RETURNING MATERIALS: 1V1531_) Place in book drop to LJBRARJES remove this checkout from .—_ your record. FINES will 7 be charged if book is returned after the date stamped below. 9 1 ’.' r e" e , r 1.1. § :- .1 3-, 1 -‘ E”? ‘ fi‘ifl"? wv ' mm 3 mm COMPARATIVE DIETARY ECOLOGY OF A COMMUNITY OF FRUGIVOROUS FOREST UNGULATES IN ZAIRE BY John Amasa Hart A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Fisheries and Wildlife 1985 Copyright by JOHN AMASA HART 1985 ABSTRACT COMPARATIVE DIETARY ECOLOGY OF A COMMUNITY OF FRUGIVOROUS FOREST UNGULATES IN ZAIRE BY John Amasa Hart Factors affecting diet and food choice in a guild of seven species of frugivorous ungulates (six duikers, genus cephalaphus. and the chevrotain. Hyemaschus aquaticus) were investigated in the Ituri Forest of Zaire. The methods used included radio-telemetry. feeding experiments with captive individuals, analyses of rumen contents of free-ranging animals in relationship to food availability and distributions of animals flushed on hunts. Radio telemetry and feeding trials contrasted a large and a small duiker species. The larger species was more mobile and had a greater digestive capacity for low quality foods. The smaller species was more selective in its choice of food but also had relatively smaller total food needs. Studies of mouth morphology. rumen contents, food availability and animal distributions involved all species in the guild. Fruits and seeds on the forest floor occurred in discrete patches. Relative to larger species, a small species of duiker selected foods of high nutritional quality but foraged on food patches of both small and large total food_weight. The larger species selected foods of variable nutritional quality but preferred foods which occurred in patches of high total food biomass. Mouth size limited the size of food items which could be ingested. Species with narrow mouths included smaller food items in their diets. Species with broad mouths avoided these. Diets of all species converged when high quality fruits and seeds were abundant in the upland forests. Diets diverged when high quality foods were scarce, but overall food diversity was high. During periods of food scarcity. two of the ungulate species foraged mainly in riverine habitats. Five species remained in upland habitats but specialized on fruits of varying nutritional and distributional characteristics. Abundances of two of these upland duikers of similar body size were negatively correlated indicating possible competition. Ungulate abundance was contrasted in two forest types, a mixed forest and a forest dominated by the single species Gilbertiadendron dawevrei. Ungulate abundance was lower in the single—species dominant forest. Compared to mixed forest. food abundance in the single-species dominant forest fluctuated strongly, but food diversity was always low. Diet overlap occurred in the mixed forest when both food abundance and diversity were low. indicating possible competition. AKNOWLEDGMENTS This study was made possible by a grant from the United States Man and the Biosphere Program (1981-1983, Federal Grant Number 4789-4) in co-operation with the Institut Zairois pour la Conservation de la Nature (IZCN). Financial support during the period of data analysis was provided by the African Studies Center and the Department of Fisheries and wildlife. both of Michigan State University. Terese B. Hart contributed in a major way to every phase of this study. Her botanical and ecological studies of the Ituri Forest served as the basis for the analysis of the food habits of the ungulate community reported here. Her field assistance proved indispensible. Her freely shared insights were a continuous source of inspiration and direction. Dr. Peter G. Waterman of the University of Strathclyde generously analyzed the bulk of the food samples for this study and has been a source of objective criticsm and insight throughout the writing. Drs. George Petrides, Thomas Struhsaker and Richard Wrangham visited the project in Zaire. There observations helped focus my efforts on specific problems. Dr. Duane Ullrey suggested I attempt feeding trials with captive duiker and Dr. M. Demment generously shared insights and unpublished papers on relationships between ruminant nutrition and ecology and criticized earlier drafts of chapters of this dissertation. Drs. J. Haufler and N. Phu provided me with laboratory space and assisted in the chemical analysis of samples from the duiker digestion trials. Dr. Peter Murphy generously made his personal computer available to me to complete the typing of the manuscript. The criticisms of Dr George Petrides and Dr. Donald Hall added substatially to the analysis and presentation of the results. This dissertaion has benefited from the comments and helpful criticisms of my doctoral committee consisting of Drs. George Petrides (chairperson). John King. Donald Hall and Jon Naufler. Throughout the research period in Zaire and subsequent analysis and writing at Michigan State University I have had the support of my entire family. My parents. Joanne and Nathaniel Hart and my parents-in—law. Kathleen and Aruther Butler provided encouragement and finacial support. I am especially grateful to my wife Terese and our daughters, Sarah and Rebekah who made it possible for me to perservere with this project. Finally, I owe my deepest gratitude and a special thanks to the Mbuti hunters of the Epulu area in the Ituri Forest. Without their enthusiastic co-operation, lively interest and continuous assistance. this project would not have been possible. This dissertation is dedicated to them in their forest home. iii TABLE OF CONTENTS List of Tables ................................................. List of Figures ................................................ CHAPTER ONE: Introduction ...................................... Literature Cited ........................................... CHAPTER TWO: Mobility. Food Handling and Digestive Capacity in Two Species of Duikers of Differing Body Size .............. Materials and Methods ...................................... Home Range and Mobility ............................... Ruminoreticular Capacity .............................. Food Handling ......... ................................ Digestion Trials ...................................... Results .................................................... Home Range and Mobility ............................... Gut Capacity .......................................... Food Handling ......................................... Digestion Trials ...................................... Discussion ................................................. Conclusions ................................................ Literature Cited ........................................... CHAPTER THREE: A Study of Factors Affecting Food Preference and Intake in Two Species of Duikers ....................... Methods .................................................... Palatability Trials ................................... Nutritional Composition of Food ....................... Results .................................................... Preference Ranks ...................................... Composition of Foods .................................. Chemical Correlates of Preference: Fruit .............. Chemical Correlates of Preference: Foliage ............ Intake on Fruit Diets ................................. iv 0) 000$” 13 13 15 18 23 27 29 32 35 35 4O 41 41 47 50 52 54 Discussion ................................................. 62 Foliage as a Food Source .............................. 67 Intake, Diet Quality and Body Size in Duikers ......... 69 Conclusions ................................................ 71 Literature Cited ........................................... 72 Appendix 3-A ............................................... 75 CHAPTER FOUR: Comparative Dietary Ecology in a Community of Frugivorous Ungulates ................................... 78 Study Site ................................................. 80 Climate ............................................... 80 Vegetation and Soils .................................. 83 Forest Ungulates and Their Status ..................... 85 Methods .................................................... 87 Body Weight and Cranial Morphology .................... 88 Food Availability ..................................... 89 Diet .................................................. 91 Nutritional Quality... ................................ 93 Diet Selectivity ...................................... 95 Animal Abundance and Distribution ..................... 96 Statistical Tests ..................................... 97 Results .................................................... 97 Cranial Morphology .................................... 97 Food Availability ..................................... 98 General Diet Composition .............................. 111 Food Selection ........................................ 119 Taxonomic Identity ............................... 121 Food Patch Weight ................................ 121 Food Item Size ................................... 123 Food Nutritional Quality ......................... 129 Diet Overlap .......................................... 134 Patterns in Abundance of Frugivorous Ungulates ........ 142 Ungulate Distributions ........................... 142 Ungulate Abundance and Patterns in Food Abundance ........................................ 144 Discussion ................................................. 149 Food Selection in Upland Duikers ................. 148 Diet Overlap and Community Structure ............. 152 Are Forest Ungulates Food-Limited? ............... 156 Literature Cited ........................................... 158 Appendix 4-A ............................................... 164 Appendix 4-B ............................................... 165 LIST OF TABLES Table CHAPTER ONE 1-1 Species of frugivorous ungulates studied in the Ituri Forest, Zaire. 1981—1983 ............................. 3 CHAPTER TWO 2—1 Measures of rumen capacity and ratios to body weight in adult blue duiker (C. monticala) and bay duiker (C.dbrsalis) ............................................... 14 2—2 Handling times for fruits by an adult female blue duiker (C. manticola) and a subadult female bay duiker (C. dbrsalis) .............................................. 16 2—3 Estimated number of fruits to fill rumen, fruit handling times and time to rumen-fill for fruits of different sizes for blue duiker (C. monticoia) and bay duiker (C. dorsalis) .............................................. 17 2-4 Nutritional composition of foods utilized in feeding trials with blue duiker (C. monticola) and bay duiker (C. dbrsaIis) ....................................... 19 2-5 Percentage dry weight composition of diets of blue duiker (C. monticola) and bay duiker (C. dorsaIis) on feeding trials containing K; gabonensis fruit and I. batata foliage (Trial 1) and R. heudelotii fruit and I. batata foliage (Trial II) ...................... 21 2—6 Digestion coefficients for blue duiker (C. monticoIa) and bay duiker (C. dorsalis) on diets of R. gabonensis fruit and I. batata foliage (Trial 1) and R. heudeIotii fruit and I. batata foliage (Trial 11) ...................... 22 2-7 Apparent lignin digestion coefficients for blue duiker (C. monticola) and bay duiker (C. dbrsalis) on diets of K. gabonensis fruit and I. batata foliage (Trial 1) and R. heudelotii fruit and I. batata foliage foliage (Trial II) ....................................... 24 vi 3-1 3-3 3-5 3-6 3-8 3-10 CHAPTER THREE Design of fruit palatability trials ........................ 36 Rank preferences of fruits and seeds as determined by palatability trials with captive blue duiker (C. monticola) and bay duiker (C. dorsalis) ............................... 42 Scores and preference ranks for foliage species offered in palatability trials with captive blue duiker (C. montjcola) and bay duiker (C. dorsalis) ................................. 45 Consistency of preference choices between replicated presentations of the same food combination within trials.... 46 Chemical composition of foods offered in palatabilu trials.. in palatability trials .................................... 48 Chemical composition of foliage offered in palatability trials ....................................... 49 Spearman's rank correlation coefficients (r ) of rank preference with chemical measures of fruits in palatability trials with blue duiker (C. monticola) and bay duiker (C. dorsalis) .............................................. 51 Spearman's rank correlation coefficients (r ) of rank preference with chemical measures of ten species of canopy foliage tested in palatability trials with blue duiker (C. monticola) and bay duiker (C. dorsalis) ..................... 53 Possible defensive characterisitics of species of understory foliage offered to a blue duiker (C. monticola) in palatability trials ........................................ 55 Percentage of total dietary dry matter contributed by I. batata foliage in diets of blue duiker (C. monticoia) and bay duiker (C. dbrsaIis) in which the fruits offered varied in preference rankings ...................... 61 CHAPTER FOUR The frugivorous ungulate fauna of the Ituri Forest. Zaire... 86 Occurrence of fallen fruits. seeds and flowers on ground transects in mixed forest ................................... 103 Food patch weight (W) class of fruits, seeds and flowers collectively. recorded on transects in the Ituri Forest Zaire. July, 1981 to May, 1983 ..................................... 106 vii. 4-4 4-5 4-6 4-7 4-8 4-9 4-10 4-11 4-12 4-14 Size classes of fruits. seeds and flowers collectivley. recorded on transects in the Ituri Forest. Zaire. July. 1981 to May, 1983 ..................................... Rumen contents of frugivorous ungulates analyzed seasonally between December, 1981 and May. 1983 ........................ Percentage composition of large (3 5 mm) particles screened from rumen contents of adult and weaned-juvenile duikers and chevrotain ...................................... Selection by four upland duiker species for foods which were abundant on transects (> 5% total weight) ................... Changes in preference by four species of upland duikers for six food species available during more than one sampling period ...................................................... Numbers of preferred and avoided food species of large patch weight (W > 100 g) and small patch weight (W_g 100 g) in diets of four species of upland duikers .................. Chi square values for tests of A) equal food size distributions of preferred food species between diets of four species of upland duikers, and B) equal size distributions of preferred and avoided food species within diets of each duiker species ................................ A) Mann Whitney U probabilities associated with tests that adjusted dry matter yield (Y ) values of preferred foods are greater than avoided foogs in diets of four species of upland duikers during five sample periods. B) Numbers of preferred food species with Y values greater than and less than average values for gvailable foods summed over five sample periods ......................................... Quality and abundance of dominant foods (I > 5%) shared in the diets of at least two of three specqes of duikers. C. leucagaster. C. caIIipygus and C. dorsalis during periods of high dietary overlap ............................. Numbers and characteristics of dominant food species (IU > 5%) unique to diets of each of three duikers. C. Jeucagaster, C. callipygus and C. dorsalis during periods of dietary divergence ............................... Values for Morisita's (I ) index for dietary overlap between two riverine species. H. aquaticus and C. nigrifrons, and between two riverine species and upland duikers ................................. viii 107 112 116 120 122 124 127 132 .137 .139 141 4‘15 A) Numbers f108h8d/ kmz. and B) percentages of duikers and chevrotains flushed on drive hunts at nine sites in the Ituri Forest. Zaire. 1981 to 1983 ...................... 143 4-16 Species diversity and total weights recorded on transects in mixed and mbau forests during seasonal fruiting cycles of Caesalpiniaceous trees ..................................... 148 LIST OF FIGURES Figure INTRODUCTION 1—1 The Ituri Forest. Zaire .................................. 2 CHAPTER THREE 3-1 Average (1 SE) wet weight eaten (g/day) versus 4-1 4-3 4-4 preference rank of each fruit species when the indicated species was the preferred choice offered .................. 57 Total fruit dry matter intake (Mean 1 SE) on fruit diets of decreasing rank preference ................. 60 Total dry matter intake (g) versus percent Ipomea batata foliage in blue duiker diets for all two—species fruit offerings in Trial A ................................ 64 CHAPTER FOUR The Epulu study area in the Ituri Forest of Zaire ......... 82 Cranium size and mouth shape (width/length) in duikers and chevrotain ............................................ 100 Diversity (Shannon index and number of species/km transect) and collective abundance (Kg/km transect) of fruits. seeds and flowers on the forest floor ............. 102 Relationship between item size class and patch weight (W) for 174 fruits, flowers and seeds collected in the Ituri Forest, Zaire. 1983 ................................. 110 Numbers of fruits, seeds and flowers in rumens of six species of duiker and the chevrotain in the Ituri Forest, Zaire ............................................. 114 Size distributions of preferred (fine stipling) and avoided (coarse stipling) foods in diets of four species of upland duikers ............................................ 126 Adjusted dry matter yield (Y ) of selected (closed circles) and avoided (open cIrcles) food species in diets of four species of upland duikers ......................... 131 4-9 Values 0f MOP181t8'8 (IM) index for dietary overlap... 136 Average ungulate abundance (1 SE) on drive areas of differing percentage mbau forest cover at two sites. K and E. Ituri Forest ........................................... xi 146 CHAPTER ONE Introduction The Ituri Forest of Zaire (Figure 1-1) contains one of the richest assemblages of forest ungulates on earth. Prominent among these is a guild (sensu Root 1967) of six species of duikers (genus Cephalophus. Bovidae) and the chevrotain (Hyemoschus aquaticus. Tragulidae) (Table 1-1). These species span a size range from 5 to 70 kilograms but share a similar diet consisting of a wide array of fallen fruits, soft seeds and flowers (Dubost 1984). Studies of this group of species in Gabon (Dubost 1979. 1984, Gautier-Hion et al. 1980, Emmons et al. 1983) have indicated that despite differences in fruit sizes eaten and in habitat and activity patterns, the duikers and chevrotain shared a number of food species. These observations indicated the possibility of competition between guild members. Mechanisms mediating coexistence. however. have been little investigated. The purpose of this study was to investigate ecological relations within the frugiovrous ungulate community by a study of factors affecting food choice and diet. This research is reported as three separate papers. In the first paper (Chapter 2). radio telemetry, morphological studies and feeding trials with captive animals were used to examine how differences in body size may be linked to physical and 2a Figure 1-1. The Ituri Forest of Zaire. - LL. p O O. on o . ...... a.“ . .. . x E. R ./ man I .J. . \ \Oém . ' I .... Sz<~24p // \ (J .\ :25... W I \ a sin..\ n. I. e r I. II I. [I .II I \ \0 \l". o..s..8~°u M J x I!“ ‘GI—Bod c v a ,. o x a. .. \\ o... .0 Pi) \- \\ '.‘.\a r \1.‘. (I ‘5.-.‘c‘. \wuoufi. Emcee-.2 .- 33.1%.. .50-IODI .- 3...: hmmmOm u:- .o.aI o {mm .mmm..ml ..|...|.| \‘ tom's. "a m’LCHU mum's... .mo...m.m. men-Om.— g m¢~‘ Cephalophus aonticola blue duiker 4.7 Hyeaoschus aquaticus chevrotain 11.2 C. niprifrans black-fronted 13.9 duiker C. leucopaster white-bellied 16.7 duiker C. callipypus Peter‘s duiker 17.7 C. dorsalis bay duiker‘ 22.0 C. sylvicultor yellow-back 68.0 duiker Average adult except for 0. body weight data from this study (Chapter 4), sylvicultor (Emmons et al. (1983). physiological constraints affecting foraging. These studies were limited to two species, the blue duiker (C. monticola). the smallest member of the guild. and the bay duiker (C. dorsaijs) one of the largest. The second paper (Chapter 3), reports the results of food preference tests with these same two duiker species. The trials were designed to investigate the relation of food nutritional quality to diet choice and food intake. The third paper (Chapter 4) expands the focus of study to the entire guild. The results of a study of the diets and distributions of free-ranging animals in relation to food availability are reported. The questions posed in this chapter include the following: What are the patterns of diet selection in free-ranging animals and how are these correlated with morphological and size differences in the animals? What are the patterns of dietary divergence and convergence within the guild and how are these related to the abundance and diversity of food reosurces. Finally, what are the patterns of ungulate distibution and how are they correlated with food abundance? LITERATURE CITED LITERATURE CITED Dubost, G. 1979. The size of African forest artiodactyls as determined by the vegetation structure. African Journal of Ecology 17: 1-17. Dubost. G. 1984. Comparison of the diets of frugivorous forest ruminants of Gabon. Journal of Mammalogy 65: 298—316. Emmons. L.H., A. Gautier-Hion. and G. Dubost. 1983. Community structure of the frugivorous-folivorous forest mammals of Gabon. Journal of Zoology (London) 199: 209-222. Gautier-Hion, A. L.H. Emmons. and G. Dubost. 1980. A comparison of the diets of three major groups of primary consumers (primates, squirrels and ruminants) of the evergreen forests of Gabon. Oecologia 45: 182—189. Root. R.B. 1967. The niche exploitation patterns of the blue-gray gnatcatcher. Ecological Monographs 37: 317-350. CHAPTER TWO Mobility. Food Handling and Digestive Capacity in Two Species of Duikers of Differing Body Size Differences in body size have been seen as important in the ecological segregation of animal species in a number of communities including ungulates (Bell 1970. Hoffmann 1973. Schoener 1974. Hanley 1980, Owen-Smith 1980, Foose 1982, Demment and Van Soest 1985. Bunnell and Gillingham in press). While differences in body size affect a number of ecologically significant parameters (Clutton-Brock and Harvey 1983). the relationship betweeen body size, foraging and diet have received special attention. The frugivorous ungulates of African forests present a potentially interesting community to study from this perspective. Up to seven species, including six duikers (Cephalophus) and the chevrotain (Hyemoschus aquaticus). may co-occur in the same forest. These species. in particular the duikers, vary to some extent in habitat and activity patterns (Emmons et al. 1983) but overall are similar in general morphology (Kingdon 1982). rumen anatomy (Hoffmann 1973) and diet (Dubost 1984). Co-occurring species do vary in body size and in the richest communities may span a range from 5 to almost 70 kg (Dubost 1979). Factors determining diet across this size range. however. remain little known. This paper presents the results of observations and experiments on two species of duikers. a small species, the blue duiker. C..monticoIa (5 kg adult weight) and a large species, C. dorsaIis (22 kg adult weight). The purpose of these studies was to develop a basis for evaluating how differences in body size might be related to differing constraints on the use of the available food supply (Chapter 4). Food availability for ungulates in tropical forests differs from that of many species in more open environments. Leafy browse in the shady forest understory may be scarceand unpalatable (Chapter 3). A high diversity of fallen fruits, seeds and flowers represent the bulk of available food but many species may be rare and dispersed. Fruits further vary in texture, shape and nutrient composition (Chapter 4). These conditions suggest that differences in the relative ability to find. swallow and digest foods of differing distribution. size and nutritional quality may be important in segregating diets of related species. Reported here are observations on relative mobility, gut capacity, food handling and diet digestibility. Diet digestibility was investigated in relationship to plant cell wall (fiber) levels in fruits. Dietary fiber is only slowly and incompletely digested. Fiber thus has been identified as a determinant of ungulate diet quality. Differences in ability to digest fiber are correlated with ruminant body size (Short 1963, Short et al. 1974, Demment and Van Soest 1985). The relationship of diet quality, digestibility and body size was examined according to three general hypotheses: 1. A large duiker species will digest a low quality diet more completely than a small species. 