AESTRACT Ihis study of the underwater behavior of harbor seals at Hopkins Marine Life Refuge comprises two papers: Fart I considers aspects of aquatic sleep patterns, while Fart II describes an underwater vocalization behavior, referred to as "rumbling". Harbor seals were found to sleep in the shallow waters near their haul-out. In the majority of cases a seal slept in a prone position amongst the surf grass. Sleep was most frequently observed at mid-tide levels. Fups were obseryed in aquatic sleep only at the water's surface. Rumbling is a highly stereotyped behavior involving both vocalizing seals (rumblers) and hon-vocalizing attendants. The low-frequency vocalization is performed by males only and in a localized portion of the study site. Attendants, also males, participate in the behavior by situating themselves alongside the vocalizing seal so that their heads are in close proximity. It is suggested that this behavior is linked to reproductive activity. INTRODUCTION Due to the general wariness of the animals, few behavioral studies have focused on the underwater activities of free-ranging harbor seals, Ehoca vitulina (see, for example, Venables and Venables 1759; Sullivan 1781; Allen 1785: Renouf and Lawson 1787). The seals of the Hopkins Marine Life Refuge, Facific Grove, California, have become acclimated to human activity on shore and in the water. While snorkeling I was able to observe patterns in underwater sleep and also noticed a highly stereotyped underwater vocalization behavior, hereafter referred to as "rumbling". MATERIOLS AND METHODS The study was conducted from April 25 to June 4, 1990 and consisted of approimately S hours of underwater observation. Figure I illustrates the study site, located in the area of Hopkins Marine Life Hefuge between Ogassiz Heach and Seal Racks. I chose this site because it is a reasonable area to observe over the time designated for this study and because it is immediately adjacent to pne of two groups of racks used as a haul-out by the approkimately 100 harbor seals resident to the refuge. The zeal population consistsof bath sexes of all age classes, though it has only become a birthing area in the past few years. The site has a granite and sand substrate predominated by Phyllospadix and algal growth; makimum depth is approximately ap m. In an etfort to insure that all portions of the site were surveyed adequately, it was divided into six manageable regions. Ihis notation also allowed me to document the approximate location of any observed behaviors. All observations were made while snorkeling during daylight hours, with notes recorded on underwater paper mounted on a plexiglass clipboard. Observations were conducted after dark on only three pecasions for reasons of visibility and personal safety. In addition, the seals are reported to be night foragers and therefore absent from the haul-out area at this time (Osborn 185). I saw no seals in the water or on the haul-out during my night sessions. FART I: SLEEF INTRODUCTION While precise definitions of sleep are difficult, the behavior can be recognized in mammals according to four criteria (Campbell and Tobler 1984): (1) a characteristic posture, (2) maintained quiescence, (3) a raised threshold Of arousal, and (4) reversibility of the behavioral state. There is no standardized method of distinguishing between "drowsiness" and "sleep" Captive gray seals have been described as sleeping at the surtace of the tank with the nostrils clear of the water, then sinking below the surface for up to four minutes betore rising again to breathe (Ridqway et al 1975). Similarly, the harbor seal is known to engage in what is known as bottling, where the animal bobs at the surface with the snout pointing directly overhead. Sullivan (1779) observed an adult female harbor seal to alternate between Bottling and sinking underwater in the same location for 68 minutes continucusly. However, the behavior of the seal onme it is underwater, including the four sleep criteria previously discussed, have never been described, MATERIALS AND METHODS I regarded as "sleep" any behavior in which the seal remained in a characteristic sleep posture for at least ten seconds, including those animals which may have been more properly described as "drowsy". Occasipnally a seal would adopt such a posture for the required time but would moye its head alertly as if looking for something in the water: these instances were not recorded as sleep. RESULTS Festure: In 887 of all recorded observations (n=42 of 48) the sleeping seal was lying prone on the substrate. There were a tew cases of a seal lying on its side or back. On four occasions a seal was observed rolling back and forth in the surge. In all cases the body was fully extended, with the foreflippers resting at the seal's side. Site: Seals