Trophic ecology and foraging behavior of Tropidurus hispidus and Tropidurus semitaeniatus ( Squamata , Tropiduridae ) in a caatinga area of northeastern Brazil

This study aimed to analyze the seasonal variation in diet composition and foraging behavior of Tropidurus hispidus (Spix, 1825) and T. semitaeniatus (Spix, 1825), as well as measurement of the foraging intensity (number of moves, time spent stationary, distance traveled and number of attacks on prey items) in a caatinga patch on the state of Rio Grande do Norte, Brazil. Hymenoptera/Formicidae and Isoptera predominated in the diet of both species during the dry season. Opportunistic predation on lepidopteran larvae, coleopteran larvae and adults, and orthopteran nymphs and adults occurred in the wet season; however, hymenopterans/Formicidae were the most important prey items. The number of food items was similar between lizard species in both seasons; however the overlap for number of prey was smaller in the wet season. Preys ingested by T. hispidus during the wet season were also larger than those consumed by T. semitaeniatus. Seasonal comparisons of foraging intensity between the two species differed, mainly in the wet season, when T. hispidus exhibited less movement and fewer attacks on prey, and more time spent stationary if compared to T. semitaeniatus. Although both lizards are sit-and-wait foragers, T. semitaeniatus is more active than T. hispidus. The diet and foraging behavior of T. hispidus and T. semitaeniatus overlap under limiting conditions during the dry season, and are segregative factors that may contribute to the coexistence of these species in the wet season.

The ecological and behavioral interactions of lizards are directly related to their foraging mode (Huey & Pianka, 1981;RocHa, 1994;WeRneck et al., 2009).In a broader context, factors related to lizards' trophic ecology include prey availability in the environment and selectivity by the forager (VRcibRadic & RocHa, 1995;RocHa & anjos, 2007).Diets are therefore an important and dynamic component of the interaction of a lizard with its environment and with other co-existing species (duffield & bull, 1998).Lizard species may differ in their diets as competition-reducing mechanisms (dunHam, 1983;Vitt & Zani, 1996), as a result of morphological differences (colli et al., 1992; mesquita et al., 2006), as well as related to different availability of food types (stamPs et al., 1981;duRtscHe, 1995;sousa & cRuZ, 2008).An additional source of variation in diet may result from different foraging behaviors, especially among phylogenetically related species (magnusson et al., 1985; duffield & bull, 1998).
The tropidurid lizards Tropidurus hispidus (Spix, 1825) and T. semitaeniatus (Spix, 1825) are found in sympatry in rocky outcrops in the caatinga of northeastern Brazil.Studies on the diet of these species were previously conducted in one caatinga area in the state of Pernambuco (Vitt, 1995), and on T. hispidus in the Amazon Forest (Vitt et al., 1996) and rocky meadows in the state of Minas Gerais (Van sluys et al., 2004).Ants, insect larvae, termites and beetles were the most frequent items in the diet of both species.In relation to behavior, the only study addressing foraging was carried out in the Amazon region of Brazil for T. hispidus (Vitt et al., 1996).In addition, the role of T. semitaeniatus as a seed disperser (RibeiRo et al., 2008), drinking and thermoregulatory behaviors (RibeiRo & fReiRe, 2009a;RibeiRo & fReiRe, 2010), as well as the reproductive cycle of this species (RibeiRo et al., 2010) were described in the caatinga.
Given that tropidurid lizards are traditionally considered sit-and-wait foragers (scHoeneR, 1971;RocHa, 1994), our initial prediction was that T. hispidus and T. semitaeniatus would have generalized diets, and that preys would be captured according to their availability in the environment.Furthermore, we expected little variation between the foraging behaviors of these species.In this study, we analyse the diet composition and foraging behavior of T. hispidus and T. semitaeniatus in an area of the Caatinga Biome in northeast Brazil.