2. Both large and small species will digest a high quality diet more completely than they will a low quality diet. 3. The digestion of a high quality diet is equally complete in both species. MATERIALS AND METHODS All studies were conducted at Epulu in the Ituri Forest between 1982 and 1983. Home Range and Mobility Four free-ranging adult and sub adult blue duiker (2 males and 2 females) and a subadult female bay duiker were caught, equipped with radio collars and followed on the same 70 hectare study area between February 26 and April 30, 1983. The study area was divided into 100 ha grids by a path system. Animal locations were monitored once an hour during recording sessions lasting 4 to 5 hours each. A total of 92 hours of observation were made at all hours of day and night. The total area area utilized by an animal during the study period was estimated by the area of a polygon connecting outermost location points on a map. Two indices of relative mobility were calculated as meters/hour (the average distance the animal moved between hourly locations) and meters/move (the average distance the animal moved when it relocated between grids). Rgginoreticular Capacity Ruminoreticular capacity was measured as weight of contents and by liquid volume fill. Adult blue and bay duiker killed by local hunters were weighed at the site of capture. The rumen and reticulum were weighed full and empty. The rumen was then filled with water to the ruminoreticular junction through an opening in the reticulum. Rumen volume (without reticulum) was determined by filling the rumen with water (to nearest 0.1 liter) while holding it submerged to avoid distention and overestimation of capacity (Demment 1982). Food Handling Food handling was tested using a captive adult female blue duiker and a nearly full-sized subadult female bay duiker. Handling time was defined as the number of seconds elapsing between the time an item was picked up until it was swallowed. During trials surplus quantities of fruits were offered in piles on the ground. Handling times were recorded for each item until the animal stopped and moved away (1 to 10 items). Six species of fruit. varying in size, shape and texture. depending on the species, were tested. Fruit size was measured by its total length. Five size classes were distinguished: I (< 0.5 cm), II (0.5-1.0 cm), III (1.1-2.5 cm), IV (2.6-5.0 cm), V (5.1-10.0 cm) and VI (> 10.0 cm). 10 'D_i_gestion Trials General procedures for digestion trials followed those outlined for domestic species on cultivated forages by Schneider and Flatt (1975) and extended to wild species on natural diets by Ullrey et al. (1969). Robbins et al. (1975) and Milton et al. (1980). Prolonged trials using only fruit originally were attempted. These were abandoned. however. after blue duikers lost weight and sickened when fed a pure Kleinedoxa gabonensis fruit diet. Diets containing a single fruit species offered with the readily available and palatable foliage of the sweet potato vine Ipamea batata Two diets were tested. a low quality diet, consisting of ripe K. gabonensis (Trial 1) and a high quality diet (Trial 11) consisting of ripe Ricinodendron headeiotii Both fruits were eaten by bay duiker and to a lesser extent by blue duiker in the wild (Chapter 4). Fruits used in the trials were all gathered from the same trees and at the same stage of ripeness. Fruits were cut into uniform pieces (3 cm) and the seeds removed. Fresh. young. fully expanded leaves of Ipamea foliage were cut from the vines and used in all trials. Every attempt was made to make food offerings as uniform as possible. Experimental animals were caught with nets in the forest near Epulu and kept in adjacent 10 m by 10 m enclosures constructed in the forest. The pens were cleared of small saplings and herbaceous growth. but all trees and larger saplings (> 2.5 cm dbh) were left. Each pen was provided with a lean-to shelter. salt block and water pan. These additions did not appear to alter feeding behavior or food choices. The ll pens were swept daily during the trials to remove extraneous food (foliage, flowers, fruit) falling from the canopy in order to assure that the experimental diets being tested were the only source of food. Two pairs (male and female) of adult blue duiker. E/K and R/D and two individual bay duiker. M (subadult female) and P (adult female) were used for each trial. except for Trial II when only one bay duiker. M, was available. Experimental animals were acclimated to a given test diet for 2-3 weeks before the 7-8 day total collection period. During the trials. ad libitum quantities of each food were presented twice a day. and the amount consumed calculated by subtraction. taking into account changes due to drying as determined by controls. Fecal collections were made every half day on a precise schedule. Urine was not collected but did not appear to contaminate the droppings. Samples of foods fed during each trial and all feces were collected and dried under low heat to a constant weight. Three to seven daily samples of feces and samples of all foods were randomly selected from each experimental unit for chemical analyses. Field dried samples were prepared for analyses by grinding in a Wiley Mill to pass a 1 mm mesh. Total dry matter (DM) was determined by dehydration at 1000 C for 24 hours. Total ash was determined by combustion of sample organic matter at 6000 C. The plant cell wall fractions extracted included neutral detergent fiber (NDF), acid detergent fiber (ADF) and sulphuric acid digested acid detergent lignin (ADL). Extractions for these components followed standard procedures of Goering and Van Soest (1970), with the modification that sodium sulfite 12 was not included in the extracting solution. The difference between NDF and ADF extractions defined the cellulose component of the sample. Total Kjeldahl nitrogen (N) was determined following methods in Horowitz (1970). Levels of condensed tannins (CT) were measured from a methanol extract using the acid hydrolosis method of Bate Smith (1973). Total phenolics (TP) were determined by the Folin-Denis method (Horowitz 1970). Further details on these methods as applied to samples in this study are contained in Oates et al. 1977. All calculations of composition were done on an ash-free dry matter basis to accommodate soil contamination in the food and fecal samples (Mould and Robbins 1982). The proportion of the diet, or of a given nutrient. which disappears during passage through the gut is a measure of apparent digestibility and can be defined by the digestion coefficient (DC) (Schneider and Flatt 1975). where DC = 1 - (DM excreted / DM ingested). Digestion coefficients for nitrogen must accommodate the presence of fecal nitrogen of metabolic origin. The proportion of fecal metabolic nitrogen can be estimated by extracting fecal NDF and determining its nitrogen content. The undigested portion of the nitrogen consumed was assumed to be contained in the NDF residue (Goering and Van Soest 1970). Fecal NDF content was used in calculating nitrogen digestion coefficients. Differences in digestibility between species and between diets were tested by analysis of variance for unequal sample sizes using a pooled error variance to which the unreplicated trial did not 13 contribute (Steel and Torrie 1980: 146). In the tests reported here, probabilities of p < .05 were considered significant. RESULTS .Hgme Range and Mobility The total area utilized by the radio-collared blue duikers over the 2 month study period ranged from 3.7 to 6.4 ha and averaged 5.0 ha. There was no significant difference in the areas covered by males (average 5.2 ha) and females (average 4.8 ha). Over the same period. the collared bay duiker ranged over an area over 3 times larger than the blue duikers (15.2 ha). The blue duikers were primarily diurnal whereas the bay duiker was nocturnal. The average distance between hourly locations for the blue duikers (during daylight recording sessions) averaged 41.3 m/hr (range 25 m - 56 m). The average distance recorded for the bay duiker during night recording sessions was 82 m (range 38 m - 138 m). When the bay duiker moved between grids it moved further than the blue duiker averaging 179 m/move, (range 75 m/move - 223 m/move). The smaller blue duikers' average movement distance was 82 m/move (range 66 m/move - 94 m /move). Gut Capacity Both the total rumen volume and the weight of ruminoreticular contents were significantly greater in the larger bay duiker than in the blue duiker (Table 2-1). On a per body weight basis, bay 14 Table 2-1. Measures of rumen capacity and ratios to body weight in adult blue duiker (C. aonticola) and bay duiker (C. dorsalis). Rumen Capacity Rumen capacity/Body weight Volume Contents Volume Contents (1) (kg) (1) (kg) Blue duiker mean 1.51 0.41 0.41 0.089 4.7 kg SE 0.36 0.02 0.01 0.004 n 19 37 21 33 Bay duiker mean 7.29 1.87 0.48 0.098 21.8 kg SE 1.10 0.12 0.01 0.010 n 5 12 6 11 Difference p<.001 p<.001 p<.01 N.8. (t-test) 15 duiker had significantly greater rumen volume than the blue duiker (p < 0.01). The relative weights of the ruminoreticular contents were not statistically different. Wet weight contents averaged 0.89 kg/kg body weight and 0.98 kg/kg body weight (8.9 X and 9.8% body weight) for blue and bay duiker respectively. Food Handling Handling times for bay and blue duiker varied between food size classes. as well as within food size class between species of fruit (Table 2-2). Both blue and bay duikers handled small items with comparable efficiency. The bay duiker had shorter handling times for many of the larger items (Class IV and Class V) than did the blue duiker, especially for fruits which were tough and fibrous such as.K. gabonensis. Indeed. many large R: gabanensis fruits (size class V) could not be handled by blue duiker. The largest fruits (Class VI). with the possible exception of unripe pods of some legumes. not included in the experiments, could not be handled by either species. The time needed for an animal to fill its rumen with items of a specific size class (rumen-fill time) was calculated from estimates of rumen capacity (above) and data on numbers of differently sized itmes in rumens of wild caught animals (Hart. unpublished data). Although these estimates would be expected to vary, depending on fruit species. they nevertheless provided a measure of major differences in relative time costs for large and small duikers to specialize on feeding on differently-sized items (Table 2-3). In both species rumen—fill time decreased with increasing 16 Table 2-2. Handling times for fruits by an adult female blue duiker (C. aonticola) and a subadult female bay duiker (C. dorsalis). Handling Time (sec) Item Slue duiker Bay duiker— size class Fruit Species n mean SD n mean 80 ll Ficus capenszs 4 10.0 5.0 5 8.9 4.1 III Ceitis adoifi-fridericii 1 14.0 -- 14 6.2 1.2 Klainedoxa gabonenszs 7 13.9 2.4 5 7.5 1.0 IV Donella pruniforais 21 16.6 2.3 18 17.3 2.4 KIainedoxa gabonensis 3 49.3 11.7 5 7.8 1.0 Ricinodendron heudelotil 3 44.0 12.7 7 23.3 6.5 Ficus aocuso 7 49.6 12.5 not tested V Klainedoxa pabonensx: ' 2 51.5 24.4 5 21.7 3.0 ' Host size size V K. gabonensis fruits were too large for blue duiker to handle. Handling times reported do not include these fruits. Table 2-3. 17 Estimated number of fruits to fill rumen, fruit handling times and time to rumen-fill for fruits of different sizes for blue duiker (C. aonticola) and bay duiker (C. dorsalis). ‘ Item Number Unit Time to Size to Fill Handling Time Rumen Fill (seconds) (minutes) Blue Duiker I 500 10 83.3 11 250 10 41.7 III 70 14 16.3 IV 40 40 26.7 V 10 49 f 8 f (many too large too handle) VI 7 too large to handle Say Duiker I 4,000 10 666.7 11 1,350 9 202.5 111 200 7 23.3 IV 50 16 13.3 V 20 22 7.3 VI 10 too large to handle Time to rumen-fill (minutes) estimated for each food size as Number to Fill x Unit Handling Time. 18 fruit size up to a point where items were too large to handle. There was a marked difference, however, between the two duikers in the relative time reduction in feeding on the smallest and largest food species. In blue duikers, estimated rumen—fill times increased 10.4 times from 8 minutes (for size class V fruits which could be swallowed) to 83 minutes for foods of the smallest size class. In the bay duiker, rumen-fill time increased 91.4 times over the same range of food sizes. It seems even from these limited data that the bay duiker would have to spend considerably more time to fill its rumen than the blue duiker when feeding on small-sized items. Digestion Trials The nutritional composition of the fruits and foliage utilized in the feeding trials is shown in Table 2—4. Total dry matter. fiber and lignin levels were higher in R3 gabonensis fruits than in R. headelotii fruits, while percentage nitrogen was lower (p < 0.05). Ipomea foliage, in contrast to the fruits. had lower dry weight content, but higher percent dry weight nitrogen. Fiber levels in Ipomea foliage were intermediate between the two fruit species and lignin content was relatively high. Condensed tannin levels in all ’ foods were relatively low. Total phenolic content of the fruits. on the other hand. was quite high, averaging almost one sixth dry weight content in Klgabonensis and nearly one third dry weight content of R. heudeiotii. The identity of the phenolic compounds has not been determined. although samples of R. heudelotii were particularly rich in hydrolysable tannins (P.G. Waterman pers comm). 19 .c-uauc >co -.e*-g-¢ mac co non-n .c- uoa_-> Logan _.- .c-uu-e >cn .ouoa co non-n cc. nu._oe-:m .uuoh we. ue.cce» name-econ .guc age so. um:.o> ..c. egos. a-u.u o.a-u co» .uco.am.>-u scene-u. ecu. new-— age -=-> . .nu.o. .o..o. .sn.o. .o.n. .5... .o.n. .eo.~. .o.o. n a c -.~.-a no.~ oo.o oo.¢ «.3 5.5. n.3u n.~ o.s. .no. ecu... cocoa. .oo.n. .uo.o. .oo.o. .o.o. ..... .n.u. .oo.o. ...o. n a : ...o.~u=ec u...n on.. no.. o.. 9.0. n.n. .o.o n.c~ 3.3.4 ca.c .ocuguuog.u.¢ .oo.¢. .ns.o. .o..o. .o... .5... .n.n. .un... ..... c a c n..eacoa-s .n.n. 00.. oo.o . 0.3 c.o~ o.~¢ n... 3.5“ u.:c4 ca.c caoe.¢.~.x mp pg 2 49¢ mac an: no. 4a¢ no ...cou-: .-.uon. noon c.=... ta... ta... uu..ocoga u:.=c-p cooocu.z peace-Lou «zoos-ace scone-use gnu coo».- ..uoh nose-econ u.u¢ u.u¢ .acuaoz c.co.. sea . co.».uoaeou ounce-ace; ..u_-neou .9. 2.3.5: >oa ecu .m~ou_.eo- .u. cox.:u o:.a g... ....c. .¢.o... 2. u......a anon. .o ca....oa.ou .aco...c.=z ..-~ ..... 20 The purpose of the trials was to test digestibility of a given diet by both duiker species. This was not possible on the low quality fruit diet (Trial I) because the two species selected significantly different portions of fruit and leaf (Table 2-5). The bay duiker averaged 72% K2 gabonensis fruits, whereas blue duikers averaged 588 fruit. The bay duiker's consumption of K. gabonensis fruit resulted in a diet higher in NDF and cellulose (p <.05) and lower in nitrogen (p (.01). Only ADL densities (6.73) were the same for both species on this diet. On the high quality fruit trial (Trial II) the composition of the diets of both duiker species was the same (Table 2-5). Both species averaged over 80% fruit on a dry weight basis. Fiber composition (16* NDF, 10% cellulose. 28 ADL) and nitrogen levels (1.6x - 1.7%) were also equivalent in both species diets and lower than on the K1 gabonensis diet. . There was no difference in dry matter digestion between the two duiker species in Trial I despite differences in the relative composition of the diets (Table 2-6). Both species digested comparable percentages of dietary NDF even though the bay duikers' diets averaged twice the levels of those in the blue duikers' diets. Cellulose digestion by the bay duikers averaged even higher than the blue duikers (67.6% versus 60.4%. p < .05, n = 2). though here again levels in the bay duikers's diets were almost three times those of the blue duikers (Table 2-5). Dry matter digestion in both species' diets did not differ in Trial II (Table 2-6). Both dry matter and cellulose digestion were Table 2-5. 2]. Percentage dry weight composition of diets of blue duiker (c. eonticola) and bay duiker (c. dorsalis) on feeding trials containing K. gabonensis fruit and I. batata foliage (Trial 1) and R. heudelotii fruit and I. batata floiage (Trial 11). A) Percent Composition of Diets blue duiker bay duiker Cellu- Cellu- Rep- Fruit NDF lose ADL N Rep- Fruit NDF lose ADL N licate licate I E/K 59.1 37.1 16.5 6.6 2.34 H 73.5 39.4 17.7 6.7 1.73 RID 56.2 36.6 16.3 6.7 2.46 P 70.8 38.8 17.5 6.7 1.84 Trial 1 . Average 57.7 36.9 16.4 6.6 2.40 72.2 39.1 17.6 6.7 1.79 II E/K 90.6 14.7 9.6 2.0 1.39 H 81.7 16.0 9.8 2.4 1.74 RID 80.7 16.1 9.9 2.4 1.78 Trial 11 Average 85.6 15.4 9.8 2.2 1.59 8) Tests of hypotheses: Percentage composition of diets‘ Cellu Fruit NDF lose ADL N 1) blue vs bay, low quality diet (Trial 1) 4 e f N.S. f 2) blue vs bay. high quality diet (Trial 11) N.8. N.8. N.S. N.8. N.S. 3) high quality diet vs low quality diet. blue 9 is *9 44 4 4) high quality diet vs low quality diet, bay N.8. 9! am 44 N.S. . NeSeg p < 0.01 difference not significant; difference siginficant, p < 0.05 (I); (99). Table 2-6. 22 Digestion coefficients for blue duiker (C. eonticola) and bay duiker (C. dorsalis) on diets of K. gabonensis fruit and I. batata foliage (Trial I) and R. heudelotii fruit and I. batata foliage (Trial II. A) Digestion Coefficients ‘ blue duiker bay duiker Dry Cellu- Dry Cellu- Trial Rep- Hatter NDF lose N Rep- Hatter NDF lose N licate licate I E/K 76.7 59.1 61.8 77.1 H 76.8 61.6 64.4 73.8 RID 74.9 57.2 58.9 75.2 P 80.0 65.5 69.6 77.4 Average 76.3 58.2 60.4 76.2 Average 78.4 63.3 67.0 75.9 11 Elk 81.0 57.2 72.2 58.3 H 83.4 62.1 86.0 66.7 RID 83.5 55.5 76.1 63.7 Average 82.3 56.4 74.2 61.0 8) Tests of Hypotheses: Digestion Coefficients.b Dry NDF Cellu N Hatter lose 1) blue vs bay, low quality diet (Trial A) N.S. N.S. 4 N.S. 2) blue vs bay, high quality diet (Trial 8) ".8. . i N.8. 3) high quality diet vs low quality diet, blue 4 N.S. 9* f 4) high quality diet vs low quality diet, bay N.S. H.8. f f Digestion Coefficients (DC) I 1-(Dry Ht Excrete/Dry Ht Ingest) times 1001. H.S., difference not significant; 4. difference significant, p < .05; ..’ p < .01. 23 higher in the blue duiker in Trial II than in Trial 1. and cellulose digestion was higher for the bay duiker . The bay duiker also digested cellulose more completely than the blue duiker on this trial. Patterns of nitrogen digestion were at variance with those of plant cell wall constituents. Both duikers exhibited comparable nitrogen digestibilities on the K; gabonensis diet (76%). On the R. heudelotii diet. however. the bay duiker's digestion coefficient fell to 67% whereas bay duikers' dropped to 61%. Lignin, a complex plant cell wall constituent is theoretically indigestible (Van Soest 1982), and ADL digestion coefficients were expected to be zero. Values close to the expected were obtained on the .K. gabonensis diet (Table 2-8). In contrast to these results, there was an accumulation of an apparent lignin-like artifact in the feces of both duiker species on the R. heudEIotii diet. Apparent fecal ADL increased 89% in the bay duiker trial and an average 126% in the blue duiker trial. DISCUSSION Though limited by small sample sizes. especially for the bay duiker, the results of the mobility and food handling studies indicated that size and spatial distribution of food is likely to have an impact on foraging and food choice in duikers of differing body size. It was not possible to relate the movements and home range use of collared duikers to specific food distributions. It is unlikely too that all animal movements entailed food searching. Nevertheless, the 24 Table 2-7. Apparent lignin digestion coefficients for blue duiker (C. eonticola) and bay duiker (C. dorsalisl on diets of K. pabonensi: fruit and I. batata foliage (Trial I) and R. heudeiotii fruit and I. batata foliage (Trial ll)‘ Blue duiker Say duiker Trial E/K R/D H P I ’15.0 ’9.0 6.0 14.0 Average -12.0 10.0 11 '160.0 '91e0 '99.0 Average -126.0 ‘ Digestion Coefficients DC = 1-(Dry Ht Excrete/Dry Ht Ingest) times 100%. Negative values indicate accumulation of lignin artefact in the feces. 25 larger home range and apparently greater mobility of the bay duiker indicated a greater potential for the larger animal to find and exploit widely-dispersed food resources. The fact that the collared bay duiker tended to make direct, relativlely long-distance moves during its activity periods suggested that the animal visited specific. known food sources. and tended to avoid patches in between. Based on the amount of time needed to manipulate and swallow a food item, both the bay and blue duikers were equally adept at handling small-sized foods. The large species required less time to handle large foods. In both species. however, rumen-fill times were minimized by feeding on the larger items up to the apparent limits of mouth size. The major difference between the two species became apparent in the relative time costs in filling the rumen with small-sized items. Based on calculations. it was estimated that it would take the larger bay duiker eight times longer than the blue duiker to fill its rumen with foods of the smallest size class (< 0.5 cm), and 4.5 times longer to fill its rumen with items of the next size class (0.5 - 1 cm). Thus while both species incurred time costs in feeding on small items, these are relativley greater for the bay duiker than for the blue duiker. These estimates assume no search time between items. The addition of search time would clearly reduce the potential utility of small items to the bay duiker even more. Diets of free-ranging animals were diverse (Chapter 4). Test diets were simplified to examine the relative importance of the fiber content of the diet in affecting preference and digestibility. An evaluation of the feeding trial results in the light of the initial 26 hypotheses, nevertheless pointed to a role for multiple factors in determining dietary quality. Hypothesis 1. The large—bodied bay duiker was hypothesized to have higher digestion coefficients than small-bodied blue duiker on the low-quality R2 gabonensis diet. This hypothesis could not be evaluated on diets of equal quality because the two duiker species selected foliage and fruit in different proportions. The fact that bay duiker replicates had equal or higher digestion coefficients for cell wall components on diets with twice the plant cell wall levels provides, nevertheless, supported for the hypothesis that larger-bodied ruminants were better able than small-bodied species to utilize foods over a range of dietary quality. (see reviews in Owen-Smith 1980 and Demment and Van Soest 1985). The results of Trial I were somewhat confounding, however, in that despite its higher fiber content, K. gabonensis fruit was favored to foliage. at least by the bay duiker. Fiber alone is thus not the only dietary component affecting food preference and quality. Hypothesis 2. Digestion coefficients for both species were predicted to converge on the high quality diet. Dry matter digestion coefficients were equivalent for both species but fiber fractions were digested more efficiently by the bay duiker. The results of this trial were not confounded by differences in diet composition, as both species ate equivalent proportions of leaf and fruit. Hypothesis 3. High quality (low fiber) diets were predicted to be more digestible than low quality (high fiber) diets. This generally proved to be the case for dry matter and at least some plant 27 cell wall components of the diet. The lower apparent nitrogen digestion on the R. heudeIotii diet, however. was unexpected. This might have been due due to high levels of phenolics in this fruit. Recent studies have ascribed a digestion inhibitory role to plant phenolics (Mould and Robbins 1982), but results are not entirely unequivocal (Horvath 1981). The accumulation of indigestible lignin-like compounds in the feces of animals on Ricinadendron diets was similar to those reported for other species fed diets high in tannins (McLeod 1974). The urine of duikers on the Ricinodendron diet was dark reddish brown. This has also been reported in elk on diets high in total phenolics (Mould and Robbins 1982). and is similar to the hematuria reported in tannin-dosed sheep (McLeod 1974). R. heudelotii fruits may well be somewhat toxic. The fact that they were also highly digestible and eaten freely is an indication that phenolic compounds will not deter feeding, though they may limit the amount of a food which may be ingested (Freeland and Janzen 1974). CONCLUSIONS Comparisons of mobility. food handling and diet digestion in two species of duikers, a large species, the bay duiker and a small species the blue duiker, demonstrated the potential for size-related differences in foraging behavior in the forest ungulate community. The bay duiker was more mobile and could better digest foods over an apparent range of quality. Relative to the blue duiker. 