66

Regional Blood Flow in Sea Turtles: Implications for Heat Exchange

in an Aquatic Ectotherm

Sandra Hochscheid1,2,*Flegra Bentivegna2

John R. Speakman1

1Department of Zoology, University of Aberdeen, TillydroneAvenue, Aberdeen AB24 2TZ, Scotland, United Kingdom;2Stazione Zoologica Anton Dohrn, Villa Comunale 1, 80121Naples, Italy

Accepted 10/26/01

ABSTRACT

Despite substantial knowledge on thermoregulation in reptiles, themechanisms involved in heat exchange of sea turtles have not beeninvestigated in detail. We studied blood flow in the front flippersof two green turtles, Chelonia mydas, and four loggerhead turtles,Caretta caretta, using Doppler ultrasound to assess the importanceof regional blood flow in temperature regulation. Mean blood flowvelocity and heart rate were determined for the water temperatureat which the turtles were acclimated (19.3�– 22.5�C) and for severalexperimental water temperatures (17�–32�C) to which the turtleswere exposed for a short time. Flipper circulation increased withincreasing water temperature, whereas during cooling, flipper cir-culation was greatly reduced. Heart rate was also positively cor-related with water temperature; however, there were largevariationsbetween individual heart rate responses. Body temperatures, whichwere additionally determined for the two green turtles and sixloggerhead turtles, increased faster during heating than duringcooling. Heating rates were positively correlated with the differencebetween acclimation and experimental temperature and negativelycorrelated with body mass. Our data suggest that by varying cir-culation of the front flippers, turtles are capable of either trans-porting heat quickly into the body or retaining heat inside thebody, depending on the prevailing thermal demands.

Introduction

Reptiles are generally defined as ectotherms. To achieve highbody temperatures (Tb’s), they migrate between different mi-

* Corresponding author. Present address: Stazione Zoologica Anton Dohrn, Villa

Comunale 1, 80121 Naples, Italy; e-mail: s.hochscheid@abdn.ac.uk.

Physiological and Biochemical Zoology 75(1):66–76. 2002. � 2002 by TheUniversity of Chicago. All rights reserved. 1522-2152/2002/7501-0069$15.00

croclimates and perform a variety of thermoregulatory behav-iours. For example, effective absorption of solar radiation en-ables the lizard Liolaemus multiformis to become active at highaltitudes even when ambient temperatures (Ta’s) are aroundzero (Pearson 1954). Early work on this subject is reviewed byAvery (1982); for more recent studies see, for example, Bauwenset al. (1996), Christian and Weavers (1996), and Seebacher(1999). However, the utility of behavioural thermoregulationis limited for reptiles that live in or near water (Seymour 1982;but see Seebacher et al. 1999). This is due to the physicalproperties of water, particularly the high thermal conductivityand absorption of infrared radiation. For example, the Gala-pagos marine iguana, Amblyrhynchus cristatus, rapidly coolsdown to Ta while it forages in water at 25�C (Bartholomew1966). Consequently, it has to warm up before and after for-aging bouts. Sometimes an animal has to interrupt foraging toemerge and warm up. These thermal constraints mean thatmarine iguanas are obliged to spend more time on land baskingin the sun than foraging in the sea (Bartholomew 1966; Trill-mich and Trillmich 1986).

Few reptiles spend significant amounts of time in water.Some freshwater turtles that seldom leave the water thermo-regulate by selecting water of particular temperatures ratherthan basking on land (Templeton 1970; Spotila et al. 1990).The absence of direct sunlight as a heat source particularlylimits behavioural thermoregulation in fully marine reptiles likesea snakes and sea turtles, which usually leave the sea only tolay eggs. Moreover, the marine environment provides a rela-tively homogeneous thermal climate so that options for theselection of optimal temperatures are reduced. Although oc-casional basking at the water surface, or on beaches, has beenobserved in some sea turtle species (green turtles [Whittow andBalazs 1982] and loggerhead turtles [Sapsford and van der Riet1979]), this behaviour does not occur frequently enough torepresent a major source of incoming energy (turtles spent onaverage 35 min basking in a 20-d observation period [Whittowand Balazs 1982; see also Sato et al. 1995]). Despite these con-straints, it has been suggested that leatherback turtles, Der-mochelys coriacea, may maintain a large difference between Tb

and Ta, even during prolonged periods of inactivity (Paladinoet al. 1990). Davenport et al. (1990) reported that leatherbackturtles are the only reptile with a peripheral layer of blubbersimilar to that of marine mammals. Moreover, other sea turtlespecies like loggerhead (Sakamoto et al. 1990; Sato et al. 1994,1995) and green turtles (Standora et al. 1982; Sato et al. 1998)are also able to sustain their Tb’s above the water temperature.

Regional Blood Flow in Sea Turtles 67

Table 1: Body masses (BM), acclimation temperature (Tac), experimental temperature (Tex), andaverage heating and cooling rates of individual turtles

Part 1 Part 2

IndividualBM(kg)

Tac

(�C)Tex

(�C)BM(kg)

Tac

(�C)Tex

(�C)Heating(�C/min)

Cooling(�C/min)