MATERIAL AND METHODS
Study area.The study was conducted at the Seridó Ecological Station (ESEC Seridó,06°34'36.2"S,37°15'20.7"W,datum: WGS84;altitude: 192 m), which encompasses a Caatinga area of 1,166.38 hectares located in the municipality of Serra Negra do Norte, Rio Grande do Norte, Brazil.The climate is semiarid (ab'sábeR, 1974), with a short wet season between March and May and rainfall ranging between 500 and 700 mm/year.Mean annual temperatures vary from 28ºC to 30ºC, but can exceed 40ºC; the minimum ranges between 17ºC and 20ºC.Relative air humidity oscillates around 30-50% in the dry season, but reaches 80-90% in the wet season (nimeR, 1972).The vegetation of the study area is arboreal-bushy hyperxerophilous (VaRela-fReiRe, 2002).Amidst this vegetation, various rocky extrusions are found that remain covered by the vegetation during the wet season, resulting in partially shaded areas.During the dry season these rocky outcrops become exposed after leaf fall (Velloso et al., 2002).
Trophic ecology.Fieldwork was conducted on monthly excursions, each lasting three consecutive days, from October 2007 to May 2008.The dry season was from October 2007 through January 2008, and the wet season lasted from February to May 2008.Although the wet season is concentrated between March and May, rainfall is irregular in the semiarid region.Indeed, February 2008 was considered rainy at ESEC Seridó due to recorded rainfall of about 90 mm.Mean monthly rainfall during the study period was 8.0 mm for the dry season and 236.7 mm for the wet season.
Lizards were captured on five rocky outcrops between 07:00 and 18:00 h, using air rifles.At the end of each collection day, the animals were dissected for stomach removal and these were stored in 70% alcohol for subsequent analysis.Animal items were identified to order level, except for hymenopterans, for which the family Formicidae was identified.For Coleoptera and Lepidoptera, the development stage of the individual (larva or adult) was also considered.Arthropod fragments were categorized as arthropod remains.Plant items were classified into the following categories: leaves, flowers and seeds; those that could not be identified, were categorized as nonidentified.
Monthly samplings of arthropods were conducted at the same time as lizard collections, to determine their availability (possible food item;stRüsmann et al., 1984;VRcibRadic & RocHa, 1995;RocHa & anjos, 2007).Thirty traps consisting of plastic trays (50.0 x 25.0 x 5.5 cm) containing water and soap were used.They were placed 2 m apart, covering the rocky surface and leaf litter around the rocky outcrops that make up the study area.The trap system in the leaf litter was set up, buried in the soil, analogous to pitfall traps.Trays were exposed from 07:00 to 17:00 h and the arthropods that fell into the traps were collected.Ten traps consisting of cardboard boxes (5.0 x 5.0 x 5.5 cm) containing canned sardines in oil were also used daily during one hour of exposure for ant collection.All arthropods found in the traps were fixed in 70% alcohol for later identification.
The following measurements were taken for all lizards collected using a manual caliper (to the nearest 0.1 mm): snout-vent length (SVL); head length (HL); jaw length (JL) and head width (HW).Classification of T. semitaeniatus individuals into adult and juvenile categories considered females with SVL ≥ 58 mm and males with SVL ≥ 64 mm (Vitt, 1995).For T. hispidus, size ranges were defined according to RibeiRo & fReiRe (2009b), adult females had SVL ≥ 65 mm and adult males had SVL ≥ 68 mm.All lizards were deposited in the Herpetological Collection of the Departamento de Botânica, Ecologia e Zoologia (CHBEZ), Universidade Federal do Rio Grande do Norte.
The frequency of occurrence (number of lizards) and the number and volume of prey were estimated for each prey type.The volume of prey was estimated using the ellipsoid formula (e.g., Vitt et al., 1996): V = 4/3π (length/2) x (width/2) 2 .The index of relative importance (I) was calculated for each food category, adding occurrence, numerical and volumetric percentages and dividing by three (mesquita et al., 2006).
The overlap in the number of food items in the diets of T. hispidus and T. semitaeniatus was determined using Pianka's Symmetric Overlap Index (O jk ) (Pianka, 1973).Values approaching zero indicate no similarity in diets, whereas values near one indicate diet overlap.Niche overlap was calculated using Ecological Methodology software (kenney & kRebs, 2000).
