Rapid solving of a problem apparatus by juvenile black-throated monitor lizards (Varanus albigularis albigularis)

Transcription

1 Anim Cogn (28) 11: DOI 1.17/s ORIGINAL PAPER Rapid solving of a prolem apparatus y juvenile lack-throated monitor lizards (Varanus aligularis aligularis) Jennifer D. Manrod Ruston Hartdegen Gordon M. Burghardt Received: 5 April 27 / Revised: 15 August 27 / Accepted: 4 Septemer 27 / Pulished online: 25 Septemer 27 Springer-Verlag 27 Astract It is widely accepted that providing stimulus enrichment is an important part of the development and maintenance of ehavior and well-eing in mammals. However, extending this idea to non-avian reptiles has arely een explored, certainly as an aid to cognitive development. Monitor lizards have a reputation for eing highly curious and intelligent lizards, ut quantitative experiments are necessary to evaluate such impressions as well as the value of providing enrichment to captive squamate reptiles. In this study eight juvenile lack-throated monitors, Varanus aligularis, were tested in their home enclosures with three presentations, at weekly intervals, of a novel task apparatus: a transparent food tue containing several prey. The food tue allowed the monitors to otain prey y using hinged doors at either end of the tue to access food. All eight lizards learned to open the tue, insert head, and capture the prey within 1 min in the Wrst trial. By the second trial, oth mean latencies to access the tue and capture the Wrst prey item decreased signiwcantly, as did the use of inevective responses such as shaking the tue. A further J. D. Manrod (&) Department of Conservation and Science, Knoxville Zoological Gardens, P.O. Box 64, Knoxville, TN 37914, USA jmanrod@knoxville-zoo.org R. Hartdegen Department of Herpetology, Dallas Zoo and Aquarium, 65 South R.L. Thronton Freeway, Dallas, TX 7523, USA ruston17@yahoo.com G. M. Burghardt Department of Psychology, The University of Tennessee, 31a Austin Peay Building, 144 Circle Drive, Knoxville, TN , USA gurghar@utk.edu slight decrease occurred in the third trial. Due to the results of this and similar studies, serious consideration should e given to further testing of cognitive ailities in squamate reptiles. Incorporating prolem solving tasks may also e useful to increase the activity level and captive well-eing of squamate reptiles, especially monitor lizards. Keywords Lizard Varanus Learning Prolem solving Enrichment Introduction There has een a resurgence of interest in the cognitive ailities of diverse species as seen in the Welds of comparative cognition, ehavioral ecology, and cognitive ethology (Shettleworth 1998). Although a numer of studies of the role of learning and experience in ectothermic reptiles has accumulated since the last thorough review of learning in this group (Burghardt 1977), studies of higher level cognitive processes and prolem solving have lagged. Testing diverent species on comparale prolems adapted to their sensory and manipulative ailities can e useful in uncovering comparative prolem solving ailities and their possile evolution (Davis and Burghardt 27). The impetus to provide environmental enrichment to captive animals to enhance their well-eing and encourage normal ehavior has also served to enhance the diversity of species studied. With a few exceptions, however (e.g., Burghardt et al. 1996; Almli and Burghardt 26), nonavian reptiles have not een emphasized in this resurgence, which has een largely devoted to nonhuman primates, carnivores, rodents, and some irds such as passerines and parrots. Non-avian reptiles were included in an early comparative study of curiosity in zoo animals toward introduced

2 268 Anim Cogn (28) 11: ojects (Glickman and Sroges 1966). An Orinoco crocodile (Crocodylus intermedius) was rated highest in response to an introduced oject followed y three lizards, two of which were monitors. Monitor lizards (Varanus) may, in fact, e excellent candidates for the study of higher cognitive ailities in squamate reptiles. The genus includes the world s largest lizards, which are often considered to e perhaps the most intelligent lizards (Pianka and Vitt 23; Burghardt et al. 22). Behaviorally, their active foraging ailities on diverse prey contriute to this impression (Kaufman et al. 1994, 1996; Phillips 1995). Varanids general physiology also contriutes to these animals eing considered more mammal-like then other non-avian reptiles. Their