Journal of Herpctology, Vol. 35, No. 3, pp. 395402, 2001 Copyright 2001 Society for the Study of Amphibians and Reptiles

Habitat Use and Activity Patterns of the Neotropical Arboreal Lizard Tropidurus ( = Uracentron) azureus werneri (Tropiduridae)

NORBERT ELLINGER,~ GERHARD SCHLATTE, NICOLE JEROME, AND WALTER HÖDL

Drpavtment of Evolutionary Biology, Institute of Zoology, Uniwusity of Vienna, Althanstvasse 14, A-1090 Vienna, Austria

ABSTRACT.-We studied habitat use, activity patterns, foraging mode, and prey spectrum of Tropidurus azureus werneri using a tower crane located in an Amazonian lowland rain forest in southern Venezuela. The lizards were strictly arboreal with a preference for the branches and twigs of the canopy. The horizontal distribution of lizards within the 1.4-ha study area was clustered and remained remarkably stable during two consecutive years. The lizards favored high and well-structured crowns with tree holes that served as shelters during the night. Activity was highest at air temperatures above 28°C measured at a height of 21 m. Activity patterns were influenced by cloud Cover and insolation. Tvopidums azureus werneri showed notable agility and tended toward more active foraging patterns than found in other tropidurid lizards. The prey spectrum was dominated by ants.

Basic ecological data on reptiles living high in the trees of tropical rain forests are sparse, pri- marily because of the difficulties of observing reptiles more than a few meters above ground level and the challenges of conducting studies within the forest canopy (Reagan, 1995). Prior to the development of safe canopy-access meth- ods and effective traps for use on trunks and in tree crowns, arboreal lizards were mainly col- lected by shooting or by searching logged trees. These methods, however, did not yield large numbers of individuals, nor did they contribute much knowledge about ecology and behavior. For example, Duellman (1978) collected only two individuals of the arboreal iguanid lizard Tropidurus (Uracentron) flmiceps (Tropiduridae)

Corresponding Author. E-mail: norbert.ellinger@ univie.ac.at

during four years of fieldwork in Ecuador (for the still disputed synonymization of the tropi- durid genera Plica and Umcentron with Tropidu- rus, See Frost, 1992; Avila-Pires, 1995; Harvey and Gutberlet, 1998). However, by using canoes to access the limbs of trees in a flooded lagoon, Vitt and Zani (1996a) were able to make more than 100 observations of T flmiceps during a sin- gle wet season, and by covering known retreats of lizards with minnow traps, they captured 23 lizards within one week (Zani and Vitt, 1995).

Basic ecological data of lizards found in the canopy are of special interest because this hab- itat differs in many ways from the forest interior near ground level. The canopy consists of a dense three-dimensional matrix of leaves and small branches and is exposed to full sun and great variability in humidity and high midday temperatures. In contrast, the forest interior is typically darker, cooler, more humid, and rela-

N. ELLINGER ET AL.

tively Open (Reagan, 1995). The majority of eco- logical studies on tropidurid lizards deals with species of open formations or of forest interior (cf. Howland et al., 1990; Vitt, 1993; Vitt et al., 1996; Cruz, 1998). Detailed studies on arboreal Tropiduridae exist for T. (U.) flaviceps (Vitt and Zani, 1996a), T. (Plica) umbva (Gasnier et al., 1994; Vitt et al., 1997), and T. (P.) plica (Vitt, 1991). Scientific reports of 7: azureus (U azuv- eum) are less extensive. Observations by Cott (1926), and Greene (1977) and the analysis of four stomach contents (Hoogmoed, 1973) sug- gested that the prey spectrum of I: azuveus is dominated by ants. Hoogmoed (1973) also pub- lished some anecdotal data on the reproduction of T azureus. Two eggs of I: azuveus zuevneui, a subspecies distributed in northwestern Ama- zonia (Avila-Pires, 1995), were found by Mäg- defrau (1991) between roots at the base of a tree at Rio Mavaca. Venezuela. We used a tower crane located in the Amazonian lowland rain forest of southern Venezuela to investigate this arboreal lizard and focused on habitat use, spa- tial distribution within the canopy, activity pat- terns, foraging mode, and type of prey.

