Prey availability and diet of maned wolf in Serra da Canastra National Park, southeastern Brazil

Acta Theriologica 2 (4): 391 42, 27. PL ISSN 1 71 Prey availability and diet of maned wolf in Serra da Canastra National Park, southeastern Brazil Diego QUEIROLO and José C. MOTTA-JUNIOR Queirolo D. and Motta-Junior J. C. 27. Prey availability and diet of maned wolf in Serra da Canastra National Park, southeastern Brazil. Acta Theriologica 2: 391 42. Analysis of 4 scats of the maned wolf Chrysocyon brachyurus (Illiger, 181), collected in Serra da Canastra National Park from November 1998 to January 2 yielded 1688 occurrences of 69 identified food items, 49.% of which were represented by plant items and.% by animal items. According to the frequency of occurrence analysis, the most important food item categories were wolf s fruit, Parinari obtusifolia (Chrysobalanaceae), small mammals and birds. When the estimated biomass consumption is considered, the animal items became more important (6.8%) than plant items (43.2%). Small mammals were consumed more in the dry season, while miscellaneous fruits and arthropods were consumed mostly in the wet season. Functional response was found only for the category miscellaneous fruits; however, seasonality analysis revealed that wolf s fruit and small mammals were consumed according to their availability in the area. Prey selection was observed for most of the small mammal species, emphasizing a high level of consumption of Necromys lasiurus, Monodelphis sp. and Gracilinanus sp. The finding that most of the diet was composed of fruits and animal items from open areas of the savannah-like Cerrado led us to suggest that the maned wolf spends most of its time preying in this habitat, indicating a need to establish conservation policies for the Cerrado. Departamento de Ecologia, Instituto de Biociências, Universidade de São Paulo. Rua do Matão 321, travessa 14, São Paulo, SP, 8-9, Brazil, e-mail: diqueirolo@yahoo.com.br (DQ) Key words: maned wolf, biomass, Cerrado, diet, prey selection, seasonality Introduction The maned wolf Chrysocyon brachyurus (Illiger, 181), is a threatened species (Ministério do Meio Ambiente 23) in need for an effective management plan for it s conservation. To have a better knowledge of the species, ecological studies on predator-prey interaction are important, particularly those that focus on feeding habits and prey availability. The diet of the maned wolf was first studied by Dietz (1984). More recently, others studies were carried out in several places throughout central and southeastern Brazil (Motta-Junior et al. 1996, Motta- -Junior 2, Aragona and Setz 21, Rodrigues 22, Bueno et al. 23, Santos et al. 23, Jácomo et al. 24). These studies have shown that the maned wolf feeds on a large range of plant and animal species, and most of its diet is [391]

392 D. Queirolo and J. C. Motta-Junior composed of wolf s fruit Solanum lycocarpum, several species of miscellaneous wild fruits, small mammals, birds and armadillos. Studies including prey availability and carnivore functional responses increase the interpretability of the results obtained (Jaksic et al. 1992, Corley et al. 199). However, for canid species in the Neotropics, limited data are available regarding functional responses and selectivity in the diet (Iriarte et al. 1989, Castro et al. 1994, Corley et al. 199). Particularly for the maned wolf, only two studies have assessed prey selectivity and functional responses (Motta-Junior 2, Bueno and Motta-Junior 26). We examined the diet of the maned wolf taking into account the poorly known aspects of prey selectivity and functional response in savannah habitat, southeastern Brazil, from November 1998 to January 2. Our objectives were to: (1) quantitatively determine the diet of the maned wolf, according to the frequency of occurrence, biomass and individual prey consumed, (2) estimate the abundance of major prey species of the maned wolf in the study area and (3) assess the selection and functional responses of the maned wolf according to the availability of major prey items. Study area Serra da Canastra National Park (SCNP) is located in southeastern Brazil, in the south-west part of the state of MinasGerais(2 2 S, 46 4 W). The total area is 2 ha and includes many small rivers and several springs. Plateaus range from 8 to 1496 m, surrounded by steep slopes forming slightly irregular relief (Instituto Brasileiro de Desenvolvimento Florestal/Fundação Brasileira para a Conservação da Natureza 1981). The climate is characterized by a cold dry season from April through September with a mean temperature of 16.7 C, and a hot rainy season from October through March with a mean temperature of 19. C. The mean annual rainfall ranges from 13 to 17 mm with a peak of rains from December to February. The vegetation is included in the Cerrado phytogeographical province, a tropical savannah found throughout central Brazil, covering approximately 22% of the area of Brazil (Coutinho et al. 22). Our study site was located in the eastern part of the SCNP, where the predominant habitats include open areas composed mainly of grasses and herbaceous plants ( campo limpo and campo umido ), interspaced with scattered or sparsely distributed tree species and shrubs ( campo sujo and campo cerrado ) and areas of high elevation meadows ( campo rupestre ) (Eiten 1994). Gallery forests and patches of forest occur along water bodies and near springs (< 2.