Acceleration data reveal the energy management strategy of a marine ectotherm during reproduction

Functional Ecology 2012 doi: 10.1111/j.1365-2435.2011.01960.x Acceleration data reveal the energy management strategy of a marine ectotherm during reproduction Sabrina Fossette*, Gail Schofield, Martin K. S. Lilley, Adrian C. Gleiss and Graeme C. Hays Department of Biosciences, College of Science, Swansea University, Singleton Park, Swansea SA2 8PP, UK Abstract 1. Maintaining a high and stable body temperature is often critical for female ectotherms during reproduction. Yet this strategy may be energetically costly, and therefore challenging, during this period of already high-energy demand. 2. Here, the 6-week deployment of tri-axial accelerometers (n = 6) on a marine ectotherm, the loggerhead turtle (Caretta caretta), reproducing at the northern limit of the species breeding range (i.e. in a thermally dynamic environment) revealed the behavioural mechanisms underlying its energy management strategy during the breeding season. 3. The estimated activity levels of female loggerheads using overall dynamic body acceleration (ODBA) were high during the breeding season, suggesting that marine turtles may not be able to remain inactive for long periods in the same manner as terrestrial ectotherms, because of the thermally dynamic nature of their environment. 4. However, activity levels were not constant throughout the season, being impacted by both ambient water temperature and female reproductive status. In cold water at the beginning of the nesting season, high levels of activity suggested that females behaviourally thermoregulated by seeking out warm water patches along the shoreline. Interactions with male turtles (courtship and or avoidance) may also explain this high level of activity. As sea temperatures warmed up and the amount of energy devoted to reproduction probably increased, the turtles spent more time resting during long sequential flat-bottomed dives, and reduced any unnecessary locomotory activity. 5. Turtles may therefore adjust their activity patterns in response to seasonal variations in abiotic (i.e. ambient temperature) and biotic (i.e. reproductive status) factors. This may help minimize activity-linked metabolic rate and maximize reproductive output over a season while breeding in thermally dynamic environments. 6. A mechanistic model gave support to these empirical results. The model revealed that actively maintaining high and stable body temperature is of clear benefit to female turtles at temperate breeding sites. While energetically costly, such active thermoregulatory behaviour may speed up egg maturation, allowing turtles to initiate nesting earlier in the season, and hence maximize reproductive output. Key-words: diving, energy expenditure, locomotory costs, microhabitat, reptiles, thermal preference, viviparous Introduction In ectotherms, body temperature has profound effects on locomotory performance, physiological functions and ultimately fitness (e.g. Huey & Berrigan 2001; Herrel, James & Van Damme 2007). Therefore, the need for ectotherms to maintain optimal body temperature is critical (Martin & Huey 2008) and, in many cases, drives their temporal and *Correspondence author. sabrina.fossette@gmail.com spatial activity patterns and behaviour, such as in marine iguanas, (Wikelski & Trillmich 1994), terrestrial turtles (Dubois et al. 2009) and saltwater crocodiles (Seebacher, Franklin & Read 2005). Reproductive status may significantly affect the behavioural thermoregulation in ectotherms, notably for species living in cool regions at the margin of their thermal range, where the selection of high temperatures by reproductive females may be particularly beneficial for embryonic development (e.g. Shine 2004; Wallman & Bennett 2006; Lourdais, Heulin & Ó 2012 The Authors. Functional Ecology Ó 2012 British Ecological Society

2 S. Fossette et al. Denardo 2008). In viviparous species, maternal thermoregulatory behaviour has been reported to accelerate embryonic development and enhance hatchling survival (e.g. Shine 2004). In oviparous species, maternal thermoregulation has been shown to speed up egg maturation rates, which may impact oviposition dates, incubation temperature and, in turn, the phenotypic traits of hatchlings (e.g. Angilletta, Winters & Dunham 2000; Wallman and Bennett 2006, Berger, Walters & Gotthard 2008; Lourdais, Heulin & Denardo 2008; Weber et al. 2011). However, high body temperature also implies increased metabolic rates in ectotherms (Hochscheid, Bentivegna & Speakman 2004; Seebacher, Franklin and Read 2005), which may severely impact an individual s energy balance. The implications of a high metabolic rate may be even more challenging during periods of high-energy demand, such as reproduction, when optimizing energy reserves is critical as females need to allocate energy to clutch production without jeopardizing their own survival (e.g. Shine 2003a; Williams, Vezina & Speakman 2009). Terrestrial ectotherms have been reported to adopt different behavioural and physiological tactics to conserve energy during the breeding period, such as sedentary behaviour or the ability to downregulate organs such as the digestive tract (e.g. Lourdais et al. 2002; Shine 2003a,b). However, little is known about the energy management strategy of female marine ectotherms during reproduction. Investigating the spatio-temporal activity patterns of reproductive female ectotherms in marine habitats may thus help improve our understanding of the behavioural mechanisms underlying their energy management strategy. Loggerhead turtles (Caretta caretta, Linnaeus, 1758, Fig. 1) are long-lived reptiles, which undertake long-distance migrations from foraging to breeding areas every 1 to 3 years (Broderick et al. 2007). Females spend up to several months at the breeding area to lay multiple clutches of 50 130 eggs (Broderick et al. 2003). Between successive clutches, females remain at sea for a minimum of 9 days (i.e. an internesting interval, Broderick et al. 2002) during which egg maturation occurs. Breeding is usually associated with long-term fasting in marine turtles, and while opportunistic (e.g. Schofield et al. 2007; Fossette et al. 2008) or regular foraging (e.g. Hochscheid et al. 1999a; Hays et al. 2002b) has been reported for some sea turtle populations during the nesting season, reproduction is theorized to be primarily fuelled from stored energy reserves (Bonnet, Bradshaw & Shine 1998). A large rookery is situated at the northern limit of the species breeding range in Laganas Bay, Zakynthos Island (Greece) in the Mediterranean Sea (Margaritoulis 2005). At this breeding ground, sea surface temperatures vary seasonally and are generally cool (between 13 and 22 C) during the first couple of months of the breeding season (Schofield et al. 2009). Being ectothermic, the loggerhead body temperature is largely driven by ambient water temperature (Hochscheid, Bentivegna and Speakman 2004), and hence, these cool temperatures at Zakynthos are likely suboptimal for reproduction (Hamann, Limpus & Read 2007). Hence, duration of the breeding season for a female loggerhead in Greece may be limited by two main Fig. 1. Female loggerhead turtle Caretta caretta equipped with a triaxial accelerometer during the 