Applied Ethnobotany: People, Wild Plant Use and Conservation

Applied Ethnobotany

Applied Ethnobotany 20/11 04/12/2000 05:27 pm Page i

PEOPLE AND PLANTS CONSERVATION MANUALS

Manual Series EditorMartin Walters

Manual Series OriginatorAlan Hamilton

People and Plants is a joint initiative of WWF, the United Nations Educational, Scientific and Cultural Organization (UNESCO)

and the Royal Botanic Gardens, Kew.

Forthcoming titles in the series

Biodiversity and Traditional Knowledge: Equitable Partnerships in PracticeSarah A Laird (ed)

Ethnobotany: A Methods Manual 2nd editionGary J Martin

The Management and Marketing of Non-Timber Forest Products: Certification as a Tool to Promote Sustainability

Patricia Shanley, Sarah A Laird, Alan Pierce and Abraham Guillén (eds)

People, Plants and Protected Areas: A Guide to In Situ Management (reissue)John Tuxill and Gary Paul Nabhan

Plant Invaders: The Threat to Natural Ecosystems (reissue)Quentin C B Cronk and Janice L Fuller

Uncovering the Hidden Harvest: Valuation Methods for Woodland and Forest Resources

Bruce M Campbell and Martin K Luckert (eds)

Applied Ethnobotany 20/11 04/12/2000 05:27 pm Page ii

Applied Ethnobotany

People, Wild Plant Use and Conservation

Anthony B Cunningham

Earthscan Publications Ltd, London and Sterling, VA

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page iii

First published in the UK and USA in 2001 byEarthscan Publications Ltd

Copyright © WWF, 2001

All rights reserved

A catalogue record for this book is available from the British Library

ISBN: 1 85383 697 4

Typesetting by PCS Mapping & DTP, Newcastle upon TynePrinted and bound in the UK by Redwood Books Ltd, Trowbridge, WilshireCover design by Yvonne BoothCover photo by A B CunninghamPanda symbol © 1986 WWF® WWF registered trademark owner

For a full list of publications please contact:Earthscan Publications Ltd120 Pentonville RoadLondon, N1 9JN, UKTel: +44 (0)20 7278 0433Fax: +44 (0)20 7278 1142Email: [email protected]://www.earthscan.co.uk

22883 Quicksilver Drive, Sterling, VA 20166–2012, USA

Earthscan is an editorially independent subsidiary of Kogan Page Ltd and publishes in associationwith WWF-UK and the International Institute for Environment and Development

This book is printed on elemental chlorine-free paper

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page iv

Contents

List of Figures, Tables and Boxes viiThe People and Plants Initiative by Alan Hamilton xiiPreface xiiiIntroduction xvPeople and Plants partners xixAcknowledgements xx

1 Conservation and context: different times, different views 1Introduction 1Historical context 3Management myths and effective partnerships 5Vegetation and change: spatial and time scales 7Human influence: landscapes and species 8

2 Local inventories, values and quantities of harvested resources 10Introduction 10Local priorities: vegetation types, resource categories and species 10Choosing the right methods 12Before starting: attitudes, time spans and cross-checking 15Taxonomy with all your senses: the use of field characters 32Potentials and pitfalls: combining skills in inventories 44Local to international units 51

3 Settlement, commercialization and change 60Introduction 60Local markets: order within ‘chaos’ 63Location and mapping of marketplaces 64Characteristics of markets 73Market schedules 78Marketing chains and types of seller 82Inventory and frequency of plants on sale 87

4 Measuring individual plants and assessing harvesting impacts 96Introduction 96Necessary equipment 97Measuring diameter, height and bark thickness 97Methods for ageing plants 115Harvesting impacts 126

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page v

5 Opportunities and constraints on sustainable harvest: plant populations 144Introduction 144Plant populations and practical constraints: selecting species 145Bridging gaps in knowledge: life forms, plant architecture and reproductive strategies 150

Plant life forms 150Costs and complexity: inventory, management and monitoring 156Yields: supply versus demand 180Population modelling using transition matrices 184

6 Landscapes and ecosystems: patterns, processes and plant use 192Introduction 192Tools for the ‘big picture’: aerial photographs and satellite images 196Distribution, degree of threat and disturbance 202Local knowledge, landscapes and mapping 212

7 Conservation behaviour, boundaries and beliefs 222Introduction 222Conservation and the ingredients for common property management 223Ecological factors, land use, tenure and territoriality 233Property rights: land and resource tenure 238Boundaries and tenure, meaning and mapping 245Ritual, religion and resource control 253Who are the stakeholders? 259

8 Striving for balance: looking outward and inward 264Introduction 264Looking outward 267Looking inward; examining innovative local approaches 269

Acronyms and abbreviations 272Further reading 274References 278Index 295

Applied Ethnobotany

vi

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page vi

List of Figures, Tables and Boxes

Figures

I.1 A patrol ranger in Mgahinga Gorilla National Park xviiI.2 A schematic continuum from wild to domesticated plant species, from

foraging of wild species to farming domesticated species xviii1.1 Distribution of the sands of the African coast and Kalahari basin 21.2 Achieving sustainable use of resources within a political and policy framework 61.3 The distribution of professional ecologists in relation to the distribution of

plant species richness 72.1 Five basic steps in dealing with plant-resource management issues 132.2 The palm wine trade and variation in volumes sold 172.3 Seasonal fluctuations in edible wild plants by people in the north-western

Kalahari 182.4 Discrepancies between forestry annual report data and weigh-bridge returns

for Prunus africana bark harvests 202.5 Comparison of the proportional species of pole harvests at Hlatikulu Forest

Reserve 212.6 Resource and population trends from ‘stick graphs’ 242.7 Reported frequency of consumption by 211 adults of 47 edible wild greens 272.8 Examples of bark characteristics 352.9 Local harvester assessments of bamboo utility in Bwindi-Impenetrable

National Park, Uganda 512.10 Local units convertible to international units of mass or volume 522.11 Commonly used local estimates of circumference for different-sized bundles

of plant products 532.12 Steps in the processing of an edible fruit, showing the low recovery rate 552.13 The selection and processing of Smilax anceps for basket making 583.1 Road construction through tropical forest in Côte d’Ivoire 623.2 The involvement of local people in palm-wine sales 653.3 Map of markets along the River Congo (Zaire River) in Central Africa

made by anthropologist Yuji Ankei 693.4 Sequence of a ‘node’ becoming what geographers call a ‘nucleated’

settlement, and developing into a town and then a city 713.5 A seller of hardwood poles and a woman harvester-seller of edible wild

spinach 723.6 An interview survey of chewing stick use in Ghana 733.7 (a) Human population density by local authority. (b) Location of markets

in the same area showing how most markets are located in high population density areas 75

3.8 Location of sellers and interchange of goods at a small market in Chesegon, Kenya 77

3.9 Seven-day market schedules are common in West Africa 79

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page vii

3.10 The sequential development from periodic (P) to daily (D) markets as paths to the marketplace develop into tracks and then into roads 80

3.11 Changes in the importance of some plant species sold in Mercado de Sonora, Mexico City 81

3.12 Patterns of market visits, with African medicinal plant sales as an example 833.13 A marketing chain in the sale of plates made from sal (Shorea robusta)

leaves harvested from woodland in West Bengal, India 853.14 Genetic variation in Dacryodes edulis fruits from a survey of market stalls in

Yaounde, Cameroon 934.1 Standard measurements of diameter at breast height (dbh) or of girth at

breast height (gbh) in trees that are leaning or strangely shaped 994.2 Use of a clinometer to measure tree height 1024.3 Two methods for measuring tree height 1054.4 A Swedish bark gauge being used to measure bark thickness 1104.5 The relationship between Rytigynia kigeziensis diameter at breast height

(dbh) and bark mass available from tree stem (up to 2m) 1114.6 With complex palisade fences, traditional Owambo housing uses a

spectacular amount of wood 1124.7 A comparison of the stem basal area/weight regression for ten southern

African savanna tree species 1144.8 Diagrammatic respresentation of three tree stems, with growth rings

matched to earlier periods 1164.9 The stem of the nikau palm in coastal forest, New Zealand 1194.10 Surface features of some grass trees and tree ferns 1204.11 The relationship of the number of leaf scars and the height of Chamaedorea

tepejilote palm stems 1224.12 Ageing methods for bulbs and corms 1254.13 Methods used for measuring growth and life span 1274.14 Crown ratings used to measure trees’ condition and general health 1304.15 Methods of latex tapping from the rubber tree 1314.16 Measurement of leaflet length in Hyphaene palms 1344.17 A seven-point scale used for bark damage ratings. The photograph shows a

harvester removing medicinal bark from an Afromontane forest tree 1364.18 The resilience or vulnerability of trees after bark removal 1384.19 Stunted Euclea divinorum shrubs after ten years of root removal for dyeing

basketing 1404.20 A seven-point scale for rating of root harvesting damage 1415.1 Acacia trees after bark removal, showing irrecoverable damage 1455.2 Comparison of density and number of Acacia trees, both outside and inside

a protected area 1465.3 The effects of harvesting on a plant population 1475.4 Musanga and Cecropia, both with a short leaf life-spans; Podocarpus has

a long leaf life-span 1555.5 (a) Specific leaf area in relation to leaf life span. (b) Relative growth rate

of young plants for a range of species from different ecosystems 1575.6 A flow chart showing an adaptive management approach to the sustainable

harvest of non-timber resources from tropical moist forest 1595.7 Four methods of sampling in an area with three vegetation types 1635.8 Possible arrangement of tiered subplots for measuring below 5cm and 20cm

diameter at breast height (dbh) thresholds on a 1ha plot 1665.9 Mopane trees outside a village, and the change in density of mopane

coppice stems in relation to distance from village 168

Applied Ethnobotany

viii

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page viii

5.10 The increase in reproductive output, age and depth of corms of the West Australian geophyte Philydrella pygmaea (Philydraceae) 170

5.11 A matched pair of photographs showing tree population decline between 1946 and 1996 172

5.12 Generalized models showing how ‘typical’ tree diameter size-class distributions can indicate the state of a species population 173

5.13 The debarking and demise, between 1982 and 1992, of an entire population of Berchemia discolor trees 174

5.14 A hypothetical example of long-term change in the population structure experienced by a tropical forest tree species 175

5.15 The palm-pilot field computer used by trackers in the Karoo National Park, South Africa 178

5.16 (a) Edible fruit yield related to tree diameter size. (b) Deadwood yield for fuel related to tree standing biomass in semi-arid savanna, southern Africa 181

5.17 A life-stage graph for mukwa, a southern African savanna tree species 1895.18 Mean growth rate and extinction probability over 100 years of an American

ginseng population 1906.1 Levels of detail will vary with spatial scale in a hierarchy from a global or

continental level to that of species levels and genetic levels within a population 193

6.2 The proposed Central American Biological Corridor 1956.3 Landsat image showing differences in vegetation cover between southern

Angola and northern Namibia 1976.4 The deforestation history of Eastern Madagascar, derived from aerial

photographs and satellite images 1986.5 Afromontane forest within a prtoected area that provides habitat for half the

world’s mountain gorilla population 1996.6 Two different distributions of two medicinal species in high demand 2046.7 Fire frequency and its effect on the survival prospects of Haemanthus

pubescens 2106.8 A conceptual model of the dynamics of a subtropical lowland forest in

southern Africa 2116.9 Land form units traversed by women on foraging excursions from two

different settlements in Western Australia 2146.10 (a) Species richness. (b) Diversity indices of food plants on major land-form

units in the Great Sandy Desert, Western Australia 2156.11 Diagrammatic representation of stages in succession of bamboo in East

African montane forest 2167.1 Much of the success or failure of local participation in conservation

programmes hinges on the social factors of relations, rights and responsibilities 225

7.2 Ecological impacts and land-use conflicts caused by the encroachment of wheat farmers on savanna and grasslands in Kenya 230

7.3 Conceptual models and variation in access nights amongst hunter-gatherers and ‘traditional’ pastoralists 235

7.4 A matrix of types of conflict over different natural resources at various institutional levels developed during a rapid rural appraisal of resource conflicts 237

7.5 A diagram drawn by Jinga villagers showing the location of sites and restrictions on land and resource use 238

7.6 Patch-burning of spinifex grasslands in Australia and a painting of spinifex landscape, rich in symbolic meaning 246


ix

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page ix

7.7 ‘Ritual topography’ at different spatial scales influences tenure and access to different people 249

7.8 Aerial photograph analysis, combined with ‘ground-truthing’ with local people helps gain insight to tenure, boundaries and local institutions 250

7.9 The frequency of reaffirming key points in the landscape is an important measure of the social significance of boundaries 255

7.10 Diagrammatic representation of the mediating role of an African ritual specialist 257

7.11 Mean percentage of people from communities adjacent to Bwindi-Impenetrable National Park (n = 978) who were involved in collecting forest products or pit-sawing, or who were affected by crop-raiding animals prior to park closure compared to those further away (n = 1405) 260

8.1 Fifteen basic steps towards resource management 2658.2 The historical cycle of forest product extraction, with examples from

Amazonia and Africa 2668.3 Consumption pressure: a measure of the burden placed on the environment

by people, 1995 2688.4 (a) World population, projected to 2150. (b) Population by region, 1995

and projection for 2050 under a medium-fertility scenario 2698.5 Encouraging conservation through drama, in Kenya 270

Tables

2.1 A comparison of fuelwood preference or avoidance using three different methods 22

2.2 Total annual wood consumption (tonnes per household per year) 542.3 Basket makers’ assessments of Hyphaene petersiana palm leaves rejected

or considered acceptable for basketry 543.1 A summary of methods used at different levels of detail in the study of

exchange and distribution 613.2 Market types in north-eastern Ghana and Guatemala classified on Skinner’s

(1964) hierarchical system 743.3 Distribution of marketplace levels by marketplace types in Guatemala 763.4 The top 10 medicinal plants sold in the markets in Ethiopia, showing number

of sellers in 3 of the 15 markets sampled, including the total sellers/species for all markets 87

3.5 Rabinowitz’s seven forms of rarity 905.1 The basic ecological characteristics of early pioneer, late secondary and

primary tropical forest species 1565.2 A portion of a random number table 1645.3 Slope corrections for different distances on slopes of varying steepness 1656.1 The overall percentage of land surface in five African regions under all

forms of conservation recognized by the IUCN, and the percentage of these conserved areas that are designated as national parks 194

6.2 The matrix for integrating biological distinctiveness and conservation status of ecoregions to assign priorities for biodiversity conservation 209

7.1 Opportunities for community-based resource management vary with rainfall, soils and land use, socio-economic factors and the composition of local communities 234

Applied Ethnobotany

x

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page x

7.2 A schematic way in which Nguni structure the world: a ‘traditional’ view which has resonance in several other agricultural societies, such as Aouan farmers in the Côte d’Ivoire and Bakiga farmers who have felled (‘domesticated’) much of Bwindi (meaning ‘dark’) Forest in western Uganda 256

7.3 Criteria for identifying the most significant stakeholders in sustainable forest management: an example from East Kalimantan, Indonesia 261

Boxes

1.1 IUCN protected area categories: the modified system of protected area categories agreed at the IV World Congress on National Parks and Protected Areas, 1992 4

2.1 Steps in questionnaire design and implementation 292.2 Collecting plant specimens: five important reminders 312.3 Plant exudates: standardizing botanical descriptions 382.4 Bulbs and corms 422.5 Multiple names, single species 462.6 Mismatch: Linnaean names to folk taxonomy and vice versa 472.7 Standardized terms for describing bark components, bark texture, patterns

and exudates 563.1 Checklist: ethnobotanical surveys of marketplaces 663.2 Ethnobotanical surveys of markets 894.1 Lengths of ‘awkward customers’: climbing palms and tilting trees 1035.1 Predictors of resilience or vulnerability to harvesting based on geographic

distribution, habitat specificity, local population size, growth rates, part ofthe plant used, variety of uses and reproductive biology 148

5.2 Although harvesting of different plant parts can be grouped into lower-impact (leaves, flowers, fruits) and higher-impact uses (bark, roots, stems, whole plant), each of these can be subdivided according to the biology of the plant species concerned 149

5.3 Characteristics across a continuum: long-lived reseeders versus resprouters 1526.1 Aerial photographs for vegetation interpretation 2006.2 The IUCN Red List categories 2076.3 Participatory mapping exercises: resources at landscape and species

population levels 2187.1 Ingredients for successful community-based natural resource management

(CBNRM) programmes 2277.2 Access rights, environment and cultural practice 2327.3 Four types of property rights 2407.4 Eleven characteristics of tenure systems 2417.5 Mapping methods: potential and pitfalls 251


xi

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xi

The People and Plants Initiative

Conservation is directly linked to people’s values and behaviour. It is therefore ironic thatthe people–conservation interface has been neglected in the past. Part of this neglect hasbeen due to a lack of appreciation of the roles that the knowledge, institutions andcultural perspectives of local people can play in resource management and conservation.To see conservation areas or natural resources through the eyes of resource users is aninstructive and important process for any conservation biologist or national parkmanager. Research in ethnobiology (of which ethnoecology and ethnobotany are parts) isa useful element in this process. Ethnobotany stands at the interface of several disci-plines, including anthropology, botany, ecology, geography, economics and others. Towork in ethnobotany applied to conservation or rural development may therefore seem adaunting intellectual challenge. Progress is greatly dependent on the ability to recognizepriorities – an ability which the Applied Ethnobotany manual is designed to promote.

People and Plants is an initiative of WWF, the United Nations Educational, Scientificand Cultural Organization (UNESCO) and the Royal Botanic Gardens, Kew. It aims toincrease the capacity for community-based plant conservation worldwide. Training isundertaken at field sites in selected countries, with case studies and other informationmade available to a wide audience through various publications, training videos and anInternet service. Publications include working papers, issues of a handbook and discus-sion papers, in addition to the People and Plants conservation manuals series, to whichthe present work contributes.

The People and Plants website can be visited at http://rbgkew.org.uk/peopleplants. Itcontains full versions of several of the smaller People and Plants publications, and contactinformation for organizations involved in applied ethnobotany.

Alan HamiltonHead, International Plants Conservation Unit

WWF-UK

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xii

Preface

Ironically, although conservation is directly linked to people’s values and behaviour, inlandscapes changed by people, the people–conservation interface is often neglected. Partof this neglect results from a lack of appreciation of the role that local people’s knowl-edge, institutions and cultural perspectives can play in resource management andconservation. To see conservation areas or natural resources through the eyes of resourceusers is an instructive and important process for any conservation biologist or nationalpark manager. Research at the interface between several disciplines, using methodsderived from anthropology, geography, economics and ecology – in what is becoming the‘new’ discipline of ethnobiology, of which ethnoecology and ethnobotany are part – is auseful part of this process.

Confusion or clarity in conservation practice?

Tropical ecosystems are the most diverse on earth, yet they have been poorly studied byscientists. Although these habitats have received increasing attention in the past decade,many species in the tropics remain undescribed. Even less is known about the biomassproduction of most tropical species, or about the ways in which species interact, so thatthe ecological impact of the loss of species through over-harvesting is difficult or evenimpossible to predict. Under these circumstances, it is no wonder that field researchersand national park managers ask themselves: how can a policy of sustainable use beimplemented when hundreds of species may be involved?

Many researchers or national park managers, whether expatriates or nationals, havegrown up in an urban environment, and end up in positions where they have to makedecisions on resource-sharing arrangements on the basis of limited theoreticalbackground or field experience. Many, like myself, were trained in university or collegesystems where undergraduate courses start at a micro-level, full of detail (cell biology,taxonomy, physiology), rather than at the macro-level of pattern and process acrosslandscapes. Many university courses have limited linkage between subjects such aszoology, botany or geology that are closely interconnected in the field, and even fewergraduates have training in both the biological and the social sciences.

It is no wonder, therefore, that graduates who end up in conservation areas are oftenconfused by the detail of the hundreds of species and life forms, or by the patchinesswithin and between vegetation types. It is even more perplexing for those working at theinterface between parks and people, trying to straddle social and biological issues in aneffort to resolve land-use conflicts or to set up sustainable harvesting systems. Where dowe start (or stop) in the information-collecting process? Do we collect everything orfocus only on key issues, and if so, what are those issues?

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xiii

The answers to these questions depend on the objectives of the research, on timeconstraints, and on the available money and manpower. In many developing countries,the problems facing conservation areas are urgent, and time, funding and trainedresearchers are scarce. Long-term monitoring therefore has to be limited to key issues,with initial guidelines set through short-term research of less than two to three years.The first step is to ‘make haste slowly’. This should be done by systematically workingthrough the social, economic and ecological components that all influence resourcemanagement and conservation at different spatial and time scales. The chapters thatfollow deal with different steps in this process.

At one stage, during the long process of writing this manual, it crossed my mind thatit would be better to produce a manual on methods which was composed of just oneZen-like sentence: ‘The only method is that there is no method.’ There would have beenmethod in this. In a field as complex as conservation, one cannot hope to produce a‘recipe book’ of methods, applicable to every situation. What is suitable in one case maybe completely unsuitable in another. Some problems are unique to a particular region,posing new challenges for innovative methods that need to be designed in the field. Forthis reason, I cover some general principles, conceptual models and methods as tools tokeep in mind when faced with particular problems, to take out where appropriate, fieldtest, modify and test again. This is far better than blindly transplanting field methodsfrom one place to another. However, any choice of methods needs to be informed bytheory, research design and an understanding of basic concepts. For this reason, the ‘howto do it’ part of methods is given in the context of the practical and theoreticalbackground to those methods.

The advantage of applied ethnobotanical research is that a great deal can be achievedwith simple, inexpensive equipment: pencils, paper, a tape measure, a compass and a ballof string. In common with any science, a healthy dose of scepticism is also an excellentingredient! In the complex world of conservation, it is unlikely that we will get to knowall the answers. It is to be hoped that this manual is a guide to asking the right questions– and also to answering some of them.

This book is dedicated to the students and young professionals from local communi-ties who work so much in isolation, but who are at the forefront of conservation effortand ethnobotanical research.

Applied Ethnobotany

xiv

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xiv

Introduction

This manual is a product of the People and Plants Initiative, a joint programme of WWF,UNESCO and the Royal Botanic Gardens, Kew. It is a companion volume to two othermethods manuals in the People and Plants series: Ethnobotany: a methods manual byGary J Martin (1995; second edition forthcoming) and People, Plants and ProtectedAreas: a guide to in-situ management by John Tuxill and Gary P Nabhan (1998; reissue2001). Gary Martin’s manual, the first in this series, provides practical guidelines forwork in regional floras of ethnobotanical importance and describes the botanical, anthro-pological, phytochemical, linguistic and ecological approaches used to collect informationon useful plants. John Tuxill and Gary Nabhan’s manual focuses on in-situ conservationof crop plant varieties and useful wild plants. Applied Ethnobotany: people, wild plantuse and conservation focuses on practical steps to develop a better understanding of thevalues, vulnerability and resource management options for wild, non-cultivated plantresources. All three manuals stress the essential collaborative nature of ethnobotany,linking scientific and folk knowledge. They also contribute to efforts to build local capac-ity for plants conservation by promoting applied research on biodiversity conservationwhich strengthens connections between biological and social sciences.

Over the past 30 years, conservation efforts have broadened from the earlier empha-sis on increasingly insular, strictly protected areas to a broader approach involving landusers in ‘bioregional’ management at an ecosystem level. This broader approach is evidentin the different World Conservation Union (IUCN) categories of protected areas whichwere developed in the mid 1980s and recently modified at the IV World Congress onNational Parks and Protected Areas (Chapter 1, Box 1.1). It also represents a changefrom one where intervention by the state (by government) through proclamation ofnational parks was seen as the solution, to one where the role of private landowners andresidents of communal lands are recognized. Despite increased awareness of environ-mental concerns and international backing for conservation, many national parks haveinadequate staff or funding to control often large protected areas.

The focus of this manual

This manual focuses on an issue crucial to rural development and conservation: theimpact of harvesting of wild plants by people. It thus covers the borderland betweencultural and biological diversity. It is intended as a practical guide to approaches andfield methods for participatory work between resource users and field researchers. Inparticular, it is aimed at African students or professionals working in conservation, ruraldevelopment or as national park managers who have to make resource managementdecisions.

The emphasis of the manual is on how to identify the most urgent problems, needsand opportunities relating to wild plant use and resource management. It also aims toprovide practical guidelines for research which interface applied ecological approaches

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xv

with the knowledge and expertise of localresource users.

The excellent ecological primerwritten by Charles Peters (1994) providedguidelines for the sustainable harvest ofnon-timber forest products, focusingprimarily on examples from South-EastAsia and Latin America. This manuallooks beyond tropical forests to othervegetation types as well, with many of itsexamples drawn from Africa.

There are three reasons for this. Thefirst is that despite the great importancewild plant use plays in African people’slives, far more attention seems to havebeen given to plant use in Asia or LatinAmerica than in Africa. Secondly, it is thecontinent where I was born and have spentmost of my life. Thirdly, plant use bypeople is an increasingly important issueto take into account at the interfacebetween conservation areas and localcommunities, and African conservationareas are a prime example of this.Understaffed, with limited money andmanpower, and sometimes overrun bywarfare, many conservation areas existvirtually only on paper. Examples over thepast decade alone are national parks inAngola, Chad, Ethiopia, Liberia,Mozambique, Rwanda, Sierra Leone,Somalia, Sudan, Uganda and theDemocratic Republic of Congo (formerlyZaire) (Figure I.1). Projections for the future of prime conservation areas such as forestsare considered to be bleak (Barnes, 1990). It will become even bleaker if planning ofprotected areas does not take local land and plant use into account. To some extent,similar problems are faced in parts of Latin America and Asia, affecting not only thefuture of land set aside for biodiversity conservation, but also the lives of peoplesurrounding national parks and nature reserves. Whether effective answers to thesequestions can be found that can benefit both people and conservation remains to be seen,and careful monitoring of resource sharing and participatory management projects isessential (Kremen et al, 1994).

Why use the term ‘wild’ plants?

Some people are uncomfortable with the term ‘wild’ in the title of this manual, feelingthat it sidesteps issues of indigenous peoples’ intellectual property. What is important isthe context in which ‘wild’ is used. As you read through this manual, you will see that Ihave stressed that there are few landscapes in the world that are not affected by human

Applied Ethnobotany

xvi

Figure I.1 Can protected areas survive underthese circumstances? A patrol ranger in

Mgahinga Gorilla National Park, a Ugandan conservation area in

the Virunga Mountains on the border ofRwanda and the Democratic Republic

of Congo

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xvi

disturbance. Some of these were deliberately burned, manipulating plant productionwith fire as a tool to ‘domesticate’ some landscapes. There is no controversy about this.Few virgin habitats exist on earth, and landscape ‘domestication’ using fire predatesplant species domestication by people by about 200,000 years. At a species level, I usethe term wild to distinguish between wild and domesticated plant species, where domes-ticated plant species are those whose breeding systems have been so changed throughgenetic or phenotypic selection that they have become dependent upon sustained humanassistance for their survival. Wild and domesticated species are at opposite ends of acontinuum (Figure I.2). ‘Wild’ is also a lot shorter than alternative terms, such as ‘tradi-tional non-domesticated plant resources’.

Source: Harris, 1989

Figure I.2 A schematic continuum from wild to domesticated plant species, from foraging of wildspecies to farming domesticated species

Problems in conservation

Seen from the outside, the problems facing conservation and resource management seeminsurmountable. Indeed, many efforts to solve these problems through interventionsplanned from the ‘outside’ by urban-based planners or policy makers have failed. Forthis reason, there has been a move away from centralized planning and identification ofproblems to a decentralized, local approach. Ethnobotanical methods are part of this

Introduction

xvii

Incr

easi

ng

sed

enti

sm (

sett

lem

ent

size

, den

sity

an

d d

ura

tio

n o

f o

ccu

pat

ion

)

Incr

easi

ng

po

pu

lati

on

den

sity

(lo

cal,

reg

ion

al a

nd

co

nti

nen

tal)

Incr

easi

ng

so

cial

co

mp

lexi

ty (

ran

kin

g ➜

str

atif

icat

ion

➜ s

tate

fo

rmat

ion

)

TIMESOCIOECONOMICTRENDS

FOOD-YIELDINGSYSTEM

ECOLOGICAL EFFECTS(SELECTED EXAMPLES)

PLANT-EXPLOITATIVEACTIVITY

PLA

NT–

FOO

D P

RO

DU

CTI

ON

Wild plant–foodprocurement

(foraging)

Wild plant–foodprocurement(with minimal

tillage)

Cultivation(with systematic

tillage)

Agriculture(farming)

Evolutionarydifferentiation of

agricultural systems

Burning vegetation

Gathering/collectingProtective tending

Reduction of competition; acceleratedrecycling of mineral nutrients; stimulation ofasexual reproduction; selection for annualor ephemeral habit; synchronization offruitingCasual dispersal of propagulesReduction of competition; local soildisturbance

Replacement planting/sowingTransplanting/sowingWeedingHarvesting

StorageDrainage/irrigation

Land clearance

Systematic soil tillage

Propagation of genotypic and phenotypic variants: DOMESTICATIONCultivation ofdomesticated crops(cultivars)

Maintenance of plant population in the wildDispersal of propagules to new habitatsReduction of competition; soil modificationSelection for dispersal mechanisms: positiveand negativeSelection and redistribution of propagulesEnhancement of productivity; soilmodification

Transformation of vegetation compositionand structureModification of soil texture, structure andfertility

Establishment of agroecosystems

I

II

III

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xvii

decentralized approach, where people contribute to solutions in resourceful ways, ratherthan being part of the problem.

Innovative, decentralized approaches also have a way of catching on and spreading.Two examples are CAMPFIRE (Communal Areas Management Programme forIndigenous Resources) in Zimbabwe (Child, 1996) and Joint Forest ManagementProgramme projects spread across India and Nepal (Poffenberger et al, 1992a, b; Fischer,1995). Although small, and begun in isolation, these programmes have built up experi-ence and common ground that have been more widely applied. Norman Reynolds’sexperience with community forestry in India, for example, led to recommendations forCommunity Land Companies (CLC), aimed at combining local people’s control ofnatural resource harvesting with effective resource management. CLCs were laterproposed for rural development in southern Africa (Reynolds, 1981) and for strengthen-ing traditional fisheries management (Scudder and Conelly, 1985).

Some creative projects give signs of hope, however, even under the bleakest of circum-stances. Two Central African examples highlight the need for training hand-picked localpeople in protected area management. One of the strongest tests of conservation strate-gies is how resilient they are to the chaos of civil conflicts. Recent tests of this stem fromconservation areas in Rwanda and the Democratic Republic of Congo, engulfed byconflict (Hart and Hart, 1997; Fimbel and Fimbel, 1997). These Central Africanexamples highlight the crucial need for appropriate training for hand-picked local peopleat various levels (rangers, technical staff, research professionals and managers) to takeresponsibility for conservation programmes. International non-governmental organiza-tions have key roles in this process, and one of these is to support this training process.In both cases, international funding was disrupted and expatriate staff left or were evacu-ated due to conflicts in or around the Nyungwe Forest Conservation Project in Rwandaand four World Heritage Sites in the Democratic Republic of Congo. What maintainedthese conservation areas during the conflicts was the presence of local people connectedto these projects.

The important lesson from both cases is summed up from the Rwandan case, whereNyungwe Forest, an Integrated Conservation and Development Project (ICDP) and apriority area for conservation, was held together in the face of lawlessness and land-grabs. Four local people with exceptional leadership qualities continued to collect andsafeguard project records and liaise with people neighbouring the park and local govern-ment representatives. Of 45 local staff, all from villages bordering the conservation area,40 remained, continuing to undertake forest patrols without salaries or communicationsfrom former supervisors or senior staff, who had fled. The main lesson is:

‘… that vehicles, buildings, and short-term consultants supported by large multi-nationals do not make a conservation project. Instead, conservation is achievedby people with commitment. Project personnel recruited from the local popula-tion who demonstrate qualities of leadership and commitment, who receiveregular hands-on training that empowers them to take responsibility for themanagement of their natural resources, are the formula proven to sustain long-term conservation efforts under difficult conditions. The combination of a fewdedicated individuals, together with the support of a non-governmental organi-zation (independent of political constraints) with a long-term commitment toconservation, is the best recipe for achieving lasting success in countries wherepolitical stability is in question, or perhaps anywhere.’ (Fimbel and Fimbel, 1997)

Applied Ethnobotany

xviii

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xviii

People and Plants Partners

The African component of the People and Plants Initiative is supported financially by theDarwin Initiative, the National Lottery Charities Board and the Department forInternational Development (DFID) in the UK, and by the Norwegian Funds in Trust.

DISCLAIMER

While the organizations concerned with the production of this manual firmly believe thatits subject is of great importance for conservation, they are not responsible for thedetailed opinons expressed.

WWFWWF (formerly the World Wide Fund For Nature), founded in 1961, is the world’slargest private nature conservation organization. It consists of 29 national organiza-tions and associates, and works in more than 100 countries. The coordinatingheadquarters are in Gland, Switzerland. The WWF mission is to conserve biodiversity, toensure that the use of renewable natural resources is sustainable and to promoteactions to reduce pollution and wasteful consumption.

UNESCOThe United Nations Educational, Scientific and Cultural Organization (UNESCO) is theonly UN agency with a mandate spanning the fields of science (including socialsciences), education, culture and communication. UNESCO has over 40 years of experi-ence in testing interdisciplinary approaches to solving environmental and developmentproblems in programmes such as that on Man and the Biosphere (MAB). An interna-tional network of biosphere reserves provides sites for conservation of biologicaldiversity, long-term ecological research and testing and demonstrating approaches tothe sustainable use of natural resources.

ROYAL BOTANIC GARDENS, KEW

The Royal Botanic Gardens, Kew, has 150 professional staff and associated researchersand works with partners in over 42 countries. Research focuses on taxonomy, prepara-tion of floras, economic botany, plant biochemistry and many other specialized fields.The Royal Botanic Gardens has one of the largest herbaria in the world and an excel-lent botanic library.

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xix

Acknowledgements

This manual is dedicated to the students and young professionals from rural communi-ties in developing countries who often work in isolation, sacrifice a great deal, and whoare at the forefront of conservation efforts and ethnobotanical research. It is alsodedicated to my family, who have spent so much time without me when I have been inthe field – or immersed in this manual.

A manual of this type is not just based on years of field work, but is a product ofdiscussions with many colleagues over a long time: too many to acknowledge individu-ally here. Nevertheless, I must thank a few people: William Bond, Charles Breen, BruceCampbell, Bekazitha Gwala, Margie Jacobsen, Jeremy Midgley, Eugene Moll, JacksonMutebi, Bev Sithole, Ken Tinley, Fiona Walsh, Rob Wild and Siyabonga Zondi for inspir-ing discussions; my colleagues in the People and Plants Initiative for their support andfor reading through parts of this manual at different stages: Alan Hamilton, Gary Martin,Robert Hoeft and Yildiz Aumerruddy. Richard Cowling, Martin Luckert, JeremyMidgley, Jack Putz and Trish Shanley also commented on sections of the manual, as didthe series editor, Martin Walters.

I must thank both Martin Walters and Alan Hamilton for their interest, patience andunderstanding during the long process of writing this manual between field trips andsupporting students. Wendy Hitchcock is thanked for drawing several of the figures.Reprinted figures are acknowledged in the text, but I must thank Charles Peters for theuse of several figures from his ecological primer, Terry Sunderland, De Wet Bösenberg,Robin Guy, Glen Mills, Fiona Walsh and Yildiz Aumeeruddy for contributingphotographs, and the McGregor Museum, Kimberley, South Africa, for permission touse an unpublished photograph (Figure 4.6) from the Duggan-Cronin collection. Wherenot acknowledged, the slides are my own.

Finally, I should like to thank the organizations that have funded the African compo-nent of the People and Plants Initiative, as this has also supported production of thismanual: funds to WWF from the Darwin Initiative, the National Lottery Charities Board(NLCB) and the Department for International Development (DFID) in the UK, and toUNESCO through the Norwegian Funds in Trust.

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page xx

Throughout the world, wild, naturalizedor non-cultivated plants provide a ‘greensocial security’ to hundreds of millions ofpeople, for example in the form of low-cost building materials, fuel, foodsupplements, herbal medicines, basketrycontainers for storage, processing orpreparation of food crops, or as a sourceof income. Edible wild foods often helpprevent starvation during drought, whileeconomically important species provide abuffer against unemployment duringcyclical economic depressions. This isparticularly important for people living inareas with drought-susceptible soils ofmarginal agricultural potential, such asthe vast areas of sub-equatorial Africacovered by leached, nutrient-poor sands(Figure 1.1). Despite the immense impor-tance of these plant resources, their valueis rarely taken into account in land-useplanning; and when it is, it is oftenassumed that these species are sustainablyharvested and that this ‘green socialsecurity’ will always be available toprovide a safety net for resource users.This is not always true. Although manyecosystems and harvested species popula-tions are resilient and have a long historyof human use, they can be pushed beyond

recovery through habitat destruction oroverexploitation.

Cultural systems are even moredynamic than biological ones, and the shiftfrom a subsistence economy to a casheconomy is a dominant factor amongst allbut the remotest of peoples. In many partsof the world, ‘traditional’ conservationpractices have been weakened by culturalchange, increased human needs andnumbers, and by a shift to cash economies.There is a growing number of cases whereresources which were traditionallyconserved, or which appeared to beconserved, are today being overexploited.The people whose ancestors hunted,harvested and venerated the forests thatare the focus of enthusiastic conservationefforts are sometimes the people who arefelling the last forest patches for maizefields or coffee plantations, often on slopesso steep that sustainable agriculture isimpossible. In other areas, local humanpopulations have decreased due toepidemic disease or even urbanization,with swidden agriculture only occurringon old secondary forest. While someresources are being overharvested due tocultural and economic change, the major-ity are still used sustainably, and the

Chapter 1

Conservation and Context: Different Times,Different Views

Introduction

Applied Ethnobotany 20/11 04/12/2000 05:28 pm Page 1

impact on others has lessened because ofsocial change. In the most extreme cases,‘islands’ of remaining vegetation, usuallycreated by habitat loss through agricul-tural clearance, then become focal pointsfor harvesting pressure, and are sites ofconflict over remaining land or resources.

For all interest groups, whetherresource users, rural development workersor national park managers, it is far betterto have proactive management and to stopor phase out destructive harvesting infavour of suitable alternatives before

overexploitation occurs, than to have the‘benefit’ of hindsight in the midst of adevastated resource. Marilyn Hoskins(1990) puts this well in her paper onforestry and food security:

All research and management byoutsiders must remember thattheir activities come and go, butfood security – land and resourcessurety – is a long-term, life anddeath issue for rural peoples.

Applied Ethnobotany

Source: Cooke, 1964

Figure 1.1 Soils are a major determinant of reliance on plant use. In Africa, for example, the distribution of leached, nutrient-poor and drought-susceptible sands of the coast (light grey

shading) and Kalahari basin (dark grey shading) affects most land users. Whether they are hunter-gatherers, pastoralists or farmers, people remain dependent to some extent upon wild plants (and

the associated edible insects) for food supplements, housing, fuel, furniture and fibre forhousehold containers

2

Equator Equator

0 km 1000

Orange

Limpopo

Na

mi b

Sa

nd

s

Mo

z am

bi q

ue

Sa

nd

sZambezi

Kasai

Congo

K a l a h a r i

S a n d s

K a l a h a r i

S a n d s


Since the 1960s, the approach to conser-vation in developing countries hasbroadened from its past emphasis onstrictly policed protected areas, or land setaside for large mammals or spectacularlandscapes. Nowadays the emphasis hasshifted to sustainable resource use and themaintenance of ecological processes andgenetic diversity and a broader approachinvolving land users in ‘bioregional’management at an ecosystem level.

This broader approach is evident in thedifferent IUCN categories of protectedareas which were developed in the mid1980s and recently modified at the IVWorld Congress on National Parks andProtected Areas (Box 1.1). It also repre-sents a change from one where interventionby the state (by government) throughproclamation of national parks was seenas the solution, to one where the conserva-tion roles of private landowners andresidents of communal lands are recog-nized. It also became widely accepted thatthe future of most conservation areaslargely depends upon the acceptance andsupport of the surrounding human popula-tions. In Africa, for example, theconsequences of political turmoil, changesof government and a ‘brain drain’ of parkbiologists and policy makers, reinforceJonathan Kingdon’s (1990) point that:

‘… the realities of power areexactly the opposite to thoseperceived by most of the partici-pants of this struggle to conservekey areas of high endemism andbiodiversity because the long-termfuture of Africa’s Centres ofEndemism lies with localpeasantries rather more than withtransient governments or enthusi-

astic conservationists; yet localsseldom receive the respect that isgenerally accorded to those thatwield power. Meanwhile, bothpopulations and resentmentsgrow.

… The conservationists’ answersshould not lie in propagandacampaigns, which are generallyseen for what they are, but in ashared growth of knowledge anddebate. The minimal demands oflocal communities will includesustained, not ephemeral,programmes of action in whichtheir own people can findmeaningful, decisive and dignifiedroles.’

At a meeting in Tanzania in the 1960s, SirJulian Huxley suggested that the means tojustify conservation as a form of land useto local people or national governmentscentred upon ‘pride, profit, protein andprestige’. Little attention was paid to wildplants and their importance to ruralpeople. This is no longer the case. There isnow a strong emphasis on sustainable useof resources, including wild plants, and theinvolvement of national governments andlocal people in conservation. Buffer zones,formed around strictly protected coreconservation areas, have been one of thetools in this process and are a characteris-tic planning tool of biosphere reservesestablished by the UNESCO Man andBiosphere programme. This approach isembodied in many recent policydocuments, such as the WorldConservation Union’s (1991) strategydocument Caring for the Earth, and morerecently, the World Resources Institute’s

Conservation and Context: Different Times, Different Views

Historical context

3


Applied Ethnobotany

4

BOX 1.1 IUCN PROTECTED AREA CATEGORIES: THE MODIFIED

SYSTEM OF PROTECTED AREAS CATEGORIES AGREED AT THE

IV WORLD CONGRESS ON NATIONAL PARKS AND PROTECTED

AREAS, 1992

1 Strict Nature Reserve/Wilderness AreaAreas of land and/or sea possessing some outstanding or representative ecosystems,geological or physiological features and/or species, available primarily for scientificresearch and/or environmental monitoring; or large areas of unmodified or slightlymodified land, and/or sea, retaining their natural character and influence, withoutpermanent or significant habitation, which are protected and managed so as topreserve their natural condition.

2 National ParkProtected areas managed mainly for ecosystem conservation and recreation. Naturalareas of land and/or sea, designated to:

• Protect the ecological integrity of one or more ecosystems for this and futuregenerations.

• Exclude exploitation or occupation inimical to the purposes of designation of thearea.

• Provide a foundation for spiritual, scientific, educational, recreational and visitoropportunities, all of which must be environmentally and culturally compatible.

3 Natural MonumentProtected areas managed mainly for conservation of specific features. Areas contain-ing one or more specific natural or natural/cultural features of outstanding or uniquevalue because of their inherent rarity, representative or aesthetic qualities or culturalsignificance.

4 Habitat/Species Management AreaProtected areas managed mainly for conservation through management intervention.Areas of land and/or sea subject to active intervention for management purposes inorder to ensure the maintenance of habitats and/or to meet the requirements ofspecific species.

5 Protected Landscape/SeascapeProtected areas managed mainly for landscape/seascape conservation and recreation.Areas of land, with coast and sea as appropriate, where the interaction of people andnature over time has produced areas of distinct character with significant aesthetic,cultural and/or ecological value, and often with high biological diversity. Safeguardingthe integrity of this traditional interaction is vital to the protection, maintenance andevolution of such areas.

6 Managed Resource Protected AreaProtected areas managed mainly for the sustainable use of natural ecosystems. Areascontaining predominantly unmodified natural systems, managed to ensure long-termprotection and maintenance of biological diversity, while providing at the same time asustainable flow of natural products and services to meet community needs.Source: IUCN


(1992) Global Biodiversity Strategy (WRI,1992) and the World Convention onBiological Diversity (see Glowka et al,1994).

Policies on sustainable development orcalls for sustainable use of resources bylocal people within protected areas (forexample, Ghimire and Pimbert, 1997) arefine on paper. The challenges arise withtheir implementation. In their review ofconservation projects, trying to make alink between parks and people throughplanning of buffer zones and links todevelopment (which they termedIntegrated Conservation and Development

Projects), Michael Wells and KatrinaBrandon (1992) found very few bufferzone models which they were convincedworked well. As usual, ‘the devil is in thedetails’: and if policies are impractical,they are worthless. If implementationresults in resource degradation rather thanthe sustainable use intended, then the self-sufficiency of resource users is furtherreduced, increasing the likelihood of land-use conflict between national parks andpeople. This manual is about one of thosedetails: the sustainable harvesting of wildplant resources.


Management myths and effective partnerships

5

Policy changes towards sustainable use ofresources in conservation areas have placedmany field researchers and national parksmanagers in a dilemma. How do we gobeyond the rhetoric of policy on humanneeds and sustainable resource use withoutjeopardizing the natural resource base orprimary goal of the conservation area: themaintenance of habitat and species diver-sity? This is no easy task. The higher thenumber of harvesters, the more uses a plantspecies has. The scarcer the resource, thegreater the chance that resource managersand local people will get embroiled in acomplex juggling of uses and demands, inan attempt at a compromise that could endup satisfying nobody.

In theory, sustainable harvesting ofplants from wild populations is possible,but is often more complex than the urbanbiopoliticians and policy makers think.Sustainable management of wild plant useby people depends as much upon anunderstanding of the biological compo-nent as it does on the social and economicaspects of wild plant use. Without anunderstanding of ecological, political and

socio-economic factors (Figure 1.2), plansfor sustainable use are likely to fail.

Sustainable use of resources by localpeople and the concept of ‘extractivereserves’ appear to have been promoted indeveloping countries on the basis of twocommonly held assumptions that:

1 Local (or indigenous) peoples havebeen harvesting these resources forthousands of years, with no detrimen-tal effects on harvested populations.Traditionally, many useful biologicalresources have been valued andconserved. Therefore if ‘resourcesharing’ between national parks andneighbouring peoples takes place inbuffer zone areas, then the people livingaround that national park will have aninterest in conserving resources for thefuture, and will harvest these resourcesin a sustainable way.

2 Wider recognition for wild plantproducts, whether from forests,savanna or wetlands, will result inmore appropriate values being placedon vegetation currently being damaged


for a few products (hardwood timber-logging, charcoal) or cleared foragriculture or pasture, ignoring otherprotective (eg watersheds) and produc-tive (nuts, oils, fibres, etc) functions.

While there is some truth in these assump-tions, they have reached almost mythicalproportions, with the result that localresource users are often considered naturalconservationists who have always usednatural resources sustainably. There is nodoubt that ‘traditional’ conservationpractices existed in many societies, andthat these have buffered the effects of

people on favoured species and in selectedhabitats. Customary restrictions can alsobe an important guide to culturally accept-able limits on the harvesting of vulnerablespecies (see Chapter 6). Equally, there aremany examples of resource overexploita-tion prior to the introduction of firearmsand more efficient hunting technologies orlarge-scale, species-specific commercialtrade.

Also often glossed over is the fact thatprotected areas, particularly those with ahigh species diversity and vulnerability tooverexploitation, require a level of detailedmanagement that is not possible with the

Applied Ethnobotany

Source: Martin, 1994

Figure 1.2 Achieving sustainable use of resources requires cross-disciplinary work at theconfluence of the social sciences, economics and ecological studies, all within a political

(and policy) framework

6

ECOLOGY

CULTURALAND SOCIAL

SUSTAINABLEUSE

ECONOMICS

The focus of mostconservation research

Component mostignored in

conservation circles

The majordriving force

Politics and policy


economic constraints that are a feature ofmany conservation departments. Despiteincreased awareness of environmentalconcerns and international backing forconservation, many national parks haveinadequate staff or funding to control oftenlarge protected areas. The level of respon-sibility faced by two young Ugandanssupported through the People and PlantsInitiative provides a typical example: oneUgandan is the only ecological monitoringofficer for Rwenzori Mountains National

Park, 1300km2 in extent. The other iswarden of Semliki National Park, 219km2

in size. While similar sized parks in the USwould have a team of ecologists or parkmanagers, these young professionals carryimmense responsibility for huge areas ontheir own. This is not an unusual situationin developing countries, and worldwidethere is an inverse relationship betweenplant species richness and numbers ofecologists (Figure 1.3).


Source: World Resources Institute, 1992

Figure 1.3 The distribution of professional ecologists in relation to the distribution of plantspecies richness

7

Vegetation change: spatial and time scales

Europe andNorth Asia

50

40

30

20

10

0

100

80

60

40

20

0North

AmericaSouth and

Southeast AsiaSub-Saharan

AfricaCentral and

South AmericaAustralia andSouth Pacific

Ecologists (per cent) Vascular plant species (thousands)

Ecologists Species

Lack of communication between disci-plines has led to a number ofmisconceptions, myths and inaccuracies inconceptual models. Limits on a ‘cross-pollination’ of concepts have occurred attwo levels: firstly, between different acade-mic disciplines, principally between thebiological and social sciences; andsecondly, between formally trainedresearchers and local peoples and resourceusers. The resultant misconceptions havehad important implications for biodiver-

sity conservation, and have prevented aclear understanding of how ecologicalsystems function and how dynamic biolog-ical and cultural systems change over time.Climate change over long time scales canbe superimposed by human-inducedchanges on vegetation, for example, whilecultural change can be rapid.

To facilitate informed decisionmaking, plant use and conservation policyhave to be seen against the backgroundinfluences of climate and human distur-


bance of ecosystems (see Chapter 6). Eachhas had a major influence on worldvegetation in the past. This will increas-ingly be the case in the future, with theeffects of high human consumption rates,high population growth rates and globalwarming.

Massive oscillations of Pleistoceneclimate, accompanied by expansions andshrinking of polar ice caps, resulted inlong, cool, dry periods, alternating withshorter, warmer, moist periods. Equatorialforests, as indicators of world climaticconditions, are believed to have expandedoutwards from, or shrunk into, Pleistocenerefugia. In several parts of the world,pollen analysis from cores, usually taken in

lakes, swamps or bogs, provides evidenceof vegetation dynamics and climate changeover long periods of time. Pollen analysisin Uganda, for example, provides a recordof vegetation history over the past40,000–50,000 years, including forestexpansion about 10,600 years BP (beforepresent) (Taylor, 1990). This also showsthat during the most recent glacial phase(pre-12,000 years BP), forests wererestricted to a few refugia, expandingoutwards with moister, warmer conditions(Hamilton, 1981). Conserving forestswhich retained forest cover during this aridphase can be extremely important sincemany have a high biological diversity.

Applied Ethnobotany

Human influence: landscapes and species

8

Human disturbance and deliberate modifi-cation of vegetation have beensuperimposed on natural disturbance,sometimes in the relatively recent past.Local people in many parts of the worldhave also favoured certain useful speciesthrough traditional conservation practices,dispersal and planting. Anthropogenicchanges caused by agricultural clearing,burning patterns or species-selectiveoverexploitation are sometimes over-looked at a policy level. Archaeologicalstudies similarly show the extinction ofmammal and bird species on islands suchas Hawaii and New Zealand, or thecomplete disappearance of forest habitatand palm woodlands on Easter Island afterthe arrival of Polynesians (Diamond,1992; Flannery, 1995). Human-induced oranthropogenic changes due to the use offire, for example, have long been recog-nized by ecologists as contributing to themaintenance of African savanna or ofprairie grasslands in North America.Archaeologists have also provided detailed

evidence of how long-lasting such changescan be in creating small-scale patcheswithin savanna woodlands, as in the 1000-year-old Cenchrus ciliaris grass patches onearly Iron Age dung accumulations inBotswana.

In tropical forests, this has been lesscommonly recognized until recently. Tomany city people who support rainforestconservation, whether they are from urbanareas of the tropics such as São Paulo orBangkok, or temperate cities such asLondon and New York, even disturbedsecondary rainforest might appear to bepristine. This is understandable, giventheir unfamiliarity with these environ-ments – but until biologists started talkingto people, they had also often been misled.Darrell Posey, for example, working withKayapo people of the Brazilian Amazon,has shown that certain ‘wild’ plants alongpaths through ‘pristine’ forest are in factplanted by the Kayapo as a source of food,medicine and other resources (Posey,1984). Expanding this approach to the


entire Amazon terra firma forest, WilliamBalee (1989), of the New York BotanicalGarden, estimates that at least 11.8 percent of this forest is anthropogenic. Evenbiologists well aware of the dynamicnature of vegetation change can besurprised by what they find in seemingly‘undisturbed’ forests in remote areas. Thusbiologist Alan Hamilton, digging soil pitsin ‘remnant’ forests of the UsambaraMountains of Tanzania, East Africa(which had been selected as soil samplesites for their ‘undisturbed’ status),regularly found charcoal and pottery; thesites had been occupied by people withiron-smelting and agricultural technologyfrom about 1800 years ago (Hamilton andBensted-Smith, 1989).

If human influence and culturallandscapes are so widespread, why then usethe term ‘wild’ plants in the title of thismanual? I raise this point because acolleague involved in policy and intellectualproperty rights issues was uncomfortablewith the term ‘wild’. The main reason forthis concern is that the term is linked to theword ‘wilderness’, usually taken to meanan uninhabited or uncultivated tract ofland. Use of the word ‘wild’ was thenconsidered to undermine the issues ofindigenous peoples’ intellectual propertyrights. The sense in which the word ‘wild’is used here is explained in the introduction.At a species level, I use the term ‘wild’ todistinguish between wild and domesticatedplant species, not to suggest that thelandscapes where they occur are virgin

land, unaffected by human influence ortenure.

Out of a global flora of 270, 000 plantspecies, relatively few are domesticated(species whose breeding systems have beenso changed through genetic or phenotypicselection that they have become dependentupon sustained human assistance for theirsurvival). The vast majority of species arewild. Others along the continuum arereplanted from wild-collected seed orseedlings, self-sown species which aremanaged or tolerated in fields, or semi-domesticates in the process ofdomestication, where phenotypic (andgenotypic) modifications have arisenthrough people deliberately selectingfavoured characteristics. The quantitativeethnobotanical studies by Alejandro Casasand Javier Caballero (1996) on selectivemanagement of Leucaena by MixtecIndians is a good example of this process.Innovative quantitative studies like theirsthat carefully document this process areextremely useful.

Also important is the development ofquantitative methods and predictivemodels, rather than lists of species oranecdotal data. Such models can lead to amore effective conservation of the remain-ing habitats. These sites may hold the wildrelatives of domesticated species, or wildspecies which are too slow growing andtake such a long time to reach reproduc-tive maturity that in situ conservation istheir only option.


9


The methods outlined in this chapter arethe steps used to get a better understand-ing of people’s preferences and the demandfor particular plant species. Althoughsome plant uses such as harvesting ofwood for fuel, building or commercialwoodcarving are more obvious, occurringthroughout the year and in large volume,wild plant gathering is often part of a‘hidden economy’ unnoticed by outsiders.Consequently, careful field observation,sensitive consultation with local harvestersand strategic planning are required beforeany monitoring takes place. Even theidentity of some commonly harvestedspecies, often well known to local people,is often poorly known to protected areamanagers or outside researchers. Each

method provides useful information on itsown, but ideally should be cross-checkedagainst data collected using differentmethods.

If alternatives are to be provided toprevent overexploitation of resources orto defuse land-use conflicts before demandexceeds supply, it is also important toknow what quantities of plant material arebeing harvested. If this is not known, it isvery easy to underestimate quantities ofthe resource required, or to provide onlypiecemeal alternatives in such smallquantity that they are of little practicalvalue. The ‘resource demand’ componentdiscussed here and in Chapter 3 leads onto the ‘resource supply’ componentscovered in Chapters 4, 5 and 6.

Chapter 2

Local Inventories, Values and Quantities ofHarvested Resources

Introduction

Local priorities: vegetation types, resource categories and species

A reversal of roles has always been implicitin ethnobotanical work. Formally trainedoutsiders, whatever their experience, havea lot to learn from the insights of localpeople who are acknowledged within theirown communities as experts on local

vegetation. As a result, local people play acrucial role at several parts in the researchprocess, including research design, speci-men and data collection, interpretation ofdata and, less commonly, the presentationof research results back to the community.


Amongst development practitioners, thistype of approach, where local people helpin conducting research, has given rise tothe term ‘participatory research’. It isimportant to realize that this does not referto a single research method, as Brian Prattand Peter Loizos (1992) point out:

‘Although some writers make itsound as though there is a separate“participatory” research method,this is misleading. The idea ofparticipation is more an overallguiding philosophy of how toproceed, than a selection ofspecific methods. So when peopletalk about participatory research,participatory monitoring andparticipatory evaluation, on thewhole they are not discussing aself-contained set of methodolo-gies, but a situation whereby themethods being used have includedan element of strong involvementand consultation on the part of thesubjects of the research. Not allmethods are equally amenable toparticipation.’

Considerable common ground for jointwork in resource management often exists.Resource users, development workers andprotected area managers often have acommon interest in cases of conflicts overvalued but vulnerable plant resources.This can be due to restrictions on harvest-ing of rare species or vegetation withinprotected areas, to overexploitationarising from demand exceeding supply foruseful wild plant species, or to conflictsbetween local harvesters and people fromoutside the community. Involvement ofresource users as research partners is anessential part of a successful conservationstrategy for useful plant species that arevulnerable to overexploitation. There arethree main reasons for this.

Firstly, the knowledge and perceptionsof resource users such as traditionalhealers, craft workers and commercialmedicinal plant harvesters providevaluable insights into the scarcity of usefulplant species. It is these resource users whowalk further or pay more for scarceresources, and are thus aware of scarcitylong before any conservation biologists.Their knowledge therefore provides a‘short-cut’, saving time and money, andenabling biologists to monitor key species.Local knowledge represents a practicaland cost-effective method for identifyingpossible key species. In some cases, as withsmall, cryptic and low-population densityplants such as Schlechterina mitostemma-toides (Passifloraceae), it provides themain evidence of occurrence as commer-cial trade items, and can direct specialistmonitoring and conservation programmes.The validity of local knowledge can alsobe tested against data in herbaria and inliterature on the geographical distribution,rarity and extent of exploitation of species.Thus, local traders’ conceptions of scarcitymay be a result of limited geographicaldistribution rather than overexploitation –for example, in the case of the medicinalplant Synaptolepis kirkii (Thymelaeaceae)in South Africa.

Secondly, dialogue with resource usersis a crucial part of developing conserva-tion and resource management proposalswith, rather than for, resource users. Thisincludes interaction with resource usersabout their perceptions as to why scarcityhas arisen, setting quotas and humancarrying capacities if practical, and identi-fying appropriate alternatives and howthese can be implemented.

Thirdly, it enables specialist usergroups to be identified. Rural communi-ties are not homogeneous, but are complexnetworks, divided on the basis of power,gender and specialist interest groups.People who specialize in harvesting

Local Inventories, Values and Quantities of Harvested Resources

11


specific resources such as medicinal plants,basketry fibres or woodcarving timber canhave a direct interest in maintaining ruralself-sufficiency and in ensuring thatfurther resource degradation (or alterna-

tively, restoration of diversity and self-sufficiency) takes place. Identifyingdifferent user groups plays an importantrole in the social side of resource manage-ment (see Chapter 7).

Applied Ethnobotany

Choosing the right methods

12

Choices of methods should be made case-by-case on the basis of preliminaryplanning, bearing in mind timetable andbudget constraints. More detail is given onthese methods in this chapter and inChapters 3 (ethnobotanical surveys ofmarkets), 4 to 6 (harvesting impacts andvegetation dynamics) and Chapter 7(tenure).

Once permission for research workhas been granted at a national and locallevel, then researchers need to decide onthe survey methods that are appropriate.Commonly used methods are:

• discussions with individual resourceusers;

• group interviews and discussions;• rapid rural appraisal (RRA), partici-

patory rural appraisal (PRA) andparticipatory assessment, monitoringand evaluation (PAME);

• social surveys using various samplingtechniques and structured or semi-structured interviews;

• participant observation;• ethnobotanical inventory methods;• sample surveys based on field records

with local resource users;• surveys of plants sold in local markets

(see Chapter 3).

In the past five years, several excellentmanuals and reviews have been publishedthat give detailed descriptions of differentmethods. Many of these are readily avail-able through organizations supporting

field work in developing countries. Ratherthan repeat the detail contained in thesemanuals, I will first give an overview ofsocial survey methods that are commonlyused, such as interview surveys andvarious participatory methods. I alsorecommend that, if at all possible,researchers or the organizations they areworking for should obtain these usefulmethods manuals that are suggested forfurther reading at the end of this chapter. Ithen describe ethnobotanical surveymethods and approaches in more detail.Ethnobotanical methods mentioned byGary Martin (1995) or by John Tuxill andGary Nabhan (1998) are then described inmore detail in this chapter. Other recentand useful reviews of methods covered inthis chapter are Oliver Phillips’s (1996)review of quantitative methods, andDarna Dufour’s and Nicolette Teufel’s(1995) description of methods for assess-ing food use and dietary intake.

There are four recent methods manualsthat I would recommend for fieldresearchers planning to use social surveymethods: Social Survey Methods: a Field-Guide for Development Workers (Nichols,1991); Choosing Research Methods: DataCollection for Development Workers (Prattand Loizos, 1992); The CommunityToolbox: the Ideas, Methods and Tools forParticipatory Assessment, Monitoring andEvaluation in Community Forestry (FAO,1990) in the FAO Community Forestryfield manual series available fromFAO/SIDA Forests, Trees and People



13

1 Resource users identify problem: which species, where,why?

2 Are background data available? Formal or informalsales data? How do trends in sales (up) compare tonatural resource supply (down)?

3 For key species, translate ‘user units’ (bundles, bags,baskets) into ‘natural resource units (leaves, stems etc➜ plants per hectare ➜ total area ➜ production rates➜ yields)

4 With resource users: map, measure, mark, evaluate(resources, harvest methods, impacts, ‘harvest per uniteffort’)

5 Identify practical alternatives (and local opinion onhow these could best be implemented, evaluation,adjustment of methods)

Palm

150

120

90

60

30

0Dye 1 Dye 2 Vine

Number of respondents

Shortage No shortage

Grass

1972

500

400

300

200

100

0

Total sales value (Pula) (thousands)

Total sales Local basket sales

Basketry materials

Botswanacraft total sales

Export basket sales

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987

Best method: palm cultivation

Etsha interviews (n=123)

Producerplants own

57%

Communalgroups29%

No preference 3%

Do not know 1%

Large plantations 8%

Individuals and group 2%

• • • • •

• No data

Figure 2.1 Five basic steps in dealing with plant-resourcemanagement issues


Programme; Participatory Learning andAction: a Trainer’s Guide (Pretty et al,1995), produced by the InternationalInstitute for Environment and Development(IIED), London, in their ParticipatoryMethodology series.

These all give detailed and well-illus-trated descriptions of social surveymethods that are very useful in ethnob-otanical work. Robert Chambers (1992)also provides a very readable account ofPRA approaches in a discussion paperproduced by the Institute for DevelopmentStudies.

Once the geographic focus of the studyhas been decided, it is useful to reviewrelevant studies that have been done in thesame region, or on the same or compara-ble resource species. For researchers fromoutside the region, it is also worthwhilefinding out whether cross-references ofvernacular to botanical names for thatregion are available. Background biologi-cal, social and economic information cancome from a wide variety of published orunpublished reports, as well as fromdiscussions with people who have livedand worked in the region for a long time.These may include, for example, previousecological or social science studiespublished in journals, or unpublishedreports of non-governmental organiza-tions (NGOs) or government departments.A considerable amount of time can besaved by consulting and obtaining datafrom annual reports, export statistics andpopulation census information in reportsand publications of departments of health,trade and industry, agriculture andforestry, survey and lands, or geological

survey. These may be available in a univer-sity, herbarium or government library. Ineach case, it is useful to have an introduc-tion to someone in these organizationswho is aware of the aims of the research.If not, despite the delays that may becaused by bureaucracy and poor filingsystems, time can be saved through visit-ing these departments in the capital city orregional centre.

If informal sector harvest and tradefeeds into a formal trade network, thenquantities may be reflected in official tradestatistics or forestry department records.Examples are the quantities of Brazil nutssold, in data summarized by Prance(1990), or the quantity of Prunus africanabark bought by a factory in Cameroon (seeFigure 2.4). Wherever possible, suchfigures need to be cross-checked foraccuracy, and one should always be scepti-cal unless proved otherwise. Whencross-checking, you may need to work‘backwards’ from marketing surveys orexport data to field studies. Althoughcertain assumptions can be made fromtrade statistics on the basis of what part ofthe plant is harvested, and from thepopulation biology, abundance and distri-bution of the species, trade data havelimited value unless they are combinedwith field studies providing a link betweenvolumes traded and what is happening tothese species in the wild. In many cases,however, plant species are traded by theinformal sector and no trade records exist:so if you need to document the quantitiesinvolved, you need to collect the datayourself (see Chapter 3).