capacity for sustained activity, eycient circulatory and pulmonary systems, and metaolic rate transcends the capailities of most other non-avian reptiles and approaches the warm-looded mammalian level (Steel 1996; Horn 1997). The high aeroic metaolic rate of many varanids is an exception among lizards and may e related to their large ody sizes. Squamate metaolic rates increase with increasing mass at a higher rate than in mammals (Andrews 1985). The relative increase in the standard metaolism of varanids may also rexect life history traits, particularly their widely foraging and primarily carnivorous diets (Thompson 1997). Ecologically, monitors are comparale to small to medium-sized species of predatory mammals. They occupy similar ecological niches in Australia and compete well with mammals of comparale size in Africa and Asia. Varanids successfully inhait diverse ecological environments including deserts, savannas, steppes, relatively dry forests, tropical humid rainforests, rivers and riveranks, and marine shores (Horn 1997). All of these physiological, life history, and ehavioral traits seem to allow monitors to e more mammal-like in their predatory strategy and communication as well as overall ehavior (AuVenerg 1978). The Dallas Zoo and the University of Tennessee Reptile Ethology La developed a joint environmental enrichment study with hatchling lack-throated monitors (Varanus aligularis aligularis). As part of this study, eight juvenile monitors raised in an enriched environment, as compared to standard housing, were systematically tested on their reactions to, and interaction with, ojects and novel conspeciwcs. All animals received the enrichment ojects in the same order, ut none provided food rewards except for the food tue prolem discussed here. This paper presents the results of the trials involving the food tue introductions, in which the animal had to learn how to otain prey they could see and smell ut had to access through a hinged door. Although it was hypothesized that the lack-throated monitors would attain and capture the prey in the food tue more quickly over trials, we did not expect to see the rapid learning oserved, and thus feel these data are noteworthy. Methods Sujects The sujects were eight monitors reared in an enriched environment at the Dallas Zoo and were from the same commercially red clutch hatched in Novemer of 2. The monitors were measured and weighed each month after hatching. Because the animals were under 1 year of age during testing, three individuals could not e sexed due to the fact that hemipene calciwcation does not usually show up on radiographs efore one year. Susequent to the experiment, most of the lizards were dispersed so further data on them are not availale. Housing and maintenance The eight lizards were individually housed ov-exhiit in four acrylic enclosures measuring cm. Each of these enclosures was divided y an opaque plywood wall into two cages measuring cm. Each cage contained 8 cm of Cyprus mulch sustrate, a horizontal shelf (aout 15 cm high and 61 cm long) that ran along the length of the ack wall, a hide ox (approx cm) that was placed under the shelf to allow more Xoor space, and painted enclosure walls so the animals could not see one another ut could see the outside activity. Each cage had an 85 W spot lamp providing a asking area of aout 3 4 C as well as a doule ul 61 cm Xuorescent Wxture with 2 W Sylvania 35BL uls. A plastic water owl measuring approximately 15 cm in diameter contained water 6 cm deep. As part of the approved enrichment protocol at the Dallas Zoo, the lizards were fed live prey, crickets and mice, throughout the experiment, with larger mice provided as the lizards grew. For carnivorous reptiles, live prey may contriute to their psychological well-eing and ehavioral development (Burghardt 1996; Almli and Burghardt 26). Live young mice were used in the tests here as they were familiar prey. 1 Test procedure and apparatus Each test session included a 15-min introduction following aout 4 min of pre (aseline) and aout 1 min of post introduction oservations. All sessions were taped from outside of the enclosure and were recorded in full. The 1 It is common practice to use live rodent prey to study predatory ehavior, as enrichment, and as a part of routine feeding in reptiles in zoo, aquarium, private, and research collections (Deufel and Cundall 23, 26; Hartdegen et al. 1999, 2; Mori 1993; Mehta 23).