2010 breeding season at Zakynthos (Greece). factors: the short window of optimal temperatures for egg maturation and incubation (Margaritoulis 2005) and female energy reserves (Broderick et al. 2003). It has been previously suggested that female turtles at this site show thermoregulatory behaviour by actively seeking down-wind, shallow patches of warm water within Laganas Bay (Schofield et al. 2009). Similar cases of maternal thermoregulation have been reported at other nesting sites and for other sea turtle species (Weber et al. 2011). It is well-established that egg maturation rates in marine turtles, and hence the duration of the internesting intervals, vary with environmental temperatures (Sato et al. 1998; Hays et al. 2002a; Weber et al. 2011). Therefore, this thermoregulatory behaviour may speed up egg maturation rates and presumably help initiate nesting at an earlier date, maximizing the number of clutches produced during the available period for breeding (Schofield et al. 2009; Weber et al. 2011). Female turtles breeding at Zakynthos therefore seem ideal candidates to investigate the behavioural mechanisms underlying the energy management strategy of marine ectotherms during reproduction. Investigating at-sea activity patterns of free-ranging sea turtles remains difficult. Direct in-water observations (e.g. Schofield et al. 2007) and the use of animal-borne video cameras (Seminoff, Jones & Marshall 2006; Fuller et al. 2009) have allowed the collection of information on activity patterns for short periods of time only (i.e. a maximum of 24 h). However, the recent development of accelerometry provides new means to assimilate detailed activity budgets over relatively longer periods of time for free-ranging individuals (Wilson et al. 2006; Wilson, Shepard & Liebsch 2008; Shepard et al. 2009; Gleiss et al. 2010). Here, we first develop a mechanistic model to show how water temperature may influence the number of clutches a turtle produces during a breeding season and how this may be linked to maternal energy reserves. We use this model to predict the optimal behavioural strategy a female turtle should adopt during the nesting season. We then use tri-axial accelerometers attached to six free-living female loggerheads to test the model predic-

Accelerometry reveals turtle energy management strategy 3 tions and investigate their behavioural strategy and activity patterns to manage their energy reserves throughout the breeding season, and successfully reproduce at the margins of their range. This study is, to our knowledge, the longest deployment (i.e. 6 weeks) of tri-axial accelerometers on a free-ranging marine species and, as such, opens a new field of investigation. Materials and methods STUDY AREA AND INSTRUMENTATION The study was conducted between May and July 2010 in Laganas Bay on the Greek island of Zakynthos (37 43 N, 20 52 E). Between 9 and 11 May, before the onset of the nesting season (Margaritoulis 2005), tri-axial acceleration data-loggers (G6a, CEFAS Technology Ltd, http://www.cefastechnology.co.uk, 40*28*16 mm, 7Æ3 g in air) were attached to six female loggerhead turtles (curved carapace length range: 76 96 cm, Table 1). Devices were set to record all three acceleration channels at a frequency of 5 Hz (12 bit resolution, range ± 8 g, resolution 72 mg), pressure every 3 s (0Æ04 m depth resolution, ±1 cm accuracy) and temperature every 5 min (±0Æ1 C accuracy). Pressure and temperature were recorded continuously over a period of 40 days, while a duty cycle of 8 h on 8 h off was set for the acceleration channels. Acceleration sensors were calibrated to g (9Æ8 ms )2 ) by rotating devices through known angles in all three spatial planes. Turtles were captured at sea from a boat using the turtle-rodeo technique (Schofield et al. 2009), and equipped with data-loggers. The data-logger was fixed to a plastic plate and the plate embedded in quick-setting two-part epoxy resin (Powers Fasteners Inc., New Rochelle, NY, USA), with wooden baffles positioned at the anterior to minimize impact to the equipment. The data-logger was positioned at the highest point of the carapace and aligned with the anterior posterior axis of the turtle body. After a minimum of 40 days, the turtles were recaptured on the nesting beaches (during the second or third nesting events), and the loggers were removed and the data downloaded. DATA ANALYSIS Data were analysed using custom-written software. Nesting events were visually identified by a drop in ambient temperature concurrent with depth measures at sea surface level for a minimum of 40 min. The date of the first nesting event varied between turtles (i.e. from 20 to 36 days after deployment, Table 1). Therefore, for each turtle, the deployment period was divided into several phases according to the reproductive status of the female relative to the first nesting event: (i) 5 weeks prior to the first nesting event (W5, n = 3 turtles), (ii) 4 weeks prior to the first nesting event (W4, n = 5 turtles), (iii) 3 weeks prior to the first nesting event (W3, n = 6 turtles), (iv) 2 weeks prior to the first nesting event (W2, n = 6 turtles), (v) the week of the first nesting event (W1, n = 6 turtles) and (vi) the first internesting interval (IP1, n = 5 turtles). A second nesting event and the beginning of the second internesting interval (IP2) were recorded for two turtles during the deployment. Acceleration data-loggers measure both dynamic acceleration (i.e. owing to animal movement) and static acceleration (i.e. related to animal posture) in three orthogonal planes: heave (dorso-ventral acceleration), sway (lateral acceleration) and surge (anterior posterior acceleration). The loggers were re-calibrated after deployment following the method described by Wilson et al. (2006). Acceleration data were first smoothed over a 3-s interval (Shepard et al. 2008) and then analysed following the methods described by Wilson et al. (2006) and Gleiss, Norman & Wilson (2011b) to yield overall dynamic body acceleration of the x (surge), y (heave) and z (sway) planes (ODBA xyz, Halsey et al. 2008b). ODBA xyz constitutes a proxy for activity level (e.g. Halsey et al. 2008b; Shepard et al. 2009; Gleiss et al. 2010). For instance periods of active swimming, gliding or resting in marine animals can be identified using ODBA xyz (e.g. Fossette et al. 2010; Gleiss et al. 2011a). In captive adult green turtles, dynamic body acceleration has been found to be highly positively correlated with oxygen consumption rates, suggesting that accelerometry is a reliable method to investigate marine turtle energetics at sea (Enstipp et al. 2011). We focused our dive analysis on U-dives (i.e. dives characterized by a long flat bottom phase, Minamikawa, Naito & Uchida 1997). U-dives are usually considered as resting dives (Minamikawa, Naito & Uchida 1997; Hochscheid & Wilson 1999b; Houghton et al. 2002), and hence, we assumed that a drop in activity level (obtained by using ODBA as a proxy) would be correlated with an increase in the time spent in U-dives and vice versa. All sequential and isolated U-dives longer than 10 min were visually identified in the data set. A bout of U-dives (i.e. sequential U-dives) was defined as a group of more than two dives occurring within a period of 10 min. For each U-dive, we recorded the