MATER~ALS AND METHODS

Study Site.-The study site is located near the junction of the river Orinoco and the blackwater tributary Surumoni, 15 km west of La Esmer- alda, Estado Amazonas, Venezuela (3'10'N 65"401W, 105 m above sea level). The plot lies within a primary lowland rain forest that has not been cleared or selectively logged in historic time and contains mostly terra firme forest. The height of canopy trees ranges from 20 m to 34 m. They form an irregularly contoured and pre- dominantly closed canopy. The average annual temperature in the study area is approximately 26"C, usually with slight variations between the coolest month (25°C) and the warmest month (26.5"C), whereas a daily range of 5-10°C fre- quently occurs. The relative air humidity resch- es almost 100% during the night and decreases to 50-60% above the canopy on sunny days. An- nual precipitation is near 3000 mm with a strong peak from May to July and a lower peak in September and October. The tower crane run- ning on rails offers access to the canopy within an area of 1.4 ha (Anhuf et al., 1999).

Observation and Cqtuve.-Haphazard searches for lizards on trunks and in tree crowns were conducted with binoculars in 1997 (55 h on 14 mornings and 11 afternoons from February to May) and 1998 (142.5 h on 19 mornings and 27 afternoons from April to July). Operational breaks of the crane resulting from change of staff prevented observations between 1200 h and 1300 h. During each observation period, the whole plot was crossed and inspected at least

twice. Observation from the tower crane was done from heights ranging from 3-32 m above ground level. Gaps were used to get near ground level. We recorded the position (specific tree, approximate height, perch diameter) of each spotted lizard and noted lizard activity for each sighting (moving or stationary; when mov- ing: foraging, interacting with other individuals, other activities). Cloud coverage was estimated visually in increments of one-eighth of full sky. Data were recorded on microcassette tapes and transcribed later. Further data concerning height and taxonomic status of trees with diameter at breast height (dbh) > 10 cm within the study plot were put to our disposal by J. Wesenberg, D. Sattler, and S. Kirmse (unpubl. data, see Ac- knowledgments). For all observations exceeding 5 min (N = 59), lizard activity was quantified as the percentage of time spent moving to total observation time. For the analysis of foraging behavior, we excluded observations of basking individuals or of specimens involved in social activities.

In 1998, branches that were frequently used by lizards or located near to holes in which liz- ards were observed previously, were encircled with nonpoisonous glue-board traps (Catch- master Maxcatcho Giant Rat Board). Twenty- eight traps were exposed on 11 days for 170.5 h, yielding 14 captures, two of which were re- captures. The lizards were removed from the boards with vegetable oil. Body mass was mea- sured with a spring balance (0.5 g). Snout-vent length (SVL) and tail length (TL) were mea- sured with a ruler tail width (TW), head width (HW), head height (HH), head length (HL), body width (BW), body height (BH), and length of hind limbs (HLL) with a caliper (see Avila- Pires, 1995). Number and position of ectopara- sites were recorded. Five individuals were killed immediately after capture. Seven individ- uals were given a unique paint code with nail polish which allowed individual recognition from a distance until shedding occurred (usu- ally 1-2 months) and released next to the site of capture. Two of them were killed after recap- ture. Prior to fixation in isopropyl alcohol (7O0lO) lizards were dissected and sexed, and stomach conteiit was analyzed. The fixed specimens were deposited in the Colecion Herpetologica del Museo de Ciencias, Caracas, and in the Mu- seo Historia Natural La Salle, Caracas.

Climatic Data.-We used visual estimations of cloud coverage and measurements of air tem- perature and of irradiation (PAR = Photosyn- thetically Active Radiation) for correlations of activity and weather. Climatic data were taken by a team of climatologists of the University of Mannheim, Germany (Anhuf et al., 1999). Air temperature ("C) was measured by a thermo-

HABITAT USE AND ACTIVITY OF TROPIDURID LIZARDS 397

TABLE 1. Morphometric characters of Tropidurus azuieus zuerneri. Snout-vent length (SVL), tail length (TL), tail width (TW), head length (HL), head width (HW), head height (HH), body width (BW), body height (BH), and hind-limb length (HLL) are in millimeters, body mass in grams.