28 however, it was more constrained by its larger total needs from effectively utilizing small food items. especially if foraging on these involved extensive searching. The digestion trials revealed that plant cell wall content was one element of fruit quality, however, digestibility appeared to be affected by phenolic compounds in fruits as well. L I TERATURE C I TED 29 LITERATURE CITED Bate-Smith. E.C. 1973. Tannins in herbaceous legumes. Phytochemnistry 12: 1809-1812. Bell, R. 1970. The use of the herb layer by grazing ungulates in the Serengeti. Pages 111-124 in A. Watson. editor. Animal populations in relation to their food resources. Symposium of the British Ecological Society. Aberdeen. UK Bunnell, F.L.. and M.P. Gillingham. In press. Foraging behavior: dynamics of dining out. In R.J. Hudson and R.G. White, editors. Bioenergetics of wild herbivores. CRC Press, Roca Baton. Florida. USA. Clutton-Brock. T.. and P. Harvey. 1983. The functional significance of variation in body size among mammals. Pages 632—663 in J. Eisenberg and D. Kleiman, editors. Advances in the study of mammalian behavior. Specical Publication No. 7, American Society of Mammalogists. Demment. M.W. 1982. The scaling of ruminoreticulum size with body weight in East African ungulates. African Journal of Ecology 20: 43—47. ‘ Demment. M.W. and P.J. Van Soest. 1985. A nutritional explanation for body-size patterns of ruminant and nonruminant herbivores. American Naturalist 125: 641-672. Dubost, G. 1979. The size of African forest artiodactyls as determined by the vegetation structure. African Journal of Ecology 17: 1-17. Dubost, G. 1984. Comparison of the diets of the frugivorous forest ruminants of Gabon. Journal of Mammalogy 65: 298-316. Emmons. L.H., A. Gautier-Hion and G. Dubost. 1983. Community structure of the frugivorous-folivorous forest mammals of Gabon. Journal of Zoology (London) 199: 209-222. Foose, T.J. 1982. Trophic strategies of ruminant versus nonruminant ungulates. Unpublished PhD thesis, University of Chicago. Freeland. W.J., and D. Janzen. 1974. Strategies in herbivory: the role of plant secondary compounds. American Naturalist 108: 269-289. Goering, H.K. and P.J. Van Soest. 1970. Forage fiber analyses (apparatus. reagents, procedures and some applications). 0.8. Department of Agricultrue. Agricultural Handbook Number 379. Hanley, T.A. 1980. The nutritional basis for food selection by 3O ungulates. Journal of Range Management 35: 146-151. Hoffmann, R.R. 1973. The ruminant stomach. East African Monographs in Biology 2. East African Literature Bureau, Nairobi. Kenya. Horowitz, W. (editor). 1980. Official methods of analysis of the Association of Official Analytical Chemists (11th edition). Association of Official Analytical Chemists, Washington, D.C., USA. Horvath. P.J. 1981. The nutritional and ecological significance of Acer-tannins and related polyphenols. Unpublished masters thesis, Cornell University, Ithaca, New York. Kingdon, J. 1982. Duikers. Cephalophinae. East African mammals: an atlas of evolution in Africa III(C) (Bovids): 263-279. McLeod, M.N. 1974. Plant tannins—- their role in forage quality. Nutrition Abstracts and Reviews 44: 803-815. Milton, K.. P.J. Van Soest. and J.B. Robertson. 1980. Digestive efficiencies of wild howler monkeys. Physiological Zoological 53: 402-409. Mould, E.D. and C.T. Robbins. 1982. Digestive capabilities in elk compared to white-tailed deer. Journal of Wildlife Management 46: 22-29. Oates. J.. P.G. Waterman. and G. M. Choo. 1980. Food selection by the south Indian leaf-monkey. Presbytis johnii in relation to leaf chemsitry. Oecologia 45: 45-56. Owen—Smith. N. 1980. Factors influencing the transfer of plant products into large herbivore populations. Pages 359-404 in B.J. Huntley. and B.H. Walker. editors. The ecology of tropical savannas. Springer-Verlag. Berlin. Robbins, C.T., P.J. Van Soest. W.W. Mautz, and A.A. Moen. 1975. Feed analyses and digestion with reference to white-tailed deer. Journal of Wildlife Management 39: 67-79. Schneider. B.H. and W.P. Flatt. 1975. The evaluation of feeds through digestibility experiments. University of Georgia Press. Athens. Georgia. USA. Schoener. T.J. 1974. Resource partitioning in ecological communities. Science 185: 27-39. Short, H.L. 1963. Rumen fermentation and energy relationships in the white-tailed deer. Journal of Wildlife Management 28: 445—458. 31 Short, H.L., R.M. Blair and C.A. Segelquist. 1974. Fiber composition and forage digestibility by small ruminants. Journal of Wildlife Management 38: 197-209. Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of statistics. Second edition. McGraw—Hill. New York, USA. Ullrey, D.E.. W.G. Youatt. H.E. Johnson, L.D. Fay, B.L. Schoepke, W.J. Magee. 1969. Digestible energy requirements for winter maintenance of Michigan white-tailed does. Journal of Wildlife Management 33:482-490. Van Soest, P.J. 1982. Nutritional ecology of the ruminant. O and 8 Books. Corvallis. Oregon, USA. CHAPTER THREE A Study of Factors Affecting Food Preference and Intake in Two Species of Duikers Six species of duikers (genus Cephalophus) and the chevrotain (Hyemaschus aquaticus) spanning a size range from 5 to 70 kg co-occur in the humid evergreen forests of central Africa (Emmons et al. 1983; see also Chapter 4). The diets of all species are dominated by fruits and seeds (Dubost 1984). Gautier-Hion et a1. (1980), in a comparative study of primary consumers, including the duikers and chevrotain, of the forests of Gabon, suggested the hypothesis that complimentarity in diets of different frugivores may reflect differences in their physical and physiological capacities in response to a range of nutritional and physical properties of the fruits. While this hypothesis suggests possible mechanisms for coexistence in related species of frugivores, it has been difficult in the field to determine how different species respond to variations in speciifc dimensions of the fruit resources. This paper reports the results of palatability trials conducted with two species of captive duiker. These trials were designed to ascertain the importance of food nutritional quality. 32 33 independent of food distribution. as a factor affecting diet choice. The duikers chosen. the blue duiker Cephalophus monticola (4.7 kg adult weight) and the bay duiker. C. dorsalis (22.0 kg adult weight) were at opposite ends of the local body-size spectrum within the guild of frugivorous ungulates. Thus the relationship of body size and food preference will also be considered. The foods tested in the trials included fruit and foliage species available to duikers in their natural environment. The questions addressed by these experiments were: 1) What are the duikers' food preferences and are they consistent from test to test? 2) What are the nutritional and chemical correlates of food preference? 3) What is the relationship between the preference for a diet and its intake? These specific questions are posed in a broader ecological framework that will also be addressed by a consideration of the following problems: How do preferences for fruits compare with those of foliage and why is foliage uncommon in the diet? How are food preference and intake related to body size in duikers? Plants contain a wide array of compounds which affect their quality as food for primary consumers (Freeland and Janzen 1974, Crawley 1983). One way to relate these to food preference and diet is to consider these compounds in terms of the costs and benefits of ingesting a given food. The major nutritional benefits of a food are 34 contained in the ready sources of protein and energy which can be contributed to the diet (Barnes and Marten 1979, Van Soest 1982, Waterman 1984). Other groups of compounds including plant cell wall components and certain defensive compounds are difficult to digest. reduce palatability. inhibit digestion or are even toxic (Rosenthal and Janzen 1979 and references therein. Waterman et al. 1980). These can be thought of as the costs of ingesting a particular food. Compounds in both these cases are concentrated in different ways in plant tissue and may determine the value of different plant parts to consumers (McKay 1979). Differences in the relative nutritional quality of foliage. fruits and seeds may be especially apparent (C. Hladik 1978. McKey et al. 1981). Leaves serve the plant as the primary organs of photosynthesis. They contain appreciable quantities of protein (Mattson 1980) but often have high levels of cell wall constituents which make them poor sources of energy (Van Soest 1982, Waterman 1984). Fleshy fruits protect the ovule as it ripens but when ripe must attract animal seed dispersers (Snow 1971. van der Pijl 1972, Janzen 1983). Relative to leaves. fruit pulp often contains less protein but may have concentrations of sugars. pectin and other readily-digested carbohydrates and may contain appreciable levels of energy-rich oils and fats (Waterman 1984). Some fruits also contain toxins (Herrera 1982). Seeds protect the embryonic plant until germination and thereafter provide energy for seedling establishment. Seeds generally 35 contain high levels of oils or starches in the endosperm or cotyledon. In some species nitrogen levels also may be high. Seeds may be protected. however. by toxins and digestive inhibitors (Janzen 1971. McKay 1979, Waterman 1984). or by hard, resistant seed coats in which case seeds are essentially inedible ballast in the diet (Herrera 1981). METHODS tflalatability Trials Experiments were conducted at Epulu in the Ituri Forest of Zaire from November. 1982 to March. 1983. See Chapter 4 for a description of the Epulu area and its fauna. Free-ranging duikers are known to eat a large number of fruit species (Dubost 1984, see also Chapter 4). It was possible to test preferences for only a minority of these, however. Five to six food species were selected for each of three trials, A, B and C conducted during the late wet, dry and early wet seasons repsectively (Table 3-1). Preferences were determined for these species and related to their nutritional composition. The nutritional correlates of preference then provided a basis for evaluating the potential food quality of other foods which were not tested. The amount of a food eaten when offered by itself without alternative choices provides an inadequate measure of food preference since animals may be able to compensate for a poor quality food in the 36 Table 3-1. Design of fruit palatability trials. Test Animals‘ Number Trial Season Food Species Blue Duiker Say Duiker A late wet 5 E/K, RID H 8 early dry 5 E/K H C early wet 6 E/K none ‘ AbbreViations refer to individual captive animals. See Chapter 2 for details. 37 diet by eating more of it. Thus the food eaten in greatest quantity is not necessarily the most favored. Cafeteria-style presentations of a number of foods simultaneously provides a means of assessing preference. Lack of adequate numbers of animals and fruit supplies in this study. however, made it impossible to test all possible combinations of even three or four food species offered at one time. As a compromise, a test protocol in which foods were offered two species at a time was used. Each combination was offered in known-weight ad libitum quantities with Ipomea batata foliage for a 12 hour period. The preferred fruit was considered that species eaten in the greater quantity. If the amount of each choice eaten differed by less than 10% both species were considered equally preferred. All possible fruit combinations were replicated (n = 2) if availability of forest fruits permitted. Combinations were presented randomly so that animals did not have the opportunity to habituate to any given choice. The rank preference of a given fruit species, 1 was based on the summed frequency of test combinations in which i was the preferred choice. Tests in which both fruit offerings were eaten in equal amounts were scored as one half in accumulating rank preference values. This protocol had several advantages. First it allowed preference for every food species to be evaluated in relation to every other in the trial. Secondly. it provided a means of examining whether food preferences were consistent. Consistent food preferences were 38 defined as tests in which the same species was preferred in both replicates of a specific two—species offering. Finally. it permitted an examination of the relationship between the quality of a diet and its intake, where a measure of diet quality is provided by the preference rank of the preferred of the two constituent food species. In testing all combinations of foods, two species at a time, animals were presented diets in which one or both fruits offered were preferred. They were also forced to choose between combinations of less favored foods. Of interest here was not only which food was selected in each combination, but how much total food was eaten. The relative ability of an animal to maintain intake over a range of diet quality provides some indication of the importance of food quality as a constraint in foraging. Although diets of free-ranging animals often contained a number of food species. between 67% and 100% of the diet by weight could generally be accounted for by one or two food species (J. Hart, unpubl. obs.. see also Chapter 4). Thus the two-species food offerings used here were not necessarily at variance with the animals' natural dietary habits. Test animals included two mated pairs of adult blue duikers identified as E/K and R/D (Male/Female). Each pair was treated as an experimental unit and compared with a single subadult female bay duiker. labelled M. The experimental animals had been captured locally and were kept in 10 m by 10 m pens in the forest. These were swept daily during experiments to remove extraneous foods fallen from the canopy. Preferences of both blue duiker pairs and the bay duiker were 39 determined in Trial A. Only blue duiker pair E/K and the bay duiker were studied in Trial B. The blue duiker pair E/K was tested alone in Trial C. Available supplies of test foods from the forest permitted two replications of almost all test combinations in Trial A. This was not possible in Trials B and C because fruits were more scarce in the forest. Nevertheless. the majority of fruit offerings were replicated in each test. Ipomea batata foliage intake was measured in Trial A, but not in Trials B and C. Most test foods were presented entire, just as they would be encountered in the forest. Some fruits of I. wombolu, K. gabonensis handle effectively. Fruits of these species were cut into 2 — 3 cm pieces and the seeds removed, before being offered to all test animals. Preference tests for forest foliage differed from those for fruits because duikers often refused to eat large quantities of many foliage species. To test preferences for foliage. an alternative protocol was established. The test animals (blue duiker pair E/K and bay duiker M) were put on a diet of unripe R). gabanensis fruits. Newly-expanded fresh leaves and petioles of different foilage species were presented to duikers in small. loose bunches of similar size. Three selection classes were established based on the amount of leaves eaten: Foliage uneaten (scored 0), less than half the bunch eaten (scored 0.5) and most or all of the foliage consumed (scored 1.0). Each species of foliage was presented on three different occasions and the scores summed and ranked. Ten species (both climax and pioneer) were tested. These comprised some of the more common species whose fallen 40 leaves were likely to be most available to free-ranging animals. Preferences for four common understory species was also tested. Nutritional Composition Samples of all fruits. seeds and foliage offered in trials were retained for analyses. In fruits which contained hard, inedible seeds which are usually regurgitated. the seeds were removed from the flesh and weighed separately. Total dry matter (DM). total phenolic content (TP), condensed tannin content (CT), acid detergent fiber content (ADF) and total nitrogen content (N) were determined for edible parts of all samples by Dr. P.G. Waterman, Department of Pharmaceutical Chemistry, University of Strathclyde, Glasgow, Scotland. Procedures followed those detailed in Oates et al. 1980. For each food species. edible dry matter yield (Y), as defined by (Herrera 1981), was calculated as Y = (T - S) x DM where T equals the total fresh weight (g). S is the inedible seed weight (g) and DM equals percentage dry matter of the edible portion. In cases where S = O. Y simply equals the grams dry matter in the sample. Ratios of nutrient to digestion inhibitors were calculated for each food tested based on nutritional composition data. Three ratios were considered: N/ADF, N/(ADF + TP) and N/(ADF + CT). Each ratio was multiplied by the edible dry matter content of the food to provide a measure of adjusted dry matter yield of the food (Y 0" Values of all nutritional measures are expressed in milligrams 41 per gram (mg/g) fresh weight of the food. The plant constituents measured in this study do not comprise the total array of nutritionally significant compounds. The chemical constituents chosen for analysis here have been associated with preference in other studies (see bleow). Their use here provides a means of comparing food preferences and diets of duikers with those of other species. Spearman's coefficient of rank correlation (rs) (Steel and Torrie 1982) was used to test association of rank preference with other variables. Probabilities of p < .10 will be considered significant. RESULTS _Prgference Ranks Rank preferences for each food (Table 3-2) were determined by four to eight test offerings to each of the animals or animal pairs being tested (Appendix 3—A). Four species of ripe fruit and one seed species were used in Trial A. Three species of ripe fruit and two species of unripe fruit were tested in Trial 8. In Trial C four unripe fruits were tested. The fruits tested in Trial C included two species tested in Trial B: Kleinedaxa gabonensis (unripe fruits tested in B. ripe fruits tested in C) and MMsanga cecropioides (full size. unripe fruits tested in B, ripe fruits and small unripe fruits tested in C). In all three trials, the first ranked species was the 42 Table 3-2. Rank preferences of fruits and seeds as determined by palatability trials with captive blue duiker (C. eonticola) and bay duiker (C. dorsalis). Blue duiker Bay duiker E/K RID H Food _________________________________ Food Species Symbol Type rank tests rank tests rank tests TRIAL A Brachystegia Iaurentii 8.1. ripe seed 1 8 1 8 1 B Canaries schweinfurthii C.s. ripe fruit 2.5 8 2 8 5 6 Donella pruniforeis D.p. ripe fruit 2.5 8 3 8 3.5 6 Ricinodendron heudelotii R.h. ripe fruit 4 7 4 8 2 6 Phyllanthus pynaertii P.p. ripe fruit 5 8 5 8 3.5 5 TRIAL B Irvingia moebolu l.w. ripe fruit 1 7 2 6 Klainedoxa gabonemsis K.g. unripe fruit 2 8 1 5 Husanga cecropioides H.u. unripe frUit 3.5 7 3 5 Ficus aocuso F.m. ripe fruit 3.5 6 4 4 Klainedoxa trillesii K.t. ripe fruit 5 7 5 4 TRIAL C Husanga cecropioides H.r. ripe fruit 1 8 Irvingia grandifolia 1.9. unripe fruit 2 8 Klainedoxa gabonenszs K.g. ripe fruit 3 6 Dacryode: edulis D.e unripe fruit 4 8 CIeistanthus eicnelsonii C.s. unripe fruit 5 7 Husanga cecropioide: H.u. unripe fruit 6 6 43 preferred choice in every test combination. The lowest ranked food offering was never the preferred choice in any combination. Foods with intermediate preference ranks were the preferred choices in some combinations but not in others. Preferred food choices were generally not eaten to the exclusion of the other fruit offering. In most cases. at least small amounts of the less preferred choices were eaten and in some combinations substantial amounts were consumed. Ipomea batata foliage was generally eaten in small amounts, and sometimes not at all except when combinations of less preferred foods were offered (see further discussion below). In Trial A, Brachystegia laurentii seed was the most preferred food by all duikers. Phyllanthus pynaertii was the least-preferred blue duiker food item, while canarium schweinfurthii was the least-preferred bay duiker food item. Food rank preferences (n = 5) of blue and bay duiker were not correlated in Trial A (E/K with M, r8 = .289, p > .10; R/D with M, rS = .205. p > .10). In Trial B, ripe fruit of Irvingia wombaIu and unripe fruit of Klainedbxa gabonensis were the first and second choices of the blue duiker with these same two species in reverse order the top ranked foods of the bay duiker. For both duikers, the least—preferred food was the ripe fruit of K. trillesii. Food rank preferences of blue and bay duikers were correlated in Trial B (rs = .872. p < .10. n = 5). In Trial 0. ripe MMsanga cecropioides was the highest ranked food of the blue duiker pair tested while small unripe fruits of 44 the same species had the lowest rank. Blue and bay duiker preference scores for canopy foliage (Table 3-3) were correlated (rS = .902, p < 0.01, n = 10). Understory foliage species had uniformly low preference scores. In the one case tested (Gilbertiadendron dewevrei). understory foliage (fully opened but still limp) had a lower preference score than canopy foliage of the same age. Replicated test combinations in which the same fruit was the preferred choice in both offerings were labelled consistent preference choices. Given the three possible outcomes of a choice between "Food A" and "Food 8": A > B, A < B, and A = B and duplicated tests, only one third of choices would be expected to be repeated if food choices were random. If more than one third of choices were repeated (Chi square test of exact probability), it was concluded that food choice was not random. In each of the trials only one or two of the blue and bay duiker replicated tests did not result in the same choice of preferred food. This was significantly less than would be expected if food choices were random (Table 3-4). Even with random presentations of food combinations, the duikers consistently favored the same fruits. The strong correlation of rank food preference values between the two blue duiker replicates in Trial A (rs = .975, p < .01) provided further evidence that food choices were consistent, and thus likely to be correlated with particular qualities of the fruits. Consistency in foliage trial scores was evaluated in the same manner. For leaves, evidence for consistent scores between replicate 45 Table 3-3. Scores and preference ranks for foliage species offered eonticola) in palatability trials with captive blue duiker (C. and bay duiker (C. dorsalis). Blue duiker Bay duiker Foliage Species Score Rank Score Rafik CANOPY FOLIABE Alstonia boonei 3.0 1 3.0 2 Ricinodendron heudelotii 2.5 2 3.0 2 Phyllanthu: pynaertii 2.0 3.5 1.5 4.5 Cola Iateritia 2.0 3.5 3.0 2 Albizzia pueeifera 1.5 6 1.0 6.5 Dacryodes edulis 1.5 6 0.5 8.5 Gilbertiodendron dewevrei 1.5 6 1.5 4.5 Canariua schweinfurthii 1.0 8 1.0 6.5 Erythrophleue suaveolens 0.5 9 0.5 8.5 Cleistanthus aichelsonii 0.0 10 0.0 10 UNDERSTDRY FOLIABE Cilbertiodendron dewevrei 1.0 1 Scaphopetalua dewevrei 0.5 2.5 Alcnornea floribunda 0.5 2.5 Brachystegia Iaurentii 0.0 4 Table 3-4. 46 Consistency of preference choices between replicated presentations of the same food combinations within trials. Animal Choices made Significance Trial Species Replicate Consistent Inconsistent Choices Choices X2 P FRUIT TRIALS A blue duiker EIK 8 1 12.50 (.005 A blue duiker RID 8 1 12.50 (.005 A bay duiker H 5 1 6.75 (.01 8 blue duiker E/K 7 1 10.50 (.005 8 bay duiker H 2 0 --- --- C blue duiker EIK 5 2 4.67 (.05 FDLIASE TRIALS blue duiker E/K 6 8 0.61 N.S. bay duiker H 6 4 3.30 (.10 47 presentations was not as strong as for fruits (Table 3-4). This may have been due to the fact that only small quantities of foliage were generally eaten at one time. It was noted, however, that animals generally sampled most leaf species when they were presented, even if they did not consume them in quantity. _Qomposition of Foods The fruits and seeds offered in the trials A - C varied in composition (Table 3—5). Dry matter yield values ranged from a low of 170 mg/g fresh weight in Ficus mucuso and Musanga cecropioides to a high of 810 mg/g fresh weight in Brachystegia laurentii seed. Total phenolic content ranged from almost 0 to 74 mg/g. condensed tannin content from 0 to 83 mg/g, fiber from 22 to 150 mg/g and crude protein (N content x 6.25) from 8.3 to 91 mg/g. In comparison with the fruits and seeds. the Ipomea batata foliage, offered during Trials A -C had a low dry matter yield (due to its high moisture content) but relatively high crude protein content. Fiber, total phenolic and tannin levels all fell within the limits exhibited by the fruits and seeds. On a fresh weight basis, rank levels of condensed tannin in fruits were positively correlated with rank fiber levels (rS = .692, p < .01 n = 16). Fiber levels were also correlated with crude protein content (rS = .596 p < .05 n = 16). Total phenolics were not significantly correlated with any other chemical component. Among the foliage species tested (Table 3-6), total phenolic levels ranged from 3 to 52 mg/g and condensed tannin levels from 0 to 48 3.3» 3..» ».. 5.» 33. 3 5. .3...3. .....3 ...... 3.3.3. .33 3.35 ».33 3.3» ».5 3.5 o 3. 3.3.. .3..33 ..3....3.3.3 .3...3. 3.35 3.5». 5... »..» 353 3 33 3.3.. 33...3 ...3...33.. .3....3....3 3.». 3.3» 5.5 3.». 