LOCOMOTOR BEHAVIOR OF NOCTURNAL GHOST CRABS ON THE BEACH: FOCAL ANIMAL SAMPLING AND INSTANTANEOUS VELOCITY FROM THREE-DIMENSIONAL MOTION ANALYSIS

Transcription

1 The Journal of Experimental Biology 198, (1995) Printed in Great Britain The Company of Biologists Limited LOCOMOTOR BEHAVIOR OF NOCTURNAL GHOST CRABS ON THE BEACH: FOCAL ANIMAL SAMPLING AND INSTANTANEOUS VELOCITY FROM THREE-DIMENSIONAL MOTION ANALYSIS RANDI B. WEINSTEIN* Department of Integrative Biology, University of California, Berkeley, CA 9472, USA Accepted 12 December 1994 Previous laboratory measurements of the energetics and biomechanics of locomotion have defined performance limits for the ghost crab Ocypode quadrata. To discover whether these animals naturally operate within these limits, a novel infrared videotaping system was used to observe nocturnally active ghost crabs in the field for threedimensional motion analysis (N=27) and focal animal sampling (N=24). Instantaneous movement velocity, movement duration, pause duration and stride frequency were determined from video tapes. Voluntarily active crabs moved at a mean instantaneous velocity of 8 cm s 1. Stressed crabs (i.e. those captured and released into the field site) moved at a mean velocity of 83 cm s 1. The mean movement and pause period durations of voluntarily active animals moving along the beach were 11.2 and 23.4 s, respectively. Stressed crabs had much shorter movement (1.4 s) and pause (7.6 s) durations. Despite the differences in mean movement and pause duration, both voluntarily Summary active and stressed crabs moved for an average of approximately 3 % of the observation period. These data indicate that voluntarily active ghost crabs primarily move at velocities that can be sustained aerobically and that their performance is not likely to be altered by moving intermittently. By contrast, stressed crabs move at faster speeds that are closer to the limits of their continuous locomotor performance (e.g. escape behavior and aggressive encounters). In the laboratory, the endurance capacity of crabs moving continuously at these rapid speeds is only a few seconds. However, in the field, the stressed crabs are able to move intermittently for more than a few seconds, yet they do not fatigue. These observations suggest that the performance limits of the stressed crabs are increased by moving intermittently. Key words: movement, Ocypode quadrata, ghost crab, locomotion, motion analysis, speed, stride frequency. Introduction The limits of locomotor performance can be identified by controlled laboratory studies. The upper limits of a terrestrial animal s locomotor performance can be described by parameters that include its maximal rate of oxygen consumption, cost of locomotion, maximum aerobic speed (MAS; the minimum speed that elicits the maximal rate of oxygen consumption; John-Alder and Bennett, 1981) and endurance capacity. Comparative studies indicate that these parameters are strongly influenced by body size (Full, 1991; Taylor et al. 197), body temperature (Full and Tullis, 199; John-Alder and Bennett, 1981; John-Alder et al. 1983; Weinstein and Full, 1994), species (Full et al. 1988) and locomotor behavior (Weinstein and Full, 1992). In the present study, I test specific predictions about natural locomotion on the basis of laboratory studies of physiology, biomechanics and behavior on a semi-terrestrial crustacean, the ghost crab Ocypode quadrata. On the basis of previous laboratory studies, I predicted that ghost crabs moving continuously over long distances would move at speeds near or below the MAS to maximize endurance capacity and to maximize the mechanical energy exchange between potential and kinetic energy of the center of mass (i.e. in a manner analogous to an inverted pendulum; see review by Full and Weinstein, 1992). In contrast, I predicted that stressed ghost crabs (e.g. those escaping or engaging in aggressive encounters) would move at speeds faster than the MAS, speeds that can only be sustained for seconds if used continuously (Full, 1987). However, if stressed crabs move intermittently in the field, rather than moving continuously, their performance limits could be altered. Locomotor behavior has been shown to affect performance limits in the laboratory. Weinstein and Full (1992) demonstrated that frequent dynamic adjustments can alter distance capacity (i.e. the total distance traveled before fatigue). When ghost crabs move intermittently, by alternating brief periods of movement with brief pauses, they can travel *Present address: Department of Environmental, Population and Organismic Biology, University of Colorado, Boulder, CO , USA.