Male (N = 5 )

Character Mean -t SE

SVL Body mass TL TW HL HW HH BW BH HLL

Range

60.1-76.0 5.0-12.0

34.341.6 8.1-11.7

15.8-19.6 9.7-15.5 7.2-19.7

10.0-17.6 6.7-11.0

39.648.1

Male subadult (N = 1)

31.0 1.0

17.0 3.5 8.9 5.7 4.3 5.8 4.0

21.7

Female (N = 1)

72.0 8.5

36.0 9.8

18.1 12.0 8.3

17.3 12.6 37.0

Unsexed (N = 5)

Mean i SE

63.1 ? 10.73 6.7 ? 2.79

35.9 ? 7.55 8.5 ? 1.02

14.6 i 3.95 10.2 It 1.51 8.5 i 1.24

12.5 t 2.96 10.7 i 3.04 36.6 i 5.61

Range

53.4-80.0 4.0-11.4

28.747.0 6.9-9.7 8.2-20.0 8.2-12.2 7.1-10.4 9.4-17.6 7.0-14.6

29.944.4

sensor fixed in the canopy of a tree (21 m height) within the study area. PAR was record- ed at a height of 41 m as photosynthetic-photon- flux density (pmolm-2s-1) by a sensor located on the crane. Data were collected continuously and stored on 21X datalogger (Campell Scien- tific, U.K.), computing 10-min averages from 30- sec sampling intervals. We abstained from mea- suring the body temperature of captured indi- 1-iduals because we were not able to estimate the amount of time a lizard had to spent on a ylueboard trap (maximally 2 h), thus being un- able to thermoregulate. We restricted the corre- lation of activity and weather to observations of 1998 because climatic data of 1997 were frag- rnentary.

Statistical Puocedures.-We regarded all obser- -.-ations as independent samples. We are aware ~f the risk of pseudoreplication, but we usually 5anged our position after we had lost visual -2ntact of an observed individual. Because of - . -.,- ---, a minimum of 2 h passed between two ob- *X-ations of possibly the Same individual, and recause we observed activities rather than in- 5::-iduals, the presumption of independence is i~missible . Means are given -C SE. We used - - - * . . I- ::\sis of variance when assumptions of ho- - :<eneity of variances (~max-tesf; Fowler and 1 : -en, 1990) and normal distribution were met '.-11-Klassen test; Zofel, 1992). Otherwise we +.X¿ nonparametric tests (Mann-Whitney U- P.-- The horizontal distribution of lizards was

r?uted using Lloyd's mean crowding mdex - :. d, 1967; Muhlenberg, 1976):

rn, = rn + (u2/" - 1)

11 Y -*I = mean crowding index, rn = mean T LI- -er of individuals per sample area of size -. = lariance of number of individuals per .I-i -. area of size z. An index r higher than 1

indicates nonrandom distribution. For all corre- lations, we used Spearman Rank Correlations.

Morpkology and Parasifes.-Snout-vent length in T azureus werneri ranged from 31-80 mm. Of the sexed individuals, the largest of six males measured 76 mm, and the only female had a SVL of 72 mm. The body was moderately flat- tened (Table 1). The dorsal part of the body was bright green with a faint reticulate or irregular pattern. Males and females did not differ in body coloration.

All handled lizards carried mites in neck folds or in the inquinal region of the hind legs (12.14 -C 9.08 mites per lizard, N = 14). One specimen was parasitized on the venter by two ticks. Number and position of parasites changed with time on the two retrapped specimens. Number of parasites and SVL were strongly correlated (Spearman Rank Correlation: N = 14, r = 0.815; P < 0.001).

Activity.-Mean number and duration of ob- servations were distinctly lower during the morning (0.43 lizards/h; 5 min/lizard) than during the afternoon (0.85 lizards/h; 13 min/ lizard). Number of observed lizards per obser- vation period increased with temperature in the aftemoon (N = 24, r = 0.582, P = 0.001, Fig. 1). We did not observe activity higher than 15% be- fore midday (Fig. 2). Activity rose significantly (Mann-Whitney U-test, NI = 14, N2 = 45, U = 181.3, P = 0.018) when air temperature mea- sured at a height of 21 m exceeded 28°C (usu- ally around 12 h). The lizards showed high var- iability in afternoon activity, ranging from rest- ing in a hole or in shade and basking to foraging and interacting with other individuals. Activity was negatively correlated with cloud Cover (Spearman Rank Correlation: N = 50, r = -0.424, P = 0.002), and positively correlated

398 N. ELLINGER ET AL.

Mean air temperature ("C)

FIG. 1. Frequency of lizard sightings per afternoon and mean air temperature in the afternoon. Solid straight line: regression line; solid curved lines: con- fidence intervals for 95% of cases; dashed lines: pre- diction intervals.

with PAR (Spearman Rank Correlation: N = 50, r = 0.368, P = 0.005).