% of the park s surface) (Instituto Brasileiro de Desenvolvimento Florestal/Fundação Brasileira para a Conservação da Natureza 1981). Material and methods Sampling and diet analysis Several techniques are used to study mammalian diet, and scat analysis is one of the most common (Putman 1984). Most studies describe scat contents as the frequency of occurrence of each food item in relation to the total number of scats collected, or to the total number of occurrences of all food items (Dietz 1984, Dalponte 1997). The frequency of occurrence is a valid parameter to determine the diet composition and the proportionofeachitemconsumed,becauseitiseasytoapplyandcanbecomparedwithother studies (Corbett 1989). Some studies complement this analysis with an estimate of the minimum number of consumed individuals of each prey item and their biomass, which provides complementary information about the relative importance of each food item in the diet of the species (Motta- -Junior et al. 1996). We collected 4 scat samples from November 1998 to January 2 along roads and trails and identified them by their characteristic odour, size/diameter (> 2 mm of diameter), texture and place of deposition (generally easily discernible, over rocks or on the middle of the road) (Dietz 1984, Jácomo et al. 24). The identification was aided by presence of wolf tracks near the scats and presence of predators hair in the scats (Santos et al. 23, Juarez and Marinho-Filho 22). The only large carnivore living in the SCNP with similar-sized scats is the puma Puma concolor, but its faeces have a different texture, shape (definite segments), odour and hair. All the scats were labelled, washed in water over two fine mesh screens (. and 1. mm), dried in an oven at 6 o C for 24 h, and stored in plastic bags. Remains of teeth, claws, scales, feathers, bills, bones, hair and seeds, among other items, were identified by comparison with a reference collection, literature, and by assessment of museum (Museum of Zoology, University of São Paulo) and herbarium (Federal University of Minas Gerais) specialists. All food items were identified to the level of species or morph species. Diet composition was described in terms of the frequency of occurrence of the food items as a function of the total number of occurrences (sensu Dietz 1984, Motta- -Junior et al. 1996). Minimum number of individual prey animals was obtained by counting teeth, mandibles, and bills (Motta-Junior et al. 1996, Ray 1998). Whenever hair, feathers and scales were found, we considered them as belonging to a single prey individual (Emmons 1987). To estimate the number of fruits within one faecal sample, we compared the number of seeds obtained in the scat to the average number of seeds per fruit collected in the field, and data from the literature (Castro et al. 1994, Motta-Junior et al. 1996). To avoid duplication in prey count, only identifiable remains were considered (Emmons 1987). To complement diet analysis, we estimated consumed biomass for each food item by multiplying the minimum

Prey availability and diet of maned wolf 393 number of individual prey animals and fruit species found in scat samples by their average weight calculated from individuals collected in the field and from the literature (Fonseca et al. 1996, Motta-Junior et al. 1996). We used the average number of seeds per fruit species to estimate the number of fruits and their biomass consumed by maned wolf according to the proportion of seeds found in the scats (Castro et al. 1994, Bueno et al. 23). The frequency of occurrence and the minimum number of individual prey provide information on the relative impact the predator plays on the prey community, whereas the estimation of consumed biomass more accurately indicates the relative importance of each item in the predator s diet (Marti 1987). Also, the frequency of occurrence overestimates the importance of smaller prey, which is corrected by calculating the estimated biomass consumption. However, although smaller prey (< 4 kg) are eaten almost completely by the maned wolf (Motta-Junior et al. 1996), larger ones may be overestimated by the estimated biomass consumption. Food items were also grouped according to the habitats in which they were found, following field observation and data obtained from the literature (Lorenzi 1992, 1998, Fonseca et al. 1996). We considered five groups: (1) species exclusive to open habitats ( campo limpo, campo sujo, campo cerrado and campo rupestre ); (2) species exclusive to forests; (3) species occurring both in open habitats and forests; (4) introduced species; and () species whose habitat could not be clearly determined. Finally, food items were also grouped into large food categories (ie, wolf s fruit, P. obtusifolia fruit, miscellaneous fruits, grasses, arthropods, reptiles, Tinamiformes, other birds, armadillos, small