2 99 R. B. WEINSTEIN as much as 2 5 times further before they fatigue than if they move continuously at the same average speed (Weinstein and Full, 1992; R. J. Weinstein and R. J. Full, in preparation). Alternatively, using different exercise and pause intervals, the distance capacity for intermittent locomotion can be reduced to one-tenth of that of continuous locomotion at the same average speed (Weinstein and Full, 1992). Therefore, intermittent locomotor performance limits are dependent upon three key variables: (1) movement velocity, (2) movement duration and (3) pause duration. Furthermore, the movement and pause intervals that alter performance limits are temperature-sensitive (R. B. Weinstein and R. J. Full, in preparation). Laboratory studies of intermittent locomotor performance limits provide the basis for more rigorous predictions about natural locomotion. Specifically, ghost crabs moving at slow speeds with alternating brief movements and pauses should have similar performance limits to those of crabs moving continuously at the same average speed. In contrast, fastermoving stressed crabs should sustain locomotion for more than a few seconds if they move intermittently. In addition, the movement durations of faster-moving stressed crabs should be shorter than those of crabs moving at slow speeds in order to minimize the metabolic disturbances associated with movement at rapid speeds. Whereas the study of continuous and intermittent locomotor performance in the laboratory defines the upper operating limits for system function (e.g. fuel and oxygen transport), discovering where animals naturally operate within their performance limits will identify the particular circumstances under which the systems currently function. Studies of natural terrestrial locomotor behavior presented in the literature examine foraging behavior, home ranges, migration distances and activity patterns (e.g. Huey and Pianka, 1981; Kenagy and Hoyt, 1989; Pietruszka, 1986; Reynolds and Riley, 1988; Wallin, 1991; Yabe, 1992). The methods typically used to measure locomotor behavior range from estimating distances by eye and using a stopwatch to record elapsed time to sophisticated telemetry techniques. Measurements reported include average speed (i.e. over a period of minutes to days), average speed of movement (i.e. while moving), movement frequency (i.e. number of moves per unit time), pause duration and the proportion of time spent moving (see McLaughlin, 1989; McLaughlin et al. 1992). While these studies have provided significant advances in animal behavior, physiological ecology and conservation biology, they are generally too limited in spatial and temporal resolution to test specific physiological and biomechanical predictions. Previous attempts (Adamczewska and Morris, 1994; Bennett et al. 1984; Garland, 1985; Hertz et al. 1988) to integrate laboratory and field studies of animal locomotion have proposed that many activities, such as long-distance foraging and searching for mates, are primarily supported by aerobic metabolism. These field studies suggest that terrestrial animals move at speeds lower than or equal to their MAS. However, none of these studies has measured the instantaneous velocity of unrestrained animals, nor have they taken into account intermittent locomotor behavior. In the present study, I designed a novel infrared videotaping system to make quantitative measurements of the nocturnally active ghost crab s natural locomotor behavior in the field. I tested specific predictions based on performance limits defined in the laboratory. To accomplish this, I measured the instantaneous velocity of ghost crabs moving freely in their natural habitat by three-dimensional motion analysis. In addition, I measured movement duration, pause duration and stride frequency of focal animals to provide quantitative descriptions of locomotor behavior. Individuals observed in the present study were categorized as being either voluntarily active or stressed. Voluntarily active crabs moved along the beach with minimal observer interference. In contrast, stressed crabs were captured and released and displayed escape behavior. By investigating both voluntarily active and stressed crabs, I could examine the range of locomotor behaviors exhibited by ghost crabs moving freely in their natural habitat. Materials and methods Field site All field data were collected during the summer (June August) at Fort Macon State Park, North Carolina. The park is located on the eastern end of the Atlantic Beach barrier island. The study site was a sandy beach on the ocean side of the island. The park is closed to visitors between sunset and sunrise. At the study site, ghost crabs are primarily active above ground from late May until October and spend the winter months underground (Wolcott, 1978). The ghost crab population is active at night, emerging from their burrows after dusk and moving and foraging actively until dawn. Some individuals migrate as far as 3 m along the beach in one night (Wolcott, 1978). Ghost crabs are rarely active at ambient temperatures below 15 2 C, even though their body temperatures are likely to be well above their lower lethal limit (6 8 C; T. G. Wolcott, personal communication). In the present study, all field measurements were conducted at night during the first 2 h after it became dark. At this time, the ghost crabs had emerged from their burrows and were slowly foraging along the beach, primarily in search of mole crabs, Emerita talpoida, and coquina clams, Donax variabilis (Wolcott, 1978). I also observed ghost crabs sprinting towards and away from other ghost crabs in their paths and away from the incoming surf, although these events were somewhat infrequent. While many ghost crabs were seen moving voluntarily along the beach, I also wanted to document the locomotor behavior of stressed (i.e. sprinting) crabs. To observe the behavior of a greater number of stressed crabs, I captured individuals near the field site. Crabs were initially spotted with a flashlight (Mag-Lite). Individuals that remained motionless in the beam of the flashlight were covered with a long-handled net. Crabs that struggled during capture were not used for videotaped observations. Once captured under the net,

3 Integrating field and laboratory studies of locomotion 991 individuals were quickly transferred to a small plastic container and carried to the field site. Within 2 5 min of being spotted, the stressed crabs were released and observed. N Burrows and dunes Field videotaping procedures Ghost crabs were videotaped in the field by two methods. First, the instantaneous velocity of individual ghost crabs was determined by three-dimensional motion analysis. Second, the movement duration, pause duration and stride frequency were determined by focal animal sampling. Videotaping method for three-dimensional motion analysis I designed a novel infrared video system to capture the movements of individual ghost crabs moving along the beach for three-dimensional motion analysis. Images were captured by two stationary black-and-white video cameras (WATEC, WAT-92) with wide-angle lenses (Tamaron, 6 12 mm). Images were recorded in SuperVHS format at a framing rate of 3 frames s 1 (6 fields s 1 ; Panasonic, PVS-77) and observed on portable video monitors (Sony, Watchman FD- 25). The video cameras were mounted approximately 2 m above the beach surface on tripods (Bogen, 336). The cameras were positioned on the beach to maximize the probability of ghost crabs voluntarily moving within the field site. The cameras were pointed towards the ocean (away from the dunes) and at a distance of several meters above the surf zone, taking the