Manual 21386829

Manual 21386829
Ecological separation of the black and blue wildebeest on Ezemvelo
Nature Reserve in the highveld grasslands of South Africa
by
Chantal Vinisia Helm
Submitted in partial fulfilment of the requirements for the degree
MAGISTER SCIENTIAE (WILDLIFE MANAGEMENT)
in the Centre for Wildlife Management
Department of Animal and Wildlife Sciences
Faculty of Natural and Agricultural Sciences
University of Pretoria
Pretoria
Supervisor: Prof. J. du P. Bothma
Co-supervisor: Prof. M.W. van Rooyen
December 2006
(The financial assitance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions
expressed and conclusions arrived at, are those of the author and are not necessarily to be attributed to the NRF).
I declare that the dissertation, which I hereby submit for the degree Master of
Science (Wildlife Management) at the University of Pretoria, is my own work and has
not previously been submitted by me for a degree at this or any other tertiary
institution.
SIGNATURE: ..…………………………….
DATE: ………………………………….
ii
Ecological separation of the black and blue wildebeest on Ezemvelo Nature
Reserve in the highveld grasslands of South Africa
by
Chantal Vinisia Helm
Supervisor: Prof. J. du P. Bothma
Co-supervisor: Prof. M.W. van Rooyen
in the Centre for Wildlife Management
Department of Animals and Wildlife Sciences
Faculty of Natural and Agricultural Sciences
University of Pretoria
Pretoria
MAGISTER SCIENTIAE (WILDLIFE MANAGEMENT)
ABSTRACT
The present study was conducted on Ezemvelo Nature Reserve in the highveld
grasslands of South Africa. The evidence for ecological separation between the black
and blue wildebeest was investigated in an area with suboptimal habitat for both
types of wildebeest. Habitat selection and separation of the black and blue
wildebeest population were investigated at three main scales. A combination of
logistic regression analysis, discriminant analysis and hypothesis testing techniques
were used to determine whether habitat separation occurred between the black and
blue wildebeest at the various scales. Seasonal, social group and weather influences
on the habitat selection of both types of wildebeest were also investigated. Black and
blue wildebeest showed resource partitioning in terms of habitat at the macro and
mesoscales but not at the microscale. The preference for open areas by the black
wildebeest and its more specialised territoriality were found to be the main driving
factors contributing to the habitat separation of the two types of wildebeest. The
population of black wildebeest was found to be decreasing while the blue wildebeest
population was found to be increasing in the study area during the study period.
Spatial overlap between the black and blue wildebeest was found to be low. Little
evidence of interference interspecific competition between the black and blue
iii
wildebeest was found. It was, however, concluded that exploitative competition
between the two types of wildebeest would be found in areas with low habitat
heterogeneity. Ecological separation between the black and blue wildebeest was
found to be incomplete. However, the coexistence of the black and blue wildebeest
was deemed possible provided habitat heterogeneity in terms of the factors found to
be important for habitat separation was high and population sizes were strictly
monitored and actively controlled. Finally, a number of additional management
recommendations for the black and blue wildebeest at Ezemvelo Nature Reserve
and for other reserves confining both types of wildebeest together based on the
results of this study were made.
iv
LIST OF CONTENTS
ABSTRACT.............................................................................................................. iii
LIST OF CONTENTS ................................................................................................ v
LIST OF FIGURES.................................................................................................. vii
LIST OF TABLES .................................................................................................. xiii
CHAPTER 1: INTRODUCTION................................................................................. 1
CHAPTER 2: THE STUDY AREA ............................................................................. 6
INTRODUCTION ................................................................................................... 6
LOCALITY ............................................................................................................. 6
GEOLOGY AND SOILS ....................................................................................... 10
CLIMATE ............................................................................................................. 12
SEASONAL DIVISIONS....................................................................................... 17
VEGETATION...................................................................................................... 18
ANIMALS ............................................................................................................. 19
RESERVE HISTORY ........................................................................................... 19
HISTORY OF THE BLACK AND BLUE WILDEBEEST POPULATIONS .............. 23
CHAPTER 3: THE BLACK AND BLUE WILDEBEEST .......................................... 27
TAXONOMY OF THE GENUS CONNOCHAETES .............................................. 27
DESCRIPTION OF THE CONNOCHAETES SPECIES ....................................... 27
HYBRIDISATION ................................................................................................. 38
CONSERVATION ................................................................................................ 43
CHAPTER 4: HABITAT SELECTION AND SEPARATION: GENERAL
METHODOLOGY .................................................................................................... 52
INTRODUCTION ................................................................................................. 52
METHODS........................................................................................................... 55
CHAPTER 5: HABITAT SELECTION AND SEPARATION: MACROHABITAT
SCALE .................................................................................................................... 65
INTRODUCTION ................................................................................................. 65
MATERIALS AND METHODS ............................................................................. 68
RESULTS ............................................................................................................ 77
DISCUSSION..................................................................................................... 103
CONCLUSION ................................................................................................... 106
CHAPTER 6:HABITAT SELECTION AND SEPARATION: MESOHABITAT SCALE
.............................................................................................................................. 107
INTRODUCTION ............................................................................................... 107
METHODS......................................................................................................... 108
RESULTS .......................................................................................................... 112
DISCUSSION..................................................................................................... 150
CONCLUSION ................................................................................................... 156
CHAPTER 7: HABITAT SELECTION AND SEPARATION: MICRO-HABITAT
SCALE .................................................................................................................. 157
INTRODUCTION ............................................................................................... 157
METHODS......................................................................................................... 159
RESULTS .......................................................................................................... 161
DISCUSSION..................................................................................................... 165
CONCLUSION ................................................................................................... 167
v
CHAPTER 8: ACTIVITY BUDGETS...................................................................... 168
INTRODUCTION ............................................................................................... 168
METHODS......................................................................................................... 169
RESULTS .......................................................................................................... 171
DISCUSSION..................................................................................................... 185
CONCLUSION ................................................................................................... 190
CHAPTER 9: NICHE BREADTH, OVERLAP AND EXPLOITATIVE INTERSPECIFIC
COMPETITION...................................................................................................... 191
INTRODUCTION ............................................................................................... 191
METHODS......................................................................................................... 194
RESULTS .......................................................................................................... 196
DISCUSSION..................................................................................................... 203
CONCLUSION ................................................................................................... 210
CHAPTER 10: BEHAVIOURAL INTERACTIONS, SPECIES ASSOCIATIONS AND
INTERSPECIFIC INTERFERENCE COMPETITION ............................................. 211
INTRODUCTION ............................................................................................... 211
METHODS......................................................................................................... 212
RESULTS .......................................................................................................... 213
DISCUSSION..................................................................................................... 220
CONCLUSIONS................................................................................................. 223
CHAPTER 11: GRAZING CAPACITY AND STOCKING DENSITY ...................... 224
INTRODUCTION ............................................................................................... 224
METHODS......................................................................................................... 225
RESULTS .......................................................................................................... 229
DISCUSSION..................................................................................................... 231
CONCLUSION ................................................................................................... 235
CHAPTER 12: POPULATION DYNAMICS ........................................................... 236
INTRODUCTION ............................................................................................... 236
METHODS......................................................................................................... 237
RESULTS .......................................................................................................... 238
DISCUSSION..................................................................................................... 246
CONCLUSION ................................................................................................... 250
CHAPTER 13: MANAGEMENT IMPLICATIONS .................................................. 252
INTRODUCTION ............................................................................................... 252
MANAGEMENT OBJECTIVES .......................................................................... 253
ACTIVE ADAPTIVE MANAGEMENT ................................................................. 253
MONITORING.................................................................................................... 253
SPECIFIC MANAGEMENT RECOMMENDATIONS .......................................... 260
CHAPTER 14: CONCLUSIONS............................................................................ 268
FUTURE RESEARCH PERSPECTIVES............................................................ 271
PREDICTIONS FOR THE FUTURE OF THE BLACK AND BLUE WILDEBEEST IN
SOUTH AFRICA ................................................................................................ 272
SUMMARY............................................................................................................ 273
ACKNOWLEDGEMENTS ..................................................................................... 276
REFERENCES ...................................................................................................... 278
APPENDICES ....................................................................................................... 298
APPENDIX 1...................................................................................................... 298
APPENDIX 2...................................................................................................... 299
vi
LIST OF FIGURES
CHAPTER 2
Figure 2.1: The location and boundaries of Ezemvelo Nature Reserve in the
Grassland Biome of South Africa, indicating its position (star) on the
border between the Gauteng and Mpumalanga provinces (Adapted from
the map of Low and Rebelo 1996)……………………………………………7
Figure 2.2: The topography and drainage at Ezemvelo Nature Reserve, South
Africa. Adapted from the topographical mapsheets 2528DB and 2529CA
(Government Printer 1996 and 1998)………………………………………..9
Figure 2.3: The geology at Ezemvelo Nature Reserve, South Africa. Adapted from
the 1:250 000 geological series map sheet 2528 Pretoria (Geological
survey staff 1978)……………………………………………………………..11
Figure 2.4: Climate diagram for rainfall (mm) and temperature (°C) as determined
following Walter’s convention (Walter 1979) from July to June based on
data obtained from the Witbank Weather Station from 1993 to 2003…...13
Figure 2.5: Long-term total annual rainfall (mm) for the Bronkhorstspruit Weather
Station (0514408X), closest to Ezemvelo Nature Reserve in the Gauteng
Province of South Africa from 1970 to 2003………………………………..15
Figure 2.6: The actual total monthly rainfall (mm) received during the study period
(January 2004 to August 2005) at Ezemvelo Nature Reserve, South
Africa……………………………………………………………………………16
Figure 2.7: The distribution of the various sections purchased since 1974 by the
Oppenheimer family to form the current Ezemvelo Nature Reserve, South
Africa. The numbers on the map indicate Section 1 to 6 as described in
the text……………………………………………………………………….....21
Figure 2.8: The number of black and blue wildebeest on the Telperion Nature
Reserve from 1991 to 2001………………………………………………….24
Figure 2.9: The number of black wildebeest on the eZemvelo section of Ezemvelo
Nature Reserve from 1998 to 2001…………………………………………25
CHAPTER 3
Figure 3.1: Current distribution of the black wildebeest in South Africa. Adapted from
Friedmann and Daly (2004)………………………………………………….30
vii
Figure 3.2: Current distribution of the blue wildebeest in South Africa. Adapted from
Friedmann and Daly (2004)……………………………………………….….31
CHAPTER 5
Figure 5.1: The broad habitat types found at Ezemvelo Nature Reserve, South
Africa……………………………………………………………………………69
Figure 5.2: The distribution of the black and blue wildebeest at Ezemvelo Nature
Reserve, South Africa from January 2004 to August 2005……………….72
Figure 5.3: Mean percentage composition of the five ecological classes of grass
species in the five broad habitats that were utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve in 2004. Bars represent the
standard error of the percentage composition. No significant differences
between habitats were found. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy
grasslands……………………………………………………………………...88
Figure 5.4: Mean (columns) and standard error (bars) of the degree of past
utilisation of the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve in 2004 as calculated from the
percentage composition of the five ecological classes of grass species
present within these habitats. No significant differences between habitats
were found. BW = Burkea woodlands, RG = rocky grasslands, OL = old
lands,
MG
=
moist
grasslands,
SG
=
sandy
grasslands…………………………………………………………………..….89
Figure 5.5: Mean (columns) and standard error (bars) of the veld condition score of
the five broad habitats utilised by the black and blue wildebeest at
Ezemvelo Nature Reserve in 2004 as calculated from the percentage
composition of the five ecological classes of grass species present within
these habitats. No significant differences between habitats were found.
BW = Burkea woodlands, RG = rocky grasslands, OL = old lands, MG =
moist grasslands, SG = sandy grasslands………………………………….91
Figure 5.6: Mean (columns) and standard error (bars) of the Shannon-Wiener
diversity index of the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve in 2004 as calculated from the
percentage composition grass species present within these habitats.
Means with the same superscripts were not significantly different from
each other. BW = Burkea woodlands, RG = rocky grasslands, OL = old
viii
lands,
MG
=
moist
grasslands,
SG
=
sandy
grasslands……………………………………………………………...………92
Figure 5.7: Mean (columns) and standard error (bars) of the plant species density of
the five broad habitats utilised by the black and blue wildebeest at
Ezemvelo Nature Reserve in 2004 as calculated from the percentage
composition of the five ecological classes of grass species present within
these habitats. Means with the same superscripts were not significantly
different from each other. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy
grasslands…………………………………………………………………...…93
Figure 5.8: Mean (columns) and standard error (bars) of the percentage bare ground
of the five broad habitats utilised by the black and blue wildebeest at
Ezemvelo Nature Reserve in 2004. Means with the same superscripts
were not significantly different from each other. BW = Burkea woodlands,
RG = rocky grasslands, OL = old lands, MG = moist grasslands, SG =
sandy grasslands……………………………………………………………...94
Figure 5.9: Mean (columns) and standard error (bars) of the percentage canopy
cover of the herbaceous layer of the five broad habitats utilised by the
black and blue wildebeest at Ezemvelo Nature Reserve over the three
ecological seasons in 2004. Means with the same superscripts were not
significantly different from each other; these are compared across the five
habitats within each season. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy
grasslands……………………………………………………………………..95
Figure 5.10: Mean (columns) and standard error (bars) of the total grass height of
the herbaceous layer in the five broad habitats utilised by the black and
blue wildebeest at Ezemvelo Nature Reserve over the three ecological
seasons in 2004. No significant differences were found between the
means of the different habitat types. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy
grasslands……………………………………………………………….…….97
Figure 5.11: Mean (columns) and standard error (bars) of the grass leaf height of the
herbaceous layer in the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve over the three ecological
seasons in 2004. No significant differences between categories were
found. BW = Burkea woodlands, RG = rocky grasslands, OL = old lands,
MG = moist grasslands, SG = sandy grasslands………………………….98
ix
Figure 5.12: Mean (columns) and standard error (bars) of the grass biomass of the
herbaceous layer in the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve over the three ecological
seasons in 2004. No significant differences between categories were
found. BW = Burkea woodlands, RG = rocky grasslands, OL = old lands,
MG = moist grasslands, SG = sandy grasslands…………………………99
Figure 5.13: Mean (columns) and standard error (bars) of the biomass concentration
of the herbaceous layer in the five broad habitats utilised by the black and
blue wildebeest at Ezemvelo Nature Reserve over the three ecological
seasons in 2004. Superscripts that are the same indicate no significant
differences. BW = Burkea woodlands, RG = rocky grasslands, OL = old
lands, MG = moist grasslands, SG = sandy grasslands……………..….100
CHAPTER 8
Figure 8.1: Daily time budgets for the black and blue wildebeest for the entire study
period at Ezemvelo Nature Reserve when combining all the social groups
and age classes……………………………………………………………...172
Figure 8.2: Daily time budgets (percentage of time spent) for adult females, territorial
bulls and calves of the black and blue wildebeest for the entire study
period at Ezemvelo Nature Reserve………………………………………174
Figure 8.3: Seasonal daily time budgets (percentage time spent) for black and blue
wildebeest at Ezemvelo Nature Reserve from January 2004 to August
2005. G=grazing, S=standing, L= Lying down, W=walking, O=other. *
indicates a significant difference between the black and blue wildebeest
within that season for that activity………………………………………….176
Figure 8.4: Diurnal activity patterns (percentage of time spent) by the black and blue
wildebeest for the entire study period at Ezemvelo Nature Reserve…..178
Figure 8.5: Comparison of the diurnal behavioural patterns of the black and blue
wildebeest expressed as a percentage of the time spent for the three time
periods in the daytime for the entire study period at Ezemvelo Nature
Reserve. * indicates a significant difference between the black and blue
ZLOGHEHHVWIRUWKDWDFWLYLW\DQGWLPHRIGD\ZLWK
.
*
*UD]LQJ6
Standing; L = Lying down; W = Walking; O = Other……………..………180
x
Figure 8.6: Diurnal activity patterns expressed as a percentage of time spent, of the
black and blue wildebeest for the late growing season at Ezemvelo Nature
Reserve……………………………………………………………………….181
Figure 8.7: Diurnal activity patterns, expressed as a percentage of time spent, of the
black and blue wildebeest for the dormant season at Ezemvelo Nature
Reserve…………………………………………………………………….…182
Figure 8.8: Diurnal activity patterns, expressed as a percentage of time spent, of the
black and blue wildebeest for the early growing season at Ezemvelo
Nature Reserve………………………………………………………………183
Figure 8.9: Periods of activity (grazing and walking) and rest (lying down and
standing), expressed as percentage of time spent, for the black and blue
wildebeest for the entire study period and for each ecological season at
Ezemvelo Nature Reserve. * Indicates significant differences between the
black and blue wildebeest within the seasons and periods of activity….186
CHAPTER 9
Figure 9.1: Percentage frequency of occurrence of the dominant plant species at the
sites of occupation of the black and blue wildebeest at Ezemvelo Nature
Reserve from January 2004 to August 2005. Data obtained during the
habitat survey collection period (Chapter 4)……………………………....204
Figure 9.2: Species composition at the feeding sites of the black and blue wildebeest
at Ezemvelo Nature Reserve for all plant species contributing more than
2% to the overall species composition in the feeding sites sampled in
March 2004. Data obtained from vegetation surveys in the feeding sites of
the black and blue wildebeest (Chapter 6)………………………………..205
CHAPTER 10
Figure 10.1: Diurnal patterns, expressed as percentage time spent, of activity
(grazing and walking) for black and blue wildebeest based on scan
samples of activity taken at 5-minute intervals throughout the daytime at
Ezemvelo Nature Reserve from March 2004 to August 2005………….215
Figure 10.2: Seasonal diurnal patterns of activity (grazing and walking), expressed
as the percentage of time spent being active for the black and blue
wildebeest based on scan samples of activity taken at 5-minute intervals
xi
throughout the daytime at Ezemvelo Nature Reserve from March 2004 to
August 2005………………………………………………………………….216
Figure 10.3: The association of the black and blue wildebeest with other types of
wildlife at Ezemvelo Nature Reserve as observed from January 2004 to
August 2005. * Black and or blue wildebeest depending on the type of
wildebeest under analysis…………………………………………………..218
Figure 10.4: The association of the black and blue wildebeest with other types of
wildlife at Ezemvelo Nature Reserve for the three ecological seasons as
observed from January 2004 to August 2005…………………………….219
CHAPTER 12
Figure 12.1: Population trends of the black and blue wildebeest at Ezemvelo Nature
Reserve as obtained from monthly counts from May 2003 to August
2005……………………………………………………………………..…….241
Figure 12.2: The broad population structure of the black and blue wildebeest
populations at Ezemvelo Nature Reserve, South Africa for the period
January 2004 to August 2005……………………………………………...244
Figure 12.3: Mean, minimum and maximum herd sizes of black and blue wildebeest
over three seasons on Ezemvelo Nature Reserve from January 2004 to
August 2005. The bars on the mean columns represent the standard
errors which are small due to the large sample sizes that were used for
these calculations……………………………………………………………245
xii
LIST OF TABLES
CHAPTER 2
Table 2.1: Large mammals found at Ezemvelo Nature Reserve……………………..20
CHAPTER 3
Table 3.1: Comparison of the morphological features of the hybrids and pure types of
wildebeest as studied by Fabricius et al. (1988)…………………………...41
Table 3.2: A list of the problems and their proposed solutions for the black and blue
wildebeest hybridisation problem in South Africa as stipulated in a
workshop on black wildebeest hybridisation held in June 2003 at the
Florisband Quaternary Research Station in the Free State province (Anon
2003a)……………………………………………………………………..……49
CHAPTER 5
Table 5.1: The percentage occurrence of the black and blue wildebeest in the five
broad habitat types indicating the utilisation of the various habitat types
over three ecological seasons and for the total data at Ezemvelo Nature
Reserve from January 2004 to August 2005……………………………….78
Table 5.2: Chi-squared test results to evaluate the hypothesis that the black
wildebeest on Ezemvelo Nature Reserve used the available broad
habitats in proportion to their occurrence by surface area. Values in
brackets indicate sample sizes of <5 and therefore the Chi-squared test
results for these entries may be invalid. + indicates a positive selection, indicates a negative selection and 0 indicates random selection. N/a
indicates that that habitat type was not utilised at all………………………...79
Table 5.3: Chi-squared test results to evaluate the hypothesis that the blue
wildebeest on Ezemvelo Nature Reserve used the available broad
habitats in proportion to their occurrence by surface area (values in
brackets indicate sample sizes of <5 and therefore the chi-squared test
results for these entries may be invalid). + indicates a positive selection, indicates
a
negative
selection
and
0
indicates
random
selection………………………………………………………………..………80
xiii
Table 5.4: Bonferroni confidence intervals calculated to determine the seasonal
broad habitat selection by black and blue wildebeest on Ezemvelo Nature
Reserve, South Africa from January 2004 to August 2005 relative to the
total land surface area of the reserve. N/a indicates those habitats where
that type of wildebeest was never encountered.…..…………………...….81
Table 5.5: Summary of the characteristics of the five broad habitat types utilised by
black and blue wildebeest at Ezemvelo Nature Reserve in 2004.……..100
Table 5.6: Summary of the herbaceous characteristics of the five habitat types
utilised by the black and blue wildebeest at Ezemvelo Nature Reserve
over
the
three
ecological
seasons
in
2004.
BC
=
biomass
concentration…………………………………………………………...…….101
CHAPTER 6
Table 6.1: The percentage observations of black and blue wildebeest social groups
over the three ecological seasons at Ezemvelo Nature Reserve obtained
by using the methods described in Chapter 4 from January 2004 to
August 2005…………………………………………………………………..113
Table 6.2: Predictor variables for the various combinations of season, social
structure, activity, time of day and weather conditions used in the PROC
LOGISTIC procedure (SAS 8.01) to determine those variables that
separate the habitats used by the black and blue wildebeest on Ezemvelo
Nature Reserve. This analysis was based on 1 558 wildebeest
observations that were collected from January 2004 to August 2005….114
Table 6.3: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 1 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………...……………115
Table 6.4: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 2 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
xiv
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005………………………………………………………………..………….116
Table 6.5: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 3 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………...……………………117
Table 6.6: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 4 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005………………………………………………………………………...…121
Table 6.7: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 5 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………………….…..123
Table 6.8: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 6 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………………...……125
Table 6.9: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
xv
category of the variables selected by Model 7 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………...……………………127
Table 6.10: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 8 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………...…………....128
Table 6.11: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 9 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………………………...130
Table 6.12: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 10 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………………….……..132
Table 6.13: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 11 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
xvi
collected on Ezemvelo Nature Reserve from January 2004 to August
2005………………………………………………………...…………………134
Table 6.14: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 13 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………...……………136
Table 6.15: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 14 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………...……..………..138
Table 6.16: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 15 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………………...…..…..140
Table 6.17: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 16 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005……………………………………………………………...…………....142
Table 6.18: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK
category of the variables selected by Model 17 in the logistic regression
xvii
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………………………...144
Table 6.19: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 18 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………..………………146
Table 6.20: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 19 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………………………...147
Table 6.21: The percentage of the probabilities (indicating the presence of either a
black (0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each
category of the variables selected by Model 20 in the logistic regression
analysis to separate between the habitat of the black and blue wildebeest.
Percentages in bold indicate those categories that were significantly
selected by the one type of wildebeest over the other (p<0.05). Data
collected on Ezemvelo Nature Reserve from January 2004 to August
2005…………………………………………………………...…………..…..149
CHAPTER 7
Table 7.1: Mean and standard errors of the characteristics of the herbaceous layer of
the feeding sites of the black and blue wildebeest, and sites that were not
utilised by either of them, that were analysed to indicate differences in the
feeding sites of the black and blue wildebeest at Ezemvelo Nature
xviii
Reserve in April 2004. Bold values indicate a significant difference and
different superscripts denote significant differences between sites…….162
Table 7.2: Results of the discriminant function analysis performed to compare the
feeding sites of the black with those of the blue wildebeest, and to
compare these sites with sites which were not utilised by either type of
wildebeest at Ezemvelo Nature Reserve in April 2004…………………..164
CHAPTER 9
Table 9.1: The index of Levin (1968) of niche breadth for each type of wildebeest,
and that of Pianka (1973) of niche overlap for each season and social
group for the habitat type choices of the black and blue wildebeest at
Ezemvelo Nature Reserve from January 2004 to August 2005. The tvalues are based on the paired t-test……………………………...………198
Table 9.2: The index of Levin (1968) for niche breadth for each type of wildebeest for
each season and for the entire study period for the abiotic and biotic
habitat factor choices of the black and blue wildebeest at Ezemvelo
Nature Reserve from January 2004 to August 2005 obtained from the
habitat survey data (Chapter 4) and based on the paired t-test……..….199
Table 9.3: The index of Pianka (1973) of niche overlap for each season and for the
entire study period for the abiotic and biotic habitat factor choices of the
black and blue wildebeest at Ezemvelo Nature Reserve from January
2004 to August 2005 obtained from the habitat survey data
(Chapter 4)……………………………………………………………………201
Table 9.4: Summary of the indices of overlap and quotients of similarity in spatial
distributions, habitat type selection and diet between black and blue
wildebeest at Ezemvelo Nature Reserve from January 2004 to August
2005. Table adapted from Anthony and Smith (1977)……………….…..209
CHAPTER 10
Table 10.1: Summary of the behavioural interactions between the black and blue
wildebeest at Ezemvelo Nature Reserve as recorded from April 2004 to
August 2005. M = males; F = females; C = calves……………….………214
Table 10.2: Summary of the Chi-squared tests performed to evaluate the hypothesis
that black and blue wildebeest at Ezemvelo Nature Reserve associated
with other species in proportion to their occurrence. Values in brackets
xix
indicate association frequencies of <5 and therefore the Chi-squared test
may not be valid. + indicates a positive selection, 0 indicates random
selection, - indicates a negative selection…………………………………221
CHAPTER 11
Table 11.1: Veld condition index and ecological grazing capacity calculation in
Grazer Units (GU) for Ezemvelo Nature Reserve, a 8 500 ha grassland
reserve in South Africa, based on the condition of the vegetation in 2004
and calculated by using the methods that were described by Bothma et al.
2004………………………………………………………………………...…230
Table 11.2: Estimated current numbers of herbivore grazers after the calving season,
and stocking densities calculated for the herbivore grazers at Ezemvelo
Nature Reserve for December 2004……………………………………….232
CHAPTER 12
Table 12.1: Mean population statistics for the black and blue wildebeest at Ezemvelo
Nature Reserve in 2005………………………………………………..……240
Table 12.2: The ratio of females per male in herds of black and blue wildebeest at
Ezemvelo Nature Reserve over the three ecological seasons in
2005…………………………………………………………………………...242
Table 12.3: Population size and density of the black and blue wildebeest at
Ezemvelo Nature Reserve as calculated from the mean monthly count
data for the entire study period from January 2004 to August 2005…...248
xx
CHAPTER 1: INTRODUCTION
The term ecological separation has been defined as the partitioning of a natural
resource such as food, among two or more species, so that each species has access
to a different part of the resource (Chapman and Reiss 1995). Species in a
community where ecological separation has occurred will utilise all the available
habitats optimally. Resource partitioning is defined as the differential use by
organisms of resources such as food and space (Schoener 1974a). Therefore
resource partitioning is the mechanism by which ecological separation is achieved.
Ecological separation explains how species can coexist even though they may have
extensive overlap in their ecological requirements (May 1973). Competition has been
considered to be the major selective force causing differential use of resources.
Therefore competition is usually cited as being limited by ecological separation (Von
Holdt 1999). However, resource partitioning may also occur through processes such
as predation and different responses of species to environmental gradients
(Schoener 1986).
African herbivores have evolved as an integrated community, whereby the available
habitats are fully utilised. Therefore, species, which occur naturally in the same
geographical area, are ecologically separated (Riney 1982). However, the
introduction of wildlife into areas where they have not naturally occurred in the past
may upset this balance.
A study of the ecological separation between two species involves the quantification
of spatial distributions, habitat selection, temporal activity patterns and feeding habits
pertinent to the niche relationships of the two species (Pianka 1973; Anthony and
Smith 1977). Additional studies of the potential competition among the species of an
area may provide further insight into the ecological separation of the species under
consideration (Scogings et al. 1990).
For the last four decades, research on African ungulates has described ecological
separation by habitat choice (e.g.: Lamprey 1963; Hirst 1975, Engelbrecht 1986;
Scogings et al. 1990; Wentzel 1990; Weaver 1995; Dekker 1996; Von Holdt 1999),
by feeding ecology (De Wet 1988; Wentzel 1990; Von Holdt 1999) and by social
behaviour (Keast 1965). Most of these and similar studies on ecological separation
have concentrated on wildlife which have naturally evolved together and which tend
1
to have obvious differences in terms of physiology, morphology, behaviour and
ecological requirements.
Presently, with the large-scale increase in the number of wildlife ranches being
developed throughout South Africa, landowners are keeping a wide range of species
on their properties (Bothma 2002a). This is primarily being done to cater for local and
overseas hunters and ecotourists. Therefore, more and more properties are confining
ecologically similar wildlife within the same area. As a general rule in wildlife
management, ecologically similar taxa should not be confined in small areas
together. Wildlife that have the potential to hybridise will produce hybrids when the
area of confinement is too small and the minimum herd size is not maintained (Du
Toit et al. 2002).
The black wildebeest Connochaetes gnou and the blue wildebeest Connochaetes
taurinus subsp. taurinus are two such ecologically similar types of wildlife that are
currently being confined on the same properties in southern Africa. These two types
of wildebeest separated from a common ancestor just over 1 million years ago (Brink
et al. 1999). They therefore still have many morphological, physiological and
behavioural characteristics in common. In addition they are able to hybridise, as
reproductive isolation has not yet evolved (Fabricius et al. 1988), and the resulting
offspring are fertile. Such hybridisation can have serious implications for the
conservation of the two types of wildebeest and warrants urgent attention.
Many vital questions about the hybridisation process between the black and blue
wildebeest remain unanswered. Factors that lead to hybridisation are still poorly
understood and the ecological and behavioural differences between the two types of
wildebeest still require in-depth investigation (Anon 2003a). Further research is also
required to quantifiably define the ecological niche of the two types of wildebeest
found in South Africa (Vrahimis 2003a). Such a study would be important in
understanding whether there is any ecological separation between them. This
information would also enable researchers to pinpoint possible ecological factors that
can lead to hybridisation in areas where the two types occur together (Vrahimis
2003b).
Ecological separation, if it occurs between the two types of wildebeest, would be a
mechanism that could limit interspecific competition and also aid in minimising
hybridisation. If, however, the ecological requirements of both types of wildebeest do
2
not differ in terms of spatial distribution, habitat, diet and behaviour, they would be
considered to be too closely related ecologically to be kept in the same area without
harming each other or the habitat, and the possibility of hybridisation would be high.
Under natural conditions, different habitat preferences are probably the main
mechanisms that are ecologically separating the black and blue wildebeest (Codron
and Brink In press). Factors that reduce the extent of contact between these two
types of wildebeest, such as large areas with enough habitat heterogeneity to provide
suitable, but separate, habitats for both types of wildebeest, or with distinctive
geographic barriers, may assist in reducing, but not necessarily eliminating, the
hybridisation risk (Vrahimis 2003a). Although areas may look structurally and
compositionally homogeneous, the two types of wildebeest may prefer them
differentially. Subtle differences in specific vegetation parameters such as grass
height, plant biomass, veld condition, species composition and grass canopy cover
may be responsible for such differential choice. However, this may not occur when
they are both confined to an area where habitat diversity is low.
Both the extant types of wildebeest in South Africa have similar mating and calving
seasons as well as a fairly similar social organisation. It has been suggested that the
width of the mouth determines the level of selection for high-quality food items
(Owen-Smith 1982) and therefore different ungulate species may be more proficient
when feeding in grass swards of different growth stages (Murray and Brown 1993).
Morphological observations do not support a trophic difference between the two
types of wildebeest as there is no statistical difference in the width of the premaxillae
(Brink et al. 1999). Measurements of the premaxillae show a mean width of 72.3 mm
in the blue wildebeest and one of 74.9 mm in the black wildebeest (Roberts 1951).
This suggests that grasses will be cropped at the same height and in essentially a
similar way by the black and the blue wildebeest. Both are also specialised grazers of
short grasses. It is, therefore, unlikely that there will be any separation in terms of
their feeding height.
Fossil evidence suggests that the morphological traits that are associated with the
distinct territorial social behaviour of black wildebeest were the first to change,
indicating that a shift in breeding behaviour (especially territoriality) accompanied the
appearance of the first ancestral black wildebeest (Brink et al. 1999.). A shift to a
more territorial behaviour is linked to the evolution of treeless grasslands in the
central interior of southern Africa over a million years ago. According to Brink (op.
3
cit.) some of the black wildebeest’s characteristic features, such as its large eye
sockets, reduced nasal area and forward pointing horns, reflect its more territorial
behaviour in comparison with the blue wildebeest. Such territorial behaviour is often
observed in captivity by zookeepers, some of whom have, in the past, been attacked
and killed by black wildebeest individuals. The evolution of larger eye sockets,
reduced nasal area and forward pointing horns in the black wildebeest are all
probably a response to the need to visually patrol and defend breeding territories in a
treeless habitat without visual obstruction. In support of this view, it has been
observed that dominant black wildebeest bulls that are kept in captivity in bushy
areas tend to remove bush and tree branches with their horns, and break down tree
canopies in an attempt to clear the area for better visibility (African Wildlife 2003). It
can therefore, be expected that territorial behavioural differences between the black
and blue wildebeest could be important as ecologically separating mechanisms
where the two types are forced to co-inhabit an area with low habitat diversity.
Ezemvelo Nature Reserve in South Africa is a property where both types of
wildebeest have been confined together. Its location on the inland plateau of South
Africa, its diversity of habitats and its relatively large size (8 468 ha), provided an
opportunity to investigate the ecological separation of the black and blue wildebeest
where they have been confined together. According to Du Plessis (1969) blue
wildebeest did not naturally occur on sourveld and Mentis and Duke (1976) found
that it was ill adapted to such conditions. Von Richter (1971b) has also indicated that
sour grassveld is not suitable habitat for black wildebeest. Therefore the mainly
sourveld nature of Ezemvelo Nature Reserve forms a sub-optimal habitat for both
types of wildebeest.
The objective of the present study was therefore to investigate the evidence for
ecological separation of the two types of wildebeest at Ezemvelo Nature Reserve. To
determine the possible ecological separation, the habitat preferences, resource
utilization and potential interspecific competition between the two types of wildebeest,
was investigated as recommended by Scogings et al. (1990). This information will be
used to provide recommendations on the management of the two types of wildebeest
in areas where they are confined together.
In view of the historical blue and black wildebeest distribution, their occasionally
overlapping populations and their morphological, physiological and ecological
similarities, it was expected that interspecific competition would occur between the
4
two types of wildebeest in areas where they may be confined together. If so, it would
act as a limiting factor to the black wildebeest mainly because of its smaller size as
compared to the blue wildebeest. The aggressiveness and higher degree of
territoriality of the black wildebeest bulls in comparison with the blue wildebeest bulls
may, however, play some role in rivalling the overall dominance of the blue
wildebeest under certain circumstances. On the other hand, the black wildebeest is
less adaptable in terms of habitat use and area selectivity than the blue wildebeest.
This characteristic makes the black wildebeest more prone to displacement by the
more versatile blue wildebeest, due to exploitative interspecific competition. It was
therefore expected that the black and the blue wildebeest were too similar
ecologically to be kept together on Ezemvelo Nature Reserve, without harming each
other or the habitat.
To arrive at the objective of investigating the evidence for ecological separation
between the black and blue wildebeest, and to develop a relevant management
proposal for the two types of wildebeest where they are confined together, the
following key questions were addressed in the present study:
1. Is there any evidence of ecological separation between the two types of
wildebeest in terms of habitat preferences, spatial distribution, temporal
activity budgets and diet?
2. Does interspecific competition occur between the two types of wildebeest in
terms of behaviour and resource use?
3. What are the population dynamics of the two types of wildebeest?
4. What is the impact of the black and the blue wildebeest on their habitat?
5. If the two types of wildebeest are to be kept on the same property, what
management
actions
should
be
implemented
to
avoid
interspecific
competition if such interspecific competition is found to occur?
6. What management actions should be implemented to avoid damage to the
habitat by the black and blue wildebeest if such damage is found to occur?
5
CHAPTER 2: THE STUDY AREA
INTRODUCTION
This study was conducted at Ezemvelo Nature Reserve in South Africa (Figure 2.1).
The name eZemvelo means “back to nature” in the Zulu language. Currently
Ezemvelo Nature Reserve is one of the largest privately owned grassland reserves in
South Africa. It is an extremely important reserve from an ecological point of view as
the grassland biome is a meagrely conserved biome due to the preponderance of
agricultural activities and urbanisation in the areas that were previously open
grasslands. Only approximately 1% of this biome is formally conserved in South
Africa and many rare and endangered species can be found in the grasslands as it
has an extremely high biodiversity (Low and Rebelo 1996).
Black and blue wildebeest have a long history of co-occurrence on certain areas of
Ezemvelo Nature Reserve. In other sections of the reserve their co-occurrence is
relatively recent. The entire reserve formed the study area for the present study.
Ezemvelo Nature Reserve supports ecotourism in the form of overnight
accommodation, hiking trails, wildlife viewing and birding. It conserves a healthy
population of oribi Ourebia ourebi, South Africa’s rarest antelope, and provides
habitat for a high diversity of grassland bird species.
LOCALITY
Ezemvelo Nature Reserve is situated 24 km northeast of Bronkhorstspruit on both
sides of the border of the Gauteng and Mpumalanga provinces in South Africa. It lies
between latitudes 25º 38’ and 25º 45’ South and longitudes 28º 55’ and 29º 03’ East
on the topographical map sheets 2528DB and 2529CA (Government Printer 1996
and 1998). Figure 2.1 shows the position of the study area in South Africa within the
grassland biome. The reserve falls within the mesic highveld grassland bioregion
(Mucina et al. 2005) on the inland plateau of South Africa. The reserve is
approximately 8 468 ha (84.68 km2) in size, with approximately 45 km of boundary
fencing. It is bounded on its southern side by Renosterpoort Private Nature Reserve
and is surrounded on all its other boundaries by private farmland where both cattle
production and crop agriculture are dominant practices.
6
Ezemvelo Nature Reserve in South Africa
To Witbank
To Bronkhorstspruit
km
Figure 2.1: The location and boundaries of Ezemvelo Nature Reserve in the
Grassland Biome of South Africa, indicating its position (star) on the border between
the Gauteng and Mpumalanga provinces (Adapted from the map of Low and Rebelo
1996).
7
PHYSIOGRAPHY
Topography
Ezemvelo Nature Reserve ranges in altitude from 1 240 m above sea level at its
lowest point to approximately 1 500 m above sea level at its highest point (Figure
2.2.). The landscape is dominated by open, grassy plains, which are broken by
wooded, rocky ridges. The grassy plains occur on undulating hilly terrain. The east to
west profile of the reserve consists of undulating hills in the east sloping gradually
down to the Wilge River which bisects the reserve and then open plains gradually
rising to the west. In the southeastern section of the reserve the topography is very
broken, consisting of rocky cliffs where the Wilge River has cut through the
landscape. The topography toward the north, slopes downward into a wide valley
which stretches across the reserve from east to west and which is bound on the
northern side by a steep ridge rising to a northern plateau. This open plateau is
bounded by steep slopes in the north. The Grootspruit forms the northern boundary
of the reserve in the west and drains eastwards into the Wilge River that continues in
a north-easterly direction to form the northern boundary of the reserve in the east.
The Wilge River forms the northern boundary of the reserve in the northeastern
section (Figure 2.2). The range in slope is from 0 to 30° and the slope shape is
mostly convex (Land Type Survey Staff 1987).
Drainage
The perennial Wilge River divides the reserve almost in half and is fed by numerous
streams that originate from higher-lying wetlands or sponge areas (Figure 2.2). The
Wilge River flows northwards and eventually joins the Olifants River to the sea. The
Grootspruit joins the Wilge River in the north and forms the northern boundary of the
reserve in the west. The Sterkfonteinspruit occurs in the south of the reserve and
joins the Wilge River in the middle of the reserve. Both of these smaller streams,
along with another one, which flows to the eastern boundary of the reserve, contain
water year round. Two large dams that are fed by perennial streams occur on the
reserve, one on the western side and one on the eastern side. One small dam, also
fed by a small stream, occurs in the northwestern part of the reserve. A few
depressions fed only by rainfall also occur on the reserve.
8
Grootspruit
Sterkfonteinspruit
km
Figure 2.2: The topography and drainage at Ezemvelo Nature Reserve, South Africa.
Adapted from the topographical mapsheets 2528DB and 2529CA (Government
Printer 1996 and 1998).
9
There are no maintained artificial watering holes on the reserve as the available
natural water sources have proved sufficient for the wildlife during normal rainfall
years.
GEOLOGY AND SOILS
The reserve lies on the Wilge River Formation of the Waterberg Group and on the
Ecca and Dwyka Formations of the Karoo Group that were formed during the
Mokolian and Palaeozoic Eras respectively (1:250 000 geological series map sheet
2528 Pretoria (Geological survey staff 1978)) (Figure 2.3). The lithology of the Karoo
Group is dominated by an arenite conglomerate which produces dystrophic or
mesotrophic soils. The tillite-arenite produces some rocky areas with miscellaneous
soils. The Wilge River Formation consists almost entirely of sedimentary rocks such
as sandstone, but in parts it is intruded by conglomerate and igneous rocks. The
sandstone parent material is rich in iron oxides and consequently has a red to
purplish colour. Its subsequent weathering has resulted in red beds that contain
characteristically deep, red soils (South African Committee for Stratigraphy 1980).
There are also small amounts of manganese oxide in these Waterberg sediments.
The slope of the Wilge River Formation is gentle to the south in the plains, but steep
in the river valleys. The depositional characteristics of the conglomerate and shale
are indicative of an alluvial environment, with transportational direction from the
northeast. Cross-layeredness in the sandstone indicates the presence of a desert or
dune-veld previously (Lurie 2001).
The red colour of the rocks represents oxidised hematite, which indicates that the
atmosphere at the time of formation was strongly oxidising. In some areas diabase
occurs, which either covers the formation as plates, or is interspersed with it. These
plates are gabbroic to diabasic, with the most important minerals being plagioclase
and hornblende (South African Committee for Stratigraphy 1980).
10
km
Figure 2.3: The geology at Ezemvelo Nature Reserve, South Africa. Adapted from
the 1:250 000 geological series map sheet 2528 Pretoria (Geological survey staff
1978).
11
The soils include Hutton, Clovelly, Katspruit and Rensburg soil forms, and are highly
weathered with diagnostic dystrophic (highly leached), red and yellow to brown,
apedal characteristics (Soil Classification Working Group 1991). The soil texture is
generally sandy to sandy loam with 10 to 20% clay, and with little structure except for
in the bottomland areas. The sandy nature of the soils makes them susceptible to
forces of water erosion especially under conditions of poor ground cover.
The red and brown soil colours indicate adequate drainage and aeration. Grey soils
that occur along the drainage lines indicate waterlogged to wet conditions (Land
Type Survey Staff 1987).
Although it is calculated that the earth is approximately 4.6 billion years old, relatively
younger rock formations are found on the reserve. The red-brown, sandstone of the
Wilge River Formation that is widespread in the reserve is approximately 1.9 billion
years old (South African Committee for Stratigraphy 1980).
CLIMATE
Rainfall data statistics were calculated from the data for the Bronkhorstspruit
Weather Station (0514408X) by using data from 1970 until November. Rainfall data
from November 2003 data was obtained from the reserve records. The climate
diagram for temperature and rainfall following Walter’s convention (Walter 1979)
(Figure 2.4.) gives an indication of the distribution of the temperatures and rainfall
throughout the year taken from a mean of 11 years (1993 to 2003) for the Witbank
Weather Station (05153208) as temperature statistics were not available from the
Bronkhorstspruit Weather Station.
Temperature
According to the long-term records at the Witbank Weather Station, the highest
temperature over the period between 1970 and 2003 was 37.5ºC while the lowest
temperature recorded during the same period was –13.1ºC. The mean daily minimum
temperature of the coldest month (July) and the mean daily maximum temperature of
the hottest month (January) were 3.9 and 26.1ºC respectively, and the mean annual
temperature was 16.3ºC. These values were similar to publications by the Land Type
Survey Staff (1987) which indicated that the mean daily minimum temperature of the
coldest month (June) and the mean daily maximum temperature
12
(b) 1320 m
(a) Witbank
(e) 675 mm
(d) 16.3°C
(i) 37.5°C
(f) 3.9°C
(g) –13.1°C
(c1: 11-c2:11)
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
140
120
o
100
80
n
60
40
m
m
20
0
J
A
S
O
N
D
J
F
M
A
M
J
Month
Temperature
a = Weather station
Rainfall
f = Mean daily minimum
temperature for the coldest month
b = Altitude
g = Absolute minimum temperature
c1 = Duration of temperature data
recording
h = Mean daily maximum
temperature for the warmest month
c2 = Duration of rainfall data
recording
i = Absolute maximum temperature
d = Mean annual temperature
m = Dry period
e = Mean annual long-term rainfall
n = Wet period
o = Perhumid period (rainfall above
100mm)
Figure 2.4: Climate diagram for rainfall (mm) and temperature (°C) as determined
following Walter’s convention (Walter 1979) from July to June based on data
obtained from the Witbank Weather Station from 1993 to 2003.
13
Rainfall (mm)
Temperature (°C)
(h) 26.1°C
of the hottest month (January) were –0.7 and 27ºC respectively . Frost occurs in the
winter months from May to August.
According to data from the Land Type Survey Staff (1987), the mean duration of the
frost period in the area in which Ezemvelo Nature Reserve is situated, is 115 days. In
addition, the earliest frost date recorded for the area was 12 April and the latest frost
date was 19 September.
Rainfall
According to the climate diagram, the wet period is from October to March and the
dry period is from May to August (Figure 2.4). Figure 2.5 gives an indication of the
rainfall pattern since 1970 as measured at the Bronkhorstspruit Weather Station.
According to this data the wettest month was generally January and the driest month
was generally July. The mean annual rainfall for the area in which the reserve is
situated as calculated from the totals obtained from the Bronkhorstspruit Weather
Station, was approximately 650 mm and ranging from 412 mm (1998) to 949 mm
(1989). The rainfall data for Ezemvelo Nature Reserve, gathered from 10 regularly
checked rain gauges, are available since November 2003. Reserve data before this
time is not reliable and therefore could not be used. The actual rainfall that fell during
the study period (January 2004 to August 2005) is presented in Figure 2.6. The
annual rainfall that fell during the study period was 644 mm indicating that it was a
normal rainfall period. The three years before the study was undertaken were below
average rainfall years with a total annual rainfall of 556 mm, 506 mm and 531 mm for
2001, 2002 and 2003 respectively, as calculated from the Bronkhorstspruit Weather
Station data.
Ezemvelo Nature Reserve falls in the summer rainfall region of South Africa.
Conditions for rainfall in the study area, which is in the highveld region of South
Africa, are most favourable when there is an anti-cyclone off the east coast of South
Africa and a low-pressure system over the interior (Preston-Whyte and Tyson 1988).
Moist air from the Indian Ocean that reaches the plateau is usually unstable. Such
atmospheric conditions give rise to conventional destructive thundershowers of high
intensity, creating a high potential for erosion. In the winter, the days are generally
cool and cloudless and the nights are frosty. This is mainly due to a high-pressure
cell over the interior and the influx of dry air over the plateau (Preston-Whyte and
Tyson 1988).
14
0
100
200
300
400
500
600
700
800
900
1000
1970
1973
1976
1979
1982
1988
Year
1985
1991
1994
1997
2000
2003
in the Gauteng Province of South Africa from 1970 to 2003.
15
Figure 2.5: Long-term total annual rainfall (mm) for the Bronkhorstspruit Weather Station (0514408X), closest to Ezemvelo Nature Reserve
Rainfall (mm)
200
180
160
140
120
100
80
60
40
20
0
J
F
M
A
M
J
2004
2005
M o n th
J
A
S
O
N
D
Reserve, South Africa.
16
Figure 2.6: The actual total monthly rainfall (mm) received during the study period (January 2004 to August 2005) at Ezemvelo Nature
R ainfall (m m)
SEASONAL DIVISIONS
Three ecological seasons were identified at Ezemvelo Nature Reserve for the
present study to mirror the effects of the phenological cycle of plants. Grasses often
start to produce new leaves and emerge from dormancy before the onset of the first
rains, indicating that a factor other than rainfall is responsible for the onset of the
growth of dormant grass plants (Anslow 1966; Larcher 1995). Seasonal variation in
grass leaf emergence largely appears to reflect changing temperatures (Tainton
1999). Light and temperature are the most important factors influencing leaf growth
(Larcher 1995; Tainton 1999). It was assumed, based on studies of temperate plants
that the cut-off temperature below which grass plants become dormant was 15°C
(Larcher 1995; Tainton 1999). Mean daily temperatures for each month, based on
the data from the Witbank Weather Station, were therefore used to determine the
months when the mean daily temperature was below 15°C and when it was equal to
or above 15°C (Figure 2.4). This enabled making a division of the months of the year
into a cold season (dormant season or winter) and a warm season (growing season
or summer). The warm season was further delineated into an early growing season
(early summer) and a late growing season (late summer) based on biomass
accumulation rates (Grossman 1982), which are related to the mean rainfall during
those months. Three seasons were therefore delineated.
The January to April period (late growing season) was characterised by abundant
green forage, high temperatures and frequent thunderstorms. Wildebeest were in
excellent condition at this time, having just calved during early December, allowing
the calves to utilise the abundant forage available and keeping the cows in good
condition to produce abundant milk. The May to August period (dormant season) was
characterised by maturation and drying of grasses, low rainfall, and mean daily
temperatures below 15°C. Forage was dry and low in nutrient content. The
September to December period (early growing season) was characterised by
increasing mean daily temperatures, increasing occurrence of rain showers, and the
growth of grasses stimulated by the rising temperatures (sprouting) and the increase
of rainfall (growth and flowering). The most critical period for wildebeest was at the
end of the dormant season and the beginning of the early growing season due to low
abundance of quality forage. Other factors beside temperature can also affect the
onset of the growing season. The growing season can be accelerated by the
occurrence of fire. At the same time the growing season can be delayed due to a
17
paucity of fire where moribund material prevents new shoots from sprouting (Tainton
1999).
VEGETATION
The overall vegetation of the reserve can be classified as the eastern variation of the
Bankenveld, which has been described by Acocks (1988). Low and Rebelo (1996)
classify the vegetation in the area occupied by the reserve as a combination of the
Moist Sandy Highveld Grassland, Rocky Highveld Grassland and the Moist Cool
Highveld Grassland variations. More recent classifications have classified the
vegetation of the area in which the reserve occurs as Rand Highveld Grassland and
Loskop Mountain Bushveld (Mucina et al. 2005).
The Bankenveld is a transitional zone between the Savanna and Grassland Biomes.
The grassland areas of the reserve cover approximately 4 688 ha and are
characterised by large open plains that are interspersed with wetland areas, old
lands, patches of alien vegetation and a few rocky outcrops. Savanna dominates the
mountainous areas of the reserve, covering an area of approximately 3 780 ha and is
characterised by the occurrence of a higher tree density than the grassland areas.
The Wilge River that bisects the reserve is bordered by dense riverine bush.
The sandstones of the Wilge River formation is usually associated with the following
indigenous trees in the rocky areas: Englerophytum magalismontanum (stem fruit),
Vangueria infausta (wild medlar), Faurea saligna (Transvaal beech), Burkea africana
(wild syringa), Combretum apiculatum (red bushwillow), Cussonia paniculata
(mountain cabbage tree), Strychnos pungens (monkey orange), and Protea caffra
(highveld protea). The diabase intrusions are characterised by thorn trees, among
these the sweet thorn Acacia caffra and by Gymnosporia species. Vegetation in such
diabase intrusions is easily seen from the air, as the areas are densely wooded,
contrasting with the scattered vegetation on the sandstones. Intrusions of igneous
rocks, especially diabase, are characterised by lowlands, wetlands, and the rockless
strips through the hills of the reserve. Trees with thorns such as the Acacia species
are found in these areas.
No detailed vegetation assessment had been conducted for the whole reserve at the
onset of this project. A detailed vegetation classification is currently being compiled
18
by A. Swanepoel at the University of Pretoria but was not available at the time of
writing.
ANIMALS
Table 2.1 provides a list of the large mammals found at Ezemvelo Nature Reserve.
Approximately 250 species of bird have also been identified on the reserve.
RESERVE HISTORY
The information for the compilation of this section was obtained through interviews
with staff and management at Ezemvelo Nature Reserve as well as from the local
inhabitants of the area. Although every care has been taken to ensure its accuracy
based on these interviews, no responsibility will be taken for any errors or
ommisions.
Ezemvelo Nature Reserve is privately owned by the Oppenheimer family and is
made up of a number of farms that were purchased over the years (Figure 2.7). In
1974, the Oppenheimer family first purchased 1 640 ha from the owners of the
magazine, Farmer’s Weekly, to build a house and develop an organic garden. This
was done in the eastern section of the present reserve and the area involved was
named Telperion Farm (Section 1).
In 1980 another 855 ha of land (Section 2), which already had a variety of wildlife
species on it, including both black and blue wildebeest, was bought from Captain P.
Grobler. This land bordered on the southern section of the land previously bought.
The area was then incorporated into Telperion Farm and called Telperion Nature
Reserve. In 1984, another 386 ha of land to the south of the nature reserve (Section
3) was purchased and the whole reserve became 2 881 ha in extent. Of this, 20 ha
was used for planting lucerne to produce bales to sell to the public.
Another 913 ha (Section 4) were bought in 1988 to increase Telperion Nature
Reserve to 3 794 ha in size. This new area was called the Tshuswane section (the
name meaning ant) and the older part was called the Isipethu section (the name
meaning little stream).
19
Table 2.1: Large mammals found at Ezemvelo Nature Reserve.
Common Name
Herbivores
Black wildebeest
Blesbok
Blue wildebeest
Burchell’s zebra
Common eland
Common reedbuck
Common warthog
Gemsbok
Giraffe
Greater kudu
Grey duiker
Impala
Klipspringer
Mountain reedbuck
Oribi
Ostrich
Red hartebeest
Springbok
Steenbok
Waterbuck
White rhinoceros
Carnivores
Aardvark
Aardwolf
African civet
Black-backed jackal
Brown hyaena
Caracal
Leopard
Serval
Primates
Baboon
Vervet monkey
Scientific Name
Connochaetes gnou
Damaliscus pygargis phillipsi
Connochaetes taurinus taurinus
Equus burchelii
Taurotragus oryx
Redunca arundinum
Phacochoerus africanus
Oryx gazella
Giraffa camelopardalis
Tragelaphus strepsiceros
Sylvicapra grimmia
Aepyceros melampus melampus
Oreotragus oreotragus
Redunca fulvorufula
Ourebia ourebi
Struthio camelus
Alcelaphus buselaphus caama
Antidorcas marsupialis
Raphicerus campestris
Kobus ellipsiprymnus
Ceratotherium simum
Orycteropus afer
Proteles cristatus
Civettictis civetta
Canis mesomelas
Parahyaena brunnea
Caracal caracal
Panthera pardus
Leptailurus serval
Papio hamadryas
Chlorocebus pygerythrus
20
4
1
6
3
2
5
km
Figure 2.7: The distribution of the various sections purchased since 1974 by the
Oppenheimer family to form the current Ezemvelo Nature Reserve, South Africa. The
numbers on the map indicate Section 1 to 6 as described in the text.
21
In 1990 the planting of lucerne stopped and a breeding centre for Nguni cattle was
started. A number of problems were experienced with the cattle as a result of the
prevalence of malignant catarrhal fever caused by the intermingling of cattle and
black and blue wildebeest on the reserve.
In 1993 the neighbouring cattle farm on the western side of the Wilge River burned
down and the owner G. Britz sold it to the Oppenheimers in 1994. The 2 590 ha area
was named Bohlokwa (the name meaning important) (Section 5). In a 15-year period
prior to the incorporation of Bohlokwa into the Telperion Nature Reserve, two
different owners had made use of the land. The first owner practised poultry
production and maize was produced on the areas suitable for cultivation. Thereafter,
a large portion of the cultivated area was re-established to permanent pasture by the
second owner. Yellow maize, feed sorghum and lucerne were grown on a limited
scale in the rest of the cultivated areas to provide conserved feed for the cattle.
Mixed agriculture was also practised. Crops such as maize, sunflower, potatoes and
groundnuts were produced with some success. Animal production systems were
practised including cattle, wildlife, sheep, goats and chickens. Both the previous
owners made use of inorganic inputs such as fertilisers.
The Oppenheimers practised only organic farming and the cattle on Telperion Nature
Reserve were moved to the Bohlokwa section. In 1997 a peach production project
was initiated at Bohlokwa. A population of 6 500 trees was cultivated in a peach
orchard until 2003 when they were cut down.
In 1998 the Van Wyk family sold a 2 084 ha wildlife park on the western boundary of
Bohlokwa known as eZemvelo to the Oppenheimers to increase the land owned by
them to 8 468 ha (Section 6). This new section contained a variety of wildlife such as
common eland Taurotragus oryx, greater kudu Tragelaphus strepsiceros, impala
Aepyceros melampus melampus, white rhinoceros Ceratotherium simum, ostrich
Struthio camelus, waterbuck Kobus ellipsiprymnus, blesbok Damaliscus pygargus
phillipsi, and black wildebeest. The whole area was then renamed Ezemvelo Nature
Reserve.
Initially eZemvelo was kept separate from the other areas by keeping the boundary
fence in place. In 2000 a flood washed away the fenceline separating the Isipethu
and Bohlokwa sections, allowing wildlife to cross the Wilge River and mix with the
wildlife on the other side. In 2002 the Nguni cattle were removed from the reserve
22
and the fence between Bohlokwa and eZemvelo was taken down. Ezemvelo Nature
Reserve has been operating as a unit since 2002 with regard to wildlife movements.
Wildlife are now free to move throughout the 8 468 ha reserve, only being restricted
from crossing the Wilge River when it is in flood, something which occurs for a few
weeks every two to four years.
HISTORY OF THE BLACK AND BLUE WILDEBEEST POPULATIONS
Since at least 1980 there were both black and blue wildebeest on the eastern side of
the Wilge River in the Isipethu section. Records of the number of wildlife for this
section only date back to 1991 and the black and blue wildebeest numbers since that
time up until 2000 when the fence separating the two sections was washed away
appear in Figure 2.8. It is clear that there was a small number of black wildebeest on
the Isipethu section that struggled to reproduce since 1991. The blue wildebeest
were, however, successful and their numbers increased over the years. No culling or
wildlife capture has ever taken place on the Isipethu section since it was taken over
by the Oppenheimer family.
However, wildlife capture and hunting operations were a regular feature in the
eZemvelo section over the years, but these records along with wildlife census data
are not available for this section before 1998.
Before 2000 the black wildebeest population on the eZemvelo section had not come
into contact with blue wildebeest. The results of the wildlife counts conducted on this
section between 1998 and 2001 are shown in Figure 2.9.
During the time of the study, there were five main herds of black wildebeest occurring
on the western side of the Wilge River. There were only seven black wildebeest left
on the Isipethu section of the reserve on the eastern side of the Wilge River. There
were four large herds of blue wildebeest on the eastern side of the Wilge River and at
least three more large herds on the western side of the river in 2004. At the start of
the study there were 98 black wildebeest and 256 blue wildebeest on the entire
reserve of 8 468 ha (Tau 2004 pers. comm.)1.
1
Mr. M. Tau. Manager, Ezemvelo Nature Reserve. P.O. Box 599, Bronkhorstspruit, 1020,
South Africa. [email protected]
23
Num be r of a m im a ls
250
200
150
100
50
0
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Ye a r
B lac k w ildebees t
B lue w ildebees t
Figure 2.8: The number of black and blue wildebeest on the Telperion Nature
Reserve from 1991 to 2001.
24
Number of animals
90
80
70
60
50
40
30
20
10
0
Aug- Oct- Dec- Feb- Apr- Jun- Aug- Oct- Dec- Feb- Apr- Jun- Aug- Oct- Dec- Feb98
98
98
99
99
99
99
99
99
00
00
00
00
00
00
01
Month and year
Figure 2.9: The number of black wildebeest on the eZemvelo section of Ezemvelo
Nature Reserve from 1998 to 2001.
25
All the blue wildebeest occurring on the western side of the Wilge River during the
time of the study had migrated there from the Isipethu section after the fence along
the Wilge River was washed away in 2000.
26
CHAPTER 3: THE BLACK AND BLUE WILDEBEEST
TAXONOMY OF THE GENUS CONNOCHAETES
Both types of wildebeest belong to the Kingdom Animalia, the phylum Chordata and
the class Mammalia. They are placed in the order Artiodactyla, which contains all the
even-toed animals. They are grouped into the family Bovidae due to them being in
possession of horns. All African antelope species are grouped into this family as well
as the various species of buffalo Syncerus caffer. The wildebeest are placed in the
subfamily Alcelaphinae, which comprises three genera namely Connochaetes
(wildebeest species), Alcelaphus (hartebeest species) and Damaliscus (tsessebe
species, blesbok and bontebok Damaliscus pygargus dorcas). The Alcelaphinae
subfamily is characterised by both sexes having horns, well-developed pre-orbital
glands and pedal glands on the front feet, which are rudimentary or absent on the
hind feet, and no inguinal glands. The females of this subfamily have a single pair of
mammae (Smithers 1986). Both wildebeest species belong to the genus
Connochaetes due to the horns being smooth throughout and directed downwards
initially. Prominent facial tufts of hairs, a neck mane and a fringe of hairs either on the
throat or on the chest between the fore limbs further characterise this genus (Ansell
1974).
Five subspecies of the blue wildebeest occur on the African continent, but
Connochaetes taurinus taurinus is the only one naturally occurring in South Africa.
There are no subspecies of the black wildebeest and the only existing species is
endemic to South Africa.
DESCRIPTION OF THE CONNOCHAETES SPECIES
Field characteristics
Black wildebeest
Also called the white-tailed gnu, the males of the black wildebeest stand
approximately 120 cm at the shoulders and weigh from 147 to 193 kg whereas the
females are lighter and smaller being from 120 to 160 kg and standing approximately
110 cm at the shoulders (Smithers 1983; Furstenburg 2002a). Both sexes carry
horns arising from expanded bases, sweeping downwards and forwards and then
curving upwards. Males have thicker and heavier horns than females. Buffy brown is
27
the general coat colour with darker faces. The males become darker brown, almost
black as they age. Juveniles have shaggy fawn-coloured coats and straight horns
that only start to curve at about 1 year of age. The tail is dark at the base, becoming
off-white toward the tip and reaching almost to the ground. The manes are high and
erect with the hairs at the base a creamy colour becoming black at the tips. A distinct
beard of long hair is present, as well as a brush of long hair between the eyes and
the nostrils (Smithers 1983; Furstenburg 2002a). Territorial bulls utter a loud call
often described as an ‘oink’ (Mills and Hess 1997). Life expectancy is about 16 years
for males and 18 years for females and the natural population growth rate is
approximately 28 to 33% annually (Furstenburg 2002a).