mammals, and medium-sized mammals), and the total number of occurrences was separated into seasons and then compared using the G test (Zar 1999). In the case of a significant result, each food category was also tested applying the G test to check for non-seasonality. Abundance of small mammals and fruits The census of terrestrial small rodent and marsupial populations was conducted from November 1998 through January 2, except November 1999. Mammals were captured in wire-cage live-traps (4 2 2 cm, Cafelândia, SP, Brazil), spaced at 1-m intervals, and baited with a mixture of peanut butter, banana, maize flour, and sardine slices. Traps were distributed in six line-transects, three in each of the main habitats of the study site in the east of the SCNP ( campo limpo and campo sujo/cerrado ). Traps were set for three consecutive nights during a trapping session, amounting to a total of 3348 trap/nights. The line-transects were at least m apart, and thesampleeffortwasevenly distributed between the two habitats. The captured individuals were marked using an ear-hole sequence (Talamoni and Dias 1999), and they were weighed and identified to species. We used the number of individuals captured for the first time per 1 trap-nights for each month as an index of small mammal abundance, which we assumed as a reflection of small mammal availability (Jaksic 1989). The forest patches were not assessed in this study, since this habitat covered very small areas in the park and is less used by the maned wolf than open areas (Juarez and Marinho-Filho 22, Jácomo et al. 24). We also recorded the fructification phenology, monitoring five plots (three 2 1 m in campo sujo/cerrado and two 1 1 m in campo limpo ) during the study period, conducting a monthly count of the number of individuals of each species bearing fleshy fruit (green and ripe fruits) located up to one meter above the ground. The phenology of the wolf s fruit was carried out separately from April 1999 through January 2 due to the difficultly of encountering enough individuals of the species. We tagged 2 individuals of wolf s fruit, counting the number of fruits per month. Functional response and small mammal selection To verify a possible functional response of the maned wolf to seasonal fluctuations of the relative abundance of its prey species, small mammals as well as fruits, we used the Spearman correlation coefficient (r s ) (Zar 1999), according to Silva et al. (199). To complement the analysis, we conducted a set of correlations between the most important categories of food items, also using the Spearman correlation coefficient. The comparison between the abundance of small mammals and their incidence in the maned wolf diet was made by sampling the small mammal species in the same region where we collected the scats (Jaksic 1979, 1989). We used G test (Zar 1999) to test a hypothesis of non-selective predation, in which the observed frequencies were the absolute values of the individual number of each prey found in scats, while the expected frequencies were calculated from the proportion of each prey in the capture sample (Jaksic 1979). According to Jaksic (1979), when some species presented values of expected frequency less than, we grouped them to obtain higher expected frequencies (Zar 1999). A possible methodological caveat regarding trapping of small mammals is that, in general, studies on prey selectivity assume trapping results reflect availability of prey to predators (Jaksic 1989). This may be not the case, however, and some caution should be used when interpreting the statistical evidence (Jaksic 1989). Results Diet composition and analysis In total we found 1688 food items occurrences, belonging to 69 taxa (26 plant items, with 83 occurrences; 43 animal items, with 83 occurrences). The average number of food items found per scat was 4.23. Plant and animal items were consumed equally by the maned wolf in terms of frequency of occurrence (Tables 1 and 2). The species with the highest frequency in the scats were Parinari obtusifolia, wolf sfruit,necromys lasiurus, Ophio-

394 D. Queirolo and J. C. Motta-Junior des sp., Cavia sp., and Rynchotus rufescens, which alone yielded 4.1% of all occurrences. Regarding estimated biomass, plant material represented 43.2%, with P. obtusifolia and miscellaneous fruits being the most important categories (Table 1, Fig. 1). Wolf s fruit also accounted for a high percentage of the maned wolf s diet by its occurrence and biomass, whereas the contribution of grasses was insignificant (Table 1, Fig. 1). The importance of animal prey increased in terms of biomass (Table 2). The incidence of medium-sized mammals and armadillos had higher proportions when biomass was analyzed. The biomass of reptiles and arthropods were very small with respect to the analysis of frequency of occurrence (Fig. 1). Based on frequency of occurrence, most of the animal prey species (87.