33. o .3 3.3.. .3...3 ...33. 3.33....3 ».u 3.33 n.n 5.3» com o .5 3.3.3 33.. 3.3.3.3335 3333...... o.o~ ..- 5.. n.3. 035 o 35 3.3.3 ea..c: -.~35.33..3 3.5....— ».3. 5.». 5.3 3.. 35. on. 53 3.3.. 33.. 3.3.3.33.3.3 .3...3. u .3... 3.5 3.5» 3.5» 3.3 335 3 35 3.3.3 33.. ......... .~33.....3 ..5. 3.35 N... 3.5 35. 5 »3 3.3.. 33.. 3.333. .33.. 3.55 5.33. 3.35 5.» 3.5 33. 33 3.3.. 33...3 ..3.3..3.3.3 .3...3. 5... 3.». 3.5 3.5» 3.» 3 33 3.3.. .3..c3 ......3..3 ..33...... ».o 5..» 5.3 5.35 035 o 35 3.3.. 33.. 3.33.33 3.3..... m 33... 3.3. 3.33 3.35 5.» 33. 33 »3 3.3.. 33.. .33...... .333....... 3.3 ..35 o 3.33 335 33. .. 3.3.. 33.. ..33..33.3 .3.3..33..3.. ».3. ..33 3.3. 3.35 33» o». 33 3.3.. 33.. .3......3.5 3...... o... 5.35 o..3 3.5. 35» 333 n. 3.3.. 33.. ..33.3....333. .3.....3 ...3 3.33. 3.53 3.». 3.3 o 5. 3... 33.. ..3...3.. 3.5....33..- 3 33... z .33 53 .5 3 33.33....33 33.3.33 c.e3o.m .mp35 3:33:35 u._oc-zm .0333: 5.9 uo-m c.33u.o: .33.u gene-neon .3365 3.3.3w 3.3.33.3o-c 33.3.35 333.333: 333.5 c3.o 5 3.3.0....: .-_-..3 533.33.33.33 :. 33.333o .3663 3o no.3.noaeou .au.aagu .3.» 3.335 49 Ieole 3-6. trials. Chemical CDMpOSltiDn of foliage offered in palatability Values expressed as mq/q of {resh weight. Dry Total Condensed Acid Det Crude Hatter Phenolic Tannin: Fiber Protein Species DH TP CT ADF N CANOPY FOLIABE Alstonie boonei .267 8.8 2.1 73.7 38.7 Caneriue schueinfurthii .274 51.5 10.7 76.5 58.1 Erythrophieul sueveolens .306 18.1 13.8 95.2 83.5 Cole lateritie .310 40.6 129.9 96.4 53.0 Ricznodendron heudeiotiz .243 11.7 2.7 56.4 67.1 Phyllentnu: pyneertii .309 3.1 .00 95.8 50.7 Decryodes eduiis .410 20.5 35.7 167.7 35.7 41612216 guleifere .274 18.4 19.2 100.0 103.8 Cloistenthu: lichelsonii .436 49.3 9.6 159.1 50.1 Gilbertiodendron deuevrei .239 44.7 98.9 94.4 39.2 UNDERSTDRY FDLIABE Brechystegie Iaurentii .385 6.9 17.3 178.3 83.2 Gilbertiodendron deuevrei .258 19.6 41.8 117.7 46.4 Scephopetelue deuevrei .188 12.2 22.9 47.6 35.3 Alchornee floribunde .224 19.3 2.2 46.1 64.5 ‘ Crude protein = N x 6.25. 50 130 mg/g. The fiber content of foliage was more uniform and on average higher than in fruits. Values ranged from 74 to 168 mg/g fresh weight. Crude protein levels were uniformly high, 36 to 104 mg/g fresh weight. Foliage dry matter yield values fell between 240 and 440 mg/g fresh weight. Rank total phenolics was correlated with rank condensed tannins (rS . .563. p < .10, n = 14). No associations of other measured chemical variables were apparent. Compared to the Ipomea batata foliage offered with the fruits and seeds in Trials A - C (Table 3-5). the forest foliage species tested had markedly higher tannin and phenolic levels in all species. Dry matter yield and fiber levels were also higher. but protein levels were comparable. Chemical Correlates of Preference: Fruit No single chemical component measured was consistently correlated with preference over all trials (Table 3-7). For the blue duiker, rank food preference in Trial A was significantly correlated with food levels of crude protein and edible dry matter. For foods offered in Trial B preference was negatively correlated with condensed tannin levels. In Trial C. there was no correlation between blue duiker preference and any of the single chemical variables measured. If ripe .MUsanga cecropioides (see discussion below) was excluded from the analyses. however. food preferences in Trial C became negatively correlated with levels of both fiber and tannins. In the bay duiker trials, the only single variable correlated with preference was condensed tannin levels in Trial B. In contrast with the measures of single chemical variables, 5]. .o. v 3 .333 .no. 3 3 .33 .o.. v a .3 .3uc3u.3.co.m 3 con. .33. .33. 30".: 303.- 33 003.: 33 333.- con. ...2 33033.3 «3..- amo.n 3uo.- n3n.u .nn.. 3.».1 303.- onn.u ...= 33.3 x.m .33.33 33.3 u .3... con. 30.3. 3.~o. 00.. com. com. 3 3.3.- cos. = .3..33 >33 0.3. 33n.o. 3.uo. 33». .no. 3n..- 333 n.3.u 33m. x.“ .3..33 33.3 a .3... .no. 3.~m. 3.~m. now. now. won. ~33. no». 2 .3..33 >33 33303. 003. 303. 333 303.. 33 ono. ooo. oo.. con. 3.: 33uso. .33. .33. 333 n.3. 3 «.3. on». 3.3. on». x.m .33.33 33.3 3 .3... .3.3.a¢..z.x> .303.g¢..2.3> .uacxz.x> > z .33 .u a» .3... .33. 3.... ....33 ..3 3.3.3.33 33.33..3> .3u.3333 3.33.3 3 ..3..33.33 .3. .3..33 >33 3:3 .3.3u..333 .3. .33.33 33.3 33.: 3.3..3 >3...3333.3n c. 333: 33.3.3 33 33.33333 .3u.3333 33.3 3333.333.3 .33. 3o . ... 3333.3.33333 33.33....33 3:3. 3.333.3333 .5.» 3.93» 52 composite measures incorporating levels of both nutrients and digestion inhibitors were consistently correlated with food preferences. Two measures of adjusted dry matter yield, Y x (N/ADF) and Y x [N/(ADF + CT)] had coefficients of rank correlation of 0.60 or above for all animals in all three trials (Table 3~7; Trial C. ripe .MUsanga cecropioldes strongly correlated with food preferences except for the two blue duiker pairs tested in Trial A. Qhemigalmgorrelates of Preference: Foliage Preferences for canopy foliage were negatively correlated with phenolic levels (blue duiker) and fiber levels (blue and bay duiker) but positively. though weakly correlated with a measure of adjusted dry matter yield, Y x [N/(ADF + TP)] (Table 3-8). The three foliage species most favored by both blue and bay duiker, Phylianthus pynaertii, Alstonia boonei and Ricjnodendron heudeiotii, all had the lowest total phenolic and condensed tannin rankings. The two most preferred species also had the lowest fiber rankings. Clejstanthus nichelsanii, Erythrophleun suaveolens and canarium schweinfurthii were the three least preferred species. All exhibited high levels of potentially toxic, or digestive inhibiting compounds. 0. licheisonii had high fiber and tannin content and low protein levels. C. schweinfurthii had low fiber and high protein contents. but also high total phenolics and a strong. resinous odor, indicating presence of essential oils. Essential oils have been shown to inhibit rumen fermentation in some tests (Nagy et al. 1964, see also 53 .33. v . ... ... v a .. ..u3.3...33.3 - ..33. .3“. .n.. 3N3.- .. ....- 3...- .mo.. ....33 ..3 ..33. a... 3... 3...- . 3.3.- 33..- .. 3...- ....33 .3.3 .....33.z..> ..3..33..z..> ..33.z..> z .3. .u .. ...3..3 .o.. 3.... ....33 ..3 3.3.3.3. ...3....3 ..u...33 ...3.3 . ..m..33.33 .3. .3..33 .33 333 ...33..33. .3. .33.33 33.3 33.3 3.3... .3...3333.33 3. 33.33. 333..o. .33333 .3 33.3333 33. .3 33.33333 .33.:333 33.3 3333.3.3.3 333. .3 .3.. 3.33.3...333 33.33.3..ou 333. 3.333.3333 .3-3 3.33. 54 references in Bryant and Kuropat 1980). E. suaveolens had high protein and low fiber levels. however, this species is renowned locally for its toxic properties. The bark was used to poison fish. None of the four understory foliage species were eaten freely by the blue duiker pair. and all four exhibited apparently effective anti-herbivore properties (Table 3—9). Alchornea floribunda had high protein and low fiber and tannin levels but leaves were never eaten. The foliage had a peculiar peppermint-like aroma, indicating possibly significant levels of secondary compounds. This species is a member of a group of Euphorbiaceae known for its toxic qualities (P.Waterman, pers. comm.). Although Scaphopetaium dewevrei foliage also had low fiber and tannin levels it too was not eaten. S. dewevrei foliage was strongly pubescent and harbored ants in small pockets in the leaf blade at the juncture with the petiole. The foliage of seedlings of both Brachystegja laurentii and G. dewevrei were high in fiber, and in the case of G. dewevrei, there were appreciable levels of condensed tannins as well. In addition, the first leaves put out by both of these seedlings characteristically become tough. fibrous and unpalatable even before the cotyledons are absorbed. .lgtake on_§ruit Diets For the blue duiker in all three trials. the total amount of fruit eaten on a diet was correlated with the rank preference of the preferred of the two food choices offered (Figure 3-1). Fruit intake on test combinations which included the top ranked food species was higher 55 Table 3-9. Possible defensive characteristics of species of understory foliage offered to a blue duiker (C. eonticola) in palatability trials.“ Species Acceptance. Evident Characteristic Alchornea floribunda saepled once strong odor Gilbertiodendron deuevrei sampled twice high fiber and tannins Scaphopetaiue copious pubescence, deuevrei sampled twice ants Brachystegia laurentii never eaten high fiber ‘ Three replicate offerings made- 56 Figure 3-1. Average (f SE) wet weight eaten (g/day) versus preference rank of each fruit species when the indicated species was the preferred choice offered. Number of tests shown below each species. For key to species abbreviations see Table 3-2. eaten preferred fruit VVIIEQIIII Wei :55 IX grams 57 TRIAL C 0|ue(E/K) 600[ .i‘i 6 I 5 200. De ‘ +\. . . ‘39 Ci'“ 1 2 3 4 5 Hints Ilue(E/K) Bey 800. 1200- /. 4 K9 r L 5 11 . Fm ‘00» '7‘, N 600% I ‘ Mu Kg 3 6 Fm6 Mu 1 2 . 1 2 3.5' 2 L 4, TRI A L A Blue (E/K) Blue (Ii/0) Bay 1000. 1000 1500 F F - Rh 4 8| a 600i 600-. 900. 8| 9 . C. 8| 5’ Op 8 Dose, 2 Dpa. Pp ‘ s 2 I 200. R35 200. N" 300. 2 1 2.5 4 1 2 3 4 1 2 3.5 58 than on diets which contained combinations of foods of lower rank preference. In other words, if blue duiker were presented with foods they preferred, they ate more of them. This was not the case with the bay duiker in either of the two trials it was tested. The preference score of a food and its intake were not correlated (Figure 3-1). Although the bay duiker exhibited consistent food preferences in both Trial A and B (see above), it was able nevertheless to consume large amounts of less—highly ranked foods when more favored choices were not presented. The bay duiker's ability to eat relatively large quantities of both preferred or less-preferred foods allowed this species to maintain relatively constant levels of edible dry matter intake over a range of diet preferences. Except for diets containing Brachystegja laurentii seed. in which dry matter intake was high (see discussion), total fruit dry matter intakes averaged between 350 to 410 g/day for all paired fruit combinations offered in Trials A and B. This was not the case for the blue duikers. Their fruit dry matter intakes declined as the animals were forced to choose between fruits of lower preference rank (Figure 3-2). When confronted with a choice between two less-preferred fruit species. blue duikers increased Ipomea batata foliage intake in their diet. In Trial A, foliage averaged only 5% to 6% of total dry matter intake when Brachystegja laurentii seed, the most preferred food was available. In contrast, when the two lowest ranked fruits were presented, foliage represented 45% — 52% of total dry matter intake (Table 3-10). Increased levels of I. batata foliage in the 59 Figure 3—2. Total fruit dry matter intake (mean 3 SE) on fruit diets of decreasing rank preference. Diet rank equals rank of the more-preferred of the two fruit choices offered. Symbols: e—~e. blue duikers E/K; O~O, blue duikers R/D: *_*. bay duiker M. 60 33.039 coo. O .e..h ue..e.e.n m .m_.h .0 scat F Q In < .m..h 30in euieid 1.301. llnll ueiee Jenem Mp 61 .33.3333 3.3.. 33..3.3.3 .3 3333 a . n n 3 a u n 3 . n n 3 3 3.. ..o 3.3 3.3 3.3. 3.3 3.3 ... -- 3.. ..n 3.. 33 3.3. 3.3 3... 3.. 3.33 ...n 3.33 3.3 3.33 3.3. 3.3. 3.3 m n.» 3.3 a . 3 n a . 3 3.. 3.3 . .3333 .3. .33.33 .33 .3.3. .33.33 33.3 .x.m. .3x.33 33.3 .333..33. 3333.3.3.3 3. 33..3> 33.3..3 33.3.. 33. 33.33 3. .u..3n.33 .3. .33.33 .33 333 .3.33..333 .3. .33.33 33.3 .3 3.3.3 3. 333..3. 3.3.33 33333. .3 33.33..3333 .33333 ..3 ..3u3.3 .333. .3 333.333.33 .o.-n 3.33. 62 diet did not offset the loss in total dry latter intake precipitated by reductions in fruit consumption on the less-preferred diets. Percentage foliage in diet and total dry matter intake were negativley correlated (Figure 3-3) (E/K, R2 = .45. p < .05, n = 20: R/D. R2 = .54. p < .001, n = 19). In the bay duiker diets the contribution of foliage to total dry latter intake was never lore than 15%. even when the least preferred fruit combinations were offered (Table 3-10). The bay duiker thus maintained fruit intake over a range of fruit quality. DISCUSSION Referring to the specific questions posed in the introduction, food choices of both the blue and bay duiker in palatability trials were not random. Preferences for specific foods were generally consistent between sinilar replicated presentations. Although both duiker species revealed clear food preferences. neither species ate even the lost favored food to the exclusion of other offerings. This result is consistent with what has been seen of herbivore food choices in general (Bunnell and Gillingham in press) and is thought to be due to nutritional inadequacies and defensive conpounds in some plants (Freeland and Janzen 1974. Westoby 1978). The results of this study are in agreement with recent findings on diets of tropical forest primates which demonstrate that nutritional correlates of food preference are not well defined by any single chemical variable (Oates et a]. 1977, 1980; Milton 1979. McKey 63 Figure 3—3. Total dry latter intake (g) versus percent [pa-ea batata foliage in blue duiker diets for all two-species fruit offerings in Trial A. 64 3332o. 33.3.3... 303.330.33. co om 0v on ON or on 0.3 on ON or 3 T 3 1 3 3 O 3 3 3 3 3 o O O J .oou 0 .oc« A O O O . m / .5 O .6 . O . O .333 / 3.333 .. 0 Ox 0 O O I a. ’ .3 3 .333 o .333 ) o 6 .33: 33...: 3 .3... 33.”... ( a). 3.3.. 33.23... 3:..0 v.\m 3.3.. 333...... 33.0 0 0 Lean Lcan 65 et al. 1981). Composite measures incorporating apparent nutritional costs and benefits of a food were better correlated with food preferences in both species. In this study ratios of N/ADF and N/(ADF + CT), each multiplied by dry matter yield (Y) provided positive correlations with food preference. McKey et al. (1981) found that ratios of nitrogen to fiber and nitrogen to fiber plus tannins were correlated with diet choice in the black colobus monkey (Colobus satanus) in Cameroon. These authors did not weight their ratios with the dry matter yield of the foods as was done in this study. Most of the important foods eaten by the colobus were either foliage or edible seeds. The proportion of these foods which was edible dry weight apparently may not have varied markedly from species to species. This was not the case for many of fruits reported here in which the relative proportions of water and inedible (reguritated) seed varied widely. As Herrera (1981) has pointed out. the relative portion of a fruit which is edible dry weight pulp can have a marked effect on frugivore preference. Not all duiker food preferences were correlated with adjusted dry matter yield. Ripe MMsanga cecropioides fruit was most preferred by blue duikers in Trial C yet it had low edible dry matter content and relatively low nitrogen to fiber ratios (Table 3—5). Preference for this fruit may have been due to high levels of readily fermented carbohydrates (sugars and pectins) not measured in the extractions. Use of this fruit by bats and cercopithecine monkeys, neither of which have developed capacities for fermentive digestion. also is evidence that it must contain readily available stores of 66 energy. The apparently high levels of fiber recorded in the chemical analyses may have been due to the mixing of both fruit and seed in the samples used for extraction. Since the seeds of this species are inedible and are voided by the duikers, their inclusion in the chemical anlayses may give a distorted measure of the true nutritonal value of the food. The blue duikers' preference for ripe Canarium. schweinfurthii fruits may also be a function of concentrated, readily available energy. These fruits are apparently rich in digestible lipids (Wu Leung 1968). It was not clear why they were avoided by the bay duiker. A number of the foods preferred by the duikers contained appreciable levels of tannins. The importance of these compounds in affecting selectivity is not clear. canarium schweinfurthii fruits had high condensed tannin levels. Its fruits were favored by the blue duiker. but not by the bay duiker. Ricinadendron heudelotii fruits had high levels of hydrolysable tannins (P. Waterman. pers comm.) and yet were eaten freely by bay duiker in the Trial A. Captive blue and bay duikers maintained on a diet dominated by R. heudelotji, however. exhibited what was apparently marked hematuria, (Chapter 2). Differences in the diets of frugivorous ungulates and primates in African tropical forests may be linked to to differences in the relative tolerances of these two taxa to tannins and perhaps other plant defensive compounds. Two studies (Wrangham and Waterman 1981. 1983) have shown that at least some species of primates which lack foregut fermentation may be sensitive to condensed tannin levels in 67 foods and will avoid foods such as some unripe fruits where levels of these compounds are high. Other studies, however, indicate that the role of tannins as feeding deterents may not be easily generalized (Zucker 1983. Waterman 1983, Hole and Waterman in press). Nevertheless, in the Ituri Forest. up to 50% or more of the foods available to ungulates on the forest floor were unripe fruits or parts of ripe fruits which showed evidence of being discarded by primates (Chapter 4). Many of these fruits were included in duiker diets and were apparently an exclusive resource for them. It is tempting to ascribe the ruminants ability to utilize these fruits to the detoxifiying and digestive capacity of foregut fermentation. Further investigation into this capacity may enlighten our understanding of trophic relationships between co-occurring frugivorous taxa. It may also help to explain the relative success and evolutionary radiation of the African forest ungulates. Foliage as a_Food Source It has been suggested (Hladik 1978. Milton 1979) that foliage may be important in the diets of frugivorous primates as a source of protein which is lacking in fruits. While this may be true for many of the fruits consumed by primates, there is evidence that this is not the case for some of the important foods in the diets of the frugivorous ungulates. In a survey of fruits, flowers and seeds available to duikers on the forest floor (Chapter 4), 19 ripe fruits, 21 unripe fruits and 5 flowers averaged 16.0 mg/g, 17.7 mg/g and 26.9 mg/g nitrogen 68 respectively. Four species of frequently eaten seeds averaged 59.8 mg/g nitrogen. essentially equivalent to the average 60.0 mg/g contained in the 10 canopy foliage species tested in the trials. At least some seeds thus may be equivalent to foliage as a source of nitrogen. Seeds have an additional advantage over foliage in that they are often concentrated sources of readily digested energy (McKey et al. 1981. Waterman 1984). The relative availability of nitrogen must also be taken into account in evaluating a food as a source of dietary protein. Many forest foliage species and some forest fruits available to ungulates are high in fiber. This would tend to reduce their digestibility and thus their value as a source of protein. Although many of the fruits eaten by duikers were lower in total nitrogen content than foliage. they were also relativley lower in fiber and thus probably more digestible. From the perspecitive of the animal, these may have been superior sources of dietary protein. Blue and bay duiker differed in their patterns of foliage intake on the test diets. The bay duiker maintained a high fruit intake over a wide range of fruit preferences. Ipomea comprised more than 15% of the bay duiker's diet even when it had only less-preferred fruits as alternative foods. In the blue duiker in contrast, I. batata were presented fruits of apparent low nutritional quality. a pattern which was also documented in digestion trials (Chapter 2). Diets dominated by foliage were correlated with low total dry matter intake. This may have been due to the low edible dry matter 69 content of I. batata foliage relative to that of many fruits and seeds (Table 3-5). Since a lowered total intake could threaten an animals energy balance, especially in a species of small body size (Milton 1979, Hokey et al. 1981). the availability of suitable fruits or seeds may be essential to the occupation of the forest understory by the duikers. It is interesting that the only small ungulate folivore in this forest, Neotragus batesi, avoids the forest understory and is partial to clearings and treefalls where presumably higher light levels allow adequate supplies of high quality foliage to grow (Freer 1979, J. Hart, unpubl. obs.). Intake, Diet Quality and Body Size in Duikers Highest total dry matter intake was achieved by both duiker species on diets containing Brachystegia laurentii seeds. Relative to all other species of fruits and all foliage tested, B. laurentii seeds had both the highest edible dry matter density and protein to fiber plus tannin ratios. The response of the two duiker species to lower quality diets varied. Like sheep on progressivly diluted concentrate diets (Baile and Forbes 1974), the bay duiker increased its total intake as the diet quiality was reduced. By varying total intake, it was able to maintain total dry matter intake at a constant level over all food combinations. Blue duiker did not show this same response to decreasing diet quality. This species was unable to increase intake on low quality diets. 70 Demment and Van Soest (1985) have proposed the hypothesis that different dietary fiber levels of plant food resources create a gradient of food quality which can be partitioned by animals of different body size as a function of their relative digestive capacity. Fiber levels alone were not the only correlate of food selectivity by duikers in captive trials. Dietary quality for duikers is likely to include relative levels of nutrients and other digestion inhibitors, as well as the edible dry matter content (see also Chapter 2). Nevertheless, the idea of size-partitioned resources based on food quality may be pertinent to an understanding of relations within the frugivorous ungulate guild. While both blue and bay duiker preferred foods with high adjusted dry matter yield values, bay duiker had a greater capacity to broaden the diet to include lower quality items. Broadening of the diet would be expected to occur when food is scarce and the larger animals relatively large total needs would not likely be met by selective foraging for small, dispersed patches (see also Chapter 2 and 4). Based on the results of these trials, the smaller blue duiker is evidently more constrained to select high quality foods than the bay duiker. Its total food needs, however, are lower, and thus this species could be expected to be better able to meet these by being selective than would the bay duiker over periods of low food availability. 71 CONCLUSIONS Two pairs of blue duiker and a bay duiker exhibited consistent preferences for fruits. Overall preferences for fruits in both species was better correlated with a measure of the edible dry matter content of the food multiplied by a ratio of protein to fiber or protein to fiber plus tannin content. Preferences for different foliage species were less apparent. Only small amounts of many foliage species were eaten. Overall, foliage is less preferred than many fruits and edible seeds. Dry matter intake was highest 0n diets containing Brachystegia laurentii seeds. Dry matter intake in the bay duiker remained constant over diets including both high ranked and low-ranked fruits. In blue duiker in contrast, dietary dry matter intake declined in all three trials as the quality of the fruits presented was decreased. Blue duikers increased Ipomea foliage intake on low quality diets, but this did not compensate for a decline in fruit intake. LITERATURE C ITED 72 LITERATURE CITED Baile. C. A. and J. H. Forbes. 1974. Control of feed intake and regulation of energy balance in ruminants. Physiological Review 54: 160-214. Barnes, R. F. and G. C. Marten. 1979. Recent developments in predicting forage quality. Journal of Animal Science 48: 1554—1561. Bryant, J .P and P. J. Kuropat. 1980. Selection of winter forage by subarctic browsing vertebrates: the role of plant chemistry. Annual Review of Ecology and Systematics 11: 261-285. Bunnell, F. L., and N. P. Gillingham. in press. Foraging behavior: dynamics of dining out. In: Hudson, R., and White, R. (editors). Bioenergetics of wild herbivores. CRC Press, Roca Baton, Florida, USA. Crawley, H.J. 1983. Herbivory: the dynamics of animal-plant interactions. Studies in Ecology (10). University of California Press. Berkeley, USA. Demment, M. W. and P. J. Van Soest. 1985. A nutritional explanation for body-size patterns of ruminant and nonrumianat herbivores. American Naturalist 125: 641-672. Dubost, G. 1984. Comparison of the diets of frugivorous forest ruminants of Gabon. Journal of Mammalogy 65: 298-316. Emmons. L.H., A. Gautier-Hion, and G. Dubost. 1983. Community structure of the frugivorous-folivorous forest mammals of Gabon. Journal of Zoology (London) 199: 209-222. Freeland, W. J. and D. H. Janzen. 