Habitat Distribution.-Tropidurus azureus wer- neri is by far the most abundant lizard species in the canopy of the study area. During the Course of the study, 167 observations of T a. wer- neri were made. The species T. (P.) plica (Tro- piduridae) was occasionally encountered sitting in head-down position at lower parts of tree trunks but never in the canopy. The only other canopy-dwelling lizards we discovered were a single specimen of Polychrus marmoratus (Poly- chrotidae) and one unidentified anole.

The horizontal distribution of 'I: a. werneri within the study plot was clustered (mean crowding index: r = 2.14; i = 20 m). The pat- terns of spatial distribution were remarkably similar in 1997 and 1998 (Fig. 3).

FIG. 3. Map of the study plot, showing the trees in which specimens of Tvopidurus azureus zuerneri were spotted in 1997, in 1998, and in both years.

The lizards preferred the highest trees in the plot (Mann-Whitney U-test, NI = 294, N, = 17, U = 759.0, P < 0.001). They were strictly arbo- real, and the majority of observations (43%) were made at heights between 18 and 22 m (Fig. 4), within a few meters below the top of the canopy. Mean height of trees on which lizards

O b C O N W < D Time of day ,- v ~ ~ ~ ~ A PI

F

FIG. 2. Mean air temperature of the observation days, measured at a height of 21 m (squares), and Height in meters mean activity (bars) of Tropiduvus azuveus werneri. Non overlapping error bars (f-corrected SE) indicate sig- FIG. 4. Height preference of Tvopidurus azureus wer- nificant differences (P < 0.05). nevi.

HABITAT USE A N D ACTIVITY OF TROPIDURID LIZARDS

Perch diameter (cm)

FIG. 5. Perch diameter used by Tropidurus azureus wernwi.

were spotted was 25.3 i 5.1 m , and mean height o f trees wi th dbh > 10 c m i n the plot was 18.5 +- 5.8 m. The highest tree i n the plot measured 33 m . The lizards were mainly sight- ed o n rather small branches w i th diameters be- tween 1 c m and 25 c m and were rarely observed on large trunks (Fig. 5). Tropidurus azureus wer- neri showed an af f ini ty for two tree species, Gou- pia glabra (Celastraceae) and Qualea trichantkera (Vochysiaceae; chi-square test: d f = 2, xZ = 23.78). The percentage o f these two species among tree species w i t h d b h >10 c m was 19% and 4%, respectively. However, 35% o f the liz- ard sites were located on G. glabra, and 20% on Q. frichanthera. For trees higher than 21 m, 47% o f lizard sites were located o n G. glabra and Q. trichantkera, which accounted for only 29% o f tall trees. Dead trees without foliage were avoided or visited only for a few minutes during phases o f very active foraging (N = 2). Tree holes, several drilled b y larvae o f cerambycid beetles, were used as shelters during the night. The length o f the holes varied between equal t o the body length o f lizards and more than 70 Cm.

Mean length o f holes was 29 2 24.3 c m ( N = 7). The diameter at the entrance varied between 1.2 and 1.8 Cm. Mean diameter at the entrance was 1.5 -t 0.25cm ( N = 7). Lizards usually could not turn i n the hole and entered it head first and left it tail first. Marked lizards were encountered regularly within the Same crown or group o f adjacent crowns and i n a specific hole for at least a two-month period. The lizards typically cxcurred singly or t o a lesser extent i n pairs 18.9% o f observations). W h e n two lizards were observed coincidently at the Same site, they usu- ally showed either aggressive or mating behav- 1or.

Foraging Behmior und Prey.-When foraging,

TABLE 2. Diet Summary for six specimens of Tm- pidurus azureus werneri. Frequency is the number of lizards containing each prey type.

