ORIGINAL PAPER Kenneth J. Goldman Regulation of body temperature in the white shark, Carcharodon carcharias Accepted: 17 February 1997 Abstract Stomach temperatures of three white sharks, Carcharodon carcharias, (one reported previously and two new individuals) were intermittently recorded by acoustic telemetry at the South Farallon Islands, central California. Temperature profiles of the water column were obtained concurrently. Stomach temperatures were elevated over ambient water temperatures by as much as 14.3 ��C. Stomach temperatures varied within a narrow range while ambient water temperature fluctuated over a much larger range, showing that this species regulates its body temperature. These data, in combination with previous work on the physiology and anatomy of white sharks, indicate that the white shark is endothermic. It appears that the heat retention system in lamnid sharks has allowed them to inhabit cold water and remain ac- tive predators of swift and agile prey. Key words Carcharodon carcharias AE South Farallon Islands AE Telemetry AE Stomach temperature AE Thermoregulation AE Endothermy Abbreviations SFI South Farallon Islands AE SST sea surface temperature Introduction White sharks, Carcharodon carcharias, belong to a very small group of lamnoid sharks (Order Lamniformes) that possess vascular counter-current heat exchangers (Bone and Chubb 1983 Carey et al. 1981). These allow individuals to retain metabolically-generated heat in- stead of losing that heat to the ambient water during oxygenation of the blood at the gill lamellae. The masses of parallel arteries and veins making up these retia mirabilia were first described in the related porbeagle shark Lamna nasus [as Lamna cornubica (Burne 1923) see also Smith and Rhodes (1983)]. They are located in the brain (orbital rete), musculature (subcutaneous lat- eral rete), and viscera (suprahepatic rete), and function to elevate the shark���s body temperature over the ambient water temperature. White sharks have been shown to possess elevated muscle, brain, eye, and stomach tem- peratures (Carey et al. 1982 Tricas and McCosker 1984 Block and Carey 1985 Carey et al. 1985 McCosker 1987 Wolf et al. 1988 Goldman et al. 1996). The first stomach temperature data were obtained from a white shark in South Australia by McCosker (1987), who suggested that ������further study might be di- rected toward the hypothesis that Carcharodon can control its stomach temperature.������ A second individual tagged o�� California in 1991 (Goldman et al. 1996) generated the hypothesis that white sharks maintain an ������optimal physiological operating temperature with minimal e��ects from ambient water temperature.������ The purpose of this paper is to examine newly obtained stomach temperature data along with all previous tem- perature and anatomical data in order to test the hy- potheses of physiological thermoregulation and endothermy in the white shark. I then briefly discuss the relationship of body temperature to the physiological ecology of this species. Materials and methods This paper presents data on stomach temperatures from three free- swimming adult male white sharks at the South Farallon Islands (SFI) (37��42��N, 123��00��W), a small group of rocky islands located 30 km west of San Francisco, California. One shark, an approxi- mately 4.3 m (total length) individual, reported in Goldman et al. J Comp Physiol B (1997) 167: 423���429 �� Springer-Verlag 1997 K.J. Goldman1 Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132, USA Present address: 1 The College of William and Mary, School of Marine Science, Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, VA 23062, USA e-mail: keng@vims.edu

(1996), was observed in 1991 and is referred to as shark #1. The two additional individuals were observed in 1993. One shark (#2) was approximately 4.9 m total length, and the second shark (#3) was approximately 3.7 m total length. All observations took place at SFI during October���November, when white shark predation on pinnipeds is common (Ainley et al. 1981 Ainley et al. 1985 Klimley et al. 1992 Anderson et al. 1996a). Gender was determined by using underwater videos obtained when the transmitter was fed to the individual or when it was attracted to an unbaited decoy which housed a video camera [see Anderson et al. (1996b) for description of decoy study]. Total lengths of individuals were estimated from repeated observations of the sharks next to the 5.4 m tracking vessel or at a decoy in close proximity to the tracking vessel (K.J. Goldman, S.D. Anderson, P. Pyle, unpublished observation). Stomach temperature data were obtained from acoustic transmit- ters that were fed to sharks during the course of natural predatory attacks on pinnipeds. All data are diurnal and were gathered in- termittently over multi-day periods. Transmitters were manufactured by VEMCO Ltd. (Halifax, Nova Scotia, Canada ��� model V4TP-8H) and operated at fre- quencies of 30.000 and 32.768 kHz. They possessed thermistors and depth sensors with ranges of 0���30 ��C and 0���200 m, had a life of 57��� 91 days, and a maximum range of 1.1 km at SFI (based on range tests conducted in the field). All transmitters were calibrated by the manufacturer I verified thermistor readings on four separate oc- casions in a digital readout waterbath, using a Fluke K/J 51 ther- mometer as a backup. These two devices were always within 0.1 ��C of each other. The VEMCO calibrations and my waterbath recali- brations were always within 0.2 ��C. One transmitter���s calibrations were further