%) were small-sized (< 1 g), and consisted mainly of small mammals,birds,andreptiles(fig.2a).however,in terms of biomass, larger prey (> 1 g) were more common, representing 73.2% of the total biomass consumed (Fig. 2b). Moreover, a higher number of food items consumed by the maned wolf, including most of the animal prey and plant species, and their frequency of occurrence and biomass, occurred in open habitats (Table 3). One hundred eighty two scats (761 occurrences) were collected in the dry season and 218 scats (927 occurrences) in the rainy season. Plants accounted for 1.2% of the food items in the rainy season and 47.3% in the dry season. Animals represented 48.8% of the food items in the rainy season and 2.7% in the dry season (Fig. 3). In terms of biomass, the percentage of plants (36.7%) in the rainy season was lower than that of animals (63.3%), and in the dry season, plants (49.7%) and animals (.3%) showed similar percentages. Frequency of food occurrence varied with season when all major groups of food items were considered (G = 46.49, p =.1, df = 1). All plant items consumed varied seasonally (G = Table 1. Number, percentage of occurrence, and biomass (total mass and percentages of estimated biomass consumed) of plant food items identified in the diet of maned wolf (n = 4 scats) from November 1998 to January 2 at Serra da Canastra National Park, Brazil. Food items Occurrence Biomass n % Mass (g) % Solanum lycocarpum 167 9.9 19 47.1 1.6 Parinari obtusifolia 196 11.6 33 4. 18.7 Miscellaneous fruits Mangifera indica 3.2 12..7 Annona crassiflora 8. 726..4 Annona coriacea 8. 64..4 Annona sp. 2.1 8. Bromelia sp. 18 1.1 2663.4 1. Melancium campestre 64 3.8 128.8.7 Byrsonima sp. 1.1 4.2 Miconia sp. 1.1.3 Campomanesia sp. 29 1.7 889.2. Psidium guajava 47 2.8 922.8.2 Psidium rufum 8. 247. 1.3 Eugenia pyriformis 7.4 88.8 Eugenia involucrata 2.1 9.6 Allagoptera campestris 47 2.8 97.. Butia sp. 1.1.. Syagrus oleraceae 4.2 126..1 Pouteria sp. 17 1. 1927. 1.1 Alibertia sessilis 22 1.3 223.1 1.2 Vitex sp. 6.4 8. Unidentified sp. (four food items) 11.7 17.6.1 Subtotal miscellaneous fruits 36 18.1 24 721.8 13.9 Grass 166 9.8 23.6 Subtotal plants 83 49. 77 197.4 43.2

Prey availability and diet of maned wolf 39 36.6, p <.1, df = 3), whereas among animal prey, only arthropods (G = 4.3, p <.34, df = 1) and small mammals (G =.313, p <.22, df = 1) showed significant seasonality (Fig. 3). Functional response and small mammal selection There was fluctuation of small mammal abundance. We captured eight species and 33 individuals (range = 3.7 1.8 individuals/1 trap-nights). Most variation occurred during the dry season (Fig. 4). The maned wolf didn t consume small mammals in relation to its availability (r s =.9, p =.63) (Fig. 4). We identified 2 species of fruits, but only 17 were potentially consumed, primarily during the wet season (Fig. ). For all fruits, there was a positive and significant correlation (r s =.69, Table 2. Number, percentage of occurrence, and biomass (total mass and percentages of estimated biomass consumed) of animal food items identified in the diet of maned wolf (n 4 scats) from November 1998 to January 2 at Serra da Canastra National Park, Brazil. Food items Ocurrence Biomass n % Mass (g) % Arthropods (ten food items) 73 4.3 142.7.1 Reptiles Bothrops sp. 1.1 4.. Colubridae 1.6..3 Unidentified Serpentes 19 1.1 9.. Mabuia sp. 1.1.. Ophiodes sp. 1 6.2 24. 1.3 Tropidurus sp. 4.3 36..2 Stenocercus sp..4 16..1 Teiidae 7.2 2..1 Unidentified Squamata 3.2 3..2 Subtotal reptiles 1 9.2 492. 2.8 Birds Tinamiformes (four food items) 91.4 22 33. 12. Passeriformes 91.4 272. 1. Unidentified birds 31 1.8 11..6 Subtotal birds 213 12.6 26 1. 14.6 Mammals Gracilinanus sp. 1.9 162..1 Monodelphis sp. 8. 39..2 Unidentified Didelphidae 1.1 3.. Necromys lasiurus 13 7.7 11 327. 6.3 Calomys tener 18 1.1 46..3 Calomys callosus.3 114..1 Holochilus sp. 6.4 168..9 Oligoryzomys sp. 14.8 374..2 Oxymycterus sp. 9. 6..3 Oryzomys subflavus 1.1 68.. Unidentified Muridae 72 4.3 231. 1.3 Clyomys sp. 6.4 8..4 Cavia sp. 4 2.4 1..6 Unidentified small mammals 68 4. 84.. Subtotal small mammals 393 23.3 29 1. 16.3 Dasypus septencinctus 13.8 17 87. 1. Unidentified Dasypodidae 2.1 2131. 1.2 Subtotal armadillos 1.9 2 6. 11.2 Subtotal medium-sized mammals 4.2 21 3. 11.9 (two food items) Subtotal mammals 412 24.4 7 461. 39.4 Subtotal animals 83. 11 673.2 6.8

396 D. Queirolo and J. C. Motta-Junior 2 Frequency of occurrence Estimated biomass 2 Percentage 1 1 Wolf, s fruit P. obtusifolia Miscellaneous fruits Grasses Arthropoods Reptiles Tinamiformes Other birds Armadillos Small mammals Medium-sized mammals Fig. 1. Percentages of frequency of occurrence and estimated consumed biomass of the food items categories found in the diet of maned wolf at Serra da Canastra National Park, Brazil. Data gathered from November 1998 to January 2. Individual number (%) 6 4 3 2 1 (a) Biomass (%) 4 3 2 1 (b).1-1 1.1-1 1.1-1 >1.1-1 1.1-1 1.1-1 >1 Prey body mass (g) Prey body mass (g) Fig. 2. Percentage of individual number (a) and biomass (b) of animal prey items as a function of prey body mass categories in the diet of maned wolf at Serra da Canastra National Park, Brazil. N = 197 individuals and 173 8 g of total biomass. Table 3. Association of maned wolf food items (animal and plant species) according to habitat types where they occur most commonly and their origin. The values are percentages