start and end time, the maximum depth reached and the Table 1. Information on six female loggerhead turtles equipped with tri-axial accelerometers for 40 days during the breeding season in 2010 at Zakynthos (Greece) T27 T28 T31 T32 T33** T34 CCL in cm 85 90 78 96 87 76 Date and time of deployment (GMT + 3 h) 9 May 2010, 13:00 10 May 2010, 13:00 11 May 2010, 14:00 9 May 2010, 14:00 10 May 2010, 14:00 10 May 2010, 15:00 Date of nesting events 2 June 2010 18 June 2010 1 July 2010* Ambient temperature in C (Mean ± SD) 9 June 2010 25 June 2010* 9 June 2010 26 June 2010* 29 May 2010 15 June 2010 1 July 2010* 16 June 2010 1 July 2010* 12 June 2010 28 June 2010* 22Æ9 ±2Æ4 22Æ4 ±2Æ3 22Æ8 ±2Æ3 22Æ4 ±2Æ2 21Æ9 ±2Æ1 22Æ3 ±2Æ4 ODBA in g (Mean ± SD) 0Æ0516 ± 0Æ012 0Æ0494 ± 0Æ013 0Æ0581 ± 0Æ017 0Æ0514 ± 0Æ013 0Æ0520 ± 0Æ017 0Æ0537 ± 0Æ023 Time spent in U-dives (h) 368 328 286 377 266 *Date of the nesting event when the logger was removed. **The pressure sensor of this turtle stopped after 1 week, and therefore the dive data were not analysed for this individual. CCL: curved carapace length. ODBA: overall dynamic body acceleration.

4 S. Fossette et al. duration of the dive. In addition, we calculated mean ODBA for 85 sequential U-dives, 24 isolated U-dives, 60 V-dives (i.e. dives without a bottom phase), 40 Type 3 dives (i.e. dives with a gradual ascent phase preceding the final ascent phase) and 28 Type 4 dives (i.e. dives with a steep ascent phase preceding a gradual ascent phase as defined by Minamikawa, Naito & Uchida 1997; based on time-depth profiles, Minamikawa et al. 2000). We examined the daily variation in ambient temperature recorded by the data-loggers deployed on the six turtles throughout the deployment period. In addition, the mean temperature depth profiles during daytime and night-time (based on local time of sunrise and sunset, http://www.timeanddate.com) in May and June 2010 were obtained from the six turtles by calculating the mean ambient temperature over 1-m depth bins. Statistical analyses were performed using SPSS 16.0 Ò (IBM Corp., Armonk, NY, USA). A generalized linear model (GLM) with a Gamma distribution and a log link function was used to evaluate the effect of ambient temperature, female reproductive status, female identity and date since deployment on mean ODBA throughout the breeding season. The significance of model effects was assessed using likelihood ratio tests. To avoid autocorrelation in the analysis, the dependant variable ODBA was averaged over each 8-h block. For the comparisons of different dive types, ODBA was averaged over the entire dive cycle, and a two-way ANOVA considering dive type and female identity as fixed and random factors, respectively, followed by a post hoc Tamhane s test was performed. Statistical significance was set at a =0Æ05, with data being presented in Eastern European Summer Time (GMT + 3). MECHANISTIC MODEL A mechanistic model was constructed to investigate how maternal energy expenditure increases through the nesting season according to water temperature and how this may impact the number of clutches a turtle can lay before her energy stores are fully depleted and or before the end of the available period for nesting. Two key parameters may vary with the ambient temperature during the nesting season: (i) the routine metabolic rate (RMR) and (ii) the duration of the internesting interval, and both parameters may influence reproductive output (Hays et al. 2002a; Hochscheid, Bentivegna and Speakman 2004). Higher RMR will cause body stores to be used faster, and therefore, a shorter time is available for reproduction before the energy stores are fully depleted. However, warmer temperatures induce a faster rate of egg maturation (Weber et al. 2011), shorter internesting intervals (Hays et al. 2002a) and therefore more clutches in a given time. We calculated whether the combination of (i) higher RMR and also (ii) shorter internesting intervals at higher temperatures would lead to fewer or more clutches being laid in a season. We assumed that (1) the initial water temperature was 17 C( stay cool model) or 20 C ( stay warm model), (2) the water temperature (Tw in C) increased during the nesting season in Greece following the equation. Tw tþ1 day ¼ 01 þ Tw t eqn 1 (Schofield et al. 2009), (3) the duration of the internesting interval in days varied with the water temperature following the equation. LogðInternesting IntervalÞ ¼ 225 043Tw eqn 2 (Hays et al. 2002a), (4) the turtle body mass (Mb) was 68 kg (Hays & Speakman 1991), (5) the turtle RMR varied with the water temperature following the equation ln _VO 2 ¼ 394 þ 0195Tw þ 0303 ln Mb eqn 3 ( _VO 2 in ml O 2 min )1 ) (Hochscheid, Bentivegna and Speakman 2004), (5) the clutch size was 125 eggs (Hays and Speakman 1991) and constant throughout the nesting season (Bjorndal & Carr 1989), but see (Broderick et al. 2003), (6) the energy content of a loggerhead egg was 165 kj (Hays and Speakman 1991), (7) the beginning of the breeding season was defined as when mating starts, that is, about 30 days before the first nesting event (Godley et al. 2002) and (8) the maximum number of clutches per female was four during a nesting season of about 80 90 days long (Margaritoulis 2005). Results CONCEPTUAL FRAMEWORK FOR THE LINK BETWEEN TEMPERATURE AND ENERGY STORES Our mechanistic model revealed that for any level of energy store at the start of the breeding season, staying warm allowed turtles to produce clutches faster. Therefore, depending on the duration of the available period for breeding, this strategy either led to (a) more or (b) the same number of clutches as staying cool, that is, the stay warm strategy was never outperformed by the stay cool strategy (Fig. 2). For example, with an initial energy store of 80 000 kj, both strategies would produce three clutches, but a stay warm strategy would produce these three clutches in 63 days and the stay cool strategy in 76 days (Fig. 2). This difference may be important in nesting sites where there is only a short window of optimal sand temperatures for egg incubation (Margaritoulis 2005). Likewise, in 80 days, a stay warm strategy would produce four clutches but the stay cool strategy only three (Fig. 2). In addition, the benefit of the stay warm strategy may also change according to the relationship between metabolic rate (MR) and water temperature, that is, the Q 10 effect. Low Q 10 values may notably lead to a lower overall integrated energy expenditure of the stay warm strategy compared with the stay cool strategy for producing the same number of clutches. This simple model therefore suggests that a passive stay warm strategy may always allow a female turtle to optimize her reproductive output over a nesting season. However, in an environment where sea temperature is relatively cool and not homogenous, female turtles may have to adopt an active behavioural thermoregulatory strategy in order to stay warm. They may thus have to trade off energy conservation for the maintenance of a high metabolism favourable to egg maturation. In this case, we predict that an active stay warm strategy would still outperform a stay cool strategy, but only if female turtles cease activity as sea temperatures increase in order to minimize energy expenditure and optimize energy stores for egg production.