1974. Strategies in herbivory by mammals: the role of plant secondary compounds. American Naturalist 108: 269-289. Freer, F. 1979. Observations ecologiques sur le neotrague de Bates (Neotragus batesi de Winton 1903) du nord—est du Gabon. Terre et Vie 33: 159-239. Gautier-Hion, A., L.H. Emmons. and G. Dubost. 1980. A comparison of the diets of three major groups of primary consumers of Gabon (primates, squirrels and ruminants). Oecologia (Berlin) 45: 182-189. Herrera, C. 1981. Are tropical fruits more rewarding to dispersers than temperate ones? American Naturalist 118: 896-907. Herrera, C. 1982. Defense of ripe fruits from pests: its significance in relation to plant-disperser interactions. American 73 Naturalist 120: 218-241. Hladik, C. 1978. Adaptive strategies of primates in relation to leaf eating. Pages 373-396, in G. Montgomery, editor. The ecology of arboreal folivores. Smithsonian Institution Press, Washington, D.C.. USA. Janzen, D.H. 1971. Seed predation by animals. Annual Review of Ecology and Systematics 2: 465-492. Janzen, D.H. 1983. Physiological ecology of fruits and their seeds. Pages 625-655 in O.L. Lange, P.S. Nobel, C.B. Osmond, H. Ziegler, editors. Physiological plant ecology III, Encyclopedia of Plant Physiology, New Series 120. Springer-Verlag, Berlin. Mattson, W. J. 1980. Herbivory in relation to plant nitrogen content. Annual Review of Ecology and Systematics 11: 119-161. McKey. D. 1979. The distribution of secondary compounds within plants. Pages 55-133 in G.A. Rosenthal and D.H. Janzen, editors. Herbivores, their interactions with secondary plant metabolites. Academic Press, New York, USA. McKey. D. B., T. J. Gartlan, P. G. Waterman, G. Choo. 1981. Food selection by black colobus monkeys (colobus satanus) in relation to plant chemistry. Biological Journal of the Linnean Society 16: 115-140. Milton, K. 1979. Factors influencing leaf choices of howler monkeys: a test of some hypotheses of food selection by generalist herbivores. American Naturalist 114: 362-378. Mole. S. and P. G. Waterman. in press. Stimulatory effects of tannins and cholic acid on tryptic hydrolysis of proteins: ecological implications. Journal of Chemical Ecology. Nagy. J. G.. H. W. Steinhoff and G. M. Ward. 1964. Effects of essential oils of sage brush on deer rumen microbial function. Journal of Wildlife Management 28: 785-790. Oates, J. F., T. Swain and J. Zantovska. 1977. Secondary compounds and food selection by colobus monkeys. Biochemical Systematics and Ecology 5: 317-321. Oates, J. F. P. G. Waterman and G. M. Choo. 1980. Food selection by the south Indian leaf-monkey Presbytis johnii in relation to leaf chemistry. Oecologia 45: 45-56. Rosenthal, G.A. and D.H. Janzen. 1979. Herbivores: their interaction with secondary plant metabolites. Academic Press, New York, USA. 74 Snow, D.W. 1971. Evolutionary aspects of fruit eating by birds. Ibis 113: 194-202. Steel, R.G.D. and J.H. Torrie. 1982. Principles and procedures of statistics. Second edition. McGraw-Hill. New York, USA. van der Pijl. 1972. Principles of dispersal in higher plants. Second edition. Springer-Verlag. Berlin. Van Soest, P. 1982. Nutritional ecology of the ruminant. O and B Books, Corvallis. Oregon, USA. Waterman, P. G. 1983. Distribution of secondary metabolites in rainforest plants: towards an understanding of cause and effect. In S.L. Sutton, T. C. Whitmore and A. C. Chadwick (editors), Tropical rainforest: ecology and management. Blackwell Scientific Press. Oxford, UK. Waterman, P. G. 1984. Food acquisition and processing as a function of plant chemistry. pp. 171-211. In: Chivers, D. J., B. A. Woodind and A. Bilsborough (eds.), Food acquisition and processing in primates, Plenum Press, London. Waterman, P. G., C. Mbi, D. McKey and J. S. Gartlan. 1980. African rainforest vegetation and rumen microbes: phenolic compounds and nutrients as correlates of digestibility. Oecologia 47: 22-33. Westoby, M. 1978. What are the biological bases of varied diets? American Naturalist 112: 627-631. Wrangham, R.W. and P.G. Waterman. 1981. Feeding behavior of vervet monkeys on Acacia tortilis and Acacia xanthoploea with special reference to reproductive strategies and tannin production. Journal of Animal Ecology 50: 715-731. Wrangham. R.W., and P.G.Waterman. 1983. Condensed tannins in fruits eaten by chimpanzees. Biotropica 15: 217-222. Wu Leung, W.T. 1968. Food composition tables for use in Africa. U.S. Department of Health, Education and Welfare and U.N. Food and Agriculture Organization. Zucker, W. V. 1983. Tannins: does structure determine function? An ecological perspective. American Naturalis 121: 335-365. APPEND I X 755 AHENDD<3> COSe --- D.p. R.h. P.p. Bey Duiker (Replicate n) S SDECIIS A: pecies B: Bole R.h. --- >> C.s. (Replicate E/K) D.p. R.h. )) >> » >> --- <3 D.p. R.h >> >) >< >> "' ) D.p. P.p. )) >> > > --- ). Preferred P.p. choxcee )) 4.0 )> 2.5 >) 2.0 >) 1.5 --- 0 Preferred P.p. choxce ‘ >> 4.0 >> 2.5 >) 2.5 >> 1.0 “‘ 0 Preferred C.s. choice‘ )) 4.0 )) 3.0 ) 1.5 > 1.5 --- 0 Preference rank 1 2 Preference rank 1 2.5 2.5 Preference rank 1 2 3.5 3.5 '76 Appendix 3-A continued UHRIPUL E3 Blue Duiker (Replicate E/K) Species B: l.w. Speczes A: Bay Duiker 13H. --- (Replicate H) Species 8: Species A: K.q. K.q. --' l.w. H.u. F.m. K.g. H.u. F.m. K.t. )> >> ) )> --- >> >< >> --- a >> --- > l.w. H.u. F.m. K.t. >> > > > -- )) > ) —-- > > --- > Preferred choice“ 4.0 2.5 1.5 1.5 0 Preferred choice‘ 4.0 Preference rank 1 2 3.5 3.5 Preference rank 1 2 77 Appatfix,}4lcnnthmed TRIAL C Blue Duiker (Replicate E/K) Species 8: Preferred Preference H.r. 1.9. K.g. D.e. C.e. n.u. chalce' rank Species A: H.r. --- >> > >= 2 >) 4.0 l 1.9. --- > >< )2 > 3.5 2 Klg. --- = )> > 2.5 4 0.2. --- >> ) 30° 3 Ce.e --- > 135 5 H.u. O-- o 6 lkkes: a For species abbreviations, see Table 3-2. b Symbols: ) , A eaten more than B; < , A eaten less than B; =, amount of A and B eaten differ by less than 10%. Each symbol records results of a single paired test. 0 Number of paired combinations in which Species A is the preferred choice. Paired tests in which A = B, or replicate tests in which A) B and Auauu note coa~=u_s~xu .9 cos can: acum-s sung udauu oz .: >n cans. oufiu -_asu- mayo» gum: aux-u goon cots uo_ooa a“ coma-«>uu stance». on» .uoxau can: so coaco>~ um anew-s >uon n .Anmoac .d. a. acosem noes accouu—n >uw>fiuua use unsung: . Haccauuoc a "acts“: ounces ucauqa o.mo .u guannsodgu> cognauws~xu nacao~qca.u flaccauuo: ounces tends: "a o." o.~m c-xasu >~n uu~nucon nacao~oca~o daccaau unoco+ vegan: cu m.~ 5.5“ taxman u.c-u-m nauxa--u nacao-caou duccaau usages uc-a= o“ o.“ n.o~ taxman conga a...-u°u=.~ nacao-ca.u mustang ”canoe gens. m , n.“ o.n« .u noncoc+uxuodn acocencuuc nacao-cauo daccauuoc nus-cu. A“ a." «.mfl cauuoc>ozu managaau- nacuuo-.x: flaccaau uuota+ uc-a= on 0.0 n.¢ costu .zdn u~ou-con “sauna-ca-o scuba-a >ua>.uu¢ u-u.n.: c .a.m can: L-fiauacco> ammo-am . .oxc seas—3 Lassa ..-c«a~ .u-ocod menu” as“ +o acaas ou._=oc: usoto>aoacw asp ."uc o_n-h 87 The community includes both distinctly nocturnal and diurnal species. One species, C. sylvicultor, is reportedly active both day and night (Emmons et al. 1983). Details on the biology of most species remain little known (Kingdom 1982). Accounts of the behavior and ecology of C; monticoia, C. dbrsaiis and H. aquaticus are available in Dubost (1975. 1978. 1980 and 1983). The major observed impacts on local duiker and chevrotain populations were hunting by nomadic Mbuti hunter-gatherers and conversion of forest to farmed bush and secondary forest by shifting cultivators. The latter was limited to a small area near the settlement. Although hunters ranged widely. remote forest areas were only infrequently hunted. Based on hunting returns and pellet counts, duiker populations may be higher in these areas than in continuously hunted areas near town (Roster and Hart in prep.). All data for this study were collected from more remote locations. METHODS Data on food availability, animal distributions and samples of animal gut contents for dietary analyses were collected by accompanying local hunters into the forest on drive hunts using dogs and nets. working with the hunters allowed access to large areas of forest and samples of animals. The net hunt technique involved 10 to 30 hunters encircling areas of forest from 4 to 12 ha (mean 8 ha) with nets about 1.2 m in height and driving any animals within the circle into the nets. Nets were then coiled and quickly moved to a new site. usually 88 less than 1 km distant. Five to seven drives were generally conducted in the same vicinity of forest in the course of one day. Field data and gut contents were collected at nine sampling sites throughout the study area (Figure 4-1). These sites were visited during 13 sampling periods between June. 1981 and May, 1983. Each sampling period encompassed 3 to 9 daily hunts and 17 to 44 drives (Appendix 4—A). _Body weight and Cranial Morphology Animals caught on net drives were weighed entire to the nearest 0.1 kg. Age (based on mandibular molar erruption) and reproductive condition were assessed. Fetal weights were subtracted from the body weight of gravid females. Skulls of one to four individuals of six of seven species of frugivorous ungulates were collected from obliging hunters. Length of cranium was measured ventrally along the surface of the skull from the anterior edge of the occipital condyles to the tips of the premaxillary bones. Midth of cranium was measured transversely on the ventral surface of the skull across the widest point of the zygomatic arches. Length of the mouth was measured longitudinally along the palatine and maxillary sutures from a transverse line connecting the posterior edge of the thrid molar. anteriorly to the tips of the premaxillary bones. Mouth width was measured transversely across the outer cusps of the tooth row at its widest point, usually the second molar. 89 Food Availability The fallen fruits, seeds and flowers eaten by duikers and the chevrotain generally occurred in discrete patches on the forest floor, generally beneath parent trees. These food patches had well—defined boundaries and were readily identifiable against a background in which the item was absent. The size, abundance and species composition of food patches were measured along transects 3 m wide which followed the perimeter of the net drive. Each time an item was encountered on the transect, an attempt was made to record the following basic data: 1. Items were identified to species. Unknowns were given a number and specimens were dried. I 2. Fruits and seed were classified as ripe or unripe and the relative proportions of fruit flesh and seed were estimated. Note was made whether the patch occurred under a parent tree's crown, whether the patch contained items which had been bitten and dropped by arboreal vertebrates (mainly squirrels and primates) and whether aborted unripe fruits had insect or other damage. 3. Fruits, seeds and flowers were classified by the lengths of their longest axis into five size classes (Chapter 2): Class 1. 0.0 - 0.5 cm; Class 2, 0.5 - 1.0 cm; Class 3. 1.0 - 2.5 cm; Class 4, 2.5 - 5.0 cm: and Class 5, 5.0 - 10.0 cm. When a patch contained items of more than one size class, the proportion in each class was estimated visually. 4. Individual specimens of 40 species of commonly-encountered fruits, seeds and flowers were weighed to nearest gram (wet weight). 90 Weights of other species were estimated by visual comparison to known species of similar size. 5. The number of items on the ground along the transect was counted (or estimated when especially numerous) and expressed by square meter density. 8. Total patch area both on and off the transect was estimated by pacing the diameter or length of the patch and converting to areal measure by classifying patch shape as circular or rectangular. This procedure was modified in specific instances. Where only an individual food item was found, it was classified as an isolate and arbitrarily assigned a patch area of 1 m2. Discrete patches could not be distinguished for flowers and released seeds of the Caesalpiniaceous dominants. B. iaurentii and G. dewevrei during mast years. The occurrence and densities of these items were recorded at 30 pace (44 m) intervals along the transect. Each time seeds or flowers were encountered, the occurrence was equated with a single tree crown and the patch area was arbitraily assigned a value of 150 m2 , corresponding with crown dimensions of these species (T. Hart, 1985). Partially-eaten fruits dropped by primates were often scattered at some distance from the parent tree. Identifications of patch boundaries were arbitrary in these cases. Discrete groups were counted as different patches, even though they may have been produced by the same tree or liane. For each food patch, the weight (W) in grams was calculated 91 as: W = U x D x A where U = item unit weight (g) D . average density (items/m2) on transect A = patch area (m2). Food patches were grouped into 4 size classes based on estimates of W: very small patches W 5,10 g; small patches, W = 10 - 100 g; large patches, W = 101-1000 g; very large patches, W > 1000 g. Weights of patches recorded on individual transects were summed for each sampling period. Fatch species diversity for a sampling period was computed by the Shannon Index (H') (Drawer and Zar 1977). Note was kept of leafing phenology of the more common tree species (J. and T. Hart, unpublished data). The abundance and weights of patches of terrestrial fungi were noted qualitatively. On net hunts, ruminoreticular contents (hereafter referred to as rumen contents) were removed from a sample of animals at the site of capture and stored in plastic bags. Rumen contents for animals dead less than four hours were washed through a 5 mm mesh screen and the larger food fragments retained in 10 s formalin for analysis. Duikers tend to swallow fruits and seeds whole and then to regurgitate and 92 masticate (ruminate) them at a later time. Large fragments separated in the rumen contents evidently comprised the duikers' last meal. This fraction could be easily identified and its analysis was probably less biased by differential digestion of readily-fermented portions than smaller fractions (Dirschl 1982. Puglisi et al. 1978). Rumen contents from unweaned animals and contents in which the 5 mm fraction weighed less than 10 g and comprised less than 10% of the total contents were not included in the analyses. Rumen contents were sorted by food type. Fruits, seeds and flowers were identified to species. Unknown items were assigned a number. Examples were dried and mounted on cards for later comparisons. For each species found in the gut, the total weight and item size class were determined. Patch weights of food species in the rumen were estimated when possible based on values obtained for the species from food transects or from visits made to fruit-bearing trees. It was found that within a given sampling period many species exhibited a characteristic patch weight, at least wihin the broad limits of the classes used in this study. The contribution of each food species or food class to the total diet was evaluated by a percentage utility index (1“) defined for each food 1 as: 10(1) - ‘91 x qi) x 100x where p1 equals the proportion of total rumens in the sample containing food item 1 and q1 equals the average proportional contribution of i to the total weight of contents in rumens in which it was recorded. Values for Iu ranged from zero in 93 cases in which no rumen contained a given item to 100.0% where a single species comprised the total contents of all rumens. Owaga (1978) has shown that the proportion of foliage measured in rumen contents may differ according to the screen size used to wash the sample. In order to test for this effect and to examine the frequency of smaller items in the gut such as insect parts, a fixed measured subsample of the total rumen contents was washed through a 2 mm mesh screen and the proportions of fragments by food type (fruit/seed, fungi. foliage, insect) were determined. Diet overlap was measured by by Morisita's (1M) index (Morisita 1959) and considered in light of discussions of Horn (1966) and Hurlburt (1978). Nutritional Quality Nutritional composition was determined for samples of 19 ripe fruits. 21 unripe fruits. four seeds and five flowers. These items included commonly-available potential foods collected from transects between October. 1982 and May. 1983. For each food item. the nutritionally significant portion (Herrera 1981. see also Chapter 3). termed edible dry matter yield (Y), was determined as: Y = (T - S) x DH where T is the total wet weight of the item (g), S is the weight of indigestible, regurgitated seeds (g) (if they occurred) and DM equals the dry weight of the assimilable portion determined by dessicating the sample at 1002 C for 24 hours. Assays for condensed tannins (CT). 94 acid detergent fiber (ADF) and crude protein (N) content were conducted for each sample by Dr. P.G. Waterman. following standard procedures described in Horowitz (1970), Goering and Van Soest (1970) and Oates et al. (1980). Levels of each component were expressed as mg/g fresh weight of the food. An estimate of food nutritional quality (YQ), targed the adjusted dry matter yield, was calculated as: Y = Y x [N / (ADF + CT)]. 0 Yb is essentially an estimate of the edible dry matter portion of a food weighted by its relative levels of nutrient and refractory or digestion-inhibiting compounds. YQ was found to be positively correlated with food preference for two species of duiker in feeding experiments (Chapter 3). The average nutritional value of all available food (YQ(T)) during a sample period was determined as: You) =2 Your) x pm. where Y°(i) equals adjusted dry matter yield for food species 1 and p(i) equals the proportion of total food biomass contributed by i as recorded on transects for the period. YQ(T) values were calculated for five sample periods between July. 1982 and May. 1983. These values accounted for 813 to 983 of total foods recorded on transects for each period. 95 Diet Selectivity Various indices have been proposed to measure an animal's differential utilization of foods in relation to their availability (Ivlev 1961. Manley et al. 1972, Petrides 1975, Cock 1978, Chesson 1983). Their use is limited. however. in tropical forests where high species diversity and scattered foods inhibit the accurate measurement of availability. Johnson (1980) has proposed a rank index of the selective use of a resource which is calculated by subtracting a rank measure of the use of the resource from a rank measure of its availability. Values greater than zero indicate positive selection. those differences less than zero indicate avoidance. Resources for which rank availability and rank use are equal are utilized in proportion to their abundance and. by definition. are not used selectively. The Johnson index is relatively robust to the problem of irregular availability. It has the additional advantage that various measures of availability and utilization can be used. so long as they can be ranked. The Johnson index was used in this study. The availability of a given food was calculated as its percentage contribution to the total weight of patches recorded on transects. Utilization of a given food was measured by it percentage occurrence (utility index value) from gut contents. Four ranked classes of use and availability were defined as: Rank 1 = < 1.0 s Rank 2 = 1.0 3 to 5.0 X Rank 3 = 5.1 X to 25.0 X 96 Rank 4 = > 25.0 3 Strong positive selection was defined as values of rank use minus rank availability of +3 and +2. Strong avoidance was defined similarly by values of -3 and -2. Weak selection was exhibited by values of +1 and -1. No selection was exhibited when rank use minus rank availability equaled zero. Animal Abundance and Distribution Total area of each drive hunt was estimated by assuming circular placement of the nets and pacing their perimeter to determine drive circumference then calculating drive area. Forest types of net drive areas were quantified by classifying forest type as mixed (including old secondary). mbau or swamp/riverine at 30 pace (44 m) intervals around the perimeter of the drive area. The identity of each animal captured on drives was recorded. The identities of animals flushed but not captured were ascertained by questioning hunters. Two independent confirmations of each sighting were sought whenever possible. Indices of animal density were calculated as the number of animals flushed per square kilometer of drive area summed over all drives at a sampling site or within a forest type. Reliability of net hunt results was verified in two ways. Indices of animal abundance from hunts were compared with densities derived from pellet group and track counts conducted in the same areas (Roster and Hart in prep.). Relative frequency of specific microhabitats sampled while accompanying net hunts was compared with similar counts made from randomly placed transects in the same area (J. 97 Hart and T. Hart,unpubl. obs.). These studies indicated that net hunt results were probably not strongly biased with regard to animal densities and specific habitat features. It was noted. however, that hunters tended to prefer mixed forest to mbau and to avoid large swampy areas. These preferences did not create a sampling bias for ths study, however, since large swampy areas were uncommon. and the results of drives from mixed and mbau forest were analyzed separately. Statistical Tests Throughout this study standard parametric statistical techniques (Steel and Torrie 1982) were employed for all analyses where the required assumptions were met. Ranked data were analyzed by techniques described in Siegel (1958). Specific tests are identified with each analysis. RESULTS Cranial Morphology. Skulls of duikers are similar in general form. Crania of C. nigrifrans and on callipygus can be differentiated only by the form of the horn sheath. The crania of C. dbrsaiis and C. Ieucagaster'are distinctive in being more rounded, with foreshortened muzzles. The skull of C. dbrsalis is especially robust with relatively large attachments for masticatory muscles, wider transverse of the mandibular condyle, and correspondingly heavier molars (Kingdom. 98 1982: 315-318). The crania of C. manticala. C. nigrifrans. C. cailipygus and the chevrotain in contrast. are more finely structured with elongated muzzles and less massive teeth. The wide crania of C. ieucogaster'and C: dbrsaiis'are associated with a broadened gapes and higher ratios of mouth width to length than in the other species (Figure 4-2). Food Availability The diversity and abundance of fallen fruits. seeds and flowers recorded along the transects in mixed forest varied seasonally over the two year study period. Abundance was greatest during the mid to late wet seasons. Peaks of diversity occurred during the mid wet seasons. During dry seasons and early wet seasons by comparison total fruit abundance and diversity were reduced (Figure 4-3). Many, but not all trees in the Ituri Forest followed the same general seasonal cycle. Flowering and fruit set occurred in the early wet season followed by fruit ripening and seed fall through the late wet season (J. and T. Hart. unpubl.~ obs.). This cycle is comparable to that reported in other tropical forests (Frankie et al., 1974, Terborgh 1983) including those of Zaire (Diterlein 1978). On average. fruit comprised most of the food biomass and the majority of the patches recorded on the transects (Table 4-2). Ripe fruits dominated the transect finds. averaging 55% of the total weight and 32% of all patches. Unripe fruits averaged 21* of total weight and 38% of all patches. Fallen flowers and seeds comprised smaller percentages of the standing crop recorded. 99 Figure 4-2. Cranium size and mouth shape (width/length) in duikers and chevrotain. Values shown are mean and standard error. Number of skulls measured indicated below each species. Species abbreviations: C m. Cephalophus monticola; H a. Hyemoschus aquaticus; C n. C: nigrifrons; C l. C. leucogaster; C c C. callipygus: C d, C. dorsaiis. Mouth 100 .56 F d— : .52 )- "P 0 Cd 4 C 0 Cl .. 48 _ 3 \ g -— 44 l- + H. 3 Cm 2 Cc On .40 l 1 L i 11 100 120 140 160 180 Length of cranium (mm) 101 Figure 4-3. Diversity (Shannon index and number of species/km of transect) and collective abundance (kg/km of transect) of fruits. seeds and flowers on the forest floor. Ituri Forest. Zaire. July 1981 to May, 1983. 102 1 O J \. J '\ e\ /— 4 O \. ‘ \. ‘0 J7 a g7 a m 4' n N Q' o lseeueaa my [Ox 1 O .’ /\. O/ ‘ O\. , \. ./ : .\. Ill \. 6‘ h o a m n a p o aseeueaa mm/eeuseds .4 \. ~-%MW§%MW?~ . 0 If \ . 1 D Q N O a) v- v- '- F O (.H) KIIIJOAIp Co'agds 1983 1982 S‘ea aor: 1981 1983 1982 Season 1981 E 1983 L D I I 1982 Season L 1981 103 .o.n. .m.¢s. .A.OL. .a.s. .s.o. .o.~d. .o.m. A..o. o.o n.n~ ..on n.1n n.o ¢.oz o.o~ o..n moats>¢ .o.~«. .m.d~. .o.os. no.0. .o.n. A~.o~. .s..u. mo.muc o.o n.n~ \.sw _.se n._ o.¢~ m.- n.~o s Ls: .Las .o.o. .o... .o.m. .a.~. .h.ss. .o.o. ...~s. .s.~n. m.o n.s o.s¢ ~.~. o.._ ".0" ~.ou m.s. u 3.3-uaz .o.~s. in.-. .m.o. .s.n. ...na. .s.mc Am.s~. .n.o~. so: ~.n~ ~.s~ m.m¢ m.ms m.~s o.~ n.o~ o.am . >_t.w .o.o. ...on. .o.m. .a.on. 10.0. .s.o.. .