Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Resource limitation and population ecology of white-eared kob Fryxell, John M. 1985

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1985_A1 F79.pdf [ 6.96MB ]
Metadata
JSON: 831-1.0096671.json
JSON-LD: 831-1.0096671-ld.json
RDF/XML (Pretty): 831-1.0096671-rdf.xml
RDF/JSON: 831-1.0096671-rdf.json
Turtle: 831-1.0096671-turtle.txt
N-Triples: 831-1.0096671-rdf-ntriples.txt
Original Record: 831-1.0096671-source.json
Full Text
831-1.0096671-fulltext.txt
Citation
831-1.0096671.ris

Full Text

RESOURCE LIMITATION AND POPULATION ECOLOGY OF WHITE-EARED KOB by JOHN M. FRYXELL B.Sc. (hon), University of B r i t i s h Columbia, Vancouver, 1978 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept th i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA 1985 @ @ John M. F r y x e l l , In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6(3/81) i i ABSTRACT In this study I • examine the e f f e c t of seasonal resource l i m i t a t i o n on the behavior and population dynamics of white-eared kob, Kobus kob l e u c o t i s , in the Boma region of the southern Sudan. This population, numbering over 800,000, migrates seasonally between savannah grasslands in areas with low r a i n f a l l and ephemerally swamped grasslands in areas with high r a i n f a l l . The aims of the study were: (1) to examine whether kob migration tracks ephemeral d i s t r i b u t i o n s of food or water resources, (2) to test the hypothesis that the Boma kob population is limited by food a v a i l a b i l i t y , (3) to determine i f c a l f production is cued to seasonal peaks in food abundance, and (4) to evaluate the e f f e c t of breeding synchrony on lekking behavior and male competition. Seasonal climatic changes produced pronounced changes in the d i s t r i b u t i o n and abundance of both green forage and water supplies. Dry season migration primarily tracked limited supplies of water. Within the dry season range, kob aggregated at high densities (over 1,000 per km2) in low-lying meadows that supported grass re-growth when l i t t l e green grass was available elsewhere in the ecosystem. However, southerly movements in the' wet season were not explainable by the resource hypothesis, since both food and water were widely d i s t r i b u t e d during the wet season. I suggest that kob may move southward in order to avoid surface flooding during the wet season. Kob mortality during the dry seasons of 1982 and 1983 was considerably higher than estimated mortality during the wet season. Unusual r a i n f a l l during the dry season of 1982 provided a "natural experiment" to test the food l i m i t a t i o n hypothesis. Adult mortality was s i g n i f i c a n t l y lower during the dry season of 1982 than during the more t y p i c a l dry season of 1983. Calf mortality did not vary s i g n i f i c a n t l y between years. Adult mortality rates were related to dry season duration. Dry season mortality was related to sub-maintenance food intake and declining fat reserves. The age structure of the kob population in 1983 suggests that large-scale mortality (ca. 40%) occurred in the 1980 drought that immediately preceded t h i s study. These findings support the food l i m i t a t i o n hypothesis. Kob exhibited a 4 month period of c a l f production during the late wet season, when food a v a i l a b i l i t y was highest. As a consequence, females continued l a c t a t i o n through the dry season period of food s c a r c i t y . I suggest that kob reproductive phenology may result from an obligatory delay during which females restore their fat reserves prior to calving or selection pressures imposed by predation during the vulnerable post-partum period. Synchronous breeding in the Boma kob was related to increased rates of aggression between males and increased color dimorphism, in comparison to the asynchronous breeding Uganda kob, Kobus kob thomasi. Male aggression served not only to e s t a b l i s h dominance relations between males on leks, but also disrupted the mating a c t i v i t i e s of neighboring males. Young adult males suffered higher age-specific mortality than females, possibly r e s u l t i n g from i n j u r i e s incurred during strenuous fi g h t i n g on leks. In order to analyze the age structure of the kob population, I devised a new method for estimating age-specific mortality rates that i s free of the r e s t r i c t i v e assumptions that underlie most conventional techniques. The proposed method has somewhat greater sampling v a r i a t i o n , but is considerably more robust, than two current methods. Moreover, the proposed method permits c a l c u l a t i o n of age-specific mortality at frequent inter v a l s during periods of population fluctuation and, under some circumstances, population numerical trends may be d i r e c t l y determined from age structure. TABLE OF CONTENTS ABSTRACT i i LIST OF TABLES v i LIST OF FIGURES v i i ACKNOWLEDGEMENTS - ix CHAPTER 1 . GENERAL INTRODUCTION 1 CHAPTER 2. SEASONAL MIGRATION IN RELATION TO RESOURCES 8 Introduction 8 Methods 10 Ae r i a l surveys 10 Ground Observations 13 The study area 17 Physical features and vegetation types 17 Climatic seasonality 20 Results 24 Ae r i a l survey results 24 Seasonal changes in green biomass 24 Seasonal changes in water a v a i l a b i l i t y .... 27 White-eared kob population numbers 29 Seasonal migration of white-eared kob 29 Dry season ground observations 35 Kob d i s t r i b u t i o n patterns 35 Factors affecting forage abundance 45 Feeding selection for plant parts 49 Food a v a i l a b i l i t y r e l a t i v e to kob requirements 53 Discussion 55 Seasonal changes in resource d i s t r i b u t i o n 55 Seasonal kob migration 56 Benefits of migration 59 CHAPTER 3. FOOD LIMITATION AND KOB MORTALITY PATTERNS 61 Introduction 61 Methods 62 Population density 62 Carcass density 63 Sex and age d i s t r i b u t i o n 65 Tooth cementum lines 66 Post-mortem examination 67 Results 68 Age d i s t r i b u t i o n in the l i v e population 68 Age d i s t r i b u t i o n at death 68 Age-specific mortality 72 Sex r a t i o 72 Seasonal changes in body condition 76 Seasonal changes in mortality rates 79 Dry season mortality in r e l a t i o n to food intake 82 Cumulative dry season mortality 85 Discussion 89 Causes of mortality during the dry season 89 Predation 89 Disease 92 Age structure of the kob population 93 Evidence for food l i m i t a t i o n 96 V CHAPTER 4. KOB REPRODUCTIVE PHENOLOGY 100 Introduction 100 Methods 101 Results 103 Discussion 111 CHAPTER 5. BREEDING SYNCHRONY AND MALE AGGRESSION 114 Introduction 114 Methods 119 Sexual dimorphism 119 Lek observations 119 Results 121 Sexual dimorphism 121 Temporal change in mating and agonistic behavior ....121 Spatial d i s t r i b u t i o n of females and agonistic encounters 124 Functions of fightin g behavior 128 Consequences of fightin g 131 Discussion 133 CHAPTER 6. AGE-SPECIFIC MORTALITY: AN ALTERNATIVE APPROACH 137 Introduction 137 The model 138 Methods 140 Calculation of age-specific mortality 141 Sampling d i s t r i b u t i o n s 144 Effects of an unstable age d i s t r i b u t i o n 146 Results 147 Discussion 151 Advantages of the proposed method 151 Comparisons between methods 154 CHAPTER 7. GENERAL DISCUSSION 158 Food l i m i t a t i o n 158 Seasonal migration 160 Breeding phenology 162 Breeding synchrony and male aggression 163 Age-specific mortality patterns 164 General conclusions 165 LITERATURE CITED 168 APPENDIX 1. PLANT SPECIES COLLECTED 183 LIST OF TABLES Table 2.1 Akobo monthly climatic normals 21 Table 2.2 Kob population estimates 1980-1983 30 Table 2.3 Dry season grass biomass estimates 39 Table 2.4 N u t r i t i o n a l analyses of meadow vs. t a l l grasses. 41 Table 2.5 Feeding selection for leaf vs. stem tissue 52 Table 3.1 Dry season weekly mortality estimates 80 Table 3.2 Mortality in r e l a t i o n to food intake 84 Table 5.1 Changes in lek behavior over the dry season 123 Table 5.2 Female movements following disruption 132 Table 6.1 Data used to calculate q x using method C 143 Table 6.2 Skewness and kurtosis of sampling d i s t r i b u t i o n s obtained using 3 alternative methods 149 v i i LIST OF FIGURES Figure 1.1 Map of the southeast Sudan 4 Figure 1.2 Lines of investigation 7 Figure 2.1 Calibration of photometer readings 14 Figure 2.2 Map of the study area 15 Figure 2.3 Map of woody vegetation cover 19 Figure 2.4 Mean monthly r a i n f a l l t o t a l s 22 Figure 2.5 North-south gradient in seasonal green biomass abundance 25 Figure 2.6 Seasonal d i s t r i b u t i o n of green biomass 26 Figure 2.7 Seasonal d i s t r i b u t i o n of water supplies. 28 Figure 2.8 Seasonal d i s t r i b u t i o n of kob 31 Figure 2.9 Kob d i s t r i b u t i o n in r e l a t i o n to nearest water. . 33 Figure 2.10 Kob dry season concentrations 36 Figure 2.11 Population densities in woodland vs. meadow habitats 37 Figure 2.12 Kob diurnal movements onto meadows 38 Figure 2.13 Dry season meadow grass growth 42 Figure 2.14 Dry season leaf and stem tissue growth 43 Figure 2.15 Kob aggregation around meadows 46 Figure 2.16 Kob abandonment of meadows 47 Figure 2.17 Dry season grass growth in r e l a t i o n to s o i l moisture 48 Figure 2.18 S o i l moisture decline over the dry season 50 Figure 2.19 Feeding concentration near water 51 Figure 3.1 Age structure of the l i v e population 69 Figure 3.2 Age structure of found carcass sample 70 Figure 3.3 Age frequencies of carcasses by sex 71 Figure 3.4 Age-specific mortality curve 73 Figure 3.5 Carcass freqencies in r e l a t i o n to distance from observers 75 Figure 3.6 Dry season decline in fat reserves 77 Figure 3.7 Dry season adult mortality rates (Ajwara) 81 Figure 3.8 Dry season c a l f mortality rates (Ajwara) 83 Figure 3.9 Adult mortality as a function of dry season duration 86 Figure 3.10 Graphical explanation of the body condition/mortality hypothesis 88 Figure 3.11 Cumulative mortality during a drought in Nairobi National Park 90 Figure 3.12 Kob population estimates 1979-1983 95 Figure 4.1 Frequencies of births by month 1.04 Figure 4.2 Female reproductive condition January to May. ..105 Figure 4.3 Dry season body condition of l a c t a t i n g vs. non-l a c t a t i n g females 107 Figure 4.4 A c t i v i t y patterns of males vs. females 108 Figure 4.5 Seasonal changes in calf/female r a t i o 109 Figure 5.1 Sexual selection in r e l a t i o n to breeding synchrony 116 Figure 5.2 Hypothetical cost-benefit relations of male fighting intensity 118 Figure 5.3 Kob l i v e weights at age 122 v i i i Figure 5.4 Changes in agonistic behavior rates over the breeding season 125 Figure 5.5 Female s p a t i a l d i s t r i b u t i o n on leks 126 Figure 5.6 Female group sizes with single males 127 Figure 5.7 S p a t i a l d i s t r i b u t i o n of fights between males. ..129 Figure 6.1 African buffalo age-specific mortality curves ..142 Figure 6.2 Sampling variation r e s u l t i n g from the use of 3 alte r n a t i v e methods 148 Figure 6.3 Age-specific mortality estimated from an unstable age d i s t r i b u t i o n 150 Figure 6.4 Sampling variation as a function of denominator value 156 ix ACKNOWLEDGEMENTS A number of people assisted in the f i e l d work reported in this study, frequently under trying circumstances: Alpayo Dani, Michael Earle, Dana F r y x e l l , Beko Ladu, Charles Loring, Ipote Luke, Yusif Mohammed, and Baba Terkumpte. Steve Cobb, Conrad Eveling, Tim Fyson, Jens Hessel, Chris Hillman, Karen Ross, P h i l Snyder, and Tim Tear assisted ably in a e r i a l surveys. In addition, a number of individuals and organisations provided invaluable l o g i s t i c support: Barbie Alle n , the Arensons, L i z Davis, Peter Garment, GTZ, the Haspels, Interfreight Ltd., Mike Norton-Griffiths, the Juba Boatyard, P h i l l i p Winter, John Olander and the mujanim. In their own ways, a l l contributed to many memorable adventures. I thank my thesis committee, Ray Hilborn, Charley Krebs, Peter Murtha, Mike P i t t , Tony S i n c l a i r , and Jamie Smith for their unhesitating attention and high-quality c r i t i c i s m . In addition to my thesis committee, Lee Gass, John Eadie, Don Ludwig, John Greever, Dave Tai t , B i l l N e i l l , and Dolph Schluter made valuable comments on e a r l i e r drafts of various parts of the thesis. Fellow graduate students, s t a f f , and faculty of the Instit u t e of Animal Resource Ecology provided a simultaneously c r i t i c a l and cooperative atmosphere that contributed greatly to the development of these ideas. This project was funded in large part by the New York Zoological Society. Additional f i n a n c i a l support was provided by the Natural Sciences and Engineering Research Council (Canada), the National Geographic Society, the Frankfurt Zoological Society, and the Charles A. Lindbergh Fund. The Ministry of W i l d l i f e Conservation of The Democratic Republic of the Sudan kindly provided permission to work in Boma National Park and assisted in obtaining other necessary documents as well. I am grateful to a l l for their assistance. F i n a l l y , I would l i k e to give special thanks to 4 people: Tony S i n c l a i r , for numerous deeds beyond the normal c a l l of duty; Chris Hillman, for his companionship, sage advice, and innumerable favors; Shirley F r y x e l l , for her steadfast support over the years; and Sue Pennant, for enduring even my wildest f o l l i e s . 1 CHAPTER 1. GENERAL INTRODUCTION Assumptions of food resource l i m i t a t i o n underlie many fundamental hypotheses in large-mammal ecology. Limited resources are commonly presumed to constrain foraging adaptations, habitat s u i t a b i l i t y , and s o c i a l behavior of individuals (Geist 1974; Jarman 1974; Owen-Smith 1982). Resources may determine population d e n s i t i e s through density-dependent effects on recruitment, mortality, or dispersal (Bobek 1977; S i n c l a i r 1977; McCullough 1979; Fowler 1981). F i n a l l y , resource scarcity may structure communities through niche p a r t i t i o n i n g or f a c i l i t a t i o n (Vesey-Fitzgerald 1960; Lamprey 1963; B e l l 1971; Jarman and S i n c l a i r 1979). Direct evidence of food l i m i t a t i o n has been documented for roe deer (Bobek 1977), African buffalo ( S i n c l a i r 1977), kangaroo (Bayliss 1985), and wildebeest ( S i n c l a i r et a l . 1985) populations. Also, there i s considerable circumstantial evidence that large-mammal populations are limited by food a v a i l a b i l i t y . Klein (1968) documented the rapid increase and catastrophic die-off of reindeer introduced onto St. Matthew Island, which Klein concluded was caused by the overgrazing of available vegetation. Caughley (1970) reviewed evidence for similar ungulate eruptions, which Caughley and Lawton (1981) later explained using a theoreti c a l model based on trophic interactions between herbivores and vegetation. Coe et a l . (1976) demonstrated that large-herbivore biomass in African game reserves was p o s i t i v e l y correlated with annual r a i n f a l l and, 2 presumably, vegetation abundance. S i n c l a i r (1977) found a similar positive c o r r e l a t i o n between African buffalo population density and annual r a i n f a l l . Leader-Williams and Ricketts (1980) and Clarke and Henderson (1981) found demographic evidence suggesting food l i m i t a t i o n in reindeer and chamois. In addition, numerous long-term studies indicate that many large-mammal populations exhibit density-dependent responses in fecundity and age at f i r s t reproduction (Woodgerd 1964; Gross 1969; Geist 1971; Fowler and Smith 1973; Gambell 1975; Lett et a l . 1981; McCullough 1979; Fowler 1981; Clutton-Brock et a l . 1982), juvenile survival (Grubb 1974; Lett et a l . 1981; McCullough 1979; Clutton-Brock et a l . 1982; Houston 1982), and adult survival (Grubb 1974; S i n c l a i r 1977; S i n c l a i r et a l . 1985). Many of these authors speculated that food a v a i l a b i l i t y was ultimately responsible for the density-dependent response. However, there is also empirical evidence that some large-herbivore populations are l i m i t e d by predators or disease. Gasaway et a l . (1983), Mech and Karns (1977), and Messier and Crete (1985) suggested that juvenile survival in moose and white-tailed deer was related to wolf population density and that predators could maintain prey populations at low levels for which food was not l i m i t i n g . Smuts (1978) suggested that l i o n predation limited wildebeest population numbers below the carrying capacity set by vegetation abundance. Caughley et a l . (1980) argued that dingo predation may "control" some kangaroo and emu populations in A u s t r a l i a . Berry (1981) argued that combined ef f e c t s of disease and 3 predation limited wildebeest population numbers in a Botswana game reserve. Christian et a l . (1960) suggested that density-dependent stress disease caused the catastrophic mortality of a herd of introduced Sika deer. F i n a l l y , results from Serengeti suggest that an introduced disease, rinderpest, limited buffalo and wildbeest populations at low numbers prior to eradication of the disease in the early 1960's ( S i n c l a i r 1977; S i n c l a i r and Norton-Griffiths 1979). Thus, while there i s evidence that at least some natural populations of large herbivores are limited by food a v a i l a b i l i t y , t h i s is by no means universally true. In this study, I investigate whether a large migratory population of an African antelope, the white-eared kob (Kobus kob leuc o t i s , Lichtenstein and Peters, 1854), is limited by food abundance and examine the eff e c t s of food l i m i t a t i o n on kob movement patterns, breeding phenology, and mating system. The central theme of my study is that food resource l i m i t a t i o n determines many of the most important l i f e history c h a r a c t e r i s t i c s of herbivores. The white-eared kob population under study numbers about 830,000 in the Boma National Park region of the southern Sudan (Fig. 1.1). The Boma ecosystem is composed of broad expanses of t r o p i c a l savannah grasslands punctuated by a patchy cover of woody vegetation. Like most African savannahs, the Boma grasslands are subject to seasonal extremes in r a i n f a l l ; mean monthly r a i n f a l l ranges from less than 20mm to over 150mm. This extreme v a r i a t i o n in monthly r a i n f a l l causes reduced food abundance during the dry season, from January to A p r i l , and at 4 Figure 1.1 Map of the S.E. Sudan. Watercourses and i n t e r n a t i o n a l boundaries indicated by s o l i d l i n e s , roads indicated by broken l i n e s , and townships indicated by boxes. 5 this time food may be l i m i t i n g for the kob. In Chapter 2, I describe seasonal climatic changes in the Boma ecosystem and consequent effects on the d i s t r i b u t i o n and abundance of green grass and water supplies. Having shown that resource d i s t r i b u t i o n changes dramatically throughout the year, I test whether kob migratory movements are responses to temporal changes in the d i s t r i b u t i o n of scarce supplies of water (Western 1975) or green forage (Pennycuick 1975). I examine movement patterns at three d i f f e r e n t scales: 1) annual migration of the entire kob population between wet and dry season ranges 200 km apart, 2) dry season movement patterns between d i f f e r e n t habitat types, and 3) foraging movements of individuals within s p e c i f i c dry season habitats. In Chapter 3,, I investigate the effect of dry season food scarcity on kob demographic patterns, p a r t i c u l a r l y mortality. I test the hypothesis that dry season food abundance determines kob population numbers through n u t r i t i o n - r e l a t e d mortality. In addition, I examine the impact of large-scale mortality that occurred at the outset of the study by evaluating the current age structure of the population. F i n a l l y , I suggest an empirical model of kob population mortality as a function of dry season duration. I examine the phenology of c a l f production in Chapter 4. I use these data to test the hypothesis (Sadleir 1969; S i n c l a i r 1983a) that production of young in mammals should be synchronized to periods of peak food abundance. If food abundance i s l i m i t i n g during the dry season, th i s hypothesis 6 predicts that kob should produce young during the wet season. In Chapter 5, I investigate the effect of breeding synchrony (resulting presumably from adaptation to seasonal changes in food abundance) on the mating system of the white-eared kob. The Boma population has a lek system (Bradbury 1981; Bradbury and Gibson 1983), with a presumably high degree of polygyny. I test Emlen and Oring's (1977) hypothesis that moderate breeding synchrony should increase the degree of polygyny and competition between males for females, by comparing published data on the conspecific Uganda kob (Kobus kob thomasi) to o r i g i n a l data for the white-eared kob. In order to evaluate the age structure of the kob population, I derived a new method for estimating age-specific s u r v i v a l . This method i s described and compared to two conventional methods in Chapter 6. As shown in Fig. 1.2, the l i n e s of investigation in t h i s study derive from a unifying assumption of periodic food l i m i t a t i o n for the white-eared kob population. In Chapter 7, I summarize my conclusions, and consider the merits and drawbacks of t h i s approach. 7 c l i m a t i c seasonality movement patterns resource " l i m i t a t i o n " breeding phenology foraging population dynamics * male competition Figure 1.2 Lines of i n v e s t i g a t i o n employed i n t h i s study. 8 CHAPTER 2. SEASONAL MIGRATION IN RELATION TO RESOURCES Introduction Long distance movements are c h a r a c t e r i s t i c of many large herbivores in African savannah ecosystems. Although some populations exhibit nomadic movements apparently unrelated to season (Delaney and Happold 1979), most large herbivore migrations follow seasonal changes in resource d i s t r i b u t i o n and abundance ( S i n c l a i r 1983b), as do migrations of species drawn from a wide variety of other taxa (Dingle 1980). African savannah ecosystems exhibit pronounced seasonal changes in grassland productivity due to periodic variation in r a i n f a l l ( P h i l l i p s o n 1975; S i n c l a i r 1975; Strugnell and Pigott 1978; McNaughton 1979). As well, many grasses decline in n u t r i t i o n a l quality as they flower and mature, and t h i s process i s usually a response to r a i n f a l l seasonality (Plowes 1957; Laredo and Minson 1973; Reid et a l . 1973; Egan 1977). Early in the growing season, grasses produce new leaves with high protein content and high d i g e s t i b i l i t y . At l a t e r growth stages, many perennial grasses cease vegetative production and translocate most of the soluble constituents back into the roots and stem bases, leaving highly l i g n i f i e d , poorly d i g e s t i b l e , low protein tissues above ground. As a re s u l t , grazing animals in savannah grasslands are faced with seasonally variable food quality and absolute abundance: in the wet season food i s abundant, widely d i s t r i b u t e d , and n u t r i t i o u s ; in the dry season food i s scarce, 9 unevenly d i s t r i b u t e d , and of r e l a t i v e l y poor q u a l i t y . The most detailed studies of large-herbivore migration in Af r i c a are on the Serengeti wildebeest population (Pennycuick 1975; Maddock 1979). These studies suggest that migration tracks seasonal changes in the d i s t r i b u t i o n and abundance of high quality forage. Wildebeest move into l o w - r a i n f a l l (400mm per year) .short grass plains during the wet season, when these areas support substantial production of nu t r i t i o u s grasses. During the dry season, wildebeest migrate northwards into high r a i n f a l l (1000mm per year) t a l l grass areas that produce limited quantities of nutritious regrowth as the result of infrequent dry season rainstorms. However, Western (1975) pointed out that in many African savannah ecosystems, water supplies become r e s t r i c t e d during the dry season, due to the evaporation of wa t e r - f i l l e d depressions. As a consequence, species that are unable to meet their metabolic water requirements solely from their forage ( i . e . most grazers) are obliged to concentrate around permanent water supplies during the dry season. In contrast, many browsers can obtain s u f f i c i e n t water from their forage to meet body requirements and, as a consequence, do not exhibit dry season movements to watering points. Thus, the a v a i l a b i l i t y of water constrains the dry season foraging options of grazers. In t h i s chapter, I examine whether the seasonal migration of white-eared kob tracks s h i f t i n g d i s t r i b u t i o n s of food, water, or both. If kob migration i s an adaptation for exploiting seasonally r e s t r i c t e d resources, I predict the following: 10 1. Resources should be unevenly distributed throughout the ecosystem. 2. Resource abundance should follow predictable seasonal trends. 3. - Kob d i s t r i b u t i o n should be correlated with resource d i s t r i b u t i o n . 