Fre- Prey N % N quency

Formicidae Pkeidole sp. Zacryptocerus sp. Pseudomyrrnex sp. Camponotus sp. Cephalotes sp. Solenopsis sp. Azteca sp. Dolickoderus sp. Other Myrmicinae Other Formicinae

Formicidae total

Ant pupae Ant eggs

Other Hymenoptera Homoptera Isoptera SUM

lizards searched the branches and twigs o f the canopy for insects. Typical foraging behavior in- cluded running sequences o f a few centimeters u p to several body lengths, and short stops wi th scanning head movements. Foraging behavior coincided w i th phases o f high activity. Foraging lizards spent 34.6 +- 19.7 percent o f their time i n movement for all observations o f foraging liz- ards exceeding 5 m i n ( N = 39), whereas for ob- servations wi th cloud cover less than half o f full sky ( N = 22) the value was 45.2 i 11.4 %. While foraging, the lizards occasionally cruised sev- eral adjacent crowns and were observed t o cov- er horizontal and vertical distances u p to 12-15 m within less than 30 min. The lizards' remark- able locomotory abilities included jumps u p t o a distance o f 1.2 m when leaping f rom a higher branch to a lower one. Jump distance between branches o f Same height was u p to 0.5 m. In addition t o foraging on the bark, the lizards at- tempted t o capture flying insects. The stomachs o f the dissected individuals contained workers o f eight ant genera, as well as remains o f other insect orders (Table 2). Ant pupae were found i n two individuals, and one o f these also con- tained ant eggs. In term o f numbers, Formicidae represented the most important food i tem (90.9%).

Activity depended on air temperature and in- solation. The negative correlation o f activity wi th cloud cover suggests that sunlight may be important for attainment and maintenance o f

400 N. ELLINGER ES AL.

activity body temperatures. A typical day of ac- tivity can be summarized by basking in the morning, followed by social interactions and foraging later in the day. In late afternoon, liz- ards again basked. On days with intermittent clouds, activity began when sun was available and persisted unless cloud Cover became dense or rain began.

Tree structure rather than its taxonomic status is relevant for the spatial distribution of T a. zuer- neri within the canopy. The presence of suitable tree holes may affect the horizontal distribution and cause the clustered occurence. Two aban- doned holes were repopulated in less than two weeks, indicating their probable restricted avail- ability. Eggs of T a. werneri have only been found on the ground (Mägdefrau, 1991). In con- trast, T flauiceps is known to deposit clutches in tree cavities (Dixon and Soini, 1975). The di- mensions of the holes used by T a. werneri in our study area are small compared to holes used by T flaviceps (Vitt and Zani, 1996a). The small dimensions of the holes suggest that, for T a. werneri, the main significance of tree holes is to provide protection against nocturnal pred- ators. Nocturnal mammals feeding on small vertebrates, such as kinkajous (Potos flauus, Pro- cyonidae) and wooliy opossums (Caluromys sp., Didelphidae), as well as arboreal snakes (Coral- lus caninus, Boidae; Leptophis ahaetulla, Colubri- dae, Oxybelis viridis, Colubridae) were encoun- tered several times within the study area (pers. obs.; C. Kirmse, unpubl. data). The narrowness of the holes restricts predator access, and the spiny tail may block predators (cf. Cooper et al., 1999). Nevertheless, we cannot exclude the pos- sibility that T a. wrner i may use tree holes as nest sites when cavities with adequate dimen- sions are available. The lizards show a preference for high and well- structured tree crowns, as offered by G. glabra and Q. trichanthem. Trees with basking sites, but also dense foliage for hiding from visually ori- ented predators, appeared to be favored by the lizards. Birds are notable predators of 7: a. wer- neri. A. Strunz (unpubl. data) reported preda- tion of T a. zuerneri by a white-necked puffbird (Notharchus rnacrorhjnchus, Bucconidae) and sev- eral captures by a nesting pair of double- toothed kites (Harpagus bidentatus, Accipitridae).

Marked individuals were spotted repeatedly at the same sites. Observations of territorial be- havior such as head bobbing, display, and chas- ing suggest the existence of territorial residents. Tropidurids are known to be territorial (Diaz- Uriarte, 1999). Tropidurus azureus werneri did not occur in groups, as has been found for T flaui- ceps (Vitt and Zani, 1996a). Because sexual di- morphism was not apparent, it is unknown whether territories were exclusively occupied by

males. However, six of seven captured and sexed lizards were males. The only female was trapped together with and next to the hole of a male. Because the traps were positioned in a nonrandomized design but at sites within an al- ready detected territory (e.g., in front of a hole), it may be that territories are mainly occupied by males.

