The proposed Parc Regional de Belomotse - Frontier-publications ...

Frontier Madagascar Environmental Research
REPORT 6
The proposed Parc Regional de
Belomotse
Biodiversity survey and conservation evaluation
Frontier-Madagascar
2003

Frontier Madagascar Environmental Research
Report 6
The proposed Parc Regional de
Belomotse
Biodiversity survey and conservation evaluation
Emmett, D. A., Fanning, E. & Olsson, A. (eds)
Frontier-Madagascar
University of Toliara
The Marine Sciences Institute
Madagascar
The Society for Environmental Exploration
UK
Toliara
2003

Suggested Technical Paper citation:
Frontier-Madagascar (2003) The proposed Parc Regional de Belomotse: biodiversity survey and
conservation evaluation. Frontier-Madagascar Environmental Research Report 6. Society for
Environmental Exploration, UK and the Institute of Marine Sciences, University of Toliara,
Madagascar.
Suggested Section citations:
Emmett, D. A. & Rasoambolaniaina, Y. (2003) Flora: In The proposed Parc Regional de Belomotse:
biodiversity survey and conservation evaluation. pp. 13 – 18. Frontier Madagascar Environmental
Research Report 6. Society for Environmental Exploration, UK and the Institute of Marine Sciences,
University of Toliara, Madagascar.
Emmett, D.A., Olsson, A, Bloxam, C., Henson, D. & Thomas, H. (2003) Fauna: In The proposed Parc
Regional de Belomotse: biodiversity survey and conservation evaluation. pp. 19 – 62. Frontier
Madagascar Environmental Research Report 6. Society for Environmental Exploration, UK and the
Institute of Marine Sciences, University of Toliara, Madagascar.
Olsson, A., Henson, D. & Emmett, D.A. (2003) Resource Use and Socio-Economics : In The proposed
Parc Regional de Belomotse: biodiversity survey and conservation evaluation. pp. 62– 64. Frontier
Madagascar Environmental Research Report 6. Society for Environmental Exploration, UK and the
Institute of Marine Sciences, University of Toliara, Madagascar.
Bloxam, C., Henson, D., Olsson, A. & Emmett, D.A. (2003) Tourism: In The proposed Parc Regional
de Belomotse: biodiversity survey and conservation evaluation. pp. 64 – 71 Frontier Madagascar
Environmental Research Report 6. Society for Environmental Exploration, UK and the Institute of
Marine Sciences, University of Toliara, Madagascar.
Emmett, D. A. & Olsson, A. (2003) Conclusions and Recommendations: In The proposed Parc
Regional de Belomotse: biodiversity survey and conservation evaluation. pp. 71 – 76. Frontier
Madagascar Environmental Research Report 6. Society for Environmental Exploration, UK and the
Institute of Marine Sciences, University of Toliara, Madagascar.
This report series was created in 2005 and incorporated previous reports published by Frontier-
Madagascar. The previous citation for this report was:
Frontier Madagascar (2003) Emmett, D. A., Fanning, E., & Olsson, A. (eds) The proposed Parc
Regional de Belomotse: biodiversity survey and conservation evaluation. Frontier Madagascar
Environmental Research Report 6. ISSN 1479-120X Society for Environmental Exploration, UK and
the University of Toliara & Institute of Marine Sciences, Madagascar.
The Frontier -Madagascar Environmental Research Report Series is published by:
The Society for Environmental Exploration
50-52 Rivington Street,
London, EC2A 3QP
United Kingdom
Tel: +44 (0)20 7613 3061
Fax: +44 (0)20 7613 2992
E-mail: research@frontier.ac.uk
Web Page: www.frontier.ac.uk
ISSN 1479-120X (Print)
ISSN 1748-3719 (Online)
ISSN 1748-5126 (CD-ROM)
© Frontier-Madagascar 2003, 2005

The Society for Environmental Exploration (SEE)
The society is a non-profit making company limited by guarantee and formed in 1989. The society’s
objectives are to advance field research into environmental issues and implement practical projects
contributing to the conservation of natural resources. Projects organised by The Society are joint
initiatives developed in collaboration with national research agencies in co-operating countries.
The University of Toliara / IHSM
The University of Toliara was established in 1976 as a centre for learning and research in the biological
sciences. The Institut Halieutique et des Sciences Marines (IH.SM) is the marine department of the
University of Toliara.
The University is studying the flora and fauna of Madagascar and is conducting research into the
maintenance and improvement of the environment and the sustainable use of the natural resources in
the region.
The WWF Madagascar Dry Forest Programme
The WWF Madagascar Dry Forest Programme was initiated in 1998. It is a 50-year programme with
the aim of conserving key habitats throughout southwest Madagascar. This eco-region is suffering
more rapid habitat loss than any other region in Madagascar. The project aims to implement
management plans for 15-25% of the region, incorporating all major habitat types and including areas
of high biodiversity.
The Frontier-Madagascar Forest Research Programme
The Society for Environmental Research, the University of Toliara and the WWF Madagascar Dry
Forest Programme have been conducting collaborative research into environmental conservation issues
since July 2001 under the title of the Frontier-Madagascar Forest Research Programme. From July
2001 until July 2002, the Programme has been working in the forests of the Sept Lacs region. This
activity was aimed at identifying core areas of biodiversity and establishing baseline biodiversity and
resource-use information in those areas. In addition, tourism feasibility surveys were undertaken.
FOR MORE INFORMATION
Frontier -Madagascar
BP41, Antsiranana, 201
MADAGASCAR
Tel/Fax: +261 (0) 20 82 23117
E-mail: frontier@wanadoo.mg
L’Insitute Halieautique et des Sciences Marine
(IHSM)
Zone Portuaire, BP 141,
Tulear 601
MADAGASCAR
Tel: +261 (0) 20 94 43552
Fax: +261 (0) 20 94 43434
E-mail: ihsm@wanadoo.mg
Society for Environmental Exploration
50-52 Rivington Street,
London, EC2A 3QP. U.K.
Tel: +44 (0) 20 7613 3061
Fax: +44 (0) 20 7613 2992
E-mail: research@frontier.ac.uk
Internet: www.frontier.ac.uk

Contents
List of Figures
List of Tables
List of Appendices
List of Abbreviations
Executive Summary
Acknowledgements
Page Number
i
i
ii
ii
iii
v
1. INTRODUCTION 1
1.1 Madagascar 1
1.2 The history of conservation in Madagascar 2
1.3 Threats to conservation 3
1.4 Rationale 4
2. DESCRIPTION OF THE SURVEY AREA 5
2.1 The region 5
2.2 Location of the proposed Regional Park 5
2.3 Local communities 6
2.4 Site history 6
2.5 Topography, geology, and climate 7
2.6 Vegetation 8
2.7 Previous research 8
3. THE STUDY SITES 9
3.1 Data collection methods 9
3.2 Zoological trapsites 10
3.3 Trapsite procedure 12
4. PLANTS 13
4.1 Introduction 13
4.2 Methods 13
4.3 Results 14
4.3.1 Gallery forest 15
4.3.2 Spiny forest 16
4.4 Discussion 16
4.4.1 Dominant species and forest structure 16
4.4.2 Plant-resource use 17
4.4.3 Threats to habitats 18
5. MAMMALS 19
5.1 Introduction 19
5.2 Methods 19
5.2.1 Pitfall trapping 19
5.2.2 Shernam and mesh trapping 20
5.2.3 Observations, tracks and calls 20
5.2.4 Mist-nets and harp traps 20
5.2.5 Owl pellets 20
5.2.6 Specimen handling 21
5.3 Results 21

5.3.1 Trapping success 21
5.3.2 Rattus rattus population composition 22
5.3.3 Observations, tracks and calls 23
5.3.4 Owl pellet analysis 23
5.4 Discussion 23
5.4.1 Ecological distribution and species richness 23
5.4.2 Introduced species 25
5.4.3 Seasonal variation 26
5.4.4 Rare species and range extensions 27
5.4.5 Threats to mammal conservation 28
6. TRIAENOPS POPULATION STUDY 29
6.1 Introduction 29
6.2 Methods 29
6.3 Results 30
6.4 Discussion 31
6.4.1 Temporal variation 31
6.4.2 Seasonal variation 31
6.4.3 Breeding status 32
6.4.4 Population density and conservation status 32
7. DIURNAL LEMUR POPULATION STUDY 32
7.1 Introduction 32
7.2 Methods 33
7.3 Results 34
7.3.1 Line transects 34
7.3.2 Direct observation 34
7.4 Discussion 35
7.4.1 Comparison of methods 35
7.4.2 Lemur density and distribution 35
7.4.3 Conservation status 36
8. BIRDS 36
8.1 Introduction 36
8.2 Methods 37
8.3 Results 37
8.4 Discussion 38
8.4.1 Rare species and endemism 38
8.4.2 Ecological variation 38
8.4.3 Threats to bird conservation 39
9. REPTILES 40
9.1 Introduction 40
9.2 Methods 40
9.3 Results 41
9.4 Discussion 41
9.4.1 Biogeography and species richness 41
9.4.2 Seasonal variation 42
9.4.3 Ecological distribution 42
9.4.4 Rare species 44
9.4.5 Range extensions and locality records 45
9.4.6 Threats to reptile conservation 45
10. AMPHIBIANS 46

10.1 Introduction 46
10.2 Methods 47
10.3 Results 47
10.4 Discussion 48
10.4.1 Biogeography and species richness 48
10.4.2 Seasonal variation 48
10.4.3 Ecological distribution 48
10.4.4 Threats to amphibian conservation 48
11. INVERTEBRATES 49
11.1 Butterflies 49
11.1.1 Introduction 49
11.1.2 Methods 49
11.1.3 Results 50
11.1.4 Discussion 50
11.1.4.1 Ecological distributions 50
11.1.4.2 New species, rare species and range extensions 51
11.2 Malaise trapping 53
11.2.1 Introduction 53
11.2.2 Methods 53
11.2.3 Results 53
11.3 Leaf litter extraction 54
11.3.1 Introduction 54
11.3.2 Methods 54
11.3.3 Results 54
11.3.4 Discussion 54
11.4 Dragonflies 56
11.4.1 Introduction 56
11.4.2 Methods 56
11.4.3 Results 56
11.5 Molluscs and millipedes 57
11.5.1 Introduction 57
11.5.2 Methods 57
11.5.3 Results 58
11.6 Spiders and scorpions 58
11.6.1 Introduction 58
11.6.2 Methods 58
11.6.3 Results 58
12. RESOURCE USE AND SOCIO ECONOMICS 59
12.1 Introduction 59
12.2 Methods 59
12.2.1 Ground-truthing and GIS data layers 60
12.2.2 Biological diversity and habitats maps 60
12.2.3 Human settlements, land, and resource use 60
12.3 Results 61
12.3.1 Biological diversity and distribution 61
12.3.2 Human populations and local management 61
12.3.3 Resource use 62
12.3.3.1 Gallery and spiny forests 62
12.3.3.2 Wetlands 62
12.3.4 Agriculture 63
12.4 Discussion 64

13. TOURISM 64
13.1 Introduction 64
13.2 Methods 65
13.3 Results and discussion 65
13.3.1 Priority areas for tourism 65
13.3.2 Transport 66
13.3.3 Campsite location 67
13.3.4 Campsite logistics 68
13.3.5 Attractions and activities 68
13.3.5.1 Guided walks 68
13.3.5.2 Guides 69
13.3.5.3 Boat trips 69
13.3.5.4 Bird hides 69
13.3.6 Amenities within the area 70
13.3.7 Potential problems 70
13.3.8 Monitoring 70
14. CONCLUSIONS AND RECOMMENDATIONS 71
REFERENCES 73

List of Figures
Figure 1:
Figure 2:
Figure 3:
Figure 4:
Figure 5:
Figure 6:
Figure 7:
Figure 8:
Figure 9:
Figure 10:
Map of the Toliara region, showing the Onilahy River. (Reproduced from
Lonely Planet, Madagascar 4 th Edition, May 2001.)
Map of the proposed Regional Park (from WWF, Toliara)
Map of the survey area from Antafoky Lake to the Sept Lacs region.
Annual trapping of small mammals
Annual composition of R. rattus composition
Triaenops nocturnal activity pattern
Diurnal lemur line-transect locations in the survey area
Proportions of crops grown in the survey area
Map of core area of Regional Park of interest to tourists
Position of proposed tourist walks and birdhides at Antafoky Lake
List of Tables
Table 1:
Table 2:
Table 3:
Table 4:
Table 5:
Table 6:
Table 7:
Table 8:
Table 9:
Table 10:
Table 11:
Summary of biodiversity and endemism of taxa surveyed
Trapsite description
Number and endemism of tree and scrub families, genera and species in
different habitats
Most cost common tree species in different habitats.
Number and density of diurnal lemurs within the survey area.
Distribution of reptile species within the survey area
Amphibian species list from the survey area
Summary of butterfly species from the survey area
Ecological distribution of butterflies within the survey area
Ant species list from the survey area
Proportion of crops grown and sold by local communities within the survey
area
i

List of Appendices
Appendix 1:
Appendix 2:
Appendix 3:
Appendix 4:
Appendix 5:
Appendix 6:
Appendix 7:
Appendix 8:
Appendix 9:
Appendix 10:
Appendix 11:
Appendix 12:
Appendix 13:
Trapsite habitat description
List of plant species found in riparian forest
List of plant species found in spiny forest
Plant numbers and specific uses – Sept Lacs (west)
Plant numbers and specific uses – Antafoky
List of mammal species found in the proposed Regional Park
List of bird species positively identified in the proposed Regional Park
List of bird species seen but not positively identified in the proposed
Regional Park
List of reptile species found in the proposed Regional Park
Butterfly species collected in the proposed Regional Park, with associated
habitat data.
Description of proposed tourist walks
Description of proposed bird hides
List of collaborators for specimen identification and deposition
List of Abbreviations
ANGAP: Association Nationale pour le Gestion des Aires Protégées
GELOSE: Gestion Locale Securisee
GIS: Geographical Information System
GPS: Global Positioning System
IH.SM: Institut Halieutique et des Sciences Marines
IUCN: The World Conservation Union
MGF: Frontier Madagascar Forest research programme
ONE: Office Nationale pour l’Environnement
SAGE: Service d’Appui a la Gestion de L’Environment
SEE: Society for Environmental Exploration
WWF: World Wildlife Fund
ii

EXECUTIVE SUMMARY
This report describes the biodiversity survey of the recently demarcated Parc Regional de
Belomotse, Toliara Province, Madagascar, conducted as part of the Frontier-Madagascar Forest
Research Programme, between July 2001 and September 2002.
The Frontier-Madagascar surveys involved a study of the extent and condition of the different
vegetation types; vegetation species composition; and the collection of baseline biodiversity and
ecological data for mammals, birds, reptiles, amphibians, butterflies and other selected
invertebrates. Additional studies investigated resource use in the area, and the potential for the
development of ecotourism.
The surveyed area was, until recently, ungazetted and did not have a formally demarcated
boundary. It has now been designated as a proposed Regional Park. Within this area is a core
zone, managed through the locally enforced Gestion Locale Securisée (GELOSE). This core
area, described for the purposes of this report as the ‘Sept Lacs region’, covers over 3,000
hectares between 40 and 250m asl, approximately 500 hectares of which are riparian forest and
wetland habitats. The area surveyed by Frontier-Madagascar extends beyond the locally managed
(Sept Lacs) core zone, and also includes a lakeside habitat, Antafoky Lake, and associated areas
of riparian forest covering approximately 150 hectares. Thus, the report details surveys of the
Sept Lacs region and the Antafoky Lake region, although both sites are contained within the
proposed Regional Park.
The surveyed area contains approximately 800 inhabitants, distributed through five villages.
These inhabitants belong to the Mahafaly, Tandroy and Masikoro tribes.
The surveyed area contains many rare species. Four species of mammal, two species of bird and
four species of reptile recorded during this survey are classified as vulnerable (IUCN, 2002).
Many of the species collected are too poorly studied to be designated IUCN status at this time.
Furthermore, several undescribed species were discovered during this survey, including a Blind
snake (Typhlops sp.), a pipistrelle bat (Pipistrellus sp.) a long-fingered bat (Miniopterus sp), and
one butterfly species (Strabena sp.). It is expected that the results of the entomological
collections will yield many new species.
The species richness and endemism of the surveyed taxa are summarised in Table 1.
Table 1. Summary of biodiversity and endemism of taxa surveyed.
Taxon: Total no. No. of Percentage endemism
of species endemics
Trees and shrubs 142 16* -
Mammals 26 21 81%
Birds 79 33 42%
Reptiles 55 50 91%
Amphibians 6 5 83%
Butterflies 61 - -
* Endemic genera (species endemism unknown)
The riparian forest and wetlands within the Toliara region are threatened with
destruction due to clearance or drainage for agricultural purposes. Charcoal production is
leading to the loss of many trees of all sizes, and larger trees are logged for the
production of pirogues (dugout canoes) or for local burial purposes. Grazing by domestic
animals leads to further damage of the disturbed areas. Due to the dry climate,
iii

egeneration is slow. Some species will become locally extinct unless the current trend in
environmental degradation is reversed.
The major underlying cause of environmental degradation is the low level of agricultural
productivity, obliging local people to exploit natural resources. Supplementary factors are
rapid population growth, low educational levels, harmful agricultural methods due to poor
soils and techniques, and specific aspects of the Malagasy culture. Ultimately, in order to
reduce exploitation of natural resources, living standards of local people must be raised.
However, this will take a long time to achieve, and in the short term it is necessary to reduce
the rate of environmental degradation by promoting sustainable agriculture and natural
resource use, raising awareness about environmental problems, and initiating appropriate
forest management schemes.
The proposed Regional Park contains some of the most important areas of riparian forest and
wetland in southwest Madagascar. However, the remaining areas are becoming fragmented.
Habitat destruction within the GELOSE-managed Sept Lacs core zone is occurring at a
comparatively low rate, and may be sustainable in the short term. However, this area contains
only a small number of wetland habitats, and many of the riparian forests are already degraded
and fragmented.
The riparian forest around Antafoky Lake is the largest unfragmented area of riparian forest in
the surveyed area. However, this area is suffering from deforestation at an alarming rate and,
unless checked, it will be mostly destroyed within three or four years. Also, this lake and the
associated forests contain the highest level of biological diversity found within the proposed
Regional Park. All vulnerable species found during the surveys were recorded at this site, as were
all but one of the undescribed mammal, reptile and butterfly species. Therefore, it is essential that
immediate support be given to this site in order for it to remain a viable and integral part of the
Regional Park.
iv

ACKNOWLEDGEMENTS
This summary report has been written based on research conducted during five Frontier-
Madagascar phases, running consecutively from July 2001 to September 2002, funded by the
Society for Environmental Exploration.
The research team comprised Frontier-Madagascar field-staff, scientists from WWF-
Madagascar, researchers from the University of Toliara, and Frontier-Madagascar volunteer
research assistants. Frontier-Madagascar is a collaboration between the Society for
Environmental Exploration, UK and the Institute of Marine Sciences, Madagascar.
This report is the culmination of the advice, co-operation, hard work and expertise of many
people. In particular, acknowledgments are due to the following:
L’INSTITUT HALIEUTIQUE ET DES SCIENCES MARINE (IHSM)
F-M Co-ordinators:
Dr. Man Wai Rabenevanana
Dr. Mara Edouard Remanevy
SOCIETY FOR ENVIRONMENTAL EXPLORATION (SEE)
Managing Director Eibleis Fanning
Development Programme Manager Elizabeth Humphreys
Programme Manager (Research) Nicola Beharrell
Programme Manager (Operations) Matthew Willson
WWF-MADAGASCAR
Regional Director
Research & Training Officer
Biodiversity Officer/Herpetologist
Sub-regional co-ordinator
Mark Fenn
Pierre Jules Rakotomalaza
Achille Raselimanana
Voninjatovo Andriamiarisoa
FRONTIER-MADAGASCAR
Project Co-ordinator: David Emmett
Research Co-ordinator:
Annette Olsson
Assistant Research Co-ordinators: Craig McCoy, Catherine Bloxam, Esther Ball, and
David Henson
Logistics Manager:
Hannah Thomas
Field Assistants:
Vomash, Ennanetse Alfred
ASSOCIATION NATIONALE POUR LE GESTION DES AIRES PROTEGEES
(ANGAP)
Regional Director
Jocelyn Rakotomalala
CHICAGO FIELD MUSEUM
Curator of Mammology
Steven Goodman
AMERICAN MUSEUM OF NATURAL HISTORY (AMNH)
Curator of Herpetology
Christopher Raxworthy
CALIFORNIAN ACADEMY OF SCIENCES (CAS)
Curator of Entomology Brian Fisher
NATURAL HISTORY MUSEUM (NHM), LONDON
v

Research Associate (Entomologist)
UNIVERSITY OF LEEDS
Postgraduate Entomologist
David Lees
Alison Cameron
UNIVERSITY OF ANTANANARIVO
DEA Mammology student
Research Assistants during the five phases of fieldwork were:
Juliana Vonimanitra Rahantavololona Rasoma
July-Sept 2001 Oct-Dec 2001 Jan-March 2002
Ross Adamson Rose Adams Geoff Allen
Holly Baldwin Leo Barasi Kathryn Baskerville
Marcus Barker Aaron Alexandra-Bloch Christopher Chapman
Andrew Britton Noa Dychtwald Katie Foster
Alastair Campbell Amanda Greenfield Steven Grzywacz
Helen Dacre Patrick Griffiths Chris Hodgson
Emily Fox Samantha Hanvey Rebecca Hughes
Tony Kavanagh GeorgeLow Lara Hunt
Sharon Ketley Ben Rayner Akiko Kumagae
Jamil Khan Cat Simmons Charlotte Lawlor
Emily Kilborn Lizzy Tennant Susan Loose
Alwyn Mortimer James Thorburn Edward Morrison
Tara O’Kelly
Kate Wilkins
Victoria Peaple
Richard Wiltshire
Anna Scaife
April-June 2002 July-Sept 2002
Philip Bowles
Janette Buckley
Thomas Carr
Emma Butterworth
Charlotte Corke
Karen Dawson
Katie Francis
Colin Higgins
Deborah Hendson
Charlotte Jones
Katherine Howard
Curstain Keltie
Melvin Jarmin
Katy Littler
Sarah Taylor
Michael Sanderson
Helen Seyler
vi

1. Introduction
1.1 Madagascar
Madagascar is the fourth largest island in the world, with a surface area of approximately
587,000 sq km. The country stretches from 12º S, 50º E in the north to 26º S, 44º E in the
south. A series of mountain ranges runs down the centre of the island from north to
south, which creates a central highland plateau. To the east steep slopes fall towards the
coast, whereas to the west the slopes are longer and gentler.
Trade winds approaching the island from the Indian Ocean to the east bring rain that falls
over the escarpment. The west coast lies in a rain shadow and receives very little rain.
Monsoon rains approach the island from the north and their strength decreases towards
the south. The north-eastern part of the island is therefore the wettest, receiving up to
4,000 mm rain per year, while the south-western part is the driest, with some areas
receiving less than 300 mm per year. The highland plateau ranges between 750 and
1,350 metres above sea level (m a.s.l.) and the Tsaratanana massif in the north rises to
2,876 metres. Numerous rivers run from their sources in these highlands towards the
coasts (Garbutt, 1999; Rogers, 2001).
Due to the differences in altitude and latitude throughout the island, the climate varies
both from area to area and throughout the year. In the central highlands the night
temperature drops to near freezing point during the dry season, and in the north-east the
day temperature often exceeds 40 º C.
The variation in geography and climate creates a wide variety of habitats. The humid
areas to the east are dominated by lowland and sub-montane evergreen rainforests.
Grasslands stretch over the highland plateau and deciduous and dry forests dominate
the western areas. Spiny forest and scrub thrive in the arid south. Along the coastal
areas, marshlands, mangroves and coastal forests are common. These different
habitats are home to a great diversity of plants and animals.
Madagascar is classified amongst the top eight mega-diversity countries on Earth. It
ranks amongst the top four of the world’s most threatened biodiversity hotspots, and
contains exceptionally high levels of endemic species, many of which are classified as
vulnerable or endangered (Myers et al, 2000).
The flora of Madagascar shows very high levels of endemism. It has been estimated
that over 96% of all trees and shrubs are endemic solely to Madagascar. This includes
approximately 161 endemic genera (Schatz, 2001). Only one other country, Brazil,
surpasses these figures. Over 95% of all plant species found in the dry forests of south
and southwest Madagascar are regionally endemic; that is, they are not only endemic
to Madagascar, they are also endemic to the dry southern eco-region.
The high levels of endemism in Madagascar are not restricted to the flora. Over 45%
of all bird species are endemic, including 5 endemic families and 35 endemic genera
(Morris & Hawkins, 1998). The figure for reptiles is even higher, with 91% species
endemism and 56% genus endemism. At present, nearly 350 reptile species are
known. At least 230 different amphibian species have been identified, and over 99%
of the species described are endemic to Madagascar and its offshore islands. This
1

epresents one of the highest numbers of endemic amphibians from any country in the
world. However, it is with the primates that Madagascar shows the highest degree of
endemism; 100% species endemism, encompassing 14 genera and 5 families (Garbutt,
1999). Madagascar contains over one third of the world’s primate families.
1.2 The History of Conservation in Madagascar
Madagascar has suffered extremely high levels of extinction since man arrived on the
island. The date (and countries of origin) of man’s first arrival in Madagascar has long
been the subject of studies. Archaeological evidence seems to indicate that man arrived
approximately 2,000 years ago (Burney et al, 1997). Since this remarkably late date,
man has precipitated a landslide of extinctions. Again, archaeological evidence validates
the ‘Blitzkrieg hypothesis’ put forward by Martin (1984), whereby as the human
population on Madagascar began to increase, man began to over-hunt the animals and
burn the land for cultivation, leading to a population crash amongst the endemic
megafauna. This in term led to the extinction of many species, including up to 15 lemur
species, incorporating 8 genera. Other losses included two genera of elephant bird and at
least two species of pygmy hippo. The extinction of these megafauna was remarkably
rapid, and relatively recent; the pygmy hippos and elephant bird are thought to have gone
extinct less than 1,000 years ago.
Madagascar has had a long history of attempted conservation efforts. After French
colonisation in 1895, prohibitive regulations were implemented (Kull, 1996). These
covered such aspects as banning the killing of lemurs, the initiation of replanting
schemes using eucalyptus and pine and, in 1927, the initiation of ten nature reserves.
However, France, through intensive logging activities combined with cash cropping of
coffee, precipitated further, rapid destruction of the remaining rainforests. This forced
many Malagasy to move into the highlands to find land for subsistence cultivation. In
turn, the French government banned tavy (slash and burn agriculture). This led the
Malagasy people to revolt, which in turn led to independence.
After independence, the lack of funding to Madagascar proved disastrous for
conservation efforts. In addition, in 1975 the president of Madagascar, Didier Ratsiraka,
established a socialist regime and expelled all foreigners from the country; effectively
halting all outside support and severely hindering progress (Richard and O’Connor,
1997). However, in the face of mounting external pressure, in 1985 an International
Conference on Conservation for Development was held in Antananarivo.
This led to a dramatic increase in external support for Malagasy conservation
development by organisations such as WWF and World Bank. It also led to an influx of
foreign scientists. In turn, a new Malagasy non-governmental organisation was set up;
the National Association for the Management of Protected Areas (ANGAP), which led
to the introduction of implemented conservation and development strategies in specific
protected areas, or National Parks. Today, ANGAP control over forty National Parks and
protected areas.
General elections in 2002 led to a new president, Marc Ravalomanana. It is to be hoped
that this will lead to further improvements in conservation development and support in
Madagascar.
2

1.3 Threats to Conservation
The high diversity of habitats and species in Madagascar is under severe threat. The
human population in Madagascar is increasing by about 3% per year and has now
reached approximately fifteen million (Bille, 2002; Durbin, 1994). With a growing
population, the need for food and other resources also increases, putting further pressure
on natural resources throughout the country.
Over the last few decades, the Malagasy culture has shifted towards a more materialistic
society. This has led to a demand for non-basic goods, leading to the need for a financial
income. In turn, this has led to excess production of food-crops and cattle for export and
sale.
Natural vegetation is being cleared by the traditional slash and burn method, tavy, for
maize cultivation. Forest and scrub areas in valleys and on hillsides are partly cut down,
and then the area is burnt. The cleared land is used to grow maize. After two or three
seasons the soil is usually too poor in nutrients for crops to grow, so the area is
abandoned and new sites are found. This method is highly destructive and makes large
areas of land barren for a long period after use. This is especially true in the dry forests
of the south and west of Madagascar, where regeneration is exceptionally slow.
Most of the crops are consumed locally and nationally, but there is now an increasing
trend for maize to be exported to mainland Africa, Mauritius, Réunion and the
Seychelles (Seddon et al., 2000). This exported maize is mainly used for animal food.
Over 80% of Malagasy households are dependent on wood or charcoal for cooking.
Certain tree species are more useful for charcoal production than others and areas are
either selectively logged or are cleared of all trees. The latter technique is becoming
more common as the demand for charcoal is rising. Often, only a small proportion of the
charcoal from a production area is used locally. The majority of the charcoal is taken for
sale to urban areas and regions where charcoal production is not possible. Charcoal is
mainly produced in areas with a fairly good infrastructure, as it is difficult to transport
the large heavy sacks without a vehicle.
Although alternative materials such as bricks and fibreglass are used in some areas for
construction of buildings and boats, wood is still the most commonly used material and
logging of specific tree species is carried out to provide planks and poles. Also, when
bricks are made, the process uses a large amount of firewood.
Zebu cattle and other livestock play an important role in Malagasy life and culture. Zebu
are seen as a sign of wealth and status by the Malagasy people (M. Fenn, pers. comm.,
2002). Large herds of zebu, goats and sheep are frequently found grazing in the forests
near rural villages. Intensive grazing prevents plant regeneration. Saplings, shrubs and
grasses are eaten, dramatically changing the structure of the forest plant communities.
Also, during periods of very dry weather the branches from large trees may be cut down
to provide additional food for the livestock.
These pressures on natural resources have led to the degradation of many habitats, and
remaining areas are often fragmented. At present, more than 90% of the original natural
vegetation has been lost in Madagascar (Myers et al., 2000).
3