~.n. .o.on. 0.0 o.¢. n.s~ «.4» 0.0 m.nn n.~s a..n a sea Lasts slats 3.3ca Lassa .uoatoa cosonu noun agate: ug.x coxo~u scum undue: on.m nun-cm cons-m nozuuon so coo—acouc-m anus-z sun-cue» so poo-ucoucoa .Aucomu-.>-u nests-u. new. acne- ... -=_-> .nood .xa: a» «mad .>~=a .utaa~ .uu-cou “can" .uaocos ucxas cm oboe-cat“ acaoco co sausage ago at... .uu.:cs ca-as +0 nun-ccauuo .m-¢ snack 104 These average trends concealed marked seasonal variability in the contribution of different food types to total food availability. High standard deviations associated with the percentage contributions of most food types, even within a given season (Table 4-2). reflect marked differences from one year to another. Much of this variation can be accounted for by the periodic occurrences of a few species. When these became available they comprised a large percentage of the the transect weight and in some cases the majority of food patches. Notable among these were the flowers and seeds of the synchronously fruiting Caesalpiniaceous dominants, Brachystegia laurentii in mixed forest and Gilbertiodendron dewevrei in mbau forest (see discussion below). Other common species which fruited gregariously on an irregular basis were several species of Landblphia‘which were available during the late wet season. 1982 and Pancovia harmsiana which fruited in 1982 but not 1981 or 1983. The ripe and unripe fruits of KJaInedbxa gabanensis and Cleistanthus michelsonii. co-dominants in mixed forest (T. Hart 1985) were also periodically abundantly available. Between 29% and 59% of all food patches recorded on transects contained items that had been bitten or showed other signs of having been handled and dropped by arboreal mammals. The damaged items included unripe fruits, ripe fruit parts and seeds. unripe pods and seeds of leguminous trees. partially eaten inflorescenses. and some foliage. Primates and large squirrels notably Protxerus stangerii. were responsible for most of this food rain from the canopy. Most of the unripe fruits which were dropped by primates were bitten once and 105 rejected uneaten. Squirrel drops included the pulp and exocarp of full-sized. but unripe fruits. from which the seeds had been extracted. Ripe fruit drops mainly consisted of portions inedible to primates, such as rinds or capsules. discarded after the contents had been eaten. Spontaneously—aborted fruits comprised a second major source of unripe fruit on the ground. Fruit abortion was especially prominent in the Irvigiaceae and Buphorbiaceae, but occurred in other families as well. Parent trees abort fruits for a variety of reasons (Stephensen 1981). In the present study. aborted fruits rarely showed evidence of insect damage. Some individual trees evidently aborted fruit over extended periods, as fruits under many crowns ranged in size up to more than half mature size. On average, 52s of food patches recorded on transects over the year were of the smaller weight classes (W < 100 g). These smaller food patches, however, represented only 8% of the total weight of available food (Table 4-3). Very large food patches (W > 1000 g) were rare and widely dispersed. These largest patches. however. accounted for almost 58% of the total weight recorded. The distribution of available food item sizes paralled that of food patch weights (Table 4-4). Small fruits. seeds and flowers (size classes 1 — 3) were numerically the most abundant, averaging 63% of all patches. Larger fruits (size classes 4 and 5). however. accounted for over 70s of the available weight. On average, the size of an item and the weight class (W) in which it was recoreded were positively correlated (rS - 0,90, p < 106 .oooou A 3 .e .ooood a» uso_ a 3 .n .aoos as so“ . 3 .u .ao« v 3 .z .u...._u s:o..: . .¢.n. .m.m. .~.«~. .¢.n. .m.m. .¢.~. .~.~. .o.o. ~.o~ °.mn u.un a.o~ c.5n n.¢n n.n m.o guaca>¢ .¢.o. .u.nu. .~.o~. .o.o~. .n.o_. .o.o~. .m.¢. .u.o. m.u" c.on 9.0. o.o~ n.ho n.5u «.n ¢.o e “a: cyan .a.o. .m.n_. .o.o. .~.m. .m.~. .o.n~. .~.N~. .m.~. n.» o.nn ¢.uc c.o~ c.5e o.un o.m« A." a u.3uu.= .o.n.. .o.n~. .d.o~. .n.¢_. .m.dn. .~.¢n. .o.n. .¢.o. as: «.ua n.nn “.mu n.nu .u.uo ~.nn u.n n.o c >~cam .o.o. .o.on. .o.s. .n.~u. .n.nn. .o.on. .o.o. .n.o. «.ou n.on ¢.o« o.n~ o.nn m.c¢ a." n.o u >ca c n u a c n u a count-a can-om cans-m noguuuu so coo-oceans; ozone: pus-cute so nooapc-ucom .A-coaua.>ou acute-u. was. ac... at. uaa~u> .nmod .>a= o» "no“ >~=a .ocuaN .u-ucou “Lava on» c" caucuses» co neuron-c .>->Luu-~ao -L.:o_~ ac. an... .uuflacs *o nua.u .3. ”gag-s :uuua noon .nnc egg.» 107 ..c.s..ss=.u 9.01 - s.n .n .0.» - o.~ .. .0." - a." .n .0.“ - 0.9 .N .n.o v .. . .L..as ..._u .ssm . .o.o. .s.a. .5... .¢.s. ...ss. .s.ms. .s.m. .n... in... o.na o.ou o.ns m.n A.sn s.nn s.». a.» n.n .o.t.>¢ .o.s. la.n~. .s.u~. .1.~. .n.s. .s.o. .m._. .s... .m.s. o..« m.«» 5.01 n.« o.~u s.un o.a~ ..n a.“ s.3 .».4 «.mu n..~ ”.m. n.n a... 0.1" ..o 9.91 ..m 3.3-usz .o.~. .o.ns. .m.m. .s.n. .a;m~. .~.m. .n.o~. .o.ns. ls... s.3 n.n _..n o.- a.» ~.o. n.na n.o~ s.n~ n.~ >_c.u .o.nz. .a... .«.n. .s.o~. .o.mn. .o.o.. .o.~. .o.o. .m.ns. 0.x" ~.ss o.«_ a..a ”.4” o.n. o.n s.o m.s sea n c a a n . n a a so...» nanny-n so accuse-need azouos pug-cut» so cognac-need .e nucouaau>09 tunucauu up... 2.3.. 0L0 0.3—.) .nmo~ .>o= ou “no" >_=n .-c_- .uuocom “can" as» as caucus-cu co u-uLou-c .>->auuo_~ou custom» ac. an... .uu.=c* +0 auouadu sumo .cuv can.» 108 .10, n - 5). This relationship was variable, however. especially for larger items. In a survey of 174 recorded items (Figure 4-4). patch weights of most smaller-size classes were low. but ranged over 1000 g in some instances. For example, the bracts and unopened flowers of MM laurentii and Gilbertiadendran dewevrei (size classes 1 and 2) and the aborted unripe fruits of Kleinedoxa gabonensis. Irvingia grandifoiia and I. wombalu (size classes 2 -3) characteristically occurred in large. dense patches. The relationship between the areal extent of a patch and its total biomass was less variable. Isolated items ranged in weight class depending upon the weight of the fruit. For example. certain of the larger unripe fruits (Landalphia app. Cola lateritia) sometimes occurred as isolated items in small scattered patches, especially when they had been carried some distance and dropped by primates. With the possible exception of primate drops and mast seed fall of the Caesalpiniaceous dominants. where contiguous crowns produce simultaneously. food patch areas were discrete and associated with the crown of a single tree. The largest patch area recorded was under 300 m2, and most patches were less than 150 m2. No patch of large area (> 90 m2) recorded during the study had a patch weight of under 100 g. In characterizing food patches. patch weight (W) provides a measure of the total food biomass available to be harvested. Patches of low weight (W < 100 g) contained small, scattered items of low unit weight or single isolated items of larger weight. Patches of larger total biomass contained larger itmes and/or smaller items in dense 109 Figure 4-4. Relationship between item size class and patch weight (W) for 174 fruits. flowers and seeds collected in the Ituri Forest. Zaire, July, 1981 to May, 1983. 110 49 # .5 ..... n .4 u “a u .I. ”1.... L... w. ._ m. . Aoucacux. «ovxaEmcu .su.e3 sou-a size class Item 111 patches. General Diet Commsition Dietary analyses were based on rumen contents from 175 animals collected during nine sampling periods from December, 1981 through May, 1983 (Table 4-5) Additional rumens collected between June and December. 1981 were analyzed for species poorly represented in this sample including ll. equations (8 rumens). C. nigrifrons (5 rumens) and C. syivicuitor (3 rumens). This brought the total collection to 191 rumens. Four duikers. including a small species, C. monticoia. two medium-sized species. C. leucogaster and C. callipygus and a large species. C. dorsalis were sampled during every season and most sampling periods. Between 240 and 250 species of fruits, flowers and seeds were recorded from the rumens of all seven ungulate species over the course of this study. The number of fruit, seed and flower species recorded per rumen varied from zero in rumens containing only foliage and/or fungi to over 23. In the majority of samples. however. the two most important food species accounted for 67% to 100% of the total sample weight. Averaged over all rumens examined during the entire study. there were no significant differences in dietary diversity (number of food species per rumen) between any of the seven ungulate species (Figure 4-5). Average dietary diversity was not correlated with body size (rs - 0.07. p > .05, n = 7). The largest and smallest species exhibited the lowest average food diversity in the rumen. .ln. carcasslaxc .o ..n. ceaseless: .8 ..m. ssc-~=s. .a .fioo— .c-n-ou-a can scan cu-xu-n uo~acnu nucoucou couac -co.u«uu¢ a — s" on cm 5 cm on adage» o c s a a s n s no 3.: mac-u o n n s o o A m 3.xnm_t.m\>ta o o o o a n o x no sea 0 a ¢ a o a a" n no as: can; 0 n a n o o m a No 3.3 u.z m o o o a o n a s. S s... it: o a a . s o m c «a “.3 >sc.m o n as s o n n“ m «a sea a u o ¢ a n o" z do us: cue; cog~au~s~xu n-~ncou naoxau-uu ceanauouao~ neoceucoue sauna-zen -ou_.co- nacao~qcacu u=¢a8~ocaou cacao~ocaeo nacao~cgaou nacao-¢auu “scone-o»: nagao~acauo coma-nob coca-m a .oc..~ .u-ccoa «Luau .noou .>~z u:— dood .cons-uoa coezuon >~Huconauu nou>~acu unusaaoca unoco>.o=cs so nae-acou cocax .nuc o~aub 113 Figure 4-5. Numbers of fruits, seeds and flowers in rumens of six species of duiker and the chevrotain in the Ituri Forest. Zaire. Values shown are means and standard deviations averaged over 13 sampling periods. December. 1981 to May, 1983. Abbreviations: C m, cephaiophus monticala; H a. Hyemoschus aquaticus; C n. C. nigrifrons; C 1, C. leucogaster; C c, C. callipygus: C d. C. dorsaiis; C s. C. sylvicuitor. Numbers of rumens sampled indicated below each species' abbreviation. 114 8 ‘11- m a v o ' 10 O 13 A D 8 UN 0'- 4 0.. ”O J 10] ON I on :N . .c 0" v 3 CG 1 ‘ z 1:“ I 42 3 >~ '6 O D E . .= : m L l j l l P 3 w o ' N o (as 3X) uemn 1 "I CHOOGB to aequanu 115 suggestive that the range of foods these species utilized was perhaps limited. Highest average diversity was found in the two nocturnal species. 1!. equations (mean 7.4 species per rumen) and C. dorsalis (mean 7.5 species per rumen). The riparian species, H. equations and especially C. nigrifrons showed the greatest variability in dietary diversity. indicating possible roles for activity pattern and habitat differences in determining diet. The two diurnal upland duikers of similar size. C. ieucagaster (6.1 species per rumen) and C. callipygus (4.6 species per rumen) differed in average dietary diversity providing evidence that these otherwise similar species differed in foraging patterns. Fruits and seeds comprised the bulk of of the diets of duikers and chevrotain throughout the year. Flowers and fungi comprised a variable small percentage of the total. Foliage was eaten in variable amounts but on average was low and sometimes absent (Table 4-6). Several authors (Gautier-Hion et al. 1980, Dubost 1984) have suggested that there is a relationship between body-size and insectivory in forest ungulates. Insect remains (mainly large ants) occurred in 62% of C. monticoia rumens. They accounted for < 1% of the 2 mm sample in all but a few cases. In no sample did insect remains surpass 3.5%. Insect remains were only found in 20% of the rumens of larger animals. and then mostly in trace amounts. While some of these insects may have been ingested accidentally. this was probably not the case with all. In particular. ants killed by a fungal disease (and sometimes sprouting small gilled mushrooms) were ingested. These ants were found dead as solitary individuals in a characteristic pose. grasping low vegetation 1:16 table 4-6. Percentage cosposition of large (1.5 as) particles screened (roe rusen contents of adult and weaned-Juvenile duikers and chevrotain. Ituri Forest, Zaire. 1981 to 1983. Values are grass wet weight. giving aeans, standard deviations (in parentheses) and ranges. Species Ripe fruit Unripe truit 8eed Flowers Fungi Foliage c. aorticola 6.6 33.6 40.9 4.5 5.8 8.1 n I 78 (7.8) (32.0) (31.9) (5.5) (0.4) (12.6) 0 - 20 3 - 92 4 - 3 0 - 16 0 - 32 0 - 39 N. aquatics: 9.2 32.7 38.0 1.7 3.2 11.4 n I 26 (4.7) (28.0) (32.2) (3.4) (6.2) (16.2) 4 - 15 0 - 85 0 - 84 0 - 10 0 - 17 0 - 37 C. niprifronr 10.7 24.0 46.4 0.3 0.4 29.0 n I 12 (15.4) (28.1) (40.0) (0.7) (1.1) (44.2) 0 - 46 0 - 73 0 - 100 0 - 2 0 - 3 0 - 100 C. leucopastar 18.1 46.2 21.1 7.7 0.7 6.2 n I 24 (13.5) (19.9) (30.6) (13.3) (1.3) (8.5) 3 - 42 1 - 72 0 - 80 0 - 38 0 - 4 0 - 24 C. caliipygu: 24.5 39.0 24.7 6.0 3.8 5.7 n I 30 (12.5) (21.6) (29.5) (11.5) (7.2) (6.2) 12 - 50 11 - 67 0 - 62 0 - 32 0 - 18 0 - 14 C. dorsalir 31.6 44.4 17.0 3.1 1.0 3.5 n I 17 (21.9) (25.3) (29.0) (3.5) (2.3) (5.5) 14 - 61 17 - 83 0 - 79 0 - 9 0 - 6 1 - 15 c. sylvicultor 96.7 4.3 0.0 0.0 0.0 3.0 n I 4 (5.5) (5.1) (0.0) (0.0) (0.0) (5.5) 90 - 100 0 - 10 0 0 0 0 - 6 117 in the understory. This is similar to the postures assumed by flies killed by Entomophthora fungi (T. Hart. pers comm). Why duikers should selectively ingest these ants was not clear. There was no further evidence for faunivory by any species except for the partial skeleton and flesh of a black mongoose Crassarchus obscura (probably scavenged) from an adult C. dorsalis rumen. Most of the ripe fruits eaten by duikers belonged to species which are seemingly specialized for seed dispersal by large terrestrial mammals (van der Pijl 1972, Alexandre 1978). Many of these fruits were quite large or of low nutritional quality. Other ripe fruit on the forest floor included the less palatable rinds and capsules discarded by primates. Ripe fruits and fruit parts were more important in the diets of the larger animals than the smaller ones. The average percentage of ripe fruits in the diet was correlated with body size across all seven ungulate species (rs . 1.00, p < 0.001, n = 7). Values ranged from 6.6% of wet weight contents in the diet of C. monticola (for which ripe fruit never averaged more than 20% of the diet in any sample period) to over 97% in that of C. syivicultor. Patterns in edible seed consumption were in the opposite direction to those of ripe fruits. Average percentage seeds in diet of all seven species was negatively correlated with body size (rS = -0.93. p < 0.01. n -= 7). Seeds accounted for more than 5% of the diet of C. monticoia during all nine sampling periods and accounted for over 40% of the diet in five samples. Edible seeds were also favored by the three red duiker species. Their importance in the diet. however, was limited to periods of mast seed fall of the 118 Caesalpiniaceous forest dominants or during periods when primates dropped or defecated seeds of Landophia spp and Cola lateritia in abundance. During the remainder of the year. seeds generally only occurred in small scattered patches and were not used by the larger animals. Arcsine transformed percentage foliage was correlated in 2 mm and 5 mm screened samples from the same rumen (R2 .. 0,35, p < 0.001. n I 46). Foliage accounted for only a small proportion of the diet of all species, except during the dry seasons. Most of the foliage which could be identified in the rumens came from canopy trees. Much of this consisted of new leaves. Minor amounts of older. dead leaves and what appeared to be r00t fibers from the forest floor also were found. Foliage of the dominant understory species was unpalatable to C. monticala in feeding trials (Chapter 3) and was absent from the rumens of all species. Foliage is apparently available year-round. Leaf flush in the evergreen canopy dominats, Brachystegia laurentii and Gilbertiodendron dewevrei occurred in all seasons. Leaf flush in most deciduous canopy species occurred in the late dry and early wet seasons. Use of foliage by the ungulates increased at this time. but foliagewas not a major componenet of the diet. Foliage was only preferred to lowest ranked fruits or seeds in palatability trials with two duiker species (Chapter 3) and apparently was not a favored food of free-ranging duikers either. 119 Food Selection Although duiker diets were often diverse, many food specis contributed only a small percentage to the total diet. While some may have been nutritionally significant (Freeland and Janzen 1974, Oates 1977). preference levels for these species were difficult to ascertain because many were also rare in the environment. These species are not further considered here. Data on food availability and use adequate to ascertain food selecion were available for 8 to 19 food species in each of nine sampling periods between December. 1981 and May. 1983 (Appendix 4-8). These species all had utility indices of at least 5% and/or comprised at least 5% of available food (total transect weight). This analysis of food selection focuses on four species of upland duikers for which data are adequate, including a small species. C. monticola. two medium-sized species. C. leucogaster and C. caIIipygus and a large species C. damalis. The abundance of a potential food on the transects was not necessarily an indication of its importance in the diet (Table 4-7). During each sample period, between 40% and 83% of species comprising at least 5% of transect weight were avoided by upland duikers. There was no significant difference between duiker species in the percentages of foods avoided (t .. 0.73. p > 0.05). All species fed selectivley. Many of the preferred foods were relatively uncommon. An analysis of the characteristics of preferred and avoided foods permits an evaluation of specific dimensions of the food resources which are important to each of the ungulate species and along 120 Table 4-7. Selection by four upland duiker species for foods which were abundant on transects ( > 51 total weight). lturi Forest, Zaire, 1981 to 1983. '. 8eason 8pecies‘ a Food date Food species typs‘ C.s. 6.1. c.c. c.a. Late let Klairedoxa paboaesris RF/UF + + + + Dec, 81 K. triilesii RF - - - - Dapaca puiaeessis Fl - - - - Dry K. triilesii RF - - - - Jan, 82 Albiaia pussifara UF - - - - K. pabowassis RF/UF - + + + Ficus sp 4 HF + + + + Early wet K. pabowewsis RF - - + + Bar, 82 Pacovia harssiana UFIB + + r I Early wet K. pabossasis RF - - Ray, 82 Pascoria harssiasa 8 + - NR IR Cleistasthus sichelsoaii UF - - Ceitis adoifi-fridaricii RF - - dildortiodasdroa dewevrai Fl - - Hid wet lrachyrtepia iaureatii Fl - - - - Aug. 82 Lasdoiphia spp UF/B _ I + + unknown 'kotou' RF - - - liiphia welwitschii RF/8 + + + Late wet K. trillesii UF - - - - Ottg .2 0e d.”.".i 8 " T T 7' I. iaurastii 8 + - + - lawdoiphia spp RF/8 + + + + I. laurewtii 8 + - t ‘ Dry K. paboaessis RF/UF + + Feb, 83 K. triliesii RF - - NR NR 8. dewavrai F1 - - 8. laurewtii 8 + + DIE wet K. pabosewsis RF - - - Feb-Har, K. triilesii UF - + - - 83 C. sichelsosii UF + + * Taraswa iaurastii UF - - - D. dewsvrei Fl - - * Early wet K. pebowessis RF - - - Ray. 83 D. iaurestii Fl - ' - ' unknown 'lipasa' UF * ' t * ‘ 8electivity: +. food selected or eaten in proportion to availability; -, foods avoided) NR, no data for duiker species. ' Food type: RF, ripe (ruit) 0F, unripe iruiti Fl. (lowers; 8. seeds. ‘ Upland duiker species: C.s., c. sosticola; 0.1.. C. leucopaster) C.c, c. caiiipypusr C.d., c. dorsalis. 121 which animals' diets might be segregated. This analysis examines selection in relationship to the taxonomic identity of the food. food patch weight. item size and nutritional quality. Taxonomic Identity Six food species known to be eaten by duikers were available during more than one sampling period (Table 4-8). None of these species was a preferred food for duikers at all times it was available. Unripe K1ainedoxa gabonensis fruits were selectively eaten by all species. ripe fruits of the same species were often avoided by the red duikers and were always avoided by C. monticola. Flowers of Gilbertiodendron dewevrei and Brachystegia laurentii were less preferred than the ripe seeds. While consistency of duiker food preference was noted with captive animals during controlled feeding trials (Chapter 3) this was not the case in the field where abundance and relative availability of alternative foods may be a factor in determining preference. While no food was consistently preferred in the diet. a number of species. regularly recorded on transects were consistently avoided, even when abundantly available. Food $93:ch Two hypotheses on the importance of food patch weight in food selection were examined for each duiker species: 1. Preferred food species occur equally in patches of both large and small weight. 2. Avoided food species occur equally in patches of both 122 Table 4-8 Changes in preference by (our species of upland duikers For six food species available during more than one sampling period. Ituri Forest, Zaire, December, 1981 to hay, 1983. Selectivityb Species Season Food Type‘ C.s. 6.1. C.c. C.d. Klainedoxa L 81 UF ++ ++ ++ ++ paboaenri: D 82 UF + ++ ++ ++ E 82 RF - - ++ - H 82 RF - - - ++ L 82 RF - - - - D 83 HF ++ ++ no data DIE 83 RF - ++ - - E 83 RF - - + + Brachystepia H 82 F1 - - - - laurentii L 82 8 ++ - - D 83 S ++ ++ no data DIE 83 S + - + - E 83 F1 + - - + Gilbertiodandron E 82 F) - - no data dewevrei L 82 S - ++ ++ ++ DIE 83 F1 + - + - Cleistanthus E 82 UP - no data sicheisonii DIE 83 UF ++ ++ - Pancovia E 82 UF/S ++ ++ ++ ++ harnsiana E 82 8 ++ - no data Ricinodendron L 81 RF - ++ ++ ++ heudeiotii H 82 RF - - - ++ L 82 RF - - - - ‘ Food type: UF, unripe fruit; RF, ripe fruit; S, seed: Fl, flower. 9 Selection: ++, preferred: +, eaten in proportion to availability; 'g .VOid.ds 6.1., C. leucopaster; Duiker species abbreviations: C.c., C. caiiipypus; C.s., C. sonticoia; C.d., C. dorsaiis. 123 large and small weight. There was no significant difference in patch weight of foods preferred by C. monticola (Table 4-9). Of 30 preferred food species in which patch weight was known, 13 (40%) occurred in large patches (W > 100 g) while 17 (57%) occurred in patches of W < 100 g. In contrast, significantly more preferred foods had large patch weights for C. ieucogaster (21 of 27 foods) and C. dorsaiis (18 of 24 foods). In diets of C. caflipygus 21 of 30 preferred foods in which patch weight was known were large. This was not significantly different than 15 which would be expected if theis species exhibited no selection for food patch weight (1:2 . 2.40, p > .10, 2 df). Although most preferred foods of C. callipygus had large patch weights. at least some foods of small patches were also preferred during every sampling period. In all four duiker species there was no significant trend in patch size of avoided food species (Table 4-9). Avoided foods included species which occurred in both large and small patches. Food_ Item Size The size distributions of preferred and avoided fruits. seeds and flowers (Figure 4-6) were examined with respect to two hypotheses: 1. The size distributions of preferred foods are equivalent across all duiker species. . 2. Within duiker species. preferred and avoided foods are of equivalent size. Species by species comparisons (Table 4—10) demonstrated 124 Table 4-9. Numbers of preferred and avoided food species of large patch weight (W > 100 g) and small patch weight (H g 100 g) in diets of four species of upland duikers. Ituri Forest, Zaire, 1982-i983. Duiker: eonticoia leucopaster callipygus dorsaiis Preferred foods ‘ Large patch weight 13 27 21 24 Seal) patch weight 17 6 9 6 Total 30 37 30 30 12 0.53 6.68 2.40 5.40 P = (4.8. ~<.01 ms. (.05 Avoided foods 9 Large patch weight 29 28 19 24 Seal) patch weight 16 21 15 16 Total 45 49 34 40 X2 1.88 0.50 0.24 0.80 P ‘ N.S. N.S. N.S. N.S. ‘ Nuebers of preferred foods of unknown patch wieght: eonticola, 5; leucogaster, 3: callipygus, 8; dorsalis 2. 9 Numbers of avoided foods of unknown patch weight: eonticola, 2; leucopester, 2: caiiipygus, 1: dorsaiis, 1. = Probabilities of p > .10 are not significant (N.S.). 125 Figure 4-6. Size distributions of preferred (fine stipling) and avoided (coarse stipling) foods in diets of four species of upland duikers. Ituri Forest, Zaire. December. 1981 to May. 1983. 126 a..ee.om .0 mm<40 a: m>a_:eo .0 wN_m DOCK .e.ssmoose_ .o I.°0_-=°E .0 SSIOBdS $0 HEBHHN 127 Table 4-10. Chi square values for tests of A) equal food size distributions of preferred food species between diets of four species of upland duikers, and 8) equal size distributions of preferred and avoided foods species within diets of each duiker speces. Duiker: eonticoia ieucogaster caliipygus dorsalis A) Between species‘ eonticola -- 11.16 6.85 16.82 (805 NISI (8°01 ieucogerter -- 2.36 1.91 N.S. N.S. callipypus -- 4.50 N.S. dorsalis -- 8) Within species ‘ Duiker eonticola leucogester cailipyqus dorsalis x2 14.85 3.19 1.75 6.38 Probability (.001 N.S. 8.8. (.05 ‘ Chi square values with 3 degrees of freedom. Probabilities of p > .10 are not significant (N.S.). 