1 17.3 21.0 16.9, 19.3, 24.8, 27.2, 31.0 24.2 16.1 28.8 .018 .0072 10.0 21.0 17.1, 18.7, 23.8, 27.5, 32.0 19.0 16.0 28.9 .026 .0093 11.8 22.1 17.0, 26.2, 29.2, 30.7 … … …4 7.6 22.1 28.3 … … …5 3.35 19.9 17.1, 24.0, 26.9, 29.5 7.1 17.0 29.1 .05 .0316a 2.29 19.9 17.5, 24.0, 27.4, 30.0 6.7 23.5 28.9 .04 .0316b 2.29 19.9 17.5, 24.0, 27.4, 30.0 6.7 16.1 28.7 .046 .0227 … … … 2.1 16.0 28.9 .061 .0378 … … … 22.6 19.4 14.9 x .0179a … … … 16.7 20.0 24.1 .026 .0119b … … … 16.7 19.4 29.0 .02 .00710a … … … 59.3 20.1 24.0 nh x10b … … … 59.3 19.3 29.0 .006 x

Note. Individuals 1 and mydas. Individuals 3– caretta. heating (or cooling, respectively)2 p Chelonia 10 p Caretta x p still

30 min after retransfer; heating after 30 min in Tex.nh p nota Experiment 1.b Experiment 2.

These observations have raised the possibility that sea turtlesare at least partially endothermic (Heath and McGinnis 1980;Standora et al. 1982; Spotila and Standora 1985; Goff and Sten-son 1988; Davenport 1997). Whatever their status, the regu-latory mechanisms of heat exchange in sea turtles are still poorlyunderstood. There is, however, evidence that they heat up fasterthan they cool down, indicating the involvement of physio-logical control mechanisms (Heath and McGinnis 1980; Smithet al. 1986).

Since the carapace is a good insulator (Standora et al. 1982),heat exchange with the surrounding water probably primarilytakes place across other parts of the body. The front flippers,for example, possess both a large surface area over which heatcan be exchanged and a relatively high resistance to blood flowsince resistance increases with distance to the heart. If thisresistance were regulated by vasoconstriction and vasodilata-tion, then the circulatory system of the flippers would be animportant thermoregulatory tool. The role of flippers and otherappendages in thermoregulation has already been shown for avariety of marine mammals, the bodies of which are insulatedby a thick fat layer (Scholander and Schevill 1955; Kvadsheimand Folkow 1997; Schmidt-Nielsen 1997; Noren et al. 1999).We hypothesised that sea turtles are capable of altering theblood flow in the flippers to either transport heat into the bodyor restrict its loss. To test this hypothesis, we studied the bloodflow in the front flippers of green turtles and loggerhead turtlesat different water temperatures using a noninvasive Dopplerultrasound technique.

Material and Methods

The present study consists of part 1 (blood flow) and part 2(Tb), which were conducted in November and December 1999and in December 2000 and January 2001, respectively. A totalof 10 turtles (two juvenile green turtles, seven juvenile and oneadult loggerhead turtle) were used (Table 1). The turtles werekept in separate tanks at the Aquarium of Naples (StazioneZoologica Anton Dohrn, Naples, Italy) and were supplied withcirculating seawater, which was pumped from the Gulf of Na-ples. Three different sizes of tanks were used. The smaller turtleswere kept in 200-L tanks, the bigger turtles were kept in 400-L tanks, and the adult turtle (turtle 10) was kept in a 1,000-Ltank. Both species were fed daily between 1000 and 1200 hours(except weekends) with anchovies, Engraulis encrasicholus.

Doppler Ultrasound and Principles of Measurements

We measured blood flow using a handheld bidirectional Dopp-ler ultrasound device (Multi Dopplex II, Huntleigh Diagnostics,Cardiff, U.K.). All measurements were made with a vascularprobe connected to the control unit, which emitted ultrasoundat a frequency of 5 MHz. However, for the two smallest log-gerhead turtles (turtles 5 and 6; Table 1), a less powerful probeof 8 MHz was used. The emitted ultrasound (fe) was reflectedby the moving blood cells, mainly the erythrocytes, at a differentfrequency (fr). The resulting Doppler shift ( ) gaveF p f � fD r e

information about direction, character, and velocity of the flow.

68 S. Hochscheid, F. Bentivegna, and J. R. Speakman

Figure 1. Schematic waveform of blood flow during one heart cycle. The y-axis is either Doppler shift frequency or blood flow velocity. S pheight of waveform during systole; height during diastole; blood flow velocity.maximum D p minimum Mean p mean

Blood flow was audible in the loudspeaker of the Doppler.Arteries were distinguished from veins because they emitted ahigh-pitched pulsatile sound, while veins emitted a nonpulsatilesound similar to rushing wind. The Doppler unit was connectedvia an RS232 interface to a laptop personal computer (ToshibaTecra 730CDT) running the Dopplex Reporter software pro-gramme (Huntleigh Diagnostics 1997). The audio signal wastransformed and displayed as a waveform on the screen (Dopp-ler shift frequency or flow velocity, respectively, vs. time). Inthis way, each measurement was recorded on-line, and datawere stored for further waveform analysis. Because of the lowheart rate (minimum 7–12 beats/min [bpm]), a measuring pe-riod of 20 s (i.e., the longest period permitted by the software)was chosen to facilitate the recording of at least two completeheart cycles.

Heart rate was calculated as the average number of cardiaccycles per minute using the reciprocal of the interbeat intervalcalculated from the complete cycles determined from the wave-form. For this, and for all other calculations (see “Data Anal-ysis”), we used a mean � SD of heart cycles (range3.2 � 1.81–8) depending on the length of the interbeat interval.

Since Doppler shift frequency is proportional to flow velocity,it was possible to calculate the blood flow velocity using theequation (Kremkau 1993)

2f # n # cosveF p f � f p , (1)D r e c

where n is the reflector speed (p blood flow velocity), v is theangle between the probe and the direction of the blood flow,and c is the speed of ultrasound in the tissue. The calculationof the flow velocity therefore depends on the angle v. The

optimum angle range for most vascular studies is recommendedto be between 30� and 60�. In this study, the best signal wasobtained when using an angle of 30�.

