ArticlePDF Available

Abstract and Figures

Bovid diets have been studied for decades, but debate still exists about the diets of many species, in part because of geographical or habitat-related dietary variations. In this study we used stable carbon isotope analyses of faeces to explore the seasonal dietary preferences of 11 bovid species from a West African savanna, the Pendjari Biosphere Reserve (PBR), along the browser/grazer (or C3/C4) continuum. We compare our carbon isotope values with those for eastern and southern African bovids, as well as with dietary predictions based on continent-wide averages derived from field studies. Oribi and reedbuck, expected to be grazers were found to be predominantly C3-feeders (browsers) in the PBR. Bushbuck, common duiker and red-flanked duiker consumed more C4 grass than reported in previous studies. When comparing wet and dry season diets, kob, roan and oribi showed the least variation in C3 and C4 plant consumed proportions, while red-flanked duiker, bushbuck, reedbuck and waterbuck showed the most marked shifts. This study shows that animals in the better- studied eastern and southern African savannas do not exhibit the full range of possible dietary adaptations. Inclusion of data from a wider geographical area to include less well-studied regions will inform our overall picture of bovid dietary ecology.
Content may be subject to copyright.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers,
academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
Stable Carbon Isotope Analysis of the Diets of West African Bovids
in Pendjari Biosphere Reserve, Northern Benin
Author(s): C.A.M.S. Djagoun , D. Codron , J. Sealy , G.A. Mensah & B. Sinsin
Source: South African Journal of Wildlife Research, 43(1):33-43. 2013.
Published By: Southern African Wildlife Management Association
BioOne ( is a nonprofit, online aggregation of core research in the
biological, ecological, and environmental sciences. BioOne provides a sustainable online
platform for over 170 journals and books published by nonprofit societies, associations,
museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates
your acceptance of BioOne’s Terms of Use, available at
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use.
Commercial inquiries or rights and permissions requests should be directed to the individual
publisher as copyright holder.
Stable carbon isotope analysis of the
diets of West African bovids in Pendjari
Biosphere Reserve, Northern Benin
C.A.M.S. Djagoun
, D. Codron
, J. Sealy
, G.A. Mensah
& B. Sinsin
Laboratory of Applied Ecology, Faculty of Agronomic Sciences, University of Abomey-Calavi, 01 BP 526 LEA-FSA Cotonou, Benin
Florisbad Quaternary Research, National Museum, P.O. Box 266, Bloemfontein, 9301 South Africa
Clinic for Zoo Animals, Exotic Pets and Wildlife, University of Zürich, 260 Winterthurestrasse, CH-8057, Switzerland
Department of Archaeology, University of Cape Town, Rondebosch, South Africa
National Institute of Agricultural Research of Benin (INRAB), 01 BP2359, Cotonou, Benin
Received 14 January 2013. Accepted 26 March 2013
Bovid diets have been studied for decades, but debate still exists about the diets of many
species, in part because of geographical or habitat-related dietary variations. In this study
we used stable carbon isotope analyses of faeces to explore the seasonal dietary preferences
of 11 bovid species from a West African savanna, the Pendjari Biosphere Reserve (PBR),
along the browser/grazer (or C
) continuum. We compare our carbon isotope values with
those for eastern and southern African bovids, as well as with dietary predictions based on
continent-wide averages derived from field studies. Oribi and reedbuck, expected to be grazers
were found to be predominantly C
-feeders (browsers) in the PBR. Bushbuck, common
duiker and red-flanked duiker consumed more C
grass than reported in previous studies.
When comparing wet and dry season diets, kob, roan and oribi showed the least variation
in C
and C
plant consumed proportions, while red-flanked duiker, bushbuck, reedbuck and
waterbuck showed the most marked shifts. This study shows that animals in the better-
studied eastern and southern African savannas do not exhibit the full range of possible
dietary adaptations. Inclusion of data from a wider geographical area to include less
well-studied regions will inform our overall picture of bovid dietary ecology.
Key words: diet, bovid, graze, browse, savanna.
African Bovidae (Mammalia: Ruminantia) represent
some of the most conspicuously diverse family of
large herbivores globally. More than 80 species
inhabit the continent, and because of their large
size make up an important structural and func
tional component of the biomass of many ecosys
tems (Kingdon 1997). Considerable field research
has focused on the feeding ecology of African
bovid species, to help understand both their evolu
tionary adaptations and their roles in ecosystem
processes (du Toit 2003; Clauss 2008; Smit 2011).
Research on diet composition and diet quality has
revealed that coexistence is achieved primarily
through resource partitioning, in particular the
separation of dietary niches along a browser/grazer
continuum, with three guilds typically recognized:
browsers, grazers and mixed-feeders (Hofmann &
Stewart 1972; Jarman 1974; Owen-Smith 1997;
du Toit 2003; Clauss 2008).
Recent reviews and evidence from stable carbon
isotopes have, however, reached dissimilar con
clusions regarding the diets of many species, and
differ in their respective interpretations of the
heuristic value of the browser/grazer classification
scheme (Gagnon & Chew 2000; Sponheimer
et al.
2003a; Cerling
et al.
2003; Codron
et al.
One source of this confusion is the vast and often
contradictory literature on bovid feeding ecology.
To address this problem, Gagnon & Chew (2000)
undertook a critical literature review in an attempt
to synthesize what is known about diets of African
bovids, an important step towards understanding
bovid trophic patterns. Yet that review was hin
dered by the nature of the available literature
which, because of the variety of methodologies
employed by researchers in the field, and the fact
that most published studies are isolated in space
and time (Owen-Smith 1988, 1997), comprised
data that are not necessarily comparable. More
recently, stable carbon isotope studies (yielding data
that are more comparable across geographical
South African Journal of Wildlife Research 43(1): 33–43 (April 2013)
*To whom correspondence should be addressed.
and temporal boundaries) revealed differences for
several species not only when compared with pre
dictions from the literature, but also between
southern and East African populations (Cerling
et al.
2003; Sponheimer
et al.
2003a; Codron
et al.
2007a). These findings imply that diets of
many species are habitat-specific, and that global
averages as often used in comparative evolu
tionary and ecological studies (Clauss 2008;
Codron & Clauss 2010) may be of limited value.
Numerous methods exist to delineate dietary
patterns of wild animal populations, all with certain
strengths and limitations. Stable isotope analysis
is a more recent, yet by now common tool for
studying the feeding ecology of wildlife (Crawford
et al.
2008), which has the advantage of providing
more comparable data for multiple populations
over various space and time scales. Stable isotope
analysis has proved especially useful for investi
gating the diets of mammalian herbivores in tropical
environments, due to the fact that their two major
food groups,
browse (trees, shrubs, forbs) and
grass (monocots), have highly distinct, non-over-
lapping carbon-isotope compositions (Deines
1980). In these environments, most browse plants
use the C
photosynthetic pathway, whereas most
grasses use the C
pathway(Vogel 1978; Cerling &
Harris 1999). C
plants are consistently depleted in
C relative to C
plants (Ehleringer
et al.
Ehleringer & Monson 1993; Codron
et al.
and this difference is recorded in the body tissues
and faeces of the herbivores that feed on them.
The application of stable isotopes in feeding ecology
has emerged as a powerful tool for partitioning
herbivore diets into browsing and grazing compo
nents, and testing ecological/evolutionary hypoth
eses related to browsing and grazing (Tieszen
et al.
1979; Cerling
et al.
2003; Sponheimer
et al.
2003b; Codron
et al.
Several studies in Africa have used the stable
isotope technique to investigate diets of diverse
herbivore taxa including antelope (Tieszen
et al.
1979; Cerling & Harris 1999; Codron
et al.
2007b,c, 2011a), elephants (Koch
et al.
et al.
2011a), and suids (Cerling & Viehl
2004). Most of this research has focused on East
and southern African savannas (Cerling
et al.
2003; Sponheimer
et al.
2003a; Codron
et al.
2006, 2007a, 2011a,b), with little emphasis placed
on savannas in West Africa. In fact, even beyond
stable isotope investigations, empirical data for
West African bovid ecology is generally lacking
et al.
1998; Kassa
et al.
2007). Given the
obvious value of resolving habitat-specific diets of
many species (Codron
et al.
2007a), and the
unique climatic and habitat conditions of West Afri
can savannas (Kassa
et al.
2007), a comparative
study of bovids between this and other savanna
regions is warranted.