Accelerometry reveals turtle energy management strategy 5 (a) Fig. 2. Model predictions for the energy expenditure (in kj) of a 68- kg female loggerhead turtle throughout the nesting season at an initial water temperature of 17 C (solid line, stay cool model) or 20 C (dashed line stay warm model) and the number of clutches (steps) this turtle is able to produce. The model also predicts the energy expenditure of a female actively thermoregulating at the beginning of the season to access warm sea temperatures (i.e. 20 C, dotted line, active and warm model). For instance, if the available period for nesting was limited to 80 days (vertical red line), both warm strategies would produce four clutches but the cool strategy would produce only three. If the initial maternal energy store was 80 000 kj (horizontal red line), all strategies would produce three clutches, but in different times (63 days for the stay warm strategies and 76 days for the stay cool strategy). The square indicates the start of the breeding season (see text for details). (b) EMPIRICAL DATA Five loggers successfully recorded pressure, temperature and acceleration for a total of 40 days each, while one (T33) stopped recording pressure after 7 days of deployment. Pressure data from this logger were omitted from subsequent analysis. The ambient temperature experienced by the turtles progressively increased throughout the deployment period (minimum: 16Æ8 18Æ3 C; maximum: 27Æ8 29Æ5 C, Fig. 3). In May, the ambient temperature was relatively cool (mean ± SD = 19Æ0 ±1Æ0 C, Fig. 3) and the temperature experienced at the surface was 3Æ4 C warmer than at 20 m. This difference progressively narrowed throughout the season (Fig. 3). No significant difference in the mean ambient temperature experienced by the turtles was found (mean ± SD = 22Æ4 ±2Æ0 C, Kruskal Wallis, H 5 =4Æ6, P =0Æ471, Table 1). Throughout the deployment period, the turtles were actively diving with only 15 dives deeper than 15 m (i.e. <0Æ2% of total time). Turtles performed classical dive types, that is, subsurface dives, V-dives, Type 3 and Type 4 dives, sequential U-dives and isolated U-dives, which were associated with different acceleration patterns and ODBA values (Fig. 4). Mean ODBA was not significantly different between sequential U-dives, isolated U-dives and Type 3 dives but was lower than ODBA recorded during Type 4 dives and V-dives Fig. 3. (a) Daily mean ambient temperature experienced by six female loggerhead turtles during their 2010 breeding season at Zakynthos (Greece) and (b) Mean temperature depth profiles of the turtles during the daytime (grey symbols) and night-time (black symbols) from 10 05 to 29 5(trianglesdown),30 05 to 12 06 (triangles up) and 13 06 to 20 06 (circles). (two-way ANOVA followed by a post hoc Tamhane s test, F types =18Æ552, d.f. = 4, P <0Æ0001; F Ind =1Æ735, d.f. = 4, P =0Æ188; F types*ind =5Æ072, d.f. = 16, P <0Æ0001, n = 237 dives, Fig. 4). Mean ODBA during sequential U-dives decreased with dive duration (y =0Æ081x )0Æ23, R 2 =0Æ297, F 83,1 =35Æ11, P <0Æ0001, Fig. 5). The mean percentage of time spent in U-dives per day was 33Æ6 ±9Æ0%, of which half (16Æ4 ±5Æ5%) was spent in sequential U-dives (Table 1). For four of the five turtles, the time spent in sequential U-dives per day was negatively corre-

6 S. Fossette et al. (a) (b) (c) Fig. 4. (a) Mean ODBA calculated over the entire dive cycle and (b) mean ODBA.min )1 of (c) five different dive types from left to right: V-dives, Type 4 dives, Type 3 dives, isolated U-dives and sequential U-dives (as defined by Minamikawa, Naito & Uchida 1997, Minamikawa et al. 2000) performed by female loggerhead turtles during the 2010 breeding season at Zakynthos (Greece). Differences in mean ODBA between dive types were statistically tested using a two-way ANOVA followed by a post hoc Tamhane s test. Different letters (a, b) indicate significant (P < 0Æ05) differences between dive types. lated with daily mean ODBA (Spearman s correlation, P <0Æ05 in all four cases, n =40daysineachcase). Over the entire period, ODBA (30 min mean) varied from 0Æ021 to 0Æ153 g, with a mean value of 0Æ053 ± 0Æ016 g for all six turtles (Table 1). A GLMM revealed that ambient temperature, female reproductive status, female identity and date since deployment all had a significant effect on the variation of mean ODBA throughout the breeding season. When considering female identity as a random factor, the effects of all three fixed factors remained significant. In addition, there was a significant interaction between ambient temperature and female reproductive status. For all six turtles, mean ODBA decreased throughout the nesting season (P < 0Æ0001) and was negatively correlated with the mean ambient temperature (P <0Æ037, Fig. 6). Significant variations in mean ODBA were found between the different reproductive stages (P <0Æ001, Fig. 6). For all six turtles, mean ODBA first regularly decreased from W5 to W2. Then, it slightly increased during W1 before decreasing again during IP1. Mean ODBA decreased by 28% between W5 and IP1. Mean ODBA during IP2 was significantly lower than during the first few weeks of the breeding season and 13% lower than during IP1, even though this difference was not significant. (a) (c) (b) Fig. 5. (a) Dive profiles of sequential U-dives and (b) variation in ODBA averaged over 10 s during this dive bout for a female loggerhead turtle during the 2010 breeding season at Zakynthos (Greece). (c) Relationship between mean ODBA and dive duration of 85 U-dives recorded by female loggerhead turtles during the breeding season at Zakynthos (Greece).