verified after it was regurgitated by shark #3 in 1993. Calibrations were unchanged during this biological excursion. It was fed to another white shark (approximately 4.0 m TL female) at the North Neptune Islands (35��14��S, 136��03��E), Spencer Gulf, South Australia on January 29, 1994 (see Results and Discussion). Each shark at SFI was fed a piece of blubber (3���4 kg) from an elephant seal, Mirounga angustirostris, with transmitter attached, that had been placed in the water during a feeding event resulting from a predatory attack. This procedure was followed so the shark would ingest it during the course of its natural predatory and feeding behavior. No attractants (e.g. blood or fish parts) were used to avoid altering the natural behavior of the sharks at SFI. Immediately upon ingestion of a transmitter, sharks were monitored using a directional hydrophone (Dukane Corporation, St. Charles Illinois, model N3OA5A) from a 5.4-m ������Boston Whaler������ ski��. Data (acoustic signal) were recorded on audio tape (Panasonic model RQ-L340) and later decoded by playing the tape through an analog to digital converter (Ultrasonic Telemetry Sys- tems, Brea, California). Four data points per minute (at 0, 15, 30, and 45 s) were read, manually by the author, from the converter and entered into a computer spreadsheet. Temperature-depth profiles of the water column were obtained by lowering another transmitter early during the tracking period and again at the end. They were conducted at the buoy on the south-south-east side of the islands, where sea conditions were su��ciently calm, allowing water temperature to be measured to a depth of 20���25 m. The two profiles never varied by more than 0.1 ��C. Sea surface temperature was taken at various times and locations around the islands during tracking sessions, and they never di��ered from the temperature-depth profile site by more than 0.1 ��C. Considering the small di��erences between water tempera- ture at the surface and at 20���25 m (maximum of 1.5 ��C), and the lack of a di��erence in sea surface temperature around the islands, the profiles provide valid estimates of water temperature at a shark���s mean swimming depth [for further description see Goldman et al. (1996)]. Comparing stomach temperature with estimates of water temperature at mean swimming depth gives a more precise representation of the true di��erence, whereas using sea surface temperature would underestimate it. During each tracking session the telemetered shark (or sharks) were followed for as long as conditions around the island permitted or until dusk. Weather permitting, attempts were made to relocate the shark the next day. Searches were conducted by checking the area where the animal was last tracked and proceeding around the island in an inshore/o��shore star pattern to maximize the chances of hearing the acoustic signal. These searches covered a radius around the island of approximately 3 km. Stomach temperatures were compared to estimated water temperatures at a shark���s mean swimming depth. On feeding days, stomach temperatures were reduced, presumably because of sea water ingested with food (and/or transmitter and bait). Tempera- tures for those days are presented but were not used in any cal- culation or statistical analysis (Tables 1 and 2). Data from the adult male white shark at SFI in 1991 (Goldman et al. 1996) were used with data from the two new individuals to test the thermoregula- tion hypothesis of Goldman et al. (1996). Results Temperature data were successfully obtained from three individual white sharks at SFI, one in 1991 (Goldman et al. 1996), and two in 1993 (new data reported here). Additional data were obtained in 1995 from an indi- vidual at the North Neptune Islands, South Australia. This shark, however, was injured and appeared to be in poor condition. Its stomach temperature was lower than would be expected: 2.1 ��C above surface water temper- ature. The data from that shark are not presented here. All individuals from SFI [Goldman et al. (1996) and the two from 1993] were uninjured and apparently in good condition. A total of 5.5 h of data were obtained from shark #1 (Goldman et al. 1996) on 5 days over the 8-day period of 17���24 October 1991. Stomach temperature for the time periods tracked ranged from 23.4 to 27.6 ��C (Table 1). Mean stomach temperatures ranged from 24.6 to 27.3 ��C, and the overall mean was 26.6 ��C (Table 2). Sea surface temperatures (SST) ranged from 13.3 to 16.4 ��C and water temperatures at mean swimming depth ranged from 12.3 to 13.9 ��C. Di��erences between mean stomach temperatures and water temperatures at mean swimming depth ranged from 10.9 to 14.1 ��C (Table 2). A total of 7.3 h of data were obtained from shark #2 on 10 days over the 17-day period of 12���28 October Table 1 Stomach temperature ranges (Ts) and sea surface tem- peratures (SST), in ��C, from three white sharks at South Farallon Islands Day no. Shark #1a Shark #2 Shark #3 Ts Range SST Ts Range SST Ts Range SST 1 23.4���25.9b 14.7 23.4���25.2b 15.9 23.8���26.1b 13.8 2 25.8���26.8 14.6 26.0���27.0 15.5 26.2���27.1 13.7 3 25.4���27.4 16.4 25.8���26.9 15.8 25.7���26.7 13.9 4 26.8���27.6 14.4 25.6���26.5 14.4 26.0���27.1 15.8 5 26.2���26.6 13.3 25.5���26.5 14.0 25.8���27.0 15.9 6 25.7���26.6 14.0 25.3���26.3 16.0 7 25.5���26.5 13.9 24.2���25.3b 16.1 8 25.9���26.8 15.8 9 25.9���26.9 15.9 10 25.8���26.7 16.0 a Data from Goldman et al. 1996 b Indicates values are low owing to probable ingestion of sea water with prey and/or transmitter and bait at a feeding event 424