

Reviewer #2: (No Response)

Reviewer #4: I have a few minor comments that could improve the manuscript further.

Ln 26 – could change ‘perceived amplitude’ to ‘sensation level’ to use standard terminology

Ln 273 – this section has ‘over distance’ in the title, but no consideration of propagation effects or distance over which they might be used are mentioned here. Consider to change the title of this section. Alternatively, you could mention that frequencies of the HFC will absorb roughly 1dB/km more than the LFC and consider the consequences of this difference on the typical detection ranges indicated in Miller et al., 2006 – At 10km distance, the HFC will have lost 10dB more than the LFC.

Ln 279 – the HFC may also not be detectable if the whales were oriented away

Ln 338: could change ‘mandatory’ to ‘obligatory’

Ln 457: change ‘the main’ to ‘a likely’ – until we are able to test these different ideas using playback experiments all of the proposed functions remain somewhat speculative.

Ln 461: the directionality means that this will not be audible as far to receivers that aren’t ahead of the signaller. This is a potential drawback to its use as a general family indicator, so it may not really be ‘especially suitable’. The combination of both components does provide a redundancy of information that is beneficial for these animals to maintain contact with preferred group members.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Reviewer #4: No

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2020 Jul 30;15(7):e0236749. doi: 10.1371/journal.pone.0236749.r004

Author response to Decision Letter 1

11 Jul 2020

Reviewer #4: I have a few minor comments that could improve the manuscript further.

Ln 26 – could change ‘perceived amplitude’ to ‘sensation level’ to use standard terminology

Response: Changed.

Response: I have added the explanation about the frequency-dependent absorption and the reference to Miller (2006).

Ln 279 – the HFC may also not be detectable if the whales were oriented away

Response: If the whale is close, the HFC will be detectable even if it is oriented away (I have seen it many times when whales were passing our boat, calling first towards, and then away from the hydrophone – the HFC was weaker in the latter case, but still detectable). Therefore, the orientation away decreases the received level of the HFC, but it becomes undetectable due to transmission loss, which is already mentioned in the text.

Ln 338: could change ‘mandatory’ to ‘obligatory’

Response: Changed.

Ln 457: change ‘the main’ to ‘a likely’ – until we are able to test these different ideas using playback experiments all of the proposed functions remain somewhat speculative.

Response: Changed.

Response: I have deleted the statement that the directionality makes these calls ‘especially suitable’ and added the statement that combination of both components provides a redundancy that is beneficial in their environment.

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(15KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0236749.r005

Decision Letter 2

William David Halliday

14 Jul 2020

Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

PONE-D-20-05337R2

Dear Dr. Filatova,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

William David Halliday, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

This manuscript can now be accepted. I did find one very minor spelling error on line 284: "loose" should be "lose". This can be changed prior to production.

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0236749.r006

Acceptance letter

William David Halliday

17 Jul 2020

PONE-D-20-05337R2

Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

Dear Dr. Filatova:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. William David Halliday

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Table. Similarity matrices for the lower- and higher-frequency components of K5 and K7 calls.

(XLSX)

Click here for additional data file.^{(945.9KB, xlsx)}

S2 Table. Measurements of the amplitude of the lower- and higher-frequency components of K5 and K7 calls.

The table also shows the time stamp of the measurement within a file, and the frequency at which the amplitude was measured.

(XLSX)

Click here for additional data file.^{(62.3KB, xlsx)}

Attachment

Submitted filename: LineReview_PONE-D-20-05337.docx

Click here for additional data file.^{(24.7KB, docx)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(15KB, docx)}

Data Availability Statement

All relevant data are within the manuscript and its Supporting Information files.