were found sleeping in thick patches of Fhyllospadix in 757 of all cases (ne of 48). The grass provides excellent camouflage as it moves with the surge and is swept over the seal. In about half of these gmass¬ covered cases (5, n-19 of 3o) the seal was actually wedged under something, a rock ledge or a railroad tie, which had grass growing on or arpund it. The remainder of observations found the seal sleeping on shorter vegetation or on the open sandy substrate. I was unable to identify individuals well enough to determine whether each seal has its own preferred, specific site. Arousal: It was not uncommon for my approach to awaken a sleeping seal. However, often times the result was merely that the seal would open its eyes and look around slowly. apparently not awaking completely. Frequently then the seal would remain in place. On many occasions I was able to swim very close to a sleeping seal without eliciting any response. Similarly, I often observed other seals to swim within a foot of a sleeping seal without causing any disturbance. The average underwater sleep duration was about 1.5 minutes (x-1.6 min., n-35, range 10 seconds to 7.3 min.). Related locomotor activities: In all cases where the seal was wedged under something. it approached its sleep site by swimming into it. Howeyer. in 527 of the total observations where the data was taken (n-12 of 23), the seal sank rather than swam to the substrate. This sinking from the surface to the substrate was achieved without any discernible motion of the fore- or hindtlippers and without releasing bubbles. The motion appears to be consistent with previous observations of sinking and resurfacing, mentioned earlier. Seals surfaced for air immediately upon leaving the sleep site in 877 (n=20 Ot 23) of all cases for which these data were recorded. Diurnal patterns: Prequency of sleep observations was related to tidal Height, the greatest frequency occurring at mid-tide, rather than time of day (Figure 2). Note on pupE: All above results reflect only the activities of seals at least one year old; no pups were observed sleeping in this manner. Instead, pups floated at the surface with the back exposed and head and tail underwater. They maintained this posture for minutes at a time and were often easily approached while doing so. DISCUSSION In summary, we see that the underwater resting behavior of harbor seals does meet the definition of sleep as outlinedlearlier. The seals maintain a characteristic prone Posture amongst the surf grass for up to seven minutes at a time and, presumably during deeper stages of sleep, are not easily aroused by the observer or other seals. he seals were observed to sleep underwater most frequently at mid-tide levels. This result is surprising for two reasons: (1) it has been hypothesized that harbor seals sleep in water as a result of being forced off the haul-out by the rising tide (Riedman 1990), and (2) there does not seem to be a significant rise in general aquatic activity at higher tides, though there are much fewer seals on the haul-out. Therefore, it would seem that many seals leave the area altogether at high tide. There are several biases introduced by the methods of this study. While not likely causing a disturbance in the general postural and site characteristics of the behavior. my presence as a foreign observer of the seals probably did interfere in the accuracy of measurements of sleep duration and in the alternating surface/substrate patterns. In addition, I am sure that I simply did not notice a signiticant amount of sleep activity because the surf grass provided such excellent camouflage. There were instances when I lost track of the location of a sleeping seal, even though I had seen it just a moment before, because of the constant motion of the grass. Finally, my study focused only on the underwater portion of sleep behavior. The results should be combined with observations from land that follow the seal as it floats and sinks in the water. In this manner a more accurate portrayal of the in-water sleeping abilities of the harbor seal can be compiled. FORT II: RUMBLING INTFODUCTION Renouf and Lawson (1987) found that adult harbor seals play socially only during the mating season, which extends from April through June in California (Bonner 1990). Ralls et al (1985) hypothesize that phocids are considered one of the least vocal pinnipeds because most of their Vocalizations occur underwater; they note that yocalization rates are probably greatest during the mating season and that sezually mature males are the most vocal animals. Yet the best documented underwater activities of the harbor seal are the clicking vocalizations thought to be associated with echolocation (see, for example, Schevill, Watkins and Hay 1953; Renouf 1989). During my studies of underwater