In this study, we use stable carbon isotope analy
sis of faeces to explore the diets of several bovid
species living in a West African savanna (Pendjari
Biosphere Reserve, Benin) along the browser/
grazer (or C
) continuum, and on a seasonal
time-scale. We compare results with dietary pre
dictions for these species based on continent-wide
averages derived from field studies (Gagnon &
Chew 2000), and with carbon isotope data for East
and southern African bovids (Cerling
et al.
et al.
2003a; Codron
et al.
The Pendjari Biosphere Reserve (PBR) is situated
in northwestern Benin (10°30 to 11°30N; 0°50 to
2°00E; Fig. 1). The PBR was declared a Game
Reserve in 1954, and upgraded to the National
Park of Pendjari (NPP) in 1961. Today it comprises
a strictly protected core area (the Pendjari National
Park covering 2660 km²) and two adjacent hunting
zones: Konkombri and Pendjari which are, respec-
tively, on the eastern and southwestern sides of
the park (covering together 1971 km²). Climate is
tropical, with a seven-month dry period between
late October and early April. The PBR is located in
the Sudanian Zone with a single wet season from
April/May to October and one dry season from
November until March. Mean annual precipitation
is 1000 mm, with 60% falling between July and
September (Sinsin
et al.
2002). The mean annual
temperature is 27°C (Verschuren 1988). During
the rainy season numerous small natural ponds
are full of water, as well as large ones in the centre
of the National Park, namely Tiabiga, Fogou,
Mondri, Diwouni, Yangouali and Bali. During the
dry season, the ponds attract a variety of animals,
especially large mammals searching for water.
The dominant vegetation type is savanna
interspersed by some patches of dry forests with
deciduous trees (Sokpon
et al.
2001). Savanna
vegetation is burnt every year by management to
provide fresh pasture to herbivores that dominate
the reserve, provide visibility to wildlife tourists and
hunters who visit mostly during the dry season,
and avoid uncontrolled mid-or late dry season fires
that spread from surrounding villages or that are lit
by poachers to mask illegal activities.
34 South African Journal of Wildlife Research Vol. 43, No. 1, April 2013
Sample collection
Faecal samples of 11 bovid species were collected
from PBR over six months, from December 2011
to May 2012, extending over both late dry and
earlier wet seasons. A total of five samples was
expected to be collected for each species per
month. However, this sample number could not be
attained for topi (
Damaliscus korrigum
), red-
flanked duiker (
Cephalophus rifulatus
), common
duiker (
Sylvicapra grimmia
) and oribi (
) due to the difficulty of observing them in
the field (see details in Table 1). Species affiliation
and freshness of each sample was determined
by an experienced ranger of the PBR according
to the method of Stuart (1994), and location of
the dung samples was recorded using a Global
Positioning System (Fig. 1). Shape and size of
most dung is species-specific and errors of identifi
cation are very unlikely. However, pellet-type
dung may be confused, particularly between
hartebeest (
Alcelaphus buselaphus
) and topi and
between reedbuck (
Redunca redunca
), bushbuck
Tragelaphus scriptus
) and within duiker species,
as suggested by Hibert
et al.
(2008, 2011). For
those species, faecal sampling involved locating
animals and then following on foot. In some cases,
dung samples were collected from bedding sites,
after waiting for animals to get up and move away.
We also collected plant samples over the same
study period to provide baseline isotopic data for
PBR vegetation, as this is the first study in the area
to use the carbon isotope approach. Following
et al.
: Stable carbon isotope analysis of the diets of bovids in Benin 35
Fig. 1. Map of Benin showing the position of Pendjari Biosphere Reserve and the faecal sampling sites.
methods similar to those employed by Codron
et al.
(2005), plant samples were collected from
five ten square metre study plots across the study
area, covering all vegetation and topographic
habitat types. Plant specimens representing the
local diversity of tree, forb and grass species were
collected at each site; we did not sample all
species but the three most dominant plant species
within each plot. Over the whole six-month sam-
pling period, a total of 270 plant specimens were
collected, comprising 20 different species of tree,
13 forb/herb species and five grass species.
Faeces were oven-dried at 60°C and ground in a
Wiley mill to pass through a 40 mm mesh screen.
Samples were then weighed into tin cups to an
accuracy of 1 microgram on a Sartorius micro
balance. The samples were combusted in a Flash
2000 organic elemental analyser and the gases
passed to a Delta V Plus isotope ratio mass spec
trometer (IRMS)
a Conflo IV gas control unit
(all three instruments manufactured by Thermo
Scientific, Bremen, Germany).
C ratios are
expressed as Ç
C values in parts per thousand
(‰), relative to the Vienna PeeDee Belemnite
(VPDB) standard., according to the formula
13 12
13 12
During the analyses, standard deviation for
repeated measurements of the in-house stan
dards was less than 0.15‰.
We estimated the percentage C
fraction (grass)
of the diet represented by each faecal sample
using the following simple linear mixing model
%C i n d i e t =
13 13 13
ants C
C represents the isotopic difference
(discrimination) between diet and faeces. Diet-
faeces discrimination data were taken from recent
controlled-feeding experiments with goats, which
revealed a mean of –0.5‰ and a range of –0.9‰
to +0.3‰ (Codron
et al.
2011a). This diet-faeces
discrimination range does not appear to change
between species with different digestive physio
logies, nor between animals fed C
grass, and C
grass diets (Sponheimer
et al.
2003b; Codron
et al.
2005; Codron
et al.
The model endpoints were based on seasonally-
specific medians of Ç
C values from the local C
and C
vegetation collected over the same time
periods as the faeces. The equation above pro
vides an estimate of C
-intake, but is not an accu
rate index of diet because of variation in Ç
C, in
carbon isotope composition of dietary baselines
(plants), and stochasticity (Post 2002; Cerling
et al.
2003; Codron
et al.
2005, 2007b).To account
for this variation in our analysis, we varied a
values randomly between the published maximum
and minimum values reported by Codron
et al.
(2011a), and C
and C
plant endpoint values
between their respective 25th to 75th interquartile
ranges. We report on means and 95% confidence
limits derived from 1000 iterations. The average
95% confidence interval derived from this proce
dure was ~18%, which is similar to an error of
36 South African Journal of Wildlife Research Vol. 43, No. 1, April 2013
Table 1. Dry-to-wet season mean faecal Ç
C values ± 1 S.D. for bovids from Pendjari Biosphere Reserve, and
estimated C
fraction of the diet.
Dry season Wet season
(‰) Mean % C
(‰) Mean % C
in diet in diet
Syncerus caffer
20 –16.5 ± 1.82 82 10 –16.0 ± 1.64 84 1.973
Tragelaphus scriptus
20 –27.2 ± 4.05 32 10 –21.5 ± 7.10 45 9.102*
Alcelaphus buselaphus
20 –15.5 ± 2.26 89 10 –16.8 ± 3.25 80 7.009
Damaliscus korrigum
19 –14.7 ± 2.19 95 7 –14.4 ± 1.19 96 1.343
Hippotragus equinus
20 –14.5 ± 1.90 96 10 –15.5 ± 2.53 88 2.286
Kobus ellipsiprymnus
20 –19.0 ± 2.91 64 10 –15.8 ± 2.93 89 8.546*
Kobus kob
20 –14.8 ± 0.94 95 10 –13.9 ± 0.62 100 2.597
Redunca redunca
20 –21.3 ± 7.80 47 10 –25.2 ± 5.13 19 8.454*
Ourebia ourebi
20 –27.3 ± 0.39 1 8 –26.6 ± 2.58 6 2.600
Sylvicapra grimmia
20 –24.5 ± 3.68 21 7 –21.6 ± 3.32 46 6.704
Cephalophus rifulatus
14 –28.6 ± 0.57 0 8 –25.1 ± 4.24 18 6.941*
*Maximum likelihood chi-square test indicates significant seasonal shift in diet (
< 0.05).
±10% estimated previously (Codron
et al.
Thus, the data were binned in intervals of 20% (
categories of <10% C
intake, 10–30%, 30–50%,
50–70%, 70–90%, and >90% C
; the <10%
and >90% categories include animals eating
only C
or only C
foods). We tested for seasonal
shifts in species diets based on the frequency
distribution of individual samples in each %C
using maximum likelihood chi-square tests. The
number of dung samples represented for each
species varies widely between months. To increase
statistical power, we grouped data into two sea
sons: dry (December to March) and wet (April and
May). We then checked whether the results for
species from PBR differed from predictions derived
from field studies (FS) in either season by compar
ing their seasonal means with means reported in
Gagnon & Chew (2000). Note that Gagnon &
Chew (2000) reported on % fruits, % dicots, and
% monocots in the diet: we assume here that
monocots are equivalent to % C
grass in the
diet (see also Cerling
et al.