Blue wildebeest
Also called the brindled gnu, adult males of the blue wildebeest stand approximately
150 cm at the shoulder and weigh from 210 to 260 kg, while females are smaller
standing 135 cm tall and weighing from 170 to 200 kg (Attwell 1977; Furstenburg
2002b). Their heads are massive and elongated, broadening out at the lips and
nostrils. A distinct beard of long black hair occurs on the chin. The overall coat
colouring is dark grey tinged with brown and with a silvery sheen. A series of dark
coloured bars occur on the neck and shoulders, extending to about the middle of the
body to give a brindled appearance. Juveniles and females are browner than the
adult males and may have more of a russet colour on the forehead, extending from
the forehead to between the eyes. Manes of long black hair, long whisks of black hair
on the ends of their tails and a fringe of long black hair occurs down the throat. Both
sexes carry horns which sweep outwards and slightly downwards and then rise
upwards to the inwardly pointing tips often directed slightly backwards (Smithers
1983). The mean life expectancy is the same as for the black wildebeest
(Furstenburg 2002b).
Morphological separation
Morphological separation of the two types of wildebeest is based on horn direction
and tail colour. The black wildebeest has horns that are directed downwards before
curving up, with the muzzle not noticeably elongated and a white tail. The blue
wildebeest has horns directed outwards and slightly downward before curving
upwards, a muzzle and nasals noticeably elongated and a black tail. The blue
wildebeest also has rudimentary pedal glands on the hind feet (Ansell 1974), which
are absent in the black wildebeest. The black wildebeest is generally smaller than the
28
blue wildebeest, with the males of black wildebeest weighing approximately 180 kg,
whereas the males of blue wildebeest weigh approximately 250 kg (Smithers 1983;
Furstenburg 2002a and b). Brindles on the neck and shoulders are present in the
blue wildebeest but not in the black wildebeest.
Distribution and status
Black wildebeest
The natural occurrence of the black wildebeest is limited to the central inland plateau
of South Africa (Figure 3.1). Therefore, it is endemic to South Africa. It was
specifically known to have occurred in the Free State province, the highveld regions
of the southern, central and northern Cape region, the southern parts of Gauteng, the
southeastern parts of North West province and marginally in the grassland regions of
KwaZulu-Natal province in the foothills of the Drakensberg Range (Von Richter
1974). The distribution coincides with the extension of the grassland and Karoo
vegetation types as delineated by Acocks (1988). It therefore occupied the central
grass and shrublands from south of Pretoria to close to the Swaziland border in the
east and the Botswana border in the west. Its distribution southwards included the
region south from the Kgalagadi Transfrontier Park in the west, south to and
including the central Karoo (Plug and Badenhorst 2001). The black wildebeest is
currently categorised as being of least concern on the Red Data List of Threatened
Animals as compiled by the International Conservation Union (IUCN) (Friedman and
Daly 2004).
Blue wildebeest
In the southern African subregion the blue wildebeest has a restricted distribution in
Namibia but it is widespread in Botswana. It occurs in parts of the North West and
Limpopo provinces, and throughout Mpumalanga province all the way south to near
the Swaziland border (Smithers 1983). It also occurs southwards in the HluhluweImfolozi Park in KwaZulu-Natal and has a marginal occurrence in the Northern and
Western Cape provinces (Figure 3.2).
It is particularly associated with open woodland where there is water (Smithers
1986). This type of wildebeest occurs in large numbers in Africa but due to loss of
habitat and illegal hunting, numbers are declining drastically throughout Africa.
Approximately 35 660 animals reside in South Africa, the largest populations existing
in protected areas, while other large populations are present on privately owned
29
Figure 3.1: Current distribution of the black wildebeest in South Africa. Adapted from
Friedmann and Daly (2004).
30
Figure 3.2: Current distribution of the blue wildebeest in South Africa. Adapted from
Friedmann and Daly (2004).
31
wildlife ranches (East 1998). The species is not under threat in South Africa and is
classified as of least concern in the latest IUCN Red Data List (Friedman and Daly
2004).
Du Plessis (1969) stated that the past geographical ranges of the black and blue
wildebeest must have overlapped in parts of the Gauteng, Free State, and Northern
and Western Cape provinces, at least seasonally. In these areas of overlap
differences in the habitat preference of the two types of wildebeest would have kept
them separate, as black wildebeest prefer short open grassland or karroid shrubland,
whereas blue wildebeest prefer grassland associated with open woodland. Today,
however, the black and blue wildebeest are confined together on many private
properties throughout South Africa.
Ontogeny and reproduction
Black wildebeest
The black wildebeest has a gestation period of 250 to 260 days and is strictly
seasonal in its calving time. The precise time of calving varies in different localities
from the middle of November to the end of December (Smithers 1983; Furstenburg
2002a). A single calf is born, usually during the morning and is able to stand on its
own within minutes of birth. They are weaned from six to nine months of age
depending on the range condition. Sexual maturity is reached by the females from 16
to 18 months and they give birth for the first time at about two years of age
(Furstenburg 2002a). Sexual maturity is reached by the males at about three years of
age, but usually only the older males can procure a territory and thus partake in the
rut (Von Richter 1974). The bond between a mother and her single calf is strong and
the calf remains close to its mother for the early part of its life, but with the birth of her
next calf the female drives it away (Smithers 1986). Day-length is the proximate
factor influencing the breeding cycle of this species (Smithers 1983).
Blue wildebeest
Gestation is about 250 days and the main calving season is from mid-November to
the end of December, with some births in May (Smithers 1983; Furstenburg 2002b).
Although there may be an inherent rhythm in the breeding activities, such activities
may be more influenced by prevailing climatic conditions (Fairall 1968). A single calf
is born, weighing about 22 kg at birth. It can run with the herd within five minutes
after birth (Smithers 1983).
32
Habitat
There are still large gaps in the literature regarding the habitat preferences of the
black and the blue wildebeest on small nature reserves under artificial circumstances
where predation is not a population control factor. Furthermore, no study has to date
compared the two types of wildebeest ecologically on the same property.
Weaver (1995) conducted a study on the habitat utilisation of selected herbivores in
the Klaserie Private Nature Reserve. In the dry season, a strong positive correlation
between the number of blue wildebeest and distance from riparian vegetation, as well
as distance from an established graded road was found. Habitats with good visibility
in all directions and with 10 to 40% bare ground were preferred. It was also found
that blue wildebeest had a clear affinity for habitats where Aristida congesta was the
visually dominant grass species, and that there was a dry season preference for
habitats where the visually dominant woody plant was a Grewia species. During the
wet season the blue wildebeest was most often found in areas where there was
heavy utilisation of the available grass biomass. These habitat preferences reflect the
dependence of the blue wildebeest on water, as well as its affinity for semi-open to
open grassland habitats and for areas that are dominated by increaser grass
species.
Fabricius (1984) found that the most important abiotic factors affecting black
wildebeest distribution in the Golden Gate Highlands National Park were slope,
aspect and grass height, with moderate slopes, northerly aspects and short grasses
being preferred. Schmidt (1988) described the habitat utilisation of black wildebeest
in the Suikerbosrand Nature Reserve, and found that it tended to aggregate on
northern slopes with an incline of 2 to 7º. Moreover, they especially concentrated on
such areas in the winter. The preference for northern slopes was related to
temperature. The black wildebeest was also found to concentrate on high-lying areas
and on plateaus, probably because of behaviour such as territory defence and
sighting predators. No study to date has found a relationship between the availability
of surface water, with the distribution of black wildebeest. A preference for open
grassland is probably associated with predation, as black wildebeest rely more on
speed than on camouflage to escape predators. Open areas do not allow for
concealment of approaching predators and this also relates to a preference for
slopes and high-lying areas where the visibility over the surrounding terrain is high.
33
Food
Black wildebeest
Black wildebeest are predominantly grazers, although they may also make use of
karroid shrubs. Their preferred habitat is the treeless, sweet grassveld and Karoo of
the central South African plateau. Sour grass species that mature quickly and
become unpalatable are only taken when fresh. Short grass veld is preferred and
areas of tall, matured grass are avoided (Von Richter 1974). The species conditions
its own preferred habitat by the tendency of herds to stay for prolonged periods in the
same areas, thereby keeping the grass in the preferred short state (Von Richter
1971a). Utilisation percentages of the different components of the veld have been
found to include 94% grass, 3% karroid shrubs and 3% herbs (Mills and Hess 1997).
Black wildebeest turn to the karroid browse after the first winter frosts when the grass
starts to lose its nutritional value (Mills and Hess op. cit.). Among the grasses
Sporobolus spp., Themeda triandra and Cynodon dactylon are important as food.
The karroid shrubs that are used as food include Nenax microphylla, Salsola
rabieana, Osteospermum leptolobum, Nolletia ciliaris and Pentzia spp. (Smithers
1983). The black wildebeest practises an extreme form of area selection in reserves
where migration is prevented (Von Richter 1971b).
Blue wildebeest
The blue wildebeest is virtually a pure grass eater. Attwell (1977) found that 96% of
the rumen content was grass with the other 4% being negligible amounts of bark and
browse. It is highly selective for leaf blade and sheaths throughout the year, taking a
higher percentage of stems during the dry season than the wet. It prefers short grass
(30 to 40 cm tall) and together with Burchell’s zebra Equus burchellii, is the first
herbivore species to appear on burnt veld (Grunow 1980). They are partial to fresh
sprouting grass on burnt areas and will move in search of fresh green grass
sprouting after rain (Smithers 1983).
Panicum spp., Digitaria argyrograpta and Themeda triandra are important in the diet
in KwaZulu-Natal, and Cynodon dactylon is utilised when other grass species are
heavily grazed (Attwell 1977). The habitat of the blue wildebeest can be described as
open grassland, floodplain grassland, open bush savanna and light, open woodland
(Smithers 1983).
34
Both types of wildebeest are dependent on water and their daily consumption is
about 8.3 litres with a drinking interval of 47 hours (Du Toit 1991). Black and blue
wildebeest also prefer natural water points to artificial waterholes.
Behaviour
Black wildebeest
This type of wildebeest is gregarious, with a social organisation involving territorial
males, female herds and bachelor groups (Smithers 1983). Territorial males alone
take part in the rut, and territories are marked by scent-marking with faeces and urine
and the various glands on the body, including the preorbital and the interdigital
glands (Furstenburg 2002a). The female herds consisting of females and their
offspring wander freely over their territories. The bachelor herds are loose
associations made up of adult, subadult and yearling males, with a lack of aggression
between the members of this herd (Von Richter 1971a). The black wildebeest is most
active in the early morning and late afternoon, resting in the middle of the day. Unlike
most animals, it does not seek shade during midday. The open grassland habitat,
which this species occupies, provides little shelter from the elements, as well as little
cover from predators. Therefore, the black wildebeest needs to be constantly vigilant
and relies on its speed when chased, rather than on camouflage to escape from
predators.
Blue wildebeest
This type of wildebeest is also gregarious and occurs in herds of up 30 animals. It
also has territorial males, female herds (nursery herds) and bachelor groups (Estes
1969). Territorially and sexually active bulls employ a number of ritual displays for
different circumstances (Smithers 1983). Young males are evicted from the nursery
herd at about 2 years of age. The social organisation of this species is much more
fluid than that of the black wildebeest, and only becomes rigid during the breeding
season. However, the males are not as territorial as they are in the black wildebeest.
Diseases
Wildebeest are prone to rinderpest, anthrax, bovine malignant catarrhal fever
(snotsiekte / BMCF), foot-and-mouth disease and heart water (black wildebeest),
amongst others. Snotsiekte and foot-and-mouth are the most important diseases
affecting wildebeest distribution and movement.
35
Snotsiekte is a viral disease that is sustained in its reservoir host, the wildebeest, and
has important consequences for the distribution of both black and blue wildebeest.
This is because the disease can be highly destructive to susceptible cattle
populations that may come into contact with infected wildebeest. Both black and blue
wildebeest (also including all members of the subfamilies Hippotraginae and
Alcelaphinae) are susceptible to the virus. Primarily the wildebeest calves excrete the
virus, which may be the main reason why outbreaks of the disease occur mostly
during the calving and weaning seasons (Du Toit et al. 1996). The virus has been
isolated from nasal and occular secretions of wildebeest calves up to 3 months of
age. It has been suggested that the virus replicates in the cornea and turbinates of
young wildebeest calves (Mushi et al. 1981). The vector is currently unknown,
therefore the mechanism of the transmission of the disease from wildebeest to cattle
is still poorly understood. Both cattle and exotic ruminants show signs of the disease,
whereas wildebeest are merely carriers of it.
Even though cattle and wildebeest graze in close proximity in many parts of South
Africa, only sporadic outbreaks of snotsiekte are recorded. For example, during 2002
only two outbreaks were recorded in the KwaZulu-Natal province resulting in the
death of nine head of cattle (Cooper 2003). This is due to the fact that although
wildebeest are continual carriers of the disease, they do not necessarily excrete the
virus and are therefore not always infectious to cattle.
Calves excrete the virus just after birth and during the stressful period of weaning.
However, any form of stress such as capture, hunting and aerial counts will promote
virus excretion, even in adult wildebeest.
The disease is limited to certain districts in South Africa (Du Toit 1991):
•
High-risk districts are Pilgrims Rest, Bela Bela, Waterberg, Thabazimbi and
Soutpansberg.
•
Medium-risk districts are Brits, Mokopane, Letaba, Lydenburg, Barberton,
Lephalale, Musina and Phalaborwa.
•
Low-risk districts are all other districts in South Africa.
Therefore the study area falls within a low risk area for snotsiekte.
Control of the disease involves the prevention of contact between wildebeest and
cattle, farm registration and the threat of litigation. However, it is realised that there
36
are no scientifically proven control measures for BMCF, except the separation of
cattle and wildebeest by wide distances and possibly the reduction of stress in
wildebeest if cattle are kept in the vicinity (Cooper 2003). Wildebeest are crawlers
and may be able to crawl under many types of game fencing. It is, therefore,
recommended that electrified fences be erected in areas where wildebeest are on
farms adjoining cattle ranches and that a separation distance of over 1 km is used as
the minimum between wildebeest and cattle (Du Toit 1991).
The smaller the farm the higher the risk of outbreaks or contacts, and the lower the
veld condition the greater the number of outbreaks (Du Toit 1991). Therefore, certain
minimum farm sizes should be set for keeping wildebeest and the veld needs to be
kept in a good condition.
Heart water is an indigenous bacterial disease that is a significant natural restrictor of
the distribution of the black wildebeest in South Africa (Du Toit et al. 1996). The black
wildebeest, like the springbok, is not indigenous to areas where the vector, the
Amblyomma tick species, occurs naturally and are thus susceptible to the disease.
Indigenous inhabitants of such areas are, however, immune.
Black wildebeest have also been found to be prone to swayback disease in areas
with copper deficiency in the soil (Penrith et al. 1996). Copper deficiency can result
directly from low dietary copper, or it can be induced in spite of adequate dietary
copper by interactions with other minerals. This can be prevented by adequate
copper supplementation in the form of licks or oral dosing. Swayback disease usually
causes death in calves, and ataxia and microscopic lesions of myelopathy in adults
(Penrith et al. 1996). This disease can have serious consequences for the survival of
isolated populations on wildlife ranches in areas where the soils are prone to
leaching and thus will have a copper deficiency.
Parasites
Parasite burdens are susceptible to wide seasonal fluctuations and in wild
populations it can be one of the main causes for declines in population size. In a
study of the blue wildebeest in the Kruger National Park and the black wildebeest in
the Golden Gate Highlands National Park in South Africa, it was found that only
Haemonchus bedfordi, a nematode and Gedoelstia hässleri, an oestrid fly larvae,
were recovered in large numbers in both types of wildebeest (Horak et al. 1983).
37
Black wildebeest appear to be fairly resistant to parasitic infections and were found to
have a much lower parasite load than the blue wildebeest. According to Horak et al.
(1983), blue wildebeest harboured 13 nematode species, four cestode species, one
trematode species, the larvae of five oestrid fly species, three lice species, seven
ixodid tick species and one mite species. Black wildebeest on the other hand carried
burdens of four nematode species, one cestode species, the larvae of five oestrid fly
species, two lice species, four ixodid tick species and one mite species. The lower
temperatures experienced by black wildebeest in their distributional range may be a
significant contributor to the lower parasite burden of the black wildebeest.
HYBRIDISATION
Threats of hybridisation
Hybridisation is possible if the two taxa in question have not diverged too far, but the
final outcome is unpredictable. It is widely believed that hybridisation is ephemeral,
leading ultimately to either speciation or to fusion of two races by introgressive
hybridisation (Moore 1977). This means that either a new species can be produced
by hybridisation or two species can be fused into one.
The black wildebeest is endemic to South Africa and hybridisation with blue
wildebeest poses a serious threat to the genetic integrity of the black wildebeest.
Hybridisation may eventually also pose a significant threat to the genetic integrity of
the blue wildebeest (Vrahimis 2003b). Black wildebeest numbers are still relatively
low in South Africa and therefore, the risk of extinction remains, which if allowed to
occur would result in a loss of endemic biodiversity (Anon 2003a).
Presently, with the large-scale increase in the number of wildlife ranches throughout
South Africa, landowners are keeping a wider range of wildlife on their properties
(Bothma 2002a). This is being done mainly to cater for local and overseas hunters
and ecotourists. The result is more and more areas accommodating both types of
wildebeest, with a concomitant potential increase in hybridisation.
Hybridisation is not only a threat to the types of wildebeest, but also to the livelihood
of wildlife producers (Vrahimis 2003b). Blesbok Damaliscus pygargus phillipsi and
bontebok Damaliscus pygargus dorcas hybridise freely, and only certified pure
bontebok and blesbok are recognised in the Safari Club International Record Book.
38
Wildlife producers would have to go through costly procedures to ensure that their
wildebeest populations are pure and if there is any possibility of hybridisation having
occurred, their animals may have reduced value. They may not even be able to sell
live animals. The only market for the hybrids may be in their meat and maybe for
hunters who would like to hunt a strange variation of the wildebeest, namely the socalled red wildebeest. This translates into serious economic implications for the value
of both types of wildebeest.
The extent of crossbreeding between black and blue wildebeest throughout South
Africa is presently unknown. However, the hunting community is reporting more and
more cases of hybridisation in different parts of the country. In order to remedy this
situation, it is of the utmost importance to impose drastic measures on a national
level (Anon 2003a).
Description of the hybrid
First generation hybrids are easily identified, but the offspring of hybrids that have
interbred with pure stock black or blue wildebeest are difficult to recognise on
appearance alone (Vrahimis 2003b). The most obvious deviation is in the shape of
the horns. Table 3.1 compares morphological features of the two types of wildebeest
and the hybrid. These features are in no way an absolute description of the hybrid
and probably only pertain to the hybrids studied by Fabricius et al. (1988).
Morphological features may vary according to the original black to blue wildebeest
ratio and the generation of the hybrid. According to Fabricius et al. (1988) the hybrids
appeared to have the same social organisation as the black wildebeest. Vocalisation
was found to consist of the grunting sound characteristic of the blue wildebeest and
the hybrid has been shown to be fecund. Conditions affecting the fertility of the
hybrids still need to be assessed. Evolution and speciation knowledge can be
improved by further study of the hybridisation process. Further studies also need to
be made on established hybrid herds in order to obtain information on the hybrid
ecology and the effects that the parentage would have on this ecology.
Factors leading to hybridisation
It is clear that the social behaviour of wildebeest, especially the habit of male blue
wildebeest to associate with animals of other species, predisposes them to
opportunities of hybridisation when confined with black wildebeest (Vrahimis 2003b).
39
In most recorded cases it appears that disruption of the normal demographic or
social structure was involved, as was seen at the Spioenkop Nature Reserve in 1995
(Langley 1995). According to Langley (op. cit.), the water level of the Spioenkop Dam
dropped during a dry period in the reserve. This allowed several blue wildebeest bulls
to cross the dam and establish territories in an area favoured by black wildebeest. It
is suspected that the larger blue wildebeest dominated the smaller black wildebeest
males during the rut and mated with black wildebeest females. It is believed,
however, that hybridisation is only likely under artificial conditions, where the two
types of wildebeest are forced together in a confined area. Hybridisation in such
areas occurs primarily as a result of the similar behaviour and the synchronised
breeding seasons of the two types of wildebeest, and due to their relatively recent
phylogenetic divergence (Anon 2003a).
Methods for identifying hybrids
The study by Corbet (1991) represented the first attempt to determine the extent of
genetic divergence between the blue and black wildebeest. A variety of molecular
and cytogenetic techniques were used to study the divergence between the two
types of wildebeest and to assess the status of the South African populations of
wildebeest with regards to inbreeding and hybridisation.
40
Table 3.1: Comparison of the morphological features of the hybrids and pure types of
wildebeest as studied by Fabricius et al. (1988)
Black wildebeest
Blue wildebeest
Hybrid
Horns curl down for half of
Horns horizontal for half of
Horns curl down at an
their length and then curl
their length and then curl
angle of 30 degrees and
upwards.
inwards towards the head.
then curl outward, away
from the head.
Colour buff brown.
Colour dark grey.
Colour dark brown.
Tail white.
Tail black.
Top half of the tail black,
bottom half white.
Mane stiff and erect, black
Mane shaggy and black.
Mane stiff and erect, black
and white.
No brindles.
Brindles on neck.
Brindles on neck.
Nose not elongated with a
Nose smooth and
Nose not elongated with a
tuft of hair.
elongate.
tuft of hair.
Shoulder height of males
Shoulder height of males
Shoulder height of males
1.2 m.
about 1.5 m.
about 1.2 m.
and white.
41
The study included cytogenetics, which included chromosome number and shape
(karyotypic) comparisons between taxa based on G- and C-banding techniques
(involving the staining of chromosomes). This gives an indication of whether stable
meiosis was possible in cases of hybridisation. Both types of wildebeest have the
same number of chromosomes (diploid number: 2n = 58) and it was not possible to
distinguish between the two by using staining techniques, indicating that meiosis
would be stable and not impair hybrid fertility (Corbet 1991). Analyses of
mitochondrial
deoxyribonucleic
acid
(mtDNA)
restriction
fragment
length
polymorphisms were also performed. The mtDNA analysis showed that there was a
2% divergence between the black and blue wildebeest, which roughly corresponds to
1 million years of divergence, a similar estimate to what has been suggested based
on the fossil record (Brink 1993). The mtDNA also showed important differences
between the two types of wildebeest with respect to the amount of within species
genetic variability.
Examination of protein variation by using one-dimensional gel electrophoresis
(allozyme or protein electrophoresis) and of variation in the nuclear genome by
utilising ribosomal DNA and DNA fingerprinting probes was also performed, but did
not reveal a diagnostic test for hybrids (Corbet 1991).
The use of DNA fingerprinting has recently been, to a large extent, replaced by
micro-satellite markers (Grobler 2003). These are short DNA repeats (of two to five
bases) that are highly variable and have the advantage that each locus is studied
independently, therefore allowing the assignment of heterozygosity. Alais (2000) and
Grobler (2003) used micro-satellite markers to address genetic variability within black
wildebeest and in the hybridisation between the two types of wildebeest. Four
potentially diagnostic markers were found and now need to be tested on known
hybrids and the two pure forms.
Grobler (2003) listed a few priorities for future research on the topic:
•
Assess the genetic diversity of the pure species by using carefully selected
reference populations.
•
Test the diagnostic loci on known hybrids (preferentially F1 hybrids and
backcrosses).
•
Assess fitness related traits (such as sperm quality) of pure forms and their
hybrids and assess adaptive genetic variation.
42
Osteological work is also being conducted, and potential markers are in the process
of being identified (Anon 2003a).
Prevention of hybridisation
It is currently thought that the only way in which to ensure that hybridisation does not
occur, is to prevent any contact between the two types of wildebeest. This is because
conservationists and scientists cannot clearly identify the factors that result in
hybridisation, and whilst there is much speculation, there are insufficient, adequately
documented cases. In order to identify these factors it is necessary to understand the
ecological and behavioural differences between the two types of wildebeest.
CONSERVATION
Historical conservation status of the black wildebeest
As a result of heavy, indiscriminate hunting pressure by travellers, hide hunters and
settlers, as well as the allocation of the best fertile land to farming in the previous
century, the number of black wildebeest had dwindled to the brink of extinction by the
1940s (Fabricius and Oates 1985). Conservation programmes were then set up to
conserve and breed this endangered species for relocation to those areas where it
had previously occurred.
At the turn of the 20th century the population of black wildebeest in South Africa had
fallen to below 1 000 (Fabricius and Oates 1985). A survey performed by Bigalke
(1947) revealed that there were approximately 1 048 black wildebeest in the Union of
South Africa in 1946. Eighteen years after the publication of the first survey by
Bigalke (1947), another survey was performed by Brand (1965) yielding a total of 1
808 black wildebeest in South Africa. In 1970, when a third survey was done by the
Orange Free State Directorate of Nature and Environmental Conservation, the
population size had risen to 3 220 animals. A survey done in 1979 showed that there
were 1 532 black wildebeest in the Transvaal alone. In 1981 another survey revealed
a population size of 6 493, and in 1988 the count revealed a total of 6 685 animals in
South Africa (Kay 1992). By 1997 the numbers of black wildebeest in South Africa
had increased to approximately 12 000 animals (Mills and Hess 1997). Presently, the
total black wildebeest population is estimated at more than 18 000 animals, of which
43
80% occur on private land and 20% in formally protected areas (Anon 2003a). This
steady increase of animals was the direct result of the intensive conservation
programmes put into place to conserve the species, as well as the dedication of a
few Free State farmers who were intent on conserving the species (Weaver 1992).
The species has been widely re-established within its former distribution range and
more recently introduced into other parts of the country and into neighbouring
countries outside its historical range (Mills and Hess 1997). On private farmland in
Namibia, importations from South Africa have led to a dramatic rise in the estimated
total numbers of black wildebeest, from 150 in 1982 to more than 7 000 in 1992 (East
1998). The megapopulation size is therefore, steadily increasing, especially on
private land. This has been a major reversal in status for this species and reflects
favourably on the conservation efforts of the past.
Regulations and policies in South Africa
Past policies in the former provinces of South Africa did not provide any regulations
on the housing of the two types of wildebeest on the same property. Recently it has
been recognised that black wildebeest are an important endemic South African
species and that hybridisation with the blue wildebeest is undesirable and a threat to
both types of wildebeest. It is suspected that with the ever-increasing number of
wildlife ranches being developed throughout South Africa, more and more properties
are keeping the two types of wildebeest together. At present, in the Free State,
Northern Cape and KwaZulu-Natal provinces alone, it is estimated that there are
more than 120 private properties that are housing black and blue wildebeest together
(Anon 2003a). This means that the hybridisation problem in South Africa can already
be far advanced.
The present policies in the various provinces of South Africa are described below,
keeping in mind that a National Translocation Policy is about to be released, which
will hopefully streamline the policies in the current nine provinces and address the
hybridisation problem.
44
Gauteng
Information for this province was supplied by Buijs 2003 (pers. comm.)2. The policy is
not to allow both types of wildebeest on one property. In addition only black
wildebeest are allowed on highveld grassland properties and blue wildebeest in the
savanna regions of Gauteng. When it is a borderline case, it is treated on merit
through the use of historical data and an ecological assessment. The species that
occur on other farms in the region are also looked at and the species most common
in the area is recommended. This attempts to prevent complications in the future
when landowners decide to amalgamate their properties. Where permits for both
species have been issued in the past, permits for the removal of live animals from
such farms will not be allowed. The only legal way to remove them will be to shoot
them.
As far as the records at Gauteng Nature Conservation go, only seven properties in
Gauteng house both types of wildebeest. It is not apparent whether they are
separated by a fence or not. Considering the size of Gauteng in relation to the other
provinces this indicates a high density of properties possessing both types of
wildebeest.
KwaZulu-Natal
Information for this province was supplied by Rushworth 2003 (pers. comm.)3. Black
wildebeest have protected status in KwaZulu-Natal. Their policy is to ensure that both
types of wildebeest do not occur in the same protected area, even where a fence or
other barrier may separate them. They also aim to maintain wildebeest in protected
areas in accordance with historical distribution and habitat suitability, with preference
being given to black wildebeest in areas of overlap in distribution and where habitat is
suitable for both types. Only certified genetically pure wildebeest are to be introduced
into protected areas. They plan to adopt a certification process for all black
wildebeest populations in the province and hope that this will be copied by the other
provinces in the country. The policy is also to prevent the introduction of both types of
wildebeest to any property in the province through permit controls.
2
Mr. D. Buijs. Regional Ecologist, Gauteng Nature Conservation. P.O. Box 8769,
Johannesburg, 2000, South Africa. [email protected]
3
Mr. I. Rushworth. Ecological Advice Co-ordinator, Ezemvelo KZN Wildlife.
PO Box 13053, Cascades, 3202, South Africa. [email protected]
45
Excess wildebeest of any type will only be sold to landowners who either have
certified pure populations or no wildebeest on their properties. Only landowners who
have adequate fencing to contain these animals and have made a commitment to
maintaining these herds as pure in the long-term will be able to purchase such
wildebeest. KwaZulu-Natal promotes and facilitates, with compensation where
necessary, the removal of hybrids or of one type of wildebeest where both have
already been introduced to a property under permit.
Northern Cape
Information for this province was obtained from Jonk 2003 (pers. comm.)4. New
translocation policies for wild animals within the Northern Cape have been approved,
which state that black and blue wildebeest are not allowed on the same property
unless they are separated by a game fence that meets the specifications of the
province for antelope. Fence specifications for the wildebeest are either a stock proof
fence or a jackal proof fence that is 1.4 m high.
Western Cape
Information for this province was supplied by Lloyd 2003 (pers.comm.)5. Originally
the mammalian translocation policy in the Western Cape did allow for the two types
of wildebeest to be kept on the same property. This was in order to cater for those
properties in the former Cape Province where the two types of wildebeest occurred
sympatrically. Once it became known how serious the hybridisation problem was,
such permission was no longer given. However, there are several properties in the
Western Cape that still have both. It is simple to deal with those that are covered by
certificates of adequate enclosure. However, there are cases of people having them
illegally without such certificates.
Other provinces
It was not possible to obtain information from the remaining provinces in South
Africa. However, it is clear that most provinces are taking the hybridisation threat
seriously and are attempting to implement regulations, which will prevent the two
types of wildebeest from occurring together on the same property. Some provinces
are stricter than others in their policies. The development of the National
4
Ms. M. Jonk. Chief Nature Conservator, Northern Cape Nature Conservation. P.O. Box 231,
Upington, 8300, South Africa. [email protected]
5
Mr. P. Lloyd. Western Cape Nature Conservation Board. Private Bag 5014, Stellenbosch,
7599, South Africa. [email protected]
46
Translocation Policy will hopefully address these differences and incorporate
wildebeest hybridisation.
Evidently, the existing regulations to prevent keeping blue and black wildebeest
together should be retained and enforced as vigorously as possible. All possible
hybrid herds on private properties and nature reserves should continue to be
regarded as such, until proven otherwise by using a molecular genetic approach
(Anon 2003a).
National conservation plan
A National Conservation Plan for the two types of wildebeest is being developed by
the Free State Department of Tourism, Environmental and Economic Affairs, the
University of Pretoria, the Animal Genetics laboratory at the Agricultural Research
Council and provincial parks’ boards and nature conservation agencies throughout
South Africa (Vrahimis 2003b). The aim of the National Conservation Plan is to
investigate the extent of hybridisation between black and blue wildebeest in South
Africa, by involving all role-players and to develop a national strategy and action plan,
which will be aimed at ensuring the genetic integrity of both these wildebeest types
(Anon 2003a).
It is important to understand that the problems being experienced with black and blue
wildebeest hybridisation are symptomatic of larger problems pertaining to
conservation and trade in wildlife and cannot be addressed in isolation. Addressing
some of the broader problems will automatically assist in addressing some of the
wildebeest issues.
A workshop on black wildebeest hybridisation was held in June 2003 at the
Florisband Quaternary Research Station in the Free State (Anon 2003a). This
workshop was the starting platform for the development of the national conservation
plan. Participants worked in two groups by focussing on the research aspects of the
hybridisation process and the regulatory and policy mechanisms that are necessary
to manage the situation (Anon 2003a). The policy and legislation working group and
the research-working group identified problems covering the core issues. The
problems were addressed and possible solutions sought (Table 3.2).
47
One of the main problems identified was the lack of uniform national legislation in
South Africa (Problem 1). Therefore, South Africa was considered not to be
complying with its obligations in terms of the Convention on Biodiversity, which this
country ratified in 1995 (Anon 2003a).
The National Translocation Policy, which is about to be released, includes specific
mention of the issue of wildebeest hybridisation and may be able to provide a
solution to this problem. Current market forces in wildlife ranching are promoting
large diversity of animals on small properties (Problem 2). These forces are not only
favouring hybridisation but also the breeding and hunting of colour variations such as
those of springbok, blesbok and impala, as well as the housing of species outside of
their natural distribution range (Anon 2003b). This is a difficult problem to tackle as it
involves the changing of attitudes among members of the wildlife ranching industry.
An incentive-disincentive scheme has to be compiled to address this problem.
48
Table 3.2: A list of the problems and their proposed solutions for the black and blue
wildebeest hybridisation problem in South Africa as stipulated in a workshop on black
wildebeest hybridisation held in June 2003 at the Florisband Quaternary Research
Station in the Free State province (Anon 2003a)
Number
Problem
Proposed Solution
1
The lack of uniform national
legislation, policy and strategy is
resulting in the inability of the state
to control existing problem areas
and prevent further problems.
Support current initiatives to develop and
fast-track the publication of the National
Translocation Policy, ensuring that this
addresses all the wildebeest issues.
2
Current market forces promote
landowners to stock many species,
often on relatively small areas,
thereby encouraging hybridisation.
1). Consult/work with hunting
organisations to get their support for not
recognising trophies hunted out of
natural distribution, hybrids/potential
hybrids, or colour variations.
2). Consult with zoological organisations
to prevent inappropriate ‘dumping’ of
exotic/hybrid animals.
3). Conservation Extension staff have to
be trained/re-trained to discourage the
desire among landowners for stocking
as many species as possible.
4). Investigate the development of
certification/registration system for pure
herds.
5). Reduce cost to individuals of
rectifying problem areas by exchanging
hybrids or swapping for one type of
wildebeest where both exist.
6). Advisory service by conservation
agencies at auctions to prevent the
unintentional spread of problems and
discourage deliberate problem creation.
7). Auctioneer incentives through
developing a code of conduct and
providing conservation ‘Stamp of
Approval’/accreditation – will assist in
controlling the live trade.
3
There is uncertainty as to the
genetic purity of existing wildebeest
populations.
1). Conduct a national survey and
develop a database of all properties with
one or both species of wildebeest, with
tentative status i.e. not confirmed by
genetic tests.
Database must:
-Include details of landowners.
-Must be available to all provinces to
assist with permitting.
-Will highlight specific, immediate
problem areas.
2). Support genetic and osteological
research, including financial support by
the state.
49
Table 3.2 (Continued):
Number
Problem
Proposed Solution
4
Lack of capacity within conservation
agencies to implement existing legislation
i.e. reducing effectiveness of the state to
regulate, control and be aware of
movement of wildebeest.
No solutions were formulated
due to time constraints.
5
Failure by South Africa to comply with the
requirements of the Convention on
Biodiversity.
No solutions were formulated
due to time constraints
6
Different organs of state are enacting
legislation in an uncoordinated manner and
that often results in contradictory/conflicting
approaches.
No solutions were formulated
due to time constraints
7
There is no integration of the existing
genetic benchmarks or in determining the
potential gaps in the existing data.
Establish the genetic
benchmarks for identifying
black and blue wildebeest
and hybrids.
8
Lack of understanding of the natural
distribution range and origin of wildebeest
within that range.
Collation and verification of
existing information on: fossil
evidence-records; historical
records; indirect indications –
habitat requirements as a
proxy.
9
There is a lack of national information on
the extent of the hybridisation problem
(number of farms and nature reserves with
one or both type).
Collate existing information
and develop a national
database on wildebeest data.
10
There is insufficient understanding of the
hybridisation process (herd dynamics and
farm practices for possibly keeping the two
species).
Collate all existing
information on the
hybridisation process and
formulate methods to fill in
knowledge gaps.
11
There is no clear understanding of black
wildebeest or hybrids as potential vectors
for snotsiekte.
Consult Onderstepoort on
existing research, or anybody
else working on the problem,
to find out what they know on
the issue.
50
Table 3.2 (Continued):
Number
12
Problem
There is insufficient dissemination of
information to relevant stakeholders.
Proposed solution
1). Disseminate information
on the problems and possible
solutions to a wide range of
stakeholders, including
wildlife ranch owners and
managers, scientists and the
public.
2). Publish information in
scientific journals, write
reports and hold training
workshops for nature
conservation scientists and
managers.
13
There is no general, easily implementable
identification system for pure individuals.
Develop a cost-effective,
rigid and easily
implementable ID system.
51
CHAPTER 4: HABITAT SELECTION AND SEPARATION: GENERAL
METHODOLOGY
INTRODUCTION
Habitat selection is one of the major components of any general ecological study of a
species (Penzhorn 1982). Understanding black and blue wildebeest habitat selection
is vital for assessing the ecological separation between the two types of wildebeest at
Ezemvelo Nature Reserve. A habitat can broadly be defined as the area that
contains all the biotic and abiotic components necessary to an animal to sustain all of
its basic life requirements (Fabricius 1989; Joubert 2002). Therefore, the presence of
an animal in a certain habitat indicates that the minimum requirements for its
existence have been met by that habitat (Riney 1982).
Habitats are selected by a species according to the specificity of its niche and the
extent of the special physical adaptations that the species has developed to
successfully exploit that niche (Ben-Shahar 1986). The Hutchinsonian concept of a
niche was used in the present study, which states that a niche is the totality of the
environmental factors in n-dimensional hyperspace acting on a species (Hutchinson
1957).
Differences in body size, mouth morphology, feeding style, and digestive systems are
usually the main reasons cited for differential niche use between sympatric African
grazing herbivores (Gordon and Illius 1989; Voeten and Prins 1999). However, when
two types of wildlife are morphologically, physiologically and behaviourally similar,
such as is the case with black and blue wildebeest, their niches would also be
expected to be similar. Consequently their habitat preferences should be similar.
Given these similarities, the question arises as to if and how these two types of
wildlife would be able to co-exist without competing for their basic resources? The
occurrence of both black and blue wildebeest in the same area provides an ideal
opportunity to provide some information which may allow deductions to be made as
to whether the two types of wildebeest have inherent niche differences, which cannot
be directly deduced from their overall morphology, physiology and behaviour.
If no niche differences exist between the two types of wildebeest, alteration of the
normal foraging and habitat use patterns by one or both types may have to occur in
52
order to avoid competition. Such adaptations in behaviour may have negative
impacts on the long-term viability and survival of an animal in an area (Rubin et al.
2002). However, evidence for such adaptations may only be found in studies where
the habitat use of the two types of wildebeest is studied in isolation, as well as where
they occur together and thus, if they are being made, the present study will only be
able to infer such adaptations, not prove them. However, if the adjustments
mentioned above are being made, then it may result in the long-term decline in one
of the two types of wildebeest that presently occur at Ezemvelo Nature Reserve.
Blue wildebeest habitat selection has been extensively studied in savanna
ecosystems in the eastern parts of South Africa (Whyte 1985; De Wet 1988; Wentzel
et al. 1991; Weaver 1995). Black wildebeest habitat selection has been studied in
detail in the Golden Gate Highlands National Park (Fabricius 1984; Kay 1992) and at
the Suikerbosrand Nature Reserve (Schmidt 1988) in South Africa. No habitat
selection studies on either type of wildebeest have been conducted on populations
inhabiting the grasslands of the highveld where Ezemvelo Nature Reserve is
situated.
Habitat separation has been demonstrated by a number of studies on niche
partitioning in sympatric species (Van Horne 1982; Dueser and Shugart 1978;
Rushworth 1992; Forsyth 2000; Wei et al. 2000; Namgail et al. 2004). A decision on
what aspects of a habitat to measure in order to determine the factors that separate
the habitat choices of two species is a complicated one. The habitat of a species
consists of both biotic (wildlife and vegetation) and abiotic (physical) factors (Joubert
2002), and an analysis of both factors is equally important. The scale of analysis is
also important. Habitat selection within a reserve context can take place at a number
of scales. The scales most commonly analysed is the broad habitat type level
(macroscale) and the feeding site level (microscale) (Novellie 1990). In addition
habitat selection can also be analysed at a scale that would incorporate all the abiotic
and biotic habitat factors such as slope, aspect and woody vegetation cover at the
sites of occupation of the species under study (mesoscale). Analysing habitat
selection at multiple levels can allow for observation of influences that may be
masked within a single level analysis (Lyons et al. 2003). Various authors have
recommended hierarchical approaches to the analysis of habitat utilisation in order to
examine habitat selection operating on different levels (Johnson 1980; Manly et al.
1993).
53
Therefore in the present study, habitat selection was firstly examined at the broad
habitat type scale to determine which broad habitat types and their associated
vegetation characteristics were preferred by which type of wildebeest (macrohabitat
scale) (Chapter 5). Secondly, habitat preferences of the black and blue wildebeest
were examined at the mesoscale incorporating both the abiotic and biotic
components of the habitat (Chapter 6). This was achieved with the creation of logistic
regression models. These models allowed for the delineation of the important
separating mechanisms operative in the habitat preferences of the two types of
wildebeest. Finally, the vegetation characteristics of the feeding sites (irrespective of
habitat type) of the two types of wildebeest were then examined to determine
whether they differed in terms of forage quality and quantity (microhabitat scale)
(Chapter 7).
In addition to the above, examination of habitat utilisation within a seasonal context is
an important component of habitat separation analysis (Fabricius and Mentis 1990;
Heitkönig and Owen-Smith 1998; Traill 2004). Many studies have investigated habitat
separation and resource overlap in only one season, usually at the time when
resources are most abundant. However, the critical season when resources are most
limiting would be expected to result in higher competition for shared resources
(Gordon and Illius 1989). For continued coexistence, mechanisms, if they exist, to
reduce this competition would be most evident during the critical season. In addition,
habitat features and requirements of herbivores may change within a daily context
and thus the time of day may impact on habitat separation between the two types of
wildebeest and provide further opportunity to avoid competition for shared resources
(Hemami et al. 2004). Certain weather conditions may also influence the habitat
preferences of a species, especially when these conditions become extreme (Pianka
1973). Different social groups may also exhibit different habitat preferences (Geist
and Petocz 1977; Przybylo and Merila 2000).
Taking the above into consideration, habitat selection and separation of the two types
of wildebeest was examined also within a seasonal context, and where possible, a
daily, weather and social group context.
It was expected that the habitat preferences of the two types of wildebeest would be
too similar for ecological separation to occur in terms of habitat use and therefore the
objective of this part of the present study, which has been detailed in Chapters 5 to 7,
was to answer the following key question:
54
•
Is there any evidence of ecological separation between the black and blue
wildebeest being achieved by segregation in habitat selection?
METHODS
The basic survey method utilised to determine habitat preferences of the two types of
wildebeest has been described in detail in this chapter. Any variations to this method
and additional methods that were applied exclusively when examining the vegetation
characteristics of the habitat types and feeding sites have been described under the
relevant chapters.
Field collection of the data
Numerous methods exist to determine the habitat preferences and interactions of
herbivores with each other and their respective habitats. The specific method chosen
is usually determined by local circumstances such as available resources and time.
The level of the investigation is important and studies can be conducted at the
population or individual level (Thomas and Taylor 1990). The present study
investigated habitat selection at the population level as individual animals were not
identified.
The most widely used method for assessing habitat selection in Africa is that of
observing the species under study by traversing road transects that are established
to incorporate all the plant communities and habitat types in the area (Weaver 1995;
Dörgeloh 1998; Von Holdt 1999; Strauss 2003; Cromhout 2006). This method
requires the measurement of certain variables at each sighting of the animals
concerned. The variables chosen for incorporation in a study are based on prior
observation of the species under study and from other relevant studies.
The method selected for the present study was that of systematic sampling in all
possible wildebeest habitats and the recording of site attributes wherever wildebeest
were located. According to Fagen (1988), ecological habitat selection theory
suggests that population densities are an indication of habitat quality. The
procedures used here were primarily those of Ferrar and Walker (1974), Melton
(1978) and Reilly (1989).
55
To avoid sampling bias, a methodical search pattern was laid down to cover the
entire study area as suggested by Pettifer and Stumpf (1981). The study area was
divided into ten blocks, each of which could be conveniently searched within 2 to 3
hours. Two or three adjacent blocks were systematically searched each day on foot
or by vehicle, following a standard route. This standard route was devised and
formalised after a ground reconnaissance session that aimed to determine the areas
on the reserve where both types of wildebeest were most likely to occur and the
areas that were inaccessible to them. This resulted in some areas of the reserve
being excluded from analysis as the terrain in these areas was found to be
inaccessible to both types of wildebeest. The route was reversed on every alternate
search to minimise observer bias (Von Holdt 1999). All wildebeest encountered
within that block were recorded. An attempt was made not to measure the habitat of
the same individuals on the same day (Manly et al. 1993). The blocks were searched
in an orderly manner from one to ten to ensure independent observations. No block
was therefore surveyed more than once in 3 days. Each block was surveyed at least
twice a month and not more than five times a month. Data collection lasted from
January 2004 to August 2005 allowing for seasonal differences to be explored.
The objective of this type of sampling was to achieve an even intensity of sampling in
all wildebeest habitats so that species frequencies would reflect abundance,
distribution and habitat preferences (Ferrar and Walker 1974; Reilly 1989). This
method was chosen over the fixed transect method utilised by many other studies
(Weaver 1995; Von Holdt 1999; Traill 2004), due to the relative seasonal stability of
the distribution of the two types of wildebeest within the study area and due to certain
areas consistently not being utilised by the wildebeest (pers. obs.). This observation
may have been due to the territoriality of both types of wildebeest (Von Richter
1971a). In addition, the road system in the reserve did not cover all the areas where
wildebeest were known to occur and the road transect method would have undersampled the entire population of both types of wildebeest.
Once an animal or group of animals was located, the point where an individual
occurred or the centre point of the herd was determined. This position was recorded
with a Global Positioning System (GPS) instrument. During approach, the dominant
activity was recorded and the site size or area of occupation of the group was noted
for the purpose of subsequent vegetation sampling. A standard site size of 5 m
radius was used for single stationary animals where all subsequent vegetation
variables were recorded (Ferrar and Walker 1974).
56
A fixed set of habitat variables was measured and their values recorded on a field
data sheet (Appendix 1). A broad-based holistic approach was considered most
valuable for this part of the present study and it was therefore decided to reduce the
accuracy of data collection, and thus the time spent at each site, rather than the area
or number of habitat factors considered (Ferrar and Walker 1974). As a result, many
of the variables were visually estimated rather than measured. Since it is expected
that the choice of a site by a type of wildebeest is likely to include variability of the
same order of magnitude as would result from such visual estimates (Ferrar and
Walker 1974), the use of this level of accuracy (visual estimate, rather than the
quantitative measurement) was considered appropriate to the conditions of the
present study.
The following variables were recorded at each sighting:
Type of wildebeest
•
Black wildebeest
•
Blue wildebeest
Date and time of observation
The date was noted to determine seasonal habitat preferences and the time was
recorded to determine whether the time of day affected habitat selection.
The seasons were categorised as follows:
•
Late growing season: January, February, March and April
•
Dormant season: May, June, July and August
•
Early growing season: September, October, November, December
The time of day was categorised into three categories:
•
Morning: >05:00 – 10:00
•
Midday: >10:00 – 14:00
•
Afternoon: >14:00 – 19:00
No night-time observations were made due to logistic constraints.
57
Location
The position at the middle of the herd or where the individual was standing was
determined by using a Garmin Global Positioning System (GPS) instrument. These
co-ordinates were used to plot wildebeest distribution on a map of the study area.
Group composition
•
Males: identified by distinct male characteristics
•
Females: identified by distinct female characteristics
•
Subadults: young animals,
• \HDU ROG EXW QRW KDYLQJ UHDFKHG UHSURGXFWLYH
maturity determined through horn development patterns
•
Calves: <1 year old.
•
Total herd size: number of animals in the herd.
Social group
Three social groups were recognised based on Von Richter (1971a):
•
Bachelor herds: Herd of at least three individuals consisting of only males of all
ages
•
Female herds: Herd of at least three individuals consisting of adult females,
subadults and/or calves.
•
Territorial bull: Single dominant bull occupying a territory.
For those observations that did not fit into these three categories, no attempt was
made to classify that particular observation into a social group category and these
were omitted from any further social analyses.
Habitat type
No formal vegetation classification had been done before this study for the entire
reserve. A number of physiognomic classifications had been performed (Bancroft
1989) and portions of the reserve had been surveyed for phytosociological
classification (Grobler 1999). A phytosociological classification of the entire reserve
was initiated half-way through the present study, but the results were not available for
use in the present study (Swanepoel 2006 pers. comm.)6. Five broad habitat types or
homogeneous units, all utilised by the wildebeest, were identified subjectively from
6
Miss. A. Swanepoel. MSc student, Department of Botany, University of Pretoria, Pretoria
0002, South Africa.
58
stereo aerial photographs and ground reconnaissance as described by Barrett
(1982):
•
Burkea woodlands
•
Moist grasslands
•
Old lands
•
Rocky grasslands
•
Sandy grasslands.
The boundaries for these habitat types were based on overall physiognomy, rock
cover, moisture regimes, dominant plant species and previous land use,
incorporating overall vegetation composition to a lesser extent. Burkea woodlands
were defined as habitats where even height stands of Burkea africana occurred in
open to moderately dense woodlands and were dominated by deep, red sandy soils
predominantly on northerly slopes. Moist grasslands were defined as habitats
occurring in wetland areas or along drainage lines with dense herbaceous vegetation
and dominated by plant species such as: Imperata cylindrica, Aristida junciformis,
Eragrostis nindensis and Paspalum urvillei. Old lands were defined as habitats
occurring on relatively flat plains, with no rock cover and where crops were cultivated
in the past. The dominant plant species in this habitat were: Digitaria eriantha,
Cynodon dactylon and Eragrostis curvula. Rocky grasslands were defined as
habitats predominantly occurring on slopes with shallow soils and with the
occurrence of the plant Xerophyta retinervis. The rock cover was
• 6DQG\
grasslands were defined as habitats occurring on rolling plains and on the plateaus of
the study area (natural grassland) where the soils were deep and red and rock cover
was <30%. The areas and locations of these broad habitat types and those areas not
utilised by the black and blue wildebeest are discussed in more detail in Chapter 5.
Topography
Aspect
The major direction towards which a slope faced was measured with a compass and
categorised into the following categories: North (•º - 90º) and South (•º - 270º).
Landscape position
The landscape unit where an individual or group was sighted was recorded. These
included plains, gentle slopes, valleys (including drainage areas), and plateaus.
59
Slope
An area that is inclined at an angle of more than 2º, but <45º from the horizontal, is
defined as a slope. The slope of the land was visually estimated and the following
broad categories were used: flat (0º), gentle (>0º - 10º), moderate (>10º - 20º) and
steep (>20º).
Erosion
The degree of erosion in the area of occupation was categorised as follows (adapted
from Theron 1991):
•
Low: Small areas with exposed soils, but with the soil mantle generally intact
•
Moderate: Larger areas with exposed soils, signs of sheet erosion, with a low
plant cover
•
High: Distinct signs of dongas and a high degree of soil loss.
Rock cover
The proportion of the surface covered by rock within the area of occupation was
estimated visually as a percentage.
Altitude
The height above sea level in metres at the site was determined from a topocadastral
map with the GPS recording used for verification.
Geomorphology
Geomorphology described the shape of the landscape and has an influence on the
drainage and erosion in an area. Three categories were used: flat, concave and
convex.
Other physical factors
Time since last burn
Records of the reserve management for accidental and block burns were used to
determine the time since last burn at the location.
Distance from water
The available water sources were noted throughout the year. The distance in metres
from the site to the nearest water source was determined from a topocadastral map.
60
Exposure
The exposure to the sun at the site was categorised as: shade, partial shade and full
sun.
Distance to shade
The distance from the position of a wildebeest to the nearest suitable shade was
estimated visually in metres.
Weather
Temperature
The screened ambient temperature (not exposed to rays of the sun) was measured
by using a digital thermometer and categorised into three classes: <15°C,
•
-
25°C, and >25°C.
Cloud cover
The percentage cloud cover was visually estimated and categorised into three
classes: 0% (Clear skies), >0 – 50% (Partly cloudy), >50% (Overcast).
Wind velocity and direction
Wind velocity was assessed from an adapted Beaufort scale as: none: 0 - 2 km/h
(smoke rises vertically); slight: >2 - 5 km/h (direction of wind shown by smoke drift,
wind felt on face, leaves rustle); moderate: >5 - 13 km/h (leaves in constant motion,
raises dust, a moderate breeze); severe: >13 km/h (trees begin to sway, fresh to
strong breeze). The wind direction was measured with a compass in degrees and
classified into the following categories: North; Northeast; East; Southeast, South;
Southwest; West; and Northwest.
Vegetation structure
Vegetation structure is the organisation in space of the plant individuals that form a
vegetation type, the primary elements of which are growth form, stratification (height
class) and cover (canopy cover).
Woody plant cover
The woody plant cover was assessed visually and categorized into the following
classes:
•
No woody vegetation: 0% cover
•
Sparse: > 0-10% cover
61
•
Open: > 10-20% cover.
Grass cover
The horizontal herbaceous structure was assessed visually and categorised into the
following classes:
•
Sparse: Grasses sparsely spread in areas, with annual grasses and forbs
•
Medium: A moderate grass canopy cover with occasional open areas
•
Dense: High grass canopy cover with little or no open areas.
Grass height
The vertical herbaceous structure was assessed at two levels. The first level was the
height of the entire grass plant, including the inflorescence, and the second level was
the height of the top grass leaves excluding the inflorescence. The mean height of
the dominant vegetation within the site of occupation was recorded with a tape
measure. The grass leaf height was categorised into four classes namely: 0 - 50 mm,
>50 - 100 mm, >100 - 400 mm and >400 mm. The total grass height was categorised
into three classes namely: 0 - 50mm, >50 - 500 mm and >500 - 800 mm.
Plant species composition
The dominant and sub-dominant plant species within the site of occupation were
identified, and specimens not identifiable in the field were collected for later
identification in the Schweickerdt Herbarium at the University of Pretoria or the
herbarium of the National Botanical Institute in Pretoria. All non-grass herbaceous
species were grouped under the category forb, and all woody species under the
category woody.
Forb: grass ratio
The mean forb: grass ratio was visually estimated on the site of occupation.
Plant utilisation
Utilisation was defined by the presence of any partially or totally eaten grass plants.
The degree of utilisation of the herbaceous layer by herbivores was subjectively
estimated and classified as follows:
•
Low: Tufts of grass recently grazed were wide apart and 1 - 10% of the
current season’s grass biomass had been utilised.
62
•
Moderate: Tufts of grass recently grazed were close together and >10 - 50%
of the current season’s grass biomass had been utilised.
•
High: Tufts of grass recently grazed were close together and >50 - 60% of the
current season’s grass biomass had been utilised.
•
Excessive: Tufts of grass recently grazed were extremely close together and
>60% of the current season’s grass biomass had been utilised.
Visibility
Visibility was determined by the mean visibility of an animal the size of a wildebeest
into each of the four main compass directions. Visibility into four directions was done
at just above the shoulder height of an average mature wildebeest. For the black
wildebeest this height was taken at 110 cm and for the blue wildebeest it was taken
at 130 cm (Smithers 1983). Visibility distances in metres were taken as the distance
from the position of the wildebeest to where visibility was first obscured. Distances
were estimated in metres.
Activity
At each observation the dominant activity of the individual or the group was recorded
and categorised based on a modification of Engelbrecht (1986), Wentzel (1990) and
Von Holdt (1999):
•
Grazing: More than half the group was grazing.
•
Lying down: More than half the group was lying down.
•
Walking: More than half the group was walking.
•
Standing: More than half the group was standing.
•
Other: Any activity that is not part of the above activities, including drinking,
running and grooming.
Association
Whenever any other animals were in the vicinity of a wildebeest and were closer than
100 m to it, the species was recorded.
Application of the methods
When applying these methods a number of assumptions were made based on those
suggested by Fagen (1988):
63
•
Wildebeest locations are representative of the entire population
•
Wildebeest are able to move freely to and between preferred habitats
•
Wildebeest will select habitats that provide the highest returns in terms of
energy investment
•
Natural resource availability is predictable and equal for both types of
wildebeest
•
Moving within each habitat costs the same in terms of energy expenditure.
This chapter serves as an introduction to Chapters 5, 6 and 7 and therefore no
results or discussion have been presented here. The results and discussion have
been incorporated under the relevant chapters for the various levels and components
of the habitat that have been examined in the present study.
64
CHAPTER 5: HABITAT SELECTION AND SEPARATION: MACROHABITAT
SCALE
INTRODUCTION
It is evident that large vegetation units such as savannas and grasslands form the
macrohabitat for different animal species over a broad geographical scale. However,
large vegetation units may have significant variation within their boundaries. Although
grasslands have an overall homogeneous nature, being defined as consisting of
grass species of about 0.3 to 1.5 m tall, there is substantial variation in plant species
composition, functional attributes of the plant species, productivity and vegetation
dynamics (Bredenkamp and Van Rooyen 1998). Where different structural habitat
types occur in the same region, the presence of a variety of herbivores with different
habitat preferences, may result in a broad spectrum of utilisation and allow for a
higher biodiversity and production per unit area (Van Rooyen et al. 1996).
At least two apparently functional habitat types have long been recognised in the
African grasslands (Cromsigt 2006): grasslands that are dominated by tall, bunch
grass communities with a caespitose growth form, and grasslands that are
dominated by short, stoloniferous lawn grass species, commonly referred to as
grazing lawns (McNaughton 1984; Archibald et al. 2005). Any variation between
these two extremes may also be found. Mixed grasslands consist of a combination of
patches of bunch grasses with patches of grazing lawns (Cromsigt 2006).
Grazing lawns have been described as areas where grazing promotes forage quality
in terms of increased nitrogen content (Ruess et al. 1983) and forage quantity in
terms of primary production (Hik and Jeffries 1990). They are defined as an expanse
of short grass in an immature state and have grass with a higher leaf to stem ratio
and a higher bulk density than that of tall grass stands (Verweij et al. 2006).
Compared with grazing lawns, the grass species that dominate the bunch grasslands
are of a relatively low forage quality (low protein and high fibre content), but offer a
high quantity of food in terms of standing grass biomass. Due to the high rainfall in
the study area, plant biomass increases during the growing season and the nutritive
value and digestibility of the forage decreases (McNaughton 1979). Therefore,
herbivores need to feed on grass swards kept in a favourable condition by repeated
grazing in order to optimise their intake (Fryxell 1991). This results in the creation of
patches within the tall grasses that can be referred to for the purposes of this study
65
as grazing sites, which have some of the characteristics of grazing lawns. No areas
in the study area could be described as pure grazing lawns as the grazing sites
referred to here were interspersed with taller grasses at varying intervals. A number
of areas in the study area could be classified as grazing sites due to the repeated
grazing by herbivores in these sites. It was therefore thought appropriate to use the
term mixed grasslands for those areas that were consistently utilised by the herbivore
species creating a patch mosaic of mostly short immature grasses and some tall
unpalatable grasses. These areas occurred throughout the study area and were
extremely patchy in extent. They could therefore not form a basis for habitat
delineation. The grazing sites (or feeding sites) within the mixed grassland areas
were, however, analysed in more detail in Chapter 7.
It has been pointed out that the a priori decisions necessary in defining habitat
boundaries can result in spurious inferences (Porter and Church 1987). Therefore
habitat division decisions should be based on overall functional differences that have
already been proven in the particular vegetation type under study.
It was therefore decided in the present study to divide the study area into broad
habitat types (hereafter referred to as habitats) based on soil moisture regime, soil
type, physiography (rock cover) and past land use. Within the grassland vegetation
type there may also be areas with localised growth of woody vegetation due to
particular environmental conditions such as rocky outcrops and the occurrence of
certain soil types and such areas could also be considered as separate habitats.
Once these habitats have been identified, a habitat utilisation study at the
macrohabitat level could be conducted to indicate which of these habitats may have
been preferred by a particular herbivore species (Novellie 1990).
Studies of the detailed vegetative characteristics of the delineated habitats could
further provide the mechanisms leading to the reasons why certain habitats are
preferred by a certain species and why others are utilised to a lesser degree than
expected. Therefore ecological separation between the black and blue wildebeest
may be shown by the differential utilisation of the different habitats that are present at
Ezemvelo Nature Reserve, which may be caused by differences at various levels in
the characteristics of their vegetation and physical character. In addition, the
identification of those habitats on which there is a potential for conflict between two
wildlife species may also be possible from such a study (Barrett 1982).
66
When a habitat is used disproportionately to its availability, use of that habitat is said
to be selective (Johnson 1980). A number of biological factors may affect habitat
selection studies, including variations among subpopulations (sex and age groups),
fluctuations in population size, lack of independence of individuals due to territoriality
or aggregation, traditional use of resources and/or the local occurrence of
competitors and predators (Thomas and Taylor 1990). An understanding of these
influences is fundamental to the interpretation of the results of such a study.
Some portions of a habitat may not be used or underutilised while others may be
selected and overutilised (Van Rooyen et al. 1996). Management practices could be
used to reduce this impact. Fire and salt licks could be used to attract animals to
underutilised areas and the opening and closing of waterpoints could be a way to
attract wildlife away from overutilised areas. Patterns of habitat use by both types of
wildebeest at Ezemvelo Nature Reserve would therefore also aid in developing
management proposals for the future of the two types of wildebeest present on the
reserve. In order to determine whether a habitat is underutilised or selected by a
particular type of wildebeest, information on the availability of the habitat in question
must be obtained (Manly et al. 1993). Even if a type of wildebeest occurs within a
habitat, this does not necessarily imply that the habitat in question is being selected
for. When comparing the occurrence of any type of wildlife and the availability in
terms of surface area, information on positive or negative selection may be obtained.
Both the physical structure and the vegetation of a habitat may influence its selection
by an animal species. By analysing the vegetation of the various habitat types in an
area, the relationship between the habitat preferences of an animal species and the
characteristics of the vegetation of that habitat may become clear (Reilly 1989). If few
species differences in the vegetation of two habitats can be discerned, physical
feature differences may be the governing factor for selection of one or the other
habitat by an animal species.
