'Mill/lily!mill/MlWilli/WI" LIBRARY 1 LIIChIPnn Sf’f}: fl“ '. i i I t.‘ :2 4%.. "" ”VP-mm... ‘3‘ a) URLJKP fy Pen-v- This is to certify that the dissertation entitled ECOLOGY AND SOCIAL ORGANIZATION OF THE MANED WOLF (CHRYSOCYON BRACHYURUS) presented by James Montgomery Dietz has been accepted towards fulfillment of the requirements for Ph.D. degree in Zoology fifiBQ ¥_,.... Major professor Date 4 August 1981 MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES “ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. ECOLOGY AND SOCIAL ORGANIZATION OF THE MANED WOLF (CHRYSOCYON BRACHYURUS) by James M. Dietz A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1981 C) Copyright by James Montgomery Dietz 1981 ABSTRACT ECOLOGY AND SOCIAL ORGANIZATION OF THE MANED WOLF (CHRYSOCYON BRACHYURUS) By James M. Dietz The maned wolf is the largest and most distinct mem- ber of the South American Canidae. Its distribution is limited tx> the grasslands and savannas of central South America. Although prized by zoological gardens and classified by conservation agencies as endangered with extinction, relatively little is known about the natural history of this species. Therefore, the objectives of this study included quantification and description of the following aspects of the ecology and social organization of the maned wolf: relative use of available habitat types, circadian and seasonal activity patterns, social organization, reproduction, mortality and other limiting factors, and relations with humans. The results of this study are based on two years of field work conducted in and around Serra da Canastra National Park in southeastern Brazil. Nine adult maned wolves were captured in large wooden box traps using live chickens as bait. Observations were James M. Dietz also recorded on the behavior of two captive pups. Cap- tured adults were immobilized, weighed, measured, examined for ectoparasites and fitted with radiotelemetric collars. Blood, urine and fecal specimens were collected for labor— atory determinations. Individuals were located daily on a scheduled basis and also were periodically monitored for several hours at a time. Availability of trophic resources was estimated by monthly censuses of small mam- mals and fruits. Scat analysis was used to determine food habits of resident maned wolves. Results suggest that maned wolves used grassland sig- nificantly less than the proportion of occurrence of that habitat. During the day, wolves most commonly rested in cerrado habitat; at night, they were most often found in forest habitat. Wolves were largely nocturnal with the majority of circadian activity occurring from dusk to mid- night. No significant differences in activity patterns were noted between bonded pairs, but males were signifi- cantly more active than females. The most important components of the diet of these maned wolves, as determined by relative volume and occur- rence in seats, were the fruit Solanum lycocarpum, small mammals and miscellaneous fruits. Consumption of these three items was consistent with their seasonal avail- ability suggesting that maned wolves were opportunistic and flexible in their feeding habits. James M. Dietz Feces were nonrandomly deposited along trails and were replaced after I collected them. Locus-specific defecation was practiced near dens and resting places. Maned wolves of the Serra da Canastra displayed facultative monogamy characterized by a long-term pair bond between male and female. The male and female, as a pair, had completely overlapping ranges, yet individuals were rarely located in close association except during the breeding season. Each bonded pair defended a territory averaging 27 kmz. Territorial boundaries were apparently inflexible except when an adjacent territory was vacated by the death of a resident wolf. Under those circum- stances, territorial invasion by the male of the adjacent pair was observed on three occasions. In three cases, vacancies created by the deaths of resident wolves were quickly filled by wolves that were presumably unlanded nomads. The greatest source of mortality of maned wolves appeared to be the result of reprisals by local ranchers for depredation on domestic fowl. In addition, six of eight examined wolves tested positive for cystinuria, an inherited metabolic disease occasionally fatal in captive wolves. Suggestions basedcxithe available literature and on the results of this study are presented concerning the conservation and management of this endangered species. ACKNOWLEDGEMENTS The Maned Wolf Project, as this research was denomi- nated in Brazil, was the result of the combined efforts of many people in several countries. It is with pleasure that I acknowledge the contributions of the following individuals who made this study possible. To Rollin H. Baker, chairman of my Advisory Committee, I extend my warmest thanks for participation in every aspect of this project. Similar thanks are also due present members of my Advisory Committee: S. C. Bromley, R. W. Hill, and G. A. Petrides; and past members of my Advisory Committee: J. H. Fitch and L. Gysel. Coordination of this research at the federal level in Brazil was not an easy task, but was handled smoothly and efficiently by Dra. Maria Teresa Padua and her competent team of professionals at IBDF, Brasilia. I thank her and Renato Petry Leal, José Carlos Duarte and Eduardo Kunze Bastos for their support during my two years in Brazil. The late Dr. Ivens P. Franqueira, past Director of IBDF in Minas Gerais State, invited me to study maned wolves in the Serra da Canastra. Without his zeal and enthusiasm for conservation, that national park would not have been iii created and my study would not have been possible. To his 'memory and to the following members of his staff, I extend my sincerest thanks: Rage, Nicia, Juliana, Marco Aurélio, Haroldo and Isaque. Dr. Mario Amoroso, Executive Director of FBCN helped with financial aspects of the study. Life in the Serra da Canastra would have been diffi- cult without the assistance of the IBDF staff in the park. I am indebted to Oliveiro A. Soares, park Director, and to members of his staff: Antonia, Amadeu, Jean-Pierre, Delmo, and Adaniel; to Hélio Lasmar and his work crew; to Guilherme and the park guards. All of the peOple of the small city of sec Roque de Minas were supportive of my study and contributed dir- ectly and indirectly to its success. To these wonderful people, I extend my thanks. Particularly helpful were Aldivino and Fiinha Fregfigia and their 14 sons and daughters. I also wish to thank Vicente Candido, his family and his ranch hand Sebastiao for loaning us sup- plies, housing and horses, and for teaching us to use the latter. The importance of the contribution of the rural resi- dents of the Serra da Canastra to this study cannot be overemphasized. Without the help of these industrious and humble people, little information concerning the maned wolf would have been collected during this project. To the fol- lowing neighbors, I extend agradecimentos g abragos; they helped make my years in the Serra da Canastra some of the iv most enjoyable in my life: Elpides and Nair Ferreira; Tonico and sons Antonio, José and Joao, and son-in-law Zezé; and Zé Dorvalino. Dra. Rosa Vieira, director of the Clinical Laboratory, School of Veterinary Medicine, Federal University of Minas Gerais gratuitously performed the laboratory analyses dur- ing the course of this research. I also thank J. O. Whitaker and N. Wilson for identification of ectopara- sites, and D. O. Straney, P. Meyers, and R. Pine for identification of small mammals. R. Ramalho and J. Carlos identified botanical specimens. P. C. Dietz, L. A. Dietz, and R. Krasnoborski deserve thanks for their translations of literature in various languages into English. The lat- ter two friends were also kind enough to read and edit the manuscript for this dissertation. This research was partially funded by generous grants from the following agencies: World Wildlife Fund/ International Union for Conservation of Nature and Natural Resources, Instituto Brasileiro para Desenvolvimento Florestal/Fundacao Brasileiro para Conservacao de Natureza, New York Zoological Society, and Michigan State University. Finally, I would like to acknowledge the contribution and participation of my wife, Lou Ann, in this study. It was she who gave up two years of her career to cook on a wood-burning stove, wash clothes in a creek, and chase maned wolves in the Serra da Canastra. It is to her that I dedicate this dissertation. TABLE OF CONTENTS ACKNOWLEDGEMENTS LIST OF TABLES . LIST OF FIGURES. INTRODUCTION . METHODS. The Study Area . Capture Procedure. Handling Procedure . . . Location of Maned Wolves by Radiotelemetry . Determination of Food Habits . Scat Collection. Scat Analysis. Census of Available Food Items RESULTS. Physical Characteristics of Captured Maned Wolves. Pelage, Weights and Measurements Hematology and Blood Chemistry . . . Description of Home Ranges in the Study Area . West Range . . . East Range . Taperao Range. Other Ranges . . Use of Available Habitat Types by .Maned Wolves Activity Patterns. . . Daily and Seasonal Variation in Activity . Observations on Individual Activity. Social Organization and Communication. Relationships of Individuals Within Pairs. Beta . . . . . . . Gamma and Delta. vi Epsilon and Lambda . Kappa, Zeta and Sigma. Vocalizations. . Deposition of Feces and Urine. Relationships Between Adjacent Pairs of wolves Floaters, Loners or Nomadic Maned Wolves Feeding Ecology. . Seasonal Availability of Trophic Resources . Food Habits of Maned Wolves in the Study Area. Exploitation of Domestic Stock by Maned Wolves Reproduction . . . . . . . . . . . . . . . . Denning. . Parturition and Litter Size. The Behavior of Two Pups Parental Care of Pups. Parasites, Diseases and Mortality. Ectoparasites. Endoparasites. Diseases . Mortality of Maned Wolves in the Serra da Canastra. . Relations of Maned Wolves with Humans. Human Disturbance Factors and Their Effects on Maned Wolves . Attitudes of Area Residents Toward Maned Wolves: Folklore . DISCUSSION . Evolution of Sociality in a Grassland Habitat. The Pair Bond in Maned Wolves. Conservation and Management. Maned Wolves in the Wild . Maned Wolves in Captivity. APPENDIX . BIBLIOGRAPHY . vii Page 77 78 84 88 93 98 99 99 105 118 121 121 122 128 133 135 135 136 140 140 142 142 144 145 147 151 159 160 161 167 170 175 Table 10. ll. 12. 