[pone.0236749.ref001] 1.Abramson JZ, Hernández-Lloreda MV, García L, Colmenares F, Aboitiz F, Call J. Imitation of novel conspecific and human speech sounds in the killer whale (Orcinus orca). Proc Royal Soc B Biol Sci. 2018; 285(1871): 20172171 10.1098/rspb.2017.2171 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref002] 2.Ford JKB. Vocal traditions among resident killer whales (Orcinus orca) in coastal waters of British Columbia. Can J Zool. 1991; 69: 1454–1483. 10.1139/z91-206 [DOI] [Google Scholar]

[pone.0236749.ref003] 3.Rendell L, Whitehead H. Culture in whales and dolphins. Behav Brain Sci. 2001; 24(2): 309–324. 10.1017/s0140525x0100396x [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref004] 4.Laland KN, Hoppitt W. Do animals have culture? Evol Anthropol. 2003; 12(3): 150–159. 10.1002/evan.10111 [DOI] [Google Scholar]

[pone.0236749.ref005] 5.Filatova OA, Ivkovich TV, Guzeev MA, Burdin AM, Hoyt E. Social complexity and cultural transmission of dialects in killer whales. Behav. 2017; 154(2): 171–194. 10.1163/1568539X-00003417 [DOI] [Google Scholar]

[pone.0236749.ref006] 6.Deecke VB, Barrett-Lennard LG, Spong P, Ford JKB. (2010). The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca). Naturwissenschaften. 2010; 97:513–518. 10.1007/s00114-010-0657-z [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref007] 7.Miller PJO. Diversity in sound pressure levels and estimated active space of resident killer whale vocalizations. J Comp Physiol A. 2006; 192: 449–459. 10.1007/s00359-005-0085-2 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref008] 8.Miller PJO, Samarra FI, Perthuison AD. Caller sex and orientation influence spectral characteristics of “two-voice” stereotyped calls produced by free-ranging killer whales. J Acoust Soc Amer. 2007; 121: 3932–3937. 10.1121/1.2722056 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref009] 9.Filatova OA, Fedutin ID, Nagaylik MM, Burdin AM, Hoyt E. Usage of monophonic and biphonic calls by free-ranging resident killer whales (Orcinus orca) in Kamchatka, Russian Far East. Acta Ethol. 2009; 12:37–44. 10.1007/s10211-009-0056-7 [DOI] [Google Scholar]

[pone.0236749.ref010] 10.Foote AD, Osborne RW, Hoelzel AR. Temporal and contextual patterns of killer whale (Orcinus orca) call type production. Ethol. 2008; 114:599–606. 10.1111/j.1439-0310.2008.01496.x [DOI] [Google Scholar]

[pone.0236749.ref011] 11.Weiß BM, Symonds H, Spong P, Ladich F. Intra-and intergroup vocal behavior in resident killer whales, Orcinus orca. J Acoust Soc Am. 2007; 122: 3710–3716. 10.1121/1.2799907 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref012] 12.Wilden I, Herzel H, Peters G, Tembrock G. Subharmonics, biphonation and deterministic chaos in mammal vocalization. Bioacoustics. 1998; 9:1 71–196. 10.1080/09524622.1998.9753394 [DOI] [Google Scholar]

[pone.0236749.ref013] 13.Riede T, Herzel H, Mehwald D, Seidner W, Trumler E, Tembrock G, et al. Nonlinear phenomena and their anatomical basis in the natural howling of a female dog-wolf breed. J Acoust Soc Am. 2000; 108: 1435–1442. 10.1121/1.1289208 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref014] 14.Volodin IA, Volodina EV. Biphonation as a prominent feature of dhole Cuon alpinus sound. Bioacoustics 2002; 13: 105–120. 10.1080/09524622.2002.9753490 [DOI] [Google Scholar]

[pone.0236749.ref015] 15.Fischer J, Hammerschmidt K, Cheney DL, Seyfarth RM. Acoustic features of female chacma baboon barks. Ethol 2001; 107: 33–54. 10.1111/j.1439-0310.2001.00630.x [DOI] [Google Scholar]