Species composition and structure are the two most important components of the
vegetation that form part of the habitat of an individual animal (Von Holdt 1999). The
species that constitute the vegetation type will determine whether or not the food
resource is potentially sufficient for herbivores (Wentzel et al. 1991). The structure of
the vegetation will determine the availability of food at certain height classes (Bothma
and Van Rooyen 1996) as well as the availability of shade for daytime resting. The
availability of shade may in turn provide opportunities for escape from climatic
67
extremes (Cromhout 2006). Plant phenology may also be an important factor
governing habitat selection as deciduous trees lose their leaves in winter and hence
less protection against extreme environmental conditions may be experienced by a
species (Krüger 1996; Cromhout 2006). Habitat selection may therefore change
through the seasons due to the availability of certain plant parts or nutrients within
the plants eaten. A survey of the herbaceous characteristics of the preferred habitat
of each type of wildebeest on Ezemvelo Nature Reserve could therefore be used to
determine if the two types of wildebeest present selected habitats that are
characterised by certain plant species or vegetation structure or not (Reilly 1989).
The null hypothesis to be tested in this part of the present study was that the black
and blue wildebeest at Ezemvelo Nature Reserve would utilise the available habitats
in proportion to their occurrence.
The following key questions were therefore addressed:
•
What are the broad habitat preferences of the two types of wildebeest
present, and what are the associated herbaceous characteristics of each
habitat?
•
Are there any seasonal differences in the broad habitat choices of the two
types of wildebeest?
•
Are there any social group differences in the broad habitat choices of the
two types of wildebeest?
MATERIALS AND METHODS
Habitat delineation
The five broad habitats described in Chapter 4 were used to assess habitat use by
the two types of wildebeest at Ezemvelo Nature Reserve. These habitats are
mapped in Figure 5.1.
The approximate areas covered by each of these broad habitats was 2 933 ha
(sandy grasslands), 2 540 ha (rocky grasslands), 744 ha (old lands), 658 ha (moist
grasslands), and 123 ha (Burkea woodlands). Of the surface area available to both
types of wildebeest, the sandy grasslands formed 42% of such area, the rocky
grasslands formed 36%, the old lands formed 11%, the moist grasslands formed 9%
and the Burkea woodlands formed 2% of the area. The availability of each habitat on
68
km
Figure 5.1: The broad habitat types found at Ezemvelo Nature Reserve, South Africa.
69
the study area was determined from recent stereo aerial photographs and
topographical maps by using Arc View 3.2®.
The old lands consisted predominantly of short stoloniferous lawn grass species
interspersed with tall unpalatable grass species at wide intervals. This habitat was
thus considered to be structurally most similar to grazing lawns but was classified as
mixed grasslands for the purposes of the present study due to the reasons
mentioned in the introduction to this chapter. The rocky grasslands, sandy
grasslands and moist grasslands represented the bunch grass communities on
Ezemvelo Nature Reserve. Spatial heterogeneity within the bunch grass community
included differences in rock cover, soil type and soil moisture regime. Certain areas
within these bunch grass communities consisted of patches of heavily grazed
grasses and could hence also be considered as mixed grassland. Given the mobility
of both types of wildebeest in the study area, it was assumed that all these habitats
were equally available physically to all wildebeest.
Two other habitats were identified at Ezemvelo Nature Reserve, ie. rocky slopes and
riverine bush (Figure 5.1). The rocky slopes were dominated by woody vegetation
and the slopes were >20°, making it inaccessible to both types of wildebeest. The
rocky slope vegetation type was evident on the rocky outcrops and the rocky areas
near the Wilge River. It comprised 15% of the surface area of the reserve. Riverine
bush comprised only 5% of the surface area of the reserve and was restricted to the
banks of the Wilge River. The riverine bush habitat did not form part of the area
utilised by either type of wildebeest except for the dry parts of the year when it was
visited for drinking water when the other watering holes were dry. It was also
excluded from the present study, as it was not found not to be an important
component of the overall habitat utilisation of the black and blue wildebeest at
Ezemvelo Nature Reserve.
The present study followed Study Design I of Thomas and Taylor (1990) where
analysis at the population level was conducted (no individuals were identified) and
where the availability of each habitat was measured. Preference for the different
habitats was analysed by using the data collected during the surveys for wildebeest
described in Chapter 4. These surveys provided frequency of occurrence of both
types of wildebeest in each habitat over time. Thomas and Taylor (1990) suggested
that at least 50 observations on at least 20 animals would be required for adequate
hypothesis testing, and this was achieved in the present study by the 1 558
70
observations that were recorded from January 2004 to August 2005. The distribution
of the two types of wildebeest within the habitats was also mapped (Figure 5.2).
The availability of a habitat is the quantity of that habitat that is accessible to a
population of wildebeest during the study period (Manly et al. 1993). Preference
implies that the preferred resource is utilised to a significantly greater proportion than
its availability would suggest (Thomas and Taylor 1990). Therefore in order to assess
preference for a particular habitat by a particular species the amount of that habitat
available must be taken into consideration to ensure that mere presence in a widely
available habitat is not inferred to mean that the particular habitat under investigation
is selected for.
Habitat herbaceous characteristics
A minimum of four and a maximum of eight vegetation plots were located at random
in each of the five habitats to assess herbaceous characteristics. The number of plots
analysed depended on the size of each habitat. The surveys were done on these
sites during April, August and December 2004, representing the end of each of the
three ecological seasons, for those characteristics which were likely to change over
the seasons (Dörgeloh 1998).
The following vegetation parameters were measured at each site: grass species
composition, grass species density (species/m2), above-ground standing crop
(kg/ha), total grass height (cm), grass tuft height (cm), grass canopy cover (%), and
grass basal cover (%). Species density, species diversity, veld condition, degree of
utilisation and biomass concentration were calculated from the above variables.
71
km
Figure 5.2: The distribution of the black and blue wildebeest at Ezemvelo Nature
Reserve, South Africa from January 2004 to August 2005.
72
Relative grass species composition, diversity, density, degree of utilisation and veld
condition
For this part of the study, the sites within each habitat type were surveyed only once
for the entire study period. At each site, 20 quadrants of 0.5 x 0.65 m (Barnes et al.
1982) were placed 5 m apart along a single north to south transect. Grass species
were identified and the percentage cover of each rooted grass species within a
quadrant was estimated visually and recorded.
Forbs were treated as a single category. Species density was used as a measure of
plant species richness and was calculated from the number of plant species per 6.5
m2, which is the total area covered by the 20 quadrants for each site. This species
density value was converted to species per m2 for analysis.
The Shannon-Wiener diversity index was calculated from the species frequencies
using the following equation (Magurran 1991):
H’ = -™Silnpi
where H’ is the Shannon-Wiener diversity index and pi is the proportion of individuals
found in the ith species. The Shannon-Wiener index combines both species richness
and evenness in a single value (Magurran 1991). This index measures the degree of
uncertainty of predicting the species of an individual that is selected at random from
the community. This uncertainty increases as the number of species and equitability
increases (Magurran 1991).
The step-point method was used to assess veld condition (Tainton 1999). The
recommended sample size for assessing veld condition with this method in
grasslands is 200 points per site (Mentis 1981; Hardy and Walker 1991). Each
transect consisted of two parallel lines of 200 m long spaced 20 m apart along a
north to south direction. At every second pace, the end of a measuring staff was
grounded and the grass plant that was hit by it was identified (Dörgeloh 1998). The
grass species were grouped into five ecological classes based on their response to
grazing, perceived grazing value, phytomass production and palatability (Fourie and
Visagie 1985; Trollope et al. 1989; Van Oudtshoorn 1999; Bothma et al. 2004).
These ecological classes were defined as follows (Bothma et al. 2004):
73
Class 1 species
Includes valuable and palatable tufted or stoloniferous grass species with a high
productivity and a high grazing value.
Class 2 species
Includes tufted, perennial grass species with an intermediate productivity and
moderate grazing value.
Class 3 species
Includes tufted, tall, perennial grass species with a high productivity but a low grazing
value.
Class 4 species
Includes generally unpalatable, annual and perennial, tufted or stoloniferous grass
species with an intermediate productivity and a low grazing value.
Class 5 species
Includes unpalatable, annual grass and forb species with a low productivity and a low
grazing value.
A modification of the ecological index method (Trollope 1990; Tainton 1999) as
described by Bothma et al. (2004) was used to calculate the veld condition index for
each site.
The degree of past utilisation (DOU) to which the herbaceous layer in this study area
had been subjected to was established by means of a utilisation index: Class 2,
Class 3, Class 4 and Class 5 grass species categories were each allotted values of
–1.0, 0.33, 0.67 and 1.0 respectively (Wentzel et al. 1991). No value was allocated to
the Class 1 grass species as their abundance can decrease under the effect of overor underutilisation. The frequency value of every grass species category was
multiplied by the index values and summed.
Grass cover and grass structure
The step-point method (Holechek et al. 1989) was used to assess grass cover and
grass structure (Dörgeloh 1998). Each transect consisted of two parallel lines 200 m
long and 20 m apart. At every second pace, the end of a measuring staff was
74
grounded and the height above the ground level of the inflorescence (total grass
height) and of the leaves (height below which 80% of the leaves occur (Shackleton
1990)) (grass leaf height) was measured with the graduated staff (Voisin 1988). It
was also recorded whether the strike was a canopy strike, or if bare ground was hit.
This enabled the calculation of the percentage canopy cover and percentage bare
ground along the transect.
Above-ground standing crop and biomass concentration
The disc pasture meter was used to measure the above-ground standing crop.
(Trollope and Potgieter 1986). At each sample site, 100 readings of the settling
height of the disc were taken along two demarcated transects, each 100 m long but
20 m apart) at every 2 m interval (Dörgeloh 1998). The relationship between the disc
height readings and the above-ground standing phytomass (kg/ha) was calculated
from published regression equations (Trollope and Potgieter 1986). The equation
utilsed was as follows:
y = -3019+2260x
where y = estimated fuel load (kg/ha) and x = mean disc height (cm).
To calculate biomass concentration (BC) (kg/m3), grass leaf height (cm) and biomass
(kg/ha) was used in the equation as an indication of the amount of leaf material
present (Dörgeloh 1998): The following equation was used:
BC=(BMASS/10 000) x (100/GLH)
where BC=Biomass concentration in kg/m3
BMASS= Biomass in kg/ha
GLH= Grass leaf height in cm.
Statistical analysis
Data from all the sites within each habitat were pooled for statistical analysis.
Observations were weighted according to the herd size variable as described in
Chapter 4. The Chi-squared goodness of fit test was used to compare the habitat use
by the two types of wildebeest and availability of the habitat statistically (Thomas and
Taylor 1990; Manly et al. 1993). Expected frequencies were calculated from the
75
available proportions of habitats. This was achieved by using the area of each habitat
and converting it to a proportion of the total area available. If the Chi-squared test
was found to be significant, the null hypothesis that all habitats were used in
proportion to their availability on the study area (no selection) was rejected.
Subsequently, the cell Chi-squared values for each habitat type were calculated. If
these values were significant, then the differences between the observed and
expected values were examined. If the observed value was greater than the
expected value, a positive selection for that habitat type was indicated. It was here
concluded that the type of wildebeest under investigation actively selected the habitat
type in question and therefore utilised it in greater proportion than its availability. If
the expected value was greater than the observed value, it was concluded that the
habitat type involved was utilised to a lesser degree than expected based on its
availability by the relevant type of wildebeest. For comparison, the Bonferroni
adjusted 100 (1-. FRQILGHQFH LQWHUYDOV IRU KDELWDW XVH ZHUH FDOFXODWHG IRU HDFK
KDELWDW W\SH ZKHUH
. LV %\HUV et al. 1984; Manly et al. 1993). A habitat was
selected for use by the wildebeest type under investigation if the lower confidence
interval for that habitat type was greater than the corresponding wildebeest
population proportion within that habitat type. Similarly, a habitat was not preferred
when the upper confidence interval for that habitat type excluded the corresponding
wildebeest population proportion in that habitat type (Namgail et al. 2004). Spatial
and temporal differences in vegetation characteristics between habitat types were
tested with general linear modelling by using the PROC GLM procedure at 95%
confidence intervals (SAS®7). This procedure involved an analysis of variance.
Where a significant difference was found between all the habitat types, differences
between individual habitat types were further investigated by using multiple
comparisons.
To detect differences in habitat use across the seasons, the data were divided into
three ecological seasons, being the late growing season (January to April), the
dormant season (May to August) and the early growing season (September to
December) as was described in Chapter 2. To detect differences in habitat use
among the three social groups of wildebeest, the social class involved in each
observation was classified as a territorial bull, a bachelor herd or a female herd
following Von Richter (1971a).
7
Integrated system of software providing complete control over data management, analysis
and presentation. Version 8.2 on UP mainframe, SAS Institute Inc. SAS Campus Drive, Cary,
North Carolina 27513
76
RESULTS
The percentage frequency of the various observations on black and blue wildebeest
appear in Table 5.1. The habitat selection results of the black and blue wildebeest at
Ezemvelo Nature Reserve based on using the cell Chi-squared calculations and
those based on the Bonferroni-adjusted confidence intervals were almost identical
(Tables 5.2, 5.3 and 5.4). Based on these results and the recommendations of
Groeneveld 2006 (pers. comm.)8, it was therefore decided to provide habitat use
results for both methods only for the seasonal and total data, and to provide the cell
Chi-squared habitat selection results for the more detailed analyses that were also
performed.
Overall habitat utilisation by all wildebeest social groups
There was a difference in the extent to which the black and blue wildebeest utilised
the five broad habitats (Tables 5.2 and 5.3). Neither type of wildebeest utilised all five
habitats in proportion to their availability in the study area (Black wildebeest: Ø2 =
194.8; df = 4; p<0.0001; Blue wildebeest: Ø2 = 549.9; df = 4; p<0.0001). Black
wildebeest selected the sandy grasslands in 62% of the observations and the moist
grasslands in 12% of the observations, showing a positive selection for both these
habitats, whereas the Burkea woodlands were never used (Table 5.2). The old lands
were used in proportion to their availability in the study area, whereas the rocky
grasslands were used to a lesser degree than expected. Blue wildebeest selected
the old lands in 27% of the observations, showing a positive selection for it and for
the Burkea woodlands. Blue wildebeest utilised the sandy grasslands in proportion to
their availability on the study area, but utilised the rocky and moist grasslands to a
lesser degree than expected (Table 5.3).
Seasonal influence
Late growing season
There was a strong association between type of wildebeest and the type of habitat
used during the late growing season (Black wildebeest: Ø2 = 70.0; df = 4; p<0.001;
Blue wildebeest:Ø2 = 484.9; df = 4; p<0.0001).
8
Prof. H. Groeneveld. Department of Statistics, University of Pretoria, Pretoria, 0002, South
Africa.
77
Table 5.1: The percentage occurrence of the black and blue wildebeest in the five
broad habitat types indicating the utilisation of the various habitat types over three
ecological seasons and for the total data at Ezemvelo Nature Reserve from January
2004 to August 2005
Size and season
Area of habitat type (ha)
Early growing season
Dormant season
Late growing season
Total data
Type of
Burkea
Moist
wildebeest
woodland
grassland
-
123
658
Black wildebeest
0.0
Blue wildebeest
Rocky
Sandy
grassland
grassland
744
2540
2933
7.5
24.7
12.9
54.9
8.2
3.1
33.8
23.6
31.4
Black wildebeest
0.0
11.0
10.7
8.3
70.0
Blue wildebeest
6.6
3.2
23.7
25.5
41.1
Black wildebeest
0.0
19.0
10.5
8.9
61.7
Blue wildebeest
10.4
6.8
23.3
20.4
39.2
Black wildebeest
0.0
12.2
15.5
10.1
62.1
Blue wildebeest
8.4
4.3
26.9
23.2
37.3
78
Old land
Table 5.2: Chi-squared test results to evaluate the hypothesis that the black wildebeest on
Ezemvelo Nature Reserve used the available broad habitats in proportion to their occurrence
by surface area. Values in brackets indicate sample sizes of <5 and therefore the Chisquared test results for these entries may be invalid. + indicates a positive selection, indicates a negative selection and 0 indicates random selection. N/a indicates that that habitat
type was not utilised at all
Social
Habitat
group
Ecological season
Late growing season
Ø
df
Selection
18.15
1
36.07
Dormant season
Ø
df
Selection
+
28.39
1
1
-
37.01
12.73
1
+
0.05
1
0
n/a
n/a
0.02
1
0
(0.48)
1
Old lands
(0.43)
Moist
Early growing season
df
Selection
+
54.95
1
+
1
-
108.92
1
-
0.02
1
0
3.10
1
0
+
34.82
1
+
18.62
1
+
n/a
n/a
n/a
n/a
n/a
n/a
n/a
1.08
1
0
0.18
1
0
0.12
1
0
0
(4.33)
1
-
(3.20)
1
0
7.42
1
-
1
0
(0.50)
1
0
(0.99)
1
0
(1.84)
1
0
(3.76)
1
0
53.42
1
+
15.57
1
+
63.56
1
+
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
(0.05)
1
0
(2.06)
1
0
(1.79)
1
0
1.76
1
0
(0.81)
1
0
(1.60)
1
0
(1.86)
1
0
(3.96)
1
-
Old lands
(6.16)
1
+
(0.05)
1
0
(0.09)
1
0
(1.78)
1
0
Moist
(0.04)
1
0
(0.02)
1
0
(0.12)
1
0
(0.0002)
1
0
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
21.21
1
+
34.79
1
+
8.34
1
+
62.29
1
+
36.30
1
-
31.10
1
-
31.13
1
-
98.32
1
-
Old lands
11.4
1
+
0.16
1
0
0.43
1
0
4.36
1
+
Moist
0.59
1
0
0.12
1
0
23.55
1
+
3.31
1
0
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
All groups
Sandy
combined
grasslands
Rocky
2
Ø
df
Selection
+
9.43
1
1
-
35.99
0.44
1
0
4.78
1
Total data
Ø
2
2
2
grasslands
Old lands
Moist
grasslands
Burkea
n/a
n/a
n/a
woodlands*
Bachelor
Sandy
herds
grasslands
Rocky
grasslands
grasslands
Burkea
n/a
woodlands*
Territorial
Sandy
bulls
grasslands
Rocky
grasslands
grasslands
Burkea
woodlands*
Female
Sandy
herds
grasslands
Rocky
grasslands
grasslands
Burkea
n/a
n/a
woodlands*
* The Burkea woodlands were never utilised by the black wildebeest in the study area
79
n/a
Table 5.3: Chi-squared test results to evaluate the hypothesis that the blue wildebeest on
Ezemvelo Nature Reserve used the available broad habitats in proportion to their occurrence
by surface area (values in brackets indicate sample sizes of <5 and therefore the chi-squared
test results for these entries may be invalid). + indicates a positive selection, - indicates a
negative selection and 0 indicates random selection
Social group
Habitat
Ecological season
Late growing season
2
df
Selection
10.91
1
35.10
Dormant season
Ø
df
Selection
-
1.03
1
1
-
13.13
397.79
1
+
20.67
1
-
20.48
1
1.35
Early growing season
Total data
Ø
df
Selection
0
0.68
1
0
0.81
1
0
1
-
42.22
1
-
83.07
1
-
57.08
1
+
39.33
1
+
389.91
1
+
9.62
1
-
0.12
1
0
19.74
1
-
+
0.02
1
0
84.66
1
+
56.37
1
+
1
0
0.73
1
0
0.35
1
0
2.32
1
0
0.48
1
0
1.04
1
0
1.38
1
0
0.29
1
0
Old lands
(1.61)
1
0
(1.52)
1
0
(1.79)
1
0
0.87
1
0
Moist
(0.58)
1
0
(0.57)
1
0
(0.48)
1
0
0.17
1
0
53.50
1
+
(0.67)
1
0
(10.80)
1
+
39.42
1
+
1.55
1
0
0.12
1
0
0.61
1
0
1.84
1
0
0.58
1
0
0.98
1
0
1.68
1
0
3.04
1
0
Old lands
10.08
1
+
5.00
1
+
7.49
1
+
21.95
1
+
Moist
(1.26)
1
0
(2.40)
1
0
(0.73)
1
0
4.24
1
-
(18.37)
1
+
(17.76)
1
+
24.08
1
+
59.37
1
+
8.07
1
-
2.57
1
0
2.06
1
0
0.001
1
0
43.04
1
-
16.77
1
-
41.17
1
-
95.18
1
-
Old lands
462.66
1
+
66.38
1
+
30.32
1
+
428.15
1
+
Moist
(19.47)
1
-
10.12
1
-
0.29
1
0
18.61
1
-
1.12
1
0
(3.40)
1
0
53.38
1
+
10.71
1
+
Ø
All groups
Sandy
combined
grasslands
Rocky
2
2
2
Ø
df
Selection
grasslands
Old lands
Moists
grasslands
Burkea
woodlands
Bachelor
Sandy
herds
grasslands
Rocky
grasslands
grasslands
Burkea
woodlands
Territorial
Sandy
bulls
grasslands
Rocky
grasslands
grasslands
Burkea
woodlands
Female
Sandy
herds
grasslands
Rocky
grasslands
grasslands
Burkea
woodlands
80
Table 5.4: Bonferroni confidence intervals calculated to determine the seasonal broad habitat
selection by black and blue wildebeest on Ezemvelo Nature Reserve, South Africa from
January 2004 to August 2005 relative to the total land surface area of the reserve. N/a
indicates those habitats where that type of wildebeest was never encountered
i
Variable
Black wildebeest
ui
oi
Blue wildebeest
Blower
Bupper
si
ui
oi
Blower
Bupper
si
Late growing season
Sandy grasslands
0.419
105.79
0.63
0.54
0.73
+
106.10
0.30
0.24
0.37
-
Rocky grasslands
0.363
13.93
0.08
0.03
0.14
-
59.86
0.17
0.12
0.22
-
Old lands
0.106
32.76
0.08
0.12
0.27
+
158.15
0.45
0.39
0.52
+
Moist grasslands
0.094
14.85
0.20
0.03
0.15
0
6.74
0.02
0.00
0.04
-
Burkea woodlands
0.018
0.00
0.00
n/a
n/a
n/a
17.61
0.05
0.02
0.08
+
1.000
167.34
1.00
348.46
1.00
Sandy grasslands
0.419
119.92
0.68
0.59
0.77
+
163.85
0.45
0.39
0.52
0
Rocky grasslands
0.363
15.44
0.09
0.03
0.14
-
89.64
0.25
0.19
0.31
-
Old lands
0.106
15.88
0.09
0.03
0.15
0
85.04
0.23
0.18
0.29
+
Moist grasslands
0.094
25.53
0.14
0.08
0.21
0
15.88
0.04
0.02
0.07
-
Burkea woodlands
0.018
0.00
0.00
n/a
n/a
n/a
6.83
0.02
0.00
0.04
0
Total
1.000
176.76
1.00
361.24
1.00
Sandy grasslands
0.419
95.56
0.57
0.47
0.67
+
148.48
0.45
0.38
0.52
0
Rocky grasslands
0.363
13.87
0.08
0.03
0.14
-
48.98
0.15
0.10
0.20
-
Old lands
0.106
18.32
0.11
0.05
0.17
0
75.27
0.22
0.16
0.28
+
Moist grasslands
0.094
39.05
0.23
0.15
0.32
+
33.05
0.10
0.06
0.14
0
Burkea woodlands
0.018
0.00
0.00
n/a
n/a
n/a
28.43
0.09
0.05
0.13
+
Total
1.000
166.80
1.00
331.21
1.00
Sandy grasslands
0.419
322.34
0.63
0.58
0.69
+
417.46
0.40
0.36
0.44
0
Rocky grasslands
0.363
43.25
0.08
0.05
0.12
-
200.74
0.19
0.16
0.22
-
Old lands
0.106
67.07
0.13
0.09
0.17
0
317.79
0.31
0.27
0.34
+
Moist grasslands
0.094
77.88
0.15
0.11
0.19
+
53.91
0.05
0.03
0.07
-
Burkea woodlands
0.018
0.00
0.00
n/a
n/a
n/a
51.24
0.05
0.03
0.07
+
Total
1.000
510.54
1.00
1041.14
1.00
Total
Dormant season
Early growing season
Total data
NRWHVi, proportion of habitat type available; ui, number of used resource units in category i (weighted according to
herd size); oi, the proportion of used units in category i; Blower, Lower Bonferroni-adjusted 95% confidence intervals for
habitat type use; Bupper, Upper Bonferroni-adjusted 95% confidence intervals for habitat type use; si, selection of
habitat category i.
81
Of the five broad habitats within the range of the wildebeest found within the study
area, the black wildebeest positively selected the sandy grasslands and the old lands
during the late growing season (Table 5.2). Black wildebeest used the rocky
grasslands and Burkea woodlands to a lesser degree than expected, but utilised the
moist grasslands in proportion to their occurrence on the study area.
Blue wildebeest significantly selected the old lands and the Burkea woodlands, but
they showed a negative selection for the sandy grasslands, rocky grasslands and
moist grasslands during the late growing season (Table 5.3).
Dormant season
During the dormant season there was a strong overall relationship between the type
of wildebeest and the type of habitat (Black wildebeest: Ø2 = 73.8; df = 4; p<0.001;
Blue wildebeest: Ø2 = 80.9; df = 4; p<0.001). Black wildebeest significantly selected
the sandy grasslands and the moist grasslands but utilised the rocky grasslands to a
lesser degree than expected (Table 5.2). Black wildebeest utilised the old lands in
proportion to their availability in the study area but never occurred in the Burkea
woodlands. Blue wildebeest significantly selected the old lands during the dormant
season and showed a negative selection for the moist grasslands and rocky
grasslands (Table 5.3).
Early growing season
During the early growing season there was a strong relationship between the type of
wildebeest and the type of habitat (Black wildebeest: Ø2 = 83.3; df = 4; p<0.001; Blue
wildebeest: Ø2 = 167.0; df = 4; p<0.0001). The black wildebeest showed a significant
selection for the sandy grasslands and moist grasslands and utilised the rocky
grasslands to a lesser degree than expected (Table 5.2). They were never found in
the Burkea woodland habitat type but selected the old lands in proportion to their
availability on the study area. Blue wildebeest significantly selected the Burkea
woodlands and old lands but utilised the rocky grasslands to a lesser degree than
expected during the early growing season (Table 5.3). The sandy grasslands and
moist grasslands were utilised in proportion to their availability on the study area.
82
Between season comparisons
A comparison of the habitat selection of the black and blue wildebeest for the five
broad habitats in the study area over the ecological seasons indicated which habitats
were selected and those which were utilised to a lesser degree than expected at the
different times of the year (Tables 5.2 and 5.3). Even though the black wildebeest
actively selected the sandy grasslands throughout the study period, it was most likely
to utilise the sandy grasslands during the dormant season (Ø2 = 28.4; df = 1; p<0.001;
Table 5.2). Blue wildebeest did not actively select the sandy grasslands but utilised
them in proportion to their availability in the study area throughout most of the study
period, except for the late growing season when they were were utilised to a lesser
degree than expected. Blue wildebeest utilised the rocky grasslands to a lesser
degree than expected throughout the study period but were most likely to do so
during the dormant season. Black wildebeest utilised the rocky grasslands to a lesser
degree than expected with an equal intensity year round. Black wildebeest did not
actively select the old lands and utilised them in proportion to their availability
throughout most of the study period, except for the late growing season when they
were selected for. Blue wildebeest actively selected the old lands throughout the
study period but were most likely to select the old lands during the late growing
season. Black wildebeest did not occur in the Burkea woodlands at any time, but blue
wildebeest actively selected the Burkea woodlands throughout the study period
except for during the dormant season when they were used in proportion to their
availability in the study area. Blue wildebeest were most likely to select the Burkea
woodlands during the early growing season. Black wildebeest actively selected the
moist grasslands throughout most of the study period except during the late growing
season when they were utilised in proportion to their availability in the study area.
Black wildebeest were most likely to select the moist grasslands during the early
growing season. Blue wildebeest avoided the moist grasslands throughout the study
period except during the early growing season.
Social group influence
Bachelor herds
There were a limited number of bachelor herds of both types of wildebeest in the
study area. Therefore the data set for this social group was not as robust as the data
sets for the other social groups. Black wildebeest bachelor herds showed a
significant positive selection for the moist grasslands and tended to utilise the rocky
83
grasslands to a lesser degree than expected (Table 5.2). They utilised the sandy
grasslands and old lands in proportion to their availability in the study area. All
habitats were utilised in accordance with their availability in the late growing season,
while in the dormant season the bachelor herds selected the moist grasslands and
showed a negative selection for the rocky grasslands. In the early growing season
the black wildebeest bachelor herds selected the moist grasslands but used all the
other habitat types in proportion to their availability in the study area.
Blue wildebeest bachelor herds were seldom encountered, but when they were
encountered they utilised all the habitats in proportion to their availability in the study
area except the Burkea woodlands, which they seemed to actively select throughout
the study period.
Territorial bulls
Black wildebeest territorial bulls seldom occurred alone and usually formed part of a
female herd. Therefore the data set for black wildebeest territorial bulls is small due
to the clouding of these data. Black wildebeest territorial bulls therefore had a similar
distribution in terms of habitat choice, as did the female herds. However, their
frequency of occurrence individually was too low for the Chi-squared tests for broad
habitat type selection to be valid (Table 5.2).
Lone blue wildebeest territorial bulls were encountered with a higher frequency than
black wildebeest bulls, but the sample sizes were also too small for detailed seasonal
analysis. The results did, however, indicate that blue wildebeest bulls tended to show
a significant selection for the Burkea woodlands and old lands and a negative
selection for the moist grasslands. Blue wildebeest territorial bulls utilised the sandy
grasslands and rocky grasslands in proportion to their availability in the study area
(Table 5.3).
Female herds
An analysis of the total data revealed that black wildebeest female herds never
occurred in the Burkea woodlands and that the frequency of occurrence in the rocky
grasslands was lower than expected. They actively sought out the sandy grasslands
and old lands and utilised the moist grasslands in proportion to their availability in the
study area. An analysis of the seasonal data indicated that the above pattern of
habitat selection occurred during the late growing season but that in the dormant
season the old lands were no longer selected but were rather utilised in proportion to
84
their availability in the study area. In the early growing season, black wildebeest
female herds showed the same pattern of habitat selection as that of the dormant
season except that they tended to select the moist grasslands instead of utilising
them in proportion to their availability in the study area.
An analysis of the total data revealed that the blue wildebeest female herds actively
selected the old lands and the Burkea woodlands but utilised the rocky grasslands
and moist grasslands to a lesser degree than expected (Table 5.3). They utilised the
sandy grasslands in proportion to their availability in the study area. An analysis of
the seasonal data indicated that during the late growing season a similar pattern to
that observed in the total data occurred, except that the sandy grasslands were
utilised to a lesser degree than expected and the Burkea woodlands were used in
proportion to their availability in the study area during the late growing season.
During the dormant season the habitats were utilised in a similar way to that
described for the total data except that the Burkea woodlands were used in
proportion to their availability in the study area instead of being positively selected. In
the early growing season the Burkea woodlands were actively selected and the moist
grasslands were utilised in proportion to their availability in the study area. The rest of
the habitats were utilised in a similar manner to that described for the total data.
Vegetation assessment by broad habitat types
During the study period a total of 51 grass species were recorded in the vegetation
surveys at Ezemvelo Nature Reserve in the broad habitats that were utilised by the
two types of wildebeest.
Ecological Class 1 grass species at Ezemvelo Nature Reserve in those habitats that
were utilised by both types of wildebeest included: Diheteropogon amplectens,
Digitaria eriantha, Monocymbium ceressiforme, Panicum natalense, Setaria
sphacelata, and Themeda triandra. Based on the results of the generalised linear
model procedures (PROC GLM), there were no significant differences in percentage
of Class 1 grass species between the broad habitats that were utilised by the black
and blue wildebeest (p = 0.4438) (Figure 5.3).
Ecological Class 2 grass species at Ezemvelo Nature Reserve in the broad habitats
that were utilised by both types of wildebeest included: Trachypogon spicatus,
Tristachya rehmannii, Schizachyrium sanguineum, Melinis nerviglumis, Imperata
85
cylindrica, Hyparrhenia filipendula, Cymbopogon excavatus, Brachiaria brizantha and
Andropogon schirensis. The generalised linear model procedures (PROC GLM)
indicated acceptance of the null hypothesis of no difference regarding the abundance
of Class 2 grass species in the respective broad habitat types (p = 0.0578) (Figure
5.3).
Ecological Class 3 grass species were generally rare at Ezemvelo Nature Reserve,
especially in the broad habitats that were occupied by the two types of wildebeest
(Figure 5.3). This indicated that tufted tall perennial grass species with a high
productivity but a low grazing value and which tend to increase with light
overutilisation are not abundant in the study area.
Ecological Class 4 grass species at Ezemvelo Nature Reserve in those broad
habitats that were utilised by both types of wildebeest included: Eragrostis
chloromelas, Eragrostis curvula, Elionurus muticus, Eragrostis racemosa, and
Sporobolus festivus. The generalised linear model procedures (PROC GLM) showed
no significant difference between the abundance of Class 4 grass species between
the broad habitat types (p = 0.1631) (Figure 5.3).
Ecological Class 5 grass species at Ezemvelo Nature Reserve in those habitats that
were utilised by both types of wildebeest included: Sporobolus africanus, Perotis
patens, Pogonarthria squarrosa, Melinis repens, Microchloa caffra, Eragrostis plana,
Eragrostis inamoena, and Eragrostis gummiflua. The results of the PROC GLM
procedures indicated that there was no significant difference in the abundance of
Ecological class 5 grass species between the habitats (p = 0.7530) (Figure 5.3).
In general, Ecological Classes 4 and 5 made a larger contribution to the grass
composition than Ecological Classes 1 and 2. Only in the rocky grasslands was there
an equal distribution of Ecological Classes 1 and 2 compared with Ecological
Classes 4 and 5. Old lands and Burkea woodlands had a low percentage of Class 2
grass species while the moist grasslands had a low percentage of Class 1 grass
species. The old lands and sandy grasslands were dominated by Class 4 grass
species, while the moist grasslands were dominated by Class 5 grass species. Class
1 grass species made a large contribution to the grass species in the Burkea
woodlands (26%), rocky grasslands (18%) and in the old lands (18%).
86
The degree of past utilisation as calculated from the ecological species composition
of each habitat type did not differ significantly between habitat types (p = 0.1281).
The rocky grasslands had the lowest degree of utilisation (~15%) of all the habitats
investigated, while all the other habitat types all were utilised at a frequency of >50%
(Figure 5.4).
The generalised linear model procedures (PROC GLM) indicated that there was no
significant difference in veld condition between the broad habitats (p = 0.9417). The
veld condition score at Ezemvelo Nature Reserve is poor (veld condition score: 350450). This indicated that the veld was generally dominated by an abundance of
Ecological Classes 4 and 5 grass species. The Burkea woodlands had the highest
veld condition score while the moist grasslands had the lowest veld condition score
of all the habitats investigated (Figure 5.5).
The Shannon-Wiener index of diversity indicated that the grass species diversity of
all the habitat types was relatively low (D<2.5), yet there was a significant difference
in grass species diversity between the habitats (p = 0.0133). The PROC GLM test
results also indicated that the grass species diversity index in the rocky grasslands
was significantly higher than that of old lands (p = 0.0025) and the moist grasslands
(p = 0.0040) (Figure 5.6).
87
P e rce nta ge com position
70
60
50
40
30
20
10
0
BW
RG
Clas s 1
OL
Clas s 2
Clas s 3
MG
Clas s 4
SG
Clas s 5
Figure 5.3: Mean percentage composition of the five ecological classes of grass
species in the five broad habitats that were utilised by the black and blue wildebeest
at Ezemvelo Nature Reserve in 2004. Bars represent the standard error of the
percentage composition. No significant differences between habitats were found. BW
= Burkea woodlands, RG = rocky grasslands, OL = old lands, MG = moist
grasslands, SG = sandy grasslands.
88
90
Degree of utilisation
80
70
60
50
40
30
20
10
0
BW
RG
OL
MG
SG
Figure 5.4: Mean (columns) and standard error (bars) of the degree of past utilisation
of the five broad habitats utilised by the black and blue wildebeest at Ezemvelo
Nature Reserve in 2004 as calculated from the percentage composition of the five
ecological classes of grass species present within these habitats. No significant
differences between habitats were found. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy grasslands.
89
The PROC GLM tests showed a significant difference in grass species density
between habitat types (p = 0.0011). According to the PROC GLM tests, the rocky
grasslands had a significantly higher grass species density than the Burkea
woodlands (p = 0.0067), the moist grasslands (p = 0.0007) and the old lands (p =
0.0003) (Figure 5.7). The old lands had a grass species density that was significantly
lower than that of the sandy grasslands (p = 0.0255). The moist grasslands also had
a significantly lower grass species density than the sandy grasslands (p = 0.0373).
The percentage bare ground present differed significantly between the habitat types
(p = 0.0004). The rocky grasslands had a significantly higher percentage of bare
ground compared with the Burkea woodlands (p = 0.0003), the moist grasslands (p =
0.0009), the old lands (p = 0.0008) and the sandy grasslands (p = 0.0002) (Figure
5.8).
All the habitat types had a grass canopy cover of >60% during all the ecological
seasons (Figure 5.9). The percentage canopy cover differed significantly between the
habitat types during the late growing season (p = 0.0461). In the late growing season
the moist grasslands had a significantly higher canopy cover than the rocky
grasslands (p = 0.0067), as did the Burkea woodlands (p = 0.0332) and the old lands
(p = 0.0234). In the dormant season, the percentage canopy cover also differed
significantly between habitat types (p = 0.0046). The moist grasslands then had a
significantly higher percentage canopy cover than the old lands (p = 0.0097) and the
rocky grasslands (p = 0.0003), while the rocky grasslands also had a significantly
lower percentage canopy cover than the sandy grasslands (p = 0.0053). In the early
growing season the percentage canopy cover also differed significantly between
habitat types (p = 0.0472). The moist grasslands had a significantly higher
percentage canopy cover than the old lands (p=0.0206) and the rocky grasslands (p
= 0.0261), while the sandy grasslands had a significantly higher percentage canopy
cover than the old lands (p=0.0352) and the rocky grasslands (p = 0.0473).
90
V e ld condition score
600
500
400
300
200
100
0
BW
RG
OL
MG
SG
Figure 5.5: Mean (columns) and standard error (bars) of the veld condition score of
the five broad habitats utilised by the black and blue wildebeest at Ezemvelo Nature
Reserve in 2004 as calculated from the percentage composition of the five ecological
classes of grass species present within these habitats. No significant differences
between habitats were found. BW = Burkea woodlands, RG = rocky grasslands, OL
= old lands, MG = moist grasslands, SG = sandy grasslands.
91
b
2.5
Species diversity index
ab
2
ab
a
a
OL
MG
1.5
1
0.5
0
BW
RG
SG
Figure 5.6: Mean (columns) and standard error (bars) of the Shannon-Wiener
diversity index of the five broad habitats utilised by the black and blue wildebeest at
Ezemvelo Nature Reserve in 2004 as calculated from the percentage composition of
grass species present within these habitats. Means with the same superscripts were
not significantly different from each other. BW = Burkea woodlands, RG = rocky
grasslands, OL = old lands, MG = moist grasslands, SG = sandy grasslands.
92
Species density ( species/m 2)
3
c
2.5
2
bc
ab
a
a
OL
MG
1.5
1
0.5
0
BW
RG
SG
Figure 5.7: Mean (columns) and standard error (bars) of the plant species density of
the five broad habitats utilised by the black and blue wildebeest at Ezemvelo Nature
Reserve in 2004 as calculated from the percentage composition of the five ecological
classes of grass species present within these habitats. Means with the same
superscripts were not significantly different from each other. BW = Burkea
woodlands, RG = rocky grasslands, OL = old lands, MG = moist grasslands, SG =
sandy grasslands.
93
b
Percentage bare ground
25
20
15
10
5
a
a
a
a
0
BW
RG
OL
MG
SG
Figure 5.8: Mean (columns) and standard error (bars) of the percentage bare ground
of the five broad habitats utilised by the black and blue wildebeest at Ezemvelo
Nature Reserve in 2004. Means with the same superscripts were not significantly
different from each other. BW = Burkea woodlands, RG = rocky grasslands, OL = old
lands, MG = moist grasslands, SG = sandy grasslands.
94
Percentage canopy cover
100
90
80
70
60
50
40
30
20
10
0
a
b
a
a
ab
a
c b
ab a
ab
b
b
a
a
BW
RG
Late growing season
OL
Dormant season
MG
SG
Early growing season
Figure 5.9: Mean (columns) and standard error (bars) of the percentage canopy
cover of the herbaceous layer of the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve over the three ecological seasons in 2004.
Means with the same superscripts were not significantly different from each other;
these are compared across the five habitats within each season. BW = Burkea
woodlands, RG = rocky grasslands, OL = old lands, MG = moist grasslands, SG =
sandy grasslands.
95
The lowest total grass height was found in the old lands and the highest in the
Burkea woodlands throughout all seasons (Figure 5.10). The highest grass leaf
height was found in the moist grasslands and the lowest grass leaf height was found
in the old lands (Figure 5.11). There was no significant difference in total grass height
between the habitat types in the late growing season (p = 0.7280). There also was no
significant difference in grass leaf height between the broad habitat types in late
growing season (p = 0.2354). In the dormant season there was no significant
difference between habitat types in total grass height (p = 0.5295). Grass leaf height
between habitat types also did not differ significantly (p = 0.0786) during this season.
The total grass height (p = 0.9563) and grass leaf height (p = 0.9243) did not differ
significantly between the habitat types in the early growing season.
The highest grass biomass was found in the moist grasslands (4 900 kg/ha) and the
lowest in the old lands (2 500 kg/ha) throughout all seasons (Figure 5.12). No
significant difference was found in biomass between habitat types (p=0.1403) in late
growing season. There also was no significant difference in grass biomass between
the habitat types in the dormant season (p = 0.3202) or in the early growing season
(p = 0.3391).
There was no significant difference in the biomass concentration between habitat
types in the late growing season (p=01971), and there was no significant difference
in grass biomass concentration between the habitat types in the dormant season (p =
0.4375), but there was a significant difference in grass biomass concentration
between habitat types in the early growing season (p = 0.0579) (Figure 5.13). During
the early growing season, the moist grasslands had a significantly higher grass
biomass concentration than the old lands (p = 0.0073) and rocky grasslands (p =
0.0459). The sandy grasslands had a significantly higher biomass concentration than
the old lands in the early growing season (p = 0.0333).
Tables 5.5 and 5.6 summarise all the characteristics of the five broad habitat types
that were utilised by the black or blue wildebeest at Ezemvelo Nature Reserve during
the study period.
96
Tota l gra ss he ight (cm )
90
80
70
60
50
40
30
20
10
0
BW
RG
Late growing s eas on
OL
Dorm ant s eas on
MG
SG
E arly growing s eas on
Figure 5.10: Mean (columns) and standard error (bars) of the total grass height of the
herbaceous layer in the five broad habitats utilised by the black and blue wildebeest
at Ezemvelo Nature Reserve over the three ecological seasons in 2004. No
significant differences were found between the means of the different habitat types.
BW = Burkea woodlands, RG = rocky grasslands, OL = old lands, MG = moist
grasslands, SG = sandy grasslands.
97
Grass leaf height (cm)
60
50
40
30
20
10
0
BW
RG
Late growing season
OL
Dormant season
MG
SG
Eraly growing season
Figure 5.11: Mean (columns) and standard error (bars) of the grass leaf height of the
herbaceous layer in the five broad habitats utilised by the black and blue wildebeest
at Ezemvelo Nature Reserve over the three ecological seasons in 2004. No
significant differences between categories were found. BW = Burkea woodlands, RG
= rocky grasslands, OL = old lands, MG = moist grasslands, SG = sandy grasslands.
98
Grass biomass (kg/ha)
6000
5000
4000
3000
2000
1000
0
BW
RG
OL
Late growing season
Dormant season
MG
SG
Early growing season
Figure 5.12: Mean (columns) and standard error (bars) of the grass biomass of the
herbaceous layer in the five broad habitats utilised by the black and blue wildebeest
at Ezemvelo Nature Reserve over the three ecological seasons in 2004. No
significant differences between categories were found. BW = Burkea woodlands, RG
= rocky grasslands, OL = old lands, MG = moist grasslands, SG = sandy grasslands.
99
Biomass concentration
1.8
c
1.6
1.4
bc
ab
ab
1.2
a
1
0.8
0.6
0.4
0.2
0
BW
RG
Late growing season
OL
Dormant season
MG
SG
Early growing season
Figure 5.13: Mean (columns) and standard error (bars) of the biomass concentration
of the herbaceous layer in the five broad habitats utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve over the three ecological seasons in 2004.
Superscripts that are the same indicate no significant differences. BW = Burkea
woodlands, RG = rocky grasslands, OL = old lands, MG = moist grasslands, SG =
sandy grasslands.
100
2540
744
658
2933
Old lands
Moist grasslands
Sandy grasslands
123
Loudetia simplex
shallow soils,
Digitaria eriantha
grasses, flat areas, no
Eragrostis curvula
Setaria sphacelata
Eragrostis
chloromelas
Rock cover <30%,
open high-lying
plains, deep sandy
soils
waterlogged
seasonally
101
% Class 1 and Class 2.
High % Class 4 and 5, low
Class 2 %, lowest Class 1 %
Aristida junciformis
Eragrostis gummiflua
drainage areas,
valleys, fine clay soils,
no rock cover,
High Class 5 %, Moderate
% Class 2
highest Class 1 %, very low
High % Class 4, second
Highest % Class 2
Class 2
Highest % Class 1, low %
Ecological composition
Eragrostis nindensis
Tall fibrous grasses,
rock cover
Eragrostis curvula
Short stoloniferous
exposed areas
Xerophyta retinervis
Burkea africana
Setaria sphacelata
species
Dominant plant
>30% rock cover,
sandy soils
rock cover, deep
Open woodland, no
characteristics
area in
ha
General
Surface
Rocky grasslands
woodlands
Burkea
Habitat
2004
50
54
55
15
55
utilisation
Percentage
363
313
418
414
447
score
condition
Veld
1.9
1.5
1.5
2.2
1.8
diversity
Grass species
2
1.4
1.3
2.4
1.5
density
species
Grass
Table 5.5: Summary of the characteristics of the five broad habitat types utilised by black and blue wildebeest at Ezemvelo Nature Reserve in
grasslands
Sandy
grasslands
Moist
Lands
Old
grasslands
Rocky
woodlands
Burkea
Habitat
89
95
82
80
71
73
64
69
77
(cm)
(cm)
83
height
height
(%)
39
43
31
38
41
leaf
grass
cover
Grass
Total
Canopy
3857
4562
2568
3224
4462
(kg/ha)
Biomass
Late growing season
1.0
1.1
0.7
0.8
1.1
BC
80
90
72
60
95
(%)
cover
Canopy
56
60
45
45
59
(cm)
height
grass
Total
102
27
41
25
20
29
(cm)
height
leaf
Grass
3965
4343
3123
2359
3733
(kg/ha)
biomass
Grass
Dormant season
Ecological season
Reserve over the three ecological seasons in 2004. BC = biomass concentration
1.5
1.0
1.2
1.2
1.3
BC
89
91
75
74
96
(%)
cover
Canopy
61
52
46
55
44
(cm)
height
grass
Total
36
34
29
31
28
(cm)
height
leaf
Grass
4386
4657
2870
3330
2858
(kg/ha)
biomass
Grass
Early growing season
1.3
1.4
0.9
1.1
1.0
BC
Table 5.6: Summary of the herbaceous characteristics of the five habitat types utilised by the black and blue wildebeest at Ezemvelo Nature
DISCUSSION
The results of the present study indicated that ecological separation between the
black and blue wildebeest does exist to some extent on the basis of broad habitat
type selection. Observations of the two types of wildebeest indicated a distinct habitat
divergence with the black wildebeest selecting the open and moist bunch grass
habitats and blue wildebeest the habitats that provided some form of cover and with
extensive grazing sites. This study therefore further supported previous observations
that blue wildebeest prefer areas where cover is readily available (Skinner and
Chimimba 2005). Previous research also supports the results of this study that black
wildebeest prefer open grasslands with little tree cover (Von Richter 1971b; Schmidt
1988). A preference for open grasslands has been attributed to predation risk, as
black wildebeest rely more on speed than on camouflage to escape predators
(Chapter 3). Moreover, open areas also do not allow for concealment of approaching
predators (Schmidt 1988). However, other studies have indicated that territorial
defence may play a larger role in the preferences of black wildebeest for open areas
(Brink et al. 1999). This theory may be more appropriate in a reserve where
predation does not play an important role such as at Ezemvelo Nature Reserve.
According to Von Richter (1971a) the black wildebeest prefers short grasslands and
avoids areas where tall mature grasses predominate. The same is true for the blue
wildebeest (Child 1968; Estes 1969). Both types of wildebeest also tend to condition
their grazing sites in such a way that they keep the grass short (Von Richter 1971a).
Past land use practices at Ezemvelo Nature Reserve, such as planted pastures and
cattle kraals as was described in Chapter 2, have left a large proportion of the
reserve as optimal habitat for wildebeest with large areas of short grass. Before
being declared a nature reserve, a large portion of Ezemvelo consisted of ploughed
lands. These lands were sown with Eragrostis curvula and Digitaria eriantha to
provide a cultivated pasture (Tau 2004 pers. comm.)9. The above two grass species
are still the most abundant plant species in the old lands, with a mean percentage
frequency of 25% and 47% respectively.
It would therefore be expected that both types of wildebeest would preferentially
select the old lands. The results of the present study indicated that the blue
wildebeest showed a strong preference for the old lands, whereas the black
9
Mr. M. Tau. Manager, Ezemvelo Nature Reserve, P.O. Box 599, Bronkhorstspruit, 1020,
South Africa. [email protected]
103
wildebeest only showed a preference for this habitat during the late growing season
and used it in a random manner throughout the rest of the year. The black wildebeest
instead tended to select the sandy grasslands that consisted mainly of tall mature
grass stands. Possible conflict between black and blue wildebeest may occur in the
late growing season when they both are selecting the old lands as a preferred
habitat.
With their habit of remaining in one area for extended periods of time (Von Richter
1971a), the black wildebeest tends to trample and overgraze certain areas. The
dominant grasses in the sandy grasslands where the black wildebeest predominated
at Ezemvelo Nature Reserve are not stoloniferous but are stemmy grasses replaced
through grazing pressure by grasses with a low palatability, such as Aristida
congesta (Van Oudtshoorn 1999). The basal cover therefore decreases over time,
after which the black wildebeest will move off to another grazing area where the
grasses have not yet been degraded. Selection of the grazing site tended to depend
on the openness of the habitat rather than the quality of the footage available. This
high affinity for open areas at the expense of forage and feeding site suitability,
suggested that forage was not the only constraint on black wildebeest habitat use.
This selection could have been as a result of a trade-off between selecting for open,
high-lying areas for territorial defence and areas that provide the most suitable
feeding grounds (Namgail et al. 2004). The differences in old land habitat use that
were found between the black and blue wildebeest may therefore well be attributable
to species-specific differences in territorial behaviour as suggested by Brink et al.
(1999) in his study of fossil evidence.
It appears that both types of wildebeest avoided the rocky grassland habitats
throughout all the seasons. This may be due to large portions of the rocky grasslands
being in areas far from the nearest water. Therefore only those sections of this
habitat type that have water in its vicinity would be utilised. The rocky grasslands had
the lowest degree of utilisation by wildebeest (15%), but also the highest proportion
of Ecological Class 2 grass species, most of which were highly unpalatable (Van
Oudtshoorn 1999). Avoidance of this habitat by both types of wildebeest may
therefore be more related to its grass species composition than the physical
rockiness of the terrain.
The moist grasslands in the drainage areas that were selected by the black
wildebeest and avoided by the blue wildebeest, tended to have the lowest veld
104
condition score, the lowest grass species diversity, the highest percentage canopy
and basal cover, and the tallest stands of grasses of all the habitat types studied
(Table 5.6). Black wildebeest tended to concentrate in the higher-lying parts of this
habitat type. They also tended to show the most preference for this habitat type
during the early growing season when it tended to be dominated by nutritious forbs
and sedges (Cyperaceae), which appeared in highest abundance after the first rains
in the moist grasslands. Observations of black wildebeest utilising these ephemeral
plant species have been recorded by Furstenberg (2002a) and data from field studies
have shown a higher browse intake by black wildebeest as compared to blue
wildebeest (Codron and Brink In press). Where this habitat occurs in higher-lying
areas, it may be used by the black wildebeest as a substitute for the old lands to
meet their nutritional requirements. This habitat type would have a higher percentage
of C3 monocotyledons such as reeds (Phragmites spp.) and sedges (Cyperaceae)
which have been shown to contribute to the diets of wildebeest and of making up 40
to 50% of the plant species in the South African central interior (Stock et al. 2004).
In total, 62% of the black wildebeest and 40% of the blue wildebeest observations
were in the sandy grasslands. This habitat type is extremely important in the study
area due to its large surface area. Spatial heterogeneity within this habitat type is
higher than the other habitat types investigated due to patches of bunch grass being
intermingled with patches of lawn grasses. Spatial heterogeneity within this habitat
could allow for use of this habitat type by both the black and blue wildebeest without
competition.
Habitat segregation between the three social groups in both the types of wildebeest
was not clearly indicated. The results did, however, indicate that the bachelor herds
of both types of wildebeest randomly selected most habitat types during all the
seasons, indicating that the bachelor herds may be forced to occupy those habitats
that were not utilised by the other social groups. This tended to support the
observations of Von Richter (1971a). The sample size for bachelor herd sightings
may, however, have been too small to demonstrate clear habitat selection, but it is
feasible that the bachelor herds would tend to roam widely through the study area
without selecting specific areas (Penzhorn 1982). The female herds of both types of
wildebeest on the other hand were more sedentary than the bachelor herds.
In order to ensure some measure of separation of the two types of wildebeest on
Ezemvelo Nature Reserve, the establishment of Burkea woodlands should be
105
encouraged and the loss of these woodlands due to fire and insect infestation must
be prevented. The Burkea woodlands provide cover for the blue wildebeest, while the
black wildebeest completely avoid this habitat. This habitat type may therefore
provide one of the main means of providing mutually exclusive habitat for the two
types of wildebeest and thus its preservation should be promoted. Other wildlife such
as red hartebeest and Burchell’s zebra also utilised this habitat to a large degree.
The dynamics of why these even age stands of Burkea woodland (cohorts) occur in
certain areas and not in others are not yet clearly understood (Wilson and Witkowski
2003). Research into these cohorts may provide details on establishment of Burkea
africana and thus ensure that these stands are actively managed for the continued
coexistence of the two types of wildebeest in the study area.
The greater the impact of a particular vegetation pattern on the foraging behaviour of
a particular herbivore species, the greater may be the divergence of impact of that
species on the areas which have different distributions of the same vegetation types
(Bailey et al. 1998). The grasslands that were dominated by stoloniferous grazing
lawn grass species were characterised by a high overall degree of utilisation by both
the black and blue wildebeest, indicating that this habitat may generally offer a better
food quality but possibly a lower food availability (standing grass biomass) than the
bunch grass communities on the sandy grasslands, rocky grasslands and moist
grasslands (Cromsigt 2006). Several studies have shown that grass production in
grazing lawns is higher than in other grassland types (Hik and Jeffries 1990).
CONCLUSION
Pianka (1978) indicated that ecologists have long considered that habitat separation
can serve to decrease both interference and exploitative competition and could
facilitate coexistence of ecologically similar species. Selection by black and blue
wildebeest of the various habitat types on Ezemvelo Nature Reserve was found to be
mainly due to the physical characteristics of the habitats and a certain degree of
ecological separation was evident through this habitat dimension. Therefore habitat
separation in terms of degree of openness and elevation may reduce competition
and facilitate the coexistence of the black and blue wildebeest at Ezemvelo Nature
Reserve as was also shown in New Zealand for Tahr and Camois by Namgail et al.
(2004).
106
CHAPTER 6:HABITAT SELECTION AND SEPARATION: MESOHABITAT SCALE
INTRODUCTION
Habitat selection does not only occur at the broad habitat scale as was analysed in
Chapter 5. The sites that are utilised by an animal may occur within a particular
habitat, but the features of those sites of utilisation may be the actual cause of
selection for that site rather than the fact that the site occurs within that particular
habitat. This requires analysis of habitat selection and separation at the meso-habitat
scale. In determining whether habitat separation occurs between two types of wildlife
there is a wide range of possible meso-habitat factors to be considered that could be
responsible for its existence. Habitats are defined by certain geomorphological
factors such as topography, geological formations and soil types, as well as weather
and vegetation (Theron 1991). It is imperative to assess as many meso-habitat
factors as possible to determine which factors are most important in differentiating
between areas selected by one type of wildlife and perhaps not by the other.
For African ungulates, the main determinants of local movements are forage
availability, forage quality in terms of mineral nutrition, water availability (Ben-Shahar
and Coe 1992) and certain landscape types and features such as topography, soil
type and vegetation composition and structure (Ben-Shahar 1995). Seasonal
movements of animals may be attributed to climatic conditions, the seasonal
phenological development of forage and the occurrence of fire (Munthali and Banda
1992). For water dependent wildlife like the black and blue wildebeest the availability
of water would be most important in habitat preference, but when water is abundant
the physical structure of the habitat would become more important. Physical aspects
such as topography, slope, geomorphology and rock cover do not change over short
periods of time and thus can be regarded as relatively constant. Vegetation structure
would then be expected to determine the suitability of a habitat for such factors as
available shade and visibility. Within this context the plant species composition would
play an important role in determining whether the food source is sufficient for the
requirements of the species (Strauss 2003).
Since it was expected that the black and blue wildebeest are too ecologically similar
to be kept in the same area, it was hypothesised that there would be no meso-habitat
separation between the black and blue wildebeest at Ezemvelo Nature Reserve.
The objectives of this part of the study were therefore to:
107
•
Determine which habitat factors (if any) separate the habitats of the two
types of wildebeest and whether the bachelor herds, territorial bulls and
female herds of the black and blue wildebeest have different separating
mechanisms.
•
Determine if the black and blue wildebeest graze in habitats with different
meso-habitat characteristics.
•
Determine if habitat separation (if found) of the black and blue wildebeest
is affected by any seasonal influences, time of day influences, daily
temperature fluctuations or daily cloud cover fluctuations
METHODS
All the data for this section of the study were obtained by using the methods
described in Chapter 4.
Statistical analysis of the data
In the past, simple qualitative descriptions of the data were used to determine the
habitat preferences of a species (Lamprey 1963). Later, quantitative techniques were
introduced. In their simplest form, quantitative studies on habitat selection expressed
habitat utilisation in terms of the proportion of animals seen in each sub-habitat or
habitat (Scogings et al. 1990). A comparison of the observed habitat use with the
expected habitat use, according to habitat availability, is an extension of these simple
quantitative techniques (Hirst 1975).
Multivariate analyses have more recently been used to quantify the relationship
between herbivores and their habitat. Studies using multivariate analysis techniques
do not require information on the amount of sub-habitat available, as a record of
habitat variables at each animal location is sufficient (Strauss 2003). Some of the
drawbacks of the traditional multivariate analysis methods, however, include the
assumption of normally distributed data, which is seldom justified in ecological data,
and the assumption of linear relationships between variables, which is often violated
as the relationships between variables is usually more complex in ecological data
(Beardell et al. 1984).
Among those techniques that are used in multivariate analyses are discriminant
function analysis (Ferrar and Walker 1974), multiple regression (Hirst 1975; BenShahar 1986), correspondence analysis (Beardell et al. 1984; Engelbrecht 1986) and
108
detrended correspondence analysis (Scogings et al. 1990). Even more recently the
categorical modelling (CATMOD) of data has been used to determine habitat
selection where categorical variables can easily be investigated (Weaver 1995; Von
Holdt 1999). This procedure, however, has proven extremely time-consuming for
previous researchers as multiple runs and variable recategorisations have to be
performed in order to obtain a model with the most significant variables (Van der
Linde 2006 pers. comm.)10. Logistic regression analysis has also been used by
researchers in habitat studies (Morrison et al. 1992; Pauley et al. 1993).
Since the aim of the present study was to determine the habitat separation between
the black and blue wildebeest, a simple analysis of which habitat factors were related
to the habitat use of each type of wildebeest, would not be sufficient to separate their
habitat choices. Thus, it was decided to use a multivariate approach to determine
predictor variables for determining the occurrence of a black wildebeest or a blue
wildebeest.
Both data management and statistical analyses were performed by using SAS
version 8.01 (SAS Institute, Inc., Cary, NC, USA). In the current study, each
observation consisted of 43 variables. The data associated with each of these
variables were investigated in groups of categories. The variable coding and
explanations can be found in Appendix 2.
The multivariate approach that was used here was that of linear logistic regression.
Logistic regression analysis is a combination of multiple discriminant function
analysis and multiple regression analysis (Hair et al. 1995). It is a robust alternative
to simple discriminant analysis (Dattalo 1994; Lottes et al. 1996). Logistic regression
analysis was performed with the PROC LOGISTIC procedure to determine predictor
variables (at the 5% level) that would separate the habitats that were utilised by the
black wildebeest from those that were utilised by the blue wildebeest. The PROC
LOGISTIC procedure was used to investigate the relationship between discrete or
binary responses and a set of explanatory variables by fitting linear logistic
regression models through the method of maximum likelihoods. The response
variable is the dependent variable. This procedure can handle both continuous or
categorical explanatory variables and can analyse large data sets. The aim of
maximum likelihood estimation is to find the parameter value(s) that makes the
10
Dr M. van der Linde. Department of Information Technology, University of Pretoria,
Pretoria, 0002, South Africa.
109
observed data most likely. This is because the likelihood of the parameters, given the
data, is defined to be equal to the probability of the data, given the parameters.
The basic logistic model as described by Ely et al. (1996) was applied:
Logit (p) = log (p/1-p) = .¶[
Where p is the probability of finding a black wildebeest
.
¶
LQWHUFHSWSDUDPHWHU
URZYHFWRURIVORSHSDUDPHWHU
x = column vector of explanatory variables
The parameters of the standard logistic model can be interpreted directly or
indirectly, after transformation to odds ratios, to a probability or to a difference in
probability (McArdle and Hamagami 1996; Groeneveld 2006 pers. comm.11). The
power value or log odds (Hall and Round 1994) for each observation was calculated
based on the mean value of each habitat variable. Probabilities (p) were calculated
from the antilog of the power value or odds. Probabilities were used to identify the
habitat variables which best described the variation between the black and blue
wildebeest habitat types. If the predicted probability was >0.5, then the prediction
was taken to relate to a black wildebeest, otherwise it related to a blue wildebeest
(Hair et al. 1995).
All the variables were categorised before input into the PROC LOGISTIC model
(Appendix 2). Before input, the variables were examined for missing values,
correlation and singularity. A number of variables were not included in the analysis
due to missing values. These were drainage (V42), dominant plant species (V37) and
sub-dominant plant species (V38). The association variable (V32) was not included in
this analysis but it was used for further analysis in the behavioural interaction section
of this study. The vegetation structure (V33) was found to be correlated to a number
of other variables and was omitted in this analysis.