Accelerometry reveals turtle energy management strategy 7 (a) (b) during the breeding period resting at very shallow depths (2 m deep on average). Marine turtles may not be able to stay inactive for long periods in the same manner as terrestrial ectotherms, perhaps because of the relatively dynamic nature of the environment that they live in. However, activity levels were not constant throughout the season but varied on a seasonal basis. SEASONAL VARIATION IN ACTIVITY LEVELS Fig. 6. (a) Mean ODBA (black dots) and (b) mean ambient temperature (black dots) and mean proportion of U-dives (histograms) of six female loggerhead turtles during their different reproductive stages of the 2010 breeding season at Zakynthos (Greece). W5: 5 weeks prior to the first nesting event, W4: 4 weeks prior to first nesting event, W3: 3 weeks prior to first nesting event, W2: 2 weeks prior to first nesting event, W1: week of the first nesting event, IP1: first internesting interval, IP2: second internesting interval. Means of ODBA for the reproductive stages were calculated using a generalized linear model (see methods), and the differences among reproductive stages were statistically tested by pairwise comparisons of the estimated means. Different letters (a, b, c, d) indicate significant (P <0Æ05) differences between reproductive stages. Discussion DIVING AND ACTIVITY PATTERNS During the breeding season, the diving patterns of female turtles and their relatively wide range in activity level (i.e. variation in ODBA from 0Æ02 to 0Æ15 g) suggest that turtles were mostly active. In contrast, many terrestrial ectotherms have been reported to adopt sedentary and cryptic behaviours within selected warm sites to conserve energy while thermoregulating during the breeding period (e.g. Shine 2003a,b; Lourdais, Heulin & Denardo 2008). Foraging has been reported for this population during the nesting season but only as sporadic unsuccessful attempts (Schofield et al. 2007), and thus cannot explain the activity patterns observed in these turtles. Here, turtles performed several types of dives, which were associated with different levels of activity. The turtles were the least active when performing long (>10 min) flat-bottomed U-dives. The activity levels during bouts of U- dives were observed to progressively decrease, suggesting that turtles rest, or even sleep, during such prolonged bouts. This supports existing published literature (Hochscheid et al. 1999a; Hochscheid & Wilson 1999b; Minamikawa et al. 2000; Hays et al. 2004; Reina et al. 2005) and notably a previous study where the recording of buccal oscillations revealed that turtles may indeed enter a phase of sleep during long benthic U-dives (Houghton et al. 2008). Therefore, turtles at Zakynthos may spend on average only a third of their time Variation in activity levels throughout the season was impacted by ambient sea temperature, female reproductive status and female identity. The first few weeks of the breeding season corresponded to the highest mean activity level of the deployment period when the mean water temperature in the bay was relatively cold (i.e. 16 18 C, G. Schofield unpublished data). This result may at first seem surprising as cool temperatures are usually associated with low metabolic rates and activity levels in reptiles (e.g. Elsworth, Seebacher & Franklin 2003; Enstipp et al. 2011). The previous deployment of GPS data-loggers on female loggerhead turtles at the same site as this study has revealed that females may actively thermoregulate at the beginning of the nesting season (Schofield et al. 2009). They select down-wind shallow warm water patches and regularly reposition themselves according to the wind conditions within Laganas Bay, or following conflicts with other females (Schofield et al. 2009). Accordingly, our model suggests that female turtles should maintain body temperatures as warm as possible during the entire breeding season, and in particular for the first few weeks of the season to maximize reproductive output. Female leopard sharks Triakis semifasciata have similarly been observed aggregating in the warmest areas of an embayment (Hight & Lowe 2007). Here, the ambient temperatures experienced by the turtles were 2 3 C warmer than the actual mean water temperature in the bay, which will have increased their body temperature. At temperate breeding sites, female loggerhead turtles may thus trade off energy conservation at the beginning of the season for the maintenance of a high metabolism, which likely speeds up the process of egg maturation before oviposition, thus maximizing reproductive output (Weber et al. 2011). The benefit of exploiting these warm patches may, however, depend on how long the turtle had to search for and remain in each patch. These are important parameters that could be estimated in future by simultaneously deploying GPS data-loggers and accelerometers on female turtles. In addition, this high level of activity might also be linked with interactions (courtship avoidance) with males (Booth & Peters 1972; Schofield et al. 2007). As the season progressed, the mean percentage of time females spent resting increased, while mean ODBA decreased by 28% between the beginning of the season and the first internesting interval, and decreased again by 13% after the second nesting event. As water temperature increases, the turtles may gradually decrease the amount of time spent searching for warm patches and progressively reduce any