2003). Based on
the analytical error inherent in carbon isotope
approaches to diet described above, we consider
two values to be different only if they differ by 10%
or more, as suggested by Codron
et al.
Finally, we made similar comparisons with data
from Cerling
et al.
(2003); Sponheimer
et al.
(2003a), and Codron
et al.
(2007a) to determine
whether bovid diets in PBR differ from those
of their counterparts in East African (EA) and
southern African (SA) savannas, respectively.
Patterns of seasonal diet variation
plants (tree/shrub and forb) and C
from PBR were as expected for plants following the
-photosynthetic pathway (mean = –27.14‰ ±
= 60), and C
-photosynthetic pathways
(mean = –14.20‰ ± 2.11,
= 30), respectively.
These figures did not change temporally, means
for C
plants ranging between –26.33‰ and
–27.74‰ across different months, and similarly
monthly means for C
grasses ranged from
–12.29‰ to –14.37‰. Further, our approach to
modelling bovid diets incorporates variations in
plant Ç
C values across months, and between C
and C
groups, and so any variations are unlikely
to influence our dietary analysis in a major way.
Savanna faecal Ç
C values in the dry season
reflected the C
-dominated diets of savanna
grazers, as in buffalo (
Syncerus caffer
) (mean =
–16.5‰ ± 1.82,
= 20, ~82% C
in diet), red harte
beest (mean = –15.5‰ ± 2.26,
= 20, ~89% C
diet), roan antelope (
Hippotragus equinus
(mean = –14.5‰ ± 1.90,
= 20, ~88% C
in diet),
kob (
Kobus kob
) (mean = –14.8‰ ± 0.94,
= 10,
~95% C
in diet), topi (mean = –14.7‰ ± 2.19,
19, ~95% C
in diet) and waterbuck (
) (mean = –19.0‰ ± 2.91,
= 20,
~64% C
in diet). Amongst these grazers only
waterbuck show a significant seasonal shift in
mean estimated % C
(Ų = 8.546, d.f. = 3,
0.035) with C
-intake reaching 89% in the wet
season (Table 1). Five bovid species had fecal
C values reflecting C
-dominated diets: oribi
Ourebia ourebi
) (mean = –27.3‰ ± 0.39,
= 20),
red-flanked duiker (
Cephalophus rifulatus
(mean = –28.6‰ ± 0.57,
= 14), bushbuck
Tragelaphus scriptus
) (mean = –27.2‰ ± 4.05,
= 20), common duiker (mean = –24.5‰ ± 3.68,
= 20) and reedbuck (mean = –21.3‰ ± 7.80,
20). Three of these species red-flanked duiker,
reedbuck and bushbuck showed significant sea-
sonal diet switches (Ų = 6.941, d.f. = 3,
= 0.008;
Ų = 8.454, d.f. = 3,
= 0.037; Ų = 9.102, d.f. = 3,
= 0.028, respectively) in mean % C
between the dry season (mean % C
intake for
red-flanked duiker 0, for reedbuck = 47, and for
bushbuck = 32) and wet seasons (mean % C
take for red-flanked duiker = 18, reedbuck = 19,
and bushbuck = 45). More extensive isotopic, and
hence dietary, variation was observed at monthly
scales (Fig. 2).
Seasonal variations were also recorded
amongst C
-feeders. Roan antelope and harte
beest consumed only 30–50% C
in January and
April, and a similarly low range of C
intake was re
corded for topi in January. At monthly scales diets
were even more variable, with lower % C
occurring in several months for buffalo (ranging
up to 50–70% C
in January) and waterbuck
(in December, January February ranging up to
50–70% C
). Kob was found to maintain a
more-or-less constant high C
-intake throughout;
however a small amount of C
-intake (10–30%)
occurred in December and February. Fidelity in
-intake (browsing) was found at monthly level for
the red-flanked duiker and oribi, with some C
grazing occurring only in April (10–30% C
) and
May (10–30% C
), respectively. Common duiker
was found to feed on the range of 50–70% C
some months (March and May). The most striking
shifts occurred in reedbuck and bushbuck, from
et al.
: Stable carbon isotope analysis of the diets of bovids in Benin 37
38 South African Journal of Wildlife Research Vol. 43, No. 1, April 2013
Fig. 2. Monthly variation of % C
categories are <10% C
, 10–30%, 30–50%, 50–70%, 70–90%, and >90% C
) in diets of 11 bovid species from Pendjari Biosphere
pure C
feeding (<10% C
) in some months to
>90% C
in others.
Comparison with continent-wide C
consumption by bovid species
Five species from PBR (buffalo, hartebeest,
roan, kob and waterbuck) had mean % C
in diet
similar to data derived from field studies (Table 2).
Another four (bushbuck, common duiker, oribi and
reedbuck) differed substantially (by >30%) from
the continent-wide averages. Two species known
to be primarily grazers elsewhere in Africa
(reedbuck and oribi) had C
-dominated diets
in PBR. A difference of more than 10% C
in diet
was also found for buffalo from PBR compared
with the same species from East Africa, and for red
hartebeest and waterbuck in comparison with both
east and southern Africa. Percent C
intake by kob
and roan antelope did not change across different
regions of Africa (Table 2). There is no difference
between PBR and Kruger National Park for buffalo,
roan and red hartebeest, even when the data are
separated into dry and wet seasons.
We analysed the proportions of C
grass in the diets of 11 bovid species from Pendjari
Biosphere Reserve (PBR), Benin, West Africa. Of
these, eight were expected to be grazers (buffalo,
reedbuck, waterbuck, hartebeest, kob, oribi, roan
and topi), and three browsers (red-flanked duiker,
bushbuck and common duiker), based on data in
the literature (
Gagnon & Chew 2000). How
ever, most literature data are derived from studies
in East and/or southern African savannas. Carbon
isotope results from faeces of PBR bovids implied
that only six species (buffalo, waterbuck, harte-
beest, kob, roan and topi) had primarily C
grass-dominated diets, whereas five species
focused largely on C
(reedbuck, oribi, red-flanked duiker, bushbuck
and common duiker). This suggests differences
between bovid species diet at PBR compared with
other African regions, including southern and East
African savannas where previous carbon isotope-
based research was carried out (Cerling
et al.
2003; Sponheimer
et al.
2003a; Codron
et al.
2007a) and thus diets are not constant across
space. The results also demonstrate temporal diet
switches at monthly and seasonal scales, indicat
ing that even supposed stenotopic grazers and
browsers can have highly variable diets.
One source of uncertainty could come from the
misidentification of faeces. This was minimized by
using a sampling protocol (see details in method
section) whereby faeces of species whose dung is
easy to misidentify were collected from animals
under observation. In the case of the two species
that presented the most unexpected results
(reedbuck and oribi), the chance of misidentifica
tion was further reduced by a
post hoc
older dung samples were compared to fresh
samples that were identified with certainty, based
on the size, shape and indentation of the pellets.
Additionally samples were then double-checked
(blind test) by experienced local trackers and
et al.
: Stable carbon isotope analysis of the diets of bovids in Benin 39
Table 2. Estimated % C
-intake of 11 bovid species from Pendjari Biosphere Reserve (PBR) compared with other
savanna regions in Africa
Comparative % C
in diet
PBR Field studies East Africa Southern Africa Kruger Kruger
Species (FS) (EA) (SA) (SA): dry (SA): wet
Syncerus caffer
83 78 100 88 89 88
Tragelaphus scriptus
38 0 0 10
Alcelaphus buselaphus
84 75 100 98 90 89
Damaliscus korrigum
Hippotragus equinus
92 85 100 91 98 95
Kobus ellipsiprymnus
76 84 92 100
Kobus kob
97 95 95
Redunca redunca
33 95 100
Ourebia ourebi
382 84 90
Sylvicapra grimmia
33 12 0 2
Cephalophus rifulatus
Data for field studies are continent-wide averages of % monocots in diet reported in Gagnon & Chew (2000). Data from East Africa (EA) and
southern Africa (SA) are means reported in Cerling
et al.