Within the sit-and-wait and active-foraging dichotomy (Huey and Pianka, 1981), tropidurid lizards are usually classified as sit-and-wait for- agers (Cooper, 1994; Colli et al., 1997). Tropidu- rus aztlreus werneri shares a suite of characters usually associated with a sit-and-wait foraging mode (Huey and Pianka, 1981; Cooper, 1994; Pough et al., 1998): feeding on mobile prey, use of visual prey discrimination, cryptic coloration, and stout body shape. Foraging modes of liz- ards are associated with lineage and are phy- logenetically quite stable (Cooper, 1995). How- ever, substantial variation occurs in the degree and pattern of movements within the two major foraging modes (McLaughlin, 1989; Cooper and Whiting, 1999; Perry, 1999). Our data indicate that 7: U . werneri spends more time in motion than T hispidus in rock outcrops (Vitt et al., 1996), T torquatus in a restinga habitat (Bergallo and Rocha, 1994), and 7: plica on large tree trunks near the ground within lowland rain for- est (Vitt, 1991). Besides that, T a. werneri are at the most mobile end of the activity spectrum of Iguania (Perry, 1999). The likely shift to more active forage patterns than found in other tro- pidurids emphasizes the potential for plasticity in lizard foraging behavior and may reflect the influence of microhabitat structure on specific foraging behavior. Narrow surface zones on branches can be scanned visually more quickly than wide surfaces on trunks or on the ground. Beyond that, the small surveyable area within a highly structured environment requires fre- quent repositioning moves to renew part or all of the scanning area (Moermond, 1979). Thus, the small diameter of perches used by 7: a. wer- neri and the small field of vision within the structurally complex canopy may result in a shift to a more active foraging mode. This is most obvious when the foraging behavior of T a. werneri is set against the foraging patterns of T plica, which is undoubtedly a typical sit-and- wait forager (Vitt, 1991; pers. obs.). Tvopidurus plica forages on the smooth trunks of large trees, often sitting vertically in a head-down posture. A lizard perching in such a position can scan a large area (Moermond, 1979).

Lizards feeding mainly on ants comprise only a small portion of arthropod-eating lizards (Caldwell and Vitt, 1999). However, ants are the basic food source for T a. werneri. It is not clear whether the ant pupae and eggs in the diet of

HABITAT USE AND ACTIVITY OF TROPIDURID LIZARDS

the lizards derive from active breaking into ant nests. If so, this would underline the potential shift to a more active foraging mode, because sit-and-wait foragers usually do not dig prey from refuges such as nests (Vitt and Zani, 1996a). Ants appear to be important in the diets of most species in the Tropiduridae (Vitt, 1996), but only I: plica, T. flauiceps, and T umbra are known to feed on ants in proportions as high as T a. werneri (Vitt, 1991; Vitt and Zani, 1996a; Vitt et al., 1997). Because all of these species live in large trees in Amazonian forest, the dominance of Formicidae in the diet of T a. werneri might result from the high abundance of ants in low- land forest canopies (Tobin, 1995). Nevertheless, other arboreal lizards in lowland rain forests eat fewer ants (Vitt and Carvalho, 1992; Vitt and Zani, 199610; Vitt et al., 1997). The diet may also reflect an evolutionary history of prey choice (Caldwell and Vitt, 1999; Vitt et al. 1999). Within tropidurid lizards, the Tropidurus clade evolved toward formicivory (Vitt, 1996), and the fact that even basal taxa feed on ants may form the base for the success of the most derived arboreal spe- cies of tropidurids in Amazonian forest (Vitt et al., 1997).

In Summary, T. a. werneri can be classified as an arboreal lizard of Neotropical lowland rain forest that prefers the branches and twigs of the canopy and mainly feeds on ants. Its ecological characteristics are similar to that of other tro- pidurid lizards, except the probable shift to more active foraging patterns.

Acknm1edgments.-The study was funded by the Austrian Science Foundation (FWFProject P- 11565BI0, project leader W. Hödl, University of Lienna). Fieldwork was carried out using the to~ver crane erected by the Austrian Academy of Science (AAS), within the Austrian-Venezue- :an Cooperation. Department PROFAUNA of the \linisterio del Ambiente y de los Recursos Na- xrales Renovables (MARNR) in Caracas, Ven- ?zuela, provided the permission to collect spec- :Yens (collecting permit 15.0157). We sincerely 'ank R. Rollenbeck and B. Schröder, University ?f 'tlannheim, for supporting climate data and 2-e topography of the canopy as basis for Figure 1 md J. Wesenberg, D. Sattler, and S. Kirmse, . . -51-ersity of Leipzig, for putting results of their --ctanical studies at our dispmal and for collab- .'rzSon in the field. N. Blüthgen, University of ?$:in, was very helpful in the determination of -:-e stomach contents of dissected lizards. We r it syateful to J. P. Caldwell, M. S. Hoogmoed, .:-I= L. J. Vitt for their careful review and con-

-&tive comments on the manuscript. R. Jehle, -. - : =der, B. Gottsberger, and E. Gruber made ..-CA suggestions during the preparation of the -

- L - ~ c r i p t .