sleep, I noticed a previously undescribed underwater vocalization behavior, which I refer to as "rumbling". performed exclusively by males and possibly associated with reproductive behavior. MOTERIALS AND METHODS Vocalizations were recorded with a hydrophone and a Sony portable recorder. Sonograms of the vocalizations were produced on the Multigon Industries Uniscan II. Vided recordings were made on two separate occasions using a Super VHS format recorder in a waterproof housing. I used the tapes primarily for illustrative purposes, only including individual discrete bouts of behavior for analysis. HESULTS There are two possible roles played by male harbor seals involved in rumbling behavior, both illustrated in figure 3: the vocalizing seal and non-vocalizing attendants. Vocalizing seals are not always attended. The age of the males involved is unknown, but in the majority of observations the rumbling seal had a pronounced swelling along the penile line, indicative of a certain degree of sexual maturity. In addition, all but three isolated observations occurred in an approximately 100 sg. m. region of the study site immediately in front of one end of the haul-out (see figure 1). The behavioral pattern seemed to be distributed through daylight hours and not related to the tidal cycle. In the stereotypical pattern of rumbling behavior, as a seal swims towards the substrate and arrives within a few feet of the bottom, it rolls onto its back so that it reste parallel to the substrate, upside down. Its throat and chest region swell and its head begins to lift slightly while the tail remains in place. Suddenly there follows a full body contraction similar to that made by a person with a severe cough. This motiom results in a loud burst of sound, like a pop, followed by a low frequency rumbling roar which lasts approximately three to five seconds. Sonograms Ot three separate vocalizations made by two unattended seals are very similar and show that the band of maximum energy occurs between 150 and 200 Hz, with no harmonics (figure 4). The seal may rumble several times while still on its back. I will refer to each discrete period which the seal spends on its back as one rumbling bout; therefore, the seal may perform several vocalizations per bout. During the vocalizations the swelling of the throat and chest decreases, though there is not necessarily an accompanying release of bubbles. In fact, there are generally few bubbles, if any, released during the vocalizations. Unce the seal completes this rumbling bout, it returns to its upright position, briefly scans the area visually. and then either rolls over to perform another bout in appruximately the same location or completes the entire rumbling episode by swimming to the surface. I define an episode as that perind of time from the first rumbling bout to the time when the seal either returns to the surface or escapes my field of vision. On average a seal will perform approximately three outs per episode( -.2 bouts/episode, range 1 to 7 bouts/episode, S.D. - 1.47 bouts/episode. n - 71). In addition, frequently the same seal will perform a long series of episodes per obeervation period, always returning to the same spot to vocalize. bserved one seal perform a continucus series ofI 10 episodes, consisting of a combined total of 116 bouts. during a one hour and forty minute period. Rumbling behavior was observed during 15 of 25 separate Observational sessions. On nine of these 15 occasions the Fumbling seal was accompanied by at least one other seal during one or more bouts. Though I could only consistently identify four rumbling seals for the duration of the study, it is possible that a total of up to thirty different individuals were observed rumbling. By my definition, an accompanying or attending seal may assume any of a wide variety of body postures as long as the following criterion is met: an attending seal alighs itself alongside the vocalizing seal so that their heads are within a tew inches of each other, if not in actual physical contact. This posture is held only for the duration of an individual bout; between bouts the accompanying seal often swims away. Of the 358 bouts observed 68.47 (n-245) were umattended, 24.97 (n=89) had one attendant, 6.17 (n=22) had two attendants, and 0.67 (n-2) had three attendants. During any one observational period, the members of a rumbling behavior group remained consistent: that is, the same Fumbling seal would have the same attending seals, though not all of those attendants would necessarily be present at all times. However, determining whether these groups remained constant across observational periods was difficult. Although there was generally very little antagonism amongst the seals involved in rumbling behavior. I did 11 observe at least one case of aggression