Many species of mammals, reptiles and birds are often hunted for food (Ferguson, 2002).
Specific reptile species are collected for their skins, or for export and sale in the pet
trade, mostly to Europe and America. Excessive hunting or mass-collections of species
can cause severe declines in local populations, which can cause extinctions at the local
and national level.
Exotic species that have been deliberately or accidentally introduced by man to
Madagascar can cause severe threats to populations of native species. Many introduced
species often adapt well to new surroundings and can be very successful competitors.
Furthermore, introduced species often carry diseases and parasites to which the native
species have no resistance, causing serious declines in affected populations.
1.4 Rationale
In 1996, WWF identified 234 ecological regions (eco-regions) around the world on
which to concentrate their present and future efforts. Five eco-regions were identified in
Madagascar. In 1997, two eco-regions were selected as priority regions for WWF-
Madagascar. These were the dry deciduous forests of the west, and the spiny forests of
the south and southwest. The spiny forest eco-region comprises over 6,612,850 hectares
of land, less than 3,000,000 hectares of which is forested. Only 3.2% of the entire ecoregion
is under protective management. Conversely, over 20% of the rainforests of
Madagascar are under protective management (Lowry et al, 1997).
Since 1998, the WWF-Madagascar Dry Forest Programme has been undertaking
reconnaissance and research throughout the region, the objectives being to identify sites
of biological interest and to collect biological and socio-economic information. The aim
has been to provide long-term protective management for 15-25% of the eco-region. To
this end, research has focussed on the identification of biodiversity hotspots and unique
or species-rich habitats. In order for long-term management to be successful, community
participation and integration is vital. WWF has also focussed on identifying the
underlying social issues and aims to address these issues at the ‘grass-roots’ level, with
involvement and support from national organisations.
The primary aim of the Frontier-Madagascar Forest Research Programme is to promote
conservation of the natural resources within selected study areas. The objectives of the
research programme aim to provide information on the biological value and
importance of surveyed areas within the spiny forest eco-region in order to identify
sites of biological conservation importance. The project will also survey resource use
and report on the status of, and threats to, the habitats in the surveyed areas.
The research programme will also help to develop GIS maps of the surveyed areas
showing priority habitats (e.g. riparian forest), as well as areas of high biodiversity
and sites of conservation importance. In addition, the programme will develop a
system of methods for monitoring aspects of biodiversity and resource-use, and assist
in the training of Malagasy personnel for long-term monitoring programmes in the
region.
The biodiversity and resource use surveys conducted by the Frontier-Madagascar
Forest Research Programme provide information that can be combined with data
4

collected by the WWF-Madagascar Dry Forest Programme and other conservation
organisations. This can be used in the identification of high priority sites of
conservation interest, and provide support for the implementation of sustainable forest
management plans and activities. These activities can then be monitored over time
based on biodiversity and ground-truthed resource use data collected during the
surveys, and can utilise personnel trained in biodiversity surveying and monitoring.
2. Description of the survey area
2.1 The Region
Southwest Madagascar is the most arid region on the island, with some areas receiving
an annual rainfall of less than 300 mm (Rogers, 2001). The area is dominated by spiny
forest composed mainly of xerophytic plants. These plants are well adapted to the dry
climate. Typical adaptations include succulent leaves, low fruit production and reduced
branching and foliage (MacPhee, 1986). The area is referred to as the spiny forest ecoregion.
This region stretches from the Mangoky River on the west coast to Fort Dauphin
in the southeast (WWF, 2002; Seddon, 2000).
The coastal town of Toliara has a population of approximately 140,000 inhabitants
(Billé, 2002), and is the largest town in southwest Madagascar. The region around
Toliara is one of the poorest in the country. The majority of the rural communities breed
cattle and poultry and cultivate crops that are adapted to the dry environment, mainly
manioc and maize.
The spiny forest eco-region is the least studied region in Madagascar, but the few
surveys that have been carried out suggest a very high diversity of animals (Nicoll &
Langrand, 1989; Raxworthy, 1995). Deforestation and habitat fragmentation are
especially prominent in the dry forests of the west and south, which are undergoing the
most rapid deforestation of any region in Madagascar (Sussman & Rakotozafy, 1994;
Dupuy & Moat, 1996; Smith, 1997). In fact, tropical dry forests are now considered to
be the most endangered of all major tropical ecosystems (Janzen, 1988). This is due to
extremely slow regeneration of the vegetation, caused by low rainfall, poor soils and the
rapid onset of erosion following land clearance for cultivation or grazing.
2.2 Location of the proposed Regional Park
The proposed Parc Regional de Belomotse is located along the Onilahy River, east of
Toliara, see figure 1 for regional map. The area is dominated by spiny forest on the
calcareous Mahafaly plateau. However, it has a core area containing lakes, marshes and
areas of riparian forest. This is called the Sept Lacs region, see figure 2. The Sept Lacs
region (S 23º 28' - S 23º 31', E 44º 04' – E 44º 10') covers over 3,000 ha between 40 and
250m asl, and is located along the north bank of the Onilahy River between the villages
of Antafoky and Ifanato. To the west of this core area is Antafoky Lake, with a large
area of associated riparian forests (S 23º 28', E 044º 03'). The lake covers approximately
150 ha, the gallery forest covers a further 150 ha.
5

Both sites are contained within the proposed boundaries of the Regional Park. The areas
surveyed by the Frontier-Madagascar Forest Research Programme include the core Sept
Lacs region, the Antafoky Lake area and the corridor of gallery forest along the Onilahy
River.
2.3 Local Communities
There are a number of villages within the park, representing several tribes. The small
village of Antafoky, located on the western border of the core area, contains villagers
that are mainly from the Masikoro tribe. Mahaleotse, in the central part of the park,
consists of three small settlements; Bevolavo, Manderano and Befas. The village of
Ifanato is located near the eastern border of the core area. Mahaleotse and Ifanato mainly
consist of people from the Mahafaly tribe.
Figure 1. Map of the Toliara region, showing the Onilahy River. (Reproduced from Lonely Planet, Madagascar
4 th Edition, May 2001.)
The villagers cultivate the surrounding land, the most common crops being rice, maize,
beans and manioc. In addition they herd livestock, and collect fish from the Onilahy
River and associated ponds. The villagers also hunt wild ducks and gather natural
products such as wood, reeds, medicinal plants and honey.
2.4 Site history
The site of the proposed Regional Park includes the last remaining fragments of
riparian forest along the northern bank of the Onilahy River (Raxworthy, 1995). In
order to preserve these habitats the Department of Eaux et Forets implemented a
system of local management, called Gestion Locale Securisee (GELOSE), in this area
in March 2000.
The principal aim of GELOSE was to provide the local communities with the authority
to manage and direct the sustainable use of the remaining forested areas. The regional
division of the National Environment Office (ONE) provides technical support in this
6

initiative. This regional department, called Service d’Appui a la Gestion de
l’Environment (SAGE), is working in conjunction with socio-economic advisors from
the WWF-Madagascar Dry Forest Programme.
Antafoky Lake and the associated riparian forests are not included within the GELOSE
management plan, although they contain large areas of riparian forest and a unique
wetland habitat.
2.5 Topography, Geology and Climate
The proposed Regional Park
lies at the western edge of the
Mahafaly Plateau, a limestone
formation that extends south
from Toliara to Cap Ste
Marie. This area borders the
north bank of the Onilahy
River. The lowest point has
an altitude of approximately
30m, rising to over 250m on
the plateau. The flanks of the
plateau are fairly gentle,
although there are many large
cliffs along the edge of the
river. Narrow, shallow valleys
extend into the hills.
The lake at Antafoky, the
Sept Lacs river system and
numerous other ponds in the
area are fed by a number of
underwater springs, with
water flowing continuously
throughout the year.
Figure 2. Map of the proposed Regional Park, outlined in red.
(Reproduced courtesy of WWF Toliara)
The region has acquired its shape due to the dominant rock type in this region,
carboniferous limestone. This has created large areas of comparatively high-standing
hills and plateau, interspersed with channels and gullies cut into the limestone and the
less erosion-resistant sandstone. A large limestone pavement can be found on the
plateau. Exposed subterraneous caverns can also be found in the region, along with
the remnants of stalagmites and stalactites, which are often found in underground cave
systems.
The local geology consists of well-bedded white sandstones, sub-rounded
conglomerates and bedded limestone. The sandstones contain a large proportion of
7

marine shell fragments (mostly coral stems, bivalves, and gastropods). The limestone
is less well bedded than the sandstone unit and is generally thicker (up to 5m). The
limestone contains an abundance of fossils, mainly gastropods, bivalves and
fragments of corals.
The climate is semi-arid, with an annual average rainfall of around 400 mm. Over
85% of this annual rainfall occurs during the rainy season, from November to March.
Average maximum temperatures vary from 26ºC in the dry season to 36ºC in the wet
season, occasionally reaching as high as 42ºC with a high humidity (Du Puy and
Moat, 1996).
2.6 Vegetation
The proposed site of the Regional Park comprises a vegetation mosaic of unique
habitats, including riparian and spiny forest. The riparian forest is limited to the
valleys, whereas the spiny forest dominates the hillsides and plateau. The riparian
forest, which occurs along the river system and around areas of wetland, has been the
most heavily disturbed by the activities of man.
In addition to the forest habitats, the proposed Regional Park contains large areas of
agricultural land and scrub. Agriculture is concentrated on the flat land along the river
edges and hilltops. Scrub areas are found where forest has been cleared around the
edge of cultivated areas, and on the plateau in areas that have been subjected to
burning.
2.7 Previous research
Southwest Madagascar has very high levels of floral and faunal diversity and
endemism. However, comparatively few biodiversity surveys have been undertaken in
the region, and very little reliable or comprehensive data have been collected
(Sussman & Rakotozafy, 1994).
Previous research in southwest Madagascar has focussed on the Beza Mahafaly
Special Reserve, the Kirindy Forest Reserve and the Mikea Forest region. There are
no published studies available for the proposed site of the Regional Park. The area has
only been formally surveyed on one occasion, during a short visit to the area by an
Earthwatch team (Raxworthy, 1995). Occasionally, students from the University of
Toliara have visited the area, but data are often inaccessible and unreliable.
8

3. The Study sites
Work was carried out in the survey areas over a period of 15 consecutive months from
July 2001 to September 2002. The surveys were divided into five periods of
fieldwork, the dates of which are given below:
Period 1 1 July to 6 September 2001
Period 2 4 October to 15 December 2001
Period 3 7 January to 20 March 2002
Period 4 4 April to 1 July 2002
Period 5 1 July to 6 September 2002
The surveys focussed on the Antafoky Lake area for fieldwork periods 1, 2 and 3.
During these three fieldwork periods, a significant proportion of the time was also
spent in the Sept Lacs region. During fieldwork period 4 the surveys were based on
the western side of the Sept Lacs region, around Manderano, see figure 3. Fieldwork
period 5 focussed on the central and eastern areas of the Sept Lacs region.
Figure 3. Map of the core region from Antafoky Lake to the Sept Lacs (produced based on GPS data collected by
the Frontier-Madagascar Forest research Programme).
3.1 Data Collection Methods
There are many methods for assessing zoological biodiversity. The methods used by
the Frontier-Madagascar Forest Research Programme are a combination of suitable
methodologies from comparative surveys in Madagascar, based on similar studies by
Frontier-Tanzania using over twelve years of biodiversity experience in coastal forests
and rainforest.
The methodologies and associated data-recording sheets were identified and compiled
through the following process:
9

A review of existing methodologies through a library search of references relating to past
biodiversity, ecological and behavioural surveys in the country and region.
The identification of methodologies that could be used, considering the environmental,
logistical and personnel constraints of the project and region.
The identification, through field trial phases, of methodologies that would be suitable.
The identification of an efficient system for data collection, storage and use.
The production of a series of accurate, simple data sheets and databases that enable
efficient data collection and utilisation.
The herpetological methodologies follow those of Raxworthy and Nussbaum (1994)
and the small mammal surveys follow those of Goodman et al. (1999). The butterfly
collection methodologies follow those of Kremen (1994) and the general invertebrate
collection methods follow those of Fisher (1999). The botanical surveys follow those
of Seddon et al (2000) and WWF methodologies (2002). The resource use surveys
follow the methods of Kremen et al (1999).
Behavioural and ecological studies were undertaken when the opportunity arose to
survey a particular, usually threatened, species. These surveys included a study of the
density and distribution of diurnal lemurs, following the methods used by Whitesides et
al. (1988) and Rabinowitz (1997), and the estimation of populations of two species of
bats, following the method used by Wilson et al. (1996).
Trapping could not be fully relied upon to provide a representative sample of species
from an area as not all species were targeted with the standardised methods. So,
opportunistic day and night searches were carried out to ensure the capture or
observation of the majority of species present.
See appendix 13 for a list of collaborators for specimen identification and deposition.
See Methodology Training Manual (Frontier Madagascar, 2003) for detailed
descriptions of the methods used.
3.2 Zoological trap sites
The biodiversity of the study area was assessed by applying a set of different methods
for different target animal groups. These included mammals, birds, reptiles, amphibians
and invertebrates. The set-up of a combination of the chosen methods in a certain place
for a set time-period was called a ‘trapsite’. The trapsites were typically established for
eight-day periods and consisted of a combination of methods including pitfall trapping,
Sherman and mesh trapping, bat collection using mist nets and harp traps, butterfly
collection using sweep-nets and canopy traps, invertebrate collection, and casual
collections and observations.
10

The trapsites were positioned such that all habitats in the study area were represented.
They were set up in suitable localities for specimen collection and observation. Ten of
the trap sites were undertaken during the dry season, one at the onset of the rains and six
during the rainy season. This optimised opportunities for collecting and observing all of
the chosen taxa.
The surveys focussed on nine trap-sites around Antafoky Lake and a further eight
trap-sites within the Sept Lacs region. See Appendix 1 and table 2 for trapsite location
descriptions.
Table 2. Description of the zoological trap sites.
Trap site Date and
(phase No.) duration
1 (013 1) 13-20 July
8 days
2 (013 2) 24-31 July
8 days
3 (013 3) 8-11 Aug.
4 days
4 (013 4) 17-24 Aug.
8 days
5 (014 1) 13-20 Oct.
8 days
6 (014 2) 23-30 Oct.
8 days
7 (014 3) 12-21 Nov.
10 days
8 (014 4) 25 Nov.-2
Dec. 8 days
9 (021 1) 21-28 Jan.
8 days
10 (021 2) 3-10 Feb.
8 days
Season Position Altitude Habitat description * Trap effort
Dry S 23º 28' 44.7"
E 44º 03' 58.0"
Dry S 23º 29' 16.0"
E 44º 04' 39.1"
Dry S 23º 28' 23.8"
E 44º 04' 22.8"
Dry S 23º 31' 41.9"
E 44º 09' 20.0"
Dry/wet S 23º 28' 44.7"
E 44º 03' 58.0"
Wet S 23º 29' 16.0"
E 44º 04' 39.1"
Wet S 23º 31' 29.2"
E 44º 09' 33.2"
Wet S 23º 31' 05.4"
E 44º 08' 18.0"
Wet S 23º 28' 44.7"
E 44º 03' 58.0"
Wet S 23º 29' 22.9"
E 44º 04' 43.1"
50 masl. Gallery forest, distance
to lake 1-60 m.
50-80 Gallery forest, distance
masl. to lake 1-70 m.
80-90 Thorny scrub, no water
masl. association
80-90 Gallery forest, distance
masl. to river 5-10 m.
50 masl. Gallery forest, distance
to water 10-100 m.
50 masl. Gallery forest, distance
to lake 2-70 m.
100-130 Transitional forest and
masl. thorny scrub
50 masl. Gallery forest, distance
to marsh 5-80 m.
50-80 Gallery and transitional
masl. forest.
50-250 Gallery forest,
masl. transitional scrub and
forest.
120 bucket nights
48 trap nights
120 bucket nights
48 trap nights
60 bucket nights
24 trap nights
120 bucket nights
48 trap nights
264 bucket nights
80 trap nights
264 bucket nights
96 trap nights
330 bucket nights
120 trap nights
264 bucket nights
96 trap nights
264 bucket nights
96 trap nights
264 bucket nights
96 trap nights
11

11 (021 3) 1-10 Mar.
10 days
Wet S 23º 31' 29.2"
E 44º 09' 33.2"
120-150
masl.
Gallery and transitional
forest
330 bucket nights
110 trap nights
12 (022 1) 4-11 May
8 days
Dry S 23º 31' 26.9"
E 44º 05' 34.3"
70-80
masl.
Gallery and transitional
forest.
264 bucket nights
48 trap nights
13 (022 2) 11-18 May
8 days
Dry S 23º 30' 36.4"
E 44º 04' 57.5"
60-70
masl.
Gallery forest
176 bucket nights
800 trap nights
14 (022 3) 22-29 May
8 days
Dry S 23º 28' 25.1"
E 44º 04' 24.7"
50-90
masl.
Gallery and transitional
forest
176 bucket nights
800 trap nights
15 (022 4) 3-10 June
8 days
Dry S 23º 29' 15.2"
E 44º 04' 43.2"
70 masl. Gallery forest 176 bucket nights
800 trap nights
16 (022 5)
16-23 June Dry S 23º 31' 26.9" 70-90 Gallery and transitional 176 bucket nights
8 days
E 44º 05' 34.3" masl. forest
800 trap nights
17 (023 1)
4-11 Aug. Dry S 23º 31' 15" 150-170 Gallery and transitional 264 bucket nights
8 days
E 44º 08' 35" masl. forest
48 trap nights
Total 136 days
3632 bucket nights
4158 trap nights
* See appendix 1 for detailed habitat descriptions of the trapsites.
3.3 Trapsite procedure
When the trapsites were assembled, and during the days of checking the traps, care was
taken not to disturb the area. This was to ensure that target species were not scared away
from the area. The sites were only visited when checking the traps.
The day of trapsite assembly was regarded as ‘day 0’. The following days were regarded
as ‘days 1 - 8’. A ‘trap night’ or ‘bucket night’ refers to the period of 24 hours during
which the traps and pitfall buckets were in use, from late afternoon after checking the
traps to the following afternoon before checking the traps. For example, eleven pitfall
buckets used for eight days would be described as 88 bucket nights.
The pitfall and box traps were checked and emptied in the morning between 06.00 and
08.00 hours, and the butterfly canopy traps were baited. Between 16.00 and 18.00 hours
in the afternoon the pitfall and butterfly traps were re-checked and the mammal traps
were opened and baited. When a trapsite was located in open exposed areas, the pitfall
traps were additionally checked at midday as the high temperatures could kill captured
specimens.
The target mammal species were mostly nocturnal, so the mammal box traps were open
throughout the night and closed during the day. This also prevented trapped animals
dying from heat exhaustion during the day.
The Sherman and mesh traps were baited with fish, fruit (mostly banana) or shrimps to
attract all possible mammal species. The butterfly canopy traps were baited with
fermented banana. Pitfall traps were not baited.
All captured individuals were taken to the base-camp for identification. Individuals that
were not preserved were returned immediately after identification and released near the
location of capture. Sherman traps were washed and returned to their previous position
after the capture of an animal.
12

4. Plants
4.1 Introduction
The Malagasy flora encompasses at least 107 families and 490 native genera of trees
and large shrubs, of which 161 genera are endemic to Madagascar and the Comoro
Islands (Schatz, 2000). The endemic tree genera are represented by 940 species. The
329 non-endemic tree genera are represented by 3,280 species, of which 95% are
endemic. In total, over 96% of the estimated 4,220 species of Malagasy trees and
large shrubs are endemic (Schatz, 2001).
Three families – Euphorbiaceae, Fabaceae and Rubiaceae – account for almost one
third of the genera of Malagasy trees and large shrubs. Of the remaining families, 52
are represented by only a single genus of trees in Madagascar, of which 14 are
endemic. There are seven tree families endemic to Madagascar – Asteropeiaceae,
Didymelaceae, Kaliphoraceae, Melanophyllaceae, Physenaceae, Sarcolaenaceae and
Sphaerosepalaceae. With over 700 species, the tree family Rubiaceae is the second
largest plant family in Madagascar, after the family Orchidaceae.
Southwest Madagascar has the highest percentage plant endemism of all domains in
Madagascar. However, in the arid regions of southern Madagascar, no detailed study
of floristic composition exists (Sussman & Rakotozafy, 1994). As a prerequisite to
conservation and monitoring, baseline data on forest composition are vital.
4.2 Methods
Sixteen quadrats were laid out randomly along a series of 500 m-long transects that
were placed through either gallery or spiny forest. The locations of these transects
were selected because two distinct plant communities were recognised and the study
was attempting to identify indicative species from each habitat and measure the
differences between them. Within the studied areas, the transects were positioned to
pass through representative forest. Representative gallery and spiny forest habitats
were identified based on a number of factors. For example, it was assumed that
comparatively undisturbed forests near water, with high canopy and Tamarindus
indica as the dominant species in a complex plant community, could be defined as
gallery forest. These forests also typically showed a strong collaboration with deep,
moist soils on fairly shallow gradients. By comparison, low canopy forests on the dry
calcareous hills and plateau could be defined as spiny forest. These forests typically
contained xerophytic, spiny plant communities on dry soils.
Two quadrats were assigned to each transect. Quadrats were placed at random along
each transect by generating random numbers and using them to correspond to the
appropriate distance along the transect. The quadrat areas were calculated using
minimum area curves. This entailed establishing a 1m x 1m quadrat and identifying
the number of different species that were present. The quadrat area was then doubled
and the number of new species recorded. This was continued to produce a species
accumulation curve. When the curve levelled off, the suitable quadrat size was
achieved. For gallery forest, this quadrat size was 16 m x 16 m. For spiny forest, the
quadrat size was 32 m x 32 m.
13

All trees, shrubs and stem herbs that were rooted within each quadrat were measured
and the Diameter at Breast Height (DBH) was recorded. Each plant was assigned to a
DBH size class (< 2.5 cm, 2.5 - 5 cm, 5 - 10 cm, 10 - 25 cm, >25 cm) and plants
within each size class were combined for analysis. The DBH was taken for each plant
to help assess the current state of the habitat and to predict previous forest use. The
DBH measurements were also used to assess the effect of species-specific plant use
on chosen trees and shrubs, compared to other tree species present within the same
quadrat. Thus, the DBH data could be used to help show previous resource use within
the area, and could indicate whether villagers selectively used certain plants or sizeclasses
of plants.
Plant identifications were made in the field. Problematic individuals that could not be
identified had their vernacular name taken for future identification. Fanaterany
Manantiry, a local village elder with extensive knowledge of the regional flora,
provided the vernacular names and also gave information regarding local plant uses.
Further information on plant use, both by man and by local fauna, was obtained from
Schatz (2001).
The plants within the quadrats were not marked or tagged. However, the co-ordinates
of the individual quadrats were recorded using GPS. The survey was intended as a
species inventory, as plant data for the region are scarce. At present, regional
emphasis is on the identification of species and habitats. However, using the GPS
points it will be possible to return to each quadrat in the future to monitor the state of
the plant communities using the species-specific size data and designated DBH sizeclasses.
4.3 Results
A total of 3,880 plants were sampled, from which 142 different tree and large shrub
species were identified. Family and genera names follow the classifications given by
Schatz (2001). The vernacular names of all specimens were also recorded (see
appendices 2 and 3). Local uses of plants are also listed (appendices 4 and 5).
4.3.1 Gallery forest
Thirty families of plants were represented over eight 16 x 16 m quadrats, of which
plants from twenty-five families were classified as trees or large shrubs (see table 3).
The family Euphorbiaceae had the most species (7), followed by Fabaceae (6) and
Rubiaceae (5). Two of the families present, Physenaceae and Sphaerosepalaceae, are
endemic to Madagascar.
A total of forty-one tree and large shrub genera were identified in the gallery forests,
seven (17%) of which are endemic. The total number of species that were collected
was sixty-one, with forty-six being identified and named to the species level. The
flora of Madagascar is too poorly known to accurately assess the level of species
endemism at this time.
14

Table 3. The number and endemism of tree and large shrub families, genera and species per habitat.
Gallery
forest
Spiny
forest
Family Genus Species Unidentified
Number Endemic Number Endemic Number Endemic Number
25 2 41 7 46 - 15
31 2 60 12 86 - 32
Throughout all surveyed areas of gallery forest, the dominant large (DBH > 25 cm)
tree species was Tamarindus indica (see table 4). This species formed a closed canopy
15-25 m in height in undisturbed areas. The tree species that was dominant through
the lower size ranges was Celtis bifida. Other common gallery forest species in the
surveyed areas included Diospyros sakalavarum and Olax lanceolata.
Table 4. Number of individuals of the two most common tree species in each size class and habitat
Riparian Forest quadrats
Spiny Forest quadrats
Species 25cm 25cm
ANACARDIACEAE
Operculicarya decaryi 16 6 11 26
BURSERACEAE
Commiphora arafy 2 5 10 8 3
Commiphora sp1 1 2 27 21 14 16 2
Commiphora sp2 43 26 21 3
CELTIDACEAE
Celtis bifida 99 47 15 9 3
EBENACEAE
Diospyros latispatula 100 44 17
Diospyros sakalavarum 68 23 12 2 13 5
EUPHORBIACEAE
Croton sp.3 178 2
Securinega seyrigii 1 45 32 14 5
FABACEAE
Alantsilodendron sp. 49 53 37 14
Tamarindus indica 5 2 20
MALVACEAE
Grewia grevei 1 140 19 3
OLACACEAE
Olax lanceolata 37 26 30 6 1 1 1
4.3.2 Spiny forest
Thirty-nine families of plants were represented over eight 32 x 32 m quadrats, of
which plants from thirty-one families were classified as trees or large shrubs. The
15

families Euphorbiaceae and Fabaceae had the most species (14 each), followed by
Malvaceae (7) and Burseraceae (6). Two of the families present, Physenaceae and
Sphaerosepalaceae, are endemic to Madagascar.
A total of sixty tree and large shrub genera were identified in the spiny forests, twelve
(20%) of which are endemic (see table 3). The total number of species that were
collected was 118, with 86 being identified and named to the species level. As with
the gallery forest species, endemism at the species level is not included.
Throughout all surveyed areas of spiny forest, there were few large (DBH > 25 cm)
tree species. The upper size ranges were dominated by Antsilodendron sp.,
Commiphora spp. and Operculicarya decaryi. These species formed an open, low
canopy 3-10 m in height in undisturbed areas. The canopy height depended on the
gradient and substrate of the area. The tree species that were dominant through the
lower size ranges were Croton sp. 3, Grewia grevia and Diospyros latispatula.
4.4 Discussion
The most significant achievement of the botanical survey was the production of a
preliminary species list for the two main natural habitats within the proposed regional
park. Few inventory studies have been undertaken in the region, so this species list
can be used for comparative studies and monitoring when further studies have been
completed in other areas. At present, the analyses of survey data are awaiting the
completion of comparative studies, in order to allow comparisons to be made and
specific regional habitats to be identified. Therefore, this discussion will focus on
forest structure, dominant species, resource use and threats to the spiny and gallery
forests.
4.4.1 Dominant species and forest structure
The two most dominant tree and shrub species within each size class and habitat are
shown in Table 4. The data show clearly that there are very few dominant habitatgeneralist
species. By that, it is meant that the dominant species within the spiny
forests are almost entirely absent in the gallery forests, and vice versa. This seems to
indicate that, within the region, habitats that may be only metres apart are clearly very
different in terms of plant dominance and forest structure. It also indicates the habitatspecific
diversity within the spiny and gallery forests of southwest Madagascar.
The results in Table 4 also show that in gallery forest, a small number of species are
dominant, but they are dominant through more than one size class. For example,
Celtis bifida was one of the two most dominant plant species within every size class.
This leads to the ground layer, shrub layer and canopy layer of gallery forest being
dominated by a small number of species.
In contrast, the spiny forests show a much greater diversity of dominant species
between the different size classes. None of the plant species that are dominant in the
lower size classes are even found in the upper size classes. This reflects the diversity
of species that are found in the ground layer, shrub layer and canopy layer, and shows
the level of niche-specificity that spiny forest plant species exhibit.
16