of absolute totals of food items, occurrence, and estimated consumed biomass. Association of food items Food items Occurrence Biomass Open habitats 34.8 6.8 63.6 Forests 2.9.2 9.2 Open habitats and forests 37.7 23.9 18. Introduced species 1.4.2.7 Undetermined 23.2 18.8 8. Total 69 1688 173 721.8 p =.34) between the consumption and availability data, suggesting a functional response by the maned wolf (Fig. ). Fructification of wolf s fruit presented less fluctuation, with peaks during the dry season (Fig. 6), but no functional response was observed (r s =.1, p =.7). However, when we compared the consumption of wolf s fruit with that of miscellaneous fruits, we observed a negative and significant correlation (r s =.873, p <.1), which may suggest a substitution between them (Fig. 7). The compar-

Prey availability and diet of maned wolf 397 Frequency of occurrence (%) 3 2 2 1 1 Wolf, s fruit Rainy season Dry season P. obtusifolia Miscellaneous fruits Grasses Arthropoods Reptiles Tinamiformes Other birds Armadillos Small mammals Medium-sized mammals Fig. 3. Percentages of frequency of occurrence of food items categories in rainy (n = 927 occurrences in 218 scats) and dry seasons (n = 761 occurrences in 182 scats), found in the diet of maned wolf from November 1998 to January 2 at Serra da Canastra National Park, Brazil. Frequency of occurrence in scats (%) 4 3 3 2 2 1 1 N D 1998 J Study area F M A M J J A S O D J 1999 2 Fig. 4. Small mammal abundance index (number of individuals per 1 trap-nights) in study area plotted against the percentage of frequency of occurrence of small mammals recovered from 4 maned wolf scats, over 14 months from November 1998 to January 2 at Serra da Canastra National Park, Brazil. Scats 18 16 14 12 1 8 6 4 2 Individuals per 1 trap-nights ison among the other food item categories (small mammals vs. wolf s fruit, small mammals vs. miscellaneous fruits, and small mammals vs. other animals) showed no significant correlations. The most commonly trapped small mammal species (N. lasiurus 67.3%, n = 222) was also the most abundant small mammal in the diet of the maned wolf (Table 2). However, Akodon lindber-

398 D. Queirolo and J. C. Motta-Junior 3 Study area 3 Number of individuals 3 2 2 1 1 Scats 2 2 1 1 Frequency of occurrence in scats (%) N D 1998 J F M A M J J A S O D J 1999 2 Fig.. Fruiting phenology (number of individual with fruits) in study area plotted against the percentage of frequency of occurrence of the same taxa recovered from maned wolf scats, over 14 months from November 1998 to January 2 at Serra da Canastra National Park, Brazil. 2 Scats 7 Frequency of occurrence in scats (%) 2 1 1 N D 1998 J Study area F M A M J J A S O N D J 1999 2 6 4 3 2 1 Individuals with fruits (%) Fig. 6. Fruiting phenology of wolf s fruit (percentage of number of individuals with fruits) in study area plotted against the percentage of frequency of occurrence of wolf s fruit recovered from 4 maned wolf scats, over 14 months from November 1998 to January 2 at Serra da Canastra National Park, Brazil. ghi (2.2%, n = 83), the second most commonly trapped, was not consumed by the maned wolf. The G-test result was highly significant (G = 23.767, df = 4, p <.1) with N. lasiurus being consumed more than expected, whereas consumption of A. lindberghi waslessthanexpected(table4).

Prey availability and diet of maned wolf 399 Frequency of occurrence of wolf, s fruit (%) 2 2 1 1 N D 1998 J Wolf, s fruit Other fruits F M A M J J A S O D J 1999 2 3 2 2 1 1 Frequency of occurrence of other fruits (%) Fig. 7. Frequency of occurrence of wolf s fruit (percentage) recovered from maned wolf scats plotted against the percentage of frequency of occurrence of miscellaneous fruits from the same 4 maned wolf scats, over 14 months from November 1998 to January 2 at Serra da Canastra National Park, Brazil. Table 4. Comparison of the observed number of individuals of each species of small mammals recovered from 4 maned wolf scats from November 1998 to January 2 with regard to expected number based on proportional abundances of the same species derived from trapping results in the same period. All the marsupials were grouped to perform G-test analysis. Small mammals species Observed Expected Gracilinanus sp. 9 1.7 Monodelphis sp. 8.9 Lutreolina crassicaudata 8. Necromys lasiurus 238 189.7 Akodon lindberghi 7.9 Oligoryzomys nigripes 17 1.7 Oxymycterus sp. 1 8.6 Total 282 282 Discussion In general, our findings regarding the diet of the maned wolf were comparable to other studies. However, some variation exists in relative importance of particular food categories. In many studies the wolf s fruit was most commonly consumed item (Dietz 1984, Motta-Junior et al. 1996, Juarez and Marinho-Filho 22, Rodrigues 22, Santos et al. 23), whereas miscellaneous fruits were more important in others (Motta-Junior 2, Aragona and Setz, 21, Bueno and Motta-Junior 24, Jácomo et al. 24). On the other hand, Silva and Talamoni (23) and Bueno et al. (23) found that small mammals were the most frequent component, as occurred in the SCNP. It is important to emphasize that contrary to other studies (Juarez and Marinho-Filho 22, Motta-Junior and Martins 22, Jácomo et al. 24) the SCNP is the only place where P. obtusifolia was consumed by wolves more frequently than the wolf s fruit. Among animal components, arthropods were not very representative in the maned wolf diet in the SCNP, despite being frequently observed in the diet of sympatric canid species (Dalponte 1997, Juarez and Marinho-Filho 22, Bueno and Motta-Junior 24). On the other hand, reptiles were relatively more important in our study than in other areas where they accounted for only 3.% of the diet (Motta-Junior 2, Aragona and Setz 21, Jácomo et al. 24). The frequency of occurrence of birds in the SCNP was similar to other sites (Dietz 1984, Rodrigues