A t-test was used to evaluate intra and interspecific differences in SVL of males and females.Differences in the mean number and volume of food items found in the stomachs of T. hispidus and T. semitaeniatus in each season, and in the diets of each species between seasons, were evaluated using the Mann-Whitney U test (ZaR, 1999).It was also applied to ascertain whether the mean size of food items ingested by T. hispidus and T. semitaeniatus differed in general, as well as in each season and between sexes of each species.The same test was applied to assess intraspecific differences between seasons for food item size.In order to analyze the effect of body size on prey size, we tested the relationship between jaw width and mean length of the five largest food items found in each stomach, using Linear Regression Analysis (ZaR, 1999).
The diets of the Tropidurus species were compared with arthropod availability in the environment, by number of prey, using Spearman's correlation (ZaR, 1999).Nonidentified arthropod and plant parts were not considered in the statistical tests and when calculating item sizes.Statistical analyses were conducted using SPSS 13.0 and the significance level adopted was 0.05.
Foraging behavior.For behavioral observations, five rocky outcrops surrounded by vegetation and distant 500 m from the lizard collection area were established.Foraging intensity is a measure of predator activity when hunting and capturing prey (RocHa et al., 2000).Behavioral data were collected on three consecutive days, from 07:00 to 10:00 h and from 14:00 to 17:00 h, divided into 6 one-hour intervals to obtain a homogeneous sample.These time intervals were selected based on periods of the highest lizard activity (Vitt, 1995).During these intervals, 10-min focal sampling sessions were carried out (altmann, 1974), alternating the species observed whenever possible, using a voice recorder and digital watch.As each animal was observed only once, at 10-minute intervals, each observation was independent of the others.Observation was initiated 5 minutes after the lizard was sighted to minimize observer interference.Because foraging intensity is expressed mainly in terms of movement, we have applied the following indices to assess the foraging intensity of T. hispidus and T. semitaeniatus: number of moves (NM), time spent stationary (TS, in seconds), distance travelled (DT, in cm; linear distance between the initial location of the lizard to the most distant point reached, using a tape measure) and number of attacks on prey items (NAP).We also discussed the foraging site characteristics and hunting strategies of both lizard species during the dry and wet seasons.
The foraging indices (NM, TS, DT, NAP) were estimated for each lizard using the arithmetic mean by species and season, following methodology adapted from stRüssman et al. (1984), magnusson et al. (1985) and gasnieR et al. (1994).The Mann-Whitney U test (ZaR, 1999) was used to evaluate whether foraging differed between T. hispidus and T. semitaeniatus.Descriptive statistics are represented in the text as mean ± 1 standard error.
The most important animal food items for T. hispidus (Tab.I) during the dry season were ants (45.5%) and isopterans (26.4%).Leaves were the plant items with the highest index of importance (12.4%).In the wet season, the most important prey items were ants (48.1%) and lepidopteran larvae (37.8%), whereas leaves were the most important plant items (32.3%).One adult female T. hispidus consumed an adult of the hylid frog Scinax x-signatus (Spix, 1824) during the wet season.For T. semitaeniatus (Tab.II), in the dry season, the most important food items were isopterans (30.3%) and ants (28.2%); in the wet season, ants (46.1%) and lepidopteran larvae (38.0%).Leaves were the most important plant items in both seasons (Tab.II).
The diets of both lizard species overlapped considerably, especially during the dry season (dry season: O jk = 0.91; wet season: O jk = 0.76).During the dry season, the diets of T. hispidus and T. semitaeniatus did not differ in number (Mann-Whitney U test, z = -1.811,P = 0.074) nor volume of the items consumed (Mann-Whitney U test, z = -1.953,P = 0.052).The number of items consumed by both species was also not different during the wet season (Mann-Whitney U test, z = -1.585,P = 0.113).However, T. hispidus ingested a larger volume of items (Mann-Whitney U test, z = -4.082,P = 0.0001).
