THE INFLUENCE OF LARGE, VEGETATED TERMITARIA AND
LARGE HERBIVORES ON SPIDER (ARANEAE) DIVERSITY IN
MIOMBO WOODLANDS
By
Lenin Dzibakwe Chari
A thesis submitted in partial fulfillment of the requirements for
the degree of Master of Science in Tropical Resource Ecology
University of Zimbabwe
Faculty of Science
Department of Biological Sciences
Tropical Resources Ecology Programme
March 2011
1
ABSTRACT
This study reports on spider species richness, abundance, and spider species community
composition, in relation to differing herbivore impacts in miombo woodlands that have large
vegetated termitaria. The chosen woodlands formed a gradient of herbivore impacts from
Chizarira National Park that had high herbivore impacts mostly due to the presence of
elephants, to Chivero Game Park that had intermediate herbivore impacts and no elephant
population, and to the Chivero Bird Sanctuary that had minimal herbivore impacts. Pitfall
traps were used to sample spiders on the large termite mounds and in the adjacent woodland
matrix, and these spiders were consequently classified into morpho-species. In order to
explain any differences in spider species richness, abundance, and spider species community
composition the habitat structure was characterized through measuring ground cover, aerial
cover, and woody plant species richness. The Welch two sample t-test revealed no significant
differences in spider abundance between termite mounds and the woodland matrix in the Bird
Sanctuary, a higher abundance in the woodland matrix in the Chivero Game Park, and a
higher abundance on the termite mounds in the Chizarira National Park. The same results
were observed with spider richness except for a higher richness in the woodland matrix in the
Bird Sanctuary.
Generalized linear models showed that no single environmental variable was responsible for
observed patterns in spider abundance and richness in all the study areas. Instead, it was
established that various variable interactions (of different orders) of the aerial cover, ground
cover, woody plant richness, and site (termite mounds/woodland matrix) influence spider
richness and abundance differently, with the number of interactions increasing across the
herbivore impact gradient, from the least impacted (the Bird Sanctuary) to the most impacted
(Chizarira National Park). The importance of termite mounds in determining spider diversity
I
also seemed to increase across this gradient. Non metric multi-dimensional scaling and
hierarchical cluster analysis revealed different spider species community composition on
termite mounds and in the surrounding woodland matrix, and this difference increased from
the lowest herbivore impact area to the highest, with distinct spider assemblages being
realized in the most impacted woodland. It was therefore concluded that termite mounds are
not hotspots for the diversity of ground spiders as they are to other animals, but rather act as
refugia for ground dwelling spiders in highly impacted miombo woodlands.
Keywords
Spiders, Large vegetated termitaria, Diversity, Community composition, Miombo woodland,
Herbivory, Refugia
II
ACKNOWLEDGEMENTS
This study was supported by a South African National Research Foundation and Department
of Science and Technology (NRF-DST) collaborative grant, through the University of Cape
Town, to Professors Graeme Cumming and David Cumming. I am deeply indebted to David
Cumming, my supervisor and mentor, for guiding me every step of the way, saving me from
countless senseless errors, and for his encouragement throughout this study.
My appreciation goes to the Director General of the Zimbabwe National Parks and Wildlife
Management Authority (ZNPWMA) for granting me the permission to carry out this research
in Chizarira National Park and Chivero Recreational Park. I would also like to thank the
ZNPWMA staff at both these parks for welcoming me and assisting me in doing this
research. The Tropical Resource Ecology Programme (University of Zimbabwe) provided me
with all the equipment I used in the field and I am grateful for that. I also thank Shakkie
Kativu, Christopher Magadza, and all my lecturers for helping me become a true ecologist.
My heartfelt thanks go to Meg Cumming for introducing me to the world of spiders and
motivating me in ways she will never know. I would like to thank Zaccheus Mahlangu,
Colleen Seymour, Grant Joseph, Ben Heermans, and Allison Skidmore for assisting me in the
field and for making every moment spent out there an enjoyable experience. To Michael Tiki,
I am truly grateful for all the support you gave me, including giving me a place to stay during
my research at the Bird Sanctuary, when I was all out of options you took me in, thank you so
much.
Warmest thanks go to my family and friends: my parents, my brother, my sister, my friends;
Edwin, Gregory, Sydney, Tatenda, Trish, my classmates and Gloria, thank you all for your
III
support, friendship and love. Many thanks also go to Edwin Tambara for positively
criticizing me all the way and helping me produce a quality thesis. I would also like to thank
David Icishahayo for assisting me in doing Generalized Linear Models.
Finally I would like to thank the Almighty God for making all this possible.
IV
TABLE OF CONTENTS
Contents
Page
ABSTRACT ............................................................................................................................... I
ACKNOWLEDGEMENTS ..................................................................................................... III
TABLE OF CONTENTS .......................................................................................................... V
LIST OF TABLES ................................................................................................................. VII
LIST OF FIGURES .................................................................................................................IX
LIST OF APPENDICES ..........................................................................................................XI
CHAPTER 1: INTRODUCTION .............................................................................................. 1
1.1 General Introduction ........................................................................................................ 1
1.2
Main Objective ............................................................................................................ 3
1.3
Specific Objectives ...................................................................................................... 3
1.4
Hypotheses .................................................................................................................. 4
CHAPTER 2:LITERATURE REVIEW .................................................................................... 5
2.1
Miombo Woodlands and the Role of Termitaria in These Woodlands....................... 5
2.2
Large Herbivores ......................................................................................................... 8
2.3
Spiders as Indicators of Habitat Change ..................................................................... 9
2.4
Effects of disturbance and habitat structure on spider diversity ............................... 10
CHAPTER 3:STUDY AREAS AND METHODS .................................................................. 14
3.1
Study Areas ............................................................................................................. 14
3.1.1 Chizarira National Park ........................................................................................... 14
3.1.2 Lake Chivero Recreational Park .............................................................................. 17
3.2
Field Methods.......................................................................................................... 19
3.2.1 Sample sites ............................................................................................................. 19
3.2.2 Pitfall trapping ......................................................................................................... 21
3.2.3 Species Sorting and Identification ........................................................................... 23
3.2.4 Vegetation cover ...................................................................................................... 24
3.2.5 Woody plant species inventory................................................................................ 25
3.3
Data Analysis .......................................................................................................... 26
3.3.1 Species accumulation curves ................................................................................... 26
3.3.2 Welch two sample t-test .......................................................................................... 26
V
3.3.3 Similarity Indices, Multi-Response Permutation Procedure, and Hierarchical
cluster analysis .................................................................................................................. 27
3.3.4 Generalized Linear Models and Non Metric Multidimensional Scaling ................. 29
CHAPTER 4:RESULTS .......................................................................................................... 32
4.1
Spider Species Community Composition ................................................................. 32
4.2
Species Accumulation Curves ................................................................................... 34
4.3
Differences in Spider Species Abundance and Species Richness, Between Termite
Mounds and the Adjacent Woodland Matrix .......................................................... 36
4.4
Similarity Analysis ................................................................................................... 38
4.5
Influence of Ground Cover, Aerial Cover, Woody Plant Species Richness and Site
in Determining Spider Species Richness and Abundance ........................................ 43
CHAPTER 5:DISCUSSION .................................................................................................... 56
5.1
Sampling Adequacy .................................................................................................. 56
5.2
Patterns of Spider Species Richness, Abundance, and Community Composition .... 57
5.3
Influence of Habitat Characteristics .......................................................................... 59
5.4
Limitations of the Study ............................................................................................ 63
CHAPTER 6:CONCLUSION ................................................................................................. 66
REFERENCES ........................................................................................................................ 67
APPENDICES ......................................................................................................................... 84
VI
LIST OF TABLES
Table 4.1: The total numbers of spiders and morpho-species, in each family, in each
location. ....................................................................................................................................33
Table 4.2: Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in the Chivero Bird
Sanctuary..................................................................................................................................36
Table 4.3: Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in the Chivero Game Park.. ........37
Table 4.4: Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in Chizarira National Park.. .......37
Table 4.5: Summary statistics of ground cover, aerial cover, and woody plant species
richness measured in the Chivero Bird sanctuary.. ..................................................................44
Table 4.6: Summary statistics of ground cover, aerial cover, and woody plant species
richness measured in Chivero Game Park. ..............................................................................44
Table 4.7: Summary statistics of ground cover, aerial cover, and woody plant species
richness measured in Chizarira National Park.. .......................................................................45
Table 4.8: Results of a Generalized Linear Model (Type III) run on the Bird Sanctuary
data, with spider abundance as the response variable and site, aerial cover, and their
interactions as the factors.. .......................................................................................................46
Table 4.9: Results of a Generalized Linear Model (Type III) run on the Bird Sanctuary
data, with spider species richness as the response variable and site, aerial cover, and their
interactions as the factors.. .......................................................................................................46
Table 4.10: Results of a Generalized Linear Model (Type III) run on the Chivero Game
Park data, with spider abundance as the response variable and site, aerial cover, ground
cover and some interactions including plant richness as the factors. .....................................47
VII
Table 4.11: Results of a Generalized Linear Model (Type III) run on the Chivero Game
Park data, with spider species richness as the response variable and site, aerial cover, and
some interactions including plant richness and ground cover as the factors. . ........................48
Table 4.12: Results of a Generalized Linear Model (Type III) run on the Chizarira
National Park data, with spider abundance as the response variable and site, aerial cover,
ground cover, plant richness and some of their interactions as the factors. .............................49
Table 4.13: Results of a Generalized Linear Model (Type III) run on the Chizarira
National Park data, with spider species richness as the response variable and aerial cover,
ground cover and some interactions including site and plant richness as the factors. .............50
VIII
LIST OF FIGURES
Figure 3.1: A map showing the study areas; Chizarira National Park, and the Lake Chivero
Recreational Park. .................................................................................................................... 15
Figure 3.2: A picture showing a large vegetated termite mound surrounded by a heavily
impacted miombo woodland with no large trees, in the Chizarira National Park ................... 16
Figure 3.3: A picture showing a large vegetated termite mound surrounded by a miombo
woodland with relatively large trees, in the Chivero Game Park ............................................ 16
Figure 3.4: A picture showing a large and densely vegetated termite mound surrounded by a
miombo woodland with very large trees, in the Chivero Bird Sanctuary ................................ 17
Figure 3.5: Schematic representation of the arrangement of the study plots and sites in the
three different study areas ........................................................................................................ 20
Figure 3.6: Schematic diagram representing the arrangement of the pitfall traps on a single
sampling site.. .......................................................................................................................... 22
Figure 4.1: Species accumulation curves and distance curves for termite mound and
woodland matrix sites sampled in the three study locations .................................................... 35
Figure 4.2: Dendogram showing species shared between sites in the Chivero Bird Sanctuary
.................................................................................................................................................. 38
Figure 4.3: Dendogram showing species shared between sites in Chivero Game Park.. ....... 40
Figure 4.4: Dendogram showing species shared between sites in Chizarira National Park.. . 41
Figure 4.5: Box plot showing the Jaccard corrected similarity indices of the three study
locations for similarities in the number of spiders and species shared by termite mounds and
matrices .................................................................................................................................... 42
Figure 4.6: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chivero Bird Sanctuary ................................ 52
IX
Figure 4.7: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chivero Game Park....................................... 53
Figure 4.8: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chizarira National Park ................................ 55
X
LIST OF APPENDICES
Appendix A: Global Positioning reference points and vegetation cover data for all sites in the
Bird Sanctuary…………………………………………………………………………….…84
Appendix B: Global Positioning reference points and vegetation cover data for all termite
mounds sampled in the Chivero Game Park………………………………………………....85
Appendix C: Global Positioning reference points and vegetation cover data for all woodland
matrix sites sampled in the Chivero Game Park…………………………………………..…86
Appendix D: Vegetation cover data for all sites sampled in the Chizarira National Park…..87
Appendix E: Ground cover and aerial cover recording sheet………………………...……..88
Appendix F: Spider recording sheet…………………………………………………...…….89
Appendix G: Plant species recorded in the Chivero Bird Sanctuary, on termitaria and in the
woodland matrix……………………………………………………….……………….…….90
Appendix H: Plant species recorded in the Chivero Game Park, on termitaria and in the
woodland matrix………………………………………………………………………….…..92
Appendix I: Plant species recorded in the Chizarira National Park, on termitaria and in the
woodland matrix……………………………………………………………………………...94
XI
CHAPTER 1: INTRODUCTION
1.1
General Introduction
Biodiversity is currently undergoing dramatic changes worldwide (Wilson 1993), and lately
much emphasis has been placed on conserving and preserving one of the last preserves of
wildlife, the protected areas. The managers of these protected areas face many challenges as
managing biodiversity is a complex task. Due to the complexity of ecosystems, it is difficult
to conserve all wildlife species. Priorities have to be set, and the „more important‟ species
have to be protected.
In Southern Africa, the problem has been further exacerbated by the current increases in large
herbivore populations, particularly, the African elephant (Loxodanta africana). Over the past
century elephant numbers in Southern Africa have increased from a few thousand to more
than 300 000 (Cumming and Jones 2005) resulting in high elephant densities and consequent
changes in habitat structure (Anderson and Walker 1974; Cumming et al. 1997; Guy 1981,
Thomson 1974). Given the potential severity of this situation, it is important to detect
biodiversity change or loss early before the changes become irreversible (Barnes 1983).
The current study is part of a larger project entitled; “Indicators of large herbivore impacts
on biodiversity in southern Africa”. This project is an attempt to develop methods of rapidly
and effectively detecting biodiversity change using large vegetated termite mounds
(termitaria). Termitaria are a conspicuous feature in many savanna ecosystems and act as
nutrient hotspots (Holdo and McDowell 2004) in largely dystrophic savanna soils. Due to
their high nutrient status they also harbour a higher diversity of animal and plant species than
the surrounding matrix (e.g. Fleming and Loveridge 2003; Moe et al. 2009).
1
Termitaria are of particular interest to the project as it has been shown that the high nutrient
status of these mounds attracts large herbivores (Loveridge and Moe 2004; Holdo and
McDowell 2004). Consequently, in areas of high large herbivore density, particularly
elephants, it is expected that the first signs of overutilization will be seen on the mounds well
before they become apparent in the surrounding area.
Previous studies have examined woody plant species structure and diversity on termite
mounds and in the surrounding matrix (Humphrey 2008, Makumbe 2009); ant and reptile
diversity (Skidmore 2010, Heermans 2010); and large termitaria as refugia for hole nesting
birds (Joseph 2008, Joseph et al. 2011). The primary aim of the current study is to investigate
the influence of termitaria and large herbivores on the diversity of spiders (Order - Araneae),
in three areas with different levels of herbivore impacts on woody vegetation.
The knowledge of spiders in Southern Africa is largely limited to species description, while
their ecology remains relatively unexplored. In the last two decades however, there has been
an increase in research into the biodiversity and ecology of spiders in Southern Africa
(Haddad and Dippenaar-Schoeman 2002). An opportunity to further explore the biodiversity
and ecology of spiders in savanna habitats was therefore presented. The current study places
emphasis on spiders as little is known about how spider species abundance and richness are
related to landscape/spatial heterogeneity that is created by the large vegetated termite
mounds in savanna ecosystems.
It is likely that direct effects of herbivores on vegetation will result in an indirect influence on
spider diversity (e.g. Warui et al. 2005). Since spiders are generalist predators abundant in
most terrestrial ecosystems (Snyder and Wise 1999; Snyder and Wise 2001) the population of
2
other invertebrate taxa is therefore expected to be an indirect function of spider population, as
foraging spiders are considered the major agent controlling insect communities in terrestrial
ecosystems (Young and Edwards 1990). An investigation into the local spider diversity could
therefore reveal the indirect impacts of herbivory on invertebrate taxa.
The general hypothesis that invertebrate diversity reduces with an in increase in disturbance
(Warui et al. 2005) or decrease in habitat complexity (Robinson 1981; and Balfour and
Rypstra 1998), was therefore investigated with spiders as the target invertebrate group. The
contribution of large vegetated termitaria to this association between spiders and the level of
herbivory was also assessed.
1.2
Main Objective
To assess the influence of the presence of large vegetated termitaria on the diversity of
spiders in miombo woodlands subjected to differing levels of herbivore impacts.
1.3
Specific Objectives
1) To determine the spider species richness and abundance on large vegetated termite
mounds and in the surrounding woodland matrix in Chizarira National Park, Chivero
Game Park, and Chivero Bird Sanctuary; areas with differing levels of large herbivore
impacts.
2) To examine the influence of vegetation cover and woody plant species richness on spider
species richness, abundance, and spider species community composition on large
vegetated termitaria and in the surrounding woodland matrix, in the three locations.
3
1.4
Hypotheses
Given that the overall objective of this study was to determine the influence of large
termitaria and large herbivores, on spider species richness and abundance, the response
variables were spider species richness and spider species abundance. The proposed major
explanatory variables were therefore location (Chizarira National Park, Chivero Game Park,
and Chivero Bird Sanctuary) and site (termite mound and woodland matrix). The different
locations and sites were expected to have different habitat characteristics due to the impacts
of large mammals on woody plant species richness, aerial cover, and ground cover.
The hypotheses on which this study is based upon are as follows:
1.
Spider species richness, abundance, and spider community species composition on
termite mounds differ from those in the adjacent woodland.
2.
Woody plant species richness, ground cover, and aerial cover on termite mounds are
different in the adjacent woodland.
3.
The interactions between site (termite mound and woodland matrix), aerial cover,
ground cover, and woody plant species richness determine spider species richness,
abundance, and spider species community composition.
4.
The interaction between ground cover and site determines spider abundance, richness,
and composition.
5.
Similarity between large vegetated termitaria and woodland matrix spider diversity
and community composition diminishes along a gradient of herbivore impacts, from
the least impacted to the most impacted area.
4
CHAPTER 2: LITERATURE REVIEW
2.1
Miombo Woodlands and the Role of Termitaria in These
Woodlands
The word miombo is a term used to describe woodlands in the central, southern and eastern
parts of Africa, which are dominated by the genera Brachystegia, Julbernardia and/or
Isoberlinia, three closely related genera from the legume family (White 1983). Brachystegia,
Julbernardia and/or Isoberlinia may be the dominant species in miombo woodlands but at
any point there can be considerable heterogeneity in plant physiognomy, structure and
diversity, reflecting the variation in soils (Campbell and Du Toit 1988), rainfall, and the
impacts of fire (Lawton 1978), land use (Robertson 1984; Chidumayo 1987), herbivory
(Anderson and Walker 1974; Thomson 1974; Guy 1981; 1989) and other disturbances.
These miombo woodlands have been described (Frost 1996) as the most extensive type of
tropical seasonal woodland and dry forest formation in Africa (perhaps even globally),
covering an estimated 2.7 million km2 in regions receiving >700 mm mean annual rainfall on
nutrient-poor soils. The miombo woodlands are characterised by high plant diversity and
endemism (White 1983) and have recently been described as one of the world‟s biodiversity
hotspots (Mittermeier et al. 2003).
This juxtaposition of infertile miombo and other, more fertile, moister and productive
vegetation types has already been described (Frost and Robertson 1987) as an important
factor in maintaining populations of large wild and domestic herbivores in miombo
woodlands, dependent on the extent and degree of interspersion of the vegetation types (Frost
and Robertson 1987). A number of workers (Wild 1952, Loveridge and Moe 2004, Holdo and
5
McDowell 2004) have cited the inclusion of habitat islands of non-miombo as an enhancer of
overall wildlife diversity in miombo woodlands.
Termites make up the greatest contribution to total soil macro-fauna biomass in many tropical
ecosystems, with values comparable to ungulate biomass in African savannas. Termites in the
tropical savannas function as ecosystem engineers as do elephants (Jones et al. 1994; Laws
1970; Dangerfield et al. 1998), by modifying the physical habitat and creating islands of high
soil fertility. They do this by influencing the spatial and temporal distribution of water,
carbon, and nutrients through their mound structures (Lavelle 1997; Brown et al. 2000;
Dangerfield et al. 1998).
The importance of Macrotermes and other Macrotermitinae lies in their dependence on
cellulose-decomposing fungi which they cultivate in their mounds. To maintain the fungi the
termites forage widely, collecting surface litter and dried grass which is carried back to the
mounds and decomposed by the fungi. Because of the ability of the fungi to produce
cellulase, almost all of this organic matter is decomposed (Jones 1990).
Termites, especially species of the genus Macrotermes construct large epigeal nests and
extensive underground gallery systems. Through their foraging behaviour, termites localise
nutrients on their mounds thus influencing nutrient flow rates and the spatial distribution of
nutrients. They also relocate soil particles for mound construction and maintenance, and as a
result influencing soil physical properties (Dangerfield et al. 1998). Large termite mounds
built by the termite (isopteran) genus Macrotermes are conspicuous in the miombo landscape
(Malaisse 1978). The presence of these large vegetated termite mounds creates heterogeneity
in a largely dystrophic landscape (Scholes and Walker 1993), on which a unique suite of
6
trees, shrubs and grasses occur (Moe et al. 2009; Traore et al. 2008; Loveridge and Moe
2004).
Termitaria function as islands of local diversity in the miombo system (Fleming and
Loveridge 2003). They are an important resource used by small mammals (Fleming and
Loveridge 2003) and large ungulates (Holdo and McDowell 2004) and support a higher
diversity of plant and animals than the surrounding matrix (Malaisse 1978). The vegetation
on the mounds is often the focus of activity for birds (e.g. Joseph et al. 2011) and other
animals, enabling these species to exist in an otherwise largely unproductive environment
(Frost and Robertson 1987).
In particular, the high nutrient content of trees on termitaria makes them attractive to mega
herbivores (Loveridge and Moe 2004) such as elephants, giraffes and other large herbivore
browsers (Ruggerio and Fay 1994) in nutrient poor savannah systems. Aside from the basic
requirement for water, animals respond to spatial variability by selecting patches or areas
which offer the highest intake of digestible nutrients (O'Reagain, and Schwartz 1995). It is
therefore expected that large herbivores will prefer to browse on these nutrient hotspots and
this preferential selection for termitaria vegetation is expected to result in varying impacts on
vegetation cover, on and off (Mobaek et al. 2005) vegetated termite mounds. Depending on
the intensity of herbivore browsing, termitaria may thus provide a simple indicator of large
herbivore impacts on biodiversity, and habitat change, and should therefore be considered as
a focus of conservation in miombo woodlands.
.
7
2.2
Large Herbivores
Large mammalian herbivores exert a direct impact on vegetation by their consumption of
plant parts, and breaking or trampling plants. One indirect impact is the removal of bark
(especially by elephants) making the plants more susceptible to fire and attack by wood
boring beetles (Owen-Smith 1988). Browsing pressure by large herbivores such as elephant
(Loxodonta africana), black rhino (Diceros bicornis) and giraffe (Giraffa camelopardalis)
can supress the regrowth of woody plants, and as a result, keeping them within the fire
susceptible zone for longer (Owen-Smith 1988).
The African elephant (Loxodonta africana) in particular, has been cited as an example of a
large, generalist herbivore which has been responsible for considerable transformation of
natural habitats in protected areas (e.g. Anderson and Walker 1974; Cumming et al. 1997;
Guy 1981; and Thomson 1974) where it has been protected from human predation
(poaching).
Elephants play a key role in the ecology of their habitats. For example, their feeding habits
open up thick bush and forest for grazing species; they also maintain waterholes and keep
open forest pathways used by wildlife and humans (Carroll 1988). Trees destroyed by
elephants are replaced by regenerating shrubs or grasses that offer more accessible foliage for
consumption by smaller herbivores. The mosaic diversity of habitats created by these impacts
on vegetation is important in promoting the coexistence of a wide diversity of other
mammalian herbivore species; a keystone role (Owen-Smith 1988).
With global human populations being on the rise and doubling in the past 40 years (Cohen
2003), areas previously occupied by wildlife have been taken over by humans (e.g. Blanc et
8
al. 2003; Gratwicke and Stapelkamp 2006) and as a result shrinking the range available to
elephants (Owen-Smith 1988) and other large mammals. The decline in elephant range,
accompanied by population increases, has resulted in high elephant densities and a loss of
habitat structure (e.g. Fenton et al. 1998), as elephants tend to impact heavily on the tall tree
component of woodlands and savannas (Laws 1970; Anderson and Walker 1974; and
Mapaure 2001). Several studies have shown that high densities of elephants can cause loss of
biodiversity (e.g. Cumming et al. 1997), substantial changes in tree species composition
together with fire (Mapaure 2001), reduction in tree density and biomass (Guy 1981),
lowering of bird and ant species richness (Cumming et al. 1997), and changes in bird species
composition (Herremans 1995), amongst others. Limited biodiversity surveys are therefore a
worrying issue, especially in protected areas.
Preserving large populations of elephants whilst at the same time maintaining biodiversity in
national parks and protected areas is therefore challenging. In Africa the problem is worsened
by the lack of funding for ecological research. Consequently, by the time that elephant and
other large herbivore impacts on woodland become obvious it is generally too late (Barnes
1983) to take any corrective measures so that developing early warning systems is important.
2.3
Spiders as Indicators of Habitat Change
According to Bouyer et al. (2007), a good ecological indicator should be sensitive to slight
ecosystem changes in a predictive manner, thereby allowing the detection and measurement
of the effect of various disturbances to the ecosystem. Invertebrates are critical components
of ecosystems and can make excellent bioindicators (Kremen et al. 1993) of ecosystem health
and change as they are sensitive to change, and can be used to index changes in the
environment at small, spatial and very short, temporal scales (Ginsberg 1993; and Schroeter
9
et al. 1993). The large numbers of invertebrates makes them more amenable to statistical
analyses than vertebrate data (Kremen et al. 1993).
Many studies, including Downie et al. (1999) and New (1999) have demonstrated that spiders
in particular are extremely sensitive to small changes in the habitat structure, including
habitat complexity, litter depth and microclimate. Spiders are generalist predators abundant in
most terrestrial ecosystems (Snyder and Wise 1999, Snyder and Wise 2001) and the
population of other invertebrate taxa is therefore an indirect function of spider population.
Thus among arthropods, spiders are probably one of the best target groups for use as indicator
species of disturbance or habitat change. They are hyper-diverse yet can be easily sampled
and sorted to morphospecies and they are probably the most abundant representatives of the
top-predators guild in many habitat types (Cardoso et al. 2008). Several studies of ecology
and biodiversity have already proven this (Bonte et al. 2004; Lambeets et al. 2007; Negro et
al. 2010). In addition to being highly diverse, spiders are abundant and inhabit a wide array
of spatial and temporal niches (Juen and Traugott 2004; Vasconcellos-Neto 2005; Entling et
al. 2007). As a result of their high abundances and insectivorous behavior spiders are useful
indicators of the ecological status of biotic communities, and of changes in habitat and
landscape structure (Warui et al. 2005; Foord et al. 2008; and Horvath et al. 2009).
2.4
Effects of disturbance and habitat structure on spider
diversity
Ecological studies of invertebrates have shown that structural habitat complexity affects
species diversity (e.g. Dean and Connell 1987; and Magagula 2003) and therefore any
disturbance results in a reduction in complexity leading to reduced species diversity. The
10
diversity of ground beetles, plant hoppers and spiders has already been related to botanical
diversity and the structural variability of vegetation (Cherrill and Rushton 1993; Downie et
al. 1999; Sanderson et al. 1995; Dennis et al. 1998; Siemann et al. 1999). In addition, other
researches (Lawton 1983; Halaj et al. 2000) have shown that more complex vegetation
provides arthropods with sites for shelter, foraging, oviposition, and mating. Such conditions
ideally support an increase in spider diversity.
Spiders in particular, have been found to be favoured by complex habitats (Robinson 1981;
Gunnarsson 1988; Balfour and Rypstra 1998; Raizer and Amaral 2001) while Dean and
Connell (1987) showed that increased structural habitat complexity promoted increase in
species diversity. Halaj et al. (2000) also reported that structural habitat complexity had a
profound effect on canopy spiders and other arthropods. Ysnel and Canard (2000)
demonstrated that the foliage orientation influences species composition of spider
communities. More work supporting importance of habitat complexity on spiders can be
found in Greenstone (1984), and Buddle and Rypstra (2003). In an attempt to explain the link
between habitat complexity and spider diversity Rypstra (1983) and Wise (1993) found that
availability of unique habitat structural features allows more efficient prey capture and may
limit spider species populations more than the availability of food itself.
Spiders are predominantly generalist feeders that primarily attack insects, but also eat other
arthropods, including spiders (Wise 1993). They possess neurotoxins that enable them to kill
prey rapidly. The prey is usually smaller than or similar in size to the spider, but many
spiders subdue prey several times their mass (Wise 1993). According to Uetz (1991), there
are several reasons why spiders should be more sensitive to structure than other organisms.
As a group, spiders perceive their environment using vibratory cues which are mediated
11
through the substrate on which they live. Web spiders must anchor their prey capture device
to the appropriate substratum and complex habitats provide appropriate sites for a greater
range of sizes and types of webs. Finally, since all spiders are predators that can potentially
consume one another, the extent to which they can coexist may strongly depend on their
ability to move around and hide in a complex environment. Spider populations are therefore
likely to be largely determined by any habitat change or disturbance.
A study by Cumming and Wesolowska (2004) in a suburban area showed that different
microhabitats host a consistent and predictable cluster of jumping spider (Salticidae) species,
although some species are habitat generalists and occupy a range of habitats. A detailed study
(Butler and Haddad 2011) of the relationship between spider assemblages and different litter
types also showed that the habitat largely influences spider composition and abundance, as
the shallower and more compact litter in shady areas was discovered to support higher
abundance and species richness of spiders with similar assemblage structure. It is possible
that different litter microhabitats have a varied influences on the microclimate (e.g.
temperature and humidity) and prey availability (Uetz 1979) resulting in different spider
communities.
A general rule therefore seems to be that as disturbance increases spider species richness
declines, as plant community structure, and ecosystem dynamics such as disturbance
influence spider assemblages (Bonte et al. 2002). Apart from the study by Haddad and
Dippenaar-Schoeman (2002) on the influence of Trinervitermes trinervoides mound structure
on spider diversity, no research on spiders associated with termite mounds has been
conducted in southern Africa. Few studies have been conducted on the association of spiders
with termites (Dippenaar and Meyer 1980; van den Berg and Dippenaar-Schoeman 1991;
12
Jocque and Dippenaar-Schoeman 1992; Dippenaar-Schoeman et al. 1996; Cumming 1993;
and Wesolowska and Cumming, 1999).
Since spiders are sensitive to changes in habitat structure, their diversity on mounds and in
the woodland matrix of savanna woodlands may be very sensitive to the effects of large
herbivore density and browsing pressure. In the current study it was expected that the
presence of herbivores would reduce the relative vegetation cover by trampling, browsing,
and grazing, thereby reducing the habitat complexity. This in turn was expected to influence
spider diversity.
13
CHAPTER 3: STUDY AREAS AND METHODS
This study was carried out in three different miombo woodland locations, with different large
herbivore systems, and characterized by the presence of large vegetated termitaria. These
areas were; Chizarira National Park, Chivero Game Park, and Chivero Bird Sanctuary with
high levels of herbivore impact in Chizarira, intermediate levels in the Chivero Game Park
and little if any large herbivore impact in the Bird Sanctuary.
3.1
Study Areas
3.1.1 Chizarira National Park
The first part of this study was carried out in Chizarira National Park during the months of
October and November 2009. The park is situated in the North West part of Zimbabwe
(17°32' - 18°15', 25°35' - 28°13') in the Zambezi Valley and it covers a total area of 1910 km²
(Thomson 1974) and has an elevation of between 700 and 1400 m (BirdLife International
2011). The park is bounded by communal lands of the Binga and Gokwe Districts and shares
a boundary with Chirisa Safari Area to the south.
The area experiences a wet season from November to April, a cool dry season from May to
July and a hot dry season from August to November. Mean annual temperature is 20–22.5°C
(maxima: October 32.5–35°C, July 22.5–25°C; Torrance 1965 in Joseph et al. 2011). Mean
annual rainfall is 600–800 mm (annual coefficient of variation of 25–30%; Lineham 1965 in
Joseph et al. 2011).
The vegetation at Chizarira is dominated by the miombo tree genera Brachystegia and
Julbernardia (Fabaceae, subfamily Caesalpiniodeae). Other common vegetation types within
the miombo ecoregion are mixed Combretum and Colophospermum mopane woodlands
14
(Campbell et al. 1996). In 1972 large areas of Brachystegia boehmii woodlands in Chizarira
were reduced by elephants and fire and converted into shrublands (Thomson 1974, Cumming
1981). Elephant densities over the past 30 years have tended to be high, at 1 per km2, and at
times as high as 3 per km2 (Cumming 1981; Dunham et al. 2006). A prominent and
conspicuous feature in a large part of the park is that large trees are confined to large termite
mounds and large trees are seldom found in the open woodland (Figure 3.2).
The impacts of elephants and fire have transformed most of the park‟s former tall, open
woodland to shrubland (Cumming 1981) comprised mainly of Combretum species and
regenerating Brachystegia boehmii in the matrix, with tall trees being largely confined to
termitaria.
Figure 3.1: A map showing the study areas; Chizarira National Park, and the Lake Chivero
Recreational Park; Zimbabwe.
15
Figure 3.2: A picture showing a large vegetated termite mound surrounded by a heavily
impacted miombo woodland with no large trees, in the Chizarira National Park (Taken by
David Cumming, 2007).
Figure 3.3: A picture showing a large vegetated termite mound surrounded by a miombo
woodland with relatively large trees, in the Chivero Game Park (Taken by Lenin Chari, 2010)
16
Figure 4.4: A picture showing a large and densely vegetated termite mound surrounded by a
miombo woodland with very large trees, in the Chivero Bird Sanctuary (Taken by Lenin
Chari, 2010).
Mammalian herbivores in the park include elephant (Loxodanta africana), buffalo
(Cyncernus caffer), sable (Hippotragus niger), greater kudu (Tragelaphus strepsiceros),
waterbuck (Kobus ellipsiprymnus), eland (Tragelaphus oryx), impala (Aepyceros melampus),
zebra (Equus quagga) and common warthog (Phacochoerus africana), with elephant and
buffalo having the highest population density (Dunham et al. 2006).
3.1.2 Lake Chivero Recreational Park
Lake Chivero Recreational Park is about 6 100 hectares in extent including the 2630 ha lake.
In order to maintain an atmosphere appropriate to different forms of park use, and so as to
minimize the conflicts between antagonistic forms of use, the park was divided into three
major zones (Parks and Wildlife Board 1975). These zones are the south bank that largely
17
consists of the game park, the lake zone that includes the lake and its islands, and the north
bank that is largely devoted to outdoor recreation (Figure 3.1). The current study took place
in the game park inside the south zone, and in the relatively undisturbed woodland in the Bird
Sanctuary on the North bank.
The Chivero area is dominated by miombo woodland; Brachystegia spiciformis and
Julbernada globiflora with associated trees such as Terminalia sericea, Parinari
curatellifolia, Monotes glaber, and Burkea africana (Malinga 2001).
The Game Park of 1867ha extends from the Tiger Bay (Figure 3.1) to the Bushman‟s point
on the South Zone. It was opened in 1962 holds a variety of large mammals; most of which
were introduced from the Hwange National Park. Browsers in the Game Park include; giraffe
(Giraffa cameleopardalis), eland (Taurotragus oryx), and greater kudu (Tragelaphus
strepsiceros). Grazers include tsessebe (Damaliscus lunatus), wildebeest (Connochaetes
taurinus), zebra (Equus burchelli), white rhino (Ceratotherium simum), sable (Hippotragus
niger), and waterbuck (Kobus ellipsipyrimnus). Mixed feeders present are impala (Aepyceros
melampus). Grey duiker (Sylvicarpra grimmmia) and common warthog (Phacochoerus
africana) are also present. A part of this study was carried out inside the Game Park within
the “Ostrich loop” (17°55.333'S, 30°48.869'E, Figures 3.1 and 3.3).
The Bird Sanctuary (17°54.740'S 30°50.438'E, Figures 3.1 and 3.4) extends into the Lake
zone and interests of avifauna are paramount in the area. The management of the sanctuary is
aimed at conserving as many birds of the widest possible range of species, particularly of
those species dependent on the aquatic environment. Facilities are also provided for bird
watching (Parks and Wildlife Board 1975). The only mammalian herbivores present within
18
this area are occasional reed buck (Redunca arundinum) sighted close to the lake shore, bush
pig (Potamochoerus porcus) and common warthog (Phacochoerus africana) sighted in the
woodland (personal observation). This area has been devoid of large browsers for the past 50
years Cumming (pers. comm.) and therefore acts as the control in this investigation where the
effect of herbivory on spider diversity is the objective. The area therefore represents an area
with very low disturbance and minimal mammalian herbivory.
The Chivero Game Park and the Bird Sanctuary were included in this study as they provided
a miombo study area with two comparable sites with large vegetated termitaria, one with
large herbivores but no elephant, and one without large herbivores.
3.2
Field Methods
3.2.1 Sample sites
Aerial photographs were used to select an area with a considerable number of termite mounds
within each of the three locations. In each location, a single termite mound was chosen
randomly and thereafter every second nearest mound from the first mound was sampled.
Sampling of every second nearest mound was done to reduce the effects of spatial
autocorrelation. A total of 20 mounds were selected in the Chivero Bird Sanctuary, 30 in
Chivero Game Park, and 12 in Chizarira National Park, providing a sample of 62 plots, each
comprising a paired mound and matrix site.
For every termite mound selected, an adjacent woodland area, of the same diameter as the
mound, was demarcated 25m from the edge of the mound (Fig. 3.2). Each matrix site edge
was at least 25m from the mound edge along a cardinal compass bearing. The bearing (N, E,
19
S or W) from the mound plot was determined using random numbers (1-4) from a random
numbers table. If another mound was less than 60m away in the chosen direction, an
alternative direction was chosen (also randomly). Figure 3.2 clearly illustrates the general
layout of study plots and sites within the three locations. The locations of all the mounds and
matrices were recorded using a hand held Global Positioning System (GPS) coordinates. The
center GPS coordinates of the matrix and the mound were the ones recorded.
Figure 3.5: Schematic representation of the arrangement of the study plots and sites in the
three different study areas, Chizarira National Park, Chivero Game Park, and Chivero Bird
Sanctuary.
20
The heights and diameters of the mounds were measured with a ranging rod and a measuring
tape, respectively. The surface area of the mounds was approximated by treating the shape
either as a cone or a half sphere and using the arithmetic formula to find the surface area of
the respective shapes. A circular plot of the same area as the corresponding mound was
demarcated in the adjacent matrix.
3.2.2 Pitfall trapping
Five pitfall traps were placed on each site, at the four cardinal points, north, south, east, and
west, and also at the centre (Figure 3.3). The four traps on the cardinal points were installed
20% inside the site boundary (circular sub-plot) and emptied every 24 hours for a 7 day
trapping period in Chizarira and 4 day trapping periods in Chivero (Bird Sanctuary and Game
Park). Each trap consisted of 2 conical cups with a top diameter of 9.5cm and bottom
diameter of 7.5cm and depth of 8.7cm that fitted into each other. Holes were dug in the
ground using a soil auger and a garden trowel at the predetermined points. Two cups were
slotted into the hole, with one inserted into the other. The mouth of the top cup was levelled
with the ground. The cup beneath was used to simply maintain the hole and the one on top
was ¼ filled with water. A few drops of detergent were added to the water so as to break the
surface tension and allow any spider (or invertebrate) that fell in to sink.
Traps were checked every morning and the catch was placed into appropriately labelled film
canisters, 30 ml in volume. All spiders collected from the 5 pitfall traps on each site were put
into one container with water to prevent the specimens from desiccating. Spiders from the
five traps at each site were pooled in order to reduce stochastic heterogeneity among samples
and homogenize sampling effort (Cardoso et al. 2008), making each mound site comparable
21
to the adjacent woodland matrix site. Upon collecting the catch from the traps the cups were
reinstalled in the ground and refilled with water and detergent added.
Figure 3.6: Schematic diagram representing the arrangement of the pitfall traps on a single
sampling site. Traps at the cardinal point were installed 20% inside the sites as a way of
standardizing the placement of traps. The same arrangement was used on all plots in the three
different study locations.
After collecting all the catches from all the traps (all the sites), the spiders were separated
from the debris and other invertebrates in the catch. This was done in a white tray with the
use of forceps to pick out the spiders, and water to wash away the debris. These spiders were
examined under a field stereomicroscope and consequently identified to family level and
assigned to morpho-species. The spider specimens were consequently put into well labelled
5ml and 2ml cryovials, in 70% ethyl alcohol, for future full taxonomic identification by
experienced personal.
22
In the Chivero area the same procedure as used in Chizarira was employed except that pitfall
trap catches on each site were not pooled but kept separate so as to test for intra site
heterogeneity at a later stage.
3.2.3 Species Sorting and Identification
Spiders collected were classified up to the family level, a measure commonly used to
examine community level patterns (e.g. Whitmore et al. 2002). Taxonomic keys used include
Filmer (1991), Leroy and Leroy (2003), and Dippenaar-Schoeman (2002).
Spiders were sorted into morphospecies, based on outward morphological characteristics.
These features included body colour, pattern, and relative sizes of segments, eye pattern and
number, leg spines, hairs and length, and the number and type of spinnerets. Each
morphospecies was allocated a family name and number after looking for groups of
morphologically indistinguishable spiders, followed by describing briefly the set of
characters unique to each group.
Juveniles were not distinguished from adults and it is appreciated that this may have resulted
in some juveniles being classified as separate morpho-species. It is well established that
juveniles often do not resemble adults and sometimes males may be morphologically distinct
from females (Derraik et al. 2002). It is for these reasons that spider abundance carries a
greater weighting than the number of species in this study. Further assistance will be sought,
in due course, in identifying all spiders sampled.
23
3.2.4 Vegetation cover
Vegetation cover was estimated in terms of percentage aerial cover and ground cover in all
three locations but different methods were employed between Chizarira National Park and the
two locations in the Lake Chivero area (the Game Park and the Bird Sanctuary).
In Chizarira National Park, three parallel transects were traversed across each mound and
across the adjacent circular matrix site. The middle transect was the diameter of the circular
site and the other two transects were equidistant from this diameter and 20% inside the
circular site. A measuring tape was laid down on each of the 3 transects. Moving along one
end of the tape, the intercept distance of any plant (woody or herbaceous) that had any part of
its body hanging over the transect, was recorded. The total distance intercepted was added
and divided by the total length of each transect, and then multiplied by 100%. This was done
for all three transects and the aerial cover estimate of each subplot was obtained by adding
and averaging the percentage cover values of the three transects. A single aerial cover
estimate was as a result used for each subplot. On the same transects used for aerial cover,
ground cover was also estimated. This was done by recording the distance covered by any
plant material on the ground, adding it up, and expressing the accumulated distance as a
percentage of the transect length. At each site the estimates for the three transects were
averaged to produce one estimate.
In the Chivero area, vegetation cover was measured using a modified pin frame method. Only
a single pin (instead of a frame with several pins), approximately 4mm thick, was used to
place predetermined points on two predetermined transects in each sub-plot (site). One
transect was set on the east-west direction and the other on a north-south direction at all sites.
These transects were placed at 90 degrees to each other, intercepting each other at the centre
24
of the site. A measuring tape was laid down on each transect and its length (diameter of subplot) was divided by 100 and the pin was dropped at each 1/100th point of the transect.
At each 1/100th point of the transect, the wire pin struck the point and a short description of
the ground and aerial cover was consequently given (see Appendix E). Plant species, dead
leaves, bare ground, wood, or any kind of ground or aerial cover hit, was recorded for each
hit. If more than one plant or plant part touched the pin it was still considered as one hit.
Relative ground and aerial cover estimates were subsequently obtained by counting the
number of hits and expressing them as percentages. Since two transects were traversed per
sub plot, the two values were averaged to give a single cover estimate (ground and aerial) for
each sub-plot.
Aerial cover was only estimated up to a height of 1m as it was assumed that ground dwelling
spiders were not likely to be affected by aerial cover greater than 1m in height. Anything
dead or alive touching the 1m long wire was taken as aerial cover and ground cover was
taken as anything that the point of the pin hit.
3.2.5 Woody plant species inventory
A complete inventory of all woody plant species in all the study sites was made and plants
were classified to species level. In order to characterise woodland structure for each site
(mound/matrix) three height classes were constructed and the height class of each woody
plant was recorded. The three height classes were as follows; >1m, 1-3m, and >3m, where
the plants below 1m were classified as shrubs and those above as trees (1-3m) and tall trees
(>3m). The data for plant structure were not used in this study.
25
Woody plant species were classified to species level with the aid of field identification guides
and the help of an expert botanist, Zaccheus Mahlangu. Where species were not identified in
the field, leaves, small fruits, and flowers were pressed for later identification at the Harare
National Herbarium. Data on plant species for 15 plots in Chivero Game Park and 15 plots in
the Chivero Bird Sanctuary were adopted from an earlier study by Makumbi (2009).
3.3
Data Analysis
Statistical analyses were carried out using: (i) PC-ORD 5.1 (McCune and Mefford 2006),
(ii) SPSS Version 16.0 for Windows (SPSS Inc. 2007), (iii) Estimates v 7.5 (Colwell 2004),
and (iv) R Version 2.12.0 (The R Foundation for Statistical Computing 2010).
3.3.1 Species accumulation curves
Species accumulation curves (Gotelli and Colwell 2001) were plotted using PC-ORD 5.1
(McCune and Mefford 2006), in order to evaluate the adequacy of sample size in the spider
community data set. Species accumulation curves and Sorenson distance curves were plotted
for each location (Chizarira National Park, Chivero Game Park, and Chivero Bird Sanctuary)
using the number of species recorded for each site to determine the adequacy of sampling for
species richness (McCune and Mefford 2006).
3.3.2 Welch two sample t-test
The Welch two sample t-test (an unequal variance t-test) was used to test for differences in
spider species richness and abundance on mounds and in the matrix, in three locations in
order to correct for differences in variance between sites. This test was run in R Version
26
2.12.0 (The R Foundation for Statistical Computing 2010), upon testing for normality (using
Q-Q plots in SPSS Version 16 for Windows) and homogeneity of the variance (Levene‟s test
in SPSS Version 16 for Windows) in the spider data sets of the mounds and the matrix.
Means and the corresponding confidence intervals were calculated to enable direct
comparison between the mounds and the woodland matrix.
At each location, ground cover estimates, aerial cover estimates and woody plant species
richness of termite mounds were compared with those of the woodland matrices. The Welch
two sample t-test was also used to ascertain the statistical significance of the differences. This
was done in order to determine if there was any variability in terms of vegetation cover and
plant species richness between the two types of sites (mound and matrix), in the three
locations.
The primary reason for using the Welch t-test (as opposed to a standard t-test) was to
compensate for differences in variance (Siegel 1956) between sites. The Welch t-test does
this by adjustsing the degrees of freedom. According to Zar (1996) the Student‟s t-test
performs badly when variances are unequal.
3.3.3 Similarity Indices, Multi-Response Permutation
Procedure, and Hierarchical cluster analysis
Comparisons of spider community structure, on and off termitaria, were achieved through the
use of similarity indices that were compared across the three different locations. The
programme ESTIMATES v 7.5 of (Colwell 2004) was used to calculate the new Jaccard
corrected index (Chao et al. 2005). This index is both abundance and probability based and
has been proved (Chao et al. 2005) to reduce under-sampling bias by estimating and
27
compensating for the effects of unseen, shared species. Chao et al. (2005) also recommended
this index (as well as the Sorensen corrected index) for assessing species composition
similarity between samples that differ in size, contain numerous rare species, and are
suspected (or known) to be undersampled. This study was complicated by all three issues
stated above.
The Multi-Response Permutation Procedure (MRPP) was used to investigate differences in
spider species composition between mounds and matrices. It was computed using a Sorenson
(Bray-Curtis) distance measure and a natural weighting. The MRPP is a non-parametric
procedure used to test the hypothesis of no difference between two or more groups of entities
(PCORD V5 McCune and Mefford 2006). The MRPP was opted for in place of discriminant
analysis as it has the advantage of not requiring the assumptions of multivariate normality
and homogeneity of variances, which according to Biondini et al. 1985, are rarely met with
ecological community data.
The MRPP uses a test statistic; A that is a descriptor of within-group homogeneity. When all
items are identical within groups A = 1 (Mielke 1984), the highest possible value for A. If
heterogeneity within groups equals expectation by chance, then A = 0 (Mielke 1984). It is
therefore expected that the more homogeneous the mounds and matrices were, the higher the
A value would be. The statistical significance (p-value) was given for each A value between
mounds and matrices, in each location.
A hierarchical cluster analysis (PCORD V5 McCune and Mefford 2006) was used to
represent any differences in spider species composition among the sites, on a dendogram. The
classification was based upon the Bray-Curtis distance measure, a statistic used to quantify
28
the compositional dissimilarity between two different sites, and the unweighted pair-group
average (UPGAMA) was the linkage method used to plot the dendrograms for each location.
3.3.4 Generalized Linear Models and Non Metric
Multidimensional Scaling
A Generalized Linear Model (GLM), Type III was performed in order to determine if the
proposed explanatory variables; site, aerial cover, ground cover and woody plant species
richness had a significant influence on spider species richness and species abundance, the
response variables.
According to Crawley (2007) when count datum is the response variable there are often lots
of zeros in the data frame and the variance may therefore increase linearly with the mean.
Regression analysis could not be used in this study as it assumes that variance is constant.
GLMs were appropriate in this study as both the response variables (spider richness and
abundance) were count data, and variances were unequal. GLMs were run with a Poisson
distribution to account for the non-normality of the count data and a log link function was
used to ensure that fitted values were positive (since it is not possible to have counts less than
zero).
The Generalized linear model was run as a factorial model taking into account the four
explanatory variables and any interactions, of different orders, between the explanatory
variables in determining spider species richness and abundance. A stepwise progression from
the maximal model through a series of simplifications to the minimal adequate model was
made on the basis of deletion tests. The maximal model consisted of all the four explanatory
variables and all 2-way, 3-way, and 4-way interactions between the variables. Model
simplification was carried out by removing the least significant terms (largest p-value) first
29
starting with the highest-order interactions. These deletion tests were chi-squared tests that
assessed the significance of the increase in deviance that resulted when a given term (variable
or interaction of variables) was removed from a particular model.
In the final model, the Wald Chi-square test was used to test the true value of the model
parameter based on the sample estimate, for each factor in the model. This test was used
precisely to test for model effects without emphasising on the actual parameter estimates. The
greater the Wald Chi-square the greater the probability of the factor being significant hence
the p-value was also given. Degrees of freedom were also displayed.
SPSS Version 16.0 for Windows (SPSS Inc. 2007) was used for the GLM analyses.
Non Metric Multidimensional Scaling (nMDS) ordinations were used to graphically represent
the relationship between sites, in terms of spider species community composition, in
multivariate space. In addition, nMDS was used to investigate any relationships between
spider species community composition and the environmental variables, aerial cover, ground
cover and woody plant species richness. According to Clarke and Warwick (1994), the
advantages of MDS include giving a good link between the original data and the final picture
and representing complex patterns correctly in low-dimensional space. Principal Component
Analysis (PCA) has the disadvantage of inflexibility of dissimilarity measure and poor
distance-preservation (Clarke and Warwick 1994).
Non Metric Multidimensional Scaling (nMDS) ordinations were constructed in PC-ORD 5.1
(McCune & Mefford 2006). This was achieved with a Bray-Curtis distance measure of 250
runs and 500 iterations and random starting configurations with a maximum of six axes.
30
Dimensionality was assessed by choosing the solution with the minimum number of axes and
lowest stress. The Monte Carlo test result stress was computed in order to compare the stress
obtained with the real data against the stress obtained for randomized data. A p-value of less
than 0.05 indicates significantly more reduction in stress than expected by chance (McCune
& Mefford 2006).
Ordination diagrams of nMDS showed two kinds of entities; the sample units (termite
mounds and the woodland matrix sites) and the environmental variables (ground cover, aerial
cover, and woody plant richness). Environmental variables were represented as lines
radiating from the centroid of the plot and the direction and length of the lines represented the
direction and the strength of the relationship with spider composition in the two different
sites.
Coefficients of determination (R2) were calculated for the correlations between ordination
distances and distances in the original 2-dimensional space in order to determine how well
the ordination represented the original data. The Pearson correlation test (r) was used to
evaluate the degree of linear association between the environmental variables and spider
community structure (species community composition).
Significance level was set at p = <0.05 for all analyses in this study.
31
CHAPTER 4: RESULTS
4.1
Spider species community composition
Sixty two plots, each with paired sites (mound and adjacent matrix) were sampled, with 20
plots in the Bird Sanctuary, 30 plots in the Chivero Game Park and 12 plots in the Chizarira
National Park. A total of 3139 spiders were caught in the three study locations, with 19
families in the Bird Sanctuary, 23 in the Chivero Game Park and 17 in the Chizarira National
Park, bringing up to resulting in a total of 28 families in all the three locations (Table 4.1).
The wolf spiders (Lycosidae) were clearly the most abundant family and were highly
abundant in all three locations (Table 4.1). Ground spiders (Gnaphosidae) and jumping
spiders (Salticidae) were the two other families that were considerably abundant in all three
locations.
Table 4.1 gives a summary of the relative abundances for each location, together with the
number of morpho-species identified for each location. Families that were found exclusively
in Chizarira National Park were Agelenidae, Clubonidae, and Deinopidae. The following
families were found exclusively in the Chivero Game Park; Nesticidae, Uloboridae,
Dictynidae and Erisidae. The only family exclusive to the Bird Sanctuary was Scytodidae.
32
Table 4.1: The total numbers of spiders and morpho-species, in each family, in each location (* = not spiders/ Araneae)
Family
Agelenidae
Amaurobiidae
Barychelidae
Caponiidae
Clubonidae
Cyrtauchenidae
Deinopidae
Dictynidae
Erisidae
Gnaphosidae
Hahnidae
Heteropodidae
Loxoscelidae
Lycosidae
Miturgidae
Nesticidae
Oonopidae
Oxyopidae
Palpimanidae
Pholcidae
Pisauridae
Salticidae
Scytodidae
Tetragnathidae
Thomisidae
Uloboridae
Zodariidae
*Solifugidae
TOTALS
CHIVERO BIRD SANCTUARY
Termite Mound
Woodland Matrix
Number Number Number Number
of
of
of
of
spiders
species
spiders
species
CHIVERO GAME PARK
Termite Mound
Woodland Matrix
Number Number Number Number
of
of
of
of
spiders
species spiders
species
114
1
25
9
2
1
144
1
2
1
1
12
1
52
10
1
2
1
2
1
1
1
168
12
2
2
136
1
1
1
10
1
2
4
1
1
3
10
1
111
1
11
1
1
1
3
1
1
1
3
17
1
2
8
1
3
5
1
1
23
61
1
5
3
1
1
5
3
409
49
11
3
482
4
3
59
1
1
100
1
1
1
260
8
1
9
12
4
1
35
104
1
3
7
1
1
3
17
1
1
1
110
4
683
1
1
1
6
2
74
1
14
1
1
16
62
2
1
1
1
4
4
CHIZARIRA NATIONAL PARK
Termite Mound
Woodland Matrix
Number Number Number Number
of
of
of
of
spiders
species spiders
species
Total
Relative
number family
of
%
spiders
14
9
2
2
3
2
2
2
2
1
2
1
1
1
1
1
6
2
45
3
10
1
26
7
6
3
84
1
14
1
17
218
82
7
2
6
6
1
2
677
12
4
6
1092
3
1
27
47
11
2
217
367
2
2
13
1
194
120
3139
1
194
1
4
1
2
246
1
1
9
2
255
1
20
12
18
1
2
4
1
1
5
2
1
3
1
1
6
1
5
156
141
5
20
27
6
17
5
3
2
1
1
1
2
1
2
68
14
941
10
3
70
81
454
1
54
18
170
1
45
33
0.5
6.9
2.6
0.2
0.1
0.2
0.2
0.0
0.1
21.6
0.4
0.1
0.2
34.8
0.1
0.0
0.9
1.5
0.4
0.1
6.9
11.7
0.1
0.1
0.4
0.0
6.2
3.8
100.0
4.2
Species Accumulation Curves
The species accumulation curves for all three locations (Figure 4.1) did not level off to a
plateau before the Sorensen distance value reached zero. Although a greater sampling effort
would have resulted in a greater number of new species, spider sampling was considered
acceptable but clearly not completely adequate. In all the curves (Figure 4.1) a gradual
deceleration in the acquisition of new species was observed after an initial rapid increase in
the number of species.
34
Figure 4.1: Species accumulation curves (rising curves) and distance curves (falling curves)
for termite mound and woodland matrix sites sampled in the three study locations. Both
curves were adjusted for random sample order. The distance curve represents the average
Sorenson distance between the whole sample and subsamples and broken lines represent
standard deviations from the mean.
35
4.3
Differences in Spider Species Abundance and Species
Richness, Between Termite Mounds and the Adjacent
Woodland Matrix
In the Bird Sanctuary, the average spider species richness and abundance were greater (Table
4.2) in the woodland matrix than on the termite mounds. The Welch two sample t-test
confirmed the differences in species richness to be statistically significant (p = <0.05) but the
apparent differences in abundance between mounds and matrices were not significant (p =
>0.05). In the Game Park (Table 4.3) both spider species richness and abundance were
significantly (p = <0.05) higher in the woodland matrix. In Chizarira National Park (Table
4.4) spider species richness and abundance were significantly (p = <0.05) greater on the
termite mounds than the adjacent woodland matrix.
Table 4.2 Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in the Chivero Bird Sanctuary. The
Welch two sample t-test was used to check for statistically significant differences.
Measure
Number of
Spiders
Welch two sample t-test
Standardd
Deviation t-value
d.f.
p-value
10.38
-1.21
36.59
0.232
8.54
Site
N Mean
Termite mound
20 20.45
Woodland Matrix 20 24.10
Termite mound
20 8.80
2.82
Number of
Species
Woodland Matrix 20 11.3
3.76
The symbol * represents a statistically significant difference
36
-2.38
35.08
0.023*
Table 4.3: Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in the Chivero Game Park. The
Welch two sample t-test was used to check for statistically significant differences.
Measure
Number of
Spiders
Site
Termite mound
Woodland Matrix
Welch two sample t-test
Standard
N Mean Deviation t-value d.f.
p-value
30 22.73 6.41
-2.42
35.80 0.021*
30 31.40 18.51
30 11.17 3.34
Number of Termite mound
-2.92
Species
Woodland Matrix 30 13.63 3.18
The symbol * represents a statistically significant difference
57.92 0.005*
Table 4.4: Mean (with standard errors and 95% confidence intervals) spider richness and
abundance on termite mounds and in the woodland matrix in Chizarira National Park. The
Welch two sample t-test was used to check for statistically significant differences.
Welch two sample t-test
Measure
Number of
Spiders
Site
Termite mound
Woodland Matrix
Standard
N Mean Deviation
12 37.75 16.52
12 14.25 7.76
12 15.17 4.12
Number of Termite mound
Species
Woodland Matrix 12 9.67
3.71
The symbol * represents a statistically significant difference
37
t-value
d.f. p-value
4.46
15.65
<0.001*
3.44
21.74
0.002*
4.4
Similarity Analysis
In the Bird Sanctuary, a cluster analysis (Figure 4.2) based on spider species community
composition showed that there were various groups of termite mounds and woodland matrix
sites clustered together.
Figure 4.2: Dendogram showing species shared between sites (n=40) in the Chivero Bird
Sanctuary. The unweighted pair-group average (UPGAMA) and Bray-Curtis similarity
measure were used to plot the dendogram. The symbol ● represents a woodland matrix (Ma)
site and
represents a termite mound (TM) site. Numbers following the initials TM and Ma
are the plot numbers.
38
It can therefore be said that in the Bird Sanctuary spider community assemblages differ
between the mounds and the woodland matrix, but there also seems to be other factors
determining spider species community composition other than just the site (termite mound or
woodland matrix).
The Chivero Game Park produced a somewhat similar dendogram (Fig 4.3) to the Bird
Sanctuary but with fewer and larger groups of mound and matrix sites. Grouping was much
more pronounced in Chizarira National Park (Fig 4.4) as all the termite mounds except
termite mound number 2 were clustered together. A closer look at the habitat characteristics
shows that mound number 2 was the only mound with no aerial cover (0% cover) and also
had a very low ground cover (Appendix D) cover. There therefore seems to be an increase in
spider species community composition similarity between the termite mounds and the
woodland matrix, from the least impacted to the most impacted miombo woodland, from
Chivero Bird Sanctuary to Chizarira National Park.
It was also observed that in the Chizarira National Park dendogram (Figure 4.4) distances
between individual termite mounds were much less as compared to those observed in the
matrices. This observation was also evident in the Game Park dendogram, but less
pronounced, and not identifiable in the Bird Sanctuary spider data. It seems likely that
grouping is much closer between termite mounds than between woodland matrix sites, and
therefore spider assemblages on termite mounds are more homogeneous and termite mounds
harbour somewhat distinct spider assemblages.
39
Figure 4.3: Dendogram showing species shared between sites (n = 60) in Chivero Game Park. The
unweighted pair-group average (UPGAMA) and Bray-Curtis similarity measure were used to plot the
dendogram. The symbol ● represents a woodland matrix (Ma) site and
(TM) site. Numbers following the initials TM and Ma are the plot numbers.
40
represents a termite mound
Figure 4.4: Dendogram showing species shared between sites (n = 24) in Chizarira National Park. The
unweighted pair-group average (UPGAMA) and Bray-Curtis similarity measure were used to plot the
dendogram. The symbol ● represents a woodland matrix (Ma) site and
represents a termite mound
(TM) site. Numbers following the initials TM and Ma are the plot numbers.
Multi-Response Permutation Procedures (MRPP) detected differences in spider community
assemblages between the termite mounds and the woodland matrix in all the three areas; the
Bird Sanctuary (A = 0.016 p = 0.005), the Chivero Game Park (A = 0.028 p = <0.001), and
the Chizarira National Park (A = 0.047 p = <0.001).
Although a gradient in similarity was established from the least impacted to the most
impacted woodland using hierarchical cluster analysis, simple calculation of the Jaccard
abundance corrected index and consequent statistical significance testing did not show any
differences in similarity between mound and matrix sites across the different herbivore
41
impact areas. The average Jaccard corrected indices for all the locations were relatively
similar (Figure 4.5). Chizarira had an average value (corrected Jaccard similarity index) of
0.45 +/- 0.06 at a 95% confidence interval, the Bird Sanctuary had a value of 0.45 +/- 0.07 at
a 95% confidence interval, and Chivero Game Park had a value of 0.48+/- 0.05 at a 95%
confidence interval. A Kruskal-Wallis test for independent samples revealed no statistically
significant (p = >0.05) difference in similarity across the three locations.
Figure 4.5: Box plot showing the Jaccard corrected similarity indices of the three study
locations for similarities in the number of spiders and species shared by termite mounds and
matrices. A Kruskal-Wallis for independent samples revealed no statistically significant (X2 =
0.143, d.f.= 2, p = 0.931) difference in similarity across the three study areas.
Nevertheless, the box plot (Figure 4.5) shows that the data sets from the Bird sanctuary and
the Game Park had much greater variability as compared to the data for Chizarira although
42
the means were relatively similar. The data for Chizarira was skewed towards the upper
values whilst the data from the Bird sanctuary and the Game Park seemed to be normally
distributed. The medians of the three locations were relatively similar but Chizarira National
Park had the greatest median followed by the game park, and the bird sanctuary, respectively.
These observations show that in the Game Park and the Bird Sanctuary some mounds shared
the same number of species and individuals as the adjacent matrices (Jaccard index = 1),
whilst some shared none at all (Jaccard index = 0). In Chizarira National Park no mound had
the same number of species and spiders as the adjacent matrix and a somewhat more
intermediate similarity was portrayed.
4.5
Influence of Ground Cover, Aerial Cover, Woody Plant
Species Richness and Site in Determining Spider Species
Richness and Abundance
In all the study areas the mean ground cover (Table 4.5 - 4.7) was highest in the woodland
matrix sites but the Welch two sample t-test indicated that all these differences were not
statistically different (p = >0.05). In the Game Park (Table 4.6) and the Bird Sanctuary (Table
4.5), aerial cover was also greatest in the matrices but only the difference in the Chivero
Game Park was statistically significant (p = <0.05). Termite mounds in Chizarira had a
higher aerial cover average value as compared to the matrices, and this difference was
statistically significant (p = <0.05). Woody plant species richness was greater on the mounds
in all three areas and only that in the Bird sanctuary and the Game Park was statistically
significant (p = <0.05).
43
Table 4.5: Summary statistics of ground cover, aerial cover, and woody plant species richness
measured in the Bird sanctuary. The significance of differences between sites was determined
using the Welch two sample t-test.
Measure
Ground cover
Aerial cover
Site
Termite Mound
Woodland Matrix
Welch two sample t-test
Standard
N Mean Deviation t-value d.f.
p-value
20 73.10 11.94
-1.61
37.91 0.117
20 79.33 12.57
Termite Mound
Woodland Matrix
20 32.98
20 35.10
9.53
10.87
Plant species Termite Mound
20 29.35 4.11
richness
Woodland Matrix 20 15.75 6.13
The symbol * represents statistical significance, p = <0.05
-0.66
37.32 0.515
8.24
33.35 <0.001*
Table 4.6: Summary statistics of ground cover, aerial cover, and woody plant species richness
measured in the Chivero Game Park. The significance of differences between sites was
determined using the Welch two sample t-test.
Welch two sample t-test
Measure
Ground cover
Aerial cover
Site
Termite Mound
Woodland Matrix
Standard
N Mean Deviation t-value
30 60.05 13.36
30 65.27 18.84
-1.24
Termite Mound
Woodland Matrix
30 24.38
30 33.82
9.91
22.62
Plant species Termite Mound
30 17.93 5.48
richness
Woodland Matrix 30 10.07 3.67
The symbol * represents statistical significance, p = <0.05
44
d.f.
p-value
52.29 0.222
-2.09
39.78 0.043*
6.57
50.63 <0.001*
Table 4.7: Summary statistics of ground cover, aerial cover, and woody plant species richness
measured in the Chizarira National Park. The significance of differences between sites was
determined using the Welch two sample t-test.
Site
Termite Mound
Ground cover
Woodland Matrix
N Mean
12 22.81
12 34.24
Welch two sample t-test
Standard
Deviation t-value d.f.
p-value
15.66
-1.37
18.76 0.188
24.39
Termite Mound
Woodland Matrix
12 53.76
12 4.66
28.54
4.54
Measure
Aerial cover
Plant species Termite Mound
12 12.50 3.43
richness
Woodland Matrix 12 10.92 3.33
The symbol * represents statistical significance, p = <0.05
5.88
11.56 <0.001*
1.15
21.98 0.262
After computing various stepwise model simplifications, a Generalized linear model (GLM)
was produced for each of the two response variables (spider abundance and richness), at each
location. Tables 4.8 to 4.13 are the resultant minimal adequate models showing the influence
of explanatory variables and their interactions, on spider species richness and abundance.
In the Bird Sanctuary none of the explanatory variables in the model (X2 = 24.173, df = 8, p =
0.002) were able to explain variations in spider richness (Table 4.9). The factors influencing
spider abundance (Table 4.8) (X2 = 6.337, df = 8, p = 0.061) were aerial cover and the two
way interactions between aerial cover and ground cover, and between site (termite mound
and woodland matrix) and woody plant richness.
45
Table 4.8: Results of a Generalized Linear Model (Type III) run on the Bird Sanctuary data,
with spider abundance as the response variable and site, aerial cover, and their interactions as
the factors. The model was constructed through a stepwise model simplification and with a
Poisson error distribution and a log link function.
Factor
(Intercept)
Site
Aerial cover
Site ● Plant richness
Aerial cover ● Ground cover
Site ● Ground cover
Site ● Aerial cover
Aerial cover● Plant richness
Site ● Aerial cover ● Ground cover
Site ● Aerial cover● Ground ● Plant richness
● Represents an interaction between variables. *
<0.05)
Wald Chi-Square
d.f.
0.001
1
0.232
1
3.974
1
6.379
2
5.078
1
3.395
2
2.39
1
2.862
1
3.502
1
3.685
2
Represents a significant difference
p-value
0.977
0.63
0.046*
0.041*
0.024*
0.183
0.122
0.091
0.061
0.158
(p =
Table 4.9: Results of a Generalized Linear Model (Type III) run on the Bird Sanctuary data,
with spider species richness as the response variable and site, aerial cover, and their
interactions as the factors. The model was constructed through a stepwise model
simplification and with a Poisson error distribution and a log link function.
Source
(Intercept)
Site
Site ● Plant richness
Site ● Aerial cover
Site ● Ground cover
Site ● Ground cover ● Aerial cover
Site ● Plant richness ● Aerial cover
Site● Ground cover ● Plant richness
Site ● Ground cover ● Plant richness ● Aerial cover
● Represents an interaction between variables.
46
Wald Chi-Square
1.196
0.136
0.202
0.517
0.666
0.654
0.204
0.635
0.571
d.f.
1
1
2
2
2
2
2
2
2
p-value
0.274
0.712
0.904
0.772
0.717
0.721
0.903
0.728
0.752
Table 4.10: Results of a Generalized Linear Model (Type III) run on the Chivero Game Park
data, with spider abundance as the response variable and site, aerial cover, ground cover and
some interactions including plant richness as the factors. The model was constructed through
a stepwise model simplification and with a Poisson error distribution and a log link function.
Source
Wald Chi-Square
d.f.
p-value
(Intercept)
0.04
1
0.842
Site
2.211
1
0.137
Ground ● Plant richness
5.486
1
0.019*
Site ● Ground
13.867
2
0.001*
Site ● Aerial
12.551
2
0.002*
Site ● Plant richness
5.104
2
0.078
Site ● Ground ● Aerial
14.435
2
0.001*
Site ● Ground ● Plant richness
5.898
1
0.015*
Site ● Plant richness ● Aerial
6.541
2
0.038*
Ground ● Plant richness ● Aerial
4.658
1
0.031*
Site ● Ground ● Plant richness ● Aerial
7.243
1
0.007*
● Represents an interaction between variables. * Represents a significant difference (p =
<0.05)
In the Game Park various two way interactions, 2 three way interactions and a four way
interaction between the explanatory variables (Table 4.10) (X2 = 43.351, df = 10, p = <0.001)
were found to significantly (p = <0.05) influence spider abundance, with the majority of the
interactions having the variable site. Nevertheless, the factor site on its own was deemed to
have an insignificant (p = >0.05) influence on spider abundance. It can therefore be said that
in the Chivero Game Park spider abundance is determined by a combination of
environmental variables, with the presence or absence of termite mounds playing an
important role in these interactions. Spider richness (Table 4.11) (X2 = 23.275, df = 10, p =
0.003) on the other hand was proved to be driven by the three way interaction between site,
ground cover, and aerial cover, and the four way interaction between site, woody plant
richness, ground cover, and aerial cover. The variable site was also a part of both the two way
47
and three way interaction factors. As with spider abundance the site was considered an
essential variable as it was in all the interactions.
Table 4.11: Results of a Generalized Linear Model (Type III) run on the Chivero Game Park
data, with spider species richness as the response variable and site, aerial cover, and some
interactions including plant richness and ground cover as the factors. The model was
constructed through a stepwise model simplification and with a Poisson error distribution and
a log link function.
Source
Wald Chi-Square d.f.
p-value
(Intercept)
1.055
1
0.304
Site ● Ground cover
4.943
2
0.084
Site ● Aerial cover
5.69
2
0.058
Site ● Plant richness
1.99
2
0.37
Site ● Plant richness ● Aerial cover
4.883
2
0.087
Site ● Plant richness ● Ground cover
4.941
2
0.085
Site ● Ground cover ● Aerial cover
8.397
2
0.015*
Site ● Plant richness ● Ground cover ● Aerial cover
7.738
2
0.021*
● Represents an interaction between variables. * Represents a significant difference (p<0.05)
The generalized linear model for Chizarira National Park (X2 =133.29; df = 15, p=0.000) also
showed that various factors influence spider abundance. The variable site was also a part of
all the interactions except the two way interaction between ground cover and aerial cover.
Site and aerial cover were proved to be a significant factor of spider abundance.
Spider richness (Table 4.13) (X2 = 30.84, d.f. = 8, p = <0.001) in Chizarira was also proved
to be significantly (p = <0.05) influenced by ground cover, and the two way interactions
between site and woody plant richness, between site and aerial cover, and between site and
ground cover. Spider richness was also significantly (p = <0.05) influenced by the three way
48
interactions between site, woody plant richness, and aerial cover; and between site, woody
plant richness, and ground cover.
Table 4.12: : Results of a Generalized Linear Model (Type III) run on the Chizarira National
Park data, with spider abundance as the response variable and site, aerial cover, ground cover,
plant richness and some of their interactions as the factors. The model was constructed
through a stepwise model simplification and with a linear distribution and a log link function.
Source
Wald Chi-Square d.f.
p-value
(Intercept)
27.653
1
<0.001*
Site
18.719
1
<0.001*
Aerial
9.665
1
0.002*
Site ● Ground cover
11.331
2
0.003*
Site ● Aerial cover
9.054
1
0.003*
Ground cover ● Aerial cover
4.093
1
0.043*
Site ● Plant richness
2.553
2
0.279
Site ● Ground cover ● Plant richness
5.021
2
0.081
Site ● Aerial cover ● Plant richness
4.003
2
0.135
Site ● Ground cover ● Aerial cover
7.177
1
0.007*
Site ● Ground cover ● Aerial cover ● plant richness 0.435
2
0.805
● Represents an interaction between variables. * Represents a significant difference (p<0.05)
49
Table 4.13: Results of a Generalized Linear Model (Type III) run on the Chizarira National
Park data, with spider species richness as the response variable and aerial cover, ground cover
and some interactions including site and plant richness as the factors. The model was
constructed through a stepwise model simplification and with a Poisson error distribution and
a log link function.
Source
Wald Chi-Square
d.f.
p-value
(Intercept)
5.04
1
0.025*
Site
0.804
1
0.370
Ground cover
4.375
1
0.036*
Site ● Plant richness
17.951
2
<0.001*
Site ● Aerial cover
7.402
2
0.025*
Site ● Ground cover
13.026
1
<0.001*
Plant richness ● Ground cover
3.736
1
0.053
Aerial cover ● Ground cover
0.1
1
0.752
Site ● Plant richness ● Aerial cover
9.788
2
0.007*
Site ● Aerial cover● Ground cover
0.095
1
0.757
Site ● Plant richness ● Ground cover
8.716
1
0.003*
Site ● Plant richness ● Aerial cover ● Ground cover 0.123
2
0.940
● Represents an interaction between variables. * Represents a significant difference (p<0.05)
In all the three locations a distinct spider species community composition was evident
between matrix and mound sites. Nevertheless, a non-metric Multidimensional Scaling plot
for the Bird Sanctuary (Figure 4.6) showed a considerable similarity in the spider
composition between mounds and matrices as compared to the plot for Chizarira National
Park (Figure 4.8). A degree of overlap in species composition between mounds and matrices
was also observed in the Game Park, but at a lesser extent than in the Bird Sanctuary.
In the Bird Sanctuary plot (Figure 4.7), ground cover and woody plant richness appeared in
opposite directions showing that in the woodland matrix spider composition is driven mainly
by ground cover in a rather weak association and on the termite mounds spider composition
is driven mainly by woody plant richness in a strong relationship. In general, the Pearson
50
correlation test (r) shows that in axis 1 ground cover (r = -0.260) had the greatest association
with spider community structure followed by aerial cover (r = -0.127) and lastly plant
richness (r = -0.024). In axis 2 all the environmental variables, ground cover (r = 0.199),
aerial cover (r = 0.159), and plant richness (r = -0.125) had more or less the same degree of
linear association with spider community structure.
Woody plant richness was also the major factor determining spider composition on the
termite mounds in the Game Park (Figure 4.6) and woodland matrix species composition was
mainly driven by aerial cover but a weak relationship was evident from the short line
representing aerial cover on the nMDS plot in Figure 4.6. Both termite mound and woodland
matrix spider species community compositions were aslo strongly associated with ground
cover. In general, the Pearson correlation test (r) shows that in axis 1 ground cover (r = 0.552), had the greatest linear association with spider species community composition
followed by plant richness (r = 0,166), and aerial cover (r = 0.063). In axis 2 plant richness (r
= 0.420), had the greatest linear association followed by aerial cover (r = -0.144), and ground
cover (r = -0.054).
51
Figure 4.6: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chivero Bird Sanctuary. The symbol ●
represents woodland matrix sites and
represents termite mounds. Ellipses were drawn
around similar communities. The group A is primarily composed of woodland matrix sites
and the group B is composed of termite mound sites. Monte Carlo test result for mean stress
is 48.185 (p = 0.0040) for axis 1 (R2 = 0.397) and 29.716 (p = 0.0040) for axis 2 (R2 = 0.237).
52
Figure 4.7: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chivero Game Park. The symbol ● represents
woodland matrix sites and
represents termite mounds. Ellipses were drawn around similar
communities. The group A is primarily composed of woodland matrix sites and the group B
is composed of termite mound sites. Monte Carlo test result for mean stress is 48.662 (p =
0.0080) for axis 1 (R2 = 0.228) and 31.589 (p = 0.0040) for axis 2 (R2 = 0.425).
53
In Chizarirra National Park the spider composition between mounds and matrices was well
separated, with no overlap (Figure 4.8). It was therefore deduced that both aerial cover and
plant richness influence species composition on the termite mounds, with aerial cover having
a very strong relationship and plant richness having a very weak one. The woodland matrix
spider composition has a very weak relationship with ground cover and wooddy plant
richness. Pearson correlation test (r) shows that in axis 1 aerial cover (r = -0.719) had the
greatest linear association with spider species community composition, followed by plant
richness (r = -0.221) and ground cover (r = 0.115). In axis 2, plant richness (r = -0.419) had
the greatest linear association followed by aerial cover (r = 0.292) and ground cover (r = 0.142). All the environmental variables therefore had an influence on spider community
structure, with aerial cover being the most influential variable in determining spider
community structure.
The nMDS plot of Chizarirra also shows tight clustering of the spider community on the
termite mounds, which shows that spiders on termite mounds in Chizarira have a strong
preference for this particular habitat.
In all the nMDS plots the solutions were stronger than expected by chance (p = <0.05) when
stress was related to dimensionality (Monte Carlo test for mean stress). All the plots are
therefore deemed reliable representations of the actual spider communities in the three
locations.
54
Figure 4.8: Non-metric Multidimensional Scaling plot showing the relationships between
spider community species composition and the explanatory variables; aerial cover, ground
cover and woody plant species richness, in the Chizarira National Park. The symbol ●
represents woodland matrix sites and
represents termite mounds. Ellipses were drawn
around similar communities. The group A is primarily composed of woodland matrix sites
and the group B is composed of termite mound sites. Monte Carlo test result for mean stress
is 49.213 (p = 0.0040) for axis 1 (R2 = 0.481) and 26.446 (p = 0.0040) for axis 2 (R2 = 0.26).
55
CHAPTER 5: DISCUSSION
5.1
Sampling Adequacy
According to Soberon and Llorente (1993) at any particular time there is only a finite
number of species in a given area and for ecological sampling to be considered adequate and
representative, a species accumulation curve should be drawn and it should reach an
asymptote or plateau where any additional sampling effort should not result in any more new
species. However, none of the species-accumulation curves (from all the study locations;
Figure 4.1a-f) in this study reached an asymptote. Ugland et al. (2003) suggest that this
asymptote may be reached for data sets of species that can be identified easily, such as of
plants and breeding birds where it is possible to obtain a count of all the species present
(Colwell & Coddington 1994). For other habitats (or taxa) one cannot expect to count all the
species. This asymptote has not been attained in several studies (eg. Erwin 1988, 1991;
Ugland et al 2003) and Thompson et al (2003) state that this asymptote need not be achieved
always for sampling to be adequate but the curve should elbow and begin to rise at a reduced
rate for sampling to be considered reasonable. Nevertheless if species richness estimates are
required from these curves then much greater sampling effort would have to be employed.
Spiders in particular, are a very diverse and highly abundant group of invertebrates and
sampling would have to take place over a much longer period of time. Many spider inventory
studies (e.g. Whitmore et al. 2002; Russell-Smith 1999; Van den Berg and DippenaarSchoeman 1991) are evidence of this.
56
5.2
Patterns of Spider Species Richness, Abundance, and
Community Composition
The results from the Bird Sanctuary, an area with minimal herbivore impacts, suggest that
termite mounds alone have no influence on the number of ground dwelling spiders. However,
in a heavily impacted woodland such as Chizarira National Park, a higher spider richness and
abundance was realized on the termite mounds. Based on these observations I therefore make
a claim that large termitaria do not serve as hotspots of diversity for ground dwelling spiders,
but rather as refugia in highly disturbed and elephant impacted woodlands.
This conclusion is supported by a study done by Joseph et al. (2011) in Chizarira National
Park, on cavity using birds. They concluded that termitaria in a dystrophic savanna system
can contribute to ecosystem resilience by providing refugia for key functional elements such
as woody plant species, as woodland matrix quality declines. As a result, a refuge for cavityusing birds is also facilitated. The current study was evidence of this for spiders and results
(Table 4.7) show that termite mounds provide greater aerial cover than the adjacent
woodland. By providing refugia for animal and plant life, large termitaria could therefore
play an important role in sustaining biodiversity in highly disturbed area such as Chizarira
National Park.
The results for the Chivero Game Park showed a greater number of spider species and
individuals in the woodland matrices than on the mounds. The greater number of species
could be due to the fact that there is only intermediate herbivore impact and disturbance in
the woodland (Makumbe 2009) and the intermediate disturbance hypothesis (Connell 1978)
could explain this. The hypothesis states that species richness is maximized at intermediate
levels of disturbance. A fundamental assumption of the hypothesis is that a trade-off exists
between the ability of a species to tolerate disturbance and its ability to compete. According
57
to this hypothesis, if disturbance occurs frequently, richness decreases because species
intolerance to disturbance becomes locally extinct. If disturbances are too infrequent, richness
decreases because dominant species occupy resources and eliminate weak competitors. The
intermediate disturbance in the game park woodland could therefore be leading to greater
spider richness in the woodland than on the termite mounds. The hypothesis of disturbance
driving diversity (Tilman 1994; Petraitis et al. 1989) therefore seems to holds in this study.
However, in the Bird Sanctuary a higher spider richness was also found in the woodland and
yet herbivore disturbance in the woodland was deemed minimal. This result seems more like
a discrepancy but it should be admitted that this study did not examine vegetation structure. A
study by Makumbe (2009) has already shown that woody plant species structure differs on
and off the termite mounds. So since invertebrate diversity is a function of habitat complexity
(Robinson 1981; Gunnarsson 1988; Balfour and Rypstra 1998; Raizer and Amaral 2001),
plant structure could have influenced diversity in the woodland in the Chivero Game Park.
Another school of thought stems from the debatable use of statistical methods to determine
significance of differences rather than assessing ecological or biological significance of
differences (Johnson 1999) but this is a discussion for another paper.
Differences in species composition between the termite mounds and the woodland matrices
were evident in all the three study areas (Figures 4.6 – 4.8). The study showed that the higher
the herbivore impact is, the greater the difference in spider communities on and off termite
mounds, as termitaria is impacted differently from the adjacent woodland matrices.
Unfortunately, a full analysis of indicator species was not carried out due to lack of expertise
in taxonomy. It was also observed that in Chizarira National Park there was a completely
different composition of spider species on the mound as compared to the matrix. Less
58
dissimilarity was observed in the woodland with intermediate herbivore impacts, the Chivero
Game Park and the least dissimilarity was observed in the minimal herbivore impact area, the
Bird Sanctuary. This study has thus provided evidence that in miombo woodlands spider
species composition can be tied to herbivore impacts such that the more disturbed the habitat
is, the more the dissimilarity in spider species composition between termite mounds and the
woodland matrices is observed.
The results of no significant (p = <0.05) difference in similarity across the herbivore impact
gradient could be interpreted to mean that either there was no pattern of similarity exhibited
by ground spiders across the three areas or that there was a highly complex pattern that was
not revealed by the corrected Jaccard similarity indices. This result emphasises the
importance of ordination methods such as nMDS in place of basic statistics, because of the
complexities of communities. The results could also be taken to mean that analysis at
community level gives a clearer picture of the state of ecosystems as opposed to simple
species (richness) and individuals (abundance) counts. In this study nMDS, an ordination
method, gave more information about the differences in spider communities in each study
area.
5.3
Influence of Habitat Characteristics
Results on habitat characteristics showed that in the Bird sanctuary only plant richness
differed on and off the mounds, plant richness and aerial cover differed in the Game Park,
and only aerial cover differed in Chizarira. Spider abundance and richness did not follow this
pattern entirely as in the game park a greater aerial cover in the matrix was followed by
greater spider abundance and richness in the woodland matrix as well. In Chizarira national
park, a greater woody plant species richness was also followed a greater abundance and
richness of spiders on termite mounds.
59
Generalized linear models (Tables 4.8 – 4.13) gave evidence that no one environmental
variable alone was responsible for the variation in spider abundance and richness. Instead,
various interactions, of different orders, of ground cover, aerial cover, plant richness, and site
were the drivers of spider richness and abundance. As a result simply trying to establish a
linear relationship between spider abundance or richness and any one of the proposed
explanatory variables would not work entirely.
Along the gradient of herbivore impacts, from the Bird Sanctuary to Chizarira National Park,
the number of interactions between the explanatory variables seemed to increase. In the Bird
Sanctuary no variables or interactions were proved to influence spider species richness. This
observation could be a result of high structural complexity in the Bird Sanctuary woodland,
which has been removed from herbivore disturbance for the past 50 years (Cumming pers.
comm.).
Many ecological studies are complicated by the unavoidable collinearity of explanatory
variables and MacNally (2000) points out that this collinearity limits regression analysis
adequacy in finding appropriate causal variables. Collinearity (or multicollinearity) is the
undesirable situation when one independent variable is a linear function of other independent
variables (Jöreskog and Sörbom 1989) and this is undesirable in models such as regression
and GLM that are based on the assumption (Jöreskog and Sörbom 1989) that explanatory
variables are not linearly related. As such, a possible explanation as to why none of the
environmental features (ground cover, aerial cover, plant species richness) in this study were
strong factors might be that other factors came into play and influenced spider diversity as
well in a somewhat linear fashion. Nevertheless, the full factorial model of GLM used in this
60
study explored all possible interactions and their influences on spider richness and
abundance.
According to Crawley (2007) straightforward linear regression methods (assuming constant
variance, normal errors) are not appropriate for count data for the following reasons: the
linear model might lead to the prediction of negative counts, the variance of the response
variable is likely to increase with the mean, the errors will not be normally distributed, and
zeros are difficult to handle in transformations. As spider abundance and richness (both count
data) were the response variables in this study it was appropriate to use generalized linear
models that are able to deal with all these difficulties.
Another possible explanation for this deviation can be pinned down on the habitat
specialization and the high diversity of spiders. According to Buchholz (2010), ecological
traits of spider in shading may be preferential either for habitat openness or for vegetation
cover. Thus, it is possible that in the current study spider species with such contrasting habitat
preferences were present. A typical example is provided by a study by Warui et al. (2005) on
the impacts of wildlife and cattle on the diversity of spiders. In their study they identified one
species, Aelurillus sp. and concluded that it was probable that the species preferred open
habitats, which are less complex because of its mode of feeding which involves hunting, and
this could become hindered by a complex habitat. No particular spider community can
therefore have all species favouring a particular habitat. This discussion therefore further
emphasises the importance of proper taxonomy in similar studies.
It should also be acknowledged that other studies (e.g. Mallis and Hurd 2005) have failed to
find convincing correlations between the environment and the occurrence of spiders. These
61
authors claimed that spider communities are stochastic assemblages and habitat conditions as
well as niche properties have little influence on their structure and dynamics. Nevertheless,
this is essentially a neutral view on spider communities.
Spider community analysis revealed that spider species composition on termite mounds is
driven by a different set of factors that determine composition in the matrices, and this
becomes more apparent with the increase in the level of herbivore impacts. nMDS results
showed that woody plant richness strongly influenced spider species community composition
in both the Game Park and the Bird Sanctuary. Previous studies have proved that termite
mounds harbour a unique suite of plants in comparison with the surrounding woodland. The
Welch two sample test in Tables 4.5 and 4.6 also showed that woody plant richness was
significantly higher on the termite mound. A higher plant richness of unique plants therefore
translates into greater habitat complexity in relation to the woodland matrix. This is well in
accordance with the hypothesis that more complex habitats provide arthropods with sites for
shelter, foraging, oviposition, and mating (Lawton 1983, Halaj et al. 2000). Results of other
authors who have investigated spider communities in relation to vegetation structure
(Maelfait and De Keer 1990; Gibson et al. 1992; Mc Ferran et al. 1994) also show this
association between spider communities and the local vegetation structure.
In addition, the niche theory states that differences between species and environmental
factors drive the distribution of species and ultimately the composition and diversity of
communities (Hutchinson 1958), as so was observed in this study.
It can therefore be said that woody plant species richness is an important driver of spider
community structure in miombo woodlands with large termitaria, but as the level of herbivore
62
impacts increase plant richness becomes less important in determining spider community
structure, as was shown in Chizarira National Park, a heavily impacted woodland. Instead
aerial cover becomes the most important driver of spider community structure on the mounds.
In a woodland matrix with minimal herbivore impacts ground cover seems to be an important
driver of spider community structure. In the presence of intermediate impacts both aerial
cover and ground cover are the drivers, but in a highly impacted woodland ground cover and
plant richness only have weak associations with spider community composition.
Generally, results clearly illustrated the advantages of using ordination methods such as
nMDS in place of basic statistics of abundances and species richness‟s, because of the
apparent complexities of spider (or invertebrate) communities. Non metric multidimensional
scaling gave more information about spider communities. The results indicate that the two
types of sites, termitaria and matrix, have unique species compositions. Additionally, there
are many environmental factors that determine the composition at a site and not simply the
habitat type.
The hypothesis that disturbance drives diversity (Tilman 1994 Petratis et al. 1989) therefore
seems to hold for ground dwelling spiders, although not in the expected linear fashion.
5.4
Limitations of the Study
One of the problems when looking at spider diversity at a coarse level of resolution e.g. at the
guild level, is the fact that it is not possible to detect the sensitivity of individual species to
disturbances (Buchhloz 2010). Lawton et al. (1998) argued that different species vary in their
requirements within a natural ecosystem. This was further supported by Goldstein (1999) and
63
Alonso (2000) who emphasized that individual species always had their unique history that
dictated their distribution. Such arguments are against analysis at a coarse level of resolution
to detect disturbances and would instead tend to support the species-level approach. The use
of morphospecies in this study clearly overlooks this species-level approach but rather
assumes that all the species grouped together will respond in a similar way to changes. The
failure to properly classify the spiders to species level therefore compromised the accuracy of
this study.
In addition, Goldstein (1999) and Alonso (2000) emphasized the need for conservation and
management plans that not only incorporate the number of species but also the identity and
biology of species present. Nevertheless, such knowledge is still lacking among many
African savanna species and the biology and or ecology of the species already identified are
still not well documented, which makes understanding of many individual species difficult.
Sampling days: Many ecological studies involving spiders are conducted over long periods of
time, due to the high diversity of these invertebrates. As time was a limitation in the current
study sampling period was relatively short but highly intensive and this has the advantage
that a more robust comparative analysis (Sørensen 2004) can be attained. A short period was
also ideal for this study as it reduces the effects of immigration or emigration (Sørensen
2004) of spiders from one patch or site to another.
The use of the morphospecies approach has been used by numerous workers (e.g. Klein 1989;
Kremen 1992; Kremen et al. 1993) and it has been suggested (Beattie and Oliver 1995) that
non-specialists may use this method to classify invertebrates to morphospecies without
compromising scientific accuracy. As a result, environmental and conservation surveys can
64
be conducted in the absence of taxonomists, that are usually difficult to come by and require
much time for the taxonomic identification of specimens. The morphospecies approach can
however be complicated as proper species separation is often possible only with a detailed
study that may include dissection of genitalia. In these cases, use of morphospecies can result
in underestimation of species richness due to lumping (Derraik 2002). Non-specialists are
likely to assume in such situations that the small variation relates to the same species.
Another problem comes from overestimation of species by splitting when there is much
intraspecific variation, such as sexual dimorphism or large morphological differences
between adult and juvenile instars (Derraik 2002). There is therefore clearly no substitute to
taxonomy, but this study has shown that in determining the importance of termitaria in
miombo woodlands of different herbivore impacts morpho-species worked well, although
proper taxonomy would yield a more accurate result.
Besides herbivore impacts, other disturbances such as fire and human activities may impact
the woodlands under question as much. It is rather unfortunate that this study was not able to
factor them in due to limitations of time and resources. However, the three woodlands
(locations) exhibited quite distinct herbivore impact levels, and this alone was sufficient in
order to establish the relationship between the level of herbivore impacts and spider richness
and abundance, on and off large termitaria.
The present study is a step toward the use of spiders as indicators in the management of
Savannah ecosystems, but a better understanding of communities will only be obtained
through long-term studies.
65
CHAPTER 6: CONCLUSION
Based on spider abundance (the most reliable measure in the study), large vegetated
termitaria are not hotspots for ground dwelling spiders in miombo woodlands, but in high
herbivore impacted miombo woodlands they become refugia for a unique suite of spiders.
Spider community composition differs from the termite mounds to the adjacent miombo
woodland, and as the level of disturbance increases in the woodland, the two communities
(on the mound and in the woodland) become more and more distinct. The level of herbivore
impact seems to have an influence on vegetation cover and woody plant species richness that
as a result, together drive spider community composition, richness and abundance. These
results therefore show the indicator value of spider species richness, abundance, and
composition, and justify the use of spiders as bioindicators of habitat change in miombo
woodlands that have large termitaria, in future studies.
Nevertheless, more extensive sampling, with full identification of spiders, on a seasonal
basis, could uncover dynamic shifts in spider diversity and community structure that could
not be detected by this short term study.
66
REFERENCES
Anderson, G.D. and Walker, B.H. 1974. Vegetation composition and elephant damage in the
Sengwa Wildlife Research Area, Rhodesia. Journal of the Southern African Wildlife
Management Association 4: 1-14.
Alonso, L.E. 2000. Ants as indicators of diversity. In: Agosti, D., Majer, J. D., Alonso, L. E.
and Schultz, T. R. (eds.). Ants standard methods for measuring and monitoring
biodiversity. Washington, Smithsonian Institute Press. Pp. 80-88.
Balfour, R.A. and Rypstra, A.L. 1998. The influence of habitat structure on spider density in
a no-till soybean agroecosystem. Journal of Arachnology 26: 221-226.
Barnes, R.F.W. 1983. Is There a Future for Elephants in West Africa? Mammal Review
29(3): 175-200.
Beattie, A.J. and Oliver, I. 1995. Reply. Trends in Ecology and Evolution 10: 203–204.
Biondini, M.E., Bonham, C.D. and Redente, E.F. 1985. Secondary successional patterns in a
sagebrush (Artemisia tridentata) community as they relate to soil disturbance and soil
biological activity. Vegetatio 60: 25-36.
BirdLife International 2011. Important Bird Areas factsheet: Chizarira National Park.
Downloaded from http://www.birdlife.org on 11/02/2011
Blanc, J.J., Thouless, C.R., Hart, J.A., Dublin, H.T., Douglas-Hamilton, I. Craig, C.J. and
Barnes, A.F.W. 2003. African Elephant Status Report 2002. An Update from the
African Elephant Database. IUCN/SSC African Elephant Specialist Group. IUCN,
Gland, Switzerland and Cambridge, UK
67
Bonte, D., Leon, B. and Maelfait, J.P. 2002. Spider assemblage structure and stability in a
heterogeneous coastal dune system. Journal of Arachnology 30:331-343.
Bonte, D., Baert, L., Lens, L. and Maelfait, J.P. 2004. Effects of aerial dispersal, habitat
specialisation, and landscape structure on spider distribution across fragmented grey
dunes. Ecography 3: 343-349.
Bouyer, J., Sana, Y., Samandoulgou, Y., Cesar, J., Guerrini, L., Kabore-Zoungrana, C. and
Dulieu, D. 2007. Identification of ecological indicators for monitoring ecosystem
health in the trans-boundary W Regional Park: A pilot study. Biological Conservation
138: 73-88.
Brown, G.G., Barois, I. and Lavelle, P. 2000. Regulation of soil organic matter dynamics and
microbial activity in the drilosphere and the role of interactions with other edaphic
functional domains. European Journal of Soil Biology 36: 177–198.
Buchholz, S. 2010. Ground spider assemblages as indicators for habitat structure in inland
sand ecosystems. Biodiversity Conservation 19:2565–2595.
Buddle, C.M. and Rypstra, A.L. 2003. Factors initiating emigration of two wolf spider
species (Araneae: Lycosidae) in an Agroecosystem. Environmental Entomology
32:82-95.
Butler, V.P. and Haddad, C.R. 2011. Spider assemblages associated with leaf litter of three
tree species in central South Africa (Arachnida: Araneae). African Journal of
Ecology. University of Free State.
Campbell, B.M. and du Toit, R.F. 1988. Relationships between wood resources and the use
of species for construction and fuel in the communal lands of Zimbabwe.
Monographs in Systemic Botany of the Misssouri, Botanic Gardens 25: 331-341.
68
Campbell, B., Frost, P. and Byron, N. 1996. Miombo woodlands and their use: overview and
key issues. In: Campbell B. (ed.). The Miombo in Transition: Woodlands and Welfare
in Africa. Center for International Forestry Research (CIFOR). Bogor, Indonesia pp 110.
Carroll, R.W. 1988. Elephants of the Dzangha-Sangha dense forest of south-western Central
African Republic. Pachyderm 10: 12-15.
Cardoso, P., Scharff, N., Gaspar, C., Henriques S.S., Carvalho R., Castro P.H., Schmidt, J.B.,
Silva, I., Szüts T., De Castro, A. and Crespo, L.C. 2008. Rapid biodiversity
assessment of spiders (Araneae) using semi-quantitative sampling: a case study in a
Mediterranean forest. Insect Conservation and Diversity 1: 71-84.
Chao, A., Chazdon, R.L., Colwell, R.K. and Shen, T.J. 2005. A new statistical approach for
assessing compositional similarity based on incidence and abundance data. Ecology
Letters 8:148-159.
Cherrill, A.J. and Rushton, S.P. 1993. The Auchenorhyncha of unimproved moorland in
northern England. Ecological Entomology 18: 95- 103.
Chidumayo, E.N. 1987. Woodland structure, destruction and conservation in the copperbelt
area of Zambia. Biological Conservation 40: 89-100.
Clarke, B. and Warwick, R.M. 1994. Change in Marine Communities: an approach to
statistical analysis and interpretation. Plymouth, Plymouth Marine Laboratory.
Cohen, J.E. 2003. Human population: The Next Half Century. Science 302: 1172-1175.
69
Colwell, R.K. 2004. EstimateS: Statistical Estimation of Species Richness and shared species
from samples. Version 7.5 Persistent URL <purl.oclc.org/estimates>
Colwell, R.K. and Coddington, J.A. 1994. Estimating terrestrial biodiversity through
extrapolation. Philosophical Transactions of the Royal Society of London Series B,
Biological Sciences 345: 101-1 18.
Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310.
Crawley, M.J. 2007. The R Book. John Wiley & Sons Ltd, West Sussex, England.
Cumming, D.H.M. 1981. The management of elephant and other large mammals in
Zimbabwe. In Jewel, P.J., Holt, S. and Hart, D. (eds.). Problems in Management of
Locally Abundant Wild Animals. Academic Press, New York, NY. pp. 91–118
Cumming, D.H.M., Fenton, M.B., Rautenbach, I.D., Taylor, R.D., Cumming, G.S.,
Cumming, M.S., Dunlop, J.M., Ford, A.G., Havorka, M.D., Johnston, D.S.,
Kalcounis, M., Mahlangu, Z. and Portfors, C.V.R. 1997. Elephants, woodlands and
biodiversity in southern Africa. South African Journal of Science 93: 231-236.
Cumming, D.H.M. and Jones, B. 2005. Elephants in southern Africa: management issues and
options. WWF-SARPO, Harare, Zimbabwe.
Cumming, M.S. 1993. Spiders preying on Odontotermes transvaalensis (Sjostedt) termites in
the chimneys of their mounds. Abstracts of the 4th Colloquim of the RGSAA.
Newsletter of the Research Group for the study of African Arachnids 7:8-9.
Cumming, M.S. and Wesolowska, W. 2004. Habitat separation in a species-rich assemblage
of jumping spiders (Araneae: Salticidae) in a suburban study site in Zimbabwe.
Journal of Zoology 262: 1–10
70
Dangerfield, J.M. McCarthy, T.S. and Ellery, W.N. 1998. The mound building termite
Macrotermes michaelseni as an ecosystem engineer. Journal of Tropical Ecology 14:
507-520.
Dean, R.L., and Connell, J.H. 1987. Marine invertebrates in an algal succession. Journal of
Experimental Marine Biology and Ecology 109: 217-247.
Dennis, P., Young, M.R. and Gordon, I.J. 1998. Distribution and abundance of small insects
and arachnids in relation to structural heterogeneity of grazed, indigenous grasslands.
Ecological Entomology 22: 253-264.
Derraik, G.B., Closs, G.P., Dickinson, K.J.M., Sirvid, P., Barratt, B.I.P. and Patrick B.H.
2002. Arthropod Morphospecies versus Taxonomic Species: a Case Study with
Araneae, Coleoptera, and Lepidoptera. Conservation Biology 16: 1015–1023.
Dippenaar, A.S. and Meyer, M.K.P. 1980. On the spiders of the African genus Ammoxenus
(Araneae: Ammoxenidae), with descriptions of two new species. Journal of the
Entomological society of Southern Africa 43:41-49.
Dippenaar-Schoeman, A.S., de Jager, M. and van den Berg, A. 1996. Ammoxenus species
(Araneae: Ammoxenidae)- specialist predators of harvester termites in South Africa.
African Plant Protection 2:103–109.
Dippenaar-Schoeman, A.S. 2002. Baboon and Trapdoor Spiders of Southern Africa: An
Identification Manual. Plant Protection Research Institute Handbook No. 13.
Agricultural Research Council, Pretoria.
71
Downie, I.S., Wilson, W.L., Abernethy, V.J., McCracken, D.I., Foster, G.N., Ribera, I.,
Murphy, K.J. and Waterhouse, A. 1999. The impact of different agricultural land-use
on epigeal spider diversity in Scotland. Journal of Insect Conservation 3: 273-286.
Dunham, K.M., Mackie, C.S., Musemburi, O.C., Chipesi, D.M., Chiweshi, N.C., Taylor,
R.D., Chimuti, T., Zhuwau, C. and Brightman, M.A.H. 2006. Aerial survey of
elephants and other large herbivores in the Sebungwe Region, Zimbabwe, 2006.
WWF-SARPO Occassional Paper No. 12. December, 2006, 90 pp. WWF-SARPO,
Harare, Zimbabwe.
Entling, W., Schmidt, M.H., Bacher, S., Brandl, R. and Nentwig, W. 2007. Niche properties
of Central European spiders: shading, moisture and the evolution of the habitat niche.
Global Ecology and Biogeography 16:440–448.
Erwin, T.L. 1988. The tropical forest canopy. In Wilson, E.O. (ed.). 1993. Biodiversity.
National Academy Press, Washington D.C.
Erwin, T.L. 1991. How many species are there?: Revisited. Conservation Biology 5: 330-333.
Filmer, M. 1991. South African Spiders: An identification guide. Struik, Capetown.
Fleming, P.A. and Loveridge, J.P. 2003. Miombo woodland termite mounds: resource islands
for small vertebrates? Journal of Zoology 259: 161-168.
Foord, S.H., Mafadza, M.M., Dippenaar-Schoeman, A.S., Van Rensburg, B.J. 2008. Microscale heterogeneity of spiders (Arachnida: Araneae) in the Soutpansberg, South
Africa: a comparative survey and inventory in representative habitats. African
Zoology 43:156–174.
72
Frost, P.G.H. and Robertson, F. 1987. The Ecological Effects of Fire in Savannas, IN Walker
B.H. (Ed.) Determinants of Tropical Savannas. IRL Press, Oxford.
Frost, P. 1996. The Ecology of Miombo Woodlands. In: Campbell B. (ed.). The Miombo in
Transition: Woodlands and Welfare in Africa. Center for International Forestry
Research (CIFOR). Bogor, Indonesia pp 11-57.
Fenton, M.B., Cumming, D.H.M., Rautenbach, I.L., Cumming, G.S., Cumming, M.S., Ford,
G., Taylor, R., Dunlop, J., Hovorka, M.D., Johnston, D.S., Portfors, C.V., Kalcounis,
M.C. and Mahlangu, Z. 1998. Bats and the loss of tree canopy in African woodlands.
Conservation Biology 12: 399–407.
Gibson, C.W.D., Hambler, C. and Brown, V.K. 1992. Changes in spider (Araneae)
assemblages in relation with succession and grazing management. Journal of Applied
Ecology 29: 132-142.
Ginsberg, H.S. 1993. Invertebrate monitoring in the national park system. Narragansett, RI:
National Park Service Coastal Research Center, University of Rhode Island,
Narragansett Bay Campus; final report; tech. report NPS/NARURUNRTR-93/02.
Goldstein, P. Z. 1999. Functional ecosystems and biodiversity buzzwords. Conservation
Biology 13: 247-255.
Gotelli, N.J. and Colwell, R.K. 2001. Quantifying biodiversity: procedures and pitfalls in the
measurement and comparison of species richness. Ecology Letters 4: 379–391.
Greenstone, M.H. 1984. Determinants of web spider species diversity: vegetation structural
diversity vs. prey availability. Oecologia 62: 299-304. Berlin
73
Gratwicke, B. and Stapelkamp, B. 2006. Wildlife Conservation and Environmental
Management in an „Outpost of Tyranny‟. Zimconservation Opinion 3: 1-39.
Gunnarsson, B. 1988. Spruce-living spiders and forest decline; the importance of needle loss.
Biological Conservation 43: 309-319.
Guy, P.R. 1981. Changes in the biomass and productivity of woodlands in the Sengwa
Wildlife Research Area, Zimbabwe. Journal of Applied Ecology 18: 507-519.
Guy, P.R. 1989. The influence of elephants and fire on a Brachystegia-Julbernardia
woodland in Zimbabwe. Journal of Tropical Ecology 5: 215-226.
Haddad, C.R. and A.S., Dippenaar-Schoeman, 2002. The influence of mound structure on the
diversity of spiders (Araneae) inhabiting the abandoned mounds of the snouted
harvester termite, Trinervitermes trinervoides. The Journal of Arachnology 30: 403408.
Halaj, J., Ross, D.W. and Moldenke, A.R. 2000. Importance of habitat structure to the
arthropod food-web in Douglas-fir canopies. Oikos 90: 139-152.
Heermans, B.C. 2010. Large vegetated termitaria and fire impacts on reptilian community
assemblage in a miombo woodland system heavily impacted by elephants. Master of
Science in Conservation Biology, PercyFitzpatrick Institute of African Ornithology,
University of Cape Town.
Herremans M. 1995. Effects of woodland modification by African elephant, Loxodonta
africana, on bird diversity in northern Botswana. Ecography 18: 440-454.
Holdo, R.M. and McDowell, L.R. 2004. Termite mounds as nutrient-rich food patches for
elephants. Biotropica 36: 231-239.
74
Horvath, R., Magura, T., Szinetar, C. and Tothmeresz 2009. Spiders are not less diverse in
small and isolated grasslands, but less diverse in overgrazed grasslands: a field study
(East Hungary, Nyirseg). Agriculture, Ecosystems and Environment 130:16–22.
Hutchinson, G.E. 1958 Concluding remarks. Cold Spring Harbour Symposium on
Quantitative Biology 22: 415–427.
Humphrey, G.J. 2008. Influence of large termitaria on species composition and tree height
structure in a modified Brachystegia/Julbernardia miombo woodland in Chizarira
National Park, Zimbabwe. Master of Science in Conservation Biology. Percy
Fitzpatrick Institute of African Ornithology, University of Cape Town.
Hurd, L.E. and Fagon, W.F. 1992. Cursorial spiders and succession: Age or habitat structure?
Oecologia 92:215–221.
Jocqué, R. and Dippenaar-Schoeman, A.S. 1992. Two new termite eating Diores species
(Araneae, Zodariidae) and some observations on unique prey immobilization. Journal
of Natural History 26: 1405-1412.
Johnson, D.H. 1999.
The Insignificance of Statistical Significance Testing. Journal of
Wildlife Management 63(3):763-772.
Jones, J.A. 1990. Termites, soil fertility and carbon cycling in dry tropical Africa: a
hypothesis. Journal of Tropical Ecology 6: 291-305.
Jones, C.G., Lawton, J.H. and Shachak, M. 1994. Organisms as ecosystem engineers. Oikos
69: 373- 386.
Jöreskog, K.G. and Sörbom, D. 1989. LISREL 7: A guide to the program and applications,
SPSS, Chicago.
75
Joseph, G.S. 2008. Large termitaria provide refugia for cavity-using birds in a modified
miombo woodland system. Master of Science in Conservation Biology. Percy
Fitzpatrick Institute of African Ornithology, University of Cape Town, South Africa.
Joseph, G.S., Cumming, G.S., Cumming, D.H.M., Mahlangu, Z., Altwegg, R. and Seymour,
C.L. 2011. Large termitaria act as refugia for tall trees, deadwood and cavity-using
birds in a miombo woodland. Landscape Ecology 26(3): 439-448.
Juen, A., and Traugott, M. 2004. Spatial distribution of epigaeic predators in a small field in
relation to season and surrounding crops. Agriculture, Ecosystems and Environment
103:613–620.
Klein, G.C. 1989. Effects of forest fragmentation on dung and carrion beetle communities in
central Amazonia. Ecology 70: 1715–1725.
Kremen, C. 1992. Assessing the indicator properties of species assemblages for natural areas
monitoring. Ecological Applications 2: 203–217.
Kremen, C., Colwell, R.K., Erwin, T.L., Murphy, D.D., Noss, R.F. and Sanjayan, M.A. 1993.
Terrestrial arthropod assemblages: their use in conservation planning. Conservation
Biology 7: 796–808.
Lambeets, K., Hendrickx, F., Vanacker, S., Van Looy, K., Maelfait, J-P. and Bonte, D. 2007.
Assemblage structure and conservation value of spiders and carabid beetles from
restored lowland river banks. Biodiversity Conservation 17:3133–3148.
Lavelle, P. 1997. Faunal activities and soil processes: adaptive strategies that determine
ecosystem function . Advanced Ecology Research 27: 93 – 132.
76
Laws, R.M. 1970. Elephants as agents of habitat and landscape change in East Africa. Oikos.
21:1–15.
Lawton, R.M. 1978. A study of the dynamic ecology of Zambian vegetation. Journal of
Ecology 66: 175-198.
Lawton, J.H. 1983. Plant architecture and the diversity of the phytophagous insects. Annual
Review of Entomology 28: 23-39.
Lawton, J.H., Bifnell, D.E., Bolton, B., Blowmers, G.F., Eggleton, P., Hammond, P.M.,
Hodda, M., Holt, R.D., Larsen, T.B., Mawdsley, N.A., Stork, N.E., Srivastava, D.S.
and Watt, A.D. 1998. Biodiversity inventories, indicator taxa and effects of habitat
modification in tropical forest. Nature 391: 72-76.
Leroy, A. and Leroy, J. 2003. The spiders of Southern Africa. Struik.
Loveridge, J.P. and Moe, S.R. 2004. Termitaria as browsing hotspots for African
megaherbivores in miombo woodland. Journal of Tropical Ecology 20: 337-343.
MacNally, R. 2000. Regression and model-building in conservation biology, biogeography
and ecology: the distinction between – and reconciliation of – „predictive‟ and
„explanatory‟ models. Biodiversity and Conservation 9: 655 – 671.
Maelfait, J.-P. and De Keer, R. 1990. The border zone of an intensively grazed pasture as a
corridor for spiders Araneae. Magagula, C.N. 2003. Changes in carabid beetle
diversity within a fragmented agricultural landscape. African Journal of Ecology 41:
23-30.
Magagula, C. N. 2003. Changes in carabid beetle diversity within a fragmented agricultural
landscape. African Journal of Ecology 41: 23-30.
77
Makumbe, M.T. 2009. Woody plant species diversity and structure on large termitaria under
different land uses in the Lake Chivero area of Zimbabwe. Master of Science in
Tropical Resource Ecology. Tropical Resource Ecology Programme (TREP),
University of Zimbabwe, Harare.
Malaisse, F. 1978. High termitaria, Biogeography and Ecology of Southern Africa 2: 12791300.
Malinga, N. 2001. Visual interpretation of 1995 aerial photographs and fieldchecks of Lake
Chivero Recreational Park. Department of National Parks and Wildlife Management,
with assistance from the Forestry Commission.
Mallis, R.E. and Hurd, L.E. 2005. Diversity among ground dwelling spider assemblages:
habitat generalists and specialists. Journal of Arachnology, 33: 101–109.
Mapaure, I. 2001. Small-scale variations in species composition of miombo woodland in
Sengwa, Zimbabwe: The influence of edaphic factors, fire and elephant herbivory.
Systematics and Geography of Plants 2: 935-947.
McCune, B. and Mefford, M.J. 2006. PC-ORD. Multivariate Analysis of Ecological Data.
Version 5.10 MjM Software, Gleneden Beach, Oregon, U.S.A.
McFerran, D.M., Montgomery, W.I. and McAdam, J.H. 1994. The impact of grazing on
communities of ground-dwelling spiders (Araneae) in upland vegetation types.
Biology and Environ 94: 119-126.
Mielke, P.W., Jr. 1984. Meteorological applications of permutation techniques based on
distance functions. In Krishnaiah, P.R. and Sen, P.K. (eds.). Handbook of Statistics,
Elsevier Science Publishers. 4: 813-830.
78
Mittermeier, R.A., Mittermeier, C.G., Brooks, T.M., Pilgrim, J.D., Konstant, W.R., da
Fonseca, G.A.B. and Kormos, C. 2003. Wilderness and biodiversity conservation.
PNAS 100: 10309-10313.
Mobaek, R., Narmo, A. K., and Moe, S. R. 2005. Termitaria are focal feeding sites for large
ungulates in Lake Mburu National Park, Uganda. Journal of Zoology 267:97102.Moe, S., Mobaek, R., and Narmo, A. 2009. Mound building termites contribute to
savanna vegetation heterogeneity. Plant Ecology 202:31–40
Negro, M., M., Isaia, C., Palestrini, and A., Rolando 2010. The impact of forest ski-pistes on
diversity of ground dwelling arthropods and small mammals in the Alps. Biodiversity
Conservation 19:1853–1870.
New, T.R. 1999. Untangling the web: spiders and the challenges of invertebrate conservation.
Journal of Insect Conservation 3:251–256.
O‟Reagain, P.J. and Schwartz. J. 1995. Dietary selection and foraging strategies of animals
on rangelands. Coping with spatial and temporal variability. In: Journet, M., Grenet,
E., Farce, M.H., Theriez, M. and Clermont-Ferrand, D.C. (eds). Proceedings IV
International Symposium on the Nutrition of Herbivores. France, September 1995.
INRA, Paris.
Owen-Smith, N. 1988. Megaherbivores: The influence of very large body size on ecology.
Cambridge University Press, Cambridge.
Palmer, M. W. 1990. The estimation of species richness by extrapolation. Ecology 71:11951198.
Parks and Wildlife Board 1975. Robert Mcilwaine Recreational Park: Policy Document.
Approved by the Minister.
79
Petraitis, P., Latham, R., Niesenbaum, R.A., 1989. The maintenance of species diversity by
disturbance. Q. Rev. Biol. 64 (4), 393–418.
R Development Core Team (2010) R: a language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org.
ISBN 3-900051-07-0
Raizer, J. and Amaral, M.E.C. 2001. Does the structural complexity of aquatic macrophytes
explain the diversity of associated spider assemblages? Journal of Arachnology 29:
227-237.
Robertson, E.F. 1984. Regrowth of two African woodland types after shifting cultivation.
Ph.D thesis, University of Aberdeen.
Robinson, J.V. 1981. The effect of architectural variation in habitat on a spider community:
An experimental field study. Ecology 62: 73-80.
Ruggerio, R.G. and Fay, M. 1994. Utilization of termitarium soils by elephants and its
ecological implications. Africa Journal of Ecology 32: 222-232.
Russell-Smith, A. 1999. The spiders of Mkomazi Game Reserve.
(eds.) Coe, M.,
McWilliam, N., Stone, G. and Parker, M. Mkomazi: the ecology, biodiversity and
conservation of a Tanzanian Savanna. London, Royal Geographic Society, pp 608.
Rypstra, A.L. 1983. The importance of food and space in limiting web-spider densities: a test
using exclosures. Oecologia 59: 312–316.
Sanderson, R.A., Rushton, S.P., Cherrill, A.J., and Byrne, J.P. 1995. Soil, vegetation and
space: an analysis of their effects on the invertebrate communities of a moorland in
north-east England. Journal of Applied Ecology 32: 506-518.
80
Scholes, R.J. and Walker, B.H. 1993. An African savanna: Synthesis of the Nylsvley study.
Cambridge University Press, Cambridge, England.
Schroeter, S.C., Dixon, J.D., Kastendiek, J. and Smith, R.O. 1993. Detecting the ecological
effects of environmental impacts: a case study of kelp forest invertebrates. Ecological
Applications 3: 331–350.
Siegel, S. 1956. Nonparametric Statistics for the Behavioral sciences. McGraw-Hill Book
Company, Inc.
Siemann, E., Tilman, D. and Haarstad, J. 1999. Abundance, diversity and body size: patterns
from a grassland arthropod community. Journal of Animal Ecology 68: 824-835.
Skidmore, M.A. 2010. Landscape heterogeneity facilitated by termitaria and its effect on ant
community composition in the miombo woodlands of Chizarira National Park,
Zimbabwe. Master of Science in Conservation Biology, PercyFitzpatrick Institute of
African Ornithology, University of Cape Town.
Snyder, W.E. and Wise, D.H. 1999. Predator Interference and the Establishment of Generalist
Predator Populations for Biocontrol. Biological Control 15: 283-292.
Snyder, W.E. and Wise, D.H. 2001. Contrasting trophic cascades generated by a community of
generalist predators. Ecology 82: 1571-1583.
Soberon, J. and Llorente, J. 1993. The use of species accumulation functions for the
prediction of species richness. Conservation Biology 7: 480–8.
Sørensen, L.L. 2004. Composition and diversity of the spider fauna in the canopy of a
montane forest in Tanzania . Biodiversity and Conservation 13: 437 – 452.
SPSS 16.0 2007. SPSS for Windows. Chicago: SPSS Inc.
81
Thompson, G.G., Withers, P.C., Pianka, E.R. and Thompson, S.A. 2003. Assessing
biodiversity with species accumulation curves; inventories of small reptiles by pittrapping in Western Australia. Austral Ecology 28: 361–383
Thomson, P.J. 1974. The role of elephants, fire and other agents in the decline of a
Brachystegia boehmii woodland. Journal of the Southern African Wildlife
Management Association 5: 11-18.
Tilman, D. 1994. Competition and biodiversity in spatially structured habitats. Ecology 75: 216.
Traoré, S., Nygård, R., Guinko, S. and Lepage, M. 2008. Impact of Macrotermes termitaria
as a source of heterogeneity on tree diversity and structure in a Sudanian savannah
under controlled grazing and annual prescribed fire (Burkina Faso). Forest Ecology
and Management 255(7): 2337‐2346.
Uetz, G. W. 1979. Influence of variation of litter habitats on spider communities. Oecologia
40: 29-42.
Ugland, K.I. Gray, J.S. and Ellingsen, K.E. 2003 The Species-Accumulation Curve and
Estimation of Species Richness The Journal of Animal Ecology 72: 888-897.
Van den Berg, A., and Dippenaar-Schoeman, A.S. 1991. Ground-living spiders from an area
where the harvester termite Hodotermes mossambicus occurs in South Africa.
Phytophylactica 23: 247-253.
Vasconcellos-Neto, J. 2005. The effects of plant structure on the spatial and microspatial
distribution of a bromeliad-living jumping spider (Salticidae). Journal of Animal
Ecology 74:12–21.
82
Warui C.M., Villet, M.R., Young, T.P., and Jocque, R. 2005. Influence of grazing by large
mammals on the spider community of a Kenyan savanna biome. Journal of
Arachnology 33:269–279
Wesolowska, W. and Cumming, M.S. 1999. The first termitivorous jumping spider (Araneae:
Salticidae). Bulletin of the British Arachnological Society 11:204-208.
Whitmore, C., Slotow, R., Crouch, T.E. and Dippenaar-Schoeman, A.S. 2002. Diversity of
spiders (Araneae) in a savanna reserve, Northern Province, South Africa. Journal of
Arachnology 30: 344-356.
White, F. 1983. The vegetation of Africa. Natural Resources Research 20, UNESCO, Paris.
Wild, H. 1952. The vegetation of Southern Rhodesia termitaria. Rhodesian Agriculture
Journal 49: 280-292.
Wise, D.H. 1993. Spiders in ecological webs. Cambridge, Cambridge University Press.
Young, O.P. Edwards, G.B. 1990. Spider in United States field crops and their potential
effect on crop pests. Journal of Arachnology 18: 1-7.
Ysnel, F. and Canard, A. 2000. Spider biodiversity in connection with the vegetation
structure and the foliage orientation of hedges. Journal of Arachnology 28: 107- 114.
Zar, JH. 1996. Biostatistical analysis. 3rd ed. London: Prentice Hall International.
83
APPENDICES
Appendix A: Global Positioning reference points and vegetation cover data for all sites in the
Bird Sanctuary. The dimensions of mound sites are also given
Site
Mound 1
Mound 2
Mound 3
Mound 4
Mound 5
Mound 6
Mound 7
Mound 8
Mound 9
Mound 10
Mound 11
Mound 12
Mound 13
Mound 14
Mound 15
Mound 16
Mound 17
Mound 18
Mound 19
Mound 20
Matrix 1
Matrix 2
Matrix 3
Matrix 4
Matrix 5
Matrix 6
Matrix 7
Matrix 8
Matrix 9
Matrix 10
Matrix 11
Matrix 12
Matrix 13
Matrix 14
Matrix 15
Matrix 16
Matrix 17
Matrix 18
Matrix 19
Matrix 20
Latitude
(S)
17°54.623'
17°54.652'
17°54.707'
17°54.681'
17°54.740'
17°54.813'
17°54.746'
17°54.765'
17°54.826'
17°54.805'
17°54.590'
17°54.759'
17°54.650'
17°54.551'
17°54.520'
17°54.781
17°54.708'
17°54.257'
17°54.280'
17°54.226'
17°54.633'
17°54.636'
17°54.726'
17°54.709'
17°54.743'
17°54.801'
17°54.767'
17°54.748'
17°54.808'
17°54.806'
17°54.597'
17°54.762'
17°54.656'
17°54.530'
17°54.507'
17°54.792'
17°54.708'
17°54.254'
17°54.265'
17°54.230'
Longitude
(E)
30°50.373'
30°50.416'
30°50.398'
30°50.440'
30°50.438'
30°50.451'
30°50.563'
30°50.503'
30°50.547'
30°50.590'
30°50.311'
30°50.651'
30°50.653'
30°50.796'
30°50.827'
30°50.461
30°50.360'
30°50.201'
30°50.312'
30°50.285
30°50.393'
30°50.403'
30°50.393'
30°50.429'
30°50.464'
30°50.437'
30°50.557'
30°50.508'
30°50.545'
30°50.571'
30°50.335'
30°50.636'
30°50.634'
30°50.798'
30°50.836'
30°50.436'
30°50.360'
30°50.219'
30°50.308'
30°50.276'
Mound
Height (m)
3.8
2.8
2.8
3.8
3.2
3.9
1.3
2.9
2.2
1.5
3.2
2.5
3.4
3.4
1.8
-
84
Diameter
1 (m)
25.8
23.9
20.3
24.2
24.7
22.5
12.25
20.79
15.1
11.9
26.9
16.2
19.6
13.63
12.82
-
Diameter 2 Ground
(m)
cover (%)
22.2
60
20.65
63
12.56
77.5
21
81.5
21.5
71.5
21.7
51
10
68
20.22
77
11.31
54.5
9.8
83
22.75
68
12.39
92.5
15.9
74.5
10.15
95.5
12.08
78
57
78.5
71.5
74
85.5
55
89.5
73
83.5
67.5
81
74.5
68.5
70
65
78
100
95
97.5
96.5
83
71.5
78.5
91.5
67.5
Aerial
cover (%)
17.5
43.5
40.5
36.5
41
31.5
18
29.5
20
26.5
24
23.5
37
44
33.5
37.5
43.5
50.5
34.5
27
35
46
51
35
33.5
43
21
29
38
13.5
45.5
12.5
35.5
31
38
30.5
42
30
54
38
Appendix B: Global Positioning reference points and vegetation cover data for all termite
mounds sampled in the Chivero Game Park. The dimensions of the mounds are also given
Site
Mound 1
Mound 2
Mound 3
Mound 4
Mound 5
Mound 6
Mound 7
Mound 8
Mound 9
Mound 10
Mound 11
Mound 12
Mound 13
Mound 14
Mound 15
Mound 16
Mound 17
Mound 18
Mound 19
Mound 20
Mound 21
Mound 22
Mound 23
Mound 24
Mound 25
Mound 26
Mound 27
Mound 28
Mound 29
Mound 30
Longitude
Latitude (S) (E)
17°55.103' 33°49.360'
17°55.188' 33°49.344'
17°55.168' 33°49.322'
17°55.091' 30°49.322'
17°55.017' 30°49.295'
17°55.106' 30°49.239'
17°55.031' 30°49.408'
17°54.991' 30°49.377'
17°55.287' 30°49.395'
17°55.181' 30°49.39'
17°54.941' 30°49.273'
17°55.125' 30°49.417'
17°55.113' 30°49.293'
17°55.072' 30°49.162'
17°55.081' 30°49.096'
17°55.162' 30°49.087'
17°55.198' 30°49.014
17°55.169' 30°48.970'
17°55.237' 30°48.100'
17°55.331' 30°48.895'
17°55.399' 30°48.909'
17°55.441' 30°48.935'
17°55.501' 30°48.872'
17°55.463' 30°48.824'
17°55.398' 30°48.842'
17°55.348' 30°48.798'
17°55.310' 30°48.850'
17°55.274' 30°48.811'
17°55.244' 30°48.778'
17°55.245' 30°48.842'
Mound
Height
(m)
2.2
3.5
3
2.7
1.4
4
3
3.1
3.4
3.2
3
2.3
3.95
3
2.8
2.2
1.5
1.85
2.3
2.8
3.4
3.3
2.8
2.9
2.5
3.25
3.1
2.1
2.3
85
Diameter
1 (m)
11.4
21.8
14.1
11.7
11
16.7
13.05
16.9
14.2
14.5
14.3
10
23.7
13.8
18.9
20.6
18.7
10.1
12.9
20.5
17.5
23.3
25.5
22.3
22.2
22.4
20.4
24
14.3
16.5
Diameter
2 (m)
11.4
16.3
11.5
9.1
8.5
16.05
10.55
12.5
13.02
13.2
11.3
9.5
19.2
12.9
16.6
19.6
15.1
10
12.8
19.4
17
21
23
20.5
21
20.5
18.6
22
14
16
Ground
cover
(%)
46
47
51
52.5
48.5
48
56.5
47.5
50.5
61
53.5
52.5
30.5
46.5
60.5
53
63
66
71.5
65
74
77
91
67.5
72
65
75
83.5
51
75
Aerial
cover
(%)
27
32
39.5
16.5
14
8
20.5
18.5
27
14
23.5
18.5
21.5
17
20.5
24
24.5
16.5
18.5
52
42.5
36.5
26
17.5
23
22
16
38
37.5
19
Appendix C: Global Positioning reference points and vegetation cover data for all woodland
matrix sites sampled in the Chivero Game Park
Site
Matrix 1
Matrix 2
Matrix 3
Matrix 4
Matrix 5
Matrix 6
Matrix 7
Matrix 8
Matrix 9
Matrix 10
Matrix 11
Matrix 12
Matrix 13
Matrix 14
Matrix 15
Matrix 16
Matrix 17
Matrix 18
Matrix 19
Matrix 20
Matrix 21
Matrix 22
Matrix 23
Matrix 24
Matrix 25
Matrix 26
Matrix 27
Matrix 28
Matrix 29
Matrix 30
Latitude (S)
17°55.076'
17°55.165'
17°55.160'
17°55.096'
17°55.031'
17°55.072'
17°55.008'
17°54.993'
17°55.279'
17°55.201'
17°54.915'
17°55.124'
17°55.110'
17°55.073'
17°55.080'
17°55.145'
17°55.188'
17°55.177'
17°55.232'
17°55.333'
17°55.389'
17°55.467'
17°55.529'
17°55.460'
17°55.405'
17°55.338'
17°55.295'
17°55.296'
17°55.245'
17°55.270'
Longitude (E)
30°49.365'
30°49.341'
30°49.323'
30°49.340'
30°49.323'
30°49.230'
30°49.407'
30°49.365'
30°49.426'
30°49.400'
30°49.321'
30°49.443'
30°49.311'
30°49.138'
30°49.114'
30°49.064'
30°49.009'
30°48.962'
30°48.922'
30°48.869'
30°48.931'
30°48.930'
30°48.864'
30°48.851'
30°48.826'
30°48.817'
30°48.855'
30°48.799'
30°48.797'
30°48.846'
86
Ground cover (%)
76
42
50.5
58
50
54
11
64.5
60.5
57.5
69
38
48
68
42
71
64
59
71
66
66
75
91.5
96.5
89.5
76
76.5
88
92.5
86.5
Aerial cover (%)
58
22.5
46.5
52.5
17.5
8.5
75
13
96
35.5
26.5
93
61.5
15
17.5
34
34.5
38
26
29
17
21.5
21
27.5
15.5
20.5
24
26
19
22.5
Appendix D: Vegetation cover data for all sites sampled in the Chizarira National Park. The
dimensions of mounds are also given
Site
Mound 1
Mound 2
Mound 3
Mound 4
Mound 5
Mound 6
Mound 7
Mound 8
Mound 9
Mound 10
Mound 11
Mound 12
Matrix 1
Matrix 2
Matrix 3
Matrix 4
Matrix 5
Matrix 6
Matrix 7
Matrix 8
Matrix 9
Matrix 10
Matrix 11
Matrix 12
Mound
Height (m)
1.8
2.3
1.6
1.85
2.2
2.1
1.65
2.8
2.6
3
1.2
2.4
-
Diameter
1 (m)
11.8
12.3
13
12.3
14.3
14.3
12.5
16.1
14.7
16.62
12.78
12.54
-
Diameter
2 (m)
11
12.1
12.85
11.9
11.6
13.05
11.8
14.46
13.45
16.12
12.72
12.4
-
87
Ground
cover (%)
7
19.9
12.7
6.2
34.8
40.2
54.9
38.6
19.7
8.95
11.7
19.1
46.6
9.6
24.5
10.2
36.5
20.5
34.5
45.1
16.8
64.3
12.3
90
Aerial
cover (%)
19.3
0
38.6
20.8
77.4
74.1
64.6
91.1
59.7
61.5
85.3
52.7
0
6
0
4.3
9.4
11.3
0
6.6
0
11
7.3
0
Appendix E: Ground cover and Aerial cover Recording sheet
DATE: …………….
Recorders: …………………………… Locality: ……………….………..
Mound / Matrix Plot No. …… GPS Coordinates: …………… South, ……………East. Photo Nos:
Transect No. ……… Position of transect: …………………………………………………………
………………………………………………………………………………………………………
Point Ground
No.
Cover
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
….
50
Aerial
Cover
Species
Point Ground
No.
Cover
Aerial
Cover
Species
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
…
100
CODES - Ground cover: BG - Bare Ground; St - Stone: Wd - Wood; Lf - Leaf; Gr – Grass; Sp –
Seed pod
Aerial cover: Dicot leaf - DLf; Twig or branch - Twg; Woody stem - WSt; Grass leaf - GLf :
Grass Stem - GSt
NOTES:
…………………………………………………………………………………………………
…………………………………………………………………………………………………
88
Appendix F: Spider recording sheet
DATE: ……………. Recorders: ………………………………….. Locality: ……………….……….. Mound / Matrix Plot No. …… Sheet No……
GPS Coordinates: …………………. South, ………………… East.
Photo Nos: ……..…………..
Mound dimensions: Height ………m, Diameter 1 ……….m Diameter 2 ………..m.
Active / Not Active Halo / No Halo
Burned/Not Burned
Diagram of Mound Profile:
Spider
No.
Vial
No.
Date
Site
(TM/Ma)
Family
Morpho-species
ID/Name
Species
ID
Identified by:
Date
NOTES:
………………………………………………………………………………………………………………………………………………...
89
Note
No.
Appendix G: Plant species recorded in the Chivero Bird Sanctuary, on Termitaria and
in the woodland matrix
Termitaria
Acacia schweinfurthii
Acacia seiberiana
Albizia amara
Albizia antunesiana
Allophylus africanus
Asparagus racemosus
Azanza garckeana
Boscia salicifolia
Brachystegia spiciformis
Bridelia mollis
Cadaba termitaria
Canthium lactescens
Capparis tomentosa
Cassia abbreviata
Cassine transvaalensis
Catunaregum spinosa
Celtis africana
Clerodendrum glabrum
Clerodendrum myricoides
Clerodendrum mytifolia
Clerodendrum transvaalense
Combretum apiculatum
Combretum collinum
Combretum molle
Combretum zeyheri
Dichrostachys cinerea
Diospyros lycioides
Dombeya rotundifolia
Dovyalis zeyheri
Duranta repens
Ehretia amoena
Ehretia rigida
Euclea crispa
Euclea divinorum
Euphorbia ingens
Ficus natalensis
Ficus zanzibarica
Flueggea virosa
Gardenia volkensii
Grewia bicolor
Grewia flavescens flavescens
Woodland Matrix
Acacia schweinfurthii
Albizia amara
Albizia antunesiana
Annona senegalensis
Asparagus racemosus
Azanza garckeana
Brachystegia glaucescens
Brachystegia spiciformis
Burkea africana
Catunaregum spinosa
Clerodendrum glabrum
Clerodendrum myricoides
Clerodendrum myrtifolia
Clerodendrum transvaalense
Combretum apiculatum
Combretum brachypetalum
Combretum collinum
Combretum molle
Combretum psidioides
Combretum zeyheri
Dichrostachys cinerea
Diospyros lycioides
Dovyalis zeyheri
Eriosema engleriana
Euclea crispa
Faurea saligna
Fluggea virosa
Gardenia volkensii
Grewia retinervis
Jasiminum fluminense
Kochia sp.
Lannea discolor
Lannea edulis
Maytenus heterophylla
Maytenus senegalensis
Monotes glaber
Mystroxylon aethiopicum
Ochna pulchra
Ochna schweinfurtiana
Ozoroa insignis reticulata
Parinari curatellifolia
90
Grewia flavescens olukondae
Grewia monticola
Grewia retinervis)
Jacaranda sp.
Jasminum fluminense
Jasminum stenolobum
Lannea discolor
Lantana camara
Maerua juncea
Maerua triphylla
Maytenus heterophylla
Maytenus senegalensis
Mystroxylon aethiopicum
Ochna pulchra
Pappea capensis
Parinari curatellifolia
Pavetta gardenifolia
Pavetta schumanniana
Peltophorum africanum
Pouzolzia mixta
Psychotria kirkii
Psydrax livida
Pterocarpus angolensis
Pterocarpus rotundifolius
Rhoicissus tridentata
Rhus longipes longipes
Rhuus tenuinervis
Schotia brachypetala
Senna singueana
Solanum delagoense
Strychnos cocculoides
Strychnos potatorum
Teclea trichocarpa
Terminalia brachystemma
Vangueria infausta
Vangueria randii
Vangueriopsis lanciflora
Vernonia amygadalina
Ximenia americana
Ziziphus mucronata
Pavetta gardenifolia
Pavetta schumanniana
Peltophorum africanum
Pouzolzia mixta
Protea angolensis
Psorospermum febrifugum
Psydrax livida
Pterocarpus angolensis
Pterocarpus rotundifolius
Rhus longipes
Rhus tenuinervis
Securidaca longipendunculata
Senna singueana
Steganotaenia araliacea
Strychnos cocculoides
Swartzia madagascariensis
Syzygium sp
Teclea trichocarpa
Terminalia brachystemma
Vangueria infausta
Vangueria randii
Vangueriopsis lanciflora
Vangueriopsis lanciflora
Vitex mombassae
Vitex payos
Ximenia americana
Ximenia caffra
91
Appendix H: Plant species recorded in the Chivero Game Park, on Termitaria and in the
woodland matrix
Termitaria
Albizia amara
Albizia antunesiana
Allophylus africanus
Ehretia amoena
Asparagus racemosus
Bauhinia thonningii
Boscia salicifolia
Brachylaena rotundata
Brachystegia spiciformis
Bridelia mollis
Burkea africana
Cadaba termitaria
Canthium lactescens
Capparis tomentosa
Celtis africana
Clerodendrum transvaalense
Clerodendrum glabrum
Clerodendrum myricoides
Clerodendrum myricoides
Combretum apiculatum
Combretum molle
Combretum zeyheri
Dichrostachys cinerea
Diospyros lycioides
Dombeya rotundifolia
Dovyalis zeyheri
Ehretia amoena
Ehretia rigida
Euclea crispa crispa
Euclea divinorum
Euphorbia ingens
Ficus natalensis
Ficus zanzibarica
Ficus thonningii
Fluggea virosa
Gardenia volkensii
Grewia bicolor
Grewia flavescens flavescens
Grewia flavescens olukondae
Grewia monticola
Grewia retinervis (Grewia flavescens)
Woodland Matrix
Albizia antunesiana
Albizia amara
Allophylus africanus
Annona senegalensis
Brachystegia spiciformis
Bridelia mollis
Burkea africana
Canthium lactescens
Clerodendrum glabrum
Combretum apiculatum
Combretum brachypetalum oatesii
Combretum molle
Combretum zeyheri
Dichrostachys cinerea
Diospyros lycioides
Ehretia amoena
Faurea saligna
Fluggea virosa
Gardenia volkensii
Grewia bicolor
Grewia flavescens flavescens
Grewia monticola
Grewia retinervis (Grewia flavescens)
Lannea edulis
Lapholaena coriifolia
Mystroxylon aethiopicum
Maytenus senegalensis
Monotes glaber
Ochna pulchra pulchra
Ozoroa insignis reticulata
Parinari curatellifolia
Pavetta schumanniana
Peltophorum africanum
Pterocarpus angolensis
Pterocarpus rotundifolius
Rhus longipes longipes
Schotia brachypetala
Securidaca longipendunculata
Senna singueana (cassia singueana)
Solanum delagoense
Solanum incanum
92
Jasiminum stenolobum
Lannea discolor
Lannea edulis
Lantana camara
Lapholaena coriifolia
Maerua juncea
Mystroxylon aethiopicum
Monotes glaber
Maytenus heterophylla heterophylla
Mystroxylon aethiopicum
Ochna pulchra pulchra
Pappea capensis
Parinari curatellifolia
Pavetta gardenifolia
Peltophorum africanum
Pouzolzia lucens
Pterocarpus rotundifolius
Rhoicissus tridentata
Rhus longipes longipes
Rhuus tenuinervis
Schotia brachypetala
Senna singueana (cassia singueana)
Solanum delagoense
Solanum incanum
Terminalia brachystemma
Terminalia trichopoda
Vangueria randii
Vangueriopsis lanciflora
Ximenia americana
Ziziphus mucronata
Psorospermum febrifugum
Strychnos cocculoides
Strychnos potatorum
Swartzia madagascariensis
Syzygium guineense
Terminalia brachystemma
Terminalia trichopoda
Vangueriopsis lanciflora
Vangueria infausta
Vitex mombassae
Ximenia caffra
Ziziphus mucronata
93
Appendix I: Plant species recorded in the Chizarira National Park, on Termitaria and in the
woodland matrix
Termitaria
Acacia nilotica
Andropogon gayanus
Allophylus africanus
Berchemia discolor
Boscia angustifolia
Boscia salicifolia
Brachystegia boehmii
Capparis tomentosa
Cassia abbreviata
Cissus cornifolia
Combretum apiculatum
Combretum collinum
Combretum hereroense
Combretum molle
Combretum mossambicense
Combretum xeyheri
Comiphora mollis
Commiphora mossambicensis
Dalbergia melanoxylon
Dichrostachys cinerea
Diospyros kirkii
Diospyros quiloensis
Diospyros senensis
Erythroxylum zambesiacum
Euclea divinorum
Feretia aeruginenscens
Flueggea virosa
Friesoldielsia obovatum
Grewia monticola
Lannea schweinfurthii
Lannea stuhlmannii
Lonchocarpus capassa
Maerua prittwitzii
Manilkara mochisia
Markhamia zanzibarica
Pterocarpus rotundifolius
Strychnos potatorum
Xeroderris stuhlmannii
Ximenia americana
Ximenia caffra
Ziziphus mucronata
Woodland Matrix
Acacia nilotica
Andropogon gayanus
Aristida bicolor
Aristida leucophaea
Brachystegia boehmii
Brachystegia spiciformis
Bridelia cathatica
Burkea africana
Catuneragum spinosa
Combretum apiculatum
Combretum collinum
Combretum hereroense
Combretum molle
Combretum zeyheri
Crossopteryx febrifuga
Dichrostachys cinerea
Diospyros kirkii
Diplorynchus condylocarpon
Flacourtia indica
Hyparrhenia filipendula
Jubinardia globiflora
Lannea edulis
Lannea discolor
Lonchocarpus capassa
Loudetia flavida
Ozoroa insignis
Pavetta schumanniana
Progonathria squarrosa
Pseudolachnostylis maprouneifolia
Sclerocarya birrea
Terminalia brachystemma
Terminalia sericea
Terminalia sternostachya
Turrea nilotica
Xeroderris stuhlmannii
Ziziphus abyssinica
94
95
Lenin Chari