tidal movements into consideration (Fig. 1). The angle between the cameras and the field site was approximately 9. Before sunset, the field site was calibrated with a 123 cm 123 cm 3 cm structure made out of.5 inch (1.3 cm) polyvinylchloride pipe that approximately filled the overlapping views of both video cameras. This structure provided six non-colinear points for video analysis. Once the calibration structure had been recorded on video tape, it was removed from the field site. After it became dark, the video cameras were synchronized to one another using an infrared light-emitting diode (LED, Radio Shack, ). Two infrared LEDs were used as reference points within the field site. The field site and the ghost crabs were illuminated by 2 cd spot/floodlights (Brinkman, Q-Beam) with infrared filters (American Optical Co., filter ). Prior to the field observations, I determined that the ghost crabs could not see the light emitted by the infrared spotlights by placing the ghost crabs in an optokinetic nystagmus drum with white or infrared light. The ghost crabs displayed a typical visual response in fluorescent and incandescent lighting, but no visual response in infrared light. Using infrared illumination, it was not necessary to mark individual crabs in the field. Therefore, all individuals moving within the field site could be readily observed and recorded. I videotaped all medium-sized (approximately 3 g) ghost crabs that moved voluntarily through the field site (N=12). Stressed (i.e. captured and released; N=15) crabs were released at the edge of the overlapping views for both video cameras and followed until they moved out of the field site. Camera 1 Camera 2 Foraging zone Three-dimensional calibration structure Ocean Fig. 1. Diagram of the arrangement for the three-dimensional videotaping method. The dimensions of the calibration structure were 123 cm 123 cm 3 cm and it approximately filled the overlapping views of both cameras. See text for additional details. Videotaping method for focal animal sampling A similar infrared video system was used in combination with focal animal sampling techniques to record the movements of individual crabs moving along the beach. Images were captured with a zoom lens (Tokina, mm). Crabs illuminated with the infrared spotlight were readily observable and therefore it was not necessary to mark animals prior to videotaping. The resolution of the videotaped image was high enough to see the movements of the ghost crabs individual legs. Ghost crabs moved freely along the beach and were followed from as great a distance as possible while maintaining clarity of the video image (N=21). Observer movement was kept to a minimum. The basic protocol for focal animal sampling was to select at random a medium-sized crab and then to observe that individual for as long a period as possible. The predetermined minimum observation time for analysis was 5 min for the voluntarily active animals. Individuals were observed until I lost sight of them, until the crab had traveled too far away from the observer for the individual legs to be viewed clearly, or until the crab remained motionless for more than 5 min. Additional medium-sized ghost crabs were captured, released and videotaped. These stressed crabs (N=3) generally moved at rapid speeds and in unpredictable directions after being released and were difficult to follow in the darkness. Therefore, the stressed crabs were observed for as long as possible and there was no minimum observation time. Body size and temperature In the laboratory, individuals (N=16) were weighed to the nearest.1 g and carapace width was measured to the nearest 1 mm. The relationship between carapace width and body mass was used to estimate the body mass of crabs in the field. Ambient and body temperatures were measured in the field

4 992 R. B. WEINSTEIN with a thermocouple (copper constantan, Teflon leads,.2 mm diameter). The thermocouple leads were connected to a microprocessor thermometer (model HH23, type K connector, Omega) and calibrated from to 4 C. Ambient temperatures were measured approximately.5 m above the ground with dry thermocouple leads. Relative humidity was measured with a sling psychrometer (Bacharach, ) at a distance of approximately.5 m above the substratum. Body temperatures of active ghost crabs were measured with a thermocouple within 2 min of capture. Immediately after capture, a small hole (1 mm) was made in the surface of the ventral carapace with a dissecting needle. The thermocouple tip was threaded through the opening to a depth of approximately 5 mm within the levator muscle adjacent to the third walking leg, and a stable body temperature was recorded. Following measurement of body temperature and then size (i.e. carapace width), crabs were released. Body temperature and size measurements were not made on the same ghost crabs that were videotaped in order to minimize disturbance to the animals prior to videotaping and to allow for focal individuals that I was unable to relocate following videotaping. While body temperature measurements were not made on the specific individuals that were videotaped, an effort was made to make these measurements at the same approximate time and location when the observations were made on the videotaped individuals. An effort was also made to make body size measurements on individuals that were similar in size to those that were videotaped. Body temperatures were measured for at least five ghost crabs during each videotaping session and generally varied by less than.5 C for a given session. Body temperatures for all crabs videotaped on a single night were assumed to be the mean of the five individuals actually measured. Body size measurements were made to assess my ability to estimate body mass by eye in the field (i.e. to select medium-sized, approximately 3 g, ghost crabs as focal animals). Video analysis Three-dimensional motion analysis Videotapes were analyzed using three-dimensional motion analysis hardware and software (PEAK Performance Technologies, 3D Package) at 3 fields s 1 (i.e. every other field). Instantaneous velocity was calculated by direct linear transformation (Biewener and Full, 1992). The average mean squared error in measurement of position was.3 cm. All medium-sized individuals moving within the field site were analyzed and portions of the videotaped record were analyzed only when the crabs were moving. Focal animal sampling Videotapes of focal animals were analyzed field-by-field (at 3 fields s 1 ; i.e. every other field) to determine movement duration and pause duration. Animals were considered to be moving if their walking legs were in motion. The total number of strides during each movement was counted. A stride frequency of 6 strides s 1 (the maximum stride frequency for medium-sized ghost crabs; Blickhan and Full, 1987) would result in 5 frames stride 1. If a crab remained motionless for at least 15 consecutive frames (.5 s), it was considered to be pausing. The duration of each movement, the average stride frequency for each movement period (total number of strides divided by the movement duration) and the duration of each pause were determined for each focal animal. Results Body size and temperature Body mass was correlated with carapace width (Fig. 2A). The mean body mass of crabs videotaped in field was 34.6±12. g (S.D.; median 33.1 g; range g; Fig. 2B). Most of these individuals (66 %) were medium-sized (2 4 g). Body temperatures were generally slightly cooler than ambient temperature (Fig. 3A). When plotted as a function of the ambient relative humidity, the depression in body Body mass (g) Number of individuals A Carapace width (mm) B Mass (g) Fig. 2. (A) Carapace width plotted as a function of body mass: log(mass) = 3.31log(carapace width) (N=16, r 2 =.97, P<.1). (B) Frequency distribution of body size for observed ghost crabs [N=76, mean 34.6±12. g (S.D.), median 33.1 g]. 5