The diet of T. hispidus did not differ significantly in number of items (Mann-Whitney U test, z = -0.732,P = 0.479) between the dry and wet seasons, whereas the mean volume of items ingested was significantly higher during the wet season (Mann-Whitney U test, z = -3.181,P = 0.001).The same occurred for T. semitaeniatus in the number of items (Mann-Whitney U test, z = -1.735,P = 0.083) and in mean volume (Mann-Whitney U test, z = -3.822,P = 0.0001) between the dry and wet seasons.
In total, 6,049 arthropod specimens were collected during sampling prey availability in the environment, with 21 prey categories recognized (Figs. 1, 2).There was no significant correlation between the diet composition of T. hispidus and food availability (number of items) in the dry season (r s = 0.339, P = 0.133).During the dry months some arthropod recorded in the environment (e.g., Araneae and Embioptera) were absent in the stomach contents of T. hispidus.However, in the wet season this correlation was significant (r s = 0.032, P = 0.035).For T. semitaeniatus, there was a significant correlation between diet composition and prey availability, in both the dry (r s = 0.473, P = 0.017), and wet seasons (r s = 0.546, P = 0.005).
Foraging Behavior.A total of 122 focal observations were carried out for foraging behavior in the dry and wet seasons of 2006-2007: 38 for T. hispidus and 84 for T. semitaeniatus.In the dry and wet seasons of 2007-2008, there were 68 focal samplings: 25 for T. hispidus and 43 for T. semitaeniatus.During the dry season, foraging intensity did not differ (P > 0.05) between T. hispidus and T. semitaeniatus, in either year (Tab.III).The only exception was that the number of moves was greater for T. semitaeniatus (Tab.III) in 2007.However, during the wet season only the distance travelled was similar for both species (Tab.III).The number of moves and attacks on prey was significantly (P < 0.05) higher for T. semitaeniatus, while time spent stationary was comparatively longer for T. hispidus (Tab.III).

DISCUSSION
Tropidurus hispidus and T. semitaeniatus are consumers of plant matter and various arthropods, although being predominantly insectivores.This diet composition is similar to that of other Tropidurids (colli et al., 1992;Van sluys, 1993;teixeiRa & gioVanelli, 1999;fialHo et al., 2000;faRia & aRaújo, 2004;meiRa et al., 2007).Nevertheless, seasonal variations in diet are expected in species that live in habitats in which local productivity is subject to humidity (janZen & scHoeneR, 1968;ballingeR & ballingeR, 1979) and rainfall cycles (RocHa, 1994;Van sluys, 1995) like the Caatingas (ab'sábeR, 1974).The diet composition of T. hispidus and T. semitaeniatus during the dry season, with ants and termites as the most important items, resembled that recorded for the respective species in another caatinga area (Vitt, 1995), and for T. hispidus in rocky outcrops in the Amazon region (Vitt et al., 1996) and the Atlantic Forest domain (Van sluys et al., 2004).
Sedentary prey items such as termites, unpredictably distributed and grouped in the environment, are considered typical of active foraging lizard diets (Huey & Pianka, 1981;magnusson et al., 1985).However, many lizard species with low movement rates also consumed termites in high frequencies (scHoeneR, 1967;Vitt & caRValHo, 1995;VRcibRadic & RocHa, 1998;fialHo et al., 2000).According to teixeiRa & gioVanelli (1999), the species saves energy by investing in small, numerically abundant prey, such as ants and termites.Furthermore, availability and ease of capture may account for the large amount of these preys in the lizard diet (floyd & jenssen, 1983).The high proportion of ants in the diet of both Tropidurus species seems to be related to their high density in the area and reinforces a well-known pattern of item consumption among tropidurids [e.g.T. spinulosus (Cope, 1862(Cope, ): colli et al., 1992;;T. etheridgei Cei, 1982: cRuZ et al., 1998;T. itambere Rodrigues, 1987: faRia & aRaújo, 2004;T. oreadicus Rodrigues, 1987: RocHa & siqueiRa, 2008;T. torquatus (Wied, 1820): caRValHo et al., 2007].In addition, the high termite consumption by these lizards, as well as characterizing predation on insects with clumped distribution, also identifies them as important sources of water (nagy et al., 1984).