8 S. Fossette et al. unnecessary locomotory activity, as they are probably able to passively maintain warm body temperatures (see also Hays et al. 1999). A shift in energy partitioning from thermoregulatory activities to reproductive activities may therefore occur as the water temperature reaches a certain threshold and the turtles start nesting. Similar behavioural adjustments, and notably decreased locomotory activity, have been previously recorded in female birds, to compensate for the cost of producing eggs (Vézina, Speakman & Williams 2006) and or to conserve energy when fasting for several weeks during courtship and egg incubation (Halsey et al. 2008a; Green et al. 2009). This same process seen in birds may also occur in marine ectotherms, such as turtles, that have a long reproductive season and or a high reproductive investment. Activitylinked metabolic rate may decrease during the nesting season, as indicated by the observed decrease in ODBA values, while routine metabolic rate may concurrently increase with water temperature and reproductive costs. NEW MECHANISTIC MODEL The energetic costs of the active thermoregulatory behaviour probably used by the females at the beginning of the nesting season to access warm sea temperatures (i.e. 20 C) were integrated in a new mechanistic model (Fig. 2). This active behaviour was assumed to be associated with an increase in metabolic rate of 1Æ5 times that of inactive turtles during the pre-nesting period and 1Æ2 times during the first internesting interval (see Fig. 6). As the temperature warmed to about 24 C, turtles seemed to minimize their activity level (see Fig. 6); hence, any increase in metabolic rate beyond this temperature was assumed to be because of Q 10 effects only. This new model suggested that, despite this cost of activity, the strategy of staying active and warm still outperformed that of staying inactive and cool at the start of the season. An active turtle may thus lay the same number of clutches as an inactive turtle but in a shorter period of time and earlier in the season (Fig. 2). If the cost of activity is <1Æ5 times that of inactivity, then this benefit of the active and warm strategy improves. Laganas Bay, where the turtles resided, is fairly small, spanning only a few kilometre. Active reproductive turtles are not akin to migratory turtles that will be swimming many 10 s of kilometre per day. Rather active turtles in Laganas Bay are probably swimming <5 km per day (Schofield et al. 2010b) which equates to <0Æ1 body lengths per second. The MR of this level of swimming is likely to be far <1Æ5 times the MR of inactive turtles. For instance, green turtles swimming in an open flow respirometer at around 0Æ6 body lengths per second had a MR of about 1Æ5 times that of inactive turtles (Prange 1976). Hence, we are confident in the robustness of our conclusion that the active and warm strategy maximizes reproductive output over a season at temperate breeding sites. However, behavioural plasticity may be present in this species and in other species of sea turtles, depending on the environmental characteristics of the breeding site (e.g. Hatase, Omuta & Tsukamoto 2007; Rees et al. 2010; Schofield et al. 2010a; Hawkes et al. 2011). For instance, at warmer breeding sites, female turtles may spend the majority of their time resting to save energy and maximize the number of clutches, while at breeding sites where food is available, females may invest energy into foraging activities to supplement their body reserves and maximize reproduction (e.g. Hays et al. 2002b). Collecting empirical data on activity patterns of female sea turtles at various breeding sites is therefore important, and mechanistic models could be built to compare the different energy management strategies used across a range of environmental conditions. Conclusion While terrestrial and marine ectotherms both need to maintain warm body temperatures and minimize energy expenditure during the breeding season, female turtles breeding at the northern limit of their range may not be able to stay inactive for long periods perhaps because of the thermally dynamic nature of their environment. Hence, they adopt an active and stay warm strategy at the beginning of the season and then adjust their activity patterns in response to seasonal variations in abiotic (i.e. ambient temperature) and biotic (i.e. reproductive status) factors to maximize their reproductive output. Future studies should simultaneously investigate individual body condition, reproductive output and longterm activity patterns to further develop our understanding of individual reproductive and energy management strategies in marine ectotherms. Acknowledgements The authors thank the National Marine Park of Zakynthos (NMPZ) for permission to conduct this research. We are also very grateful to S. Vandenabeele for her assistance in the field. We thank NMPZ for boats and drivers for fieldwork and Archelon volunteers for logger retrieval assistance. We thank two anonymous reviewers for constructive comments on an earlier version of this manuscript. Financial and logistical support was provided by NERC, AXA Research Fund, NMPZ and the European Science Foundation THERM- ADAP program. GCH, SF and GS conceived the project and designed the study with contributions from ACG. GS, SF and MKSL performed the fieldwork; SF led the data analysis and wrote the manuscript with contributions from all authors. References Angilletta Jr, M.J., Winters, R.S. & Dunham, A.E. (2000) Thermal effects on the energetics of lizard embryos: implications for hatchling phenotypes. Ecology, 81, 2957 2968. Berger, D., Walters, R. & Gotthard, K. (2008) What limits insect fecundity? Body size- and temperature-dependent egg maturation and oviposition in a butterfly Functional Ecology, 22, 523 529. Bjorndal, K.A. & Carr, A. (1989) Variation in clutch size and egg size in the green turtle nesting population at Tortuguero, Costa Rica. Herpetologica, 45, 181 189. Bonnet, X., Bradshaw, D. & Shine, R. (1998) Capital versus income breeding: an ectothermic perspective. Oikos, 83, 333 342. Booth, J. & Peters, J.A. (1972) Behavioural studies on the green turtle (Chelonia mydas) in the sea. Animal Behaviour, 20, 808 812. Broderick, A.C., Glen, F., Godley, B.J. & Hays, G.C. (2002) Estimating the number of green and loggerhead turtles nesting annually in the Mediterranean. Oryx, 36, 227 235.

Accelerometry reveals turtle energy management strategy 9 Broderick, A.C., Glen, F., Godley, B.J. & Hays, G.C. (2003) Variation in reproductive output of marine turtles. Journal of Experimental Marine Biology and Ecology, 288, 95 109. Broderick, A.C., Coyne, M.S., Fuller, W.J., Glen, F. & Godley, B.J. (2007) Fidelity and over-wintering of sea turtles. Proceedings of the Royal Society B: Biological Sciences, 274, 1533. Dubois, Y., Blouin Demers, G., Shipley, B. & Thomas, D. (2009) Thermoregulation and habitat selection in wood turtles Glyptemys insculpta: chasing the sun slowly. Journal of Animal Ecology, 78, 1023 1032. Elsworth, P.G., Seebacher, F. & Franklin, C.E. (2003) Sustained swimming performance in crocodiles (Crocodylus porosus): effects of body size and temperature. Journal of Herpetology, 37, 363 368. Enstipp, M.R., Ciccione, S., Gineste, B., Milbergue, M., Ballorain, K., Ropert- Coudert, Y., Kato, A., Plot, V. & Georges, J.Y. (2011) Energy expenditure of freely swimming adult green turtles (Chelonia mydas) and its link with body acceleration. The Journal of experimental biology, 214, 4010 4020. Fossette, S., Gaspar, P., Handrich, Y., Maho, Y.L. & Georges, J.Y. (2008) Dive and beak movement patterns in leatherback turtles Dermochelys coriacea during internesting intervals in French Guiana. Journal of Animal Ecology, 77, 236 246. Fossette, S., Gleiss, A.C., Myers, A.E., Garner, S., Liebsch, N., Whitney, N.M., Hays, G.C., Wilson, R.P. & Lutcavage, M.E. (2010) Behaviour and buoyancy regulation in the deepest-diving reptile: the leatherback turtle. Journal of Experimental Biology, 213, 4074 4083. Fuller, W.J., Broderick, A.C., Hooker, S.K., Witt, M.J. & Godley, B.J. (2009) Insights into habitat utilization by green turtles (Chelonia mydas) during the inter-nesting period using animal-borne digital cameras. Marine Technology Society Journal, 43, 51 59. Gleiss, A.C., Norman, B. & Wilson, R.P. (2011b) Moved by that sinking feeling: variable diving geometry underlies movement strategies in whale sharks. Functional Ecology, 25, 595 607. Gleiss, A.C., Dale, J.J., Holland, K.N. & Wilson, R.P. (2010) Accelerating estimates of activity-specific metabolic rate in fishes: testing the applicability of acceleration data-loggers. Journal of Experimental Marine Biology and Ecology, 385, 85 91. Gleiss, A.C., Jorgensen, S.J., Liebsch, N., Sala, J.E., Norman, B., Hays, G.C., Quintana, F., Grundy, E., Campagna, C. & Trites, A.W. (2011a) Convergent evolution in locomotory patterns of flying and swimming animals. Nature Communications, 2, 352 358. Godley, B.J., Broderick, A.C., Frauenstein, R., Glen, F. & Hays, G.C. (2002) Reproductive seasonality and sexual dimorphism in green turtles. Marine Ecology Progress Series, 226, 125 133. Green, J.A., Boyd, I.L., Woakes, A.J., Warren, N.L. & Butler, P.J. (2009) Evaluating the prudence of parents: daily energy expenditure throughout the annual cycle of a free-ranging bird, the macaroni penguin Eudyptes chrysolophus. Journal of Avian Biology, 40, 529 538. Halsey, L.G., Butler, P.J., Fahlman, A., Woakes, A.J. & Handrich, Y. (2008a) Behavioral and physiological significance of minimum resting metabolic rate in king penguins. Physiological and Biochemical Zoology, 81, 74 86. Halsey, L.G., Green, J.A., Wilson, R.P. & Frappell, P.B. (2008b) Accelerometry to estimate energy expenditure during activity: best practice with data loggers. Physiological and Biochemical Zoology, 82, 396 404. Hamann, M., Limpus, C.J. & Read, M.A. (2007) Vulnerability of marine reptiles in the Great Barrier Reef to climate change. Climate Change and the Great Barrier Reef: A Vulnerability Assessment (eds J.E. Johnson & P. Marshall). pp. 465 496, Great Barrier Reef Marine Park Authority and Australian Greenhouse Office, Australia. Hatase, H., Omuta, K. & Tsukamoto, K. (2007) Bottom or midwater: alternative foraging behaviours in adult female loggerhead sea turtles. Journal of Zoology, 273, 46 55. Hawkes, L.A., Witt, M.J., Broderick, A.C., Coker, J.W., Coyne, M.S., Dodd, M., Frick, M.G., Godfrey, M.H., Griffin, D.B.B. & Murphy, S.R. (2011) Home on the range: spatial ecology of loggerhead turtles in Atlantic waters of the USA. Diversity and Distributions, 17, 624 640. Hays, G.C. & Speakman, J.R. (1991) Reproductive investment and optimum clutch size of loggerhead sea turtles (Caretta caretta). Journal of Animal Ecology, 60, 455 462. Hays, G.C., Luschi, P., Papi, F., Del Seppia, C. & Marsh, R. (1999) Changes in behaviour during the inter-nesting period and post-nesting migration for Ascension Island green turtles. Marine Ecology Progress Series, 189, 263 273. Hays, G.C., Broderick, A.C., Glen, F., Godley, B.J., Houghton, J.D.R. & Metcalfe, J.D. (2002a) Water temperature and internesting intervals for loggerhead (Caretta caretta) and green (Chelonia mydas) sea turtles. Journal of Thermal Biology, 27, 429 432. Hays, G.C., Glen, F., Broderick, A.C., Godley, B.J. & Metcalfe, J.D. (2002b) Behavioural plasticity in a large marine herbivore: contrasting patterns of depth utilisation between two green turtle (Chelonia mydas) populations. Marine Biology, 141, 985 990. Hays, G.C., Metcalfe, J.D., Walne, A.W. & Wilson, R.P. (2004) First records of flipper beat frequency during sea turtle diving. Journal of Experimental Marine Biology and Ecology, 303, 243 260. Herrel, A., James, R.S. & Van Damme, R. (2007) Fight versus flight: physiological basis for temperature-dependent behavioral shifts in lizards. Journal of Experimental Biology, 210, 1762 1767. Hight, B.V. & Lowe, C.G. (2007) Elevated body temperatures of adult female leopard sharks, Triakis semifasciata, while aggregating in shallow nearshore embayments: evidence for behavioral thermoregulation? Journal of Experimental Marine Biology and Ecology, 352, 114 128. Hochscheid, S., Bentivegna, F. & Speakman, J.R. (2004) Long-term cold acclimation leads to high Q10 effects on oxygen consumption of loggerhead sea turtles Caretta caretta. Physiological and Biochemical Zoology, 77, 209 222. Hochscheid, S. & Wilson, R.P. (1999b) A new method for the determination of at-sea activity in sea turtles. Marine Ecology Progress Series, 185, 293 296. Hochscheid, S., Godley, B.J., Broderick, A.C. & Wilson, R.P. (1999a) Reptilian diving: highly variable dive patterns in the green turtle Chelonia mydas. Marine Ecology Progress Series, 185, 101 112. Houghton, J.D.R., Broderick, A.C., Godley, B.J., Metcalfe, J.D. & Hays, G.C. (2002) Diving behaviour during the internesting interval for loggerhead turtles Caretta caretta nesting in Cyprus. Marine Ecology Progress Series, 227, 63 70. Houghton, J.D.R., Cedras, A., Myers, A.E., Liebsch, N., Metcalfe, J.D., Mortimer, J.A. & Hays, G.C. (2008) Measuring the state of consciousness in a free-living diving sea turtle. Journal of Experimental Marine Biology and Ecology, 356, 115 120. Huey, R.B. & Berrigan, D. (2001) Temperature, demography, and ectotherm fitness. American Naturalist, 158, 204 210. Lourdais, O., Heulin, B. & Denardo, D.F. (2008) Thermoregulation during gravidity in the children s python (Antaresia childreni): a test of the preadaptation hypothesis for maternal thermophily in snakes. Biological Journal of the Linnean Society, 93, 499 508. Lourdais, O., Bonnet, X., Shine, R., Denardo, D.F., Naulleau, G. & Guillon, M. (2002) Capital-breeding and reproductive effort in a variable environment: a longitudinal study of a viviparous snake. Journal of Animal Ecology, 71, 470 479. Margaritoulis, D. (2005) Nesting activity and reproductive output of loggerhead sea turtles, Caretta caretta, over 19 seasons (1984 2002) at Laganas Bay, Zakynthos, Greece: the largest rookery in the Mediterranean. Chelonian Conservation and Biology, 4, 916 929. Martin, T.L. & Huey, R.B. (2008) Why suboptimal is optimal: Jensen s inequality and ectotherm thermal preferences. The American Naturalist, 171, E102 E118. Minamikawa, S., Naito, Y. & Uchida, I. (1997) Buoyancy control in diving behavior of the loggerhead turtle, Caretta caretta. Journal of Ethology, 15, 109 118. Minamikawa, S., Naito, Y., Sato, K., Matsuzawa, Y., Bando, T. & Sakamoto, W. (2000) Maintenance of neutral buoyancy by depth selection in the loggerhead turtle Caretta caretta. Journal of Experimental Biology, 203, 2967 2975. Prange, H.D. (1976) Energetics of swimming of a sea turtle. Journal of Experimental Biology, 64, 1 12. Rees, A.L.F., Saady, S.A., Broderick, A.C., Coyne, M.S., Papathanasopoulou, N. & Godley, B.J. (2010) Behavioural polymorphism in one of the world s largest populations of loggerhead sea turtles Caretta caretta. Marine Ecology Progress Series, 418, 201 212. Reina, R.D., Abernathy, K.J., Marshall, G.J. & Spotila, J.R. (2005) Respiratory frequency, dive behaviour and social interactions of leatherback turtles, Dermochelys coriacea during the inter-nesting interval. Journal of Experimental Marine Biology and Ecology, 316, 1 16. Sato, K., Matsuzawa, Y., Tanaka, H., Bando, T., Minamikawa, S., Sakamoto, W. & Naito, Y. (1998) Internesting intervals for loggerhead turtles, Caretta caretta, and green turtles, Chelonia mydas, are affected by temperature. Canadian Journal of Zoology, 76, 1651 1662. Schofield, G., Katselidis, K.A., Dimopoulos, P., Pantis, J.D. & Hays, G.C. (2007) Behaviour analysis of the loggerhead sea turtle Caretta caretta from direct in-water observation. Endangered Species Research, 3, 71 79. Schofield, G., Bishop, C.M., Katselidis, K.A., Dimopoulos, P., Pantis, J.D. & Hays, G.C. (2009) Microhabitat selection by sea turtles in a dynamic thermal marine environment. Journal of Animal Ecology, 78, 14 21.

10 S. Fossette et al. Schofield, G., Hobson, V.J., Fossette, S., Lilley, M.K.S., Katselidis, K.A. & Hays, G.C. (2010a) Fidelity to foraging sites, consistency of migration routes and habitat modulation of home range by sea turtles. Diversity and Distributions, 16, 840 853. Schofield, G., Hobson, V.J., Lilley, M.K.S., Katselidis, K.A., Bishop, C.M., Brown, P. & Hays, G.C. (2010b) Inter-annual variability in the home range of breeding turtles: implications for current and future conservation management. Biological Conservation, 143, 722 730. Seebacher, F., Franklin, C.E. & Read, M. (2005) Diving behaviour of a reptile (Crocodylus johnstoni) in the wild: interactions with heart rate and body temperature. Physiological and Biochemical Zoology, 78, 1 8. Seminoff, J.A., Jones, T.T. & Marshall, G.J. (2006) Underwater behaviour of green turtles monitored with video-time-depth recorders: what s missing from dive profiles? Marine Ecology Progress Series, 322, 269 280. Shepard, E.L.C., Wilson, R.P., Halsey, L.G., Quintana, F., Laich, A.G., Gleiss, A.C., Liebsch, N., Myers, A.E. & Norman, B. (2008) Derivation of body motion via appropriate smoothing of acceleration data. Aquatic Biology, 4, 235 241. Shepard, E.L.C., Wilson, R.P., Quintana, F., Laich, A.G. & Forman, D.W. (2009) Pushed for time or saving on fuel: fine-scale energy budgets shed light on currencies in a diving bird. Proceedings of the Royal Society B: Biological Sciences, 276, 3149 3155. Shine, R. (2003a) Locomotor speeds of gravid lizards: placing costs of reproduction within an ecological context. Functional Ecology, 17, 526 533. Shine, R. (2003b) Reproductive strategies in snakes. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270, 995 1004. Shine, R. (2004) Incubation regimes of cold-climate reptiles: the thermal consequences of nest-site choice, viviparity and maternal basking. Biological Journal of the Linnean Society, 83, 145 155. Vézina, F., Speakman, J.R. & Williams, T.D. (2006) Individually variable energy management strategies in relation to energetic costs of egg production. Ecology, 87, 2447 2458. Wallman, H.L. & Bennett, W.A. (2006) Effects of parturition and feeding on thermal preference of Atlantic stingray, Dasyatis sabina (Lesueur). Environmental Biology of Fishes, 75, 259 267. Weber, S.B., Blount, J.D., Godley, B.J., Witt, M.J. & Broderick, A.C. (2011) Rate of egg maturation in marine turtles exhibits universal temperature dependence. The Journal of Animal Ecology, 80, 1034 1041. Wikelski, M. & Trillmich, F. (1994) Foraging strategies of the Gala pagos marine iguana (Amblyrhynchus cristatus): adapting behavioral rules to ontogenetic size change. Behaviour, 128, 255 279. Williams, T.D., Vezina, F. & Speakman, J.R. (2009) Individually variable energy management during egg production is repeatable across breeding attempts. Journal of Experimental Biology, 212, 1101 1105. Wilson, R.P., Shepard, E.L.C. & Liebsch, N. (2008) Prying into the intimate details of animal lives: use of a daily diary on animals. Endangered Species Research, 4, 123 137. Wilson, R.P., White, C.R., Quintana, F., Halsey, L.G., Liebsch, N., Martin, G.R. & Butler, P.J. (2006) Moving towards acceleration for estimates of activity-specific metabolic rate in free-living animals: the case of the cormorant. Journal of Animal Ecology, 75, 1081 1090. Received 2 May 2011; accepted 16 December 2011 Handling Editor: Dan Costa