and/or competition between most of the possible combinations of rumbling roles: between two attendants, between an attendant and rumbler. between two rumblers, and between a rumbler and a seal not participating in the behavior at all. While two or more attendants never fought during bouts, they frequently fought between bouts, though never to the extent of drawing blood. For example, during one morning observation session I noticed that one attendant was consistently head-thrusting and scratching a second attendant. As the session continued, the more aggressive attendant was frequently the only accompanying seal. When I returned to observe in the atternoon, the same rumbling seal was performing and the more aggressive attendant was still accompanying, but the sther attendant never again appeared. Un a separate occasion I observed a seal which consistently had attendants. One of these attendants twice performed a rumbling episode while the one it had been attending was at the surface. However, this attendant¬ turned-rumbler vocalized alone, while the original rumbling seal soon returned to perform more episodes and was still attended by the thers of the group. Humblers were also aggressive at times. On one occasion I observed an unattended, vocalizing seal roll onto its stomach, look around, spot another rumbling seal about 15 teet away, and approach this second seal as if to become an attendant. Instead it began to scratch at the thrat of 12 the second seal with its foreflippers until that seal swam away. Ihen the first seal began another rumbling bout in this new location. In another case of aggression on the part of the rumbler, a male which had been attended while rumbling earlier was rumbling alone. (This was the same seal which I was able to observe for an hour and forty minutes continuously.) After vocalizing, he righted himself. retracted his head so that his neck appeared thick, then Burst upwards towards one of a pair of seals which had been interacting at the surface. I was unable to determine the sex of either of these seals. After a brief encounter. the attacked seal left the area and the rumbler foreflipper splashed (as described by Sullivan 1981) a short distance away. However, the pair at the surface regrouped and was attacked in the same manner by the rumbler three more times. with the rumbler foreflipper splashing after each encounter. Ihere were only two cases which definitely involyed female seals, though not as attendant or rumbler. In one instance a female rested under a nearby rock ledge while rumbling occurred. However, she did nut seem to notice the activity and the males behaved according to the standard patterm. During another session a male and remale were interacting far from the usual rumbling site. They were tumbling around each other near the surface, biting each uther's neck and grabbing each other with the foreflippers. s I approached, the female swam away and the male rolled onto his back and began swimming upside-down at a fairly rapid pace. Occasionally as he swam he rumbled, which was quite different from the basically stationary pattern I had seen betore. He swam in a wide circle without surfacing or righting himself, eventually arriving back at the original site to which the female had returned. He righted himself as soon as he saw her and they began playing again. This pattern was repeated twice, after which the male hauled out. DISCUSSION There is a wide variety of evidence to suggest that the rumbling behavior is linked to mating. A somewhat similar behavior known as ballooning is performed by a closely related species, the spotted seal, Ehoca largha, exclusively during its mating season (Beier and Wartzok 1979). Ballooning is a male behavior in which the seal lies on the bottom of the tank on his stomach and retracts his head so that the thoracic region appears swollen. No vocalizations appear to e associated with the behavior. However. this same study also shows that the spotted seal des haye underwater vocalizations which occur only during the mating season, though Stirling (1973) found no such vocalizations Tor the ringed seal, Fhoca hispida. Evelyn Hanggi (personal communication), recording underwater vocalizations of harbor seals at Foint Lobos. California, has heard only the rumbles, and variations on them, and has found that they are only prominent during May and June. Similarly, I observed rumbling behavior primarily during May and the beginning of June. For a review of pinniped vocalizations, refer to Riedman (1790). Given that these studies took place during the breeding season for this species, it is possible that this behavior is a form of lekking. Bradbury (1981) defines a lek species as one in which 1) males contribute only gametes to the next generation; 2) males gather at a site to which