<%BANNER%>

Forest, Fragments, and Fruit: Spatial and Temporal Variation in Habitat Quality for Two Species of Frugivorous Primates ...


PAGE 1

FOREST, FRAGMENTS, AND FRUIT: SPATIAL AND TEMPORAL VARIATION IN HABITAT QUALITY FOR TWO SPE CIES OF FRUGIVOROUS PRIMATES ( Cercopithecus mitis AND Lophocebus albigena ) IN KIBALE NATIONAL PARK, UGANDA By CEDRIC ODRISCOLL WORMAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Cedric ODriscoll Worman

PAGE 3

This document is dedicated to all those who ha ve helped protect Kibale Forest and to the people of Karugutu.

PAGE 4

ACKNOWLEDGMENTS First and foremost I thank my family for not only the tradition of scholarship, but also a passion for it. My field assistant, Amooti Katusabe Swaibu, is a man of great integrity and honor in addition to possessing a wonderful knowledge of the forest, far beyond what I could have hoped for. Without his skills, my work could not have been done. Without his quick mind and original ideas, my work would not be half as good as I hope it is. Without his tolerant camaraderie, my work would have been just that: work. Webaale muno, Amooti! I thank my advisor, Colin Chapman, for all the help, advice, and facilitation he has given to me during this project, and my committee members, Graeme Cumming and Bob Holt, for their patience and suggestions. I thank the other researchers at Kanyawara for making my stay there interesting and sociable, most especially Claudia Stickler for being there. I thank Mike Wasserman and Lisa Danish for help with the laboratory analysis. Lastly, I thank Manjula Tiwari, John Poulsen, Toshinori Okuyama, Greg Pryor, Tristan Kimball, and all the other graduate students in the Department of Zoology for all their help with my research and writing. Funding was provided by an Alumni Fellowship from the University of Florida. Permission to conduct research in Kibale National Park was granted by the Makerere University Biological Field Station, Uganda Wildlife Authority, the National Council on Science and Technology, and the Office of the President of Uganda, and Kibale National Park. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.........................................................................................................................x CHAPTER 1 INTRODUCTION........................................................................................................1 Study Site......................................................................................................................4 Study Species................................................................................................................5 Blue Monkeys........................................................................................................5 Gray-cheeked Mangabeys.....................................................................................6 Overview.......................................................................................................................7 2 DENSITIES OF TWO FRUGIVOROUS PRIMATES WITH RESPECT TO FOREST AND FRAGMENT TREE SPECIES COMPOSITION AND FRUIT AVAILABILITY..........................................................................................................8 Introduction...................................................................................................................8 Methods......................................................................................................................10 Study Site.............................................................................................................10 Forest transects.............................................................................................11 Fragments.....................................................................................................14 Primate Diets................................................................................................14 Primate Densities.................................................................................................14 Forest Composition.............................................................................................19 Fruit and Flower Availability..............................................................................21 Results.........................................................................................................................22 Primate Densities.................................................................................................22 Forest Composition.............................................................................................24 Fruit Availability.................................................................................................29 Discussion...................................................................................................................38 Primate Densities.................................................................................................38 Forest Composition .............................................................................................39 v

PAGE 6

Fruit Availability .................................................................................................43 Transect 1: A Primate Perspective ......................................................................43 Summary.....................................................................................................................45 3 SEASONAL VARIATION IN THE QUALITY OF A TROPICAL RIPE FRUIT AND THE EFFECT ON THE DIETS OF THREE FRUGIVORES .........................49 Introduction.................................................................................................................49 Methods ......................................................................................................................50 Results.........................................................................................................................54 Discussion...................................................................................................................56 4 CONCLUSION...........................................................................................................63 APPENDIX: POTENTIAL FACTORS INFLUENCING HABITAT USE BY Cercopithecus mitis AND Lophocebus albigena .......................................................66 Unexamined Possibilities............................................................................................66 Light Levels, Wind and Disease..........................................................................66 Human Impacts....................................................................................................67 Predation..............................................................................................................69 Arthropod Abundance .........................................................................................70 Examined Possibilities................................................................................................71 Temperature Extremes and Evaporative Potential ..............................................71 Forest Composition and Fruit Availability..........................................................73 Nutritional Quality of Vegetable Foods ..............................................................73 Arthropod Foraging Substrate .............................................................................73 LITERATURE CITED......................................................................................................76 BIOGRAPHICAL SKETCH .............................................................................................87 vi

PAGE 7

LIST OF TABLES Table page 2-1 Species specific food items constituting 4% of the annual diet of Kanyawara C. mitis and L. albigena groups. ...........................................................15 2-2 Group densities of the arboreal monkeys of Kanyawara by transect.......................23 2-3 One-tailed Pearsons correlations between primate densities and relative densities and the basal area densities of various tree categories and fruit availability (average summed standardized fruiting intensity)................................23 2-4 Groups of the arboreal monkeys of Kanyawara sighted/km walked by transect during the monkey surveys......................................................................................24 2-5 The basal area densities (cm2 tree basal area/m2 land) of trees 10 cm dbh in transect 1 and the forest section of transect 1 compared to the 97.5% confidence limits of the fragments and the other transects.................25 2-6 The top ten tree species by basal area density of individuals 10 cm dbh in the forest fragments and the transects characterizing the main forest block near Kanyawara, Kibale National Park..........25 vii

PAGE 8

LIST OF FIGURES Figure page 1-1 Location of the Makerere University Biological Field Station (MUBFS), Kibale National Park, Uganda...............................................................................................4 2-1 The locations of the 8 transects in Kibale National Park, Uganda...........................13 2-2 Standardized fruiting intensity for species whose fruits comprise 4% of the annual diet of C. mitis .............................................................................................30 2-3 Standardized fruiting intensity for species whose fruits comprise 4% of the annual diet of L. albigena .........................................................................................34 2-4 C. mitis territories in relation to the unused area of forest.......................................47 3-1 The relationship between the monthly dry matter lipid content of ripe Celtis durandii fruit and the summed average daily rainfalls of the concurrent and previous months at the Makerere University Biological Field Station, Kibale National Park............................................................................................................54 3-2 The relationships between reported dietary use of Celtis durandii fruit and lipid content of the fruit, average daily rainfall of the previous and concurrent months (a predictor of lipid content), and fruit availability..................................................55 3-3 Changes in the percentage of Celtis durandii ripe fruit in the diet of a Cercopithecus ascanius group relative to changes in fruit availability and fruit lipid content..............................................................................................................57 3-4 Changes in the percentage of Celtis durandii fruit in the diet of a Cercopithecus mitis group relative to changes in fruit availability and estimated fruit lipid content......................................................................................................................58 3-5 Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group relative to changes in fruit availability and estimated fruit lipid content ......................................................................................................................59 3-6 Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group relative to changes in estimated fruit lipid content..........................60 viii

PAGE 9

A-1 The intraand inter-measurement period variation of 10 evaporation potential monitors from the same location..............................................................................72 ix

PAGE 10

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FOREST, FRAGMENTS, AND FRUIT: SPATIAL AND TEMPORAL VARIATION IN HABITAT QUALITY FOR TWO SPECIES OF FRUGIVOROUS PRIMATES (Cercopithecus mitis AND Lophocebus albigena) IN KIBALE NATIONAL PARK, UGANDA By Cedric ODriscoll Worman August, 2004 Chair: Colin A. Chapman Major Department: Zoology Conservation of wildlife populations requires extensive knowledge of their habitat requirements, efficient methods of evaluating habitat quality, and an understanding of the habitat value of fragments and edges. In this study I first evaluate the relationships between the basal area densities of several types of important food trees and fruit availability with the densities of two species of frugivorous monkeys (Cercopithecus mitis, the Blue monkey, and Lophocebus albigena, the Gray-cheeked mangabey) that have varying densities throughout the study area, Kibale National Park, Uganda, which includes a region of forest that is unused by both species. The density of C. mitis was most strongly correlated with the basal area density of all types of food trees combined. The density of L. albigena was not correlated with the basal area densities of any category of food trees or with fruit availability. However, an index of density, number of groups seen per kilometer walked, was marginally correlated to fruit availability. The x

PAGE 11

lack of a satisfactory relationship between the basal area densities of food trees and L. albigena density may be due to a mismatch in scale between this study and their large home ranges. The unused area of forest was then compared to the other areas of the forest and the forest fragments outside the park that are also uninhabited by the two species in question. It was found to have higher basal area densities in all food tree categories for both species than the forest fragments and lower basal area densities of most categories than the other parts of the forest, indicating that the forest fragments are very poor quality and would be unused even if dispersal were likely. The fragments were also found to have a higher frequency of exotic and edge tree species than the forest transects. Because fruit availability is so often used to evaluate habitat, but its quality is generally assumed to be constant, I examined nutritional components of a constantly available fruit important in the diets of many birds and mammals (Celtis durandii). The lipid content varied from 0.03-30.8% over 6 months and was predicted by the summed average daily rainfalls of the previous and concurrent months. Using 4 existing data sets, it was shown that lipid levels affected the intake of this fruit by at least 3 primate species, with more fruit being eaten when it contained more lipids. It appeared that fruit availability was being driven by consumption and not, as is often assumed, the other way around. This indicates that the seasonality of resources may be underestimated in tropical forests depending on the strategies of the fruiting trees xi

PAGE 12

CHAPTER 1 INTRODUCTION The continued loss of contiguous forest blocks via deforestation and fragmentation and the resulting dependence on preserves and parks for the survival of remaining forest species (Marsh 2003) has led to a triage system of conservation in which protection priority is defined by biotic value (various levels of biodiversity, presence of endangered or important organisms or habitats, etc.) (Prendergast et al. 1993, Krupnick and Kress 2003, Luck et al. 2004, Ortega-Huerta and Peterson 2004) or function (Poiani et al. 2000). This often opposes the more traditional conservation by consumption when areas are preserved simply because they are the least valuable for human use (e.g., mountains, deserts, disease ridden areas, etc.) (Willock 1964, Rodriguez-Toledo et al. 2003, Chapman et al. 2003b) or scarce and only valuable if protected (e.g., hunting/forestry reserves) (Jedrzejewska et al. 1994, Williams 2000). An assumption in the triage system is that non-protected habitat will be lost and protected areas will be isolated from each other (Newmark 1993); thus the ability of a protected area to support viable long-term populations is of paramount interest (Gurd et al. 2001). To fulfill the goals of protecting the biologically richest areas of (at least) the minimum required size to ensure healthy populations and managing them properly, habitat requirements must be known and the total area providing those requirements calculated (Kouki et al. 2001). The requirements of species that have broad distributions over many habitat types and flexible diets are difficult to define, but a comparison of differentially used areas within a single habitat type may be the first step in a broader understanding of their 1

PAGE 13

2 requirements (Chapman et al. 2002a). Ideally, similar studies should be done throughout a species range and compared for commonalities; however, depending on the species, this could be a daunting task. A more achievable, if less certain, method is to use knowledge of local conditions (diet, available tree species, etc.) to tailor the findings from one area to others. An understanding of habitat requirements is also vital to detailed estimations of available habitat and expected carrying capacity of a given area. Since it is unlikely that resources will be spread evenly throughout a protected area or even a habitat type, estimations of carrying capacities need to be based on the actual resource landscape and not on the richest (and often the most studied) areas. Even if an average for the entire area is used, the spatial distribution of resources or resource types may still be an important consideration (Pope et al. 2000). For example, two types of resources may be spatially separated requiring frequent travel and high energy expenditure or reducing the number of possible territories. This applies well to the many species of primate that have diets that are spatially and temporally variable and are widespread over multiple habitat/forest types. Kibale National Park provides an ideal situation to study the habitat requirements of two widespread species: Cercopithecus mitis (Blue monkey) and Lophocebus albigena (Gray-cheeked mangabey) because they both have variable densities throughout the forest, including an unused edge adjacent to frequented areas (C. Chapman, pers. comm.) and because of the extensive research trail system and detailed, long-term dietary and ranging information available from previous research (Struhsaker 1997). Additionally, there are forest fragments outside the main forest block which are unused by C. mitis and L.

PAGE 14

3 albigena (Onderdonk and Chapman 2000) and an examination of the fragments in relation to the used and unused areas of the main forest might help in understanding how fragmentation affects these two frugivores. Tropical high forests currently form 4.5% of Ugandas land cover, though over 30% of this is considered degraded (Kayanja and Byarugaba 2001). Outside of parks and reserves (which contain 14% of Ugandas area, (Howard et al. 2000) and hold 62% of the existing forest, (Kayanja and Byarugaba 2001)), tropical high forests are mainly relegated to small fragments in a human dominated matrix. For the successful conservation of forest species, the habitat value of these fragments and their possible importance in assisting demographic connection between metapopulations in the main forests must be more fully understood. If edges, fragments, degraded areas or other forest types offer undesirable habitat, the effective areas available to populations of forest species, like C. mitis and L. albigena, is much less than the total forest cover. If species are unable to use forest fragments to move between forest blocks, each local population may be isolated with higher than assumed risks of extinction (Brashares 2003) and loss of genetic variation (Segelbacher et al. 2003). The goal of this study was to contribute to the understanding of habitat requirements, and ultimately conservation, of C. mitis and L. albigena through the evaluation of potential habitat quality indicators with observed patterns of habitat use in different areas of the forest, and to evaluate the habitat potential of the forest fragments outside Kibale National Park thereby gaining a better understanding of what types of habitat characteristics potentially limit the distribution and density of these species over wider areas.

PAGE 15

4 Figure 1-1: Location of th e Makerere University Biol ogical Field Station (MUBFS), Kibale National Park, Uganda. Study Site Kibale National Park, Uganda (Figure 1-1) is mostly comprised of mid-altitude moist tropical forest (elevation: 920-1590 m, rainfall from 1990-2001: 1749 mm/year, latitude: 0 13’-0 41’ N) and has two ra iny and two dry seasons each year (though the length and severity varies greatly). The fr uiting of tree species in Kibale is usually synchronous, though the timing of fr uiting can be irregular, sub-annual, annual, or superannual depending on the species (Chapman et al. 1999b). Prior to becoming a national park, the ar ea was a forest reserve with the last logging in the natural forest occurring in 1969. This study took place in an unlogged

PAGE 16

5 compartment (K30) and a relatively untouched section of a lightly logged compartment (K14) (Chapman et al. 2000). Study Species Blue Monkeys Blue monkeys (Cercopithecus mitis stuhlmanni) belong to the C. (nictitans) superspecies, which includes both the Putty-nosed monkeys (C. (n.) nictitans) of central and west African evergreen forests and the Gentle monkeys (C. (n.) mitis) of eastern and southern African forests. Gentle monkeys are strongly arboreal and rarely descend from the trees, but have been observed foraging at ground level (Kingdon 1974, Beeson 1989). Their social structure normally consists of territorial matrilineal female-bonded groups with a single resident male (Rudran 1978, Butynski 1990). However, if many females come into estrus at once, the resident male may be temporarily unable to expel transient males from the group (Cords and Rowell 1986). The Gentle monkey is often considered frugivorous in spite of the catholic and variable diets of the subspecies that include ripe and unripe fruits, young and mature leaves, invertebrates, flowers, seeds, and, occasionally, bark (Kingdon 1974, Struhsaker 1978, Wrangham et al. 1998). Dietary variation is both spatial (between territories, forests, and subspecies) and temporal (monthly, seasonally, and yearly) and can be quite extreme (Rudran 1978, Lawes et al. 1990, Chapman et al. 2002b, Fairgrieve and Muhumuza 2003). Because of their diverse and variable diets, extensive geographic distribution, and broad climatic and habitat tolerance (Lawes et al. 1990), Gentle monkeys are considered classic generalists (Kingdon 1974, Struhsaker 1978, Butynski 1990, Fairgrieve 1995).

PAGE 17

6 The Blue monkey subspecies (C. m. stuhlmanni) occurs in eastern Congo, Uganda, western Kenya, and northern Tanzania (Lernould 1988). In Uganda, Blue monkeys are found on the eastern border near Mt.Elgon, on the north eastern border near Kidepo National Park, and extensively on the western side of the country including Budongo Forest Reserve, Ruwenzori Mountains National Park, Maramagambo Forest, and Kibale National Park. The southern border of Uganda is occupied by other subspecies, the Golden monkey (C. m. kandti) and the Silver monkey (C. m. doggetti). In Kibale National Park, Blue monkeys occur at lower densities (Chapman et al. 2000) and with larger home ranges (Rudran 1978, Butynski 1990) than in close by, more northerly, Budongo Forest (Aldrich-Blake 1970, Fairgrieve 1995) and are not found in any of the fragments outside the main forest block of Kibale (Onderdonk and Chapman 2000). Also, Blue monkey densities decline as one moves from the northern to the southern end of Kibale, suggesting a possible gradient in habitat suitability (Chapman and Lambert 2000). Gray-cheeked Mangabeys Gray-cheeked mangabeys (Lophocebus albigena johnstoni) are closely related to the baboons and, though strictly arboreal, retain many similarities with them such as multi-male groups, complex and frequent vocalizations, and eostrus advertisement with swollen and brightly colored perineal skin (Kingdon 1974). Gray-cheeked mangabeys specialize on intense fruiting events and consequently have large home ranges (Waser and Floody 1974). However they also eat other vegetative matter, including bark, young leaves, mature leaves, pith, and flowers (Olupot 1998) and spend a substantial amount of time foraging for arthropods (Waser 1975). Much like the Blue monkey, Gray-cheeked mangabeys appear to forage for more sedentary arthropods in, as well as on, substrates.

PAGE 18

7 They often rip apart rotten wood, tear off dead bark, and split twigs as well as searching through moss and epiphyte layers (Struhsaker 1978). Grey-cheeked mangabeys range from the Central African coast to the Nile, with their distribution in Uganda limited to a fairly narrow band from Semuliki National Park in the west to Mabira Central Forest Reserve in the east (Kingdon 1974). Overview To gain an understanding of habitat requirements and gauge the effectiveness of measures of habitat quality, different measures of habitat quality were done and related to primate use of the areas measured both in the forest and in the forest fragments. Nutritional analysis of food items was also done to test whether the measures of habitat quality were adequate indicators of actual resource availability. Chapter 2 deals with the composition of the forest (basal area density of food trees) and fruit availability and how well different measures of food availability are correlated to observed monkey densities, as well as the composition of the fragments and how they relate to the forest transects in terms of habitat for C. mitis and L. albigena. Chapter 3 explores the temporal variability in the nutritional quality of the fruit of a common tree species and its impacts on the diets of three species of primate. This chapter discusses the importance of considering changes in fruit quality and not just fruit availability when describing changes in habitat quality and resource availability over time. Chapter 4 concludes and summarizes.

PAGE 19

CHAPTER 2 DENSITIES OF TWO FRUGIVOROUS PRIMATES WITH RESPECT TO FOREST AND FRAGMENT TREE SPECIES COMPOSITION AND FRUIT AVAILABILITY Introduction Primates are currently of great interest to conservation not only because of their potentials for acting as flagship species (Karanth 1992, Vargas et al. 2002) but also because half of the worlds primate species are in trouble for a variety of reasons (Chapman and Peres 2001). While hunting is an important and widespread threat (Peres 1990, Chapman et al. 1999a), the dependence of most primate species on tropical forests (Mittermeier and Cheney 1987) and the continuing devastation of these forests on a global scale (DeFries et al. 2002) make an understanding of primate habitat requirements, limitations, and flexibilities in relation to heterogeneity in primary and degraded forests paramount for conservation. Forest composition has often been studied as the major factor determining the abundance and distribution of forest dwelling primates. These studies have sometimes been conducted in reference to the changes initiated by logging (e.g., Wilson and Wilson 1975, Skorupa 1988, Plumptre and Reynolds 1994, Rao and van Schaik 1997, Chapman et al. 2000, Olupot 2000, Fairgrieve and Muhumuza 2003), or with reference to the unique characters of fragmented forests (Woodwell 2002; e.g., Medley 1993, Granjon et al. 1996, Onderdonk and Chapman 2000, Umapathy and Kumar 2000, Marsh and Loiselle 2003, Norconk and Grafton 2003). In both cases, changes in food availability are often seen as a main driving force behind changes in primate densities (for an 8

PAGE 20

9 exception see Onderdonk and Chapman 2000). In fragments, changes in food availability can be caused by many processes, such as greater windfall, the cutting of firewood, an increase in proportion of unfavorable microclimate, demographic stocasticity, distance from a main forest block, or the extinction of primary dispersers (Cordeiro and Howe 2001). Many previous studies are limited with respect to identifying key habitat requirements because of the time lag between a dramatic disturbance (logging or fragmentation) and a response by the species in question (Brooks et al. 1999, Chapman et al. 2000, Gonzalez and Chaneton 2002). In the case of fragmentation, this limitation is compounded by the difficulty of separating habitat requirements from dispersal abilities of a species. If a species of interest is absent from fragments, it could be due either to insufficient resources or the inability of the species to transfer from the main forest or among fragments. The situation in Kibale National Park, Uganda can be used to address the issue of habitat requirements by relating habitat characteristics and natural heterogeneity of use in areas of the forest that have not been impacted significantly by commercial logging, thereby avoiding the time lag issue. The dispersal issue can be circumvented by examining a fragment-like area (it abuts human settlement, and has a large proportion of edge) completely avoided by the species in question (Cercopithecus mitis and Lophocebus albigena) but contiguous with inhabited forest. It is expected that, within Kibale National Park, the observed densities of C. mitis and L. albigena will be related to the abundance of their respective foods. Given that there are forest fragments near Kibale inhabited by the same diurnal primates as the

PAGE 21

10 avoided area (including an absolute lack of C. mitis and L. albigena groups), with the exception of Cercopithecus lhoesti (Onderdonk and Chapman 2000), the question arises as to whether these fragments are unused by C. mitis and L. albigena because of a lack of resources or due to poor dispersal abilities in these species. By comparing the availability of resources within the fragments with their availability in the used and unused portions of the contiguous forest, the habitat suitability of the fragments can be gauged. However, if the fragments have poor habitat quality, the hypothesis that distance from the main forest block may limit C. mitis and L. albigena use of the fragments cannot be rejected, yet is moot from a conservation perspective unless the fragments increase in habitat quality in the future. Methods Study Site Kibale National Park, Uganda is mostly comprised of mid-altitude moist tropical forest (elevation: 920-1590 m, rainfall from 1990-1998: 1778 mm/year, latitude: 0 13-0 41 N) and has two rainy and two dry seasons each year (though the length and severity varies greatly). The fruiting of tree species in Kibale is usually synchronous, though the timing of fruiting can be irregular, sub-annual, annual, or super-annual depending on the species (Chapman et al. 1999b). Prior to becoming a national park, the area was a forest reserve with the last logging in the natural forest occurring in 1969. This study took place in an unlogged compartment (K30) and a relatively untouched section of a lightly logged compartment (K14) (Chapman et al. 2000). Outside the main forest block contained in the national park, community owned forest fragments provide habitat for many of the forest primate species (Onderdonk and Chapman 2000).

PAGE 22

11 Forest transects A system of transects was established in the existing research trail grid near the Makerere University Biological Field Station (MUBFS), near the Kanayawara village, Kibale National Park, Uganda in an area that has not been impacted much by commercial logging (C. Chapman, pers. comm.). Because edges are increasingly prevalent sources of habitat heterogeneity, one transect (transect 1) was placed along the avoided forest edge and another along an edge known to be used by both C. mitis and L. albigena (transect 8). Both edge transects were paired with a nearby interior transect (transect 1 was paired with transect 3 and 8 with 6) and each pair was connected at both ends by two transitional transects (transects 2, 4, 5, and 7) (Figure 2-1). Due to the lay of the land and limitations of the existing trail system, transects ranged from 453 m to 1205 m long, with a mean of 767 m. Because transect 1 ventures beyond the forest, for some analyses only the area of the transect in the forest was considered to ensure that any observed differences were not due simply to the open and unique character of the field station. The section of transect 2 that ran through the area avoided by C. mitis and L. albigena (represented by transect 1) was discarded for analysis to remove the influence of the avoided area and make transect 2 representative of a utilized forest area. Due to the spatial arrangement of the transects, there is an area of overlap at some of the corners, but this is quite small, accounting for less than 0.5% of the total area measured, therefore each transect is considered independent.

PAGE 23

Figure 2-1: The locations of the 8 transect s in Kibale National Park, Uganda. The transects are shown with thick lines. Roads are shown with thin lines. The discarded area of transect 2 is shown with a dashed line. The Makerere University Biological Field Station (MUB FS) is also labeled. The satellite image is a Quickbird (high resolutio n, 2.4 m) from DigitalGlobe, Inc., Longmont, CO, USA.

PAGE 24

13

PAGE 25

14 Fragments The human dominated matrix outside Kibale National Park consists mainly of small-scale non-industrialized agriculture, large-scale tea plantations, pastures and fallow land with forest fragments surviving on agriculturally undesirable topography such as steep slopes and swampy lowlands (Chapman et al. 2003b). The fragments are actively used by the surrounding people for fuelwood, poles, livestock fodder, and medicinal and food plants (Chapman et al. 2003b). Many of these fragments are inhabited by the same primates as the main forest block in the national park with the notable exceptions of C. mitis and L. albigena (Onderdonk and Chapman 2000). Primate Diets Only specific dietary items (i.e., a specific part from a particular species) that constituted 4% of the total annual diet according to Rudran (1978) and Butynski (1990) for C. mitis or Waser (1975) and Olupot (1994) for L. albigena were considered. This produced 10 specific dietary items for C. mitis and 7 for L. albigena (Table 2-1). Though Olupot (1994) reports liana fruits as the second most frequent dietary item for L. albigena (8.34%), they were ignored in this study because there are a number of species, liana species in general are poorly known and difficult to identify, and the dominance of lianas measured in basal area is not comparable to that of trees. Primate Densities Primate group densities were estimated using a line transect method, which is appropriate for easily detectable diurnal primates (National Research Council 1981, Chapman et al. 2000). The transects were surveyed for primates once a fortnight for a year and a half (7/01 to 12/02). Observers walked the transects at a speed of about

PAGE 26

15 Table 2-1. Species specific food items constituting 4% of the annual diet of Kanyawara C. mitis and L. albigena groups. Butynskis (1990) numbers are the averages of study groups 1-4. Rudrans (1978) numbers are for groups 1 (not in parentheses) and 2 (in parentheses). Olupot (1994) and Rudran (1978) reported percentages of vegetable diet. Their numbers have been recalculated as percentages of total diet. The order of sources is constant throughout the table with Butynski and Olupot above and Rudran and Waser below for all species. Percent of annual diet Species Part C. mitis L. albigena Celtis africana fruit 4.1 (Butynski 1990)* 6.6(6.8) (Rudran 1978) -(Olupot 1994) -(Waser 1975) Celtis durandii fruit -4.1(5.2) 12.4 6.3 Croton macrostachys fruit --4.3 -Diospyros abyssinica fruit --6.2 22.6 Diospyros abyssinica leaves 4.8 ---Ficus bracylepis fruit -4.9 4.2 6.1 Ficus exasperata fruit -(11.2) 4.6 -Markhamia lutea young petioles 8.8 ---Pancovia turbinata fruit -6.6 -4.1 Parinari exelsa invertebrates ---5.0 Premna angolensis flowers -4.3 --Teclea nobilis fruit -5.3 --Uvariopsis congensis fruit -5.3(8.5) --*Butynski (1990) only reported foods used by Kanyawara groups and ignored at Ngogo. 1 km/hr during the morning hours (7:30-10:30) and recorded the method with which each primate troop was first detected, the observer to animal distance (from the first individual seen), the perpendicular trail to animal distance, the height of the observed individual, the tree species being occupied by the troop, and any observed feeding. Solitary animals were noted, but excluded from analysis. All distances and heights were estimated by trained and practiced field assistants. Information for groups

PAGE 27

16 at transect intersections was recorded separately if the group was seen from both transects (or was likely to have been seen but moved at the approach of the observers). Unfortunately, there were not enough sightings of C. mitis or L. albigena groups to use density estimating computer programs like Distance (Buckland et al. 1993) and there is debate over the most appropriate method of estimating densities from line transect counts for primates (Chapman et al. 2000). Unlike large ungulates, the landscape available for travel by arboreal primates is fully three dimensional which can lead to unusual sighting-distance histogram profiles which in turn are thought to overestimate group densities (National Research Council 1981). To control for this, observer to animal distance is often used instead of perpendicular distance (National Research Council 1981). This overestimates the area being sampled and can exclude observations of groups seen in the sampled area but from a long distance, therefore lowering density estimates. However, the histogram generated by this study using perpendicular distance and 5 m intervals was a classic Kelker histogram with a plateau extending from 0 to a distinct shoulder and subsequent steep decline in sightings. Therefore perpendicular distances were used to estimate densities. Because there were so few sightings of C. mitis or L. albigena, all primate group sightings were used to produce the Kelker histogram with the assumption that all species are equally visible. Even when unusual species (i.e., ground dwellers like Papio anubis (Olive baboon) and Cercopithecus lhoesti (LHoests monkey), species normally found in small quiet groups like Colobus guereza (Black-and-white colobus), and Pan troglodytes (Chimpanzee), which often exhibits both characters) were excluded, the profile of the histogram did not change meaningfully.

PAGE 28

17 One of the assumptions of the Kelker method is that 100% of the animals (or groups) within a certain distance are observed. That distance (multiplied by the transect length) is used to delineate the area sampled for density calculations. Ideally, the cutoff distance is determined by the sharp drop-off of observations after an initial plateau in the sighting distance histogram. It is assumed that the animals are distributed randomly in space (at least with respect to the transect) so the drop-off in observations is caused by a drop-off in the sighting probability. Previous studies have often used an objective method for determining the cutoff distance (Chapman et al. 2000) but this assumes a sudden decline in the probability of being seen with distance from transect instead of a slow, consistent drop. Using the same objective criteria as Chapman et al. (2000), the cut-off for this study would be 44 m, which is almost twice the subjectively determined cutoff (24 m) and seems unreasonable for this forest. The violation of the assumption that no groups within the cutoff are missed underestimates densities. There was no correlation between distance from transect and number of groups seen per meter within the 24 m cutoff (Pearsons r = -0.157, p = 0.227, N = 25) indicating that this assumption is not significantly violated. Pearsons correlations were used to detect relationships between the transect densities of C. mitis and the average summed standardized fruiting intensity (see Fruit Availability below) for all fruit species, and the basal area density of all trees, species specific food trees, fruit trees, leaf trees, and flower trees. In these analyses one-tailed probabilities were used since clear a priori predictions were previously made. Pearsons correlations were also used to detect relationships between transect densities of L. albigena and the average summed standardized fruiting intensity for all fruit species, and

PAGE 29

18 the basal area density of all trees, species specific food trees, fruit trees, and arthropod trees (tree species used as invertebrate foraging substrate). Only the forest area was used to represent transect 1 for these analyses. Comparing a relative index of density (groups sighted/km walked during the surveys) to the density estimations for each species can be used to strengthen confidence in the density estimations (Chapman et al. 2000). In this case, the small numbers of sightings may make density estimates particularly susceptible to the influence of a single sighting and its inclusion or exclusion due to distance from the transect, therefore the Pearsons correlations were recalculated using the index of relative density for both primate species. Checking the results in this manner was considered more important for L. albigena because the relative index of density was less strongly correlated with calculated density for L. albigena (r = 0.736, one-tailed p = 0.019, N = 8) than for C. mitis (r = 0.986, one-tailed p < 0.001, N = 8). Though invertebrates generally form an important part of the diet of C. mitis (Twinomugisha et al. in press), this is not used as a dietary category because no particular species of tree was used enough for invertebrate foraging to qualify as an important food species. Therefore the lack of an invertebrate foraging substrate category in food trees should not be taken to mean that invertebrates are unimportant, but as an indication that C. mitis may not be as selective about the species in which they forage for invertebrates as they are about the species in which they forage for fruit, leaves, or flowers. The results for the P. troglodytes, C. lhoesti, and P. anubis are not given as they have more terrestrial habits (making line transect surveys less appropriate) and were

PAGE 30

19 rarely seen during surveys, despite the frequent observation of the animals or their sign at other times. Forest Composition Free standing woody plants (including strangler figs) with a dbh 10 cm and within 10 m of either side of a transect were identified. Each stem was measured (dbh), the height estimated, and it was noted if the plant was a snag, was bent or broken, or if the main stem was dead. Snags were excluded from the analyses. If a plant was in an area of overlap between transects, it was counted in each separately. Attempts were made to measure above any pronounced buttressing, and, although this was not always feasible, only a few of the largest trees of rarer species presented this sort of problem. Multiple stemmed plants were treated as if each stem were a separate individual. A fig with a diameter of more than 10 cm 3 m up on the host and with a substantial crown could be supported by only small roots at breast height, making dbh a meaningless measure. If multiple roots converged into a single trunk before the trunk started branching, the diameter of the single trunk was measured or estimated. If the fig had large roots connecting above the branching point (essentially multiple trunks), the main trunk and several of the large roots were measured and combined to give a single dbh estimate. These huge figs are widely spaced and were captured by the transects very few times. Most of the fig trees measured presented no unusual difficulties. The dbh of each plant was converted to basal area and summed for each species to produce the total basal area by species for each transect. The total species basal area was divided by the total area covered by a given transect, producing a basal area density (cm2/m2) for each species for each transect.

PAGE 31

20 In the fragments, a complete census of every tree 10 cm dbh was done. The forest fragments and transects were compared using bootstrapping analysis in order to see if the forest composition of the unused transect was more similar to that of the fragments or the rest of the forest. The variables used to characterize the transects and fragments for C. mitis were the basal area densities of all trees, all C. mitis food trees, fruit trees, leaf trees, and flower trees. Those for L. albigena were the basal area densities of all trees, all L. albigena food trees, fruit trees, and arthropod foraging trees. It was assumed that transect 1 would fall in between the fragments and the forest so the upper 97.5% confidence limit was calculated for the fragment characters and the lower 97.5% confidence limit was calculated for the transect characters and compared to transect 1 and transect 1 forest only. For each variable, a sample population equal in size to the original population was created by sampling the original population with replacement 10,000 times. The 97.5% confidence limits for each variable were calculated from the distributions of the means of the sample populations. To determine if the habitat of the fragments mirrored that of the main block of forest, the top ten trees by basal area density for each fragment and transect were classed as edge/savanna if they were described in Hamilton (1991) as only, primarily or commonly occurring in forest edges or savannas, otherwise they were classed as forest trees. Eucalyptus sp. was the only exotic in the top ten lists and was treated as an edge species because it is either planted by humans or disperses from the human dominated matrix outside the fragments. 2 analyses were done on the frequency of edge/savanna and/or exotic species in the ten most dominant species in the fragments versus the transects. All of transect 1 was included in the transect group in order to be as

PAGE 32

21 conservative as possible. One species, Markhamia lutea (formerly M. platycalyx), defined by Hamilton (1991) as, [a] very common forest edge species, sometimes found within forests, either where the canopy is fairly open or where there has been a large gap, is characterized by Struhsaker (1997) as a late successional or old growth forest species. This disparity may be due to site specific conditions so a second set of 2 analyses were done with Markhamia lutea defined as a forest interior tree. Fruit and Flower Availability Each month for 6 months (June, 2002-November, 2002), phenological data (presence of ripe and unripe fruit and fruiting/flowering intensity) were collected from all transects, including the dbh of the fruiting trees ( 5 cm dbh) important in the diet of either species within 10 m of either side of the transects. The fruiting intensity was determined by assigning each tree to a numerical class (1-5) based on the fruit density in the crown, with each class indicating about twice the density as the next lower class and 3 being the average fruit density of the species. Because basal area is strongly correlated with crown size, the basal area of each tree was multiplied by a factor (, 1, 2, or 4, respectively) based on the fruit density class to estimate the total fruit load on each tree. The fruit loads were summed (trees having both ripe and unripe fruits were considered to have ripe and unripe fruit) and divided by the total area sampled to produce a standardized monthly fruiting intensity for each species and transect for both ripe and unripe fruit. The only flowering species eaten frequently, Premna angolensis, flowered during the study, but synchronously for a short duration and was missed by the surveys.

PAGE 33

22 Results Primate Densities L. albigena was found to have the lowest average density over all the transects with C. mitis the next lowest; however, densities varied between transects. Cercopithicus ascanius (Redtail monkey) and Piliocolobus tephrosceles (Red colobus) had the highest overall densities (Table 2-2). The density of C. mitis on the transects was correlated most strongly with total food tree basal area density, but was also correlated positively with fruit tree basal area density and leaf tree basal area density. C. mitis density was not positively correlated with average summed standardized fruiting intensity, total tree basal area density, or with flowering tree basal area density (Table 2-3). The C. mitis index of relative density (groups observed/km walked) (Table 2-4) was also correlated positively with total food tree basal area density, fruit tree basal area density, and leaf tree basal area density but not with average summed standardized fruiting intensity, total tree basal area density, nor with flowering tree basal area density (Table 2-3). L. albigena density was not positively correlated with average summed standardized fruiting intensity, total tree basal area density, total food tree basal area density, fruit tree basal area density, nor with invertebrate foraging tree basal area density (Table 2-3). Similarly, the L. albigena relative index of density (groups observed/km walked) (Table 2-4) was not positively correlated with total tree basal area density, total food tree basal area density, fruit tree basal area density, nor with invertebrate foraging tree basal area density. However there was a marginally significant positive correlation between the relative index of mangabey density and average summed standardized fruiting intensity (Table 2-3).

PAGE 34

23 Table 2-2. Group densities of the arboreal monkeys of Kanyawara by transect. 1 Field Station is the portion of transect 1 running through the Makerere University Biological Field Station. 1 Forest is the portion of transect 1 that is in the forested area. Transect 1 is the entirety of the transect. The means were calculated with all of Transect 1. Presence noted with opportunistic observations. Presence noted during surveys but at distances of >24 m. Group densities (groups/km2) Species Transects C. mitis L. albigena C. ascanius C. guereza P. tephrosceles 1 0.0 0.0 4.1 11.6 15.8 1 Field Station 0.0 0.0 6.8 25.5 15.3 1 Forest 0.0 0.0 2.3 2.3 16.1 2 5.9 0.0* 24.8 10.6 16.5 3 5.3 3.9 14.4 3.9 7.9 4 5.2 2.1 8.3 8.3 8.3 5 3.4 0.9 7.7 4.3 6.8 6 0.0 1.6 3.9 0.8 2.3 7 1.1 1.1 2.8 1.7 7.2 8 0.5 0.0 6.4 3.5 4.4 Mean 2.7 1.2 9.1 5.6 8.7 Table 2-3. One-tailed Pearsons correlations between primate densities and relative densities and the basal area densities of various tree categories and fruit availability (average summed standardized fruiting intensity). Positive correlations significant at = 0.05 are in bold font. Positive correlations significant at = 0.10 are underlined. In all cases N = 8. C. mitis density C. mitis relative density L. albigena density L. albigena relative density Basal area density and fruit avail. r p r p r p r p Total tree 0.306 0.231 0.371 0.183 0.032 0.470 0.042 0.406 Food tree 0.745 0.017 0.824 0.006 -0.159 0.353 -0.126 0.383 Fruit tree 0.662 0.037 0.741 0.018 0.212 0.307 0.134 0.376 Leaf tree 0.630 0.047 0.685 0.031 Flower tree -0.519 0.094 -0.572 0.069 Invert. tree -0.441 0.137 -0.304 0.232 Fruit avail. 0.239 0.285 0.237 0.289 0.399 0.164 0.596 0.060

PAGE 35

24 Table 2-4. Groups of the arboreal monkeys of Kanyawara sighted/km walked by transect during the monkey surveys. 1 Field Station is the portion of transect 1 running through the Makerere University Biological Field Station. 1 Forest is the portion of transect 1 that is in the forested area. Transect 1 is the entirety of the transect. Means were calculated with all of Transect 1. Presence noted with opportunistic observations. Relative group density index (groups sighted/km walked) Species Transects C. mitis L. albigena C. ascanius C. guereza P. tephrosceles 1 0.0 0.0 0.36 0.92 1.18 1 Field Station 0.0 0.0 0.57 1.63 1.31 1 Forest 0.0 0.0 0.22 0.44 1.10 2 0.34 0.0* 1.25 0.85 1.30 3 0.25 0.19 0.76 0.32 0.76 4 0.30 0.10 0.60 0.60 0.60 5 0.16 0.12 0.57 0.29 0.70 6 0.04 0.11 0.37 0.04 0.22 7 0.08 0.19 0.29 0.16 0.59 8 0.05 0.07 0.40 0.24 0.36 Mean 0.15 0.10 0.58 0.43 0.71 Forest Composition The 97.5% basal area density confidence intervals of the tree types (all trees, C. mitis fruit, leaf, flower and total food trees and L. albigena fruit, invertebrate foraging and total food trees) of the fragments and the forest transects never overlapped and were separated by large gaps, except in the case of L. albigena invertebrate foraging trees. Transect 1 and the forest part of transect 1 had higher basal area densities than the fragments with regards to every category. They had lower basal area densities than the other transects in every category except C. mitis flower trees, L. albigena total food trees, and L. albigena invertebrate foraging trees (Table 2-5). The 10 species with the highest basal area densities were more often exotic (Eucalyptus sp.) (df = 1, 2 = 5.6, p 0.025), edge/savanna species (excluding Eucalyptus sp.) (df = 1, 2 = 5.1, p 0.025), or both (df = 1, 2 = 6.0, p 0.025) in the fragments than in the forest transects even with the inclusion of all of transect 1 and

PAGE 36

25 Table 2-5. The basal area densities (cm2 tree basal area/m2 land) of trees 10 cm dbh in transect 1 and the forest section of transect 1 compared to the 97.5% confidence limits of the fragments and the other transects. Fragments: Upper 97.5% confidence limit (mean) Transect 1 Transect 1 forest only Transects 2-8: Lower 97.5% confidence limit (mean) All trees (10.4) 14.9 22.0 30.4 30.8 (36.7) Total C. mitis food trees (0.6) 0.8 2.2 3.6 10.7 (14.5) C. mitis fruit trees (0.3) 0.6 0.9 1.4 6.0 (8.4) C. mitis leaf trees (0.2) 0.3 0.9 1.5 3.6 (5.7) C. mitis flower trees (0.001) 0.002 0.4 0.7 0.1 (0.4) Total L. albigena food trees (1.1) 2.8 6.1 10.2 8.7 (10.8) L. albigena fruit trees (0.2) 0.3 1.4 2.2 6.5 (8.7) L. albigena invertebrate foraging trees (0.9) 2.5 4.8 8.0 0.6 (2.1) Table 2-6. The top ten tree species by basal area density of individuals 10 cm dbh in the forest fragments and the transects characterizing the main forest block near Kanyawara, Kibale National Park. Entries in bold indicate an important species in the diet of C. mitis. Underlined entries indicate an important species in the diet of L. albigena. *Described as occurring in forest edges or savannas by Hamilton (1991). Exotic Bugemba (4.7 ha) Durama (CK) (4.9 ha) Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Neoboutonia melleri 3.41* 34.8% Pseudospondias microcarpa 1.98 21.9% Ficus dawei 0.94 9.6 Aningeria altissima 0.62 6.8 Mitragyna rubrostipulata 0.70 7.2 Ficus vallis 0.56 6.3 Prunus africana 0.63* 6.4 Mitragyna rubrostipulata 0.56 6.2 Parinari excelsa 0.43 4.4 Prunus africana 0.55* 6.1 Mystroxylon aethiopicum 0.41* 4.2 Mimusops bagshawei 0.55 6.1 Symphonia globulifera 0.35 3.6 Rauvolfia sp. 0.44 4.9 Sapium ellipticum 0.34* 3.5 Macaranga schweinfurthii 0.33 3.6 Fagara sp. 0.34 3.4 Diospyros abyssinica 0.33 3.6 Eucalyptus sp. 0.27 2.7 Strombosia scheffleri 0.30 3.3 Top ten total 7.82 79.8% Top ten total 6.22 68.9%

PAGE 37

26 (Table 2-6. Continued) Kiko1trading (6.2 ha) Kiko2tea office (5.0) Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Neoboutonia melleri 0.59* 30.3% Eucalyptus sp. 0.75 28.0% Eucalyptus sp. 0.55 28.2 Neoboutonia melleri 0.31* 11.8 Sapium ellipticum 0.14* 6.9 Sapium ellipticum 0.30* 11.3 Erythrina abyssinica 0.13* 6.7 Cleistanthus polystachyus 0.21 7.7 Maesa lanceolata 0.12* 6.2 Symphonia globulifera 0.18 6.8 Alangium chinense 0.07* 3.4 Erythrina abyssinica 0.16* 6.2 Macaranga schweinfurthii 0.07 3.4 Macaranga schweinfurthii 0.13 4.7 Mitragyna rubrostipulata 0.06 2.9 Prunus africana 0.12* 4.4 Albizia sp. 0.05* 2.5 Alangium chinense 0.07* 2.5 Croton macrostachyus 0.04* 2.1 Bridelia bridelifolia 0.06* 2.3 Top ten total 1.81 92.5% Top ten total 2.28 86.0% Kiko3school fragment (1.7) Kiko4church (1.2 ha) Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Neoboutonia melleri 2.64* 29.0% Parinari excelsa 9.39 32.2% Eucalyptus sp. 1.44 15.8 Eucalyptus sp. 4.80 16.4 Sapium ellipticum 1.36* 15.0 Fagara sp. 2.20 7.5 Parinari excelsa 0.78 8.6 Neoboutonia melleri 2.16* 7.4 Macaranga schweinfurthii 0.69 7.6 Aningeria altissima 1.81 6.2 Erythrina abyssinica 0.37* 4.1 Newtonia buchananii 1.71 5.9 Bridelia bridelifolia 0.32* 3.5 Ficus capensis 1.12* 3.9 Markhamia lutea 0.30* 3.3 Symphonia globulifera 0.94 3.2 Prunus africana 0.17* 1.9 Prunus africana 0.81* 2.8 Mitragynia rubrostipulata 0.13 1.4 Erythrina abyssinica 0.58* 2.0 Top ten total 8.19 90.2% Top ten total 25.5 87.3%

PAGE 38

27 (Table 2-6. Continued) Rutoma (Big) (4.9 ha) Rwaihamba (2.4 ha) Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Aningeria altissima 1.15 24.7 Eucalyptus sp. 8.84 55.9% Pseudospondias microcarpa 0.71 15.3 Syzyguim sp. 2.41 15.3 Ficus dawei 0.35 7.5 Neoboutonia melleri 0.65* 4.1 Prunus africana 0.32* 6.9 Pseudospondias microcarpa 0.64 4.0 Diospyros abyssinica 0.20 4.2 Ficus dawei 0.57 3.6 Strombosia scheffleri 0.19 4.1 Polyscias fulva 0.56* 3.6 Fagara sp. 0.18 3.9 Ficus vallis 0.42 2.7 Cordia abyssinica 0.17* 3.6 Markhamia lutea 0.29* 1.9 Mitragyna rubrostipulata 0.12 2.6 Croton macrostachyus 0.28* 1.8 Ficus brachylepis 0.12 2.6 Cordia abyssinica 0.19* 1.2 Top ten total 3.51 75.4% Top ten total 14.9 94.1% Transect 1 Transect 1 forest Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Parinari excelsa 4.75 21.6% Parinari excelsa 7.96 26.2% Funtumia latifolia 1.97 9.0 Funtumia latifolia 3.20 10.5 Olea welwitschii 1.73* 7.9 Olea welwitschii 2.88* 9.5 Prunus africana 1.58* 7.2 Chrysophyllum gorungosanum 2.20 7.2 Chrysophyllum gorungosanum 1.31 6.0 Strombosia scheffleri 1.71 5.6 Strombosia scheffleri 1.02 4.6 Trilepsium madagascariense 1.34 4.4 Maesa lanceolata 0.94* 4.3 Doispyros abyssinica 1.22 4.0 Trilepsium madagascariense 0.80 3.6 Prunus africana 1.06* 3.5 Diospyros abyssinica 0.73 3.3 Mimusops bagshawei 0.92 3.0 Mimusops bagshawei 0.56 2.5 Newtonia buchananii 0.86 2.8 Top ten total 15.4 70.1% Top ten total 23.3 76.7%

PAGE 39

28 (Table 2-6. Continued) Transect 2 Transect 3 Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Ficus exasperata 7.04 14.9% Celtis durandii 4.94 14.7% Strombosia scheffleri 5.74 12.2 Markhamia lutea 4.65* 13.8 Funtumia latifolia 4.70 10.0 Strombosia scheffleri 4.47 13.3 Markhamia lutea 4.68* 9.9 Funtumia latifolia 3.95 11.7 Celtis durandii 4.00 8.5 Diospyros abyssinica 3.15 9.4 Celtis africana 3.13 6.6 Olea welwitschii 1.66* 4.9 Olea welwitschii 2.53* 5.4 Celtis africana 1.64 4.9 Parinari excelsa 2.14 4.5 Blighia unijugata 1.44* 4.3 Ficus polita 2.08 4.4 Mimusops bagshawei 1.19 3.6 Diospyros abyssinica 1.81 3.8 Trilepsium madagascariense 1.18 3.5 Top ten total 37.8 80.1% Top ten total 28.3 84.1% Transect 4 Transect 5 Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Markhamia lutea 7.25* 21.6% Strombosia scheffleri 5.63 15.7% Ficus exasperata 4.78 14.3 Parinari excelsa 3.95 11.0 Diospyros abyssinica 3.61 10.8 Aningeria altissima 3.81 10.6 Celtis durandii 3.35 10.0 Dombeya kirkii 2.69 7.5 Olea welwitschii 3.10* 9.2 Celtis durandii 2.63 7.3 Strombosia scheffleri 2.48 7.4 Funtumia latifolia 1.92 5.3 Funtumia latifolia 2.23 6.7 Chaetacme aristata 1.67* 4.7 Trilepsium madagascariense 1.14 3.4 Cordia abyssinica 1.58* 4.4 Celtis africana 0.78 2.3 Mimusops bagshawei 1.56 4.3 Ehretia cymosa 0.48* 1.4 Albizia grandebracteata 1.19* 3.3 Top ten total 29.2 87.1% Top ten total 26.6 74.2%

PAGE 40

29 (Table 2-6. Continued) Transect 6 Transect 7 Species Basal area density (cm2/m2) % of total basal area Species Basal area density (cm2/m2) % of total basal area Funtumia latifolia 5.81 13.7% Celtis durandii 5.77 13.8% Strombosia scheffleri 5.04 11.9 Strombosia scheffleri 4.93 11.8 Parinari excelsa 4.47 10.6 Parinari excelsa 4.25 10.2 Ficus natalensis 3.77 8.9 Olea welwitschii 3.01* 7.2 Celtis durandii 3.36 7.9 Aningeria altissima 2.68 6.4 Chrysophyllum gorungosanum 2.99 7.1 Diospyros abyssinica 2.21 5.3 Diospyros abyssinica 2.21 5.2 Funtumia latifolia 1.90 4.6 Aningeria altissima 1.78 4.2 Uvariopsis congensis 1.51 3.6 Blighia unijugata 1.67* 3.9 Ficus dawei 1.48 3.6 Markhamia lutea 1.46* 3.5 Symphonia globulifera 1.44 3.4 Top ten total 32.6 77.0% Top ten total 29.2 70.1% Transect 8 Species Basal area density (cm2/m2) % of total basal area Diospyros abyssinica 4.97 22.0% Albizia grandebracteata 3.51* 15.5 Celtis africana 1.98 8.8 Celtis durandii 1.89 8.4 Chaetacme aristata 1.60* 7.1 Dombeya kirkii 1.23 5.4 Cordia abyssinica 1.21* 5.4 Markhamia lutea 1.12* 5.0 Mimusops bagshawei 1.13 5.0 Ficus natalensis 0.94 4.1 Top ten total 19.6 86.8% transect 8 (Table 2-6). A second set of analyses with Markhamia lutea defined as an interior species confirmed the differences between the fragments and transects (edge/savanna species excluding Eucalyptus sp.: df = 1, 2 = 6.8, p 0.01; edge/savanna/exotic species: df = 1, 2 = 10.2, p 0.01). Fruit Availability Transect 1 had several species fruit during the study (e.g. one small Monodora myristica, several large Parinari excelsa, and two individuals in a grove of Diospyros abyssinica), but none of the important dietary species for C. mitis fruited on this transect.

PAGE 41

30 For L. albigena the only important tree species to fruit was D. abyssinica which fruited from July until at least November, leaving May and June with no available fruit from important species. For both primate species transects 3 and 7 ha d the highest fruiting intensity overall, with transect 2 having a large pulse from a single Ficus exasperata in November. Celtis durandii was the only fruit available in every month in all transects (except of course for transect 1) (Figure 2-2 a nd Figure 2-3). 0 1 2 3 4 5 6 7 8 9 10C. durandii U. congolensis C. durandii C. durandii C. durandii C. durandii T. nobilis U. congolensis C.durandii C.africana F. exasperata U. congolensis C. durandii F. exasperata MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 2* Figure 2-2: Standardized fruiting intensity for species whose fruits comprise 4% of the annual diet of C. mitis Transect 1 is not shown as no important trees fruited on this transect during the study. Note the consistent fruiting of C. durandii in all months and transects and the pulse of a single F. exasperata tree in transect 2. The F. exasperata pulse on transect 7 resulted from two individuals whose fruiting overlapped in Septembe r. Solid black bar sections indicate ripe fruit. Striped areas indicate unripe fruit. Fr uit was present but at extremely low intensity.

PAGE 42

31 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii C. durandii C. durandii C. durandii C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 3 0 1 2 3 4 5 6 7 8 9 10C. durandii F. exasperata C. durandii C. durandii C. durandii C. durandii T. nobilis C. durandii C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 4 Figure 2-2. (Continued)

PAGE 43

32 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii C. durandii C. durandii C. durandii C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 5 0 1 2 3 4 5 6 7 8 9 10C. durandii U. congolensis C. durandii C. durandii C. durandii C. durandii C. durandii C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 6 Figure 2-2. (Continued)

PAGE 44

33 0 1 2 3 4 5 6 7 8 9 10C. durandii U. congolensis C. durandii C. durandii F. exasperata C. durandii F. exasperata C. durandii F. exasperata C. durandii F. exasperata C. durandii F. exasperata MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 7 0 1 2 3 4 5 6 7 8 9 10C. africana C. durandii C. africana C. durandii T. nobilis C. africana C. durandii C. africana C. durandii C. africana C. durandii C. africana C. durandii C. africana C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityTransect 8* *** * Figure 2-2. (Continued)

PAGE 45

34 Transect 1 0 1 2 3 4 5 6 7 8 9 10D. abyssinica D. abyssinica D. abyssinica D. abyssinica D. abyssinica MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting Intensity Transect 2 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii C. durandii C. durandii D. abyssinica C. durandii D. abyssinica F. exasperata C. durandii D. abyssinica F. exasperata MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting Intensity Figure 2-3: Standardized fruiting intensity for species whose fruits comprise 4% of the annual diet of L. albigena Note the consistent fruiting of C. durandii in all months and transects (except transect 1), the pulse of a single F. exasperata tree in transect 2, and the synchronized fruiting of D. abyssinica in the latter part of the study in all transects. The F. exasperata pulse on transect 7 resulted from two trees whose fruiting overlapped in September. Solid black bar sections indicate ripe fruit. Striped areas indicate unripe fruit.

PAGE 46

35 Transect 3 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting Intensity Transect 4 0 1 2 3 4 5 6 7 8 9 10C.durandii F.exasperata C. durandii C. durandii D. abyssinica C. durandii D. abyssinica C. durandii Diospyros abyssinica C. durandii D.abyssinica C. durandii D. abyssinica MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityFigure 2-3. (Continued)

PAGE 47

36 Transect 5 0 1 2 3 4 5 6 7 8 9 10C. durandii C.durandii C. durandii D. abyssinica C. durandii C. durandii C. durandii C. durandii MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting Intensity Transect 6 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D.abyssinica C. durandii D. abyssinica MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityFigure 2-3. (Continued)

PAGE 48

37 Transect7 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii C. durandii D. abyssinica F. exasperata C. durandii D. abyssinica F. exasperata C. durandii D. abyssinica F. exasperata C. durandii D. abyssinica F. exasperata C.durandii D.abyssinica F. exasperata MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting Intensity Transect 8 0 1 2 3 4 5 6 7 8 9 10C. durandii C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C. durandii D. abyssinica C.durandii D. abyssinica MayJuneJulyAug.Sept.Oct.Nov. Standardized Fruiting IntensityFigure 2-3. (Continued)

PAGE 49

38 Discussion Primate Densities C. mitis and L. albigena groups were never observed or re ported to have been in the field station or the adjoining forest, despite the constant presence of potential observers over many years. The other primates are regul arly seen in this area. The correlations between C. mitis density and food tree ba sal area density, fruit tr ee basal area density, and leaf tree basal area density sugge st that, in the case of this species, and potentially other generalists and folivores, habitat quality might be indexed dependably and quickly with a knowledge of local diet and corresponding basal area measurements of only a few important food species (see also Siex and Struhs aker 1999). It is notable that total food tree basal area density is more highly correla ted with monkey density than that of any dietary component alone. This is not surp rising considering the generalist nature of C. mitis There are several possibly cont ributing factors involved in the lack of correlations between the basal area densities of the tree categories and L. albigena density. As mentioned above, liana fruit was the sec ond most important mangabey food item of Olupots (1994) study, but is not accounted fo r here, weakening the predictive power of the food plant index. Additionally, L. albigena home ranges are large (410 ha), an order of magnitude larger than C. mitis home ranges (61 ha) (Struhsaker and Leland 1979), and the scale of the system of transects is pr obably not large enough to match the scale at which the mangabeys travel. One of the reasons for such wide ranging in L. albigena is their exploitation of rare, intense fruiting events incl uding those produced by widely separated large Ficus spp. (Waser 1975). The area of the transects is most likely insufficient to accurately repr esent the occurrence of huge, widely-spaced, and important

PAGE 50

39 figs. Adjusting the size of a study area to include multiple home ranges could alleviate this problem. The average summed standardized fruiting in tensities were not anticipated to have a strong relationship with the densities of either monkey species because of the assumptions inherent in calculating a fruit av ailability index for each transect over a long period. By summing all fruiting species, it is necessarily assumed that the normal fruit load on individual trees of a speci es is equivalent to that of any other species in terms of value to the frugivores. It is also assumed that the value of the fruit does not change from one month to the next. Given that the assump tion of normal fruit load interchangability is almost certainly violated and that the a ssumption of unchanging monthly fruit value is violated (Chapter 3), the near signific ance of the correlation between number of mangabey groups seen per km walked during the primate surveys and average summed standardized fruiting intensity is as much as can be expected. In view of the specialization of L. albigena on concentrated fruiting events compared with the more generalized diet of C. mitis it is logical that habitat use by L. albigena would be more easily related to fruit availability and habitat use by C. mitis would be more related to the occurrence of all types of f ood trees (see also Beeson 1989). Fruit availability might be an acceptable indicator of habitat quality for fruit specialists like L. albigena but given the temporal variation in fruit load, long-term studies w ould be necessary. Forest Composition The fragments have lower basal area dens ities of trees in general and a higher proportion of the top ten trees by basal area de nsity is composed of edge/savanna trees, making the fragments different from Kibale forest in more ways than just being physically isolated. This in itself does not imply that the fragments are unsuitable habitat

PAGE 51

40 for resident or transient groups of C. mitis or L. albigena as both are found widely through equatorial Africa in diverse habitats (Kingdon 1997). Similarly, the fact that the fragments are much smaller in area than C. mitis and L. albigena home ranges at Kanyawara does not automatically exclude these species as other primates, C. ascanius and P. troglodytes are able to move among fragment s to meet their needs (Chapman et al. 2003b). While this may have an im pact on the use of the fragments by C. mitis which rarely descends to the ground (Kingdon 1974, Rudran 1978, Beeson 1989) or crosses open areas (Devos and Omar 1971, Fairgrieve 1995, Lawes 2002), L. albigena has been reported to do as well by utilizing several fragments as inhabiting contiguous forest (Tutin et al. 1997). Additionally, there are larger fr agments not available for the analysis of forest composition, Kasisi (130 ha) and Lake Mwamba (28.7 ha), that also lack both species (Onderdonk and Chapma n, 2000) despite being large enough to contain the home range of at least one C. mitis group (25-44 ha; Butynski, 1990), though not the entire home range of an L. albigena group (441 ha; Waser 1984). In any case, neither differences in forest structure nor size apply to the area of transect 1 since it could easily be used by a group from the adjoining areas. The fact that the unused area could be used in concert with populated adjacent areas and th erefore would not be required to provide all the resources necessary to support a group indicates th at the unused area does not produce resources at high enough levels to attract even casual use. Examining the basal area densities of the important food trees for both C. mitis and L. albigena it becomes clear how different the frag ments are from the inhabited forest. The area of transect 1 generally falls betw een the two. The only exceptions to this pattern are given by minor (though not necessari ly unimportant) dietary components. In

PAGE 52

41 the case of C. mitis transect 1 has the same basal area density of flower-producing species as the mean of the other transects. Ex cluding the field stati on simply increases the flower-producing basal area density. This s uggests that lack of flower-producing trees can be ruled out as a possible cause for the avoidance of this area by C. mitis Likewise, the total L. albigena food tree basal area densit y is only different from the other transects when the field station is in cluded. The forest al one fits in well with the rest of the transects in this characterist ic and actually had a higher basal area density of invertebrate foraging species than any other transect. However, these seem to be driven by the extremely hi gh basal area density of Parinari excelsa (the one important invertebrate foraging species for L. albigena ) in the area of transect 1. The high density of P. excelsa in one of the fragments also drives th e overlap of the fragment and transect confidence intervals for invertebrate foraging trees. This suggests that invertebrate foraging substrate is, in this case, not limiting the range of L. albigena The hypotheses that the observed pattern of C. mitis and L. albigena occurrence is driven by the basal area density of fr uit-producing trees or, in the case of C. mitis leafproducing trees or a combinati on of the two, are supported by these findings. It is not possible to rule out this patte rn being driven by the basal ar ea density of all tree species, but this seems unlikely given that both m onkey species use areas around Kanyawara that lack a closed canopy (the transect 1 forest has a closed canopy). The observed lack of fruiting basal area is probably not due to a lack of fruit dispersers in the fragments (as suggested by Cordeiro and Howe 2001) as large bodied frugivores are still present (Onderdonk and Chapman 2000), and almost certainly not due

PAGE 53

42 to a lack of dispersers in the area of transect one. In this case, it is the lack of fruiting trees that leads to a lack of fr ugivores and not the other way around. There is ample evidence to conclude that the fragment s are poor habitat and do not contain enough food trees to support their use by C. mitis or L. albigena The fact that transect 1 has higher food tree basal areas than the fragments and yet is completely unused by either species though they inhabit th e adjoining forest indicates how poor the fragments really are. This suggests that even if the matrix between the fragments and the forest were conducive to dispersal by C. mitis and L. albigena the fragments would likely be as unused as they are currentl y. The importance of human impacts on the fragments is indicated by Eucalyptus spp. being one of the top ten trees in 8 of the 12 fragments. If native trees important to frugivor es were found to be su itable alternatives to Eucalyptus spp. for human use and planted instead, the habitat value of the fragments might be greatly increased even with no change in extraction practices. Though managing the smaller fragments for C. mitis or L. albigena would be overly optimistic, wildlife species that currently depend on the fragments could be expe cted to benefit from management aimed at increasing nativ e forest trees at the expense of Eucalyptus spp. and edge/savanna trees that are common in the surrounding matrix. Moving management away from heavy use of Eucalyptus spp. for the purpose of improving wildlife habitat would be diffi cult given the long establishment of Eucalyptus spp. use in Uganda, the general focus of ex tension programs on a few exotic species while ignoring natives (Katende et al. 1995), the benefits of Eucalyptus spp. including quick production of above-ground biomass (Harmand et al. 2004) and coppicing ability (Little and Gardner 2003), and the widespread hostility towards wild life because of crop-

PAGE 54

43 raiding (pers. obs.). However, the recogniti on and exploitation of the tourism value of primates and the forest fragments themselves (shown by multiple ecotourism projects in the area based on, or offering access to, the fr agments) (pers. obs.), the common (albeit sometimes reluctantly admitted) knowledge that not all species of wildlife are serious agricultural pests (pers. obs., A. Lepp unpublished data, Hill 1997), and the problems associated with Eucalyptus growth, like high water demands (Radersma and Ong 2004) and allelopathic leaf litter (Bernhard-Reversat 1999), may pr ovide valid motivations for management change in some of the fragments. Fruit Availability With respect to the avoidance of transect 1 by C. mitis and L. albigena even more telling than the forest composition data, is the fruit availability data. Both species are frugivorous with fruit forming the plurality, if not the majority, of their diets. The fact that transect 1 had no im portant trees fruiting for C. mitis during the study and low levels of fruiting by a single common species for L. albigena during only five of the eight months suggests that the unused area does not produce enough fruit to attract either species. Transect 1: A Primate Perspective The area represented by transect 1 is low quality habitat for both C. mitis and L. albigena but this does not automatically exclude incidental use of or travel through the area by either species. To understand the absolute lack of use by troops of these monkeys, it is important to take a deep er look at their natural history. C. mitis groups have been studied at Kanya wara since the early 1970s (Rudran 1978) and identified by consistent numerical labels (Butynski 1990) and inhabit stable territories (although groups 4 and 5 seem to be switched between the 1970s and 1980s).

PAGE 55

44 Even in this study, the areas where C. mitis observations were made match the home ranges given by Butynski (1990). He characterized the Kanyawara C. mitis population as being stable and it would be useful to ponder how important traditional home range boundaries are in determining use. If home ranges are stable for three decades in a relatively stable population, one wonders if th e perturbations of gr oup fission (Cords and Rowell 1986, Lwanga 1987) or marked ecosystem change (logging, fire, etc. ) are perhaps necessary for the exploration and colonization of new areas. It has been noted more than once that C. mitis home ranges tend to have obvious visual landmarks as boundaries, such as exotic plantations, swamps, roads, streams, and, of course, forest edges (Aldrich-Blake 1970, Rudran 1978, Butynski 19 90). It is logical that visually oriented primates, which do not use scent marks to define territories and whose main territorial beha vior (calls) indicates pres ence, not boundaries, might use natural, easily seen habitat changes and features for boundari es. The area around transect 1 is cut off from the rest of the forest by a belt of swamp-marsh. It is this belt that delineates the unused area and is the limit for C. mitis territories in Butynski (1990) and sightings during this study (Figure 24). Thus, from the perspective of a C. mitis group, the area of transect 1 is not simply an ar ea of low food but an area of low food cut off from the rest of the forest by a convenient and distinct boundary. To the collective eyes of C. mitis the area of transect 1 might appear to be a fragment outside (even if contiguous with) the forest, more accessible than the other fragments, but not much more desirable. The case for L. albigena is less conclusive. They commonly use marshes and swamps in other areas (Poulsen et al. 2001) and were seen in the middle of the swamp-

PAGE 56

45 marsh belt during this study in a multispecies group (including C. guereza which visits an open pool there occasionally). On e of their major fruit trees ( Diospyros abyssinica ) did fruit to a certain extent on transect 1 during the st udy, and their most important invertebrate foraging tree has an extremely hi gh basal area density ther e. It is possible that Diospyros abyssinica rarely fruits in the area, rarely did so in the past, or that the fruiting intensity is not worth the trip cons idering how common that fruit was throughout the forest during the synchronous fruiting of Diospyros abyssinica However, given the use of fragments by L. albigena in other areas, the presence of at least some important fruit, and their willingness to use swamps and marshes in this and other areas, L. albigena is probably the more likely of these two frugivor es to incorporate the area of transect 1 into a home range now or in the future as vegetation changes take place. Summary The density of C. mitis was most strongly correlated with the basal area density of total food trees and was also co rrelated with the basal area densities of fruit and leaf providing trees, but it was not correlated with fruit availability, total tree basal area density, or flower tree basal area density. The density of L. albigena was not correlated with the basal area density of any class of trees or with fru it availability. However, number of mangabey groups seen per km walked was marginally co rrelated to fruit availability. When comparing the occupied forest to th e fragments, the fragments were found to have lower basal area densities of all food tree classes except for L. albigena invertebrate foraging trees. The unused area was found to have lower basal area densities than the occupied forest ar eas for all tree classes except for C. mitis flower trees, total L. albigena food trees, and L. albigena invertebrate foraging trees. The unused area had higher basal area densities than the fragments in all classes. The fragments had higher frequencies of e xotic and edge/savanna species than the forest even with the edge transects included. The basal area density of food trees is an adequate habitat quality indicator in the case of C. mitis The failure of basal area density of food trees to account for the

PAGE 57

Figure 2-4: C. mitis territories in relation to the unused area of forest. Territories are outlined in thick lines (redrawn from Butynski, 1990). The unused area of forest is outlined with a white dashed line. Note that it is separated from surrounding territories by a band of swamps Roads are shown with thin lines. The Makerere University Biological Fi eld Station (MUBFS) is labeled. The satellite image is a Quickbird (high reso lution, 2.4 m) from DigitalGlobe, Inc., Longmont, CO, USA.

