Satisfying Giant Appetites
Mechanisms of small scale foraging by large African
herbivores
Yolanda Pretorius
Promotores
Prof. Dr. H.H.T. Prins
Professor of Resource Ecology
Prof. Dr. R. Slotow
Professor of Biological Sciences, University of Kwa-Zulu Natal (South Africa)
Co-promotor
Dr. W.F. de Boer
Lecturer at the Resource Ecology Group
Thesis committee
Dr.A.D. Rijnsdorp
Wageningen University (Netherlands)
Prof.Dr. S. Verhulst
University of Groningen (Netherlands)
Prof.Dr. H. Udo de Haes
Leiden University (Netherlands)
Dr. A.G. Toxopeus
International Institute for Geo-information Science and Earth Observation Enschede (Netherlands)
This research was conducted under the auspices of the C.T. de Wit Graduate School for
Production Ecology and Resource Conservation
Satisfying Giant Appetites
Mechanisms of small scale foraging by large African
herbivores
Yolanda Pretorius
Thesis
submitted in partial fulfillment of the requirements for the degree of doctor
at Wageningen University
by the authority of the Rector Magnificus
Prof. Dr. M.J. Kropff
in the presence of the
Thesis Committee appointed by the Doctorate Board
to be defended in public
on Monday 23 November 2009
at 11 AM in the Aula.
Yolanda Pretorius
Satisfying giant appetites: mechanisms of small scale foraging by large African herbivores,
141 pages.
PhD-thesis, Wageningen University, Wageningen, the Netherlands (2009)
with summaries in English, Dutch and Afrikaans.
ISBN 978-90-8585-478-4
Preface
The true meaning of the phrase ‘you are what you eat’ only became clear to me during the
course of this study, except that for herbivores I think it should be ‘you are what you are
constrained to eat’. This study started off rather broadly with the aim of determining the
mechanisms behind small scale foraging by large herbivores. With guidance, trial and error I
soon realized that a holistic approach leads to confusion and increases the chances of
misinterpretation of results as it is impossible to account for all of the variability present, even
at small scales. Eventually, after adopting a reductionist approach and using field
experimentation as a tool to reduce variability, the study evolved to focus on how body mass
differences determine mouth morphology, patch selection and daily diets of large African
herbivores. I dedicate this thesis to my parents, Danie and Jorina and to the four mentors who
have guided and inspired me throughout my academic career including my grandmother
Drienie Smal, Jock McMillan, Paul Hendrik and Herbert Prins. This PhD journey has truly
been an enriching experience both personally and in terms of my career. I hope you enjoy
reading this work as much as I, despite occasional hardships, enjoyed creating it.
Contents
CHAPTER 1
General introduction……………………………………………………………….
1
CHAPTER 2
Why elephant have trunks, giraffe have long tongues & rhino have broad lips:
mechanisms for how plants shape large herbivore mouth morphology …………..
9
Y. Pretorius, K. Kortekaas, M. van Wijngaarden, W.F. de Boer, R. Slotow, &
H.H.T. Prins
CHAPTER 3
Soil nutrient status determines how elephant utilize trees and shape
environments ……………………………………………………………………... 25
Y. Pretorius, W.F. de Boer, R. Slotow, C. van der Waal & H.H.T. Prins
CHAPTER 4
Large herbivore responses to nutrient heterogeneity in an African savanna……… 39
Y. Pretorius, W.F. de Boer, M.B. Coughenour, H.J. de Knegt, H. de Kroon,
C.C. Grant, I. Heitkonig, E. Kohi, E. Mwakiwa, M.J.S. Peel, A.K. Skidmore,
R. Slotow, C. van der Waal, F. van Langevelde, S.E. van Wieren & H.H.T.
Prins
CHAPTER 5
Diet selection of African elephant over time shows changing optimization ……... 57
currency
Y. Pretorius, J.D. Stigter, W. F. de Boer, S.E. van Wieren, C.B. de Jong, H.J. de
Knegt, C.C. Grant, I. Heitkönig, N. Knox, E. Kohi, E. Mwakiwa, M.J.S. Peel, A.K.
Skidmore, R. Slotow, C. van der Waal, F. van Langevelde, H.H.T. Prins
CHAPTER 6
Synthesis: why size matters at small scales………………………………............. 77
Y. Pretorius
References ……………………………………………………………………...... 101
Summary …………………………………………………………………............ 120
Samenvatting – Dutch Summary ………………………………………………... 124
Samevatting – Afrikaans Summary……………………………………………… 130
Affiliation of coauthors ……………………………………………………….....
135
Acknowledgements ……………………………………………………………...
136
Curriculum Vitae ……………………………………………………………....... 139
PE&RC PhD Education Certificate ……………………………………………... 141
Abstract
Pretorius Y., 2009. Satisfying giant appetites: mechanisms of small scale foraging by large
African herbivores
Variation in body mass allows for resource partitioning and co-existence of different species.
Body mass is also seen as the main factor governing nutrient requirements in herbivores as
metabolic rate and requirements have often been found to scale to ¾ power of body mass.
Although the consequences of body mass on foraging behaviour of herbivores has been
extensively studied, the mechanism behind how body mass differences determines the small
scale foraging patterns of especially larger herbivores, has up to now been unclear. In this
study, I looked at how body mass and small scale vegetation characteristics shaped the mouth
morphology of herbivores and how body mass of a herbivore affects the scale at which intake
is maximized. More over, I looked at what spatial scales a mega-herbivore can select nutrient
rich patches, and at the trade-offs between quality and quantity of plant species as forage to be
included in the diet of different size herbivores. The results indicate that the dilution of plant
mass and more specifically leaf mass in space requires that mega-herbivores have enlarged
soft mouth parts to compensate for this dilution. In the analysis of patch selection by
herbivores, I introduce the novel concept of nutrient load, which provides a way of expressing
the total available nutrients to a herbivore per grain of a specific spatial scale. I show that
solitary animals and/or animals with extreme mouth sizes are able to select the aggregation of
patches with either the largest patch sizes or highest local nutrient concentrations which
together yield the highest total nutrient loads at a spatial scale of 2500m2. Finally, I
demonstrate, using linear programming techniques with multiple nutrients as constraints, how
a mega-herbivore’s daily diet choice is determined by forage abundance whereas a small
herbivore is more constrained by fibre.
Keywords: patch selection, allometry, spatial scaling, body mass, mouth morphology,
linear programming
CHAPTER 1
GENERAL INTRODUCTION
1
Changes in abiotic environmental factors such as atmospheric temperature and moisture
levels have been present since the origin of life. These changes occur at both a macro and
micro scale leading to spatial and temporal heterogeneity in the occurrence of biotic elements.
Organisms either go extinct or adapt to environmental changes through the process of natural
selection, which is the basis of speciation and the principle by which variation in a trait that
enhances survival and reproduction is preserved (Darwin 1859). Millions of years of
evolution have selected the fitness maximization strategies that extant species possess.
However, over a much shorter time, humans have heavily impacted on and artificially
changed environmental conditions (Prins & Gordon 2008) so that extant species’ ability to
adapt has been challenged to the extreme. For example, during the last two centuries Africa’s
large herbivores were first threatened by extinction from uncontrolled hunting activities and
poaching (Pringle 1982), followed by unprecedented habitat conversion for agricultural
practices (Olff et al. 2002) only to be left with a future where the effects of global warming
might present an even bigger challenge to survival.
Grazing ecosystems are among the earth’s most endangered terrestrial habitats (Frank et
al. 1998) and with increasing CO2 levels, C4 grasses, which dominate savannas, are expected
to decrease (Bond et al. 2003) with concomitant effects on wild grazing species (Prins &
Gordon 2008). Although predictions on the possible effects of global warming have received
much attention during the last decade (Root et al. 2003 Google, cited 1038 times), the
immediate threat of habitat loss through land-use change by humans is often neglected
(Brooks et al. 2002 Google, cited 284 times). Nearly half the world’s vascular plant species
and one third of terrestrial vertebrates are endemic to 25 biodiversity “hotspots”, which
historically covered 12 % of the land’s surface but today less than 1.4 % of intact habitat
remains (Brooks et al. 2002).
The inevitable question is, with increasing human impact, will large herbivores be able to
adapt and survive or will they face another mass extinction? To prevent such an extreme
event we need to know how large herbivores satisfy their daily nutrient requirements and how
they are adapted to do so. More over, prediction of the distribution of large herbivore species,
either for the purpose of conservation or management, requires an understanding of the
mechanisms behind resource use by these species. Historically, studies often focused on large
scale habitat use and distribution of large African herbivores using a descriptive approach
(McNaughton & Georgiadis 1986). However, small scale foraging processes can significantly
affect these large scale distribution patterns (Shipley 2007). In this study, I have used an
experimental approach to determine the mechanisms behind small scale foraging by large
African herbivores.
2
Contribution to foraging ecology
Optimal foraging theory, which states that organisms forage in such a way as to maximize
their energy intake per unit time, developed from the late 1960’s (MacArthur & Pianka 1966,
Stephens & Krebs 1986) and has been central to foraging ecology ever since (Belovsky 1997,
Illius et al. 2002, Prins & van Langevelde 2008). Nevertheless, classical optimal foraging
theory has some shortcomings. First, it assumes that fitness is maximized by maximizing
daily food intake, subject to physical and physiological constraints. This assumption lacks
support because fitness is more likely to be maximized by balancing benefit and costs over
the organism’s lifetime (Illius et al. 2002). Secondly, energy has been used as the exclusive
currency to model optimal foraging whereas in reality animals have to satisfy requirements of
multiple nutrients. Although more recent studies have recognized this fact, little support of
theoretical models are available from field studies and some disparity still exists on whether
animals aim to only satisfy daily nutrient requirements or whether the intake of particular
nutrients is maximized (Raubenheimer & Simpson 1999, Voeten & Prins 1999, Prins & van
Langevelde 2008, Treydte et al. 2009, Felton et al. 2009, Hengeveld et al. 2009).
Finally, classical optimal foraging theory has neglected the issue of scale (Prins & van
Langevelde 2008). Nonetheless, it is well recognized that animals select forage resources at a
range of temporal and spatial scales (Spalinger & Hobbs 1992, Ball et al. 2000). In 1987
Senft and co-workers proposed a spatial hierarchy of foraging components with selection
units ranging in scale from individual plants through communities to entire landscapes
embedded within regions. Since then, other studies have developed theoretical frameworks to
explain the spatial scaling of foraging (Bailey et al. 1996, Ritchie & Olff 1999) but support
from field research in the ‘natural’ systems is scarce (Cromsigt & Olff 2006). In this study, I
aimed to contribute to the larger body of knowledge on optimal foraging theory by conducting
field experiments incorporating the spatial scaling of foraging as well as the foraging
strategies adopted by large herbivores at small spatial scales including satisficing and
maximization of the intake of multiple nutrients.
Study subjects
All animals included in this study are commonly referred to as large herbivores, defined as
terrestrial mammals heavier than 5 kg in weight, which obtain most of their food resources
from vegetative plant parts (Prins & Olff 1998, van Langevelde & Prins 2008). More
specifically, the study focused on large African herbivore species including: steenbok
(Raphicerus campestris), duiker (Sylvicapra grimmia), bushbuck (Tragelaphus scriptus),
impala (Aepyceros melampus), warthog (Phacochoerus aethiopicus), blesbok (Damaliscus
dorcas phillipsi), nyala (Tragelaphus angasii), oryx (Oryx gazelle), red hartebeest
(Alcelaphus caama), kudu (Tragelaphus strepsiceros), wildebeest (Connochaetes taurinus),
3
waterbuck (Kobus ellipsiprymnus), zebra (Equus burchellii), eland (Taurotragus oryx),
buffalo (Syncerus caffer), giraffe (Giraffa camelopardalis), white rhinoceros (Cerathorium
simum) and elephant (Loxodonta africana). For comparison, species were distinguished from
each other according to their body mass, type of digestive system, forage type, mouth size and
herd size (Figure 1.1).
Body mass
Herbivores included covered a body mass range of more than three orders magnitude (10 kg –
4850 kg). The most renowned implications of body mass differences are that basal metabolic
rate is proportional to ¾ power of body mass (Kleiber 1932, Prins & van Langevelde 2008).
The metabolic requirements of a large animal like an elephant will thus be lower in relation to
body mass compared to a small animal like an impala. From the Bell-Jarman principle it
follows that because body mass and gut size increase isometrically, larger animals with their
lower metabolic energy requirements, can tolerate lower quality diets and can therefore afford
to be less selective (Bell 1971 and Jarman 1974). Hence, because patch selection at small
scales by the largest terrestrial mammal will provide strong support for the ability of large
herbivores to select nutrient rich patches in general, the elephant was used as a model animal
for a large part of the study (Chapter 3, 5 & 6). More over, vigilance behaviour, which is often
difficult to assess, could to a large extent be ignored because elephant, due to their enormous
size, are less prone to predation. Finally, the characteristic signs of foraging by elephant on
trees, which even when old are easy to detect, were thought to enhance the quality of data
collected.
Digestive system type
Herbivores were further distinguished according to the digestive system type of the suborder
Ruminantia versus other herbivores (Figure 1.1). Ruminants typically re-chew their food and
have specialized compartments to host micro organisms that aid in the digestion of plant cell
walls, which mammalian digestive juices generally cannot (Foose 1982, van Soest 1994).
Although, monogastric herbivores of other suborders can also host micro organisms for
fermentation of plant cell walls in their hind guts, the efficiency of this type of digestion is not
as good as in ruminants (Foose 1982). Plant cell walls consist of hemi-cellulose, cellulose and
lignin, often cumulatively referred to as fibre (McDonald et al. 1995). However, irrespective
of digestive system type, digestibility is strongly correlated with the fibre content of plants
(van Soest 1994) and is a good indicator of the metabolizable energy value of a plant to an
animal (McDonald et al. 1995). Hence, unless analyzed for multiple nutrients, the fibre
content of forage in this study served as a proxy of its nutrient value and quality to an animal.
As a measure of fibre, forage was analysed for its Acid Detergent Fibre (ADF) content that
included the cellulose and lignin fraction and not hemi-cellulose, which can be digested much
easier.
4
Forage type
Various authors have categorized herbivores as browsers, grazers or mixed feeders depending
on the amount of dicot and monocot plant species in their diets (Hofman & Stewart 1972, van
Soest 1994, Gagnon & Chew 2000, Gordon & Prins 2008). Similarly, herbivores included in
the current study were categorized as depicted in Figure 1.1. Grasses, unlike browse plants,
form a continuous layer with small differences in the leaf and stem fibre content within a
plant. The fibre fraction of grasses is higher than in browse species but consist mostly of
cellulose that can be digested by micro organisms in the gut whereas browse contain more
indigestible lignin (Searle & Shipley 2008).
Implications of these forage type differences is that herbivores foraging on browse have
developed narrow muzzles to better select leaves in between browse stems whereas grazers
developed broader muzzles and adopted a strategy involving higher bite rates than browsers
(Gordon & Illius 1988, Drescher 2003, Searle & Shipley 2008). Larger animals, which tend to
be grazers, have a proportionately larger gastrointestinal tract than do small animals, which
tend to be browsers (Van Soest 1994). Because small animals require more energy per unit
weight to fuel a higher mass-specific metabolism, they must obtain a high rate of energy
return per gram of food ingested. Large herbivores are thus better suited to extract energy
from high-fiber grasses while small animals from the cell contents of browse (Demment and
Van Soest 1985).
Study area
All data were collected between June 2005 and June 2008 in the north-eastern parts of South
Africa. This area falls within the savanna biome as described by Rutherford and Westfall
(1986). Foraging observations on herbivores, reported on in Chapter 2, were collected at four
sites as all herbivores included for this part of the study were not common at all sites. White
rhinoceros and some blue wildebeest measurements were collected at Mabula Private Game
Reserve (MGR) in the Limpopo Province (latitudes: 24°42 to 24°50 S, longitudes: 27°50 to
27°58 E) which has a mean annual rainfall of about 600 mm and is classified as a sour- and
mixed bushveld dominated by tree species such as Combretum apiculatum, Terminalia
sericea, Burkea africana and Acacia caffra (Low & Rebelo 1996). Measurements on nyala,
bushbuck and some kudu were collected in the north of Kruger National Park (KNP)
(latitudes: 22°20 to 22°32 S, longitudes: 30°53 to 32°02 E), whereas most zebra, buffalo,
waterbuck and wildebeest measurements were collected in the centre of the park around the
Satara rest camp. Annual rainfall in KNP ranges between 300-1000 mm and whereas the
north is especially dominated by Colophospermum mopane woodlands and shrubveld, the
centre is characterized more by open savannas dominated by Acacia and Combretum species
(Venter et al. 2003).
5
Figure 1.1: Classification of the large African herbivores used in this thesis along an axis of
body mass (a, steenbok; b, duiker; c, bushbuck; d, impala; e, warthog; f, blesbok; g, nyala;
h, oryx; i, red hartebeest; j, kudu; k, blue wildebeest; l, warthog; m, zebra; n, eland; o,
buffalo; p, giraffe; q, white rhinoceros; r, elephant). Figures surrounded by dark grey
indicate animals that pre-dominantly browse whereas light grey indicate mixed feeding
and white, grazing. Numbers beside letters indicate the chapters in which data of each
species were used.
All observations on elephant, steenbok, warthog and impala and some observations on
giraffe, waterbuck and kudu were conducted along the eastern border of KNP within the
Associated Private Nature Reserves (APNR), which included Timbavati, Umbabat, Klaserie
and Balule reserves. Fences between the APNR and the Kruger National Park were removed
in 1995, thus allowing free movement of animals between these areas. These reserves, as with
the KNP and MGR, are characterized by a dry season stretching from April to November
(winter) and a wet season from October to March (summer). Annual precipitation within the
APNR ranges between 200 and 1100 mm and vegetation is dominated by mixed Combretum/
6
Terminalia sericea woodland, Combretum/ Colophospermum mopane woodland and some
thornveld on gabbro (Greyling 2004, Peel et al. 2007). To address the research questions in
Chapter 3 and 4, a large scale field fertilization experiment, consisting of thirty 50 m x 50 m
plots were set-up in December 2004 in the Timbavati (24˚14'11"S; 31˚22'32"E) (*see also van
der Waal in prep). The area was re-fertilized two years later. The experimental study site and
the area from which the data were collected for Chapter 5 was dominated by a homogeneous
layer of C. mopane shrub veld, with a herbaceous layer dominated by Urochloa
mosambisencis and Bothriochloa spp. on granitic soils.
Thesis outline
The thesis focused mainly on how animal characteristics such as mouth size, digestion type,
forage type and herd size in relation to a herbivore’s body mass determines the scaling of
foraging mechanisms which range from the bite to the daily foraging range (Figure 1.2).
Spatial scale is characterized by grain size, which is the smallest measurement unit used at a
specific scale and area. This includes the entire area of interest (Kotliar & Wiens 1990).
Patches are localities that are more or less homogenous in their plant species composition,
biomass and quality (Prins & van Langevelde 2008). Patches may also vary in size and
number within or beyond the specified grain size of a chosen scale.
Chapter2 begins at the bite-size scale to investigate whether instantaneous intake rate
scales to body mass (BM) in a similar fashion as metabolic rate as BM0.75 (Kleiber 1932, Prins
& van Langevelde 2008). However, because allometric mass-space relationships in plants
indicate that leaf:stem ratios of plants decrease as volume increase and that the average
number of plants per unit area decrease as the average plant mass increases (Enquist et al.
1998, Niklas 2004), I further investigated how mega-herbivores in particular are adapted to
cope with the dilution of forage resources in space thus enabling them to achieve their
required intake rates. I hypothesize that intra-dental mouth volume, determined as the product
of incisor width, muzzle width and jaw length, should scale linearly with body mass and that
the volume added by soft mouth parts such as lips and tongue, should cause the total bite
volume to scale larger than one using body mass.
In Chapter 3, I scale up from bite-size, to determine within the daily foraging range, at a
pre-determined scale, whether herbivores as large as elephant are able to select grains with the
highest concentration of nutrients irrespective of the heterogeneity of patches within the
grains. I further tested whether elephant were able to select nutrient rich patches at even
smaller scales and whether the level of impact on trees differed between nutrient rich and
poor patches.
*C. van der Waal, Ph.D. candidate. Resource Ecology Group, Wageningen University, the Netherlands
7
Figure 1.2: Outline of thesis showing how animal characteristics such as mouth size are used
in relation to body mass to determine the scaling of mechanisms of foraging ranging from
bites to daily foraging ranges.
Body mass is the most well studied animal characteristic and a reliable universal predictor
of foraging behaviour in herbivores (West & Brown 2005). Other foraging characteristics,
which are essentially functions of body mass that have been recognised to significantly
influence large herbivore foraging, include mouth size, herd size and digestive system type
(van Soest 1996). Although the functions of these characteristics are well understood, the way
in which they affect a herbivore’s response to spatial heterogeneity in forage resource
distribution is less clear. In Chapter 4, I analysed how body mass, mouth size, digestive
system type and herd size affect the selection of the spatial scale at which herbivores select
nutrient rich patches to ensure maximum intake. I hypothesized that smaller animals, animals
occurring in large herds, and ruminant species will select areas of high nutrient concentration
at small spatial scales whereas small-mouthed animals will select areas with high nutrient
content at larger spatial scales.
In Chapter 5, using linear programming techniques I tested which strategies best explain
the daily diets of African elephant. I used daily requirements of the macro-nutrients nitrogen,
phosphorous, potassium, sodium, calcium, magnesium and metabolizible energy from
literature as constraints in the model. Further, I expressed the availability of each potential
forage species in the study area as the amount of nutrients and metabolizable energy available
per plant that could be gained after energy required finding each species has been accounted
for. In Chapter 6, I synthesize the results from this study by looking at the implications of
body mass differences in large herbivores when foraging at scales ranging from bites to daily
foraging range.
8
CHAPTER 2
WHY ELEPHANT HAVE TRUNKS, GIRAFFE HAVE
LONG TONGUES & RHINO HAVE BROAD LIPS:
MECHANISMS FOR HOW
PLANTS SHAPE LARGE HERBIVORE MOUTH
MORPHOLOGY
Y. Pretorius, K. Kortekaas, M. van Wijngaarden, W.F. de Boer, R. Slotow,
& H.H.T. Prins
9
Abstract
From universal scaling laws, plant mass has been shown to dilute in space as volume0.75, as
does leaf mass to stem mass0.75. Even though metabolic requirements of herbivores scale to
body mass (BM) as BM0.75, which means that larger herbivores can tolerate lower quality
foods, these animals still have high absolute food requirements for which the dilution of plant
resources in space will some how have to be compensated for. We investigate whether mass
and morphological spatial patterns in plants possibly induced the development of enlarged
soft mouth parts in especially mega-herbivores. We used power functions and geometric
principles to explore allometric relationships of both morphological and foraging
characteristics of mammalian herbivores from the savannas of South Africa, covering a body
mass range of more than three orders magnitude. Our results show that, although intra-dental
mouth volume scaled to a power slightly less than one to body mass, actual bite volume, as
measured in the field, scaled to body mass with a factor closer to 1.75. However, when
including the volume added to intra-dental mouth volume by soft mouth parts, such as tongue
and lips (or trunks in elephant), mouth volume scaled linearly with actual bite volume and in a
similar fashion as actual bite volume to body mass. Bite mass and bite leaf mass scaled
linearly with body mass. We conclude that these scaling relationships indicate that large
herbivores use their enlarged soft mouth parts to not only increase bite volume and thereby
bite mass, but also to select soft plant parts, and thereby increase the leaf mass fraction per
bite. This argument is strengthened by our findings of instantaneous intake rate, expressed as
forage mass ingested per unit of time, scaling linearly to body mass, and of instantaneous
intake rate of digestible cell mass ingested per unit of time scaling to BM0.75, which is in
accordance to Kleiber’s allometry. We conclude by hypothesizing that past extinctions of
mega-herbivores occurred because these herbivores did not have suitable mouth morphologies
to cope with the dilution of food resources in space.
Keywords: large African herbivores, soft mouth parts, allometry, bite mass, instantaneous
intake rate, feeding station
10
Introduction
Allometry describes the disproportionate changes in shape, size or function observed when
comparing separated isolated features in animals (or plants) spanning across a range of body
mass (Lindstedt & Schaeffer 2002). In both plant and animal ecology, fractal geometry and
basic physical laws are used to develop principles for allometric scaling. For example, the
surface area of an animal’s body scales to 2/3 power of its body volume whereas linear
measurements of body parts scale to 1/3 of body mass (Hutchinson 1959, Schmidt-Nielson
1984). As these allometric scaling laws provides strong predictions based on logic geometric
principles, the quest to find universal scaling laws of body mass with biological
characteristics of foraging is ongoing (West & Brown 2004, 2005). In this paper, we first
investigate how foraging characteristics such as bite mass, bite rate and subsequently
instantaneous intake rate scale with herbivore body mass. We then focus on the mechanisms
behind these scaling relationships by looking at how the small scale distribution of forage
resources affect the allometry between herbivore body mass and mouth morphology.
Body mass influences habitat selection and facilitates species coexistence because a wider
food quality tolerance by larger herbivores, as predicted by Kleiber’s allometry (Kleiber
1932), will allow them to use a higher diversity of habitat types (Prins & Olff 1998).
Conversely, small scale foraging characteristics such as bite, feeding station and patch size,
can also shape large scale distribution patterns of herbivores (Bailey et al.1996, Shipley 2007).
Hence, it is reasonable to expect small scale requirements and intake rates to scale to body
mass in a similar fashion as larger scale requirements and intake rates.
The relationship of instantaneous intake rate with available plant biomass is generally
used to describe the functional response of an animal (Solomon 1949, Holling 1959), where
intake is the result of bite mass, bite rate and feeding time (Spedding et al. 1966, Hodgson
1985). Bite mass has been proposed to scale linearly with body mass (Clutton-Brock &
Harvey 1983, Owen-Smith 1985), which is an assumption that can also be derived from the
fact that gut capacity scales linearly with body mass (Demment and van Soest 1985). In
contrast, support for the existence of a scaling relationship between bite rate and body mass is
rare and studies that have investigated this relationship have found none (Shipley et al. 1994).
Therefore, as bite rate is depressed by bite mass through competition (Spalinger et al. 1988,
Spalinger & Hobbs 1992) and bite mass is believed to have a larger influence on
instantaneous intake rate than bite rate (Spalinger & Hobbs 1992, Shrader et al. 2006), it can
be expected that intake rate like bite mass should scale linearly with body mass. However, in
the functional response, herbivores may trade-off between biomass against forage digestibility
because of digestive constraints (Beekman & Prins 1989, Wilmhurst et al. 1995, Iason & van
Wieren 1999). Fibre is the main determinant of digestibility of forage as it affects the
throughput rate of food (Demment & van Soest 1985). A frequently used measure of the fibre
11
content of forage is Acid Detergent Fibre (ADF), which includes the cellulose and lignin
fractions of a cell (McDonald et al.1995). As trees are generally higher in lignin and grasses
higher in cellulose (Searle & Shipley 2008), ADF should be a good indicator of digestible cell
mass across herbivore species of different feeding types.
Bite rate (Spalinger et al. 1988) and chewing time (Beauchemin 1991) are both affected
by the fibre content of bite forage (Beekman & Prins 1989). More over, as chewing rate scales
negatively as BM-0.25 (Fortelius 1985) and chewing rate is linearly correlated to bite rate
(Demment & Greenwood 1988), bite rate might be expected to scale as BM-0.25. Therefore,
alternatively, when including digestibility (or ADF as its proxy) in the calculation of
instantaneous intake rate as the product of the percentage digestible cell content (100 – ADF
%), bite rate (BM-0.25) and bite mass (BM1), we expect intake rate to scale to BM0.75 as
predicted by Kleiber’s allometry (Kleiber 1932, Prins & van Langevelde 2008).
The mechanisms behind the scaling relationships between body mass and the before
mentioned forage characteristics should be affected by the spatial distribution of forage
resources. Thus, the laws governing the availability of these forage resources first have to be
investigated. According to the plant thinning law the number of plants per unit space will
decrease at a rate of ¾ power as average plant mass increase (Enquist et al. 1998). Similar to
this, standing leaf biomass has been found to scale to ¾ power of standing stem biomass
(Niklas 2004) whilst maximum plant height is proportional to 2/3 power of trunk diameter
(buckling height: McManhon 1973). These scaling relationships in plants have consequences
for herbivores because even though larger herbivores can tolerate food of lower quality, they
still have high absolute food requirements for which the dilution of plant resources in space
will some how have to be compensated for. Therefore, we further investigate the adaptations
that especially mega-herbivores have evolved to adapt to spatial patterns in plants.
At the bite scale, forage intake behaviour basically represents a compromise between
mastication, which increases passage rate, and taking a new bite, which increases intake.
However, both these processes depend on the herbivore’s mouth morphology (Demment &
Greenwood 1988). Generally, grazers have larger incisor widths and muzzles than browsers
which allow the forager to obtain large bite sizes from the surface of uniformly distributed
grass swards (Gordon & Illius 1988). However, browsers, with their smaller muzzles and
prehensile lips, can obtain large bites by stripping many leaves from one stem in a sideways
motion (Searle & Shipley 2008). Hence, we included both grazers and browsers in our study
and reason that differences in the dimensions of mouth parts and methods of feeding between
these animals will be compensated for, which will result in similar scaling of foraging
characteristics such as bite volume and bite mass across species of different body mass.
Bite mass is not only regulated by plant characteristics such as sward height, presence of
stems, bulk density or biomass (Stobbs 1973, Black & Kennedy 1984, Prins 1996,
12
WallisdeVries et al.1998, Benvenutti et al. 2006, Heuermann 2007), but also by the
morphology of the animal’s mouth (Illius & Gordon 1987, Shipley et al. 1994). Previous
studies on mouth morphology and its relation to body mass, focused on skull measurements
such as incisor width, jaw length and cheek teeth surface. From these studies, we know that
most of these parameters do indeed scale allometrically to body mass (e.g. linear dimensions
as BM0.33 and surface dimensions as BM0.67) (Fortelius 1985, Illius & Gordon 1987, Gordon
& Illius 1988, Wilson & Kerley 2003). However, even though the function of the tongue and
lips (e.g., trunk of elephant) have been widely recognised as foraging extensions that increase
the bite area (Ungar et al. 1991, Shipley et al. 1994, Drescher 2003, Hongo & Akimoto 2003,
Griffiths 2006), no study has included measurement of these soft mouth parts and related
them to body mass and foraging efficiency.
The spatial characteristics of foraging are determined by volume of the available forage
and the resource mass contained within that volume. For example, at the bite scale, bite mass
is determined by bite volume and the bulk density of the herbage in that volume (Ungar et al.
2001). According to Clutton-Brock and Harvey (1983) and Owen-Smith (1985), bite mass
should scale linearly with intra-dental mouth volume. However, in Canadian geese, where
soft mouth parts outside the bill are absent, bill length increase much faster with body mass
than expected from the general allometric relationship of BM0.33 (Heuermann 2007). We
reason that intra-dental mouth volume, calculated as the product of jaw length (BM0.33),
incisor width (BM0.33) and muzzle diameter (BM0.33) should scale linearly with body mass, but
that when including soft mouth parts in the equation, bite volume should increase much faster
with body mass than expected in order for larger herbivores to compensate for the dilution of
plant mass in space. Bite volume refers to the effective volume of the plant from which forage
is removed, which consists of three dimensions: bite depth and bite surface area (Ungar et al.
1991, Gordon & Lascano 1993). The bite area covered can be summarized as the extension of
the tongue (or lips) together with the gape area of an open mouth, which defines the total area