(2003) and Sponheimer
et al.
(2003a), respectively, and are derived from analyses of
bones, teeth and hair. Figures for Kruger National Park in South Africa are from Codron
et al.
(2007a) and are derived from faeces.
finally checked again using Stuart & Stuart (2000).
Since faeces consist largely of the undigested
portion of an animal’s diet, geochemical proxies
from this material might be expected to be biased
towards less-digestible food items. This does not,
however, appear to be a significant problem for
carbon-isotope studies (Codron
et al.
2007b), as
controlled-feeding trials have shown that the Ç
values of faeces and herbage ingested are closely
related (Sponheimer
et al.
2003a; Codron
et al.
2011a). Additionally, Codron & Codron (2008)
have shown that spatio-temporal patterns of varia
tion in plant and faecal Ç
C are very strongly cor
related, demonstrating fidelity with which faecal
C reflects the isotopic composition of foods con
sumed. However, sources of variation such as
differences between plant species, plant parts,
and small differences in diet-consumer (faeces)
C-discrimination between species and environ
ments may influence results (Sponheimer
et al.
2003b; Codron
et al.
2005). For these reasons,
the use of mixing models that take into account
these variations, as done in this study, inspire
greater confidence in results. Further, in this study
we made use of data bins resulting from mixing
models, rather than focusing on specific values
of C
intake as they are only estimates within ±
10% error. Finally, we acknowledge that compari-
sons between our data and those derived from
East and southern Africa may be complicated
because PBR data are derived from faeces,
representing only short-term dietary information,
whereas some of the other studies are based on
longer-term body tissues like hair, bones and
teeth. Comparative data from Kruger National
Park in South Africa are based on faeces, and are
in broad agreement with results from hard tissues,
suggesting that there should at least be general
congruence amongst datasets at broad, multi-
species scales such as employed in the current
study (Codron
et al.
The analysis of short-term materials, such as
faeces, revealed subtle changes in diets of brows
ers and grazers from PBR at seasonal (wet/dry
season) and monthly time scales. Grazing species
that showed very little evidence of temporal diet
switching were topi, kob and roan antelope: these
species consumed high proportions of C
throughout. These results contradict an earlier
study of West African grazers, based at Nazinga
ranch in Burkina Faso, which reported that the
proportion of browse plants in the diets of grazer
species increased during dry seasons as grasses
reach senescence (Schuette
et al.
1998). Although
in PBR the grass layer also became substantially
reduced over the dry season, palatable grasses
remained in abundance along the natural pond
and river margins for highly water-dependent
species such as kob, which need to drink daily and
forage not far away from water sources (Estes
1991; Kingdon 1997; Smit 2011). Also, after
grasses had dried out during the hot, dry season,
fires during the middle of the cool, dry season
removed old growth and promoted re-growth of
some perennial grasses (Van de Vijver
et al.
et al.
2005; Klop
et al.
2007; Sensenig
et al.
savanna 2010), which may have been
utilized by the topi and roan antelope to supply
the dry season demand. Studies elsewhere in
Africa revealed that roan antelope are monoto
nous C
-grazers throughout the seasonal cycle
(Codron & Codron 2008; Codron
et al.
2009). Less
is known about topi diets, although they are gener
ally perceived to be obligate grazers because of
their preference for grassland habitats (Murray &
Brown 1993). Although feeding habits of the topi
are little known, those of its close relative, the
tsessebe (
Damaliscus lunatus
), are better under-
stood. Tsessebe are reported to consume pure or
near-pure C
grass-based diets throughout the
year (~98% C
) (Codron
et al.
2009). Thus the lack
of seasonality in topi diets at PBR appears to be
similar to what has been observed in its close rela-
tives in other African savannas. Buffalo and harte
beest are also grazers, and though they show
slightly more seasonal variation than topi, kob and
roan antelope, this remains very limited. Similar
results have been reported for buffalo from Kruger
National Park in South Africa (Codron
et al.
Among species with predominantly C
diets, the only species at PBR to show significant
seasonal variation in the proportion of C
ingested is the waterbuck, which changed from a
diet comprising approximately 89% C
grass in the
wet season to 64% C
grass in the dry season. A
recent study on the feeding ecology of waterbuck
in the Pendjari Biosphere Reserve has also shown
browse-based feeding behavior during the dry
season, with waterbuck switching to annual
grasses (
Oryza barthii, Diheteropogon amplec
tens, Hyparrhenia involucrata
) only after the first
rains (Kassa
et al.
2007). The drop in grass nutri
tional quality in dry seasons (
et al.
2006) would account for the observed diet shift in
waterbuck. The absence of a seasonal diet shift in
40 South African Journal of Wildlife Research Vol. 43, No. 1, April 2013
grazers other than waterbuck could be due to the
fact that waterbuck require about four times more
protein than other similar-sized grazers (Taylor
et al.
1969), and are thus more likely to be more
compelled to switch during the dry seasons.
Amongst expected browsing bovid species from
PBR (bushbuck, red-flanked duiker, and common
duiker), bushbuck and red-flanked duiker increased
their % C
intake significantly from dry to wet
season. Common duiker did not show a directional
seasonal shift, but their diets did vary across
months. A possible explanation for the change in
bushbuck and red-flanked duiker could be that
both species need high quality food all the time,
and therefore have adapted to optimize their diet to
maintain high protein content and low fibre (Van
Rooyen 1992; Meissner
et al.
1996; Sponheimer
et al.
2003a; Codron
et al.
2006). Thus, despite
being primarily browsers, both species are able to
tolerate a wider variety of foods, including grass
during the wet season, while ecological diet plas-
ticity was most pronounced in the common duiker.
Reedbuck and oribi from the PBR had substan-
tially more negative faecal Ç
C than expected from
the literature (Gagnon & Chew 2000; Cerling
et al.
2003; Sponheimer
et al.
2003a) reflecting predom-
inantly C
diets. Previous studies have reported
small amounts of dicots in the diets of oribi, espe-
cially during the dry season (Reilly
et al.
et al.
2006), but the PBR oribi were
almost pure C
browsers, with the least variation
between wet and dry seasons of any of the species
studied. Identification of the dicots consumed was
not possible in this study. We suspect that they
consist mainly of forbs (legumes and others)
rather than leaves from woody plants, but this
needs to be confirmed by further studies.
Given the discrepancies in % C
in the diets of
bovids from PBR compared with literature for other
parts of Africa, it is clear that there is spatio-temporal
dietary flux in many bovid species. Bovid diets
are, to some extent, habitat-specific and averaging
of % monocots in diet from continent-wide field
studies does not adequately represent dietary
diversity among African bovid species. Further
research using material such as bones, teeth and
hair, which preserve longer-term records, could
help us to understand the nature and extent of
dietary variation amongst African bovids.
The field work for this study was funded by
LOEWE-Biodiversity and Climate Research Cen
tre (BiK-F). Analytical work was carried out in the
Stable Light Isotope Laboratory in the Department
of Archaeology at the University of Cape Town,
supported by UCT and the South African National
Research Foundation. The first author’s visit to
Cape Town was funded by the Centre for African
Origins at the University of Cape Town, supported
by the Vice-Chancellor’s Strategic Initiative. We
are grateful to the local communities who partici
pated in this research. Our acknowledgments also
go to our field guide Mr Tchabi, for his invaluable
help with field work and bovid faecal pellet identifi
FAIRBANKS, D.H.K. 2005. Shaping the landscape:
fire-grazer interactions in an African savanna.
15: 96–109.
CERLING, T.E. & HARRIS, J.M. 1999. Carbon isotope
fractionation between diet and bioapatite in ungulate
mammals and implications for ecological and paleo
ecological studies.
120: 347–363.
CERLING, T.E. & VIEHL, K. 2004. Seasonal diet
changes of the forest hog (
Hylochoerus meinertz-
) based on the carbon isotopic composition of
Afr. J. Ecol.
42: 88–92.
Diets of East African Bovidae based on stable isotope
J. Mammal.
84: 456–470.
CLAUSS, M. 2008. The morphophysiological adapta-
tions of browsing and grazing mammals. In: I.J.
Gordon & H.H.T. Prins (Eds),The ecology of browsing
and grazing. Springer, Heidelberg.
2011a. Landscape-scale feeding patterns of African
elephant inferred from carbon isotope analysis of
165: 89–99.
HEIMER, M. & DE RUITER, D. 2005. Animal diets in
the Waterberg based on stable isotopic composition
of faeces.