13. 14. LIST OF TABLES Chronological outline of procedures per- formed during handling of maned wolves for the first time . Weights and measurements of captured maned wolves Hematologic values for 10 maned wolves . Biochemical serum values for seven maned wolves Use of three habitat types by maned wolves . Use of habitat types by maned wolves grouped by bonded pairs. . . . . . . . Seasonal use of three habitat types by maned wolves . . . . . Use of available habitat types with respect to their relative occurrence . . . . . Levels of activity of eight maned wolves grouped by bonded pairs. Levels of activity of eight maned wolves grouped by sex . . . . Capture dates for maned wolves in three home ranges. Substrate used by maned wolves for the deposition of feces. Animal material identified in 740 seats of maned.wolves . Plant material identified in 740 seats of maned wolves viii 21 39 42 44 55 56 58 6O 65 66 7O 91 109 110 Table 15. l6. 17. 18. 19. 20. 21. 22. 23. Volume and occurrence of selected food items in 740 seats of maned wolves Seasonal occurrence of selected food items in 740 seats of maned wolves Occurrence of selected food items in 740 seats of maned wolves in three home ranges Weights and measurements of two captured maned wolf pups. Endoparasites observed in fecal samples from eight maned wolves The results of selected laboratory exami- nations performed on urine samples from eight captured maned wolves. Woody plants identified in the grassland and cerrado habitats, Serra de Canastra National Park . Trees identified in the riparian forest habitat, Serra da Canastra National Park . Identification of small mammals captured in the Serra da Canastra National Park. ix 117 129 138 139 170 171 173 Figure 10. 11. LIST OF FIGURES Adult maned wolf iii the Serra da Canastra National Park, Brazil . Location of the Serra da Canastra National Park, Minas Gerais State, Brazil, South America . Wooden live-trap used to capture maned wolves. Trap is baited with a live chicken . Temperature, length of day and rainfall as criteria for definition of wet and dry seasons Locations of home ranges of maned wolves in the study area. Percentages of all radiotelemetric locations made during each six-hour time block and classified according to level of activity of observed maned wolves A chronological representation of the rela- tionships among wolves inhabiting the East, West and Taperao Ranges . . . . . . East Range occupied sequentially by maned wolves Delta and Gamma; Epsilon and Lambda. West Range occupied by maned wolves Kappa, Zeta and Sigma. . . . . . Percent of radiotelemetric locations per month in which Lambda and Epsilon were found in close association. Number of nights per month during which maned wolves were heard vocalizing in the East Range and in all ranges in the study area. . . . . . 13 18 33 50 62 72 75 8O 83 87 Figure 12. 13. 14. 15. l6. 1?. Areas of dispute between maned wolves residing in the West Range and in the East Range. . . . . . . . . Solanum lycocarpum, "fruit of the maned wolf" . Seasonal availability of fruit species in the study area. . . . . Monthly captures of individual small mammals . . . Estimated months of birth of 21 litters of maned wolves seen by residents of the Serra da Canastra . Numbers of pups in 17 maned wolf litters observed by residents of the Serra da Canastra. . . xi Page 96 101 104 107 124 127 INTRODUCTION The maned wolf (Chrysogyon brachyurus Illiger, 1815; Figure l) is the largest and most distinct member of the South American Canidae. The first detailed description of the species and its habits was published by the famed naturalist Don Félix D'Azara in 1801. He referred to the maned wolf by its Paraguayan name Agouara Gouazou (his spelling) which he translated as "large fox." Simpson (1941) added that this name originated in the language of the Tupi-Guarani Indians and has been shortened to guara. Rengger (1830) suggested that the name Aguara-guazu was onomatopoeic for the vocalization of the maned wolf. The animal is now known throughout Brazil as lobo guara, in Bolivia as boroche, aguaré guazu in Paraguay and Argentina, and 1929 a; grin in Bolivia and Peru. The taxonomic relationships within the family Canidae have been relatively unstable since Gray (1825) coined the family name in its present form. Within this context, classification at the generic level of the South American forms has been the most unstable (see Cabrera, 1931, and Clutton-Brock, 1976, for reviews). Fortunately the maned wolf has largely escaped these problems of "vexatious FIGURE 1. Adult maned wolf in the Serra da Canastra National Park, Brazil. Figure 1. technical nomenclature of the canids” (Simpson, 1941) and has not been challenged as a distinct genus since clari- fication of its current scientific name by Osgood in 1919 and 1934. The maned wolf is monotypic, and no geographic races are recognized. The question which canid groups constitute the maned wolf's nearest relatives still remains unanswered: Clutton-Broek et al. (1976) using morphological character analysis concluded that the maned Wolf is weakly related to the Dusicyon group. Chiarelli (1975) writes that the karyotype of Chrysocyon is very similar to that of Canis. However, the diploid number of the former is 76, while that of the latter is 78 (Newnham and Davidson, 1966). Interbreeding between these two genera is not likely (Bastos, 1978). Gel electrophoresis has been used to dif- ferentiate between the serum proteins of various carni- vores including the maned wolf and domestic dog (anig familiaris) (Durr and Schmidt, 1970) but a relatively low incidence of polymorphic loci has generally limited the effectiveness of this technique in the Canoidea (Seal, 1975). Morphological data collected from free-living maned wolves are relatively rare. Recently-published measure- ments can often be traced back through the literature to Azara (1801), who recorded the dimensions of only one cap- tive wolf. Detailed morphological information will be presented later in this dissertation; therefore, only a brief description will be presented here. The adult maned wolf weighs about 23.3 kg, stands nearly 90 cm at the shoulder, and has a total length of about 147 cm, of which approximately 45 cm is contributed by the tail. The pelage is long and buff-red over most of the body. The legs are black as is part of the dorsal erectile mane. White markings are found on the pinnae, under the chin and on the tip of the tail. Poeoek (1927) and Langguth (1969) have described respectively the external characters and the morphology of the maned wolf and have compared these data with those.from other South American canids. Studer (1904) made similar comparisons using cranial measurements. Two additional facts concerning the anatomy of the maned wolf are of interest: first, whereas the caeca of most canids are convoluted, the caecum of the maned wolf is short and cylindrical (Flower, 1879), as it is in the three species of Brazilian canids occupying forest habi- tats: the bush dog (Speothos venaticus), the crab-eating fox (Cerdocyon thous) and the short-eared dog (Atelocynus microtis) (Langguth, 1969). Second, the long legs of the maned wolf make it one of the tallest members of the Canidae. Sokolowsky (1927), speculating on the adaptive significance of these long legs, surmised that maned wolves are not adapted for swift running, as stated by Stains (1975), but rather for efficient travel and vision in tall vegetation. That the maned wolf, as well as the other South American canids, evolved from one or more North American prototypes is now generally accepted (Simpson, 1980). Since no fossil remains of the maned wolf have been found outside of the Brazilian Central Highlands, it has been assumed that the maned wolf evolved in that region (Langguth, 1975). However, the question of when this radiation took place is still a matter of some contention. Simpson (1980) believes that all South American canids evolved from North American forms that arrived during the faunal exchange brought about by the closing of the Panamanian land bridge at the beginning of the Pleistocene Epoch. Langguth (1969) argues that the maned wolf and the three Brazilian forest foxes descended from a common ancestor that must have been in Brazil well before the Pleistocene. The present species range of the maned wolf includes the following areas: all geographic regions of Brazil (excluding the Amazon Basin, parts of the arid Northeast and the Atlantic coast); the provinces of Formosa, Corrientes and Chaco in Argentina; most of Paraguay; eastern Bolivia; and the Pampas del Heath in Peru (Breyer, in prep.; Cabrera, 1958; Cabrera and Yepes, 1960; Hofmann et a1., 1975-1976; Paiva, 1973; Schaller, 1975; Schaller and Vasconcelos, 1976). Because the maned wolf is a species highly valued by zoological gardens, the management and breeding of these animals in captivity is relatively well documented. Acosta (1972) described hand-rearing a litter of maned wolf pups. These data were used by Bekoff and Jamieson (1975) to compare their physical development with that of other canids. The growth parameters for pups were estab- lished by Acosta (1972), Encke et al. (1970), and by Seidel (1972). Information on the type of enclosure and diet necessary to maintain maned wolves successfully in captivity