swept (Ungar et al.1991). In our study, mouth volume is calculated as the sum of intra-dental
mouth volume and the volume covered by the soft mouth parts, which are assumed to be the
length of the longest mouth part raised to the power of three.
We expect bite volumes as measured from locations where herbivores actually fed in the
wild to scale linearly with mouth volume and in a similar linear fashion as mouth volume with
body mass. Generally, leaves have higher quality than stems as their fibre content is less
(Demment & van Soest 1985). Therefore, although larger herbivores might have larger bites
in the bitten off food mass, bite quality will decrease because the fraction of leaf biomass is
expected to decrease with increasing bite size according to plant mass-space laws (Niklas
2004). As the function of tongue and lips have also been implicated to facilitate the selection
13
of soft plant parts like leaves (Hongo & Akimoto 2003, Searle & Shipley 2008), we further
predict that bite leaf mass will scale higher than ¾ powers and closer to one with body mass.
Methods
All data were collected in the northern parts of South Africa, northeast of Pretoria and
included six morphological characteristics, most commonly associated with feeding, from
both carcasses and skulls of dead or immobilized animals (Table 2.1). Herbivore species
included covered more than three orders magnitude of body mass (10-4850 kg) (Table 2.2).
Direct observations on herbivore foraging
Field observations on the foraging behaviour of herbivores were conducted at two spatial
scales: the bite and the feeding station (Bailey & Provenza 2008). Direct observations
typically involved locating foraging herbivores, picking a focal animal not more than 50 m
from the observer of which the head and front part of the body was clearly visible and
recording feeding activity of this individual using a digital camcorder. We only included
observations where the animal had to have been feeding for more than 60 sec and needed to
have visited at least two feeding stations. If the animal had not left within 5 min, it was chased
off to prevent consumption of all the forage. A bite was defined as the amount of forage
removed from a plant in a single cropping motion (Searle et al. 2005), and a feeding station as
the array of plants available to a herbivore from which bites could be taken without moving
its front legs (Novellie 1978). Video recordings were analysed to determine the number of
bites per feeding station and the bite rate.
At one of the actual feeding stations, fresh bite marks were identified and counted and
video material aided in finding back all bites within the feeding station. Further measurements
included the size of one of the bite marks the animal left behind, measured in three
dimensions to obtain the bite volume. Based on these bite measurements, a similar place was
located within the foraged plant, and a simulated bite with the same dimensions and at the
same feeding height was removed by hand. Simulated bite samples were stored in paper bags
and dried at 70˚C for 24 h after which leaf and stems were separated and weighed. Following
this, the bite samples were grounded through a 1mm sieve for further chemical analysis at the
laboratory of the Resource Ecology Group, Wageningen University (Netherlands). Dry matter
(DM) and organic matter (OM) contents were determined first by drying samples at 105°C
overnight (for DM content) and afterward ashing the same samples at 505°C for 3 h (for OM
content). ADF was measured using the ANKOM200 filter bag technique (ANKOM
Technology, Macedon, NY, USA). The Acid Detergent Solution was prepared following
Goering and van Soest (1970).
14
Data analysis
From observations on foraging, we tested whether bite biomass and bite leaf mass scaled
linearly with body mass and whether bite rate scaled as BM-0.25:
Bite rate (bites/s) = number bites per feeding station / time between stepping between two
feeding stations
where feeding station was defined as the array of plants available to a herbivore from which
bites could be taken without moving its front legs (Novellie 1978). Instantaneous intake rates
were calculated in two ways. First, we used the conventional method where intake is
calculated as:
Intake rate (g forage ingested /s) = Bite mass (g) x Bite rate (bites/s)
Next, we included the effects of ADF and calculated instantaneous intake rate as:
Digestible cell mass (g) = Bite mass (g) x ((100 - %ADF) /100)
Intake rate (g digestible cell mass ingested /s) = Digestible cell mass (g) x Bite rate
(bites/s)
Finally, both intake rates were scaled against body mass.
To test whether larger herbivores had relatively bigger mouth sizes, and longer lips (or trunks),
tongues and necks, we used power functions of body mass scaled against the different body
parts associated with feeding, as geometric principles predict that all linear measurements
should scale to BM0.33. Following this, at the bite scale, we tested whether intra-dental mouth
volume, mouth volume and actual bite volume, as measured in the field, scaled linearly with
body mass according to the same geometric principles. Volumes from morphological
measurements were calculated as follows:
Intra-dental mouth volume (cm3) = Incisor width (cm) x Jaw length (cm) x Muzzle diameter
(cm)
Mouth volume (cm3) = Intra-dental mouth volume (cm3) + Maximum soft mouth part length
(lip or tongue) (cm)3
15
All power law equations used for scaling were calculated as follows:
Y = a BMb
where Y is expressed as a dependent function of body mass BM, a is a constant (a = 1 when
BM = 1) and b is the scaling factor (slope of the regression line).
For statistical analysis, curve estimations using power functions were used in SPSS followed
by an ANOVA analysis to test the significance of the fitted curves. To prevent bias because of
uneven sampling, mean values for each sex of each species were used in all analysis.
Table 2.1
Morphological forage characteristics of herbivores measured in this study
Characteristic
Description/method of measurement
Incisor width (cm)
Linear width of the incisor arc in the lower jaw. For animals without
incisors such as elephant the width of the gums were measured
Muzzle circumference
Circumference of the muzzle measured form just above the nasal cavity
(cm)
and around the lower jaw
(Muzzle diameter = Muzzle circumference/ (2π))
Tongue protrusion length
The length the tongue protruded outside the mouth, measured from the
(cm)
incisors/lower gum to the tongue tip, without too much resistance when
pulling the tongue out for measurement
Upper inner lip length
The length of the upper lip measured from the tip of the upper inner lip
(cm)
to the edge of the gum (for elephants this was measured as the length of
the trunk)
Jaw length (cm)
The distance between the gonial angle in the lower jaw and the mental
protuberance
Results
Most linear morphological measurements associated with feeding such as jaw length, neck
length and muzzle diameter scaled as predicted close to BM0.33, except for neck length which
scaled closer to 0.25, and maximum soft mouth part length with a scaling factor of 0.66
(Table 2.3, Figure 2.1).
16
Table 2.2
Large African herbivore data collected and used for analysis
Sample sizes used in analysis
from
Direct
Carcass Skull
observed
1
15
5
0
15
9
Herbivore Species
Sex
Adult body
mass (kg)
Steenbok
(Raphicerus campestris)
♂
♀
10
11
Duiker
(Sylvicapra grimmia)
♂
16
1
0
0
Impala
(Aepyceros melampus)
♂
♀
62
42
11
15
8
18
10
9
Warthog
(Phacochoerus africanus)
♂
81
1
0
1
Bushbuck
(Tragelaphus scriptus)
Blesbok (Damaliscus dorcas
phillipsi)
Nyala
(Tragelaphus angasii)
♂
♀
65
45
0
0
3
3
1
0
♂
70
1
0
0
♂
♀
♂
120
80
180
3
0
1
0
2
0
8
4
0
♂
180
1
0
0
♂
♀
♂
♀
221
160
200
150
1
0
3
3
1
3
0
0
5
10
5
3
♂
270
1
0
4
♂
♀
♂
♂
♂
♀
♂
♀
♂
♀
315
302
500
574
1184
827
2500
1600
4850
3500
0
1
3
6
4
0
0
1
16
3
2
0
0
3
0
2
1
0
0
0
5
6
0
4
6
10
6
6
15
10
Oryx (Oryx gazelle)
Red Hartebeest
(Alcelaphus caama)
Kudu
(Tragelaphus strepsiceros)
Wildebeest
(Connochaetes taurinus)
Waterbuck
(Kobus ellipsiprymnus)
Zebra (Equus burchelli)
Eland (Taurotragus oryx)
Buffalo (Syncerus caffer)
Giraffe
(Giraffe camelopardis)
White Rhinoceros
(Ceratotherium simum)
Elephant
(Loxodonta africana)
At the bite scale, intra-dental mouth volume scaled to body mass with a factor slightly
less than 1, whereas both mouth volume and actual bite volume as measured in the field
scaled to BM1.75 (Table 2.3, Figure 2.2).
17
Table 2.3
Results on the allometric scaling of morphological features and forage observations from large
herbivores
One
dimensional
morphological
features
Bite scale
characteristics
Feeding station
scale
characteristics
Instantaneous
intake rate
Dependent
variable
Incisor
width
Jaw length
Independent
variable
Body mass
Scaling
factor b
0.312
Constant
a
0.917
R2
Df
F
P
0.785
87.632
<0.01
Body mass
0.295
5.46
0.876
162.338
<0.01
Muzzle
diameter
Maximum
soft mouth
part length
Intra-dental
mouth
volume
Mouth
volume
Actual bite
volume
Bite mass
Body mass
0.308
1.065
0.837
87.445
<0.01
Body mass
0.626
0.345
0.787
1,
24
1,
23
1,
17
1,
18
66.484
<0.01
Body mass
0.888
6.889
0.859
1,
16
97.798
<0.01
Body mass
1.723
0.255
0.820
77.773
<0.01
Body mass
1.698
0.027
0.859
122.062
<0.01
Body mass
0.969
0.002
0.690
49.040
<0.01
Bite leaf
mass
Actual bite
volume
Bite rate
Body mass
0.997
0.001
0.705
52.507
<0.01
Mouth
volume
Body mass
0.879
0.280
0.855
76.928
<0.01
-0.024
0.37
0.007
1,
17
1,
20
1,
22
1,
22
1,
13
1,
20
0.142
0.710
Intake rate
(g / s)
Intake rate
(digestible
cell (g / s)
Body mass
0.903
0.001
0.689
47.440
<0.01
Body mass
0.751
0.001
0.573
1,
20
1,
20
26.837
<0.01
As expected bite mass and bite leaf mass scaled linearly with body mass. In accordance with
predictions, conventional instantaneous intake rate scaled linearly to body mass, whereas
instantaneous intake rates, including the effects of ADF, scaled as BM0.75 (Table 2.3, Figure
2.3).
18
(b)
In c is o r w id th (c m )
14
80
70
12
Jaw length (cm)
(a)
10
8
6
4
0.312
y = 0.9174x
2
R = 0.785
2
40
30
20
0.2951
y = 5.4598x
2
R = 0.8759
0
1000
2000
3000
4000
5000
6000
14
12
10
8
6
4
0.3084
y = 1.0646x
2
R = 0.8372
2
250
0
1000
2000
3000
4000
5000
6000
Body mass (kg)
Body mass (kg)
16
(d)
0
Maximum soft mouth part
length (cm)
18 0
M u z z le d ia m e te r (c m )
50
10
0
(c)
60
200
150
100
50
0.6255
y = 0.3453x
2
R = 0.7869
0
0
1000
2000
3000
4000
5000
6000
Body mass (kg)
0
1000
2000
3000
4000
Body mass (kg)
Figure 2.1: The scaling of body mass with linear morphological foraging characteristics of
large African herbivores including, incisor width (a), jaw length (b), muzzle diameter (c)
and maximum soft mouth part length (d).
19
5000
6000
(a)
14000
y = 6.889x0.888
R2 = 0.8594
3
Mouth volume (cm )
12000
10000
8000
.
6000
4000
2000
0
(b)
3
Predicted bite volume (cm )
10000000
0
1000
2000
1.7235
Body
y = 0.2546x
9000000
3000
4000
5000
6000
5000
6000
5000
6000
mass (kg)
2
R = 0.8206
8000000
7000000
6000000
5000000
4000000
3000000
2000000
1000000
0
(c)
180000
0
2000
3000
4000
Body mass (kg)
1.6979
160000
y = 0.0269x
2
R = 0.8592
140000
3
Bite Volume (cm )
1000
120000
100000
80000
..
60000
40000
20000
0
0
1000
2000
3000
4000
Body mass (kg)
Figure 2.2: The scaling of body mass of large African herbivores with intra-dental mouth
volume (a) and mouth volume (b) as calculated from animal morphology, and with mean
bite volume as measured in the field (c).
20
2.5
(a)
Intake rate (g DM / second)
0.9031
y = 0.0009x
2
R = 0.7034
2
1.5
1
0.5
0
(b)
0
1000
2000
0.7
Intake rate (g digestible
cell /second)
0.7507
y = 0.001x
2
R = 0.573
0.6
3000
4000
5000
6000
5000
6000
Body mass (kg)
0.5
0.4
0.3
.
0.2
0.1
0
0
1000
2000
3000
4000
Body mass (kg)
Figure 2.3: The scaling of body mass (kg) with intake rate expressed as dry matter mass (g/s)
(a) and intake rate expressed as digestible cell mass (g)/sec; (b).
21
Discussion
Although most mouth morphological features scaled geometrically as BM0.33, maximum soft
mouth part length scaled higher, closer to BM0.66 in contrast to what we believed on basis of ,
for example, Hutchinson (1959). As a consequence of this and as predicted by Clutton-Brock
and Harvey (1983) and Owen-Smith (1985), bite mass scaled linearly with body mass. Other
studies conducted in the past on this type of scaling contradict our findings. For example,
Shipley and co-workers (1994) found bite mass to scale to BM0.72, whereas Fortin (2006)
found a scaling factor of 0.64. However, these studies were conducted in zoos, which did not
allow for animals to make choices between plants occurring in their natural habitats. Bite rate
did not scale with body mass possibly because larger animals such as white rhinoceros do not
move their heads when feeding but instead use their lips to increase bite rates (Owen-Smith
1988) probably because muscle movement required to move the head for such a large animal
would require too much energy resulting in decreased bite rates. We found intra-dental mouth
volume to scale to body mass with a factor slightly less than one, but because mouth volume
and actual bite volume both scaled as BM1.75 and bite volumes of elephant were even higher
than the predicted curves, we reason that the increase in bite volume in larger animals is a
result of elongated soft mouth parts.
The linear scaling of bite leaf mass with body mass at first sight seems to defy logic. To
understand this relationship we have to first understand what happens with plant mass in
space. According to scaling laws found in plant ecology, average plant mass scales as plant
density0.75 and standing leaf biomass as standing stem biomass0.75 (Enquist et al. 1998, Niklas
2004). Similarly, in South African savannas Drescher (2003) found the proportion of leaves
and average nitrogen content in plants to decrease with increasing plant mass. Large bites
usually encompass more fibrous plant parts, which reduces digestibility (Shipley & Spalinger
1995), because structural tissues of plants increase from the distal to the proximal parts of the
plant (Hjeljord 1987, Hubbert 1987) and nutritional quality of twigs decrease with increasing
diameter (Palo et al. 1992, Jia et al. 1995). Bite diameter cropped by African thicket browsers
scales with body mass in the same allometric manner as incisor breadth, as BM0.3 (Wilson &
Kerley 2003). The implications of these findings and scaling relationships for herbivores, is
that larger herbivores will not only have less plant biomass available to them in relation to
their body mass but will also take in forage of a lower quality as the fraction of stem material
will be larger and the leaf mass fraction will be lower. We reason that to compensate for the
dilution of plant mass and quality in space, enlarged soft mouth parts not only allows larger
animals to be more selective but also to cover a bigger area in one bite, which increase both
leaf mass and total bite mass.
As expected, instantaneous intake rate, calculated in the conventional way as the product
of bite rate and bite mass, scaled linearly with body mass. In disagreement, and closer to what
22
we expected when using digestible cell mass per bite to calculate instantaneous intake rate,
Shipley and co-workers (1994) found maximum instantaneous intake rate to scale as BM0.71.
However, this might be explained by the fact that digestibility was already artificially
corrected for in Shipley’s zoo experiment as all animals were fed alfalfa plants of the same
quality. In the wild, herbivores of different body mass are presented with a wide variety of
plants with different leaf to stem ratios, and a different spatial structure of the forage biomass.
Stems have been found to serve as both vertical (Ginnett et al. 1999) and horizontal (Drescher
et al. 2006) barriers reducing bite depth, bite mass and therefore instantaneous intake rate. As
stems contain more structural compounds, such as lignin and cellulose than leaves, they
decrease bite quality (Searle & Shipley 2008). Recently, van Langevelde and co-workers
(2008) developed a model that for the first time included quality as a factor affecting the
functional response of animals. Our results for instantaneous intake rate as obtained from the
inclusion of ADF to calculate the digestible cell mass available to a herbivore, supports this
notion of including quality in the calculation of instantaneous intake rate. ADF is frequently
used as a proxy for fibre and fibre is the most common variable used to predict energy content
of feeds (Weiss 1993). Hence, the scaling of digestible cell mass intake to BM0.75 is in
accordance with the scaling of metabolizable energy requirements to BM0.75 (Kleiber 1932).
The spatial distribution of a herbivore’s habitat, and food resources can be described well
with fractal geometry as this distribution is often self-similar across ecologically relevant
ranges of scales (3-4 orders magnitude) (Ritchie & Olff 1999). In both plants and animals,
species richness versus organism size follows a left-skewed, unimodal distribution because
larger species are limited by the maximum patch size in the environment (Ritchie & Olff
1999). Although some plants have adapted to herbivory, it is generally accepted that the
evolution of plants preceded mammals (Janis 1993), meaning that the same chemical and
mass characteristics and patterns in plants should also be apparent in the animals feeding on
them. Evidence of these similarities exists in the universal allometric scaling of resource use
and metabolic rate in both animals and plants to BM0.75 (West et al. 1997, Enquist et al. 1998).
Development of morphological features such as trunks takes millions of years (Shoshani
1993). More over, sudden climate changes such as those that occurred around the end of the
Eocene and the beginning of the Oligocene, which have been implicated to have caused
changes in plant abundance and greater differentiation of fibre content between plant leaf and
stem, coincide with extinctions of mega-herbivores (Janis 2008). We hypothesize that these
extinctions occurred because large herbivores did not have suitable mouth morphologies to
cope with the dilution of food resources in space. We conclude that mega-herbivores of today
such as giraffe, white rhinoceros and elephant have extreme mouth morphologies and sizes as
a result of natural selection to adapt to the spatial mass and quality distribution patterns of
plants.
23
24
CHAPTER 3
SOIL NUTRIENT STATUS DETERMINES HOW
ELEPHANT UTILIZE TREES AND SHAPE
ENVIRONMENTS
Y. Pretorius, W.F. de Boer, R. Slotow, C. van der Waal & H.H.T. Prins
25
Abstract
Establishment of the mechanism determining the spatial scale of patch selection by herbivores
have been complicated by the way in which resource availability at a specific scale is
expressed and by vigilance behaviour of the herbivores themselves. To reduce these
complications we chose to study patch selection by an animal with negligible predation risk,
the African elephant. Further, we introduce the concept of nutrient load as the product of
patch size, number of patches and local patch nutrient concentration, which offers a novel
way to express the total available nutrients a herbivore can select from per grain of a specific
scale. We hypothesized that elephant will be able to select nutrient rich patches based on the
nutrient load per 2500 m2 down to the individual plant scale and that this selection will
depend on the nitrogen and phosphorous content of plants. Further, we predicted that elephant
will cause more injurious impact to trees of lower value to them in order to reach plant parts
with higher nutrient concentrations such as bark and root but will maintain nutrient rich trees
by inducing coppicing of trees through re-utilization of leaves. Elephant patch selection was
measured in a homogenous tree species stand by manipulating the spatial distribution of soil
nutrients in a large field experiment using NPK fertilizer. Our results showed that elephant
were able to select nutrient rich patches and utilized Colophospermum mopane trees inside
these patches more than outside, at scales ranging from 2500 m2 down to 100 m2. Although
both nitrogen and phosphorus content of leaves from C. mopane trees were higher in fertilized
and selected patches, nitrogen had the strongest correlation with patch choice. As for the
impact of elephant on trees, stripping of leaves occurred more in nutrient rich patches,
whereas injurious impact such as uprooting of trees occurred more in nutrient poor areas. Our
results shed some light on how future studies should interpret patch selection by herbivores at
different scales and how elephant foraging behaviour can be used as indicators of changes in
the availability of nutrients.
Keywords: patch selection, Colophospermum mopane, nitrogen, nutrient load, spatial scale,
re-use
26
Introduction
An herbivore that is selective and able to discriminate between food of good and bad quality
should have a selective advantage (Fryxell 2008). However, food selection is complicated in
that, unlike with a carnivore, a large herbivore’s food is much less concentrated and is
distributed as a nested hierarchy of aggregated resources, which vary widely in nutrient
composition and mass (Senft 1987, Kotliar & Wiens 1990, Bailey et al. 1996, Searle et al.
2005, Fryxell 2008). These aggregated resources can be defined as patches that are discrete
spatial units differing from their surroundings in nature and/or appearance and that cause
changes in a herbivores foraging behaviour (Kotliar & Wiens 1990, Searle et al. 2005).
Patch selection by herbivores further depends on the spatial scales at which the
environment is perceived, where ‘grain’ is the smallest scale at which a herbivore responds to
patch structure and ‘extent’ the largest scale of heterogeneity to which a herbivore responds
(Kotliar & Wines 1990). In this study, we will test whether a large herbivore is able to select
nutrient rich patches at multiple scales as boundaries between the sub-units within different
hierarchical scales should be defined by the animals perceptions and foraging responses
(Senft et al. 1987). However, within a specific spatial scale, patches are not simply distributed
uniformly but vary in size, number and local nutrient concentration (Bailey & Provenza 2008).
Therefore, at each scale, the resource value per grain should be expressed as the total nutrient
load, whether it be areas with large nutrient rich patches, areas with many small nutrient rich
patches or areas with patches with very high local nutrient concentrations. In our study spatial
scales will range from 2500 m2, where a herbivore select between areas with different total
nutrient loads, to selection between individual neighbouring plants of the same size and
species (Figure 3.1).
Studying the mechanisms behind patch use is complicated by the effects of predation risk.
For example, the Marginal Value Theorem (MVT) predicts that foragers will depart from a
single patch when their instantaneous intake rate drops below the average rate of intake
attainable in all patches (Charnov 1976). However, increased predation risk should affect
MVT in that giving up densities should increase (Brown 1999). Therefore, as a study animal,
we chose the largest terrestrial herbivore, the African elephant, for which predation risk in our
study area is negligible.
Studies on patch selection of herbivores in relation to its nutrient status have shown that
herbivores are not only able to select nutrient rich patches but are also able to do so at
different scales (Wallis deVries 1999, Ball et al. 2000, Cromsigt & Olff 2006, Chapter 2).
Elephant in particular have been found to be able to select vegetation growing on termite
mounds at very small scales (Holdo & McDowell 2004). Therefore we predict that
27
elephant will be able to select nutrient rich areas at scales ranging from 2500 m2 down to the
individual plant scale.
Figure 3.1: A schematic representation of the study set-up illustrating the spatial scales at
which experiments were conducted.
Many studies show that savannas are either limited by N or P or co-limited by both (du
Toit et al. 1940, Weir 1969, Ludwig et al. 2001, Snyman 2002, Augustine et al. 2003, Cech et
al. 2008). Because non-ruminants such as the elephant cannot make use of microbial protein
and have high food passage rates, they can be expected to incur higher losses of N in the
faeces (Foose 1982). Selection for N and P by elephant have been described by Jachmann &
Bell (1985) and from diet studies on elephant at our study site in South Africa, we know that
elephant maximize N and P intake depending on the time of the year (Chapter 4). Hence, we
expect elephant to select patches and plants of the same species depending on their N and P
content.
28
The impact African elephant have on trees have been a topic of great controversy
(Wiseman et al. 2004, de Beer et al. 2006, Lawes & Chapman 2006, O’Connor et al. 2007,
Chafota & Owen-Smith 2009, Scholes & Mennell 2008). Elephant have been implicated as
one of the key factors maintaining low tree-grass ratios in savannas (van de Koppel & Prins
1998). However, few studies have investigated the causal mechanisms of elephant impact,
especially the interactions with nutrient availability and distribution (Skarpe et al. 2004).
Injurious impact on trees when bark and roots are consumed is not unique to elephant and has
been described in voles (Microtus spp.) (e.g., Sullivan et al. 2004), Sika deer (Cervus nippon)
(e.g., Yokoyama et al. 2001) or red deer (Cervus elaphus) (e.g., Verheyden et al. 2006).
However, consumption of bark and roots has mostly been attributed to the lack of good
quality alternative food sources, especially during winter, rather than to the high nutritional
quality of these plant parts (Servello 1984, Bucyanayandi et al. 1992, Verheyden et al. 2006).
Similarly, elephant impact on trees also is especially prevalent during the dry season when
elephant switch from a diet dominated by grass to browse because grass quality decrease to
below the animal’s maintenance requirements during this time (Barnes 1982, Beekman &
Prins 1989, Kos et al. 2008). Hence, according to O’Connor and co-workers (2007), an
increased consumption of woody material indicates nutritional stress. We predict that elephant
will have a larger impact on trees that represent low quality food through, for example,
increased utilization of roots, whereas impact on trees growing in nutrient rich patches will be
less, with leaves being utilized more.
Maintenance of nutrient rich patches (so called grazing lawns: Vesey-FitzGerald 1969) is
especially common among grazing animals like hippopotamus (Hippopotamus amphibious)
and white rhinoceros (Ceratotherium simum) (Verweij et al. 2006, Waldram et al. 2008) and
consumption of fresh re-growth along with short re-visitation intervals have also been
observed during spring time for grazers such as Brent geese (Branta bernicla) (Prins et al.
1980) and buffalo (Syncerus caffer) (Prins 1996). Elephant have been shown to prefer
previously browsed trees (Anderson & Walker 1974; Jachmann & Bell 1985; Lewis 1991)
and forage quality and quantity have been shown to increase with repeated herbivory (du Toit
et al. 1990, Rooke et al. 2004, Fornara & du Toit 2007). For Colophospermum mopane trees,
Smallie and O’Connor (2000) found that elephant prefer previously hedged trees and induce
coppicing of these trees. We thus predict that elephant maintain nutrient rich patches of C.
mopane trees and expect coppiced trees, occurring as a result of previous hedging by elephant,
to occur more in nutrient rich areas and to have more fresh signs of leaf utilization than in
nutrient poor areas. To test these predictions, we experimentally manipulated the spatial
distribution of nutrients in a homogenous stand of plant species dominated by C. mopane
trees.
29
Methods
In December 2004 we fertilized thirty 50 m x 50 m plots in the Timbavati Private Nature
Reserve (TPNR), South Africa (24˚14'11"S; 31˚22'32"E) with NPK (3:2:1) fertilizer. The area
was re-fertilized two years later at the same fashion. The experimental study site was
dominated by a homogeneous layer of C. mopane shrub veld on granitic soils. The herbaceous
layer was dominated by Urochloa mosambisencis and Bothriochloa spp. (for a full
description of the vegetation see van der Waal, in prep). The warm, rainy season stretched
from October to March and during the study period the mean annual rainfall was 420 mm.
Data were collected from October 2005 to March 2008.
Surface water was not constraining plot use, as it was available within 0.5 km north
and south from the closest experimental plot. The experimental layout followed a randomized
block design including three replicates of seven different treatments and nine controls. One
control was allotted to each scale treatment and replicated three times. The fertilizer was
applied at each 50 x 50 m plot at either one of three different spatial configurations: one large
50 x 50 m patch, five 10 x 10 m patches, or 25 patches of 2 x 2 m, using either one of three
different nitrogen concentrations (30.0 g/m2, 6.0 g/m2, 1.2 g/m2). This resulted in seven
combinations and one of three different total nutrient loads per 50 x 50 m plot (0.6 kg N, 3 kg
N, 15 kg N), as the lightest and heaviest fertilizer concentrations were excluded, as these were
expected to generate too faint signals or toxic effects.
Data collection on the type of nutrients selected by elephant and the scale of nutrient
selection consisted of annual measurements during the wet seasons from 2006 to 2008 on the
leaf nutrient content and accumulated signs of elephant utilization on 600 marked C. mopane
trees and direct observations on elephant selecting between two C. mopane trees of the same
size within areas surrounding the experiment, which were also dominated by C. mopane. Data
on the way elephant utilized trees and the occurrence and utilization of coppiced C. mopane
trees in and outside fertilized patches were obtained from an elephant utilization assessment
on all trees larger than 1.5m within the experimental plots at the end of the study.
Annual mopane tree measurements
Tree responses were monitored in 600 marked (with aluminium tags) C. mopane trees. In
controls and in whole-plot fertilizer treatments, C. mopane trees (> 1m height) closest to
twenty randomly picked points were selected throughout each plot. In heterogeneous
treatments (patch size either 10 x 10 m or 2 x 2 m), ten trees with stems within 2 m distance
of fertilized patches were randomly selected, together with ten trees throughout the
unfertilized plot area (> 2m distance from fertilized patches). Leaf samples from C. mopane
trees were collected during the growing seasons of 2006, 2007 and 2008. Five fully expanded
leaves were randomly collected from the canopies of the marked trees. In homogeneous
treatments, two pooled samples were analysed per plot. In heterogeneous treatments, samples
30
were pooled for leaves collected from the plants in and outside fertilized patches. This
resulted in sixty samples per year. Prior to milling (through a 1 mm sieve), C. mopane leaves
were dried to constant weight at 600C and weighed. All marked C. mopane trees were
assessed for elephant impact (percentage canopy volume reduction) using the scale of
Anderson & Walker (1974). In 2006 the height and diameter of marked trees were calculated
from digital photographs. At the end of the experiment in 2008, a selection of trees was remeasured to determine tree height changes in relation to visual elephant impact scores. For
analysis, as ten trees were measured per patch, elephant utilization was also expressed as the
proportion of trees utilized per patch.
Direct observations
All elephant observations conducted in areas dominated by C. mopane surrounding the
experiment were of randomly encountered adult solitary bulls or bulls in small bachelor
groups. All data were collected during the wet season of 2006 and were spatially and
temporally independent as only one paired sample were collected at each sighting of a
particular elephant or group of elephant on a particular day. Each elephant was located using
the extensive road network on the reserve and observed from a vehicle for between 5 and 10
min (or until the animal moved out of site) using binoculars. For an observation to be
accepted, an elephant had to be foraging in a homogeneous area of C. mopane trees. Only
once an elephant moved from the tree it was first observed feeding on, to feed on the next tree,
was the next tree used for measurement. Measured trees were all C. mopane, between 3-6m,
and the elephant had to have taken at least five bites from the canopy, before the tree was
considered a appropriate tree to be sampled. For each of the measured trees eaten from, a
discarded paired control tree was selected. This discarded tree had to be a C. mopane of
similar height and canopy size, which the elephant had to have walked past (within 2 m) and
not selected while moving towards the tree eaten from. About 50 g of leaf material was
collected from various branches on each tree at the same height range at which the elephant
fed from. Leaves from both the tree eaten from and the discarded tree were stored in separate
paper bags, dried at 60 ºC for 24 h and ground through a 1mm sieve and stored for further
analysis.
Chemical analysis of leaf samples
All dried leaf samples were analyzed for dry matter, ash, neutral detergent fibre, calcium,
sodium, potassium, magnesium, nitrogen and phosphorus content at the laboratory of the
Resource Ecology Group, Wageningen University (the Netherlands). Total Ca, Na, Mg, K, N
and P content were measured with a Skalar San-plus auto analyzer, after destruction with a
mixture of H2SO4, Selenium and salicylic acid. (Novozamsky et al. 1983). Neutral detergent
fibre of dry leaf was determined using the ANKOM filter bag procedure (ANKOM
Technology Macedon. NY, USA) with omission of the sodium sulphite and the heat resistant
31
α-amylase. A neutral Detergent Solution was prepared following Goering & van Soest (1970).
Condensed tannins were measured with the Proanthocyanidin method and total polyphenol
with the Folin-Ciocalteu method (Waterman et al. 1994).
Elephant impact assessment
All trees higher than 1.5 m in 50 x 50 m plots containing five fertilized patches of 10 x 10 m
each were assessed during March 2008: at each tree the position of the tree in the 50 x 50 m
plot was recorded within a 2 x 2 m grid. Fertilized patches were marked at the beginning of
the study with iron stakes, to indicate whether a tree occurred within a fertilized patch or not.
The tree species, type of treatment within the 50 x 50 m plot, tree height and canopy width,
the occurrence of coppicing as a result of visible previous impact by elephant, percentage of
each type of utilization, total percentage impact and estimation of the age of impact were
recorded. The area on a branch impacted on by elephant becomes grey after a year through a
rainy season (Ben-Shahar 2002). Therefore, the age of impact on a tree was classified as fresh
when scars had not yet turned grey and old, when it had. The type of utilization was
categorized into five classes: leaf stripping, bark stripping, impact on branches, uprooting, or
breaking of the main tree trunk. The percentage leaf stripping, bark stripping and impact on
branches were estimated in relation to the availability of the plant part across the entire tree.
Uprooting and breaking of the main tree trunk were recorded as 100% impact.
Data analysis
All data were first tested for normality using the Kolmogorov-Smirnov test and where
proportions were used data were first arcsine transformed. Two types of analysis were used to
test elephant selection for specific nutrients. First, we used a one-way ANOVA to test
whether N, P, K, Mg, Ca and Na were higher in leaves from trees in fertilized patches
compared to unfertilized patches and whether elephant utilized these trees in fertilized patches
