1 Title: Extreme diversity in the songs of Spitsbergen’s bowhead whales 2 Authors: K.M. Stafford1*, C. Lydersen2, Ø. Wiig3, K.M. Kovacs2 3 4Affiliations: 51Applied Physics Laboratory, University of Washington, Seattle, WA, USA 62Norwegian Polar Institute, Tromsø, Norway 73Natural History Museum, 8*Correspondence University of Oslo, Oslo, Norway to: kate2@uw.edu 9 10Abstract: Almost 11complex all mammals communicate using sound, but few species produce songs. Two baleen whales sing complex songs that change annually, though 12only the humpback 13attention. This 14(Balaena 15184 16in whale (Megaptera novaeangliae) has received much research study focuses on the other baleen whale singer, the bowhead whale mysticetus). Members of the Spitsbergen bowhead whale population produced different song types over a 3-year period, based on duty-cycled recordings from a site Fram Strait in the NE Atlantic. Distinct song types were recorded over short periods, 17lasting at most some months. This song diversity could be the result of population 18expansion, 19from each 20within 21the or immigration of animals from other populations that are no longer isolated other by heavy sea ice. However, this explanation does not account for the season and annual shifting of song types. Other possible explanations include that extraordinary diversity in songs might be the result of weak selection pressure for 22interspecific identification 23strong or for maintenance of song characteristics or, alternatively, pressure for novelty in a small population. 24 25Keywords: Bowhead whale, Arctic, Balaena mysticetus, song 26 27Background Complex “song” in mammals is rare. While many mammalian taxa produce 28 29repetitive “calls,” 30vocal displays 31modulated 32have been documented in only a few mammalian species, including some bats 34capensis), 35(Balaena 36females gibbons (Hylobatidae), mice (Scotinomys spp), rock hyraxes (Procavia and two great whales, humpback (Megaptera novaeangliae) and bowhead mysticetus) whales [3,5-8]. With the exception of gibbons, in which males and duet [5], complex songs in mammals are thought to be produced only by males 37[6,9-11]. Male mammals are thought to sing to defend territories, advertise their quality, 38attract mates, or some combination of these functions [5,11]. Song in baleen whales has been studied extensively only in humpback whales, 39 40which sing similar songs within a season across a whole population. The structure of that gradually evolves over the season in unison [12] and transfer of song types has been 42documented 43[13]. akin to bird song, which is defined by multiple frequency and amplitude- elements combined into phrases and organized in long bouts [4]. Such songs 33(Chiroptera), 41song sometimes called advertisement songs [1-3], few mammals produce to occur directionally from one population to another over a period of years Humpback whale songs are composed of a hierarchy from units to sub-phrases to 44phrases to themes [12]. 45 Less is known about the songs of bowhead whales compared to humpback 46whales, but bowhead whale songs generally consist of a single phrase that includes 47amplitude48sounds 492009 and frequency-modulated elements repeated in bouts, with two different often produced simultaneously [14]. A pilot study from the Fram Strait in 2008- provided the first indication that tens of song types were produced by bowhead 50whales in this region within a single overwinter period [8]. No year-round studies of song 51diversity exist 52single for other bowhead whale populations although multiple song types in a year have been documented for two other populations [15,16]. Herein, we 53document extremely high 54Spitsbergen inter- and intra-annual diversity in mammalian song from the bowhead whale population. 55Methods Omni-directional hydrophone recorders were deployed and redeployed annually 56 57from 2010-2014 58Strait, in September on a long-term oceanographic mooring in western Fram at about 78°49N, 5°W (figure S1). The mooring deployed in 2011was not 59recovered. Instruments recorded acoustic data for the first 14-17 min h-1 throughout the 60year. Upon recovery, the data were downloaded and spectrograms (10-4000 Hz, 2048 61point FFT, 50% overlap, Hann window) were created for each data file. Spectrograms 62were then 63by eye 64songs based on time-frequency characteristics of each song type [17]. Files with “loud” (possessing distinct units, song bandwidth exceeding 500 Hz) were analysed 65further 66a reviewed visually for the presence of bowhead whale songs that were classified to determine unit structure and song composition. Individual songs were assigned unique number if more than one iteration occurred (e.g. Fram2012-11 was the eleventh 67song recorded in the 2012-13 season while Fram2012-54 was the fifty-fourth, ESM 1-4). 