S. Afr. J. Wildl. Res.
35: 43–52.
CODRON, D. & CLAUSS, M. 2010. Rumen physiology
constrains diet niche: linking digestive physiology
and food selection across wild ruminant species.
Can. J. Zool.
88: 1129–1138.
CODRON, D. & CODRON, J. 2008. Reliability of δ
C and
N in faeces for reconstructing savanna herbivore
Mamm. Biol.
74: 36–48.
Dietary variation in impala
Aepyceros melampus
recorded by carbon isotope composition of feces.
Acta Zool. Sin.
52: 1015–1025.
HEIMER, M., GRANT, C.C. & BRINK, J.S. 2009.
Stable isotope evidence for nutritional stress, compe
tition, and loss of functional habitat as factors limiting
recovery of rare antelope in southern Africa.
J. Arid.
. 73: 449–457.
et al.
: Stable carbon isotope analysis of the diets of bovids in Benin 41
BERNASCONI, S.M. & CLAUSS, M. 2011b. When
animals are not quite what they eat: diet digestibility
C-incorporation rates and apparent
discrimination in a mixed-feeding herbivore.
Can. J.
89: 453–465.
R. & FOURIE, N. 2007b. Diets of savanna ungulates
from stable carbon isotope composition of faeces.
J. Zool.
27: 21–29.
CODRON, J. 2007c. Stable carbon isotope recon
struction of ungulate diet changes through the sea
sonal cycle.
S. Afr. J. Wildl. Res.
37: 117– 125.
CODRON, J., DE RUITER, D. & BRINK, J.S. 2007a.
Significance of diet type and diet quality for ecological
diversity of African ungulates.
J. Anim. Ecol.
2008. Applications of stable isotope techniques to the
ecology of mammals.
Mamm. Rev.
38: 87–107.
DEINES, P. 1980.The isotopic composition of reduced
organic carbon. In: P. Fritz & J.C. Fontes (Eds), Hand
book of environmental isotope geochemistry. Else-
vier, Amsterdam.
DU TOIT, J.T. 2003.Large herbivores and savanna heter-
ogeneity. In: J.T. du Toit, K.H. Rogers & H.C. Biggs
(Eds), The Kruger experience: ecology and manage-
ment of savanna heterogeneity. Island Press, Wash-
EHLERINGER, J.R. & MONSON, R.K. 1993. Evolution-
ary and ecological aspects of photosynthetic path-
way variation.
Ann. Rev. Ecol. Syst.
24: 411–439.
PEARCY, R.W. 1991. Climate change and the evolu
tion of C
Trends. Ecol. Evol.
ESTES, R.D. 1991. The behavior guide to African mam
mals, including hoofed mammals, Carnivores, Pri
mates. University of California Press: Berkeley.
GAGNON, M. & CHEW, A. 2000.Dietary preferences in
extant African Bovidae.
J. Mamm.
81: 490– 511.
DULIEU, D. 2008. Morphological criteria to identify
faecal pellets of sympatric ungulates in West African
savanna and estimates of associated bias.
Afr. J.
46: 523–532.
ABDOU, H.N. & WINTERTON, P. 2011. Ageing of
ungulate pellets in semi-arid landscapes: how the
shade of color can refine pellet-group counts.
Eur. J.
Wildl. Res.
57: 495–503.
HOFMANN, R.R. & STEWART, D.R.M. 1972. Grazers
and browsers: a classification based on the stomach
structure and feeding habits of East African rumi
36: 226–240.
JARMAN, P.J. 1974. The social organization of antelope
in relation to their ecology.
48: 215–266.
KASSA, B., LIBOIS, R. & SINSIN, B. 2007. Diet and food
preference of the waterbuck (
Kobus ellipsiprymnus
) in the Pendjari National Park, Benin.
Afr. J.
46: 303–310.
KINGDON, J. 1997. The Kingdon field guide to African
mammals. Academic Press, London and New York
Resource selection by grazing herbivores on post-
fire regrowth in a West African woodland savanna.
Wildlife. Res.
34: 77–83.
Isotopic tracking of change in diet and habitat use in
African elephants.
267: 1340–1343.
McCANN, K., DALY, B. & FRIEDMANN, Y. 2006.Oribi an
telope (
Ourebia ourebi
). The Conservation Breeding
Specialist Group (CBSG IUCN/SSC), CBSG South
ern Africa Endangered Wildlife Trust Oribi Working
Group, Ezemvelo KwaZulu-Natal Wildlife
1996. Seasonal food selection by male impala
Aepyceros melampus
in two habitats.
S. Afr. J. Wildl.
26: 56–63.
MURRAY, M.G. & BROWN, D. 1993. Niche separation of
grazing ungulates in the Serengeti: an experimental
J. Anim. Ecol.
62: 380–389.
OWEN-SMITH, N. 1997. Distinctive features of the nutri
tional ecology of browsing versus grazing ruminants.
Z. Säugetierk.
62: 176–191.
OWEN-SMITH, R.N. 1988. Megaherbivores. The influ-
ence of very large body size on ecology. Cambridge
University Press, Cambridge.
POST, D.M. 2002. Using stable isotopes to estimate
trophic position: models, methods, and assumptions.
REILLY, B.K., THERON, G.K. & BOTHMA, J. du P. 1990.
Food preferences of oribi
Ourebia ourebi
in the Golden
Gate Highlands National Park.
33: 55–61.
JENKS, J.A. 1998.Diets of hartebeest and roan ante-
lope in Burkina Faso: support of the long-faced
. J. Mamm.
79: 426–436.
Allometric scaling predicts preferences for burned
patches in a guild of East African grazers.
2002. Abundance and species richness of larger
mammals in Pendjari National Park in Benin.
66: 369–380.
SMIT, I.P.J. 2011. Resources driving landscape-scale
distribution patterns of grazers in an African savanna.
34: 67–74.
C. & BARBIER, N. 2001. Inventaire et caractérisation
des formations végétales du complexe national de la
Pendjari, zone cynégétique de la Pendjari et de
l’Atacora: région de Konkombri. Rapport. Université
Nationale du Bénin/Bénin.
HEIDELBERGER, C.& MARCUS, W. 2003a. Diets of
southern African Bovidae: stable isotope evidence.
J. Mamm.
84: 471–479.
42 South African Journal of Wildlife Research Vol. 43, No. 1, April 2013
2003b. An experimental study of carbon-isotope frac
tionation between diet, hair, and feces of mammalian
Can. J. Zool.
81: 871–876.
STUART, C. 1994. A field guide to tracks and signs of
southern and East African wildlife. Southern Book
Publishers, Halfway House, South Africa.
Water relations of the waterbuck, an East African
Am. J. Physiol.
217: 630–634.
J. & IMBAMBA, S. 1979. Use of δ
C values to deter
mine vegetation selectivity in east African herbivores.
37: 351–359.
1999. Causes of increased nutrient concentrations in
post-fire regrowth in an East African savanna.
214: 173–185.
VAN ROOYEN, A.F. 1992. Diets of impala and nyala in
two game reserves in Natal, South Africa.
S. Afr. J.
Wildl. Res.
22: 98–101.
VERSCHUREN, J. 1988. Notes d’écologie, principale
ment des mammifères du Parc National de la
Pendjari, Bénin.
58: 185–206.
VOGEL, J.C., FULS, A. & ELLIS, R.P. 1978. The geo
graphical distribution of Kranz grasses in South
S. Afr. J. Sci.
74: 209–215.
et al.
: Stable carbon isotope analysis of the diets of bovids in Benin 43
Corresponding Editor: M.R. Perrin
... Diet, niche partitioning, and community structure of African mammal herbivore assemblages has been studied extensively in East and southern African savannas, but less attention has been paid to West African assemblages (Schuette et al., 1998) despite the rich diversity of ungulates and inherent environmental (climate, vegetation) uniqueness of this region (Kassa et al., 2007;Assédé et al., 2012). Recently we showed, using stable carbon isotope analysis of faeces, that the diets of several West African herbivore taxa differ fundamentally in terms of browse/grass composition in comparison with the same (or closely-related) species elsewhere on the continent (Djagoun et al., 2013a). Therefore, the role of body size in structuring these assemblages is even less well-known than it is in East and southern African environments. ...