ANHUF, D., T. MOTZER, R. ROLLENBECK, AND J. SZAR- ZYNSKI. 1999. Water budget of the Surumoni-cra- nesite. Selbyana 20:179-185.

AVILA-PIRES, T. C. S. 1995. Lizards of Brazilian Ama- zonia (Reptilia: Squamata). Zool. Verh. Leiden 299: 1-706.

BERGALLO, H. G., AND C. F. D. ROCHA. 1994. Spatial and trophic niche differentiation in two lizards (Tvopiduvus torquatus and Cmidophovus ocellifev) with different foraging tactics. Austr. J. Ecol. 19:72- 75.

CALDWELL, J. P., AND L. J. VITT. 1999. Dietary asym- metry in leaf litter frogs and lizards in a transi- tional northern Amazonian rain forest. Oikos 84: 383-397.

COLLI, G. R., A. K. PERES, AND M. G. ZATZ. 1997. For- aging mode and reproductive seasonality in trop- ical lizards. J. Herpetol. 31:490499.

COOPER, W. E. 1994. Prey chemical discrimination, foraging mode and phylogeny. In L. J. Vitt and E. R. Pianka (eds.), Lizard Ecology: Historical and Experimental Perspectives, pp. 95-116. Princeton Univ. Press, Princeton, NJ.

. 1995. Foraging mode, prey chemical discrim- ination, and phylogeny in lizards. Anim. Behav. 50: 973-985.

COOPER, W. E., AND M. J. WHITING. 1999. Foraging modes in lacertid lizards from southern Africa. Amphib.-Reptilia 20:299-311.

COOPER, W. E., J. H VAN WYK, AND P. L. F. N. MOUTON. 1999. Incompletely protective refuges: selection and associated defences by a lizard, Covdylus cov- dylus (Squamata: Cordylidae). Ethology 105:687- 700.

COTT, H. B. 1926. Observations on the life habits of some batrachians and reptiles from the lower Am- azon and a note on some mammals from Marajo Island. Proc. Zool Soc. Lond. 1926:1159-1178.

CRUZ, F. B. 1998. Natural history of Tropiduvus spi- nulosus (Squamata, Tropiduridae) from the dry chaco of Salta, Argentina. Herpetol. J. 8:107-110.

D~Az-UMARTE, R. 1999. Anti-predator behaviour changes following an aggressive encounter in the lizard Tropidurus hispidus. Proc. R. Soc. Lond. B Biol. Sci. 266:2457-2464.

DIXON, J. R., AND P. SOINI. 1975. The reptiles of the upper Amazon basin, Iquitos region, Peru. I. Liz- ards and amphisbaenians. Milwaukee Publ. Mus. Contr. Biol. Geol. 4:1-58.

DUELLMAN, W. E. 1978. The biology of an equatorial herpetofauna in Amazonian Ecuador. Misc. Publ. Univ. Kans. Mus. Nat. Hist. 65:l-352.

FOWLER, J., AND L. COHEN. 1990. Practical Statistics for Field Biology. Open Univ. Press, Buckingham, U.K.

FROST, D. R. 1992. Phylogenetic analysis and taxon- omy of the Tropiduruc group of lizards (Iguania: Tropiduridae). Am. Mus. Novit. 3033:l-68.

GASNIER, T. R., W. E. MAGNUSSON, AND A. P. LIMA. 1994. Foraging activity and diet of four sympatric lizard species in a tropical rainforest. J. Herpetol. 28:187-192.

402 N. ELLINGER ET AL.

GREENE, H. W. 1977. Lizards of the genus Uracentron Kaup in Eastern Ecuador. Herpetologica 24:253- 255.

HARVEY, M. B., AND R. L. GUTBERLET. 1998. Lizards of the genus Tropidurus (Iguania: Tropiduridae) from the Serrania de Huanchaca, Bolivia: new spe- cies, natural history, and a key to the genus. Her- petologica 54:493-520.