68Each song type was characterized by frequency, duration, and amplitude and/or 69frequency modulation, 70within and between years to determine the minimum number of different song types 71detected 72 number of units and phrases and compared to all other song types and to describe the diversity of songs in this species at this location. 73Results Bowhead whale songs were detected 24 h per day throughout most of the winter 74 75every year (figure S2). The greatest number of different song types was recorded in 76December and January (figure 1). A total of 184 different song types were recorded in the 77three-year study period. Each song type was recorded in only one deployment period. There was interannual variation in the number and timing of songs. The fewest 78 79songs 80(69 were recorded in 2010-2011 (39 song types total, 895 recordings). Both 2012-2013 song types total, 1338 recordings) and 2013-2014 (76 song types total, 998 81recordings) had approximately twice as many different songs (figure 1). Most songs occurred only within short time periods – from a few hours to a few 82 83days – and were seldom recorded in more than one month (figure 2). However, in each 84year, some few song types occurred throughout the winter. The overall trend for all years 85was a progression of song types appearing and then disappearing over time with the 86greatest within-year Of the 3,231 recordings containing songs in the three years, slightly over half 87 88(53%) 89the diversity occurring in January for all three years examined. contained only a single song type, while two different songs occurred in 37% of recording periods. Few recording periods included greater song diversity (3 (9%) or 4 90(<1%) distinct songs). 91Discussion 92 93year The diversity and interannual variability in songs of bowhead whales in this threestudy is rivaled only by a few species of songbirds [4]. Among other mammalian 94singers, mice and gibbons tend to produce highly stereotyped and repetitive songs with 95few elements [e.g. 3,5]. Variation in rock hyrax and bat songs is primarily through 96changes 97that are in the arrangement of units [6,11]. Humpback whales produce complex songs similar within a year [7,12,13]. Although the repertoire of any one individual 98bowhead whale in this study cannot be determined, the catalogue of song types (184) is 99remarkably varied. 100 It is not known whether individual bowhead whales sing multiple song types in a 101season, but some are known to share the same song type in the same period in the Bering- 102Chukchi-Beaufort 103whales (BCB) population [14,16]. Nor is it known if individual bowhead maintain the same song throughout their lifetime or if they switch within and/or 104between years. One explanation for the very high song diversity in the Spitsbergen bowhead 105 106whale population could be that the animals occupying this area in modern times include 107immigrants from both the BCB and the eastern Canada-western Greenland bowhead 108populations. 109other 110past Until recently, these populations have been assumed to be isolated from each due to extensive, impenetrable sea ice cover in the High Arctic. However, in the few decades, extreme declines in sea ice extent and thickness may have facilitated 111contact between these populations [18]. However, even if this region contains bowhead 112whales from multiple populations, this does not fully explain the high numbers of 113different song 114year 115 116of types recorded in this study or the lack of recurrence of song types from to year. It is plausible that the bowhead whales in the Fram Strait are simply a remnant the original Spitsbergen population that survived the extreme historical levels of 117exploitation [19]. The influence of small population size on song diversity is conflicted; 118some studies suggest song diversity increases in smaller populations, although