The data set submitted to the PROC LOGISTIC procedure included the following
explanatory variables: habitat type (V15), aspect (V16), slope analysis (V17),
distance to water (V19), woody vegetation cover (V20), grass cover (V21), rock cover
(V26), total grass height (V27), grass leaf height (V28), plant utilisation (V29),
11
Prof. H. Groeneveld. Department of Statistics, University of Pretoria, Pretoria 0002, South
Africa.
110
visibility (V30), distance to shade (V31), erosion (V34), altitude (V35), time since last
burn (V36), exposure (V39), geomorphology (V40), and forb : grass ratio (V41). The
herd size (V13) was used as a weighting factor.
A stepwise selection procedure was then followed to select variables that would be
significant in differentiating between habitats that were utilised preferentially by black
and blue wildebeest. The probability modelled was for the presence of a black
wildebeest. The model was tested for significance by the –2 log-likelihood statistic,
explanatory power by the maximum rescaled r2 value and its capacity to successfully
discriminate black or blue wildebeest habitat use by the c value (While and McArthur
2005). The Wald statistic was used to test the significance of effects of each
independent variable. Once these variables were selected they were rerun through a
PROC LOGISTIC analysis process with no selection and the probabilities associated
with each category of each selected variable were calculated. When the significant
variables are examined it is possible to identify which categories are preferred by
which type of wildebeest when all the other factors are held constant. This was done
by examining p-values from 0 to 0.2 indicating preference by blue wildebeest and pvalues from 0.8 to 1 indicating preference by black wildebeest. The percentage of the
observations with p-values from 0 to 0.2 and from 0.8 to 1 were tabulated to indicate
the preferred categories within each variable for each model by each type of
wildebeest. Preference for a particular category within a variable by a type of
wildebeest was shown by the presence of a high percentage of probabilities
indicating the likelihood of finding a certain type of wildebeest within that category.
Therefore if 80% of the 0.8-1.0 probabilities (indicating a high probability of finding a
black wildebeest) were within the south-facing category of the aspect variable, then
this indicates that the black wildebeest preferred (or were most likely) to be found in
this category of the aspect variable.
This process was repeated to obtain predictor variables for the three ecological
seasons. The same variables were thereafter submitted to PROC LOGISTIC analysis
by the type of activity (V14), time of the day (V3), cloud cover (V22), temperature
(V23), and finally by social group (V43).
111
RESULTS
A total of 1 558 wildebeest observations were made. Of these, 24% (371
observations) were of black wildebeest and 76% (1187 observations) were of blue
wildebeest. The number of observations reflects the number of herds and individuals
of each type of wildebeest that was present on the reserve. There were
approximately 256 blue wildebeest in seven main herds, and approximately 98 black
wildebeest in five main herds at Ezemvelo Nature Reserve. There were many more
single blue wildebeest bulls than there were single black wildebeest bulls, and
therefore the blue wildebeest data set has also been increased due to this effect. The
observations were evenly spread through the three seasons.
Bachelors comprised 10% of the data, female herds made up 32% of the data and
territorial bulls made up 58% of the data. The percentage sightings per season and
per social group per type of wildebeest are shown in Table 6.1.
The logistic regression procedure (PROC LOGISTIC) was able to provide models for
predictor variables that could separate out the black and blue wildebeest habitat
choices at most of the levels analysed (Table 6.2). All the models ran provided good
predictive capacity for separating out the black and blue wildebeest habitat choices.
Entire study period analysis
Model 1
The logistic regression correctly classified 93.4% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.936).
Twelve predictor variables were identified to be significant in discriminating between
the sites occupied by the two types of wildebeest during the entire study period. The
important habitat separating variables were: habitat type (V15), aspect (V16), slope
(V17), distance to water (V19), woody vegetation cover (V20), total grass height
(V27), plant utilisation (V29), distance to shade (V31), altitude (V35), time since last
burn (V36), geomorphology (V40) and forb : grass ratio (V41) (-2 log likelihood = 1
339.99; df; = 33; p<0.0001) (Table 6.3).
112
Table 6.1. The percentage observations of black and blue wildebeest social groups
over the three ecological seasons at Ezemvelo Nature Reserve obtained by using the
methods described in Chapter 4 from January 2004 to August 2005
Type of
Social group
wildebeest
Black wildebeest
Blue wildebeest
Late growing
Dormant season
season
Early growing
Overall
season
Bachelor herds
11.02
19.69
17.95
16.17
Territorial bulls
39.37
25.20
35.90
33.42
Female herds
49.61
55.12
46.15
50.40
Bachelor herds
6.33
9.00
8.40
7.92
Territorial bulls
69.22
64.72
64.04
66.13
Female herds
24.05
26.28
27.56
25.95
113
Table 6.2: Predictor variables for the various combinations of season, social
structure, activity, time of day and weather conditions used in the PROC LOGISTIC
procedure (SAS 8.01) to determine those variables that separate the habitats used
by the black and blue wildebeest on Ezemvelo Nature Reserve. This analysis was
based on 1 558 wildebeest observations that were collected from January 2004 to
August 2005
Activity
Not specified
Time of the day
Not specified
Season
Not specified
Model
1
Not specified
Not specified
Not specified
Grazing
Not specified
Not specified
Not specified
Not specified
Late growing
Dormant
Early growing
Not specified
2
3
4
5
Grazing
Grazing
Grazing
Not specified
Not specified
Not specified
Not specified
Before 10:00
Late growing
Dormant
Early growing
Not specified
6
7
8
9
Not specified
Not specified
10:00-14:00
After 14:00
Not specified
Not specified
10
11
Predictor variables
V15, V16, V17, V19, V20, V27, V29,
V31, V35, V36, V40, V41.
V26, V31
V17, V26, V28, V29, V31, V35, V36
V15, V16, V31, V35, V36
V15, V16, V17, V21, V27, V31, V35,
V36
V16, V26, V31
V17, V21, V26, V28, V31, V35, V36
V15, V16, V31, V35, V36
V15, V16, V17, V20, V31, V35, V36,
V40
V15, V26, V31, V36
V15, V16, V31
Temperature
<15°C
15 – 25°C
Cloud cover
Not specified
Not specified
Season
Not specified
Not specified
Social group
Not specified
Not specified
Model
12
13
>25°C
Not specified
Not specified
Not specified
14
Not specified
0%
Not specified
Not specified
15
Not specified
>0-50%
Not specified
Not specified
16
Not specified
Not specified
>50%
Not specified
Not specified
Not specified
Not specified
Female herds
17
18
Not specified
Not specified
Not specified
Territorial bulls
19
Not specified
Not specified
Not specified
Bachelor herds
20
Predictor variables
None
V15, V16, V17, V18, V20, V21,
V27, V31, V35, V36
V15, V16, V27, V31, V35, V36,
V40, V41
V16, V17, V26, V27, V29, V30,
V31, V35, V36, V40
V15, V16, V18, V21, V27, V31,
V35, V36, V40, V41
V15, V31, V35, V36
V15, V16, V17, V19, V26, V31,
V36, V40
V16, V20, V31, V34, V35, V36,
V40, V41
V31, V41
c-value
0.94
0.83
0.95
0.91
0.91
0.75
0.95
0.92
0.93
0.92
0.84
c-value
0.91
0.94
0.93
0.93
0.92
0.98
0.91
0.83
Note: The c-value indicates the discriminatory power of the model. Appendix 2
provides an explanation of the variable codes listed in the above table
114
Table 6.3: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 1 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Slope
Distance to water (m)
Woody vegetation cover
Total grass height (mm)
Plant utilisation
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Geomorphology
Forb : grass ratio
Category
Black wildebeest
Blue wildebeest
0.00
14.88
13.64
6.20
65.29
32.64
67.36
19.83
26.45
53.72
12.40
21.49
33.88
32.23
100.00
0.00
0.00
5.37
45.45
42.98
6.20
2.89
28.10
21.07
47.93
0.00
7.02
18.60
74.38
18.60
34.30
17.77
29.34
98.76
0.00
0.00
1.24
0.00
45.04
30.17
24.79
23.55
49.17
25.21
2.07
10.13
3.83
28.61
24.98
32.45
72.96
27.04
2.75
46.61
50.64
12.68
37.27
20.16
29.89
53.00
33.04
13.96
16.03
35.99
25.37
22.62
8.55
23.80
20.55
47.10
16.91
60.37
21.24
1.47
37.36
23.80
18.68
20.16
66.37
1.77
2.36
28.71
0.79
19.96
77.38
2.65
17.99
38.35
39.04
4.62
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Flat
Gentle
Moderate
0-50
>50-300
>300-500
>500
None
Sparse
Open
0-50
>50-500
>500-800
>800
Low
Moderate
High
Excessive
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
0:100
10:90
30:70
50:50
115
The variables that contributed the most to the separation of the habitat of the two
types of wildebeest were distance to shade (Wald $2 = 156.4764; df = 3; p<0.0001),
habitat type (Wald $2 = 68.5961; df = 4; p<0.0001) and time since last burn (Wald $2
= 65.0; df = 4; p<0.0001). Woody vegetation cover contributed the least to the
analysis (Wald $2 = 9.9045; df = 2; p = 0.0071).
The results indicated that black wildebeest were more likely to be found in sandy
grasslands than any other habitat type while blue wildebeest were equally likely to be
found in old lands and sandy grasslands. Black wildebeest were more likely to be
found in south facing aspects while blue wildebeest were more likely to be found on
north-facing aspects. The blue wildebeest was most likely to be found on gentle or
moderate slopes with a northerly aspect. Black wildebeest, however, were more
likely to occur on moderate slopes with a southerly aspect. Blue wildebeest favoured
distances of >50 to 300 m from the nearest water, whereas black wildebeest
preferred distances >300 m from the nearest water. Blue wildebeest showed a
preference for total grass height ranging from >50 to 500 mm, while black wildebeest
showed a clear preference for sites where the total grass height was >50 to 800 mm.
Both types of wildebeest were more likely to occur on sites where the herbaceous
layer had been heavily utilised. The blue wildebeest strongly preferred habitats where
the distance to the nearest shade was much less than what the black wildebeest
preferred (>5 to 100 m as opposed to >600 m). The blue wildebeest also tended to
be more likely to occur at lower altitudes than the black wildebeest (” P DV
opposed to 1341 to 1360 m). The blue wildebeest preferred recently burnt areas,
whereas the black wildebeest were more likely to occur at sites that had not been
burnt in a while.
The blue wildebeest most frequently utilised sites with a convex geomorphology,
whereas the black wildebeest tended to utilise all types of geomorphology equally,
although it tended to favour a concave geomorphology to a certain degree. The blue
wildebeest was more likely to utilise habitats where the forb : grass ratio was higher
than that used by the black wildebeest.
Seasonal analyses
Model 2: Late growing season
The logistic regression correctly classified 78.0 % of the habitat samples according to
type of wildebeest. The discriminating power of this model was good (c = 0.832).
116
Only two predictor variables were identified to be significant in discriminating
between the sites occupied by the two types of wildebeest during the late growing
season. The important habitat separating variables were: rock cover (V26) and
distance to shade (V31) (-2 log likelihood = 305.8227; df = 5; p<0.0001) (Table 6.4).
Distance to shade contributed the most to the analysis (Wald $2 = 105.9141; df = 3;
p<0.0001).
The late growing season is when resources are abundant and animals are in a peak
physical condition. Therefore few mechanisms separating the habitat of the black and
blue wildebeest would be expected. The only separation found was that the blue
wildebeest were most likely to occur in habitats with no rock cover while black
wildebeest were more likely to utilise areas where the rock cover was <30%. Black
wildebeest showed a strong preference for areas that were >600 m away from the
nearest shade, while blue wildebeest preferred areas where the shade was ”100 m
away.
Model 3: Dormant season
The logistic regression correctly classified 94.6% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.947).
Seven predictor variables were identified to be most significant in discriminating
between the sites occupied by the two types of wildebeest during the dormant
season. The important habitat separating variables were: slope (V17), rock cover
(V26), grass leaf height (V28), plant utilisation (V29), distance to shade (V31),
altitude (V35) and time since last burn (V36) (-2 log likelihood = 484.74367; df = 23;
p<0.0001) (Table 6.5).
Distance to shade contributed the most to separating the habitat choices of the two
types of wildebeest during the dormant season (Wald $2 = 83.6146; df = 3;
p<0.0001). Plant utilisation contributed the least to the analysis (Wald $2 = 14.2395;
df = 3; p = 0.0026).
117
Table 6.4: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D blue (0”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 2 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Rock cover (%)
Distance to shade (m)
Category
Black wildebeest
Blue wildebeest
None
1-30
>30
0-5
>5-100
>100-600
>600
62.16
37.84
0.00
0.00
0.00
0.00
100.00
73.93
3.72
22.35
18.62
59.03
22.35
0.00
118
Table 6.5: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 3 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Slope
Rock cover (%)
Grass leaf height (mm)
Plant utilisation
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
Flat
Gentle
Moderate
None
1-30
>30
0-50
>50-100
>100-400
>400
Low
Moderate
High
Excessive
0-5
>5-100
>100-600
>600
13.68
18.95
67.37
27.37
69.47
3.16
22.11
17.89
47.37
12.63
12.63
2.11
22.11
63.16
0.00
13.68
26.32
60.00
12.63
35.79
23.16
28.42
95.79
0.00
0.00
2.11
2.11
2.13
45.74
52.13
53.99
21.54
24.47
35.90
9.04
30.59
24.47
25.53
5.85
31.65
36.97
12.77
61.17
24.73
1.33
34.84
23.14
17.82
24.2
70.48
2.39
1.33
23.67
2.13
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
119
The dormant season is when resources become limiting and it was therefore
expected that there would be more factors that would participate in separating the
habitats utilised by black and blue wildebeest to minimise competition. Blue
wildebeest preferred gentle and moderate slopes with equal frequency, but black
wildebeest were more likely to select moderate slopes than any other slope category.
Blue wildebeest were most likely to utilise areas with no rock cover while black
wildebeest most frequently selected areas with a rock cover of 1 to 30%. Blue
wildebeest were most likely to utilise areas with grass leaf heights of <50 mm or
those >100 to 400 mm. Black wildebeest, however, were most likely to select sites
where the grass leaf height was >100 to 400 mm. Black wildebeest also were more
likely to select heavily utilised areas while blue wildebeest were more likely to select
areas where the use pressure ranged from moderate to excessive. Blue wildebeest
preferred habitats where the nearest shade was >5 to 100 m away, whereas black
wildebeest favoured areas where the nearest shade was >600 m away, as they did in
the late growing season.
Blue wildebeest showed no clear preference for any altitudinal range, whereas black
wildebeest were most likely to utilise habitats at altitudes >1340 m above sea level.
Blue wildebeest were ten times more likely to utilise recently burnt areas than black
wildebeest.
Model 4: Early growing season
The logistic regression correctly classified 90.2% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.911).
Five predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest during the early growing season. The
important habitat separating variables were: habitat type (V15), aspect (V16),
distance to shade (V31), altitude (V35) and time since last burn (V36). (-2 log
likelihood = 426.0372; df = 14; p<0.0001) (Table 6.6).
Distance to shade contributed the most to separating the habitats of the black and
blue wildebeest (Wald $2 = 77.9892; df = 3; p<0.0001) followed by the time since last
burn variable (Wald $2 = 46.5893; df = 4; p<0.0001). Aspect contributed the least to
the analysis (Wald $2 = 28.6237; df = 1; p<0.0001).
120
Table 6.6: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 4 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Habitat type
Aspect
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
23.75
7.50
8.75
60.00
36.25
63.75
0.00
0.00
10.00
90.00
25.00
33.75
21.25
20.00
97.50
0.00
0.00
2.5
0.00
13.31
8.12
26.30
14.61
37.66
80.52
19.48
21.10
54.87
21.43
2.60
44.81
24.35
14.94
15.91
59.09
0.97
0.97
38.96
0.00
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
121
In the early growing season there was a separation in aspect choice between the
black and blue wildebeest, with the blue wildebeest being more likely to select
northerly slopes while the black wildebeest was more likely to select southerly
slopes. Blue wildebeest were also more likely to occur in areas with shade within 100
m while black wildebeest were most likely to utilise areas with shade at >600 m
away. Blue wildebeest tended to favour low-lying areas while black wildebeest were
more likely to occur at higher altitudes. Blue wildebeest were more likely to utilise
recently burnt areas, whereas black wildebeest tended to prefer areas that had not
been burnt recently.
Analyses of data where wildebeest were grazing
Model 5: Overall
When only those observations where the wildebeest were grazing were taken into
account, the logistic regression was able to discriminate between black and blue
habitat variables (c = 0.914) and correctly classified 91.2% of the habitat samples
according to type of wildebeest. Eight predictor variables were identified which were
significant in discriminating between the sites occupied by the two types of
wildebeest while they were grazing. The important habitat separating variables were:
habitat type (V15), aspect (V16), slope (V17), grass cover (V21), total grass height
(V27), distance to shade (V31), altitude (V35) and time since last burn (V36) (-2 log
likelihood = 595.9007; df = 22; p<0.0001) (Table 6.7). The variables that contributed
most to the separation in the habitat choices of the two types of wildebeest were
distance to shade (Wald $2 = 122.2178; df = 3; p<0.0001) and time since last burn
(Wald $2 = 51.5282; df = 4; p<0.0001). Slope contributed the least to the analysis
(Wald $2 = 9.7449; df = 2; p = 0.0077).
When grazing the black wildebeest were most likely to occur in sandy grasslands and
moist grasslands, while the blue wildebeest preferred old lands and rocky
grasslands. Grazing activities for the blue wildebeest were most likely to occur on
northerly slopes, whereas the black wildebeest showed an equal preference for both
northerly and southerly slopes.
122
Table 6.7: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 5 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Habitat type
Aspect
Slope
Grass cover
Total grass height (mm)
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
22.12
8.85
7.96
61.06
43.36
56.64
15.04
33.63
51.33
13.27
50.44
36.28
7.08
40.71
47.79
4.42
0.00
0.00
13.27
86.73
17.70
45.13
21.24
15.93
93.81
0.00
0.00
5.31
0.88
2.94
5.28
27.59
29.75
34.44
72.41
27.59
2.35
43.84
53.82
19.37
46.77
33.86
14.29
34.05
27.01
24.66
4.31
72.21
22.50
0.98
35.03
25.24
17.42
22.31
66.93
1.57
2.74
27.79
0.89
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Flat
Gentle
Moderate
Sparse
Medium
Dense
0-50
>50-500
>500-800
>800
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
123
Both black and blue wildebeest were most likely to occur on moderate slopes. The
blue wildebeest was least likely to occur on flat areas. Black and blue wildebeest
preferred areas where the grass cover was medium, but the blue wildebeest were
more likely than the black wildebeest, to occur where the grass cover was sparse.
Blue wildebeest preferred areas where the grass sward was >500 to 800 mm tall and
avoided grass swards of <50 mm and >800 mm tall. Blue wildebeest were more likely
to occur across the full range of grass sward heights, but they were most likely to
graze in grass sward heights of >50 to 500 mm tall. Black wildebeest showed a
strong preference for areas where the nearest shade was >600 m away and avoided
grazing in areas where they were <100 m away from shade.
Blue wildebeest were most likely to graze in habitats where the nearest shade was
>5 to 100 m away. Black wildebeest were most likely to graze at altitudes of >1341 to
1360 m, while blue wildebeest tended to prefer to graze at lower altitudes, thus being
more likely to graze at altitudes of <1340 m above sea level.
Model 6: Early growing season
The logistic regression correctly classified 72.8% of the habitat samples according to
type of wildebeest. The discriminating power of this model was medium (c = 0.761).
Three predictor variables were identified to be significant in discriminating between
the sites occupied by the two types of wildebeest while they were grazing during the
early growing season. The important habitat separating variables were: aspect (V16),
rock cover (V26) and distance to shade (V31) (-2 log likelihood = 135.4589; df = 6;
p<0.0001) (Table 6.8). Distance to shade contributed most to this analysis (Wald $2 =
49.9599; df = 3; p<0.0001) while rock cover contributed the least (Wald $2 = 10.8034;
df = 2; p = 0.0045).
When grazing, blue wildebeest preferred northerly slopes while black wildebeest
showed no preference in their grazing patterns in respect of aspect chosen during
the early growing season. Blue wildebeest were most likely to graze in areas where
there was no rock cover, whereas black wildebeest preferred to graze in areas where
the rock cover was 1 to 30%. Blue wildebeest were most likely to graze where the
distance to shade was >5 to 100 m, while black wildebeest were most likely to graze
in areas where the nearest shade was >600 m away.
124
Table 6.8: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 6 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Aspect
Rock cover (%)
Distance to shade (m)
Category
Black wildebeest
Blue wildebeest
45.45
54.55
54.55
42.42
3.03
0.00
0.00
15.15
84.85
69.46
30.54
65.87
16.77
17.37
5.39
77.25
17.37
0.00
North-facing
South-facing
None
1-30
>30
0-5
>5-100
>100-600
>600
125
Model 7: Dormant season
The logistic regression correctly classified 93.8% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.939).
Seven predictor variables were identified to be significant in discriminating between
the sites occupied by the two types of wildebeest while they were grazing during the
dormant season. The important habitat separating variables were: slope (V17), grass
cover (V21), rock cover (V26), grass leaf height (V28), distance to shade (V31),
altitude (V35) and time since last burn (V36) (-2 log likelihood = 242.3448; df = 19;
p<0.0001) (Table 6.9). Distance to shade contributed the most to separating the
habitats of the black and blue wildebeest while grazing during the dormant season
(Wald $2 = 37.5790; df = 3; p<0.0001), while grass cover (Wald $2 = 12.0868; df = 2;
p = 0.0024) and altitude (Wald $2 = 13.0609; df = 3; p = 0.0045) contributed the least.
During the dormant season black wildebeest preferred moderate slopes while blue
wildebeest were most likely to occupy gentle and moderate slopes. Blue wildebeest
preferred to graze where there was no rock cover or where the rock cover was
>30%, whereas black wildebeest were most likely to graze in habitats where the rock
cover was 1 to 30%. Blue wildebeest showed an equal preference for areas with all
grass leaf height classes, but preferred not to graze in areas with grass leaf heights
of >50 to 100 mm. Black wildebeest showed a clear preference for grass leaf heights
of >100 to 400 mm when grazing. When grazing, blue wildebeest preferred distances
of >5 to 100 m away from the nearest shade, while black wildebeest were most likely
to utilise habitats >600 m from the nearest shade although they showed some
tendency to graze in habitats that were >100 to 600 m from the nearest shade during
the dormant season. Blue and black wildebeest showed no specific preference for a
specific altitudinal range when grazing in the dormant season. Blue wildebeest were
more likely than black wildebeest to graze in recently burnt areas.
Model 8: Early growing season
The logistic regression correctly classified 91.0% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.917).
Five predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest while they were grazing during the late
growing season. The important habitat separating variables were: habitat type (V15),
aspect (V16), distance to shade (V31), altitude (V35) and time since last burn (V36)
(-2 log likelihood = 245.3268; df = 14; p<0.0001) (Table 6.10).
126
Table 6.9: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 7 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Slope
Grass cover
Rock cover (%)
Grass leaf height (mm)
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
Flat
Gentle
Moderate
Sparse
Medium
Dense
None
1-30
>30
0-50
>50-100
>100-400
>400
0-5
>5-100
>100-600
>600
10.87
17.39
71.74
30.43
50.00
19.57
32.61
65.22
2.17
17.39
21.74
47.83
13.04
0.00
10.87
32.61
56.52
13.04
34.78
23.91
28.26
95.65
0.00
0.00
0.00
4.35
3.63
44.04
52.33
22.28
56.48
21.24
45.08
22.80
32.12
29.53
9.33
34.20
26.94
3.63
70.98
24.35
1.04
31.09
25.39
18.13
25.39
70.98
1.04
1.04
24.87
2.07
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
127
Table 6.10: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 8 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
34.09
4.55
11.36
50.00
47.73
52.27
0.00
0.00
6.82
93.18
25.00
43.18
25.00
6.82
90.91
0.00
0.00
9.09
0.00
4.57
11.43
22.29
21.14
40.57
79.43
20.57
5.14
70.29
21.71
2.86
41.14
23.43
12.57
22.86
64.57
1.14
1.14
33.14
0.00
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
128
Distance to shade contributed the most to separating out the black and blue
wildebeest in this analysis (Wald $2 = 38.7172; df = 3; p<0.0001) while habitat type
contributed the least (Wald $2 = 12.7942; df = 4; p = 0.0123).
Blue wildebeest preferred to graze on northerly slopes and were more likely to do so
than black wildebeest. Black wildebeest grazed on all aspects without preference.
During the early growing season blue wildebeest were most likely to be found grazing
at distances >5 to 100 m from the nearest shade but on occasion they showed a
tendency to graze >100 m away from the nearest shade. Black wildebeest showed a
strong preference for grazing >600 m from the nearest shade. Blue wildebeest
tended to graze with the most likelihood at low altitudes during the early growing
season, while black wildebeest preferred to graze at altitudes >1340 to 1360 m. Blue
wildebeest were more likely to graze at the high altitudes than the black wildebeest.
Blue wildebeest were also more likely to utilise recently burnt areas than black
wildebeest.
Time of the day analyses
Model 9: >05:00 – 10:00
The logistic regression correctly classified 92.3% of the habitat samples according to
type of wildebeest based on the time of the day. The discriminating power of this
model was excellent (c = 0.928). Eight predictor variables were identified to be
significant in discriminating between the sites occupied by the two types of
wildebeest during this time of the day. The important habitat separating variables
were: habitat type (V15), aspect (V16), slope (V17), woody vegetation cover (V20),
distance to shade (V31), altitude (V35), time since last burn (V36) and
geomorphology (V40) (-2 log likelihood = 504.2736; df = 21; p<0.0001) (Table 6.11).
Distance to shade contributed the most to separating the black from the blue
wildebeest in this analysis (Wald $2 = 63.0404; df = 3; p<0.0001) followed by the time
since last burn (Wald $2 = 40.4576; df = 4; p<0.0001). Woody vegetation cover
contributed the least (Wald $2 = 9.3081; df = 2; p = 0.0095).
129
Table 6.11: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 9 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Habitat type
Aspect
Slope
Woody vegetation cover
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Geomorphology
Category
Black wildebeest
Blue wildebeest
0.00
9.41
9.41
9.41
71.76
38.82
61.18
11.76
34.12
54.12
100.00
0.00
0.00
0.00
9.41
27.06
63.53
9.41
27.06
25.88
37.65
98.82
0.00
0.00
1.18
0.00
44.71
41.18
14.12
8.70
3.82
28.03
28.03
31.42
72.40
27.60
3.82
48.20
47.98
52.23
32.91
14.86
13.80
64.76
20.59
0.85
40.98
23.14
16.99
18.90
65.61
1.49
2.12
29.72
1.06
20.59
75.58
3.82
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Flat
Gentle
Moderate
None
Sparse
Open
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
130
The sandy grassland habitat type was favoured by black wildebeest in the morning
(>05:00 to 10:00), while blue wildebeest were equally likely to occur on old lands, on
rocky grasslands and in sandy grasslands during this time. Blue wildebeest were
most likely to utilise northerly slopes while black wildebeest were most likely to utilise
southerly slopes. Blue wildebeest were most likely to occur in habitats with no woody
vegetation. Black wildebeest showed a strong preference for habitats where there
was no woody vegetation cover. Blue wildebeest preferred areas of >5 to 100 m from
the nearest shade, but were almost as likely to utilise areas >100 m away from the
nearest shade. Black wildebeest were most likely to occur where the nearest shade
was >600 m away, but also were as likely to utilise habitats where the nearest shade
was >100 to 600 m away. Blue wildebeest were more likely to utilise lower altitudes
than black wildebeest, and also most likely to occur in habitats with a convex
geomorphology, while black wildebeest showed an equal preference for both convex
and concave areas.
Model 10: >10:00 – 14:00
The logistic regression correctly classified 90.9% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.918).
Four predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest during this time of the day. The
important habitat separating variables were: habitat type (V15), rock cover (V26),
distance to shade (V31) and time since last burn (V36) (-2 log likelihood = 491.8874;
df = 13; p<0.0001) (Table 6.12). Distance to shade contributed the most to the
analysis (Wald $2 = 89.0344; df = 3; p<0.0001) while habitat type contributed the
least (Wald $2 = 21.7419; df = 4; p = 0.0002).
Blue wildebeest were equally likely to occur on old lands, rocky grasslands and
sandy grasslands, while black wildebeest preferred sandy grasslands and were more
likely than blue wildebeest to occur in moist grasslands. During midday (>10:00 to
14:00), the hottest part of the day, blue wildebeest were more likely to frequent
habitats where there was no rock cover. Black wildebeest were least likely to use
areas where the rock cover was >30%, but showed an equal preference for habitats
where there was no rock cover and where it was ”30%.
131
Table 6.12: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 10 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Rock cover (%)
Distance to shade (m)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
21.09
17.01
2.72
59.18
49.66
49.66
0.68
0.00
0.00
12.93
87.07
99.32
0.00
0.00
0.00
0.68
11.99
5.24
27.72
26.59
28.46
58.43
15.36
26.22
22.47
54.31
21.72
1.50
53.18
3.37
2.25
40.45
0.75
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
None
1-30
>30
0-5
>5-100
>100-600
>600
2001 or earlier
2002
2003
2004
2005
132
Blue wildebeest were most likely to occur at distances >5 to 100 m away from the
nearest shade. Black wildebeest consistently favoured habitats where the shade was
>600 m away.
Model 11: >14:00 – 19:00
The logistic regression correctly classified 80.8% of the habitat samples according to
type of wildebeest. The discriminating power of this model was good (c = 0.829).
Three predictor variables were identified to be significant in discriminating between
the sites occupied by the two types of wildebeest during this time of the day. The
important habitat separating variables were: habitat type (V15), aspect (V16) and
distance to shade (V31) (-2 log likelihood = 195.5085; df = 8; p<0.0001) (Table 6.13).
Distance to shade contributed the most to separating the black from the blue
wildebeest in this analysis (Wald $2 = 50.8243; df = 3; p<0.0001) while habitat type
contributed the least (Wald $2 = 28.4887; df = 4; p<0.0001).
After 14:00 black wildebeest were more likely to utilise old lands than other times of
the day but they were still most likely to be found on sandy grasslands. Blue
wildebeest were most likely to be found on old lands and sandy grasslands during
this time. Blue wildebeest preferred northerly slopes, while black wildebeest showed
a strong preference for southerly slopes. Blue wildebeest preferred habitats >5 to
100 m from the nearest shade, while black wildebeest showed a strong preference
for distances >600 m from the nearest shade.
Temperature analyses
Model 12: Temperatures <15 °C
There were not enough observations in this temperature range to discriminate
between sites at this level.
Model 13: Temperatures •15-25 °C
The logistic regression correctly classified 90.9% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.913).
Ten predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest when the temperatures were •15-25
°C.
133
Table 6.13: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 11 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Distance to shade (m)
Category
Black wildebeest
Blue wildebeest
0.00
17.39
26.09
4.35
52.17
0.00
100.00
0.00
0.00
0.00
100.00
11.99
0.37
34.83
21.72
31.09
85.02
14.98
19.48
57.68
22.85
0.00
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
0-5
>5-100
>100-600
>600
134
The important habitat separating variables were: habitat type (V15), aspect (V16),
slope (V17), landscape position (V18), woody vegetation cover (V20), grass cover
(V21), total grass height (V27), distance to shade (V31), altitude (V35), and time
since last burn (V36) (-2 log likelihood = 453.8716; df = 27; p<0.0001) (Table 6.14).
Distance to shade contributed the most to the analysis (Wald $2 = 66.2119; df = 3;
p<0.0001), while slope (Wald $2 = 6.8763; df = 2; p = 0.0321) and woody vegetation
cover (Wald $2 = 8.1849; df = 2; p = 0.0167) contributed the least.
When the temperature was from 15 to 25°C, blue wildebeest showed an equal
likelihood to utilise old lands and rocky grasslands, but were most likely to utilise
sandy grasslands. Black wildebeest were most likely to be found in the sandy
grasslands. Blue wildebeest were twice as likely to occur on northerly slopes than
black wildebeest, as the black wildebeest showed a strong preference for southerly
slopes. Blue wildebeest were most likely to occur on plains and also with an equal
likelihood on gently sloping landscapes, while black wildebeest preferred the plains
and to some extent the plateaus. Black wildebeest showed a strong preference for
habitats where there was no woody vegetation, while blue wildebeest were equally
likely to occur in habitats without woody vegetation and in habitats where the woody
vegetation cover was sparse or open. Blue wildebeest were equally likely to spend
their time in areas containing all grass height categories, while black wildebeest
preferred only habitats where the total grass height was >50 to 800 mm. Blue
wildebeest were most likely to occur at distances of >5-100 m from the nearest
shade, while black wildebeest preferred distance of >600 m from the nearest shade.
Blue wildebeest were most likely to occur at altitudes of <1340 m, while black
wildebeest tended to utilise with equal likelihood altitudes >1340 to 1360 m.a.s.l. and
>1380 m.a.s.l. Both types of wildebeest were most likely to be found in habitats that
had not been recently burnt but blue wildebeest were much more likely to utilise
recently burnt habitats than black wildebeest.
135
Table 6.14: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 13 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Slope
Landscape position
Woody vegetation cover
Grass cover
Total grass height (mm)
Distance to shade (m)
Altitude (m)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
14.29
4.76
11.90
69.05
39.29
60.71
14.29
30.95
54.76
2.38
64.29
14.29
19.05
100.00
0.00
0.00
19.05
46.43
34.52
4.76
40.48
51.19
3.57
0.00
4.76
28.57
66.67
17.86
30.95
16.67
34.52
98.81
0.00
0.00
0.00
1.19
7.98
4.49
25.69
25.69
36.16
70.57
29.43
2.74
44.14
53.12
29.18
55.36
2.74
12.72
54.61
32.67
12.72
21.20
48.63
30.17
18.95
29.43
25.19
26.43
12.97
64.34
20.20
2.49
36.16
20.45
18.70
24.69
62.34
3.24
2.00
31.17
1.25
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Flat
Gentle
Moderate
Gentle slopes
Plains
Plateau
Valley
None
Sparse
Open
Sparse
Medium
Dense
0-50
>50-500
>500-800
>800
0-5
>5-100
>100-600
>600
>1340
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
136
Model 14: Temperatures >25 °C
The logistic regression correctly classified 93% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.935).
Eight predictor variables were identified to be significant in discriminating between
the sites occupied by the two types of wildebeest when the temperatures were
>25°C. The important habitat separating variables were: habitat type (V15), aspect
(V16), total grass height (V27), distance to shade (V31), altitude (V35), and time
since last burn (V36), geomorphology (V40) and forb : grass ratio (V41) (-2 log
likelihood = 823.2523; df = 23; p<0.0001) (Table 6.15). Distance to shade contributed
the most to separating black from blue wildebeest in this analysis (Wald $2 =
90.9291; df = 3; p<0.0001) while aspect contributed the least (Wald $2 = 10.0159; df
= 1; p = 0.0016).
At temperatures >25 °C, blue wildebeest were most likely to occur on the old lands,
while black wildebeest preferred sandy grasslands. Blue wildebeest were most likely
to occur on northerly aspects while black wildebeest preferred southerly aspects.
Blue wildebeest were most likely to utilise habitats where the total grass height was
>50 to 500 mm, while black wildebeest were most likely to utilise habitats where the
total grass height was >50 to 800 mm. Blue wildebeest preferred distances >5 to 100
m away from the nearest shade, whereas black wildebeest preferred distances >600
m away from the nearest shade and were least likely to occur at distances <100 m
away from the nearest shade. Blue wildebeest preferred areas with a convex
geomorphology and were least likely to utilise flat areas, while black wildebeest were
most likely to occur in areas with a concave geomorphology, but also utilised areas
with convex and flat geomorphologies with an equal likelihood to each other but less
than that for concave areas. Blue wildebeest were more likely to occur in habitats
with a greater percentage of forbs making up the herbaceous layer than did black
wildebeest.
137
Table 6.15: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 14 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Total grass height (mm)
Distance to shade (m)
Altitude (m.a.s.l)
Time since last burn
Geomorphology
Forb : grass ratio
Category
Black wildebeest
Blue wildebeest
0.00
18.23
17.68
5.52
58.56
38.12
61.88
4.97
46.96
40.33
7.73
0.00
5.52
15.47
79.01
20.44
38.12
19.34
22.10
96.13
0.00
0.00
3.31
0.55
46.96
28.18
24.86
25.97
47.51
25.41
1.10
13.08
3.22
34.00
22.33
27.36
76.46
23.54
15.29
42.05
23.94
18.71
22.94
54.73
20.93
1.41
36.62
27.77
19.32
16.30
67.00
0.60
3.02
29.18
0.20
17.51
79.28
3.22
11.07
36.42
46.48
6.04
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
0-50
>50-500
>500-800
>800
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
0:100
10:90
30:70
50:50
138
Cloud cover analyses
Model 15: No cloud cover (Clear skies)
The logistic regression correctly classified 93.1% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.927).
Ten predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest when there was no cloud cover. The
important habitat separating variables were: aspect (V16), slope (V17), rock cover
(V26), total grass height (V27), plant utilisation (V29), visibility (V30), distance to
shade (V31), altitude (V35), time since last burn (V36), and geomorphology (V40) (-2
log likelihood = 496.8198; df = 26; p<0.0001) (Table 6.16). Distance to shade
contributed the most to separating the black from the blue wildebeest in this analysis
(Wald $2 = 71.1455; df = 3; p<0.0001), while visibility (Wald $2 = 7.3556; df = 3; p =
0.0614) and plant utilisation (Wald $2 = 7.9505; df = 3; p = 0.0470) contributed the
least.
During days with clear skies, blue wildebeest were most likely to occur on northern
slopes while black wildebeest preferred southern slopes. Blue wildebeest were also
most likely to occur on moderate and gentle slopes and were least likely to utilise
areas that had no slope, while black wildebeest preferred moderate slopes and were
less likely to utilise gentle slopes and flat areas. Blue wildebeest preferred habitats
free of rock cover and were less likely to utilise areas with a rock cover of >30%.
Black wildebeest, however, were most likely to occur in habitats with a rock cover of
<30% and least likely to utilise areas where it was >30%. Blue wildebeest were
equally likely to occur at all classes of total grass height, while black wildebeest were
least likely to occur in habitats where the total grass height was <50 mm or >800 mm.
Blue wildebeest preferred habitats where the visibility was relatively low, while black
wildebeest were most likely to occupy habitats with a high degree of visibility. Blue
wildebeest preferred habitats where the distance to shade was >5 to 100 m, with the
highest likelihood of occurrence in the 0 - 5 m shade class.
139
Table 6.16: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 15 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Aspect
Slope
Rock cover (%)
Total grass height (mm)
Plant utilisation
Visibility (m)
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Geomorphology
Category
Black wildebeest
Blue wildebeest
35.96
64.04
16.85
21.35
61.80
33.71
60.67
5.62
3.37
48.31
43.82
4.49
3.37
30.34
21.35
44.94
0.00
12.36
44.94
42.70
0.00
5.62
21.35
73.03
16.85
41.57
15.73
25.84
100.00
0.00
0.00
0.00
0.00
50.56
29.21
20.22
72.80
27.20
1.76
48.61
49.62
56.17
20.65
23.17
20.40
36.52
20.65
22.42
6.55
26.20
23.68
43.58
7.30
34.01
31.49
27.20
16.62
57.93
23.17
2.27
35.52
21.91
20.91
21.66
61.21
2.77
0.76
33.75
1.51
21.91
76.07
2.02
North-facing
South-facing
Flat
Gentle
Moderate
None
1-30
>30
0-50
>50-500
>500-800
>800
Low
Moderate
High
Excessive
0-50
>50-100
>100-200
>200
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
140
Black wildebeest were most likely to occur at distances of >600 m from the nearest
shade. Blue wildebeest preferred low altitudes of <1340 m above sea level, while
black wildebeest preferred areas that had an altitude >1340 to 1360 m above sea
level. Blue wildebeest were most likely to occur in areas with a convex
geomorphology,
while
black
wildebeest
preferred
areas
with
a
concave
geomorphology, but they were also found in convex and flat areas with equal
likelihood if slightly less than the likelihood for concave areas.
Model 16: >0 – 50% cloud cover (Partly cloudy)
The logistic regression correctly classified 92.4% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.927).
Ten predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest when there was >0 to 50% cloud cover.
The important habitat separating variables were: habitat type (V15), aspect (V16),
landscape position (V18), grass cover (V21), total grass height (V27), distance to
shade (V31), altitude (V35), time since last burn (V36), geomorphology (V40) and
forb : grass ratio (V41) (-2 log likelihood = 520.8920; df = 28; p<0.0001) (Table 6.17).
Distance to shade (Wald $2 = 57.7461; df = 3; p<0.0001) and habitat type (Wald $2 =
38.7367; df = 4; p<0.0001) contributed the most to separating black from blue
wildebeest in this analysis, while grass cover (Wald $2 = 6.490; df =2; p = 0.0389),
landscape position (Wald $2 = 9.110; df =4; p = 0.0279) and forb : grass ratio (Wald
$2 = 9.2272; df = 3; p = 0.0264) contributed the least.
With partly cloudy skies, blue wildebeest were most likely to occur on old lands and
sandy grasslands, while black wildebeest preferred sandy grasslands but also
showed some preference for moist grasslands. Blue wildebeest were most likely to
occur on northern slopes while black wildebeest preferred southern slopes. Blue
wildebeest were least likely to occur on plateaus, while black wildebeest were least
likely to utilise gentle slope landscapes. Blue wildebeest were most likely to occur in
habitats where the total grass height >50 to 500 mm and also with some likelihood in
areas where the grass was >500 mm tall. Black wildebeest preferred areas where
the total grass height was >50 to 800 mm and were least likely to occur where it was
>800 mm tall. Blue wildebeest occurred with the highest likelihood at distances of >5
to 100 m away from the nearest shade, but were also likely to occur at distances <5
m away and >100 to 600 m away with an equal likelihood. Black wildebeest preferred
distances >600 m from the nearest shade and were least likely to occur at distances
<5 m away.
141
Table 6.17: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 16 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Landscape position
Grass cover
Total grass height (mm)
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Geomorphology
Forb : grass ratio
Category
Black wildebeest
Blue wildebeest
0.00
20.51
17.95
8.97
52.56
38.46
61.54
6.41
58.97
19.23
15.38
16.67
46.15
37.18
10.26
44.87
37.18
7.69
0.00
12.82
10.26
76.92
20.51
32.05
17.95
29.49
98.72
0.00
0.00
1.28
0.00
50.00
19.23
30.77
23.08
51.28
24.36
1.28
11.17
2.73
31.76
24.81
29.53
76.92
23.08
30.77
56.33
1.49
11.41
24.81
45.91
29.28
12.90
40.45
25.81
20.84
19.11
59.80
20.10
0.99
36.97
24.57
17.62
20.84
70.97
0.50
3.97
24.32
0.25
17.12
79.90
2.98
12.66
35.98
45.66
5.71
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Gentle slopes
Plains
Plateau
Valley
Sparse
Medium
Dense
0-50
>50-500
>500-800
>800
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
0:100
10:90
30:70
50:50
142
Blue wildebeest were most likely to occur in areas with a convex geomorphology and
least likely in flat areas. Black wildebeest preferred areas with a concave
geomorphology but were just as likely to spend time on flat areas. Black wildebeest
were more likely than blue wildebeest to occur in habitats with a low forb : grass ratio.
Model 17: >50% cloud cover (Overcast)
The logistic regression correctly classified 91.2% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.918).
Four predictor variables were identified to be significant in discriminating between the
sites occupied by the two types of wildebeest when there was >50% cloud cover.
The important habitat separating variables were: habitat type (V15), distance to
shade (V31), altitude (V35) and time since last burn (V36) (-2 log likelihood =
325.9301; df = 14; p<0.0001) (Table 6.18). Distance to shade contributed the most to
separating out the black and blue wildebeest in this analysis (Wald $2 = 59.3605; df =
3; p<0.0001) and altitude contributed the least (Wald $2 = 23.7720; df = 3; p<0.0001).
During overcast conditions, blue wildebeest were most likely to occur on old lands,
rocky grasslands and sandy grasslands but least likely to occur on moist grasslands.
Black wildebeest were most likely to occur on sandy grasslands and least likely to
occur in Burkea woodlands or rocky grasslands. Blue wildebeest were most likely to
utilise distances of >5 to 100 m away from the nearest shade, while black wildebeest
preferred distances >600m away from the nearest shade. Blue wildebeest preferred
the low altitudes of <1340 m, while black wildebeest were most likely to be found at
altitudes of >1340 to 1360 m above sea level.
Social group analyses
Model 18: Female herds
The logistic regression correctly classified 98.3% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c = 0.982).
143
Table 6.18: The percentage of the probabilities (indicating the presence of either a
black (0.8”3” RU D EOXH ”3” ZLOGHEHHVW DVVRFLDWHG ZLWK HDFK FDWHJRU\ RI
the variables selected by Model 17 in the logistic regression analysis to separate
between the habitat of the black and blue wildebeest. Percentages in bold indicate
those categories that were significantly selected by the one type of wildebeest over
the other (p<0.05). Data collected on Ezemvelo Nature Reserve from January 2004
to August 2005
Variables selected by the
model
Habitat type
Distance to shade (m)
Altitude (m.a.s.l.)
Time since last burn
Category
Black wildebeest
Blue wildebeest
0.00
16.25
16.25
2.50
65.00
0.00
0.00
25.00
75.00
16.25
40.00
18.75
25.00
98.75
0.00
0.00
0.00
1.25
6.73
1.44
28.85
26.44
36.54
12.98
70.19
15.87
0.96
38.94
25.96
14.42
20.67
64.42
2.88
2.40
29.81
0.48
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
0-5
>5-100
>100-600
>600
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
144
Eight predictor variables were identified to be significant in discriminating between
the sites occupied by the female herds of the two types of wildebeest. The important
habitat separating variables were: habitat type (V15), aspect (V16), slope (V17),
distance to water (V19), rock cover (V26), distance to shade (V31), time since last
burn (V36) and geomorphology (V40) (-2 log likelihood = 496.9715; df = 21;
p<0.0001) (Table 6.19). Distance to shade contributed the most to separating the
black from the blue wildebeest female herds in this analysis (Wald $2 = 33.8164; df =
3; p<0.0001), followed by habitat type (Wald $2 = 18.2259; df = 4; p<0.0001). Rock
cover (Wald $2 = 6.9409; df = 2; p = 0.0311) contributed the least.
Female herds of blue wildebeest were most likely to occur on northerly slopes while
black wildebeest females preferred southerly slopes. Blue wildebeest female herds
were most likely to utilise moderate slopes and least likely to utilise flat areas, while
black wildebeest female herds preferred moderate slopes, but they were also likely to
utilise flat areas and gentle slope. Blue wildebeest female herds were most likely to
occur >50 to 300 m away from the nearest water, whilst black wildebeest female
herds preferred distances away from water of >300 m. Blue wildebeest female herds
were most likely to be found in habitats with no rock cover while black wildebeest
female herds were most likely to utilise areas where the rock cover was <30%. Blue
wildebeest female herds preferred distances of >5 to 100 m away from the nearest
shade, whereas the female herds of black wildebeest preferred distances >600 m
away from the nearest shade and were least likely to use areas where the nearest
shade was <100 m away. Blue wildebeest female herds were most likely to utilise
areas with a convex geomorphology, while black wildebeest preferred areas with a
concave geomorphology. Whereas blue wildebeest were least likely to utilise flat
areas, black wildebeest were most likely to utilise areas with both a flat and a convex
geomorphology.
Model 19: Territorial bulls
The logistic regression correctly classified 91.2% of the habitat samples according to
type of wildebeest. The discriminating power of this model was excellent (c=0.909).
Eight predictor variables were identified to be significant in discriminating between
the sites occupied by the territorial bulls of the two types of wildebeest. The important
habitat separating variables were: aspect (V16), woody vegetation cover (V20),
distance to shade (V31), erosion (V34), altitude (V35), time since last burn (V36),
geomorphology (V40) and forb : grass ratio (V41) (-2 log likelihood = 298.4180, df =
20, p<0.0001) (Table 6.20).
145
Table 6.19: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 18 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Habitat type
Aspect
Slope
Distance to water (m)
Rock cover (%)
Distance to shade (m)
Time since last burn
Geomorphology
Category
Black wildebeest
Blue wildebeest
0.00
11.92
16.56
7.28
64.24
28.48
71.52
22.52
23.18
54.30
12.58
18.54
33.11
35.76
33.11
60.93
5.96
0.00
9.27
17.22
73.51
98.01
0.00
0.00
1.32
0.66
43.71
29.14
27.15
3.93
3.93
36.79
16.07
39.29
70.71
29.29
2.50
37.50
60.00
12.50
42.86
26.07
18.57
69.64
15.71
14.64
10.36
60.00
27.14
2.50
50.71
2.86
2.86
41.79
1.79
20.71
76.79
2.50
Burkea woodland
Moist grassland
Old land
Rocky grassland
Sandy grassland
North-facing
South-facing
Flat
Gentle
Moderate
0-50
>50-300
>300-500
>500
None
1-30
>30
0-5
>5-100
>100-600
>600
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
146
Table 6.20: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 19 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Aspect
Woody vegetation cover
Distance to shade (m)
Erosion
Altitude (m.a.s.l.)
Time since last burn
Geomorphology
Forb : grass ratio
Category
Black wildebeest
Blue wildebeest
15.00
85.00
100.00
0.00
0.00
0.00
0.00
15.00
85.00
10.00
60.00
30.00
0.00
50.00
15.00
35.00
100.00
0.00
0.00
0.00
0.00
60.00
20.00
20.00
20.00
65.00
15.00
0.00
72.85
27.15
47.74
36.35
15.91
19.81
60.69
17.78
1.72
15.29
70.67
14.04
41.34
17.78
19.34
21.53
70.98
1.72
2.65
24.18
0.47
16.07
80.03
3.90
14.66
36.51
43.06
5.77
North-facing
South-facing
None
Sparse
Open
0-5
>5-100
>100-600
>600
Low
Moderate
High
”
>1340-1360
>1360-1380
>1380
2001 or earlier
2002
2003
2004
2005
Concave
Convex
Flat
0:100
10:90
30:70
50:50
147
Distance to shade contributed the most to the separation of black and blue
wildebeest in this analysis (Wald $2 = 46.9783; df = 3; p<0.0001) followed closely by
altitude (Wald $2 = 40.1248; df = 3; p<0.0001), while aspect contributed the least
(Wald $2 = 5.7413; df = 1; p = 0.0166).
Blue wildebeest territorial bulls were most likely to occur on northerly aspects while
those of the black wildebeest preferred on southerly aspects. Blue wildebeest
territorial bulls were equally likely to occur in habitats with no woody vegetation and
those with sparse woody vegetation. Black wildebeest territorial bulls showed a
strong preference for areas where there was no woody vegetation. Blue wildebeest
territorial bulls were most likely to utilise areas at distances of >5 to 100 m away from
the nearest shade, and least likely to utilise areas at distances >600 m, while black
wildebeest territorial bulls were most likely to occur at distances >600 m away from
the nearest shade. Blue wildebeest territorial bulls were more likely to occur where
the erosion was moderate, whereas black wildebeest were more likely to occur
where the degree of erosion was high. Blue wildebeest territorial bulls preferred
altitudes of <1340 m, while black wildebeest territorial bulls were most likely to occur
at altitudes >1340 to 1360 m above sea level. Blue wildebeest territorial bulls were to
some extent likely to make use of recently burnt areas but those of black wildebeest
were least likely to utilise recently burnt areas. Blue wildebeest territorial bulls
preferred areas with a convex geomorphology and were least likely to occur in flat
areas, while those of black wildebeest preferred areas with a concave
geomorphology but were also likely to utilise areas with a convex and flat
geomorphology to some degree. Blue wildebeest territorial bulls were more likely
than those of black wildebeest to utilise areas where the forb : grass ratio was 10:90.
Model 20: Bachelor herds
The logistic regression correctly classified 81.2% of the habitat samples according to
type of wildebeest. The discriminating power of this model was good (c = 0.838). Two
predictor variables were identified to be significant in discriminating between the sites
occupied by the bachelor herds of the two types of wildebeest. The important habitat
separating variables were: distance to shade (V31) and forb : grass ratio (V41) (-2
log likelihood = 73.4035; df = 6; p<0.0001) (Table 6.21). Distance to shade
contributed the most to differentiating black from blue wildebeest habitat in this
analysis (Wald $2 = 22.958; df = 3; p<0.0001), while the forb : grass ratio contributed
the least (Wald $2 = 14.5510; df = 3; p = 0.0022).
148
Table 6.21: The percentage of the probabilities (indicating the presence of either a black
(0.8”P”1.0) or a blue (0”P”0.2) wildebeest) associated with each category of the variables
selected by Model 20 in the logistic regression analysis to separate between the habitat of the
black and blue wildebeest. Percentages in bold indicate those categories that were
significantly selected by the one type of wildebeest over the other (p<0.05). Data collected on
Ezemvelo Nature Reserve from January 2004 to August 2005
Variables selected by the
model
Distance to shade (m)
Forb : grass ratio
Category
Black wildebeest
0-5
>5-100
>100-600
>600 m
0:100
10:90
30:70
50:50
0.00
0.00
0.00
100.00
0.00
65.38
34.62
0.00
149
Blue wildebeest
24.29
55.71
20.00
0.00
38.57
8.57
42.86
10.00
Blue wildebeest bachelor herds tended to favour distances of >5 to 100 m away from
the nearest shade, while those of black wildebeest were least likely to venture <600
m away from the nearest shade. Bachelor herds of the blue wildebeest were more
likely to utilise areas with a forb : grass ratio of 30:70 than those of black wildebeest.
DISCUSSION
Habitat separation has been demonstrated to be one of the most common forms of
resource partitioning in sympatric species (Cody 1978; Werner and Hall 1979;
Reinert 1984; Wei et al. 2000). The present study has demonstrated that habitat
separation at least at the meso-habitat scale could aid the co-existence of black and
blue wildebeest at Ezemvelo Nature Reserve.
The present study showed that the habitats selected by the black and blue
wildebeest could be separated in terms of a number of mainly physical habitat
factors. Therefore, black and blue wildebeest used the habitat available differentially,
and thus accomplished resource partitioning. This pattern of resource separation
varied across the different ecological seasons, times of the day and weather
conditions, indicating that behavioural adjustments were being made depending on
the circumstances encountered and differences between the two types of wildebeest
were not merely inherent. The selection of different habitat features by different types
of wildlife has been related to antipredator strategies, protection against adverse
climatic conditions, reduction of interspecific competition, and establishment of routes
to reproductive and feeding sites (Alvarez-Cardenas et al. 2001). It is, however,
based on the results of the present study, expected that the pattern of resource
partitioning observed could possibly be explained by means of specific differences in
territoriality, body size, and temperature tolerance between the two types of
wildebeest (Brink et al. 1999; Skinner and Chimimba 2005; Codron and Brink In
press).
All the analyses showed that the distance to shade was an extremely important
variable separating the habitat use of the black and blue wildebeest at Ezemvelo
Nature Reserve. Distance to shade is affected by the vegetation structure. Blue
wildebeest were shown to be much more dependent on nearby shade than black
wildebeest and actively sought out shade during the hot parts of the day especially
during the hot months in the late growing season and the early growing season in the
study area (Pers. obs.). Blue wildebeest prefer open savanna habitats but readily
150
make use of other habitats under certain conditions (Hirst 1975). Therefore it is
possible to find the blue wildebeest in open habitats during certain times of the year
when weather conditions permit it. They, however, tend to remain in habitats with
cover nearby unless certain factors such as lack of suitable habitat, or overpopulation
forcing bachelors into sub-optimal areas, prevent it. It would be these bachelors,
which could be the cause of hybridisation between the two types of wildebeest
(Vrahimis 2003b). Due to their body conformation, black wildebeest are able to
withstand direct sunlight at all irradiances and thus shade-seeking behaviour does
not normally form part of their behavioural repertoire at any time of the year (Skinner
and Chimimba 2005). For black wildebeest an open habitat is a necessity, while for
blue wildebeest a habitat with cover is optimal but not a necessity for survival.
Therefore, in areas where there is no cover, this separating mechanism between the
black and blue wildebeest would no longer be operative.
Distance to shade is related to visibility, exposure and woody vegetation structure.
Visibility only separated the habitat use of black from blue wildebeest when the skies
were clear, with black wildebeest preferring habitats with a higher visibility than blue
wildebeest. The results, therefore indicate that both the black and blue wildebeest
require habitats with a high visibility, which has been suggested to be related to
antipredator behaviour (Hirst 1975).
Exposure never featured as a separating mechanism as all wildebeest were mostly
found in full sun. Woody vegetation cover proved to be an important habitat
separating mechanism throughout the entire study period (Model 1). It was also
important for the data representing the territorial bulls (Model 19), early in the
morning (Model 9) and at moderate temperatures (Model 13). Black wildebeest
territorial bulls only utilised habitats where there was no woody vegetation, while blue
wildebeest territorial bulls were more inclined to utilise habitats with sparse woody
vegetation.
The above findings can be related to the fossil record which suggests that
morphological aspects which are associated with the distinct territorial social
behaviour of the black wildebeest were the first to change after the geographical
separation of the common wildebeest ancestor 1 million years ago, indicating that a
shift in breeding behaviour (especially territoriality) accompanied the appearance of
the first ancestral black wildebeest (Brink et al. 1999). This shift to a more territorial
behaviour is linked to the evolution of treeless grasslands in the central interior of
151
southern Africa over a million years ago. This is borne out by the present results
which indicated that black wildebeest, but most notably the territorial bulls, selected
treeless habitats while blue wildebeest tended to be less selective for woody
vegetation cover.
Territorial bulls of the two types of wildebeest were also separated in terms of aspect
utilisation, erosion levels, altitude and forb : grass ratio (Model 19). In contrast, the
habitats of the female herds of the two types of wildebeest (Model 18) were not
separated in terms of any of these factors. This may possibly indicate that these
differentiating factors are solely the result of differences in the territorial behaviour of
the black and blue wildebeest.
Black wildebeest territorial bulls mainly utilised the southerly aspects while blue
wildebeest utilised the northerly aspects. Southerly aspects on the study area are
usually cooler and wetter than the northerly aspects. In contrast, northerly aspects
tended to be hot and dry (Tainton 1999). Since black wildebeest mainly occupied
treeless habitats with no shade at distances <600 m, another technique for cooling
may be needed. This could be achieved by the utilisation of the cooler southerly
aspects as opposed to the hotter northerly aspects. Blue wildebeest occurred close
to shade and in sparse woody vegetation and would therefore not need to
compensate by selecting the southerly aspects that would maximise cooling. The
orientation of slopes may afford protection against adverse weather conditions as it
determines temperature and wind differences during the day or among seasons
(Alvarez-Cardenas et al. 2001).
Territories occupied by black wildebeest territorial bulls were often more heavily
eroded than those occupied by blue wildebeest bulls. Due to increased territoriality,
black wildebeest spent more time in their territory and hence more time displaying
than blue wildebeest bulls, thus increasing the erosion levels locally. The black
wildebeest tends to overgraze and trample the areas where they stay for prolonged
periods (Von Richter 1971b). As water is easily available throughout the year there is
no incentive to move within the reserve. The size of the reserve makes seasonal
movements impossible. The behaviour of territorial bulls pawing and horning the
ground aggravates the condition by removing the soil cover in the favoured areas in
the reserve as was described in other areas by Von Richter (1971b).
152
Black wildebeest bulls never occurred in the low-lying areas of the reserve, and
tended to occur mostly at high-lying locations, while blue wildebeest bulls favoured
low-lying areas and were much less likely to utilise high-lying areas. The higher the
altitude, the more exposed the landscape, the lower the temperatures and the higher
the visibility (Tainton 1999). Fabricius (1984) found that black wildebeest chose
northern gentle slopes and at higher altitudes. This was related to increased visibility
for territorial defense of territories. Altitude was found to be an important habitat
separating mechanism between the black and blue wildebeest during the dormant
season and the early growing season but not during the late growing season. This
result is due to the blue wildebeest utilising a wider range of altitudes during the late
growing season. The late growing season is when the rut takes place and the
normally relaxed breeding behaviour of the blue wildebeest becomes more rigid and
territoriality becomes more important.
When grazing, black and blue wildebeest were no longer separated by distance to
water, woody vegetation cover, plant utilisation, geomorphology and forb : grass ratio
as they were when all the activities combined were analysed. However, when grazing
the grass cover became an additional separating mechanism. With their remarkably
wide dental pad and incisor row, it has been suggested that wildebeest obtain a high
intake of forage on short and leafy grass swards (Owen-Smith 1985). They are also
most efficient at harvesting grass swards with a high biomass of green leaf (Murray
and Brown 1993).
Seasonal differences were found in the resource separation patterns between the
two types of wildebeest. During the late growing season, which by definition has a
high forage quantity (Chapter 2), only rock cover and distance to shade separated
the habitats utilised by the two types of wildebeest. This indicated that both types of
wildebeest were utilising a broad range of habitats during the late growing season.
Due to the abundance of resources during the late growing season, there may be no
incentive for one type of wildebeest to choose habitats where the other type of
wildebeest does not occur. Therefore during the period when resources are
abundant, it may become profitable to utilise resources other than the ones for which
the phenotype has been specifically selected (Gordon and Illius 1989). Niche breadth
and overlap measures will be analysed in Chapter 9 to study this aspect further.
However, the quality of the grass layer during the late growing season would
probably already be low (Tainton 1999). Therefore wildlife able to utilise a higher
153
quantity of forage per time period would be better able to maintain their condition
during this season.
It has been suggested that during periods of low resource abundance, selection
resulting from interspecific competition is likely to result in adaptations most suited for
resources that are used exclusively by a species (Schoener 1986). This will result in
both types of wildebeest concentrating in areas that would provide the most efficient
use of their time. However, if these areas are the same as those preferred by the
other type of wildebeest it would be expected that greater competition would result
from this greater selectivity. Therefore behavioural adjustments would have to be
made by one or both type of wildebeest to ensure greater separation in resource use
so that food and space are used optimally during the critical season when it is cold
and dry and the vegetation is dormant (dormant season). The rut takes place during
the dormant season too. Slope, grass height, plant utilisation, and altitude are the
additional factors which separate the habitats of the two types of wildebeest during
the dormant season which do not appear to separate their habitat use during the late
growing season. Slope and altitude are related to increased territoriality of the black
wildebeest as discussed above during the rut.
Black wildebeest concentrated on heavily utilised short grass areas during the
dormant season (critical season), while blue wildebeest tended to occur in areas with
taller grass and where the grass sward was less intensively utilised. According to a
model proposed by Illius and Gordon (1987), short grass swards impose greater
limitations on the food intake of larger herbivores than on smaller ones. Quite small
differences in body size are expected to cause exclusion of larger herbivore species
from swards that are able to sustain the smaller species, suggesting that this
mechanism may be important in the common phenomenon of ecological separation
in grazing species (Clutton-Brock and Harvey 1983; Gordon and Illius 1989).