5 Integrating field and laboratory studies of locomotion 993 temperature increased as the relative humidity decreased (Fig. 3B). The magnitude of the difference between body and ambient temperature was greatest (approximately 2 C) at the lowest relative humidity (approximately 63 %) and was smallest (<.5 C) at the highest relative humidity (approximately 96 %). However, at an ambient temperature of 2 C, body and ambient temperatures were similar, despite the low relative humidity (Fig. 3A,B). On the night when the ambient temperature was 2 C, the air temperature 5cm below the mouth of the burrow was 22.3±.5 C (N=4). Mean body temperatures of the videotaped animals ranged from approximately 23 to 27 C. Mean body temperature of the crabs videotaped for the determination of instantaneous velocity was 25. C (range C; see Table 1). The mean body temperatures for the voluntarily active and stressed focal animals were 24.8 and 25.2 C, respectively (range C; see Table 2). Instantaneous velocity Twelve ghost crabs moved voluntarily through the field site. A second set of fifteen individuals was captured and released T a T b ( C) T b ( C) T a ( C) Relative humidity (%) A 28 1 Fig. 3. (A) Body temperature (T b) in the field as a function of ambient temperature (T a; N=82). The line represents T b=t a. (B) Body temperature depression (T a T b) as a function of relative humidity: T a T b= (relative humidity) (N=82, r 2 =.17, P<.1). B into the site. In general, the voluntarily active crabs entered the field site, moved short distances, paused once or twice, and often changed direction. In contrast, the captured crabs, once released, generally sprinted rapidly out of the field site with little change in direction. The instantaneous velocity for each voluntarily active individual increased and decreased as a function of time, indicating that the crabs accelerate and decelerate frequently rather than moving at a constant speed (Fig. 4A,B,C). In addition, the ghost crabs did not move in straight lines and sometimes moved in circuitous routes (Fig. 4D). The total observation time for each voluntarily active crab was brief (<9 s), limited primarily by the size of the overlapping viewing areas for the video cameras (approximately 123 cm 123 cm 3 cm, the size of the calibration structure). The mean instantaneous velocity for all of the voluntarily active crabs was 8.3 cm s 1 (range cm s 1 ; Table 1). The maximum instantaneous velocity for voluntarily active crabs ranged from 22.6 to cm s 1. Stressed crabs moved at significantly faster mean and maximum (Mann Whitney U-test, P<.1) velocities than voluntarily active crabs. The mean instantaneous velocity of the stressed crabs ranged from 18.5 to 176.4cm s 1. The mean instantaneous velocity for all of the stressed crabs was 82.9 cm s 1, which was an order of magnitude faster than that for the voluntarily active animals (Table 1). The maximum instantaneous velocity for stressed individuals ranged from 52.9 to cm s 1. The shorter observation time for captured and released crabs (3 s), compared with that for the voluntarily active crabs (36 s), can be explained by a combination of the fixed size of the field site and the faster instantaneous velocities of each individual. Movement duration, stride frequency and pause duration Twenty-one voluntarily active focal animals were observed for at least 5 min, the pre-determined minimum observation period. Observations on these individuals were terminated because (1) I lost sight of the crab (52 % of the trials), (2) the crab had traveled too far away from the observer for the individual legs to be viewed clearly (34 % of the trials), or (3) the crab remained motionless for more than 5 min (14 % of the trials). When the focal animals were first observed, 48% of the individuals (1 out of 21) were facing away from the observer and 52 % (11 out of 21) were facing towards the observer. In general, the crabs continued in the same initial direction as they moved along the beach. Each voluntarily active individual alternated brief movements with brief pauses throughout the entire observation period, rather than moving continuously (Fig. 5A). The mean movement duration was 11.2s (range s) and the mean pause duration was about twice as long (23.4 s; range s; Table 2). The median movement duration was only 8.5 s, indicating that most of the movement periods were brief (i.e. shorter than the mean), while a small number were significantly longer than the mean movement duration. The longest movement duration was 74 s. Similarly, the median

6 994 R. B. WEINSTEIN Instantaneous velocity (cm s 1 ) A MAS Time (s) B MAS Time (s) C MAS Time (s) D E F 2 cm Fig. 4. (A C) Examples of instantaneous velocity versus time for individual ghost crabs. (A) Crab walking around its burrow. (B) Crab moving within the field site. (C) Crab sprinting within the field site. The dashed line represents the maximum aerobic speed (MAS) determined in the laboratory (2 cm s 1 ; Full, 1987). (D F) The path of each individual as if viewed from above. 2 cm pause period (i.e. 6.7 s) was shorter than the mean pause duration, indicating a predominance of short pause periods. The duration of each movement was not directly correlated with the subsequent pause duration in 19 out of 21 voluntarily active individuals. However, while the ratio of the mean movement duration to mean pause duration varied considerably for each individual (Fig. 6), most of the voluntarily active focal crabs were moving for between 1 and 5 % of the observation period (81 % of the trials; Fig. 6). The movement velocity of the focal animals was estimated from the mean stride frequency for each movement period. For medium-sized ghost crabs, stride frequency is directly related to velocity at stride frequencies less than 6 strides s 1 (approximately 9 cm s 1 ; Blickhan and Full, 1987). At faster speeds, stride length increases while stride frequency remains constant. The walk trot transition for a ghost crab occurs at a stride frequency of approximately 3.7 strides s 1 (approximately 4 cm s 1 ; Blickhan and Full, 1987). When traveling at the fastest mean stride frequencies (and therefore the fastest velocities), the movements were of short duration (<5 s; Fig. 7). In contrast, the longest movement durations (>15 s) were made at low mean stride frequencies (<3 strides s 1 ; Fig. 7). A stride frequency of 3 strides s 1 is equivalent to a speed of approximately 27 cm s 1 (Blickhan and Full, 1987). The mean stride frequency for the voluntarily active crabs was 2.1 strides s 1 (range strides s 1 ). Three stressed crabs were observed by focal sampling. These captured and released individuals were difficult to follow since they moved quickly in unpredictable directions. The average observation period for the stressed animals was 1.4 min (range min; Table 2). The stressed animals made movements that were of shorter duration and faster velocity, and with shorter pauses, than the voluntarily active animals (Figs 5B, 6; Table 2). The mean movement duration was 1.4 s (range s), approximately one-tenth of the movement duration of the voluntarily active crabs. The mean Table 1. Instantaneous velocity for voluntarily active and stressed ghost crabs Instantaneous velocity T b Mean Median Maximum Observation time Group ( C) (cm s 1 ) (cm s 1 ) (cm s 1 ) (s) Voluntarily active (N=12) 25.±.1 8.3± ± ± ±6.5 Stressed (N=15) 25.± ± ± ± ±.6 Values are reported as means±standard error. T b, body temperature.