others 119have found that reduced or isolated populations exhibit a reduction in song diversity and 120produce simpler songs [20,21]. In some species, females appear to prefer a diverse song repertoire [22-25], 121 122suggesting that increased complexity of singing might confer reproductive advantages. A 123recent study of 124reproductive 125males, howler monkeys (Alouatta spp) documented tradeoffs in male characteristics based on (temporary) social structure: in groups with fewer or smaller social groups, males invested more in vocal displays as a reproductive 126tactic [26]. Normally, testes size and vocal repertoire (or other reproductive displays) are 127considered evolutionary trade-offs: depending upon social context, one of these may 128provide a selective advantage for individuals within a population over the other. For 129example, humpback whales have relatively small testes, and engage in physical 130competition 131spp) 132and as well as producing complex song displays, while right whales (Eubalaena have enormous testes, are drawn to “surface active groups” by a vocalizing female, lack any apparent male acoustic display [27-29]. However, bowhead whales have 133both large testes and large vocal repertoires [30]. Bowhead whales are the only High Arctic resident baleen whale. Thus, 134 135interspecific identification 136bowheads 137song that it might for other cetacean species. This could reduce selection pressure on stereotypy, allowing for greater variation in song types as a result of long-term 138cultural 139 via song may not confer the same selective advantage for mutation in songs, or song novelty itself might confer an advantage [4,20,22,23]. Because bowhead whales sing underwater, in heavy ice during the polar night, a 140nuanced understanding of the variable syntax of this species will be difficult to obtain. 141Nevertheless, 142distinct the singing behaviour of Spitsbergen bowhead whales, in which tens of song types are produced annually, makes them remarkable among mammals. 143References 144 145 146 147 148 149 150 1. Clutton-Brock TH, Albon SD. 1979 The roaring of red deer and the evolution of honest advertisement. Behaviour 69, 145–170 2. Watkins W. 1981 Activities and underwater sounds of fin whales. Sci. Rep. Whales Res. Inst. 33, 83–117. 3. Miller JR, Engstrom MD. 2007 Vocal stereotypy and singing behavior in Baiomyine mice. J. Mammal. 88, 1447–1465. 4. Kroodsma DE. 2004 The Diversity and Plasticity of Birdsong. In Nature’s 151 Music: the Science of Birdsong (eds Marler P, H Slabbekoorn), pp. 108-131. 152 Elsevier. 153 5. Cowlishaw G. 1992 Song function in gibbons. Behaviour. 121, 131–153. 154 6. Kershenbaum A, Ilany A, Blaustein L, Geffen E, 2012. Syntactic structure and 155 geographical dialects in the songs of male rock hyraxes. Proc. R. Soc. B: Biol. 156 Sci. 279, 2974–2981. 157 158 7. Payne RS, McVay S. 1971 Songs of humpback whales. Science, New Ser. 173, 585–597. 159 8. Stafford KM, Moore SE, Berchok CL, Wiig Ø, Lydersen C, Hansen E,  160 Kalmbach D, Kovacs KM. 2012 Spitsbergen’s endangered bowhead whales sing 161 through the polar night. Endang. Species Res. 18, 95–103. 162 163 164 165 166 167 168 9. Winn HE, Winn LK. 1978 The song of the humpback whale Megaptera novaeangliae in the West Indies. Mar. Biol. 47, 97-114. 10. Davidson SM, Wilkinson GS. 2004 Function of male song in the greater white­ lined bat, Saccopteryx bilineata. Anim. Behav. 67, 883–891. 11. Bohn KM, Schmidt-French B, Schwartz C, Smotherman M, Pollak GD. 2009 Versatility and stereotypy of free-tailed bat songs. PLoS ONE 4, e6746. 12. Cholewiak DM, Sousa-Lima RS, Cerchio S. 2012 Humpback whale song 169 hierarchical structure: Historical context and discussion of current classification 170 issues. Mar. Mamm. Sci. 29, E312–E332 (2012). 171 13. Garland EC, Goldizen AW, Rekdahl ML, Constantine R, Garrigue C, Hauser  172 ND, Poole MM, Robbins J, Noad MJ. 2011 Dynamic horizontal cultural  173 transmission of humpback whale song at the ocean basin scale. Curr. Biol.  174 21,687–691. 175 14. Würsig B, Clark CW. 1993. Behavior. Pages 157–199 In The Bowhead Whale 176 (eds JJ Burns, JJ Montague, CJ Cowles), Special Publication Number 2, The 177 Society for Marine Mammalogy. 178 179 180 15. Stafford KM, Moore SE, Laidre KL, Heide-Jørgensen MP. 2008 Bowhead whale springtime song off West Greenland. J. Acoust. Soc. Am. 124, 3315–3323. 16. Johnson HD, Stafford KM, George JC, Ambrose Jr. WG, Clark CW. 2014 Song 181 sharing and diversity in the Bering-Chukchi-Beaufort population of bowhead 182 whales (Balaena mysticetus), spring 2011. Mar. Mamm. Sci. 31, 902–922. 183 184 185 186 187 17. Kershenbaum A, et al. 2014 Acoustic sequences in non-human animals: a tutorial review and prospectus. Biol. Rev. 91,13–52. 18. Heide-Jørgensen MP, Laidre KL, Quakenbush LT, Citta JJ. 2012 The Northwest Passage opens for bowhead whales. Biol. Lett. 8, 270–273. 19. Vacquié-Garcia J, Lydersen C, Marques TA, Aars J, Ahonen H, Skern-Mauritzen 188 M, Øien N, Kovacs KM. 2017 Late summer distribution and abundance of ice- 189 associated whales in the Norwegian High Arctic. Endang. Species Res. 32, 59– 190 70. 191 192 193 194 195 196 20. Hamao S, Ueda K. 2000 Simplified song in an island population of the bush warbler Cettia diphone. J. Ethology 18,53–57. 21. Laiolo P, Vögeli M, Serrano D, Tella JL. 2008 Song diversity predicts the viability of fragmented bird populations. PLoS ONE 3, e1822. 22. Byers BE, Kroodsma DE. 2009 Female mate choice and songbird song repertoires. Anim. Behav. 77, 13–22. 197 23. Chabout J, Sarkar A, Dunson DB, Jarvis ED. 2015 Male mice song syntax 198 depends on social contexts and influences female preferences. Front. Behav. 199 Neurosci. 9, 1–16. 200 24. Sullivan B, Hinshaw S. 1992 Female choice and selection on male calling 201 behaviour in the grey treefrog Hyla versicolor. Anim. Behav. 44, 733–744. 202 25. Pasteau M, Nagle L, Kreutzer M. 2004 Preferences and predispositions for intra- 203 syllabic diversity in female canaries (Serinus canaria). Behaviour 141,571–583. 204 26. Dunn JC, Halenar LB, Davies TG, Cristobal­Azkarate J, Reby D, Sykes D,  Dengg S, Fitch WT, Knapp LA. 2015 Evolutionary trade-off between vocal tract 205 206 207 and testes dimensions in howler monkeys. Curr. Biol. 25, 1–7. 27. Brownell RL Jr, Ralls K. 1986 Potential for sperm competition in baleen whales.  208 209 Rep. Int. Whal. Commn. Spec. Iss. 8, 97–112. 28. Tyack P, Whitehead H. 1982 Male competition in large groups of wintering 210 211 humpback whales. Behaviour 83,132–154. 29. Parks SE. 2003 Response of North Atlantic right whales (Eubalaena glacialis) to 212 playback of calls recorded from surface active groups in both the north and South 213 214 Atlantic. Mar. Mamm. Sci. 19, 563–580. 30. Haldiman JT, Tarpley RJ. 1993 Anatomy and Physiology. In The Bowhead Whale (eds JJ Burns, JJ Montague, CJ Cowles), pp. 71-156. The Society for 215 216 217 Marine Mammalogy. 31. Stafford KM, Lydersen C, Wiig Ø, Kovacs KM. 2018 Data from: Extreme  218 diversity in the songs of Spitsbergen’s bowhead whales. Dryad Digital  219 220 Repository. (doi:10.5061/dryad.1ck400f). 221Ethics 222 223The project was approved by the Norwegian Nation ethics committee for animal use in 224Research (Mattilsyn) and permits were issued by the Governor of Svalbard (Sysselman). 225 226Acknowledgments 227Kristen Fossan serviced the Fram Strait hydrophone annually. 228Author contributions 229All authors collected the data and conceived the idea for analysis. KMS analysed the 230data. All 231to authors contributed to interpreting the results, writing the manuscript and agree its publication. All authors are accountable for all aspects of the work. 232Data accessibility 233Exemplar song files are deposited in Dryad Digital Repository [31]. 234Competing 235We have interests no competing interests 236Funding 237This work was funded by NPI, The Svalbard Environmental Protection Fund, Svalbard 238Science Forum, 239(grant the Fram Centre Incentive Fund and the Norwegian Research Council number 244488/E10). 240 241Figure 1. Total numbers of bowhead whale song types recorded in each month (bars) and 242cumulative number of song types (dashed lines) by year. The greatest number of different 243song 244for 245 types occurs in December and January, presumed to be the peak of mating season bowhead whales. Figure 2. Total number of hours and months during which each song type was recorded 246by year. In most cases, a song type was only recorded in one month, though in some 247instances 248song the same song type was recorded in 2-4 different months. a) 2010-2011: 38 types were recorded; b) 2012-13: 69 song types recorded; c) 2013-14: 76 song types 249were recorded. 250 251 252 253 254