... Based on competition theory, we expect that species of similar body size would have more similar dietary (and hence isotopic) niches compared with differently-sized syntopic taxa. Using the carbon (␦ 13 C) data published in Djagoun et al. (2013a), we test the hypothesis that isotopic niche segregation amongst sympatric species is non-random, reflecting differential use of dietary niches between browsers and grazers, as well as within these guilds. We then present previously unpublished faecal stable nitrogen isotope data, to further resolve the niche breadths of each taxon, and ultimately to compare patterns of isotopic similarity with body size similarity across taxa. ...
... Of the various tissues that can be used for isotopic diet reconstructions of living animals, faeces are usually the most readily available, since they can be obtained entirely non-invasively. The faecal samples in this study are the same as those for which the carbon isotope values were reported in Djagoun et al. (2013a). The sample comprises 11 bovid species inhabiting the PBR, collected monthly from December 2011 to May 2012. ...
Understanding the mechanisms of species coexistence within local assemblages can play a crucial role in conservation of a species. There is little understanding of how large mammalian bovid species from West Africa partition diet resources, and to what extent they may vary their diet and habitat selection seasonally in order to coexist. Here we studied an assemblage of eleven bovid species in Pendjari Biosphere Reserve, West Africa and used faecal stable isotopes of carbon (δ¹³C) and nitrogen (δ¹⁵N) to test the impact of body mass diet partitioning at a seasonal scale. We found a significant positive relationship between isotopic niche similarity and body size similarity both in dry (p < 0.001) and wet (p < 0.001) season. Partitioning of carbon isotope niches is at least partly due to interactions amongst species rather than historical effects. Our findings also show numerous patterns in resource partitioning amongst the 11 bovid species studied, suggesting that different species used dietary resources in contrasting ways. In practice, actual resource competition between bovid species is difficult to demonstrate, but there exists much overlap in diet along the stable carbon isotope axis for most of the studied species. However we conclude that in our study area, especially in the wet season, niche breadth and diet overlap remain large. Abundant resources and low herbivore densities mean there is no need for herbivores to specialize, because they do not have to compete over scarce resources.
... They are most successful at maintaining their protein intake on semi-mature grass swards, because these contain abundant green grass leaves. Meeting their feeding requirements in semi-arid savanna systems marked by large changes in resource availability may be more difficult in coexisting systems where topi has to share resources with other, more abundant sympatric bulk grazers such as hartebeest (Alcelaphus buselaphus) and roan (Hippotragus equinus) with similar dietary requirements (Djagoun et al., 2013). Therefore, all these species could contribute to topi decline through depletion of food resources during the critical more pronounced dry season. ...
... As the three species topi, hartebeest and roan are obligate grazers, gregarious, and feed on similar resources (Djagoun, 2013), strategies such as low trophic niche overlap are expected, in order to promote their co-existence. However, the resource availability concepts suggest that relatively rare species have the capacity to exploit a wide range of resources, but are restricted to places where resources remain unused by superior competitors (Rosenzweig and Lomolino, 1997), resulting in the selection of a low-quality diet. ...
... The stable isotope analysis performed on tooth collagen confirmed that the study species are grazers and their diet is composed of C4 plants with consistent N (d 15 N and %N) levels (Codron et al., 2009). This is consistent with the findings of Djagoun et al. (2013) in Pendjari, who studied the diet of the same species using feces samples. The d 15 N value of herbivore body tissue is controlled by that of the diet (Hartman, 2011). ...
Understanding factors limiting rare species is a major concern in conservation ecology. Little is known about how large mammalian bovid species from West Africa partition diet resources, and to what extent the rare topi antelope and other sympatric bulk grazers can cope with this co-existing system. Using tooth collagen stable isotope data, we studied an assemblage of three sympatric grazers (topi, hartebeest and roan) with different population density. Our findings showed numerous patterns in resource partitioning amongst the three sympatric species studied, suggesting that the species used dietary resources in contrasting ways. There were significant differences in both δ13C and δ15N signatures among topi, hartebeest and roan. A very narrow range could be noticed for the rare topi antelope (−6.05‰ to −7.63‰) for δ13C values and no significant differences at individual scale. The more abundant species such as hartebeest and roan were found to use a high proportion of the isotopic niche space of topi (more than 70%), while topi used only a narrow range of hartebeest and roan niche space (less than 10%). Our study provides the first evidence for the resource partitioning amongst high and low density ungulates in West African savanna. Expansion of the niche noted in the roan and hartebeest is much more related to the strong variation of the niche at the individual level. Topi, as subordinate species narrow their niche instead of expanding the dietary niche as expected. Our findings regarding the absence of dietary niche flexibility observed in the topi despite it being considered as subordinate species were discussed in being one of the factor limiting its population recovery in West African grazer assemblages. Keywords: Diet partitioning, Grazers, Low density, Nutritional stress, Isotopic niche
... The hypothesis that the morphophysiological differences were based on the functional relevance of stratified rumen content for the retention of particles of grass or browse diets (Clauss et al., 2003) was also refuted by empirical testing (Lechner et al., 2010). Additionally, it was increasingly realized that ruminant species often did not follow the strict feeding type categorisation, with animals of a certain morphophysiology also ingesting other forages than expected (Djagoun et al., 2013;Marchand et al., 2013;Clauss and Hofmann, 2014;Przybylo et al., 2019). Such apparent discrepancies led to the statement used in a publication on ruminant dental anatomy that 'morphology is not destiny' (Gailer et al., 2016). ...
Rumen content stratification and the degree of dissociation of particle and fluid retention in the reticulorumen differ between ‘moose-type’ and ‘cattle-type’ ruminant species. These differences are not strictly linked to diet, except for a seeming limitation of ‘moose-type’ ruminants to a browsing niche. Nevertheless, these differences can be plausibly linked to other observed differences in ruminants, such as the intraruminal papillation pattern, or the size of the omasum. However, many of the corresponding measures are still only available for a restricted number of species. Here, we investigated the dry matter (i.e., the inverse of the moisture) concentration in forestomach contents of 10 blackbuck (Antilope cervicapra) and 7 Arabian sand gazelle (Gazella subgutturosa marica), and quantified the rumen papillation pattern. The blackbucks had distinct rumen contents stratification, with more moisture in ventral than in dorsal contents (difference 3.6% units, P < 0.001), whereas this difference was much less pronounced in the sand gazelles (0.6% units, P = 0.227). While reticulum contents were particularly moist in both species, omasum contents were particularly dry in sand gazelles, but did not differ in moisture from rumen contents in the blackbuck. This species is an outlier among ruminants due to its extremely small omasum. The intraruminal papillation pattern did not differ between blackbucks and sand gazelles and showed a surface enlargement factor (SEF) in the dorsal rumen of 27–28% of the SEF in the Atrium ruminis. Compared to data on digesta retention in the same species, the findings are in line with the overall concept of a high fluid throughput causing a distinct stratification of rumen contents and intraruminal papillation, and necessitating a large omasum for fluid re-absorption. However, the data also show that individual species may not correspond to all the assumptions of the concept, suggesting taxon-specific differences between species. Reasons for these differences cannot be linked to a dietary grass-browse spectrum, but may lie in evolutionary contingency.
... Gambia (Jallow, Touray, & Jallow, 2004), few quantitative studies are currently available (i.e. in Benin: Djagoun, Codron, Sealy, Mensah, & Sinsin, 2013;Kassa et al., 2008), and little is known of the population dynamics of any West African waterbuck population. ...