A useful method to characterise blood flow is to use definedparameters of the recorded waveform to calculated indices suchas Pourcelot’s resistance index (RI; Evans et al. 1989). To cal-culate this index, the maximum (S) and minimum (D) heightsof the waveform during systole and diastole, respectively, mustbe known (Fig. 1). The RI gives a numerical value of the re-sistance opposed to the blood flow and is calculated as

S � DRI p , (2)

S

where . The advantage of taking such a ratio is that0 ! RI ≤ 1both the numerator and denominator include the cosine of theangle between the Doppler probe and the blood vessel (see eq.[1]). The cosine term therefore cancels, and the index is in-dependent of angle. We therefore used RI to control for theeffect of probe angle on the measured blood flow velocities.

Blood Flow Measurements on Sea Turtles

Before the Doppler measurements, the water level in the tankwas lowered so that the central part of the carapace and a smallarea on the neck and head were above the water surface. Theturtles then remained in a relaxed state with their head sub-merged, and blood flow was measured without any disturbancedue to movements. The Doppler probe was positioned on theventral surface of the proximal front flipper, where the skinwas thin enough for the ultrasound to penetrate. The large ovalscale that is adjacent to the third large scale on the posterior

Regional Blood Flow in Sea Turtles 69

edge of the flipper was taken as an orientation point to ensurethat the same area of approximately 1 cm2 proximal of thisscale was always examined. The flipper was taken as an ana-tomic plane to which the probe was oriented at a 30� angle.Presuming that the vessels under examination run parallel tothe flipper surface, we took this 30� angle for all calculationsof blood flow velocity. However, because the probe was hand-held and values of calculated flow velocity from observed FD

are only as accurate as the estimated Doppler angle v, we al-lowed a deviation from this angle of �5�. In this case, a de-viation of 5� would result in an accuracy of �5%. Precedingtest runs during which both flippers were examined showedthat there was no difference in blood flow between the twofront flippers. Subsequently, only the left front flipper of eachturtle was used. In addition, the blood flow in the neck ofturtles 5 and 6 was measured on the left lateral side of the neck(relatively proximally to the carapace).

In all the measurements, the probe was held so that it pointedtoward the body core. The time between the first placement ofthe Doppler probe and the first recorded blood flow variedaccording to the strength of the blood flow signal. Generally,in relatively warm water, the blood flow signal was strongenough that an artery was found in !30 s. However, in relativelycold water, there were only weak blood flow signals, whichmade it difficult to find an artery. In these situations, it tookup to a maximum of 5 min to complete the first recordedmeasurement. Once a clear signal was received from an artery,recordings of blood flow were taken almost continuously; theonly interruptions were due to the closing of the recordingwindow after 20 s and the opening of a new recording windowof the Reporter software programme. A Doppler session wasterminated when an average of at least seven waveforms wasrecorded. The time until this was achieved varied due to in-terruptions of the measurements caused by movements of theturtle. In the latter, the measurements could only be continuedwhen the turtles rested again. However, the average durationof one Doppler session seldom exceeded 15 min; in a few cases,the maximum duration was 30 min.

In November, 2000, we took turtle 5 to Vincenzo MonaldiHospital, Naples, to examine the flipper vessels using ColourDoppler Echography (Acuson Sequoia). With the help of thissophisticated ultrasound machine, it was possible to visualisethe local arrangement of arteries and veins in a cross sectionof 16 mm2. The investigation was undertaken with a 7.5-MHzprobe on the unanesthetised turtle while it was lying on itscarapace.

Transfer Trials: Part 1

Turtles were exposed to a sudden temperature change by trans-ferring them from their home tank into an adjacent tank ofdifferent water temperature (Tw). The temperature in the trans-fer tank was manipulated in such a way that differences to the

acclimation temperature (Tac) of ca. 3�, 6�, and 9�C were es-tablished. Since the turtles were subject to natural variation inseawater temperature, Tac and, hence, experimental temperature(Tex) were slightly different for each turtle. All correspondingtemperatures are listed in Table 1.

Two of the turtles were also transferred into a tank thatcontained water of the same temperature as the home tank.Blood flow measurements conducted on these control animalswere taken to evaluate the handling effect on the turtles.

The six turtles used in this experimental part were dividedinto three groups (group 1: turtles 3 and 4; group 2: turtles 1and 2; group 3: turtles 5 and 6) that were examined in con-secutive experimental periods (duration each: 4–5 d). Eachturtle was exposed only once per day to a Tex. Blood flow wasmeasured as soon as the turtle rested calmly in one corner ofthe tank. Usually the first measurement after a transfer wasinitiated immediately after the turtle had been placed into thetransfer tank. In some cases, however, especially when trans-ferred into colder water, the turtles moved around for 1–2 minbefore they calmed down and the measurements could be un-dertaken. The turtles raised their heads to breathe once or twiceduring such a session. The mean length of apnoeic intervalswas 17.4 min for loggerhead turtles in 20�C (F. Bentivegna, S.Hochscheid, and C. Minucci, unpublished data). Similar ap-noeic interval lengths were observed for the green turtles. Sinceblood flow measurements during breathing were normally af-fected by the respiratory movements of the turtle (e.g., thelifting of the head above the water line), only blood flow dataduring apnoea were used in this present analysis except wherestated otherwise.