3 Anim Cogn (28) 11: three food tue trials were conducted weekly over a 3- week period. The prolem task involved presentation of a clear heavy plastic tue cm with four rows of holes (.5 cm in diameter) and two hinged doors. Every other row had Wve holes drilled in a parallel line along the tue. Two rows had three holes followed y a hinged door ( cm) on opposite ends and opposite sides of the tue (Fig. 1). Three or four live small mice (fuzzies) were placed inside the tue. To otain the mice the monitor had to rotate the tue and open the door y using either its foreclaws, snout, or, if the tue was rolled the correct way, gravity. The lizard then had to insert its head, and often its forequarters, into the tue to gra the mouse, after which it always exited the tue efore swallowing the prey whole. The mice did not appear highly reactive to the presence of the monitors efore capture, and any stress they experienced seemed far less than in most current mouse ehavioral testing and phenotyping protocols. The monitors were 6 7 months old during the food tue trials and were at that time eing fed 2 3 live young mice once a week. Trial days were spaced etween feedings. Usually the monitors would go one weekend without food (Friday through Sunday), tested on Monday, and fed on Tuesday. Fig. 1 Food tue. The clear heavy plastic tue with four rows of holes (.5 cm in diameter) and two hinged doors (one shown open) Ethogram Ten ehavior patterns were prevalent during the food tue trials. These ehavioral categories were divided into four states and six events (Lehner 1996). Descriptions of the state and event ehaviors, the areviation for the ehavior, and a rief description of the movement or position involved in each ehavior are provided in Tales 1 and 2. The Oserver program (Noldus) was used to quantify the ehavior in the trials through a repeated viewing of the tapes. Tongue Xick (TF) frequencies are often used as a surrogate measure of the overall activity in highly chemosensory squamate reptiles (Burghardt and Pruiit 1975). They rexect how active the monitors were efore the oject was introduced (pre-introduction), during the trial, and after the oject was removed (post-removal). Statistics Comparisons were made across trials and among the individuals. Both mean occurrence and percent duration (to adjust for diverences in trial times) were compared. Data for all ehaviors oserved in the trials were analyzed in three ways, including duration, occurrence, and changes over trials. Due to its roustness to violations of normality, MANOVA was used in some analyses to compare the trials and repeated measure MANOVAs were conducted on the tongue Xick data to compare the within trial tongue Xick counts over 2- min intervals to evaluate within-session changes that may have occurred. The majority of analyses relied on the nonparametric Wilcoxon Matched Pairs Signed Rank and Friedman tests to compare changes etween trials over individuals for latencies to enter the food tue, to capture prey, for changes in use of ehavior patterns, and to compare tongue Xick rates in the pre, during, and post intervals. The tongue Xick data and ehaviors were also used to compare the Wrst, second, and third food tue trial. Analyses were conducted using SPSS 11.5 (SPSS Inc., ) using an alpha of Tale 1 State ehaviors oserved during videotape analysis in the food tue trials Behavior pattern Description Approach (AP) Movement directed toward a speciwc item in the enclosure. This ehavior was most often enacted when the monitor was approaching the tue (APo) or when the monitor was approaching the glass of the enclosure (APot) Moving (MO) Apparent purposeful physical movement of the introduced oject y the monitor. For example, the monitor would pick up or push the oject usually with the head/mouth and move it around the enclosure Head insertion (HI) Head inserted at least two-thirds into the oject for at least 5 s. This occurred with the tue (HIo) ut rarely with the sustrate (HIot). Interact (IN) Interacting with oject. Monitor spent more than 5 s in physical contact and/or chasing and eing attentive to oject introduced

4 27 Anim Cogn (28) 11: Tale 2 Event ehaviors oserved during videotape analysis in the food tue trials Behavior pattern Tongue Xick (TF) Fore scrape (FS) Mouth gra (MG) Shake (SH) Bite (BI) Capture prey (CP) Description A single extension and retraction of the tongue from the mouth. The fast tongue Xick was recorded when just the simple motion of ringing the tongue in and out in a fairly quick manner occurred Moving the forearm and claws onto the oject and then pulling ack and out along the sustrate A ite at the oject with mouth open followed y the oject eing held in the mouth for more than.1 s. It was typically followed y a shake or head insertion With the oject held in the mouth, the monitor would shake the head ack and forth. Shaking was categorized in two forms. Shake fast (SHf) occurred in.1 s or less. Shake slow (SHs) occurred in greater than.1 s Quick snap towards oject with mouth open, though oject is not held with the mouth as in mouth gra Recorded at the time the monitor Wrst captured prey, followed y ingestion.5 and were