These results seem to show that the spiny forests contain a high diversity of plant
species within a complex community. The gallery forests seem to show a lesser
degree of plant diversity, with a more simple plant community dominated by
relatively fewer species.
4.4.2 Plant-resource use
Appendices 4 and 5 show some of the uses of specific tree and shrub species, both by
man and by specific fauna. As can be seen, trees from gallery forest that are typically
used for construction purposes by man were found in low numbers and only in the
lower size classes. For example, Bivinia jalberti and Colvillea racemosa, both of
which are used for the construction of houses and pirogues, were found in low
numbers and only as saplings. Conversely, many of the trees from the spiny forest that
could be used for construction purposes were found in fairly high numbers, often as
large trees. It therefore appears that specific tree species of a particular size are being
targetted and removed from areas of gallery forest for the purposes of construction,
leading to the loss of large trees from this habitat. However, the effect of speciesspecific
tree removal from areas of spiny forest appears to be comparatively less. This
may be due to the relative difficulties of access to areas of spiny forest, or it may
reflect a preference for the larger trees found within gallery forest. It may also be that
the surveyed areas of spiny forest were not representative.
Many of the plants listed in appendices 4 and 5 are used by man for purposes other
than construction. For most of these plant species, such purposes either require large
trees, such as for the collection of edible fruits, or, the plant is used in a nondestructive
way, such as the collection of leaves for medicinal purposes. The results
appear to show comparatively large numbers of tree and shrub species of varying size
classes that are utilised for resources other than wood. Without comparative data, it is
not possible to state whether these species have been adversely affected by local use.
However, there appear to be significant numbers of these plants, in all size ranges,
both in gallery forest and spiny forest. For example, Tamarindus indica produces
fruits that are collected and eaten or sold by local villagers. These trees are dominant
in the upper size ranges in all surveyed areas. It therefore appears that, based on
current data, the non-destructive use of certain plants has had a negligable negative
effect on the numbers of these plants within their habitat. It may even be that, in
certain circumstances, specific plants may benefit from their ‘usefulness’ to local
villages.
The fruits of some tree species are edible to lemurs. In particular, Tamarindus indica,
Ficus spp. and Diospyros spp. are important food sources for lemurs within the
region. Indeed, Ficus spp. and Tamarindus indica could be classified as keystone
species, such is their importance to the regional fauna. These species are found in
comparatively large numbers, often as adult trees. This is partially due to the use of
their fruits by man, and partly due to cultural beliefs. Tamarindus indica and Ficus
spp. have important spiritual significance in southwest Madagascar; in many areas
large trees of these species are sacred. This may partially explain the dominance of
these tree species in the upper size classes within the areas of gallery forest in the
region, along with the apparent low incidence of clearance of adult trees of these
17

species for timber. It is fortunate, or perhaps well judged, that the two keystone
species within gallery forest are provided with such protection.
4.4.3 Threats to habitats
The areas of spiny forest and gallery forest are very different in terms of plant
composition and structure, but the threats to both of these habitats are similar. Large
areas of gallery and spiny forest have been cleared cultivation, predominantly maize
and cassava. Many trees have been felled for timber, and areas have been cleared for
the production of charcoal. Grazing of cleared areas has prevented regeneration.
These destructive uses of plant resources have led to the clearance of almost all
significant areas of gallery forest along the south bank of the Onilahy River, mainly
for cultivation purposes. The gallery forests contained within the surveyed areas
represent the last significant forested areas along the Onilahy River. On the hillsides,
large areas of spiny forest have been cleared for cultivation and charcoal production.
Within all surveyed gallery forest quadrats, there was clear evidence of previous
forest clearance for the purposes of cultivation or charcoal production. The gallery
forest quadrats around Lake Antafoky contained evidence of fairly recent fire damage
and slash-and-burn activities. In fact, the area was observed in the process of
clearance in September 2001. The need for agricultural land is great around Lake
Antafoky, so the threat to the areas of gallery forest in this area is severe. The gallery
forest quadrats to the west of the Sept Lacs region (around the village of Mahaleotse)
showed evidence of past use for small-scale charcoal production. However, the areas
that were cleared for charcoal production appeared to be fairly small and old, and
regeneration was evident. Unfortunately, goats were regularly seen in the gallery
forests throughout the surveyed area, which will have affected the forest structure and
reduced the rate of regeneration considerably.
The areas of spiny forest around Lake Antafoky showed comparatively low levels of
disturbance within the quadrats and on the hillsides, although large areas of cleared
forest were visible on the hilltops. These areas had been cleared several years
previously for the cultivation of maize. The spiny forest within the Sept Lacs region
was fairly intact within the surveyed quadrats and on the hillsides, but an area of
cleared forest was located on the plateau. This area was extremely large, extending for
several kilometres across the plateau, and had been cleared primarily for maize
cultivation.
Taking into account the relative areas both of the spiny forest and gallery forest, it is
evident that the habitat of primary concern is gallery forest, particularly around Lake
Antafoky. The remaining areas of gallery forest within the region are fairly small and
are gradually becoming degraded through selective logging and localised clearance.
The structure and quality of these forests are therefore affected, and they are
becoming fragmented. In comparison, the spiny forests within the surveyed region
show large areas of relatively intact forest on the hillsides, although clearance for
cultivation has severely impacted the spiny forest in large areas across the plateau.
18

5. Mammals
5.1 Introduction
Mammals occur in a wide variety of terrestrial and aquatic habitats throughout the world,
and constitute an essential aspect of the fauna in Madagascar. Because of the long period
of isolation of Madagascar from other landmasses, the mammal fauna has evolved
separately and created species and communities endemic to Madagascar (Hutcheon,
1994; Cook et al, 1991). As with many islands, Madagascar lacks many familiar higher
taxa such as the wild cats, dogs, monkeys, apes and the large herbivores that dominate
the plains of mainland Africa (Cook et al, 1991; Garbutt, 1999). Only six orders are
present on Madagascar - Chiroptera, Insectivora, Rodentia, Carnivora, Primates and
Artiodactyla. Of all the native species, of which more than 90 % are endemic to the
island, the best known are the primate order of lemurs and insectivore order of tenrecs.
Aside from the charismatic lemurs, most mammal species are poorly studied, indeed
the suborder of microchiroptera are one of the most diverse orders and yet the least
studied (Stephenson, 1994; Goodman, 1996). There have been no comprehensive
small mammal studies in the region of the proposed Parc Regional de Belomotse. A
short survey was conducted in this area (Raxworthy, 1995), but no mammals were
captured. No other reported mammal studies have been undertaken in the region prior
to this survey, so the mammal fauna is comparatively un-studied and unknown.
Small mammals have important roles in their ecosystems because they are central to the
food chain. They feed on vegetation and invertebrates and are preyed upon by
carnivorous mammals, reptiles and birds. Furthermore, mammals often function as seed
dispersers and are therefore important for plant species distribution and regeneration.
Small mammals are comparatively easy to catch, handle, and identify. They have
short life spans so changes in diversity, density, and community composition over
time are relatively easy to detect and monitor. Small mammals are a useful ecological
indicator group. Data from repeated surveys can be used to monitor the level of
disturbance to an area and thereby contribute to the management of these areas.
As with most other groups of animals in Madagascar, the mammals are threatened by
habitat destruction, excessive hunting and competition by introduced species.
5.2 Methods
As mammals have adapted to many different habitats, several methods are necessary to
reveal the species occurring within an area.
5.2.1 Pitfall trapping
Small insectivores and rodents were captured using pitfall traps. Pitfall traps are a cheap
and effective method, used to sample ground dwelling small mammal, reptile and
amphibian species. The method consists of buckets sunk into the ground along a
plastic sheeting drift fence, that channels the animals into the buckets (commonly
called bucket lines). The buckets were 275 mm deep, with a top diameter of 290 mm.
Three 100-metre long bucket lines, each with 11 buckets, were used at every trapsite.
19

5.2.2 Sherman and mesh trapping
Rodents, large insectivores and mouse lemurs were captured using either Sherman or
mesh traps. The Sherman traps were of the standard size, 225 mm x 87 mm x 75 mm.
The mesh traps measured 400 mm x 150 mm x 150 mm. The traps were placed in
suitable habitats at each trapsite, typically with two of each type of trap near each pitfall
line. A total of six Sherman traps and six mesh traps were used at each trapsite apart
from trapsites 13-16, when 80 Sherman traps and 20 mesh traps were used. Suitable
locations for placing the traps were in areas with a dense understorey, along fallen and
rotting logs, or near obvious food sources such as fruiting trees. Traps were situated as
far away from the pitfall lines as possible, with a minimal distance of 10 metres. They
were typically positioned at least 50 m apart. For the capture of arboreal species, traps
were fastened with tape to a suitable branch or vine.
5.2.3 Observations, tracks and calls
Diurnal and nocturnal observations, tracks, dung and vocalisations were used to identify
mammal species that could not be trapped. Specimens were photographed wherever
possible to provide confirmation of species identification. Identifications of species
through photographs, tracks and calls were made using the book ‘Mammals of
Madagascar’ (Garbutt, 1999).
5.2.4 Mist nets and harp traps
Microchiroptera were captured using mist nets and harp traps. The traps were set in
suitable locations along flight-paths and around roosts or feeding sites. Ideal places were
across forest clearings, forest edges, paths, roads, streams and ponds. Mist nets were
placed in enclosed areas with overhanging and surrounding vegetation whenever
possible, increasing the capture rate by channelling the bats into the nets. The mist net
composition depended on the location and habitat. The number of mist nets used varied
depending on the habitat, but typically consisted of three or four nets. Short nets were
suitable in dense forest and on narrow paths and roads, whereas long nets were suitable
for open spaces and over water. Harp traps were ideal for trapping around roosts where
there were a large number of bats, as it was much easier and more efficient to collect
large numbers of individuals in a harp trap than to untangle them from a mist net. Harp
traps used in conjunction with mist nets ensured a high probability of capturing the
majority of species that were present. Some bats learn to avoid the mist nets and harp
traps, so trapping was undertaken for no more than 2-3 days at each location.
5.2.5 Owl pellets
Owl pellets were collected. Most owls feed partly on small mammals. They usually
swallow their prey whole, and after digesting the flesh, the indigestible fur and bones are
regurgitated as compact pellets. Owl pellets can be broken apart and analysed for prey
species. Small mammals can be readily identified from their skulls, which are fairly
intact in the pellets. The analysis of owl pellets can contribute to identifying the small
mammal species present within a locality. Owl pellet analysis is particularly useful for
detecting the presence of species that are rare or not readily collected with the used
methods. Useful owl species include the Madagascar long-eared owl, Asio
20

madagascariensis and the Barn Owl, Tyto alba. Both species feed mainly on mammals
and return after foraging to their roosting sites, where they digest and regurgitate pellets.
Owl pellets were collected frequently and regularly during the survey period, so the
data could be used to identify seasonal variation in small mammal populations.
5.2.6 Specimen handling
Captured specimens were measured, weighed and marked for individual recognition.
The sex and breeding status of the captured individuals was also recorded, based on
visible sexual characteristics. Specimens were defined as juvenile, sub-adult or adult,
depending on their size, weight and level of sexual maturity.
For a detailed study of Rattus rattus, captured specimens were allocated a specific age
class. A juvenile was defined as having a head-body length of less than 100 mm and a
weight of less than 50 g. A sub-adult was defined as having a head-body length between
100 mm and 150 mm and a weight of less than 100 g, whereas an adult was defined as
having a head-body length of over 150 mm and a weight of over 100 g.
See Frontier-Madagascar Forest Research Programme Methodology Training Manual
(Frontier Madagascar, 2003) for detailed descriptions of all methodologies used.
5.3 Results
Seventeen mammal species, representing 4 orders and 6 families, were captured
during a total of 3,632 pitfall nights, 4,158 trap nights, and 1,829 mist net hours. A
further 9 species, representing a further 3 orders and 5 families, were identified either
from direct observation, or from tracks and dung. In total, 26 species representing 7
orders and 11 families were found in the surveyed areas during the study.
Classifications follow Garbutt (1999).
The orders represented include Macrochiroptera (1 species), Microchiroptera (7
species), Insectivora (6 species), Rodentia (3 species), Carnivora (2 species),
Artiodactyla (1 species) and Primates (6 species). See Appendix 6 for species list.
5.3.1 Trapping success
Sherman and mesh traps traps proved most successful in the dry season, during the
months of July and August, with capture rates of 12.5% and 11% respectively.
Conversely, the capture rate of mammals in pitfall traps was highest in the wet season,
during the months of January and February, with capture rates of 18% and 10%
respectively. Mist nets and harp traps were successful throughout the year, with
gallery forest habitats proving most successful. The spiny forest habitats were
extremely damaging to the mist nets and the bats were dispersed over a wide area,
whereas the gallery forest canopy provided a more suitable environment for trapping.
Traps placed across water proved to be successful, as did trapping near to roost sites
(see Section 6).
The number of rodents captured was highest in July (24 captures), whereas the number
of insectivores captured was highest in January (50 captures). Rodents were most
frequently caught in the dry season (June to August) where they made up 89-100% of
21

the total live-trap captures (see Figure 4). Conversely, insectivores were most
frequently caught in the wet season (January and February) when they made up 96-
100% of the total captures.
The most frequently captured mammal species throughout the survey was Rattus
rattus, with thirty-three caught individuals from an accumulated total of fifty captures.
Composition and annual patterns in the surveyed populations could therefore be
detected. Of the 19 adult individuals captured, 10 were males and 9 were females,
indicating an even sexual distribution in the population.
Annual mammal trapping
No. of captures
60
50
40
30
20
10
0
July
August
October
November
January
February
Month
March
May
June
Figure 4. Annual trapping of rodents and insectivores
120
100
80
60
40
20
0
% of all captures
rodent
captures
Insectivore
captures
% rodents of
all captures
%
insectivores
of all
captures
5.3.2 Rattus rattus population composition
In October, most captured individuals of R. rattus were breeding adults, whereas in
May most captured individuals were juveniles (see Figure 5). This indicated that the
majority of the mature population mated around the onset of the rainy season. As food
is most abundant during the wet season, this would provide the necessary resources
for developing juveniles. A small number of breeding adults and sub-adults were
captured throughout the survey. This indicates that, although there is a clear ‘peak
season’, this species does continue to breed throughout the year.
Annual R. rattus population composition
12
10
No. of
individuals
8
6
4
2
Adult
breeding
Adult
Sub-adult
0
July/August October November May June
Month
Juvenile
Figure 5. Annual composition of the Rattus rattus population.
22

5.3.3 Observations, tracks and calls
Observations, tracks and vocalisations accounted for the identification of nine species
that were not captured using standard trapping methods (see Appendix 6 for details).
This included five species of lemur. Primates are protected species, and specific
permission is necessary in order for specimens to be collected. Therefore,
identifications based on observations are most suitable.
The bush pig, Potamochoerus larvatus; the Indian civet, Vivvericula indica; and the
fosa, Cryptoprocta ferox, are large, shy nocturnal animals who are rarely trapped.
Tracks were seen, as were droppings. The Indian civet was sighted on a small number
of occasions.
5.3.4 Owl pellet analysis
The analysis of owl pellets is a time-consuming process and at the time of writing this
report the analyses have not been completed. The results will be added in a later
version of this report.
5.4 Discussion
5.4.1 Ecological distribution and species richness
Many of the plants and animals of Madagascar display an exceptionally high level of
endemism, and the mammals are no exception. Of the twenty-six species found in the
surveyed areas, twenty-one species (81%) were endemic. The infraorder
Lemuriformes, of the order Primate, are endemic to Madagascar. Six lemur species
were identified during the surveys. Five species that were present in the surveyed
areas are not endemic to the country, at least three of which are thought to have been
introduced by human colonists. Although this is a fairly low number, these species
can damage the native ecosystems and are seen as pests in Madagascar.
Compared to other tropical and sub tropical regions throughout the world, the
mammals of Madagascar show low diversity (Lundahl & Olsson, 2000). Also, the
humid areas of the east coast show a higher diversity of mammals than the arid areas
of the west. However, the surveyed areas within the proposed Parc Regional de
Belomotse showed high diversity compared to the findings of other surveys in similar
habitats (see Raxworthy, 1995). The high number of mammal species found in the
survey area may be due to the diversity of habitats in the surveyed areas. It may also
reflect the fact that the areas were surveyed over a long period of time, covering both
the wet and the dry season, so the probability of catching and observing many
different species was high.
The order of chiroptera is the most diverse of Madagascar’s small mammal fauna. The
taxonomy for many genera and species is still being revised and new species and
range extensions are frequently being found. During the survey, seven species of
microchiroptera were identified, two of which are undescribed species. Three species
of megachiroptera occur in Madagascar. The two largest species, Pteropus rufus and
Eidolon dupreanum, have previously been found in the southwest. During this survey,
fruit bats were observed roosting in caves or overhangs on cliff faces, and feeding in
23

fruiting trees at night. The specimens appeared to be Eidolon dupreanum. As no
specimens were captured, this identification is based on the fact that E. dupreanum
typically roosts in caves, whereas Pteropus rufus typically roosts in trees. Roosting
colonies seemed to vary in size depending on the availability and capacity of suitable
caves. Colonies with as few as ten individuals were observed. The roosting animals
were often semi-active and noisy throughout the day, emerging at dusk.
All species of the larger tenrecs; Tenrec ecaudatus, Setifer setosus and Echinops
telfairi, are known from southwest Madagascar. All of these species were collected
from the surveyed areas. T. ecaudatus was heard foraging in the undergrowth at night
and young individuals were caught in pitfall traps. This species was found both in
gallery forest near to water and in the arid spiny forest on the hills.
Large-eared tenrecs and shrew tenrecs, belonging to the genera Geogale and
Microgale consecutively, are known from southwest Madagascar. Geogale aurita is
restricted to the southwest region where it is typically found in dry deciduous and
spiny forests. The genus Microgale contains approximately twenty species that are
found in many different habitats throughout the island. During the surveys, G. aurita
was captured in low numbers in the dry spiny forest on the hillsides and plateau. One
species of Microgale was found during the survey. See section 5.4.4 for a discussion
of this species. M. pusilla has previously been collected from the Toliara region
(Garbutt 1999), though it was not identified in the surveyed areas. This species is
typically found in comparatively high numbers in suitable habitats, so it is suggested
that this species does not occur in the surveyed area.
Four species from the family Cheirogaleidae (Mouse and Dwarf Lemurs) were
identified within the surveyed areas. The genus of mouse lemurs, Microcebus, was
previously assumed to consist of two species, with one of them occurring in the west,
but recent studies suggest that at least seven species are found in western Madagascar
(Rasoloarison et al. 2000). Two species, Microcebus murinus and M. griseorufus,
were identified during the survey based on observations, DNA samples and the
preliminary analysis of owl pellets (Goodman, pers. comm). M. murinus is distributed
along the west coast of Madagascar, where it appears to be common and
comparatively abundant. During the survey, individuals were typically observed in the
dry transitional scrub and spiny forest. M. griseorufus was observed in similar
habitats.
The fat-tailed dwarf lemur, Cheirogaleus medius, is distributed along the west coast
of Madagascar in dry deciduous forest and gallery forest. During the survey, it was
observed frequently between the months of October and March. This species was
fairly inactive during the dry season from May to September, only occasionally
emerging. C. medius appears to be comparatively abundant, although it was only
observed in gallery forest habitats. Coquerel’s dwarf lemur, Mirza coquereli, is
known from the north-west of Madagascar in the Sambirano region, and also from
several locations in the west. Previous unconfirmed sightings suggested the existence
of populations south of this range, along the north bank of the Onilahy River. The
species was positively identified in this area during the surveys. The presence of M.
coquereli in this area probably defines the most southern limit to the range of this
species, as no sightings have been made south of the Onilahy River. In comparison to
24

Cheirogaleus medius, the Coquerel’s dwarf lemur was not seen to reduce its activity
levels during the dry season. It was observed all year round, though only in gallery
forest habitats.
Verreaux’s sifaka, Propithecus verreauxi verreauxi, and the ring-tailed lemur, Lemur
catta, were found in large numbers throughout the surveyed areas. Both species have
been recorded in the Toliara region and were known to be present in the Sept Lacs
region. They were typically found in gallery forest and spiny forest. However, all
observed groups of both species appeared to have the core area of their home ranges
in the galley forest. Both species spent much more time in the gallery forest during the
dry season than during the wet season, when they often foraged in the spiny forest on
the hillsides during the day. The two species co-existed in the same areas, often
occupying the same trees.
The elusive fosa, Cryptoprocta ferox, is distributed throughout much of the island in
many different habitats. It is very shy and difficult to observe. No confirmed direct
observations were made of the fosa during the survey. However, on several occasions,
unconfirmed sightings were made of a large animal by field staff and volunteers.
Also, dung was found containing the fur and spines of Tenrec ecaudatus. This species
of tenrec could only have been eaten by either C. ferox or a large dog (S. Goodman,
pers. comm), although domestic dogs were not seen in the remote location where the
dung was found. Furthermore, local villagers have pointed the species out from
photos, stating that the species has been seen around their village at night. It is
therefore tentatively assumed that the fosa is present in the surveyed area.
5.4.2 Introduced Species
The impact of introduced species on the native wildlife of Madagascar is not yet fully
understood, and although relatively few mammal species have been introduced to
Madagascar, further studies are needed to identify the direct and indirect effects that
these species are having on populations of native species.
Two introduced species from the sub-family Murinae were found during the survey,
Rattus rattus and Mus musculus. Both were found in high densities. Both species were
probably introduced to Madagascar within the last 2,000 years, since man arrived on
Madagascar. Being generalist and opportunistic, both species appear to have adapted
well to the new surroundings. Both R. rattus and M. musculus breed rapidly and
populations quickly reach high densities. These density levels cause extensive
resource competition with those native species occupying the same guilds. Without
exception, R. rattus now inhabits all habitats on the island, even penetrating into the
most dense forested mountains far from human habitation (Goodman, 1995). Previous
surveys revealed low numbers of R. rattus in the Toliara region (Raxworthy, 1995).
This led to the conclusion that R. rattus may not be as abundant in native habitats in
this region as elsewhere in Madagascar. However, during this study R. rattus was
caught frequently in all habitats surveyed, showing that the species is very abundant
and widespread. Furthermore, it was frequently observed around the campsites near
tents and food sources. It is well known by the local people, who consider it to be a
pest.
25

Rattus rattus has been shown to be a harmful pest, bringing several diseases to
Madagascar. One of these diseases, plague, which is carried by the fleas of R. rattus,
only occurred on the island after the introduction of rats. Today, the disease is still a
problem, affecting approximately 800 humans every year (S. Goodman, pers. comm.,
2002). Furthermore, other parasites and viruses carried by this animal have caused
severe problems for native rodent species. With little immunity to these exotic
diseases, populations have decreased and several species may have been driven close
to extinction as a result. The endemic rodent Eliurus myoxinus was found in low
numbers in the surveyed sites compared to the introduced rodents R. rattus and Mus
musculus. Comparatively little is known about the populations of E. myoxinus in the
region, and further surveys need to clarify whether the low density observed is normal
for the species, or whether it is suffering from competition with the introduced
species.
The pygmy musk shrew, Suncus madagascariensis, from the family Soricidae, has
been found during the survey. This species is the world’s smallest terrestrial mammal,
with a head-body length of 50- 55 mm and a weight of 3-4 g (Garbutt, 1999).
However, the individuals that were captured during this survey were all considerably
smaller, with a head-body length of 35-40 mm and a weight of approximately 2 g.
The endemic status of this species is uncertain, but it is currently assumed that the
populations on Madagascar are introduced (S. Goodman, pers comm). This species
was commonly found both in gallery and spiny forest habitats, but the impact of the
species on native species of rodents and tenrecs is not known. Being insectivorous, it
may compete with other small insectivores for food.
The introduced ungulate Potamochoerus larvatus larvatus occurs along the west coast
of Madagascar. Spoors were found both in marshy and in forested areas within the
surveyed areas. This species poses a threat to some native species in these habitats. It
forages vigorously in the undergrowth, digging up the soil as it feeds on roots, fruits,
invertebrates, vertebrates and bird eggs. It is also known to feed on tortoise eggs.
The introduced civet, Viverricula indica, is potentially harmful as it preys upon a wide
variety of native animal species, particularly small vertebrates. V. indica was observed
in several locations throughout the surveyed areas, often near to villages.
5.4.3 Seasonal variation
Due to the arid climate in the region, some animal groups tended to lower their
activity during the dry season, when much of the vegetation dried out and food
sources became scarce. However, rodents appeared to stay active all year round and
were observed regularly throughout the year in all surveyed areas. They were trapped
in higher numbers during the dry season than during the wet season, see figure 4.
Being omnivorous and opportunistic in their feeding behaviour, rodents can survive
on a wide variety of food items, e.g. fruits, seeds and invertebrates, which will allow
them to be active at times where other groups are inactive. Stephenson (1994) found
the same levels of activity in a year-round trapping survey in central Madagascar. He
suggested that due to the abundance of ‘natural’ food sources in the wet season, the
trap bait seem less attractive to the animals, so trapping success for rodents is reduced
at this time. This hypothesis explains the high trap success during the dry season in
the present survey, with the accompanying dip in trapping during the wet season.
26

However, only two year-round trapping surveys have been carried out in Madagascar
and more long-term studies are needed.
Insectivores were mainly caught during the wet season, with almost total absence
during the dry season (July and August). During these dry periods, insectivores lower
their metabolism and enter a state of torpor, hiding in burrows, caves, fallen trees and
other such sites. Insectivores are highly dependent on the availability of insects and
other invertebrates as a food source. During the dry season the temperature can be as
low as 25º C during the day and below 10º C at night. Combined with the lack of
precipitation, this leads to a seasonal climate that is unfavourable for invertebrates.
The numbers of invertebrates during the dry months are therefore low and may not be
sufficient for the insectivorous species.
Although little is known about the ecology and behaviour of the musk shrews Suncus
murinus and S. madagascariensis, they are thought to be active and breed throughout
the year (Garbutt, 1999; S. Goodman, pers. comm., 2002). During this survey,
trapsites were repeated in similar locations during both the dry and wet seasons. S.
madagascariensis was the only insectivorous species collected during the months of
June and July, though only in small numbers. This species was only captured in July
or August on one occasion (after light showers), whereas between October and March
it was captured in comparatively high numbers. This indicates that the species has a
longer activity period than the other collected insectivore species, and also that it does
not hibernate for a long period, if at all.
5.4.4 Rare species and range extensions
Two species of Triaenops bats, Triaenops furculus and T. rufus, were found in the
surveyed area. Both have previously been found in lowland areas of northern and
western Madagascar (Garbutt, 1999). The identification of both of these species
within the region of the proposed Parc Regional de Belomotse, combined with nonconfirmed
sightings further south, suggest that the ranges of these species are greater
than previously assumed. T. furculus is regarded as vulnerable, with a decreasing
population (IUCN, 2002), so the presence of a large population of this species within
the survey area is of high importance to the conservation of the species, see section 6.
A species of insectivorous bat from the genus Miniopterus was captured from a
number of locations throughout the surveyed areas. This species has characteristics
unlike any others from this genus and is thought to be an undescribed species (S.
Goodman, pers. comm., 2002). Further specimen collections, with associated tissue
samples, are needed to confirm the identity of this species.
The taxonomy of the bat genus Pipistrellus is unclear in Madagascar. One species
from this genus was collected during the surveys. This appears to be an undescribed
species (S. Goodman, pers. comm., 2002).
Microgale brevicaudata is distributed throughout a number of habitats on the
northeast and west coast of Madagascar. It is also found south of Toliara. An
unknown Microgale species was captured within riparian and spiny forest habitats
throughout the region of the proposed Parc Regional de Belomotse. The species has
27

proved difficult to identify, as the measurements and external characteristics do not
resemble those of M. brevicaudata. This may be an undescribed species of Microgale,
or a previously undescribed regional variant of M. brevicaudata. Specimens and DNA
samples have been sent to Chicago Field Museum for analysis. Results are not yet
available. Until formal identification has been completed, the species has been
tentatively assigned as M. brevicaudata.
The Coquerel’s dwarf lemur, Mirza coquereli, is currently listed as vulnerable with a
decreasing population (IUCN, 2002). During this survey this species was frequently
observed, foraging high in the canopy in gallery forest. It was not observed in
degraded gallery forest with low or fragmented canopies, nor was it observed in spiny
forest. This indicates that the species is dependent on undisturbed riparian forest.
Furthermore, the species is solitary with large home range sizes (Sussman et al.,
1985), so large areas of continuous forest are necessary in order to sustain viable
populations. Comparatively little gallery forest is left in Madagascar, and the
remaining areas are highly fragmented. These factors are part of the reason for the
classification of M. coquereli as vulnerable. The unexpected existence of
comparatively large populations in the surveyed area is a positive sign and an
important find for the conservation of this species.
5.4.5 Threats to mammal conservation
A number of factors affect mammal species density and distribution throughout
Madagascar. Several of these factors impact severely on specific mammal species,
causing declines in their overall populations and threatening their survival.
Throughout Madagascar, large areas of land are being cleared for agriculture. In
addition, forested areas are cleared for the production of charcoal and the collection of
firewood and building materials. Some endemic mammal species, such as Microcebus
murinus, are distributed throughout a wide variety of habitats and adapt well to
changing environments. However, many endemic mammal species are sensitive to
change, so populations are often adversely affected by habitat degradation. Many
mammal species are limited to specific habitats and comparatively small areas in
Madagascar. Clearance and fragmentation of the habitats and areas can be reflected by
a decrease in the populations of these mammals, which can lead to species becoming
vulnerable or endangered with extinction.
The remaining areas of gallery forest within the region of the proposed Parc Regional
de Belomotse are under severe threat through clearance for cultivation and charcoal
production. The corridor of gallery forest between Antafoky Lake and the Sept Lacs
region is very fragmented and it is doubtful whether it can be classified as a corridor
for certain mammal species. The small populations of many species that have become
isolated by fragmentation of habitats may be in danger of local extinction. An
example of this would be the dwarf lemur, Mirza coquereli, which is sensitive to
habitat fragmentation. This species was only found in gallery forest and now appears
to form three fragmented populations; one around Lake Antafoky, a second in an area
of forest near to Mahaleotse and the last in the upper reaches of the Sept Lacs region.
Further degradation of these sites will limit the range of this species still further.
28