Transfer Trials: Part 2

A second set of transfer trials was conducted to determine theeffect of a 30-min transfer experiment (as described in “TransferTrials: Part 1”) on Tb (see Table 1). Turtles 1, 8, 9, and 10 werefed miniature temperature loggers (DS1921 Thermocron 1-Wire iButton, Dallas Semiconductor, Dallas) that recorded Tb

in 1-min intervals (resolution: 0.5�C; accuracy: �1�C).Tb of the other turtles was taken in the cloaca using a digital

thermometer (accuracy: �0.3�C; Checktemp, Hanna Instru-ments, Leighton Buzzard, United Kingdom), which was alsoused to measure Tw. Cloacal Tb of each of these turtles wasmeasured three times: (1) before the transfer, (2) after a periodof 30 min during which the turtle remained in the transfertank, and (3) 30 min after the turtle was returned to the hometank. The corresponding heating (kh) and cooling (kc) rates(listed in Table 1) during a 30-min period were calculated usingthe equation

ln (T � T ) � ln (T � T )ex b0 ex b30k p (3)h t

70 S. Hochscheid, F. Bentivegna, and J. R. Speakman

for heating and the equation

ln (T � T ) � ln (T � T )b�re b0 b30�re b�rek p (4)c t

for cooling. Tex was the experimental water temperature thatwe assumed was also the asymptotic Tb the turtles were headingtoward during the heating process (Tb measurements on seaturtles have shown that there is no difference between Tb andTw; see “Results,” but see also Read et al. 1996). Tb0 was thebody temperature at the start of the transfer, and was theTb�re

body temperature just before the retransfer; Tb30 was the bodytemperature after 30 min in the experimental temperature, and

was body temperature 30 min after the retransfer. TheTb30�re

period t for which the heating and cooling rates were deter-mined was 30 min.

Data Analysis

S, D, and mean blood flow velocity (M) were determined foreach recorded heart cycle using the Dopplex Reporter program.RI was calculated from the obtained values for S and D, andboth RI and M were averaged over the total number of analysedheart cycles at each temperature.

We analysed the factors influencing the blood flow param-eters (M, RI, and heart rate) using generalised linear modelling(GLM; Minitab 11, Minitab, State College, Pa.). We includedindividual as a factor in the analyses to account for the repeatedmeasurements involved and Tw (includes both Tac and Tex) asa covariate. We analysed the changes in Tb over time usingregression analysis (Minitab 11).

Results

The control experiment ( ) revealed no changes inT p Tex ac

heart rate for either turtle (turtle 2: 5 bpm before and after thetransfer; turtle 3: 8 bpm before and after the transfer). M in-creased slightly from 0.2 cm/s to 0.4 cm/s in turtle 3 but de-creased in turtle 2 from 1.4 cm/s to 1.0 cm/s. These results anddirect observation of the animals indicate that the turtles werenot disturbed in a manner that affected blood flow parameters.This was further supported by the bidirectional change in bloodflow (see below) induced by Tex, in contrast to the expectationof a unidirectional elevation of blood flow induced by stress.

Blood Flow in the Front Flippers

When examining a flipper with the Doppler probe, we foundthat a minimal change in probe position resulted in the signalof either a vein or an artery. The Colour Doppler Echographyverified that veins are located in the vicinity of arteries. Wefound smaller veins (diameter ca. 0.15 mm) surrounding largerarteries (diameter ca. 0.3–0.5 mm), but there was also one

major vein running along a major artery (both ca. 0.3 mm indiameter).

The decrease in Tac due to seasonal change in seawater tem-perature also resulted in different blood flow patterns in theindividual turtles. It was possible to obtain a clear Dopplersignal from turtles 1–4 in their home tank ( �–22�C).T p 21ac

However, blood flow in turtles 5 and 6 was heard only as verylow pulses during breathing periods in water of their Tac (20�C).The signal was too weak to be recorded by computer and,moreover, was likely to be affected by movements of the turtles.During apnoea, circulation in their flippers was below theDoppler detection threshold (K0.1 cm/s).

There was no blood flowing in the flipper arteries at Tac atthe end of the diastolic phase ( ; top graph, Fig. 2a, 2b).D p 0M was between 0.4 and 0.5 cm/s and considered to be 0 inthose cases where blood flow could not be recorded (e.g., turtles5 and 6). Maximum recorded blood flow velocities at systolicpeak (S) were between 2.5 and 2.8 cm/s. Transfer into warmerwater resulted in an elevated blood flow velocity (bottom graph,Fig. 2a, 2b). Some of the measurements of the elevated bloodflow were obtained within 30 s after the turtle had been exposedto the warmer Tw. A series of good waveforms was recordedfor one turtle during a warm water exposure at a time of 0.5,2, 5, and 20 min after transfer. All waveforms were similar,with similar mean blood flow velocities, so that no change inblood flow over exposure time was detected.

The maximum recorded S was 13.5 cm/s in water of 30�C,and M was between 3.1 and 11.1 cm/s at temperatures between30� and 32�C. When the turtles were transferred into colderwater, the blood flow was greatly reduced, and in three turtles,it virtually ceased (!0.1 cm/s). In water of 17�C, M was 0.1cm/s, and S was between 1.4 and 1.9 cm/s in those turtles forwhich blood flow data at this low temperature could beobtained.