7 Integrating field and laboratory studies of locomotion 995 Stride frequency (Hz) A (Pause) < > Fig. 5. Examples of focal animal movement and pause sequences. (A) Voluntarily active crab. (B) Stressed crab. Filled bars represent movements and open bars represent pauses. The horizontal axis indicates duration so that each row represents 15 s. Shading represents movement intensity: (1) light stippling, <2 strides s 1 ; (2) medium stippling, 2 3 strides s 1 ; (3) dark stippling, 3 4 strides s 1 ; and (4) black shading, >4 strides s 1. Note that the voluntarily active crab made longer movements with longer pauses, and with lower stride frequencies, than the stressed crab Time (s) B Time (s) pause duration was 7.6 s (range s), approximately one-third of the pause duration of the voluntarily active animals. However, the percentage of time spent moving was similar for the voluntarily active and stressed animals (i.e. between 1 and 5 % of the observational period; Fig. 6). The mean stride frequency of the captured and released crabs (3.6 strides s 1 ; range strides s 1 ) was significantly higher than that of the voluntarily active crabs (Mann Whitney U-test, P<.1; Fig. 5; Table 2). The ghost crab s stride frequency reaches a maximum at approximately 6 strides s 1 (Blickhan and Full, 1987). For the stressed crabs, the mean stride frequency of 18 out of a total of 51 movement periods achieved this maximum rate. Discussion This study provides the first measurements of instantaneous velocity determined by three-dimensional motion analysis on an unrestrained terrestrial animal in its natural habitat. In combination with quantitative measurements of locomotor Table 2. Focal animal sampling summary for voluntarily active and stressed ghost crabs Movement duration Pause duration Observation Time spent Mean stride T b Mean Median Mean Median time moving frequency Group ( C) (s) (s) (s) (s) (min) (%) (strides s 1 ) Voluntarily active 24.8± ±1. 8.5± ±3. 6.7±1. 12.± ± ±.1 (N=21) Stressed 25.2±.3 1.4±.2 1.2±.3 7.6±4. 5.3± ± ± ±.4 (N=3) Values are reported as means±standard error. T b, body temperature.

8 996 R. B. WEINSTEIN 5 6 Mean pause duration (s) <1 % 1 5 % >5 % Stride frequency (s 1 ) Run Walk Mean movement duration (s) Fig. 6. Mean pause duration as a function of mean movement duration for all focal animals. Open symbols represent voluntarily active crabs and filled symbols represent stressed crabs. The voluntarily active animals made longer movements alternated with longer pauses than the stressed crabs. Mean pause duration was not correlated with mean movement duration. Most animals (2 of 24) were active for between 1 and 5 % of the observation period. behavior, these data test predictions about natural locomotion based on laboratory studies of physiology, biomechanics and behavior. Voluntarily active crabs moved intermittently at mean velocities that were less than their MAS (Figs 5A, 6). The performance limits of the voluntarily active crabs are not likely to be altered if they move intermittently at these slow speeds rather than walking continuously at the same average speed. By contrast, the stressed crabs moved at speeds that are greater than their MAS. Stressed crabs moved intermittently, showed shorter movement and pause periods, and generally exhibited brief movement periods relative to pause periods compared with voluntarily active crabs (Figs 5B, 6). The intermittent locomotor behavior of the stressed crabs may increase performance limits since the crabs were able to move for longer periods (i.e. more than a few seconds) than predicted on the basis of their limited endurance capacity for continuous locomotion at these rapid speeds. To evaluate these conclusions, it is necessary to consider where the ghost crabs operate in nature with respect to the factors that affect locomotor performance limits. Therefore, to facilitate the integration of laboratory and field measurements, animals selected for observation in the field were of comparable body size and body temperature to those studied in the laboratory. Body size and temperature Carapace width, a simple measurement made in the field, was a good indicator of body mass. The body masses and corresponding carapace widths measured in the present study are similar to previously published values (Fig. 2A; Wolcott, 1978). The ghost crabs that were videotaped in the present study were primarily medium-sized crabs (approximately 2 4 g; Fig. 2B) Movement duration (s) Fig. 7. Mean stride frequency as a function of movement duration for voluntarily active focal animals (N=485 data points representing 21 individuals). The dashed line represents the stride frequency at the walk run transition determined in the laboratory (approximately 3.7 strides s 1 ; Blickhan and Full, 1987). Crabs moved at low mean stride frequencies (<3 strides s 1 ) during long movements (>2 s). The highest mean stride frequencies (>4 strides s 1 ) were sustained only during the shortest movements (<5 s). A stride frequency of approximately 3 strides s 1 is equivalent to 2 m s 1 (MAS; Blickhan and Full, 1987; Full, 1987). The mean body temperatures of the videotaped ghost crabs were close to 25 C (Tables 1, 2). Body temperature of the ghost crabs can differ by several degrees from ambient temperature, even at night (Fig. 3). Ghost crabs are cooled by evaporative water loss if the relative humidity is low enough to allow for water loss through the body surface (Weinstein et al. 1994). As predicted by controlled laboratory studies (Weinstein and Full, 1994), body temperature depression in the field was greatest at the lowest relative humidity (Fig. 3B). Although moderate dehydration decreases locomotor