one-tailed when a directional hypothesis was eing tested. Non-parametric tests were ased on Siegel (1956) and Conover (1999). Results We measured prolem solving through latency to Wrst head insertion into the tue (Fig. 2) and to capture the Wrst prey (Fig. 3). There were signiwcant diverences across trials in oth measures (Tale 3). Using the conservative Wilcoxon matched-pairs signed rank test, head insertion (HI) time decreased etween the 1st and 2nd tue trials (T =5, n =8, P <.5) and the 1st and 3rd tue trials (T =1, n =8, P <.1). This result suggests that y the second trial, the monitors had learned to enter through the hinged doors more quickly. A similar, though even more dramatic, change was seen in the latency to Wrst prey capture (CP). The latency decreased etween the 1st and 2nd tue trials (T =, n =8, P <.5) and the 1st and 3rd tue trials (T =, n =8, P <.5). On average, the monitors took 6.5 min.53 SE to capture any one prey in the Wrst food tue introduction. In the second and third food tue introductions, the monitors took only 2 3 min to capture any one prey (2nd SE, 3rd SE min). In oth latency to the Wrst HI and the Wrst CP, the diverences etween the second and third introductions were not signiwcant (Figs. 2 and 3). Evaluation of the ehavior patterns used y the lizards with each successive trial helps to explain the shorter latencies. By the second trial, the monitors ecame more prowcient at using fore scrape (FS) at the hinged doors to open the tue and attain and capture prey (CP) (Fig. 4). In the third set of trials, the individuals captured the prey (CP) within the Wrst 2 min and then graed the food tue y the open doors and moved it around the enclosure, if they continued to interact with the tue at all. Comparing these two measures, head insertion (HI) and capture prey (CP), the data indicate that the monitors were ale to learn to gain Mean Latency to Head Insert (Minutes) + SE a 1st Food Tue 2nd Food Tue 3rd Food Tue M ean Latency to Capture Prey (Minutes) + S E a 1st Prey a 1st FT 2nd FT 3rd FT 3rd Prey Fig. 2 Mean latency per individual per trial to head insertion in the 1st, 2nd, and 3rd food tue trials. SigniWcant diverences (P <.5 trial 1 and 2; P <.1 trial 1 and 3, see text for details) indicated y letters Fig. 3 Mean latency per individual per trial to capture 1st and 3rd prey in the 1st, 2nd, and 3rd food tue trials. SigniWcant diverences (P <.5, see text for details) indicated y letters

5 Anim Cogn (28) 11: Tale 3 Univariate tests (etween-sujects) of latency to head insert (HI) and capture (CP) in food tue trials *SigniWcant Source Dependent variale Type III sum of squares df Mean square F P Corrected model HI* CP* Intercept HI* <.1 CP* <.1 FT trial HI* CP* Error HI CP Mean Occurrence + SE Mean Occurrence + SE SN BI SH Behavior Pattern IN Behavior Pattern Fig. 4 Mean frequency of interaction (IN) event (top) and most frequent state and event (ottom) ehavior patterns in the 1st, 2nd, and 3rd food tue trials. Areviations of ehavior patterns as in Tales 1 and 2 access to the mice y inserting their head into the food tue and ecome more prowcient at catching and consuming the prey inside. In all of the trials, there were signiwcantly more tongue Xicks during the oject introductions than in the pre or post interval (F 2,16 = 5.662, P <.1) and there were signiwcantly more tongue Xicks in the post interval as compared to the pre interval (F 2,16 = 4.11, P <.6) (Fig. 5). These results demonstrate that the oject did have a positive (increase) evect on the monitor s activity level during the introduction as well as after the oject was removed. FS FT1 FT2 FT3 FT1 FT2 FT3 Minute Mean Frequency Per st Food Tue 2nd Food Tue 3rd Food Tue Time (Minutes) Fig. 5 Mean frequency SE of tongue Xicks within 2-min intervals in the aseline ( 5 min, Wrst arrow), oject present (5 2 min), and post trial (2 3 min, second arrow) phases in the 1st, 2nd, and 3rd food tue trials The moving mice and their odors quickly attracted the lizards (AP). The most frequent mode of interaction with the food tue involved head insertion (HI) which was the ehavior elicited in order to capture (CP) the prey within the tue. Before and during HI the monitors would also fore scrape (FS) Wrst at the hinged doors, then inside the tue, as they searched for the prey. There were also outs of ites (BI) efore and after prey was ingested. BI was usually directed at the hinge of the doors and was fairly rapid and repeated. Shake (SH) took place when the lizard it the tue, lifted, and moved it from side to side. Overall the monitors were very active in the trials. During the Wrst food tue trials, predatory ehaviors towards the tue were elicited and these changed over the three trials. The frequency of FS and interaction (IN) ehaviors increased (Fig. 4). These ehaviors most likely increased due to the greater prowciency y the monitors of oth attaining entry to the tue and capturing the prey. The FS ehavior was utilized to open the hinged doors, and then the monitors would insert their heads into the tue in order to attain and ingest the prey inside. Over time there was also less time spent locomoting and searching for the prey.