en The waterbuck (Kobus ellipsiprymnus), though widespread throughout Africa, is suspected to be declining overall. Data on population numbers and structure are lacking for many parts of its range, especially in West Africa, where the subspecies defassa is found. The aim of the present study was to evaluate the abundance, distribution and attributes of waterbuck populations from the Nazinga Forest Reserve, southern Burkina Faso. We investigated waterbuck population trends in the park using transect data collected in 1985–2019. For the more detailed analyses of population structure and distribution of the animals, we used census data gathered during 2019. Most animals were adults (46.6%), and the sex ratio was heavily skewed towards females (5:1). Most animals were concentrated along the larger rivers. There was no influence of poacher activity on waterbuck distribution. In the long term (1985–2019), the population dynamics of waterbuck can be roughly divided into two main periods: a phase of population increase from 1985 to 2005, and one of ongoing population collapse from 2007 to 2019. Although the declining population trend was obvious, coefficients of determination were low indicating that the years explained poorly the number of individuals and the number of sightings obtained. Waterbuck numbers in the Nazinga Forest Reserve are declining, but we found no single reason to explain this trend. It is likely that a combination of factors, including global warming (increased aridity) and illegal activities such as poaching, is responsible. Because there are probably multiple reasons for the observed waterbuck population decline in our study area, we suggest that a multifaceted approach should be adopted in order to enhance the conservation status of the local waterbuck populations. Résumé fr Bien que répandu dans toute l'Afrique, le cobe à croissant (Kobus ellipsiprymnus) semble être d’une manière générale en déclin. Les données sur les effectifs et la structure de la population font défaut pour de nombreuses parties de son aire de répartition, en particulier en Afrique de l'Ouest, où se trouve la sous‐espèce defassa. Le but de la présente étude était d'évaluer l'abondance, la répartition et les attributs des populations de cobes à croissant de la réserve forestière de Nazinga, au sud du Burkina Faso. Nous avons étudié les tendances de la population de cobes à croissant dans le parc à l'aide des données des transects recueillies en 1985 et 2019. Pour les analyses plus détaillées de la structure de la population et de la répartition des animaux, nous avons utilisé les données du recensement recueillies en 2019. La plupart des animaux étaient des adultes (46.6%) et le rapport des sexes observé a fait apparaître une forte proportion de femelles (5:1). La plupart des animaux étaient concentrés le long des grandes rivières. Aucune influence de l’activité des braconniers sur la répartition des cobes à croissant n’a été constatée. Sur le long terme (1985–2019), la dynamique de la population de cobes à croissant peut être approximativement divisée en deux périodes principales : une phase d'augmentation de la population de 1985 à 2005 et une phase de diminution continue de la population de 2007 à 2019. Bien que la tendance à la baisse de la population était évidente, les coefficients de détermination étaient faibles, indiquant que les années ne constituaient pas une explication suffisante du nombre d'individus et du nombre d'observations obtenues. Le nombre de cobes à croissant présents dans la réserve forestière de Nazinga est en baisse, mais nous n'avons trouvé aucune raison expliquant cette tendance. Il est probable qu'une combinaison de facteurs, notamment le réchauffement climatique (aridité accrue) et des activités illégales telles que le braconnage, en soit la cause. Comme il existe probablement de multiples raisons expliquant le déclin de la population de cobes à croissant observé dans notre zone d'étude, nous suggérons qu'une approche à multiples facettes soit adoptée afin d'améliorer l'état de conservation des populations locales de cobes à croissant.
... Isotope technique is also able to detect subtle differences in diet between family groups and/or herds, even within one fairly small reserve and a small sample size (Codron & Brink 2007). Moreover Djagoun et al. (2013) recently found a close similarity in waterbuck and kob diet from Pendjari Biosphere Reserve and East and southern Africa studies, although one was based on the faeces representing only short-term dietary information and the other one on longer-term body tissues like hair, bones and teeth. The distribution pattern of waterbuck was predicted both in the dry and wet season by grass and canopy cover, which is in accordance with the diet preference found for this species in the PBR. ...
Niche theory suggests differential use of shared resources facilitates coexistence of species in a community. In this study we used the faecal stable isotope analysis with observations along transect lines perpendicular to the Pendjari River. This was to examine seasonal habitat features and diet partitioning between two sympatric bovid species waterbuck (Kobus ellipsiprymnus) and western kob (Kobus kob) in the Pendjari Biosphere Reserve. In support of niche partition hypothesis, diets of western kob and waterbuck diverged significantly along both faecal selection axes (δ¹³C and δ¹⁵N) during the resource-limited period of the dry season as opposed to wet season when there is resource abundance. Western kob and waterbuck resource partitioning does not occur only on the basis of diet segregation but also some habitat variables play an important role in the coexisting system. Findings support the niche partition hypothesis, where morphologically, ecologically and closely related sympatric species segregate at least in one of the niche dimensions to allow coexistence. The two bovid species were observed to diverge largely along distance to water source gradient. The results provided empirical evidence that habitat features acts as an additional dimension over which herbivores partition resources.
... For example, Robbins et al. (1995) found no significant difference in the ability of browsers and grazers to digest fibre, whereas Van Wieren (1996) used a larger dataset to show that grazers digest fibre more efficiently than browsers. Second, most studies make use of average data on body mass and diet type; where these averages are taken from too great a geographic range, the results may be misleading (Djagoun et al. 2013). Third, where data come from captive animals or feeding trials they may not be appropriate to a freeranging context (Van Wieren 1996; Pérez-Barbería et al. 2004). ...
Many theories attempt to explain patterns of community organisation among large herbivores. We explored the role of body size, diet type and residence time on habitat use in a community comprising four metatherians (western grey kangaroo, Macropus fuliginosus; eastern grey kangaroo, M. giganteus; red-necked wallaby, Notamacropus rufogriseus; swamp wallaby, Wallabia bicolor) and two eutherians (red deer, Cervus elaphus; European rabbit, Oryctolagus cuniculus) in south-eastern Australia. We used camera traps to estimate habitat occupancy, quantified habitat specialisation using relative entropy, and ran regressions using percentage grass consumed, log(mass) and log(time at site) as predictor variables and relative entropy as the response. If body size influenced habitat use, we predicted smaller species would occupy fewer habitats. If diet type influenced habitat use, we predicted intermediate feeders would use more habitats. If the time that a species had been present at a site predicted community organisation, newer species would use more habitats. None of these theories explained habitat use in our community. Red deer used a narrower range of habitats than expected, perhaps due to the poor suitability of habitats available in the Grampians. While interactions between our hypotheses are likely to be important, the body size model deserves further attention in this community.
... In African savannas and other subtropical environments, stable carbon isotopes track herbivore feeding styles, because of the well-known bimodal distribution of carbon isotope compositions amongst C 3 (mainly browse) and C 4 (grass) vegetation (Vogel, 1978;Vogel et al., 1978;Tieszen et al., 1979;Cerling and Harris, 1999). Thus carbon isotopes have been successfully used to test hypotheses about browsing and grazing in African savannas across taxa, across populations through space and time, and through fossil lineages over evolutionary time (Tieszen et al., 1989;Koch et al., 1995;Cerling et al., 2003Cerling et al., , 2015Sponheimer et al., 2003a;Botha and Stock, 2005;Codron et al., 2006Codron et al., , 2008aCodron et al., , 2011bCodron et al., , 2012bDjagoun et al., 2013;Lehmann et al., 2013;Radloff et al., 2013). Large mammal carnivores have more complex diets, the composition of which is determined largely by predator-prey body size relationships (Carbone et al., 1999(Carbone et al., , 2007Gervasi et al., 2015). ...
Full-text available
Large mammal ecosystems have relatively simple food webs, usually comprising three—and sometimes only two—trophic links. Since many syntopic species from the same trophic level therefore share resources, dietary niche partitioning features prominently within these systems. In African and other subtropical savannas, stable carbon isotopes readily distinguish between herbivore species for which foliage and other parts of dicot plants (13C-depleted C3 vegetation) are the primary resource (browsers) and those for which grasses (13C-enriched C4 vegetation) are staples (grazers). Similarly, carbon isotopes distinguish between carnivore diets that may be richer in either browser, grazer, or intermediate-feeding prey. Here, we investigate levels of carbon and nitrogen isotopic niche variation and niche partitioning within populations (or species) of carnivores and herbivores from South African savannas. We emphasize predictable differences in within-population trends across trophic levels: we expect that herbivore populations, which require more foraging effort due to higher intake requirements, are far less likely to display within-population resource partitioning than carnivore populations. Our results reveal generally narrower isotopic niche breadths in herbivore than carnivore populations, but more importantly we find lower levels of isotopic differentiation across individuals within herbivore species. While these results offer some support for our general hypothesis, the current paucity of isotopic data for African carnivores limits our ability to test the complete set of predictions arising from our hypothesis. Nevertheless, given the different ecological and ecophysiological constraints to foraging behavior within each trophic level, comparisons across carnivores, and herbivores, which are possible within such simplified foodwebs, make these systems ideal for developing a process-based understanding of conditions underlying the evolution of intra-specific, individual-level separation of ecological niches.
... However, the pellets of certain bovid species may be misidentified, particularly between hartebeest and topi pellets, and between reedbuck and bushbuck (Hibert et al. 2008(Hibert et al. , 2011. Therefore, we used additional information such as hoof prints, animal vocalizations, and when possible, direct observations to enhance accuracy (Djagoun et al. 2013). The habitat type was recorded for each sampling quadrat. ...