capacity in the ghost crab (Weinstein et al. 1994), several factors prevent even moderate dehydration in the field. First, the ambient relative humidity is generally high (above 75 %) when the crabs are most active. Second, the locations where the ghost crabs burrow and forage are able to support high rates of water uptake. The burrows of ghost crabs often extend below the water table, providing a humid and moist retreat. While foraging, the ghost crabs are often directly exposed to shallow water and spray from incoming waves. Finally, ghost crabs can extract water from soil that is less than 5 % water through setal tufts at the base of their walking legs (Wolcott, 1976, 1984). While even moderate dehydration is unlikely in the beach habitat of the ghost crab, the risk of dehydration probably restricts their distribution on land (Bliss, 1968, 1979). Interestingly, the crabs exposed to an ambient temperature of 2 C did not exhibit a body temperature depression, despite the low ambient relative humidity. One explanation may be that, at this low ambient temperature, crabs only emerge from their burrows for brief periods. Under controlled conditions in the laboratory, the half-time to reach a stable body temperature

9 Integrating field and laboratory studies of locomotion 997 is approximately 15 min (Weinstein and Full, 1994). The temperature in the burrows was warmer than that of the air outside the burrow at night, even just a few centimeters from the mouth of the burrow. If the ghost crabs had equilibrated with the ambient air temperature (2 C), their body temperatures would have been approximately 18 C. Assuming a Q 1 of 2 over this temperature range, the MAS would have decreased by 13 %. The Q 1 over this range may be closer to 4. (Weinstein and Full, 1994), which would give a 24 % decrease in the MAS. In any case, cooling from 2 to 18 C would result in a substantial decrease in locomotor performance capacity. Instantaneous velocity As determined by three-dimensional motion analysis, the voluntarily active crabs moved at a mean instantaneous velocity of 8.3 cm s 1 (Table 1). Similarly, the mean stride frequency of the voluntarily active focally sampled animals (2.1 strides s 1 ; Fig. 7; Table 2) is equivalent to a velocity of approximately 1cm s 1 (Blickhan and Full, 1987). Therefore, most of the time, the ghost crabs were moving at speeds below their MAS (approximately 2 cm s 1, approximately 3 strides s 1 ; Blickhan and Full, 1987; Full, 1987). In the laboratory, the preferred speed range of ghost crabs moving freely on a track is 1 3 cm s 1 (Blickhan and Full, 1987). Mechanically, these voluntarily active ghost crabs were primarily using a walking gait since the mean stride frequency was less than the stride frequency at the walk run transition (Fig. 7; Blickhan and Full, 1987). The metabolic energy for these slow speeds is probably supplied by aerobic metabolism, and although lactate concentrations were not measured in the field, they are likely to be close to resting levels (Full, 1987). Endurance capacity for continuous exercise at 8.3 cm s 1, the mean instantaneous velocity measured in the field, is greater than 2 h (Table 1; Full, 1987). By contrast, the stressed crabs moved at a mean instantaneous velocity (82.9 cm s 1 ; Table 1) that is greater than their MAS. Mechanically, the stressed crabs were moving at speeds that correspond to a running gait, since the stride frequency estimated from the mean instantaneous velocity would be greater than the stride frequency at the walk run transition (Fig. 7; Blickhan and Full, 1987). The mean stride frequency of the stressed crabs determined by focal animal sampling (3.6 strides s 1 ; Table 2) is equivalent to a velocity of only approximately 38 cm s 1 (Blickhan and Full, 1987). However, the mean stride frequency measured for each stressed individual underestimates its mean velocity because the mean stride frequencies of many of the movement periods were approximately 6strides s 1. The velocity of an individual moving at a stride frequency of 6 strides s 1 can be above 9 cm s 1, depending upon the stride length. The stressed crabs may be moving at speeds faster than 9 cm s 1 by increasing their stride length rather than their stride frequency, although this was not measurable by the focal sampling techniques used in this study. In any case, the stressed focal animal data are in general agreement with the mean instantaneous velocities determined by three-dimensional motion analysis and support the result that the stressed animals moved at a faster mean velocity than the voluntarily active crabs. In contrast to the slow, aerobically sustained movements of the voluntarily active crabs, the rapid movements of the stressed crabs could not be supported aerobically and would require additional energy from non-oxidative energy sources. The lactate levels of these stressed crabs, although not measured in this study, are likely to be elevated above resting levels (Full and Weinstein, 1992). Endurance capacity for continuous exercise at 82.9 cm s 1 is less than a few seconds, corresponding to a distance of less than 2 m (Table 1; Full, 1987; Full and Prestwich, 1986). Despite the low endurance capacity for continuous exercise at the mean instantaneous velocities measured in the field, the stressed crabs were observed to move over distances greater than 2 m, yet they did not appear to be completely fatigued (i.e. subsequent movements were observed following brief pauses). The greater distance capacity observed in the field, compared with the predicted distance capacity based on continuous locomotion, suggests that the performance limits of the stressed crabs are increased