more than unfertilized patches. Second, we conducted a multiple linear regression analysis to
test for correlations between the concentrations of different nutrients as independent variables
and the proportion of trees utilized by elephant per patch as dependent variable. Annual
elephant utilization measurements on marked C. mopane trees were used in an ANOVA to
test whether elephant were able to select fertilized patches at a 50 x 50 m scale and at a 10 x
10 m scale. For analysis at the plant scale between neighbouring trees, paired sample t-tests
were used to compare nutrient levels in leaves between eaten and uneaten trees. Because leaf
stripping (%) and signs of elephant utilization on branches (%) were not normally distributed,
the non-parametric Mann-Whitney test was used to detect differences between fertilized and
unfertilized patches. As the number of coppiced trees and trees killed by elephant are rare
events with many zero values, a generalized linear model with a negative binomial error
distribution was used to test for differences of these variables inside and outside fertilized
patches.
32
Results
The application of the NPK fertilizer significantly changed the chemical composition of
leaves from mopane trees at our experimental site. Both nitrogen (F1,58 = 35.566, p < 0.001)
and phosphorus (F1,58 = 28.226, p < 0.001) content of tree leaves were significantly higher in
fertilized patches compared with unfertilized patches, whereas the opposite was true for the
tannin content of leaves (F1,58 = 13.628, p < 0.001). However, no significant difference could
be found for K (even though the fertilizer contained K), or Ca, Mg, Na or NDF concentration
in leaves. Irrespective of patch size, the proportion of mopane trees utilized by elephant per
patch was significantly higher in fertilized patches as compared to unfertilized patches (F1,58 =
10.95, p < 0.05) although the proportion of trees utilized by elephant per patch was only
positively correlated to the nitrogen concentration in leaves and not with the other nutrients
Proportion trees utilized/patch
(linear regression model: R2 = 0.462, F7,52 = 6.368, p < 0.001) (Figure 3.2).
A
1.0
1.00
A
0.8
0.80
A
A
0.6
0.60
A
0.4
0.40
A
0.2
0.20
A
A
A
A
A
A
A
AAAA A
A
2.0
A
A A
AA
A
A
AA A AA
A A
A
AA
A A
1.8
A
AAAA
AA
A
A
AA
A
A
2.2
2.4
Percent nitrogen (DM)
Figure 3.2: The relationship between the proportion Colophospermum mopane trees utilized
by elephant per patch and the percentage nitrogen (on a dry matter basis) in tree leaves
(n = 60).
At both the 2500 m2 (F3,56 = 4.564, p < 0.05) and 100 m2 (F1,22 = 9.743, p < 0.05) scales
elephant were able select the plots with the highest nutrient loads as utilization of mopane
trees were significantly higher at these plots (Figure 3.3). At plant scale, although N and P
content from leaves of eaten trees were higher than leaves from neighbouring trees of the
same species that were not eaten, these differences were not significant (N: t14 = -1.242, p =
0.235, P: t14 = -1.61, p = 0.13).
For the elephant utilization assessment at the 100 m2 scale, utilization of tree leaves via
leaf stripping occurred more inside fertilized patches than outside (Z = -4.694, n = 786, p <
33
0.001), whereas total impact on branches did not differ between patches (Z = -0.946, n = 786,
p = 0.344) (Figure 3.4a & 3.4b). Tree killing, through uprooting of trees or breaking main tree
trunks, occurred more outside fertilized patches than inside (Wald chi-square = 3.818, df = 1,
p < 0.05) (Figure 3.4c).
Tree coppicing occurred equally frequent inside fertilized patches as in the surrounding
areas (Wald chi-square= 3.181, df=1, p=0.075). As expected, coppiced trees had significantly
more signs of fresh leaf stripping than non-coppiced trees (Wald chi-square=5.398, df=1,
p<0.05) (Figure 3.5).
b
n=9
(b)
0.6
]
0.4
0.2
]
n= 9
n= 15
0.0
Absent
Proportion trees utilized/patch
(a)
Proportion trees utilized/patch
n = 15
1.0
]
ab
ab
0.5
a
Present
0
n = 42
0.4
]
0.3
0.2
]
0.1
0.00
Absent
120
600
3000
concentrationgrams
Nutrient
load (g N / 10 x 10 m)
(d)
b
Proportion trees utilized/patch
Proportion trees utilized/patch
n = 18
]
0.0
NPK fertilizer
NPK Fertilizer
in 10 x 10 m
(c)
]
]
0.6
ab
ab
]
0.4
]
]
600.00
3000.00
3000
a
0.2
]
0.0
0.00
0
1.00
Present
600
15000.00
15000
Nutrient
Nutrient load
load (g
(g N /2500m2)
/ 50 x 50 m)
fertilizerscale
NPK Fertilizer
in 50 x 50 m
Figure 3.3: Proportion of Colophospermum mopane trees used by elephant in fertilized and
unfertilized patches (panels a and c) and in patches with increasing nitrogen loads
2
(panels b and d) at 100m2 (panels a and b) and 2500 m (panels c and d) scale (error
bars show 95% confidence interval of the mean and different letters in the figures denote
significant difference).
34
Percentage
leaf
strip
strip
leaf
Percentage
(a)
2.5
]
2.0
1.5
1.0
0.5
]
(c)
Percentage
impact
branch impact
Percentage branch
(b)
0.00
Absent
Fertilizer
14
13
]
12
]
11
10
9
6
0.06
rtion trees
Propo
Percentage
treeskilled
killed
1.00
Present
0.00
Absent
1.00
Present
Fertilizer
]
0.04
4
0.02
2
]
0.00
0
0.00
Absent
1.00
Present
Fertilizer
2
Figure 3.4: Differences in elephant utilization at a 100 m scale between fertilized (n=176)
and unfertilized (n=610) patches, distinguishing different use-categories: (a) leaf stripping,
(b) impact on branches
and (c) killing of trees respectively (error bars show 95%
confidence interval of the mean).
35
0.15
15
15
1.5
(b)
]
0.10
10
]
5
0.05
n=176
n=610
Percentage
Proportion leaf strip
trees
coppicingtrees
Percentage
Proportioncoppicing
(a)
1.0
0.5
]
]
n=757
n=29
0.00
No
1.00
Yes
0.0
0.00
Absent
1.00
Present
Fertilizer
Coppicing
Coppicing tree
tree
Figure 3.5: Differences in the percentage coppicing trees in the presence or absence of
fertilizer (a) and signs of fresh leaf stripping on coppiced and non-coppiced trees (b) (error
bars show 95% confidence interval of the mean).
Discussion
Two decades ago, Senft and co-workers (1987) introduced a hierarchy theory into herbivore
foraging ecology to integrate foraging decisions at different spatio-temporal scales as the
application of traditional optimal foraging theory were problematic. Since then, many studies
have illustrated how herbivores of various body mass are able to select nutrient rich patches at
a range of spatial scales (WallisdeVries 1999, Durant et al. 2004, Cromsigt & Olff 2006). In
deed, in this study we also found that elephant are able to select nutrient rich patches at scales
ranging from 2500 m2 to 100 m2.
Contradictory to the theoretical predictions of Ritchie and Olff (1999) that larger species
will not be able to detect food patches at fine scales, we found that even the largest terrestrial
mammal are able to select nutrient rich plant parts at fine scales within a homogeneous tree
species stand. To our knowledge this has never been shown for elephant. At a regional scale
(100 km), in a study conducted in C. mopane woodlands in Botswana, no relationship could
be found between elephant impact on C. mopane trees and leaf N content even though soil N
levels and leaf N were related and leaf N differed significantly between sampling sites (BenShahar & McDonald 2002). However, this latter study was conducted at larger scales than our
study, no analysis was done on the relationship between the type of elephant impact and the
nutrient status of the plants and soil and leaf samples were collected in the dry season when
elephant are more limited by energy (Chapter 4) and thus less selective of other nutrients such
as N and P.
36
Selection of plant species high in N have been described for elephant within their daily
foraging ranges (Jachmann & Bell 1985), for barnacle geese (Branta leucopsis) in fertilizer
experiments at 250 m2 scale (Ydenberg & Prins 1981) and at even smaller scales black
colobus monkeys (Colobus satanas) in the rain forests of Cameroon have been found to select
plant parts in relation to their nitrogen content (McKey et al. 1981). Some authors reason that
switching between plant species or plant parts is a behavioural adaptation to cope with
declines in N content (Mattson 1980) and that elephant in particular switch between plant
species and plant parts to sequester the greatest amount of digestible protein per unit time
(O’Connor et al. 2007). For example, cambium in the bark and roots of plants contain
relatively high levels of N (Mattson 1980) and Hiscocks (1999) found that cambium samples
from bark and root that elephant utilized were higher in N than cambium from unutilized trees.
Therefore, uprooting of trees in nutrient poor areas, such as was found in our study, could be a
result of low leaf N levels, which encourages elephant to utilize other plant parts with higher
N content.
Irrefutably, our study’s biggest contribution to the current body of knowledge on optimal
foraging theory is the concept of patch selection by a herbivore based on the total nutrient
load within the grain size of a specific scale. In the real world patch size do not simply
increase uniformly but patches vary in size and can be scattered or aggregated depending on
the scale of observation. More over, the local nutrient concentration between patches can also
vary widely (Fryxell 2008). Thus when studying patch selection by a herbivore at a particular
scale, the available nutrients must be expressed as total nutrient load per grain size, which is a
product of the number of patches, local nutrient concentration of patches and patch sizes. We
found that at a scale of 2500 m2, elephant are able to select the patch combinations resulting
in the highest total nutrient loads, whether it was many small patches of high local nutrient
concentration per 2500 m2 or few large patches of lower local nutrient concentration per 2500
m2. These findings may not only have implications for the way patch selection by herbivores
are interpreted at different scales, but also offers a possible explanation why some studies find,
contrary to the Marginal Value Theorem, that herbivores select sub-optimal patches or leave
patches when there is still much food left (Schaefer & Meissier 1995).This dependence of
Marginal Value Theorem predictions on the scale at which a herbivore perceives patches,
which is very difficult to determine, has also been noted by Kotliar and Wiens (1990) and
Searle and co-workers (2005).
Finally, our study illustrate how elephant can possibly facilitate other species by
maintaining tree ‘islands of fertility’(Ludwig et al. 2008, Treydte et al. 2009) through re-use
of trees and dung deposition, which may potentially initiate positive feedbacks into the abiotic
environment (van der Waal, in prep). Jachmann and Bell (1985) and subsequently Smallie
and O’Connor (2000) hypothesized that by continuously hedging C. mopane trees elephants
37
can maintain a supply of preferred coppice growth through time. The increased proportion of
coppiced mopane trees due to previous breaking of branches by elephant inside fertilized
plots and the higher occurrence of leaf stripping on these trees indicate that the fertilizer in
our experiment might have triggered a positive feedback response that may persist through
time. In a fenced off clipping experiment conducted in northern Botswana elephant broke into
the experimental site three years after initiation. Assessment of elephant impact one month
after the break-in revealed higher utilization of trees subjected to simulated browsing three
years before compared to control trees (Makhabu 2006).
Often, the problem with studying keystone species in the wild is that their effects on other
organisms only become prevalent once the keystone species are removed as was demonstrated
by the removal of impala by disease followed by a subsequent increase in woodland cover in
East Africa (Prins & van der Jeugd 1993). The advantage of an experimental approach is that
the mechanisms behind these effects become clearer without removal of keystone species.
Both soil fertility and elephant have been identified as major determinants of tree-grass ratios
in savannas (Scholes 1990, Fritz et al. 2002, O’Connor et al. 2007, Sankaran et al. 2008).
However, the interaction between these two factors and more specifically how soil nutrient
status affect the way in which elephant utilize their environments were until recently unclear
(Olff et al. 2002, Skarpe et al. 2004). We aimed to better understand this mechanism but more
importantly our results shed some light on how future studies should interpret patch selection
by herbivores at different scales.
By concentrating limiting resources, patchiness may play a critical role in maintaining
ecosystem productivity (Aguiar & Sala 1999) and herbivory have been shown to create spatial
heterogeneity and induce vegetation patterning through the process of self facilitation (Groen
2007). We conclude that more emphasis should be placed on the nutrient status of soils when
trying to understand the impact elephant have on trees (Fritz et al. 2002) as changes in
elephant foraging behaviour can be used as indicators of changes in the availability of
nutrients.
38
CHAPTER 4
LARGE HERBIVORE RESPONSES TO NUTRIENT
HETEROGENEITY IN AN AFRICAN SAVANNA
Y. Pretorius, W. F. de Boer, M.B. Coughenour, H.J. de Knegt, H. de
Kroon, C.C. Grant, I. Heitkonig, E. Kohi, E. Mwakiwa, M.J.S. Peel, A.K.
Skidmore, R. Slotow, C. van der Waal, F. van Langevelde, S.E. van
Wieren, H.H.T. Prins
39
Abstract
Understanding how animals respond to spatial heterogeneity in resource distribution is central
to animal ecology. In this paper, we present a study to analyse at what spatial scale herbivores
select nutrient rich patches to maximize intake, and how an animal’s digestive system, body
mass, mouth size and herd size affects this selection. We tested whether the selection of areas
with the highest nutrient content occurs at different spatial scales for different herbivore
species. We hypothesized that smaller animals, animals occurring in large herds and ruminant
species will select areas of high nutrient concentration at small spatial scales whereas smallmouthed animals will select areas with high nutrient content at larger spatial scales. Using
fertilizer to manipulate soil nutrient levels, we tested our hypotheses. Nutrients were applied
in 2500 m2 plots at three different patch sizes and three different nutrient concentrations. This
resulted in differences in nutrient concentrations at the local scale and at the 2500 m2 plot
scale. Dung and spoor counts served as measures of animal abundance and indicators for
patch selection. Consistent with our hypotheses, large animals selected areas of high nutrient
concentration at larger spatial scales, whereas the abundance of ruminants decreased as
nutrient concentrations at large scales increased. As for individual species, duiker, warthog
and elephant selected for high nutrient concentrations at large scales while buffalo, zebra and
impala did so at the local point scale. Beyond selection of nutrient rich areas at small scales,
buffalo and zebra were the only species to increase in abundance with increasing patch size.
We explain the reaction of species to nutrient spatial heterogeneity using a conceptual model
that divides herbivores into two groups of animals based on their mouth to body mass ratios
and the total metabolic weight of a herd. The model shows that herd animals, with their
medium sized mouths, select for higher nutrient concentrations at small scales whereas
animals with extreme mouth to body mass ratios select for higher nutrient concentrations at
larger scales. We reason that due to competition and the risk of resource depletion, herd
animals may be forced to select areas of high nutrient concentration at the local point scale.
Keywords: herd size, metabolic weight, mouth size, patch selection, spatial scale
40
Introduction
Understanding foraging strategies that herbivores adopt at different spatial and temporal
scales is central to animal ecology (Fryxell, Wilmshurst & Sinclair 2004; Prins & Van
Langevelde 2008). However, these scaling effects have often been neglected in the past
(Owen-Smith 1994, van Wieren 1996; Illius et al. 1999). One of the first studies in the wild
on maximization of energy at different temporal scales, showed that wood bison (Bison bison
athabascae) maximize their energy intake in the short-term but not over longer time scales
(Bergman et al. 2001). More recent studies have attempted to understand the effects of spatial
scale on herbivore forage selection (Cromsigt & Olff) but some disparity still exist. In this
paper, we assessed how herbivores respond to changes in nutrient heterogeneity and at what
spatial scale a herbivore with a particular mouth, herd and body mass, will select areas with
higher nutrient concentrations.
Spatial scaling depends on the extent (range that study covers) and grain size (unit size
area are divided into) selected to study an area (Wiens 1989). Various spatial scales of forage
selection have been described ranging from bites, feeding stations and patches at small scales
to feeding sites at larger scales (Novellie 1978; Jiang & Hudson 1993; Laca, Ungar &
Demment 1994; Bailey et al. 1996). It has been shown that abiotic factors such as soil nutrient
content (Seagle & McNaughton 1992) and biotic factors such as forage quality and quantity
(Bailey et al. 1996) regulate the distribution and abundance of large herbivores (Augustine,
McNaughton & Frank 2003). The spatial distribution of nutrients is heterogeneous and varies
with both the size of nutrient patches and the concentration in these patches. An increase in
patch size could occur as result of redistribution of nutrients, for example by latrines of large
herbivores (Moe & Wegge 2008, van der Waal et al. 2009) and termite activity (Loverigde &
Moe 2004), which could lead to an increase in the area of plants with both higher biomass and
quality. An increase in nutrients at very small spatial scales results only in a local increase in
plant quality and biomass (Cromsigt & Olff 2006). The relative effects of patch size and
nutrient concentration on the selection of forage by herbivores are difficult to disentangle
(Cromsigt 2006). Different combinations of patch size and local nutrient concentrations will
result in different total nutrient loads at larger scales. For example, a high local nutrient
concentration will not necessarily result in a high nutrient concentration at larger scales
because the total nutrients at larger scales also depend on the number and sizes of nutrient rich
patches.
We reason, similar to Schoener (1969) that animals can only choose one of two foraging
strategies. Firstly, an animal selects the combination of patch sizes and local nutrient
concentrations at small spatial scales that will yield the highest nutrient concentrations at
larger scales. Under these conditions, the animal may need more time foraging because the
local small scale nutrient concentration may be low, but at larger scales, there will be more
41
nutrients available in total. Bison and elk, for example, have been found to feed randomly
within patches but select feeding sites based upon forage abundance at broader, landscape
scales (Wallace et al. 1995). Alternatively, an animal selects, at a small scale, for the highest
local nutrient concentration that may not necessarily yield the highest nutrient concentrations
at larger spatial scales. Because the animal will meet its nutrient requirements faster with high
small scale nutrient concentrations, it may prefer to spend more time on other activities at
larger spatial scales. We expect that body mass determine which foraging strategy herbivores
will follow. Along with body mass, the most widely recognized animal characteristics
influencing foraging include gastro-intestinal tract type, feeding type and mouth size (Van
Soest 1996). To our knowledge, the relationship between these foraging traits and how they
shape an animal’s perception of spatial heterogeneity have not been studied before.
Because body mass and gut size increase allometrically and metabolic requirements scale
with body mass to the power 0.75 (Kleiber 1932, Prins & van Langevelde 2008), it is
commonly accepted that where forage quantity is limiting small animals outperform larger
ones, but larger species do better on low quality feeding sites (Bell 1971, Jarman 1974).
Ritchie and Olff (1999) predicted that larger species perceive and use less spatial detail of
heterogeneously distributed resources than do smaller animals. We therefore expect that
smaller animals will select areas of high nutrient concentration at smaller spatial scales than
larger animals. We further predict that body mass will be positively related to the size of
selected patches, so that large animals will occur more frequently at larger patches than small
ones.
The influence of differences in gastro-intestinal tract type on foraging behaviour has been
well-studied (Bell 1971; Foose 1982; Demment & Van Soest 1985; Gordon & Illius 1988;
Illius & Gordon 1992 & 1994). Most of these studies suggest that forage quality is more
important than quantity to the ruminant digestive system as the throughput rate in ruminants is
limited by the digestive process. In contrast, animals with monogastric digestive systems can
increase their intake to compensate for the low nutrient content as long as the quantity of the
forage is sufficient. Because ruminants need higher quality with every bite, we expect them to
select areas of high nutrient concentration at a small spatial scale. Monogastric animals can
get more nutrients by consuming more food and we thus expect them to choose areas with
high total nutrient content at larger spatial scales even though local nutrient concentration at
these scales might be low.
According to Clutton-Brock and Harvey (1983) intake rate is controlled by bite rate,
incisor morphology and mouth dimensions. Although some studies suggest that animals with
very large mouths can sustain their nutrient requirements by taking large bites in spots with
low biomass but high quality (Owen-Smith 1988), it is generally accepted that animals with
small mouths are capable of feeding more selectively than animals with larger mouths
42
(Jarman 1974). Small-mouthed animals do not have to increase bite quality as much as large
mouthed animals, because small-mouthed animals are able to select high quality bites.
Therefore, we predict that small-mouthed animals, in order to obtain more nutrients over time,
will select areas with high nutrient content at larger spatial scales. In contrast, large-mouthed
animals with an equal body mass will be forced to select areas with high nutrient content at
small scales to increase the quality of their bites.
We also expect that the size of the herd in which the herbivores forage determine their
response to the spatial scale at which nutrients are distributed. From studies up to date, the
general understanding is that herd size tends to increase with habitat openness, although the
underlying mechanisms for this selection are still unclear (Estes 1974, Pays et al. 2007).
Beauchamp (2005) found that group foragers obtained food at a lower rate than solitary
foragers, because of the increased cost associated with searching for new food patches. Living
in a herd can also increase the costs associated with exploitative or interference competition
(Smallegange et al. 2006), but this could be compensated for by the benefits derived from
reduced individual vigilance behaviour and information sharing of high quality patches (Prins
1989). Because of interference competition and the increased resource depletion risk for
individuals within a herd, we predict that animals occurring in large herds will select areas of
high nutrient concentration at a small spatial scale, whereas solitary animals can afford to
sacrifice selecting high nutrient areas at small scales in order to obtain higher total nutrient
loads at larger scales. We further predict that herd size will be positively related to patch size
so that animals living in large herds will occur more at larger patches than small ones.
To test these predictions, we experimentally manipulated the spatial distribution of
nutrients in a large fertilization experiment (30 ha) under open field conditions. This is a rare
but powerful approach, which allows for the manipulation of one environmental variable (e.g.,
soil nutrients) at a variety of spatial scales (Ball, Danell & Sunesson 2000). We used
fertilizers to increase forage quality and quantity and measured differences in visitation
frequency and grass utilization by the various herbivore species at our study site.
Methods
In December 2004 and 2006 we fertilized thirty 50 m x 50 m plots in the Timbavati Private
Nature Reserve (TPNR), South Africa (24˚14’11,760”S; 31˚22’32,184”E) with NPK (3:2:1)
fertilizer. The study area was dominated by Colosphermum mopane shrub veld but also
included trees such as Combretum apiculatum. The herbaceous layer was dominated by
Urochloa mosambisencis and Bothriochloa spp. (for a full description of the vegetation see
Van der Waal 2008, unpublished). The rainy and warm season stretched from October to
March. Data was collected from October 2005 to January 2008. During the study period the
mean annual rainfall was 420 mm.
43
At the study site, surface water was available within half a kilometre north and south from
the closest experimental plot. The experimental layout followed a randomized block design
including three replicates of seven different treatments and nine controls. One control was
allotted to each scale treatment and replicated three times. The fertilizer was applied at each
50 m x 50 m plot across either one of three different surface areas (50 x 50 m; N = 1), (10 x
10 m ; N = 10), (2 x 2 m ; N = 25) and at either one of three different nitrogen concentrations
(30.0 g/m2, 6.0 g/m2, 1.2 g/m2). This resulted in seven combinations (Figure 4.1), as the
lightest fertilizer concentrations with a signal too faint to pickup, and heaviest fertilizer
concentrations with a risk of a toxic effect were excluded.
Figure 4.1: Experimental plot layout and NPK fertilizer treatment. The values in the white
cells represent the local nitrogen concentration, calculated from the total nitrogen per
experimental plot divided by the product of fertilized patch size and number of patches
per plot.
44
Fences between the TPNR and the Kruger National Park were removed in 1995, allowing
free movement of animals across borders. Large herbivore species commonly found in the
area include elephant (Loxodonta africana), buffalo (Syncerus caffer), impala (Aepyceros
melampus), warthog (Phacochoerus aethiopicus), duiker (Sylvicapra grimmia), zebra (Equus
burchellii) and steenbok (Raphicerus campestris), whilst predators include lion (Panthera
leo), spotted hyena, (Crocuta crocuta) and leopard (Panthera pardus). Due to their low
abundance at the experimental plots, data for kudu (Tragelaphus strepsiceros) and giraffe
(Giraffa camelopardalis) was not used for individual species analysis. Sporadic visitors such
as, waterbuck (Kobus ellipsiprymnus), wildebeest (Connochaetes taurinus) and white
rhinoceros (Cerathorium simum) were excluded from all analysis.
Spoor counts
A total of eight 1 m x 10 m spoor plots were laid out around each experimental plot (two plots
10 m apart along each of the four sides of the 50 m x 50 m experimental plot). Two replicates
of spoor counts were conducted consecutively every three months at all plots. Because spoors
of elephant and small animals like duikers were found to become unrecognizable after two
weeks, a ten-day interval was used between raking plots and counting spoor.
Dung counts
Total dung counts were conducted in each experimental plot every three months. Dung was
marked at each repetitive count with different coloured stones (2 – 4 cm in diameter) to
determine the degradation rate and spotting accuracy. The probability of dung detection was
calculated by dividing the number of previously marked dung piles re-spotted at the current
count with the number of dung piles known to be marked during the previous count. Stones
used for marking dung were picked up and dung counted as degraded if there was no dung
left at a stone dung marker. The dung degradation over time was calculated by dividing the
number of degraded dung piles by the total number of marked dung piles of the same age.
Ten sets of spoor and eight sets of dung counts were conducted between October 2005
and January 2008. For the analysis, each dung and spoor count at each plot represented the
abundance of each animal species per 50 m x 50 m plot.
Grass utilization and herbaceous above ground biomass
Herbaceous above ground biomass was measured using a calibrated disc pasture meter
(Zambatis et al. 2006). Nine parallel transects were laid out in each 50 m x 50 m plot. The
biomass, utilization and dominant species were recorded at 11 points along each transect in a
0.5 m x 0.5 m quadrate for both the wet and dry seasons of 2006 and 2007. Utilization was
indexed according to the eight-point ‘Walker’ scale, where a score of zero represented no
utilization and seven removals of 100% of the standing crop (Walker 1976).
Herd sizes for each species in the analysis were obtained from aerial game counts
conducted by the TPNR in 2003 and 2004, as well as from road censuses conducted around
45
the experimental area in 2006. Because elephant and rhinos do not have lower incisors we
used muzzle circumference as measure of mouth size. Muzzle circumferences were measured
from fresh carcasses of hunts conducted in the northern parts of South Africa between 2006
and 2008. Muzzle circumference of elephants was measured in the same way as for other
species. This measurement is probably an overestimate because elephants feed more with the
tips of their trunks, which has a smaller circumference. Body mass for each species was
calculated as the average value between adult male and female weights obtained from
literature (Skinner & Smithers 1990; Estes 1991; Nowak 1991 and Kingdon 1997). Metabolic
weight was calculated as BM0.75.
Data analysis
As spoor and dung were only measured at the entire 50 m x 50 m extent of an experimental
plot the only scales that could be analysed was nutrient concentrations at the smallest local
scale and total nutrient concentration at the 50m x 50m scale. For this reason and because
nutrient concentrations were heterogeneous within a plot when using a 10m x10m grain size,
analysis at scales with 10m x 10m grain size were not possible (Figure 4.2). Therefore, patch
size was seen as an emerging property which allowed animals to select for larger nutrient rich
areas beyond the effects of local small scale nutrient concentration. Local nutrient
concentration was calculated as the amount of nitrogen that was applied per m2 of fertilized
area. An increase in this resulted in an increase in both local forage quality and quantity.
Local nutrient concentration multiplied by the accumulated size of all sub-plots fertilized
within an experimental plot equalled the total nutrient load. Therefore, total nutrient load (kg
N) is calculated from the total amount of fertilizer that was applied per 50 m x 50 m
experimental plot. For example, in figure 1, total nutrient load and local nutrient concentration
increase vertically along the table if patch size is kept constant, whereas if total nutrient load
is kept constant, local nutrient concentration decreases as patch size increases, moving
horizontally across the table. When animal abundance increased with total nutrient load, it
was interpreted as selection of areas with high nutrient concentration at a large spatial scale
(50 m x 50 m). When animal abundance increased as local nutrient concentrations increased,
it was interpreted as selection of areas with high nutrient concentrations at a small spatial
scale.
For analyzing dung degradation and detection probability, data were tested for normality
using the Kolmogorov-Smirnov test. If data were not normally distributed after
transformation, non-parametric tests were used. Because non-parametric tests showed that
dung detection probability was significantly different between the months in which dung was
counted, we could not use survival analysis for analyzing dung degradation (Kruskal-Wallis:
chi-square=21.253, d=6, p=0.002). In stead, only dung piles with an age ≤ 1 year were used in
a univariate analysis to test for differences in the rate of dung degradation between species.
46
Figure 4.2: An illustration of the effect of spatial scale at three different grain sizes on nutrient
concentration as could be perceived by a herbivore at each of the seven plot treatments.
47
We found that differences between dung degradation rates of animal species were small,
and no significant difference could be found between species (Tukey: p=0.272, df=9).
Therefore, we did not have to correct for the number of dung piles for further interspecies
comparisons.
Our hypothesis testing was carried out in R with dung or spoor density as dependent
variables. A generalized linear model (GLM) with a negative binomial error distribution was
used, as over dispersion of the data was detected under the assumption of a Poisson
distribution (White 1996). Independent treatment factors in this GLM were local nutrient
concentration, patch size and total nutrient load. For analysis of animal characteristics, gastrointestinal tract type, body-, mouth- and herd size were added independently. All interactions
between and within independent treatment factors and independent animal characteristics
were included in the model. In the analysis, we corrected for distance from the closest water
point, grass biomass and plot distance from the mid point of the experiment. The selection
between the different candidate models was carried out by comparing their AIC values.
Results
Dung and spoor counts both increased with increasing grass utilization, showing that the
abundance of dung and spoor served as good indicators for actual vegetation consumption
(GLM dung: Z=2.094, p=0.036; GLM spoor: Z=3.93, p<0.001).
Duiker, warthog and elephant abundance increased significantly as total nutrients/ 2500
m2 plot increased (Figure 4.3, Table 4.1) indicating that these animals select areas of high
nutrient concentration at the largest spatial scales. Buffalo, zebra and impala abundance
increased significantly with local nutrient concentration and buffalo numbers further
increased with increasing patch size (Figure 4.3, Table 4.1). Steenbok did not show any
preference for any of the treatments possibly because both males and females are territorial
causing dung and spoor to occur equally throughout all plots.
48
Table 4.1
Significant results from generalized linear models on the abundance of herbivores as indicated
by dung and spoor in response to changes in local nutrient concentration, patch size and total
Gastro-intestinal
Tract Type
5.62
1
13
Browser
Ruminant
Duiker
16
8.00
1
13.8
Browser
Ruminant
Impala
49
18.52
16
17.9
Mixed
Ruminant
Warthog
68
23.68
3
46
Grazer
Monogastric
Kudu
189
50.97
4
26
Browser
Ruminant
Zebra
308
73.52
6
33.9
Grazer
Monogastric
+
Buffalo
552
113.88
30
52.2
Grazer
Ruminant
+
Giraffe
1010
179.16
3
43.5
Browser
Ruminant
Elephant
4750
572.16
5
222.7
Mixed
Monogastric
Type Feeder
10
Body Mass (kg)
Mouth Size
(circumference,
cm)
Patch Size
Steenbok
Animal Species
Total
Local Nutrient
Herd Size
Metabolic Weight
(kg)
nutrient loads
Concentration
Nutrients/
2
(m )
2
2
(g/m )
Dung
2500m plot
Spoor
Dung
Spoor
Dung
Spoor
+
+
++
_
++
+
-
+
-
++
(‘+’ indicates a positive relationship at 95% significance, ‘++’ positive at 99% and ‘-’ negative at 95%;
empty cells had either no prediction or no significant relationship; N=24)
49
Figure 4.3: Responses of elephant, buffalo, warthog and impala to different fertilizer
treatments as measured from dung and spoor counts (N=66).
50
Larger animals occurred more at plots with higher total nutrient loads and smaller animals