In many areas, species are hunted for food or collected for the pet trade. Mammal
species that are impacted by these activities include lemurs and tenrecs. The large
tenrec species are seen as a staple food source in many areas. For example, Tenrec
ecaudatus is frequently hunted throughout much of its range. However, the species
reproduces often and has large litters, so declines in distribution and population levels
have not been reported, although this may be the result of a lack of research and
monitoring. Relatively high numbers of T. ecaudatus were caught during the survey
and population densities seemed to be high. However, as the human population (both
regionally and nationally) increases, the long-term impact of hunting and collection
on this species needs to be monitored throughout Madagascar. The fruit bats Eidolon
dupreanum and Pteropus rufus are hunted for food, potentially causing local
extinctions and threatening the future of these species in certain areas (Hutcheon,
1994). E. dupreanum was found throughout the surveyed areas. However, because of
the inaccessible locations of many of the roosting sites, the impact of hunting on this
species does not seem severe.
Ring-tailed lemurs, Lemur catta, are collected from the proposed park area for the pet
trade. Juveniles of this species are sold to hotels and restaurants in Toliara and nearby
tourist resorts. Infant mortality in the wild is high for this species, so such collections
may not severely impact upon the populations. However, if the mothers are killed to
facilitate the collection of juveniles, or if collection rates increase, this could impact
upon ring-tailed lemur populations in the area.
6. Triaenops population study
6.1 Introduction
Preliminary surveys within the proposed Parc Regional de Belomotse revealed three
species of microchiroptera roosting within a cave system in the Sept Lacs region. The
identified species were Hipposideros commersoni, Triaenops rufus and T. furculus.
Very little is known about the ecology, distribution or behaviour of these species in
Madagascar. In particular, little is known about the endemic species T. furculus,
which is classified as vulnerable with a decreasing population, or T. rufus, which is
classified as data deficient (IUCN, 2002).
A population study was undertaken in order to ascertain the approximate number of
individuals within the cave system. Associated behavioural data were also recorded.
The study aimed to contribute to the basic knowledge and conservation status of these
species.
6.2 Methods
In August 2002, an eight-day mark-recapture survey was undertaken, following Wilson
et al. (1996), to estimate the populations of the species present within the cave system.
The cave consisted of a large subterraneous cavern in limestone that opened to the
outside through a narrow sinkhole. This hole measured approximately 2m x 3m.
Previous explorations of the cave system by qualified cavers had identified a flask-
29

shaped limestone cave with an extremely deep, branching central tunnel. The small,
single exit to the cave provided an ideal opportunity to comprehensively survey the
entire bat community that lived within the cave system.
Two 1 m x 1 m harp traps were used in conjunction with two mist nets (9 m and 12 m),
which were placed in suitable locations around the cave entrance. Trapping was
maintained throughout the night from dusk until dawn. The time of each capture was
recorded, then captured individuals were identified to species, sexed, measured, marked
(by cutting the fur between the shoulders) and released. Individuals were marked over a
total of 6 nights, encompassing approximately 144 harp trap hours and 1,260 mist net
hours. During the following 2 nights, trapping was continued but no specimens were
marked. Instead, the time of capture and numbers of marked and unmarked individuals
were recorded. This two-night trapping period encompassed approximately 42 harp trap
hours and 480 mist net hours.
6.3 Results
The survey revealed large populations of Triaenops rufus and T. furculus. No
individuals of the species Hipposideros commersoni were captured or observed, even
though previous trapping had revealed a population of this species within the cave
system.
Comparing the ratio of marked to unmarked individuals captured during the final twonight
collection period, the total population of each Triaenops species was estimated.
Using the Lincoln index, the population of T. rufus was estimated to be approximately
32,000 individuals, with a standard error of 3,060. The population of T. furculus was
estimated to be approximately 9,000 individuals, with a standard error of 1,900. The
combined population totals of both Triaenops species equate to approximately 41,000
individuals within the cave system.
Approximately 2,700 individuals in total were captured, marked and released over 6
nights, with consistently high numbers of individuals captured every night.
The sex ratio of captured individuals was almost equal for both species, although T.
rufus had a slightly higher number of males and T. furculus had a slightly higher
number of females.
30

6.4 Discussion
6.4.1 Temporal variation
At dusk and during the first hour of darkness (between 6.00 and 7.00 p.m.) many
thousands of bats emerged from the sinkhole. Between 6.00 and 6.30 p.m. the
majority of captured individuals were T. rufus. Large numbers of T. furculus emerged
from the sinkhole slightly later, between 6.30 and 7.30 p.m. Between approximately
7.00 until 9.00 p.m., few bats were seen or captured. After this time, a small number
of bats began to return. Between approximately 9.00 p.m. and midnight the bats
showed another peak in activity, with large numbers of both species returning to the
cave. After this time, activity was low until approximately 4.30 a.m. when the
remaining bats returned to the cave prior to sunrise. See figure 6.
Nightly activity pattern
Number of individuals
1000
900
800
700
600
500
400
300
200
100
0
rufus total
Furculus total
18:00-19:00
20:00-21:00
22:00-23:00
Time
00:00-1:00
2:00-3:00
4:00-5:00
Figure 6. Triaenops rufus and T. furculus nocturnal activity pattern
6.4.2 Seasonal variation
During November 2001, a preliminary Frontier study of the species that inhabited the
cave system had identified a population of Hipposideros commersoni. A number of
individuals were captured and released at that time. However, no individuals of this
species were observed or collected during the mark-recapture study in August 2002.
It may be that the population no longer roosts within the cave, or it may be that the
species shows a reduction in activity during the dry season. During discussions with
villagers and local guides within the area, many stated that the ‘large bat’ hibernates
deep within the cave during the dry season. H. commersoni is one of the largest
insectivorous bats in Madagascar, with a weight of 40-80g (Garbutt, 1999). This large
body weight may enable the species to retain the necessary reserves to remain inactive
during the dry season. Alternatively, it may be that the cave is only used by H.
commersoni as a communal nursery colony during the breeding season.
31

6.4.3 Breeding status
Many of the captured female T. furculus had swollen genitalia, indicating that the
species was in the breeding season. During a previous survey, in November 2001,
many of the captured females of both Triaenops species were pregnant. Most of the
other mammal species that occur in the surveyed areas also had a breeding season
towards the end of the dry season. This would mean that the births would coincide
with the onset of the wet season, which would ensure the provision of food for the
offspring.
6.4.4 Population density and conservation status
The survey showed exceptionally high densities of Triaenops rufus and T. furculus
within the cave system. Previous studies of these species have identified populations
roosting within caves or equivalent sites, often in mixed-species colonies with
Miniopterus, Myotis and Otomops species. These populations typically contain several
hundred individuals (Garbutt, 1999).
T. furculus is classified as vulnerable, with a decreasing population. T. rufus is data
deficient (IUCN, 2002). Considering the given status of these species, this survey has
identified important populations of both Triaenops species. The presence of
approximately 9,000 individuals of a vulnerable mammal species within a single site
warrants specific conservation attention. In conjunction with the exceptional numbers
of T. rufus and the suspected presence of Hipposideros commersoni, the site should be
classified as one of the most important known microchiroptera roost sites in southwest
Madagascar.
7. Diurnal lemur population study
7.1 Introduction
Preliminary biodiversity studies within the proposed Parc Regional de Belomotse
identified the presence of two species of diurnal lemur; Verreaux’s sifaka, Propithecus
verreauxi verreauxi, and the ring-tailed lemur, Lemur catta. Both species are classified
as vulnerable (IUCN, 2002). While there have been extensive studies of both species
across their ranges in west and south-west Madagascar, no previous studies had been
undertaken in the proposed park area. Whilst lemurs are not necessarily reliable
indicators of forest biodiversity (Ganzhorn 1999), their endemism and status as
Madagascar’s only primate species makes them an extremely important family to
monitor both for biodiversity and conservation management. In addition, they are
charismatic flagship species of importance for tourism, raising public awareness and
providing support to fund-raising.
A population distribution and density study was undertaken in order to ascertain the
range and approximate number of individuals of both species of diurnal lemur within
the surveyed areas. The study aimed to contribute both to conservation and tourism
development within the region, and to provide a basis for future monitoring.
32

7.2 Methods
Between May and August 2002, densities of diurnal species were calculated and
compared in surveys of three separate areas within the proposed regional park. Due to
the topographical and logistical constraints of surveying steep spiny forest, the survey
habitats were stratified and lemur densities were calculated for gallery forest alone.
Given the dependence of both diurnal lemur species on gallery forest during the dry
season and taking into account seasonal changes in home ranges, a study of lemur
density within gallery forest during the dry season should still reflect the general
population distributions and group dynamics within the area.
Line-transect methods outlined by Whitesides et al. (1988), Chapman et al. (1988)
and Rabinowitz (1997) were used to establish a survey design. Twenty 500-metre
transects were walked twice daily for a minimum of five days each. All sightings of
Verreaux’s sifaka and ring-tailed lemur groups were recorded. Date, time, bearing of
the transect, bearing of the spotted group or individuals, distance to the group, lemur
species, group height and numbers of individuals were all noted down for each
sighting.
Three areas within the proposed park area were surveyed. The first area, around
Antafoky Lake, comprised a fairly substantial band of gallery forest. The second
survey area investigated the thin strip of gallery forest along the north bank of the
Onilahy River and the forested areas within small tributary valleys. This area formed
a narrow, fragmented corridor between the first area and the third area in the Sept
Lacs region. The third area surveyed the sections of gallery forest located along the
A
B
Antafoky Lake
C
D
A
B
Antafoky
track road
source
agriculture A
B
Andranomite
A
(sacred spring)
A
MAHALEOTSE
Bevoalavo
D
ONILAHY RIVER
C
Camp 1
ManderanoC D B
sand
C
Camp 2
B
D
SEPT
Ifanato
Ifanato River and lake system, within the Sept Lacs region (see figure 7).
Figure 7. Positions of the twenty lemur transects. The red and green transects were within the Antafoky
Lake area. The Purple transects were within the Sept Lacs area and the blue and orange transects were
located within the forest corridor.
33

Line transect methods have been developed for estimating the density of primates in
continuous forest and have not been adapted for fragmented habitats such as those found
within the survey area. However, the survey team spent long periods of time in each
area, allowing for repeated observations of the lemur groups. Therefore, total counts of
both species were also estimated from these direct observations.
7.3 Results
The survey revealed populations of Propithecus verreauxi verreauxi and Lemur catta
in all of the three surveyed areas (see table 5).
7.3.1 Line transects
Assessments of lemur population numbers using line transects identified high
numbers of ring-tailed lemurs around Antafoky Lake and along the corridor of forest
between the lake and the Sept Lacs region, with a comparably lower number within
the Sept Lacs region. The density of ring-tailed lemurs appeared to be highest around
Antafoky Lake.
A comparable pattern was observed for Verreaux’s sifaka, with high numbers around
Antafoky Lake and along the forest corridor, and comparatively lower numbers within
the Sept Lacs region. The density of sifaka appeared to be considerably greater around
Antafoky Lake than at the other two sites.
7.3.2 Direct observation
Assessments of ring-tailed lemur populations by direct observation identified high
population numbers around Antafoky Lake and along the forest corridor, with lower
numbers within the Sept Lacs region. The density of ring-tailed lemurs appeared to be
slightly higher around Antafoky Lake.
A similar pattern was observed with Verreaux’s sifaka, with high numbers observed
around Antafoky Lake. The density of sifaka appeared to be greater around Antafoky
Lake than at the other two sites.
Table 5. Minimum numbers and density of diurnal lemurs at three locations within the survey area,
using two different methods.
Lemur catta
Area of
gallery forest
Min. No. by
direct count
Density (No. / ha.)
by direct count
Min. No. by line
transect
Density (No. / ha.)
by line transect
Antafoky Lake 140 ha. 138 1 299 2.1
Sept Lacs 110 ha. 75 0.7 158 1.4
Corridor 220 ha. 150 0.7 260 1.2
P. v. verreauxi
Antafoky Lake 140 ha. 42 0.3 110 0.8
Sept Lacs 110 ha. 25 0.2 17 0.2
Corridor 220 ha. 25 0.1 52 0.2
34

7.4 Discussion
7.4.1 Comparison of methods
The two lemur density assessment methods produced very different results. The direct
observation method obtained results through repeated sightings and countings, and
gave a minimum figure for the number of individuals present within the surveyed
areas. However, it is probable that not all groups and individuals were observed and
counted. Conversely, the line transect method produced much higher figures for lemur
numbers and densities than the direct observations, in some cases the estimates were
over twice as high.
Based on the relative accuracy of repeated counts based on direct observations within
fairly small areas of forest, the estimates produced by the line transects appear to be
too high. The areas of gallery forest surveyed were fairly narrow and a large field
team spent a long period of time in each survey area. The presence of such large
numbers of diurnal lemurs would have been observed during direct observations.
The actual number of individuals for each of the diurnal lemur species is probably
closest to the figures given by direct observation. Previous surveys within similar
areas and habitats have shown that the line transect method is difficult to apply, and
that counts based on direct observations have been found to be more accurate (J.
Durbin, pers. comm., 2002).
7.4.2 Lemur density and distribution
The areas of gallery forest differ considerably in size between the three surveyed areas
(see Table 5). This appears to have been reflected in the numbers and densities of
lemurs that were observed in these habitats. High densities of both diurnal lemur
species were estimated for the area of forest around Antafoky Lake where, despite
recent forest clearance for crop cultivation, the forest is both large and fairly
undisturbed. With an estimated area of approximately 140 ha, this is the single largest
fragment of gallery forest within the proposed park area.
The corridor of gallery forest contained fairly high numbers of P. verreauxi verrauxi
and very high numbers of L. catta. However, the density of lemurs was low. The
combined areas of gal1ery forest along this corridor were estimated to be
approximately 220 ha in size, although the quality of the forest varied considerably.
The small tributary valleys contained some areas of relatively undisturbed forest,
whereas the gallery forest that bordered the river was narrow, fragmented and
extremely degraded in some areas. It appears that this fragmentation to the gallery
forest along the corridor, combined with the fairly high levels of degradation in some
areas, may have led to the comparatively low lemur densities that were observed.
Low numbers of both diurnal lemur species were recorded within the Sept Lacs
region. However, the densities of both species were comparatively high. The forest
was less disturbed than the corridor area but only covered approximately 110 ha in a
narrow band along the Ifanato River. This may have adversely affected lemur density.
However, several other factors should be taken into account, such as plant species
composition and detection bias (lemurs could not be heard over the sound of the
35

waterfalls). Also, the Sept Lacs region is visited by a small number of tourists, which
may have affected the behaviour of the diurnal lemurs. Therefore, further studies may
be needed to clarify these findings.
7.4.3 Conservation status
The ring-tailed lemur and Verreaux’s sifaka are both classified as vulnerable (IUCN,
2002). Both species are distributed throughout southwest Madagascar and are found
within several protected areas. However, both species rely on restricted habitats that
are known to be diminishing rapidly due to clearance for agriculture, fuel wood and
building materials. In addition, both species are hunted for food throughout much of
their range and are collected for the pet trade.
Large populations of both species were recorded within the proposed Parc Regional
de Belomotse. The survey data provide a good baseline estimate of diurnal lemur
density and can go some way to reflecting the abundance of lemurs in this relatively
unstudied area. However, considering the fact that some areas of gallery forest are
under severe threat within the proposed park area, more work needs to be done to
maintain and improve the remaining forested areas. Further conservation efforts
should also focus on the maintenance and improvement of the forest around Antafoky
Lake and the corridor along the Onilahy River. Studies should also focus on the
identification of the combination of factors that affect the densities of both lemur
species throughout the area.
8. Birds
8.1 Introduction
The long isolation of Madagascar from mainland Africa has not been such a
restriction to colonisation for early bird species as for the mammals, reptiles and
amphibians. As a consequence, birds show a lesser degree of endemism than these
other taxonomic groups, but one that is still remarkably impressive by global
standards. Madagascar boasts at least four endemic families – the mesites
(Mesitornithidae), vangas (Vangidae), ground-rollers (Brachypteraciidae), and
cuckoo-rollers (Lepstosomatidae), as well as the subfamily of asities (Eurylaimidae).
A total of thirty-seven genera are endemic to Madagascar and the Comoros islands.
Several of these genera are monotypic, including the Madagascar Harrier Hawk and
the Helmet Vanga. At the species level, 59% of breeding birds are endemic.
This significant level of endemism unfortunately correlates with the level of
threatened bird species in Madagascar. The island’s extraordinary biogeography
means that habitat loss has a huge impact on bird species that have become
evolutionarily specialised, particularly wetland and rainforest species. Thirty-three of
the thirty-seven endemic genera are only found in forested areas. Several are under
threat from forest clearance. In total, twenty-seven endemic bird species are currently
threatened (IUCN, 2002), of which five are ‘critically endangered’. Four of these are
wetland species. Six species are ‘endangered’ and the remaining sixteen are
considered ‘vulnerable’. A further fourteen bird species are classified as near
36

threatened. Several species, including the recently discovered Red-shouldered Vanga
(Calicalicus rufocarpalis), are classified as ‘data deficient’ and further studies are
needed to ascertain their IUCN status.
Protected areas form a crucial element to bird conservation since nearly all threatened
forest species are found in these areas. However, wetland species receive virtually no
protection and only two of the critically endangered waterbirds have populations or
pairs in protected areas. Permanent wetland areas in the south and west of Madagascar
are scarce. The proposed Parc Regional de Belomotse contains forest and wetland
habitats that are suitable for a wide variety of bird species. The proposed park area
also contains some of the last remaining areas of riparian forest along the Onilahy
River. The Onilahy River appears to be a biogeographical limit for many bird species,
which makes it an interesting area to study.
The primary aim for studying the bird communities in the proposed park area was to
create a species inventory. Furthermore, the study identified and compared the bird
species that occupied different habitats, to provide baseline information that could
support habitat-specific conservation recommendations.
8.2 Methods
Bird observations were recorded and a list of species was produced, with
accompanying ecological data. All positive identifications were confirmed from
captures and sightings, as field-staff were not experienced enough to identify species
solely from vocalisations.
Bird observations were mostly opportunistic. All members of the group made these
incidental bird records, although one staff member confirmed the majority of positive
identifications. In order to minimise the margin of error caused by the use of
volunteers when making field observations, a strict protocol was observed to confirm
a positive identification. Information was recorded at the time of the sighting. If the
recorded distinguishing features of the bird matched the distinguishing features of one
specific species in the field-guides, a positive identification could be made and the
species was recorded as a confirmed sighting. If a description was made that did not
match any species in the book, the description was recorded in a logbook and further
observations were made in the area.
Species were identified using the following field guides: ‘Birds of Madagascar - A
Photographic Guide’ by Pete Morris and Frank Hawkins (1998), and ‘Birds of the
Indian Ocean Islands’ by Ian Sinclair and Oliver Langrand (1998).
See Frontier-Madagascar Methodology Training Manual (Frontier Madagascar,
2003) for detailed description of methods.
8.3 Results
Over 300 man-hours of field observations revealed 1,971 bird sightings, the vast
majority of which were confirmed to species level.
37

Seventy-nine species were positively identified in the area. See Appendix 7 for the
complete species list. A further eleven birds were not identified to species, but were
distinct from the positively identified species, see Appendix 8. Classifications follow
those used by Morris & Hawkins (1998) and Sinclair & Langrand (1998).
66% of the positively identified birds were found both around Antafoky Lake and
within the Sept Lacs region, most probably because they contain similar habitats. A
total of 26.5% of bird species were found only around Antafoky Lake, although this
may reflect the higher birding effort in this area rather than a habitat-specific
difference in species distributions. Many of the birds that were identified from only
one location were positively identified on only one or two occasions.
8.4 Discussion
8.4.1 Rare species and endemism
Of the positive identifications made, two were classified as vulnerable (IUCN, 2002);
the Humblot’s Heron (Ardea humbloti) and the Madagascar Pond Heron (Ardeola
idae). One near threatened species was identified, the Madagascar Sparrowhawk
(Accipiter madagascariensis).
The Peregrine falcon (Falco peregrinus) was observed on several occasions. This
species is noted as uncommon throughout Madagascar (Morris & Hawkins, 1998;
Sinclair & Langrand, 1998). A number of other observed bird species are noted as
uncommon in southwest Madagascar, including the Madagascar wagtail (Motacilla
flaviventris), the Madagascar long-eared owl (Asio madagascariensis) and the
Madagascar white-eye (Zosterops maderaspatana). Many water birds, including the
Black-Crowned Night Heron (Nycticorax nycticorax), Great egret (Egretta alba) and
Hammercop (Scopus umbretta) are also stated as uncommon in the dry south and
western regions.
A total of thirty-three species (41% of all species found) were endemic or endemic
breeders to Madagascar. A total of twenty-one species (27%) were regionally
endemic.
8.4.2 Ecological variation
Sixty bird species (76%) were observed in gallery forest, many of which are reliant on
the presence of this habitat. For example, the cuckoo-roller (Leptosomus discolor) was
observed in gallery forest and is said to be restricted to this habitat in the south
(Morris & Hawkins, 1998). Five bird species were observed solely in gallery forest, of
which four were observed only in the gallery forest around Lake Antafoky (see
Appendix 7). This may reflect the quality and size of this forested area. Many bird
species were observed in disturbed areas of gallery forest, as well as comparatively
undisturbed forest. It may be that this is indicative of the habitat restrictions imposed
on bird species by the fragmented areas of gallery forest, rather than typical habitat
preferences of these species.
38

Nine species were observed solely in spiny forest throughout the surveyed areas. All
other bird species that were observed in this habitat were also seen in disturbed
habitats.
A total of fifteen species of water bird were observed within the surveyed areas, many
of which are recorded as rare in southwest Madagascar Madagascar (Morris &
Hawkins, 1998; Sinclair & Langrand, 1998). The Humblot’s Heron and the
Madagascar Pond Heron, both classified as vulnerable with a decreasing population,
were regularly seen at Antafoky Lake and the marshes along the Onilahy River. The
variety of water birds observed reflects the large areas of wetland that occur in the
area. The proposed park area contains a variety of different wetland habitats including
lakes, ponds, marshes and stream edges. These habitats are distributed as a series of
core areas linked by the Onilahy River. These are the only permanent wetlands in
approximately 2,500 sq km, providing a unique and vital habitat for water birds in the
region.
8.4.3 Threats to bird conservation
Many bird species are threatened through habitat destruction due to forest clearance
for agriculture, charcoal production and other resource uses. The gallery forests within
the surveyed areas are undergoing variable rates of degradation. The area of forest
around Antafoky Lake is undergoing a rapid rate of deforestation, which will reduce
the area of viable habitat for many species.
The wetlands are under threat from drainage for rice cultivation. In addition, the reeds
are collected for use as building materials. Two vulnerable bird species, Ardea
humbloti and Ardeola idae, rely on these wetlands. The destruction of these habitats
could lead to a further decline in the populations of these species.
Many bird species are threatened through hunting exploitation. In particular, the
White-faced Whistling duck and Comb duck are hunted around the wetlands within
the surveyed areas, often by visitors from Toliara. Villagers within the region also
hunt birds and have been seen hunting for buttonquail, ground couas and guinea fowl.
This can cause dramatic reductions in populations and can cause noise disturbance
leading to migration of other species from the sites.
Introduced fauna can adversely affect bird communities within the area. The bush-pig,
Potamochoerus larvatus, was present within the surveyed areas. It was particularly
common in the upper reaches of the Ifanato River, within the core Sept Lacs area.
This area contains many marshes. P. larvatus plunders the nests of ground-laying
birds and disturbs water-birds and their nests, affecting populations within the area.
The small indian civet, Viverricula indica, has been seen throughout the surveyed
area. This introduced animal preys on small vertebrates and will take birds and eggs.
The introduced rat, Rattus rattus, can have an impact on breeding populations of
birds, destroying eggs of ground-laying and arboreal-nesting bird species.
39

9. Reptiles
9.1 Introduction
The Field-Guide to the Reptiles and Amphibians of Madagascar (Glaw & Vences,
1994) lists 290 species of reptiles for Madagascar, although this figure has now risen
to 346 known species (Raselimanana, pers. comm). This figure is likely to increase
further, as undescribed species are frequently being found and many of the genera are
in need of revision. The Order reptilia exhibits extremely high levels of diversity and
endemism in Madagascar. Of the 346 known species, 314 are endemic to Madagascar.
A recent global survey of biodiversity hotspots found just two other regions
(Caribbean and Meso-America) that had a greater diversity of endemic reptile species
(Myers et al., 2000).
Southwest Madagascar has become one of the worlds leading herpetological hotspots,
with an extremely high proportion of endemic species. This is mainly a result of the
region’s long isolation, its variable and often extreme conditions, and the relative lack of
competition from other animal groups (such as mammals). However, due to the
inaccessibility and difficult working conditions, surprisingly little work has been carried
out on the herpifauna of the region. Many species are known from only one or very few
specimens so there is extremely little information on their physical dimensions,
population densities, ecology, behaviour and distribution. Because of the lack of
information, comparatively few reptile species have IUCN status.
Reptiles are worthy of study not only because so little is known about them in
Madagascar, but also because they can be good indicators of habitat types and levels
of disturbance. During the study, a species list was compiled and ecological,
biogeographical and physiological data were collected.
9.2 Methods
Reptiles were observed and collected using systematic and opportunistic methods at
all selected trapsites. Pitfall traps were set up in representative habitats to target
ground-dwelling and fossorial species. In addition, fossorial species were collected by
opportunistic digging and leaf-litter searches in suitable habitats. Crevice searches
were also undertaken. This included looking under bark, fallen logs, rocks, and
looking inside vegetation. Large diurnal snakes and lizards were captured by hand or
with the use of sweepnets. At night, representative habitats were targeted for nightwalks,
using head-torches to locate specimens. In addition, ongoing casual collections
were made during other survey activities.
Representative specimens were collected and ecological and geographical data
recorded for each capture site. Specimen colouration was recorded in life, prior to
killing the animals with an injection of saturated methane sulphate (MS222) solution.
Once killed, the specimens were fixed in a 5-7 % formalin solution. DNA samples
were taken from all species collected.
40

9.3 Results
A total of fifty-five different reptile species, belonging to thirty-four genera and ten
families, were recorded during the study period. These are listed in Appendix 9 with
associated ecological data. IUCN status is also listed (IUCN, 2002). The classification
and endemic status is based on Glaw & Vences (1994) and Raselimanana (pers.
comm., 2002). Appendix 9 also shows the habitat localities for observed and collected
specimens during this survey. However, this may not necessarily reflect the general
habitat distribution of these species.
The asymptote for reptile species accumulation curves were reached both for the
Antafoky Lake area and the Sept Lacs region (see figure 8). It is therefore assumed
that the majority of species present were collected using the methods described above.
60
Antafoky Lake area
60
Sept Lacs region
50
50
No. of Species
40
30
20
No. of Species
40
30
20
10
10
0
0
0 50 100 150 200 250 300 350 400 450
0 50 100 150 200 250 300 350 400
Effort
Effort
Figure 8. Reptile species accumulation curves for the surveyed areas.
9.4 Discussion
9.4.1 Biogeography and species richness
The biodiversity survey produced an extremely comprehensive reptile species list for
the site of the proposed Regional Park. Prior to this survey, only one short
herpetological study had been undertaken in this area (Raxworthy, 1995). During that
survey, twenty-nine species were recorded, all of which were also recorded during
this study.
Of the fifty-five species collected, fifty (91%) are endemic, of which nineteen (35%)
are limited to the dry forests of southwest Madagascar, see Appendix 9. The nonendemic
species that were collected or observed included the Nile crocodile
(Crocodylus niloticus), two species of freshwater turtle (Pelusios castanoises and
Pelomedusa subrufer), and two gecko species (Hemidactylus mercatorius and H.
mabouia). All of these species are also found on the African mainland. The means by
which these non-endemic species arrived on Madagascar is not clear, although it is
probable that man introduced the house gecko, Hemidactylus mabouia.
The high level of species richness in the surveyed areas can be attributed to the
diversity of habitats in this area. These habitats include riparian and spiny forest,
lakes, temporary marshlands, open sandy grassland, cultivated areas and limestone
41

cliffs. In addition, many areas contain habitats with a negligible level of human
disturbance.
9.4.2 Seasonal variation
The region has an extremely variable climate. This provides seasonal variation of
conditions and habitats, to which many of the reptile species have adapted. Many
reptiles in southwest Madagascar exhibit behavioural adaptations during the dry
season (May-October).
Certain species were dormant during the dry season, but were collected, often in large
numbers, following the onset of the rains. These species included two species of
freshwater turtle, Pelomedusa subrufer and Pelusios castanoides; a water-dependant
snake, Liopholidophis lateralis; and a burrowing snake, Typhlops decorsei. All other
species were collected throughout the year, although the density of active specimens
of many of these species varied considerably during the dry season.
During the months of May – October, the temperature in the region is comparatively
cool, reaching as low as 10ºC during the night. There is little vegetation on many of
the plants, particularly those found in the dry habitats on the hillsides and plateau.
Also, the number of active invertebrate species is comparatively low. These factors
have a significant effect on the levels of activity of the reptile populations in the
region at this time. Many reptile species appeared to reduce their levels of activity
during the dry season. In particular, nocturnal reptiles became less active. This
reduction in activity was particularly noticeable for nocturnal species from the
families Boidae and Colubridae. Also, chameleons from the genus Furcifer were less
frequently observed during the dry season. During this time, several individuals of
Furcifer verrucosus were observed in a dormant state, buried in shallow soil. This
behaviour has been noted before in this region (C. Raxworthy, pers. comm., 2002). It
appears that many reptile species reduce their activity during the dry season as an
adaptation to one or more factors. These factors include a reduction in temperature, a
reduction in available food, and an increase in levels of predation.
Not all reptile species reduced their activity during the dry season. In particular, many
species and specimens from the genera Mabuya and Paroedura were collected at this
time. This did not appear to represent an increase in the number of active individuals,
though, as comparable numbers were also collected during the rainy season. However,
an apparent increase in the levels of activity of some small colubrids was noted. All
but one specimen of the colubrid species Liophidium trilineatum, L. torquatum and
Pseudoxyrhopus quinquilineatus were collected in pitfall traps during the dry season.
The increase in numbers of specimens collected at this time may reflect an increase in
leaf-litter activity due to a reduction in available niches during the dry season, or it
may reflect an increase in the level of feeding activity necessary with the associated
reduction in numbers of invertebrates.
9.4.3 Ecological distribution
Reptiles were found in all of the major habitats in the region. The majority of species
collected from the surveyed areas were found both in riparian and spiny forest. Out of
a total of 22 species (40%) that were found in both habitats throughout the surveyed
42

areas, all of them were found in the Sept Lacs region and all but one were found in the
Antafoky Lake area, see Table 6. This included species from eight different families.
This shows that many species have adapted to a variety of habitats, and that these
species are comparatively widespread throughout the surveyed area.
Table 6. Ecological distribution of reptile species in the proposed Regional Park
(Percentages in perentheses).
No. of Species
Location
Riparian
forest
Riparian and
spiny forest
Spiny forest
Water
Antafoky
Lake area
Sept Lacs
region
17 (35%) 21 (44%) 8 (17%) 2 (4%)
18 (37%) 22 (46%) 6 (13%) 2 (4%)
Both 21 (38%) 22 (40%) 9 (16%) 3 (6%)
A total of twenty-one reptile species (38%) were found only in riparian forest in the
surveyed areas. Of these, fourteen were snake species, many of which typically
depend on trees (e.g. Ithycyphus oursi, Langaha madagascariensis, and the
Madagascar tree boa, Sanzinia madagascariensis), or live in the leaf-litter (e.g.
Liophidium trilineatum, L. torquatum and Pseudoxyrhopus quinquilineatus). Other
species of note included Paroedura vahiny, Androngo trivittatus and Mabuya dumasi,
all of which are regionally endemic. The pygmy chameleon, Brookesia brygooi, was
found only in riparian forest, as were Geckolepis maculata and G. petiti. These
species are dependant on riparian forest and rainforest throughout their range (Glaw &
Vences, 1994).
A total of 9 reptile species (16%) were found only in spiny forest habitats. All of these
species have adapted to the dry environment, and include the radiated tortoise,
Geochelone radiata, and Standingi’s day gecko, Phelsuma standingi, both of which
are classified as vulnerable (IUCN, 2002). Other species include the burrowing snake,
Typhlops decorsei, of which very little is known, and the iguanas Chalarodon
madagascariensis and Oplurus cyclurus. Also collected in spiny forest were the skink
Mabuya gravenhorstii, the geckos Paroedura picta and Blaesodactylus sakalava, and
the gerrhosaurid Tracheloptychus madagascariensis. All of these species are
dependant on dry forests throughout their range (Glaw & Vences, 1994).
A total of 3 aquatic reptile species were observed or collected. The Nile crocodile,
Crocodylus niloticus, was observed in Antafoky Lake and in the Onilahy River
around Manderano. The freshwater turtle Pelusios castanoides was collected at
Antafoky Lake, where it was recorded in large numbers during the wet season. This
species is found in ponds and lakes throughout Madagascar. The turtle Pelomedusa
43

subrufer was collected from flooded marshes and temporary water bodies in the Sept
Lacs region. This species is recorded as inhabiting temporary areas of water
throughout Madagascar.
9.4.4 Rare species
The radiated tortoise, Geochelone radiata, was observed in areas of spiny forest on
the hillsides around Lake Antafoky. It was not observed in the Sept Lacs Region,
although extensive searches were carried out in suitable habitats. This endemic
species is classified as Vulnerable (IUCN, 2002) and is promoted as a flagship species
throughout south and southwest Madagascar. Very few specimens were seen, and the
population seems to consist of isolated, low density populations.
Standingi’s day gecko, Phelsuma standingi, was observed and collected from the
spiny forest on the hilltops around Lake Antafoky. It was not observed in the Sept
Lacs Region. This endemic species is classified as Vulnerable by IUCN (2002). Only
a single specimen was collected, as the species is apparently very uncommon in the
surveyed areas.
The endemic skink, Mabuya dumasi, has a localised distribution. It has only been
collected from areas of riparian forest along the Onilahy River. Specimens were
collected from the forests around Lake Antafoky and from the Sept Lacs region. In
both localities, it was commonly found.
The nocturnal snake, Madagascarophis ocellatus, is a regional endemic. It is
restricted to the dry forests of southwest Madagascar. Specimens were observed and
collected from the forests around Lake Antafoky and from the Sept Lacs region. In
both localities, it was commonly observed.
The nocturnal, endemic snake, Lycodryas gaimardi, is rarely collected in the Toliara
Region of southwest Madagascar. One specimen was collected in riparian forest
around Lake Antafoky, and a further two specimens were collected from the Sept
Lacs region.
The endemic burrowing snake, Typhlops decorsei, is rarely collected. It is limited to
the dry regions of southwest Madagascar. Only few specimens have been collected,
and the ecology of the species is poorly understood. Specimens were collected from
both surveyed sites, though only from spiny forest areas. During and soon after
rainfall, the species was comparatively easy to collect and appeared to be fairly
common.
A second endemic burrowing snake, Typhlops sp., was collected from both surveyed
sites. This snake is similar in appearance to Typhlops arenarius, but differs in midbody
scale count and pigmentation. Specimens of this potentially undescribed species
were commonly collected in areas of riparian forest around Antafoky Lake and
throughout the Sept Lacs region. It was the most commonly collected colubrid snake
in these areas.
44

9.4.5 Range extensions and locality records
Range extensions were recorded for several species. In addition, new locality records
were made for many species at the surveyed sites.
The endemic snake, Liophidium trilineatum, has only previously been recorded south
of the Onilahy River. Specimens of this species were observed and collected in
riparian forest around Lake Antafoky, on the north bank of the Onilahy River. This
species was not collected from the Sept Lacs region, and therefore appears to have a
localised distribution in the area.
The endemic pygmy chameleon, Brookesia brygooi, has previously been recorded as
far south as Sakaraha. A single specimen collected from riparian forest in the Sept
Lacs region represents a range extension of approximately 160km. This cryptic
species is difficult to observe. However, extensive searches only revealed one
specimen, so it can be assumed to be uncommon in the area.
The surveys conducted by Frontier identified twenty-six species that had not been
previously collected in the surveyed regions. These represent new locality records for
the area.
9.4.6 Threats to reptile conservation
The most significant threat to reptile species diversity in the surveyed region is habitat
loss. In particular, the riparian forests in the region are rapidly becoming fragmented.
Twenty-one species of reptile were collected only from areas of riparian forest, and
are vulnerable to the loss of this habitat. In particular, species of gecko from the genus
Geckolepis were only collected from areas of undisturbed riparian forest, showing
sensitivity to disturbance. The fragmented distribution of several riparian forest
dependant species (eg Liophidium trilineatum and Brookesia brygooi) may indicate
that localised extinctions of forest dependent species are already occurring.
Nine species of reptile showed dependence on the dry spiny forests, large areas of
which are being cleared for cultivation. Forested areas are cleared by burning, which
severely impacts the reptile populations, particularly the radiated tortoise, Geochelone
radiata. Also, large trees from the spiny forests are selectively felled for use as
building materials. These trees are the preferred habitat of the rare Standingi’s day
gecko, Phelsuma standingi.
The pet trade has affected the populations of many reptiles in Madagascar,
particularly chameleon, tortoise and day gecko species. There is not a recent history of
reptile collection in the surveyed areas, as regional collections have focussed on the
dry forests inland from Ifaty, where Phelsuma standingi have been collected in large
numbers in the past. However, if specimen collection for the pet trade were to occur in
the area, it would severely affect the populations of Geochelone radiata and Phelsuma
standingi, which have been found in such low densities that their populations already
appear to be critically threatened.
The area supports populations of two boa species, Sanzinia madagascariensis and
Boa dumerili. These species are both collected for the pet trade. These species are also
45

collected for their skins, which are sold in tourist markets throughout the country.
These species were common throughout the surveyed areas, but both species have
slow reproductive rates and the populations of these snakes could be threatened if
collecting was to occur.
The Nile crocodile, Crocodylus niloticus, is also present in the surveyed areas, and is
under threat throughout Madagascar due to hunting. Crocodiles are killed for their
skins and for food. However, the crocodiles in Lake Antafoky are locally considered
to be fady and are not hunted.
Local communities around Antafoky Lake have been observed trapping and collecting
the freshwater turtle, Pelusios castanoides, for food. Also, shell fragments of the
radiated tortoise, Geochelone radiata, have been observed around the village of
Antafoky, including the shell of an individual that had been marked and released by
Frontier earlier in the year. This specimen had been released over two kilometres from
the village, showing that tortoise collection occurs over a fairly large area. Only five
radiated tortoises were found in the surveyed areas over a period of fifteen months, so
it appears that this species is in danger of becoming locally extinct.
10. Amphibians
10.1 Introduction
The Field-Guide to the Reptiles and Amphibians of Madagascar (Glaw & Vences,
1994) lists 187 described species of amphibian for Madagascar, although as with the
reptiles, the actual figure is proving to be much higher. At least 230 different species
have been identified, and the status of more than 45 additional forms remain to be
clarified (Raselimanana, pers. comms). The amphibians show high levels of endemism,
with over 99% of the species endemic to Madagascar and its offshore islands. This
represents one of the highest percentages of amphibian species endemism in the world.
Only a few other countries, such as Brazil, Columbia and Mexico, have more endemic
amphibian species (Raselimanana, pers. Comm., 2002).
Southwest Madagascar contains a low number of amphibian species, compared to other
regions of Madagascar. This is mainly due to the arid climate and short rainy season.
However, several of those species present are endemic to the region and show high
levels of adaptation. For example, the genus Scaphiophryne has adapted well to the dry
habitats of south and southwest Madagascar and several species can be found.
The amphibian fauna of the proposed site of the Regional Park was studied as part of
the Frontier-Madagascar biodiversity survey. A species list was compiled, and
ecological, biogeographical and physiological data were collected. As with the
reptiles, amphibians can be good indicators of general habitat types and levels of
disturbance.
46

10.2 Methods
Amphibians were observed and collected using both systematic and opportunistic
methods at all selected trapsites. Pitfall traps were set up in representative habitats to
target ground-dwelling species. After rain, typical amphibian habitats were targeted
for sampling. In addition, ongoing casual collections were made during other survey
activities.
Specimens were collected whenever possible, and ecological and geographical data
recorded for each capture site as outlined in the Methodology Training Manual
(Frontier Madagascar, 2003). Specimen colouration was recorded in life, prior to
killing the animals in methane sulphate (MS222) solution. Once killed, the specimens
were fixed in a 5 % formalin solution. DNA samples were taken from all species
collected.
10.3 Results
A total of 6 different amphibian species, belonging to 5 genera and 3 families, were
recorded during the study period. These are listed in Table 7 with associated
ecological data. The classification and endemic status is based on Glaw & Vences
(1994), Glaw & Vences (2002), and Raselimanana (pers. Comm., 2002.). Table 7 also
shows the habitat localities for observed and collected specimens during this survey,
although this may not necessarily reflect the general habitat distribution of these
species.
Many specimens were captured in pitfall traps. However, nearly all Boophis doulioti
and all Mantidactylus curtus specimens were recorded from opportunistic collections.
Table 7. List of amphibian species found in the Sept Lacs region.
No. SPECIES
LOCATION HABITAT
ENDEMIC
Antafoky Sept Lacs Riparian Spiny
Lake Region
STATUS
forest scrub
Family: Ranidae
1 Ptychadena mascareniensis N
Family: Mantellidae
2 Mantidactylus curtus E
3 Boophis doulioti RE
4 Laliostoma labrosum E
Family: Microhylidae
5 Scaphiophryne brevis RE
6 Scaphiophryne calcarata RE
Total 4 6
KEY TO ABBREVIATIONS.
Endemic status:
• RE - Regional Endemic: Endemic species with limited distribution, occurring only in southwest Madagascar.
• E - Endemic: Occurring only in Madagascar.
• N – Non-endemic.
47

10.4 Discussion
10.4.1 Biogeography and species richness
The most significant achievement of the amphibian survey was the production of a
comprehensive species list for the proposed site of the Regional Park. Prior to this
survey, the only inventory of this region was produced following a short study by
Raxworthy (1995). A total of three species were recorded, all of which were also
recorded in this study.
Five endemic species were recorded, three of which are endemic to the dry coastal
forests of south and southwest Madagascar (see Table 7). The non-endemic species,
Ptychadena mascareniensis, originated from mainland Africa and has spread
throughout Madagascar. It can be found in almost every large body of fresh water
throughout the country.
The low level of species richness in the Sept Lacs region can be attributed to the dry
climate and low number of suitable habitats for amphibians. The species collected
during this survey are typical for the dry habitats of southwest Madagascar.
10.4.2 Seasonal variation
The seasonal distribution for several of the amphibian species within the survey area
was characterised by a single peak in species abundance during the rainy season. The
species showing this pattern were Scaphiophryne brevis, S. calcarata and Laliostoma
labrosum. These species burrow into the ground at the end of the rainy season and
remain dormant during the dry season, emerging again after the first rains. The other
species collected during the survey showed a high affinity to water, and the
abundance of these species appeared to be fairly constant throughout the year.
10.4.3 Ecological distribution
Amphibians were found in all of the natural habitats within the surveyed area,
although the majority of species were located only in close proximity to permanent
bodies of water. During the dry season, no specimens were collected in area of spiny
forest. However, following the onset of the rains, Laliostoma labrosum,
Scaphiophryne brevis and S. calcarata were collected from areas of spiny forest that
were fairly close to water. The other species were collected around permanent water
bodies. The most frequently recorded amphibian species was Ptychadena
mascareniensis, which was collected from every area of permanent water within the
surveyed areas, as was Boophis doulioti. However, the distribution of Mantidactylus
curtus was limited to the fast-moving stream within the core area of the Sept Lacs
region.
10.4.4 Threats to amphibian conservation
The most significant threat to amphibian species diversity in the location of the
proposed Regional Park is habitat loss. All species show affinity to water bodies and
48

the associated riparian forest, the deterioration or loss of which will threaten their
continued survival.
However, all species are present within the Sept Lacs region, managed by GELOSE.
Also, all species are present within the core area of the Sept Lacs region, which
currently has the highest degree of protective management. In particular,
Mantidactylus curtus is only found within this area. It is hoped the designation of the
area with reserve status will allow the continued existence of this isolated population.
11. Invertebrates
11.1 Butterflies
11.1.1 Introduction
Butterflies are often considered to be important indicators of overall biodiversity in an
area (Ehrlich, 1988). However, at present little work has been done on the butterflies
of Madagascar. There are no field guides specific to the country, and species lists are
incomplete. Ecological, biogeographical and seasonal data and lacking for many
species. As yet, the presence or absence of specific species cannot be used as an
indication of habitat quality until further studies have identified these links.
The first comprehensive list of the butterflies of Madagascar is in the process of
completion, the author of which has provided the taxonomic identifications here. The
specimens and ecological and seasonal data collected by the Frontier-Madagascar
Forest Research Programme will be incorporated into this work.
The primary aims of this survey were to produce a species inventory for the Sept Lacs
region, and provide data on distribution with respect to habitat type and season.
11.1.2 Methods
Canopy traps and sweep-nets were used for the collection of butterflies. Canopy traps
were an integral part of the biodiversity trapsites. The traps were placed at different
heights and in different habitats (e.g. near the ground in a clearing, or high up in thick
vegetation). The bait used in canopy traps was fermented banana. The banana bait was
made in advance and left in a suitable container in the sun for at least one week to allow
it to ferment. Traps were baited in the morning and checked at noon and at dusk.
Sweep-netting focussed on the peak activity time for most butterfly species, around
midday. However, in order to maximise the chances of collecting all species present,
opportunistic sweep-netting was undertaken throughout the day from dawn to dusk.
Representative specimens were collected, and ecological and geographical data
recorded for each capture site as outlined in the Methodology Training Manual
(Frontier Madagascar, 2003). Specimens were killed either by pinching the thorax or
by placing the butterfly in a container with ethyl acetate vapours. The specimens were
49

stored in butterfly envelopes and kept dry. DNA samples were taken from many of the
species collected.
11.1.3 Results
During the survey, a total of 340 specimens were retained for taxonomic purposes.
These represented 61 species from 37 genera and 5 families; Hesperiidae, Lycaenidae,
Nymphalidae, Papilionidae and Pieridae (see Table 8). See Appendix 10 for the full
species list. The classifications follow those of Lees (Pers. Comm. 2002).
11.1.4 Discussion
11.1.4.1 Ecological distribution
A definitive field-guide to the butterflies of Madagascar has yet to be produced, and
as such, the current ecological distributions of many of the butterfly species of
Madagascar are unknown. Therefore, it is not always possible to compare trends in
ecological distribution that have been identified during this survey with results from
other regions of Madagascar. What is apparent, however, is that there are many
butterfly species in the Toliara region that appear to have restricted ecological
distributions.
Table 8. Summary of butterfly specimens collected.
Family No. of genera No. of species
Hesperiidae 2 4
Lycaenidae 7 7
Nymphalidae 16 31
Papilionidae 3 5
Pieridae 9 14
Total 37 61
The surveyed region contained two major habitats: riparian forest and spiny forest. A
total of 45 species were collected solely from areas of riparian forest. Four species
were collected solely from areas of spiny forest, and 12 species were collected from
both habitats. See Table 9 for a summary of the ecological distribution of butterfly
species.
As a general rule, riparian forest is associated with water, whereas spiny scrub is dry.
This may lead to incorrect assumptions about the number of riparian forest dependant
species, as there are very real grounds for suggesting that many of the species are
associated with the water source, not the riparian forest plants, and that these species
can survive in any moist environment. Indeed, it may be that several of the species
categorised as gallery forest dependant in this survey will also be found in areas of
open or cultivated land near to water sources in the region. However, several of the
species collected show distinct patterns of distribution and, based on current
50

knowledge, are linked solely to riparian and rainforest habitats. These species include
Aterica rabena, Leptosia alcesta, Mylothris phileris and Tagiades insularis. Other
species that were collected solely in riparian forest, for which little data are known to
suggest whether there is actual habitat dependency, are Charaxes andara, C. cacuthis,
Apaturopsis paulianii, Libythea tsiandava, Melanitis leda, Neptis kikideli, and
Pseudacraea imerina.
As shown in Table 10, the families Hesperiidae and Papilionidae appear to show a
high degree of habitat dependency, with all species from these families being found
only in riparian forest. However, it should be noted that many of the species from
these two families are attracted to damp soil, particularly species from the genera
Papilio and Coeliades. Further study is needed to identify whether these and other
species collected show dependency on the gallery forest and associated plant species,
rather than dependency on water and shade.
Table 9. Ecological distribution of butterfly species in the proposed Regional
Park.
Habitat
Family Riparian forest Spiny forest Both
Hesperiidae 4 0 0
Lycaenidae 3 1 3
Nymphalidae 23 2 6
Papilionidae 5 0 0
Pieridae 10 1 3
TOTAL 45 4 12
Several species were only collected from the hillsides and plateau, in spiny forest.
These species appear to show dependency on this habitat. However, further studies on
the food-plants of the larvae are necessary before they can be defined as spiny forestdependant.
The species found only in spiny forest were Acraea encedon, A. zitja,
Colotis amata and Hemiolaus cobaltina.
A further twelve species were collected both in riparian and spiny forest, many of
which are widespread throughout Madagascar and mainland Africa. These included
Leptotes pirithous, Eurema floricola, Danaus chrysippus aegyptius and Zizula hylax.
It appears that these species are able to live in a variety of habitats.
It is to be hoped that ecological data collected from this and other studies in
Madagascar will be used to identify indicator species for different habitats. This will
be of use for rapid assessments of survey sites, allowing target species to be identified
and collected, and assumptions and decisions made based on the presence or absence
of these species.
11.1.4.2 New species, rare species and range extensions
A total of sixty-one butterfly species were collected and identified during this study.
These included 1 undescribed species, 1 possible new sub-species and 7 rare species
51

(D. Lees, pers. comm, 2002.). In addition, data and DNA collected are to be used in
the revision of species complexes.
A specimen was collected at the Antafoky Lake site that was tentatively identified as
Strabena goudoti. Further examination of this species by Dr Lees (NHM) showed the
specimen to be an undescribed species. An individual of this species was also
discovered from a site further north at around the same time. The specimen has been
labelled as morphospecies 77 at this time, and is awaiting formal identification.
Several of the species that were collected are extremely rare, or rarely collected. They
shall be discussed here separately.
Two specimens of the species Acraea obeira were collected from the Antafoky Lake
region, a male and a female. This species is widespread throughout Madagascar, but
rare throughout its range. Few specimens have been collected, and the Natural History
Museum has only two specimens.
During the dry season, two specimens of the species Apaturopsis paulianii were
collected from the Sept Lacs region, a male and a female, with a DNA sample. Only
one specimen of this species had ever been collected from southwest Madagascar. It is
a very interesting find of an extremely rarely collected butterfly (D. Lees, pers.
comm., 2002).
Three specimens of the species Byblia anvatara anvatata were collected, with DNA
samples. The specimens were collected both at the Antafoky Lake site and the Sept
Lacs region. The specimens conform to the form seriata collected in ‘Morondawa’ by
Rothschild and Jordan in 1903. Only that initial collection of 10 specimens was
known, until they were collected during this study. This represents a range extension
of nearly 300km, and a rediscovery of the species.
Two specimens of Charaxes cacuthis were collected from riparian forest around
Antafoky Lake, a male and a female. This species is rare in the west of Madagascar,
and may even be a new subspecies (D. Lees, pers. comm., 2002). These specimens,
and the DNA sample collected, will be used to help ascertain subspecies status.
One specimen of the species Neptis kikideli was collected from the Sept Lacs region.
This species is very rarely collected, and nothing was known of its habitat preference.
It was found in riparian forest.
One specimen of the species Pseudacraea imerina was collected from riparian forest
in the Antafoky Lake area. This is a very rare species (D. Lees, pers. comm., 2002). It
is of great interest for DNA studies, so the specimen will be extremely useful.
One specimen of the species Tagiades insularis was collected from the Sept Lacs
region. This also is a very rare species, with few specimens having been collected in
southwest Madagascar.
52

11.2 Malaise Trapping
11.2.1 Introduction
Invertebrate taxonomy is complex. Very few studies have been undertaken on the
invertebrate fauna of Madagascar, and even fewer surveys collect flying invertebrates.
Many studies that have been undertaken have focussed mainly on large and easily
captured species, such as butterflies and dragonflies.
Malaise trapping in the proposed location of the Regional Park was designed to
survey flying invertebrates. Malaise traps typically collect large numbers of
specimens from a wide range of orders, including Lepidoptera (butterflies and moths)
and Odonata (dragonflies). They also collect many Hymenoptera (particularly wasps,
flying ants and parasitoids of other insect groups). Also, vast numbers of Diptera
(flies) are collected, particularly when the malaise trap is set up close to water, when
large numbers of Nematocera (craneflies, mosquitoes and gnats) are captured. Malaise
trapping also produces large numbers of Coleoptera (beetles).
11.2.2 Methods
Malaise traps are large, tent-like stationary mesh nets that are placed in suitable habitats.
The insects meet a vertical barrier underneath a pitched canopy and instinctively fly
up towards the sky where they are directed towards the collecting bottle. Once in the
bottle they are unlikely to escape and soon fall into the collecting fluid - usually 70-
95% alcohol. These traps were located in natural corridors for flying invertebrates, such
as the edges of dense vegetation, gaps through vegetation, paths, or the edges of streams.
The traps were set up according to the methods outlined in the Methodology Training
Manual (Frontier Madagascar, 2003). They were placed in representative habitats over
all seasons, in order to collect a wide variety of invertebrate species. Traps were placed
in suitable habitats for a period of 2-4 days in each location. After that time, specimens
were removed and the trap was taken down.
Specimens were stored in at least 75 % ethanol. All specimens from the same malaise
trapping-location were stored together. Ecological and geographical data were
recorded for each collection, as was the date and duration of the trapping period.
11.2.3 Results
The malaise trapping proved a successful way of collecting flying invertebrates. Large
numbers of specimens from the focal orders were collected. However, while awaiting
the identification of the specimens, little can be ascertained regarding the invertebrate
fauna of the region.
53

11.3 Leaf litter Extraction
11.3.1 Introduction
The biota of the leaf-litter and soil are a major component of the number of
extinctions caused by tropical deforestation (Wilson, 1992). The number of leaf-litter
and soil dwelling species that have not been discovered or described has been
estimated in the millions, and includes insects, spiders, molluscs and millipedes.
These organisms play a very limited role in conservation decisions due to the lack of
information on the majority of such species (Kremen, 1994).
Leaf-litter extraction surveys in the proposed location of the Regional Park were
designed to survey ground-dwelling invertebrate Orders. This would include
Hymenoptera (ants and solitary bees), Coleoptera (beetles), Diptera (fly larvae),
Lepidoptera (moth and butterfly larvae), Diplopoda (millipedes), Chilopoda
(centipedes), Orthoptera (crickets), and primitive wingless groups such as Collembola
(springtails).
11.3.2 Methods
Leaf-litter extractions sampled the forest floor invertebrate community. Leaf-litter
samples were systematically collected along 100 m transects in representative
microhabitats throughout the survey area. The samples were filtered, and the
invertebrates extracted. See Methodology Training Manual (Frontier Madagascar, 2003)
or Fisher (1999) for full details of the extraction techniques.
Specimens were stored in at least 75 % ethanol. All specimens from the same leaflitter
transect were stored together. Ecological and geographical data were recorded
for each collection, as was the date of the trapping period.
11.3.3 Results
The leaf-litter extractions proved a successful way of collecting ground-dwelling
invertebrates. Large numbers of specimens from all focal orders were collected.
While awaiting the identification of the specimens, little can be ascertained regarding
the invertebrate fauna of the region. However, ant specimens that were collected have
been identified to the genus level, see Table 10. Specimens from nineteen genera of
ants were collected, of which one genus is dependant on the dry forests of western
Madagascar and three genera are described as rarely found in southwest Madagascar (B.
Fisher, pers. comm, 2002.). These genera were Aphaenogaster, Cerapachys, Platythyrea
and Pyramica respectively. Work is underway to identify the specimens to species level.
11.3.4 Discussion
Specimens from the genus Aphaenogaster were collected from the Sept Lacs region.
This genus has only been collected from the dry forests of western Madagascar. The
specimens were collected both in riparian forest, closely associated with water, and on
the hillsides in dry spiny scrub.
54

The genus Cerapachys is rarely found. Species from this genus are extremely
specialised, feeding only on the eggs of other ant species. A single collection of
specimens from this genus was made from an area of disturbed riparian forest in the
Antafoky Lake area, see table 10.
Table 10. Ecological distribution of ant genera from sampled areas in the proposed
Regional Park.
Antafoky Lake area Sept Lacs region
Disturbed
Riparian
Riparian
Riparian
Forest
Forest
Genus
Forest
Spiny forest
Anochetus 7
Aphaenogaster 3 3
Camponotus 1
Cardiocondyla 1
Cerapachys 1
Hypoponera 2 7 8 15
Leptogenys 1
Monomorium 18 18 19 20
Oligomyrmex 3 6 7 2
Pachycondyla 6 5
Paratrechina 4 9
Pheidole 3 5 18 19
Plagiolepis 2 2 12 14
Platythyrea 1 1
Pyramica 1 6 8 2
Simopone 1
Strumigenys 2 6 1
Tapinoma 1
Technomyrmex 2 2
Tetramorium 13 17 19 19
Tetraponera 1 1 8
TOTAL 11 11 14 15
Specimens from the genus Platythyrea were collected from the Sept Lacs region. Species
from this genus are rare in southwest Madagascar. However, specimens were found both
in riparian forest and on the hillsides in spiny scrub.
The genus Pyramica is also rarely found in southwest Madagascar. Specimens from this
genus were collected from the Antafoky Lake area and the Sept Lacs region. In both
sites, they were found in riparian forest. They were also collected from spiny scrub in the
Sept Lacs region.
From the results shown here, the Sept Lacs area appears to have a greater diversity of ant
genera than the Antafoky Lake area. However, further leaf-litter transects have been
undertaken in these sites, and the results of these collections must be known before
conclusions can be drawn.
55

11.4 Dragonflies
11.4.1 Introduction
Dragonflies, from the order Odonata, show extremely high levels of diversity in
Madagascar. However, few studies of dragonflies have previously been undertaken in
many regions. The arid southwest of Madagascar has been very poorly studied. There
are no taxonomic keys or field-guides to the dragonfly fauna of Madagascar, and
identification is extremely difficult.
The study aimed to sample the dragonfly communities within the survey areas, and
identify the species present.
11.4.2 Methods
Dragonfly specimens were collected with the use of a sweep-net. Collections were made
along watercourses and in areas of wetland. Specimens were collected in the early
morning or late evening, when the dragonflies were flying slowly or resting on
vegetation near the water.
Specimens were killed using the vapours of ethyl acetate. They were stored in
butterfly envelopes and kept dry. Ecological and geographical data were recorded for
each collection location, as described in the Methodology Training Manual (Frontier
Madagascar, 2003).
11.4.3 Results
A total of at least twenty-five species of dragonfly were collected. While awaiting the
identification of the specimens, little can be ascertained regarding the dragonfly fauna
of the region.
56

11.5 Molluscs and Millipedes
11.5.1 Introduction
Madagascar has a land snail fauna currently considered to consist of about 1,200
species, although estimates suggest a figure as high as 4,000 species (Emberton,
2001). It is considered one of the most interesting snail faunas in the world. Twentythree
families and seventy-one genera are known. The dominant families are entirely
different from those found in Africa, and consist of the families Cyclophoridae,
Pomatiasidae, Acavidae and Ariophantidae.
In general, the humid calcareous areas of the north, west and south of Madagascar
have the richest mollusc fauna. Most species are found on or not far from the coast.
Few species survive in the arid regions. Southwest Madagascar is inhospitable to
terrestrial molluscs due to the intense heat and low rainfall. However, those species
that are found in the dry environments of this region show a high degree of adaptation
to this harsh environment. The most common adaptation is that of dormancy during
the dry season. Many species bury deep into the soil after the rains have ended, to
emerge following the onset of the next rains. This is especially true of species from
the genus Clavator, which are found in the sandy regions of southwest Madagascar.
Land snails can be useful indicators for conservation assessment, due to their
restricted distributions and ecological fragilities (Emberton, 2001).
The millipedes of Madagascar are very poorly studied. Collections are needed to
clarify the level of diversity and endemism at the regional and national level.
11.5.2 Methods
Mollusc and millipede surveys were undertaken prior to and during the onset of the
rains. Molluscs were sampled in three 1m x 1m quadrats per trapsite, over eight
trapsites. Millipedes were sampled in three 3m x 3m quadrats per trapsite, also over
eight trapsites. Each quadrat was established near a pitfall trap-line. They were
located in order to encompass a range of microhabitats.
Two man-hours were spent searching for molluscs in the leaf litter and top 15cm of
soil within each quadrat. All molluscs encountered were collected. Four man-hours
were spent searching for millipedes in the leaf litter and top 15cm of soil within each
quadrat. All millipedes encountered were collected. In addition, specimens were
collected opportunistically during other survey activities.
Live molluscs were placed in tobacco-water to kill them and to draw their bodies partway
out of the shell for ease of preservation and identification. Molluscs and
millipedes were stored in 70-95 % ethanol.
Ecological and geographical data were recorded for each quadrat, as described in the
Methodology Training Manual (Frontier Madagascar, 2003). Specimens were sent for
formal identification to Dr Brian Fisher of the Californian Academy of Sciences.
57

11.5.3 Results
A total of at least nine species of mollusc were collected. Field identifications showed
the presence of the giant African snail, Acatina fulica. This species has been
introduced from mainland Africa. Also present were unidentified species from the
genera Clavator and Tropidophora. While awaiting formal identification of the
mollusc specimens, little can be ascertained regarding the mollusc fauna of the region,
or their usefulness as indicators of habitats.
A total of at least sixteen species of millipede were collected, including two species of
pill millipede. While awaiting the identification of the specimens, little can be
ascertained regarding the millipede fauna of the region.
11.6 Spiders and Scorpions
11.6.1 Introduction
The spiders and scorpions of Madagascar are extremely diverse, yet comparatively
poorly studied. Little is known about the distribution of these species, many of which
are still undescribed.
The study aimed to identify spider and scorpion species present within the proposed
Parc Regional de Belomotse, in an attempt to increase the knowledge of spider and
scorpion diversity within the region.
11.6.2 Methods
Spider and scorpion specimens were collected through a variety of different methods.
Pitfall trapping provided the ideal framework to produce a comprehensive collection
with minimal effort, and specimens were collected during each trapsite. Crevice searches
also yielded specimens, as did daytime and night-time walks (the eyes of a spider reflect
blue or orange in torchlight).
Specimens were placed in at least 75 % ethanol to kill and preserve them, and stored
according to the locality from which they were collected. Ecological and geographical
data were recorded for each collection, as described in the Methodology Training
Manual (Frontier Madagascar, 2003).
Specimens were sent for formal identification to Dr. Brian Fisher of the Californian
Academy of Sciences.
11.6.3 Results
A total of at least four species of scorpion and thirty-two species of spider were
collected. While awaiting the identification of the specimens, little can be ascertained
regarding the scorpion and spider fauna of the region.
58

12. Resource use and Socio economics
12.1 Introduction
Madagascar’s economy is based mainly on agriculture, which employs about 80% of
the workforce, provides 70% of the countries export earnings and generates about one
third of the GDP. Madagascar produces a wide variety of crops both for consumption
and for export, predominantly rice, maize, cassava and sugar cane. Until 1971,
Madagascar produced enough rice to feed itself and have surplus for export. But in
1976, with the return of the president Didier Ratsiraka to power, rice fields were
nationalised. Although this lasted only 8 years, the production of rice dropped
dramatically. Combined with an increase in soil erosion due to deforestation, this has
meant that yields and production are so low that Madagascar now imports rice.
Madagascar was once the biggest exporter of vanilla in the world, which
predominantly went to the U.S for ice cream production. However, the world demand
for vanilla has decreased. This, combined with the Comoros undercutting the market,
has left coffee and cotton as Madagascar's main agricultural exports.
Madagascar has sold its large territorial fishing waters to Japan in return for money
and supplies of vehicles. However, prawns are still sufficiently plentiful to provide a
good source of export revenue.
Traditional communities throughout the island use resources from the surrounding
areas. The arid south-western areas of Madagascar are the least fertile, with only spiny
scrub growing naturally on the limestone plateau. However, along the few lakes and
rivers in the area water is constant and plentiful, creating a good basis for agriculture.
The Onilahy River is wide with broad strips of flat land lined with marshes and
riparian forest. Many of these areas have been cleared or drained, and cultivated by
the local communities. In addition, areas of spiny forest have been burned and cleared
for the cultivation of maize, cassava and peanuts.
The forested areas of southwest Madagascar are also under threat from clearance for
charcoal production and logging. In addition to these high-impact uses of resources,
hunting is also common throughout the region. Species of hunting interest include
birds, lemurs and tenrecs. Fishing is also a common practise in lakes, ponds and rivers
throughout the region, as well as intensive saltwater fishing. Medicinal plants and
natural food products, such as honey and fruits, are collected.
12.2 Methods
Areas of visible resource use were mapped, as were representative vegetation habitats
within the surveyed areas. In addition, non-visible uses of forest resources were
identified through semi-structured interviews with suitable individuals or groups from
local villages. Also, interviews were held to discuss and identify the agricultural
practices used within the region
59

12.2.1 Ground-truthing and GIS data layers
Ground-truthing and mapping data were collected for incorporation into geographic
information systems (GIS) as data layers. Specific layers of information were
collected.
Elements of human land use were mapped including village territories, settlements,
and access routes (roads, tracks, paths, and rivers). Wetlands and forested areas were
mapped. This included gallery forest, spiny forest, rivers, lakes, ponds, and marshes.
Areas of high biodiversity were mapped, including home ranges for diurnal lemur
species. Areas of land use were mapped including agricultural land on river flood
plains, areas of forest cleared for cultivation, and areas of forest impacted through
charcoal production and logging.
Surveyed areas were digitised from global positioning systems (GPS), with data
collected using Garmin 12’s.
Ground-truthing of forest edges and specific zones of resource use was conducted on
foot within the core area of the proposed Parc Regional de Belomotse, between the
villages of Antafoky and Ifanato. GPS points were collected at approximately 0.05 to
0.2 km intervals. The points were plotted onto FTM 1/100,000 Carte de Madagascar
regional maps. Forest boundaries and zones of resource use were traced. The map was
used to identify sites for further work and to provide maps for other survey and
reporting activities.
12.2.2 Biological diversity and habitat maps
Species inventory studies were conducted in different habitats and environments
likely to influence species distributions. These were gallery forests, spiny forests,
wetland areas and disturbed habitats. Altitudinal ranges were considered too small to
influence species distributions.
Trapsites were established at seventeen different localities within these habitats across
the surveyed areas. These sites were mapped and the faunal data were recorded for the
specific habitat, including mammal, reptile, amphibian, bird and invertebrate data.
This information was recorded for each of the trapsites, which were mapped, which
will allow distribution maps of faunal data to be produced throughout the region.
12.2.3 Human settlements, land, and resource use
Human settlements were mapped, with GPS points taken at the centre of each village
and around the periphery of the village. Surrounding agricultural or cleared land was
also mapped. Information was recorded for each village, including the number of
households, village population, and village status (permanent or temporary).
Socioeconomic and agricultural data were obtained from semi-structured interviews
with village groups and individuals from five villages. Data collected included aspects
of resource use, agricultural practices and restrictions on land use, including locally
managed areas, or areas protected by fady.
60

12.3 Results
12.3.1 Biological diversity and distribution
High levels of habitat specificity were observed from the biological inventory studies.
Species distribution ranges were identified for indicator taxa throughout the proposed
park area in different habitats. Faunal zonation was observed between spiny forest and
gallery forest habitats, with each habitat having a complement of restricted species.
The faunal species distribution data were examined to assess and identify the habitats
and areas of conservation management importance, if diversity and representivity is to
be maintained within the proposed regional park. In particular, species of special
concern were identified (locally endemic species or those currently classified as
vulnerable by IUCN) and their distributions were assessed and compared.
One locality (Antafoky Lake) was identified that contained all vulnerable mammal,
reptile and butterfly species. This locality also contained the highest diversity of bird
species. A second locality (the Sept Lacs region) was identified that contained all the
vulnerable mammal species. In total, three localities within the proposed regional park
contained all of the locally endemic, rare or vulnerable mammal, bird, reptile,
amphibian and butterfly species. These localities encompassed all vegetation zones,
geographical features and elevations.
This information is structured to allow incorporation into the GIS map, providing a
data layer for habitat-specific diversity and distribution of key faunal taxa.
12.3.2 Human populations and local management
A total of 5 permanent villages were observed within the core area of the proposed
regional park, all of which were situated along the Onilahy River. The average
number of inhabitants of these villages was approximately 120, with the largest
village, Ifanato, containing approximately 280 inhabitants. The total number of
inhabitants for these villages is approximately 800 individuals.
Village territories typically included areas of flat land on the Onilahy flood plain and
within forested valleys. The extent of the village territory was based on closeness to
other villages, and traditional and cultural factors such as the age of the village and
the location of fady areas and tombs. Four of the five village territories included
significant areas of agricultural land. The fifth village, Antafoky, has suffered from
high levels of erosion by the Onilahy River due to the location of the village on an
outer bend on the river. This has led to the recent loss of large areas of agricultural
land during floods.
The villages and associated areas of agricultural land were mapped and this
information recorded to allow incorporation into the GIS map.
Four of the five villages managed the surrounding forests through a programme of
local management, Gestion Local Securisee (GELOSE). The fifth village, Antafoky,
was outside of the designated GELOSE area.
61

12.3.3 Resource use
12.3.3.1 Spiny and gallery forests
Large areas of spiny forest on the plateau had been cleared for the cultivation of maize
and cassava. Access routes were found between villages and the agricultural areas on
the plateau. These are used to transport drinking water to villagers working in the
area, and to transport harvested crops down from the plateau.
Areas of gallery forest had been cleared for cultivation, mostly around Lake
Antafoky. All areas of gallery forest east of the village of Ifanato have been cleared
for at least 12 km along the river, leaving only small patches of trees around the edges
of agricultural land. The same is true of the gallery forest between the villages of
Ambohimahavelona and Antafoky. This means that the areas of gallery forest around
Lake Antafoky and along the river between Antafoky and the Sept Lacs region
represent the last remaining significant areas of gallery forest on the north bank of the
Onilahy River.
Many areas of gallery forest had been selectively logged for pirogues, building
material or charcoal production. Cut poles and charcoal pits were frequently observed.
Between the villages of Antafoky and Ifanato, within the GELOSE managed area,
almost all charcoal pits were old. However, recent or unburnt charcoal mounds were
regularly seen in the gallery forest and spiny forest around Antafoky Lake.
Recently cleared areas of forest were often further impacted by grazing from cattle
and goats, which were observed in dry and riparian forest in most of the surveyed
areas.
Natural forest products such as tamarind fruits and tubers were collected to
supplement the local diet and for sale. Plants were collected for medicinal use from
the spiny and gallery forests. However, due to the low population in the area, the
amount collected is low, so impacts are minimal. Honey was collected throughout the
year and was dependent on gallery forest and spiny forest at different times of the
year. Honey produced in the dry forests is regarded as very good quality and it is
collected for sale in Toliara.
Several species of animals are collected or hunted, either for food consumption locally
or for sale as food or pets. Hunting was observed in all surveyed areas. Specific
targets included birds, tenrecs and freshwater turtles. Young ring-tailed lemurs were
occasionally collected, often by children. These were kept in the villages or taken to
Toliara for sale.
12.3.3.2 Wetlands
All permanent villages within the surveyed areas are located near the Onilahy River.
The inhabitants of these villages fish in the river. During the rainy season, prawns are
collected in large numbers from the river.
Wetlands along the Onilahy River are threatened with destruction by drainage so that
the land can be used to cultivate rice. However, many permanent marshes and ponds
were found to be in good condition within the GELOSE managed area.
62

Reeds are collected from the shallow marshy areas. They are dried and used for
roofing. However, this activity seems to be sustainable at current levels. Selective
harvesting was observed from the expansive marshes around the village of
Manderano.
Lake Antafoky is deep and has steep-sloping banks around most of the lake,
preventing the cultivation of rice in these areas. The villagers from Antafoky catch
several species of fish from the lake, as well as freshwater turtles. Fish are sold to
nearby villages. There are several fady about the lake, often involving the small
population of resident Nile crocodiles. This has prevented excessive exploitation of
the resources from the lake.
All areas of gallery forest and wetlands were mapped and this information was
structured in order to allow incorporation into the GIS map as a habitat data layer.
12.3.4 Agriculture
The land along the Onilahy River is fertilised annually by alluvial deposits from the
river. Much of this land is used for agricultural purposes. The wet season floods
provide these areas with water at a known time. However, water can be a limiting
factor. If the rains are insufficient or excessively plentiful this will have a major
impact on yields. Many different crops are grown in the fertile floodplains, usually
giving a good yield to the local villages.
Table 11. Proportion of different crops grown, and the percentage that is sold.
Crop
Percentage of all Percentage sold
grown
Maize 33 70
Manioc 16 0
White beans 10 75
Sweet potatoes 10 30
Rice 8 0
Butter beans 5 70
Sugar cane 5 80
Tomatoes 4 90
Peanuts 4 55
Kidney beans 3 70
Pumpkin 1 17
Bananas 1 70
Maize is the most important crop grown in the area. It is produced in high quantities
and the surplus is sold. Manioc is also grown in large quantities and is the staple diet
of the villagers. Both of these crops are predominantly grown in cleared areas of spiny
forest on the plateau. Manioc is also grown in dry valleys.
Beans and other vegetables are grown on the Onilahy River flood plain. They are
grown in comparatively small quantities (see Table 11 and Figure 8). A proportion of
the harvested produce is consumed locally and the surplus is sold.
63

Proportional Crop Agriculture
MANIOC
RICE
MAIZE
WHITEBEANS
TOMATOES
SWEET POTATO
BUTTERBEANS
KIDNEY BEANS
SUGAR CANE
PUMPKIN
BANANAS
PEANUTS
Figure 8. Proportions of crops grown by the local villagers within the central Sept Lacs area of the
proposed Parc Regional de Belomotse.
All areas of agricultural land along the river between the villages of Antafoky and
Ifanato have been mapped. Areas of spiny and gallery forest that have been cleared
have also been mapped. This information was structured in order to allow
incorporation into the GIS map as a resource use data layer.
12.4 Discussion
The natural resources within the proposed regional park are used in a variety of
different ways. The sustainability of many of these resource use activities needs to be
assessed and the effects monitored. In particular, clearance of gallery forest needs to
be restricted and the remaining areas must be carefully monitored. The same is true of
the spiny forest on the plateau, although the loss of gallery forest will have longerterm
effects on the biodiversity of the region.
The areas of gallery forest around Antafoky Lake are suffering rapid rates of
clearance for agriculture, due to the lack of agricultural land along the Onilahy River.
This area has exceptional levels of biodiversity, both faunal and floral. The area is
outside of the GELOSE managed zone. With the current rate of deforestation, the
gallery forests around Lake Antafoky will rapidly become too small and fragmented
to harbour viable populations of many species. The use of forest resources around the
village of Antafoky must therefore be monitored, and alternative, sustainable
activities identified.
13. Tourism
13.1 Introduction
The proposed Parc Regional de Belomotse has great potential for tourism due to its
spectacular scenery, interesting geological features, varied terrestrial and wetland
habitats, abundant wildlife and fascinating local community and culture. The site,
which is close to the city of Toliara, has a relatively good infrastructure. Within the
proposed regional park is an existing small-scale tourism attraction, the ‘Sept Lacs
64

egion’. The development of the regional park would focus tourism on a larger area
with more features and involvement from more local communities.
Tourism has the potential to provide a workable alternative to non-sustainable
resource use activities and could contribute to sustainable conservation in the region.
However, in order for this to work it is vital that the operation is set up in a way that is
going to have minimal negative impacts on the environment and culture. In turn, the
proposed tourism operation must directly benefit the community that primarily relies
on the land for survival, and must be sufficiently beneficial to prevent the people
wanting to use the land in a more profitable way.
Despite the presence of unique and fascinating flora, fauna and scenery, the Toliara
region receives comparatively few tourists. Toliara is easily reachable both by road
and by air, and has a wide range of tourist accommodation both in Toliara and in
tourist resorts along the coast. The region contains many beautiful and interesting
natural sites that have great potential to become excellent tourist attractions if given
sufficient support, development, advertising and management. Increasing the number
of visitors to the area would contribute to the general income for the region and could
benefit the local communities.
13.2 Methods
During a series of field studies, the potential for tourism within the core area of the
proposed park region was assessed. This core area encompassed gallery forest, lakes
and rivers between the villages of Antafoky and Ifanato. All existing footpaths were
walked and notes were taken regarding the duration, level of difficulty, scenery and
wildlife. As a result of the systematic bird observations undertaken during the survey,
information regarding suitable locations for bird hides was easily extracted. Options
for possible campsites were assessed as the project moved around during the course of
the field studies and established several long-term base camps and short-term satellite
camps. Conducting field studies over a period of approximately 15 months provided
the opportunity to assess the infrastructure and logistical support available upon which
a tourist operation could be based.
The volunteer research assistants working on the project during the field studies
originated from a variety of backgrounds and countries. Many had never travelled
outside of Europe or the United States, which placed them in an excellent position to
identify features and attractions that would appeal to potential tourists, and to identify
those features that needed development or would not be suitable. These ‘eco-tourists’
contributed with objective opinions relating to potential tourism issues.
13.3 Results and Discussion
13.3.1 Priority areas for tourism
Within the region of the proposed Parc Regional de Belomotse, several areas were
located with potential for tourism. In particular, two areas were identified that would
be of interest to the vast majority of visitors. These were Antafoky Lake and the Sept
Lacs region (see Figure 9).
65

The Antafoky Lake area contains comparatively large, intact areas of gallery forest.
Many wildlife species that are of interest to tourists were found here – sightings of
Verreaux’s sifaka and ring-tailed lemurs could be virtually guaranteed. Dry spiny
forest and vast sandstone cliffs surround the lake and gallery forest, presenting a
dramatic biotic and abiotic contrast to the lake and surrounding forest. The spiny
forest contains many interesting bird and reptile species and geological formations.
The lake itself is a picturesque attraction, with opportunities for boat trips to observe
water birds, crocodiles and lemurs.
Camp 1
Antafoky
Ankotrofoty
and Toliara
Antafoky
track road
Befasy
Bevoalavo
MAHALEOTSE
Manderano
Andranomite
(fresh water
spring)
ONILAHY RIVER
Camp 2
Bat Colony
Walk 6
Walk 5
Antsoha
Sept Lacs
region
Ifanato
Figure 9. The core area of the park, of major interest to tourists
The Sept Lacs region is an existing tourist attraction that is currently visited by a low
but regular number of visitors throughout the year. The forest and lakes are of great
aesthetic appeal and swimming is permitted. A visit could easily be combined with
walks or boat-trips to one or more of the large colonies of bats that roost nearby.
The track that links Antafoky Lake with the Sept Lacs region is easy to walk,
scenically beautiful, and shaded for the most part. It also presents opportunities to see
local villages, lemurs and water birds en-route.
Given the spatial location of the main attractions of the park, locations for campsites
have been suggested both for Antafoky Lake and the Sept Lacs region, with a possible
improvement of pre-existing amenities en-route to make the walk between sites more
enjoyable for visitors.
13.3.2 Transport
A rough road runs through the proposed park from east to west, following the north
bank of the Onilahy River. This road is passable in a 4 x 4 vehicle, quad bike or
motorbike all the way to Sept Lacs, although it is in fairly poor condition in some
places. For those on a smaller budget, there is public transport from the centre of
Toliara to the village of Ankotrofoty, which is approximately 3 hours walk from
Antafoky Lake. Once in Antafoky, the transfer to the campsite could take place by
66

oat or on foot. From the lake, the walk to the Sept Lacs region takes a further 4
hours. Alternatively, it is possible to travel by zebu cart. The development of “luxury
zebu carts” would be encouraged, with the addition of padding and a sunshade. This
would enhance the travelling experience, while providing traditional transport to
tourists. These zebu chariots could collect tourists from the village of Ankotrofoty and
provide a system of transport throughout the core tourist area.
The walk from Antafoky Lake to the Sept Lacs region is fairly easy, with a low level
of physical fitness required. The route follows a flat road through gallery forest along
the Onilahy River. There are good opportunities to see wildlife along the route,
particularly ring-tailed lemurs, Verreaux’s sifaka, and a variety of reptiles and water
birds. Spectacular views across the river can be seen along this route. There are
several resting places for walkers, including a natural, spring-fed swimming pool just
before the village of Manderano (approximately 1 hour after Antafoky Lake) and a
freshwater spring at a shrine approximately 1 hour from the Sept Lacs region.
Luggage could be sent on ahead by zebu cart and the walk could be undertaken
unguided. The campsite suggested in the Sept Lacs region is adjacent to the road,
easing logistical difficulties for arriving tourists.
13.3.3 Campsite location
Providing an environmental experience should override the need for provision of
luxurious accommodation or sophisticated facilities. Many facilities could be kept
simple, to minimise local disturbance and long term impact.
When considering a campsite location, key criteria were considered. These included:
• Accessibility - the ease of getting to the campsite from the road
• Proximity to biodiversity – the level of biodiversity near the campsite and the ease
of observation
• Surroundings - the beauty of the scenery and immediate surroundings
• Level of disturbance – the amount of work needed to develop the campsite and the
short and long-term effects of this disturbance on local flora and fauna
Campsite 1 at Antafoky Lake, see Figure 9, is not easily accessible on foot, but
visitors and their luggage can easily be transported from the road to the campsite by
boat. It is possible to reach the proposed campsite on foot, although this would
involve wading through water on a number of occasions. Building walkways or
raising existing footpaths could alleviate this. The proposed campsite is within an area
of gallery forest that is rich in biodiversity. The area contains five lemur species and
many birds, and the surroundings provide natural beauty and shade. The site is in a
pre-existing clearing in gallery forest, but currently lacks any facilities.
Campsite 2, within the Sept Lacs region, see figure 9, is located near to the road on
the side of a small hill, near to the Ifanato River and associated lakes. This site has
been recently developed by WWF and has tent space for approximately twenty tents,
with simple shower and toilet areas. There is a small restaurant in the village of
67

Ifanato, 2 km from the campsite. The village also contains a small guesthouse
(hotely), providing an alternative for those who do not wish to camp.
13.3.4 Campsite design
To minimise disturbance to the fragile ecosystems within the proposed Parc Regional
de Belomotse, campsites should be kept small and fairly simple. Ideally, the campsites
would contain a series of areas for tents (open or covered), a covered cooking area
and table, washing facilities and a simple toilet. A covered pit latrine or basic toilet
would be suitable. A simple covered shower area would also be needed. Water could
be pumped in, or collected from the nearby watercourse in buckets.
Tourists could bring bottled drinking water from Toliara, or lake or river water could
be treated prior to drinking. All litter should be collected and taken out of the park
area for disposal, preferably by the tourists or tour guides.
There would not be a necessity for electrical power. A simple level of lighting, either
from candles or kerosene lamps, could provide illumination for the communal eating
and cooking areas.
13.3.5 Attractions and activities
13.3.5.1 Guided Walks
A variety of walks and wildlife-watching excursions are possible from both proposed
campsites, see Figures 9 and 10. In order to keep the disturbance on the environment
to a minimum, pre-existing paths can be used.
The proposed walks were identified based on several criteria. These included the
potential for viewing wildlife and other points of interest, the level of
ease/accessibility, the length and duration, and the predicted level of disturbance and
environmental impact. See Appendix 11 for a summary of the proposed walks.
Viewpoints were identified and photographed. See Frontier-Madagascar Tourism
Feasibility Study Report (in prep) for further details of walks, viewpoints and
viewable wildlife.
Due to the difficult terrain, fragile ecosystems and potentially dangerous wildlife
(there are crocodiles in Lake Antafoky), it is strongly recommended that guides are
compulsory for tourist groups. Also, the forests contain several areas of cultural
significance, including sacred tombs. It is imperative that the cultural beliefs and
traditions of the local communities are respected, so tourists should avoid these areas.
Guides would be necessary to identify these sacred areas and highlight the
significance of the sites. This would ensure that the villagers’ traditions are respected,
as well as providing cultural information for tourists.
68

Figure 10. The approximate position of walks and bird hides at
Antafoky Lake.
13.3.5.2 Guides
Two local guides, Enanetse Alfred and Mre Silvestre, have been identified and trained
in simple English and French. In addition, they have had wildlife observation training
and have had basic tourism etiquette explained. Both guides know the areas of forest
within the Sept Lacs area and the corridor of forest between Antafoky and the Sept
Lacs.
13.3.5.3 Boat Trips
There are several options for boat trips within the proposed park area. At Antafoky
Lake, boat-trips could be co-ordinated with several of the walks listed above, or they
could form a separate excursion. Many bird and lemur species can be easily be
observed and approached from the water, providing a different perspective to the
walks. A second option would be a boat-trip from the village of Ifanato, within the
Sept Lacs region, to a large colony of fruit bats within a limestone cave, in an area of
gallery forest on the south bank of the Onilahy. The cave can only be approached
from the water and the bats are easily visible. A third option would be to use pirogues
as a return route to Toliara. Tourists could sail down-river in a large pirogue to St.
Augustin, where they could take a taxi-brousse back to Toliara. Alternatively, from St
Augustin tourists could extend their tour by travelling south to the coastal tourist
resort of Anakao or Lac Tsimanampetsotse.
13.3.5.4 Bird Hides
Initially, the choice of bird hide location was to be quantified by comparing the time
spent at a particular site with the number and variety of bird species observed.
69

However, birding experience suggested that species in forest habitats were most easily
observed on foot, enabling birds to be followed. Also, the high density of vegetation
in the hillside scrub provided poor visibility for bird watching. Therefore, all of the
suggested bird hide locations are around Antafoky Lake.
See Appendix 12 for a summary description of the proposed bird hides.
13.3.6 Amenities within the area
There are a series of pre-existing amenities within the proposed park area that could
be adapted slightly to effectively provide tourist support. This would also provide a
mode by which local inhabitants could financially benefit from tourism.
Small shops and stalls are located in the villages of Antafoky, Mahaleotse and Ifanato,
selling basic items such as soap, cigarettes and sweets. At present, these shops are not
well indicated and visiting tourists may well not notice them. A simple sign could
solve this problem. The same applies to the basic “hotelys” which exist in Antafoky,
Mahaleotse and Ifanato, selling coffee and basic food. These could provide good
resting points for tourists travelling between Antafoky Lake and the Sept Lacs region,
or during a guided walk. A simple guesthouse and restaurant already exists in Ifanato.
These pre-existing facilities could be highlighted with park publicity and could
develop as the number of visiting tourists increase.
13.3.7 Potential Problems
One of the few current logistical problems facing the development of tourism within
the area is the location of the village toilet in Antafoky. At present, this is beside the
lake, along a path that tourists would use. It is assumed that tourists would be
extremely unhappy about walking through an open toilet, thus making a large
proportion of the lake inaccessible to visitors on foot. A possible relocation of the
toilet could be discussed with the villagers. A second toilet area is located along the
road, away from the proposed tourist circuits, so this could be developed further. This
would also probably enhance the health of the village, as the current toilet location is
polluting the drinking water of the village.
The access road passes through a corridor of gallery forest between Lake Antafoky
and the Sept Lacs region. This road is in poor condition and is in need of
improvement. However, the development of the road will cause further disturbance to
the gallery forest, which is already becoming degraded due to erosion caused by the
existing road. There is no other location for the road, so improvements need to be
undertaken with care. Also, improvements to the road could lead to an increase in the
rate of charcoal production and maize cultivation in the area, due to the improved
transportation route. These activities need to be monitored.
13.3.8 Monitoring
In order for the tourism operation to run in a sustainable manner with minimal
negative environmental and cultural impacts and maximum community benefits,
monitoring must take place. Without a monitoring system, negative impacts caused by
70

the development of tourism and associated infratructures may go unnoticed and
potential positive impacts may go unrecognised and undeveloped.
With a good monitoring system in place, key aspects of the tourism operation can be
observed. As a result of this system, an assessment could be made as to whether the
current management was working, and steps could be taken to mitigate any negative
impacts and alter aspects of operation management that are not effective.
See Frontier-Madagascar Forest Research Programme Tourism Feasibility Report
(in prep.) for further details.
14. Conclusions and recommendations
The proposed Parc Regional de Belomotse is an important site for conservation
management, supporting the maintenance of biodiversity and representivity within the
region and the conservation of biodiversity in Madagascar. The regional park, if
implemented, would protect approximately 470 ha of gallery forest. This fragile habitat
has been highlighted for conservation priority in the government Environmental Action
Plan for the region. Nationally, large areas of gallery forest are represented in very few
protected areas, specifically Beza Mahafaly Special Reserve and Berenty Private
Reserve. The proposed regional park would represent one of the largest areas of
protected gallery forest in Madagascar. In, addition, the proposed park area contains
significant areas of wetland, a habitat type that is uncommon in the dry forests of
southwest Madagascar.
The proposed regional park contains many rare, locally endemic or vulnerable species,
including flagship species such as Verreaux’s sifaka, Propithecus verreauxi verreauxi;
the ring-tailed lemur, Lemur catta; and the radiated tortoise, Geochelone radiata.
The surveys identified four principal direct threats to biodiversity in the core section
of the proposed regional park. These were slash-and-burn agriculture, both in the
gallery forests and spiny forests; drainage of wetlands for agricultural purposes; overharvesting
of specific tree species for pirogues, building material or charcoal
production; and over-harvesting of non-timber forest products (animals and plants).
Of these, slash-and-burn agriculture has the most long-term and severe effects,
causing destruction of spiny forest that will not regenerate. Large areas of grassland
on the plateau show that the progression of the habitat following spiny forest
clearance is conversion to grassland. In addition, the loss of gallery forest severely
threatens the survival of the many faunal species that are dependant on this habitat.
The gallery forests within the proposed regional park are the last fragments on the
north bank of the Onilahy River, so many dependant species would not be able to
migrate to other areas if these areas were cleared and fragmented. The loss of these
gallery forests would therefore cause the extinction of significant populations of
vulnerable and rare species.
71

The forests on the hillsides and plateau have an important role in watershed protection,
both for local people and the Onilahy River system. Erosion caused by deforestation
on the plateau will cause run-off that will have long-term affects on the ecosystem. In
addition, run-off will cause the sedimentation of the reefs that are found near the river
estuary.
Fortunately, there are a comparatively small number of villages and inhabitants within
the core region of the proposed regional park. In addition, much of this area is under
local GELOSE management, which appears to be maintaining a sustainable level of
resource use in the gallery forest and wetland areas. However, even in these areas the
spiny forest on the plateau is rapidly becoming degraded. In addition, the single largest
area of gallery forest, around Antafoky Lake, is not currently under GELOSE
management. This area contains the highest levels of biodiversity within the entire
proposed park area. If this area could be included within the GELOSE management
scheme, the gallery forest could be managed and monitored to prevent or reduce further
fragmentation.
The core region of the proposed regional park has great potential as a destination for
domestic and foreign tourists. This could create an alternative source of sustainable
revenue for local communities, while promoting environmental awareness within the
regional park. However, if tourism is to help support the conservation of the area,
sensitive management is needed incorporating the interests and concerns of all the
relevant stakeholders.
Many areas within the proposed regional park have suffered from human disturbance
in the past, although many of the disturbed areas still have the potential to return to a
natural condition. Replanting activities in specific areas could merge fragmented
habitats. In particular, specific disturbed areas of gallery forest around Lake Antafoky
and along the corridor between Antafoky and the Sept Lacs region could be the focus
of replanting activities, allowing the corridor to function effectively and linking
isolated forested areas.
There are certain issues, such as the inclusion of Antafoky within GELOSE
management and the sustainable management of the remaining areas of gallery forest,
which must be addressed rapidly. However, providing that this takes place, the
proposed Parc Regionale de Belomotse has a promising future.
72

References
Billé, R. & Mermet, L. 2002. Integrated coastal management at the regional level:
lessons from Toliary, Madagascar. Ocean & Coastal Management 45, 41-58.
Burney, D.A., James, H.F., Rafamantanantsoa, J., Ramilisonina, Wright, H.T., &
Cowart, J.B. 1997. “Environmental change, Extinction, and Human Activity:
Evidence from Caves in NW Madagascar”. Journal of Biogeography, Vol. 24.
Chapman, C., Fedigan, L. M. & Fedigan, L. 1988. A Comparison of Transect
Methods of Estimating Population Densities of Costan Rican Primates. Brenesia 30:
67-80.
Cook, J. M., Trevelyan, R., Wall, s. s., Hatcher, M & Rakotondraparany, F. 1991. The
ecology of Hypogeomys antimena, an endemic Madagascan rodent. Journal of
Zoology, London 224, 191-200.
Durbin, J. C., & Ralambo, J. A. 1994. The Role of Local People in the successful
Maintenance of Protected Areas in Madagascar. Environmental Conservation, vol. 21,
Nr. 2.
Dupuy, D. J. & Moat, J. 1996. A refines classification of the primary vegetation of
Madagascar based on the underlying geography: using GIS to map its distribution and
to assess its conservation status. In: Biogéographie de Madagascar. (ed. Laureço, W.
R.) Paris.
Ehrlich, P.R. 1998. Tomorrow’s world: why saving biodiversity is today’s priority.
World Birdwatch 10 (2): 6-7.
Emberton, K. C., (2001). Exploratory phylogenetic and biogeographic analyses within
three land snail families in southeastern-most Madagascar. Biological Journal of the
Linnean Society, 72: 567-584.
Fergusson, B. 2002. Opportunities for Community Based Conservation in the Androy
Region of Madagascar’s Southern Spiny Forests. Non published paper.
Fisher, B.L., 1999. Improving inventory efficiency: a case study of leaf litter ant
diversity in Madagascar. Ecological Applications, 9: 714-731.
Frontier Madagascar 2002. Methodology Training Manual. Edited by Emmet, D.,
Olsson, A. and Bloxam, C. Non published document.
Ganzhorn, J. U., Fietz, J., Rakotovao, E., Schwab, D. & Zinner, D. 1999. Lemurs and
the Regeneration of Dry Deciduous Forest in Madagascar. Conservation Biology, vol.
13, No. 4, pp. 794-804.
Garbutt, N. 1999. Mammals of Madagascar. Pica Press.
73

Glaw, F., & Vences, M. 1994. A Field Guide to the Reptiles and Amphibians of
Madagascar. Pica Press.
Glaw, F., & Vences, M. 2002. Molecular phylogeny of Boophis tephraeomystax: a
case test for east-west vicariance in Malagasy anurans (Amphibia, Anura,
Mantellidae). Spixiana, Vol 25 (1): 79-84.
Goodman, S. M. 1996. Results of a Bat Survey of the Eastern Slopes of the Réserve
Naturelle Intégrale d’Andringitra, Madagascar. In: Goodman, S. M., ed., A floral and
faunal inventory of the eastern slopes of the Réserve Naturelle Intégrale
d’Andringitra, Madagacsar: With reference to elevational variation. Fieldiana:
Zoology, n.s., 85: 1-319.
Goodman, S.M. 1995. Rattus on Madagascar and the Dilemma of Protecting the
Endemic Rodent Fauna. Conservation Biology Volume 9, No. 2, April 1995.
Goodman, S.M., M.D. Carleton & M. Pidgeon, 1999. Rodents of the Reserve
Naturelle Integrale d’Andohahela, Madagascar. Pp. 217-249, In Goodman, S.M.
(Ed.), A floral and faunal inventory of the Reserve Naturelle Integrale d’Andohahela,
Madagascar: with reference to elevational variation. Fieldiana: Zoology, new series,
94.
Hawkins, F. & Morris, P. 1998. Birds of Madagascar, A Photographic Guide. Pica
Press.
Hutcheon, J. M. 1994. The Great Red Island: A Future for its Bats BATS, Vol. 12,
No. 2.
Kremen, C., 1994. Biological inventory using target taxa: a case study of the
butterflies of Madagascar. Ecological Applications, 4(3), pp. 407-422.
Kremen , C., Razafimahatratra, V., Guillery, R. P., Rakotomalala, J., Weiss, A. and
Ratsisompatrarivo, J. 1999. Designing the Masoala National Park in Madagascar
Based on Biological and Socioeconomic Data. Conservation Biology Vol. 13, No. 5.
Kull, Christian A. 1996. “The Evolution of Conservation efforts in Madagascar.”
International Environmental Affairs, Vol. 8, No. 1.
IUCN 2002. Red List of Threatened Species. IUCN, Gland, Switzerland.
Janzen, D.H. 1988. In E. O. Wilson & F. M. Peter (eds), Biodiversity, pp. 130-137.
National Academy Press, Washington, D.C.
Lowry, P.P., Schatz G.E., & P.B. Phillipson. 1997. The Classification of Natural and
Anthropogenic Vegetation in Madagascar. pp. 93-123. In: Goodman, S.M. and B.D.
Patterson, eds., Natural Change and Human Impact in Madagascar. Smithsonian
Institution Press, Washington, D.C.
74

Lundahl, S. L. L. & Olsson, A. 2000. Small Mammal Communities in Lowland
Rainforest of Krau Wildlife Reserve, Peninsular Malaysia- with reference to human
disturbance. MsC. Thesis, unpublished. University of Copenhagen.
Martin, P. (eds.) 1984. Prehistoric Overkill: The global Model. In Quaternary
Extinctions: A prehistoric Revolution. Tucson: University of Arizona Press.
MacPhee, R. D. E. 1986. Environment, Extinction, and Holocene Vertebrate
Localities in Southern Madagascar. National Geographic Research 2 (4):441-455.
Morris, P. & Hawkins, F. 1998. Birds of Madagascar. A Photographic Guide. Pica
Press.
Myers, N., Mittermeier, R.A., Mittermeier, C.G., de Fonseca, G.A.B., Kent, J. 2000.
Biodiversity Hotspots for Conservation Priorities. Nature, 403, 853-858.
Nicoll, M. E. & Langrand, O. 1989. Madagascar: Revue de la Conservation et des
Aires Protégées. WWF, Gland, Switzerland.
Rabinowitz, A. 1997. Wildlife Field Research and Conservation Training Manual.
Wildlife Conservation Society
Rasoloarison, R. M., Goodman, S. M. & Ganzhorn, J. U. 2000. Taxonomic Revision
of Mouse Lemurs (Microcebus) in the Western Portions of Madagascar. International
Journal of Primatology, Vol. 21, No. 6.
Raxworthy, C.J., Nussbaum, R. A. 1994. A rainforest survey of Amphibians, Reptiles
and Small mammals at Montagne D’Ambre, Madagascar. Biological Conservation
69, pp 65-73.
Raxworthy, C. J. 1995. Field survey of Amphibians, Reptiles and Small Mammals
from the Toliara Region, Madagascar. Unpublished field Report.
Richard, Alison F., and Shiela O’Connor. 1997. “Degradation, Transformation, and
Conservation: The Past, Present, and Possible Future of Madagascar’s Environment.”
In Natural Change and Human Impact in Madagascar, edited by Steven M. Goodman
and Bruce D. Patterson. Washington, D.C.: Smithsonian Institution Press.
Rogers, H. (Eds) 2001. Madagascar. Lonely Planet Publications Pty Ltd.
Seddon, N., Tobias, J., Yont, J. W., Ramanampamonjy, J. R., Butchart, S. &
Randrianizahana, H. 2000. Conservation issues and priorities in the Mikea Forest of
south-west Madagascar. Oryx Vol 34 No. 4.
Schatz, G.E. 2000. Endemism in the Malagasy tree flora. Pp. 1-9, In: W. R. Lourenco
& S. M. Goodman (eds), Diversite et Endemisme a Madagascar, Memoires de la
Societe de Biogeographie, Paris.
75

Schatz, G. E. 2001. Generic Tree Flora of Madagascar. Royal Botanical Gardens,
Kew & Missouri Botanical Garden. Cromwell Press.
Sinclair, I. & Langrand, O. 1998. Birds of the Indian Ocean Islands. Struik Pblishers
(Pty) Ltd. Cape Town.
Smith, A. P. 1997. Deforestation, fragmentation and reserve design in western
Madagascar. In: Tropical Forest Remnants, Ecology, Management and Conservation
of fragmented Communities (eds. Lawrence, W. & Bierregaard, O. W.). University of
Chicago press, Chicago.
Stephenson, P. J. 1994. Seasonality effect on small mammal trap success in
Madagascar. Journal of Tropical Ecology 10:439-444.
Sussman, R. W. & Rakotozafy, A. 1994. Plant Diversity and Structural Analysis of a
Tropical Dry Forest in Southwestern Madagascar. Biotropica 26 (3): 241-254.
Sussman, R. W., Richard, A. F. & Ravelojaona, G. 1985. Madagascar: Current
Projects And Problems In Conservation. Primate Conservation, No. 5.
Whitesides, G. H., Oates, J. F., Green, S. M. & Klubendanz, R. P. 1988. Estimating
Primate Densities from Transects in a West African Rain Forest: A Comparison of
Techniques. Journal of Animal Ecology, 57, 345-367.
Wilson, E. O. (Ed), 1992. The Diversity of Life. Harvard University Press,
Cambridge.
Wilson, D. E., Cole, F. R., Nichols, J. D., Rudran, R. & Foster, M. S. 1996.
Measuring and Monitoring Biological Diversity. Standard Methods for Mammals.
Smithsonian Instistution.
WWF 2002. Action Plan 2002-2005. WWF-Madagascar Dry Forest Programme,
Toliara.
76

Appendix 1.
Trapsite (TS) Habitat Descriptions
TS 1:
Antafoky (Andamalamy). Secondary riparian forest. Some old selective logging.
Dominated by tamarind (Tamarindus indica) and fig (Ficus) trees. Some large (>30
cm DBH) trees. Good regeneration with many saplings.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 1-60 metres.
TS 2:
BL 1+2:
BL 3:
Antafoky (Tongomboa).
Secondary riparian forest. Some old selective logging. Dominated by tamarind
(Tamarindus indica) and fig (Ficus) trees. Some large (>30 cm DBH) trees. Good
regeneration with many sapling.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 1-70 metres.
Cultivated area. Slash and burn, two to three years old maize field. No live trees, soil
powdery and mostly ash. Ground, scrub, and canopy layer 50 %. Canopy height 15-20 metres. Distance to water
(stream/river) 0-50 metres.
TS 5:
Antafoky (Andamalamy). Secondary riparian forest. Some old selective logging.
Dominated by tamarind (Tamarindus indica) and fig (Ficus) trees. Some large (>30
cm DBH) trees. Good regeneration with many sapling.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 10-100 metres.
TS 6:
BL 1+2:
BL 3:
Antafoky (Tongomboa).
Secondary riparian forest. Some old selective logging. Dominated by tamarind
(Tamarindus indica) and fig (Ficus) trees. Some large (>30 cm DBH) trees. Good
regeneration with many saplings.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 2-60 metres.
Riparian forest. Highly disturbed from recent slash and burn. Some live trees.
Ground layer < 10 %, scrub layer < 10 %, canopy layer 10-50 %. Distance to water ~
70 metres.
TS 7: Sept Lacs.
BL 1+2: Transitional forest, some selective logging. Ground layer < 10 %, scrub layer 10-50
%, canopy layer > 50 %. Canopy height 5-10 metres. Distance to water (stream/river)
2-100 metres.
BL 3: Thorny scrub, hilltop, some selective logging. Ground layer < 10%, scrub layer >
50%, canopy layer < 50%. Canopy height 3-4 metres. No water association.
77

TS 8:
Sept Lacs. Tall riparian forest, some selective logging. Valley along road. Some large
(DBH > 30 cm.) trees, good regeneration. Ground layer < 10%, scrub layer < 10 %,
canopy layer > 50 %. Canopy height 15-20 metres. Distance to marsh 5-80 metres.
TS 9:
BL 1+3:
BL 2:
Antafoky (Andamalamy).
Secondary riparian forest. Some old selective logging. Dominated by tamarind
(Tamarindus indica) and fig (Ficus) trees. Some large (>30 cm DBH) trees. Good
regeneration with many saplings.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 3-50 metres.
Transitional scrub/forest. Some selective logging. Rocky and dry.
Ground layer < 10 %, scrub layer 10-50 %, canopy layer > 50 %. Canopy height ~ 5
metres. No water association.
TS 10:
BL 1:
BL 2:
BL 3:
Antafoky (Tongomboa)
Secondary riparian forest. Some old selective logging. Dominated by tamarind
(Tamarindus indica) and fig (Ficus) trees. Some large (>30 cm DBH) trees. Good
regeneration with many sapling.
Ground layer < 10 %, scrub layer < 10 %, canopy layer > 50 %. Canopy height 15-20
metres. Distance to water (lake) 3-100 metres.
Transitional scrub. Hillside. Some old charcoal production. Sign of Zebu grazing.
Ground layer < 10 %, scrub layer > 50 %, canopy layer ~ 50 %. Canopy height ~ 3-4
metres. No water association.
Transitional forest. Hilltop. Some old selective logging. Ground layer < 10 %, scrub
layer 10-50 %, canopy layer > 50 %. Canopy height 5- 7 metres. No water
association.
TS 11:
Sept Lacs. Transitional forest. Rocky area, some succulents. No tamarind/fig. Mostly
small, new trees (DBH < 10), but some large trees (DBH > 10). Ground layer < 10%,
scrub layer 10-50 %, canopy layer > 50%. Emergents 10-15 metres. Distance to
stream/river 3-200 metres.
TS 12: Manderano.
BL 1+2: Tall riparian forest, some selective logging and some grazing. Large tamarind and
fig trees, some saplings and regeneration. Ground layer < 10%, scrub layer < 10%,
canopy layer > 50%. Canopy height 15-20 metres. Distance to marsh ~ 20-40 metres.
BL 3: Transitional forest, some selective logging. Ground layer < 10%, scrub layer < 10%,
canopy layer 10-50%. Canopy height 5-10 metres. No water association.
TS 13: Manderano, valleys along road towards Antafoky.
BL 1: Riparian forest. Some selective logging. Ground layer < 10%, scrub layer < 10%,
canopy layer > 50%. Canopy height 10-15 metres. Distance to marsh ~ 50 metres.
BL 2: Riparian forest. Disturbance from logging and charcoal burning. Ground layer < 10%,
scrub layer < 10%, canopy layer > 50%. Canopy height 10-15 metres. Distance to
marsh ~ 50 metres.
TS 14:
BL 1:
BL 2:
Antafoky (Andamalamy)
Secondary riparian forest. Some old selective logging. Dominated by tamarind
(Tamarindus indica) and fig (Ficus) trees. Some large (>30 cm DBH) trees. Good
regeneration with many saplings. Ground layer < 10 %, scrub layer < 10 %, canopy
layer > 50 %. Canopy height 15-20 metres. Distance to water (lake) 1-60 metres.
Transitional scrub/forest. Some selective logging. Rocky, dry.
78

Ground layer < 10 %, scrub layer 10-50 %, canopy layer 50 %. Canopy height ~ 5
metres. No water association.
TS 15:
BL 1:
BL 2:
Antafoky (Tongomboa).
Secondary riparian forest. Some old selective logging. Partly maize cultivation in
between riparian forest trees. Ground layer < 10 %, scrub layer < 10 %, canopy layer
10-50 %. Canopy height 10-15 metres. Distance to water (lake) 1-70 metres.
Old cultivated area. Slash and burn, two to three years old maize field. Some live
trees, soil powdery and mostly ash. Ground, scrub, and canopy layer 10-50 %. No
water association.
TS 16: Manderano.
BL 1: Tall riparian forest, some selective logging and some grazing. Large tamarind and
fig trees, some saplings and regeneration. Ground layer < 10%, scrub layer < 10%,
canopy layer > 50%. Canopy height 15-20 metres. Distance to marsh ~ 20-40 metres.
BL 2: Transitional forest, some selective logging. Rocky. Ground layer < 10%, scrub layer
< 10%, canopy layer 10-50%. Canopy height 5-10 metres. No water association.
TS 17 :
BL 1:
Sept Lacs.
Transitional forest. Some selective logging. Rocky. Ground layer

Appendix 2.
Numbers of individuals of plants encountered in eight 16 x 16m quadrats in
gallery forest. (Nomenclature based on Schatz, 2001.)
Tree DBH
Vernacular name Family and species
ANACARDIACEAE
Sohihy Abrahamia grandidieri 1
ASCLEPIADACEAE
Bokabe Mardenia cordifolia 1
BIGNONIACEAE
25cm
Somontsoy Fernandoa madagascariensis 5 2 1
BORAGINACEAE
Varo Cordia varo 3 2
BRASSICACEAE
Paky Boscia madagascariensis 1
Kalaogna Crateva excelsa 1 2 2
BURSERACEAE
Daro Commiphora sp1 1 2
CELASTRACEAE
Reampy Buxiella sp 3 1
Tsingilofilo Maytenus linearis 4
CELTIDACEAE
Tsiambanilaza Celtis bifida 99 47 15 9 3
COMBRETACEAE
Vahifoty Combretum sp. 1
Talinala Terminalia rhopalophora 1
EBENACEAE
Maintifototse Diospyros sakalavarum 68 23 12 2
ERYTHROXYLACEAE
Montsoy Erythroxylum sp. 17 1
EUPHORBIACEAE
Hazombalala Croton mahafaliens 7 1 3
Pisopiso Croton sp1 5
Hazondimohy Drypetes sp. 16 16 8 2
Laro Euphorbia laro 1 1 3 2
Voafogna Flueggea virosa virosa 1 1
Hazomena Securinega seyrigii 1
Fitrekantsifake 19 12 4 3 1
80

Appendix 2. Continued.
FABACEAE
Robontsy Acacia morondavensis 2
Mendoravy Albizia greveana 1 1
Bagnake Bauhinia hildebradtii 6 9 1
Sarongaza Colvillea racemosa 4
Magnary Dalbergia greveana 1
Kily Tamarindus indica 5 2 20
HYPPOCRETACEAE
Vahimpindy Hyppocratea angustipetala 4 2
LAMIACEAE
Tsilaitsora Capitanopsis angustifolia 39 2
LOGANIACEAE
Relefo Strychnos decussata 5 1
Bakoa Strychnos madagascariensis 8 4 1
MALVACEAE
Malimatsy Grewia bororum 1
Katepoky Grewia grevei 1
MELIACEAE
Hompy Quivisianthe papinae 1
Mampisaraky Turraea sericea 1 3 1
MORACEAE
Mandresy Ficus mormorata 1 1 1
Fihamy Ficus pyrifolia 1
OLACACEAE
Tsilaitsy Olax lanceolata 37 26 30 6
Tanjaky Olax pervilleana 12 5 7
PANDANACEAE
Fandra Pandanus sp. 1
PAPILIONACEAE
Fangitsy Dolichos fangitsy 1
PHYSENACEAE
Fandriandambo Physena sessiliflora 13 4
RHAMNACEAE
Roihavitse Scutia sp. 3
Tapisaky Xericisyos dangui 1
81

Appendix 2. Continued.
RUBIACEAE
Beholitse Hymenodictyon decaryum 1
Voafotake Rothmannia' sp. 2 2
Mantsake Tarenna pruinosum 7 1 4
Lamotimboay Catunaregam spinosa 1
Litsake 2 1
RUTACEAE
Monongo Zanthoxylum tsihanimposa 1
SALICACEAE
Lalipito Bivinia jalberti 3 3 3
SAPINDACEAE
Matitihena Stadmania oppositifolia 18 3 2
SPHAEROSEPALACEAE
Talafoty Rhopalocarpus lucidus 1 3 1
VITACEAE
Laza Cyphostema laza 1 1
UNIDENTIFIED
Bika 1
Diamainty 1
Filankibo 14 4
Hazomposa 1
Malaignarety 1
Roimena 1
Sandrisandry 11 10
Sarevaro 1 1
Tagnatagna 1
Tagnatagnanala 2
Tambakobe 1
Tamenaky 4
Tandrokaondrilahy 1
Timbatse 1
Tsimatindrahavelo 1
82

Appendix 3.
Numbers of individuals of plants encountered in eight 32 x 32m quadrats in
Spiny forest. (Nomenclature based on Schatz, 2001.)
Tree DBH
Vernacular
name
Family and species 25cm
ANACARDIACEAE
Jabihy Operculicarya decaryi 16 6 11 26
Sakoakomoke Poupartia birrea caffra 2 8
ASCLEPIADACEAE
Ranga Cynancum perieri 4
Try Cynancum sp. 1
Taritariky Leptadenia madagascariensis 1
Bokabe Mardenia cordifolia 1
ASTERACEAE
Malaignevotsy Psiadia altissima 1 1
BIGNONIACEAE
Somontsoy Fernandoa madagascariensis 4 2 1
Mangarahara Stereospermum euphoroides 3
BORAGINACEAE
Lambotaho Erhetia ambovombes 17 6 3
BRASSICACEAE
Paky Boscia madagascariensis 14 8 1 3
Maharoake Boscia sp. 8 3 1 2
Mahafangalitse Thilachium monophyllum 5 1
BURSERACEAE
Mamy Canarium madagascariensis 1
Arofy Commiphora arafy 2 5 10 8 3
Sengatsy Commiphora simplicifolia 8 1 3 7
Daro Commiphora sp1 27 21 14 16 2
Tariby Commiphora sp2 43 26 21 3
Daromena Commiphora sp3 1
CELASTRACEAE
Reampy Buxiella sp 1
COMBRETACEAE
Kapikala Combretum grandidieri 1 2
Talinala Terminalia rhopalophora 1 4 1 2
Taly Terminalia tricristata 3 3
CONNARACEAE
Vahimainty Agelaea peutagyna 2
CONVALLARIACEAE
Tavevola Dracaena sp 31 2
83

Appendix 3. Continued.
CYCADACEAE
Fatra Cycas thouarsii 27 18 3 1
EBENACEAE
Hazonta Diospyros humbertiana 9
Kobaintsihotse Diospyros latispatula 100 44 17
Maintifototse Diospyros sakalavarum 13 5
ERYTHROXYLACEAE
Montsoy Erythroxylum sp. 2 3
EUPHORBIACEAE
Hazombalala Croton mahafaliens 14 17 9
Andriambolafotsy Croton sp.1 70 1 1
Somoro Croton sp.2 43 6
Pisopiso Croton sp.3 178 2
Karimbola Croton sp.4 32 5 1
Teloravy Dalechampia elematidifolia 1 1
Antso Euphorbia antso 23 9 9 7
Pira Euphorbia intisy 37 24 15 1
Laro Euphorbia laro 3 1 1
Betondro Euphorbia oncoclada 11 6 4
Famata Euphorbia sp 3 1
Avoha Jatropha curcus 37 5 9 1
Katratra Jatropha mahafaliensis 28 13 4 2
Hazomena Securinega seyrigii 45 32 14 5
FABACEAE
Fatipatike Acacia angustifolia 9 3 2
Ambilazo Alantsilodendron sp. 49 53 37 14
Mendoravy Albizia greveana 6 2 2 1 1
Hafobakondry Bauhinia grandidieri 101 11
Bagnake Bauhinia hildebradtii 6 2
Kitomba Caesalpinia sp 1 1
Sarongaza Colvillea racemosa 1 2
Fengoke Delonix sp 2 4 1 1
Lovanafy Dicraeopetalum mahafaliense 35 21 2 1
Roy Mimosa sp. 15 7 1
Talamena Neoapaloxylon sp 5 2 3 2
Sofasofa Ormocarpopsis sp 10 2
Vaovy Tetrapterocarpon geayi 30 7 3 1
Tsikidrakitsy Voganhia mahafaliensis 6 6
Kapaipoty
HERNANDIACEAE
Gyrocarpus americanus
capuronianus
1 3 5 1
HYPPOCRETACEAE
Vahimpindy Hyppocratea angustipetala 6
LILIACEAE
Vaho Aloe divaricata 1
Vahontsoy Aloe sp. 2 1
84

Appendix 3. Continued.
LOGANIACEAE
Relefo Strychnos decussata 2 1
LYTHRACEAE
Lelaosy Woodfordia fructicosa 2 1
MALVACEAE
Tainkafotsy Dombeya ambongoensis 12 8 4
Satro Dombeya sp 16 5 5 1
Malimatsy Grewia bororum 13 4 4 1
Katepoky Grewia grevei 140 19 3
Selibohoky Grewia sp. 1 2
Sely Grewia sp1 2 1
Ampelambatotse Helmiopsiella madagascariensis 1
MELIACEAE
Handimbohitse Neobeguea mahafaliensis 3
MORACEAE
Fihamy Ficus pyrifolia 1
Nonokamboa Ficus sp. 1
OLACACEAE
Tsilaitsy Olax lanceolata 1 1 1
Tanjaky Olax pervilleana 11 3 1
PANDANACEAE
Fandra Pandanus sp. 1
PAPILIONACEAE
Fangitsy Dolichos fangitsy 2
Malaikifafa Indigofera tinctoria 1
Kifafa Plubango aphylla 6
PASSIFLORACEAE
Hola Adenia holaboainaris 1 1
PEDALIACEAE
Farehitse Uncarina sp. 8 2
PHYSENACEAE
Fandriandambo Physena sessiliflora 3
PTAEROXYLACEAE
Katrafaifilo Cadrelopsis grevei 2 1
RHAMNACEAE
Roihavitse Scutia sp. 1
Tapisaky Xericisyos dangui 3 2
RUBIACEAE
Lamotimboay Catunaregam spinosa spinosa 1
Voligejy Gardenia rutenbergiana 4
85

Appendix 3. Continued.
RUBIACEAE (cont.)
Kaposera Poederum grevei 50 22
Voafotake Rothmannia' sp. 1 4 1
Lavaravy Tarenna sp. 1
RUTACEAE
Katrafay Cedrelopsis gracilis 121 25 7 3
Katrafaintsifake Cedrelopsis simplifolia 1
Monongo Zanthoxylum tsihanimposa 3 1
SALICACEAE
Lalipito Bivinia jalberti 2
SAPINDACEAE
Matitihena Stadmania oppositifolia 3 1 1
SPHAEROSEPALACEAE
Talafoty Rhopalocarpus lucidus 4 1 3
Lombiry Rhopalocarpus sp 2
UNIDENTIFIED
Andriambelo 4 5 1
Atolimpoly 28 4
Berotsy 7 3 1
Boramena 2 1
Eringitse 12
Filankibo 2
Hatambogna 5 3
Hazofio 33 12 5 1
Hazofoty 51 16 2 1
Hazomamy 7
Hazondragnaty 2 7 1
Hazonkoaky 1
Hazovola 5 1
Katrafay 43 8 3
Kopitse 3
Lamendra 1
Malaitambio 102 9 1
Roimena 1
Sandrisandry 89
Saresakoa 10 4 4 2
Sariboy 1
Sp1 22
Tainkoaky 5
Tambarikosy 3 1
Tandrokaondrilahy 1
Timbatse 1
Tsianagnampo 1
Tsimarefe 1 2
Tsimatindrahavelo 1
Vahimena 2 2
Valae 1
Volomboto 6 13 6 1
86

Appendix 4.
Plant numbers and specific uses – Sept Lacs (west)
Family Genus and species Notes
< 2.5cm 2.5 - 5cm 5 - 10cm 10 - 25cm >25cm
Gallery forest
SALICACEAE Bivinia jalberti Wood - construction 1
BRASSICACEAE Boscia madagascariensis Edible fruit, leaves - rheumatism 1
FABACEAE Colvillea racemosa Wood - construction, pirogues 4
BURSERACEAE Commiphora sp1 Wood - construction 1 2
FABACEAE Dalbergia greveana Wood - construction & furniture 1
EBENACEAE Diospyros sakalavarum Fruit - edible for lemurs 63 20 11 2
MORACEAE Ficus mormorata Fruit - edible for lemurs 1 1 2
LOGANIACEAE Strychnos madagascariensis Fruit - edible for lemurs and man 8 4 1
FABACEAE Tamarindus indica Fruit - edible for man and lemurs 1 3 17
Spiny forest
BRASSICACEAE Boscia madagascariensis Edible fruit, leaves - rheumatism 9 4 1 2
FABACEAE Colvillea racemosa Wood - construction, pirogues 1 2
BURSERACEAE Commiphora arafy Wood - construction 2 5 10 7 3
BURSERACEAE Commiphora simplicifolia Wood - construction 6 1 3 7
BURSERACEAE Commiphora sp1 Wood - construction 3 6 2 4 1
BURSERACEAE Commiphora sp2 Wood - construction 1
BURSERACEAE Commiphora sp3 Wood - construction 2 6 6 1
FABACEAE Delonix sp Resin - glue, wood - pirogues 1 3 1 1
EBENACEAE Diospyros humbertiana Fruit - edible for lemurs 3
EBENACEAE Diospyros latispatula Fruit - edible for lemurs 66 31 7
EBENACEAE Diospyros sakalavarum Fruit - edible for lemurs 4
EUPHORBIACEAE Jatropha mahafaliensis Whole plant - living fence 3 3 2
FABACEAE Neoapaloxylon sp Root - edible, bark - rope 5 2 3 2
FABACEAE Ormocarpopsis sp Wood - construction & tool handles 4 2
ANACARDIACEAE Poupartia birrea caffra Fruit - edible 4 8
PEDALIACEAE Uncarina sp. Roots - medicinal 8 2
87

Appendix 5.
Plant numbers and specific uses – Antafoky
Family Genus and species Notes
< 2.5cm 2.5 - 5cm 5 - 10cm 10 - 25cm >25cm
Gallery forest
SALICACEAE Bivinia jalberti Wood - construction 2 3 3
EBENACEAE Diospyros sakalavarum Fruits - eaten by lemurs 5 3 1
FABACEAE Tamarindus indica Fruits - eaten by lemurs and man 4 3
Spiny forest
SALICACEAE Bivinia jalberti Wood - construction 2
BRASSICACEAE Boscia madagascariensis Edible fruit, leaves - rheumatism 5 4 1
BRASSICACEAE Boscia sp. Edible fruit, leaves - rheumatism 8 3 1 2
BURSERACEAE Canarium madagascariensis Wood - pirogues, fruit - eaten by lemurs 1
BURSERACEAE Commiphora arafy Wood - construction 1
BURSERACEAE Commiphora simplicifolia Wood - construction 2
BURSERACEAE Commiphora sp.3 Wood - construction 34 19 15 2
BURSERACEAE Commiphora sp2 Wood - construction 7 1
BURSERACEAE Commmiphora sp 1 Wood - construction 24 15 12 12 1
FABACEAE Delonix sp Resin - glue, wood - pirogues 1 1
EBENACEAE Diospyros humbertiana Fruits - eaten by lemurs 6
EBENACEAE Diospyros latispatula Fruits - eaten by lemurs 34 13 10 1
EBENACEAE Diospyros sakalavarum Fruits - eaten by lemurs 9 5
MORACEAE Ficus pyrifolia Fruits - eaten by lemurs 1
MORACEAE Ficus sp. Fruits - eaten by lemurs 1
EUPHORBIACEAE Jatropha mahafaliensis Whole plant - living fence 15 10 4
FABACEAE Ormocarpopsis sp Wood - construction and tool handles 6
COMBRETACEAE Terminalia rhopalophora Wood - construction 1 4 1 2
88

Appendix 6.
List of mammal species found in the proposed Parc Regional de Belomotse.
HABITAT
No. SPECIES Method Gallery Transitional Endemic
forest scrub / forest Status
Family: Pteropopidae
1 Eidolon dupreanum O E V
Family: Hipposideridae
2 Hipposideros commersoni C, O E
3 Triaenops furculus C E V
4 Triaenops rufus C E
Family: Vespertilionidae
5 Myotis goudoti C E
6 Miniopterus sp C E
7 Pipistrellus sp C E
Family: Molossidae
8 Tadarida leucogaster* C E
Family: Tenrecidae
9 Tenrec ecaudatus C, O E
10 Setifer setosus C E
11 Echinops telfairi C E
12 Geogale aurita C E
13 Microgale sf brevicaudata C E
Family: Soricidae
14 Suncus madagascariensis C N
Family: Muridae
15 Eliurus myoxinus C, O E
16 Rattus rattus C, O N
17 Mus musculus C, O N
Family: Viverridae
18 Cryptoprocta ferox T E En
19 Viverricula indica O N
Family: Cheirogaleidae
20 Microcebus murinus** C, O E
21 Microcebus griseorufus O E
22 Cheirogaleus medius O E V
23 Mirza coquereli O, T E V
Family: Lemuridae
24 Lemur catta O, T E V
Family: Indriidae
25 Propithecus verreauxi
verreauxi
Family: Suidae
26 Potamochoerus larvatus
larvatus
IUCN
Status
O, T E V
T N
KEY
Methods: C= captured, O= observed, T= tracks, calls, dung.
Status: E= endemic to Madagascar, N = Non-endemic. V = vulnerable, En = endangered.
* Colony roosting in building in village. ** Identified from the colouration of observed individuals.
89

No.
Appendix 7
List of bird species found in the proposed Parc Regional de Belomotse
SPECIES
LOCATION
Antafoky
Lake
Sept Lacs
region
Gallery
forest
HABITAT
Spiny
forest
Disturbed
ENDEMIC
STATUS
IUCN
STATUS
1 Nycticorax nycticorax
N
2 Ardeola idae RE V
3 Bubulcus ibis N
4 Butorides striatus
N
5 Egretta dimorpha
N
6 Egretta alba
N
7 Ardea purpurea
N
8 Ardea cinerea N
9 Ardea humbloti E V
10 Scopus umbretta
N
11 Dendrocygna viduata
N
12 Sarkidiornis melanotos
N
13 Aviceda Madagascariensis
E
14 Milvus aegyptius N
15 Polyboroides radiatus E
16 Accipiter madagascariensis
E NT
17 Accipiter francesii
RE
18 Buteo brachypterus
E
19 Falco newtoni
RE
20 Falco concolor N
21 Falco peregrinus N
E
22 Margaroperdix
madagascariensis
23 Numida meleagris
N
24 Turnix nigricollis E
25 Dryolimmas cuvieri RE
26 Gallinula chloropus
N
27 Actitis hypoleucos
N
28 Sterna caspia
N
29 Pterocles personatus
E
30 Sreptopelia picturata RE
31 Oena capensis N
32 Treron australis
RE
33 Coracopsis vasa
RE
34 Coracopsis nigra
RE
35 Agapornis canus
E
36 Cuculus rochii EB
37 Coua gigas E
38 Coua coquereli
E
39 Coua cursor
E
40 Coua ruficeps
E
41 Coua cristata
E
90

Appendix 7. Continued.
42 Centropus toulou RE
43 Tyto alba N
44 Otus rutilus
RE
45 Ninox superciliaris
E
46 Asio madagascariensis
E
RE
47 Caprimulgus
madagascariensis
48 Cypsiurus parvus
N
49 Alcedo vintsioides
RE
50 Merops supersciliosus
N
51 Eurystomus glaucurus
N
52 Leptosomus discolor N
53 Upupa marginata E
54 Phedina borbonica
RE
55 Motacilla flaviventris
E
56 Coracina cinerea
RE
57 Hypsipetes madagascarensis
RE
58 Vanga curvirostris E
59 Xenopirostris xenopirostris E
60 Falculea palliata
E
61 Artamella viridis
E
62 Leptopterus chabert
E
63 Saxicola torquata
N
64 Copsychus albospecularis E
65 Thammornis chloropetoides E
66 Newtonia brunneicauda
E
67 Newtonia archboldi
LE
68 Cisticola cherina
RE
69 Neomixis tenella
E
70 Neomixis striatigula E
71 Terpsiphone mutata RE
72 Nectarinia souimanga
RE
73 Zosterops maderaspatana
RE
74 Lonchura nana
E
75 Ploceus sakalava
E
76 Foudia madagascariensis E
77 Acridotheres tristis N
78 Dicrurus forficatus
RE
79 Corvus albus
N
73 58 60 50 58
KEY:
Status: E= endemic to Madagascar, EB – Endemic Breeder, RE – Regional Endemic, LE – Locally endemic, N = Non-endemic,
V = vulnerable, NT = Near Threatened.
91

Appendix 8.
Bird species that are different from all other positive identifications but that could not
be confirmed to species level.
LOCATION
HABITAT
Common
No. SPECIES
Antafoky Sept Lacs Gallery Spiny Disturbed name
region forest forest
1 Phoenicopterus sp Flamingo
2 Anas sp Teal
3 Charadrius sp Plover
4 Sterna Tern
5 Anous stolidus Brown
noddy
6 Zoonavena grandidieri Madagascar
spinetail
7 Apus barbatus African
black swift
8 Monticola sp Rock thrush
9 Tylas eduardi Tylas vanga
10 Nestillas Warbler
11 Hartlaubius auratus Madagascar
starling
92

Appendix 9.
List of reptile species found in the proposed Parc Regional de Belomotse.
No. SPECIES
Antafoky
Lake area
Family: Testudinidae
LOCATION
Sept Lacs
region
Riparian
forest
HABITAT
Spiny forest
ENDEMIC
STATUS
IUCN &
CITES
STATUS
1
Geochelone radiata
RE V,
CITES I
2 Pelusios castanoides N
3 Pelomedusa subrufa N CITES III
Family: Boidae
4
Acrantophis dumerili
E V,
CITES I
5
Sanzinia madagascariensis
E V,
CITES I
Family: Typhlopidae
6 Typhlops sp 1 RE
7 Typhlops decorsei RE
Family: Colubridae
8 Dromicodryas bernieri E
9 Heteroliodon occipitalis RE
10 Ithycyphus oursi E
11 Langaha madagascariensis E
12 Leioheterodon geayi RE
13 Leioheterodon madagascariensis E
14 Leioheterodon modestus E
15 Liophidium torquatum E
16 Liophidium trilineatum RE
17 Liophidium vaillanti E
18 Liopholidophis lateralis E
19 Lycodryas gaimardi E
20 Madagascarophis colubrinus E
21 Madagascarophis meridionalis RE
22 Madagascarophis ocellatus RE
23 Mimophis mahfalensis E
24 Pseudoxyrhopus quinquelineatus E
Family: Crocodilia
25 Crocodylus niloticus N CITES I
Family: Gekkonidae
26 Blaesodactylus sakalava RE
27 Geckolepis maculata E
28 Geckolepis petiti E
29 Hemidactylus mabouia N
30 Hemidactylus mercatorius N
31 Lygodactylus tolampyae E
32 Lygodactylus tuberosus E
93

LOCATION
No. SPECIES Antafoky GELOSE
protected
area
Riparian
forest
HABITAT
Spiny forest
ENDEMIC
STATUS
IUCN &
CITES
STATUS
33 Paroedura bastardi E
34 Paroedura pictus RE
35 Paroedura vahiny RE
36 Phelsuma mutabilis E CITES II
37
Phelsuma standingi
RE V,
CITES II
Family: Gerrhosauridae
38 Tracheloptychus madagascariensis E
39 Zonosaurus karsteni RE
40 Zonosaurus laticaudatus E
Family: Scincidae
41 Amphiglossus igneocaudatus E
42 Amphiglossus ornaticeps E
43 Androngo trivittatus RE
44 Mabuya dumasi RE
45 Mabuya elegans E
46 Mabuya gravenhorstii E
47 Mabuya vato RE
48 Voeltzkowia fierinensis RE
Family: Iguanidae
49 Chalaradon madagascariensis E
Family: Opluridae
50 Oplurus cyclurus RE
51 Oplurus quadrimaculatus E
Family: Chameleonidae
52 Brookesia brygooi E CITES II
53 Furcifer lateralis E CITES II
54 Furcifer oustaleti E CITES II
55 Furcifer verrucosus E CITES II
TOTAL 48 48
KEY TO ABBREVIATIONS
Endemic status:
• RE - Regional Endemic: Endemic species with limited distribution, occurring only in southwest Madagascar.
• E - Endemic: Occurring only in Madagascar and associated islands (Comoros, Reunion, Mauritius).
• N – Non-endemic.
IUCN status:
• EN - Endangered
• V - Vulnerable
CITES status
• I – Export and Import authorisations are necessary
• II – Export authorisations are necessary
• III – Exportation must include documentation stating the country of origin of the specimen
94

Appendix 10.
List of butterfly species found in the proposed Parc Regional de Belomotse.
No. SPECIES Riparian forest Spiny forest
Family: Hesperiidae
1 Coeliades ernesti
2 Coeliades rama
3 Coeliades ramanatek ramanatek
4 Tagiades insularis insularis
Family: Lycaenidae
5 Chilades miniscula
6 Cupidopsis jobates jobates
7 Deudorix antalus
8 Hemiolaus cobaltina
9 Leptomyrina phidias
10 Leptotes pirithous
11 Zizula hylax
Family: Nymphalidae
12 Acraea dammii
13 Acraea encedon encedon
14 Acraea igati
15 Acraea lia
16 Acraea obeira obeira
17 Acraea ranavalona
18 Acraea turna
19 Acraea zitja
20 Apaturopsis paulianii
21 Aterica rabena
22 Byblia anvatara anvatara
23 Charaxes andara
24 Charaxes antamboulou
25 Charaxes cacuthis
26 Charaxes zoolina betsimisaraka
27 Danaus chrysippus aegyptius
28 Eurytela dryope lineata
29 Heteropsis ankaratra
30 Heteropsis narcissus fraterna
31 Junonia hierta paris
32 Junonia rhadama
33 Libythea tsiandava
34 Melanitis leda
35 Melanitis leda helena
36 Neptidopsis fulgurata fulgurata
37 Neptis kikideli
38 Neptis saclava saclava
95

No. SPECIES Riparian forest Spiny forest
39 Pseudacraea imerina imerina
40 Sallya amazoula
41 Sallya madagascariensis
42 Morphospecies 77
Family: Papilionidae
43 Graphium evombar
44 Papilio demodocus demodocus
45 Papilio epiphorbas epiphorbas
46 Papioio erithonioides
47 Pharmacophagus antenor
Family: Pieridae
48 Belenois creona prorsus
49 Belenois grandidieri
50 Catopsilia thauruma
51 Colotis amata crowleyi
52 Colotis evanthe
53 Colotis mananhari
54 Colotis zoe
55 Dixea charina narena
56 Eurema floricola
57 Eurema floricola floricola
58 Gideona lucasi
59 Leptosia alcesta sylvicola
60 Mylothris phileris
61 Pinacopteryx eriphia mabillei
96

Appendix 11.
Descriptions of Suggested Walks
Walk 1
Begins at campsite 1 and follows the edge of the lake to Antafoky village. The time
needed is approximately 1 hour. A medium level of fitness would be required as there
is sometimes a steep gradient and footing may be difficult on the narrow pathway.
Visitors can expect to see a variety of wildlife - in particular the bird life is very
prevalent and obvious, especially water birds (cattle and great egrets, Madagascan
pond, purple, green-backed and night crowned herons), which are abundant. The walk
provides a wonderful view of the spectacular limestone cliffs that frame the lake. At
Antafoky village it is possible to rest and have simple refreshments under the mango
trees overlooking the Onilahy River.
Return is via the same route or boat.
Possible hazards encountered in this walk are narrow paths and loose stones that may
require careful footing. The path also passes through several shallows around the lake
and tourists would need to wade up to waist deep in water.
Walk 2
Begins in Antafoky, goes up the steep hill behind the village and then descends down
into the valley to join the road back to Antafoky. The time needed is approximately
five hours. This walk is fairly arduous and requires a high level of fitness due to the
steepness.
Reptiles are abundant and radiated tortoises have been seen here. This walk also
encompasses a wide range of vegetation types - gallery forest, transitional forest,
thorny scrub and some cultivated areas. A spectacular dried limestone waterfall, dried
up riverbed, caves and huge cliffs can be seen. The walk includes the village of
Antafoky and a tomb on the hillside which visitors may find interesting.
A pirogue can be taken back to the camp or Walk 1 will take walkers back.
The hike is fairly arduous and loose rock and thorny scrub may cause minor injuries
such as scratches. Plenty of water, good walking boots, long trousers and hat and
suncream are recommended. There are plenty of picnic sites along the route - the best
being the limestone waterfall due to the spectacular viewpoint it provides. Certain
areas of the path need to be cleared.
Walk 3
This walk starts at campsite 1 and travels up the valley before going up a small hill
and then curving around and returning to camp. The estimated required time is two
hours. The level of fitness required is fairly low to medium.
97

Many species of birds can be seen, as well as lemurs on the cliff face and many reptile
species. The landscape features include a fossilised coral reef, sandstone cliffs and a
limestone plateau.
The path leads through a cultivated area where there is a lack of shade and the path
needs to be cleared of rocks and vegetation. This area could be the focus of a
replanting scheme.
Walk 4
This is a stunning walk but fairly strenuous. The route begins at campsite 1 and goes
around the side of some impressive sandstone cliffs where a spectacular view across
the lake can be seen. From here, the path ascends up a fairly steep path that levels out
on the plateau to give a magnificent view of the surrounding valleys. The sandstone
cliffs, a dried-up waterfall and rock amphitheatre are just some of the geological
points of interest to note as you pass along the riverbed towards the lake. This walk
takes 8 hours and requires a moderate to high level of fitness.
Several areas of the path need to be cleared, but this would only need minor work, and
possibly benches could be placed at strategic places around the walk for bird watching
or just resting and admiring the viewpoints.
Walk 5
This walk begins at campsite 2 at the Sept Lacs and goes up the valley taking in all
seven lakes. At the final pool, a track leads to a zebu-cart path that can be followed
back down to the main Antafoky – Ifanato road, and hence back to camp.
This walk would take around an hour and a half, but most people would want to stop
and swim in at least one of the pools. There are numerous places to do this and there
are many ideal picnic spots.
The walk could be done by anyone with a moderate level of fitness. The path passes
in and out of the river and may need some careful clearance and management.
Walk 6
This walk begins at campsite 2 and can either begin by following walk 5 and turning
right when the path meets up with the zebu track, or it can follow the zebu track
directly from the road.
The walk winds up through the dry forest, affording spectacular views over the
Onilahy River valley, and then passes down through the Sept Lacs River until the
valley opens out again into a patch of gallery forest. A track then continues up to the
left out of the gallery forest into spiny forest, passing a sinkhole that contains a
sizeable colony of bats.
This walk would take around 4 hours. A medium level of fitness would be necessary
for the majority of the way, and it passes through the river for a short distance.
98

Ring-tailed lemurs and Verreaux’s sifaka are abundant, as are many bird and reptile
species. The bat colony in the sinkhole comprises approximately 40,000 individuals of
three different species. The bats begin to leave their roosts around dusk (about 6 p.m).
The first hour is the most spectacular and would be of great interest to many tourists.
Walk 7
This walk starts at campsite 2 and follows walk 6 until the path reaches the area of
gallery forest. From here, follow a small path that runs parallel to the river, taking in a
further three lakes, including an extremely deep pool that is perfect for swimming and
high diving. The path and river continue for over five kilometres.
The time needed is approximately 5 hours. A medium level of fitness is required,
although the path is fairly level.
Birds are extremely abundant in the upper sections of the river, particularly birs of
prey. Ring-tailed lemurs and Verreaux’s sifaka can be seen along the route.
The path needs to be cleared and clearly signposted in places.
99

Appendix 12.
Descriptions of proposed bird hides
Hide 1
Surroundings – Located at Antafoky Lake in water hyacinth raft area with views
across lake to opposite hillsides.
Features – Good potential for crocodile watching.
Disturbance – General lake activity (fishing and transportation by pirogue).
Visibility – Excellent unobstructed views across lake and water hyacinth rafts.
Special hide requirements – approach to hide to be constructed of floating wooden
boards secluded by existing aquatic vegetation growth. Hide could be floating to
manage changes in lake levels.
Suggested combined activities - Close proximity to walk 4 to valley with natural
amphitheatre, plunge pool and cliff nesting sites for raptors and other bird species.
Birds – Potentially high variety on lake - gallery forest transition. Grey Heron and
Great Egret regularly seen on water hyacinth rafts.
Means of Access – By foot using walk 1 from suggested camp via Antafoky village.
Hide to be accessed via floating boarded walk from lake side path.
Time from proposed camp – 1hr30min by foot.
Hide 2
Surroundings – Located at Antafoky Lake – gallery forest transition with view across
water hyacinth raft area and surrounding hillsides.
Features – Good potential for crocodile watching.
Disturbance – General lake activity (fishing and transportation by pirogue) and lake
path disturbance.
Visibility – Excellent unobstructed views across lake and water hyacinth rafts.
Suggested combined activities - Close proximity to walk 4 to valley with natural
amphitheatre, plunge pool and cliff nesting sites for raptors and other bird species.
Birds – Potentially high variety on lake-gallery forest transition. Grey Heron and
Great Egret regularly seen on water hyacinth rafts.
Means of Access – By foot using walk 1 from suggested camp via Antafoky village.
Time from proposed camp – 1hr30min by foot
100

Hide 3
Surroundings – Located in gallery forest commanding views of forest canopy and
lakeside near proposed campsite 1.
Features – Good potential for lemur watching.
Special bird features – Canopy views afford good views of Vasa Parrots and other
canopy bird species.
Disturbance – General lake activity (fishing and transportation by pirogue).
Visibility – Good view of canopy and lake.
Suggested combined activities - Walk 4 to sandstone cliff and further walk to valley
with natural amphitheatre, plunge pool and cliff nesting sites for raptors and other bird
species.
Birds – Potentially high variety on lake-gallery forest transition. Cliff location –
potential of seeing Rock Thrushes.
Means of Access – By foot using walk 1 from suggested camp via Antafoky village.
Time from proposed camp – 1hr30min by foot.
Hide 4
Surroundings – Located at Antafoky Lake – gallery forest transition at lake bend with
view of opposite scrub hillside and raft of lake water hyacinths.
Features – Good potential for watching crocodiles and lemurs in forest opposite.
Disturbance – General lake activity (fishing and transportation by pirogue). Path to
hide currently not in regular use as a main village footpath.
Visibility – Some overhanging vegetation and small bushes at the lakeside obscure the
view from the hide.
Suggested combined activities - Close proximity to walk 4 to valley with natural
amphitheatre, plunge pool and cliff nesting sites for raptors and other bird species.
Birds – Potentially high variety on lake-gallery forest transition. Grey Heron and
Great Egret regularly seen on water hyacinth rafts.
Means of Access – By foot using walk 1 from suggested camp via Antafoky village.
Time from proposed camp – 1hr50min by foot.
Difficulty – Path to bird hide fairly difficult due to boulders and narrow path.
101

Hide 5
Surroundings – Located at Antafoky Lake – gallery forest transition at lake bend with
view of opposite scrub hillside and raft of lake water hyacinths.
Features – Good potential for watching crocodiles and lemurs in forest opposite.
Disturbance – General lake activity (fishing and transportation by pirogue). Path to
hide currently not in regular use as a main village footpath.
Visibility – Some overhanging vegetation and small bushes at the lakeside obscure the
view from the hide.
Suggested combined activities - Close proximity to walk 4 to valley with natural
amphitheatre, plunge pool and cliff nesting sites for raptors and other bird species.
Birds – Potentially high variety on lake-gallery forest transition. Grey Heron and
Great Egret regularly seen on water hyacinth rafts.
Means of Access – By foot using walk 1 from suggested camp via Antafoky village.
Time from proposed camp – 1hr50min by foot.
Difficulty – Path to bird hide fairly difficult due to boulders and narrow path.
Accommodation in hides:
It is envisaged that small groups (5 people maximum) with guide will visit hides at
any one time. Each hide should adequately accommodate at least 2-3 persons at any
given time. Seating (e.g. wooden bench) is required in addition to a shelf below
viewing slits for standing telescopes and other equipment. Hides should be
waterproof.
Hide materials, construction and design:
Hides should be constructed with minimal impact to the local environment. It is
suggested that all hides are built from wood incorporating simple design features
including a concealed entrance door and viewing slits on three hide sides. Use of local
labour and skills in both construction and maintenance is desirable.
Birdwatching information :
Tourists will require specific information on what bird species can be seen in the area
and may require general information such as bird field guides to Madagascar. It is
suggested that bird species-lists and a map of local bird features are made available
for tourists. Local knowledge of the guide will complement these sources of
information.
Birdwatching equipment :
Visitors will have to bring their own equipment, such as binoculars, guide-books,
torches (for night walks), tripods and telescopes.
102

Appendix 13.
Specimen identification and deposition
For identification purposes, selected specimens are sent to the following taxonomists
and institutions:
Botany:
Pierre Jules Rataromalaza WWF Toliara WWF Toliara, B.P. 220
Toliara (1), Madagascar
Zoology:
Small mammals:
Steven Goodman Field Museum of Natural History Field Museum of Natural
History, 1400 Roosevelt Road,
Chicago, Illinois 60605, USA
Daniel Rakotondravony Dept of Animal Biology University of Antananarivo,
B.P. 906, Antananarivo,
Madagascar
Reptiles and Amphibians:
Chris J. Raxworthy US Museum of Natural History American Museum of Natural
History, New York, USA
Achille Raselimanana WWF Antananarivo Biodiversity Officer, WWF
Antananarivo
Butterflies:
David Lees and Natural History Museum Dept of Emtomology, The
Alison Cameron
Natural History Museum,
Cromwell Road, London,
SW7 5BD
Ants and other invertebrates:
Brian Fisher Dept of Entomology California Academy of
Sciences, Golden gate
Park, San Francisco,
California 94118, USA
103