has been summarized by Brady and Ditton (1979), Carvalho (1976), and by Crandall (1964). The longevity of maned wolves in captivity has been estimated at from 12 to 15 years (Silveira, 1969). Reproduction by maned wolves in captivity is rarely successful. In most cases the neonates are killed by one of their parents. This, perhaps, has led to the publish— ing of several articles concerning the reproductive behav- ior of this species in captivity. Female maned wolves are monoestrous and may copulate several times during an estrus period of about five days (Kuhme, 1975; Lippert, 1973). Copulatory behavior has been described by Lippert (1973). Copulation by captive maned wolves has been observed during the months of October through February in the Northern Hemisphere (Altmann, 1972; Encke, 1971), and from August through October in South America (Acosta, 1972). The birth of from two to five pups, a normal lit- ter size for this species, follows a gestation period of about 65 days (Faust and Scherpner, 1968). A record of seven pups in one litter was recorded in the 850 Paulo Zoo (Carvalho, 1976). Parturition has been described by Hammerling and Lippert (1975), and the studbook record for the species has been published by Roeben (1975). Another problem associated with the breeding and raising of maned wolves in captivity is that of parasites and diseases. Renal parasites in maned wolves were early described by Azara (1801) in Paraguay, and later by Magalhaes (1939) for maned wolves in Brazil. The nematode described by these authors was later-identified as the giant kidney worm (Dioctgphyma renale) (Giovannoni and Molfi, 1960; Lamina and Black, 1966; Matera et a1., 1968), commonly fatal in captive and probably also in free—living maned wolves. Silveira (1969) suggested that the Solanaceae eaten by maned wolves might have an inhibiting action on this parasite, but the methods used to test this hypothesis were questioned by Carvalho (1976). Other endoparasites found in maned wolves are mentioned by Carvalho (1976), and by Encke et a1. (1970). Bush (in press) reviewed the diseases and medical management of maned wolves in captivity. Maned wolves in captivity lead essentially solitary existences except during the reproductive period. Pairs composed of animals of opposite sexes are apparently easier to house in the same enclosure than like-sexed animals. Individuals maintain separate preferred resting sites and do not usually rest in close association (Altmann, 1972; Encke, 1971; Kuhme, 1975). Kleiman (1972) contrasted the solitary behavior of the maned wolf with the highly gregarious nature of the bush dog (Speothos venaticus). The vocalizations of maned wolves have been described by Brady (in prep.), and the behavior of pups by Encke et a1. (1970). Information concerning the ecology and natural history of free-living maned wolves is scarce and often anecdotal. The wolf feeds mainly on small vertebrate animals and on fruits (Azara, 1801; Cabrera and Yepes, 1960; Carvalho, 1976). Dennler de la Tour (1968) and Miller (1930) describe the stiff-legged pounce used by maned wolves capturing small prey. Roosevelt (1925) and Krieg (1948) write that the wolf often preys on small domestic stock but this allegation is refuted by Cabrera and Yepes (1960). Krieg (1928, 1940 and 1948) refers to the wolf as the "savanna stroller” and comments on its solitary nature. Dennler de la Tour (1968) also noted that the wolf is usually seen alone and hints that individual territories may be maintained. The maned wolf is classified by the International Union for the Conservation of Nature and Natural Resources as "vulnerable" (IUCN Red Data Book, 1976), and classified by agencies of the Brazilian government as "endangered" (Carvalho, 1968; Coimbra Filho and Magnanini, 1968). Silveira (1968) estimated that only l,500-2,200 maned 2 wolves remain in 650,000 km in Brazil, and Meritt (1973) details pathways by which wolves captured in Paraguay and lO northern Argentina were exported out of Paraguay in 1971. However, in spite of what appears to be a rather bleak future for the maned wolf, little research is currently underway on this species, and no objective suggestions have been formulated concerning its conservation. In consideration of the above, the general objective of the present study was to obtain sufficient data about the ecology and social organization of the maned wolf so that rational decisions could be made with respect to the conservation and management of this endangered species in its natural range. Within this context I attempted to achieve the following specific objectives: (1) measure quantitatively the relative use by maned wolves of avail- able habitat types in the Serra da Canastra National Park, Brazil; (2) describe diurnal and seasonal activity pat- terns; (3) describe the social organization of the wolves of the Serra da Canastra and the behavioral mechanisms by which social relationships are maintained or reinforced; (4) quantify seasonal food habits, including the exploita- tion of domestic stock; (5) make observations on the reproduction of maned wolves in the Serra da Canastra; (6) identify direct sources of mortality of wolves, and also factors such as diseases and parasites that may limit population expansion; and (7) portray the relations between humans and maned wolves residing in the Serra da Canastra. METHODS The Study Area "Serra da Canastra," loosely translated, means "moun- tain range shaped like a foot-locker,” and is the regional name given to an area of hills and ridges located 300 km west of Belo Horizonte, capital of the state of Minas Gerais, Brazil (Figure 2). The hills rise abruptly from surrounding agricultural land to elevations of nearly 1,500 m and descend into valleys of about 800 m elevation. The areas of higher elevation are characterized by poorly drained soils supporting seasonal grasslands. Subject to heavy grazing and to local pasture improvement techniques, these grasslands are composed of a variety of species of Gramineae such as Aristida Sp., Tristachya sp., and the exotic Melinis minutiflora. Also present are occasional shrubs and small trees such as Vellozia sp. and Arnica montana. Grassland habitat comprises perhaps 50 percent of the area of the Serra da Canastra. Approximately 35 percent of the area was composed of cerrado vegetation. This habitat type is typified by well-spaced, thick-barked trees (see Table 21, Appendix) less than about 7 m in height. Trees of the cerrado are 11 12 FIGURE 2. Location of the Serra da Canastra National Park, Minas Gerais State, Brazil, South America. 13 .NMMUWB LS? Brazil Brasilia . Minas l Gerais o Paulo " (3 io de Janeiro SERRA DA CANASTRA NATIONAL PARK Figure 2. l4 typically drought- and fire-resistant with thick tortuous stems, limbs and branches. Understories are composed of a variety of annual plant species (e.g., Gramineae, Cyperaeeae and Leguminosae) whose presence is often influ- enced by disturbance factors such as annual burning or clearing of brush to "improve" pasture for cattle grazing. Many streams and rivers, including the Rio Séo Francisco, originate in the highlands of the Serra da Canastra. Along these waterways and at their headwaters are found narrow strips of seasonal trepical forest. This riparian habitat type accounts for perhaps 15 percent of the area and is composed of a wide variety of tree species (Table 22, Appendix). In this habitat type, understories vary from relatively open ones in drier areas to thick tangles of vines, herbaceous and shrub species in more mesic areas. The climate of the region is subtropical with well- defined wet and dry seasons. November through February are typically the wettest months and have a mean tempera- ture of about 22° C. June through August are the driest months and have a mean temperature of approximately 18° C. Frosts occur rarely in the Serra da Canastra. Yearly rainfall varies from about 1,300 to 1,700 mm. Winds aver- age 18.9 km/hr and are predominantly from the East (Pédua, 1978). A final characteristic common to all regions of the Serra da Canastra is the ubiquitous presence of humans and 15 their activities. Accessible parts of each range have been annually burned and heavily grazed by cattle during the past several generations. Cowboys tending cattle ride or walk over much of the area on a daily basis. Temporary and permanent human dwellings, as well as dirt roadways, are scattered throughout the area. Within this hill country (20° South latitude, 46° West longitude) is locatd in the Serra da Canastra National Park. The park was created by Brazilian federal decree in 1972, but at the time of my study had not been paid for and was still being used for cattle grazing by local ranchers. The park is 715 km2 in area and is fenced and regularly patrolled by guards in automobiles. A dirt road traverses the park from the city of Séo Roque de Minas on the east to Sacramento to the west. Upland seasonal grassland is the most common habitat type within the park. The study area for my research consisted of the eastern half of this park and the areas adjacent to it to the east and north. My wife and I, with help from workmen of the Brazilian Forest Service (IBDF), rebuilt a stone house originally occupied by slaves during the 18003. We lived in this house in the park from July 1978, through May 1980. Our living conditions were pleasant, and by main- taining a large flock of chickens and a garden, we were largely self-sufficient. We soon became friends with several rural families living outside the park and relied 16 on them for help in collecting much of the information presented in this dissertation. Capture Procedure The first critical problem faced during this field research was that of devising a noninjurious method to capture maned wolves in the Serra da Canastra. Political considerations as well as the likelihood of breaking the long and apparently fragile limb bones of captured maned wolves precluded the use of steel leg-hold traps. Exten- sive interviews with area residents produced several models of live-traps purported to have been successful in capturing wolves. From these suggestions, I designed a wooden live-trap (Figure 3), and hired a local carpenter to build seven to my specifications. The traps were constructed of durable native hard- woods and weighed about 110 kg assembled. The sides of the trap were held in place by bolts to permit disassembly for transport. The trap design was an open-ended box with a falling door released by a treadle mechanism located at the back of the trap. A live chicken supported on a plat- form above the treadle was used as bait. The chicken was provided with a four-day supply of corn and water in liter cans and was separated from the inside of the trap by a wall of light chicken-wire. A second critical problem was trap placement. Since initial attempts using random distribution of traps within 17 FIGURE 3. Wooden live-trap used to capture maned wolves. Trap is baited with a live chicken. 