[pone.0236749.ref016] 16.Brown CH, Alipour F, Berry DA, Montequin D. Laryngeal biomechanics and vocal communication in the squirrel monkey (Saimiri boliviensis). J Acoust Soc Am. 2003; 113: 2114–2126. 10.1121/1.1528930 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref017] 17.Riede T, Owren MJ, Arcadi AC. Nonlinear acoustics in pant hoots of common chimpanzees (Pan troglodytes): frequency jumps, subharmonics, biphonation, and deterministic chaos. Am J Primatol 2004; 64:277–291. 10.1002/ajp.20078 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref018] 18.Tyson RB, Nowacek DP, Miller PJO. Nonlinear phenomena in the vocalizations of North Atlantic right whales (Eubalaena glacialis) and killer whales (Orcinus orca). J Acoust Soc Am. 2007; 122(3): 1365–1373. 10.1121/1.2756263 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref019] 19.Quick N, Callahan H, Read AJ. Two‐component calls in short‐finned pilot whales (Globicephala macrorhynchus). Mar Mam Sci, 2018; 34(1): 155–168. 10.1111/mms.12452 [DOI] [Google Scholar]

[pone.0236749.ref020] 20.Fitch WT, Neubauer J, Herzel H. Calls out of chaos: the adaptive significance of nonlinear phenomena in mammalian vocal production. Anim Behav 2002; 63: 407–418. 10.1006/anbe.2001.1912 [DOI] [Google Scholar]

[pone.0236749.ref021] 21.Aubin T, Jouventin P, Hildebrand C. Penguins use the twovoice system to recognize each other. Proc R Soc Lond B Biol Sci. 2000; 267: 1081–1087. 10.1098/rspb.2000.1112 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref022] 22.Volodina EV, Volodin IA, Isaeva IV, Unck C. Biphonation may function to enhance individual recognition in the dhole, Cuon alpinus. Ethol. 2006; 112(8): 815–825. 10.1111/j.1439-0310.2006.01231.x [DOI] [Google Scholar]

[pone.0236749.ref023] 23.Kriesell HJ, Elwen SH, Nastasi A, Gridley T. Identification and characteristics of signature whistles in wild bottlenose dolphins (Tursiops truncatus) from Namibia. PLoS One. 2014; 9(9): e106317 10.1371/journal.pone.0106317 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref024] 24.Papale E, Buffa G, Filiciotto F, Maccarrone V, Mazzola S, Ceraulo M, et al. Biphonic calls as signature whistles in a free-ranging bottlenose dolphin. Bioacoustics. 2015; 24(3): 223–231. 10.1080/09524622.2015.1041158 [DOI] [Google Scholar]

[pone.0236749.ref025] 25.Deecke VB, Ford JKB, Spong P. Dialect change in resident killer whales: implications for vocal learning and cultural transmission. Anim Behav. 2000; 60: 629–638. 10.1006/anbe.2000.1454 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref026] 26.Yurk H, Barrett-Lennard L, Ford JKB, Matkin CO. Cultural transmission within maternal lineages: vocal clans in resident killer whales in southern Alaska. Anim Behav. 2002; 63: 1103–1119. 10.1006/anbe.2002.3012 [DOI] [Google Scholar]

[pone.0236749.ref027] 27.Strager H. Pod-specific call repertoires and compound calls of killer whales, Orcinus orca Linnaeus, 1758, in the waters of northern Norway. Can J Zool. 1995; 73: 1037–1047. 10.1139/z95-124 [DOI] [Google Scholar]

[pone.0236749.ref028] 28.Wellard R, Pitman RL, Durban J, Erbe C. Cold call: the acoustic repertoire of Ross Sea killer whales (Orcinus orca, Type C) in McMurdo Sound, Antarctica. Royal Soc Open Sci. 2020; 7(2): 191228 10.1098/rsos.191228 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref029] 29.Filatova OA, Miller PJO, Yurk H, Samarra FIP, Hoyt E, Ford JKB, et al. Killer whale call frequency is similar across the oceans, but varies across sympatric ecotypes. J Acoust Soc Amer. 2015; 138: 251–257. 10.1121/1.4922704 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref030] 30.Morin PA, Archer FI, Foote AD, Vilstrup J, Allen EE, Wade P, et al. Complete mitchondrial genome phylogeographic analysis of killer whales (Orcinus orca) indicates multiple species. Genome Res. 2010; 20: 908–916. 10.1101/gr.102954.109 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref031] 31.Bigg MA, MacAskie I, Ellis G. Photo-identification of individual killer whales. Whalewatcher. 1983; 17(1): 3–5. [Google Scholar]