more at low total nutrient loads (Z=2.360, p=0.018; Table 4.2) supporting our hypothesis that
larger animals select areas with high nutrient content at larger scales. Ruminant animal
numbers decreased as total nutrient loads increased, but increased as local nutrient
concentrations increased, indicating that these animals select areas of high nutrient
concentration at the smallest spatial scales (Z=-3.920, p<0.001). Contrary to expectations
large-mouthed animals selected areas with high nutrient concentrations at larger spatial scales
than small-mouthed animals (Z=2.384, p=0.017). Herd size increased with increasing local
nutrient concentration (Z=3.134, p=0.002) and patch size (Z=2.290, p=0.022) but decreased
with increasing total nutrient loads (Z=-4.036, p<0.001).
Table 4.2
Predicted and observed significant results from generalized linear models on dung counts
showing how species characteristics such as digestive system type, metabolic weight, mouth
size and herd size determine selection of experimental plots with different patch sizes, local
nutrient concentrations and total nutrient loads
2
Patch Size (m )
Local Nutrient
Total Nutrients/
2
2
Concentration (g/m )
2500m plot
Animal
Predicted
Observed
Predicted
Observed
Predicted
Observed
+
+
Characteristic
Metabolic Weight
-
Ruminants
+
+
+
+
Mouth Size
---
-
+
-
---
+
Herd Size
+
++
+
+
(‘+’ indicates a positive relationship at 95% significance, ‘++’ positive at 99%, ‘+++’ significance
above 99% and ‘-’ negative significance; empty cells had either no prediction or no significant
relationship; N=24)
51
Discussion
We assessed how herbivores respond to changes in nutrient heterogeneity and at what spatial
scale a herbivore with a particular mouth, herd and body mass, will select areas with higher
nutrient concentrations. Our results showed that larger animals selected areas of high nutrient
concentration at larger spatial scales by selecting plots with higher total nutrient loads than
smaller animals. Body mass has been linked with the spatial scale of resources to explain the
co-existence of species by predicting that larger species perceive and use less spatial detail of
heterogeneously distributed resources (Ritchie & Olff 1999). Field experiments, however,
contradicted this hypothesis and found that smaller species, such as warthog and impala,
selected for larger patches (Cromsigt & Olff 2006). Although we could not find any
relationship of body mass with patch size for the total herbivore assemblage at our
experimental plots, buffalo and zebra selected for larger patches compared to warthog, which
is in agreement with Ritchie and Olff’s (1999) original hypothesis. It is possible that Cromsigt
and Olff (2006) could not detect selection of larger patches by buffalo, because the patch sizes
in their study only ranged from 1-64 m2 compared to 4-2500 m2 in our study. In a patch
selection experiment with steers, WallisDeVries, Laca and Demment (1999) found that in a
fine-grained environment (2 x 2 m), where patch size is equivalent to an animal’s body length,
the cost of frequently turning to find the most profitable patch becomes too high. They
concluded that large herbivores would avoid small patches and concentrate on the larger ones
where foraging costs are low, such as found for buffalo and zebra in our study.
Large-mouthed animals, as like large-bodied animals, selected areas of high nutrient
concentration at large scales. This contradicted what we predicted. Although mouth size has
been found to be highly correlated with body mass (Hanley 1982), large extremes exists in
mouth to body mass ratios (Pretorius 2008, unpublished). Arsenault and Owen-Smith (2008)
found that the scaling of mouth width relative to body mass might be the main factor
governing grass height selection, rather than body mass alone. When looking at the mouth to
body mass ratios of individual species in our experiment we found that both extremes, i.e.,
small animals with large relative mouth sizes, like duiker and warthog, and large animals with
small relative mouth sizes, like elephant, selected areas of high nutrient concentration at
larger spatial scales.
A large mouth to body mass ratio together with high metabolic requirements of a small
animal will make it very difficult to select foods at a small scale that will deliver a bite that is
rich enough in energy. The best option would thus be to select to have more bites and more
nutrients available at larger spatial scales, even though the nutrients may be more diluted,
rather than to have less bites of higher quality at the small local scale. All the larger, herd
living animals in our experiment had intermediate mouth to body mass ratios and intermediate
absolute mouth sizes. Under these conditions, selection of high nutrient areas at larger scales
52
becomes difficult. The larger number of animals in a herd forces an individual to select the
patch that satisfy its most important immediate requirements before the food gets depleted by
other herd members. Thus, every bite counts and as expected we found herd animals to select
areas of high nutrient concentration at a small local scale. On top of this scale selection, herds
further preferred larger patches. This is in agreement with a study conducted by Hester et al.
(1999) who found that solitary sheep selected smaller patches than red deer that foraged in
small groups, possibly due to the increased absolute forage requirements of a larger number
of animals.
Following the results of the individual species analysis, we developed a conceptual model
from our data. The model (Figure 4.4) illustrates how mouth to body mass ratios can divide
herbivores into two groups: (a) animals with extreme mouth to body mass ratios (either small,
such as elephant or large such as duikers) that select high nutrient areas at large scales, and (b)
herd animals with intermediate mouth to body mass ratios that select high nutrient areas at
small scales.
An animal with a small mouth to body mass ratio that do not have to increase bite quality
but has a large absolute forage intake requirement will be better of maximizing the total
amount of nutrients gained at large scales. Although animals like elephant also live in herds,
their small mouth to body ratios may override the herd effect, allowing them to select areas of
high nutrient concentration at larger scales. Stokke (1999) found elephant family units to
discriminate between patches in their surroundings, and to select patches offering the highest
density of palatable species. Small animals, like duiker, on the other hand will not be able to
herd together in great numbers, as a small absolute mouth size with increased absolute forage
requirements for the herd, will force individuals to select high nutrient areas at a small local
scale. This will mean that individuals will not be able to meet their relatively high metabolic
requirements with a large mouth to body mass ratio. In a snow melting experiment using 2 x 2
m plots, Svalbard reindeer (Rangifer tarandus platyrhynchus) was found to select plots of
highest plant biomass rather than of highest plant quality, and thereby maximized both the
total amount of nitrogen and energy uptake simultaneously (Van der Wal et al. 2000). These
results are consistent with our findings of small animals with large mouth to body mass ratios
selecting combinations of patch sizes and local nutrient concentrations that yielded the highest
total nutrients at larger scales. The Svalbard reindeer has an average group size of 2.2
(Alendal, de Bie & van Wieren 1979), weighs 5.8 kg with an incisor width of 2.41 cm
(Mathiesen et al. 2000) which, compared to the duiker weighing 16 kg (Skinner & Smithers
1990) with an incisor width of 2 cm, yields an even larger mouth to body mass ratio. Soay
sheep (Ovis aries) have been found to be more efficient at exploiting clustered patches with a
high energy reward than a matrix of homogeneously distributed patches of lower local energy
reward, consistent with our findings for impala (Pérez-Barbería, Walker & Marion 2007).
53
Soay sheep weighs 23 kg on average and has a muzzle width of 6 cm similar to the 5.7
cm mouth width of impala (Pérez-Barbería et al. 2007). Although the body mass of Soay
sheep is closer to a 16 kg duiker than to a 49 kg impala, sheep, like impala tend to herd
together in larger numbers, which increases the total metabolic weight and absolute nutrient
requirements of the herd.
Figure 4.4: A conceptual model demonstrating how the relationship between mouth to body
mass ratio and total metabolic weight of the herd determines the scale at which nutrient
rich areas is selected. Circles indicate small scale nutrient maximizers and diamonds,
large scale nutrient maximizers. Black dots represent animals with significant results in
our analysis; grey, individuals with small sample sizes and white, rare visitors to the
experiment (1=steenbok, 2=duiker, 3=warthog, 4=impala, 5=kudu, 6=wildebeest,
7=waterbuck, 8=zebra, 9=giraffe, 10=white rhinoceros, 11=elephant, 12=buffalo).
54
Between white rhinoceros, zebra, warthog and impala, Cromsigt and Olff (2006) found
only warthog to respond significantly to both fertilizer and mowing induced increases in
nutrient availability and patch size. This combination of mowing, fertilizing and increasing
patch size yielded the maximum total nutrient load per plot and supports our finding for
warthog selecting areas with the highest nutrient concentration at larger scales, whereas zebra
and impala do so at smaller scales. Warthog has an extreme mouth to body mass ratio
compared to herd species like impala and zebra, which have medium mouth to body mass.
Although white rhinoceros only visited our plots sporadically and could not be included in the
analysis, we predict that with a body mass of 1600 kg and a muzzle circumference of 96 cm, a
white rhinoceros will maximize nutrient intake at the same scale as a warthog (Figure 4).
In accordance with our hypothesis, ruminant animals selected nutrient rich areas at small
scales, although non-ruminants varied in their response. When looking at our results of
individual species however, no clear pattern emerges for predicting the scale at which animals
with different digestive systems select areas of high nutrient concentration. Duiker, warthog
and elephant, which all selected areas at larger scales, have different digestive systems and
feed on different food types. What these animals have in common is their extreme mouth to
body mass ratios. Small-scale selectors like buffalo, impala and zebra on the other hand also
differ in their digestive systems and type of food they eat, but tend to herd together and have
medium mouth to body mass ratios.
Evidence on the spatial scales at which nutrient intake is being maximized for different
animal species is rare. We suggest that, at the scale we measured, large mammalian
herbivores select areas with the highest nutrient concentrations at either the local scale or at
the larger 2500 m2 plot scale depending on a combination of their body, mouth and herd sizes.
Our results provide valuable foraging selection criteria for large herbivores, which can be
incorporated into models that predict species distribution at larger scales.
55
56
CHAPTER 5
DIET SELECTION OF AFRICAN ELEPHANT OVER
TIME SHOWS CHANGING OPTIMIZATION
CURRENCY
Y. Pretorius, J.D. Stigter, W. F. de Boer, S.E. van Wieren, C.B. de Jong,
H.J. de Knegt, C.C. Grant, I. Heitkönig, N. Knox, E. Kohi, E. Mwakiwa,
M.J.S. Peel, A.K. Skidmore, R. Slotow, C. van der Waal, F. van
Langevelde, H.H.T. Prins.
57
Abstract
Multiple factors determine diet selection of herbivores. However, in many diet studies
selection of single nutrients is studied or optimization models are developed using only one
currency. In this paper, we use linear programming to explain diet selection by African
elephant based on plant availability and chemical composition over time. Our results indicate
that elephant at our study area maximized intake of phosphorus throughout the year, possibly
in response to the deficiency of this nutrient in the region. After adjusting the model to
incorporate the effects of this deficiency, elephant were found to maximize nitrogen intake
during the wet season and energy during the dry season. We reason that the increased energy
requirements during the dry season can be explained by the changes in water availability and
forage abundance in the environment. As forage abundance decrease into the dry season,
elephant struggle to satisfy their large absolute food requirements. Adding to this restriction is
the simultaneous decrease in plant and surface water availability, which force the elephant to
seek out scarce surface water sources at high energy costs. During the wet season when food
becomes more abundant and energy requirements are satisfied easier, elephant aim to
maximize nitrogen intake for growth and reproduction.
Keywords: linear programming, elephant diets, nitrogen maximization, phosphorous
deficiency
58
Introduction
How abundant does a plant species need to be and what nutrient content does it have to have
in order to be included in a free ranging herbivore’s diet? Theoretically, this depends on the
physiological and morphological forage adaptations and nutrient requirements of the animal,
which in turn depend on the animal’s body mass, life history stage and current nutritional
status (Hanley 1997, Shipley et al.1999, Brown et al. 2004). However, due to the difficulty in
determining these parameters, the theoretical considerations are hardly supported by field data.
Scientific knowledge on nutrition of large herbivores mostly comes from studies on livestock
under controlled conditions (Robbins 1993). Whereas the animal’s ideal diet is in the hands of
the nutritionist in agriculture, what wild herbivores eat is a result of natural selection and
adaption to the environment. The challenge is not only to determine the nutrient requirements
of wild animals but also to understand how animals compile their own ideal diets.
From studies in biochemistry and physiology, we know that some macro minerals such as,
sodium (Na), phosphorous (P), nitrogen (N), potassium (K), magnesium (Mg), calcium (Ca)
and carbon (C) are essential to animals (McDonald et al. 1995). Phosphorous, Mg and Ca can
readily be stored in bone whereas K and Na are not stored to a great extent in animal tissue.
Although K is generally readily available from plants (Robbins 1993), Na, like P and N, often
occurs at low levels in plants in southern Africa (du Toit et al. 1940, Weir 1969, Snyman
2002). Carbon based energy rich compounds can be stored as fat, but energy constantly needs
to be replenished to sustain metabolic demands (Brown et al. 2004). Nitrogen forms the
building blocks of muscle, it has a high turn-over rate leading to endogenous losses that need
to be replenished on a daily basis and it can also serve as a less efficient source of energy
(McDonald et al. 1995). Adding to the complexity of the multiple nutrients that need to be
selected for in the wild is the presence of chemical deterrents such as condensed tannins in
plants that negatively affect the ability of an animal to digest nutrients (Robbins et al. 1987).
In this paper, we aim to provide a mechanistic framework for plant selection by a wild
animal based on plant availability, nutrient and deterrent content. Nutrients are assumed to be
selected for based on their availability in the environment in relation to the animal’s nutrient
requirements (Hengeveld et al. 2009). To emphasize the trade-offs between plant quality and
quantity we chose to study the largest terrestrial mammalian herbivore, the African elephant
(Loxodonta africana). Some experimental studies on elephant nutrition have been conducted
in zoos (Clauss et al. 2003a & 2005). However, most conventional studies in the wild are
descriptive and case sensitive as nutrient selection is determined using consumption rates of
plant species high in a specific nutrient in relation to its availability in the environment. From
these studies elephant have been shown to select for calcium (Greyling 2004, Stokke 1999),
sodium and nitrogen (Jachmann & Bell 1985). Because no clear evidence exists for the
59
selection of only one currency by elephant, we will test several postulated hypotheses to see
which nutrient selection strategies best fit the observed diets.
Due to the enormous size of an elephant, its absolute forage requirements are very high.
Since the Pleistocene, the morphological and behavioural forage adaptations of the elephant’s
ancestors have closely followed changes in the abundance of forage on earth (Hofmann &
Stewart 1972, Ehleringer et al. 1997, Cerling 1998). As elephant are sometimes classified as
non-selective bulk feeders (van Soest 1981) and have the largest daily intake of all terrestrial
herbivores, the best null model would be to expect that modern day elephant randomly
include plant species in their diet, i.e., according to availability. Alternatively, we can argue
that an elephants’ high absolute energy requirements will hamper the animal from not only
consuming too many rare plants that will cost more energy to find, but will also force the
animal to select plants of higher quality and digestibility so that energy intake can be
maximized. If this is the case, we can expect elephant to maximize energy intake by including
the most abundant plant species with the highest metabolizable energy value in their diet.
Besides energy maximization being used as the ultimate model for explaining plant
selection, there is growing support for models satisfying requirements of multiple nutrients
(Yearsley et al. 2001). Raubenheimer and Simpson (1993, 1997 & 1999) found that insects
could adjust the amounts of food ingested from different food sources in order to keep the
balance between different nutrients, and consistently reach the same target, i.e. their daily
nutrient requirements. In the wild, animals also need to aim to reach targets of nutrient
requirements for maintenance, growth, mobility and reproduction. As an alternative foraging
strategy, we expect elephant to select the combination of plants in the diet that will yield
nutrients that is as close to the minimum requirement for each macronutrient as possible.
Recently, Ludwig et al. (2008) and Treydte et al. (2009) demonstrated, using linear
programming models, how grazing ungulates can meet their daily nutrient requirements by
consuming the right combinations of grass from under tree canopies and in open grasslands
before reaching their maximum intake as determined by grass fibre concentrations. Even
though these studies took a multi-nutrient objective approach to understand how an animal
satisfies its requirements, they did not consider variation in availability of different plant
species. Although linear programming models have been used in foraging ecology for a long
time (Westoby 1974, Belovsky 1978), they stand under criticism because of the sensitivity of
the model’s parameters (Huggard 1994; Yearsley et al. 2001). In calculating diets for the
current study, we not only specified dietary constraints but also included goal functions to
define foraging strategies. The foraging strategies used included before mentioned predictions
as well as alternative strategies such as maximization of N, P, Ca and Na intake, and
minimization of condensed tannin intake. The deviation between the nutrient content of
60
observed and predicted diets, calculated from linear programming models, is used as a
relative measurement to test the hypotheses on diet choices.
Methods
The study was conducted in the northern parts of the Timbavati- and the Umbabat Private
Nature Reserves, South Africa (24˚14’11,760”S; 31˚22’32,184”E) between December 2005
and 2007. A homogeneous layer of Colophospermum mopane shrubveld on granite soils
dominates the area. The rainfall season normally extends from October to March and during
the wet season of 2005/2006 it rained 427 mm and of 2006/2007 390 mm.
At the beginning of the study, thirteen plant species were selected and their leaf biomass
and chemical composition measured over a period of two years. These species represented
forage available for elephant to choose a daily diet from and their selection was based on their
abundance within the study area and on preliminary forage observations on herbivores in the
field and from literature (Stokke 1999; Greyling 2004). Tree species selected for
measurement consisted of the six most common and frequently utilized browse species
including, Colophospermum mopane, Grewia monticola, Acacia nigrescens, Combretum
apiculatum, Lannea schweinfurthii, and Dichrostachys cinerea and the rare but well utilized
browse species Maerua parvifolia. Herbaceous species selected for measurement consisted of
the three most common and frequently utilized grass species Panicum maximum, Digitaria
eriantha and Urochloa mosambicensis, the rare but well utilized annual grass species
Brachiaria deflexa, a common herb (Indigofera spp.) and a rare and frequently utilized herb
(Cyperaceae spp.). The study was divided into five periods at three-month intervals stretching
over a wet-, late wet/early dry- and dry season and two late dry/early wet seasons. The total
nutrient content of observed and predicted diets, compiled from the available thirteen plant
species, were calculated for each period using the following variables:
Plant structural characteristics
Nine monitoring sites were randomly selected throughout the study area in vegetation
typically representative of the dominant Colophospermum mopane shrubveld. At each site,
the area used for sampling was about 10 ha in size. Leaf cover and biomass measurements of
the thirteen selected plant species were conducted every month at all nine sites and the data
pooled at three-month intervals. For trees, intra-plant species distance, height, canopy height,
maximum canopy diameter and perpendicular canopy diameter of each species were
determined at the beginning of the study using the Point Centre Quadrate (PCQ) method
(Mueller-Dombois & Ellenberg 1974). This entailed recording the distance to the closest
species and to the selected tree species in each of the four quadrates at 20m intervals along a
100m transect at each of the nine study sites. For each of the seven selected tree species at
these PCQ transects, we calculated available browse volume per tree according to the
61
Biomass Estimates from Canopy VOLume (BECVOL) method (Smit 1996). To determine the
intra-plant species distance and ground biomass of herbaceous species the Dry-weight-rank
(Dekker et al. 2001) and Comparative Yield methods (Haydack & Shaw 1975) were used and
repeated at monthly intervals along the transects at each of the nine study sites.
Along with herbaceous biomass estimations in 0.5 m x 0.5 m x 0.5 m metal frames at
each site each month, a representative sample of each selected browse, herb and grass species
was cut in the same size frame resulting in nine frame cut samples per species per month.
These samples were dried at 70 °C for 24h, the leaf material separated from the stem and
weighed. For each grass and herb species, the average leaf mass per frame for each period
was weighted. For each browse species, the average leaf mass determined for each species per
frame was used along with the available browse volume estimates from BECVOL to calculate
the total available leaf mass per tree per period.
Leaf nutrient and deterrent content of plant species
The dried leaf samples collected each month per frame for each of the thirteen plant species
were pooled per species at a three monthly interval and milled through a 1 mm sieve. Quality
of the leaf samples were determined at the laboratory of the Resource Ecology Group
(Wageningen University, the Netherlands) through analysis of the macronutrients N, P, Ca,
Na, K, Mg and deterrents including acid detergent fibre (ADF), acid detergent lignin (ADL)
and condensed tannins. Dry matter content was determined by drying the samples for 8h at
70˚C. Total N, Ca, Mg, Na, K and P content were measured with a Skalar San-plus auto
analyzer, after destruction with a mixture of H2SO4, Selenium and salicylic acid
(Novozamsky et al. 1983). Condensed tannins content were measured using the
Proanthocyanidin method (Waterman & Mole 1994) and ADF content by using an ANKOM
Fibre Analyzer followed by treatment with 72 % H2SO4 and ignition at 525˚C for 3h to
determine ADL.
Metabolizable energy content of forage
Intake of elephant on a dry matter basis has been estimated by Owen-Smith (1988) to be
between 1 and 2 % of body weight. We assumed 1.5 % intake of body mass for a 3500 kg
elephant, which equates to 53 kg dry matter per day. Metabolizable energy values of plant
species were calculated from gross energy values from literature, taking into account forage
digestibility and estimated methane and urine energy loss. Digestibility of food measured in
zoos range from 22 % to 36 % (Rees 1982, Pendlebury 2005) and in the wild from 30 % to 45
% (Owen-Smith 1988, Meissner et al. 1990). To determine digestible energy values of
individual plant species from their ADF content, a regression equation was calculated for
digestibility of ADF from the ADF content of food (Digested ADF = 6.665e0.0246(ADF diet),
R2=0.87, N=8) and of digestible energy from digested ADF (Digested energy =
64.85*Digested ADF-0.205, R2=0.6, N=8) using data from feeding trials of elephant in zoos
62
conducted by Benedict (1936), Ulrey et al. (1979), Tomat (1999), Foose (1982), Rees (1982)
and Clauss (2003a).
Gross energy values for plants do not vary much and generally range between 16.8 kJ/g
and 21 kJ/g (Robbins 1993). Values for Panicum and Brachiaria species, for example, have
been estimated at 16.8 kJ/g in (Tiemann 2008, Adu & Adamu 1982) and for Colosphermum
mopane between 18.5 kJ/g to 20.4 kJ/g (Styles & Skinner 1997, 2000). We used an 18 kJ/g
gross energy content for grass species and 20 kJ/g for browse and herb species. Dry matter
digestibility was calculated from digested energy values as explained above (Digested energy
= 0.984*DM digestibility + 1.991, R2=0.92, N=8). These dry matter digestibility values for
each plant species were then compared with values presented by Meissner et al. (1990) to
estimate urine and methane energy loss. Metabolizable energy content of each plant species
(ME) was calculated as follows:
ME
(MJ/kg)
= Gross Energy x Digestible Energy – Methane & Urinary Energy Loss
(MJ/kg)
(%)
(MJ/kg)
Nutrient requirement estimation
Most research on animal’s nutritional requirements has been undertaken using animals in the
agriculture industry such as cattle and pigs. Consequently, extrapolated values from feeding
trails on these livestock species under controlled environments were applied to wild species.
The problem with this method is that little is known about the scaling of requirements for
different nutrients in relation to body mass. For example, a linear extrapolation of
requirements for an element like copper, which is the technique most commonly used, from a
horse to an elephant, have led to fatalities (Rucker 2007). We used a combination of
extrapolations from values stipulated by the Agriculture Research Council (ARC 1989) for
horses, zoo studies and studies in the wild on elephant to compile a set of minimum and
average daily nutrient requirements for an adult elephant (Table 5.1).
It is generally accepted that energy scales with a factor of 0.75 to body mass (Kleiber
1932, Prins & van Langevelde). Daily energy requirements were calculated from equations
obtained from Taylor et al. (1982), Robbins (1993) and Prins & van Langevelde (2008) so
that daily basal metabolic energy requirements (Ebmr) equals:
Ebmr (kJ /day) = 293 x BW0.75
For a 3500 kg elephant this would be 133327 kJ/day. Energy expenditure for foraging (Ef) is:
Ef = 0.54 x Ebmr,
which for an elephant equals 72000 kJ. Energy for standing (Es) is:
Es = 0.2 x Ebmr,
which for an elephant equals 26665 kJ. Thus requirements for basal metabolic rate, standing
and foraging for a 3500 kg elephant equal 232 MJ/day. This amount excludes energy required
63
for walking to find food, which will vary depending on the availability of food. The energy
expenditure per km of walking per unit mass (Ew) is:
Ew =10.75 x BW-0.316
The energy expenditure per day (Ebw) including moving over a certain distance (D) and
feeding is:
Ebw = Ebmr + (Ew x D x BW) + Ef + Es
Where D is the distance travelled in km (adjusted from Prins & van Langevelde 2007). The
average distance walked per day over a two-year period for 30 different collared elephants at
our study area was 8.3 km. Therefore, the energy expenditure for an elephant of 3500 kg per
day for walking, standing and feeding is estimated at 258 MJ. However, this value excludes
energy required for digestion and assumes that the elephant is moving across a flat landscape.
Alternatively Robbins (1993) suggests that free-ranging terrestrial eutherians that are not
actively breeding have daily energy expenditures of 2.3 times basal metabolic rate, which
amounts to 307 MJ for the elephant. From field studies on elephant, Meissner (1982)
calculated the metabolizable energy requirements of a fifty year old bull of 3700 kg to be 310
MJ, close to our calculated value, while Owen-Smith (1988) estimated energy requirements of
a female elephant at 259 MJ. As energy requirements estimated for elephant vary in literature,
we replicated calculation of predicted diets three times using the following energy values: 258
MJ, 280 MJ and 307 MJ
Plant species composition of observed diets
Leaves from the thirteen measured plant species were sampled for an epidermis reference
collection. Plant parts were cleared in household bleach for 24 h and washed out in water.
Leaf epidermis fragments were stripped off and mounted in glycerol on slides. Photographs of
these slides were used to identify the epidermis and cuticle fragments present in the faeces
(De Jong et al. 2004).
Every month ten fresh (less than 12 h old) adult elephant faecal samples were collected
throughout the study area. Samples were pooled together at three month intervals. Each
pooled sample was analysed individually. Samples were heated in a pressure cooker from
115-125°C in water for at least 2 h and left to soak overnight. In order to separate inner tissue
from epidermis and cuticle, a 5 g subsample was washed in a blender with tap water, strained
first over a 1.5 mm sieve to eliminate course fibre, then over a plankton sieve (0.01 mm), and
stored in 70 % ethanol (De Jong et al. 2004). The subsample was transferred into a Petri dish
and allowed to settle. Ten random grab samples of the residue were taken with a Pasteur
pipette. Each droplet was put on a glass slide, spread out evenly and covered with a cover slip.
On each slide, ten epidermis fragments were identified using the reference material and their
surface areas measured through a grid of 0.01 mm2 squares in the microscopic eyepiece (De
Jong et al. 2004).
64
The results of the ten grab samples were pooled for every faecal sample. Plant epidermal
fragments that could not be identified down to species, genus or family level were categorized
as either ‘monocotyledon – unidentified’ or ‘dicotyledon – unidentified’. The abundance of
each forage type and species was represented as a percentage of the total measured fragment
area (De Jong et al. 2004). The proportion of each plant species in the observed elephant diets
was assumed to be equal to the proportion of fragments of each plant species in the faeces
(Stewart 1967). Because >50 % of fragments consisted of the thirteen plant species selected,
the percentage of fragments for each of these selected species were extrapolated to calculate a
total observed daily diet of 53 kg for each period.
Diet calculations
For comparison with the observed diets, seven possible foraging strategies were used to
calculate predicted diets in MATLAB (Version 6, The MathWorks, Inc.) using linear
programming. These strategies included random foraging, energy maximization, nutrient
satisficing, nitrogen maximization, phosphorous maximization, calcium maximization,
sodium maximization and condensed tannin minimization. The proportions of each plant
species that needed to be included in the random diet were equal to the relative abundance of
each plant species in the study area. The maximum energy diet was calculated as a 100 %
consumption of the most abundant plant species in the environment with the highest
metabolizable energy value, which would therefore require the least amount of energy to find.
Each of the remaining strategies was defined as the goal function in the linear models and the
diets calculated under the constraints that minimum energy, or P, K, Mg, N, Ca and Na
requirements, or only energy requirements had to be satisfied.
After preliminary tests of the model we realised that predicted diets for the wet season
deviated greatly from the observed diets for most nutrients, possibly because animals
replenish stores of many of the nutrients during these times of abundance. Because the model
calculated only one optimal solution, small changes in some of the specified constraints
resulted in completely different diets. We used the deviation of species composition of
predicted diets from observed diets to indicate which nutrient constraints were most sensitive.
For example, if Colophospermum mopane had high levels of Ca but low levels of P and Ca
levels were lower in most of the other species, a small increase in the constraints specified for
P would not only increase the chances of mopane being excluded from a diet but also
decrease the total Ca gained in a diet. Through sensitivity analysis we determined changes in
energy constraints to be the most sensitive. As the exact nutrient requirements of an adult
elephant were unknown and to avoid the pitfall of generating constraints from the data, all
seven theoretical diets, excluding the random diet, were calculated six times using three
different energy values (257 MJ, 280 MJ, 307 MJ) and one of two sets of dietary requirements
that were in agreement with literature (Table 5.1: minimum and average). This resulted in 43
65
predicted diets in total for each of the observed diets in each of the five periods. Because
elephant are known to satisfice their Na requirements from other sources such as soil and
water (Holdø et al. 2002), constraints for Na were set to zero.
The linear programming technique only allowed one goal function to be specified at any
run. As phosphorous was deficient in the area, requirements for P, as specified in the
constraints, could not be reached for most of the diets. Therefore, we assumed maximization
of P first by setting the constraints to the maximum value of P that could be reached in the
study area for a specific period. This value was determined by removing all other nutrient
constraints, except minimum energy requirements, and then calculating a diet that maximized
P intake. For each diet the nutrient content for each food type (plant species) was multiplied
by the proportion of the food type in the diet and summed over all species to yield the total
content for each nutrient in the whole diet.
The energy gain of each diet was calculated by subtracting the energy spent walking to
find each of the plant species in the diet with the metabolizable energy content of each species
at the proportions that they occurred in the diet. The energy expenditure for each km of
walking per unit mass (Ew) was calculated as:
Ew = 10.75 x BW-0.316 x Distance to find plant species (km) x BW (adapted from Taylor
et al. 1982).
Data analysis
Plant species and nutrient composition of each predicted diet were viewed as co-ordinates in a
species or nutrient space respectively. Therefore, the differences between each predicted diet
and the observed diets were calculated as absolute distances, as follows:
Diet Xi where i = 1,2,3,…37
Nutrients Nj where j = 1,2,3,…7
 N 
 P 