as a result of intermittent movement. Locomotor behavior Ghost crabs moving freely in their natural habitat moved intermittently (Fig. 5). The movement periods for the voluntarily active focal animals were brief (mean 11.2 s; Fig. 6; Table 2). The median movement duration was slightly shorter than the mean movement duration. The longest movement period duration for any individual was 74 s, although most were much shorter. The mean movement duration of the stressed crabs was only one-tenth of that of the voluntarily active crabs (Fig. 6; Table 2). The mean movement durations for the voluntarily active and stressed crabs were shorter than the time required to attain 5% of the steady-state rate of oxygen consumption, as determined in the laboratory for continuous locomotion (t 1/2,on =3 s; Full, 1987) and the time required to reach steady-state oxygen consumption (approximately 9 12 s; Full, 1987). In addition, ghost crabs did not move at a constant speed during each movement period. Instead, they accelerated and decelerated and started and stopped frequently (Fig. 4). The ghost crabs did not move for a long enough period to attain steady-state oxygen consumption during each movement period. However, Weinstein and Full (1992) demonstrated that the rate of oxygen consumption oscillates around a pseudo-steady state if intermittent exercise is carried out for long periods (more than 1 min). The voluntarily active ghost crabs made frequent short pauses and occasional longer pauses. This pattern is indicated by tremendous variation in pause period duration and a large difference between the median (6.7 s) and mean (23.4 s) pause duration. In the laboratory, it takes approximately 15 min for the rate of oxygen uptake to fall from steady-state exercise levels to resting rates (Full, 1987). It is unlikely that the ghost crabs pause long enough to recover completely from each

10 998 R. B. WEINSTEIN movement period. The stressed crabs, with faster movements and even shorter pauses than the voluntarily active crabs (Figs 5, 6), must recover from a greater metabolic disturbance (i.e. as a result of moving at faster instantaneous velocities) in a shorter time. Recovery in the field is probably a gradual process that is cumulative over an entire bout of activity. Although the duration of the pause period was not directly related to the previous exercise period for any individual, it may be more important physiologically for the mean pause period duration to be related to the mean movement duration. On average, the pause periods were twice as long as the exercise periods for the voluntarily active crabs and five times as long as the exercise periods for the stressed crabs (Fig. 6; Table 2). Integration of field and laboratory studies of locomotion As predicted by laboratory studies of continuous locomotor performance limits, ghost crabs generally moved at speeds below their MAS. Movement at the speeds recorded for the voluntarily active crabs confers energetic and biomechanical advantages to these highly active animals. A ghost crab moving continuously at 8.3 cm s 1, a speed that is 46 % of its MAS, will have an endurance capacity of more than 2 h (at a body temperature of 24 C; Full, 1987). In addition, intermittent locomotion is not likely to alter the ghost crab s performance limits at these slow speeds. If the ghost crab moves intermittently, alternating 11.2 s movement periods with 23.4 s pause periods, the average speed is only 2.7 cm s 1 (15 % of MAS). The endurance capacity for continuous exercise at 2.7 cm s 1 will also be greater than 2 h. In addition, energy for continuous locomotion at this slow speed will be supplied primarily by aerobic metabolism and will not result in a net accumulation of lactate. Starting and stopping frequently may increase the metabolic cost of movement at these slow speeds but is not likely to constrain the ghost crab s distance capacity (Weinstein and Full, 1992). In contrast to the voluntarily active crabs, the stressed crabs moved at average speeds greater than their MAS. According to predictions from laboratory studies of continuous locomotor performance, movement at these speeds would reduce endurance capacity and increase the metabolic cost of locomotion. A ghost crab moving continuously at 83 cm s 1, a speed that is 42 % of its MAS, would have an endurance capacity of less than 1 s (Full, 1987). Continuous locomotion at this speed, a sprinting speed, would be likely to result in the accumulation of lactate. However, the stressed ghost crabs moved intermittently, alternating 1.4 s movement periods with 7.6 s pause periods. As a result, the average speed was only 12.9 cm s 1 (65 % of MAS). Endurance capacity for continuous locomotion at 12.9 cm s 1 is approximately 55 min, corresponding to a total distance of 41 m. On the basis of laboratory data and limited field observations, the stressed ghost crabs have an increased distance capacity compared with crabs moving continuously at the same absolute speed (83 cm s 1 ). Although this specific combination of exercise speed (83 cm s 1 ), exercise duration (1.4 s) and pause duration (7.6 s) has not been tested in the laboratory, the stressed crabs may also have an increased distance capacity compared with crabs moving continuously at the same average speed (12.9 cm s 1 ). This expectation is based on the observation that, when the ghost crabs move at speeds greater than their MAS (i.e. the stressed crabs), their locomotor behavior (i.e. movement and pause duration and ratio of movement to pause) is different from that of crabs moving at speeds less than their MAS (i.e. the voluntarily active crabs). These behavioral changes have been