4. Kob should remain in locations only as long as those locations have greater abundance of resources than surrounding areas. In t h i s chapter, I f i r s t describe seasonal climatic c h a r a c t e r i s t i c s of the Boma ecosystem. Then, I examine seasonal changes in the d i s t r i b u t i o n of food, water, and the kob population, as determined from a e r i a l surveys. F i n a l l y , I test the proposed predictions against findings from both a e r i a l surveys and ground observations in the dry seasons of 1982 and 1 983. Methods Ae r i a l surveys The d i s t r i b u t i o n and abundance of large mammal populations, forage, water resources, and habitat c h a r a c t e r i s t i c s of the entire study area were measured by systematic a e r i a l surveys. A f u l l description of the a e r i a l survey procedures employed is given in Norton-Griffiths (1978). A e r i a l surveys of the whole study area were conducted in 11 early A p r i l 1980, October 1980, and July 1982. In addition, a survey of a smaller portion of the study area was performed in May 1982. A e r i a l survey teams consisted of a p i l o t , a front seat observer, and two rear seat observers. The p i l o t was responsible for correct orientation of transect l i n e s , maintenance of constant f l y i n g height, and for s i g n a l l i n g the start and f i n i s h of transect l i n e s . At 1 minute in t e r v a l s , the front seat observer recorded photometer and radar altimeter readings, observations of water sources, and estimates of woody vegetation cover and woody species composition. The rear seat observers counted a l l animals observed within a s t r i p of fixed width on both sides of the a i r c r a f t . Transect width was delineated by 2 fibreglass rods attached to either wing s t r u t . When rear seat observers encountered groups of more than 25 individuals, they took color photographs of the entire group and, time permitting, also made a v i s u a l count of group numbers. A l l rear seat observations were recorded immediately on a tape recorder as well as the time i n t e r v a l of the observation. Transect width was ca l i b r a t e d at the end of each survey by f l y i n g at known height perpendicular to an a i r s t r i p marked off in 20m intervals by white sand bags. As the plane passed over the markers, the rear seat observers counted the number of bags seen inside the rods. Several passes, at a wide range of heights, were taken. A regression l i n e was drawn through the resultant s t r i p width estimates at given heights; t h i s regression was used to calculate transect width during the actual survey from height above ground data. 12 Transect width varied between 150 and 300m for the three surveys. Transects were spaced at regular 10km intervals over the study area, with an east-west orientation. The f i r s t 2 surveys were flown at a height of 91m above ground, the f i n a l 2 surveys at 76m above ground. F l i g h t speed varied between 150 and 200km per hour. During the f i r s t 2 surveys, f l i g h t position was determined using a Global Navigation System. During the f i n a l 2 surveys, f l i g h t l i n e s were oriented to known landmarks, such as prominent rocky outcrops. White-eared kob population estimates for each survey were calculated as the observed t o t a l for a l l transects divided by the r a t i o of the sample area to the t o t a l area. Confidence l i m i t s for individual surveys were calculated according to J o l l y ' s method 2 for unequal sampling units ( J o l l y 1969). The white-eared kob population estimate for the entire study period was calculated from the weighted population estimates and variances of the four surveys (Norton-Griffiths 1978). Seasonal d i s t r i b u t i o n patterns were obtained by summing animal counts for 10km segments along each transect l i n e , then dividing t h i s count by the area observed to ar r i v e at the density per km2. Green biomass was estimated from d i g i t a l photometer readings. The photometer consists of two probes with f i l t e r s allowing passage of l i g h t at 800nm (i n f r a red) and 675nm (red) wavelengths respectively. Red l i g h t i s absorbed by chlorophyll in green vegetation, while i t i s re f l e c t e d by non-green matter. Infrared l i g h t is r e f l e c t e d by both green and non-green vegetation. Thus, the r a t i o between these reflectance values i s 1 3 correlated with the abundance of green biomass (Tucker et a l . 1973; McNaughton 1979). Photometer readings were calibr a t e d against measurements of the dry weight of green matter obtained in 0.25m2 c l i p plots (Fig. 2.1). The resulting regression l i n e was used to estimate green biomass from d i g i t a l photometer readings obtained during a e r i a l surveys. Ground Observations At five study sites (Fig. 2.2), I measured grass biomass and kob population density along 5km transects throughout the dry season (January 1 to A p r i l 30). I measured grass biomass within 0.25m2 c l i p plots taken at 0.5km intervals along transects. Subsequent to c l i p p i n g , grass samples were sorted into green le a f , green stem, and brown fractions, weighed on a triple-beam balance, dried in the sun for 8 hrs, and re-weighed. Preliminary observations confirmed l i t t l e weight change following 8 hrs of sun drying. Kob population density was estimated by counting a l l kob encountered within a s t r i p 100m either side of a moving vehicle. S t r i p width was measured using a range-finder at the beginning of each transect; thereafter observers estimated the 100m transect s t r i p width. Animal t o t a l s were subsequently summed for 0.5km transect int e r v a l s . In order to reduce the eff e c t of diurnal movements on observed d i s t r i b u t i o n patterns, I performed a l l transects in the early morning. At approximately monthly intervals during the dry season, I made similar measurements of kob population density along a 25km transect that bisected the 14 dry green biomass (gm/m 2) Figure 2.1 C a l i b r a t i o n of photometer readings. The r a t i o of reflectance values recorded by the 800nm and 675nm probes was p o s i t i v e l y r e l a t e d to 2 green grass dry weight (r =0.812; p<0.05). 15 Figure 2.2 Map of the study area. Study sites are indicated by triangles (l=Gom swamps; 2=Wangchira; 3=Ajwara; 4=Neubari; 5=Ungwala). 16 kob dry season range. Diurnal changes in kob d i s t r i b u t i o n pattern were recorded using two methods. F i r s t , at the Ajwara study s i t e I made t o t a l counts of kob present throughout the day on a 0.5km2 grassy meadow on selected dates over the course of the dry season. Second, at the Gom Swamp study s i t e I repeated a 4km transect at intervals throughout the day. Grass growth over the course of the dry season was determined from 0.25m2 c l i p plots taken inside exclosures constructed in both swampy meadows and wooded grassland areas. Grass growth rates were calculated from regressions of green biomass over time. S o i l moisture content throughout the dry season was recorded for 7 meadow and 1 wooded grassland exclosures at the Ajwara study s i t e . S o i l samples (300-500gm) were obtained at a depth of 10cm below the surface, weighed, sun-dried for 2 days, and re-weighed. The difference between wet and dry weight indicated r e l a t i v e water content. Kob feeding rates were calculated from the difference between green biomass measurements inside and outside exclosures. Since at a l l study s i t e s no other grazing ungulates were observed, I assume that a l l grass biomass removed was taken by kob. In order to determine whether kob were feeding on particular plant parts, I measured the length of leaves and stems inside and outside exclosures a f t e r kob had been feeding for several weeks. The n u t r i t i o n a l q u a l i t y of stoloniferous grasses was 1 7 estimated by analysis of: (1) crude protein content, (2) in  v i t r o d i g e s t i b i l i t y , and (3) detergent fiber content (Johnson et a l . 1964; Van Soest 1963a, b; Van Soest and Marcus 1964). A l l chemical analyses were performed by the W i l d l i f e Habitat Laboratory, Washington State University, Pullman, WA. Plant species i d e n t i f i c a t i o n s were made by the East African Herbarium, Nairobi, Kenya. The study area Physical features and vegetation types Physical features of the Boma region are depicted in F i g . 2.2. The study area, covering some 28,000 km2, i s bordered by the Boma escarpment in the south-east, the Kangen/Pibor River system in the west, and the Akobo River to the north-east. The entire region i s in a watershed that ultimately empties into the Nile River via the Sobat River. Terrain i s f l a t over most of the area, except in the v i c i n i t y of the central Maruwa H i l l s and around the Boma escarpment. Most of the plains are characterized by chernozems, or black-cotton s o i l s , which are nutrient r i c h , have a high clay content, and range in consistency from being highly adhesive in the rains to rock-hard and cracked in the dry season. These s o i l s seal quickly after being wetted, and, as a r e s u l t , subsequent r a i n f a l l puddles on the surface. The s o i l s at the base of the Boma escarpment are red-colored, volcanic loam, and are considered nutrient r i c h 18 (Willimott 1956). Red-colored sandy s o i l s are also present near basalt monoliths and small h i l l s that are scattered in the southern and eastern parts of the study area. Termitaria are present over most of the region, and are p a r t i c u l a r l y abundant in the north. Much of the study area has some degree of woody vegetation cover, being p a r t i c u l a r l y dense (>25% cover) in the north-east and around the Boma escarpment (Fig. 2.3). Intermediate woody cover (10-25%) occurs along the margins of the Kong-kong River, in the far south, and in patches of the north-eastern woodlands. Light cover (< 10%) characterizes much of the rest of the area, except for expanses of open grassland in the north and south central parts of the study area. Most of the areas of dense woody cover are composed of a community of deciduous, broadleaf trees, dominated by Combretum  fraqrans. The Kong-kong River area i s characterized by low-growing thicket species, dominated by Ziziphus mauritiana and Balanites aegyptiaca. The l i g h t l y wooded grasslands in the north are composed of scattered Balanites aegyptiaca, Acacia  seyal, and Acac ia sieberiana. The southern l i g h t l y wooded grasslands are dominated by extensive stands of Acac ia  Z a n z i b a r i c a . The far southern bushed grasslands are characterized by scattered patches of low-growing Acac ia  melli fera, Acac ia polycantha, and an unidentified Combretum species. V i r t u a l l y the entire Boma region supports substantial grass cover, excluding areas of gallery forest and thin s t r i p s of 19 Figure 2.3 Map of woody vegetation cover, as estimated during aerial surveys (shaded=25% cover; hatched=10-25%; stippled=l-10%; open 1%) . Each grid square is 10 x 10 km. 20 riparian woodland along the margins of the Akobo River. Even the extensive Combretum woodlands support substantial grass growth. The predominant grass species over much of the region i s Hypparhenia rufa, which grows to heights of 2-3m each year. In the far south and in sandier areas, a wide variety of grasses are common: Heteropoqon contortus, Setaria incrassata, Sporobolus ioclados, and Panicum ioclados. A stoloniferous grass species, tentatively i d e n t i f i e d as Echinochloa  pyramidalis, characterizes the flood p l a i n margins of major watercourses. Due to surface flooding, i t was not feasible to v i s i t study s i t e s during the wet season, and the absence of reproductive structures during the dry season f i e l d study period permitted only tentative i d e n t i f i c a t i o n s of grass species. A f u l l l i s t of a l l i d e n t i f i e d vegetation samples i s given in Appendix 1. Climatic seasonality Mean monthly records of temperature, r e l a t i v e humidity, wind d i r e c t i o n , and evaporation piche from the nearby Akobo township are given in Table 2.1. Mean da i l y maximum temperatures range between 38.9°C in March to 30.8°C in August. Mean da i l y minimum temperatures range between 23.2°C in A p r i l and 19.9°C in December. Relative humidity i s highest during the rainy season, from May to November, but i s considerable throughout the year. Seasonal changes in evaporation rate follow seasonal temperature v a r i a t i o n . Seasonal changes in r a i n f a l l are i l l u s t r a t e d in F i g . 2.4. Table- 2.1 Monthly climatic normals for Akobo township (see Fig. 1.1). Data are mean values for 27 years of records between 1941 and 1970 (from the Meteorological Department, Government of the Sudan).. max min relative humidity wind evaporation month temp temp 6:00 12:00 direction piche Jan. 36.6 20.0 47% 27% NE 9.7 Feb. 37.9 21.6 41 24 NE 11.2 Mar. 38.9 23.1 49 27 SE 10.4 Apr. 37.5 23.2 62 35 SE 8.2 May 34.9 22.2 74 47 SE 5.7 Jun. 32.6 21.5 82 55 SE 4.1 Jul. 30.9 21.5 87 62 SE 2.9 Aug. 30.8 21.1 88 65 SE 2.6 Sep. 32.0 22.7 86 59 SE 2.9 Oct. 33.5 21.8 82 54 SE 3.5 Nov. 34.6 21.2 75 46 SE 4.7 Dec. 35.4 19.9 59 34 NE 7.1 * Pachalla (1232mm) 1 5 0 1 0 0 6 0 • 1 . 5 1 . 0 0 . 5 Boma (1430mm) 2 0 0 E 1 5 0 E « 1 0 0 " CO 5 0 -2 . 0 O 1 . 5 ^ C C © o wE 1 . 0 — .5 •*- Z! CD CO O > 0 . 5 ° 1 5 0 i Loelli (603mm) T 1 . 5 1 0 0 -5 0 1 . 0 0 . 5 N D J F M A M J J A S O month Figure 2.4 Mean monthly r a i n f a l l totals for Pachalla, Boma Plateau, and L o e l l i (see Fig. 2.2) indicated by histograms. Triangles indicate coefficient of variation (S/mean) for monthly r a i n f a l l totals. 23 At a l l s i t e s , r a i n f a l l i s highest between A p r i l and November, inclu s i v e . Between year v a r i a t i o n in r a i n f a l l (measured as the c o e f f i c i e n t of variation) i s highest during the dry season, indicating that r a i n f a l l is most unpredictable at t h i s time of year. Seasonal r a i n f a l l patterns in the southern Sudan are clo s e l y related to movements of the Inter-Tropical Convergence Zone (ITCZ) ( G r i f f i t h s 1972). The ITCZ is the interface between south-easterly and north-easterly winds in the t r o p i c s . This zone moves north during the northern hemisphere summer, and south of the equator during the northern hemisphere winter. After the ITCZ passes through the southern Sudan each March or early A p r i l , p r e v a i l i n g winds s h i f t from a northerly to a southerly d i r e c t i o n (Table 2.1). These winds bring in moisture-laden a i r from the A t l a n t i c coast. "During the dry season, sporadic r a i n f a l l results from thunderstorms that arise from the Ethiopian Highlands. However, during most of the dry season hot, dry winds are t y p i c a l . Mean annual r a i n f a l l t o t a l s for the Pachalla and L o e l l i s i t e s (Fig. 2.4) suggest a gradient from high r a i n f a l l in the north (1232 mm per year) to low r a i n f a l l in the south (603 mm per year). This trend i s consistent with published r a i n f a l l isohyet maps for the entire southern Sudan ( G r i f f i t h s 1972). R a i n f a l l on the Boma escarpment i s unusually heavy because of i t s high elevation. 24 Results A e r i a l survey results Seasonal changes in green biomass There was considerable seasonal and s p a t i a l v a r i a t i o n in green biomass over the study area (Fig. 2.5). In the early wet season (July), the northern half of the region supported biomasses in excess of 300 gm/m2, while southern areas supported less than 300 gm/m2. In the lat e wet season, the northern region supported more than 200 gm/m2 while southern areas supported less than 200 gm/m2. In both the early and late wet seasons, green biomass increased from south to north (early wet: ANOVA,F=42.8, p<0.05; late wet: ANOVA, F=60.5, p<0.05). In the dry season, green biomass abundance was low in a l l areas (ANOVA, F=2.0, n.s.), but was greatest along the margins of the Kong-kong, Kangen, and Oboth watercourses and on the Boma escarpment (Fig. 2.6). Less than 5% of the study area had dry season green biomasses in excess of 50 gm/m2, mostly around the Boma escarpment. In a l l seasons, the s p a t i a l d i s t r i b u t i o n of green biomass was patchy, possibly indicating that r a i n f a l l was unevenly d i s t r i b u t e d . These results suggest that annual grass production rates were highest in the north and on the Boma escarpment, and lowest in the southern areas. This trend i s consistent with the north-25 500-, 300-dry E co CO CO co E o B c CD CD i_ CO c to CD E 100-500-, T 1 r — 300-100-late wet ~ i 1 1 1 1 1 1 1 1 1 r 500 300 100 early wet i 1 1 1 1 1 1 1 1 1 1 i i 1 5 9 13 17 21 25 south transect north Figure 2.5 North-south gradient in seasonal green biomass abundance, estimated from d i g i t a l photometer readings during aerial surveys. Shaded bars indicate transects in which kob were observed during specific aerial surveys (dry: April 1980; late wet: October 1980; early wet: July 1982). late wet o OOO O O o OOO 0 0 0 ° early wet O O O O O O ^ ^ O O o OOO OO O O o o oXt 0-25 gin/m1 o 26-50 o 61-100 O 101-200 O 201-300 O 301-400 O >*oo dry: April 1980 survey late wet: October 1980 early wet: July 1982 Figure 2.6 Seasonal distribution of green biomass, estimated from d i g i t a l photometer readings during aerial surveys. 27 south r a i n f a l l gradient described e a r l i e r . In the dry season, green biomass values were generally low across the ecosystem, but were somewhat greater along the margins of northern watercourses and at higher elevations. Seasonal changes in water a v a i l a b i l i t y The seasonal d i s t r i b u t i o n of water supplies is shown in F i g . 2.7. During the rains, water supplies were widely d i s t r i b u t e d throughout the region. The north-central plains were inundated; standing water covered several thousand km2. Swamping did not occur south of the Maruwa H i l l s , near the Boma escarpment, or in the Combretum woodlands to the east of the plains, but during the rains there were large numbers of f u l l watercourses and waterholes scattered throughout these areas. By the late wet season, most of the ephemeral swamped grasslands in the north dried out and water was c h i e f l y r e s t r i c t e d to watercourses and waterholes. The south was less well supplied than the north, possibly as a result of lower r a i n f a l l . In the dry season, water was mainly r e s t r i c t e d to the northern watercourses: the Oboth, Neubari, and Akobo Rivers. The largest remaining swamps, c a l l e d the Gom Swamps, are part of the Oboth drainage system. Some waterholes remained in the Combretum woodlands and around the Boma escarpment. Figure 2.7 Seasonal distribution of water supplies, estimated during aerial surveys. Solid dots indicate standing water observed at least once per 10km in a given transect. Boundaries of the study area as indicated in Fig. 2.6. 29 White-eared kob population numbers During 1980-82, the Boma kob population numbered approximately 830,000, making i t the second largest migratory ungulate population in A f r i c a , after the Serengeti wildebeest ( S i n c l a i r and Norton-Griffiths 1982). There was no s i g n i f i c a n t change in kob population numbers over the course of the study (Table 2.2; t-tests for a l l pairs of census t o t a l s , n.s.). Seasonal migration of white-eared kob The seasonal d i s t r i b u t i o n pattern of white-eared kob i s shown in F i g . 2.8. During the rains, kob were concentrated in large herds south of the Kangen River, in l i g h t l y wooded grasslands where r a i n f a l l i s low. By October, kob moved northwards along the Kangen River, crossing the ri v e r at several points. Large numbers of kob proceeded as far north as Pibor Post before crossing the Kangen River. A e r i a l reconnaissance f l i g h t s and f i e l d observations indicated that kob moved into the dry season range by early January. The dry season range, adjoining the Neubari/Oboth River system, was maintained u n t i l the onset of rains. The southward migration was sporadic; the herds often remained stationary for a number of days and even reversed d i r e c t i o n on occasion. By May, migratory kob herds were in the v i c i n i t y of the Maruwa H i l l s and the Boma escarpment. The wet season range was reached by early July. The dry season and wet season ranges were 150 to 200km apart. The Oboth River system was the only watercourse that showed 30 Table 2.2 Boma white-eared kob population estimates from aerial censuses (1980-1982). The mean population estimate for the entire period was calculated from the weighted estimates of individual censuses (Norton-Griffiths 1978) . survey date population estimate 95% confidence interval April 1980 October 1980 May 1982 July 1982 888,880 651,020 869,553 964,932 + 70 % + 93 % + 47 % + 104 % mean 831,081 + 31% 3 1 Figure 2.8 Seasonal d i s t r i b u t i o n of white-eared kob. Size of the s o l i d 2 dots i n d i c a t e s kob density per km . 32 continued flow during the dry season, although l i m i t e d supplies of water were available in woodland waterholes during part of the dry season. A e r i a l surveys suggest that the dry season range was better endowed with water than surrounding areas. F i g . 2.9 depicts the r e l a t i v e frequency d i s t r i b u t i o n of kob in r e l a t i o n to distance to the nearest water supplies. Kob were concentrated closer. to water supplies than expected i f d i s t r i b u t i o n was random (Kolomogorov-Smirnov t e s t , D=0.49, p<0.05). This histogram indicates that few kob (<5%) were located farther than 20km from the nearest known water supplies, and most kob (75%) were within 10km. During the wet season, no locations in the study area were further than 20km from the nearest water supplies. If kob migration tracks the seasonal a v a i l a b i l i t y of food, then one would expect a pos i t i v e correlation between kob population density and green biomass. As shown in F i g . 2.5, kob vacated high biomass northern areas and concentrated in the south during the early wet season (July). By the late wet season (October), kob began to move north into the high biomass swamped grasslands. By the late dry season ( A p r i l ) , northern areas occupied by the kob did not d i f f e r s i g n i f i c a n t l y from southern areas in terms of green biomass. A e r i a l survey results showed no s i g n i f i c a n t c o r r e l a t i o n between kob population density and green biomass during the dry season (r 2=0.0l, n.s.). Since kob are r e s t r i c t e d in their choice of habitats by the a c c e s s i b i l i t y of water supplies, I examined the same data, excluding a l l ecosystem locations 3 3 100 8 0 c o t j 6 0 00 3 a o a n o 2 0 4 0 2 0 ca a 4 0 CO >. "D 3 6 0 CO 80 1 0 0 J < 1 0 2 0 - 2 9 4 0 - 4 9 6 0 - 6 9 1 0 - 1 9 3 0 - 3 9 5 0 - 5 9 distance to water (km) Figure 2.9 Kob d i s t r i b u t i o n i n r e l a t i o n to the nearest water supplies during the dry season (April) a e r i a l survey. 34 further than 20km from water. Results indicate there was s t i l l no c o r r e l a t i o n (r 2=0.0l, n.s.). However, forage d i s t r i b u t i o n patterns observed in the late dry season may not r e f l e c t forage d i s t r i b u t i o n patterns available at the time kob actually migrated i n t o the dry season range, since kob quickly reduce grass biomass to low l e v e l s . A clearer test might be to consider the abundance of green biomass in a l l regions at the end of the wet season, before kob had an opportunity to graze i t down. After doing t h i s , I found a substantially higher c o r r e l a t i o n between kob population density during the dry season and the amount of green biomass that was available in the late wet season, prior to migration into the dry season range (r 2=0.l87; p<0.05). Thus, while there was weak evidence that kob migration tracked forage abundance during the t r a n s i t i o n from wet to dry seasons, there was no indication that migration tracked s h i f t i n g d i s t r i b u t i o n s of green biomass at other times of the year. There was more substantial evidence that the dry season d i s t r i b u t i o n of kob was correlated with the a v a i l a b i l i t y of nearby water supplies. 35 Dry season ground observations Kob d i s t r i b u t i o n patterns Habitats in the kob dry season range may be lumped into three main types: 1) wooded grassland with a scattered woody cover of Acac ia sieberiana, Combretum fraqrans, Balanites  aeqyptiaca, and Ziziphus mauritiana , and an understory of Hyparrhenia rufa grass; 2) t a l l H.rufa open grasslands; and 3) swampy meadows with dense low growing stoloniferous grass, Echinochloa pyramidalis, interspersed with occasional patches of t a l l e r clump-forming grasses. Data from 25km transects show that kob were unevenly d i s t r i b u t e d across the dry season range, and were concentrated around the swampy meadows (Fig. 2.10). Much lower population densities were found in areas of t a l l open grassland or in the wooded grasslands that surround the swampy meadows (Fig. 2.11). Although substantial numbers remained in swampy meadows throughout the day, most kob moved off into surrounding woodlands during the hottest part of the day (Fig. 2.12). This diurnal movement pattern became more pronounced, as the dry season progressed. Large numbers of kob continued to feed on swampy meadows throughout the night. C l i p plot measurements (Table 2.3) show that there was nine times as much green biomass in swampy meadows than in surrounding wooded grasslands ( t - t e s t , t=3.66, p<0.05) even though t o t a l biomass was similar ( t - t e s t , t=0.10, n.s.). The w ater water 7 9 1 1 13 15 17 19 21 23 25 distance along transect (km) Figure 2.10 Kob dry season aggregation near swampy meadows, as indicated by mean numbers seen per km during 25km transects from January to April, 1983 (n=3). 37 Jan . F e b . date Mar. Figure 2.11 Kob population d e n s i t i e s i n wooded grassland (broken l i n e ) vs. swampy meadow ( s o l i d l i n e ) habitats at the Ajwara study s i t e during the 1983 dry season ( v e r t i c a l bars=95% confidence l i m i t s ) . 38 8:00 12:00 time of day 16:00 Figure 2.12 Kob di u r n a l movements onto meadows. Maximum kob numbers were present i n the early morning and l a t e afternoon. Note the increasing tendency to abandon meadows at mid-day as the dry season progressed from January to March. Table 2.3 Dry season grass standing crops in meadows vs. wooded grasslands at the Ajwara study site in 1983. Meadows had more green biomass than wooded grasslands (t-test; t=3.66; p<0.05), but no other significant differences were observed. MEADOWS 2) mean dry wt. (gm/0.25m location date brown green total sample size Neubari 2-3-83 7.7 9.1 16.0 7 Neubari 23-3-83 1.0 2.3 3.3 2 Gom 14-3-83 5.9 7.9 13.8 5 Gom 1-4-83 1.7 4.9 6.6 5 Wangchira 13-2-83 1.8 3.7 5.5 11 Wangchira 21-3-83 9.5 17.3 26.8 8 Ungwala 24-3-83 42.1 7.5 49.6 • 6 Ajwara 16-1-83 14-4-83 to 7.5 15.8 23.3 18 mean 9.7 8.5 18.2 8 ( SE 4.5 1.8 5.0 WOODED GRASSLANDS Neubari 2-3-83 29.7 0.4 30.1 7 Neubari 23-3-83 24.3 0.1 24.4 6 Gom 14-3-83 14.1 0.8 14.9 6 Gom 1-4-83 0.3 1.1 1.4 6 Wangchira 13-2-83 1.1 0.9 2.0 11 Ungwala 24-3-83 6.2 1.1 7.3 5 Ajwara 16-1-83 38.4 2.1 40.5 17 mean 16.3 0.9 17.2 7 ( SE 5.2 0.2 5.2 40 stoloniferous grasses remained green throughout the dry season long after t a l l e r grasses had flowered and senesced. Since green grasses are generally more protein r i c h than senescent grasses (Wilson 1976), kob that fed in the meadows should' have obtained forage higher in protein than those that fed in surrounding wooded grasslands. Small ruminants require food with higher protein content than larger ruminants, since they need to process food at a faster rate to compensate for their smaller gut capacity (Owen-Smith 1982; Demment and Van Soest 1985). Following Owen-Smith's (1982) cal c u l a t i o n s , I estimate that kob require forage with a minimum 9% crude protein content. Stoloniferous grass samples col l e c t e d from meadows contained on average 17% crude protein, well above the minimum requirements for kob (Table 2.4). In contrast, senescent Hyparrhenia grass in surrounding wooded areas was probably too low in protein to be u t i l i z e d by kob, since other studies in nearby grasslands report crude protein values of 2% in mature Hyparrhenia grass (Jonglei Investigation Team, unpublished). Since kob rarely fed on t a l l Hyparrhenia grass, I conclude that dry season grazing was largely confined to meadows because only these areas offered forage of s u f f i c i e n t q u a l i t y . Moreover, r e s u l t s from exclosure plots indicate that stoloniferous grasses continued to produce new green growth throughout the dry season (Fig. 2.13; ANOVA, F=178.7, p<0.0l), when t a l l grasses were no longer productive. This growth was manifested in elongation of stem and leaf tissue (Fig. 2.14; stems: ANOVA, F=50.8, p<0.05; leaves: ANOVA, F=106.8, p<0.05). 