 Ca 


N j =  Na 
 K 


 Mg 
 Energy


But
 0.68 
 0.05 


0.132


N j Requirement = 0.005
 0.16 


 0.05 
 257 


for a 53 kg DM diet
[ ]
Xi = N j
7
Dev( X i ) = ∑
j =1
 N j Pr edicted 2 − N j Observed 2 


2


N j Re quirement


66
The predicted diet with the smallest nutrient deviation from the observed diet was
considered to be the strategy followed by the elephant. The same approach was used to
determine deviation between species composition of predicted and observed diets by
replacing nutrients in the above equation with plant species.
Diet Xi where i = 1,2,3,…37
Plant species Sj where j = 1,2,3,…7
Colophsope rmum 
 Combretum 



 Grewia
Acacia




Lannea


 Dichrostac hys 


Maerua
Sj = 


 Urochloa

 Panicum



 Digitaria

 Brachiaria



 Indigofera
 Cyperaceae 


[ ]
But X i = S j
13
Dev( X i ) = ∑
j =1
(S
j
Pr edicted 2 − S j Observed 2
)
Results
Predicted diets for the elephant show that at the beginning of the wet season, a diet aimed at
maximizing energy intake would force an elephant to only feed on Combretum apiculatum,
thereby leading to a deficiency in Na. However, a diet minimizing condensed tannin would
mainly consist of grass species like Urochloa mosambiscensis, resulting in a very high Na
content but Ca levels below the average requirements (Figure 5.1a). For this period, the
predicted diet with the smallest deviation in nutrient composition from the observed diet had a
composition where maximum P could be gained under a 280 MJ energy constraint, N was
maximized, and average requirements for Mg, K, Na and Ca were satisfied. Towards the end
of the wet season the best fit with the observed diet was the same as for the beginning of the
wet season but at a lower energy constraint of 257 MJ.
67
During the dry season, most of the diets that best fitted the observed diet included more
than 75 % Colophospermum mopane. The diet with the smallest deviation included 100 %
Colophospermum mopane and maximized energy intake. However, the N content of this diet
was below the minimum requirement. The second closest diet maximized energy and P under
a 307 MJ energy constraint whilst satisfying minimum requirements for N, Ca, Na, Mg and K
(Figure 5.1b).
The strategy followed by elephant at the switch between the dry and wet season appeared
to be affected by the quantity of the first rains. Between September and October 2006, it had
only rained 50 mm whereas 120 mm of rain was recorded for the same period in 2007. For the
drier period during 2006, elephant appeared to maximize P intake at high energy constraints
(307 MJ) whilst satisficing minimum requirements of N, Mg, Na, K and Ca. However, in
2007 during the same period elephant followed a similar strategy as they did for the early wet
season by maximizing N intake at energy constraints of 280 MJ.
For each period, predicted diets with the smallest deviation from observed diets in terms of
nutrient composition also had the smallest deviation in species composition (Figure 5.2).
Discussion
In this study, we aim to understand the strategy by which elephant compile their daily diets.
Our approach is novel in that we considered both the satisficing of multiple nutrients as well
as the possibility of maximization of particular nutrients and avoidance of deterrents. We
found that elephant at our study area did not feed at random or exclusively maximized intake
of a single nutrient. Instead, elephant used a combination of satisfying requirements of some
nutrients whilst maximizing intake of others depending on the availability of resources in the
environment. As nutrient availability changed over the seasons, the foraging strategy of the
elephant also changed to maintain their metabolic requirements.
Metabolic rate in heterotrophs equals the rate of respiration, which is the main process
that sources energy (Brown et al. 2004). Therefore, when determining the value of a plant
species to an animal, not only does the energy content of the plant need to be considered but
also the increased respiration and energy cost to find it. This becomes especially important for
a large animal with high absolute energy demands, like the elephant. During the dry season,
both surface water availability and plant water content decrease. Buffalo have been found to
obtain all their water requirements from the plants they eat in the wet season whereas they
become more dependent on surface water in the dry season as the water content in plants
decrease (Prins 1987). Therefore, during the dry season, the elephant does not only need to
reduce searching costs for food so that its net energy gain is still positive, but also to find
water. This was clearly illustrated at our study site when elephant increased their energy
68
intake in the dry season by consuming more of the most abundant food species
Colophospermum mopane.
Using surface water demands as an explanation for higher energy demands in the dry
season can further be supported by the sodium content of the diet during this time. Elephant
have been known to select soils and water sources with high sodium content (Weir 1969,
Holdo 2002). As the elephant becomes more dependent on surface water sources to fulfil its
water requirements its sodium requirements can simultaneously be satisfied in the same way.
In Kruger National Park sodium levels in water averages 280 ppm (Leyland & Witthüser
2008). If an elephant consumes 160 litres of water a day (Guy 1975) it will gain 4.5 % Na,
which is much higher than the estimated average requirement of 0.2%. Therefore, the sodium
obtained through foraging does not have to be high, which is reflected in our results of the dry
season diet.
In the wet season, as surface water availability and plant water content and abundance
increase, energy restrictions and demands of elephant also decrease. During this period,
nutrient reserves are replenished causing requirements for all nutrients to be higher (Parker et
al. 2009). Throughout the wet season elephant maximized their nitrogen intake. Nitrogen is
the building blocks for structural proteins in an animal’s body whereas plants consist more of
carbohydrates (Mattson 1980). For this reason, herbivores need to spend a great deal of time
eating and processing food to correct for the low ratio of nitrogen to fibre and carbohydrate in
their food (White 1978, Mattson 1980).
Because utilization of the microbes in the herbivores gut requires digestion by acidic
juices in the stomach, pregastric fermenters, like ruminants, can utilize this source of N
efficiently, whereas post-gastric fermenters tend to lose this microbial resource in the faeces
(Foose 1982). Through digestive processes, great amounts of protein are lost because of
erosion of cells from the gut and residues of digestive juices. The amount of endogenous
losses will be proportional to the amount of plant material passed through the gut (Foose
1982). Illius and Gordon (1992) quantified the influence of body weight and gut capacity on
time available for digestion of food for hindgut fermenters as: Mean Retention Time (MRT) =
9.4 Body Weight0.255. According to this the MRT of a 3 ton elephant should be around 72 h
although in reality it varies between 12-50 h (Owen-Smith 1988). Because non-ruminants
cannot make use of microbial protein and have high food passage rates, they can be expected
to incur higher losses of protein in the faeces (Foose 1982). This might explain why nitrogen
intake of elephant was maximized during the wet season at our study site. The metabolic rate
of an animal is determined by the accumulation rate of nutrients required for the biochemical
reactions that drives metabolism (Brown et al. 2004).
69
(a)
(b)
70
Figure 5.1: (a) Plant species (top graph) and nutrient composition, expressed as part of a 53
kg daily dry matter intake, for observed (left most column) and each of nine predicted
diets of elephant during the wet season: random food selection, energy maximization,
energy maximization with minimum nutrient requirement constraints, satisficing of
minimum nutrient requirements under energy constraints of at least 257 MJ, minimization
of condensed tannin with constraints satisfying minimum nutrient requirements and at
least 257 MJ energy, maximization of N with constraints satisfying average nutrient
requirements and at least 280 MJ energy, maximization of Ca with constraints satisfying
average nutrient requirements and at least 280 MJ energy, maximization of Na with
constraints satisfying average nutrient requirements and at least 280 MJ energy, and
maximization of P with constraints satisfying average nutrient requirements and at least
280 MJ energy.
(b) Plant species (top graph) and nutrient composition, expressed as part of a 53 kg daily
dry matter intake, for observed and each of nine predicted diets of elephant during the dry
season: random, energy maximization, energy maximization with minimum nutrient
requirement constraints, satisficing of minimum nutrient requirements under energy
constraints of at least 257 MJ, minimization of condensed tannin with constraints
satisfying minimum nutrient requirements and at least 257 MJ energy, maximization of P
with constraints satisfying minimum nutrient requirements and at least 307 MJ energy,
maximization of N with constraints satisfying minimum nutrient requirements and at least
307 MJ energy, maximization of Ca with constraints satisfying minimum nutrient
requirements and at least 307 MJ energy and maximization of Na with constraints
satisfying minimum nutrient requirements and 307 MJ energy. Lower and upper dotted
lines represent minimum and average nutrient requirements for elephant respectively, as
derived from literature (see Table 5.1).
71
Figure 5.2: Deviation in nutrient (top graph) and species (lower graph) composition between
observed and predicted diets (random, energy maximization, energy maximization with
minimum nutrient requirement constraints, satisficing of minimum nutrient requirements at
energy constraints of at least 257 MJ, minimization of condensed tannin with constraints
satisfying minimum nutrient requirements and at least 257 MJ energy, maximization of N
with constraints satisfying minimum nutrient requirements and at least 257 MJ energy,
maximization of Ca with constraints satisfying minimum nutrient requirements and at least
257 MJ energy, maximization of Na with constraints satisfying minimum nutrient
requirements and at least 257 MJ energy, maximization of P with constraints satisfying
minimum nutrient requirements and at least 257 MJ energy, maximization of N with
constraints satisfying minimum nutrient requirements and at least 307 MJ energy,
maximization of P with constraints satisfying minimum nutrient requirements and at least
307 MJ energy, maximization of N with constraints satisfying average nutrient
requirements and at least 257 MJ energy, maximization of N with constraints satisfying
average nutrient requirements and at least 280 MJ energy, maximization of P with
constraints satisfying average nutrient requirements and at least 280 MJ energy. Arrows
indicate predicted diet with the smallest deviation from the observed diets.
72
These nutrients are required in fixed amounts and ratios, and according to Elser (2000) the
growth of herbivore populations can be limited directly by minerals such as P, Ca and N. N
and P have also been shown to co-limit vegetation productivity on nutrient poor bushland
sites (Augustine et al. 2003). Therefore, a deficiency in one of the essential nutrients will
necessitate selection for that nutrient. When investigating nutrient selection of herbivores in a
particular area, the nutrient contents of plants in the area are seldomly placed in context of
nutrient availability at larger scales. Phosphorous can be regarded as deficient in our study
area as can be seen from comparisons of P levels of plant species at our study site with
minimum P requirements of large herbivores and similar plant species elsewhere in Africa
(Figure 5.3). Responding to this deficiency, the elephant at our study area maximized
phosphorous intake for all observed diets throughout the year. We suggest that a general
deficiency in a particular nutrient, like P, can be used as a predictor of what plant species
elephant are most likely to consume.
Condensed tannin content of the elephant’s diet in our study area were always much
higher than predicted from the theoretical diet where minimization of tannin content was used
as a foraging strategy. Tannin binds to protein and renders it unavailable for digestion. Thus,
in a pregastric fermenter the N from protein cannot be incorporated into microbial protein,
which negatively effects digestion. However, in monogastric animals, food first enters the
stomach where extreme acidic conditions cause for the tannin-protein complexes to dissociate
and the protein to become available for digestion (McArthur et al. 1991). Therefore,
monogastric animals could be less sensitive to high tannins levels in their diet, and as such
explain the large deviation between predicted diets with low tannin contents and the observed
diets with unexpected high tannin concentrations.
73
(a)
(b)
Figure 5.3: A comparison of (a) phosphorous and (b) nitrogen content of plant species in the
current study with the same species in other parts of Africa (from the database of the
Resource Ecology Group, Wageningen University, the Netherlands). Dotted lines
represent minimum and average nutrient requirements for elephant as derived from
literature.
74
During the last decade there has been growing support in nutritional ecology to not only
consider single nutrients or deterrents as the determining forces behind forage selection but to
look at the satisficing of multiple nutrients (Raubenheimer & Simpson 1999, Ludwig et al.
2008, Prins & van Langevelde 2008, Raubenheimer et al. 2009). Developments using a
geometric framework by Simpson & Raubenheimer (1999) have proved to be very useful in
understanding foraging of insects and more recently primates (Felton et al. 2009). However,
the limitation of this approach is that a foraging strategy aimed towards reaching a target
intake is assumed, limiting the ability to explore selection of specific nutrients beyond this
target when restrictions on daily food intake is also taken into consideration. Linear
programming resembles the geometric framework approach in that it too considers selection
of food types based on the content of multiple nutrients. However, the big advantage of linear
programming is that on top of satisficng requirements of multiple nutrients, different
strategies of maximization of specific nutrients can also be explored.
In this study, by using linear programming models we have identified possible
mechanisms by which elephant compile their daily diets. We recognize that linear
programming should be used with caution as each outcome is highly dependent on the
specified constraints. As the exact nutrient requirements of an elephant are unknown, we
reason that the constraints can be fine-tuned as long as the nutrient requirements used as
constraints can be justified from literature and are not derived from field data.
75
76
Table 5.1
CHAPTER 6
SYNTHESIS:
WHY SIZE MATTERS AT SMALL SCALES
Y. Pretorius
77
The saying ‘size matters’ elegantly summarizes the theme of this chapter in that the focus is
on the implications of body mass differences in herbivores as determined by environmental
constraints and how a herbivore of a particular size is able to satisfy its nutrient requirements
at small scales. As a contribution to the field of large herbivore foraging ecology, I start of by
explaining some important aspects on the evolution of body mass in herbivores, the trends
and patterns in the distribution and abundance of their forage resources and how this may be
used to explain the upper limits of body mass. Next, I demonstrate the implications of the
results at each spatial scale by comparing the foraging strategies of impala and elephant,
which are both mixed feeders but with large differences in morphology and body mass.
Finally, using a conceptual model I expalin the implications of forage selection by a large
herbivore on other species as a possible example of bottom-up and top-down effects
cascading through different trophic levels. In conclusion I propose how different foraging
mechanisms act at spatial scales ranging from bites to daily foraging ranges.
Evolution of being big
The first question of interest when looking at animals of different body mass, is ‘what are the
advantages of becoming bigger/larger??’ Various explanations for increases in body mass
have been proposed and a growing number of studies indicate that fecundity selection in
females and sexual selection in males are the major evolutionary forces that select for larger
body mass in many organisms (Blanckenhorn 2000). Other advantages of being large include
greater physiological homeostasis, reduced predation risk and the ability to utilize resources
out of reach of other herbivores (Owen-Smith 1988, Brown & Sibly 2006). But if bigger is
better the next question that I ask is ‘why aren’t there any terrestrial mammals larger than an
African elephant?’
Constraints to the upper limit in body mass of terrestrial mammals have been attributed to
surface-to-volume ratios (Alexander 1989) and the metabolic cost of gravity (Economos
1981). This upper limit has been calculated as 20 tons (Economos 1981), which is similar to
the weight of the largest terrestrial mammal ever thought to be alive, Indricotherium
transouralicum (Fortelius & Kappelman 1993). Clauss and co-workers (2003) formulated an
interesting hypothesis explaining the physiological upper limits of fermentation which results
in the largest ruminant grazer being smaller than the largest ruminant browser and the largest
ruminants that ever existed not being significantly larger than extant ones. The lower quality
food consumed by larger herbivores and especially grazers requires shorter retention times.
This is more easily achieved in hindgut fermenters but this requires an increase in the size of
the reticulo-rumen in foregut fermenters with further limits the space for other organs (van
Soest 1994, Clauss et al. 2003). However, these physiological explanations for the upper
78
limits of body mass still do not explain why the largest extant terrestrial mammal today
weighs around 4000 kg.
The fibre and food trap
A further clue to the reason for an upper limit in body mass might be found in the major
extinctions of mega-herbivores during the history of life on earth. Although some relate these
extinctions to climate change (Prins 1998, Janis 2008) and others to the extirpation of
predators by humans, which precipitate population booms in herbivores and subsequent
ecosystem crashes (Whitney-Smith 2008), most postulated explanations agree that a change in
the quality and abundance of forage resources ultimately leads to these extinctions (Prins &
van Langevelde 2008). As examples I will look at the extinction of mega-herbivores during
the Cretaceous and Tertiary as well as the status of extant herbivores.
Cretaceous Period (144 – 65 Mya)
During the early Cretaceous period precipitation was high and the productivity of the earth
was nearly double that of today (Beerling et al. 1994). The globe was dominated by Jurassic
ferns and gymnosperms (Ehleringer et al. 2005). Dinosaurs represented the main herbivore
assemblage of which sauropods made up a significant part. Sauropods were the largest
animals ever to occur on earth and typically had wide mouth openings and no cheek muscles
for mastication which facilitated a high intake capacity without the need to chew their food.
Their diets consisted mostly of gymnosperms and soft-tissued plants such as ferns and
ginkgoes (Hummel et al. 2008).
As CO2 levels and precipitation started to decline during the mid-Cretaceous period
angiosperms appeared and expanded in conjunction with a decline in sauropod diversity
(Coria & Salgado 2005). It is believed that CO2 starvation during the Cretaceous contributed
to the taxonomic diversification of angiosperms (McElwain et al. 2005). At lower CO2
concentrations plants have to keep their stomata open for longer periods, the consequence of
which is greater water losses through transpiration. Unlike with gymnosperms that have
softwood and no vessels in their xylem, angiosperms developed vessels for the more efficient
transport of water. These vessels consisted of elongated perforate cells with lignified
secondary walls and no protoplasm (Logan & Thomas 1987). Although it was previously
thought that there was a co-evolution between dinosaurs and angiosperms Butler (2009) who
tested this hypothesis extensively, found more evidence for the opposite being true in that
dinosaurs started to decline when angiosperms increased. Therefore, I reason that an increase
in the structural tissue of plants, which translates to an increase in fibre, and a decline in the
abundance of plant biomass of especially gymnosperms could have lead to the decline in
sauropod diversity as their wide mouths and inability to select nutrient rich plant parts would
have forced them to consume plants and plant parts with a higher fibre content.
79
Tertiary Period (65 – 1.8 Mya)
The Cenozoic era (65 Mya – today) is commonly referred to as the ‘age of mammals’ as this
period marked the major evolution and diversification of mammals (Janis 1993). The
vegetation of the Tertiary was dominated by angiosperms with their smaller leaves and higher
fibre content than the angiosperms from the mid-Cretaceous (McTwain et al. 2000). This may
have stimulated the appearance of the first ruminants during the Eocene. These small
browsers would have had an advantage over monogastric animals in coping with the more
fibrous forage of lower biomass by using their specialized compartments to host micro
organisms which could digest fibrous plant cell walls (Janis 1989). Mega-herbivores such as
Brontotheres went extinct at the end of the Eocene as the earth cooled down and became
even drier during the Oligocene period (33.7-23.8 Ma) (Janis 2008). Plant biomass decreased
further and tropical forests turned into sub-tropical forests (Pearson & Palmer 2000). More
over, seasonality increased and this led to a greater differentiation of fibre content between
plant leaves and stems (Janis 2008). Mega-herbivores such as browsing Indricotherium
species roamed the earth during this time but disappeared towards the beginning of the
Miocene (Lucas & Sobus 1989) just as CO2 levels declined again and the first plants using a
C4 photosynthetic pathway appeared (Pearson & Palmer 2000, Christin et al. 2008,
Bouchenak-Khelladi et al. 2009).
Only the most advanced families of angiosperms, which mostly consist of grasses and
some sedges, evolved C4 photosynthetic pathways to cope with climate change (Sage 2004).
Significantly, the appearance of C4 grass and ungulate adaptations to C4 dominated habitats
occurred at a similar time (Bouchenak-Khelladi et al. 2009). C3 grasses have higher levels of
protein and lower levels of structural carbohydrates than C4 grasses (Barbehenn et al. 2004)
but by the late-Miocene C4 grasses contributed significantly to vegetation globally resulting in
a diversification of ruminant grazers with superior fibre digestive efficiencies (Janis 1989).
During the late Miocene browse quality decreased progressively again causing another
decline in large bodied browsers (Janis et al. 2000). Ruminants, although restricted by a
physiological upper limit in size (van Soest 1994, Clauss et al. 2003), had an advantage over
other animals by hosting micro organisms specialized to digest cellulose which are especially
high in C4 grasses. As with dinosaurs during the Cretaceous I reason that most of the
extinctions of mega-herbivores during the Tertiary were caused by a decline in plant biomass
and an increase in the structural components of plants.
Present day
A significant part of the earth’s vegetation biomass currently consists of C4 plant species,
which are all angiosperms that mostly include monocotyledonous grasses. Unsurprisingly
therefore, global diversity and the biomass of extant large herbivore species are dominated by
80
ruminant grazers (van Soest 1994, Fritz & Loison 2006, Prins & Fritz 2008). When looking at
global species density over body mass, the species distribution curve for mammals is left
skewed with a slight under abundance of animals around 1 kg and an over abundance of
animals around 300 kg (Clauset & Erwin 2008). Similar patterns have also been found at
regional scales for large herbivores as well as for plants (Richie & Olff 1999). It has long
been known that larger herbivores can tolerate lower quality food but need a higher biomass
than smaller herbivores because the metabolic rate scales to ¾ powers of body mass (Kleiber
1932) whereas gut capacity scales isometrically with body mass (Bell-Jarman principle: Bell
1971 & Jarman 1974). Therefore, as plant available moisture reduces the nutrient content of
plants but increases productivity, and plant available nutrients increase both these factors, the
highest herbivore diversities should and does occur in locations with intermediate moisture
and high nutrients, which benefits both small and large animals (Olff et al. 2002). These
findings suggest that forage resource abundance and quality somehow regulate body mass
distributions.
Results from this thesis re-confirm that larger animals are restricted more by food
availability while smaller animals need higher quality forage (Chapter 5). I also provide an
alternative explanation for the upper limits of body mass in terms of the challenges faced by
herbivores due to the dilution of forage resources in space and how they have adapted by
changing their mouth morphologies (Chapter 2). Past and present climate trends and habitat
types may explain herbivore body mass distributions and morphologies (Janis 1993). I
conclude that increased CO2 levels, declines in precipitation and more prominent seasonality
will lead to a decreases in plant biomass, larger differences between the quality of plant parts
and an increase in structural compounds such as fibre. The latter are strongly linked to the
diversity of large herbivore taxa as well as the upper limits of herbivore body mass (Prins &
Gordon 2008). In the following sections I demonstrate how abundance and fibre content of
forage constrain the intake of mega-herbivores in particular.
The mouth size paradox
Compared to sauropods, mammalian herbivores have molars and cheek muscles to allow
mastication of food. Even though the largest sauropod, Brontotheres which was the size of
extant elephant, had unique upper molars with W-shaped ectolophs to help with shearing of
food, they did not have extreme sized soft mouth parts such as trunks (Mihlbachler 2008).
Indricotheriums, which diversified mainly when Bronthotheres went extinct, had tusk-like,
elongated incisors of which the upper pair pointed straight down and the lower pair pointed
outward (Lucas & Sobus 1989). This type of dentition along with a long neck allowed these
animals to strip leaves from trees at levels that were out of reach of other animals. Therefore,
as the vegetation deteriorated through time it seems that mega-herbivores adapted their mouth
81
morphologies in particular on order to grow to the sizes they did. However, sudden changes in
climate may have caused extinctions as there was no time to adapt their mouth morphologies.
The ancestors of extant elephant, Moeritherium, which occurred during the same time as
Brontotheres in the Eocene, were small pig-like animals without trunks but with elongated
tusk-like incisors in both jaws (Palmer 1999). These animals would not have been affected as
badly as larger herbivores by the decline in forage biomass, and due to their small size would
have been able to select more nutrient rich plant parts even if the structural components
increased. Further, as with extant wild boars, the tusk-like incisors would have allowed the
animal to dig up the otherwise out of reach roots and tubers below the ground.
Following the Eocene, body mass and trunk development of proboscideans gradually
increased and the development of high-crowned molars indicated a diet which included
increasing amounts of grass (Cerling et al. 1999). Although more recent studies caution
against the assumption that the development of hypsodonty is an evolutionary response to
diets including more grass, this assumption is accepted for the time being as many extant
grazers have high-crowned teeth and there is little evidence that this is in fact not the case
(Strömberg 2006). Unsurprisingly, the Miocene diversification of the proboscidean taxa
correlates well with the appearance and expansion of especially C4 grasses (Shoshani 1998).
Due to increasing atmospheric CO2 levels at the end of the Pleistocene period, C4 grasslands
started to decline again (Ehleringer et al. 1997, Cerling et al. 1998), which may explain why
most extant African elephant browse more than graze.
To illustrate the importance of mouth size and morphology for satisfying nutrient
requirements of different size herbivores, I modeled and compared the forage intake of impala,
elephant and Indricotherium on grass and browse. Further, for each animal I compared
nutrient gains per intra-dental mouth volume and per total mouth volume (see Chapter 2) to
illustrate how elongated soft mouth parts of mega-herbivores are necessary for survival.
Model calculations
For each of the herbivores, intra-dental mouth volume (IMV) and total mouth volume (TMV)
were calculated from their body masses according to the scaling equations derived in Chapter
2 (IMV = 6.89 (BM0.89), TMV = 0.26 (BM1.72)). Because the plant thinning law predicts that
the number of plants per unit space scales to 0.75 power of average plant mass (Enquist et al.
1998) and similarly that standing leaf biomass scales to 0.75 power of standing stem biomass
(Niklas 2004), I assumed that stem mass scales to volume to the power 0.75 which means that
leaf mass should scale with volume to 0.56 (as the product of 0.75 x 0.75). However, these
relationships should only hold true when an animal takes a bite using its intra-dental mouth
volume but not any soft mouth parts. This is because soft mouth parts are used to increase the
leaf mass per bite so that the stem mass included in the mouth volume, when the volume
covered by soft mouth parts are also included, should have a ratio with leaf mass of higher
82
than 0.75 and closer to one. Thus for total mouth volume, stem mass should, as with leaf mass,
scale to volume to the power of 0.56. Indeed, data from Chapter 2 on the bite mass to bite
volume relationships/ratio? reveal a similar scaling factor.
Because stem mass per unit volume is heavier than leaf mass for the same size volume
and even more so for trees than for grasses, mass gain per bite volume was calculated for
browse and grass for each animal species using the equations in Figure 6.1. “Compensated”
stem mass indicated that stem mass included in a bite when soft mouth parts were added to
the intra-dental mouth volume. Acid Detergent Fibre (ADF) were used as a proxy for
calculating the digestible cell mass, and average ADF values for browse and grass were
obtained from chemical analysis conducted on leaf and stem material throughout the study
(Chapter 2, 4, & 5). Digestible cell mass per bite was calculated from stem and leaf mass
derived from Figure 6.1 as:
Browse Stem Mass / bite (g) = 0.002 Intra-dental mouth volume0.75
Browse Leaf Mass / bite (g) = 0.002 Intra-dental mouth volume0.56
Digestible cell mass / bite = [Browse leaf mass / bite x (100 – % browse leaf ADF)] +
[Browse stem mass / bite x (100 - % browse stem ADF)]
Digestible cell mass per bite for grass and for mouth volumes including soft mouth parts were
calculated in the same way. A bite rate of 0.8 bites per second was used as a baseline for the
maximum bite rate achieved by a 10 kg animal. Because bite rates on grass were consistently
higher than on browse, the maximum bite rate for a 10 kg animal on browse was estimated at
half that of grass (0.4 bites/s). Because chewing rate scales negatively as BM-0.25 (Fortelius
1985) and chewing rate is linearly correlated to bite rate (Demment & Greenwood 1988), bite
rate might be expected to scale as BM-0.25. Hence, bite rates on browse were calculated as:
Bite rate = 0.715 (BM-0.25) and on grass as:
Bite rate = 1.43 (BM-0.25)
Finally, instantaneous intake rate was calculated as the product of digestible cell mass per
bite and bite rate. Because the equation derived for instantaneous intake rate (IR) from
Chapter 2 (IR = 0.001 (BM0.75)) represented 0.1 % of daily metabolic requirements (MJ) (MJ
= BM0.75)), I assumed that any percentage less than that calculated for the current model
would mean that the animal would not be able to satisfy its daily metabolic requirements.
Model results and implications
According to the model, Indricotherium would have only been able to satisfy its daily nutrient
requirements if it had elongated soft mouth parts and fed on grass (Table 6.1). However, when
using the equation from Chapter 2 for the scaling relationship between soft mouth part length
83
(SL) and body mass to calculate the mouth size required by Intricotherium to cover the
predicted bite volume used in the model, lip and/or tongue length would have had to be 1.69
m long (SL = 0.345 (BM0.63)). This might have put too much strain on the already
compensated for long neck (Seymour 2009). More over, the model shows that an elephant
will not be able to survive and satisfy its daily nutrient requirements without a trunk.
Although I realize that the comparison between nutrient gain from grass and browse rests
heavily on the assumptions of mass-space relationships and differences in bite rates between
the two forage types, the model adequately illustrates that nutrient gains for these different
forage types are comparable and that browsers and grazers do not have to be kept separately
when comparing animals of different body mass.
Hummel and co-workers (2008) recently used chemical analysis to demonstrate how a
sauropod could have survived on these plants using extant gymnosperms. However they only
used leaf material in their analysis of scaling relationships thus ignoring the dilution of plant
mass in space as well as the difficulty that such a big animal would have had to select leaves
among stems. Although it seems to have taken longer to evolve, enlarged soft mouth parts
allow mega-herbivores greater flexibility in coping with a decline in plant abundance and
quality. Therefore, at the relatively deteriorated state of extant forage, compared to the
Cretaceous and Tertiary, mega-herbivores such as the African elephant are still able to survive.
300
0.75
Browse Stem Mass = 0.002V
250
Mass (g)
200
150
0.75
Grass Stem Mass = 0.001V
Compensated Browse Stem
0.56
Mass = 0.005V
(-------)
100
Compensated Grass Stem
0.56
Mass = 0.003V
50
(-------)
0.56
Browse Leaf Mass = 0.002V
0.56
Grass Leaf Mass = 0.001V
0
0
1E+06 2E+06 3E+06 4E+06 5E+06 6E+06 7E+06
Volume (cm3)
Figure 6.1: Proposed mass-space scaling relationships for leaves and stems of grass and
browse plants in the South African savanna. “Compensated” stem mass indicated the
stem mass included in a bite when soft mouth parts were added to the intra-dental mouth
volume.
84
Herding
In the previous section I discussed how mouth morphology of especially mega-herbivores is
an adaptation to past and current environmental conditions through the process of evolution.
However, body mass as a function of resource availability is not only restricted to the
individual, as a herd of animals also possesses a total mass with metabolic requirements that
need to be satisfied. Ecologists have proposed many explanations for herbivores grouping
together including reducing predation risk (Jarman 1974, Molvar & Bowyer 1994), sharing
information about resources (Prins 1989, Dall et al. 2005) and maintaining high quality food
resources (e.g., grazing lawns, Vesey-FitzGerald 1960, McNaughton 1984, Fryxell 1991).
However, herding also has its disadvantages as intra-specific competition increases (Shrader
et al. 2006, Smallegange et al. 2006). Cromsigt and Olff (2006) suggested that group size
differences among species might be very important in testing allometric hypotheses such as
patch selection but to my knowledge no study has illustrated how group size affects the scale
at which herbivores select nutrient rich patches. In the following model I explore how herd
living animals benefit more from maximizing instantaneous intake rate rather than selecting
nutrient rich areas at larger scales.
Model calculations
For the model, 50 bites each by steenbok and impala antelope in three plot types were used
for comparison as in the experiment in Chapter 4 and 5 (Table 6.2). The first plot consisted of
25 small 4 m2 patches (100 m2 fertilized area in total) per 2500 m2 plot with a high local
nutrient concentration per patch, resulting in a medium total nutrient load per plot. The second
plot consisted of 5 patches of 10 m2 each (500 m2 fertilized area in total) with a medium local
nutrient concentration per patch, resulting in a medium total nutrient load per plot and in the
last plot the whole 2500 m2 area was fertilized with a medium local nutrient concentration that
resulted in a high total nutrient load.
The percentage digestible material in high local nutrient concentration patches was
assumed to be 70 %, in medium local nutrient patches 55 % and in unfertilized patches 50 %.
Bites were divided among patches according to the total area fertilized per plot and for herd
animals it was assumed that half of the bite locations had already been visited by another
animal. Therefore, at these previously visited locations, bite mass was assumed to be 10 %
less and nutrient concentration 50 % less due to selection for nutrient rich plant parts by the
previous animal. Bite mass was calculated according to the equation derived from Chapter 2
(0.002BM). The digestible cell mass per patch was calculated as the product of bite mass, the
number of bites per patch and the percentage digestible cell mass. Next, all the patch gains for
each plot were added to obtain the total digestible cell mass gained with 50 bites from a plot.
For comparison, gain per bite was calculated.
85
86
As digestibility was assumed to scale similarly to body mass as metabolic rate (BM0.75), I
back transformed the digestible cell mass per bite by using the bite mass equation (0.002BM)