Therefore the utilisation of short grass areas by the black wildebeest would enable it
to sustain itself through the dormant season (critical season), and in addition reduce
possible competition with the blue wildebeest during this critical period. Competition
could be reduced or minimised because the blue wildebeest with its larger body size
may find it difficult to sustain itself in these heavily utilised areas with short grass
during the critical period.
An analysis of data available for grazing wildebeest from the present study indicated
that grass cover was also an important habitat-separating variable between the two
154
types of wildebeest during the dormant season. The grazing sites that were occupied
by the black wildebeest during the dormant season tended to have a lower grass
cover than the areas occupied by the blue wildebeest. This may be due to the higher
impact of the black wildebeest on such areas during this time.
Limited separation between the habitats selected by the black and blue wildebeest
was observed during the early growing season (Model 4). This was the season when
calves were dropped by both types of wildebeest in the study area (Pers. obs.) and
resources are abundant.
Black and blue wildebeest showed a greater separation in habitat use early in the
morning, and this degree of separation decreases as the day progressed. From
sunrise until 10:00, slope, woody vegetation cover, altitude and geomorphology
separated the habitats used by the black and blue wildebeest. None of these factors
separated the two types of wildebeest later in the day. Temperatures were generally
lower, moisture levels were higher, the sun was in an easterly position and animals
are usually more active at periods of the day before 10:00 as compared to periods
after 10:00.
At moderate temperatures, black and blue wildebeest habitats were separated by
slope, landscape position, woody vegetation cover, and grass cover which did not
separate the two types of wildebeest at temperatures >25°C. Geomorphology and
forb : grass ratio were additional variables that separated the habitats of the two
types of wildebeest at high temperatures but not at moderate temperatures.
When the skies are clear, black and blue wildebeest habitat use was separated
based on slope, rock cover, plant utilisation, and visibility, none of which separated
the habitat use of the two types when cloud cover was >0%. During overcast
conditions (cloud cover >50%) fewer factors separated the habitat use of the two
types of wildebeest than when cloud cover was low or absent.
Black wildebeest did not utilize areas that were recently burnt if such areas did not
occur on the open plains. Blue wildebeest in contrast would totally change their
distribution patterns to make use of recently burnt areas and were usually the first
type of wildlife to be found on burnt areas (Melton 1978). High quality grass in the
post-burn areas and its attraction to herbivores has been repeatedly shown (Tomor
and Owen-Smith 2002). The observation that blue wildebeest would utilise recently
155
burnt areas no matter where their occurrence and black wildebeest would not, just
reaffirms the extreme form of area selection practiced by the black wildebeest (Von
Richter 1971b).
CONCLUSION
The question asked here was whether separation in terms of habitat utilisation
occurred between the black and blue wildebeest. The data presented here have
concluded that separation in terms of meso-habitat does occur between the black
and blue wildebeest. The present study clearly showed that there was always some
habitat factor causing resource partitioning between the two types of wildebeest and
that subtle differences in the way in which the two types reacted to the challenges
posed by the different ecological seasons, times of the day and weather conditions
may be sufficient to reduce competition between the two types of wildebeest in the
study area at the current population levels. Selection for specific environmental
parameters contributes to the ecological separation of the black and blue wildebeest
at Ezemvelo Nature Reserve.
156
CHAPTER 7: HABITAT SELECTION AND SEPARATION: MICRO-HABITAT
SCALE
INTRODUCTION
Analysis of the five broad habitats at Ezemvelo Nature Reserve revealed some
evidence for habitat separation between the black and blue wildebeest. Habitat
separation was also demonstrated at the mesohabitat scale. It was found that blue
wildebeest preferred habitats where cover was in the near vicinity and would select
areas where the grasses were short and in an immature state such as the old lands.
Black wildebeest tended to trade-off forage quality for an open habitat and therefore
selected the sandy grasslands where visibility was high. Grass cover, grass height,
plant utilisation, woody vegetation cover, forb : grass ratio, time since last burn,
distance to shade, visibility, and habitat type are all directly related to the vegetation.
Factors such as aspect, rockiness, and altitude probably show a correlation to the
vegetation type.
In habitats that may seem structurally and compositionally homogeneous, herbivores
may select some parts of these habitats over others in a non-random patchy way
(Novellie 1990). Due to the broad nature of the five identified habitats, it was
expected that there would be abundant small-scale variations in physical factors,
plant species composition, grass height and grass cover within these habitats. High
spatial heterogeneity within grassland habitats might be essential for maintaining
high wildlife species richness and abundance in relatively small nature reserves
(Owen-Smith 2004). Therefore, as well as making habitat use decisions at higher
scales, such as macro and mesohabitat scales, wild animals must also make
decisions at the finer microhabitat scale (While and McArthur 2005). These fine-scale
decisions would most frequently be influenced by either predation or foraging
requirements.
In a natural situation, a trade-off between high quality food patches and predation is
frequently made (While and McArthur 2005). Anecdotal evidence (Von Richter
1971b) suggests that black and blue wildebeest employ different strategies to deal
with the threat of predation. Black wildebeest prefer to outrun the predators and blue
wildebeest rely more on escape cover. However, with the low predation risk at
Ezemvelo Nature Reserve where the present study was done, the selection of
feeding sites was expected to be based on factors other than predation. In the
157
absence of predators, or when vulnerability to predation is low, grazers are expected
to choose feeding sites that offer the highest net energy gain per unit time spent
(Edwards 1983; While and McArthur 2005).
When food resources occur as discrete items such as fruits or seeds, resource
partitioning can be accommodated by selecting different food size classes. For large
grazers such as the black and blue wildebeest, the grass layer does not consist of
easily distinguishable discrete items such as seeds or fruits. However, the grass
sward has several characteristics that are related to quantity, such as grass biomass
and grass height, and quality such as grass species composition (Voeten and Prins
1999) that can form the basis for selection. These characteristics can be evaluated
and any differences found between feeding sites may indicate mechanisms for
resource partitioning at the microhabitat scale.
The major variables affecting food intake rate, and hence energy maximisation by
grazing herbivores, are the structural characteristics of the vegetation such as height,
density and the vertical distribution of biomass (Burlison et al. 1991; Illius et al. 1992).
Past research on assemblages of African grazing herbivores have indicated that the
grass sward structure, forage production, plant species composition, grass leaf
height and plant biomass, amongst others, are important factors determining
resource partitioning between the different grazing species (Bell 1971; Grobler 1983;
McNaughton 1985; Novellie and Strydom 1987; Novellie 1990; Wentzel et al. 1991;
Heitkönig and Owen-Smith 1998).
Both types of wildebeest at Ezemvelo Nature Reserve tended to concentrate their
grazing activities on certain patches (feeding sites) within their range. In order to
determine why such selection was taking place, a detailed study of the herbaceous
characteristics of these feeding sites (microhabitat patch scale) was conducted. It
was hoped that such a study would differentiate between the feeding sites selected
by black wildebeest and those selected by blue wildebeest at the habitat patch scale.
Through their grazing activity, trampling, defaecation and urination in these feeding
sites, wildebeest may affect the nutrient flow, vegetation community dynamics as well
as related fauna (Hester et al. 1999). Comparing sites that were utilised by
wildebeest and sites which were apparently structurally and compositionally similar,
but which were not utilised by wildebeest, could provide information which could be
158
used to indicate whether the wildebeest have a negative impact on their preferred
grazing sites or not.
Since both black and blue wildebeest have a similar mouth morphology, body size
and digestive system (Skinner and Chimimba 2005), it was expected that they would
graze at the same height and trophic level and thus little difference was expected to
exist between the feeding sites selected by each type of wildebeest at the habitat
patch scale. The present study will test this hypothesis.
The following key questions were therefore examined:
1. What herbaceous characteristics can be used to discriminate between a
feeding site selected by a black wildebeest and one being selected by a blue
wildebeest?
2. What similarities and differences exist between feeding sites selected by
black and blue wildebeest and sites that seemed to be suitable but which
were consistently not utilised?
METHODS
Seven black wildebeest and 11 blue wildebeest feeding sites were selected to study
their resource partitioning at the feeding site scale. A preferred feeding site was
defined as an area where most of the members of a female herd of a specific type of
wildebeest were found to be feeding actively at the same locality during at least three
of five consecutive transect investigations (Wentzel et al. 1991) which were
conducted during the habitat survey phase of the present study (Chapter 4).
Vegetation sampling of these sites was done by using the centre of the herd as the
centre of the sampling point. These centres were selected solely on the basis of high
animal densities with no consideration for vegetation composition as suggested by
Novellie (1990).
The following vegetation parameters were measured at each site by using the same
methodology as was applied for the measurement of the herbaceous characteristics
of the different habitat types (Chapter 5): Grass species composition, above-ground
standing crop (kg/ha), grass height (cm), grass leaf height (cm) and grass canopy
cover (%). Plant species density and diversity, veld condition, degree of utilisation
and grass biomass concentration were calculated from the above variables using the
same equations and methodology as was applied in Chapter 5.
159
In addition, 17 sites were surveyed where wildebeest, either black or blue, were
never recorded but where they were expected to occur. These non-utilisation sites
were selected by using a non-random, stratified sampling approach (Novellie 1990).
For this part of the study, and due to the time-consuming nature of this type of
analysis and the logistical constraints, it was decided to conduct an analysis of the
feeding sites on one occasion only and not to repeat the analysis over the three
ecological seasons. Surveys were therefore only done in the late growing season.
The ideal season for analysis of the vegetation characteristics would have been the
dormant season as this would have been when the food resources would have been
most limiting. However, the grass species in these feeding sites would have been
very difficult to identify and thus severely restricting the number of variables that
would have been possible to measure during the dormant season.
Statistical analysis
An ANOVA test was done (PROC GLM) to determine whether there were statistical
differences in vegetation characteristics between the sites that were utilised by black
wildebeest, blue wildebeest or not at all by either type of wildebeest. These tests
were done for each of the herbaceous characteristic variables listed above.
The herbaceous layer variables listed above were submitted to a step-wise
discriminant analysis (STEP DISCRIM) with site type (either blue wildebeest utilised,
black wildebeest utilised or not utilised) as a class variable, to determine any
significant variables that could separate the three site types. Discriminant analysis
was used since all the variables measured were continuous and the fact that it has
the ability to identify predictor variables from potentially useful environmental factors
(Marnell 1998). Discriminant function analysis therefore is a multivariate technique
that is particularly useful in habitat use separation studies (Ferrar and Walker 1974).
The step-wise approach enters variables into the discriminant function analysis
individually, and the variable that minimises the overall Wilks’ lambda for the function
is selected for entry at each step. The process is repeated until no additional
increase in the accuracy of the discriminant function was achieved.
It was also decided to analyse the data for black and blue wildebeest separately, and
then to conduct a separate analysis comparing black wildebeest with not utilised
sites, blue wildebeest with not utilised sites and both types of wildebeest combined
with not utilised sites.
160
Standardised canonical discriminant function coefficients and correlations between
discriminating variables and canonical discriminant functions can be used to estimate
the relative contribution of each selected variable to the power of the discriminant
function (Wei et al. 2000). Larger absolute values of correlations or coefficients
indicate stronger contributions to the power of the function for the relevant variables
(Cooley and Lohnes 1971). For example, a value of –2.4 for the standardised
canonical discriminant function coefficient would indicate that that variable had a
stronger contribution to the power of the function than a value of 1.1.
RESULTS
In the present study, the mean and standard error of the 17 herbaceous layer
characteristics for the black and blue wildebeest feeding sites and the not utilised
sites are shown in Table 7.1. The ANOVA tests revealed no significant differences
between the feeding sites utilised by the black and blue wildebeest (Table 7.1).
Differences were found between sites that were not utilised by either type of
wildebeest and those that were utilised by both types of wildebeest (Table 7.1). Oneway analysis of variance detected four variables that differed significantly between
the three types of site (p” 7KH JUDVV KHLJKW DQG JUDVV OHDI KHLJKW YDULDEOHV
were tallest in the not utilised areas, indicating that the feeding sites of both types of
wildebeest tended to be under some grazing pressure that decreased the grass
height to a more preferred grazing level.
The mean grass biomass in the not utilised sites was significantly higher than in the
black wildebeest feeding sites, but not significantly different to the blue wildebeest
feeding sites. Grass biomass concentration followed the same trend as grass
biomass, being significantly higher in the not utilised sites than in black wildebeest
feeding sites, but not significantly different from the sites that were utilised by the
blue wildebeest.
161
Table 7.1: Mean and standard errors of the characteristics of the herbaceous layer of
the feeding sites of the black and blue wildebeest, and sites that were not utilised by
either of them, that were analysed to indicate differences in the feeding sites of the
black and blue wildebeest at Ezemvelo Nature Reserve in April 2004. Bold values
indicate a significant difference and different superscripts denote significant
differences between sites
Variable
Black wildebeest
Blue wildebeest
Not utilised
P-value
Species diversity
Species density
Class 1 (%)
Class 2 (%)
Class 3 (%)
Class 4 (%)
Class 5 (%)
Invaders (%)
Bare ground (%)
Veld condition score
Degree of utilisation
Total grass height (cm)
Grass leaf height (cm)
Canopy cover (%)
Biomass (kg/ha)
3
Biomass concentration (kg/m )
1.8 ± 0.11
1.7 ± 0.17
8.8 ± 5.01
10.2 ± 4.06
0.0 ± 0.00
22.2 ± 6.85
44.6 ± 10.83
2.8 ± 2.79
11.4 ± 4.26
310.9 ± 75.71
49.2 ± 12.71
a
59.5 ± 6.21
a
31.0 ± 3.71
88.5 ± 4.22
a
2193.1 ± 518.33
a
0.7 ± 0.14
1.7 ± 0.14
1.5 ± 0.15
18.4 ± 8.58
3.2 ± 1.57
0.0 ± 0.00
31.5 ± 8.12
36.6 ± 8.71
0.3 ± 0.29
10.0 ± 3.46
369.4 ± 71.29
54.5 ± 7.82
a
66.0 ± 5.85
a
33.6 ± 4.39
89.1 ± 3.2
ab
3469.2 ± 541.17
ab
0.9 ± 0.11
1.9 ± 0.11
2.0 ± 0.15
12.7 ± 3.5
16.6 ± 4.63
0.0 ± 0.00
36.1 ± 6.06
22.1 ± 3.98
7.1 ± 4.48
5.3 ± 1.71
417.1 ± 33.65
29.8 ± 8.99
b
78.9 ± 1.87
b
43.3 ± 1.61
85.1 ± 4.71
b
4605.2 ± 278.75
b
1.1 ± 0.06
0.37
0.08
0.57
0.09
0.44
0.07
0.44
0.26
0.42
0.15
0.007
0.01
0.77
0.002
0.04
162
The feeding sites of the black and blue wildebeest could be discriminated by using
the step-wise disciminant analysis function, based on the percentage of Class 2 plant
species, mean grass biomass (kg/ha), mean grass leaf height and the mean
percentage of invaders present (Table 7.2). This discriminant function analysis of the
two types of wildebeest was significant (Eigenvalue = 1.873, Likelihood ratio = 0.348,
df = 4, p = 0.0138) which suggested that the two types of wildebeest exhibited
different patterns in habitat use in terms of these identified variables. Absolute
standardised coefficients of the selected variables ranged from 0.649 to 1.977, with
the mean grass biomass contributing the most to the power of the discriminant
function and the percentage of invaders contributing the least (Table 7.2).
The sites that were utilised by the black wildebeest and those that were not utilised
by either the black or blue wildebeest were discriminated based on the grass leaf
height and the percentage of Class 4 grasses present (Table 7.2). The discriminant
function analysis of the sites that were utilised by the black wildebeest and those that
were not utilised by either type of wildebeest was significant (Eigenvalue = 5.975,
Likelihood ratio = 0.143, df = 3, p<0.0001) suggesting that there were certain
environmental factors that made a feeding site more suitable for a black wildebeest.
Absolute standardised coefficients of the selected variables ranged from 1.194 to
2.673 (Table 7.2) with the grass leaf height contributing the most to the power of the
discriminant function and the percentage of Class 4 grasses contributing the least.
Sites that were utilised by the blue wildebeest and those that were not utilised by
either type of wildebeest were discriminated based on grass leaf height only (Table
7.2). The discriminant function analysis of the sites that were utilised by the blue
wildebeest and those that were not utilised by either type of wildebeest was
significant (Eigenvalue = 0.224, Likelihood ratio = 0.817, df = 1, p = 0.023).
Since the ANOVA test revealed no statistical differences between the feeding sites of
the black and blue wildebeest (Table 7.1), it was decided to combine all of the sites
utilised by any type of wildebeest by combining the data for black and blue
wildebeest feeding sites, and then to compare them with the not utilised sites. The
sites that were utilised by wildebeest and those that were not utilised by either type of
wildebeest were discriminated based on grass species density, percentage Class 5
grasses and grass biomass (Table 7.2).
163
Class 2 (%)
Biomass (kg/ha)
Grass leaf height (cm)
Invaders (%)
0.806
-1.977
1.556
0.649
Grass leaf height (cm)
Class 4 (%)
164
2.673
1.194
coefficient
coefficient
Standardised canonical
discriminant function
Variable
discriminant function
Standardised canonical
utilised sites
blue wildebeest sites
Variable
Black wildebeest sites compared with not
Black wildebeest sites compared with
Grass leaf height (cm)
Variable
coefficient
1.086
discriminant function
Standardised canonical
not utilised sites
Blue wildebeest sites compared with
coefficient
0.508
-0.514
0.893
discriminant function
Standardised canonical
not utilised sites
Species density
Class 5 (%)
Biomass (kg/ha)
Variable
combined compared with
Black and blue wildebeest sites
and to compare these sites with sites that were not utilised by either type of wildebeest at Ezemvelo Nature Reserve in April 2004
Table 7.2: Results of the discriminant function analysis performed to compare the feeding sites of the black with those of the blue wildebeest,
The discriminant function analysis of the two types of site was significant (Eigenvalue
= 0.7090, Likelihood ratio = 0.585, df = 3, p = 0.0013). Absolute standardised
coefficients of the selected variables ranged from 0.508 to 0.893 with grass biomass
contributing the most to the power of the discriminant function and grass species
density contributing the least (Table 7.2).
Black wildebeest occurred at sites with a lower grass biomass, a higher percentage
of Class 2 grass species, a lower grass leaf height and grass height, with more
invader species, and a lower biomass concentration than the blue wildebeest. Sites
that were selected by both types of wildebeest showed a lower grass biomass, a
higher percentage of Class 5 grass species and a lower grass species density than
sites that were not utilised by either type of wildebeest. Black wildebeest tended to
have a higher impact on their feeding sites, decreasing the grass leaf height and
percentage Class 4 grass species, and increasing the percentage basal cover, while
the blue wildebeest decreased the grass leaf height in their feeding sites compared
to the sites that were not utilised by either type of wildebeest.
DISCUSSION
Black and blue wildebeest showed some discrimination in their feeding site selection
based on the grass biomass and the grass height of the herbaceous layer. Further
feeding site differentiation between the black and blue wildebeest was obtained
through differences in grass species composition. The other variables were not
selected in any of the discriminant analyses as they were either correlated with a
variable that had been selected already or they could not explain additional variation.
Both types of wildebeest in the present study tended to select feeding sites with a
higher percentage of Class 5 grass species than what was found in sites that were
not utilised by either type of wildebeest. Class 5 grass species increase with heavy
over-utilisation (Bothma et al. 2004) and in the study area may indicate that in the
feeding sites of the wildebeest, ecologically better plant species were being replaced
with those that increase with over-utilisation. The grass species density of the feeding
sites that were utilised by wildebeest as feeding sites was also less than that found in
the sites that were not utilised by either type of wildebeest, indicating that wildebeest
may have been selecting some grass species in their feeding sites, hence
decreasing the overall grass diversity. As some areas are being overutilised, such as
in the feeding sites of the wildebeest in the present study, grass species richness will
165
decrease with an accompanying increase in the dominance of a few more
unpalatable species (Morrison et al. 1992). All these factors indicate that the
wildebeest on the study area were overutilising the grass in the areas in which they
were feeding, or that they were modifying these areas to make them more suitable
for their grazing habits which require short grass areas. In order to confirm this
observation, before and after studies of these sites should be conducted.
Black wildebeest feeding sites tended to have a higher percentage of Class 2 grass
species than those of blue wildebeest. Class 2 grass species increase with underutilisation. Black wildebeest feeding sites also had a higher percentage of invader
plant species than those of blue wildebeest. The feeding sites that were utilised by
the black wildebeest had a herbaceous composition made up mainly of 44% Class 5
grass species and a few Class 1 (8%) and Class 2 (10%) grass species. Blue
wildebeest feeding sites had a herbaceous composition consisting of 36% Class 5
grass species, 32% Class 4 grass species and 18% Class 1 grass species.
The removal of the above-ground grass biomass stimulates regrowth that produces
young plant material that is more digestable and nutritious than older plant material.
Repeated grazing during the growing season therefore increases the quality of the
forage (Ydenberg and Prins 1981). The creation and maintenance of grazing sites by
black wildebeest in particular, improves the quality and digestability of such areas
(Augustine et al. 2003). These grazing areas are expanses of short grass in an
immature state, have grasses with higher stem:leaf ratios, and a higher bulk density
than that of tall stands (Cromsigt 2006). The higher bulk density means a potentially
higher food yield per bite (McNaughton 1984). The results of the present study,
however, indicated that the grazing sites that were utilised by the black wildebeest at
Ezemvelo Nature Reserve had a lower bulk density than the taller stands that were
not utilised. This may indicate that the grasses that were utilised by the black
wildebeest were not the type of grasses that would react to grazing to produce a
traditional grazing lawn, but would instead decline until the patch became denuded
and the wildebeest were forced onto another area.
The percentage canopy cover has been used to provide a rough indication of the
quantity of forage available in a given area (Novellie and Strydom 1987). The
phytomass available is important in determining the feeding habits and habitat
utilisation of herbivores (Kinyamario and Macharia 1992). Annual consumption of
plant material by large herbivores may be limited by the amount of herbage available.
166
A grass height of less than 30 mm would be the minimum grazing height for most
grazing ungulates (Dörgeloh 1998). Grass heights <30 mm would not provide
sufficient herbage for the maintenance of a healthy body weight. Some large
herbivore species select grass swards that are dominated by Class 1 grasses (with a
consequent high veld condition score) while other species favour a grass sward
structure that is dominated by Classes 2 to 5 grasses, with a consequent low veld
condition score (Novellie 1990).
On the scale of a feeding site, studies have shown that variation in the size, spatial
detail and quality of these food patches influence the selectivity of grazer species
differentially and could potentially determine large herbivore coexistence and
diversity at such a small scale (Hester et al. 1999). However, the majority of
herbaceous layer characteristics at the feeding sites did not differ between the two
types of wildebeest or between those sites utilised by wildebeest species and those
not utilised by them. The overall pattern at Ezemvelo Nature Reserve indicated that
some degree of species-specific difference did exist between the black and blue
wildebeest with regards to the grass phytomass levels, grass sward structure and
grass species composition of the feeding sites. These differences were, however, not
considered to be large enough to allow for the coexistence of the black and blue
wildebeest at this fine scale. Therefore, it is suspected that the black and blue
wildebeest in the study area did not partition the food resources at the feeding site
scale, and if confined to areas with no habitat variation, they may compete for
feeding sites. However, without a detailed floristic analysis to the species level, these
results remain inconclusive.
CONCLUSION
The results of the present study indicated that the feeding sites of the black and blue
wildebeest were only discriminated based on differences in grass quantity and grass
species composition. These differences did not prove to be significantly different
when classical hypothesis testing was applied. A greater difference was however
detected between the black wildebeest feeding sites and those sites not utilised by
either type of wildebeest. These differences were based on grass structure and grass
quantity.
167
CHAPTER 8: ACTIVITY BUDGETS
INTRODUCTION
The daily activity patterns of ungulates are affected by environmental factors such as
temperature, cloud cover, wind velocity, moon cycle, as well as the presence of other
animals (Berry et al. 1984; Theron 1991; Vrahimis and Kok 1992; Vrahimis and Kok
1993). Activity patterns also tend to vary between species experiencing the same
environmental factors in the same area, indicating a compromise to a number of
factors that act concurrently on the animals (Leuthold 1977). Disparity in the activity
patterns occurring between species inhabiting the same area may therefore, mirror
their individual physiological adaptations to the prevailing environmental conditions
(Ben-Shahar and Fairall 1987).
Animals may also make behavioural adjustments to their natural activity patterns in
response to competition from other species in the near vicinity (Pianka 1973).
Differences detected between two species in terms of their relative activity patterns
may allow for the exploitation of different resources at different times. Such temporal
separation of activities may reduce the extent of competition between these two
species. Therefore, subtle temporal differences in daily and seasonal activity patterns
may allow for coexistence (Pianka 1973).
By quantifying and comparing the activity patterns of the black and blue wildebeest at
Ezemvelo Nature Reserve, differences in the requirements of the two types of
wildebeest may be evaluated (Ben-Shahar and Fairall 1987). This comparison and
quantification would also provide information on the behavioural adjustments made
by each type of wildebeest, which may be in response to competition or due to the
suitability of their environment.
It has been suggested that the thermal tolerance of the black wildebeest is high due
to their body conformation, dark pelage and thick coat adapting them to an open
habitat with no shade where they are exposed to the sun throughout the day
(Vrahimis and Kok 1992; Skinner and Chimimba 2005). In contrast, other studies
have shown that the thermal tolerance of the blue wildebeest is lower than that of the
black wildebeest due to their paler pelage and thinner coat (Hofmeyer 1981; BenShahar and Fairall 1987) resulting in shade-seeking behaviour and a concentration of
168
active periods in the early and later parts of the day when temperatures are low (BenShahar and Fairall 1987).
The two types of wildebeest at Ezemvelo Nature Reserve occupy an area outside
their historical distribution range, which has sub-optimal habitat for wildebeest, with a
long history of human activity and an absence of large natural predators (Chapter 2).
These factors may all require some compensatory behaviour by both types of
wildebeest in order to survive and reproduce effectively. Combined with the influence
that competition may have on activity patterns, it is expected that the black and blue
wildebeest activity patterns will be dissimilar. This hypothesis was tested here by the
following key questions:
•
What are the daytime activity patterns of the two types of wildebeest over the
entire study period and over the three ecological seasons?
•
What are the different activity patterns adopted by the various social groups
of each type of wildebeest?
METHODS
Field collection of data
Observations on the daily activity patterns of the black and blue wildebeest at
Ezemvelo Nature Reserve were made from March 2004 to August 2005. Direct field
observations were conducted on a monthly basis from a parked vehicle or other
vantage point by using a pair of 16 x 50 binoculars during daylight hours. Most of the
observations were done at distances of 100 – 500 m. During each observation
period, the dominant activity of each individual visible from the observation point was
recorded at 5-minute intervals (Grimsdell and Field 1976) by using the scan-sampling
method (Altmann 1973). Activities observed were classified into five categories,
namely grazing, standing, walking, lying down and other activities (Ben-Shahar and
Fairall 1987). The latter included all activities that did not feature strongly in the
general activity pattern, such as grooming, running, defaecating and urinating.
Following Von Richter (1971a) and Berry (1980) three basic social groups were
recognized for both types of wildebeest, namely breeding herds consisting of females
and their calves, bachelor herds consisting of non-breeding males and territorial
males. During most sampling sessions it was attempted to keep a female herd with a
territorial bull under continuous observation. It was considered that the female herd
169
would be the most representative group to indicate patterns in daily activity of the two
types of wildebeest and it also provided for more activity records (Winterbach 1999).
Black wildebeest territorial bulls were closely associated with the female herds but
blue wildebeest territorial bulls tended to occur around the edges of breeding herds.
Thus for the blue wildebeest, a number of nearby territorial bulls could be observed
concurrently with the herd. For comparative purposes, activity records for calves,
female adults and territorial bulls were recorded separately.
If the presence of the observer caused the herd at any time to appear uneasy for
more than 15 minutes, or it ran off for more than 100 m, observations on that herd
were discontinued on the assumption that normal activities were interfered with
(Vrahimis and Kok 1993; Winterbach 1999). As the number of individuals per
observation varied, even between consecutive observations, all the observations
were standardised to percentages before analysis to remove the effect of group size
(Winterbach 1999).
In addition, the cloud cover, wind speed and wind direction were estimated every two
hours. Temperature was recorded in the shade outside the car every 15 minutes.
Statistical analysis of the data
Observations were distributed as equally as possible over the different age and sex
classes, thus minimising bias towards observation of the more conspicuous
individuals. Imbalances within the data set and potential serial correlations between
observations would have severely restricted the options for testing the influence of
biological and physical factors on activity patterns (Groeneveld 2006 pers. comm.)12.
Nonparametric tests were therefore applied to some subsets of the data, but use of
the complete data set in a multivariate analysis was not feasible. Frequencies of
each activity were calculated by dividing the number of observations by the total
observations in each hour. For analysis these percentages were log transformed.
The Kruskal-Wallis test (Zar 1984) was used to test the hypothesis that there was no
difference between the times spent per activity between black and blue wildebeest.
Analyses were performed on various subsets and groupings of the overall data set.
Each of the five activities for the black and blue wildebeest was compared. Four of
12
Prof. H. Groeneveld. Department of Statistics, University of Pretoria, Pretoria, 0002, South
Africa.
170
the five activities were further grouped into two categories namely active (feeding and
walking) and inactive (standing and lying down) and compared between the two
types of wildebeest (the “other” category was left out as it included both active and
inactive activities). Seasonal data were analysed separately, as were social groups,
while the day was divided into three time periods (<10:00; •10:00-14:00; and >14:00)
and each time period was analysed separately. Due to the nature of the data,
statistical analysis of the daily activity pattern was not feasible.
The generalised linear model procedure (PROC GLM) utilising a number of ANOVA
tests was performed to test the null hypothesis that no differences occurred among
the seasons for each type of activity, followed by calculation of the Least Square
Means to determine categories which were significantly different or not (Zar 1984).
RESULTS
A total of 405 activity hours were recorded for black and blue wildebeest at Ezemvelo
Nature Reserve. In total, 198 activity hours were recorded for the black wildebeest,
92 hours during the late growing season, 44 hours during the dormant season and 62
hours during the early growing season. A total of 207 activity hours were recorded for
the blue wildebeest, with 65, 62 and 80 hours during the late growing season, the
dormant season and the early growing season respectively.
Entire study period daily time budget (all data)
The relative proportion of daily activities as shown by the black and blue wildebeest
over the entire study period, combining all age classes and social groups is
summarized in Figure 8.1.
The largest part of the day for black wildebeest was spent grazing (35.4%) followed
closely by lying down (32.6%). Standing, often not considered a dominant activity
(Vrahimis and Kok 1993), formed a substantial portion (26.9%) of the daily time
budget. Walking and other activities combined only formed 5.1% of the daily time
budget.
Blue wildebeest spent most of their daily time budget by grazing (44.5%). Equal time
was spent standing (23.3%) and lying down (23.1%) during the day. Walking, which
171
27%
35%
Other
Walking
Lying down
Standing
Grazing
Activity
23%
23%
7%
2%
Blue wildebeest
45%
all the social groups and age classes.
172
Figure 8.1: Daily time budgets for the black and blue wildebeest for the entire study period at Ezemvelo Nature Reserve when combining
33%
3%
2%
Black wildebeest
is often not considered a dominant activity formed 7.0% of the daily time budget and
other activities represented a minor proportion (2.1%).
According to the Kruskall Wallis test, blue wildebeest spent a significantly greater
percentage of their time grazing than did black wildebeest (Ø2 = 7.6464; df = 1; p =
0.0057). Blue wildebeest also spent a significantly greater percentage of their time
walking than did black wildebeest (Ø2 = 12.8569; df = 1; p = 0.0003).
Figure 8.2 shows the relative proportion of the daily time spent on the various
activities by territorial bulls, adult females and calves throughout the study period for
black and blue wildebeest. Blue wildebeest adult females spent a significantly higher
proportion of their time grazing (46.3%) than did black wildebeest adult females
(37.2%) (Ø2 = 7.6863; df = 1; p = 0.0058). However, black wildebeest adult females
spent significantly more time lying down (31.1%) and standing (27.3%) than did blue
wildebeest adult females (22.3% and 22.6% respectively) (Ø2 = 4.4625; df = 1; p =
0.0346 and Ø2 = 5.4913; df = 1; p = 0.0191). Blue wildebeest adult females in turn
spent significantly more time walking (7.3%) than did black wildebeest adult females
(3.1%) (Ø2 = 4.4625; df = 1; p = 0.0346).
Black wildebeest calves spent most of their daily time budgets by lying down (46.0%)
while blue wildebeest calves spent most of their time in grazing (40.4%). Blue
wildebeest calves spent more of their daily time walking than did black wildebeest
calves (7.4% vs 3.2%) (Ø2 = 7.3237, df = 1; p = 0.0068), but this was the only
significant difference that could be detected between the activities of the calves of the
two types of wildebeest.
Black wildebeest territorial bulls spent most of their time standing (42.3%) while blue
wildebeest territorial bulls spent most of their time grazing (43.5%). The only
significant difference between the activities of the territorial bulls was that blue
wildebeest bulls spent more time grazing (43.5%) than did the black wildebeest bulls
(28.3%) (Ø2 = 4.0585; df = 1; p = 0.0439).
173
Blue wildebeest
Black wildebeest
2%
Adult females
3%
37%
31%
7%
2%
G raz ing
S tanding
22%
46%
Ly ing down
W alk ing
O ther
23%
27%
1%
Calves
3%
7%
32%
2%
G raz ing
40%
S tanding
Ly ing down
46%
30%
W alk ing
O ther
18%
21%
Territorial bulls
4% 2%
6% 2%
28%
G raz ing
24%
S tanding
19%
43%
Ly ing down
W alk ing
O ther
30%
42%
Figure 8.2: Daily time budgets (percentage of time spent) for adult females, calves
and territorial bulls of the black and blue wildebeest for the entire study period at
Ezemvelo Nature Reserve.
174
Seasonal daily time budgets
There were not sufficient replicates for an adequate seasonal analysis. Most of the
tests conducted indicated non-significant differences between the black and blue
wildebeest. This could probably be attributed to the low number of replicates per
season. Therefore the results for the seasonal analysis had to be interpreted with
caution. Age or sex comparisons were not considered in this section due to the low
number of samples available for analysis. Figure 8.3 shows the percentage of time
spent in conducting each type of activity in each of the three ecological seasons.
Late growing season
Blue wildebeest spent significantly more time in grazing (40.3%) than black
wildebeest (29.9%) during the late growing season (Ø2 = 4.8348; df = 1; p = 0.0279)
and they also spent significantly more time in walking (6.5%) than the black
wildebeest (2.5%) (Ø2 = 6.4821; df = 1; p = 0.0109). Black wildebeest, however, spent
significantly more time lying down (39.1%) during this season than the blue
wildebeest (24.1%) (Ø2 = 4.8348; df = 1; p = 0.0279).
Dormant season
During the dormant season the proportional allocation of time to different activities
did not differ significantly between the black and blue wildebeest (Figure 8.3).
175
60
P ercentage
50
**
*
40
*
30
20
**
**
10
0
G
S
L
W
O
L a te g ro win g se a so n
G
S
L
W
O
D o rma n t se a so n
G
S
L
W
Ea rly g ro win g se a so n
S eason and activity
B la ck wild e b e e st
B lue wild e b e e st
Figure 8.3: Seasonal daily time budgets (percentage time spent) for black and blue
wildebeest at Ezemvelo Nature Reserve from January 2004 to August 2005.
G=grazing, S=standing, L= Lying down, W=walking, O=other. * indicates a significant
difference between the black and blue wildebeest within that season for that activity.
176
O
Early growing season
The blue wildebeest spent significantly more time in walking (6.77%) than the black
wildebeest (4.6%) during the early growing season (Ø2 = 3.6923; df = 1; p = 0.0547).
There was no difference in the time allocation patterns of the other activities between
the black and blue wildebeest.
Between season comparisons
The general linear model procedure (PROC GLM) indicated that black wildebeest
spent significantly less time in grazing during the late growing season than during the
dormant season (p = 0.0010) and the early growing season (p = 0.0014). There were
no seasonal differences in terms of the percentage time spent standing for the black
wildebeest. Black wildebeest spent significantly more time lying down during the late
growing season than during the dormant season (p = 0.0194) and they also spent
less time walking during the late growing season than during the early growing
season (p = 0.0214). No seasonal differences were found in terms of percentage
time spent conducting other activities by the black wildebeest.
The PROC GLM procedure that was done indicated that there were no seasonal
differences in terms of the percentage of time spent in grazing by the blue
wildebeest. Blue wildebeest spent significantly more time standing during the late
growing season than during the early growing season (p = 0.0179) and also more
time standing during the dormant season than during the early growing season (p =
0.0184).
No significant seasonal differences were found in terms of the percentage of time
spent lying down and walking in the blue wildebeest. However, the blue wildebeest
spent more time conducting “other” activities during the late growing season than
during the early growing season (p = 0.0012) and also more time in conducting
“other” activities during the dormant season than during the early growing season (p
= 0.0050).
Entire study period: diurnal behavioural patterns (all data)
Figure 8.4 illustrates the overall diurnal activity budget for the black and blue
wildebeest throughout the study period. Most of the lying down by the black
wildebeest occurred just after midday and continued until approximately 15:30 in the
afternoon.
177
B lack w ildebeest
100
O ther
P e rce nta ge
80
W alk ing
60
Ly ing down
40
S tanding
20
G raz ing
0
:0
0
17
16
:0
0
:0
0
15
:0
0
14
:0
0
13
12
:0
0
:0
0
11
:0
10
00
9:
00
8:
00
7:
6:
00
0
Tim e
B lue w ildebeest
100
O ther
P e rce nta ge
80
W alk ing
60
Ly ing down
40
S tanding
20
G raz ing
0
17
:0
0
16
:0
0
15
:0
0
14
:0
0
13
:0
0
12
:0
0
11
:0
0
:0
10
00
9:
00
8:
00
7:
6:
00
0
Tim e
Figure 8.4: Diurnal activity patterns (percentage of time spent) by the black and blue
wildebeest for the entire study period at Ezemvelo Nature Reserve.
178
The main grazing periods when more than 40% of a herd was found grazing were in
the morning from sunrise until about 07:30 and again in the afternoon from
approximately 15:30 until shortly after sunset.
Another smaller peak in grazing occurred at midday, but was not as marked as the
other two peaks. The lowest incidence of grazing occurred from 13:00 to 14:00, as it
was the main diurnal resting period. Most walking occurred in the early morning, with
a general movement to the daytime resting place and in the late afternoon with a
general movement to the night-time resting place. No movement to water was
observed during the day. The highest incidence of standing was associated with the
period before the daytime resting period (09:00 to 12:00).
Most lying down in blue wildebeest occurred from 14:00 to 16:00 in the afternoon and
a smaller peak in lying down from 10:00 to 12:00. Three main peaks in grazing were
observed. The highest peak in grazing occurred in the morning from 06:00 to 09:00.
Another peak of grazing occurred from 12:00 to 14:00 and the third peak from 16:00
till sunset. Wildebeest tended to stand more than lie down from 09:00 to 12:00.
Walking activity was relatively evenly distributed throughout the day with a slight drop
off towards the afternoon. Most walking activity in the morning was associated with a
movement to water fro drinking.
Figure 8.5 illustrates the results of dividing the daytime into three equal periods
<10:00, 10:00 to 14:00, >14:00. The Kruskal Wallis Test indicated that blue
wildebeest spent significantly more time grazing in the mornings (<10:00) than did
black wildebeest (Ø2 = 5.0370; df = 1; p = 0.0248). Black wildebeest spent
significantly more time standing in the mornings than did blue wildebeest. Blue
wildebeest spent significantly more time walking during midday (10:00 to 14:00) and
the afternoon (>14:00) than the black wildebeest (Ø2 = 3.8991; df = 1; p = 0.0483 and
Ø2 = 6.2267; df = 1; p = 0.0126 respectively). No other significant activity differences
between the black and blue wildebeest were found.
Seasonal analysis of diurnal behavioural patterns
The diurnal activity patterns that were recorded for the black and blue wildebeest at
Ezemvelo Nature Reserve during the three ecological seasons are illustrated in
Figures 8.6 to 8.8.
179
60
*
*
Pe rce ntage
50
*
40
30
*
*
20
*
10
*
0
G
S
L
W
O
G
< 10:00
S
L
W
O
10:00-14:00
G
S
L
W
O
> 14:00
T ime of day and activ ity
B lac k wild e b e e s t
B lue wild e b e e s t
Figure 8.5: Comparison of the diurnal behavioural patterns of the black and blue
wildebeest expressed as a percentage of the time spent for the three time periods in
the daytime for the entire study period at Ezemvelo Nature Reserve. * indicates a
significant difference between the black and blue wildebeest for that activity and time
RIGD\ZLWK
.
*
*UD]LQJ6
6WDQGLQJ/
Other.
180
/\LQJGRZQ:
:DONLQJ2
Black wildebeest
100
O ther
Pe rce ntage
80
W alking
60
Lying dow n
40
S tanding
20
Grazing
:0 0
17
:0 0
:0 0
16
:0 0
15
14
:0 0
13
:0 0
12
:0 0
11
10
:0 0
0
0
9 :0
0
8 :0
7 :0
6 :0
0
0
T ime
B lue w ildebeest
Pe rce n tag e
100
O ther
80
W alking
60
Lying down
40
Standing
20
G razing
0
17
:0
0
16
:0
0
15
:0
0
14
:0
0
13
:0
0
12
:0
0
11
:0
0
:0
10
00
9:
00
8:
00
7:
6:
00
0
Time
Figure 8.6: Diurnal activity patterns expressed as a percentage of time spent, of the
black and blue wildebeest for the late growing season at Ezemvelo Nature Reserve.
181
100
90
80
70
60
50
40
30
20
10
0
O ther
W alk ing
Ly ing dow n
S tanding
0
:0
0
17
:0
0
16
:0
0
15
:0
0
14
:0
0
13
:0
0
12
:0
0
11
:0
10
00
9:
00
8:
00
G raz ing
7:
6:
00
P e rce nta ge
B lack w ild eb eest
Tim e
B lu e w ild eb eest
P e rce n ta g e
100
80
O ther
60
W alk ing
Ly ing dow n
40
S tanding
20
G raz ing
0
17
:0
0
:0
0
16
15
:0
0
14
:0
0
:0
0
13
12
:0
0
11
:0
0
:0
00
10
9:
00
8:
00
7:
6:
00
0
T im e
Figure 8.7: Diurnal activity patterns, expressed as a percentage of time spent, of the
black and blue wildebeest for the dormant season at Ezemvelo Nature Reserve.
182
B lack w ildebeest
100
O ther
Pe rcenta ge
80
W alk ing
60
Ly ing down
40
S tanding
20
G raz ing
:00
:00
17
16
:00
15
:00
14
:00
13
:00
12
:00
11
10
:00
0
0
9:0
0
8:0
7:0
6:0
0
0
Tim e
B lue w ildebeest
100
O ther
P e rce nta ge
80
W alk ing
60
Ly ing down
40
S tanding
20
G raz ing
0
17
:0
0
16
:0
0
15
:0
0
14
:0
0
13
:0
0
12
:0
0
11
:0
0
:0
10
00
9:
00
8:
00
7:
6:
00
0
Tim e
Figure 8.8: Diurnal activity patterns, expressed as a percentage of time spent, of the
black and blue wildebeest for the early growing season at Ezemvelo Nature Reserve.
183
During the late growing season the black wildebeest had two main peaks of grazing
when more than 40% of the herd was found grazing. These occurred from 06:00 to
08:00 and again in the evening from 16:00 to 18:00.
Another less intense grazing peak occurred from 10:00 to 13:00. During the dormant
season no small grazing peak occurred at midday and most grazing was
concentrated in the hour before sunset. During the early growing season there are
three main grazing peaks through the diurnal period where more than 40% of the
herd was found grazing. The first was from 06:00 to 08:00, the second from 12:00 to
13:00 and the last from 15:00 to 17:00.
During the late growing season one marked resting period occurred in the day where
more than 40% of the herd was lying down. This occurred from 13:00 to 15:00 in the
afternoon. Another less distinct grazing period occurred from 08:00 to 12:00. During
the dormant season only one marked peak in resting activity occurred from 12:00 to
16:00 where more than 40% of the animals were lying down. Lying down was less
than 10% for the rest of the day during this season. During the early growing season
no major peaks in resting activity occurred and lying down featured evenly
throughout the day accept for from 06:00 to 08:00 when few animals were lying
down.
In the late growing and early growing seasons, walking was concentrated in two main
bouts, one in the early morning and another in the late afternoon just before sunset.
This pattern was not distinct during the dormant season. No drinking behaviour was
observed by black wildebeest during any of the activity budget surveys, indicating
that drinking must be restricted to night time in the study area.
During the late growing, season the blue wildebeest showed three main peaks of
grazing activity when more than 40% of the herd was grazing. The first peak was
from 06:00 to 09:00, the second from 12:00 to 14:00 and a smaller peak from 16:00
to 18:00. Three main but slightly longer, grazing peaks were also found in the
dormant season. The first occurred from 07:00 to 10:00, the second from 11:00 to
14:00 and the last from 16:00 to 18:00. The early growing season showed the same
pattern as that of the late growing season with three main grazing peaks from 06:00
to 09:00, from 11:00 to 13:00 and another from 16:00 to 18:00.
184
In the late growing season, the blue wildebeest were found to spend more time
walking in the mornings than the black wildebeest (Ø2 = 4.44; df = 1; p = 0.0350).
Blue wildebeest spent significantly more time in grazing and walking during midday
than the black wildebeest (Ø2 = 3.333; df = 1; p = 0.0679 and Ø = 4.0333; df = 1; p =
0.0446 respectively) during the late growing season. During the dormant season, the
black wildebeest spent significantly more time standing in the mornings than the blue
wildebeest (Ø2 = 3.1527; df = 1; p = 0.0758). Blue wildebeest spent significantly more
time walking in the afternoons than black wildebeest during the dormant season (Ø2 =
4.0833; df = 1; p = 0.0433). During the early growing season the only significant
difference indicated by the Kruskal Wallis test was that black wildebeest spent
significantly more time standing in the mornings than the blue wildebeest (Ø2 = 4.800;
df = 1; p = 0.0285).
Periods of activity and rest
Blue wildebeest were found to be significantly more active than the black wildebeest
throughout the entire study period (Ø2 = 11.1727; df = 1; p = 0.0008) (Figure 8.8).
Blue wildebeest spent 53% of their daily time being active while black wildebeest
were only active for 43% of their overall daily time. Blue wildebeest were also found
to be significantly more active than the black wildebeest during the late growing
season and the early growing season (Ø2 = 6.4821; df = 1; p = 0.0109 and Ø2 =
3.6923; df = 1; p = 0.0547 respectively). No significant difference between the black
and blue wildebeest in terms of time spent active was found for the dormant season.
DISCUSSION
Like the blue wildebeest, the black wildebeest was a migratory animal that occurred
in large herds when totally wild (Von Richter 1971b). The black wildebeest has never
been studied in its natural habitat while interacting with its natural predators. This
opportunity is lost in South Africa as migration is no longer possible, and probably will
never be again. Therefore, the social organisation of both types of wildebeest in most
parts of South Africa (blue wildebeest in the Kalahari still migrate from time to time)
reflects a permanently sedentary phase, consisting of a pattern of permanently
established territories, with separate and small (in relation to the migratory herd)
female herds and segregated bachelor herds (Jarman 1974). A single male defends
a territory. A central trampled and heavily grazed core area of use occurs in each
territory, and it is associated with much dung deposition (Estes 1969).
185
Percentage time spent
80
70
60
50
40
30
20
10
0
*
*
*
Active
Inactive
Late growing
season
Active
Inactive
Dormant season
*
Active
Inactive
*
*
Active
Early growing
season
Inactive
Overall
Season and type of activity
Black wildebeest
Blue wildebeest
Figure 8.9: Periods of activity (grazing and walking) and rest (lying down and
standing), expressed as percentage of time spent, for the black and blue wildebeest
for the entire study period and for each ecological season at Ezemvelo Nature
Reserve. * Indicates significant differences between the black and blue wildebeest
within the seasons and periods of activity.
186
Territorial behaviour in the blue wildebeest was studied by Estes (1969) and in the
black wildebeest by Von Richter (1971a). Their studies and others indicate that the
blue wildebeest may have a much more fluid breeding behaviour than the black
wildebeest. This may, however, be because most blue wildebeest studies have been
conducted on large migratory herds in eastern Africa. Studies on sedentary
populations are few. The last study was conducted by Knight (1991) on a blue
wildebeest population in the Kalahari, which still has the ability of migrating from time
to time.
The activity patterns of the black wildebeest have been studied in detail by Vrahimis
and Kok (1993). Direct observations were conducted on a monthly basis from sunrise
to sunset in both the dry and the wet season. Black wildebeest were found to spend
most of the day lying down, followed by grazing (Vrahimis and Kok 1993). Territorial
and bachelor males spent more time in grazing, standing and performing other
activities but less time in lying down than the females.
The time that was devoted to grazing by the blue wildebeest as it was studied by
Berry et al. (1982) was 33%. The predicted foraging time in relation to body mass is
28% (Owen-Smith 1982). Both black and blue wildebeest foraging time at Ezemvelo
Nature Reserve was greater than this percentage. The lower foraging time for the
black wildebeest (30%) compared with the blue wildebeest (40%) could be as a
result of the smaller stomach size of the black wildebeest, resulting in more time
spent in ruminating.
The physiological limitations of the blue wildebeest result in the effective use of the
woodland areas within the study area where shade is available (Ben-Shahar and
Fairall 1987), while the black wildebeest can survive on open grass plains with no
shade. Ambient temperature can be related to changes in activity (Ben-Shahar and
Fairall 1987). In the present study, the black wildebeest seemed to be more inactive
than blue wildebeest during the day and spent equal proportions of their time in
grazing and lying down while ruminating. The blue wildebeest spent more time in
grazing than in lying down. While the blue wildebeest responds to environmental
pressure in the form of heat stress (Ben-Shahar and Fairall 1987) the black
wildebeest does not. The thicker, darker coat of the black wildebeest enables it to
tolerate greater heat stress than is allowed by the thinner pelage of the blue
wildebeest.
187
The major differences between the two types of wildebeest as was found in the
present study relate to the amount of time spent in grazing, and the tendency of the
blue wildebeest to be more active than the black wildebeest. While the blue
wildebeest seemed to respond to environmental pressure in the form of heat stress,
the black wildebeest did not.
In both types of wildebeest, the mean time spent feeding per day was longer in the
adult females than in the calves. Calves of both types of wildebeest spent more time
lying down than the adults. Adult females spent more time grazing than territorial
bulls for both types of wildebeest, but the differences were more marked for black
wildebeest than for blue wildebeest. Black wildebeest tended to be more territorial
than blue wildebeest and hence it would be expected that black wildebeest territorial
bulls would spend less time feeding and more time standing and viewing their
territories (Vrahimis and Kok 1993). The results of the present study suggest sexrelated differences in either the mode of food-gathering and processing and/or in
food and nutrient requirements (Leuthold and Leuthold 1978).
It is also possible that the differences found in feeding time between the black and
blue wildebeest reflected the lack of nocturnal observations. Compensation for any
feeding deficit incurred during the day could have been made during the night
(Leuthold and Leuthold 1978).
The majority of animals preferred to lie down rather than stand during the heat of
midday in the present study as was also observed by Vrahimis and Kok (1993). This
could be a result of the wildebeest attempting to reduce the impact of reflected
radiation from the ground (Jarman 1977). Berry et al. (1984) and Vrahimis and Kok
(1992) found that body orientation was related to sun and wind direction in both the
black and blue wildebeest.
Connochaetes species have precocial young which show a well-developed following
response and the calves accompany their mothers from the moment when they first
gain their feet, and are able to run within minutes of birth (Estes 1966; Von Richter
1971a). Calving is also strictly seasonal and highly synchronised, with the bulk of the
young being dropped within a 3-week period (Estes 1966; Von Richter 1971a;
Skinner and Chimimba 2005). Vrahimis and Kok (1994) studied the diurnal activity of
early post-natal black wildebeest calves and found that they spent most of their time
lying down. The results of the present study agreed with these observations for the
188
black wildebeest. However, the blue wildebeest calves spent more time feeding than
lying down. These differences may be related to an innate means of predator
avoidance in the black wildebeest that tends to occur in more open habitats than the
blue wildebeest (Vrahimis and Kok 1994). The tawny coat of the young calf of both
types of wildebeest is completely different from that of the older animals, thereby
improving its concealment in its natural environment (Estes 1974).
The black wildebeest in the present study spent less time lying down (33%) than the
black wildebeest that were studied by Vrahimis and Kok (1993) (40%). The black
wildebeest in the present study also spent more time standing (27%) than in the
study of Vrahimis and Kok (op. cit.) (12%). Less time was also spent grazing (35%) in
the present study than in the study of Vrahimis and Kok (op. cit) (40%).
Interspecific differences could be an inherent part of each species’ behaviour but
they may also be linked to the climatic factors prevailing in different areas of study
(Vrahimis and Kok 1993). An increase in static activities such as lying down and
standing usually appears under conditions of high heat load (Leuthold 1977; Berry et
al. 1982).
According to Owen-Smith (1982), foraging time tends to increase with increasing
body mass in large herbivores, but factors such as the location of the study area, the
availability of grazing and the foraging behaviour of the different species may all play
a vital role in the amount of time spent grazing (Vrahimis and Kok 1993). The time
spent in standing by ungulates may be influenced by a variety of circumstances,
including weather conditions (Leuthold 1977) and external disturbances. Territorial
bulls show a long time spent in standing. When standing, territorial males have an
improved view of the surrounding terrain, making it easier to spot potential intruders
or possible mating partners (Vrahimis and Kok 1993).
The small proportion of the time that was spent in walking by the black wildebeest in
the present study can partly be attributed to the pronounced tendency of black
wildebeest to remain in their concentration areas (Vrahimis and Kok 1993).
Blue wildebeest tended to spend more time grazing than black wildebeest. This may
be attributed to black wildebeest concentrating in areas where the grass layer has
been modified to such an extent that its standing biomass is less and thus the black
wildebeest do not need to graze as much to obtain the same nutritional intake. Also
189
the smaller body size may require less food intake. The open environment may
require more time to be spent lying down to reduce heat gain than in the more
shaded habitats of the blue wildebeest.
Seasonal differences indicate that the time spent grazing is greater in the dormant
season when the grasses are dormant and low in nutritional value. More time is
spent lying down in the hotter seasons than in the cooler seasons. More time is spent
walking in the early growing season than the late growing season and dormant
season.
CONCLUSION
The activity patterns of the black and blue wildebeest at Ezemvelo Nature Reserve
differed significantly in terms of the time spent being active and resting. This has
been attributed to the differences in size between the two types of wildebeest and the
area selectivity of the black wildebeest that did not require much moving about.
Movement of the black wildebeest during the night requires further study. The activity
patterns differed from the same type of wildebeest as found in other study areas
where similar categories were used. This could be attributed to human disturbance in
the study area affecting the activity patterns to a certain degree. It could also be
attributed to differences in the quality of the vegetation available to the two types of
wildebeest at Ezemvelo Nature Reserve.
190
CHAPTER 9: NICHE BREADTH, OVERLAP AND EXPLOITATIVE INTERSPECIFIC
COMPETITION
INTRODUCTION
The niche, as defined according to the Hutchinsonian concept (Hutchinson 1957), is
accepted as the region in n-dimensional space where the fitness of an individual of a
species is said to be positive. The utilisation of multiple resources allows for resource
partitioning between the species occupying a specific area. Resource partitioning
would in turn, result in niche differentiation and therefore the coexistence of a number
of different species in an area would be facilitated (Schoener 1974b).
Resource partitioning patterns in ungulate communities have been extensively
studied (Gordon and Illius 1989; Voeten and Prins 1999; Forsyth 2000; Johnson et
al. 2000; Bagchi et al. 2003; Hemami et al. 2004; Namgail et al. 2004), but the
mechanisms that produce such partitioning remain poorly understood.
The observed pattern of resource partitioning between two co-evolved species could
be ascribed to ecological forces (i.e. inherent ecological requirements, current
competition and / or predation) and evolutionary history (Connell 1980). The pattern
can, however, only be tested and interpreted in terms of current ecological forces as
the past processes cannot be determined in co-evolved ecosystems (Namgail et al.
2004). This has elicited much criticism of studies that have concluded that
competition was a main mechanism of resource partitioning (Forsyth 2000).
Animals partition resources in three fundamental ways: temporally, spatially and
trophically (Pianka 1973). They may differ in the times when they are active, the
places which they exploit and even the foods which they eat. Such differences,
separate niches, reduce competition and presumably allow for the coexistence of a
variety of species in one area. The various niche dimensions along which partitioning
may take place include habitat, diet, temporal activity and spatial distribution. Habitat
is the most common niche dimension to be partitioned, followed closely by food
resources (Hemami et al. 2004). Temporal partitioning becomes important in
environments where resources are renewed rapidly (Bagchi et al. 2003).
In studies of species interactions it is useful to quantify the degree to which two
species overlap in their use of space, habitat or other resources (Hurlbert 1978).
191
Overlap indices measure the similarity of two species’ use of resources (Loman
1986). Many studies have cited overlap in the use of resources as evidence for
competition between ungulates (e.g. Gordon and Illius 1989). However, interpreting
these studies is problematic for a number of reasons, the most important being that a
high degree of overlap will result in competition only when that resource becomes
limiting (Forsyth 2000). The role of competition in ungulate communities has been the
subject of considerable debate (Putman 1996). The relative lack of evidence for the
effects of interspecific competition among ungulates arises largely from the difficulty
of conducting replicated manipulations of density in the field (Caughley and Sinclair
1994). Hence, interspecific competition is difficult to prove (Putman 1996).
Interspecific competition between two species is possible only when three separate
conditions are met (Putman 1996; Traill 2004):
•
There is habitat overlap
•
There is overlap in forage consumed by the two species within those shared
habitats
•
The shared dietary resources are limiting.
Therefore, to prove that interspecific competition is occurring a reduction of fitness of
one of the competitors needs to be demonstrated with the above conditions being
met.
To demonstrate that competition is occurring, a reduction in fitness of one of the
competitiors needs to be found, and mere overlap may not indicate the presence of
competition.
The seasonal context is also extremely important since it is essential to understand
resource partitioning between two species during the critical season when resources
are limiting (Riney 1982; Traill 2004).
When competition seems to occur as a result of extensive overlap in the use of a
limiting resource, a finer level of analysis may reveal separation being achieved by a
more fine-grained division of the environment (Dunbar 1978). Ecological overlap may
be reduced in quite subtle ways, which may not be obvious and immediately
apparent. In addition, species may be able to tolerate much greater levels of
192
ecological overlap in more diverse or richer habitat types than in simple and resource
poor ones (Dunbar 1978).
Competition is not expected to play a role in two species that have a low resource
overlap. Such species may separate spatially or overlap extensively in a random
manner and depending on the distribution of their preferred resources (Hofer et al.
2004). Where a high degree of resource overlap is evident ecological competition
hypotheses can be formulated.
It is predicted that if two species were similar on any one of the resource dimensions
(e.g. habitat, diet, spatial distribution), they would be segregated on some other
dimension (Pianka 1973). For coexistence, segregation has to occur on at least one
dimension. Therefore, if they showed similarity in habitat use, they may differentiate
along another dimension, such as spatially or behaviourally, or even in terms of diet
in order to coexist successfully.
Taking into account the historical distributions, occasionally overlapping populations
and the morphological, ecological and physiological similarities between the black
and blue wildebeest, interspecific competition is expected to occur between the two
types of wildebeest in areas where they have been confined together. Interspecific
competition is here defined as the act of two species seeking the same space and
food (exploitation), which are in short supply, or interacting in such a way that their
growth and survival are affected (interference) (Anthony and Smith 1977).
Interference competition includes active or passive social interactions such that one
or both species avoids the other, whereas exploitative competition exists when one
species uses a resource, making it unavailable to another species (Putman 1996).
This chapter explores the evidence for exploitative competition between the black
and blue wildebeest and Chapter 10 explores evidence for interference competition.
It is hypothesised that exploitative competition between the black and blue
wildebeest will take place.
The objectives of this part of the present study were therefore to:
•
Describe quantitatively the respective niches of the two types of wildebeest along
the habitat, spatial distribution and dietary dimensions
•
Explore the potential for exploitative interspecific competition between the two
types of wildebeest based on their overlap in the respective niche dimensions.
193
METHODS
Habitat niche dimension
Niche breadth was calculated for each type of wildebeest by using the different
habitat types as resource states and for a separate analysis the different habitat
factors were also used as resource states. The following equation of Levins (1968)
was used to calculate the niche breadth and hence to quantify the niches of the black
and blue wildebeest along the habitat dimension:
B=1/™nipj2
where B is Levin’s measure of niche breadth, pj the proportion of the observations
found in resource state j. B is maximum when the observation percentages are
similar in each resource state and would be minimal if all observations occur only in
one resource state. B was standardised to BA, ranging from 0 to 1, based on the
following equation of Hurlbert (1978):
BA =B-1/n-1
where n is the number of resource states.
Overlap in resource use was assessed by using Pianka’s niche overlap index
(Pianka 1973):
Orm = ™n(pjr x pjm) / (™Sjr2 x ™Sjm2)1/2
where Orm is the index of niche overlap between black and blue wildebeest; pjr is the
black wildebeest proportionate use of the resource unit j (calculated as: percentage
of the observations that were in habitat j, as a proportion of the total observations
across all habitat types); pjm is the blue wildebeest proportionate use of resource unit
j; n is the total number of habitat types. This is a symmetrical measure that ranges
from 0 when no habitats are used in common to 1 when there is complete habitat
overlap.
For comparison another measure of niche overlap was utilised. The Sorenson’s
Quotient of Similarity (QS) as used by Churchfield et al. (1999) was therefore also
194
applied to the habitat type data and the individual habitat factor data. The equation
used was as follows:
QS = 2j/(a+b)
where j = the total number of resource states common to both types of wildebeest
being compared; a = the total number of resource states found in species a (black
wildebeest); b = the total number of resource states found in species b (blue
wildebeest).
Spatial distribution niche dimension
It was difficult to provide quantitative measurements of the distributions of the black
and blue wildebeest at Ezemvelo Nature Reserve as individuals were not identified.
A descriptive analysis has therefore been provided to give some indication of the
spatial overlap between the two types of wildebeest at Ezemvelo Nature Reserve.
The study area was divided into grids of equal size and the number of black and blue
wildebeest within each grid was calculated and converted to a percentage of the total
number of animals of that type of wildebeest. To calculate indices of spatial overlap
by using these percentages the method as described by Anthony and Smith (1977)
was applied. For each grid, if for example the black wildebeest occurrence was 5%
and the blue wildebeest occurrence was 7%, then the overlap, Yi , would be 5%. The
total overlap in spatial distribution for a particular season was calculated from the
sum of the overlaps for each individual grid according to the following equation:
™<i ,
n
i=1
where n equals the total number of grids utilised. Sorenson’s Quotient of similarity
was also applied to this data to indicate the degree of similarity between the areas
utilised by the black and blue wildebeest.
Estimated dietary niche dimension
Since results from a detailed faecal analysis of the diet of the black and blue
wildebeest were not yet available, the diet of both types of wildebeest had to be
inferred from floristic data at the sites of their occurrence. It was assumed that a large
proportion of the diet would be made up of the dominant plant species that were
found within the feeding sites of each type of wildebeest. During the habitat surveys
195
(Chapter 4), data were collected at each site of occurrence of a black or a blue
wildebeest on the dominant and sub-dominant plant species present within a 5 m
radius of the site where the wildebeest was observed. Only data for the grazing
observations were extracted to estimate the diet. Further data from the vegetation
surveys that were done at the feeding sites of the black and blue wildebeest (Chapter
6) were also utilised to indicate the grass species with the highest percentage
occurrence in the chosen feeding site of either type of wildebeest. This estimated diet
was then used to determine the estimated dietary overlap between the black and
blue wildebeest. Overlap in diet was calculated by using the same method as was
described for spatial overlap. Sorenson’s Quotient of similarity was also applied to
this data to indicate the percentage of shared plant species in the diets of the black
and blue wildebeest.
RESULTS
Habitat type niche dimension
Habitat type niche breadth and overlap indices of black and blue wildebeest for each
ecological season are given in Table 9.1. Niche breadths of all the black wildebeest
social groups combined differed significantly from that of blue wildebeest as
assessed by the paired t-test (t = -12.19; df = 2; p = 0.007). Niche breadths also
differed significantly between the female herds of the black wildebeest and those of
the blue wildebeest (t = -5.27; df = 2; p = 0.034). Black wildebeest female herds
tended to have a significantly lower niche breadth than the blue wildebeest female
herds (0.27 vs 0.53). The territorial bulls of the black wildebeest did not have a
significantly lower niche breadth than those of the blue wildebeest (t = -3.15; df = 2; p
= 0.087). Territorial bulls had a higher niche breadth than both bachelor herds and
female herds for both types of wildebeest.
Niche breadth was the lowest during the dormant season (0.25) and the highest
during the early growing season (0.37) for the black wildebeest for all social groups
combined. For the blue wildebeest, niche breadth was maximised during the early
growing season (0.63) and at its lowest during the late growing season (0.50) for all
social groups combined.
Niche overlap between the two types of wildebeest was least during the late growing
season (0.77), while maximal niche overlap occurred during the early growing
196
season (0.94) for all social groups combined. Territorial bulls had the highest niche
overlap during the late growing season (0.94), while female herds had the highest
overlap during the early growing season (0.94). The lowest overlap for female herds
was during the late growing season (0.73).
In terms of habitat type utilisation, overlap between the black and the blue wildebeest
was generally high for all the seasons (0.87). Similarly, the Sorenson’s Quotient of
Similarity indicated a degree of similarity of 0.89 in terms of the habitat type utilisation
of the black and blue wildebeest.
Habitat factor niche dimension
Niche breadth and overlap indices for the habitat factors for black and blue
wildebeest appear in Tables 9.2 and 9.3. Niche breadth of the black wildebeest only
differed significantly from that of the blue wildebeest in terms of distance to water (t =
4.94; df =2; p = 0.039), woody vegetation cover (t = -14.76; df = 2;p = 0.005), total
grass height (t = -7.84; df =2; p = 0.016), and geomorphology (t = 11.7; df = 2; p =
0.007). The black wildebeest (0.87) were more likely to utilise a wider variety of
distances from water than the blue wildebeest (0.83) and had a narrow niche breadth
in terms of different woody vegetation covers (0.00), while the blue wildebeest tended
to utilise a wider variety of woody vegetation covers (0.53). The blue wildebeest
(0.88) was less selective in terms of the total grass height at the site of occupation
than the black wildebeest (0.47), and the black wildebeest (0.88) utilised a wider
range of geomorphology than the blue wildebeest (0.35). Although they were not
found to be significantly different between the black and blue wildebeest, the rock
cover and exposure habitat factors also indicated that the blue wildebeest had a
higher niche breadth than the black wildebeest (Table 9.2).
197
Table 9.1: The index of Levin (1968) of niche breadth for each type of wildebeest,
and that of Pianka (1973) of niche overlap for each season and social group for the
habitat type choices of the black and blue wildebeest at Ezemvelo Nature Reserve
from January 2004 to August 2005. The t-values are based on the paired t-test
Late growing
Dormant
Early growing
Entire study
season
season
season
period
Black wildebeest
0.30
0.25
0.37
0.31
Blue wildebeest
0.50
0.52
0.63
0.59
Black wildebeest
0.26
0.20
0.36
0.27
Blue wildebeest
0.39
0.47
0.57
0.53
Black wildebeest
0.51
0.34
0.43
0.46
Blue wildebeest
0.54
0.48
0.83
0.63
Black wildebeest
0.55
0.13
0.23
0.36
Blue wildebeest
0.73
0.64
0.74
0.71
All data
0.77
0.89
0.94
0.87
Female herds
0.73
0.90
0.95
0.87
Bachelor herds
0.76
0.55
0.85
0.72
Territorial bulls
0.94
0.79
0.80
0.87
t-value
P-value
-12.13
0.007
-5.27
0.034
-1.74
0.225
-3.15
0.087
Niche breadth
All data
Female herds
Bachelor herds
Territorial bulls
Niche overlap
198
Table 9.2: The index of Levin (1968) for niche breadth for each type of wildebeest for
each season and for the entire study period for the abiotic and biotic habitat factor
choices of the black and blue wildebeest at Ezemvelo Nature Reserve from January
2004 to August 2005 obtained from the habitat survey data (Chapter 4) and based on
the paired t-test
Habitat variable
Type of
Late growing
Dormant
Early growing
Entire study
season
season
season
period
Black wildebeest
0.83
0.99
0.98
0.97
Blue wildebeest
0.74
0.48
0.64
0.65
Black wildebeest
0.61
0.58
0.69
0.63
Blue wildebeest
0.58
0.52
0.48
0.52
Black wildebeest
0.38
0.52
0.37
0.43
Blue wildebeest
0.40
0.46
0.38
0.42
Black wildebeest
0.84
0.78
0.98
0.87
Blue wildebeest
0.76
0.65
0.91
0.83
Black wildebeest
0.01
0.00
0.00
0.00
Blue wildebeest
0.54
0.46
0.58
0.53
Black wildebeest
0.48
0.76
0.65
0.73
Blue wildebeest
0.82
0.73
0.98
0.88
Black wildebeest
0.45
0.50
0.66
0.53
Blue wildebeest
0.57
0.75
0.76
0.62
Black wildebeest
0.43
0.46
0.50
0.47
Blue wildebeest
0.69
0.81
0.91
0.88
Black wildebeest
0.56
0.60
0.98
0.76
Blue wildebeest
0.81
0.85
0.80
0.88
Black wildebeest
0.50
0.35
0.68
0.52
Blue wildebeest
0.55
0.68
0.36
0.58
Black wildebeest
0.53
0.51
0.49
0.70
Blue wildebeest
0.55
0.68
0.76
0.74
Black wildebeest
0.45
0.50
0.26
0.41
Blue wildebeest
0.38
0.44
0.59
0.48
Black wildebeest
0.29
0.36
0.44
0.37
Blue wildebeest
0.43
0.37
0.60
0.47
Black wildebeest
0.86
0.88
0.87
0.93
Blue wildebeest
0.95
0.96
0.84
0.94
Black wildebeest
0.00
0.00
0.00
0.00
Blue wildebeest
0.12
0.06
0.15
0.11
Black wildebeest
0.92
0.80
0.89
0.88
Blue wildebeest
0.39
0.35
0.28
0.35
Black wildebeest
0.42
0.47
0.39
0.57
Blue wildebeest
0.46
0.63
0.54
0.64
Black wildebeest
0.05
0.15
0.33
0.18
Blue wildebeest
0.04
0.05
0.15
0.08
Black wildebeest
0.48
0.51
0.56
0.55
Blue wildebeest
0.54
0.55
0.60
0.59
wildebeest
Aspect
Slope
Landscape position
Distance to water
Woody vegetation cover
Grass cover
Rock cover
Total grass height
Grass leaf height
Plant utilisation
Visibility
Distance to shade
Erosion
Altitude
Exposure
Geomorphology
Forb : grass ratio
Drainage
Combined niche breadth
199
t-value
p-value
2.49
0.130
1.80
0.213
0.62
0.597
4.94
0.039
-14.76
0.005
-1.69
0.233
-3.48
0.074
-7.84
0.016
-0.74
0.538
-0.10
0.927
-2.19
0.160
-0.53
0.647
-2.21
0.157
-1.15
0.369
-3.99
0.058
11.70
0.007
-2.95
0.099
1.89
0.200
-7.00
0.020
The combined niche breadth (considering all habitat factors together) of the blue
wildebeest was significantly higher than the black wildebeest (t = -7.00; df = 2; p =
0.020) (Table 9.2).
For the black wildebeest, the highest niche breadth was in aspect (0.97) and altitude
(0.93), while for the blue wildebeest the highest niche breadth occurred in the grass
leaf height (0.88) and grass cover (0.88) habitat factors. Black wildebeest had their
lowest niche breadth in the woody vegetation cover (0.00), exposure (0.00) and
drainage (0.19) habitat factors. Blue wildebeest similarly had the lowest niche
breadth in the drainage (0.08) and exposure (0.11) habitat factors.
The highest niche breadth in terms of all habitat factors combined was during the
early growing season for both the black and blue wildebeest. The lowest niche
breadth for both black and blue wildebeest was found during the late growing
season. This is an almost identical pattern to that found for the habitat type analysis
in the previous section. The only significant habitat factor that differed from this
pattern was geomorphology where the highest niche breadth for both black and blue
wildebeest was during the late growing season.
Niche overlap between the black and blue wildebeest was complete for the plant
utilisation (1.00), erosion (1.00), exposure (1.00) and drainage (1.00) habitat factors,
and very high for the visibility (0.99), forb : grass ratio (0.98), grass cover (0.98),
grass leaf height (0.96), slope (0.96), and landscape position (0.95) habitat factors.
Niche overlap for the black and blue wildebeest was lowest for the distance to shade
(0.46) habitat factor. The highest niche overlap between the black and blue
wildebeest was during the dormant season followed closely by the late growing
season and the lowest niche overlap was during the early growing season (Table
9.3).
The Sorenson’s Quotient of Similarity between the black and blue wildebeest for
each habitat factor was very high and mostly equalled 1 as the categories for each
habitat factor were subjectively constrained based on wildebeest occurrence and
thus these values have not been presented here.
200
Table 9.3: The index of Pianka (1973) of niche overlap for each season and for the
entire study period for the abiotic and biotic habitat factor choices of the black and
blue wildebeest at Ezemvelo Nature Reserve from January 2004 to August 2005
obtained from the habitat survey data (Chapter 4)
Late growing
Dormant
Early growing
season
season
season
Aspect
0.79
0.83
0.86
0.85
Slope
0.94
0.96
0.94
0.96
Landscape position
0.96
0.94
0.96
0.95
Distance to water
0.84
0.90
0.95
0.93
Woody vegetation cover
0.92
0.92
0.89
0.91
Grass cover
0.96
0.99
0.87
0.98
Rock cover
0.74
0.72
0.94
0.74
Total grass height
0.93
0.95
0.76
0.92
Grass leaf height
0.97
0.97
0.88
0.96
Plant utilisation
1.00
0.93
0.85
1.00
Visibility
0.99
0.94
0.94
0.99
Distance to shade
0.45
0.53
0.39
0.46
Erosion
0.99
1.00
0.98
1.00
Altitude
0.95
0.93
0.89
0.94
Exposure
1.00
1.00
1.00
1.00
Geomorphology
0.86
0.77
0.82
0.82
Forb : grass ratio
0.97
0.98
0.97
0.98
Drainage
1.00
1.00
0.98
1.00
Combined overlap
0.90
0.90
0.88
0.91
201
Overall
Spatial distribution niche dimension
It is evident that there is a high degree of spatial separation between the two types of
wildebeest. The overall spatial overlap between the black and the blue wildebeest at
Ezemvelo Nature Reserve was only 14%. Black wildebeest were primarily found on
the open plains on the northwestern side of the reserve on the high-lying areas. The
only place where the blue wildebeest did not occur in large numbers was on these
open plains that were favoured by the black wildebeest. Overlap between the black
and blue wildebeest occurred primarily on the edges of the open plains where blue
wildebeest territorial bulls were invading territories that were favoured by the black
wildebeest bulls.
Since there were some seasonal changes in the spatial distributions of both types of
wildebeest, the data were also analysed on a seasonal basis. The total spatial
overlap for the late growing season was 11%, for the dormant season 8% and for the
early growing season 13%. Thus spatial overlap was highest during the early growing
season and the lowest during the critical dormant season. Only 60% of the study
area was utilised by any type of wildebeest. The remaining 40% is made up of
unsuitable habitat for both types of wildebeest, consisting mainly of rocky slopes that
are covered with dense woodland.
The degree of similarity of the areas occupied by the black and blue wildebeest
based on the Sorenson’s Quotient of Similarity indicated that it was 23% during the
late growing season, 21% during the dormant season and 27% during the early
growing season.
Estimated dietary niche dimension
The plant species present in the feeding sites and their relative percentage frequency
of occurrence are shown in Figure 9.1 and 9.2. The niche breadth of the black
wildebeest was found to be 0.64, which was higher than the niche breadth for the
blue wildebeest, which was found to be 0.39. The most common plant species in the
feeding sites of the black and blue wildebeest was Eragrostis curvula, and the
dominant plant species in the sites occupied by the black wildebeest was Cynodon
dactylon and Eragrostis curvula for the sites occupied by the blue wildebeest. Black
wildebeest had a larger variety of plant species than the black wildebeest (Figure 9.2)
in their feeding sites indicating the possibility of a more diverse diet.
202
An overall overlap in the species composition of the feeding sites of the black and
blue wildebeest was found to be 55%. Assuming that the composition of the feeding
sites gives an indication of the diet of the black and blue wildebeest, the estimated
dietary overlap is therefore 55%. It is acknowledged that a quantitative Sorenson’s
Index may have provided more accurate results.
The shared plant species in the diet of the black and blue wildebeest as calculated by
the Sorenson’s Quotient of Similarity was 88% for the feeding sites (Chapter 6).
According to the dominant plant species in the areas occupied by the black and blue
wildebeest and determined in the habitat surveys (Chapter 4) the shared plant
species in the diet of the black and blue wildebeest was 81%.
DISCUSSION
The present study has shown that the niche breadth of the black wildebeest is
smaller than that of the blue wildebeest in terms of the spatial and habitat
dimensions. However, preliminary results on the possible diet of the two types of
wildebeest indicate that the black wildebeest has a higher niche breadth than the
blue wildebeest in terms of the dietary dimension. This is supported by other studies
that have indicated that the natural distributional range of the two types of wildebeest
tends to follow differences in their habitat tolerances that do not reflect differences in
their trophic behaviour (Codron and Brink In press).
Black and blue wildebeest have not until recently been studied in the same area
before and therefore details on their joint resource partitioning are lacking. The
knowledge that is available which can be utilised to predict the outcome of
competition between the two types of wildebeest is based on overall physiology,
morphology and individual habitat choices, all of which have been studied in
isolation. Codron and Brink (In press) conducted a study on the feeding niches and
trophic ecology of the black and blue wildebeest using stable carbon and nitrogen
isotope data from faeces and tooth dentine collagen and concluded that speciation of
the black and blue wildebeest was not driven by resource competition. The results of
this study, however, indicated that there were different trophic behaviours between
the two types of wildebeest but that sympatric coexistence of the two types of
wildebeest was facilitated by differential niche occupation at herd levels rather than
between species.
203
Forb
Sporobolus africanus
Paspalum urvillei
urvellei
Eragrostis plana
Eragrostis nindensis
Setariasphacelata
sphacelatasubsp.
subp. sphacelata
Sphacelata
Setaria
Plant species
Elionurus muticus
Blue wildebeest
Melinis repens
Black wildebeest
Aristida
barbicollis
Aristidacongesta
congesta subsp.
subsp. barbicollis
Schizachyrium sanguineum
Loudetia simplex
Aristida
congestasubsp.
subsp.congesta
congesta
Aristida congesta
Aristida junciformis
Eragrostis chloromelas
Tristachya rehmannii
Digitaria eriantha
Eragrostis gummiflua
Cynodon dactylon
Eragrotis curvula
0
5
10
15
20
25
Percentage of observations
Figure 9.1: Percentage frequency of occurrence of the dominant plant species at the
sites of occupation of the black and blue wildebeest at Ezemvelo Nature Reserve
from January 2004 to August 2005. Data obtained during the habitat survey
collection period (Chapter 4).
204
Brachiaria brizantha
Imperata cylindrica
Eragrostis chloromelas
Pogonarthria squarrosa
Perotis patens
Setaria sphacelata
sphacelata var. sphacelata
Setaria
sphacelata
Eragrotis racemosa
Trachypogon spicatus
Elionurus muticus
Plant species
Aristida transvaalensis
Aristida
congestasubsp.
subsp.congesta
congesta
Aristida congesta
Sporobolus africanus
Eragrostis gummiflua
Blue wildebeest
Black wildebeest
Themeda triandra
Melinis repens
Eragrostis inamoena
Paspalum
Paspalum urvellei
urvillei
Aristida junciformis
Cynodon dactylon
Aristidacongesta
congestasubsp.
barbicollis
subsp. barbicollis
Aristida
Digitaria eriantha
Forb
Eragrostis curvula
0
2
4
6
8
10
12
14
16
Percentage composition
Figure 9.2: Species composition at the feeding sites of the black and blue wildebeest
at Ezemvelo Nature Reserve for all plant species contributing more than 2% to the
overall species composition in the feeding sites sampled in March 2004. Data
obtained from vegetation surveys in the feeding sites of the black and blue
wildebeest (Chapter 6).
205
Black wildebeest are smaller than blue wildebeest, both are ruminants with similar
mouth morphologies and are selective short grass feeders. The present study has so
far indicated that there is a large degree of habitat and dietary overlap and spatial
separation is high. In addition it appears that the black wildebeest population in the
study area has been decreasing over the last 3 years (2003 to 2005), while the blue
wildebeest population has increased in density over the same time period. Does the
above information indicate the presence of sufficient ecological separation between
the two types of wildebeest to reduce competition to such an extent that they can
coexist successfully when confined together? The response of two ecologically
similar types of wildlife to habitat heterogeneity through time and space may be
adequately dissimilar to result in an overall shift in trophic position (Cordon and Brink
2006). Therefore, the answer to the question posed would depend on the
heterogeneity of the habitat in which the two types of wildebeest are confined and
whether that habitat is able to cater for the specific requirements that separate the
two types of wildebeest.
The two types of wildebeest may be able to tolerate greater levels of ecological
overlap in more diverse or richer habitats (Dunbar 1978) as such habitats would have
more niches available for exploitation that could allow for their coexistence even if
most of those niches are utilised by both types of wildebeest. In richer habitats there
will always be some niche that could be utilised by one type but not by the other.
Had black and blue wildebeest evolved sympatrically, they may have been expected
to have diverged or occupied different niches at such an extent that they were able to
coexist. However, being brought together recently, their similarity may be testing the
limits of coexistence (Bryce et al. 2002), especially in areas with low habitat
heterogeneity.
Codron and Brink (In press) conclude that since trophic partitioning is an important
mechanism allowing sympatric species to avoid competition, it is anticipated that a
trophic shift would have accompanied the divergence of the black from the blue
wildebeest. However, as has been stated previously trophic adaptation between the
black and blue wildebeest do not differ. Therefore it is thought that the two types of
wildebeest may have been able to co-exist in some areas by temporally shifting their
feeding niches over short time scales (Brink et al. 1999; Codron and Brink In press).
206
Studies of niche overlap and ecological separation amongst coexisting shrews
concluded that body size has an important role in effecting ecological separation in
multi-species communities where a high degree of morphological and ecological
similarity occurs between members (Churchfield et al. 1999). Black wildebeest are
smaller than blue wildebeest and therefore would be expected to require a higher
energy diet due to increases in metabolic rate (Codron and Brink In press). This has
been disproven and studies have shown that the feeding ecology of the black
wildebeest is not dissimilar to the blue wildebeest.
Black wildebeest evolved in the high-lying open treeless grasslands of South Africa,
while the blue wildebeest evolved in the savannas of the low-lying portions of
southern Africa. Therefore, from their evolutionary histories, it may be expected that
the black and blue wildebeest would have different habitat preferences. Both types of
wildebeest are selective grazers of short grass and therefore it would be expected
that they would utilise the grass species that are available in a similar way. The black
wildebeest has a dark, thick coat which is adapted to continual exposure to full sun,
while the blue wildebeest has a thin, dark pelage that is less adapted to sun
exposure. Hence the blue wildebeest will seek shade when it is available. Therefore,
these physiological differences indicate that there may be differences in the way in
which the two types of wildebeest distribute their activities during the daytime and
that would influence the temporal distribution of their active periods.
Niche breadth indices based on habitat type (Table 9.1) indicated that the blue
wildebeest had more generalist habitat associations than the black wildebeest. The
more detailed analyses of habitat selection incorporating the various abiotic and
biotic factors of the habitat also suggested that, at the spatial scales considered in
the present study, blue wildebeest showed fewer preferences than black wildebeest.
Although the two types of wildebeest showed different preferences in their habitat
use, their use of the habitat types and factors showed considerable overlap. At the
time of the present study the spatial overlap between the two types of wildebeest was
low. However, there was an indication that the blue wildebeest bulls were moving into
the areas that were traditionally occupied by the black wildebeest, at least during
periods of the year when the heat load was not as high, therefore allowing the blue
wildebeest to venture into areas where shade was not in the near vicinity.
207
Table 9.4 summarises the indices of overlap and quotients of similarity calculated in
the present study. The indices of niche overlap that were found in the present study
indicated that the overlap in spatial distribution was low for the entire study period
(14%) but was lowest during the dormant season (8%). This critical season was
when food shortage may play the greatest role. Black wildebeest may be able to
maintain themselves in their chosen habitats, but the invading blue wildebeest may
be unable to maintain their body weight in these areas. Such areas may also be too
exposed to the cold during the dormant season and thus also make them unsuitable
for the blue wildebeest. Since only the blue wildebeest encroaches on the black
wildebeest habitats, it is the blue wildebeest that are the cause of the observed
spatial overlap. The reasons why the blue wildebeest bulls have been entering this
habitat need to be elucidated. Possible explanations include that it is only bachelors
with no territory that have been forced out of the preferred blue wildebeest habitat
into the black wildebeest habitat; or that the sex ratio of the blue wildebeest is
skewed and hence blue wildebeest bulls are seeking potential territories in other
habitats; or that the population size of the blue wildebeest has outgrown its preferred
habitat and hence is expanding into less favourable habitat. The highest spatial
overlap is during the early growing season. This is the season when the calves are
born and fresh grass is sprouting after the first rains.
Spatial segregation is less at finer spatial scales than that of broad habitat types. If
competition is a strong structuring force, then a non-random spatial distribution can
be expected (Hofer et al. 2004). The spatial overlap in the present study was lower
than expected by chance. This could be as a result of interspecific competition for
similar resources, and the result of this reduction of spatial overlap would decrease
the intensity of competition (Hofer et al. 2004).
Overall, indices of niche breadth indicate that black and blue wildebeest habitat
resources were likely to be very similar at the study area and the overlap indices
indicated that the two types of wildebeest overlapped considerably in terms of habitat
choices. Regardless of the scale, the extent of overlap was considerable and greater
than the 0.5 proposed by Levins (1968) to prohibit coexistence.
208
Table 9.4: Summary of the indices of overlap and quotients of similarity in spatial
distributions, habitat type selection and diet between black and blue wildebeest at
Ezemvelo Nature Reserve from January 2004 to August 2005. Table adapted from
Anthony and Smith (1977). (QS is the quotient of similarity)
Spatial
Season
QS for
Habitat
QS for
Diet
QS for
overlap
shared
type
shared
overlap
shared plant
(S)
areas of
selection
habitat
(F)
species in
occupation
(H)
types
Late growing season
0.11
0.23
0.77
0.89
0.55
-
Dormant season
0.08
0.21
0.89
0.89
0.55
-
Early growing season
0.13
0.27
0.94
0.89
0.55
-
Seasonal mean
0.11
0.24
0.86
0.89
0.55
0.88
209
diet
The indices utilised to calculate niche breadth and overlap in the present study may
have reduced biological interpretability and their suitability as a foundation for
discussion of resource utilization strategies, competition, and species packing
(Hurlbert 1978). An index and its interpretation depends on a number of factors as
listed by (Hurlburt op. cit.) as: how the resource states were defined: whether they
are arbitrary units or discrete natural entities; and whether or not empty resource
states are excluded from the analysis. Overlap indices often fail as measures of
competition and this needs to be taken into consideration when interpreting the
results of the present study.
Comparing these results with the results of the logistic analysis, which separated the
habitats of the black and blue wildebeest (Chapter 7), it is possible to conclude that
habitat differentiation and spatial differentiation between the two types of wildebeest
provide the main mechanism for coexistence at Ezemvelo Nature Reserve and
trophic differences play little or no role.
The black and blue wildebeest need to differ on one dimension for ecological
separation to occur. This dimension was found to be the habitat factor dimension of
distance to shade. The tools required to measure this dimension properly may not
have been applied correctly in the present study and hence further detailed analysis
of this dimension is required.
CONCLUSION
Although black wildebeest are more selective than blue wildebeest, resource
partitioning between the two types of wildebeest was found to be incomplete.
Considerable overlap in the use of key resources such as habitats and possible food
species, but little overlap in spatial distribution and temporal activities was found.
Overlap in resource use tended to be lowest during the dormant season when food
resources were most limiting. The results of the present study indicated that
exploitative competition was being avoided at the current population levels at
Ezemvelo Nature through the partitioning of mutually exclusive resources such as
woody vegetation cover and altitude and space. This partitioning was relaxed during
the late growing and early growing seasons when the food resources were less
limiting than during the dormant season.
210
CHAPTER 10: BEHAVIOURAL INTERACTIONS, SPECIES ASSOCIATIONS AND
INTERSPECIFIC INTERFERENCE COMPETITION
INTRODUCTION
Interactions with other species may indicate facilitative or competitive behaviour
between individuals (Arsenault and Owen-Smith 2002). An investigation into the
interspecific behaviour of coexisting black and blue wildebeest may reveal evidence
of interference which may further indicate that competition could play a role in the
continued existence or decline of one or both of the two types of wildebeest at
Ezemvelo Nature Reserve. Instances of both intolerance and mutual tolerance and
their evaluation in terms of random expectation may help in revealing such
competition (Krämer 1973). Avoidance behaviour may allow antagonistic species to
coexist in common habitats without stressful interactions (Anthony and Smith 1977).
Observations of interspecific behaviour are expected to reveal interspecific
dominance if present (Krämer 1973) and to shed some light on the question of
hybridisation between the black and blue wildebeest.
Evidence for intolerance or mutual tolerance of the two types of wildebeest has not
been investigated. Only anecdotal reports of associations between black and blue
wildebeest are available (Vrahimis 2004 pers. comm.)13.
It has already been found that spatial overlap between the black and blue wildebeest
at Ezemvelo Nature Reserve is low (11%) (Chapter 9). Overlap occurred where blue
wildebeest bulls have entered the preferred areas of the black wildebeest during
times of high food abundance. No cases of black wildebeest entering blue wildebeest
areas were observed. It was hypothesised that interference competition would be
evident between the two types of wildebeest
The objectives of this part of the present study were therefore to:
•
Investigate the interspecific behaviour of the coexisting black and blue wildebeest
at Ezemvelo Nature Reserve for evidence of interference competition
•
Determine the associations of each type of wildebeest with other wildlife species
present within the study area.
13
Prof. S. Vrahimis. Free State Department of Tourism, Environmental and Economic Affairs,
Bloemfontein, Free State, 0004, South Africa.
211
METHODS
Behavioural interactions
When black and blue wildebeest were located within 45 m of each other in the
present study, notes were taken on any behavioural interactions between the two
types of wildebeest. Behavioural interactions between black and blue wildebeest
were evaluated in only those encounters in which individuals or groups of individuals
of the two types of wildebeest came close enough for antagonism or avoidance to be
displayed by either type (Anthony and Smith 1977). Passive dominance refers to
those situations in which one type of wildebeest was dominant but no overt
aggression was observed. Active dominance refers to those situations in which one
type of wildebeest was dominant by way of overt aggression. Overt aggression was
displayed by horning and other territorial displays as described by Von Richter
(1971a) and Estes (1969). In addition, in the habitat surveys (Chapter 4), whenever a
wildebeest was located, any association with other species was also recorded if the
two involved were within 100 m of each other.
Temporal activity overlap
The diurnal distribution of activity (grazing and walking) data obtained during the
activity budget surveys (Chapter 8) was used to determine the likelihood of mutual
interference among the two types of wildebeest by averaging the conjoint
probabilities of being active at the same time across the hours of the day as was
described by Dunbar (1978).
Species associations
The Chi-squared goodness of fit was used to statistically compare species
associations of the two types of wildebeest and the number of individuals of each
species present on the reserve (Thomas and Taylor 1990; Manly et al. 1993).
Expected frequencies were calculated from the available proportions of types of
wildlife obtained from the counts conducted within the study area. If the Chi-squared
test was found to be significant, the null hypothesis that all the associations were in
proportion to the number of individuals of the associated species on the reserve (no
selection) was rejected. Subsequently, the cell Chi-squared values for each species
212
category were calculated. If these were significant the difference between the
observed and expected values was examined. If the observed value was greater
than the expected value, a positive association with that species was concluded. If
the expected value was more than the observed value, it was concluded that that
species category was avoided by the relevant type of wildebeest.
RESULTS
Behavioural interactions
Encounters between the two types of wildebeest were most frequent during the
rutting season (30%) (Table 10.1). In 49% of the encounters between black and blue
wildebeest, dominance could not be determined. When dominance could be
determined (51% of the encounters), blue wildebeest were dominant in all but one
encounter. Dominance of the blue wildebeest over the black wildebeest did not
appear to depend upon the sex or group composition of either species (Table 10.1).
The dominance of blue wildebeest over black wildebeest was displayed mostly by
passive mechanisms (68%) and occasionally by active mechanisms (32%). Passive
dominance usually occurred when the two types of wildebeest were grazing together.
Temporal activity overlap
The diurnal distribution of activity (grazing and walking) for the overall data and by
season is shown in Figures 10.1 and 10.2. The mean conjoint probability that black
and blue wildebeest will be active at the same time during the day was 0.225 for all
the data, 0.170 for the late growing season, 0.229 for the dormant season and 0.283
for the early growing season. These results indicated that the likelihood of
interference between the black and blue wildebeest when feeding was low and that
direct competition for forage was thus relatively unlikely. Black wildebeest tended to
concentrate their active periods during the early mornings and the late afternoons.
Additional observations have indicated that the black wildebeest was more active at
night than the blue wildebeest, which may allow for further separation in the active
period of each type of wildebeest. Blue wildebeest were much more likely to be
active over the midday period (12:00 to 14:00) than the black wildebeest.
213
Table 10.1: Summary of the behavioural interactions between the black and blue
wildebeest at Ezemvelo Nature Reserve as recorded from April 2004 to August 2005.
M = males; F = females; C = calves
Date
Black wildebeest
Blue wildebeest
Dominant species
Type of dominance
21/04/2004
1M, 8F, 2C
1M
Blue wildebeest
Passive
26/04/2004
5M
1M
Blue wildebeest
Passive
27/04/2004
9M
1M
Not determined
20/07/2004
1M
1M, 20 F, 6 C
Blue Wildebeest
Passive
16/08/2004
5M
1M
Blue wildebeest
Passive
17/08/2004
1 M, 25 F, 6 C
1M
Not determined
20/08/2004
1M
1M
Not determined
21/08/2004
4 M, 25 F, 5 C
1M
Blue wildebeest
Active
03/09/2004
1M, 7F, 2C
1M
Black wildebeest
Active
05/09/2004
4M
1M
Not determined
16/09/2004
2M
1M
Blue wildebeest
Passive
21/09/2004
9M
1M
Blue wildebeest
Active
30/09/2004
1M, 16F, 3C
1M
Not determined
26/10/2004
1M
1M, 16F, 7C
Blue wildebeest
Passive
10/11/2004
1M
1M
Blue wildebeest
Passive
13/11/2004
1M, 9 F, 1C
1M
Blue wildebeest
Passive
04/01/2005
2M, 1F, 1C
1M
Not determined
04/01/2005
1M
1M
Blue wildebeest
10/01/2005
1M, 28F, 7C
1M
Not determined
11/01/2005
2M, 1F
1M
Not determined
13/01/2005
2M, 9F, 3C
1M
Not determined
02/02/2005
1M, 18F, 3C
1M
Not determined
25/02/2005
1M
1M
Blue wildebeest
Passive
01/03/2005
1M
1M
Blue wildebeest
Passive
15/03/2005
1M, 10 F, 3C
1M
Not determined
18/03/2005
2M
1M
Blue wildebeest
15/04/2005
3 M, 1F
1M
Not determined
18/04/2005
1M
1M
Blue wildebeest
Passive
22/04/2005
1M
1M
Blue wildebeest
Passive
22/04/2005
1M
2M
Not determined
11/05/2005
1M
1M
Not determined
23/05/2005
1M, 4F
2M
Not determined
10/06/2005
3M, 1F
1M
Not determined
18/06/2005
3M
2M
Not determined
23/07/2005
2M
2M
Blue wildebeest
Passive
26/08/2006
1M, 7F, 1C
3M
Blue wildebeest
Active
30/08/2006
1M, 7F, 1C
2M
Not determined
214
Passive
Passive
Percentage tim e spent active
100
80
60
40
20
0
06:00
08:00
10:00
12:00
14:00
16:00
Tim e of the day (hour s )
B lac k wild e b e e s t
B lue wild e b e e s t
Figure 10.1: Diurnal patterns, expressed as percentage time spent active (grazing
and walking), for black and blue wildebeest based on scan samples of activity taken
at 5-minute intervals throughout the daytime at Ezemvelo Nature Reserve from
March 2004 to August 2005.
215
100
80
active
P ercen tag e tim e sp e
Late growing season
60
40
20
0
06:00
08:00
10:00
12:00
14:00
16:00
T im e o f th e d ay (h ou rs)
B la c k w ilde be e s t
B lue w ilde be e s t
Percentage tim e
spent active
D ormant season
100
80
60
40
20
0
0 6 :0 0
0 8 :0 0
1 0 :0 0
1 2 :0 0
1 4 :0 0
1 6 :0 0
T ime of the day (hours)
B la ck w ild e b e e st
B lue w ild e b e e s t
Percentage tim e
spent active
Ea rly g row in g s e a s on
100
80
60
40
20
0
06:00
08:00
10:00
12:00
14:00
16:00
T ime of the day (hours)
B lac k wild e b e e s t
B lue wild e b e e s t
Figure 10.2: Seasonal diurnal patterns, expressed as the percentage of time spent
active (grazing and walking), for the black and blue wildebeest based on scan
samples of activity taken at 5-minute intervals throughout the daytime at Ezemvelo
Nature Reserve from March 2004 to August 2005.
216
Blue wildebeest were least active around 10:00 in the morning and from 14:00 to
15:00 in the afternoon. Black wildebeest, however, tended to be least active from
09:00 to 14:00.
Species associations
Black wildebeest were observed not to be associated with any other species in
45.3% of the observations while blue wildebeest occurred alone in 61.0% of the
observations (Figure 10.3). On those occasions where wildebeest were associated
with other species, the black wildebeest was most likely to be associated with the
blesbok (28.9%), while the blue wildebeest was most likely to be associated with
Burchell’s zebra (19.6%). In general blue wildebeest were associated with a wider
diversity of species (13) than the black wildebeest (9). No seasonal variations in this
pattern were observed (Figure 10.4).
Blue wildebeest were less likely to be associated with black wildebeest than black
wildebeest were likey to be associated with blue wildebeest. This is indicated by the
result that, of all the blue wildebeest association observations, only 4.2 % were with
black wildebeest and of all the black wildebeest association observations, 7.2 % were
with blue wildebeest. The results also indicated that either a black or a blue
wildebeest was more likely to associate with the other type of wildebeest (i.e. a black
wildebeest was more likely to associate with a bluek wildebeest or a blue wildebeest
was more likely to be associated with a black wildebeest), than with the greater kudu,
common eland, gemsbok Oryx gazella, ostrich, springbok Antidorcas marsupialis,
waterbuck, common warthog Phacochoerus africanus, red hartebeest Alcelaphus
buselaphus caama and impala. This observation is most likely due to different habitat
choices of these wildlife and due to the varying number of individuals of these
species available for association.
The Chi-squared tests indicated that black wildebeest were associated with blesbok
to a much higher degree than what was expected (Table 10.2). The results also
indicated that black wildebeest were associated with the common warthog, Burchell’s
zebra, impala and red hartebeest to a much lesser extent than what was expected.
Blue wildebeest tended to associate with Burchell’s zebra to a much greater extent
than what was expected, and also with the blesbok and interestingly with the black
wildebeest. Blue wildebeest tended to avoid most other species.
217
65
55
50
45
40
35
30
25
20
15
10
5
e
on
N
er
th
O
st
ee
eb
ild
W
R
ed
ha
r te
Im
be
pa
es
*
t
la
a
br
th
ar
W
Ze
og
d
an
El
es
bo
k
0
Bl
Percentage of observations
60
Associate d spe cie s
Blac k wilde bees t
Blue wildebees t
Figure 10.3: The association of the black and blue wildebeest with other types of
wildlife at Ezemvelo Nature Reserve as observed from January 2004 to August 2005.
* Indicates a black or blue wildebeest depending on the type of wildebeest under
analysis. (Eland = common eland; warthog = common warthog; zebra = Burchell’s
zebra).
218
P ercentage of o bservat
L a te g ro w in g s e a s o n
ed
e
W
h
ild
a
e
N
O
th
on
e
st
ee
b
e
rt
Im
r
*
st
e
be
a
p
b
e
Z
W
a
la
ra
g
o
h
rt
la
E
B
le
sb
n
o
d
k
65
60
55
50
45
40
35
30
25
20
15
10
5
0
R
*
Asso ciate d sp e cie s
B lac k w ild e b e e s t
B lue w ild e b e e s t
e
on
e
N
O
th
e
r
st
ee
b
e
rt
R
*
W
ed
h
ild
a
Z
*
st
e
be
a
p
e
b
h
la
ra
g
o
Im
B
W
a
E
rt
la
sb
n
o
d
k
60
55
50
45
40
35
30
25
20
15
10
5
0
le
P ercen tag e o f o b servat
D o rm a n t s ea s o n
Asso ciate d sp e cie s
B la c k w ild e b e e s t
B lue w ild e b e e s t
e
e
th
on
N
e
W
ild
O
ee
b
e
R
ed
h
a
rt
Im
r
st
*
st
e
a
p
b
e
Z
be
la
ra
g
o
h
W
a
rt
la
E
le
B
n
o
d
k
65
60
55
50
45
40
35
30
25
20
15
10
5
0
sb
P ercentage of observa
E a rly g ro w in g s e a s o n
*
Associate d spe cie s
B la c k w ild e b e e s t
B lue w ild e b e e s t
Figure 10.4: The association of the black and blue wildebeest with other types of
wildlife at Ezemvelo Nature Reserve for the three ecological seasons as observed
from January 2004 to August 2005. * Indicates black or blue wildebeest depending
on the type of wildebeest under analysis. (Eland = common eland; warthog =
common warthog; zebra = Burchell’s zebra).
219
Minimal seasonal differences in this pattern were observed and hence seasonal
analyses were not repeated here.
DISCUSSION
The dominance of the blue wildebeest over the black wildebeest in many of the
interspecific encounters observed throughout the present study was not surprising as
the blue wildebeest is almost 1.5 times the size of the black wildebeest (Furstenburg
2002a and b). This size difference may account for the blue wildebeest’s dominance
in all but one circumstance observed during the present study, but other factors such
as age, period of sexual cycle and group composition may also influence dominance
in interspecific interactions (Anthony and Smith 1977).
Many encounters between black and blue wildebeest involved either a single blue
wildebeest bull with a female herd of black wildebeest, or a number of blue
wildebeest bulls with a number of black wildebeest bulls (Table 10.1). It appears that
the blue wildebeest tends to be the instigators of such interspecific encounters.
Competitive interference is operative only if it affects one or both species by their
exclusion from an area or by detrimental behavioural interactions (Miller 1967). No
evidence of either of these effects could be found on Ezemvelo Nature Reserve.
The extent to which population densities and species ratios affected the results of
this analysis remains undetermined. However, it can tentatively be concluded that
interference competition appears not to play a significant role in the association
between the black and blue wildebeest at Ezemvelo Nature Reserve.
Blue wildebeest tended to be associated with Burchell’s zebra, a roughage and bulk
grass feeder, which accepts both tall and short grasses and does not have a patchselective feeding style (Bothma et al. 2002). Blue wildebeest themselves are
selective grazers who prefer short grass and have a patch-selective feeding style as
do black wildebeest (Bothma et al. 2002). Black wildebeest were associated with the
blesbok on many occasions, also a selective grazer, which prefers short grass and
has a patch selective feeding style. These differences in the species associations
may be indicative of differences in the degrees to which black and blue wildebeest
are able to tolerate areas with tall bunch grasses.
220
Table 10.2: Summary of the Chi-squared tests performed to evaluate the hypothesis
that black and blue wildebeest at Ezemvelo Nature Reserve associated with other
species in proportion to their occurrence. Values in brackets indicate association
frequencies of <5 and therefore the Chi-squared test may not be valid. + indicates a
positive selection, 0 indicates random selection, - indicates a negative selection
Late growing season
Type of
Species
wildebeest
Chi-
df
Selection
square
Black wildebeest
Dormant season
Chi-
df
Selection
square
Early growing season
Chi-
df
Selection
square
Overall
Chi-
df
Selection
square
Blesbok
179.60
1
+
139.61
1
+
151.14
1
+
470.19
1
+
Common
(2.39)
1
0
(1.74)
1
0
0.40
1
0
1.68
1
0
(3.87)
1
-
(1.54)
1
0
(3.53)
1
0
(8.74)
1
-
9.16
1
-
5.89
1
-
9.89
1
-
24.76
1
-
Impala
(7.96)
1
-
(6.64)
1
-
(7.25)
1
-
(21.85)
1
-
Red
0.009
1
0
(2.70)
1
0
(3.20)
1
0
3.97
1
-
Wildebeest*
0.86
1
0
0.17
1
0
0.36
1
0
0.19
1
0
Other
0.32
1
0
0.07
1
0
1.27
1
0
1.29
1
0
Blesbok
7.10
1
+
16.72
1
+
0.73
1
0
20.36
1
+
Common
1.72
1
0
(4.64)
1
-
0.00
1
0
4.23
1
-
1.24
1
0
0.00
1
0
0.05
1
0
0.23
1
0
26.95
1
+
9.12
1
+
21.70
1
+
55.51
1
+
(9.60)
1
-
0.07
1
0
6.73
1
-
10.29
1
-
2.04
1
0
2.86
1
0
1.96
1
0
6.72
1
-
Wildebeest*
6.22
1
+
9.74
1
+
8.49
1
+
24.57
1
+
Other
20.48
1
-
25.21
1
-
21.48
1
-
71.23
1
-
eland
Common
warthog
Burchell’s
zebra
hartebeest
Blue wildebeest
eland
Common
warthog
Burchell’s
zebra
Impala
Red
hartebeest
* Indicates one of the two types of wildebeest depending on which comparison is
involved
221
Black wildebeest occurred in areas on the reserve that were not favoured by
Burchell’s zebra. Burchell’s zebra therefore occurred in similar habitats to the blue
wildebeest and thus they may facilitate the feeding behaviour of blue wildebeest by
reducing the height of the grass layer in areas that may not seem suitable for a
patch-selective grazer (Bell 1970; McNaughton 1976). Therefore, the close
association of the blue wildebeest and Burchell’s zebra may be facilitative rather than
competitive. In contrast, the close association between the black wildebeest and
blesbok, often seen grazing selectively on the same patch within a bunch grass
community, may be a competitive one. Black wildebeest and blesbok have both
evolved similar feeding styles and similar tolerances for open habitats and thus
competition between the two species may be high (Skinner and Chimimba 2005).
It is clear that the social behaviour of wildebeest, especially the habit of male blue
wildebeest to associate with animals of other species, predisposes them to
opportunities of hybridisation when confined with black wildebeest (Vrahimis 2003a).
Of all the blue wildebeest observations where an association with another wildlife
species was recorded, 40% of these were of a lone territorial bull. Of all recorded
cases it appears that disruption of the normal demographic or social structure was
involved, as was seen at the Spioenkop Nature Reserve in 1995 (Langley 1995). At
Ezemvelo Nature Reserve there has not been any disruption of the normal
demographic or social structure of the black or blue wildebeest. However, lone blue
wildebeest bulls are in excess (26% of the blue wildebeest population) (Chapter 12).
These bulls tend to be forced into less favourable habitats by the breeding territorial
bulls, especially during the rutting season (Von Richter 1971a). This will cause
increased encounters between these bulls and the black wildebeest herds. During
the rut, these bulls may find themselves in a suitable position to mate with the black
wildebeest females since the black wildebeest males would either be too small to
fend them off, or too few to prevent mating from occurring. The large number of lone
blue wildebeest bulls at Ezemvelo Nature Reserve is an indication that all the
suitable territories have been occupied. From this it can be inferred that the
maximum stocking density for blue wildebeest has been attained or exceeded. These
lone bulls relegated to unfavourable habitats may interfere with the social structure of
the black wildebeest. These bulls have to be removed from the black wildebeest
habitat on an annual basis to ensure that hybridisation does not take place.
It is believed that hybridisation only occurs under artificial conditions, where the two
species are forced together in a confined area (Vrahimis 2003a). This may become
222
the situation at Ezemvelo Nature Reserve if the population size of the blue
wildebeest is allowed to increase beyond acceptable levels.
It was already stated in Chapter 3 of the present study that it was currently thought
that the only way to ensure that hybridisation between the black and blue wildebeest
does not occur, was to prevent any contact between the two types. This was
because conservationists and scientists could not clearly identify the factors that
resulted in hybridisation. It was suggested that in order to identify these factors, an
understanding of the ecological and behavioural differences between the two types of
wildebeest needed to be attained.
The present study has gone some way towards reaching an understanding of these
differences and has shown that the ecological requirements of the black and blue
wildebeest differ to such an extent that the chances of the two types of wildebeest
cross-breeding is reduced. This is, however, dependent on the population sizes of
the wildebeest present in an area, the level of habitat heterogeneity available, and in
providing a mixture of open grassland and savanna as habitat. The underlying
premise for this conclusion is that the two types of wildebeest are ecologically
separated and pose little threat of crossbreeding under natural conditions.
CONCLUSIONS
The present study has indicated that there was little evidence for interference
competition between the black and blue wildebeest at Ezemvelo Nature Reserve.
Black wildebeest were most commonly associated with a possible competitor, the
blesbok, while the blue wildebeest was most commonly associated with a possible
facilitator, Burchell’s zebra. In the future, the threat of hybridisation between the black
and blue wildebeest at Ezemvelo Nature Reserve could increase unless the
population size of the blue wildebeest is reduced and blue wildebeest lone bulls are
removed from the black wildebeest core habitats.
223
CHAPTER 11: GRAZING CAPACITY AND STOCKING DENSITY
INTRODUCTION
To develop an effective grazing management policy for wildlife on a reserve, detailed
information on the veld condition, grazing capacity and the response of plant species
to grazing pressure is required (Bredenkamp and Theron 1978). This information
provides a basis for the calculation of recommended stocking densities for an area.
The ecological capacity of an area is the potential of that area to support herbivores
through grazing and/or browsing over an extended period without the deterioration of
the ecosystem (Bothma et al. 2004). It is a characteristic of the entire habitat of which
the vegetation, herbivores and their predators all form a part.
Grazing capacity and browsing capacity are used to determine the stocking density
for a wildlife ranch or nature reserve. Stocking density is an important management
aspect on any reserve and relies on reliable predictions of the grazing and browsing
capacities of the specific area under investigation (Vorster 1999). The stocking
density is an estimate of an allowable land to animal relationship which would provide
the most beneficial returns in terms of a given management objective.
The ecological capacity is a product of the quantity and quality of the natural
resources present, while the stocking density is based on personal preference and
the objectives of an area (Von Holdt 1999). The stocking density should, however,
never exceed the ecological capacity (Bothma et al. 2004) and should preferably be
conservative to allow for variable rainfall conditions and changes in the quality and
quantity of the natural resources in an area during the critical time of the year.
Veld condition assessments should thus be conducted on a regular basis to evaluate
the vegetation’s response to current management practices such as the stocking
density (Donaldson and Vorster 1989; Trollope 1990). Management practices can
then be adapted, if necessary, according to the observed trends in the different plant
communities.
Animal-plant interactions in terms of habitat preference, grazing ecology and the
feeding category of a specific herbivore species, play an important role in the setting
of appropriate stocking densities. Some herbivores posses the ability to change
vegetation in order to provide in their specific habitat needs (Van Rooyen 2002).
224
Specific animals prefer specific types and structures of vegetation. These
preferences usually correlate with the various anatomical adaptations of a herbivore
species, which enables it to optimally utilise the preferred stratum or structure of a
vegetation type. The feeding habit of one animal can change the vegetation to such
an extent that it is more suitable for another herbivore species with different feeding
requirements and preferences (Van Rooyen 2002).
To ensure the continued survival and coexistence of the black and blue wildebeest at
Ezemvelo Nature Reserve it was necessary to determine the quality and quantity of
grazing available to the grazers at Ezemvelo Nature Reserve. The aims of this study
were therefore to determine the potential grazing capacity and hence to see whether
the current stocking density in the study area was optimal or whether it was
overstocked. It was hypothesised that Ezemvelo Nature Reserve was overstocked in
terms of the number of grazing wildlife present.
The objectives of this study were:
•
To quantify the frequency of occurrence of the different ecological classes of
the grasses present in order to compile a veld condition index for every
habitat type
•
To use the data from the veld condition index to calculate a realistic prediction
of the grazing capacity for every habitat type
•
To determine the potential black and blue wildebeest stocking density at
Ezemvelo Nature Reserve based on the grazing capacity, veld condition and
stocking density of the present wildlife in the reserve during a year with a
mean or near mean rainfall.
METHODS
Veld-condition assessment
A veld condition index was calculated according to procedures laid out by Bothma et
al. (2004). To determine the percentage of occurrence of each grass species in the
herbaceous layer for each habitat type, the step–point technique as described by
Donaldson and Vorster (1989) and Vorster (1982) was used. At each site a transect
consisting of two parallel lines of 200 m long and 20 m apart along a north to south
direction were surveyed. At every second pace, the end of a measuring staff was
grounded and the grass plant that was nearest to the point was identified. If no
225
herbaceous plant occurred within 0.5 m of the step-point, the point was classified as
bare soil. The grass species were grouped into five ecological classes as described
by Bothma et al. (2004) and as repeated in Chapter 5. All non-grassy herbaceous
species were classified as forbs. These classes were based on a subjective
assessment of the grazing value, phytomass production and palatability and
response to grazing of each grass species (Bothma et al. 2004).
A total of 102 step–point surveys were conducted throughout the five habitat types
that were available to the black and blue wildebeest at Ezemvelo Nature Reserve in
order to get a reliable estimation of the relative frequency of the different grass
species and the cover of the herbaceous layer. These habitat types covered a large
proportion of the reserve and most of the other wildlife present utilised these five
habitat types almost exclusively. The remaining areas on the reserve were
inaccessible to most types of wildlife accept for perhaps the greater kudu, due to
steep slopes and dense vegetation.
Once every grass species was allocated to a specific ecological status class, the total
percentage frequency for every ecological class in each sample plot could be
calculated from data collected during the step–point survey. This percentage was
multiplied by the grazing value of that specific ecological class. These weighted
constants were 10 for class 1, 7 for class 2, 5 for class 3, 4 for class 4, 1 for class 5
and 0 for bare soil (Bothma et al. 2004). When added, the sum of all the values
calculated for every ecological class gave a veld condition score with a maximum
value of 1000.
The veld condition score was converted to a veld condition index by expressing it as
a percentage of the maximum score. Veld condition indices can be interpreted as
follows (Bothma et al. 2004):
•
<40%
- veld in an extremely poor condition
•
40–59%
- veld in a poor to moderate condition
•
60–80%
- veld in a good condition
•
>80%
- veld in an excellent condition
226
Grazing capacity
Grazing capacity refers to the production potential of veld and was originally defined
as the area of veld needed to sustain a livestock unit for a year in a good productive
condition without being detrimental to the vegetation (Fourie et al. 1985). Grazing
capacity can therefore be expressed either as hectare per Large Stock Unit (LSU), or
as Large Stock Units per hectare. In the past this concept was also applied in wildlife
ranching (Bothma et al. 2004).
Recently, grazing capacity for wildlife has been redefined by Bothma et al. (2004) for
better application in wildlife ranching. The concept reflects the ecological production
potential of the grazeable portion of a homogeneous vegetation unit. Grazing
capacity for wildlife is expressed as the area of land (hectares) that is required to
maintain a single Grazer Unit over an extended number of years without deterioration
to the vegetation or soil. A Grazer Unit refers to a blue wildebeest with a mass of 180
kg. The grazing capacity for wildlife is expressed as the number of Grazer Units per
100 hectares.
The ecological capacity for herbivores of a habitat generally refers to the maximum
number of grazers that the given habitat can sustain (Bothma et al. 2004). When
stocked at ecological capacity, neither the animals nor the vegetation on an area will
be in a particular good condition (Behnke and Scoones 1993). Therefore, it is often
reduced by 20 to 30% to achieve an economic grazing capacity (Bothma et al. 2004).
The equation as used in the present study to calculate grazer units is based on the
relationship between the recent and mean annual rainfall, veld condition, percentage
grass cover, habitat accessibility and the influence of fire on plant production and is
described as follows (Bothma et al. 2004):
Grazer Units per 100 ha = 0.547 * [(c+(r-500) * 0.23 * a * f * (log (g)-1)0.4]
c
=
veld condition index
r
=
rainfall over the past two years at the site (mm)
g
=
percentage grass cover
a
=
accessibility of habitat to plains wildlife on a scale of 0.1 to
1, with 0.1 = totally inaccessible and 1 = totally accessible
f
=
fire factor on a scale of 0.8 to 1, with 1 = absence of fire
227
500
=
mean annual rainfall (mm) for the larger region based on long-
term means for the rocky highveld grassland vegetation type (Low and
Rebelo 1996).
A mean annual rainfall of 675 mm was used for Ezemvelo Nature Reserve in the
calculation. The accessibility of the habitat to plains wildlife was taken as 1 on the
scale as explained above for basically all the terrain described by the five habitat
types identified to be available to both types of wildebeest on Ezemvelo Nature
Reserve is accessible for all wildlife. Since fire has been excluded from most habitats
for more than 3 years on Ezemvelo Nature Reserve, the fire factor (f) was set at 1 for
the sandy grasslands, old lands and moist grasslands. An accidental fire passed
through a large proportion of the rocky grasslands and Burkea woodlands in late
2004 and hence the fire factor was set at 0.8 for these two habitat types
Stocking density
A preliminary stocking density for the habitat types potentially available to the black
and blue wildebeest at Ezemvelo Nature Reserve was calculated from the grazing
capacity of the potential wildebeest habitats.
Herbivores were categorised according to their feeding behaviour, into the following
categories (Van Hoven 2002):
•
Bulk or low selectivity feeders
•
Highly selectivity feeders
•
Mixed feeders
•
Browsers
The feeding category of a specific animal is a good prediction of the ratio of
browse:graze in the animal’s diet. Bulk feeders mostly graze, whereas browsers feed
almost exclusively on leaves, twigs and seedpods of fodder trees. Bulk and browse
feeders are the two extremes, with concentrate- and mixed feeders in between.
However, the graze:browse ratio varies considerably between species. Therefore,
following Bothma et al. (2004) the percentage grazing and browsing in the diet was
considered and GU equivalents calculated based on the relevant diets to set stocking
densities for each type of herbivore.
228
For productive wildlife ranching the economic grazing capacity for an area is usually
set at 70% of the ecological grazing capacity (Bothma et al. 2004). Therefore the
economic grazing capacity was also calculated for the present study.
For comparison the rainfall method of Coe et al. (1976) was applied to calculate the
possible large herbivore biomass in kg/m2 based on a mean annual rainfall of 675
mm for Ezemvelo Nature Reserve utilising the following equation:
Large herbivore biomass (kg/km2) = 8.684 x mean annual rainfall (mm) – 1205.9
RESULTS
Ecological grazing capacity
All the grass species recorded and identified on Ezemvelo Nature Reserve during the
step–point survey were classified into their different ecological classes and listed in
Appendix 3.
The characteristics of the five habitat types that were utilised by the black and blue
wildebeest at Ezemvelo Nature Reserve have been described in Table 11.1 The veld
condition index of each of these habitat types ranged from 31 to 45%, indicating that
the overall veld condition at Ezemvelo Nature Reserve was poor. The Burkea
woodlands had the highest veld condition score (448), while moist grasslands had
the lowest veld condition score (313).
According to the rainfall method of Coe et al. (1976), a total of 26 GU/100 ha could
be supported at Ezemvelo Nature Reserve. This estimate is a first approximation of
the ecological capacity of an area for herbivores and does not consider local
temporal and spatial variations in the habitats within a specified area (Van Rooyen
2002). Therefore, the ecological grazing capacity was calculated separately for each
of the five habitat types available to the wildlife at Ezemvelo Nature Reserve by using
the method of Bothma et al. 2004. The value obtained was decreased by 30% to
provide an economic grazing capacity for Ezemvelo Nature Reserve.
229
Table 11.1: Veld condition index and ecological grazing capacity calculation in
Grazer Units (GU) for Ezemvelo Nature Reserve, a grassland reserve in South
Africa, based on the condition of the vegetation in 2004 and calculated by using the
methods that were described by Bothma et al. 2004
Characteristics
Habitat types
SG
RG
OL
2933
2540
744
658
123
Class 1
7
19
18
4
26
Class 2
12
19
2
17
3
Class 3
0
0
0
0
0
Class 4
43
17
44
25
32
Class 5
33
22
28
49
37
Bare soil
4
21
6
2
5
Veld condition score (maximum 1000)
363
414
418
313
448
Veld condition index (%)
36.3
41.4
41.8
31.3
44.8
Grass cover (%)
80
60
72
90
83
675
675
675
675
675
1
1
1
1
1
1
0.8
1
1
0.8
GU/100 ha
28.1
22.6
29.6
26.9
25.2
Total GU
825
575
220
177
31
Size (ha)
Contribution of the ecological classes
Fire factor
BW
a
Mean rainfall (mm/year)
Topography index of accessibility
MG
b
c
Economic grazing capacity at mean annual rainfall:
SG = Sandy grasslands, RG = Rocky grasslands, OL = Old lands, MG = Moist grasslands, BW = Burkea woodlands
a
Ecological classes
1.
Valuable and palatable tufted or stoloniferous grass species with a high productivity and high grazing value
2.
Tufted, perennial grass species with an intermediate productivity and moderate grazing value
3.
Tufted, tall perennial grass species with a high productivity but low grazing value
4.
Generally unpalatable annual and perennial tufted or stoloniferous grass species with an intermediate
productivity and low grazing value
5.
Unpalatable annual grass and forb species with an intermediate productivity and low grazing value
b
Topography index of accessibility: 0.1 = Inaccessible to plains wildlife, 1.0 = fully accessible to plains wildlife
c
Fire factor: 0.8 = recent fires; 1.0 = No recent fires.
230
The old lands could therefore support the most Grazer Units per 100 hectares (30
GU/100 ha) of all the habitats available. For all the habitat types available to both
types of wildebeest in the study area, a total of 1 823 GU could be supported, giving
a mean economic grazing capacity of 27 GU/100 ha for Ezemvelo Nature Reserve
(Table 11.1). This value compares well with the recommendation obtained from the
rainfall equation suggested by Coe et al. (1976).
It was estimated that these five habitat types encompassed 90% of the grazing
available in the reserve for utilisation by grazing herbivores. The remaining habitats
(rocky slopes and riverine vegetation) that were considered to be inaccessible to the
majority of the wildlife on the reserve, consisted mainly of steeply sloping rocky areas
covered in shrubs and trees and dense riverine vegetation. These inaccessible were
not included in any calculations of the grazing capacity in this part of the present
study.
Stocking density
During the study period Ezemvelo Nature Reserve was stocked at 26 GU/ha (1475
GU) (Table 11.2). This figure compares favourably with the economic grazing
capacity of 1823 GU for the accessible habitats. The wildlife at Ezemvelo Nature
Reserve that utilised the grazing resources were made up of 34% low selectivity
feeders, 38% high selectivity feeders and 28% mixed feeders. The Burchell’s zebras
at Ezemvelo Nature Reserve, which are bulk feeders, utilised 39% of the available
grazing capacity. Black wildebeest utilised 4% of the available grazing, while the blue
wildebeest utilised 12% of the available grazing in the study area. Therefore black
and blue wildebeest combined utilised 16% of the available grazing in the study area.
DISCUSSION
The recommended stocking density of 1 828 GU for Ezemvelo Nature Reserve was
an estimate and the veld condition, rainfall and physical condition of the wildlife
should be monitored to make fine adjustments through active adaptive management
(Bothma et al. 2004). This should be repeated yearly. By comparing the actual
wildlife numbers present on the reserve with those that are recommended based on
the available plant resources, any overstocking observed can be corrected (Bothma
et al. 2004).
231
Table 11.2: Estimated current numbers of herbivore grazers after the calving season,
and stocking densities calculated for the herbivore grazers at Ezemvelo Nature
Reserve for December 2004
Type of wildlife
Number of
Grasses in
Mean
Number of
Grazer units
Number
Percentage
animals
the diet (%)
mass
grazing
(GU / animal)
of. GU
of grazing
(kg)
animals
capacity
Low selective grazers
White rhinoceros
2
100
1727
2
5.5
11
1
Burchell’s zebra
583
93
260
542
1.32
716
39
Ostrich
93
80
69
74
0.5
37
2
Total
678
764
42
High selective grazers
Blesbok
166
90
65
149
0.5
75
4
Blue wildebeest
250
87
180
218
1.0
218
12
Black wildebeest
93
90
160
84
0.8
67
4
Gemsbok
18
75
210
14
1.1
15
1
Red hartebeest
178
75
120
134
0.7
93
5
Waterbuck
62
84
205
52
1.1
57
3
Total
767
525
29
Mixed Feeders
Springbuck
96
32
37
31
0.3
9
1
Common eland
121
50
460
61
2.0
121
6
Impala
249
45
41
112
0.3
34
2
Common warthog
106
70
30
74
0.3
22
1
Total
572
186
10
All grazing herbivores
2017
1475
81
232
The vegetation at Ezemvelo Nature Reserve is classified as grassland with
Bankenveld areas in the koppies. This means that there will be little browse available
to browsers on the reserve, especially during the winter (dormant season) when the
trees lose their leaves. The number and type of wildlife that can be supported at
Ezemvelo Nature depends on the availability of browse during the dormant season.