128 significant differences in size distributions of preferred foods only between C. monticoia and the two red duikers C. ieucogaster and C. dorsalis. Sizes of preferred foods were not significanlty different between C. monticola and C. callipygus or between the three red duiker species. Within species comparisons demonstrated that the sizes of preferred and avoided food species were significantly different in diets of C. monticola and C. dorsaiis but not in diets of C. calflpygus or C. Jeucogaster. Preferred food species were under-represented in the largest size class in the diet of C. monticola while avoided food species were over represented. In C. dorsaiis the reverse trend was apparent. Most preferred foods were of large size while relatively few avoided species were of this class (Figure 4-6). Patterns of food size selection revealed that both mouth size and shape contributed to the distribution of food sizes an animal will ingest. Overall, larger animals with larger mouths can ingest larger foods. This restricts C. monticola to smaller food sizes. Two species. C. manticola and C. callipygus of differing body size but similar narrow mouth shape included small-sized items among the foods they selected. Cephalophus callipygus, the largest of these two species also included a larger proportion of large food sizes in its diet as well. Species with relatively broad mouths, C. Ieucogaster and C. dorsaus preferred larger foods and avoided smaller food items. The largest of these two species, C. dorsaiis selectively fed on fruits of the largest size class. Many of these, 129 including the ripe fruits of Klainedoxa gabonensis, and Irvingia grandifalia were tough and fibrous. Their inclusion in the diet was associated with this species' relatively heavy jaw musculature (Kingdon 1982) as well its broad mouth. Food Nutritional Quality Palatability trials with captive C. montiocia and C. dorsalis (Chapter 3) deomonstrated taht preferred foods generally had high values of adjusted dry matter yield (Yo). Selection for food nutritonal quality (YO) in free-ranging animals was investigated according to two hypotheses: 1. Adjusted dry matter yield (‘10) of selected foods is greater than avoided foods. 2. Values 01' YO of preferred foods are greater than the average available value (YQ(T))' The adjusted dry matter yield of available foods (YQ(T)) varied widely over the five sampling periods in which it was measured (Figure 4-7). Food values of Y0 analyzed on a seasonal basis revealed that in diets of C. manticola. preferred foods had significantly higher values than avoided foods ((1 test. p < .10) in four of the five sampling periods (Table 4-11). Values of preferred and avoided foods were not singinificantly different only during the early wet season of 1983. Average food value (YQ(T)) at this time was low and high quality foods were rare. The three upland red duikers species were not as consistently selective for 1’0008 With high YQ values as was C. monticola 130 Figure 4-7. Adjusted dry matter yield (YQ) of selected (closed circles) and avoided (open circles) food species in diets of four species of upland duikers in comparison with average adjusted dry matter Vi81d (YQ(T)) of apparently available fruits, seeds and flowers on transects (connected squares). Ituri Forest, Zaire. October. 1982 to May, 1983. o ._ s . o m N m 0 a V n .2: o 18... o o 468 C—...s°fi .0 20m Avoided Duiker Hid Wet Late Wet Dry Dry/Early Wet Early Wet 1982 1982 1983 1983 1983 C. eonticola (.05 (.10 (.01 (.01 N.S. C. leucogaster (.05 N.S. (.10 N.S N.S. C. callipypus N.S. <.10 no data (.05 N.S. C. dorsali: (.10 N.S. .no data N.S. N.S. 8) Preferred > Average C. aonticola C. ieucoqaster C. callipygus C. dorsalis > Average 11 9 8 11 < Average 5 8 8 7 X“ 2.25 0.03 0.00 0.88 Probability N.S. N.S. N.S. N.S. ‘ Data are values for food items shown in Figure 4-9. N.S. , difference not significant, p > .10. 133 (Figure 4-7). Preferred foods included species with high values, but also species with low values. Values of Y0 were significantly higher than avoided foods in only two of five sampling periods for C. leucogaster. two of four samples for C. callipygus and only one of four samples for C. dorsalis (Table 4-11 A). Summed over all five sample periods. values of Y0 of preferred foods were not significantly greater than average available fOOd value (YQiTl) in any duiker species (Table 4-11 B). An examination of the trends on a sample by sample basis, however (Figure 4-7). reveals differences between the species in preferences for foods 01' high Yo values. During periods when high quality foods were abundant and average nutritonal value ’of available foods was high (late wet season. 1982 and dry season, 1983, during macystegia laurentii seed—fall), selected foods of all duikers included species of below average value. During periods when average food values were lower (mid wet season, 1982 and early wet season. 1983) YQ values of preferred foods in diets of C. monticola were singificantly greater than average available values (10 of 11 food species. x2 x 7.36, p < .05). The upland red duikers. C. Jeucogaster, C. callipygus and C. dorsalis selected foods of less than average value during these same three sample periods. Four of 14 preferred foods in diet of C. leucogaster had YQ values less than average. Three of seven preferred foods in diet of C. cailipygus and four of seven in diets of C. dorsalis were similarly below average quality (X2 values 2.58, 0.14. 0.14 for each species respecitively, p >> .10). 134 Diet Overlap Dietary overlap (In) between the three species of upland red duikers. C. leucogaster, C. cailipygus and C. dorsalis ranged from zero to over 0.90 (mean 0.55) over seven sample periods for which there were adequate data (Figure 4-8 A). The diets of all three species converged during four sample periods and exhibited marked divergence during three periods. Dietary overlap between the three red duiker species and C. monticola ranged from zero to 0.60 (mean 0.43) over this same period (Figure 4-8 8). Rank values of 1M for diets of C. callipygus and C. manticala were significantly higher (mean 0.55) than between C. leucogaster and C. manticola (mean 0.39) or between C. dorsaiis and C. monticala (mean 0.32) (Friedman 2-way ANOVA, p < 0.001). There were, however. no periods of marked dietary convergence. Overall. dietary overlap between C. monticola and the red duikers was never as high as between the three red duikers species. During periods of dietary convergence. the dominant foods in the diets of the red duikers were both high quality and abundantly available (Table 4-12). Among the shared food species at this time were species with high values of V0 as well as fruits which had high preference ranks in feeding trials with captives (Chapter 3). Most of these foods comprised at least 4 % of total food transect weight and up to 28% of total patches recorded. By comparison, during periods of dietary divergence. high quality foods were scarce. although total food diversity was high. This 135 Figure 4-8. Values of Morisita's (1M) index for dietary overlap between: A) three species of upland red duikers. C. leucagaster (l). C. callipygus (c) and C. dorsalis (d) and 8) three species of upland red duikers and C. monticoia (m), for seven sample periods. Ituri Forest, Zaire, December. 1981 to May, 1983. Lfi mmomw «map >co \awa x xW/////x \o/////M moo VOA 14in $6 use. o.ex.sm to... as... .0 r 1.94 Us. 7» 06 on. .W 0 two 0W /// 01 aces.:m sec.< .Qo 04 is xepu: delieno ("1) 137 Table 4-12. Duality and abundance of dominant foods (In > ax: shared in the diets of at least two of three species of duikers, C. leucopaster, C. callipypus and C. dorsalis during periods of high dietary overlap. Ituri Forest, Zaire, December 1981 to Hay, 1983. Season Species Abundance on transect Food Duality X weight I patches Late Klainedoxa preferred in trials Wet gabonenSiS high Ya 1981 unripe fruit 51.2 10.4 Ricinodendron preferred in trials heudelotii high Ya ripe fruit not recorded Dry Klainedoxa preferred in trials not recorded 1981 gabonensis high Ya unripe fruit fungi low quality 7 abundant low dry matter yield low quality foliage most not preferred in trials not measured Hid Landoiphia spp high Ya Wet unripe fruit 3 35.7 7.3 seeds Croton aubanga high Yo ripe fruit not recorded Flighia weiwitschii high Ya ripe fruits & seeds 4.3 1.2 Late Cilbertiodendron high Yo Wet dewevrei 1982 ripe seeds 6.7 1.2 Landolphie spp high Ya ripe fruit 5 seeds 10.5 27.6 138 is evident by examining the range of foods selected by each of the duikers which were not found in the diets of the other species. (Table 4-13) . Most of the domianant foods (IU > 5%) eaten by each species during these periods were unique to the diet and not shared with the other two species. The only exception to this trend being the C. callipygus diet during the early wet season of 1982. Characterisitics of these unique foods indicated that each species appeared to adopt differing foraging and feeding behaviors. C. leucagaster evidently was highly mobile. It appeared to attend primates and its diets were dominated by widely dispersed foods which occurred in ephemeral patches such as flowers and soft figs (Ficus). Diets of C. callipygus were broadened during these periods to include increased percentages of scattered food patches of smaller patch weight. This indicates that this species may have had restricted movements and made increasing use of foods which were ignored at other seasons. The diet of the largest red duiker. C. dorsaiis consistently included very large and in some cases low quality fruits including ripe fruits of Kleinedaxa gabonensis, Irvingia grandlfolia and I. womboiu These fruits were difficult for smaller species to handle. An apparent exception to this trend occurred during the dry season sample of 1982 (Table 4-12). Unripe fruits of K. gabonensis occurred in all diets but the fruits were rare in the forest. Foliage and fungi were dominant in the diets. but both were of apparent low quality (see discussion). A further feature of food availability during this period was that both total availability and food diversity 139 Table 4-13. Numbers and characteristics of dominant food species (In > 5%) unique to diets of each of three duikers, C. leucopaster, C. callipypus and C. dorsalis during periods of dietary divergence. Ituri Forest, Zaire, March, 1982 to Hay, 1983. Number Number dominant unique Sample Duiker foods foods Characteristics Early ieucogaster 6 3 primate drops Wet 1982 callipygus 2 0 None dorsalis 5 2 large hard fruits Dry/ leucogaster 7 4 primate drops, low Ya, Early 2 spp soft Ficus Wet 1983 callipygus 4 2 primate drops, small size dorsalis 7 5 large, hard, low quality Early ieucopaster 2 2 flowers, aborted fruits Wet 1983 callipygus 5 4 small patch, hard Ficus dorsalis 4 3 large, tough fruits/capsules 140 were lower than during any other sample period recorded (Figure 4-4). Dietary overlap between the riverine species and upland duikers was variable but often high (Table 4-14). Values of IM ranged from 0.50 to 0.97 (mean 0.70) for H. equations and C. montiooia and from 0.00 and 0.77 (mean 0.46) between H. equations and the three red duikers species. Dietary overlap between C.nigrifrans and upland red duikers ranged from 0.01 to 0.22 (mean 0.22) and between 0.06 and 0.95 (mean 0.50) with C. montiooia. Overlap values for diets of the two riverine species were comparable to those between riverine and upland species. Periods of high overlap in the riverine species paralled those in upland duikers and occurred mainly during periods of food abundance in upland forests. Diets diverged during periods when these foods were absent (Table 4-14). These patterns demonstrate that neither C. nigrit‘rons nor H. equations were confined to riparian habitats for foraging. Instead it appears that these habitats were a refuge, at least for C. nigritronsmuring periods of potential competiton with upland species. This was less clearly the case for H. equations Dietary overlap between H. equations and C. monticoia was consistently high over all sample periods it was measured. H. equations forages in upland forests at night and apparently only retreats to the water—side for shelter during the day (see also Dubost 1978). Although data on diet and behavior of H. aquations were limited. it was neither as active nor as selective a feeder . . as C. montioola. (J. Hart, unpubl. obs.. see also 1J41 Table 4-14. Values for Norisita's (In) index for dietary overlap between two riverine species, N. aquaticus and C. nigrifrons, and between two riverine species and upland duikers. lturi Forest Zaire, 1981 -l983.‘ N. aquaticus C. niprifrons Available Riverine blue upland red Blue Upland red Date food“ species duiker duikers duiker duikers Jun, 81 0.04 0.63 0.33 0.06 0.01 Sep, 81 Fruit 0.52 0.50 -- 0.82 -- Oct, 81 Rest -- 0.72 0.62 -- -- Dec, 81 Fruit 0.92 0.97 0.55 0.95 0.61 Jan, 82 -- 0.83 0.77 -- -- Her, 82 -- -- -- 0.31 0.26 Hay, 82 Fruit -- 0.78 0.00 -- -- Oct, 82 Heat -- 0.72 0.62 -- -- Feb, 83 Heat 0.52 0.50 -- 0.82 -- Ray, 83 0.04 0.63 0.33 0.06 0.01 Average 0.41 0.70 0.46 0.50 0.22 ‘ No value shown (--) indicates one or both species absent and comparison could not be made. C. leucopaster, C. cailipypus and C. dorsalis. Overlap value for red duikers is average for ° Periods of abundant high quality food in upland forest: fruit, rips or unripe fruit: mast, Caesalpineaceous mast seed fall. 142 Dubost 1978. 1975). Patterns in Abundance of Frggivorogs Ungggates Ungglate Distributions Four species. C. monticola, and the upland red duikers C. Jencogaster, C. cellipygns and C. dorsalis, were recorded at all nine sample sites and accounted for the largest percentage of total animals flushed (Table 4-15). The smallest species. C. manticole was the most abundant, averaging 52% to 65% of total observations with 14.9 animals flushed/kmz. The upland red duikers. C. Ienoogaster. C. oaliipygus and C. dorsaiis. accounted for most of the remaining observations and together averaged 7.4 animals flushed/kmz. Two riverine species. C. nigrifmns and H. equations were encountered on drives in the vicinity of streams. Cephalophns sylvionltor was irregularly and less frequently recorded in upland forest. Excluding the riverine species. average abundance and body weight were significantly and negatively correlated (R = -0.92. p < .01. n = 5). “0 red duiker species of similar body size. C. Iencogester and C. oaliipygus varied in aspects of their social behavior and in their distributions. Average group size (number of animals flushed together) in C. oaiiipygns was 1.3 with 24 % of flushes including more than one animal. Group size in the similarly-sized C. iencogaster was 1.1. with only 12.5 % of flushes containing more than one animal. 143 Table 4-15. A) Nuebers flushed! he', and 8) 1981 to 1983. percentages of duikers and chevrotains flushed on drive hunts at nine sites in the lturi Forest, Zaire, Site‘ Standard Species .K R E N 8 A P 8 1 mean deviation A) Abundance (number flushed! is”) C. aoeticoia 15.6 13.5 11.9 10.3 17.5 11.9 13.7 15.7 24.2 14.9 4.14 8. equaiticus 2.7 0.6 N.R. 1.7 0.6 N.R 0.3 0.8 5.5 1.4 1.78 C. eiprifroes 0.5 0.2 N.R 1.7 N.R. 1.1 0.3 N.R. 1.4 0.6 0.65 C. ieucopaster 4.4 2.5 4.0 2.6 1.2 3.4 2.7 2.4 1.4 2.7 1.07 C. caiiipypus 2.0 2.9 0.8 2.1 6.2 4.0 0.6 6.3 4.1 3.2 2.10 C. dorsaiis 1.5 1.9 3.2 0.4 0.8 1.1 2.1 2.4 1.2 1.5 0.73 C. sylvicuitor 1.0 1.2 N.R. 0.9 0.6 N.R. 0.9 1.6 8.8 0.7 0.58 Total flushes 27.7 22.8 19.9 19.7 26.9 21.5 20.6 29.2 37.8 25.1 5.90 per he“ Total flushes 168 120 25 46 128 33 72 38 33 663 Total area 6.06 5.26 1.26 2.33 4.76 1.54 3.58 1.30 0.87 26.96 sampled tea 8) Percentage total animals flushed C. aoeticoia 57.1 60.8 60.0 52.2 64.1 54.5 65.3 52.6 60.6 63.0 5.1 N. equations 9.5 2.5 N.R. 8.7 2.3 N.R. 1.4 2.6 12.1 5.6 4.4 C. aiprifroes 1.8 0.8 N.R. 8.7 N.R. 6.1 1.4 N.R. 3.0 3.6 3.1 C. ieucopaster 15.5 10.0 20.0 13.0 4.7 15.2 13.9 7.9 3.0 12.3 5.8 C. caiiipypus 7.1 12.5 4.0 10.9 23.4 21.2 2.8 21.1 9.1 13.3 7.9 C. dorsaiis 5.4 7.5 16.0 2.2 3.1 3.0 11.1 7.9 9.1 7.7 4.5 C. syivicultor 3.6 5.8 N.R. 4.3 2.3 N.R. 4.2 5.3 W.R 4.6 1.3 recorded in sample. For location of sites, see Figure 4-1. N.R. indicates no individuals 144 C. callipygus appeared to occur in pairs or as family parties. These assemblages were irregularly distributed over the study area. Indices of abundance for C. oeilipygus varied from 0.6/km2 at site P. to 6.3/km2. at site S. No differences were noted in the mixed forest composition at locations where this species was present and where it was absent. Averaged over all nine sampling sites, the abundances of C. oeIIipygus and C. lencogaster were not significantly different (2.7 and 3.2 animals flushed 002. t = 0.87, p >> 0.10 n = 9). On a site by site basis. however. the relative abundance of the two species V was negatively correlated (R = -0.66, p = 0.05, n = 9). Unflilate Abundance and Patterns in Food Abandance There were few discernible trends in total upland duiker abundance between sites in mixed forest of comparable hunting hisotry. Differences in abundance between the two mature forest types on the study area, mixed forest and the mbau forest were apparent. At two sites. R and E, where drive samples included both forest types. the average abundance of all upland species combined was negatively correlated with the percentage mbau forest in the drive area (Figure 4-9; r 8 -0.96. p (.01. n = 5). S Overall patterns of food abundance in borth forest types were affected by the irregular cycles of synchronous flowering and seed-fall of the Caesalpiniaceous dominants of each forest type Braohystegia 145 Figure 4-9. Average ungulate abundance (1 SE) on drive areas of differing percentage mbau forest cover at two sites. K and E. Ituri Forest. Zaire. Sample sizes indicated for each percentage class. 146 T 25;. d:- u: 1 (D +' 20 _ ,_ T 18 IX v ‘1- N ‘1- 5 15 . 8 r \ ‘U -- .1:— 9 1 n w 10 )- 7 a 7" " s m m: h 1 E s . 4 o r: ‘ o l L g L 1 0-19 20-39 40-59 60-79 80-100 Percent mbau forest in drive area 147 laurentii in mixed forest and Gilbertiodendran dewevrei in mbau forest. Most B. laurentii were inactive over the study area in 1981 but fruited synchronously and widely in 1982. Abundant seed-fall in the late wet season of 1982 extended into the succeeding dry season. This contributed to a nearly three-fold difference in dry season food abundance in mixed forest in 1983 than in the non-mast dry season of 1982 (Figure 4-3). Flowering and seed-fall of G. dewevrei in the two major stands of mbau forest on the study area were not synchronized. Simultaneous flowering occurred in the Eboyo mbau forest in 1981 but not in 1982. Flowering and seed production occurred in Mangbara in 1982 but not to any large extent in 1981. The flowering and fruiting of the caesalpiniaceous dominants created pulses of abundant food resources (Table 4-16). This was especially evident in the mbau forest where fallen fruits of other species were scarce. Total food abundance recorded on transects varied from no foods recorded on 0.7 km of transect druing the dry season of 1983. to almost 70 kg/km recorded during the G. dewevrei seed-fall of 1981 at Eboyo. High transect weights were also recorded during periods of G. dewevrei flowering and seed ripening when fallen flower bracts and unripe seeds dropped by foraging primates were common. Despite peaks in apparent food abundnace. however. overall food diversity in mbau forest was consistently low. corresponding with the low tree species diversity in this forest type. In the mixed forest, the seasons of mast seed—fall of B. Ienrentii were also periods of greatest food availability. Levels of 0.050— LOIOLQ B...- L-N. zinc“ >Um-L->.D cue—=20 s 5000000 0000 05.0 05.0 00.00 50..000 00.0 00.0 05.0 00 .000 00000000 ou.s 0n.0 00.o0 No 0000 00.90 no.0 00.0 0a .009 0000 00-0 00.0 00.0 «0.00 no .000 00.0 «0.0 0~.0 00 .000 00000000 0000 mu 00.0 00.0 00.0 50 .500 00.05 05.0 00.0 00 .00: 000000000 1 00500 00000000 00500 00000000 0000000 000» 000000000 0000000000 00 0000 000000000 0000000000 00 0000 00-000 0000: 000000 :00: 00000000 00 000000000 .000000 00000 00 000005000 000000000050 0000000 000- 00 00500000 0050000000500000 000000 00000000 .0000 .500 00 0000 0000 .00005 .000000 00000 000 00 00000 00000000000000.0u 00 00000 00000000 00000000 000 000000 0000000 0000 000 0000- 00 00000000» 00 00000000 000000: 0000» 000 000000000 0000-00 .00-. 0000» 149 food bionass recorded on transects, however. did not reach levels recorded in nbau forest during periods of G. dewevrei seed-fall. This was offset by a decreased disparity in food abundance between nast and non-nast seasons. The seasonally nore equitable food availability in nixed forest corresponds with the lower doninance of B. Jaurentu in nixed forest and the higher representation of other fruit bearing species in the canopy. Overall. nixed and nbau forest did not differ in the weight of fallen fruits, seeds and flowers on transects. The diversity of these potential foods, however, was significnatly greater in nixed forest than in nbau forest (t = 7.59, p < .001. n - 5). DISCUSSION Food Selection in Upland Duikers Dubost's (1984: 311) conclusion that duikers were unselective polyphagic foragers and that there "appears to be no najor deternining factor for the trophic differentiation of these frugivorous runinants." is not supported by the results of this study. Although the taxononic diversity in the diets of all ungulates for which there were adequate data was high. at least four species studied in depth exhibited narked patterns of selection along other dinensions of their food. In this study. food patch weight. food iten size and food nutritional quality were differntially selected by aninals of differing body size and nouth norphology. The snall species preferred foods of high nutritional quality and snaller iten size but selected equally foods of both large 150 and snall food patch weight. Two nediun-sized species and a large species selected foods over a range of relative nutritional quality. One of the niddle—sized species with a broad nouth preferred foods of both large size and generally large patch weight. The other nediun-sized species had a narrow nouth and included both large and snall food itens and a greater proportion of foods from snaller food patches in its diet. Along the species selected by the largest species were a nunber of the largest and toughest fruits. Differences in selectivity of the duikers were evident even when all species fed on the sane food species. Fruits with enbedded seeds were typically found in fallen Blighia welwitchii, Cola .lateritia and Landolphia app. and were eaten by all upland duiker species. Whereas the larger species, C. dorsalis characteristically ate both fruit and seed together, the snallest species, C. nonticola and the narrow-nouthed nediun species C. callipygus selected seeds fron the fruits andleft the capsules and rinds. In the case of Landolplua spp, and perhaps 0. lateri ta as well, the selectivity of C. nonticala apparently extended to finding and eating dispersed seeds defecated by prinates. perhaps even eating them fron their droppings. Selection for different parts of the sane food species has parallels in ungulate connunities of nore open environnents. Snaller species are reported to select for specific plant parts while larger species are restricted to foraging on large. but sonetines low quality swards (Bell 1971, Owen-Snith 1980. Bunnell and Gillinghan in press). Differences in foraging and diet reflect differing constraints 151 associated with the body size of the animal (Denment and Van Soest 1985). Snall aninals are required by higher relative netabolic needs to forage on nore readily digested, nutrient-rich foods. Larger species, though having lower relative needs and hence a greater capacity to digest poor quality food, are nevertheless constrained by large total food needs which nay not be net searching for snall, dispersed food itens. The relationship between body size and the relative ability to exploit diffuse versus concentrated food patches has energed as a cannon thene in the study of a nunber of size-distributed guilds of consuners, including seed eating heteronyid rodents (Brown 1975, Price 1981'. Harris 1984) and frugivorous prinates (Terborgh 1983). As Terborgh has shown, large species nay be unable to effectively exploit snall, scattered food patches, thus providing for an exclusive resource for snaller species. The results of the present study support this theory in denonstrating that the larger duikers preferred food patches with high food weights and avoided patches with low food weight. One problen with the patch use analysis reported here, however, is that patch weight is a conposite value incorporating neasures of both food item density and areal extent. Qualitatively very different food patches thus can have the sane total weight. Both Lewis (1980) and Schluter (1982) have suggested that the relative value of a food patch to a consuner is a function of the rate of resource harvest fron the patch. Patch value then is a function not only of the value of the individual itens, but how easily they can be 152 found and ingested. With the ungulates, food nutritonal quality is potentially an inportant deterninant of food patch value. Large dense patches of fruits nay be potentially easily