shown to alter performance limits in the laboratory. Furthermore, the stressed crabs did not fatigue as they moved intermittently in the field. Instead, they were able to continue to escape. The long pause period duration, relative to the short movement duration, may allow the stressed crabs to recover partially from their brief high-intensity movements and may explain their increased performance capacity. Dr Robert J. Full provided constructive criticism at all phases of this project. I would like to thank S. Conova, K. Reinsel and D. Sanders for assistance with field data collection. The Duke University Marine Laboratory and Fort Macon State Park provided logistical support for the field component of this study. This work was supported by a National Science Foundation Graduate Fellowship, grants from the Lerner-Gray Fund for Marine Research and Sigma Xi, and an NSF Dissertation Improvement Grant DEB References ADAMCZEWSKA, A. M. AND MORRIS, S. (1994). Exercise in the terrestrial Christmas Island red crab Gecardoidea natalis. II. Energetics of locomotion. J. exp. Biol. 188, BENNETT, A. F., HUEY, R. B. AND JOHN-ALDER, H. (1984). Physiological correlates of natural activity and locomotor capacity in two species of lacertid lizards. J. comp. Physiol. B 154, BIEWENER, A. AND FULL, R. J. (1992). Force platform and kinematic analysis. In Biomechanics: Structures and Systems A Practical Approach (ed. A. Biewener), pp New York: IRL at Oxford University Press. BLICKHAN, R. AND FULL, R. J. (1987). Locomotion energetics of the ghost crab. II. Mechanics of the center of mass. J. exp. Biol. 13, BLISS, D. E. (1968). Transition from water to land in decapod crustaceans. Am. Zool. 8, BLISS, D. E. (1979). From sea to tree: Saga of a land crab. Am. Zool. 19, FULL, R. J. (1987). Locomotion energetics of the ghost crab. I. Metabolic cost and endurance. J. exp. Biol. 13, FULL, R. J. (1991). The concepts of efficiency and economy inland locomotion. In Efficiency and Economy in Animal Physiology (ed. R. W. Blake), pp Cambridge: Cambridge University Press. FULL, R. J., ANDERSON, B. D., FINNERTY, C. M. AND FEDER, M. E. (1988). Exercising with and without lungs. I. The effects of metabolic cost, maximal oxygen transport and body size on terrestrial locomotion in salamander species. J. exp. Biol. 138, FULL, R. J. AND PRESTWICH, K. N. (1986). Anaerobic metabolism of walking and bouncing gaits in ghost crabs. Am. Zool. 26, 88A. FULL, R. J. AND TULLIS, A. (199). Capacity for sustained terrestrial

11 Integrating field and laboratory studies of locomotion 999 locomotion in an insect: Energetics, thermal dependence and kinematics. J. comp. Physiol. B 16, FULL, R. J. AND WEINSTEIN, R. B. (1992). Integrating the physiology, mechanics and behavior of rapid running ghost crabs: Slow and steady doesn t always win the race. Am. Zool. 32, GARLAND, T. (1985). Physiological and ecological correlates of locomotory performance and body size in lizards. PhD thesis, University of California at Irvine. 21pp. Chapter 5. HERTZ, P. E., HUEY, R. B. AND GARLAND, T. (1988). Time budgets, thermoregulation and maximal locomotor performance: Are reptiles Olympians or boy scouts? Am. Zool. 28, HUEY, R. B. AND PIANKA, E. R. (1981). Ecological consquences of foraging mode. Ecology 62, JOHN-ALDER, H. B. AND BENNETT, A. F. (1981). Thermal dependence of endurance and locomotory energetics in a lizard. Am. J. Physiol. 241, R342 R349. JOHN-ALDER, H. B., LOWE, D. H. AND BENNETT, A. F. (1983). Thermal dependence of locomotory energetics and aerobic capacity of the Gila monster (Heloderma suspectum). J. comp. Physiol. 151, KENAGY, G. J. AND HOYT, D. F. (1989). Speed and time energy budget for locomotion in golden-mantled ground squirrels. Ecology 7, MCLAUGHLIN, R. L. (1989). Search modes of birds and lizards: Evidence for alternative movement patterns. Am. Nat. 133, MCLAUGHLIN, R. L., GRANT, J. W. A. AND KRAMER, D. L. (1992). Individual variation and alternative patterns of foraging movements in recently-emerged brook charr (Salvelinus fotinalis). Behaviour 12, PIETRUSZKA, R. D. (1986). Search tactics of desert lizards: how polarized are they? Anim. Behav. 34, REYNOLDS, D. R. AND RILEY, J. R. (1988). A migration of grasshoppers, particularly Diabolocatantops axillaris (Thunberg) (Orthoptera: Acrididae), in the West African Sahel. Bull. ent. Res. 78, TAYLOR, C. R., SCHMIDT-NIELSEN, K. AND RAAB, T. L. (197). Scaling of energetic cost of running to body size in mammals. Am. J. Physiol. 219, WALLIN, H. (1991). Movement patterns and foraging tactics of a caterpillar hunter inhabiting alfalfa fields. Funct. Ecol. 5, WEINSTEIN, R. B. AND FULL, R. J. (1992). Intermittent exercise alters endurance in an eight-legged ectotherm. Am. J. Physiol. 262, R852 R859. WEINSTEIN, R. B. AND FULL, R. J. (1994). Thermal dependence of locomotor energetics and endurance capacity in the ghost crab, Ocypode quadrata. Physiol. Zool. 67, WEINSTEIN, R. B., FULL, R. J. AND AHN, A. N. (1994). Moderate dehydration decreases locomotor performance of the ghost crab, Ocypode quadrata. Physiol. Zool. 67, WOLCOTT, T. G. (1976). Uptake of soil capillary water by ghost crabs. Nature 264, WOLCOTT, T. G. (1978). Ecological role of ghost crabs, Ocypode quadrata (Fabricius) on an ocean beach: Scavengers or predators? J. exp. mar. Biol. Ecol. 31, WOLCOTT, T. G. (1984). Uptake of interstitial water from soil: Mechanisms and ecological significance in the ghost crab Ocypode quadrata and two gecarcinid land crabs. Physiol. Zool. 57, YABE, T. (1992). Sexual difference in annual activity and home range of the Japanese pond turtle, Mauremys japonica, assessed by mark recapture and radio-tracking methods. Jap. J. Herpetol. 14,