The significant correlation between the diet composition of the two Tropidurus species and arthropod availability indicates that individuals captured prey according to their occurrence in the environment.This was confirmed for other lizard species ( Van sluys, 1993;Vitt, 1993;RocHa, 1996;RocHa & anjos, 2007).In contrast to the T. semitaeniatus, the absence in the diet of T. hispidus of some arthropod taxa recorded in the environment, such as spiders and webspinners (Embioptera), may have favored the non-significance of correlations between prey availability and diet composition of this species during the dry months.
During the wet period, even though ants were predominant in the diet of both lizard species, a considerable volume of lepidopterans, coleopterans and orthopterans was also consumed, characterizing opportunistic predation on arthropods with reproductive cycles concentrated in the rainy months.We can therefore suppose that T. hispidus and T. semitaeniatus consumed fewer of these prey in the dry season owing to their lower availability in the environment.Our data on arthropod availability in the ESEC Seridó support these predictions, since such preys (e.g.larvae and adult lepidopterans, orthopterans and coleopterans) were significantly represented during the wet season.
Both species consumed a considerable volume of plant material, characterising them as omnivorous species, since more than 10% of plant material was ingested (cooPeR & Vitt, 2002).In Exu, another area of caatinga in Pernambuco, plant matter constituted nearly 30% by volume of the diet of T. semitaeniatus and more than 63% of T. hispidus (Vitt, 1995).Tropidurus torquatus is also known for its consumption of plant matter, mainly fruits (fialHo et al., 2000).The fact that plant parts (leaves, flowers, fruits and seeds) account for a sizable portion of the total volume ingested in the diet of Tropidurus species (caRValHo et al., 2007;RocHa & siqueiRa, 2008), suggests that ingestion is not accidental, and that plants are indeed a common food item for these lizards.At ESEC Seridó, in addition to invertebrates and plant material, the stomach contents of an adult female included an adult hylid frog Scinax x-signatus (6,280 mm 3 in volume) (RibeiRo & fReiRe, 2009c).Published diets of T. hispidus include small vertebrates, such as frogs (Vitt et al., 1996) and other lizards (Vitt, 1995;costa et al., 2010) and reinforce the opportunistic habits of these lizards.
Tropidurus hispidus generally feeds on significantly larger prey than T. semitaeniatus, which may reflect their difference in body size.The two species represent opposite extremes in body size, T. semitaeniatus being one of the smallest species of Tropidurus, and T. hispidus the largest species of the genus (RodRigues, 1987).The difference in body size between the syntopic lizards T. spinulosus and T. oreadicus in the cerrado characterized the difference between the sizes of their ingested prey (colli et al., 1992).In contrast, similar sized species T. itambere and T. oreadicus in another cerrado area in central Brazil ingested similar sized prey (faRia & aRaújo, 2004).
Although in T. hispidus significant result was not found in the regression of mean prey size on jaw width, in T. semitaeniatus the prey size varied significantly with jaw width.Significant associations between lizard body size, including head measurements, and prey dimensions have been reported in Tropidurus, as in T. hispidus and T. montanus Rodrigues, 1987(Van sluys et al., 2004), and T. itambere and T. oreadicus (Van sluys, 1993;faRia & aRaújo, 2004).Conversely, for T. torquatus (caRValHo et al., 2007) and for another population of T. oreadicus (RocHa & siqueiRa, 2008) the authors failed to find an effect of lizard size on prey dimensions.Thus, factors other than lizard body size or head measurements may also be important in determining the size of prey consumed by these lizards.