in combination with the scaling of metabolic rate to see if all body mass requirements would
be met (Digestible cell mass = 0.002BM0.75 -> BM = (Digestible cell mass/0.002)1.33).
Model results and implications
From this model it is first of all clear that small antelope such as steenbok would not be able
to satisfy the metabolic requirements for their body mass when living in a herd (Table 6.2).
Secondly, solitary animals seem to gain more by selecting plots with the highest total nutrient
loads whereas herd living animals have to select plots where the local nutrient concentration
per patch is highest. From this I conclude that the phenomenon of herbivores aggregating
together in large numbers would be more common in larger animals that forces these animals
to maximize their instantaneous intake rates to overcome the effects of intra-specific
competition.
Nutritional trade-offs of being big versus being small
After demonstrating how large herbivores have adapted either physically, through their mouth
morphologies, or behaviourally by being able to select nutrient-rich patches at small scales, I
focused on the environmental constraints that force an animal of a particular body mass to eat
what it eats. Metabolic rate in heterotrophs equals the rate of respiration, which is the main
process that sources energy (Brown et al. 2004). Therefore, when determining the value of a
plant species to an herbivore, not only does the energy content of the plant need to be
considered but also the increased respiration and energy expended to find it. This becomes
especially important for a large animal with high absolute energy demands like the elephant.
To better illustrate the trade-offs between forage abundance and quality for different size
animals, I compared the diets of elephant and impala as calculated using the linear
programming model developed in Chapter 5.
An impala of 50 kg requiring a daily dry matter intake of 2.5 % of body weight consumes
1.3 kg of food per day (Pieterson et al. 1993). When using acid detergent lignin (ADL)
content of plant species from our study, the in vitro digestibility of forage organic matter
(IVDOM) can be calculated and these values used to estimate metabolizable energy value of
plant species for impala (Pieterson et al. 1993). When using the model to calculate energy
gain under the assumption that both elephant and impala eat 100 % of a rare, low fibre plant,
like Panicum maximum, elephant turn out to be deficient in energy whereas impala easily
satisfy their minimum requirements for energy (Table 6.3). However, if both elephant and
impala eat 100 % of an abundant, high fibre plant, like Combretum apiculatum, impala
become deficient in energy whilst elephant gain sufficient energy to satisfy minimum energy
requirements.
87
Table 6.2
An example of the nutrients gained from 50 bites in one of three 2500 m2 plots by steenbok
when living alone (a) or in a herd (b) and by impala when living in a herd (c). When
calculating the amount of body mass requirements satisfied (last column) from the average
digestible cell mass gained per bite using scaling relationships of bite mass with body mass,
this value should be equal to or more than actual body mass (fourth column).
Animal
Species
a
Steenbok
Plot
Nr
Patch
Size
(m 2)
Body
Mass
(kg)
1
100
10
2
500
10
3
2500
10
b
Steenbok
1
100
10
2
500
10
3
2500
10
c
Impala
1
100
50
2
500
50
3
2500
50
Bite
Mass
(0.002*
BM)
Patch
0.020
0.020
0.020
0.020
In fresh
Out fresh
In fresh
Out fresh
5
45
10
40
70
50
55
50
0.070
0.450
0.110
0.400
0.020
In fresh
50
55
0.550
0.020
0.018
0.020
0.018
0.020
0.018
0.020
0.018
0.020
0.018
In fresh
In eaten
Out fresh
Out eaten
In fresh
In eaten
Out fresh
Out eaten
In fresh
In eaten
5
5
20
20
10
10
15
15
25
25
70
35
50
28
55
28
50
25
55
28
0.070
0.032
0.200
0.101
0.110
0.050
0.150
0.068
0.275
0.126
0.100
0.090
0.100
0.090
0.100
0.090
0.100
0.090
0.100
0.090
In fresh
In eaten
Out fresh
Out eaten
In fresh
In eaten
Out fresh
Out eaten
In fresh
In eaten
5
5
20
20
10
10
15
15
25
25
70
35
50
28
55
28
50
25
55
28
0.350
0.158
1.000
0.504
0.550
0.252
0.750
0.338
1.375
0.630
%
#
NonBites
ADF
Digestible
Cell Mass
(g) / Patch
Type
Digestible
Cell Mass
(g) / 50
bites in
2500m 2
Plot
Digestible
Cell Mass
(g) / bite
Body
Mass
Satisfied
(kg)
0.520
0.010
9
0.510
0.010
9
0.550
0.011
10
0.402
0.008
6
0.378
0.008
6
0.401
0.008
6
2.012
0.040
55
1.890
0.038
50
2.005
0.040
54
* Plot 1 consist of 25 x 4 m2 patches with high local nutrient concentration per patch resulting in a
medium total nutrient load for the plot, plot 2 consist of 5 x 100 m2 patches with medium local
nutrient concentrations resulting in a medium total nutrient load for the plot and in plot 3 the whole
2500 m2 have medium local nutrient concentration resulting in a high total nutrient load (see Chapter
3 &4 for detail).
88
Recently, Prins and van Langevelde (2008) re-introduced the use of linear programming to
explain how large herbivores assemble their diets from different places. They generated a
hypothesis that predicts that animals that are dependent on surface water during the dry
season will better be able to assemble a diet during the wet season when resources are more
available compared to the dry season. Indeed, from the findings in Chapter 5 of this thesis the
89
maximization of energy intake and the constraints of forage abundance during the dry season
on elephant seem to support this idea. The inclusion of grain size in Prins and van
Langevelde’s (2008) model reveals that inter-patch distance affects smaller animals more than
larger ones. However, even though the landscape seem homogenous at larger scales, much
heterogeneity in quality of plant species still exists at small scales so that smaller animals do
not need to travel as far as large herbivores because they are not as constrained by forage
biomass as large herbivores are. Secondly, the various plant species used in the model from
Chapter 5 can be seen as patches because the abundance of each species was expressed in
distance together with the amount of energy required to obtain each species.
However, the model of Prins and van Langevelde (2008) included the energy required to
cover the distance between two patches in the total energy requirements of the animal and not
explicitly as the value of each patch to the animal. This means that their model could not
account for more than two patches of the same kind or of different qualities and biomass.
Hence, contrary to the model of Prins and van Langevelde (2008), the model from Chapter 5
illustrates that a small herbivore is able to live solely off a rare plant with large inter-patch
distances whereas animals as large as an elephant cannot. The fact that very large animals are
more restricted by biomass and small animals by quality not only re-confirms the Bell-Jarman
principle (Bell 1971 & Jarman 1974) but also provides an explanation for mega-herbivores
being less common in areas of low resource abundance and why the greatest diversity of
herbivores occurs in areas of intermediate resource quality and quantity (Olff et al. 2002).
A tribute to Hutchinson’s homage to Santa Rosalia
Half a century ago, Hutchinson (1959) wrote a seminal paper proposing an explanation for
animal diversity. Through many examples he illustrated that within a trophic guild and in one
habitat no two species of animals have the same body mass and morphology. Since then the
debate on co-existence of species and niche separation through resource partitioning has been
ongoing (Brown 1981, Prins & Olff 1998, Ritchie & Olff 1999, Prins & van Langevelde
2008). In a recent study on grazing ungulates in South Africa, resource partitioning by grass
height could not be explained by body mass alone but indicated the importance of the scaling
of mouth width relative to body mass and hence metabolic demands (Arsenault & OwenSmith 2008). Bite size may contribute to habitat segregation among species because small
differences in available bite sizes can generate larger differences in intake rates among
herbivores of different sizes (Shipley 2007). As a tribute to Hutchinson and a contribution to
the general body of knowledge on this subject I will illustrate the conditions necessary for
herbivores of different body mass to survive under the constraints of their mouth
morphologies.
90
If bite mass increases linearly with body mass (Chapter 2) and quality per bite required by
small animals is higher than for large animals, the trade-offs between quality and quantity can
be depicted as in Figure 6.2 (a). This also follows from Kleiber’s allometry (Kleiber 1932,
Prins & van Langevelde 2008) and from the Bell-Jarman principle that states that as body
mass and gut size increase isometrically, larger animals with their lower metabolic energy
requirements can tolerate lower quality diets and can therefore afford to be less selective (Bell
1971, Jarman 1974, van Soest 1996).
Further, according to plant mass distribution patterns in space (Enquist et al. 1998, Niklas
2004), plant mass can be plotted to increase with volume at a decreasing rate whereas plant
quality is plotted to decrease with volume because leaf to stem ratios decline resulting in
similar patterns as was previously explained for animal requirements in relation to their body
mass (Figure 6.2 (b)).
When the forage availability and herbivore requirement graphs are combined an
indication is obtained of whether a particular habitat is able to adequately provide the
metabolic requirements of a herbivore (Figure 6.2 (c) & (d)). More specifically, if we assume
that browse is heavier than grass per unit space and that browse quality decreases more
quickly in space, because the stem of the browse plant comprises an increasing amount of the
plant with higher quality differences between browse stem and leaf, than grass stem to leaf,
another interesting property emerges. The shaded areas in Figure 6.2 (c) and (d) where the
distance between the animal’s requirements and forage availability is greatest is larger with
grass than with browse. I hypothesize that the greater biomass and quality of food available in
these regions, allows animals within the body mass range covered by this shaded area to herd.
Therefore, because this region is bigger for grass it might explain why grazers tend to
aggregate together in larger numbers than browsers.
However, simply combining the forage availability and animal requirement graphs
reveals that larger animals cannot satisfy their forage requirements. By combining the animal
requirement and forage abundance graphs with a graph on the relationship of volume with the
body mass of animals (Chapter 2), a three dimensional model can be derived that includes the
compensation of larger bite volumes by mega-herbivores. This model further illustrates that if
forage biomass is low, larger animals cannot satisfy their metabolic requirements (Figure 6.3
(a)), whereas if forage quality is low small herbivores cannot satisfy their metabolic
requirements (Figure 6.3 (b)).
91
Figure 6.2: Illustration of the trade-offs between quality and quantity of forage requirements
for different sized animals (a) and the distribution of forage mass and quality in space (b).
When a forage resource distribution graph of browse or grass (b) is combined with the
forage requirement graph of animals (a), two better explanatory graphs for browse (c) and
grass (d) emerges. The grey regions in these graphs (c & d) where the availability of
either browse or grass in relation to requirements is greatest may allow animals within the
body mass range covered by this region to group together.
92
(a)
(b)
Figure 6.3: A three dimensional conceptual model on the relationship between metabolic
requirements of a herbivore, forage availability and space. Because of the addition of bite
volume against body mass (compared to Figure 6.2), the compensation of bite volume by
mega-herbivores allow them to gather enough forage resources (elephant in b). However,
when forage biomass in the environment is limiting (a) large animals such as elephant are
not able to satisfy their requirements whereas where forage quality is limiting (b) small
animals such as impala cannot satisfy their metabolic requirements. The little animal
figures indicate the corresponding points on the axis of the graphs; the grey figures
indicate when requirements are not met and the grey lines serve as guidelines.
93
In a synthesis of the progress in the field of resource ecology up to date, Prins and van
Langevelde (2008) hypothesized that the relative abundance of large animals (in contrast to
small) in assemblages increases with spatial variation and sudden fluctuations in resource
availability that result in longer time periods between foraging events. I partially agree but
reason that within the constraints of metabolic requirements for a specific body mass a
herbivore’s perception of spatial scale is first of all as small as its mouth morphology allows it
to be and secondly as large as the variability in the abundance and quality of its forage
resources allows it to be. As all living organisms have to abide by the same physical laws
(Gould 1970) and evidence exist for the universal scaling of biological rates (West et al. 1997,
Enquist et al. 1998) across plant and animal species, species-coexistence and diversity within
a trophic level is simply a product of evolution to adapt to bottom-up effects from changes in
the spatial distribution of resources at lower trophic levels. In the next section, I demonstrate
how these bottom-up effects determine a large herbivores foraging pattnerns and how
foraging behaviour can feedback through top-down effects to lower trophic levels.
The bigger picture
To provide context, I developed a conceptual model on the implications of large herbivores
forage selection. Herbivores select forage resources of high quality and abundance at multiple
scales. In the past much controversy existed on whether primary control throughout trophic
levels is due to resources (bottom-up control) or predators (top-down control) (Power 1992).
Instead of exclusive top-down or bottom-up control more recent studies indicate shifting
control, with resource limitations dominating in dry years and biotic interactions dominating
in wet years (Meserve et al. 2003) or where top-down factors such as herbivory shape
vegetation patterns although bottom-up factors such as rainfall still sets the upper bound of
what productivity is attainable in a system (Groen 2007). To better understand primary
control throughout trophic levels, the dynamic feedbacks between adjacent and non-adjacent
trophic levels have to be addressed (Power 1992).
Cascading effects are especially prevalent for ecosystem engineers, which are organisms
such as elephant that directly or indirectly modulate the availability of resources to other
species by causing changes in the physical state of biotic and/or aboitic components (Jones et
al. 1994). The impact that African elephant have on trees has been a topic of great
controversy (Wiseman et al. 2004, de Beer et al. 2006, Lawes & Chapman 2006, O’Connor et
al. 2007, Chafota & Owen-Smith 2009, Scholes & Mennel 2008) and some studies speculate
that the occurrence of high, detrimental impact on trees might be related to the availability of
surface water (de Beer 2006) or sudden changes in the availability of preferred forage species
(Chafota & Owen-Smith 2008).
94
In Chapter 3, I used an experimental approach to demonstrate the possible consequences
of very large animals being able to select nutrient rich patches and even plant parts at very
fine scales. I now expand on this using a conceptual model of the possible bottom-up and topdown effects of elephant foraging cascading up and down through trophic levels and feeding
back into the abiotic environment (Figure 6.4).
I reason that under nutrient poor conditions elephant will impact on trees randomly with
the predominant impact being uprooting and pollarding. As there are few nutrients available
in this scenario, tree seedlings will outcompete grasses (van der Waal et al. 2009) and replace
the killed trees (Figure 6.4). The phenomenon of elephant re-utilizing trees and inducing
coppicing might lead to some heterogeneity in the vertical structure of the tree layer. At high
elephant densities most trees will be killed leading to large bare areas with low cover of
nutrient poor herbaceous plants that establish under reduced competition from trees. Evidence
of such a scenario has been described on the sandy nutrient poor soils of Etosha (Strohbach
2000) by de Beer (2006) who found that areas that were heavily utilised by elephant also
experienced lower rates of woody plant survival and that, coupled with low rainfall and high
elephant impact, decreased tree recruitment even further.
In a woodland where there is a patchy distribution of nutrients, elephant will select
nutrient rich patches particularly where patches are aggregated, larger or have higher nutrient
concentrations. Because elephant re-utilize these nutrient rich patches more and mostly
consume leaves, leaf density will increase (Smallie & O’Connor 2000) and more nutrients
will be added, for example from dung deposition, which can potentially initiate a positive
feedback loop (van der Waal, unpublished). The increased presence of elephant in areas of
high nutrient concentration also means that neighbouring nutrient poor trees have a higher
chance of being killed. Under a scenario with a patchy nutrient distribution and high elephant
densities, small patches might become over utilized and nutrients depleted, which may lead to
the complete removal of woody patches and replacement by palatable grass species. High
quality grass establish particularly in the vicinity of trees due to the trees’ ability to enhance
nutrient availability as nutrient pumps (Ludwig et al. 2008, Treydte et al. 2009). Under
uniform nutrient rich conditions, elephant will utilize trees at random but because of
preference for previously browsed nutrient rich trees, utilization will become more
pronounced at these spots. This might lead to a re-distribution and concentration of nutrients .
Under high elephant densities, the trees that have become less nutrient rich will be killed and
replaced by nutrient rich grasses.
Elephant are renowned for being ecosystem engineers (Jones et al. 1994) and are
implicated as one of the main biotic factors determining the tree grass ratios of savannas (van
de Koppel & Prins 1998). By concentrating limiting resources, patchiness may play a critical
role in maintaining ecosystem productivity (Aguiar & Sala 1999).
95
Figure 6.4: Conceptual model on the predicted effects elephant will have on woodlands
depending on the nutrient status of the soil.
96
Although elephant may contribute to homogenization of the landscape under nutrient poor
conditions, in nutrient rich areas the opposite is more likely with elephant enhancing
landscape heterogeneity. As this study shows, elephant are able to select nutrient rich areas at
a range of spatial scales and thus provides some insight as to why spatial vegetation patterns
in savannas are so large and difficult to predict. Thus future research in this field should take
the spatial scaling effects of elephant foraging into consideration.
Scaling conclusions
To conclude this synthesis I look at the spatial scales at which the mechanisms that operate to
allow an animal of a particular body mass to survive in a particular area. Ritchie and Olff
(1999, 2004) developed a theory with a novel approach to explain the diversity and coexistence of species using fractal geometry. They showed that co-existence through resource
partitioning is possible as larger herbivores perceive less spatial detail than smaller animals.
However, the only field experiments conducted to support this theory could not find any
relationship between body mass and the spatial scale of patch selection (Cromsigt & Olff
2006). Despite this there was some support in Chapter 4 for larger animals preferring larger
patches, but smaller animals were also able to maximize nutrient intake at larger scales and
mega-herbivores such as elephant were able to select nutrient rich patches at very small scales
(Chapter 3).
Ultimately it is the animal’s perceptions and foraging responses that must define the
boundaries between the subunits within hierarchical scales (Senft et al. 1987). As this is
difficult to determine accurately, Senft and co-workers (1987) proposed a theory of ecological
hierarchies for large herbivore foraging and suggested that the significance of interactive
resources to a herbivore is highest at intermediate spatial and temporal scales. Subsequently,
Kotliar and Wiens (1990) characterized spatial scale by grain and extent where grain is the
smallest measurement unit used and extent the total area of interest. I propose a mechanism
by which an herbivore’s mouth size and morphology determines the grain size with which it
measures its environment and its daily metabolic requirements as the extent of its daily
foraging range (Figure 6.5). At the bite scale, variation in the quality of plant parts such as
leaf and stem is large but herbivores are constrained by their mouth morphologies to select
bites that satisfy their metabolic requirements (Chapter 2).
Van Langevelde (2008) recently emphasized that an issue that needs more thought is
whether consumers are selected for maximizing instantaneous intake or daily intake. Given
the results of this thesis where I look at scales ranging from bite size to daily foraging range, I
suggest that maximization of nutrient intake with the associated fitness benefits takes place
between the bite scale and the daily foraging range because mouth morphology and metabolic
requirements restrict selection at these boundaries (Figure 6.5).
97
Figure 6.5: Illustration of how fitness is maximized (left vertical axis) at the patch scale in
between the mouth morphological and metabolic constraints of the bite scale and daily
foraging range. However, at these scales, between bite scale and daily foraging range,
selection will also be more difficult as variability in quality and quantity between patches of
forage of the same species composition will be lower.
98
Variation in quality and quantity between plant species is large but the inclusion of a plant
species in the diet depends on its abundance and quality in relation to the body mass and
metabolic requirements of the animal (Chapter 5). Once the constraints of metabolic
requirements are satisfied the animal can afford to select patches with higher levels of for
example N and P (Chapter 3 &4). However, selection at these intermediate patch scales will
be more difficult as variation in quality and quantity within vegetation types of the same plant
species composition will be small.
I conclude that size, as a characteristic which includes mass and morphology, matters
because it first of all determines whether a herbivore will be able to survive in a particular
habitat or not and hence the likelihood of it going extinct when climate changes. Secondly, it
determines whether herbivores are able to herd or not which further determines the spatial
scale at which nutrient intake will be maximized. Future research needs to better define the
scaling of spatial patterns in plants to enable a testing of the hypothesis on the conditions
necessary for large herbivores to herd. Finally, for the time being, like Hutchinson, I propose
to be content with an appreciation of the beauty and magnificence of large herbivore diversity.
“I cannot refrain from pointing out the immense scientific importance of obtaining a
really full insight into the ecology of the large mammals of Africa while they can still
be studied under natural conditions. It is indeed quite possible that the results of
studies on these wonderful animals would in long-range through purely practical
terms pay for the establishment of greater reservations and National Parks than at
present exist” – (Hutchinson 1959)
99
100
REFERENCES
Adler, P.B., D.A. Raff & W.K. Lauenroth. 2001. The effect of grazing on the spatial
heterogeneity of vegetation. Oecologia. 128, 465-479.
Adu, E.F. & A.M. Adamu. 1982. The nutritive value and utilization of three tropical grass
hays by sheep. Tropical Grasslands. 16(1), 29-33.
Aguiar, M.R. & O.E. Sala. 1999. Patch structure, dynamics and implications for the
functioning of arid ecosystems. TREE. 14(1), 273-277.
Alendal, E., S. de Bie & S.E. van Wieren. 1979. Size and composition of the wild reindeer
Rangifer tarandus platyrhynchus population in the southeast Svalbard Nature Reserve.
Holarctic Ecology. 2, 101-107.
Alexander, R.M. 1989. Dynamics of dinosaurs and other extinct giants. Columbia University
Press, New York.
Anderson, G.D. & B.H. Walker. 1974. Vegetation composition and elephant damage in
Sengwa Wildlife Research Area, Rhodesia. Journal of the South African Wildlife
Management Association. 4, 1-14.
Arsenault, R. & Owen-Smith, N. 2008. Resource partitioning by grass height among grazing
ungulates does not follow body mass relation. Oikos. 117, 1711-1717.
Augustine, D.J. & D.A. Frank. 2001. Effects of migratory grazers on spatial heterogeneity of
soil nitrogen properties in a grassland ecosystem. Ecology, 82(11), 3149-3162.
Augustine, D.J., S.J. McNaughton & D.A. Frank. 2003. Feedbacks between soil nutrients and
large herbivores in a managed savanna ecosystem. Ecological Applications. 13(5), 13251337.
Bailey, D.W, J.E. Gross, E.A. Laca, L.R. Rittenhouse, M.B. Coughenour, D.M. Swift & P.L.
Sims. 1996. Mechanisms that result in large herbivore grazing distribution patterns.
Journal of Range Management. 49(5), 386-400.
Bailey, D.W. & F.D. Provenza. 2008. Mechanisms determining large-herbivore distribution.
In: H.H.T. Prins & F. van Langevelde (Eds.). Resource ecology: spatial and temporal
dynamics of foraging. Springer, Netherlands.
Ball, J.P., K. Danell & P. Sunesson. 2000. Response of a herbivore community to increased
food quality and quantity: an experiment with nitrogen fertilizer in a boreal forest.
Journal of Applied Ecology. 37, 247-255.
Barnes, R.F.W. 1982. Elephant feeding behaviour in Ruaha National Park, Tanzania. African
Journal of Ecology. 20, 123-136.
Barbehenn, R.V., Z. Chenw, D. N. Karowez & A. Spickard. 2004. C3 grasses have higher
nutritional quality than C4 grasses under ambient and elevated atmospheric CO2. Global
Change Biology. 10, 1565-1575.
101
Beauchamp, G. 2005. Does group foraging promote efficient exploitation of resources? Oikos.
111, 403-407.
Beauchemin, K.A. 1991. Effects of dietary neutral detergent fiber concentration and alfalfa
hay quality on chewing, rumen function, and milk production of dairy cows. Journal of
Dairy Science. 74, 3140-3151.
Beekman, J.H. & H.H.T. Prins. 1989. Feeding strategies of sedentary large herbivores in East
Africa, with emphasis on the African buffalo, Syncerus caffer. African Journal of Ecology.
27, 129-147.
Beerling, D.J. 1994. Modelling palaeophotosynthesis: late Cretaceous to present.
Philosophical Transactions: Biological Sciences. 346(1318), 421-432.
Bell, R.H.V. 1971. A grazing ecosystem in the Serengeti. Science America. 225, 86–93.
Belovsky, G.F. 1978. Diet optimization in a generalist herbivore: the moose. Theoretical
Population Biology. 14, 105-134.
Belovsky, G.E. 1997. Optimal foraging and community structure: the allometry of herbivore
food selection and competition. Evolutionary Ecology. 11, 641-672.
Benedict, F.G. 1936. The Physiology of the Elephant. Carnegie Institution of Washington,
Washington.
Ben-Shahar, R. & D.W. Macdonald. 2002. The role of soil factors and leaf protein in the
utilization of mopane plants by elephants in northern Botswana. BMC Ecology.
Benvenutti, M.A., I.J. Gordon & D.P. Poppi. 2006. The effect of the density and physical
properties of grass stems on the foraging behviour and instantaneous intake rate by cattle
grazing an artificial reproductive tropical sward. Grass and Forage Science. 61, 272-281.
Bergman, C.M., J.M. Fryxell, C.C. Gates & D. Fortin. 2001. Ungulate foraging strategies:
energy maximization or time minimization? Ecology. 70, 289-300.
Black, J.L. & P.A. Kennedy. 1984. Factors affecting diet selection by sheep. II. Height and
density of pasture. Australian Journal of Agricultural Research. 35, 565-578.
Blanckenhorn, W.U. 2000. The evolution of body mass: what keeps organisms small? The
Quarterly Review of Biology. 75(4), 385-407.
Bond, W. 2001. Keystone species: Hunting the snark? Science. 292(5514), 63-64.
Bond, W.J., G.F. Midgley & F.I. Woodward. 2003. The importance of low atmospheric CO2
and fire in promoting the spread of grasslands and savannas. Global Change Biology. 9(7),
973-982.
Bouchenak-Khelladi, Y., G.A. Verboom, T.R. Hodkinson, N. Salamin, O. Francois, G.
Nichonghaile & V. Savolainen. 2009. The origins and diversification of C4 grasses and
savanna adapted ungulates. Global Change Biology. Online,1-21.
102
Brooks, T.M., R. Mittermeier, C.G. Mittermeier, G.A.B. Dafonseca, A.B. Rylands, W.R.
Konstant, P. Flick, J. Pilgrim, S. Oldfield, G. Magin & C. Hilton-Taylor. 2002. Habitat
loss and extinction in the hotspots of biodiversity. Conservation Biology. 16(4), 909-923.
Brown, J.H. 1981. Two decades of homage to Santa Rosalia: Toward a general theory of
diversity. American Zoologist. 21(4), 877-888.
Brown, J.H., J.F. Gillooly, A.P. Allen, V.M. Savage & G.B. West. 2004. Toward a metabolic
theory of ecology. Ecology. 85(7), 1771-1789.
Brown, J.H. & R.M. Sibly. 2006. Life-history evolution under a production constraint. PNAS.
103(47), 17595-17599.
Bucyanayandi, J.D., J.M. Bergeron, J. Soucie, D.W. Thomas & Y. Jean. 1992. Differences in
nutritional quality between herbaceous plants and bark of conifers as winter food for the
vole Microtus pennsylvanicus. Journal of Applied Ecology 29, 371-377.
Butler, R.J., P.M. Barrett, P. Kenrick & M.G. Penn. 2009. Diversity patterns amongst
herbivorous dinosaurs and plants during the Cretaceous: implications for hypotheses of
dinosaur/angiosperm co-evolution. The Natural History Museum, European Society for
Evolutionary Biology. 22, 446-459.
Cech, P.G., T. Kuster, P.J. Edwards & H.O. Venterink. 2008. Effects of herbivory, fire and
N2-fixation on nutrient limitation in a humid African savanna. Ecosystems. 11, 991-1004.
Cerling, T.E., J.R. Ehleringer & J.M. Harris. 1998. Carbon dioxide starvation, the
development of C4 ecosystems, and mammalian evolution. Philosophical Transactions of
the Royal Society. 353, 159-171.
Cerling, T.E., J.M. Harris and M.G. Leakey. 1999. Browsing and grazing in elephants: the
isotope record of modern and fossil proboscideans. Oecologia. 120, 364-374.
Chafota, J. & N. Owen-Smith. 2009. Episodic severe damage to canopy trees by elephants:
interactions with fire, frost and rain. Journal of Tropical Ecology. 25, 341-345.
Christin, P.A, G. Besnard, E. Samaritani, M.R. Duvall, T.R. Hodkinson, V. Savolainen & N.
Salamin. 2008. Oligocene CO2 decline promoted C4 photosynthesis in grasses. Current
Biology. 18, 37–43.
Clauset, A. & D. H. Erwin. 2006. The evolution and distribution of species body mass.
Science. 321, 399-401.
Clauss, M., W. Loehlein, E. Kienzle & H. Wiesner. 2003a. Studies on feed digestibilities in
captive Asian elephants (Elephas maximus). Journal of Animal Physiology and Animal
Nutrition. 87, 160-173.
Clauss, M., R. Frey, B. Kiefer, M. Lechner-Doll, W. Loehlein, C. Polster, G.E. Rössner &
W.J. Streich. 2003b. The maximum attainable body mass of herbivorous mammals:
morphophysiological constraints on foregut, and adaptations of hindgut fermenters.
Oecologia. 136, 14-27.
103
Clauss, M., N. Robert, C. Walzer, C. Vitaud & J. Hummel. 2005. Testing predictions on body
mass and gut contents: dissection of an African elephant Loxodonta africana Blumenbach
1797. European Journal of Wildlife Research. 51, 291-294.
Clauss, M., J.C. Castell, E. Kienzle, E.S. Dierenfeld, E.J. Flach, O. Behlert, S. Ortmann, W.J.
Streich, J. Hummel & J.M. Hatt. 2006. Digestion coefficients achieved by the black
rhinoceros (Diceros bicornis), a large browsing hindgut fermenter. Journal of Animal
Physiology and Animal Nutrition. 90, 325-334.
Clutton-Brock, T.H. & P.H. Harvey. 1983. The functional significance of variation in body
mass among mammals. American Society of Mammalogists. 7, 632-663.
Coria, R.A. & L. Salgado. 2005. Mid-Cretaceous turnover of saurischian dinosaur
communities: evidence from the Neuque´n Basin. Geological Society of London Special
Publication. 252, 317–327.
Cromsigt, J.P.G.M. & H. Olff. 2006. Resource partitioning among savanna grazers mediated
by local heterogeneity: an experimental approach. Ecology. 87(6), 1532-1541.
Cromsigt, J.P.G.M. 2006. Large herbivores in space, resource partitioning among savanna
grazers in a heterogeneous environment. PhD thesis, Community and Conservation
Ecology Group, University of Groningen, Haren, the Netherlands.
Dall, S.R.X., L. Giraldeau, O. Olsson, J.M. McNamara & D.W. Stephens. 2005. Information
and its use by animals in evolutionary ecology. TREE. 20(4), 187-193.B
Darwin, C. 1859. On the origin of species by means of natural selection, or the preservation
of favoured races in the struggle for life. John Murray, London.
De Beer, Y., W. Kilian, W. Versfeld & R.J. van Aarde. 2006. Elephants and low rainfall alter
woody vegetation in Etosha National Park, Namibia. Journal of Arid Environments. 64,
412-421.
De Jong, C.B., S.E. van Wieren, R.M.A. Gill & R. Munro. 2004. Relationships between diet
and liver carcinomas in roe deer in Kielder forest and Galloway forest. Veterinary
Record. 155, 197-200.
Dekker, B., K.P. Kirkman & S.I. du Plessis. 2001. Use of the dry-weight-rank method of
botanical analysis in semi-arid savanna communities. African Journal of Range & Forage
Science. 18, 63-66.
Demment, M.W. & P.J. van Soest. 1985. A nutritional explanation for body mass patterns of
ruminant and nonruminant herbivores. The American Naturalist. 125(5), 641-672.
Demment, M.W. & G.B. Greenwood. 1988. Forage ingestion: effects of sward characteristics
and body mass. Journal of Animal Science 66, 2380-2392.
Dierenfeld, E. S., R. du Toit & W.E. Braselton. 1995. Nutrient composition of selected
browses consumed by black rhinoceros (Diceros bicornis) in the Zambezi Valley,
Zimbabwe. Journal of Zoo and Wildlife Medicine. 26, 220-230.
104
Dierenfeld, E.S., R.E.C. Wildman & S. Romo. 2000. Feed intake, diet utilization and
composition of browses consumed by the Sumatran rhino (Dicerorhinus sumatrensis) in a
North American zoo. Zoo Biology. 19, 169-180.
Drescher, M. 2003. Grasping complex matter: large herbivore foraging in patches of
heterogeneous resources. Ph.D. Thesis, Wageningen University, the Netherlands.
Drescher, M., I.M.A. Heitkönig, J.G. Raats & H.H.T. Prins. 2006. The role of grass stems as
structural foraging deterrents and their effects on the foraging behaviour of cattle. Applied
Animal Behaviour Science. 101, 10-26.
du Toit, P.J., J.G. Louw & A.I. Malan. 1940. A study of the mineral content and feeding value
of natural pastures in the Union of South Africa. Onderstepoort Journal of Veterinary
Science. Anim. Ind. 14, 123-127.
du Toit, J.T., J.P. Bryant & K. Frisby. 1990. Re-growth and palatability of Acacia shoots
following pruning by African savanna browsers. Ecology 71, 149-154.
Economos, A.C. 1981. The largest land mammal. Journal of Theoretical Biology. 89, 211-215.
Ehleringer, J.R., T.E. Cerling & B.R. Helliker. 1997. C4 photosynthesis, atmospheric CO2,
and climate. Oecologia. 112, 285-299.
Ehleringer, J., T.E. Cerling & D. Dearing. 2005. A history of atmospheric CO2 and its effects
on plants, animals and ecosystems, (Ecological studies, Vol.177). Springer.
Elser, J. J., W. F. Fagan, R. F. Denno, D. R. Dobberfuhl, A. Folarin, A. Huberty, S. Interlandi,
S. S. Kilham, E. McCauley, K. L. Schulz, E. H. Siemann, & R. W. Sterneret. 2000.
Nutritional constraints in terrestrial and freshwater food webs. Nature. 408, 578–580.
Enquist, B.J., J.H. Brown & G.B. West. 1998. Allometric scaling of plant energetics and
population density. Nature. 395, 163-165.
Estes, R.D. 1974. Social organization of the African Bovidae. In: V. Geist & F. Walther
(Eds.). The behaviour of ungulates and its relation to management 1 & 11, 166-215.
Estes, R.D. 1991. The behaviour guide to African mammals. The university of California
Press, Berkeley, Los Angeles and London.
Felton, A.M., A. Felton, D. Raubenheimer, S.J. Simpson, W.J. Foley, J.T. Wood, I. R. Wallis
and D.B. Lindenmayer. 2009. Protein content of diets dictates the daily energy intake of a
free-ranging primate. Behavioural Ecology (in press).
Foose, T.J. 1982. Trophic strategies of ruminant versus nonruminant ungulates. Ph.D. Thesis,
University of Chicago, United States.
Fornara, D.A. & J.T. du Toit. 2007. Browsing lawns? Responses of Acacia nigrescens to
ungulate browsing in an African savanna. Ecology. 88(1), 200-209.
Fortelius, M. 1985. Ungulate cheek teeth: developmental, functional, and evolutionary
interrelations. Acta Zoologica Fennica 180, 1-76.
105
Fortelius, M. & J. Kappelman. 1993. The largest land mammal ever imagined. Zoological
Journal of the Linnean Society. 107, 85-101.
Fortin, D. 2006. The allometry of plant spacing that regulates food intake rate in mammalian
herbivores. Ecology. 87(7), 1861-1866.
Frank, D.A., S.J. McNaughton & B.F. Tracy. 1998. The ecology of the earth’s grazing
ecosystems. BioScience. 48(7), 513-521.
Fritz, H., P. Duncan, I. J. Gordon, & A. W. Illius. 2002. Megaherbivores influence
trophic guilds structure in African ungulate communities. Oecologia. 131, 620–625.
Fritz H. & A. Loison. 2006. Large herbivores across biomes. In: K. Danell, R. Bergstrom, P.
Duncan & J. Pastor. (Eds.). Large herbivore ecology, ecosystem dynamics and
conservation. Cambridge University, Cambridge.
Fryxell, J.M. 1991. Forage quality and aggregation by large herbivores. The American
Naturalist. 138 (2), 478-498.
Fryxell, J.M., Wilmshurst, J.F. & Sinclair, A.R.E. 2004. Predictive models of movement by
Serengeti grazers. Ecology, 85(9), 2429-2435.
Gagnon, M. & A.E. Chew. 2000. Dietary preferences in extant African bovidae. Journal of
Mammalogy. 81(2), 490-511.
Ginnett, T.F., J.A. Dankosky, G. Deo & M.W. Demment. 1999. Patch depression in grazers:
the roles of biomass distribution and residual stems. Functional Ecology. 13, 37-44.
Goering, H.K. & P.J. van Soest. 1970. Forage Fiber Analyses (Apparatus, Reagents,
Procedures and Some Applications). Agricultural Handbook 379, United States
Department of Agriculture, Washington DC.