18 . U 4§§§§§§§ $§§§§§V§§x§§§$ . x \3 \§\s§§§ . (A\\§x\\§\\\mlx\\m\\\\\\x§\\\\\\\\\\\\\\\\\\\\\\,§\.\\.\\\“W\\N. ... 3a m\\. . \ \\ . : .... _\ \- A i 1/. Figure 3. 19 the study area produced no captures, I again interviewed area residents and determined the exact locations where maned wolves had recently been observed. At these areas, I estimated quantities of urine, feces and tracks to determine which trails were currently being used by maned wolves. All traps set within a few meters of these major trails were successful in capturing resident maned wolves. Traps were placed in pasture, tall grass or scrub brush and were not camouflaged or hidden. When a particular trap repeatedly captured the same wolves, that trap was temporarily closed or was moved to a new area of wolf activity. A third critical problem was daily trap inspection. Because the objective of trapping was to capture and place radiotransmitters on all resident maned wolves in the study area, sign of wolf activity and not an area grid design selected a priori was used to determine trap place- ment. When traps were in place, the size and topo- graphical irregularity of the area, and the general lack of roads absolutely disallowed my examination of all traps in a single day. Therefore it became necessary to enlist the help of interested area residents in this respect. Each trap was assigned to a resident of the study area and was placed such that he could check that trap from a dis- tance during the course of his daily activities. These volunteers informed me immediately when a wolf was dis- covered in a trap. For this service, I paid one day's 20 wages (approximately US$2.00) for the capture of a wolf with a radiocollar, and eight days' wages for each wolf captured for the first time. Handling Procedure The information collected from a trapped maned wolf usually depended on whether that wolf had been captured previously. Recaptures of wolves were relatively common, and the associated handling procedure was not complex. While the wolf was still in the trap, I tested its trans- mitter for proper functioning and examined the wolf for signs of external injuries. Attempts were made to photo- graph the wolf before and after it left the trap. In every case the captured wolf tore a hole in the chicken wire and consumed the chicken used for bait. The handling procedure associated with the first cap- ture of any wolf was more complex and involved the collec— tion of information and biological materials according to a standardized chronological outline (Table 1). The first step in this procedure was collection of a urine sample for use in testing that wolf for cystinuria. I noticed that trapped maned wolves urinated when first approached and I capitalized on this to collect fresh urine samples from eight wolves. I used one stick to distract the wolf land simultaneously inserted a second stick, with a piece of filter paper attached, between the legs of the wolf and collected the urine as it was voided. The urine absorbed 21 TABLE 1. Chronological outline of procedures performed during handling of maned wolves captured for the first time. Activity Assign name to wolf Record date Trap type, bait and person.in.charge NOte location and habitat description Collect urine sample Apply innobilizing drugs and note doses Record time to immObilization Sex of captured wolf Apply ear tags and.record numbers Nete identifying marks Affix.transmitter and record frequency Start transmitter Collect whole blood and mix half with anticoagulant Apply antibiotic injection.and note dose Record other treatment Describe tooth wear incisors canines cheek teeth Measure length of canines upper lower NOte reproductive condition testes vulva lactating nipples gravid Record condition and color of pelage Measure the following weight total length tail ear right hind.foot forefoot to scapula hindfoot to iliun white on tail blaCk on front foot blaCk on hind foot blaCk on mane 22 TABLE 1 (continued) Activity Determine heart rate Determine respiration rate Collect all ectoparasites Note time elapsed before wolf is alert Collect fresh feces fran the trap Photograph capth'ed wolf Record names of field personnel present 23 onto the filter paper was stored for future analysis by paper chromatography. The following combination of immobilizing drugs was injected intramuscularly by means of a syringe recessed into a hole bored down the end of a broomstick: 20 mg Sernylan (Parke-Davis); 7.5 mg Rompun (Haver-Lockhart); 0.75 mg Acepromazine (Haver-Lockhart). This dose usually rendered the wolf immobile within 10 minutes and allowed about two hours of working time before the animal began to regain consciousness. No permanent adverse effects were noted with this combination of immobilizing drugs, and all immobilized maned wolves recovered physical control quickly and apparently completely. Whole blood was collected from a superficial vein of a limb and a portion was mixed with anticoagulant (KEDTA) for serological determinations. Blood samples and feces from the trap were immediately sent for analysis to the School of Veterinary Medicine, Federal University of Minas Gerais, Belo Horizonte. Tooth wear (after Gier, 1957) was used to classify each wolf according to a relative age class: young, middle-aged, or old. A11 ectoparasites found were counted and collected for future identification. Any breaks in the skin were treated with a topical spray antiseptic (Topazone, Eaton Labs.) and a prophylactic dose of 2 ml Longicil Fortified Antibiotic (Penicillin G; Fort Dodge Products) was applied subcutaneously. Chloromycetin Ophthalmic Ointment (Fort Dodge Products) was used to 24 reduce drying of the eyes during immobilization. A single Rototag (size no. 5; Dalton, England) was affixed to the lateral margin of each pinna. A radiotransmitting collar was fitted around the neck of each wolf. Transmitters (Wildlife Materials) measured 4 cm by 11 cm, were equipped with external whip antennae 26 cm in length, weighed about 250 g and operated in the 151 MHz frequency range. These transmitters were equipped with bimodal activity-monitoring switches to constantly trans- mit information about the relative levels of activity of radiocollared wolves. Transmitters were fastened to col- lars of nylon webbing which were riveted around the necks of captured wolves. The sizes and shapes of collars were based on models of the necks of maned wolves from a zoological garden (courtesy of Dr. P. T. Robinson, San Diego Zoo). Various types of receiving antennae were tested, and a collapsible three-element yagi model (Wildlife Materials) proved most satisfactory under field conditions in the Serra da Canastra. Location of Maned Wolves by Radiotelemetry The objectives expedited by radiotelemetry were the determination on a year-round basis of home ranges, acti- vity patterns, habitat preference, and the monitoring of social behavior of maned wolves residing in my study area. These objectives were achieved by locating radiocollared wolves according to a daily schedule based on four six-hour 25 time blocks: 0601-1200, 1201-1800, 1801-2400, and 0001- 0600 hours. Temporal and economic constraints precluded the daily location of all wolves during each time block. In addition, the accentuated tepographic relief in the area severely restricted my capacity to travel at night and often reduced the effective radius of transmission of a transmitter to a few hundred meters. Nonetheless, after all resident wolves were fitted with radiotransmitters, an effort was made to locate each wolf at least once in each 24-hour time period. Attempts to radiolocate particular wolves were dispersed over time so that each time-block was represented in a given month. I used triangulation based on two or more radio- telemetric vectors to determine the location of each wolf (e g., Craighead and Craighead, 1972; Craighead et al., 1973; Seidensticker et al., 1973). For each of these locations, the time of day, level of physical activity of the wolf and habitat type were recorded. Based on 30 seconds of monitoring at the time of location, I assigned that wolf to one of three distinct classes: inactive (no activity recorded during the interval), intermittently- aetive (activity not continuous during the interval), or constantly-active. These activity classes were established by affixing radiocollars to domestic dogs and observing the effects of various levels of physical activity on transmitted pulse rates. The type of habitat where a wolf 26 was located was classified grassland, cerrado (see Description of Home Ranges, this dissertation) or forest. Radiotelemetric triangulation also facilitated the location of maned wolves for subsequent visual monitoring of individual and social behavior. This type of observa- tion was attempted periodically when a wolf was known to be active or in a location where visual observation was likely to be successful. However, since maned wolves were largely nocturnal, extensive visual observation of their behavior was not often possible. If a wolf had not been visually observed for approxi- mately a month, or if it was inactive and in the same location for several days, radiotelemetry was used to approach the animal to determine if it was in good physi- cal condition. This activity often forced the wolf to abandon its resting place and to take flight at my approach. Observations on the behavior and physical con- dition of the wolf were recorded at this time. The con- stant movement of cattle and ranch hands throughout my study area often displaced these maned wolves in a similar manner; therefore, it is unlikely that my interference had any significant effects on their behavior or activity pat- terns . 