Full-text available
Multiple land uses including tourism, hunting, and agriculture around protected areas can be a serious complication for wildlife management. We calculated habitat selection indices (Manly's alpha) for 10 bovid species in the Pendjari Biosphere Reserve in Benin, west Africa, to assess if habitat use differed in each bovid species between hunting and non-hunting zones. Presence/absence data was used in resource-selection functions based on a generalized linear mixed effect model to examine factors that explained bovid species distribution. We observed stronger avoidance of open habitat types in the hunting zone than in the non hunting zone for the hartebeest Alcelaphus buselaphus, oribi Ourebia ourebi, roan Hippotragus equines, kob Kobus kob, waterbuck Kobus ellipsiprymnus defassa and reedbuck Redunca redunca. In contrast, in grey duiker Sylvicapra grimmia, red-flanked duiker Cephalophus rufilatus, bushbuck Tragelaphus scriptus and buffalo Syncerus caffer we found no differences in habitat use between hunted and non-hunted areas. This may indicate that the latter species show more pronounced ecological and behavioural plasticity. Further, resource selection of bovid species on a small scale was influenced by other factors such as habitat structure, landscape characteristics, and human disturbance. This preliminary assessment of bovid habitat relationships in west Africa suggests that human hunting activities may cause species to alter their habitat selection. We therefore suggest habitat models may need to incorporate this source of variation if they are to accurately predict habitat use or distribution of a species.
Full-text available
Prehistoric and recent extinctions of large-bodied terrestrial herbivores had significant and lasting impacts on Earth’s ecosystems due to the loss of their distinct trait combinations. The world’s surviving large-bodied avian and mammalian herbivores remain among the most threatened taxa. As such, a greater understanding of the ecological impacts of large herbivore losses is increasingly important. However, comprehensive and ecologically-relevant trait datasets for extinct and extant herbivores are lacking. Here, we present HerbiTraits , a comprehensive functional trait dataset for all late Quaternary terrestrial avian and mammalian herbivores ≥10 kg (545 species). HerbiTraits includes key traits that influence how herbivores interact with ecosystems, namely body mass, diet, fermentation type, habitat use, and limb morphology. Trait data were compiled from 557 sources and comprise the best available knowledge on late Quaternary large-bodied herbivores. HerbiTraits provides a tool for the analysis of herbivore functional diversity both past and present and its effects on Earth’s ecosystems.
Stable isotope analyses provide insights into large herbivore diets. Carbon isotopes are particularly useful for analysing diet given their ability to reflect C3 or C4 photosynthetic pathways and, therefore, distinguish between browser and grazer diets. Although isotopes are useful for comparing diets across a range of spatial and temporal scales, most studies have focused on either interspecific differences within habitats or single species averages across their geographic range. In this study, we use both new and existing carbon isotopic data to compare large mammal herbivore diets across habitat types in southern and east Africa. We examined whether species in the grassland biome of central South Africa had the same carbon isotopic signatures as counterparts from savannas in southern and east Africa. Within the grassland biome, isotopic signatures from faecal samples were strongly correlated with signatures from bone and dentine collagen samples, supporting our approach of comparing faecal data from the grasslands with savanna data obtained mainly through analysis of skeletal materials. Regression analysis comparing grasslands and savannas confirmed that species' isotopic niches within communities were conserved across different geographic regions. However, herbivore δ¹³C values were generally lower in the grassland compared with the savanna biome, highlighted particularly by the lower δ¹³C values of several grazer species (but not browsers), indicating greater consumption of C3 vegetation in grasslands. This finding is counter-intuitive and suggests that non-woody C3 vegetation may be an important component of grazer diets in the grassland biome. We speculate on various environmental factors, related to food abundance that could underline the pattern.
Full-text available
Diet quality of Aepyceros melampus and Tragelaphus angasi was determined in Mkuzi and Ndumu Game Reserves, Natal, respectively. Dicotyledon content of both species' diets varied seasonally and reached a peak during the dry season at 52.0% for impala whereas nyala diets contained significantly more dicotyledons at 83.2%. During this time diet quality declined as reflected by higher crude fibre content and a decline in dietary protein. After onset of the rainy season both species changed to a diet of predominantly monocotyledons, 86.9% and 81.5% for impala and nyala respectively. Diet quality improved during this period as protein content increased and fibre decreased. -Author
Full-text available
Diets of hartebeest (Alcelaphus buselaphus) and roan antelope (Hippotragus equinus) were assessed at the Nazinga Game Ranch in southern Burkina Faso, West Africa. Microhistological analysis of feces indicated that dietary overlap was high during the rainy (X̄ = 73.7%) and cool-dry (68.2%) seasons, low during the hot-dry season (48.2%), and lowest during the last month of the hot-dry season (31.5%). As the hot-dry season progressed and food presumably became less available, diets of the two antelopes diverged. Hartebeest maintained a high percentage of grass in their diet, but roan antelope switched from being predominantly grazers (>95% grass) to mixed feeders (<50% grass). As grass feeders, both antelopes have skeletal features that facilitate acquisition and grinding of highly fibrous diets, but 11 of 12 mass-relative indices of the skull morphology of hartebeest exceeded those of roan antelope. Because of those differences in skull morphology, and in keeping with the "long-faced" hypothesis, hartebeest were apparently more capable than roan antelope of acquiring and masticating scarce regrowth of perennial grasses when availability of forage was lowest. Such divergence within a single foraging class of African bovids, such as grass feeders, should reduce competition and perpetuate coexistence.
Full-text available
The carbon-isotope composition of hair and feces offers a glimpse into the diets of mammalian herbivores. It is particularly useful for determining the relative consumption of browse and graze in tropical environments, as these foods have strongly divergent carbon-isotope compositions. Fecal δ13C values reflect the last few days consumption, whereas hair provides longer term dietary information. Previous studies have shown, however, that some fractionation occurs between dietary δ13C values and those of hair and feces. Journal Article
Dietary and nutritional distinctions between browsing and grazing ruminants are reviewed and related to feeding types identified by Hofmann and Stewart (1972) from the morphology of the digestive tract. African ruminants show a clear distinction in terms of graminoid-dicot proportions in the diet, while among ruminants inhabiting temperate regions few exclusive grazers are represented. Among browsers, the duikers (Cephalophinae) represent a distinct frugi-folivore category. Most browsers consume a mix of woody plant leaves and shoots and forbs, and include fruits when available. Grazers consume certain pods but not succulent fruits. The diets of browsers are higher in protein and lower in cell wall constituents than those of similar-sized grazers for most of the year, but metabolizable energy yield may be similar for both feeding types. Characterizing the staple food types of larger browsers as 'concentrates' is misleading. Browsers consumed higher levels of tannins and other allelochemicals than grazers. The fibrous cell walls of grasses, particularly those with the C4 photosynthetic pathway, restrict the effective use of this food type by browsers. Grazers are limited in their capacity to cope with the allelochemicals prevalent in dicots. Large salivary glands producing tannin-complexing secretions seem typical of frugifolivores that may consume substantial amounts of unripe fruits.
1. The niche separation of three species of alcelaphine antelope (wildebeest, topi and hartebeest) with similar body size was compared by measuring bite weight, bite rate, intake rate and selectivity of tame animals in plots containing grass at different growth stages. 2. On growing swards, hartebeest had a smaller bite weight and lower intake rate, and were also less selective of green leaf, than either topi or wildebeest. On senescent swards, hartebeest were more selective of leaf than the other two species. 3. Wildebeest had a faster bite rate than either topi or hartebeest on swards with low biomass and high protein content of green leaf (green flush). Bite weight and intake rate of wildebeest and topi were similar despite the difference in breadth of their incisor rows. 4. Topi were significantly more selective of green leaf than the other two species and were the only species to maintain a rapid bite rate on swards with high green leaf biomass. 5. The feeding experiments did not reveal significant cross-overs between species in the rate of food intake on different grass types, but each species was most proficient either in leaf selection or bite rate when feeding on grass swards in a particular growth stage. We suggest that growth stage is a primary determinant of niche separation. 6. In Serengeti, grazing ungulates which migrate are specialists of the earlier growth stages of grass which tend to be transient, while those that are residential specialize on late growth stages which are more enduring. The mobility of species, and the spatial and temporal dynamics of pastures containing different growth stages of grass, contribute to niche separation.