Gordon, I.J. & A.W. Illius. 1988. Incisor arcade structure and diet selection in ruminants.
Functional Ecology. 2(1), 15-22.
Gordon, I.J. & C.E. Lascano. 1993. Foraging strategies of ruminant livestock on intensively
managed grasslands: Potential and constraints. In: M.J. Baker (Ed.). Grasslands for our
world. SIR Publishing. Wellington, New Zealand.
Gordon, I. & H.H.T. Prins (Eds.). 2008. The ecology of browsing and grazing. SpringerVerlag, Berlin.
Gould, S.J. 1970. Evolutionary paleontology and the science of form. Earth-Science Reviews.
6, 77-119. - Printed in The Netherlands
Greyling, M.D. 2004. Sex and age related distinctions in the feeding ecology of the African
elephant, Loxodonta africana. Ph.D Thesis, University of Witwatersrand, Johannesburg,
South Africa.
Griffiths, W.M. 2006. Plant-animal mechanics & bite procurement in grazing ruminants. In:
A. Herrel, T. Speck & N.P. Rowe (Eds.). Ecology and biomechanics: A mechanical
approach to the ecology of animals and plants. CRC Press, Boca Raton.
106
Groen, T.A. 2007. Spatial matters: How spatial processes and patterns affect savanna
dynamics. Ph.D. Thesis, Wageningen University, the Netherlands.
Guy, P.R. 1975. The daily food intake of the African elephant Loxodonta africana
Blumenbach, in Rhodesia. Arnoldia. 26, 1-6.
Hanley, T.A. 1982. The nutritional basis for food selection by ungulates. Journal of Range
Management. 35(2), 146-151.
Hanley, T.A. 1997. A nutritional view of understanding and complexity in the problem of diet
selection by deer (Cervidae). Oikos. 79(2), 209-218.
Haydock, K.P. & N.H. Shaw. 1975. The comparative yield method. Australian Journal of
Experimental Agriculture and Animal Husbandry. 15, 663-670.
Hengeveld, G.M., F. van Langevelde, T.A. Groen & H.J. de Knegt. 2009. Optimal foraging
for multiple resources in several food species. American Naturalist. 174, 102-110.
Hester, A.J., I.J. Gordon, G.J. Baillie & E. Tappin. 1999. Foraging behaviour of sheep and red
deer within natural heather grass mosaics. Journal of Applied Ecology. 36, 133-146.
Heuermann, N. 2007. Tall swards and small grazers: competition, facilitation and
coexistence of different-sized grazers. Ph.D. Thesis, Wageningen University, the
Netherlands.
Hiscocks, K. 1999. The impact of an increasing elephant population on the woody vegetation
in southern Sabi Sand Wildtuin, South Africa. Koedoe. 42(2), 47-55.
Hixon, M.A. 1982. Energy maximizers and time minimizers: theory and reality. The
American Naturalist,. 119(4), 596-599.
Hjeljord, O. 1987. Research on moose nutrition in the Nordic countries. Swedish Wildlife
Research, Suppl.I, 389-397.
Hodgson, J. 1985. The control of herbage intake in the grazing ruminant. Proceedings of the
Nutrition Society. 44, 339-346.
Hofmann, R.R. & D.R.M. Stewart. 1972. Grazer or browser: a classification based on the
stomach structure and feeding habits of East African ruminants. Mammalia. 36, 226-240.
Holdø, R.M., J.P. Dudley & L.R. McDowell. 2002. Geophagy in the African elephant in
relation to availability of dietary sodium. Journal of Mammalogy. 83(3), 652-664.
Holling, C.S. 1959. Some characteristics of simple types of predation and parasitism.
Canadian Entomology. 41, 385-398.
Hongo, A. & M. Akimoto. 2003. The role of incisors in selective grazing by cattle and horses.
Journal of Agricultural Science. 140, 469-477.
Hubbert, M.E. 1987. The effect of diet on energy partitioning in moose. Ph.D. Thesis,
University of Alaska, Fairbanks.
Huggard, D.J. 1994. A linear programming model of herbivore foraging: imprecise, yet
successful? Oecologia. 100, 470-474.
107
Hummel, J., C.T. Gee, K. Südekum, P.M. Sander, G. Nogge & M. Clauss. 2008. In vitro
digestibility of fern and gymnosperm foliage: implications for sauropod feeding ecology
and diet selection. Proceedings of the Royal Society. 275, 1015-1021.
Hutchinson, G.E. 1959. Homage to Santa Rosalia or why are there so many kinds of animals?
The American Naturalist. 93(870), 145-159.
Illius, A.W. & I.J. Gordon. 1987. The allometry of food intake in grazing ruminants. Journal
of Animal Ecology. 56, 989-999.
Illius, A.W. & I.J. Gordon. 1992. Modelling the nutritional ecology of ungulate herbivores:
evolution of body mass and competitive interactions. Oecologia. 89(3), 428-434.
Illius, A. W. & I.J. Gordon. 1994. The functional significance of the browser-grazer
dichotomy in African ruminants. Oecologia. 98, 167-175.
Illius, A.W., I.J. Gordon, D.A. Elston & J.D. Milne. 1999. Diet selection in goats: a test of
intake-rate maximization. Ecology. 80(3), 1008-1018.
Illius, A.W., B.J. Tolkamp & J. Yearsley. 2002. The evolution of the control of food intake.
Proceedings of the Nutrition Society. 61, 456-472.
Iason, G.R. & S.E. van Wieren. 1999. Digestive and ingestive adaptations of mammalian
herbivores to low quality forage. In: H. Olff, V.K. Brown, R. Brent (Eds.). Herbivores:
between plants and predators, Proceedings of the Symposium of the British Ecological
Society. Blackwell Scientific Publications, Oxford.
Jachmann, H., & R. H. V. Bell. 1985. Utilization by elephants of the Brachystegia woodlands
of the Kasungu National Park, Malawi. African Journal of Ecology. 23, 245–258.
Janis, C.M. 1989. A climatic explanation for patterns of evolutionary diversity in ungulate
mammals. Palaeontology. 32, 463-481.
Janis, C.M. 1993. Tertiary mammal evolution in the context of changing climates, vegetation
and tectonic events. Annual Review of Ecology. 24, 467-500.
Janis, C.M., J. Damuth & J.M. Theodor. 2000. Miocene ungulates and terrestrial primary
productivity: where have all the browsers gone? PNAS. 97(14), 7899-7904.
Janis, C.M. 2008. An evolutionary history of browsing and grazing ungulates. In: I.J. Gordon
& H.H.T. Prins (Eds.).The ecology of browsing and grazing. Springer, Berlin Heidelberg.
Jarman, P.J. 1974. The social organisation of antelope in relation to their ecology. Behaviour.
48,215–266.
Jia, J, P. Niemela & K. Danell. 1995. Moose Alces alces bite diameter selection in relation to
twig quality on four phenotypes of Scots pine Pinus sylvestris. Wildlife Biology. 1, 47–55.
Jian, Z. & R.J. Hudson. 1993. Optimal grazing of wapiti (Cervus elaphus) on grassland: patch
and feeding station departure rules. Evolutionary Ecology. 7, 488-498.
Jones, C.G., J.H. Lawton & M. Shachak. 1994. Organisms as ecosystem engineers. Oikos.
69(3), 373-386.
108
Kingdon, J. 1997. The Kingdon Field Guide to African Mammals. Academic Press, London
and New York.
Kirkham, F.W. & R.J. Wilkins. 1994. The productivity and response to inorganic fertilizers of
species-rich wetland hay meadows on the Somerset Moors: the effect of nitrogen,
phosphorus and potassium on herbage production. Grass & Forage Science. 49, 163-175.
Kleiber, M. 1932. Body mass and metabolism. Hilgardia. 6, 315-349.
Kotliar, N.B. & J.A. Wiens. 1990. Multiple scales of patchiness and patch structure: a
hierarchical framework for the study of heterogeneity. Oikos. 59(2), 253-260.
Laca, E.A., E.D. Ungar & M.W. Demment. 1994. Mechanisms of handling time and intake
rate of a large mammalian grazer. Applied Animal Behaviour Science. 39, 3-19.
Lawes, M.J. & C.A. Chapman. 2006. Does the herb Acanthus pubescens and/or elephants
suppress tree regeneration in disturbed Afrotropical forest? Forest Ecology and
Management. 221, 278-284.
Lawley, L. 1994. The World of Elephants. Friedman/Fairfax Publishers, New York, United
States of America.
Lewis, D.M. 1991. Observations of tree growth, woodland structure and elephant damage on
Colophospermum mopane in Luangwa Valley, Zambia. African Journal of Ecology. 29,
207-221.
Leyland, R.C. & K.T. Witthüser. 2008. Regional description of the groundwater chemistry of
the Kruger National Park. Report to the Water Research Commission by Department of
Geology, University of Pretoria, South Africa.
Lindstedt, S.L. & P.J. Schaeffer. 2002. Use of allometry in predicting anatomical and
physiological parameters of mammals. Laboratory animals 36, 1-19.
Logan, K.J. & B.A. Thomas. 1987. The distribution of lignin derivatives in fossil plants. New
Phytologist. 105(1), 157-173.
Loveridge, J.P. & S.R. Moe. 2004. Termitaria as browsing hotspots for African
megaherbivores in miombo woodland. Journal of Tropical Ecology. 20, 337-343.
Low, A.B. & A.G. Rebelo (Eds.). 1996. Vegetation of South Africa, Lesotho and Swaziland.
Department of Environmental Affairs and Tourism, Pretoria, South Africa.
Lucas, S. G. & J.C. Sobus.1989. The systematics of Indricotheres. In: D.R. Prothero. & R.M.
Schoch. (Eds.). The evolution of Perissodactyls. Oxford University Press, Oxford,
England.
Ludwig, F., H. de Kroon, H.H.T. Prins & F. Berendse. 2001. Effects of nutrients and shade on
tree-grass interactions in an East African savanna. Journal of Vegetation Science. 12(4),
579-588.
Ludwig, F., H. de Kroon & H.H.T. Prins. 2008. Impacts of savanna trees on forage quality for
a large African herbivore. Oecologia. 155, 487-496.
109
MacArthur, R.H. & E.R. Pianka. 1966. On optimal use of a patchy environment. The
American Naturalist. 100(916), 603-609.
Makhabu, S.W. & C. Skarpe. 2006. Rebrowsing by elephants three years after simulated
browsing on five woody plant species in northern Botswana. South African Journal of
Wildlife Research. 36(11), 99-102.
Mathiesen, S.D., W. SØrmo, Ø.E. Haga, H.J. Norberg, T.H.A. Utsi & N.J.C. Tyler. 2000. The
oral anatomy of arctic ruminants: coping with seasonal change. Journal of the Zoological
Society of London. 251, 119-128.
Mattson, W.J. 1980. Herbivory in relation to plant nitrogen content. Annual Review of
Ecology and Systematics.11, 119-161.
McArthur, C., A.E. Hagerman & C.T. Robbins. 1991. Physiological strategies of mammalian
herbivores against plant defenses. In: R.T. Palo & C.T. Robbins. (Eds.). Plant defenses
against mammalian herbivory. CRC Press, Boca Raton, FL.
McDonald, P., R.A. Edwards, J.F.D. Greenhalgh & C.A. Morgan. 1995. Animal nutrition.
Edition 5. Longman Group Limited, Harlow, United Kingdom.
McElwain, J.C., K.J. Willis & R. Lupia. 2005. Cretaceous CO2 decline and the radiation and
diversification of angiosperms. In: J. Ehleringer, T.E. Cerling & D. Dearing (Eds.). A
history of atmospheric CO2 and its effects on plants, animals and ecosystems, (Ecological
studies, Vol.177). Springer.
McMahon, T. 1973. Size and shape in biology. Science. 179(4079), 1201-1204.
McNaughton, S.J. 1984. Grazing lawns: animals in herds, plant form, and coevolution.
American Naturalist. 124, 863.
McNaughton, S.J. & N.J. Georgiadis. 1986. Ecology of African grazing and browsing
mammals. Annual Review of Ecology, Evolution, and Systematics. 17, 39-65.
Meissner, H.H. 1982. Theory and application of a method to calculate intake of wild southern
African ungulates for purposes of estimating carrying capacity. South African Journal of
Wildlife Research. 12(2), 41-47.
Meissner, H.H., E.R. Spreeth, P.A. De Villiers, E.W. Pietersen, T.A. Hugo & B.F. Terblanch.
1990. Quality of food and ad libitum intake by elephant as measured by lignin index.
South African Journal of Wildlife Research. 20(3), 104-110.
Meserve, P.L., D.A. Kelt, W. B. Milstead & J.R. Gutierrez. 2003. Thirteen years of shifting
top-down and bottom-up control. BioScience. 53(7), 633-646.
Meyer, H. and M. Coenen. 2002. Pferdefütterung. Parey Verlag, Berlin.
Mihlbachler, M.C. 2008. Species taxonomy, phylogeny, and biogeography of the
Brontotheriidae (Mammalia: Perissodactyla). Bulletin of the American Museum of
Natural History. 311, 1-475.
110
Moe, S.R. & P. Wegge. 2008. Effects of deposition of deer dung on nutrient redistribution
and on soil and plant nutrients on intensively grazed grasslands in lowland Nepal.
Ecological Research. 23, 227-234.
Molvar, E.M. & R.T. Bowyer. 1994. Costs and benefits of group living in a recently social
ungulate: the Alaskan moose. Journal of Mammalogy. 75(3), 621-630.
Mouissie, A.M., R.V. Diggelen, M.E.F. Apol, & G.W. Heil. 2004. Stability and dimensions of
grazing-induced vegetation patterns. In: Seed dispersal by large herbivores: implications
for the restoration of plant biodiversity, edited by University of Groningen, Groningen.
Mueller-Dombois, D. & H. Ellenberg. 1974. Aims and Methods of Vegetation Ecology. John
Wiley and Sons, Inc., New York.
National Research Council. 1989. Nutrient Requirements of Horses, 5th Rev. Ed. National
Academy Press, Washington, DC.
Niklas, K.J. 2004. Plant allometry: is there a grand unifying theory? Biological Reviews. 79,
871-889.
Novellie, P.A. 1978. Comparison of the foraging strategies of blesbok and springbok on the
Transvalal highveld. South African Journal of Wildlife Research. 8, 137–144.
Novozamsky I, V.J.G. Houba, R. Vaneck, & W. Vanvark. 1983. A novel digestion technique
for multi-element plant analysis. Communications in Soil Science and Plant Analysis. 14,
239-248
Nowak, R. M. 1991. Walker's Mammals of the World. Fifth Edition. The Johns Hopkins
University Press, Baltimore.
O’Connor, T.G., P.S. Goodman & B.Clegg. 2007. A functional hypothesis of the threat of
local extirpation of woody plant species by elephant in Africa. Biological Conservation.
136, 329-345.
Olff, H., M.E. Ritchie & H.H.T. Prins. 2002. Global environmental controls of diversity in
large herbivores. Nature 415, 901-904.
Owen-Smith, N. 1985. Niche separation among African ungulates. In: E.S. Vrba (Ed.).
Species and speciation. Transvaal Museum Monograph 4. Transvaal Museum, Pretoria.
Owen-Smith, N. 1988. Megaherbivores: the influence of very large body mass on ecology.
Cambridge University Press, Cambridge
Owen-Smith, N. 1994. Foraging responses of kudus to seasonal changes in food resources:
elasticity in constraints. Ecology. 75(4), 1050-1062.
Palmer, D. (Ed.). 1999. The Marshall Illustrated Encyclopedia of Dinosaurs and Prehistoric
Animals. London: Marshall Editions.
Palo, R.T., R. Bergstrom & K. Danell. 1992. Digestibility distribution of phenols and fiber at
different twig diameters of birch in winter: implications for browsers. Oikos. 65, 450-454.
111
Parker, K.L., P.S. Barboza & M.P. Gillingham. 2009. Nutrition integrates environmental
responses of ungulates. Functional Ecology. 23, 57-69.
Pays, O., S. Benhamou, R. Helder & J. Gerard. 2007. The dynamics of group formation in
large mammalian herbivores: an analysis in the European roe deer. Animal Behaviour. 74,
1429-1441.
Pearson, P. N. & M. R. Palmer. 2000. Atmospheric carbon dioxide concentrations over the
past 60 million years. Nature. 406, 695–699.
Peel, M.J.S., J.M. Kruger & S. MacFadyen. 2007. Woody vegetation of a mosaic of protected
areas adjacent to the Kruger National Park, South Africa. Journal of Vegetation Science.
18, 807-814.
Pendlebury, C., N.E. Odongo, A. Renjifo, J. Naelitz, E.V. Valdes & B.W. McBride. 2005.
Acid-insoluble ash as a measure of dry matter digestibility in captive African elephants
(Loxodonta Africana). Zoo Biology. 24, 261-265.
Pérez-Barbería, F.J., D.M. Walker & G. Marion. 2007. Maximizing intake under challenging
foraging conditions at two spatial scales in Soya sheep. Animal Behaviour. 73, 339-348.
Peters, R.H. 1983. The ecological implications of body mass. Cambridge, England,
Cambridge University Press.
Pietersen, L.M., H.H. Meissner & E.W. Pietersen. 1993. Food selection and intake by male
impalas in the Timbavati area. Suid-Afrikaanse Tydskrif Natuurnavorsing. 23(1), 6-11.
Power, M.E. 1992. Top-down and bottom-up forces in food webs: do plants have primacy?
Ecology. 73(3), 733-746.
Pringle, J. 1982. The conservationists and the killers. T.V. Bulpin & Books of Africa, Cape
Town.
Prins, H.H.T., R.C. Ydenberg & R.H. Drent. 1980. The interaction of Brent geese Branta
bernicla and sea plantain Plantago maritime during spring staging: field observations and
experiments. Acta Botanica Neerlandica. 29,585-596.
Prins, H.H.T. 1987. The buffalo of Manyara: the individual in the context of herd life in a
seasonal environment of East Africa. Ph.D. Thesis, University of Groningen, Groningen,
the Netherlands.
Prins, H.H.T. 1989. Buffalo herd structure and its repercussions for individual African buffalo
cows. Ethology, 81, 47-71.
Prins, H.H.T. 1996. Ecology and behaviour of the African buffalo: social inequality and
decision making. Wildlife Ecology and Behaviour series 1. Chapman & Hall, London.
Prins, H.H.T. 1998. Origins and development of grassland communities in northwestern
Europe. In: M.F. Wallis DeVries, J.P. Bakker & S.E. van Wieren. (Eds.). Grazing and
conservation management. Kluwer, Boston.
112
Prins, H.H.T. & H. Olff. 1998. Species-richness of African grazer assemblages: towards a
functional explanation. In: D.M. Newberry, H.H.T. Prins & N.D. Brown. (Eds.).
Dynamics of tropical communities. The 37th Symposium of the British Ecological Society.
London. Blackwell Science.
Prins, H.H.T. & H. Fritz. 2008. Species diversity of browsing and grazing ungulates:
consequences for the structure and abundance of secondary production. In: I. Gordon &
H.H.T. Prins. (Eds.). The Ecology of Browsing and Grazing. Springer-Verlag, Berlin.
Prins, H.H.T. & I.J. Gordon. 2008. Introduction: Grazers and browsers in a changing world.
In: I. Gordon & H.H.T. Prins (Eds.). The Ecology of Browsing and Grazing. SpringerVerlag, Berlin.
Prins, H.H.T. & F. van Langevelde (Eds.). 2008. Resource ecology: spatial and temporal
dynamics of foraging. Springer, the Netherlands.
Prins, H.H.T. & F. van Langevelde. 2008b. Assembling a diet from different places. In:
H.H.T. Prins & van Langevelde (Eds.). Resource ecology: spatial and temporal dynamics
of foraging. Springer, the Netherlands.
Provenza, F.D. 1995. Postingestive feedback as an elementary determinant of
food selection and intake in ruminants. Journal of Range Management. 48, 2–17.
Raubenheimer, D. & S.J. Simpson. 1997. Integrative models of nutrient balancing:
application to insects and vertebrates. Nutrition Research Reviews. 10, 151-179.
Raubenheimer, D. & S.J. Simpson. 1999. Integrating nutrition: a geometrical approach.
Entomologia Experimentalis et Applicata. 91, 67-82.
Raubenheimer, D., S.J. Simpson & D. Mayntz. 2007. Nutrition, ecology and nutritional
ecology: toward an integrated framework. Functional Ecology. 23, 4-16.
Rees, P.A. 1982. Gross assimilation efficiency and food passage time in the African elephant.
African Journal of Ecology. 20(3), 193-198.
Ritchie, M.E. & H. Olff. 1999. Spatial scaling laws yield a synthetic theory of biodiversity.
Nature. 400, 557-560.
Ritchie, M. E. & H. Olff. 2004. Resource partitioning and biodiversity in fractal
environments, with application to dryland communities. In: M. Shachak, J. R. Gosz, A.
Perevolotsky & S. T. A. Pickett. (Eds.). Biodiversity in drylands: towards a unified
framework. Oxford University Press, Oxford, UK.
Robbins, C.T. 1987. Role of tannins in defending plants against ruminants: reduction in dry
matter digestion? Ecology. 68(6), 1606-1615.
Robbins, C.T. 1993. Wildlife feeding and nutrition.2nd Ed. Academic Press, New York, NY.
Rooke, T., R. Bergström, C. Skarpe & K. Danell. 2004. Morphological responses of woody
species to simulated twig-browsing in Botswana. Journal of Tropical Ecology. 20, 281289.
113
Root, T.L., J.T. Price, K.R. Hall, S.H. Schneider, C. Rosenzweig & J.A. Pounds. 2003.
Fingerprints of global warming on wild animals and plants. Nature. 421, 57-60.
Rucker, R.B. 2007. Allometric scaling, metabolic size and interspecies comparisons of basal
nutritional requirements. Journal of Animal Physiology and Animal Nutrition. 91(3), 148156.
Rutherford, M.C. & R.H. Westfall. 1986. Biomes of Southern Africa: an objective
categorization. Memoirs of the Botanical Survey of South Africa. 54, 1-98.
Sage, R.F. 2004. The evolution of C4 photosynthesis. New Phytologist. 161, 341-370.
Sankaran, M., J. Ratnam & N. Hanan. 2008. Woody cover in African savannas: the role of
resources, fire and herbivory. Global ecology and biogeography. 17, 236-245.
Schmidt-Nielsen, K. 1984. Scaling: Why is animal body mass so important? Cambridge
University Press, Cambridge.
Schoener, T.W. 1969. Optimal size and specialization in constant and fluctuating
environments; an energy-time approach. Brookhaven Symposium of Biology. 22, 103-114.
Scholes, R.J. 1990. The influence of soil fertility on the ecology of southern African dry
savannas. Journal of Biogeography. 17, 415-419.
Scholes, R.J. & K.G. Mennell. 2008. Elephant management: a scientific assessment for South
Africa. Wits University Press, Johannesburg, South Africa.
Seagle, S.W. & S.J. McNaughton. 1992. Spatial variation in forage nutrient concentrations
and the distribution of Serengeti grazing ungulates. Landscape Ecology. 7(4), 229-241.
Searle, K.R., N.T. Hobbs & L.A. Shipley. 2005. Should I stay or should I go? Patch departure
decisions by herbivores at multiple scales. Oikos. 111(3), 414–424.
Searle, K.R., & L.A. Shipley. 2008. The comparative feeding behaviour of large browsing
and grazing herbivores. In: I.J. Gordon & H.H.T. Prins (Eds.). The ecology of browsing
and grazing. Springer Series, Ecological Studies.
Senft, R.L., M.B. Coughenour, D.W. Bailey, L.R. Rittenhouse, O.E. Sala & D.M. Switft.
1987. Large Herbivore Foraging and Ecological Hierarchies. BioScience. 37(11), 789-799.
Servello, F. 1984. Pine vole diet quality in relation to apple tree root damage. Journal of
Wildlife Management. 48(2), 450-455.
Seymour, R.S. 2009. Raising the sauropod neck: it costs more to get less. Biology Letters. 5,
317-319.
Shipley, L. A., J.E. Gross, D.E. Spalinger, N.T. Hobbs & B.A. Wunder. 1994. The scaling of
intake rate in mammalian herbivores. The American Naturalist. 143(6), 1055-1082.
Shipley, L.A. & D.E. Spalinger. 1995. Influence of size and density of browse patches on
intake rates and foraging decisions of young moose and white-tailed deer. Oecologia. 104,
112-121.
114
Shipley, L.A., A.W. Illius, K. Danell, N.T. Hobbs & D.E. Spalinger. 1999. Predicting bite
size selection of mammalian herbivores: A test of a general model of diet optimization
Oikos. 84(1), 55-68.
Shipley, L. A. 2007. The influence of bite size on foraging at larger spatial and temporal
scales by mammalian herbivores. Oikos. 119, 1964-1974.
Shoshani, J. 1998. Understanding proboscidean evolution: a formidable task. TREE. 13(12),
480-487.
Shrader, A.M., N. Owen-Smith & J.O. Ogutu. 2006. How a mega-grazer copes with the dry
season: food and nutrient intake rates by white rhinoceros in the wild. Functional Ecology.
20, 376-384.
Sikes, S.K. 1971. The natural history of the African elephant. Weidenfeld and Nicolson,
London, UK.
Simpson, S.J. & D. Raubenheimer. 1993. A multi-level analysis of feeding behaviour: the
geometry of nutritional decisions. Philosophical Transactions of the Royal Society. 342,
381-402.
Skarpe, C., P.A. Aarrestad, H.P. Andreassen, S.S. Dhillion, T. Dimakatso, J.T. du Toit, D.J.
Halley, H. Hytteborn, S. Makhabu, M. Mari, W. Marokane, G. Masunga, D. Modise, S.R.
Moe, R. Mojaphoko, D. Mosugelo, S. Motsumi, G. Neo-Mahupeleng, M. Ramotadima, L.
Rutina, L. Sechele, T.B. Sejoe, S. Stokke, J.E. Swenson, C. Taolo, M. Vandewalle & P.
Wegge. 2004. The return of the giants: ecological effects of an increasing elephant
population. Ambio. 33, 276–282.
Skinner, J.D. & H.N. Smithers. 1990. The mammals of the southern African sub-region.
University of Pretoria, South Africa.
Smallegange, I.M., J. van der Meer & R.H.J.M. Kurvers. 2006. Disentangling interference
competition from exploitative competition in a crab bivalve system using a novel
experimental approach. Oikos. 113, 157-167.
Smallie, J.J. & T.G. O’Connor. 2000. Elephant utilization of Colophospermum mopane:
possible benefits of hedging. African Journal of Ecology. 38, 352-359.
Smit, G.N. 1996. Becvol: Biomass Estimates from Canopy/Volume (Version 2) Users Guide
(Unpublished manual). University of Orange Free State, Bloemfontein, South Africa.
Snyman, H.A. 2002. Short-term response of rangeland botanical composition and productivity
to fertilization (N and P) in a semi-arid climate of South Africa. Journal of Arid
Environments. 50, 167-183.
Solomon, M.E. 1949. The natural control of vertebrate populations. Journal of Animal
Ecology. 18, 1-35.
Spalinger, D.E, T.A. Hanley & C.T. Robbins. 1988. Analysis of the functional response in
foraging in the Sitka black-tailed deer. Ecology. 69, 1166-1175.
115
Spalinger, D.E. & N.T. Hobbs. 1992. Mechanisms of Foraging in Mammalian Herbivores:
New Models of Functional Response. The American Naturalist. 140(2), 325-348.
Spedding, C.R.W., R.V. Large & D.D. Kydd.. 1966. The evaluation of herbage species by
grazing animals. Proceedings of the 10th International Grassland Congress, Helsinki, p.
479-483.
Stephens, D. W., & J.R. Krebs. 1986. Foraging theory. Princeton University Press, USA.
Stevenson, M.F. & O. Walter. 2006. Management guidelines for the welfare of zoo animals.
Elephants, Loxodonta Africana and Elephas maximus. British and Irish Association of
zoos and aquariums.
Stewart, D.R.M. 1967. Analysis of plant epidermis in faeces: A technique for studying the
food preference of grazing herbivores. The Journal of Applied Ecology. 4(1), 83-111.
Stobbs, T.H. 1973. The effect of plant structure on the intake of tropical pastures. I. Variation
in the bite size of grazing cattle. Australian Journal of Agricultural Research. 24, 809819.
Stokke, S. 1999. Sex differences in feeding patch choice in a megaherbivore: elephants in
Chobe National Park, Botswana. Canadian Journal of Zoology, 77, 1723-1732.
Strohbach, B. 2000. Vegetation degradation trends in the northern Oshikoto Region: 11. The
Colophospermum mopane shrublands. Dinteria. 26, 63-75.
Strömberg, C.A.E. 2006. Evolution of hypsodonty in equids: testing a hypothesis of
adaptation. Paleobiology. 32(2), 236-258.
Styles, C.V. & J.D. Skinner. 1997. Seasonal variations in the quality of mopane leaves as a
source of browse for mammalian herbivores. African Journal of Ecology. 35(3), 254.
Styles, C.V. & J.D. Skinner. 2000. The influence of large mammalian herbivores on growth
form and utilization of mopane trees, Colosphermum mopane, in Botswana’s northern
Tuli game reserve. African Journal of Ecology. 38, 95-101.
Sullivan, T.P., D.S. Sullivan, D.G. Reid & M.C. Leung. 2004. Weasels, voles and trees:
influence of mustelid semiochemicals on vole populations and feeding damage.
Ecological Applications 14(4), 999-1015.
Tomat, L., B. Schumann, J.L. Atkinson and E.V. Valdes. 1999. Digestibility studies with
captive African elephants (Loxodonta africana). First European Zoo Nutrition Meeting.
Rotterdam Zoo Veterinary Faculty of Utrecht, Rotterdam, the Netherlands.
Treydte, A.C., I.M.A. Heitkönig & F. Ludwig. 2009. Modelling ungulate dependence on
higher quality forage under large trees in African savannahs. Basic and applied ecology.
10(2), 161-169.
Ullrey, D. E., P.T. Robinson & P.A. Whetter. 1979. Comparative digestibility studies with
zoo herbivores. In: The American Association of Zoo Veterinarians Annual Proceedings,
Denver.
116
Ullrey, D.E., S.D. Crissey & H.F. Hintz. 1997. Nutrition advisory group handbook. Elephants:
Nutrition and dietary husbandry.
Ungar, E.D., A. Genizi & M.W. Detriment. 1991. Bite dimensions and herbage intake by
cattle grazing short hand-constructed swards. Agronomy Journal. 83, 973-978.
Ungar, E.D., N. Ravid & I. Bruckental. 2001. Bite dimensions for cattle grazing herbage at
low levels of depletion. Grass and Forage Science. 56, 35-45.
Van de Koppel, J. & H.H.T. Prins. 1998. The importance of herbivore interactions for the
dynamics of African savanna woodlands: an hypothesis. Journal of Tropical Ecology. 14,
565-576.
Van der Waal, C., H. de Kroon, W.F. de Boer, I.M.A. Heitkönig, A.K. Skidmore, H.J. de
Knegt, F. van Langevelde, S.E. van Wieren, R.C. Grant, B.R. Page, R. Slotow, E.M. Kohi,
E. Mwakiwa & H.H.T. Prins. 2009. Water and nutrients alter herbaceous competitive
effects on tree seedlings in a semi-arid savanna. Journal of Ecology. 97, 430-439.
Van der Wal, R., N. Madan, S. van Lieshout, C. Dormann, R. Langvatn & S.D. Albon. 2000.
Trading forage quality for quantity? Plant phenology and patch choice by Svalbard
reindeer. Oecologia. 123, 108-115.
Van Dijk, J.G.B & Duijns, S. 2008. Effect of interference competition on the foraging
behaviour of mallards (Anas platyrhynchos). Masters thesis, Wageningen University, the
Netherlands.
Van Langevelde, F., M. Drescher, I.M.A. Heitkönig & H.H.T. Prins. 2008. Instantaneous
intake rate of herbivores as function of forage quality and mass: Effects on facilitative and
competitive interactions. Ecological Modelling. 213, 273-284.
Van Langevelde, F.M. & H.H.T. Prins. 2008. Introduction on resource ecology. In: H.H.T.
Prins & F.M. van Langevelde. (Eds.). Resource ecology: spatial and temporal dynamics
of foraging. Springer, the Netherlands.
Van Soest, P.J. 1981. Impact of feeding behaviour and digestive capacity on nutritional
response. In: Animal genetic resources and management. Proceedings of the fao/unep
technical consultation. Food and agriculture organistion of the united nations. Rome.
Van Soest, P.J. 1994. Nutritional ecology of the ruminant, 2nd edition. Cornell University
Press, Ithaca, New York.
Van Soest, P.J. 1996. Allometry and ecology of feeding behavior and digestive capacity in
herbivores: A Review. Zoo Biology. 15, 455-479.
Van Wieren, S.E. 1996. Do large herbivores select a diet that maximizes short-term energy
intake rate? Forest Ecology and Management. 88, 149-156.
Venter, F.J., R.J. Scholes & H.C. Eckhardt. 2003. The abiotic template and its associated
vegetation pattern. In: J. du Toit, K. Rogers, and H. Biggs (Eds.). The Kruger Experience.
Island Press, Washington DC.
117
Verheyden, H., P. Ballon, V. Bernard & C. Saint-Andrieux. 2006. Variations in bark-stripping
by red deer Cervus elaphus across Europe. Mammal Review. 36(3), 217-234.
Verweij, R.J.T., J. Verrelst, P.E. Loth, I.M.A. Heitkönig & A.M.H. Brunstig. 2006. Grazing
lawns contribute to the subsistence of mesoherbivores on dystrophic savannas. Oikos. 114,
108-116.
Vesey-FitzGerald, D.F. 1960. Grazing succession among East African game animals. Journal
of Mammalogy. 41(2), 161-172.
Voeten, M.M. & H.H.T. Prins. 1999. Resource partitioning between sympatric wild and
domestic herbivores in the Tarangire region of Tanzania. Oecologia. 120, 287-294.
Waldram, M.S., W.J. Bond & W.D. Stock. 2008. Ecological engineering by a mega-grazer:
white rhino impacts on a South African savanna. Ecosystems. 11, 101-112.
Walker, B.H. 1976. An approach to the monitoring of changes in the composition and
utilization of woodland and savanna vegetation. South African Journal of Wildlife
Research. 6, 1-32.
Wallace, L.L., M.G. Turne, W.H. Romme, R.V. O’Neill & Y. Wu. 1995. Scale of
heterogeneity of forage production and winter foraging by elk and bison. Landscape
Ecology. 10(2), 75-83.
WallisDeVries, M.F., E.A. Laca & M.W. Demment. 1998. From feeding station to patch:
scaling up food intake measurements in grazing cattle. Applied Animal Behaviour Science.
60, 301-315.
WallisDeVries, M.F., E.A. Laca & M.W. Demment. 1999. The importance of scale of
patchiness for selectivity in grazing herbivores. Oecologia. 121, 355-363.
Waterman, P. G., & S. Mole. 1994. Analysis of phenolic plant metabolites. Blackwell
Scientific, Oxford, UK.
Weir, J.S. 1969. Chemical properties and occurrence on Kalahari sand of salt licks created by
elephants. Journal of Zoology. 158, 293-310.
Weiss, W.P. 1993. Predicting energy values of feeds. Journal of Dairy Science. 76(6), 18021811.
West, G.B., J.H. Brown & B.J. Enquist. 1997. A general model for the origin of allometric
scaling laws in biology. Science. 276, 122-126.
West, G.B. & J.H. Brown. 2004. Life’s universal scaling laws. Physics Today p. 36-42.
West, G.B. & J.H. Brown. 2005. The origin of allometric scaling laws in biology from
genomes to ecosystems: towards a quantitative unifying theory of biological structure and
organization. The Journal of Experimental Biology. 208, 1575-1592.
Westoby, M. 1974. An analysis of diet selection by large generalist herbivores. The American
Naturalist. 108(961), 290-304.
118
White, T.C.R. 1978. The importance of a relative shortage of food in animal ecology.
Oecologia. 33, 71-86.
White, G.C. 1996. Analysis of frequency count data using the binomial distribution. Ecology.
77(8), 2549-2557.
White, C.R. & R.S. Seymor. 2004. Does basal metabolic rate contain a useful signal?
Mammalian BMR allometry and correlations with a selection of physiological ecological,
and life-history variables. Physiological and biochemical zoology. 77(6), 929-941.
Whitney-Smith, E. 2008. The evolution of an ecosystem: Pleistocene extinctions. In: A.A.
Minai & Y. Bar-Yam (Eds.). Unifying themes in complex systems IV, Proceedings of the
fourth international conference on complex systems. Springer, Berlin Heidelberg.
Wiens, J.A. 1989. Spatial scaling in ecology. Functional Ecology. 3(4), 385-397.
Williamson, B.R. 1975. The condition and nutrition of elephant in Wankie National Park.
Arnoldia. 7, 43831.
Wilmhurst, J.F., J.M. Fryxell & R.J. Hudson. 1995. Forage quality and patch choice by wapiti
(Cervus elaphus). Behavioural Ecology. 6, 209-217.
Wilson, S.L. & G.I.H. Kerley. 2003. Bite diameter selection by thicket browsers: the effect of
body mass and plant morphology on forage intake and quality. Forest Ecology and
Management. 181, 51-65.
Wiseman, R., B.R. Page & T.G. O’Connor. 2004. Woody vegetation change in response to
browsing in Ithala Game Reserve, South Africa. South African Journal of Wildlife
Research. 34(1), 25-37.
Ydenberg, R.C. & H.H.T. Prins. 1981. Spring grazing and the manipulation of food quality by
Barnacle geese. Journal of Applied Ecology. 18, 443-453.
Yearsley, J., B.J. Tolkamp & A.W. Illius. 2001. Theoretical developments in the study and
prediction of food intake. Proceedings of the Nutrition Society. 60, 145-156.
Yokoyama, S., I. Maeji, T. Ueda, M. Ando & E. Shibata. 2001. Impact of bark stripping by
sika deer, Cervus nippon, on subalpine coniferous forests in central Japan. Forest Ecology
and Management. 140, 93-99.
Zambatis, N., P.J.K. Zacharias, C.D. Morris & J.F. Derry. 2006. Re-evaluation of the disc
pasture meter calibration for the Kruger National Park, South Africa. African Journal of
Range and Forage Science. 23(2), 85-97.
119
SUMMARY
Introduction
Large mammalian herbivores utilize about half of the earth’s land surface and contribute
significantly to the survival and well-being of the human race, whether it is through
consumption or admiration. Predicting the distribution of large herbivore species, either for
the purpose of conservation or management, requires an understanding of the mechanisms of
resource use by these species. Historically, many studies used a descriptive approach to
determine herbivore diet and habitat use. Subsequently, with the development of optimal
foraging theory around the 1970s, the field of foraging ecology became more theoretical and
researchers started using experimentation to determine the mechanisms behind the individual
components of foraging. However, up until recently optimal foraging theory did not address
the issue of scale. Small scale foraging decisions by herbivores have been shown to
significantly alter their large scale distribution patterns. By focusing on the mechanisms
driving small scale foraging by large herbivores, this study attempts to determine how large
herbivores are adapted and able to satisfy their nutritional requirements at scales ranging from
bites to the daily foraging range.
More specifically, the following research questions are addressed:
- Do instantaneous intake rates scale with metabolic requirements and if so how are large
herbivores adapted to achieve these rates?
- Within the daily foraging range, at pre-determined scales, are large herbivores able to
select grains with the highest concentration of nutrients irrespective of the heterogeneity of
patches within grains and if so, are they able to select nutrient rich grains at smaller scales?
- Which of the following animal characteristics best determine the spatial scale at which
herbivores aim to maximize their nutrient intakes: body mass, mouth morphology, herd
size and gastro-intestinal tract type?
- How is a herbivore, of a specific body mass, able to satisfy its daily nutrient requirements
and does a herbivore maximize the intake of specific nutrients beyond satisficing its daily
metabolic requirements?
To answer these questions field experiments were set-up and data collected on wild large
herbivores in the northern parts of the South African savanna.
Adaptations at the bite scale
Allometric scaling relationships of body mass (BM) provide a powerful but relatively simple
tool to explain foraging mechanisms. I hypothesized that bite mass and instantaneous intake
rate (IR) of large herbivores should scale isometrically with body mass, but when digestibility
is incorporated in the calculation of IR, IR should scale to metabolic rate as BM0.75, following
120
Kleiber’s allometry. However, allometric mass-space relationships in plants indicate that
leaf:stem ratios of plants decrease as volume increase and that the average number of plants
per unit space decreases as average plant mass increase. The implications of these
relationships for large herbivores is that if mouth size scales linearly with body mass, megaherbivores will not be able to satisfy their requirements. Therefore, I further hypothesized that
intra-dental mouth volume, determined as the product of incisor width, muzzle width and jaw
length, should scale linearly with body mass and that the volume added by soft mouth parts
such as lips and tongue, should increase total bite volume to scale larger than one with body
mass.
To test the hypothesis, mouth morphological measurements were collected from carcasses
and immobilized large African herbivores covering a body mass range of three orders
magnitude. Bite masses and intake rates were determined from foraging observations on the
same range of species in the wild. Results confirmed the hypotheses and revealed that the
elongated soft mouth parts of mega-herbivores are responsible for that fact that bite volume
scales to body mass, as BM1.7, hence with a constant larger than 1. I conclude by
hypothesizing that past extinctions of mega-herbivores occurred because these herbivores did
not have suitable mouth morphologies to cope with the dilution of food resources in space.
Scaling patch selection
Spatial scale is characterized by grains size, which is the smallest measurement unit used at a
specific scale, and extent, which includes the total area of interest. Patches are resource
locations that are more or less homogenous in their plant species composition, biomass and