27 Determination of Food Habits Scat Collection The food habits of maned wolves residing in my study area were determined by the collection and analysis of fecal droppings (seats). To reduce the possible confusion of maned wolf seats with those of other predators, only those fecal specimens 25 mm or larger in diameter were collected. This criterion was based on the measurement of seats known to have been deposited by a maned wolf and of seats known to have been deposited by other species of canids also occurring in my study area. Seats of ques- tionable origin either were not collected or were analyzed separately from those probably deposited by maned wolves. After several months of seat collection, additional characteristics of maned wolf seats also became evident. These included odor, texture and site. All maned wolf seats found were collected for analy- sis. However, scheduled collection was limited to the three home ranges of the maned wolves that I intensively studied. A goal for collection of at least 50 seats per home range per month was arbitrarily set and was achieved with varying degrees of success depending mainly upon climatic variables. The physical determination of these ranges and their mutual exclusivity was made empirically by radiotelemetric locations of resident wolves. Seats collected from the interface between two ranges were 28 omitted from analyses relating to home ranges but were included in other analyses. Collected seats were placed in small paper bags and the following information recorded for each sample: date, location, freshness (moist and apparently intact; dry and apparently intact; or dry and disintegrating), elevation from Surrounding ground level, substrate, and habitat type (grassland, cerrado or forest). All feces deposited in a single location (approximately 0.5 m in diameter) were classified as a single seat except when obvious differ- ences in dates of deposition were evident. In this latter case, feces deposited at a single location were divided according to seat freshness criteria and were analyzed separately. Scat Analysis Seats were air-dried and then stored in paper bags until analysis. A measure of seat diameter for each sample was based on the largest diameter of a randomly chosen continuous segment. After recording all identi- fiable components, each seat was disassociated into the following 16 major food types: (1) reptiles and amphi— bians; (2) birds; (3) arthropods; (4) mollusks; (5) fish; (6) fruits except Solanum lycocarpum; (7) foliage; (8) small mammals; (9) armadillos; (10) rabbits; (11) Solanum.lycocarpum; (12) soil; (13) paca (Agouti paca); (14) giant anteater (Myrmecophaga tridactyla); 29 (15) collared anteater (Tamandua tamandua); (16) unidenti- fiable material. These categories were selected g priori based on my previous experience in examining seats of maned wolves in other geographic regions of Brazil. Identification of fecal constituents was facilitated by the use of a reference collection that I assembled in the study area. I measured the volume of the major components of each seat by placing the material in a graduated cylin- der and compacting it with a close-fitting dowel rod. Although I attempted to compact each component equally, I recognize the potential error attributable to differential compaction of components varying in amount and degree of compressibility. The assumption that proportions of food items in fecal droppings accurately reflect the ingested propor- tions of those food items is probably as unfounded for maned wolves as it is for other canids (e.g., Murie, 1946). The quantitative problems of this method are well docu- mented and include differential digestibility of types and sizes of food items (Floyd et al., 1978; Meriwether and Johnson, 1980). Thus, my results probably do not accur- ately reflect the relative proportions and relative impor- tance of prey consumed by maned wolves. The nonmanipula- tive alternatives to determination of food habits by scat analysis include direct observation of feeding success (e.g., Brady, 1979; Schaller, 1972) or the examination of stomach contents (e.g., Korschgen, 1980). Unfortunately, 30 neither of these alternatives was feasible in the present study. The nocturnal and extraordinarily wary nature of the maned wolves at the Serra da Canastra made intensive behavioral observations largely impractical. Furthermore, the endangered status of the species precluded taking or risking the lives of wolves in my study area. A quantitative evaluation of the relative importance of each component food item identifiable in fecal drop- pings is customarily expressed in one or more of three manners: relative volume of each component, calculated as Total Volume of Component 1 . (1) Total Volume of All Components’ frequency of occurrence, based on numbers of occurrences, calculated as Number of Seats in Which Component i Occurred_ (2) Total Number of Occurrences of All Components’ or frequency of occurrence, based on the number of seats, calculated as Number of Seats in Which Component 1 Occurred. (3) Total Number of Seats ’ The disadvantages of each method considered alone and the lack of quantitative data concerning the trophic habits of this species suggested the use of the first two of the above formulas in the enumeration of the data from this study. The third expression, frequency of occurrence 31 as a function of the total number of seats, can be obtained by multiplying Formula 3 by Formula 4: Total Number of Occurrences of All Components (4) Total Number of Seats Volumetric data from scat analyses were transcribed directly onto Fortran coding forms and were subsequently punched onto computer cards. Data analyses were performed using a series of SPSS (Statistical Package for the Social Sciences, Version 8, Vogelback Computing Center, Northwestern University) programs on the Control Data Corporation Cyber 170, Model 750 computer at the Michigan State University Computer Laboratory. To investigate seasonal changes in the food habits of the maned wolves of the Serra da Canastra, it was neces- sary to define the climatic seasons in that region for the two years of this study. Maximum and minimum.temperature were measured daily; precipitation, actual temperature, relative humidity, cloud cover, wind speed and direction were determined at intervals of 12 hours (approximately 0630 and 1830 hours). Mean monthly minutes of sunlight were estimated from astronomical tables (World Almanac and Book of Facts, 1980, Newspaper Enterprise Association, New York). Total monthly precipitation and monthly averages for minimum temperature and minutes of sunlight were selected as criteria for determination of seasons (Figure 4). Other measured climatic variables were also seasonal FIGURE 4. 32 Temperature, length of day and rainfall as criteria for definition of wet and dry seasons. Curves: temperature = the mean of daily mini- mum temperatures for each calendar month; length of day = the mean number of minutes of daylight for the days in each month; rainfall = total measurable precipitation per month. O '0' I n 33 (m0) 11V:INIVH O O 10 D20 z 8 s - o” (I) .0 I > (o a: \ I c. "\ I 3. Q... .9.‘ .’.‘ z ’o I O O' l’ a) < f I m ,, t“ l .- ‘ w \\v§\ ““ B \ ““\‘ \ e \ ~ 5 ‘ g 4 x I \ s“ \ \\ \ \‘ z ‘ I o \ a) \l < A :11 wt l >. g ’ " l’ O I ’ ” I t , ’ 1' B‘ I o ‘5‘. .O’ x?‘ ’ \ I z: \I 8 ’ ‘\ 6 o 5” 7“ co 5 3‘ ‘b g. 3 0‘ O' m 5 \ 0' a: 3 E \g’ 3 Q 0 \ “ fl .— _l ‘ C \ \ x ‘\ O g (min) xva so H19N3‘l man 3 m h 0 '0 V '9 N z: ,0. m a: (0.) BUMVUSdWSJ. WRWNIW NVBW 7 10 12 6 MONTH Figure 4. 