PAGE 58

47

PAGE 59

48 density of L. albigena may be related to scaling issues. The fragments outside Kibale National Park offer wretched habitat for these two species and dedicated management to incr ease native fruit producing trees, while potentially increasing the value of the frag ments to other species, is unlikely to impact C. mitis or L. albigena due to the size of most of the fragments and the human dominated matrix surrounding them. The unused area had no important fruit for C. mitis produced during the study and low levels of fruiting by a single importan t species during part of the study for L. albigena The other areas had vari able fruiting levels with Celtis durandii ubiquitously fruiting in all months. Because of naturally occurring boundaries around the area of transect 1 and apparently stable home ranges, C. mitis groups probably view this area in the same way as a low quality fragment contiguous with the forest. Because of L. albigena s use of swamps and the presen ce of an important fruit in the unused area, mangabeys are the more likel y species to use the area of transect 1.

PAGE 60

49 CHAPTER 3 SEASONAL VARIATION IN THE QUALITY OF A TROPICAL RIPE FRUIT AND THE EFFECT ON THE DIETS OF THREE FRUGIVORES Introduction Seasonality of climate and the correspondi ng temporally uneven distribution of resources place hardships on wildlife, which can limit populations and act as major forces in natural selection (Beeson 1989, Rich ardson 1991, Brown and Brown 2000, Fiksen 2000, Brugiere et al. 2002). Torpor, migration, and re source switching are some of the common solutions to predictable periods of resource scarcity (van Schaik et al. 1993); however, the effects of lean seasons can be exacerbated by climatic events (Hafner et al. 1994, Muri 1999) or human disturbance (Laurance and Williamson 2001). While the effects of seasonality are more well known in temperate zones and dry tropical zones, the less distinct seasons of tropical moist forest s and the corresponding vegetational changes have been long r ecognized (Janzen 1967, Karr 1976) with much work being conducted describing the cha nging patterns of fruit availability (Sun et al. 1996, Chapman et al. 1999b, Larue et al. 2002, Schaefer and Schmidt 2002, De Walt et al. 2003). Since the nutritional values of speci es differ, the nutritional makeup and the availability patterns of each sp ecies can be combined to get a better idea of what seasons and/or nutrients are potentially lim iting for frugivores (Conklin-Brittain et al. 1998, Rode et al. 2003). While there often appear to be peri ods of fruit or nutrien t scarcity, previous studies have assumed that the nutritional value of any given t ype of food remains unchanged through time (Conklin-Brittain et al. 1998, Gupta and Chivers 1999).

PAGE 61

50 Therefore many months or years of phenology and diet data are often combined with nutritional data from fruit collected once. This is, in fact, an assumption inherent to any study that ignores possible nutritional changes of dietary items over ti me, even those that do not include nutritional analyses. If this assumption of constant nutritional qualities is not a valid one, dietary choices of wildlife may appear random or illogical and the seasonality of tropical mois t forests may be seriously underestimated making the understanding of foraging strate gies much more difficult. The objectives of this study were to e xplore the temporal va riation in dietary quality of a common fruit and investig ate the relationships between frugivore consumption and the availability and quality of the fruit. By focusing on the quality and consumption of a common species I show the value of incorporating the possibility of temporal variation of fruit qual ity in ecological study designs. Methods Phenological data and ripe fruit samp les were collected over 6 months and compared with rainfall (the main seasonal change) in the moist tropical forest of Kibale National Park, Uganda (0 13 0 41 N a nd 30 19 30 32 E). The forest is a mature, mid-altitude, moist, semi-deciduous and evergreen forest with a mean annual rainfall of 1734 mm (1990-2000), a mean daily minimum temperature of 15.5 C, and a mean maximum daily temperature of 23.7 C (Chapman et al. 2003a). There are two rainy seasons: March-May and September-N ovember, with the later having the higher rainfall. One of the more important fruit bearing species in Kibale, in terms of both its abundance and contribution to the diets of frugivores, is Celtis durandii (Struhsaker 1997). The fruit is a small, yellow, ovoid dr upe (~0.8 cm) containing a single hard seed.

PAGE 62

51 C. durandii is an extremely common and prolif ic fruiter in many areas of Kibale (Struhsaker 1997, Barrett and Lowen 1998, Chapman et al. 1999b), and it was the only species that was constantly fruiting throughout the study (May-November, 2002). Though it has annual fruiting peaks, fruit is usua lly available to some extent in many, if not all, months of any given year (Struhs aker 1997), making it a potentially important staple or fallback food in times of general fr uit scarcity. The fru it is eaten by many birds (Struhsaker 1997) and is important in the diets of Chimpanzees ( Pan troglodytes ) (Ghiglieri 1984), Blue monkeys ( Cercopithecus mitis ) (Rudran 1978), Redtail monkeys ( Cercopithecus ascanius ) (Stickler 2004), Gray -cheeked mangabeys ( Lophocebus albigena ) (Olupot 1994), Black-and-white colobus monkeys ( Colobus guereza ) (Struhsaker 1978), Red colobus monkeys ( Piliocolobus tephrosceles ) (Struhsaker 1975) and probably many other animals. Ripe C. durandii fruit was collected each month for 6 months (June-November, 2002). A fruit was defined as ripe if it had changed from green to at least partially yellow. The fruit was peeled off the seed, dr ied in a dehydrator (35 C), and stored in plastic bags for shipment back to the Universi ty of Florida. The fruit was redried in a drying oven at 50C overnight before being ground in a Wiley mill. The percent lipid content of the fruit was determined with a microwave assisted extr action technique. The dried, ground samples were weighed, placed in petroleum ether, and exposed to microwaves in a MARS 5 machine (CEM Co rporation, Matthews, NC) to subject the samples to extreme heat and pressure, allowing the lipids to be extracted quickly and efficiently. The samples were rinsed in pe troleum ether, dried, reweighed and the lipid content calculated from the difference. The data for the highest lipid percentages are

PAGE 63

52 probably underestimations as the more oily samples lost oil on the drying paper and sample labels during storage and processing. Samples were also analyzed for total ethanol soluble carbohydrates, sa ponins, acid detergent fiber, protein, and alkaloids, but none of these showed an independent seasonal pattern so they are not discussed further. Each month, phenological data (presence of ripe and unripe fruit, fruiting intensity, and diameter at breast height (dbh) of fruiting trees) were collected, along 20m wide fixed strip transects through the forest. The fruiting intensity was determined by subjectively assigning each tree to a numerical class (1-5) base d on the fruit density in the crown with each class indicati ng about twice the density as the next lower class and 3 being the average fruit density of the species Because dbh is strongly correlated with crown size (Anderson et al. 2000) and fruit production (Chapman et al. 1992) within a species, the basal area of each tree was multiplied by a factor (, 1, 2, or 4, respectively) based on the fruit density class to estimate the total fruit load on each tree. The fruit loads were summed (trees having both ripe and unripe fruits were considered to have ripe and unripe fr uit) and divided by the tota l area sampled to produce a standardized monthly fruiting intens ity for both ripe and unripe fruit. The average lipid levels in each month we re compared to the sum of the average daily rainfalls of the concurrent and previous months (the best pr edictor months) using linear regression. To see if the fruit lipid level is related to the intake of C. durandii fruit by frugivores, monthly dietary information was gleaned from existing sources (1 C. mitis 1 C. ascanius and 2 L. albigena data sets were available). The C. ascanius data set (C. Stickler, unpub. data) was from the same lo cation and time as the phenology and lipid

PAGE 64

53 data and could be directly compared; however, the C. mitis data set was from 1973-4 (Rudran 1978) and the L. albigena data sets were from 1972-3 (Waser 1975) and 1992-3 (Olupot 1994) and could only be compared with lipid levels i ndirectly through the summed average daily rainfall of the c oncurrent and previous months. Rudran (1978), Waser (1975), and Olupot ( 1994) did not different iate between ripe and unripe fruit, but that should not have a ma jor effect on the results as the lipid content of the two fruit types is highly correlated (r=0.99, p < 0.001) and show s a similar pattern of change as the ripe fruit with differences between ripe and unrip e fruits of the same month smaller than the seasonal differe nces. Wasers (1975) reported monthly percentages of C. durandii in the diet include both fruit and insects. However, the use of insects from C. durandii is negligible compared to the range of the fru it. Additionally, Olupot (1994) only provides m onthly dietary data from the top 5 foods. For 5 of 9 months, C. durandii is not in the top 5. Instead of assuming C. durandii fruit was not eaten at all, a more conservative ta ck was taken and it was assumed that C. durandii fruit was a close 6th in every such month ( i.e. the fifth ranks percentage minus 0.01%). This reduced the correlation coefficient and increase d the p-value slightly compared with the assumption of no consumption during months when C. durandii was not in the top 5. Availability indices for C. durandii fruit from each study were also compared with the importance of the fruit in each diet, except for Olupot (1994) where there was none given and in the case of C. ascanius for which the average standardized fruiting intensity of ripe C. durandii fruit for the same area as the C. ascanius study was used. All data sets were analyzed with Pearson s correlations except for that of C. mitis which violated

PAGE 65

54 the assumption of normality and therefore was analyzed using Spearmans ranked correlations. Results The lipid levels in ripe Celtis durandii varied over two orders of magnitude among months (0.3%-30.8%) and are predicted by th e summed average daily rainfall of the previous and concurrent months (r2 = 0.91, p = 0.003, N = 6) (Figure 3-1). 0% 5% 10% 15% 20% 25% 30% 35% 40% 024681012141618Rainfall (mm)Lipid content Figure 3-1: The relationship between the monthly dry matter lipid content of ripe Celtis durandii fruit and the summed average daily rainfalls of the concurrent and previous months at the Makerere University Biological Field Station, Kibale National Park, Uganda for June 2002-November 2002 (r2=0.91, p=0.003, N = 6). The error bars represent +/one standard error. C. ascanius monthly consumption of ripe C. durandii fruit at the same time and place is marginally correlated with the fruit lipid content (r = 0.88, p = 0.059, N = 4, onetailed) (Figure 3-2 a), and is also ma rginally correlated with rainfall (r = 0.78, p = 0.061, n = 5, one-tailed) (Figure 3-2 b).

PAGE 66

55 0 5 10 15 20 25 30 35 40 45 5%15%25%35% Lipid content of ripe fruit% of redtail dieta 0 5 10 15 20 25 30 35 40 45 81012141618 Rainfall (mm)% of redtail dietb 0 5 10 15 20 25 30 0510152025 Rainfall (mm)% blue monkey dietc 0 5 10 15 20 25 30 35 40 45 051015 Rainfall (mm)% of mangabey dietd 0 10 20 30 40 50 60 0102030 Rainfall (mm)% mangabey diet (max)e 0 5 10 15 20 25 30 35 40 45 0510152025 Availability (0-40)% of mangabey dietf Figure 3-2: The relationships between reported dietary use of Celtis durandii fruit and lipid content of the fruit, average daily rainfall of the previous and concurrent months (a predictor of lipid content), and fruit availability. All data sets are from Makerere University Biological Field Station, Kibale National Park, Uganda. a) the relationship between lipid content and Cercopithecus ascanius diet (C. Stickler, unpub. data) for August 2002-November 2002 (r = 0.88, p = 0.059, N = 4, one-tailed). b) the relationship between rainfall and C. ascanius diet (C. Stickler unpub. data) fro m August 2002 December 2002 (r = 0.78, p = 0.061, N = 5, one-tailed). c) the relationship between rainfall and Cercopithecus mitis diet (Rudran 1978) from February 1973 May 1974 (rs = 0.56, p = 0.013, N = 16, one-tailed test). d) the relationship between rainfall and Lophocebus albigena diet (Waser 1975) from May 1972 April 1973 (r = 0.64, p = 0.009, N = 13, one-taile d test). e) the relationship between rainfall and L. albigena diet (Olupot 1994) from October 1992 June 1993 (r = 0.63, p = 0.034, N = 9, one-tailed test). f) the relationship between fruit availability and L. albigena diet (Waser 1975) from May 1972-April 1973 (r = -0.61, p = 0.035, N = 12, two-tailed test).

PAGE 67

56 The C. mitis monthly dietary percentage of C. durandii fruit described by Rudran (1978) shows a correlation with rainfall (rs = 0.56, p = 0.013, N = 16, one-tailed test) (Figure 3-2 c). Wasers (1975) L. albigena data shows a correlation between the monthly proportion of C. durandii fruit in the diet and rain fall (r = 0.64, p = 0.009, N = 13, onetailed test) (Figure 3-2 d), as does Olupot s (1994) data (r = 0.63, p = 0.034, N = 9, onetailed test) (Figure 3-2 e). Neither the C. ascanius nor the C. mitis consumptions of C. durandii fruit are related to fruit availability. Wasers (1975) L. albigena data set has a negative correlation between availability and consumption (r = -0.61, p = 0.035, N = 12, two-tailed test) (Figure 3-2 f) with all species seemed to s how decreases in availa bility occurring at the same time as the highest consump tion (Figures 3-3, 3-4, and 3-5). Discussion While nutritional components of leaves and fruits are known to vary between species, location, position on th e tree, stage of developm ent, and/or time of day (Woodwell 1974, Marquis et al. 1997, Fernandez-Escobar et al. 1999, Nergiz and Engez 2000, Klages et al. 2001, Chapman et al. 2003a), and yearly fru it production and quality have been shown to be dependent on envi ronmental conditions of earlier months (temperature, rainfall, amount of sunli ght, etc.) (Sams 1999, Woolf and Ferguson 2000), to my knowledge this is the first time nutriti onal quality of a conti nuously available ripe fruit has been shown to be seasonally variab le and related to rainfall. These results suggest that the seasonality of tropical moist forests could be more marked than previously thought. Though ripe C. durandii fruit was constantly available during the

PAGE 68

57 study and at relatively high levels throughout th e forest, the lipid resource base provided by the fruit was negligible in the dry season but abundant in the rainy season. Resource availability estimates based sole ly on observed fruit loads and nutritional analyses of samples taken at one time may be misleading if seasonal changes in nutritional values commonly occu r. The ripe fruits of C. durandii in the wet and dry seasons appear almost identical to human obser vers, only being plumper and juicier in the 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 MayJuneJulyAug.Sept.Oct.Nov.Dec.Standardized fruiting intensity (cm^2/m^2)0% 5% 10% 15% 20% 25% 30% 35% 40% 45%% of diet / % lipid Ripe fruit availability % of Redtail diet % lipid Figure 3-3: Changes in the percentage of Celtis durandii ripe fruit in the diet of a Cercopithecus ascanius group relative to changes in fruit availability and fruit lipid content. All data sets were from the forest around Makerere University Biological Field station, Kibale National Park, Uganda, 2002.

PAGE 69

58 0.0 0.5 1.0 1.5 2.0 2.5Jan-73 Feb-73 Mar-73 Apr-73 May-73 Jun-73 Jul-73 Aug-73 Sep-73 Oct-73 Nov-73 Dec-73 Jan-74 Feb-74 Mar-74 Apr-74 May-74 Fruit availability index0% 10% 20% 30% 40% 50% 60%% of diet / % lipid Fruit availability index % of blue monkey diet Estimated % lipid Figure 3-4: Changes in the percentage of Celtis durandii fruit in the diet of a Cercopithecus mitis group relative to changes in fruit availability and estimated fruit lipid content. The lipid content was calculated from rainfall records. The peaks ar e probably unrealistically high but the basic pattern should reflect reality. The diet data were from Rudran (1978), Makerere University Biological Field Station, Kibale National Park, Uganda, January 1973 May 1974. wet season. Despite this, the observed variati on of lipid content is apparently important, as indicated by foraging effort, to at least three frugivorous primates. During dry months when the lipid levels were negligible, C. durandii fruit was rarely eaten even when most abundant, but was heavily consumed when lipid levels increased. This may also be important to birds as they can distinguish and prefer lipid content as little as 2% higher (Schaefer et al. 2003). The fact that the data sets span four decades suggests that the pattern holds true over long periods of time. Rudran (1978) also notes that another of his groups (for which he did not present de tailed data) had a si milar pattern of C. durandii fruit consumption as the group whose data are used here.

PAGE 70

59 0 5 10 15 20 25 Apr72 May72 Jun72 Jul72 Aug72 Sep72 Oct72 Nov72 Dec72 Jan73 Feb73 Mar73 Apr73Availability index (0-40)0 5 10 15 20 25 30 35 40 45% in diet / % lipid Availability % of mangabey diet Calculated % lipid Figure 3-5: Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group relative to changes in fruit availability and estimated fruit lipid content. The lipid content was calculated from rainfall records. The diet data are from Waser (1975), Makerere University Biological Field Station, Kibale National Park, Uganda, May 1972 April 1973. Though the relationships were assumed to be linear for the purpos es of analysis, the true relationships between lipid content and rainfall, lipid content and consumption, and consumption and rainfall are probably best represented by logistic functions. The minimum lipid content of the fruit is limite d both by 0 and by some low level needed for cellular function, while the maximum should al so be limited by physiologic mechanisms. Consumption of the fruit is likewise li mited by 0 as a minimum and by nutritional demands for substances not f ound at high enough levels in C. durandii fruit as a maximum. Even though it is eas y to visualize S-shaped curves as appropriate for some of the data sets, the small sample sizes, especi ally at the higher end of the ranges, make fitting reasonable logistic curves questionable.

PAGE 71

60 None of the other nutrit ional components in the C. durandii fruit or the fruit of any other of the 9 species analyzed showed a sign ificant relationship with rainfall. During this study, other species flushed with fruit ei ther synchronously or independently for one to four months and tended to show slight increases in total ethanol soluble carbohydrates (sugars) as the crops ripened ( i.e. what was collected as ripe when the first fruits started to ripen was not as rip e as the fruits at the end of the fruiting interval). 0% 10% 20% 30% 40% 50% 60% Oct-92Nov-92Dec-92Jan-93Feb-93Mar-93Apr-93May-93Jun-93% Diet(max) / % lipids (calculated) % Mangabey diet Calculated % lipids C.d. dietary maximum Figure 3-6: Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group relative to changes in estimated fruit lipid content. The lipid content was calculated from rainfall records. The diet data were from Olupot (1994), Makerere University Biological Field Station, Kibale National Park, Uganda, October 1992 June 1993. Since only the top 5 foods were reported, the percent of the fruit in the diet from February on was assumed to be the maximum possible (see text for full explanation). As 1992 was a very wet year, the lipid content of the fruit is probably estimated at impossibly high levels. However, this phenomenon was probably not the case for C. durandii as 1) it fruited throughout the study with different trees coming into and going out of fruit at different times, 2) any individual fruit probably does not stay ripe fo r more than a month, 3) the unripe fruit showed a similar increase in lipi d levels with rainfall, and 4) the amount of

PAGE 72

61 variation in lipid content obs erved is not only over two orde rs of magnitude but greater than the variation between ripe and unripe fruits of the same month. This study did not cover the minor rai ny season in March-May. Lipid levels estimated with rainfall had two peaks a y ear, but consumption by the primates only showed one major peak in a year (Figures 3-4, 3-5, and 3-6). If consumption is a good indicator of lipid levels in general, it appear s that lipid peaks may not be directly driven by moisture availability, but instead may o ccur only once a year during the September to November rains. The negative correlation between fruit ava ilability and monthly dietary importance for Wasers (1975) data and th e observation that the months of high usage occur at the same time as drops in availability (Figures 3-3, 3-4, and 3-5) suggest that the availability of the fruit may be driven by consumption and not the other way around as is usually assumed. This may explain the observation that C. durandii has irregular flowering, but annual fruiting peaks (Chapman et al. 1999b). Instead of annual fruit production peaks, C. durandii may have consumer driven annual fruit load lows ultimately caused by changes in lipid content. If this is true, it implies that competition for this resource is intense for only a few months out of the year and is driven ultimately by resource quality. Celtis durandii seems to have a strategy of producing fruit year round, even in the difficult seasons, but altering the value of the reward provided to seed dispersers. This strategy could be relying on several principles: i) either it tricks certain dispersers into eating a fruit with low nutritional value, ii) is eaten and dispersed onl y as a last resort by desperate frugivores when nothi ng else is available, or iii ) is ignored by all frugivores during the off-season and is dispersed by gr avity or seed predators. Alternately, C.

PAGE 73

62 durandii could be able to produce lipid rich frui ts in all seasons, but the chances of seedling survival might be low in drier season s, thus the investment in more attractive fruits at a time when seed dispersal could be mo st beneficial. In any case, it is clear that the abundance of this fruit is not related to its value as f ood, and resource availability studies based on an assumption of the constant quality of specific foods may be missing an important component. Primate diets are often diverse and variab le with seemingly random changes that appear inexplicable to the human researcher. However, the use of C. durandii fruit by at least three species can be explained by the ch anging nutrient levels found in the fruit and not by the changing abundance levels of the fruit. If this strategy of producing low quality fruits during certain seasons or under certain conditio ns is shared by a number of tropical tree species, it could go a long way towards explaining changing and enigmatic dietary preferences in frugi vores and be another indicati on that the tropics are more seasonal than originally thought.

PAGE 74

63 CHAPTER 4 CONCLUSION The conservation of wildlife populations requi res the ability to assess the quality of remaining habitats, but it is often difficult to determine habitat requirements in species that have temporally and spatially variable diets. An understanding of how edges and fragmentation affects habitat quality is incr easingly important in these days of wanton habitat destruction and fragmentation. This study of C. mits and L. albigena was undertaken to understand how habitat quality for these two species might be assessed efficiently and evaluate the habitat potentia l of unoccupied forest fragments near the intact forest. This study showed that: The density of C. mitis was most strongly correlated with the basal area density of total food trees. The number of mangabey groups seen per km walked was marginally correlated to fruit availability. The fragments were found to have lower ba sal area densities of all food tree classes than the occupied forest, except for L. albigena invertebrate fora ging trees. The unused area within the park was found to have lower basal area densities than the occupied forest areas for all tree classes except for C. mitis flower trees, total L. albigena food trees, and L. albigena invertebrate foraging trees. The unused area had higher basal area densities than the fragments in all classes. The fragments had higher frequencies of e xotic and edge/savanna species than the forest even with the edge transects included. The basal area density of food trees is an adequate habitat quality indicator in the case of C. mitis The failure of basal area density of food trees to significantly predict the density of L. albigena may be related to scale differences between the study and home ranges of mangabeys or foraging strategy.

PAGE 75

64 The fragments outside Kibale National Park offer extremely poor habitat for these two species and dedicated management to increase native fr uit producing trees, while potentially increasing the value of th e fragments to other species, is unlikely to impact C. mitis or L. albigena due to the size of most of the fragments and the human dominated matrix surrounding them. The unused area had no important fruit for C. mitis produced during the study and low levels of fruiting by a single importan t species during part of the study for L. albigena The other areas had vari able fruiting levels with Celtis durandii ubiquitously fruiting in all months. Because of naturally occurring boundaries around the area of transect 1 and apparently stable home ranges, C. mitis groups probably view this area as a low quality fragment contiguous with the forest. Because of L. albigena s use of swamps and the presen ce of an important fruit in the unused area, mangabeys may be more likely than C. mitis to use the area of transect 1. Many studies have assumed that the nutrien t content of food items remains constant over time. Here, I documented that the lipid content of Celtis durandii ripe fruit varied tremendously during the study and was correlated with rainfall. The estimated variation of lipid content was shown to be correlated with the dietary intake of C. durandii fruit in three primate species with four data sets. C. durandii fruit availability was a poor predic tor of dietary use and seemed to be driven by consumption which seemed in tu rn to be driven by lipid content. Therefore common use of use of fruit av ailability as an index of resource availability should be more carefully scrutinized. This study has shown the validity of asse ssing habitat quality by quantifying the basal area densities of locally impor tant food trees for the generalist C. mitis This method quickly and easily meas ures a slowly changing ( sans logging) aspect of the environment and is therefore more efficient th an evaluating habitat quality with indices of fruit availability which may change quickly and seasonally and be onl y loosely related to the availability of resources in the habitat. Basal area density was a worse predictor of habitat use by L. albigena than fruit availability, perhap s because of the mismatch of

PAGE 76

65 scales between home ranges and this study, or perhaps because these mangabeys specialize on intense fruiting events. The fragments outside Kibale National Park have extremely low habitat quality for these species and would therefor e be of little use to these monkeys either as habitat or stepping stones connecting intact forest popul ations even if the matrix surrounding them were conducive to dispersal. The unused area, though contiguous w ith used forest and higher in habitat quality th at the fragments, is probably avoided because of the combination of population stability, conveni ent boundaries surroundi ng that area, and poor habitat quality. It is likely that this area will be used by L. albigena before C. mitis The variation in the lipid content of C. durandii and the observation that lipid content, and not availability, seems to drive consumption of this fruit indicates that seasonality of resources might be more pronounced in tropical forests depending on the strategies of component fruiting species. Di etary studies should be designed to account for this possibility and assess resource avai lability and not just fruit availability.

PAGE 77

66 APPENDIX POTENTIAL FACTORS INFLUE NCING HABITAT USE BY C ercopithecus mitis AND Lophocebus albigena With the complete avoidance of th e area of transect 1 (an edge) by Cercopithecus mitis and Lophocebus albigena in mind, many possible reasons for differential habitat use especially with regards to edges were gene rated. Logistical considerations limited the number of factors that could be examined thus many possibly important edge effects were excluded from consideration, such as: Light levels Wind Disease Human impacts Predation Arthropod abundance However, many other factors were examined: Temperature extremes Evaporative potential Tree species composition Fruit availability Nutritional quality (fruits and leaves) Arthropod foraging substrate availability Unexamined Possibilities Light Levels, Wind and Disease Light, wind, and botanical diseases would affect the m onkeys indirectly by either altering the composition of the fore st or the availability or quality of foods. However, it is unlikely that these are important in this case as increased light levels and wind are found on edges and open forest types and would therefore have a similar impact at least in the immediate study area a nd exclude the monkeys from other areas, which is not

PAGE 78

67 apparent.. A simian disease ha s been thought to be able to control the di stribution of C. mitis in other parts of Uganda (Haddow et al. 1947), but in that case there was an extreme ecological gradient that contro lled the distribu tion of the disease vect or which is not true here. Also, a disease would most likely have a larger or more temporary effect than that observed. Therefore, thes e potential factors are not considered further. Human Impacts Since the edge in question bor ders on a field station, certain human impacts in the area are relatively high. Even without dir ect persecution, some extremely sensitive species may shy away from constant human presence. However, both C. mitis and L. albigena groups in the area have pr eviously been fully or part ially habituated (depending on the group) and generally appe ared less excited when approached than some of the other more heavily studied species (e.g., Piliocolobus tephrosceles ) and are present in rest of the forest area that is heavily used by both researchers and loggers. Additionally, C. mitis are known to be pests in pine plantations and maize fields in other parts of Africa (Devos and Omar 1971, Maganga and Wri ght 1991), and mangabeys are found in Magombe Swamp (bordering Kibale on the east) that is surrounded by homes and small scale cultivation (pers. observ ation). It is thus unlikely that human presence alone is disturbing these monkeys enough to cau se them to avoid this edge. Hunting could extirpate these species from an area, but is unlikely to have done so in this case. Firstly, the area is so small it is doubtful that hunters would focus their efforts there and ignore nearby forest areas. Secondly, the surrounding tribe, the Batoro, do not eat monkeys, their use of monkey skins, for royal regalia ( Colobus guerza ) and dancing (possibly baboon ( Papio anubis ) but not C. mitis or L. albigena ), is limited (pers. observation; (Haddow 1952), and Rutoro lacks a specific name for C. mitis indicating

PAGE 79

68 limited contact and cultural importance (Onderdonk and Chapman 2000, S. Katusabe pers. com). Kingdons (1974) reported name of C. mitis in Runyoro, a closely related dialect, nkima simply means monkey in Rutoro (S Katusabe pers. com.). Onderdonk and Chapman (2000) also state that there is no Rutoro name for L. albigena but it has been reported to me as jugujugu an onomatopoeia mimicking the mangabey whoopgobble (S. Katusabe pers. comm.). However, this word may be unknown or unused except in places where cont act is common (e.g., Magombe Swamp to the south). A nearby people, the Bakonzo, who do re gularly hunt monkeys, including C. mitis (pers. observation; Haddow 1952), have been accused of poaching with firearms and possibly lowering primate populations in Kibale befo re being expelled by the Batoro in 1962 (or 1964) (Struhsaker 1975, Ghiglieri 19 84). However the only evidence of this seems to be local stories (Ghiglieri cites he arsay and Struhsaker cites no one) It is true that starting in 1962 intertribal tension boiled over into violence, arson (Ssembeguya et al. 1962), and eventually a rebellion that lasted for d ecades in the Rwenzori Mountains (Ingham 1975, Syahuka-Muhindo 1991), which could easily have ejected the Bakonzo from the Kibale area (~10% of the population of the subcounty Rutete was Bakonzo in 1959 (Lubowa et al. 1963)). However, the presence of hunters do es not automatically cause low densities and extirpation (specifically of C. mitis ). For example, the cu ltural stronghold of the Bakonzo, the Rwenzori Mountains, still supports its fu ll compliment of primates in spite of their continued exploitation (Lwanga 1987) Finally, there has been no reported hunting in over 40 years so it is highly unlikel y that hunting is responsible for the current absence of C. mitis and L. albigena in this area or paucity of C. mitis in the southern region of the park.

PAGE 80

69 Predation Predation can be important in controlling the abundance and range of some species in certain areas. In Kibale, the main predat or of monkeys is the African Crowned Eagle ( Stephanoaetus coronatus ) (Mitani et al. 2001) with chimpanzees focusing on Red colobus (Mitani and Watts 2001). Skorupa (1989) found that both C. mitis and L. albigena made up 9% of the observed prey items of S. coronatus and C. guereza was the most common at 39%. Struhsaker and Leakey (1990) found that C. guereza and L. albigena were killed more than expected and ad ult males of all monkeys were preferred much more than other age/sex classes. They concluded that eagle predation has little impact on the monkey populations at Kanyawara, as did Rudran (1978). If the presence of an edge increases the area in the canopy vulnerable to eagle attack, monkeys might choose to avoid the ed ge. However, they do not avoid just the trees on the edge of the forest but also the trees surrounded by other trees in that area. Additionally, the Mikana area just to the north of Kanyawa ra was heavily logged and still retains an extremely broken and uneven canopy with most of the large trees well exposed on all sides. This canopy morphology should expose primates to eagle predation more than a closed canopy or edge but the area is frequently used by C. mitis and L. albigena (pers. observation). Though C. ascanius C. guereza and P. tephrosceles frequent the field station, are often in exposed trees, and surrounded by potential observers, eagles have not been reported hunting there. This may be because they are deterred by the presence of humans, making the field station potentially safer th an the rest of the forest. Because of these reasons predation was given no further consideration.

PAGE 81

70 Arthropod Abundance Since arthropods form a large proportion of the diet of both C. mitis and L. albigena any discussion of habita t suitability must address the availability of this important resource. Unfortunately, due to the small size of most arthropods, the speed at which they are ingested, the distance from the observers, and the immense diversity of arthropods in tropical forests, the ingested species have ne ver been identified for these monkeys. In fact, even the ge neral type of arthropod is extr emely rarely identified by direct visual means and identif ication of food items in fecal material is often difficult, leaving inference from fora ging style as the main method for identifying the general arthropod resource utilized (Was er 1977, Struhsaker 1978, Poulsen et al. 2001). C. mitis tends to use slower capture me thods than the closely related Cercopithecus ascanius which are often seen stalking and pouncing on prey, implying that C. mitis more commonly target slow or stationary arth ropods (Cords 1986). They also forage on and in specific substrates (foliage and epiphyt es especially layers of moss and lichen) in specific tree species (Rudran 1978, Struhsaker 1978, Cords 1986). Similarly, Gray-cheeked mangabeys use fora ging tactics appropriate for the capture of slow or stationary prey (Waser 1984), probably because their size makes large and quick movements dangerous in the canopy (Struhsaker 1978). However, because of their size, they are also able to break apart dead wood and rip off large sections of dead bark (Struhsaker 1978), exploiting a prey base that is not available to the smaller monkeys. Since the specific arthropod items in the diets of C. mitis and L. albigena are unknown and the difficulties involved in syst ematically collecting arthropods from the correct foraging substrates are prohibitive, indices of available foraging substrates (epiphytes and dead wood) we re used as substitutes.

PAGE 82

71 Examined Possibilities Temperature Extremes and Evaporative Potential It is well known that forest edges have more extreme temperatures and are more xeric than the interior (Burke and Nol 1998, Didham and Lawton 1999, Gehlhausen et al. 2000). These conditions could pot entially affect the distribu tion of primates directly through their own tolerances or indirectly through the toleran ces of or indirect impacts on their food, predators, competitors, or di sease organisms (Murcia 1995). Kingdon (1974) notes that captive C. mitis appear uncomfortable in hot sun and in the wild will often abandon a sunny canopy as the morning warms up (also Aldrich-Blake 1970). Furthermore, Kingdon (1974) mentions that L. albigena does not tolerate cold temperatures in captivity. This means that temperature extremes near an edge could affect both species, albeit in opposite ways. However, if the edge were directly affecting the monkeys through temperature, they would be expected to only avoid the edge during exceptionally hot or cold periods and their pa tterns of use should change diurnally and seasonally, which they do not. Temperature and evaporative potential mon itoring stations were set up on two edge to interior transects (one for each edge-inter ior set). A station consisted of a min/max thermometer with the wire temperature probe ti ed to the evaporative potential monitor. Evaporative potential monitors followed Didham and Lawton (1999), but were altered by having reduced 1cm2 wick areas above the tubes and rain covers because they were left out continuously and checked every other da y. It was discovered that using natural broom straws (palm midribs) to support the wicks of ten evaporat ive potential monitors in the same location caused excessive variab ility in measurements (mean CV: 0.337), so plastic was used (mean CV: 0.101) (Figure A-1).

PAGE 83

72 0 5 10 15 20 256/21 / 02 6 /28/02 7/5/02 7/1 2/ 02 7/1 9/ 02 7 /26/02 8/2/02 8/ 9/ 0 2 8/1 6/ 02 8/2 3/ 02 8 /30/02 9/6/02 9/ 13/ 02 9/2 0/ 02 9/27/02 1 0/4/02Datemm H2O Figure A-1: The intraand inter-measurement period variation of 10 evaporation potential monitors from the same location. The dotted line indicates wh en the wicks were first changed. The solid line indicates when the wicks were changed an d the wick supports were changed from broom straws to plastic. The dashed lines indicate when the wicks were changed and the water was changed instead of just being topped off. Note the high variation while using broom straws or water that had been used with the broom stra ws and the reduction of variation once plastic supports were used with regular water changes. Considering the possible indirect affect s of both temperature and evaporative potential (as they are tightly interrelated), the expectation that th e unused edge varies substantially from both the utili zed interior and the utilized edge is not met. In fact topography was found to have a large effect, w ith a forested hilltop being warmer than and as dry as the edge and a valley being mois ter than the interior. Therefore, while the microclimate of forest edges may have indirect impacts on C. mitis and L. albigena in this case microclimate edge effects are not cons idered to play an important direct role in the observed habitat use patterns and are not discussed further.

PAGE 84

73 Forest Composition and Fruit Availability Forest species composition and fruit avai lability were judged to be the most important factors producing the observed habi tat usage patterns and are discussed in detail in Chapter 2. Nutritional Quality of Vegetable Foods Nutrient availability and qua lity of food has been shown to be important in primate distribution and population densities (Chapman et al. 2002a, Wasserman and Chapman 2003). Food items could exist in the unused ar ea in the same amounts as in the used areas but if the nutritional quality is low this might explain the dearth of frugivorous monkeys in the unused area. The nutritional components examined in this study can be divided into two categories: those presumed desirable to th e monkeys (lipids, protein, and sugar) and those presumed undesirabl e (alkaloids, fiber, and saponins). Fruit and leaves were collected each m onth for 6 months (June, 2002-November, 2002). Non-fig fruits were define d as ripe if it had at least partially changed to the ripe color. The ripeness of figs was determined by smell and touch. Non-fig fruits were peeled off the seed and dried in a dehydrator before being stored in plastic bags for shipment back to the University of Florida. Figs were quartered before drying with no attempt to remove the seeds. Preliminary analysis showed that there were few strong patterns (detailed in Chapter 3) and none that supporte d the nutritional components hypothesis (i.e. the unused area had less of a desirable component or more of an undesirable component than the other areas). Arthropod Foraging Substrate Once at the end of the March-May rainy se ason and once at the end of the JuneAugust dry season, all transects were surveyed for the extent of moss/epiphyte foraging

PAGE 85

74 substrate available to the monkeys. Every tr ee belonging to an impor tant insect foraging substrate species (see Chapter 2 for the defini tion of important food species) within 10 m of the transects and larger than 5 cm dbh was measured (dbh) and placed into one of four moss/epiphyte coverage categories (0-24% 25-49%, 50-74%, or 75-100%). Only the parts of the tree above 5 m were taken into c onsideration. If the ep iphytes were large and thick (cabbage-like) and provided more substr ate than their covera ge of the trunk would imply, their contribution to the overall covera ge was estimated by considering what it would be if the leaves were spread out, e dge to edge, on the surface of the trunk. Lichen was considered only if foliose or fruticose and formed moss-like mats on the tree (crustose growth forms, while extremely co mmon, were deemed to be as poor foraging substrate as the bark they so resemble). The basal area for each species and coverage category for each transect was added to produce an index of foraging substrate availability. Because the unused area did not have no ticeably lower levels of epiphyte cover than the other transects, as would be e xpected if arthropod foraging substrate were driving habitat use, it was deci ded that further analysis s hould focus on basal area and not epiphyte cover of arthropod trees. At the same time the moss/epiphyte surv eys were being done, all transects were surveyed for dead wood foraging substrate av ailable to the monkeys All observed dead wood above 5 m in height on important invert ebrate foraging species was placed into a size category: twig (~2 cm diameter), leafy twig, branch (~10 cm), limb (~20 cm), or trunk (~30 cm or greater) and the total end to end length of wood in each category was estimated for each tree. Unfortunately, it was im possible to include dead bark in the dead

PAGE 86

75 wood index. This weakness, plus my genera l impression that the index was extremely sensitive to visibility th rough the canopy and position of the observer, led to the conclusion that this index is a poor a nd unreliable one and should not be used.

PAGE 87

76 LITERATURE CITED Aldrich-Blake, F. P. G. 1970. The ecol ogy and behavior of the Blue monkey Cercopithecus mitis stuhlmanni PhD. dissertation University of Bristol, Bristol. Anderson, S. C., J. A. Kupfer, R. R. Wilson, and R. J. Cooper. 2000. Estimating forest crown area removed by selection cutting: a linked regression-GI S approach based on stump diameters. Forest Ecology and Management 137 :171-177. Barrett, L., and C. B. Lowen. 1998. Random wa lks and the gas model: spacing behaviour of Grey-cheeked Mangabeys. Functional Ecology 12 :857-865. Beeson, M. 1989. Seasonal dietary stress in a forest monkey ( Cercopithecus mitis ). Oecologia 78 :565-570. Bernhard-Reversat, F. 1999. The leaching of Eucalyptus hybrids and Acacia auriculiformis leaf litter: laboratory experi ments on early decomposition and ecological implications in congolese tr ee plantations. Applied Soil Ecology 12 :251261. Brashares, J. S. 2003. Ecological, behavior al, and life-history correlates of mammal extinctions in West Afri ca. Conservation Biology 17 :733-743. Brooks, T. M., S. L. Pimm, and J. O. O yugi. 1999. Time lag between deforestation and bird extinction in tropical fore st fragments. Conservation Biology 13 :1140-1150. Brown, C. R., and M. B. Brown. 2000. Weathe r-mediated natural selection on arrival time in cliff swallows ( Petrochelidon pyrrhonota ). Behavioral Ecology and Sociobiology 47 :339-345. Brugiere, D., J. P. Gautier, A. Moungazi, and A. Gautier-Hion. 2002. Primate diet and biomass in relation to vegetation com position and fruiting phenology in a rain forest in Gabon. Internati onal Journal of Primatology 23 :999-1024. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance Sampling: Estimating abundance of biol ogical populations. Chapman and Hall, London. Burke, D. M., and E. Nol. 1998. Edge and fragment size effects on the vegetation of deciduous forests on Ontario, Canada. Natural Areas Journal 18 :45-53.

PAGE 88

77 Butynski, T. M. 1990. Comparative Ecology of Blue Monkeys (Cercopithecus-Mitis) in High-Density and Low-Density Sub populations. Ecological Monographs 60 :1-26. Chapman, C. A., S. R. Balcomb, T. R. Gillesp ie, J. P. Skorupa, and T. T. Struhsaker. 2000. Long-term effects of logging on Afri can primate communities: a 28year comparison from Kibale National Pa rk, Uganda. Conservation Biology 14 :207-217. Chapman, C. A., L. J. Chapman, K. A. Bjorndal, and D. A. Onderdonk. 2002a. Application of protein to fiber ratios to predict col obine abundance on different spatial scales. Internationa l Journal of Primatology 23 :283-310. Chapman, C. A., L. J. Chapman, M. Cords, J. M. Gathua, A. Gautier-Hion, J. E. Lambert, K. Rode, C. E. G. Tutin, and L. J. T. White. 2002b. Variation in the diets of Cercopithecus species: Differences within forests, among forests, and across species. in M. E. Glenn and M. Cords, ed itors. The Guenons: Diversity and Adaptation in African Monkeys. Kluwer Academic/Plenum Publishers, New York. Chapman, C. A., L. J. Chapman, K. Rode, E. M. Hauck, and L. R. McDowell. 2003a. Variation in the nutritional value of primate foods: Among trees, time periods, and areas. International J ournal of Primatology 24 :317-333. Chapman, C. A., L. J. Chapman, R. W. Wrangham, K. D. Hunt, D. Gebo, and L. Gardner. 1992. Estimators of fruit ab undance of tropical trees. Biotropica 24 :527531. Chapman, C. A., A. Gautier-Hion, J. F. Oa tes, and D. A. Onderdonk. 1999a. African primate communities: Determinants of stru cture and threats to survival. Pages 1-37 in J. G. Fleagle, C. H. Janson, and K. E. Reed, editors. Primate Communities. Cambridge University Press, Cambridge. Chapman, C. A., and J. E. Lambert. 2000. Ha bitat alteration and the conservation of African primates: Case study of Kibale National Park, Uganda. American Journal of Primatology 50 :169-185. Chapman, C. A., M. J. Lawes, L. Naughton-Treves, and T. R. Gillespie. 2003b. Primate Survival in Community-owned Forest Fragments: Are Metapopulation Models Useful Amidst Intensive Use? Pages 404 in L. K. Marsh, editor. Primates in Fragments. Kluwer Academic/Plenum Publishers, New York. Chapman, C. A., and C. A. Peres. 2001. Pr imate conservation in the new millennium: The role of scientists. Evolutionary Anthropology 10 :16-33. Chapman, C. A., R. W. Wrangham, L. J. Chapman, D. K. Kennard, and A. E. Zanne. 1999b. Fruit and flower phenology at two sites in Kibale National Park. Journal of Tropical Ecology 15 :189-211.

PAGE 89

78 Conklin-Brittain, N. L., R. W. Wrangham, and K. D. Hunt. 1998. Dietary response of chimpanzees and cercopithecines to seas onal variation in fruit abundance. II. Macronutrients. International Journal of Primatology 19 :971-998. Cordeiro, N. J., and H. F. Howe. 2001. Low recruitment of trees dispersed by animals in African forest fragments. Conservation Biology 15 :1733-1741. Cords, M. 1986. Interspecific and Intraspecific Variation in Diet of 2 Forest Guenons, Cercopithecus-Ascanius and C-Miti s. Journal of Animal Ecology 55 :811-827. Cords, M., and T. E. Rowell. 1986. Group Fi ssion in Blue Monkeys of the Kakamega Forest, Kenya. Folia Primatologica 46 :70-82. De Walt, S. J., S. K. Maliakal, and J. S. Denslow. 2003. Changes in vegetation structure and composition along a tropical forest chr onosequence: implica tions for wildlife. Forest Ecology and Management 182 :139-151. DeFries, R. S., R. A. Houghton, M. C. Hanse n, C. B. Field, D. Skole, and J. Townshend. 2002. Carbon emissions from tropical deforest ation and regrowth based on satellite observations for the 1980's and 1990's. Proc. Nat. Acad. Sci. 99 :14256-14261. Devos, A., and A. Omar. 1971. Territories and Movements of Sykes Monkeys (Cercopithecus-Mitis-Kolbi Neuman) in Kenya. Folia Primatologica 16 :196-&. Didham, R. K., and J. H. Lawton. 1999. Edge structure determines the magnitude of changes in microclimate and vegetation st ructure in tropical forest fragments. Biotropica 31 :17-30. Fairgrieve, C. 1995. The comparat ive ecology of Blue monkeys ( Cercopithecus mitis stuhlmannii ) in logged and unlogged forest, B udongo Forest Reserve, Uganda: the effects of logging on habitat and population density. PhD. University of Edinburgh, Edinburgh. Fairgrieve, C., and G. Muhumuza. 2003. F eeding ecology and dietary differences between blue monkey ( Cercopithecus mitis stuhlmanni Matschie) groups in logged and unlogged forest, Budongo Forest Reserve, Uganda. African Journal of Ecology 41 :141-149. Fernandez-Escobar, R., R. Moreno, and M. Garcia-Creus. 1999. Seasonal changes of mineral nutrients in olive leaves during the alternate-bearing cycle. Scientia Horticulturae 82 :25-45. Fiksen, O. 2000. The adaptive timing of diapau se a search for evolutionarily robust strategies in Calanus finmarchicus Ices Journal of Marine Science 57 :1825-1833. Gehlhausen, S. M., M. W. Schwartz, a nd C. K. Augspurger. 2000. Vegetation and microclimatic edge effects in two mixe d-mesophytic forest fragments. Plant Ecology 147 :21-35.

PAGE 90

79 Ghiglieri, M. P. 1984. The chimpanzees of Kibale Forest. Columbia University Press, New York. Gonzalez, A., and E. J. Chaneton. 2002. He terotroph species extin ction, abundance and biomass dynamics in an experimentally fragmented microecosystem. Journal of Animal Ecology 71 :594-602. Granjon, L., J. F. Cosson, J. Judas, and S. Ringeut. 1996. Influence of tropical rainforest fragmentation on mammal communities in French Guiana: Short-term effects. Acta Oecologica--Internatio nal Journal of Ecology 17 :673-684. Gupta, A. K., and D. J. Chivers. 1999. Biom ass and use of resources in south and southeast Asian primate communities. in J. G. Fleagle, C. H. Janson, and K. E. Reed, editors. Primate Communities. Cambri dge University Press, Cambridge. Gurd, D. B., T. D. Nudds, and D. H. Riva rd. 2001. Conservation of mammals in eastern North American wildlife reserves: How sm all is too small? Conservation Biology 15 :1355-1363. Haddow, A. J. 1952. Field and laborat ory studies on an African monkey Cercopithecus ascanius schmidti (Matschie). Proceedings of the Zoological Society of London 122 :297-394. Haddow, A. J., K. C. Smithburn, A. F. Ma haffy, and S. C. Bugher. 1947. Monkeys in relation to epidemiology of yellow fever in Bwamba County, Uganda. Transcripts of the Royal Society of Tr opical Medicine and Hygiene 40 Hafner, H., O. Pineau, and Y. Kayser 1994. Ecological determinants of annual fluctuations in numbers of breeding Little egrets ( Egretta garzetta L) in the Camargue, S. France. Revue D Ecologie La Terre et la Vie 49 :53-62. Hamilton, A. 1991. A Field Guide to Uganda n Forest Trees. Makerere University Printery. Kampala. Harmand, J. M., C. F. Njiti, F. Bernhard-Reversat, and H. Puig. 2004. Aboveground and belowground biomass, productivity and nutrient accumulation in tree imporved fallows in the dry tropics of Came roon. Forest Ecology and Management 188 :249265. Hill, C. M. 1997. Crop-raiding by wild vertebrates: the farmer's perspective in an agricultural community in western Uga nda. International Journal of Pest Management 43 :77-84. Howard, P. C., T. R. B. Davenport, F. W. Ki genyi, P. Viskanic, M. C. Baltzer, C. J. Dickinson, J. S. Lwanga, R. A. Matthews, and E. Mupada. 2000. Protected area planning in the tropics: Uganda's nati onal system of forest nature reserves. Conservation Biology 14 :858-875.

PAGE 91

80 Ingham, K. 1975. The Kingdom of Toro in Uganda. Methuen and Co. Ltd., London. Janzen, D. H. 1967. Synchronization of sexual reproduction of trees within the dry season in Central America. Evolution 21 :620-637. Jedrzejewska, B., H. Okarma, W. Jedrzejew ski, and L. Milkowski. 1994. Effects of exploitation and protection on forest structure, ungulate density and wolf predation in Bialowieza Primeval Forest, Poland. Journal of Applied Ecology 31 :664-676. Karanth, K. U. 1992. Conservation prospects for lion-tailed macaques in Karnataka, India. Zoo Biology 11 :33-41. Karr, J. R. 1976. Seasonality, resource availabi lity, and community diversity in tropical bird communities. The American Naturalist 110 :973-994. Katende, A. B., A. Birnie, and B. Tengnas. 1995. Useful Trees and Shrubs for Uganda: Identification, Propogation and Management for Agricultural and Pastoral Communities. Regional Soil C onservation Unit, Nairobi. Kayanja, F. I. B., and D. Byarugaba. 2001. Disappearing forests of Uganda: The way forward. Current Science 81 :936-947. Kingdon, J. 1974. East African Mammals: An atlas of evolution in Africa. The University of Chicago Press, Chicago. Kingdon, J. 1997. The Kingdon Field Guide to African Mammals. Academic Press, London. Klages, K., H. Donnison, J. Wunsche, and H. Boldingh. 2001. Diurnal changes in nonstructural carbohydrates in l eaves, phloem exudate and fruit in 'Braeburn' apple. Australian Journal of Plant Physiology 28 :131-139. Kouki, J., S. Lofman, P. Martikainen, S. Rouvinen, and A. Uotila. 2001. Forest fragmentation in Fennoscandia: LInking ha bitat requirements of wood-associated threatened species to landscape and habitat changes. Scandinavian Journal of Forest Research Supplement 3 :27-37. Krupnick, G. A., and W. J. Kress. 2003. Hots pots and ecoregions: a test of conservation priorities using taxonomic data. Biodiversity and Conservation 12 :2237-2253. Larue, M., S. Ringeut, D. Sabatier, and P. M. Forget. 2002. Fruit richness and seasonality in a fragmented landscape of French Guia na. Revue D Ecologie La Terre et la Vie Supplement 8 :39-57. Laurance, W. F., and G. B. Williamson. 2001. Positive feedbacks among forest fragmentation, drought, and climate cha nge in the Amazon. Conservation Biology 15 :1529-1535.

PAGE 92

81 Lawes, M. J. 2002. Conservation of Fragmented Populations of Cercopithecus mitis in South Africa: the Role of Reintroduc tion, Corridors and Metapopulation Ecology. Pages 375-392 in M. E. Glenn and M. Cords, editors. The Guenons: Diversity and Adaptation in African Monkeys. Kluwer Academic/Plenum Publishers, New York. Lawes, M. J., S. P. Henzi, and M. R. Perrin. 1990. Diet and Feeding-Behavior of Samango Monkeys ( Cercopithecus mitis labiatus ) in Ngoye Forest, South-Africa. Folia Primatologica 54 :57-69. Lernould, J.-M. 1988. Classificat ion and geographical distri bution of guenons: a review. in A. Gautier-Hion, F. Bourliere, J. P. Gautier, and J. Kingdon, editors. A Primate Radiation: Evolutionary Biology of the African Guenons. Cambridge University Press, Cambridge. Little, K. M., and R. A. W. Ga rdner. 2003. Coppicing ability of 20 Eucalyptus species grown at two high-altitide sites in Sout h Africa. Canadian Journal of Forest Research 33 :181-189. Lubowa, L., B. M. Byanyima, and P. N. Kavuma. 1963. Report of the Commision of Inquiry into the Administration by the Gove rnment of the Kingdom of Toro of the services for which it is responsible in certain counties of the Kingdom. Uganda Government, Entebbe. Luck, G. W., T. H. Ricketts, G. C. Dail y, and M. Imhoff. 2004. Alleviating spatial conflict between people and biodiversity. PNAS 101 :182-186. Lwanga, J. S. 1987. Group fission in Blue Monkeys ( Cercopithecus mitis stuhlmanni ): Effects on the socioecology in Kibale Forest Uganda. Master of Science. Makerere University, Kampala. Maganga, S. L. S., and R. G. Wright. 1991. Bark-stripping by Blue Monkeys in a Tanzanian forest plantation. Tropical Pest Management 37 :169-174. Marquis, R. J., E. A. Newell, and A. C. Villegas. 1997. Non-structural carbohydrate accumulation and use in an understory rain-forest shrub and relevance for the impact of leaf herbivory. Functional Ecology 11 :636-643. Marsh, L. K. 2003. The Nature of Fragmentation. Pages 404 in L. K. Marsh, editor. Primates in Fragments: Ecology and C onservation. Kluwer Academic/Plenum Publishers, New York. Marsh, L. K., and B. A. Loiselle. 2003. R ecruitment of black howler fruit trees in fragmented forests of Northern Belize. International Journal of Primatology 24 :6586. Medley, K. E. 1993. Primate conservation along the Tana River, Kenya--An examination of the forest habitat. Conservation Biology 7 :109-121.

PAGE 93

82 Mitani, J. C., W. J. Sanders, J. S. Lw anga, and T. L. Windfelder. 2001. Predatory behavior of crowned hawk-eagles ( Stephanoaetus coronatus ) in Kibale National Park, Uganda. Behav Ecol Sociobiol 49 :187-195. Mitani, J. C., and D. P. Watts. 2001. Why do chimpanzees hunt and share meat. Animal Behavior 61 :915-924. Mittermeier, R. A., and D. L. Cheney. 1987. C onservation of primates and their habitats. Pages 477-490 in B. B. Smuts, D. L. Cheney, R. Seyfarth, R. W. Wrangham, and T. T. Struhsaker, editors. Primate Societie s. Chicago University Press, Chicago. Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution 10 :58-62. Muri, H. 1999. Weather situation, aspects of reproduction and population density in roe deer ( Capreolus capreolus L.). Zeitschrift fur Jagdwissenschaft 45 :88-95. National Research Council. 1981. Techni ques for the study of primate population ecology. National Academy Press, Washington, D.C. Nergiz, C., and Y. Engez. 2000. Compositional variation of olive fruit during ripening. Food Chemistry 69 :55-59. Newmark, W. D. 1993. The role and design of wildlife corridors w ith examples from Tanzania. Ambio 22 :500-504. Norconk, M. A., and B. W. Grafton. 2003. Cha nges in Forest Composition and Potential Feeding Tree Availability on a Small Landbridge Island in Lago Guri, Venezuela. Pages 404 in L. K. Marsh, editor. Primates in Fragments: Ecology and Conservation. Kluwer Academic/Plenum Publishers, New York. Olupot, W. 1994. Ranging patterns of the grey-cheeked mangabey Cercocebus albigena with special reference to food finding and f ood availability in Kibale National Park. Master of Science. Makerere University, Kampala. Olupot, W. 1998. Long-term variation in mangabey ( Cercocebus albigena johnstoni Lydekker) feeding in Kibale Nationa l Park. African Journal of Ecology 36 :96-101. Olupot, W. 2000. Mass differences amoung male mangabey monke ys inhabitting logged and unlogged forest compartm ents. Conservation Biology 14 :833-843. Onderdonk, D. A., and C. A. Chapman. 2000. C oping with forest fragmentation: The primates of Kibale National Park, Uga nda. International Journal of Primatology 21 :587-611. Ortega-Huerta, M. A., and A. T. Pete rson. 2004. Modelling spatial patterns of biodiversity for conservation prioritization in North-easte rn Mexico. Diversity and Distributions 10 :39-54.

PAGE 94

83 Peres, C. A. 1990. Effects of hunting on western Amazonian primate communities. Biological Conservation 54 :47-59. Plumptre, A. J., and V. Reynolds. 1994. The ef fect of selective logging on the primate populations in the Budongo Forest Reserv e, Uganda. Journal of Applied Ecology 31 :631-641. Poiani, K. A., B. D. Richter, M. G. A nderson, and H. E. Richter. 2000. Biodiversity conservation at multiple scales: Functional sites, landscapes, and networks. Bioscience 50 :133-146. Pope, S. E., L. Fahrig, and N. G. Merriam. 2000. Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81 :2498-2508. Poulsen, J. R., C. J. Clark, and T. B. Smith. 2001. Seasonal variation in the feeding ecology of the Grey-cheeked mangabey ( Lophocebus albigena ) in Cameroon. American Journal of Primatology 54 :91-105. Prendergast, J. R., R. M. Quinn, J. H. Lawton, B. C. Eversham, and D. W. Gibbons. 1993. Rare species, the coincidence of diversity hotspots and conservation strategies. Nature 365 :335-337. Radersma, S., and C. K. Ong. 2004. Spatial dist ribution of root leng th density and soil water of linear agroforestry systems in sub-humid Kenya: implications for agroforestry models. Forest Ecology and Management 188 :77-89. Rao, M., and C. P. van Schaik. 1997. The beha vioral ecology of Sumatran orangutans in logged and unlogged forest. Tropical Biodiversity 4 :173-185. Richardson, J. S. 1991. Seasonal food limitation of detritivores in a montane stream an experimental test. Ecology 72 :873-887. Rode, K., C. A. Chapman, L. J. Chapman, and L. R. McDowell. 2003. Mineral resource availability and consumption by colobus in Kibale National Park, Uganda. International Journal of Primatology 24 :541-573. Rodriguez-Toledo, E., M, S. Mandujano, a nd F. Garcia-Orduna. 2003. Relationships Between Forest Fragments and Howler Monkeys ( Alouatta palliata mexicana ) in Southern Veracruz, Mexico. Pages 404 in L. K. Marsh, editor. Primates in Fragments. Kluwer Academic/Plenum Publishers, New York. Rudran, R. 1978. Socioecology of the Blue Monkeys ( Cercopithecus mitis stuhlmanni ) of the Kibale Forest, Uganda. Smithsonian Contributions to Zoology 249 Sams, C. E. 1999. Preharvest factors affecti ng postharvest texture. Postharvest Biology and Technology 15 :249-254.

PAGE 95

84 Schaefer, H. M., and V. Schmidt. 2002. Vertical stratification and cal oric content of the standing fruit crop in a tropical lowland forest. Biotropica 34 :244-253. Schaefer, H. M., V. Schmidt, and F. Bairle in. 2003. Discrimination abilities for nutrients: which difference matters for choosy birds and why? Animal Behavior 65 :531-541. Segelbacher, G., J. Hoglund, and I. Storch. 2003. From connectivity to isolation: genetic consequences of population fr agmentation in Capercaillie across Europe. Molecular Ecology 12 :1773-1780. Siex, K. S., and T. T. Struhsaker. 1999. Ec ology of the Zanzibar red colobus monkey: Demographic variability and habitat stabi lity. International Jo urnal of Primatology 20 :163-192. Skorupa, J. P. 1988. The effect of selective timber harvesting on rainforest primates in Kibale Forest, Uganda. Univer sity of California, Davis. Skorupa, J. P. 1989. Crowned eagles Stephanoaetus coronatus in rainforest Observations on breeding chronology and diet at a nest in Uganda. Ibis 131 :294298. Ssembeguya, F. C., G. O. B. Oda, and J. M. Okae. 1962. Report of the Commission of Inquiry into the Recent Disturbances amongst the Baamba and Bakonjo People of Toro. Uganda Government, Entebbe. Stickler, C. M. 2004. The effects of selectiv e logging on primate-habitat interactions: A case study of redtail monkeys ( Cercopithecus ascanius ) in Kibale National Park, Uganda. Master of Science. Univ ersity of Florida, Gainesville. Struhsaker, T. T. 1975. The red colobus monkey. University of Chicago Press, Chicago. Struhsaker, T. T. 1978. Food habits of five monkey species in the Kibale Forest, Uganda. Pages 225-248 in D. J. Chivers and J. Herbert, editors. Recent Advances in Primatology. Academic Press, London. Struhsaker, T. T. 1997. Ecology of an African Rainforest. University of Florida Press, Gainesville. Struhsaker, T. T., and M. Leakey. 1990. Pr ey selectivity by Crowned hawk-eagles on monkey in the Kibale Forest, Uganda. Behav Ecol Sociobiol 26 :435-443. Struhsaker, T. T., and L. Leland. 1979. Soci oecology of Five Sympatric Monkey Species in the Kibale Forest, Uganda. Pages 158-228 in J. Rosenblatt, R. A. Hinde, C. Beer, and M. C. Busnel, editors. Advances in the Study of Behavior. Academic Press, New York.

PAGE 96

85 Sun, C., B. A. Kaplin, K. A. Kristense n, V. Munyaligoga, J. Mvukiyumwami, K. K. Kajondo, and T. C. Moermond. 1996. Tree pheno logy in a tropical montane forest in Rwanda. Biotropica 28 :668-681. Syahuka-Muhindo, A. 1991. The Rwenzururu Move ment and the Democratic Struggle. Working Paper No. 15 Centre for Basic Research, Kampala, Uganda. Tutin, C. E. G., L. J. T. White, and A. Machanga-Missandzou. 1997. The use by rain forest mammals of natural forest fragme nts in an equatorial African savanna. Conservation Biology 11 :1190-1203. Twinomugisha, D., C. A. Chapman, M. J. Lawe s, C. O. Worman, and L. M. Danish. in press. How does the Golden Monkey of the Virungas cope in a fruit scarce environment? in Primates of Uganda. Umapathy, G., and A. Kumar. 2000. The occurr ence of arboreal mammals in the rain forest fragments in the Anamalai Hills, south India. Biological Conservation 92 :311-319. van Schaik, C. P., J. W. Terborgh, and S. J. Wright. 1993. The phenology of tropical forests: Adaptive significance and consequences for primary consumers. Annual Review of Ecology and Systematics 24 :353-377. Vargas, A., I. Jimenez, F. Palomares, and M. J. Palacios. 2002. Di stribution, status, and conservation needs of the golden-crowned sifaka ( Propithecus tattersalli ). Biological Conservation 108 :325-334. Waser, P. 1975. Monthly variations in feed ing and activity patte rns of the mangabey, Cercocebus albigena (Lydekker). East African Wildlife Journal 13 :249-263. Waser, P. 1977. Feeding, ranging and group size in the mangabey Cercopithecus albigena in T. H. Clutton-Brock, editor. Prim ate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes. Academic Press, London. Waser, P. 1984. Ecological differences and be havioral contrasts between two mangabey species. Pages 195-216 in P. S. Rodman and J. G. H. Cant, editors. Adaptations for Foraging in Non-Human Primates. Colu mbia University Press, New York. Waser, P., and O. Floody. 1974. Ranging patterns of the mangabey Cercopithecus albigena in the Kibale Forest, Uganda Zeitschrift fur Tierpsychologie 35: 85-101 Wasserman, M. D., and C. A. Chapman. 2003. Determinants of colobine monkey abundance: The importance of food energy, pr otein and fibre cont ent. Journal of Animal Ecology 72 :650-659. Williams, M. 2000. Dark ages and dark areas: Global deforestation in the deep past. Journal of Historical Geography 26 :28-46.

PAGE 97

86 Willock, C. 1964. The Enormous Zoo: A profile of the Uganda national parks. Harcourt, Brace and World, Inc., New York. Wilson, C. C., and W. L. Wilson. 1975. The in fluence of selective logging on primates and some other animals in East Kalimantan. Folia Primatologica 23 :245-274. Woodwell, G. M. 1974. Variation in th e nutrient content of leaves of Quercus alba Quercus coccinea and Pinus rigida in the Brookhaven forest from bud-break to abscission. American Journal of Botany 61 Woodwell, G. M. 2002. On purpose in scien ce, conservation and government The functional integrity of the Earth is at issue not biodiversity. Ambio 31 :432-436. Woolf, A. B., and I. B. Ferguson. 2000. Postha rvest responses to high fruit temperatures in the field. Postharv est Biology and Technology 21 :7-20. Wrangham, R. W., N. L. Conklin-Brittain, and K. D. Hunt. 1998. Dietary response of chimpanzees and cercopithecines to seas onal variation in fruit abundance. I. Antifeedants. Internationa l Journal of Primatology 19 :949-970.

PAGE 98

87 BIOGRAPHICAL SKETCH Cedric ODriscoll Worman received an ex cellent elementary school education from the Minnesota public school system (wanderi ng the forests, mountains, and deserts of North America with his family did not hur t either) and attended North Hollywood High School Magnet for Biological and Mathemati cal Sciences. He earned his Bachelor of Science at Iowa State University with specializations in restoration ecology and mammalian behavior. After graduation he joined the Peace Corps and was sent to Uganda where he learned more than he rea lly wanted to. Before entering graduate school, he taught in and explored Korea.


Permanent Link: http://ufdc.ufl.edu/UFE0006740/00001

Material Information

Title: Forest, Fragments, and Fruit: Spatial and Temporal Variation in Habitat Quality for Two Species of Frugivorous Primates (Cercopithecus mitis and Lophocebus albigena) in Kibale National Park, Uganda
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006740:00001

Permanent Link: http://ufdc.ufl.edu/UFE0006740/00001

Material Information

Title: Forest, Fragments, and Fruit: Spatial and Temporal Variation in Habitat Quality for Two Species of Frugivorous Primates (Cercopithecus mitis and Lophocebus albigena) in Kibale National Park, Uganda
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006740:00001


This item has the following downloads:


Full Text












FOREST, FRAGMENTS, AND FRUIT: SPATIAL AND TEMPORAL VARIATION
IN HABITAT QUALITY FOR TWO SPECIES OF FRUGIVOROUS PRIMATES
(Cercopithecus mitis AND Lophocebus albigena) IN KIBALE NATIONAL PARK,
UGANDA















By

CEDRIC O'DRISCOLL WORMAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Cedric O'Driscoll Worman


































This document is dedicated to all those who have helped protect Kibale Forest and to the
people of Karugutu.















ACKNOWLEDGMENTS

First and foremost I thank my family for not only the tradition of scholarship, but

also a passion for it. My field assistant, Amooti Katusabe Swaibu, is a man of great

integrity and honor in addition to possessing a wonderful knowledge of the forest, far

beyond what I could have hoped for. Without his skills, my work could not have been

done. Without his quick mind and original ideas, my work would not be half as good as I

hope it is. Without his tolerant camaraderie, my work would have been just that: work.

Webaale muno, Amooti! I thank my advisor, Colin Chapman, for all the help, advice, and

facilitation he has given to me during this project, and my committee members, Graeme

Cumming and Bob Holt, for their patience and suggestions. I thank the other researchers

at Kanyawara for making my stay there interesting and sociable, most especially Claudia

Stickler for being there. I thank Mike Wasserman and Lisa Danish for help with the

laboratory analysis. Lastly, I thank Manjula Tiwari, John Poulsen, Toshinori Okuyama,

Greg Pryor, Tristan Kimball, and all the other graduate students in the Department of

Zoology for all their help with my research and writing. Funding was provided by an

Alumni Fellowship from the University of Florida. Permission to conduct research in

Kibale National Park was granted by the Makerere University Biological Field Station,

Uganda Wildlife Authority, the National Council on Science and Technology, and the

Office of the President of Uganda, and Kibale National Park.

















TABLE OF CONTENTS

page

ACKNOW LEDGM ENTS ........................................ iv

LIST OF TABLES ................... ...... ..................... ................. vii

LIST OF FIGURES .............. ...................... ..... ...... ..............viii

ABSTRACT ........... ................... ........... .. ................x

CHAPTER

1 INTRODUCTION ................... .................. .............. .... ......... .......

Study Site...................................... .................. ................ ........4
Study Species................................. ........5
Blue Monkeys........................................ .........5
Gray-cheeked Mangabeys ..............................................6
Overview .................... ...................................... ................ ........ 7

2 DENSITIES OF TWO FRUGIVOROUS PRIMATES WITH RESPECT TO
FOREST AND FRAGMENT TREE SPECIES COMPOSITION AND FRUIT
A V A ILA B ILITY ....... ... ...... .... ........ .......... .......... .. ..8

Introduction .......................... .. .............. ........8
Methods ........................................ .............................. ........10
Study Site............................................10
Forest transects........................ ................... ...............1
Fragments.............. ..... ......... ....... ... .........14
Prim ate D iets .................................................................................. 14
Primate Densities........................................ .........14
Forest Composition .................................................. ........19
Fruit and Flower Availability ................................. ........21
Results............. ....... .................................22
Primate Densities........................................ .........22
Forest Composition .................................................. ........24
Fruit Availability .................................................... ........29
Discussion ......... ........................... ........... ...............38
Primate Densities........................................ .........38
Forest Composition .................................................. ........39


v










Fruit Availability .......................................... ...... ... ....43
Transect 1: A Primate Perspective ..........................................43
Summary ........... ....................... ..............................45

3 SEASONAL VARIATION IN THE QUALITY OF A TROPICAL RIPE FRUIT
AND THE EFFECT ON THE DIETS OF THREE FRUGIVORES ......................49

Introduction ........... ......... ......... ............ .....49
M methods ...................................................... ........ 50
Results ........................... .... ................ ..54
Discussion ....................... .... .... ............ .56

4 CONCLUSION................... .. ....... .................63

APPENDIX: POTENTIAL FACTORS INFLUENCING HABITAT USE BY
Cercopithecus mitis AND Lophocebus albigena ........................................ ...66

Unexamined Possibilities......................... ........... ......... 66
Light Levels, W ind and D isease................................ .................. 66
Human Impacts........................... .............. 67
Predation........................................ ........69
A rthropod A bundance ............................................... ............... 70
Examined Possibilities........................ ............. ...............71
Temperature Extremes and Evaporative Potential ......................................... 71
Forest Composition and Fruit Availability....................................................73
Nutritional Quality of Vegetable Foods ....................................................73
A rthropod Foraging Substrate ...................................................................... 73

L ITER A TU R E C ITED ...............................................................76

BIOGRAPHICAL SKETCH .................................................. ............... 87
















LIST OF TABLES


Table page

2-1 Species specific food items constituting >4% of the annual diet of
Kanyawara C. mitis and L. albigena groups. ................ .................. ............15

2-2 Group densities of the arboreal monkeys of Kanyawara by transect.....................23

2-3 One-tailed Pearson's correlations between primate densities and relative densities
and the basal area densities of various tree categories and fruit
availability (average summed standardized fruiting intensity). .............................23

2-4 Groups of the arboreal monkeys of Kanyawara sighted/km walked by transect
during the m onkey surveys .............................................. ............... 24

2-5 The basal area densities (cm2 tree basal area/m2 land) of
trees > 10 cm dbh in transect 1 and the forest section of transect 1 compared
to the 97.5% confidence limits of the fragments and the other transects..............25

2-6 The top ten tree species by basal area density of
individuals > 10 cm dbh in the forest fragments and the transects
characterizing the main forest block near Kanyawara, Kibale National Park..........25
















LIST OF FIGURES


Figure page

1-1 Location of the Makerere University Biological Field Station (MUBFS), Kibale
N national Park, U ganda. ............................. ........................... .. .4

2-1 The locations of the 8 transects in Kibale National Park, Uganda...........................13

2-2 Standardized fruiting intensity for species whose fruits comprise > 4% of the
annual diet of C m itis .............. ........................................... ... .......... ......... ....... 30

2-3 Standardized fruiting intensity for species whose fruits comprise > 4% of the
annual diet of L. albigena. ................................................ ............... 34

2-4 C. mitis territories in relation to the unused area of forest...............................47

3-1 The relationship between the monthly dry matter lipid content of ripe Celtis
durandii fruit and the summed average daily rainfalls of the concurrent and
previous months at the Makerere University Biological Field Station, Kibale
National Park..................... ................ ......... 54

3-2 The relationships between reported dietary use of Celtis durandii fruit and lipid
content of the fruit, average daily rainfall of the previous and concurrent months
(a predictor of lipid content), and fruit availability ............ ...........55

3-3 Changes in the percentage of Celtis durandii ripe fruit in the diet of a
Cercopithecus ascanius group relative to changes in fruit availability and fruit
lipid content. .................. .... ................. ...... ........57

3-4 Changes in the percentage of Celtis durandii fruit in the diet of a Cercopithecus
mitis group relative to changes in fruit availability and estimated fruit lipid
content. ............................. .................... ........ 58

3-5 Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus
albigena group relative to changes in fruit availability and estimated fruit lipid
content ...................... .... .. .................... .59

3-6 Changes in the percentage of Celtis durandiifruit in the diet of a Lophocebus
albigena group relative to changes in estimated fruit lipid content..........................60









A-i The intra- and inter-measurement period variation of 10 evaporation potential
monitors from the same location. ................................ ............... 72
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

FOREST, FRAGMENTS, AND FRUIT: SPATIAL AND TEMPORAL VARIATION
IN HABITAT QUALITY FOR TWO SPECIES OF FRUGIVOROUS PRIMATES
(Cercopithecus mitis AND Lophocebus albigena) IN KIBALE NATIONAL PARK,
UGANDA
By

Cedric O'Driscoll Worman

August, 2004

Chair: Colin A. Chapman
Major Department: Zoology

Conservation of wildlife populations requires extensive knowledge of their habitat

requirements, efficient methods of evaluating habitat quality, and an understanding of the

habitat value of fragments and edges. In this study I first evaluate the relationships

between the basal area densities of several types of important food trees and fruit

availability with the densities of two species of frugivorous monkeys (Cercopithecus

mitis, the Blue monkey, and Lophocebus albigena, the Gray-cheeked mangabey) that

have varying densities throughout the study area, Kibale National Park, Uganda, which

includes a region of forest that is unused by both species. The density of C. mitis was

most strongly correlated with the basal area density of all types of food trees combined.

The density of L. albigena was not correlated with the basal area densities of any

category of food trees or with fruit availability. However, an index of density, number of

groups seen per kilometer walked, was marginally correlated to fruit availability. The









lack of a satisfactory relationship between the basal area densities of food trees and L.

albigena density may be due to a mismatch in scale between this study and their large

home ranges.

The unused area of forest was then compared to the other areas of the forest and the

forest fragments outside the park that are also uninhabited by the two species in question.

It was found to have higher basal area densities in all food tree categories for both species

than the forest fragments and lower basal area densities of most categories than the other

parts of the forest, indicating that the forest fragments are very poor quality and would be

unused even if dispersal were likely. The fragments were also found to have a higher

frequency of exotic and edge tree species than the forest transects.

Because fruit availability is so often used to evaluate habitat, but its quality is

generally assumed to be constant, I examined nutritional components of a constantly

available fruit important in the diets of many birds and mammals (Celtis durandii). The

lipid content varied from 0.03-30.8% over 6 months and was predicted by the summed

average daily rainfalls of the previous and concurrent months. Using 4 existing data sets,

it was shown that lipid levels affected the intake of this fruit by at least 3 primate species,

with more fruit being eaten when it contained more lipids. It appeared that fruit

availability was being driven by consumption and not, as is often assumed, the other way

around. This indicates that the seasonality of resources may be underestimated in tropical

forests depending on the strategies of the fruiting trees














CHAPTER 1
INTRODUCTION

The continued loss of contiguous forest blocks via deforestation and fragmentation

and the resulting dependence on preserves and parks for the survival of remaining forest

species (Marsh 2003) has led to a triage system of conservation in which protection

priority is defined by biotic value (various levels of biodiversity, presence of endangered

or important organisms or habitats, etc.) (Prendergast et al. 1993, Krupnick and Kress

2003, Luck et al. 2004, Ortega-Huerta and Peterson 2004) or function (Poiani et al.

2000). This often opposes the more traditional conservation by consumption when areas

are preserved simply because they are the least valuable for human use (e.g., mountains,

deserts, disease ridden areas, etc.) (Willock 1964, Rodriguez-Toledo et al. 2003,

Chapman et al. 2003b) or scarce and only valuable if protected (e.g., hunting/forestry

reserves) (Jedrzejewska et al. 1994, Williams 2000). An assumption in the triage system

is that non-protected habitat will be lost and protected areas will be isolated from each

other (Newmark 1993); thus the ability of a protected area to support viable long-term

populations is of paramount interest (Gurd et al. 2001). To fulfill the goals of protecting

the biologically richest areas of (at least) the minimum required size to ensure healthy

populations and managing them properly, habitat requirements must be known and the

total area providing those requirements calculated (Kouki et al. 2001).

The requirements of species that have broad distributions over many habitat types

and flexible diets are difficult to define, but a comparison of differentially used areas

within a single habitat type may be the first step in a broader understanding of their









requirements (Chapman et al. 2002a). Ideally, similar studies should be done throughout

a species' range and compared for commonalities; however, depending on the species,

this could be a daunting task. A more achievable, if less certain, method is to use

knowledge of local conditions (diet, available tree species, etc.) to tailor the findings from

one area to others.

An understanding of habitat requirements is also vital to detailed estimations of

available habitat and expected carrying capacity of a given area. Since it is unlikely that

resources will be spread evenly throughout a protected area or even a habitat type,

estimations of carrying capacities need to be based on the actual resource landscape and

not on the richest (and often the most studied) areas. Even if an average for the entire

area is used, the spatial distribution of resources or resource types may still be an

important consideration (Pope et al. 2000). For example, two types of resources may be

spatially separated requiring frequent travel and high energy expenditure or reducing the

number of possible territories.

This applies well to the many species of primate that have diets that are spatially

and temporally variable and are widespread over multiple habitat/forest types. Kibale

National Park provides an ideal situation to study the habitat requirements of two

widespread species: Cercopithecus mitis (Blue monkey) and Lophocebus albigena (Gray-

cheeked mangabey) because they both have variable densities throughout the forest,

including an unused edge adjacent to frequented areas (C. Chapman, pers. comm.) and

because of the extensive research trail system and detailed, long-term dietary and ranging

information available from previous research (Struhsaker 1997). Additionally, there are

forest fragments outside the main forest block which are unused by C. mitis and L.









albigena (Onderdonk and Chapman 2000) and an examination of the fragments in

relation to the used and unused areas of the main forest might help in understanding how

fragmentation affects these two frugivores.

Tropical high forests currently form 4.5% of Uganda's land cover, though over

30% of this is considered degraded (Kayanja and Byarugaba 2001). Outside of parks and

reserves (which contain 14% of Uganda's area, (Howard et al. 2000) and hold 62% of the

existing forest, (Kayanja and Byarugaba 2001)), tropical high forests are mainly relegated

to small fragments in a human dominated matrix. For the successful conservation of

forest species, the habitat value of these fragments and their possible importance in

assisting demographic connection between metapopulations in the main forests must be

more fully understood. If edges, fragments, degraded areas or other forest types offer

undesirable habitat, the effective areas available to populations of forest species, like C.

mitis and L. albigena, is much less than the total forest cover. If species are unable to use

forest fragments to move between forest blocks, each local population may be isolated

with higher than assumed risks of extinction (Brashares 2003) and loss of genetic

variation (Segelbacher et al. 2003).

The goal of this study was to contribute to the understanding of habitat

requirements, and ultimately conservation, of C. mitis and L. albigena through the

evaluation of potential habitat quality indicators with observed patterns of habitat use in

different areas of the forest, and to evaluate the habitat potential of the forest fragments

outside Kibale National Park thereby gaining a better understanding of what types of

habitat characteristics potentially limit the distribution and density of these species over

wider areas.
































Kibale National Park


Figure 1-1: Location of the Makerere University Biological Field Station (MUBFS),
Kibale National Park, Uganda.

Study Site

Kibale National Park, Uganda (Figure 1-1) is mostly comprised of mid-altitude

moist tropical forest (elevation: 920-1590 m, rainfall from 1990-2001: 1749 mm/year,

latitude: 00 13'-0o 41' N) and has two rainy and two dry seasons each year (though the

length and severity varies greatly). The fruiting of tree species in Kibale is usually

synchronous, though the timing of fruiting can be irregular, sub-annual, annual, or super-

annual depending on the species (Chapman et al. 1999b).

Prior to becoming a national park, the area was a forest reserve with the last

logging in the natural forest occurring in 1969. This study took place in an unlogged









compartment (K30) and a relatively untouched section of a lightly logged compartment

(K14) (Chapman et al. 2000).

Study Species

Blue Monkeys

Blue monkeys (Cercopithecus mitis stuhlmanni) belong to the C. (nictitans)

superspecies, which includes both the Putty-nosed monkeys (C. (n.) nictitans) of central

and west African evergreen forests and the Gentle monkeys (C. (n.) mitis) of eastern and

southern African forests. Gentle monkeys are strongly arboreal and rarely descend from

the trees, but have been observed foraging at ground level (Kingdon 1974, Beeson 1989).

Their social structure normally consists of territorial matrilineal female-bonded groups

with a single resident male (Rudran 1978, Butynski 1990). However, if many females

come into estrus at once, the resident male may be temporarily unable to expel transient

males from the group (Cords and Rowell 1986).

The Gentle monkey is often considered frugivorous in spite of the catholic and

variable diets of the subspecies that include ripe and unripe fruits, young and mature

leaves, invertebrates, flowers, seeds, and, occasionally, bark (Kingdon 1974, Struhsaker

1978, Wrangham et al. 1998). Dietary variation is both spatial (between territories,

forests, and subspecies) and temporal (monthly, seasonally, and yearly) and can be quite

extreme (Rudran 1978, Lawes et al. 1990, Chapman et al. 2002b, Fairgrieve and

Muhumuza 2003). Because of their diverse and variable diets, extensive geographic

distribution, and broad climatic and habitat tolerance (Lawes et al. 1990), Gentle

monkeys are considered classic generalists (Kingdon 1974, Struhsaker 1978, Butynski

1990, Fairgrieve 1995).









The Blue monkey subspecies (C. m. stuhlmanni) occurs in eastern Congo, Uganda,

western Kenya, and northern Tanzania (Lernould 1988). In Uganda, Blue monkeys are

found on the eastern border near Mt.Elgon, on the north eastern border near Kidepo

National Park, and extensively on the western side of the country including Budongo

Forest Reserve, Ruwenzori Mountains National Park, Maramagambo Forest, and Kibale

National Park. The southern border of Uganda is occupied by other subspecies, the

Golden monkey (C. m. kandti) and the Silver monkey (C. m. doggetti). In Kibale

National Park, Blue monkeys occur at lower densities (Chapman et al. 2000) and with

larger home ranges (Rudran 1978, Butynski 1990) than in close by, more northerly,

Budongo Forest (Aldrich-Blake 1970, Fairgrieve 1995) and are not found in any of the

fragments outside the main forest block of Kibale (Onderdonk and Chapman 2000).

Also, Blue monkey densities decline as one moves from the northern to the southern end

of Kibale, suggesting a possible gradient in habitat suitability (Chapman and Lambert

2000).

Gray-cheeked Mangabeys

Gray-cheeked mangabeys (Lophocebus albigenajohnstoni) are closely related to

the baboons and, though strictly arboreal, retain many similarities with them such as

multi-male groups, complex and frequent vocalizations, and eostrus advertisement with

swollen and brightly colored perineal skin (Kingdon 1974). Gray-cheeked mangabeys

specialize on intense fruiting events and consequently have large home ranges (Waser

and Floody 1974). However they also eat other vegetative matter, including bark, young

leaves, mature leaves, pith, and flowers (Olupot 1998) and spend a substantial amount of

time foraging for arthropods (Waser 1975). Much like the Blue monkey, Gray-cheeked

mangabeys appear to forage for more sedentary arthropods in, as well as on, substrates.









They often rip apart rotten wood, tear off dead bark, and split twigs as well as searching

through moss and epiphyte layers (Struhsaker 1978).

Grey-cheeked mangabeys range from the Central African coast to the Nile, with

their distribution in Uganda limited to a fairly narrow band from Semuliki National Park

in the west to Mabira Central Forest Reserve in the east (Kingdon 1974).

Overview

To gain an understanding of habitat requirements and gauge the effectiveness of

measures of habitat quality, different measures of habitat quality were done and related to

primate use of the areas measured both in the forest and in the forest fragments.

Nutritional analysis of food items was also done to test whether the measures of habitat

quality were adequate indicators of actual resource availability.

Chapter 2 deals with the composition of the forest (basal area density of food trees)

and fruit availability and how well different measures of food availability are correlated

to observed monkey densities, as well as the composition of the fragments and how they

relate to the forest transects in terms of habitat for C. mitis and L. albigena.

Chapter 3 explores the temporal variability in the nutritional quality of the fruit of a

common tree species and its impacts on the diets of three species of primate. This

chapter discusses the importance of considering changes in fruit quality and not just fruit

availability when describing changes in habitat quality and resource availability over

time.


Chapter 4 concludes and summarizes.














CHAPTER 2
DENSITIES OF TWO FRUGIVOROUS PRIMATES WITH RESPECT TO FOREST
AND FRAGMENT TREE SPECIES COMPOSITION AND FRUIT AVAILABILITY

Introduction

Primates are currently of great interest to conservation not only because of their

potentials for acting as flagship species (Karanth 1992, Vargas et al. 2002) but also

because half of the world's primate species are in trouble for a variety of reasons

(Chapman and Peres 2001). While hunting is an important and widespread threat (Peres

1990, Chapman et al. 1999a), the dependence of most primate species on tropical forests

(Mittermeier and Cheney 1987) and the continuing devastation of these forests on a

global scale (DeFries et al. 2002) make an understanding of primate habitat requirements,

limitations, and flexibilities in relation to heterogeneity in primary and degraded forests

paramount for conservation.

Forest composition has often been studied as the major factor determining the

abundance and distribution of forest dwelling primates. These studies have sometimes

been conducted in reference to the changes initiated by logging (e.g., Wilson and Wilson

1975, Skorupa 1988, Plumptre and Reynolds 1994, Rao and van Schaik 1997, Chapman

et al. 2000, Olupot 2000, Fairgrieve and Muhumuza 2003), or with reference to the

unique characters of fragmented forests (Woodwell 2002; e.g., Medley 1993, Granjon et

al. 1996, Onderdonk and Chapman 2000, Umapathy and Kumar 2000, Marsh and

Loiselle 2003, Norconk and Grafton 2003). In both cases, changes in food availability

are often seen as a main driving force behind changes in primate densities (for an









exception see Onderdonk and Chapman 2000). In fragments, changes in food availability

can be caused by many processes, such as greater windfall, the cutting of firewood, an

increase in proportion of unfavorable microclimate, demographic stocasticity, distance

from a main forest block, or the extinction of primary dispersers (Cordeiro and Howe

2001).

Many previous studies are limited with respect to identifying key habitat

requirements because of the time lag between a dramatic disturbance (logging or

fragmentation) and a response by the species in question (Brooks et al. 1999, Chapman et

al. 2000, Gonzalez and Chaneton 2002). In the case of fragmentation, this limitation is

compounded by the difficulty of separating habitat requirements from dispersal abilities

of a species. If a species of interest is absent from fragments, it could be due either to

insufficient resources or the inability of the species to transfer from the main forest or

among fragments.

The situation in Kibale National Park, Uganda can be used to address the issue of

habitat requirements by relating habitat characteristics and natural heterogeneity of use in

areas of the forest that have not been impacted significantly by commercial logging,

thereby avoiding the time lag issue. The dispersal issue can be circumvented by

examining a fragment-like area (it abuts human settlement, and has a large proportion of

edge) completely avoided by the species in question (Cercopithecus mitis and

Lophocebus albigena) but contiguous with inhabited forest.

It is expected that, within Kibale National Park, the observed densities of C. mitis

and L. albigena will be related to the abundance of their respective foods. Given that

there are forest fragments near Kibale inhabited by the same diurnal primates as the









avoided area (including an absolute lack of C. mitis and L. albigena groups), with the

exception of Cercopithecus Ihoesti (Onderdonk and Chapman 2000), the question arises

as to whether these fragments are unused by C. mitis and L. albigena because of a lack of

resources or due to poor dispersal abilities in these species. By comparing the

availability of resources within the fragments with their availability in the used and

unused portions of the contiguous forest, the habitat suitability of the fragments can be

gauged. However, if the fragments have poor habitat quality, the hypothesis that distance

from the main forest block may limit C. mitis and L. albigena use of the fragments cannot

be rejected, yet is moot from a conservation perspective unless the fragments increase in

habitat quality in the future.

Methods

Study Site

Kibale National Park, Uganda is mostly comprised of mid-altitude moist tropical

forest (elevation: 920-1590 m, rainfall from 1990-1998: 1778 mm/year, latitude: 00 13'-0o

41' N) and has two rainy and two dry seasons each year (though the length and severity

varies greatly). The fruiting of tree species in Kibale is usually synchronous, though the

timing of fruiting can be irregular, sub-annual, annual, or super-annual depending on the

species (Chapman et al. 1999b).

Prior to becoming a national park, the area was a forest reserve with the last

logging in the natural forest occurring in 1969. This study took place in an unlogged

compartment (K30) and a relatively untouched section of a lightly logged compartment

(K14) (Chapman et al. 2000). Outside the main forest block contained in the national

park, community owned forest fragments provide habitat for many of the forest primate

species (Onderdonk and Chapman 2000).









Forest transects

A system of transects was established in the existing research trail grid near the

Makerere University Biological Field Station (MUBFS), near the Kanayawara village,

Kibale National Park, Uganda in an area that has not been impacted much by commercial

logging (C. Chapman, pers. comm.). Because edges are increasingly prevalent sources of

habitat heterogeneity, one transect transectt 1) was placed along the avoided forest edge

and another along an edge known to be used by both C. mitis and L. albigena transectt 8).

Both edge transects were paired with a nearby interior transect transectt 1 was paired

with transect 3 and 8 with 6) and each pair was connected at both ends by two transitional

transects transectss 2, 4, 5, and 7) (Figure 2-1). Due to the lay of the land and limitations

of the existing trail system, transects ranged from 453 m to 1205 m long, with a mean of

767 m.

Because transect 1 ventures beyond the forest, for some analyses only the area of

the transect in the forest was considered to ensure that any observed differences were not

due simply to the open and unique character of the field station. The section of transect 2

that ran through the area avoided by C. mitis and L. albigena (represented by transect 1)

was discarded for analysis to remove the influence of the avoided area and make transect

2 representative of a utilized forest area. Due to the spatial arrangement of the transects,

there is an area of overlap at some of the corners, but this is quite small, accounting for

less than 0.5% of the total area measured, therefore each transect is considered

independent.









Figure 2-1: The locations of the 8 transects in Kibale National Park, Uganda. The
transects are shown with thick lines. Roads are shown with thin lines. The
discarded area of transect 2 is shown with a dashed line. The Makerere
University Biological Field Station (MUBFS) is also labeled. The satellite
image is a Quickbird (high resolution, 2.4 m) from DigitalGlobe, Inc.,
Longmont, CO, USA.


















.. .. .























O'A












.0 4 0









Fragments

The human dominated matrix outside Kibale National Park consists mainly of

small-scale non-industrialized agriculture, large-scale tea plantations, pastures and fallow

land with forest fragments surviving on agriculturally undesirable topography such as

steep slopes and swampy lowlands (Chapman et al. 2003b). The fragments are actively

used by the surrounding people for fuelwood, poles, livestock fodder, and medicinal and

food plants (Chapman et al. 2003b). Many of these fragments are inhabited by the same

primates as the main forest block in the national park with the notable exceptions of C.

mitis and L. albigena (Onderdonk and Chapman 2000).

Primate Diets

Only specific dietary items (i.e., a specific part from a particular species) that

constituted > 4% of the total annual diet according to Rudran (1978) and Butynski (1990)

for C. mitis or Waser (1975) and Olupot (1994) for L. albigena were considered. This

produced 10 specific dietary items for C. mitis and 7 for L. albigena (Table 2-1). Though

Olupot (1994) reports liana fruits as the second most frequent dietary item for L. albigena

(8.34%), they were ignored in this study because there are a number of species, liana

species in general are poorly known and difficult to identify, and the dominance of lianas

measured in basal area is not comparable to that of trees.

Primate Densities

Primate group densities were estimated using a line transect method, which is

appropriate for easily detectable diurnal primates (National Research Council 1981,

Chapman et al. 2000). The transects were surveyed for primates once a fortnight for a

year and a half (7/01 to 12/02). Observers walked the transects at a speed of about











Table 2-1. Species specific food items constituting >4% of the annual diet of Kanyawara C. mitis and L.
albigena groups. Butynski's (1990) numbers are the averages of study groups 1-4. Rudran's
(1978) numbers are for groups 1 (not in parentheses) and 2 (in parentheses). Olupot (1994)
and Rudran (1978) reported percentages of vegetable diet. Their numbers have been
recalculated as percentages of total diet. The order of sources is constant throughout the table
with Butynski and Olupot above and Rudran and Waser below for all species.

Species Part Percent of annual diet
C. mitis L. albigena

s africana frut 4.1 (Butynski 1990)* -- (Olupot 1994)
Celtis africana fruit
6.6(6.8) (Rudran 1978) -- (Waser 1975)

Celtis durandii fruit 12.4
4.1(5.2) 6.3

Croton macrostachys fruit
-c\ -)
-- 6.2
Diospyros abyssinica fruit -- 22.6

Diospyros abyssinica leaves 4.8 --
-- 4.2
Ficus bracylepis fruit 4.9 6.1

-- 4.6
Ficus exasperata fruit
(11.2) --
Markhamia lutea young petioles --

Pancovia turbinata fruit 6.6 4.1
6.6 4.1

Parinari exelsa invertebrates 5.0

Premna angolensis flowers

Teclea nobilis fruit 5.3

Uvariopsis congensis fruit 5.3(8.5) -

*Butynski (1990) only reported foods used by Kanyawara groups and ignored at Ngogo.


1 km/hr during the morning hours (7:30-10:30) and recorded the method with

which each primate troop was first detected, the observer to animal distance (from the

first individual seen), the perpendicular trail to animal distance, the height of the

observed individual, the tree species being occupied by the troop, and any observed

feeding. Solitary animals were noted, but excluded from analysis. All distances and

heights were estimated by trained and practiced field assistants. Information for groups









at transect intersections was recorded separately if the group was seen from both transects

(or was likely to have been seen but moved at the approach of the observers).

Unfortunately, there were not enough sightings of C. mitis or L. albigena groups to

use density estimating computer programs like Distance (Buckland et al. 1993) and there

is debate over the most appropriate method of estimating densities from line transect

counts for primates (Chapman et al. 2000). Unlike large ungulates, the landscape

available for travel by arboreal primates is fully three dimensional which can lead to

unusual sighting-distance histogram profiles which in turn are thought to overestimate

group densities (National Research Council 1981). To control for this, observer to

animal distance is often used instead of perpendicular distance (National Research

Council 1981). This overestimates the area being sampled and can exclude observations

of groups seen in the sampled area but from a long distance, therefore lowering density

estimates. However, the histogram generated by this study using perpendicular distance

and 5 m intervals was a classic Kelker histogram with a plateau extending from 0 to a

distinct shoulder and subsequent steep decline in sightings. Therefore perpendicular

distances were used to estimate densities.

Because there were so few sightings of C. mitis or L. albigena, all primate group

sightings were used to produce the Kelker histogram with the assumption that all species

are equally visible. Even when unusual species (i.e., ground dwellers like Papio anubis

(Olive baboon) and Cercopithecus Ihoesti (L'Hoest's monkey), species normally found in

small quiet groups like Colobus guereza (Black-and-white colobus), and Pan troglodytes

(Chimpanzee), which often exhibits both characters) were excluded, the profile of the

histogram did not change meaningfully.









One of the assumptions of the Kelker method is that 100% of the animals (or

groups) within a certain distance are observed. That distance (multiplied by the transect

length) is used to delineate the area sampled for density calculations. Ideally, the cutoff

distance is determined by the sharp drop-off of observations after an initial plateau in the

sighting distance histogram. It is assumed that the animals are distributed randomly in

space (at least with respect to the transect) so the drop-off in observations is caused by a

drop-off in the sighting probability. Previous studies have often used an objective

method for determining the cutoff distance (Chapman et al. 2000) but this assumes a

sudden decline in the probability of being seen with distance from transect instead of a

slow, consistent drop. Using the same objective criteria as Chapman etal. (2000), the

cut-off for this study would be 44 m, which is almost twice the subjectively determined

cutoff (24 m) and seems unreasonable for this forest. The violation of the assumption

that no groups within the cutoff are missed underestimates densities. There was no

correlation between distance from transect and number of groups seen per meter within

the 24 m cutoff (Pearson's r = -0.157, p = 0.227, N = 25) indicating that this assumption

is not significantly violated.

Pearson's correlations were used to detect relationships between the transect

densities of C. mitis and the average summed standardized fruiting intensity (see Fruit

Availability below) for all fruit species, and the basal area density of all trees, species

specific food trees, fruit trees, leaf trees, and flower trees. In these analyses one-tailed

probabilities were used since clear a priori predictions were previously made. Pearson's

correlations were also used to detect relationships between transect densities of L.

albigena and the average summed standardized fruiting intensity for all fruit species, and









the basal area density of all trees, species specific food trees, fruit trees, and arthropod

trees (tree species used as invertebrate foraging substrate). Only the forest area was used

to represent transect 1 for these analyses.

Comparing a relative index of density (groups sighted/km walked during the

surveys) to the density estimations for each species can be used to strengthen confidence

in the density estimations (Chapman et al. 2000). In this case, the small numbers of

sightings may make density estimates particularly susceptible to the influence of a single

sighting and its inclusion or exclusion due to distance from the transect, therefore the

Pearson's correlations were recalculated using the index of relative density for both

primate species. Checking the results in this manner was considered more important for

L. albigena because the relative index of density was less strongly correlated with

calculated density for L. albigena (r = 0.736, one-tailed p = 0.019, N = 8) than for C.

mitis (r = 0.986, one-tailed p < 0.001, N = 8).

Though invertebrates generally form an important part of the diet of C. mitis

(Twinomugisha et al. in press), this is not used as a dietary category because no particular

species of tree was used enough for invertebrate foraging to qualify as an important food

species. Therefore the lack of an invertebrate foraging substrate category in food trees

should not be taken to mean that invertebrates are unimportant, but as an indication that

C. mitis may not be as selective about the species in which they forage for invertebrates

as they are about the species in which they forage for fruit, leaves, or flowers.

The results for the P. troglodytes, C. Ihoesti, and P. anubis are not given as they

have more terrestrial habits (making line transect surveys less appropriate) and were









rarely seen during surveys, despite the frequent observation of the animals or their sign at

other times.

Forest Composition

Free standing woody plants (including strangler figs) with a dbh > 10 cm and

within 10 m of either side of a transect were identified. Each stem was measured (dbh),

the height estimated, and it was noted if the plant was a snag, was bent or broken, or if

the main stem was dead. Snags were excluded from the analyses. If a plant was in an

area of overlap between transects, it was counted in each separately. Attempts were

made to measure above any pronounced buttressing, and, although this was not always

feasible, only a few of the largest trees of rarer species presented this sort of problem.

Multiple stemmed plants were treated as if each stem were a separate individual.

A fig with a diameter of more than 10 cm 3 m up on the host and with a substantial

crown could be supported by only small roots at breast height, making dbh a meaningless

measure. If multiple roots converged into a single trunk before the trunk started

branching, the diameter of the single trunk was measured or estimated. If the fig had

large roots connecting above the branching point (essentially multiple trunks), the main

trunk and several of the large roots were measured and combined to give a single dbh

estimate. These huge figs are widely spaced and were captured by the transects very few

times. Most of the fig trees measured presented no unusual difficulties.

The dbh of each plant was converted to basal area and summed for each species to

produce the total basal area by species for each transect. The total species basal area was

divided by the total area covered by a given transect, producing a basal area density

(cm2/m2) for each species for each transect.









In the fragments, a complete census of every tree > 10 cm dbh was done. The

forest fragments and transects were compared using bootstrapping analysis in order to see

if the forest composition of the unused transect was more similar to that of the fragments

or the rest of the forest. The variables used to characterize the transects and fragments

for C. mitis were the basal area densities of all trees, all C. mitis food trees, fruit trees,

leaf trees, and flower trees. Those for L. albigena were the basal area densities of all

trees, all L. albigena food trees, fruit trees, and arthropod foraging trees. It was assumed

that transect 1 would fall in between the fragments and the forest so the upper 97.5%

confidence limit was calculated for the fragment characters and the lower 97.5%

confidence limit was calculated for the transect characters and compared to transect 1 and

transect 1 forest only. For each variable, a sample population equal in size to the original

population was created by sampling the original population with replacement 10,000

times. The 97.5% confidence limits for each variable were calculated from the

distributions of the means of the sample populations.

To determine if the habitat of the fragments mirrored that of the main block of

forest, the top ten trees by basal area density for each fragment and transect were classed

as edge/savanna if they were described in Hamilton (1991) as only, primarily or

commonly occurring in forest edges or savannas, otherwise they were classed as forest

trees. Eucalyptus sp. was the only exotic in the top ten lists and was treated as an edge

species because it is either planted by humans or disperses from the human dominated

matrix outside the fragments. X2 analyses were done on the frequency of edge/savanna

and/or exotic species in the ten most dominant species in the fragments versus the

transects. All of transect 1 was included in the transect group in order to be as









conservative as possible. One species, Markhamia lutea (formerly M. platycalyx),

defined by Hamilton (1991) as, "[a] very common forest edge species, sometimes found

within forests, either where the canopy is fairly open or where there has been a large

gap," is characterized by Struhsaker (1997) as a late successional or old growth forest

species. This disparity may be due to site specific conditions so a second set of X2

analyses were done with Markhamia lutea defined as a forest interior tree.

Fruit and Flower Availability

Each month for 6 months (June, 2002-November, 2002), phenological data

(presence of ripe and unripe fruit and fruiting/flowering intensity) were collected from all

transects, including the dbh of the fruiting trees (> 5 cm dbh) important in the diet of

either species within 10 m of either side of the transects. The fruiting intensity was

determined by assigning each tree to a numerical class (1-5) based on the fruit density in

the crown, with each class indicating about twice the density as the next lower class and 3

being the average fruit density of the species. Because basal area is strongly correlated

with crown size, the basal area of each tree was multiplied by a factor (1/4, 1/2, 1, 2, or 4,

respectively) based on the fruit density class to estimate the total fruit load on each tree.

The fruit loads were summed (trees having both ripe and unripe fruits were considered to

have 1/2 ripe and 12 unripe fruit) and divided by the total area sampled to produce a

standardized monthly fruiting intensity for each species and transect for both ripe and

unripe fruit.

The only flowering species eaten frequently, Premna angolensis, flowered during

the study, but synchronously for a short duration and was missed by the surveys.









Results

Primate Densities

L. albigena was found to have the lowest average density over all the transects with

C. mitis the next lowest; however, densities varied between transects. Cercopithicus

ascanius (Redtail monkey) and Piliocolobus tephrosceles (Red colobus) had the highest

overall densities (Table 2-2).

The density of C. mitis on the transects was correlated most strongly with total food

tree basal area density, but was also correlated positively with fruit tree basal area density

and leaf tree basal area density. C. mitis density was not positively correlated with

average summed standardized fruiting intensity, total tree basal area density, or with

flowering tree basal area density (Table 2-3). The C. mitis index of relative density

(groups observed/km walked) (Table 2-4) was also correlated positively with total food

tree basal area density, fruit tree basal area density, and leaf tree basal area density but

not with average summed standardized fruiting intensity, total tree basal area density, nor

with flowering tree basal area density (Table 2-3).

L. albigena density was not positively correlated with average summed

standardized fruiting intensity, total tree basal area density, total food tree basal area

density, fruit tree basal area density, nor with invertebrate foraging tree basal area density

(Table 2-3). Similarly, the L. albigena relative index of density (groups observed/km

walked) (Table 2-4) was not positively correlated with total tree basal area density, total

food tree basal area density, fruit tree basal area density, nor with invertebrate foraging

tree basal area density. However there was a marginally significant positive correlation

between the relative index of mangabey density and average summed standardized

fruiting intensity (Table 2-3).









Table 2-2. Group densities of the arboreal monkeys of Kanyawara by transect. 1 Field
Station is the portion of transect 1 running through the Makerere University
Biological Field Station. 1 Forest is the portion of transect 1 that is in the
forested area. Transect 1 is the entirety of the transect. The means were
calculated with all of Transect 1. Presence noted with opportunistic
observations. tPresence noted during surveys but at distances of >24 m.


Group densities (groups/km2)
Species
Transects Sei
C. mitis L. albigena C. ascanius C. guereza P. tephrosceles
1 0.0 0.0 4.1 11.6 15.8
1
Field 0.0 0.0 6.8 25.5 15.3
Station
1
e1 0.0 0.0 2.3 2.3 16.1
Forest
2 5.9 0.0* 24.8 10.6 16.5
3 5.3 3.9 14.4 3.9 7.9
4 5.2 2.1 8.3 8.3 8.3
5 3.4 0.9 7.7 4.3 6.8
6 0.0V 1.6 3.9 0.8 2.3
7 1.1 1.1 2.8 1.7 7.2
8 0.5 0.0W 6.4 3.5 4.4
Mean 2.7 1.2 9.1 5.6 8.7


Table 2-3. One-tailed Pearson's correlations between primate densities and relative
densities and the basal area densities of various tree categories and fruit
availability (average summed standardized fruiting intensity). Positive
correlations significant at a = 0.05 are in bold font. Positive correlations


gis nificant at a


0 10 are underlined I 8


Basal area C. mitis density C. mitis relative L. albigena L. albigena
density and density density relative density
fruit avail. r p r p r p r p
Total tree 0.306 0.231 0.371 0.183 0.032 0.470 0.042 0.406
Food tree 0.745 0.017 0.824 0.006 -0.159 0.353 -0.126 0.383
Fruit tree 0.662 0.037 0.741 0.018 0.212 0.307 0.134 0.376
Leaf tree 0.630 0.047 0.685 0.031 -
Flower tree -0.519 0.094 -0.572 0.069 -
Invert. tree -0.441 0.137 -0.304 0.232
Fruit avail. 0.239 0.285 0.237 0.289 0.399 0.164 0.596 0.060










Table 2-4. Groups of the arboreal monkeys of Kanyawara sighted/km walked by transect during
the monkey surveys. 1 Field Station is the portion of transect 1 running through the
Makerere University Biological Field Station. 1 Forest is the portion of transect 1
that is in the forested area. Transect 1 is the entirety of the transect. Means were
calculated with all of Transect 1. Presence noted with opportunistic observations.
Relative group density index (groups sighted/km walked)
Transects Species
C. mitis L. albigena C. ascanius C. guereza P. tephrosceles
1 0.0 0.0 0.36 0.92 1.18
1 0.0 0.0 0.57 1.63 1.31
Field Station
Ss 0.0 0.0 0.22 0.44 1.10
Forest
2 0.34 0.0* 1.25 0.85 1.30
3 0.25 0.19 0.76 0.32 0.76
4 0.30 0.10 0.60 0.60 0.60
5 0.16 0.12 0.57 0.29 0.70
6 0.04 0.11 0.37 0.04 0.22
7 0.08 0.19 0.29 0.16 0.59
8 0.05 0.07 0.40 0.24 0.36
Mean 0.15 0.10 0.58 0.43 0.71


Forest Composition

The 97.5% basal area density confidence intervals of the tree types (all trees, C.

mitis fruit, leaf, flower and total food trees and L. albigena fruit, invertebrate foraging

and total food trees) of the fragments and the forest transects never overlapped and were

separated by large gaps, except in the case of L. albigena invertebrate foraging trees.

Transect 1 and the forest part of transect 1 had higher basal area densities than the

fragments with regards to every category. They had lower basal area densities than the

other transects in every category except C. mitis flower trees, L. albigena total food trees,

and L. albigena invertebrate foraging trees (Table 2-5).

The 10 species with the highest basal area densities were more often exotic

(Eucalyptus sp.) (df = 1, X2= 5.6, p < 0.025), edge/savanna species (excluding

Eucalyptus sp.) (df = 1, X2 = 5.1, p < 0.025), or both (df = 1, X2 = 6.0, p < 0.025) in the

fragments than in the forest transects even with the inclusion of all of transect 1 and










Table 2-5. The basal area densities (cm2 tree basal area/m2 land) of trees > 10 cm dbh in transect
1 and the forest section of transect 1 compared to the 97.5% confidence limits of the
fragments and the other transects.
Fragments:
Fragments: -Transects 2-8:
Upper 97.5% Transect 1 Lower 97.5
Transect 1 Lower 97.5%
confidence limit forest only confidence limit (mean)
confidence limit (mean)
(mean)
All trees (10.4) 14.9 22.0 30.4 30.8 (36.7)
Total C food (0.6)0.8 2.2 3.6 10.7 (14.5)
trees
C. mitis fruit trees (0.3) 0.6 0.9 1.4 6.0 (8.4)
C. mitis leaf trees (0.2)0.3 0.9 1.5 3.6 (5.7)
C. mitis flower (0.001) 0.002 0.4 0.7 0.1 (0.4)
trees
Total L. albigena (1.1)2.8 6.1 10.2 8.7(10.8)
food trees
L. albigena fruit
albigena fruit (0.2) 0.3 1.4 2.2 6.5 (8.7)
trees
L. albigena
invertebrate (0.9)2.5 4.8 8.0 0.6(2.1)
foraging trees

Table 2-6. The top ten tree species by basal area density of individuals > 10 cm dbh in the forest
fragments and the transects characterizing the main forest block near Kanyawara,
Kibale National Park. Entries in bold indicate an important species in the diet of C.
mitis. Underlined entries indicate an important species in the diet of L. albigena.
*Described as occurring in forest edges or savannas by Hamilton (1991). tExotic
Bugemba (4.7 ha) Durama (CK) (4.9 ha)
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (cm2/m2) area

Neoboutonia melleri 3.41* 34.8% Pseudospondias 1.98 21.9%
microcarpa
Ficus dawei 0.94 9.6 Aningeria altissima 0.62 6.8
Mitragyna 0.70 7.2 Ficus vallis 0.56 6.3
rubrostipulata

Prunus africana 0.63* 6.4 Mitragyna 0.56 6.2
rubrostipulata
Parinari excelsa 0.43 4.4 Prunus africana 0.55* 6.1
Mystroxylon
Mystroyon 0.41* 4.2 Mimusops bagshawei 0.55 6.1
aethiopicum
Symphonia globulifera 0.35 3.6 Rauvolfia sp. 0.44 4.9

Sapium ellipticum 0.34* 3.5 Macaranga 0.33 3.6
schweinfurthii
Fagara sp. 0.34 3.4 Diospyros abyssinica 0.33 3.6
Eucalyptus sp. 0.27_ 2.7 Strombosia li, tl i 0.30 3.3
Top ten total 7.82 79.8% Top ten total 6.22 68.9%










(Table 2-6. Continued)
Kikoltrading (6.2 ha) Kiko2tea office (5.0)
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (m2/m2) area
Neoboutonia melleri 0.59* 30.3% Eucalyptus sp. 0.75' 28.0%
Eucalyptus sp. 0.55 28.2 Neoboutonia melleri 0.31* 11.8
Sapium ellipticum 0.14* 6.9 Sapium ellipticum 0.30* 11.3
Cleistanthus
Erythrina abyssinica 0.13* 6.7 Clestanthus 0.21 7.7
polystachyus
Maesa lanceolata 0.12* 6.2 Symphonia 0.18 6.8
globulifera
Alangium chinense 0.07* 3.4 Erythrina abyssinica 0.16* 6.2
Macaranga 0.07 3.4 Macaranga 0.13 4.7
schweinfurthii schweinfurthii
Mitragyna 0.06 2.9 Prunus africana 0.12* 4.4
rubrostipulata
Albizia sp. 0.05* 2.5 Alangium chinense 0.07* 2.5
Croton macrostachyus 0.04* 2.1 Bridelia bridelifolia 0.06* 2.3
Top ten total 1.81 92.5% Top ten total 2.28 86.0%

Kiko3school fragment (1.7) Kiko4church (1.2 ha)
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (m2/m2) area
Neoboutonia melleri 2.64* 29.0% Parinari excelsa 9.39 32.2%
Eucalyptus sp. 1.44T 15.8 Eucalyptus sp. 4.80W 16.4
Sapium ellipticum 1.36* 15.0 Fagara sp. 2.20 7.5
Parinari excelsa 0.78 8.6 Neoboutonia melleri 2.16* 7.4
Macaranga
achenf 0.69 7.6 Aningeria altissima 1.81 6.2
schweinfurthij
Erythrina abyssinica 0.37* 4.1 Newtonia buchananii 1.71 5.9
Bridelia bridelifolia 0.32* 3.5 Ficus capensis 1.12* 3.9
Markhamia lutea 0.30* 3.3 Symphonia 0.94 3.2
globulifera
Prunus africana 0.17* 1.9 Prunus africana 0.81* 2.8
Mitragyma 0.13 1.4 Erythrina abyssinica 0.58* 2.0
rubrostipulata
Top ten total 8.19 90.2% Top ten total 25.5 87.3%










(Table 2-6. Continued)
Rutoma (Big) (4.9 ha) Rwaihamba (2.4 ha)
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (m2/m2) area
Aningeria altissima 1.15 24.7 Eucalyptus sp. 8.84Z 55.9%
Pseudospondias 0.71 15.3 Syzyguim sp. 2.41 15.3
microcarpa
Ficus dawei 0.35 7.5 Neoboutonia melleri 0.65* 4.1

Prunus africana 0.32* 6.9 Pseudospondias 0.64 4.0
microcarpa
Diopyros abyssinica 0.20 4.2 Ficus dawei 0.57 3.6
Strombosia fili__, 0.19 4.1 Polyscias fulva 0.56* 3.6
Fagara sp. 0.18 3.9 Ficus vallis 0.42 2.7
Cordia abyssinica 0.17* 3.6 Markhamia lutea 0.29* 1.9
Mitragyna 0.12 2.6 Croton 0.28* 1.8
rubrostipulata macrostachyus
Ficus brachylepis 0.12 2.6 Cordia abyssinica 0.19* 1.2
Top ten total 3.51 75.4% Top ten total 14.9 94.1%

Transect 1 Transect 1 forest
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (cm2/m2) area
Parinari excelsa 4.75 21.6% Parinari excelsa 7.96 26.2%
Funtumia latifolia 1.97 9.0 Funtumia latifolia 3.20 10.5
Olea welwitschii 1.73* 7.9 Olea welwitschii 2.88* 9.5

Prunus africana 1.58* 7.2 ( j 2.20 7.2
gorungosanum
1.31 6.0 Strombosia 1.I. li, 7 1.71 5.6
gorungosanum

Strombosia b1it7 1.02 4.6 Trilepsium 1.34 4.4
madagascariense
Maesa lanceolata 0.94* 4.3 Doipyros abyssinica 1.22 4.0
Trilepsium
lepsi 0.80 3.6 Prunus africana 1.06* 3.5
madagascariense
Diospyros abyssinica 0.73 3.3 Mimusops bagshawei 0.92 3.0
Mimusops bagshawei 0.56 2.5 Newtonia buchananii 0.86 2.8
Top ten total 15.4 70.1% Top ten total 23.3 76.7%










(Table 2-6. Continued)
Transect 2 Transect 3
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2/m2) area (cm2/m2) area
Ficus exaserata 7.04 14.9% Celtis durandii 4.94 14.7%
Strombosia hld. it,, 5.74 12.2 Markhamia lutea 4.65* 13.8
Funtumia latifolia 4.70 10.0 Strombosia ^d /),c k i 4.47 13.3
Markhamia lutea 4.68* 9.9 Funtumia latifolia 3.95 11.7
Celtis durandii 4.00 8.5 Diosyros abyssinica 3.15 9.4
Celtis africana 3.13 6.6 Olea welwitschii 1.66* 4.9
Olea welwitschii 2.53* 5.4 Celtis africana 1.64 4.9
Parinari excelsa 2.14 4.5 Blighia i 1.44* 4.3
Ficus polita 2.08 4.4 Mimusops bagshawei 1.19 3.6
Diospyros abyssinica 1.81 3.8 Trlepsium 1.18 3.5
-madagascariense
Top ten total 37.8 80.1% Top ten total 28.3 84.1%

Transect 4 Transect 5
Basal % of Basal % of
area total area total
Species Species
density basal density basal
(cm2m2) area (cm2m2) area
Markhamia lutea 7.25* 21.6% Strombosia .l. h., ,d i/ 5.63 15.7%
Ficus exasperata 4.78 14.3 Parinari excelsa 3.95 11.0
Diopyros abyssinica 3.61 10.8 Aningeria altissima 3.81 10.6
Celtis durandii 3.35 10.0 Dombeya kirkii 2.69 7.5
Olea welwitschii 3.10* 9.2 Celtis durandii 2.63 7.3
Strombosia Ldq<,,H 2.48 7.4 Funtumia latifolia 1.92 5.3
Funtumia latifolia 2.23 6.7 Chaetacme aristata 1.67* 4.7
Trilepsium
sci 1.14 3.4 Cordia abyssinica 1.58* 4.4
madagascariense
Celtis africana 0.78 2.3 Mimusops bagshawei 1.56 4.3
Ehretia cymosa 0.48* 1.4 Albia 1.19* 3.3
grandebracteata
Top ten total 29.2 87.1% Top ten total 26.6 74.2%










(Table 2-6. Continued)
Transect 6 Transect 7
Basal % of total Basal % of
% of total
area area total
Species basal Species
density density basal
(cm2Im 2 area (cm2/m2) area
Funtumia latifolia 5.81 13.7% Celtis durandii 5.77 13.8%
Strombosia ,- .. ill.. 1 5.04 11.9 Strombosia ,- .. ill.. 1 4.93 11.8
Parinari excelsa 4.47 10.6 Parinari excelsa 4.25 10.2
Ficus natalensis 3.77 8.9 Olea welwitschii 3.01* 7.2
Celtis durandii 3.36 7.9 Aningeria altissima 2.68 6.4
(/iol. i-,h) M 2.99 7.1 Diospyros abyssinica 2.21 5.3
gorungosanum
Diospyros abyssinica 2.21 5.2 Funtumia latifolia 1.90 4.6
Aningeria altissima 1.78 4.2 Uvariopsis congensis 1.51 3.6
Blighia unijugata 1.67* 3.9 Ficus dawei 1.48 3.6
Markhamia lutea 1.46* 3.5 Symphonia globulifera 1.44 3.4
Top ten total 32.6 77.0% Top ten total 29.2 70.1%

Transect 8
Basal
% of total
area
Speciesarea basal
density
(c m2I 2 area
Diospyros abyssinica 4.97 22.0%
Albizia grandebracteata 3.51* 15.5
Celtis africana 1.98 8.8
Celtis durandii 1.89 8.4
Chaetacme aristata 1.60* 7.1
Dombeya kirkii 1.23 5.4
Cordia abyssinica 1.21* 5.4
Markhamia lutea 1.12* 5.0
Mimusops bagshawei 1.13 5.0
Ficus natalensis 0.94 4.1
Top ten total 19.6 86.8%


transect 8 (Table 2-6). A second set of analyses with Markhamia lutea defined as an

interior species confirmed the differences between the fragments and transects


(edge/savanna species excluding Eucalyptus sp.: df = 1, X2 = 6.8, p < 0.01;

edge/savanna/exotic species: df = 1, X2= 10.2, p < 0.01).

Fruit Availability

Transect 1 had several species fruit during the study (e.g. one small Monodora

myristica, several large Parinari excelsa, and two individuals in a grove of Diospyros

abyssinica), but none of the important dietary species for C. mitis fruited on this transect.










For L. albigena, the only important tree species to fruit was D. abyssinica, which fruited

from July until at least November, leaving May and June with no available fruit from

important species. For both primate species, transects 3 and 7 had the highest fruiting

intensity overall, with transect 2 having a large pulse from a single Ficus exasperata in

November. Celtis durandii was the only fruit available in every month in all transects

(except of course for transect 1) (Figure 2-2 and Figure 2-3).


Transect 2


Ila .1*


.4
^b *
Ci C)
bJ b
0) a c


May


s b 'e ',e

i ii i '
d: o i d
0 0
U) U) -


June July Aug.


Sept.


Nov.


Figure 2-2: Standardized fruiting intensity for species whose fruits comprise > 4% of the annual diet of C.
mitis. Transect 1 is not shown as no important trees fruited on this transect during the study.
Note the consistent fruiting of C. durandii in all months and transects and the pulse of a single
F. exasperata tree in transect 2. The F. exasperata pulse on transect 7 resulted from two
individuals whose fruiting overlapped in September. Solid black bar sections indicate ripe
fruit. Striped areas indicate unripe fruit. Fruit was present but at extremely low intensity.










Transect 3


May


c May June July Aug.
May June July Aug.


Oct. Nov.

Oct. Nov.


Figure 2-2. (Continued)


6
Cu
June


6
Cu
July


6
Cu
Aug.


Cu
6
Oct.


Transect 4













m" n P 1 1


Cu

Nov.
Nov.


Cu
e
Sept.


Si

Sept.


i I I I I i,












Transect 5


0 0 0 0 0 0 0


May June July Aug. Sept. Oct. Nov.




Transect 6


May



May


June


July


Aug.


Sept.


Figure 2-2. (Continued)


Oct. Nov.


L 7 m M pm ^















9

8

77

6-

5-




C 3



1

0







May June July Aug. Sept Oct. ov.




10
Transect 8
9

8 -

7 7

6 -

45
LI



3 3

!) 2

m a* P

0







May June July Aug. Sept. Oct. Nov.
C10 C------C------C------C------C---b--C--b
Transect
9( ------------------------------------

May June---July ---Aug.---Sept-----Oct.---Nov.-


Figure 2-2. (Continued)













Transect 1


May June July Aug. Sept. Oct Nov.




Transect 2


May

May


'b


June


'b


July


Aug.


Aug.


'b .
bS (
d
S i
Sept.


0 i LC
O ct
Oct.


;bS (C (

0 i LC

Nov.


Figure 2-3: Standardized fruiting intensity for species whose fruits comprise > 4% of the annual diet of L.
albigena. Note the consistent fruiting of C. durandii in all months and transects (except
transect 1), the pulse of a single F. exasperata tree in transect 2, and the synchronized fruiting
of D. abyssinica in the latter part of the study in all transects. The F. exasperata pulse on
transect 7 resulted from two trees whose fruiting overlapped in September. Solid black bar
sections indicate ripe fruit. Striped areas indicate unripe fruit.












U
Transect 3
9

8

7

6

5

4

3

2

1

0







May June July Aug. Sept. Oct. Nov.



10
Transect 4
n


-~


2



LL
May


June


July


Aug.


Q. :)
Q

Sepo
Sept.


Figure 2-3. (Continued)


Nov.











Transect 5


May June July Aug. Sept. Oct. Nov.



Transect 6


May June


CU Cu^ ( Q Q^ (
o o o

bQ
bc c
* -Q -Q- Q-
c~ C C U ) U ) CUC) C


July


Aug.


Sept.


Figure 2-3. (Continued)


Nov.


I n, W, m,,, i,












10
Transect7
9

8 -

7 7





NU 4

.3 -
C 3

U2 -

1

0





Q



10
Transect 8
9

8

7-7

6

5
L.

. 4

3 3

B 2

1

0

Cs Cs s CO o C U c C U s (o -s Co C
c C c c c 2 c c c5

d U mU mU 'mU 'm U ) )



May June July Aug. Sept. Oct. Nov.


Figure 2-3. (Continued)









Discussion

Primate Densities

C. mitis and L. albigena groups were never observed or reported to have been in the

field station or the adjoining forest, despite the constant presence of potential observers

over many years. The other primates are regularly seen in this area. The correlations

between C. mitis density and food tree basal area density, fruit tree basal area density, and

leaf tree basal area density suggest that, in the case of this species, and potentially other

generalists and folivores, habitat quality might be indexed dependably and quickly with a

knowledge of local diet and corresponding basal area measurements of only a few

important food species (see also Siex and Struhsaker 1999). It is notable that total food

tree basal area density is more highly correlated with monkey density than that of any

dietary component alone. This is not surprising considering the generalist nature of C.

mitis.

There are several possibly contributing factors involved in the lack of correlations

between the basal area densities of the tree categories and L. albigena density. As

mentioned above, liana fruit was the second most important mangabey food item of

Olupot's (1994) study, but is not accounted for here, weakening the predictive power of

the food plant index. Additionally, L. albigena home ranges are large (410 ha), an order

of magnitude larger than C. mitis home ranges (61 ha) (Struhsaker and Leland 1979), and

the scale of the system of transects is probably not large enough to match the scale at

which the mangabeys travel. One of the reasons for such wide ranging in L. albigena is

their exploitation of rare, intense fruiting events including those produced by widely

separated large Ficus spp. (Waser 1975). The area of the transects is most likely

insufficient to accurately represent the occurrence of huge, widely-spaced, and important









figs. Adjusting the size of a study area to include multiple home ranges could alleviate

this problem.

The average summed standardized fruiting intensities were not anticipated to have

a strong relationship with the densities of either monkey species because of the

assumptions inherent in calculating a fruit availability index for each transect over a long

period. By summing all fruiting species, it is necessarily assumed that the normal fruit

load on individual trees of a species is equivalent to that of any other species in terms of

value to the frugivores. It is also assumed that the value of the fruit does not change from

one month to the next. Given that the assumption of normal fruit load interchangability is

almost certainly violated and that the assumption of unchanging monthly fruit value is

violated (Chapter 3), the near significance of the correlation between number of

mangabey groups seen per km walked during the primate surveys and average summed

standardized fruiting intensity is as much as can be expected. In view of the

specialization of L. albigena on concentrated fruiting events compared with the more

generalized diet of C. mitis, it is logical that habitat use by L. albigena would be more

easily related to fruit availability and habitat use by C. mitis would be more related to the

occurrence of all types of food trees (see also Beeson 1989). Fruit availability might be

an acceptable indicator of habitat quality for fruit specialists like L. albigena, but given

the temporal variation in fruit load, long-term studies would be necessary.

Forest Composition

The fragments have lower basal area densities of trees in general and a higher

proportion of the top ten trees by basal area density is composed of edge/savanna trees,

making the fragments different from Kibale forest in more ways than just being

physically isolated. This in itself does not imply that the fragments are unsuitable habitat









for resident or transient groups of C. mitis or L. albigena, as both are found widely

through equatorial Africa in diverse habitats (Kingdon 1997). Similarly, the fact that the

fragments are much smaller in area than C. mitis and L. albigena home ranges at

Kanyawara does not automatically exclude these species as other primates, C. ascanius

and P. troglodytes, are able to move among fragments to meet their needs (Chapman et

al. 2003b). While this may have an impact on the use of the fragments by C. mitis, which

rarely descends to the ground (Kingdon 1974, Rudran 1978, Beeson 1989) or crosses

open areas (Devos and Omar 1971, Fairgrieve 1995, Lawes 2002), L. albigena has been

reported to do as well by utilizing several fragments as inhabiting contiguous forest

(Tutin et al. 1997). Additionally, there are larger fragments not available for the analysis

of forest composition, Kasisi (130 ha) and Lake Mwamba (28.7 ha), that also lack both

species (Onderdonk and Chapman, 2000) despite being large enough to contain the home

range of at least one C. mitis group (25-44 ha; Butynski, 1990), though not the entire

home range of an L. albigena group (441 ha; Waser 1984). In any case, neither

differences in forest structure nor size apply to the area of transect 1 since it could easily

be used by a group from the adjoining areas. The fact that the unused area could be used

in concert with populated adjacent areas and therefore would not be required to provide

all the resources necessary to support a group indicates that the unused area does not

produce resources at high enough levels to attract even casual use.

Examining the basal area densities of the important food trees for both C. mitis and

L. albigena it becomes clear how different the fragments are from the inhabited forest.

The area of transect 1 generally falls between the two. The only exceptions to this

pattern are given by minor (though not necessarily unimportant) dietary components. In









the case of C. mitis, transect 1 has the same basal area density of flower-producing

species as the mean of the other transects. Excluding the field station simply increases the

flower-producing basal area density. This suggests that lack of flower-producing trees

can be ruled out as a possible cause for the avoidance of this area by C. mitis.

Likewise, the total L. albigena food tree basal area density is only different from

the other transects when the field station is included. The forest alone fits in well with

the rest of the transects in this characteristic and actually had a higher basal area density

of invertebrate foraging species than any other transect. However, these seem to be

driven by the extremely high basal area density of Parinari excelsa (the one important

invertebrate foraging species for L. albigena) in the area of transect 1. The high density

of P. excelsa in one of the fragments also drives the overlap of the fragment and transect

confidence intervals for invertebrate foraging trees. This suggests that invertebrate

foraging substrate is, in this case, not limiting the range ofL. albigena.

The hypotheses that the observed pattern of C. mitis and L. albigena occurrence is

driven by the basal area density of fruit-producing trees or, in the case of C. mitis, leaf-

producing trees or a combination of the two, are supported by these findings. It is not

possible to rule out this pattern being driven by the basal area density of all tree species,

but this seems unlikely given that both monkey species use areas around Kanyawara that

lack a closed canopy (the transect 1 forest has a closed canopy).

The observed lack of fruiting basal area is probably not due to a lack of fruit

dispersers in the fragments (as suggested by Cordeiro and Howe 2001) as large bodied

frugivores are still present (Onderdonk and Chapman 2000), and almost certainly not due









to a lack of dispersers in the area of transect one. In this case, it is the lack of fruiting

trees that leads to a lack of frugivores and not the other way around.

There is ample evidence to conclude that the fragments are poor habitat and do not

contain enough food trees to support their use by C. mitis or L. albigena. The fact that

transect 1 has higher food tree basal areas than the fragments and yet is completely

unused by either species though they inhabit the adjoining forest indicates how poor the

fragments really are. This suggests that even if the matrix between the fragments and the

forest were conducive to dispersal by C. mitis and L. albigena, the fragments would

likely be as unused as they are currently. The importance of human impacts on the

fragments is indicated by Eucalyptus spp. being one of the top ten trees in 8 of the 12

fragments. If native trees important to frugivores were found to be suitable alternatives

to Eucalyptus spp. for human use and planted instead, the habitat value of the fragments

might be greatly increased even with no change in extraction practices. Though

managing the smaller fragments for C. mitis or L. albigena would be overly optimistic,

wildlife species that currently depend on the fragments could be expected to benefit from

management aimed at increasing native forest trees at the expense of Eucalyptus spp. and

edge/savanna trees that are common in the surrounding matrix.

Moving management away from heavy use of Eucalyptus spp. for the purpose of

improving wildlife habitat would be difficult given the long establishment of Eucalyptus

spp. use in Uganda, the general focus of extension programs on a few exotic species

while ignoring natives (Katende et al. 1995), the benefits of Eucalyptus spp., including

quick production of above-ground biomass (Harmand et al. 2004) and coppicing ability

(Little and Gardner 2003), and the widespread hostility towards wildlife because of crop-









raiding (pers. obs.). However, the recognition and exploitation of the tourism value of

primates and the forest fragments themselves (shown by multiple ecotourism projects in

the area based on, or offering access to, the fragments) (pers. obs.), the common (albeit

sometimes reluctantly admitted) knowledge that not all species of wildlife are serious

agricultural pests (pers. obs., A. Lepp unpublished data, Hill 1997), and the problems

associated with Eucalyptus growth, like high water demands (Radersma and Ong 2004)

and allelopathic leaf litter (Bernhard-Reversat 1999), may provide valid motivations for

management change in some of the fragments.

Fruit Availability

With respect to the avoidance of transect 1 by C. mitis and L. albigena, even more

telling than the forest composition data, is the fruit availability data. Both species are

frugivorous with fruit forming the plurality, if not the majority, of their diets. The fact

that transect 1 had no important trees fruiting for C. mitis during the study and low levels

of fruiting by a single common species for L. albigena during only five of the eight

months suggests that the unused area does not produce enough fruit to attract either

species.

Transect 1: A Primate Perspective

The area represented by transect 1 is low quality habitat for both C. mitis and L.

albigena but this does not automatically exclude incidental use of or travel through the

area by either species. To understand the absolute lack of use by troops of these

monkeys, it is important to take a deeper look at their natural history.

C. mitis groups have been studied at Kanyawara since the early 1970's (Rudran

1978) and identified by consistent numerical labels (Butynski 1990) and inhabit stable

territories (although groups 4 and 5 seem to be switched between the 1970's and 1980's).









Even in this study, the areas where C. mitis observations were made match the home

ranges given by Butynski (1990). He characterized the Kanyawara C. mitis population as

being stable and it would be useful to ponder how important traditional home range

boundaries are in determining use. If home ranges are stable for three decades in a

relatively stable population, one wonders if the perturbations of group fission (Cords and

Rowell 1986, Lwanga 1987) or marked ecosystem change (logging, fire, etc.) are perhaps

necessary for the exploration and colonization of new areas.

It has been noted more than once that C. mitis home ranges tend to have obvious

visual landmarks as boundaries, such as exotic plantations, swamps, roads, streams, and,

of course, forest edges (Aldrich-Blake 1970, Rudran 1978, Butynski 1990). It is logical

that visually oriented primates, which do not use scent marks to define territories and

whose main territorial behavior (calls) indicates presence, not boundaries, might use

natural, easily seen habitat changes and features for boundaries. The area around transect

1 is cut off from the rest of the forest by a belt of swamp-marsh. It is this belt that

delineates the unused area and is the limit for C. mitis territories in Butynski (1990) and

sightings during this study (Figure 2-4). Thus, from the perspective of a C. mitis group,

the area of transect 1 is not simply an area of low food but an area of low food cut off

from the rest of the forest by a convenient and distinct boundary. To the collective eyes

of C. mitis, the area of transect 1 might appear to be a fragment outside (even if

contiguous with) the forest, more accessible than the other fragments, but not much more

desirable.

The case for L. albigena is less conclusive. They commonly use marshes and

swamps in other areas (Poulsen et al. 2001) and were seen in the middle of the swamp-









marsh belt during this study in a multispecies group (including C. guereza which visits an

open pool there occasionally). One of their major fruit trees (Diospyros abyssinica) did

fruit to a certain extent on transect 1 during the study, and their most important

invertebrate foraging tree has an extremely high basal area density there. It is possible

that Diospyros abyssinica rarely fruits in the area, rarely did so in the past, or that the

fruiting intensity is not worth the trip considering how common that fruit was throughout

the forest during the synchronous fruiting of Diospyros abyssinica. However, given the

use of fragments by L. albigena in other areas, the presence of at least some important

fruit, and their willingness to use swamps and marshes in this and other areas, L. albigena

is probably the more likely of these two frugivores to incorporate the area of transect 1

into a home range now or in the future as vegetation changes take place.

Summary

* The density of C. mitis was most strongly correlated with the basal area density of
total food trees and was also correlated with the basal area densities of fruit and leaf
providing trees, but it was not correlated with fruit availability, total tree basal area
density, or flower tree basal area density.

* The density of L. albigena was not correlated with the basal area density of any
class of trees or with fruit availability. However, number of mangabey groups seen
per km walked was marginally correlated to fruit availability.

* When comparing the occupied forest to the fragments, the fragments were found to
have lower basal area densities of all food tree classes except for L. albigena
invertebrate foraging trees. The unused area was found to have lower basal area
densities than the occupied forest areas for all tree classes except for C. mitis flower
trees, total L. albigena food trees, and L. albigena invertebrate foraging trees. The
unused area had higher basal area densities than the fragments in all classes.

* The fragments had higher frequencies of exotic and edge/savanna species than the
forest even with the edge transects included.

* The basal area density of food trees is an adequate habitat quality indicator in the

* case of C. mitis. The failure of basal area density of food trees to account for the









Figure 2-4: C. mitis territories in relation to the unused area of forest. Territories are
outlined in thick lines (redrawn from Butynski, 1990). The unused area of
forest is outlined with a white dashed line. Note that it is separated from
surrounding territories by a band of swamps. Roads are shown with thin lines.
The Makerere University Biological Field Station (MUBFS) is labeled. The
satellite image is a Quickbird (high resolution, 2.4 m) from DigitalGlobe, Inc.,
Longmont, CO, USA.


















:LM
tinoote









density of L. albigena may be related to scaling issues.

* The fragments outside Kibale National Park offer wretched habitat for these two
species and dedicated management to increase native fruit producing trees, while
potentially increasing the value of the fragments to other species, is unlikely to
impact C. mitis or L. albigena due to the size of most of the fragments and the
human dominated matrix surrounding them.

* The unused area had no important fruit for C. mitis produced during the study and
low levels of fruiting by a single important species during part of the study for L.
albigena. The other areas had variable fruiting levels with Celtis durandii
ubiquitously fruiting in all months.

* Because of naturally occurring boundaries around the area of transect 1 and
apparently stable home ranges, C. mitis groups probably view this area in the same
way as a low quality fragment contiguous with the forest.

* Because of L. albigena's use of swamps and the presence of an important fruit in
the unused area, mangabeys are the more likely species to use the area of transect 1.














CHAPTER 3
SEASONAL VARIATION IN THE QUALITY OF A TROPICAL RIPE FRUIT AND
THE EFFECT ON THE DIETS OF THREE FRUGIVORES

Introduction

Seasonality of climate and the corresponding temporally uneven distribution of

resources place hardships on wildlife, which can limit populations and act as major forces

in natural selection (Beeson 1989, Richardson 1991, Brown and Brown 2000, Fiksen

2000, Brugiere et al. 2002). Torpor, migration, and resource switching are some of the

common solutions to predictable periods of resource scarcity (van Schaik et al. 1993);

however, the effects of lean seasons can be exacerbated by climatic events (Hafner et al.

1994, Muri 1999) or human disturbance (Laurance and Williamson 2001).

While the effects of seasonality are more well known in temperate zones and dry

tropical zones, the less distinct seasons of tropical moist forests and the corresponding

vegetational changes have been long recognized (Janzen 1967, Karr 1976) with much

work being conducted describing the changing patterns of fruit availability (Sun et al.

1996, Chapman et al. 1999b, Larue et al. 2002, Schaefer and Schmidt 2002, De Walt et

al. 2003). Since the nutritional values of species differ, the nutritional makeup and the

availability patterns of each species can be combined to get a better idea of what seasons

and/or nutrients are potentially limiting for frugivores (Conklin-Brittain et al. 1998, Rode

et al. 2003). While there often appear to be periods of fruit or nutrient scarcity, previous

studies have assumed that the nutritional value of any given type of food remains

unchanged through time (Conklin-Brittain et al. 1998, Gupta and Chivers 1999).









Therefore many months or years of phenology and diet data are often combined with

nutritional data from fruit collected once. This is, in fact, an assumption inherent to any

study that ignores possible nutritional changes of dietary items over time, even those that

do not include nutritional analyses. If this assumption of constant nutritional qualities is

not a valid one, dietary choices of wildlife may appear random or illogical and the

seasonality of tropical moist forests may be seriously underestimated making the

understanding of foraging strategies much more difficult.

The objectives of this study were to explore the temporal variation in dietary

quality of a common fruit and investigate the relationships between frugivore

consumption and the availability and quality of the fruit. By focusing on the quality and

consumption of a common species I show the value of incorporating the possibility of

temporal variation of fruit quality in ecological study designs.

Methods

Phenological data and ripe fruit samples were collected over 6 months and

compared with rainfall (the main seasonal change) in the moist tropical forest of Kibale

National Park, Uganda (00 13' 0' 41' N and 300 19' 300 32' E). The forest is a

mature, mid-altitude, moist, semi-deciduous and evergreen forest with a mean annual

rainfall of 1734 mm (1990-2000), a mean daily minimum temperature of 15.50 C, and a

mean maximum daily temperature of 23.70 C (Chapman et al. 2003a). There are two

rainy seasons: March-May and September-November, with the later having the higher

rainfall.

One of the more important fruit bearing species in Kibale, in terms of both its

abundance and contribution to the diets of frugivores, is Celtis durandii (Struhsaker

1997). The fruit is a small, yellow, ovoid drupe (-0.8 cm) containing a single hard seed.









C. durandii is an extremely common and prolific fruiter in many areas of Kibale

(Struhsaker 1997, Barrett and Lowen 1998, Chapman et al. 1999b), and it was the only

species that was constantly fruiting throughout the study (May-November, 2002).

Though it has annual fruiting peaks, fruit is usually available to some extent in many, if

not all, months of any given year (Struhsaker 1997), making it a potentially important

staple or fallback food in times of general fruit scarcity. The fruit is eaten by many birds

(Struhsaker 1997) and is important in the diets of Chimpanzees (Pan troglodytes)

(Ghiglieri 1984), Blue monkeys (Cercopithecus mitis) (Rudran 1978), Redtail monkeys

(Cercopithecus ascanius) (Stickler 2004), Gray-cheeked mangabeys (Lophocebus

albigena) (Olupot 1994), Black-and-white colobus monkeys (Colobus guereza)

(Struhsaker 1978), Red colobus monkeys (Piliocolobus tephrosceles) (Struhsaker 1975)

and probably many other animals.

Ripe C. durandii fruit was collected each month for 6 months (June-November,

2002). A fruit was defined as ripe if it had changed from green to at least partially

yellow. The fruit was peeled off the seed, dried in a dehydrator (350 C), and stored in

plastic bags for shipment back to the University of Florida. The fruit was redried in a

drying oven at 500C overnight before being ground in a Wiley mill. The percent lipid

content of the fruit was determined with a microwave assisted extraction technique. The

dried, ground samples were weighed, placed in petroleum ether, and exposed to

microwaves in a MARS 5TM machine (CEM Corporation, Matthews, NC) to subject the

samples to extreme heat and pressure, allowing the lipids to be extracted quickly and

efficiently. The samples were rinsed in petroleum ether, dried, reweighed and the lipid

content calculated from the difference. The data for the highest lipid percentages are









probably underestimations as the more oily samples lost oil on the drying paper and

sample labels during storage and processing. Samples were also analyzed for total

ethanol soluble carbohydrates, saponins, acid detergent fiber, protein, and alkaloids, but

none of these showed an independent seasonal pattern so they are not discussed further.

Each month, phenological data (presence of ripe and unripe fruit, fruiting intensity,

and diameter at breast height (dbh) of fruiting trees) were collected, along 20m wide

fixed strip transects through the forest. The fruiting intensity was determined by

subjectively assigning each tree to a numerical class (1-5) based on the fruit density in the

crown with each class indicating about twice the density as the next lower class and 3

being the average fruit density of the species. Because dbh is strongly correlated with

crown size (Anderson et al. 2000) and fruit production (Chapman et al. 1992) within a

species, the basal area of each tree was multiplied by a factor (1/4, 1/2, 1, 2, or 4,

respectively) based on the fruit density class to estimate the total fruit load on each tree.

The fruit loads were summed (trees having both ripe and unripe fruits were considered to

have 1/2 ripe and /2 unripe fruit) and divided by the total area sampled to produce a

standardized monthly fruiting intensity for both ripe and unripe fruit.

The average lipid levels in each month were compared to the sum of the average

daily rainfalls of the concurrent and previous months (the best predictor months) using

linear regression.

To see if the fruit lipid level is related to the intake of C. durandii fruit by

frugivores, monthly dietary information was gleaned from existing sources (1 C. mitis, 1

C. ascanius, and 2 L. albigena data sets were available). The C. ascanius data set (C.

Stickler, unpub. data) was from the same location and time as the phenology and lipid









data and could be directly compared; however, the C. mitis data set was from 1973-4

(Rudran 1978) and the L. albigena data sets were from 1972-3 (Waser 1975) and 1992-3

(Olupot 1994) and could only be compared with lipid levels indirectly through the

summed average daily rainfall of the concurrent and previous months.

Rudran (1978), Waser (1975), and Olupot (1994) did not differentiate between ripe

and unripe fruit, but that should not have a major effect on the results as the lipid content

of the two fruit types is highly correlated (r=0.99, p < 0.001) and shows a similar pattern

of change as the ripe fruit with differences between ripe and unripe fruits of the same

month smaller than the seasonal differences. Waser's (1975) reported monthly

percentages of C. durandii in the diet include both fruit and insects. However, the use of

insects from C. durandii is negligible compared to the range of the fruit. Additionally,

Olupot (1994) only provides monthly dietary data from the top 5 foods. For 5 of 9

months, C. durandii is not in the top 5. Instead of assuming C. durandii fruit was not

eaten at all, a more conservative tack was taken and it was assumed that C. durandii fruit

was a close 6th in every such month (i.e., the fifth rank's percentage minus 0.01%). This

reduced the correlation coefficient and increased the p-value slightly compared with the

assumption of no consumption during months when C. durandii was not in the top 5.

Availability indices for C. durandii fruit from each study were also compared with

the importance of the fruit in each diet, except for Olupot (1994) where there was none

given and in the case of C. ascanius for which the average standardized fruiting intensity

of ripe C. durandii fruit for the same area as the C. ascanius study was used. All data

sets were analyzed with Pearson's correlations except for that of C. mitis which violated










the assumption of normality and therefore was analyzed using Spearman's ranked

correlations.

Results

The lipid levels in ripe Celtis durandii varied over two orders of magnitude among

months (0.3%-30.8%) and are predicted by the summed average daily rainfall of the

previous and concurrent months (r2= 0.91, p = 0.003, N = 6) (Figure 3-1).

40%


35%


30%


25% -


8 20%


15%


10%


5%


0%
0 2 4 6 8 10 12 14 16 18
Rainfall (mm)
Figure 3-1: The relationship between the monthly dry matter lipid content of ripe Celtis durandii fruit and
the summed average daily rainfalls of the concurrent and previous months at the Makerere
University Biological Field Station, Kibale National Park, Uganda for June 2002-November
2002 (r2=0.91, p=0.003, N = 6). The error bars represent +/- one standard error.

C. ascanius monthly consumption of ripe C. durandii fruit at the same time and

place is marginally correlated with the fruit lipid content (r= 0.88, p = 0.059, N = 4, one-

tailed) (Figure 3-2 a), and is also marginally correlated with rainfall (r= 0.78, p = 0.061,

n = 5, one-tailed) (Figure 3-2 b).













4U
1 35
2 30
. 25
20
* 15
S10
5
0
5



30
|25
S20
o 15
E
C 10
5
0




8 60
E 60
. 40
$ 30
S20

E 10
0


% 15% 25% 350,
Lipid content of ripe fruit


*








e *
0 5 10 15 20 2
Rainfall (mm)


e




+


46
40
5 36
S30
S 26
20
0 16
o 10
5
0



45
S 40
35
30
c 25
a 20
E 15
0 10

0
5


46
Z 40
36
&30
g 26
a 20
E 16
0 10
6
0


Rainfall (mm)


30 0 6 10 16 20 26
Availability (0-40)


Figure 3-2: The relationships between reported dietary use of Celtis durandii fruit and lipid content of the
fruit, average daily rainfall of the previous and concurrent months (a predictor of lipid
content), and fruit availability. All data sets are from Makerere University Biological Field
Station, Kibale National Park, Uganda. a) the relationship between lipid content and
Cercopithecus ascanius diet (C. Stickler, unpub. data) for August 2002-November 2002 (r=
0.88, p = 0.059, N = 4, one-tailed). b) the relationship between rainfall and C. ascanius diet
(C. Stickler unpub. data) from August 2002 December 2002 (r= 0.78, p = 0.061, N = 5,
one-tailed). c) the relationship between rainfall and Cercopithecus mitis diet (Rudran 1978)
from February 1973 May 1974 (r, = 0.56, p = 0.013, N = 16, one-tailed test). d) the
relationship between rainfall and Lophocebus albigena diet (Waser 1975) from May 1972 -
April 1973 (r = 0.64, p = 0.009, N = 13, one-tailed test). e) the relationship between rainfall
and L. albigena diet (Olupot 1994) from October 1992 June 1993 (r = 0.63, p = 0.034, N =
9, one-tailed test). f) the relationship between fruit availability and L. albigena diet (Waser
1975) from May 1972-April 1973 (r = -0.61, p = 0.035, N = 12, two-tailed test).


8 10 12 14 16 1
Rainfall (mm)






+d





0 5 10 1V
Rainfall (mm)


f *









The C. mitis monthly dietary percentage of C. durandii fruit described by Rudran

(1978) shows a correlation with rainfall (r, = 0.56, p = 0.013, N = 16, one-tailed test)

(Figure 3-2 c).

Waser's (1975) L. albigena data shows a correlation between the monthly

proportion of C. durandii fruit in the diet and rainfall (r = 0.64, p = 0.009, N = 13, one-

tailed test) (Figure 3-2 d), as does Olupot's (1994) data (r = 0.63, p = 0.034, N = 9, one-

tailed test) (Figure 3-2 e).

Neither the C. ascanius nor the C. mitis consumption of C. durandii fruit are

related to fruit availability. Waser's (1975) L. albigena data set has a negative correlation

between availability and consumption (r = -0.61, p = 0.035, N = 12, two-tailed test)

(Figure 3-2 f) with all species seemed to show decreases in availability occurring at the

same time as the highest consumption (Figures 3-3, 3-4, and 3-5).

Discussion

While nutritional components of leaves and fruits are known to vary between

species, location, position on the tree, stage of development, and/or time of day

(Woodwell 1974, Marquis et al. 1997, Fernandez-Escobar et al. 1999, Nergiz and Engez

2000, Klages et al. 2001, Chapman et al. 2003 a), and yearly fruit production and quality

have been shown to be dependent on environmental conditions of earlier months

(temperature, rainfall, amount of sunlight, etc.) (Sams 1999, Woolf and Ferguson 2000),

to my knowledge this is the first time nutritional quality of a continuously available ripe

fruit has been shown to be seasonally variable and related to rainfall. These results

suggest that the seasonality of tropical moist forests could be more marked than

previously thought. Though ripe C. durandii fruit was constantly available during the











study and at relatively high levels throughout the forest, the lipid resource base provided

by the fruit was negligible in the dry season but abundant in the rainy season.

Resource availability estimates based solely on observed fruit loads and nutritional

analyses of samples taken at one time may be misleading if seasonal changes in

nutritional values commonly occur. The ripe fruits of C. durandii in the wet and dry

seasons appear almost identical to human observers, only being plumper and juicier in the


1.8


1.6

E 1.4

E


S1.0


I 0.8


a 0.6


0.4


0.2


0.0


45%


40%


35%


30%


25% 4

20% -
o

15%


10%


5%


0%


May June July Aug. Sept. Oct. Nov. Dec.

Figure 3-3: Changes in the percentage of Celtis durandii ripe fruit in the diet of a Cercopithecus ascanius
group relative to changes in fruit availability and fruit lipid content. All data sets were from
the forest around Makerere University Biological Field station, Kibale National Park, Uganda,
2002.












2.5 60%
Fruit availability index
% of blue monkey diet
S Estimated % lipid
S* 50%
2.0 A

x / \
0g al 40% U




0.0 0\ %
2 0 PI-

0.5 10 r
/ \000

0.0 '000 0%

a) MOCOO CO Z a) UO a) M a) M M

Figure 3-4: Changes in the percentage of Celtis durandii fruit in the diet of a Cercopithecus mitis group
relative to changes in fruit availability and estimated fruit lipid content. The lipid content was
calculated from rainfall records. The peaks are probably unrealistically high but the basic
pattern should reflect reality. The diet data were from Rudran (1978), Makerere University
Biological Field Station, Kibale National Park, Uganda, January 1973 May 1974.

wet season. Despite this, the observed variation of lipid content is apparently important,

as indicated by foraging effort, to at least three frugivorous primates. During dry months

when the lipid levels were negligible, C. durandii fruit was rarely eaten even when most

abundant, but was heavily consumed when lipid levels increased. This may also be

important to birds as they can distinguish and prefer lipid content as little as 2% higher

(Schaefer et al. 2003). The fact that the data sets span four decades suggests that the

pattern holds true over long periods of time. Rudran (1978) also notes that another of his

groups (for which he did not present detailed data) had a similar pattern of C. durandii

fruit consumption as the group whose data are used here.













25 45
Availability
_% of mangabey diet --* 40
2 Calculated % lipid

20\
x 1535




-20
10 ._
/ 15
B F S Ki\al 20N5





5

0 0
Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr-
72 72 72 72 72 72 72 72 72 73 73 73 73

Figure 3-5: Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group
relative to changes in fruit availability and estimated fruit lipid content. The lipid content was
calculated from rainfall records. The diet data are from Waser (1975), Makerere University
Biological Field Station, Kibale National Park, Uganda, May 1972 April 1973.

Though the relationships were assumed to be linear for the purposes of analysis, the

true relationships between lipid content and rainfall, lipid content and consumption, and

consumption and rainfall are probably best represented by logistic functions. The

minimum lipid content of the fruit is limited both by 0 and by some low level needed for

cellular function, while the maximum should also be limited by physiologic mechanisms.

Consumption of the fruit is likewise limited by 0 as a minimum and by nutritional

demands for substances not found at high enough levels in C. durandii fruit as a

maximum. Even though it is easy to visualize S-shaped curves as appropriate for some of

the data sets, the small sample sizes, especially at the higher end of the ranges, make

fitting reasonable logistic curves questionable.










None of the other nutritional components in the C. durandii fruit or the fruit of any

other of the 9 species analyzed showed a significant relationship with rainfall. During

this study, other species flushed with fruit either synchronously or independently for one

to four months and tended to show slight increases in total ethanol soluble carbohydrates

(sugars) as the crops ripened (i.e., what was collected as "ripe" when the first fruits

started to ripen was not as "ripe" as the fruits at the end of the fruiting interval).




6O%
% Mangabey diet
Calculated % lipids
50% C.d. dietary maximum


40%
In /
30%





10%


0% %1
Oct-92 Nov-92 Dec-92 Jan-93 Feb-93 Mar-93 Apr-93 May-93 Jun-93

Figure 3-6: Changes in the percentage of Celtis durandii fruit in the diet of a Lophocebus albigena group
relative to changes in estimated fruit lipid content. The lipid content was calculated from
rainfall records. The diet data were from Olupot (1994), Makerere University Biological
Field Station, Kibale National Park, Uganda, October 1992 June 1993. Since only the top 5
foods were reported, the percent of the fruit in the diet from February on was assumed to be
the maximum possible (see text for full explanation). As 1992 was a very wet year, the lipid
content of the fruit is probably estimated at impossibly high levels.

However, this phenomenon was probably not the case for C. durandii as 1) it fruited

throughout the study with different trees coming into and going out of fruit at different

times, 2) any individual fruit probably does not stay ripe for more than a month, 3) the

unripe fruit showed a similar increase in lipid levels with rainfall, and 4) the amount of









variation in lipid content observed is not only over two orders of magnitude but greater

than the variation between ripe and unripe fruits of the same month.

This study did not cover the minor rainy season in March-May. Lipid levels

estimated with rainfall had two peaks a year, but consumption by the primates only

showed one major peak in a year (Figures 3-4, 3-5, and 3-6). If consumption is a good

indicator of lipid levels in general, it appears that lipid peaks may not be directly driven

by moisture availability, but instead may occur only once a year during the September to

November rains.

The negative correlation between fruit availability and monthly dietary importance

for Waser's (1975) data and the observation that the months of high usage occur at the

same time as drops in availability (Figures 3-3, 3-4, and 3-5) suggest that the availability

of the fruit may be driven by consumption and not the other way around as is usually

assumed. This may explain the observation that C. durandii has irregular flowering, but

annual fruiting peaks (Chapman et al. 1999b). Instead of annual fruit production peaks,

C. durandii may have consumer driven annual fruit load lows ultimately caused by

changes in lipid content. If this is true, it implies that competition for this resource is

intense for only a few months out of the year and is driven ultimately by resource quality.

Celtis durandii seems to have a strategy of producing fruit year round, even in the

difficult seasons, but altering the value of the reward provided to seed dispersers. This

strategy could be relying on several principles: i) either it "tricks" certain dispersers into

eating a fruit with low nutritional value, ii) is eaten and dispersed only as a last resort by

desperate frugivores when nothing else is available, or iii) is ignored by all frugivores

during the off-season and is dispersed by gravity or seed predators. Alternately, C.









durandii could be able to produce lipid rich fruits in all seasons, but the chances of

seedling survival might be low in drier seasons, thus the investment in more attractive

fruits at a time when seed dispersal could be most beneficial. In any case, it is clear that

the abundance of this fruit is not related to its value as food, and resource availability

studies based on an assumption of the constant quality of specific foods may be missing

an important component.

Primate diets are often diverse and variable with seemingly random changes that

appear inexplicable to the human researcher. However, the use of C. durandii fruit by at

least three species can be explained by the changing nutrient levels found in the fruit and

not by the changing abundance levels of the fruit. If this strategy of producing low

quality fruits during certain seasons or under certain conditions is shared by a number of

tropical tree species, it could go a long way towards explaining changing and enigmatic

dietary preferences in frugivores and be another indication that the tropics are more

seasonal than originally thought.














CHAPTER 4
CONCLUSION

The conservation of wildlife populations requires the ability to assess the quality of

remaining habitats, but it is often difficult to determine habitat requirements in species

that have temporally and spatially variable diets. An understanding of how edges and

fragmentation affects habitat quality is increasingly important in these days of wanton

habitat destruction and fragmentation. This study of C. mits and L. albigena was

undertaken to understand how habitat quality for these two species might be assessed

efficiently and evaluate the habitat potential of unoccupied forest fragments near the

intact forest.

This study showed that:

* The density of C. mitis was most strongly correlated with the basal area density of
total food trees.

* The number of mangabey groups seen per km walked was marginally correlated to
fruit availability.

* The fragments were found to have lower basal area densities of all food tree classes
than the occupied forest, except for L. albigena invertebrate foraging trees. The
unused area within the park was found to have lower basal area densities than the
occupied forest areas for all tree classes except for C. mitis flower trees, total L.
albigena food trees, and L. albigena invertebrate foraging trees. The unused area
had higher basal area densities than the fragments in all classes.

* The fragments had higher frequencies of exotic and edge/savanna species than the
forest even with the edge transects included.

* The basal area density of food trees is an adequate habitat quality indicator in the
case of C. mitis. The failure of basal area density of food trees to significantly
predict the density of L. albigena may be related to scale differences between the
study and home ranges of mangabeys or foraging strategy.









* The fragments outside Kibale National Park offer extremely poor habitat for these
two species and dedicated management to increase native fruit producing trees,
while potentially increasing the value of the fragments to other species, is unlikely
to impact C. mitis or L. albigena due to the size of most of the fragments and the
human dominated matrix surrounding them.

* The unused area had no important fruit for C. mitis produced during the study and
low levels of fruiting by a single important species during part of the study for L.
albigena. The other areas had variable fruiting levels with Celtis durandii
ubiquitously fruiting in all months.

* Because of naturally occurring boundaries around the area of transect 1 and
apparently stable home ranges, C. mitis groups probably view this area as a low
quality fragment contiguous with the forest.

* Because of L. albigena's use of swamps and the presence of an important fruit in
the unused area, mangabeys may be more likely than C. mitis to use the area of
transect 1.

* Many studies have assumed that the nutrient content of food items remains constant
over time. Here, I documented that the lipid content of Celtis durandii ripe fruit
varied tremendously during the study and was correlated with rainfall.

* The estimated variation of lipid content was shown to be correlated with the dietary
intake of C. durandii fruit in three primate species with four data sets.

* C. durandii fruit availability was a poor predictor of dietary use and seemed to be
driven by consumption which seemed in turn to be driven by lipid content.
Therefore common use of use of fruit availability as an index of resource
availability should be more carefully scrutinized.

This study has shown the validity of assessing habitat quality by quantifying the

basal area densities of locally important food trees for the generalist C. mitis. This

method quickly and easily measures a slowly changing (sans logging) aspect of the

environment and is therefore more efficient than evaluating habitat quality with indices of

fruit availability which may change quickly and seasonally and be only loosely related to

the availability of resources in the habitat. Basal area density was a worse predictor of

habitat use by L. albigena than fruit availability, perhaps because of the mismatch of









scales between home ranges and this study, or perhaps because these mangabeys

specialize on intense fruiting events.

The fragments outside Kibale National Park have extremely low habitat quality for

these species and would therefore be of little use to these monkeys either as habitat or

stepping stones connecting intact forest populations even if the matrix surrounding them

were conducive to dispersal. The unused area, though contiguous with used forest and

higher in habitat quality that the fragments, is probably avoided because of the

combination of population stability, convenient boundaries surrounding that area, and

poor habitat quality. It is likely that this area will be used by L. albigena before C. mitis.

The variation in the lipid content of C. durandii and the observation that lipid

content, and not availability, seems to drive consumption of this fruit indicates that

seasonality of resources might be more pronounced in tropical forests depending on the

strategies of component fruiting species. Dietary studies should be designed to account

for this possibility and assess resource availability and not just fruit availability.















APPENDIX
POTENTIAL FACTORS INFLUENCING HABITAT USE BY Cercopithecus mitis
AND Lophocebus albigena

With the complete avoidance of the area of transect 1 (an edge) by Cercopithecus

mitis and Lophocebus albigena in mind, many possible reasons for differential habitat use

especially with regards to edges were generated. Logistical considerations limited the

number of factors that could be examined thus many possibly important edge effects

were excluded from consideration, such as:

* Light levels
* Wind
* Disease
* Human impacts
* Predation
* Arthropod abundance
However, many other factors were examined:

* Temperature extremes
* Evaporative potential
* Tree species composition
* Fruit availability
* Nutritional quality (fruits and leaves)
* Arthropod foraging substrate availability
Unexamined Possibilities

Light Levels, Wind and Disease

Light, wind, and botanical diseases would affect the monkeys indirectly by either

altering the composition of the forest or the availability or quality of foods. However, it

is unlikely that these are important in this case as increased light levels and wind are

found on edges and open forest types and would therefore have a similar impact at least

in the immediate study area and exclude the monkeys from other areas, which is not









apparent.. A simian disease has been thought to be able to control the distribution of C.

mitis in other parts of Uganda (Haddow et al. 1947), but in that case there was an extreme

ecological gradient that controlled the distribution of the disease vector which is not true

here. Also, a disease would most likely have a larger or more temporary effect than that

observed. Therefore, these potential factors are not considered further.

Human Impacts

Since the edge in question borders on a field station, certain human impacts in the

area are relatively high. Even without direct persecution, some extremely sensitive

species may shy away from constant human presence. However, both C. mitis and L.

albigena groups in the area have previously been fully or partially habituated (depending

on the group) and generally appeared less excited when approached than some of the

other more heavily studied species (e.g., Piliocolobus tephrosceles) and are present in

rest of the forest area that is heavily used by both researchers and loggers. Additionally,

C. mitis are known to be pests in pine plantations and maize fields in other parts of Africa

(Devos and Omar 1971, Maganga and Wright 1991), and mangabeys are found in

Magombe Swamp (bordering Kibale on the east) that is surrounded by homes and small

scale cultivation (pers. observation). It is thus unlikely that human presence alone is

disturbing these monkeys enough to cause them to avoid this edge.

Hunting could extirpate these species from an area, but is unlikely to have done so

in this case. Firstly, the area is so small it is doubtful that hunters would focus their

efforts there and ignore nearby forest areas. Secondly, the surrounding tribe, the Batoro,

do not eat monkeys, their use of monkey skins, for royal regalia (Colobus guerza) and

dancing (possibly baboon (Papio anubis) but not C. mitis or L. albigena), is limited (pers.

observation; (Haddow 1952), and Rutoro lacks a specific name for C. mitis indicating









limited contact and cultural importance (Onderdonk and Chapman 2000, S. Katusabe

pers. com). Kingdon's (1974) reported name of C. mitis in Runyoro, a closely related

dialect, nkima, simply means "monkey" in Rutoro (S. Katusabe pers. com.). Onderdonk

and Chapman (2000) also state that there is no Rutoro name for L. albigena but it has

been reported to me asjugujugu, an onomatopoeia mimicking the mangabey "whoop-

gobble" (S. Katusabe pers. comm.). However, this word may be unknown or unused

except in places where contact is common (e.g., Magombe Swamp to the south). A

nearby people, the Bakonzo, who do regularly hunt monkeys, including C. mitis (pers.

observation; Haddow 1952), have been accused of poaching with firearms and possibly

lowering primate populations in Kibale before being expelled by the Batoro in 1962 (or

1964) (Struhsaker 1975, Ghiglieri 1984). However the only evidence of this seems to be

local stories (Ghiglieri cites hearsay and Struhsaker cites no one). It is true that starting

in 1962 intertribal tension boiled over into violence, arson (Ssembeguya et al. 1962), and

eventually a rebellion that lasted for decades in the Rwenzori Mountains (Ingham 1975,

Syahuka-Muhindo 1991), which could easily have ejected the Bakonzo from the Kibale

area (-10% of the population of the subcounty Rutete was Bakonzo in 1959 (Lubowa et

al. 1963)). However, the presence of hunters does not automatically cause low densities

and extirpation (specifically of C. mitis). For example, the cultural stronghold of the

Bakonzo, the Rwenzori Mountains, still supports its full compliment of primates in spite

of their continued exploitation (Lwanga 1987). Finally, there has been no reported

hunting in over 40 years so it is highly unlikely that hunting is responsible for the current

absence of C. mitis and L. albigena in this area or paucity of C. mitis in the southern

region of the park.









Predation

Predation can be important in controlling the abundance and range of some species

in certain areas. In Kibale, the main predator of monkeys is the African Crowned Eagle

(Stephanoaetus coronatus) (Mitani et al. 2001) with chimpanzees focusing on Red

colobus (Mitani and Watts 2001). Skorupa (1989) found that both C. mitis and L.

albigena made up 9% of the observed prey items of S. coronatus and C. guereza was the

most common at 39%. Struhsaker and Leakey (1990) found that C. guereza and L.

albigena were killed more than expected and adult males of all monkeys were preferred

much more than other age/sex classes. They concluded that eagle predation has little

impact on the monkey populations at Kanyawara, as did Rudran (1978).

If the presence of an edge increases the area in the canopy vulnerable to eagle

attack, monkeys might choose to avoid the edge. However, they do not avoid just the

trees on the edge of the forest, but also the trees surrounded by other trees in that area.

Additionally, the Mikana area just to the north of Kanyawara was heavily logged and still

retains an extremely broken and uneven canopy with most of the large trees well exposed

on all sides. This canopy morphology should expose primates to eagle predation more

than a closed canopy or edge but the area is frequently used by C. mitis and L. albigena

(pers. observation). Though C. ascanius, C. guereza, and P. tephrosceles frequent the

field station, are often in exposed trees, and surrounded by potential observers, eagles

have not been reported hunting there. This may be because they are deterred by the

presence of humans, making the field station potentially safer than the rest of the forest.

Because of these reasons predation was given no further consideration.









Arthropod Abundance

Since arthropods form a large proportion of the diet of both C. mitis and L.

albigena, any discussion of habitat suitability must address the availability of this

important resource. Unfortunately, due to the small size of most arthropods, the speed at

which they are ingested, the distance from the observers, and the immense diversity of

arthropods in tropical forests, the ingested species have never been identified for these

monkeys. In fact, even the general type of arthropod is extremely rarely identified by

direct visual means and identification of food items in fecal material is often difficult,

leaving inference from foraging style as the main method for identifying the general

arthropod resource utilized (Waser 1977, Struhsaker 1978, Poulsen et al. 2001).

C. mitis tends to use slower capture methods than the closely related Cercopithecus

ascanius, which are often seen stalking and pouncing on prey, implying that C. mitis

more commonly target slow or stationary arthropods (Cords 1986). They also forage on

and in specific substrates (foliage and epiphytes especially layers of moss and lichen) in

specific tree species (Rudran 1978, Struhsaker 1978, Cords 1986).

Similarly, Gray-cheeked mangabeys use foraging tactics appropriate for the capture

of slow or stationary prey (Waser 1984), probably because their size makes large and

quick movements dangerous in the canopy (Struhsaker 1978). However, because of their

size, they are also able to break apart dead wood and rip off large sections of dead bark

(Struhsaker 1978), exploiting a prey base that is not available to the smaller monkeys.

Since the specific arthropod items in the diets of C. mitis and L. albigena are

unknown and the difficulties involved in systematically collecting arthropods from the

correct foraging substrates are prohibitive, indices of available foraging substrates

epiphytess and dead wood) were used as substitutes.









Examined Possibilities

Temperature Extremes and Evaporative Potential

It is well known that forest edges have more extreme temperatures and are more

xeric than the interior (Burke and Nol 1998, Didham and Lawton 1999, Gehlhausen et al.

2000). These conditions could potentially affect the distribution of primates directly

through their own tolerances or indirectly through the tolerances of or indirect impacts on

their food, predators, competitors, or disease organisms (Murcia 1995). Kingdon (1974)

notes that captive C. mitis appear uncomfortable in hot sun and in the wild will often

abandon a sunny canopy as the morning warms up (also Aldrich-Blake 1970).

Furthermore, Kingdon (1974) mentions that L. albigena does not tolerate cold

temperatures in captivity. This means that temperature extremes near an edge could

affect both species, albeit in opposite ways. However, if the edge were directly affecting

the monkeys through temperature, they would be expected to only avoid the edge during

exceptionally hot or cold periods and their patterns of use should change diurnally and

seasonally, which they do not.

Temperature and evaporative potential monitoring stations were set up on two edge

to interior transects (one for each edge-interior set). A station consisted of a min/max

thermometer with the wire temperature probe tied to the evaporative potential monitor.

Evaporative potential monitors followed Didham and Lawton (1999), but were altered by

having reduced 1cm2 wick areas above the tubes and rain covers because they were left

out continuously and checked every other day. It was discovered that using natural

broom straws (palm midribs) to support the wicks of ten evaporative potential monitors

in the same location caused excessive variability in measurements (mean CV: 0.337), so

plastic was used (mean CV: 0.101) (Figure A-i).











25



20



15-



10 -I



5 -111



0



Date

Figure A-1: The intra- and inter-measurement period variation of 10 evaporation potential monitors from
the same location. The dotted line indicates when the wicks were first changed. The solid
line indicates when the wicks were changed and the wick supports were changed from broom
straws to plastic. The dashed lines indicate when the wicks were changed and the water was
changed instead of just being topped off. Note the high variation while using broom straws or
water that had been used with the broom straws and the reduction of variation once plastic
supports were used with regular water changes.

Considering the possible indirect affects of both temperature and evaporative

potential (as they are tightly interrelated), the expectation that the unused edge varies

substantially from both the utilized interior and the utilized edge is not met. In fact

topography was found to have a large effect, with a forested hilltop being warmer than

and as dry as the edge and a valley being moister than the interior. Therefore, while the

microclimate of forest edges may have indirect impacts on C. mitis and L. albigena, in

this case microclimate edge effects are not considered to play an important direct role in

the observed habitat use patterns and are not discussed further.









Forest Composition and Fruit Availability

Forest species composition and fruit availability were judged to be the most

important factors producing the observed habitat usage patterns and are discussed in

detail in Chapter 2.

Nutritional Quality of Vegetable Foods

Nutrient availability and quality of food has been shown to be important in primate

distribution and population densities (Chapman et al. 2002a, Wasserman and Chapman

2003). Food items could exist in the unused area in the same amounts as in the used

areas but if the nutritional quality is low this might explain the dearth of frugivorous

monkeys in the unused area. The nutritional components examined in this study can be

divided into two categories: those presumed desirable to the monkeys (lipids, protein,

and sugar) and those presumed undesirable (alkaloids, fiber, and saponins).

Fruit and leaves were collected each month for 6 months (June, 2002-November,

2002). Non-fig fruits were defined as ripe if it had at least partially changed to the ripe

color. The ripeness of figs was determined by smell and touch. Non-fig fruits were

peeled off the seed and dried in a dehydrator before being stored in plastic bags for

shipment back to the University of Florida. Figs were quartered before drying with no

attempt to remove the seeds. Preliminary analysis showed that there were few strong

patterns (detailed in Chapter 3) and none that supported the nutritional components

hypothesis (i.e. the unused area had less of a desirable component or more of an

undesirable component than the other areas).

Arthropod Foraging Substrate

Once at the end of the March-May rainy season and once at the end of the June-

August dry season, all transects were surveyed for the extent of moss/epiphyte foraging









substrate available to the monkeys. Every tree belonging to an important insect foraging

substrate species (see Chapter 2 for the definition of important food species) within 10 m

of the transects and larger than 5 cm dbh was measured (dbh) and placed into one of four

moss/epiphyte coverage categories (0-24%, 25-49%, 50-74%, or 75-100%). Only the

parts of the tree above 5 m were taken into consideration. If the epiphytes were large and

thick (cabbage-like) and provided more substrate than their coverage of the trunk would

imply, their contribution to the overall coverage was estimated by considering what it

would be if the leaves were spread out, edge to edge, on the surface of the trunk. Lichen

was considered only if foliose or fruticose and formed moss-like mats on the tree

(crustose growth forms, while extremely common, were deemed to be as poor foraging

substrate as the bark they so resemble). The basal area for each species and coverage

category for each transect was added to produce an index of foraging substrate

availability.

Because the unused area did not have noticeably lower levels of epiphyte cover

than the other transects, as would be expected if arthropod foraging substrate were

driving habitat use, it was decided that further analysis should focus on basal area and not

epiphyte cover of arthropod trees.

At the same time the moss/epiphyte surveys were being done, all transects were

surveyed for dead wood foraging substrate available to the monkeys. All observed dead

wood above 5 m in height on important invertebrate foraging species was placed into a

size category: twig (~2 cm diameter), leafy twig, branch (-10 cm), limb (-20 cm), or

trunk (-30 cm or greater) and the total end to end length of wood in each category was

estimated for each tree. Unfortunately, it was impossible to include dead bark in the dead






75


wood index. This weakness, plus my general impression that the index was extremely

sensitive to visibility through the canopy and position of the observer, led to the

conclusion that this index is a poor and unreliable one and should not be used.
















LITERATURE CITED


Aldrich-Blake, F. P. G. 1970. The ecology and behavior of the Blue monkey
Cercopithecus mitis stuhlmanni. PhD. dissertation University of Bristol, Bristol.

Anderson, S. C., J. A. Kupfer, R. R. Wilson, and R. J. Cooper. 2000. Estimating forest
crown area removed by selection cutting: a linked regression-GIS approach based
on stump diameters. Forest Ecology and Management 137:171-177.

Barrett, L., and C. B. Lowen. 1998. Random walks and the gas model: spacing behaviour
of Grey-cheeked Mangabeys. Functional Ecology 12:857-865.

Beeson, M. 1989. Seasonal dietary stress in a forest monkey (Cercopithecus mitis).
Oecologia 78:565-570.

Bemhard-Reversat, F. 1999. The leaching of Eucalyptus hybrids and Acacia
auriculiformis leaf litter: laboratory experiments on early decomposition and
ecological implications in congolese tree plantations. Applied Soil Ecology 12:251-
261.

Brashares, J. S. 2003. Ecological, behavioral, and life-history correlates of mammal
extinctions in West Africa. Conservation Biology 17:733-743.

Brooks, T. M., S. L. Pimm, and J. 0. Oyugi. 1999. Time lag between deforestation and
bird extinction in tropical forest fragments. Conservation Biology 13:1140-1150.

Brown, C. R., and M. B. Brown. 2000. Weather-mediated natural selection on arrival
time in cliff swallows (Petrochelidon pyrrhonota). Behavioral Ecology and
Sociobiology 47:339-345.

Brugiere, D., J. P. Gautier, A. Moungazi, and A. Gautier-Hion. 2002. Primate diet and
biomass in relation to vegetation composition and fruiting phenology in a rain
forest in Gabon. International Journal of Primatology 23:999-1024.

Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance
Sampling: Estimating abundance of biological populations. Chapman and Hall,
London.

Burke, D. M., and E. Nol. 1998. Edge and fragment size effects on the vegetation of
deciduous forests on Ontario, Canada. Natural Areas Journal 18:45-53.









Butynski, T. M. 1990. Comparative Ecology of Blue Monkeys (Cercopithecus-Mitis) in
High-Density and Low-Density Subpopulations. Ecological Monographs 60:1-26.

Chapman, C. A., S. R. Balcomb, T. R. Gillespie, J. P. Skorupa, and T. T. Struhsaker.
2000. Long-term effects of logging on African primate communities: a 28- year
comparison from Kibale National Park, Uganda. Conservation Biology 14:207-217.

Chapman, C. A., L. J. Chapman, K. A. Bjomdal, and D. A. Onderdonk. 2002a.
Application of protein to fiber ratios to predict colobine abundance on different
spatial scales. International Journal of Primatology 23:283-310.

Chapman, C. A., L. J. Chapman, M. Cords, J. M. Gathua, A. Gautier-Hion, J. E. Lambert,
K. Rode, C. E. G. Tutin, and L. J. T. White. 2002b. Variation in the diets of
Cercopithecus species: Differences within forests, among forests, and across
species. in M. E. Glenn and M. Cords, editors. The Guenons: Diversity and
Adaptation in African Monkeys. Kluwer Academic/Plenum Publishers, New York.

Chapman, C. A., L. J. Chapman, K. Rode, E. M. Hauck, and L. R. McDowell. 2003a.
Variation in the nutritional value of primate foods: Among trees, time periods, and
areas. International Journal of Primatology 24:317-333.

Chapman, C. A., L. J. Chapman, R. W. Wrangham, K. D. Hunt, D. Gebo, and L.
Gardner. 1992. Estimators of fruit abundance of tropical trees. Biotropica 24:527-
531.

Chapman, C. A., A. Gautier-Hion, J. F. Oates, and D. A. Onderdonk. 1999a. African
primate communities: Determinants of structure and threats to survival. Pages 1-37
in J. G. Fleagle, C. H. Janson, and K. E. Reed, editors. Primate Communities.
Cambridge University Press, Cambridge.

Chapman, C. A., and J. E. Lambert. 2000. Habitat alteration and the conservation of
African primates: Case study of Kibale National Park, Uganda. American Journal
of Primatology 50:169-185.

Chapman, C. A., M. J. Lawes, L. Naughton-Treves, and T. R. Gillespie. 2003b. Primate
Survival in Community-owned Forest Fragments: Are Metapopulation Models
Useful Amidst Intensive Use? Pages 404 in L. K. Marsh, editor. Primates in
Fragments. Kluwer Academic/Plenum Publishers, New York.

Chapman, C. A., and C. A. Peres. 2001. Primate conservation in the new millennium:
The role of scientists. Evolutionary Anthropology 10:16-33.

Chapman, C. A., R. W. Wrangham, L. J. Chapman, D. K. Kennard, and A. E. Zanne.
1999b. Fruit and flower phenology at two sites in Kibale National Park. Journal of
Tropical Ecology 15:189-211.









Conklin-Brittain, N. L., R. W. Wrangham, and K. D. Hunt. 1998. Dietary response of
chimpanzees and cercopithecines to seasonal variation in fruit abundance. II.
Macronutrients. International Journal of Primatology 19:971-998.

Cordeiro, N. J., and H. F. Howe. 2001. Low recruitment of trees dispersed by animals in
African forest fragments. Conservation Biology 15:1733-1741.

Cords, M. 1986. Interspecific and Intraspecific Variation in Diet of 2 Forest Guenons,
Cercopithecus-Ascanius and C-Mitis. Journal of Animal Ecology 55:811-827.

Cords, M., and T. E. Rowell. 1986. Group Fission in Blue Monkeys of the Kakamega
Forest, Kenya. Folia Primatologica 46:70-82.

De Walt, S. J., S. K. Maliakal, and J. S. Denslow. 2003. Changes in vegetation structure
and composition along a tropical forest chronosequence: implications for wildlife.
Forest Ecology and Management 182:139-151.

DeFries, R. S., R. A. Houghton, M. C. Hansen, C. B. Field, D. Skole, and J. Townshend.
2002. Carbon emissions from tropical deforestation and regrowth based on satellite
observations for the 1980's and 1990's. Proc. Nat. Acad. Sci. 99:14256-14261.

Devos, A., and A. Omar. 1971. Territories and Movements of Sykes Monkeys
(Cercopithecus-Mitis-Kolbi Neuman) in Kenya. Folia Primatologica 16:196-&.

Didham, R. K., and J. H. Lawton. 1999. Edge structure determines the magnitude of
changes in microclimate and vegetation structure in tropical forest fragments.
Biotropica 31:17-30.

Fairgrieve, C. 1995. The comparative ecology of Blue monkeys (Cercopithecus mitis
stuhlmannii) in logged and unlogged forest, Budongo Forest Reserve, Uganda: the
effects of logging on habitat and population density. PhD. University of Edinburgh,
Edinburgh.

Fairgrieve, C., and G. Muhumuza. 2003. Feeding ecology and dietary differences
between blue monkey (Cercopithecus mitis stuhlmanni Matschie) groups in logged
and unlogged forest, Budongo Forest Reserve, Uganda. African Journal of Ecology
41:141-149.

Fernandez-Escobar, R., R. Moreno, and M. Garcia-Creus. 1999. Seasonal changes of
mineral nutrients in olive leaves during the alternate-bearing cycle. Scientia
Horticulturae 82:25-45.

Fiksen, 0. 2000. The adaptive timing of diapause a search for evolutionarily robust
strategies in Calanusfinmarchicus. Ices Journal of Marine Science 57:1825-1833.

Gehlhausen, S. M., M. W. Schwartz, and C. K. Augspurger. 2000. Vegetation and
microclimatic edge effects in two mixed-mesophytic forest fragments. Plant
Ecology 147:21-35.









Ghiglieri, M. P. 1984. The chimpanzees of Kibale Forest. Columbia University Press,
New York.

Gonzalez, A., and E. J. Chaneton. 2002. Heterotroph species extinction, abundance and
biomass dynamics in an experimentally fragmented microecosystem. Journal of
Animal Ecology 71:594-602.

Granjon, L., J. F. Cosson, J. Judas, and S. Ringeut. 1996. Influence of tropical rainforest
fragmentation on mammal communities in French Guiana: Short-term effects. Acta
Oecologica--International Journal of Ecology 17:673-684.

Gupta, A. K., and D. J. Chivers. 1999. Biomass and use of resources in south and south-
east Asian primate communities. in J. G. Fleagle, C. H. Janson, and K. E. Reed,
editors. Primate Communities. Cambridge University Press, Cambridge.

Gurd, D. B., T. D. Nudds, and D. H. Rivard. 2001. Conservation of mammals in eastern
North American wildlife reserves: How small is too small? Conservation Biology
15:1355-1363.

Haddow, A. J. 1952. Field and laboratory studies on an African monkey Cercopithecus
ascanius schmidti (Matschie). Proceedings of the Zoological Society of London
122:297-394.

Haddow, A. J., K. C. Smithburn, A. F. Mahaffy, and S. C. Bugher. 1947. Monkeys in
relation to epidemiology of yellow fever in Bwamba County, Uganda. Transcripts
of the Royal Society of Tropical Medicine and Hygiene 40.

Hafner, H., 0. Pineau, and Y. Kayser. 1994. Ecological determinants of annual
fluctuations in numbers of breeding Little egrets (Egretta garzetta L) in the
Camargue, S. France. Revue D Ecologie La Terre et la Vie 49:53-62.

Hamilton, A. 1991. A Field Guide to Ugandan Forest Trees. Makerere University
Printery. Kampala.

Harmand, J. M., C. F. Njiti, F. Bernhard-Reversat, and H. Puig. 2004. Aboveground and
belowground biomass, productivity and nutrient accumulation in tree imporved
fallows in the dry tropics of Cameroon. Forest Ecology and Management 188:249-
265.

Hill, C. M. 1997. Crop-raiding by wild vertebrates: the farmer's perspective in an
agricultural community in western Uganda. International Journal of Pest
Management 43:77-84.

Howard, P. C., T. R. B. Davenport, F. W. Kigenyi, P. Viskanic, M. C. Baltzer, C. J.
Dickinson, J. S. Lwanga, R. A. Matthews, and E. Mupada. 2000. Protected area
planning in the tropics: Uganda's national system of forest nature reserves.
Conservation Biology 14:858-875.









Ingham, K. 1975. The Kingdom of Toro in Uganda. Methuen and Co. Ltd., London.

Janzen, D. H. 1967. Synchronization of sexual reproduction of trees within the dry season
in Central America. Evolution 21:620-637.

Jedrzejewska, B., H. Okarma, W. Jedrzejewski, and L. Milkowski. 1994. Effects of
exploitation and protection on forest structure, ungulate density and wolf predation
in Bialowieza Primeval Forest, Poland. Journal of Applied Ecology 31:664-676.

Karanth, K. U. 1992. Conservation prospects for lion-tailed macaques in Karnataka,
India. Zoo Biology 11:33-41.

Karr, J. R. 1976. Seasonality, resource availability, and community diversity in tropical
bird communities. The American Naturalist 110:973-994.

Katende, A. B., A. Birnie, and B. Tengnas. 1995. Useful Trees and Shrubs for Uganda:
Identification, Propogation and Management for Agricultural and Pastoral
Communities. Regional Soil Conservation Unit, Nairobi.

Kayanja, F. I. B., and D. Byarugaba. 2001. Disappearing forests of Uganda: The way
forward. Current Science 81:936-947.

Kingdon, J. 1974. East African Mammals: An atlas of evolution in Africa. The University
of Chicago Press, Chicago.

Kingdon, J. 1997. The Kingdon Field Guide to African Mammals. Academic Press,
London.

Klages, K., H. Donnison, J. Wunsche, and H. Boldingh. 2001. Diurnal changes in non-
structural carbohydrates in leaves, phloem exudate and fruit in 'Braeburn' apple.
Australian Journal of Plant Physiology 28:131-139.

Kouki, J., S. Lofman, P. Martikainen, S. Rouvinen, and A. Uotila. 2001. Forest
fragmentation in Fennoscandia: LInking habitat requirements of wood-associated
threatened species to landscape and habitat changes. Scandinavian Journal of Forest
Research Supplement 3:27-37.

Krupnick, G. A., and W. J. Kress. 2003. Hotspots and ecoregions: a test of conservation
priorities using taxonomic data. Biodiversity and Conservation 12:2237-2253.

Lame, M., S. Ringeut, D. Sabatier, and P. M. Forget. 2002. Fruit richness and seasonality
in a fragmented landscape of French Guiana. Revue D Ecologie La Terre et la Vie
Supplement 8:39-57.

Laurance, W. F., and G. B. Williamson. 2001. Positive feedbacks among forest
fragmentation, drought, and climate change in the Amazon. Conservation Biology
15:1529-1535.









Lawes, M. J. 2002. Conservation of Fragmented Populations of Cercopithecus mitis in
South Africa: the Role of Reintroduction, Corridors and Metapopulation Ecology.
Pages 375-392 in M. E. Glenn and M. Cords, editors. The Guenons: Diversity and
Adaptation in African Monkeys. Kluwer Academic/Plenum Publishers, New York.

Lawes, M. J., S. P. Henzi, and M. R. Perrin. 1990. Diet and Feeding-Behavior of
Samango Monkeys (Cercopithecus mitis labiatus) in Ngoye Forest, South-Africa.
Folia Primatologica 54:57-69.

Lemould, J.-M. 1988. Classification and geographical distribution of guenons: a review.
in A. Gautier-Hion, F. Bourliere, J. P. Gautier, and J. Kingdon, editors. A Primate
Radiation: Evolutionary Biology of the African Guenons. Cambridge University
Press, Cambridge.

Little, K. M., and R. A. W. Gardner. 2003. Coppicing ability of 20 Eucalyptus species
grown at two high-altitide sites in South Africa. Canadian Journal of Forest
Research 33:181-189.

Lubowa, L., B. M. Byanyima, and P. N. Kavuma. 1963. Report of the Commision of
Inquiry into the Administration by the Government of the Kingdom of Toro of the
services for which it is responsible in certain counties of the Kingdom. Uganda
Government, Entebbe.

Luck, G. W., T. H. Ricketts, G. C. Daily, and M. Imhoff. 2004. Alleviating spatial
conflict between people and biodiversity. PNAS 101:182-186.

Lwanga, J. S. 1987. Group fission in Blue Monkeys (Cercopithecus mitis stuhlmanni):
Effects on the socioecology in Kibale Forest, Uganda. Master of Science. Makerere
University, Kampala.

Maganga, S. L. S., and R. G. Wright. 1991. Bark-stripping by Blue Monkeys in a
Tanzanian forest plantation. Tropical Pest Management 37:169-174.

Marquis, R. J., E. A. Newell, and A. C. Villegas. 1997. Non-structural carbohydrate
accumulation and use in an understory rain-forest shrub and relevance for the
impact of leaf herbivory. Functional Ecology 11:636-643.

Marsh, L. K. 2003. The Nature of Fragmentation. Pages 404 in L. K. Marsh, editor.
Primates in Fragments: Ecology and Conservation. Kluwer Academic/Plenum
Publishers, New York.

Marsh, L. K., and B. A. Loiselle. 2003. Recruitment of black howler fruit trees in
fragmented forests of Northern Belize. International Journal of Primatology 24:65-
86.

Medley, K. E. 1993. Primate conservation along the Tana River, Kenya--An examination
of the forest habitat. Conservation Biology 7:109-121.









Mitani, J. C., W. J. Sanders, J. S. Lwanga, and T. L. Windfelder. 2001. Predatory
behavior of crowned hawk-eagles (Stephanoaetus coronatus) in Kibale National
Park, Uganda. Behav Ecol Sociobiol 49:187-195.

Mitani, J. C., and D. P. Watts. 2001. Why do chimpanzees hunt and share meat. Animal
Behavior 61:915-924.

Mittermeier, R. A., and D. L. Cheney. 1987. Conservation of primates and their habitats.
Pages 477-490 in B. B. Smuts, D. L. Cheney, R. Seyfarth, R. W. Wrangham, and T.
T. Struhsaker, editors. Primate Societies. Chicago University Press, Chicago.

Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation.
Trends in Ecology and Evolution 10:58-62.

Muri, H. 1999. Weather situation, aspects of reproduction and population density in roe
deer (Capreolus capreolus L.). Zeitschrift fur Jagdwissenschaft 45:88-95.

National Research Council. 1981. Techniques for the study of primate population
ecology. National Academy Press, Washington, D.C.

Nergiz, C., and Y. Engez. 2000. Compositional variation of olive fruit during ripening.
Food Chemistry 69:55-59.

Newmark, W. D. 1993. The role and design of wildlife corridors with examples from
Tanzania. Ambio 22:500-504.

Norconk, M. A., and B. W. Grafton. 2003. Changes in Forest Composition and Potential
Feeding Tree Availability on a Small Land-bridge Island in Lago Guri, Venezuela.
Pages 404 in L. K. Marsh, editor. Primates in Fragments: Ecology and
Conservation. Kluwer Academic/Plenum Publishers, New York.

Olupot, W. 1994. Ranging patterns of the grey-cheeked mangabey Cercocebus albigena
with special reference to food finding and food availability in Kibale National Park.
Master of Science. Makerere University, Kampala.

Olupot, W. 1998. Long-term variation in mangabey (Cercocebus albigenajohnstoni
Lydekker) feeding in Kibale National Park. African Journal of Ecology 36:96-101.

Olupot, W. 2000. Mass differences among male mangabey monkeys inhabitting logged
and unlogged forest compartments. Conservation Biology 14:833-843.

Onderdonk, D. A., and C. A. Chapman. 2000. Coping with forest fragmentation: The
primates of Kibale National Park, Uganda. International Journal of Primatology
21:587-611.

Ortega-Huerta, M. A., and A. T. Peterson. 2004. Modelling spatial patterns of
biodiversity for conservation prioritization in North-eastern Mexico. Diversity and
Distributions 10:39-54.









Peres, C. A. 1990. Effects of hunting on western Amazonian primate communities.
Biological Conservation 54:47-59.

Plumptre, A. J., and V. Reynolds. 1994. The effect of selective logging on the primate
populations in the Budongo Forest Reserve, Uganda. Journal of Applied Ecology
31:631-641.

Poiani, K. A., B. D. Richter, M. G. Anderson, and H. E. Richter. 2000. Biodiversity
conservation at multiple scales: Functional sites, landscapes, and networks.
Bioscience 50:133-146.

Pope, S. E., L. Fahrig, and N. G. Merriam. 2000. Landscape complementation and
metapopulation effects on leopard frog populations. Ecology 81:2498-2508.

Poulsen, J. R., C. J. Clark, and T. B. Smith. 2001. Seasonal variation in the feeding
ecology of the Grey-cheeked mangabey (Lophocebus albigena) in Cameroon.
American Journal of Primatology 54:91-105.

Prendergast, J. R., R. M. Quinn, J. H. Lawton, B. C. Eversham, and D. W. Gibbons.
1993. Rare species, the coincidence of diversity hotspots and conservation
strategies. Nature 365:335-337.

Radersma, S., and C. K. Ong. 2004. Spatial distribution of root length density and soil
water of linear agroforestry systems in sub-humid Kenya: implications for
agroforestry models. Forest Ecology and Management 188:77-89.

Rao, M., and C. P. van Schaik. 1997. The behavioral ecology of Sumatran orangutans in
logged and unlogged forest. Tropical Biodiversity 4:173-185.

Richardson, J. S. 1991. Seasonal food limitation of detritivores in a montane stream an
experimental test. Ecology 72:873-887.

Rode, K., C. A. Chapman, L. J. Chapman, and L. R. McDowell. 2003. Mineral resource
availability and consumption by colobus in Kibale National Park, Uganda.
International Journal of Primatology 24:541-573.

Rodriguez-Toledo, E., M, S. Mandujano, and F. Garcia-Orduna. 2003. Relationships
Between Forest Fragments and Howler Monkeys (Alouatta palliata mexicana) in
Southern Veracruz, Mexico. Pages 404 in L. K. Marsh, editor. Primates in
Fragments. Kluwer Academic/Plenum Publishers, New York.

Rudran, R. 1978. Socioecology of the Blue Monkeys (Cercopithecus mitis stuhlmanni) of
the Kibale Forest, Uganda. Smithsonian Contributions to Zoology 249.

Sams, C. E. 1999. Preharvest factors affecting postharvest texture. Postharvest Biology
and Technology 15:249-254.









Schaefer, H. M., and V. Schmidt. 2002. Vertical stratification and caloric content of the
standing fruit crop in a tropical lowland forest. Biotropica 34:244-253.

Schaefer, H. M., V. Schmidt, and F. Bairlein. 2003. Discrimination abilities for nutrients:
which difference matters for choosy birds and why? Animal Behavior 65:531-541.

Segelbacher, G., J. Hoglund, and I. Storch. 2003. From connectivity to isolation: genetic
consequences of population fragmentation in Capercaillie across Europe. Molecular
Ecology 12:1773-1780.

Siex, K. S., and T. T. Struhsaker. 1999. Ecology of the Zanzibar red colobus monkey:
Demographic variability and habitat stability. International Journal of Primatology
20:163-192.

Skorupa, J. P. 1988. The effect of selective timber harvesting on rainforest primates in
Kibale Forest, Uganda. University of California, Davis.

Skorupa, J. P. 1989. Crowned eagles Stephanoaetus coronatus in rainforest -
Observations on breeding chronology and diet at a nest in Uganda. Ibis 131:294-
298.

Ssembeguya, F. C., G. 0. B. Oda, and J. M. Okae. 1962. Report of the Commission of
Inquiry into the Recent Disturbances amongst the Baamba and Bakonjo People of
Toro. Uganda Government, Entebbe.

Stickler, C. M. 2004. The effects of selective logging on primate-habitat interactions: A
case study of redtail monkeys (Cercopithecus ascanius) in Kibale National Park,
Uganda. Master of Science. University of Florida, Gainesville.

Struhsaker, T. T. 1975. The red colobus monkey. University of Chicago Press, Chicago.

Struhsaker, T. T. 1978. Food habits of five monkey species in the Kibale Forest, Uganda.
Pages 225-248 in D. J. Chivers and J. Herbert, editors. Recent Advances in
Primatology. Academic Press, London.

Struhsaker, T. T. 1997. Ecology of an African Rainforest. University of Florida Press,
Gainesville.

Struhsaker, T. T., and M. Leakey. 1990. Prey selectivity by Crowned hawk-eagles on
monkey in the Kibale Forest, Uganda. Behav Ecol Sociobiol 26:435-443.

Struhsaker, T. T., and L. Leland. 1979. Socioecology of Five Sympatric Monkey Species
in the Kibale Forest, Uganda. Pages 158-228 in J. Rosenblatt, R. A. Hinde, C. Beer,
and M. C. Busnel, editors. Advances in the Study of Behavior. Academic Press,
New York.









Sun, C., B. A. Kaplin, K. A. Kristensen, V. Munyaligoga, J. Mvukiyumwami, K. K.
Kajondo, and T. C. Moermond. 1996. Tree phenology in a tropical montane forest
in Rwanda. Biotropica 28:668-681.

Syahuka-Muhindo, A. 1991. The Rwenzururu Movement and the Democratic Struggle.
Working Paper No. 15 Centre for Basic Research, Kampala, Uganda.

Tutin, C. E. G., L. J. T. White, and A. Machanga-Missandzou. 1997. The use by rain
forest mammals of natural forest fragments in an equatorial African savanna.
Conservation Biology 11:1190-1203.

Twinomugisha, D., C. A. Chapman, M. J. Lawes, C. 0. Worman, and L. M. Danish. in
press. How does the Golden Monkey of the Virungas cope in a fruit scarce
environment? in Primates of Uganda.

Umapathy, G., and A. Kumar. 2000. The occurrence of arboreal mammals in the rain
forest fragments in the Anamalai Hills, south India. Biological Conservation
92:311-319.

van Schaik, C. P., J. W. Terborgh, and S. J. Wright. 1993. The phenology of tropical
forests: Adaptive significance and consequences for primary consumers. Annual
Review of Ecology and Systematics 24:353-377.

Vargas, A., I. Jimenez, F. Palomares, and M. J. Palacios. 2002. Distribution, status, and
conservation needs of the golden-crowned sifaka (Propithecus tattersalli).
Biological Conservation 108:325-334.

Waser, P. 1975. Monthly variations in feeding and activity patterns of the mangabey,
Cercocebus albigena (Lydekker). East African Wildlife Journal 13:249-263.

Waser, P. 1977. Feeding, ranging and group size in the mangabey Cercopithecus
albigena. in T. H. Clutton-Brock, editor. Primate Ecology: Studies of feeding and
ranging behaviour in lemurs, monkeys and apes. Academic Press, London.

Waser, P. 1984. Ecological differences and behavioral contrasts between two mangabey
species. Pages 195-216 in P. S. Rodman and J. G. H. Cant, editors. Adaptations for
Foraging in Non-Human Primates. Columbia University Press, New York.

Waser, P., and 0. Floody. 1974. Ranging patterns of the mangabey Cercopithecus
albigena, in the Kibale Forest, Uganda. Zeitschrift fur Tierpsychologie 35: 85-101.

Wasserman, M. D., and C. A. Chapman. 2003. Determinants of colobine monkey
abundance: The importance of food energy, protein and fibre content. Journal of
Animal Ecology 72:650-659.

Williams, M. 2000. Dark ages and dark areas: Global deforestation in the deep past.
Journal of Historical Geography 26:28-46.









Willock, C. 1964. The Enormous Zoo: A profile of the Uganda national parks. Harcourt,
Brace and World, Inc., New York.

Wilson, C. C., and W. L. Wilson. 1975. The influence of selective logging on primates
and some other animals in East Kalimantan. Folia Primatologica 23:245-274.

Woodwell, G. M. 1974. Variation in the nutrient content of leaves of Quercus alba,
Quercus coccinea, and Pinus rigida in the Brookhaven forest from bud-break to
abscission. American Journal of Botany 61.

Woodwell, G. M. 2002. On purpose in science, conservation and government The
functional integrity of the Earth is at issue not biodiversity. Ambio 31:432-436.

Woolf, A. B., and I. B. Ferguson. 2000. Postharvest responses to high fruit temperatures
in the field. Postharvest Biology and Technology 21:7-20.

Wrangham, R. W., N. L. Conklin-Brittain, and K. D. Hunt. 1998. Dietary response of
chimpanzees and cercopithecines to seasonal variation in fruit abundance. I.
Antifeedants. International Journal of Primatology 19:949-970.















BIOGRAPHICAL SKETCH

Cedric O'Driscoll Worman received an excellent elementary school education from

the Minnesota public school system (wandering the forests, mountains, and deserts of

North America with his family did not hurt either) and attended North Hollywood High

School Magnet for Biological and Mathematical Sciences. He earned his Bachelor of

Science at Iowa State University with specializations in restoration ecology and

mammalian behavior. After graduation he joined the Peace Corps and was sent to

Uganda where he learned more than he really wanted to. Before entering graduate

school, he taught in and explored Korea.