HOOGMOED, M. S. 1973. Notes on the Herpetofauna of Suriname N. The Lizards and Amphisbaenians of Surinam. D. Junk, The Hague, The Netherlands.

HOWLAND, J. H., L. J. VITT, AND P. T. LOFEZ. 1990. Life on the edge: the ecology and life history of the tropidurine iguanid lizard Uranoscodon supevcilios- um. Can. J. Zool. 68:1366-1373.

HUEY, R. B., AND E. R. PIANKA. 1981. Ecological con- sequences of foraging mode. Ecology 62:991-999.

LLOYD, M. 1967. Mean crowding. J. Anim. Ecol. 36: 1-30.

MÄGDEFRAU, H. 1991. Zu den Tafelbergen Venezue- las. Zusammenfassungen. Jahrestagung der Deutschen Gesellschaft für Herpetologie und Ter- rarienkunde, 18-22 Sept. 1991, Bonn:9-10.

MCLAUGHLIN, R. L. 1989. Search modes of birds and lizards: evidence for alternative movement pat- terns. Am. Nat. 133:654-670.

MOERMOND, T. C. 1979. The influence of habitat struc- ture on Anolis foraging behavior. Behavior 70:147- 167.

M~HLENBERG, M. 1976. Freilandökologie. UTB 565. Verlag Quelle & Meyer, Stuttgart, Germany.

PERRY, G. 1999. The evolution of search modes: Eco- logical versus phylogenetical perspectives. Am. Nat. 153:98-109.

POUGH, F. H., R. M. ANDREWS, J. E. CADLE, M. L. CRUMP, A. H. SAVITZKY, AND K. D. WELLS. 1998. Herpetology. Prentice Hall Inc., Englewood Cliffs, NJ.

REAGAN, D. P. 1995. Lizard ecology in the canopy of an island rain forest. In M. D. Lowman and N. M.

Nadkarni (eds.), Forest Canopies, pp. 149-164. Ac- ademic Press, San Diego, CA.

TOBIN, J. E. 1995. Ecology and diversity of tropical forest canopy ants. In M. D. Lowman and N. M. Nadkarni (eds.), Forest Canopies, pp. 129-148. Ac- ademic Press, San Diego, CA.

VITT, L. J. 1991. Ecology and life history of the scan- sorial arboreal lizard Plica plica (Iguanidae) in Am- azonian Brazil. Can. J. Zool. 69:504-511.

. 1993. EcoIogy of isolated open formation Tro- pidurus (Reptilia: Tropiduridae) in Amazonian lowland rain forest. Can. J. Zool. 71:2370-2390.

. 1996: Biodiversity of Amazonian lizards. Occ. Publ. Mildred E. Mathias Bot. Gard. 1:89-108.

VITT, L. J., AND C. M. CARVALHO. 1992. Life in the trees: the ecology and life history of Kentropyx striatus (Teiidae) in the Lavrado area of Roraima, Brazil, with comments on the life histories of trop- ical teiid lizards. Can. J. Zool. 70:1995-2006.

VITT, L. J., AND A. ZANI. 1996a. Ecology of the eh- sive tropical lizard Tropidurus (= Uracentvon) flmi- ceps in lowland rainforest of Ecuador. Herpetolo- gica 52:121-132.

. 199613. Organization of a taxonimically di- verse lizard assemblage in Amazonian Ecuador. Can. J. Zool. 74:1313-1335.

VITT, L. J., P. A. ZANI, AND J. CALDWELL. 1996. Be- havioural ecology of Tropidurus hispidus on isolated rock outcrops in Amazonia. J. Trop. Ecol. 12:81- 101.

VITT, L. J., P. A. ZANI, AND T. C. C. AVILA-PIRES. 1997. Ecology of the arboreal tropidurid lizard T~opidu- rus (- .~lica) umbra in the Amazon region. an. J. Zool. 75:1876-1882.

VITT, L. J., P. A. ZANI, AND M. C. ESPOSITO. 1999. His- torical ecology of Amazonian lizards: implications for community ecology. Oikos 87:286-294.

ZANI, P. A., AND L. J. VITT. 1995. Techniques for cap- turing arboreal lizards. Herpetol. Rev. 26:136-137.

ZOFEL, P. 1992. Statistik in der Praxis. Gustav Fischer Verlag, Stuttgart, Germany.

Accepted: 2 October 2000.