34 but were perhaps less important in the regulation of pri- mary productivity in the study area. Dry seasons were defined as the months of April through September and wet seasons as the months of October through March. Census of Available Food Items A list of the food items typically consumed by maned wolves has certain academic value, but limited practical relevance without the additional knowledge of what food items were available to choose from. More specifically, any resource-oriented management strategy would certainly benefit from knowledge of the extent to which maned wolves use available resources. A consensus taken from the pub- lished accounts of the food habits of maned wolves indi- cated that small mammals and native fruits were perhaps the items most commonly exploited by this species (e.g., Cabrera and Yepes, 1960; Carvalho, 1976), The avail- ability of these food items was therefore selected as an indicator of the relative abundance of trOphic resources available to the maned wolves in the study area. Avail- ability of these resources was estimated on a monthly basis. Small-mammal populations were sampled in each of the three major habitat types in the park using wooden-based rat traps. Fifty traps baited with masticated rolled oats were set for three consecutive nights during the last week of each month in each habitat type. The area to be 35 sampled each month within each habitat type was selected randomly from a list of candidate areas accessible from my field headquarters. To diminish the sampling problems associated with the depletion of populations in small areas, I adopted the convention of not sampling the same area twice consecutively. Traps were placed four meters apart in a straight line following a randomly chosen vector within each habitat type. Marginal areas between different habitat types were avoided. Study skins were made from the majority of captured small mammals, and the remainder were preserved in formalin. All specimens were deposited in the collections of The Museum, Michigan State University (Table 23, Appendix). The trap lines described above were also used as transect lines for the estimation of relative diversity and abundance of fruits available to maned wolves. The 5 ha area sampled in each habitat type was arbitrarily defined by a strip 10 m wide running the length of the trap line. Any fruits larger than 1 cm in diameter and less than 1 m above the ground level were defined as available to maned wolves. Relative abundance of a fruit species was estimated by the number of individual plants in fruit in the sampled area. The above sampling scheme estimated the relative monthly diversity and abundance of all available fruits but indicated nothing of the extent to which these might be used by maned wolves. To determine which fruit species 36 might be eaten by maned wolves, I interviewed area resi- dents and from their responses made a list of fruits likely to be consumed. The months when these species were in fruit were noted and examples were incorporated into a reference collection used in the analysis of seats of maned wolves. Finally, the fruits found in seats of maned wolves were recorded, and the season in which these species were in fruit was documented. RESULTS Physical Characteristics of Captured Maned Wolves Various authors (e.g., Azara, 1801; Cabrera and Yepes, 1960; Langguth, 1969) have described in detail certain physical attributes of the maned wolf such as pelage, weights and measurements. My observations are in general agreement with these previously published descriptions and therefore are not stressed in this dissertation. However, as published data on the hematology and blood chemistry of this species are unavailable for free-living individuals, I have considered these aspects in greater detail. Pelagg, Weights and Measurements Examination of close-up photographs of the nine adult maned wolves captured in the Serra da Canastra revealed almost no observable variation in pelage coloration between these individuals. One male, Lambda, was slightly darker on the cheeks and muzzle, but at distances greater than about 50 m I found it impossible to identify any of the individuals in my study area without the aid of radio- telemetry. The guard hair on the bodies of captured wolves was about 8 cm in length, straight and golden-red in color. 37 38 No underfur was present on these adult wolves. The distal portion of the muzzle was black as was the antebrachium distal from the elbow, and the hindlimb distal from the tarsus. Of the dorsal erectile mane, approximately the cranial half was black. Black from the mane also radiated laterally over the scapulae. The inner aspects of the pinnae were white as were the intermandibular and throat regions. The distal half of the tail was also white. The length of white on the tail and of black on the forelimb, hindlimb, and on the mane were measured on five of the cap- tured wolves (Table 2). Weights and selected measurements of nine captured adult wolves are also presented in Table 2. All compari- sons of these values between age classes, and all but one comparison between sexes were not significantly different at the p = .05 level of significance. A Student's t-test for unpaired observations (Steel and Torrie, 1960) indi- cated that sampled male wolves had significantly longer pinnae than sampled females (p < .05) in this small group. Hematology and Blood Chemistry I collected whole blood for hematologic and serum analysis from eight adult maned wolves in the Serra da Canastra. In addition, the veterinarian at the Belo Horizonte Zoological Garden contributed blood samples from two maned wolves (Zoo 1 and Zoo 2) recently acquired by 39 2.0 A 0.3 0.3 0.3 0.00 0.00 0.00 -- l -- l 03600880 8 #83 mo £08; 00 0 + 0 0H 0 H0 0 00 0 0H 0 0 0 00 -- .. -- -- 323 moammuwafl 00.0 A 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 $3300.80 0.5050 00.0 A 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 $500.80 4508.0 00.0 A 0.00 0.00 0.00 0.00. 0.00 0.00 0.00 0.00 0.00 0.00 £0.80 080 0:01 00. A 0.3 0.2 0.: 0.3 0.00 0.3 0.2 0.: 0.: 0.00 £080 .80 00.0 A. 0.3 0.3 0.00 0.00 0.8 0.00 0.00 0 0.00 0.00 £080 :00. 50.0 H 0.000 0.0: 0.000 0.00 0.000 0.02 0.000 0.000 0.00 0.00 A50 5080 .008 09.. 08: 24 u 0.00 0.00 0.00 0.00 0.00 0.00 0.8 0.00. 0.00 0.00 30 ”£ng 2 z 0 w w o x z z s8008000mmflo uw< 00 A. m 30.2 £02 30: 30.8.0 £0.80 30.8.0 3...: 30: 30.8.0 8000 0.me SE 005 823m 330 850 800 202 mz00 000000000 000 0008 009000 00 0000000000000 500.0 000000000 003 00000030 000 00900000 00 00000 00800900: 0000000.. .0003 000800 8000 009000 00000 000.0% 03.00 0.003 .. 0.00 0.00 0.00 0.00 .: 0.000 @0330 .500 03.0 0.30 2. 0.0.0 0.00 0.3 0.00 -- 0.30 00.03500 0000 00.0 0.3 0.3 0.03 0.3 2. 0.2 0.0 -- 00.0 0.53330 00.0 0.00 0.00 0.00 0.3 -- 0.00 0.00 -- 00.0 0 30330 :3 0.3 0.3 0.03 0.3 2. 0.3 0.03 -- 00.0 00 30830 00.0 0.03 0.3 0.0 0.: .. 0.3 0.0 -- 00.0 00 33330 00.0 0.00 0.3 0.00 0.: -- 0.00 0.00 -- 00.0 @502 00. 0.0 0.0 0.0 0.0 -- 0.0 0.3 -- Se 00:00 00380 380.0 3 Z 9 3 mm m W: m. m m m. .m. M. 800E008 m; _.L B ) \oz. ) \w. m ) D. {K \I 9. 1K .0 l.\ O. /\ ) {\ ( w 0.0000003 000008 00>00 000 m0300> 0500 00008050000 .0 590.0. 45 hematocrit and erythrocyte count for wolves of the Serra da Canastra is corroborated by the fact that similar determinations from wolves in the National Zoological Park, Washington (Bush, in press), resulted in mean corpuscular volumes above the mean for normal dogs. Numbers of mature and immature neutr0phils were within limits for normal dogs for all tested wolves except Delta. The relatively high percentages of band and seg- mented cells in the blood of this individual may indicate that an inflammation or suppurating wound was present at the time the blood sample was collected. In fact, Delta was suffering at that time from severe fly strike in one inner ear and surrounding tissues. This infestation, per- haps accompanied by a secondary bacterial infection, was possibly responsible for the elevated neutrophil count. With the exception of maned wolf Beta, eosinophil counts were normal, as judged by dog standards, for all tested wolves. The phagocytic role of eosinophils is often involved in antigen-antibody reactions to parasitic infestation or re-exposure to a foreign protein (Rich, 1974). The fact that four speciescyfnematode larvae were found in the feces of Beta lends some support to an aller- gic reaction to endoparasites as the explanation of this eosinophilia. Parasitization by the giant kidney worm, common in maned wolves, might also have contributed to an elevated eosinophil count if that parasite were present in this wolf. The presence of this parasite could not have 46 been determined by the diagnostic technique used in this study. Lymphocyte counts were above normal levels for dogs only for the two wolves residing in the Belo Horizonte Zoological Garden. Lymphocyte function is involved in chronic immune responses, and elevated counts often indi- cate lymphosarcoma or recovery from an infection. I have no indication as to what may have caused the elevated counts in these zoo animals. However, it is interesting to note that the lymphocyte counts for the captive maned wolves tested by Bush are also higher than values for nor- mal dogs. With the exception of maned wolf Epsilon, total pro- tein in blood serum samples from wolves tested in this study was within established limits for normal dogs. The cellular component of the blood sample from Epsilon had apparently hemolyzed prior to removal of the serum frac- tion and was therefore not considered in calculation of group means. The a1, 02, and y globulin components of the serum protein from tested wolves were within the ranges for nor- mal dogs. However, although the normal values of albumin and B globulin are respectively about 50 percent and 25 percent in dogs, these values in tested wolves averaged about 23 percent and 39 percent. If these values are not artifactual, then eithertfluanormal values for these serum proteins in maned wolves are different from those of the 47 normal values in dogs, or an extrinsic factor such as chronic infection caused a mobilization of protein from albumin to B globulin in all tested maned wolves. How- ever, the consistency of these values, as indicated by the small standard deviations for these proteins (2.22 and 3.20, respectively), tends to refute the latter explana— tion. Serum glutamic pyruvic transaminase (SGPT) and serum glutamic oxalacetic transaminase (SGOT) were measured by the Reitman-Frankel procedure in each of five wolves. Although SGPT values from tested wolves were higher than those for normal dogs, the consistency of the values from tested wolves (SD = 4.18), and the fact that the values for these enzymes often vary significantly among the laboratories performing the analysis suggests that this difference may be an artifact of laboratory procedure. In normal dogs, SGPT and SGOT values are of about equal magni- tude. Maned wolves Gamma and Lambda showed SGOT values moderately greater (45 and 37 percent, respectively) than SGPT values; the SGOT value for Delta was 149 percent greater than the SGPT value. Myositis associated with capture and restraint could have contributed to the ele- vated SGOT in these wolves. The very high SGOT value for Delta could be associated with muscle damage resulting from the fly strike in her inner ear and surrounding tis- sues . 48 Description of Home Ranges in the Study Area During the two years of this study, I captured eight maned wolves in the Serra da Canastra and three wolves elsewhere. Radiotelemetric and visual observations of the eight wolves residing in the study area allowed the definition of three mutually exclusive home ranges. I named these ranges the West Range, the East Range and the Taperao Range. Portions of each range were located within the boundaries of the Serra da Canastra National Park. Each range was occupied by a single male-female bonded pair of maned wolves. West Range The West Range was located northwest of my field headquarters and east of the village of sac Joao Batista (Figure 5). Connecting peripheral points on a map of plotted radiotelemetric locations yields a polygon with an area of 21.7 km2. This estimate, however, is probably smaller than the area actually used by the resident pair of wolves. This is because one series of ridges and val- leys, measuring about 4 kmz, was so difficult to reach from my field headquarters that I never obtained more than one radiotelemetric vector from a wolf in that area. Thus, I was unable to quantify the extent to which that area was used by wolves of the West Range. 49 FIGURE 5. Locations of home ranges of maned wolves in the study area. .m 000w0m II|| I I I ll / .2020: 0:00:00 00 / / ”5&0” / l / \ \ II‘ «tab .02.; 00:00:30.. 00:08sz 00:00:32. 5505. I mm02> amz<2 51 East Range Directly northeast of my field headquarters and adja- cent to the West Range was the East Range (Figure 5). The pairs of wolves that successively used this range were more easily located than the wolves of the other two ranges. Therefore, my estimate of the area of this range, as based on radiotelemetric locations, is perhaps the most accurate of the three under consideration. Connection of the peripheral points on a map of radiotelemetric loca- tions for wolves using this area show that the total occupied area was 30.0 kmz. Taperao Range The Taperao Range (Figure 5) was located about 2 km southwest of the city of Sao Roque de Minas, and about 28 km east of my field headquarters. Ranchers and farmers were particularly knowledgeable about the movements of the maned wolves in the Taperao Range and declared that the extent of the movements of the resident wolves had remained unchanged for many years. With the help of these farmers, I easily located paths frequently used by wolves and trapped male maned wolf Beta on 10 September 1978. On many occasions, Beta was seen in the company of a small female wolf. Although Beta was recaptured six times dur- ing the course of the study, I was never able to trap and put a transmitter on his mate. 52 The region of the Taperao Range was characterized by extremes of topographic relief. Only one road was usually passable by automobile. Any of a number of large ridges effectively blocked the signal from Beta's transmitter. Therefore, the area of his home range, 12 km2 as calcu- lated from radiotelemetric data, underestimates the area actually used by this wolf. A more realistic estimate, based on single radiotelemetric vectors and visual sight- ings is about 24 km2 for the Taperao Range. The problems with radio reception in this area, the long distance from my field headquarters, and the fact that I never succeeded in capturing a female wolf there combined to make the Taperao Range unsuitable for attaining several of the objectives of my study. I therefore elected to spend the majority of my limited temporal and financial resources in the West and East Ranges. With the exception of scat col- lection, data were collected only occasionally or oppor- tunistically in the Taperao Range. Other Ranges In addition to the three home ranges previously men- tioned, I was also aware of several other ranges in the region (Figure 5). One such range abutted the West Range to the north, and a second range was adjacent to, and east of, the East Range. Between this second range and the city of 850 Roque de Minas was a third range, while a fourth existed to the south of $50 Roque and east of the 53 Taperao Range. Because of the problems associated with access to these additional ranges, I made no attempts to capture their resident maned wolves. I have little infor- mation on how many wolves occupied these ranges other than the facts that pairs of wolves were sighted in three of the four ranges, and reproduction took place in all four ranges. Reports by area residents of the numbers of wolves vocalizing in these areas, and my own subjective estimates of the amounts of wolf-sign present, suggest that two wolves probably occupied each of these additional ranges. Use of Available Habitat Types by Maned Wolves I determined the extent to which available habitat types were used by the maned wolves of the Serra da Canastra by first establishing criteria to distinguish functionally between the various vegetation types in my study area. Overstory density, composition and height were used to classify habitats either as grassland, cerrado or forest. I defined grassland habitat as areas dominated by annual grass species up to 1 m in height, and with only occasional trees or shrubs. Cerrado was char- acterized by an open, woody overstory up to 5 m in height and composed of small, tortuous trees or shrubs. I defined forest habitat by the presence of a closed over- story composed largely of trees greater than 5 m in height. The relative percentages of these habitat types in the 54 study area as previously mentioned, were 50, 35 and 15 respectively. The following questions of interest were addressed: (1) What are the relative frequencies with which maned wolves use these available habitat types? (2) Do these frequencies differ between sexes of wolves? (3) Do these frequencies differ between pairs of wolves? (4) Do these percentages vary with seasonal change? (5) Are habitat types used in pr0portion to their occurrence? Tabulation of 458 radiotelemetric locations (Table 5) for which habitat types were known indicated that maned wolves spent approximately 34 percent of their time in grassland, 43 percent in cerrado and 24 percent in forest. These data were transformed to reflect equal sample sizes for both sexes, and a chi-square test of goodness of fit was applied to test for differences between sexes in the use of each habitat type. Results of these tests suggest that male maned wolves used grassland habitat significantly more than did females (p < .05), and that females used cerrado habitat significantly more than did males (p < .025). No significant difference between sexes was found in use by maned wolves of forest habitat. Sufficient radiotelemetric observations (n = 458) were made to allow the comparison of habitat use between three bonded pairs of maned wolves (Table 6). G-tests of independence (Sokal and Rohlf, 1969) were performed for each habitat type. Results indicated that these three 55 0.000 000 0.000 mg 0.80 wqm 12.090. m . mm m00 m . 0N 0m 0 . mm mm 00000.0 0 . N0 mom w . 3 000 m . mm mm 00000000 0 . mm 80 m . mm 00 0 .mm 8 0000000000 00000090 A5 0000090 A5 0000000 A5 000000000 000000000 000000000 logmm logmm Iogmm NE “MUN—”g 005.090 0.00 mg mg .00>003 00:05 .8. 000.00 000000: 00050 00 0000 .m ”500.90. 56 o . 000 00 o .000 mm m . mm own 0009. N00 om 0.0m 0m 0.00 No 000000 06m «N 0.00 00 93 $0 0000000 0.0m R mdm . 00 0.0m «00 00000000 0000000 AOV 0000000 A5 0000000 00V 000000000 0000 00000 000000000 00000 00000 00000 0000. 0000000 m 0000 N 0000 . 00000 000000 00 0000000 00>003 00008 00 00000 0000000 00 003 .0 0.00000. 57 male-female pairs were not independent of the use of cerrado (p < .005) or of forest habitat (p < .005). Seasonal differences in habitat use were determined by dividing the radiolocations for which habitat type was known into those occurring during the wet and dry seasons (Table 7). Chi-square tests were used to test the null hypothesis that there was no seasonal difference in the use of each habitat type of these wolves. Results of these tests indicate that grasslands were used more often during the wet season (p < .025), and forest habitat was used more often during the dry season (p < .025). A pos- sible biological significance of these data is that grass- land habitat may support a lower relative abundance of prey species during the dry season when pastures are burned by local ranchers. The preceding analyses are biased to some extent in two ways: first, the majority of the radiolocations made during this study occurred during the day. Thus, because these wolves were largely nocturnal (see the following section of this dissertation), habitat types associated with daytime rest sites would tend to be overrepresented. Second, although the three defined habitat types did not differ among home ranges with respect to relative avail- ability, those three habitat types were not equally abun- dant in the study area. Therefore, the contribution of smaller habitat types would be overemphasized in the above analyses. In an attempt to compensate for these biases, 58 063 :3 063 Now 0.03 mmm ASE. m.m~ m2 NS 3 m .3. mo ummnom odq now 5.3 8H m .3 3H opmfiwo m . mm m3 H .3 8 o . R 8 @5335 unwoumm A5 ”Emohom A5 unmouom A5 9.5383 $83.83 $53.33 9m low—Hug Iofiumm 10.25% m .H umuwflmm mzomfim 4.2 mzomHg Eng: %D mug UNUHQQS @993 MO mmfl ngmwmm .N g 59 I divided those radiolocations for which habitat type was known, into those recorded during the day (1601-1800 hours), and those observations made by night (1801-0600). A correction factor was applied to adjust for differences in relative abundance of the three habitat types. I placed a confidence interval using the Bonferroni z-statistic (after Neu et al., 1974) on the observed pro- portion of occurrence of wolves in each habitat type to test the hypotheses that habitat types were used in direct proportion to their occurrence (Table 8). Results of this analysis suggest that maned wolves of the Serra da Canastra were less likely to be found in grassland habi- tat than in other types of habitat at any time. During the day, wolves were most often found in cerrado; at night, they were most commonly located in forest habitat. Activity Patterns Daily and Seasonal Variation in Activity The daily and seasonal patterns of temporal activity of the maned wolves of the Serra da Canastra were deter- mined in two manners: (l) classification by time block and activity level of daily locations of radiocollared wolves; and (2) periodic monitoring _._.0< .l m>_._.0< >4hzwhtimmhz. / g m>_._.o._._>_._.0< njO>> sz<_>_ 10v [8 loo. . o ouawwh SNOILVAUBSSO :IO .LNSOHad 63 contributed by many consecutive locations of the same wolf during a single time period by randomly selecting a maxi- mum of two locations per wolf per time block for inclusion in this analysis. The activity-monitoring devices incorporated into the transmitters worn by maned wolves in the study area con- sisted of glass tubes each containing a drOp of mercury. When the mercury was stationary at one end of the tube, the transmitted pulse rate was about 50 signals per minute. Alternatively, when the mercury closed the circuit at the opposite end of the tube, the transmitted signal rate was approximately doubled. The high and low pulse rates dif- fered between collars and were therefore useful in identifying individual wolves. When wolves were physi- cally active, the drop of mercury vacillated between contact points thereby producing a sync0pated pulse rate highly distinct from that of a transmitter on an inactive wolf. In most cases, the two activity classes mentioned above were distinguishable from intermittent activity, the activity class in which wolves made only occasional postural changes or other erratic movements. Radiotelemetric data collected during the hours of daylight (0601-1800) indicate that radiocollared wolves were active during 34.7 percent of this time period, or, 35.0 percent of the six hours prior to noon and 33.8 per- cent of the six hours following noon. The majority (53.8 percent) of these observations of wolves active during the 64 day were classified as intermittently-active, and the remainder (46.2 percent) as constantly-active. That is, wolves spent approximately 115 minutes travelling from place to place during the hours of daylight. The maned wolves of the Serra da Canastra were con- siderably more active at night than during the day. Ninety-one percent of the radiotelemetric observations made between 1801 and 0600 hours were of constantly- or intermittently-active wolves. Between the hours of 1801 and midnight, 96 percent of the radiolocations of active wolves were classified as constantly active. During the six hours following midnight, 48 percent of the radio- locations were of intermittently-active wolves and 52 per- cent were of constantly-active wolves. In other words, wolves of the Serra da Canastra spent about eight hours (five hours before midnight and three hours after midnight) moving from place to place at night. Although a chi-square test of independence did not indicate a dependent relationship between level of activ- ity and four male-female pairs (p = .33; Table 9), the same test did suggest a possible relation between activity and sex (p = .012; Table 10). Further, when numbers of intermittently-active, and of constantly-active observa- tions were pooled to form a single class for each sex, a G-test of independence indicated a significant level of dependence between activity and the sex of these wolves (p < .01). Again, using combined numbers of observations 65 .8. u a r: o .30 u was, 356.09. 0.03 mm 0.03 003 0.03 on 0.03 00¢ A433. 0.3 0 0.NN mm 0.00 3 N63 00 ”083950 ¢.NN 3 0.03 mm 0.3 n N6m 00 nauudauoufi H60 00 0.00 33 0.00 mm 0.30 ~00 mfiuomfi unmonmm A8 ”.50qu A5 unmoumm A8 unmohmm A5 moon.” ”.803 983803 95.." 8803 303803 96.3 fidfiuo -03 -330 -330 .0305 H . . < 0 am 0 320 N an 3 ~33 «6.3mm 8950 E 39.on mot/HQ; 0mg: “imam mo 33300 mo 0396.3 .0 H505. 66 .H0. v a . .mfi H .mmé u mama. uwmuuu 455980 «Hm mcouflmawmno m>Huum $383800 98 m>Huom$HucmuuHfiumucH 5:3 .NH0. u m . .90 N .000 u 8Hm> ummu mac??? 0.03 «on 0.03 Rm 0 .03 50¢ 45.8. m . mH HmH 0.3 mm N.0H 00 ucmumaoo H . 0H 3H N . 0N Nu N . 0H «m unmuufifioufi H .3 Sq 0 . 0m 00w 0 .3 now 9.38.5 uaoumm A5 ucmonmm A5 ufimoumm A5 @0383 0 95383 30383 .A IOHUNM IOMUMM IOHUNM H93 Hunk/Hug mm>§ 4?. mg mg «48m An 0&5on mm>H03 0988 uanm mo 3.33% mo mHm>QH .0H mama? 67 of constantly-active, and intermittently-active wolves, females were active 34.4 percent of the time and males were active 44.0 percent of the time. Observations on Individual Activity In addition to radiotelemetric observations across blocks of time, data collected during 16 monitoring ses- sions of from two to six hours in length added descriptive detail to the picture of temporal relationships among maned wolves. During daylight hours, the wolves rested in thick cover -- often in riparian forest or in the grass and shrubs at the headwaters of a small stream. Individuals usually remained in the same patch of cover or moved only a few meters during the day. At intervals of about 45 minutes, however, each wolf could be expected to move for a few seconds before again settling down. Although I never saw a wolf perform these brief movements, the trans- mitted radio signals suggested postural changes or perhaps standing up and turning around before again lying down. Shortly after sunset (between the hours of 1750 and 1845), the wolves left the areas where they had been resting dur- ing the day and began to hunt and travel throughout their ranges. These movements continued for much of the night with occasional pauses of up to about 20 minutes. On some occasions, observed wolves travelled from one extreme to another of a home range with only brief stops. 68 Alternatively, an entire night might be spent hunting in a small area. Usually by 0700 hours, but occasionally as late as 0840 hours, the wolves again retired to thick cover for the day. Although I did not statistically test the relationship between weather conditions and the activ- ity patterns of maned wolves, wolves appeared to remain active longer during rainy or foggy days than during days of bright morning sunshine. Social Organization and Communication Relationships of Individuals Within Pairs The maned wolves of the Serra da Canastra displayed facultative monogamy (see Kleiman, 1977) characterized by a long-term pair bond between male and female. Although the male and female, as a pair, had completely overlapping ranges, very little time was invested in close association. According to local observers, family groups were rarely seen. It is noteworthy to mention that of the three per- cent of mammal species described as monogamous, only 13 species demonstrate facultative monogamy (Kleiman, 1977). This type of monogamy has not previously been described for any species of the Carnivora. The above findings are based on 1,537 trap-nights, 828 precise radiotelemetric locations, nearly as many single directional vectors, and approximately 50 visual sightings of eight maned wolves in the Serra da Canastra study area. Traps placed in locations featuring an 69 abundance of spoor resulted in the capture of eight wolves on a total of 47 occasions (Table 11) in the Serra da Canastra. One female and her two pups were taken in another area and will be discussed separately. Using daily radiotelemetry, I was able to estimate the bound- aries of home ranges of the wolves in the Serra da Canastra. As estimates improved with increased numbers of observations, they are therefore likely to underestimate the areas used by wolves tracked for a short period of time. Trapping efforts were initiated in September 1978. By May 1979, at least one wolf from each of three home ranges had been captured and fitted with a radiocollar. Beta This wolf, a middle-aged male, was captured in the Taperao Range on 10 September 1978 (Figure 7). He was observed on numerous occasions in the company of a small female wolf that apparently shared that range. Due to previously discussed problems intrinsic to that geographic area, trapping efforts there were discontinued shortly after Beta was captured and radiotelemetric location of that wolf was attempted only occasionally. Little can be said about these wolves other than that they shared part if not all of the same home range. I estimate the area of this home range to be about 24 kmz. During the 20 months that Beta carried a transmitter, he was observed either alone or with the same female wolf. Bums 95“. EB wags mangam gm H83 m. .3 8:? $3 303 a? .me 8x05 m €008 mo £28» m 9... 8% 303 mg 6on gum mfi ummH HmEHdm on... was”. Home—6mm mg 5 .oumfi mg no MHDS 3.5 £55.. umoH mg pomucoo 3.53038vame 70 N 83 .82 mm RS .82 m: 2m: saw ”.33 H :35 .82 8.. 8 ES .30 2 R2 32. m 2...: 3mm name .6 82 .8: mm ES .Hz 3 Same mg 0%: mm 83 .3: 3 E3 was 2 3m: 895 “mam n 83 mm: mm RS ...a< «N £98m 82am 03m N .355 Ag mm ES .93. 3 mamas «38 3mm e .55 «5. mm 33 d8 3 3m: 280 name a 83 a: «N m5 gem 3 3m: 33 885. 8938 mo 5.8m mo 0800 H3005 “am 982 mg 3952 .8 8383 we 88 £03 85: #3 mo 88 .mwwqmu m8: 853 an" wo>Hos 095: How moumv mhfiamo .HH was. 71 FIGURE 7. A chronological representation of the rela- tionship among wolves inhabiting the East, West and Taperao Ranges. Symbols: C = the date of the initial capture of a maned wolf; D = the date by which a wolf was known to have died. 'Maned wolf Zeta died some time during the period indicated by the dotted line. ‘ 72 E3 v u ou> a. Z a. a o h o v a :> a. Z 3:? in J: 605: omcoamh J: =—..... *9 9 6931625 A I: ill: l". mu J: 605m «mam .n mustm -88.: .3. bumo .sH muswfim a£p u . 20mma 20m