4 D. Queirolo and J. C. Motta-Junior 22, Santos et al. 23, Jácomo et al. 24). However, armadillos were less consumed compared to findings from other studies. This can be partially explained by differences in prey densities among areas. The maned wolf is known to feed on medium-sized mammals, mostly deer (Motta-Junior et al. 1996, Bestelmeyer and Westbrook 1998, Juarez and Marinho-Filho 22, Jácomo et al. 24). In the SCNP, only remains of Cuniculus paca and an unidentified carnivore (either from the Mustelidae or Procyonidae family) were found in the scats. However, as these prey average 4 kg or more of body weight, their biomass may be overestimated. The importance of wolf s fruit didn t change when we compared frequency of occurrence and consumed biomass, contrary to what was observed by Juarez and Marinho-Filho (22). It was apparently caused by the fact that mean fruit weight in Juarez and Marinho-Filho (22) study was 6 g, whereas in our study site it averaged only 29 g. In SCNP the biomass of Tinaniformes was more important than frequency of occurrence, but the opposite was reported by Juarez and Marinho-Filho (22). This difference was likely due to larger species of Tinamidae that were found in the wolf diet in SCNP (eg, Rynchotus rufescens, g), as compared to those in Jaborandi municipality (average 1 g: Juarez and Marinho-Filho 22). This study, like those of Dietz (1984), Aragona and Setz (21), and Juarez and Marinho-Filho (22) confirms the omnivorous feeding habit of the maned wolf. In some sites, the consumption of plant material was higher than that of animals, but it was never over 6.% of the total food item occurrences (Motta-Junior 2, Rodrigues 22). We found a higher diversity of food items (n =69 species or morphospecies), compared to values between 3 and 44 reported in other areas (eg Aragona and Setz 21, Juarez and Marinho- -Filho 22, Jácomo et al. 24). The seasonality of the maned wolf diet in SCNP, in terms of frequency of occurrence and biomass, agrees with other studies, with exception of research by Aragona and Setz (21). Small mammals were mostly consumed during the dry season, while arthropods and fruits were more consumed in the rainy season, according to their availability. The same pattern was found by Dietz (1984), Motta-Junior et al. (1996), and Bueno and Motta-Junior (24). As in the Cerrado Region mature miscellaneous fruits (other than wolf s fruit) and insects are more abundant during the wet season (Dietz 1984, Rocha et al. 1994, this study) and small mammals are more abundant during dry months (Dietz 1984, Alho et al. 1986, this study), the maned wolf appears to be a temporal trophic opportunist with regard to these large groups of food. The qualitative analysis of the maned wolf diet indicates that it consumes a wide range of items, both in number and variety, however, quantitatively this predator have a relatively specialized diet in both seasons. The maned wolf consumed a few items in high percentages and a great variety of items in low percentages. Only seven items had more than.% frequency of occurrence each (wolf s fruit, P. obtusifolia, grasses, Ophiodes sp., Tinamiformes, Passeriformes and N. lasiurus), together totalling 6.% of the total occurrences and 1.% of total estimated biomass. The intense consumption of food items characteristic of open areas reinforces the urgency of preserving open habitat in the Cerrado, threatened by the spread of agriculture and cattle ranches (Klink and Moreira 22). Dietz (1984) studied the maned wolf by radiotelemetry in SCNP and found that it spent 76.% of its time in open habitats, and the remaining time in forested patches. In addition, Jácomo et al. (24), using a camera trapping technique, found that the maned wolf predominantly used the grassland habitat in Emas National Park, a protected area similar to SCNP. The negative correlation between consumption of wolf s fruit and miscellaneous fruits suggests that wolves use the different fruits to maintain an approximately constant rate of fruit intake in the face of seasonal variation in community-wide fruit availability (Motta-Junior and Martins 22). Apparently no functional responses were observed for wolf s fruit and small mammals, as reported for other areas in southeastern Brazil (Motta-Junior 2, Bueno and Motta-Junior 26). Conversely, in our study site, small mammal consumption was pos-

Prey availability and diet of maned wolf 41 itively and marginally correlated with its abundance, showing a tendency towards a functional response. These results, however, should be interpreted with caution, since trapping procedures may not be reflecting the real availability of food items in the field. Selection of small mammals by maned wolves observed in our study along with other research conducted in southeastern Brazil (Motta-Junior 2, Bueno and Motta-Junior 26) suggest that the maned wolf is not an opportunistic species, at least in relation to this group of animals. It seems to prey selectively on Necromys lasiurus a small terrestrial mammal that uses all cerrado physiognomies (Alho et al. 1986). In contrast, Akodon lindberghi despite it is very common, was apparently less vulnerable to the wolves, possibly due to the semi-fossorial character of these mammals (Emmons 199). Additionally, A. lindberghi was almost exclusively captured in campo umido (Herskovitz 199, D. Queirolo and M. A. Granzinolli, in prep.), which may decrease the probability of encounters with wolves. The constant rate of fruit intake by the maned wolf during the year, irrespective of its availability, suggests that this canid needs fruits on a regular basis. Because most of the fruit species consumed by the wolves occur exclusively on open cerrado, it should be taken in consideration in habitat management policies. These data on functional response and prey selection are relatively recent in the literature about the maned wolf and open new perspectives to future research on this species, especially those that may focus on its dietary opportunism (sensu Jaksic 1989). Additionally, further studies of the maned wolf diet selection should include more detailed data on behavioural and morphological traits of prey as a way to understand their differential predation (Corley et al. 199). Acknowledgements: We thank A. A. Bueno, E. Wang and G. Dotta for assistance in fieldwork and the Instituto Brasileiro do Meio Ambiente e Recursos Naturais Renováveis (IBAMA) for permission to study in the area. We also thank A. Percequilo,G.I.Ximenes,M.J.J.Silva,V.FagundesandJ. M. Pagnozzi for identification of small mammals, M. Martins for identification of reptiles, and J. A. Lombardi for identification of botanical specimens. We thank G. Dotta for critically reviewing early version of the manuscript. The English was revised by A. Kepple. F. H. G. Rodrigues, P. Ciucci, and an anonymous reviewer made invaluable comments and suggestions. Conselho de Aperfeiçoamento de Pessoal do Ensino Superior (CAPES) and WWF Brasil (CSR 123-99) provided respectively a grant and financial supporttod.queirolo.j.c.motta-junior received financial support from Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP 97/69-7). References Alho C. J. R., Pereira L. A. and Paula A. C. 1986. Patterns of habitat utilization by small mammals populations in Cerrado Biome of central Brazil. Mammalia : 447 46. Aragona M. and Setz E. Z. F. 21. Diet of the Maned wolf, Chrysocyon brachyurus (Mammalia: Canidae), during wet and dry seasons at Ibitipoca State Park, Brazil. Journal of Zoology, London 24: 131 136. Bestelmeyer S. V. and Westbrook C. 1998. Maned wolf (Chrysocyon brachyurus) predation on pampas deer (Ozotoceros bezoarticus) Central in Brazil. Mammalia 62: 91 9. Bueno A. A., Belentani S. C. S. and Motta-Junior J. C. 23. Feeding ecology of the maned wolf, Chrysocyon brachyurus (Illiger, 181) ( Mammalia: Canidae), in the ecological station of Itirapina, São Paulo State, Brazil. Biota Neotropica 2(2): http://www.biotaneotropica. org.br/v2n2/pt/download?article+bn182222+item Bueno A. A. and Motta-Junior J. C. 24. Food habits of two syntopic canids, the maned wolf (Chrysocyon brachyurus) and the crab-eating fox (Cerdocyon thous), in southeastern Brazil. Revista Chilena de Historia Natural 77: 14. Bueno A. A. and Motta-Junior J. C. 26. Small mammals selection and functional response in the diet of the maned wolf, Chrysocyon brachyurus (Mammalia: Canidae), in southeast Brazil. Mastozoología Neotropical 13: 11 19. Castro S. A, Silva S. I., Meserve P. L., Gutiérrez J. R., Contreras L. C. and Jaksic F. M. 1994. Frugivoría y dispersión de semillas de Pimiento (Schinus molle) por el zorro Culpeo (Psedalopex culpaeus) en el Parque Nacional Fray Jorge (IV Región, Chile). Revista Chilena de Historia Natural 67: 169 176. Corbett L. K. 1989. Assessing the diet of Dingoes from feces: a comparison of 3 methods. The Journal of Wildlife Management 3: 343 346. Corley J. C., Fernandez G. J., Capurro A. F., Novarro A. J., Funes M. C. and Travaini A. 199. Selection of cricetine prey by the Culpeo fox in Patagonia: a differential prey vulnerability hypothesis. Mammalia 9: 31 32. Coutinho L., Miranda H. S. and Morais H. C. de 22. [The Cerrado biome and the fire: I- characteristics.] Instituto de Estudos Avançados, Universidade de São Paulo, São Paulo, Coleção Documentos, série Ciências Ambientais, 2: 1 48. [In Portuguese] Dalponte J. C. 1997. Diet of the hoary fox, Lycalopex vetulus, in Mato Grosso, Central Brazil. Mammalia 61: 37 46. Dietz J. M. 1984. Ecology and social organization of the Maned wolf (Chrysocyon brachyurus). Smithsonian Contributions to Zoology 392: 1 1.

42 D. Queirolo and J. C. Motta-Junior Eiten G. 1994. [Vegetation of cerrado]. [In: Cerrado: characteristics,useandperspectives.m.n.pinto,ed].editora Universidade de Brasília, Brasília: 17 73. [In Portuguese] Emmons L. H. 1987. Comparative feeding ecology of felids in a neotropical rainforest. Behaviour, Ecology and Sociobiology 2: 271 283. Emmons L. H. 199. Neotropical rainforest mammals: a field guide. The University of Chicago Press, Chicago: 1 281. Fonseca G. A. B. da, Herrmann G., Leite Y. L. R., Mittermeier R. A., Rylands A. B. and Patton J. L. 1996. Lista anotada dos mamíferos do Brasil. Occasional Papers in Conservation Biology, Conservation International 4: 1 38. Herskovitz P. 199. Mice of the Akodon boliviensis size class (Sigmodontinae, Cricetidae), with the description of two new species from Brazil. Fieldiana, Zoology 7: 1 3. Instituto Brasileiro de Desenvolvimento Florestal/Fundação Brasileira para a Conservação da Natureza 1981. [Management plan: Serra da Canastra National Park]. Instituto Brasileiro de Desenvolvimento Florestal, Fundação Brasileira para a Conservação da Natureza, Brasília. [In Portuguese] IriarteJ.A.,JimenezJ.E.,ContrerasL.C.andJaksicF. M. 1989. Small-mammal availability and consumption by the fox, Dusicyon culpaeus, in central Chilean scrublands. Journal of Mammalogy 7: 641 64. Jácomo A. T. A., Silveira L. and Diniz-Filho J. A. F. 24. Niche separation between the maned wolf (Chrysocyon brachyurus), the crab-eating fox (Dusicyon thous) and the hoary fox (Dusicyon vetulus) in central Brazil. Journal of Zoology, London 262: 99 16. Jaksic F. M. 1979. Técnicas estadísticas simples para evaluar selectividad dietária en Strigiformes. Medio Ambiente 4: 114 118. Jaksic F. M. 1989. Opportunism vs. selectivity among carnivorous predators that eat mammalian prey: a statistical test of hypotheses. Oikos 6: 427 43. Jaksic F. M., Jiménez J. E., Castro S. A. and Feisenger P. 1992. Numerical and functional responses of predators to a long-term decline in mammalian prey at a semi- -arid Neotropical site. Oecologia 89: 9 11. Juarez K. M. and Marinho-Filho J. 22. Diet, habitat use, and home ranges of sympatric canids in central Brazil. Journal of Mammalogy 83: 92 933. Klink C. A. and Moreira A. G. 22. Past and current human occupation, and land use. [In: The Cerrados of Brazil: ecology and natural history of a Neotropical savanna.p.s.oliveiraandr.j.marquis,eds].columbia University Press. New York: 69 88. Lorenzi H. 1992. [Brazilian trees: manual of identification and cultivation of Brazilian native trees.] Editora Plantarum, São Paulo, Vol. 1: 1 382. [In Portuguese] Lorenzi H. 1998. [Brazilian trees: manual of identification and cultivation of Brazilian native trees.] Editora Plantarum, São Paulo, Vol. 2: 1 368. [In Portuguese] Marti C. D. 1987. Raptor food habits studies. [In: Raptor Management Techniques Manual. B. A. Pendleton, B. A. Millsap, K. W. Clinc and D. M. Bird, eds]. National Wildlife Federation, Washington, D. C., Scientific Technical Series, 1: 67 8. Ministério do Meio Ambiente 23. [National List of the Brazilian Threatened Species.] Diário Oficial da União, 11: 88 97. [In Portuguese] Motta-Junior J. C. 2. [Temporal variation and prey selection in the diet of the maned wolf Chrysocyon brachyurus (Mammalia: Canidae), at Luiz Antonio Ecological Station, SP, Brazil]. [In: Estudo Integrados em Ecossistemas. Estação Ecológica de Jataí. J. E. Santos and J. S. R. Pires, eds]. Rima Editora, São Carlos, São Paulo, Vol. 1: 331 346. [In Portuguese] Motta-Junior J. C. and Martins K. 22. The frugivorous diet of the Maned Wolf, Chysocyon brachyurus, in Brazil: ecology and conservation. [In: Seed dispersal and frugivory: ecology, evolution and conservation. D. J. Levey, W. R. Silva and M. Galetti, eds]. CABI Publishing, Wallingford: 291 33. Motta-Junior J. C., Talamoni S. A., Lombardi J. A. and Simokomaki K. 1996. Diet of the maned wolf, Chrysocyon brachyurus, in Central Brazil. Journal of Zoology, London 24: 277 284. Putman R. J. 1984. Facts from faeces. Mammal Review 14: 79 97. Ray J. C. 1998. Temporal variation of predation on rodents and shrews by small African Forest carnivores. Journal of Zoology, London 244: 363 37. RochaI.R.D.,CavalcantiR.B.,MarinhoFilhoJ.S.and Kitayama K. 1994. [Fauna of the Federal District.] [In: Cerrado: characteristics, use and perspectives. M. N. Pinto, ed]. Editora Universidade de Brasília, Brasília: 4 431. [In Portuguese] Rodrigues F. H. G. 22. [Biology and Conservation of Maned Wolf at Águas Emendadas Ecological Station, DF]. PhD thesis, Universidade de Campinas, Campinas: 1 96. [In Portuguese] Santos E. F., Setz E. Z. F. and Gobbi N. 23. Diet of the Maned wolf (Chrysocyon brachyurus) and its role in seed dispersal on a cattle ranch in Brazil. Journal of Zoology, London 26: 23 28. Silva J. A. and Talamoni S. A. 23. Diet adjustments of maned wolves, Chrysocyon brachyurus (Illiger) (Mammalia, Canidae), subjected to supplemental feeding in a private natural reserve, southeastern Brazil. Revista Brasileira de Zoologia 2: 339 34. Silva S. I., Lazo I., Silva-Aranguiz E., Jaksic F. M., Meserve P. L. and Gutiérrez J. R. 199. Numerical and functional response of Burrowing Owls to long-term mammal fluctuations in Chile. Journal of Raptor Research 29: 2 2. Talamoni S. and Dias M. M. 1999. Population and community ecology of small mammals in southeastern Brazil. Mammalia 63: 167 181. Zar J. H. 1999. Biostatistical analysis. Prentice Hall, Upper Sandle River, New Jersey: 1 663. Received 3 October 26, accepted 17 June 27. Associate editor was Krzysztof Schmidt.