females come and where most mating occurs; 3) this display site contains no resources found exclusively at the site and necessary to the females, except for the males themselves: and 4) the females have a choice of mates. Lek behavior has already been suggested for another pinniped, the walrus. Udcbenus rosmarus. The male walrus exhibits acoustical and visual displays in the water near groups of females, subadulte, and young resting on the pack ice during their breeding season (Ray and Watkins 1975). The authors also suggest that the display may be a means of establishing underwater territories or dominance hierarchies. The amecdotal evidence found in my studies regarding interactions between vocalizers and attendants might indicate that the harbor seal rumbling behavior is a competitive yet nen-violent means of creating a male hierarthy amongst the seals which vocalize, are attended, or are attendants. 15 ACKNOWLEDGMENTS I would like to thank my advisor, Chuck Baxter, for his helpful suggestions and the loan of various equipment. and Alan Baldridge for providing much enthusiasm and encouragement and repeatedly pointing me in the right direction. Thanks to John Lee, Mark Denny, and Khosrow Lashkari for their collective attempts to explain acoustics to me; to Marianne Reidman for her advice and the loan of her very helpful book; and to Ken Norris and the members of his lab for allowing me to use the Multigon. I am particularly thankful to Nicole Crane for the amazing video and the even more amazing feat of accessing functioning sonogram equipment. U course, many thanks go to the other students of 175H. I've never met such a great group! Finally, for my incredible friend Diane Scott. thank you for understanding. LITERATURE CITED Allen, S.G. 1985. Mating behavior in the harbor seal. Marine Mammal Science i:84-87. Beier, J.C. and D. Wartzok. 1979. Mating behaviour of captive spotted seals (Fhoca langha). Animal Behavior 27: 781. Bonner, W.N. 1990. The natural history of seals. Facts on File, Inc., New York, NY. Bradbury, J.W. 1981. The evolution of leks. Fages 138-167 in H.D. Alexander and D.W. Tinkle, eds. Natural selection and social behavior: Research and new theory. Chiron Fress, New York. NY. Campbell, S.S. and I. Tobler. 1784. Animal sleep: O review of sleep duration across phylogeny. Neurpscience & Biobehavioral Reviews 8:267-700. Osborn, L.S. 1985. Fopulation dynamics, behavior, and the etfects of disturbance on haulout patterns of the harbor seal Phoca vitulina richardsi. Senior thesis, University of California, Santa Cru. Ralls, K., F. Fiorelli and S. Gish. 1785. Vocalizations and vocal mimicry in captive harbor seals, Ehoca vitulina. Canadian Journal of Zoology 6:1050-105a. Hay. G.C. and W.A. Watkins. 1975. Social function of Underwater sounds in the walrus Odobenus rosmarus. Hapports et Froces-verbaus des Heunions Conseil Fermanent International pour l'Exploration de la Mer 169: 524-526. Renouf, D. and J.W. Lawson. 1987. Quantitative aspects of harbour seal (Ehoca vitulina) play. Journal of Zoology (London) 212:267-273. Renouf, D. 1989. Sensory function in the harbor seal. Scientific American 260:90-95. Ridgway, S.H., R.J. Harrison, and F.L. Joyce. 1975. Sleep and cardiac rhythm in the gray seal. Science 187:SS SS5. Riedman, M. 1990. The pinnipeds: Seals, sea lions, and walruses. University of California Press, Berkeley, CO. Schevill, W.E., W.A. Watkins and C. Ray. 1763. Underwater sounds of pinnipeds. Science 141:50-53. Stirling, I. 1973. Vocalization in the ringed seal (Ehoca hispida). Journal Fisheries Research Board of Canada J0: 1592-1594. Sullivan, R.M. 1979. Behavior and ecology of harbor seals, Ehoca vitulina, along the open coast of northern Lalifornia. Master's thesis, Humboldt State University, Sullivan, R.M. 1781. Aquatic displays and interactions in harbor seals, Ehoca vitulina, with comments on mating systems. Journal of Mammalogy o2:825-831. Venables, U.M. and S.V. Venables. 1957. Vernal coition of the seal, Fheca vitulina, in Scotland. Froceedings of the loological Society of London 132:665-667. FIGURE LEGEND Figure 1: Map of study site Figure 2: Graph of mean number of underwater sleep obseryations per hour vs. tidal height. The sine-wave trace represents a typical daily tide cycle. Figure 3: Typical posture of rumbling seal (below) and one attendant. Figure 4: Sonograms of unattended rumbles. (A) and (B) were recorded from the same seal. Passiz OO SEAL ARocks X C A AGASSIZ BEACA — Q BOAT WORKS FIGURE — O L A X vel O C O Q () IHSISH Tvoil + O +o E D L O — — — FIGURE 3 ++- --. 14 14 + 4 t 1 141 4.. T 4 â 1003" * . 1 os à k- (2H9) XoNanoaad - L 6 L E 1 (2HY —+ - * — -.. -. ....... .. ....13. . ...... 2. . 1 .. : 4 ... . P 1008• 4 . 4 B F K E t 47 --1 XONSnosa L + (2HY) --. .* + 1 50t —- —.-. c.7r . .. — J — XONSROS L 2 L L