The present wildlife on the reserve makes up 589 BU, which would require a
browsing capacity for the reserve of 8 BU/100ha. The only habitat types that would
support this browsing pressure would be the rocky grasslands, Burkea woodlands
and the rocky slopes and the riverine vegetation habitats, which make up 38% of the
surface area of the reserve.
The calculations as stipulated here indicate that Ezemvelo Nature Reserve is
presently under-stocked. These values should be examined with caution, as there
are large areas within the reserve where the grazing capacity would be considered to
be almost zero. There is a large amount of evidence indicating that large portions of
the habitat available to the grazing herbivores are not suitable. These include:
•
The appearance of distinct overgrazed patches in many of the old lands in the
study area
•
The loss of condition of the high-selective feeders during the dormant season.
Approximately 35% of the sandy grasslands are unavailable for grazing due to the
presence of dense stands of Stoebe vulgaris in this habitat type. This would severely
reduce the grazing capacity of this habitat and needs consideration. If this factor is
taken into account then the grazing capacity of the reserve will be reduced to 1 539
GU. Other factors that would have a negative influence on the grazing capacity
calculation include the following:
•
The encroachment of Stoebe vulgaris in many of the previously overgrazed
grassland areas of the reserve. This encroachment is estimated at 35% of the
grassland areas from aerial photographs
•
The rocky grasslands constituted 2 540 ha of the study area (36%) and were
not readily utilised by the black and blue wildebeest (Chapter 5) due to the
relative inaccessibility of the terrain within this habitat type
•
The Burkea woodlands constituted only 2% of the area available to the black
and blue wildebeest but were heavily utilised by the blue wildebeest due to its
high grazing value and provision of cover. It therefore suffered severe
233
overgrazing by wildlife at Ezemvelo Nature Reserve, making it unavailable
during the dormant season.
•
Large areas of the study area within the delineated habitats were invaded by
alien plants such as Acacia mearnsii and Acacia dealbata. These areas were
estimated to cover 25% of the area that would otherwise be available to
grazers.
Black and blue wildebeest only formed part of the grazing community at Ezemvelo
Nature Reserve, but their stocking densities in the study area could have an
influence on the ability of the two types of wildebeest to co-exist as well as an
influence on the condition of the habitats which they utilize.
Black wildebeest
Black wildebeest prefer open grassland areas where water is freely available. They
prefer short grass, both for feeding and visibility (Apps 1996). They are natural
migraters and pose a severe threat of patch overutilisation when their numbers are
too high and they are confined to fenced land. Black wildebeest can also be
productive with a mean population growth rate of 30 to 38% (Bothma et al. 2002).
Black wildebeest on Ezemvelo Nature Reserve are already posing a threat of patch
overutilisation. Preferred areas within the herd’s territory are being utilised intensively
and are often similar to a blesbok’s preferred grazing spots. There are only a certain
number of suitable habitat areas for the black wildebeest on the reserve. Black
wildebeest are also not dependent on shade during the hot time of the day like most
other herbivore species (Apps 1996).
During the study period, Ezemvelo Nature Reserve was stocked with 93 black
wildebeest, which equates to 74 GU. Territorial behaviour sets a limit to the numbers
of competing black wildebeest that can co-exist in an area. According to Furstenburg
(2002a), the range size for a black wildebeest bull is approximately 400 ha while that
for a cow varies from 200 to 500 ha. A territory for a bull is 2 to 6 ha in size. The
stocking density of black wildebeest should not exceed 10 animals per 100 ha and
the minimum group sex ratio is recommended at one bull for every 10 cows
(Furstenburg 2002a).
234
Blue wildebeest
The habitat of the blue wildebeest consists of open woodland, scrub and grassland,
with access to permanent water (Apps 1996). Blue wildebeest seek shade during the
midday heat and will stand under a tree during this time (Apps 1996; Bothma et al.
2002). They prefer short grass up to 15 cm tall, but browse can consist of up to 13%
of the diet (Bothma et al. 2002). Blue wildebeest are selective of plant parts, but to a
lesser extent of the plant species if the majority of the species are sweetveld species.
In marginal habitat, blue wildebeest will overutilise preferred grass species, which
can result in severe vegetation damage (Furstenburg 2002b).
During the study period, Ezemvelo Nature Reserve was stocked with 250 blue
wildebeest, which equates to 250 GU. According to Furstenburg (2002b), the range
size for a blue wildebeest bull varies from 600 to 1 800 ha while that for a cow varies
from 1000 to 2500 ha. A territory for a bull is 0.5 to 1.5 ha in size. The stocking
density of blue wildebeest should not exceed 7 animals per 100 ha (Furstenburg
2002b) and the minimum group sex ratio is recommended at one bull of more than 4
years for every 6 to 10 cows of more than 2 years of age. Many wildlife species are
limited by territorial behaviour. It sets a limit to the number of competing bulls that
can co-exist in an area (Bothma et al. 2004).
CONCLUSION
The results of this part of the present study indicated that the total ecological grazing
capacity of Ezemvelo Nature Reserve was not exceeded. This conclusion needs to
be considered carefully along with a number of other habitat factors that may
influence the grazing capacity. The populations of black and blue wildebeest at
Ezemvelo Nature Reserve have reached saturation levels based on grazing capacity
and social behaviour. The populations of both types of wildebeest should not be
allowed to increase. However, only through active adaptive management where the
veld condition is monitored and the stocking density adjusted can trends in the
vegetation be related to the stocking density and correction measures be made.
235
CHAPTER 12: POPULATION DYNAMICS
INTRODUCTION
Berryman (1981) defined a population as a group of individuals of the same species
that occur together in the same place and at the same time. The individual animal is
the distinct unit used for providing the basic information to determine the dynamics in
a population of animals (Delany and Happold 1979). The dynamics of a population
are determined from the birth and mortality rates as well as from immigration,
emigration and the interaction of these parameters with the population’s age and sex
ratio. A change in any of these parameters tends to influence population size and the
rates at which these change will impact the rate of increase or decrease in a
population (Herbert 1970; Caughley and Sinclair 1994). Therefore, by noting changes
in the rate of population size change, changes in the fecundity rate, mortality rate and
age distribution of the population may also be identified (Caughley and Sinclair
1994).
In areas where predation is not a significant population regulatory factor, it has been
suggested that the main mechanisms of population regulation in ungulates may be a
reduction in reproductive success or juvenile survival (Turchin 1995). The quality and
quantity of food in an animal’s habitat may affect birth rates as nutritional stress can
lead to decreases in pregnancy rates and rebreeding frequency (Shaw 1985).
Recruitment and overall population changes can be used to measure survival rate
(Krebs 1999).
The sex ratio of a population corresponds with the type of reproduction system and
the bond between the sexes (Leuthold 1977). An imbalance in the sex ratio of
animals often leads to poor mating frequency, especially in species where males
have a harem of females such as impalas and territorial species such as both types
of wildebeest. Age structure is also important since the reproductive potential of an
individual depends on its age (Bothma 2002b).
In nature, populations of wild animals have developed a social structure that
promotes the optimum production of young (Bothma 2002b). However, within fenced
nature reserves, this natural social structure may be modified due to predation or
inadequate food supply (Tambling and Du Toit 2005). Biological monitoring of the
population dynamics of a species is essential to ensure that populations are
236
maintaining demographic and genetic viability within a reserve situation (Walpole et
al. 2001) and is essential for management decisions for a particular species (Bothma
2002b).
Changes in the growth rate of a population may be as a result of the detrimental
effect of interspecific competition acting through a combination of fecundity and
survivorship and the effects are expected to be density dependent (Begon et al.
1996). Therefore, competition between the black and blue wildebeest may result in
the decrease of the weaker competitor, which in the present study is thought to be
the black wildebeest due to its smaller size and area dependence. Evidence of
population decline in one of the types of wildebeest may provide some support for
the hypothesis that competition between the two types of wildebeest at Ezemvelo
Nature Reserve happens and from that, the further inference that no ecological
separation exists between the two types of wildebeest, if all the other factors have
been considered. These other factors may include the possibility that the habitat is
not suitable for one or both types of wildebeest as described in Chapter 1. It was
hypothesised here that the population size of the black wildebeest is negatively
affected by the blue wildebeest population.
The objectives of this part of the study were therefore to:
•
Determine the size and growth rate of the black and blue wildebeest populations
at Ezemvelo Nature Reserve
•
Determine the sex ratio of the black and blue wildebeest populations at Ezemvelo
Nature Reserve
•
Determine the age structure of the black and blue wildebeest populations at
Ezemvelo Nature Reserve
•
Determine the grouping behaviour and mean herd sizes of the black and blue
wildebeest populations at Ezemvelo Nature Reserve
•
Make inferences based on the above results on evidence for competition and
hence evidence for ecological separation between the black and blue wildebeest.
METHODS
A monthly count was conducted on the reserve to count all animals of all types of
wildlife present. Animals were sexed and aged during these counts by using
binoculars. These counts used the road transect technique (Collinson 1985) and
237
provided a repeatable count with repeatable results. This technique, according to
Collinson (1985), is able to provide precise counts but may be inaccurate as the
population size may be under- or overestimated. However, due to the high visibility
on the reserve and the detailed knowledge of all wildebeest herds on the reserve
these inherent inaccuracies of the method were regarded to be negligible in this
study. Additional population data were obtained from the habitat survey data as
described in Chapter 4. These counts provided information on the population sizes
and growth rates, group composition, sex ratios and age structure of the populations
of black and blue wildebeest at Ezemvelo Nature Reserve.
The differences between the sexes in the adults were determined by using a
combination of horn structure, presence or absence of a penile sheath, and general
build (Von Richter 1971a; Attwell 1977; Skinner and Smithers 1990; Estes 1991).
Young calves were difficult to sex and therefore were not sexed for both types of
wildebeest.
Age structure and group composition were established by determining the number of
males, females and juveniles in each group encountered. Juveniles were classified
as such until the next breeding season when new calves were dropped. Therefore,
all animals over 1 year of age were classified as adults during these counts.
Population performance is expressed as the reproductive success rate (Riney 1982)
or annual birth rate (calf:cow ratio) (Caughley and Sinclair 1994). The mean calf:cow
ratio was calculated for both black and blue wildebeest to determine the reproductive
success rates of the populations of black and blue wildebeest at Ezemvelo Nature
Reserve. Calves were taken as less than 1 year of age and adult females as > 1.5
years based on mean first ages for calving by wildebeest. Sexual maturity in blue
wildebeest is reached at 3.5 years and in black wildebeest at 1.2 years of age
(Furstenburg 2002a and b).
RESULTS
Population size and growth
There were 2.4 times more blue wildebeest on Ezemvelo Nature Reserve than there
were black wildebeest (Table 12.1). With their larger body size the blue wildebeest
238
formed 3.2 times more biomass on Ezemvelo Nature Reserve than did the black
wildebeest.
Figure 12.1 shows the trend in population growth of the black and blue wildebeest
over the period May 2003 to August 2005. The blue wildebeest population showed a
finite growth rate of 3% while the black wildebeest population decreased by 2% over
the period. No wildlife capture operations had taken place on Ezemvelo Nature
Reserve since 2002, except for five black wildebeest males which were removed in
July 2004 to be taken to the Voortrekker Monument Reserve in Pretoria as they were
donated by the owners of Ezemvelo Nature Reserve (Tau 2004 pers. comm.)14. Eight
blue wildebeest bulls were culled in 2005, three of these due to injury and the rest for
biltong production.
Sex ratio
The results of investigating sex ratios of adult blue and black wildebeest are
presented in Table 12.2. Sex ratios were based on breeding animals. Since the entire
population of both black and blue wildebeest was known on the reserve, the sex ratio
for the entire population was calculated for the whole study period and the sex ratio
for the separate herds was calculated for each season.
The mean ratio of females per male at Ezemvelo Nature Reserve for the entire
population of both black and blue wildebeest was 1.56:1 during the study period.
Blue wildebeest were observed in herds where the mean female to male ratio was
9.73:1, while the black wildebeest were observed in herds where the mean female to
male ratio was 9.93:1.
14
Mr. M. Tau. Manager. Ezemvelo Nature Reserve. P.O. Box 599, Bronkhorstspruit, 1020,
South Africa. [email protected]
239
Table 12.1: Population statistics for the black and blue wildebeest at Ezemvelo
Nature Reserve in 2005
Wildebeest type
Population
Mean mass
size
(kg)
Biomass (kg)
Percentage
change in
3 years
Black wildebeest
98
150
14 700
-2
Blue wildebeest
236
200
47 200
3
240
300
Numbers
250
200
Blue
150
Black
100
50
M
ay
-0
3
Au
g03
No
v03
Fe
b04
M
ay
-0
4
Au
g04
No
v04
Fe
b05
M
ay
-0
5
Au
g05
0
Date
Figure 12.1: Population trends of the black and blue wildebeest at Ezemvelo Nature
Reserve as obtained from monthly counts from May 2003 to August 2005. The black
lines indicate trend lines, while the blue diamonds inficate blue wildebeest and the
pink squares indicate black wildebeest.
241
Table 12.2: The ratio of females per male in herds of black and blue wildebeest at
Ezemvelo Nature Reserve over the three ecological seasons in 2004 and 2005
Wildebeest type
Late growing
Dormant season
season
Early growing
season
Black wildebeest
9.03:1
10.58:1
10.19:1
Blue wildebeest
9.53:1
11.17:1
8.50:1
242
Age structure
Both types of wildebeest had populations consisting of fewer than 40% young
individuals (Figure 12.2). According to Bothma (2002b) a productive wildlife
population consists of 40% young individuals. As seen in Figure 12.2, 89% of the
black wildebeest population were adults and only 11% were young individuals. The
blue wildebeest population had 83% adults and 17% young individuals.
Grouping behaviour and herd size
Territorial blue wildebeest bulls tended to be more solitary than black wildebeest
territorial bulls. Only 33% of the black wildebeest observations were of a black
wildebeest territorial bull on its own, while 66% of the blue wildebeest observations
were of a single blue wildebeest territorial bull. Therefore, most black wildebeest bull
territories were occupied by female herds and the bull would closely associate with
the herds throughout the year. In contrast, most blue wildebeest bull territories were
not occupied by female herds, and therefore there was generally no close
association between the territorial bulls and the female herds except during the
rutting season.
The herd sizes of black wildebeest tended to remain relatively constant throughout
the year, while the blue wildebeest tended to have a more fluid herd size and groups
tended to separate and regroup on various occasions. Blue wildebeest had larger
herds than black wildebeest in the late growing season and the dormant season.
There was no difference between the herd sizes of black and blue wildebeest in the
early growing season. Black wildebeest herd size did not differ over the seasons and
remained constant at a mean of 12 animals throughout the year. The mean blue
wildebeest herd size was largest during the late growing season (17 animals) and
lowest during the early growing season (12 animals). During the dormant and late
growing season, the black wildebeest maximum herd size was 35 animals, and 39
during the early growing season. Blue wildebeest maximum herd size during the late
growing season was 45, 37 during the dormant season and 42 in the early growing
season (Figure 12.3).
243
Percentage of
population
60
50
40
30
20
10
0
Black wildebeest
Blue wildebeest
Type of wildebeest
Males
Females
Juveniles
Figure 12.2: The broad population structure of the black and blue wildebeest
populations at Ezemvelo Nature Reserve, South Africa for the period January 2004
to August 2005.
244
herd size
50
45
40
35
30
25
20
15
10
5
0
Black
Blue
Black
Blue
Black
Blue
wildebeest wildebeest wildebeest wildebeest wildebeest wildebeest
Late growing season
Dormant season
Early growing season
Season and wildebeest type
mean
min
max
Figure 12.3: Mean, minimum and maximum herd sizes of black and blue wildebeest
over three seasons on Ezemvelo Nature Reserve from January 2004 to August 2005.
The bars on the mean columns represent the standard errors, which are small due to
the large sample sizes that were used for these calculations.
245
Recruitment rate
The mean blue wildebeest calf:cow ratio was 0.52:1 while the black wildebeest had a
much lower mean calf:cow ratio of 0.21:1 (Table 12.3).
DISCUSSION
The current negative rate of increase of the black wildebeest population at Ezemvelo
Nature Reserve suggests that the population has probably reached its optimum
stocking density. Calf mortality through food limitation during the dormant season and
some predation by leopard could be limiting growth in this population.
The calving rate of the black wildebeest was also relatively low (21%). Ezemvelo
Nature Reserve is situated in sour grassland with inferior nutrition during the winter
months and is considered marginal habitat for both black and blue wildebeest.
Populations of black wildebeest in the Giant’s Castle Game Reserve and Golden
Gate National Park, all with marginal habitat, showed a reproductive rate for black
wildebeest of between 47% and 68% (Von Richter 1971a). The values of Von Richter
(1971a) represent a population that at best is maintaining itself. Blue wildebeest at
Ezemvelo Nature show a positive rate of increase of 3% but their low reproductive
rate of 52% is an indication that this population is just managing to sustain itself. Blue
wildebeest in the Serengeti were reported to have a reproductive rate of 96%
(Watson 1969). The Serengeti is in optimal habitat under natural conditions unlike the
habitat at Ezemvelo Nature Reserve.
The breeding potential of a wildlife population can be indicated by its sex ratio (Giles
1978). Under natural circumstances the number of females per male at adulthood for
black wildebeest is 1.5:1 to 2.0:1 and for blue wildebeest it is 1.5:1 to 2.2:1 (Bothma
2002b). At Ezemvelo Nature Reserve the ratio was in accordance with this ratio and
therefore the breeding potential based on the sex ratio is adequate. Estes (1969)
listed a number of territorial species which consistently showed a dominance of
females in a variety of habitats, suggesting that vigorous condition-depleting rutting
activities may have adverse effects on the survival of males during the winter months
(Von Richter 1971a). Unmanaged herds in suitable habitats showed a sex ratio of
close to parity in studies conducted by Von Richter (1971a) while herds in marginal
habitats showed a strong preponderance of males. One would expect a
preponderance of males at Ezemvelo Nature Reserve due to the marginality of its
246
habitats but the results showed that there was a dominance of females over males
for both types of wildebeest. This may indicate that the habitat is not as marginal as
may have been thought. However, the blue wildebeest population did have a greater
percentage of its population being made up of territorial bulls than did the black
wildebeest population (Table 12.3), indicating that there may be more suitable habitat
available and more open niches for blue wildebeest to utilise than there is for black
wildebeest in the study area.
In addition to sex ratio, the age structure is also important for assessing the
productivity of a population. For the population to remain productive, a stable age
structure made up of 30 to 40% young should be maintained in a natural area
(Bothma 2002b and c). The low percentage of young in the black wildebeest
population at Ezemvelo Nature Reserve indicates that the black wildebeest are not
highly productive. The low calf:cow ratio of the black wildebeest could indicate that
the black wildebeest population is mainly made up of old individuals past their
reproductive peak (pers. obs.) and this could be the reason for the declining
population of black wildebeest in the study area. Young animals in the Golden Gate
National Park formed 33% of the population (Von Richter 1971a).
The black and blue wildebeest at Ezemvelo Nature Reserve were strict seasonal
breeders and the majority of the calves were dropped within 3 weeks from the
beginning of December. This agrees with studies on black wildebeest as done by
Von Richter (1971a and b) and on blue wildebeest by Estes (1966).
The black wildebeest population at Ezemvelo Nature Reserve has been decreasing
since both black and blue wildebeest were kept together (Chapter 2). No records on
the population trends of the black wildebeest on the eZemvelo section of the reserve
were available and therefore it cannot be known for sure whether this population was
increasing or decreasing before blue wildebeest began invading the area due to the
lowering of fences.
The reserve has no supplemental feeding programme (Tau 2005 pers. comm.)15 and
only salt licks are distributed throughout the reserve during the dormant season.
15
Mr. M. Tau. Manager, Ezemvelo Nature Reserve. P.O. Box 599, Bronkhorstspruit, 1020,
South Africa. [email protected]
247
Table 12.3: Population size and density of the black and blue wildebeest at Ezemvelo
Nature Reserve as calculated from the mean monthly count data for the entire study
period from January 2004 to August 2005
Item
Black wildebeest
Size of available area (ha)
8 468
Population size
Blue wildebeest
8 468
98
236
0.01
0.03
Wildebeest: percentage of total
6
14
Wildebeest in breeding herds (% of total
66
60
16
8
10
28
Wildebeest per ha
population)
Wildebeest in the bachelor herds (% of total
population)
Territorial bulls (% of total population)
248
Thus being in a marginal habitat with no supplemental food and no management
burns has created a habitat where the black wildebeest population is decreasing.
With the addition of possible competition from the blue wildebeest, which is showing
an increasing population trend, the black wildebeest population at Ezemvelo Nature
Reserve may not recover and may possibly become extinct in the future.
There can be a number of possible reasons for the black wildebeest decline at
Ezemvelo Nature Reserve. These range from food shortages as a result of
competition with other grazers, changed rainfall patterns (Dunham et al. 2003), and
habitat degradation (Harrington et al. 1999) due to Stoebe vulgaris invasion in the
preferred habitat of the black wildebeest. It is unknown whether this decline is
restricted to Ezemvelo Nature Reserve as little information was available from nearby
areas that had populations of black wildebeest. Renosterpoort Nature Reserve,
which is on the southern boundary of Ezemvelo Nature Reserve, also had a decline
in their black wildebeest numbers but this was thought to be due to the initial herd
size that was introduced being too small to form a viable population, rather than there
being any form of food shortage (Anon 2004). No sick animals or carcasses of black
wildebeest were found during the study period in the study area.
There is evidence that an increase in the density of the shrub Stoebe vulgaris could
be contributing to the decline of the black wildebeest at Ezemvelo Nature Reserve.
Shrubs compete with grasses for water and thus reduce the grass standing crop
(O’Connor 1985). Stoebe vulgaris cover has almost doubled over the last 10 years in
the areas inhabited by the black wildebeest (Tau 2004 pers. comm.)16. These shrub
invasions have decreased the area of grassland and increased the grazing pressure
on the grasslands remaining, without a decrease in the stocking density of the wildlife
in the study area. There is no work available on the affect of shrub invasions on the
dry season availability of green grass (Dunham et al. 2003).
The potential competitors for black wildebeest at Ezemvelo Nature Reserve were
blue wildebeest, blesbok, Burchell’s zebra, red hartebeest, common warthog,
common eland, ostrich, springbok and impala as they all tended to occur in the same
habitats as the black wildebeest and grass formed a large proportion of their diets
(Skinner and Chimimba 2005). An increase in the number of one species causing the
16
Mr. M. Tau. Manager, Ezemvelo Nature Reserve. P.O. Box 559, Bronkhorstspruit, 1020,
South Africa. [email protected]
249
decline of another species can be an indication of interspecific competition if there is
overlap in diet, this overlap occurs in a shared habitat and the food supply is limited
(Borner et al. 1987). Black wildebeest numbers were negatively correlated with blue
wildebeest (r = -0.32), Burchell’s zebra (r = -0.36), impala (r = -0.12) and ostrich
(r = -0.28) numbers. None of these correlations were found to be significant.
Therefore it can be concluded that the decline of black wildebeest is not solely
caused by interspecific competition with other grazers. The type of wildlife that the
black wildebeest was most commonly associated with was the blesbok, and the
number of blesbok in the study area was also decreasing over the last 3 years. The
correlation of black wildebeest with blesbok was negative (r = -0.23).
All bones and carcasses of black wildebeest that were found on the reserve were of
old individuals and therefore it was concluded that there was only a limited young
adult black wildebeest mortality on the reserve. The reason for the low recruitment of
black wildebeest should therefore lie in the low fecundity levels. This could be a
result of low nutrition of the adult females and age (Owen-Smith 1990; Harrington et
al. 1999). From the evidence that the veld condition is poor to medium (Chapter 5)
and the invasion of Karoo shrubs that is decreasing the grazing capacity further, it
appears that nutritional stress may well be the reason why the black wildebeest
numbers are declining. Due to their need for an open habitat, they are unable to
move to other areas of the reserve where the shrub invasion is lower because the
visibility, altitude, aspect and openness of these habitats would not be suitable for
their territorial behaviour.
Black and blue wildebeest have coexisted on Ezemvelo Nature Reserve for a
number of years (Chapter 2). Their continued coexistence will rest on their different
habitat requirements and preferences, the correct population size of each of the two
types of wildebeest at Ezemvelo Nature Reserve as well as on habitat improvement
of the areas that have been invaded by Stoebe vulgaris in the preferred habitats of
the black wildebeest.
CONCLUSION
This study investigated possible reasons for the decline of the black wildebeest
population at Ezemvelo Nature Reserve. The possible factors for this decline could
be marginal habitat in association with competition with the blue wildebeest. The
strength of the competition between the black and blue wildebeest has been
250
discussed in chapter 10 and 11 and found to be present but weak in nature. Without
corrective management action the black wildebeest may not be able to withstand
both pressures and may become locally extinct at Ezemvelo Nature Reserve.
However, to substantiate this, more detailed investigations of the age-specific
mortalities and age structure of both types of wildebeest have to be made.
251
CHAPTER 13: MANAGEMENT IMPLICATIONS
INTRODUCTION
Some individuals and agencies operate on the premise that a population’s
persistence can be ensured with the simple protection of a suitable portion of the
preferred habitat of an animal population. However, the conservation of biodiversity,
or any part of it, usually requires active adaptive wildlife management. Wildlife
management is a science and an art of making land produce populations of wildlife
(Bailey 1984) and has confines set by legislation and official policy. Operating on the
“look after itself” premise has resulted in overutilisation and land degradation
(Cromhout 2006). Wildlife population numbers can increase to such levels within
confined reserves that they may change the vegetation composition and cover to the
detriment of other species. This further results in soil erosion followed by a reduction
in grazing for grazer species. Wildlife- proof fences and the lack of predators prevent
natural migrations and natural population degradation and areas with these
characteristics can by no means be considered as natural self-sustaining systems.
Therefore management is essential.
Management can be of two types: conservation management and preservation
management (Thomson 1992). Conservation management includes the sustainable
utilisation of wildlife for the benefit of man, while preservation management includes
the protection of endangered species and does not allow sustainable utilisation of the
wildlife resources in the protected area. Utilisation can be consumptive and nonconsumptive. Ezemvelo Nature Reserve currently obtains most of its income from
ecotourism. This is a non-consumptive type of utilisation. Little consumptive
utilisation of the wildlife has taken place on the reserve over time, except for some
wildlife sales and the harvesting of a few animals for biltong production. The
aesthetic value of healthy wildlife is important from a tourism point of view. A delicate
balance exists in managing for both abundance of healthy large herbivores and
maintaining the scenic beauty of the vegetation. To achieve this balance,
conservation management is required.
The purpose of this section is to present some broad management recommendations
for Ezemvelo Nature Reserve, with specific guidelines for the management of the
black and blue wildebeest. These recommendations are based on the results of the
present study. All the recommendations on stocking densities, habitat manipulation
252
and general management should be made from a balanced holistic perspective (Von
Holdt 1999).
MANAGEMENT OBJECTIVES
Well-defined management objectives are essential for any wildlife management plan
(Mentis and Collinson 1979). Management is futile if unambiguous goals are not
defined. The objectives for the black and blue wildebeest at Ezemvelo Nature
Reserve were to maintain a free-ranging, self-sustaining population of each type of
wildebeest on the reserve while at the same time ensuring that hybridisation between
the two types did not occur at any time. Black and blue wildebeest are waterdependent selective grazers and therefore both water and forage quality may limit
their populations. However, where water is abundantly available, forage quality would
be the main limiting factor. Since the two types of wildebeest have been confined
together artificially, and since they are able to hybridise, special attention should be
given to the active management of both populations of wildebeest.
ACTIVE ADAPTIVE MANAGEMENT
Adaptive management is a term that is used to describe the system of making
management decisions by learning from one’s past mistakes (Stuart-Hill 1989). It is a
useful form of management where management decisions have to be made without
having all the facts at hand. Adaptive veld management depends on three important
monitoring programmes:
•
Recording environmental conditions and the management systems that are
being applied
•
Measuring the performance of the animals
•
Measuring changes in the vegetation.
MONITORING
Monitoring of the habitat aims at the purposeful and repeated examination of the
state or condition of the habitat in relation to external stress, and involves the
frequent testing of the differences between baseline or initial surveys and follow-up
surveys (Bothma and Van Rooyen 2002). Obtaining regular, repeatable ecological
data is important for the successful management and utilisation of any wildlife area.
The aim of monitoring is to observe trends in animal populations and the habitat over
253
time. A monitoring programme serves as an early warning system and it aims to
detect changes or trends that occur as a result of management actions or natural
events. It is important to adapt the management programme in good time when and
where it is necessary. The influence of the management strategy on the following
should be monitored regularly (Bothma and Van Rooyen 2002): veld condition,
grazing capacity, browsing capacity, affects of water provision, affects of bush
encroachment and its control, and the affects of habitat reclamation measures such
as soil erosion control. Opportunistic management has to be undertaken and
continual monitoring and flexibility in the management plan is essential.
Monitoring is the most important aspect of any wildlife management programme and
should therefore be standard procedure on any wildlife area. Monitoring and adaptive
wildlife management go hand in hand and allow wildlife managers to make proper
decisions. Long-term monitoring allows for the measurement of changes over time
and these changes can be evaluated against the area’s long-term objectives to
provide an indication as to whether a specific management action needs to be
altered. The following section discusses the three important monitoring programmes
which should form the basis of a wildlife management plan for Ezemvelo Nature
Reserve.
Environmental monitoring
Certain components of the habitat can be regarded as key components as they are
reliable indicators of the condition of the habitat. Aspects related to these key
components should be monitored regularly. The key components on a wildlife area
are: rainfall, soil erosion, permanent natural surface water, fire, aspects of vegetation
structure, plant biomass production, vegetation cover and composition, and the
productivity, growth rate and numbers of the animal population (Bothma and Van
Rooyen 2002). In practice it is difficult to monitor all these components annually.
Therefore, for practical management purposes the following components should be
regularly monitored: rainfall, temperature, water quality, soil erosion, habitat
(herbaceous component, woody component) and animals.
Rainfall
Long-term rainfall is important for the determination of trends. The graph of the longterm rainfall is supplied for Ezemvelo Nature Reserve (Figure 2.3). Rainfall received
should be monitored daily if possible. Rainfall figures for over 20 years are required
254
to make reliable deductions for a specific area (Bothma and Van Rooyen 2002).
Rainfall has the greatest influence on the productivity of vegetation and ecological
capacity of a reserve (Coe et al. 1976). Rainfall records are important for the
adjustment of stocking densities. There are currently 10 rainfall gauges placed at
strategic spots at Ezemvelo Nature Reserve. Another two rain gauges are
recommended for the high-lying northern plateau areas on the eZemvelo section and
another gauge in the low-lying areas near the bridge crossing the Wilge River to
provide a more even spread of rain gauges throughout the reserve. Accurate records
should be kept and trends in the rainfall patterns analysed regularly.
Temperature
The minimum and maximum temperature should be measured daily at 08:00 at a
standard height of 1 to 2 m above the ground surface in a shaded and well-ventilated
area (Bothma and Van Rooyen 2002). This could be done at the reception area on
the reserve. A Stevenson screen weather station could be set up to record this
information, which could be made available for future researchers and for burning
purposes.
Relative humidity
Relative humidity of the air and air temperature has major affects on fire intensity
(Trollope et al. 2004). The use of these two factors can be used to determine
optimum periods for management burns for the creation of quality forage for grazing
herbivores. This becomes extremely important during the fire season and when
firebreaks are being burnt and will help ensure that runaway fires do to inappropriate
burning times are avoided. Relative humidity of the air can be measured by using a
hygrometer.
Water quality
Inorganic and organic constituents in water can supplement an animal’s mineral
requirements or aggravate/induce a mineral imbalance in an animal (Meyer and
Casey 2002). The susceptibility of toxic and palatability hazards differs from animal
species to animal species.
255
It is recommended that key watering points be monitored on a quarterly basis. The
Institute for Soil, Climate and Water of the Agricultural Research Council17 of South
Africa offers a water analysis package that gives information on a wide range of
water constituents (Meyer and Casey 2002). Water quality assessments should also
be performed regularly.
At Ezemvelo Nature Reserve, there are no artificial watering points at present. The
wildlife utilise the banks of the Wilge River to drink as well as the various streams
that cross the reserve. There are a number of points that are regularly utilised and
the quality of the water at these points should be measured in order to determine
what is present in the water. This information will aid in the setting up of future
supplementation programmes.
Soil erosion
Many wildlife areas experience some form of soil erosion, especially if they originated
from trampled livestock ranches. Soil erosion is generally caused by wind and water.
Wildlife are also known to cause localised erosion. The main culprits are dassies,
arid-zone mongooses, ground squirrels, yellow mongooses, bat-eared foxes,
porcupine, field mice, rats and moles (Snyman 1999). Insects may also be blamed
for some localised erosion. The springbok is one of the herbivores, excluding small
stock, which has the greatest influence on the vegetation. It is responsible for bare
patch formation, pan formation and general erosion (Roux and Opperman 1986). At
Ezemvelo Nature Reserve there are a number of areas that have been severely
eroded due to the action of wildlife. Black wildebeest and blesbok are responsible for
creating bare patches in the sandy grasslands where the herds intensively
concentrate their grazing in one patch for an extended period of time. Blue
wildebeest territorial bulls also tend to create bare patches that could potentially
result in erosion but the population is much less patch selective than the black
wildebeest.
The nature and quality of the vegetation plays an important role in preventing soil
erosion. Vegetation provides a protective layer that is responsible for holding the soil
in place and protecting it against the erosive activity of wind and water. The erosion
process is accelerated if this protective layer is damaged. Stands of perennials are
17
Institute for Soil, Climate and Water, Agricultural Research Council, Private Bag X79,
Pretoria, 0001, South Africa. Tel: 012 310 2500.
256
much more effective than stands of annuals at preventing erosion (Snyman 1999).
Grasses are more effective at preventing erosion than shrubs because grasses have
a larger basal cover and a network of roots close to the surface that bind the soil. In
veld where grass has been replaced by bushes, erosion is prevalent. The areas at
Ezemvelo Nature Reserve that have been invaded by Stoebe vulgaris may become
vulnerable to erosion in the future due to the reduction in the grass cover. This is
another reason why this invader should be controlled in the study area.
Different soil types react differently to erosion. Van Schalkwyk (1984) lists the main
factors that have an influence on the erodability of a soil as:
•
Texture: soils containing a high percentage of fine sands and silts are more
erodable than those with a high percentage of clay and coarse sands.
•
Structure: soils with a coarse blocky, platey or massive structure are more
prone to erosion than those with a fine granular structure.
•
Organic material: soils with a high organic content are more resistant to
erosion than those with a low organic content. The organic matter in the soil is
important for soil structure and water infiltration and it has been found that in
semi-arid rangelands it is normally below 2.5% and decreases with veld
degradation and increased aridity (Snyman 1999).
•
Profile permeability: soils with a high permeability are more resistant to
erosion than those with a low permeability.
The deep sandy soils at Ezemvelo Nature Reserve have a high percentage of sands
and are thus expected to be easily erodable. The soils are adequately aerated and
consist of a loose granular structure that is easily affected by wind erosion. The
organic material content in the soils at Ezemvelo Nature Reserve is generally low, as
the soils tend to dry out quickly and not retain their moisture. The profile permeability
is high due to the sandy nature of the soils, which causes the soils to be highly
leached.
Soil erosion at Ezemvelo Nature Reserve varies from moderate to severe. The main
types of erosion include sheet erosion, gully erosion and channel erosion. Bare
patches are the starting point of sheet erosion and erosion gullies. Episodic floods
sometimes contribute to serious sheet erosion, the incision of pediments, and gully
formation.
257
Gullies are visible manifestations of land-use malpractices and lead to increased
denudation of the soil and increased runoff (Van Schalkwyk 1984). Gullies are
caused by the destruction of the vegetation in drainage ways by fires or overgrazing
and animal trails, labour paths or vehicle tracks. Dongas usually occur near the
bottom of slopes. There are a number of deep gullies at Ezemvelo Nature Reserve.
Serious attention needs to be paid to these gullies. It is recommended that the
wildlife be attracted to the edges of these gullies with molasses so that their hooves
can break down the sides of these gullies. They should be filled with debris and
stabilised with gabions. A programme needs to be implemented for the stabilisation
of the gullies at Ezemvelo Nature Reserve to prevent them from increasing in size.
The consequences of erosion can be measured by the development of
microtopography: sheet erosion, gulley and drainage systems, accumulation of
sediment, decrease in soil fertility, changes in soil structure and texture, changes in
soil moisture status, salinization and compaction of the soil, water runoff and lowering
of the water table (Roux and Opperman 1986). These factors should be measured
regularly at Ezemvelo Nature Reserve.
The management of the reserve should do all that is possible to prevent man-made
erosion. It is important to have knowledge of the areas on the ranch that may be
sensitive to erosion, as well as areas that are already degraded due to erosion. Veld
reclamation programmes should be in place for the prevention and reclamation of
eroded areas. Roads should not be placed in erosion sensitive areas of the reserve.
Vehicles should only be allowed on the designated roads. Bare patches and any
signs of erosion should be carefully monitored so that any erosion can be prevented
in the early stages of development. To ensure the continued stability of an
ecosystem, the loss of abiotic components such as mineral nutrients (and soil) must
be minimised (Bothma 1996).
Habitat monitoring
To conduct monitoring surveys of the habitats, fixed points sites should be
established in each habitat. Thirty-four fixed points were established during the
present study. It is recommended that at least three representative sites from these
34 sites be selected in each habitat at Ezemvelo Nature Reserve. The species
composition and biomass of the herbaceous vegetation should be surveyed annually.
The same methods as described in the present study should be utilised to provide
258
information on these parameters. Such data will also provide information on fuel load
for fires and allow the detection of any changes in species composition. Aerial
photographs taken at constant intervals and altitudes will also provide an effective
way of evaluating the impact of herbivores on the vegetation over time. Fixed-point
photography at the selected sites can also be used to provide a subjective evaluation
of the trends in vegetation of the area over time.
Monitoring of wildlife
In order to have a thorough knowledge of the wildlife populations at Ezemvelo Nature
Reserve a number of factors related to the wildlife populations need to be measured.
The seasonal distribution and numbers of wildlife should be recorded continually and
population growth rates calculated (Bothma and Van Rooyen 2002). The age and
sex ratio of the animals should be monitored annually, as well as the natural rate of
mortality and the cause of death. The physical condition (at the end of the wet and
dry seasons), diseases and parasites of animals should also be recorded (Bothma
and Van Rooyen 2002). For the purpose of determining trends in the populations,
repeatable counts should also be conducted.
The current wildlife censusing techniques utilised at Ezemvelo Nature Reserve
should be continued. These should be complemented with annual aerial counts and
drive counts in the rocky areas for rare animals. Night spotlight counts for nocturnal
species should also be conducted regularly. These counts will assist in determining
population trends of the wildlife. To reduce the growth rate of a wildlife population,
the most productive females and the mature males could be removed. The
populations of black and blue wildebeest need to be monitored carefully and the
growth rate of the blue wildebeest population curbed.
It is also important to monitor the habitats that are preferred by the different types of
wildlife in an area. Seasonal wildlife movements can have a considerable impact on
the habitat and knowledge of such movements will aid in setting realistic stocking
densities.
With Ezemvelo Nature Reserve not being optimal habitat for both black and blue
wildebeest, black and blue wildebeest may be more susceptible to disruptive
pressures especially in confined areas where they cannot escape harmful ecological
factors. It would be advantageous if some form of rotational grazing could be
259
encouraged. Local movements at Ezemvelo Nature Reserve tended to be controlled
by availability and quality of forage. Selection by the blue wildebeest was further
enhanced by the presence of burnt areas. The breeding female component of the
population is the most important component (Caughley and Sinclair 1994). They
often select the best grazing areas while the bachelor herds are forced into suboptimal areas by the territorial bulls. Habitat selection by the female herds would
therefore be a good indication of optimal habitats within the confines of the reserve.
The present study showed that the female herds of the blue wildebeest were highly
selective for the old lands and the black wildebeest for the sandy grasslands during
the early growing season. To optimise Ezemvelo Nature Reserve for the coexistence
of the black and blue wildebeest, the vegetation should be managed specifically to
suit the needs of both types of wildebeest. The habitats should be monitored for
structural and nutritional adequacy (Dörgeloh 1998). Patch burning throughout winter
would improve forage quality during the critical dry period.
A population viability analysis should be conducted on the black wildebeest at
Ezemvelo Nature Reserve. A comprehensive and accurate population composition
analysis needs to be conducted. Herds could be attracted to an artificial feeding site
during the winter of each year and age-specific sex ratios, fecundities and mortalities
monitored. Replacement of some animals with unrelated subadult females to
maintain the demographic balance and genetic variation should be implemented.
SPECIFIC MANAGEMENT RECOMMENDATIONS
Ecological capacity and stocking density
As was discussed in detail in Chapter 11, the ecological grazing capacity of
Ezemvelo Nature Reserve has not yet been exceeded. However, the populations of
the grazing ungulates need to be monitored to ensure that the capacity is not
exceeded in the future. The browsing component in the study area is limited and a
detailed study of the browsing capacity needs to be conducted to determine stocking
densities for the browsers that are present on the reserve. The browsing component
on the reserve is currently made up of mainly greater kudu and common eland.
These animals tend to jump the fences in winter to search out areas where browse is
more readily available. If adjustments to the stocking density are to be made the first
animals that could be removed include Burchell’s zebra, blue wildebeest, red
hartebeest and common eland.
260
Habitat manipulation
It is clear that some form of ecological separation exists between the black and blue
wildebeest at Ezemvelo Nature Reserve. In addition, the loss of grazing areas due to
the encroachment of Stoebe vulgaris and the increase in the blue wildebeest
population has also led to the increased potential for competition for food resources
between the two types of wildebeest. Keeping in mind that one of the objectives of
Ezemvelo Nature Reserve is to prevent hybridisation between the black and blue
wildebeest, such competition should be removed. This can be done by maintaining
conservative stocking densities, by habitat manipulation or by a combination of both
methods.
In addition, manipulation of the habitat to suite specific species can be considered.
This would involve removing trees to provide more grassland for grazers and also the
mowing of grassland to simulate fire. Habitat diversity is essential for providing a
variety of habitats for different species of wildlife and the greater the habitat diversity,
the more the different types of wildlife that can be kept on the reserve.
The following specific steps are recommended to assist in the manipulation of certain
habitats on Ezemvelo Nature Reserve so as to increase the grazing capacity of the
area and to reduce the potential for competition between the black and blue
wildebeest:
•
The most important action is the control of the blue wildebeest stocking
densities as well as the total grazer stocking densities. Specific
recommendations have been given in Chapter 11.
•
Fire can be used as a habitat manipulation tool by drawing animals away
from overgrazed patches and creating areas of better quality forage during
critical times of the year.
•
The enlargement of the property may provide further suitable habitat for the
black wildebeest if chosen in such a way.
•
A Stoebe vulgaris control and monitoring programme should be implemented
in an attempt to regain the large portions of grazing land lost to this
encroaching shrub.
261
•
Roads have important ecological effects in any wildlife area and therefore
consideration needs to be given to road placement and construction as well
as road use (Du Toit and Van Rooyen 1996). A road is a disturbance to the
natural vegetation as it compacts the soil, increases runoff in such areas and
causes soil erosion in those soils prone to erosion (Du Toit and Van Rooyen
1996). Poorly planned roads lead to soil erosion and habitat degradation.
They also lead to disturbance of the wildlife in certain areas.
•
Removal of all alien and problem plants from waterways and from open
areas should be conducted to improve the grazing capacity in these areas.
The plants of concern are black wattle Acacia mearnsii, grey poplar Populus
x canescens, silver wattle Acacia dealbata, Argemone ochroleuca, and
Sesbania punicea.
•
The survival of the Burkea woodlands shoud be ensured through their
protection from fire, and their establishment should be encouraged.
Genetics management and hybridisation
Inbreeding leads to the loss of genetic fitness, increased mortality in young animals,
reduced fertility and depressed growth (Du Toit et al. 2002). The number of breeding
animals in the herd influences the rate of inbreeding at each generation. The sex
ratio also plays an important role in the flow of genetic material in a population (Du
Toit et al. 2002). For healthy population growth to occur, a genetically viable
population is essential. It therefore, remains sound policy to obtain breeding males
from another genetic source from time to time as reserve fences prevent the
exchange of genetic material between animals of bordering reserves.
The random nature of genetics and the lack of previous genetic studies on wildlife,
makes it extremely difficult for researchers to recommend genetic management
policies (Bothma 2002a). The genetic norm among different species varies greatly
and some animals are naturally interbred. In formulating a management strategy for
the maintenance of genetic diversity in wildlife populations on reserves, it is
suggested that the conservation of pure populations should take precedence over the
maintenance of high diversity values (Anon 2003a). Genetic diversity should be
considered as an integral part of biodiversity. Maintaining large-scale biodiversity is
essential in keeping an ecosystem healthy.
262
The principle of managing several smaller populations as a meta-population with
artificially induced gene flow remains a viable strategy (Grobler 2003). Gene flow
needs to be maintained in order to prevent inbreeding and thus avoid detrimental
genetic effects. The first step towards meta-population management is to identify
suitable subpopulations elsewhere (Bothma 2002a). Artificially induced gene flow
requires a critical selection of which populations should be part of the programme.
Ideally, only populations that have been analysed through molecular methods and
proven to be pure individuals, should be used for new genetic material when
exchanging animals between different populations (Grobler 2003). The smaller an
animal population, the more frequently would the stock have to be translocated to
mimic natural migration patterns (Bothma 2002a).
It is therefore recommended that since it is thought that the reason for the low
productivity of the black wildebeest is the large proportion of older animals in the
herds, new young black wildebeest cows and bulls should be brought in to increase
this productivity. This will hopefully halt the population decline of the black wildebeest
at Ezemvelo Nature Reserve while at the same time increasing the genetic diversity
of the population.
Reserve management should only keep those wild animals that are ecologically
adapted to a region and are known to have occurred previously in that region (Du
Toit et al. 2002). This will ensure that competition between ecologically equivalent
animals is eliminated. Therefore it is recommended that only black wildebeest should
be kept at Ezemvelo Nature Reserve.
It is advisable never to mix animals of different subspecies. Wild animals will
hybridise on a wildlife area when the area is too small and minimum herd sizes are
not maintained (Du Toit et al. 2002). Black and blue wildebeest hybridise and
produce fertile hybrids. Red hartebeest and blesbok can also hybridise to give
infertile hybrids.
Supplementary feeding
In order to balance their diets and meet their nutritional requirements, wild herbivores
evolved over time with the behaviour of migrating as the seasons change. However,
today wild herbivores are increasingly being confined to reserves (Maskall and
Thornton 1996; Thornton 2002). Wild herbivores are therefore dependent on the
vegetation in a relatively small area to provide their required nutrients for normal
263
reproduction and physiological processes (Thornton 2002). To balance a system in
terms of nutrient flow (inputs versus outputs) the correct nutrient supplementation on
nature reserves is becoming all the more essential (Maskall and Thornton 1996;
Whitehead 2000). However, the size of Ezemvelo Nature Reserve is sufficient to
allow for some movement of wildlife and supplementary feeding should only become
necessary when the area is overstocked or when the habitat is not suitable for the
particular species that requires the feeding.
Ezemvelo Nature Reserve is characterised by sourveld plant species, which
generally lose their nutritional value during the dormant season. Most of the grass
species on the reserve are unpalatable and stemmy even when available.
Immediately after the rainy season in March, when enough grazing of high quality is
available throughout the reserve, wildlife select their most preferred habitat types in
terms of structure and plant species. At the end of the winter months the wildlife
clearly select the mountainous areas within the reserve. This is especially true for
Burchell’s zebra that congregate on the rocky hills and mountains of the reserve,
utilising just about all the available plant species during the winter months. This
foraging behaviour further indicates that soil in the mountain veld is probably more
fertile than soil in the sandy regions. After the depletion of available food in the
mountain veld areas, wildlife will start selecting plants more for volume than for
quality.
When wildlife are in a poor condition during the winter months some form of
supplementation is required. During the winter months protein and energy are likely
to be deficient in sourveld regions (Schmidt and Snyman 2002). Therefore it is
recommended that protein licks be placed out for the wildlife at Ezemvelo Nature
Reserve.
The substandard quality of the available vegetation for both types of wildebeest could
lead to delayed puberty, resulting cows having their first breeding season later in life.
It could also lead to females taking longer to reach the required target mass for
conception, and a reproductive cycle could be skipped. Calf survival could also be
impeded and abortions or stillbirths are possible. This may be one of the contributing
factors to the decline of the black wildebeest population at Ezemvelo Nature Reserve
causing a low calf:cow ratio as was found in Chapter 12.
264
Salt licks have been regularly provided for the wildlife at Ezemvelo Nature Reserve
during the winter months. These licks are heavily utilised by the wildlife populations
on the reserve indicating that there is a need for some form of supplementation in
order to get the populations through the winter period in a good condition. Since the
main objective of the reserve is for ecotourism, the presence of extremely thin and
unhealthy animals is not recommended. This may detract from the aesthetic beauty
of the reserve.
Disease management
The recent development of the wildlife ranching industry together with the lifting of
restrictions on the movement of wildebeest in 1993 has caused concern amongst
cattle producers with regards to the increased incidence of wildebeest-associated
bovine malignant catarrhal fever in cattle (Cooper 2003).
There are a set of proposed control measures that have been drawn up by the Red
Meat Producers Organisation and the South African Game Ranchers Organisation
(SAGRO) (Cooper 2003). These include that this disease should be declared a
controlled disease under the new Animal Health Act 7 of 2002. All farms presently
keeping wildebeest should be registered through a statutory procedure, to be
prescribed in the Regulations under the new Animal Health Act 7 of 2002. For all new
registrations, applicants must obtain the written consent of all directly adjoining
neighbours. All existing farms, new farms and facilities of agents, auctioneers and
wildlife capturers, where wildebeest are being kept or will be kept, must be
registered. A registration certificate will be issued and will be valid only for the land
specified on the certificate and can be withdrawn if the holder thereof is convicted of
an offence in terms of the new Animal Health Act, concerning the registration,
keeping or the movement of wildebeest. Movement without state veterinary permit
control will be allowed only between farms/holdings registered according to the
prescribed procedure. SAGRO will apply to issue the movement permits and handle
the recording and administration process as an assignee under the Animal Health
Act 7 of 2002. It will be the responsibility of the buyer to produce a registration
certificate before wildebeest can be purchased privately or at an auction.
Ezemvelo Nature Reserve is directly adjacent to, on a number of sides, farms which
carry cattle. Possible conflict may results in the future if the wildebeest at Ezemvelo
Nature Reserve are found to be the cause of a snotsiekte outbreak on adjacent
265
properties. Therefore note should be taken by management of these proposed
controlled managements to avoid any future problems.
Translocation of wildebeest
Currently legislation is also being implemented which controls the movement and
keeping of wildebeest. The National Environmental Biodiversity Act of South AfricaAct 10 of 2004 (Anon 2004) regulates the translocation of indigenous species to
areas where they are locally exotic in South Africa. Translocation of wildlife in South
Africa has become increasingly easy and results in private landowners bringing
animals together into locally exotic areas. Such practices have many associated
risks. These risks include the possibility of outbreeding depression, of hybridisation
between species, between subspecies, and the mixing of ecotypes with the possible
resultant loss of local genetic fitness due to the modification or loss of local gene
adaptations (or alleles). A further great risk is the transmission of diseases and
parasites to areas where they previously did not exist and these pose considerable
threats to the wildlife, and in some instances even to domestic livestock, of the areas
into which they have been introduced. Yet another major risk is that an introduced
taxon often has the potential to cause considerable irreversible ecological damage
whether it is in the form of substrate or habitat destruction, or even outcompeting
local taxa.
By introducing both black and blue wildebeest into a reserve together, the
hybridisation risk is the greatest followed by habitat destruction. There is thus a need
for all those interested in the maintenance of natural biodiversity to establish
guidelines and principles for the translocation of biota so that no one element poses
a conservation threat to any other. The guidelines state that translocated mammalian
herbivore taxa must not threaten the genetic integrity of naturally occurring local taxa,
i.e. they must originate from genetically identifiable and appropriate populations and
not be susceptible to hybridisation. Existing hybrids should be removed. This applies
directly to Ezemvelo Nature Reserve. South African endemic and threatened taxa will
be afforded priority protection and will be restricted to their natural distribution ranges
(for certain taxa extralimital populations may be considered). Taxa may only be
translocated to areas where suitable and adequate habitat exists. Properties where
wild herbivores occur will be accredited by provincial conservation authorities
according to the taxa present on the property, with those supporting only historically
266
appropriate taxa under natural conditions enjoying the highest status and qualifying
for incentives.
Black wildebeest are placed in category 2 of the translocation categories for South
African mammalian herbivores. This means that this type of wildebeest may be
translocated within their natural geographic ranges and conditionally to areas within
South Africa outside their natural range. Only individuals of approved origin will
qualify. These conditions, however, will include appropriate genetic origin, adequate
enclosure and registration as a zoological institution. This category includes a
number of mammalian herbivore taxa that have unfortunately been extensively
translocated in the past to destinations outside their natural ranges, thereby setting
precedents of major numeric proportions, but which can, at least temporarily, be
intensively managed through a variety of conditions, ranging from adequate
enclosure prescriptions to registration of properties as zoological gardens. The aim
with this category of animals is to persuade owners to remove them from the areas
where they do not belong, possibly through the introduction of a system of
incentives/-disincentives. If not, then the properties will be downgraded within the
accreditation system, or will possibly be required to be registered as zoological
gardens.
Blue wildebeest are in category 3. This is the category of South African mammalian
herbivore taxa that may only be translocated within their natural geographic ranges
within South Africa. This category includes mammalian herbivore taxa with a
relatively wide historic range, but which have close relatives elsewhere; this implies
that each related taxon is restricted to its own natural range.
Where blue and black wildebeest; or tsessebe and red hartebeest and / or blesbok;
or greater kudu, nyala and bushbuck; or waterbuck and other Reduncinae,
historically occurred sympatrically, they may no longer be kept on the same property
in order to prevent hybridisation, unless an inspection reveals that in those areas of
sympatry there is sufficient suitable habitat for each taxon, and the property itself is
greater than 10 000 ha. All individuals of the taxon can be freely translocated within
its natural range. Since Ezemvelo Nature Reserve is currently smaller than this size
hybridisation is most definitely a cause for concern.
267
CHAPTER 14: CONCLUSIONS
The overall aim of the present study was to determine whether ecological separation
existed between the black and blue wildebeest at Ezemvelo Nature Reserve. It was
predicted that the black and blue wildebeest would be too close ecologically to be
kept together in the same area without harming each other or the habitat.
Ecological separation was studied in terms of habitat separation at three different
scales: macroscale, mesoscale, and microscale. Separation in habitat use was found
at the macroscale and at the mesoscale but not at the microscale. The type of
vegetation in the different habitats was not the factor governing habitat selection by
the black and blue wildebeest. Instead the physical features of the habitat were the
main driving factors of habitat selection. Factors such as distance to shade, woody
vegetation cover, aspect, and altitude were the most important separating factors.
The diversity of habitats at Ezemvelo Nature Reserve offers mutually exclusive areas
for the black and blue wildebeest thus allowing for effective spatial separation of the
two types of wildebeest on the reserve. As long as these habitats are maintained
intact, the coexistence of the black and blue wildebeest at Ezemvelo Nature Reserve
at the current population levels can be maintained without the threat of hybridisation.
Black wildebeest are willing to trade-off nutritional quality for an open habitat and
therefore may require supplementation in their preferred habitats to ensure the longterm viability of the population. The habitat offering the most high quality forage for
both types of wildebeest is the old lands. This is the habitat where possible conflict
between the black and blue wildebeest may occur during the critical season.
Mesoscale habitat separation is the highest as it is at this scale that the physical
features of the site of occupation become most important. The heightened territorial
behaviour of the black wildebeest as compared to the blue wildebeest makes it
dependent on high-lying open areas. Such areas will be chosen before the quality of
the forage is considered. Therefore, areas with sufficient open areas at high altitudes
and with high visibility are required for black wildebeest. Blue wildebeest require
cover to be in the near vicinity of their feeding sites. Therefore, an area with a mosaic
of open habitats and more densely vegetated areas will provide suitable
circumstances for the coexistence of both types of wildebeest without competition,
provided that the sizes and demographics of both populations are carefully
monitored. The lack of habitat separation at the microscale was expected as it has
been found that there was no difference in the way that the two types of wildebeest
feed and hence no trophic difference between the two types. The feeding sites of the
268
black and blue wildebeest were similar and showed little difference in terms of the
vegetation characteristics that were measured. Slight differences in terms of grass
quantity and grass species composition were found which could be attributed to the
area selective nature of the black wildebeest compared to the more mobile blue
wildebeest. In addition only limited suitable habitat was available for the black
wildebeest on the reserve forcing them to concentrate for longer on certain patches
whereas the more versatile blue wildebeest was able to utilise a wider variety of
habitats.
The differences in the activity patterns of the black and blue wildebeest can be
attributed to the differences in the mobility of the two types of wildebeest and to the
differences in the openness of the habitats selected by either type of wildebeest. Blue
wildebeest were much more active than the black wildebeest and spent less time
resting than the black wildebeest. This could also be attributed to the smaller size
and hence digestive capacity of the black wildebeest as compared to the blue
wildebeest.
Resource partitioning between the two types of wildebeest was found to be
incomplete. Considerable overlap in the use of key resources such as habitats and
possible food species occurs between the black and blue wildebeest. In
homogeneous landscapes with little habitat variation this finding would indicate that
the two types of wildebeest would be in direct competition for their basic resources. If
the study area consisted of only open plains, the black and blue wildebeest would not
be able to coexist without harming each other or the habitat. The minimal overlap in
terms of spatial distribution and temporal activities at Ezemvelo Nature Reserve is a
direct result of the presence of a diversity of habitats that serve to provide mutually
exclusive areas that can be exploited independently by either type of wildebeest.
Seasonal differences in habitat use were identified in the present study. As expected,
overlap in resource use tended to be lowest during the dormant season when food
resources were most limiting. This critical season prevents the members of either
type of wildebeest from expanding their niche dimensions, as only a limited supply of
resources is available. During the other seasons when resources are readily
available it may be possible to exploit a wider breadth of resources but during the
critical season niche breadth decreases and animals become more specialised. In
terms of the possibility for hybridisation, the rutting period in the late growing season
is the most crucial for the implementation of ecological separation between the two
269
types of wildebeest. For hybridisation to be avoided, ecological separation should
therefore be the greatest during the late growing season. Evidence for this was found
in the present study.
Due to the spatial separation of the black and blue wildebeest at Ezemvelo Nature
Reserve, there is little opportunity for interference competition between them.
However, in a homogeneous area, interference competition could become a problem.
The encroachment of blue wildebeest bachelor males into black wildebeest habitat is
a clear indication that the population size of the blue wildebeest has reached
saturation levels in its suitable habitat. Population regulation of the blue wildebeest
population is imperative to ensure that hybridisation does not occur.
The black wildebeest population at Ezemvelo Nature Reserve was found to be
declining. No clear reason for this decline was found, but it was suggested that the
quality of the available suitable habitat was declining due to encroachment by Stoebe
vulgaris, lack of burning causing the build up of moribund material and the sourveld
nature of the vegetation requiring supplementation in the critical season.
It is concluded that the introduction of both types of wildebeest into the same area is
not recommended, but if done requires intensive management to prevent
hybridisation and competition. Only certain areas would be suitable for such an
introduction where the habitat heterogeneity is able to supply a suitable mixture of
open habitats and cover. The owner of such properties has a responsibility to ensure
that hybridisation does not occur and by implementing the recommendations
supplied in this study with continuous monitoring may be able to conserve pure
populations of black and blue wildebeest in the same area. Black and blue
wildebeest are not ecologically separated to such a degree that they will be able to
coexist without management action.
For the situation at Ezemvelo Nature Reserve it is recommended that one of the
types of wildebeest be removed from the reserve. This will require the destruction of
the animals, as live animals can no longer be sold due to provincial regulations
discussed in Chapter 3. Since this action may seem too drastic, the populations at
Ezemvelo Nature Reserve need to be intensively managed to ensure that suitable
habitat is available for both black and blue wildebeest. This will require management
action in terms of Stoebe vulgaris control, patch burning and population control in the
270
form of hunting or culling. The black and blue wildebeest populations at Ezemvelo
Nature Reserve cannot be left to “sort themselves out” as this will inevitably result in
either hybridisation or the loss of the black wildebeest population.
FUTURE RESEARCH PERSPECTIVES
The aim of this section is to briefly outline possible future work associated with the
present study that could be done at Ezemvelo Nature Reserve and on other
properties where black and blue wildebeest are confined together. These include:
•
A detailed study of the dietary requirements of both types of wildebeest
•
Detailed age structure and population dynamics analysis of the black wildebeest
population to determine the exact cause of its decline over the last three years as
found in Chapter 12.
•
Establishment and growth of the even cohorts of Burkea africana in the study
area. There is a need to understand the population dynamics of these trees and
how and why they grow in the areas they do. Ensure their continued survival of
the woodlands as they provide an important habitat for many of the wildlife
species on the reserve.
•
Range size analysis of identified individuals of both types of wildebeest to show
movements and activities and behaviour especially during the rutting season to
continue monitoring the whether any occurrence of interbreeding between the two
types of wildebeest will occur in the future.
•
An in depth study of the ecological separation of the grazers because these
animals compete for the same food resources, especially in terms of quantity
during the dormant season.
•
Genetic studies should be conducted on the black and blue wildebeest to
determine the degree of genetic variation, inbreeding and whether hybridisation
has taken place in the past. This detailed data could be included into a population
viability analysis to predict the viability of the populations on the reserve.
•
A study of coexisting populations of black and blue wildebeest in areas with low
habitat heterogeneity would be able to confirm the conclusions reached in the
present study.