41 Table 2.4 Nutritional analyses of above-ground tissues from stoloniferous vs. Hyparrhenia grasses. No differences were observed between stoloniferous grass fractions except in crude protein content (ANOVA; p<0.05) . Statistical comparisons were not possible between Hyparrhenia and stoloniferous grasses because variances were not stated in Mefit-Babtie report (1983). CP=crude protein; IVDDM= iri vitro digestible dry matter; NDF=neutral-detergent soluble fraction; ADF=acid-detergent soluble fraction; LIG=lignin content; AIA= acid insoluble ash. grass type part n CP IVDDM NDF ADF LIG AIA stoloniferous green 19 21.7 39.3 66.8 30.1 8.8 4.8 leaf stoloniferous green 17 16.1 45.7 62.8 30.9 12.1 4.7 stem stoloniferous brown 10 18.2 43.3 64.8 32.5 9.2 4.4 Hyparrhenia'*' * total n.a. 1.7 n.a. n.a. n.a. n.a. n.a. rufa plant 1. From: Mefit-Babtie, 1983 (April). Development Studies in the Jonglei Canal Area. Range Ecology Survey, Livestock Investigations, and Water Supply. A report to the Jonglei Executive Organ, Government of the Sudan, Vol. 1-10. 42 Figure 2.13 Dry season meadow grass growth, Ajwara study si t e . Values 2 determined by 0.25m clip-plots from exclosures erected in mid-January, 1983 (r2=0.78; p<0.05) . 43 Figure 2.14a Dry season leaf growth, Ajwara study site (r =0.82; p < 0 . 0 5 ) . 44 y = 1.32x + 21.7 200n time (days) Figure 2.14b Dry season stem growth, Ajwara study s i t e (r =0.68; p<0.05). 45 These r e s u l t s suggest that kob concentrate at high d e n s i t i e s d u r i n g the dry season in feeding r e f u g i a (swampy meadows) that continue to produce high p r o t e i n food when other grasslands have l i t t l e green growth. In 1983 there was e s s e n t i a l l y no dry season r a i n f a l l i n the kob dry season range. By c o n t r a s t , i n 1982 approximately 70mm of r a i n f e l l i n l a t e February, i n the middle of the dry season. While t h i s r a i n f a l l was i n s u f f i c i e n t to form new waterholes, s u b s t a n t i a l amounts of green regrowth appeared i n the t a l l Hyparrhenia open g r a s s l a n d s and i n the wooded g r a s s l a n d s . A comparison between 1982 and 1983 t r a n s e c t d i s t r i b u t i o n s suggest that kob d i s p e r s e d away from the swampy meadows i n response to the widespread occurrence of t h i s unusual f l u s h of green growth ( F i g . 2.15). C i r c u m s t a n t i a l o b s e r v a t i o n s from s e v e r a l of these dry season r e f u g i a a l s o i l l u s t r a t e the importance of nearby water s u p p l i e s . While the meadows continued to produce s u b s t a n t i a l regrowth, kob abandoned these areas as soon as nearby waterholes d r i e d up ( F i g . 2.16). Thus, high q u a l i t y food alone was i n s u f f i c i e n t t o meet kob dry "season h a b i t a t requirements; they a l s o needed nearby s u p p l i e s of water. F a c t o r s a f f e c t i n g forage abundance Dry season grass growth r a t e s i n swampy meadows depended on r e s i d u a l s o i l moisture ( F i g . 2.17; ANOVA, F=16.0, p<0.05). T o t a l l y s a t u r a t e d s o i l s were approximately 40% water by weight. T a l l Hyparrhenia wooded g r a s s l a n d s surrounding swampy meadows 50- , 3 0 -10 1982 before rain E n o 6 c 50 -i 3 0 -10 1982 after rain 70 -i 5 0 -30 1 0 -1983 1 2 3 4 5 6 7 8 distance from water (km) Figure 2.15 Kob aggregation near meadows, as affe c t e d by dry season r a i n f a l l . Data from 25km transects show that kob dispersed away from meadows following 70mm of r a i n f a l l i n the 1982 dry season. In 1983, kob remained concentrated near meadows throughout the dry season. Figure 2.16 Kob abandonment of meadows following disappearance of water supplies. The mean number of kob present on meadow declined following evaporation of the central water hole (solid line), despite the fact that food intake over the same period was increasing (broken lin e ) . 48 Figure 2.17 Dry season stoloniferous grass growth rate as a function of 2 s o i l moisture content, Ajwara study site (r =0.86; p<0.05) . 49 had s o i l moisture values of less than 18% during the dry season and none of the exclosures in the wooded grasslands produced dry season grass re-growth. S o i l moisture declined even in the meadows over the course of the dry season (Fig. 2.18; ANOVA, F=17.2, p<0.05). Kob maintained a uniformly short sward (stems < 25mm) in the swampy meadow areas: new regrowth was immediately u t i l i z e d as forage. This explains why the dry season grass biomass in meadows was l i t t l e d i f f e r e n t from the unproductive surrounding wooded grasslands. While new regrowth occurred over the entire meadow, kob concentrated their feeding close to the margins of drying waterholes (Fig. 2.19; males: "XL2 = 212.7, p<0.0l; females: X2=788.2, p<0.05). Since grass growth was related to s o i l moisture content, kob probably concentrated in those areas in order to exploit faster growing grasses, although I have no data to test t h i s d i r e c t l y . Females showed a greater tendency to concentrate close to the water margin than did males ( %2=39.2, p<0.05). Feeding selection for plant parts Comparisons between grass structure inside and outside exclosures suggest that at times kob selected for certain parts of the grass plants (Table 2.5). When feeding on young t a l l grasses (l-2m), kob selected green leaves but fed l i t t l e on stems. However, kob fed on both stems and leaves of stoloniferous grasses. There was a small difference in protein 164 20 40 60 days since Jan. 19 80 Figure 2.18 Soil moisture decline over the dry season in meadows at the 2 Ajwara study site (r =0.44; p<0.05) . 51 30- females n=1193 O o o 10 30 10 males n=802 0 15 30 45 60 75 distance from water (meters) Figure 2.19 Kob feeding concentration near drying waterholes, Ajwara study site. Although both sexes concentrated in the region 15-30m from the water's 9 edge, this tendency was most pronounced in females (X =39.2; p<C0.05). Table 2.5 Kob dry season feeding selection for leaf vs. stem tissue. There was no significant difference in t a l l grass stem lengths inside and outside exclosures, while there were significant differences in th lengths of t a l l grass leaves, short grass stems, and short grass leave (t-tests; p<0.05) . mean stem or leaf length plant part inside outside t t a l l grasses stems t a l l grasses leaves short grasses stems short grasses leaves 961 + 235 175 + 20 137 + 14 290 + 23 874 + 180 31 + 4 22 + 4 36 + 6 1.3 , n.s. 30.4 , * 20.3 , * 27.5 , * 53 content between stoloniferous grass stems and leaves (Table 2.4). However, Owen-Smith (1982) has documented elsewhere a substantial difference in protein content and d i g e s t i b i l i t y between leaves and stems of t a l l grasses. Thus, kob apparently selected leaves when feeding on t a l l , presumably highly l i g n i f i e d grasses while they showed no such selection when feeding on stoloniferous grasses. Food a v a i l a b i l i t y r e l a t i v e to kob requirements There i s l i t t l e information available on food requirements of free-ranging African ungulates, so I estimated kob food requirements from data on domestic sheep, which are similar to kob in size and feeding habits. The maintenance energy requirements of rams and non-lactating ewes i s 536 kJ/W (NRC 1975). During l a c t a t i o n , energy requirements increase to 1335 kj/W0"75. The energy content of stoloniferous grass during the dry season was 16.53 kJ/gm (n=21; SE=0.042) and the dry matter d i g e s t i b i l i t y was 42.3% (n=14; SE=2.1). Thus, the di g e s t i b l e energy available in forage was 6.94 kJ/gm (=16.53 kJ/gm x 0.42). The r a t i o of energy required over d i g e s t i b l e energy available provides an estimate of minimum forage requirements (Demment and Van Soest 1985). On this basis, males should require 1.56 and non-lactating females 1.23 kg of forage per day. A large proportion of kob females continued l a c t a t i n g throughout the dry season (Chapter 4). These females would require 3.08 kg/day to meet their energy needs. Kob d a i l y food intake during the 1983 dry season averaged 54 0.93 kg/individual (n=5; SE=0.20). This rate of food intake was s i g n i f i c a n t l y less than requirements of males (t=-3.15; p<0.05) and l a c t a t i n g females (t=-l0.80; p<0.05), but not of non-l a c t a t i n g females (t=-1.49; n.s.). Thus, even though kob selected the best feeding locations available in the dry season range, they were s t i l l unable to meet their d a i l y forage requirements. A series of simple calculations indicates that food a v a i l a b i l i t y during the wet season probably exceeded kob requirements. The green biomass available in the early wet season range was approximately 0.l3kg/m2. Since kob were di s t r i b u t e d over an area of 1500km2 for no more than 90 days, the food available per individual per day was at least 2.7kg, or almost twice the da i l y kob requirement. This is a conservative estimate, since i t ignores grassland productivity during the wet season and the p o s s i b i l i t y of kob foraging in adjoining areas. Similar calculations in the late wet season indicate that food a v a i l a b i l i t y was at least 4.2kg/individual/day. While other factors, such as sward structure (Bell 1971), might constrain actual forage a v a i l a b i l i t y , these calculations demonstrate that, in theory, there was s u f f i c i e n t green biomass available during the wet season to meet kob energetic needs. 55 Discussion Seasonal changes in resource d i s t r i b u t i o n It has been often assumed that the tropics are less seasonal than temperate regions (Dobzhansky 1950). This is untrue, p a r t i c u l a r l y in t r o p i c a l savannahs. While annual temperature variation i s often minor, most t r o p i c a l savannahs have marked seasonal v a r i a t i o n in r a i n f a l l . In A f r i c a , r a i n f a l l seasonality i s linked to trans-equatorial movements of the Inter-Tropical Convergence Zone (ITCZ). Resultant s h i f t s in wind patterns produce increased r a i n f a l l in the wake of the ITCZ. As a result of highly seasonal r a i n f a l l , resources essential to the white-eared kob population varied considerably in d i s t r i b u t i o n and abundance. In t h i s study, both green forage and water supplies became increasingly r e s t r i c t e d during the dry season. During the wet season, abundant supplies of both forage and water were available throughout the Boma region. There was a north-south gradient in annual r a i n f a l l and peak green biomass in the Boma ecosystem. During the wet season, green biomass was highest in the north and lowest in the south. During the dry season, swampy meadows that produced continued grass re-growth occurred along the major northern watercourses, as well as permanent water supplies. L i t t l e green grass or water was available elsewhere during the dry season. These results are consistent with the f i r s t two predictions 56 o u t l i n e d a t the b e g i n n i n g of t h i s c h a p t e r . Both food and water were unevenly d i s t r i b u t e d throughout the ecosystem, as a r e s u l t of an i n c r e a s i n g r a i n f a l l g r a d i e n t from s o u t h t o n o r t h . Resource abundance e x h i b i t e d pronounced v a r i a t i o n throughout the year because of s e a s o n a l c l i m a t i c changes. Se a s o n a l kob m i g r a t i o n The Boma w h i t e - e a r e d kob p o p u l a t i o n m i g r a t e d each year from s o u t h e r n wooded g r a s s l a n d s i n the wet season t o the n o r t h e r n Neubari/Oboth R i v e r system i n the dry season. Comparisons between o b s e r v a t i o n s from t h i s s t u d y and a n e c d o t a l o b s e r v a t i o n s by e a r l y European v i s i t o r s t o the Boma r e g i o n suggest t h a t the m i g r a t o r y phenology has changed l i t t l e s i n c e the 1940's (Cave and C r u i k s h a n k 1940; L y e t h 1947; Weeks 1947; Anderson 1949; Za p h i r o 1949). As w e l l , a l l o b s e r v e r s commented on the e x t r a o r d i n a r y numbers of kob i n t h e m i g r a t o r y h e r d s , s u g g e s t i n g t h a t kob have been abundant i n the r e g i o n f o r ' a t l e a s t f o r t y y e a r s . F i n a l l y , these h i s t o r i c a l r e c o r d s of kob m i g r a t i o n were a l l from s o u t h e r n a r e a s , but i n w i d e l y s e p a r a t e d l o c a t i o n s . T h i s s u g g e s t s t h a t the wet season range may v a r y i n l o c a t i o n from year t o y e a r . S i n c e kob c o n c e n t r a t e i n the dry season a l o n g one s p e c i f i c r i v e r system, one would expect the dry season range t o v a r y c o m p a r a t i v e l y l i t t l e i n l o c a t i o n from year t o y e a r . D u r i n g the dry season, permanent water s u p p l i e s were c h i e f l y r e s t r i c t e d t o the Neubari/Oboth R i v e r system. Kob c o n c e n t r a t e d i n l o c a t i o n s a d j o i n i n g t h i s w a t e r c o u r s e , 57 p a r t i c u l a r l y in the Gom Swamps. Dry season movement towards permanent water supplies was consistent with Western's (1975) hypothesis that most grazers must migrate seasonally in order to obtain s u f f i c i e n t water supplies. Laboratory t r i a l s indicate that the conspecific Uganda kob requires water on a d a i l y basis (Schoen 1971). However, within their dry season range, kob concentrated at high densities (over 1,000/km2) in feeding refugia that produced nutritious regrowth throughout the dry season when surrounding grasslands showed l i t t l e green biomass. Many of these swampy meadows were old meander channels of the Oboth River that have been subsequently cut off from the main flow. These low-lying depressions remained flooded longer than surrounding t a l l Hyparrhenia grasslands. As a consequence, the s o i l in the meadows retained moisture longer than surrounding s o i l s . This residual s o i l moisture supported substantial regrowth of stoloniferous grasses high in protein throughout the dry season. Thus, migration to ephemerally swamped grasslands during the dry season ensured the a v a i l a b l i t y of both n u t r i t i o u s forage and water when these resources were scarce elsewhere in the ecosystem. These findings are consistent with my t h i r d prediction, that kob d i s t r i b u t i o n should be correlated with resource abundance. During the wet season, however, kob migrated south into low r a i n f a l l areas that supported lower biomasses of green forage than in the north. This i s inconsistent with the hypothesis that migration f a c i l i t a t e s the exp l o i t a t i o n of s h i f t i n g 58 d i s t r i b u t i o n s of important resources. Why don't kob simply remain in the northern grasslands on a year-round basis? I suggest that kob can not do so because much of the north becomes flooded during the wet season. Kob exhibited signs of restlessness following extensive rains, even during the dry season, as was seen in February 1982. Following 70mm of r a i n f a l l , kob aggregated into larger groups and began moving in columns southwards, vacating areas that several days previously had supported upwards of 1000 individuals per km2. When these temporary rains ceased, kob returned in large numbers to the swampy meadows in the north. By contrast, in February and March of 1983, northern areas received no ra i n , while 50mm of rain f e l l around the Boma plateau, producing luxurient grass regrowth. Kob showed no signs of migratory restlessness u n t i l the north received heavy rains in the middle of A p r i l . Thus, kob may u t i l i s e the onset of rains in thei r location as the proximate cue to begin the migration south. These observations are consistent with the hypothesis that kob vacate northern areas in order to avoid waterlogged s o i l s and surface flooding. The fourth prediction outlined at the beginning of this chapter i s that kob should remain in a given location only as long as food and water resources are more abundant than in surrounding areas. Dry season observations were consistent with th i s prediction, since kob dispersed away from swampy meadows either when nearby water supplies dried up or when dry season r a i n f a l l produced new green re-growth (with presumably high protein content) in surrounding Hyparrhenia grasslands. 59 In conclusion, kob movements from the wet season to dry season ranges apparently tracked increasingly scarce supplies of both food and water. However, wet season movements were inconsistent with the resource exploitation hypothesis, since kob moved into grasslands with lower green biomass and less abundant water supplies than the areas they had l e f t . It seems l i k e l y that another factor, such as avoidance of surface flooding, explains this southward migration. Benefits of migration Numerous studies suggest grazers can improve their n u t r i t i o n a l intake by selective feeding. Studies on domestic stock suggest that the rate of food intake i s related to the protein content of forage, due to effects on forage d i g e s t i b i l i t y , rumen retention time, and possibly appetite suppression (Laredo and Minson 1973; Reid et a l . 1973; Egan 1977). This i s p a r t i c u l a r l y important in savannah grasslands because t r o p i c a l grasses generally have lower d i g e s t i b i l i t i e s than temperate grasses (Minson and McLeod 1970; Reid et a l . 1973). Grazers have minimum forage protein requirements below which they begin to lose weight, because they are forced to metabolize their own body reserves to compensate for n u t r i t i o n a l i n s u f f i c i e n c y (Chalmers 1961; Bredon and Wilson 1963). P a s t o r a l i s t s l i v i n g in the Boma region also migrate with their c a t t l e herds between wet season ranges west of the Pibor River to the Kong-kong River and Gom Swamps during the dry 60 season. Similar p a s t o r a l i s t migrations were once common in much of the Sahel region of A f r i c a ( S i n c l a i r and F r y x e l l 1985). Thus, t r a d i t i o n a l methods of animal husbandry, presumably designed to maximize secondary production, apparently mimic evolved behaviors in natural herbivore populations. These findings suggest that migration i s a p a r t i c u l a r l y successful strategy for coping with resource scarcity in highly seasonal envi ronments. 61 CHAPTER 3. FOOD LIMITATION AND KOB MORTALITY PATTERNS Introduction Numerous studies suggest that food a v a i l a b i l i t y l i m i t s many ungulate populations (Klein 1968; Caughley 1970; Bobek 1977; S i n c l a i r 1977; McCullough 1979; Leader-Williams and Ricketts 1980; Fowler 1981; Clutton-Brock et a l . 1982). Many of these studies i l l u s t r a t e changes in recruitment or survival rates as a function of population size (e.g. McCullough 1979; Fowler 1981; Clutton-Brock et a l . 1982). However, comparatively few ungulate population studies have actually documented that resources are l i m i t i n g or measured the effect on consumer populations (Bobek 1977; S i n c l a i r 1977; Bayliss 1985; S i n c l a i r et a l . 1985). Thus, while ungulate populations do exhibit density dependent responses, there i s l i t t l e evidence that food l i m i t a t i o n i s responsible. An a l t e r n a t i v e view i s that herbivore populations can not be limited by food abundance because only a small fracti o n of p o t e n t i a l l y edible food i s ever consumed (Hairston et a l . 1960; Slobodkin et a l . 1967; Van Valen 1973). Moreover, at least some large-mammal populations may be maintained below the vegetation l i m i t e d carrying capacity by predation (Mech and Karns 1977; Smuts 1978; Caughley et a l . 1980; Gasaway et a l . 1983), or disease (Christian et a l . 1960; S i n c l a i r and Norton-G i f f i t h s 1979; Berry 1981). If white-eared kob are limited by food a v a i l a b i l i t y , I 62 predict the following: 1. Survival and/or recruitment rates should decline during years of l i m i t i n g food abundance and during seasonal periods of food s c a r c i t y . 2. There should be a direct correspondence between body condition indices and survival rates. 3. Survival in s p e c i f i c locations should be p o s i t i v e l y correlated with food a v a i l a b i l i t y . In t h i s chapter I examine these predictions. F i r s t , I consider trends in population numbers and age structure of the population in relation to a large-scale drought that occurred in 1980, immediately preceding this study. Second, I report on f i e l d observations of mortality and recruitment of yearlings for 1982 and 1983. Third, I relate dry season mortality rates to rates of food intake and indices of body condition. F i n a l l y , I compare my findings against the predictions of the food l i m i t a t i o n hypothesis. Methods Population density I recorded population density and carcass numbers at four study s i t e s , from February to A p r i l , 1982 and at f i v e study s i t e s from January to A p r i l , 1983 (Fig. 2.2). Population numbers were estimated in two ways: at the Ajwara study s i t e I counted t o t a l numbers of kob present at various times throughout 63 the day on a 0.5km2 grassy meadow; at a l l other study s i t e s I estimated population density from 5 km transects that bisected each study s i t e . Two observers stood in the back of a moving vehicle, counted animals seen within a 100m s t r i p either side of the vehicle, and recorded a l l observations at 0.1km interv a l s into a tape-recorder. The driver maintained a straight course and c a l l e d out 0.1km intervals to the observers. S t r i p width was estimated during transects, following c a l i b r a t i o n with a range-finder at the outset of each transect. Observers were consistently able to estimate the 100m s t r i p width to within 10m accuracy. High population densities made the use of a range-finder impractical during transects, but bias introduced by inclusion of animals from outside the legitimate s t r i p width was l i k e l y to be inconsequential. In order to minimize bias introduced by diurnal movement patterns (see Chapter 2), a l l transects were performed in the early morning (07:00-09:00). On one occasion, I performed 4 re p l i c a t e transects throughout the day, to test the repeatability of the density estimates. Since t o t a l s varied r e l a t i v e l y l i t t l e between these replicates ( c o e f f i c i e n t of variation = 0.28), I assume that transects were a r e l i a b l e index of kob population density. Carcass density Carcass density was determined in two ways. At Ajwara, t o t a l counts of carcasses were made over a 0.5km2 area by a team of four observers on foot, separated by distances of 40 meters. 64 Thus, each observer was responsible for searching a s t r i p 20 meters to either side. A l l carcasses encountered were marked with a numbered s t r i p of red tape around a horn or limb. Of the marked adult carcasses, 61% were subsequently recovered 6 weeks following i n i t i a l observation, while 39% of carcasses of calves were recovered, suggesting that while many carcasses remained from survey to survey, mortality estimates were biased downwards by carcass disappearance or mutilation by scavengers. At a l l other study s i t e s , carcasses were counted during 5km transects, by a two observer team from the back of a moving vehicle. In a l l cases, carcasses were counted along the same transect l i n e s used for l i v e population density estimates. I measured distance to each carcass using a tape measure, and a l l carcasses located outside 50 meters were disregarded in subsequent analyses. Weekly mortality rates (m) were estimated by: m = c / n t where, m=finite weekly mortality rate c=number of carcasses/km 2 since previous survey n=mean kob population density t=time in weeks 65 Sex and age d i s t r i b u t i o n At regular intervals throughout the 1982 and 1983 dry seasons, I recorded sex r a t i o , calf/female r a t i o , and proportions of yearling and 2-year old males from large groups of kob. Sexes were ea s i l y distinguished since mature males have dark brown to ebony pelage and lyre-shaped horns, while females are hornless and tawny brown in color. Yearling males have short straight horns as long as the ears (Buechner 1974) and are s l i g h t l y smaller than females. Two-year old males are larger than females (Fig. 6.3) and have s l i g h t l y curved horns two to three times as long as the ears (Buechner 1974). Both yearling and 2 year old males are similar in color to females. I estimated age structure of the population older than 2 years from a random shot sample of 38 individuals, since I could detect no v i s i b l e differences between adults of d i f f e r e n t age. The usual hunting protocol involved selecting a group of kob from which to sample, slowly approaching in a vehicle, and then shooting the nearest animal of the sex I wanted. High population densities and ease of approach made sampling unusually simple. Under the circumstances, I consider i t unlikely that hunter preference or avoidance behavior by certain sex or age classes greatly biased the r e s u l t s . I estimated the ages of 155 kob (from both the shot sample and from carcasses) by counting the cementum annuli from f i r s t i n c i s o r s and f i r s t molars. 