6 272 Anim Cogn (28) 11: The frequency of SH decreased, as did the latency to HI. These results demonstrated that the monitors spent less time in the second and third trials trying to gra and shake the tue efore opening the doors and capturing the prey. Mouth gra of the tue (MG) signiwcantly declined etween the initial (mean = 1.6) and each susequent trial (mean of trial 2 = 2.6, trial 3 = 3.) (T =2, n =7, P <.5; T =1, n =7, P <.25). As shown in Fig. 4, SH also declined, ut as many animals did not perform SH, the change was not signiwcant. Activity across successive food tue trials as measured y tongue Xick (TF) rates during oject presence did not signiwcantly vary across trials (χ 2 =1.75, df =2) due to the variaility among lizards. Nevertheless, the results suggest that the overall activity levels of the monitors increased after the Wrst trial and then declined in the third trial (the mean frequency of tongue Xicks after the tue was introduced to its removal in the Wrst trial was SE, for the second trial the mean was SE, and in the third trial the mean was SE. The tongue Xick data were also compared over 2-min intervals within the oject introduction trials (Fig. 5). Comining the trials, the 2-min intervals were not signiwcantly diverent when all were compared to each other in an ordered repeated MANOVA. However, there was an increase during the Wrst interval after tue introduction (at aout 4 min), after which the tongue Xicking remained at a consistently high rate. This increase was not signiwcant in the Wrst trial (T =7, n =7, P >.1), ut was signiwcant in food tue trials 2 (T =, n =8, P <.1) and 3 (T =, n =8, P <.1). This stailization was followed, after the oject was removed, y a progressive, ut erratic, decrease of tongue Xicking to near the aseline rate (Fig. 5). Interestingly, this decrease was most dramatic in the third trial. Discussion The monitors clearly ecame more evective in opening the tue and otaining prey in the prolem device. Not only did they rapidly learn to otain food in the tues, drop less evective ehavior patterns, ut also developed high rates of tongue Xicking and activity when the oject was introduced. They also showed a decrease in activity after the tue was removed in later trials. Oservations of this species in a food selection task indicated that animals learned when no more food was forthcoming (Kaufman et al. 1996) and they might even have counting-like skills (Pianka and Vitt 23). A possile criticism is that the lizards may have een initially neophoic to the tue and thus the learning was primarily haituation. This seems unlikely as the lizards generally approached the oject (AP) quickly. In fact, out of the 24 trials (3 for each of 8 lizards), in 19 of them the lizards approached the tue within 1 s after introduction and in 21 trials within 17 s or less. Three lizards (612, 613, and 621) did take a minute or more to approach the tue on test day 1, ut even discounting these animals, the decreases in latency over trials for oth HI and CP were signiwcant. Head insertion time diverences were due primarily to the dropping of inevective ehaviors and increasing evective ones. Thus, the animals were not avoiding the tue so much as initially interacting less evectively with it in terms of gaining food. Head insertion time was less variale than successful capture time. The ehaviors oserved here are typically seen in feeding in captive and wild monitors. Both MG and SH were oserved and descried y AuVenerg (1981) from Weld studies on Komodo monitors. The SH occurred when the prey was slung side-to-side and sometimes pounded against the ground (AuVenerg 1981). It appears to e derived from a predatory movement for stunning, killing, or further tearing the prey. The MG was descried as a ite followed y gripping with the teeth (AuVenerg 1981). Overall, most of these ehaviors have een oserved and documented in Weld studies of wild monitors. From these descriptions, the ehaviors oserved in captivity resemle those in the wild. Behavior seen in captivity can lead us to reassess ehavior reported in wild monitors as well as to look for novel foraging tactics. This study demonstrated evidence of learning and the potential of enrichment activities that involve cognitive tasks in monitor lizards. Perhaps most signiwcantly, renewed interest in the cognitive ailities of ectothermic reptiles may e rewarding for the scientists as well as the animals themselves. Acknowledgments This work was supported y the Dallas Zoo. All research was conducted under the approval of the Dallas Zoo s Animal Care and Use Committee and complies with the laws of the United States. We thank the following people for their support: B. Aucone, C. Bennett, K. Bradley, R. Burger, W. Card, L. Mitchell, R. Reams and C. Watson. We also thank the University of Tennessee and the Reptile Research Fund. References Almli L, Burghardt GM (26) Environmental enrichment alters the ehavioral prowle of ratsnakes (Elaphe). J Appl Anim Welf Sci 9:85 19 Andrews RM (1985) Metaolism of squamate reptiles: allometric and ecological relationships. Physiol Zool 58(2): AuVenerg W (1978) Social feeding ehavior in Varunus komodoensis. In: Greenerg N, McLean PD (eds) Behavior and neurology of lizards. Government Printing OYce, Washington, pp AuVenerg W (1981) Behavioral ecology of the Komodo monitor, University of Florida Press, Gainesville Burghardt GM (1977) Learning processes in reptiles. In: Gans C, Tinkle DW (eds) The iology of the reptilia. Academic, London, 7:

7 Anim Cogn (28) 11: Burghardt GM (1996) Environmental enrichment of controlled deprivation? In: Burghardt GM, Bielitski JT, Boyce JR, Schaefer DO (eds) The well-eing of animals in zoo and aquarium sponsored research. Scientists Center for Animal Welfare, Greeenelt, pp Burghardt GM, Pruitt CH (1975) The role of tongue and senses in feeding of neworn garter snakes. Physiol Behav 14: Burghardt GM, Ward B, Rosscoe R (1996) Prolem of reptile play: environmental enrichment and play ehavior in a captive nile soft-shelled turtle, Trionyx triunguis. Zoo Biol 15: Burghardt GM, Chiszar D, Murphy JB, Romano J, Walsh T, Manrod J (22) Behavioral complexity, ehavioral development, and play. In: Murphy JB, CioW C, de La Panouse C, Walsh T (eds) Komodo Dragons: iology and conservation. Smithsonian Institution Press, Washington, pp Conover WJ (1999) Practical nonparametric statistics. Wiley, New York Davis KM, Burghardt GM (27) Training and long-term memory of a novel food acquisition task in a turtle (Pseudemys nelsoni). Behav Processes 75: Deufel A, Cundall D (23) Feeding in Atractaspis (Serpentes: Atracaspididae): a study in conxicting functional constraints. Zoology 16:43 61 Deufel A, Cundall D (26) Functional plasticity of the venom delivery system in snakes with a focus on the poststrike prey release ehavior. Zool Anz 245: Glickman SE, Sroges RW (1966) Curiosity in zoo animals. Behaviour 26: Hartdegen RW, Chiszar D, Murphy JB (1999) Oservations on the feeding ehavior of captive Varanus eccari. Amphi-reptil 2(3): Hartdegen RW, Roerts DT, Chiszar D (2) Laceration of prey integument y Varanus prasinus (Schlegel 1839) and V. eccarii (Doria 1874). Hamadryad 25(2): Horn HG (1997) Evolutionary eyciency and success in monitors: a survey on ehavior and ehavioral strategies and some comments. Mertensiella 11: Kaufman JD, Burghardt GM, Phillips JA (1994) Density-dependent foraging strategy of a large carnivorous lizard, the savanna monitor (Varanus aligularis). J Comp Psychol 18: Kaufman JD, Burghardt GM, Phillips JA (1996) Sensory cues and foraging decisions in a large carnivorous lizard, Varanus aligularis. Anim Behav 52: Lehner PN (1996) Handook of ethological methods, 2nd edn. University Press, New York Mehta RS (23) Prey handling ehavior of hatchling Elaphe helena (Coluridae). Herpetologica 59: Mori A (1993) Prey handling ehavior of neonate rat snakes, Elaphe taeniura and E. dione (Coluridae). Jpn J Herpetol 15(2):59 63 Pianka ER, Vitt HG (23) Lizards: windows to the evolution of diversity. University of California Press, Los Angeles Phillips JA (1995) Movement patterns and density of Varanus aligularis. J Herpetol 29: Shettleworth SJ (1998) Cognition, evolution, and ehavior. Oxford University Press, New York Siegel S (1956) Nonparametric statistics for the ehavioral sciences. McGraw-Hill, New York Steel R (1996) Living dragons. A natural history of the world s monitor lizards. Ralph Curtis Books, Saniel Island Thompson GG (1997) Standard and maximal metaolic rates of goannas (Squamata: Varanidae). Physiol Zool 7:37 323