[pone.0236749.ref032] 32.Burdin AM, Hoyt E, Sato H, Filatova OA. 2006. The Killer Whales of Eastern Kamchatka. Alaska SeaLife Center; 2006. [Google Scholar]

[pone.0236749.ref033] 33.Ivkovich TV, Filatova OA, Burdin AM, Sato H, Hoyt E. The social organization of resident-type killer whales (Orcinus orca) in Avacha Gulf, Northwest Pacific, as revealed through association patterns and acoustic similarity. Mamm Biol. 2010; 75: 198–210. 10.1016/j.mambio.2009.03.006 [DOI] [Google Scholar]

[pone.0236749.ref034] 34.Filatova OA, Fedutin ID, Burdin AM, Hoyt E. The structure of the discrete call repertoire of killer whales Orcinus orca from Southeast Kamchatka. Bioacoustics. 2007; 16: 261–280. 10.1080/09524622.2007.9753581 [DOI] [Google Scholar]

[pone.0236749.ref035] 35.Filatova OA, Burdin AM, Hoyt E, Sato H. A catalogue of discrete calls of resident killer whales (Orcinus orca) from the Avacha Gulf of Kamchatka Peninsula. Zoologicheskii Journal. 2004; 83: 1169–1180 [Google Scholar]

[pone.0236749.ref036] 36.Filatova OA, Deecke VB, Ford JKB, Matkin CO, Barrett-Lennard LG, Guzeev MA, et al. Call diversity in the North Pacific killer whale populations: implications for dialect evolution and population history. Anim Behav. 2012. 83: 595–603. 10.1016/j.anbehav.2011.12.013 [DOI] [Google Scholar]

[pone.0236749.ref037] 37.Deecke VB, Janik VM. Automated categorization of bioacoustic signals: avoiding perceptual pitfalls. J Acoust Soc Amer. 2006; 119(1): 645–653. 10.1121/1.2139067 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref038] 38.Szymanski MD, Bain DE, Kiehl K, Pennington S, Wong S., Henry KR. Killer whale (Orcinus orca) hearing: Auditory brainstem response and behavioral audiograms. J Acoust Soc Am. 1999; 106: 1134–1141. 10.1121/1.427121 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref039] 39.Branstetter BK, Leger JSt, Acton D, Stewart J, Houser D, Finneran JJ, et al. Killer whale (Orcinus orca) behavioral audiograms. J Acoust Soc Amer. 2017; 141(4): 2387–2398. 10.1121/1.4979116 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref040] 40.Miller PJO. Mixed-directionality of killer whale stereotyped calls: A direction of movement cue? Behav Ecol Sociobiol. 2002; 52(3): 262–270. 10.1007/s00265-002-0508-9 [DOI] [Google Scholar]

[pone.0236749.ref041] 41.Branstetter BK, Moore PW, Finneran JJ, Tormey MN, Aihara H. Directional properties of bottlenose dolphin (Tursiops truncatus) clicks, burst-pulse, and whistle sounds. J Acoust Soc Amer. 2012; 131: 1613–1621. 10.1121/1.3676694 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref042] 42.Filatova OA, Samarra FIP, Barrett-Lennard LG, Miller PJO, Ford JKB, Yurk H, et al. Physical constraints of cultural evolution of dialects in killer whales. J Acoust Soc Amer. 2016; 140(5): 3755–3764. 10.1121/1.4967369 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref043] 43.Hodges RP. Underwater acoustics: Analysis, design and performance of sonar. John Wiley & Sons; 2011. [Google Scholar]