quality. More over, patches can vary in size and number within or beyond the specified grain
size of a chosen scale. Using a novel approach, I set-up a large fertilization experiment at a
pre-defined spatial scale of grain size 2500 m2 and an extent of 30 hectares. Within each 2500
m2 plot (grain), I fertilized patches at one of three size (4 m2, 100 m2, 2500 m2) and one of
three local nutrient concentrations (1.2 g N/ m2, 6 g N/ m2, 30 g N/ m2). These combinations
resulted in three total nutrient loads per 2500 m2 plot (0.6 ton N, 3 ton N, 15 ton N). Through
measuring elephant utilization of trees I tested whether elephants could select the plots with
the highest total nutrient loads, irrespective of the number of patches, patch sizes and patch
nutrient concentrations in each plot. Further, I tested whether elephants were able to select
nutrient rich patches at even smaller scales, and whether the type of impact on trees differed
between nutrient rich and poor patches.
My results showed that at a spatial scale with a grain size of 2500 m2, elephant are able to
select grains with the highest total nutrient loads and can even do so at smaller scales. More
over, uprooting of trees occurred more outside nutrient rich patches, whereas leaf utilization
was more common inside these patches. This means that elephant foraging behaviour can be
121
used as indicators of changes in the availability of nutrients. Hence, the findings of this
research shed some light on how future studies should interpret patch selection by herbivores
at different scales.
Determinants of the scale of intake
Body mass is probably the best studied animal characteristic and a reliable universal predictor
of foraging behaviour in herbivores. Other foraging characteristics, which are essentially
functions of body mass, and that have been recognised to significantly influence large
herbivore foraging, include mouth size, herd size and gastro-intestinal tract type. Although the
functions of these characteristics are well understood, how they affect a herbivore’s response
to spatial heterogeneity in forage resource distribution is less clear. I analysed how body mass,
mouth size, digestive system type and herd size affect selection of the spatial scale at which
herbivores select nutrient rich patches to maximize intake. I hypothesized that smaller animals,
animals occurring in large herds, and ruminant species select areas of high nutrient
concentration at small spatial scales, whereas small-mouthed animals select areas with high
nutrient content at larger spatial scales.
Data on spoor and dung counts of large herbivores visiting the large scale fertilization
experiment described before were used to test these hypotheses. I found that large animals
selected areas of high nutrient concentration at larger spatial scales, whereas the abundance of
ruminants decreased as nutrient concentrations at large scales increased. As for individual
species, duiker, warthog and elephant selected for high nutrient concentrations at large scales
while buffalo, zebra and impala did so at the local point scale. From the response of
individual species it seems as though herd living animals maximize nutrient intake at the local
point scale in stead of at larger scales, possible due to intra-specific competition.
Satisfying large herbivore requirements
Classical optimal foraging models only use energy as the currency of maximization. However,
in reality animals have to satisfy requirements for multiple nutrients. Although recent studies
have recognized this fact, some disparity still exists on whether animals aim to only satisfy
daily nutrient requirements or whether the intake of particular nutrients is maximized. Using
linear programming techniques I tested which strategies best explain the daily diets of African
elephant. I used daily requirements of the macro-nutrients nitrogen, phosphorous, potassium,
sodium, calcium, magnesium and metabolizable energy from literature as constraints in the
model, and incorporated availability of each potential forage species in the study area as the
amount of nutrients and metabolizable energy available per plant that could be gained, taken
into account the amount of energy that is required to find each species. Surprisingly, results
from the model show that elephant use different optimization currencies, depending on the
122
time of year and hence the availability of resources. Energy intake is maximized during the
dry season possibly due to changes in water availability and forage abundance in the
environment. During the wet season when food becomes more abundant and energy
requirements are satisfied easier, elephant maximize nitrogen intake for growth and
reproduction.
Conclusion
Results from this study not only re-confirm the effects of the universal scaling of metabolic
rate as BM0.75 on foraging of large herbivores, but also provides new insights for how
especially mega-herbivores are adapted to cope with the heterogeneity in the spatial
distribution of their forage resources. Moreover, from the main theme of this study on the
mechanisms of small scale foraging by large herbivores, new questions for future research
emerged, including:
-
What are the morphological adaptations required by animals in relation to body mass to
cope with the effects of climate change?
-
Is the body mass range of extant large herbivore species, which tend to aggregate together
in large numbers, a reflection of a greater forage resource availability at small spatial
scales compared to herbivores with body masses outside of this range?
I conclude that size, as a characteristic which includes mass and morphology, matters because
it first of all determines whether an herbivore will be able to survive in a particular landscape
and secondly, whether herbivores are able to aggregate together or not which further
determines the spatial scale at which nutrient intake will be maximized.
123
SAMENVATTING – Dutch summary
Introduction
Grote zoogdierlijke herbivoren komen voor op ongeveer de helft van het landoppervlak van
de Aarde en dragen bij aan de overleving en het welzijn van het menselijke ras, zowel door
middel van consumptie als door waardering. Het voorspellen van de verspreiding van grote
herbivoren, ofwel voor natuurbescherming of voor beheer, vereist een goed begrip van de
mechanismen waarmee deze soorten voedselbronnen gebruiken. Historisch gezien hanteren
veel studies een beschrijvende aanpak om het dieet en habitat gebruik van herbivoren te
bepalen. Na de totstandkoming van de optimaal foerageren theorie in the jaren 1970, werd het
vakgebied ecologie meer theoretisch georiënteerd en begonnen onderzoekers met
experimenten om de mechanismen achter de verschillende componenten van foerageren vast
te stellen. Echter, tot voorkort besteedde de optimaal foerageren theorie geen aandacht aan de
kwestie van schaal. Kleinschalige foerageer beslissingen door herbivoren kunnen een enorme
invloed hebben op hun verspreidingspatroon op grote schaal. Door te focussen op de
mechanismen achter het foerageren van herbivoren op en kleine schaal probeert dit onderzoek
vast te stellen hoe grote herbivoren zijn aangepast en in staat zijn om hun nutritionele
behoeften te voldoen op schalen variërend van een hap tot aan hun dagelijks
verspreidingsgebied. Meer specifiek, de volgende onderzoeksvragen komen aan bod:
- Schalen instantane voedselinname snelheden met metabolische behoeften, en zo ja, hoe zijn
grote herbivoren aangepast om deze snelheden te bereiken?
- Binnen het dagelijks verspreidingsgebied, op vooraf vastgestelde schalen, zijn grote
herbivoren in staat om korrels met de hoogste concentratie van nutriënten, ongeacht de
heterogeniteit van patches binnen korrels te selecteren, en zo ja, zijn ze in staat om
nutriëntrijke korrels op kleinere schalen te selecteren?
- Welke van de volgende kenmerken bepalen het beste de ruimtelijke schaal waarop
herbivoren ernaar streven hun inname van nutriënten te maximaliseren: lichaamsgewicht,
mond morfologie, kudde grootte of het type van het maag-darmkanaal?
- Hoe is een herbivoor, met een specifiek lichaamsgewicht, in staat om aan haar dagelijkse
nutritionele behoeften te voldoen en maximaliseert een herbivoor de inname van specifieke
nutriënten meer dan zij nodig heeft om in haar dagelijkse behoefte te voldoen?
Om deze vragen te beantwoorden werden veldexperimenten opgezet en gegevens verzameld
over wilde herbivoren in de noordelijke delen van de Zuid-Afrikaanse savanne.
124
Aanpassingen op schaal van een hap
Allometrische schaalrelaties van lichaamsgewicht (BM) bieden een krachtige maar relatief
eenvoudige mannier om foerageer mechanismen te verklaren. Ik had als hypothese dat hap
grootte en instantane opname snelheid (IR) van grote herbivoren isometrisch zullen schalen
met een lichaamsgewicht, maar als verteerbaarheid wordt opgenomen in de berekening van
IR, zal IR schalen naar stofwisseling als BM0.75, zoals Kleiber's allometrie. Echter,
allometrische gewicht-volume verhoudingen van planten geven aan dat de blad:stengel ratio
van planten afneemt met toenemend volume en dat het gemiddeld aantal planten per eenheid
ruimte daalt als het gemiddelde gewicht van planten hoger wordt. De implicatie van deze
relaties voor grote herbivoren is dat als mond grootte lineair schaalt met een lichaamsgewicht,
mega-herbivoren niet in staat zullen zijn om aan hun behoeften te voldoen. Daarom had ik de
hypothese dat intra-tand mond volume, bepaald als het product van snijtand breedte, mond
breedte en lengte van de kaak, lineair zal schalen met lichaamsgewicht en dat het volume dat
wordt toegevoegd door zachte monddelen zoals de lippen en de ton, het totale hap volume zal
doen toenemen zodat het groter schaalt dan 1 met lichaamsgewicht. Om deze hypothese te
testen, werden mond morfologische metingen gedaan van karkassen en geïmmobiliseerd grote
Afrikaanse herbivoren, over een reeks in lichaamsgewicht die meer dan drie orders van
grootte spant. Hap massa en inname snelheid waren vastgesteld op basis van foerageer
observaties van dezelfde soorten in het wild. De resultaten bevestigen de hypothesen en lieten
zien dat de langwerpige zachte monddelen van mega-herbivoren verantwoordelijk zijn voor
het feit dat hap volume schaalt met lichaamsgewicht, als BM1.7, dus met een constante groter
dan 1. Ik sluit af met de verwachting dat er in het verleden uitsterving van mega-herbivoren
optrad omdat deze herbivoren een ongeschikte mond morfologie hadden om goed om te
kunnen gaan met de verdunning van voedselbronnen in de ruimte.
Het schalen van patch selectie
Ruimtelijke schaal wordt gekenmerkt door korrelgrootte, dat is de kleinste meeteenheid die
gebruikt wordt op een bepaalde schaal, en de uitgestrektheid, namelijk de oppervlakte van het
totale gebied. Patches zijn locaties met resources die min of meer homogeen zijn in hun
planten soortensamenstelling, biomassa en kwaliteit. Bovendien kunnen patches variëren in
grootte en aantal ten opzichte van de korrelgrootte van een gekozen schaal. Met behulp van
een nieuwe benadering heb ik een groot bemestingsexperiment opgestart, met een vooraf
bepaalde ruimtelijke schaal wat betreft korrelgrootte van 2500 m2 en een omvang van 30
hectare. Binnen elk 2500 m2 plot (korrelgrootte), bemestte ik patches met een van de drie
groottes (4 m2, 100 m2, 2500 m2) en een van de drie lokale concentraties nutriënten (1,2 g N /
m2, 6 g N / m2, 30 g N / m2). Deze combinaties resulteerde in drie totale nutriënten
125
hoeveelheden per 2500 m2 plot (0,6 ton N, 3 ton N, 15 ton N). Door het meten van het gebruik
van bomen door olifanten heb ik getest of olifanten de plots met de hoogste totale nutriënten
hoeveelheid zouden kiezen, ongeacht het aantal patches, patch grootte en patch nutriënten
concentraties in elk plot. Verder testte ik of de olifanten in staat waren om nutriëntrijke
patches te selecteren op nog kleinere schalen, en of de aard van de gevolgen voor de bomen
verschilden tussen nutriëntarme en nutriëntrijke patches.
Mijn resultaten toonden aan dat, op een ruimtelijke schaal met een korrelgrootte van 2500 m2,
olifanten in staat zijn om korrels met de hoogste totale nutriënten hoeveelheid te selecteren,
en dat ze dit zelfs kunnen op kleinere schalen. Bovendien kwam ontworteling van bomen
meer voor buiten nutriëntrijke patches, terwijl consumptie van blad meer gebruikelijk was
binnen deze patches. Dit betekent dat het foerageergedrag van olifanten kan worden gebruikt
als indicator voor veranderingen in de beschikbaarheid van voedingsstoffen. De bevindingen
van dit onderzoek werpen dus enig licht op hoe toekomstige studies patch selectie door
herbivoren moeten interpreteren op verschillende schalen.
Determinanten van de schaal van inname
Lichaamsgewicht is waarschijnlijk de best bestudeerde karakteristiek van dieren en een
universele en betrouwbare voorspeller van het foerageergedrag van herbivoren. Andere
foerageer kenmerken, die in wezen functies zijn van het lichaamsgewicht en die een grote
invloed hebben op het foerageren van herbivoren, zijn grootte van de mond, grootte van de
kudde en het type maag-darmkanaal. Hoewel de functies van deze kenmerken al goed worden
begrepen, is de invloed van deze kenmerken op de reactie van een herbivoor op de ruimtelijke
heterogeniteit in de verspreiding van voedsel minder duidelijk. Ik analyseerde daarom hoe
lichaamsgewicht, mond grootte, het type spijsvertering en grootte van de kudde invloed
hebben op de selectie van de ruimtelijke schaal waarop herbivoren nutriëntrijke patches
selecteren om nutriënt inname te maximaliseren. Ik veronderstelde dat kleinere dieren, dieren
in grote kuddes en herkauwers gebieden selecteren met een hoge concentratie voedingsstoffen
op kleine ruimtelijke schalen, terwijl dieren met een kleine mond gebieden selecteren met een
hoog gehalte aan nutriënten op een grotere ruimtelijke schaal.
Gegevens
over
sporen
en
feces
van
grote
herbivoren
in
het
grootschalige
bemestingsexperiment zoals eerder beschreven werden gebruikt om deze hypothesen te testen.
Ik vond dat grote dieren gebieden selecteerden met een hoge concentratie nutriënten op
grotere ruimtelijke schalen, terwijl het aantal herkauwers daalde naarmate de nutriënt
concentratie op grote schaal toenam. Duikers, wrattenzwijnen en olifanten selecteerden hoge
concentraties nutriënten op een grote schaal, terwijl buffels, zebra's en impala’s dit deden op
een lokale schaal. Uit de respons van individuele soorten lijkt het alsof kudde dieren
126
nutriëntinname maximaliseren op de lokale schaal in plaats van op grotere schaal, mogelijk
als gevolg van binnen-soortelijke concurrentie.
Het voldoen aan de behoeften van grote herbivoren
Klassieke optimaal foerageer modellen gebruiken alleen energie als de eenheid van
maximalisatie. Echter, in werkelijkheid moeten dieren voldoen aan de behoefte voor meerdere
nutriënten. Hoewel recente studies dit feit erkennen, bestaat er nog steeds enige ongelijkheid
over de vraag of dieren het doel hebben om alleen aan de dagelijkse behoeften te voldoen dan
wel de inname van bepaalde nutriënten te maximaliseren. Met behulp van lineaire
programmering testte ik welke strategieën het beste het dagelijkse dieet van de Afrikaanse
olifant verklaren. Ik gebruikte de dagelijkse behoeften voor de macro-nutriënten stikstof,
fosfor, kalium, natrium, calcium, magnesium en metaboliseerbare energie uit de literatuur als
beperkingen in het model, en introduceerde de beschikbaarheid van elke plantensoort in het
studiegebied als de hoeveelheid nutriënten en metaboliseerbare energie die beschikbaar is per
plant en die kan worden geconsumeerd, rekening houdend met de hoeveelheid energie die
nodig is om elke soort te vinden. Verrassend genoeg blijkt uit de resultaten van het model dat
olifanten gebruik maken van verschillende optimalisatie eenheden, afhankelijk van de tijd in
het jaar en daarmee de beschikbaarheid van voedselbronnen. Energie-inname wordt
gemaximaliseerd tijdens het droge seizoen, mogelijk als gevolg van veranderingen in de
beschikbaarheid van water en planten in de omgeving. Tijdens het natte seizoen, wanneer
voedsel overvloediger wordt en aan energiebehoeften makkelijker wordt voldaan,
maximaliseren olifanten de inname van stikstof voor groei en voortplanting.
Conclusie
De resultaten van deze studie bevestigen niet alleen opnieuw de effecten van het universeel
schalen van metabolische activiteit als BM0.75 op het foerageergedrag van grote herbivoren,
maar bieden ook nieuwe inzichten voor de wijze waarop vooral mega-herbivoren zijn
aangepast voor het omgaan met heterogeniteit in de ruimtelijke verspreiding van
voedselbronnen. Bovendien, uit het hoofdthema van deze studie over de mechanismen van
kleinschalig foerageren door grote herbivoren zijn nieuwe vragen voor toekomstig onderzoek
naar voren gekomen, waaronder:
- Wat zijn de morfologische aanpassingen van dieren in relatie tot lichaamsgewicht voor het
omgaan met de gevolgen van klimaatsverandering?
- Is de spreiding in het lichaamsgewicht van uitgestorven grote herbivoren, die vaak in grote
aantallen aggregeerden, een weerspiegeling van een grotere beschikbaarheid van
voedselbronnen op kleine ruimtelijke schalen in vergelijking met herbivoren met een ander
lichaamsgewicht?
127
Ik concludeer dat grootte, als een kenmerk dat gewicht en morfologie omvat, van belang is,
omdat het zaken bevat die in de eerste plaats bepalen of een herbivoor in staat zal zijn te
overleven in een bepaald landschap en, in de tweede plaats, of herbivoren al dan niet zullen
aggregeren, wat weer de ruimtelijke schaal waarop nutriëntinname zal worden
gemaximaliseerd beïnvloed.
128
129
SAMEVATTING - Afrikaans summary
Inleiding
Groot soogdier herbivore benut die helfte van die aarde se land oppervlakte en dra grotendeels
by tot die welvaart en oorlewing van die mens, of dit nou is deur te dien as n’bron van voedsel
of deur blote bewondering. Vir navorsing en bestuur vereis die voorspelling van groot
herbivoor verspreiding n’ begrip van die meganismes wat die benutting van hul natuurlike
hulpbronne dryf. Histories het die meeste studies n’ beskrywende benadering gebruik om
herbivore se dieet en habitat benutting te bepaal. Vervolgens, met die ontwikkeling van die
optimale voedingsteorie rondom 1970, het die veld van voedingsekologie meer teoreties begin
raak en het navorsers meer eksperimentering begin gebruik om die meganismes agter die
individuele komponente van voeding te bepaal. Ten spyte hiervan het die optimale voedings
teorie tot onlangs nie die kwessie van skaal aangespreek nie. Kleinskaalse voedingskeuses
deur herbivore is aangetoon om hul grootskaalse verspreidingspatrone merkwaardig te
beinvloed. Deur te fokus op die meganismes wat kleinskaalse voeding van groot herbivore
dryf, probeer die huidige studie om vas te stel hoe groot herbivore aangepas is en hoe dit
moontlik is om hul voedingsbehoeftes te bevredig by skale wat strek vanaf individuele happe
tot die daaglikse voedingsarea.
Meer spesifiek sal die volgende navorsingsvrae aangespreek word:
-
Kan oombliklike tempos van inname voorspel word deur metaboliese behoeftes en indien
wel hoe is groot herbivore aangepas om hierdie tempos te bereik?
-
Binne die daaglikse voedingsruimte en by n’vooraf bepaalde skaal, is groot herbivore
instaat om areas met die hoogste konsentrasies nutriente onafhanklik van die
heterogeniteit van posisies binne daardie areas te selekteer en indien wel is hulle instaat
om dieselfde op kleiner skale te doen?
-
Watter van die volgende dierlike eienskappe bepaal die ruimtelike skaal waarby herbivore
mik om hul nutrient innames te maksimiseer die beste: liggaamsmassa, mond morfologie,
kudde grootte en/ tipe verteringsstelsel?
-
Hoe is n’ herbivoor met n’ spesifieke liggaamsmassa instaat om sy daaglikse nutrient
behoeftes te bevredig en behalwe vir die bevrediging van daaglikse metaboliese behoeftes,
probeer herbivore om inname van ander spesifieke nutriente te maksimiseer?
Om hierdie vrae te beantwoord is veld eksperimente opgestel en data versamel op wilde groot
herbivore in die noordelike dele van die Suid-Afrikaanse savanna.
Aanpassings by die skaal van individuele happe
Allometriese skaal verhoudings van liggaamsmassa (LM) voorsien n’ kragtige maar relatief
eenvoudige manier om voedingsmeganismes te verduidelik. Ek stel deur n’ hipotese voor dat
130
hapmassa en die oombliklike tempo van inname (IT) van groot herbivore isometries moet
skaal met liggaamsmassa maar dat as verteerbaarheid in ag geneem word in die berekening
van IT, IT behoort te skaal met die tempo van metabolisme volgens Kleiber se allometrie
(LM0.75). Aan die ander kant, dui allometriese massa-ruimte verhoudings in plante daarop dat
blaar:stingel verhoudings van plante afneem soos wat volume toeneem en dat die gemiddelde
hoeveelheid plante per eenheid ruimte afneem soos wat gemiddelde plant massa toeneem. Die
gevolge van hierdie verhoudings vir groot herbivore is dat indien mondgrootte liniêr met
liggaamsmassa toeneem, mega-herbivore nie hul behoeftes sal kan bevredig nie. Daarom stel
ek met n’volgende hipotese voor dat intra-dentale mondvolume wat bepaal word deur die
produk van snytand wydte, snoet wydte en kaakbeen lengte, liniêr moet toeneem met
liggaamsmassa en dat die volume wat bygevoeg word met sagte monddele soos lippe en die
tong, die totale hap volume sal laat toeneem sodat dit met n’faktor groter as een tot
liggaamsmassa skaal.
Hierdie hipotesis is getoets deur mond morfologiese metings te versamel vanaf karkasse en
geimmobiliseerde groot Afrika herbivore wat in totaal n’ liggaamsmassa reeks van meer as
drie ordes gedek het. Hap massas en inname tempos is vasgestel vanaf voedingswaarnemings
op dieselfde reeks grootte herbivoor spesies in die veld. Resultate bevestig die hipotesis en
openbaar verder dat die verlengde sagte monddele van mega-herbivore verantwoordelik is vir
die feit dat hap volume skaal tot liggaamsmassa as LM1.7 en dus met n’ faktor groter as een.
Ek kom tot gevolgtrekking met n’ verdere hipotese wat uitspel dat uitsterwings van megaherbivore in die verlede plaasgevind het omrede hierdie diere nie geskikte mond morfologiese
aanpassings gehad het om die verdunning van hul voedingsbronne in ruimte te kon hanteer
nie.
Skaling van posisie seleksie
Ruimtelike skale word gekenmerk deur graan grootte wat die kleinste metingseenheid is wat
gebruik word vir n’ spesifieke skaal en area grootte wat die totale area van toepassing insluit.
Kolle in die konteks van hierdie studie is plekke waar voedingsbronne min of meer homogeen
is in hul plant spesies samestelling, biomassa en kwaliteit. Daarby kan kolle varieer in grootte
en hoeveelheid binne of verby die wydte en breedte van die vooraf bepaalde graan grootte van
die gekose skaal. Met n’ nuwe benadering het ek in hierdie studie n’ groot skaalse
bemestingsprojek gebruik met n’ vooraf bepaalde ruimtelike skaal van graan grootte 2500 m2
en n’ area grootte van 30 hektaar. Binne elke 2500 m2 plot (graan) het ek kolle bemes by een
van drie groottes (4 m2, 100 m2, 2500 m2) en een van drie lokale nutrient konsentrasies (1.2 g
N/ m2, 6 g N/ m2, 30 g N/ m2). Hierdie kombinasies het gelei tot drie totale nutrient vragte per
2500 m2 plot (0.6 ton N, 3 ton N, 15 ton N). Deur die benutting van bome deur olifante te
meet het ek getoets of olifante die plotte kon selekteer met die hoogste totale nutrient vragte,
131
ongeag die hoeveelheid kolle, kol grootte en die nutrient konsentrasie van kolle in elke plot.
Verder het ek getoets of olifante in staat was om die nutrient ryke kolle by selfs kleiner skale
kon selekteer en of die tipe impak op bome verksil tussen nutrient ryk en arm kolle.
My resultate dui daarop dat olifante in staat is om die plotte (grane) met die hoogste totale
nutrient vragte by ruimtelike skale met n’ graan grootte van 2500 m2 of selfs kleiner te
selekteer. Bo en behalwe hiervoor het die ontworteling van bome meer buite nutrient ryke
kolle plaas gevind terwyl die benutting van blare meer binne hierdie kolle geskied het. Dit
beteken dat olifant voedingsgedrag gebruik kan word as n’ indikator van veranderings in die
beskikbaarheid van nutriente. Hiermee bied die vindinge van hierdie navorsing meer
duidelikheid vir toekomstige studies oor hoe om die seleksie van nutrient kolle deur herbivore
by verskillende skale te intrepeteer.
Bepalende faktore vir die skaal van inname
Liggaamsmassa is sekerlik die bes bestudeerde dierlike eienskap en n’ betroubare universele
voorspeller vir die voedingsgedrag van herbivore. Ander voedingseienskappe wat in
werklikheid funksies van liggaamsmassa is en wat erken is om voeding van groot herbivore
merkwaardig te beïnvloed sluit mondgrootte, kudde grootte en spysverteringstelsel tipe in.
Alhoewel die funksies van hierdie eienskappe welom bekend is, is die manier waarop hulle n’
herbivoor se reaksie teenoor die ruimtelike heterogeniteit van voedingsbron verspreiding
beïnvloed
minder
duidelik.
Ek
het
bepaal
hoe
liggaamsmassa,
mondgrootte,
spysverteringstelsel tipe en kudde grootte die ruimtelike skaal waarby herbivore nutrient ryke
kolle kies om hul innames te maksimiseer beïnvloed. Ek het voorgestel dat kleiner diere, diere
wat in groot troppe voorkom en herkouer spesies areas selekteer met hoë nutrient
konsentrasies by klein ruimtelike skale waar andersins diere met klein monde areas selekteer
met hoë nutrient konsentrasies by groter ruimtelike skale.
Data van spoor en mis tellings van groot herbivore wat die groot skaalse
bemestingseksperiment wat hierbo beskryf is besoek het was gebruik om die hipotese te toets.
Ek het gevind dat groot herbivore areas met hoë nutrient konsentrasies by groter ruimtelike
skale selekteer terwyl die hoeveelheid herkouers afgeneem het onder hierdie kondisies. Wat
die individuele spesies betref het duikers, vlakvarke en olifante areas met hoë nutrient
konsentrasies by groot skale geselekteer waarteenoor buffels, zebras en rooibokke nutrient
ryke areas geselekteer het by die lokale punt skaal. Vanuit die reaksie van individuele spesies
lyk dit asof trop lewende diere nutrient inname maksimiseer by die lokale punt skaal moontlik
as gevolg van intra-spesifieke kompetisie.
132
Versadiging van groot herbivoor behoeftes
Klassieke modelle oor die optimale voedingsteorie gebruik slegs energie as die eenheid vir
maksimisering. Nietemin, moet diere in die werklikheid behoeftes vir verskeie nutriente
bevredig. Alhoewel meer huidige studies hierdie feit erken bestaan daar nog onduidelikheid
oor of diere daarop mik om slegs daaglikse behoeftes te bevredig en of die inname van
spesifieke
nutriente
gemaksimiseer
word.
Deur
gebruik
te
maak
van
liniere
programmeringstegnieke het ek getoets watter strategieë die daaglikse dieet van die Afrika
olifant die beste verduidelik. Ek het die daaglikse behoeftes vir die makro nutriente stikstof,
fosfor, kalium, natrium, kalsium, magnesium en metaboliseerbare energie vanuit literatuur
geneem en gebruik as beperkings in die model en het verder die beskikbaarheid van elke
potensiële voer spesie in die studie area geinkorporeer as die hoeveelheid nutriente en
metaboliseerbare energie wat gewin kon word per plant met inagname van die hoeveelheid
energie wat benodig word om elke plantspesie te vind. Onverwags het die resultate van die
model daarop gedui dat olifante verskillende optimiseringseenhede gebruik afhangende van
die tyd van die jaar en dus die beskikbaarheid van voedselbronne. Energie inname word
gemaksimiseer gedurende die droë seisoen moontlik as gevolg van veranderinge in die
beskikbaarheid van water en voer in die omgewing. Gedurende die nat seisoen wanneer voer
meer beskikbaar raak en energie behoeftes makliker versadig kan word maksimiseer olifante
hul stikstof innames om te groei en te reproduseer.
Gevolgtrekking
Resultate van hierdie studie herbevestig nie net die uitwerking van die universele skaling van
metaboliese tempo as LM0.75 op die voeding van groot herbivore het nie, maar verskaf ook
nuwe insig oor hoe veral mega-herbivore aangepas is om die heterogeniteit in die
verspreiding van hul voedselbronne te hanteer. Vanuit die hoof tema van hierdie studie oor
die meganismes agter die klein skaalse voedingskeuses van groot herbivore ontstaan nuwe
vrae vir toekomstige navorsing insluitende:
-
Watter morfologiese aanpassings is noodsaaklik vir diere in verhouding tot hul
liggaamsmassas om die uitwerking van klimaatveranderings te kan hanteer?
-
Is die liggaamsmassa reeks van bestaande herbivoor spesies wat tipies saamdrom in groot
getalle, n’ refleksie van hoër voedingsbron beskikbaarheid by klein skale in vergelyking
met herbivore met liggaamsmassas buite hierdie reeks?
Ek kom tot die gevolgtrekking dat grootte as n’ eienskap wat massa en morfologie insluit saak
maak omdat dit eerstens bepaal of n’ herbivoor kan oorleef in n’ spesifieke landskap of nie en
tweedens of herbivore kan saam drom of nie wat verder die ruimtelike skaal waarby nutrient
innames gemaksimiseer word bepaal.
133
134
AFFILIATION OF COAUTHORS
Resource Ecology Group, Wageningen University*:
H.J. de Knegt, PhD candidate
Dr. I.M. Heitkönig, lecturer
E. Kohi, PhD candidate
K. Kortekaas, MSc
E. Mwakiwa, PhD candidate
C. van der Waal, PhD candidate
Dr.ir. F. van Langevelde, lecturer
Dr. S.E. van Wieren, lecturer
M. van Wijngaarden, MSc
*(Droevendaalsesteeg 3a, 6708 PB, Wageningen, the Netherlands)
Systems and Control Group, Wageningen University*:
Dr.ir. J.D. Stigter, lecturer
*(Droevendaalsesteeg 1, 6708 PB, Wageningen, the Netherlands)
International Institute for Geo-information Science and Earth Observation*:
N. Knox, PhD candidate
Prof. Dr. A.K. Skidmore, head of department of Natural Resources
*(PO Box 6, 7500 AA, Enschede, the Netherlands)
Department of Experimental Plant Ecology, Radboud University Nijmegen*:
Prof. Dr. H. de Kroon, head of department
*(P.O. Box 9010, 6500 GL Nijmegen, the Netherlands)
Natural Resource Ecology Laboratory, Colorado State University*:
Dr. M.B. Coughenour, researcher & lecturer
*(Fort Collins, Colorado 80523 USA)
Scientific Services, Kruger National Park*:
Dr. C.C. Grant, senior researcher
*(Private Bag X402, Skukuza 1350, South Africa)
Agricultural Research Council - Animal Production Institute*:
Dr. M.J.S. Peel
*(Institute, PO Box 7063, Nelspruit 1200, South Africa)
135
ACKNOWLEDGEMENTS
When you start your journey as a PhD candidate you have no idea of what to expect and in
hindsight, even my wildest dreams (and sometimes nightmares) could not have prepared me
for the road ahead. However, at the end of it all you come to the realization that the key to
success does not lie in yourself but in the ones who help you along the way. First and
foremost I would like to thank God for granting me the strength to complete this study, for
guiding me every step of the way and for blessing me with the wonderful people who have
shared this experience with me. Even with the best opportunities and technical support in the
world I realized early on that completing this research could not be possible without moral
support and people who truly believe in me. In this regard I would first like to thank my
parents, Danie and Jorina and my best friend Paul for their unconditional love and support,
further, my entire family right through from my grandmothers and grandfather to my aunt
Retha and Dorie, my brother and his wife as well as all my friends out in the field including
Garth, Liz, Emma, Bruce, Judy, Ziggy, Rina, Anna, Nikky, Rebecca and Jock.
Next, I would like to thank the following invaluable people that helped me collect data and
sweat it out in the field under sometimes harsh conditions: field technicians Tian, Rina, Emma,
Jan and especially Floris and Christine, who served the project for the longest period,
volunteers Anne, Daniel, Kirsty, Eve and Steve, MSc students Martine, Arno, Kim and
Machiel and the ever loyal field rangers Reis and Nyati. Further, I am forever grateful to all
the members of the TEMBO programme, especially Hans Stigter, Emmanuel Mwakiwa, Rina
Grant, Mike Peel and Sip van Wieren, who all helped shape the articles to be published from
this thesis, my co-supervisor Rob Slotow for his advice and all members of the Resource
Ecology Group at Wageningen University who constantly provided me with inspiration
during the write-up phase of this thesis.
I think a good supervisor can be categorized as someone who knows his field of science, who
gives quick feedback, who creates an atmosphere where questions can be asked and ideas
shared with comfort and who has the ability to guide when necessary but also to step back and
let the student make his/her own mistakes when necessary. I think Fred de Boer possesses all
these qualities and I feel myself incredibly fortunate to have had him as my daily supervisor
(thank you Fred, and know that I will always be grateful for what you have done for me).
Further, I would like to thank the stakeholders of Mabula Game Reserve, Kruger National
Park and the Associated Private Nature Reserves for allowing me access to these beautiful
areas to collect data, and especially the wardens Scott, Paul, Jacques, Errol, Gustav and Colin,
owners oom Tokkie, oom Piet, oom Wilkens, Andy, Winston and Van Zyl, researchers Steve,
136
Michelle, Rina, Mike and Stephanie and reserve personnel Steve, Cathy and Michelle for all
their assistance as well as the professional hunters, Rudi and Siegfred, who worked with me
to get data on the mouth morphology of different herbivores. Finally, this study would not
have been possible without the financial support from the WOTRO fund, Dr Marie Luttig
trust, Shell-SA, Mr. T. Scholtz and Omnia.
Once, and maybe more than once if you are very fortunate, someone enters your life with the
ability to look right through the barriers, which each of us have, and speak to you in a
language that the essence of your being understands, thereby fuelling the thirst for knowledge
and stimulating the longing to distinguish right from wrong. This is what Herbert Prins has
meant to me during the last five years and as my mentor I will always be grateful to him for
that (Herbert, please know that your teachings have become one of the voices of my
conscience and that this makes me better equipped for the road ahead).
137
138
CURRICULUM VITAE
Yolanda Pretorius was born on the 10th of October 1978 in Pretoria, South Africa (SA). She
matriculated at Verwoerdburg High School, Centurion (SA) in 1996 in the same area where
she grew up. Her academic career started in 1997 at the University of Pretoria (SA) where she
completed her pre-graduate studies in agriculture and animals science (BSc(Agric) Animal
science) and was awarded the SASAS prize for most outstanding under graduate in animal
science in 2000. In 2001 she obtained her Baccalaureus Scientiae Honores degree in wildlife
management at the Centre for Wildlife Management (University of Pretoria). This study
included a small thesis on feed supplementation for the African elephant at Mabula Private
Game Reserve under supervision of Mr. B. Orban, Prof.dr. W. van Hoven and Prof.dr. JduP.
Bothma. Following this, she moved to the University of Kwa-Zulu Natal (SA) to initiate her
masters studies.
In 2003, under the supervision of Prof.dr. R. Slotow, Yolanda completed her Masters
degree in Biology with a behavioural study and thesis on tourism and disrupted social
structure as possible causes of stress for African elephant on small game reserves. Whilst
moving out of academia and formal
education
temporarily
during
2004,
Yolanda acted as field researcher for the
University
subsequently
of
Kwa-Zulu
for
the
Natal
and
governmental
organization, Ezemvelo-KZN wildlife, to
determine the population dynamics of the
elephant at Ithala Game Reserve.
Since 2005, Yolanda has been
registered as a PhD student at the Resource Ecology Group, Wageningen University
(NL) with joint partnership from the University of Kwa-Zulu Natal. She has been part
of the TEMBO programme initiated by the Resource Ecology Group in co-operation
with academic institutions and organizations in South Africa, Australia and the United
States. The main aim of this multi-disciplinary programme is to develope a spatial
explicit cost benefit analysis of elephant management in the Greater Kruger National
Park, South Africa. Yolanda’s function in the programme has been to contribute in the
understanding of the mechanisms governing small-scale foraging choices of large
herbivores. Her study involved two and a half years of field work at a remote location
within the Greater Kruger National Park followed by a year and a half of processing
her data and completing this thesis.
139
140
PE&RC PhD Education Certificate
With the educational activities listed below the PhD candidate
has complied with the educational requirements set by the C.T.
de Wit Graduate School for Production Ecology and Resource
Conservation (PE&RC) which comprises of a minimum total
of 32 ECTS (= 22 weeks of activities)
Review of Literature (4.2 ECTS)
- Investigating mechanisms which determines small scale foraging by herbivores (2008)
Writing of Project Proposal (2 ECTS)
- Small scale foraging by large African herbivores (2005)
Laboratory Training and Working Visits (1 ECTS)
- Plant chemical analysis of N, P, Ca, Mg,, K & tannin; Wageningen University (2008)
Post-Graduate Courses (5.5 ECTS)
-
Survival analysis; FE and PE&RC (2005)
The art of modelling; SENSE Research School (2006)
Introduction of R for statistical analysis; PE&RC (2008)
Spatio-temporal models in ecology: an introduction to Integro-difference equations; PE&RC
(2009)
Deficiency, Refresh, Brush-up Courses (2.8 ECTS)
- Ecological methods; Resource Ecology Group (2005)
Competence Strengthening / Skills Courses (2.5 ECTS)
-
The art of writing; CENTA (2005)
Techniques for writing and presenting a scientific paper; PE&RC (2008)
Discussion Groups / Local Seminars and Other Scientific Meetings (7 ECTS)
-
Ecological theory and application discussion group; Wageningen University (2005 and 2008)
Research progress discussions at the Associated Private Nature Reserves; South Africa (20052008)
PE&RC Annual Meetings, Seminars and the PE&RC Weekend (1.2 ECTS)
-
PE&RC Weekend (2005)
PE&RC Day (2008)
International Symposia, Workshops and Conferences (8.3 ECTS)
-
Elephant Managers and Owners Association, EMOA, annual symposia; South Africa (2006
and 2007)
Scientific services Kruger National Park Network meeting (2006, 2007 and 2008)
Pathways to success conference – Integrating human dimensions into fisheries and wildlife
management; Colorado (2008)
Supervision of MSc Students (4 students, 40 days)
-
Elephant and impala diet switches between grass and browse
Foraging dynamics within patches of browsing ruminant African ungulates
141