66 Tooth cementum l i n e s Whole teeth used for counting cementum annuli were f i r s t fixed in 10% formalin for f i v e minutes prior to shipment from the f i e l d . They were then d e c a l c i f i e d in 5% n i t r i c acid u n t i l they reached a rubbery consistency (3-5 days) and immediately cut into 10-15 micr thick longitudinal sections through the pulp cavity, using a freeze microtome. After a i r drying for 24 hours, teeth were stained in Harris haematoxylin for 4 minutes, rinsed in tap water for 4 minutes, bathed in Scott solution for 3 minutes, and f i n a l l y rinsed in tap water again for 4 minutes. After a i r drying for 24 hours, s l i d e s were then mounted in Permount. I examined tooth sections under a compound microscope at both 10x and 40x magnification. The three clearest sections out of at least 9 sections per tooth were counted and recorded. A mean annuli count was determined for each tooth and rounded to the nearest whole number. Ring counts from yearling kob indicate that the f i r s t cementum rest l i n e i s l a i d down during the dry season following b i r t h , thereafter followed by a single l i n e per year. Cementum l i n e counts required some degree of judgement, since l i n e s were frequently d i f f i c u l t to di s t i n g u i s h . In order to assess the r e l i a b i l i t y of the technique, I recounted cementum l i n e s from a random 20% sub-sample of the i n i t i a l 155 individuals that were aged. Close c o r r e l a t i o n between the two counts (r 2=0.85, p<0.05) indicated that counting c r i t e r i a remained consistent throughout the study. 67 Post-mortem examination A l l shot animals and fresh carcasses encountered along transects were subjected to post-mortem examination. I weighed the t o t a l carcass, rumen (with and without contents), and the remaining alimentary tract using a spring balance. I measured head-tail length, heart g i r t h , shoulder height, and t i b i a length. I also measured the horn length and distance between horn t i p s for a l l males. Body condition in ungulates i s commonly assessed by measurement of the fat content of bone marrow ( S i n c l a i r and Duncan 1972; Hanks 1981; Riney 1982). I estimated the fat content of femur bone marrow using a modification of standard procedures (Hanks 1981; Riney 1982). F i r s t , I extracted 5-10 gm of bone marrow from the center of the femur. This sample was weighed in an aluminum tray prior to being heated for a set time over hot coals. This procedure evaporated a l l the water present in the bone marrow sample, leaving a residue of fat and protein. Since mobilized marrow fat i s replaced by water, the r a t i o of post-heating weight over pre-heating weight r e f l e c t s the bone marrow fat content. Preliminary t r i a l s indicated l i t t l e subsequent weight change following the f i r s t 5 minutes of heating; thereafter 5 minutes was maintained as a standard. 68 Results Age d i s t r i b u t i o n in the 1ive population The d i s t r i b u t i o n of ages in the 1983 l i v e population i s shown in F i g . 3.1. Data were pooled from both sexes, due to the small sample size. Individuals 5 to 7 years formed the largest age classes. No indivi d u a l s older than 10 years were found in the shot sample, although in d i v i d u a l s up to 13 years old were observed among carcasses. Frequency of the 3 year age class was somewhat smaller than adjacent age classes. Recruitment of young to age 1 was approximately 10% of the t o t a l population in both 1982 and 1983. Age d i s t r i b u t i o n at death The age d i s t r i b u t i o n of carcasses (Fig. 3.2) was somewhat similar in shape to the standing age d i s t r i b u t i o n . The most frequent age classes were from 5 to 7 years. However, subadults and young adults (1-4 years) were less common among carcasses than in the l i v e population, while older age classes were more common. Age d i s t r i b u t i o n at death d i f f e r e d between sexes (Fig. 3.3, -X.2 = 15.0, p<0.0l). A greater proportion of 4 to 7 year old males died than similar aged females. Conversely, smaller numbers of old males were found than old females. These differences suggest that males suffered s i g n i f i c a n t l y greater 69 24 20 C 16 + O D 12 + a o a jjo 8 -4 • live population 6 7 age 8 9 10 1 1 1 2 13 Figure 3.1 Age structure of the 1983 live population (1 year and older). Frequencies of individuals aged 1-3 estimated directly from observations of herd composition. Frequencies of age classes 3 years and older estimated from a random shot sample. Data f rom h>oth sexes pooled (n—38) • 24 carcasses 20-C 16 O CO a o a 8 -4 -o 1 ••I 1 1i " * 2 3 4 5 6 7 8 9 10 1 1 1 2 1 3 age Figure 3.2 Age structure of the found carcass sample (1 year and older). Data from both sexes pooled (n=10 3). 20 12 • 4 -1 2 -20 4-5 6-7 8 - 9 age classes 10-11 Figure 3.3 Age frequencies of male vs. female carcasses (males=50 females=45). Female total was adjusted to equal male total for 2 s t a t i s t i c a l comparison ( X =15.0; p<0.05) . 72 risk of mortality as young to middle-age adults than females. Age-specific mortality The age-specific mortality, or q x , curve i s a useful means of i l l u s t r a t i n g differences in v u l n e r a b i l i t y to potential mortality agents according to age (Caughley 1966). The Boma kob q x curve indicates high mortality of youngsters, r e l a t i v e l y l i t t l e mortality of young to middle-aged adults, and increasing mortality for older age groups (Fig. 3.4). This general form is consistent with previous published mammalian q^ curves (Caughley 1966). The c u r v i l i n e a r increase in q with increasing age appears more similar in form to published curves for domestic sheep, African buffalo, waterbuck, and d a l l sheep (Deevey 1947; Hickey 1963; S i n c l a i r 1977; Spinage 1982) than the linear increase seen in thar (Caughley 1966). Sex r a t i o The sex r a t i o in the l i v e population (excluding calves) in both 1982 and 1983 was skewed in favor of females (males=2550,females=3434; X 2 = 1 3 0' 8' P<0.001). If we assume an equal sex r a t i o at b i r t h , then mortality must d i f f e r between the sexes. Carcass to t a l s during the dry season indicated that more males died than females. However, these results may have been biased since there was some evidence that male carcasses were more l i k e l y to be seen during transects. The frequency of both ca l f and female carcasses declined s i g n i f i c a n t l y at increasing 7 3 age Figure 3 . 4 Kob age-specific mortality curve, from pooled data for both sexes. Age-specific mortality rates were calculated using equation 1, as explained in Chapter 6 . 74 distance from the vehicle (Fig. 3.5)(females: %2=]0.9, p<0.05; calves: "Y_2 = 96.8, p<0.05), while the number of male carcasses observed at d i f f e r e n t distances from the vehicle varied l i t t l e ( "X2 = 5.9, n.s.). This suggests that males were more v i s i b l e than females and calves during carcass transects, and that observed carcass totals for females and calves were biased downwards. In order to account for t h i s s i g h t a b i l i t y bias, I multiplied carcass t o t a l s for youngsters and females by correction factors of 2.99 and 1.77 respectively. These correction factors were derived by assuming that during transects a l l individuals less than 10m from observers were seen, consequently t o t a l s from greater distances should be equivalent in frequency to observations from,the 0-10m range. After applying the correction factor, there were s t i l l s i g n i f i c a n t l y more male and less female carcasses than one would expect from the proportion of each sex in the l i v e population (males=85, females(adjusted)=94; -X2 = 4.18, p<0.05). In addition, circumstantial evidence suggests that the skewed sex r a t i o in the l i v e population may also result from p r e f e r e n t i a l hunting of kob males by l o c a l tribesman. During 2 communal hunts that I observed, s i g n i f i c a n t l y more males and less females were taken than expected (males=l34, females=72, X 2 =54.8, p<0.001). These results suggest that the skewed sex r a t i o in the l i v e population results from d i f f e r e n t i a l natural and hunting mortality between sexes. 60 40 J 20 CO C 60 males n= 69 CO > 40 © CO O 20- | "ca o females n s 39 c£ 60-1 40 20 10 20 30 40 distance (m) 50 calves n = 99 Figure 3.5 Numbers of carcasses observed at increasing distance from observers. 76 Seasonal changes in body condition Several previous studies indicate that when ungulates are unable to maintain adequate food intake t h e i r fat reserves decline (Hanks 1981). Animals f i r s t use up t h e i r subcutaneous fat reserves, before mobilizing bone marrow fat at advanced stages of emaciation ( S i n c l a i r and Duncan 1972; Hanks 1981). Body condition, as measured by bone marrow fat content, declined dramatically during the course of the 1983 dry season (Fig. 3.6b; Kruskal-Wallis, H=16.4, p<0.05). Males remained in comparatively better condition than females (Mann-Whitney, U=90, p<0.05), but both sexes showed similar rates of decline in body condition. Bone marrow fat content in carcasses compared to shot individuals suggests that individuals in poorest body condition were also the most l i k e l y to die. Decline in body condition probably r e f l e c t s submaintenance supplies of high protein food; thus, stressed individuals were forced to mobilize body fat reserves to meet metabolic requirements. Indices of body condition in the dry season d i f f e r e d between years. In 1982 (Fig. 3.6a), body condition indices did not decrease to l e v e l s seen in 1983 (Fig. 3.6b). In 1982, unusual amounts of r a i n f a l l (over 70mm) f e l l in late February, during the middle of the dry season. This r a i n f a l l produced a temporary flush of green growth over large areas in the kob dry season range (Chapter 2). Increased abundance of green grass resulting from this unusual dry season r a i n f a l l in 1982 may have increased kob green forage intake, preventing the precipitous decline in fat reserves seen in 1983. Figure 3.6a Dry season measurements of bone marrow fat content (1982). Data combined for both sexes. Solid line connects monthly mean values. 78 Figure 3.6b Dry season measurements of bone marrow fat content (1983). Data combined for both sexes. Solid line connects monthly mean values. 79 Seasonal changes in mortality rates Weekly mortality rates for d i f f e r e n t l o c a l i t i e s and periods during the dry seasons of 1982 and 1983 are given in Table 3.1. During the 1982 dry season, adult mortality rates, expressed as a proportion of the i n i t i a l adult population, varied between 0.0030 and 0.0044 per week, while 1983 values ranged from 0.0027 to 0.0195 per week. Calf mortality rates varied from 0.0070 to 0.0779 per week in 1983. Thus, there was a four f o l d difference in weekly mortality rates between calves and adults. Adult mortality rates in the dry season d i f f e r e d between years. The mean weekly mortality rate in the 1982 dry season (0.0039) was s i g n i f i c a n t l y lower than the mortality rate in 1983 (0.0076)(Table 3.1; Mann-Whitney, U=7, p<0.05). This is consistent, with the hypothesis that the dry season produces s t r e s s f u l conditions that are r e f l e c t e d in elevated mortality rates. The unusual dry season r a i n f a l l of 1982 caused increased production of green grass over much of the dry season range, which led to improved body condition and decreased adult mortality rates. In 1983, mortality rates of adults increased s i g n i f i c a n t l y over the course of the dry season (Table 3.1; Mann-Whitney, U=6, p<0.05). Data from the most intensively monitored study s i t e , Ajwara, i l l u s t r a t e t h i s trend (Fig. 3.7). Adult mortality rates at Ajwara increased by a factor of 4 during the course of the dry season, at an apparently constant rate of increase. Estimates of c a l f mortality suggest a similar trend although not s t a t i s t i c a l l y s i g n i f i c a n t (Table 3.1, Mann-Whitney, U=3, n.s.). 80 Table 3.1 Dry season weekly mortality estimates (1982 and 1983). Mean adult mortality rates differed between years (Mann-Whitney; U=7; p<0.05), but there was no significant difference between calf mortality rates in the 2 years. location date adult weekly mortality rate calf weekly mortality rate Gom Gom Wangchira Wangchira Neubari Neubari Ungwala Ungwala Ajwara Ajwara Ajwara Feb.-March 1983 March-April 1983 Jan.-Feb. 1983 Feb.-March 1983 Jan.-Feb. 1983 Feb.-March 1983 Jan.-Feb. 1983 Feb.-March 1983 Jan. 1983 Feb. 1983 March-April 1983 x SE 0.0032 0.0068 0.0061 0.0195 0.0027 0.0054 0.0061 0.0060 0.0054 0.0081 0.0143 0.0076 0.0032 x SE 0.0155 0.0627 0.0650 0.0741 n.a. n.a. n.a. n.a. 0.0070 0.0779 0.0634 0.0522 0.0245 Ajwara Ajwara Gom Ungwala Jan.-Feb. 1982 March-April 1982 Jan.-Feb. 1982 Jan.-March 1982 x SE 0.0030 0.0039 0.0043 0.0044 0.0039 0.0009 x SE 0.0206 n.a. 0.0508 n.a. 0.0357 0.0107 y=0.00106x+ 0.0026 0.016 dry season residence time (weeks) Figure 3.7 Dry season adult mortality rates at Ajwara (1983). Means indicated by solid dots, 95% confidence intervals indicated by vertical 2 bars (r =0.98; p<0.05) . 82 Calf mortality at the Ajwara study s i t e increased rapidly during the dry season (Fig. 3.8). What proportion of annual mortality occurs outside the dry season? I was unable to measure mortality rates d i r e c t l y during the wet season. However, simple c a l c u l a t i o n s suggest that mortality during the wet season was probably much lower than that during the dry season. The average mortality rate of adults during the 1982 and 1983 dry seasons was 0.0039 and 0.0076 per week. If one applies these mortality rates over the entire year, then, at the 1982 rate, t o t a l mortality would have been 18%, and at the 1983 rate, t o t a l mortality would have been 33%. Since recruitment of young into the population was approximately 10% per year, annual adult mortality greater than 10% would have produced a decline in population numbers. Since censuses show no evidence of population decline over th i s period (Table 2.2), I argue that mortality rates were probably substantially higher during the dry season than during the remainder of the year. Dry season mortality in r e l a t i o n to food intake Food intake during the 1983 dry season was greater at Gom than Ajwara (Table 3.2). Estimated mortality of both adults and calves was higher at the location with lower food a v a i l a b i l i t y , although t h i s inference could not be s t a t i s t i c a l l y tested due to the limited number of replicates (Ajwara n=3; Gom n=2). While inconclusive, these results provide additional supporting evidence that mortality rates during the dry season were related 0.10-, Figure 3.8 Dry season calf mortality rates at Ajwara (1983). Means indicated by solid dots, 95% confidence intervals indicated by vertical 84 Table 3.2 Dry season mortality in relation to green grass consumption rates at Gom and Ajwara study sites. location grass consumed (kg/ind/day) adult mortality calf mortality n Gom x 1.27 SE 0.07 0.0050 0.0013 0.0391 0.0167 Ajwara x 0.71 SE 0.26 0.0093 0.0022 0.0494 0.0177 85 to food a v a i l a b i l t y . Cumulative dry season mortality A linear increase in weekly mortality rates of adults implies a c u r v i l i n e a r increase in t o t a l adult deaths as the dry season progresses. The regression l i n e drawn through estimates of weekly mortality rates from the Ajwara study s i t e means that the number of adult deaths r e l a t i v e to i n i t i a l numbers (y) was related to the dry season residence time (t) by: dy = 0.00l06t + 0.0026 dt so that y = 0.00053t2 + 0.0026t + c where y= adult d e a t h s / i n i t i a l adult population size t= time in weeks c=0.0, by d e f i n i t i o n The r e s u l t i n g plot of t o t a l adult mortality as a function of dry season duration i s shown in F i g . 3.9. Note, that while t o t a l adult mortality changed r e l a t i v e l y l i t t l e over the f i r s t 5 weeks of the dry season, thereafter mortality rapidly escalated. The exact parameters of t h i s model should be strongly affected by year to year differences in both kob population density and habitat conditions. However, as long as there i s a constant increase in weekly mortality rates as the dry season progresses, a c u r v i l i n e a r increase in t o t a l population mortality necessarily y = .00053x2 + .0026x dry season duration (weeks) Figure 3.9 Cumulative adult mortality over the 1983 dry season, us Ajwara data. 87 follows. This c u r v i l i n e a r increase in cumulative adult mortality may have resulted from increasing v u l n e r a b i l i t y to mortality agents as the dry season progressed, due to decline in body condition. Let us assume that at a given time the body condition of individuals within a population has a bell-shaped frequency d i s t r i b u t i o n ; that i s , some individuals are in much better condition and some individuals are in much poorer condition than the bulk of the population. I f , as shown in F i g . 3.6b, mean body condition of the population declines steadily during periods of food scarcity, then an increasing proportion of the population would become vulnerable to mortality agents, due to their weakened condition (Fig. 3.10). The vulnerable portion of the population would increase at an increasing rate, as in Fi g . 3.9. This hypothesis predicts that cumulative adult mortality should increase c u r v i l i n e a r l y u n t i l 1/2 of the population has died (assuming a symmetrical d i s t r i b u t i o n of body condition), thereafter cumulative adult mortality would show a diminishing rate of increase. Note that the hypothesis does not demand that food supplies become increasingly scarce, only that available food i s less than the maintenance requirements of individuals within the population. Unfortunately, there are no data to test the hypothesis adequately. Catastrophic d i e o f f s of ungulate populations have been documented a number of times (Christian et a l . 1960; Klein 1968; Child 1972; Keep 1973) but none of these studies recorded mortality at intervals over the period of demise. Hillman and 88 O c CD 3 rj body condition index Figure 3.10 Graphical representation of the body condition/mortality hypothesis. As the body condition of individuals within a population declines during periods of food scarcity (T=l to 3), an increasing proportion of the population becomes vulnerable to mortality agents (shaded portion under the curve). 89 Hillman (1977) estimated mortality at monthly intervals during an extended drought in Nairobi National Park (Fig. 3.11). Their results show the predicted c u r v i l i n e a r increase in cumulative adult mortality for kongoni and wildebeest, and possibly for zebra. Discussion Causes of mortality during the dry season I now consider the three most obvious sources of kob mortality during the dry season: predation, disease, and malnutrition. Predation Predation by wild carnivores was probably inconsequential over most of the Boma kob dry season range. While a variety of carnivores occur in the area, including l i o n , hyena, leopard, and cheetah, their numbers were low. Leopard and cheetah sightings were confined to forested areas near the Boma Plateau. Kob used these areas during the southward migration, but only for a few days each year. Lion and hyena occurred in savannah areas frequented by kob during the dry season. However, in most locations, l i o n and hyena were rarely seen or heard. Social systems of l i o n and hyena preclude migration with the herds, because of the need to feed non-mobile young (Schaller 1972; (from Hillman and Hillman 1077) ' T 1 1 1 1 r-0 2 4 6 dry season length (months) Figure 3.11 Cumulative adult mortality of kongoni, wildebeest, and zebra during a drought in Nairobi National Park, Kenya (from data in Hillman and Hillman 1977). 91 Hanby and Bygott 1979). Resident prey numbers were low in the northern grasslands, possibly because these areas become flooded each wet season. This suggests that Boma predators were rare because of a low biomass of prey available on a year-round basis. An alternative hypothesis is that carnivores are themselves a c t i v e l y persecuted by l o c a l tribesmen. This mortality i s probably small, since l o c a l p a s t o r a l i s t s tolerate carnivores unless they actually k i l l livestock or people (J.Arenson, pers. com.). The low carnivore density means that predation on kob was probably i n s i g n i f i c a n t . Predation of kob by man may be more substantial. During migration, 5000 kob were k i l l e d in communal hunts near Pibor Post by Murle tribesmen in 1982 and 1983. In many years, however, the migration does not come close enough to Murle settlements for successful communal hunts. During the course of the dry season, subsistence hunting by both Murle and Anuak t r i b e s occurred on a regular basis. There were 6 major v i l l a g e s with approximately 500 persons each that were in close contact with kob herds. Assuming that an adult kob would provide 20kg of meat, that each person ate 1kg of meat per day, and that kob hunters replenished th e i r larders as needed, over a 3 month dry season perhaps 15,000 kob might have been slaughtered. In the wet season, kob were largely isolated from subsistence hunters, so offtake was probably minimal. Thus, I estimate that no more than 20,000 kob were harvested per year by human predators. By comparison, results indicate that at least 44,000 and 84,000 adult kob died of natural causes 92 during the respective dry seasons of 1982 and 1983. Thus, I estimate that 2-4 times as many kob died from natural causes as died from hunting. Two points need to be stressed about hunting mortality. F i r s t , i t i s unlikely that the substantial mortality observed in the f i e l d was affected by hunting. Carcasses were only counted when there was no evidence of injury or mutilation by hunters. Second, i t i s f a i r to assume that subsistence hunters take r e l a t i v e l y constant numbers of kob per year. There is no means of adequately preserving meat, there is no commercial trade in meat, and a l l hunt k i l l s are for subsistence purposes. It is r e l a t i v e l y easy to k i l l kob, because of their sheer abundance and behavior. Thus, i t i s l i k e l y that there i s l i t t l e density dependent predation by man. Di sease There was no evidence that mortality during the dry season resulted from unusual disease occurrence. Post-mortem examinations of shot individuals showed no unusually high parasite loads or ubiquitous disease symptoms. There are comparatively few diseases of s u f f i c i e n t pathogenicity and high transmission rates to have large-scale e f f e c t s , the few that do include bovine tuberculosis, pneumonia, rinderpest, and anthrax (L. Karstad, pers. com.). Since a l l study s i t e s showed similar increases in mortality rates over the dry season, i t i s unlikely that widely separated kob herds would have been equally affected by disease outbreak. 93 Many pathogens are r e l a t i v e l y harmless unless the n u t r i t i o n a l condition of an affected individual declines to a c r i t i c a l l e v e l (Scrimshaw et a l . , 1968). Thus, disease may act as the proximate cause of death, when n u t r i t i o n a l i n s u f f i c i e n c i e s are ultimately responsible for making individuals vulnerable. Age structure of the kob population The age d i s t r i b u t i o n of the kob population in 1983 showed a pronounced bulge in the 5-7 year old age groups. If a population has reached a stationary age d i s t r i b u t i o n , the frequency of any given age cl a s s must be equivalent to or smaller than preceding age classes. Since 5 to 7 year old individuals were markedly more common than individuals 1 or 2 years old, this implies that the population must have undergone sizeable change previous to 1983. This bulge in the age d i s t r i b u t i o n could have arisen from increased recruitment of young in 1976 to 1978, with a decrease in subsequent years. A l t e r n a t i v e l y , i f the population decreased between 1978 and 1983 due to increased r e l a t i v e mortality of both calves and old adults, t h i s would produce an increased r e l a t i v e proportion of young adult age classes in the surviving population. Without further information we can not discriminate between these hypotheses (Caughley 1977). Much of the southern Sudan experienced drought conditions in 1979-80, immediately preceding t h i s study. As a res u l t , many 94 usually r e l i a b l e watering points dried up and p a s t o r a l i s t s in the Pibor township west of Boma National Park reported heavy c a t t l e losses due to inadequate forage and lack of water (J.Arenson, pers. com.). Large numbers of dead kob near drying waterholes were reported by independent observers in February 1980 (G. Schaller and D. Western pers. com.). I estimated that there were over 80,000 kob carcasses during an a e r i a l census in March 1980, late in the drought. A e r i a l censuses suggest no subsequent change in kob population numbers since March 1980, (Table 2.2; t - t e s t , n.s.). Thus, i t is l i k e l y that a sizeable reduction in kob numbers resulted from drought conditions in 1979 and 1980. The white-eared kob q x curve indicates that mortality risk is highest for both very young and very old kob. Under s t r e s s f u l environmental conditions these may be the most heavily affected age groups. Therefore, I suggest that the bulge in the standing age d i s t r i b u t i o n observed in 1983 was due to increased mortality of both calves and older age classes during the 1980 population decline rather than a temporary increase in c a l f survival or b i r t h rates from 1976 to 1978. Current frequencies of 5 to 7 year old age classes suggest a mean minimum recruitment in 1976-1978 of 145,000 yearlings per year (range=128,000 to 179,000). If one assumes that prior to 1980 yearlings s i m i l a r l y composed 10% of the t o t a l population (excluding calves), then t h i s implies that the kob population prior to 1980 was about 1.5 m i l l i o n (Fig. 3.12). The 1980 drought may, therefore, have caused a 40% reduction in kob 95 Figure 3.12 Kob population estimates 1979-1983. Population estimates after 1979 from aerial censuses. 1979 population estimate as explained in the text. 96 population numbers. Evidence for food l i m i t a t i o n My results support the hypothesis that food a v a i l a b i l i t y l i m i t s the Boma kob population. At the beginning of this chapter I outlined a series of predictions to test the hypothesis. The f i r s t prediction of the food hypothesis i s that survival rates should decline during years of below-average food abundance and during seasonal periods of food s c a r c i t y . There was evidence that adult mortality rates varied considerably from year to year, depending on the severity of dry season conditions. The short dry season of 1982 produced considerably less adult mortality than the long dry season of 1983. I observed no s i g n i f i c a n t differences between years in juvenile mortality or recruitment rates. Moreover, large-scale mortality during the 1979-80 drought i s consistent with the food l i m i t a t i o n hypothesis. During drought conditions, grazers are forced to concentrate near the few remaining waterholes and consequently can not obtain adequate food supplies (Corfield 1975). Thus, in many droughts, herbivores die of inadequate food rather than lack of water. Mortality rates of both adults and calves increased during the dry season. As shown in Chapter 2, food intake during the dry season was i n s u f f i c i e n t to meet kob energetic requirements. Many previous studies have suggested that the dry season is the period of greatest n u t r i t i o n a l stress in African savannah 97 grasslands (e.g. Western 1975; S i n c l a i r 1977). In one of the few studies that have actually measured both food a v a i l a b i l i t y and mortality rates, S i n c l a i r et a l . (1985) found that dry season mortality in the Serengeti wildebeest was closely correlated with food a v a i l a b i l i t y . The second prediction of the food hypothesis i s that there should be an inverse c o r r e l a t i o n between mortality rates and body condition of herbivores. I found dir e c t correspondence between body condition and elevated dry season mortality rates in the Boma kob population, which suggests that i n s u f f i c i e n t supplies of high protein food led to declining fat reserves, ultimately increasing the risk of mortality. Individuals in the poorest condition were also the most l i k e l y to die. Similar findings from African buffalo, wildebeest, and topi suggest th i s cause-effect relationship i s common among African ungulates (Duncan 1975; S i n c l a i r 1975; S i n c l a i r 1977). The t h i r d prediction of the food l i m i t a t i o n hypothesis i s that mortality rates should be correlated with food intake. Mortality rates d i f f e r e d between the 2 study s i t e s for which I was able to estimate rates of food intake, but due to the small sample sizes, a s t a t i s t i c a l comparison was not possible. These observations suggest that dry season food a v a i l a b i l i t y l i m i t s kob population numbers. Both adult and c a l f mortality rates increased during the dry season, but adult mortality was more c l e a r l y related to food a v a i l a b i l i t y than was c a l f mortality. A limited number of previous studies (Klein 1968; Grubb 1974; S i n c l a i r 1977; S i n c l a i r et a l . 1985) have 98 also found that food a v a i l a b i l i t y determines mortality rates of adults. There i s considerably more evidence that n a t a l i t y and/or c a l f survival are related to food a v a i l a b i l i t y (Woodgerd 1964; Gross 1969; Geist 1971,-Grubb 1974; McCullough 1979; Clutton-Brock et a l . 1982; Houston 1982). White-eared kob calves may be p a r t i a l l y buffered against year to year variation in forage abundance because their mothers continue to provide milk throughout the dry season (Chapter 5). In contrast, a l l of the studies showing density-dependent responses in juvenile survival come from temperate populations that terminate lacta t i o n prior to the winter period of food s c a r c i t y . One of the most intriguing results from t h i s study is the apparently constant increase in mortality rates of adults as the dry season progresses. This would tend to confound any attempt to measure density-dependent mortality e f f e c t s d i r e c t l y . Mortality i s not only affected by the number of individuals competing for scarce resources, but also the length of time individuals are dependent upon sub-maintenance supplies. The longer individuals must rely on their own body reserves to meet a n u t r i t i o n a l d e f i c i t , the lower these reserves become, and the more vulnerable they are to various potential mortality agents. One of the consequences of a d i r e c t r e l a t i o n s h i p between the duration of periods of food s c a r c i t y and mortality rates i s that many large-mammal populations may be prone to large-scale population fluctuations (Klein 1968; Caughley and Lawton 1981; S i n c l a i r and Norton-Griffiths 1979; Bayliss 1985). Populations similar to the Boma kob may be affected substantially by even 99 moderate droughts. Thus, resource l i m i t a t i o n should come into play at frequent intervals (Schoener 1982) and we should expect to see large fluctuations over time in herbivore populations l i v i n g in highly variable environments. 100 CHAPTER 4. KOB REPRODUCTIVE PHENOLOGY Introduction Many ungulate populations show seasonal peaks in ca l f production corresponding to periods of pronounced food abundance (reviews in Sadleir 1969; Schaller 1977; Bunnell 1982; S i n c l a i r 1983a). Since high calf mortality i s common in ungulates (Caughley 1966; S i n c l a i r 1977; McCullough 1979; Fowler 1981; Clutton-Brock et a l . 1982), breeding synchrony may have evolved as a means of improving the survival of vulnerable young. Females supply a l l the food for th e i r newborn offspring through la c t a t i o n and as a result, l a c t a t i n g females frequently suffer diminished fat reserves during the calving period r e l a t i v e to both non-lactating females and males in the population ( S i n c l a i r 1977; Clutton-Brock et a l . 1982). By timing c a l f production to periods of the year when food i s in greatest abundance, females should be better able to meet the demands of feeding dependent young, thereby increasing o f f s p r i n g s u r v i v a l . Thus, there i s presumably strong selective pressure for populations l i v i n g in seasonal environments to produce calves during the period of greatest food abundance. In t h i s chapter, I compare the reproductive phenology of the white-eared kob, Kobus kob leu c o t i s , to the conspecific Uganda kob, Kobus kob thomasi. The migratory white-eared kob l i v e s in savannah grasslands of the southern Sudan with markedly seasonal r a i n f a l l patterns (Chapter 2), while the Uganda kob i s 101 non-migratory and l i v e s in savannah areas with less pronounced seasonality (Buechner et a l . 1966; Modha and Eltringham 1976). In response to climatic seasonality, the amount of green biomass available as food to white-eared kob varies predictably throughout the year (Chapter 2). During the rains, from A p r i l to October, forage i s both abundant and widely d i s t r i b u t e d . However, during the dry season, from November to March, green grass becomes increasingly r e s t r i c t e d to ephemeral swamps of limited extent. As a consequence of resource s c a r c i t y , mortality of both adult and juvenile white-eared kob is pronounced during the dry season (Chapter 3). In Chapter 2 I showed that food a v a i l a b i l i t y i s highest during the late wet season. On the basis of this information, kob should time the production of young to coincide with the late wet season. Methods Seasonal changes in female reproductive condition were determined from a random shot sample. A l l shot females were examined for evidence of lactat i o n (by mammary palpation and measurement of mammary gland wet weight), and pregnancy (by dissection of u t e r i ) . I measured f e t a l t o t a l weight and length, and described the f e t a l developmental stage. Conception dates were subsequently estimated from the regression of f e t a l weight on age reported by Buechner et a l . (1966). B i r t h dates were estimated by extrapolating the 240 day gestation length of the Uganda kob (Buechner et a l . 1966) from estimated conception 102 dates. I also noted the condition of the ovaries, p a r t i c u l a r l y the presence of corpora lutea, based on the description in Morrison (1971 ). Seasonal changes in the proportion of calves (individuals < 12 months old) in the kob population were determined from calf/female ratios in samples of the population observed both on the ground (December to A p r i l ) , and from photographs of migrating herds taken from low-flying a i r c r a f t (November and May). Sample sizes for ground observations ranged from 96 to 867 individuals, while a e r i a l sample sizes ranged from 243 to 339 individuals. Adult males were ea s i l y distinguishable from females because of pronounced color dimorphism and presence of s-shaped horns (Chapter 6). Males 1 to 3 years old were also readily distinguished from females during ground observations due to presence of horns, even though males do not develop dark adult coloration u n t i l after 3 years. It was d i f f i c u l t to distinguish immature males from females on photographs, so I confined calculations to those counts from well defined, close up a e r i a l photos (taken at f l y i n g heights of less than 30m above ground). Calves less than 6 months old are approximately one t h i r d the size of adult females (Chapter 6) and were ea s i l y distinguished both on the ground and from a e r i a l photographs. Diurnal a c t i v i t y budgets of adult kob during the dry season were estimated from large aggregations on swampy meadows. At lOmin i n t e r v a l s , each individual within 40m was recorded as feeding, ruminating, r e c l i n i n g , standing, or other. 1 03 Results Sixteen out of 17 fetuses examined were conceived between January and A p r i l (Fig. 4.1), implying a b i r t h peak between September and December. The b i r t h peak occurred at the end of the wet season, as predicted. Female reproductive condition (Fig. 4.2) ref l e c t e d the observed seasonal trend in c a l f production. In January and February, approximately 80% of females were l a c t a t i n g . Since kob lactate for approximately 180 days (Buechner 1974), Fig. 4.2 suggests that 80% of the females observed in January had produced calves in the preceeding 6 months. The proportion of l a c t a t i n g females declined over the course of the dry season (January to A p r i l ) , reaching approximately 25% by May. In January and February, none of the sampled females were pregnant (Fig. 4.2). By the end of A p r i l , approximately 80% of the females were pregnant. These findings are somewhat inconsistent with the conception period (Fig. 4.1) estimated by the regression of f e t a l weight on age. It i s possible that some of the females sampled in January and February might have had small fetuses that were undetected. A l t e r n a t i v e l y , the sampling procedure may have been biased such that non-pregnant females were more l i k e l y to be sampled than pregnant females. In any case, both Figs. 4.1 and 4.2 indicate that most conceptions occurred between January and May. Approximately 80% of the female population conceived each year, and only one offspring was produced per female. Females were capable of conception as yearlings (2 females of age 2 were 104 co o c CO cr co E E 6-4 -2-0 200 = 100-CO c CO 1 h N 0 month Figure 4.1 Upper histogram - frequencies of b i r t h s by month, estimated from f e t a l weight (n=17). Lower histogram - mean monthly r a i n f a l l t o t a l s f o r Pachalla township (1953-1962). 80 -40 • 0 40 -80 J n=5 F M A n=13 n=9 n=ll month M n=4 Figure 4.2 Female reproductive condition (January-May) from pooled data (1982 and 1983). Monthly sample sizes as indicated below histograms. 1 06 sampled and both were la c t a t i n g , indicating that they conceived as yearlings). I observed no instances of multiple fetuses, in accordance with records from Uganda kob (Buechner et a l . 1966). Like the Uganda kob (Buechner 1961b) white-eared kob embryos were invariably implanted in the right horn of the uterus. Reproductive females were exposed to considerable n u t r i t i o n a l stress during the l a t t e r stages of the l a c t a t i o n period, as i t extended well into the dry season (Fig. 4.3). Lactating females in the dry season were in poorer body condition (as measured by bone marrow fat content; S i n c l a i r and Duncan 1972; Hanks 1981), than non-lactating females (Mann-Whitney, u=14, p<0.0l). Lactating females were also in poorer body condition than adult males (Mann-Whitney, u=90, p<0.05) despite the fact that females devoted a greater portion of the daylight hours to feeding than did males (Fig. 4.4). These findings indicate that l a c t a t i o n i t s e l f imposed increased demands on fat reserves at a time of year when kob were p a r t i c u l a r l y vulnerable to n u t r i t i o n a l stress (Chapter 3). During the dry season, calves were subject to high mortality: between 1.5-7.5% per week from late January to A p r i l (Chapter 3). Calf mortality in the dry season may have resulted in part from the li m i t e d capacity of females to feed young adequately. Although calves began to forage on their own before weaning, thi s food source was i n s u f f i c i e n t to make up for the loss of mother's milk. Seasonal changes in calf/female r a t i o are shown in F i g . 4.5 (curve A). There was a steady increase in the proportion of Figure 4.3 Dry season body condition of lactating (solid dots) vs. non-lactating (open dots) females (1983 data). Lines connect monthly mean values. 108 100 CD c o cd > CD 0} o 3 100 o o 60 -I 20-I other stand \j iimlnale recline - feed males 8:00 1 2:00 16:00 time of day Figure 4.4 A c t i v i t y budgets of male vs. female kob on meadows (male observations=1025; female observations=1379). Note that males spent 2 s i g n i f i c a n t l y l e s s time feeding than d i d females ("X =269.6; p< 0.001) 109 Figure 4.5 Seasonal changes in calf/female ratio, estimated in two ways. Curve A indicates calf/female ratio observed from f i e l d observations of herd composition (open triangles= 1981-1982; open circles= 1982-1983). Curve B indicates calf/female ratio predicted from frequencies of births by month (Fig. 4.1) and f i e l d estimates of calf mortality rates (Fig. 3.8). Lower histogram - mean monthly r a i n f a l l totals for Pachalla (1953-1962). 1 1 0 females with calves from October to mid-March, and a sharp decline in calf/female ratios from mid-March to May. Since v i r t u a l l y a l l calves were born between September and December (Fig. 4.1), one would expect the calf/female r a t i o to have increased only u n t i l the end of the calving period (December), in contrast to f i e l d observations. This discrepancy may have arisen because new-born kob hide in thickets and t a l l grass for some time, a common behavior in young of many African ungulates (Leuthold 1966). As a result, f i e l d observations of calf/female r a t i o may have been biased downwards prior to the date when a l l calves joined the adult herds following the hiding out period. I calculated an expected seasonal curve of calf/female r a t i o s (Fig. 4.5, curve B) from the frequency of births by month (Fig. 4.1), the fecundity rate among reproductive age females (80%, F i g . 4.2), and f i e l d estimates of mortality rates of calves during the dry season (Chapter 3). The expected curve peaks at the conclusion of the calving period (December) and declines over the dry season to level s close to observed calf/female r a t i o s by the end of the dry season ( A p r i l ) . S i m i l a r i t y between curves A and B in the late dry season "suggest that by A p r i l a l l young have joined the adult herds. Calf mortality was apparently highest in the l a t t e r dry season period, l e v e l l i n g off with the onset of the rains. S i m i l a r i t y between f i e l d observations of calf/female r a t i o (Fig. 4.5) in 1981-82 (open tria n g l e s ) and 1982-83 (open c i r c l e s ) , suggests that n a t a l i t y rates and reproductive timing varied l i t t l e between the two years. 111 Discussion White-eared kob l i v i n g in seasonal grasslands of the southern Sudan showed synchronous breeding r e s u l t i n g in c a l f production in the late wet season (September to December). There was a post-parturition i n t e r v a l of approximately 4 months before females conceived again. As a result, individual females gave b i r t h to a single offspring per year. By contrast, Uganda kob produce calves throughout the year and females conceive within 2 months following p a r t u r i t i o n (Buechner et a l . 1966; Buechner 1974). Consequently, female Uganda kob produce on average 1.35 young per individual per year (Buechner 1974), while white-eared kob females produced 0.80 young per year. Observations of l a c t a t i o n in 2 year old females in t h i s study confirmed Buechner et a l . ' s (1966) view that Kobus kob females are capable of conception as yearlings. Thus, seasonal breeding reduced the annual reproductive potential of white-eared kob r e l a t i v e to Uganda kob. White-eared kob c a l f production occurred in the late wet season, when food was both abundant and widespread. These results are consistent with a large number of studies on other ungulates l i v i n g in seasonal environments (Sadleir 1969; S i n c l a i r 1983a). However, females continued l a c t a t i o n well into the dry season, when food supplies were scarce. As a consequence, l a c t a t i n g females experienced greater depletion of fat reserves than non-lactating females. Since abundant food supplies were available early in the wet season, i t i s therefore puzzling that c a l f production did not occur e a r l i e r in the wet 1 12 season. One possible explanation i s that females require a substantial period to recover fat reserves l o s t during lact a t i o n before giving b i r t h and feeding new young ( S i n c l a i r 1983a). If kob calves were born in the early wet season, females would have just emerged from the s t r e s s f u l dry season, possibly with inadequate fat reserves to sustain lact a t i o n even under abundant food conditions. Thus, the observed timing may r e f l e c t a trade-off between producing young during the wet season and ensuring that females have adequate recovery time prior to the next l a c t a t i o n . This hypothesis could be tested by measuring fat reserves of females throughout the wet season period. If females recover fat reserves quickly after the onset of the rains, t h i s would be inconsistent with the condition recovery hypothesis. A l t e r n a t i v e l y , i t i s possible that there are other selective forces acting on reproductive phenology. The white-eared kob population migrates each year from low r a i n f a l l , short grass areas in the south each wet season to high r a i n f a l l , ephemeral swamped grasslands in the north in the dry season Chapter 2). Calf production takes place largely during the northward migration from October to December. The observed timing of c a l f production may function as an anti-predator adaptation. There are apparently few predators present in the northern grasslands, possibly due to the lack of year round resident prey species (Chapter 3). Thus, c a l f production during migration into the dry season range may reduce the l i k e l i h o o d of 1 1 3 predation on vulnerable young and females. There is a clear way in which the predation hypothesis might be tested. In the Boma ecosystem, several ungulate species have north-south migratory movements similar to the white-eared kob (e.g. tiang, zebra, and eland), while others are non-migratory (e.g. buffalo, lelwel, and o r i b i ) . If kob produce calves during the northward migration as an anti-predator adaptation, then I predict that other migratory species should adopt a similar strategy, while non-migratory species should exhibit calving peaks in the early wet season. Thus, there should be a clear difference in the timing of c a l f production between non-migratory and migratory populations. 1 14 CHAPTER 5. BREEDING SYNCHRONY AND MALE AGGRESSION Introduction Four features characterize lek mating systems (Bradbury 1981): (1) males congregate at communal display s i t e s for the sole purpose of at t r a c t i n g females; (2) males do not care for young; (3) display sites contain no defensible resources of interest to females; and (4) females choose males to mate with from the display s i t e . As a consequence of these conditions, lek systems constitute an extreme form of polygyny and, since female choice depends mainly on the display q u a l i t i e s of males, there is presumably intense sexual selection on males (Emlen and Oring 1977; Borgia 1979; Bradbury and Gibson 1983). Thus, lekking species are p a r t i c u l a r l y well suited to testing hypotheses about the evolution and adaptive function of mate choice (Emlen and Oring 1977; Borgia 1979; Wittenberger 1978; Wrangham 1980; Bradbury 1981; Bradbury and Gibson 1983; Foster 1983). A c r i t i c a l element in lek systems i s that males compete aggressively for occupation of preferred lek t e r r i t o r i e s , or courts. In most lekking populations, matings are largely r e s t r i c t e d to males that are lek members, and s a t e l l i t e males are excluded from breeding a c t i v i t i e s . Moreover, most matings are achieved by holders of a small number of central courts within the lek (Buechner 1961a; Wiley 1973; L i l l 1974, 1976; Robel and Ballard 1974; but see Leuthold 1966; Floody and Arnold 115 1975). As a re s u l t , there is a highly skewed d i s t r i b u t i o n of mating success among males in the population (Emlen and Oring 1977; Borgia 1979; Bradbury and Gibson 1983). Among lekking ungulates, males can control preferred central courts for only a few days, while they can control peripheral courts for several weeks (Buechner et a l . 1966). This difference in s i t e occupancy could arise from central males expending more energy on defense and courting of females (Buechner and Schloeth 1965). There i s some evidence that only a few males have access to central positions because dominant males return repeatedly for brief periods of tenure within leks following temporary periods of association with surrounding bachelor herds (Buechner et a l . , unpublished MS, c i t e d in Bradbury and Gibson 1983). L i t t l e attention has been paid to the effect of synchronized breeding on lek mating systems. Emlen and Oring (1977) suggested that extreme breeding synchrony should diminish the degree of polygyny, since individual males can only obtain a few matings when many females come into breeding condition simultaneously (Fig. 5.1). Pronounced polygyny may be expected at intermediate degrees of breeding synchrony, since individual males would be able to obtain many matings during the short period of tenure in a preferred lek po s i t i o n . Asynchronous populations should be less polygynous, since i t would be d i f f i c u l t for males to retain a dominant position for extended periods of time, providing new opportunities for other males. Therefore, one would predict more polygyny and intense sexual 116 Figure 5.1 Hypothetical r e l a t i o n s h i p between degree of breeding synchrony and degree of polygyny (from Emlen and Oring 1977). 1 1 7 selection in populations with intermediate breeding synchrony. Two predictions arise from this hypothesis. F i r s t , i f sexual selection i s greater in moderately synchronous populations, sexual dimorphism should also be more pronounced. Second, dominant males should obtain greater benefits under conditions favoring polygyny ( i . e . for moderately synchronous populations). Natural selection should favor individuals that are able to obtain greater benefits r e l a t i v e to costs, so competing males on leks should be more aggressive in moderately synchronous populations (Fig. 5.2). In order to test t h i s hypothesis, I examined lek behavior in a migratory population of the white-eared kob, Kobus kob  leu c o t i s , in the southern Sudan. The behavior of the white-eared kob on leks is s i m i l a r to that of the Uganda kob, Kobus  kob thomasi (Buechner and Schloeth 1965), so I s h a l l concentrate on quantitative differences between the two races. The chief difference i s that mating i s r e s t r i c t e d to a 4 month period in white-eared kob (Chapter 5) while Uganda kob breed year-round (Buechner 1974). Thus, a comparison of the populations w i l l test the hypothesis that sexual selection i s more intense in moderately synchronous than in asynchronous lekking populations. This hypothesis predicts a greater degree of sexual dimorphism in the white-eared kob and more intense competition among males. I also investigated the consequences of aggression between males on leks. Previous studies on the Uganda kob suggest only that lek aggression determines dominance relations between males (Buechner 1961a). However, recent work suggests that disruptive 118 Figure 5.2 Postulated male cost/benefit curves i n r e l a t i o n to f i g h t i n g i n t e n s i t y (b^ = benefits to males i n highly synchronized or asynchronous populations; b^ = benefits to males i n moderately synchronized populations; c = costs to males; = f i g h t i n g i n t e n s i t y at which b^-c i s greatest; and f = f i g h t i n g i n t e n s i t y at which b -c i s greatest). 119 behavior i s common in some lekking species (Wrangham 1980; Foster 1983; T r a i l 1985). This suggests an alte r n a t i v e hypothesis, that male aggression may serve to disrupt the mating a c t i v i t i e s of neighboring males. Methods Sexual dimorphism I measured size dimorphism in adult kob from a random shot sample of 18 males and 24 females. Animals were weighed, and a variety of other body measurements were recorded. A non-random (with respect to age) sample of 10 immature males and 11 calves were also taken. These age groups were e a s i l y recognizable by horn shape, horn length, and coat c o l o r . Ages of adults were determined by counts of tooth cementum annuli (Chapter 3). Lek observations Observations of lek behavior were compiled from 8 hours of observations per month for January to A p r i l , 1983. Because kob populations were transient, no leks were active throughout the entire 4 month period. I therefore combined observations from 5 dif f e r e n t leks, each numbering between 20 and 65 males. I made recorded observations for 10 minute periods on a tape recorder, separated by 10 minute i n t e r v a l s during which I transcribed the records onto data sheets. For each observation 1 20 period I counted the t o t a l number of males and females present at the beginning of the i n t e r v a l , and noted the locations of females on the lek. During the observation periods I noted a l l incidences of the following behaviors: (a) mount attempts - a mount in which the penis i s erect and the male actually makes physical contact with the female, although penetration may not occur; (b) sexual s o l i c i t s - male exhibits p r e c o i t a l behaviors such as "prancing", " l i p - c u r l i n g " , or " l e g - s t r i k i n g " (Buechner and Schloeth 1965); (c) threats - males engage in agonistic displays with heads lowered and ears extended, and prepare for complete sparring a c t i v i t y ; (d) chases - male chases another male, usually when a second male moves onto the court of the f i r s t , or as a result of threat or f i g h t behavior; (e) fights males make physical contact with the horns, usually after threat behavior. Only the most escalated agonistic behavior was recorded for each incident. Since v i r t u a l l y a l l fights terminated in a retreat by the loser and chase by the v i c t o r , I simply scored such events as a f i g h t . In addition, I recorded the location of a l l fights and mount attempts within 3 concentric zones of equal width on the lek: c e n t r a l , intermediate, and peripheral. Since leks were roughly c i r c u l a r in shape, the r a t i o s of the r e l a t i v e areas of the three zones were approximately 1:3:5. I used these ratios to calculate expected values for the s p a t i a l d i s t r i b u t i o n of females and of aggressive encounters on the lek. For example, i f females were dis t r i b u t e d at random on the lek, one would predict f i v e times as many females in the perimeter zone as in 121 the central zone. Results Sexual dimorphism Adult white-eared kob d i f f e r markedly from Uganda kob in degree of color dimorphism. Adult leucotis males are ebony colored, with highly contrasting white patches on the face, ears, and throat, and are e n t i r e l y white v e n t r a l l y . By contrast, thomasi males are an even tawny brown color. Females of both races resemble thomasi males in color. The pronounced color dimorphism in leucotis f i r s t appears at 3 years of age, when males are approaching f u l l body size and sexual maturity (Buechner et a l . 1966). There is no difference in weight dimorphism between the two races (Fig. 5.3)( leucotis adult male/female weight = 1.37; thomasi male/female weight = 1.41 (Buechner and Golley 1967) ). White-eared kob are, however, smaller than Uganda kob (mean l i v e weight of leucotis males = 55kg, mean l i v e weight of thomasi males = 90kg). Temporal change in mating and agonistic behavior The number of females observed on leks declined s l i g h t l y between' January and A p r i l (Table 5.1). However, mount attempts per hour did not change s i g n i f i c a n t l y over the 4 month breeding 122 70-, Oi 5 0 Ui CD CD > 30 10 - «' (4) (11) ( 6 ) (4) (3) ~i r birth 1 i 3 (4) ( 2 ) C) ( 6 ) ( 6 ) males females (5) i 5 i 7 age Figure 5.3 Kob l i v e weights at age (males=solid dots; females=open dots). Mean values indicated by dots, range by v e r t i c a l bar, and sample sizes i n brackets. 123 Table 5.1 Changes in mating behavior over the breeding season (January-April, 1983)(mean values per hour, SE of mean shown in parentheses). Differences between means evaluated by Kruskal-Wallis Test (df=3;*:p< 0.05; **:p<0.001) . differences between January February March April means? no. males 28.7(1.7) 26.4(0.9) 51.5(4.2) 18.4(0.7) H=42.5,** no. females 7.6(1.7) 7.0(0.7) 4.8(0.3) 5.3(0.4) H=11.0,* no. mounts 21.1(4.0) 20.6(3.9) 17.1(3.5) 19.6(5.1) H=3.5,n.s. no. fights 2.06(0.33) 0.98(0.17) 0.39(0.07) 0.42(0.03) H=30.7,** per male no. chases 0.64(0.09) 0.26(0.05) 0.12(0.03) 0.13(0.03) H=26.6,** per male no. threats 0.25(0.08) 0.09(0.03) 0.13(0.03) 0.08(0.03) H=4.7,n.s. per male 1 24 season. These results suggest that matings were evenly spread over the entire breeding period, as in the Uganda kob (Buechner 1974). However, the mean number of fights per male per hour did decrease s i g n i f i c a n t l y over the breeding season, ranging from a peak of 2.06 fights per hour in January to 0.39 fights per hour in March (Fig. 5.4). The mean value for the 4 month period was 1.07 fights per hour. When fights are combined with a threat rate of 0.13 per hour, there was a mean agonistic encounter rate of 1.20 per hour over the breeding season, which i s much greater than the value of 0.38 per hour for the Uganda kob, calculated in the same way from Floody and Arnold's (1975) data. Agonistic encounters were therefore three times as frequent in the white-eared kob population as the Uganda kob population. However, the agonistic encounter rate declined to a comparable value by the end of the white-eared kob breeding season. Spat i a l d i s t r i b u t i o n of females and agonistic encounters More than 3 times as many females as expected concentrated in the central area of the display grounds (Fig. 5.5, ^ 2=296, df=2, p<0.00l). This i s consistent with the behavior of female Uganda kob (Buechner 1961b; Floody and Arnold 1975), as well as that of most other lekking species (Bradbury 1981). Although central areas were preferred, and as many as 12 di f f e r e n t females were on occasion located with a single male, females were generally not highly clumped (Fig. 5.6). Clumping of females was even less common away from central or intermediate 125 3.2-, O JC Q) CO E -»-• CO i_ c o CO CO CD i_ O) D) CO c CO CD E 2.44 1 . 6 4 0.8 0.0-J a n Feb Mar date Apr Figure 5.4 Changes in (dots=means, vertical aggressive behavior over the breeding bars=95% confidence intervals). season 126 "D CD 1 6 0 > CD CO o central intermediate perimeter O c CD 3 rj CD 8 0 -0 "D CD O CD CL X CD >» O c CD CT CD 8 0 -1 6 0 -2 4 0 -location of females Figure 5.5 Female spatial distribution on leks. Upper histogram shows observed values, lower histogram shows expected frequencies under the null hypothesis that females choose position at random. 127 1204 80 4 40 4 no. females with single male o c CD 3 cr Figure 5 . 6 Female group si z e s with i n d i v i d u a l courting males. 1 28 areas. A mean of 5.7 females were located on the lek for each observation period, and these females were on average d i s t r i b u t e d among 3.2 d i f f e r e n t males. Agonistic encounters were more frequent than expected on ce n t r a l positions on the lek (Fig. 5.7, ?C 2 = 116.1, df=2, p<0.00l), although males also fought in peripheral areas. Spot checks of male d i s t r i b u t i o n indicated that males were spread out r e l a t i v e l y evenly over the lek (based on expected numbers in the 3 concentric zones). Thus, the increased agonistic encounter rate in the center of the lek did not simply result from more males being present. These results are consistent with Floody and Arnold's (1975) observations of a close co r r e l a t i o n between the number of mating attempts and the number of agonistic encounters by i n d i v i d u a l Uganda kob males. Both behaviors were concentrated in c e n t r a l areas in Uganda kob, as i s the usual case for other lekking species (Bradbury 1981). In addition, fights in central areas tended to be both more prolonged and more energetic than those in more peripheral areas. Functions of f ighting behavior Buechner (1961a) suggested that aggressive behavior in the kob contributes to the establishment of dominance positions on the display ground. Males compete strenuously for possession of central courts and agonistic encounters frequently escalate into f u l l - s c a l e f i g h t s . In 204 of 305 fi g h t s observed (67%), no females were present with either combatant, suggesting that 129 central intermediate 0 > 1 2 0 -i _ 0 CO .Q O >» O c 0 6 0 -c r CD perimeter 0 T 5 0 60 O CD a x 0 O c 0 CJ 0 120 180-location of fights Figure 5.7 Spatial distribution of male fights. Upper histogram shows observed values, lower histogram shows the expected frequencies under the null hypothesis that the number of fights is related to the number of males in zones of different sizes. 130 these fights were concerned solely with deciding the lek positions of the males involved. However, in the other 33% of f i g h t s , at least one female was present in the court of one of the combatants. This suggests that many fights may disrupt mating a c t i v i t i e s of neighboring males. To examine this hypothesis, I compared the r e l a t i v e p r o b a b i l i t i e s of a male being engaged in a fight given that a female was or was not present. If f i g h t i n g was not oriented towards disruption, figh t i n g should be equally l i k e l y in both cases. I calculated the conditional p r o b a b i l i t i e s of a fight occurring, given that a female was or was not present, according to Bayes' rule (Scheaffer and Mendenhall 1975), using the following p r o b a b i l i t i e s calculated from 91 observations periods: A = p(fight) = 0.18 A' = p(no fight) = 0.82 B = p(female present fight) = 0.33 B' = p(female present no fight) = 0.07 C = p(no female present fight) = 0.67 C = p(no female'present no fight) = 0.93 (A)(B) p(fight female present) = = 0.509 (A)(B) + (A')(B') (A)(C) p(fight no female present) = = 0.137 (A)(C) + (A')(C) Thus, the probability of a fight occurring was approximately four times greater i f a female was present than not. This suggests that neighboring males may i n i t i a t e fights in order to either disrupt mating or to steal females. 131 During 14 out of 91 observation periods (15%) females changed location subsequent to f i g h t s . On these 14 occasions, the male that was challenged was courting on average 4.1 females before the fight and 0.9 females after the f i g h t . As a result of t h i s disruptive behavior, a t o t a l of 57 females moved to new locations, while 22 females changed position during the 77 observation periods (85%) characterized by no disruptive behavior. Subsequent movements of females are summarized in Table 5.2. Nearly twice as many disrupted females moved to a neighboring male (40%) than returned to their o r i g i n a l partner (21%), and 39% l e f t the lek e n t i r e l y . The majority (64%) of movements of non-disrupted females resulted in females leaving the lek, while 36% resulted in females moving to neighboring males. These findings indicate that disruptive behavior greatly increased the frequency of movements by females on the lek, although a substantial proportion of females returned to their o r i g i n a l partners. The data are inadequate to test whether the challenging male subsequently mated with females scattered by f i g h t i n g . At worst, however, disrupting males temporarily forced females back into a "pool" from which they might la t e r gain. Consequences of fighting Vigorous competition for access to females may expose males to substantial r i s k s . During 16 hours of observations at leks, I recorded 4 instances of serious wounding during 305 fights; Table 5.2 Movements of disrupted vs. non-disrupted females (proportion total observations in parentheses). returned moved to total no. to origin neighbour l e f t lek females disrupted 12 (0.21) 23 (0.40) 22 (0.39) 57 not disrupted 0 (0.00) 8 (0.36) 14 (0.64) 22 1 33 a l l of these were horn wounds to the sides or abdomen of combatants. Agonistic encounters between 2 males were frequently joined by a t h i r d male, who often directed his i n i t i a l blows to the unprotected flanks and abdomen of the other combatants. Also, 3 out of 5 male carcasses found near leks showed symptoms of recent horn wounds. Moreover, the d i s t r i b u t i o n of ages at death (obtained during the breeding season) d i f f e r e d s i g n i f i c a n t l y between the sexes (Fig. 3.8, X 2 = 15.0, df = 3, p<0.0l)'. A larger proportion of male than female carcasses were young to mid-aged adults, and these are the age classes most active on the display ground (Leuthold 1966). Aggressive competition between males may explain the skewed sex rati o s observed in both white-eared kob (58-62% females) and Uganda kob (59% females; Buechner 1974) populations. Pi scussion Emlen and Oring's (1977) hypothesis that species with moderate breeding synchrony should be more polygynous than either asynchronous or highly synchronous breeders predicts that sexual selection should be greater for the moderately synchronous white-eared kob than the asynchronous Uganda kob. One consequence of increased sexual selection i s increased sexual dimorphism in size and coloration. White-eared kob d i f f e r e d markedly from the conspecific Uganda kob in the degree of color dimorphism, but not in size dimorphism. Dark coloration has no known adaptive value other than male advertisement. Few savannah-dwelling antelope species are 1 34 characterized by extreme melanism, presumably because of the disadvantages of being highly v i s i b l e to predators. Moreover, dark pelage absorbs considerably more solar radiation than l i g h t pelage, making individuals more prone to heat stress in t r o p i c a l c l i m a t i c conditions (Finch 1972). Mean dai l y maximum temperatures in the Sudan study area range from 30-39° C , so the potential for thermal stress i s considerable. Previous studies in hot savannah grasslands suggest strong selection against dark coat color in African c a t t l e herds (Finch and Western 1977). I argue that contrasting black and white coloration of males may serve largely for sexual advertisement. Recent t h e o r e t i c a l work has shown that otherwise non-adaptive t r a i t s of males may evolve, i f they are used by females as a r b i t r a r y mating cues (Lande 1980, 1981; Kirkpatrick 1982). Thus, dark coloration in white-eared kob males may consitute an example of "runaway" sexual selection (Fisher 1958). A second prediction from increased sexual selection i s that white-eared kob males should compete more vigorously than Uganda kob for possession of preferred lek courts. Results from t h i s study suggest that the o v e r a l l frequency of agonistic encounters is s i g n i f i c a n t l y higher for the moderately synchronous white-eared kob than for the asynchronous Uganda kob, supporting the hypothesis. However, fi g h t i n g frequency declined substantially over the course of the breeding season, and by the end of the period the white-eared kob aggression rate was similar to values reported for Uganda kob. The significance of this decline" i s not c l e a r . It may simply r e f l e c t that dominance relations on 1 35 the display grounds are gradually established over the course of the breeding period. A l t e r n a t i v e l y , males may simply have less to gain by supplanting males in preferred courts late in the mating season when most females have already been served, while the risks of injury remain the same. Thus, the most successful male mating strategy may involve f i g h t i n g vigorously early in the mating season when potential f i t n e s s gains are highest, and less vigorously l a t e r on. The data from this study are inadequate to discriminate between these hypotheses. Lekking males incur a high r i s k of mortality as a consequence of intrasexual competition for access to females. I observed several instances of serious wounding, and lekking males frequently attacked the undefended flanks of other males engaged in t e r r i t o r i a l disputes. D i s t r i b u t i o n of ages at death d i f f e r e d s i g n i f i c a n t l y between the sexes, suggesting that strong intrasexual selection may result in predictable changes in age-s p e c i f i c mortality patterns and a skewed sex r a t i o . Clutton-Brock et a l . (1985) have recently suggested that juvenile mortality i s higher for males than females in species with high sexual dimorphism. I argue, in addition, that mortality of reproductive-aged males may be higher than that of females under conditions of strong sexual s e l e c t i o n . This may explain the adult sex r a t i o s skewed towards females frequently observed in ungulates (Cowan 1950; Clutton-Brock et a l . 1982). Aggressive behavior not only influences dominance relations of males on display grounds, but i t also disrupts the mating a c t i v i t i e s of neighboring males. Males that were courting 136 females were much more l i k e l y to be engaged in fights than unattended males. Thus, much of the observed aggression was directed towards successfully mating males. Females were scattered immediately following disruptive f i g h t s , although a substantial proportion of females subsequently returned to their o r i g i n a l male partners. Disruptive f i g h t i n g p e r i o d i c a l l y redistributes females onto new courts, which is advantageous to challenging males. Such disruptive behavior has not been previously described for mammalian leks, although i t occurs in some insect and bird species (Foster 1983; T r a i l 1985). I suggest that disruption may be a sucessful strategy in lekking populations l i k e the white-eared kob with r e s t r i c t e d breeding seasons. In species that breed year-round, peripheral males on leks may do better by waiting for the owners of preferred central courts to weaken, rather than incur the substantial risks of f i g h t i n g . Recent reviews of lek behavior have examined the importance of disruptive behavior in the evolution of lek mating choice systems (Wrangham 1980; Foster 1983). Results from this study suggest that disruptive behavior may be more prevalent than previously assumed, p a r t i c u l a r l y under conditions of increased sexual selection associated with a r e s t r i c t e d mating season as in white-eared kob. 137 CHAPTER 6. AGE-SPECIFIC MORTALITY: AN ALTERNATIVE APPROACH Introduct ion Age-specific survivorship i s basic to many population models used by ecologists ( L e s l i e 1945; Beverton and Holt 1957; Gadgil and Bossert 1970; Schaffer 1974; Stearns 1976; Gulland 1977). Since i t is rarely possible to monitor the fates of true cohorts over their entire l i f e spans (Lowe 1969; Connell 1970; Sherman and Morton 1984), many studies estimate survivorship from samples of the population age structure (Seber 1973; Ricker 1975; Caughley 1977). These i n d i r e c t methods of ca l c u l a t i n g " s t a t i c " l i f e tables assume that the population rate of increase (r) has remained constant for several generations, producing a stable age d i s t r i b u t i o n . Despite the widespread use of these methods, few published l i f e table studies meet adequately the underlying assumptions on which they are based (Caughley 1966). Two of the most frequently used methods estimate age-sp e c i f i c survivorship from (A) the standing age d i s t r i b u t i o n ( s x , the number of age x in d i v i d u a l s in a population at a specified time r e l a t i v e to the number of new-born young), or (B) the standing d i s t r i b u t i o n of ages at death (s x , the r e l a t i v e frequencies of individuals of age x that die over a specified time i n t e r v a l , usually 1 year). For method A, the standing age d i s t r i b u t i o n i s converted to an l x schedule by multiplying s x values by e , to discount the e f f e c t s of the population rate of increase on the observed age d i s t r i b u t i o n . For method B, the standing d i s t r i b u t i o n of ages at death is likewise converted to 138 a 6^  schedule by multiplying the s^ values by e . This d x schedule is then used to calculate the corresponding 1 schedule, and the r a t i o d x / l x i s used to estimate the age-s p e c i f i c mortality rate, . Both methods A and" B rely on the assumption that the population has a stable age d i s t r i b u t i o n , and require precise estimates of population numbers over several generations in order to calculate r. I propose an al t e r n a t i v e method of estimating age-specific mortality rates based on Pielou's (1977) approach, that i s free of these r e s t r i c t i v e assumptions, and compare i t to methods currently in use. The model Simply put, q x values indicate the probability of dying for individuals in each age class over the time interval from t to t+1. Thus, i t is the proportion of individuals of age x that die over a given time i n t e r v a l r e l a t i v e to the number a l i v e at the beginning of the i n t e r v a l . As indicated by Pielou (1977), % = — ' a. where c = number of age x individuals that die during the year, and n = i n i t i a l number of age x individuals. In f i e l d studies i t i s rarely possible to count a l l 1 39 individuals of known age in a population or even to count a l l those that die in a given year. However, these values may be estimated by multiplying the sampled frequencies of each age clas s for both l i v e and dead populations by ove r a l l population estimates. Note that c x = C f x and ry. = N g. where C = N = fx 9x £ c x , =• sampled proportion at age x among carcasses, and = sampled proportions at age x among l i v e i n i t i a l population. Thus k f. Qx = equation 1 where k = C / N . In some f i e l d situations, the i n i t i a l standing age d i s t r i b u t i o n may be unavailable for one reason or another. However, i t i s possible to derive these values from the standing age d i s t r i b u t i o n at the end of the i n t e r v a l . Suppose h x + 1 = sampled proportions at age x+1 among l i v e population subsequent to annual mortality and N' = t o t a l population size at end of i n t e r v a l . Then n x = N'h x + 1 + Cf x and equation 1 can be rewritten as 1 40 C f 'x equation 2 N' h x+l + C f x In order to calculate a q x schedule using equations 1 or 2, one requires an estimate of annual mortality, standing d i s t r i b u t i o n of ages at death, and the standing age d i s t r i b u t i o n either at the beginning or the end of the period in which mortality was monitored. Equation 2 also requires an estimate of population numbers at the end of the observation period. There are no further assumptions of a stable age d i s t r i b u t i o n and constant rate of population increase. To investigate the u t i l i t y of the proposed technique, three approaches were employed. F i r s t , I compared estimates of age-s p e c i f i c mortality derived by the proposed technique to published values ( S i n c l a i r 1977) for the African buffalo, Syncerus c a f f e r . Second, I performed a series of computer simulations, using S i n c l a i r ' s buffalo data, to estimate the sampling v a r i a t i o n resulting from three alternative methods of estimating age-specific mortality. Third, I tested the accuracy of the three methods when the underlying assumption of a stable age d i s t r i b u t i o n i s violated, by c a l c u l a t i n g age-specific mortality following a temporary perturbation of a hypothetical African buffalo population. Methods 141 Calculation of age-specific mortality S i n c l a i r (1977) derived a l i f e table for female buffalo in the Serengeti ecosystem from a found sample of 246 carcasses 2 years and older. Survival of calves (age 0 to 1) and yearlings (age 1 to 2) was estimated independently by counts of these age classes per 100 females from a e r i a l photographs ( S i n c l a i r 1977). At the time of the study the population was increasing at a constant rate (r = 0.077) , so S i n c l a i r calculated cL^  by multiplying the standing d i s t r i b u t i o n of ages at death by e r x to incorporate the effect of population increase. The resultant q x schedule is shown in F i g . 6.1. S i n c l a i r also shot a random sample of 80 females 2 years and older over the same three year period (1967-1969) that the s k u l l sample was col l e c t e d ( S i n c l a i r unpublished). The ages of the shot sample and the carcass sample were determined from counts of tooth cementum annuli. The standing age d i s t r i b u t i o n from the l i v e population was converted to a g x d i s t r i b u t i o n by dividing the frequency of each age group in the shot sample by the t o t a l sample s i z e , as shown in Table 6.1. The standing d i s t r i b u t i o n of ages at death was converted to an f x d i s t r i b u t i o n by d i v i d i n g the number of carcasses of age x in the found sample by the t o t a l number of carcasses (246). For subsequent calculations of age-specific mortality, I smoothed both the and the g^  d i s t r i b u t i o n s by taking a three point running average for sequential age classes. I estimated q x from the r a t i o of f x over g x, mul t i p l i e d by the annual mortality rate of adults (0.0596) averaged over the 142 Figure 6.1 African buffalo age-specific mortality curve estimated using methods B and C. Solid dots indicate q estimates obtained using method B x (data from Sinclair 1977). Open dots indicate q^ estimated using the new method, C. 143 Table 6.1 Age distribution data for female African buffalo, from a found carcass sample (Sinclair 1977) and a live sample (Sinclair personal communication). Age-specific mortality rate, q^, calculated according to the new method, C (*: estimated from polynomial regression of q values 2 2 X for age classes 2-13; q = 0.0564 + 0.0214x + 0.00290x ; r =0.950). age carcass sample live sample smoothed f X smoothed gx q x 2 9 5 0.037 0.075 0.029 3 9 7 0.045 0.109 0.025 4 15 14 0.046 0.154 0.018 5 10 16 0.057 0.163 0.021 6 17 9 0.052 0.141 0.022 7 11 9 0.062 0.096 0.038 8 18 5 0.072 0.084 0.051 9 24 6 0.083 0.054 0.092 10 19 2 0.099 0.038 0.155 11 29 1 0.100 0.029 0.206 12 25 4 0.100 0.025 0.238 13 19 1 0.081 0.021 0.230 14 16 0 0.065 n.a. 0.325* 15 13 0 0.049 n.a. 0.388* 16 7 0 0.032 n.a. 0.456* 17 4 1 0.016 n.a. 0.531* 18 1 0 0.004 n.a. 0.611* 1 44 three year period of study ( S i n c l a i r 1977). Although the standing age d i s t r i b u t i o n was not c o l l e c t e d p r i o r to the found sample, equation 1 i s s t i l l appropriate because the buffalo population had been increasing at a constant rate for several years prior to the c o l l e c t i o n period, presumably producing a stable age d i s t r i b u t i o n . Due to the r e l a t i v e l y small size of the shot sample, only one individual older than 13 years was taken. I therefore estimated q values for age classes 14-18 using a polynomial curve f i t t e d to data for ages 2-13 (r 2=0.950). Sampling d i s t r i b u t i o n s The methodology employed follows Polacheck's (1985) procedure for simulating sampling d i s t r i b u t i o n s for s p e c i f i e d survival vectors. A series of computer simulations were executed to assess the sampling variation of age-specific mortality estimates derived from three d i f f e r e n t methods: (A) age-specific survivorship estimated from the r a t i o of adjacent age frequencies from the standing age d i s t r i b u t i o n ; (B) age-s p e c i f i c mortality estimated from the standing d i s t r i b u t i o n of ages at death; and (C) age-specific mortality estimated from equation 1 above. Demographic data from S i n c l a i r ' s (1977) African buffalo study were used for each method. Age-specific mortality rates were calculated according to the following equations (Caughley 1977): 1 45 method A: q x = 1 - (sx/sx_1 ) X-l method B: q = d / l - ) _ d . x x i method C: q x = k f x /g x S i n c l a i r ' s (1977) schedules for 1 and d , corrected for x x ' the population rate of increase, were used to calculate the expected standing age d i s t r i b u t i o n ( s x ) and standing d i s t r i b u t i o n of ages at death (s' ). The smoothed f and a X X " X d i s t r i b u t i o n s were taken as the expected values for method C. Simulated age di s t r i b u t i o n s were created by generating pseudorandom numbers from a uniform d i s t r i b u t i o n between 0 and 1.0 and assigning each random number to a s p e c i f i c age class by comparing the random number to the cumulative r e l a t i v e frequencies of s x , s x , f x , or . For example, i f a random number between 0 and 0.138 was drawn, one individual would be assigned to age class 0 in the simulated standing age d i s t r i b u t i o n , i f a number between 0.138 and 0.230 was drawn, one individual would be assigned to age class 1, and so on. For methods A and B, age di s t r i b u t i o n s were simulated for 1000 sampled individuals. For method C, each of the f and g x d i s t r i b u t i o n s were simulated for 500 individuals, giving a t o t a l sample of 1000. Age-specific mortality rates were calculated for each simulated age d i s t r i b u t i o n , and th i s procedure was repeated 400 times to generate sampling d i s t r i b u t i o n s of age-s p e c i f i c mortality rates using each of the three methods. A l l 146 simulations were programmed in C Language and were executed on a D i g i t a l VAX 11/750 computer, running under a Berkeley UNIX 4.2 operating system. Eff e c t s of an unstable age d i s t r i b u t i o n I f i r s t calculated the stable d i s t r i b u t i o n that would result from S i n c l a i r ' s (1977) l x and d x schedules, assuming a constant rate of increase (r=0.077). Then, for a hypothetical female population of 50,000 buffalo, I calculated the standing age d i s t r i b u t i o n produced by 2 years of increased mortality, followed by a t h i r d year in which age-specific mortality rates returned to normal. During the 2 years of increased mortality, I a r b i t r a r i l y set each q x at twice the values shown in Table 1. This scenario produced (1) a temporary decrease in buffalo population numbers from 50,000 to as low as 45,000, before increasing in the t h i r d year to 49,000', and (2) an unstable age d i s t r i b u t i o n . This scenario would mimic the e f f e c t s of temporarily harsh weather conditions, for example a drought, on an otherwise increasing population. I then calculated q x using methods A, B, and C for the hypothetical population at the end of the t h i r d year. 147 Results Age-specific mortality estimates derived using methods B (from S i n c l a i r 1977) and the new method C are shown in F i g . 6.1. There i s close agreement between the two q x curves, indicating that the proposed method (C) y i e l d s comparable results to the t r a d i t i o n a l method (B). The f i t between the two curves is close for age classes up to age 13, and diverges thereafter for older age classes whose q x values were estimated using a f i t t e d polynomial curve. F i g . 6.2 shows the standard deviations for q x derived by the three a l t e r n a t i v e methods. Method A had the greatest v a r i a t i o n , p a r t i c u l a r l y for younger age classes, while method B had the lowest v a r i a t i o n at a l l ages. Method C was comparable to B for age classes 2 through 9, but sharply increased for older age classes. Method A had an order of magnitude greater v a r i a b i l i t y than method B, even for younger age classes. Both methods A and C produced sampling d i s t r i b u t i o n s that were s i g n i f i c a n t l y skewed and kurtotic (Table 6.2). A l l of the sampling d i s t r i b u t i o n s for methods A and C were p o s i t i v e l y skewed. 71% of method C and 44% of method A d i s t r i b u t i o n s were s i g n i f i c a n t l y k u r t o t i c . In contrast, d i s t r i b u t i o n s derived from method B were neither skewed or kurtotic. Thus, sampling d i s t r i b u t i o n s derived using method B are approximately normal, while methods A and C produce d i s t r i b u t i o n s that, w h i l e ' b e l l -shaped, show some departure from normality. The comparative robustness of the three methods i s demonstrated in F i g . 6.3. Age-specific mortality curves 148 149 Table 6.2 Skewness and kurtosis of simulated sampling distributions, as indicated by g^ and g^ statistics (Snedecor and Cochrane 1967). Values significantly different from 0.0 (p'< 0.05) indicated; by.*. Method A Method B Method C Age <J1 g 2 g^ ^ g 2 g± g 2 0 0.60* 0.52* 0.03 -0.19 n.a. n.a. 1 0.33* 0.14 -0.01 -0.25 n.a. n.a. 2 0.37* 0.25 -0.03 -0.06 0.31* -0.03 3 0.41* 0.43 0.23 0.07 0.39* -0.02 4 0.72* 0.87* 0.02 -0.38 0.31* -0.06 5 0.71* 1.64* 0.32* 0.02 0.57* 0.15 6 0.46* 0.20 0.15 -0.05 0.60* 0.62* 7 0.22 -0.30 0.23 -0.15 0.70* 0.52* 8 0.44* 0.26 0.16 -0.33 0.48* -0.13 9 0.52* 0.39 0.10 -0.22 0.65* 0.67* 10 0.50* 0.56* 0.02 -0.06 1.42* 3.64* 11 0.56* 0.37 0.05 -0.04 1.29* 2.82* 12 0.61* 0.31 0.03 0.09 1.74* 6.08* 13 1.14* 2.05* -0.05 0.11 2.04* 7.66* 14 0.82* 0.81* 0.02 0.01 1.50* 1.51* 15 2.81* 14.57* -0.06 -0.17 1.26* 0.87* 150 Figure 6.3 Age-specific mortality curves estimated from an unstable age d i s t r i b u t i o n , as described i n the text, using 3 a l t e r n a t i v e methods, (assuming that the fates of a l l members of the population are known, i . e . , there i s no sampling v a r i a t i o n ) . 151 estimated using methods A or B were biased by a sl i g h t departure from the underlying assumption of a stable age d i s t r i b u t i o n , while method C was not affected. The scenario producing F i g . 6.3 involved a short (2 year) period of increased mortality, much l i k e the presumed effect of a temporary drought, with a subsequent return to normal conditions. Such perturbations may be frequent in many natural populations (Wiens 1977, Schoener 1982). To the degree that such perturbations do occur, we should expect few natural populations to exhibit a stable age d i s t r i b u t i o n , seriously l i m i t i n g the usefulness of methods A and B. Di scussion Advantages of the proposed method Under circumstances familiar to most f i e l d b i o l o g i s t s , l i f e table calculations using conventional techniques are rarely j u s t i f i e d because of r e s t r i c t i v e assumptions. One way to avoid t h i s problem i s to use independently obtained estimates of the standing age d i s t r i b u t i o n and the standing d i s t r i b u t i o n of ages at death. Because age-specific mortality rates are calculated d i r e c t l y from population deaths r e l a t i v e to those p o t e n t i a l l y vulnerable, there i s no need to know long-term trends in population si z e , or whether the age d i s t r i b u t i o n i s stable. As 152 in other methods of estimating age-specific mortality, there is an underlying assumption that sampled age frequencies are representative of the population as a whole. If s o c i a l factors promote the uneven s p a t i a l d i s t r i b u t i o n of certain age classes, then either a shot sample or a carcass sample may be biased. Equations 1 and 2 do require an estimate of t o t a l population mortality, expressed as a proportion of i n i t i a l numbers. This is often a d i f f i c u l t s t a t i s t i c to obtain. However, i t requires less e f f o r t to obtain a single precise estimate of t o t a l annual mortality than to measure precisely population numbers over an extended time period. Moreover, there i s no underlying assumption that the population has reached a stable age d i s t r i b u t i o n . One advantage of the proposed technique i s that the shape of the q x curve i s determined only by the r a t i o of the f x and d i s t r i b u t i o n s . If k ( t o t a l annual mortality) is estimated inaccurately, the o^ . curve w i l l be displaced upwards or downwards, but the shape of the curve w i l l remain the same. As Caughley (1966) suggested, the shape of the q^ curve is perhaps the most robust basis for comparisons of age-specific selection pressures experienced by d i f f e r e n t populations. From this point of view, the proposed method i s preferable to t r a d i t i o n a l techniques that rely on underlying assumptions that can fundamentally change the shape of the curve. Another benefit of the proposed technique i s that i t allows the c a l c u l a t i o n of age-specific mortality rates at frequent i n t e r v a l s during periods of rapid population fluctuation, which 153 i s not possible using current methods. Such data may y i e l d new insights into changes in age-specific selection pressures coincident with population fluctuations. Using the proposed method C i t should also be possible to determine whether a population i s increasing or decreasing simply from the age d i s t r i b u t i o n s and a single estimate of annual mortality. Let us assume that data on both the standing age d i s t r i b u t i o n and standing d i s t r i b u t i o n of ages at death are col l e c t e d for a given population, and a q^ curve is estimated using method C. Then, suppose that the s^ data i s then used to calculate q^ values using method B, under the n u l l hypothesis that population numbers have remained constant. If the q x curve generated using method B l i e s below the curve generated using method C ( i . e . age-specific mortality under the assumption that r=0.0 i s less than an unbiased estimate of age-specific mortality), then t h i s implies that the population i s decreasing. Conversely, i f q^ values calculated under the n u l l hypothesis of r=0.0 are greater than the unbiased estimates, then this implies that the population i s ac t u a l l y increasing. This reasoning i s v a l i d only i f age-specific n a t a l i t y remains unchanged. If n a t a l i t y rates increase at the same time that age-specific mortality rates are increasing, then by using the above procedure one might f a l s e l y conclude that a population was decreasing when, in fact, i t was remaining constant or even increasing. Such s i t u a t i o n s , however, may be rare in nature. If increased age-specific mortality rates result from a r e l a t i v e shortage of resources (e.g. during a drought or as a result of 1 54 large-scale habitat changes) then this process, i f anything, should s i m i l a r l y cause a decrease in n a t a l i t y . The only r e a l i s t i c circumstances under which such a scenario seems l i k e l y is when mortality changes are due to an unusual event, such as the outbreak of an epidemic or introduction of predators, coincidental with an increase in n a t a l i t y rates. Comparisons between methods The simulated sampling d i s t r i b u t i o n s allow us to judge the r e l a t i v e precision of the three procedures for estimating age-s p e c i f i c mortality. Method B, based on the d x schedule derived d i r e c t l y from the standing d i s t r i b u t i o n of ages at death, i s less subject to sampling variation than either method A or the proposed method C. Moreover, using method B i t i s impossible to generate b i o l o g i c a l l y u n r e a l i s t i c values ( i . e . greater than 1.0 or less than 0). Method A can generate q x values less than 0 and method C can generate q^ values greater than 1.0. S t r i c t l y on the basis of sampling v a r i a t i o n , method B should be preferred to method C, which i s in turn preferable to method A. Method C i s similar in precision to Method B for younger age classes, but is less precise for older age classes. There are several reasons why imprecise estimation of survival for older age classes using method C may not cause serious problems. F i r s t , in most populations the number of individuals in the older age groups i s much less than numbers in younger age groups, so that sampling errors would be small r e l a t i v e to the 1 55 whole population. Second, in many species older individuals are no longer reproductively active, so inaccuracies in estimating their frequencies w i l l have l i t t l e bias effect on population models. Third, i f q x values for younger age classes are known with precision, c u r v e - f i t t i n g procedures may be employed to discount the effect of high sampling variation for older age classes (e.g. Chapman 1964, Caughley 1966, 1977). Sampling variation becomes more pronounced for older age classes using a l l three methods. This is because age-specific mortality i s derived from rat i o s of two frequencies, that are themselves subject to sampling v a r i a t i o n . The denominator in each of the equations used for methods A, B, and C, becomes increasingly smaller as a function of age. The smaller that the expected frequency i s , the greater the e f f e c t of random variation (Fig. 6.4). As a consequence, sampling variation increases for older age classes. Methods A and B produce inaccurate q x curves when a stable d i s t r i b u t i o n has not been reached, while the proposed method C is not affected. This problem w i l l be most serious for long-l i v e d species, since i t would take many years at a constant rate of population increase to produce a stable age d i s t r i b u t i o n . Thus, method C is more robust than either methods A or B. When the underlying assumptions are met, q^ estimation from the standing d i s t r i b u t i o n of ages at death (method B), i s the most precise procedure a v a i l a b l e . However, when the underlying assumptions of a stable age d i s t r i b u t i o n and constant rate of population increase are not met, method C offers an alternative 156 0 . 4 7 -0.33H 0 . 1 9 ^ A- A A A A AAA cr* 0 . 0 5 -c o • - 0.08-1 CO > CO I "D 0 . 0 4 -0.03 0 . 0 9 0 . 1 5 B CO "D C CO CO 0 . 0 0 -0.42H 0 . 2 0 0 . 6 0 1 . 0 0 6 . 2 8 -0 . 1 4 -0 . 0 0 - A A a A_ 0 . 0 4 0 . 1 2 denominator 0 . 2 0 Figure 6.4 Sampling variation of q x estimates as a function of denominator value in the equations used. approach with a tolerable loss in precision. 1 58 CHAPTER 7. GENERAL DISCUSSION Food l i m i t a t i o n Several factors are known to l i m i t ungulate populations: (1) food a v a i l a b i l i t y (Bobek 1977; S i n c l a i r 1977; McCullough 1979; Fowler 1981; Clutton-Brock et a l . 1982; Houston 1982), (2) predation (Mech and Karns 1977; Caughley et a l . 1980; Gasaway et a l . 1983), and (3) disease (Christian et a l . 1960; S i n c l a i r 1977; Berry 1981). Moreover, a given population may be li m i t e d at d i f f e r e n t times by d i f f e r e n t factors (May 1977; Peterman et a l . 1979). For example, long-term studies indicate that Serengeti wildebeest and buffalo populations were limited by rinderpest p r i o r to a successful eradication program in the early 1960's ( S i n c l a i r 1977; S i n c l a i r and Norton-Griffiths 1979). A temporary period of increase followed rinderpest eradication, and these populations are now apparently limited by food a v a i l a b i l i t y ( S i n c l a i r 1977; S i n c l a i r and Norton-Griffiths 1982; S i n c l a i r et a l . 1985). S i m i l a r l y , Gasaway et a l . (1983) reported that some Alaskan moose populations are regulated at low population densities by wolf predation, following a series of harsh winters that depressed prey population numbers. These results indicate that multiple e q u i l i b r i a may be more prevalent than previously assumed. At the outset of t h i s study, extreme seasonal climatic v a r i a t i o n suggested that the white-eared kob population might be lim i t e d by food a v a i l a b i l i t y during the dry season period. I 159 tested this hypothesis by attempting to f a l s i f y a number of i t s predictions. Results indicated that kob mortality during the dry season, when food intake was below requirements, was considerably greater than at other times of the year. Fat reserves (a measure of body condition) declined dramatically during the dry season at the same time that mortality was increasing. F i n a l l y , unusual r a i n f a l l during the 1982 dry season, which produced increased food a v a i l a b i l i t y , resulted in decreased adult mortality. On the basis of t h i s evidence, there is no reason to reject the food hypothesis. Food a v a i l a b i l i t y during the dry season i s affected by both environmental conditions and herbivore population density. If environmental conditions were to remain constant for an extended period of time, increased mortality re s u l t i n g from i n t r a -s p e c i f i c competition for food should serve to regulate kob population numbers. However, most savannah ecosystems show considerable year to year v a r i a t i o n in r a i n f a l l . Results from thi s study suggest that t o t a l adult mortality i s highly sensitive to the duration of sub-maintenance dry season conditions. This predicts that savannah herbivore populations should fluctuate considerably from year to year. 1 60 Seasonal migration Two hypotheses have been advanced to explain the adaptive function of migration by large herbivores. The f i r s t hypothesis suggests that herbivores migrate in order to take advantage of ephemeral di s t r i b u t i o n s of food supplies (Pennycuick 1975; Maddock 1979). The second hypothesis (Western 1975) suggests that herbivores unable to meet their metabolic water requirements from forage are obliged to migrate during dry periods to permanent water supplies. During wet periods, herbivores disperse into surrounding areas to forage at w i l l . Migration patterns of white-eared kob were consistent with both hypotheses, since during the dry season water was r e s t r i c t e d to the same areas that offered the greatest abundance of green forage. Like many problems in ecology (Krebs 1985), the examination of d i s t r i b u t i o n patterns was greatly affected by the scale of observation. Data from a e r i a l surveys indicated poorly that kob dry season movements tracked the a v a i l a b i l i t y of food and water supplies, while ground surveys demonstrated t h i s r elationship more c l e a r l y . Choice of the wrong scale of observation could easily lead one to f a l s e l y reject either of the resource acquisition hypotheses. In addition to the problems of scale, i t i s often d i f f i c u l t to measure resources in terms that are meaningful for the animals involved. Not a l l vegetation may be available to herbivores, due to the e f f e c t s of vegetation structure (Bell 1971). Furthermore, as shown in th i s study and elsewhere (McNaughton 1984), continual cropping i s a common feature of 161 many grazing ecosystems. Thus, simple measures of resource abundance are not necessarily v a l i d indices of resource a v a i l a b i l i t y . Many studies have concluded that animal migration tracks s h i f t i n g resource d i s t r i b u t i o n s (Dingle 1980; S i n c l a i r 1983). However, most studies (including t h i s one) have been based on ind i r e c t , c o r r e l a t i v e evidence. Animal migration may have an en t i r e l y d i f f e r e n t primary function, such as avoiding predators or reducing parasite loads by periodic movement into new habitats, and only secondarily track resources. Consider an hypothetical population under strong s e l e c t i v e pressure to remain transient. It should then be of se l e c t i v e advantage to secondarily "choose" a migratidn route that also tracks resources. For example, results from t h i s study and elsewhere in A f r i c a (Pennycuick 1975; Western 1975) suggest that migration tracks resources. This pattern i s most apparent in the dry season, when resources are scarce. However, there i s currently no convincing explanation why these same populations move elsewhere during the wet season, since the dry season ranges also have abundant forage during the wet season. There is some supporting evidence that migratory ungulates in the Serengeti can not be regulated by predators, since the predators are unable to follow the herds because of the need to feed immobile young during a long period of dependency (Schaller 1972; Hanby and Bygott 1979). Thus, migration may have evolved primarily as an anti-predator adaptation, with a secondary s e l e c t i o n pressure 1 62 to cue movements to resource d i s t r i b u t i o n . Since in these same ecosystems other species (and sometimes population sub-groups) pe r s i s t without migration, i t i s clear that migration in search of scarce resources i s not obligatory for s u r v i v a l . Breeding phenology Mammals l i v i n g in seasonal environments should time production of young to the period of the year when food i s most abundant (Sadleir 1969; S i n c l a i r 1983a). Results from this study are consistent with t h i s hypothesis, since kob gave b i r t h during the late wet season, when food a v a i l a b i l i t y was highest. However, white-eared kob continued to lactate throughout the dry season, when food l i m i t a t i o n was most pronounced. As a consequence, females were exposed to considerable n u t r i t i o n a l stress, and had lower fat reserves than either non-lactating females or males in the population. Since food was also abundant early in the wet season, i t i s puzzling that c a l f production did not occur e a r l i e r . It i s possible that females are not able to produce young in the early wet season because of an obligatory time lag for replenishing fat reserves following the onset of the rains ( S i n c l a i r 1983a). This hypothesis implies that kob reproductive phenology has evolved to ensure adequate female body condition prior to b i r t h rather than food intake during lactation ( S i n c l a i r 1983a). Al t e r n a t i v e l y , b i r t h timing may r e f l e c t other selection pressures. Kob calves are born largely during the northward migration to the dry season range. I propose that 1 63 calving at t h i s time may serve as an anti-predator strategy, since there are few predators present in the northern areas. By delaying c a l f production u n t i l the late wet season, kob may avoid predation on vulnerable young and females in the immediate post-calving period. Breeding synchrony and male aggression The white-eared kob has a lek mating system (Bradbury 1981), in which males aggressively compete for position on specialized display grounds, from which females choose their mates. Emlen and Oring's (1977) breeding synchrony hypothesis predicts that male aggression and sexual dimorphism should be higher in the moderately synchronous white-eared kob than in the asynchronous Uganda kob. Results from th i s study were consistent with these predictions: white-eared kob males were more aggressive ( p a r t i c u l a r l y in the early part of the breeding season) and exhibit pronounced color dimorphism' r e l a t i v e to the conspecific Uganda kob. In addition, I considered the consequences of male aggression. Previous studies (Buechner 1961a) have suggested that lek aggression in ungulates serves primarily to es t a b l i s h dominance relations among lek males, but recent evidence for other species (Foster 1983) suggests that male aggression may also disrupt the breeding a c t i v i t i e s of neighboring males. Results from th i s study showed that the majority of male c o n f l i c t s occurred in the absence of females, and were presumably concerned with establishing dominance r e l a t i o n s . 1 64 However, males fought more frequently when females were present, causing a r e d i s t r i b u t i o n of females on the lek, indicating that disruption i s an important feature in kob leks. Fighting was related to increased mortality of young adult males, suggesting that competition for mates causes the skewed adult sex r a t i o in the kob population. Age-specific mortality patterns Most conventional methods of estimating age-specific mortality rates rely on r e s t r i c t i v e assumptions of a constant rate of population increase and a stable age d i s t r i b u t i o n . In practice, these assumptions are seldom met (Caughley 1966), l i m i t i n g the usefulness of these approaches. I devised an alternative method for estimating age-specific mortality rates based on the annual population mortality rate and age d i s t r i b u t i o n s from both the l i v e population and naturally occurring deaths that i s free of these r e s t r i c t i v e assumptions. The method was i l l u s t r a t e d using demographic data for the African buffalo ( S i n c l a i r 1977). I compared the proposed to two conventional methods using Monte Carlo simulation techniques. The advantages of the proposed method are: (1) i t i s more robust than conventional methods because i t has less r e s t r i c t i v e underlying assumptions, (2) i t can be applied at frequent intervals to fluctuating populations, and (3) under some circumstances i t may be used to estimate whether a population i s increasing or decreasing. The proposed method i s , however, less 165 precise than c a l c u l a t i n g values d i r e c t l y from the age d i s t r i b u t i o n of carcasses, provided that the underlying assumptions are adequately met. General conclusions Evidence from this study suggests that the Boma white-eared kob population i s limited by the a v a i l a b i l i t y of green forage during the dry season. White-eared kob exhibit a number of adaptations in response to t h i s selection pressure. Seasonal migration allows kob to u t i l i z e refuge areas that supply green forage and water when these resources are scarce elsewhere in the ecosystem. Year-round residency i s precluded by flooding of these refuge areas during the wet season. Kob time the production of young to coincide with peak forage abundance in the late wet season. Thus, adaptation to seasonal resource l i m i t a t i o n explains several important aspects of kob l i f e history. These findings suggest that other savannah-dwelling herbivores may be limited s i m i l a r l y by food abundance. However, there are 2 features of the Boma ecosystem that are probably uncommon in most other savannah ecosystems. F i r s t , during the dry season adequate forage i s r e s t r i c t e d to a r e l a t i v e l y small proportion of the Boma ecosystem. Kob population densities in the dry season range commonly exceed 1000 individuals/km 2. Very few other ungulate populations occur at such high densities, even in natural reserves. Second, predators are rare in the Boma ecosystem, unlike most other savannah ecosystems in A f r i c a . 166 Thus, in many other savannah ecosystems predation may exert a much greater impact on herbivore numbers. In the course of this study, several useful l i n e s of future work became apparent. F i r s t , although food inadequacy was ultimately responsible for much of the dry season mortality, the proximate cause of death in many cases was probably n u t r i t i o n -related disease. There has been a good deal of research into the rel a t i o n s h i p between n u t r i t i o n and resistance to disease in man and domesticated animals (Scrimshaw et a l . 1968) but l i t t l e work has been done on natural ungulate populations. Such research might provide a useful means of predicting how a given herbivore population w i l l respond to changes in food abundance. Second, Caughley (1966) has shown that many vertebrate populations have similar age-specific mortality curves. However, we know l i t t l e about how age-specific mortality rates are affected during periods of rapid population change. For example, in t h i s study I speculated that during the 1980 drought both very young and very old kob were probably more affected than young adults. Using the new technique I outlined in Chapter 6, i t should be possible to determine the r e l a t i v e impact of perturbations on s p e c i f i c age groups. This information would be of p a r t i c u l a r use in age-structured population models. Third, although natural predators in the Boma ecosystem are rare, hunting pressure may be of much greater s i g n i f i c a n c e . I inferred that hunting i s probably less important than n u t r i t i o n -related mortality in the Boma kob, but t h i s assertion needs to 167 be tested. There have been few studies to my knowledge that have examined man as a predator, but thi s information may be of pa r t i c u l a r importance in the many ecosystems, l i k e Boma, where harvesting by man i s e s s e n t i a l l y unrestricted. 1 68 LITERATURE CITED Anderson, A.B. 1949. Seasonal game migration. Sudan W i l d l i f e and Sport. J_: 1 6- 1 8 . Bayliss, P. 1985. The population dynamics of red and western grey kangaroos in a r i d New South Wales, A u s t r a l i a . I. Population trends and r a i n f a l l . J. Anim. Ecol. 54:111-125. B e l l , R.H.V. 1971. A grazing ecosystem in the Serengeti. S c i . Am. 224:86-93. Berry, H.H. 1981. Abnormal leve l s of disease and predation as li m i t i n g factors for wildebeest in the Etosha National Park. Madoqua 12:242-253. Beverton, R.J.H. and S.J. Holt. 1957. On the dynamics of exploited f i s h populations. Fisheries Investigations, London, Ser. 2. Bobek, B. 1977. Summer food as the factor l i m i t i n g roe deer population size. Nature 268:47-49. Borgia, G. 1979. Sexual selection and the evolution of mating systems. In: Sexual selection and competition in insects (Blum, M.S. and N.A. Blum, eds.). p. 19-80. Academic Press, N.Y. Bradbury, J.W. 1981. The evolution of leks. In: Natural selection and so c i a l behavior: recent research and new theory (Alexander, R.D. and D.W. Tinkle, eds.). p. 138-169. Chiron Press, N.Y. Bradbury, J.W. and R.M. Gibson. 1983. Leks and mate choice. In: Mate choice (Bateson, P., ed.). p. 109-138. Cambridge Univ. Press, Cambridge. Bredon, R.M. and J . Wilson. 1963. The chemical composition and n u t r i t i v e value of grasses from semi-arid areas of Karamoja, as related to ecology and types of s o i l . E. Afr. agric. For. J. 29:134-142. 1 69 Buechner, H.K. 1961a. T e r r i t o r i a l behavior in Uganda kob. Science 133:698-699. Buechner, H.K. 1961b. U n i l a t e r a l implantation in the Uganda kob. Nature 190:738-739. Buechner, H.K. and R. Schloeth. 1965. Ceremonial mating behavior in Uganda kob (Adenota kob thomasi Neumann). Z. Tierpsychol. 22:209-225. Buechner, H.K., J.A. Morrison and W. Leuthold. 1966. Reproduction in Uganda kob with special reference to behavior. In: Comparative biology of reproduction in mammals (I.W.Rowlands, ed). p. 69-88. Sym. Zool. Soc. Lon. Academic Press, London. Buechner, H.K., Leuthold, W. and Roth, H.D. (Unpublished MS) Lek t e r r i t o r y occupancy in male Uganda kob antelope. In: Bradbury and Gibson (1983). Buechner, H.K. and F.B. Golley. 1967. Preliminary estimation of energy flow in Uganda kob. In:. Secondary productivity of t e r r e s t r i a l ecosystems (K. Petrusewicz, ed.). p. 243-254. Pol. Acad. S c i . Inst. Ecol. Buechner, H.K. 1974. Implictions of s o c i a l behavior in the management of Uganda kob. In: The behaviour of ungulates and i t s re l a t i o n to management (Geist, V. and F. Walther, eds.). p. 853-870. Vol. 2. I.U.C.N. Morges, Switz. Buechner, H.K. and Roth, H.D. 1974. The lek system in Uganda kob antelope. Amer. Zool. 14:145-162. Bunnell, F.L. 1982. The lambing period of mountain sheep: synthesis, hypotheses, and t e s t s . Can. J. Zool. 60:1-14. Caughley, G. 1966. Mortality patterns in mammals. Ecology. 47: 906-918. Caughley, G. 1970. Eruption of ungulate populations with emphasis on Himalayan thar in New Zealand. Ecology 51:53-72. 1 70 Caughley, G, 1977. Analysis of vertebrate populations. John Wiley and Sons. N.Y. Caughley, G., G.C. Grigg, J . Caughley, and G.J.E. H i l l . 1980. Does dingo predation control the densities of kangaroos and emus? Aust. W i l d l . Res. 7:1-12. Caughley, G. and J.H. Lawton. 1981. Plant-herbivore systems. In: Theoretical ecology: p r i n c i p l e s and applications (May, R.M., ed.). p. 132-166. Blackwell, Oxford. Cave, F.O. and A. Cruickshank. 1940. A note on game migration in the southeastern Sudan. Sudan Notes and Records 23:341-344. Chalmers, M.I. 1961. Protein synthesis in the rumen. In: Digestive physiology and n u t r i t i o n in the ruminant (Lewis, D., ed.). p. 205-222. Butterworth, London. Chapman, D.G. 1964. A c r i t i c a l study of the P r i b i l o f fur seal population estimates. Fisheries B u l l e t i n 63:657-669. Child, G. 1972. Observations on a wildebeest dieoff in Botswana. Arnoldia 5:1-13. Christ i a n , J . J . , V. Flyger, and D.E. Davis. 1960. Factors in the mass mortality of a herd of Sika deer, Cervus nippon . Chesapeake S c i . U79-95. Clarke, C.M.H. and R.J. Henderson. 1981. Natural regulation of a non-hunted chamois population. N.Z. J. Ecol. 4:126-127. Clutton-Brock, T.H., S.D. Albon and F.E. Guiness. 1985. Parental investment and sex differences in juvenile mortality in birds and mammals. Nature 313:131-133. Clutton-Brock, T.H., F.E. Guiness, and S.D. Albon. 1982. Red deer: behavior and ecology of two sexes. Edinburgh Univ. Press, Edinburgh. 171 Coe, M.J., D.-H. Cummings, and J. P h i l l i p s o n . 1976. Biomass and production of large African herbivores in rel a t i o n to r a i n f a l l and primary production. Oecologia 22:341-354. Connell, J.H. 1970. A predator-prey system in the marine i n t e r t i d a l region. I. Balanus qlandula and several predatory species of Thais. Ecological Monographs 40:49-78. C o r f i e l d , T.F. 1975. Elephant mortality in Tsavo National Park, Kenya. E. Afr. W i l d l . J. jM:339-368. Cowan, I.McT. 1950. Some v i t a l s t a t i s t i c s of big game on overstocked mountain range. Trans. N. Am. Wi l d l . Conf. 15: 581-588. Deevey, E.S. 1947. L i f e tables for natural populations of animals. Quart. Rev. B i o l . 22:283-314. Delaney, M.J. and D.C.D. Happold. 1979. Ecology of African mammals. Longman, London. Demment, M.W. and P.J. Van Soest. 1985. A n u t r i t i o n a l explanation for body-size patterns of ruminant and nonruminant herbivores. Am. Nat. 125:641-672. Dingle, H. 1980. Ecology and evolution of migration. In: Animal migration, orientation, and navigation (Gauthreaux, S.A., ed.). p. 1-101. Academic Press, London. Dobzhansky, T. 1950. Evolution in the tropi c s . Amer. S c i . 38: 209-221. Duncan, P. 1975. Topi and their food supply. PhD. thesis. Univ. of Nairobi. Egan, H.R. 1977. Nu t r i t i o n a l status and intake regulation in sheep. VIII. Relationships between the voluntary intake of herbage by sheep and the protein/energy r a t i o in the digestion products. Austr. J. Agric. Res. 28:907-915. 1 72 Emlen, S.T. and L.W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223. Finch, V.A. 1972. Energy exchanges with the environment of two East African antelopes, the eland and the hartebeest. In: Comparitive physiology of desert animals (Maloiy, G.M.O., ed.). Sym. Zool. Soc. Lon. 31:315-326. Finch, V.A. and D. Western. 1977. C a t t l e colours in pastoral herds: natural selection or s o c i a l preference. Ecology 58: 1384-1392. Fisher, R.A. 1958. The genetical theory of natural selection (2nd e d i t i o n ) . Dover Publications, N.Y. Floody, O.R. and A.P. Arnold. 1975. Uganda kob ( Adenota kob thomasi ) : t e r r i t o r i a l i t y and the s p a t i a l d i s t r i b u t i o n s of sexual and agonistic behaviors at a t e r r i t o r i a l ground. Z. Tierpsychol. 37:192-212. Foster, M.S. 1983. Disruption, dispersion, and dominance in lek-breeding birds. Am. Nat. 122:53-72. Fowler, C.W. 1981. Density dependence as related to l i f e history strategy. Ecology 62:602-610. Fowler, C.W. and T. Smith. 1973. Characterizing stable populations: an application to the African elephant population. J. Wildl. Mgmt. 37:513-523. Gadgil, M. and W.H. Bossert. 1970. L i f e h i s t o r i c a l consequences of natural s e l e c t i o n . Am. Nat. 104:1-24. Gambell, R. 1975. Variations in reproduction parameters associated with whale stock s i z e s . Int. Whaling Comm., Rep. 25:182-189 Gasaway, W.C, R.O. Stephenson, J.L. Davis, P.E.K. Shepherd, and O.E. Burris. 1983. Interrelationships between wolves, prey, and man in i n t e r i o r Alaska. W i l d l . Monogr. 84:1-50. 1 73 Geist, V. 1971. Mountain sheep: a study in behavior and evolution. Univ. of Chicago Press. Chicago. Geist, V. 1974. On the r e l a t i o n s h i p of s o c i a l evolution and ecology in ungulates. Amer. Zool. 44:205-220. G r i f f i t h s , J.F. 1972. Wet and dry t r o p i c s . In: Climates of A f r i c a ( G r i f f i t h s , J.F., ed.). p. 221-258. Elsevier Pub. Co., Amsterdam. Gross, J.E. 1969. Optimal y i e l d in deer and elk populations. Trans. N. Am. Wildl. Conf. 34:372-386. Grubb, P. 1974. Population dynamics of the Soay sheep. In: Island survivors: the ecology of the soay sheep of St. Kilda (Jewell, P.A., C. Milner, and J.M. Boyd, eds.). p. 242-272. Athlone Press. A t l a n t i c Highlands, N.J. Gulland, J.A. 1977. The analysis of data and development of models. In: Fish population dynamics (Gulland, J.A., ed.). John Wiley and Sons, London, UK. Hairston, N.G., F.E. Smith, and L.B. Slobodkin. 1960. Community structure, population c o n t r o l , and competition. Am. Nat. 94:421-425. Hanby, J.P. and J.D. Bygott. 1979. Population changes in lions and other predators. In: Serengeti: dynamics of an ecosystem ( S i n c l a i r , A.R.E. and M. Norton-Griffiths, eds.). Univ, of Chicago Press, Chicago. Hanks, J. 1981. Characterization of population condition. In: Dynamics of large mammal populations (Fowler, C.W. and T.D. Smith, eds.). p. 47-74. John Wiley and Sons. N.Y. Hillman, J.C. and A.K.K. Hillman. 1977. Mortality of w i l d l i f e in Nairobi National Park during the drought of 1973-1974. E. Afr. W i l d l . J. 15:1-18. Hickey, F. 1963. Sheep mortality in New Zealand. New Zealand A g r i c u l t u r a l i s t . 15:1-3. 1 74 Houston, D.B. 1982. The northern Yellowstone elk: ecology and management. Macmillan. N.Y. Jarman, P.J. 1974 The s o c i a l organisation of antelopes in rela t i o n to their ecology. Behaviour 48:215-267. Jarman, P.J. and A.R.E. S i n c l a i r . 1979. Feeding strategy and the pattern of resource p a r t i t i o n i n g in ungulates. In: Serengeti: dynamics of an ecosystem ( S i n c l a i r , A.R.E. and M. Norton-Griffiths, eds.). p. 130-163. Univ. of Chicago Press, Chicago. Johnson, R.R., B.A. Dehority, K.E. McClure, and J.L. Parsons. 1964. A comparison of in v i t r o fermentation and chemical s o l u b i l i t y methods in estimating forage n u t r i t i v e value. J. Anim. S c i . 23:1124-1128. J o l l y , G.M. 1969. Sampling methods for a e r i a l census of w i l d l i f e populations. E. Afr. agric. For. J. 34:46-49. Keep, M.E. 1973. Factors contributing to a population crash in Nduma Game Reserve. The Lammergeyer 19:16-23. Kirkpatrick, M. 1982. Sexual selection and the evolution of female choice. Evolution 36:1-12. Klein, D.R. 1968. The introduction, increase, and crash of reindeer on St. Mathews Island. J. Wildl. Manage. 32:350-367. Krebs, C.J. 1985. Ecology: the experimental analysis of d i s t r i b u t i o n and abundance (Third e d i t i o n ) . Harper and Row. N.Y. Lamprey, H. 1963, Ecological separation of the large-mammal species in the Tarangire Game Reserve, Tanganyika. E. Afr. ag r i c . For. J. J_:63-92. Lande, R. 1980. Sexual dimorphism, sexual selection, and adaptations in polygenic characters. Evolution 34:292-305. 1 75 Lande, R. 1981. Models of speciation by sexual selection on polygenic t r a i t s . Pro. Nat. Acad. S c i . USA 78:3721-3725. Laredo, M.A. and D.J. Minson. 1973. The voluntary intake, d i g e s t i b i l i t y , and retention time by sheep of leaf and stem fractions of fi v e grasses. Aust. J . agr i c . Res. 24: 875-888. Leader-Williams, N. and C. Ricketts. 1980. Population dynamics and mortality of reindeer introduced onto South Georgia Island. J. Wildl. Manage. 44:640-657. L e s l i e , P.H. 1945. On the use of matrices in certain population mathematics. Biometrika 33:183-212. Lett, P.F., R.K. Mohn, and D.F. Gray. 1981. The impact of the determination of density-dependent processes on the management strategy for the north-western A t l a n t i c harp seal. In: Dynamics of large mammals (Fowler, C.W. and T.D. Smith, eds.). p. 135-158. John Wiley and Sons, N.Y. Leuthold, W. 1966. Variations in t e r r i t o r i a l behavior of Uganda kob, Adenota kob leucotis (Neumann,1896). Behaviour 27: 214-257. Leuthold, W. 1977. African ungulates. Springer-Verlag. B e r l i n . L i l l , A. 1974. Social organization and space u t i l i z a t i o n in the lek-forming white-bearded manakin, M. manacus t r i n i t a t i s . Z. Tierpsychol. 36:513-530. L i l l , A. 1976. Lek behavior in the golden-headed manakin, Pipra  erythrocephala , in Trinidad (West Indies). Adv. in Ethol. 18:1-83. Lowe, V.P.W. 1969. Population dynamics of the red deer (Cervus  elaphus L.) on Rhum. J . Anim. Ecol. 38:425-457. Lyeth, R.E. 1947. The migration of game in the Boma area. Sudan Notes and Records. 30:191-192. 1 76 Maddock, L. 1979. The "migration" and grazing succession. In: Serengeti: dynamics of an ecosystem ( S i n c l a i r , A.R.E. and M. Norton-Griffiths, eds.). p. 104-128. Univ. of Chicago Press. Chicago. May, R.M. 1977. Thresholds and breakpoints in ecosystems with a m u l t i p l i c i t y of stable states. Nature 269:471-477. McCullough, D.R. 1979. The George reserve deer herd: population ecology of a K-selected species. Univ. of Michigan Press. Ann Arbor, Michigan. McNaughton, S.J. 1979. Grassland-herbivore dynamics. In: Serengeti: dynamics of an ecosystem ( S i n c l a i r , A.R.E. and M. Norton-Griffiths, eds.). p. 46-81. Univ. of Chicago Press, Chicago. McNaughton, S.J. 1984. Grazing lawns, animals in herds, plant form, and coevolution. Am. Nat. 124:863-886. Mech, L.D. and P.D. Karns. 1977. Role of the wolf in a deer decline in the Superior National Forest. U.S. Dept. Agric. Serv. Res. Pap. NC-148, North Cent. For. Exp. Stn., St. Paul, Minn. 23 p. Messier, F. and M. Crete. 1985. Moose-wolf dynamics and the natural regulation of moose populations. Oecologia 65:503-512. Minson, D.J. and M.N. McLeod. 1970. The d i g e s t i b i l i t y of temperate and t r o p i c a l grasses. Pro. 11th Int. Grassl. Cong. p. 719. Modha, K.L. and S.K. Eltringham. 1976. Population ecology of the Uganda kob ( Adenota kob thomasi Neumann) in rel a t i o n to the t e r r i t o r i a l system in the Rwenzori National Park, Uganda. J. Appl. Ecol. 13:453-473. Morrison, J.A. 1971. Morphology of corpora lutea in the Uganda kob antelope, Adenota kob thomasi (Neumann). J. Reprod. Fert. 26:297-305. 177 National Research Council. 1975. Nutrient requirements of sheep. National Research Council, National Academy of Sciences, Washington, DC. Norton-Griffiths, M. 1978. Counting animals. (Second e d i t i o n ) . Handbook no. 1, African W i l d l i f e Leadership Foundation, Nairobi, Kenya. 139 p. Owen-Smith, N. 1982. Factors influencing the consumption of plant products by large herbivores. In: Ecology of tr o p i c a l savannas (Huntley, B.J. and B.H. Walker, eds.). p. 359-404. Springer-Verlag, B e r l i n . Pennycuick, L. 1975. Movements of migratory wildebeest population between 1960 and 1973. E. Afr. W i l d l . J . 13:65-87. Peterman, R.M., W.C. Clark, and C.S. Ho l l i n g . 1979. The dynamics of r e s i l i e n c e : s h i f t i n g s t a b i l i t y domains in f i s h and insect systems. In: Population dynamics (Anderson, R.M., B.D. Turner, and L.R. Taylor, eds.). p. 321-341. Blackwell, Oxford. P h i l l i p s o n , S. 1975. R a i n f a l l , primary production, and "carrying capacity" of Tsavo National Park (East), Kenya. E. Afr. Wildl. J. 13:171-201. Pielou, E.C. 1977. Mathematical ecology. John Wiley and Sons. New York. Plowes, D.C.H. 1957. The seasonal v a r i a t i o n of crude protein in twenty common veld grasses at Matopos, Southern Rhodesia, and related observations. Rhod. Agric. J. 54:33-55. Polacheck, T. 1985. The sampling d i s t r i b u t i o n of age-specific survival estimates from an age d i s t r i b u t i o n . J. W i l d l . Manage. 49:180-184. Reid, R.L., A.J. Post, F.J. Olsen, and J.S. Mugerwa. 1973. Studies on the n u t r i t i o n a l q u a l i t y of grasses and legumes in Uganda. I. Application of in v i t r o d i g e s t i b i l i t y techniques to species and stage of growth e f f e c t s . Trop. A g r i c , T r i n . 50:1-16. 178 Ricker, W.E. 1975. Computation and interpretation of b i o l o g i c a l s t a t i s t i c s of f i s h populations. B u l l e t i n of the Fisheries Research Board of Canada 191. Riney, T. 1982. Study and management of large mammals. John Wiley and Sons. N.Y. Robel, R.J. and W.B. Ballard. 1974. Lek s o c i a l organization and reproductive success in the greater p r a i r i e chicken. Amer, Zool. 14:121-128. Sadleir, R. 1969. Ecology of reproduction in wild and domestic mammals. Methuen and Co. London. Schaffer, W.M. 1974. Selection for optimal l i f e h i s t o r i e s : the effe c t s of age structure. Ecology 55:291-303. Schaller, G.B. 1972. The Serengeti l i o n : a study of predator-prey r e l a t i o n s . Univ. Chicago Press, Chicago. Schaller, G.B. 1977. Mountain monarchs. Wild sheep and goats of the Himalaya. Univ. Chicago Press, Chicago. Scheaffer, R.L. and W. Mendenhall. 1975. Introduction to pr o b a b i l i t y : theory and applications. Duxbury Press, North Scituate, Mass. Schoen, A. 1971. The eff e c t of heat stress and water deprivation on the environmental physiology of the bushbuck, the reedbuck, and the Uganda kob. E. Afr. a g r i c . For. J. 37:1-7. Schoener, T.W. 1982. The controversy over i n t e r s p e c i f i c competition. Am. S c i . 70 : 586-595. Scrimshaw, N.S., C E . Taylor, and J.E. Gordon. 1968. Interactions of n u t r i t i o n and in f e c t i o n . Geneva: WHO no. 57. 1 79 Seber, G.A.F. 1973. The estimation of animal abundance and related parameters. Hafner Press, New York, New York, USA Sherman, P.W. and M.L. Morton. 1984. Demography of Belding's ground s q u i r r e l s . Ecology 65;1617-1628 . S i n c l a i r , A.R.E. 1975. The resource l i m i t a t i o n of trophic l e v e l s in t r o p i c a l grassland ecosystems. J. Anim. Ecol. 44:497-520. S i n c l a i r , A.R.E. 1977. The African buffalo. Univ. of Chicago Press. Chicago. S i n c l a i r , A.R.E. 1983a. The adaptations of African ungulates and their e f fects on community function. In: Tropical savannas (Bourliere, F. ed). Ecosystems of the world 13:401-426. S i n c l a i r , A.R.E. 1983b. The function of distance movements in vertebrates. In: The ecology of animal movements (Greenwood, P.J. and I.R. Swingland, eds.). Clarendon Press, Oxford. S i n c l a i r , A.R.E., H. Dublin, and M. Borner. 1985. Population regulation of Serengeti wildebeest: a test of the food hypothesis. Oecologia 65:266-268. S i n c l a i r , A.R.E. and P. Duncan. 1972. Indices of condition in t r o p i c a l ruminants. E. Afr. Wildl. J. 10:143-149. S i n c l a i r , A.R.E. and J.M. F r y x e l l . 1985. The Sahel: ecology of a dis a s t e r . Can. J. Zool. 63:987-994. S i n c l a i r , A.R.E. and M. Norton-Griffiths. 1979. Serengeti: dynamics of an ecosystem. Univ. of Chicago Press. Chicago. S i n c l a i r , A.R.E. and M. Norton-Griffiths. 1982. Does competition or f a c i l i t a t i o n regulate ungulate populations in the Serengeti? A test of hypotheses. Oecologia 53:364-369. 180 Slobodkin, L.B., F.E. Smith, and N.G. Hairston. 1967. Regulation in t e r r e s t r i a l ecosystems, and the implied balance of nature. Am. Nat. 101:109-124. Smuts, G.L. 1978. Interelations between predators, prey, and their environment. Bioscience 28:316-320. Snedecor, G.W. and W.G. Cochrane. 1967. S t a t i s t i c a l methods (6th ed i t i o n ) . Iowa State Univ. Press, Ames. Spinage, CA. 1982. A t e r r i t o r i a l antelope: the Uganda waterbuck. Academic Press, London. Stearns, S.C 1976. L i f e - h i s t o r y t a c t i c s : a review of the ideas. Quart. Rev. of B i o l . 51:3-47. Strugnell, R.G. and CD. Pigott. 1978. Biomass, shoot production, and grazing of two grasslands in the Rwenzori National Park, Uganda. J. Ecol. 66:73-76. T r a i l , P. 1985. Courtship disruption modifies mate choice in lek-breeding b i r d . Science 227 : 778-780. Tucker, C.J., L.D. M i l l e r , and R.L. Pearson. 1973. Measurement of the combined ef f e c t of green biomass, chlorophyll, and leaf water on canopy spectroreflectance of the short-grass p r a i r i e . Proc. of the Second Annual Remote Sensing of Earth Resources Conf., Univ. of Tenn., Tullahoma, Tenn. Van Soest, P.J. 1963a. Use of detergents in the analysis of fibrous feeds. I. Preparation of fi b e r residues of low nitrogen content. J. Assn. O f f i c i a l Agr. Chem. 46:825-829. Van Soest, P.J. 1963b. Use of detergents in the analysis of fibrous feeds. I I . A rapid method for the determination of fiber and l i g n i n . J. Assn. O f f i c i a l Agr. Chem. 46:829-835. Van Soest, P.J. and W.G. Marcus. 1964. Method for the determination of c e l l - w a l l constituents in forages using detergents and the re l a t i o n s h i p between this f r a c t i o n and voluntary intake and d i g e s t i b i l i t y . J . Dairy S c i . 47:704-705. Van Valen, L. 1973. Patterns and the balance of nature. Evolutionary Theory J_:31-49. Vesey-Fitzgerald, D.F. 1960. Grazing succession amongst East African game animals. J. Mammal. 41:161-17 0. Weeks, J.T. 1947. Letter to editor. Sudan Notes and Records 30: 218-219. Western, D. 1975. Water a v a i l a b i l i t y and i t s influece on the structure and dynamics of a savannah large-mammal community. E. Afr. W i l d l . J. 13:265-286. Wiens, J.A. 1977. On competition and variable environments. Amer. S c i . 65:590-597. Wiley, R.H. 1973. T e r r i t o r i a l i t y and non-random mating in sage grouse, Centrocercus urophasianus . Animal Behavior Monogr. 6:87-169. . Willimot, S.G. 1956. S o i l s and vegetation of the Boma Plateau and eastern d i s t r i c t Equatoria. Sudan Notes and Records. 39:9-20. Wilson, J.R. 1976. Varia t i o n in leaf c h a r a c t e r i s t i c s with l e v e l of insertion on a grass t i l l e r . I. Development rate, chemical composition and dry matter d i g e s t i b i l i t y . Aust. J. a g r i c . Res. 27:343-354. Wittenberger, J.F. 1978. The evolution of mating systems in grouse. Condor 80:126-137. Woodgerd, W. 1964. Population dynamics of bighorn sheep on Wildhorse Island. J . Wi l d l . Manage. 28:381-391. Wrangham, R.W. 1980. Female choice of least costly males: a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-367. Zaphiro, D. 1949. Notes on L o e l l i game. Sudan W i l d l i f e and Sport. 1:6-17. 1 83 APPENDIX 1. PLANT SPECIES COLLECTED Acacia nubica Benth. Acacia polyacantha Willd. Acac ia Senegal (L.) Willd. Acac ia seyal DC. Acac ia sieber iana DC. var. sieberiana Acacia Z a n z i b a r i c a (S. Moore) Taub. Achyranthes aspera L. Alysicarpus glumeceus (Vahl) DC. Asparagus africanus Lam. Balanites aegyptiaca (L.) Del. Balanites r o t u n d i f o l i a (Van Tiegh.) Blatter Butyrospermum paradoxum (Gaertn.) Hepper ssp. niloticum (Kotschy) Hepper Cadaba farinosa Forssk. ssp. farinosa Capparis tomentosa Lam. Celt is toka (Forssk.) Hepper & Wood Combretum fragrans F. Hoff. Cordia sinensis Lam. forma v e l sp. a f f . Crateva adansoni i DC. Dichrostachys cinerea (L.) Wight & Arn. ssp. cinerea Diospyros mespiliformis A. DC. Dobera glabra (Forssk.) R. Br. Echinochloa pyramidalis (Lam.) Hitchc. & Chase Echinochloa staqnina (Retz.) P. Beauv. Mimosaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Amaranthaceae Leguminosae Li l i a c e a e Balanitaceae Balanitaceae Sapotaceae Capparaceae Capparaceae Ulmaceae Combretaceae Boraginaceae Capparaceae Mimosaceae Ebenaceae Salvadoraceae Gramineae Gramineae 184 Eragrostis c i 1 i a n e n s i s ( A l l . ) F.T. Hubb. Ficus lutea Vahl Ficus sur Forssk. forma vel sp. a f f . Fuerstia africana T.C.E. Fr. Gardenia t e r n i f o l i a Schumach. & Thonn. Gardenia volkensi i K. Sch. Grewia bicolor Juss. Grewia mollis Juss. Grewia tenax (Forssk.) F i o r i Harrisonia abyssinica O l i v . Heteropogon contortus (L.) Roem. & Schult. Hyparrhenia f i l i p e n d u l a (Hochst.) Stapf Hyparrhenia rufa (Nees) Stapf. Hyperthelia dissoluta (Nees ex Steud.) W.D. Clayton Lonchocarpus l a x i f l o r u s G u i l l . & Perr. Loudetia arundinacea (A. Rich.) Steud. Maerua angolensis DC. Maerua o b l o n g i f o l i a A. Rich. Maerua pseudopetulosa (Gilg & Bened.) De Wolf Maytenus senegalensis (Lam.) Ex e l l Mimosa pigra L. Nauclea l a t i f o l i a Smith Ozoroa i n s i q n i s Del. ssp. insign i s Panicum coloratum L. Piliostigma thonningii (Schumach.) Milne-Redh. Pseudocedrela kotschyi (Schweinf.) Harms Gramineae Moraceae Moraceae Labiatae Rubiaceae Rubiaceae Ti1iaceae Ti 1iaceae Ti l i a c e a e Simaroubaceae Gramineae Gramineae Gramineae Gramineae Papilionaceae Gramineae Capparaceae Capparaceae -Capparaceae Celestraceae Mimosaceae Rubiaceae Anacardiaceae Gramineae Caesalpiniaceae Meliaceae 185 Rhus natalensis Krauss Securineqa virosa (Willd.) B a i l l . Sehima nervosum (Willd.) Stapf Setaria incrassata (Hochst.) Hack. Sorghum purpureo sericeum (Hochst. ex A. Rich.) Aschers & Schweinf. Sporobolus helvolus (Trin.) Dur. & Schinz. Sporobolus ioclados (Trin.) Nees Steganotaenia araliacea Hochst. forma v e l . a f f . Stereospermum kunthianum Cham. Tamarindus indica L. Terminalia brevipes Pampan Triumfetta rhomboidea Jacq. Ximenia c a f f r a Sond. Ziziphus mauritiana Lam. Anacardiaceae Euphorbiaceae Gramineae Gramineae Gramineae Gramineae Gramineae Araliaceae Bignoniaceae Caesalpinaceae Combretaceae Tili a c e a e Olacaceae Rhamnaceae 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0096671/manifest

Comment

Related Items