[pone.0236749.ref044] 44.Erbe C. Underwater noise of whale‐watching boats and potential effects on killer whales (Orcinus orca), based on an acoustic impact model. Mar Mam Sci. 2002; 18(2): 394–418. 10.1111/j.1748-7692.2002.tb01045.x [DOI] [Google Scholar]

[pone.0236749.ref045] 45.Holt MM, Noren DP, Veirs V, Emmons CK, Veirs S. Speaking up: killer whales (Orcinus orca) increase their call amplitude in response to vessel noise. J Acoust Soc Amer. 2009; 125(1): EL27–EL32. 10.1121/1.3040028 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref046] 46.Veirs S, Veirs V, Wood JD. Ship noise extends to frequencies used for echolocation by endangered killer whales. PeerJ. 2016; 4; e1657 10.7717/peerj.1657 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref047] 47.McKenna MF, Ross D, Wiggins SM, Hildebrand JA. Underwater radiated noise from modern commercial ships. J Acoust Soc Amer. 2012; 131(1): 92–103. 10.1121/1.3664100 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref048] 48.Brown JC. Mathematics of pulsed vocalizations with application to killer whale biphonation, J Acoust Soc Amer. 2008; 123(5): 2875–2883. 10.1121/1.2890745 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref049] 49.Foote AD, Griffin RM, Howitt D, Larsson L, Miller PJO, Hoelzel AR. Killer whales are capable of vocal learning. Biol Lett. 2006; 2: 509–512. 10.1098/rsbl.2006.0525 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref050] 50.Filatova OA, Samarra FIP, Deecke VB, Ford JKB, Miller PJO, Yurk H. Cultural evolution of killer whale calls: background, mechanisms and consequences. Behav. 2015; 152(15): 2001–2038. 10.1163/1568539X-00003317 [DOI] [Google Scholar]

[pone.0236749.ref051] 51.Filatova OA, Burdin AM, Hoyt E. Is killer whale dialect evolution random? Behav Proc. 2013; 99: 34–41. 10.1016/j.beproc.2013.06.008 [DOI] [PubMed] [Google Scholar]

[pone.0236749.ref052] 52.Mundinger PC. Animal cultures and a general theory of cultural evolution. Ethol Sociobiol. 1980; 1: 183–223. 10.1016/0162-3095(80)90008-4. [DOI] [Google Scholar]

[pone.0236749.ref053] 53.Lumsden CJ, Wilson EO. The relation between biological and cultural evolution. J Soc Biol Struct. 1985; 8: 343–359. 10.1016/0140-1750(85)90042-9. [DOI] [Google Scholar]

[pone.0236749.ref054] 54.Nousek AE, Slater PJ, Wang C, Miller PJO. The influence of social affiliation on individual vocal signatures of northern resident killer whales (Orcinus orca). Biol Lett. 2006; 2: 481–484. 10.1098/rsbl.2006.0517 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0236749.ref055] 55.Yurk H. Vocal culture and social stability in resident killer whales (Orcinus orca). Dissertation, University of British Columbia; 2005. [Google Scholar]

PERMALINK

Independent acoustic variation of the higher- and lower-frequency components of biphonic calls can facilitate call recognition and social affiliation in killer whales

Olga A Filatova

Roles

Abstract

Introduction

Material and methods

Ethics statement

Data collection

Similarity of contours of the lower and higher frequency components

Detectability of the lower- and higher-frequency components over distance

Fig 1.

Results

Similarity of contours of the lower and higher frequency components

Fig 2.

Table 1. Distances (mean ± SD) between pairs of calls calculated through dynamic time warping for lower- and higher-frequency components of K5 and K7 call types between the same families, between different families from the same pods, and between different pods.

Detectability of the lower and higher frequency components over distance

Fig 3.

Discussion

Fig 4. Biphonic K5 call type showing heterodyne frequencies above and below the higher-frequency component.

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

William David Halliday

Roles

Author response to Decision Letter 0

Decision Letter 1

William David Halliday

Roles

Author response to Decision Letter 1

Decision Letter 2

William David Halliday

Roles

Acceptance letter

William David Halliday

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases