Citation
The Amphibians and reptiles of Kibale Forest, Uganda

Material Information

Title:
The Amphibians and reptiles of Kibale Forest, Uganda herpetofaunal survey and ecological study of the forest floor litter community
Creator:
Vonesh, James R., 1969- ( Dissertant )
Lillywhite, H. ( Thesis advisor )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1998
Language:
English

Subjects

Subjects / Keywords:
Amphibians ( jstor )
Dry seasons ( jstor )
Fauna ( jstor )
Forest habitats ( jstor )
Forest litter ( jstor )
Forests ( jstor )
Frogs ( jstor )
Rain forests ( jstor )
Reptiles ( jstor )
Species ( jstor )
Dissertations, Academic -- Zoology -- UF ( lcsh )
Zoology thesis, M.S ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent) ( marcgt )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )
Spatial Coverage:
Uganda--Kibale

Notes

Abstract:
The amphibians and reptiles of Kibale National Park in western Uganda were inventoried over an 18-mo. period in 1995 and 1996-97. A total of 75 species, including 28 amphibians and 47 reptiles, were collected or observed. Comparison with other equatorial African herpetofaunas confirms that the Kibale fauna is most similar to those of southwest Uganda and eastern Congo-Zaire, both hypothesized Pleistocene forest refugia. Comparison with a West Africa fauna also shows a fair degree of overlap, while almost no overlap was observed between Kibale and the forests of coastal East Africa. This confirms that the Kibale herpetofauna is an extension of the Guinea-Congolean forest faunas. Randomly placed 5 x 5 m plots were used to sample the herpetofauna of the forest leaf-litter layer in unlogged forest, logged forest, and a neighboring exotic pine plantation. A total of 18 amphibian and reptile species were captured in the litter, a number similar to that observed in mid-elevation tropical forests in Central America and Southeast Asia. Density at Kibale was much lower than most previous studies. Analysis of the feeding ecology of the most abundant litter species showed that most diurnal litter frogs are active foragers of hard-bodied prey such as ants; sit-and-wait predators of larger soft-bodied prey are curiously absent. Plots sampled under fruiting Ficus natalensis trees showed significantly higher prey densities, but litter amphibians and reptiles did not seem to respond to this increase. Of the physical and biotic factors measured in each plot, seasonal changes in soil moisture were most closely correlated with the patterns of herpetofauna abundance observed in the forest. This is consistent with the fact that Kibale receives less rain than any site where the ecology of the litter herpetofauna has been studied, and that most of the species present in Kibale are believed to have evolved in the wetter forests of eastern Congo-Zaire.
Thesis:
Thesis (M.S.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (p. 94-100).
General Note:
Title from first page of PDF file.
General Note:
Document formatted into pages; contains x, 101 p.; also contains graphics (some in color).
General Note:
Vita.
Statement of Responsibility:
by James R. Vonesh.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
81493781 ( OCLC )
002424957 ( AlephBibNum )
AMD0037 ( NOTIS )

Downloads

This item has the following downloads:


Full Text












THE AMPHIBIANS AND REPTILES OF KIBALE FOREST, UGANDA:
HERPETOFAUNAL SURVEY AND ECOLOGICAL STUDY
OF THE FOREST FLOOR LITTER COMMUNITY












By

JAMES R. VONESH


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


1998





































To my mother (who helped me through 3rd grade long division),
my father (who funded my science fair projects),
and Helen Burch and Mrs. Mansfield
(two teachers who made a difference early on)
















ACKNOWLEDGMENTS


Not enough can be said about the role Drs. Lauren and Colin Chapman played in

facilitating this project. Certainly without their support I would never have been able to conduct

this research in Uganda. Both provided invaluable comments on early proposals and later

manuscript drafts, facilitated getting Ugandan research clearance, and provided essential logistic

support while I was in the field. But perhaps most importantly, they helped make much of the

experience just plain fun.

Dr. Lillywhite, serving as the thesis committee chair, and Dr. Max Nickerson, have also

provided valuable comments and suggestions throughout the process of developing this thesis,

and have helped me to significantly improve the quality of the work.

Dr. Robert Drewes, Jens Vindum, and the staff of the Herpetology Department at the

California Academy of Sciences, provided invaluable first hand knowledge of Ugandan frogs and

superb accommodations during visits to use their collections. Visits to the collection for the

purpose of identifying the Kibale specimens were financially supported by the Charles S. Steams

Grant-in-Aid of Herpetology, California Academy of Sciences, and by the California Academy of

Sciences, Department of Herpetology.

I would also like to thank Christopher Amooti and Katusabe Swaibu who worked with

me in the field during the leaf-litter study, and proved to be both excellent company as well as a

reliable and observant field crew.

Finally, I would like to thank my fiancee, field companion, and (truth be known) initial

reason for going to Uganda, Sophia Balcomb. Sophia was instrumental in providing me with the










opportunity to work in Uganda, helped me maintain my sanity throughout the field season,

supported me during the difficult times, and proved a worthy cribbage adversary.

This research was supported in part by the Chicago Zoological Society.

















TABLE OF CONTENTS
page

A C K N O W L E D G M EN T S ................................................................................. .............. ....... iii

LIST O F TA B LE S.................................................................... .......................... ................. vi

L IST O F F IG U R E S .................................................................................................. ......... ...... v iii

A B S T R A C T ................... ....... .... ........................ .. ............ .................................. . .... ix

CHAPTER

1 IN T R O D U C T IO N ............................................. .. .............................. ............ ........ 1

2 THE HERPETOFAUNA OF KIBALE NATIONAL PARK, UGANDA: SPECIES
COMPOSITION AND BIOGEOGRAPHY........................ ..... ........................... 5
In tro d u ctio n ................................................................ 5
M methods .............................................................................. 8
Results ......................... ...... ............... 13
Discussion ......................... ...... ............... 17
Conclusions ......................... ...... ............... 22

3 ECOLOGICAL CORRELATES OF LITTER HERPETOFAUNA: RICHNESS AND
ABUNDANCE.................................................... 38
Introduction ..................................... ............. ............... 38
M eth o d s ................................................................................................................... .............. ...... 4 0
Results ......................... ...... ............... 46
D discussion .......... ............................................ 53
Conclusions ......................... ...... ............... 64

4 CON CLU SION S ............................. .................... 80

APPENDICES

A AMPHIBIANS FROM EIGHT TROPICAL AFRICAN LOCALITIES ............................. 83

B REPTILES FROM EIGHT TROPICAL AFRICAN LOCALITIES. ................................... 88

LIST OF REFERENCES........................................................ 94

B IO G R A PH IC A L SK E T C H .................................................................................................... 101






V
















LIST OF TABLES


Tableage

2.1. Ecological characteristics of the 29 species of amphibians known to
occur in and around Kibale forest, Uganda. .............................................. ............... 23

2.2. Ecological characteristics of the 50 reptile species found in and around
K ibale forest, U ganda ............... ........... ........................... ..... ............................. ..... 26

2.3. Eight equatorial African sites included in herpetofaunal comparisons............................... 32

2.4. Coefficients of faunal similarity for eight equatorial African herpetofaunas.....................33

2.5. A comparison of the "forest-dependent" herpetofauna species richness
in tropical mid-elevation-montane forests of Africa, Central America,
and the Philippines................................... .. ...... ........... 35

3.1. Reptiles and amphibians of the forest floor leaf-litter layer, Kanyawara, Kibale National
Park, Uganda, based on 15 mo of collecting and Pitman (1974)....................................... 66

3.2. Numbers and species of amphibians and reptiles captured in the leaf-litter
of pristine, logged, and pine forest during the wet and dry seasons, Kibale
N national Park, U ganda .............................. ................ ................................ ..................... 67

3.3. Density, richness, evenness, diversity, and similarity of the leaf-litter
herpetofauna calculated for pristine, logged, and pine plantation
forest types at Kibale National Park, Uganda. ....................................................... 68

3.4. Differences among the three forest types in environmental variables
within each plot during the wet and dry seasons, expressed as the mean
and 1 SD, and the P-values associated with each as determined by the
K ruskal-W allis A N O V A .................................................................... ........................... 69

3.5. Results of stepwise logistic regression of the presence or absence of amphibians and
reptiles in leaf litter plots (Kibale National Park, Uganda) predicted by habitat variables. .70

3.6. Numbers of adults and juveniles of the five most common leaf-litter
herpetofauna in the three forest types during the wet and dry seasons.............................. 71

3.7. Results from the stomach content analysis of the six most common leaf-litter herpetofauna
in K ib ale N action al P ark ................................................ ................................................. 72










table page

3.8. Horn's modified Morisita's similarity indices for the six most common litter anurans....... 73

3.9. Results of plot sampling around Ficus natalensis (n = 5) under the canopy
and away from the canopy during the height of the fruiting season
1 m o later. ........................................................... ........ .......... ...... 74

3.10. Repeated measures analysis of variance for the differences in arthropod
abundance, herpetofauna abundance, and litter depth at Ficus natalensis trees
(Kibale National Park) during and after fruiting, under and away from the
canopy. Mauchley's criterion W = 1 in cases with only two sampling intervals ............ 75

3.11. Summary of the quantitative plot studies of mid-elevation tropical litter herpetofaunas
and Scott's (1982) study in lowland W est Africa................ ........................ .............. 77
















LIST OF FIGURES


Figure page

2.1. A map of Kibale National Park, Western Uganda, East Africa, showing
Kanyawara, the primary sampling site, and five supplementary amphibian
and reptile collection sites: Sebatoli, Ngogo, Lake Nyabikere, Dura River, and Mainaro.. 36

2.2. A map of equatorial Africa showing the eight sites that were used in
comparing tropical amphibian and reptile faunas. Dots indicate estimated
extent of tropical African rain forest based on Hughes (1983) and dashed
ines indicate national boundaries.................................................. ............................ 37

3.1. Feeding ecology of the six most abundant leaf-litter species in Kibale
N action al P ark U g an da ..................................................... ............................................... 79
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

THE AMPHIBIANS AND REPTILES OF KIBALE FOREST, UGANDA:
HERPETOFAUNAL SURVEY AND ECOLOGICAL STUDY OF THE FOREST FLOOR
LITTER COMMUNITY

By

James R. Vonesh

December, 1998


Chairman: Dr. Harvey B. Lillywhite
Major Department: Zoology

The amphibians and reptiles of Kibale National Park in western Uganda were

inventoried over an 18-mo period in 1995 and 1996-97. A total of 75 species, including 28

amphibians and 47 reptiles, were collected or observed. Comparison with other equatorial

African herpetofaunas confirms that the Kibale fauna is most similar to those of southwest

Uganda and eastern Congo-Zaire, both hypothesized Pleistocene forest refugia. Comparison

with a West Africa fauna also shows a fair degree of overlap, while almost no overlap was

observed between Kibale and the forests of coastal East Africa. This confirms that the Kibale

herpetofauna is an extension of the Guinea-Congolean forest faunas.

Randomly placed 5 x 5 m plots were used to sample the herpetofauna of the forest leaf-

litter layer in unlogged forest, logged forest, and a neighboring exotic pine plantation. A total of

18 amphibian and reptile species were captured in the litter, a number similar to that observed in

mid-elevation tropical forests in Central America and Southeast Asia. Density at Kibale was

much lower than most previous studies. Analysis of the feeding ecology of the most abundant










litter species showed that most diurnal litter frogs are active foragers of hard-bodied prey such as

ants; sit-and-wait predators of larger soft-bodied prey are curiously absent. Plots sampled under

fruiting Ficus natalensis trees showed significantly higher prey densities, but litter amphibians

and reptiles did not seem to respond to this increase. Of the physical and biotic factors measured

in each plot, seasonal changes in soil moisture were most closely correlated with the patterns of

herpetofauna abundance observed in the forest. This is consistent with the fact that Kibale

receives less rain than any site where the ecology of the litter herpetofauna has been studied, and

that most of the species present in Kibale are believed to have evolved in the wetter forests of

eastern Congo-Zaire.
















CHAPTER 1
INTRODUCTION

Rain forests currently cover about 7% of the Africa continent, and represent slightly

more than one fifth of the total remaining tropical forest worldwide. While rain forests

everywhere are under severe and increasing pressures, a recent survey indicates that African

forests, relative to those of Asia and Latin America, are the most depleted, representing only

about one-third of their historical extent (Collins, 1992). In East Africa, the moist forest

diminishes as the climate becomes increasingly more arid. Forests in East Africa are found

primarily in isolated patches of higher moisture; along rivers, on mountains, or among the coastal

hills. These fragmented forest patches are particularly vulnerable to encroachment and

exploitation, and yet these forests are home to rich plant and animal communities. Relative to

the savanna ecosystem and its charismatic megafauna, the forests of East Africa have received

little attention from zoologists, and what interest they have attracted has been focused primarily

upon their primate, and to a lesser extent bird faunas. Our knowledge of the amphibian and

reptile species of these forests, until recently, was primarily a result of the work of Arthur

Loveridge, of the Harvard Museum for Comparative Zoology (e.g., Loveridge, 1935, 1942a,b,c).

The initial expeditions of Loveridge and the later work on treefrogs by Schiotz (1975) revealed

two distinct forest herpetofaunas in East Africa. The first fauna is restricted to the coastal ranges

of Tanzania, and is characterized by a high degree of endemism (Howell, 1993). The second

fauna is the eastern-most extension of the Guinea-Congolean forests, and is characteristic of the

forests of Uganda and western Kenya. Little research has followed Loveridge's work, with the

notable exception of Schiotz (1975). A few of these forests have been recently surveyed for

amphibians and reptiles (Drewes and Vindum, 1991; Drewes and Rotich, 1995), and Howell










(1993) has summarized his work and that of others in the forests of Tanzania. Even with these

efforts, the amphibian and reptile faunas of most East Africa forests remain poorly studied, or

completely unknown. Considering the extreme pressures that face many of these forests, our

opportunities to learn of these faunas may be limited.

The first objective of this study was to conduct a survey of the herpetofauna of Kibale

National Park in western Uganda. Kibale is a transitional lowland-montane moist forest at the

eastern foot of the Ruwenzori Mountains. It is one of the better studied forests in all of tropical

Africa. Research focusing primarily on primates and forest ecology has been ongoing at Kibale

since the late 1960s, yet little is known about its herpetofauna; and no amphibian or reptile

species list exists for the park. In Chapter 2, I present the results of a survey of the herpetofauna

of Kibale National Park conducted during two trips to Uganda between 1995 and 1997.

Sampling focused on the forest habitat around the Makerere University Biological Field Station

at Kanyawara, but five supplementary sites, encompassing the northern and southern extremes of

the park, were also sampled. The natural history of both the amphibian and reptile assemblages

are summarized based on accounts from literature sources and my observations in the field.

According to previous researchers (Schiotz, 1976; Howell, 1993), the fauna of Kibale

would be expected to share many species with the Guinea-Congolean forests to the west. This

was examined by using Duellman's (1965) Coefficient of Biogeographic Resemblance to

compare the herpetofauna of Kibale with that of eight other equatorial African sites. The

similarity of these faunas is considered in the context of Pleistocene refugia theory.

The zoogeography of tropical African forests is thought to be strongly influenced by

forest expansion, associated with wet interglacial periods, and forest retraction, associated with

colder drier conditions during periods of glaciation. Much attention has focused on the last

glacial maximum, at 18,000 years BP, when the extent of tropical forest was greatly reduced,

creating hypothesized isolated forest refugia (Hamilton, 1976; Moreau, 1969). This vicariance










promoted allopatric speciation in these refugia. Warmer wetter conditions since 12,000 BP have

been associated with expansion of rain forests, which reached their maximum extent at

approximately 7,000 BP, and allowed mixing of previously isolated forest faunas. A number of

authors have used modem patterns of distribution of forest species in order to elucidate past

forest history, and have argued that patterns of richness and endemism suggest two principle core

refugia, one in Cameroon and Gabon, and another in Eastern Zaire, with smaller refugia in West

Africa and coastal East Africa (Hamilton, 1976, 1992; Moreau, 1969). These refugia are

characterized by high species richness and a high proportion of endemic species. Kibale, located

on the plateau between eastern Africa's two great rift valleys, is at an elevation approximately

800 m higher then the lowland rain forest of eastern Congo-Zaire. Due to its greater elevation,

Kibale is expected to be a somewhat species poor neighbor of the hypothesized eastern Zaire

forest refugia, which lies less than 100 km to the west.

Through the work of Loveridge (1935, 1942a,b,c) and Pitman (1974), and more recently

Drewes and Vindum (1991), Howell (1993), Broadly and Howell (1991) and this study, we are

beginning to get a better understanding of amphibian and reptile distributions in the forests of

East Africa. However, few studies have included any detailed examination of the ecology of

these herpetofaunas. One of the objectives of ecology is to determine what factors are important

in determining species' distributions, both on a regional and more local scale.

The leaf-litter of most tropical forests supports a rich herpetofauna that may include

frogs, salamanders, caecilians, lizards, snakes, amphisbaenids, and turtles. Comparative studies

of litter herpetofaunas of southeast Asia and Central America have revealed interesting

differences between the regions. The lowland forests of Central America support a similar

number of species, but at much higher densities (often ten times greater) than the forest of

southeast Asia. There has been a good deal of speculation as to what is responsible for this

difference, and where the litter herpetofaunas of Africa fall in comparison to these regions (May,










1980). Unfortunately, comparisons with Africa at this stage are premature. For, with the

exception of two preliminary studies of Coastal West Africa (Toft, 1982; Scott, 1982), the litter

communities of tropical Africa have been overlooked. Neither Scott (1982), or Toft (1982)

quantitatively addressed the importance of physical, biotic, or anthropogenic factors in

structuring the litter herpetofaunal communities at their sites in West Africa. However, studies

in Central America and Southeast Asia identified a number of physical and biotic factors that

seem to influence litter herpetofaunal abundance and composition on a local scale. These studies

have focused on the litter faunas' use of macrohabitat, microhabitat, food type, food size, diel

time, and seasonal time (Toft, 1985). In the third chapter, I examine aspects of how the litter

herpetofauna of Kibale Forest uses these resource categories and then relate the findings to

existing studies from other tropical regions.

To achieve this objective, I first examine potential habitat correlates of abundance in

three differently managed forest types during wet and dry seasons. This provides information on

the physical and biotic factors most important in structuring the Kibale community and how it

responds to anthropogenic disturbance. Next, I describe the feeding ecology of the six most

common litter anurans, and use dietary overlap indices to provide an indication of the strength of

interspecific interactions along this resource axis. Finally, I examine the local-scale response of

litter arthropods and herpetofauna to large fruiting trees. Specifically, I test the hypothesis that

litter arthropod numbers increase under fruiting fig trees, and that the litter herpetofauna

increases locally in response to increased arthropod abundance. Taken collectively, the different

elements of this study address the initial question of what factors are important in structuring this

community. I compare these results with those from other tropical leaf-litter herpetofauna

studies to see whether the Kibale fauna is similar in diversity and abundance to faunas from

Central and South America and Southeast Asia, and whether the same factors appear to be

important in structuring these communities.
















CHAPTER 2:
THE HERPETOFAUNA OF KIBALE NATIONAL PARK, UGANDA: SPECIES
COMPOSITION AND BIOGEOGRAPHY




Introduction


Africa's tropical moist forests extend from Senegal, West Africa, to montane forests of

eastern-most Somalia (Collins, 1992) and are home to an estimated 333 amphibian (Duellman,

1993), 105 snake (Hughes, 1983), 95 lizard, 16 turtle, and three crocodilian species (Bauer,

1993). While these estimates illustrate the richness of African rain forest herpetofaunas, our

knowledge of these faunas is far from complete. The faunas of tropical Africa and South

America have been studied since the mid-1700s, and until the beginning of this century species

discovery rates were about the same for both continents (Duellman, 1993). However, after 1960

the species discovery rate in South America increased dramatically, while that of Africa

remained roughly the same. In South America the increase in species descriptions coincides with

an increase in the number of South American herpetologists, a phenomenon without African

parallel. The relative paucity of studies from tropical Africa suggests that many species of the

African rain forest are as of yet unknown to science.

The forests of Africa, like those worldwide, are disappearing at an alarming rate.

Deforestation in West Africa is particularly severe, averaging 90% loss of the original forest

cover from Sierra Leone to Nigeria. The forest of central Africa is also threatened from all

directions, with an estimated 57% of the forest of central Congo-Zaire already being lost (World

Resources Institute, 1994). Other regions of Central and East Africa are experiencing similar










loss. Deforestation has been severe in Uganda, with an estimated 86% loss of tropical moist

forest (World Resources Institute, 1994). The remaining forests are primarily isolated fragments,

which are vulnerable to encroachment and exploitation by a rapidly expanding, predominantly

rural population.

Relative to other vertebrate groups, amphibians and reptiles in East Africa have been

poorly studied, and future opportunities may be limited by the threats facing East Africa's

forests. The need for baseline herpetological research in tropical Africa has been pointed out by

a number of authors (e.g., Mittermeier et al., 1992; Lawson, 1993; Drewes and Vindum, 1997),

who discussed the difficulty in conserving faunas we have so little information on.

The herpetofaunas of East African forests are generally thought to contain two non-

overlapping faunas (Schiotz, 1976). The first faunal element is regarded as an eastern extension

of the Congo Forest block stretching from Cameroon to Kakamega Forest in western Kenya. The

second herpetofauna is that of the East African coastal forests. The herpetofaunas of the latter

area have been reviewed recently by Howell (1993), however, few studies have examined the

herpetofaunas of the Central African relict forests in East Africa since Loveridge (1935, 1942a,b,

1957). Such relicts include the Budongo, Bwamba, Kibale, Bwindi, Mbira, and Mt. Elgon

forests in Uganda and reach their eastern limit in the Kakamega forest of Kenya. Of these

forests, only Bwindi-Impenetrable Forest in southwestern Uganda has been inventoried

specifically for herpetofauna (Drewes and Vindum, 1991).

Probably the most thoroughly studied forest in East Africa is Kibale Forest in western

Uganda (Fig. 2.1). Research at Kibale, ongoing since the late 1960s, has resulted in

approximately 160 scientific publications. The vast majority of these focus on Kibale's primate

community (12 species); other animal taxa studied include fish, birds, small mammals, and

elephants. Amphibians and reptiles have received little attention and are the only vertebrate

groups at Kibale for which these are no species lists. However, there have been a few










herpetological collecting expeditions that have visited Kibale. In 1938, Harvard biologist Arthur

Loveridge visited Kibale Forest and camped along the Dura River, one of the sites included in

this study. He collected eight frog species during his 10-d stay (Loveridge, 1942c). The snakes

of Uganda have been well documented primarily through the efforts of Uganda's first game

warden, Captain Charles Pitman, who published "A Guide to the Snakes of Uganda" in 1938 and

a revised edition in 1974. This comprehensive volume included records of 15 snake species as

occurring in Kibale Forest. The tree frogs of East African have been examined in depth by Arne

Schiotz (1975) who visited Uganda in 1968 and lists six tree frogs from Kibale Forest. Based on

the work of these four researchers, 14 frogs, 15 snakes and two lizards were known from Kibale

prior to this study.

The first objective of this paper is to present the results of a herpetofauna survey of

Kibale Forest conducted between 1995 and 1998. In combination with previous records from

Kibale this study provides an overview of the amphibian and reptile richness and basic natural

history of this area. Such presence-absence data are time consuming to collect, and the lack of

such baseline information is likely to be one of the primary factors that has discouraged research

on these taxa at Kibale. Furthermore, in light of mounting evidence of worldwide amphibian

declines, it is increasingly important that baseline information on amphibian distributions be

established, particularly for the poorly studied forests of tropical Africa. A second goal of this

paper is to compare the herpetofaunal composition of eight equatorial African sites, one from

West Africa, four from Central Africa, and three from East Africa, to reexamine the distribution

of tropical forest amphibians and reptiles in the context of proposed past forest refugia, to

establish which regions are most similar to Kibale, and to compare these results with those

observed for other taxonomic groups.










Methods


Description of Study Sites

Kibale National Park is located in western Uganda (013' to 0041'N and 30019' to

30032'E) near the eastern base of the Ruwenzori Mountains (Fig. 2.2). Kibale was established as

a national park in 1993. Prior to that it was managed as a forest reserve with selective timber

extraction and a number of exotic softwood plantations. Currently, approximately 766 km2 is

protected. The park is primarily characterized as a moist evergreen forest, transitional between

lowland rain forest and montane rain forest (Struhsaker, 1997; Skorupa, 1988), but a variety of

habitats including swamp, grassland, woodland thicket, and colonizing scrub are also represented

(Struhsaker, 1997). Kibale lies on the plateau adjacent to the eastern edge of the western Great

Rift Valley. The area was influenced by drifting and volcanic eruptions during the Pleistocene,

and numerous crater lakes lie to the west and northwest of Kibale (Struhsaker, 1997). Soils are

variable among sites, but valley bottoms often have deep, waterlogged, and dark soils

characterized by low pH and fertility; hillslopes often have deep, red sandy loam; and hilltops

have shallow, often rocky soils, or are covered in deep laterite (Lang-Brown and Harrop, 1962).

Rainfall is typically concentrated during two distinct wet seasons, March through May and

September through November. The mean annual maximum temperature measured at Kanyawara

is 23.30C, and mean annual minimum temperature is 16.20C (Struhsaker, 1997). The

management and research histories, fauna, and flora of Kibale have recently been reviewed in

detail by Struhsaker (1997).

Five forest sites were sampled, Sebatoli, Kanyawara, Dura River (near Kanyanchu),

Mainaro, and Ngogo. The first four sites are approximately 10-15 km apart along a N-S gradient

while Ngogo is 12 km south-east of Kanyawara (Fig. 2.1). All four areas consist of a series of










moderately undulating valleys which result in a topographical relief of 150-200 m (Chapman et

al., 1997).

A sixth site, Lake Nyabikere, is a crater lake surrounded by areas cleared for agriculture

and a few small forest patches. Lake Nyabekere lies approximately 1-2 km outside the forest

along the road from Kanyawara to Kanyanchu. This site represents typical "farmbush" habitat

outside the park and was sampled on January 12-13, 1997.

Kanyawara is the site of the Makerere University Biological Field Station and the area of

most intensive sampling. It is situated at an elevation of 1500 m and is characterized by the

steepest terrain, with an average slope of 8.70 (Chapman et al., 1997). The most abundant tree

species are Uvariopsis congensis, Markhamia platycalyx, and Bosqueia phoberos, while large

emergent trees include Parinari excelsa and Pseudospondias microcarpa which reach heights of

30 m (Chapman et al., 1997). Rainfall at this site averages approximately 1600 mm per year.

Aquatic amphibian breeding habitat surveys focused on four Kanyawara sites: the Lower Camp

Well, a permanent artificial pool -15 x 14 m on the forest edge in a swamp forest patch; the K30

Forest Pool, a shallow natural pond -20 m in diameter along the Nykagera stream in unlogged

mature forest; the Mikana stream and seasonal flooded swamp forest; and the Karumbi Road

ephemeral pools, tire ruts on an old logging road that seasonally fill with water. Forest litter

sampling was also concentrated at Kanyawara with 100-140 plots in unlogged forest, selectively

logged forest, and exotic pine plantation.

The Dura River site near Kanyanchu tourist center was the second most intensively

sampled area. This riparian forest at 1250 m elevation along the Dura River is characterized by

the lowest relief (mean slope 5.90) and is dominated by the trees Celtis durandii, Uvariopsis

congensis, and Bequaertiodendron oblanceolatum (Chapman et al., 1997). Several habitats were

sampled at Dura River including eight 5 x 5-m litter plots on the North bank of the river, dip net

sampling of seasonal pools along the river, and visual searches of several stream tributaries of the










Dura and a grassy forest gap (the elephant wallow) at the Kanyanchu tourist center.

Approximately 14 d were spent sampling at this site in 1996-97, including three nocturnal

samples.

Mainaro, at 1200 m, the site lowest in elevation and furthest south, is characterized by

Cynometra alexandri forest along the Dura River (Chapman et al., 1997). Here, the river

seasonally inundates adjacent forest forming pools that become isolated from the river. Three

day time visits were made to this site, in July-August, 1997.

Sebatoli was the northern most site visited. This site, at a slightly higher elevation of

-1590 m (Struhsaker, 1997), is characterized by riparian forest along the Mpanga River, one of

two major rivers in Kibale. During heavy rains the forest is inundated. Three daytime visits

were made to this site, January 13-15, 1997.

The Ngogo field station, at an elevation of 1350 m (Chapman, 1997), was visited for 3 d

in early December, 1997. Nocturnal and diural searches were conducted along the Kanyanchu

stream and surrounding forest. Ngogo receives less rain than Kanyawara, approximately 1490

mm per year. The forest is characterized by moderate topographical relief (slope 6.00) and is

dominated by Uvariopsis congensis, Diospyros abyssinica, and C hii,, phiiilmn albidum

(Chapman et al., 1997).



Sampling

Field work was conducted in and around Kibale from May to August 1995 and

November 1996, to December 1997. Collections of amphibians and reptiles were made using a

variety of techniques. At Kanyawara, four amphibian breeding sites were surveyed twice weekly

(nocturnal and diural) between November 15, 1996 and December, 1, 1997, resulting in

approximately 500 total hours of active searching in these habitats. In addition, 340 5 x 5 m leaf-

litter plots were sampled between March and November 1997 (See Chapter 2). Each plot was










searched for at least 1 person-hr and plots were randomly assigned locations within the forest

using the Kanyawara trail grid map. Opportunistic visual searches of suitable habitats

supplemented these sampling methods at Kanyawara and was the primary sampling method at

the five additional sites, unless otherwise noted above.

Snakes and lizards occur at lower densities than most amphibians and are less frequently

observed in the West African tropics than in comparable Neotropical forest habitats (Lawson,

1993). This is also true in Kibale, and opportunistic road cruising was the most reliable method

for obtaining snake specimens. Most snakes collected in this manner were found between

Kanyawara and Fort Portal or between Kanyawara and the Dura River site (Fig. 2.2). The first

route runs through agricultural and farmbush habitats while the second runs through agricultural

and forest habitat.

In addition to specimens collected during my field work in 1995-1997, a number of

specimens were collected by Drs. Colin and Lauren Chapman between 1990-1995. Most of these

specimens were the results of by-catch in minnow traps used in their studies of the Kibale fish

fauna. In total approximately 500 specimens were collected from Kibale and neighboring areas.

Collection of vouchers was limited to three specimens per gender per locality within the national

park boundaries. Many more individuals were observed than were collected. Specimens were

preserved in the field with 10% buffered formalin and were subsequently transferred to 70%

ethanol. Specimens are shared between the Makerere University Zoological Collection,

Kampala, Uganda, and the California Academy of Sciences, San Francisco.



Faunal Comparisons

Herpetofaunal lists for the eight equatorial African sites compared in this study were

compiled primarily from published lists: Korup National Park, Cameroon (Lawson, 1993); Parc

National des Virunga, Congo-Zaire (Amphibians Laurent, 1972; Reptiles De Witte, 1941);










Parc National de la Garamba, Congo-Zaire (Amphibians- Inger, 1968; Reptiles De Witte,

1966); Kibale National Park, western Uganda (this study); Bwindi-Impenetrable National Park

(Amphibians Drewes, 1991; Reptiles Drewes and Vindum, 1997 unpub. report), Usambara

Forest Reserve, northeast Tanzania (Howell, 1993); Uzungwa Forest Reserve, Tanzania (Howell,

1993); and the Arabuko-Sokoke Forest Reserve, coastal Kenya (Drewes and Rotich 1995). The

nomenclature used in several of these studies is antiquated, and attempts were made to update

genera and species titles. Frost (1985) was followed for most amphibians; however, I followed

Duellman and Trueb (1986) and Drewes and Vindum (1991) in regarding arthroleptid frogs as a

subfamily of Ranidae. I also retained the genus Hylarana in keeping with other African workers

(e.g., Perret, 1977; Drewes and Vindum, 1991; Lawson, 1993). Lizard nomenclature follows

Broadly and Howell (1991) for many species, and snake nomenclature follows Meirte (1992),

Pitman (1974), and Hughes (1981, 1985).

For faunal comparisons I use Duellman's (1965) modification of Pirlot's (1956) formula



CBR = 2C/(N1 + N2)



where CBR is the Coefficient of Biogeographic Resemblance, C is the number of species two

areas share in common, N1 is the number of species in the first area, and N2 is the number of

species in the second area. This algorithm was used because it takes into account the size of the

larger fauna and has been used in numerous comparisons of Neotropical herpetofaunas (e.g.,

Dixon, 1979; Hoogmoed, 1979; Duellman, 1990).










Results


Herpetofaunal inventory and natural history

Fourteen amphibian species, 13 lizard species, and 15 snake species previously unknown

from Kibale were collected and/or observed during this study. The resulting herpetofauna

consists of 28 frogs, 15 lizards, and 32 snakes (Table 2.1). In addition, one amphibian and six

snakes are listed as expected for the park based on their presence in other nearby Ugandan

forests.

All amphibians observed belonged to the order Anura (frogs). Salamanders (Caudata)

are not known from sub-Saharan Africa, and caecilians (Gymnophiona) have never been

collected in Uganda. Among anurans, three aquatic species (11% of fauna) of Xenopus were

collected; X wittei and X. victorianus were often found in sympatry. The terrestrial leaf litter

frog assemblage consisted of nine species (32% of fauna) from three genera, Bufo,

Phrynobatrachus, and Schoutedenella. Of the Bufonids, only B. funereus and B. kisoloensis

were forest-dwelling, while B. maculatus was abundant outside the park in farmbush habitat.

Five species of Phrynobatrachus are known from Kibale, but two are very rare.

Phrynobatrachus dendrobates was collected by Loveridge in 1938, but has not been collected

since; and I collected the only specimen of P. versicolor known from Kibale. Of the remaining

three species, P. graueri was the most commonly heard and seen frog at most sites in the park

and was often found breeding at the same sites as the smaller P. parvulus. Phrynobatrachus

auritus was uncommon at Kanyawara but frequently observed at the Dura Site. Hylarana

albolabris and Rana angolensis were common both in and outside the forest in association with

water bodies, and could best be classified as semi-aquatic species (7%). Both Ptychadena

species were associated with forest gaps, though P. mascarenensis (hylaea) appeared to be more

strictly limited to the forest. Twelve arboreal or semi-arboreal frogs (43%) of the genera










Afrixalus, Leptopelis, Hyperolius, Phlyctimantis, and Chiromantis were collected or observed in

the park. Of these, four Hyperolius species, Leptopelis christyi, and Phlyctimantis verrucosus

were reported as occurring in Kibale by Schiotz (1975). The new additions to the Kibale tree

frog list were Afrixalus laevis, known from a single specimen collected at Sebatoli; Hyperolius

platyceps langi; Leptopelis kivuensis; and the rhacophorid, foam-nesting tree frog Chiromantis

rufescens (M. Cherry, pers com.). This is the second Ugandan record for C. rufescens, a West

and Central African species which reaches its eastern limit in Budongo Forest Reserve and

Kibale in western Uganda.

Little is known about the feeding ecology of most of these frog species. Frogs typically

prey upon invertebrates, but other taxa are sometimes taken. Xenopus wittei preys primarily

upon insects while the largerX. 1. victorianus has been observed to feed on insects, other frog

larvae, and small fish (pers. obs.). The feeding ecology of the litter assemblages is examined in

greater detail in Chapter 2. The two forest bufonids and the Schoutedenella are active foragers

which feed primarily on ants, P. graueri feeds on Collembola, and the tree frogs L. kivuensis and

H. lateralis appear to be sit-and-wait predators on soft-bodied insects.

Five reproductive strategies are exhibited by Kibale's frogs. Most species (61%) deposit

their eggs in water, and eggs hatch into aquatic larvae. Seven species (25%) from three genera

(one Phrynobatrachus, two Afrixalus, four Hyperolius) deposit their eggs on vegetation above

water, which subsequently hatch and drop into the water as aquatic larvae. The two Leptopelis

(7%) species both bury their eggs in the moist soil of depressions that later fill with water, upon

which aquatic larvae hatch, and Chiromantis rufescens (4%) deposits eggs in arboreal foam nests

from which aquatic larvae hatch. Only the arthroleptid S. schubotzi exhibits direct development

(4%).

Lizards were uncommon within the forest. The most abundant species were those

associated with the forest edge or disturbed farmbush habitats (e.g., H. mabouia, S. atricollis, M.










straita, C. ellioti, Table 2.2). Within the forest, the two lacertids A. africanus and A. vauereselli

were the most conspicuous, actively foraging in the leaf litter and basking in sun-flecks. Of these

two species, A. africanus appeared to be more arboreal. The highly arboreal lacertid H.

guentheri was observed only once, but its habits make it extremely difficult to collect. Of the

forest chameleons, Rhampoleon boulengeri was the most common species, and was usually

observed in low shrubs. The other two forest chameleons, C. adolfifriderici and C. ituriensis,

appear to be more arboreal in their habits.

Of the 37 species of snakes that occur or potentially occur in Kibale, 11% are fossorial or

burrowing, 38% are primarily terrestrial, 51% are primarily arboreal, and less than 3% are

primarily aquatic (Table 2.2). Most species are probably nocturnal or active during both day and

night. The feeding ecology of Kibale's snakes can be surmised based on diets reported in Pitman

(1974), the stomach contents of specimens from this study, and field observations. Of the 37

species, 8% feed primarily on invertebrate prey, 8% feed on fish, 24% prey on lizards or other

snakes, 51% prey on frogs, 37% feed on small mammals such as rodents, 14% feed on birds, and

5% (Bitis gabonica, Python sebae) are capable of taking larger mammals as prey (categories add

to >100% because many species prey on several taxa). Among the most specialized feeders are

Dasypeltis atra, which feeds exclusively on bird eggs, Duberria lutrix, which preys on slugs,

Causus lichtensteinii, which feeds exclusively on forest toads, and Thelotornis kirtlandii and

Lycophidion ornatum which feeds on lizards.



Herpetofaunal Comparison

The herpetofauna (excluding crocodilians) from eight equatorial African sites were

compared using Duellman's (1965) Coefficient ofBiogeographic Resemblance (CBR). These

sites ranged from Coastal West Africa (Korup 804'E) to Coastal Kenya (Arabuko 39030'E),

and from 50N (Korup) to 7050'S (Uzungwa) latitude (Fig. 2.2). Several broad habitat categories










are found among these sites. Korup and Virunga are composed primarily of Guinea-Congolean

rain forest. Korup is mostly lowland rain forest (97% < 850 m) while the elevational range in

Virunga extends from 710 m to over 5000 m in the Ruwenzori Mountains (Table 2.3). Garamba

is characterized primarily by moist mixed woodlands and savanna, with gallery forests along

larger streams and in deeply entrenched ravines (Inger, 1968). Both Kibale and Bwindi in

western Uganda are situated upon the central African plateau between the western and eastern

Rift Valleys. For that reason the lowest elevations at both sites are approximately 1100 m.

However, only Bwindi, with a maximum elevation of 2607 m has true montane rain forest. Both

the Tanzanian sites, the Usambaras and the Uzungwa Mountains, are characterized by moderate

elevation to montane rain forest, representing a similar range of elevations as Kibale and Bwindi

in Uganda. The final site, Arabuko-Sokoke, is comprised of relatively dry coastal forest. The

combined species lists of these sites include 191 amphibians (APPENDIX A.), 10 turtles, 98

lizards, and 134 snake species (APPENDIX B.). This represents roughly 55% of the total

estimated frog fauna of African rain forests, and a higher proportion of African forest reptiles.

All of these sites were sampled over a period of at least 3 mo, and most were sampled for

considerably longer. Total species richness ranged from 167 species in Korup to 33 species in

Uzungwa (forest dependent species only). The areas of these parks vary considerably; therefore,

it is also valuable to consider richness controlling for area. Arabuko-Sokoke has the highest

richness per area (2.4 spp/l000ha), followed by Bwindi (2.0), and Korup and Kibale (both 1.3,

Table 2.3).

Kibale N. P., Bwindi N. P., Virunga N. P. exhibit the greatest similarity among the eight

sites (Table 2.4). Kibale and Bwindi are separated by about 200 km along a N-S gradient in

western Uganda, and share 20 species of amphibians and reptiles (Table 2.4). Both sites show

strong similarity to the Virunga site (23 and 24 shared species respectively), from which they are

separated by less than 100 km (Table2.4). The lower CBR values for Kibale versus Virunga and










Bwindi versus Virunga compared to Kibale versus Bwindi are due to the larger size of the

Virunga fauna. These three neighboring Central African forests exhibit the greatest similarity

among the eight sites.

The herpetofauna of Korup in Cameroon is most similar to that of Virunga, 2300 km to

the East, and least similar to the montane forests of Tanzania and the coastal forests of Kenya,

over 3300 km distant. The Usambaras and Uzungwas are most similar to each other, and show

very little similarity to any other sites. Likewise, Arabuko-Sokoke shows little similarity to other

forests, however, it is interesting that it is more similar to Garamba and Virunga, over 1400 km

distant, than the nearby forests of the Usambara Mts. (220 km), Uzungwa Mts. (740 km), or the

intervening Ugandan forests. Reptile faunas were more similar (mean CBR 0.19) than

amphibian faunas (mean CBR 0.14) among the eight areas (Table 2.4).

A comparison of mid-elevation to montane tropical rain forests from Africa, Central

America, and southeast Asia show that "forest dependent" herpetofaunas of higher elevation sites

range between 20 and 67 species (Table 2.5). The Virunga N. P. and Monteverde, Costa Rica

supported the richest higher elevation forest herpetofaunas, N. P. and Bwindi N. P. support an

intermediate number of species, while Cueros de Negros, Philippines, and the Tanzanian

Uzungwa Mts. had the lowest species richness. Reptiles outnumbered amphibians at all sights

except the Uzungwas and Korup (at Korup this is likely due to low sampling effort in higher

elevations). Amphibian richness was greatest in Korup, and reptile richness was highest in

Virunga N. P.




Discussion


The Kibale herpetofauna consists of two main assemblages, those species which seem to

be forest-dependent and those that do well in a variety of habitats, both inside and outside the










forest. Because of the lack of ecological, physiological, and behavioral studies for African forest

amphibians and reptiles, it is often difficult and somewhat subjective to decide which species are

truly "dependent" upon forest habitat. This difficulty has been discussed for the forest

amphibians of Tanzania (Howell, 1993), East African amphibians and reptiles (Loveridge, 1935,

1957), tree frogs (Schiotz, 1967, 1975), snakes (Hughes, 1981), and birds (Moreau, 1966).

In Kibale, the non-forest dependent, or farmbush, component of the herpetofauna is

characterized by the frogs X 1. victorianus, B. maculatus, S. schubotzi, H. albolabris, P.

parvulus, P. chrysogaster, P. mascarenensis, R. angolensis, A. quadrivittatus, H.

cinnamomeoventris, H. kivuensis, H. nasutus, and H. n i,,,hi,,' i,, roughly 46% of the frog fauna.

Five lizards, H. mabouia, C. ellioti, M striata, M maculilabris, and M megalura are found

primarily outside the forest, and two, S. atricollis and L. fernandi, are edge species that were

never observed in the forest interior. Of the remaining eight species, two were collected only

from exotic pine plantations adjacent to the forest, and locality records for these, C.

quattuorseriatus from Bwindi (Drewes and Vindum, 1997) and L. aloysiisabaudiae from

Garamba (De Witte, 1966), suggest that these species also are not forest-dependent.

Approximately 60% of the lizard fauna should be considered part of the farmbush assemblage.

Hughes (1981) reviewed the forest, farmbush, and savanna snake faunas of tropical Africa, and I

follow his classification. Accordingly, of the 37 known and expected snakes in Kibale, 32% (12)

are common in both forest and savanna habitats, A. irregularis, B. gabonica, C. hotamboeia, D.

lutrix, L. ornatum, L. lineatus, N. melanoleuca, P. angolensis, P. phillipsii, P.sebae, T.

angolensis, and T. punctatus. Philothamnus semivariegatus, known from a single specimen

collected outside the park, is considered strictly a savanna form by Hughes (1981, 1985).

The transitional nature of Kibale between lowland and montane forest is evident in the

herpetofauna. Among the forest-dependent species in Kibale there are elements of both a higher

elevation montane fauna, consisting primarily of species endemic to Central Africa, and a










lowland forest fauna which often ranges to West Africa. Wide ranging species, characteristic of

the lowland rain forest of the Guinea-Congolean forest include 14% of the frogs (B. funereus, P.

auritus, A. laevis and C. rufescens), 20% of the lizards (L. fernandi, A. africanus, and H. g.

guentheri) and 46% of the snakes (17 species). The montane forest-dependent species

characteristic of Central Africa make up 28% of the Kibale frog fauna (X. wittei, X vestitus, B.

kisoloensis, P. dendrobates, P. graueri, P. versicolor, P. verrucosus and L. kivuensis), 13% of

the lizard fauna (A. vauereselli, C. adolfifriderici) and 9% of the snake fauna (Philothamnus

heterodermus ruandae, A. nitschei, and Dasypeltis atra, Laurent, 1974). The remaining species

represent mid-low elevation forest forms that have distributions limited to Central Africa. This

category includes four frogs (H. lateralis, H. platyceps, H. kivuensis, and L. christyi) the Ituri

forest chameleon, C. ituriensis, and four snakes (P. christyi, G. depressiceps, T. jacksonii, D.

jamesoni).

Clearly the farmbush assemblage represents a significant proportion of the Kibale

herpetofauna. Bwindi-Impenetrable N. P., the site most similar to Kibale, is characterized by a

higher proportion of forest-dependent species. At Bwindi, 70% of the frogs (versus 60% at

Kibale) are forest-dependent. This is largely due to the existence of five montane forest tree frog

species and the montane endemic R. ruwenzorica which do not occur in Kibale. The lizards

show the same pattern: 50% are forest dependent at Bwindi (vs 40% at Kibale), and again this is

largely due to the presence of two endemic montane chameleons and two endemic montane

skinks. The greater representation of farmbush species in the Kibale herpetofauna is further

illustrated by the similarity of the Kibale fauna to that of Garamba, a non-forest, moist savanna

woodlands site. The herpetofauna of Garamba, which is characterized by moist savanna

woodlands, shares 30 species in common with Kibale, but only 16 in common with Bwindi.

Hamilton (1976, 1981) and Struhsaker (1981) hypothesized that Bwindi functioned as a forest

rufugia during the arid conditions of the last glacial maximum. Thus, the larger number of










Albertine Rift endemic frogs and reptiles at Bwindi relative to Kibale is probably a function of

its greater elevational range and age.

The zoogeography of tropical African forests is thought to be strongly influenced by

forest expansion, associated with wet interglacial periods, and forest retraction, associated with

colder drier conditions during periods of glaciation. Much attention has focused on the last

glacial maximum, at 18,000 years BP, when the extent of tropical forest was greatly reduced,

creating hypothesized isolated forest refugia (Hamilton, 1976; Moreau, 1969) This vicariance

promoted allopatric speciation in these refugia. Warmer wetter conditions since 12,000 BP have

been associated with expansion of rain forests, which reached their maximum extent at

approximately 7,000 BP, and allowed mixing of previously isolated forest faunas. A number of

authors have used modem patterns of distribution of forest species to elucidate past forest

history, and have argued that patterns of richness and endemism suggest two principle core

refugia, one in Cameroon and Gabon, and another in Eastern Zaire, with smaller refugia in West

Africa and coastal East Africa (Hamilton, 1976, 1992; Moreau, 1969).

Those critical of refugia theory have argued that modem distribution patterns reflect

recent environmental conditions and tell us little about the past. The argument most commonly

used to support this criticism is based on the positive correlation between areas of high rainfall

and high species richness in tropical Africa (Hamilton, 1992). Three primary responses rebut

this criticism. First, Hamilton (1992) suggests that areas of current rainfall are likely to have

been areas of past heavy rainfall because the broad atmospheric circulation patterns over Africa

at 18,000 BP were similar to those of today. Second, the refugia are not only species and

endemic rich, but they represent isolated populations of disjunct species distributions, such as the

gorilla (Hamilton, 1992). Finally, few species appear to be endemic to forest regions outside the

proposed refugia (Diamond and Hamilton, 1980).










Faunal comparisons among the eight sites examined here are generally consistent with

forest refugia theory. The high richness of Korup N. P. and Virunga N. P. is consistent with

hypothesized core refugia (Hamilton, 1988), though lack of distribution data from intervening

areas precludes finer scale examination of potential gradients of declining diversity between

these areas (Hamilton, 1988). Kibale, Bwindi, and Virunga are more similar to each other than

to the fauna of West Africa, represented by Korup. This is consistent with the hypothesis that

these areas have been connected more frequently or over a longer time period. Schiotz (1976)

suggested that the forests of Uganda and western Kenya are depauperate outliers of Virunga, the

eastern Zaire core area. The lower total richness observed at these two national parks is probably

a function of both smaller area and their increased elevation and corresponding loss of many

lowland rain forest forms present on the Zaire side of the Albertine Rift. However, when the

number of species per area is considered, both Kibale and Bwindi exhibit higher values then

Virunga N. P. In fact, the smaller parks consistently exhibited higher species per area values.

This may be correlated with habitat degradation (L. Chapman, pers. com.), and illustrates the

importance of conserving these remaining forests. These forests are now small islands, but they

may have maintained many of the species that were present when they were part of larger,

continuous forest.

The montane forests of Tanzania are believed to represent a minor refugia that has been

isolated from the Congo forest block for millions of years, a conclusion supported by the high

endemism of many taxa in these forests. In the Usambara Mountains, 87% of the amphibians

and 55% of the reptiles are endemic to Tanzanian montane rain forest (Howell, 1993). Faunal

comparisons with Kibale also suggest a long isolation period. The Kibale herpetofauna is five

times more similar to Korup, 2420 km distant, than to the Usambaras, 1050 km away; and there

is no overlap between the Uzungwa fauna and any site except Usambara.










These results also demonstrate the unique fauna of the coastal dry forest represented by

the Arabuko-Sokoke site. Coastal forest is one of the most imperiled African habitats and is

home to a distinctive herpetofauna unlike that of the wetter tropical forest. Interestingly,

Arabuko has the greatest species overlap with moist savanna woodlands of Garamba.




Conclusions


Kibale forest supports a rich herpetofauna of more than 70 species. This relatively high

species richness is derived from overlap of two primary faunal elements, farmbush species and

the forest-dependent species. The relatively high number of species observed is probably related

to the heterogeneous habitat of forest, swamp, grassland, and agricultural areas characteristic of

Kibale. Despite the large number of farmbush species, the number of forest-dependent species in

Kibale is comparable to that of other mid-elevation forests in Africa, Central America, and the

Philippines. None of these species are endemic to Kibale, and only a few are restricted to the

Albertine Rift. Most of the endemic amphibian and reptile species of the Albertine Rift are

characteristic of higher elevation, montane forest not present at Kibale. Previous researchers

have noted that the forest herpetofaunas of East Africa consist of two, non-overlapping groups.

One group represents the extension of the Guinea-Congolean rain forest fauna, the second is

largely endemic to the coastal mountains of Tanzania. The Kibale herpetofauna is clearly an

example of the first group, as suggested by the high species overlap with the Parc National des

Virunga, in eastern Congo-Zaire, and secondarily to Korup National Park in Cameroon. The

high similarity among Virunga, Bwindi, and Kibale relative to the similarity of these forests to

the forests of West and East Africa is consistent with the hypothesized core forest refugia of

eastern Zaire.












TABLE 2.1. Ecological characteristics of the 29 species of amphibians known to occur in and around Kibale forest, Uganda. (x = Albertine Rift
endemic)


Taxon Abuna Sitesb Habitatc Dield Microe Foodf Repro.! General Distributionh


Pipidae

Xenopus laevis victorianus

Xenopus wittei

Xenopus vestitus

Bufonidae

Bufo funereus

Bufo kisoloensis

Bufo maculatus

Ranidae, Arthroleptinae

Schoutedenella schubotzi x

Ranidae, Petropedetinae

Phrynobatrachus auritus

Phrynobatrachus dendrobates x

Phrynobatrachus graueri


C

U

R



U

I

C-U


1,2

1,2

2



1-4,5

1

6


FI, FE, A

FI, FE

FI?



FI

FI

A, FE


Aq

Aq

Aq



L,U

L, S

L


C-U 1,2 FE, FI, A D L, U


1-3,5

2

1-4


FI

FI

FI, FE


L,R

L,R

L, S


I, F?

I

I?



A,I

A,I

A,I?


E. Zaire W. Kenya

E. Zaire, Uganda, Rwanda

E. Zaire, Uganda, Rwanda



forests; Angola to Uganda

forests; E. Zaire and E. Africa

tropical Africa


A TrDd E. Zaire Uganda, Rwanda?


Aq

ArAq

Aq


forests; W. Africa Uganda

forests; E. Zaire W. Uganda,

forests; Zaire Uganda











TABLE 2.1


Taxon Abun Sites Habitat Diel Micro Food Repro General distribution


Phrynobatrachus parvulus

Phrynobatrachus versicolor x

Ranidae, Raninae

Hylarana albolabris

Ptychadena chrysogaster

Ptychadena mascarenensis

Rana angolensis

Hyperoliidae

Afrixalus laevis

Afrixalus osorioi

Afrixalus quadrivittatus

Hyperolius cinnamomeoventris

Hyperolius kivuensis

Hyperolius lateralis

Hyperolius nasutus


C-U

R


FI, FE

FI


L, S

L, R


FE, FI, A

FE

FI

FI, FE



FI

FI, FE

FI, FE

FI, FE

FI, FE

FI

A, FE


ArAq

ArAq

ArAq

ArAq

ArAq

ArAq

Aq


Angola W. Tanzania

montane forest; E. Zaire, Uganda



Liberia Uganda

mid-elevations; Uganda, Rwanda

forests; W. Africa Uganda

sub-Saharan Africa



Cameroon Uganda

E. Zaire Uganda

Cameroon Kenya

Cameroon Kenya

Zaire W. Kenya

forests; Zaire-Kenya

grasslands; tropical Africa











TABLE 2.1.


Taxon Abun Sites Habitat Diel Micro Food Repro General distribution

Hyperolius platyceps langi I 1,2 FI N AH ? ArA forests; E. Zaire Uganda

Hyperolius v,,/,,i/,,ir, bayoni U 1,6 A, FI N AL ? Aq E. Zaire Uganda

Leptopelis christyi U 1-5 FI N AH ? TrAq forests; E. Zaire Uganda

Leptopelis kivuensis x U 1-3 FI N AH I TrAq forests; E. Zaire Uganda

Phlyctimantis verrucosus U 1,3 FI N AL ? Aq forests; E. Zaire Uganda

Rhacophoridae

Chiromantis rufescens R 1 FE N A ? FoAq Nigeria Uganda

Notes: a Abund, relative abundance: C = common-one can find many specimens; U = usual-one can find when looking in the proper habitat
during the appropriate season; I = infrequent-not predictable; R = rarely seen; E = expected-not yet reported from Kibale but presence
seems very likely
b Sites, collected or observed: 1 = Kanyawara; 2 = Dura River near Kanyanchu; 3 = Mpanga River at Sebatoli; 4 = Ngogo field station;
Mainaro; 6 = Lake Nyabikere
c Habitat: FI = forest interior; FE = forest edge, or opening; A = agricultural areas outside the forest, farmbush
d Diel, time of activity: D = diurnal; N = nocturnal; DN = variably active either night or day
e micro, microhabitat and vertical position: Aq = aquatic; L = forest leaf litter; T = terrestrial at forest edge or opening; A = arboreal; U=
upland; S = swamp forest or valley bottoms; R = riparian or streamside; AL = semi-arboreal, in low vegetation; AH = highly arboreal, in
tree canopy
fFood: I = insects; A = ants, F = fish. Most preferred indicated first, if known.
g Rep. mode, reproductive strategy: Aq = eggs and larvae aquatic; ArAq = eggs arboreal, larvae aquatic; TrDd = eggs terrestrial, larvae
direct developing; Tr Aq = eggs terrestrial, larvae aquatic after flooding; FoAq = eggs in arboreal foam nest, larvae aquatic
h General distribution: references include Inger (1968); Schmit and Inger (1968); Laurent (1972); De Witte (1941); Schiotz (1975)










TABLE 2.2. Ecological characteristics of the 50 reptile species found in and around Kibale forest, Uganda. (x = Albertine Rift endemic)


Taxon Abuna Sitesb Habitatc Dield Microe Foodf General distribution


Sauria, Gekkonidae

Hemidactylus mabouia

Cnemaspis quattuorseriata

Agamidae

Stellio (Agama) atricollis

Chamaeleonidae

Chamaeleo adolfifriderici x

Chamaeleo ellioti

Chamaeleo ituriensis

Rhampoleon boulengeri x

Scincidae

Mabuya striata

Mabuya maculilabris

Mabuya megalura


1,2,6

1


C 1,5,6


1

1,6,7

1,3

1



1,6,7

5?

5?


FE, A

A, FE


FE, A



FI

A, FE

FI

FI, FE



A, FE

A, FE

A, FE


tropical Africa

forests?, E. Zaire, Uganda, Kenya, Ethiopia


D A


A

AL

A

L, AL



L

L


? Angola-Ethiopia & Natal


forests, E. Zaire, W. Uganda

S. Sudan, E. Zaire, W. Kenya, Uganda

forests, E. Zaire, W. Uganda

forests, E. Zaire, W. Uganda


I E. Zaire-Ethiopia & Zimbabwe

? West & Central Africa











TABLE 2.2.


Taxon


Leptosiaphos aloysiisabaudiae

Lygosomafernandi

Lacertidae

Adolfus africanus

Adolfus vauereselli

Holoaspis guentheri guentheri

Serpentes, Typhlopidae

Typhlops angolensis

Typhlops punctatus congestus



Boidae

Python sebae sebae

Colubridae

Bothrophthalmus 1. lineatus

Crotaphopeltis hotamboeia


Abun

R

U



U

U

R?



I

E






I-R



U

R


Sites

1

1,7



1

1

1



7


Habitat

FE, A

FE, A



FI, FE

FI, FE

FI



A, FE

FI


Diel

D

D



D

D

D



N

N


Micro

L

L



A,L

L,A

AH



B

B


Food

?.

?.



I

I

I



W, I

W, I


General Distribution

E. Zaire-W. Uganda

Bioko & Cameroon, through to Uganda



forests, Cameroon, Zaire, Uganda

forests, E. Zaire, W. Uganda, W. Tanzania,

forests, Sierra Leone to Uganda



montane grasslands, Cameroon to Kenya

forests, Liberia to Kenya. Uganda localities:

Bwamba, Budongo, Kigezi forests


1,7 FI, FE, A N T, Aq B, M Liberia to Kenya


FI

FE, A


T, Aq?

T, Aq


forests, Angola & Guinea to Uganda

Central and East Africa











TABLE 2.2.


Taxon Abun Sites Habitat Diel Micro Food General Distribution


Lamprophis lineatus

Lamprophis olivacea

Lycophidion ornatum

Mehelya poensis


Mehelya stenopthalmus

Duberria lutrix atriventris

Geodipsas d. depressiceps

Psammophis phillipsiii

Natriciteres olivacea

Philothamnus angolensis

Philothamnus carinatus

Philothamnus heterodermus

Philothamnus semivariegatus

Philothamnus nitidus


A, FE

FI

FI, FE

FI


8

1,7

1

5

1,5

1,2,6?,7

1

1

7

1


FE

A, FE

FI

A, FE

FE, A

A, FE

FI, FE

FI, FE

FE, A

FE


Ro, Re

Ro

Re

Re


N

N

DN

D

DN

DN

DN?

DN

DN?

DN?


T

T

T

T

T, Aq

A

A

A

A

A


Re

Ga

Am

Ro, Am

Am, F

Am

Am

Am

Am

Am


grasslands, Seirra Leone to Uganda

forests, Liberia to Uganda

forests, Zaire, Uganda

forests, Liberia to Uganda.

Uganda: Budongo, and Kigezi forests

W. Africa to Uganda

grasslands, Uganda, Kenya, Rwanda

forests, Nigeria, Cameroon to Uganda

Liberia to Uganda

grasslands, Ghana to Kenya

Cameroon to Uganda

forests, Liberia to Kenya

forest, Zaire, Uganda, Rwanda, Burundi

East Africa

Liberia to Uganda











TABLE 2.2.


Taxon Abun Sites Habitat Diel Micro Food General Distribution


Hapsidophrys lineatus

Gastropyxis smaragdina


Boiga blandingii

Boiga pulverulenta

Dipsadoboa u. unicolor

Thrasops j. jacksonii

Rhamnophis aethiopissa

Thelotornis kirtlandii

Polemon christyi

Dasypeltis atra

Viperidae

Causus lichtenstienii

Atheris s. squamiger

Atheris n. nitschei x


2,8 FE, FI DN?

FI DN?


1,8

8

7,8

1,2,7,8

2,8

1,2,7,8

7

1,8



1,5

1,2,8

1


FE, FI

FI, FE

FE, FI

FE, FI, A

FI, FE

FE, FI, A

FE, FI, A

FI



FI, FE

FI, FE

FI, FE?


DN

N

N

DN

DN?

DN

N?

N



DN

N

N


Am

Am, Re


A

A

A

A

A

A

B

T,A



T

A

A


B, M

M

Am

B, M, Re

Am

Re

Re

B eggs



Am

Ro, Am

Ro, Am


forests, Liberia to Kenya

forests, Angola & Guinea to Uganda.

Uganda: Bwamba, Chambura, Mbira forests

forests, Liberia to Kenya

forests, Liberia to Kenya

forests, Liberia to Kenya

forests, Zaire to Kenya

forests, Liberia to Kenya

forests, Liberia to Kenya

forests, Zaire to Kenya

montane forests, Zaire to Kenya



forests, Liberia to W. Kenya

forests, Ghana to Kenya

montane forests, Zaire and Uganda











TABLE 2.2.


Taxon Abun Sites Habitat Diel Micro Food General Distribution


Bitis nasicornis

Bitis gabonica


U 1,3,7,8 FI, FE, A N


T, AL Ro, Am, F forests, Equatorial Guinea to Kenya


N T M, B, Am


Cameroon to Kenya; Uganda: Budongo,

Bwamba forests, Queen Elizabeth N. P.


Atractaspididae


Atractaspis irregularis


FE, FI


N B Ro E. Zaire to Kenya; Uganda: Budongo,


Bwamba, Mbira, Kigezi forests


Elapidae


Naja melanoleuca

Pseudohaje goldii


I 1,5,8


FE, FI DN T


F, Re, Am wide-spread through rain forest region


FI N A Am forests, Ghana to W. Kenya below 1,500 m


Uganda: Bwamba & Mbira forests


Dendroaspis jamesoni


U 1,7,8 FE, FI, A N A, T Ro, B forests, Zaire, W. Kenya, Uganda


Notes: a Abun, subjective relative abundance based on frequency of encounters and number of specimens: C = common, one of the most
frequently seen snakes; U = usual, sighted several times over the year, more than one specimen; I = infrequent, seen only once or twice,
may only have a single voucher; R = rare, reliable report, or a single specimen; E = expected, based on habitat preference and Ugandan
distribution







31


Table 2.2
Notes:
b Sites, collected and observed: 1 = Kanyawara station; 2 = Dura River near Kanyanchu; 3= Sebatoli; 4 = Ngogo; 5 = Mainaro, or
southern art of park; 6 = Lake Nyabikere; 7= agricultural area between Ft. Portal and Kanyawara; 8 = reported from Kibale by Pitman
(1974).
c Habitat: FI = forest interior; FE = forest edge or opening; A = agricultural areas outside forest, farmbush
d Diel: D = Diurnal; N = Nocturnal; DN = variably active, day and night
e Micro, microhabitat and vertical position: Aq = aquatic; B = burrowing; L = forest litter; T = terrestrial; A = arboreal; AL = arboreal in
low vegetation; AH = Arboreal in tree canopy
f Food: I = insect; W = annelids; Ga = Gastropods; F = fish; Am = amphibians; Re = reptiles; B = birds; M = general mammals; Ro =
rodents; most information on snake diet based on Pitman (1974) and personal observations
g General distribution: primarily providing East-West range; most snake range data from Pitman (1974), lizards references include
Loveridge (1942a) and Arnold (1989),











TABLE 2.3. Eight equatorial African sites included in herpetofaunal comparisons


Site Latitude Elevation (m) Rain (m/yr) Area (ha) Amphibians Reptiles spp/area Reference


Korup N. P., Cameroon

Garamba N. P., Zaire


Virungas N. P., Zaire



Kibale N. P., Uganda

Bwindi N.P., Uganda



Usambara F.R.,Tanzania

Uzungwa F. R..,Tanzania

Arabuko-Sokoke F. R.

E. Kenya


50N

40 N



loS



<1N

10 S



50 S

7050' S

3030' S


0-1950

710-1061


798-5119



1100-1590

1160-2607



1000-2286

300-2800

-10-70


2.3-5.3

1.5


0.5-3.0



1.5

1.4



0.7-2.0

-2.0

0.6-0.9


124000

490000


780000



56000

31000



621000

100000

36000


1.3 Lawson (1993)

0.24 DeWitte (1966)

Inger(1968)

0.21 DeWitte (1941)

Laurent (1972)

1.3 Pitman (1974)

2.0 Drewes (1991)

Drewes (1997)

0.08 Howell (1993)

0.3 Howell (1993)

2.4 Drewes (1995)

Drewes (1996)











TABLE 2.4. Coefficients of faunal similarity for eight equatorial African herpetofaunas; Korup
(KOR), Virungas (VIR), Garamba (GAR), Kibale (KIB), Bwindi (BWI), Usambaras (USA),
Uzungwa (UZU), and Arabuko-Sokoke (ARA). Values > 0.50 in bold.


Amphibians

KOR


VIR GAR KIB BWI USA UZU ARA


0.26

0.10

0.11

0.09

0.01

0

0.02


18




0.26

0.56

0.54

0

0

0.05


6

13




0.26

0.15

0

0

0.15


KOR

VIR

GAR

KIB

BWI

USA

UZU

ARA

Reptiles



KOR

VIR

GAR

KIB

BWI

USA

UZU

ARA


0.37

0.37

0.18

0.08

0

0.09


0.47

0.55

0.46

0.06

0

0.19


0.34

0.20

0.04

0

0.25


6 5

24 23

9 5

- 20


0.73

0

0

0.04


0.04

0

0.04


KIB BWI

23 10

43 33

21 11

- 27


0.67

0.08

0

0.09


0.13

0

0.08


KOR VIR GAR

- 33 27

0.36 43


1

0

0

0

1




0.5

0.04



USA

4

4

2

3

4




0.27

0.09


0

0

0

0

0

10




0




UZU

0

0

0

0

0

6




0


ARA

6

16

5

5

4

4

0











TABLE 2.4.
Amphibians and Reptiles

KOR VIR

KOR 51


VIR

GAR

KIB

BWI

USA

UZU

ARA


0.32

0.24

0.25

0.14

0.05

0

0.06


0.40

0.56

0.49

0.04

0

0.14


GAR

33

56




0.32

0.18

0.02

0

0.22


KIB BWI

29 15

67 56

30 16

- 47


0.69

0.05

0

0.07


0.09

0

0.07


USA

5

4

0

3

5




0.38

0.07


ARA

7

18

22

6

5

5

0


UZU

0

0

0

0

0

16




0











TABLE 2.5. A comparison of the "forest-dependent" herpetofauna species richness in tropical
mid-elevation-montane forests of Africa, Central America, and the Philippines.


Site Elevation Amphibians Reptiles Totals

Usambara F. R., Tanzania 1000-2286 m 23 29 52

Uzungwa F. R., Tanzania 300-2800 m 19 16 35

Kibale N. P., Uganda0 1400-1550 m 15 29 44

Bwindi N. P., Ugandab 1200-2600 m 20 21 41

Virungas N. P., Congo' 1300-3000 m 27 -40 67

Korup N. P., Cameroond 1080-1768 33 11 44

Monteverde, Costa Ricae 1300-1470 m 25 36 61

Cuernos de Negros, 1050-1350 m 7 13 20

Philippinesf

Notes: a only Kibale species restricted to the forest interior are included
b species listed as extralimital by Drewes and Vindum (1991) are excluded
c only forest species found above 1300 m are included
d only species found at Mt. Yuhan (to 1079 m); Rumpi Hills (1,000-1,768),
and Nta Ali (to 1,200) are included. Lawson (1993) comments that these
elevations were under sampled.
e only species restricted to Timmerman's (1981) pre-montane and lower
montane zones (2-5) are included.
only specimens in submontane and montane forest zones are included
from Brown and Alcala (1961).























Lake













UGANDA







Scale (km)

FIGURE 2.1. A map of Kibale National Park, Western Uganda, East Africa, showing
Kanyawara, the primary sampling site, and five supplementary amphibian and reptile collection
sites: Sebatoli, Ngogo, Lake Nyabikere, Dura River, and Mainaro.





























500 km


FIGURE 2.2. A map of equatorial Africa showing the eight sites that were used in
comparing tropical amphibian and reptile faunas. Dots indicate estimated extent of
tropical African rain forest based on Hughes (1983) and dashed lines indicate national
boundaries
















CHAPTER 3
ECOLOGICAL CORRELATES OF LITTER HERPETOFAUNA: RICHNESS AND
ABUNDANCE




Introduction


The leaf-litter layer of most tropical rain forests support diverse herpetofaunas which

may include frogs, salamander, caecilians, amphisbaenids, lizards, snakes, and turtles. These

often sizable faunas are composed of species which may be ecologically similar and

phylogenetically closely related. Given these similarities, it is natural to ask, how do these

species coexist? A growing number of researchers have attempted to address this question by

trying to determine which physical and biotic factors are important structuring these

communities, and how these communities vary among tropical regions (Lloyd et al., 1968; Scott,

1976; Toft, 1980a; Allmon, 1991; Heinen, 1992; Gascon, 1996).

Comparative study of litter herpetofaunas of Southeast Asia and Central America has

revealed interesting differences between the regions. The lowland forests of Central America

support a similar number of species, but at much higher densities (often 10 times as great) as the

forests of Southeast Asia. Unfortunately, comparisons with Africa at this stage are premature.

For, with the exception of two preliminary studies of coastal West Africa (Toft, 1982; Scott,

1982), the litter communities of tropical Africa have been overlooked.

Neither Scott (1982) nor Toft (1982) quantitatively addressed the importance of physical,

biotic, or anthropogenic factors in structuring the litter communities at their sites in West Africa.

However, studies in Central America and Southeast Asia have identified a number of physical










and biotic factors that seem to influence litter herpetofaunal abundance and composition on a

local scale. Scott (1976), Fauth et al. (1989), and Heinen (1992) found that litter depth was

correlated with herpetofaunal abundance and diversity and Allmon (1991) found a similar

relationship with dry litter mass. Physical characteristics such as elevation, slope, and moisture

(Scott, 1976), and biotic characteristics such as prey abundance (Toft, 1980a; Lieberman, 1986;

Guyer, 1988) and juvenile recruitment (Allmon, 1991) also correlate with litter herpetofauna

abundance in some systems. Both physical and biotic factors may vary seasonally, causing

corresponding changes in the litter community either directly (increased desiccation risk during

the dry season) or indirectly (greater prey abundance during the dry season). Heatwole (1982)

reviewed the structuring of herpetofaunal assemblages and found that species richness decreases

with increasing latitude and altitude and with decreasing availability of moisture, while the

animal density is highest at intermediate elevations.

Anthropogenic factors, such as implementation of different forest management

strategies, may also affect in the distribution of litter amphibian and reptile species. In Costa

Rica, Lieberman (1986) found increased abundance but lower diversity and evenness of litter

species in anthropogenically disturbed sites relative to primary forest. Similar patterns have been

reported from South America (Miyata, 1980) and Malaysia (Inger, 1980b).

Habitat, food, and time are three traditional categories of resources, which may be

further divided into macrohabitat, microhabitat, food type, food size, diel time, and seasonal time

(Toft, 1985). In this study I will examine aspects of how the litter herpetofauna of Kibale

National Park, Uganda, uses each of these different resources and then relate these findings to

existing studies from other tropical regions.

To achieve this objective, I first examine potential habitat correlates of litter

herpetofaunal abundance in three differently managed forest types during both wet and dry

seasons. I will use these data to identify the physical and biotic factors most important in










structuring the Kibale community and how it responds to anthropogenic disturbance. Next, I

describe the feeding ecology of the six most common litter anurans, and use dietary overlap

indices to provide an indication of the strength of interspecific interactions along this resource

axis. Finally, I examine the local-scale response of litter arthropods and herpetofauna to large

fruiting trees. Specifically, I test the hypothesis that litter arthropod numbers increase under

fruiting fig trees, and that the litter herpetofauna increases locally in response to increased

arthropod abundance. Taken collectively, the different elements of this study will address the

initial question: "What factors are important in structuring this community". The next step is to

compare these results with those from other tropical leaf-litter herpetofauna studies to get a

better idea of whether the Kibale fauna is similar in diversity and abundance to faunas from

Central and South America and Southeast Asia, and whether factors that appear to be important

in structuring those communities are similar to those important in Kibale.





Methods


Study site

This study was conducted at the Makerere University Biological Field Station, Kibale

National Park, in western Uganda (00 13' 00 41'N and 300 19' 300 32'E). The field station is

located near the eastern base of the Ruwenzori mountains at an elevation of 1530 m. The forest

is transitional between lowland and montane rain forest with a typical canopy height of between

20 30 m (Struhsaker, 1997). The general topography of the study area consists of moderately

undulating hills and valleys. Mean annual rainfall at this site between 1977 and 1991 was 1622

mm, most of which fell during the two rainy seasons, March May and September November

(Struhsaker, 1997).










Kibale received national park status in 1993. Prior to that it was managed as a forest

reserve, and various compartments of the forest were commercially exploited to different

degrees. As a result, the forest surrounding the field station is comprised of a mosaic of

disturbed and undisturbed sections. For this study, I selected an unlogged study site, the K 30

forest compartment, which is comprised of approximately 300 ha of undisturbed, mature,

Parinari forest (Struhsaker, 1997). The K 15 timber compartment served as the heavily logged

study site. K 15 was selectively logged in 1968 69 at an intensity of 7.4 commercial stems per

hectare (Skorupa, 1988). The areas adjacent to K 15 include lightly logged and heavily logged

forest, and conifer plantations. My third study area, Nyakatojo, is a monotypic stand of mature,

exotic pine (Pinus caribaea) on a low ridge draining into the Dura and Nyakagera Rivers,

directly abutting the unlogged K 30 compartment. This area was planted with pines between

1963 and 1965, and at that time it was dominated by elephant grass Pennisetum purpureum

(Kingston, 1967; Chapman and Chapman, 1996). Such grassland areas in Kibale are believed to

be the result of past human clearing of forest which were then maintained by frequent

arthropogenic burning (Kasenene, 1987). At the time of this study, Nakatojo was characterized

by a mature pine canopy, while the understory vegetation included many indigenous forest tree

species (Chapman and Chapman 1996; Zanne, 1998). The greatest distance between any two of

these three areas is less than 5 km.



Site and Season Comparisons

I follow Heinen (1992) in defining the litter herpetofauna as all reptiles and amphibians

that live directly in the leaf litter, or were found within 0.5 m of the forest floor, and feed on

other litter organisms (e.g., arthropods, frogs, lizards). Using a map of the trail system in each of

the three areas, I randomly generated plot sampling locations at the beginning of the study. Since

statistical variances for mean densities approach an asymptote when based on 50 or more










quadrats, even in areas where densities are low (Lloyd et al., 1968; Heyer et al., 1994), I sampled

50 plots each season in each of the three forest types. No plot was sampled more than once. Wet

season plots were sampled between March 11 May 15, and dry season plots were sampled

between June 12 July 26, with four to eight plots being sampled in each area per week. Each

plot was 5 m x 5 m (Fauth et al., 1989; Allmon, 1991; Heinen, 1992) and enclosed with a clear

plastic fence 50 cm high immediately prior to sampling. The bottom of the fence was secured

using large nails. Once the fence was in place, 2 4 (usually 3) people would search the entire

plot on hands and knees, overturning all litter and logs, and sifting through the first few

centimeters of top-soil. A mean of 58.5 (33 108) person-minutes was spent searching each plot.

On occasions when the a priori plot location fell in flooded swamp-forest, extremely dense

vegetation, or in a forest gap, an alternative location was randomly generated.

All amphibians and reptiles captured were identified to species, measured (snout vent

length), and classified as adult or subadult. After voucher specimens of each species were

preserved, subsequent captures were identified, measured, and released. Voucher specimens are

deposited in the zoological collection of Makerere University, Kampala, Uganda, and in the

herpetological collection of the California Academy of Sciences, San Francisco, U.S.A.

For each plot, I recorded: (1) air and wet bulb temperatures, (2) slope in degrees

(clinometer); (3) percent soil humidity (soil humidity meter); (4) soil pH (soil pH meter); (5) wet

litter mass of a randomly selected 0.5 m2 subplot; (6) mean litter depth, measured to the nearest

0.5 cm, 1 m in from each comer of the plot; (7) estimated percent of low ground vegetation cover

under 1 m high; (8) estimated percent shrub (1 3 m) cover; (9) percent canopy cover, measured

using a spherical densitiometer; (10) number of logs > 10 cm diameter; and (11) number of trees

> 10 cm dbh (diameter at breast height).

Among forestry compartments, I compared the observed total number of species in an

area (s), the Jackknife estimate of species richness (S, Heltshe and Forrester, 1983), diversity










(H'), and evenness (J'), as suggested by Hayek (1994) and Krebs (1989). To facilitate

comparison with previous studies, the logarithmic base 10 was used to calculate both Shannon's

H' and Pielou's J'. The Jackknife richness estimate (S), rather than the number of species

observed (s) was used to estimate the maximum value of H' when calculating J'. Morisita's index

of similarity (C, Morisita, 1959) was used to measure the similarity of the herpetofaunas in each

forest type. Krebs (1989) recommends this index for quantitative data because, unlike many

other commonly used indices, it is not affected by sample size. Data on the abundance of forest-

floor arthropods in the unlogged, logged, and pine plantation areas for the wet season sampling

period months of March May and the dry season months of June August were available from

Nummelin (1989).

Among-site and within-site and season comparisons of physical and biotic variables were

made using Kruskal Wallis nonparametric analysis of variance with a Bonferroni adjusted alpha

level for 12 tests (SPSS, 1997). For significantly different tests among sites multiple

comparisons were conducted using the Mann Whitney test. Nonparametric tests were used in

the analysis because many of the data to be compared had unequal variances. The relationship

between these physical and biotic variables and the presence or absence of amphibians and

reptiles was examined within and among sites and seasons using forward stepwise logistic

regression (Trexler and Travis, 1993; SPSS, 1997), which included the following variables:

slope, pH, wet litter mass, litter depth, number of logs, ground vegetation cover, shrub cover,

canopy cover, number of trees, and hill category. The hill category ranged from (5) = valley

bottom to (1) = hill-top.



Feeding Ecology

I examined the diets of the six most abundant litter anuran species by analyzing the

stomach contents of the voucher specimens from this study as well as the stomachs of other










museum specimens previously collected in Kibale forest between 1990-1997 by various other

researchers. For three species, specimens from Bwindi-Impenetrable National Park, in

southwestern Uganda were included in the analysis in order to increase the sample size. For

these three species, the prey of Bwindi and Kibale specimens were compared to determine if

large differences in the feeding ecology of these species exist. Because of the paucity of material

available, and the desire to minimize the number of animals sacrificed, specimens collected in all

the forest compartments adjacent to the field station during both wet and dry seasons were

combined. In addition, for all six species, both genders and all age categories were also

combined. All specimens were were fixed in 10% formalin then stored in ethanol. The six

species examined were Phrynobatrachus graueri (n = 46), Bufofunereus (n = 17), Bufo

kisoloensis (n = 11 Kibale N. P., n = 6 Bwindi N. P.), Schoutedenella schubotzi (n = 16 Kibale N.

P., n = 3 Bwindi N. P.), Leptopelis kivuensis (n = 16 Kibale N. P., n = 33 Bwindi N. P.), and

Hyperolius lateralis (n = 14). Stomachs were excised and contents were identified to at least

Order. For each prey taxon, the frequency of occurrence (%F = number of stomachs containing

that taxon divided by number of stomachs with contents), relative abundance (/oN = number of

individual prey items of that taxon divided by the total number of prey items), and percent of

prey volume (%V = total estimated volume of that taxon divided by total prey volume for all

specimens) were calculated for each species. Prey volume for each prey type was estimated

based on the volume of a cylinder, and was calculated using an average of the length and width

(nearest 0.1 mm) of several intact specimens of each prey type. The Relative Importance Index

(RI) of each prey taxon was calculated using the following George and Hadley (1979), where:

AI = /oN + %V + %F,

n
RI= 100 AI/YAI
1










(AI) is the 'absolute importance index' for each prey taxa and n is the number of different prey

categories. This index provides a single estimate of dietary importance by combining relative

abundance (%N), frequency of occurrence (%F), and percent volume (%V). Percentage RI values

were calculated by summing the RI values for each prey taxa among all species and then dividing

by the number of species. Values for this index range from 0 to 1.0. The diet of each species

was then compared using Horn's modification of the Morisita's index of similarity for

percentage RI values for prey taxa:

CH = (2XP1Plk)/(yP2l + P21k)

In this equation CH is the simplified Morisita's index of similarity (Krebs, 1989) between species

j and species k. P, and Pk is the proportion of the prey taxa consumed by the two species.



Fruiting Fig Trees and Litter Herpetofauna Distribution

In order to determine the effect of large fruiting events on the distribution and abundance

of leaf-litter arthropods and herpetofauna, I sampled five Ficus natalensis during their fruiting

peak and again 1 mo later (between 5 September and 1 November 1997). Six randomly placed

plots were sampled at each tree, three under the canopy (within the fruit-rain) and three

approximately 3 m out from the canopy edge. The location of each plot was marked with

flagging, and the same area was sampled 30 d post-fruiting. By 30 d post-fruiting, no fallen fruit

remained in the litter, and no evidence of our first sampling was visible. All plots were sampled

during the rainy season. In addition, arthropod abundance was estimated in each plot using three

50 cm2 sticky traps set on the litter surface for 18 hr, two in opposite covers and one in the

center of each plot. A general linear model repeated measures ANOVA (SPSS, 1997) was used

to test for a significant fruiting effect, canopy effect, and fruiting-canopy interaction effect on

arthropod abundance, herpetofauna abundance, and litter depth.










Results


Site and Season Comparisons

During the study, a total of 211 animals were sampled in plots, consisting of 10 frog

species, five lizard species, and three species of snakes. Based on 15 mo of regular collecting

and literature records (Pitman, 1974), the complete litter herpetofauna at the field station

probably contains between 29 and 32 species of amphibians and reptiles (Table 3.1). This is

consistent with the jackknife species richness estimate, based on all 300 plots, of 25 5 species.

However, plot sampling did not appear to sample all herpetofaunal taxa equally. Frogs and

lizards appear to have been sampled fairly well,

with the exception of the riparian Phrynobatrachus species. Phrynobatrachus auritus, P.

dendrobates, and P. versicolor appear to be limited to the stream-side habitats. The latter two

species are abundant in similar habitats at higher elevations in Bwindi National Park in

southwestern Uganda (Drewes and Vindum, 1991). Only three of 13 possible species of

terrestrial snakes were observed during the study.

Five frog species (Bufofunereus, B. kisoloensis, Phrynobatrachus graueri, P. parvulus,

and Leptopelis kivuensis) were found in all three forest types (Table 3.2). Three species were

observed in two forest types. Schoutedenella schubotzi was absent from the logged forest

samples; Leptopelis christyi was absent from the unlogged plots; and Hyperolius lateralis was

found only in the logged and unlogged forests. A single Phlyctimantis verrucosus individual was

observed in the pine plantations during the study, and a single Rana angolensis was found in the

logged forest.

The distribution of lizard species was much less even, relative to the amphibians. Three

of the five lizard species were found only in the pine plantations, (Cnemaspis quattuorseriata,

Rhampoleon boulengeri, and Leptosiaphos aloysiisabaudiae, Table 3.2). The lacertid Adolfus










africanus was found in the pines as well as unlogged forest, and single A. vauereselli was

collected from the logged forest. Observations of snakes in the plots were rare (four individuals),

and only one species, Geodipsas depressiceps depressiceps was captured more than once. The

other two species, Bothropthalmus lineatus and Dasypeltis atra were both found in the logged

forests. It is important to note that the absence of a species from the plots in a particular forest

type is more a reflection of its lower abundance in that site relative to the others, not its absence.

Based on opportunistic sightings in these areas, only Cnemaspis quattuorseriata seems to be

limited to one forest type, the pines. The fossorial skink, L. aloysiisabaudiae, is also known only

from the pines, but that is likely due to its secretive nature and rarity.

Over the entire study, more animals were captured in the logged forest than in the other

two areas combined (Table 3.3). The highest density of animals (5.84 animals/100 m2) was

observed in the logged forest in the wet season. Density decreased in all sites during the dry

season, ranging from a 29% decrease in animals captured in the unlogged forest to a 55%

decrease in the logged forest. A total of 12 species were captured in both the pines and the

logged forest, while only nine were captured in the unlogged section. Richness was similar in

both seasons, except in the pine plantation where it dropped from 10 species in the wet season to

seven in the dry season. The decrease in abundance, while richness remained static or declined,

resulted in a higher evenness (J') in the unlogged and logged sites during the dry season sample.

The pine plantation and unlogged forest had similar evenness measures, and both were higher

than that of the logged forest (Table 3.3). The pine plantation had the highest overall diversity

(H'), followed by the unlogged, then logged forests. Seasonally, diversity increased in the dry

season, except in the pines where it remained essentially the same. Morisita's similarity indices

(C ) show a high degree of similarity between the unlogged and logged forests, and little

similarity between the pine plantation and the other two habitats.










Kruskal -Wallis ANOVA and post hoc Mann Whitney multiple comparison tests were

used to compare the physical and biotic variables among the three sites during the wet and dry

seasons. During the wet season, the unlogged site was significantly steeper than the other two

areas, and had significantly less litter (Table 3.4). The logged forest had significantly less low

ground vegetation relative to the other sites (Table 3.4). The pine plantation had significantly

heavier and deeper litter, more low vegetation, less shrub cover and canopy cover than the other

sites. The pines also had significantly more reptiles per plot (Table 3.4). Within the dry season

samples, the unlogged remained the steepest area, and had the lowest litter mass. The logged

area again had the lowest ground cover, but in the dry season had significantly higher shrub

cover and the fewest trees per plot. The pines had the lowest dry season canopy cover, and

significantly higher litter mass, litter depth, and number of trees per plot relative to the other

areas. Seasonal changes in these three areas followed a similar pattern (Table 3.4): soil

humidity, litter mass, and litter depth all decreased and soil pH increased in the dry season.

There were no significant differences in the number of frogs or reptiles among these

areas during the dry season. The decrease in the number of frogs for all three sites combined in

the dry season was significant (Kruskal-Wallis, x2 = 12.9, P < 0.001). However, within sites the

decrease was not statistically significant except in the logged forest (K-W, x2 = 10.1, P < 0.01).

Stepwise logistic regression was used to identify significant environmental predictors of

the presence of amphibians and reptiles for all three areas combined in the wet season and the

dry season. For all three areas combined in the wet season the only significant predictor of

herpetofaunal presence was soil humidity (Table 3.5). The goodness of fit of this model is

indicated by the overall percent of plots correctly classified in the classification table. In this

model the presence or absence of herpetofauna in plots was correctly predicted in 59.1% of 150

cases. The R statistic indicates the importance of the individual contribution of each independent

variable, values for the R statistic range between -1 and +1. A positive value indicates that as the










variable increases, so does the likelihood that the plot will contain herpetofauna. A small R

value for a variable indicates that the variable has a small partial contribution to the model

(SPSS, 1997). Soil humidity was positively related to herpetofauna presence in plots during both

seasons (R = 0.106, Table 3.5). Significant predictors ofherpetofaunal abundance in the dry

season included soil humidity, wet litter mass, number of logs and hill category (Table 3.5). Wet

litter mass, soil humidity, number of logs, and shrub cover increase in a plot, so does the

likelihood that the plots will contain herpetofauna (Table 3.5). It also indicates that hill category

is a significant predictor ofherpetofaunal presence. Of these, litter mass, hill category, and soil

humidity are the best predictors, and the overall model successful classified 80% of the plots

sampled.

When forest types were analyzed separately for wet season data, there were no

significant predictors ofherpetofaunal presence (Table 3.5). During the dry season tree number

and soil pH were both negatively related to herpetofaunal presence in unlogged forest, and

overall the model successfully classified the presence or absence of herpetofauna in plots in 76%

of the 50 plots. In logged forest, significant predictors included both low ground and shrub

vegetation (Table 3.5). Amphibians were more likely to be found in plots with greater amounts

of these vegetation types, with low vegetation being the more important of the two, and overall

the model successfully classified 80% of the plots. In the pine plantation in the dry season, both

canopy cover and number of logs were significant positive predictors of herpetofaunal presence

in plots and overall the model successfully classified the presence or absence of animals in 79%

of the 50 plots sampled.

In all three forest types, the abundance of adults and juveniles of the common species

declined during the dry season, with the exception of B. kisoloensis juveniles, which were

captured only during the dry season (Table 3.6). Across all sites and excluding B. kisoloensis,

the adult abundance of the five most common species decreased by an average of 30% during the










dry season, while juveniles decreased by an average of 70%. A similar pattern was observed

within areas; the proportional decline in juveniles during the dry season in each site was always

higher than the decline of adults of the common species. The logged forest had the greatest

decline in the overall abundance of animals during the dry season. Phrynobatrachus graueri

juveniles were very abundant at this site during the wet season, but decreased by 82% during the

dry season. The abundance of the most common species decreased by 42% in the pines, largely

due to the fewer S. schubotzi juveniles captured, and the unlogged forest appeared least affected,

with a 26% decline in common species.



Feeding Ecology

Of the 46 P. graueri stomachs examined (15 males, mean SVL 20 mm; 17 females, mean

SVL 25 mm; 12 subadults, mean SVL 15 mm) 25 (54%) were empty (Table 3.7). The stomach

contents of this species consisted of 22 invertebrate prey categories (Fig 3.1). Among frogs that

contained prey, the average stomach contents consisted of 7.3 invertebrates (range = 1 to 44), and

the average prey item volume was 14 mm3. Springtails (Order Collembola) were the prey group

with the highest Relative Importance Index (RI = 34). They were found in 13 of the stomachs

which contained prey (%F = 62%) and accounted for 64% of the prey items and comprised 34%

of the total prey volume. Beetles (RI = 15; %F = 30; /oN = 5, %V = 36) and ants (RI = 14; %F =

52; /oN = 12; %V = 3) were the second and third most important prey taxa.

Of the 17 B. funereus stomachs examined (8 males, mean SVL 51 mm; 8 females, mean

SVL 62 mm; 1 subadult, SVL 30 mm) 15 (88%) contained prey. The average number of prey in

those 15 stomachs was 8.9 (range = 2 to 27), and the average individual prey volume was 221

mm3 (Table 3.7). Of the 12 invertebrate prey taxa, ants were clearly the most important (RI = 36;

Fig. 3.1), being present in 93% of the stomachs and accounting for 69% of the prey items and

25% of the prey volume. The index for the Order Coleoptera (RI = 32) was similar to that of










ants. Beetles accounted for 17% of the prey items and 36% of the prey volume. All other prey

taxa had RI indices less than seven.

Of the 17 B. kisoloensis stomachs examined (8 male, mean SVL 57 mm; 7 female, mean

SVL 71 mm; 2 subadult, mean SVL 40 mm) 13 (76%) contained prey. Five of those 13 were

from Bwindi specimens, and at the level of the prey categories examined here, there were no

differences in the common prey taxa observed between sites. There was an average of 18 prey

per full stomach (range = 1 to 51), and the average individual prey size was 131 mm3 (Table 3.7).

Again, ants (RI = 36) and beetles (combined RI = 34) were the most important of the nine

invertebrate prey taxa. Ants were found in 100% of the stomachs that contained prey and they

accounted for 69% of the prey items and 31% of the prey volume. Coleopterans were found in

70% of the stomachs and accounted for 20% of the prey items, and comprised 47% of the prey

volume.

Of the 19 S. schubotzi stomachs examined (5 male, mean SVL 21 mm; 7 female, mean

SVL 20mm; 7 subadult, mean SVL 16 mm) 14 (74%) contained prey representing nine prey taxa.

Three of these 14 were from Bwindi specimens, and both Kibale and Bwindi specimens appear

to prey predominantly upon the same taxa. Stomachs contained an average of 15 prey items

(max. 47) with the average prey volume being 2 mm3. Small ants (< 3 mm length) were found in

all stomachs with prey items and accounted for 80% of the prey items and 58% of the prey

volume resulting in an importance index of 49 (Fig. 3.1). This was the highest index value

observed for any prey taxa for any frog species. Beetles were of second greatest importance

(combined RI = 19) in S. schubotzi diet, accounting for 65% of prey items and 16% of prey

volume.

Only 16 of the 49 L. kivuensis stomachs (36 male, mean SVL 33 mm; 9 female, mean

SVL 45 mm; 4 subadult, mean SVL 26) examined contained prey items (33%). Six of the 16 full

stomachs were from Bwindi specimens. At both sites the three most common prey taxa were










taken in very similar proportions. Leptopelis kivuensis had the lowest number of prey items per

stomach with contents (1.4 prey/frog; range = 1 to 3) and the largest average prey item volume

(489 mm3). The most common prey items were orthopterans (RI = 40; %F = 38; /oN = 27; %V =

62), lepidopteran larvae (RI = 21; %F = 25; /oN = 18, %V = 24), and spiders (RI = 18; %F = 31;

/oN = 23; %V = 3). Many of the prey species found in the diet ofL. kivuensis were associated

with understory vegetation (e.g., kaytidids, lepidopteran larvae) while many of the terrestrial

species were absent from the diet.

Eight male (mean SVL 23 mm) and six female (mean SVL 26 mm) H. lateralis were

examined, of which ten had prey remains in their stomachs. Those ten stomachs contained an

average of four prey individuals (range = 1 to 7) with an average individual prey volume of 34

mm3 (Table 3.7). Prey from 12 taxa were represented in the diet (Fig. 3.1), with dipterans (RI =

24; %F = 80; /oN = 20; %V = 19), hemipterans (RI = 17; %F = 40; /oN = 10; %V = 37%), and

ants (RI = 13; %F = 40; /oN = 20; %V = 5) appearing most prominently (Fig. 1.2). The presence

of invertebrates that are associated exclusively with the forest floor (e.g., Collembola) indicates

that this species does forage in the litter.

Bufo kisoloensis, B. funereus, and S. schubotzi all show a high degree of similarity in

their diets, as all three prey predominately upon ants (Table 3.8). However, when ants are

divided into two size categories (> and < 5 mm length), this similarity is much less pronounced,

particularly between S. schubotzi, which eats small ants, and the two toads, which feed on larger

ants. The diet of B. kisoloensis was remarkably similar to that of B. funereus; and many of the ant

and beetle species that these two toads preyed upon appeared to be the same. However, while B.

funereus preyed exclusively upon larger ants, the diet of B. kisoloensis included ants of all sizes

(Table 3.8). The lack of similarity between L. kivuensis and all the other species, even the other

hyperoliid H. lateralis, again suggests that this species may not be feeding in the litter layer.










Fruiting Fig Trees and Litter Herpetofauna Distribution

The results of sampling under Ficus natalensis indicate that arthropod density increased

during fruiting periods (Table 3.9), however, this increase was not entirely limited to the area

directly underneath the tree canopy. With regard to herpetofauna, six anuran species (B.

kisoloensis, B. funereus, P. graueri, H. lateralis, L. kivuensis, S. schubotzi) totaling 16

individuals were captured during plot sampling at Ficus natalensis. Herpetofaunal density was

significantly higher in plots under the canopy, but there was no significant effect of fruiting

(Table 3.10). Frogs were found more frequently under F. natalensis trees than 3 m out from the

canopy during, as well as after, a fruiting event. Litter depth was greater under the canopy and

greater during the fruiting period, however the interaction of these two terms was not significant

(Table 3.10).




Discussion


Habitat Correlates of Herpetofaunal Abundance

Several studies in the American tropics have found that seasonality is important in

structuring leaf-litter herpetofaunas through seasonal changes in prey abundance (Toft, 1980a;

Lieberman and Dock, 1982), litter depth (Scott, 1976), and juvenile recruitment (Allmon, 1991).

The results of this study suggest that seasonal rainfall and topography interact to create moisture

gradients that may be important in determining the patterns of abundance and distribution at

Kibale.

For all study areas combined, there was a significant decrease in soil humidity and the

number of amphibians and reptiles sampled during the dry season. During both wet and dry

seasons, soil humidity was a significant predictor of herpetofaunal presence in plots. Various

aspects of cover (litter mass, number of logs, ground vegetation, shrub vegetation) become more










important during the dry season, presumably due to greater desiccation risk. Upper slopes and

hill tops were always drier than lower slopes and valleys, but this difference was greater during

the dry season. During the wet season it was not uncommon to observe amphibians in hill-top

plots. However, none were sampled from the upland habitats during the dry season. Therefore it

appears that lack of moisture may limit the use of these habitats during the dry months. This may

reflect a seasonal shift in habitat use similar to that observed by Toft (1980a), who found

relatively fewer amphibians used ridge tops compared to stream bottoms during the dry season at

a relatively dry site in Panama.

In contrast to this study, several others have observed higher litter herpetofaunal

densities in the dry season (Toft, 1980a; Lieberman, 1986; Allmon, 1991). Both Toft (1980a)

and Lieberman and Dock, (1982) found that prey densities (litter arthropods) were highest in the

dry season when herpetofaunal abundance was greatest. Nummelin (1989) concluded that in

Kibale forest-floor arthropod numbers peak in the late wet season, which would initially suggest

that arthropod and litter herpetofaunal densities may be positively correlated. However, a closer

look at Nummelin's data for unlogged, logged, and pine plantation sites during the short rains

and subsequent dry season reveals a more complicated scenario. The site with the highest

arthropod abundance (pines) is not the site with the highest herpetofaunal densities (logged).

Nor does arthropod abundance show a similar pattern of increase or decrease among all sites

during the shift from short wet to dry seasons, as herpetofaunal densities do (all decrease).

However, one similarity between arthropod and herpetofaunal numbers bears mentioning. The

unlogged site showed the least seasonal variation in both arthropod abundance and herpetofaunal

abundance, while the logged site showed the greatest variation. While this is suggestive that

prey abundance may contribute to the pattern of abundance observed in litter herpetofaunas, this

is only speculation based on arthropod densities recorded more than ten years earlier. Arthropod

densities at Kibale exhibit considerable year to year variation (Nummelin, 1989), and










simultaneous sampling of arthropod and herpetofaunal abundance across a range of habitats and

across seasons will be necessary to test for an effect of prey abundance on herpetofaunal

numbers.

Allmon (1991) concluded that juvenile recruitment was the most important factor

responsible for seasonal variation in species abundance; however his study did not sample

arthropods. The short duration of my study prohibits careful evaluation of the importance of

recruitment. All of the amphibian species sampled except one have aquatic tadpoles and breed

primarily during the two wet seasons (Vonesh, unpubl. data). However, juvenile abundance for

all species declined, not increased, from the wet to dry season, with the exception of Bufo

kisoloensis. B. kisoloensis juveniles were captured only during the dry season, and therefore may

represent new recruitment. It is possible that the large number of juveniles observed for most

species during the wet season may represent late metamorphs from the October-December rainy

season, which is longer and is likely to be a period of greater anuran reproductive activity.

Although the number of adult animals decreased, the percentage decrease was much higher for

juveniles of the common species.This suggests either higher mortality among juveniles over the

study period, or some aspect of juvenile behavior (e.g., aggregating in valley bottom pools,

retreating to burrows) that biases against their capture in the dry season. A longer study would

be necessary in order to understand the importance of seasonal recruitment in structuring these

communities.

In summary, it appears that seasonal variation in soil moisture and the interaction of

moisture and topography, are the most important ecological variables in structuring the Kibale

litter herpetofauna. The lower number of dry season captures may be due to clumping of animals

in valley bottoms around water bodies, dry season mortality, and/or retreating into burrows or

other refuges. All three responses suggest moisture as the limiting resource. Moisture may be

particularly important, since Kibale typically receives less than 1600 mm of rain per year, one of










the lowest annual rainfalls reported for any tropical litter herpetofaunal study. Unlike a number

of other studies, arthropod abundance, based on Nummelin (1989), does not seem to account for

the seasonal and spatial variation in herpetofauna abundance that was observed, although this

conclusion must be entertained with caution. The role of juvenile recruitment in shaping the

patterns of abundance and diversity can not be determined unequivocally, but over the short

period of this study it played a minor role, through the dry season recruitment of B. kisoloensis

juveniles.



Comparison of Disturbed and Undisturbed Forests

Several studies have examined how logging and conversion into tree plantations

influences native herpetofaunas (Miyata, 1980; Lieberman, 1986; Heinen, 1992; Aukland et al.,

1997). Miyata (1980) found increased herpetofaunal densities in cacao and rubber plantations in

Ecuador relative to primary forest, and Lieberman (1986) and Heinen (1992) found a similar

pattern between primary forest and cacao plantations in Costa Rica. Heinen's (1992) study

further suggested that the high abundance observed in recently disturbed sites decreases with

time. In contrast to density trends, Heinen (1992) found that undisturbed forests in Costa Rica

had higher diversity (H') and greater equitability (J') than young or older cacao plantations.

Aukland et al., (1997) found a very different situation in her studies in western Uganda, with

higher litter frog densities and lower diversity in unlogged forest than forest that had been

selectively logged in the last 5 yr. The selectively logged forest studied at Kibale was logged

nearly 30 years ago, and the tree plantations are of a similar age with a dense, predominately

native flora regenerating underneath the pines. My results indicate that herpetofaunal

communities respond differently to different forest management strategies. Compared to the

unlogged forest, the plantation had higher densities as well as greater diversity and evenness, and

the logged forest had greater densities but lower diversity and evenness.










The Nyakatojo pine plantation was established on derelict land dominated by elephant

grass (Pennistum purpureum) adjacent to unlogged forest. This grassland is believed to have

been formed through human clearing and maintained by fires which inhibit the establishment of

native trees (Kasenene, 1987). Several studies (Lugo, 1992; Chapman and Chapman, 1996;

Zanne, 1997) suggest that exotic tree plantations can be used to facilitate forest regeneration in

cases where natural succession is very slow or arrested. Grasslands not converted to plantations

in the 1960's are still dominated by elephant grass today, while many native trees and shrubs are

reestablishing in the pine plantation understory (Chapman and Chapman, 1996; Zanne, 1997).

However, the ability of exotic tree plantations to aid in the restoration of native herpetofaunal

communities is unclear. On one had, the pine plantation site at Kibale is home to the most

diverse herpetofauna of the three sites studied. In contrast to Heinen (1992) who concluded that

primary forests were an important refuge for rare species, the pine plantation, not the adjacent

unlogged forest, harbored the greatest number of rare (captured 1-2 times) litter lizards and frogs

(5 spp. vs 2 spp.). The greater number of rare species, higher diversity, and greater equitability

all suggest that the pines' fauna is successionally mature. However, the pine fauna has a very

low similarity to the unlogged forest, while the logged forest and unlogged forests show a high

degree of overlap. A comparison of Heinen (plantations, 1992) and Aukland et al., (logged

forest, 1997) suggest that logged forests are generally more similar to undisturbed native forests

than plantations. In fact, though abundance decreases and diversity and evenness increase with

the age of Heinen's (1992) cacao plantations, the older plantation is not any more similar to the

original forest than the new plantation.

This has several implications from the perspective of faunal restoration as discussed in

Heinen (1992). First, it may be inappropriate to expect abandoned plantations, be they pine or

cacao, to revert to a state similar to that of the original forest, despite the availability of nearby

source forests. If that is the case, what measuring stick should be used to evaluate a faunas'










relative degree of "restoration", particularly if indices of community richness, diversity, and

equitability in the plantations may, in some cases, exceed that of the original forest. The

implications for selective logging are different. While diversity and evenness may decrease (this

study) or increase (Aukland et al.., 1997) with selective logging, the similarity of the disturbed

fauna to that of the original may be high. In the case of selective logging then, it may be more

realistic to discuss restoration of the fauna to a state similar to that of undisturbed forests.



Feeding Ecology

Several studies in the past (Toft, 1980a,b; Lieberman, 1986) have examined the

questions of how the feeding ecology of litter species relate to their ability to coexist. Their

results suggest that both diet and activity period may be important in determining the number of

species that can share a habitat. Litter species can be broadly categorized as either diurnal or

nocturnal, and are either active foragers, which prey upon large numbers of hard bodied

arthropods such as ants or beetles which are slow to escape, or are sit-and-wait predators which

tend to prey upon fewer, larger, often soft-bodied prey species (Toft, 1981). At Kibale, the six

most abundant anuran species, B. funereus, B. kisoloensis, S. schubotzi, P. graueri, L. kivuensis,

and H. lateralis, accounted for 80% of the total number of reptiles and amphibians captured. Of

these the first four are primarily diurnal while the two hyperoliid species are nocturnal.

A closer look at the diurnal species suggests that B. funereus, B. kisoloensis, and S.

schubotzi are active foragers, as indicated by the high proportion of these species that had prey in

their stomachs at the time of examination and the relatively high number of prey items per

stomach. All three of these species prey primarily on hard-bodied arthropods, namely ants of

different species and beetles. Without prey availability data, I am unable to say whether these

three species are highly selective of ants as prey or that ants are simply the most frequently

encountered prey species. Regardless, ants are clearly the most important prey taxa. This is not










surprising for the two Bufo species; Bufo species around the world typically eat high proportions

of ants (Inger and Marx, 1961; Toft, 1980b, 1981). What is less clear is how two species coexist

with such similar diets. The Morisita's similarity index for these two species is 0.69, the highest

of any species pair, and much higher than the average for all six species studied, 0.33. One

subtle difference in diet is that B. kisoloensis preys upon a wider range of ants size classes than

B. funereus. Thirty-five percent of the ants in B. kisoloensis' diet are smaller than 5 mm in

length, while only 3% of the ants in B. funereus' diet fall in the smaller size category. Plot

results and personal observations also suggest that these two toads may use different parts of the

forest. Of the nine B. kisoloensis captured, 7 (78%) were found in valley bottom plots, and these

were all juveniles, possibly dispersing from larval habitats. No B. kisoloensis were captured on

hill-tops. Only 38% of the B. funereus were captured in bottom plots, and 28%, including

several adults, were found on hill tops. These data support personal observations that suggest B.

kisoloensis adults are most common in dense, often seasonally flooded, vegetation in valley

bottoms along streams, while B. funereus is more generally distributed throughout the forest.

There may also be some temporal separation in the seasonal abundance of these species, B.

kisoloensis were only captured in plots during the dry season, while B. funereus was twice as

abundant in plots during the wet season. This is similar to Toft's (1980b) findings with the

species Eleutherodactylus toftae and Andenomera andreae, two non-ant specialists of similar

size. E. toftae was only observed in the dry season, when A. andreae was not observed.

S. schubotzi had the highest Relative Importance Index for ants of any species in this

study. Interestingly, in a phylogenetic analysis based on osteological and myological characters,

Ford (1989) suggested that the family Artholeptidae, which includes the genus Schoutedenella, is

sister group of the South American Dendrobatidae, and that the vicariance that gave rise to these

two groups probably coincides with the separation of South America from Africa. The

dendrobatids are also ant specialists, and derive potent defensive toxins from the ants they eat.










To the best of my knowledge, the presence or absence of defensive toxins in African

arthroleptids has not been examined, but toxins in these species are probably not well developed

given that they are usually cryptically colored.

The diet of S. schubotzi was most similar to that of the similarly sized P. graueri, the

fourth diural species. Ants, beetles, and collembolans are the primary prey of both these

species. However, the proportion in which they prey upon them is quite different.

Schoutedenella, as mentioned, prey primarily upon small ants with beetles being the prey with

the second highest RI index, while the important prey for P. graueri, based on RI indices, are

collembolans, beetles, and ants, in that order. The fact that P. graueri preys primarily upon

collembolans is rather unique. Lieberman and Dock (1982) found that almost no terrestrial

anurans preyed upon Collembola, despite the fact that they were the third most abundant

arthropod taxa in the litter. She concluded that they were probably very unpalatable or preferred

by some non-anuran litter insectivore. Also, none of the species in Toft's studies (1980a, 1980b)

preyed upon collembolans. The different body morphotypes of the collembolans in P. graueri's

diet suggest that it is feeding on Collembola species which dwell on the litter surface, as well as

those that dwell within or under the litter.

In contrast to the diural species, the two nocturnal hyperoliids prey primarily upon soft-

bodied arthropods. Leptopelis kivuensis appears to fit the model of the sit-and-wait predator, as

indicated by the low number of non-empty stomachs (%33) and the large mean prey size. The

diet of this species bears little resemblance to those of the other species studied, preying

primarily upon orthopterans and lepidopteran larvae, which may or may not have been in the

litter. Several personal observations of this species feeding suggest that it is a nocturnal forager,

and probably captures most of its prey in low shrubs. It is possible that this species uses the

moist litter layer only as a daytime refuge, and therefore should not really be considered part of

the litter assemblage in terms of feeding ecology.










Hyperolius lateralis also preys primarily upon soft-bodied insects, in this case dipterans

and hemipterans. The proportion of full stomachs suggests that this species may forage more

actively than L. kivuensis. In addition, the presence of ants and Collembola in the diet indicate

that this species does forage in the litter. In fact, the diet is much more similar to that of P.

graueri than to L. kivuensis.

My results differ in some respects from those of Toft (1982), the only other study to

examine the feeding ecology of an African leaf-litter anuran fauna. She examined the diets of the

four most abundant litter frogs in Makokou, Gabon. These four included Arthroleptis sylvatica,

Cardioglossa leucomystax, both arthrolepitids, Bufo camerunensis, and the ranid,

Phrynobatrachus batesi. The bufonid, as in my study, preyed primarily upon ants, but the other

three species all preyed primarily upon isopterans (termites), whereas none of the species at

Kibale preyed upon termites. Also, because of the preponderance of isopteran in the diets of

these species, Toft (1982) observed much higher dietary similarity indices than were observed in

the Kibale study. Isopteran abundance tends to be highly seasonal, and it is possible that the

short duration of both this and Toft's (1982) study over or de-emphasize the importance of this

taxa in the diet of African litter frogs. Toft (1982) also remarked about the absence of sit-and-

wait, soft-bodied arthropod specialists in the diurnal litter frog fauna at Makokou. She found that

38% of the diurnal frog fauna specialized in non-ant prey. In this regard, Kibale resembles

Makokou; all four common diurnal frogs eat hard-bodied arthropods (ants and Collembola).



Frogs and Figs

While seasonal litter arthropod abundance has been shown, in several tropical forests, to

be positively correlated with herpetofaunal abundance, few studies (Guyer, 1988) have examined

this on a finer scale. One question of interest is how does forest fruit-fall influence small scale

changes in litter arthropod abundance, and do litter amphibians and reptiles respond to such local










scale changes, if they do exist? Inger, (1980a) hypothesized that the low litter herpetofaunal

densities observed in the forest of Southeast Asia were due to the mast fruiting cycle of the

dominant dipterocarp trees. He suggested that during non-mast years, the litter arthropod

numbers decrease and are unable to support large amphibian and insectivorous reptile

populations. However, to my knowledge, no study has demonstrated that forest-floor arthropods

or herpetofauna respond in any way to forest fruit-fall. Unlike the forests of tropical Asia,

Kibale is not characterized by dipterocarp trees or mast fruiting events, however litter

herpetofauna might be expected to respond on a local scale to increases in availability, if only to

decrease foraging effort. My results suggest that arthropods do increase under fruiting fig trees,

and that this increase extends beyond the canopy edge. It does not appear that litter herpetofauna

respond to this change, as herpetofaunal abundance is greater under fig trees during both fruiting

and post-fruiting samples. Therefore it is possible that frogs are simply attracted to some other

microhabitat characteristic associated with tree bases, such as deeper litter.



Regional Comparisons

The number of mid-elevation forest litter herpetofaunal studies is limited, relative to

studies of lowland faunas, but comparison with these few is of interest. Studies from Costa Rica,

the Philippines, and Africa indicate that the upland forests of these three areas support a similar

number of leaf litter species (Table 3.11). San Vito, Costa Rica, appears to be an exceptional

case, supporting a tremendously rich fauna (Scott, 1976), but results from a study in progress at

nearby Fila Cruces (1300 m) have found a more modest 18 leaf-litter species assemblage

(Schlaepfer, 1998). Animal densities vary widely within and across regions. The widely cited

trend observed in lowland faunas, where densities are much higher in Central American relative

to Southeast Asia, is not obvious in the mid-elevation faunas.










While the species richness of the Kibale litter fauna is comparable to that from other

areas, the combined wet and dry season animals density, particularly in the unlogged forest is

very low, similar to that of the depauperate lowland forests of Borneo (Lloyd et al., 1968) and

Thailand (Inger and Colwell, 1977). The litter herpetofauna of lowland Cameroon (Scott, 1982),

which shares two frog species, and four genera with Kibale, has a fauna of similar richness, but

much higher densities than either Kibale or Budongo. This contradicts the trend reported by

Scott (1976) that diversity decreases, and abundance increases with increasing elevation.



Limitations

As Heinen (1992) pointed out, there are several limitations to the generalities that can be

drawn from the available studies of litter herpetofaunas, including this one. First, as the areas

being compared may be adjacent to one another they may not be truly independent. For example,

the unlogged site in this study may serve as a source for the pines litter fauna, and studies of

isolated pine plantations may yield results different to those presented here. A recent study of

floral regeneration in Kibale pine plantations found that the indigenous tree species regenerating

underneath the pines were more similar to natural forest species in pine plantations surrounded

by forest compared with isolated plantations (Zanne, 1998). Another problem with this and other

studies is pseudoreplication (Hurlburt, 1984), taking multiple samples from a single

representative forest type, rather than sampling in forest type replicates. There are a number of

studies at Kibale that have compared the same three areas examined here, and while

pseudoreplication may limit generalization, these studies none-the-less provide valuable

information and suggest directions for more rigorous, manipulative studies. Furthermore, the

patterns observed in comparing the unlogged and logged forests are very similar to those

observed by Heinen (1992) and Lieberman (1986) in Costa Rica, Miyata (1980) in Ecuador, and










Inger (1980b) in Malaysia. The similarity of the results of these studies argues for their validity

and general applicability.




Conclusions


A total of 18 species were captured in the leaf-litter layer of the forest floor at Kibale, a

number similar to that observed in other studies of tropical litter herpetofaunas from mid-

elevation forests. The density of animals at Kibale, however, was much lower than that reported

from other sites.

During the wet season, the selectively logged forest was characterized by a greater

abundance of herpetofauna, but lower overall diversity and species equitability compared with

the unlogged forest. This pattern has been observed in a number of studies that have examined

litter herpetofaunas from disturbed and undisturbed sites. Interestingly, the exotic pine

plantation exhibited the highest species diversity among all three forests. The faunal

composition of the pine plantation was very different from that of the native forests, being

characterized by the dominance of species adapted to more drier conditions, namely reptiles and

direct developing frogs.

The most common diural anuran species in the litter are active foragers of small hard-

bodied prey such as ants. There are no sit-and-wait predators which prey on larger soft-bodied

prey. This absence of this guild is an interesting contrast with Neotropical herpetofaunas, in

which predators of soft-bodied arthropods make up a large proportion of the litter anuran fauna.

Two nocturnal treefrogs that are often captured in the litter appear to be sit-and-wait predators,

but only one of these actively feeding in the litter, the other appears to only use it as a diural

refuge.










Prey abundance was found to increase significantly beneath fruiting fig trees. Frog

abundance was also higher under fig tree canopies then away from them, but this was the case

during both fruiting and post-fruiting sampling periods. Therefore it seems likely that frogs are

attracted to figs trees, but not necessarily (or only) due to increased prey availability, but perhaps

due to the deeper leaf litter found under the figs.

All three forest types showed a decrease in the number of animals captured during the

dry season, in contrast with most Neotropical litter herpetofaunas which reach their peak

densities in the dry season. This combined with the absence of animals from the drier hilltop and

upper slope forest habitats suggests that moisture is one of the most important factors in

determining local patterns in herpetofauna abundance. This is not all that surprising, as Kibale is

relatively dry compared to most tropical forest where litter herpetofaunas have been studied and

many of the species present at Kibale are believed to have originated from the wetter lowland

forests of eastern Congo-Zaire.











TABLE 3.1. Reptiles and amphibians of the forest floor leaf-litter layer, Kanyawara, Kibale
National Park, Uganda, based on 15 mo of collecting and Pitman (1974). Eighteen species were
sampled during the study. The jackknife species richness estimate ( 95% confidence interval)
based on all 300 plots is 25.0 + 5.0 species. (Q indicates species found primarily in upland
habitats; # indicates species restricted to streamside habitats; ? indicates species found at least
occasionally in the litter, but the total proportion of their life that is spent in the litter layer is
unknown; indicates fossorial species; t indicates species expected to be in Kibale forest, but
no record as yet; X indicates those species that were successfully sampled using plots)


Order Anura
X Bufofunereus
X Bufo kisoloensis
X Schoutedenella schubotzi f
Phrynobatrachus auritus #
Phrynobatrachus dendrobates #
X Phrynobatrachus graueri
X Phrynobatrachus parvulus
Phrynobatrachus versicolor #
X Rana angolensis #
X Phlyctimantis verrucosus ?
X Leptopelis christyi ?
X Leptopelis kivuensis ?
X Hyperolius lateralis ?


Order Sauria
X Cnemaspis quattuorseriatus I ?
X Rhampoleon boulengeri
X Adolfus africanus ?
X Adolfus vauereselli
X Leptosiaphos aloysiisabaudiae 8
Lygosomafernandi


Order Serpentes
Typhlops punctatus *
Causus lichtenstienii
Bitis nasicornis
X Bothropthalamus lineatus
Lamprophis olivacea
Mehelya poensis t
Mehelya stenophthalmus
X Geodipsas depressiceps
Polemon christyi *
X Dasypeltis atra
Lycophidion ornatum
Bitis gabonica t
Atractaspis is I gI,,, r" t







67




TABLE 3.2. Numbers and species of amphibians and reptiles captured in the leaf-litter of
pristine, logged, and pine forest during the wet and dry seasons, Kibale National Park, Uganda.
Fifty plots of 25 m2 were sampled in each forest type during each season.


Forest types
Pristine Logged Pine
Family and species wet dry wet dry wet dry total
Bufonidae
Bufofunereus 4 4 7 2 5 1 23


B. kisoloensis
Ranidae
Schoutedenella schubotzi
Rana angolensis
Phrynobatrachus graueri
P. parvulus
Hyperoliidae
Phlyctimantis verrucosus
Leptopelis kivuensis
L. christyi
Hyperolius lateralis
Gekkonidae
Cnemaspis quattuorseriata
Lacertidae
Adolfus africanus
Adolfus vauereselli
Chamaelonidae
Rhampoleon boulengeri
Scincidae
Leptosiaphos aloysiisabaudiae
Colubridae
Bothropthalmus lineatus
Geodipsas depressiceps
Dasypeltis atra


0 0 0 0 10 4 14


0 0 0 0 0 1 1

0 0 0 0 1 0 1











TABLE 3.3. Density, richness, evenness, diversity, and similarity of the leaf-litter herpetofauna
calculated for pristine, logged, and pine plantation forest types at Kibale National Park, Uganda.


No. individuals


Density (animals/100 m2)

Richness (no. spp.)

Diversity (H')


Unlogged
wet dry total wet
28 20 48 73


Logged Pine plantation
dry total wet dry total
33 106 38 19 57


2.24 1.60 1.92 5.84 2.64 4.24 3.04 1.52 2.28

7 7 9 9 9 12 10 7 12

0.60 0.78 0.77 0.49 0.80 0.63 0.77 0.76 0.86


0.41 0.53 0.52


Evenness (J')

Similarity (C)


0.33 0.54


0.94


0.43 0.52 0.51 0.58


0.24


0.42












TABLE 3.4. Differences among the three forest types in environmental variables measures
within each plot during the wet and dry seasons, expressed as the mean and 1 SD, and the P-
values associated with each as determined by the Kruskal-Wallis ANOVA. Similar superscripts
for pairs of values signify that that variable was not significantly different between those two
forests. (* indicates a significant difference in that variable between wet and dry season; NS
indicates value does not significantly vary between seasons, # K-W p-values compared against
the Bonferroni adjusted x = 0.004, for 12 tests)


Variable Pristine Logged Pine X2 K-W
P-value#


Wet season:
Slope
Soil humidity
Soil pH
Wet litter mass
Litter depth
Ground
vegetation
Shrub cover
Canopy cover
No. logs/plot
No. trees/plot
No. frogs/plot
No. reptiles/plot

Dry season:
Slope
Soil humidity
Soil pH
Wet litter mass
Litter depth
Ground
vegetation
Shrub cover
Canopy cover
No. logs/plot
No. trees/plot
No. frogs/plot
No. reptiles/plot


14.9 + 5.3
76.6 + 13.7a
6.2 0.5a
0.9 0.5
2.9 + 1.0a
41.9 16.0a

42.3 + 12.7a
93.8 + 1.4a
0.3 + 0.6
1.8 + 1.5
0.6 + 1.0
0.02 + 0.1a


13.1 5.0aNS
57.2 + 20.8**
6.6 + 0.4**
0.4 + 0.2a**
2.3 + 0.9a**
41.5 15.2aNS

37.2 13.4aNS
94.0 1.7aNS
0.3 0.6NS
1.4 + 1.2aNS
0.4 + 0.8NS
0.02 + 0.14NS


8.0 5.5b
71.9 13.7a
6.3 + 0.4a
0.7 0.3b
3.2 0.9b
30.7 13.1b

45.7 13.4a
93.5 + 2.3a
0.4 + 0.8
1.5 1.2
1.5 + 4.2
0.02 0.1a


7.8 + 5.4bNS
60.4 + 14.8**
6.7 0.3**
0.5 0.2b**
2.6 1.1a**
28.1 + 10.1bNS

44.7 12.0b
94.0 1.8aNS
0.2 0.4NS
0.9 1.0b**
0.7 1.8**
0.06 + 0.24NS


8.9 5.7b
72.7 + 17.05a
6.1 + 0.9a
1.3 + 0.6c
5.1 + 1.0c
49.3 + 19.2a

34.2 + 14.3b
88.8 2.4b
0.5 1.0
1.9 1.2
0.5 + 0.7
0.2 0.5b


9.5 + 4.7bNS
52.3 + 12.6**
6.7 + 0.3**
0.9 + 0.3c**
4.3 0.9b**
36.3 + 13.5a**

32.4 + 15.5aNS
89.5 + 2.7bNS
0.4 + 0.7NS
2.0 + 1.2cNS
0.3 + 0.5NS
0.1 + 0.3NS


34.8
5.2
5.2
40.3
81.9
27.8

17.2
74.2
1.3
2.0
5.9
12.6


24.6
9.2
1.4
57.1
71.6
23.7

18.3
73.2
4.1
22.4
0.04
2.8


<0.001*
0.073
0.075
<0.001*
<0.001*
<0.001*

<0.001*
<0.001*
0.525
0.361
0.052
0.002*


<0.001*
0.01
0.505
<0.001*
<0.001*
<0.001*

<0.001*
<0.001*
0.13
<0.001*
0.978
0.24












TABLE 3.5. Results of stepwise logistic regression of the presence or absence of amphibians
and reptiles in leaf litter plots (Kibale National Park, Uganda) predicted by habitat variables.
Independent variables included: Slope, soil humidity, soil pH, wet litter mass, litter depth, logs,
% low vegetation cover, % medium vegetation cover, canopy cover, number of trees, hill
category. Hill categories were: 1 = valley bottom, 2 = lower slope, 3 = upper slope, 4 = hilltop.


Habitat Season Classification R -2 log likelihood Predictors

table % correct x2 (P value)

All 3 habitats wet 59.1 0.106 4.71 (0.03) soil humidity

dry 80.0 0.162 21.41 (0.0003) hill category


0.180

0.128

0.129

0.105


Logged forest






Pine plantation


- 0.292

-0.165




0.311

0.237




0.176

0.266


30.55 (<0.0001)

35.34 (<0.0001)

41.07 (<0.0001)

45.20 (<0.0001)


8.82 (0.003)

12.95 (0.002)




8.26 (0.004)

15.55 (0.0004)




6.62 (0.01)

11.59 (0.003)


wet litter mass

number of logs

soil humidity

med. vegetation

none




soil pH

no. trees

none

low vegetation

med. vegetation

none

canopy

no. logs


Unlogged

forest


76.0






80.0






79.0












TABLE 3.6. Numbers of adults and juveniles of the five most common leaf-litter herpetofauna
in the three forest types during the wet and dry seasons. Adults identified by SVL as follows:
Phrynobatrachus graueri > 19.0 mm; Bufofunereus > 35.0 mm, Bufo kisoloensis > 40.0 mm;
Schoutedenella schubotzi > 19.0 mm; Leptopelis kivuensis > 35.0 mm.


Species
P. graueri-adult

P. graueri-juv.

B. funereus-adult

B. funereus-juv.

B. kisoloensis-adult

B. kisoloensis-juv.

S. schubotzi-adult

S. schubotzi-juv.

L. kivuensis-adult

L. kivuensis-juv.


Unlogged
wet dry
5 5

9 1

4 2

0 2

0 0

0 5

0 1

2 0

0 0

3 1


Logged
wet dry
13 5

39 7

4 2

4 0

0 0

0 4

0 0

0 0

0 0

7 5


Pines
wet dry
0 3

1 0

1 0

4 2

0 0

0 3

8 4

6 0

0 0

1 0


Totals
wet dry
18 13

49 8

9 4

8 4

0 0

0 8

8 5

8 0

0 0

11 6


totals
31

57

13

12

0

8

12

8

0

17


Totals 23 17 67 21 21 12 111 50 161











TABLE 3.7. Results from the stomach content analysis of the six most common leaf-litter
herpetofauna in Kibale National Park. SVL: mean length of all individuals of that species for
which stomach content data was collected. N: number of stomachs examined. No. Full: number
of stomachs examined which held prey remains. No. Prey: total number of prey individuals
recovered. No. Items: average number of prey individuals per stomach. Prey Volume: mean
prey size expressed in mm3




Species SVL (mm) N No. Full No. Prey No. Items Prey Volume (mm3)

P. graueri 20 46 21 154 7.0 14

B.funereus 55 17 15 134 9.0 221

B. kisoloensis 61 17 13 234 18.0 131

S. schubotzi 19 19 14 215 15.0 2.0

L. kivuensis 35 49 16 22 1.0 489

H. lateralis 24 14 10 40 4.0 34.0











TABLE 3.8. Horn's modified Morisita's similarity indices for the six most common litter
anurans. Values in parentheses indicate the similarity of those to species when ants are divided
into a small (< 5 mm length) and large (> 5 mm) categories.


S. schubotzi B. funereus B. kisoloensis L. kivuensis H. lateralis

P. graueri 0.58 0.48 (0.11) 0.48 (0.3) 0.13 0.52

S. schubotzi 0.89 (0.16) 0.88 (0.41) 0.05 0.42

B. funereus -0.91 (0.69) 0.16 0.43 (0.15)

B. kisoloensis 0.21 0.41(0.22)

L. kivuensis 0.20











TABLE 3.9. Results of plot sampling around Ficus natalensis (n = 5) under the canopy and
away from the canopy during the height of the fruiting season and 1 mo later.




Treatment Arthropod densities Herpetofauna density Litter depth
(animals/50cm2) (animals/100m2) (cm)

Fruiting/Under 70.8 27.7 2.1 2.0 4.4 + 1.4

Fruiting/Away 40.2 14.0 0.5 0.7 2.9 + 0.7

Post-fruiting/Under 36.4 + 8.8 1.3 + 2.3 3.0 + 0.3

Post-fruiting/ Away 31.4 + 8.4 0.3 0.6 2.3 + 0.8












TABLE 3.10. Repeated measures analysis of variance for the differences in arthropod
abundance, herpetofauna abundance, and litter depth at Ficus natalensis trees (Kibale National
Park) during and after fruiting, under and away from the canopy. Mauchley's criterion W = 1 in
cases with only two sampling intervals.


df SS F P R2


Source of variation

Arthropods

Between subjects effects

canopy

error

Within subjects effects

time (fruiting)

time canopy

error (date)

Herpetofauna abundance

Between subjects effects

canopy

error

Within subjects effects

time (fruiting)

time canopy

error (time)


4.36






12.06

4.234








19.262






0.344

0.086


0.07






0.008

0.07








0.025






0.574

0.777


29.0

7.0



44.1

15.5

3.6






17.1

17.9



2.6

0.7

61.6


1584.20

363.00



2332.80

819.20

193.50






8.71

9.158



1.352

0.338

31.470










TABLE 3.10.

Source of variation df SS F P R2

Litter depth

Between subjects effects

canopy 1 5.618 6.015 0.04 31.1

error 8 0.934 5.2

Within subjects effects

time (fruiting) 1 5.202 7.437 0.026 28.8

time canopy 1 0.722 1.032 0.339 3.9

error (time) 8 5.596 30.9











TABLE 3.11. Summary of the quantitative plot studies of mid-elevation tropical litter
herpetofaunas and Scott's (1982) study in lowland West Africa. Elevation is given in meters.
For richness the number of species estimated to be in the litter fauna is given first, followed by
the number of those species sampled successfully using the methods referenced in study are
given. Densities are given in animals per 100 m2. (*) frogs only, or (t) includes a few snakes


Location
Philippines
Cuernos de Negros

Cuernos de Negros


Costa Rica
San Vito

Monteverde


Cameroon
Lombe, Lac Tissongo


Uganda
Kibale Unlogged

Kibale, Logged

Budongo, Unlogged


Elevation Richness Density Reference


1350

1450



1200

1500


30



1500

1500

1100


12(8)

4 (4)



27 (13)

15 (5)


13 (8)



19(9)

19(9)

?? (6)


1100 ?? (6)


11.3 Brown and Alcala
(1961)
15.0 Brown and Alcala
(1961)


58.7 Scott (1976)

6.7 Fauth et al. (1989)


9.4 Scott (1982)



1.9t This study

4.2t This study

6.7* Aukland (1997)

5.9* Aukland (1997)


Budongo, Logged






78





50 Phrynobatrachus graueri Schoutendella schubotzi
40
30
20
10
I. 1 1 nn In nn
-3 50
B ufo funereus Bufo kisoloensis
a 40
e 30
C 20
0
o 10

S 50 Leptopelis kivuensis Hyperolius lateralis
4 40
g 30
20
10
0o: in -I -. -=._ I nn I
.~ OO OO< -4U0 Z OO 00 O 0 U4





FIGURE 3.1. Feeding ecology of the six most abundant leaf-litter species in Kibale National Park, Uganda. Relative Importance
indices for 23 prey categories are given for P. graueri, S. schubotzi, B. funereus, B. kisoloensis, L. kivuensis, and H. lateralis
ssi-s s_"'Ct3CC:II -jOSO.E L)S
FIUR .Fedngeolg o hesx mot budntleflite peie n ibl Ntina ar, gnd. eatveIpotac
















CHAPTER 4
CONCLUSIONS




Few studies have examined the herpetofaunas of East African forests. However, the

need for such studies is clear given rapidly expanding human populations, the accelerating loss

and disturbance of tropical forests, and the mounting recent evidence that amphibians are

declining worldwide. The first objective of this study was to survey the amphibians and reptiles

of the Kibale Forest and to determine its biogeographic associations by comparing the

herpetofauna of Kibale to those of other tropical African sites. The second objective of this

study was to examine the ecology of the leaf-litter component of the Kibale herpetofauna in

order to discover which physical and biotic factors are most important in determining local

patterns of herpetofauna abundance within the forest. The most important results of this study

are summarized below.

1) Kibale Forest supports a rich herpetofauna that includes at least 75 species, including

28 frog species, 15 lizard species, and 32 species of snakes. This makes it one of the richest

herpetofaunas in Uganda, with 13 more species than reported from Bwindi-Impenetrable Forest

in southwest Uganda. This richness is derived from Kibale's transitional nature between lowland

and montane forest and the mosaic of forest and grassland habitat within the park. Elements of

both the endemic Central African montane and lowland Congolean herpetofaunas are present at

Kibale, as are forest dependent species and farmbush/moist savanna species.

2) Comparisons with eight other equatorial African sites demonstrates the high degree of

similarity among Kibale and Bwindi in Uganda, and the Virunga National Park in adjacent

Congo-Zaire. These Central African forest herpetofaunas show a fair degree of overlap with the










West African fauna of Korup in Cameroon, but almost no overlap withthe coastal forests of

Kenya and the coastal mountains of Tanzania. These results support those that have focused on

other taxa, such as primates and birds, and demonstrate that the Kibale herpetofauna is best

described as an eastern extension of the Congolean fauna, as had been suggested by Schiotz

(1976), Pitman (1974), and Loveridge (1942a,b) for all the forests of western Uganda.

3) A total of 18 species were captured in the leaf-litter layer of the forest floor at Kibale,

a number similar to that observed in other studies of tropical litter herpetofaunas from mid-

elevation forests. The density of animals at Kibale, however, was much lower than that reported

from other sites.

4) During the wet season, the selectively logged forest was characterized by a greater

abundance of herpetofauna, but lower overall diversity and species equitability compared with

the unlogged forest. This pattern has been observed in a number of studies that have examined

litter herpetofaunas from disturbed and undisturbed sites. Interestingly, the exotic pine

plantation exhibited the highest species diversity among all three forests. The faunal

composition of the pine plantation was very different from that of the native forests, being

characterized by the dominance of species adapted to more drier conditions, namely reptiles and

direct developing frogs.

5) The most common diural anuran species in the litter are active foragers of small

hard-bodied prey such as ants. There are no sit-and-wait predators which prey on larger soft-

bodied prey. This absence of this guild is an interesting contrast with Neotropical herpetofaunas,

in which predators of soft-bodied arthropods make up a large proportion of the litter anuran

fauna. Two nocturnal treefrogs that are often captured in the litter appear to be sit-and-wait

predators, but only one of these actively feeding in the litter; the other appears to only use it as a

diurnal refuge.










6) Prey abundance was found to increase significantly beneath fruiting fig trees. Frog

abundance was also higher under fig tree canopies then away from them, but this was the case

during both fruiting and post-fruiting sampling periods. Therefore it seems likely that frogs are

attracted to figs trees, but not necessarily (or only) due to increased prey availability, but perhaps

due to the deeper leaf litter found under the figs.

7) All three forest types showed a decrease in the number of animals captured during the

dry season, in contrast with most Neotropical litter herpetofaunas which reach their peak

densities in the dry season. This combined with the absence of animals from the drier hilltop and

upper slope forest habitats suggests that moisture is one of the most important factors in

determining local patterns in herpetofauna abundance. This is not all that surprising, as Kibale is

relatively dry compared to most tropical forest where litter herpetofaunas have been studied and

many of the species present at Kibale are believed to have originated from the wetter lowland

forests of eastern Congo-Zaire.















APPENDIX A
AMPHIBIAN SPECIES OF EIGHT TROPICAL AFRICAN LOCALITIES

Amphibian species of tropical Africa based on lists from Korup National Park, Cameroon (KOR;
Lawson, 1993), Virungas National Park, Zaire (VIR; Laurent, 1972), Garamba National Park,
northeastern Zaire (GAR; Inger, 1968), Kibale National Park, western Uganda (KIB; this study),
Bwindi National Park, southeastern Uganda (BWI; Drewes, 1991; Drewes, 1998), East and West
Usambara mountains, Tanzania (USA; Howell, 1993); Uzungwa National Park, Tanzania (UZU;
Howell, 1993), Arabuko-Sokoke Forest Reserve (ARA; Drewes, 1995).

Taxon KOR VIR GAR KIB BWI USA UZU ARA
Gymnophiona,
Scolecomorphidae
Crotaphatrema bornmuelleri X
Scolecomorphus kirkii X
Scolecomorphus vittatus X
Caeciliidae
Boulengerula boulengeri X
Geotrypetes seraphini X
Herpele multiplicata X
Herpele squalostoma X
Idiocranium russeli X
Anura, Bufonidae
Bufo brauni X X
Bufo camerunensis X X
Bufo funereus X X X
Bufo gracilipes X
Bufo gutteralis X
Bufo kisoloensis X X X
Bufo latifrons X X
Bufo maculatus1 X X X X X
Bufo superciliaris X X
Bufo steindachneri X X
Bufo tuberosus X
Didynamipus sjostedti X
Mertensophryne micranotis X
Nectophryne afra X X
Nectophryne batesii X X
Nectophryne tornieri X X
Nectophryne viviparus X
Werneria mertensiana X
Wolterstorffina parvipalmata X
Hyperoliidae, Hyperoliinae
Alexteroon obstetricans X











APPENDIX A.


KOR VIR GAR KIB BWI USA UZU ARA
X X


X
X
X X
X


X
X X


Taxon
Hyperolius alticola2
Hyperolius argus
Hyperolius castaneus
Hyperolius cinnamomeoventris
Hyperolius concolor3
Hyperolius chrysogaster
Hyperoliusfusciventris
Hyperoliusfrontalis4
Hyperolius kivuensis
Hyperolius lateralis
Hyperolis mitchelli
Hyperolius nasutus
Hyperolius ocellatus
Hyperolius pardalis
Hyperolius parkeri
Hyperolius phantasticus
Hyperolius platyceps
Hyperolius puncticulatus
Hyperolius pusillus
Hyperolius maria
Hyperolius rubripes
Hyperolius schoutedeni
Hyperolius spinigularis
Hyperolius sylvaticus
Hyperolius tanneri
Hyperolius tuberculatus
Hyperolius tuberilingus
Hyperolius ,, ,, '1tyl,,i, 15
Hyperolius xenorhinus
Kassininae
Afrixalus dorsalis
Afrixalusfornasinii
Afrixalus laevis
Afrixalus leucostictus
Afrixalus orophilus
Afrixalus osorioi
Afrixalus paradorsalis
Afrixalus pygmaeus
Afrixalus quadrivittatus
Afrixalus uluguruensis
Afrixalus weidholzi
Kassina senegalensis
Kassina maculosa
Kassina maculata
Opisthothylax immaculatus
Phlvctimantis leonardi


X X


X X


X X


X X X


X X
X
X
X


X X

X
X


X X X


X X


X
X X
X


X X










APPENDIX A.
Taxon


KOR VIR GAR KIB BWI USA UZU ARA
X
X X X


X X


Phlyctimantis keithae
Phlyctimantis verrucosus
Leptopelinae
Leptopelis aubryi
Leptopelis argenteus
Leptopelis barbouri
Leptopelis boulengeri
Leptopelis brevirostris
Leptopelis calcaratus
Leptopelis christyi
Leptopelisfenestratus
Leptopelis flavomaculatus
Leptopelis karissimbensis
Leptopelis kivuensis
Leptopelis millsoni
Leptopelis modestus
Leptopelis notatus
Leptopelis ocellatus
Leptopelis oryi
Leptopelis parkeri
Leptopelis rufus
Leptopelis uluguruensis
Leptopelis vermiculatus
Leptopelis viridis
Microhylidae
Callulina kreffti
Holophryne rogersi
Parhoplophryne usambarica
Phrynomantis bifasciatus
Phrynomantis microps
Probreviceps macrodactylus
Spelaeophryne methneri
Pipidae
Hymenochirus boettgeri
Silurana tropicalis
Xenopusfraseri
Xenopus laevis
Xenopus muelleri
Xenopus vestitus
Xenopus wittei
Ranidae
Arthroleptinae
Arthroleptis adelphus
Arthroleptis affinis
Arthroleptis adolfifriderici
Arthroleptis poecilonotus
Arthroleptis reichei


X X


X
X
X X


X
X X


X X











APPENDIX A.
Taxon KOR VIR GAR KIB BWI USA UZU ARA
Arthroleptis stenodactylus X X


Arthroleptis tanneri X
Artholeptis tuberosus X
Arthroleptis variablis X X
Astylosternus diadematus X
Astylosternusfallax X
Astylosternus laurenti X
Astylosternus schioetzi X
Cardioglossa elegans X
Cardioglossa escalerae X
Cardioglossa gracilis X
Cardioglossa leucomystax X X
Cardioglossa nigromaculata X
Leptodactylodon bicolor X
Leptodactylodon ovatus X
Nycitbates corrugatus X
Schoutedenella schubotzi X X X
Schoutedenella sylvaticus X X
Schoutedenella xenodactylus X
Scotobleps gabonicus X
Trichobatrachus robustus X
Hemisinae
Hemisus guineensis X X
Hemisus marmoratus X X
Hemisus olivaceus X
Petropedetinae
Arthroleptides martiensseni X X
Diii ii gioith1bii,, africanus X
Petropedetes cameronensis X
Petropedetes johnstoni X
Petropedetes newtoni X
Petropedetes parkeri X
Petropedetes perreti X
Phrynobatrachus acridoides X
Phrynobatrachus auritus X X X
Phrynobatrachus bequaerti X
Phrynobatrachus calcaratus X X
Phrynobatrachus cornutus X
Phrynobatrachus cricogaster X
Phrynobatrachus cryptotis X
Phrynobatrachus dendrobates6 X X X
Phrynobatrachus graueri X X
Phrynobatrachus kreffti X
Phrynobatrachus natalensis X X
Phrynobatrachus parvulus7 X X X
Phrynobatrachus parkeri X











APPENDIX A.
Taxon KOR VIR GAR KIB BWI USA UZU ARA


Phrynobatrachus perpalmatus
Phrynobatrachus scapularis
Phrynobatrachus ukingensis
Phrynobatrachus uzungwensis
Phyrnobatrachus versicolor
Phrynodon sandersoni
Raninae
Aubria subsigillata
Conraua crassipes
Conraua robusta
Euphlyctis occipitalis
Hylarana albolabris
Hylarana amnicola
Hylarana bravana
Hylarana galamensis
Hylarana lepus
Ptychadena c.f aequiplicata
Ptychadena chrysogaster
Ptychadena huguettae
Ptychadena maccarthyensis
Ptychadena mascareniensis
Ptychadena oxyrhynchus
Ptychadena perreti
Ptychadena straeleni
Ptychadena taenioscelis
Ptychadena tournieri
Ptychadena trinodis
Pyxicephalus edulis
Rana angolensis8
Rana ornata
Rana ruwenzorica
Rhacophoridae
Chiromantis rufescens
Chiromantis xerampelina


x
X


X X


X
X X


X X


X X


X* X*


X X


X X

X X X
X
79 58 42 28 27 24 16 25


Notes: 1 includes Bufo regulars
includes Hyperolius discodactylus after Drewes (1991)
3 includes Hyperolius balfouri
includes Hyperolius diaphanus
includes subspecies H. v. bayoni, H. v. pitmani, H. v. ornatus, H. v. ,,uni iPi, ,ii .
H. v. pachydermus
includes Phrynobatrachus minutus (Laurent, 1972)
includes Rana daesageri (Laurent, 1972)















APPENDIX B
REPTILES OF EIGHT TROPICAL AFRICAN LOCALITIES.


Reptiles of eight tropical African localities. Species lists were from the following sources;
Korup National Park (KOR; Lawson, 1993); Virungas National Park (VIR; DeWitte, 1941);
Garamba National Park (GAR; DeWitte, 1966); Kibale National Park (KIB; this study; Pitman,
1974); Bwindi National Park (BWI; Drewes, 1998); Usambara mountains (USA; Howell, 1993);
Uzungwa National Park (UZU; Howell, 1993); Arabuko-Sokoke Forest Reserve (ARA; Drewes
and Rotich, 1995).

Taxon KOR VIR GAR KIB BWI USA UZU ARA
Testudines, Testudinidae
Kinixys erosa X X
Kinixys homeana X
Kinixys belliana X X
Geochelone pardalis X
Pelomedusidae
Pelomedusa subrufa X X
Pelusios castaneus X X
Pelusios gabonensis X
Pelusios niger X
Pelusios nigricans X
Trionychidae
Trionyx triunguis X
Squamata, Agamadae
Agama agama X X
Agama atricollis X X X X
Agama cyanogaster X
Chamaeleonidae
Bradypodionfisheri X
Bradypodion oxyrhinum X
Bradypodion spinosum X
Bradypodion tenue X
Chamaeleo adolfifriderici X X X
Chamaeleo bitaeniatus X
Chamaeleo carpenter X
Chamaeleo cristatus X
Chamaeleo deremensis X
Chamaeleo dilepis X
Chamaeleo eisentrauti X
Chamaeleo ellioti X X X
Chamaeleo gracilis X X
Chamaeleo goetzei X
Chamaeleo ituriensis X











APPENDIX B.
Taxon KOR VIR GAR KIB BWI USA UZU ARA


X X


Chamaeleo johnstoni
Chamaeleo laterispinis
Chamaeleo montium
Chamaeleo oweni
Chamaeleo rudis
Chamaeleo senegalensis
Chamaeleo tempeli
Chamaeleo werneri
Chamaeleo xenorhinus
Rhampoleon boulengeri
Rhampoleon brevicaudatus
Rhampoleon kirstenii
Rhampholeon spectrum
Rhampoleon temporalis
Cordylidae
Cordylus tropidosternum
Gekkonidae
Cnemaspis africana
Cnemaspis dickersoni
Cnemaspis koehleri
Cnemaspis spinicollis
Cnemapsi quattuorseriatus
Cnemaspis uzungwe
Hemidactylus ansorgii
Hemidactylus brookii
Hemidactylus echinus
Hemidactylus fasciatus
Hemidactylus ituriensis
Hemidactyus mabouia
Hemidactylus platycephalus
Hemidactylus squamulatus
Lygodactylus conradti
Lygodactylus cfconraui
Lygodactylus gravis
Lygodactylus picturatus
Lygodactylus williamsi
Pachydactylus bibronii
Urocotyledon wolterstorffi
Gerrhosauridae
Gerrhosaurus major
Gerrhosaurus flavigularis
Gerrhosaurus nigrolineatus
Lacertidae
Adolfus africanus
Adolfus jacksoni
Adolfus vauereselli


X X


X X


X X X X


X X


X X


X X X
X
X X


X X











APPENDIX B.


KOR VIR GAR KIB BWI USA UZU ARA
X


Taxon
Eremias nitida
Gastropholis prasina
Heliobolis spekii
Holoaspis guentheri
Ichnotropis capensis
Latastia longicaudata
Scincidae
Feylinia currori
Lygosoma afer
Lygosomafernandi
Lygosoma modestum
Lygosoma pembanum
Lygosoma sundevallii
Leptosiaphos aloysiisabaudiae
Leptosiaphos blochmanni
Leptosiaphos graueri
Leptosiaphos hackarsi
Leptosiaphos luberoensis
Leptosiaphos meleagris
Leptosiaphos rhomboidalis
Mabuya affinis
Mabuya blandingii
Mabuya boulengeri
Mabuya brevicollis
Mabuya maculilabris
Mabuya megalura
Mabuya perroteti
Mabuya planifrons
Mabuya quinquetaeniata
Mabuya striata
Mabuya sudanesis
Mabuya varia
Melanoseps ater
Proscelotes eggeli
Varanidae
Varanus albigularis
Varanus niloticus
Varanus exanthematicus
Zonuridae
Chamaesaura tenuior
Serpentes
Atractaspidae
Atractaspis aterrima
Atractaspis bibronii
Atractaspis corpulenta
Atractaspis irregularis'


X X
X

X X
X
x


X X
X


X X


X X X
X


X X
X


X X X X X











APPENDIX B.
Taxon KOR VIR GAR KIB BWI USA UZU ARA
Atractaspis reticulata X


Boidae
Calabaria (Eryx) rienhardtii
Eryx colubrinus
Python sebae
Python regius
Colubridae
Afronatrix anascopus
Aparallatus guentheri
Aparallatus modestus
Aparallatus turner
Aparallatus werneri
Boiga blandingii
Boiga pulverulenta
Bothropthalmus lineatus
Calamelaps cf unicolor
Crotaphopeltis hotamboeia
Crotaphopeltis tornieri
Dasypeltis atra
Dasypeltis fasciata
Dasypeltis macrops
Dasypeltis medici
Dasypeltis scabra
Dipsadoboa duchesnii
Dipsadoboa elongata
Dipsadoboa unicolor
Dipsadoboa werneri
Dispholidus typus
Dromophis lineatus
Duberria lutrix
Gastropyxis smaragdina
Geodipsas depressiceps
Geodipsas vauerocegae
Gonionotophis brussauxi
Grayia caesar
Grayia ornata
Grayia smithii
Grayia tholloni
Hapsidophrys lineatus
Hemirhagerrhis nototaenia
Homonotus (Lamprophis)
modestus
Lamprophisfuliginosus
Lamprophis lineatus
Lamprophis olivaceus
Lamprophis virgatus


X X
X


X

X X


X X X
X X
X X X
X
X X X


X X


X X


X X X X


X X
X
X
X X X
X


X X

X X


X X
X
X


X X