Overall, blood flow velocity in the front flippers of bothspecies increased significantly with Tw (GLM after loge trans-formation of the data: ; ; ; Fig. 3).F p 134.99 df p 26 P ! 0.001The response of the blood flow to Tw was not significantlydifferent between the individuals ( ; ) or be-F p 1.19 P 1 0.05tween the two species ( ; ).F p 1.28 P 1 0.05

The analysis of the angle-independent RI gave a similar sig-nificant relationship (GLM after arcsine transformation: F p

; ; ), thus supporting the accuracy of the79.19 df p 26 P ! 0.001velocity data. The only difference was that RI, as expected,decreased with increasing Tw. Differences between the individ-ual responses were not significant (slopes not significantly dif-ferent: ; ), and the intercepts of the responsesF p 1.43 P 1 0.05also did not differ significantly between individuals (F p

; ). In a few exceptional cases, it was possible to2.71 P 1 0.05hold the Doppler probe in place while the turtle was breathingso that blood flow could be recorded. M and heart rates duringbreathing and during apnoea are presented in Table 2.

Figure 2. Arterial blood flow in the left front flipper of a loggerhead turtle (a) and a green turtle (b) at different temperatures; dashed lineindicates mean blood flow velocity.

72 S. Hochscheid, F. Bentivegna, and J. R. Speakman

Figure 3. Mean blood flow velocity in the front flipper arteries ofloggerhead and green turtles at varying water temperatures. Differentsymbols mark individual turtles as shown in the key above. T pac

temperature at which turtles were acclimatised.water

Table 2: Mean blood flow velocity (M) and heart rate (HR)during breathing and during apnoea in three turtles

Turtle

Breath Apnoea

Tac

(�C)M(cm/s)

HR(beats/min)

M(cm/s)

HR(beats/min)

1 .9 14 .1 14 21.11 1.4 28 .2 13 22.71 2.8 17 .3 9 22.72 1.6 23 .5 10 22.83 .8 19 .3 18 22.4

Note. temperature.T p acclimationac

Blood Flow in the Neck

It was generally more difficult to measure blood flow in theneck since the turtles retracted their head once the probe camein contact with the skin. Consequently, we did not measureneck blood flow routinely, but we were able to obtain somemeasurements on turtles 5 and 6, which did not retract theirnecks. The waveforms looked different from those that weremeasured in the flippers because blood flow was generally re-stricted to the systolic period (Fig. 4). Figure 4 shows a coupleof waveforms for each of the following situations: a coolingtransfer into 17�C, in the home tank of C, and aT p 20�ac

warming transfer into 27�C. M increased only moderately be-tween 17� and 27�C (4.7-fold as compared to eightfold in theflipper). In contrast to the absent circulation in the flippers ofthese turtles at Tac of 20�C (see above), it was still possible toobtain blood flow recordings from their neck arteries. The samewas true for the coldest Tex. Flow in the diastolic period wasonly recorded during warming in water of 27�C.

Heart Rate

Mean heart rates of all turtles during submergence in the hometanks were between 8.5 and 15.8 bpm. During breathing epi-sodes, heart rate was 1.5- to 2.5-fold the heart rate duringapnoea. Occasionally (eight out of 40 sessions), there was amarked arrhythmia in heart rate consisting of “double beats”with one short cycle (around 2.5 s) followed by one long cycle(around 5.5 s). However, seven of these eight arrhythmic heart

beats occurred in the same turtle, and the blood flow data inall eight cases were not considered for further analysis. In thetransfer experiments, heart rate increased significantly with in-creasing Tw (GLM: ; ; ; Fig. 5), al-F p 17.48 df p 26 P ! 0.01though the individuals showed different responses (differentslopes: ; ; Fig. 5).F p 3.83 P ! 0.05

Body Temperature

Tb’s (range: 16�–23.9�C) of all turtles varied linearly in corre-spondence with variations in Tac (range: 16�–23.5�C) duringthe study period (regression equation: ;T p 1.07 # T � 0.8b ac

; ANOVA: ; ; ). None of2r p 0.9792 F p 282.32 df p 7 P ! 0.001the turtles established a Tb that was equal or similar to Tex after30 min. Tb’s obtained from the ingested data loggers showedthat these turtles were still in the linear heating or cooling phaseduring the 30 min after exposure to the changed temperature(Fig. 6). This could not be ascertained for the smallest turtle(turtle 7) for the lack of continuous Tb measurements. However,even though turtle 7 heated at the fastest rate of 0.06�C/min,it still had a Tb that was 2�C less than Tex at the end of thetransfer period. When the turtles were returned to their hometank after heating, they all cooled down again but at a signif-icantly lower rate than that at which they had heated up (pairedt-test: ; ; ; Table 1).t p 8.25 N p 8 P ! 0.001

Discussion

Among the present marine reptiles, sea turtles are the mostwidely distributed. Although they are mainly tropically andsubtropically distributed, they can also experience colder watertemperatures (e.g., green turtles encounter cooler oceanic waterwhen migrating between Brazil and Ascension Island; Mortimerand Portier 1989; Luschi et al. 1998). Apart from these long-distance migrations, sea turtles also encounter vertical tem-perature differences while diving (Sakamoto et al. 1990). Sincesea turtles are mainly confronted with varying Tw when theyare actively swimming, the question arises as to whether theirTb has to stay within a certain range for them to sustain their

Regional Blood Flow in Sea Turtles 73

Figure 4. Arterial blood flow in the neck of a loggerhead turtle (turtle 5). Vertical lines indicate onset of a measurement in water of differenttemperature; dashed lines are mean blood flow velocities.

Figure 5. Heart rate of turtles 1–6 at various experimental water tem-peratures. Symbols as in Figure 3. Dashed lines indicate exemplarydifferences between individuals.

activity. If this is the case, they should possess adaptations toregulate heat flow.

In this study, we found that when Tw increased above Tac,blood circulation in the front flippers increased. This wasachieved by a faster blood flow velocity and by a lower resistanceopposed to the flow (e.g., widening of the vessels). In contrast,

blood flow came virtually to a halt when Tw dropped belowTac. This may reduce heat exchange, helping to retain heat insidethe body.

Such a state could not be sustained indefinitely because thetissues would be deprived of oxygen. This problem was prob-ably avoided by the circulatory response to breathing. For ashort period during and after the turtle surfaced to breathe,heart rate as well as blood flow velocity increased, presumablyfacilitating replenishment of oxygen and removal of toxic me-tabolites. It might be argued that normally most of the heat isexchanged during these breathing-induced increases in circu-lation. However, these periods accounted only for about 5%of the time (e.g., 3 min in 1 h), which is representative of thatobserved in free-living sea turtles (Renaud and Carpenter 1994;van Dam and Diez 1996). Although obtaining accurate mea-surements during breathing was difficult, heart rate was, if atall, only elevated 1.9–2.3 times during these short periods, andblood velocity was elevated three- to ninefold. Heat flow willdepend on a number of characteristics in addition to bloodflow velocity, but assuming these other unmeasured parametersremain unaffected, heat exchange during breathing might beexpected to account for around 20% of the total heat exchange(average five times greater flow for 5% of the time).

Circulation to the head never ceased, even at the lowest Tw,suggesting that the brain was continuously supplied with ox-ygen. This may be part of the phenomenon that has beendescribed as the “dive response” (Butler and Jones 1997; Elsner1999); most aquatic air-breathing animals have reduced cir-culation while they are diving, providing only the most im-

74 S. Hochscheid, F. Bentivegna, and J. R. Speakman

Figure 6. Increase in body temperature (Tb) of turtle 1 during heatingafter a transfer from 16.1�C water into 28.8�C water and decrease ofTb during cooling after the retransfer into 16.1�C water (retransfer wasinitiated 30 min after transfer).

portant organs (e.g., brain) with oxygen. The mechanism servesprimarily to reduce total oxygen consumption and thus to max-imise the time spent underwater (Handrich et al. 1997). Inaddition to this function, the reduced peripheral circulationmay also assist in keeping heat inside the body core. However,when the animal warms up during periods of high activity, itmay need to dissipate heat via increased blood flow to theperiphery. The different blood flow responses in neck and flip-pers to changing environmental temperatures support our hy-pothesis that the flippers play a role in heat exchange.

The vascular net of the flippers is organised similarly to thatfound in the limbs of humans (M. Scherillo, personal com-munication). Even though Mrosovsky (1980) reported that nowell-developed countercurrent heat exchangers were found inthe flippers of a loggerhead hatchling, it is possible that thereis some heat exchange between adjacent arteries and veins.Because of this and the large surface to volume ratio, the flippersare more suitable for modulating the heat exchange than is,for example, the carapace with its much lower thermal con-ductivity (Heath and McGinnis 1980).

The method used in this study did not allow us to recordchanges in blood flow continuously. Although it appeared asif elevated blood flow during warm water exposure remainedbasically on the same level, this could not be established quan-titatively. However, since none of the turtles reached a newstable Tb during the transfer experiment, it can be inferred thatblood flow remains elevated at least until a new Tb is stabilised.

A great deal of our knowledge of reptilian thermal biology

has been obtained from heating and cooling experiments. Insuch trials, the study animal is placed in experimentally ma-nipulated Ta. Under these controlled conditions, it is possibleto monitor physiological data such as heart rate, Tb, oxygenconsumption, blood flow, and so forth. One important featureof these experiments is that a relatively rapid change in tem-perature is forced on the animal, which does not have theopportunity to acclimate to the given temperature(s) or tomove to another place of more favourable temperature. Becauseectotherms, in contrast to endotherms, lack significant amountsof external insulation, they are more immediately affected byexternal temperature changes, and hence their Tb drops or risesaccordingly. However, heating and cooling experiments on Ga-lapagos marine iguanas, Amblyrhynchus cristatus (Bartholomewand Lasiewski 1965), American alligators, Alligator mississip-pinesis (Smith 1976 cited in Smith 1979), aquatic turtles, Pseu-demys floridana and Chelydra serpentina (Weathers and White1971), green turtles, Chelonia mydas (Heath and McGinnis1980; Smith et al. 1986; this study), and loggerhead turtles (thisstudy) revealed that all these animals warm up faster than theycool down. This implicates the involvement of a physiologicalregulatory control mechanism. Morgareidge and White (1969)inferred that vasomotor control of cutaneous blood flow en-ables the Galapagos marine iguana to retain thermal stabilityunder extreme environmental conditions (e.g., when movingfrom hot lava substrate into cold water to forage). Other factors,such as thermal variation in blood viscosity, may play a roleof as yet unknown importance. Sea turtles live in a compar-atively more stable environment, yet Smith et al. (1986) claimedthat they are efficient at regulating their heating and coolingrates. This interpretation is supported by the circulatorychanges presented in this article. A 10�C difference in Tw causeda greater than 100-fold increase in blood flow velocity from0.1 cm/c to 11.1 cm/s.

In many heating and cooling experiments, heart rate is con-sidered to be an important, variable factor that accompaniesdifferent heating and cooling rates. Heart rate is typically fasterduring heating than during cooling (lizards: reviewed by Bar-tholomew 1982; green turtles: Smith et al. 1986). Although wealso observed an average increase, we observed significant in-dividual variation in the response of heart rate to increasingTex. Our data suggest that heart rate does not necessarily reflectregional heat-exchange regulation. Whereas the turtle in Figure2a almost doubled its heart rate in 10� warmer water, the turtlein Figure 2b had almost the same heart rate during heating asit had at Tac. Despite this, there was a dramatic increase inflipper blood flow in both turtles. These data do not supportSmith et al.’s (1986) statement that changes in peripheral bloodflow alter heart rate because heart rate was independent of Tex

in some turtles, yet in other animals we determined parallelalterations in peripheral blood flow with Tex (see also Morgar-eidge and White 1969; Baker et al. 1972).

In summary, using a simple, noninvasive method, we have

Regional Blood Flow in Sea Turtles 75

shown that externally induced heating rates are accompaniedby faster blood flow velocities in the front flippers of bothloggerhead and green turtles. This response was more pro-nounced at higher experimental temperatures. In contrast,cooling resulted in drastically reduced blood flow. This leadsus to the conclusion that the front flippers play a previouslyundescribed role in the heat exchange between sea turtles andtheir environment.

Acknowledgments

S.H. was supported by a research grant of the Deutscher Aka-demischer Austauschdienst. The Doppler equipment was kindlyprovided by Huntleigh Diagnostics. We are especially gratefulto Gianfranco Mazza as well as Mariapia Ciampa, Angela Pag-lialonga, and Isabella D’Ambra for their help during the ex-periments. We greatly appreciated the medical advice and tech-nical support of Marino Scherillo and his enthusiasticcolleagues of the Divisione di Cardiologia, Ospedale VincenzoMonaldi, Naples. Further, we acknowledge the excellent adviceand assistance of Andrew Fairhead from the Department ofBiomedical Physics, University of Aberdeen, and Ian Cadlefrom the Vascular Laboratory, Aberdeen Royal Infirmary, aswell as John Gowers and Sharon Thomas-Jones from HuntleighDiagnostics. Many thanks are due to David Gremillet for hispatience and many helpful discussions. We would also like tothank an anonymous referee for his helpful comments thathelped to improve earlier versions of the manuscript.

Literature Cited

Avery R.A. 1982. Field studies of body temperatures and ther-moregulation. Pp. 93–166 in C. Gans and F.H. Pough, eds.Biology of the Reptilia. Vol. 12. Academic Press, London.

Baker L.A., W.W. Weathers, and F.N. White. 1972. Temperatureinduced peripheral blood flow changes in lizards. J CompPhysiol 80:313–323.

Bartholomew G.A. 1966. A field study of temperature relationsin the Galapagos marine iguana. Copeia 1966:241–250.

———. 1982. Physiological control of body temperature. Pp.167–211 in C. Gans and F.H. Pough, eds. Biology of theReptilia. Vol. 12. Academic Press, London.

Bartholomew G.A. and R.C. Lasiewski. 1965. Heating and cool-ing rates, heart rate and simulated diving in the Galapagosmarine iguana. Comp Biochem Physiol 16:573–582.

Bauwens D., P.E. Hertz, and A.M. Castilla. 1996. Thermoreg-ulation in a lacertid lizard: the relative contributions of dis-tinct behavioral mechanisms. Ecology 17:1818–1830.

Butler P.J. and D.R. Jones. 1997. Physiology of diving of birdsand mammals. Physiol Rev 77:837–899.

Christian K.A. and B.W. Weavers. 1996. Thermoregulation of

monitor lizards in Australia: an evaluation of methods inthermal biology. Ecol Monogr 66:139–157.

Davenport J. 1997. Temperature and the life-history strategiesof sea turtles. J Therm Biol 22:479–488.

Davenport J., D.L. Holland, and J. East. 1990. Thermal andbiochemical characteristics of the lipids of the leatherbackturtles Dermochelys coriacea: evidence of endothermy. J MarBiol Assoc UK 70:33–41.

Elsner R. 1999. Living in water: solutions to physiological prob-lems. Pp. 73–116 in J.E. Reynolds and S.A. Rommel, eds.Biology of Marine Mammals. Smithsonian Institution, Wash-ington, D.C.

Evans D.H., W.N. McDicken, R. Skidmore, and J.P. Woodcock.1989. Doppler Ultrasound—Physics, Instrumentation, andClinical Applications. Wiley, New York.

Goff G.P. and G.B. Stenson. 1988. Brown adipose tissue inleatherback sea turtles: a thermogenic organ in an endo-thermic reptile? Copeia 1988:1071–1075.

Handrich Y., R.M. Bevan, J. Charrassin, P.J. Butler, K. Puetz,A.J. Woakes, J. Lage, and Y. Le Maho. 1997. Hypothermiain foraging king penguins. Nature 388:64–67.

Heath M.E. and M. McGinnis. 1980. Body temperature andheat transfer in the green sea turtle, Chelonia mydas. Copeia1980:767–773.

Huntleigh Diagnostics. 1997. Dopplex Reporter. Version 3.00.Huntleigh Technology, Cardiff.

Kremkau F.W. 1993. Diagnostic Ultrasound—Principles andInstruments. Saunders, Philadelphia.

Kvadsheim P.H. and L.P. Folkow. 1997. Blubber and flipperheat transfer in harp seals. Acta Physiol Scand 161:385–395.

Luschi P., G.C. Hays, C. Del Seppia, R. Marsh, and F. Papi.1998. The navigational feats of green sea turtles migratingfrom Ascension Island investigated by satellite telemetry.Proc R Soc Lond B 265:2279–2284.

Morgareidge K.R. and F.N. White. 1969. Cutaneous vascularchanges during heating and cooling in the Galapagos marineiguana. Nature 223:587–591.

Mortimer J.A. and K.M. Portier. 1989. Reproductive homingand intenesting behavior of the green turtle (Chelonia mydas)at Ascension Island, South Atlantic Ocean. Copeia 1989:962–977.

Mrosovsky N. 1980. Thermal biology of sea turtles. Am Zool20:531–547.

Noren D.P., T.M. Williams, P. Berry, and E. Butler. 1999. Ther-moregulation during swimming and diving in bottlenose dol-phins, Tursiops truncatus. J Comp Physiol B 169:93–99.

Paladino F.V., M.P. O’Connor, and J.R. Spotila. 1990. Metab-olism of leatherback turtles, gigantothermy, and thermoreg-ulation of dinosaurs. Nature 344:858–860.

Pearson O.P. 1954. Habits of the lizard Liolaemus multiformismultiformis at high altitudes in southern Peru. Copeia 1954:111–116.

Read M.A., G.C. Grigg, and C.J. Limpus. 1996. Body temper-

76 S. Hochscheid, F. Bentivegna, and J. R. Speakman

atures and winter feeding in immature green turtles, Chelniamydas, in Moreton Bay, Southeastern Queensland. J Herpetol30:262–265.

Renaud M.L. and J.A. Carpenter. 1994. Movements and sub-mergence patterns of loggerhead turtles (Caretta caretta) inthe Gulf of Mexico determined through satellite telemetry.Bull Mar Sci 55:1–15.

Sakamoto W., I. Uchida, Y. Naito, K. Kureha, M. Tujimura,and K. Sato. 1990. Deep diving behavior of the loggerheadturtle near the frontal zone. Nippon Suisan Gakkaishi 56:1435–1445.

Sapsford C. and M. Van der Riet. 1979. Uptake of solar radi-ation by the sea turtle, Caretta caretta, during voluntary sur-face basking. Comp Biochem Physiol A 63:471–474.

Sato K., Y. Matsuzawa, H. Tanaka, T. Bando, S. Minamikawa,W. Sakamoto, and Y. Naito. 1998. Internesting intervals forloggerhead turtles, Caretta caretta, and green turtles, Cheloniamydas, are affected by temperature. Can J Zool 76:1651–1662.

Sato K., W. Sakamoto, Y. Matsuzawa, H. Tanaka, S. Minami-kawa, and Y. Naito. 1995. Body temperature independenceof solar radiation in free-ranging loggerhead turtles, Carettacaretta, during internesting periods. Mar Biol 123:197–205.

Sato K., W. Sakamoto, Y. Matsuzawa, H. Tanaka, and Y. Naito.1994. Correlation between stomach temperatures and am-bient water temperatures in free-ranging loggerhead turtles,Caretta caretta. Mar Biol 118:343–351.

Schmidt-Nielsen K. 1997. Animal Physiology—Adaptationand Environment. 5th ed. Cambridge University Press,Cambridge.

Scholander P.F. and W.E. Schevill. 1955. Countercurrent vas-cular heat exchange in the fins of whales. J Appl Physiol 8:279–282.

Seebacher F. 1999. Behavioural postures and the rate of bodytemperature change in wild freshwater crocodiles, Crocodylusjohnstoni. Physiol Biochem Zool 72:57–63.

Seebacher F., G.C. Grigg, and L.A. Beard. 1999. Crocodiles as

dinosaurs: behavioural thermoregulation in very large ec-totherms leads to high and stable body temperatures. J ExpBiol 202:77–86.

Seymour R.S. 1982. Physiological adaptations to aquatic life.Pp. 1–51 in C. Gans and F.H. Pough, eds. Biology of theReptilia. Vol. 13. Academic Press, London.

Smith E.N. 1976. Heating and cooling rates of the Americanalligator, Alligator mississippinesis. Physiol Zool 49:37–48.

———. 1979. Behavioural and physiological thermoregulationof crocodilians. Am Zool 19:239–247.

Smith E.N., N.C. Long, and J. Wood. 1986. Thermoregulationand evaporative water loss of green sea turtles, Chelonia my-das. J Herpetol 20:325–332.

Spotila J.R., R.E. Foley, and E.A. Standora. 1990. Thermoreg-ulation and climate space of the slider turtle. Pp. 288–298in J.W. Gibbons, ed. The Life History and Ecology of theSlider Turtle. Smithsonian Institution, Washington, D.C.

Spotila J.R. and E.A. Standora. 1985. Environmental constraintson the thermal energetics of sea turtles. Copeia 1985:694–702.

Standora E.A., J.R. Spotila, and R.E. Foley. 1982. Regional en-dothermy in the sea turtle, Chelonia mydas. J Therm Biol 7:159–165.

Templeton J.R. 1970. Reptiles. Pp. 167–221 in G.C. Whittow,ed. Comparative Physiology of Thermoregulation. Vol. 1.Academic Press, New York.

Trillmich K.G.K. and F. Trillmich. 1986. Foraging strategies ofthe marine iguana, Amblyrhynchus cristatus. Behav Ecol So-ciobiol 18:259–266.

van Dam R.P. and C.E. Diez. 1996. Diving behavior of immaturehawksbills (Eretmochelys imbricata) in a Caribbean cliff-wallhabitat. Mar Biol 127:171–178.

Weathers W.W. and F.N. White. 1971. Physiological thermo-regulation in turtles. Am J Physiol 221:704–710.

Whittow G.C. and G.H. Balazs. 1982. Basking behaviour of theHawaiian green turtle (Chelonia mydas). Pac Sci 36:129–139.