271
PREDICTIONS FOR THE FUTURE OF THE BLACK AND BLUE WILDEBEEST IN
SOUTH AFRICA
National policy has been implemented which will aid in discouraging landowners from
keeping the black and blue wildebeest on the same property and outside of their
natural distribution ranges. Recent press reports have indicated, however, that the
hybridisation problem is not being taken seriously by the game industry as hybrid
wildebeest were sold at a game auction in the Free State province in 2006 (African
Indaba 2006). The genetic history of all populations of black and blue wildebeest is
not known and it is entirely possible that many populations are the result of offspring
of hybrids. This requires serious study. All populations should be analysed for genetic
purity before sales are allowed. As long as there is a market for hybrids, the threat of
hybridisation and doubts for the future of pure black and blue wildebeest populations
in South Africa remain. Hunting regulations discouraging hunting of rare hybrids
should be strongly implemented. The politics surrounding the hunting industry need
regulation. This is currently being implemented at a national level. The loss of
revenue due to keeping both black and blue wildebeest on the same property should
not be offset from the income that may be obtained from hunting a hybrid. It is the
responsibility of the hunting industry to discourage such practices.
The Stern report on global climate change indicates that climate change may also
impact the black and blue wildebeest populations in South Africa. Since black
wildebeest are endemic to South Africa, this type of wildebeest should be given
priority protection. Climate change impacts predict that the highveld grasslands of
South Africa will be encroached by Karoo vegetation in the future due to increased
temperatures and lower rainfall. The decreasing size of the grasslands in South
Africa will decrease the habitat available for black wildebeest. With all factors
increasingly piling up against the black wildebeest all efforts should be put into
conserving this type of wildebeest and the policies surrounding the prevention of
hybridisation should become more strict and implemented on a fine system rather
than an incentive/disincentive scheme.
272
The ecological separation of the black and blue wildebeest on Ezemvelo Nature
Reserve in the highveld grasslands of South Africa
by
Chantal Vinisia Helm
Supervisor: Prof. J. du P. Bothma
Co-supervisor: Prof. M.W. van Rooyen
in the Centre for Wildlife Management
Department of Animal and Wildlife Sciences
Faculty of Natural and Agricultural Sciences
University of Pretoria
Pretoria
MAGISTER SCIENTIAE (WILDLIFE MANAGEMENT)
SUMMARY
This study was conducted at Ezemvelo Nature Reserve on the boundary between
the Gauteng and Mpumalanga provinces in the central grasslands of South Africa.
The reserve covered an area of 8 468 ha. The area forms part of the grassland
biome in the rocky highveld grassland region and receives a mean of 675 mm of
rainfall annually.
The ecological separation of the black and blue wildebeest was investigated with an
emphasis on habitat separation, activity patterns and feeding ecology all within a
seasonal context. Habitat separation was analysed at three scales namely the
macro, meso and microscales. The black and blue wildebeest showed clear resource
partitioning in terms of habitat at the macro and mesoscales, but not a clear
separation at the microscale.
The main factors determining the between the black and blue wildebeest separation
as determined by the application of logistic regression analysis, was distance to
shade, aspect and altitudes indicating that black wildebeest occupy the high-lying
open north facing niches and the blue wildebeest occupy the low-lying, niches which
273
have a high availability of cover for protection from the heat which is not required by
the black wildebeest.
Five broad habitats were delineated through the reserve and the vegetation
characteristics for each habitat were measured. Black wildebeest were found to
utilise the habitats which were the most open in terms of visibility and tree cover and
which were high-lying. Blue wildebeest selected habitats with short grass and a
history of cultivation as well as areas that ensured the close proximity to shade.
The feeding sites of the black and blue wildebeest were also analysed and compared
in terms of their vegetation characteristics utilising discriminant analysis. Of the
vegetation characteristics measured, only biomass and grass height proved to
differentiate between the feeding sites of the black and blue wildebeest. Due to their
similar trophic ecology, it was concluded that the black and blue wildebeest do not
differ in terms of their microhabitat selection.
The activity budgets of the black and blue wildebeest were also compared. The black
wildebeest was found to spend more time resting than the blue wildebeest. This was
found to be due to the higher mobility of the blue wildebeest as compared to the
extreme form of area selectivity practiced by the black wildebeest.
The population dynamics of the black and blue wildebeest was also investigated. It
was found that the black wildebeest population at Ezemvelo Nature Reserve was
declining over the last 3 years. The possible reasons for this decline were due to
suboptimal habitat and thus decreased calf:cow ratios. The blue wildebeest
population on the other hand was found to have been increasing. It was concluded
that efforts needed to be made to prevent the further increase of the blue wildebeest
population on the reserve so as to prevent the further encroachment of blue
wildebeest bachelor males into black wildebeest territory.
Evidence for interspecific exploitative and interference competition was investigated.
Due to the high spatial separation of the black and blue wildebeest at Ezemvelo
Nature Reserve, little interference competition was observed accept for isolated
cases where blue wildebeest bachelor males were encroaching on black wildebeest
territory. Encounters between black and blue wildebeest usually showed that the blue
wildebeest was dominant over the black wildebeest. Exploitative competition was
274
found to be possible due to the high overlap in terms of habitat niche use and feeding
niches.
Ecological separation between the black and blue wildebeest was not found to be
complete. Certain differentiating factors such as a preference for open areas by black
wildebeest and a preference for areas in the vicinity of suitable shade by blue
wildebeest can be utilised to allow for the coexistence of black and blue wildebeest in
an area with a high habitat heterogeneity. Homogeneous areas with low habitat
diversity will not be suitable for the coexistence of black and blue wildebeest as
habitat is the main differentiating mechanism between the two.
It was concluded that without the active management of the black and blue
wildebeest populations at Ezemvelo Nature Reserve, the future of the black
wildebeest population at least is not optimal. In the long term it was predicted that the
black wildebeest population would continue to decline and the blue wildebeest would
continue to increase utilising the habitats previously exclusively occupied by the
black wildebeest. With the increase of alien vegetation providing further shade for the
blue wildebeest in these habitats, this was considered entirely possible. Management
recommendations for the black and blue wildebeest populations at Ezemvelo Nature
Reserve were made and dicussed in detail.
275
ACKNOWLEDGEMENTS
I would like to acknowledge the following people and institutions for their
contributions during my term of study:
Thanks to my supervisor, Prof. J. du P. Bothma of the Centre for Wildlife
Management at the University of Pretoria. His encouragement and support of my
study is much appreciated. I am indebted towards him for his critical examination of
my dissertation. His comments and advice helped improve all aspects of my study.
A special thanks to my co-supervisor, Prof. M.W. van Rooyen from the Department of
Botany at the University of Pretoria. She provided me with endless guidance, support
and advice for the write up of my dissertation and her critical evaluation of my work is
greatly appreciated.
Thanks is extended to the Oppenheimer family who allowed this study to take place
at Ezemvelo Nature Reserve and for free access to this reserve. Thanks also to
Duncan MacFadyen from E. Oppenheimer and Son for helping to get this study
approved and for his logistic support and guidance during my stay on Ezemvelo
Nature Reserve. The accommodation and logistical support on the reserve was
greatly appreciated.
Special thanks to the staff at Ezemvelo Nature Reserve, especially Maroti Tau and
Jackson Kone, the reserve manager and conservationist respectively, who were
always willing to help me while I was in the field and were at all times friendly. Their
assistance in completing this study is greatly appreciated. Thanks is also extended to
them for allowing my involvement in various projects on the reserve during my stay
there and for always making me feel welcome.
I am indebted to Dr. M. van der Linde and Prof. H. Groeneveld from the Department
of Statistics at the University of Pretoria for their statistical support and valuable
advice regarding the statistical analysis of my data with SAS. Many statistics
meetings and data manipulation and analysis with Dr. van der Linde were required to
analyse all the data from this study. His hard work and extended hours put into this
project are greatly appreciated.
276
Thanks to the Centre of Wildlife Management who provided me with a quad bike for
my fieldwork. A special thanks to Liset Swanepoel from the Centre for wildlife
Management for her logistical support, advice and organisational skills.
Thanks to Prof. G.J. Bredenkamp from the Department of Botany at the University of
Pretoria for his advice for the vegetation surveys and to Liesl Beukes for her help in
the field with the vegetation surveys.
The financial assistance of the National Research Foundation (NRF) towards this
research is hereby acknowledged. Without this support, this project would not have
been logistically possible.
Acknowledgment is also extended to the University of Pretoria for financial support.
Appreciation is also extended to my family for their support and encouragement
during my period of study and to Francois Swiegers for his loving support.
277
REFERENCES
(The references have been formatted according to the South African Journal of
Wildlife Research accept for the requirement for abbreviations for the journal titles)
ACOCKS, J.P.H. 1988. Veld types of South Africa (pp. 54-55). Memoirs of the
Botanical Society of South Africa No. 57.
AFRICAN WILDLIFE, 2003. Cross-breeding threatens black wildebeest. African
Wildlife 57: 45.
ALAIS, M. 2000. Population genetic structure of black wildebeest based on
mitochondrial DNA sequences and microsatellite loci. Unpublished B.Sc.
Honours Project, University of Pretoria, Pretoria.
ALTMANN, J. 1973. Observational study of behaviour: Sampling methods. Behaviour
49: 227-265.
ALVAREZ-CARDENAS,
S.,
GUERRO-CARDENAS,
I.,
DIAZ,
S.,
GALINA-
TESSARO, P. & GALLINA, S. 2001. The variables of physical habitat
selection by desert bighorn sheep (Ovis Canadensis weemsi) in the Sierra del
Muchudo, Baja California Sur, Mexico. Journal of Arid Environments 49: 357374.
ANON 2003a. Black wildebeest (Connochaetes gnou) hybridisation action planning
workshop. Unpublished Workshop Report. Conservation Breeding Specialist
Group and the Endangered Wildlife Trust, Bloemfontein.
ANON 2003b. Black wildebeest hybridisation workshop report. African Indaba
Newsletter Vol. 1. No. 5.
ANON 2004. Act number 10 of 2004: National Environmental Management:
Biodiversity Act, 2004. Government Gazette 467 (26436), Government
Printer, Cape Town.
ANSELL, W.F.H. 1974. Order Artiodactyla. In: J. Meester & H.W. Setzer (Eds). The
mammals of Africa: an identification guide (pp. 1-84). Smithsonian Institution
Press, Washington DC:
ANSLOW, R.C. 1966. The rate of appearance of leaves of Graminea. Herbal
Abstracts 36: 149-155.
ANTHONY, R.G. & SMITH, N.S. 1977. Ecological relationships between mule deer
and white-tailed deer in southeastern Arizona. Ecological Monographs 47:
255-277.
APPS, P. 1996 (Ed.). Smithers’ Mammals of Southern Africa: A Field Guide. Second
edition. Swan Hill Press, England.
278
ARCHIBALD, S., BOND, W.J., STOCK, W.D. & FAIRBANKS, D.H.K. 2005. Shaping
the landscape: fire-grazer interactions in an African savanna. Ecological
Applications 15: 96-109.
ARSENAULT, R. & OWEN-SMITH, N. 2002. Facilitation versus competition in
grazing herbivore assemblages. Oikos 97: 313-318.
ATTWELL, C.A.M. 1977. Reproduction and population ecology of the blue wildebeest
Connochaetes taurinus taurinus in Zululand. Ph.D. thesis, University of Natal,
Pietermaritzburg.
AUGUSTINE, D.J., MCNAUGHTON, S.J. & FRANK, D.A. 2003. Feedbacks between
soil nutrients and large herbivores in a managed savanna ecosystem.
Ecological Applications 13: 1325-1337.
BAGCHI, S., GOYAL, S.P. & SANKAR, K. 2003. Niche relationships of an ungulate
assemblage in dry tropical forest. Journal of Mammalogy 84: 981-988.
BAILEY, J.A. 1984. Priniciples of Wildlife Management. John Wiley and Sons, New
York.
BAILEY, D.W., DUMONT, B. & WALLIS DE VRIES, M.F. 1998. Utilisation of
heterogeneous grasslands by domestic herbivores: Theory to management.
Annales de Zootechnie 47: 321-333.
BANCROFT, C.M. 1989. The habitat utilisation of eland, impala and kudu in the
Zemvelo Game Park. Unpublished B.Sc. Honours Report, University of
Pretoria, Pretoria.
BARNES, D.L., ODENDAAL, J.J. & BEUKES, B.H. 1982. Use of the dry-weight-rank
method of botanical analysis in the eastern Transvaal highveld. Proceedings
of the Grassland Society of Southern Africa 17: 79-82.
BARRETT, R.H. 1982. Habitat preferences of feral hogs, deer and cattle on a Sierra
Foothill Range. Journal of Range Management 35: 342-346.
BEARDELL, G.M., JOUBERT, S.C.J. & RETIEF, P.F. 1984. An evaluation of the use
of correspondence analysis of herbivore-habitat selection. South African
Journal of Wildlife Research 14: 79-88.
BEGON, M., MORTIMER, M. & THOMSON, D.J. 1996. Population Ecology: A
Unified Study of Animals and Plants Third Edition. Blackwell Science Ltd.,
Cambridge.
BEHNKE, R.H. & SCOONES, I. 1993. Rethinking range ecology: Implications for
rangeland management in Africa. In: R.H. Behnke, I. Scoones & C. Kerven,
(Eds). Range ecology at disequilibrium: New models of natural variability and
pastoral adaptation in African savannas. Overseas Development Institute,
London.
279
BELL, R.H.V. 1970. The use of the herb layer by grazing ungulates in the Serengeti.
In: A. Watson (Ed.). Animal populations in relation to their food resource. (pp.
111-123). Blackwell Scientific Publications, Oxford.
BELL, R.H.V. 1971. A grazing ecosystem in the Serengeti. Scientific American 225:
86-93.
BEN-SHAHAR, R. 1986. The conservation of endangered ungulates in the
Waterberg, Transvaal. M.Sc. Dissertation, University of Pretoria, Pretoria.
BEN-SHAHAR, R. 1995. Habitat classification in relation to movements and densities
of ungulates in a semi-arid savanna. African Journal of Ecology 33: 50-63.
BEN-SHAHAR, R. & COE, M.J. 1992. The relationship between soil factors, grass
nutrients and the foraging behaviour of wildebeest and zebra. Oecologia 90:
422-428.
BEN-SHAHAR, R. & FAIRALL, N. 1987. Comparison of the diurnal activity patterns
of blue wildebeest and red hartebeest. South African Journal of Wildlife
Research 17: 49-54.
BERRY, H.H. 1980. Behavioural and ecophysiological studies on blue wildebeest
Connochaetes taurinus at the Etosha National Park. Ph.D. Thesis, University
of Cape Town, Cape Town.
BERRY, H.H., SIEGFRIED, W.R. & CROWE, T.M. 1982. Activity patterns in a
population of free-ranging wildebeest Connochaetes taurinus at Etosha
National Park. Zeitschrift fur Tierpsychologie 59: 229-246.
BERRY, H.H., SIEGFRIED, W.R. & CROWE, T.M. 1984. Orientation of wildebeest in
relation to sun angle and wind direction. Modaqua 13: 297-301.
BERRYMAN, A.A. 1981. Population Systems: A General Introduction. Plenum Press,
London.
BIGALKE, R. 1947. The status of the black wildebeest Connochaetes gnou (Zimm.)
in the Union of South Africa. South African Journal of Science 43: 213-220.
BORNER, M., FITZGIBBON, C.D., BORNER, M.O., CARO, T.M., LINDSAY, W.K.,
COLLINS, D.A., & HOLT, M.E. 1987. The decline of the Serengeti
Thompson’s gazelle population. Oecologia (Berlin) 73: 32-40.
BOTHMA, J. DU P. 1996. Erosion control and soil reclamation. In: J. du P. Bothma
(Ed.). Game Ranch Management. Third edition. (pp. 573-578). J.L. van
Schaik, Pretoria.
BOTHMA, J. DU P. 2002a (Ed.). Game Ranch Management. Fourth edition. J.L. van
Schaik, Pretoria.
BOTHMA, J. DU P. 2002b. Counting wild animals. In: J. du P. Bothma (Ed.). Game
Ranch Management. Fourth edition. (pp. 335-357). J.L. van Schaik, Pretoria.
280
BOTHMA, J. DU P. & VAN ROOYEN, N. 1996. Suitable game. In: J. du P. Bothma,
(Ed.) Game Ranch Management. Third edition. (pp. 143-159). J.L. van
Schaik, Pretoria.
BOTHMA, J. DU P. & VAN ROOYEN, N. 2002. Ecological Monitoring. In: J. du P.
Bothma (Ed.). Game Ranch Management. Fourth Edition. (pp. 30-33). J.L.
van Schaik, Pretoria.
BOTHMA, J. DU P., VAN ROOYEN, N., & DU TOIT, J.G. 2002. Antelope and other
small herbivores. In: J. du P. Bothma (Ed.). Game Ranch Management.
Fourth edition. (pp. 149-159). J.L. van Schaik, Pretoria.
BOTHMA, J. DU P., VAN ROOYEN, N. & VAN ROOYEN, M.W. 2004. Using diet and
plant resources to set wildlife stocking densities in African savannas. Wildlife
Society Bulletin 32: 840-851.
BRAND, D.J. 1965. Present numerical status of the white-tailed gnu. Quarterly
Magazine of the National Zoological Gardens of South Africa (Zoon) 5: 1-5.
BREDENKAMP, G.J. & THERON, G.K. 1978. A synecological account of the
Suikerbosrand Nature Reserve, 1. The phytosociology of the Witwatersrand
geological system. Bothalia 12: 513-529.
BREDENKAMP, G.J. & VAN ROOYEN, N. 1998. Grassland Biome. In: A.B. Low &
A.G. Rebelo (Eds). Vegetation of South Africa, Lesotho and Swaziland. (pp.
38-52). Department of Environmental Affairs and Tourism, Pretoria.
BRINK, J.S. 1993. Postcranial evidence for the evolution of the black wildebeest
Connochaetes gnou: An exploratory study. Palaeontologica Africana 30: 6169.
BRINK, J.S., BERGER, L.R. & CHURCHILL, S.E. 1999. Mammalian fossils from
erosional gullies (dongas) in the Doring River drainage, central Free State
Province, South Africa. In: Historia Animalium ex Ossibus: Beiträge zür
Palaoanatomie, Archaologie, Agyptologie, Ethnologie und Geschichte der
Tiermedizin. Festschrift für Angela von den Driesch zum 65. Geburtstag.
Verlag Marie Leidorf GmbH, Rahden/Westf.
BRYCE, J., JOHNSON, P.J. & MACDONALD, D.W. 2002. Can niche use in red and
grey squirrels offer clues for their apparent coexistence? Journal of Applied
Ecology 39: 875-887.
BURLISON, A.J., HODGSON, J. & ILLIUS, A.W. 1991. Sward canopy structure, and
the bite dimensions and bite weight of grazing sheep. Grass and Forage
Science 46: 29-38.
281
BYERS, C.R., STEINHORST, R.K. & KRAUSMAN, P.R. 1984. Clarification of a
technique for analysis of utilization-availability data. Journal of Wildlife
Management 48: 1050-1053.
CAUGHLEY, G. & SINCLAIR, A.R.E. 1994. Wildlife Ecology and Management.
Blackwell Scientific Publications, Boston.
CHAPMAN, J.L. & REISS, M.J. 1995. Ecology: Principles and Applications.
Cambridge University Press, Cambridge.
CHILD, G. 1968. Behaviour of Large Mammals during the Formation of Lake Kariba.
Kariba studies, Trustees National Museums, Rhodesia.
CHURCHFIELD, S., NESTERENKO, V.A. & SHVARTS, E.A. 1999. Food niche
overlap and ecological separation amongst six species of coexisting forest
shrews (Insectivora: Soricidae) in the Russian Far East. Journal of Zoology
(London) 248: 349-359.
CLUTTON-BROCK, T.H. & HARVEY, P.H. 1983. The functional significance of body
size among mammals. Special Publications of the American Society of
Mammalogists 7: 633-663.
CODRON, D. & BRINK, J.S. In Press. Trophic ecology of two savanna grazers, blue
wildebeest Connochaetes taurinus and black wildebeest Connochaetes gnou.
European Journal of Wildlife Research: In Press.
CODY, M.L. 1978. Habitat selection and interspecific territoriality among the sylviid
warblers of England and Sweden. Ecological Monographs 48: 351-396.
COE, M.J., CUMMING, D.H. & PHILLIPSON, J. 1976. Biomass and production of
large African herbivores in relation to rainfall and primary production.
Oecologia 22: 341-354.
COLLINSON, R.H.F. 1985. Selecting Wildlife Census Techniques. Monograph 6.
Institute for natural resources, University of Natal, Pietermaritzburg.
CONNELL, J.H. 1980. Diversity and the co-evolution of competitors, or ghost of
competition past. Oikos 35: 131-138.
COOLEY, W.W. & LOHNES, P.R. 1971. Multivariate Data Analysis. John Wiley &
Sons, New York.
COOPER, D. 2003. Wildebeest derived malignant catarrhal fever (WBMC). Appendix
3. In: Black Wildebeest Hybridisation Action Planning Workshop. Workshop
Report. (pp. 44-45). Bloemfontein: CBSG and EWT.
CORBET, S.W. 1991. Genetic divergence in the wildebeest, Connochaetes taurinus
and Connochaetes gnou: A molecular and cytogenic study. M.Sc.
Dissertation, University of Pretoria, Pretoria.
282
CROMHOUT, M. 2006. The ecology of the African buffalo in the eastern Kalahari
region, South Africa. M.Sc. (Wildlife Management) Dissertation, University of
Pretoria, Pretoria.
CROMSIGT, J.P.G.M. 2006. Large herbivores in space: Resource partitioning among
savanna grazers in a heterogeneous environment. Ph.D. Thesis, University of
Groningen, Netherlands.
DATTALO, P. 1994. A comparison of discriminant analysis and logistic regression.
Journal of Social Services Research 19: 121-144.
DE WET, S.F. 1988. Habitat- en voedingsvoorkeure van die blouwildebees
(Connochaetes taurinus taurinus (Burchell, 1923)) en ander diersoorte in die
sentraal suid-oostelike gebied van die Nasionale Krugerwildtuin. M.Sc.
Dissertation, University of Pretoria, Pretoria.
DEKKER, B. 1996. Ekologiese skeiding tussen groot herbivore van die MessinaProefplaas, Limpopovallei. M.Sc. Dissertation, University of Pretoria, Pretoria.
DELANY, M.J. & HAPPOLD D.C.D. 1979. Ecology of African Mammals. Longmann,
New York.
DONALDSON, C.H. & VORSTER, M. 1989. Veld management in the Karoo. Grazing
Management: A Strategy for the Future. Agriforum ’89, Department of
Agriculture and Water Supply.
DÖRGELOH, W.G. 1998. Modelling the habitat requirements and demography of a
population of roan antelope (Hippotragus equinus). Ph.D. Thesis, University
of Pretoria, Pretoria.
DUESER, R.D. & SHUGART, JR. 1978. Microhabitats in forest-floor small mammal
flora. Ecology 59: 89-98.
DU PLESSIS, S.F. 1969. The past and present geographical distribution of the
Perissodactyla and Artiodactyla in Southern Africa. M.Sc. Dissertation,
University of Pretoria, Pretoria.
DU TOIT, J.G. 1991. An evaluation of “snotsiekte” in South Africa for insurance
purposes. Appendix 4. Black Wildebeest Hybridisation Action Planning
Workshop. Workshop Report, pp 46-52. CBSG and EWT, Bloemfontein.
DU TOIT, J.G. & VAN ROOYEN, J. 1996. Roads. In: J. du P. Bothma (Ed.). Game
Ranch Management. Third Edition. (pp. 113-114). JL. van Schaik, Pretoria.
DU TOIT, J.G., MELTZER, D.G.A., PENZHORN, B.L. & VAN HEERDEN, J. 1996.
Bacterial, viral and protozoal diseases. In: J. du P. Bothma (Ed.) Game
Ranch Management. Third edition. (pp. 499-509). J.L. van Schaik, Pretoria.
283
DU TOIT, J.G., VAN MARLE-KÖSTER, E. & BOTHMA, J. DU P. 2002. Hybridisation
and inbreeding. In: J. du P. Bothma (Ed.) Game Ranch Management. Fourth
edition. (pp. 287-291). J.L. van Schaik, Pretoria.
DUNBAR, R.I.M. 1978. Competition and niche separation in a high altitude herbivore
community in Ethiopia. East African Wildlife Journal 16: 183-199.
DUNHAM, K.M., ROBERTSON, E.F. & SWANEPOEL, C.M. 2003. Population decline
of Tsessebe antelope (Damaliscus lunatus lunatus) on a mixed cattle and
wildlife ranch in Zimbabwe. Biological conservation 113: 111-124.
EAST, R. 1998. African antelope database 1998. IUCN/SSC Antelope Specialist
Group Report.
EDWARDS, J. 1983. Diet shifts in moose due to predator avoidance. Oecologia 60:
185-189.
ELY, J.W., DAWSON, J.D., MEHR, D.R. & BURNS, T.L. 1996. Understanding
logistic regression analysis through example. Family Medicine 28: 134-140.
ENGELBRECHT, J.S. 1986. Ekologiese skeiding van die rooibok Aepyceros
melampus (Lichtenstein), waterbok Kobus ellipsiprymnus (Ogilby) en die
koedoe Tragelaphus strepsiceros (Pallas) in die sentrale gebied van die
Nasionale
Krugerwildtuin.
M.Sc.
(Wildlife
Management)
Dissertation,
University of Pretoria, Pretoria.
ESTES, R.D. 1966. Behaviour and life history of the wildebeest (Connochaetes
taurinus Burchell). Nature 5066: 999-1000.
ESTES, R.D. 1969. Territorial behaviour of the wildebeest (Connochaetes taurinus
Burchell, 1823). Zietschrift für Tierpsychologie 26: 284-370.
ESTES, R.D. 1974. Social organisation of the African Bovidae. In: V. Geist & F.
Walther (Eds). The Behaviour of Ungulates and its Relation to Management.
(pp. 166-205). IUCN Publications, Series No. 24, Morges.
ESTES, R. D. 1991. The Behavior Guide to African Mammals. University of California
Press, London.
FABRICIUS,
C.
1984.
Abiotiese
habitatvoorkeure
van
swartwildebees
(Connochaetes gnou, Zimm.) in die Golden Gate Hoogland Nasionale Park.
Unpublished B.Sc. Honours Project, University of Pretoria, Pretoria.
FABRICIUS, C. 1989. Habitat suitability assessment for indigenous browsing
ungulates in the Northern Cape. M.Sc. Dissertation, University of the
Witwatersrand, Johannesburg.
FABRICIUS, C. & MENTIS, M.T. 1990. Seasonal habitat selection by eland in arid
savanna in southern Africa. South African Journal of Zoology 25: 238-244.
284
FABRICIUS, C. & OATES, L.G. 1985. Status of the black wildebeest in the
Transvaal. Fauna and Flora 42: 22-25.
FABRICIUS, C., LOWRY, D. & VAN DEN BERG, P. 1988. Fecund black wildebeest x
blue wildebeest hybrids. South African Journal of Wildlife Research 18: 35-37.
FAGEN, R. 1988. Population effects of habitat change: a quantitative assessment.
Journal of Wildlife Management 52: 41-46.
FAIRALL, N. 1968. The reproductive seasons of some mammals in the Kruger
National Park. Zoologica Africana 3: 189-210.
FERRAR, A.A. & WALKER, B.H. 1974. An analysis of herbivore/habitat relationships
in Kyle National Park, Rhodesia. Journal of the southern African Wildlife
Management Association 4:137-147.
FORSYTH, D.M. 2000. Habitat selection and coexistence of the Alpine chamois
(Rupicapra rupicapra) and Himalayan tahr (Hemitragus jemlahicus) in the
eastern Southern Alps, New Zealand. Journal of Zoology (London) 252: 215225.
FOURIE,
J.H.,
VAN
NIEKERK,
J.W.
&
FOUCHÉ,
H.J.
1985.
Weidingskapasiteitsnorme in die Vrystaatstreek. Glen Agric 14: 4-7.
FOURIE, J.H. & VISAGIE, A.F.J. 1985. Weidingswaarde en ekologiese status van
grasse en karoobossies in die Vrystaatstreek. Glen Agric 14: 14-18.
FRIEDMAN, Y. & DALY, B. 2004. Red Data Book of the Mammals of South Africa: A
Conservation Assessment: CBSG Southern Africa, Conservation Breeding
Specialist Group (SSC/IUCN), Endangered Willdife Trust, South Africa.
FRYXELL, J.M. 1991. Forage quality and aggregation by large herbivores. American
Naturalist 138: 478-498.
FURSTENBURG, D. 2002a. Swartwildebees “White-tailed Gnu”. Game and Hunt 8:
8-9, 39, 43.
FURSTENBURG, D. 2002b. Blouwildebees Brindled gnu: Connochaetes taurinus
taurinus. Game and Hunt 8: 8-9, 43.
GEIST, V. & PETOCZ, R.G. 1977. Bighorn sheep in winter: do rams maximise
reproductive fitness by spatial and habitat segregation from ewes? Canadian
Journal of Zoology 55: 1802-1810.
GEOLOGICAL SURVEY STAFF 1978. 1:250 000 geological series map sheet 2528
Pretoria. Government Printer, Pretoria.
GILES, R.H. 1978. Wildlife Management. W.H. Freeman and Company, San
Francisco.
GORDON, I.J. & ILLIUS, A.W. 1989. Resource partitioning by ungulates on the Isle
of Rhum. Oecologia 79: 383-389.
285
GOVERNMENT PRINTER 1996. 2528DB Sokhulumi. Fourth Edition. Chief
Directorate: Surveys and Land Information, Pretoria.
GOVERNMENT PRINTER 1998. 2529CA Sehlakwana. Third Edition. Chief
Directorate: Surveys and Land Information, Pretoria.
GRIMSDELL, J.J.R. & FIELD, C.R. 1976. Grazing patterns of buffalo in the Rwenzori
National Park, Uganda. East African Wildlife Journal 14: 339-344.
GROBLER, A. 1999. The phytosociology and veld condition of Ezemvelo Nature
Reserve. Unpublished B.Sc. Honours Report. University of Pretoria, Pretoria.
GROBLER, J.H. 1983. Feeding habits of the Cape mountain zebra Equus zebra
zebra Linn. 1758. Koedoe 26: 159-168.
GROBLER, P. 2003. Summary of genetic screening of black wildebeest on the Abe
Bailey Nature Reserve. Appendix 1. In: Black Wildebeest Hybridisation Action
Planning Workshop. Workshop Report, (pp 33-38). CBSG and EWT,
Bloemfontein.
GROSSMAN, D. 1982. Primary production of rangeland: practical and interpretative
problems. Proceedings of the Grassland Society of Southern Africa 17: 76-78.
GRUNOW, J.O. 1980. Feed and habitat preferences among some large herbivores
on African veld. Proceedings of the Grassland Society of Southern Africa 15:
141-146.
HAIR, J.F., ANDERSON, R.E., TATHAM, R.L. & BLACK, W.C. 1995. Multivariate
data analysis. Prentice-Hall, Englewood-Cliffs.
HALL, G.H. & ROUND, A.P. 1994. Logistic regression - explanation and use. JR
Collins Physicians London 28: 242-246.
HARDY, M.B. & WALKER, R.S. 1991. Determining sample size for assessing
species composition in grassland. Journal of the Grassland Society of
Southern Africa 8: 70-73.
HARRINGTON, R., OWEN-SMITH, N., VILJOEN, P.C., BIGGS, H.C., MASON, D.R.
& FUNSTON, P. 1999. Establishing the causes of the roan antelope decline
in the Kruger National Park, South Africa. Biological Conservation 90: 69-78.
HEITKÖNIG, I.M.A. & OWEN-SMITH N. 1998. Seasonal selection of soil types and
grass swards by roan antelope in a South African savanna. African Journal of
Ecology 36: 57-70.
HEMAMI, M.R., WATKINSON A.R. & DOLMAN, P.M. 2004. Habitat selection by
sympatric muntjac (Muntiacus reevesi) and roe deer (Capreolus capreolus) in
a lowland commercial pine forest. Forest Ecology and Management 194: 4960.
286
HERBERT, H.J. 1970. The population dynamics of the waterbuck, Kobus
ellipsiprymnus (Ogilby, 1883), in the Sabi Sand Wildtuin. M.Sc. Dissertation,
University of Pretoria, Pretoria
HESTER, A.J., GORDON, I.J., BAILLIE, G.J. & TAPPIN, E. 1999. Foraging
behaviour of sheep and red deer within natural heather grass mosaics.
Journal of Applied Ecology 36: 133-146.
HIK, D.S. & JEFFRIES, R.L. 1990. Increases in the net above-ground primary
production of a salt-marsh forage grass- a test of the predictions of the
herbivore-optimisation model. Journal of Ecology 78: 180-195.
HIRST,
S.M.
1975.
Ungulate-habitat
relationships
in
a
South
African
woodland/savanna ecosystem. Wildlife Monographs 44:1-60.
HOFER, U., BERSIER, L. & BORCARD, D. 2004. Relating niche and spatial overlap
at the community level. Oikos 106: 366-376.
HOFMEYER, M.D. 1981. Thermal physiology of selected African ungulates with
emphasis on the physical properties of the pelage. Ph.D. Thesis, University of
Cape Town, Cape Town.
HOLECHEK, J.L., PIEPER, R.D. & HERBEL, C.H. 1989. Range Management:
Principles and Practices. Prentice-Hall, Engelwood Cliffs.
HORAK, I.G., DE VOS, V. & BROWN, M.R. 1983. Parasites of domestic and wild
animals in South Africa. XVI. Helminth and arthropod parasites of blue and
black
wildebeest
(Connochaetes
taurinus
and
Connochaetes
gnou).
Onderstepoort Journal of Veterinary Research 50: 243-255.
HURLBERT, S.H. 1978. The measurement of niche overlap and some relatives.
Ecology 59: 67-77.
HUTCHINSON, G.E. 1957. Concluding remarks. Cold Spring Harbour Symposium on
Quantitative Biology 22: 415-427.
ILLIUS, A.W. & GORDON, I.J. 1987. The allometry of food intake in grazing
ruminants. Journal of Animal Ecology 56: 989-999.
ILLIUS, A.W., CLARKE, D.A. & HODGSON, J. 1992. Discrimination and patch choice
by sheep grazing grass clover swards. Journal of Applied Ecology 61: 183194.
JARMAN, P.J. 1974. The social organisation of antelope in relation to their ecology.
Behaviour 48: 215-267.
JARMAN, P.J. 1977. Behaviour of Topi in a shadeless environment. Zoologica
Africana 12: 101-111.
287
JOHNSON, B.K., KERN, J.W., WISDOM, M.J., FINDHOLT, S.L. & KIE, J.G. 2000.
Resource selection and spatial separation of mule deer and elk during spring.
Journal of Wildlife Management 64: 685-697.
JOHNSON, D.H. 1980. The comparison of usage and availability measurements for
evaluating resource preference. Ecology 61: 65-71.
JOUBERT, S.C.J. 2002. Animal behaviour. In: J. du P. Bothma (Ed.) Game Ranch
Management. Fourth edition. (pp. 280-286). J.L. van Schaik, Pretoria.
KAY, C.M.E. 1992. The impact of large herbivores on the vegetation of the Golden
Gate Highlands National Park, with special reference to the black wildebeest
Connochaetes gnou. M.Sc. Dissertation, University of Pretoria, Pretoria.
KEAST, J. 1965. Interrelationships of the two zebra species in an overlap zone.
Journal of Mammalogy 46: 53-66.
KINYAMARIO, J.I. & MACHARIA, J.N.M. 1992. Above-ground standing crop, protein
content and dry matter digestibility of a tropical grassland range in the Nairobi
National Park, Kenya. African Journal of Ecology 30: 33-41.
KNIGHT, M.H. 1991. Ecology of the gemsbok Oryx gazella gazella (Linnaeus) and
the blue wildebeest Connochaetes taurinus (Burchell) in the southern
Kalahari. Ph.D. thesis, University of Pretoria, Pretoria.
KRÄMER, A. 1973. Interspecific behaviour and dispersion of two sympatric deer
species. Journal of Wildlife Management 37: 288-300.
KREBS, C.J. 1999. Ecological Methodology. Second Edition. Addison-Welsey
Educational Publishers, Inc., New York.
KRÜGER, J. 1996. Socio-ecology of the African buffalo (Syncerus caffer):
Implications for game ranching. In: B.L. Penzhorn (Ed.). Proceedings of a
Symposium on the African Buffalo as a Game Ranch Animal. (pp. 1-5). The
Wildlife Group, South African Veterinary Association, Onderstepoort.
LAMPREY, H.F. 1963. Ecological separation of the large mammal species in the
Tarangire Game Reserve, Tanganyika. East African Wildlife Journal 1: 63-93.
LAND TYPE SURVEY STAFF 1987. Land types of the Maps 2526 Rustenburg, 2528
Pretoria. Memoirs on the Agricultural and Natural Resources of South Africa.
No. 8.
LANGLEY, N. 1995. Black wildebeest x blue wildebeest hybrid at Spioenkop Nature
Reserve. Unpublished Natal Parks Board Report, 6pp.
LARCHER, W. 1995. Physiological plant ecology. Third Edition. Springer, New York.
LEUTHOLD, B.M. & LEUTHOLD, W. 1978. Daytime activity patterns of gerenuk and
giraffe in Tsavo National Park, Kenya. East African Wildlife Journal 16: 231243
288
LEUTHOLD, W. 1977. African ungulates. A comparative review of their ethology and
behavioural ecology. Zoophysiology and Ecology Vol. 8. Springer Verlag,
Berlin.
LEVINS,
R.
1968.
Evolution
in
changing
environments:
some theoretical
explorations. Princeton University Press, Princeton.
LOMAN, J. 1986. Use of overlap indices as competition coefficients: tests with field
data. Ecological Modelling 34: 231-243.
LOTTES, I.L., ADLER, M.A. & DEMARIS, A. 1996. Using and interpreting logistic
regression: a guide for teachers and students. Teaching Sociology 24: 284298.
LOW, A.B. & REBELO, A.G. 1996. Vegetation of South Africa, Lesotho and
Swaziland. Department of Environmental Affairs and Tourism, Pretoria.
LURIE, J. 2001. South African Geology for Mining, Metallurgical, Hydrological and
Civil Engineering. Eighth Revised Edition. Lupon Publishing, Johannesburg.
LYONS, A.L., GAINES, W.L., & SERVHEEN, C. 2003. Black bear resource selection
in the northeast Cascades, Washington. Biological Conservation 113: 55-62.
MAGURRAN, A.E. 1991. Ecological Diversity and its Measurement. Chapman and
Hall, London.
MANLY, B.F.J., MCDONALD, L.L. & THOMAS D.L. 1993. Resource Selection by
Animals. Chapman and Hall, London.
MARNELL, F. 1998. Discriminant analysis of the terrestrial and aquatic habitat
determinations of the smooth newt (Triturus vulgaris) and the common frog
(Rana temporaria) in Ireland. Journal of Zoology (London) 244: 1-6.
MASKALL, J. & THORNTON, I. 1996. The distribution of trace and major elements in
Kenyan soil profiles and implications for wildlife nutrition. Environmental
Geochemistry and Health. Geological Society Special Publication No. 113:
47.
MAY, R.M. 1973. On the theory of niche overlap. Theoretical Population Biology 5:
297-332.
MCARDLE, J.J. & HAMAGAMI, F. 1996. Multilevel models from a multiple group
structural equation perspective. In: G.A. Marcoulides & R.E. Schumacker.
(Eds). Advanced Structural Modelling: Issues and Techniques. (pp. 89-124).
Lawrene Erlbaum Associates Publishers, Mahwah.
MCNAUGHTON, S.J. 1976. Serengeti migratory wildebeest: facilitation of energy
flow by grazing. Science 191: 92-94.
MCNAUGHTON, S.J. 1979. Grazing as an optimisation process: grass-ungulate
relationships in the Serengeti. American Naturalist 113: 691-703.
289
MCNAUGHTON S.J. 1984. Grazing lawns: animals in herds, plant form and
coevolution. American Naturalist 124: 863-886.
MCNAUGHTON, S.J. 1985. Ecology of a grazing ecosystem: the Serengeti.
Ecological Monographs 55: 259-294.
MEISSNER, H.H. 1999. Nutrient supplementation of the grazing animal. In: N.M.
Tainton (Ed.). Veld Management in South Africa. (pp. 50-64). University of
Natal Press, Pietermaritzburg.
MELTON, D.A. 1978. Ecology of the waterbuck Kobus ellipsiprymnus in the
Umfolozi. Ph.D. Thesis, University of Pretoria, Pretoria.
MENTIS, M.T. 1981. Evaluation of the wheel-point and step-point methods of veld
condition assessment. Proceedings of the Grassland Society of Southern
Africa 16: 89-94.
MENTIS, M.T. & COLLINSON, R.F.H. 1979. Management goals for wildlife reserves
in grassveld and bushveld. Proceedings of the Grassland Society of Southern
Africa 14: 71-74.
MENTIS, M.T. & DUKE, R.R. 1976. Carrying capacities of natural veld in Natal for
large wild herbivores. South African Journal of Wildlife Research 6: 65-74.
MEYER, J.A. & CASEY, N.H. 2002. Water quality. In: J. du P. Bothma (Ed.). Game
Ranch Management. Fourth edition. (pp. 95-97). J.L. van Schaik, Pretoria.
MILLER, R.S. 1967. Pattern and process in competition. In: J.B. Cragg (Ed).
Advances in Ecological Research, Vol. 4. (pp. 1-74). Academic Press,
London.
MILLS, G. & HESS, L. 1997. The Complete Book of Southern African mammals. The
Struik Publishing Group (Pty) Ltd., Cape Town.
MOORE, W.S. 1977. An evaluation of narrow hybrid zones in vertebrates. Quarterly
Review of Biology 52: 263-277.
MORRISON, M.L. MARCOT, B.G. & MANNAN, R.W. 1992. Wildlife-habitat
Relationships: Concepts and Applications. The University of Wisconsin Press,
Madison.
MUCINA, L., RUTHERFORD, M.C. & POWRIE, L.W. (Eds) 2005. Vegetation Map of
South Africa, Lesotho and Swaziland, 1:1 000 000 scale sheet maps. South
African National Biodiversity Institute, Pretoria.
MUNTHALI, S.M. & BANDA, H.A. 1992. Distribution and abundance of the common
ungulates of Nyika National Park, Malawi. African Journal of Ecology 30: 203212.
MURRAY, M.G. & BROWN, D. 1993. Niche separation of grazing ungulates in the
Serengeti: an experimental test. Journal of Animal Ecology 62: 380-389.
290
MUSHI, E.Z., RURANGIRWA, F.R. & KARSTAD, L. 1981. Shedding of malignant
catarrhal fever virus by wildebeest calves. Veterinary Microbiology 6: 281286.
NAMGAIL, T., FOX, J.L. & BHATNAGAR, Y.V. 2004. Habitat segregation between
sympatric Tibetan argali Ovis ammon hodgsoni and blue sheep Pseudois
nayaur in the Indian Trans-Himalaya. Journal of Zoology (London) 262: 5763.
NOVELLIE, P. 1990. Habitat use by indigenous grazing ungulates in relation to
sward structure and veld condition. Journal of the Grassland Society of
Southern Africa 7: 16-23.
NOVELLIE, P. & STRYDOM, G. 1987. Monitoring the response of vegetation to use
by large herbivores: an assessment of some techniques. South African
Journal of Wildlife Research 17: 109-117.
O’CONNOR, T.G. 1985. A synthesis of the field experiments concerning the grass
layer in the savanna regions of southern Africa. South African National
Scientific Programmes Report 114. CSIR, Pretoria.
OWEN-SMITH, N. 1982. Factors influencing the consumption of plant products by
large herbivores. In: J. Huntley & B.H. Walker (Eds). Ecology of Tropical
Savannas. (pp. 359-404). Springer-Verlag, London.
OWEN-SMITH, R.N. 1985. Niche separation among African ungulates. Species and
speciation In: E.S. Vrba (Ed.) Transvaal Museum Monograph No 4. (pp. 161171). Transvaal Museum, Pretoria.
OWEN-SMITH, N. 1990. Demography of a large herbivore, the greater kudu, in
relation to rainfall. Journal of Animal Ecology 59: 893-913.
OWEN-SMITH, N. 2004. Functional heterogeneity in resources within landscapes
and herbivore population dynamics. Landscape ecology 19: 761-771.
PAULEY, G.R. PEEK, J.M. & ZAGER, P. 1993. Predicting white-tailed deer habitat
use in northern Idaho. Journal of Wildlife Management 57: 904-913.
PENRITH, M.L., TUSTIN, R.C., THORNTON, D.J. & BURDETT. P.D. 1996.
Swayback in blesbok (Damaliscus dorcas phillipsi) and a black wildebeest
(Connochaetes gnou). Journal of the South African Veterinary Association 67:
93-96.
PENZHORN, B.L. 1982. Habitat selection by Cape mountain zebra in the Mountain
Zebra National Park. South African Journal of Wildlife Research 12: 48-54.
PETTIFER, H.L. & STUMPF, R.H. 1981. An approach to the calculation of habitat
preference data: Impala on Loskop Dam Nature Reserve. South African
Journal of Wildlife Research 11: 5-13.
291
PIANKA, E.R. 1973. The structure of lizard communities. Annual review of ecology
and systematics 4: 53-74.
PIANKA, E.R. 1978. Evolutionary Ecology. Second edition. Harper and Row, New
York.
PLUG, I. & BADENHORST, S. 2001. The distribution of macromammals in southern
Africa over the past 30 000 years as reflected in animal remains from
archaeological sites. Transvaal Museum Monograph No. 12, Pretoria.
PORTER, W.F. & CHURCH, K.E. 1987. Effects of environmental pattern on habitat
preference analysis. Journal of Wildlife Management 51: 681-685.
PRESTON-WHYTE, R.A. & TYSON, P.D. 1988. The Atmosphere and Weather of
Southern Africa. Oxford University Press, Cape Town.
PRZYBYLO, R. & MERILA, J. 2000. Intersexual niche differentiation in the blue tit
(Parus caeruleus) Biological Journal of the Linnean Society 69: 233-244.
PUTMAN, R.J. 1996. Competition and resource partitioning in temperate ungulate
assemblages. Chapman & Hall, London.
REILLY, B.K. 1989. Ecology of a population of oribi Ourebia ourebia ourebia
(Zimmerman, 1783) in the Golden Gate Highlands National Park. M.Sc.
Dissertation. University of Pretoria, Pretoria.
REINERT, H.K. 1984. Habitat separation between sympatric snake populations.
Ecology 65: 478-486.
RINEY, T. 1982. Study Management of Large Mammals. John Wiley, New York.
ROBERTS, A. 1951. The Mammals of South Africa. Central News Agency, Pretoria.
ROUX, P.W. & OPPERMAN, D.P.J. 1986. Soil Erosion. In: R.M. Cowling, P.W. Roux
& A.J.H. Pieterse (Eds). The Karoo Biome: a Preliminary Synthesis. Part 1Physical environment. CSIR Report No. 124. (pp. 102-133), CSIR, Pretoria.
RUBIN, E.S., BOYCE, W.M., STERMER, C.J. & TORRES, S.G. 2002. Bighorn sheep
habitat use and selection near an urban environment. Biological Conservation
104: 251-263.
RUESS, R.W., MCNAUGHTON, S.J. & COUGHENOUR, M.B. 1983. The effects of
clipping on nitrogen source and nitrogen concentration on the growthresponses and nitrogen uptake of an East-African sedge. Oecologia 59:253261.
RUSHWORTH, I.A. 1992. Biophysical determinants of grazing ungulate habitat use.
Grassland Science Research Project 490: 2-22.
SCHMIDT, A.G. 1988. Black wildebeest habitat utilisation. Unpublished B.Sc.
Honours Report, University of Pretoria, Pretoria.
292
SCHMIDT, A.G. & SNYMAN, D.D. 2002. Nutritional and mineral deficiencies, and
supplementary feeding. In: J. du P. Bothma (Ed.). Game Ranch Management.
Fourth Edition (pp. 267-274). J.L. van Schaik, Pretoria.
SCHOENER, T.W. 1974a. Resource partitioning in ecological communities. Science
185: 27-39.
SCHOENER, S.H. 1974b. Some methods for calculating competition coefficients
from resource utilization spectra. American Naturalist 108: 332-340.
SCHOENER, T.W. 1986. Resource partitioning. In: D. Anderson & J. Kikkawa (Eds).
Community Ecology- Pattern and Process (pp. 91-126). Blackwell Publishing,
Oxford.
SCOGINGS, P.F., THERON, G.K. & BOTHMA, J. DU P. 1990. Two quantitative
methods of analysing ungulate habitat data. South African Journal of Wildlife
Research 20: 9-13.
SHACKLETON, C.M. 1990. Seasonal changes in biomass concentration in three
coastal grassland communities in Transkei. Journal of the Grassland Society
of southern Africa 7: 265-269.
SHAW, J.H. 1985. Introduction to Wildlife Management. McGraw-Hill, New York.
SKINNER, J.D. & SMITHERS, R.H.N. 1990. The Mammals of the Southern African
Subregion. Second edition. University of Pretoria, Pretoria.
SKINNER, J.D. & CHIMIMBA, C.T. 2005. The Mammals of the Southern African
Subregion. Third Edition. Cambridge University Press, Cambridge.
SMITHERS, R. H. N. 1983. The Mammals of the Southern African Subregion.
University of Pretoria, Pretoria.
SMITHERS, R. H. N. 1986. Land mammals of southern Africa: A field guide.
Macmillan South Africa (Publishers) (Pty) Ltd., Johannesburg.
SNYMAN, H.A. 1999. Soil Erosion and Conservation. In: N. Tainton (Ed.). Veld
Management in South Africa. (pp. 355-380). University of Natal Press,
Pietermaritzburg.
SOIL CLASSIFICATION WORKING GROUP. 1991. Soil Classification. A taxonomic
system for South Africa. Memoirs on the Agricultural Natural Resources of
South Africa No. 15. The Department of Agricultural Development, Pretoria.
SOUTH AFRICAN COMMITTEE FOR STRATIGRAPHY (SACS) 1980. Stratigraphy
of South Africa. Part I. (Comp. L.E. Kent). Lithostratigraphy of the Republic of
South Africa, South West Africa / Namibia and the Republics of
Bophutatswana, Transkei and Venda. Handbook of the Geological Survey of
South Africa, 8.
293
STOCK, W.D., CHUBA, D.K. & VERBOOM, G.A. 2004. Distribution of South African
C3 and C4 species of the Cyperaceae in relation to climate and phylogeny.
Austral Ecology 29: 313-319.
STRAUSS, W.M. 2003. An ecological study of reintroduced Arabian oryx in the ‘Uruq
Bani Ma’arid Protected Area of the Kingdom of Saudi Arabia. M.Sc.
Dissertation (Wildlife Management), University of Pretoria, Pretoria.
STUART-HILL, G.C. 1989. Adaptive Management: the only practical method of veld
management. In: W.E. Danckwerts & W.R. Teague (Eds). Veld management
of the Eastern Cape. (pp. 4-7). Government Printers, Pretoria.
TAINTON, N. 1999. Veld Management in South Africa. University of Natal Press,
Pietremaritzburg.
TAMBLING, C.J. & DU TOIT, J.T. 2005. Modelling wildebeest population dynamics:
Implications of predation and harvesting in a closed system. Biological
Conservation 99: 237-243.
THERON, A. 1991. The behavioural ecology of two zebra species on game farms in
the Waterberg, northern Transvaal. M.Sc. Dissertation, University of Pretoria,
Pretoria.
THOMAS, D.L. & TAYLOR, E.J. 1990. Study designs and tests for comparing
resource use and availability. Journal of Wildlife Management 54: 322-330.
THOMSON, R. 1992. The Wildlife Game. The Nyala Wildlife Publication Trust,
Westville.
THORNTON, I. 2002. Geochemistry and the mineral nutrition of agricultural livestock
and wildlife. Applied Geochemistry 17: 1017
TOMOR, B.M. & OWEN-SMITH, N. 2002. Comparative use of grass regrowth
following burns by four ungulate species in the Nylsvley Nature Reserve,
South Africa. African Journal of Ecology 40: 201-204.
TRAILL, L.W. 2004. Seasonal utilisation of habitat by large grazing herbivores in
semi-arid Zimbabwe. South African Journal of Wildlife Research 34: 13-24.
TROLLOPE, W.S.W. 1990. Development of a technique for assessing veld condition
in the Kruger National Park using key grass species. Journal of the Grassland
Society of Southern Africa 7: 46-51.
TROLLOPE, W.S.W. & POTGIETER, A.L.F. 1986. Estimating grass fuel loads with a
disc pasture metre in the Kruger National Park. Journal of the Grassland
Society of Southern Africa 2: 148-152.
TROLLOPE, W.S.W., DE RONDE, C. & GELDENHUYS, C.J. 2004. Fire Behaviour.
In: J.G. Goldammer & C. de Ronde (Eds). Wildland Fire Management
294
Handbook for Sub-Saharan Africa. (pp. 200-201). ComPress, Global Fire
Monitoring Centre.
TROLLOPE, W.S.W., POTGIETER, A.L.F. & ZAMBATIS, N. 1989. Assessing veld
condition in the Kruger National Park using key grass species. Koedoe 32:
67-93.
TURCHIN, P. 1995. Population Regulation: Old Arguments and a new synthesis. In:
N. Cappuccino & P.W. Price (Eds). Population Dynamics: New Approaches
and Synthesis. (pp. 19-37). Academic Press, New York.
VAN HORNE, B. 1982. Niches of adult and juvenile deer mice (Peromyscus
maniculatus) in seral stages of coniferous forests. Ecology 63: 992-1003.
VAN HOVEN, W. 2002. The natural diet of wild herbivores. In: J. du P. Bothma (Ed.).
Game Ranch Management. Fourth edition. (pp. 243-253). J.L. van Schaik,
Pretoria.
VAN OUDTSHOORN, F. 1999. Guide to Grasses of Southern Africa. Briza
Publications, Pretoria.
VAN ROOYEN, N. 2002. Veld management in the savannas. In: J. du P. Bothma
(Ed.). Game Ranch Management. Fourth Edition. (pp. 571-613). J.L. van
Schaik, Pretoria.
VAN ROOYEN, N., BREDENKAMP, G.J. & THERON, G.K. 1996. Veld Management,
In: J. du P. Bothma (Ed.). Game Ranch Management. Third edition. (pp. 539572). J.L. van Schaik, Pretoria.
VAN SCHALKWYK, C.J. 1984. A Primer on Soil Conservation. Bulletin No. 399.
Division of Agricultural Engineering, Department of Agriculture, Pretoria.
VERWEIJ, R.J.T., VERRELST, J., LOTH, P.E., HEITKÖNIG, M.A. & BRUNSTING,
M.H. 2006. Grazing lawns contribute to the subsistence of mesoherbivores on
dystrophic savannas. Oikos 114: 108-116.
VOETEN, M.M. & PRINS, H.H.T. 1999. Resource partitioning between sympatric wild
and domestic herbivores in the Tarangire region of Tanzania. Oecologia
(Berl.) 120: 287-294.
VOISIN, A. 1988. Grass Productivity. Island Press, Washington DC.
VON HOLDT, A.L. 1999. Ecological separation by browsers on the Lewa Wildlife
Conservancy, Kenya. M.Sc. Dissertation, University of Pretoria, Pretoria.
VON RICHTER, W. 1971a. The black wildebeest (Connochaetes gnou). Nature
conservation Miscellaneous Publication No. 2. Orange Free State Provincial
Administration, Bloemfontein.
295
VON RICHTER, W. 1971b. Observations on the biology and ecology of the black
wildebeest (Connochaetes gnou). Journal of the South African Wildlife
Management Association 1: 3-16.
VON RICHTER, W. 1974. Connochaetes gnou. Mammalian Species No 50 (pp. 1-6).
The American Society of Mammalogists, New York.
VORSTER, M. 1982. Development of the ecological index method for assessing veld
condition in the Karoo. Proceedings of the Grassland Society of Southern
Africa 17: 84-85.
VORSTER, M. 1999. Karoo and the associated sweet grassveld. In: N. Tainton (Ed.)
Veld Management in South Africa. (pp. 293-311). University of Natal Press,
Pietermaritzberg.
VRAHIMIS, S. 2003a. Crossbreeding threatens the black wildebeest. African Wildlife
57: 45-45.
VRAHIMIS, S. 2003b. Wildebeest: Black, blue and red? Can you assist in the
development of a National Conservation Strategy to conserve an endemic
species? Game and Hunt, August 2003.
VRAHIMIS, S. & O.B. KOK. 1992. Body orientation of black wildebeest in a semi-arid
environment. African Journal of Ecology 30: 169-175.
VRAHIMIS, S. & O.B. KOK. 1993. Daily activity of black wildebeest in a semi-arid
environment. African Journal of Ecology 31: 328-336.
VRAHIMIS, S. & KOK, O.B. 1994. Notes on the diurnal activity of early post-natal
black wildebeest calves. Koedoe 37: 109-113.
WALPOLE, M.J., MORGAN-DAVIES, M., MILLEDGE, S., BETT, P. & LEADERWILLIAMS, N. 2001. Population dynamics and future conservation of a freeranging black rhinoceros (Diceros bicornis) population in Kenya. Biological
Conservation 99: 237-243.
WALTER, H. 1979. Vegetation of the earth and ecological systems of the geobiosphere. Second edition (pp. 79). Springer-Verlag, New York.
WATSON, R.M. 1969. Reproduction of wildebeest, Connochaetes taurinus
albojubatus Thomas in the Serengeti Region, and its significance to
conservation. Journal of Reproductive Fertility, Supplement 6: 287-310.
WEAVER, S.M. 1992. Back from the brink. Man Magnum 17: 21-24, 26.
WEAVER, S.M. 1995. Habitat utilisation by selected herbivores in the Klaserie
Nature Reserve, South Africa. M.Sc. Dissertation, University of Pretoria,
Pretoria.
WEI, F., FENG, Z., WANG, Z. & HU, J. 2000. Habitat use and separation between
the giant panda and the red panda. Journal of Mammalogy 81: 448-455.
296
WENTZEL, J.J. 1990. Ekologiese skeiding van geselekteerde herbivoor spesies in
die benede Sabie-Krokodilrivier gebied van die Nasionale Krugerwildtuin.
M.Sc. (Wildlife Management) Dissertation, University of Pretroria, Pretoria.
WENTZEL, J.J., BOTHMA, J. DU P. & VAN ROOYEN, N. 1991. Characteristics of
the herbaceous layer in preferred areas of the six herbivore species in the
south-eastern Kruger National Park. Koedoe 34: 51-59.
WERNER, E.E. & HALL, V. 1979. Foraging efficiency and habitat switching in
competing sunfishes. Ecology 60: 256-264.
WHILE, G.M. & MCARTHUR, C. 2005. Foraging in a risky environment: a
comparison of Bennett’s wallabies Macropus rufogriseus (Marsupialia:
Macropodidae)
and
red-bellied
pademelons
Thylogale
billiardierii
(Marsupialia: Macropodidae) in open habitats. Austral Ecology 30: 756-764.
WHITEHEAD, D.C. 2000. Nutrient Elements in Grassland: Soil–Plant–Animal
Relationships. Cambridge University Press, Cambridge.
WHYTE, I.J. 1985. The present ecological status of the wildebeest (Connochaetus
taurinus taurinus Burchell 1823) in the central district of the Kruger National
Park. M.Sc. Dissertation, University of Natal, Pietermaritzburg.
WILSON, B.G. & WITKOWSKI, E.T.F. 2003. Seed banks, bark thickness and change
in age and size structure (1978-1999) of the African savanna tree, Burkea
africana. Plant Ecology 167: 151-162.
WINTERBACH, H.E.K. 1999. Habitat utilisation, activity patterns and management of
Cape Buffalo in the Willem Pretorius Game Reserve. M.Sc. Dissertation.
University of Pretoria, Pretoria.
YDENBERG, R.C. & PRINS, H.H.T. 1981. Spring grazing and the manipulation of
food quality by barnacle geese. Journal of Applied Ecology 18: 443-453.
ZAR, H.J. 1984. Biostatistical analysis. Second edition. Prentice Hall, New Jersey.
297
APPENDICES
APPENDIX 1
Habitat Selection Field Data Sheet
1
2
3
4
Date
Time
Co-ordinates
Grass
cover
Woody
density
Landscape
position
Activity
Group
compos.
Type of wildebeest
Males
Females
Subadults
Calves
Total herd size
Grazing
Lying down
Walking
Standing
Other
Habitat type
Aspect
Slope/Gradient
Altitude
Geomorphology
Plains
Gentle slopes
Valley
Plateaus
None
Sparse
Open
Sparse
Medium
Dense
Total grass height
Grass leaf height
Distance to water
Dominant plant species
Sub-dominant species
Vegetation structure
Plant utilisation
Forb : grass ratio
Cloud cover
Temperature
Wind velocity
Wind direction
Drainage
Exposure
Rock cover
Visibility
Distance to shade
Associations
Erosion
Time since last burn
298
5
6
7
8
9
10
APPENDIX 2
The variables and their categories that were utilised in the PROC LOGISTIC procedure to determine the level of meso-habitat separation
between the black and blue wildebeest at Ezemvelo Nature Reserve from January 2004 to August 2005
Variable
V20
Description
Woody vegetation cover
Number of categories
3
V21
Grass cover
3
V22
Cloud cover
3
3
V23
Temperature
3
V24
Wind speed
4
V25
Wind Direction
6
V26
Rock Cover
6
V27
Total grass height
3
V28
Grass leaf height
4
299
Categories and abbreviations used
1. None
2. Sparse
3. Open
1. Sparse
2. Medium
3. Dense
1. 0% (Clear skies)
2. 1-50% (Partly cloudy)
3. >50% (Overcast)
1. 0%
2. 1-50%
3. >50 %
1. <15ºC
2. •15-25ºC
3. >25ºC
1. None (0-2 km/h)
2. Slight (>2 – 5 km/h)
3. Moderate (>5 – 13 km/h)
4. Severe (>13 km/h)
1. North
2. Northeast
3. East
4. Southeast
5. West
6. Northwest
1. None
2. 1-30%
3. >30%
1. 0-50 mm
2. >50-500 mm
3. >500-800 mm
1. 0-50 mm
2. >50- 100 mm
3. >100-400 mm
4. >400 mm
APPENDIX 2 Continued.
Variable
V29
Description
Plant utilisation
Number of categories in group
4
V30
Visibility
4
V31
Distance to shade
4
V34
Erosion
3
V35
Altitude
V36
Date of last burn
5
V39
Exposure
3
V40
Geomorphology
3
V41
Forb : grass ratio
4
V43
Social structure
3
300
Categories and abbreviations used
1. Low
2. Moderate
3. High
4. Excessive
1. 0-50 m
2. >50-100 m
3. >100-200 m
4. >200 m
1. 0-5 m
2. >5-100 m
3. >100-600 m
4. >600 m
1. Low
2. Moderate
3. High
1. ”1340 m
2. >1340-1360 m
3. >1360-1380 m
4. >1380
1. 2001 or earlier
2. 2002
3. 2003
4. 2004
5. 2005
1. Shade
2. Partial shade
3. Full sun
1. Flat
2. Concave
3. Convex
1. 0:100
2. 10:90
3. 30:70
4. 50:50
1. Bachelor herds
2. Female herds
3. Territorial bulls
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement