|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 SOCIAL ECOLOGY OF THE WHITE-LIPPED PECCARY ( Tayassu pecari ) IN CALAKMUL FOREST, CAMPECHE, MEXICO By RAFAEL ANGEL REYNA HURTADO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
2 2007 Rafael Angel Reyna Hurtado
3 To Aranza y Edith that represent th e best and the beauty of my life To my mother which is a biologist in her hearth To my father who did always love nature To the people of Calakmul
4 ACKNOWLEDGMENTS I am deeply indebt with my advisor George W. Tanner PhD for his continuous support of my professional development. He was involved in all the stages of my dissertation from writing the proposal document to the final editorial work. I really appreciate the time and effort he put on this work. The members of my committee Drs. Lyn Branch, Colin Chapman, Sophie Calm, Eduardo Naranjo and Mary Christman did a great j ob and they were very helpful and with their advise I was able to further refine the chapter into potential published papers. Edith Rojas-Flores, my wife helped me w ith the peccary capture protocol and worked also in the field for several weeks during th e capture phase. I was lucky to work with a exceptional field guide, Nicols Ar ias Dominguez from Nuevo Becal ejido Nico and I set up the main study site, open trails, captured pecc aries (with the help of Edith Rojas), followed peccaries, camped, cooked, biked and walked a lot of km under the forest for the first year of my study time. His help, expertise, and company were invaluable for the beginning of the project. Latter I worked with Gilberto Arias which proved to be a very helpful person as well and very knowledgeable of the tropical forest. Ernesto Gu terrez and Hermilo Ramrez were also very helpful several times when the job requi res participation of more people. The mayors of three ejidos (Nuevo Becal, 20 de Noviembre and Xbonil) gave me permission to enter their lands and conduct the fiel dwork; also they provid ed me useful data on the history and the social components of their ejidos Fernando Durand Siller, the Director of the Calakmul Biosphere Reserve, and his personnel were very kind in giving permission to enter the Reserve and allowing me to use the facilities in Calakmul. Workers from the INAH (National Institute of the History and Anthropology Studies) al lowed me to use their installations while the fieldwork was centered around the old Calakmul Mayan city. Gerardo Garcia-Gil (ECOSUR) provided the Calakmul Biosphere Reserve map used for this study.
5 This project was possible thanks to the generosity of CONACYT (Mexicos National Council of Science and Technology) from whom I had a scholarship (number 150332) from (CONACYT) that allowed me to study at the Univ ersity of Florida for four years from August 2002 to August 2006 to pursue the doctoral degr ee. The Wildlife Conservation Society (WCSResearch Fellowship Program) funded the logisti cal costs of the fieldwork thorough the Research Fellowship Program. IDEA WILD pr ovided invaluable field equipmen t. I was able to flight and locate groups of peccaries with a volunteer pilot from Light Hawk organization. The Department of Wildlife Ecology and Conservation of the Univ ersity of Florida through my advisor George Tanner, was very kind in provi ding some equipment necessary in the field. The Tropical Conservation and Development Program (TCD) from the Center of Latin American studies of the University of Florida also provided me fi nancial support to work on the data and write the dissertation. I want to thank to several people of Calakmul ejidos that were very kind to answered questions and helped me in different ways during the course of the field work. Finally, I appreciate the compa ny of my wife Edith Rojas and my daughter Aranza. Edith was always supported of every idea I had on mind. Edith and I lived two be autiful years in Zoh Laguna village. During that time our daughter Aranza was born. After that, the beautiful and relaxed life we have had until then changed and just become much better!!
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......11 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION..................................................................................................................15 Social Behavior in Ungulates with Emphasis on Peccaries....................................................15 The White-Lipped Peccary ( Tayassu pecari ).........................................................................16 Description.................................................................................................................... ..16 Ecological Features.........................................................................................................17 Distribution and Status....................................................................................................17 Current Knowledge.........................................................................................................18 The Great Calakmul Region...................................................................................................20 2 HOME RANGE AND HABITAT SELECTION OF WHITE-LIPPED PECCARY GROUPS ( Tayassu pecari ) IN A SEASONAL TROPICAL FOREST OF THE YUCATAN PENINSULA, MEXICO....................................................................................24 Introduction................................................................................................................... ..........24 Material and Methods........................................................................................................... ..26 Study Area..................................................................................................................... ..26 Study Design...................................................................................................................28 Capture........................................................................................................................ ....28 Radio-telemetry...............................................................................................................29 Home Range Data Collection..........................................................................................30 Habitat use Data Collection.............................................................................................30 Food and Water Availability...........................................................................................31 Analysis of Home Range Data........................................................................................32 Analysis of Habitat Use Data..........................................................................................33 Results........................................................................................................................ .............35 Home Range....................................................................................................................35 Habitat Use.................................................................................................................... ..38 Food and Water Availability...........................................................................................39 Discussion..................................................................................................................... ..........41 Home Range....................................................................................................................41 Habitat Use.................................................................................................................... ..44 Conservation Implications...............................................................................................47
7 3 DO MOVEMENTS AND GROUP SIZE IN WHITE-LIPPED PECCARY ( Tayassu pecari ) FIT THE PREDICTIONS OF THE ECOLOGICAL CONSTRAINT MODEL?.....55 Introduction................................................................................................................... ..........55 Methods........................................................................................................................ ..........57 Study Area..................................................................................................................... ..57 White-Lipped Peccary Groups Locations and Behavior.................................................58 Food and Water Availability...........................................................................................60 Group size and Precipitation...........................................................................................61 Statistical Analyses..........................................................................................................61 Results........................................................................................................................ .............62 Discussion..................................................................................................................... ..........66 Habitat Quality................................................................................................................66 Water Dependence...........................................................................................................67 Group Sizes and Rainfall.................................................................................................70 4 SEARCHING STRATEGIES OF WHITE-LIPPED PECCARY ( Tayassu pecari ) GROUPS IN A SEASONALLY-DRY TROPICAL FOREST..............................................78 Introduction................................................................................................................... ..........78 Material and Methods........................................................................................................... ..80 Study Area..................................................................................................................... ..80 White-Lipped Peccary Data Collection...........................................................................81 Distribution of Resour ces on the Study Area..................................................................82 Lvy-walk Pattern...........................................................................................................83 Central Place Foraging....................................................................................................85 Travel Behavior...............................................................................................................85 Results........................................................................................................................ .............86 Distribution of Resources................................................................................................86 Lvy-walk Pattern...........................................................................................................87 Central Place Foraging Model.........................................................................................88 Coordinate Travel Behavior............................................................................................88 Discussion..................................................................................................................... ..........90 5 HUNTING PATTERNS, POPULATION DE NSITY, GROUP SIZE AND THE CONSERVATION OF WHI TE-LIPPED PECCARY ( Tayassu pecari ) IN THE CALAKMUL REGION OF MEXICO...................................................................................98 Introduction................................................................................................................... ..........98 Methods........................................................................................................................ ........100 Study Area.....................................................................................................................100 Group sizes, Age Structure and Breeding Season.........................................................101 Density Estimation........................................................................................................103 Seasonally of the Hunting Patterns................................................................................104 Statistical Analyses........................................................................................................104 Results........................................................................................................................ ...........105 Group Size in Hunted and Non-Hunted Areas..............................................................105
8 Group Age Structure......................................................................................................105 Breeding Season............................................................................................................105 Seasonality of Hunting..................................................................................................106 Density Estimations.......................................................................................................106 Discussion..................................................................................................................... ........107 Group Size.....................................................................................................................107 Age Structure and Breeding Season..............................................................................109 Hunting Seasons............................................................................................................109 Density........................................................................................................................ ...110 Conservation Implications.............................................................................................111 6 CONCLUSIONS..................................................................................................................118 Social Ecology of Wh ite-lipped Peccary ( Tayassu pecari )..................................................118 Conservation of White-Lipped Peccary in Calakmul Region..............................................121 LIST OF REFERENCES.............................................................................................................123 BIOGRAPHICAL SKETCH.......................................................................................................131
9 LIST OF TABLES Table page 2-1 Home range sizes (km2) as estimated by three methods of the four groups of whitelipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico..................................49 2-2 Home range sizes (km2) for dry and rain season of 2005 and 2006 using Fixed Kernel with 95% of the observations fo r the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico............................................................49 2-3 Overlap among the home range of the f our white-lipped peccary groups during the dry and rain season of 2005. Calakmul Bi osphere Reserve, Campeche, Mexico.............49 2-4 Compositional analysis of the study ar ea and home ranges (second order selection Johnson 1980) for the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico..............................................................................................50 2-5 Compositional analysis of MCP home ranges and usage (third order selection Johnson 1980) of forest types for the four white-lipped peccary groups. Calakmul Biosphere Reserve, Campeche, Mexico............................................................................50 2-6 Compositional analysis of seasonal MC P home ranges and usage (third order selection Johnson 1980) of forest types for the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico............................................................51 3-1 Mean group size, monthly area used (km2) and test for differences among groups for both variables using analysis of varian ce on groups of white-lipped peccary in Calakmul Biosphere Reserve, Mexico...............................................................................72 3-2 Models selected developed by all-ti me and 2005 study periods of the factors affecting area used by groups of white-l ipped peccary in Calakmul Biosphere Reserve, Mexico................................................................................................................72 3-3 Selection of models explaining the area used by WLP in Calakmul Biosphere Reserve, Mexico, using Akaike Informati on Criterion for small sample sizes AICc using Maximum Likelihood Estimation Method...............................................................73 3-4 Mean and standard errors of the pro portions of the three vegetation types in the monthly home ranges of white-lipped peccary groups in Calakmul Biosphere Reserve, Mexico................................................................................................................73 3-5 Group sizes of white-lipped peccary as re ported along the range of the species, and associated annual rainfalls. Sources are orde red in increased order of annual rainfall.....74 4-1 Regression slopes coefficients ( ) for log10 of frequency of hourly and daily displacement length versus log10 of 2k logarithmic bin with normalization, staring at
10 20, 50 and 100 m for hourly data, and 100 a nd 200 m for daily data, for groups of white-lipped peccary ( Tayassu pecari ), Calakmul Biosphere Reserve,Mexico................95 5-1 Group size, group age composition and area classification for white-lipped peccary groups observed in Calakmul Region, Campeche, Mexico.............................................113 5-2 Proportions (%) of forest types for the ar ea used and shared by the four groups and the southern area of the Calakmul Bi osphere Reserve, Campeche, Mexico...................114
11 LIST OF FIGURES Figure page 1-1 Localization of the Calakmul Biosphere Re serve (in dark gray) and study area (white oval) in Mexico (light gray)...............................................................................................23 2-1 All seasons fixed kernel (95 and 50 % represented as to tal area, and circles inside respectively) home ranges for the four groups of white-lipped peccary in the Calakmul Biosphere Reserve, Campeche, Mexico............................................................52 2-2 Cumulative home range sizes (km2) using MCP for the four groups of white-lipped peccary in the Calakmul Biosphere Reserve, Campeche, Mexico....................................52 2-3 Fixed kernel home range estimation for the dry season of 2005 for the four groups of white-lipped peccary in the Calakmul Bi osphere Reserve, Campeche, Mexico...............53 2-4 Fixed kernel home range estimation for th e rain season of 2005 for the four groups of white-lipped peccary in the Calakmul Bi osphere Reserve, Campeche, Mexico...............53 2-5 Monthly fruit abundance index (# occurre nce/km) on the forest floor for the four most abundant species in the Calakmul Biosphere Reserve, Campeche, Mexico.............54 2-6 Water availability for the study period (rainfall in mm, water in flooded forest and Calakmul pond has two values, 100=prese nce, 0=absence) in the Calakmul Biosphere Reserve, Campeche Mexico.............................................................................54 3-1 Water availability for the study period in Calakmul Biosphere Reserve, Mexico. Rainfall is in mm and presence of water in flooded forest and Calakmul pond has two values, 100=present, 0=absent....................................................................................75 3-2 Index of abundance per km of the thr ee most abundant fruit species for the study period. Calakmul Biosphere Reserve, Mexico..................................................................75 3-3 Linear relationship between area used and group size, % of flooded forest in home range for the all time study peri od, and between area used and B. alicastrum abundance index for the dry-wet 2005 peri od, for groups of white-lipped peccary in the Calakmul Biosphere Reserve, Mexico.........................................................................76 3-4 Linear relationship between rainfall a nd group size of white-lipped peccary for 11 sites from Mexico to Argentina al ong the species distribution range................................77 4-1 Minimum convex polygon of all the local izations from the four groups of whitelipped peccary ( Tayassu pecari ), and forest type patch distribution in Calakmul Biosphere Reserve, Mexico...............................................................................................95 4-2 All-seasons frequency of steps length per hour with a bin size of 100 m, for whitelipped peccary................................................................................................................. ...96
12 4-3 Regression slope for all-seasons hourly data with a bin size of 100 m and 2k logarithmic binning with normalization method................................................................96 4-4 All-seasons frequency of steps length per day with a bin size of 200 m, for whitelipped peccary................................................................................................................. ...97 4-5 Regression slope for all-seasons daily data with a bin size of 200 m with 2k logarithmic binning with normalization method................................................................97 5-1 Calakmul Biosphere Reserve and site s where white-lipped peccary groups were sighted in Calakmul region, Campeche, Mexico.............................................................115 5-2 Median of white-lipped group size record ed in the Calakmul Biosphere Reserve (Non-Hunted Area) and four sites (Hunt ed Areas) where hunting and other human activities are taking places in Calakmul region, Campeche, Mexico..............................115 5-3 Age structure observed in groups of wh ite-lipped peccary in the Calakmul Region, Campeche, Mexico..........................................................................................................116 5-4 Seasonality of the birth events observe d in groups of white-lipped peccary in the Calakmul region, Campeche, Mexico..............................................................................116 5-5 Area used and shared for four groups wher e density was estimated (encircled by light green line) and area where density was extra polated (encircled by the blue line) of white-lipped peccary in the southern area of the Calakmul Biosphere Reserve, Campeche, Mexico..........................................................................................................117
13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOCIAL ECOLOGY OF THE WHITE-LIPPED PECCARY ( Tayassu pecari ) IN CALAKMUL FOREST, CAMPECHE, MEXICO By Rafael Angel Reyna Hurtado December 2007 Chair: George W. Tanner Major Department: Wildlife Ecology and Conservation The white-lipped peccary (WLP) is a social ungulate that forms the largest groups documented for any tropical fore st ungulate species. In the la st 20 years WLP have become increasingly rare in Mexico and Central Ameri ca and several researchers have suggested that more frequent reporting of smaller groups is re lated to an increased hunting pressure. Here I studied patterns of movement, ho me range size, and habitat pref erences of WLP in a seasonally dry tropical forest, the Calakm ul Biosphere Reserve (CBR) in Yucatan Peninsula, Southern Mexico. Additionally, I presented information rele vant to the conservation of this species like group sizes and structure, breed ing season, population density a nd hunting patterns of WLP for the Calakmul region of southern Mexico. Home range was among the largest reported ever on the literature for this specie s. White-lipped peccary preferre d ponds, medium semi-perennial forest and low-flooded forest while avoided lo w-dry forest. Peccaries moved over a large range in the wet season and they performed long movements seasonally. WLP movements are influenced by water availability at least during the dry season and they behave like central place foragers around water sources during this ti me of the year. Groups performed searching strategies that resembles the Lvy-walk movement pattern, as well as coordinate traveling. These
14 strategies seem to be very efficient for sear ching resources that are non-uniform distributed on the landscape like was demonstrated for the CBR. Groups of WLP were on average larger in the protected area than in four hunted sites, but groups were generall y smaller than those reported in other forests along the geographic range of this sp ecies. The smaller groups in the hunted areas are a detrimental signal for the conservation of this specie s in the Calakmul region. Population density is smaller than some estimates being us ed for legal hunting purposes and subsistence as well as sport hunting occurs on the dry season main ly when this species is breeding and visiting the water-bodies, and consequently is more vulnerable.
15 CHAPTER 1 INTRODUCTION Social Behavior in Ungulates with Emphasis on Peccaries There are several explanati ons to why animals live in groups. The advantages of group living have been broadly classified into thr ee main categories, avoiding predation, optimizing resource use, and recently, avoidance of c on-specific threats (Wilson 1975, Kie 1999, Chapman and Chapman 2000a, Krause and Ruxton 2002). Several mechanisms have been stated as means to accomplish these benefits, for example, living in groups increases the overall vigilance rate while allowing the individual more time to fora ge, decreases the indivi dual probability to be predated, creates a major confusion when escapi ng a predator, and serves as defense against predators or other con-specifi c groups. Advantages of group foraging, include access to food otherwise not available for solit ary individuals, cumulative know ledge of the member of the group about food temporal and sp atial distributio n and a more efficient use of resources by avoiding patches previously visited (Jar man 1974, Wilson 1975, Kiltie and Terborgh 1983, Treves and Chapman 1996, Kie 1999, Chapman a nd Chapman 2000a, Krause and Ruxton 2002). However, living in groups may also present disa dvantages due to increased sexual and resource access competition (Kie 199, Chapman and Chap man 2000a). Competing for access to limited resources has been proposed as one of the main constraints for increased group size. Ungulates (refers to species from the Arti odactyla and Perissodact yla taxonomic order), living in open areas form larger social groups th an those living in closed habitats (Jarman 1974, Kie 1999). Kie (1999) suggest that the correlation between group si ze and habitat structure could be the result of strategies of avoiding pred ation and accessing resources. Generally, ungulate species living in open areas are subjected to c oursing predators while those living in closed habitats to stealth predators. Therefore, ungulat es living in open areas form large groups to
16 increase the overall rate of vi gilance without increasing the in dividual rate of vigilance and maximizes the energy ingest. In the other hand, t hose living in closed habitats, live in small groups or solitary and behave secretively a nd spend much time on i ndividual vigilance, consequently, minimizes the time spent foraging (Kie 1999). In this context, the White-Lipped Peccary ( Tayassu pecari ) the subject of this dissertation, is an unusual social ungulate that lives in closed habitats whil e forming large groups, generally from 10 to 300 individuals, but ther e are historic reports of groups larger than a thousand animals (Mayer and Wetzel 1987). Their social behavior is only matched temporary for the Bearded pig of Borneo ( Sus barbatus ) which form temporary groups or ag gregations which travel searching for fruit mast (Caldecott et al. 1993). The main differences is th at the Bearded pig groups are temporary while the white-lipped peccary are pe rmanent, remaining cohesive all year around (Fragoso 1994, Sowls 1997, Chapter 2this disserta tion). The other two known living species of peccaries, the collared peccary ( Pecari tajacu ) and the Chacoan peccary ( Catagonus wagneri ) form smaller groups that range from 2 to 50 indi viduals (with a mean of 8 individuals) in the former and 2 to 8 (with a mean of 4 individuals) in the latter (Sowls 1997). The collared peccary groups present fusion-fission beha vior and herds are not cohesi ve. The Chacoan peccary on the other hand remain stable in the small gr oups all year long (T aber et al. 1993). The White-Lipped Peccary ( Tayassu pecari ) Other names. Spanish: Jabal, Pecar Labios Blan cos, Saino, Hauilla, Senso (Donkin 1985). Description. The white-lipped peccary ( Tayassu pecari Link 1795, Order: Artiodactyla, Fam: Tayassuidae) is one of the three recognized li ving species of peccaries (Sowls 1997). Whitelipped peccary is the largest of the peccary species of Mexico. It is 1,100 mm in body length,
17 and the weight range from 25 kg. Their body co lor is brown-black with a distinctive white beard on the chest of adult individuals. Ecological Features. White-lipped peccary is a diurnal species, wh ich forms the largest groups of ungulates in the Neotropical forest ranging from 10 to 300 an imals or even more. The groups are very cohesive and remain together all year, travel ing within sight of each other (Emmon and Feer 1990, Fragoso 1994, Fragoso 1998). Usually they tr avel long distances in non-predictable movements, moving in a line with and alpha indi vidual leading the group. Evidence of seasonal movements have been found for some places, ho wever these movements are inside large home ranges (Fragoso 1998, Altrichter et al. 2002). In other areas its pr esence is episodic and unpredictable (Bodmer 1991). The white-lippe d peccary likes to wallow in muddy soils around water ponds and rivers, especially during the dry season. They are mainly frugivorous and can eat hard nuts not consumed by other species, e.g., Buriti palm ( Mauritia flexuosa ; Kiltie and Terborgh 1983). White-lipped peccary as with co llared peccary, are prey of great cats like puma ( Puma concolor ) or jaguar ( Panthera onca ), which often follow the herds, waiting for an opportunity to catch one. Small herds of white-li pped peccary are seen in areas where they seem to be disappearing (Leopold 1959, Em mons and Feer 1990). This sp ecies shows little tolerance for humans and avoids or disappears quickly from highly populated areas when the habitat changes dramatically (Leopold 1959, Alvar ez del Toro 1991, Sowls 1997, Reyna-Hurtado and Tanner 2007). Distribution and Status. White-lipped peccary distribution once range d from Veracruz (Mexico) to northern Argentina. However, this species has been redu ced in its distribution range from several regions due to over hunting and forest fragmentation. The white-lipped peccary is on CITES Appendix
18 II, but its status is poorly known. The IUCN Red-List (2007, www.iucnredlist.org ) classified it as Lower Risk/Least Concern due to the wide range of the species. However, in a recent workshop organized by the Wildlife Conservation Society (WCS 2005), experts on this species discuss its present and historic st atus and it was clear that the sp ecies is facing rapid decline in several parts of its range with the worst part in Mesoamerica (Mexico and Central America). In this region the white-lipped peccary is threatened by habitat dest ruction and overhunting. It is becoming extremely rare in Mexico since it is the first large species to disappear when the humans colonize a new area. More than 40 y ears ago, Leopold (1959) poi nted out the reduction in numbers and range of this species in Me xico due to habitat loss and excessive hunting (Leopold 1959, Emmons and Feer 1990). So true wa s his predictions that in Mexico white-lipped peccary have disappeared from Veracruz, Tabasc o and Yucatan states, and isolated population remains in Oaxaca and Quintana Roo, and the only stable populations can be found only in Chiapas and Campeche (March 1993, Naranjo 2002, WCS 2005). Despite this reduction of the range, the species was not classified in the na tional list of endangered species (NOM-Norma Oficial Mexicana de Especies en Peligro de Extin cion) issued by the Envi ronmental Secretary of Mexican Government (SEMARNAT), and sport hun ting is allowed in Campeche and Quintana Roo states under the official scheme of C onservation and Management Units (UMAs for its Spanish initials). The current situation is even worst when we considered that subsistence hunting (Naranjo 2002, Reyna-Hurtado 2002, Webe r et al. 2006, Reyna-Hurtado and Tanner 2007) and forest fragmentation are affecting this species outside the protected areas without any conservation plan at any s cale (regional or national). Current Knowledge White-lipped peccary has been studied in severa l parts of its geogra phic range. There have been many studies/reports on this species, but the ones that focus entirely on this species started
19 with Kiltie (1981) and Kiltie and Terborgh (1983) who studied th e ecology of the groups and the differences between collared peccary and white-l ipped peccary in Manu National Park in the Peruvian Amazon forest. They stat ed that differences in masticator anatomy forces allow the white-lipped peccary to access to different reso urces than the collare d peccary. Kiltie and Terborgh (1983) also were the first that ask the question: Why do white-lipped peccary form herds? And their main conclusions were that avoiding predation was the main force for this species to live in large groups. Fragoso (1994, 1997, 1998, 1999, 2004) has conducte d the longest study on the field with more than 15 years of monitoring a populati on of white-lipped peccary on Maraca Island in Roraima state, Brazil. This researcher was the fi rst on radiocollared individuals and determines a home range that is the largest known with 109-200 km2 for a large group. Another important result was the supporting evidence of dis eases transmission between domestic pigs ( Sus scrofa ) and white-lipped peccaries that can lead the la tter to suffer population crashes. White-lipped peccary has been studied also in Corcovado Natio nal Park, Costa Rica by A ltrichter et al. (2001, 2002) and Carrillo et al. (2002). Th is study provided the first detail ed accounted of movements, daily habits and food habits of groups on this species. Home range was also determined and an super-herd concept was developed where groups do not behave individually but belong to a super-herd that aggregated a nd divided temporary. Keuroghlia n et al. (2004) studied whitelipped peccary on fragments of th e Atlantic forest (Southern, Br azil) and determine the species can tolerate high levels of forest fragmentation if the fragment is protected. Habitat use and home range were determined here too. A. Keuroghlian (p ers. comm.) is also studying this species in the Pantanal area of Brazil with a high success of capture and developing models of age structure.
20 Recently, Altrichter (2005) and Altrichter a nd Boaglio (2004) studied the distribution of this species and the hunting patterns for peasants in the dry forest of the Argentine Chaco. In South America to my knowledge at todays da te (December 2007) there are only a couple of studies more in Bolivia (Madidi National Park and Chaco dry-forest, by Wildlife Conservation Society researchers) and one study in Peru (L os Amigos-Madre de Dios Parks, by World Wildlife Fund researchers). In Mesoamerica, th ere is one ongoing study in the Peten forest, which forms a continuing forest with Cala kmul Biosphere Reserve in Mexico (WCSGuatemala). In Mexico, March (1990) did a se minal study on distribution of the species and habitat available on the country. Naranjo and his team (Naranjo 2002) captured and radiocollared two animals but they could not follow fo r a long time. This researcher also studied white-lipped peccarys population abundance in hunted and non-hunted areas as part of the ungulates guild of the Lacandon forest in Chiapa s, Mexico (Naranjo 2002, Naranjo and Bodmer 2007). Quijano (2001) studied patterns of subsiste nce hunters for a Maya community in Quintana Roo on peccaries. Reyna-Hurtado and Tanner (2007) studied habitat prefer ences of the ungulates guild of Calakmul including the white-lipped peccar y by counting tracks in different forest types and under different hunting pressure. There is one on-going study of habitat use in the Lacandon forest, Chiapas using camera-trapping method (K. Tavera, comm. pers.) and one finished study on diet and food habits of the two species of pe ccaries in the Calakmul region (S. Perez-Cortez and Reyna-Hurtado). This dissertation is the firs t ecological study on Mexico that focuses entirely on this species where wild animals are captured a nd successful followed by radio-telemetry. The Great Calakmul Region This study was conducted in the southern area of the Calakmul Biosphere Reserve (CBR) in the Mexican state of Campeche (Fig.1). The CBR is the largest protecte d tropical forest in
21 Mexico with 7,238 km2 and was decreed as protec ted area in 1989. The CBR is part of the Great Calakmul Region (Galindo 1999) that includes th e Maya Biosphere Reserve in Guatemala and the Rio Bravo-Dos-Milpas conservation area in Belize, which together conform one of the largest tropical forests in Meso-America with more than 20,000 km2. The CBR is divided into two core areas with mitigation areas around them The CBR is bordered on the west, north and east sides by more than 100 human communitie s (ejidos) with a total population around of 35,000 people (INEGI 2005). The reserve is divide d east-west by a highway along which some ejidos are causing a belt of deforest ation that probably limits animal dispersal. The southern area of the CBR is contiguous with the Maya Biosphere Reserve in Guatemala without any dispersal barrier between them. According to Kppen (modified by Garcia 1988) the Calakmul climate is warm and subhumid (Aw), with a mean annual temperature of 24.6o C. There is seasonal rainfall, mainly in summer and early fall, with an annual average of 1,076.2 mm. Of the different forest associations (Pennington and Sarukhan 1998), four are widely distributed: Medium Sub-Perennial Forest (Medium), the more humid of the region, where trees are between 15 to 25 m high; Low-Flooded Forest (Flooded) that gets seasonally inundated af ter two to three 3 months of heavy rains, and where trees are between 5 to 15 m high; and the Medium and Low Semi-Deciduous Forests, which both can be classified as dry forest (D ry) where trees range from 8 to 25 m high, but species composition differs from that of the Medi um Semi-Perennial Forest. These four types of forest are highly intermingled w ithin the area, although the humidity from northwest (driest) to southeast (wettest) has an impact on the forest ty pes too. The areas topography is very flat with some gently rolling hills. Mean elevation is 250 m above sea level with some hills that reach 340 m. The water in the area is obtained through precipitation since there is no permanent river
22 system. Most of the rainfall percolates through th e limestone, but some drains superficially and stores in ponds. These ponds constitute the onl y source for water for w ildlife through the dry season. Despite the presence of some jaguar-hunters, ch icle-tappers and archeo logical looters, this region has remained almost undisturbed since the Mayans abandoned it 1,100 years ago. In the 1940s the colonization of the area began with the creation of the Zoh Laguna village as a center for logging operations and a base-camp for the ex traction of the chicle gum (Ericson 1997). In the 1970s when the Mexican government encouraged the colonization of th e last frontier in Mexicothe humid tropics Calakmul as well as other parts of Mexi cos tropical forests, received a large influx of people from the centr al and southern states of Mexico. This colonization process brought environmental cha nges to the region, and a municipality that presently includes 114 human settlements and an estimated population of 30,000 persons (INEGI 2005). Then in 1989 the CBR was granted its prot ection status. Calakmul Biosphere Reserve (CBR) is now the second largest reserve and th e largest protected tropi cal forest in Mexico. Today, the CBR is the hope for the conservation of the tropical forest in Mexico. A mosaic of social conditions and land tenur es surround it, and hunting as well as other extractive activities are common in the area (Reyna-Hurtado et al. 1999, Escamilla et al. 2000, Weber 2000, Weber et al. 2006, Reyna-Hurtado and Tanner 2007).
23 Figure 1-1. Localization of the Ca lakmul Biosphere Reserve (in da rk gray) and study area (white oval) in Mexico (light gray). Study Area 100 km
24 CHAPTER 2 HOME RANGE AND HABITAT SELECTI ON OF WHITE-LIPPED PECCARY GROUPS ( Tayassu pecari ) IN A SEASONAL TROPICAL FOREST OF THE YUCATAN PENINSULA, MEXICO Introduction Selection of home range area, habitat t ypes and resources are key decisions that enable a wildlife species to survive in a dete rmined site. These deci sions greatly influence the ecological relationships between that speci es and its environment. Knowing when and how far an animal moves, and how it uses the resources available to satisfy its requirements are important information that will enable to understand how the species have adapted to specific ecological conditions (Kernohan et al. 2001). Johnson (1980) stated that selection of resources for a wildli fe species occurs at di fferent levels and he defined four orders of select ion that goes from the selec tion of the geographic area (1st order), selection of a home range area (2nd order), selection of ha bitat types within the home range (3rd order) and selection of specific resour ces inside a partic ular habitat type (4th order). For social species that move in groups across the landscape these selections are even more interesting because additional factors like group size, group leadership, and probably a cumulative knowledge of the member s of the group are i nvolved (Boinski and Garber 2000, Chapman and Chapman 2000) The white-lipped peccary (WLP; Tayassu pecari ) is a social ungulate that fo rms the largest cohesive gro ups of any species living in dense tropical forest. Groups up to 300 ar e common on the Amazon forest (Kiltie and Terborgh 1983, Fragoso 2004) and there are an ecdotic reports of 700 and more than 1,000 individuals in a single group (Mayer a nd Wetzel 1987). WLP is a very mobile species that performs large scale moveme nts through the landscape searching for food
25 patches and water bodies (Kilt ie and Terborgh 1983, Fragoso 1999, Altrichter et al. 2002, Carrillo et al. 2002, Keuroghlian et al. 2004). WLP has been found to highly select areas where fruit abundance is high and water is not a limitation (Fragoso 1998, Altrichter et al. 2002, Reyna-Hurtado and Tanner 2005). Where WL P home range has been determined it spanned between 18.7 km2 for a forest fragment in the Atlantic Forest, Brazil (Keuroghlian et al. 2004) to more than 200 km2 for a large virtually unlimited forest in Roraima State also in Brazil (Fragoso 1998). Consequently, the social and ecological characteristics make this species a unique opportunity to study how highly cohesive groups living under dense forest select habitats types, move throughout the landscape, and how these decisions are affected by group size and habitat features. Almost 50 years ago Leopold (1959) stat ed that the WLP was among the first species to disappear when a forest became di sturbed. So true was hi s prediction that in Mexico as well as in Centro America, WLP have disappeared from most of its distribution range during the la st 20 years (WCS 2005). The Calakmul Biosphere Reserve (CBR, hereafter), the largest pr otected tropical forest in Mexi co, is one of the few places where WLP survive. Calakmul is a seasonall y tropical forest that lies close to the northern limit of the WLP histor ic distribution range and re present different conditions from the areas where WLP have been studied. In this study I determined annual and s easonal home range sizes and investigated selection of home range and habitat types (which correspond to the second and third order resources selection according to Johns on 1980) for four sympatric groups of WLP living in the semi-dry forest of the CBR. To gain insight into the factors promoting WLP movements, I incorporated additional info rmation such as group size, overlap among
26 groups and food and water availability. This provided a detailed understanding of the selective pressure shaping habitat selection and group si ze in a unique species with exceptionally large group sizes. This is the first study involving wild groups of WLP in Mexico. The goal was to obtain insights into the ecological strategies that this highly mobile, social ungulate has developed to surviv e in a resource-limited forest such as the Calakmul forests and valuable information for conservation purposes of this endangered species in Mexico. Material and Methods Study Area This study was conducted in the southe rn area of the Calakmul Biosphere Reserve (CBR) in the Mexican state of Campeche (See map in Chapter 1). The CBR is the largest protected tropical forest in Mexico with 7,238 km2 and was decreed as protected area in 1989. The CBR is part of the Great Calakmul Region (Galindo 1999) that includes the Maya Biosphere Reserve in Guatemala and the Rio Bravo-Dos-Milpas conservation area in Belize, which together conf orm one of the largest tropical forests in Meso-America with more than 20,000 km2. The CBR is divided into two core areas with mitigation areas around them. The CBR is borde red on the west, north and east sides by more than 100 human communities ( ejidos ) with a total popu lation around of 30000 people (INEGI 2005). The reserve is divided east-west by a highway along which some ejidos are causing a belt of deforestation that probably limits animal dispersal. The southern area of the CBR is contiguous with the Maya Biosphere Reserve in Guatemala without any dispersal barrier between them. According to Kppen (modified by Garcia 1988) the Calakmul climate is warm and sub-humid (Aw), with a mean annual temperature of 24.6o C. There is seasonal rainfall,
27 mainly in summer and early fall, with an annual average of 1076.2 mm. Of the different forest associations (Pennington and Sarukhan 1998), four are widely distributed: Medium Sub-Perennial Forest (Medium), the more hum id of the region, wh ere trees are between 15 to 25 m high; Low-Flooded Forest (Flooded) that gets seasonally inundated after two to three 3 months of heavy rains, and wher e trees are between 5 to 15 m high; and the Medium and Low Semi-Deciduous Forests, whic h both can be classified as dry forest (Dry) where trees range from 8 to 25 m high, but species composition differs from that of the Medium Semi-Perennial Forest. These four types of forest are highly intermingled within the area, although the hum idity from northwest (driest) to southeast (wettest) has an impact on the forest types too. The areas topography is very flat with some gently rolling hills. Mean elevation is 250 m above se a level with some hills that reach 340 m. The water in the area is obtained through pr ecipitation since there is no permanent river system. Most of the rainfall percolates through the limestone, but some drains superficially and stores in ponds. These ponds constitute the only source for water for wildlife through th e dry season. Capture of WLP for this study took place at a pond near the Calakmul archeological site. This pond is the largest in seve ral kilometers and is believed to have been modified by the Mayans more than 1,100 years ago to augment storage of water. The whole study area is lo cated at latitude of 18 07N and longitude of 89 48W, and lies in the heart of the southern area of the CBR. The area historically has been isolated since the Mayan abandoned it more than 1,100 years ago, and only some chicle tappers, jaguar hunters and ar cheological looter s visited occasionally there before archeological studies began in Calakmul on a permanent basis about 25 years ago. The
28 study area is protected eff ectively against hunting and ot her human activities by two checkpoints along the only existing narrow roa d. During the whole time there, no recent signs of human activities inside the fore st were detected. Human activity was only present at the Archeological site and on the ro ad leading to it, and was limited to tourism and archeological research. Study Design I captured and radio-coll ared individuals from four groups of WLP in the dry season of 2005. Given the high group cohesiveness in this species, I only radio-collared between two and three individua ls in each group to keep contact with the group. I used radio-telemetry mainly to locate the groups, but observations about habitat preferences as well as home range localizations were collect ed mostly by walking toward the animals (homing) after receiving a signal and followi ng the groups. Simultaneously, I collected monthly data on food and water availability by walking between 14 to 20 km of a transects grid building in a semi -randomly way thorough the study area. Capture Seventeen WLP were captured by using a re mote injection system (Dan-Inject Inc. and Telinject Inc.) between March and A ugust 2005. They were captured when they approached the pond, and that usually ha ppened early in the morning to midday. They were anesthetized using a comb ination of Ketamine and Xila zine Hydrochloride at doses of 7.72 and 4.34 mg/kg respectively. All veterinarian procedures were performed by a field wildlife veterinarian The induction time was 3 minutes approximately. Following a protocol established in advance (IACUC pe rmit # D594) I weighted, aged (by examining teeth wear following Leo Maffei, WCS-Bolivia ) sexed and measured each individual. I fitted all 17 individuals with white radiocollars weighting 400 g (Telonics Mod 400,
29 Signal range: 150/154 MHz). Animals usua lly recovered within 1 to 3 hours after which they started searching for their group. Individu als were able to rejoin their group as soon as the same day in some cases, but others spent more than 5 days separated from their groups before rejoining. I followed the animal s by foot the same day as far as possible until the animal was strong enough to walk continuously. Radiotelemetry Following capture I intended to locate each radio-collared animal twice at month using a two-element handheld Yagui antenna (Model: RA-14, Telonics) and portable receivers (Model TR-4, Telonics ). Because the area is rela tively flat and WLP groups range over huge areas, I used the two highest Maya Temples located in the Calakmul Archeological Site as the main points to retrieve directions of the groups and then locate and follow the groups for as much time as possible. I found that the maximum detection range of the signal was between 8 to 9 km from the Maya Temples and 1.5 km when locating from the forest floor. Due to the la rge range of detection and because the two Maya Temples are separated by only 700 m, I found that the triangulation error was larger than 500 m. Consequently I used the te mples just to find directions. When animals traveled further than 9 km from the Temple s (that happened in rain season mostly) I searched for them from a network of transect s, high hills and large trees. By September 2005 I had two additional high sites (Hill 1 and Hill 2) from where I found radio signals of the groups, and several temporal points (hi gh trees mostly) from where I were able to locate the animals by triangulati on with an error less than 100 m. Hills 1 and 2 were 13 and 15 km, respectively, away from the Maya Temples; they allowed me to document the seasonal ranging area of three of the four groups. On two occasions I had the opportunity to use airplanes (Cessna bi-motor airplane from SEMARNAT-Mexico; and an ultra-light
30 airplane from Light Hawk Vol unteers) to search for the gr oups, but searching time was limited to 20 minutes and 1 hour, respectively. Home Range Data Collection Home range data were colle cted for 18 months. When gr oups were detected, I tried to maximize data collection by the method of homing. Once directions from the Maya Temples were obtained, I walked into the forest to locate the group. Once the group was contacted, ranging locations were collected every 15 min with a GPS (Garmin XL, Global Position System) device. The group wa s identified as the Red, Blue, Green and Yellow group. The group sizes were 31, 25, 20, and 25 respectively. I also recorded ad libitum group cohesiveness, food habits and ot her interesting beha viors whenever the vegetation allowed us to have fu ll or partial view of the animals. The two field assistants were skilled hunters, and knew how to approach the groups against the wind to reduce the possibility of being detected. When groups b ecame aware of me presence they would run but typically they stopped approximately 100 m away. When I scared a given group off twice in a day, I left and came back another day. On such days only the original location was recorded for home range purposes. Howeve r, after some time, at least three groups seemed to tolerate my presence. I could s it and walk between 20 to 40 meters from the group. Using the homing method I collected 70 % of the home range fixes as GPS points, while the other 30 % were obtained through ra dio-telemetry triangulation from temporal points, and in some rare occasions from th e highest points (Maya Temples and Hills 1 and 2). Habitat use Data Collection To estimate habitat preferences for each gr oup, I recorded the forest type where the group was observed and thereafter at 15-minute intervals while in contact with the group.
31 I never attempted to estimate forest type base d on the radio-telemetry locations; all forest type use data were obtained by direct obser vations of the four groups. Forest types availability in the study area and inside the home ranges was estimated by using a supervised classification of a satellite image (2000) using ERDAS Imagine 8.7and Arc View 3.3. The training points were obtained fr om coordinates of the most typical forest types I encountered during fieldwork. Food and Water Availability Food availability was estimated monthly for 15 months by walking the transects (for the first 7 months there we re only 14 km of transects, and 20 km of transects for the remaining 8 months) to get an index of fr uit abundance on the forest floor. I followed Altrichter et al. (2001) index to estimate fruit availability whereby I record ed all fruits found within 1 meter of the transect line, then located the parent tr ee and counted all the ripped, larger than 5 mm, fruits that were in 2 m2 under the tree. I also recorded the forest type (Medium, Flooded or Dry fo rest) where fruits and parent trees were present. At the end I corrected for the differences in di stances walked among months by getting an average of fruit abundance index per km walked. I also collected information on presence or absence of earthworms and other invertebra tes under the leaf litter in the three forest types (Medium, Flooded and Dry) by sampling three 5 m2-square plot in each of these habitats monthly. I did this only for the last dr y and rain seasons after I learned that these invertebrates were an important source of f ood for WLP. Finally, I kept monthly records of water presence in the flooded forest, in a ll the ponds I knew, as well as in some stones (locally called sartenejas ) that naturally store water. I le arned, after followed the groups, that peccaries visited these sartenejas and that they are very important water source. I
32 used monthly precipitation figures obtained fr om the closest weather station 40 km north of the archeological site (C NA, Mexico technical report). Analysis of Home Range Data Home range sizes were estimated for each group by the Fixed-kernel, Adaptivekernel, and Minimum-convex-polygon (MCP) methods (Kernohan et al. 2001). Fixedkernel has proved to perform better and ha s a lower bias than the adaptive-kernel (Seaman and Powell 1996); however when the fi xed-kernel was performed for the whole sample size in each group with the method of least square cross validation (LSCV) to estimate the smoothing parameter, I found that home ranges were highly skewed towards highly sampled areas (close to the pond a nd more accessible areas) whereas the areas where sample size was low home ranges were underestimated. Cons equently, I randomly reduced the data in the oversampled areas to equal monthly sample size. After performing this procedure, 95 % and 50 % fixed-kernel estimates were obtained in Arc View 3.3 using the Animal Movement Analyst Ex tension (Hooge and Eichenlaub 1997). The smoothing parameter ( h ) obtained by LSCV method for the reduced data set for each group was subsequently used to estimate seasonal home ranges. Because fixed-kernel were performed with a reduced data set that was close to the minimum requiredsuggested sample size (30-50 fixes, Seaman 1999), I also estimated home range sizes by the adaptive-kernel method (at 95 %) using the Calhome program (Kie 1996). I performed this method because it is better suit ed when locations have different estimation errors, e.g., for fixes obtained at the edge of the home range (Kernohan et al. 2001), and because for some groups I could use larger portion of the sample size than in fixedkernel. To perform adaptive-kernel I reduced the data to a single randomly selected location per day, and analyses were perfor med per group. Finally, I estimated home range
33 by the Minimum Convex Polygon (MCP) (100 %) me thod, which is a simple method that encompasses the whole area where animals ha ve been located. However, although MCP does not give any estimate of preferences of areas inside the polygon, it is not subjective, uses the whole sample size, and is useful for comparisons with other data from the literature. Seasonal home range estimates were also obtained for the dry and rainy seasons of 2005 for the four groups, for the dry and rai ny season of 2006 for the Blue group and for the rainy season of 2006 for the Red group. The degree of overlap was calculated for the dry and rainy seasons of 2005, when seasona l home ranges of the four groups were available as percentage of area shared by two groups usi ng the 95 % fixed-kernel home range estimates. I compared dry and wet season home range size estimates for each group with a Studentt test for paired means. Since the re duced data consist in few observations per month for the fixed-kernel and one fix da ily for the adaptive-kernel, I consider the autocorrelation issue irrelevant in this case. Moreover, I estimated that WLP groups can travel up to 3 km per hour (Chapter 4), so that they can cover their home range in 24 hours, which is considered as a raw measur e to define independence among observations (Rooney et al. 1998). Additionally, Kernel methods have proved to be robust to autocorrelation (De Solla et al. 1999, Kernohan et al. 2001). Analysis of Habitat Use Data I used Compositional-Analysis (Aebischer et al. 1993) as the main method for contrasting the use of habitat types by WLP groups versus habitat availability. Compositional-analysis is a technique that us es the individual as the experimental unit. Due to the high cohesion in WLP groups, I pooled all the data for individuals of the same group, and used the group as the experimental unit. In this method, estimates of use and
34 availability come from the individual (here group) home ranges instead of the study area. By using differences in log ratios between use and availability for any given habitat, compositional-analysis accounts for the problem when a habitat highly used will invariably lead to an apparent avoidance of other habitat types (A ebsicher et al. 1993). Compositional-analysis provides ranks of se lection (analogous to the Johnson Ranking Method; Johnson 1980), and the differences between habitats can be tested by a Studentst test with the null hypothe sis of no preference. To test for Johnson (1980) second order habitat selection for WLP groups, habitat availability in compositional-analysis was estimated from a study area defined as a 30 x 30 km area that encompassed the four home ranges and intermediate areas where the protection status and the fore st types are similar to the home ranges. Forest type composition was estimated for the whole study ar ea and contrasted with that of the 95 % (total home range) and the 50 % fixed -kerne l (core areas) estimates, but not with the forest composition of MCP estimates, as expect ed. I performed this analysis because FK is per se a selection of space wher eas MCP is only the area circumscribing the space where the animal has been found. For the third order selection (Johnson 1980) I contrasted forest type composition within the 100% MCP home range area of each group with the percentage of observations of this group in each forest type. I used MCP home ranges since they estimate the whole area that a WLP group visited, including these areas that are not used, therefore the actual use of forest types insi de this polygon would be an estimation of the actual groups preferences among the forest types.
35 Finally, to confirm results the overall study area forest proportions were tested versus group preferences by using the program HABUSE (Neu et al. 1974, Byers et al. 1984) with Chi-square analyses and Bonferroni interv als to test selection within individual forest types. Finally I also ranked prefer ences using the Johnson Ranking Method (Johnson 1980) which contrasts the rank s of preferences and availability of the different habitat types. To deal with the autocorrelation issu e regarding the use of forest types, I only considered obser vations 30 minutes apart, give n that forest types are so intermingled in the CBR, and that WLP can move up to 3 km per hour (Chapter 4) so they can move easily among them. All statisti cal analyses were pe rformed to the 0.05 confidence level either in Microsoft Excel-X P or SPSS (Statistical Package for Social Science1997). Results Home Range During the 18 months of the study I en countered the groups on 203 days. The radio-collared animals were always in their respective groups despite times when the four groups were within 5 m of each other. I di d not detect mixing of individuals among the groups and only in two occasions I sighted soli tary individuals that I could not assigned to any group at all. Consequently I used the radio-marked animals as a true indicator of group movements and habitat preferences. I ma intained contact with the groups from 10 to 17 months. The Blue group was followed 17 months with the exception of January 2006; the Green group was followed 12 continuous months until February 2006. The Red group was followed 11 continuous months from March 2005 to January 2006; after that they disappeared for 5 months and reappear ed for the last 2 months of the study. The Yellow group was followed 7 months during a 10 month period from March to December
36 2005. Given the different durations over wh ich individual groups were followed, I considered annual home range es timates are only well represented for the Blue and Green groups. Data for Red and Yellow groups were considered seasonal and partial estimates of their home range sizes (Table 2-1). I estimated fixed-kernel, adaptive-kernel and MCP home range sizes based in fixes collected during this time (Table 2-1). A ccording with Seaman and Powell (1996) and Seaman et al. (1999), I considered the 95% fi xed-kernel estimates to best represent the home range of the groups for their respective period of time (Fig. 2-1). With 95% fixedkernel the Blue group had the large home range followed by the Green group. The adaptive-kernel method provided relatively simi lar results but tended to uplift estimates for the two groups with larger home ranges, as compared with th e fixed-kernel method. The MCP estimates showed the whole area th e groups used over the respective periods they were followed, with the Green group using an area more than five times the size of that used by the Yellow group (Table 2-1, Fig. 2-1). The smoothing parameter obtained for each group in the 95% fixed-kernel anal yses was similar among the groups; therefore, this estimate was used for subsequent anal yses of seasonal and monthly home ranges. For none of the groups did the MCP home range estimates reach an asymptote when sampling stopped (Fig. 2-2). Thus, all groups were moving into new areas, as was illustrated in the dry season of 2006 (January 2006) when the Calakmul pond dried up and the Red and Green groups entered new ar eas after having a relatively stable home range for several months. Only the Blue gr oup showed a pattern of seasonal movements by visiting the Calakmul pond during the two dry seasons and then traveling 17 km to the North during the peak of the two rain seas ons. In 2005, the four groups increased their
37 home range substantially duri ng the rain season with a maximum increase of three times in the Blue group (Table 2-2). The differen ces between the 2005 dry and rain seasons for the four groups estimates were statistically significant (Students ttest for paired means = 3.05, df = 3, p = 0.02). The Red group home range for the rain season of 2006 was not only larger than that of the dry season of 2005, but also larger than the rain season of 2005. On the other hand, the Blue group showed similar sizes for the two seasons in 2006, with the home range of the rain season of 2006 almost half the size of that of 2005 (Table 2-2). Groups shared the same space almost completely during the dry season of 2005 with a 91.7% degree of overlap for the f our groups combined (Table 2-3; Fig.2-3). During the rain season of 2005, the Blue, Green and Yellow groups traveled to different areas to the north whereas the Red group stay ed in the dry season area. The degree of overlap dropped to 34.5% for the rain season of 2005 (Table 2-3; Fig. 2-4) for all the groups. In 2005, the Blue and Green groups travel ed between their respective northern areas and the Calakmul pond several times until the rain season peaked and the flooded forest was indeed flooded. Thereafter, they stay ed in the northern areas until the next dry season, when the Blue group came back clos e to the Calakmul pond. The Green group traveled to different areas to the east and then to the south where I lost contact with them during the dry season of 2006. The Yellow group moved to the north at the beginning of the rain season of 2005 but ne ver returned during the rest of the 10 month tracking period. The Red group was the only who stay ed close to the Calakmul pond for 11 months, but it moved eventually farther than 10 km to the south when the pond dried up
38 completely in 2006. No radio contact was made for 5 months after this southerly move, even with the aid of an airplane. In Ju ly 2006, however, the Red group returned to the Calakmul pond area where it rema ined in the 2006 rain season. Habitat Use I performed compositional-analysis for J ohnsons (1980) second order selection to compare habitat types within the fixed-kern el home range for each group and the whole study area and (Fig. 2-5). When using 95% es timates from the four groups there were no differences among the forest types (Table 24a). However, when contrasting the study area vs. 50% fixed-kernel (the core area of activity of the groups), I found a significant preference for Ponds and Flooded forest, with the least preferred be ing the Dry forest (Table 2-4b). Compositional-analysis for Johnsons (1980) third order selection (MCP home ranges vs. group localizations) showed that inside their home ranges areas all groups significantly preferred Ponds and Medium forest and avoided the Dry forest (Table 2-5a). I performed an additional analysis taking away observations in Ponds; the results consistently showed that all WLP groups significantly preferred Medium forest and avoided Dry forest (Table 2-5b). Flooded fore st was also very important for the four groups and for the Green group was ranked even above Medium forest, just below Ponds. I ran compositional-analysis for the dry (f our groups) and rain (only Red and Green groups) seasons 2005 separately, and found, not surprisingly, that Ponds were the preferred habitat type in bot h seasons. However, in the Dry season they also selected Medium forest and avoided Flooded and Dry fo rest totally (Table 2-6a). In the rain season selection of Flooded fore st outranked that of the Medi um forest (Table 2-6b) at
39 least for the Red and Green groups. Not enough data were collected in the rain season for the other two groups to be in cluded in the analysis. Finally, to confirm results from the com positional-analysis, I contrasted the group localizations with the MCP estimates for each group, and all the observations combined versus the proportions of ha bitat types for the whole area using two methods: Neu X2 analyses (Byers et al 1984) and Johnson Ranking Method (Johnson 1980). In general they confirmed the strong sele ction of Ponds and Medium fo rest with the Dry forest being the least preferred ha bitat type (Table 2-7). Food and Water Availability I found around 54 plant species along the tr ansects within the study. Four species known to be consumed by WLP ( Brosimum alicastrum, Manilkara zapota, Talisia olivaeformis, and Cordia dodecandra; Sowls 1997, Fragoso 1998, Altrichter et al. 2001, Beck 2006) accounted for the majority (73.6 %) of the fruits on the forest floor. Brosimum alicastrum and Manilkara zapota accounted for 48.15 and 18.37 %, respectively, and have been found to be heavily consumed by WLP in the Calakmul region (S. PerezCortez and R. Reyna-Hurtado, in prep.). Seasonal availability of fruits varies among species. During the study B. alicastrum fruited in the early rain season (June to September) and Manilkara zapota fruited in both seasons with a peak in the dry season (February to May). Talisia olivaeformis fruit was present in la rge amounts only in May, and Cordia dodecandra produced low quantities of fruit in May-June (Fig. 2-5). Fruit production also varied among forest types. I found B. alicastrum mainly in the Medium forest (83%), while Manilkara zapota was abundant in several forest types with the majority in the Dry forest (56%).
40 Earthworms, an invertebrate food item used seasonally by WLP, occurred in the rain season and mainly in the Flooded forest (68 %). I observed WLP spending extended periods of time during the rain season rooti ng under the leaves in the Flooded forest searching for earthworms. I did not see any other food item in that specific time in this type of forest. I also found that during th e dry season WLP fed on a species of eel fish ( Ophisternon aenigmaticum Fam: Synbranchidae ) that inhabits the mud at pond edges. Water was available during the dry s eason of 2005 only at the Calakmul pond. Despite some early rains in May 2005 (F ig. 2-6), I only saw water in other ponds, Flooded forest and the sartenejas stones after late June 2005. These sartenejas have depressions that retain water long after ra infall has ceased. They appear to be used frequently by WLP given the trampling of th e forest floor around them and the prominent trails leading to them. Water was widely available until November 2005, when the flooded forest and the ponds started to drai n out. By December the flooded forest had nearly completely dried and by January-Febr uary 2006, all ponds had completely dried; even the Calakmul pond dried up for the fist time in 16 years (Calakmul Archeological Site guards, pers. comm.). Th ere was water only in some s artenejas at that time. During those months I observed that the Blue group performed straight movements as far as 4 km aimed to locate sartenejas I assume they previously knew the locations of these stones given the precision of their movements (Chapter 4) The last two weeks of April and the first two weeks of May 2006 we re the driest because even the sartenejas dried up. The rain season of 2006 began abundantly in late May.
41 Discussion Home Range The home ranges of the four WLP groups in this study were estimated from different periods of time, ranging from 10 to 17 months. Nevertheless, the four groups home range continuously increased throughout the study period, suggesting that WLPs move over large areas and that these movement s cannot be predicted on a yearly basis. In fact it may take several years for the WLPs to return to previously visited areas, as has been documented elsewhere (Siono-Secoya area in Ecuador, Vickers 1991; Yuqui area in Bolivia, Stearman 1992). WLP at Calakmul, with group sizes that never exceeded 31 individuals, had one of the largest reported home ranges, only be ing surpassed by a very large group (>200 individuals) in Maraca Island, Roraima, Brazil (Fragoso 2004). WLP home ranges estimated elsewhere were fairly smaller than my estimates. For example, in Corcovado National Park in Costa Rica, Carrillo et al. (2002) estimated the largest annual home range of a herd of between 40 animals as 37.2 km2. Keuroghlian et al. (2004) estimated annual home ranges for groups (m ean 41.7 individuals per herd) living in fragments of the Atlantic fore st in Brazil to average 18.71 km2. Groups in Corcovado and the Atlantic forest were bigger than groups in Calakmul and interchanged repeatedly individuals, s uggesting to the authors to name them as superherds composed by several sub-herd s (Carrillo et al. 2002, Keuroghlian et al. 2004). My study, as well as Fragosos (1998) stu dy in Roraima, did not reveal frequent switching of individuals between groups. If migration of individuals o ccurred between groups it happened in very low numbers that were undetected.
42 In the two studies where herd mixing occurred, the study areas were a small fragmented forest (Atlantic fragment forest with 21.78 km2; Keuroghlian et al. 2004) or a medium-sized reserve (Corcovado National Park with 500 km2; Carrillo et al. 2002). On the other hand both studies wher e no mixing of individuals o ccurred happened in sites where habitat size seemed not be a constrai nt (Maraca Island, Roraima Brazil with more than 1,100 km2 plus 100,000 km2 of adjacent forest belonging to Indigenous groups, Fragoso 2004; and Calakmul Biosphere Reserve with 7,231 km2 plus 25,000 km2 of adjacent forest in Mexico, Guatemala and Be lize, this study). These observations suggest that when habitat is limited or fragmented, gr oups remaining close to each other for long periods of time exchange individuals more frequently than groups roaming over large areas. I did not detect fission-fussion among the groups. However, some behavioral patterns that I observed might be associated with the ability to form a super-herd because of the great degree of overlap among home ranges both spatially and temporally. When the four groups met at the Calakmul pond, they would often rest as close as 5 meters each other. Additionall y, I saw twice two groups travel together for several days (Chapter 4). Given that WLPs may have knowledge of comm on areas, and the fact they tolerated each other and travel ed together, suggests the possi bility that they may have belonged to an original single group that at some point in time split into subgroups. It would be interesting to see if contiguous groups on the area present a similar behavioral pattern. Groups of WLP substantially increased th eir home ranges when the rain season arrived to the CBR suggesting that water avai lability was more important for determining
43 movements than food availability. During the dry season the Blue, Green and Yellow groups left the pond only during rains but return ed after 2 or 3 days if the rain stopped. These three groups left the pond definitely in late May 2005 when the rain abundantly fell. Although I observed large amounts of Talisia olivaeformis fruit in early May on the forest floor, the groups were eating herbs and rooting for eel at the edges of the pond without venturing far from water. The same pattern was repeated for the Red group in the beginning of the dry season 2006, and for th e Blue group throughout the whole 2006 dry season, when both groups were observed eat ing herbs and roots around the pond without traveling to different areas in the forest to search for food. Apparently, three of the four groups (B lue, Green and Yellow) that were found sympatric in the dry season of 2005, traveled from northern areas to the Calakmul pond (the only source of wate r at that time). When the rain seas on arrived they returned to their original areas with the Red group being the only group to stay close to the pond. The next year, in the dry season of 2006, the Calakmul pond dried for the first time in 16 years. I believe this event forced the Red group to leave the area in January 2006 when the pond had very little water and the dry season was just starting. They traveled south to an undetermined area (radio contact was lost) but re turned after the rain season has filled the Flooded forest. This event had even more dram atic consequences for the Blue group that returned in March 2006 when there was no water in Calakmul pond. For 10 weeks they were observed searching desp erately in the few ponds and sartenejas for the remaining water. They were in critical physical condition by late Apr il when there was no water in the area. I could then sit within 4 meters fr om the group without the animals to run away. Not other group showed up at the Cala kmul pond during the dry season of 2006.
44 The larger home ranges and the smaller group s sizes reported in this study can be interpreted as an adaptation st rategy of the species to the particular conditions of the Calakmul forests. These forests are drier than Corcovado National Park, the Atlantic forests and the Amazon forest. According to several authors the av ailability of high energy food, like fruits, is higher in primary humi d forest than in secondary or dry forests (Bodmer 1989, Sowls 1997, Altrichter et al 2001, Beck 2006). The WLP small group sizes in our study agree with th e Ecological Constraint Model (ECM), which predicts that species feeding on patches and living in gr oups will have to travel further or more frequently as the group size increase; otherw ise they would have to split in smaller groups (Chapman and Chapman 2000, Gillespie and Chapman 2001). WLP groups survival strategies in Calakmul agree with the predictions of the ECM because they have increased their range and live in smaller groups than elsewhere. Additionally, the increase in home range de monstrates that this species has the ability to perform movements at the landscap e scale as was predicted by Fragoso (1999). This ability appears to be matched with some kind of spatial memory or orientation skills that allow them to search and find specific landmarks such as ponds or sartenejas in this large and relatively featurele sslandscape (pers. obs.). Habitat Use WLP groups in Calakmul as well as in several parts of th eir range (Sowls 1997, Fragoso 1998) depend on water bodies (ponds) to wallow, refresh and forage. I observed that WLP visited the ponds da ily unless there was water in the Flooded forest. In the CBR, aside from the ponds WLP highly preferre d patches of Medium forest throughout the year. During the rain season, however, the Flooded forest was preferred. The Dry forest that is very abundant in the area was used mainly to travel between patches of
45 preferred forest or ponds. Those preferences we re self-evident when following the groups when, on repeated occasions, I observed them travel in a single file through the Dry forest but then disperse to forage when they encountered a Medium or Flooded forest (pers. obs). Data suggest that the ideal habi tat composition for this species in Calakmul would be Medium forest intermixed with Flooded forest and permanent ponds. In effect, the four groups spent almost all the time vis iting an area where there are mixed patches of Medium and Flooded forests to the north of the Calakmul pond, but they rarely visited the huge patch of Flooded forest that lies to the west of the Calakmul pond. Likewise, the groups never visited the huge patch of Medium forest located to the southeast of the Calakmul pond. Both of these areas are in a close proximity to the mixed forest and are part of the extended home ranges of the Red, Blue, and Green groups home ranges. The selection for Flooded forest is highli ghted in the analysis of a second order selection when using the 50% of the home range, and for the rain season in the third order selection analysis. The Medium forest was selected all year long and by the four groups. These results clearly matched the food habits of this species. WLP depends largely on fruits (Kiltie and Terborgh 1983, Bodmer 1989, Sowls 1997, Fragoso 1998, Altrichter et al. 2001, Beck, 2006), and it has been confirme d that in the study area fruits compose 81% of their diet (S. Perez-Cortez and R. Reyna-Hurtado in prep.). It has been demonstrated also that they consume animal protein and feed on i nvertebrates and fish (Sowls 1997, Fragoso 1998). I observed that they used the Medium forest mainly to find fruits ( B. alicastrum, M. zapota, T. olivaeformis and C. dodecandra ) and to rest during midday (Medium forest is taller than Flooded and Dry forests; consequently it is cooler
46 than the shorter forests), and visited the Flooded forest mainly when earthworms were abundant (rain season) and to wallow on mud patches when they were flooded. Ponds were visited all year round, but es pecially during dry season. Besides the water and the mud, Ponds also pr ovide herbs at the edges, eels and other invertebrates in the soft mud around the open water. On se veral occasions I observed the peccaries moving in a single day between the Medium forest, where they ate B. alicastrum to the Flooded forest where they searched for earthworms and took mud baths. These observations suggest that the amount of Medi um forest patches in termixed with Flooded forest and the number of Ponds or sartenejas can determine the density and the group sizes in this species in the area. The forests of the Yucatan Peninsula s how a strong northwestern-southeastern ecocline, where the forests are drier to the nor th and more humid to the south. Therefore, the pattern of habitat use documented in this study would lead me to predict that in the northern areas of the CBR the density of WLP will be lower and group size will be smaller, while the opposite will occur in the southern forests of the CBR and in the neighboring Guatemala forests. This prediction seems to be true for the North because in the north core area of the CBR (north of the Escarcega-Chetumal road) where the Dry forest is dominant, WLP seems to be absent from most parts of the Reserve. During informal conversations for the past five ye ars with informants who know the area, I was told that no WLP lives in large portions of this forest. Additionally, WLP group sizes living in the southern area are larger that groups living in the northern edge were ejidos are closer (Chapter 5) Howe ver, it is not known if this is a consequence of hunting pressure or habitat quality.
47 WLP is unique in that it is a highly sel ective, social species that moves at the landscape scale. This study demonstrated thei r long distance moveme nts and their use of the forest and of ponds at the landscape scale. Also, I found that they can survive in a semi-dry landscape where patches of preferre d forest are interspe rsed with non-optimal forest by increasing their movements, decreasi ng the group size, and taking advantages of many sources of food other than fruits, such as earthworms, eels, and herbs, and by using their amazing sense of orientation, or spat ial memory, to locate and use additional sources of water as small as a stone ( sartenejas) Conservation Implications Groups of WLP are smaller and range ove r a larger area in Calakmul forests suggesting that density is probably lower than in other, more humid forests where the species lives. Thus to sustain viable populati ons of this species it will be necessary to conserve larger areas in this ecosystem. Additionally, WLP must be able to visit a source of water on a daily basis WLP also showed a high preference for Medium and Flooded forest intermixed in the landscape, and us ed complementary resources from these two habitat types. CBR and the surrounding forests (in Guatemala and Belize) form the single largest forest in Central America, and as such represent the largest contiguous habitat available for the WLP. However, hunting pr essure exists in the communities that surround these reserves, and WLP populations are not doing well near human settlements (Weber 2000, Escamilla et al. 2000, Weber et al. 2006, Reyna-Hurtado and Tanner 2007). To conserve WLP in northern Central Am erica it will be necessary to preserve intact this piece of the Maya forest and the attributes that allow WLP to survive in it. Conservation measures will likely include e ffective protection against hunting, no further road development to maintain the area isolated, the availa bility of ponds and/or other
48 sources of water, and the maintenance of a landscape composed of interspersed Medium and Flooded forests. Active conservation meas ures would be to assure WLP groups have access to water sources (could even be worth to strategically locate artificial water containers for the extreme dry years as the dry season 2006). Also, conserving large intact patches of forest would main tain WLP populations on human dominated landscapes if hunting pressure can be controlle d. Finally, maintaining the continuity of the forest that currently extends from th e study area to Guatemala and Belize seems the major conservation priority for species th at, like WLP, move over the landscape.
49 Table 2-1. Home range sizes (km2) as estimated by three methods of the four groups of whitelipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico. Groups Fixed Kernel Adaptive Kernel Minimum Convex Polygon Months 95% 50% H 95% 100% N Red* 43.9 (n=80) 5.9 1200 37.7 (n=57) 43.6 (n=286) 13** Blue 97.5 (n=33) 10.4 1264 109.7 (n=68) 77.2 (n=214) 17 Green 82.4 (n=60) 10.9 1364 105.0 (n=60) 121.8 (n=137) 12 Yellow* 38.8 (n=18) 5.7 1213 32.7 (n=41) 23.2 (n=41) 10** Partial annual home range. **Non continuous months Table 2-2. Home range sizes (km2) for dry and rain season of 2005 and 2006 using Fixed Kernel with 95% of the observations for the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico Groups Dry season 2005 Rain season 2005 Dry season 2006 Rain season 2006 Red 17.8 24.7 Out of area 33.8 Blue 19.3 61.6 36.1 31.9 Green 22.6 50.1 Lost Contact Lost Contact Yellow 19.8 36.2 Lost Contact Lost Contact Table 2-3. Overlap among the home range of the four white-lipped peccary groups during the dry and rain season of 2005. Calakmul Biosphere Reserve, Campeche, Mexico. Groups Combination Dry Season 2005 % overlap Rain Season 2005 % overlap Blue/Red 89.7 36.9 Blue/Green 99.8 40.6 Blue/ Yellow 91.7 6.0 Green/Blue 85.0 49.9 Green/Red 78.4 41.3 Green/ Yellow 84.8 19.3 Red/Blue 97.0 92.0 Red/Green 99.6 83.9 Red/ Yellow 99.2 0.0 Yellow/Blue 89.0 11.3 Yellow/Green 96.6 29.7 Yellow/Red 88.9 0.0 All groups overlap 91.6 34.25
50 Table 2-4. Compositional analysis of the study area and home ranges (second order selection Johnson 1980) for the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico. a) Study area vs. 95% kernel home ranges Habitat types Habitat types Medium Flooded Dry Ponds Rank Medium NA 0 Flooded + NA + + 3 Dry + NA 1 Pond + + NA 2 b) Study area vs. 50% kernel home ranges Habitat types Habitat types Medium Flooded Dry Ponds Rank Medium NA --+ --1 Flooded +++ NA +++ --2 Dry --NA --0 Ponds +++ +++ +++ NA 3 + and signs denote preference or avoidance and a tr iplicate sign means statistical significance Table 2-5. Compositional analysis of MCP home ranges and usage (third order selection Johnson 1980) of forest types for the four white-lipped peccary groups. Calakmul Biosphere Reserve, Campeche, Mexico. a) MCP home range vs. group locations for all four habitat types Habitat types Habitat types Medium Flooded Dry Ponds Rank Medium NA + +++ --2 Flooded NA + --1 Dry --NA --0 Ponds +++ +++ +++ NA 3 b) MCP home range vs. group location in the three forest habitats Habitat types Habitat types Medium Flooded Dry Rank Medium NA + +++ 2 Flooded NA + 1 Dry --NA 0
51 Table 2-6. Compositional analysis of seasonal MCP home ranges and usage (third order selection Johnson 1980) of forest types for the four groups of white-lipped peccary. Calakmul Biosphere Reserve, Campeche, Mexico. a) Compositional analysis of the third or der selection for the dry season 2005 (all groups) Habitat types Habitat types Medium Flooded Dry Ponds Rank Medium NA + +++ --2 Flooded NA --0 Dry --NA --0 Ponds +++ +++ +++ NA 3 b) Compositional analysis of a third order selection for the rain season 2005 (only for Red and Green groups) Habitat types Habitat types Medium Flooded Dry Ponds Rank Medium N/A +++ --1 Flooded + N/A + --2 Dry --N/A 0 Ponds +++ +++ +++ N/A 3 Table 2-7. Habitat selection with Neu X2 and Johnson Ranking Methods analyses for individual groups and all groups of white lipped peccary combined. Calakmul Biosphere Reserve, Campeche, Mexico. Groups Neu X2 Analysis Johnson Ranking Analysis Medium Flooded Dry Ponds X2, p value Medium Flooded Dry Ponds Red = + + X2=93570 p=0.00 1 1 1 -3 Blue + + X2=26332 p=0.00 -2 1 2 -2 Green = + X2=8463 p=0.00 1 0 2 -3 Yellow = + X2=53315 p=0.00 0 0 3 -3 All groups + + X2=47530 p = 0.00 1 0 2 -3 In Johnson ranking method negative values mean preference while pos itive values indicate avoidance
52 Figure 2-1. All seasons fixed kern el (95 and 50 % represented as total area, and circles inside respectively) home ranges for the four groups of white-lipped peccary in the Calakmul Biosphere Reserve, Campeche, Mexico. 0 20 40 60 80 100 120 140 Mar05 Apr05 May05 Jun05 Jul05 Aug05 Sep05 Oct05 Nov05 Dec05 Jan06 Feb06 Mar06 Apr06 May06 Jun06 JulAug 06Km 2 Red Blue Green Yellow Figure 2-2. Cumulative ho me range sizes (km2) using MCP for the four groups of white-lipped peccary in the Calakmul Biosphere Reserve, Campeche, Mexico.
53 Figure 2-3. Fixed kernel home ra nge estimation for the dry season of 2005 for the four groups of white-lipped peccary in the Calakmul Bi osphere Reserve, Campeche, Mexico. Figure 2-4. Fixed kernel home ra nge estimation for the rain season of 2005 for the four groups of white-lipped peccary in the Calakmul Bi osphere Reserve, Campeche, Mexico.
54 0 10 20 30 40 50 60 70 80 90 May05 Jun05 Jul05 Aug05 Sep05 Oct05 Nov05 Dec05 Jan06 Feb06 Mar06 Apr06 May06 Jun06 Jul06Abundance Inde x Brosimum alicastrum Manilkara zapota Talisia olivaeformis Cordia dodecandra Figure 2-5. Monthly fruit abundanc e index (# occurrence/km) on the forest floor for the four most abundant species in the Calakmul Biosphere Reserve, Campeche, Mexico. 0 50 100 150 200 250Mar05 Apr05 May05 Jun05 Jul05 Aug05 Sep05 Oct05 Nov05 Dec05 Jan06, Feb06 Mar06 Apr06 May06 Jun06 Jul06Milimeter s Rainfall Water in Flooded Forest Water in Calakmul Pond Figure 2-6. Water availability for the study period (rainfall in mm, water in flooded forest and Calakmul pond has two values, 100=presence, 0=absence) in the Calakmul Biosphere Reserve, Campeche Mexico
55 CHAPTER 3 DO MOVEMENTS AND GROUP SIZE IN WHITE-LIPPED PECCARY ( Tayassu pecari ) FIT THE PREDICTIONS OF THE ECOL OGICAL CONSTRAINT MODEL? Introduction Benefits of group living have been broadly clas sified into three main categories, namely avoidance of predation, optimization of resource use, and recently, avoidance of con-specific threats (Wilson 1975, Kie 1999, Chapman and Ch apman 2000a, Krause and Ruxton 2002). Several mechanisms have been stated as means to accomplish these benefits. For example, living in groups increases the overall vigilance rate while allowing i ndividuals more time to forage, decreases the individual probability to be predat ed, creates a major confusion when escaping a predator, and serves as defense against predator s or from con-specifics for those species where aggressive encounters between groups may occu r. Advantages of group foraging include access to food otherwise not available for solitary in dividuals probably due to a cumulative knowledge of the members of the group about the temporal and spatial distribution of food, and a more efficient use of resources by avoiding patc hes previously visited (Jarman 1974, Wilson 1975, Kiltie and Terborgh 1983, Treves and Chap man 1996, Kie 1999, Chapman and Chapman 2000a, Krause and Ruxton 2002). However, living in groups also presents disadvantages due to an increased sexual and resource access compe tition (Kie 1999, Chapman and Chapman 2000a). Competition for access to limited resources has been proposed as one of the main constraints on group size (Chapman and Chapman 2000a). The Ecological Constraint Model (Chapman 1990, Wrangham et al 1993, Chapman and Chapman 2000a, Gillispie and Chapman 2001) deal s with the disadvantages of group living, such as within-group competition. This model was developed to explain how food competition can limit group size, and is better suited for sp ecies that feed on patchy distributed food while
56 living in cohesive groups (Chapman and Chapma n 2000a). The model predicts that when group size increases, then patch depletion time decreases; consequently, groups have to travel farther or faster to visit another patch. When increase in movement rate reaches a limit where the energy spent on traveling is not matched by the energy ga ined on food consumption then is better for the group to split into smaller groups (Chapman and Chapman 2000a). The Ecological Constraint Model fits very well the behavior of several species of primat es and carnivores (Chapman 1990, Wrangham et al. 1993, Chapman et al. 1995, Chapman and Chapman 2000a, 2000b; Gillispie and Chapman 2001, Ganas and Robins 2005), but some exceptions exist, especially in folivourous species (Chapman and Pa velka 2005, Snaith and Chapman 2007) Group living among ungulates has evolved mainly for species living in open habitats (e.g., grasslands, savannas; Jarman 1974, Kie 1999). Therefore, the white-lipped peccary (WLP; Tayassu pecari ), a Neotropical ungulate, represents a unique social phenomenon by living in the largest and most cohesive groups for ungulates inhabiting dense tropical forests. WLP groups usually contain from 20 to 300 individuals (Leopold 1959, Kiltie and Terborgh 1983, Emmons and Feer 1990, March 1993, Fragoso 1994, Sowls 1997, Altrichter et al 2000, Carrillo et al. 2002) with anecdotal reports of up to 1000 memb ers (Mayer and Wetzel 1987). In addition, WLP groups travel long distances in such a way that their movements have been called nomadic, migratory, episodic, unpredictable (Donkin 1985, Vickers 1991, Hernandez et al. 1995, Sowls 1997, Bodmer et al. 1997), seasona l (Altrichter and Almeida 2002) or seasonal within a huge home range (>100 km2 for Maraca Island; Fragoso1998, 2004). Groups of white-lipped peccaries are highly cohesive and are know to demonstrate a dietary prefer ence for fruit in excess of 60 % of their diet (Sowls 1997). All over the distribution range they depend largely on palm fruits (up to 37 palm species, Beck 2006) that usua lly have a patchy sp atial distribution.
57 It is not clear yet why this species perfo rms large scale movements, although they are believed to be a response to food scarcity (Fragoso 1998, Altrichter et al. 2001). The longdistance movements showed by this species, its constantly increasing rarity of the species due to human encroachment (Leopold 1959, WCS 2005) a nd living in dense trop ical forest have prevented long-term detailed studies on this spec ies with some valuable exceptions (Maraca Island, Brazil, Fragoso 1994, 1998, 1999, 2004; Corcova do National Park, Costa Rica, Altricher et al. 2000, 2002; Carrillo et al. 200 2; Atlantic Forest, Brazil, Keur oghlian et al. 2004). Altrichter et al. (2001) demonstrated that WLP moves among habitats ty pes following fruit production. However, the unpredictable nature of some long distance movements of WLP hints at food patch depletion as a better explanation than scattere d food resources, making the Ecological Constraint Model very suitable for testing several hypothese s regarding WLP movements. In the semi-dry tropical forests of Calakmul Biosphere Reserv e where this study was car ried out, not only food but water too can be scarce or even absent seas onally. Water scarcity exercises an additional constraint on WLP groups, also adding complexity to the relationship of the species with its environment. Using the Ecologi cal Constraint Model as a theo retical framework, I explore the following questions: What are the stronger factors related to differences in area used by groups of WLP living in the semi-dry forest of Calakmul Biosphere Reserve (CBR)? Can the relationship among area used, group size, and food, and water availability on WLP be explained by the Ecological Constraint Model? And, is th ere a positive correlation between group size of WLP and rainfall across the distri bution range of this specie? Methods Study Area This study took place in the southern portion of the Calakmul Biosphere Reserve, a large protected area (7,238 km2) located in the Mexican state of Campeche at latitude 18 07N and
58 longitude 89 48W (See map in Chapter 1). The area re presents one of the northernmost and driest distribution area for the species. All the ar ea is covered by semi-dry tropical forest with three highly intermingled major vegetation types (the medium semi-perennial forest; the low semi-deciduous forest; and the low flooded fore st; named hereafter: Medium, Dry and Flooded respectively). The weather is classified as warm and sub-humid (Aw), with a mean annual temperature of 24.6o C. There is seasonal summer rainfall, with an annual average of 1,076.2 mm (CONAGUA, -Mexican Agency for Waterunpublishe d data). No major rivers transverses the region; rain water is stored in ponds that become the only sour ce of water during the dry season. The landscape is characterized by gentle rolli ng hills that reach 320 m above sea level. The southern portion of Calakmul Biosphere Reserve forms a continuous forest along with the Maya Biosphere Reserve in Guatemala. No human communities exist on the study area; the only human activity within th e study site is limited to tourism (a bout 8 cars per day. Calakmul Guards Pers. Comm.) and archeological resear ch in Calakmul Archeological City. White-Lipped Peccary Groups Locations and Behavior As part of an integral study that included ot her aspects of the ecology of this species, individuals from four groups of white-lipped peccary (named Blue, Red, Green and Yellow) were captured and radio tagged (d etailed capture and radio-teleme try description can be found in Chapter 2). The high cohesiveness of the groups of this species allowed me to monitor group movements by having only two or three individu als with radio-collars in each group. Groups were monitored basically by the homing method (White and Carroll 1990) that consisted in obtaining the broad directions of the groups from the only existing elevated points, two Mayan temples from Calakmul Archeological City, an d then walking towards the groups. Once groups were visually contacted, initial and subsequent locations were r ecorded every 15 minutes with a GPS device (Garmin eTrex), while c ontact lasted (from 1 to 6 hours) Seventy percent of the data
59 was recorded by homing, but 30 % of the data are triangulated radio-teleme try fixes as the only way to determine group locations. Groups were m onitored for different period of times during the 18 months that encompassed the study time (from March 2005 to August 2006). The Blue group was contacted for 17 months, the Red group fo r 11 continuous months, then left the area to an undetermined area for 5 months, until contac t was made again for another 2 months. The Green group was contacted for 12 continuous mont hs, and the Yellow for 7 months within a 10 months period. Groups of white-lipped peccary move at a large scale over the landscape, with some of the largest home ranges for ungulate species (Fra goso 1998, 1999, 2004; Chapter 2-this dissertation) Consequently, collecting detailed continuous movement data (every 15 minutes) was very challenging logistically and not al ways available for all group and months. Therefore, I estimated movements at the landscape scale by using lo cation data to estimate a minimum convex polygon (MCP) that represent the area us ed every month for the groups. I contrasted average polygon area for dry and rain season with the available data on movement rates for the same seasons to test if these variables were positively correlated. In this study I observed high fi delity of members to their gr oups and I did not see mingling of individuals between groups despite periodic close encounters, as close as 5 m in some occasions. However, group sizes varied mainly as a consequence of mortality and breeding during the study period (pers. obs.). Therefore, I kept a record of group size by accurately counting individuals in each group monthly while in contact with them (with exception of July 2006 for the Blue and Red group, February 2006 for the Green group and September 2005 for the Yellow group). Contact over time was maintained until the last radio-coll ared animal died or
60 when I lost contact with that group. While in contact with groups, I also recorded observations on feeding and drinking habits. Food and Water Availability Simultaneously to the collection of group lo cation data, I estimated fruit abundance on forest floor by counting fru its on randomly placed 2 m2 plots under each fruiting tree that was encountered on a 20 km network of transects. This grid of transects was semi-randomly (by using some transects placed at random and othe rs used to access areas where peccaries have moved) placed on the whole study area and enco mpassed the home range of the four groups. Two fruit species were included in analy zes because I found that these fruits ( Brosimum alicastrum Ramon, Manilkara zapota Zapote, and Talisia olivaeformis Guaya) accounted for the majority of the fruits on the forest floor (73.6 %) and because the former two species have been found to be heavily consumed by WLP in the Cala kmul region (S. Perez-Cortez and R. ReynaHurtado, unpublished data). Fruit abundance wa s calculated by summing the number of fruits counted in each 2 m2 plot in a given vegetation type on a given month. An index of abundance was then expressed as the number of fruits per kilometer of transect. Water availability was estimated by rainfa ll records from the nearest weather station located to the north of the st udy area. Flooded forests provided a significant source of water during the wet season, while ponds were basica lly the only source during the dry season. Therefore, I kept a record of the presence of water in ponds and flooded forests over time. Additionally, I recorded the presen ce / absence of invertebrates in the forest soil. I observed that during the wet season, white-lipped peccaries spent large amounts of time in the flooded forest eating earthworms and other invertebrate s. Data were collected on three 15 m2 plots in each of the three vegetation types for the last dry and wet season. I found that these invertebrates were only present in the wet soils of the Flooded forest during the wet season. Therefore, I calculated
61 the percentage of flooded forest inside each poly gon as an index of animal protein component on the diet of white-lipped peccary during the wet se ason. This percentage was calculated using Arc View 3.3 on a 2000 satellite imag e classified with ERDAS. Group Size and Precipitation Finally, I collected white-lipped peccary group size estimates and associated rainfall at 11 sites within their distribution ra nge from Mexico to Argentina (Tab le 3-5). Nine of these counts came from published studies where researchers ha d the opportunity to record repeated full counts of group sizes. A couple more observations were included from projects in progress where accurate counts have been provided, too. I explored the relationship between group size an d rainfall on other areas of the geographic range by using rainfall as an indicator of fore st productivity under the assumption that primary wetter forest produces more fruit than dryer fore st by having more primary humid tall forest with more tall-large fruit trees and by the demonstrated positive relationship between trees basal area and fruit productivity (Chapman et al. 1992, Wallace and Painter 2002), between rainfall and fruit productivity (Chapman et al. 2005), and be tween primary humid forest and fruit production (Altrichter et al. 2001, this study ). I collected published and n on-published data from typical areas where WLP has been studied and where co nfident estimates of group size have been reported. As a second step, from these 11 site s I selected only large areas (n=9; >1000 km2) where humans are absent, and frag mentation did not play a major role on group size. The list is not exhaustive but encompasses the more extreme areas (in terms of rainfall) where this species occurs (Table 3-5). I did not take in consider ation anecdotic or historic counts of group sizes. Statistical Analyses I used analysis of variance (ANOVA) and Post Hoc Tukey statistics to test if monthly estimates of group size and area used varied amo ng groups. I used multiple linear regressions to
62 determine the relationships between the area used monthly (as a repeated measure from each group) and several independent va riables: group size (# of indivi duals), monthly rainfall (mm), presence / absence of water in flooded forest us ing dummy variables (0 = absence, 1 = presence), and abundance index of fruits of B. alicastrum M. zapota and T. olivaeformis (# fruits per km of transect), and percen tage of flooded forest on the gr oups home range. These independent variable were tested for co rrelation among them and I found B. alicastrum was positively correlated with rainfall and wa ter availability on flooded fo rest and negatively with M. zapota and T. olivaeformis so I used only B. alicastrum on the analyzes because I observed that was the most consumed fruit and appears to be th e most important food resource for WLP. B. alicastrum was included with group sizes, and percentage of flooded forest in home range as final independent variables on the analyzes. The fit an d parsimony of the resulting models were tested by the Akaike Information Criterion for small sample size (AICc; Burnham and Anderson 1998). Subsequently, I ran additional re gressions by removing the non (or the less) significant variables one at a time in order to attribut e more variance to the remaining significant variables (Ganas and Robbins 2005). I estimated the standardized regression coefficient, which describes the change on the response variable caused by a change in on e of the independent variable (measured on SD units) while all other variables are held consta nt (Ganas and Robbins 2005). To explore the relationship between group size of WLP and annua l rainfall from other areas of the species geographic range, I used simple linear regr ession. Analyzes were performed in SPSS 11.5 (Statistical Package for So cial Science 1997) and SAS 9.1. Results Water availability in Calakmul forest was hi ghly seasonal, with no water available even at the Calakmul pond at the peak of the 2006 dry season (April-May 2006). After June in both years rainfall was present, and the flooded forests started to fill up (Fig. 3-1). Food availability as
63 represented by the three main fruits c onsumed by WLP was also highly seasonal. Talisia olivaeformis fruited once a year, in April-May, but fruit production was much higher in 2005 than 2006. B. alicastrum fructification coincided with the ra in season peak on both years. Fruits of Manilkara zapota were available almost all year round, but a peak was clearly distinguished during the dry season (Fig. 3-2). Earthworms and other invertebrates were only present in the wet season and mainly in the Flooded forest. The four studied groups of white-lipped peccary differed significantly in size: the largest group (Red) had at one point 31 members, wherea s the smallest (Green) had only 15 individuals left when the study finished. Testing groups monthly sizes I found the difference among groups was highly significant (F=88.7, p<0.01, n=4; Table 3-1) A post hoc Tukey test showed that the Red was different from the other three groups (p values <0.01). Th ere was a significant differences also in area used among groups where the two smallest groups (Green and Blue) used an area on average larger than the other two groups (F=3.1, p=0.03, n=4; Table 3-1). A post hoc Tukey test showed that the main difference was between the Blue and the Yellow groups, although not significant (p=0.06). Both the Blue and Green groups increased the ar ea they used in the dry season compared to the wet season in 2005, whereas the areas used by the Red and Yellow groups remained unchanged. For the 2006 dry season th e three groups remaining in contact (Red, Blue and Green) performed huge movements. These large movement s were most probably in order to locate water given that the original source of water (Calakmul Pond) dried up for first time in more than 16 years. For the 2006 wet season, only the Blue and Red groups remained in contact. The Blue group maintained a similar range area to that of the 2005 wet season, while the Red group ranged over a larger area than during the 2005 wet season (Table 3-1).
64 I tested if my estimates of area used trul y represent the movement rate on which the Ecological Constraint Model is ba sed, in order to be able to us e the former as a proxy for the latter (Ganas and Robbins 2005). I used the availabl e data for movement rate, and found first that hourly movement rate was identical for dry and wet season (dry season mean: 338 m/hour, SE=38.3 m, n=93; wet season mean: 339 m/hour, SE=42.4 m, n=80; Chapter 4). Second, when contrasting daily displacements, I found that WLP moved more during the wet season than during the dry season (dry season mean: 930 m/ day, SE=156.2 m, n=49; wet season mean: 1266 m/day, SE=247.5 m, n=33; Chapter 4). A similar resu lt was obtained using the average area used by the four groups during dry and wet season s 2005 (dry season monthly mean MCP: 3.33 km2, SE=0.6, n=11; wet season monthly mean MCP: 11.40 km2, SE=2.63, n=20). Thus I conclude that WLP moved at the same speed in both s easons, but spent more time on the move daily during the wet season, and that these differences were accurately represented by the area used. Therefore area used is a confid ent proxy for movement rate. I pooled the data from the four groups and te sted the effects of i ndependent variables on area used for all the time study period ( 2005 and 2006), then for 2005 dry and wet season combined, and for these seasons separated. Insuffici ent location data in 2 006 precluded statistical analyses among seasons for that year; contact with the Yellow group was lost and the Green group was in contact only during the dry season. The non-correlated variables included in the final models were group size, B. alicastrum abundance index (who was positively correlated with rainfall, water availability in flooded forest and negatively with M. zapota and T. olivaeformis ), and percentage of flooded forest on home range. When these models were developed we found that group size was significant ne gatively related to area used in both 2005 and 2005-2006 periods; B. alicastrum has a positive effect on area used for both, all time and
65 2005 period but was only statistica lly significant for the 2005 peri od, and percentage of flooded forest was only significant positively related w ith area used for the all-time period (Table 3-2; Fig. 3-3). For the 2005 wet and dry seasons sepa rately, only group size was significantly and negatively related to area used for both seasons (Table 3-2). When I reduced one variable at a time, the best parsimonious model according with AICc values, was the one including the three variables for the all-time study period and for the dry and wet season of 2005 the better model selected includes only group size and B. alicastrum index (Table 3-3) The negative relationship between group size and area used sugge sted that larger groups might range in habitat of better quality. I tested this possibility by us ing home range habitat composition to evaluate home range habitat qua lity. I found that compared to the other three groups, the largest group (Red) had the lowest am ount of Dry forest a forest type previously detected in this same study as the least preferre d by the same groups (Chapter 2), and had similar proportions of Medium and Flooded forest (the two forest type s that were highly selected by the same groups, Chapter 2) (Table 3-4). On the other side, the other three groups had lower proportions of Flooded forest (Green and Yellow) or Medium forest (Blue). A simple linear regression on the relations hip between group size from 11 sites across the geographic range of WLP and rainfall demonstrated that there is a positive relationship, where a small but steady increases in rainfall is associated with an increase in group size (Fig. 3-4; n=33, =0.048, R2=0.43, p<0.01). I ran a second analysis le aving out areas lesser than 1,000 km2 where habitat fragmentation may affect group size (i.e Caetetus Ecological St ation, Brazil, with 21.78 km2, Keuroghlian et al. 2004; and Corcovado National Park, Costa Rica, with 467.74 km2, Altrichter and Almeida 2002, Carrillo et al. 2002 ). The results were very similar and only slightly stronger (n=30, =0.053, R2=0.48, p<0.01).
66 Discussion Our data on areas used by white lipped peccari es in relation to food and water availability did not support the predictions of the Ecologica l Constraint Model as originally stated by Chapman and Chapman (2000a). The ECM predicts that given certain food availability, an increase in group size would cause an increase in movement rate for that group. In this study, smaller groups ranged over larger areas than larger groups in Calakmul Biosphere Reserve. Additionally, movements were str ongly associated with the search for water in the seasonally dry, water limited forest of Calakmul. Does that mean that the Ecologica l Constraint Model is worthless for this species? I beli eve not and there are several expl anations why in this case WLP group movement behavior did not fit well the model. Habitat Quality The Ecological Constraint Model assumes that food availability must remain similar for groups of different sizes to detect an increase in movement rate in larger groups. Despite the fact that the groups of WLP in this study shared th e same space temporarily, they occupied different areas at other times and had different home range s on the annual scale. Because areas differences might arise from differences in habitat quality, in particular food avai lability, I tested this possibility by quantifying the proportion of th e three vegetation types in the monthly home ranges for each group. I had determined previously that fruit abundance index was highest in the Medium forest (Chapter 2) while the Flooded forest provided earthworms and other invertebrates in high density duri ng the wet season. I had determined also that the Dry forest was the least preferred forest type, which they onl y crossed when moving between patches of other forest types (Chapter 2). The largest groups (Red) home range had the le ast amount of Dry forest. Conversely, the home ranges of the other groups not only include d higher proportions of Dry forest, but also a
67 lower proportion of the preferre d Flooded forest (Green and Yellow groups) or Medium forest (Blue group). In other words, the Red group seemed to range in a better combination of forest types as well as having the Calakmul pond in the center of its home range. These differences in forest types, i.e. potential food availabilities, within the home ranges could explain partially the lack of fit to the predictions of the Ecological Constraint Model and why group sizes were always negatively related with area used. Water Dependence White-lipped peccaries need to visit water sources daily (Sowls 1997, Fragoso 1998). It has been demonstrated that even on the driest areas where they i nhabit, such as the Chaco, they forage around the few existing water bodies (Sowls 1997). Clearly, several of the long movements and range behavior I observed were aimed at locating water sources. I also observed that the four groups were living sympatrical ly around the Calakmul pond during the dry season of 2005, most probably because it was the only av ailable source of water at that time. During water scarcity time groups only traveled to clos e areas where they could return daily to the Calakmul pond. They even foraged on sub-optimal foods items like herbs and flowers around the edge of the pond. When rain was abundant enough to fill the flooded forest three groups (Blue, Green and Yellow) returned to what I called thei r original areas, and onl y the largest (Red) group remained around year, Calakmul pond. Behavior on the wet season was contrasting. WLP groups then traveled farther and did not return to the pond for weeks, staying in the flooded forest where they satisfied their water requirements and fed on invertebrates, and tr aveling to the Medium forest where they fed on fruits ( B. alicastrum mainly). Additional evidence of water dependence in this species came duri ng the dry season 2006, when as result of a weak 2005 rainy season, Calakm ul pond dried up for first time in more than 16 years (Archeological City guards pers. comm.) .The Red group, which had stayed there for 11
68 months, then traveled to an undetermined southern site (as was determined by the direction they headed to before signal was lost), and did not return until the 2006 rain season. The Blue group returned to the Calakmul pond during the dry se ason 2006, when only a little water remained that later dried up. They stayed around for 10 weeks, and it was evident that almost all their movements were aimed at locating water. Severa l times they walked 2-3 km straight paths, ignoring food encountered (mainly M. zapota ), to visit some stones locally known sartenejas that usually stored water. At that time, however, these stones only had mud at the bottom. It was evident the WLP knew the location of the stones gi ven the directionality of their movements and how highly steeped the soil was around the stones, as well as the several paths that originated from there. Water dependence was also supported by the statistical model with the positive significance of B. alicastrum index for the dry-wet 2005 period. It is good to remember that this index was highly positively correlated with rainfa ll and water presence in the Flooded forest, so it is maybe possible that the increase in ar ea used were due to a combination of B. alicastrum and water availability during the rain season. There was also a time lag between rainfall events and accumulation of water on the forest floor and in ponds (2 months approximately, Fig. 3-1). Therefore, it is highly probable that the lands cape-scale movements performed by white-lipped peccary groups in Calakmul were linked to the spa tial and temporal distribution of water sources. This caused food availability to play a seconda ry role, and the Ecologi cal Constraint Model probably only apply at a short te mporal and spatial scale that was undetectable during this study. Group Sizes and Home Ranges Not only the smallest WLP groups in my study had the largest home range sizes (Blue and Green groups: respectively 97.5 and 82.4 km2 for 95 % with Fixed Kernel Method and 77.3 and 121.8 km2 for 100% Minimum Convex Polygon; Chapter 2), but these estimates were the largest
69 ever reported in the literature, with the singl e exception of a large gr oup on Maraca Island (134200 individuals; home range: 109.6 200 km2; Fragoso 1998, 2004). These large estimates are totally unusual for small groups like the ones ranging in Calakmul. For example in Corcovado National Park in Costa Rica, Carrillo et al. ( 2002) estimated the largest annual home range of a herd of between 40 animals as 37.2 km2; in the Atlantic forest of Brazil Keuroghlian et al. (2004) estimated annual home ranges for groups (mean 41.7 individuals per herd) living in fragments to average 18.7 km2. Therefore, the small group sizes and large home ranges found in my study lead us to hypothesize that the Ca lakmul forest landscape has strongly limiting resources for WLP, especially water. The larg e home range behavior observed in these small groups could be an adaptation to the spatial a nd temporal distribution of water as well as preferred food on the Calakmul landscape. A dditional support for the hypothesis of food limitation comes from the fact that the Calakmul Biosphere Reserve lacks large palm trees species, one of the most important food items for white-lipped peccary (Kiltie and Terborgh 1983, Fragoso 1994, Beck 2006). In Calakmul, only a small palm ( Chamaedora spp ) inhabits the area. The most important and abundant fruits in Calakmul forest for the white-lipped peccary, as well as for other species such as black howler monkeys Alouatta pigra (Rivera and Calm 2006), are B. alicastrum and M. zapota. The availability of the former is highly seasonal and coincides with the rain season. Therefore, the positive effect of this fruit abundance on group movement (Table 3-3) is confounded by the simu ltaneous availability of water in the Flooded forest that also entices group movement On the other hand the effect of the M. zapota and T. olivaeformis which are negatively correlated with B. alicastrum remained undetected here, although these species could be potential key re source for the white-li pped peccary and other
70 frugivorous species given that their fruit production peaks during the dry season (March-May) when other resources are limited. Group Sizes and Rainfall For areas where estimates of home range were absent but full counts of groups of whitelipped peccary were available, I found a positive correlation of th is latter variab le with annual rainfall. Assuming annual rainfall tr uly represents forest productivit y, at least in terms of fruit production (the food type major represented in WLP diet; Kiltie and Terborgh 1983, Sowls 1997), and water is not a limiting factor in areas w ith high rainfall, then this finding matches the predictions of the Ecological C onstraint Model that larger group sizes should exist in areas with greater food abundance. This prediction, in comb ination with the fact that the home ranges reported in literature are almost all (with the exception mentioned above) smaller than in Calakmul despite groups being larger elsewhere, implies that the increase in group size associated with increased rainfall is li kely due to an increase in food abundance. I conclude that this study supports the Ecol ogical Constraint Model as applied for whitelipped peccary on a broad, distribution-wide scale, ev en if the effect was not directly evident for the groups ranging over the Calakmul Biosphere Reserve because di fferential habitat quality per group and due to the effect of wa ter scarcity. Thus, the scale at which the model is evaluated is very important. Additionally, this study highlig hted the importance of including other limiting resources such as water availabil ity to fully evaluate the model. I showed that food availability plays a secondary role when water is scarce, a nd that landscape-scale movements in Calakmul could be more a consequence of searching for water than food. Living in cohesive but small groups and rangi ng over large areas allo ws the white-lipped peccary to survive in a forest like Calakmul Bios phere Reserve where resources (water and food) are seasonally limited and scattered over the la ndscape. Groups living in Calakmul Biosphere
71 Reserve represent an example of the extreme variation in the soci o-ecological behavior of this species and a successful strategy for cohesive groups that depend on patchily distributed resources. The Ecological Constraint Model prov ided a well suited theoretical framework for the analysis of the movements and group size of WLP. Scale issues and data collection need to be carefully considered to better understand the relationship between group size, travel rate and food availability in this species.
72 Table 3-1. Mean group size, monthly area used (km2) and test for differences among groups for both variables using analysis of varian ce on groups of white-lipped peccary in Calakmul Biosphere Reserve, Mexico. Groups Variables Blue Red Green Yellow F P Mean group size (# of indiv.) 19.0+/-0.7 N=16 30.0+/-0.3 n=12 17.0+/-0.5 n=12 20.0+/-0.8 n=5 88.7 <0.001 Mean monthly area used in the all study period 16.3+/-2.8 n=16 6.6+/-3.3 N=12 12.0+/-3.2 N=12 2.2+/0.0 n=5 3.1 0.03 Mean monthly area used in dry season 2005 4.3+/-1.8 n=3 1.5+/-0.0 n=3 4.8+/-0.5 n=3 2.2+/-0.0 n=2 Mean monthly area used in wet season 2005 20.9+/-4.8 n=6 1.6+/-0.3 n=6 14.8+/-4.3 N=6 2.1+/-0.0 n=3 Mean monthly area used in dry season 2006 17.9+/-5.4 N=5 14.9+/-11.8 n=2 13.7+/-10.0 n=3 -Mean monthly area used in wet season 2006 16.6+/-0.0 n=2 35.7+/-0.0 n=1 --Table 3-2. Models selected deve loped by all-time and 2005 study peri ods of the factors affecting area used by groups of white-lipped peccary in Calakmul Biosphere Reserve, Mexico. Time frame Response variables F n df adj. R2 std. P All study period MODEL Group size B. alicastrum (index) % of flooded forest 4.32 45 3 0.185 ---0.400 0.257 0.333 0.010 0.007 0.069 0.023 Dry and Wet 2005 MODEL Group size B. alicastrum (index) % of flooded forest 4.83 31 3 0.277 ---0.477 0.333 0.181 0.008 0.005 0.043 0.265 Dry 2005* Group size 5.19 11 1 0.295 -.605 0.049 Wet 2005* Group size 5.15 20 1 0.180 -.472 0.036 Notes: For the dry and wet 2005 separately, the only variable that has a significant correlation with area used was group size. For the dry and wet 2006 separated not enough sample size was collected.
73 Table 3-3. Selection of models explaining th e area used by WLP in Calakmul Biosphere Reserve, Mexico, using Akaike Informati on Criterion for small sample sizes AICc using Maximum Likelihood Estimation Method. Time frame Independent variab les included in the model AICc All study Group size; Ramon fruit; % flooded forest 760.3 Group size; % flooded forest 761.3 Group size; Ramon fruit 763.5 Group size 763.6 2005 Group size; Ramon fruit; % flooded forest 515.7 Group size; % of flooded forest 517.8 Group size; Ramon fruit 514.3 Group size 515.8 Table 3-4. Mean and standard errors of the pr oportions of the three vegetation types in the monthly home ranges of white-lipped peccary groups in Calakmul Biosphere Reserve, Mexico. Group ID Medium forest Flooded forest Dry forest Blue 0.23 +/0.04, n=16 0.38 +/0.03, n=16 0.37 +/0.02, n=16 Red 0.42 +/0.06, n=12 0.34 +/0.06, n=12 0.21 +/0.02, n=12 Green 0.53 +/0.05, n=12 0.08 +/0.02, n=12 0.36 +/0.03, n=12 Yellow 0.40 +/0.01, n=5 0.23 +/0.01, n=5 0.35 +/0.00, n=5
74 Table 3-5. Group sizes of whitelipped peccary as reported along the range of the species, and associated annual rainfalls. Sources are orde red in increased orde r of annual rainfall. Site Source Group size average (range) Vegetation Types Annual rainfall average (mm) Chaco Forest, Argentine Altrichter 2005 23 (7-50) Medium-tall xerophilous forest 575 Chaco Forest, Argentine Sowls 1997 45 (40-50) Medium-tall xerophilous forest 575 Calakmul Biosphere Reserve, Mexico This study 21 (14-35) Semi-dry tropical forest 1043 Caetetus Ecological Station, Brazil Keuroghlian et al 2004 42 (33-51) Tropical semideciduos, mesophytic, broadleaf forest 1400 Venezuela Hernandez et al 1995 42 (14-60) Seasonally dry forest 1469 Maraca Island, Brazil Fragoso 1998 114 (39-200) Primary tropical evergreen lowland rain forest 1950 Madidi National Park, Bolivia L. Maffei (pers. Comm..) 105 Tall evergreen tropical forest 2000 Lacandon Forest, Mexico Naranjo 2002 20 (5-60) Tall evergreen tropical forest 2226 Corcovado National Park, Costa Rica Altrichter and Almeida 2002; Carrillo et al. 2002 55 (40-70) Primary tropical evergreen lowland rain forest 2300 Los Amigos Biol. Station, Peru S. Carrillo (pers. comm..) 100 Tall evergreen lowland Amazon forest 2446 Manu National Park, Peru Kiltie and Terborgh 1983 106 (90-138) Tropical moist forest 2500
75 0 50 100 150 200 250Mar05 Apr05 May05 Jun05 Jul05 Aug05 Sep05 Oct05 Nov05 Dec05 Jan06, Feb06 Mar06 Apr06 May06 Jun06 Jul06 Rainfall Water in Flooded Forest Water in Calakmul Pond Figure 3-1. Water availability for the study peri od in Calakmul Biosphere Reserve, Mexico. Rainfall is in mm and presence of water in flooded forest and Calakmul pond has two values, 100=present, 0=absent. 0 10 20 30 40 50 60 70 80 90Mar05 Apr05 May05 Jun05 Jul05 Aug05 Sep05 Oct05 Nov05 Dec05 Jan06, Feb06 Mar06 Apr06 May06 Jun06 Jul06 B. alicastrum M. zapota T. olivaeformis Figure 3-2. Index of abundance per km of the th ree most abundant fruit species for the study period. Calakmul Biosphere Reserve, Mexico.
76 0 1000 2000 3000 4000 5000 010203040Group sizeArea Use d 0 1000 2000 3000 4000 5000 00.20.40.60.81% of Flooded Forest in Monthly Home RangeArea Used 0 1000 2000 3000 4000 5000 0102030405060Brosimum alicastrum Abundance indexArea Used Figure 3-3. Linear relationship be tween area used and group size, % of flooded forest in home range for the all time study peri od, and between area used and B. alicastrum abundance index for the dry-wet 2005 peri od, for groups of white-lipped peccary in the Calakmul Biosphere Reserve, Mexico.
77 0 50 100 150 200 250 050010001500200025003000 RainfallGroup Siz e Figure 3-4. Linear relationship between rainfa ll and group size of white -lipped peccary for 11 sites from Mexico to Argentina al ong the species distribution range.
78 CHAPTER 4 SEARCHING STRATEGIES OF WHITE-LIPPED PECCARY ( Tayassu pecari ) GROUPS IN A SEASONALLY-DRY TROPICAL FOREST Introduction A very important question in ecology is how animals find resources they need for their survival on heterogeneous natural landscapes (Stephen and Krebs 1986). Different searching strategies have been developed for finding spec ific resources. Centra l place foraging is a searching strategy where animals forage on an area but return periodically to a specific site (Orians and Pearson 1979, Schoener 1979). This m odel has received scien tific attention and qualitative and quantitative support since the la st 25 years (Giraldeu a nd Kramer 1982, Stephen and Krebs 1986). It has been demonstrated that animals behave like central-place foragers when they returned to a nest, burrow or sleeping site periodically, or that they can have multiple sites ( Ateles geoffroyi Chapman et al. 1989). Recently, the Lvy-walk model, a model that originated in physics for describing the behavior of moving particles, has been applied to animal mo vements (Shleinger and Klafter 1986, Viswanathan et al. 1996). Lvy-walks are a sp ecial type of random walks where the length of steps is not constant but chosen from a powe r-law tail distribution (V iswanathan et al. 1996), which confers the ability of searching at different spatial scales because of its scale-invariant properties. Lvy-walk scale invariance movements have been demonstrated in nature in species like Wandering Albatross ( Diomedea exulans ; Viswanathan et al. 1996), Jackal ( Canis adustus ; Atkinson et al. 2002), Reindeer ( Rangifer tarandus ; Mrrel et al. 2002), and more recently in Spider Monkeys ( Ateles geoffroyi ; Ramos-Fernandez et al. 2004), and Artic seals ( Halichoerus grypus ; Austin et al. 2004). Additionally, it has been demonstrated by simulation that searching following Lvy-walk movements confers advantag e of finding more reso urces than Brownian
79 (random) walk movements (Viswanathan et al. 1996). Additionally, some modifications that include random walks with Lvy-walks have be en proposed (Levy-modulated correlated random walk; Bartumeus et al. 2005). However, recentl y the existence of Lvy-walk movements has been challenged, and it has been proposed that Lvy-walk patterns could be consequence of classical random (Brownian) walk patterns generated by living in a patchy environment (Benhamou 2007). Additionally, there is a growi ng sense that the classical method of simple log-log plot does not provide the best way to detect Lvy-walk pattern and some alternatives methods are provided (Sims et al. 2007; see materials and methods). Finding resources for species that live in groups on unpredictable environments, like tropical forest, is a problem that has been solv ed differently. Some speci es split in sub-groups presenting the fusion-fission behavior (Chimpanzee, Pan troglodytes Balcom et al. 2000; Spider monkey, A. geoffroyi Chapman et al. 1989, Ramos-Fernandez et al. 2004), while other species remain together all year-round (Gorillas, Gorilla beringei Ganas and Robbins 2005; Howler monkeys, Alouatta pigra Chapman and Pavelka 2005; White-lipped peccary, Tayassu pecari Fragoso 1999; Chapter 3 -this dissertation, but se e, Carrillo et al. 2002, Keuroghlian et al. 2004). White-lipped peccary ( Tayassu pecari ; WLP hereafter forms the largest groups of ungulates living on dense tropical forest. Thes e groups perform long and continuous movements proposedly to allow the groups to find enough food while maintaining group cohesion. This species feeds on high energy food like fruits and nuts that generally have a patchy distribution on the landscape and requires visi t water-bodies on an almost daily basis (Kiltie and Terborgh 1983, Atrichter et al. 2001). It has been demonstrated also that WLP range over the largest home range for ungulate species living on tropi cal forest (Fragoso 1998, Chapter 2-this dissertation).
80 However, it is not known how this species fi nds specific resources that have a much localized distribution on the land scape, or how they cope with environmental seasonality and unpredictability which characterizes some tropical forests. The Calakmul Biosphere Reserve (CBR hereafter) in Southern Mexico is one of the ideal places to study how WLP groups search for resources in a heterogeneous landscape. Th is area is a seasonal forest where water and resources are temporarily scarce and highly local ized over a huge landscape and represents very different conditions from where WLP has been studied before. In this study I explored the searching stra tegies that WLP groups performed to find essential resources for its survival using a combination of radio-telemetry and direct observations. First, I described th e distribution of the important re sources for this species, such as preferred habitat type and water sources, on the Calakmul la ndscape. WLP movement patterns were then analyzed under the Lvy-walk (Viswanathan et al 1996) and the central place foraging models (Orians and Pearson 1979, Schoener 1979). Finally, I described and documented for first time on this species coordi nate travel behavior among groups and discussed it possible role as impor tant searching strategy. Material and Methods Study Area This study was conducted in the southern area of the Calakmul Biosphere Reserve (CBR) in the Mexican state of Campeche (see map in Chapter 1). The CBR is the largest protected tropical forest inMexico with 7,238 km2 and was decreed as protected area in 1989. The center of the study area is located at latitude of 18 07N and longitude of 89 48W, and lies in the heart of the southern area of the CBR. Accord ing to Kppen (modified by Garcia 1988) the Calakmul climate is warm and sub-humid ( Aw), with a mean annual temperature of 24.6o C. There is seasonal rainfall, mainly in summer a nd early fall, with an annual average of 1076.2
81 mm. Of the different forest associations (P ennington and Sarukhan 1998), four are widely distributed: Medium Sub-Perennial Forest (Medium), the more hum id of the region, where trees are between 15 to 25 m high; Low-Flooded Forest (F looded) that gets seas onally inundated after two to three 3 months of hea vy rains, and where trees are be tween 5 to 15 m high; and the Medium and Low Semi-Deciduous Forests, which bot h can be classified as dry forest (Dry) where trees range from 8 to 25 m high, but specie s composition differs from that of the Medium Semi-Perennial Forest. These four types of fo rest are highly intermingled within the area, although the humidity from northwest (driest) to southeast (wettest ) has an impact on the forest types too. The water in the area is obtained th rough precipitation since there is no permanent river system. Most of the rainfall percolates th rough the limestone, but some drains superficially and stores in ponds. These ponds constitute the on ly source for water for wildlife through the dry season. The study area is protected effectively against hunting and other human activities by two checkpoints along the only existing narrow road. Du ring the whole time there, no recent signs of human activities inside the forest were det ected. Human activity was only present at the Archeological site and on the road leading to it, and was limited to tourism and archeological research. For purposes of the analyses here, I further defined a more accurate study area as the Minimum Convex Polygon (MCP) that encompasses all the locations obtai ned from the study of four WLP groups (Fig 4-1). This MCP was made in Arc View 3.3 with the Animal Movement Analyst Extension (Hooge and Eichenlaub 1997). White-Lipped Peccary Data Collection. With the assistance of two persons a tota l of 17 white-lipped peccaries from four groups (named Red, Blue, Green and Yellow) were cap tured during the 2005 dry s eason and fitted with radio transmitters (details of capture and radi o-marked are in Chapter 2). These groups were living sympatrically around the Calakmul pond. Gi ven the cohesiveness of WLP groups, I only
82 radio-collared between two or three individuals in each group to keep contact with the group. Groups were monitored using the homing method (White and Garrot 1990) that consisted on getting directions to the groups from the only existing high points the two Mayan temples of the Calakmul Archeological city, and then walki ng towards the group. Once groups were contacted, initial and subsequent locations were recorded as frequently as every 15 minutes with a GPS device (Garmin eTrex) while cont act lasted (from 1 to 6 hours). Initially, groups responded to our presence by fleeing up to 100 m. When this happened twice in a day I left the groups and came back the next day. Subsequently, groups were more tolerant to our presence to the degree that I could sit as close as 4 m from them. The four groups were monitored for different periods of times during 18 months that encompassed the study time. WLP groups have some of the largest home ranges for ungul ate species (109-200 km2 for groups living in Maraca Island, Brazil, Fragoso 1998, 2004; and 98-121 km2 for the groups in this study, Chapter 2) Consequently, it was very difficult to locate a nd follow groups in regula r basis. Therefore, detailed continuous movement data (every 15 minutes) was not always available for all group and months. Distribution of Resources on the Study Area As part of a complementary study with the same WLP groups using Compositional Analysis (Aebischer et al. 1993), I found th ese WLP groups to selectively use ponds and Medium and Flooded forest while avoiding the tw o types of Dry forest (Chapter 2). The selection of Medium forest is mainly due to the presence of Ramon fruit ( Brosimum alicastrum ). Groups of WLP were observed foraging on B. alicastrum several times under the Medium forest. It was also demonstrated that B. alicastrum is produced in 83 % on the Medium forest (Chapter 3) and that this species is very important item in WLP diets (S. Perez-Cortez and R. ReynaHurtado, unpublished data, Sowls 1997, Fragoso 1998, Altrichter et al. 2001, Beck 2006). WLP
83 visited the Flooded forest, especially on the wet season, to forage on earthworms and other larvae present at that time, and to wa llow in the flooded areas (pers. obs.). Therefore, to know the spatial distribution of these resources I used also a previous classification of the forest t ypes and other features of the landscape made in GIS (Gerardo Garcia-Gil, ECOSUR-Chetumal) to identify the ponds existing within the MCP-derived group home ranges on the study area. I estimated the density of ponds by dividing the number of ponds over the MCP area. Additionally, the mean distances from ponds to the middle point of the MCP was obtained by using Arc View 3.3 with the An imal Movement Analyst Extension (Hooge and Eichenlaub 1997). In the course of the study I found that WLP also used some stony outcrops that store water temporarily ; these stones are called sartenejas I documented and geo-referenced the ones found with WLP use present. I used ERDAS and Arc View 3.3 to do a supervis ed classification of a satellite image to portray the spatial distribution of the two forest types (Fig. 4-2). I measured number of patches, density, the coefficient of variat ion of patch area (CV=100 standard deviation / mean), and the Contagion Index per forest type by using FR AGSTATS 3.3 (McGarigal and Marks 1994). The Contagion index measured the degree of aggrega tion and distribution of patches on the landscape (McGarigal and Marks 1994, Gergel and Turner 2002). Contagion indi ces have values from 0 to 1; lower values mean patches are dispersed uni formly over the landscape, while higher values result from patches with a high de gree of aggregation (clumpiness). Lvy Walk Pattern Detailed data on group movement and sequent ial data of many locations were very difficult to obtain given the large sc ale over which the groups moved (>100 km2, Chapter 2). Therefore, I could only use hourly and daily displ acement as the step length for analyses of the Lvy-walk pattern.
84 For data on hourly displacement I selected al l the consecutive locations from the four groups where I knew the position before and afte r one hour and the distances between those locations using Arc View 3.3. These hourly displ acement data were gathered between 8:00 am and 5:00 pm, which encompasses the time of the da y that I was able to follow the groups given the remote locations where they usually forage d. Localizations were recorded with a GPS (Garmin eTrex, Global Position System Device) that has accuracy between 5 -20 m. Daily displacements were recorded as the step length (measured using Arc View 3.3) between the two locations where I made first co ntact with the group in two consecutive days. Usually the first contact was made early in the morning, after an initial search for groups that lasted from 1 to 3 hours. I used 173 hourly displacements and 82 distances from consecutives days of the four groups. I analyzed the presence of Lvy-walk pattern in these two time scales by obtaining the regression slopes of the log10 transformation of frequency vs. log10 of the displacement distances divided in bins. I used the 2k logarithmic binning with prior no rmalization method (Sims et al. 2007). This method recently was proved to be a better method to minimize error at detecting Lvy-walk pattern than the other three methods (the simple log-log transformation plot, the 2k binning without transformation and the cu mulative distribution function). The 2k binning with prior normalization method consisted in defining the bins used to plot the frequencies by a 2k logarithmic way, where the bin k increase logarithmically (e. g.: instead of 1,2,3,4, etc. they increase 1,2,4,8, etc.). The accuracy of this met hod is further refined by a normalization of the data by dividing the frequency per logarithmic bin by bin width and n (the total number of steps) to obtain the probability density of each bin (Sims et al. 2007). I performed this analysis to all the hourly and daily displacements separately, and then I divided th is set of data for the dry and
85 the wet season respectively to dete ct possible seasonal differences on movement patterns. I tested the differences between the regression slopes with an F test (Sokal and Rohlf 1994). Central Place Foraging To determine if WLP group movements in relati on to a focal water source (the Calakmul pond) were following the Central Place Foragi ng Model, I constructed for each group a 2x2 contingency table to perform Chi Square analyses between the total days that I did contact with any groups in dry and the wet season and the nu mber of times they visited the Calakmul pond during these two seasons. A dditionally, I performed a t test between the mean distances from the locations where the four groups were found to this pond during the 2005 dry season (n=181) versus the distances from the 2005 wet season locations to the same pond (n=316). I also observed that during the 2005 dry season they perfo rmed occasional trips far away on days when there was some precipitation, but later returned to this pond. Ther efore, I performed a regression analysis of the monthly distances of the four groups from the pond versus rainfall. Travel Behavior While following the groups during the whole st udy period, I recorded several additional behaviors and interactions among groups. By havi ng animals radio-marked in the groups, I was able to confirm the identity of the groups and how thes e interactions were performed. In this way I documented instances where groups travel together in a coordinate way. I described how these interactions took place on time and form. All statistical analyzes we re performed on Microsoft-Ex cel (Microsoft XP), SPSS 11.5 (Statistical Package for So cial Science 1997) and SAS 9.1.
86 Results Distribution of Resources I found a very low density (1 pond every 9.3 km2) of water sources (ponds) within the MCP-derived home ranges of the four WLP groups (242.3 km2). Additionally, the temporal factor of water availability is also important given th at not all the ponds stored water every year. The Calakmul pond stored water during the 2005 dr y season but dried completely (for first time in at least 16 years; Archeological City Guards pers. comm.) during the harsh time (April-May) of 2006 dry season. I detected the same pattern in at least three other ponds that I visited regularly. The ponds were also dispersed over the landscape with a mean distance from ponds to the central point of the MCP of 6.5 km (SE=0.7 km; n=26). I documented and geo-referenced eight sartenejas that were visite d by the WLP groups when they were searching for water, especially during the 2006 dry season. These eight sartenejas were only detected by following the gr oups during the dry season; however, it is possible that others exists a nd were visited but remained und etected. Therefore, I did not estimate density or spatial dist ribution of them. However, th e eight known stony outcrops were dispersed over the landscape with a maximu m distance of 13.5 km between the two most separated. These reservoirs of water also were temporary; while some of them stored a little water even after the Calakmul pond dried up, others dried up at the same time as all ponds. These sartenejas are important because they can be filled with a small amount of precipitation while the ponds need significant large amounts of rain to start filling up (pers. obs.). Medium and Flooded forest presented a clum ped spatial distribution as was showed by the Contagion index of 0.70 and 0.74, respectively. Mean patch size for Medium forest was 5.6 ha (SD=219.8 ha, CV= 3,896), and for Flooded fo rest 2.1 ha (SD=57.2, CV=2,705). There were 2,032 patches of Medium forest w ith a density of 8.3 patches/ km2, representing in total 47.3 %
87 of the MCP, while there were 2,596 patches of Flooded forest with a density of 10.7 patches/ km2, representing 22.6 % of the total MCP area. The remaining area was composed for dry forest and some open areas. Lvy Walk Pattern Hourly displacements for WLP ranged fo r 20 to 2,656 m with a mean of 339 (SE=28, n=173). During the dry season the mean was 338 m (SE=38.3 m, n=93), and during the wet season the mean was 339 m (SE=42.4 m, n=80). Da ily displacement data ranged from 80 to 5,179 m with a mean of 1,056 m (SE=136 m, n=82). Daily movements were greater during the wet season than in the dry season (dry season mean: 930 m/day, SE=156.2 m, n=49; wet season mean: 1,266 m/day, SE=247.5 m, n=33). I found evidence of the Lvy-walk pattern in the two time scale m ovements (hourly and daily; Table 4-1). Because the data were collected with a GPS that has an accuracy of 5 to 20 m, I analyzed the hourly data at three starting bin size: 20, 50 and 100 m and all of them showed a coefficient ( ) that fit the Lvy-walk pattern (Table 4-1, Fig.4-2, 4-3). Based on the higher R2 value I decided to use 100 m as the start bin to perform additiona l analyses on the hourly data comparing dry and wet season. I found a very sim ilar coefficient between the seasons for the hourly data (F=35.15, p=0.00, df=1; Table 4-1 Fig. 4-4, 4-5). Daily movement data were analyzed at 100 a nd 200 m start bins and both of them showed the coefficient ( ) fits well under the expected valu es for the Lvy-walk pattern (1< <3; Table 41, Fig 4-6, 4-7). However, the coefficient of th e regression slope was lo wer than the hourly displacement, showing that more frequent long st eps are present on the daily displacements than for the hourly movements. Due to the better fit of the 200 m start bin (higher R2) I compared wet and dry season movements using the 200 m start bin. I found that during the wet season the WLP
88 performed longer steps than in the dry season. The difference between th e two regression slopes was statistically significant (F= 33.04, p=0.001, df=1; Table 4-1, 4-8, 4-9) Central Place Foraging Model WLP behaved like central place foragers around the Calakmul pond during the dry season of 2005 when this pond was the only source of wa ter. During this time the WLP foraged around the pond on low-quality food, like herbs, but oc casionally visiting patches of Zapote ( Manilkara zapota ) and Guaya fruits ( Talisia olivaeformis ) present in the forest (Chapter 2). Chi Square analyzes of the number of times groups were contacted on ponds versus non ponds on both seasons during 2005 showed that WLP visited p onds disproportionately during the dry season (Red group: X2=12.82, p=0.001, df=1, n=63; Blue group: X2=13.81, p=0.001, df=1, n=65; Green group: X2=18.19, p=0.001, df=1, n=59; Yellow group: X2=11.25, p=0.001, df=1, n=28), with the Calakmul pond being the one more visited (>90 %) for all the groups. Also, I found that during 2005, the mean distance from where the WLP groups foraged to the Calakmul pond was smaller for the dry season (mean= 595 m, SE=59 m, n= 181) than for the wet season (mean=2560 m, SE=170, n=316), showing that they stayed closed to this water source during the dry season. Finally, I found that from March to Sept ember 2005, the mean distance from where all the groups foraged to the Calakmul pond increase d positively with the amount of precipitation that fell during these months (F=23.68, p=0.004, R2=0.82, n=7). After that month, precipitation was not longer a good predictor of the distances from the pond. Coordinate Travel Behavior The four groups were living sympatrically around the Calakmul pond when I first contacted them during the dry season 2005. While maintaining individual group cohesion, the groups tolerated each other very well when mee ting at the pond edges. Groups were observed to rest 5 m apart at times. I documented on two o ccasions the green group joined another group and
89 travel together for periods gr eater than 20 days but then separated again. The Green group probably joined the Blue group around 4 May 2005, when some rain allowed them to leave the pond and travel to a patch of Medium forest 8 km to the north. I located th e two groups together for first time in the Medium forest. For the next 33 days they traveled together with the Green group always behind and following the Blue group fr om 40 to 100 m. In that way they visited the Calakmul pond several times and foraged in ar eas around a radius of 8 km from the pond. On 6 June 2005, the groups split and they remained on the same numbers and with their respective radio-marked animals as before. After these events the Blue, Green and Yello w groups left the area and travel to three separate areas in the North (sep arated by 17 km approximately). I believe these areas were their original home areas. The Red group remained around the Calakmul pond. However, between 23 to 25 September 2005, the Green group traveled back to the Calakmul pond and joined the Red group in the same way that they did with th e Blue group before. They stayed close from 28 September to 19 October 2005, when they separa ted again keeping the same numbers and the same radio-marked animals. The groups traveled in a radius of 6 km from the Calakmul pond without visiting the pond given it was the core of the wet season. The way they traveled was very similar to before, with the Green group behind 40 to 100 m the Red group. I observed that when together both groups (Blue-Green and Red-Gree n) not only traveled together but behaved similarly, resting and foraging at the same time. It is worth mentioning that at the time the Green joined the Blue group the former group size was 20 to19 individuals, while the latter was 25 to 23 individuals in size. When the Green group joined the Red group this latter group was 31 individuals, while the form er group was only 16 in size.
90 Discussion In this study I demonstrated that important re sources such as water reservoirs for WLP are dispersed and non-uniformly spatially-distribute d over the landscape at Calakmul Biosphere Reserve and can be temporally scarce. Preferred forest types that host pa tches of selected food like fruits (Medium forest: Brosimum alicastrum Talisia olivaeformis and Manilkara zapota ) or earthworms (Flooded forest) are also dispersed ov er the landscape and in termingled with other less preferred forest such as Dry forest. Whitelipped peccary must visit water on a daily basis and they feed on high energy food such as fruits and invertebrates for more than 60% of their diet (Kiltie and Terborgh 1983, Fr agoso 1994, Sowls 1997, this study). To access water in Calakmul Biosphere Reserve during the dry season WLP must travel long distances or stayed close to the few existin g storage sites. Availabl e food also was widely distributed spatially and seasonally scarce. Add itionally, this species liv es in highly cohesive groups that feed, travel and rest together all year long. So a question arises: How do WLP groups survive in a place like Calakmul where the resour ces are temporarily scarce, spatially dispersed, while keeping the groups cohesiveness? It is possible that the three s earching strategies I focused on in this study play an important role on the WLP survival while preventing grou ps from splitting. I found evidence that WLP groups move over the landscape in a pattern resembling the Lvy-walk model. Lvy-walk movements are scale invariant (Shlesinger and Klafter 1986, Visw anathan et al. 1996; Bartumeus et al. 2005, Benhamou 2007), th is allows the group to search fo r resources in detail at the small scale (with lots of small steps), while occasio nally making extended searches into the landscape scale (with few long movements) (Atkinson et al. 2002, Mrrel et al. 2002, Ramos-Fernandez et al. 2004, Austin et al. 2004). In my study the ev idence for Lvy-walk pattern in WLP movements is present at two time scales (hou rly and daily), implying that this pattern is scale invariant as
91 groups move in different spatial scale. The m ovements of WLP over the study time confirm this pattern of movement. Some WLP groups ranged over 100 km2 area (Blue and Green, and possible Red group; Chapter 2) and performed direct movements as long as 17 km in 2 or 3 days. When WLP move, they can travel to areas far away that have not b een visited recently; therefore, they access patches of resources that are temporarily new for them It is possible then, that by performing Lvy-walk movements this searching stra tegy allows them to reach patches of forest or water-bodies in a landscape scale of tens of km, and by living in cohesive groups make an efficient uses of resources by visiting patches that have less probability of being visited by other groups, and uses spatial memory in conjunction (Kiltie and Terborgh 1983). I did not find seasonal differences in the L vy-walk pattern for the hourly data. However, the daily data showed a lower coefficient fo r the wet season which means the group performed longer trips during the wet season than the dry season. This could be consequence of groups not being constrained by water availability since the flooded forest is fille d, and/or searching for B. alicastrum fruits that are available in the rain se ason and are present almost entirely in the Medium forest or invertebrates in the Flooded forest. The diffe rences between hourly and daily data could be interpreted as WLP moving at a sa me hourly pattern (speed, distance, frequency) either in dry or wet season, but during the wet season they keep moving longer (more hours) than during the dry season, which refl ects at the daily scale. I also demonstrated that at least for the 2005 dry season the four WLP groups behave like central place foragers around the only existing water source, the Calakmul pond. They performed occasional long trips (>8 km) in search of res ources, but returned always to the pond until the rainfall was enough to fill the Fl ooded forest and other ponds. When they were constrained to the pond they were seen even eating low-quality food like herbs around th e edges of the pond.
92 Rainfall seems to be a good predictor of how far the groups will venture from the Calakmul pond during a dry season, at least from March to September 2005. Initiation of the rain season potentially stimulated the groups to move farther from ponds, maybe visiting sartenejas which are easy to fill at that time, eventually movi ng to the areas where they spend the rest of the wet season. This pattern was demonstrated by the Yellow and Blue group after the Flooded forest was filled which started around July a nd lasted until November same year. After the Flooded forest was filled they not longer depend on the rainfall pattern so this explains why the relationship was not significant after September. This time lag between rainfall and movements has also been documented for elephant moveme nts in Botswana where some herds performed movements aimed to visit patches of vegetation onl y after 3 to 4 weeks of rainfall (Cushman et al. 2005). The combination of central place foraging and Lvy-walk movements seems to be an efficient strategy for WLP living in Calakmul Bi osphere Reserve. Because they have access to resources that are secure (such as water) they ca n also explore from time to time areas far away and monitor when the conditions change that allow them to move away. WLP need in addition some kind of spatial me mory to find resources very localized and small like ponds or even smaller like sartenejas I found evidence this trait while following the groups. Several times during the core of the dr y season they performed straight movements between 2-4 km moving in a single line, ignori ng fruit patches, until they arrived at some sartenejas or ponds. Additionally, finding these sartenejas did not seems to be a fortuitous event, given that there was physical ev idence of high WLP use of the area around them (the soil was compacted from hooves prints, several paths appr oached from different directions, and the vegetation was absent around the st ony outcrops). Use of ponds and sartenejas also indicate
93 multiple central place foraging behavior around these water sources. Multiple central place foraging was demonstrated for Spider monkeys in Costa Rica where they selected different sleeping sites and temporarily behaved like centr al place foragers around them, but moving after some time to a new site (Chapman et al. 1989). This model remains to be tested for WLP when more evidence is collected. If this spatial knowledge is st orage within the groups memory, then living in large groups confers advantages in this sense. Probably, th e association of two groups while traveling, as I documented here, increased the ch ances of finding these highly local ized resources. It could be that the Green group, which was always the smalle r-sized group, associated with larger groups to learn where and when the resources were av ailable. Additional evidence supporting this hypothesis is the fact that the Green group always followed the larger Blue or the Red group. In total the four groups moved ove r a huge area of almost 250 km2. Although they showed different home ranges inside this area, they shared common knowledge of areas around the Calakmul pond (Chapter 2). However, not a ll groups seemed to know the whole area. For example when the Calakmul pond dried up in 2006, the Red group left the area and traveled to the South to an undetermined ar ea, and returned 5 months late r (as determined for the radiotelemetry fixes when they left and when they returned). During this time the Blue and Yellow groups headed north and the Green traveled to the southeast where I lost c ontact with them after 12 months. Another evidence of lim ited knowledge of resource availability was when in the dry season 2006 the Blue group returned to the Ca lakmul pond. They found that it dried up, but instead of leaving the area, they spent around 10 weeks visiting some sartenejas that had little water (some only had mud at the bottom at that time). By the end of that dry season the group was in a very bad physical condition and in re duced numbers (group size=16). When the rain
94 season started they immediately headed to the no rth area. It seemed like this group was trapped without water in an area that always have had water but that was located on the border of their spatial knowledge and they do not know where else to go. Ramos-Fernandez et al. (2004) interpreted th e Lvy-walk pattern found in Spider monkeys as movements aimed to localize fr uiting trees that are in the de tection range. However, the authors pointed that their movements also indicate that they were searching for trees that were found out of this immediate detection range and where spatial memory could also play a role, similarly to what I found in this study with the WLP. Apparen tly a combination of searching strategies like Lvy-walk movements, central place foraging and travel association between groups allows WLP groups to fo rage efficiently over a landscape where resources are spatially dispersed, temporarily available and where unpredictability also plays a role. The searching strategies showed here seems to be complementary strategies with living in small groups and ranging over large areas as was demonstrated for these groups in comparison with much larger groups of the same species inhabiting other fore sts (Chapter 3). The combination of all these strategies allows WLP to survive in a sub-optimal forest like Calakmul Bi osphere Reserve. If we want to conserve WLP in Calakmul Biosphere Reserve and maintain th is area as the stronghold for the conservation of the species in Mexico, as has been previously identified (March 1990; Chapter 5this dissertation), we need to assure that the natural conditions of the landscape are preserved for the species to perform all these amazing strategies for its survival. A better understanding of the possible role of the memory and the spatial knowledge related to detailed group movement would be very interesting ecological research on this species.
95 Table 4-1. Regression slopes coefficients ( ) for log10 of frequency of hourly and daily displacement length versus log10 of 2k logarithmic bin with normalization, staring at 20, 50 and 100 m for hourly data, and 100 a nd 200 m for daily data, for groups of white-lipped peccary ( Tayassu pecari ), Calakmul Biosphere Reserve,Mexico. Time Scale Period and sample Bin Size (2k Binning w/normal.) N sample Hour All-Seasons 20 m -1.61 ( r2= 0.84) 173 50 m -1.72 ( r2= 0.90) 100 m -2.19 ( r2= 0.95) Dry Season 100 m -2.16 ( r2= 0.93) 93 Wet Season 100 m -2.22 ( r2= 0.96) 80 Day All-Seasons 100 m -1.22 ( r2= 0.81) 82 200 m -1.57 ( r2= 0.94) Dry Season 200 m -1.65 ( r2= 0.93) 49 Wet Season 200 m -1.46 ( r2= 0.94) 33 Figure 4-1. Minimum convex polygon of all the lo calizations from the four groups of whitelipped peccary ( Tayassu pecari ), and forest type patch distribution in Calakmul Biosphere Reserve, Mexico.
96 Figure 4-2. All-seasons frequency of steps length per hour with a bin size of 100 m, for whitelipped peccary. -6 -4 -2 0 01234 Logarithmic Bin Size (Log10)Frequency (Log10) Figure 4-3. Regression slope for all-seasons hourly data with a bin size of 100 m and 2k logarithmic binning with normalization method = 2.19 0 10 20 30 40 50 60 135791113151719212325272931 Bin Size 100 mFrequenc y
97 0 5 10 15 20 25 30 1357911131517192123252729 Bin Size 200 mFrequenc y Figure 4-4. All-seasons frequency of steps length per day with a bin size of 200 m, for whitelipped peccary. -6 -4 -2 001234Logarithmic Bin Size (Log10)Frequency (Log10) Figure 4-5. Regression slope for all-seasons daily data with a bin size of 200 m with 2k logarithmic binning with normalization method. = 1.57
98 CHAPTER 5 HUNTING PATTERNS, POPULATION DE NSITY, GROUP SIZE AND THE CONSERVATION OF WHI TE-LIPPED PECCARY ( Tayassu pecari ) IN THE CALAKMUL REGION OF MEXICO Introduction The white-lipped peccary (WLP; Tayassu pecari, L. 1795), a social Neotropical ungulate, forms large groups typically ranging in size from 10 to 300 animals; however, anecdotal sightings of more than 1,000 individuals have been reported (Leopold 1959, Mayer and Wetzel 1987, Emmons and Feer 1990, Alvarez del Toro 1991, Bodmer et al. 1997, Sowls 1997, Fragoso 1998). This is an exceptional behavior for an ungul ate that lives in dens e tropical forests, and only the bearded-pig ( Sus barbatus ) of Borneo has been observed to form similar groups, as large as 300 individuals (C aldecott et al. 1993). In the last 20 years WLP has become increasin gly rare in Mexico and Central America. Leopold (1959) noted more than 40 years ago th at in Mexico WLP was among the first species to disappear when a forest becomes perturbed, and that smaller herds were a consequence of human perturbation. Today, WLP have disappeared from the Ve racruz, Tabasco, and Yucatan states and survive only in a few reserves of Oaxaca, Chiapas, Campeche, and Quintana Roo (March 1993, WCS 2005). Some researchers argue th at in areas where this species seems to be disappearing, small herds of fewer than 10 ar e common (Emmons and Feer 1990). March (1993) states that frequent reporting of smaller groups in some areas is probably correlated with increased hunting pressure. In some instances hunters can eliminate an entire group as Peres (1996) reported when a group of 20 hunters kill ed 82 WLP in one day. Reyna-Hurtado (2002) documented that 13 individuals were killed in a single day by a group of hunters in the Calakmul region of Mexico.
99 The Calakmul Region in Campeche State is one of the few places where this species survives in Mexico, and it has been documented that the WLP is among the top five preferred species for subsistence hunters (Escamilla et al. 2000, Weber 2000, Reyna-Hurtado et al. 1999, 2002, Weber et al. 2006). The relative abundance of th is species was estimated to be three times lower in the hunted areas than in the protected area (Reyna a nd Tanner 2007). Additionally, legal sport hunting on this species is taking pl ace under the UMAs scheme (Unit for Wildlife Management, SEMARNAT 1997). However, there is evidence that UMAs are suffering several flaws in their administration, that population estim ates are weak and that there is a lack of information regarding breeding season and other sound biological informa tion that are essential to manage the species (Weber et al. 2006). To provide an initial understanding of the na tural range of group sizes of WLP and the status of groups and population in areas under hu man influence, I quantif ied WLP group sizes in the largest protected tropical forest in Mexico, the Calakmul Biosphere Reserve (CBR), where hunting and other human activities have been o fficially prohibited since 1989, and, for three adjacent communal forests ( ejidos ) where subsistence hunting, as well as logging and other extractive activities, are curre ntly taking place. A fourth hunted place (Calakmul-limit) was included that lies in an ar ea between the Reserve and one ejido where illegal hunting has been documented occasionally. Additionally, I estima ted WLP density in the CBR and collected information on the birth season and age structure of the groups. These data constituted the first report of WLP group size, density, breeding season, and age structure for the area and one of the few in Mexico (see Leopold 1959, Alvarez del Toro 1991, March 1990, March 1993, Naranjo 2002).
100 Methods Study Area This study was conducted in the municipality of Calakmul, Campeche, Mexico (See map in Chapter 1). The Calakmul region is located in the southern-central part of the Yucatan Peninsula (19o15-17o45 N, and 90o1089o15W). The region is mostly a protected tropical forest surrounded by areas cleared for the slash and burn agriculture. Calakmul is a mosaic of different types of tropi cal forest, ranging from low-deciduous forest in the north to tall-evergreen forest in the extrem e southeast. These differences in habitat are associated with a moisture gradient that increases from north to south a nd less clearly from west to east (Ucan et al. 2000). Desp ite the occasional pres ence of jaguar-hunters, chicle-tappers, and archeological looters during the fi st half of the past century, this region has remained largely undisturbed since the Mayans abandoned it 1,100 y ears ago. In the 1940s the colonization of the area began with the creation of the Zoh Laguna village as a center fo r logging operations and a base-camp for the extraction of chicle -gum (Ericson 1997). In the 1970s, when the Mexican government encouraged the colonization of the last frontier in Mexico, the humid tropics, Calakmul received a large influx of people from the central and southern states. This colonization process brought environmental change s, and a municipality that presently includes 114 human communities and an estimated population of 30,000 people (INEGI 2005). In 1989, 7,231.85 km2 of the forest in the region were decreed as biosphere reserve. Calakmul Biosphere Reserve (CBR) is now the second largest reserve and the largest protected tropical forest in Mexico. The CBR is adjacent to the Maya Biosph ere Reserve in Guatemala and together with the Rio Bravo-Milpas conservation area in Belize, they conform the larg est protected tropical forest in Mesoamerica with almost 20,000.00 km2. Today a mosaic of so cial conditions and land tenure systems surround the reserve, and spor t and subsistence hunting, as well as other
101 extractive activities, are comm on (Reyna-Hurtado et al. 1999, Weber 2000, Escamilla et al. 2000, Reyna-Hurtado 2002). Group Sizes, Age Structure and Breeding Season Group sizes, age structure and breeding season of WLP were recorded during field surveys at five sites (Fig. 5-1). Three sites were communal forests (formally called ejidos : Nuevo Becal: 520 km2; 20 de Noviembre: 280 km2; and Xbonil: 400 km2) where subsistence hunting, as well as selective logging and extraction of forest products (palm leaf, honey, mahogany seed, etc.) are common activities. The fourth site (Calakmullimit) encompasses some ponds located in the border between the southern core of the CBR and an ejido in which hunting occurred. In this site I detected recent huntin g activity (fresh cartridges) on the area bordering the CBR and the ejido I also include observations from another pond whic h lies 4 km south of this border because I determined, based in the radio-telemetry study, that WLP groups can range over 120 km2 in the CBR (Chapter 2). Therefore, is hi ghly probable that groups visiting this pond also travel near the ejido and become subject to hunting pressure there. The last site was the southern core area of the CBR (with approximately 3,500 km2). This site was at least 40 km south of the nearest human population center and no hunting occurred there due to the natural isolation of the area and to the law enforcement w ith two check points along the on ly road existing on the area The number of individuals and age compositions of several groups were collected for each area. Age was estimated based on size, and animals were categorized into three classes: juvenile (from 1 day old to half si ze of an adult), sub-adult (from half size of an adult to slightly smaller than adults), and adults (full gr own). I recorded birth season (at least the month when the birth took place) by noting the first time when newbor ns were observed and in a few occasions by examining fetus from animals hunted and projec ting a time to be born basis on their develop.
102 Data on group sizes were obtained from different sources. First, I searched for and located groups of WLP with the assistance of two local resident hunters on th e CBR and Calakmul-Limit site. Searches and data collection were made between January to May 2001 and for a second period between March 2005 and August 2006 (Chapt er 2). Searches were made primarily between 7:00 and 14:00 and concentrated on perm anent and temporary water sources, the main source of water for wildlife in the region duri ng the January to May dry season. I primarily stayed near water sources and waited for the peccaries to come to the water. WLP are not territorial and more than one group can be seen together (Fragoso 1994, Carrillo et al. 2002, Chapter 4-this dissertation). For example, in Corcovado National Park, Costa Rica, Carrillo et al. (2002) found several gr oups being sympatric in what appear to be a single super herd. For purposes of a radio-telemetry study I radi o-marked individuals in four groups in the 2005-2006 period (Chapt er 2), so I could differentiate the individual groups. In the CBR I found spatial and temporal overlap among th e four groups use patt erns and for the groups to travel together. I found groups to be very cohe sive without exchange of individuals (Chapter 2). Consequently, to reduce the possibility of counting the same group tw ice only sightings made in different locations or years were recorded. When several groups were seen at the same time and place, only groups who could clearly be differentiated (through direct observations or because they were radio-marked) were counted. In addition to field observations information of group sizes was collected also from hunters in the three ejidos (January-May 2001) and from direct observations made by me while doing fieldwork and from my assistants (Mar ch 2005-August 2006) while performing their daily subsistence activities (raising co rn, collecting honey, cutting timbe r, etc). I only considered observations made by hunters and assistants in op enings (e.g., ponds, forest clearings) and when,
103 based on a structured interview, I concluded that the observer was confident of group size. In several cases, observations were made by more than one person, so I interviewed each separately and decided to keep or delete the sighting base d on the consistency of the information. Seven reports were discarded since the observer was not confident of the group size or they saw only a portion of the group. In all cases I asked questions to confirm that WLP was the species sighted and not collared peccary. On three occasions I saw skulls, hair and body parts of the animals hunted and positively identified them as WLP re mains (Aranda 1981). Finally, a single sighting of a group made by me in May 2000 at one of the water sources in the CBR was also included in the results. Density Estimation I estimated WLP density for the CBR base d on a Minimum Convex Polygon constructed from location fixes obtained while searching the area several times every month (from the period of 2005-2006) when trying to make contact with the four radio-marked groups (named Red, Blue, Green and Yellow) during a 10 months period. I favored this estimate because from March to December 2005, I had continuous contacted with the four groups and had precise counts of the size of at least three groups (with the ex ception of the Yellow group). Additionally, during intense field work on these 10 months I did not se e any WLP sign that could not attributed to the four groups. Every time I encountered a WLP sign I confirmed the identity of the group traveling there by using the radio-marked animals in each group. Group observations for more than 200 days confirmed high fidelity of marked indi viduals to their respect ive group (Chapter 2). After this 10 month period I lost radio cont act with the Yellow group. At that time, I was not confident if any unmarked group sighted wa s the Yellow or a new group. Additionally, two groups (Green and Red) left th e area and traveled to an unide ntified area during the dry season 2006. Therefore, I calculated a minimum estimate of density by dividing the sum of individuals
104 from the four groups over the whole area sear ched (that encompassed the four groups home ranges) during this period of tim e in which I was confident no additional group was living there. Additionally I estimated the numbe r of individuals and groups exis ting on the Southern area of the CBR by extrapolating these estimates to a larger polygon that include the study area and almost all the southern area of the CBR, with an exception of 5 km buffer from the border with the ejidos This extrapolated area shares similar prot ection status and the ha bitat type have not changed dramatically. Seasonally of the Hunting Patterns Data about subsistence hun ting events in the three ejidos were collected on an opportunistic way during the pe riod 2005-2006, by informal talks wi th subsistence hunters and through the help of field assistants who lived ther e. I also included data that were collected in 2001 (Reyna-Hurtado 2002) by systematic surveys on the same three comm unities. Additionally, for the sport-legal hunting I c onsulted the people in charge of the program (SEMARNAT, Campeche-office, Mexico) to obtain the date s about legal times to hunt WLP under UMAs guidelines (SEMARNAT 1997). Here I focus in the season when WLP hunting took place. Statistical Analyses A Mann-Whitney U -test was used to compare group si ze from non-hunted vs hunted areas together (hunted and Calakmul-limit). A subseque nt analysis where the three different areas (non-hunted, hunted and Calakmul-limit) were compared was done using Kruskal Wallis test for differences among groups. Home range estimates were obtained in Arc View 3.3 using the Animal Movement Analyst Extension (Hooge a nd Eichenlaub 1997). All st atistical analyses were performed on SPSS (Statistical Package for Social Sciences, 1997).
105 Results Group Size in Hunted and Non-Hunted Areas The sizes of 24 WLP groups were recorded in the four sites from 2001 to 2007 (Table 5-1). Ten observations were obtained from ejidos 9 from the CBR and 5 from Calakmul-limit. Considering Calakmul-limit as a hunted site, groups were larger on the non-hunted area (Median=25, SE=1.84 n=9), than in the hunted areas (Median=16, SE=1.84, n=15) (Fig. 5-2) and the difference was significant ( Mann-Whitney test p=0.02, n=24). If I compare ejidos Calakmullimit and the CBR independently, groups were smaller on Calakmul-limit (Median=15, SE=0.86, n=5) than in the ejidos (Median=20, SE=2.41, n=10) and in the non-hunted areas (Median=25, SE=1.84, n=9), and the differences were significant ( Kruskal Wallis test p=0.01). When contrasting only observations from ejidos versus those from the CBR, groups were bigger on the CBR but the differences were not statistically significant ( Mann-Whitney test p=0.18, n=19). Group Age Structure Age structure was recorded in detail from 21 groups. These 21 groups summed a total of 444 WLP individuals observed. I found that ad ults comprised 90.3 % (401) of the population, sub-adults 3.2 % (14), and juveni les 6.5 % (29) (Fig. 5-3). Eigh t of these groups contained only adults. In 6 groups I observed thr ee age classes: 78.9 % adults 8.8 % sub-adults and 12.3 % juveniles. In 6 I observed only adults (89.4 %) and juveniles (10.6 %), and just in 1 group I observed adults (84.6 %) and sub-adults ( 15.4 %) without juven iles (Table 5-1). Breeding Season From 2001 to 2007, I recorded 19 birth events on at least 17 different groups from all the study sites except from one of the ejidos (20 de Noviembre). In 18 cases the birth events took place in a period between late De cember and May with the peak of the breeding season being in
106 JanuaryFebruary (Fig.5-4). The single excep tion was a group with ne wborns sighted in October. In 5 times I observed that fe males have given birth to twins. Seasonality of Hunting I recorded 21 hunting events during the study period. These events were done by subsistence hunters (19) on the three ejidos and sport hunters (2) on one ejido where legal hunts of this species are allowed through the UMAs system. All the hunting events but three took places in the dry season, from late December to mid May. The exceptions were two hunting events recorded on November 2002 and October 2004 in the Nuevo Becal ejido and one from an undetermined season. The legal time to hunt this species is always from early April to mid May (Direccion General de Vida Silvestre, Campeche, pers. comm.). Of these 21 events 17 were successful and hunters took betw een 1 to 13 white-lipped pecc aries per event. Sport and subsistence hunters did not discrimi nate against females, whether th ey are lactating or pregnant (pers. obs.). Finally, I recorded that in two months of 2005, subs istence hunters in combination with sport hunters harvested 27 WLP from a group of 29 individuals, leavi ng a single mother and her newborn (Weber et al. 2006). This occurren ce documented the almost complete elimination of a group in an ejido where a previous survey done 4 months before had detected just 3 to 4 groups inhabiting the whole ejido (R. Reyna-Hurtado unpublished data). Density Estimation The entire area searched where only the four radio-marked groups lived during a 10 month period of continuous radio contact was estimated to be 236.7 km2. The sizes of the four groups were 31, 25, 25 and 20 at the beginning of the radio-telemetry study for a total of 101 white-lipped peccaries. Therefore, the estimated density was 0.43 / km2, or 1 WLP/2.34 km2. I used this density to estimate WLP density on the whole southern area of the Calakmul Biosphere Reserve. For the extrapolation exercise, I defined a larger area that encompassed almost all the
107 southern area of the CBR except a buffer of 5 km between the reserve and the ejidos (Fig. 5-5). This area selected comprised 3,509.41 km2 and has the same status of legal protection, similar level of isolation and, more importantly, they ha ve similar proportions of forest types (according to analyzes performed in Arc View 3.3 using the classification originated in 1995 by ECOSURG. Garcia-Gil unpublished data; Table 5-2). Ther efore, assuming the white-lipped peccaries equally disperse over this area, I estimate 1,493 individuals on the whole southern area of Calakmul Biosphere Reserve. Considering that the mean size of the 9 groups recorded in the CBR was 25.7 individuals per group, then approx imately 58 groups could exist on the total southern area of the Calakmul Biosphere Reserve. Discussion Group Size The data reported here are the first for th e Calakmul region and among the first data for WLP groups in Mexico (Leopold 1959, Alvarez del Toro 1991, March 1993, Naranjo 2002). The largest group I saw was 35 animals; therefore, WLP groups in Calakmul region appear to be smaller than in other parts of their range. Calakmul groups ar e smaller than those from Peru, Brazil and Bolivia where group sizes larger than 100 are common (Kiltie and Terborgh 1983, Fragoso 1994, Peres 1996, Sowls 1997, L. Maffei pers. comm., R. Bodmer pers. comm., S. Carrillo pers. comm), but are with in the range reported for groups in a seasonally dry forest of Venezuela (14-60; Hernandez et al. 1995), for the dry forest of the Chaco in northern Argentine (7-50; Altrichter 2005) and for the Lacandon forest in Chiapas, Mexico (5-60; Naranjo 2002). There is evidence that the differences in group size across the geographic range for areas where humans do not seems to have an impact could be a consequence of differences associated with rainfall pattern, where a higher rainfall could incr ease habitat quality in terms of fruit abundance (Chapter 3).
108 Among groups of WLP in the Calakmul regi on, groups were bigger on the CBR than in the hunted sites. Several authors have stated that sightings of smalle r groups can be a consequence of hunting pressure as well as differences in eco logical factors like habitat quality (Leopold 1959, Kiltie and Terborgh 1983, Emmons and Feer 1990, March 1993, Naranjo 2002). Assuming the habitat types and quality, as well as other ecol ogical factors such as humidity and weather are similar for the Calakmul region as a whole, th e fact that WLP groups in communal forest are smaller than in the CBR suggests that this spec ies is being affected by human factors in the communal forests. It has been demonstrated that hunting is one of the main activities that affect population abundance in this sp ecies (Altrichter and Boaglio 2004, Naranjo and Bodmer 2007, Reyna-Hurtado and Tanner 2007). The cohesive behavior of the groups when facing danger (Peres 1996, Sowls 1997, Local hunters Pers. Co m.) makes the group vulnerable to human hunting because a hunter can kill several indivi duals at one time (Peres 1996; Reyna-Hurtado 2002). Furthermore, this species is highly prized for hunters, to the point that when a group is spotted hunters organize short trips (two or th ree days long) to hunt them (Reyna-Hurtado 2002). Besides reducing group size by physically removing i ndividuals, it is no t known what specific effects hunting has on the social st ructure. It is likely that th e chasing associated with hunting can cause groups to split (Local hunters, pers. co mm.). The results of this study suggest hunting is affecting WLP sub-populat ions living in communal fore st by reducing group size and depleting complete groups in some cases, ther efore reducing the density of this species. Additionally, at least in two ejidos (Nuevo Becal and 20 de Noviembre) groups are becoming isolated from the population living in the CBR. In these areas a road that runs North-South and the growing croplands that have been appearing along this road are separating the reserve from the communal forests. It is not known what effects this growi ng isolation may have on these
109 population, but forest fragmentation could be an other important cause for group size reduction given the requirements of large area s for groups of this species. Th is is especially true for the Calakmul groups where it has been demons trated they range ove r large areas (>100 km2), a home range size that is among the larg est reported in the li terature for this species (Chapter 2). Age Structure and Breeding Season The proportion of non-adults (juveniles and suba dults) in WLP groups was similar to that reported for groups in the Lacandon forest (Nar anjo and Bodmer 2007). I did not detect differences in non-adult proporti ons between groups in hunted and non-hunted areas. However, my field observations of hunting patterns indicate that hunting could have a potential effect on age structure of groups. For example, I observed three cases where hunters killed females that were pregnant or had newborns. In two cases th e hunters took the newborns and tried to raise them with no success (pers. obs.). Data from this study indicate that the breeding is seasonal, peaking at the beginning of the dry season (January-February) and extending into the dry s eason. Unfortunately, in Calakmul the dry season is when WLP are more vulnerable to hunting losses because they become attached to the only remaining water sources (Chapter 3). In this time is very easy to find and hunt peccaries once a water-body has been localized and fresh signals of this species have been detected there (Local hunter, pers. comm.). Hunting Seasons Hunting of WLP in Calakmul appears to be a dry season activity when the groups are easy to encounter at the few remaining water-bodies. This situation can be of particularly danger for the species in a place like Calakmul where th e rainfall patterns and wa ter availability differ year to year. For example it has been observed that some ponds can storage water some years and not the following year (Chapter 4), making the s earch for water a priority for the species at a
110 degree that in few occasions they have been obser ved close to large villages that have permanent ponds nearby (Zoh Laguna residents pers. comm.). During this time subsistence hunters can kill as many individuals as they want given that groups are easily detected. Additionally, sport hunters are allowed to hunt this species during the last two months of the dry season (April-May) a time of potential maximum stress on WLP groups I have documented sport and subsistence hunters to almost completely exterminate a grou p in a single dry seas on (Weber et al. 2006). Density Estimated WLP population density in the Calakmul Biosphere Reserve is among the lowest estimates across the dist ribution range on this species. The only similar estimate came from the Argentine Chacoan dry forest (0.33 ind/km2 for hunted sites and 1.04 ind/km2 for nonhunted sites; Altrichter 2005). This similarity is relevant given that th e Chaco dry forest and Calakmul represent the drier areas where this sp ecies is found. Estimates from more humid areas are larger. For example, in the Lacandon fore st population density varies from 1.08 ind/km2 in persistently hunted sites to 7.93 ind/km2 in slightly hunted sites (N aranjo and Bodmer 2007). In two fragmented sites of the Atlantic forest in Br azil, density of this species was determined at 6.3 and 6.9 ind/km2 (Cullen 1997, and Keuroghlian et al. 2004 re spectively). Interestingly enough, in the Peten forest of northern Guatemala (which forms a continuous forest with Calakmul) a density of 9.59 ind/km2 was determined for a non-hunted si te using line transect methodology (Novack et al. 2005). These differences could be associated with the method used; in line transects the same group could be counted se veral times because of their large home range. Another explanation could be the increase in rainfall that makes the Peten forest taller and more humid than the drier forest of Calakmul. Fi nally, the WLP population in Calakmul may have cyclical changes in density that span decad es. Fragoso (1994; 2004), studying a WLP population inhabiting the Maraca Island in Ro raima State, Brazil, reported density to vary from 139-542
111 ind/km2 in 1988 to 1.4-8.3 ind/km2 in 1992 after an apparent population crash following a diseases transmission (Fragoso 2004). However, I lack long-term data for Calakmul region needed to determine if long-term population cy cles exist in that region. A more plausible explanation could be that this species is living in Calakmul in one of the extremes of its distribution range, and the habi tat conditions there may be sub-optimal for the species. Recently, WLP population densities in the Cala kmul region were estimated using line transects for purposes of developing sport harves t guidelines under the UMAs scheme (Direccion General de Vida Silvestre, Campeche state, te chnical report). However, several flaws were detected in the estimation method (See Weber et al. 2006 for a more detailed description) that includes non-systematic sampling. These survey s maybe overestimated WLP populations. For example for a 500.00 km2 ejido that lies just north of the CBR, the WLP population was estimated to be from 1,350 to 4,000 individuals (D ireccion General de Vida Silvestre, Campeche state, technical report). These numbers highly contrast with my estimate of almost 1,500 individuals for the whole southern area of Cala kmul which is seven times larger than this ejido This overestimation can lead authorities to a llow a high harvest rate, resulting in potential negative effects of sport hunting, especially when combined with subsiste nce hunting harvest, an activity not regulated at all in the Calakmul region. I believed the density estimation of my study is more accurate because it was obtained from th e actual range of the groups and the area shared and used and should be used as a conserva tive reference point for other estimates. Conservation Implications WLP groups in the Calakmul region inhabit a fore st that lies in the northern extreme of it distribution and where conditions are sub-optimal for its surviv al. These conditions make WLP vulnerable to hunting and other human induced effects because the low density and smaller group sizes occurring there. Additi onally, the breeding season coinci de with the time of the year
112 when they become more vulnerable to hunting. It is a priority that the government institution in charge of sport hunting (Direcci on General de Vida Silvestre), as well as other NGOs involved, consider banning the sport hunting on this spec ies until more sound biological information is obtained, or allowing hunting se asons only during the rainy seas on and limiting the number of individuals harvested from any one group. Also, development of WLP conservation plans must actively involve subsistence hunters from ejidos, or it is highly probably that the WLP will disappear from communal forests in a near future. The fu ture disappearance of WLP on communal forests could have sign ificant consequences for the ecological processes of these forests and for the remnant WLP population in the CBR. Conservation of WLP only within a few, large, well protected reserves, like the CBR, will not represent the best scenario for conservation of this species in Mexico.
113 Table 5-1. Group size, group age composition and area classification for white-lipped peccary groups observed in Calakmul Region, Campeche, Mexico. Area Classification Group Size Group Composition Date Place Source Non-hunted area 30 25 adults 5 juveniles April 3, 2001 Pond near km 46 in road to Calakmul Archeological Site (CBR) Biologist, students, custodian Non-hunted area 35 30 adults 2 sub-adults 3 juveniles April 25, 2001 Pond 2 km South of Calakmul Archeological Site (CBR) R. ReynaHurtado Non-hunted area 26 26 adults April 27, 2001 Pond near km 46 in road to Calakmul Archeological Site (CBR) R. ReynaHurtado Non-hunted area 20 18 adults 2 juveniles May 3, 2000 Pond in Heliport (CBR) R. ReynaHurtado and others Non-hunted area 31 28 adults 2 sub-adults 1 juveniles March 7, 2005 Pond in Heliport (CBR) R. ReynaHurtado and others Non-hunted area 25 20 adults 3 sub-adults 2 juveniles March 7, 2005 Pond in Heliport (CBR) R. ReynaHurtado and others Non-hunted area 25 20 adults 2 sub-adults 3 juveniles March 7, 2005 Pond in Heliport (CBR) R. ReynaHurtado and others Non-hunted area 20 18 adults 2 juveniles March 7, 2005 Pond in Heliport (CBR) R. ReynaHurtado Non-hunted area 19 19 adults December 18, 2005 Pond in Heliport (CBR) R. ReynaHurtado Close to hunted area 14 8 adults 1 sub-adult 5 juvenile February 28, 2001 Pond 2 km East of checkpoint at km 20 in road to Calakmul Archeological Site (Calakmul limit) R. ReynaHurtado and others Close to hunted area 11 10 adults 1 juvenile April 14, 2001 Pond 2 km East of checkpoint at km 20 in road to Calakmul Archeological Site (Calakmul limit) Field assistant Close to hunted area 16 16 adults April 2005 Pond called Ramonal (Calakmul limit) Biologist Close to hunted area 15 12 adults 2 sub-adults 1 juveniles April 2006 Pond called Oxpemul (Calakmul limit) R. ReynaHurtado Close to hunted area 15 14 adults 1 juvenile February 2007 Pond called Oxpemul (Calakmul limit) Biologist Hunted area 25 25 adults March 2001 Pond El Verdin in the communal forest of 20 de Noviembre Ejido Local hunter Hunted area 12 12 adults April 15 2001 Pond Los Muecos in the communal forest of Xbonil Ejido Local hunter Hunted area 15 15 adults May 2001 Pond El Giro in the communal forest of Xbonil Ejido Field assistant
114 Table 5-1.(continued) Hunted area 30 27 adults, 3 juveniles October 2004 Pond Nosaya Nuevo Becal Ejido Local and sport hunters Hunted area 35 Undete rmined November 2005 Area known as Area Semillera Nuevo Becal Ejido Local hunter Hunted area 20 Undetermined December 26, 2005 Area of El Naranjalito Nuevo Becal Ejido Local hunter Hunted area 27 27 adults April 2006 Pond located at 400 meters to south of Km 92 of the road Escarcega Chetumal Xbonil Ejido Local hunter Hunted area 13 11 adults, 2 subadults April 6, 2005 Pond Chumakil in the communal forest of Nuevo Becal Ejido Field assistant Hunted area 17 Undetermined April 25, 2006 Aguada 2 Lagartos Nuevo Becal Ejido Field assistant Hunted area 20 20 adults Apr il 15, 2005 Pond Dos Guineas in the communal forest of Nuevo Becal Ejido Field assistant Table 5-2. Proportions (%) of forest types for the ar ea used and shared by th e four groups and the southern area of the Calakmul Bi osphere Reserve, Campeche, Mexico. Forest Types Area Used-Shared for the Four Groups (MCP) Southern Area of Calakmul Biosphere Reserve Medium semi-perennial forest 77.76 63.00 Low flooded forest 14.14 19.31 Low dry forest 00.07 00.04 Ponds and other waterbodies 08.00 13.8 Others 00.00 03.80 Total 99.9 % 99.9 % Area total (km2) 236.66 km2 3 509.41 km2
115 Figure 5-1. Calakmul Biosphere Reserve and s ites where white-lipped peccary groups were sighted in Calakmul region, Campeche, Mexico. 0 5 10 15 20 25 30 Non-Hunted AreasHunted AreasGroup size (Median ) Figure 5-2. Median of white-lip ped group size recorded in the Calakmul Biosphere Reserve (Non-Hunted Area) and four sites (Hunt ed Areas) where hunting and other human activities are taking places in Ca lakmul region, Campeche, Mexico. Calakmul Limit Xbonil ejido Nuevo Becal ejido 20 de Noviembre ejido Calakmul Biosphere Reserve-CBR
116 0 10 20 30 40 50 60 70 80 90 100 AdultsSub-AdultsJuveniles Age Category% Figure 5-3. Age structure observed in groups of white-lipped peccary in the Calakmul Region, Campeche, Mexico. 0 1 2 3 4 5 6 7 JulAugSepOctNovDecJanFebMarAprMayJunBirths Figure 5-4. Seasonality of the birth events obser ved in groups of white-lipped peccary in the Calakmul region, Campeche, Mexico.
117 Figure 5-5. Area used and shared for four groups where density was estimated (encircled by light green line) and area where density was extra polated (encircled by the blue line) of white-lipped peccary in the southern area of the Calakmul Biosphere Reserve, Campeche, Mexico
118 CHAPTER 6 CONCLUSIONS Social Ecology of White-Lipped Peccary ( Tayassu pecari ) The social behavior and ecological aspects of white-lipped peccarie s were studied for 18 months on the Calakmul Biosphere Reserve in S outhern Mexico. This was the first study with wild animals in Mexico. Among the main resu lts I found that WLP in Calakmul, with group sizes that never exceeded 35 individuals, had one of the largest reported home ranges, only being surpassed by a very large group (>200 individual s) in Maraca Island, Roraima, Brazil (Fragoso 2004). WLP home ranges estimated elsewhere were fairly smaller than my estimates. WLP substantially increased their home ranges when the rain season arrived to the CBR suggesting that water availability was more important fo r determining movements than food availability. The larger home ranges and the smaller group s sizes reported in this study can be interpreted as an adaptation stra tegy of the species to the partic ular conditions of the Calakmul forests. These forests are drier than other fo rest where home ranges have been estimated like Corcovado National Park (Costa Rica; Carrillo et al. 2002), and the Atla ntic forests and the Amazon forest (Brazil; Fragoso 1998, Keuroghlian et al. 2004). According to several authors the availability of high energy food, lik e fruits, is higher in primary humid forest than in secondary or dry forests (Bodmer 1989, Sowls 1997, Altricht er et al. 2001, Chapman et al. 2005, Beck 2006). Additionally, the increase in ho me range demonstrates that th is species has the ability to perform movements at the landscape scale as was predicted by Fragoso (1999). This ability appears to be matched with some kind of spatial memory or orientation skills that allow them to search and find specific la ndmarks such as ponds or sartenejas in this large and relatively featurelesslandscape of the CBR (pers. obs.).
119 WLP groups in Calakmul as well as in seve ral parts of their ra nge (Sowls 1997, Fragoso 1998) depend on water bodies (ponds) to wallow refresh and forage. I observed that WLP visited the ponds daily unless there was water in the Flooded forest. In the CBR, aside from the ponds WLP highly preferred patche s of Medium forest throughout the year. During the rain season, however, the Flooded forest was preferred. The Dry forest that is very abundant in the area was used mainly to travel betwee n patches of preferred forest or ponds. Regarding ranging area and the Ecological Constr aint Model, in this study, smaller groups ranged over larger areas than la rger groups in Calakmul Biosphere Reserve. Additionally, water was strongly associated with move ments on the seasonally dry forest of Calakmul. Therefore, it is highly probable that the landscape-scale movements performed by white-lipped peccary groups in Calakmul are linked to the spatial a nd temporal distribution of water sources. Food availability may play a secondary role causing th e Ecological Constraint Model to apply only at a short temporal and spatial s cale. Groups living in Calakmul Biosphere Reserve represent an example of the extreme variation in the soci al-ecological behavior on this species and a successful strategy for cohesive groups that depend on patchy distri buted resources. The Ecological Constraint Model provid ed a well suited theoretical framework for the analysis of the field and published data on WLP. More refineme nt needs to be done on scale issues and data collection to better understand th e relationship between group size, travel rate and food availability on this species. It was found that important resources such as water reservoirs for WLP are dispersed and non-uniformly spatially-distributed over the la ndscape at Calakmul Bios phere Reserve and can be temporally scarce. I found evidence that WL P groups move over the landscape in a pattern resembling the Lvy-walk model. Lvy-walk mo vements are scale invariant (Shlesinger and
120 Klafter 1986, Viswanathan et al. 1996, Bartumeu s et al. 2005, Benhamou 2007). This allows the group to search for resources in detail at the small scale (with lots of small steps), while occasionally making extended searches into th e landscape scale (with few long movements). Additionally, groups behave like central-place foragers around the Calakmul pond during the dry season 2005. The combination of central place foraging and Lvywalk movements seems to be an efficient strategy for WLP living in Calakmul Biosphere Reserve. Because they have access to resources that are secure (such as water) whil e exploring from time to time areas far away to monitor when the conditions change that allow them to move away. I found that groups travel coordinately, in a couple of times the smallest group (Green), followed one larger group (Blue and Red) and pe rform similar behavior and move to the same areas without mixing individuals, then after a period of 33 and 20 days respectively they separated. If this species is using some kind of spatial knowledge stor age within the groups memory, then living in large groups confers advantages in this se nse. Probably, the association of two groups while traveling, as I documented here, increased the ch ances of finding these highly localized resources. It could be that th e Green group, which was always the smaller-sized group, associated with larger groups to learn where and when the resources are available. Apparently a combination of searching strate gies like Lvy-walk movements, central place foraging and travel association between groups al lows WLP groups to forage efficiently over a landscape where resources are spatially dispersed, temporarily available and where unpredictability also plays a role. The combination of all these strategies allows WLP to survive in a sub-optimal forest like Calakmul Biosphere Reserve. When contrasting size of groups between the CBR and communal forest where hunting pressure exists, I found that groups were bigge r on the CBR than in the hunted sites, which
121 suggests that this species is be ing affected by human f actors in the communal forests. Data from this study indicate that the breed ing is seasonal, peaking at th e beginning of the dry season (January-February) and extending into the dry season. Unfortunatel y, in Calakmul the dry season is when WLP are more vulnerable to hunting loss es because they become attached to the only remaining water sources. In this time is very easy to find and hunt peccaries once a water-body has been localized and fresh signals of this specie s have been detected there (Local hunter, pers. comm.). Hunting of WLP in Calakmul appears to be a dry season activity when the groups are easy to encounter at the few remaining water-bodies. This situation can be of particularly danger for the species in a place like Calakmul where th e rainfall patterns and wa ter availability differ year to year. During this time subsistence hunters can kill as many individuals as they want given that groups are easily detected. Additionally, sport hunters are allo wed to hunt this species during the last two months of the dr y season (April-May) a time of potential maximum stress on WLP groups. Conservation of White-Lipped Peccary in Calakmul Region To conserve WLP in northern Central America it will be necessary to preserve intact this piece of the Maya forest and th e attributes that allow WLP to survive in it. Conservation measures will likely include effective protecti on against hunting, no further road development to maintain the area isolated, mainta ining the availability of ponds and/or other sources of water, and the maintenance of a landscape composed of interspersed Medium and Flooded forests. Active conservation measures would be to assu re WLP groups have access to water sources (could even be worth to strategi cally locate artificial water contai ners for the extreme dry years as the dry season 2006). Also, cons erving large intact patches of forest would maintain WLP populations on human dominated landscapes if hunting pressure can be controlled.
122 WLP groups in the Calakmul region inhabit a fore st that lies in the northern extreme of it distribution and where conditions are sub-optimal for its surviv al. These conditions make WLP vulnerable to hunting and other human induced effects because the low density and smaller group sizes occurring there. Additi onally, the breeding season coinci de with the time of the year when they become more vulnerable to hunting. It is a priority that the government institution in charge of sport hunting (Direcc ion General de Vida Silvestr e), as well as other NGOs and academic institutions involved, consider banning the sport hunting on this species until more sound biological information is obtained, or allowing hunting s easons only during the rainy season and limiting the number of individuals ha rvested from any one group. Also, development of WLP conservation plans must activ ely involve subsistence hunters from ejidos, or it is highly probably that the WLP will disappear from comm unal forests in a near future. The future disappearance of WLP on communal forests coul d have significant consequences for the ecological processes of these forests and for the remnant WLP population in the CBR. Conservation of WLP only with in a few, large, well protected reserves, like the CBR, will not represent the best scenario for the co nservation of this species in Mxico.
123 LIST OF REFERENCES Aebischer, N., P. A. Robertson, and R. E. Ke nward. 1993. Compositional analysis of habitat use from animals radio-tracking data. Ecology 74:1313-1325 Altrichter, M., E. Carrillo, J. Saenz and T. K. Fuller. 2001. White-lipped peccary ( Tayassu pecari Artiodactyla :Tayassuidae) diet and fruit av ailability in a Costa Rican rain forest. Revista de Biologia Tropical, 49 :1183-1192 Altrichter, M. and R. Almeida. 2002. Exploita tion of white-lipped peccaries Tayassu pecari (Artiodactyla :Tayassuid ae) on the Osa Peninsula, Costa Rica. Oryx 36:126-132 Altrichter, M., C. Drews, J. C. Senz y E. Carrillo. 2002. Presupuesto de tiempo del Chancho Cariblanco ( Tayassu pecari ) en un bosque hmedo de Costa Rica. Biotropica 34: 136143. Altrichter, M., and G. I. Boaglio. 2004. Distri bution and relative abundance of peccaries in the Argentine Chaco : associations with huma n factors. Biological Conservations 116 :217225 Altrichter, M. 2005. The sustainability of subs istence hunting of peccaries in the Argentine Chaco. Biological Conservation, 126 :351-362 Alvarez del Toro, M. 1991. Los mamferos de Ch iapas. 2da ed. Gobierno del Estado, Tuxtla Gutierrez, Chiapas, Mxico. 133 pp. Aranda, M. 1981. Rastros de los mamferos silv estres de Mxico. INIREB, Ediciones Maccio. Xalapa Veracruz. Mxico. 198 pp. Atkinson, R. P. D, C. J. Rhodes, D. W. Mac donald, R. M. Anderson. 2002. Scale free dynamics in the movement patterns of jackals. Oikos 98:134-140 Austin. D., W. D. Bowen, J. I. McMillan. 2004. In traspecific variation in movement patterns: modeling individual behavior in a la rge marine predator. Oikos 105:15-30 Balcomb, S. R., C.A. Chapma n, and R.W. Wrangham. 2000. Rela tionship between chimpanzee ( Pan troglodytes ) density and large, fleshy-fruit tree density: Conservati on implications. American Journal of Primatology 51:197-203. Bartumeus, F., M. G. E. Da Luz, G. M. Visw anathan, J. Catalan. 2005. Animal search strategies: a quantitative random-walk analysis. Ecology, 86:3078-3087 Beck, H. 2006. A review of peccary-palm interac tions and their ecological ramifications across the Neotropics. Journal of Mammalogy 87:519-530. Benhamou, S. 2007. How many animals rea lly do the Levy walk? Ecology 88:1962-1969
124 Bodmer, R. 1990. Responses of ungulates to se asonal inundations in the Amazon floodplain. Journal of Tropical Ecology 6:191-201. Bodmer, R. 1991. Influence of digestive mor phology on resources partitioning in Amazonian ungulates. Oecologia 85:361-365. Bodmer, R., and L. K. Sowls. 1993. The collared peccary ( Tayassu tajacu ) Pp 7-13. In: Oliver, W. L. R. (ed).1993. Pigs, peccaries and hippos. IUCN. Gland, Switzerland. 202 pp Bodmer, R., J. F. Eisenberg and K.H. Redfor d. 1997. Hunting and the likelihood of extinction of Amazonian mammals. Cons ervation Biology 2:460-466 Boinski, S., E. Ehmke, L. Kauffman, S. Sc het and A. Vreddzaam. 2005. Dispersal patterns among three species of squirrel monkeys ( Saimiri oersteddi, S. boliviensis and S. sciureus ): II. Within-species and loca l variation. Behavior 142:633-677 Boinski, S., L Kauffaman, A. Westoll, C.M. Stickler, S. Cropp and E. Ehmke. 2003. Are vigilance, risk from avian predators and gr oup size consequence of habitat structure? A comparison of three species of squirrel monkey ( Saimiri oersteddi, S. boliviensis and S. sciureus ). Behavior 140:1421-1467 Boinski, S., and P. A. Garber (eds). 2000. On the move: how and why animals travel in groups 2000. The University of Chicago Pr ess, Chicago, Ill. USA, 811 pp. Burnham, K. P., and D. R. Anderson. 1998. M odel Selection and Inference: A Practical Information-Theoretic Approach. Springer-Verlag, New York, NY, USA Byers, C. R., K. Steinhorst and P. R. Krausman. 1984. Clarification of a te chnique for analysis of utilization-availability data. Journa l of Wildlife Manage ment 48(3): 1050-1053. Caldecott, J. O., R. A. Blouch, and A. Macdonald. 1993. T he bearded pig (Sus barbatus). Pp. 136-144. In: Oliver, W. L. R. (Ed). 1993. Pigs, peccaries and hippos. IUCN. Gland, Switzerland. 202 pp Cardille J.A., and M. G. Turner. 2002. Understa nding landscape metrics I. pp 85-101. In: Gerger SE, Turner MG (eds). 2002. Learning landscap e ecology: A practical guide to concepts and techniques. Springer USA Carrillo, E., J. C. Saenz and T. K. Fuller. 2002. Movements and activ ities of white-lipped peccaries in Corcovado National Park, Cost a Rica. Biological Conservation 108: 317324. Chapman, C.A., L. J. Chapman, R. L. Mc Laughlin. 1989. Multiple central place foraging by spider monkeys: travel consequences of us ing many sleeping sites. Oecologia 79:506-511 Chapman, C.A. 1990. Ecological constraint on gr oup size in three species of Neotropical primates. Folia primatologica 55:1-9
125 Chapman, C.A., Chapman, L.J., Wrangham, R., Hunt, K., Gebo, D., Gardner, L., 1992. Estimators of fruit abundance of tropicaltrees. Biotropica 24, 527. Chapman, C.A., R. W. Wrangham, and L.J. Ch apman. 1995. Ecological cons traint on group size: an analysis of spider monkey and ch impanzee subgroups. Behavioral ecology and sociobiology 36:59-70 Chapman, C. and L. J. Chapman. 2000a. Determinan ts of group size in pr imates: the importance of travel costs. Pp. 24-42. In: Boinski, S ., and P. A. Garber (eds). 2000. On the move: how and why animals travel in group. The Un iversity of Chicago Press, Chicago, Ill. USA, 811 pp Chapman, C.A., and L.J. Chapman. 2000b. Constrai nts on group size in red colobus and redtailed guenons: examining the generality of th e ecological constraint model. International Journal fo Primatology, 21:565-585 Chapman, C. and M.S.M. Pavelka. 2005. Gr oup size in folivorous primates: ecological constraints and the possible influence of social factors. Primates 46:1-9 Chapman, C. L. A. Chapman, T. T. Struhsaker, A. Zanne, C. J. Clark and J. R. Poulsen. 2005. A long term evaluation of frui ting phenology: importance of climate change. Journal of Tropical Ecology 21:1-14 Comisin Nacional del Agua (CNA), Serv icio Meteorolgico Nacional. Mxico. http://smn.cna.gob.mx Cushman S.A., M. Chase, C. Griffin. 2005. El ephants in space and time. Oikos 109 :331-341 De Solla, S. R., R. Boundariansky, and R. J. Brooks. 1999. Eliminating autocorrelation reduces biological relevance of home range estim ates. Journal of Animal Ecology, 68:221-234. Donkin, R. A. 1985. The Peccary with observations on the introduction of pigs to the New World. The American Philosophical Society. Philadelphia, USA. 152 pp. Emmons, L. H., and F. Feer. 1990. Neotropical rainforest mammals. A field guide. The University of Chicago Press. Chicago, IL. USA. 281 pp Ericson J. 1997. Regional assessment Calakmul population-environment initiative. Preliminary report. Pronatura Peninsula de Yucatan, M xico. WWF and Univer sity of Michigan, USA. 25 pp Escamilla, A., M. Sanvicente, M. Sosa and C. Galindo-Leal. 2000. Habitat mosaic, wildlife availability, and hunting in the tropical fo rest of Calakmul, Mxico. Conservation Biology 14:1592-1601.
126 Fragoso J. M. V. 1994. Large mammals and the community dynamics of an Amazonian rain forest. Ph.D. Dissertation. The University of Florida, Gainesville, Fl. USA. 210 pp Fragoso, J. M. 1997. Desapariciones locales del baquiro labiado ( Tayassu pecari ) en la Amazona: migracin, sobre-cosecha o epidem ia?. En: Fang, T. G., Bodmer, R., Aquino, R., y Valqui, M. (Eds). 1997. Manejo de fauna silvestre en la Amazona. UNAP, University of Florida UNDP/GEF e Institut o de Ecologia. La Paz, Bolivia. Pp 309-312. Fragoso J. M. V. 1998. Home range and move ment patterns of white-lipped peccary ( Tayassu pecari ) herds in the Northern Brazilian Amazon. Biotropica 30(3): 458-469. Fragoso, J. M. V. 1999. Perception of scale and resource partitioning by peccaries: Behavioral causes and ecological implications. Journal of Mammalogy, 80(3): 993-1003 Fragoso, J. M. V. 2004. A Long-Term Study of White-Lipped Peccary ( Tayassu pecari ) population fluctuation in Northern Amazonia, Pp. 286-296. In: Silvius, K, R. E. Bodmer, and J.M.V. Fragoso. 2004. People in Nature, W ildlife Conservation in South and Central America. Columbia University Press, New York, USA, 463 pp. Fuller, T. K., E. Carrillo, and J. Senz. 2002. Su rvival of protected wh ite-lipped peccaries in Costa Rica. Canadian Jo urnal of Zoology 80:586-589. Galindo-Leal, C. 1999. La gran regin de Calakm ul: prioridades biolgicas de conservacin y propuesta de modificacin de la Reserva de la Biosfera. Reporte fi nal al World Widlife Fund. Mexico D.F. Mexico 40 pp. Ganas, J., and M. M. Robbins. 2005. Rangi ng behavior of the mountain gorilla ( Gorilla beringei beringei ) in Bwindi Impenetrable National Pa rk, Uganda: a test of the ecological constraint model. Behavioral Ecology and SocioBiology, 58:277-288 Garca, E., 1988. Adaptacin del sistema climat olgico de Koppen a la Repblica Mexicana. Mxico, D.F. 76 pp. Gillespie, T. R. and C.A. Chapman. 2001. Determin ants of group size in theRed colobus monkey ( Procolobus badius ): An evaluation of the generality of the ecological constraints model. Behavioral Ecology and Sociobiology 50:329-338. Giraldeu, L. A., and D. L. Kramer. 1982. The marg inal value theorem: A quantitative test using load size variation in a central pl ace foragers, the eastern chipmunk, Tamias striatus Anim. Behav 30:1036-1042 Hernandez, O. E., G. R. Barreto and J. Ojasti. 1995. Observations of be havioural patterns of white-lipped peccaries in the wild. Mammalia 59 (1):146-148 Hooge, P. N. and B. Eichenlaub. 1997. Animal movement extension to Arc View, ver. 1.1. Alaska Biological Science Center, U. S. Geological Survey, Anchorage, AK, USA.
127 INEGI (Instituto Nacional de Estadstica Geografia e Informtica). 2005. Conteo socioeconmico del 2005. Estado de Campeche. INEGI. Campeche, Mxico. Jarman, P.J. 1974. The social organization of an telope in relation to their ecology. Behavior, 48:215-267 Johnson, D. H. 1980. The comparison of usage a nd availability measurements for evaluating resources preference. Ecology 61: 65-71 Kernohan, B. J., R. A. Gitzen, and J. J. Mi llspaugh. 2001. Analysis of Animal Space Use and Movement. Pp. 125-166, In: Millspaugh, J. J. and J. M. Marzluff (eds). 2001. Radio Tracking and Animal Populations. Academic Press, USA, 474 pp. Keuroghlian, A., D. P. Eaton, and W. S. Longland. 2004. Area use by white-lipped and collared peccaries ( Tayassu pecari and Tayassu tajacu ) in a tropical forest fragment. Biological Conservation 120:411-425 Kie, J. G., J. A. Baldwin, and C. J. Evans. 1996. CALHOME: a program for estimating animal home ranges. Wildlife So ciety Bulletin, 24:342-344 Kiltie, R.A. 1981. The significance of interlocking can ines in the forest peccaries (Tayassuidae). Journal of Mammalogy 62:459-469. Kiltie, R. A., and J. Terborgh. 1983. Observations on the behavior of rain forest peccaries in Per: Why do white-lipped peccaries form herds? Z. Tirpssychol. 62:241-255. Krause, J. and G. D. Ruxton. 2002. Living in Groups. Osford Universiity Press, 210 pp. Leopold, A. S. 1959. Fauna Silvestre de Mxico. Instituto Mexicano de Recursos Naturales Renovables. Mxico D.F. 608 pp. Mandujano, S. 1999. Variation in herd size of coll ared peccaries in a Mexican tropical forest. The Southwestern Naturalist 44:199-204. March I, 1990. Evaluacin de h bitat y situacin actual del pecar de labios blancos Tayassu pecari en Mxico. Tesis de Maes tra de la Universidad Nacional de Costa Rica March, I. J. 1993. The white-lipped peccary ( Tayassu pecary ) In: Oliver, W.L.R. (Ed). 1993. Pigs, peccaries and hippos. IUCN. Gland, Switzerland. 13-22 pp Mrrel, A., J. P. Ball, A. Hofgaard. 2002. Foraging and movement paths of female reindeer: insights from fractal analys is, correlated random walks and Levy flights. Canadian Journal of Zoology. 80:854-865 Mayer, J. J. and R. M. Wetzel. 1987. Tayassu pecari. Mammalian Species. 293, 1-7 pp.
128 McGarical, K., B. J. Marks. 1994. FRAGSTATS : Spatial pattern analysis program for quantifying landscape structure. Forest Scie nce Department. Oregon State University Millspaugh, J. J. and J. M. Marzluff (eds). 2001. Radio Tracking and Animal Populations. Academic Press, USA, 474 pp. Naranjo, E. J. 2002. Population ecology and conser vation of ungulates in the Lacandon forest, Mxico. Ph.D. Dissertation. The University of Florida, Gainesville, Fl, USA. 146 Neu, C. W., C. R. Byers, and J. M. Pee k. 1974. A technique for analysis of utilizationavailability data. Journal of Wildlife Management 38: 541-545. Orians, G. H., N. E. Pearson.1979. On the theory of central place foraging. In: Horn, D.J., Rd. Mitchell Rd, G. R. Stairs. (eds) Analysis of ecological systems. Ohio University Press, Ohio, pp 155-177 Pennington, T. D., and J. Sarukhn. 1998. rbol es Tropicales de Mxico. UNAM. Fondo de Cultura Econmica. Mxico D.F. Mxico Peres, C. A. 1996. Populati on status of white-lipped Tayassu pecari and collared peccaries T. tajacu in hunted and unhunted Amazonian fore st. Biological Conservation 77:115-123. Quijano, E., H. 2001. Ecologa, aprovechamiento y c onservacin de la fauna silvestre en Tres Reyes, Quintana Roo. Tesis de Maestra. EC OSUR (El Colegio de la Frontera Sur), Chetumal, Q. Roo, Mxico. 66 pp. Ramos-Fernandez G, J. L. Mateos, O. Miramontes G. Cocho, H. Larralde, and B. Ayala-Orozco. 2004. Levy walk patterns in the foraging movements of spider monkeys ( Ateles geoffroyi ). Behav Ecol. Sociobiol. 55:223-230 Reyna-Hurtado, R., M. Sanvicente, S. Calm, G. Escalona, and J. A. Vargas. 1999. Estudio de fauna silvestre en la comunidad de Xbonil, Calakmul, Campeche. SEMARNAP. ECOSUR. Campeche, Camp. Mxico. 88 pp. Reyna-Hurtado, R. 2002. Hunting effects on the ungulat es species in Calakmul Forest, Mexico. Master Thesis, University of Florida. Gainesville, Fl. 91 pp. Reyna-Hurtado, R., and G. Tanner. 2005. Habitat Preferences of an Ungulate Community in Calakmul Forest, Campeche, Mexico. Biotropica. 37:676-685 Reyna-Hurtado, R. M. Altrichter and E. A. M. Junior. 2006. Methods for the study of peccaries in the wild. IUCN-PPHSG Newslette r. Suiform Soundings. Vol. 6:26-34 Reyna-Hurtado, R., and G. W. Tanner. 2007. U ngulate relative abundance in hunted and nonhunted sites in Calakmul Forest (Souther n Mexico). Biodiversity and Conservation 16:743-757
129 Rivera, A. and S. Calm 2006. Forest fragmenta tion and the changes in the feeding ecology of the Black Howler Monkey ( Alouatta pigra ) in Calakmul region. In: Estrada A., P. Garber, M. Pavelka, and L. Luecke. (eds.) New perspectives on the distribution, ecology and conservation of Mesoamerican Pr imates. Springer, New York. Pp. 189-213. Rooney, S. M., A. Wolfe, and T. J. Hayden. 1998. Au tocorrelated data in te lemetry studies : time to independence and the problem of be havioural effects. Mammal Rev. 28:89-98 Schoener T. W. 1979. Generality of the size-distance rela tionships in models of optimal feeding. Am Nat 114:902-914 Seaman, D. E. and R. A. Powell. 1996. An eval uation of the accuracy of kernel density estimators for home range analysis. Ecology 77 :2075-2085 Seaman, D. E., J. J. Millspaugh, B.J. Kernohan, G. C. Brundige, K. J. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63 :739-747 SEMARNAP. 2000. Programa de manejo de la Rese rva de la Biosfera de Calakmul. Mxico. Instituto Nacional de Ecologa. Mxico D.F. 268 pp. Shleinger, M. F., and J. Klafter. 1986. Levy walk versus Levy flight. In: Stanley, H. E., and N. Ostrowsky. (eds). On growth and form: fr actal and non-fractal pa tterns in physics. Martinus Nijhoff, Boston pp 279-283 Silvius, K, R. E. Bodmer, and J.M.V. Fragos o. 2004. People in Nature, Wildlife Conservation in South and Central America. Columbia Un iversity Press, New York, USA, 463 pp. Sims, D.W., D. Righton, J. W. Pitchford. 2007. Mi nimizing errors in identifying Levy flight behavior of organisms. Jour nal of Animal Ecology 76:222-229 Snaith, T.V., and C. A. Chapman. 2007. Primat e group size and interpre ting socioecological models: do folivores really play by differe nt rules?. Evolutionary Anthr. 16:94-106 Sokal, R. R., and F. J. Rohlf.1994. Biometry, 3rd Edition. Freman, New York Sowls, L. K. 1997. Javelinas and the other pe ccaries: their biology, management and use. 2nd. Ed. Texas A&M University Press. College Station, Tx, USA. 325 pp. SPSS. 1997. Statistical package for the social sc iences. Graduate pack version 8.0 for windows. SPSS Inc. Chicago, Illinois. USA. Stearman, A. M. 1992. Neotropical hunters and their neighbors: Sirion, Chimane and Yuqu hunting on the Bolivian frontier. Pp 108-128. In Redford, K., and C. Padoch (Eds). 1992. Conservation of Neotropical Fo rest (working from traditiona l use).Columbia University Press, New York, USA. 475 pp.
130 Stephens, D.W., and J. R. Krebs. 1986. Foraging theory. Princeton Univ.Press, Princeton 248 pp. Taber, A. 1993. The Chacoan peccary (Catagonus wagneri) In: Oliver, W. L. R. (Ed). 1993. Pigs, peccaries and hippos. IUCN. Gland, Switzerland. 22-27 Treves, A. and C.A. Chapman. 1996. Conspecific threat, predation avoidance, and resource defense: Implications for grouping in langur s. Behav.Ecology and Sociobiology 39:43-53 Viswanathan, G. M., V. Afanasyev, S. V. Buldyr ev, E. J. Murphy, P. A. Prince, and H. E. Stanley.1996. Levy flight search patterns of wandering albatrosses. Nature 381:413-415 Viswanathan, G. M., S. V. Buldyrev, S. Havlin, M. G. E. Da Luz, E. P. Raposo, and H. E. Stanley. 1999. Optimizing the success of random searches. Nature 401:911-914 Vickers, W.T. 1991. Hunting yields and game co mposition over ten years in an Amazon Indian territory. In Robinson, J. G., and K. Re dford. (Eds) Neotropi cal wildlife use and conservation. The University of Chicago Press. 520 pp Ucan, E., L. M. Ortega, J. J. Ortiz, J. Tun and S. Flores. 1999. Vegetacin y flora. Pp. 59-64. In: Folang, W. J., M.C. Snchez and J. M. Garc a. 1999. Naturaleza y cultura en Calakmul, Campeche. Universidad Autnoma de Campeche. Centro de investig aciones histricas y sociales. Campeche, Campeche, Mxico. 176 pp. Wallace, R. B. and R. L. Painter. 2002. Phenol ogical patterns in a southern Amazonian tropical forest: implications for sustainable mana gement. Forest Ecology and Management 160:19-33 WCS -Wildlife Conservation Soci ety. 2005. Taller sobre: Identif icacin de Prioridades de Conservacin para el Tapir de Tierras Baja s y el Pecar de Labio Blanco. Santa Cruz, Bolivia, April, 2005. Weber, M. 2000. Effects of hunting on tropical deer populations in Southe astern Mxico. M.Sc. Thesis. Royal Veterinary College. University of London. London, U K. 80.pp Weber, M., G. Garca-Marmolejo and R. Reyna-Hurtado. 2006. The tragedy of the commons Mexican style: A critique to the Mexican UMAs concept as applied to wildlife management and use in south-eastern Me xico. Wildlife Society Bulletin 34:1480-1488 White, G. C. and R. A. Garrot. 1990. Analysis of Wildlife Radio-Tracking Data. Academic Press, San Diego, CA, USA Wilson, O. E. 1975. Sociobiology: the modern synthe sis. Harvard University Press, Cambridge Wrangham, R.W., J. L. Gittleman, and C.A. Chapman. 1993. Constraint on group size in primates and carnivores: population density and day-range as assa ys of exploitation competition. Behavioral Ecology Sociobiology 32:199-209
131 BIOGRAPHICAL SKETCH Rafael Reyna was born in a small town in Mi choacan State in Central Mexico. He was the seventh of an eight member fa mily. Two things accounted for hi s love of nature: his frequent travels in company of his father and brothers to the forest su rrounding the town, and the animal book collection his father had in his private office. Rafael decided to study biology at the State University (Universidad Michoacana de San Nicols de Hidalgo). During and after school he spent six years of professional experience in the State Zoological Park (Parque Zoolgico Benito Jurez) and three years in a fieldwork study on the Calakmul Biosphere Reserve in Campeche State in Southern Me xico, where he fell in love with the tropical forest. He was married in 1998 and in 2000 he went to the University of Florida to pursue a masters degree in wildlife ecology and conservati on. In 2002, he finished his masters degree and immediately started the PhD program. From 2004 to 2006, he and his wife lived in the village of Zoh Laguna while doing fieldwork. During that time of capturing and following peccaries his daughter Aranza was born. The family moved to Gainesville in late 2006 to finish the PhD program. Plans for future include going back to any place on south Mexico and working on interesting research with tropical animals, while opening a rain forest caf in Zoh Laguna village, where Aranza and her possible brother/sister can liv e an interesting life in contact with nature and lovely people.