Foraging in Tropidurus reflects a history of sit-and-wait foraging involving a set of associated characteristics in the family Tropiduridae and most of the Iguania (cooPeR, 1994).In spite of the fact that there is a strong phylogenetic effect on lizards' foraging mode (cooPeR, 1995), species can adjust their hunting strategies according to pressures imposed by the environment, such as seasonality (Huey & Pianka, 1981;RocHa, 1994).In this sense, during the dry season, the vegetation of the Caatinga becomes completely dry, reducing foraging sites of T. hispidus and T. semitaeniatus to rocky formations that serve as observation points (waiting sites), which are visited alternately during displacements.The lizards wait at these sites with their heads pointing downward and quickly jump onto potential preys sighted moving through the leaf litter accumulated in the rock crevices or around the rocks.On the other hand, during the wet season, the Caatinga vegetation is invigorated and produces an abundance of flowers, acting as insect baits for lizards that remain motionless near them.Furthermore, when foraging during the wet season, T. hispidus, a habitat generalist, benefits from including the flowers on the trees above the rocky surfaces for foraging.In contrast, T. semitaeniatus, typically saxicolous, does not climb to the highest parts of the trees and limits itself to capturing prey items in the flowers of plants at the level of the rock surfaces.Tropidurus semitaeniatus also uses rapid movements to attack highly mobile prey such as flies (Muscidae) and leaf hoppers (Cicadellidae) when they land on rocky surfaces.This behavior was not observed for T. hispidus.These differences support the idea that foraging intensity may vary between species, even when they exhibit the same foraging mode (gasnieR et al., 1994).
Lizards have traditionally been classified into two categories according to foraging mode: active foragers and sit-and-wait foragers (Vitt & congdon, 1978;Huey & Pianka, 1981).Nevertheless, the generalized diet composition of the two Tropidurus species, consisting of low (termites and insect larvae) and high-mobility (ants and orthopterans) prey, the foraging behavior described here and the indices of foraging intensity, such as the low number of movements and short distances traveled (Huey & Pianka, 1981), are consistent with the categorisation of T. hispidus and T. semitaeniatus as sit-and-wait foragers.This is similar to all other Tropidurids studied (e.g.RocHa & beRgallo, 1990;Vitt, 1991;colli et al., 1992;Vitt & caRValHo, 1995).
Nonetheless, focal observations, especially in the wet season, revealed differences in the foraging intensity between T. hispidus and T. semitaeniatus for most indices evaluated, indicating that T. semitaeniatus is more active than T. hispidus.Furthermore, we found a similar mean number of moves and attacks on prey for T. hispidus in the caatinga, particularly during the wet season, as was observed in a population of T. hispidus studied in the Amazon region (Vitt et al., 1996).
We conclude that the feeding behavior and diet composition of T. hispidus and T. semitaeniatus are affected by the marked rainfall seasonality in the Caatinga.Both species are opportunistic and generalist predators that prey upon arthropods and consume plant material.In summary, the more adverse conditions in the dry season constrain both species to similar hunting strategies, but in the wet season, the different vegetation characteristics and higher food availability result in different hunting strategies.
Observations were made from October to December 2006 (dry season 2006; mean monthly rainfall: 1.6 mm), October to December 2007 (dry season 2007; mean monthly rainfall: 1.2 mm) as part of the dry season, with the wet season represented between April and June 2007 (wet season 2007; mean monthly rainfall: 68.7 mm) and April and June 2008 (wet season 2008; mean monthly rainfall: 141.7 mm).
Tab. III.Measures of foraging intensity during ten-minute focal observations of Tropidurus hispidus (Th) and Tropidurus semitaeniatus (Ts), in the dry (October to December 2006-2007) and wet (April to June 2007-2008) seasons at ESEC Seridó, Serra Negra do Norte municipality, Rio Grande do Norte, Brazil [NM, number of moves; TS, time spent stationary (seconds); DT, distance travelled (cm); NAP, number of attacks on prey items; n, sample size].Data are represented as mean ± 1 SE (range) and P values are based on the Mann-Whitney U test.Significant differences are highlighted with an asterisk.Trophic ecology and foraging behavior of Tropidurus... Iheringia, Série Zoologia, Porto Alegre, 101(3): 225-232, 30 de Setembro de 2011

Figs. 1
Figs.1, 2.  Estimates of arthropod availability in the environment during the dry (white bars) and wet (black bars) seasons at ESEC Seridó, Serra Negra do Norte municipality, Rio Grande do Norte, Brazil: 1, up to 50 specimens collected; 2, more than 50 specimens collected.