harvested, but their value is decreased or conpletely negated if th fruits are unpalatable or poorly digested (Chapter 2 and 3). While fruits and seeds available to the ungulates are patchily distributed on the forest floor, it renains to be denonstrated what dinensions of these patches aninals attend to. In the case of the diurnal speices at least at sone seasons, the aninals nay not select specific food patches but rather accconpany nobile "patches" of prinates which generate food patches for the duikers as they nove through the canopy. Qiet Overlap and anunity Structure Species specific differences in food choice can not be related directly to species diversity in this ungulate connunity. Patterns of dietary convergence and divergence, however, do indicate that conpetiton for food was possible during periods when high quality food was scarce and a diversity of alternative foods was low. Recent studies of prinary consuners in tropical forests have denonstrated that diets of different species nay converge on abundant resources and diverge during periods when these foods are scarce Heithaus et al. 1975, Ennons 1980, Gautier-Hion 1980, Terborgh 1983, see also Flening 1979). In this study diets converged when high quality foods such as the seeds of B. launentii the fruits and seeds of Landolphia ssp or the unripe fruits of K. gabonensis were 153 abundantly available. Conpetition for food during these periods was unlikely as these high quality foods were evidently available in surplus and were recorded on the fruit transects. Dietary divergence occurred during periods when high quality foods preferred by all species were absent but a diversity of other foods renained available. During these periods, species-specific patterns of foraging and food choice were nost evident. In addition to the differences in body size and nouth norphology investigated in this study, there was evidence for other factors differentiating the ungulates. Two species utilized restricted water-side habitats. Differences in activity pattern nay have also played a role as diets of nocturnal species were on average 'nore diverse than those of diurnal species. There was also evidence for spatial segregation without apparent habitat separation in two species of sinilar body size. This latter case suggests sone forn of direct interspecific interference (see Terborgh 1971, Terborgh and Weske 1975 Dianond 1975 and Noon 1981 for descriptions and discussions of this phenonenon in birds). Other than this case, it was not clear that the other species were actively conpeting during periods when diets diverged. as the food resources available to sone species were not available to other species for reasons related to the constraints of body size, relative nobility and perhaps habitat and activity pattern. Periods during which both high quality foods were scarce and total food diversity was low were periods when conpetiton was nost likely. During the course of this study, the conbination of low food abundance and low food diversity was nost likely to occur during dry 154 season which were not preceeded by B. laurentii seed fall. These conditions apparently obtained during the dry season of 1982. During this period fruit levels on transects were very low. The diets of the three upland duiker species all contained unripe fruits of K. gabonensis as well as foliage and fungi. The diets of C. nonticola contained few fruits but were dominated by foliage and fungi. Both foliage and fungi were generally not preferred foods of any ungulate species. Both were available year round but rarely conprised nore than 5% of the diet. Foliage was only eaten in quanitiy by duikers in feeding trials when offered with low qulaity fruits (Chapter 3). A nunber of fungi, known to be eaten in the wild were rejected by captive duikers when offered with nore preferred unripe Klainedoxa gabonensis fruits (J. Hart, unpubl obs.). Fogel and Trappe (1978) and Blair et al. (1984) report fungi as inportant in the diets of a nunber of nannals including ungulates. The authors argued that fungi nay contain appreciable levels of protein and digestible carbohydrates on a dry weight basis and they nay represent a high quality food source. The evidence fron the diets of the forest ungulates does not support this view, however. Although fungi nay be nutritous on a dry natter basis, they routinely average 7096 and even up to 90% water (Fogel and Trappe 1978). As a result, their edible dry natter yield is low. Results of this study (see also Chapter 3) indicate that dry natter yield is an inportannt conponent of food quality for ungulates. Favored foods contain concentrated sources of nutrients. Low runen content weights and low kidney fat levels, 155 recorded during the period when fungi dominated C. nonticola diets is further evidence that this diet was suboptinal. There is circumstantial evidence that the blue duiker's linited diet during the 1982 dry season was due in part to conpetiton with red duikers. Unripe K. gabonensis fruits were preferred foods of blue duikers and were eaten freely during other sanple periods when they were abundant (late wet season, 1981). Captive blue duiker selected unripe K. gabonensis fruits during palatability trials (Chapter 3). Fruits of K. gabonensis were not recorded on transects during the dry season, however. Their occurrence in nost red duiker runens indicated that the larger aninals were better able than the blue duiker to find and exploit the evidently linited quantities which were available. Because both the diversity and abundance of renaining foods was low at this tine, all duikers, and the blue duiker in particular, were forced to broaden their intake to include foods which were otherwise generally ignored. The particular configuration of low abundance of favored foods and low diversity and abundance of alternatives was recorded only once in over two years observations in nixed forests on the study area. This conbination, however, nay characterize food availability nore frequently and for longer periods in nbau forest and nay be an inportant reason why upland duiker densities are overall lower in this forest type. The evidence from the diet studies indicated that low food diversity nay not be a problen to foraging duikers during periods when favored food were abundant. Limited food diversity, however, nay preclude divergence of diets during periods when high quality foods are 156 not available. Are Forest Unfilates Food—linited? Reduced abundance of upland duikers in nbau forest denonstrates the potential inportance of food availabilty as a deterninant of duiker abundance in nixed forest as well. If duikers are food linited, this presents the apparent paradox that nany species of fallen fruits are not eaten (see results above) and rot on the forest floor. No hypotheses can be presented to account for the apparently uneaten food. One is that duiker densities are controlled by periodic bottlenecks in food availability and populations can not track fluctuating resource levels. This hypothesis has been developed to explain patterns of species co-occurrence in acne tenperate avian connunities (Wiens 1977, 1984). It has not. however, been used to exanine tropical forest connunities (see however Leigh et al. 1982). A second hypothesis is that all that was recorded as potential food on the transects was not in fact really available to the ungulates. Fruits, seeds and flowers vary in nutritional quality and nay contain appreciable levels of toxins and digestion inhibitors. A nunber of fruits collected fron the forest floor were not preferred by captive duikers in palatability trials (Chapter 3) Furthernore, even the nost preferred species were not eaten to the exclusion of other foods. While the runinants with foregut fernentation can eat nany foods which were unpalatable or unused by other frugiovres in the forest, notably prinates, this capacity is not unlinited. 157 Further nore careful stuides deternining what can and cannot be eaten and factors deternining inclusion of a potential food in the diet will be needed before we can evaluate these two hypotheses. L I TERATURE C I TED APPENDICES 158 LITERATURE CITED Alexandre, D.Y. 1978. Le role disseninateur des elephants en forét de Tai, Cote d'Ivoire. Terre et Vie 32: 47-72. Bailey, R.C., and N.R. Peacock. in press. Efe pygneis of northeast Zaire: subsistence strategies in the Ituri Forest. In I. de Carine and G.A. Harrison, editors. Uncertainty in the food supply. Canbridge University Press. Bell, R.H.V. 1971. A grazer ecosysten in the Serengeti. Scientific Anerican 225: 86—93. Blair, R.M., R. Alcaney, and F. Hershelf. 1984. Yield, nutrient conposition and runinant diegestibility of fleshy fungi in southern forests. Journal of Wildlife Hanagenent 48: 1344-1352. Bourliére, F., and J. Verschuren. 1980. Introduction a l'écologie des ongulés du Parc National Albert. In Exploration du Parc National Albert, Mission F. Bourliére et J. Verschuren. Bruxelles. Brower, J.B., and J.T. Zar. 1984.‘ Field and laboratory nethods for general ecology. Second edition. Brown Publishers, Dubuque, Iowa. Brown, J.H. 1975. Geographical ecology of desert rodents. Pages 315-341, 1n.H.L. Cody and J.H. Dianond, editors. Ecology and evolution of connunities. Harvard University Press, Canbridge, MA, USA. Bultot, F. 1971. Atlas clinatique du bassin congolais, preniere partie. Publications I.N.B.A.C., hors eerie, Bruxelles. Bunnell, F.L.. and M.P. Gillinghan. in press. Foraging behavior: dynanics of dining out. In, R.J. Hudson and R.G. White, editors. Bioenergetics of wild herbivores. CRC Press, Roca Baton, Florida, USA. Chesson, J. 1983. The estination and analysis of preference and its relationship to foraging nodels. Ecology 84: 1297-1304. Clutton-Brock, T., and P. Harvey. 1983. The functional significance of variation in body size anong nannals. Pages 832-863 in, J. Eisenberg and D. Kleinan, editors. Advances in the study of nannalian behavior. Special Publication No. 7, Anerican Society of Hannalogists. Cock, H.J.W. 1978. The assessnent of preference. Journal of Animal Ecology 47: 805-818. 159 Dieterlen, F. 1978. Zur Phanologie des Equatorialen regenwaldes in ost-Zaire (Kivu). Dissertations Botanical Band 47. J. Carner Publishers. Dennent, H., and P.J. Van Soest. 1985. A nutritional explanation for body-size patterns of runinant and nonruninant herbivores. Anerican Naturalist 125: 641-672. Dianond, J. 1975. Assenbly of species connunities. Pages 342-444 in, H.L. Cody and J.H. Dianond, editors. Ecology and evolution of connunities. Harvard University Press, Canbridge, MA, USA. Dirschl, H.J. 1963. Seive nesh size in relation to analysis of antelope runen contents. Journal of Wildlife Hanagenent 26: 327-328. Dubost, G. 1975. Le conportenent du chevrotain africain Hyenoschus aquaticus’Ogilby (Artiodactyla, Runinatia). Zeitschrift fur Tierpscychologie 37: 403-501. Dubost, G. 1978. Un apercu sur l'écologie du chevrotain africain Eye-oschus aquaticus Ogilby. Hannalia 42: 1-62. Dubost, G. 1979. The size of African forest artiodactyls as deternined by the vegetation structure. African Journal of Ecology 17: 1-17. Dubost, G. 1980. L'écologie et la vie sociale du céphalophe bleu (Cepahlophus nonticola Thunberg) petit runinanat forestier africain. Zeitschrift fur Tierpsychologie 54: 205—266. Dubost, G. 1983. Le conportenent de cepahlophus nonticola Thunberg et Cbpahalophus dorsalis Gray, et la place des céphalophes au sein des runinants. Hannalia 47: 141-177 and 281-310. Dubost, G. 1984. Conparison of the diets of frugivorous forest runinants of Gabon. Journal of Hannalogy 85: 298—316. Ennons, L.H. 1980. Ecology and resource partitioning anong nine species of African rain forest squirrels. Ecological Monographs 50: 31-54. Ennons, L.H., A. Gautier-Hion, and G. Dubost. 1983. Connunity structure of the frugivorous—folivorous forest nannals of Gabon. Journal of Zoology, London 199: 209-222. Flening, T.H. 1979. Do tropical frugivores conpete for food? . Anerican Zoologist 19: 1157-1172. Fogel, R., and J.H. Trappe. 1978. Fugus consunption by snall aninals. Northwest Science 52: 1-31. 160 Freeland, W.J. and D.H. Janzen. 1974. Strategies in herbivory by nannals: the role of plant secondary conpounds. Anerican Naturalist 108: 289-289. Frankie, G.W., H.G. Baker, and P.A. Opler. 1974. Conparative phenological studies of trees in tropical wet and dry forests in the lowlands of Costa Rica. Journal of Ecology 82: 881-919. Gautier-Hion. A. 1980. Seasonal variations of diet related to species and sex in a connunity of cercopithecus nonkeys. Journal of Aninal Ecology 49: 237-289. Gautier-Hion, A., L.H. Ennons, and G. Dubost. 1980. A conparison of the diets of three najor groups of prinary consuners of Gabon (prinates, squirrels and runinants). Oecologia 45: 182-189. Goering, H.E. and P.J. van Soest. 1970. Forage fiber analyses (apparatus, reagents, procedures and.some applications). U.S. Department of Agriculture, Agricultural Handbook No. 379. Gautier-Hion, A. 1980. Seasonal variations of diet related to species and sex in a connunity of cercopithecus nonkeys. Journal of Aninal Ecology 49: 237-269. Gautier-Hion, A., L.H. Ennons, and G. Dubost. 1980. A conparison of the diets of three najor groups of prinary consuners of Gabon (prinates, squirrels and runinants). Oecologia 45: 182-189. Hart, T. B. 1985. 'lhe ecology of a single-species-datfinant forest and of a.ndxed forest in Zaire,.Africa. Unpublished PhD Dissertation,JMiChigan State university, East Lansing,.MI. Heithaus, R., T.H. Flening, and P.A. Opler. 1975. Foraging patterns and resource utilization in seven species of bats in a seasonal tropical forest. Ecology 58: 841-854. Herrera, C. 1981. Are tropical fruits nore rewarding to dispersers than tenperate ones? Anerican Naturalist 118: 896—907. Hoffnann, R. R. 1973. The runinant stonach. East African nonographs in biology 2. East Afrcian Literature Bureau, Nairobi, Kenya. Horn, H.S. 1986. Heasurenent of "overlap" in conparative ecological studies. Anerican Naturalist 100: 419-424. Horowitz, (editor). 1970. Official nethods of analysis of the association of official analytical chenists. 12th edition. Association of Official Analytical Chenists, Washingotn, D.C.. USA. 161 Hurlburt, S.H. 1978. The measurement of nich overlap and some relatives. Ecology 59: 87-77. Ivlev, v.s. 1981. Experimental ecology of the feeding of fishes. Yale University Press, New Haven Connetcticut, USA. Johnson, D.H. 1980. The comparison of usage and availability neasurenents for evaluating resource preference. Ecology 61: 85-71. Kingdon, J. 1982. Duikers and Cephalophini. East African mammals: an atlas of evolution in Africa III (C) (Bovids): 283-279. Academic Press, New York, USA. Lebrun, J. and G. Gilbert. 1954. Due classification écologique des forets du congo. I.N.E.A.C. Série Scientifique, number 63. Bruxelles. Leigh, E.G., A.S. Rand and D.H. Windsor. Ecology of a tropical forest. seasonal rhythms and long-term changes. Smithsonian Institution Press, Washington, D.C.. USA. Leighton, M, and D. Leighton. 1982. The relationship of size of feeding aggregate to size of food patch: howler nonkeys (Alouatta palliata) feeding in Trichilia Jipo fruit trees on Barro Colorado Island. Biotropica 14: 81-90. Lewis, A.R. 1980. Patch use in gray squirrels and optimal foraging. Ecology 81: 1371-1379. Manley, B.F.J., Miller, P, and L.M. Cook. 1972. Analysis of selective predation experiments. American Naturalist 106: 719-736. Morisita, M. 1959. Measuring interspecific association and similarity between communities. Menoires of the Faculty of Science, Kyushu University Series E (Biology) 3: 65-80. Noon, B. 1981. The distribution of an avian guild along a temperate elevational gradient: the importance and expression of competiton. Ecological Monographs 51: 105-124. Oates, J. 1977. The guereza and its food. Pages 278-321 inT.H. Clutton-Brock, editor. Primate ecology: studies of feeding and ranging behavior in lemurs, monkeys and apes. Academic Press, London, UK. Oates, J.F., P.G. Waterman, and G.M. Choo. 1980. Food selection by the south Indian leaf-monkey Presbytis johnsii in relation to leaf chemistry. Oecologia 45: 45-58. Owaga, M.L.A. 1978. The effect of sieve mesh size on analysis of rumen contents. Journal of Wildlife Management 42: 893-697. 162 Owen-Smith, N. 1980. Factors influencing the transfer of plant productis into large herbivore populations. Pages 359-404 in B.J. Huntley, and B.H. Walker, editors, Dynamic changes in savanna ecosystems. CISRO, Pretoria, South Africa. Petrides, G.A. 1975. Principle foods versus preferred foods and their relation to stocking rate and range condition. Biological Conservation 7: 181-169. Price, M. 1984. Ecological consequences of body size: a model for patch choice in desert rodents. Oecologia (Berlin) 59: 384-392. Puglisi, H.J., S.A. Liscinsky, and R.B. Harlow. 1971. An improved methodology of rumen content analysis for white-tailed deer. Journal of Wildlife Management 42: 397-403. Schluter, D. 1982. Seed and patch selection by Galapagos ground finches: relation to foraging efficiecy and food supply. Ecology 83: 1108-1120. Siegel, S. 1958. Nonparametric statistics for the behavioral sciences. McGraw Hill. New York, New York, USA. Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of statistics, second edition. McGraw-Hill, New York, New York, USA. Stephenson, A.G. 1981. Flower and fruit abortion: proximate causes and ultimate functions. Annual Review of Ecology and Systematics 12: 253-279. Terborgh, J. 1971. Distribution on environmental gradients: theory and a preliminary interpretation of distributional patterns in the avifauna of the Cordillera Vilcabamba, Peru. Ecology 51: 23-40. Terborgh, J. 1983. Five New World prinates: a study in comparative ecology. Princeton University Press, Princeton, New Jersey, USA, 260 pp. Terborgh, J., and B. Weske. 1975. The role of competition in the distribution of Andean birds. Ecology 56: 582-578. van der Pijl, L. 1972. Principles of dispersal in higher plants, second edition. Springer-Verlag, Berlin. Walther, H. 1973. Vegetation of the earth in relation to climate and the eco-physiological conditions, second edition (translated, J. Wieser). The English Universities Press, Ltd., Springer-Verlag, London. Wiens, J. 1977. On competiton and variable environnents. American 163 Scientist 65: 590-597. Wiens, J. 1984. On understanding a non-equilibrium world: myth and reality in community patterns and processes. Pages 439-457 in D.R. Strong, D. Simberloff, L.G. Abele, and A.B. Thistle, editors. Princeton University Press, Princeton, N.J., USA. 164 APPENDIX 4-A Locations and seasonal distributions of drive hunts which were acconpanied to sanple food availability and collect runen contents. Ituri Forest, laire, 1981 to 1983. Sample Total Forest nunber Hap Date Seasonb Days Drives transect types Location' (kn) 1 K Jun 81 H 4 21 12.3 nbau, mixed 2 8 Jul 81 H 9 44 25.4 nixed 3 K Sep 81 L 4 28 8.2 nbau, nixed 4 E Oct 81 L 4 22 8.3 nbau, nixed 5 H Dec 81 L 4 19 11.5 nixed b 8 Jan 82 D 5 23 14.6 nixed 7 A Her 82 E 4 24 9.8 nixed 8 K Hay 82 E 5 34 10.0 nbau, nixed 9 P Aug 82 H 5 25 11.8 nixed 10 8 Oct 82 L 5 21 10.1 nixed 11 K Feb 83 D 5 31 8.5 nbau, nixed 12 S Feb-har 83 D/E 5 32 5.7 nixed 13 1 Hay 93 E 3 17 5.0 nixed ‘ Locations shown on Figure 4-1. 8 Season abbreviations: D, dry season; E, early wet season; H, nid wet season; L., late wet season. 165 RPPENDIX 4-8 Bharacterisitcs of selected and avoided {ood species in diets of {our species oi upland duikers during nine sanpling periods between Deceeber, 1981 and Hay, 1983, lturi Forest, Zaire. ‘ Food Patch iten Selection Species type Y0 Weight Size Ce.- CIlI COCO Old. Decenber, 1981 Klainedoxa gabonensis UP 62 L 2-4 + + + K. gabonensi: RF 33 L 5 - - - Ricinodendron heudelotii RF 112 L 4 + + + Alstonia boonei UF 8 5 + + Thonningia sanguinea Fl 53 S 4 + + Ficus 196 RF 33 L 3 - - + - Croton nubanga R/UF 73 8 3 + + + Unknown 192 UF 3 + Uapaca guineensis Fl L 3 - - - - U. guineensis UF 10 S 3 - - - Ananjenje R/UF 3 - - - Unknown 205 8 8 3 + Annonaceae sp 1 8 100 S 3 + January, 1982 00111: adoifi-fridericii RF 236 S 3 + Klainedoxa gabonensis UP 62 S 4 + + + Kiaiaedoxa trillesii RF 13 L 4 - - - - Foliage 8 + + + Fungi 8 2-4 + + + Thonningia sanguinea Fl 59 S 4 + + Anthocleista schwainfurthii UF 4 + Ficus 220 R/UF 41 L 5 + Unknown 219 UP 3 Coabretua 3p UF 8 4 - - - - Albizzia guaaifera UF 8 5 - - - - Canthiun 5p 8 S 4 - - Appendix 4-8. Continued 166 Food Patch lten Selection Species type Y0 Weight Size C.s. 0.1. C.c. C.d. Harch, 1982 Kleinedoxa gabonansis UF 43 L 5 - - + + Irvingia woeboiu UF 61 L 5 + Ricinodendron heudelotii UP 94 L 4 + Ficus 4 RF 41 L 5 + Blighia weiwitschii UF/S 178 L 4 + Kusanga cercropioides UP 30 L 5 + + Tarenna iaurentii UP 44 L 5 + Pancovia harssiana UF/S S 3 + + + + Drypetes sp Fl 5 2 0 Kokou UF 13 S 4 - - - - Uapaca guineensis UF 10 S 3 - - - - Celtis adolfi-fridericii UF S 2 - - - - hay, 1982 Pancovia haresiana S S 3 + - no data Klainedoxa gabonensis RF 33 L 5 - - “ " Cleistanthus nichaisonii UF 42 L 3 - - u " Celtis adolfi-fridericii R/UF L 3 - - “ “ Gilbertiodendron dewevrei F1 32 L 2 - - “ “ Unknown Sapindaceae RF/S 4 + n H Unknown 242 RF/S 3 + u " Ricinodendron heudeiotii HF 92 L 4 + n " lrvingia wonboiu UP 67 L 4 + u “ Diospyros crassifiora F1 86 L 4 + n u Appendix 4-8. Continued 167 Food Patch Iten Selection Species type Y0 Weight Size C.n. 8.1. C.c. C.d. August 1982 Croton aubanga RF 73 S 3 + + + + Cola iateritia S 72 S 3 + + C. lateritia RF 40 S-L 4 + Blighia welwitschii RF/S 178 L 3 + + + + Klainedoxa gabonensis RF 33 L 5 + Brachystegia laurentii UF S 4 - - - 8. laurentii Fl 50 S 2 - - - - Landoiphia spp UF/S 48 L 5 - + + + Ricinodendron heudeiotii RF 112 L 4 - - + Unknown 271 UP 3 Kokou RF 13 S 4 - - - - Koroso UF 8 4 - - - Uapaca guineensis RF 10 L 3 - - - - Pancovia haresiana S S 3 Dasylepsis seretii UF 8 3 - - - Tarenna laurentii UF 44 L 5 - - - - Nauclea xanthoxylon UF 3 + October, 1982 Bracystegia Iaurentii S 317 L 3 + - + - Gilbertiodendron dewevrei S 120 L 5 + + + Landolphia spp S 101 L 3 + + Landoiphia spp RF 47 L 5 + + Irvingia wonbolu UF 67 L 5 Kiainedoxa trillesii RF 13 L 4 - - - Kleinedoxa pabonensis RF 33 L 5 - - - - Coabreiun sp UF 8 5 - - - - Kokou RF 13 S 4 - - - - Duboscia viridifolia RF L 4 - - - - Canariua schweinfurthii UP 32 3 + Ricinodendron heudelotii RF 112 L 4 - - - - Unknown 6 RF 3 + Chrsophyilua sp RF L 4 - - - Unknown 272 S S 3 + Unknown 195 RF/S 3 + 168 Appendix 4-8. Continued Food Patch iten Selection Species type Y0 Height Size Co‘s Cele CeCe Cede February, 1983 Brachystegia laurentii S 317 L 3 + + no data Klainedoxa gabonensis UF 62 L 3 + a K. gabonansis RF 33 L 5 - ” “ Kleinedoxa triilesii RF 13 L 4 - - " “ Annonaceae sp 1 S 90 S 3 + u . Doneiia pruniforais RF 45 L 5 + u “ Tarenna laurentii UF 41 L 5 - - “ “ nusanga cecropioides UF 33 L 5 - - “ ' February, Harch, i983 Bracyhstagia laurentii S 317 S 3 + - + - 8. laurentii Fl 50 S 2 + Cieistanthus nichelsonii UF 124 L 3 + - Kleinedoxa gabonensis RF 33 L 5 - + Kleinedoxa triliesii UF 16 L 4 - + - Chlorophora excelsa UF 200 L 4 + + Biighia welwitschii UF/S 178 L 4 + + Irvingia grandifoiia UF 231 L 4 + + Ricinodendron heudeiotii UP 92 L 4 + Dilbertiodendron dewevrei Fl 32 L 2 - - + - Tetracarpidiua canophorue UF 4 + Tarenna laurentii UF 44 L 5 - - + Diospyros crassifiora Fl 86 L 2 Uapaca guineensis UF 10 L 3 - - - Ficus 2 RF 30 S 2 - - - Syzgiua staudtii RF/S 3 Ficus 4 RF 44 L 4 + Ficus 3 RF 33 3 + Husanga cecropioides UF 30 L 5 + 169 Appendix 4-8. Continued Food Patch iten Selection Species type Y0 Height Size C.n. 0.1. E.C. C.d. Hay, 1983 Dacryoides eduiis UF 55 S 3 + Anthocleista schweinfurthii UF 4 + Cola lateritia UP 40 L 4 + Lipasa UF/S 31 L 4 + + + Phyllanthus pynaertii UF 16 L 4 Klainedoxa gabonensis RF 33 L 5 - - - + Ficus 4 UF 41 L 4 + lrvingia grandifoiia UF 231 L 4 + Blighia welwitschii UF 178 L 4 + Brachystegia laurentii Fl .50 L 2 - - - - Grewia oiigoneura UF 45 L 3 - - - - Kokou UP 13 S 4 - - - - Canthiua sp 8 S 4 + Unknown 234 UF 3 + Diospyros crassiflora Fl 86 L 2 + Unknown 279 UR 3 + Canariun UF 32 L 4 - - - - schwcinfurthii UP 32 L 4 - - - - Unknown Sapotaceae UF L 4 - - - - Diospyros sp RF 4 + Unknown legune S S 4 + ‘ Notes and abbreviations. Food type: UF, unripe fruit; RF, ripe fruit; S, ripe/unripe seed; Fl, flower. Y0 equals adjusted dry natter yield: Y0 0 Y N/(ADF + Cl) where Y I edible dry natter density N I nitrogen content (ng/g wet weight) ADF 0 acid detergent fiber content (ng/g wet weight) CT 0 condensed tannin content (eg/g wet weight). Where no values are shown, Yo was not deternined. 170 Appendix 4-8. Continued Notes and abbreviations continued. Patch weight: 8, small Food patch weight (W < 100 g); L, large Food patch weight (W Z 100 g). Where no synbol is shown, Food patch weight was not known. It.” '12.. 2g<10°‘ 3, 10° - 2.51 4. 2.5 - 500' 5, 5.0 - 10.0 centineters. Selection: Duiker species synbolsi C.n., C. aonticola; C.l., C. leucogaster; C.c., C. callipygus; C.d., C. dorsalis. + indicates selected Food species) - indicates avoided Food species) no synbol indicates use in proporiton to availability and/or Food rare in Forest and ninor in diet, selection could not be assessed.