Applied Ethnobotany

14


Identifying local values or quantities ofresources used is extremely difficultwithout community support: somethingthat requires courtesy, consideration andtime. Local support, in turn, is influencedby the social survey methods used, and theapproach and attitude of the researcher,whether local or not. Ideally, it is best tocarry out surveys with the help of a team,which should also include localresearchers. In reality, however, this maynot always be possible, and it is then evenmore important to ensure that all effortsare made to obtain as much relevantbackground information as possiblebefore undertaking a survey. Before start-ing any field work, it is also worththinking about the social context of theresearch: ‘who watches whom’ and ‘whosepriorities’ are questions worth bearing inmind before field work begins.

In contrast to research studies whosehypotheses and objectives are set in urbanlaboratories, the objectives and methodsfor resource management research are bestdecided on in the field through preliminarywork with local resource users andresource managers. In this process, it isimportant to consider the following:

• Who are the resource users: are theymen, women or children and are theyspecialist plant users such as herbal-ists, weavers or midwives?

• What is their socio-economic andformal educational status?

• Is harvesting for commercial or subsis-tence purposes, or a combination ofboth?

• Which species or resource categories(eg fuelwood, thatch) are most indemand or most valued (culturally,economically, nutritionally)?

• When, where and how does collectiontake place – for instance, season,vegetation types (and patches within avegetation type): what skills andtechnology are required?

• What are the effects of harvesting onplant populations and which speciesare most vulnerable to overexploita-tion?

• Are overexploitation and increasedresource scarcity of concern withinlocal communities (rich vs poor),nationally or internationally?

• In the case of multiple-use species,what are the effects of harvesting onespecies on the availability of otherdesired natural products?

Field observation: who watcheswhom?

With few exceptions, the resource usersyou are working with are not only percep-tive observers of the environment, they areequally good observers and judges ofhuman nature. Furthermore, they areoften aware that outsiders, whetherresearchers, government officials or peoplefrom NGOs, may have worked in this ornearby communities in the past and madepromises which were never kept, arrivedwith hidden agendas or took far more thanthey gave. Be well aware that this is alearning period on both sides, and acrucial one that can set a positive ornegative tone for later work.

Field surveys with local people aremore than just asking about uses and localnames of plants; they also enable localpeople to ask questions of the researchers,such as: what attitudes do researchersshow to local people and to one another?How serious and interested are they in


15

Before starting: attitudes, time spans and cross-checking


addressing the problems raised at thecommunity meetings preceding theresearch? How much do they know aboutthe local vegetation? How do theymeasure up in working, camping or merelywalking through the forests or woodland?Do they act like a bunch of city slickers orlike people used to the bush? If they arecompletely ignorant, what hope is there oftheir resolving the problems that havebeen raised? Demonstrating some of yourknowledge does not mean that you shouldshift away from your role as a personquietly questioning and stimulating discus-sion so that local people, as experts intheir own right, contribute to the discus-sions. Your field knowledge will beapparent to local resource users from thequestions you ask, the local terminologyyou use or the approaches you follow inidentifying plants.

The ‘walk in the woods’ approach inthe first stages of field work is an impor-tant opportunity to work in the field withthe local people who know it best, as astimulus for discussions and an opportu-nity for field observation. The emphasis ison relaxed and open-minded fieldwork,avoiding repetitive questioning andencouraging free-ranging discussion onplant uses and plant ecology. It may alsobe a time to observe and discuss signs ofharvesting or patterns of plant distributionin relation to soils and disturbance. It alsogives the local people the opportunity toobserve and get to know the researchers,which is particularly important if they areoutsiders; this is an important step whichis often lost in ‘rapid’ surveys.

Short-term ‘snapshots’ or ‘long-term surveys’

Do the aims of the research match up withthe time and funding available? Howaccurate do you want to be? Fluctuationsin volume of wild plant resources used,

with season, site differences of vegetationor markets as well as cultural factors,make estimates from short-term‘snapshots’ difficult, and the results willbe doubtful, whatever the short-termtechnique used.

In the early 1980s, for example, Istarted a study of the palm-wine trade,developing appropriate forms and traininglocal enumerators at four palm-wine salepoints. All palm-wine sales weremonitored (the sale days being Monday,Wednesday, Friday); as the first fewmonths of data accumulated, I excitedlydid rough calculations, extrapolating frommonthly sales volumes to the whole year.These preliminary estimates of totalannual sales based on a few months’ datawere interesting, but with the benefit ofhindsight, completely wrong. Once I had12–18 months of sales data, I realized theextent of seasonal variations, with palm-wine sales rising in early summer(October–December), then plummetingbetween January–March when a muchtastier local beer from Sclerocarya birreafruits was available (see Figure 2.2b).

In addition, there were variationsbetween markets and changes in yields ofpalm sap to take into account (see Figure2.2c). If such survey work had been limitedonly to ‘snapshots’, the results would havebeen very different. If I had estimated totalannual volume of palm wine sold on thebasis of three months of sales records forOctober–December or, alternatively,records from January–March, they wouldhave given totally different results, bothincorrect. The same danger of extrapolat-ing from short-term surveys also applies toother resources, including edible plants (seeFigure 2.3), fuelwood or building materi-als. Long-term monitoring of quantitiessold is expensive and time consuming,however, so you also need to ask yourself‘what level of precision is required?’

Applied Ethnobotany

16



Figure 2.2 The palm-wine trade and variation in volumes sold. (a) Daily measurements of sapyields to an individual tapper, keeping different ‘batches’ of sap separate. (b) Regional volumessold at a single sales point. (c) Variation in yields from different groups of palms tapped by a

single tapper

17

0

30

20

10

0

Yield (l)

(c)

50 100 150 200 250 300 350 400

Time (days)

Batch number 1 2 3 4 5 6 7 8 9 10a 10b 11 12 13

50

40

30

20

10

0

Volume sold (thousand litres)(b)

O N D J F M A M J J A S O N D J F M A M J1981 1982 1983

Peak Sclerocarya birreafruiting season

Peak Sclerocarya birreafruiting season

(a)


Voucher specimens

If ‘rapid’ surveys are being done, forexample by using participatory methods(PRA, RRA or PAME), then field collec-tion of voucher specimens may be limitedto species identified by vernacular name.However, all surveys must allow time forcross-checking of information and collec-tion of good voucher specimens withflowers and fruits (see ‘EthnobotanicalInventories’ below). Time constraints limitthe accuracy of work in complex culturaland biological situations, and ‘cautionarytales’ abound, illustrating errors whenneither cross-checking nor collection ofvoucher specimens took place.Unfortunately, when these errors creepinto published literature, they tend to getperpetuated.

One such example is a paper on theethnobotany of Hambukushu people,

published by an anthropologist afterextensive field work in Botswana. Despitehis supposed fluency in Hambukushu, thepaper is packed with errors that wouldhave been avoided through just twothings: cross-checking and the collectionof reasonable voucher specimens. Neithertook place. As a result, the paper lists localnames for incorrectly identified treespecies that do not occur within the areaat all. Vernacular names meaning ‘flower’,‘fruit’ or even kataratara, a frame forkeeping thatch or reeds away fromtermites, are unfortunately published asspecies-specific terms, all presumably fromcases where the researcher showed a localassistant a plant specimen, asking: ‘What’sthat?’, and expecting a species-specificname in return. Instead, and quitecorrectly, the answer was: ‘that’s aflower…a fruit…a drying frame for reeds’(instead of the local name for the reed

Applied Ethnobotany

Source: adapted from Wilmsen, 1978

Figure 2.3 Seasonal fluctuations in edible wild plants eaten by /ai /ai zu/ oasi San people in thenorth-western Kalahari savanna, southern Africa, showing data from a survey over a year

18

35

30

25

20

15

10

5

0D

Kilograms

J F M A M J J A S O N D

Months

Cucumbers

Beans

Fruits

NutsBerries


species). Similarly, throughout the worldlocal terms for ‘I don’t know’ have beenpublished as species-specific names. Becareful!

Cross-checking

In addition to cross-checking the scientificand folk names of plants through collect-ing voucher specimens, it is important tocross-check information with differentpeople and compare the results fromdifferent methods.

Whatever method or set of methodsyou use, it is important to consider theaccuracy of the responses you receive.How appropriate (or inappropriate) arethe methods and questions? No singlemethod has all the answers – all haveadvantages and disadvantages. Nor canyou always expect the answers you aregiven by local people to reflect yourmeasures of time or quantity. ‘Informant’accuracy, the responses you get from thepeople you interview or discuss thingswith, can vary greatly according to howthey view your intentions. They may alsosee the issue in a different way (seeBernard et al, 1984). If someone says shewalked for six kilometres to fetchfuelwood, then did she? If she reallywalked 3.7km to collect fuelwood, is thisaccurate enough or not? Is our use of‘kilometres’ as a measure of effort appro-priate or not, and if not, what is a morevalid measure? It is crucial to cross-checkinformation from a mix of differentmethods, even if you only compare theresults from just two methods. If everyresearcher did this, there would be farfewer misunderstandings than iscommonly the case.

Five cautionary examples are givenhere, where cross-checking showeddiscrepancies between methods. The firstexample deals with information fromofficial records; the second compares

records from conservation permit datawith field studies of pole-wood cutting;and the third is a comparison of interviewdata with two sets of field data. The fourthexample compares results from PRA andinterview methods, and the final caseshows how and why the role of wild plantfoods in diet were underestimated.

Official data versus other sources

In some cases, long-term data are availablefrom official statistics, such as where thereare commercial sales for export, or whereharvesting is allowed on the basis ofpermits. Scott Mori and Ghillean Prance(1990), for example, were able to obtaindata from government reports for Brazilnuts over more than 50 years(1933–1985). If at all possible, this type ofdata needs cross-checking, since officialstatistics may be rather unreliable.Comparisons of quantities of medicinalPrunus africana bark exported fromCameroon from forestry departmentannual reports showed large discrepanciescompared to primary data cross-checkedfrom weigh-bridge returns from the singlefactory processing the bark (see Figure2.4). The same applies to data fromforestry permits or social surveys, whetherinterviews or PRA approaches.

Conservation department permitrecords versus cut bundles of poles

Working in Hlatikulu forest, South Africa,Dirk Muir (1990) assessed permit dataaccumulated by the conservation depart-ment where local woodcutters specifiedwhich tree species they would harvest forbuilding poles. He compared these dataagainst a large sample of poles already cutand bundled by the same woodcutters(593 poles, 2647 laths). This showed greatdiscrepancies between the two sets of data(see Figure 2.5). Permit data on poleharvesting implied, for example, that 18


19


species were harvested, with the top 10species carrying 90 per cent of the utiliza-tion. By contrast, the cut-wood sampledata showed that 37 species wereharvested, with the top 10 species carrying74 per cent of the utilization. Twenty-threeof these 37 species are canopy tree species;this implied that there would be a long-term impact on the forest canopy fromoverexploitation of poles.

Interviews versus cut bundlesversus cut stumps

Charlie Shackleton’s (1993) study offuelwood use also provides a cautionaryexample for field workers relying on infor-mation from a single method. Shackletonused three different methods (interviews,assessment of fuelwood bundles andidentification of cut stems) to assess whichfuelwood species were avoided orpreferred (see Table 2.1).

PRA versus interview methods

Comparisons are made when different setsof people in the same area give very differ-ent responses to the same method. Thisoften generates more questions thananswers. PRA enthusiasts, for example,like to promote the use of PRA in record-ing quantitative as well qualitative data toexpress the numbers of people, cattle orrelative amounts of resource collected.This may even be used to estimate annualor seasonal consumption rates. Workingin Zimbabwe, Allison Goebel (1996)compared the results of PRA surveys withgroups of people with information fromindividual interviews about the sale ofdifferent plant resources (fuelwood, poles,thatch-grass, herbal medicines, fruits, andwild collected and garden vegetables). Shefound that with the exception of gardenvegetables, the PRA exercise greatlyoverestimated the extent to which local

Applied Ethnobotany

Figure 2.4 Maintain a healthy skepticism and cross-check where possible: discrepancies betweenforestry annual report data and weigh-bridge returns for Prunus africana bark harvests in

South-West Province, Cameroon

20

3500

3000

2500

2000

1500

1000

500

01980/81

Tonnes of bark

81/82 82/83 83/84 84/85 85/86 86/87 87/88 88/89 89/90 90/91 91/92

Financial years

Annual reports: SWP Plantecam weigh-bridge


plant resources were sold. As a result, shesuggested that it is extremely risky toaccept PRA data as quantitatively reliable.

Methods for surveying gatheredplant foods

Food lists and food records are commonlyused for studying patterns of food selectionand consumption. A useful reference on


21

Source: Muir, 1990

Figure 2.5 Comparison of the proportional species of pole harvests at Hlatikulu Forest Reservefrom permit data and counts from a large sample of poles (n = 593) showing the inaccuracy of

responses given in permit forms

Ch

ion

anth

us

fove

ola

ta(i

siN

XIM

AN

E)

Percentage

Permit data Cut-wood data

Do

mb

eya

cym

osa

(iC

IBO

)

Stry

chn

os

hen

nin

gsi

i(u

mN

ON

O)

Stry

chn

os

usu

mb

aren

sis

(um

BA

MB

AN

DO

DA

)

Dry

pet

es g

erra

rdii

(um

HLA

KEL

A)

Prem

na

mo

oie

nsi

s(i

siTH

ATH

ATH

A)

Ptae

roxy

lon

ob

liqu

um

(um

THA

THE)

Du

vern

oia

ad

nat

od

oid

es(u

LOTH

AB

E)

Och

na

arb

ore

a(u

mSH

ELEL

E)

Cry

pto

cary

a w

oo

dii

(uxT

HU

MG

WA

NE)

Ole

a ca

pen

sis

mac

roca

rpa

(um

HLW

ATH

I)

Rin

ore

a an

gu

stif

olia

(isi

HLO

LA)

Sco

lop

ia z

eyh

eri

(isi

NG

ON

GO

NG

O)

Oth

ers

8 sp

p27

sp

p

30

25

20

15

10

5

0


methods for studying people’s diet is thechapter by Darna Dufour and NicoletteTeufel in the excellent book on commonstandards for data collection in studyinghuman societies by Emilio Moran (1995).In a systematic study of the extent to whichwild greens and fruits are used in diet,Anne Fleuret (1979) gathered data amongthe Shambaa people of the UsambaraMountains, Tanzania. Several differentmethods were used, including interviews at40 different periodic markets, to gather thenames of wild foods consumed, combinedwith counts of the wild foods and fruitssold at the markets. In addition, 200women were asked, in open-ended inter-views, to list the wild greens their familiespreferred, and to estimate the frequency oftheir consumption.

The key finding of this study was thatwild plants are important in the local dietand that they contribute significantly tonutrition, even though earlier studies,using the ‘24-hour recall method’, hadimplied otherwise, and had indeed

concluded that the Shambaa diet wasdeficient in many major nutrients. It isparticularly noteworthy that wild foliageplants accounted for some 81 per cent ofinstances of michicha (wild spinach)consumption, being found in 45 per centof all meals eaten. Village surveys involvedhouse visits by local assistants who askedthe householders what was eaten theprevious day. These village surveys wereconducted at three different seasons –April, June and October. Despite employ-ing local assistants to do the villagesurveys, however, they missed fruitconsumption, since fruit is not usually apart of meals; rather, it is an occasionalsnack taken during the day at any time.Observations of children, ranging from 6months to 14 years, revealed that a rangeof fruits was eaten in this way, typicallygathered by the older children as they wentabout other tasks.

Applied Ethnobotany

Table 2.1 A comparison of fuel wood preference or avoidance using three different methods

Data source Preferred species Avoided species

1 Cut stems Combretum collinum Dichrostachys cineraDiospyros mespiliformis Lantana camaraMaytenus senegalensis ‘pooled species’Terminalia sericea

2 Fuelwood bundles Acacia swazica Dichrostachys cineraCombretum collinum Lantana camaraCombretum hereroense Lonchocarpus capassaPeltophorum africanum Maytenus senegalensisTerminalia sericea Strychnos madagascariensis

‘pooled species’3 Interviews Acacia swazica Lonchocarpus capassa

Albizia harveyi Maytenus senegalensisCombretum collinum Pilotstigma thoningiiDalbergia melanoxylon Sclerocarya birreaDichrostachys cineraDiospyros mespiliformisPeltophorum africanumStrychnos spinosaTerminalia sericea

Source: Shackleton, 1993

22


Participatory methods with groupsof people

Participatory methods have become verypopular for use in conservation and ruraldevelopment projects. Many people usingthis manual are probably already familiarwith the terms rapid rural appraisal(RRA), participatory rural appraisal(PRA), participatory assessment, monitor-ing and evaluation (PAME) orparticipatory action research (PAR). Theseapproaches developed from a need toinvolve local communities in analysingtheir own circumstances. Respect for localknowledge and the desire to move awayfrom ‘top-down’ approaches to conserva-tion or development make these attractivemethods to use. Pratt and Loizos (1992)point out the need, however, to maintain ahealthy scepticism and critical view ofprocesses described as ‘participatory’: theword has been used to describe anythingfrom obligatory, through to genuinelydemocratic and enthusiastic involvementin a research project. It is crucial that localparticipation is genuine. It is pointless tobring local people into a data-gatheringexercise which is of no interest to them inan effort to legitimize research through‘participation’.

Very useful and detailed descriptionsof a wide range of participatory methodsare given in The Community Toolbox: theIdeas, Methods and Tools forParticipatory Assessment, Monitoring andEvaluation in Community Forestry (FAO,1990) and in Participatory Learning andAction: a Trainer’s Guide (Pretty et al,1995). Good examples are also given inthe Joint Forest Management manuals(Poffenberger et al, 1992) and in the recentmanual by John Tuxill and Gary Nabhan(1998). The most useful PRA methods forwork at the interface between ethnob-otany and resource management are listedbelow. All of these can stimulate local

insights that may have arisen during infor-mal discussions or during interviewsurveys, but some (the last two methods,in particular) may be too sensitive to usein short-term surveys unless you have agreat deal of local credibility.

Mapping

This includes mapping of land or resourcetenure, resource distribution, and socialmaps showing where different resource-user groups stay in a village, or mappingthe flow of resources after harvesting. Thismethod and some of its disadvantages arecovered in more detail in Chapter 6.

Transect walks

Transect walks combine well with initialethnobotanical surveys and discussions.These are usually done with key infor-mants through the area of interest byasking, observing, identifying differentvegetation types and land-use impacts, andby indicating problems or possiblesolutions.

Time lines

Time lines can lead on from or be devel-oped during transect walks, and identifyimportant historical events which peopleremember. It can be useful to correlatethese with known dates, which are thenrelated to historical trends. Examples ofthis are the ‘stick graphs’ showing trendsin resources and population numbersaround Bwindi Forest from 1940–1990(see Figure 2.6).

Seasonal calendars

Preferences and demand for differentproducts will also change according toseason. Seasonal calendars are a usefulPRA technique, where local seasons formone axis of a matrix and products theother, enabling local people to rank


23


Applied Ethnobotany

Source: Wild, 1996

Figure 2.6 Resource and population trends from ‘stick graphs’ made by a group of people fromthe Nteko area adjacent to Bwindi-Impenetrable National Park, Uganda. Sticks are gathered and,after explanation by a facilitator, broken into lengths representing relative abundance (long sticks)

or scarcity (short sticks) of a resource

24

1940s

• 1940s – people were few;• 1950s – people came from Rubuguli, encouraged by the County chief;• 1960s – immigration from Rubanda;• 1970s – people left the area;• 1980s – immigration from Rwanda and Zaire;• 1990s – more immigrants and families increasing.

(b) Human population

(a) Herbal medicines outside forest reserve

1950s 1960s 1970s 1980s 1990s

Time period

1940s 1950s 1960s 1970s 1980s 1990s

Time period

Rel

ativ

e am

ou

nt

Rel

ativ

e am

ou

nt

• 1940s – no need to go to forest reserve for medicines, as enough on own land;• 1950s – about half amount of 1940s due to land clearance for farming;• 1960s – as people left, herbs on farms increased;• 1970s–1980s – people migrated from Rwanda and they know herbs well;• 1990s – no forest medicinal species left outside forest reserve – they were finished.


harvesting or availability of products byseason.

Matrix ranking and scoringexercises

These can be done on the ground or onpaper, using symbols, picture cards ornames for different resource categories orspecies. The first step is to list thecategories or range of species that areavailable or which you have decided torank. On the basis of their experience inIndia with this method, Poffenberger et al(1992) suggest a listing of 5 to 15 productsfor ranking. Different sized markers, suchas small stones or seeds, are then placedunder each category, building up a matrixof preferences based on different positiveand negative qualities of each species, ormultiple uses of each species or even ofdifferent vegetation types. Examples ofqualities could be flavour of edible plants,durability and regeneration capacity ofbuilding timber, flexibility of differentcraft-work fibres, and so on. Rankingexercises can also list and rank problemswith harvesting or resource availability,and the reasons for these problems.

Venn (or ‘chapatti’) diagrams

Understanding the relationships betweendifferent institutions and how communitymembers relate to them is an importantissue in resource management (see Chapter7). Venn diagrams can give very interest-ing (and counter-intuitive) insights. Thepower of a local chief may be far less influ-ential than one is led to believe, forexample. Circles are used to representpeople, groups and institutions. These arearranged by local participants to showtheir perceptions of overlap. Lines can bedrawn between the different circles, withthick lines showing strong relationships,or thinner lines weaker ones.

Wealth ranking

Where households are listed, their namesare written on separate cards; then localresearch participants can be identified.After discussion with each person aboutlocal perceptions of wealth, the cardsrepresenting each household are sortedinto piles or wealth classes. Discussionsthen revolve around the main aspects ofeach household’s livelihood strategy andthe differences between the differentwealth classes. This can give useful insightinto which groups rely most on (orcontrol) plant resource use.

Village mapping

Participants are invited to draw a sketchmap of their neighbourhood, either on theground using stones, tins, etc as symbols,or on paper. It can be useful to divideparticipants into groups, depending upontheir gender or background. Maps varyaccording to perceptions and knowledge,and are a useful discussion point on howlong people have been in an area, wherethey came from and what they do for aliving.

Matrix ranking data can also be cross-checked or compared with informationfrom field observations, discussions,market surveys or social survey methodsthat may have been carried out in similarvegetation or social situations. It is oftenwise to work with a homogeneous groupselected on the basis of their interest in,and knowledge of, that particular resourcecategory, such as groups of woodcuttersor carvers (generally men) who haveknowledge of building or carving timber,the production of fuelwood or edible wildgreens, and the women who gather theseresources. Herbalists or midwives can beconsulted on traditional medicines. Inother cases, it can be very useful tocompare the insights of different usergroups within or between communities or


25


vegetation types. Men, women andchildren, for example, may have verydifferent insights on the values of ediblefruits, with children eating a much widerrange of species as ‘snacks’ than adults, orwomen who gather ‘top of the range’species most favoured for size, flavour andabundance for home consumption.

Individual interview surveys

Depending on time, funding and the aimsof the research study, individual interviewsmay be case studies, where the researchergains insights from discussions withindividuals who ‘typify’ a particular situa-tion. These do not yield quantitative data,but can give detailed insights into people–plant relationships.

Alternatively, discussions could bewith key research participants, who havedetailed knowledge of the topic of theresearch. Nichols (1991) distinguishes casestudies and key participant interviewsfrom a third category, individual indepthinterviews, as the former have a widerscope and are more open-ended in natureand do not follow a strictly set pattern. Allthree of these approaches are importantexercises to undertake before designing theformat for a larger-scale structured inter-view survey, just as the inventory phase isimportant in recording local vernacularnames and uses of harvested plants.Individual indepth interviews prior tolarger-scale structured social surveys maybe unstructured or semi-structured:

‘In an unstructured interview, theperson interviewed is free to voicetheir own concerns, and to sharein directing the flow of conversa-tion. The interviewer relies onopen questions to introduce topicsof interest. The aim is, literally, an“inter-view”: a mutual explo-ration of the issues, without the

researcher imposing his or herideas. In a semi-structured inter-view, the researcher has a preparedlist of topics – though still not aseries of questions. Interviewersdeal with the topics in any order,and phrase questions as they thinkbest in the circumstances’(Nichols, 1991).

In individual semi-structured interviews ateach homestead, for example, a womancould be asked which edible wild greensshe gathered in different seasons, leavingher to list the species she collects (by localname). In a study on edible wild foods inSwaziland (Ogle and Grivetti, 1985), forexample, interviews were done by sevensiSwati-speaking students studying homeeconomics, using a questionnaire whichtook 40 to 60 minutes to complete. Adultsand children from four different ecologicalzones were interviewed separately in semi-structured interviews, which askedrespondents first to identify species avail-able near their homesteads, then howfrequently each of these was consumed(see Figure 2.7).

Food lists, such as the 24-hour recallmethod, are commonly used in dietarysurveys (Dufour and Teufel, 1995). Foodlists take less time than food records,which require descriptions of the foodeaten at the time of consumption, withrecords made either by an outside observeror by a trained household member. Foodlists may be unreliable, however, largelybecause they depend upon memory. Theyinvolve asking people to recall the typesand amounts of food consumed over aparticular time period, often 24 hours.Interviews are best conducted in the areawhere the food is prepared or consumed,and all items, including drinks, noted. Acommon problem arises from the fact thatparticipants may be reluctant to admit topoor feeding habits or low-standard

Applied Ethnobotany

26


meals. The presence of visitors (includingresearchers!) may prompt the preparationof special, untypical food, or food may betaken from a shared vessel, making theassessment of amounts consumed difficult.Like other methods, this is worth cross-checking against other methods, as AnneFleuret (1979) did in her study of the

Shambaa people’s diet in Tanzania.Structured interview surveys can be

appropriate survey methods in some situa-tions but not in others, where informationcan be more efficiently gathered throughfield and participant observation, discus-sions, RRA/PRA methods or by marketsurveys. Interview surveys use a carefully


Source: Ogle and Grivetti, 1985

Figure 2.7 Reported frequency of consumption by 211 adults of 47 edible wild greens

27

Percentage100

80

60

40

20

0

Bid

ens

(2 s

pp

)

Co

rch

oru

s (3

sp

p)

Am

aran

thu

s (1

sp

)

Alt

ern

anth

era

(1 s

p)

Am

aran

thu

s (2

sp

p)

Sola

nu

m (

2 sp

p)

Cic

ho

riu

m (

1 sp

), L

actu

ca (

1 sp

)So

nch

us

(1 s

p)

Mo

rmo

rdic

a (2

sp

p)

Mo

rmo

rdic

a (1

sp

)

Oxa

lis (

2 sp

p)

Asc

lep

ias

(1 s

p)

Xys

mal

ob

ium

(2

spp

)

Gre

wia

(3

spp

)

Port

ula

ca (

1 sp

p)

Oxa

lis (

2 sp

p)

Ru

mex

(1

sp)

An

nes

orh

iza

(2 s

pp

)Pe

uce

dan

um

(1

sp)

Port

ula

ca (

1 sp

)

Lap

ort

ea (

1 sp

)

Ch

eno

po

diu

m (

1 sp

)

Spar

man

nia

(1

sp)

Op

hio

glo

ssu

m (

1 sp

)

Asc

lep

ias

(3 s

pp

)X

ysm

alo

biu

m (

1 sp

)

Am

aran

thu

s (1

sp

)

Rio

creu

xia

(2 s

pp

)

Co

loca

sia

(1 s

p)

Zan

ted

esch

ia (

1 sp

)

Asc

lep

ias

(1 s

p)


designed questionnaire and aim methodi-cally to get typical and reliableinformation from a selected sample ofpeople. Interview surveys may not yieldreliable results when covering highly sensi-tive and very personal issues such assexual, specialist medicinal or illegalharvesting activities, unless the intervieweris highly skilled and has a great degree oflocal credibility and trust. They can bevery useful, however, in quantitativesurveys of relatively large numbers ofpeople on topics which are less controver-sial and in which there is local interest; inscenarios such as this, people often giveclear answers. Design, field testing andsample techniques in questionnairesurveys are discussed in detail by PaulNichols (1991) and Brian Pratt and PeterLoizos (1992), and researchers are encour-aged to look at these two most usefulOXFAM manuals. It will also be useful tofollow the checklist covering a sequence ofevents leading to social surveys usingquestionnaires (see Box 2.1).Questionnaire surveys are generallyneither quick nor low-cost, so it is impor-tant that they are carefully designed.

Participant observation

Participant observation is an approachcommonly used by anthropologists,sociologists and ethnobiologists whochoose to live within the communitywhere their research is taking place, partic-ipating in local events, includingharvesting of natural resources.Participant observation usually takes placeover a relatively long period of time,enabling detailed observations to be madeand informal discussions held on the topicof the research. In common with anymethod, it has advantages and disadvan-tages that need to be weighed up on acase-by-case basis.

A long research period enables

researchers to record detailed informationon social and ecological issues, includingseasonal changes in harvesting, thus avoid-ing the seasonal bias of short-term work.If carefully planned on the basis of prelim-inary studies, it can also incorporatestudies on the response of selected plantspecies under different harvesting regimeswhere local villagers or resource users maybe involved in harvesting experiments.Furthermore, it facilitates research onmore sensitive topics; researchers will havethe opportunity to develop local credibil-ity – for example, when detailedmeasurements of quantities such as dietaryintake are required. Nevertheless, there areseveral disadvantages.

• The long research period is slow, andoften focused on a single community,missing out broader scale issues; itrequires extensive research trainingand often the need to learn the locallanguage. As a result, it may take yearsbefore results are available.

• Few researchers have both the anthro-pological and ecological trainingrequired for participant observationstudies relating to natural resourcemanagement. Consequently, theresults of the study can be biasedtowards one academic discipline oranother.

• It is difficult for another researcher tocross-check information due to differ-ences in timing of the research anddifferences in the network of researchcontacts (Pratt and Loizos, 1992).

For these reasons, participant observationstudies on resource management in thepeople/protected areas situation wouldeither be part of, or develop from, post-graduate academic study. In cases whereanswers are needed for urgent resourcemanagement problems, the time andexpense required for participant observa-

Applied Ethnobotany

28



29

BOX 2.1 STEPS IN QUESTIONNAIRE DESIGN AND IMPLEMENTATION

Questionnaire Construction: Content and Form

The content of the questionnaire is determined by what information is sought andfrom whom. What is the best way to elicit answers? Aspects for consideration includethe following.

• Question construction: does the question pinpoint the issue? Will its meaning beclear to all respondents (see Nichols, 1991; Pratt and Loizos, 1992)?

• Length of the questionnaire: although this depends upon the type of questionnaireand the skill of the interviewer, many people feel that an hour is the maximumamount of time that interviewees are prepared to spend answering questions.

• Language of the questionnaire: it is usually desirable to have interviewers speakthe same language as the interviewee, rather than work through an interpreter;however, this may be unavoidable in some cases. The questionnaire should prefer-ably be in the language of the interview to ensure as much consistency as possibleamong interviewers. It is also preferable to have the answers recorded in thelanguage in which they will be analysed. Decisions on these issues need to be madein terms of what local information is sought, the number of interviews and thelinguistic abilities of the interviewers.

• Amount of interview coding required of interviewers during interviews: unless verybasic, all coding requiring any discrimination or decision making should be leftuntil processing of the questionnaire data occurs. The issue of open-endedquestions needs to be carefully taken into account in cases where the researcher isunable to anticipate likely responses or is trying to raise new issues.

• Clarity and simplicity of layout: the layout of questions and responses should enablesmooth, efficient progress through the questionnaire.

• Pilot testing: this should be completed, and all major alterations made before inter-viewer training. The training period then provides the opportunity fordouble-checking on the effectiveness of questions when asked by interviewers andto check translations.

Questionnaire Implementation: Interviewers and Interviewing

Aspects to consider are the following.

• Type of interview: will the interviews be structured, providing a precise list ofquestions to the interviewer, or semi-structured, listing the general agenda of issuesto be covered? Structured interviews are necessary if the project requires a repeat-able, uniform approach, and if it uses several interviewers over a relatively largesample population.

• Who and when to interview: choice of ‘unit of analysis’. Does the informationrequired relate to the individual, family, household or the community? Who is thebest person to respond to questions in the survey? When would it be most conve-nient to be interviewed?

• Establishing an appropriate sample frame: this is very project-specific and you mayneed to get statistical advice. Nichols (1991) gives a good introduction to sampleselection, which may be random, systematic, stratified or clustered, dependingupon the type of project and resources you are dealing with.


tion can be a major problem. For thisreason, it may be possible to compromiseby working with the same community, bymaking a series of short-term visits, andby participating in local events andharvesting expeditions in appropriatelytimed visits over a number of years.

Ethnobotanical inventories

Inventory of plant and animal species is acommon basic step in field surveys,whether this includes all species or islimited to identifying useful or uniquespecies. Collecting plant specimens is animportant step in this process so that

voucher specimens can be identified byscientific as well as local names. If at allpossible, it is important to collect goodquality herbarium specimens which notonly have leaves and stems, but alsoflowers and/or fruits and, where necessary,samples of bark, wood or the roots orbulbs that characterize the species. Thesehave to be well preserved, and accompa-nied by detailed field notes on thecollection locality, characteristics of theplant, its local uses and vernacular namesand their meanings.

Duplicate specimens need to becollected and, as voucher specimens,should be located in a recognized herbar-

Applied Ethnobotany

30

• Selection, training and supervision of interviewers:1 Selection: this will depend upon: (a) the local acceptability and previous experi-

ence of potential interviewers; (b) the attitudes, empathy and sensitivity of theinterviewers; (c) linguistic ability; (d) literacy and numeracy. It may be better towork with a smaller number of skilled interviewers than with a larger numberof less skilled ones – but if so, beware of the problems of bias. Ideally, finalselection of interviewers should be made after an initial training period hasenabled potential interviewers to learn exactly what the project requires andthe project manager/researcher has had the opportunity to observe potentialinterviewers during training.

2 Training: all questionnaires differ, and thorough training is usually required.Resist the urge to move into the field too quickly. Training should includedetailed discussions of each question and/or probe question. This can also serveas an opportunity to double-check wording and meanings of questions. Makesure that interviewers have a clear understanding of administrative details:expected work schedules, amount and frequency of pay, transport arrange-ments, anticipated length of interviews.

3 Supervision: this is essential. It is also time consuming. Cross-checking ofcompleted questionnaires should not be delayed, so that incomplete or incon-sistent information can be corrected when necessary.

Computer Processing of Questionnaire Responses

The availability of good statistical computer packages and portable computers can facil-itate faster processing of data. If researchers are not familiar with a particular computerprogram, they need to get advice from the first stage of questionnaire construction.They also need to consider: keeping a code book which shows a number or letter foreach possible answer; and checking that the computer can read the codes used. Poorclassification of responses can have a disastrous influence on survey results.

Sources: Prinsloo, 1982, with reference to Nichols, 1991; Pratt and Loizos, 1992



31

BOX 2.2 COLLECTING PLANT SPECIMENS: FIVE IMPORTANT

REMINDERS

1 Specimen Size and Quantity

Make sure you collect plant samples that are the right size – enough to fit across astandard newspaper page (30cm wide x 45cm long) – and that there is enough mater-ial for duplicates, so that a specimen is retained in the national herbarium. If thematerial is unusual, or from an undercollected region, it can be sent for identificationto a regional herbarium and to an international herbarium. Four specimens may berequired from an unusual plant species collected, for example, in western Uganda: onefor the field-station herbarium, one for the national herbarium, one for the East AfricanHerbarium in Nairobi and one for a large herbarium which specializes in African plants,such as those at either Missouri or the Royal Botanical Gardens, Kew.

2 Quality

Ideally, your specimens should be ‘fertile’ (with flowers and fruits), rather than being‘sterile’ – material consisting of leaves and twigs or underground plant parts. In somecases, however, ethnobotanists (and ecologists), no matter how hard they look forfertile specimens, are faced with the necessity of collecting sterile specimens or none atall. Although this drives many taxonomists to distraction, if you have no choice, it isbetter to collect good sterile material than not to collect any specimens at all. It isessential that the specimens are pressed properly, each carefully flattened on individual‘flimsies’– the sheets of paper that support the specimen when you regularly changethe drying papers. Particular care should be taken with certain specimens. Flowers ofplant species with very delicate flowers (such as in the Aristolochiaceae, Asclepiadaceae,Balsaminaceae, Bignoniaceae, Commelinaceae, Curcurbitaceae, Passifloraceae andScrophulariaceae) should also be preserved in spirit preservative such as 50 to 79 percent ethanol, FAA (formalin-acetic acid alcohol) or ‘Kew cocktail’ (Bridson and Forman,1992). Specimens of plants such as fig species (Moraceae) and succulents (Cactaceae,Euphorbiaceae and Aloe species), which grow well from cuttings and are toughsurvivors, should be killed off quickly by immersing them in boiling water or preserva-tive (ethanol or even petrol).

3 Thorough Documentation

Specimens without thorough documentation have little value. Field notes on character-istics of the plant such as smell, sap colour and bark slash can be very important inidentifying sterile material. In addition to information required for standard botanicalspecimens – collection locality (ideally with latitude and longitude coordinates); colourof fresh flowers/fruit/sap/leaves; ecological notes on habitat; collector’s name; collec-tion number and date of collection – ethnobotanical collections should includeinformation on the vernacular name(s) of the species, their meaning, part(s) used,method of preparation and whether commercially traded or not, and the name of theperson who supplied the information. In addition to writing this information on labels,you also need to maintain a field notebook where these records are written alongsideyour collection number for each specimen.


ium. Three recent manuals give detailedinformation on how to collect, preserveand label herbarium specimens and Irecommend that you read at least one ofthese: Diane Bridson and LeonardForman’s The Herbarium Handbook(1992), Miguel Alexiades’s SelectedGuidelines for Ethnobotanists (1996) orGary Martin’s Ethnobotany: a MethodsManual (1995). For those who do nothave access to any of these manuals, I havesummarized five important points to

remember when collecting plant specimens(Box 2.2). I then give more detail on twoimportant, linked, aspects of ethnobotani-cal inventory work. Firstly, there is theneed to document and use a wider rangeof field characters than are normally usedby formally trained botanists; andsecondly, it is important to be aware ofboth the potential and the pitfalls of folktaxonomy and terminology in field workwith local resource users.

Applied Ethnobotany

Taxonomy with all your senses: the use of fieldcharacters

32

4 Proper Curation

When good quality specimens have been identified, they are poisoned, mounted andare usually placed in a recognized herbarium where they are accessible to otherresearchers. The location of your voucher specimen, indicated by an internationallyused list of herbarium codes, the Index Herbariorum (Holmgren, 1990), should berecorded in scientific publications in addition to the specimen collection number.

5 Sensitivity to Conservation and Cultural Concerns

Although collecting duplicate specimens and accurately recording localities, plant uses,methods of preparation and the name of the local person providing the information isthe ideal, you need to be sensitive to conservation and cultural issues. If a plant is veryrare and the population small, you have to be aware of the dangers posed by overcol-lecting and drawing the attention of commercial collectors to the site by giving localityinformation that is too precise. You should be equally sensitive to concerns the localpeople may have regarding religious aspects of plant uses or fears about commercial-ization of the information they provide. These concerns need to be taken into account,and the ethical guidelines proposed by professional organizations such as the Societyfor Economic Botany and the International Society for Ethnobiology should be followedby researchers (Cunningham, 1993,1996).

Although plant taxonomists rely stronglyon flower or fruit characters, it is veryuseful to be able to describe and use thecharacteristics of sterile material as an aidto identification. As this manual is for fieldresearchers, I have concentrated on describ-ing macroscopic characters that you willfind useful, not microscopic ones. Many

field workers in the tropics and subtropicswill already be familiar with some of theexcellent field guides based on the vegeta-tive characteristics of woody plants, such asthose by Al Gentry (1993) for north-westSouth America, Eugene Moll’s (1981) guideto 700 tree species in KwaZulu/Natal,South Africa, or Alan Hamilton’s (1983)


guide to trees of Uganda. Although plantgrowth form and leaf characteristics suchas simple or compound leaves, arrangedalternately or opposite, are a basis of fieldidentification, so too are other vegetativecharacteristics. Learning about the vegeta-tive characteristics of plants also makesfield work more interesting and enablesfield workers to recognize many plantfamilies from the combination of three orfour characteristics. This can be of practicalvalue when identifying the family or genusof a species from sterile material –something which led the late Al Gentry(1993), a superb botanist and field workerin the most diverse tropical forest area ofthe world, to observe that:

‘Most neotropical plants aresurprisingly easy to identify tofamily, even in sterile condition.Indeed, in many ways it is proba-bly easier to identify woodytropical plants in sterile conditionto family than it is identify thefertile material to which manysystematic botanists tend torestrict themselves. This is trueboth because of the strong conver-gence by many different familiesthat share a common disperser orpollinator, and because the techni-cal characters on which plantfamilies are defined are so obscureand esoteric (typically involving adetermination of ovule number,placement and orientation) thatthey are of limited practicaluse…vegetative characters, on theother hand, are always available,mostly macroscopically obvious,and at least in the rainforest,apparently have been subjected tomuch less of the kind of conver-gence-inducing selection ontaxonomically useful charactersthan have flowers and fruits.’

The same applies in other tropical forestareas. For these reasons, it is very impor-tant to use all your senses to record fieldcharacters based on vegetative criteriasuch as smell, texture, sap colour, skinirritant qualities or taste. It is also usefulto describe characteristics of fresh anddried plant material to assist identifica-tion. In some cases, it can be useful toconstruct your own key for the mostcommonly used species, based on bark,bulb, root or wood characters. Part of akey to medicinal bulbs commonly sold insouthern Africa is given as an example inBox 2.4.

Local people’s knowledge is an impor-tant guide to these characteristics. Incontrast with most taxonomists, whousually concentrate on dried specimens ofleaves, flowers and fruits in herbaria, localpeople harvest and work with live, wholeplants through different seasons. Theyconsequently have the opportunity toperceive important characteristics of theplants, other than those commonly usedby taxonomists. These are very useful torecord during field work in addition tocollecting voucher specimens, and thereare a number of other reasons for this.Firstly, as discussed earlier, it may be diffi-cult to obtain fertile plant specimens thatbear flowers or fruits, or sometimes evenleaves of a particular plant. They may beinaccessible, such as on the top branchesof rainforest trees. Alternatively, you maybe working during a time of the year whenno leaves, flowers or fruits are available,such as during the dry season in aridzones, deserts and savanna, or in the coldseason of alpine or arctic sites or temper-ate woodland. Similarly, in village andurban markets, medicinal plants andchewing sticks are often sold without anyleaves, flowers or fruits attached.

Wood anatomists are an exception tothe taxonomist’s normal focus on herbar-ium specimens which consist of leaves,


33


flowers and fruits. Their work provides anoutstanding example of how macroscopiccharacteristics of wood (which can be seenwith the naked eye or with a ‘x 10’ handlens), well known to local resource users,can best be combined with microscopiccharacters to form definitive keys to woodidentification (Gregory, 1980; IAWA,1981, 1989; Miller, 1981). At an earlystage, however, it is possible that systemsof bark, root and bulb identification couldsimilarly combine the best of indigenousand formal scientific approaches, usingmacro and microscopic characters todevelop identification keys similar to thosedeveloped for wood identification.

Local people have an excellent knowl-edge of bark, root or bulb characters andmake slashes in bark or roots with amachete to determine the identity of foresttrees, rather than use leaf or flower charac-teristics (see Figure 2.8). Some of these areso characteristic that they are referred toin the local names for that species. InZulu, for example, two Afromontane treesin the Rutaceae, Clausena anisata andZanthoxylum capense, are called respec-tively umnukambiba (‘smells like a stripedfield mouse’) since its crushed leaves smelllike mouse urine, and umnungamabele,since the knobs on its trunk are shapedlike breasts. Identification of species by afragment of bark, roots or stem on thebasis of a combination of scent, sap,colour or texture has its parallels in urbanindustrial society. People employed as‘noses’ by perfume companies, forexample, can identify a single perfumevariety from hundreds of others. Similarly,‘wine tasters’ are able to identify the originand year of production of a particularvariety of wine. Descriptions of the smellof bark, roots, wood or leaves of differenttree species are reminiscent of the way inwhich wine varieties are described.

Be careful to record whether these arecharacters of fresh or dried bark, roots,

wood or leaves, as some features charac-terize dry rather than fresh material. Theshiny calcium oxalate crystals are best seenin the broken cross-section of dried ratherthan fresh Bersama (Melianthaceae) bark,for example, and the leaves of pressedspecimens in the Scrophulariaceae,Loranthaceae and Olacaceae often turnblack (sometimes olive) only when theydry out. Examples of field characters youneed to look out for in bark, bulb, rootand wood identification are listed below.

Colour of roots, bark and wood

This can be a useful first step in identify-ing unknown samples of harvested plants.Wood colour characteristics are welldocumented, with characteristics of rootsand bark better known by ‘undergoundbotanists’ – herbalists and hunter-gather-ers. The roots of many Celastraceae(Maytenus, Pleurostylia, Salacia), forexample, are covered in orange, flaky‘root-bark’, as are the roots of Cardiogyneafricana (Moraceae). The roots of manyEbenaceae, such as Euclea and Diospyrosspecies used for dyes and ‘chewing sticks’(traditional toothbrushes), are character-ized by an almost black outer root-barkand a yellow or orange cross-section in theinner root. The unusual colours of thecross-section of roots of parasitic plantssuch as Hydnora (Hydnoraceae, pink) andCycnium (Scrophulariaceae, purple) areuseful guides when all you see on herbaltraders’ shelves are root material. You alsoneed to take note of colour changes, asexudates or inner roots oxidize onexposure to air. As an aid to identification,for example, Tembe-Thonga herbalistssniff the roots of Albertisia delagoense(Menispermaceae), which are used to treattoothache, and watch as the inner rootcolour darkens to a yellow-brown whentwisted open.

Applied Ethnobotany

34



35

Figure 2.8 Examples of bark characteristics. (a) Smooth pale inner bark of Cassine transvaalensis(Celastraceae). (b) Bright-yellow bark slash (Enantia chlorantha (Annonaceae)). (c) Ndumbe Paulchecks the onion-like smell of a bark slash on Afrostyrax lepidophyllus (Huaceae). (d) ‘Warts’ on

the bark of older Ocotea bullata (Lauraceae) trees (arrow), but not the smooth younger bark(centre), with both characterized by the pig-dung aroma when the bark is broken and the smooth

chocolate-brown inner bark (left). (e) Two species harvested for commercial export as one:Pausinystalia macroseras bark, characterized by extensive lichen growth (arrow (i)) distinguished

from Paunsinystalia johimbe bark (arrow (ii)), with limited lichen growth (both Rubiaceae). ThreeMyrtaceae, different bark texture: (f) Smooth bark (Eucalyptus citriodora); (g) Loose bark flakes

(Eucalyptus leucoxylon); (h) Irregular papery bark flakes (Melaleuca quinquinerva)


Scent

In your field notes, think of the best wayof describing the scent of the bark, leaves,roots or wood you are examining. If youhave the opportunity, crush or cut them toget a better scent. Some will already befamiliar. Crushed leaves of many Fabaceaesmell like green beans and the leaves ofmany African Myrtaceae like guava fruits.The bark of several Croton species(Euphorbiaceae), for example, smells likepepper; Prunus africana (Rosaceae) barkand leaves like almonds (or cyanide); theroots of some Polygalaceae, such asSecuridaca longipedunculata in Africa orPolygala paniculata in Fiji, smell likemethyl salicylate (or ‘wintergreen’ointment); Ocotea bullata (Lauraceae)bark smells like pig-dung; Clausena anisataleaves smell like striped field-mouse urine;Maesopsis eminii (Rhamnaceae) barksmells like cold, cooked chicken; whileOlinia usambarensis (Oliniaceae) bark hasa burned smell.

Some woods also have a characteristicfragrance, particularly trees in theLauraceae (Aniba, Cinnamomum,Ocotea), Santalaceae (Santalum) and thegenera Spirostachys (Euphorbiaceae),Cedrela (Meliaceae) and Viburnum(Caprifoliaceae). The most familiar ofthese are the smell of camphor (fromCinnamomum camphora), cinnamon(from C. zeylanicum, C. aromaticum andothers) and the incense smell of Santalumspecies. These characteristics are oftenencountered in the roots and leaves aswell. In some cases they are absent fromthe fresh leaves, for example in Mondiawhitei (Periplocaceae), a commonly usedappetizer and aphrodisiac root which hasa ‘fresh’ smell of cinnamon in the roots butnot in the leaves. As in any field work,however, you need to be cautious. Youwouldn’t want to crush Mucuna(Fabaceae) leaves or Euphorbia stems to

smell them (and get covered in irritanthairs or toxic sap). Based on his work inSouth American forests, Al Gentry empha-sized the need to take care when you smellanise (liquorice), as this can be the smellof a liverwort which lives on tree leaves,rather than the tree itself; so cross-check.

Texture

The texture of bark, roots, corms, bulbs orwood, when combined with other charac-ters, can be a useful guide to identification,whether you are working in local marketswhich sell medicinal plants or in tropicalforest where the trees are so tall you cannotcollect any leaves. For many tropicalbotanists and foresters, wood characteris-tics and bark texture, combined with thecolour, odour and exudate from smallblazes (slashes) in the bark are very impor-tant means of identifying tall forest trees(see Figure 2.8, the following section ondescribing bark characters and Box 2.7).For this reason, bark characteristics forman important component of field guidessuch as Polak’s (1992) guide to timber treesof Guyana, Whitmore’s (1962) studies ofthe Dipterocarpaceae, or Tailfer’s (1989)guide to trees in tropical Africa. Nobodywould fail to be impressed by a master fieldworker, whether forester or herbalist, whois dwarfed by a towering forest tree andmakes an accurate identification on thebasis of a quick slash, a sniff and a pauseto look at colour and exudate. It is impor-tant to keep bark slashes to a minimumand avoid them where possible, however,as some rare forest trees in the Proteaceaeand Podocarpaceae, such as the SouthAfrican endemic Faurea macnaughtonii,are very susceptible to fungal attackfollowing a deep bark slash.

Sound

Even your hearing can be an aid to identi-fication. There is an audible ‘squeaking’

Applied Ethnobotany

36


when Parinari excelsa (Chrysobalanaceae)bark is slashed, as air is drawn into thexylem tubes, apparently a characteristic ofsome other tropical tree species as well. InBurma, the resin-impregnated bark ofCanarium resiniferum (Burseraceae) is sohard that it makes a ringing sound whenhit with the back of a machete.

Taste

Many people would be familiar with thetaste of spices such as cinnamon bark orthe bitter flavour of quinine (Cinchonabark). Bark can also have a hot, pepperyflavour (Warburgia species, Canellaceae)or taste sweet and aromatic (root bark ofthe climber Mondia whitei). Some plantsare toxic and/or taste absolutely awful –so be careful.

Exudates

Certain families are characterized by theexudates that seep from the inner bark orleaves when they are damaged. Milky latexis well known as a character of the bark,leaves and roots of many members of theApocynaceae, Asclepiadaceae, Caricaceae,Euphorbiaceae, Moraceae and Sapotaceaefamily. Also watch for any colour changewhen the exudate is exposed to air. Thewhite exudate of Trilepsium madagas-cariense (Moraceae), for example, oxidizesfrom white to a russet brown after a fewminutes. Bright orange or yellow latex iscommon in the Guttiferae family, red sap acharacteristic of Pterocarpus species(Fabacaeae) and many Myristicaceae.Exudates are a feature of fresh rather thanthe dry plant material that you wouldencounter when working with a herbalistor during a market survey. They may bestill visible, however, in resin canals orcongealed lumps. While these will havechanged colour, they can still be a usefulclue in the combination of characters thathelp identify a specimen to family or genus.

The elastic threads that form whenlatex-containing roots, leaves or bark arebroken and gently pulled apart are usedby botanists and local herbalists as usefulfield characters. Elastic threads character-ize some roots (such as Asclepiascucculata, which is used as a love charmin Zulu traditional medicine), bulbs (suchas Crinum species) or in bark and leaves(such as Maytenus acuminata). Plantexudates are typically classified as sap,gum, latex and resins, but this can beconfusing, as it is sometimes difficult totell what is sap, resin or latex. For thisreason, it is useful to describe exudates bytheir physical characteristics and source(see Box 2.3).

Ash and charcoal

If you think of the hundreds of species andtonnes of trees burned by local people eachyear in household fires, or when clearingforest or woodland for swidden agricul-ture, it is not surprising that insightfullocal people are familiar with the ash orcharcoal characteristics of particularwoody plants. Wood anatomists also useash and charcoal characteristics. They firstcheck whether a wood splinter burns toash, or whether it burns to charcoalinstead (a characteristic of Shoreanegrosensis (Dipterocarpaceae) wood.Then they look at the colour of the ash orcharcoal: is it grey or white; black; brown;or none of these options (Miller, 1981)?Just as local names for plants often saysomething about their colour or smell, sonames can reflect their ash characteristics.A southern African example is the treeAntidesma venosum (Euphorbiaceae),known in Zulu as isibangamlotha, whichliterally means ‘ash-causer’ due to thequantity of white-coloured ash itproduces. This local knowledge needs tobe better documented and used by ethno-botanists.


37


Applied Ethnobotany

38

BOX 2.3 PLANT EXUDATES: STANDARDIZING BOTANICAL

DESCRIPTIONS

Plant exudates are typically classified by botanists either as sap, gum, latex or resins,which can be confusing. For practical purposes, it is better to use Leo Junikka’s (1994)system of describing exudates by their physical characters. It is best to break off freshleaf material and check the petiole for the presence or absence of exudates. This avoidshaving to make a bark slash and wait for the exudate to be produced. If you make abark slash, however, look to see whether the blaze (slash) is dry (no exudate and feelsdry when touched) or wet (slight exudation and feels moist). Is the exudation abundant(profuse for a while) or scanty (some families or individual plants supposedly character-ized by exudate only produce minute amounts)? Also be aware that it can take severalminutes for the exudate to appear.

The rate of exudation also depends on the season. Exudates may be clear (trans-parent) or opaque, coloured (white, yellow, golden, red, brown, blackish), maydiscolour within a few minutes on contact with air, be frothy (forming a foam whenyou rub it with your fingers), liquid (flowing readily, often transparent) or viscous(flowing slowly, but not necessarily sticky), sticky (sticking to your fingers) or non-sticky.Does the exudate have a smell or not? Although deciding whether something smellspleasant or not can be subjective (do you like the smell of garlic or not?), you can groupexudates according to whether they are odorous (smelling pleasant, like incense) orwhether they are smelly (smelling like excrement, urine or rotten eggs).

Source: Balle and Daly, 1990

The hierarchical system of exudate classification used by the Ka’apor Indians of the BrazilianAmazon. Plant families represented are the Burseraceae (Protium and Trattinnickia),

Sapotaceae (Manilkara), Guttiferae (Symphonia) and Fabaceae (Hymenaea)

‘Plant parts’ (covert)

Hik(‘resin, sap, latex’)

Kanei (covert)

Rank 6

Rank 5

Rank 4

Rank 3

Rank 2

Rank 1

‘Plant fluids’ (covert)Ho (leaf)

Kanei hik(Protium)

Kaneiape hik(Protium spp)

Yeta-hik(Hymenaea)

Trapai-hik(Hymenaea)

Irati-hik(Symphonia)

Kirihu-hik(Trattinnickia)

Irikiwa-hik(Manilkara)

Kanei-aka-hik(P. sagotianum)

Ara-kanei-hik(P. altsonii)

Kanei-pitaq-hik(P. decandrum)

Kanei-tuwir-hik(Protium spp)

Tikwer(‘watery plant fluid’)

Non-kanei (covert)

Hapo (root)


Crystals

Wood anatomists commonly use thepresence and type of calcium crystals, silicabodies and cystoliths (calcium carbonatedeposits) as important microscopic charac-teristics of wood (IAWA, 1981). Prismaticcrystals are common in the wood ofTerminalia (Combretaceae) and Diospyros(Ebenaceae), for example, but are usuallyabsent from Dipterocarpus, Betula andTilia wood. Cystoliths have only beenfound in wood of Opiliaceae,Sparattanthelium (Hernandiaceae) andTrichanthera (Acanthaceae) (Miller, 1981).While these microscopic characters do nothelp a field botanist with just a hand lens,shiny calcium oxalate crystals can be seenwith the naked eye in some cases. Thepresence of calcium oxalate crystals, forexample, is used by herbalists to identifythe bark of medicinal species being sold inurban African markets. Calcium oxalatecrystals are a useful character that are bestseen glinting in the broken cross-section ofdry bark in full sunshine. This helps todistinguish the crystal-packed bark of treessuch as Cassine papillosa (Celastraceae)and Bersama species (Melianthaceae) frombark with a similar thickness and colour.

Describing bark characters

Whether you are working with people inthe field and observing fresh bark or are

observing dried bark collected by herbal-ists or bought in village markets, it isuseful to record bark characters that willaid identification at a later stage. Whilewood anatomists clarified the terminologyused in wood identification (IAWA, 1957,1989), the terms used to describe macro-scopic features of bark were notstandardized and were confusing. This hasbeen corrected in a recent publication byLeo Junikka (1994), a Finnish botanist, onthe basis of his field experience in South-East Asia and an extensive literaturereview.

Very few herbarium specimens recorddetails of bark characteristics.Ethnobotanists have a great opportunityto change this, since they work with localpeople who are not only knowledgeableabout bark, root or bulb characters, butwho often have insightful folk classifica-tion systems for bark texture or characterssuch as exudates. Because of the value tofield workers of a standard system fordescribing macroscopic bark morphology,the terms Junikka (1994) suggests aresummarized in Box 2.7. I would alsorecommend his publication for additionalreading.

His first step was to standardize theterms used to describe the main compo-nents of bark, and to clarify terms such ascork (the trade product from cork oak –Quercus suber), bast (any fibres from theouter part of the plant, but mainly from


39

Folk classification of exudates can also be very detailed (see the figure above). In theirstudy amongst the Ka’apor Indians of the Brazilian Amazon, for example, Bill Balle andDoug Daly (1990) found that although the Ka’apor had a detailed hierarchical classifi-cation based on their uses (particularly inflammability for lighting purposes) thisdiffered from their classification of the plants themselves. This leads to distinct differ-ences between Ka’apor and Linnaean classification. For example, morphologicallydifferent species (Protium giganteum, P. pallidum and P. spruceanum) are given thesame folk-specific name due to the similar properties of their exudates. Protium gigan-teum and Protium decandrum, which look very similar, had different folk names due tothe different properties of their exudates.


secondary phloem), and phloem fibre, thetaxonomically important and oftenconspicuous fibre which can occur in thesecondary phloem, often making toughbark. He then suggested terms whichcould be used to describe bark texture(consistency), bark patterns (see Box 2.7)and exudates (see Box 2.3). Bark variesconsiderably from species to species in itsthickness and texture. For a particularspecies, bark thickness also varies withtree age, rate of growth, genotype andlocation of the tree (see Figure 2.8). Insouthern Africa, for example, Rauvolfiacaffra trees growing along the coast havea very different outer bark texture fromthose growing in upland sites. Examplesof other bark features are the smell,presence or absence of latex or oxalatecrystals, or elastic threads seen when thebark is broken, or the appearance of inneror outer bark and its cross-section.

Some local people, notably woodcut-ters and herbalists, have an excellentknowledge of bark characters and takesmall bark slashes with a machete to deter-mine the identity of forest trees, ratherthan using leaf or flowers. Be careful torecord whether these are characters offresh bark, dried bark or both, as somebark features are more evident in either adry or a fresh state. An obvious examplewould be the presence or absence of latex,which is clearly evident in freshly slashedbark, but less so in the dry bark that youmight encounter when working with aherbalist or during a market survey. At thisstage, latex will not be exuded when thebark is cut, but may be seen in resin canalsor as congealed lumps and will often havechanged colour.

Underground botany:identification of bulbs, corms and

roots

If, in your field work with herbalists, food

gatherers or ethnobotanical surveys oflocal markets you are unable to identifyroots, corms, bulbs or tubers, do not feelalone! With the emphasis that Linnaeantaxonomy has placed on flowers, fruitsand leaves, and because above-groundplant parts are easier to observe, formallytrained botanists and plant ecologistsgenerally have limited knowledge ofunderground plant parts. Ironically, inlarge areas of Africa, and possiblyelsewhere in frequently burned tropicalsavannas, underground plant biomass isgreater (and often more selectively used)than the above-ground biomass. FrankWhite (1976), for example wrote of the‘underground forests’ in the vast Kalaharisands region, stretching south of theDemocratic Republic of Congo into south-ern Africa and dominated by woody dwarfshrub genera (suffrutices). In the redsyringa (Burkea africana) sandy savanna,for example, which is on Kalahari sands,underground plant biomass is 2.2 tonnesper hectare compared to total above-ground biomass of leaves, stems, flowersand fruits of 1.7 tonnes per hectare(Huntley, 1977).

In contrast with many formally trainedtaxonomists, craft workers, fishermen,herbalists and food gatherers frequentlyhave an excellent knowledge of the charac-teristics of the roots and tubers used fordyes, fish toxins, floats and netting fibre,medicines or food. Underground plantparts can be very distinctive (see Box 2.4).In common with above-ground plant partssuch as shoots and branches, root struc-ture and patterns of root architecture areto some extent genetically determined, andit is often useful to record these importantmorphological characters in ethnobotani-cal work. As you will frequently find –since underground plant parts are sold inmarkets with no leaves, flowers or fruitsattached – the characteristics used by localresource users are important to record and

Applied Ethnobotany

40


use in field work. They may also haveadded value in formal taxonomic work(Pate and Dixon, 1982).

Obermeyer (1978), for example, usesbulb characteristics to distinguish certainOrnithogalum species, also noting howbulb morphology varies with climate andhabitat. Ornithogalum bulbs from winterrainfall areas are generally small inrelation to plant size, while bulbs insummer rainfall areas are large, firm andglobose. The two Ornithogalum species,which grow in permanently wet habitats,also differ markedly from other species asthey do not form swollen bulbs at all,presumably due to a lesser need for storageof food and moisture. The morphology ofunderground plant parts has also beenused to distinguish between reseeder andresprouter categories of dicotyledons andRestionaceae in western Australia (Pate etal, 1990; Pate et al, 1991). Insights fromsuch studies provide valuable informationrelevant to resource management and howplant species will respond to disturbanceand harvest (see Chapters 4 and 5).

Some of the characteristics of under-ground plant parts are shared by the barkand leaves of the same species. Moredetailed examples are given in the section‘Taxonomy with All Your Senses’ and inBox 2.7. Such details include sap occur-rence/absence, colour and odour;root-bark texture and colour; elasticthreads; and so on. As expected, thisapplies to the presence and colour of latexof the Apocynaceae, Asclepiadaceae andClusiaceae family or to the strong, fibrousroot-bark of Thymelaeaceae. Root-bark ofWarburgia, for example, has the samepeppery flavour as its bark and leaves.

Although many bark and root charac-teristics are shared, other characteristicsare not and these need to be carefullynoted. Examples are the characteristic‘cracks’ along roots when they dry out (asseen in Acridocarpus (Malphigiaceae)

roots sold in village markets), the way theytwist, or their shape, size, elastic threadsor cross-sectional appearance. Similarly,bulbs and corms are characterized by acombination of colour, shape, thicknessand structure of scale leaves, latex, occur-rence of irritant chemicals, and markingson the compressed stem that forms the‘base plate’ of the bulb or corm.

Discussions with local resource userscan facilitate development of detailed fieldnotes or even keys (see Box 2.4) to differ-ent categories of fibrous roots or bulbs andcorms. Although keys of this type can bevery useful in improving communicationand understanding between researchersand resource users, such as herbalists, it isvery important that verification and cross-checking are done. In some cases, you cangrow bulbs bought at markets. In others,as with pieces of root, you need to confirmroot characters and make fresh voucherspecimens while collecting with resourceusers in the field.

Describing macroscopic characteristics of wood

Wood anatomists have developed sophisti-cated ways of classifying wood andcharcoal based on macroscopic and micro-scopic characters, primarily due to thegreat economic importance of timber.Wood collections made for this purpose,found in many parts of the world, arelisted in the Index Xylariorum (Stern,1978). Early identification guides towoods using dichotomous keys (based ona choice of one of two characters) havegiven way to systems of hardwood identi-fication using multi-state characters (setsof alternative options for a feature such asheartwood colour, crystal types orgeographic origin) which are used incomputerized keys in addition to dichoto-mous characters.

These well-developed keys for wood


41


Applied Ethnobotany

42

BOX 2.4 BULBS AND CORMS

Corms, for example from Crocus and Gladiolus, are short, swollen food-storing stemssurrounded by protective scale leaves. One or more buds in the scale-leaf axils producenew foliage leaves. A bulb, of which the onion is a well-known example, consists of amodified shoot with a short flattened stem. A bulb is covered on the outside by paperyscale leaves which surround swollen leaf bases. A terminal bud sits at the centre of theupper surface and produces the foliage leaves and flowers. This is an example of partof a key to bulbs and corms commonly sold for traditional medicinal purposes, illustrat-ing characters that may be worth recording in similar studies. In this case, bulbs andcorms were divided into two main groups (group 1: exterior with fleshy or thin scaleleaves; group 2: exterior without scale leaves apparent). Only part of the key to group1 is shown here.

1 Outer surface of bulb covered in dense fibrous ‘hairs’ obscuring scale leaves inGcino (Scilla nervosa)Outer surface without dense fibrous ‘hairs’ see 2

2 Scale leaves fleshy see 3Scale leaves thin, not fleshy see 4

3 Exterior generally smooth; usually pale green in colour, small bulbs almost translucent; scale leaves only conspicuously visible towards the top of the bulb iGibisila (Bowiea volubilis)Exterior not smooth; scale leaves conspicuously visible see 4

4 Scale leaves on exterior conspicuously pointed uMhlogolosi (Urginia spp)Scale leaves on exterior not pointed see 5

5 Scale leaves with persistent leaf-base fibres uMaphipha-intelezi (Albuca fastigiata)Scale leaves without leaf-base fibres iCubudwana (Ledebouria spp)

6 Inner cut surface reddish or light purple in colour; sap stings skin inDongana-zibomvana (Drimia spp)

or isiKlenamaInner cut surface green, white, yellow or cream in colour see 7

7 Bulbs with wide, compressed stem base bearing conspicuous striations see 8Bulbs without compressed stem base bearing conspicuous striations. see 9

8 Elongated bulb with length about twice the size of base diameter; exterior with glossy brown scale leaves inGuduza (Scilla natalensis)Bulb with length approximately equal to base diameter; uMathunga (Eucomis exterior with dark-brown scale leaves; cut interior surface autumnalis)reveals yellowish leaf bases and white compressed stem or iMbola

9 Inner cut surface without sap; large elongate brown bulb iNcotho (Boophane distica)

Inner cut surface with sap see 10

10 Sap very mucilaginous and stings skin uMababaza (Ornithogalum spp)Sap milky-white, not sticky and does not sting skin umDuze (Crinum spp)

Source: Tait and Cunningham, 1988


classification provide a challenge forethnobotanists and innovative taxono-mists. Wood identification systems havebeen developed using excellent collectionsof voucher specimens of woods from manytree species. As a result, wood identifica-tion guides provide a standard for whichethnobotanists working on non-timberproducts should strive to develop for bark,roots, bulbs and corms. The advantagesthat ethnobotanists have in achieving asimilar standard of identification for bark,roots, bulbs and corms are, firstly, thatmany of the characteristics of wood, suchas odour, frothiness, fluorescence, andtypes of crystals, also apply to bark and tosome bulbs, corms and roots. Secondly,local uses already provide information onwhat characteristics would be expected.Saponin-containing fish poisons are welldocumented (Acevedo-Rodriguez, 1990),so that frothiness (a wood identificationcharacter) can also be used as a fieldcharacter to identify bark, roots, or fruitsof many Sapindaceae, Pittosporaceae andTheophrastaceae, for example. The sameapplies to local knowledge of plant dyes(wood, exudate, bark or root colour),odour or ash colour.

To avoid confusion, and enable inter-national cooperation, the InternationalAssociation of Wood Anatomists (IAWA)has developed a standard list, terminologyand computer codes for characters used inwood identification (IAWA, 1981, 1989;Miller, 1981). These and SherwinCarlquist’s (1991) review of the woodanatomy of vines and lianas are recom-mended reading for ethnobotanistsstudying wood use. The IAWA woodidentification system needs both macroand microscopic characters to be used totheir full extent, and the IAWA (1981)standard list of characters, together withMiller’s (1981) explanations, should bereferred to for full details.

The most basic questions you need to

ask yourself (and local people) in trying toidentify a cut wood sample are: wheredoes the wood come from, and what areits vernacular name(s), use(s) and growthform (is it a tree, shrub or vine?). Thesecan all help narrow down the optionsabout the identity of the wood. Bear inmind that boot polish or stains producelook-alike substitutes for scarce, morevalued woods (such as ebony – Dalbergiamelanoxylon). In these cases, the heavinessof an ebony carving compared to the light-weight black boot polish alternative is oneindicator. For greater precision, woodanatomists use basic specific gravity(based on green volume and oven-dryweight) as one of the tools in wood identi-fication.

The next questions you need to askyourself are: is the heartwood coloursimilar to the sapwood (or not), and whatcolour is the heartwood? In mostFlacourtiaceae and Sapotaceae, or well-known trees such as Polyscias (Araliaceae)or Tilia (Tiliaceae), there is no differentia-tion between heartwood and sapwood,whereas it is very distinct in manyFabaceae, such as Dalbergia, Acacia andRobinia. Although brown and pale whitewoods are found in many genera, heart-wood colour can be a good macroscopiccharacter. Yellow heartwood, for example,is a complete giveaway, found only inBerberis species (Berberidaceae), a sourceof medicine and dye. Red heartwood isalso uncommon, characterizing Berchemiazeyheri (Rhamnaceae) and several speciesof Pterocarpus (Fabaceae), Brosimum(Moraceae) and Simira (Rubiaceae). Otherexamples of macroscopic characters whichcan be used in wood identification includethe following.

• Are growth rings distinct or absent?• Is the wood ring-porous, diffuse-

porous or semi ring-porous? Takecare, however: the same species can


43


sometimes be ring-porous and at othertimes, diffuse-porous. Woods that arering-porous also have distinct growthrings. Most Proteaceae have pores in acharacteristic festoon arrangement.

• Does the wood include phloem(interxylary phloem which is embeddedin the secondary xylem of the stem orroot) or intruded phloem? Includedphloem commonly characterizes generawithin the Fabaceae, Loganiaceae,Nyctaginaceae, Menispermaceae,Solanaceae and Urticaceae family.Included phloem is usually destroyedby drying, so small wood samples haveto be preserved in 70 per cent alcoholor formalin-acetic acid alcohol (FAA).

• Is there any odour (or not)?• Is the parenchyma banded, aliform

(wing-shaped) or confluent (joinedtogether)?

Other creative methods used by woodanatomists are the wood characteristicsunder fluorescent (long-wave ultraviolet)light, the colour of wood ash, water andethanol extracts or frothiness (due tosaponins). Simple tests for frothiness(which indicates natural saponins in thewood) and the presence or absence offluorescence (heartwood, water andethanol extracts) are also useful in identi-fying some Fabaceae, Sapotaceae andRutaceae. The test for frothiness hasadded ethnobotanical interest, as plantswith saponins have many uses, including

use as fish poisons, molluscicides and inherbal medicine. Use enough heartwoodshavings (rather than splinters or chipswhose saponin extraction takes too long)to cover the bottom of a clean vial (2cm x7cm). The shavings are then covered withdistilled water, the vial blocked with a corkand then shaken vigorously. If a largequantity of saponins is present, then afroth will form on the surface of the water.If you have access to long-wave ultravioletlight, the same sample should immediatelybe placed under the light to check forfluorescence. Extracts of Zanthoxylumflavum (Rutaceae) and Pterocarpusindicus (Fabaceae), for example, willfluoresce a brilliant blue (Miller, 1981).The same vial with the sample can then beplaced on a hotplate and brought to theboil, with the liquid then checked forcolour. While brown or colourless extractsare not particularly useful in identifyingan unknown sample, reddish (Brasilettia)or yellowish (some Albizia species) can bediagnostic characters.

You could also check on the fluores-cence colour of the freshly sanded orshaved heartwood, water extract andethanol extract under long-wave ultravio-let light. Some will not fluoresce at all,whereas others have a characteristicfluorescence colour. Wood of severalgenera in the Fabaceae (Enterolobium,Robinia) glow yellow-green, for example,while Symphonia (Clusiaceae) andVatairea (Fabaceae) fluoresce orange.

Applied Ethnobotany

Potentials and pitfalls: combining skills in inventories

44

Taxonomy is the branch of biology dealingwith the naming and classification ofliving things. Internationally, biologists usewhat is known as the Linnaean classifica-tion system. Each species is given a nameconsisting of two words. This, the

binomial, consists of the genus or genericname, followed by the specific epithet.Both words together denote the species,and the binomial is conventionally writtenin italic font. The Linnaean classificationsystem is an international link in naming


plants with considerable precision. Allcultures have their own classificationsystems to facilitate communication fromperson to person in naming a particularspecies. Alternatively, specialists such aslocal healers may use special names forplants to obscure the identity of specieswith religious, medicinal or ritual signifi-cance from people who have notundergone initiation processes or special-ist training. Detailed studies have beencarried out in the field of ethnotaxonomyor folk classification systems by BrentBerlin (1992) and are covered in detail inGary Martin’s manual (1995), so will notbe covered in detail here.

Biological inventories of even the best-known national parks in the tropics andsubtropics are often incomplete or inade-quate. In most planning exercises,inventories are done on the basis ofLinnaean taxonomy. This is a perfectlyvalid approach. It can also be useful tocarry out surveys which link with the skillsof knowledgeable local people on the basisof folk taxonomies, particularly whenskilled biologists and taxonomists are ascarce resource. Folk taxonomic knowl-edge can be invaluable in inventory workfor conservation purposes, whether forbotanical or zoological surveys. Baker andMutitjulu Community (1992), forexample, point out from their work inUluru National Park, central Australia,that:

‘In a number of instances Ananguprovided names of animals whichdid not match any that werecaught at the survey sites. Onesuch animal was tjakura. Ananguprovided a detailed description ofthis reptile and brought the animalto the scientists, who identified itas the great desert skink (Egerniakintorei). This species was notcaught on any of the survey sites,

and was found to be restricted toa particular locality within thePark. The woma python(Aspidites ramsayi) and Stimsonspython (Bothrochilus stimsoni)were also recorded for the surveyonly by Anangu.’

It is important to be aware of some of thepitfalls in this process. Firstly, you need toavoid confusion of local plant names withlocal names for plant parts, such as‘flower’, ‘fruit’ or ‘leaf’, or with generalcategories for ‘tree’, ‘shrub’ or ‘vine’. Youwill be surprised how many botanistsunfamiliar with the local language havepublished names which are supposed to bethe specific local name for a plant species,but which actually refer to a plant part ora life form. Embarrassingly, some haveeven recorded the response: ‘I don’t know’as a ‘local name’ from a perfectly honestlocal helper! Sometimes folk biologicalclassification systems may also havetotemic links, so that on Groote Eylandt,Australia, an Anindilyakwa man seeing ared-winged parrot (wurruweba –Aprosmictus erythropterus) flyingoverhead may say, ‘There goes my brother-in-law’ (Waddy, 1982), which could wellconfuse a biologist unaware of the totemicconnection of the species!

You also need to be aware that whilesome species may be what linguisticanthropologists term underdifferentiated,others are overdifferentiated (Martin,1995). The nine Zulu names for the medic-inal plant species Curtisia dentata, inSouth Africa, are an example of overdif-ferentiation, a fairly common occurrencewith species of great cultural importance(see Box 2.5). With underdifferentiation, asingle local name can be a generic term forseveral different plant species. InAfromontane forest in Uganda, forexample, the single Rukiga name bitindiapplies to two Memecylon species – M.


45


jasminoides and an undescribedMemecylon species only found in Bwindiforest; omushabarara applies to at leastthree Drypetes species, including the rareDrypetes bipindensis; omurara applies toat least four Macaranga species withdifferent regional distributions; andnyakibazi applies to three locally endemicRytigynia species in one area, as well as tothe widespread tree Bersama abyssinica.For this reason alone, it is very importantto collect voucher specimens and recordboth local and scientific names for species.

Using several methods in inventorywork can be useful. Firstly, there is whatBrian Boom (1989) terms the‘artefact/interview’ technique, which is acommon approach used by anthropolo-gists, where people are asked the names ofmaterials used to make a particular item.Secondly, there is the reverse approach,termed the ‘inventory/interview’ technique,which involves the active collection ofplant specimens and subsequent interview-

ing of informants about their names anduses (Boom, 1989). This has been appliedin several studies where specimens arelinked to one-hectare forest plot surveys(Prance et al, 1987; Boom, 1989). Thirdly,there is the ‘walk in the woods’ approach,where records are made with key helpersdirectly from whole plants while in thefield, rather than in subsequent interviewsfrom fresh specimens. This avoids misiden-tification which is a danger in the‘inventory/interview’ technique in forestareas, where formally trained botanistscollect botanical specimens (leaves,flowers, fruits) to show to local resourceusers who often use criteria of bark, rootsor stem morphology as the main classifica-tion criteria.

These approaches are outlined in moredetail in Gary Martin’s (1995) EthnobotanyManual, published in this series. On theirown, neither vernacular nor botanicalnames are sufficient for thorough work,although the value of using agreed scientific

Applied Ethnobotany

46

BOX 2.5 MULTIPLE NAMES, SINGLE SPECIES

The tree Curtisia dentata (Cornaceae), endemic to Afromontane ‘islands’ of forest, isone example. It is a multiple-use species with hard, durable timber but is best knownamongst traditional healers for its red bark, used by sorcerers to cast spells(ukuthakatha), causing hysteria. Even within a small part of its range in Natal province,South Africa, I recorded eight Zulu names for this tree. Three names (inkunzi-twalitshe,ijundu-mlahleni and umphephelelangeni) are hlonipha (respect) names used only byherbalists. One name (igejalibomvu) is only known by people (specialists and non-specialists alike) within a few districts near the Nkandla and Qudeni forests; one(umgxina) is known only in the southern KwaZulu/Natal, where the name is morecommonly used amongst Xhosa-speaking people. Only three names are more commonregionally (umlahleni and umlahleni-sefile, and umagunda). Lastly, a ninth name,umhlibe, was recorded by ethnobotanist-linguist and missionary extraordinaire JacobGerstner in the 1930s, but not at all in my regional medicinal plants survey 50 yearslater. All have meanings alluding to collection or use, additional factors that can beconfusing to people from outside the immediate area of use or outside the profes-sional ranks of traditional healers. In discussions, umlahleni and umlahleni-sefile mean‘throw him away’ or ‘throw him away, he’s dead’; ijundu-mlahleni also refers to aperson being ‘thrown away’, and ijundu to medicine that kills a person outright.Umphephelelangeni means ‘something that escapes to the sun’ and igejalibomvu refersto a ‘red hoe’, and so on.


names is paramount, particularly for inter-national communication and forpublications. What we also need to knowin the supply-demand balance of sustain-able use is which species are most favoured,or which individuals within a speciespopulation are selected for harvest. This isan important issue in setting conservationor resource-use policies for plant species.

Although individuals with the richesttraditional knowledge often live in remoteareas and are frequently marginalized byurban-industrial society, their skills areoften recognized within their own commu-nities. As Gary Nabhan and his colleagues(1991) point out:

‘More often than not, conservationbiologists are unaware that indige-nous people have detailedknowledge of a particular rare plantor animal that is under study. Even

when not biased against traditionalpeoples, these biologists often lackthe skill of ethnographic interview-ing and the incentives to learn fromindigenous people who live in closeproximity to rare species. To date,only one US endangered speciesrecovery plan, which recommendedlocal Navajo participation in thehabitat protection and plantpopulation recovery efforts forCarex specuicola, has includedethnobotanical data derived fromindigenous people.’

Similarly, social scientists with training incross-cultural communication lack thetaxonomic skills or ecological insight ofconservation biologists. For this reason, ateam approach involving local traditionalexperts, biologists and social scientists canbe very useful.


47

BOX 2.6 MISMATCH: LINNAEAN NAMES TO FOLK TAXONOMY

AND VICE VERSA

Discrepancies between Linnaean and folk taxonomy are widespread, and these differ-ences must be taken into account. Firstly, they may refer to chemical, genetic ormorphological differences not considered by Linnaean taxonomists, and which indicateuseful qualities for plant breeding or phytochemistry. They also indicate the need toget scientists to take note of folk taxonomy. Secondly, they are important whenresource management priorities are being discussed – for example, species recordedunder a single name which actually represent more than one species of different conser-vation status, as when an endemic species with restricted distribution has the samename as a widespread, non-endemic species. This can apply to both Linnaean and folktaxonomic systems. In the early 1980s, I collected several voucher specimens of acommonly eaten wild spinach in the Ingwavuma district, South Africa. These were allidentified scientifically as Asystasia gangetica, yet locally identified as two separatespecies with different habitat preferences and local Tembe-Thonga names. The first,known as isihobo, was widespread, growing in fallow fields and along forest margins,with thin leaves that were not particularly tasty. The second local name, umaditing-wane, referred to a robust, fleshy-leaved species growing on coastal dunes with leaves‘as good as meat’ to eat. These were even selected for cultivation at home becausethey were so tasty. A few years later, on the basis of this local knowledge, umaditing-wane was more carefully examined and described as a ‘new’ separate species, Asystasiapinguifolia – a regional endemic along the Mozambique coastal plain – in belatedaccordance with the local folk taxonomy.


Matching folk taxonomy withLinnaean taxonomy

A cross-reference of vernacular to botani-cal names provides an invaluable guidewhen using the methods that follow, aswell as in general discussions or interviewsurveys. There is no substitute for work inthe field with a range of local resourceusers – men, women or children – who arelocally acknowledged to be experts ondifferent categories of plant uses.Inventory work can often be done by localpeople themselves, after guidance on therequirements for good herbarium speci-mens. Researchers need to be aware ofgender issues that may be involved in fieldwork. Work with midwives, for example,is often best done by female researchers,or work with hunters done by men. Thereis also a need to be aware of differencesand complexities in different namingsystems.

‘Rapid’ approaches advocated byrapid rural appraisal (RRA) and participa-tory rural appraisal (PRA) practitionersthat fail to match local names with inter-national taxonomic nomenclature are thepoorer for it. Information from the inven-tory and field-based discussion phase witha few key resource users can be invaluablefor cross-checking information fromranking and scoring exercises (see‘Participatory Methods with Groups ofPeople’). If species are identified byvernacular name only, communicationamongst local people within a district orregion may be limited, and this limits localpeople’s access to information on uses ofthe same species in most published formsas well.

Some vernacular names may only beused by specialists, others only in a verylimited area, while use of some vernacularnames may be widespread. In certaincases, this obscurity is for a good reason,as in traditional medicines used with

powerful and magical symbolism (see Box2.5). In others, there is a need to improvecommunication and link local names tointernationally used botanical names (seeBox 2.6). In any language, words can beused both to communicate or to obscuremeaning, and you have to be aware of thisin discussions and interviews. Althoughthey may initially sound awkward, there isno getting away from using Latin namesto refer to plant species; even though theymay seem strange at first, they are part ofan internationally used naming system.This enables you to work from vernacularnames and opens the door to publishedinformation on distribution, qualities,uses, population biology, conservation andcultivation which may be highly relevantto local agroforestry, health care or conser-vation programmes.

Quantitative methods: species use values

Although individual or group surveymethods are required to determine prefer-ences for some plant uses, such as foredible or medicinal plants, in several casesquantitative assessments of wild plant usecan be made from work with localresource users. In other cases, quantitativeassessments can be done as part of ethno-botanical surveys of marketplaces (seeChapter 3).

Three quantitative ethnobotanicalmethods are commonly used: informantconsensus, subjective allocation and usestotalled. A fourth, less commonly used,interview/resource assessment method isuseful for resource management purposes,since it records resource users’ assessmentsof individual plants in terms of theirutility. These quantitative approaches adda new and important component toethnobotanical work. The informantconsensus method, in particular, allows forhypothesis testing. These methods also

Applied Ethnobotany

48


enable comparisons of use between vegeta-tion types or ecological zones, betweendifferent cultural groups or betweenpeople of different ages, gender or occupa-tion within or between communities.Several of these approaches have recentlybeen reviewed by Phillips and Gentry(1993a,b) and Phillips (1996).

Informant consensus

In this method, the relative importance ofeach use is calculated directly from thedegree of consensus in the responses of thepeople interviewed. The advantages arethat this yields data which can be testedstatistically, and it is relatively objective.However, it is time consuming, as individ-uals or households must be interviewedseparately. Oliver Phillips used the follow-ing formula to analyse the results of astudy in Peru:

UVis = ΣUisnis

UVis stands for the use value (UV) attrib-uted to a particular species (s) by oneinformant (i). This value is calculated byfirst summing (indicated by the symbol Σ)all of the uses mentioned in each interviewevent by the informant (Uis), and dividingby the total number of events in whichthat informant gave information on thespecies (nis).

Phillips and Gentry (1993a) worked inplots of six types of forest in the ZonaReservada Tambopata in southern Peru.They recorded the uses of trees and vinesof 10cm diameter at breast height (dbh) ormore in a total area of 6.1ha. Use datawere recorded from 29 mestizo (mixedSpanish-Indian descent) people who wereinterviewed in the forest plots or in theircommunities. In terms of data analysis,each act of interviewing a local person onone day about the local names and uses of

one species was classified as an ‘event’. Ifa species was encountered more than oncein a single day, the person’s responses werecombined. During 12 months of fieldwork spread over 5 years, the researchersand local people participated in 1885independent events.

Results from different events can beadded to use-values derived from otherlocal people (Σi UVis). This is then dividedby the total number of people interviewedabout that particular species ns to yield theoverall use-value (UVs), as indicated in thefollowing formula:

UVs = Σi UVisns

Although this statistic was used initiallyon results from interviews that took placein forest plots, it could be applied to anydata-gathering technique in which numer-ous people give information on a range ofplant resources. For example, if you workwith local collectors who make a largenumber of ethnobotanical collections,each voucher specimen with its accompa-nying data sheet could be considered an‘event’. It is likely that each species will beencountered numerous times by eachcollector, so the number of uses on eachdata sheet can be added together to obtainUVis, the individual use-value. These canbe summed for all collectors to calculatethe overall use-value for a particularspecies (UVs).

Subjective allocation

This is a semi-quantitative method, wherethe relative importance of each use issubjectively assigned by the researcher onthe basis of the assessment of its culturalsignificance (Lee, 1979; Turner, 1988). Itsadvantage is that it is quicker than theinformant consensus method; however, theresults are not so amenable to statistical


49


analysis since they are more subjective. Italso does not allow for more than one usefor each species within each category(Phillips, 1996).

Turner (1988) developed an index ofcultural significance for each plant species.The index is made up of a range of plantuse attributes as recognized by theresearcher. The index of cultural signifi-cance (ICS) for each species is thencalculated as:

ICS = Σ(qie)ui

where, for use u, q is quality value, i isintensity value and e is exclusivity value.Therefore, the use value of each species isthe total of all (q x i x e) calculations foreach use. This index allows an indepthanalysis of species usefulness.

Uses totalled

In this method, no attempt is made toquantify the relative importance of eachuse, the numbers of uses simply beingtotalled, by species, category of plant use,or vegetation type. This method is fairlyquick and simple. Oliver Phillips (1996)points out two problems with this method.Firstly, it does not distinguish between therelative importance of different uses orspecies. Secondly, the results are notweighted by the intensity of samplingeffort. As a result, the quantity of usefulplants reported can be a result of researcheffort rather than reality. He considers thatthese problems may be less important forcountry or ecosystem-wide comparisons(such as Toledo et al, 1992), but that theymay become a problem with small-scalestudies.

Interview/resource assessmentmethod

A simple method that gives useful infor-mation for resource management purposes

is where local resource users grade theusefulness of plants within plots. Thisexercise can be repeated independentlywith several resource users (or with smallgroups of people), giving insights intowhich species are most favoured withinbroader use categories (such as trees forfuel, building poles or carving) and whyfavoured individuals are selected withinthose species based on size class, shape,health of the plants or possibly evengenetic factors. As each stem (or leaf orwhatever is the focus of the study) ismeasured by the researcher, the localresource user rates it either very good,good, acceptable, poor or not useable,giving the reason(s) why (see Figure 2.9).Cut stems (or leaves) that have alreadybeen harvested or have been damaged inother ways (such as eaten by animals) arealso recorded.

Essentially combining standardecological plot methods with the addedbenefit of work with local harvesters, thismethod not only gives insight into whathas been harvested, but also into selectioncriteria, what level of harvesting might beexpected and whether this will be sustain-able or not. As this can be timeconsuming, either due to high density ofstems of a single species or high speciesdiversity within plots, it is important tohave time available and to select an appro-priate plot size. Alternatively, assessmentsbased on transects or quantitative plotscan be done in sites where harvesting hasalready taken place, such as from theproportion of cut stumps or root or barkdamage assessments within plots (seeChapters 4 and 5).

In some cases, it may be necessary towork with small groups of harvesters.Although this means that independentratings by individual harvesters are notpossible, it has the advantage, in commonwith participatory-group interviewmethods, of being able to listen to

Applied Ethnobotany

50


harvesters debate the merits of differentqualities of the resource. In Botswana, forexample, I worked with groups of four tofive basket makers who assessed thesuitability of leaves of Hyphaene peter-siana palms which were experimentallycultivated as an alternative basketry fibresupply source. The unopened leavesusually harvested for basketry werechecked by women basket makers in twostudy plots on different soils: 70 palmstems in one plot and 74 palms in another.

Women in each group inspected eachpotentially suitable leaf (one per palmstem), had a brief discussion, thenconcluded why the leaf was acceptable forbasketry or not. The majority were not

considered suitable because they had poorqualities: ‘too hard as they hadn’t beensoftened by frequent harvesting’ (33 percent and 51 per cent respectively); ‘toothick’ (8.5 per cent and none); ‘too short’(23 per cent and 13 per cent); ‘leaves ayellowish colour’ (1.7 per cent and 3 percent); and ‘leaf apex skew’ (none and 1.4per cent). The remainder had good quali-ties: ‘soft and pliable’ (43 per cent); ‘notrough (good texture, fewer spines)’ (6.7per cent and 5.7 per cent) and ‘sharpstraight tip’ (none and 1.4 per cent). Theseinsights would probably be missed in theinformant consensus approach whichrecords whether the species (rather thanindividuals within the species) is useful.


Local to international units

51

Source: Cunningham, 1996b

Figure 2.9 Local harvester assessments of bamboo utility in Bwindi-Impenetrable National Park,Uganda, showing the high proportion of stems that are unsuitable for building purposes and the

reasons why, the low level of harvesting and number of stems eaten by non-human primates

100

80

60

40

20

0

Number of stems

Shoots Full height Dead Cut

Bamboo category

Dead or cut

Too small

Bamboo : plot 4resource value (n = 189)

Good for building

Borer

Crooked

Damaged

Eaten

Undamaged

Measuring the quantity of plant productssold can be valuable in linking with otherquantitative studies on plant biomassproduction, such as fruit or sap yields perplant or per hectare (see Chapter 4). To beof most value, however, the measurements

need to accurately reflect the quantitiesthat really are being harvested, consumedor sold. Inaccurate measurements ofvolumes used are of limited value.Quantitative market surveys are expensiveand time consuming, just as interviews are


Applied Ethnobotany

Figure 2.10 Local units, whether bundles, baskets, bicycle or vehicle loads, can be useful units ofmeasure convertible to international units of mass (tonnes, kilograms) or volume (litres, cubic

metres) for comparison to other studies on biomass production or yields

52


a more time-consuming and expensivesocial survey method than RRA or PRAmethods, yielding a different type of datathat will take longer to process. It is there-fore important to focus on key resourcesto ensure that the research funds and avail-able time are well used.

The size, shape, and weight of the‘units’ in which plant resources are trans-ported are very variable, coming inbundles, bags or bottles of varying types(see Figure 2.10). Units will also be deter-mined by which storage or transportcontainers are most commonly available,and this will often vary with time. Despitethis variation, some local ‘units’ may befairly constant over a wide area since somebundle sizes are dictated by the easiest wayof transporting that particular product.Thus, similar-sized reed bundles are soldthroughout southern Africa, measured by‘arm-circumference’ diameters. Otherbundles, such as thatch grass or weavingmaterial, may be measured in subunitswhich can be useful in discussions andPRA work, such as hand-circumferencebundles or bundles the same circumferenceas thumb and forefingers (see Figure 2.11).Measuring large samples of local ‘units’

(at least 30 of each local unit) gives themean mass or volume and standard errorfor each unit, enabling conversion to inter-national units and estimates of quantitiesused.

Recovery rates and quantity used

It is important to avoid under or overesti-mating the quantities of plant materialbeing used. One way of avoiding this is toassess the quantity harvested at source, sothat you measure unprocessed amounts,which can then be related to wild-harvested material. If this is not possible,such as when you are working from long-term sales data, you need to take recoveryrates into account, whether due to air-dryversus wet weight of material harvested orto loss of volume during harvesting.

In some cases, processed products canrepresent a high proportion of harvestedmaterial. In other situations there may bea deceptive increase in volume – forexample, when palm wine is diluted beforeresale (Cunningham, 1990). You also needto watch for cases where recycling ofharvested resources takes place. Somestudies of wood consumption by house-


Figure 2.11 Commonly used local estimates of circumference for different-sized bundles of plant products

53

5

15

50


holds have been biased through not takingnote of the recycled use of old timber usedfor housing construction as fuel. In herstudy of this in Zimbabwe, JoannMcGregor (1991) showed that 1.3 tonnesof wood per household per year wererecycled as fuel (see Table 2.2).

In woodcarving or in the case of someedible foods, very little of the unprocessedplant material becomes the final product(see Figure 2.12). Ebony (Dalbergiamelanoxylon), for example, is a primewoodcarving timber used in the touristtrade. It is also the main wood used forseveral musical instruments, includingclarinets, flutes and recorders. Since it isoverexploited in many parts of Easternand South-Central Africa and is a veryvaluable timber, it is useful to determinethe volume of cut timber exported formusical instruments. The only data avail-able during a survey of ebony use in

Tanzania were final product volumes andaverage recovery rate. Only 7 per cent oftotal log input was recovered as a ‘finalproduct’ of wooden blocks or billetsexported for the musical instrument indus-try (Moore and Hall, 1987). Data from theSawmill Industry of 125 m3/yr during1987 were then converted, to give anestimate of the volume of total log inputused annually:

Final Product Volume = 125 m3 = 1785 m3

Average 0.07 of logs/yrRecovery Rate

In other cases, recovery rates can be high.This facilitates a far more accurate assess-ment to infer quantities of raw materialsharvested. Weaving material is one suchexample. In a study of commercial tradein crafts in South-eastern Africa, for

Applied Ethnobotany

Table 2.2 Total annual wood consumption (tonnes per household per year)

Live wood Recycled from Dead wood Total domestic construction (including leftovers) fuel

Domestic fuel 1.8 1.3 2.7 5.8Brewing 1.3Brick burning 0.4Construction 1.3Total live wood 4.8

Source: McGregor, 1991

Table 2.3 Basket makers’ assessments of Hyphaene petersiana palm leaves rejected or consideredacceptable for basketry

Reasons given ETSHA-5 ETSHA-8 Reasons given ETSHA-5 ETSHA-8% (No) % (No) % (No) % (No)

a) Good qualities b) Poor qualities‘Soft’ (pliable) 43.4% (26) 30.1% (21) ‘Too hard’ 33.3% (20) 51.4% (36)

‘Too thick’ 8.3% (5) 0

‘Not rough’ (goodtexture, fewer spines) 6.7% (4) 5.7% (4) ‘Too short’ 23.3% (14) 12.8% (9)

Sharp tip 0 1.4% (1) Yellow colour 1.7% (1) 2.9% (2)Leaf apex skew 0 1.4% (1)

Source: Cunningham, 1992

54



55

Figure 2.12 Low recovery rates from an edible fruit: steps in the processing of (a) Strychnosmadagascariensis fruits – (b) cracked open after harvesting, (c) seeds and pulp removed and

sun-dried, (d) then smoked and (e) the pulp finally removed for storage or consumption


Applied Ethnobotany

56

BOX 2.7 STANDARDIZED TERMS FOR DESCRIBING BARK

COMPONENTS, BARK TEXTURE, PATTERNS AND EXUDATES

Bark Texture (Consistency)

This is the composition of the bark, mainly resulting from the characteristics of the cellsmaking up the tissue and the extent of decay of the outer bark (rhytidome). Bark texturemay be corky (like cork), fibrous (the outer and/or inner bark dominated by fibres),brittle (outer or inner bark hard, breakable), loose (outer and/or inner bark breaks upon cutting into coarse or fine grains or flakes), granular (inner bark mainly composed ofsclereids), mealy (outer bark falls off like powder), homogeneous (either fibres orsclereids occur) versus heterogeneous, soft (outer and/or inner bark is soft and easy tocut versus hard), laminate (layers in the phloem formed by sclerenchyma).

Bark Patterns

Four aspects of bark texture provide useful macroscopic characteristics: firstly, thoseseen in cross-section or in a bark slash (blaze); secondly, bark fissuring; thirdly, barkscaling; and finally, the external appearance of the bark.

Patterns in cross- and tangential section

This includes the following: corrugations (inner bark surface corrugated, matchingsimilar pattern on sapwood), dilatation (growth) (patterns from the process of tangen-tial widening of the bark during growth seen in bark slash), flame marks (a pattern likeflames seen in bark cross-section, formed by phloem rays), phloem, mottled (colouredspots seen in bark slash), phloem, scalariform (in cross-section, the phloem rays form aladder-like structure, with radial ‘rungs’), ripple marks (fine parallel horizontal linesseen in the bark slash), streaks (striations on the surface of the bark slash usually formedby phloem rays and sclerenchymatic tissue), these may be longitudinal, as in Wormiatriquetra (Dilleniaceae) bark, which has dark streaks like coconut wood, or reticulate(regular or wavy lines), like the bark of the West African timber tree Terminalia superba.

Bark fissuring

Fissured bark is cracked lengthwise into fissures (generally longitudinal grooves)between ridges in the outer bark (or rhytidome). Bark fissures may be parallel (usuallylong and regular), reticulate (with grooves joining each other and dividing again) oroblique (short or long grooves, joining and splitting again – anastomosing, but not asregularly or distinctly as with reticulate fissures). Fissures can also be classified accord-ing to depth and length: deep (at least half as deep as total bark thickness), shallow(less than half as deep as total bark thickness), boat-shaped (oval or elliptical fissureswhich are not continuous), short (<15cm long), long or elongated (>15cm long), V-shaped (sharp V-shaped cross-section), round (concave in cross-section), square-shaped(fissures flat-bottomed in cross-section), irregular (different-sized gaps and furrows),compound (anastomosing shallow fissures in the bottom of main fissures), wavy (coarselongitudinal grooves with irregular, wavy faces). Also take a good look at the barkridges, the raised part of the outer bark between the fissures. These can be flattened,hollow (concave in cross-section), rounded (convex in cross-section), V-shaped or reticu-late, joining each other, irregularly dividing and enclosing non-continuous fissures.



57

Bark scaling

Many tree species have flaky, ‘shaggy’ or ‘scaly’ older outer bark which becomesdetached, such as mvule (Milicia excelsa) in Africa or many Australian Melaleuca andEucalyptus species. Following Wyatt-Smith’s (1954) work in Malaysia, bark flakes arethose patches of outer bark more than 7.5cm long which become detached, whereasbark scales are less than 7.5cm long. Flakes and scales can vary in shape or thicknessand may be rectangular, irregular, circular, papery, scrolled (thin flakes rolled up attheir edges) or shaggy (loosened, usually curved rectangular or irregular flakes whichmay hang for a while on the stem). In addition, bark scales can be flat-sided (one orseveral layers thick), chunky (with irregular rough faces and an irregular shape), scallop-shaped (thickest in the middle, tapering to the edges, leaving a scalloped-shapeddepression on the tree stem when they drop off).

Bark may be heterogeneous (with more than one type of bark on the same stem),as in the miombo tree, Brachystegia bussei, patchy (usually with two colours dominat-ing, often with lighter blotches due to irregular dehiscence), powdery (with a finepowder-like crust which rubs off easily), usually found on smooth barks like the fever-tree, Acacia xanthophloea, stringy (thick, loose-fibred bark, never deciduous), surfacerotten (bark has short fissures, varying in depth and cross-section, scaly, rugose orsmooth, with variable bark scales – small, adherent, chunky or flat-sided; in cross-section, the inner edge of the outer bark follows the surface shapes and is not parallelto the cambium), tessellated (with fairly regular, square or oblong plates or blocks onthe bark surface, which remain on the stem for a long time) or ring-bark (a type ofouter bark where periderms are formed parallel to the first one, resulting in concentriccylinders of outer bark which detach, often annually, in large sheets).

External markings

The outer bark surface may be dippled (covered with shallow, usually circular depres-sions >1cm in diameter which are the scars of the scaled-off old bark), pock-marked(covered in depressions <1cm in diameter), scribbly (with characteristic ‘scribble’patterns caused by insect larvae), rugose (which looks smooth from a distance, but iscovered by wrinkles, depressions, shallow fissures or irregular scales), rough (with anuneven surface, such as scaly, flaky or fissured barks), smooth (thin and unbroken,although it may have lenticels), lenticellate (covered with raised, sometimes corky spotsor lenticels).

These vary in their shape, size, frequency, grouping and consistency. Lenticels maybe linear (like button-holes), round, stellate (star-shaped) or diamond-shaped, and canalso be described according to their diameter as large (>5mm), medium (3 to 5mm) orsmall (<3mm). They may be numerous (or scarce), solitary or compound, occurring invertical, horizontal or oblique lines, and may be soft, powdery or compact. Bark mayalso be characteristically marked by eye-marks (eye-shaped marks which often corre-spond to old leaf or lateral branch scars) common on Bridelia micrantha, hoops (raisedtransverse rings which partially or completely circle the trunk), ring-grooves (transversegrooves which partially or completely circle the trunk) as in Alberta magna, scroll marks(similar to dippled bark, but where sinuous marks are formed between depressionswhen scales are detached), prickles (which can be detached without tearing the wood),spines (which tear the wood if detached), and may be simple (versus branched), straight(versus curved), hard (versus supple), tapering or conical. Finally, the tree trunk orbranches may be characterized by burls (hard woody outgrowths), cankers (localizedlesions on the bark or cambium) and warts (bumpy outgrowths which are not spines,prickles or burls).


Applied Ethnobotany

58

Figure 2.13 Be careful not to make incorrect assumptions of recovery rates from quick surveys,and take part in harvesting expeditions if at all possible to determine selection procedures. (a) Smilax anceps – but vertical stems are not selected. (b) Processed horizontal runners.

(c) Completing the final product


example, I was able to use sales recordsfrom the sole craft work organization inthe area to determine annual sales, whichfortunately had been kept according to thematerials from which the crafts weremade. All craft items were weighed toconvert records of numbers of baskets soldto mass of material. Recovery rates werethen determined by working with localbasket makers. With the exception of palmleaves, other plant fibres were not reducedin volume after they had been selected byleaf size class and dried. Ilala palm(Hyphaene coriacea) leaves were alsoamongst the most favoured material (seeTable 2.3). Recovery rates were high asonly the leaf mid-rib and petiole werediscarded. Air-dry weight of whole leavesof favoured length was 310g (SE ± 97g).This provided 249g (SE ± 95g) of weavingmaterial. Taking this recovery rate intoaccount, the estimated number of leavesused per year varied: 5554 leaves (1978),6622 (1979), 5884 leaves (1980), 10,584leaves (1981), 7963 (1982) and 9902

leaves (1983). This could then be relatedto leaf-production rates in the palmsavanna mapped from aerial photographs.

Other cases are not as simple, empha-sizing the importance of field observationand participation in harvesting expedi-tions before conclusions are made onquantities of material used. Smilax anceps(Smilacaceae), known locally as ensuli, isthe most favoured basketry material inwestern Uganda and is easy to identify andfrequently pointed out by local people (seeFigure 2.13). What is less obvious, unlessone participates in harvesting, is that it isnot the vertical stems, as one might expect,that are used. Basket makers carefullyselect horizontal ‘runners’, and any infer-ence on impact of harvesting has to takethis into account.

It is in these cases that participantobservation and careful discussion byecologically aware researchers can avoidsome of the inaccuracies that may resultfrom rapid surveys.


59


Understanding trade networks is a key todesigning practical conservation, resourcemanagement or rural developmentprogrammes for species in trade. Knowingwhich species are sold, how much they sellfor, or the quantities sold, is not enough.It is equally important to know who isinvolved in sales along often complexmarketing chains, how this is organizedand how it is changing, where source areasused to be or are today, and how demandand supply are likely to change in thefuture. In this chapter, I give a generalintroduction to ethnobotanical surveys ofmarkets, then discuss the theoreticalbackground to studying the location andclassification of markets, market schedulesand the different sellers within them, andsuggest a checklist of guidelines to assistresearchers in conducting market surveys.

Rural people, moving from a subsis-tence lifestyle to a cash economy, haverelatively few options for generatingincome. They can sell agricultural orpastoral produce, work for a cash wage inagriculture or industry, or retail goods inlocal or regional marketplaces. For therural poor without land or livestock,harvesting of wild plant resources is acommon option, particularly for people in

ecosystems with low arable potential. Agrowing demand from urban areas cataly-ses this trade, drawing in resources fromrural areas to towns and cities, often forfavoured fuelwood, building materials,medicinal or edible wild fruit species.

Many people who harvest and sellwild plants are from what is termed theinformal sector: self-employed people,generally unrecognized in official statis-tics, who have little access to capital andwho earn money from labour-intensiveenterprises. From first harvest to final sale,the trade in wild or naturalized plants forlocal, national or regional trade forms partof an informal ‘hidden economy’ sector.International trade in these plant productsis more obvious, as middlemen link theinformal sector to an export sector forwhich export or import records aresometimes kept.

This trade provides an importantopportunity for systematic ethnobotanicalsurveys and a rich source of informationfor conservation, rural development andresource management programmes. Thereare several reasons for this.

Firstly, the species that are sold are a‘short-list’ of a much wider range ofspecies that are available which can be

Introduction

Chapter 3

Settlement, Commercialization and Change


cross-checked with information fromsocial surveys (see Table 3.1, and Chapter2). If demand for a species or resourcecategory such as fuel, basketry fibre orherbal medicine is high and supplies arestill available, then these species orresource categories will be sold in manymarketplaces. Conversely, a species orcategory of plant use in low demandwould be less common in marketplaces.

Systematic ethnobotanical surveys of localmarkets not only classify the species onsale, but also arrange them into hierarchi-cal levels – which may reflect their relativedemand. There is an important exception:some of the most useful and popularspecies no longer feature in markets dueto overexploitation.

Secondly, the shift from subsistenceuse to commercial sale can have important


Table 3.1 A summary of methods used at different levels of detail in the study of exchange and distribution

Level Type of data Information obtained Problems

Minimum Government records and Government price records and other Reliabilitypublished information market information; market calendars

Key informant Trader types; commodity types; Reliability andinterviews overview of organization of knowledge of key

exchange – who, where, when, what, informant(s)how; transportation information; units of measures and standard of valule; bargaining methods

Second Household surveys What is bought/sold/exchanged Resource limitationsin carrying outsurveys

Trader surveys Demographic information; Design ofinformation on organization of appropriateexchange; social relationships, credit questionsbargaining; movements of traders

Market surveys (where Number of vendors and types of Determination ofmarketplaces are main vendors; number of commodities set markets tosites for exchange) and types of commodities; study

functional groupings

Third Market surveys Movement of people (vendors, Resourceconsumers), movement of vehicles limitationsand commodities

Number of commodities (functions) Modify central-and vendors as basis for determining place methods forcentral-place hierarchy of market specific system,system especially separate

wholesale and retail functions

Most Observation and in-depth Details on social organization of Time and intensivecomprehensive interviews exchange – bargaining, customer work by research

relations, social networks, etc;additional information on basicquestions (who, when, what, how)

Source: Trager, 1995

61


implications for resource managementsince it results in larger volumes beingharvested, a higher frequency and inten-sity of harvesting, and can affect resourcetenure. In some cases, commercial harvest-ing strengthens resource tenure and theincentive to conserve individual plants.Commercial sale of wild fruits, forexample, maintains the incentive toconserve wild fruit-bearing trees in partsof Africa where development of a socialstigma against gathering wild fruits as afood resource is undermining the ‘tradi-tional’ practice of conserving wild fruittrees. In other cases, the shift from subsis-tence use to commercial harvestingweakens resource tenure and underminescustomary controls of resource use.Harvesters are often people with lowincomes and few resources in reserve. Insome cases, the species producing the tasti-est fruits, strongest fibres, the mosteffective medicinal plants, or finest timbers– those with the greatest value to localpeople, and with potential for national or

international horticultural or industrialdevelopment as new crops – are thosemost likely to be overexploited.

As roads reach further and further intoremote resource-rich regions, initiatingsettlement and clearing habitat, commer-cial harvest and trade usually increase (seeFigure 3.1). Improved transport networksstrengthen the link between ruralresources and urban demand. They alsoresult in an influx of outsiders, frequentlydisrupting traditional resource tenuresystems and increasing the scramble foreconomically valuable resources. Speciesidentified in market surveys are an impor-tant component in a matrix of biologicaland socio-economic criteria that can beused to prioritize species for conservationpurposes (see Box 3.2).

Thirdly, systematic surveys of localmarkets are a good way of identifyinginformation on indigenous species whichhave already been domesticated or whichhave a high potential for domestication.

Applied Ethnobotany

Figure 3.1 As roads reach further and further into remote resource-rich regions, initiatingsettlement and clearing habitat, commercial harvest and trade usually increase. This photograph

shows road construction through tropical forest in Côte d’Ivoire, West Africa

62


An outsider, walking into a busy rural orcity marketplace for the first time, is facedwith unusual scents, exotic sounds, brightcolours and densely packed crowds. Atfirst sight, marketplaces may appear totallychaotic. To local people, however, there isa clear pattern in the crowded marketindicated by types of sellers, what they areselling, where they come from and by thelocation of particular categories of goodsor sellers within the marketplace. Inaddition to these patterns within market-places, there is generally also a pattern inwhere markets are located within rural orurban areas, and when these markets areheld.

Before deciding on how or where toperform detailed ethnobotanical surveysin markets, it is useful to view markets ina stepwise, systematic way – firstly, fromthe perspective of an economic geographeror anthropologist, and then as a botanist.The question is: where do we start (orstop) collecting information? Just asharvesting impacts have to be viewed atdifferent scales, from a regional landscapelevel to species-populations and individualplants (see Chapters 4 to 6), with mappingand an understanding of vegetationdynamics as tools in this process, so tradeand markets need to be viewed from asystems perspective. One of the toolsfrequently used by geographers to under-stand how markets are organized isderived from central place theory, whichuses a hierarchical approach to systemati-cally classify market systems. One of thebest-known examples of this approach isSkinner’s (1964, 1965) seminal study ofmarkets in China. Despite the shortcom-ings of central place theory pointed out bysome archaeologists (Crumley andMarquardt, 1987) and geographers

(Vance, 1970), it continues to be the basisfor many studies of exchange and distrib-ution (Trager, 1995). To understandmarket networks as systems, we need tostep back from the exotic and seeminglychaotic and crowded market and workdownwards from a regional landscape (oreven global scale in the case of the inter-national export trade) to finer levels ofdetail at smaller spatial or time scales.

Survey methods will also depend uponwhether you are doing a short- or long-term survey, on the gender of the sellerswith whom you will be working, and onthe types of products and markets onwhich your survey will focus. Detailedsurveys can cross-reference with informa-tion collected from other methods, such asparticipatory mapping, seasonal calendarsand ranking of products described inChapter 2. If you are working in a largestudy area, you will need to decide onwhich markets to survey, when to surveythem and what to survey (see Box 3.1).You will notice factors that influence yoursurvey, such as the following.

• Markets are often located atpredictable places in the landscape,such as at stopping-off points ontransport routes along major rivers orroads, in villages and at predictablesites within urban areas (taxi ranks,bus stops).

• Within markets, sellers are oftendistributed according to the type ofgoods they are selling. Their locationwithin markets and the types of goodsthey sell are also indicators of theirsocio-economic status.

• The timing of markets will varyaccording to a variety of factors,including the size of the village,


Local markets: order within ‘chaos’

63


seasonal availability of the resource,or daily activity. For this reason, youneed to be well aware of the bias thatcan occur due to poorly timed surveys.

• Observe, discuss and think carefullyabout what units you will use tomonitor which products in moredetailed surveys. Survey forms mayalso need to be designed according tothe literacy levels of people involvedin monitoring, such as the palm-winesurvey form in Figure 3.2.

• Where do plant resources sold in themarket come from? Some items, andthe people selling them, may travellong distances from other vegetationtypes, with very specific products.

• Is species substitution occurring and ifso, why? In some cases, less scrupu-lous harvesters mix in similar lookingspecies to bulk up bundles ofharvested plant material when they sellto unsuspecting traders. In other cases,species substitution is a warning ‘flag’for increasing scarcity. In South Africa,for example, Zulu women harvestingmedicinal plants find it increasinglydifficult to obtain aromatic bark of theforest tree Ocotea bullata (Lauraceae).As a result, bark of two forest trees inthe same family (Cryptocarya latifoliaand C. myrtifolia), each with a similar

bark aroma to Ocotea bullata, aresubstituted and sold by urban herbtraders as the real thing.

As the volume of products sold varies fromone market to another, as well as byseason or on a daily and weekly basis,quick surveys may indicate the mainproducts sold, but may miss out importantdetails such as competition between differ-ent products. In a detailed medium-term(12 to 18 months) daily-sales survey ofpalm-wine sales in South-Eastern Africa,‘primary sale points’ were located, asexpected, within the Hyphaene palmecological zone. These were distinguishedfrom palm-wine sales outside the zone,where palm wine from primary salespoints was diluted with water for resale(see Figure 3.2). Monitoring resale pointswould have given a complete overestimateof sales volumes. Only later in the study,however, was it clear how palm-wine salesvaried not only weekly, but also with adistinctive decrease in sales when afermented beer from the popular wild fruitSclerocarya birrea was available – and thisdid not appear in markets at all, as it istraditionally never sold due to its culturaland religious importance (Cunningham,1990a,b; see Figure 2.2).

Applied Ethnobotany

Location and mapping of marketplaces

64

Marketplaces of different sizes andfunction in rural areas, or in cities, townsor villages of different sizes, can be identi-fied from maps or aerial photographs,depending upon the size of the study area.Initially, you might work from 1:250,000or 1:50,000 scale. In cities or towns, workon a smaller 1:10,000 scale.

Local knowledge is very important inidentifying rural periodic markets that are

not located in places that are obvious toan outsider – for example, at periodiccattle sales or on days when pensions arepaid out in rural areas. During thismapping process, you will also have theopportunity to categorize marketplaces indifferent ways. One example is the map ofmarkets along the Zaire River in CentralAfrica made by anthropologist Yuji Ankei(1985) (see Figure 3.3). For this reason, it



Figure 3.2 Case study example: survey form and involvement of selected local people inmonitoring palm-wine sales

65


Applied Ethnobotany

66

BOX 3.1 CHECKLIST: ETHNOBOTANICAL SURVEYS OF

MARKETPLACES

Whether you are dealing with the rural harvest (‘supply’) end or the urban retail(‘demand’) end of the marketing chain (or possibly supply and demand at points alongthe marketing chain), a systematic approach is useful in developing an understanding ofmarketplaces and the networks of resource supply and demand that support them. Thischecklist assumes a stepwise process, from a regional ‘coarse grain’ approach downwardsto more detailed surveys. The cut-off point will be determined by time and funding. Insome cases, you may be starting from scratch. In others, you may already havebackground information on which species are commercially harvested, either from fieldobservation in local markets, discussions, participatory rural appraisal (PRA) methods,standard interview surveys or informant indexing methods of species/vegetation utilityor cultural importance (see Chapter 2). This can put information from field observationsas well as rapid, broad-scale (PRA, RRA) or detailed (interview, informal indexing)methods into geographical perspective. The following steps are recommended.

Map the location and names of marketplaces in your study area

This should be transferred onto an appropriate scale map which shows transportnetworks and topography. Regional studies may be 1:500,000 – 1:250,000 scale; smallerstudies 1:50,000 scale. With urban studies, a 1:2500 –1:10,000 is more appropriate. Makea preliminary note on market location compared to size of settlement (‘central place’).Are these likely to be regional, central, intermediate, standard or minor markets? Doany of the bulking centres operate for international trade? International trade for exportmarkets needs to be distinguished as a special (and often anomalous) category. If so, arethere trade statistics, such as from private companies, customs or forest departments?

Record market schedules for each market

Record the time of day, week and month. Local knowledge is very useful in this process– for example, from traders or buyers, or according to boat schedules or timing ofmarkets at periodic cattle sales. Plan your visits according to these market schedules.Are they periodic or permanent daily markets? If they are periodic markets, on whichdays are they open and with what frequency? On open days, what time of day wouldmarket activity peak?

Visit the marketplaces before designing survey forms

Have a look at what is being sold, by whom and where. Bear in mind that this is an initialvisit that will give you a good idea of the scale of sale and the time/funding commitmentthat would be required. If possible, meet the person controlling the market to explainthe reason for your visit. If the focus of your study is trade that is only marginally legal(for example, it contravenes forestry or conservation regulations), you will need to planeven this initial survey well in advance with local support and credibility. Do wild ormanaged plant resources feature at all, and if not, why not? Are they sold at small special-ist sale points (charcoal, building materials, fuel) or in larger markets dominated by goodssuch as clothing or foodstuffs? If you are in large markets, it would be too time consum-ing to count all sellers at this stage, although you may be able to count the number ofpeople selling the products that are the focus of your study (herbal medicines, wild fruits,



67

crafts). Where are sellers positioned in the market: in permanent stalls, under temporaryshelters, on the ground in the market or outside of the market (where they may not haveto pay fees)? Are the sellers men or women? Are they harvester-sellers, long-distancetraders, itinerant traders, travelling merchants or permanent sellers? Do the quantitieslook as if this is a bulk-sale centre (wholesale) or bulk-breaking centre (retail)? Would itbe possible to categorize a range of ‘standard’ bundles/bags/bottle sizes? Keep in mindcriteria that characterize different types of markets and sellers.

Logistics, project planning

Plan your work on the basis of available time and funding. At this stage, you mayalready start to wonder whether you are being too ambitious or not. Using informa-tion from the guidelines above, make an initial grouping of markets according tolocation, size and market schedule. You will then have a more realistic basis on whichto plan more detailed work. Are you being too ambitious regarding the number ofmarketplaces? If so, you may need to: focus on wholesale (bulk buying) markets andsubsample retail markets; or scale down the size of the study area.

Have you allowed for seasonal differences that affect wild plants on sale and theirprices? You may be recording alone in a broad-scale survey or, if you have the funds,you may decide to recruit reliable local assistance with the market surveys. Work outpractical forms for the survey. Depending upon the sensitivity and credibility of yourstudy, you may/may not want to record uses and prices at this stage – they create suspi-cion. Prices are also likely to be inflated for outsider researchers and are best recordedby local researchers/assistants. Follow up initial visual surveys with longer-term detailedwork. Count and categorize sellers. Categorize the type of market. Survey species beingsold by each seller and collect voucher specimens, parts of plants used and local names;record whether the material is sold fresh or dried. Compare with data from other surveymethods. Select key species. You need to decide what your focus will be: value, volumeor vulnerability, or a combination of these. Key issues in group/individual interviews arethe interrelated issues of price, scarcity and popularity, source areas and long-distancetrade. Is there a ‘moving front’ of depletion where harvesters overexploit a resource-rich frontier then move on, or not?

Choose ‘indicator’ species for field assessment and/or monitoring

Focus on populations of species that are indicators of demand which exceed their capac-ity to regenerate after harvesting. In some cases, you will be able to overlay the map ofmarketplaces onto a map showing distribution of these ‘indicator’ species (see Box 3.2and Chapters 4 to 6). Field check harvester-seller or trader perceptions of scarcity in thefollowing different sites:

• the major harvest areas where resources are considered to be ‘finished’;• ‘frontier’ sites which have recently become a focus of harvest; and• ‘indicator’ species populations beyond the ‘frontier’, where commercial harvest has

yet to take place.

For a discussion of changes in resource tenure with commercialization and competitionfrom outsiders, see ‘Social Surveys’, Chapter 7; for methods used for damage assess-ment, biomass comparison, size-class selection, increased intensity/frequency of harvest,see Chapters 4 and 5.


is important to select local people as keyhelpers to find out about the location andschedules of markets, who sells whichproducts, and where they, and theproducts they sell, come from.

In some cases, marketplaces andmarket schedules are easily identified. Inothers, location of marketplaces can beextremely difficult, particularly wheremajor river systems rather than roads arethe main transport route. AnthropologistChristine Padoch, for example, found thatthere were almost no stable marketplaceswholesaling wild-collected plants inIquitos in the Peruvian Amazon: mostwholesale buying was done on movingriver boats before they entered any port.Informal markets also shifted with theseasonal rise and fall of the river (Padoch,1988). A similar situation probablyapplies on the Zaire River.

Local knowledge, including advicefrom buyers and sellers themselves, can beuseful in resolving these challenges to fieldresearch. Despite the complexity ofmarketing in Iquitos, Christine Padochwas able to focus on key species such asthe aguaje (Mauritia flexuosa) palm, andidentify the wholesale (land-based) marketfor these fruits where they were sold inopen 50kg sacks. This enabled her toquantify the more than 700 metric tonnesof aguaje palm fruit sold each day inIquitos and estimate that about 500people, primarily women, relied upon thissale as a source of funds.

Records of product prices recordedwithin the same season over longer timespans can also tell a story, sometimes‘flagging’ overharvesting as prices risewith increasing scarcity. Working in thesame market in Iquitos, for example,botanist Rodolfo Vasquez has also beenable to carry out regular surveys of wild-collected plants, and the harvestingmethods and the retail prices paid for themover a period of more than ten years,

highlighting species such as aguaje(Mauritia flexuosa) palms in this process(Vasquez and Gentry, 1989).

Conceptual framework: locationtheory, central places and resource

use

Two simple but useful theoretical modelsare still widely used by economic geogra-phers to explain the location of centralplaces in a landscape and the activitiesaround them. These two models haveprimarily been used by geographers inter-ested in urban or agricultural planning.Despite the simplicity of these models andthe much greater complexity of harvestingwild plant resources compared to agricul-tural or urban expansion, they are a usefulbasis for considering patterns of resourceharvesting in response to demand fromgrowing towns and cities.

The first, the von Thünen model, wasdeveloped in the early 1800s by a youngGerman farmer called Johann von Thünento determine the best way of using land onhis farming estate. This led to the publica-tion of a book called The Isolated State,which developed a theory of the mosteconomical land use around a city thatformed the only market in an isolated, flatplain. Assuming that farmers wanted tomaximize their income and that transportcosts increased with distance, he visualizeda series of concentric land-use zones aroundthe city, with the most intensive activityclosest to the market. He also considered asituation where transport along a river orgood road would influence land use, inbands of decreasing intensity in parallelzones away from the river or road.

The second, the Christaller-Löschmodel, was independently developed inGermany by two economic geographers,Walter Christaller and August Lösch in the1930s (Christaller, 1966; Lösch, 1954).While von Thünen was interested in

Applied Ethnobotany

68



Source: Ankei, 1985

Figure 3.3 Map of markets along the Zaire River in Central Africa

69

50km0KASONGO

KISOMBO

KINDU

Kasongo-RiveMulenda

Lusuna

Lubenga

Kankumba

Malekeza

Lufubu

Malungu

Nyangimongi

Kabondo

Libende

Kimbombo-Rive

Kiyungi

Kasuwa

LowerKatalama

Kambelembele

Lumbuyo

Lueki

Mulamba-Shomola

KamimbiMakula

Katangela

MuumbiOkoko

Kitulu

Mulumbila

Lopokele

Elkila BukindiYengola

Mukoko

Lufaya

Kisubi

LotemoLokandu

Zaire River

Markets practising more or less barter

Cash sale markets

Central markets in capital towns

Rapids

Rapids

ZaireRiver

Railway


agricultural activity, Christaller and Löschwere interested in patterns of retail andwholesale markets. The Christaller-Löschmodel formed the basis of a detailed analy-sis of the distribution of agrarian marketsacross landscapes by anthropologist GWilliam Skinner, underpinned by researchon rural markets and social organizationwhich he conducted in China in1949–1950. In his research, described intwo seminal papers (Skinner, 1964, 1965),he was able to draw upon the long histor-ical records of rural markets to developconceptual models of markets and market-ing, and through this to develop ourunderstanding of the transformation ofagricultural societies into modern indus-trial society. One of the steps he took wasto develop and work from a simpleconceptual model of how markets areorganized across the landscape.

In common with von Thünen’s concep-tual model, Skinner assumed that themarkets developed on a flat plain whereresources were evenly distributed. Underthese circumstances, market towns wouldalso be evenly distributed across thislandscape, with the marketing area increas-ing until the service areas of each centralplace formed a series of regular hexagons.

How do these models apply topatterns of wild plant use? Although theassumption of a flat landscape withhomogeneous vegetation rarely occurs, thevon Thünen type model has been appliedwhere there is high-volume harvest arounda central point. Patterns of intensivegrazing around permanent water points,or fuelwood harvesting in arid and semi-arid areas, are two examples. Thisconceptual model has also become gener-ally accepted as a feature of fuelwoodoverexploitation, where an ‘urban energycrisis’ is created as treeless zones expandoutwards around cities that depend mainlyon fuelwood, with fuel getting more andmore expensive. One example of its appli-

cation is in Turi Digernes’s (1979) study offuelwood use around a town in Sudan,where fuelwood harvesters move increas-ingly outwards to meet urban fuelwoodneeds.

As economic geographers and anthro-pologists using these models haveacknowledged, however, things are gener-ally more complicated than this. Firstly,the towns or cities that create the marketfor agricultural products or fuelwood arerarely located on an isotropic plain (a flatarea where resources are evenly spread inall directions). Patterns of resource deple-tion are affected by topography, climate,soil fertility and vegetation type. They arealso affected by transport routes, whetherthese are paths, rivers or roads, that makeit easier to transport marketable resources.Lastly, socio-political factors, such asresource and land tenure, play a veryimportant role in whether overexploita-tion occurs or not. This is as important asthe biological factors covered in Chapters4 and 5, or the commercial harvesting thatis the focus of this chapter. For this reason,resource tenure is the major focus ofChapter 7.

Hierarchical classification: marketsizes, schedules and sellers

Systematic classification of market areas,market schedules and the traders thatsupply them is an important step in under-standing the network of harvesting andtrade in wild plant resources. Although thefocus of ethnobotanical or resourcemanagement studies will be on marketsthat sell wild plant resources, an under-standing of market networks within thestudy area can raise interesting questionsand lead to useful insights into why somemarkets sell wild plant resources andothers do not.

The process of mapping and classify-ing markets can help to guide research

Applied Ethnobotany

70


design, providing a basis for decidingwhether to sample all markets in a smallstudy area or which markets to subsampleif there are too many involved for detailedethnobotanical surveys. It can also impartsome predictive value on what to expectfrom cultural, socio-economic or environ-mental change. Although many moredetailed studies of markets have beenconducted in agricultural societies than ofwild plants in trade, they provide usefulbackground to developing a systematicapproach to ethnobotanical surveys inmarketplaces.

Marketplaces often evolve from barterand trade at small ‘nodal points’ at rivercrossings or road junctions. Some of these‘nodes’ become what geographers callnucleated settlements, developing intotowns and cities (see Figure 3.4).

The towns, cities and other settlementswhich support markets are called centralplaces by economic geographers. Ascentral places grow, the markets withinthem exert a stronger pull on rural

resources. Over the past century, there hasbeen an unprecedented flood of peoplemoving from rural to urban areas. In1800, for example, there were only 50cities of more than 100,000 people. By1950, there were 900, and by the end ofthe century more than a quarter of theglobal population will live in cities exceed-ing one million people.

Today, the highest rate of urbaniza-tion, 6 per cent per year, is in sub-SaharanAfrica. In South Asia, the rate is 4 per centper year. Godoy and Bawa (1993) suggestthat economic development ‘encouragesrural to urban migration, lowers popula-tion growth, and supports moreproductive agriculture, all of which shoulddecrease pressure on the forest as a sourceof livelihood’.

The urban-rural divide is rarely clearcut, however, and cultural and family tiesto rural communities and rural resourcesare often strong. Urbanization may resultin a reduction in the diversity and quantityof some wild plant resources in trade as


Figure 3.4 Sequence of a ‘node’ becoming what geographers call a ‘nucleated’ settlement, anddeveloping into a town and then a city

71

1900 1930 1960

1970 1990


people enter the cash economy and alter-native foods, utensils or roofing materialsbecome available. In other cases, however,the continued cultural and economicimportance of wild plants to urban peopleis evident in continued trade in plantresources from rural to urban areas.Common examples in many cities inAfrica, Asia and Latin America are theinformal sector trade in fuelwood,charcoal, medicinal plants, basketry andconstruction materials such as bamboo orhardwood poles (see Figure 3.5).

In these cases, urbanization has tendedto increase rather than reduce the demand

for wild plant resources, resulting in acommercial trade that stimulates overex-ploitation. Cities and towns arecharacterized by the people that live inthem, and part of this identity is reflectedin the types of plant resource that are usedand sold. A good example is the interviewsurvey of chewing stick use in Ghana(Adu-Tutu et al, 1979), which showed thehigh proportion of chewing stickscommercially harvested from the wild tosupply people from different-sized settle-ments and from different educationalbackgrounds (see Figure 3.6).

Applied Ethnobotany

Figure 3.5 (a) A seller of hardwood poles from mangrove and coastal forest tree species at a dailymarket in Malindi, Kenya. (b) A woman harvester-seller of edible wild spinach, Cucumella

cinerea, at a market in Maputo, Mozambique

72


Depending upon the size of your studyarea, several methods could be used incombination to identify where thesemarkets are located, what they are called,when they are held, and which market-places sell wild plant resources.

Size of market areas

In general, the types of goods and functionsof a market increase in proportion to thenumber of people in the market area. Assettlements get bigger, more and morepeople are attracted from greater andgreater distances into the central marketpoint. In a study of rural marketingsystems in rural China, using informationfrom interviews and historical records,Skinner (1964, 1965) developed a five-levelclassification of markets, based on theirservice areas; this has been widely used byeconomic geographers.

The basis for this regional analysis wasthe central place theory developed by

Christaller (1966) and Lösch (1954),described earlier in this chapter. Althoughcentral place theory is a theory of retaildistribution centres, rather than of whole-sale distribution (which is often a feature ofthe initial sale of wild plant products afterharvest), it still provides useful backgroundto understanding patterns of supply anddemand for wild plant resources. Skinnerconcentrated on analysing rural marketsand the flow of agricultural produce fromrural areas to city markets, or the flow ofgoods from cities to rural markets. Herecognized different levels of market in anapproach which has since been used bymany other economic geographers seekingto understand marketplaces within anexchange system:

• regional markets, which cover thelargest area and which generallysupport several marketplaces;

• central markets, usually found at astrategic point in the transportation


Characteristics of markets

73

Source: Adu-Tutu et al, 1979

Figure 3.6 An interview survey of chewing stick use in Ghana showed the high proportion ofchewing sticks commercially harvested from the wild to supply people from (a) different-sized

settlements and (b) different educational backgrounds

100

80

60

40

20

0

Percentage

Villages

Buying

Collecting

(a) Settlement type

Markets Cities

100

80

60

40

20

0

Percentage

Illiterate

(b) Educational status

Primary,middle school

Post middleschool

Buying

Collecting


network, where wholesaling takesplace and which also can be the site ofseveral marketplaces;

• intermediate markets, which werenamed for their intermediate positionin the flow of goods downwards forlocalized rural use, and upwards tocentral and regional markets;

• standard markets, which are the endpoint for sale of imported items fromcities and towns, in addition to beingplaces where local exchange takesplace; standard markets are also thestarting point for the flow of agricul-tural goods and crafts into largercentral or regional markets;

• minor markets, characterized by localexchange of goods between localpeople, which Skinner termed‘horizontal exchange’; minor marketsdeal with few goods that are importedinto the area – Skinner also calledthese ‘incipient standard’ markets asthey represented an early stage in thedevelopment of markets.

It is well worth checking your interpreta-tion of the regional hierarchy of marketswith the perceptions of local key infor-mants. Local people in rural or urban areascan be an invaluable source of informationabout the hierarchical classification ofmarket areas and market types, and aboutmarket schedules and functions. U L Ukwuused this approach in studying and classify-ing markets in eastern Nigeria, describinghow ‘in every locality the principal markets

are known by common repute and areusually ranked by persons interviewed inmuch the same order…Responses toquestionnaires on market visiting habitsalso revealed the fact that certain marketsstand well above others in the extent oftheir service area.’

Skinner’s hierarchical classificationsystem fitted well in the study thateconomic geographer Wayne McKim(1972) made of periodic markets in north-eastern Ghana (see Figure 3.7 and Table3.2), with many minor markets formingthe base of a hierarchical pyramid and, inthis case, with a single regional market atthe top.

Not all marketing systems fit Skinner’shierarchical model as neatly, but it isthrough a process of systematic analysisthat ‘anomalous’ cases are recognized,raising useful questions and insights intowhy they do not fit an expected pattern.As a result, anthropologist Carol A Smith(1985) found that regional markets inGuatemala were ‘anomalous’ (see Table3.3) and did not fit the type of hierarchyshown in Table 3.2, with a higher numberof central and intermediate level marketsthan expected.

Systematic analysis of marketplacescan therefore lead to useful questionsabout their social and cultural featureswhich may otherwise have not been asked.As Smith (1985) points out in a very usefulpaper describing the methods she used toanalyse regional market systems:

Applied Ethnobotany

74

Table 3.2 Market types in north-eastern Ghana and Guatemala classified on Skinner’s (1964)hierarchical system

Market type Ghana Guatemala Size Frequency Species sold

Regional 1 1Central 3 15Intermediate 10 26Standard 20 32Minor 34 78

Sources: McKim (1972), Smith (1985), Skinner (1964)


‘The measures I used for discrimi-nating among marketplaces weresimple ones, each of which reduceda great deal of information to a fewcrucial elements. The first andmost basic measure was that of sizeor level in the marketing hierarchy.But the information contained inthis measure, while important andnecessary for all aspects of theanalysis, was not sufficient toaccount for either the layout or thefunctioning of Guatemala’s marketeconomy. To achieve the latterresult, I also had to consider theseller composition of marketplaces.

This measure involved only asimple trichotomization (divisioninto three) of sellers by theirfunctions, which could be used tocategorize market centres bydominant seller type (leading tofour basic marketplace types), onthe first step; and an even simplerdichotomization (division intotwo) of the commodity movement,which could be used to character-ize market centres by wholesalefunction (leading to two wholesalemarket types), on the second step.But while simple, this proceduredivided market centres into two


Source: McKim, 1972

Figure 3.7 (a) Human population density by local authority: 1 Western Gonja; 2 Eastern Gonja;3 Namumba; 4 Eastern Dagomba; 5 Tamale; 6 Western Dagomba; 7 South Mamprusi; 8 Builsa;

9 Kasani; 10 Frafra; 11 Kusasi. (b) Location of markets in the same area showing how mostmarkets are located in high population density areas. The regional market (level 5) is in Tamale

and the central markets (level 4) are in Bolgatanga, Bawku and Yendi

75

Navrongo

Bolgatanga

Walewale

Nalerigu

Bawku Fusiga

Chereponi

Gushegu

TAMALE YendiYendi

Saboba

Zabzugu

Bimbila

TAMALE

Bawku

89 10 11

7

6

5

4

1

23

1000 km

Under 10

10–19

20–38

39–58

59–78

Urban

Persons persquarekilometre

Level of hierarchy

3 day markets

6 day markets

1 2 3

4 5

(a) (b)


groups that made sense in terms ofGuatemala’s production character-istics, ethnic characteristics, rural–urban characteristics and trading(wholesaling versus retailing)characteristics.’

This characterization of types of seller as away of distinguishing different types ofmarketplace is described in ‘MarketingChains and Types of Seller’ below.

Marketplace micro-geography:politics and use of space

Categorizing how markets are organizedin time and across landscapes or withinurban areas provides a useful basis forplanning surveys within markets. Whenyou are working within each marketplace,it can be helpful to discern patterns ofwhere people are located and why.

It is also important to start by findingout who is in charge of the marketplace.Even in the most informal and apparentlychaotic marketplaces, there is often formalcontrol and pattern in the location ofsellers. In common with any ethnobotani-cal research, detailed market surveys canrarely be done efficiently without localpermission.

As centres of economic activity andvillage life, markets are often controlled bylocal government or traditional leadership

through a system of fees. Barter marketsalong the Zaire River, for example, arecontrolled by the traditional chief (Ankei,1985). In Uganda, many marketplaces arecontrolled by local committees linked tovillage-level government. Obtainingpermission to work in the market is impor-tant. The process can also lead todiscussions and useful insights into marketfees, market schedules and the history orgeography of the marketplace. Whilepeople may seem to be selling goods atrandom, this is rarely the case. Instead,they are located in different parts of themarketplace for a variety of reasons. Sellersmay be located according to whetherpeople can afford permanent stalls or not,on the basis of where they come from, orwhat they are selling. They may also belocated to avoid social conflict. If beer issold in markets (and this may be forbiddenin some areas), then beer sellers are locatedin a place where patrons are least likely todisrupt the rest of the market.

Less obviously, people selling tradi-tional medicines may also be located on theperiphery of the market to avoid socialdisruption resulting from ‘bodily pollution’caused by proximity to supernaturallypotent medicines.

Travelling sellers from the same districtand cultural background frequently also sittogether, although this may not be obviousto the outsider. These patterns give impor-

Applied Ethnobotany

76

Table 3.3 Distribution of marketplace levels by marketplace types in Guatemala

Type Type 1: >50% Type 2: >50% Type 3: >50% Type 4: <50% Totallevel long-distance producer- middlemen of any seller type

traders sellers

Regional 1 0 0 0 1Central 0 6 2 7 15Intermediate 0 12 5 9 26Standard 0 21 6 5 32Minor 0 42 30 6 78Totals 1 81 43 27 152

Source: Smith, 1985



Source: Kurita, 1985

Figure 3.8 Location of sellers and interchange of goods between different cultural groups withdifferent systems of land use at a small market in Chesegon, Kenya

77

P

10 m

etre

s0

010

0

500

1000

Shill

ing

s p

er w

eek

Ch

eseg

on

Milk

Dri

ed m

eat

Liq

uo

r

Ban

anas

Tom

ato

es

Cas

sava

Tob

acco

Sug

ar c

ane

Oth

er

Mar

akw

etPa

sto

ral P

oko

t

Ag

ri P

oko

t

Ban

ana

selle

r

Cab

bag

e se

ller

Cas

sava

sel

ler

Mai

ze s

elle

r

Sug

ar c

ane

selle

r

Salt

sel

ler

Tom

ato

sel

ler

Tob

acco

sel

ler

Tree

or

its

roo

ts

PTh

e p

asto

ral P

oko

t co

nce

ntr

ated

pla

ce

VEHICLERO

A

D

toLo

mot

to

Tot

Ch

es

eg

on

Riv

er


tant background to designing or interpret-ing the results of sampling surveys. For thisreason, it can also be useful to make asketch map showing the location of differ-ent types of seller within individual

markets, even if it is only for your ownfield notes, as Kazauaki Kurita did for asmall rural market at Chesegon, Kenya (seeFigure 3.8).

Applied Ethnobotany

Market schedules

78

Identifying the hierarchy of market areasin terms of size makes it a lot easier togroup markets according to when they areheld. Timing of markets provides essentialbackground information for timing yourresearch. Before you start detailedresearch, you need to find out the mostappropriate times of the day, week ormonth to visit marketplaces.

If you are collecting plant specimensfor identification, it can be useful to timeyour visits at the start of the market, whenthe plant material is still fresh and beforesellers are too busy. The same applies ifyou are counting the numbers of bags,bundles or bottles of goods for sale. It ispointless arriving when half the goods aresold, fresh medicinal plants are wilted orwhen people are on their way home.

If you are counting sellers, you have totime your visits at the peak sale time.Whether you are studying periodic orpermanent daily markets, you also need tomake sure that the surveys are done at theappropriate time of day or night. Fewmarkets are all-day affairs with no peaktime. Peak times may occur in the earlymorning, at midday or at night. Whengeographer R J Bromley (1974) and hislocal assistants surveyed urban markets inQuito, Ecuador, for example, all marketcounts were done on market days whenactivity was at its peak between 9.00 amand 12.00 am. In a long-term study ofpalm-wine trade in rural South-EasternAfrica, sales peaked around 10.00 am sothat the palm wine could be transported,

diluted and resold before it fermented toomuch.

Bulk sales of the major medicinalplants at Umlazi, near Durban, SouthAfrica, also took place early in themorning so that urban herb traders couldhave their pick of fresh plant material.Sales also peaked on a Sunday or Monday,when over 300 gatherer-sellers arrivedfrom rural areas, with sales and numbersof people declining later in the week aswell as later each day. A smaller bulk-salepoint for medicinal plants in the same areastarted in the pitch dark, once a week,from 4.30 to 7.00 am, each seller with apile of medicinal plants lit by a tiny lampor candle. This was done so that gatherer-sellers could avoid legal problems from theconservation department and so that thesellers, all of them women, could get backhome, often to remote areas, to continuewith household and farming activities.

Economic geographers classifymarkets partly according to their periodic-ity – the frequency with which they areheld. Three basic types are recognized byBromley (1971) and are widely applicable.These are: special markets, which are theleast frequent and most irregular; periodicmarkets, which are held regularly on afixed schedule such as every few days,weekly or monthly; and daily (or perma-nent) markets, which are the higher levelmarkets characteristic of larger towns andurban centres.

Periodicity of markets depends uponwhether traditional or Western calendars


are used. Some parts of the world, such assouthern China or West Africa, whichhave historically high human populationdensities, centralized political structuresand concentrated settlements, also have along history of well-developed long-distance trade. In these cases, marketschedules can be based on ancient tradi-tional calendars such as lunar or solarcycles. In southern China, for example,Skinner (1964) found that market cycleswere based on the lunar hsün (‘decade’)which averaged 9.8 days and on solar orlunar fortnights, in addition to otherschedules. At the time his study wasconducted, market schedules were mostcommonly based on the lunar hsün or a12-day (duodenary) cycle on a fixedsequence of twelve chih (‘branches’).These then determine the marketingschedule (and the timing of research inmarkets). The duodenary cycle, for

example, would be a regular systemleading to 3-day, 6-day or 12-day weeks.William Skinner also found that market-ing cycles, in turn, were characteristic ofparticular regions of southern China.Duodenary cycles were limited to theupper drainage basins of the Hsi andHung River systems, while schedules ofmarkets along downstream plains anddeltas followed lunar (hsün) cycles.

By contrast, the market schedules ofmany African countries have been influ-enced by the calendar introduced duringthe colonial period, based on a 7-day week(with a day off on Sunday). In a study of154 markets in Uganda, for example,geographer Charles Good (1975) foundthat 81 per cent of markets occurred every14 days, 16 per cent every 7 days and 3per cent every 28 days, with 14-daymarkets attended by the most people andhaving the widest range of goods.


Source: Hill and Smith, 1972

Figure 3.9 Seven-day market schedules are common in West Africa

79

Mixed

Eight-day

Seven-day

Six-day

Five-day

Four-day

Three-day

Two-day

6000 kmAbidjan

Dakar

Accra

Lagos

Freetown

BENINLIBERIA

GHANACOTED’IVOIRE

BURKINAFASO

M A L I

N I G E R I A

CAMEROON

TOGO

SENEGALTHEGAMBIA

GUINEA-BISSAU GUINEA

SIERRALEONE

N I G E R


In South-Eastern Africa, major saledays for palm wine were on a Monday,Wednesday and Friday, with only localsales taking place on intermediate days,and virtually none on a Sunday(Cunningham, 1990a, b). Seven-day marketschedules are also common in West Africa(see Figure 3.9), with Islamic limitations onmarket activity on religious days and mainmarket activity on other days of the week(Hill and Smith, 1972). Markets in EastAfrica are similarly influenced by theseven-day week or may take place every 7,14 or 28 days. A seven-day schedule intro-duced by Belgian colonists is also stillfollowed in barter markets in Zaire, withonly the trace of a traditional market cycleremaining. Anthropologist Yuji Ankei(1985), in a detailed study of bartermarkets along the Zaire River (see Figure3.3), found that the majority of marketswere held every 7 or 14 days, with only onefollowing a traditional four-day cycle. It isthese schedules that are followed by theitinerant traders or travelling merchantsdescribed further on in this chapter.

As human populations grow, anddemand becomes more concentrated, somarket schedules are compressed.Improved transport networks and highernumbers of people ‘shrink’ market sched-

ules, shifting them from being periodic topermanent daily markets (see Figure 3.10).

Traditional market days are remark-ably resilient, however, and this also needsto be borne in mind when conductingsurveys in large urban centres. In a studyof periodic markets in South Korea,Siyoung Park (1981) found that thenumbers of people in daily, permanentmarkets increased several times on a five-day schedule. This trend is also a featureof urban markets in Lagos, Nigeria (Sadaet al, 1978).

Seasonal schedules and annualtrends of plant products and

harvesters

You also need to be aware of the seasonalbias that can result from short-termsurveys of marketplaces, and ideally youshould visit markets at regular weekly ormonthly intervals during the year. The linkwith fruiting times of wild fruits isobvious, but seasonality also applies to thetiming and sale of items such as fresh wildspinach, wild mushrooms or insects. Intheir study of the Mercado de Sonora inMexico City, for example, ethnobotanistsBob Bye and Edelmira Linares (1985)found that while some cultivated edible

Applied Ethnobotany

Source: Hodder and Lee, 1974

Figure 3.10 The sequential development from periodic (P) to daily (D) markets as paths to themarketplace develop into tracks and then into roads

80

P P D P D D

Path

P = periodic market D = daily market


species such as ‘nopal’ (Opuntia spp) wereavailable throughout the year, others, suchas the aromatic food additive ‘pericon’(Tagetes lucida), were only seasonallyavailable, while Yucca and Agave flowersappeared in the market for only a fewweeks (see Figure 3.11).

To avoid this problem, Peruvianbotanist Rodolfo Vasquez visited theIquitos market virtually every week fornine years (Vasquez and Gentry, 1989),making valuable records of species soldand price fluctuations. If his team had notdone this they would have missed record-ing some of the most importantwild-collected fruits, which appeared onthe market between June and October.

Seasonal differences may also occurfor less obvious reasons than growingseason. Sales of the popular traditionalmedicines Alepidea amatymbica and

Siphonochilus aethiopicus declinedsharply at the wholesale marketplaces Isurveyed in Durban, South Africa, inspring and summer. This was due to ataboo against collecting these speciesduring the growing season, reinforced bythe belief that breaking the taboo wouldresult in lightning striking the household.

Conversely, sales of bulbous speciessuch as Scilla natalensis and Eucomisautumnalis in South Africa increased insummer when plants could be readilylocated and dug up from mountain grass-lands; but occurrence in markets declinedsharply in the winter dry season when theleaves of these species died or were burnedoff in fires. Social issues also affected palm-wine sales in the region. Firstly, there was ahigher consumption of palm wine atChristmas time, not only because of the hotsummer days but because it was a time


81

Note: The importance value is relative and ranges from 0 (absent) to 5 (abundant).Source: modified from Bye and Linares, 1985

Figure 3.11 Changes in the importance of some plant species sold in Mercado de Sonora, Mexico City, over a year

6

5

4

3

2

1

0

Importance value (relative)

J

Nopal (Opuntia species)

Edible flowers of maguey (Agave)and yuca (Yucca) species

F M A M J J A S O N D

Piru (Schinus molle)

Flor de manita (Chiranthodendronpentadactylon flowers)

Fresno (Fraxinus species)

Pericon (Tagetes lucida fresh)


when many men employed as migrantworkers returned home. Secondly, palm-wine sales dropped drastically during themarula (Sclerocarya birrea) fruiting season,

when a tastier and more popular beer isbrewed by most households (see Figure 2.2).

Applied Ethnobotany

Marketing chains and types of seller

82

Informal sector traders usually have littleaccess to capital and earn money fromlabour-intensive enterprises, forming partof a ‘hidden economy’ from first harvestto final sale. International trade in wildplant products is generally less complex,with fewer species harvested and greater‘visibility’ in the records of private compa-nies, customs or forest departments.

When harvesting takes place forcommercial trade, there is often a largenumber of people involved, from initialharvest through processing, sale and resale.This is known as a marketing chain.Different types of seller feature at differentparts of the marketing chain, from initialbulk gathering of plants from the wild,through to processing, sale and resale.

As mentioned earlier, the extent ofdemand for goods influences the size andtiming of markets, and these factors inturn influence the types of seller atmarkets. Several economic geographersand anthropologists have categorized themarketplace, using types of sellers asindicators. This method has mainly beenapplied in studies of agrarian societies.The principles and approaches can also beapplied, however, in studies of wild plantharvest, sale and resale.

Siyoung Park (1981), in a study ofperiodic rural markets in South Korea,distinguished three main types of seller,representing a transition from mobile tosedentary traders. Firstly, there were itiner-ant traders, who move from one smallmarket to another, returning home at theend of the market cycle. This is not an easylife, but one of the few ways in whichitinerant traders are able to stay in

business is through travel to periodicmarkets over a wide area (see Figure3.12a). Itinerant traders generally occur inareas with low human population densityor where there is less demand for a special-ist resource.

The second category consisted oftravelling merchants, based at home butwho commute to two to four dailymarkets, which generally are a feature oftowns or cities rather than small villages.Park found that 35 per cent of the travel-ling merchants visited three differentmarkets, 30 per cent visited two differentmarkets, 19 per cent visited four differentmarkets and only 16 per cent visited adifferent market each day of the five-daymarket week.

The third category comprised perma-nent sellers, who travel from home topermanent stalls at the market (see Figure3.12b), and who are a feature of largermarketplaces rather than smaller ones.

In her analysis of regional marketsystems in Guatemala, Carol Smith (1985)first categorized marketplaces accordingto the quantity and variety of goods beingsold. She then recognized three categoriesof trader, each of whom deals withcommodities in different ways. The firstcategory, at the start of the marketingchain, comprised producer-sellers, whosold local goods, mainly food items, mostof which they grew themselves (ratherthan buying them from other markets).The second category comprised middle-men (and women) who sold goods thathad been produced elsewhere; and thethird category consisted of the long-distance traders.



Figure 3.12 Patterns of market visits, with African medicinal plant sales as an example. (a) Anitinerant trader – a herbalist selling at village level, Malawi. (b) Permanent market seller atOwino herbal market, Kampala, Uganda. (c) Breaking down bulk medicinal plant supplies

collected by women at a market in Bouake, Côte d’Ivoire. (d) A permanent herb trader’s shop(next to doctor’s quarters!) in Pietermaritzburg, South Africa

83


She then went on to categorizemarketplaces according to the proportionsof different traders represented withinthem, using these three types of traderindicators. Markets with many producer-sellers indicated markets in an area of localexchange, or where there was bulking ofgoods redistribution upwards along themarketing chain. Marketplaces dominatedby local middlemen generally sold goodsfor local use or distribution down themarketing chain, while marketplacesdominated by long-distance tradersindicated places where there was whole-sale exchange among these traders. Onthis basis, Smith distinguished four typesof market:

• type 1: markets with >50 per centlong-distance traders;

• type 2: markets with >50 per centproducer-sellers;

• type 3: markets with >50 per centmiddlemen;

• type 4: markets where there weresellers of all sorts, and where nocategory of trader was dominant.

This facilitated the grouping of market-places into two functional types:

• bulking centres, which are mainlyrural marketplaces of types 1 and 2,supplied by producer-sellers or long-distance traders;

• bulk-breaking (or dispersing) centres,which receive goods from bulkingcentres and disperse them from there,characterized by types 3 or 4 market-places; bulk-breaking is the term usedto describe the division of thecommodity into smaller amounts.

By bringing together both economic andethnographic components, this systematicclassification enabled Smith to developseveral hypotheses about marketplace

specialization in the Guatemalan studyarea, illustrating the close relationshipbetween the role of Indian or ‘Ladino’Guatemalans in the economy.

The principles behind this systematicanalysis of regional marketing systems arecertainly applicable to the harvesting ofwild plants, and they also need to be morewidely applied in ethnobotanical surveysof market networks. Several categories ofseller are evident in the African traditionalmedicine trade – for example, in market-places of different size or centrality.

Firstly, there are the middlemen, repre-sented by itinerant traders doing a circuitof several periodic markets (see Figure3.12a). Next would be categories ofpermanent seller, who visit a singlemarketplace. Rural people (often women)regularly supply large regional or centralmarkets as bulk-supply wholesalers (seeFigure 3.12c). There are also two types ofpermanent retail seller: specialist sellers incentral markets, and permanent traderswith large storage space who work inretail and bulk-breaking in regionalmarkets (see Figure 3.12d).

Marketing chains and wild plantresources

Although hierarchical classification ofmarkets by size and centrality, or ofmarketplaces by function and seller type,is useful, in the case of ethnobotanicalsurveys – which have a focus on wildplants instead of on all products sold – weare faced with a greater level of complex-ity on at least two levels.

Firstly, for wild or managed species,particularly those traded to regional orcentral markets, the marketing chains maybe complex and very long. The marketingchain from harvest to final use of sal(Shorea robusta) leaves harvested fromwoodland in West Bengal, India, to makeplates is a good example of how long such

Applied Ethnobotany

84


a chain can be (see Figure 3.13).Secondly, although species entering

commercial trade represent a ‘short list’ ofa far greater diversity of species used inrural areas, the number of species involvedin species-specific harvest is far greaterthan an economic geographer or anthro-pologist would encounter in studying thesale of clothing, manufactured goods oragricultural crops. In their study of ediblefruits sold in the marketplaces of Iquitos inthe Peruvian Amazon, for example,Vasquez and Gentry (1990) recorded over57 wild-collected fruit species being sold.In Germany, Lange and Schippmann(1997) have documented 1543 medicinalplant species comprising 854 genera in 223families in the import or export trade. InSouth Africa, 400 to 500 species are sold

for traditional medicines (Cunningham,1990; Williams, 1996), and in north-westChina alone, Pei-Sheigji, Li Yanhui and YinShuze (1990) recorded 574 medicinal plantspecies traded in local markets. Bycontrast, the considerably fewer speciessold for building material, thatch, palmwine, woodcarving and fuelwood orcharcoal make analysis easier.

The complexity of marketing chains isimportant to keep in mind when recordingprices for wild plant resources being sold,or when classifying markets on the basisof the types of sellers within them. Thepeople you see selling wild plant resourcesmay be way down the chain of sale andresale, influencing your records of theprices, gender of sellers, quantities sold orinformation on where they obtained the


Source: Poffenberger et al, 1991

Figure 3.13 A marketing chain in the sale of plates made from sal (Shorea robusta) leavesharvested from woodland in West Bengal, India

85

1Leaf

collection

2Leaf

stitching

3Middleman’s

depot

6Retailers

5Wholesalers

4Contractor

7Consumers

No charge for collection

Rs8 per bundle(1000 plates)

Rs8.50 per bundle

Royalty toforest departmentat Rs416 per truck

Rs17–18 perbundle

(100 thals)

Rs20–24 perbundle

(100 thals)

Rs14–16 perbundle

(100 thals)


plants. Prices will obviously increase alongthe chain, with the lowest prices beingpaid for plant products to harvesters. It isuseful to record price increases along themarketing chain and the reasons. Thesemay include time or monetary costs ofprocessing, transport, fees at markets orrental of trading premises. Althoughincome from Prunus africana bark sales isan important source of revenue to villagersin Madagascar (in some cases generatingmore than 30 per cent of village revenue),the price paid to collectors is negligiblecompared to Madagascan middlemen.Price paid to bark harvesters also varies by60 per cent at different points of sale(Walter and Rokotonirina, 1995). InMexico, Paul Hersch-Martinez (1995)found that medicinal plant collectors onlyreceived an average 6.17 per cent of themedicinal plant consumer price. For thisreason, Shankar et al (1996) have recom-mended an alternative flow of amla(Phyllanthus emblica) fruit in India fromthe forest source area to the Indianconsumer, improving economic benefits tothe Soliga people involved as a means ofimproving household income while reduc-ing overharvesting of fruits.

Market vendors: number, genderand change

From the mapping exercise and initialvisits, you should have a good idea aboutwhich markets have the largest number ofsellers, and which are selling plants whole-sale or are retail centres. Unless themarkets are very large, you may also beable to count or estimate the total numberof vendors in each marketplace. With localassistance you should be able to get an ideaof the cultural and socio-economic statusof people visiting the market to buy thecategories of wild plants (such asfuelwood, building poles, crafts, charcoal,wild fruits, wild greens, medicinal plants)

on which your study will focus (see Box3.1).

Table 3.4 is an example of this stagefrom an ethnobotanical survey by geogra-pher Helmut Kloos and three Ethiopianstudents in their study of medicinal plantssold in marketplaces in central Ethiopia(Kloos, 1976/1977). Fifteen of the market-places they visited were in Addis Ababa,which at that time had a total populationof 800,000 people. As would be expectedin such a large regional centre, these weredaily markets. They also visited periodicmarkets in rural towns, with populationsas small as 1500 people.

Trends of increase or decrease innumber, gender or type of seller in therange from minor to regional markets canprovide useful insights on how demand islikely to change. As you can see fromTable 3.4, these marketplaces varied bothin size and in the cultural diversity ofpeople frequenting them. These are impor-tant features to record in marketplaces ofdifferent size or centrality, or as a market-place grows or shrinks in size over time. Itcan also be useful during a marketplacesurvey to group sellers in appropriatecategories, as Park and Smith did in theirstudies in South Korea and Guatemala (see‘Marketing Chains and Types of Seller’).Record whether people selling wild plantresources are permanent sellers, itineranttraders or travelling merchants. Are theyselling only one category of wild plantresource or are they selling wild plantsamongst other items?

In addition to socio-economic status,sellers frequently differ by ethnicity andgender. Helmut Kloos (1976/1977) notedthat the majority (95 per cent) of thepeople selling medicinal plants atEthiopian markets were women, most ofthem Gurage rather than Amhara women,due to the low social status of kosso(Hagenia abyssinica, Rosaceae) sellers intraditional Amhara society.

Applied Ethnobotany

86


Similarly, large numbers of women sellherbal medicines in the large regionalmarkets of Africa, such as in Abidjan,Côte d’Ivoire, or in South Africa. One ofthe main reasons for this is that men stopnon-specialist herbal medicine sales whenit becomes an increasingly marginaleconomic activity, persisting only as sellersof traditional medicines from animals, anactivity often closed to women. It is usefulto record trends of increase in number,changes in gender or type of sellers as onegoes from minor to regional markets. Thisprovides useful insights into how demandand marketing networks are changing.

In many developing countries, urban

growth has been particularly rapid sincethe 1950s. This has been accompanied bychanges in the number, gender and types ofsellers at marketplaces. In 1929 in Durban,South Africa, for example, there were onlytwo herbal traders and most people sellingherbal medicines at the eMatsheni marketwere men. By 1987, Durban was a regionalmarket for medicinal plants, with morethan 100 herbal traders supplied by over300 people commercially harvestingmedicinal plants from the wild. Virtuallyall of these harvester-sellers were women.By 1991 their numbers had increased toover 500 due to rising unemployment andrural poverty.


Inventory and frequency of plants on sale

Table 3.4 The top 10 medicinal plants sold in the markets in Ethiopia, showing number of sellersin 3 of the 15 markets sampled, including the total sellers/species for all markets, showing the

percentage of sellers of the 4 most popular species in the 3 main markets

Plant species Three markets in Addis Ababa Total (% ofMerkato Kirgos Selassie total)

Hagenia abyssinica 115 33 18 229 (72%)Embelia schimperi 101 26 14 219 (64%)Glinus lotoides 113 28 16 194 (73%)Silene macroselene 115 20 11 180 (85%)Echinops sp 112 17 8 178Withania somnifera 109 19 10 162Lepidum sativum 105 15 6 160Thymus serrulatus 95 14 12 155‘dingetenya’ 104 14 7 140Myrsine africana 87 9 3 115

Source: derived from Kloos, 1973

87

In ethnobotanical surveys of marketplacesaimed at identifying plant species underthreat, the question is not just ‘what isbeing sold?’ but also, ‘which species arebeing depleted by commercial trade?’ Ifdifferent species populations are beingdepleted, the next question is: ‘how candifferent species be prioritized?’ This lastquestion is introduced here and is thencovered in Chapters 4 to 6. Information

on the uses of wild-collected species whichare (or were) commonly being sold canprovide important insights into the socialissues which need to be addressed as partof the solution to overexploitation of wildstocks. Conversely, such insights alsoindicate the social benefits from plantsthat will no longer be available if overex-ploitation occurs. The medicinal plantspecies most frequently recorded by Kloos


(1976/1977, see Table 3.4), for example,are mainly used to treat internal parasiteinfestations (Hagenia abyssinica (‘kosso’);Embelia schimperi (‘enkoko’); Glinuslotoides (‘metere’); Croton macrostachys(‘bisanna’); Myrsine africana (‘kechemo’)).This reflects the social circumstances (diet,housing density, sanitation), and linksplant conservation to primary health careissues.

Demand is most likely to exceedrenewable supplies of species which aredestructively harvested and which are slowgrowing; reproduction of these species islimited, while their habitat requirementsare very specific. The shortlist of commer-cially harvested species can be furtherprioritized on the basis of this informa-tion, selecting commercially harvestedspecies with limited geographical distribu-tion that are most likely to be subjected todestructive harvesting (see Box 3.2).Marketplace surveys add to this informa-tion, enabling rapid assessments thatinclude species from a wide geographicalarea, and highlight species which shouldbe the focus of monitoring programmes.

In addition to knowing which speciesare sold, it is also important to glean infor-mation from traders on the sources,habitat and price per local unit (for laterconversion to a price/kg basis) of priorityspecies. This information can be analysedin different ways to study flow patternsfrom source areas to urban sales points,frequency of sales for different species orwhether the species sold are from wild,managed or domesticated plant popula-tions. Information on source areas (andareas not being harvested) is useful fordeciding on the best places to set up plotsto assess harvesting impact or to monitorthe effects of harvesting on indicatorspecies. Usually, this would be carried outthrough carefully located sample surveysto enable comparison of impacts onindicator species where: there was protec-

tion from human impact (for example incore conservation areas); harvesting wasunregulated; and where managed harvesttook place, if this occurred at all (seeChapters 5 and 6). Whether this informa-tion is gathered during the inventory ofspecies on sale or not can only be decidedon a case-by-case basis. In some cases, it ispossible to record information about theuses of commercially harvested species. Inothers, such as collection of informationon medicinal plant species, it can be highlysensitive and would best be left until alater period, or obtained from otherpublished studies.

In her survey of crop plants sold inGuatemala, Carol Smith used systematicmarketplace surveys to record the specieson sale, and then arranged them intohierarchical levels which reflected theirrelative demand (Smith, 1985). The samecan apply to wild plant species, with onecomplication: sometimes the most popularspecies no longer appear in marketsbecause they have been overexploited. Forthis reason, it is as important to determinewhich species are in demand but are nolonger sold, and to distinguish these fromthose currently being sold. It is useful toask vendors to free list species which theyconsider most expensive, are becomingincreasingly difficult to obtain and wheresubstitution of one species with another isoccurring, and why.

Helmut Kloos (1976/77) and hiscoworkers, in their study of medicinalplants in Ethiopian markets (see Table3.4), counted the number of sellers whowere selling different species in eachmarket, recording the total number ofpeople selling each species for all markets.Their data also illustrate a lesson which iswidely applicable for people who havelimited time and funds and need to focusmarket survey effort. All species soldwithin all markets surveyed were recordedwithin just four main markets, and the

Applied Ethnobotany

88



89

BOX 3.2 ETHNOBOTANICAL SURVEYS OF MARKETS

Ethnobotanical surveys of markets are the first steps in identifying species which are aconservation or resource management priority. They also are useful in identifyingpopular, higher-priced species which have the potential for agroforestry production orwhich are already managed or domesticated by local farmers and which may escapenotice in social surveys. It is important that local values and views of resource scarcityand conservation priority are not overshadowed by categories developed for interna-tional application (see steps 5 and 6 below). Identifying the ‘most valued, mostvulnerable’ subset of species at the local level and at a national or international levelprovides an opportunity to stimulate resource management action at two levels: onelocal, the other national/international. In some cases these will overlap.

Step 1: Identify Species in Commercial or Highest Demand

An important focus is species used in high volume locally (building poles, fuelwood) orin smaller volumes in highly species-specific trade (crafts, medicines, edible plants) (seeChapter 2 for more on local values and volumes). The identification of species in tradecan be done at ‘both ends’: in source areas and in sites where they are used (or on sale).Correct identification is best performed in source areas. It is extremely important thatthis is done through collection and expert identification of good voucher specimens(see Chapters 5 and 6). If working from ethnobotanical studies of markets that arelinked to informal trade networks, it is useful to survey the largest (regional and central)markets which carry the widest range of species; then work ‘upstream’ to source areasidentified on the basis of discussions with commercial collectors and traders in order tocollect fresh voucher specimens (see Chapters 2 and 6). In the case of internationalexport trade, this could be from listings of exporting companies or from customs dataand phytosanitary certificates.

Step 2: Prepare a Shortlist of Species

The shortlist should include species which are:

• destructively harvested (bark, roots, bulbs, stems, wood, whole plants);• slow growing (separation on the basis of life form can be useful);• most popular and/or most expensive, or sold in greatest number (small plants)

and/or volume, in local marketplaces;• considered to have become, or be in the process of becoming, scarce by market

traders or commercial collectors. Species substitution (often due to scarcity) can bea useful ‘flag’ in this case.

Step 3: Identify Species that May Require Special Conservation Effort

Conservation biologist Reed Noss (1990) has suggested five categories of species thatmay need special attention.

• Ecological indicator species: such species signal the impact of events that will affectother species with similar habitat requirements. Afro-alpine plants such as giantlobelias and giant senecios, which will be affected by global warming, are a goodexample.

• Keystone species: these species play a pivotal role in the community or ecosystem(such as fig species whose fruits support many primate, bird and invertebratespecies, but which are exploited on a large scale for making drums and beerbrewing troughs in Uganda).


Applied Ethnobotany

90

• Umbrella species: these species have large area requirements; if given enoughprotection, they will enable the conservation of many other species in the samearea. The plant equivalents of eagles and large mammalian carnivores would bedioecious tropical tree species which occur at low densities and require large areasof forest to maintain viable populations.

• Flagship species: these consist of popular, charismatic species which are symbolic ofthe need for conservation and stimulate conservation initiatives. Several medicinalplants, such as the Madagascan rosy periwinkle (Catharanthus roseus) have beenused as ‘flagships’. Culturally important species can also be ‘flagships’.

• Vulnerable species: these comprise rare species with low reproductive ability andlow genetic variation. This category would include species that are prioritized byother steps 4 to 6, which are particularly vulnerable to human impacts.

Step 4: Shortlist Species Further on the Basis of Commonness or Rarity

This is based on species’ characters of geographic distribution, habitat requirementsand local population size. From an international (and often local) perspective, thehighest priority would be given to a species with narrow geographical distribution, arestricted habitat and small population size.

Table 3.5 Rabinowitz’s seven forms of rarity

Geographic range Large SmallHabitat specificity Wide Narrow Wide Narrow

Local population sizeLarge, Locally abundant Locally abundant Locally abundant Locally abundantdominant in several habitats in a specific in several habitats in a specific habitatsomewhere over a large habitat over a small over a small over a large

geographic area geographic area geographic area geographic area

Small, Constantly sparse Constantly sparse Constantly sparse Constantly sparsenon-dominant in several habitats in a specific habitat in several habitats in a specific habitat

over a large over a large over a small over a smallgeographic area geographic area geographic area geographic area

Source: Rabinowitz et al, 1986; Pitman et al, 1999

Step 5: Set Priorities on the Basis of Phylogenetic Distinctiveness

Within the resulting shortlist, the highest priority should be given to the followingspecies (in descending order).

• species in a monotypic family (highest priority);• species in a monotypic genus;• species in a segregate genus, subgenus or section of a medium to large genus;• species in a small genus (two to five species);• species in a medium to large genus;• species which are part of a species complex;• infraspecific taxon in a medium-size to large genus (lowest priority).

Step 6: Prioritize Species According to IUCN Categories of Threat

In common with step 5 above, these priorities were developed for application on aglobal scale, such as judging the extinction risk of the whole species. In many cases, thiswill differ from the local perspective of resource users. It is important that local,national and international perspectives are taken into account.


most commonly sold species in thesemarkets were usually sold through all themarkets surveyed.

With hindsight, the lesson is that if youare making an inventory of species soldand do not have the opportunity of visitingall marketplaces in a region, but are ableto visit a few markets regularly, then selectlarge (regional or central) markets ratherthan smaller ones. Size and number ofmarketplaces in developing countries isgenerally a function of city size. Cities arealso more likely to have more culturallydiverse populations, drawn in from manyrural communities. Diversity of speciessold decreases with decreasing size ofmarketing area. The most species are soldin regional markets, fewer in centralmarkets, still fewer in intermediatemarkets, yet fewer in standard markets andleast in minor markets.

This also depends upon the impor-tance placed on certain categories of plantuse in large urban areas. Regional andcentral markets, which bring together awide array of species from a large area, aretherefore important sites for an inventoryof which species enter commercial trade.Alternatively, surveys of smaller ruralperiodic marketplaces in more remote(and often more resource-rich sites) nearto conservation areas can provide animportant ‘early warning system’ for localresources which may need to be monitoredfor the impact of an emerging commercialtrade.

Data can then be compared with infor-mation from historical records, discussions,participatory (PRA, RRA) surveys andindividual interviews, indexing methods(see Chapter 2), or social survey data – forexample, on major health problems in thestudy area. Data from markets onfrequency of sale can also be compared tospecies preference data from social surveysor quantitative ethnobotanical methodsthat link plant-disease combinations (Johns,

Kokwaro and Kimanani, 1990; see alsoChapter 2), or compared to statistical dataon health, housing or the availability ofalternative energy sources to fuelwood orcharcoal. In Ghana, for example, where107 woody plant species have beenrecorded as used for chewing sticks indental care, one may wonder which speciesshould be short listed for monitoring.Interview surveys with a sample of 887people showed that just six species (knownby four local names) accounted for 86 percent of all chewing sticks used for dentalcare and the bulk of commercial sales (Adu-Tutu et al, 1979).

Although you will initially record theprices for which different species are soldon the basis of local units (see Chapter 2),it is important that this is converted to aprice per dry mass, as unit sizes can vary.Prices for products sold in markets areuseful for several reasons. Firstly, they areuseful for assessing economic returnsregionally, according to different peoplealong the marketing chain or the economicviability of cultivation. Secondly, pricereflects resource supply in relation todemand. Records of price changes overtime can ‘flag’ increasing scarcity. Locallycommon species are rarely sold in localmarketplaces unless it is bulk sale forprocessing or retail elsewhere.

When a popular species is scarce, dueto geographical distribution or overex-ploitation, then trade occurs fromresource-rich areas to the places wherethere is demand, but little or no supply. Asscarcity increases, so does the price. Whenalternatives are not available, the higherthe price, the greater the incentive to gofurther and further afield for a scarcespecies. Improved roads and cheapertransport reduce this cost. As a result,internal marketing systems change in twoways, each shortening the marketingchain. Firstly, cheaper transport enablesrural people to get to larger centres to sell


91


their products. Secondly, better roadsimprove the access that outsiders have tomore remote plant resources. Outsidersfrequently have more buying power thanlocal people in remote, resource-rich areas.If this takes place and resource tenurestarts to break down, then this hastens thescramble for resources in high demand.

When scarcity results in higher prices,this can stimulate a shift from high-density, resource-rich patches tolow-density, less accessible or marginalareas where resource densities are lower.Where alternatives are available, thiscontinues until price capping occurs: inthis instance, prices reach a point wherealternatives are cheaper. For highlyspecies-specific uses, such as traditionalmedicinal plants, prices continue to risebecause only that species will suffice in atraditional remedy, for symbolic ormedical purposes. This stimulates a tradeover very long distances. In between thesesituations is one in which there is a rippleeffect, where overexploitation of onespecies results in a shift in harvesting toother species.

Knowing which species are sold and,of these, which are most commonly tradedis useful, but it is also important to knowthe source of these species by habitat andby location. In their study of the SantaCatarina del Monte market in Mexico, forexample, Bob Bye and Elemira Linares(1985) found that of the 114 species sold,28 species were gathered from wildhabitat, 52 species gathered from anthro-pogenic vegetation types, 32 species weredomesticated and 2 species were non-domesticated species in cultivation. Of the1560 species identified in trade inGermany, 70 to 90 per cent are primarilyharvested from the wild (Lange, 1997). InKwaZulu/Natal, South Africa, over 99 percent of the 400 medicinal plant species arewild harvested (Cunningham, 1988).

Ethnobotanical surveys of markets can

also go beyond grouping species as towhether they are wild collected, managedor domesticated, to focus on genotypicvariation within species and local prefer-ences for particular qualities representedby this variation. The widest variation isdisplayed by domesticated indigenousspecies in markets, some of them littleknown as crop plants outside that region.In West Africa, for example, agroforestersRoger Leakey and David Ladipo (1996)surveyed local markets in Cameroon to getthe vendors’ opinions on what fruit quali-ties were preferred by people buying fruitof the ‘bush plum’ (Dacryodes edulis), anindigenous tree species which has beendomesticated in West Africa as a tree crop.Apart from the useful information, thiscost-effective, short survey provided infor-mation on preferred qualities ofpulp-to-seed ratio, flavour and cookingqualities; their analysis of fresh fruit massand pulp–seed ratio shows the extent ofvariation which can occur through domes-tication by local farmers (see Figure 3.14).

Volumes sold; sources of supplyand demand

Before you spend a considerable amount oftime and money quantifying the amount ofmaterial sold, you need to be certain thatthis is going to result in the answers youneed. In some cases it may be more costeffective to select the major source areas onthe basis of social surveys, such as partici-patory rural appraisal (PRA) or interviewmethods, and to use these as a basis forcarrying out damage assessments of harvestimpacts on populations of indicator species;this will avoid spending a large amount oftime and resources quantifying the amountof plant material sold.

If the objective of your study is todetermine the value of wild plants in trade,or to establish the amount of an alterna-tive supply that must be provided to take

Applied Ethnobotany

92


the pressure off wild stocks, then it maybe necessary to quantify the volume ofplant material sold. In this case, you needto carefully pick the marketplaces and salepoints where this is done.

If you are measuring volume sold orharvested, it is important to monitor andidentify this at the start of the chain. InSouthern Africa, for example, sale offermented palm sap involved a marketingchain which began with the initial tappingby local men, moved to primary sales bywomen, transport by entrepreneurs, andfinally to resale at households outside thepalm savanna zone. At this point, in orderto make a profit, the women who resold thepalm wine doubled the volume by diluting

it with water and adding sugar. Monitoringvolume sold at this stage without under-standing the process involved wouldobviously provide a gross overestimate.

Do not be too ambitious. Instead,focus on the species or resource categorysold in the most volume or which is mostvulnerable. It is usually best to do this atwholesale or ‘bulk-breaking’ centreswhere ‘units’ of sale are larger, such aslarge sacks of fruit, medicinal plants orcharcoal, truckloads of fuelwood orcontainers of undiluted palm sap.

Are the ‘units’ in which plant productsare sold consistent within or betweenmarkets of the same type? If not, your endresult may yield unreliable data, after


Note: Different symbols represent different fruit lots. Symbols in bold are fruits being sold for more than 500 CFAfrancs/kg.Source: Leakey and Ladipo, 1996

Figure 3.14 Genetic variation in Dacryodes edulis (Burseraceae) fruits from a survey of marketstalls in Yaounde, Cameroon

93

9

8

7

6

5

4

3

2

1

0

Pulp:seed ratio

0 20 40 60 80 100 120 140Fresh fruit weight (g)


considerable effort and expense. Makesure that you have identified the range oflocal unit sizes, have community supportfor the work, have selected reliableenumerators and that you have designedand field tested any appropriate forms (seeChapter 2).

Where the wild plant resource isharvested in high volume (rattan bundles,charcoal, fuelwood, building poles), or isdifficult to transport (such as palm wine),then bulking centres are usually close tothe resource areas were they are harvested.

Where transport is easier, due toharvest and demand for smaller quantitiesof plant material (traditional medicines,weaving fibre, edible fruits), bulkingcentres may be located a long way fromthe source habitat through long-distancebulk trade. In South-Eastern Africa, forexample, the bulking centre for the palm-wine trade was at a cross-roads within theHyphaene palm savanna, from where itwas transported outside the palm savannazone by women who earned money fromdilution and resale of palm wine (seeFigure 3.2).

The distance that wild plant resourcesare transported also depends upon theperishable nature of the plant product.The rapid fermentation of palm wine, forexample, meant that it was only trans-ported 60 to 70 km from source to finalsale. A similar constraint is placed on theharvesting and sale of perishable fruits orof freshly gathered medicinal leaves orstems, affecting the distance that theseproducts are transported. ‘Perishability’also increases the risk that wholesaleharvester-sellers face in selling harvestedmaterial before it deteriorates.

By contrast, bark, roots and bulbs aregenerally far less perishable. For thisreason, dried bark, roots or bulbs, or driedbundles of whole plants, are a commonfeature of the long-distance trade acrossvegetation zones. In West Africa, for

example, dried roots of Entada abyssinica,locally called terenefou, are transported800km from the dry savanna source areasof Burkina Faso to the urban markets ofAbidjan, Côte d’Ivoire, in the tropicalforest zone.

However, if prices and profits are highenough, local traders will make remark-able use of efficient transport networks toget perishable species to the market. Asroad networks extend into more and moreremote rural areas, so commercialharvesters or middlemen flow in, andfavoured plant species flow out. Even airfreight is used to transport edible andmedicinal plants, regionally or interna-tionally: ‘bush plums’ (Dacryodes edulis)and eru (Gnetum africanum) leaves arebought by West Africans living in Franceor Belgium and Chinese traditionalmedicines are sold in Europe and NorthAmerica.

Due to its perishable nature, theAfrican medicinal plant khat (Cathaedulis) is a good example. Remarkably, fora product in long-distance trade, theyoung leaves of Catha edulis need to bechewed while still fresh for maximumeffect – and for this reason, the price ofkhat rapidly drops with time. As a result,the trade has to be highly organized to getleaves from the farm to the end-user assoon as possible. Even at the height of therecent conflict in Somalia, light aircraftfilled with carefully packed bundles ofkhat would fly into Mogadishu fromNairobi’s Wilson Airport and the bundleswould be whisked away to the Mogadishumarket in Somalia. In Kenya, most khat isgrown in Meru district just north ofMount Kenya. Packed into fast motor carswith dare-devil drivers, the khat is thendriven to Wilson Airport outside Nairobias fast as possible, packed into a lightaircraft and flown to Mogadishu. It ispacked into vehicles again and driven tothe Mogadishu market for sale.

Applied Ethnobotany

94


International air transport is also thereason why expatriate Yemeni or Somalicommunities as far afield as Australia,Holland, Italy, England, Canada or the USare able to buy khat leaves to chew(Beekhuis, 1997). A recent survey amongst70 Somali people living in Liverpool, UK,for example, found that 43 per cent ofmen had used khat, with 39 per centchewing it on a daily basis.

In the Catha edulis case, the high-valueleaves led to this species being cultivatedhundreds of years ago. In other cases, tradeleads to an unsustainably high impact onsome species, particularly when supplies ofslow-growing, destructively harvestedspecies have been diminished by habitatdegradation. Evidence for unsustainableharvest comes from the observations oflocal people, including gatherers andtraders. Rural communities in many partsof Africa, Asia, Central Europe and the

Americas are increasingly concerned aboutlosing self-sufficiency as their local wildpopulations of favoured, popular speciesare dug up, bagged and transported to far-away regional markets.

In addition, many medicinal specieshave multiple uses, some of which have afar greater impact than harvesting formedicinal purposes. From detailed studiesin Belem markets, in the BrazilianAmazon, Patricia Shanley and Leda Luz(in press) showed that, in addition to the9000kg of Tabebuia bark sold for medici-nal purposes, over 5500m3 of Tabebuiatimber were exported annually fromBelem. The important questions are: whatimpact is this is having on species? Howcan harvesting impacts be measured?What are the options for sustainable use?The following three chapters set out toanswer these questions.


95


The social survey methods and ethno-botanical work in local markets describedin Chapters 2 and 3 are the first stepstowards understanding patterns ofdemand for particular plant species. Thenext three chapters are a ‘nested progres-sion’ covering methods for studying thesupply of plants which are a focus of thatdemand.

Although it is important to considerharvesting impacts on plants at the largerspatial scales of plant populations (seeChapter 5) and landscapes (see Chapter 6),we also need to understand how individ-ual plants respond to harvesting. Usually,we see and measure things at the individ-ual plant level first. When you walkthrough forest, savanna or grassland withlocal harvesters, it is likely that you wouldsee signs of harvesting: stumps of cut trees,debarked trees or signs of root or tuberremoval for food or medicine. You mayknow which species have been harvested,or if not, would collect good-qualityherbarium specimens to enable identifica-tion. But this is just a first step. You mayalso know how much is harvested (seeChapter 3); but what size are the harvested

plants (or plant parts) and how muchharvestable material is there? From marketand social surveys or field observation,you may know what range of products hasbeen, or is likely to be, harvested from theplant; but how long does it take to reachharvestable size? How does size or agerelate to production of leaves, bark orother non-timber products? What effectdoes harvesting these non-timber productshave on individual plants? All of these areimportant questions from a resourcemanagement viewpoint.

This chapter deals with the methodswhich can be used to answer thesequestions. Chapter 5 then puts harvestinginto a plant population dynamics perspec-tive. The methods described in this chapterare linked to a ‘bird’s eye view’ of vegeta-tion dynamics, maps and aerialphotographs in Chapter 6, which dealswith the interplay between plant popula-tion dynamics and disturbance. Thesecommunity and landscape-scale factorsprovide the crucial context for under-standing plant harvesting at thepopulation and individual plant levels.

Chapter 4

Measuring Individual Plants and AssessingHarvesting Impacts

Introduction


The great thing about applied ethnob-otany is that you can do good-quality fieldwork without buying a lot of expensiveequipment. The main skills are in under-standing what you see in the field or hearfrom local resource users, and in knowingthe key measurements necessary for aparticular study. In the methods describedbelow, a great deal of very useful data canbe collected using the following:

• a regular tape measure or a 3m diame-ter (dbh) tape or forestry callipers;

• a can of paint for marking measuredstems;

• aluminium tags for marking plants;• aluminium nails;• measuring scales, with size depending

upon what you need to measure,ranging from 100g or 1kg balances to5kg, 25kg or even larger hangingscales;

• Swedish bark gauge or a Verniercalliper;

• a panga or machete;• field notebooks, graph paper, ruler,

pencils;• plant presses and specimen labels;• a hand lens.

In some cases you may need more expen-sive equipment such as a globalpositioning system (GPS), a direct readinghypsometer (also known as a clinometer)for measuring tree height, a Swedish incre-ment borer (used to extract wood fromtrees to determine age), or a batteryoperated electronic balance for measuringfresh bark or leaf mass. Only under excep-tional circumstances would you requirethe very expensive equipment sometimesused in commercial forestry research, sucha Relaskop, which is used for opticalmeasurement of tree diameters andheights, or an infra-red gas analyser tomeasure photosynthetic rates. You wouldalso need equipment for setting up plotsas described in Chapter 5. I have tried toavoid describing methods that are notavailable to most field workers. In a fewcases, however, I explain how to sectionperennial corms or tree stems in order toage trees. I also mention the use of anelectronic balance (fresh bark mass, leafmass), leaf area meter and the sphericalcrown densiometer, which is used tomeasure tree or shrub canopy closure.

Measuring Individual Plants and Assessing Harvesting Impacts

97

Necessary Equipment

Measuring diameter, height and bark thickness

With their historic focus on commercialtimber, foresters have used systematicmethods to obtain measurements for treediameter, height, volume, and cross-sectional area, and many well-establishedmethods are available (Philip, 1994). Thesame does not apply to non-timberproducts or harvesting foliage, bark, resinsor roots from trees. Nor does it apply tovines, lianas or deadwood. Until recently,

for example, many foresters consideredlianas a nuisance that suppressed timbertree production, and were more interestedin systematically removing lianas thanmeasuring them or assessing their value tolocal people. In the past, most forestersalso had little interest in standing deadtrees or fallen logs. Conservation biolo-gists see their ecological importance andlocal people value deadwood for fuel and


prize lianas as multipurpose bindingmaterial. As a result of these changingviews, some methods used to measurenon-timber products or to assess harvest-ing impacts on plants are relatively new.

Measuring individual plants and divid-ing them into size classes according todiameter or length is an important aspectof collecting inventory data, estimatingyields or developing population matrixmodels or survivorship curves (seeChapter 5). Use of stem diameter or lengthsize classes can also be related back torecords of size-class selection by localharvesters collected during householdsurveys, in markets or from bundles ofharvested plant material (see Chapter 2).Bark thickness also relates to stem diame-ter, increasing with tree size.Measurements of stem diameter (orlength) are made on the basic assumptionthat stem diameter (trees, bulbs or corms)or stem height (palms, tree ferns) reflectplant age. Size is often only poorly corre-lated with age, so this assumption must betreated with caution (see Chapter 5). A 2-metre tall sapling, for example, may be 5years or 50 years old, depending upongrowing conditions. Methods of ageingplants are therefore very important (see‘Methods for Ageing Plants’ below).

One of the reasons for using stemdiameter or height classes is thataccurately ageing plants is difficult formost species, particularly in the tropicsand subtropics. Tree stems, bulbs andcorms generally get thicker as these plantsgrow older, and diameters are thereforeused as the most appropriate measure forgrouping them into size classes. Mostpalms and tree ferns have an apical meris-tem on an unbranched stem, growingupwards (longer) as they grow older,rather than increasing in diameter. Rattanpalms, for example, show a great increasein length for very little increase in stemdiameter. For these reasons, stem length

rather than stem diameter is a moreaccurate measure for assessing the popula-tion structure of palms, cycads, grass treesand tree ferns.

Diameter: stems, bulbs and corms

Diameter measurements of trees areconventionally taken at a set height of1.3m (‘breast height’) and this is expressedeither as diameter at breast height (dbh) orcircumference (girth at breast height) (gbh).This is the most commonly used treemeasurement in forest inventory work, andwill vary according to the shape andgrowth form of trees (see Figure 4.1).When you need to calculate tree volume,then diameter measurements are taken atregular intervals along the trunk, so thattree-trunk volume calculations are madefor each trunk subsection as a way ofminimizing error as the trunk tapers (see‘Stem Mass and Volume’ below). Basaldiameter and dbh measurements can eitherbe done with a forestry ‘diameter tape’which enables the dbh to be read directlyfrom a girth measurement, with forestrycallipers, or with a standard tape measure,later converting from circumference (girth).

Diameter at breast height (dbh) is thekey measurement used to calculate basalarea (ba), the area occupied by a cross-section of the stem, usually expressed asm2 per ha. This is used to get an estimateof stand biomass of different tree specieswithin a known area. As an alternative toclimbing trees to make these diametermeasurements, two expensive opticalinstruments, either the Spiegel Relaskop orTele-Relaskop invented by WalterBitterlich, are used by foresters toaccurately measure tree diameter whenlight conditions are good and there is aclear view of the tree trunk. It is also possi-ble to measure tree height, distance andassess basal area using a Relaskop. Inpermanent sample plots, the increase in

Applied Ethnobotany

98


tree diameter (dbh) is usually measured onsuccessive intervals with a tape measure.

Increases and seasonal changes in treegirth can be measured using a Vernier girthband, usually made from steel oraluminium (Hall, 1944; Alder andSynnott, 1992). Girth bands can be madelocally at little cost using the metal bandsfound on packing cases, which are then

marked using a template so that accuratereading to an accuracy of 0.1mm can bemade. This process is described by Liming(1957). Girth bands are not suitable fortrees smaller than 7cm in diameter as theydo not allow enough room for the springand scale on the girth band. Limits on thenumber of girth bands can lead to poorsample size and a false sense of accuracy.


Source: Alder and Synott, 1992

Figure 4.1 Standard measurements of diameter at breast height (dbh) or of girth at breast height(gbh) in trees that are leaning or strangely shaped

99

1.3m

1.3m

1.3m

1m

1m

Buttress

End ofbuttress

On a slope:measure from uphill side

Leaning tree:measure from inside lean

perpendicular to stem

Forked tree:measure as two trees 1m above fork

Buttressed tree:measure 1m above buttress


For this reason, it is better to make succes-sive measurements on marked trees usinga tape measure.

In contrast to most foresters, whowork primarily with tall, single-stemmedtrees, ethnobotanists working with localresource users need to measure multi-stemmed trees and shrubs. As TauberTietema, who worked in Botswanawoodland of low vegetation height (threeto five metres), wryly commented, ‘In suchvegetation, measuring basal area at ankleheight is more practical than at breastheight.’ To solve this problem, basal areacalculations were based on measurementstaken ‘at ankle height’ of five to tencentimetres above ground level, just abovethe basal swelling. With tall multi-stemmed trees, it is useful to measure thediameter of each stem at 1.3 metres and, ifthe stems are linked to a basal stem, tomeasure basal diameter as well, recordingthe number of stems per plant.

Diameter size classes can also be usedto apply to harvesting of long-lived bulbousspecies or of corms, which are commonlyexploited for food and medicine in south-ern Africa (see Chapter 5). Bulb and cormdiameter measurements can be performedmore accurately using a Vernier calliper,measuring across the widest part of thebulb or corm. In some cases, however, largebulbs and corms grow fairly deep in the soiland destructive sampling is often undesir-able, particularly where rare species areconcerned. If destructive sampling is likely,it may be necessary to harvest a subsampleto measure bulb depth, diameter and freshbulb mass. With some bulbs, corms andtubers there is also the opportunity to ageplants by counting leaf scales (someAmaryllidaceae and Liliaceae) or to countepidermal sheath layers which accumulatearound stem tubers (some Droseraceae);these measurements can be related back tofresh mass bulb or diameter (see ‘AgeingPalms, Tree Ferns and Grass Trees’ below).

Stem length or height

Before you start measuring stem length orheight, ask yourself: ‘What level ofaccuracy is required for the purposes ofthe study?’ For many studies, heightclasses, such as <1m, 1 to 2m, 2 to 5m, 5to 10m, 10 to 20m and >20m may besufficient. In other cases, you may needmore accurate measurements.

Three main methods are used forassessing the vertical height (or stemlength) of trees and palms: direct estimatesusing a height pole; geometric methodsusing a ruler, a stick or Christen’s hypsome-ter; and trigonometric methods using aclinometer or hypsometer. Height can alsobe calculated trigonometrically frommeasurements taken with an abney level,but as this equipment is more expensiveand the calculations more time consumingthan using a direct-reading clinometer, thisis not described in detail here.

Because they loop or curve into theforest canopy, climbing palms, tilted trees,vines and lianas all pose methodologicalproblems. These problems and somesolutions are discussed in Box 4.1. Ideally,what is needed is a repeatable methodwith a low level of observer bias. Eachmethod has its disadvantages, however.For example, visual estimation methodsare vulnerable to variation betweenobservers, and the equipment needed fortrigonometric measurements is expensive.In a comparison of height measurementmethods, however, Mary Stockdale(1994) found that the ruler method wasas accurate as any other in measuringstraight rattan palm stems and that it wasthe most accurate method for estimatingthe length of curved or looped rattanstems (see Box 4.1).

Direct estimation

This uses a measured pole clearly markedat 0.5m intervals and usually 3 to 4m long.

Applied Ethnobotany

100


This is held vertically at the base of thetree, enabling an observer standing farenough away to see the base and the topof the stem. The number of pole lengths isthen counted by eye to estimate the heightof the tree.

Geometric methods

The most commonly used of these is theruler method. This uses cheaper equip-ment but is more difficult to learn than theclinometer method. The ruler methodneeds at least two people to carry outmeasurements. One person stands at thebase of the tree stem holding a pole orranging rod which is marked into 1-, 2-and 3-metre intervals. A second personstands far enough away from the tree sothat they can see the top and bottom ofthe tree at a distance where regular inter-vals on the ruler (such as every 3, 6 and9cm) correspond to the 1-, 2- and 3-metreintervals on the pole. Using the ruler as aguide for scale, the height of the tree isread off from a distance against the ruler,converting the known 1-metre intervals onthe scale against the tree into an estimateof tree height.

Trigonometric methods usingclinometers (or hypsometers)

Examples are different models of Suuntoclinometers, and Haga and Blume-Liesshypsometers. Clinometers (and diametertapes) are available from companies whichsupply the forestry industry. There are twobasic types of clinometer. Direct-readingclinometers incorporate a prism andenable quicker measurements to be made,but are more expensive. As their nameimplies, tree height can be read off theinstrument directly. With more basicclinometers, tree height has to be calcu-lated. Tree heights are determined on thesame principle with both types of clinome-ter, taking measurements from a set

distance (usually 15m or 20m) away fromeach tree, so that there is an imaginarytriangle between the person doing themeasurement and the tree (see Figure 4.2).The first step is to make sure you are theset distance away from the stem (at least15m). In dense forest, it is often impossi-ble to get a line of sight 15m or 20m awayfrom the tree. The alternative is to pick thebest vantage point, measure the distanceto the tree and angles to the base and topof the tree and then calculate height. If youare working in tropical forest, you willrealize that foresters who work in planta-tions have it easy: they rarely have to cut atrail for a sight line to the base of eachtree!

You also need to take slope intoaccount. Measurements are always takenfrom eye level. In most cases you will findthat your eye level is slightly higher orlower than the base of the tree and this hasto be taken into account. If you are onlevel ground, you need to add your heightto the reading given to tree-top height. Ifyou are on a slope below the tree base, youneed to perform the following calculation.Sight onto the tree base and take a reading(eg 2.5m); then take a reading to the treetop (eg 15m); subtract the two to get treeheight (stem length) (12.5m). If you arestanding on a slope above the tree base, gothrough the same calculation but add,rather than subtract, the two readings.

With palms you need to decide andclearly state in your methods which sight-ing you used in measuring length: the topof the stem or the total length includingthe leaves. As this second measurement isinfluenced by leaf harvesting, or will beless relevant to studies of palm-stemharvesting, it is best to sight to the top ofthe stem and not include the leaves abovethis point. Again, it is much easier to takeslope into account with a direct-readingclinometer (see Figures 4.2 b and c); butthis has to be calculated trigonometrically


101


with basic clinometer models. If youencounter a tree with a broken tip, youshould note that the stem is broken, recordthe height of the break and, based on treesof a similar size for this species, estimatethe height the tree would have been beforethe tip broke off.

Measuring biomass and volume

Ethnobotanists are interested in measuringa far greater diversity of plant resourcesthan traditional foresters, whose maininterest has been measuring ‘timber height’– the length of the tree bole being the mainsource of sawn timber. In general, however,the tree bole represents only roughly 30 per cent of the biomass of a tree.

Ethnobotanists involved in resourcemanagement are not only interested in treestems, but also in measuring the manyharvested products from the remaining 70per cent of dry mass, including roots (30 to55 per cent), twigs and leaves (about 10 percent) or branches (about 15 to 30 per cent),as well as exudates and bark. In some casesall the above-ground or below-groundbiomass is measured so that correlationscan be made between shoot:root ratios orbetween fresh tree biomass and other tree(or shrub) dimensions, such as stem diame-ter, basal area or canopy diameter.

Regression equations derived fromcorrelations between total biomass andfactors such as stem basal area or stemheight are useful tools to estimate biomass,

Applied Ethnobotany

Figure 4.2 Use of a clinometer to measure tree height. (a) On level ground one simply adds theobserver’s height (here 1.6m) to the tree height. (b) and (c) The reading obtained is from the

observer’s eye level, so on sloping ground suitable adjustments need to be made

102

12m

(a) (b)

(c)

1.6m

20m

20m

3.5m

12.5m16m

3.25m

10.25m

13.5m

20m



103

BOX 4.1 LENGTHS OF ‘AWKWARD CUSTOMERS’: CLIMBING PALMS

AND TILTING TREES

Clinometers are easily used by foresters who have the luxury of working in plantationsor coniferous forests, where most trees are straight. Working in tropical forests andsavanna is more challenging, particularly if we encounter tilted trees or are interestedin non-timber plant products such as climbing palms (rattans) or lianas that curve ortwist as they climb into the forest canopy. Each of these cases will be dealt withseparately. In each scenario, the methodological dilemma results from two factors.Firstly, there is a need to avoid destructive harvesting. In the past, most studies thatmeasured rattans resulted in pulling them out of the canopy. This is time consumingand destructive; and if you are pulling an ant-associated rattan such as Calamus deera-tus, you are also likely to get covered in angry, biting ants! Particularly if you areworking in a national park or on scarce resources valued by local people, it is importantto minimize destructive harvesting. Secondly, it is also crucial to use a method formeasuring length of these ‘awkward customers’ in an accurate, repeatable way thatminimizes the bias between field workers using that method.

Working in Brunei, forest researcher Mary Stockdale (1994) carried out a very inter-esting test of the accuracy of four different methods for measuring rattan palm lengths.She compared visual estimates of length with the ruler method, the clinometer methodand length estimation based on internode counts. The internode method counts thenumber of internodes by eye (or with binoculars) and multiplies by the mean internodelength (based on five randomly selected, measured internodes for each stem).

In order to get around the problem of rattan stems curving into the forest canopy,a three-metre pole was held vertically so that its tip touched the rattan stem (Stockdale,1994). This formed an important reference point for the ruler and clinometer methods.Electrician’s tape was used to mark a point three metres below where the marker poletouched the stem, dividing the rattan stem into two sections: ground length, from theroot collar to the tape, and above-ground length, from the tape to the base of thepetiole of the top-most leaf of the stem (see Figure 4.3a). If the stems were verycrooked, observers had to imagine where the top would have reached if the stem hadbeen straight. Visual and ruler methods both followed similar procedures to thosedescribed above for straight stems. With the clinometer method, the angles weremeasured to the top of the stem (θtop) and base of the height pole (θbase). The next stepwas to measure the distance (X) from the observer’s eyes to the base of the rattan stemusing a tape measure. To calculate the horizontal distance from the observer’s eyes tothe stem, Stockdale used the formula:

D = X(cos θbase)

The above-ground length (L) was then calculated using the formula:

L = D(tan θtop + tanθbase)

In assessing the accuracy of these four methods, Stockdale also took into account howlong it took to take measurements and factors that commonly affect height measure-ments, such as variation between observers performing the measurements, topographyand light levels in the forest. Her results were surprising and encouraging to fieldworkers who cannot afford a clinometer. Contrary to Philip’s (1994) report that geomet-


usually expressed in kilograms or tonnesper hectare. In this way, harvested quanti-ties (see Chapter 2) can be comparedagainst the standing stock (biomass). Bycontrast, with detailed work on a fewforestry plantation species, a variety ofdifferent regression equations have beendeveloped by researchers in naturalsubtropical woodlands in Asia and Africa.Examples of different regressions arebiomass regressed against stem basal areain Southern African savanna (Rutherford,1982; Tietema, 1993), against stemcircumference in dry tropical forest inIndia (Singh and Singh, 1991) and againststem diameter in Somalia (Bird andShepherd, 1989).

Although different methods are usedto measure biomass or volume, fresh massof plant material is commonly standard-ized to an oven-dry mass equivalent (driedat 80°C until no more mass is lost), sincethere is a significant seasonal variation inthe amount of moisture in fresh (or evenair-dried) plant material. The ratio of fresh

mass:oven-dry mass is based on subsam-ples of plant material – sample discs ofstems and branches or of bark, roots,leaves or browse (stems/leaves) – as youcannot expect to fit large quantities ofplant material into the drying oven!

Leaf measurements

The type of leaf measurements you choosedepends upon the aim of the study and thetype of leaf resource being harvested. Ifselective harvesting of leaves takes place,then you may combine counts of thenumber of harvested (or harvestable)leaves (see ‘Harvesting VegetativeStructures’ below) with measurements ofleaf, leaflet, culm length or petiole width.These are usually made to assess leaf size-class selection, primarily for long-livedleaves harvested for fibre (Agavaceae,Cyperaceae, Juncaceae, Palmae). WithCyperaceae and Juncaceae, the wholeculm is usually measured. With palms, thisdepends upon what harvesters use, and the

Applied Ethnobotany

104

ric methods such as the ruler method are less accurate than trigonometric methodsusing a clinometer, the ruler method was the most accurate and quickest for measuringcurved or looped stems. The internode counting method was the least accurate (15.3per cent mean error) and also the most time consuming. The only disadvantage wasthat the ruler method took longer to learn. However, the field assistant in the experi-ment who found it most difficult to learn the ruler method was as accurate with theruler method as with a clinometer. The main reason for the lower accuracy of theclinometer method in measuring curved stems was that while the ruler method onlyneeded one measurement, the calculation of length needed to measure two anglesand the distance from the observer to the base of the stem.

When trees are leaning, the height is usually estimated (B’D) as an average of tworeadings, each taken from positions (A and C) opposite one another and exactly thesame distance away from the base of the tree (Philip, 1994). A more accurate measure-ment of tree length (BD) can only be made if you measure the angle at which the treeis leaning (θ) and calculate the true length (BD) as:

BD = DB’/cosθ

Another alternative for measuring crooked, forked or bent trees is to use a range-finding dendrometer (Grosenbaugh, 1991), but these are very expensive andunavailable to most field workers.


correlation between palm stem size andleaf or leaflet size. Palm leaflet length hasbeen commonly used in studies ofHyphaene palms in Southern Africa(Cunningham, 1988; Cunningham andMilton, 1987), but in her work on thepalm Sabal uresana in Sonora, Mexico,Elaine Joyal found that petiole widthprovided a statistically significant correla-tion with palm size class and this was usedinstead of leaf length (Joyal, 1996).

Foliage mass, on the other hand, isused as a measure of the quantity of leavesharvested for livestock fodder (inkilograms) or as edible greens (in grams).This is often done on the basis of localunits such as bundles (see Chapter 2),usually measured as fresh (‘green’) mass.

This is then converted to dry weight equiv-alents on the basis of fresh mass sampleswhich are oven dried and reweighed.

A third method of measuring leaves isspecific leaf area (SLA) – the ratio of leafarea to dry leaf mass. Measurements ofSLA provide useful insights into thebiology of plants and are discussed inChapter 5.

Measuring the plant canopy:biomass, volume, area, density and

crown position

A common direct measure of foliage (orforage) mass is to clip and weigh foliageto get fresh mass and then to obtain oven-dry mass, relating total amount measured


105

Sources: (a) Stockdale and Power, 1994; (b) Philip, 1994

Figure 4.3 (a) Dividing the rattan stem into two sections: ground length, from the root collar tothe tape, and above-ground length, from the tape to the base of the petiole of the topmost leaf of

the stem. (b) The isosceles triangle method of measuring tree height

3m height pole

reference point

groundlength

3m

marked point

above groundlength Tree height above

eye level = Lm

Line of sight = Lm

(a)

(b)


to plant size such as basal diameter ordiameter at breast height. MikeRutherford (1979) has reviewed themethods used to assess the quantity ofavailable browse, and many of thesemethods are applicable to uses of leaf orleaf/stem material by people as well aslivestock.

Quantitative measurements of thecanopy (or crown) area or volume ofwoody plants are commonly used byresearchers who are interested in timberproduction, and, increasingly, in assessingthe effects of lopping, pruning or flower-picking on plants. Forestry researchers, forexample, have used tree-crown surface areaor, alternatively, crown volume as predic-tors of individual tree growth, since crownsurface area relates closely to the mostactive photosynthetic area of the tree madeup of younger leaves (Philip, 1994). Crowndiameter or area can also be a useful predic-tor of fruit yields. Measurement of plantcanopy size enables useful comparisons tobe made between harvested and unhar-vested populations of trees or shrubs as anindicator of available browse or the effectsof plants lopped for forage. It has also beenused in studying the effects of commercialflower picking on Proteaceae in Australia(Banksia) or South Africa (Protea,Leucodendron).

Foresters interested in timber produc-tion most commonly measure crowndiameter, crown depth or clear bole length(to the lowest live branch or lowestcomplete whorl of branches). While shrubsare short enough for direct measurementsof crown height and depth, a hypsometeris often used to take these measurementsfor tall trees. Crown diameter (width) isused to calculate the crown (canopy) area.The crown diameter measurement used inthis calculation is commonly the averageof two crown diameter measurementstaken at right angles, either taken atrandom or in a predetermined direction

(such as north-south and east-west diame-ters) – the average of the widest diameterand the diameter at right angles to this, orthe average of twice the maximum andtwice the minimum radius from the centreof the trunk to the edge of the crown(Philip, 1994; Tietema, 1993; Witkowski,Lamont and Obbens, 1994). Forestersworking in plantations have based theircalculations of crown volumes of conifersand young Eucalyptus trees on the modelof a cone. The methods manual by MichaelPhilip (1994), which is widely used bystudents in East Africa, discusses thesemeasurements in greater detail and isrecommended for additional reading.

Calculations to estimate crown volumeobviously depend upon the characteristicshape of the tree species; and as you knowfrom field experience, this varies widely,from flat-topped Acacia trees to the roundcanopies of some woodland shrubs. Intheir study on the effects of flower pickingon the sclerophyllous shrub Banksiahookeriana (Proteaceae), for example,Witkowski et al (1994) randomly selectedBanksia shrubs in each of six sites andcalculated canopy area and canopy volumeusing the formulae:

Canopy area = πW1W2 = 0.7854W1W2__ __2 2

Canopy volume = 4 πW1W2H= 0.5236W1W2H__ __ __ __3 2 2 2

where W1 was the widest canopy diame-ter; W2 the perpendicular diameter to this;and H the canopy height.

In addition, they measured theopenness of each shrub’s canopy using aforest (or spherical crown) densiometer. Adensiometer is an instrument that deter-mines forest canopy density, used byforesters in forest thinning operations orto assess light requirements for forest

Applied Ethnobotany

106


regeneration. Densiometers have a convexor concave mirror reflector covered bygrid squares, marking out an overheadplot. Placing it under the canopy andcounting the number of grid squaresshaded gives a measure of the percentageshading. The results of their study showedthat Banksia shrubs which had not beenharvested had a significantly greatercanopy area (1.59 times), canopy volume(1.78 times) and were taller (1.1 times)than shrubs in sites where flower pickinghad taken place (Witkowski, Lamont andObbens, 1994). In addition, seed storageand seed production per individual plantwere 57 per cent and 50 per cent lowerrespectively in the harvested plants: anissue that will be discussed in the nextchapter, dealing with this issue at a plant-population level.

Access to light is a crucial factor whichneeds to be taken into account if you aremeasuring plant growth in forests (such asstem diameter increments in terms ofgrowth rates or palm leaf productionrates). If you do not have a densiometer tomeasure forest overstorey density or a luxmeter for measuring light intensity, youcan use Dawkins’s field classification oftree crown position (Dawkins, 1958)which he developed in Uganda. TheDawkins crown classification system,which has also been used in West Africaand South-East Asia, rates tree crownposition according to the following scale:

5 = emergent: crown plan fully exposedto overhead light and free from lateralcompetition (this is defined as beingexposed to overhead light at leastwithin the 90° cone of an imaginaryinverted cone with its point touchingthe base of the tree crown;

4 = full overhead light: crown plan fullyexposed to light from above (verti-cally) but next to tree crowns of equalor greater height within the 90° cone;

3 = some overhead light: crown planpartly exposed to overhead light, butpartly shaded by other crowns;

2 = some side light: crown plan fullyshaded from above, but exposed tosome direct light from the side, comingthrough a gap or past the edge of theoverhead canopy;

1 = no direct light: crown completelyshaded from above and from the sides.

Although this is subjective, it has shownto be a method giving consistent results(Wyatt-Smith and Vincent, 1962). Inaddition, a study of indigenous tropicalhardwood growth rates in forest in Ghanafound that Dawkins’s classification ofcrown position also correlates well withtree increment (Alder and Synott, 1992).It is also a useful field method for study-ing forest palm-leaf production rates, sincepalm-leaf production rates vary consider-ably with shading (see ‘Ageing Palms, TreeFerns and Grass Trees’ below, andMartinez-Ramos, 1985).

Flower, fruit and seed production

The simplest method of measuring thenumber of flowers or fruits per plant is bydirect counting. This is a suitable methodfor plants which produce relatively fewlarge flowers, such as the Proteaceae, orlarge fruits such as palms, but becomesimpractical with very tall trees, or whenshort plants produce huge quantities ofsmall fruits. When direct counts areimpractical, small, circular fruit traps(usually 0.5m2 – 79.8cm diameter to 1m2

– 112.8cm diameter) or square plots (1 x1m square) are commonly used to subsam-ple fruit fall from trees. Each trap or plothas to be numbered. Circular traps consistof netting which is fixed loosely across awire frame placed on wooden ‘legs’ 0.5 to1m off the ground. Square plot traps areeasier to make with a wooden frame. The


107


netting needs to sag (about 30cm deep inthe centre of the trap). If it is too tight, thefruits may bounce out!

This method assumes that most of thefruits fall under the tree crown and doesnot take fruit removal by birds or otheranimals into account. As Charles Peters(1996) points out in his review of methodsfor assessing fruit production, measure-ment of what is left after frugivores havehad their fill is not necessarily a disadvan-tage for an ethnobotanical study of fruityield since it gives a realistic estimate ofwhat would be available for harvest.However, they still need to be visited everyfew days before fruits rot and to limitanimals taking fruit from the sample plotsor traps.

The first step in sampling fruit fall isto estimate the ‘shadow’ of the treecrown: the area it would cover on theground. This is a lot easier in savannathan in tropical forest, where the treecrowns overlap. This is measured, drawnon graph paper and the area calculated.As fruit fall under trees is seldom even, itis best to place the fruit traps or plotswithin four quadrants around each tree.These are established by dividing the‘crown shadow’ into four, with linesdrawn at right angles to each otherextending out from the tree trunk. Fruittraps or plots are placed in a stratifiedrandom design within each quadrant.Either a constant proportion of the crownarea can be sampled or a constantnumber of traps can be placed under eachtree regardless of crown area. Bothmethods have disadvantages. A constantnumber of traps results in more intensivesampling of small trees than large ones.Varying the number of traps with differ-ences in crown area complicates somestatistical tests as sample sizes will vary.Based on his studies of measuring fruityields, Charles Peters (1996) suggests thatif a fixed sampling percentage is needed,

then you should use enough fruit traps orplots to cover 10 per cent of the totalcrown area, and if a constant number oftraps is used then you need 8 to 12 trapsor plots per tree. As you would need tosample about five to ten fruiting trees ina range of stem diameter classes (fortrees) or stem-height (in the case ofpalms) size classes, you need to choosecarefully which species you study, as thisamounts to a lot of work.

Methods for estimating annual fruitand seed production from trees have beenreviewed by Green and Johnson (1994)and by Peters (1996), both of which arerecommended reading.

Bark: thickness and mass versustree diameter

By contrast with the low diversity of planta-tion trees harvested for bark, such ascinnamon (Cinnamomom verum, C.aromaticum), black wattle (Acacia mearn-sii) and cork oak (Quercus suber), localcommunities harvest bark from tens ofthousands of tree and shrub species formany purposes (medicine, fibre, fishpoisons, spices). Bark harvesting is oftenselective for particular stem size classes orbark quality. Inner bark fibre for bindingpurposes is stripped from youngBrachystegia (Leguminosae) saplings whichhave smooth bark, rather than older treeswith rougher, thicker bark.

In Southern Africa, most herbalistspreferred to harvest thick bark from oldertrees, as this was considered more potent.In Uganda, Maud Kamatenesi found thatherbalists were even more selective. Notonly did they select mature Rytigyniakigeziensis (Rubiaceae) trees, but theypreferred trees with small yellow-greenleaves growing close to or on top of hills,rather than in valleys.

Seasonal factors can also be a factor inthe timing of bark removal as this is often

Applied Ethnobotany

108


easier during the growing season. A goodexample is the seasonal removal of barkfrom Brachystegia trees for making beehives in the miombo woodland of South-Central Africa. Other factors can alsoinfluence tree selection. In Zambia, forexample, beekeepers cut small test blocksfrom miombo woodland trees prior tobark removal, selecting for 10 to 60 percent (mean = 34 per cent) of Brachystegia,Julbernardia and Cryptocephalum treeswith cross-grained inner bark (Clauss,1992). Measuring bark thickness enablesus to correlate the tree diameter (dbh) withbark thickness for individual plants and todetermine potential bark yields. It is alsoan important link to records collected inethnobotanical surveys of local markets(see Chapter 3). Bark varies considerablyfrom species to species in its thickness andtexture. Between and within species, barkthickness also varies with tree size or age,rate of growth, genotype and location ofthe tree. In Southern Africa, for example,Rauvolfia caffra trees growing at the coasthave a very different outer bark texturefrom those growing in upland sites.

Bark thickness

This can be measured in two ways whichdiffer in cost and in their impact on thetree: either using a bark gauge or a Verniercalliper. Bark measurements should betaken at breast height (1.3m), with fourseparate measurements taken around thetrunk to get a mean bark thickness pertree. Bark gauges (Swedish and Finnishtypes) are available from suppliers offorestry equipment. The bark gauge has aninner ‘chisel’ which is pushed through thebark to the surface of the wood (see Figure4.4). As the thin shank of the gauge ispushed in, the base plate of the gauge ispushed outwards, enabling a measurementof bark thickness (in mm). This methodminimizes bark damage, an important

factor when dealing with rare trees suchas Faurea macnaughtonii, which aresusceptible to fungal attack. If a barkgauge is unobtainable, then the moredestructive method of cutting out a smallblock of bark from four points at breastheight (1.3m) can be done to get a meanmeasurement of bark thickness. Barkthickness can be measured using a Verniercalliper. If you do this, you need to avoidlarge bark slashes which damage the tree.You also need to avoid inaccuratemeasurements which might occur whenlayers splay out as a result of cutting witha blunt knife or panga (machete).Alternatively, carefully use a sharp chiselto push through the bark in the same waythat a bark gauge would be used, markingthe bark thickness and measuring itdirectly once the chisel blade is pulled out.

Bark mass per tree

It is important to take accurate measure-ments of bark thickness, particularly if youare calculating bark mass for each tree. Ina single species, trees with the same heightand diameter but with different bark thick-ness will have very different bark yields. Ingeneral, if height and dbh are constant, a1mm difference in bark thickness willcause an increase or decrease of about 10per cent in bark mass (Schonau, 1973).

In most trees, bark mass per treeincreases with increasing dbh and treeheight (Schonau, 1982; Kamatenesi,1997). As few data are available for wildspecies, bark yields from cultivated treessuch as black wattle are very useful inplacing cultivation and bark productioninto perspective as an alternative tooverexploitation of wild stocks. Based onstudies of over 1300 trees, Schonau (1972)developed a multiple regression for Acaciamearnsii of bark mass on dbh, height, barkthickness:


109


log BM = 1.87253 (log D) = 0.72118 (logH) + 0.152919 (BT) – 0.11767 (BT x logD) + 0.037728 (BT x log H) – 2.04586

where BM = total fresh bark mass per treeto a stem tip diameter of 5cm underbarkin kg; D = dbh in cm; H = total height inm; and BT = bark thickness at breastheight in mm.

Although based on Acacia mearnsii, thisequation has also proved useful in calcu-lating single tree bark mass in other species(Rytigynia kigeziensis, Prunus africana).Although fresh bark mass has been used bySchonau (1972), a forester who has doneextensive work on Acacia mearnsii barkyields – since he found fresh bark mass amore useful independent variable thanoven-dry bark mass – use of fresh bark

mass is usually fraught with problems. Fortwo reasons, it is more prudent to convertfresh mass to oven-dry bark mass. Firstly,the moisture content of bark generallyvaries seasonally and between sites.Secondly, oven-dry bark mass provides astandard against which to compare theprice per kilogram of oven-dried bark (notair-dried) samples from local markets,which is very useful if you want to comparedifferent bark (or root or leaf) prices perkilogram of different species. In Acaciamearnsii, bark moisture content variedbetween 48 to 52 per cent, with a mean of50 per cent (Schonau, 1973); in Prunusafricana (42 to 50 per cent) and inRytigynia kigeziensis it averaged 59 percent (Kamatenesi, 1997). Fresh bark massshould be measured as soon as possible inthe field using either a mechanical O’Haus

Applied Ethnobotany

Figure 4.4 A Swedish bark gauge being used to measure the thickness of Prunus africana bark

110


balance or a battery-operated electronicbalance. Each sample should be carefullylabelled. The bark samples were then driedat 80°C until dry in a laboratory oven andreweighed to determine the range andmean moisture content.

A good example of the practical valueof these data is provided by part of a studyfrom the multiple-use managementprogramme in Bwindi-ImpenetrableNational Park, Uganda. In this case, MaudKamatenesi’s study aimed to determine thebark mass available from the medicinalshrub Rytigynia kigeziensis (see Figure4.5). These data were then compared withtree densities within multiple-use zones,and with the quantities local herbalistsexpected to be able to harvest.

Stem mass and volume

For many years, ethnobotanists interestedin wood consumption for fuel, building orcarving have been weighing bundles offuelwood, or individual tree stems cut for

building purposes, using 5 to 50kg hangingscales. Unless you can chop trees intosections and weigh them, wood volume hasto be calculated for large trees which aretoo heavy to lift. If logs have beenharvested for woodcarving or use forhousing or fencing, you will not want tooffend anyone by cutting them up! Instead,calculate the volume of wood on the basisof diameter and height measurements.

In Namibia, for example, Antii Erkkilaand Harri Siiskonen (1992) calculated thevolume of Colophospermum mopane andCombretum stems used for traditionalOwambo home construction (see Figure4.6). Using measurements of the heightand diameter of the poles, they calculatedthat the 21,599 poles used to make MrLazarus Uugwanga’s homestead wouldhave a combined volume of 69.5m3. Withan additional 20 per cent of this toaccount for debarking and wood losswhen shaping the poles (13.9m3), and theassumption that an additional 10 per centof this was trimmed off after felling, the


111

Source: Kamatenesi, 1997

Figure 4.5 The relationship between Rytigynia kigeziensis diameter (dbh) and bark mass availablefrom tree stem (up to 2m)

1.78

0.32

0.06

0.01

Bark mass (kg)

0 5 10 15 20 25Girth at breast height (cm)


result was that 91.7m3 of wood were usedto construct the homestead. Fortunately,they could take measurements directly.The only highly accurate way of measur-ing volume of wood is with a xylometer.This measures how much water isdisplaced when the tree trunk (or whateveritem is being measured) is fully loweredinto the xylometer tank. This equipment isgenerally not available for the fieldresearcher, so unless you can improviseand use a water tank instead of a xylome-ter, the next best option is to find thevolume by direct measurement.

Due to their interest in assessingtimber volumes, the most detailedapproaches to measuring the volume oftree stems have been developed byforesters on the assumption that different

parts of the tree stem are similar togeometric shapes. The simplest exampleis the main branch-free trunk, which isassumed to be a truncated cone on acubical parabloid. To avoid the biascaused by slight tapering, it is importantto make a series of diameter measure-ments along the length of the trunk. Thebest way to do this with a felled tree is tomark off sections with chalk and measurethe diameter at 1m to 2m intervals. Thevolumes are calculated for each subsec-tion and summed to get the total volumeof the trunk. It is important to take aseries of measurements along the trunk toavoid overestimating timber volume.Even then, you can expect an overesti-mate due to irregularity of the tree stemand bark.

Applied Ethnobotany

Source: Duggan-Cronin collection, reproduced with permission from McGregor Museum, Kimberley, South Africa

Figure 4.6 With complex palisade fences, traditional Owambo housing used a spectacularamount of wood

112


To date, many ethnobotanical studieson the use of timber for building purposes,such as Christine Liengme’s (1983) studyof wood use in Gazankulu, South Africa,have estimated wood volumes assumingthat the poles or logs are cylindrical. Othergeometric shapes used in forestry to calcu-late tree volume are the cone, cubicparabolic and quadratic parabolic.Detailed information on these calculationsare given in Michael Philip’s (1994)manual on measuring trees. Volumes oflogs are usually estimated using Huber’sformula, which works well for logs whichare cylindrical or shaped like a quadraticparabolic which has had its upper portioncut off parallel to the base. BecauseHuber’s formula underestimates volume inlogs that taper sharply, it is best to baseyour calculations on short, measuredsections as described above. Huber’sformula is:

volume (v) = πLd2m——-

4

where dm = diameter of the log at midlength; and L = log length.

In some cases you may want to comparedata on wood volumes (in m3) with dataon wood mass (in kg). To do this, you needto know the density of the wood beingused (in kg/m3). Be aware that this varieswithin trees, depending upon whetherwood is from the top or base of the tree orfrom heartwood or sapwood. Forexample, if you have calculated that amukwa (Pterocarpus angolensis) stemfelled by a woodcarver has a volume of 1.9m3, and you know that the wood densityis 650 kg/m3, then the estimated woodmass before carving would be 1235kg.

Assessing the quantity of deadwoodper tree is an important issue in studies offuelwood availability, yet much of the

deadwood that local people collectcomprises crooked smaller branches forwhich accurate volume calculations areimpractical. For this reason, weighingdeadwood is a more practical option. Inaddition to assessing the quantity ofdeadwood per tree, you also need to weighthe woody ‘litter’ that has fallen onto theground.

As part of his study on fuelwood avail-ability in Southern African savanna,Charlie Shackleton (1993) first assesseddeadwood availability and the effects ofwood harvesting on individual trees ofdifferent species. He then used these datato determine standing deadwood biomassper hectare and the deadwood yield perhectare per year for harvested and unhar-vested sites (see Chapter 5). In addition tothe standard approaches of recording thespecies, stem circumference and height ofeach tree or shrub, he visually estimatedof the amount of deadwood as a propor-tion of the whole tree and the proportionof wood chopped from the tree. He alsorecorded whether chopped trees hadresprouted (coppiced), were dead, or werealive but had not resprouted. Repeatingthese measurements for trees and shrubswithin randomly selected plots showedthat only 6 per cent of stems showed signsof severe chopping (>50 per cent removal).This suggested that when wood was cutfrom standing trees, chopping was severe,since almost 90 per cent of chopped treeshad more than 50 per cent of biomassremoved. Harvesting also focused onlarger trees, with 36 per cent of trees withstems more than 16cm in circumferencehaving lost over 50 per cent of woodybiomass. However, the data he collectedon whether resprouting had taken place ornot illustrated the resilience of most trees,as 77 per cent had resprouted and, of theremainder, 19 per cent were dead and 4per cent were alive but had not resprouted(Shackleton, 1993).


113


Whole plant biomass

As long as you have an accurate balance,weighing smaller plants is straightforward.Whole trees, however, pose a logisticchallenge! For this reason, there arerelatively few studies outside of forestryplantation research which have correlatedthe dimensions of trees and shrubs coveredearlier in this chapter, such as canopydiameter or dbh, with total above-groundtree biomass. This information is veryuseful for studies of fuelwood availabilityor harvesting impacts in the field, orthrough the interpretation of aerialphotographs. For these reasons, TauberTietema (1993) carried out a study inBotswana which measured total biomassof 14 tree species, correlating the resultswith crown area, stem basal area andheight. Trees were selected to get a repre-sentative sample of different size classes,

and biomass was measured with a springbalance mounted in a crane fitted onto avehicle. Tree height and crown diameterwere measured before each tree was cut.With multi-stemmed trees, each stem wasweighed individually. Tree mass mostclosely correlated with stem basal area (seeFigure 4.7), but correlations between treemass and tree height or crown basal areawere less significant. When he comparedthe single regression curve combined forall trees in his sample with similar workon tree species from Africa, India andEurope, Tauber Tietema found that to alarge extent this described the relationshipbetween basal area and tree mass for thesespecies as well.

What measurements of biomass,volume or diameter do not tell you is howlong the plants took to grow. This crucialissue in resource management is coveredin the next section.

Applied Ethnobotany

Source: Tietema, 1993

Figure 4.7 A comparison of the stem basal area/weight regression lines for ten Southern Africansavanna tree species. Solid lines = Acacia species, dotted lines = Boscia albitrunca,

Colophospermum mopane, Combretum apiculatum, Croton gratissimus

114

103

102

10

1

10–1

Biomass (kg/tree)

1 10 102 103 104

Basal area (cm2)

A. mellifera

A. tortilis

A. erioloba

A. karroo

C. gratissimus

A. erubescens

B. albitrunca

C. apiculatum

C. mopane

A. luederitzii


Potentially, information on the age ofharvested plants is a key to many issues inresource management. It also leads to abetter understanding of plant life histories.In many cases, however, the age-determi-nation key is missing or does not quite fit,particularly with tropical and subtropicalplants. There are exceptions, however, andmore and more tropical trees are beingaged using tree rings (Jacoby, 1989).Reasons for difficulties in ageing some treespecies are the variation in growth rates ofthe same species in the same site orbetween populations, indistinct or non-existent annual growth rings, leaf scars ornodes. These are discussed in more detailbelow. Wherever possible, it is importantto make comparisons with plants ofknown age or with plants which have beenmarked at specific times, so that ‘annual’rings and growth can be cross-checked tosee if they really are annual or not. Whereit is possible to age perennial plants,however, this provides valuable informa-tion for resource users, managers andresearchers in predicting yields.

Slow-growing, slowly reproducingplants are known to be vulnerable tooverexploitation, yet we rarely know howold individual plants are or how long theylive. This information is not only of greatinterest in developing resource manage-ment programmes, but is of value to localresource users, who often underestimatethe age of slow-growing (and thereforevulnerable) plant species, and who areoften amazed to find out that the tree theyare carving or using for building is threeor four times older than they are! Beingable to age individual plants is also ofgreat value in developing matrix modelsof plant populations by providing accurateinformation on recruitment, the time

taken to shift from one size class or stageto another, and on plant life spans (seeChapter 5). Although some methods ofageing plants (particularly dendrochronol-ogy) require laboratory work, this ispossible for some field workers and so isincluded here. The following basic stepsare generally involved.

• Select plant species that have potentialfor ageing and are important from aresource management perspective.This could be done on the basis ofprevious studies or through the twosteps below. Although annual rings ofsome tree species can be seen with thenaked eye (macroscopically), these canbe deceptive and microscopic identifi-cation should be performed to avoiderrors where growth rings are indis-tinct (Lilly, 1977).

• Cross-check with plants of known age(usually from known date of plantingof bulbs – Ruiters, McKenzie andRaitt, 1989), corms (Werner, 1978;Levins and Kerstner, 1978) or treeswhere the cambium has been markedat known annual intervals by hammer-ing successive nails into the trunk tomark the cambium (Shiokura, 1989;Grundy, 1995); and cross-check withtrees that are marked by climaticextremes (drought, cold) of knowndate.

• Age trees based on cores taken fromthe trunk using a Swedish incrementcorer or from cross-sections of stems,where rings are counted in sanded,polished wood or stained corm cross-sections (Werner, 1978); or examineleaf-base counts from longitudinalsections of bulbs (Ruiters, McKenzieand Raitt, 1989). Destructive sampling


115

Methods for ageing plants


should be avoided whenever possible.There is an extensive literature onnon-destructive sampling methods forwood samples (see, for example,Swart, 1980).

• Take site differences into account; ifyou are ageing plants from differentsites, be aware that differencesbetween sites and populations willneed to be assessed before extrapolat-ing data from one site to another.

Counting tree rings

Ageing trees by counting tree rings(dendrochronology), seen in stem cross-sections, was substantially developedearly this century by A E Douglass (1914,1936). Douglass made the fundamentaldiscovery that information on pastclimatic variation was reflected in tree-

ring patterns and that these could bematched between trees. He used thismethod to build up a time series(chronology) by matching successivelyolder tree rings (see Figure 4.8). Thispioneering work was developed furtherby H C Fritts (1971), who identified sixprinciples for minimizing non-climaticinfluences that obscured ring-widthvariations due to climatic variation.Computer sequences of these tree ringscan be developed to cover very longperiods, and very long, accuratechronologies have been constructed. Thelongest of these is based on Pinuslongaeva, a species in which individualtrees live to 2000 years. This enabled thedevelopment of a chronology extendingback 8000 years, and was so accurate itwas used to recalibrate the radio-carbontime scale (Ferguson, 1970; Lilly, 1977).

Figure 4.8 Diagrammatic representation of three tree stems where the distinctive growth ringsindicating specific years have been matched for the periods 1900–1910 and 1920–1930

Applied Ethnobotany

116


Most dendrochronological work hastaken place in temperate Europe, NorthAmerica and Asia. In some cases, this isdone by counting rings in cut cross-sectionsof tree trunks. A less destructive method isto use an increment borer. These are T-shaped tools comprising the handle and aborer bit, 4.3 to 12mm in diameter, whichis screwed into the tree trunk to extract acore of wood. Thicker (12mm diameter)cores are often used for quantitative analy-sis, but taking cores from hardwood treescan be very difficult. For two reasons,ageing trees is not as simple in tropical andsubtropical areas. Firstly, temperategymnosperms are particularly suited totree-ring dating. Although oak (Quercus)and ash (Fraxinus) trees are easy to ageusing tree rings, some angiosperm wood ismore complex and difficult to age.Secondly, dendrochronologists select treesfrom arid or very cold sites which showmarked variation in tree-ring thickness.They avoid studying trees that grow in siteswith enough water throughout the year, astheir tree rings are likely to be uniformlywide with little variation between rings.They also avoid trees that grow denselytogether, since competition between treesobscures changes in growth rings due toclimatic change. These conditions are moredifficult to find in well-watered, warmtropical areas, dominated by angiospermswith less seasonal growth than in cooltemperate areas.

Although it has been widely acceptedthat growth rings are not a reliable methodof ageing many tropical and subtropicaltree species, surveys of woody plants fromsouthern Africa (Lilly, 1977), southernAustralia (Schweingruber, 1992) and thetropics (Jacoby, 1989) show that manyspecies have potential for ageing based ontree-rings. Based on an examination of 108Southern African tree species, for example,five species were considered promising fordendrochronological work (Albizia forbe-

sii and Burkea africana (Leguminosae),Ekebergia capensis (Meliaceae),Zanthoxylum davyi and Vepris undulata(Rutaceae)). More recent studies inSouthern Africa have also shown thatseveral species which Lilly (1977) gave avery poor rating for dendrochronologicalwork, such as Acacia karroo, produce ringswhich do correlate with age (Gourlay andBarnes, 1994; Prior and Gasson, 1990).One reason for this is the marked dry-season leaf fall and wet-season flushingcommon in deciduous woodlands inSouthern Africa and probably also in othersubtropical areas. In addition, trees whichproduce annual rings, but are less suited todendrochronological work, can still bevery useful for resource managers in deter-mining stem age, estimating annualincrements and developing practicalcutting rotations. A good example is IslaGrundy’s (1994) study of Brachystegiaspiciformis in Zimbabwe, a tree specieswidely used for building poles. Her studiesof growth rings showed that this species setdown annual rings which were very usefulin developing local woodland managementbased on coppice rotations.

Counting scars: trees andreiteration

Long-lived species which experienceannual leaf flushes will also show visiblescars in smooth-barked (usually younger)stems. These scars can be used as a fieldmethod for estimating stem age. Manysubtropical and tropical tree species inareas with highly seasonal rainfall aredeciduous, losing their leaves during thedry season (or the longer of two dryseasons in equatorial areas with bimodalrainfall). A new flush of leaves is producedat the start of the wet season (for instance,in several Acacia, Brachystegia andErythrina species). The period ofdormancy is long enough to induce an


117


anatomical change evident as a wrinkled‘bud-scar’ on tree stems and brancheswhere the new season’s growth starts. Anannual flush of leaves and consequentscarring is also evident in evergreen treessuch as Podocarpus and Afrocarpus inAfromontane forest. In addition to itsvalue in ageing younger (less than 30- to40-year-old) shade-tolerant reseeders suchas Podocarpus (often harvested as build-ing poles due to its straight growth form),this is a good method for showing studentshow poorly diameter (dbh) and age corre-late between trees in forest canopy gaps orshade. When long-term growth records arenot available, it is useful to count pastseasons’ flower heads in serotinous speciesand to examine bud scars marking annualiteration; this can help to develop realistictransition matrix models and to assess theprobability of saplings and young trees inmaking the transition from one size class(stage) to the next (see Chapter 5).

The concept of ‘reiteration’, the repli-cation of part the tree’s basic architecturalstructure with the addition of new‘modules’, is the basis for the plant archi-tectural models developed by FrancisHallé and R A Oldeman (1970) (seeChapter 5, and Bell, 1998). In some treespecies, the scars formed at the end of eachgrowing season during this process canremain visible for years (or even decades)until obscured when rough bark developson the stem. This is evident on trees withgrowth from a single meristem (monopo-dial) and those with sympodial growth(successive lateral meristems). This offersthe opportunity for ageing younger (10- to40-year-old) tree stems in the field bycounting the number of scars. In the CapeFloral Kingdom (fynbos – fine leavedshrublands) many serotinous shrub speciesin the Proteaceae and Bruniaceae alsomark this reiteration with annual orbiennial flowering. Since the flower headsare retained on the plants, these can be

counted to give field estimates of age aslong as you know the phenology of thespecies you are studying.

Ageing palms, tree ferns and grass trees

Since palm and tree-fern stems areharvested for building purposes, grass-treeand tree-fern wood for lathe-turned bowlsand many cycad species for horticulturaluse, ageing provides useful insights forconservation management and populationstudies. This ageing method has also beenused to assess habitat disturbance historybased on ageing palms affected by treefalls in tropical forest in Mexico(Martinez-Ramos et al, 1988) and the firefrequency of vegetation in south-westernAustralia (Lamont and Downes, 1979)(see Chapter 6).

Many tree ferns, cycads and arbores-cent (tree-like) monocotyledons, such asAustralian grass trees (Xanthorrhoeaceae)and palms, have a single stem with a singlegrowing shoot (apical meristem) at the tip.This type of architecture (Corner’s model;see Chapter 5) offers the opportunity forage estimates based on leaf scars (in palms,tree ferns and cycads) or the depressionsand ridges left by annual growth flushes(grass trees). Palms and many tree fernsproduce leaves singly in regular acropetalorder (developing from below upwards)from a single growing shoot (apical meris-tem) at the end of their stems. These leaveslive for a few years, die and in many casesdrop off and leave distinct leaf scars on thestem (see Figure 4.9). Leaf number istherefore a useful marker of growthevents. Counting leaf scars has beenwidely used in short-term populationstudies to age individual palms (Bullock,1980; Enright and Watson, 1992;Sarukhan et al, 1984; Tomlinson, 1963;Pinard, 1993). Knowing that tree ferns,such as Cyathea (Cyatheaceae) and

Applied Ethnobotany

118


Leptopteris (Osmundaceae), produce asingle growth flush of new leaves eachseason enabled James Ash to derive ageestimates and population models fromcounting the scars left by leaf (or stipe)scars on tree-fern trunks (Ash, 1986,1987; see Figure 4.10c). In Australia,Byron Lamont and Susan Downes (1979)used annual fluctuations in the diameterof the stems of grass trees (Xanthorrea andKingia) to age the stems (see Figures 4.10aand b). With Xanthorrea plants living upto 350 years and Kingia up to 650 years,this was a very useful tool for studyingfrequency of flowering and of incidence offire. They found, for example, that somegrass trees were at least 200 years oldbefore they flowered for the first time, andthat fire frequencies prior to Europeansettlement were far lower than the two- tofour-year frequency previously suggested.

Age estimations, as one step in devel-oping a population matrix model (seeChapter 5) can be useful in putting theharvesting of palm or tree-fern stems intoperspective. Working in Sian Ka’anBiosphere reserve in Yucatan, Mexico,Ingrid Olmsted and Elena Alvarez-Buyulla(1995) used leaf scars to estimate the ageand growth rates of Thrinax radiata(‘chit’) and Coccothrinax readii (nakax)palms harvested to make lobster traps andhouses. Over 480 adult Coccothrinaxpalms are used by Mayan fishermen tobuild a single hut. Ageing clearly demon-strated the slow growth rates andconsequent vulnerability of these twospecies. To get to just 3 metres high tookThrinax palms between 31 and 55 years,with adult palms living 100 to 145 years.Coccothrinax readii palms were evenslower growing, taking 63 years to get 3metres high and living over 145 years. Thesteps generally used in this methodcomprise the following.

• Assess the number of palm or tree-fernleaves produced each year. There aretwo methods, depending upon howleaves are shed from the palm(Tomlinson, 1963). With arecoid palmsthat have ‘self-cleaning’ trunks – cleanlyshedding their dead leaves (see Figure4.9) – a painted mark is made directlybelow the tubular base of the oldestleaf. Before you do this, you may needto rub off the waxy coating so that thepaint sticks to the palm stem.Alternatively, with palms that retain


119

Figure 4.9 The stem of the nikau palm(Rhopalostylis sapida) in coastal forest, New

Zealand, a species whose populationdynamics have been well studied by Enrightand Watson (1992) using counts of the frond

scars clearly visible on the stem


their leaves on the trunk for many years(or to use a consistent method for bothcategories of palm), you can tie analuminium or plastic tag to the petioleof the most recently fully emerged leaf.Mark a sample of all stem size classes(ideally, 30 stems per size class) over anumber of years (two to four years) tomeasure leaf production rates. Annualcounts are made of the number ofleaves produced per stem each year toget the mean number of leavesproduced by each size class per year.The same tagged leaves can also be used

to assess palm-leaf life spans, which arerelevant to leaf resource management.

• Count the number of leaf scars oneach palm or tree-fern stem. Directcounts are made of the number of leafscars for each height segment.

• Take the length of the seedling andestablishment phases into account(when no stem is visible), so that thiscan be included in the age estimate.The seedling phase comprises theduration when the embryo emergesfrom the seed and becomes indepen-dent of the food reserves in the seed

Applied Ethnobotany

Sources: (a) and (b) Lamont and Downes, 1979; (c) Ash, 1986

Figure 4.10 Grass trees and tree ferns. (a) Surface features of a Kingia australis stem afterremoval of the leaf bases, showing rings of aerial roots (cut back to their bases). The depression(d) and ridge (r) of an annual flush of vertical growth, the remnants of aborted (i) and mature

(m) inflorescences and of fire-response flowering (f) are also shown. (b) Xanthorrea preissii stemshowing the depression (d) and ridge (r), indicating annual vertical growth. The remnant of

spikes (s) and the associated ring of smooth tissue (t) are also shown. Background scale for grass-tree stems in centimetres. (c) Stem of the Fijian forest tree fern Leptopteris wilkesiana with

arrows indicating annual bands corresponding to growth flushes

120

(a) (b) (c)


endosperm by putting out the firstroots and leaves. The establishmentphase can be a long period of earlydevelopment when the seedling diame-ter grows downward to form a stembase at, or below, ground level beforegrowing upwards. In their study of thepalm Sabal palmetto, for example,Kelly McPherson and KimberlynWilliams (1996) found that theminimum length of the establishmentphase was 14 years and the fastest-growing 1 per cent, 10 per cent and 50per cent of plants would respectivelytake 33, 42 and 59 years to develop anabove-ground trunk. The time thistakes will also vary depending onwhether the palm is produced fromseed (a genet), which takes longer, orby clonal sprouting (a ramet) fromlateral buds on the parent plant, whichis quicker since the establishmentgrowth phase is much shorter.

Working in Chico Mendes ExtractiveReserve, Brazil, Michelle Pinard (1993)used leaf scars and known leaf-productionrates to age palm stems in her study of theimpact of stem harvesting on Iriarteadeltoidea palms. She measured palm-stemheights and counted leaf scars on palmstems that were 5 to 10m high, which hadconsistent, longer internodes. Leaf-scarcounts were also made on a sample offallen palms greater than 10m long, asinternode lengths were shorter in upper-stem sections of these tall stems. Thenumber of leaf scars was then estimatedon the basis of the mean number of scarsper metre using two different meannumbers of leaf scars per metre: one forstems less than 10m and the other forstems greater than 10m high.

Although this method has proveduseful in short-term demographic (popula-tion) studies of palms, it should be usedwith care as it can lead to incorrect age

estimates in palm species with variablegrowth rates (Oyama, 1993). Problemswith this method include the following.Firstly, it assumes that there is little varia-tion in growth rate of the palm species. Thisis not always the case. As a result, you needto study leaf-production rates for differentpalm size classes (seedlings, saplings,juveniles and adult size classes) in differenthabitats within your study area. RobinChazdon, for example, was able to use thismethod to age cana de danta (Geonomacongesta) palms in Costa Rica, as variationin rates of leaf production and leaf drop(abscission) was statistically insignificant.Over a three-year period, markedGeonoma congesta palms in a range of sizeclasses produced an average of 10.1 newleaves, and abscised 9.7 leaves (Chazdon,1992). In other cases, there is variation inpalm or tree-fern leaf production andgrowth rates, resulting in plants of the sameheight having very different numbers of leafscars (see Figure 4.11) (Ash, 1987; Oyama,1993). Palms growing in forest gaps inVeracruz, Mexico, for example, were foundto produce twice the number of leaves andfruits relative to palms in the same size classas palms in shady ‘mature’ forest(Martinez-Ramos, 1985). Dawkins’s classi-fication system for tree crown position (seeearly section on ‘Measuring the PlantCanopy’) is a useful field method for devel-oping an understanding of the effects ofshading on leaf production rates.

Secondly, young palms with under-ground stems and short internodes pose aproblem, as the leaf scars cannot be seenand seedlings vary in the time that theytake to form a trunk. Thirdly, leaf scarsmay be difficult to detect on older palms,leading to an underestimate of the age ofolder stems. Scars are difficult to counttowards the top of very tall palms, socounts are often done from a subsampleof fallen or felled adult palms.


121


Ageing bulbs, corms and stemtubers

Many plants use store nutrients, but somehave specially adapted below-groundroots or stems. These are classified asbulbs, corms, stem tubers or root tuberson the basis of their structure. A bulb isreally a large bud, with swollen modifiedleaves or ‘scales’, attached to a small,compressed stem which bears adventitious

roots. Food is stored in the thickened scaleleaves. Onions (Allium spp) and manyother plants in the Liliaceae are familiarexamples. Corms look similar to bulbs butconsist mainly of compressed stem tissue,with much thinner scale leaves and thebulk of storage within the compressedstem. Many plants in the Iridaceae, suchas Watsonia and Gladiolus, have corms.The potato, a swollen, starch-filled under-

Applied Ethnobotany

Source: Oyama, 1993

Figure 4.11 The relationship between the number of leaf scars and height of Chamaedoreatepejilote palm stems (n = 148 palms) in Mexico showing stems with (a) fewer than 25 scars and

(b) more than 25 scars

122

250

200

150

100

50

0

Height (cm)

0 10 15 20 25Number of stem scars

(a) Palms with less than 25 scars

5

500

400

300

200

100

0

Height (cm)

20 40 50 60 70Number of stem scars

(b) Palms with more than 25 scars

30


ground stem, is a familiar stem tuber.These differ from root tubers, such assweet potatoes (Ipomoea), which areswollen storage roots. Relatively fewstudies have been conducted on ageingunderground storage organs; but this is avery important tool in developing sustain-able harvesting rates of a large group ofplant species of great importance for foodor medicine. Many geophytes are surpris-ingly long lived and some are vulnerableto overexploitation. Although all three ofthe methods discussed below requiredestructive sampling (which shouldobviously be avoided with rare species),they are useful in ageing subsamples ofindividuals in a population or for assess-ing the age of plants that have alreadybeen harvested.

Leaf-base counts

In some bulbous species, new scale leavesare produced seasonally from the centre ofthe bulb, while formation of new adventi-tious roots each season depletes bulbreserves of the outer scale leaves. Bulbsusually increase in diameter with age.These two factors offer an opportunity forageing individual bulbs and studying bulbpopulation dynamics. For reliability, it iscrucial that any ageing is cross-checked onthe basis of leaf production rates, bulbmass and flowering patterns with knownage populations. One of the main reasonsfor this is that the number of leaves,rosettes of leaves or flowers produced bythe same bulb species per year may varybetween populations or seasons. If this isthe case, then ageing may not be possible.Even if the bulbous species you want to ageis known to produce just one leaf a year, orto flower once a year, this can be deceptive.There are two reasons for this. Firstly,many geophytes in the Amaryllidaceae,Liliaceae and Iridaceae family only flowerwhen they reach ‘critical bulb mass’ (Rees,

1969; Ruiters et al, 1993). Secondly, leafproduction rates in some Amaryllidaceaeand Liliaceae can be at one constant ratefor up to ten years in pre-reproductiveplants, then double to another constantrate when the bulbs reach reproductivematurity (Ruiters et al, 1993; Kawano etal, 1982). These have to be taken intoaccount when ageing bulbs. If not, majorover- or underestimates of bulb age willresult.

Cornelius Ruiters et al’s (1993) studyof bulbs of the medicinal plantHaemanthus pubescens (Amaryllidaceae)in coastal fynbos in South Africa is a goodexample of how ageing can be performedon the basis of leaf-base counts (see Figure4.12a) and a thorough knowledge of bulbbiology.

On the basis of studies of markedplants and of bulb mass, Ruiters and hiscoworkers knew that individuals were tenyears or older when they flowered. Theyalso knew that juvenile (pre-reproductive)plants (one to nine years old) producedone leaf per year, while reproductivelymature plants (ten years or older)produced two leaves per year. This enabledthem to avoid errors in ageing plants onthe basis of leaf-base counts since theyknew that the first nine leaf bases eachrepresented a year, but after that two leafbases were produced per year, from yearten onwards. A similar pattern of leafproduction in juvenile and reproductivelymature corms has also been recorded inErythronium japonicum (Liliaceae), agenus whose corms are used medicinallyand as a source of edible starch (Kawanoet al, 1982).

Annual rings in perennial corms

The corms of many well-known horticul-tural plants such as cyclamen(Primulaceae), crocus and gladiola(Iridaceae) are short lived, producing a


123


new corm after flowering while the previ-ous season’s corm decays. Some corms areperennial, however, with old leaf bases andstorage tissues staying within an outerlayer called a tunic. Very few studies havebeen conducted on ageing perennialcorms. Although the example given belowuses microscopic methods similar to tree-ring counts rather than a macroscopicmethod which may be more practical for afield biologist, I am including it here sincethis method could be important in ageingsome of the wide range of perennial cormsharvested for medicinal purposes. Somespecies of Liatris, such as the Kansasgayfeather (L. spicata), are used medici-nally. In studies of transverse sections ofknown-age Liatris aspera corms, PatriciaWerner (1978) showed that annualmarkers could be found in the xylem tissuein the vascular bundles if the corm wascarefully sectioned and stained withsafronin and fast green, two dyescommonly used in laboratory work.

There is, however a cautionary note.Patricia Werner’s microscopic ageingmethod followed up on an earlier study byHarold Kerstner, which showed that cross-sections of juvenile (less thanfour-year-old) corms had pigmented ringsvisible to the naked eye which conformedto known ages of the individuals. Theseconsisted of sclerenchymatous tissue. Inher follow-up study, Patricia Wernershowed that even with young corms, thenumber of rings depended upon where thecross-section was taken. In one case, athree-year-old corm had 16 rings near thecentre, 9 rings at the apex and 10 at thebase. This problem, as well as differencesbetween sites and plant populations, needsto be borne in mind if the annual ring-count method is used (Werner, 1978; Levinand Kerstner, 1978). Despite the complexstructure of older corms, this methodoffers a great opportunity to ethno-

botanists interested in resource manage-ment. Key requirements to check thereliability of this method for any speciesyou are investigating is to assess knownage, marked plants and the differencesbetween sites and populations.

Counting spent remains of annualcorms and stem tubers

After initially starting from seed, manycorm-producing species and some specieswith stem tubers, clonally (vegetatively)produce a new ‘daughter’ tuber or cormeach season. In many cases, the fibrousremains of the past season’s corm or stemtuber remain in the soil. These clonalsuccessions of ‘daughters’ may proceed forten years or more. Flowering may not takeplace in the first few years after recruitmentfrom seed. As the years since germinationincrease, so successive daughter corms ortubers are generally heavier, deeper in thesoil and, once at a flowering stage, producemore fruits per plant. These factors areimportant to take into account whenstudying harvesting and its impact on plantpopulations. For this reason, the innova-tive methods used by John Pate andKingsley Dixon (1982) in ageing threeWest Australian geophytes – Philydrellapygmaea corms, and two tuberous sundew(Drosera) species, D. bulbosa and D.erythrorhiza – offer an opportunity formore widespread application to otherharvested species. Tuberous Droseraspecies, for example, were gathered as afood source by Aboriginal people in south-western Australia (Hammond, 1933).Many plant species with seasonally clonalcorms are locally traded for medicinalpurposes (Cunningham, 1993) and at leastthree Drosera species are in internationaltrade for medicinal purposes (Lange andSchippmann, 1997). Careful investigationof hundreds of Drosera plants showed thatin the two tuberous sundew species, the

Applied Ethnobotany

124


replacement (‘daughter’) tuber formed onthe inner side of the previous season’s stemtuber. As each tuber is surrounded by apersistent outer skin (epidermal sheath),Pate and Dixon were able to count thenumber of epidermal sheaths to assess theage of each plant. In Drosera bulbosa, thisinternal replacement continued for up to17 years, and in Drosera erythrorhiza, forup to 60 years on lateritic soils, but only to15 years or so in sandy soils.

Leaf life spans

In contrast to most edible greens with theirsoft, short-lived and tasty leaves, wildplants selected for durable qualities have ahigh lignin or fibre content (as a defenceagainst herbivory and mechanicaldamage), and often have low rates of leafproduction and long-lived leaves.Examples would be leaf selection forweaving fibre, such as many palms(Arecaceae), agave species (Agavaceae),


Sources: (a) Ruiters et al, 1993; (b) Werner, 1978; (c) Pate and Dixon, 1982

Figure 4.12 Ageing methods for bulbs and corms. (a) Median longitudinal section of a 16-year-old Haemanthus pubescens (Amaryllidaceae) bulb with numbered leaf bases to show how the

bulb was aged. (b) Idealized transverse section of a perennial corm of Liatris aspera (Compositae)showing the location of the collateral vascular bundles which can be used to determine the age ofindividual corms. (c) Ten-year-old corm of Philydrella pygmaea (Philydraceae) showing the lateral

accumulation of spent corms, which can be dissected and counted to assess the years since anindividual plant was recruited as a seedling

125

Scale (cm)0

1

2

InflorescenceLeaf

Soil surface

Previous season’sinflorescence stalks

Parenttuber

Replacementtuber

(c) 10-year-old plant

(b)

ParenchymaVascular bundle

Sclerenchyma

(a)

Scale (cm)

20

1357910

11

12

13

1415

16 2 1514

131211

10

86

42

Roots

Stem stock

Leaf base (numbersshow year producedsince establishment)

10987654321

Seasons since initialcorm establishment


and bromeliad species (Bromeliaceae) usedin the tropics. Durable, long-lived leavesare also being commercially harvestedfrom the wild for florists’ greens.Examples are various Chamaedorea palmsin Costa Rica, Mexico and Guatemala,Rumohra ferns and many Proteaceae inSouthern Africa.

Long-lived leaves characterize slow-growing plants which are less tolerant ofdefoliation. They are often fibrous, likepalm leaves, or sclerophyllous (containingsclerenchyma cells with a high lignincontent) or both, often characterizingslow-growing, shade-tolerant plant species(Reich et al, 1992; Midgley et al, 1995; seeChapter 5, Figure 5.4). From a resourcemanagement perspective, it is useful toknow the age of long-lived leaves that areharvested and their natural life spans. Interms of nutrient inputs, long-lived leavesare expensive for the plant to make. Theyare also at their most valuable to the plant

when they are young, as this is when theirphotosynthetic rates are highest. The costto an ilala (Hyphaene) palm when a young‘sword’ leaf is harvested for basketry, forexample, is much greater than if a mature,two- to three-year-old leaf is harvestedfrom the same palm.

Due to their large size, newly emerged,unopened palm ‘sword’ leaves (see Figure4.13c) can be easily marked to study theirnatural life spans. Smaller leaves have tobe marked with tags (see Figure 4.13a).Once the leaves have been tagged, you canreturn periodically to record when theleaves in the sample start to die off(senesce) and when they have turnedcompletely brown. These tags also enablean assessment of annual leaf-productionrates, which is useful in field assessmentsof leaf damage – for example, in studyingpalm-leaf harvesting (see ‘HarvestingVegetative Structures’ below).

Applied Ethnobotany

126

Harvesting impacts

The effect of harvesting on individualplants will obviously vary according towhat part of the plant is used – indeed,sometimes the whole plant is removed,making it very difficult to measure impactsunless harvesting occurred within a perma-nent plot (see Chapter 5). Harvestingimpact on a plant also depends upon thefrequency and intensity of harvest.Harvesting of leaves, fruits or flowersclearly has far less impact on individualplants than does damage to roots, bark,stems, or removal of the whole plant.

Whether recording damage to individ-ual plants or plant populations, it is usefulto have a systematic way of measuringindividual plants, and field methods forassessing the intensity and frequency ofharvest. This depends upon the funding

and time you have available, as well as onplant population biology and growthform. In some cases, you may be guidedby methods used by other researchers,while in other cases you may need todevelop assessment methods yourself. It isalso useful to learn from experimentalstudies that have measured the impact ofharvesting on individual plants. In thissection, I describe different field ratingsystems and the results of harvestingexperiments (defoliation, debarking andbark regeneration, stem cutting andresprouting) which enable a better evalua-tion of the consequences of harvesting onindividual plants. Although dealt withseparately, it is important to bear in mindthe direct links between harvesting thevegetative parts of the plant (stems, bark,



127

Sources: (a) and (b) S J Milton; (c) and (d) author

Figure 4.13 (a) Seven-weeks fern (Rumohra adiantiformis) with metal tag, micrometer in front ofthe sampling grid. (b) Information provided by harvesters on season, size class and area harvested

is very useful in designing defoliation experiments. (c) Mokola palm (Hyphaene petersiana)marked with paint. (d) Lawrence Mbatha and Sam Ncube taking a monthly measurement of lala

palm (Hyphaene coriacea) leaf growth


roots, buds and leaves). Frequent and/orintense harvests of any of these vegetativestructures will deplete the plants’ carbohy-drate reserves or disrupt water andnutrient flows.

Harvesting reproductive structures(flowers, fruits and seeds)

When you see wild-collected fruits, flowersor seeds in baskets at a marketplace, youmay first assume that these were harvestedwith no impact on the individual plant. Atfirst sight, harvesting of fruits, flowers orseeds may seem to be as close as people canget to sustainable harvest. Do not be tooquick to make this assumption. Althoughharvesting of flowers or fruits generally hasa low impact on individual plants, destruc-tive harvesting – ironically of the mostfavoured species – has been recorded inmany cases, with cutting of branches oreven felling plants to collect flowers, fruitsor seeds.

If neither pruning nor felling takeplace, then the main concern for sustain-able harvesting of reproductive parts is ata species-population level (see Chapter 5)rather than concern for individual plants.The mental ‘alarm belIs’ of the resourcemanager should go off loudest when thefruits of dioecious and monocarpic(hapaxanthic) reseeders are commerciallytraded. In this section, I discuss factorsthat lead to destructive harvesting offruits, flowers or seeds, and methods formeasuring the impact of this at the individ-ual plant level.

To a certain extent, the ‘behaviouralbottleneck’, as Rodolfo Vasquez and AlGentry (1989) called the felling of trees fortheir fruits for commercial trade, ispredictable on the basis of fruiting phenol-ogy, whether plants are dioecious ormonoecious, on fruit accessibility, ondemand for the fruits and on tree tenure.All of these should be noted in the field.

The first three of these are biologicalfactors and are discussed below.Commercial sale of fruits or flowers is agood indicator of which fruits or flowersare in highest demand (see Chapter 3). Themajority of destructively harvested flowersor fruits enter commercial trade, but someare collected only for home consumption.In the Peruvian Amazon, for example, theseeds of hambre huayo (Gnetum leyboldiiand Gnetum nodiflorum) vines were notrecorded sold in Iquitos market, but thevines were down-pulled out of the forestto collect the edible seeds (Vasquez andGentry, 1989). From a plant biologyperspective, it is useful to record thefollowing as predictive factors of destruc-tive harvest of fruits.

Fruiting phenology

The timing of fruit release is a major influ-ence on harvesting method. If fruits falland can be collected from the ground, thenharvesting impacts are likely to be low. Ifthe plants are tall and the fruits thereforedifficult to reach, then felling for favouredfruits is likely. For this reason, you need tocheck whether the fruits are:

• shed as soon as they are ripe (orsometimes just before, so that finalripening takes place on the ground),such as in Brazil nut (Bertholletiaexcelsa – Lecythidaceae) or marula(Sclerocarya birrea – Anacardiaceae)trees;

• slowly released over a period of weeksor even months, the rest of the fruitsdisplayed to potential dispersal agents(birds, primates) in the canopy,something common with many palmspecies;

• serotinous, where seeds are held for 1to 30 years in canopy seed stores, areproductive strategy recorded for atleast 530 species in 40 genera of woody

Applied Ethnobotany

128


plants (Lamont et al, 1991), mostcommonly in the Bruniaceae, Ericaceae,Myrtaceae, Pinaceae and Proteaceae;when serotinous species are harvestedfor flowers or fruits, stems (withfoliage) are often cut as well, with theadded impact of removing new, long-lived and metabolically active leaves.

This issue is clearly illustrated in OliverPhillips’s (1993) survey of fruit accessibil-ity and local harvesting methods for over30 species of preferred edible fruit-bearingtrees in the Peruvian Amazon. Where l0 percent or fewer of edible fruits fell onto theground and fruit access height was between8 and 23 metres, then trees were felled.

Dioecious, monoecious orhermaphroditic?

Dioecious species bear male and femaleflowers on separate plants. Monoeciousplants have male and female flowers ondifferent parts of the same plant.Hermaphroditic plants have flowers withboth stamens and carpels. Dioecy iscommon amongst long-lived perennialplants, many of which produce large,edible fruits (such as the Anacardiaceaeand Palmae family). This has importantimplications for fruit and flower harvest-ing at the individual plant and populationlevels. At the individual plant level, whendestructive harvest of fruits (eg Palmae) orflowers (eg Proteaceae, Bruniaceae) takesplace in dioecious species, it is obviouslyselective of female plants. At a plantpopulation level, overexploitation offemale plants can totally disrupt breedingsystems. In addition, plant speciesharvested for flowers, particularly those inthe Bruniaceae and Proteaceae, are suscep-tible to fungal infection.

Access height

Are fruits accessible or out of reach of

human harvesters? Tall plants bearingpopular but inaccessible fruits are likely toget felled or, if vines or lianas, pulled out ofthe forest canopy. It is a good idea to recordhow harvesting behaviour is influenced byaccess height. In a study in the Amazon, forexample, Oliver Phillips (1993) recordedfruiting phenology, access height anddivided harvesting methods into fivecategories: Ground = collect felled fruitsfrom the ground; Picked = fruits picked byhand; Pole = fruits knocked (or pulled)down with a hooked pole; Climb = treeclimbed and fruits cut or shaken off; andCut = whole tree cut down for fruits. In myexperience, these are widely applicable inAsia and Africa as well.

Direct assessment of nutrient depletionor susceptibility to fungal attack requiresfield experiments backed up by laboratorywork. In short-term surveys, these indirectimpacts can be assessed in terms of crowndie-back or death (see Figure 4.14). Fieldresearchers have also used a rating systemin short-term assessments of the effects ofpicking serotinous flowers or fruits, whichrequires cutting branch stems as well. In astudy of the effects of foliage removal onmulti-stemmed, 1m to 3m tall Bruniaalbiflora (Bruniaceae) plants, Tony Rebeloand Pat Holmes (1988) rated plantsaccording to plucking intensity on a six-point scale:

1 dead with no evidence of plucking (K);2 killed by harsh plucking (D);3 alive, but harshly plucked (>90 per

cent of the estimated original foliageremoved);

4 alive and heavily plucked (>50 percent and <90 per cent of the estimatedoriginal foliage removed);

5 alive and lightly plucked (<50 per centof estimated original foliage removed);

6 alive and not plucked (U).

In addition, in each population of Brunia


129


albiflora, they recorded plant height andstem diameter at 30cm above ground level,and the number of individual seed heads(infructescences) produced in that year, forall plants in a series of one-metre widerandom transects. Knowing that groupedflower heads (conflorescences) areproduced on main vertical stems of Bruniaalbiflora plants every alternate year duringApril and May, and that these are retainedon the plant for at least four to six years,they were also able to estimate when theharvesting of seed heads had taken place.

More detailed studies on flowerharvesting in the Proteaceae in South Africa(Protea, Leucodendron) (Mustart andCowling, 1992) and Australia (Banksia)(Witkowski, Lamont and Obbens, 1994)have both shown the extent to whichcommercial flower picking results in adecline in plant canopy volume and seedproduction, affecting both individual plantsand plant populations (see Chapter 5).

Harvesting plant exudates (gums,resins and latexes)

Latex is a rubbery exudate tapped fromthe bark of trees and lianas in theApocynaceae (Dyera, Couma,Landolphia), Euphorbiaceae (Euphorbia),Moraceae (Ficus), Sapotaceae (Manilkara,Palaquium) and other plant families.Gums are water-soluble exudates, primar-ily from woody plants in the Leguminosae(Acacia, Astragalus) and Sterculiaceae(Sterculia). Resins are divided into twogroups, both tapped from tree bark. Thereare hard resins, such as copal from trees inthe Araucariaceae (Agathis) and damarfrom Dipterocarpaceae (Shorea, Hopea,Vatica), and soft resins (or balsams), suchas benzoin from Styrax (Styracaceae) andcopaiba from Copaifera (Fabaceae). Mostcommercially traded gums, resins andlatexes are produced from deliberatelydamaging tree or shrub bark. Exceptions

Applied Ethnobotany

130

Figure 4.14 Although used by Dawkins (1958) to rate tree-crown condition, due to variousreasons, including competition, crown ratings also reflect tree health generally, for instance as a

secondary response to bark or root use

0PERFECTComplete

circle

1GOOD

Irregularcircle

2TOLERABLE

Halfcrown

3POOR

Less thanhalf crown

4VERY POOROne or few

leafy branches

5DEAD

Definition of crown ratings0 = Perfect: excellent size and development, wide, symmetrical and generally circular in plan.1 = Good: slightly asymmetrical with some dead branch tips (‘silviculturally satisfactory’ to foresters).2 = Tolerable: markedly asymmetrical, some dieback.3 = Poor: extensive die-back, leaves form less than half original crown size.4 = Very poor: badly damaged, unlikely to survive.5 = Dead.


are gums and resins produced fromtapping roots (asafoetida from severalFerula species Umbelliferae and gumtragacanth from several Astragalusspecies) or fruits (East Indian dragon’sblood resin from the fruits of severalrattan palm Daemonorops species). Resinand latex exude from existing resin canalsin the bark and wood. In gum-producingspecies, the gum ducts only form after thebark is slashed (lysigenous ducts).

Despite low exudate yields, commercialtrade is amazingly high. In 1993, forexample, Sudan exported 11,410 tonnes ofgum arabic from Acacia senegal, Indonesia13,285 tonnes of jelutong latex from hilljelutong, Dyera costulata, and swampjelutong, D. lowii, trees (Apocynaceae), andIndia 1443 tonnes of karaya gum fromSterculia urens and S. villosa(Sterculiaceae). Given that exudate yieldsare relatively low – Acacia yields 250 gramsper tree per season, Dyera yields 3.5kg of

coagulated latex per tree per month, andSterculia yields 1 to 5 kg per tree per season– the number of trees involved in commer-cial exudate tapping is considerable.

In common with the fruit harvestingexample in the previous section, it is easy toassume that plant exudates can be sustain-ably harvested. This is certainly possible,offering the opportunity to tap trees byperiodically making cuts in the bark tomaintain a flow of exudate. In this way,trees and shrubs are maintained in steppe,savanna or forest, thus generating acommercial resource without felling.However, tapped plants do pay a high physi-ological price when nutrient-rich exudatesare removed, growing much slower andsometimes not fruiting at all (see Figure4.15b). In cases where communal or privateownership is established for exudateproducing plants, then sustainable harvest-ing generally takes place. Strict tenure isestablished in parts of Somalia, for example,


131

Source: Dijkman (1951) in Peters, 1994

Figure 4.15 (a) Methods of latex tapping from the rubber tree (Hevea brasiliensis). (b) Comparison of plantation rubber tree growth rates under two different tapping regimes

compared against untapped trees

35

30

25

20

15

10

5

0

Diameter growth (%)

1.0 5.0Years

2.0 3.0 4.0

No tapping

Normal tapping

Heavy tapping

(a) (b)


where Boswellia (Burseraceae) shrubs aretapped for frankincense (gum olibanum),with Boswellia stands belonging to extendedfamily groups (Coppen, 1995). Other caseswhere tenurial rights are established forexudate production are examples fromSudan (gum arabic from Acacia senegal)and southern Sumatra where the damargardens (kebun damar) of Shorea javanicaare a model of tropical agroforestry.

Unfortunately, cases of destructiveharvesting for exudate are common. Theclassic example is a Ferula species knownas ‘silphion’, tapped for an aromatic,medicinal gum for trade from North Africato the Roman Empire, and providing oneof the earliest examples of plant extinctionthrough overexploitation. Other cases aremore recent. Industrial companies estab-lished in Mozambique and South Africa inthe late 19th and early 20th centuries toextract latex from Landolphia kirkii lianastems all collapsed, overexploiting aresource which had been sustainablyharvested for its fruits for centuries(Cunningham, 1985). In Indian savanna,mukul trees (Commiphora wightii) and thesource of karaya gum (Sterculia urens)have also been heavily exploited.

The methods that can be used tomeasure harvesting impacts vary accord-ing to the tapping method and the timeavailable. The impact of the least damag-ing form of tapping is best measuredthrough long-term comparative studies oftree growth rates under different tappingfrequencies and intensities. In South-EastAsia, for example, high frequency tappingreduced the growth rates of rubber trees(Hevea brasiliensis) by up to 50 per centover a five-year period (see Figure 4.15b).

A similar approach could be taken in along-term comparison of growth incre-ments of tapped and untapped tropicalforest trees, such as Garcinia xanthochy-mus (Guttiferae), tapped for an orangelatex (gamboge) in order to dye Buddhist

robes in South-East Asia, and Copaiferaspecies (Leguminosae), tapped for resin inCentral America and South America.Copaifera trees are tapped for oleo-resinsin Brazil and Colombia by boring into thetrunk to tap resin-filled cavities. Indirectmeasurements of oleo-resin yields showthat although first tapping yields are high(up to 3 litres), these progressively declineover a 3.5-year period (Alencar, 1982). Theeffects of this process on tree growth orfruit production have not been studied.

Harvesting vegetative structures(leaves, roots, bark, stems)

Leaf harvesting

It is useful to think of the impact of leafharvesting in terms of plant growth ratesand life spans (see earlier section on ‘Leaflife spans’). At one extreme is the use ofedible wild greens and, at the other, leavesharvested for fibrous and long-lasting quali-ties. Collection of edible leafy greens(‘spinaches’) is generally selective for theyoung leaves of crop plants such asCurcurbitaceae or fast-growing plantspecies in recently disturbed sites(Amaranthaceae, Capparaceae, Tiliaceae),or leaves from annual or biennial above-ground stems and leaves that are producedfrom a perennial tuberous root, such as insome Cucurbitaceae and Acanthaceae.Edible leaves are generally gathered fromfast-growing species from early successionalstages, often in nutrient-rich sites such asagricultural fields, alluvial terraces alongriver banks, forest margins or canopy gaps,roadsides, or old livestock pens.

Long-lived leaves characterize long-lived, slow-growing plants (see Chapter 5).These plants invest more in physical orchemical defences. Harvesting leaves fortheir fibrous qualities (basketry, thatching,twine) involves leaf gathering from plantswith long-leaf life spans and slower

Applied Ethnobotany

132


growth rates, typically from later succes-sional stages. Two methods can be used toassess the impact of defoliation on plants.Ideally, both should be used, the first as aquicker ‘reality check’ and the other as acontrolled experiment to assess plantresponse to harvesting in a rural situationwhere intensive management outside of anexperiment may well be unlikely. The twomethods are listed below.

Retrospective counts of harvestedleavesPlants such as palms and ferns whichproduce large leaves are ideal for under-taking an initial ‘first approximation’ ofdefoliation rates based on a knowledge ofannual leaf-production rates from taggedleaves on plants in different size classes(see Figure 4.13). It is important to beaware that young palm seedlings (genets)or sprouts (ramets) may have differentlyshaped ‘sword’ leaves compared to olderplants. Harvesters generally harvest largerleaves from older palms, however, ratherthan seedling leaves, since in most caseslong leaflets are preferred for basketry orbuilding. Once the seedling stage is passed,palms generally bear bigger leaves as theyget older, with a relationship betweenleaflet size class (see Figure 4.16a) andstem size class. In some cases, however,petiole width can provide a statisticallysignificant correlation with palm-leaf sizeclass, and this has been used instead of leaflength (Joyal, 1996).

Unless you are working with palmswhich show no difference in leaf produc-tion rate with increase in stem size, youneed to take the increased annual leaf-production rate into account to avoidbiased results in a retrospective assessmentof the number of leaves harvested per year.In his study of the vegetable ivory palm(Phytelephas seemannii) in Colombia, forexample, Rodrigo Bernal (1998) devel-oped a regression equation to calculate

annual leaf-production rate (y), based onstem length (x), where y = 5.8 + 0.0065xas a way of avoiding the inaccuracy ofusing average annual leaf-production ratesfor plants of all sizes.

If you look at the youngest emergingshoot in the centre (apex) of a palm stem,you will notice that leaves are produced ina radial sequence. The youngest shoot isjust emerging, a newly emerged ‘swordleaf’ is next to it, and next to that a fullyopened leaf, and so on. Let us assume, forexample, that the stem size class you arelooking at produces five leaves per year. Ifyou count five leaves ‘backwards’ from themost recently emerged ‘sword leaf’, thiswould represent the annual leaf produc-tion over the past year. This is showndiagramatically in Figure 4.16b. Youshould also record the number of leavesthat have been harvested, browsed bycattle or damaged in some other way.Browsing by animals is recognizable dueto the jagged defoliation pattern comparedto the clean cut from harvesting of theleaves with a panga or sharp knife (seeFigure 4.16c and d).

Repeating this process for many stemsin a sample population enables you toassess relatively quickly the proportion ofannual leaf-production cut by localharvesters and the size-class selection forparticular plants, leaf sizes or leaf ages. Italso provides good insight into thecomplexity of multiple-use of the sameplants. The same palm, for example, mightbe used for its young leaves (weaving), oldleaves (housing), grazing by livestock orcut for palm hearts. A disadvantage is thatthis is not a long-term experiment whichenables the response of plants to defolia-tion to be compared against a control. Youalso need to bear in mind that the effectsof defoliation on individual stems of clonalramets is more complicated than on singleplants grown from seed. Clonal plants canpartially compensate for loss of leaf area


133


by reallocating nutrients from one part ofthe clone to another.

Simulated defoliationThis is applicable to any plant and hasbeen applied to woody plants, grasses andpalms. A typical case would be to select20 to 30 plants for a control treatmentwith no (0 per cent) defoliation, and thesame number in increasing, successivelevels of annual defoliation. These arecommonly 25 per cent, 50 per cent and100 per cent defoliation (Oyama andMendoza, 1990; Ratsirarson et al, 1996)or 30 per cent, 60 per cent and 100 per

cent defoliation (Mendoza et al, 1987). Agood example is the well-designed studyby Ana Mendoza and her colleagues onthe effects of defoliation on the palmAstrocaryum mexicanum (Mendoza et al,1987). In their study, they used threedifferent defoliation rates for palms in fourstages (seedlings, juveniles, immature andmature) over a four-year period. Theeffects of defoliation were measured interms of changes in the rate of leaf produc-tion, fruit production, palm survival andthe probability of reproducing.Furthermore, it is possible to compare theeffects of defoliation on male versus

Applied Ethnobotany

134

Figure 4.16 (a) Measurement of leaflet length in Hyphaene palms, showing point of measurementwhere the leaf base is oblique or ‘normal’. (b) Diagrammatic representation of ‘retrospective’ leafcounts viewing the palm stem apex from above for a palm producing five leaves per year. (c) and

(d) A straight cut across the petiole of the young ‘sword’ leaf or across the unopened leaf isclearly visible in retrospective leaf counts even two to three years later, and is also distinguishable

from the jagged browsing of palm leaves by livestock

Leafletlength

(a) (b)Year 3

Year

2

Year 1

(c) (d)


female palms, as Ken Oyama and AnaMendoza (1990) did with the dioeciouspalm, Chamaedorea tepejilote. If you areuncertain about experimental design, it isimportant that you consult a statistician.Useful text books on experimental designand statistical techniques are by Maxwelland Delany (1989), Sokal and Rohlf(1987) and Zar (1998).

An advantage of the experimentaldefoliation method is that you are able toassess the response of different size classesof plants to different levels of defoliation.A disadvantage is that defoliation experi-ments take a considerable amount of timeand effort if they are carried out properly.Defoliation treatments need to be contin-ued for several years – for example, toavoid being misled by short-term ‘compen-satory growth’ of leaves in the first year ortwo, which gives the impression that highrates of defoliation merely stimulate leafor fruit production.

The effects of defoliation are stronglyinfluenced by factors such as plant physi-ology, the intensity and frequency ofharvest, and habitat. Woody plants showgreat variation in their tolerance topruning (Lay, 1965; Trlica et al, 1977).Tolerance of defoliation also dependsupon whether defoliation takes placewhen total non-structural carbohydratereserves in the root and stem are low(Menke and Trlica, 1981). Plant speciespalatable to animals appear to be moretolerant of leaf removal. Defoliation oflong-lived large leaves from plants canhave a high impact on the plant, particu-larly if leaf removal takes place shortlyafter the leaf has matured. At high (greaterthan 50 per cent) defoliation levels, thisfrequently results in subsequent depressionof leaf productivity. Slow-growing ferns,for example, are vulnerable to defoliation.Milton (1987), in a study on the effects ofharvesting several fern genera (Blechnum,Polystichum and Rumohra), which are

harvested from Afromontane forest inSouthern Africa, found that defoliationresulted in smaller frond size and but nochange in frond production rate. Evenafter five years, seven-weeks ferns(Rumohra adiantiformis) had not recov-ered from defoliation for florists’ greenery(Milton, 1991).

Bark harvesting

Bark, produced by a thin layer of cambiumcells, surrounds the xylem tissues thattransport water and nutrients to and fromthe roots and leaves. Bark protects plantsagainst fire, fungal and insect attack, andbark removal can therefore have seriousconsequences for the plant. Bark harvest-ing is often highly selective for family,genera or species, based on particular barkqualities. Many species in theBombacaceae, Malvaceae, Moraceae,Sterculiaceae, Thymeleaceae, Tiliaceae andUrticaceae, for example, are used to maketwine for trapping, fishing, weaving orhome construction. Bark is often rich insecondary plant chemicals, and manyspecies in the Apocynaceae, Euphorbiaceaeand Loganiaceae, for instance, are widelyused for traditional medicines. Similarly,the Ebenaceae (Diospyros, Euclea) are arich source of phenolic compounds used asdyes. For this reason, species-specific barkremoval takes place for medicinal purposesor dyes.

Bark damage may remain visible formany years, enabling field assessments ofthe extent of damage to species popula-tions due to bark removal. Figure 4.17shows a simple seven-point scale for ratingdebarking damage on tree trunks. This canbe combined with diameter (dbh)measurements to assess bark damagewithin a tree population. It is important torecord factors that influence the intensityof bark removal, such as the following.


135


Applied Ethnobotany

136

Figure 4.17 A seven-point scale used for bark damage ratings. The photograph shows a harvesterremoving medicinal bark from an Afromontane forest tree

0No

damage

1<10%

210–25%

326–50%

451–75%

5RB*

(any %)

6TotalRB*

Definition of bark damageratings

0 = No damage1 = Small patches removed

(<10% of trunk bark) (usuallyby traditional healers forlocal use)

2 = Larger patches removed(10–25% trunk bark). Usuallyscarce species in high localdemand or small-scale,emerging commercial trade

3 = Large strips (26–50% of trunkbark), generally frompopular, open-access speciesin commercial trade

4 = Extensive bark removal(51–75% of trunk bark),popular species in large-scalecommercial trade, easy andopen access

5 = Ringbarking or girdling,where bark is completelyremoved around the trunk.This leads to death of manytree species, regardless oflevel of girdling

6 = Complete girdling, all trunkbark removed. At this stagetrees or large branches maybe felled or trees climbed tomaximize bark removal


• What is the purpose for which thebark is removed? For example, tradi-tional healers harvesting bark for theirown use, rather than for commercialpurposes, generally only require smallquantities of bark. In contrast, barktwine production requires long stripsof bark and large slabs of bark areneeded for making bee hives, canoes,cloth and aboriginal bark paintings.

• Is the outer or inner bark removed andis the cambium exposed? Cork, usedfor flooring and bottle stoppers, isproduced from the outer bark of thecork oak (Quercus suber), enablingbark removal on a rotational basis fromthe same trees. By contrast, theAustralian black wattle (Acacia mearn-sii), which is cultivated for itstannin-rich bark for the leather tanningindustry, and the cinnamon tree(Cinnamomum zeylanicum), which isgrown for its spicy bark, are both killedif bark is removed from the main trunk.For this reason, black wattle trees areused for fuelwood after being strippedof bark, the next generation growingup from seed, while cinnamon trees aregrown on a coppice rotation, based ona succession of individual stems that arefelled and stripped of bark.

• From what part of the plant is the barkremoved (trunk, main branches,secondary branches)? Depending uponthe tree species and bark quality, barkfor twine may be removed fromsecondary branches rather than themain branches or tree trunk. Barkfrom thinner branches of the hippo fig(Ficus trichopoda), for example, is asource of strong twine that is easier tostrip and roll into twine than barkfrom the trunk or main branches.

Research participants from the localcommunity will often know the reasonwhy bark has been removed. Be careful not

to confuse casual bark slashes encounteredalong paths through forest or woodlandwith bark removed for other purposes, orwith bark removed by animals. In manycases, careful inspection of the tree trunkor remaining bark will show teeth marksof bark-eating rodents such as porcupines,or tusk marks where elephants haveremoved bark. Similarly, cut marks from amachete or removal of more regularlyshaped blocks of bark often distinguishbark removal by people.

Although many people believe thatmost tree species regenerate their bark afterit has been damaged, this is not a commonresponse to bark removal. The effects ofbark removal on the plant depend uponplant physiology, bark chemistry andexudates, and the intensity and frequencyof bark removal as well as season, habitatand microclimate. In the field, you willnotice that a tree’s ability to withstandbark removal varies with different plantfamilies, species and even between individ-ual plants. Many Proteaceae andPodocarpaceae are highly susceptible tofungal infection or attack by wood-boringinsects. The endemic Afromontane foresttree Faurea macnaughtonii (see Figure4.18b) is one such example. In Australia,the cinnamon fungus (Phytophthora cinna-moni) poses a major conservation threat tomany endemic Proteaceae. On the otherhand, plants in some families, such as theEuphorbiaceae, Moraceae andCanellaceae, show a marked resilience tobark removal, in part because the cambiumis protected by exudates after barkremoval, enabling bark to regenerateoutwards from the scarred area rather thaninwards from the sides of the wound. TheAfrican baobab (Adansonia digitata) isvery unusual (and resilient) to barkdamage, with parenchyma cells rightthrough the old wood forming a callus-liketissue that seals off the wound andproduces new bark (Fischer, 1981).


137


Applied Ethnobotany

138

Source: (c) and (d) with permission from Robin Guy

Figure 4.18 (a) Resilience to bark removal due to bark regeneration by secondary xylem: baobabtrees (Adansonia digitata) debarked for twine in Malawi. (b) Vulnerability to removal of even asmall patch (five-centimetre long slash), showing fungal infection after small-scale bark removal

in the Afromontane endemic canopy tree Faurea macnaughtonii. (c) Complete girdling ofBrachystegia trees is a common seasonal practice in miombo woodland through much of Easternand South-Central Africa. (d) Dead Brachystegia – ironically, also the major source of honey to

bees and beekeepers


Frequency of bark removal is also animportant factor, but very few studies havebeen performed on the long-term effectsof periodic bark removal from tree speciesthat regenerate bark after it has beenremoved. There is no doubt, however, thatthe stress of bark removal, which resultsin loss of nutrients and moisture, fungalinfection or insect attack, even if it doesnot kill the tree, reduces the tree’s growthrate and life span. Discussions with barkharvesters can also yield useful insightsinto whether repeated bark removaloccurs and how harvesters feel that thisinfluences bark quality. In Uganda, Natalfigs (Ficus natalensis) may be debarked asmany as 30 or 40 times within the life of asingle tree (Picton and Mack, 1989), withharvesters considering that bark qualitieswere improved by harvesting, due toproduction of thinner, flexible bark.

Bark regrowth also varies with sitedifferences, season and microclimate.Trunk-bark regrowth in windswept, driersites is often poor compared withregrowth on trees in shady, moist sitesprotected from wind. A common EastAfrican example of sustainable barkharvesting based on bark regrowth andproduction is from Ficus natalensis, oftenplanted in fields as a boundary marker andbark source. The process of bark removaland regeneration, described by Picton andMack (1989), illustrates local knowledgeof the importance of microclimate and ofprotecting the trunk from desiccation afterbark removal:

‘The bark itself is easily removed.First the outer surface is scrapedoff, usually with a knife; thencircular incisions are made at theupper and lower levels of the maintrunk and a single vertical incisionis run down between the two. Atool made from a sharpened

section of a banana leaf spine, orin some cases a section of the fruit-bearing stem of the banana, isworked in behind the exposedlayer of bark which is carefullypeeled away in a cylindrical strip.The average dimensions for such astrip would be about 12 feet by 2feet. This process of stripping thebark does not damage the tree asthe whole of the exposed trunk isimmediately swathed in a bandageof banana leaves to protect it fromexposure to the sun and the wind.After about a week the tree hasalready regenerated sufficiently forthis protective covering to beremoved. Any subsequent damageto the newly forming bark may betreated by the application of apoultice of sheep’s dung.’

In common with the less damaging effectsof defoliation, it is likely that two otherfactors, season or phenological stage,influence the impact of bark removal,although few data are available on howbark harvesting is influenced by thesefactors. It is likely, however, that barkremoval during spring or summer wouldhave a higher impact on the tree.Knowledge of the time when it is easiestto debark Brachystegia trees, for example,is widespread amongst beekeepersthroughout the miombo (Brachystegia andJulbernardia) woodlands of Central andSouthern Africa, although it is unlikelythat Brachystegia trees would survivedebarking in any season. In species suchas Ficus natalensis, the easiest time todebark the tree, when the bark is looserduring spring or summer, may in fact bethe worst time for the tree itself.


139


Underground ethnobotany: roots,tubers, bulbs and corms

Nutrients and secondary plant chemicalsare often concentrated in roots and bark.For this reason, selected species areharvested for use as medicine or food.Selection also depends upon size and/orage, shape, chemistry (medicinal plants,traditional dyes, fish poisons or food),structure (qualities such as fibrous rootbark for twine or ‘corky’ root qualities forfishing floats) or function (as in well-developed underground storage organs forfood or medicines). Assessing the extent ofharvesting of underground plant parts isdifficult for two reasons. Firstly, removalof bulbs, corms, tubers or tap roots oftenmeans that the whole plant is dug out. Atmost, the only sign that there was a plantmay be a hole in the ground, which in

many cases will fill in with soil or beobscured by regrowth of other vegetation.Secondly, even when more obvious lateralor tap-root removal has occurred, withsigns of digging around the base of thestem, partial root removal is also difficultto assess as holes dug at an earlier stagemay have filled in, obscuring the trueextent of root damage.

When root harvesting is recent, it ispossible to use a field rating scale for rootdamage. In many cases, however, this isnot as useful as the rating system for barkdamage, since root damage is oftenobscured when the holes from which rootswere removed fill in with soil – forexample, when trees fall over after lateralroot removal or when shrubs are entirelyremoved (see Figure 4.19). Nevertheless,where root removal is recent, it is possibleto rate damage to trees and shrubs using

Applied Ethnobotany

140

Figure 4.19 One of the problems of damage assessment and monitoring that illustrates the needfor permanent plots: after a ten-year history of root removal for dyeing basketry, these stunted

Euclea divinorum shrubs, pointed out by local research participant Julius Rivero, have beencompletely uprooted, obscuring any future signs that they were ever there


the rating systems shown in Figure 4.20.Basal diameter (bd) or diameter at breastheight (dbh) should be recorded for eachplant. Unless root removal is recent andclearly visible, however, this approach islikely to lead to an underestimate ofdamage to both individual plants and ofoverall damage to the species population,as dead trees or shrubs killed by earlierroot damage may not be found.

Where damage is older, tree-crowndeath or complete tree die-off can takeplace after root removal, and it is useful tohave a parallel assessment rating rootdamage (see Figure 4.20) and crowncondition as a result of uprooting damage(see Figure 4.14).

Roots serve several importantfunctions in most plants. They anchor theplant in the soil, or on host plants in thecase of epiphytes or parasites, and theyalso absorb water and minerals. Foodsproduced in the leaves through photosyn-thesis move through the phloem tissue tobe stored in the roots. Although the rootsof many tropical and temperate plantspecies are harvested by herbalists, craft

workers or fishermen for use as medicines,dyes, fibres or fish poisons, experimentalwork on the effects of root removal hasbeen limited to studies on temperate anddeciduous fruit crops, such as apples andpeaches. One example is root pruning, anold-fashioned technique used to manipu-late the growth rate and leaf and fruitproduction of trees. As roots are not onlystorage organs, but also a place wheresome growth hormones are producedbefore being moved through the plant tothe shoots and leaves, the effects of rootharvesting can be very damaging to theplant. The impact of root damage dependson the following.

• Type of roots harvested: which rootsare removed? Unless infection occurs,many trees and shrubs appear tosurvive removal of some lateral roots.Tap-root removal, however, has a highimpact on most plants. Unless vegeta-tive reproduction takes place, bulb orcorm harvesting generally representsremoval of the whole plant. Vegetativereproduction, through formation of


141

Figure 4.20 A seven-point scale used for rating root harvesting damage. This is not as widelyapplicable as the bark damage scale, but is useful where recent destructive harvesting has takenplace when a ‘frontier’ population has rapidly been exposed to high intensity or frequency of

harvesting and where the soil shows clear signs of uprooting

0No rootremoval

1Removal of

<15% of thicklateral

2Removal of15–25% ofthick lateral



5Removal of

>75% ofthick lateral

6Total lateral

root and taprootremoval(d h)


clonal populations in a variety ofways, is an important factor that canbuffer the effects of root exploitation.

• Plant physiology: resilience to rootremoval varies with species. Forexample, damage to roots can stimu-late regeneration in some plants, butlead to fungal infection in others.Members of the Proteaceae andLauraceae family, in particular, aresusceptible to infection by the fungusPhytophthora cinnamoni after rootdamage. There is little known aboutthe more subtle effects of root pruning,such as an increase in secondary infec-tions through root damage.

• Rooting depth and distance: rootingdepth, and the distance that rootsspread, influence both how easily rootsare removed and the extent to whichroot removal affects the anchorage ofthe plant. Roots, bulbs or other under-ground plant parts near to the soilsurface are clearly easier to remove,and these will be selected unless thereis an incentive to remove more deeplyrooted plants. Harvesting often focuseson lateral roots, rather than on taproots, and on younger, more shallow-rooted corms, bulbs or tubers.

• Intensity and frequency of damage:this depends not only upon the type ofroots which are harvested, but on howoften root removal takes place, and onwhat proportion of the roots isremoved. Just as plants need tomaintain a certain leaf area for foodproduction through photosynthesis,they also need to maintain enoughroot surface area for absorption ofwater and nutrients, as well as foranchorage and food storage. Fieldobservation of lateral root removal fordyes, chewing sticks or traditionalmedicines shows die-off of trees whenmost lateral roots are removed. This isnot an uncommon situation where

there is commercial demand for theseresources, leading to local extinctionthrough uprooting – for example, withBerchemia discolor and Euclea divino-rum trees uprooted for dyes forcommercial basketry production inBotswana (Cunningham and Milton,1987; Cunningham, 1988), and withseveral Garcinia species (G. mannii,G. epunctata, G. afzelii) used forchewing sticks in West Africa.

The interactions between a lowlevel of root removal on leaf, floweror fruit production are less obviousand less well known, although somepossible effects can be inferred fromfruit crops. Root pruning in appletrees reduces fruit yield, growth andleaf production (Schupp, 1992). Inapple orchards, this is sometimes usedto reduce fruit drop and to improvethe colour and quality of the remain-ing fruits.

Field assessment: tree canopy(crown) health

During field work, you may have seentrees or shrubs showing die-back of thecrown due to age, bark or root removal,or secondary factors such as fungal infec-tion as a result of bark or root damage.This can be assessed quantitatively orvisually, as foresters do when using thesystem of ‘pre-emptive mortality’ used byAmin Seydack and his colleagues (Seydacket al, 1995a, b) in Afromontane forest inSouth Africa, which marks trees forharvesting shortly before they would dieanyway. The crown health rating is avisual rating method used by foresters(Philip, 1994; see Figure 4.14). It is also auseful indirect indicator of the effect ofbark or root harvest on what were oncehealthy trees. The crown health ratingmethod can be used together with directratings of bark or root damage describedearlier in this chapter.

Applied Ethnobotany

142


Stem harvesting

Tree, climber or shrub species and popula-tions vary in their vulnerability to stemremoval. Important questions to ask in thefield are: is the species characteristicallysingle stemmed or multi-stemmed? Whatare the selection criteria, if any? Do plantsregenerate after stem cutting or not? Doesthis vary between sites or with stem age ordiameter? Does the species have multipleuses and do these uses affect recruitmentby exploiting different stem size classeswithin the same population?

Where stem harvesting occurs,resprouting adds resilience to the individ-ual plant and plant population, so it isimportant to consider the importance ofseeds or sprouts (vegetative shoots) in theregeneration of harvested plants.Regeneration after stem harvesting can befrom seed or by sprouting (or both). Stemremoval of single-stemmed plants canreduce reproductive output of the speciespopulation, but this is less often the casewith multi-stemmed plants where severalstems remain due to selective harvestingor where rapid resprouting occurs. Byworking with knowledgeable local people,it is also possible and useful to identify cutstumps to species level and to record howthey were cut (with an axe, saw, chain-saw), the height of cut, the response tocutting (number and size of sprouts), andadditional impacts affecting recovery(such as browsing by livestock or wildlife).In addition, measure basal diameters todetermine the structure of sample popula-tions before harvesting occurred.

Most trees resprout to some extent,some very vigorously, some only weakly,and a few, such as various Podocarpus andRaphia species, not at all. The ability toresprout is also affected by stump size.Many tree species which resprout vigor-ously when the stumps are small do notresprout from the large stumps of maturetrees.

Growth rates after resprouting are a keyfactor in determining rotation times ifcoppice rotations are proposed as a manage-ment strategy. The ability to sprout inresponse to cutting does not automaticallyinfer an advantage to harvested individualsor populations over reseeders. Dirk Muir(1990), for example, who worked with Zuluwoodcutters in Hlatikulu forest, SouthAfrica, compared two tree species with verydifferent responses to stem removal for localbuilding purposes. The first species,umphatawenkosi (Strychnos usambarensis,Loganiaceae) regenerated very well (79 percent of cut stems) and the second, idlabatega(Chionanthus foveolatus, Oleaceae), regen-erated poorly, with only 24 per cent of cutstems resprouting. Populations of bothspecies were resilient to harvesting, however,presumably due to rapid seed regenerationin C. foveolatus.

Although quantitative studies of thestanding stock of a plant resource or of theeffects of harvesting start at the individualplant level, they are usefully put inperspective at the plant population levelthrough studies of harvesting impacts andyields. This synthesis is the goal of the nextchapter.


143


Although the response of individual plantsto harvesting impacts provides usefulinformation (see Chapter 4), it is crucialto avoid getting sidetracked when witness-ing destructive harvest at the individualplant level. Harvesting has to be seen fromthe perspective of plant populationdynamics, and harvested plant popula-tions in turn need to be viewed in terms oftheir abundance, distribution and howthey are influenced by disturbance at thelandscape level (see Chapter 6). Aseemingly low impact use, such as harvest-ing of fruits, for example, may have a highlong-term effect on populations of thatspecies, either because of long-term impacton seedling recruitment or because fruitcollection involves tree felling. On theother hand, even if harvesting bark, roots,or stems kills some individual plants, itmay have little impact on the populationsof fast-growing, fast-reproducing species.

Harvesting of sweet-thorn (Acaciakarroo) bark is a good example. At theindividual plant level, local peoplecommonly girdle trees for bark for makingrope, which may kill the tree (see Figure5.1). At a species population level, however,subsistence requirements for bark twine areeasily met, as dying Acacia karroo trees are

replaced by young, fast-growing plantsfrom a soil seed bank. In addition, humanand livestock impacts on the landscapefavour Acacia populations, as intensegrazing and exclusion of fire favour thesetrees and reduce competition from grasses.This is clearly seen in Figure 5.2, whereAcacia populations have encroached uponwhat was open woodland outside aprotected area, compared to the protectedarea itself, where open woodland has beenmaintained through controlled burning andlower levels of grazing.

Chapter 4 discussed methods formeasuring plant size, age and the impactsof harvesting on individual plants. Thischapter first discusses the biologicalfactors that influence the resilience orsensitivity of plant populations of differ-ent species to harvesting and how thissensitivity should guide decisions on whatspecies to monitor, including where andhow. It then describes methods used toassess what is available (standing stock),harvesting impacts and yields at the plantpopulation level. The necessary equip-ment, in addition to the items listed inChapter 4, includes: measuring tapes(ideally a 30m or 50m tape marked in cm),calipers, field data sheets, pencils, flagging

Chapter 5

Opportunities and Constraints onSustainable Harvest: Plant Populations

Introduction


tape and a compass. If you are going toestablish permanent plots, you will alsoneed aluminium tree tags, 7cm longaluminium nails, hammers, cable ties, andbright paint (blue or yellow). Aerialphotographs and, if available, large-scale

vegetation maps are also useful. If youdecide to develop population projectionmatrix models, you will also need a desk-top or lap-top computer and anappropriate spreadsheet programme.

Opportunities and Constraints on Sustainable Harvest: Plant Populations

145

Plant populations and practical constraints: selecting species

Figure 5.1 Good or bad, sustainable or not? For these Acacia trees, impact is high. They will notrecover from bark removal for twine. This impact also needs to be seen, however, against a

background of species population biology, where high growth rates and recruitment from a soilseed bank provide resilience against the death of debarked trees

When choosing which species will be thefocus of density, yield or harvesting-impactstudies, two important factors need to betaken into account. Firstly, is harvesting

considered from the perspective of locallivelihoods or conservation – species lossthrough overexploitation benefits neitherlocal people nor conservation in the long


term. Unrestricted access to a valued butvulnerable species may provide a highinitial harvest, but this will merely be atemporary ‘bonanza’ followed by loss oflocal self-sufficiency and higher effort orprices to get the species elsewhere. Whenconservation becomes the focus of high-impact harvesting, overexploitation alsoundermines the primary goal of anyprotected area: the maintenance of habitatand species diversity.

Secondly, the assessments of standingcrop, plant densities, yields or harvestingimpacts can be expensive and timeconsuming. While participatory ruralappraisal (PRA) methods provide usefulbackground information for studies ofplant population dynamics, there are no‘quick fixes’. At the same time, while thereis no substitute for quantitative studies ofplant population dynamics as a basis for

making decisions for conservation orresource management, these data can takedecades to collect. Recruitment of youngplants and mortality rates of plant popula-tions, for example, are generally derivedfrom long-term census data on markedplant populations in permanent sampleplots (PSPs). The dilemma many fieldworkers face is that decisions on resourcemanagement are often urgently required,yet hundreds of species are harvested, withlittle or no published data on their growthrates, biomass or demography. Goodexamples are the hundreds of medicinalplant species sold at markets in SouthAfrica (500 species) or Brazil (600species), including every plant life form,from herbs to forest trees. We thereforeneed to use whatever ‘short cuts’ we can,including conceptual ‘filters’ for choosingpriority species.

Applied Ethnobotany

Figure 5.2 Acacia population dynamics: high recruitment and high tree growth rates – coupledwith a habitat factor of disturbance due to removal of grass cover, with high cattle numbers andexclusion from fire – result in a high density and number of Acacia trees outside (left) compared

to inside (right) a protected area (Mkuze Game Reserve, Natal, South Africa)

146


‘Filters’ for choosing priorityspecies for monitoring

Whether you are walking through a marketor on a forest trail, knowing the scientificnames of the species you encounter isundoubtedly important (see Chapter 2). Itis also useful to think of grouping plantstogether in alternative ways, and several ofthese have been covered already. Harvestedplant species can be grouped according todemand – for example, the quantityharvested (see Chapter 2) – or according towhether they are used for subsistence orcommercial purposes (see Chapter 3). Theycan also be grouped in relation to the part

of the plant which is harvested, or byharvesting intensity and frequency (seeChapter 4). We know, for example, thatspecies in low demand or species harvestedat a low frequency and intensity are of lessconcern compared to species where barkor roots are in high demand (see Chapter3), and that a shift from subsistence tocommercial trade often results in an influxof harvesters, breakdown of local tenuresystems and an increased intensity andfrequency of harvesting (see Chapter 7).We also know that the likelihood ofoverexploitation and population declineincreases with how much is harvested and


147

Source: modified from Bennett, 1992

Figure 5.3 The effects of harvesting on a plant population depend upon what part of the plant isharvested and on the quantity, intensity and frequency of harvesting. Most harvesting has someeffect, but local extinction is infrequent and extinction of the harvested species even rarer. When

extinction occurs, it is usually a product of habitat destruction coupled with commercialharvesting of restricted range species

HIGH

LOW

Inte

nsi

ty a

nd

fre

qu

ency

of

trav

el

None

Sustainable

Populationdecline

Localextinction

Completeextinction

Quantity harvestedLOW HIGH


Applied Ethnobotany

148

BOX 5.1 PREDICTORS OF RESILIENCE OR VULNERABILITY TO

HARVESTINGPredictors of resilience or vulnerability to harvesting based on geographic distribution, habitat specificity, localpopulation size, growth rates, part of the plant used, variety of uses and reproductive biology. These need to belinked to social and economic factors (commercial versus. subsistence use) (see Chapters 2 and 3) and strength oflocal land or resource tenure systems (see Chapter 7). An additional ‘flag’ from an international plant conserva-tion perspective (but not relevant to village-level harvesters) would be phylogenetic distinctiveness.

Opportunity for sustainable harvest Predictors and Criteria High Medium Low information sources

Geographic Wide Limited Restricted Herbarium and distribution distribution records1

Habitat Broad Restricted Highly Life-form (eg widespread specificity specific ruderals (therophytes), clonal

sprouters, nitrogen fixing or not)Local Often Large to Everywhere Herbarium and population sizes large medium small distribution records

Growth Rapid Fairly rapid Slow Tree seedling Relative Growth rates Rates (RGRs) and Specific Leaf

Area (SLA) ratios,2,3 data on productivity, growth rates and biomass production

General Leaves, Exudates, Whole plant, bark, Case studies of resource resource flowers, deadwood roots, apical resilience or crashgroup fruits phloem sap meristems (non-

sprouting palms)Single Single use or non- Few uses, Multiple,conflicting Local market and household versus competing use low conflict uses of several size surveys, impact and multiple species (eg fruits, between uses classes (eg different regeneration studiesuse leaves) stem size classes,

bark and roots)Reproductive biology4

Pollination Wind and other Common biotic Highly specific Pollination ecology and study ofabiotic pollination. pollinators pollinator mutualisms the mutualism between plants

Many viable seeds or (insects, (eg temperate and pollinators or dispersers5

asexual reproduction birds) shrublands of SW (clonal sprouters) Australia and South

Africa (specific flies,bees, beetles, butterflies)

or bats/nectar feedingbirds in tropical

rainforests)Dispersal Wind, water Common, generalist Reseeders/weak Ecological studies of dispersal,

(or other) biotic dispersers resprouters, dispersal local knowledge of animal/birdabiotic dispersal (small birds, by large (frequently feeding behaviour

small mammals) overhunted) mammals or large birds

Costs and Vegetation Savanna or High diversity Time required per ha forcomplexity: dominated by few woodland with (>50 spp/ha) low adequate monitoring andmonitoring (1–3) species. low diversity biomass/species, management of sustainableand High plant (less than 10 long plant lifespans, harvestmanagement biomass and tree species high impact

production (eg >10cm dbh/ha) harvest (bark,grassland or swamp or grassland roots, stems,

dominated by with timber, whole Poaceae, Cyperaceae low medicinal plants)

or Juncaceae or geophyte‘oligarchic’ forest diversity

(eg palm dominated))Phylogenetic Species in large Species in Species in monotypic Taxonomic reviews6

distinctiveness genera (eg medium to families andEuphorbia, large genera monotypic generaAstragalus)

Notes: 1 Rabinowitz, Cairns and Dillon, 1986; 2 Reich, Walters and Ellsworth, 1992; 3 Midgely, Everard and vanWyk, 1995; 4 Peters, 1994; 5 Bond, 1994; 6 Mabberley, 1987


the intensity and frequency of harvest.These simple categorizations, summarizedin Figure 5.3, help prioritize species forstudies of density, regeneration rate, yieldor harvesting impact at a plant populationlevel.

Think of this as a series of ‘filters’which help to sift out species that arelikely to be resilient or more vulnerable tooverharvesting. In Box 5.1, for example,plant resources are grouped according to

their potential for sustainable harvest,from high potential (leaves, flowers, fruits)to low potential (bark, roots or the wholeplant).

Each of these coarse filters can besubdivided again on the basis of functionalecological attributes and field observation(see Box 5.2).

Despite the coarseness of these filters,thinking about plant species in this way isa useful process towards a ‘first approxi-


149

BOX 5.2 VULNERABILITY OR RESILIENCE TO HARVESTING

Although harvesting of different plant parts can be grouped into lower-impact (leaves, flowers, fruits) andhigher-impact uses (bark, roots, stems, whole plant), each of these can be subdivided according to the biology ofthe plant species concerned. While tapping of exudates or harvesting deadwood can be classed as lower impactuses, this varies so much according to harvesting method (eg selective cutting versus felling the entire tree) thatthese are excluded here.

Opportunity for sustainable harvest1

Criteria High Medium Low Predictors

Lower impactLeaves Rapid leaf growth Slow leaf Very slow leaf Leaf-life spans, Specific

and production rates. production rates, production rates, Leaf Area (SLA) ratiosDeciduous long leaf lifespans. very long leaf

Evergreen lifespans and few leaves (1–10 leaves/yr)

Fruits/flowers Many small Medium size Few, large flowers/ Fruit size, serotiny,flowers or fruits flowers or fruits, fruits, serotinous seedling bank versus

produced annually periodic (1–2 yr) or mast-fruiting, soil seed bank of orthodoxand abundant dioecious and/or seeds, dioecy versus

recalcitrant monoecyHigher impactBark Sap protects Good regeneration Cambium dies in Field observation, impact

cambium rapid in favourable sites area where bark is assessments in sites withbark regeneration (eg Warburgia removed, sensitivity high harvest levels

(eg some Moraceae) (Canellaceae) to fungal or insectAdansonia attack (eg Proteaceae,

(Bombacaceae) Podocarpaceae)Stems Vigorous resprouters, Resprouters, Reseeders, no/weak Reproductive biology,

small size classes used small/medium resprouts, growth rates, impact(= rapid rotation times) size classes hapaxanthic, dioecious, assessment in high impact

or stems of large cut large individuals sitesPoaceae (eg bamboo) cut

Roots High production Intermediate Low root production Field observation, impactrates, sap production rates, sensitive to assessment in high harvestseals wounds. Sexual fungal attack. Slow level sites. Reproductivereproduction (from growing resprouters biology and growth ratesseed) common as that rarely reproduce are useful predictors

well as being clonal from seed are veryresprouters vulnerable

Whole plants Common, fast Intermediate (eg Long-lived, slow Life-form, reproductivegrowing ruderals some Cyperaceae growing, hapaxanthic biology, plant(eg Asteraceae, and Fabaceae) species with long architecture,2 growthAcanthaceae, times to reproductive rates

Amaranthaceae, maturityPoaceae, Chenopodiaceae)

Notes: 1 modified from Peters, 1994; 2 Halle and Oldeman, 1970


mation’ of plant species where there is ahigh or low opportunity for sustainableharvest. The next sections describe thetheoretical background to some of these‘filters’. Readers who are not directly

concerned with these issues may prefer togo directly to the next section: ‘Costs andComplexity: Inventory, Management andMonitoring’.

Applied Ethnobotany

Plant life forms

150

Bridging gaps in knowledge: life forms, plant architectureand reproductive strategies

Developing an ecological perspective thatcuts across taxonomic boundaries at aspecies level is a useful tool for groupingharvested plant species and setting priori-ties for detailed studies of density, yield orharvesting impacts. In almost any society,people classify plants according to theirgrowth form – for example, as trees,shrubs, vines, lianas, epiphytes or grasses.For centuries, botanists and ecologistshave developed classification systemsbased on plant attributes that group plantsinto different categories. Several of thesemay already be familiar to you, such asplant life form (Raunkiaer, 1934), plantarchitecture (Hallé and Oldeman, 1970),leaf characteristics (Reich et al, 1992),reproductive strategies (reseeders versusresprouters) or seed characteristics (ortho-dox versus ‘recalcitrant’ seeds) In forests,it is possible to identify ‘guilds’ of plantsbased on successional stage (early pioneer,late secondary, primary forest species), orto group plant species according to their

light requirements, such as shade-tolerantversus light-demanding species, or ‘gap’versus ‘non-gap’ species. As you walkthrough grassland, savanna or along aforest trail, coming out of deep shade anddappled light into sun-filled gaps edgedwith vines and large-leaved shrubs andyoung trees, it is a useful exercise to thinkabout how the plant species you see fitinto these categories and how different‘guilds’ of species are influenced by distur-bance, such as fire, drought, disease, treefalls or clearing of habitat.

These classifications get away from thedetail of species-level identifications andenable plants to be sorted into functionalecological groups; they give a betterinsight into plant population dynamicsand improve our ability to predict theimpact of harvesting. For this reason, theyare described here for field researcherswho want background information onwhy these ‘ecological filters’ are useful.

In the early 1900s, the Danish botanistChristian Raunkiaer categorized plantsinto ‘life forms’ (Raunkiaer, 1934). Thiswidely accepted classification systemgrouped different vascular plants accord-ing to the height of mature individuals, thetype of shoot systems (such as woody orherbaceous, climbing or self-supporting,

and whether plants were found in wateror on land). Raunkiaer also took intoaccount the form and location of the budsor storage structures, such as bulbs, corms,rhizomes or tubers, which enable peren-nial plant species to persist from oneseason to the next. Although he usedrather cumbersome names for different life



151

forms, thinking about plants in terms oflife-form categories is useful in establish-ing basic resource-management principles,since these represent a sequence from largetrees (‘mega-phanerophytes’) and shrubs(‘micro-phanerophytes’) through toannual herbs (‘therophytes’). This helps tobridge the gap in knowledge about plantpopulation dynamics, enabling a firstapproximation of categories of vulnerabil-ity to destructive harvesting. Large trees(‘mega-phanerophytes’) commerciallyharvested for their medicinal bark, forexample, represent a vulnerable categorywhen there is selection for thick bark fromlarge (old) plants which have a long periodto reproductive maturity, a low ratio ofproduction to biomass and specializedhabitat requirements.

Plant architectural models

In trying to predict different impacts onthe vast range of plant species harvested,one of the ways of simplifying things is togroup the species according to theirgrowth pattern and form. Plant architec-tural models were developed throughpioneering studies by French botanistsFrancis Hallé and Roelf Oldeman (1970).Adrian Bell’s recent book (1998) is anexcellent illustrated guide to plant archi-tectural models. Architectural modelsbegan with studies of tropical trees as asystem for explaining their structure andgrowth, but has been expanded to includetemperate and tropical shrubs, lianas andherbs. Architectural models can provide auseful predictive tool for studies onharvesting in several ways:

• They provide a theoretical basis forunderstanding different levels of sensi-tivity to stem harvesting or harvestingof apical meristem – the tissueinvolved in stem growth (such asharvesting palm-hearts).

• Plant species are chosen whose archi-tecture and rhythmic growth patternsenable age estimates in a range of verydifferent plant families, such as thosein ‘Corner’s model’ – for instance,palms, tree ferns and grass trees(Xanthorrhoeaceae), whose architec-ture lends itself to field assessments ofleaf production and leaf harvestingrates, assessments of how stem harvest-ing impacts branching (see Chapter 4),past disturbance events by fire(Lamont and Downes, 1979) or treefalls (Martinez-Ramos et al, 1988).

• Tree stems are aged based on scars thatare visible from modular stem growth.In developing architectural models,great interest was shown in ‘iteration’,the repeated process of stem growthand branching as trees matured fromsaplings to maturity (‘Massart’smodel’; Podocarpaceae and someAraucariaceae; see Chapter 4).

Resprouters, reseeders andresilience

The way in which plants reproduce is animportant issue in conservation andresource management. Categorizing plantspecies in terms of where they are on thecontinuum from ‘reseeders’ (which regen-erate primarily from seed) to ‘resprouters’(which reproduce clonally through theproduction of new shoots) gives betterinsight into the potential for sustainedyield harvest or regeneration studies (seeBox 5.3).

An investment of time and effort inmonitoring the impact of fruit harvestingmay be very appropriate in tall tropicalforest, for example, where many canopytrees regenerate from seed (Peters, 1994).A similar focus on regeneration from seedwould not be a priority in short subtropi-cal thicket or forest in southern Africa,where all common canopy species (with


Applied Ethnobotany

152

BOX 5.3 CHARACTERISTICS ACROSS A CONTINUUM: LONG-LIVED

RESEEDERS VERSUS RESPROUTERS

Reseeders

• Reseeders regenerate from seed, some maintaining canopy seed banks (‘serotiny’).• Examples are common in the Proteaceae, Pinaceae, Ericacae and Podocarpaceae

families.• Reseeders are single stemmed, not multi-stemmed. Examine smaller shrubs closely.

Some reseeders are single stemmed, but branch off close to the ground, giving theincorrect impression that they are multi-stemmed reseeders.

• They do not resprout when the stem is cut.• They usually are self-pollinated or have diverse pollinators. If there are exceptions

dependent upon specialist pollinators or seed dispersers, then they are vulnerableto extinction.

• Their seeds often germinate faster than those of resprouters.• They produce abundant seedlings (a large ‘seedling bank’).• They have higher stem growth rates than resprouters, since they allocate nutrient

resources into growing upwards, rather than into underground storage organs. Asa result, reseeder species in a particular vegetation type tend to be taller thanresprouters.

• Most reseeders are short lived compared to clonal resprouters.• They often are habitat specialists (wetlands, moist montane sites, cool temperate

forests).• Their annual reproductive output is generally higher than in resprouting species.

Resprouters

• Resprouters maintain ‘bud banks’ rather than seed banks, regenerating clonally bysprouting rather than from seeds.

• They are often multi-stemmed, some shedding stems as they get older.• They produce new stems from buds which are above or below ground level (basal

or upper trunk sprouting).• Resprouters’ cut stems show obvious signs of resprouting (but be careful: resprout-

ing vigour declines with tree size or age).• Resprouters may have large, underground storage organs (rhizomes, tubers, ligno-

tubers) or lateral runners (eg many forest lianas).• Recruitment from seed is infrequent and irregular.• They may be pollinator-limited, but can still maintain long-lived clonal populations

consisting of a genetically identical clonal organism (the genet), which is made upof ramets, sprouted from buds, each of which has the potential to grow and repro-duce as an independent, individual plant.

• Few seedlings occur in the population; most small plants are ramets.• Resprouters grow more slowly than reseeders, since they have to put resources into

underground storage organs and into production and protection of buds.• Resprouters are usually generalists, found in a wide variety of habitats, rather than

habitat specialists.

Note: Recommended reading on this subject includes the following texts: Bond (1994); Harper (1977); Kruger,Midgley and Cowling (1997); Lamont et al (1991); Midgley (1996); Sakai et al (1997); Schutte et al (1995).


the exception of tall succulent species suchas Aloe and Euphorbia) reproduce bysprouting, and the few seedlings bear norelation to frequency of those species inthe forest or thicket canopy (Midgley andCowling, 1993; van Wyk et al, 1996).

Plants that reproduce solely from seedfollow a very risky lifestyle, particularly ifthey have very specific pollinators or seeddispersers. In most cases, however, thisrisky lifestyle is rare, and reseeders usuallyhave generalist pollinators or are windpollinated (Bond, 1994). Unlike animals,where small populations are particularlyvulnerable to extinction, small populationsize in plants is less closely linked to localextinction because so many plants canresprout. Plants that are able to persistthrough resprouting are far less dependentupon new recruits from seed. The rareSouth African fynbos shrub, Ixianthesretziodes (Scrophulariaceae), is a goodexample, still persisting through resprout-ing despite the extinction of its pollinator,an oil-collecting bee (Steiner, 1993). Rootharvesting of plants which depend uponresprouting for reproduction must be‘flagged’ as a potential threat.

Reseeders either build up seed storesin the soil or through canopy seed storage.Canopy seed storage where mature seedsare held for some time before they arereleased is termed serotiny. Unlikeresprouters, reseeders typically do notsurvive disturbance events, such as fire,but recruit from seed. Many Proteaceae,for example, wait for a fire to passthrough and stimulate seed release. Asreseeder populations can be seed limited,special care has to be taken in harvestingtheir flowers, fruits and seeds, particularlyin species producing few, large seeds. Sincemany reseeders in the Proteaceae,Bruniaceae (Brunea) and Ericaceaefamilies produce spectacular flowers, theyare harvested for a massive cut-flowerindustry. Much of this harvesting is still

from wild populations. For this reason,controlled harvesting of flowers and seedheads of reseeders is a very real manage-ment issue in South Africa and Australia(see also Chapter 4).

Although reseeders produce greaternumbers of vigorous seedlings and aretherefore more likely to recover from local-ized overutilization of saplings, they arevulnerable to overexploitation of the largetrees which produce the most seeds. This isclearly seen in the predictable overexploita-tion and slow recovery of Podocarpus andAfrocarpus (Podocarpaceae) populationsafter logging in Africa. Palms that generateprimarily from seed rather than byresprouting are also vulnerable to popula-tion declines through stem use. TheSouth-East Asian rattan palm, Calamusmanan, is a good example. Single stemmed,and growing up to 171 metres long (twicethe height of a California redwood), theywere once widespread in Malaysia andIndonesia and very abundant in west-central Sumatra (Dransfield, 1974; Siebert,1991). By the late 1980s, only a fewthousand mature individuals were left(Dransfield, 1987), many populationshaving been eliminated by overexploitation.The most vulnerable reseeders of all arecommercially harvested, relatively long-lived monocarpic (or hapaxanthic) palmsand trees. The term ‘monocarpic’ refers totheir strategy of flowering once, settingseed, and then dying. If mature stems arecut prior to setting seed, then the plant’sentire investment in reproduction is lost.

Rather than build up a seed bank asreseeders do, resprouters develop a ‘budbank’ and respond to damage by producingnew stems from these buds. In someinstances all the above-ground stems areremoved during disturbance, for exampleby fire, herbivores, humans or hurricanes.Most woody plants in tropical and subtrop-ical savanna, woodland and short forestsare resprouters, and many tropical and


153


subtropical palms survive through resprout-ing in the hurricane-prone monsoonaltropics. In fire-prone systems such asAfrican and Australian grasslands, savan-nas and woodland, many resprouters havelarge roots, rhizomes or specialized ligno-tubers (James, 1984; Hansen, Pate andHansen, 1991). These store starch andnutrients, and after disturbance such as fire,produce new stems from dormant buds.

Resprouters are not only found infrequently disturbed vegetation types,however. Large-scale disturbance due tofire, hurricanes or mass damage by largemammals such as elephants are rare eventsin temperate and many tropical forests, yetmany trees in these forests are resprouters(van Wyk et al, 1996; Sakai, Sakai andAkiyama, 1997). Although resilient to stemdamage, resprouters are poor colonizers ifoverutilization causes local populationdeclines. Root harvesting, for example, fordye or medicinal purposes, has seriousimplications for their survival becauseresprouters often produce few viable seedsor seedlings. Once the parent plant has diedthrough root removal, re-establishmentfrom seed will be a rare event. Anotherfactor is that the seedlings of resproutersgrow very slowly because some resourcesare allocated to storage products, protec-tion, dormant buds and a number of stems.

Light and leaves, gaps and growth rates

One of the commonest ways in whichforest trees are grouped is according totheir light requirements, such as shade-tolerant versus light-demanding species or‘gap’ versus ‘non-gap’ species. Alsocommon is to group plants on the basis ofsuccessional stage (eg early pioneer, latesecondary, primary forest species). Aclassic example of similar looking treeswhich occupy similar positions in a forestis the large-leaved pioneer Musanga (tropi-

cal Africa) and Cecropia (tropical LatinAmerica) (see Figure 5.4a). At the otherextreme are shade-tolerant, slow-growingspecies such as Podocarpus (see Figure5.4b). These groups of forest plants haveparticular ecological characteristics incommon (see Table 5.1), many of whichinfluence the opportunity for sustainableharvesting, as summarized in Box 5.1.

As you become more familiar with thespecies in your study area, it is useful tothink about the harvested species in termsof their ecological characteristics. You mayalso find anomalies to the three basiccategories in Table 5.1. The Brachylaenahuillensis trees which support the bulk ofthe Kenyan woodcarving industry are oneanomalous example: light-demanding,regenerating from small wind-blown seedsin canopy gaps and with weak resproutingability, one would expect from this habitand from the plant family (theCompositae) that they would be fastgrowing with light timber and a short lifespan. On the contrary, the beautiful goldentimber is dense, the growth rate slow andthe trees very long lived, taking 100 yearsjust to reach 40cm in diameter and livingfor over 200 years.

From field observation, it is clear thatlight-demanding ‘pioneer’ trees often havelarge, thin, short-lived leaves, and thatshade-tolerant forest trees often have thickleathery leaves with longer leaf life spans.The leaves of pioneer species are alsocommonly eaten by insects, mammals orpeople, while the leaves of many shade-tolerant species are often unpalatable andrarely eaten. In some cases, field observa-tion is not enough and quantitativemethods for studying leaf characteristicsare used. One method that has been usedfor its predictive value in plant populationbiology is specific leaf area (SLA), the ratioof leaf area to oven-dry leaf mass.Measurements of SLA have been rarelyused in applied ethnobotany, but have

Applied Ethnobotany

154


important predictive value for resourcemanagement for individual plants, plantpopulations (see Chapter 5) and distur-bance. There is a significant correlation,for example, between SLA and leaf lifespan, relative growth rate, photosyntheticrate and leaf nitrogen content (Reich et al,1992; Midgley, 1995). Fast-growing treesgenerally have thin, short-lived leaves andhigh SLA values. Slow-growing trees havethick (often sclerophyllous), long-livedleaves, slow growth rates and low SLAvalues (see Figure 5.5).

To date, SLA measurements haveapparently only been applied to plants withsmall- to medium-sized leaves. For broadsurveys, leaf area is measured for a sample

number of leaves (usually six to ten)collected from well lit branches (generallyin gaps or ecotones) of each tree speciesbeing investigated. SLA is calculated bymeasuring the area of fresh leaves (cm2) anddividing this by oven-dry leaf mass (g)(Midgley et al, 1995). This is done for allthe leaves in the sample in order to get amean for each species. Leaf area can bemeasured with an electronic leaf area meter.As a field exercise, leaf area can also bemeasured by tracing the outline of a leafonto graph paper, counting the squares andconverting this to area (cm2). Leaf mass hasto be measured with an accurate balance.

A rapid way of getting an approximateindex of shade-tolerance is to collect infor-


155

Figure 5.4 (a) Musanga (left) and Cecropia (right), large-leaved, light-demanding, fast-growingpioneer trees with short leaf life-spans (often less then one year) and high specific leaf areas,

occupy ecologically similar positions in tropical forest in Africa and South America. Herephotographed in Cameroon, where Cecropia has become an invasive species. (b) Podocarpus, at

the other extreme, is a shade-tolerant, long-lived tree with a long leaf life-span (four to five years)and a low specific leaf area

a b


mation on specific leaf area (SLA)(Midgley, van Wyk and Everard, 1995). Todetermine SLA, fully expanded young sunleaves must be collected and pressed. Theymust then be dried and weighed. The areaof each leaf must be determined (using aleaf-area meter, graph paper or gravimetri-cally). SLA is then the leaf mass divided byarea. Evergreen trees in the familyPodocarpaceae (white pines or yellowwood trees) typically have very low SLAvalues. By contrast, fast-growing decidu-ous species have larger, lighter leaves andmuch larger SLA values. Studies in SouthAmerica have also shown that SLA valuesare closely linked to leaf life span (Reich etal, 1991; Reich, Walters and Ellsworth,1992). Trees with low SLA values areshade-tolerant species and have long leaflife spans. Podocarpus leaves, for example,live for four to five years. Trees with high

SLA values are shade intolerant with short-lived leaves and fast growth rates, typicallyoccurring in canopy gaps. In a survey ofover 100 tree species in southern Africa,the highest SLA values were found indeciduous legumes (Erythrina, Acacia andErythrophleum), with short (less than oneyear) leaf life spans. Other leaf attributes,besides high SLA and short leaf life span,which indicate fast growth rates and shadeintolerance, are if the leaf is compound, ifthe leaf is large and if the leaf has highlevels of nitrogen (most of which are inphotosynthetic enzymes) (Midgley, vanWyk and Everard, 1995). At the scale ofdifferent forest types, the southern Africasurvey of SLA values grouped trees intothose with low SLA values (Afromontaneforests) and high SLA values (subtropicalcoastal forests on sands).

Applied Ethnobotany

Table 5.1 The basic ecological characteristics of early pioneer, late secondary and primarytropical forest species

Character Early pioneer Late secondary Primary

Distribution very wide very wide usually restricted;many endemics

Seed dormancy well-developed slight to moderate noneSeed or fruit size small small to intermediate largeSeed dispersal birds, bats, wind mainly wind, but also mammals, birds

mammalsShade tolerance very intolerant intolerant seedlings very tolerant,

later intolerantGap size required large intermediate smallSeedling abundance very scarce usually scarce abundantGrowth rate very fast fast slow to very slowWood density light light to medium very hardLife span 10 to 25 years 40 to 100 years, 100+ years

sometimes more

Source: Peters, 1994

156

Costs and complexity: inventory, management and monitoring

Once you have decided on a group ofpriority species, you need to map the areaavailable for harvesting, perhaps with

aerial photographs, and use this as a basisfor sampling populations of the selectedspecies. Sampling can be fairly quick and


straightforward or very time consuming,depending upon the diversity of thevegetation and biology of the speciesinvolved. What is often underestimatedare the costs of monitoring programmes.The complexity and costs of inventory,management and monitoring all increaserapidly with an increase in the diversity ofspecies and uses, number of users, quantityharvested and number of harvesters perunit area. When time and funds arelimited, before you design a sampling ormonitoring system you need to askyourself several questions, such as:

• What is the overall objective (egmaintaining biodiversity or catchmentvalues)?

• What question(s) are you trying toanswer?

• How precise do you want to be (egstatistical precision to 5 per cent, 10per cent, etc)?

• Who will do the work, and what train-ing needs are there before this can start?

• What is the control (ie compared towhat – for example, comparisons

along harvesting gradients)?• Who will analyse the data?• Who will act on the results (and who

will translate these into a suitableformat for decision makers)?

• What is the spatial and temporal scale(ie time and rate of change; how bigand where)?

• What other factors are also affectingthe same resource (and how can thesebe distinguished from what you aremonitoring)?

• How long will it be before decisionsregarding management options can bemade?

It is not worth spending a huge amount oftime and resources on relatively unimpor-tant questions. Monitoring systems that donot take these factors into account are liter-ally designed to fail as they will beunsustainable in terms of either funds, time,technology or personnel requirements.

The simplest situation is with densestands dominated by a single species withone main use such as Cymbopogon thatch-ing grass, or wetlands dominated by


157

Source: Reich, Walters and Ellsworth, 1992

Figure 5.5 (a) Specific leaf area in relation to leaf life span. (b) Relative growth rate of youngplants for a range of species from different ecosystems

1000

100

10

Specific leaf area (cm2/g)

1 10 100Leaf lifespan (months)

r2 = 0.54

10,000

1000

100

0

Relative growth rate per week (mg/g)

10 100 1000Leaf area ratio (cm2/g)

r2 = 0.68

(a) (b)


Phragmites reeds or Cyperaceae. Thesespecies usually have a wide distribution,producing stands with a high biomassproduction of annual stems resilient toharvesting. As a result, there is a ‘widemargin for error’ between sustainable useand overexploitation. For this category ofspecies, disturbance events such as fire ordrought can have far more significantinfluences than harvesting itself, and mustbe taken into account in resource manage-ment and monitoring (see Chapter 6). Theoccurrence of large single-species, single-use stands also makes the assessment ofresource stocks and yields far simpler. Atthe other extreme is vegetation with a veryhigh diversity of multiple-use, long-livedspecies. This poses a particularly complexcase which can be expensive in terms oftime and funds for thorough inventory andmonitoring. Sustainable harvest of timberfrom Afromontane forest, which has arelatively low species diversity comparedto tropical forest in South-East Asia orAmazonia, provides a good example. Withjust a single product (timber) and fewspecies involved, a team of one forester andtwo staff selecting trees greater than 30cmdbh for logging is only able to cover 5haper day (Seydack et al, 1995). Costs alsoincrease rapidly with an increase in speciesdiversity and a decrease in the diameter sizeclass of trees that need to be monitored.

The most attention, therefore, needsto be given to species which are in highdemand, slow growing and habitat-specific, and which are harvested at a highfrequency and intensity. An adaptivemanagement approach should be taken tothis category of species, with harvest levelsbased on yield studies and where harvest-ing impacts and regeneration aremonitored and the results used to makeharvest adjustments. Figure 5.6 shows aflow chart for sustainable harvest of non-timber resources from tropical moist forestwhich uses this approach.

If sustained use is not possible at all,then the emphasis needs to be placed onidentifying and developing alternatives toactually or potentially overexploitedspecies. This can be done throughagroforestry or plantation production orappropriate technology alternatives suchas fuel-efficient stoves to reduce fuelwoodconsumption. Dirk Muir (1990), whoworked with local woodcutters inAfromontane forest in southern Africa,showed that cultivating alternative sourcesof building material outside of indigenousforest was over ten times cheaper than thecost of an intensive monitoring programmefor sustainable use of building poles fromindigenous forest.

Insight into impacts: studyingharvesting effects on plant

populations

Ideally, the effects of harvesting need to bestudied on the same sample populationover time through establishment andsubsequent resurvey of permanent plots.In many cases, you will find that there wasno previous interest in ‘your’ focal speciesand consequently no permanent plots forcomparative work. One reason for this isthat ethnobotanists often focus on locallyimportant species such as medicinal plantsor lianas in which there has been littlenational or international interest. Anotherreason is that, despite their importance,permanent plots are still rare in much ofthe tropics. Under these circumstances,two methods can be used to get insightinto the effect of harvesting withouthaving to wait for decades for results frompermanent plots.

Firstly, in cases where more spectacular,long-lived large plants such as palms,cycads, cacti or Pachypodium species areharvested from open habitats (desert, grass-lands, savanna), it is possible that you maybe able to use matched photographs or

Applied Ethnobotany

158


fixed point photographs as a monitoringtool (see Figure 5.11). For most harvestedspecies, particularly in forests, the use ofmatched photographs will not be an option.A second and far more widely applicablemethod is to measure the size classes ofharvested populations. The measurementsof plants from sample populations are thengrouped into different size classes toindicate the population structure of theharvested species and to get insight intoregeneration patterns. A useful way ofdoing this is within plots at different pointsalong a gradient from places where harvest-ing impacts are high to where they are low(or absent). Each of these methods isdescribed in the sections that follow. Later

in the chapter, a third method, populationmatrix modelling, is discussed. This isinappropriate for field workers who do nothave access to a computer, but can be auseful tool to understand harvestingimpacts. Before this is done, however, it isimportant to decide where to sample andwhat sampling method to use.

Vegetation sampling: where andwhat method?

Having decided on which species will be thefocus of a detailed inventory, monitoringprogramme or study on harvesting impacts,you need to decide on where the study areawill be located, whether you will work with


159

Source: from Peters, 1994

Figure 5.6 A flow chart showing an adaptive management approach to sustainable harvest ofnon-timber resources from tropical moist forest

Monitoring

Baselinedata

Species selection

Forest inventory

Yield studies

Periodic regenerationsurveys

Periodic harvestassessments

Adequateregeneration?

Harvest adjustments

Adequateproductivity? Harvest

controls effective?

Yes Yes

NoNo


local harvesters (or not), which sample typeand size would be used in what arrange-ment (random or systematic) and samplingintensity, and what minimum size class ofplants would be measured.

Methods commonly used by botanistsand foresters in quantitative inventory ofvegetation are compared and described inCampbell (1989) and Philip (1974). Alsorecommended is the excellent ecologicalprimer by Charles Peters (1994), which isavailable from Biodiversity SupportProgramme, and his recent review ofecological methods for ethnobotanists(Peters, 1996), which deals specificallywith sampling methods in studying non-timber plant resources from tropicalforests. Permanent sample plot techniquesfor tropical forests are also covered indetail in the practical manuals by Alderand Synnott (1992) and Dallmeier (1992).Permanent plot methods used in Ugandaare reviewed in a very useful paper byDouglas Sheil (1995). Good sources ofstatistical information on sampling designcan be found in Sokal and Rohlf (1981)and Zar (1998). With these referencesreadily available, sampling methods willonly briefly be described here. What isdealt with in more detail are practicalways of identifying priority species andsites, and rating harvesting damage todetermine the effects of harvesting onsampled populations using thesetechniques.

Location of study sites

Many factors influence where peopleharvest: topography, location of paths,roads or tenure, or the density of favouredspecies or favoured size classes within aspecies population. You need to take thisinto account before locating any plots ortransects. Local harvesters’ knowledge isan important step in this process and theuse of participatory mapping methods in

mapping places where key species wereharvested in the past, or are currentlybeing harvested, is an important early stepin deciding where to sample harvestedspecies (see Chapters 2 and 6). Withoutthis, it is also easy to overlook crypticspecies, or to select inappropriate sites forevaluation of harvesting impacts, or towrongly locate ‘control’ plots. Mappingthe vegetation types where the harvestingand harvested species occur will providethe basis for locating transects or sampleplots for studying resource stocks, harvest-ing impacts or yields.

Working with local people

To be effective (and cost effective), it isoften useful to combine quantitativebotanical or forestry methods withmethods that incorporate the insights oflocal people into ecological and socio-economic issues. These inputs caninfluence the choice of sampling methodand the statistical validity of results,particularly if plot location is subjective.

If local resource users are involved inresource inventories or in monitoringimpacts, then it can be worthwhile to usetransects rather than plots (see Figure 5.7).Local resource users with low or nonumeracy or literacy skills can becomefrustrated by plot-based samples whenfocal species are not measured becausethey occur outside of the plot boundaries.In these circumstances, whether sampleplots are systematic or random, some localpeople feel that they ‘waste time’ (due tothe time required to set up the plots) or‘were not accurate’ (as trees were missedout, outside the plots). For these reasons,belt transects would be more productiveunder these circumstances. The limitationthat this places on statistical analysis dueto lack of random plots is often repaid bythe insights of local resource users duringjoint field work. Issues such as size-class

Applied Ethnobotany

160


selection can be linked to the practicalfield assessments of stem and leaf harvest-ing, root removal, bark damage or treecrown condition described in Chapter 4.

Local resource users can also add tothe quality of the data collected by identi-fying cut stumps which have died, whatimplements were used for felling, whyfelling took place, or which animals havebrowsed back resprouting stems. Thequality of the information derived fromjoint resource assessments with localharvesters therefore adds predictive valueon likely impacts of harvesting. This is notonly useful information for resourcemanagement purposes, but can add to andsometimes be cross-checked against infor-mation from ‘informant consensus’ and‘inventory-interview’ methods described inChapter 2. Field work with resource usersin plots or walking through the forest orsavanna is also an opportunity to getinsights into local people’s views of vegeta-tion dynamics. You will find that localpeople in a wide variety of environmentshave an excellent knowledge of the causesof disturbance, routinely identifying treesthat have died due to lightning strikes,elephant damage, fire, fungal infection andso on. In some cases, for example in post-fire or agricultural succession, they arealso able to estimate the time elapsed sincethis has occurred.

Type of sample unit

The most commonly used types of samplesare the following.

Square, rectangular or circular plotsChoice of plot shape depends upon theobjectives of the survey and concernsabout the influence of plot perimeter(‘edge’) effects per unit area sampled (seePeters, 1996). Rectangular plots (and belttransects) have the most edge, and circularplots the least edge. A large amount of

edge usually means that more habitats arecovered, with inventory results morerepresentative of the study area. Oneproblem is that a long plot edge means agreater chance of error in decidingwhether plants on the border of the plotare ‘in’ or ‘out’, which can soon lead tobiased results. For this reason, it is essen-tial that plot boundaries are carefullymeasured. For trees directly on the border-line, an unbiased rule can be decided onbeforehand. For example, trees locateddirectly on the border of two edges of theplot would be considered ‘in’ and those onthe other two edges ‘out’.

Belt transectsBelt transects sample vegetation on eitherside of a transect line (see Figure 5.7a).Transects are also a commonly usedmethod for assessing harvestable plantresources – for example, in narrowtransects at right angles to forest paths. Thesize of plots or length of transect lines willdepend upon the size and abundance of thespecies and individuals you are sampling.Belt or line transects cover a wider spectrumof microhabitats and can show subtlechanges in the density or structure ofsample populations compared to quadratsor circular plots. An advantage of usingtransects when conducting resource assess-ments with local resource users is that insome cases they considered belt transectsmore useful compared to quadrat methods,which they felt ‘missed out’ plants in thespaces between plots. Transects have twomain disadvantages. however. Firstly, sincetransect lines are long and narrow andcover a range of microhabitats, they areinefficient as a method of characterizingdiverse vegetation types or varied vegeta-tion (Campbell, 1989). A second problemis that even if a transect is as long as onekilometre, it still only represents a samplesize of one; therefore, it generally is betterto have more short transects than a few


161


very long ones. In diverse vegetation types,one solution is to break transects up intohomogeneous units, such as a series of100m x 5m transects.

Point-centred quarter (PCQ)The PCQ method enables sampling ofmicrohabitats, determining patterns ofspecies distribution from analyses ofassociated plants within patches (near-neighbour analysis). This method is oftenconsidered very efficient in characterizingvegetation. It also minimizes damage tothe forest understorey. The main disadvan-tages are, firstly, that unlike quadrats orbelt transects, the PCQ method does notcover an exact area each time, so thatspecies richness cannot be related to areasampled. Secondly, since only four treesare sampled per sampling interval, thismakes the method rather labour intensiveif large sample sizes are required(Campbell, 1989).

Arrangement of samples

Sample units can be arranged systemati-cally at regular intervals (see Figures 5.7 aand b), or randomly within the entire studyarea (see Figure 5.7c), or in a stratifiedrandom sample within different habitats ofthe study area (see Figure 5.7d). The choiceof sampling design will strongly influencethe statistical validity of your results andhow long it takes to establish (or relocate)the sampling units. Random selection isdone to avoid bias. In practice, this can beperformed through drawing lots, pullingnumbers out of a hat or, ideally, usingrandom number tables, which are availableas appendices to many statistical textbooks– such as Fischer and Yates’s (1945)Statistical Tables for Biological,Agricultural and Medical Research, whichhas been reprinted many times.

Part of a random number table is givenbelow as an example. Assume, for

example, that you have drawn a grid overthe aerial photograph or map of thevegetation you want to sample and needto select 20 sample units at random fromthe total of 81 numbered sampling unitslocated in the grid. Start by randomlyselecting a number in the table of randomnumbers by blindly touching the page witha pencil.

Once this is done, there is no need tocontinue at random as the numbers arealready in random order. Instead, proceedsystematically along the row or down thecolumn; when that row (or column) isfinished, continue with the next one. Inthis case, since you have a choice of 81sampling units, include all numbers from00–80 (counting 81 as 00) and reject allnumbers between 81–99. Any numberdrawn twice is also rejected. If the firstnumber the pencil touched was 30, and thenext ones were 77, 40, 44, 22, 78, 84 and26, for example, 84 would be rejected andthe others included, continuing until 20sample units are selected. You then needto locate these sample units, measure andsample them.

Random plots are preferable from thepoint of view of statistical analysis, butmay be very time consuming to locate inrough, forested terrain. Systematicsampling may be quicker since plots ortransects are easier to relocate, but becausesampling is not random, it does not allowfor any statistical assessment of precisionor sampling error (Peters, 1996). Insystematic plot sampling, the first plotmay be selected at random, but all the restfollow a set spacing. This would be accept-able if the plants being sampled werelocated at random; but in the naturalworld, plants are rarely spaced indepen-dently of each another. For these reasons,the use of regular (non-random) samplesshould be avoided whenever possible.

If plots are located on sloping ground,then there has to be a correction for slope,

Applied Ethnobotany

162


as the distance between two pointsmeasured along a slope is always greaterthan the corresponding horizontal

distance. The formula used to calculatethis correction (Durr et al, 1988) is:



Figure 5.7 Sampling in a hypothetical 100ha area with three vegetation types (I, II, III) showingcontour lines (dotted lines) and a river (black line). (a) Systematic transect samples. If these were10m wide and 1000m long, separated by 200m, they would give a sample intensity of 5 per cent.

(b) Systematic plot sampling. If circular plots 12.62m in diameter (each with a sample area of500m2) are used, then 100 plots would give a sample intensity of 5 per cent. (c) Simple random

sample, with plots located at random through the whole area (note that only two plots fell withinvegetation type III). (d) Stratified random sample where the number of plots allocated to each

vegetation type is based on the percentage of the total area represented by that type. This methodlocated seven plots in vegetation type III

163

III

II

I

II

II

(a) (b)

(c) (d)

III

II

II

II

I

II

II

II

IIII

II

II

II

IIII


Slope correction = 1/cosine arctan (%slope/100)

The result is then multiplied by thehorizontal distance of the plot to get aslope-corrected ground measure betweencorners of the quadrat. For example, if youwere using 20m x 20m plots and the slopewas 25 per cent, then the correction factorwould be 1 cosine arctan (0.25) = 1.031,and the slope-corrected ground measure-ment would be 1.031 x 20m = 20.62m.The angle of the slope needs to bemeasured using a clinometer. One way ofspeeding things up in the field is first todevelop a slope correction table whichgives correction factors for the plot sizeyou are using for a range of slope angles.An example of a table of slope correctionsfor different slopes and horizontaldistances is given below.

Plot size, sampling intensity andminimum size class

Quantitative ethnobotanical surveys areoften more detailed than those used by

commercial foresters. Commercial forestinventories, which focus on inventory oftimber trees for logging purposes, usuallyhave a minimum cut-off diameter at breastheight (dbh) size of 40cm or 60cm and uselarge plot sizes. In some cases – forexample, where large trees are cut forcanoes, beer troughs or drum-making –resource assessments with local harvestersmay have a similarly high dbh cut-offpoint to forestry studies. Local people notonly use more species, however, but alsomuch smaller tree size classes, such asbuilding poles (5 to 10cm dbh) or withies,and stakes or weaving material which maybe less than one cm in diameter. Dawkins(1958), who worked in Ugandan forests,calculated that if the minimum size of treesmeasured in a 1ha plot in tropical forest is40cm, then a single forestry working partywould take about 5 hours to mark andmeasure the trees in the plot (excludingmarking the access line or plot bound-aries). With each 10cm reduction in sizelimit, working time doubles, so that evenwith two working groups, measurement to

Applied Ethnobotany

164

Table 5.2 A portion of a random number table

70 29 17 12 13 40 33 20 38 26 13 89 51 03 7456 62 18 37 35 96 83 50 87 75 97 12 25 93 4799 49 57 22 77 88 42 95 45 72 16 64 36 16 0016 08 15 04 72 33 27 14 34 09 45 59 34 68 4931 16 93 3 43 50 27 89 19 20 15 15 37 00 4968 34 30 13 70 55 74 30 77 40 44 22 78 84 2674 57 25 65 76 59 29 97 68 60 71 91 38 67 5427 42 37 86 53 48 55 90 65 72 96 57 69 36 1000 39 68 29 61 66 37 32 20 30 77 84 57 03 2929 94 98 94 24 68 49 69 10 82 53 75 91 93 3016 90 82 66 59 83 62 64 11 12 67 19 00 71 7411 27 94 75 06 06 09 19 74 66 02 94 37 34 0235 24 10 16 20 33 32 51 26 38 79 78 45 04 9138 23 16 86 38 42 38 97 01 50 87 75 66 81 4131 96 25 91 47 96 44 33 49 13 34 86 82 53 9166 67 40 67 14 64 05 71 95 86 11 05 65 09 6814 90 84 45 11 75 73 88 05 90 52 27 41 14 8668 05 51 18 00 33 96 02 75 19 07 60 62 93 5520 46 78 73 90 97 51 40 14 02 04 02 33 31 0864 19 58 97 79 15 06 15 93 20 01 90 10 75 06


a 10cm minimum size would take about 3days (Dawkins, 1958).

The size of plots needs to vary accord-ing to the size of the plants being sampled.For statistical analysis, more small plotsare preferable to few large plots: but withsmall plots there is a greater chance oferror in terms of estimating plant density.Due to the level of detail required and theconsequent cumulative time per plot, itcan be useful to use a system of tiered plotsto subsample smaller diameter size classes(Alder and Synnott, 1992) (see Figure 5.8).It is important to spend an equal amountof time sampling each of the size classes ofinterest. For example, if in a series ofnested plots there are 1000 seedlings in100m2 but 10 big trees in 1000m2,decrease the former and increase the latter.Also bear in mind that if your plots are toosmall, you may get many plots without theplants in which you are interested.

Sample intensities of 5 to 10 per cent ofthe study area are common in forestrysurveys. There are statistical formulae forcalculating how many sample units are

needed to achieve a given level of precision,but these require a random sampling design(Philip, 1994; Peters, 1996), which can posepractical problems in ethnobotanical work.

Comparisons along harvestinggradients

Where no base-line data are available frompermanent plots, it is useful to compareharvesting impacts over a gradient fromheavily harvested to unharvested popula-tions at the resource-rich ‘frontier’. Firstly,an unharvested ‘frontier’ site is wherepopulations of plant species in high demandhave been protected from harvesting, suchas places protected from exploitation undercustomary law or in remote areas awayfrom transport routes – where harvesterpopulation densities are low – or inprotected areas. Secondly, sites where high-intensity harvesting takes place are sites ofhigh harvester density, easy access or focalpoints of commercial harvesting. It mayalso be useful to select an intermediate sitebetween these two extremes.


165

Table 5.3 Slope corrections for different distances on slopes of varying steepness

Horizontal distance (m)Slope (%) 5 10 15 20 25

10 5.02 10.05 15.07 20.10 25.1215 5.06 10.11 15.17 20.22 25.2820 5.10 10.20 15.30 20.40 25.5025 5.15 10.31 15.46 20.62 25.7730 5.22 10.44 15.66 20.88 26.1035 5.30 10.59 15.89 21.19 26.4940 5.39 10.77 16.16 21.54 26.9345 5.48 10.97 16.45 21.93 27.4150 5.59 11.18 16.77 22.36 27.9555 5.71 11.41 17.12 22.83 28.3560 5.83 11.66 17.49 23.32 29.1565 5.96 11.93 17.89 23.85 29.8270 6.10 12.21 18.31 24.41 30.5275 6.25 12.50 18.75 25.00 31.2580 6.40 12.81 19.21 25.61 32.02

Note: Slopes in excess of 80 per cent are not shown. The values in the table show the distance along a slope thatmust be measured to obtain the horizontal distance indicated by the column heading.Source: Peters, 1996


Studies in comparable sites along agradient from high to low harvestingimpacts can provide insight into differentlevels of resilience or sensitivity toharvesting.

The simplest case is to compare twospecies size-class distributions at differentsites representing different harvestingintensities or frequencies. This is where thesimple methods described in Chapter 4 formeasuring plant size or rating damage areuseful. You also need to be confident thatvalid comparisons can be made betweensites in terms of structure and density ofsample populations, soil type and aspect.This is far easier in low-diversity vegeta-tion such as wetlands or palm forest, orsavanna-dominated areas on flat sites.Comparisons become more and moredifficult with increasing complexity ofsites in terms of species diversity, distur-bance patterns, soils and topography (seeChapter 6).

Studying harvesting impacts along agradient will enable you to assess theeffects of harvesting on resource supply. Italso enables harvested species to begrouped in terms of their resilience orvulnerability to harvesting. This can

provide very useful insights into opportu-nities or constraints on sustainable harvestand may further prioritize which speciesrequire more intensive monitoring ormanagement. Assessing the response ofdifferent species along gradients from highto low disturbance in South Africansavanna enabled ecologist CharlieShackleton and his colleagues (1994) togroup woody plant species according totheir response to disturbance (harvesting,trampling, grazing) into four groups:invasive species, which increased indensity with increased disturbance; toler-ant species; an intermediate group; andsensitive species, whose density declinedunder high levels of disturbance.

The resilience or vulnerability of differ-ent plant species populations to a highfrequency or intensity of harvesting is adirect reflection of their regenerationcharacteristics and growth rates. Theextent to which resprouter tree species canpersist was shown by ecologist TauberTietema’s study of high and low harvestingimpact sites in mopane (Colophospermummopane) woodland, a tree species highlyfavoured in southern Africa for buildingpoles and fuelwood (Tietema et al, 1991).

Source: Alder and Synnott, 1992

Figure 5.8 Possible arrangement of tiered subplots for measuring below 5cm and 20cm diameterat breast height (dbh) thresholds on a 1ha plot

Applied Ethnobotany

166

Trees over 20cm dbh measured

Poles 5–20cm measured

Saplings 1.5m high to 5cm dbhmeasured

Seedlings measured on strips


The method Tietema used was to count thedensity of coppice stems in plots along atransect, starting close to Morwa village insouth-eastern Botswana. Near the village,there were over 2700 mopane coppicestems per hectare, but a kilometre away,fewer than 200 coppice stems per hectare(see Figure 5.9b). What drastically changeddue to cutting of this vigorous resprouterwas the structure of the mopane woodland,from fewer tall trees to short multi-stemmed resprouts (see Figure 5.9a). Thiscan be clearly shown by assessing popula-tion structure.

Harvest of whole plants: herbs,bulbs, corms and tubers

In her classic book about the !Kung-Sanpeople in the Kalahari desert, anthropolo-gist Lorna Marshall (1976) describes howtwo !Kung-San women, Di!ai and !Ungka,selected edible tubers:

‘They set out while the morningwas still cool and walked brisklyfor about 50 minutes. They eachbegan by picking about 2 poundsof n≠a berries from a clump ofbushes near the area, eating berriesas they picked. They then went tothe part of the area where theyknew they would find /ga [Cocciniarehmannii], n≠wara [Trochomeriadebilis], /dobi [vine, scientific nameunknown], and !ama [Ceropegiatentaculata]. That day Di!ai found34 roots; 34 times she sat down,dug a root, stood up, picked up herson, walked on to look for anotherroot to dig. She dug 17 /ga [C.rehmannii], 9 /dobi, 2 n≠wara[Trochomeria debilis], 4 !ama[Ceropegia multiflora], 1 /ama[Ceropegia pygmaea], 1 !gau[Cyperus sp]. Of these roots, 1 /ga[C. rehmannii] was too immature

to take, 2 /ga, 4 /dobi, and the /ama[C. pygmaea] were thrown awayafter they were dug because theywere too old, bitter and pithy. Onen≠wara [T. debilis] and 2 /dobiwere so deep in the ground Di!aiabandoned them in the holes afterstruggling with each of them for 10or 15 minutes. This left 23 roots forDi!ai to carry back to the encamp-ment, 14 /ga, 3 /dobi, 1 n≠wara, 1!gau, and 4!ama.’

This useful account of a foraging expedi-tion gives a good idea of the effort thatgoes into food gathering, the species andquantities gathered and the selection crite-ria used in collecting edible tubers. It alsoillustrates a dilemma faced by ethnob-otanists interested in measuring the effectsof harvesting on harvested populationswhere whole plants are removed and littleevidence remains after harvest, or whereharvesting has been going on for so longthat the remains of harvested plants (cutstumps, debarked trees) have rotted away.For smaller woody plants and herbaceousor bulbous species, the only evidence thatremains is local knowledge of where thesespecies used to occur and the increasingtime it takes harvesters to collect the samequantity of those species. In these cases, itis useful to set up permanent plots at the‘resource-rich frontier’ and follow whathappens as harvesting increases over time.

By contrast with the tuber removal forwater or food described by LornaMarshall (1976), commercial harvesters ofbulbs, tubers, corms and rhizomes for thetraditional medicine trade generally selectthe largest (oldest) individuals from multi-ple-aged populations to obtain the thickbark and large bulbs, tubers or roots thathave highest economic value. In order toobtain the most bark, bulbs or roots perunit time, gatherers also appear to selectfor dense stands. Once the plants have


167


Applied Ethnobotany

168

Source: (b) Tietema et al, 1991

Figure 5.9 (a) Mopane (Colophospermum mopane) trees outside a village in Botswana showingmultiple stems resprouting from the stumps of trees felled for building poles and fuel wood. (b)

Change in density of mopane coppice (resprout) stems in relation to distance from Morwa villagenear Mochudi in south-eastern Botswana

3000

2000

1000

0

Number of resprout stems s/ha

0

Distance from village (metres)

1000 2000

(b)

a


been removed, there is little sign that theywere there at all. In these cases, it may bedesirable to set up permanent plots in sitesselected on the basis of discussions withresource users or predictions on whereharvesting intensity and frequency arelikely to focus in the future. The size ofplots would depend upon the plant sizeand its population density. Permanentplots in forests would clearly be largerthan those for bulbous species in grass-lands, and small plants at low populationdensity would require more replicate plots.

If local harvesting patterns andpractices are not affected by the presence ofoutsiders, then this can be a good opportu-nity to record information on species andsize-class selection of underground rootsand tubers. You may also have the oppor-tunity of measuring the sizes of bulbs,corms or tubers after harvest, at markets,in herb wholesaler shops or in bags confis-cated by forestry or conservationauthorities. Patrick Nantel and hiscolleagues (1996), for example, measuredbulb size and mass in samples of edible wildleek bulbs which had been confiscated fromillegal collectors in Quebec, Canada (see‘Population Modelling Using TransitionMatrices’ below), using this as one of thesources of information to develop matrixpopulation models to assess the impact ofharvesting. In other cases, it may be possi-ble to return to the site where gathering hastaken place and to assess the proportion ofthe population that has been harvested.

Although harvesting of corms, bulbsor tubers usually means that the wholeplant is removed, the impact of destructiveharvesting at a plant population leveldepends upon bulb, corm or tuber size-class selection. Few data are available forsize-class distributions of wild populationsof species harvested for bulbs or roots;however, bulb or ligno-tuber diameter canbe a useful measure of population size-class distributions in sample populations

in a similar way to foresters’ use of diame-ter at breast height (dbh) for studying treepopulations (see Chapter 4 and Figures5.10 and 5.12).

Large plants produce larger quantitiesof seed than smaller plants. As geophytesgrow older, they also grow deeper in thesoil (see Figure 5.10). This results in betterprotection from fire and drought (Beadle,1940; Weaver and Albertson, 1943),enabling geophytes and ligno-tuberousspecies to thrive and have a competitiveadvantage in savanna and fire-maintainedgrasslands. Soil temperatures are high onthe surface during a fire, but declinerapidly with depth (Beadle, 1940). Larger,deeper-rooted geophytic plants havegreater stored moisture, and are rootedwithin cooler and possibly moister subsur-face soil where they are able to reachwater reserves untapped by shallow-rooted grasses.

The size class of geophytes that areharvested and the proportion of thepopulation removed have an importantinfluence on recruitment of young plantsand the ability of the species population tosurvive fires or drought. In Africa, harvest-ing of geophytes for medicinal purposes orfor moisture or food provides examples oftwo extremes. The toxicity, starch or waterreserves of geophytes that provide protec-tion against herbivores, drought and firemay also attract harvest by people ratherthan protect these species. Local peoplewho gather medicinal plants from the wildfor commercial trade in traditionalmedicines generally select the largestindividuals from within each population,since these are most favoured for sale inurban medicinal plant markets. Smallerindividuals are usually less favoured. As aresult, there often is a progressive declineover time in bulb or corm diameter ofheavily exploited species, such as theLiliaceous medicinal bulbs Eucomisautumnalis, Bowiea volubilis and Scilla


169


Applied Ethnobotany

170

Source: Pate and Dixon, 1982

Figure 5.10 The increase in reproductive output, age and depth of corms of the west Australiangeophyte Philydrella pygmaea (Philydraceae)

(a) 2 year plant 6 year plant 10 year plant

Leaf

Seedlingcorm

Initialplacement

corm

Roots

1 2

0

1

cm0

3

cm

2

1

Spentcorms

Deadroots

Parentcorm

Roots

Replacementcorm

1 2 3 4 5 6

0

3

cm

2

1

1 2 3 4 5 6 7 8 9 10Seasons since initial corm establishment

Leaf

Inflorescence leaf

Previous season’sinflorescence stalks

Parenttuber

Replacementtuber

(b)

40

30

20

10

0

Number of plants

19181716151413121110987654321

Number of plants – 197Mean age – 7.8 years

(c)

6

5

4

3

2

1

0

Mean fruits per reproductive plant

19181716151413121110987654321

(e)

40

30

20

10

0

Depth of corm (mm below surface)

19181716151413121110987654321

(d)

60

50

40

30

20

10

0

F Wt of 1980 corm (mg)

19181716151413121110987654321

Presumed age (years)

Presumed age (years) Presumed age (years)

Presumed age (years)


natalensis in South Africa. Selectiveharvesting of large medicinal bulbs orcorms can seriously affect recruitmentfrom seed and the resilience of the popula-tion against random events such as fire,drought or landslides which can damageplant populations (termed stochasticevents by ecologists) (Chapter 6). Bycontrast, hunter-gatherers who dig tubersfor food or water may avoid the largesttubers in a population because they are toodeeply rooted, pithy or bitter.

At the other extreme to these smallherbaceous or bulbous plants are somespectacularly large plant species. Thesespecies offer the opportunity of usingmatched photographs of the same plantpopulation in the absence of base-line datafrom long-term plots. This is described inthe next section.

Matched and fixed-pointphotographs

Matching historical photographs withrecent ones can be used to study changes inthe abundance of large, long-lived plants asa result of habitat destruction or harvest-ing. In South Africa, for example,‘long-term’ studies of cycad populations areonly available for a few species and onlyspan 20 years or less. For this reason, JohnDonaldson (in press) used matchedphotographs as a method to study changeson South African cycad numbers over time.By searching through specialist collections,publications and through appeals to thegeneral public, he obtained 210photographs taken between 1906 and 1986and relocated 92 of the sites where thehistorical photographs were taken. He thencarefully examined the site to identify newplants or to see what had happened to theplants identified in the older photographs.

Before each site was rephotographed,the position of the camera was recordedwith a global positioning system (GPS) and

marked with a metal stake. Donaldsonthen supplemented records from relocatedsites with information from written reportsand interviews with local farmers andconservation officials. Analysis of thesedata showed that illegal collection ofindividual plants for horticultural purposeswas the main reason for a decline or disap-pearance of cycads (60 per cent of sites),compared to decline as a result of naturaldeath (20 per cent of sites), habitat destruc-tion (12 per cent of sites) and cycad barkremoval for traditional medicinal purposes(5 per cent of sites).

The advantages of using historicalphotographs are that they provide ways ofassessing long-term changes in the numberof selectively harvested species such ascycads, cacti or palms when little or nobaseline data exist. Such photographs cancover a far longer time span in which toassess vegetation change (from at least1880 onwards) than is available fromaerial photographs (generally from 1930onwards). Taking fixed-point photographsas part of a supplementary baseline studyis also a relatively inexpensive methodwhich can be applied for selected speciesas a way of monitoring species-selectiveharvesting of adult plants (see Figure5.11). Disadvantages are, firstly, thathistorical photographs usually focus onadult plants, with limited informationprovided on the recruitment or survival ofjuvenile plants. Secondly, the method isonly applicable to large, long-livedspectacular species in sites which havebeen photographed in the past and is moreappropriate to relatively open vegetationtypes rather than to thickets or forests.Thirdly, because it requires resampling ofthe same site, the method assumes thatcycad populations are static and does notprovide information on new populationsthat may have developed nearby. Despitethese disadvantages, this method canprovide a useful record of changes in the


171


Applied Ethnobotany

172

Source: Donaldson, in press

Figure 5.11 An example of a matched pair of photographs showing decline in a population ofEncephalartos frederici-guiliemi from the time of the original photograph, taken by R A Dyer in1946, to the matched photograph taken by De Wet Bösenberg in 1996. The decline was due to

exploitation for traditional medicine and for horticultural purposes

a

b


number and survival of adult plants whichcould be applicable to long-lived, largeplants such as palms, tree ferns,Pachypodiums, aloes and cacti.

Size-class distributions,survivorship and population

structure

Information on how a plant population isregenerating provides valuable data forresource management purposes and iswidely used in planning for sustainablemanagement of uneven-aged, mixed-speciesforests. By contrast with life tables foranimal populations, which are usuallybased on age, studies of plant populationsare generally based on size-class distribu-tions. These take plants within a samplepopulation and group them into size classesbased on stem diameter (trees, bulbs) orstem length (palms, tree ferns). These size-class distributions are a way of showing

plant population structure and indicatingthe chance of plants in one size class tosurvive into the next size class. They areused as a tool for understanding plantpopulation dynamics, most commonly fortrees.

Three ideal types of size-class distribu-tion are usually recognized for trees inuneven-aged, mixed-species forests (seeFigure 5.12). The assumption is made thatsize-class distributions give an indication ofsize-specific mortality and therefore thestatus of the population. The reverse J-shaped (or negative exponential) curve isconsidered to indicate species tolerant ofshade or competition from other species.Recruitment is continuous and populationsare expanding under these circumstances.The second size-class distribution is for aspecies with equal amounts of regenerationto mature individuals. A bell-shaped curveor a unimodal (flat) curve indicates shade-intolerant or competition-intolerant species


173

Source: Geldenhuys, 1992

Figure 5.12 Generalized models showing how ‘typical’ tree diameter (dbh) size-class distributionsare usually considered to indicate whether a species population is expanding, stable or declining.

Species 1 shows the typical reverse J-shaped or negative exponential curve indicative ofcontinuous recruitment of young stems. Curves shown by species 2 (bell-shaped) and species 3

(flat) both show low numbers of seedlings but different numbers of saplings due to differences inreproductive strategies and site requirements, such as canopy gaps

Stems/ha

Size

Sp 1

Sp 2

Sp 3


or low numbers of seeds due to an unusualreproductive strategy. Species with the thirdsize-class distribution shown in Figure 5.12may be pioneer species which regenerate incanopy gaps or other disturbed sites, repre-sented either by large canopy trees or densestands of seedlings in gaps. The flat curve,showing small numbers of seedlings whichdecline through sapling and adult stages,may be due to levels of light, competitionor irregular reproduction. Be cautious inmaking these assumptions, however (seeChapter 6, Figure 6.8). Remember that inpractice you will encounter curves that arefar from the ideal curves shown in Figure5.12, including multi-modal curves whichmay show survival or reproduction indifferent cohorts.

These different distribution levelssuggest different sensitivities to harvesting.

Shade- or competition-intolerant speciesshowing a flat curve are often much morevulnerable to overexploitation whenyoung seedlings or saplings are harvestedthan species exhibiting a reverse J-shapedcurve. Species with size-class distributionsfollowing a flat curve indicate the scarcityand great importance of young plants forrecruitment. Cutting many young stemsfor building purposes, for example, limitsrecruitment of these species into future sizeclasses, particularly if regeneration fromsprouting is weak or if resprouts are killedby browsing mammals. The reverse J-shaped curve, on the other hand, isindicative of a high mortality of stems insmaller size classes. Harvesting some ofthese individuals may have little impact onpopulation structure, since some woulddie anyway.

Applied Ethnobotany

174

Figure 5.13 The debarking and demise, between 1982 and 1992, of an entire population ofBerchemia discolor trees due to bark harvesting for basketry dye. These occurred scattered ontermitaria over a 10ha area in what had been a favoured dye harvesting area. As this species is

widespread in southern Africa, bark overexploitation posed more of a problem for basket makersin the form of loss of local self-sufficiency than as a species conservation issue

8

6

4

2

0

Number of trees

0–1 10–20 20–30 30–40 40–50 50–601–10 10–20 20–30 30–40 40–50 50–601–10

Total bark removalRing barked>75%51–75%26–50%<10%No damage

Basal diameter (cm)

Bark removal below 2m

1982 1992


Use of size-class distributions is apractical field method for recording assess-ments of harvesting impacts and forillustrating the response of plant popula-tions to harvesting. Figure 5.13, forexample, shows the impact of commercialharvesting of bark for basketry dye on thesame Berchemia discolor tree populationover an 11-year period, and the resultantdie-off of trees in what had formerly beena favoured harvesting area. Changingpopulation structure indicated by size-class distribution may also be used as afield measure of subtle effects, such as theintensive harvesting of large fruits fromreseeder species (see Figure 5.14).

The use of survivorship curves has twomain weaknesses, however. The first is theassumption that the tree stem diameter (or

palm height) always reflects plant age. Thisis not always the case and can be a danger-ous assumption (see Chapter 4). Growth ofunderstorey trees, for example, can besuppressed when the young trees are inshade. Alternatively, tree growth rates arefaster when gaps form in the forest canopy.In addition, a few individual trees within aninitially even-aged stand, such as in adisturbed site or canopy gap, may growfaster than others, eventually dominatingthe gap (Harper, 1977). The second assump-tion is that the shape of the survivorshipcurve reflects harvesting impact alone,rather than a combination of vegetationdisturbance dynamics in addition to harvest-ing. It is crucial that disturbance dynamicsare taken into account, in order to preventmisinterpretation of data on size-class distri-


175


Figure 5.14 A hypothetical example of long-term change in the population structure experiencedby a tropical forest tree species due to intensive harvesting of fruits, showing changes in the

number and population structure of trees in the harvested population at 30-year intervals. Sizeclass intervals are 10cm

1 2 3 4 5 6 7 8 9 10 11 12

Time 0100

80

60

40

20

0

Number of individuals

Size class1 2 3 4 5 6 7 8 9 10 11 12

Time 120

16

12

8

4

0


Size class

1 2 3 4 5 6 7 8 9 10 11 12

Time 220

16

12

8

4

0


Size class1 2 3 4 5 6 7 8 9 10 11 12

Time 320

16

12

8

4

0


Size class


butions of plants. A good example of this isgiven in Chapter 6, where G F van Wyk andhis colleagues noticed how the size-classdistributions of species varied according tothe successional stage of subtropicallowland forest after disturbance (van Wyket al, 1996). Despite these weaknesses, size-class distributions are considered usefulpredictive tools (Geldenhuys, 1992).

Inventory and monitoring by localcommunities

The likelihood of natural resource manage-ment or conservation occurring in anysociety increases when people value aresource or resource area, realize that it isbecoming scarce and have local socialgroups or organizations who can dosomething to halt the decline in that area orresource (see Chapter 7). For this reason,inventory and monitoring are essentialcomponents in any conservationprogramme, including community-basednatural resource management programmes.

While local knowledge assessed byparticipatory rural appraisal (PRA) orinterview methods is useful in illustratinggeneral trends in resource availability, it isno substitute for quantitative data. Wherekey resources are concerned, more rigor-ous methods need to be used. In theBrazilian Amazon, for example, ethno-botanist Patricia Shanley worked withlocal villagers to obtain estimates of thedensity and fruit yields of forest trees wellknown to people whose families had livedin the forest for generations. Estimates offruit yields varied over tenfold, and localestimates of tree abundance were highlyinaccurate. A very knowledgeable localhunter, for example, estimated that 1000mature piquia (Caryocar villosum) treesoccurred in their 1500ha forest, at densi-ties up to 20 trees per hectare. In contrast,their tree census found just 149 matureCaryocar villosum trees in this same forest,

about 0.1 trees per hectare. In short,quantitative studies have to be done andwill take time.

In regions where trained personnel arescarce and detailed information onresource distribution and use is held in theminds of people with low or no literacy ornumeracy skills, resource inventory andmonitoring systems need to be accessibleto local people. The challenge is how toproduce quality data of significance tolocal (community) resource managers, aswell as to professionals and policy makers.Ideally, a monitoring system designed foruse in resource management programmesshould use low-cost, robust, low-mainte-nance equipment which is simple to useand locally (or at most nationally)repairable, and locally available.

If periodic monitoring is to besustained, then the people carrying out thatmonitoring need to be sufficientlymotivated to do the monitoring.Appropriate equipment and training ontheir own are not enough. For this reason,a clear distinction has to be made betweenmonitoring performed by local peopleemployed to work in national parks aspatrol rangers, forest guards or trackers,and monitoring by resource users from thelocal community who are not employed toundertake formal resource monitoring.Where local resource users are involved inresource monitoring, then they must havea personal incentive. Resource users areusually busy with other day-to-day activi-ties, such as tending children, farming, andcollecting fuelwood. For this reason,however ‘user-friendly’ the monitoringsystems are, they need to be relevant andlimited to very few species (or resources)which are important to the people doingthe monitoring – such as food, fuel orwater resources – and, ideally, fun to do.There also has to be a good reason for localpeople to go beyond the way in which theyinformally monitor gathered resources:

Applied Ethnobotany

176


that is, through noting a changing ‘catchper unit effort’ – in fishing terms – sinceresource users spend more (or less) time tocollect the same resource. Examples of keyresources in which local people may havean interest in monitoring in a more formalway are major food resources or the mainindigenous plant species with commercialvalue which are nearing complete extinc-tion (see Figure 5.3).

Biological inventory and monitoring ofa wide range of plant and animal resourcesby local people is becoming morewidespread. In the Brazilian Amazon, forexample, Marli Mattos, Daniel Nepstad andIma Vielra (1992) trained local people onhow to assess the size classes, harvestablevolumes and values of timber on their land,and have produced an illustrated handbookin Portuguese so that these methods arelocally accessible. In forests as far apart asthe Peruvian Amazon, West Kalimantan,Indonesia and the Gulf province of PapuaNew Guinea, ethnobotanist Charles Petershas trained local resource users to monitorthe regeneration of tropical tree species(such as Spondias mombin and Shoreaspecies, which produce respectively uvosand illipe nuts) which are favoured sourcesof large, commercially collected edible fruits(Peters, 1996b). Villager-level monitoringalso extends to marine resources. In theVerata area of Fiji, for example, localvillagers chose to count and measure saltwa-ter cockles (kaikoso), which are a favouredfood, and compare the results in harvestedand non-harvested tabu sites (areas tradi-tionally closed to harvesting) (BiodiversitySupport Programme, 1998). These initia-tives have produced encouraging results.John Parks, the Biodiversity SupportProgramme (BSP) officer associated withthis example, points out how:

‘The government officials saw thatthe village residents were perfectlycapable of doing a fairly sophisti-

cated level of quantitativemonitoring and that such effortscould clearly complement policylevel actions…they were amazedat the ability of the Verata peopleto monitor their resources andexplain the results. Some admittedthey thought such skills could onlybe developed through formaluniversity education.’

Nevertheless, methods used for monitor-ing by people with no literacy ornumeracy, but who have a good knowl-edge of the resources in question, have tobe carefully developed.

Interfacing different worlds: localskills and field computers

Can local people, with low or no literacyor numeracy but detailed knowledge ofnatural history and their local landscape,use computers in inventory and monitor-ing? One school of thought promotes theview that computers should not attemptto replace humans, but rather enhancehuman skills and expertise. This approachdiffers fundamentally from the artificialintelligence (AI) school of thought, whichattempts to develop computers that repli-cate (and therefore potentially replace)human skills. Examples include computersthat can play chess better than humans do,or robots that replace humans in theworkplace. An interesting example wherelocal skills and field computers are linkedtogether is the Cybertracker system(http://www.icon.co.za/~ctracker), devel-oped by computer scientist LindsaySteventon and tracking expert LouisLiebenberg.

A pilot study using the Cybertrackersystem was employed in the KarooNational Park, South Africa, to monitorwildlife (see Figure 5.15a). The objectivewas to enhance the unique skills of expert


177


wildlife trackers with no literacy andnumeracy, and it provides an example ofthis new school of thought in computerscience. Two trackers, Karel Benadie and

James Minye, both employees of theconservation department, have beentesting the field computer on an ongoingbasis. Although they cannot read or write,

Applied Ethnobotany

178

Source: (a) and (c) Liebenberg and Steveton, 1998

Figure 5.15 The palm-pilot field computer (a) used by low/no literacy trackers in the KarooNational Park, South Africa, for recording animals, with (b) a focus on black rhino activity and (c)

feeding records by species indicated by different icons, with observations classed according towhether these were from tracks or direct observation (Liebenberg and Steveton, 1998). A similar

system (d) has been developed for local community monitoring of harvesting impacts on trees, using(e) visual rating systems for bark harvest or (f) tree crown health example, where (g, h, i) icons

based on one to five dots, triangles, squares and diamonds can be used to represent tree diametersize classes (eg 0–4.9cm); these icons are on the pilot field computer and a ruler (or tape measure)


they have used the field computer torecord their observations in the field anddownload the data onto the nationalpark’s personal computer (PC) bythemselves. The data collected in theKaroo National Park are very detailed andcomprise, for example, the home ranges orfeeding behaviour of rhino, showing thechange in plant species favoured asbrowse, and shifting from the rainy seasonthrough to the dry season. In addition, theresearchers have been recording tracks ofrare or nocturnal species that are notnormally monitored, making it possible toobserve species that would not otherwisebe noticed at all. Initial field tests indicatethat a tracker can generate more than 100observations in one day. The highestnumber so far has been 266 observationsin one day, and 473 observations over a 3-day period. The approach which shouldbe taken in communal areas is very differ-ent, with fewer observations recorded eachday on one or two key species.

Due to the potential advantages of thissystem, we recently tested the Cybertrackersystem in community-based monitoring ofwild plants rather than animals, andcompared it with more standard methodsfor measuring and assessing harvestingimpact. The species chosen for thismonitoring experiment was the bird plum(Berchemia discolor) tree, whose barkprovides the major source of dye forcommercial basketry production inBotswana, Namibia and Zimbabwe. Thechallenge is to develop practical methodswhich basket makers can use, and whichdo not put knowledgeable basket makerswith low literacy skills at a disadvantage.One of the easiest ways of doing this is touse pictures illustrating rating systemswhich represent harvesting impacts. In thiscase, icons for tree-diameter size classes,tree crown health, tree crown size andbark-damage rating systems (see Chapter4) were transferred to a hand-held (palm-

pilot) computer in a similar way to theicons used for monitoring animals (seeFigure 5.15). Tree diameters weremeasured using a ruler which uses symbolsto indicate different tree-stem diameter sizeclasses instead of numbered measurements.These symbols (‘icons’) consisted of one tofive dots, triangles, squares and diamondsto represent 0–4.9cm, 5.0–9.9cm, 10.0–14.9cm size classes, etc. These icons arealso on the palm-pilot field computer sothat they are linked to dbh measurementsmade with the icons on the ruler (or tapemeasure). Using the touch-sensitive screen,people can point to the appropriate icon totake them to different screens and thenpress stop (which is like a street stop sign)to complete the record; a global position-ing system (GPS) reading can be made atthe same time.

Advantages of the palm-pilot fieldcomputer that are apparent from theKaroo National Park case include thefollowing. It is quick and easy to use,including use by people with low/no liter-acy; automatic GPS readings can be takenfor every observation; it stimulates interestin monitoring; it enables quick and easydata processing and storage; it provideseasy retrieval of large amounts of data anddata analysis for vast areas over long timeperiods. With these advantages, theCybertracker system provides the oppor-tunity for resource users, national parkpatrol rangers or forest guards to recordtheir observations easily. At the same time,it enables observations to be stored in adatabase for subsequent analysis. Thismeans that the collective observations of alarge number of local people (resourceusers, patrol rangers, wildlife trackers) canbe analysed over large areas and longperiods of time, even after many of thetrackers are no longer working in the area.Long-term trends can then be established.

Disadvantages compared to collectingdata in written (paper-based) format


179


In trying to reach a balance betweendemand and resource supply, we need toknow how much of a resource is producedwithin a specific area. The methodsdescribed in Chapter 4 for measuring plantsize (diameter, length), volume, age, stemor foliage biomass, and bark volume, orfor directly counting annual leaf or fruitproduction, are useful tools in this process.

In even-aged stands of fast-growingspecies with annual above-ground produc-tion, such as thatching grass or reeds, anestimate of annual yield can be a relativelysimple task, particularly when there is justa single use, or where harvest impacts donot conflict with one another. In mostcases, however, yield assessment is morecomplex, requiring measurement of yieldsof products from marked plants in differ-ent size classes, as well as plant density and

size-class data from inventories to extrapo-late annual yields to an area basis (egtonnes/ha/year). In addition to a variationin yield according to plant size class, yieldsoften vary from year to year with sitedifferences. For this reason, yields need tobe measured over several years. Yieldcurves can then be developed to predictestimated annual production of harvestedproducts according to plant size class or ofyields on a standing biomass per area basis(see Figure 5.16). This information is ofgreat practical value in making resourcemanagement decisions. Involving localharvesters in yield assessments, forexample, can lead to a greater awarenessof the limits to resource yields compared todemand. This, in turn, can lead to develop-ment (or reassessment) of local rules whichset limits on who or how many people will

include the relatively high cost and theneed for technical support and regularaccess to electrical power – ‘paper datacollection sheets don’t need batteries’, asZimbabwean basket makers said, in a wrycomment on the use of palm-pilot fieldcomputers. It also requires access to adesk-top or lap-top computer so that datacan be transferred from the palm-pilots forstorage, mapping or processing. For thisreason, the system is most appropriate toconservation programmes or rural devel-opment projects with an equipped office.The main disadvantage of theCybertracker field computer system in theshort term is the cost of the computerequipment. However, the cost of hand-held computers is expected to decreaserapidly in the near future.

The cost of this equipment also needsto be compared with the cost of monitor-ing systems which require much higher

levels of professional staff time, often insituations where there are few professionalstaff, many of whom have other commit-ments. Aerial census of wildlife, forexample, costs about US$200 per hour.Intensive monitoring of plant resourcescan also be expensive. Training localpeople, including those working inprotected areas, to use the palm-pilotsystem enables patrol rangers to record arange of issues in the course of normalpatrols, such as the occurrence of traps,broken fences and pit-sawing activity, orto focus on monitoring rare animal or keyplant species. Far more local staff can beinvolved in the process for the same costas a single qualified professional person. Ifimplemented, the resulting employmentopportunities will provide an additionalbenefit for protected areas or community-based conservation programmes.

Applied Ethnobotany

180

Yields: supply versus demand



181

harvest from a set area and on harvestingmethods (Chapter 7), or a reassessment oflocal decisions on land-use options.

Ecologist Charlie Shackleton, forexample, set up long-term experiments toassess the production of deadwood used forfuel in semi-arid southern African savanna.In the first study, he marked 9500 trees in51 sites to monitor annual mortality ratesof trees. In the second study he monitoredannual fuelwood production in 28 plotswithin three rainfall zones. Tree mortalitywas 3.6 ± 0.6 per cent in the first year and5.2 ± 0.7 per cent in the second year. Asexpected from the reverse J-shaped popula-tion curve (which characterizes manysavanna tree species), more small-sizedstems died than large ones. Taking intoaccount the preference of fuelwoodharvesters for stems over 10cm in circum-ference, he used allometric equationsdeveloped for southern African savanna(Rutherford, 1979) to calculate that thesemortality rates represented fuelwood

production rates of 216 ± 61 kg/ha in thefirst year and 418 ± 92 kg/ha in the secondyear. Deadwood production estimates fromcleared plots provided similar results, withyields of 378 ± 53kg/ha in the most arid site(480mm/year), 270 ± 48 kg/ha in the semi-arid site (670 mm/year), and 438 ± 69 kg/hain the mesic savanna site (870mm/year). Hefound a strong statistical relationshipbetween standing biomass and annual yieldof dead fuelwood, which was summarizedas: deadwood yield (kg/ha = a x standingbiomass (t/ha), where a varied in differentyears between 16.7 (r2 = 0.87; p < 0.00001)and 17.7 (r2 = 0.56; p < 0.0001)(Shackleton, 1998; see Figure 5.16b).

The simplest case is with dense standsof common, single-use species which havea perennial root stock but annual above-ground production (some Poaceae andCyperaceae; see Box 5.1). In the early1980s, for example, I worked with localenumerators who had recorded how manybundles of Phragmites reeds were being

Sources: (a) Peters and Hammond, 1990; (b) Shackleton, 1998

Figure 5.16 (a) Edible fruit yield related to tree diameter size. (b) Deadwood yield for fuel relatedto tree standing biomass in semi-arid savanna, southern Africa

Fruit yield (kg)15

10

5

05 6 7 8 9 10

dbh (cm)

Y = –9.123 + 1.813 Xr = 0.78, df = 11, p = 0.01

Fruit yield (kg)60

45

30

15

00 10 20 30 40

dbh (cm)

Y = –4.783 + 1.444 Xr = 0.60, df = 64, p = 0.01

(a)Deadwood yield (kg/ha)

1000

750

500

250

00 10 20 30 40

Standing biomass (t/ha)

p < 0.0001r2 = 0.87

50 60

(b)


Applied Ethnobotany

182

cut from a long, narrow wetland that wassoon to be fenced within a national parkon the South Africa/Mozambique border.Over 19,000 bundles of reeds were soldper year. We also knew that the height ofthe reeds was a key factor and that if reedstands were too short (less than 3m high),they were unsuitable for harvesting forlocal building purposes. What we alsoneeded to know was what area of reedswas available for harvesting and how longthis was likely to meet local needs.

The first step was to map out the totalarea of Phragmites reeds using 1:10,000aerial photographs (orthophotos). Most ofthe reedbeds (296ha) were within theproposed conservation area, 38ha of whichwere short and unsuitable for cutting and80ha were near elephant watering holes.This left 135ha available for harvesting justinside the boundary of the proposedconservation area, in addition to 43haoutside the boundary. Other publisheddata showed that reed yields were 500bundles/ha, so we could work out totalyield for the wetland and compare thisagainst the estimated local demand for thebuilding material. Assuming that buildingpreferences remained constant, and therewas a block rotation so that each blockwas rested for a year then cut the follow-ing year, the 178ha of reedbed available forharvesting would sustain local demand forreeds (which was increasing at 4.37 percent per year) for a further 36 years(Cunningham, 1985).

Things get more complex when distur-bance affects yields. A good example isbotanist Sheona Shackleton’s (1990) studyof Cymbopogon validus thatch grassproduction in Mkambati Game Reserve,South Africa, which was a major source ofthatch for local Pondo and Xhosa people,most of whom travelled 10 to 40km to cutthatch. Starting off with the same steps ofstudying demand based on local units(standard head-load bundles of thatch

known as isithunga) and mapping the totalarea of Cymbopogon thatch grass, she alsohad to take into account thatCymbopogon thatch grass production wasinfluenced by the biennial burning ofselected blocks of the reserve as part of afire management programme. To do thisshe worked with local harvesters in 20m x20m sample plots in grass stands with apost-burn age of one year, two years andsites such as forest margins which had notbeen burned for even longer. The bundlesof thatch yielded per plot were counted,air dried, and then weighed. Extrapolatedto a hectare basis, Cymbopogon thatchgrass yields were only 97 bundles/ha oneyear after a burn, increasing to 200bundles/ha two years after burning, thendecreasing to 150 bundles/ha thereafter.

Single-species, single-use stands withan annual above-ground production ofharvested product such as the above twoexamples enable yields (in bundles/ha orkg/ha) to be related to area. Yields frommultiple-use species become more complexwhen one type of harvest influencesanother. A classic case is the coppicemanagement system which was practisedin Europe for thousands of years toproduce three products: sticks andfuelwood from younger resprouts (fromunderwood or coppice), timber fromstandard trees growing amongst theunderwood, and pasture for deer orlivestock from grasses, sedges and herba-ceous plants (Peterken, 1993). Coppicewas cut on a block rotation (usually 5 to25 years) and each area was divided intoblocks (or coupes) depending upon therotation time. If a species required a ten-year coppice rotation, for example, thenthere would be ten coupes, with one coupecut each year. Tall upright trees (termedstandards) were grown on a multiple ofthe coppice rotation, with timber treesusually produced in less than 100 years. Ifa standard rotation was 50 years, for


example, then one fifth of the trees in thecoupe would be felled at the same timethat the coppice was cut. If 15 trees werecut for timber, then the same number (ormore) of saplings were conserved to growinto tall trees. Each coupe was usuallyclosed to grazing for four to seven yearsuntil the resprouts had grown enough toavoid being damaged by deer or domesticanimals.

Different uses and interacting impactswere taken into account in developingmanagement recommendations for ilalapalm (Hyphaene coriacea) savanna. In thiscase, livestock grazing, palm-sap tapping,edible palm-heart harvesting, leaf harvest-ing for basketry fibre and the infestationof tapped palm stems by edible palmweevil larvae were all involved, but neededto be quantified. Three potentiallyconflicting palm uses were of limitedsignificance in this case, since harvestingof palm-hearts and cattle browse of palmleaves were negligible and only 0.3 percent (3 stems) of 902 tapped stems wereaffected by beetle larvae. Palm-sap tappingcan influence palm-leaf production rates,however, so we worked with palm-winetappers and marked palms to assessannual leaf production. Sap yields fromseparate batches of a known number oftapped palms were measured daily over a12-month period to determine sap yields(see Chapter 2). Sap yields were very lowcompared to other palms, ranging from anaverage of 4.4 litres/palm (n = 90 palms)to 14.5 litres/palm (n = 18 palms), due tothe small size of available palms.

Although tapping had a negativeeffect on individual palm stems (ramets),only 3.7 per cent (7 stems) of 155 markedHyphaene palms (genets) died aftertapping, 13.8 per cent (26 stems) regrewfrom the same apical meristem, and themajority sprouted new branch stems thatwould be tapped again after 6 to 8 years(Cunningham, 1990a, b). Over hundreds

of years, this has changed the structure ofpalm savanna from one dominated by tallpalms with few stems to one with short,multi-stemmed palms. This change isassociated with decreased sap yields perpalm, but it also stimulated a multi-stemmed growth pattern, thus increasingleaf yields per hectare per year. Leafyields and palm-leaf size class increasedwith palm size. Palms in the 80 to 99cmleaf size class, for example, produced3.15 ± 0.45 leaves/stem/year while thosein the 100 to 119cm leaf size classproduced 3.79 ± 0.59 leaves/stem/year.Using this as a basis for ‘retrospectivecounts’ to estimate leaf harvesting rates(see Chapter 4), harvesting rates werevery low, ranging from no leaf harvestingto a maximum of 8 per cent of annualleaf production in palm populations atfour sample sites. Based on the conserva-tive recommendation that every third leafcould be cut (30 per cent of annual leafproduction) and a palm density of 91.4palms/ha (483.6 stems/ha), there was amassive unharvested surplus of leavessuitable for basketry in the 17,600ha ofpalm savanna in this part of South Africa(Cunningham, 1988). In contrast, palm-sap tapping was nearing its limit, as theaverage palm-wine tapper used 712palms (902 stems) per year, producingover 4800 litres of palm wine per year,but requiring 4200 to 5600 palms tocontinue tapping on a sustainable basis.Based on annual palm sap yields (4800litres/year) and total regional palm-winesales (980,000 litres/year), I estimatedthat 200 full-time tappers were involved.Each commercial palm-wine tapper tapsan average of 4200 to 5600 palms peryear. On this basis, the 1.63 millionpalms that are estimated from palm-density data to be in the area undercommercial palm-wine tapping wouldsupport a maximum of 300 to 400 palm-wine tappers.


183


Models can be useful tools for examiningwhy plant populations change. Matrixprojection models were introduced over50 years ago to study organisms classifiedby age (Leslie, 1945) and later developedto describe the population growth oforganisms grouped by stage (rather thanage) (Lefkovitch, 1965). Over the past 20years, population matrix models classify-ing plants by stage (or size) have gainedacceptance as a useful tool in understand-ing plant population dynamics and theimpact of harvesting. These modelsrequire more sophisticated analysis thanthe size-class distributions describedearlier in ‘Size-Class Distributions,Survivorship and Population Structure’and are a step beyond this manual’s focuson applied field-based methods. However,with the increased availability of lap-topcomputers, they are briefly outlined here.

For those interested in more detail, thefollowing references are suggested inFurther Reading at the end of this book:Caswell (1979); Enright and Watson(1992); Desmet, Shackleton and Robinson(1996) and Bernal (1998).

Population matrix models take advan-tage of the fact that the life cycle of anyplant species can be divided into a few(generally five to seven) stage classes andthe associated probability of moving fromone stage class to the next. In the case offorest species, for example, the stageclasses might be: seeds, seedlings (less than0.5m high), saplings, juveniles and adults.Additional study may show that 10 percent of seeds become seedlings and anaverage adult produces 250 seeds. Theopportunity to age plants using methodsdescribed in Chapter 4 can lead to moreaccurate models. These data can conve-

If neither the data nor the time forlong-term monitoring are available, then itmay still be possible to use other publisheddata to answer the demand versus supplyquestion. This is possible, for example, inlow diversity woodlands such as miombo(Brachystegia/Julbernardia) and mopane(Colophospermum mopane) woodlands insouthern Africa, or sal (Shorea robusta)forest in India. In Zimbabwe, for example,Isla Grundy and her colleagues (1993)wanted to know whether the level of wooduse for building purposes by 155 house-holds in the Mutanda communal area wassustainable. They knew from their studiesof construction-timber use that averageannual consumption of timber by the 155households was 1400 tonnes/year. Most ofthis was from Colophospermum mopanetrees, with average household consump-tion of 3.4 tonnes per household per year,

an annual requirement of 527 tonnes peryear.

From vegetation mapping they knewthat harvesting was from an area of6000ha of woodland, about 1300ha ofwhich was Colophospermum mopane.They were fortunate that an earlier studyhad studied biomass production inZimbabwean miombo woodland (1m3 perha per year) and that data were availableon air-dry mass (750kg per 1m3). On thisbasis, the household construction-timberconsumption of 1400 tonnes per year waswell within the sustainable yield of 4500tonnes per year from the 6000hawoodland area, based on a yield of 0.75tonnes per ha per year. The householdneed for 527 tonnes of mopane wood wasalso within the yield of 850 tonnes peryear from the 1300ha of mopanewoodland.

Applied Ethnobotany

184

Population modelling using transition matrices



185

niently be put into a tabular matrixformat. A further advantage is that bymultiplying the probability matrix by thepresent state of a population, a picture canbe obtained of what future generationsmay look like (assuming probabilities donot change with time). In particular, thematrix can give some idea of whether thepopulation is growing or declining, andwhether the proportions of each class arelikely to change with time.

Like any method, matrix populationmodels have advantages and disadvan-tages (Desmet et al, 1996). Advantages arethat matrix models can easily be manipu-lated to examine the life histories ofharvested plant species over time as a toolin answering ‘what if’ questions about theshort-term responses of plant populationsto harvesting. Resource managerscommonly ask, for example: ‘What wouldhappen if 80 per cent of seed productionis collected? What if uncontrolled harvest-ing removes 60 per cent of individuals in aparticular size class? What will the popula-tion be like in 100 years’ time if we startwith a total population of 400 plants, with10 per cent of bulbs harvested and onlylarge bulbs taken?’

Mathematical background tomatrix modelling

The mathematical background to thetransition matrices given here is takenfrom Desmet et al (1996) and is based ontheir study of harvesting mukwa(Pterocarpus angolensis) in South Africa.

The general form of the populationmatrix (Lefkovitch, 1965) is:

An(t) = n(t + 1)

where n(t) is the population vector (in theform of a matrix of a single column ofnumbers) whose components, ni(t),describe the size-class distribution of the

sample population at the present time (t).A is a square matrix, the populationprojection or transition matrix, with irows and j columns, which describes thenumber of offspring produced to eachstage that survive time period t, theproportion of individuals that remain inthat stage and those that survive and enterthe next stage. Multiplying the presentsize-class distribution by a matrix oftransition probabilities gives the expectedpopulation size-class distribution after onetime period has passed, n(t + 1).

The transition matrix contains valuesfor the stage-specific fecundity (F) on thefirst row of the matrix; values for theprobability of surviving and remaining inthe same stage per time period (R) on thediagonal; and the probability of survivingand growing into the next stage class orany other stage class per time period (G)on the subdiagonal or any other cell in the4 x 4 matrix not occupied by F and Rvalues. In this way, the elements of matrixA characterize the properties of a popula-tion by taking into account its fecundity,mortality and growth rates for each stagein its life history. These variables cannotbe arbitrary. They must be relevant to theorganism concerned. The simplest statevariable is the number of individuals in thepopulation – clearly an impractical taskfor very large populations and an impor-tant one for small threatened ones.

A population can be projected into thefuture for any number of time periods, k:

n(t + k) = Akn(t)

In a constant environment, as the value ofk increases, the predicted size distributionof the population approaches stability,such that the proportion of individuals ineach size class becomes constant. This isknown as the stable-stage distribution(SSD) and can be calculated algebraicallyas the dominant right-hand eigenvector


Applied Ethnobotany

186

(w) of the transition matrix correspondingto λ defined by (Crouse et al, 1987):

Aw = λw

where λ is the dominant eigenvalue of thematrix.

The SSD describes the populationstructure as it would develop if the transi-tion probabilities were constant throughtime. Comparisons of the SSD withobserved initial size-class distribution(ISD), described by n(t – 0), will showwhether the population is in balance, withmeasured rates of growth, survival andfecundity. Differences between the twodistributions may indicate past variationsin, for example, recruitment, growth ormortality. Discrepancies between observedand predicted size-class distributions aremost likely to arise from estimation errorswhere rates for growth, survival andfecundity are measured over short periodsor are poorly known.

Once the population-stage distributionhas stabilized, the rate of populationchange from one time period to the nextmust also be stable for both the individualsize classes and the population as a whole.This rate can be described as:

λ = n(k)/n(k – 1)

where λ is the finite rate of naturalincrease, and n(k) denotes the projecteddistribution vector k time periods into thefuture, such that a stable stage distributionhas been reached. The value of k requiredto reach this may vary markedly anddepends very much upon how close theISD is to the SSD. Solved for algebraically,λ is the dominant eigenvalue (not vector,as for SSD) of the transition matrix.

The finite rate of increase, λ, has avalue of 1.0 when the total populationremains constant through time; is greaterthan 1.0 when the population is increas-

ing; and is less than 1.0 when the popula-tion is declining. It is related to the intrinsicrate of natural increase, r, such that:

r = lnλ

where ln is the natural logarithm.Estimates of λ and SSD provide

insights into the current status of speciespopulations under a given set of environ-mental conditions (Hartshorn, 1975).These population parameters can be usedto compare separate sample populationsfor which transition matrices have beenderived – for example, along an environ-mental gradient.

It must be remembered that the SSD isthe dominant right-hand eigenvector andλ the dominant eigenvalue of the transi-tion matrix. They are, therefore, entirelyindependent of the ISD. For any startingpopulation vector, the values of SSD and λwill always be the same, provided thetransition matrix is unchanged.

There are a number of differentmethods whereby one can calculate thesetwo population parameters. The obviousmethod is to use a computer-based mathe-matical package capable of working withmatrices and performing eigen-analyses.Another approach is to use a computerprogramme specifically developed forecologists to deal with matrix populationmodels. For most biologists, the simplestapproach is to use a spreadsheet. Thisapproach does involve some knowledge ofmatrix multiplication, so that one canenter as formulae the actual matrix multi-plication implied by the general Lefkovitchequation (see previous page). These valuescan be copied down the spreadsheet forhowever many time periods one wants toiterate the model. Naturally, care must betaken when entering and copying theformulae so that the n(f) value referred tois in the previous row, namely, n(t – 1),and the transition matrix value is locked,


thus referring to the same transitionprobability for each time period or itera-tion. Therefore, each row would contain asuccessive population vector n(t = 0, 1, 2,3…x), where x represents the maximumnumber of time periods.

For example, a spreadsheet for a 4 x 4transition matrix would probably have thefollowing format. In column A are valuesfrom 0 to x to keep track of time (t); incolumns B to E are the values for n1(t),n2(t), n3(t), and n4(t), respectively; in row1 are the observed numbers in each stageclass (ISD); in row 2 are the first matrixmultiplication formulae. The transitionmatrix could be anywhere else in thespreadsheet. Equation 4, the finite rate ofincrease, can be entered in column F,where n in the formula represents the sumof columns B to E.

Following the value of λ through time,one will notice that it will eventually stabi-lize. This is time k. At this point, the valuefor λ is similar to that which would bederived if performing a formal eigen-analysis and can therefore be referred toas the dominant eigenvalue of the transi-tion matrix. Similarly, we have reached theSSD, which can be calculated by convert-ing the population vector at t = k to avector of proportions summing to 1.

Desmet et al (1996) sound a note ofcaution here about decimal places. Theyfound that λ is sensitive to seeminglyinsignificant decimal places in the valuesof the transition matrix and that workingto an accuracy of four places will suffice.Researchers will also notice that with thespreadsheet method, λ never really stabi-lizes. As t increases, λ is constantlychanging, but at a decimal accuracybeyond that required for a simple popula-tion model. Desmet et al also suggestanother useful population parameter thatcan be calculated for a transition matrix:the reproductive vector. This is the relativecontribution of a given stage class to

future population growth (Enright andWatson, 1991). As with SSD, the repro-ductive vector is only valid once the stabledistribution has been reached because onlythen is the structure of the populationconstant in time. It is calculatedalgebraically as the dominant left-handeigenvector (v) of the transition matrixdefined by (Crouse et al, 1987):

v1A = λv1

In Desmet et al’s spreadsheet model, thecalculation is performed simply by trans-posing the transition matrix. At time k,instead of being left with a vector repre-senting the SSD, one now has thereproductive vector. The time to stability(k) will not have changed, nor will thevalue of λ.

The one major downfall of matrixpopulation models is that the prediction offuture population size is generally of littlerelevance. If this is your goal, then Desmetet al (1996) suggest you start searching fora different modelling approach now! Inmost matrices, mean rates for fecundity,survival and growth are used. More impor-tantly, deterministic and/or stochasticfunctions or factors are not implicit in thetransition matrix and there are also nopopulation-regulation mechanisms orfeedback loops. Generally, one tends tofind that when the model is iterated, if λ isgreater than 1.0, the population growsexponentially, or, if λ is less than 1.0, thepopulation declines exponentially.

One use of matrix models lies in beingable to examine the behaviour of thetransition matrix and population vectorelements over time (Enright and Watson,1991). The added advantage of thisapproach is that it quantitatively predictsthe behaviour of an ideal population, andprovides a means to determine actualdirection of change in a particular popula-tion at a particular time, and how this


187


change might respond to changes in fecun-dity, growth or mortality (Caswell, 1989).These types of quantitative analysis areonly possible when one is confident thatthe derived transition matrix for a samplepopulation is an accurate reflection ofreality. Such transition matrices requiredetailed long-term population life-historydata, a luxury for most organisms studied.

Are matrix models of any use if wecannot predict future population sizes orstates? The answer is most definitely yes.Generally, we are interested in short-termresponses of organisms to some form ofperturbation, whether human or other-wise. We are interested in the ‘what if’questions. What if there is no recruitmentnext year? What if poaching eliminates 50per cent of the individuals in a certain sizeclass? This is where matrix models comeinto their own by providing a simple,easily manipulated, approach to prisingopen the life history of the organismconcerned.

The intuitive approach would be toperform a sensitivity analysis on thepopulation matrix, in order to test howsensitive the population growth is to varia-tions in fecundity, growth and survival;this is performed by simulating changes inthe parameters and then determiningcalculation λ of the new matrix. Bysimulating the same proportional changefor each stage successively, one cancompare the relative effect on the differentlife-history stages (Crouse et al, 1987).

One disadvantage of simulation experi-ments of this sort is that the results aredependent upon the chosen perturbation ofthe original matrix. Secondly, for largematrices, the process is very tedious.Analytical methods avoid these difficultiesby calculating the sensitivity of λ to changesin life-cycle parameters. Here we are inter-ested in the proportional sensitivity orelasticity of λ – that is, the proportionalchange in λ caused by proportional change

in one of the life-cycle parameters. Theseproportional sensitivities can be calculated,given the SSD (w) and the reproductivevector (v). The proportional sensitivity (forelasticity) of λ to a change in each matrixelement aij is given by:

aij vj wieij = __ ( —— )λ <vw>

where < > denotes the scalar product andv is a row vector with j columns, and w acolumn vector with i rows (Crouse et al,1987).

Elasticity analysis provides an index ofthe relative contribution of each transitionmatrix element to the value of λ (Enrightand Watson, 1991). The elasticities of λwith respect to fecundity, growth andsurvival sum to 1.0. Therefore, the relativecontribution of the matrix elements can becompared both between elements of thesame matrix or other matrices represent-ing different sample populations ororganisms. The larger the value e for anelement of the transition matrix, thegreater that element’s influence on thevalue of λ. Thus, one can isolate thosematrix elements most sensitive to change.When analysing a population matrix, thebest approach would be, firstly, to performan elasticity analysis to isolate thosematrix elements or life-history processesmost sensitive to change, followed bytraditional sensitivity analysis or scenariotesting (‘what if’ questions) centred onthese elements.

Analysis of data in matrices is usuallydone using commercially availablecomputer spreadsheet programmes, suchas Microsoft Excel, Borland QUATTRO-PRO, Mathcad 3.1 (Mathsoft Inc,Cambridge, Massachusetts), orprogrammes developed for stage-basedmodelling such as RAMAS/Stage Software(Exeter Software, Setauket, New York).

Applied Ethnobotany

188


Models

In their study, Paul Desmet and colleaguesused a simple four-stage model to repre-sent all possible transitions an individualplant could make in any of the stageclasses in one time period (see Figure5.17). In terms of the future of the mukwapopulation, they concluded that if thecurrent harvesting rate of 5.6 per cent perannum is maintained (1.1 stems per ha peryear or a total of 373 stems for the entirestudy area), then after 20 years the numberof trees removed will decline to 0.56 stemsper ha per year (187 stems in total), andafter 40 years to 0.29 stems per ha peryear (98 stems in total). On the basis ofthese projected declines, they concludedthat the harvesting intensity they encoun-tered was unsustainable. At a 25 per centreduction in the harvesting rate, to 0.825

stems per ha per year (276 stems in total),the population will be exhausted in 35years, and at a 50 per cent reduction (0.55stems per ha per year – a total of 184stems), the harvested population will beexhausted in 60 years.

Based on this modelling exercise, theysuggest that when trying to isolate elementsof the matrix most sensitive to change, thebest approach is first to identify the parts ofthe plant life history most sensitive tochange and only then to follow this with‘what if’ questions, testing different harvest-ing scenarios (sensitivity analysis) (Desmetet al, 1996). Elasticity or sensitivity analy-ses help to point out which are the criticalphases for maintaining growing popula-tions. They also help to identify the mostsensitive phases of the plant life cycle andwhich phases should, therefore, be avoidedfor utilization. In the case of Pterocarpus


189

Source: Desmet, Shackleton and Robinson, 1996

Figure 5.17 A life-stage graph for mukwa (Pterocarpus angolensis), a southern African savannatree species commonly harvested for woodcarving and timber. The arrows show all possible life-

history transitions. F represents fecundity (the number of offspring), R the probability ofremaining in the same size class, and G the probability of progressing to the next size class within

the specified time period of the model (in this case, one year or one complete growth season)

1 2 3 4

R1 R2 R3 R4

G1 G2 G3

F2

F3

F4


angolensis trees, the most importantrequirement for population survival was thecontinued presence of mature, reproductivetrees in the population.

Simulating different ‘what if’ situationsusing matrix models has the advantage ofstimulating counter-intuitive answers andprompting questions otherwise not antici-pated. A disadvantage is that realisticpopulation projections require long-termpopulation life-history data: somethingthat is rarely available for most harvestedplant species. Another disadvantage is thatrandom disturbance events such as fire,drought or landslides are not accounted forby simple transition matrices, nor are therepopulation-regulation mechanisms orfeedback loops. Matrix modelling of fire-prone populations is complicated, forexample, because transition probabilitieschange as the result of a fire (Bond and vanWilgen, 1989).

Two exceptions to more simple transi-

tion matrices are the models PatrickNantel and his colleagues (1996) devel-oped when studying the viability ofAmerican ginseng (Panax quinquefolium)and wild leek (Allium tricoccum) harvest-ing in Canada, and Elena Alvarez-Buylla’s(1994) matrix models for the tropicalforest pioneer tree Cecropia obtusifolia(Moraceae). Both can be used to assess theeffects of different harvesting regimes andare briefly described here. Both studies arerecommended reading for those interestedin more detail on matrix modelling.

Patrick Nantel et al (1996) estimatedthe extinction thresholds (the populationsize below which no population increase ispossible) and minimum viable populations(MVPs) of wild leek and American ginsengusing transition matrices with differentharvesting rates, different harvest rotationtimes and seasonal variations in growingconditions. Due to lack of relevant data,their population projections did not take

Applied Ethnobotany

190

Source: Nantel et al, 1996

Figure 5.18 (a) Decline of the mean growth rate (λ) of an American ginseng (Panaxquinquefolium) population under different harvesting rotations. Mean population growth rates(λ) were computed from 200-year stochastic projections using four transition matrices. (b) Thelarger the population, the lower the chance of extinction: extinction probability over 100 years

for American ginseng as a function of starting population size

1-yearrotation

3-yearrotation

5-yearrotation

10-yearrotation

Equilibrium

1.06

1.04

1.02

1.00

0.98

0.96

0.94

0.92

0.90

0.88

0.860.0 0.05 0.10 0.15 0.20 0.25 0.30

American ginseng harvesting rate

Population mean growth rate (λ)(a)

American ginseng extinctionthreshold = 91 (560)

1.00

0.90

0.80

0.70

0.60

0.50

0.40

0.30

0.20

0.10

0400 600 800 1000 1200 1400

Starting population size N(0),including seeds

Extinction probability over 100 years(b)


random catastrophes into account. Fortheir model of wild leek harvesting, theymodelled two harvesting strategies basedon analysis of confiscated bulbs: ‘choosy’harvesters, who selected large bulbs, and‘busy’ harvesters who took many morebulbs across a wide size-class range.Simulated harvests used a randomsequence of matrices and were run on atime horizon of 200 years. This provideda basis for suggesting that the sustainableharvest rate of 15.8 per cent, under theassumption of a favourable, unchangingenvironment, should be reduced to a farmore conservative 5 per cent if harvestswere to be sustainable. In addition, usingan extinction threshold of 91 plants (560,including seeds) gave an estimated MVPfor American ginseng of 172 plants (1068,including seeds; see Figure 5.18). Apopulation of this size would have a 4 percent chance of going extinct in 100 years –yet only a dozen American ginsengpopulations in Canada have more than170 plants. This is a good example of howstochastic events need to be taken intoaccount as they can significantly reducesustainable harvest levels.

In her study, Elena Alvarez-Buylla

(1994) developed four separate modelswhich incorporated different combinationsof density-dependence and habitat distur-bance resulting from canopy gapformation, and applied these to Cecropiaobtusifolia, a canopy-gap dependent treespecies which is widespread in Central andSouth America. This modelling approachenabled her to test situations where harvest-ing was, or was not, linked to forestdisturbance. Where harvesting was linkedto gap formation (as one would expectwhen the harvested trees are felled), shefound that Cecropia obtusifolia popula-tions would be expected to reach maximumpopulation levels since Cecropia needscanopy gaps in which to regenerate.

Different responses of differentspecies to disturbance and stochasticevents is a critical factor to take intoaccount in applied ethnobotany. Randomevents such as landslides, wind throws orfires, which can cause catastrophicdamage to populations of some species,may also provide the conditions underwhich other species regenerate. This is animportant factor to take into account inresource management, and is the focus ofthe next chapter.


191


Most landscapes have been influenced bynatural disturbances. Disturbance bypeople – for example, by burning, farmingor the effects of livestock – has also playeda role. As a result, vegetation patternsreflect complex interactions betweenphysical factors such as topography,rainfall and geology with biological andsocio-economic factors. Rate of loss,fragmentation, dynamics and degree ofprotection of vegetation are also takeninto account in conservation programmes.It is equally important to recognize thegreat differences within and betweenforest, savanna and desert systems in termsof patterns of disturbance, plant use andour perceptions of these in the develop-ment of resource management orconservation programmes. It is widelyrecognized now that protected areas haveto be seen in the context of the surround-ing landscape and land use.

This chapter emphasizes how patternsof harvesting relate to vegetation dynam-ics and disturbance, and what methodscan be used to take this into account indeveloping conservation and resourcemanagement plans. To make sure that you‘see the wood for the trees’ in developingsuch programmes, it is essential to view

harvesting in the context of wider spatialscales and longer time scales. This is bestdone by working from the ‘topdownwards’, starting at the landscapelevel, moving to the community-ecosystemlevel and eventually looking at the species-population and genetic levels (Noss, 1990;see Figure 6.1). The ‘mental maps’ whichharvesters have of the past or presentdistribution, or seasonality, of plantresources can provide useful insights intoprocesses of vegetation change.Participatory methods discussed inChapter 2, such as mapping, transectwalks and time lines with knowledgeablelocal people which can be combined withuse of aerial photographs or satelliteimages, are a useful step in this process.

In addition to starting with this ‘bird’seye view’, it is important to understandhow and why plant populations changeover long time scales. Landscape ecology isa well-established field of study, influencedby geography and the biological sciences,which has emphasized the dynamic natureof natural processes (Turner, 1989). Theclose association between climate, soils,vegetation and land forms makes landscapeclassification a useful tool in understandingand setting priorities for land and resource

Chapter 6

Landscapes and Ecosystems: Patterns,Processes and Plant Use

Introduction


Landscapes and Ecosystems: Patterns, Processes and Plant Use

193

Source: Noss, 1990

Figure 6.1 Levels of detail will vary with spatial scale in a hierarchy from a global or continentallevel to that of species levels and genetic levels within a population

Genetic level

a b

Individual plants

Patchy distributionwithin habitat

Species population Community ecosystem

Continental level:Afromontane islands

Regional landscape


use, based on links to human populationdensities. These relationships are evidenteven at a continent-wide scale. EcologistRichard Bell points out, for example, thatareas of Africa which are moist and fertilegenerally have dense populations of peoplewho grow crops and keep livestock, whilepastoralists are found in low rainfall, morefertile parts of Africa. Cultivating societiesare found in moist, less fertile areas (Bell,1982). This basic template of climate andsoils also determines many aspects of wildplant and animal use which should be takeninto account in conservation planning.

A landscape and ecosystem-scaleperspective is important for severalreasons. Firstly, landscape-level and histor-ical perspectives influence the way thathabitats and species are managed orconserved, both inside and outsideprotected areas. Thirty years ago, forexample, people commonly spoke of ‘thebalance of nature’ or of ‘primary’, ‘undis-turbed’ or ‘virgin’ forests (Clark, 1996).Today, ecologists have realized that veryfew plant or animal communities are in anundisturbed ‘balanced’ state. The lifehistories of most species are linked todisturbance events at different time scalesand of different sizes, forming a ‘shiftingmosaic’ of dynamic patches (Clark, 1996;Pickett and White, 1985). Very frequentdisturbance over long periods or at exces-sively large spatial scales can certainlyreduce biodiversity and plant productivity,

sometimes to levels that go beyond recov-ery. The opposite is also true. Reducingdisturbance frequencies (such as throughloss of large mammals – for example, theelephant) or fire exclusion can haveequally serious consequences for somespecies. What is needed in each case is toidentify the appropriate scale andfrequency of disturbance for a particularspecies or habitat.

Secondly, if approaches such as biore-gional planning, ecosystem managementand Integrated Conservation andDevelopment Projects (ICDPs) are tosucceed, then patterns of harvesting andland use have to be taken into account ata landscape level. The problem is that theproportion of land area under all forms ofconservation that is recognized by theIUCN falls short of the 10 per cent idealsuggested by conservation agencies (seeBox 1.1). An even smaller percentage ofthis is in national parks. The Africanregion is a good example (see Table 6.1).

If viable populations of large mammalsthat require large home ranges are to beconserved, then buffer zones and corridorslinking core conservation areas have to betaken into account. Populations of manyrare or restricted-range plant species arealso found outside formally protectedareas. A further problem is that manyprotected areas are far apart from eachother, requiring long, linking corridors. Intheir landscape-level analysis of African

Applied Ethnobotany

194

Table 6.1 The overall percentage of land surface in five African regions under all forms of conser-vation recognized by the IUCN, and the percentage of these conserved areas that are designated

as national parks

Region % of region conserved % in national parks

Southern Africa 8.4 74South-central Africa 9.1 39Western Africa 4.2 63Eastern Africa 6.6 39Northern Africa 3.3 53

Source: Siegfried, Benn and Gelderblom, 1998


protected areas, conservation biologist RoySiegfried and others (1998) found that theaverage distance between protected areasin Eastern Africa was greater than 290km2,and in Northern Africa a massive 2396km2. In many cases, therefore, the systemof corridors and buffer zones linking coreconservation areas covers vast areas ofland which are occupied by rural farmersor pastoralists. Examples are the WildlandsProject in North America, the CentralAmerican Biological Corridor (see Figure6.2) and the Greater Serengeti Ecosystemin East Africa (Miller, 1996; Mann andPlummer, 1993).

In terms of harvesting plants, a declinein the area covered by vegetation types

with characteristic species associations ishighly significant to conservation andresource management programmes.Firstly, this clearly represents a decline inavailable vegetation or key species thatwould have been used by local people.Secondly, it means that the remainingblocks of vegetation become the focus formore frequent and intensive harvesting ofhigh value species.

Thirdly, a landscape-level perspectiveon patterns of harvesting or processes ofvegetation disturbance can influence howquantitative sampling is planned and howits results are interpreted (see Chapter 5).We usually first notice the effects ofharvesting at the individual plant level (see


195

Source: Miller, 1996

Figure 6.2 The proposed Central American Biological Corridor

BELIZE

GUATEMALA

EL SALVADOR

HONDURAS

NICARAGUA

COSTA RICAPANAMA

100km0

Protected areas

N


Chapter 4) and only later consider impactsat the plant-population level, followingthis up with quantitative studies of plantpopulations using small sampling units(quadrats, transects) (see Chapter 5).

Unfortunately, ecologists frequently ‘scaleup’ the results of small-scale studies andapply them – often uncritically – to thewhole study area (MacNally and Quinn,1998).

Applied Ethnobotany

196

Tools for the ‘big picture’: aerial photographs andsatellite images

Aerial photographs or satellite images areboth extremely useful in any inventory ofharvested plant resources, and enablerecognition of vegetation patterns lesseasily perceived from the ground. Loss ofwoodland cover in northern Namibia dueto the effects of small-scale agriculture,high cattle densities and tree felling forhousing and fuel, for example, is difficultto see on the ground, but clearly visible ina Landsat image, and may be comparedwith the less densely populated, bettervegetated area across the border fence insouthern Angola (see Figure 6.3).

On the ground, one gets a moredetailed perspective of a culturallandscape influenced by increasingnumbers of livestock and people andchanges in their distribution. Additionalinfluences are the ‘traditional’ Owamboarchitectural style with its spectacularwood requirements (see Chapter 4), andwoodland clearance for millet fields.‘Ground-truthing’ in the field also givesthe opportunity to recognize where andwhy some tree species are conserved bylocal people in small areas (graveyards) orover large areas (for edible fruits andshade). A major reason for maintainingwoody cover in this area has been the‘traditional’ conservation of favourededible fruit-bearing trees (particularlySclerocarya birrea, Berchemia discolor,Diospyros mespiliformis, Ficus sycamorusand the palm Hyphaene petersiana) aspart of the food production system.

This process of getting a bird’s eye

view leads to a better understanding ofwhich vegetation occurs where (and why)over large areas and fairly long time scales.Aerial photographs are available for someparts of Africa from the 1930s and 1940s(see Box 6.1). Although it is difficult toobtain aerial photographs in somecountries for security reasons, this isgenerally the exception rather than therule. Even in economically poor countries,aerial photographs are usually availablefrom the department of lands and survey(or its equivalent). Satellite imagery – suchas Landsat, Système pour l’Observation dela Terre (SPOT) and Advanced Very HighResolution Radiometer (AVHRR) fromNOAA-AVHRR weather satellites – coversmuch larger areas than aerial photographsbut has been available for less time.Landsat data for southern and south-central Africa from the early 1970s, forexample, can be bought from the USNational Landsat Archive (Eros DataCentre, Sioux Falls, South Dakota, US),and NOAA-AVHRR data have been avail-able for Africa since 1981. Majorproviders of satellite imagery are theNational Aeronautics and SpaceAdministration (NASA) in the US and theCentre National d’Études Spatiales(CNES) in France. One of the problems forfield researchers is the high cost of high-resolution satellite data. In most cases,however, aerial photographs, which areavailable at comparatively low cost, aremore appropriate for most appliedethnobotanical field work.


Satellite imagery can be extremelyuseful for understanding land-use patternsat a landscape scale and as a planning (andpredictive) tool for conservation and ruraldevelopment work. Two good examplesare firstly, use of the daily NOAA-AVHRRdata by the International Geosphere-

Biosphere Programme’s (IGBP) miombonetwork to understand burning patternsin miombo woodland. Secondly, Danishornithologist Jon Fjeldså and hiscolleagues used ten years’ daily NOAA-AVHRR data to recognize tropical Africanmontane forests with ecoclimatic stability


197

Source: NASA, in Marsh and Seely (1992)

Figure 6.3 Landsat image showing differences in vegetation cover between southern Angola(north) and northern Namibia due to higher numbers of cattle and people, with consequent

impact on woodland and perennial grasses in Namibia as a result of the impacts of grazing andclearing woodland for millet fields. The white arrow shows the line of the international border

and the black arrow a fenced site at Ongongo Agricultural College


as an indicator of ‘museum’ sites withconcentrations of evolutionarily ‘old’species, and less stable sites which were‘hot spots’ of ‘new’ species (Fjeldså et al,1997). These data have an importantpredictive value for conservation and landmanagement at the ecosystem andlandscape level.

Anthropologist Robert Sussman andhis colleagues (1994) combined ethno-graphic studies with maps generated fromLandsat images at 1:1,000,000 scale, avegetation map prepared from 1949/1950aerial photographs and a map showing

slope (called a digital terrain model) tomap the deforestation rate, and to under-stand why it had occurred and where itwas likely to take place in the future (seeFigure 6.4). One of their main conclusionswas that if forest conservation is to besuccessful, conservation agencies have towork outside of core conservation areas aswell, focusing effort in cooperating withlocal people at ‘deforestation fronts’,which were identified from this mappingprocess. In making comparisons betweendifferent photographs or satellite imageseries, it is essential to remember that this

Applied Ethnobotany

198

Source: redrawn from Sussman, Green and Sussman, 1994

Figure 6.4 The deforestation history of eastern Madagascar, derived from aerial photographs andsatellite images. (a) Estimated original extent of forest. (b) Forest extent in 1950. (c) Forest extentin 1985. (d) Computer-generated map predicting the possible future extent of rainforest in eastern

Madagascar if all forest on slopes <7° were to be cleared

Before peoplearrived(1500–2000years ago)

1950s 1985 Possible futureextent (all forest

on <7° slopesdestroyed)

(a) (b) (c) (d)


can only be done if the different imagesare brought to the same scale. One way ofdoing this is with a zoom stereoscope (seeBox 6.1).

The scarcity or abundance of particularvegetation types that are classified accord-ing to biogeographic units stands out evenat a continental scale; depending on thefocus of the study, one can work from abroad scale towards increasing levels ofdetail (see Figure 6.1), with differentmethods appropriate at each spatial scale.The seminal paper by conservation biolo-gist Reed Noss (1990) is recommendedreading on hierarchical approaches to

understanding biodiversity.Periodic assessment of the extent and

rate of loss (or expansion) of habitat at alandscape level using aerial photographsor satellite images can be a cost-effectiveway to monitor the success or failure ofconservation programmes, but it does notgive the full picture. Forest, grassland orwoodland cover may not change at all –but underneath the canopy, populations ofhigh-value, vulnerable species can bedisappearing due to overexploitation. Forthis reason, monitoring at a large spatialscale needs to be combined with monitor-ing of a few high-value ‘indicator’ species


199

Figure 6.5 (a) High resource value, high-diversity Afromontane forest within a protected area(Bwindi-Impenetrable Forest, Uganda) that provides habitat for half the world’s mountain gorillapopulation and is considered to be a Pleistocene forest refugium. Highly dynamic forest on steepslopes due to natural disturbance through canopy gaps created by tree falls, yet with good treecover, altered to (b) agricultural fields with a few remnant Agauria salicifolia trees retained forfuel wood. The boundary of the protected area is on the crest of the ridge. As shown in Figure

3.3, this clearing has taken place since 1954. What does this represent in terms of localavailability of fuel wood, building timber, medicinal plants (roots, bark)? What does it represent

for local availability of annual or biennial ‘weedy’ species used as edible greens or medicines(leaves, whole plants) and favoured by disturbance?


Applied Ethnobotany

200

BOX 6.1 AERIAL PHOTOGRAPHS FOR VEGETATION INTERPRETATION

Aerial photographs used for the production of base maps are taken vertically from afixed-wing aircraft, using special vertical air cameras along a series of set flight paths.In the past, most aerial photographs were taken with black-and-white panchromaticfilm (see figure below), with black-and-white infra-red film used for some vegetationstudies. More recently, a range of colour film types are employed.

Miombo (Brachystegia and Julbernadia) woodland in Zambia, showing elongate dambowetlands, the circular patterns created by characteristic chitemene agriculture, where trees arepollarded and burned to fertilize agricultural fields. More recent settlements are adjacent to

the road. Margin of photograph shows date, altitude and series number

In order to locate the photographs you are using within a base map, it is important torefer to the original flight plan. The scale of aerial photographs taken for the produc-tion of base maps can vary considerably. In studying long-term changes in savannatrees, for example, Viljoen (1988) used aerial photographs which varied from 1:20,000


at a population level (see Chapter 5) togive a comprehensive picture.

The broad perspective given by aerialphotographs or satellite imagery neverthe-less gives an important background toinformation collected from local people(discussions, interview surveys, PRA

methods) or botanical surveys (transects,plots) at different spatial scales. InUganda, for example, the majority offormerly forested areas have been clearedfor small-scale agriculture by the highdensity (150–350 people/km2) of localfarmers; the remaining forests are found


201

for the 1940 series, 1:30,000 for the 1974 and 1977 series and 1:60,000 for the 1965photographs of the same area in different years. For comparisons to be made, thesehad to be optically enlarged to the same scale using a zoom stereoscope. This enabledViljoen to compare tree densities determined from aerial photographs over a 5560 km2

area and compare these with data on elephant numbers in the same area (Viljoen,1988). Similar analyses could be done comparing anthropogenic effects in savannawoodlands.

Each photograph is taken so that it overlaps with the next one. Overlap is usually60 per cent to allow for stereoscope use. A stereoscope is a relatively simple opticaltool, with a pair of lenses on a stand which, when placed over a pair of photographs inthe same flight series and the correct distance apart, gives a three dimensional appear-ance. This makes recognition of patterns within vegetation much easier, facilitatingvegetation mapping. In combination with checking in the field, including insightsprovided through participatory ground mapping exercises by local people, characteris-tic species associations can be marked out on the photograph and then traced toproduce a vegetation map. Aerial photographs are also useful in recognizing relictforest patches traditionally conserved for their use as burial sites and settlementpatterns, and trends or plant species associations with geological features such as deepsands, limestone outcrops or termitaria.

Although aerial photographs on a scale of 1:20,000–1:60,000 are very useful formapping vegetation types, and have been used for monitoring populations of largesavanna trees, the scale is generally too large for more detailed vegetation surveys,where a scale of 1:1200–1:4500 is usually preferred. Photographs on this scale alsoenable finer detail of disturbance or unusual vegetation patterns to be recognized,unless tree canopies are close together, obscuring understorey species. Photographs onthis scale are rarely commercially available in developing countries. In special caseswhere there is the opportunity to take photographs at this scale, you need to ensurethat you have made the correct preparations and have the correct film and equipment.This is described by Connah and Jones (1983). You also need to decide what scale yourequire. This is determined by the equation:

lens focal lengthscale = ———————————————

flying height above the ground

You may find the reference by Burnside (1979) helpful. When using aerial photographsin the field, be sure to keep them flat, and protect them from moisture or dust. It isoften convenient to place the photographs in a plastic sleeve and keep them clipped toa board. Be conscious that use of aerial photographs or maps may be viewed with suspi-cion by local militia or police in some countries. You need to take local advice on howto avoid this problem.


mainly within forest reserves or nationalparks. Penny Scott (1993), for example,compared 1954 and 1990 aerialphotographs to show how much foresthad been cleared outside Bwindi-Impenetrable National Park – the highestbiodiversity site in East Africa. The loss offorest and replacement by farmlandevident from the comparison of theseaerial photographs gives a clear indicationof the declining availability of all wildplant resources of old-growth vegetation(see Figure 6.5a) and the expansion offields and weedy species (see Figure 6.5b).

The reverse can also occur. Working inthe forest-savanna transition zone of

Guinea, West Africa, where there is amuch lower density of people (10–50people/km2), social anthropologists JamesFairhead and Melissa Leach (1996)compared aerial photographs from theearly 1950s and recent SPOT images toshow that ‘islands’ of forest, and thicketswithin the savanna, had increased substan-tially, not decreased. Combined withhistorical records and the knowledge oflocal people, they showed how peoplehave created these forest islands in thesavanna and that, counter-intuitively,forest area has in this case increased withhuman population growth.

Applied Ethnobotany

Distribution, degree of threat and disturbance

202

How much is out there? Is a harvestedspecies common or rare, and why?Balancing supply and demand is an impor-tant part of resource management.Methods for assessing biomass or volumeof specific products or yields at theindividual plant and population levelshave been covered in Chapters 4 and 5.Before getting involved in the detail at theplant population level, it is important touse records of the geographical distribu-tion of a species as a ‘first cut’ in decidingwhich species should be the focus of moredetailed work – and which are of lessconcern from a conservation or resource-management perspective. When specieshave restricted distributions, implying thatthere may be a limited supply, and arepotentially threatened by overharvestingwith high demand, it is equally importantto know why those species are rare.

Distribution and degree of threat

In terms of geographical distribution, thecommonness or rarity of a species is

widely publicized as a global or nationalissue expressed in Global Action plans orRed Data books at a global or nationalscale. Distribution on a global scale andthe focus on small populations of endemicspecies (species with limited distribution,often with restricted habitats) are bothundoubtedly important. A variety ofspatial scales must be taken into account,however, so that resource-managementplans retain relevance at a local or regionallevel. A restricted-range endemic of globalimportance may be considered commonand of little concern to villagers livingadjacent to its only locality in the world.Conversely, a useful species which iswidespread may be of great concern tolocal people because of destructive harvestcoupled with a limited supply; yet thespecies may be of little conservation inter-est globally.

Let us assume that you have a list ofharvested species for a particular area,based on ethnobotanical studies ofmarkets (see Chapter 3), with their correctlocal and current scientific names. The


next step is to use information from threemain sources, listed below.

Published sourcesPublished sources include standard workson the flora of the region (eg Flora ofTropical East Africa, Flora Zambesiaca)or country (such as Flore du Cameroun);taxonomic monographs, checklists,revisions (eg Legumes of Africa, Lock,1989); comprehensive field guides on anational scale (eg Kenya Trees, Shrubs andLianas, Beentje, 1994) or on particularfamilies (eg A Field Guide to the Proteasof southern Africa, Rebelo, 1995); RedData lists (eg Hilton-Taylor, 1996) andconservation actions plans (eg for palms:Johnson, 1996, or cacti and other succu-lents: Oldfield, 1997).

Herbarium collections and computerdatabasesIn contrast with species-poor temperatecountries whose flora is extremely wellknown (such as Britain with its circa 1800species and its many published field guidesand several atlases), published records formany high-diversity tropical or subtropi-cal countries may not be available; in thiscase you need to get distribution recordsfrom national, regional herbaria (eg theEast African Herbarium) or herbariawhich have large international collectionsand have specialized on plants from partic-ular continents (such as the Africancollections at Royal Botanic Gardens,Kew, or the Missouri Botanical Gardens).This enables distribution records takenfrom many herbarium sheets to be mapped(see Figures 6.6a and b). This can be verytime consuming. In addition, visits tointernational herbaria are impractical formost field workers. For these reasons, theincreasing availability of electronicdatabases is very useful. Examples of theseare: the threatened plants database devel-oped at the World Conservation

Monitoring Centre (WCMC), Cambridge;the International Legume Database andInformation Service (ILDIS), coordinatedthrough the Department of Biology,University of Southampton; or regionaldatabases such as LEAP (List of EastAfrican Plants) covering Kenya, Ugandanand Tanzania, developed at the EastAfrican Herbarium, the Plants of SouthernAfrica Database and SARARES (SouthernAfrican RARES database).

Information from local peopleSome small, unspectacular species arerarely collected by botanists, yet are wellknown to local people such as traders andharvesters, who can often provide infor-mation to supplement scanty publishedinformation or herbarium records –sometimes with embarrassing accuracy! Inthe mid 1980s, for example, I carried outa conservation assessment of commerciallytraded medicinal plants in KwaZulu-Natalprovince, South Africa, cofunded byconservation agencies, herbalists and herbtraders. I went through the normal processof identifying voucher specimens collectedin markets and with herbalists in the field.A year later, mid-way through the study,with most species identified and withherbarium and published distributionrecords, I disputed with herb traders theirassertion that the parasitic plant Hydnoraafricana (umavumbuka, Hydnoraceae)occurred in the province. Thinking thatherb traders were saying this because ofconservation legislation against transport-ing some medicinal species acrossprovincial boundaries, I confidentlyinsisted that the specimens in the market-place must have come from the adjacentprovinces of eastern Cape or theTransvaal, where Hydnora had beenrecorded. A few months later, I realizedhow wrong I was: specimens werecommonly being collected in several partsof KwaZulu-Natal by herbalists and


203


Applied Ethnobotany

Figure 6.6 Geographical distribution helps prioritize species in high demand that are likely to bein short supply, which could be the focus of more detailed work. These maps, prepared from

herbarium records as part of a study on conservation of medicinal plants (Cunningham, 1988),show two traditional commercially traded medicinal species, both exploited for their woodytubers, both in high demand but with very different distributions. (a) Pentanisia prunelloides(icishamlilo, Rubiaceae), with a wide geographic distribution in South Africa (and occurring

northwards to Tanzania), is found in large populations in grassland. (b) Of greater concern is thecycad Stangeria eriopus (imfingo, Stangeriaceae), a monotypic genus with a narrower distributionand smaller populations, which is endemic to coastal grassland and forest in south-eastern Africa

204

Tropic of Capricorn

28°E

24°S

16°E12°E 24°E 32°E 36°E

28°S

32°S

20°E

Tropic of Capricorn

28°E20°E

24°S

16°E12°E 24°E 32°E 36°E

28°S

32°S


commercial gatherers (but rarely bybotanists).

Information from these sources can beused to grade species according tocommonness or rarity. Although this canbe done by one person, it may be subjec-tive, so it is far better for a group ofexperts to assign species to differentcategories of commonness or rarity. In amedicinal plants survey carried out in the1980s, for example, I assigned commer-cially traded medicinal plants to afive-point scale. In addition to groupingcommercially traded medicinal plantspecies according to life form and whichpart of the plant was used, I also catego-rized each species into three basic groups(rare, uncommon or common) based ontheir geographical distribution. Thesecond two categories (uncommon orcommon) were subdivided further accord-ing to whether they were ‘uncommon butwidespread’ or ‘uncommon and localized’,or common species which were either‘widespread’ or ‘locally common’. Thesecriteria were then used to prioritize speciesfrom the 400 medicinal species in commer-cial trade in that part of South Africa(Cunningham, 1988, 1991). A bettersystem was used by Deborah Rabinowitzand her colleagues (1986) to examinecommonness and rarity in the British floraon the basis of geographic distribution,habitat requirements and local populationsize. Highest conservation status is givento a species with narrow geographicaldistribution, a restricted habitat and smallpopulation size (see Box 3.1, step 4).

The geographical distribution ofspecies changes as populations expand intonew areas, or contract, with habitat lossand/or overexploitation. More than 3000wild-collected Stangeria eriopus cycadshave been recorded sold per month at thetraditional medicine market in Durban,South Africa (Osborne et al, 1994). At that

rate, it is likely that the distribution mapshown in Figure 6.6b will shrink alarm-ingly over a decade. The extent of declinein populations of restricted-range species isused to assign species to the IUCN Red Listcategories to indicate the degree of threatto them (see Box 6.2). The old version ofthe IUCN Red List categories (Davis et al,1986) was considered to be too subjectiveand has been modified (Mace et al, 1993;IUCN, 1994) so that a plant species can berated as Critically Endangered (CR),Endangered (EN) or Vulnerable (VU),based on specific criteria which are identi-fied within the following five maincategories:

Criterion aProportion of the population loss: areduction of at least 80 per cent forCR, 50 per cent for EN and 20 percent for VU, observed, estimated,inferred or suspected over the last tenyears or three generations (whicheveris longer), or projected or suspected tobe met within the next ten years orthree generations.

Criterion bArea occupied by the population:extent of occurrence estimated to beless than 100km2 for CR, 5000 km2

for EN and 20,000 km2 for VU, orarea of occupancy estimated to be lessthan 10km2 for CR, 500 km2 for ENand 2000 km2 for VU. Estimates ofsevere fragmentation, continuingdecline and extreme fluctuation arealso considered.

Criterion cRates of population decline andfragmentation: fewer than 250 matureindividuals for CR, 2500 for EN and10,000 for VU. Estimates of continu-ing decline, severe fragmentation(minimum numbers within subpopula-


205


tions) and all individuals existing in asingle subpopulation are also consid-ered.

Criterion dPopulation size: fewer than 50 matureindividuals for CR, 250 for EN and1000 for VU. A VU taxon could alsobe characterized by an acute restric-tion in area of occupancy (less than100 km2) or number of locations(fewer than 5).

Criterion eProbability of extinction in the wild:at least 50 per cent within ten years orthree generations for CR, 20 per centwithin 20 years or five generations forEN and 10 per cent within 100 yearsfor VU.

In many cases, however, the informationrequired on current population levels andrate of decline is not available for heavilyharvested, restricted-range tropicalspecies.

Nevertheless, this provides a globallyaccepted system for setting conservationpriorities on a global scale. One way inwhich conservation priorities are being setin the face of information gaps are CAMP(Conservation Assessment and ManagementPlan) workshops. The CAMP workshopsgenerally bring together 10 to 40 expertsto evaluate the status of a chosen group ofspecies within a country or region, and aredesigned to reduce bias and favour objec-tive assessment of each species. Goodexamples of this are the CAMP workshopsfor endangered medicinal plants of SouthIndia, held recently in Bangalore, India. Onthe basis of ethnobotanical surveys ofmarkets, market value, habitat availability,degree of trade and records from the RedData book for Indian plants, the CAMPworkshops held in Bangalore set conserva-

tion priorities for 256 medicinal plantspecies (Molur and Walker, 1996). In mostcases, habitat loss is the major reason forpopulation decline, but harvesting is super-imposed on this as an increasinglysignificant factor.

The total loss of habitat, the rate ofhabitat loss, the number and size of blocksof vegetation that remain, and their degreeof fragmentation, degradation and level ofprotection are important factors in settingconservation priorities to habitat types. Aspart of an exercise to set conservationpriorities on a global scale, for example,Eric Dinerstein and his colleagues (1995)used five indicators of landscape integrityto classify ecoregions according to theirconservation status, in the same way thatthe IUCN assigns conservation Red Listcategories to species (see Table 6.2).

An ecosystem approach to protectedareas requires that sizable enough terrainis set aside to maintain viable ecosystemswithout outside interference and withminimal management input (Shafer,1990). The problem is that many ecosys-tems have been transformed and have lostmany of their large predators and herbi-vores. Maintaining smaller systems thanthis ideal requires more intensive manage-ment, including the need to understandwhy species are common or rare, and whatconditions they need in order to regener-ate. Plant species may be rare for severalreasons. Firstly, the habitat of that speciesis only a small proportion of thelandscape. Secondly, the factors needed forsuccessful regeneration (such as fire-stimu-lated seed release, pollination or dispersal,or canopy-gap formation) occur infre-quently and may not have occurred forsome time. Thirdly, the species may haverecently arrived in that landscape fromanother area. This requires an understand-ing of disturbance.

Applied Ethnobotany

206



207

BOX 6.2 THE IUCN RED LIST CATEGORIES

Extinct (EX)

A taxon is extinct when there is no reasonable doubt that the last individual has died.

Extinct in the Wild (EW)

A taxon is extinct in the wild when it is known only to survive in cultivation, in captiv-ity or as a naturalized population (or populations) well outside its past range. A taxonis presumed extinct in the wild when exhaustive surveys in known and/or expectedhabitat, at appropriate times (diurnal, seasonal, annual) throughout its historic range,have failed to record an individual. Surveys should be over a time frame appropriate tothe taxon’s life cycle and life form.

Critically Endangered (CR)

A taxon is critically endangered when it is facing an extremely high risk of extinction inthe wild in the immediate future, as defined by any of Criteria a–e for CR species.

Endangered (EN)

A taxon is endangered when it is not critically endangered but is facing a very high riskof extinction in the wild in the near future, as defined by any of Criteria a–e for ENspecies.

Vulnerable (VU)

A taxon is vulnerable when it is not critically endangered or endangered but is facing ahigh risk of extinction in the wild in the medium-term future, as defined by any ofCriteria a–e for VU species.

Lower Risk (LR)

A taxon is lower risk when it has been evaluated, but does not satisfy the criteria forany of the categories critically endangered, endangered or vulnerable. Taxa included inthe lower risk category can be separated into three subcategories:

1 Conservation Dependent (CD)These are taxa which are the focus of a continuing taxon-specific or habitat-specificconservation programme targeted towards the taxon in question, the cessation ofwhich would result in the taxon qualifying for one of the threatened categories abovewithin a period of five years.

2 Near Threatened (NT)Taxa which do not qualify for conservation dependent, but which are close to qualify-ing for vulnerable.

3 Least Concern (LC)Taxa which do not qualify for conservation dependent or near threatened.


Disturbance

The beginning of this chapter emphasizedthat the life histories of most species arelinked to disturbance events (such as fire,hurricanes, droughts, landslides, disease orhuman-induced disturbance) at differenttime and spatial scales. These result in a‘shifting mosaic’ of recovering patches ofdifferent ages. Some of these can be veryextensive and others small, such as branchfalls from trees or animals digging holes orbuilding mounds (Hobbs and Mooney,1991; Thorsten et al, 1997). Natural distur-bance by fire is frequent in grassland,heathland, savanna and subtropicalwoodland, but rare in desert, tundra orforest. What is important is the appropriatescale, frequency and intensity of distur-bance for a particular species or habitat.

Natural fires in moist savannas tendto occur once every 1 to 2 years, in drysavannas every 3 to 5 years or so, andevery 15 to 20 years in the shrubby heath-lands of Australia and South Africa. Bycontrast, fires are rare in moist tropicalforest, but they do occur naturally during

exceptionally dry periods (Geldenhuys,1994; Hart, 1994). Teresa Hart and hercolleagues’ (1994) remarkable survey ofcharcoal found in soil pits in the Ituriforest of the Democratic Republic ofCongo indicated 4 or 5 fires in the past2000 years. This gives a valuable perspec-tive on how this influences patterns ofplant use. Fruits of the Caesalpinoid tree,Gilbertiodendron dewevrei (mbau), aregathered in large quantity by Mbutihunter-gatherers where this slow-growing,shade-tolerant tree species dominates largeareas of the Ituri forest today. Study ofcharcoal records representing 4000 yearsof vegetation history show that this specieswas either absent or very rare in the past(Hart et al, 1994): a great example of howcontemporary vegetation patterns and useneed to be placed in perspective.

Vegetation dynamics, harvesting andthe population structure of componentspecies are closely linked. Removal of long-lived bulb species in grassland, ordebarking and die-off of forest trees, forexample, can ‘reset’ small patches to earliersuccessional states, particularly when older

Applied Ethnobotany

208

Data Deficient (DD)

A taxon is data deficient when there is inadequate information to make a direct, orindirect, assessment of its risk of extinction based on its distribution and/or populationstatus. A taxon in this category may be well studied, and its biology well known, butappropriate data on abundance and/or distribution are lacking. Data deficient is there-fore not a category of threat or lower risk. Listing of taxa in this category indicates thatmore information is required and acknowledges the possibility that future research willshow that threatened classification is appropriate. It is important to make positive useof whatever data are available. In many cases great care should be exercised in choos-ing between DD and threatened status. If the range of a taxon is suspected to berelatively circumscribed, and if a considerable period of time has elapsed since the lastrecord of the taxon, threatened status may well be justified.

Not Evaluated (NE)

A taxon is not evaluated when it is has not yet been assessed against the criteria.

Sources: Mace et al, 1993; IUCN, 1994


plants are targeted. These events can alsocreate openings for the germination andrecruitment of other species, includinginvasive-introduced species.

Methods such as charcoal analysisused in the above example, or pollenanalysis, are useful tools for understand-ing vegetation change over long time scales(see Chapter 1). Methods described inChapter 4 for ageing plants or measuringplant size or biomass ratios can explaindisturbance over shorter time scales. Firefrequencies, for example, have beenstudied in coniferous forest by cross-dating fire scars, and age of trees has beendetermined on the basis of tree rings(Wagener, 1961), or in the case of eucalyp-tus woodland using stem features of grasstrees (Lamont and Downes, 1979), and incoastal heathlands using bulb leaf-basecounts (Ruiters et al, 1993; see Figure 6.7).Ageing methods for palm stems bent bytree falls have also been used as a tool inunderstanding gap formation and distur-bance in tropical forest (Martinez-Ramoset al, 1988).

It is essential that changes in plantpopulation structure due to disturbance

are also taken into account when assessingharvesting impacts on a species over time(see Chapter 5). Working in subtropicallowland forest in southern Africa, forexample, van Wyk and colleagues (1996)noticed how species population structure,indicated by the number of stems withindifferent size classes, varied according toforest successional stage after disturbance(see Figure 6.8).

A useful way of characterizing distur-bance in forests is by thinking aboutdifferent forest types in terms of the spatialscale of disturbance or ‘grain’. On the basisof work in southern African forests, ecolo-gist Jeremy Midgley has characterizedAfrican forests as ‘fine-grained’ or ‘coarse-grained’ in terms of the regenerationrequirements of gap (shade-intolerant) andnon-gap (shade-tolerant) species. ‘Grain’refers to the spatial context of disturbance.Some simple ways he used to classify theidea of ‘grain’ were to:

• Investigate whether large and smallstems of con-specifics (individuals ofthe same species) occur in small plots(such as 400m2-sized plots).


209

Table 6.2 The matrix for integrating biological distinctiveness and conservation status ofecoregions to assign priorities for biodiversity conservation

Final conservation statusBiological Critical Endangered Vulnerable Relatively Relatively distinctiveness stable intact

Globally outstanding I I I I II

Regionally outstanding I I I II III

Bioregionally outstanding II II III III IV

Locally important III III IV IV IV

Note: The Roman numerals indicate biodiversity conservation priority classes:Level I = Highest Priority at Regional Scale (shaded area)Level II = High Priority at Regional ScaleLevel III = Moderate Priority at Regional ScaleLevel IV = Important at National ScaleSource: Dinerstein et al, 1995


• Investigate size-class distributions ofdominant species. Those with inverseJ-shaped distributions are likely to befrom fine-grained systems.

• Use only canopy or potential canopyspecies to compare the overstoreycomposition with understorey trees inplots. In coarse-grained forests there islow similarity between canopy andunderstorey, suggesting radicalcompositional change with time.

On this basis, Jeremy Midgley et al (1995)characterized the relative shade-toleranceof tree species in Afromontane forests nearKnysna, South Africa. Most of theseexhibited reverse J-shaped curves, leadingthem to describe this forest type as havinga ‘fine grain’ created by small canopy gapsformed by slowly dying trees. Most (70per cent) of trees in the Knysna forest diedstanding, for example, and average gap

size was small (35m2) (Midgley, Cameronand Bond, 1995). In a fine-grained system,dominance by shade-tolerant speciesmeans that con-specific individuals ofvarious sizes will be found close together.By contrast, Dave Everard (1992),working in subtropical coastal forest inSouth Africa, found that most species wereshade-intolerant, with few seedlings orsaplings in the understorey, suggesting aspecies composition characteristic of large-scale disturbance events – perhaps throughlarge, infrequent cyclones, fire or distur-bance by large mammals (eg elephants).However, since succession may take placeover a scale of centuries, some short cutsare needed to gain insight into disturbanceregimes and grain. One way of gettinginsight into tree life histories is to collectdata on specific leaf area (see Box 5.2).

Thinking of forests in terms of ‘grain’,shade-tolerant or shade-intolerant species,

Applied Ethnobotany

Source: (a) Ruiters et al, 1993; (b) Burgman and Lamont, 1992

Figure 6.7 Fire frequency and survival. (a) Age distribution and bulb biomass of the medicinalplant Haemanthus pubescens in a single sample population (n = 783 bulbs). Flower number

increases with bulb age and, like many ‘fynbos’ heathland species, flowering phenology appearsto be synchronized to a fire frequency of 15 to 20 years. (b) The probability of extinction (dashedline) and mean and 95 per cent confidence limits for population size (solid line) as a function of

fire frequency (number of fires in 50 years)

210

50

40

30

20

10

0

1000

100

10

1

0.1

Frequency Bulb mass (g fresh mass; log)

1 3 5 7 9 11 13 15 17 19 21 23

Fire

Fire

(a)

Bulb age (years)

1.0

0.8

0.6

0.4

0.2

0

160

120

80

40

0

Probability ofextinction Population size

0 1 2 43 5 6

(b)

Fire frequency (number on 50 years)

Probability of extinctionPopulation size

Bulb fresh mass (g)FloweringNon-flowering


or regeneration from seeding or resprout-ing, gives useful insights for protectedarea management and past history ofnatural or anthropogenic disturbanceevents. Fine-grained forests are in someways the easiest for low-level utilizationof stems because there is generally anabundance of recruits to replace harvestedindividuals. Coarse-grained systems aremore complex in terms of stem harvest-ing. Early successional species may existas single-aged cohorts and are likely to bereplaced by more shade-tolerant speciesin the absence of a large-scale disturbance,such as a blow down or clear-felling.Coarse-grained systems can also influencethe choice of natural forest loggingsystems and how these influence non-

timber forest products. In one of the fewcomparisons of the effects of differentlogging systems on production and repro-duction of non-timber forest products,ethnobotanist Jan Salick (Salick, 1992;Salick et al, 1995) found that theHartshorn Strip Clearcut Systemincreased vine diversity, but significantlyreduced palm density and diversitycompared to a selective logging system(Hutchinson Liberation Silviculture). TheHartshorn Strip Clearcut System, forexample, which was used by the YaneshaForestry Cooperative in Peru, is geared to‘coarse-grain’ forest (Hartshorn, 1989).Like the woodland coppice systems usedin Europe for centuries, strip-rotationsystems depend upon resprouting species.


211

Source: van Wyk et al, 1996

Figure 6.8 A conceptual model of the dynamics of a subtropical lowland forest in southern Africa(Dukuduku forest). Group 1 species refer to the developing forest species, group 2 species to

peripheral forest species and group 3 species are typical core forest species

Developing Peripheral Core Future core

Height

Nu

mb

er o

f st

ems

Group 1 species

Hymenocardia ulmoidesAlbizia adianthifolia

Canthium inermeApodytes dimidiata

Group 2 species

Strychnos gerrardiiChaetachme aristata

Celtis africanaBalanites maughamii

Group 3 species

Large scaledisturbance

(fire)

Diospyros natalensisMimusops obovataMaytenus undata

Strychnos decussata


Local knowledge, landscapes and mapping

By contrast, the Senility Criteria YieldRegulation System, used in the Knysna(Afromontane) forest, South Africa, tosustainably harvest high-value hardwoods(eg Ocotea bullata), has been designed for‘fine-grain’ forest (Seydack et al, 1995).In order to minimize change from naturaldisturbance patterns and to lessenmanagement inputs, the Senility CriteriaYield Regulation System is based onpreempting mortality, judged by factorssuch as stem rot and tree crown health(see Chapter 4). Based on a ten-yearfelling cycle, this system is consideredsustainable, firstly because amountsremoved are within the range of produc-tivity (basal increment) of the stand, andsecondly, because timber harvesting iswithin the natural disturbance regime andreflects life histories of the componentspecies (Seydack, 1995; Seydack et al,1995). Trees naturally die standing andsingly, as blow downs are rare in this area.Harvested trees are topped (that is, thecrown is removed before the tree ischopped down), and abundant regenera-tion of most canopy species ensuresreplacement probabilities remainconstant. Opening large gaps in ‘fine-grain’ forests can profoundly changeregeneration patterns. One of the mostserious legacies of the uncontrolled small-scale logging which took place inBwindi-Impenetrable National Park,Uganda, during the Amin era is that thiscreated large gaps in which regenerationis effectively ‘frozen’ by mono-dominantAcanthaceae or bracken (Pteridium aquil-

inum), preventing seedling regenerationthrough intense shading.

What is less predictable is howrelatively low-level habitat disturbancethrough species-selective exploitation canaffect things at the landscape level. Subtleeffects of bark damage to savanna trees byporcupines (Yeaton, 1988), or anthro-pogenic disturbance, can certainly cause amarked increase in tree mortality incombination with fire and periodic winds.Removal of bark patches results in adeadwood scar which is periodicallyburned and enlarged by successive fires.This causes lop-sided growth and hollow-ing out of the trees over 20 years or so,resulting in a chimney effect which speedsup the hollowing-out process, eventuallyleading to snapping of the weakened stemsin periodic windstorms. Debarking anddie-off of forest trees for the traditionalmedicines trade is quicker and moreobvious, greatly increasing the proportionof forest in the canopy-gap phase(Cunningham, 1991). The question iswhether it also has subtle effects. In thesouth-eastern US, small-scale die-off ofindividual trees due to lightning strikesmay be spread to a landscape level bybeetles. When conditions are favourablefor the beetles (and stressful for trees),beetle damage to trees can spread fromindividual dead trees to become anepidemic where beetle-induced die-offcreates large gaps in the forest (Rykiel etal, 1988). Does this occur in Africanforests? We do not know: but asking theright questions may help us find out.

Applied Ethnobotany

212

Linking different mapping methods, suchas aerial photograph analysis and partici-patory mapping, with local knowledge ofresource preferences can help to predict or

explain harvesting patterns within landtypes. Useful questions to bear in mindrelate to resource abundance as well as tophysical (paths, roads) and social (bound-


aries influencing tenure, territoriality)access to resources. Such questions includethe following.

• How do local people classify landtypes? Are there seasonal differencesin the use of particular land types? Docurrent land uses differ from those inthe past, and why?

• How do these landscapes or vegetationtypes ‘work’ in terms of natural distur-bance, the effects of livestock orpeople’s use of fire or species-selectiveharvesting?

• How does disturbance influencespecies composition and density offavoured resources or other species?

• How is resource abundance influencedby land form or vegetation type?

• How do topographic features affectresource quality? Position in thelandscape can affect resource qualitiessuch as height or length (of reeds,thatch, weaving fibre or bamboo) oreven the flavour and fruit size in thecase of edible species (such as bitterfruits versus sweet fruits from individ-uals of the same tree species in swampforest versus savanna).

• How are harvesting or human-induceddisturbance affected by seasonality?

• Where is access influenced by land formor vegetation type, and where is itaffected by social boundaries related totenure or territoriality (see Chapter 7)?

These factors might initially be determinedthrough ‘walks in the woods’ withresource users, informal discussions, inter-views, participatory mapping, or whatevercombination of interactive surveys ischosen as most appropriate. They can thenbe followed up with more detailed andlonger-term work with resource users toconfirm, refute or add more detail to infor-mation derived from mapping exercises.Fiona Walsh (1993), for example, working

with Martu people in the Western Desert,Australia, describes how men and womenchose to forage in different land formunits. Records of the time that Martuwomen spent foraging in different landforms reflect selection of species andresource richness within patches, with lowspecies richness in the uplands and highspecies richness in burned sandplains or inwetland sites (see Figure 6.9). In contrast,men usually hunted opportunistically,frequently using vehicles to do so andconsequently travelling longer distances;but they were more constrained to bushtracks.

At a finer level of detail, informaldiscussions and smaller-scale participatorymapping can be linked to quantitativesurveys of vegetation at different stagesafter disturbance within these land forms,whether due to fire, rainfall or agriculturalpractice. In another study in the GreatSandy Desert, Western Australia, FionaWalsh (1990) surveyed burned andunburned vegetation in different land formtypes in conjunction with field observa-tions during foraging excursions withMartu women (see Figure 6.10a, b). Thetime that women spent foraging in differ-ent vegetation types was taken as anindication of which vegetation types werepreferred, with burned sites generallypreferred due to their high species diver-sity and resource richness. This approachcan be applied to the harvesting of a widevariety of resources, whether food ormedicinal plants, building timber or craft-work material.

Bear in mind, however, that selectionof harvesting sites also occurs on smallerand more subtle scales of space and time.What we may have defined as a singlevegetation type due to the obviousdominance of one or a few species may, infact, contain a variety of subtle smallerpatches within it. Selection may be based,for example, on stem or culm height


213


within almost pure stands of single thatch,bamboo, sedge or reed species whenharvesting for thatch, weaving fibre or forbuilding purposes. Bamboo harvesting anddynamics are a good example. In 1955, aspart of an early attempt to take localpeople’s harvesting into account in tropi-cal forest management, the Uganda ForestDepartment mapped the area underbamboo (Synarundinaria alpina) in whatwas then the Mgahinga Central ForestReserve. In order to prevent ‘overcutting’of bamboo, a series of four coupes(rotational management blocks) of equalsize was demarcated. These were succes-sively opened for cutting for one year, thenclosed for three years (Kingston, 1967).Eighty basket makers were registered by

the forest department. This rotationalharvesting started in 1956 (coupe 1),rotating through to coupe 4 by 1960, andrestarting in coupe 1 over again.

Apart from concerns that bambooharvesting was destroying gorilla habitatand the ultimate closure of this area tobamboo harvesting, there was a fundamen-tal flaw in the standard forestry approachto mapping and demarcating coupes: a lackof input from local harvesters. If this hadbeen done, the distribution, size andseasonal use of the coupes would have beenvery different for three reasons. Firstly,bamboo harvesting for basketry and build-ing poles is not evenly spread, but isconcentrated at sites that are well knownfor their tall, flexible and long-internode

Applied Ethnobotany

Source: Walsh, 1993

Figure 6.9 Land form units traversed by women on foraging excursions from two different settlements (Punmu – 18 excursions; Parnngurr – 31 excursions) in the Great Sandy Desert,

Western Australia

214

Wetlands

River channel

Landform unit

0 10 20 30 40 50 60 70

Punmu

Parnngurr

Percentage of total number of excursions

Sand substrates

Calcrete

Uplands

Outwash

Old playa bed

Sandplain

Short dunes

Long narrow dunes

Long wide dunes

Calcrete with claypans

Colluvial slopes

Hilly terrain

Flat-topped hill

Major valley


bamboo – usually moister sites with deepersoil. This is not an uncommon situation. InIndonesia, for example, Sundanese villagersprefer harvesting Gigantochloapseudoarundinacea bamboo culms (stems)from hill slopes rather than from valleys,and for good reason. Comparison ofbamboo culms showed that the tensilestrength and specific gravity of bamboofrom hill slopes were significantly higherthan those in valleys (Soeprayitno et al,1988). Secondly, the uprights (warp) of thebasket weave (locally termed inkingi) canbe collected throughout the year, but themore flexible weft material (imitamu) hasto be collected seasonally before thebamboo stems are fully mature, but notwhen they are too young. Thirdly, bamboostands themselves change over time inresponse to synchronous flowering and die-off and disturbance (fire) (see Figure 6.11).

Similarly, in savanna woodlandsdominated either by mopane (Colopho-spermum mopane) or mogonono(Terminalia sericea) in southern Africa,

woodcutters wanting tall building polesmay choose tall, even-aged stands thatreflect soil type or post-fire regeneration.Within these patches, individual trees areselected for straightness, diameter size classand height. The use of ecotones betweenthe vegetation types is another importantfactor. Walsh (1990), for example, recordsthat ecotones between wetland andsandplain, or between the range andsandplain, were frequently chosen forresource harvesting when Martu womenwere moving between patches or tempo-rary settlement sites.

Incorporating local knowledge intomapping

Use of geographic information systems(GIS) for computer-based mapping andprogrammes such as ARC/INFO havebecome standard practice in somecountries but can be far less appropriatein others due to high costs. One way ofdescribing GIS systems is as ‘electronic


215

Source: Walsh, 1990

Figure 6.10 (a) Species richness. (b) Diversity indices of food plants on major land form units inthe Great Sandy Desert, Western Australia

Range Sandplain WetlandLandform

20

18

16

14

12

10

8

6

4

2

0

Species richness

Burnt

Unburnt

Range Sandplain WetlandLandform

3.0

2.5

2.0

1.5

1.0

0.5

0

Shannon-Weiner diversity index

Burnt

Unburnt

n=

5

n=

4

n=

5

n=

4

n=

4

n=

4

n=

5

n=

4

n=

5

n=

4

n=

4

n=

4

(a) (b)


Applied Ethnobotany

216

Source: (a) Agnew, 1985

Figure 6.11 Patchy distribution of resources across landscapes and time scales must be taken intoaccount in inventory, monitoring and management. (a) Diagrammatic representation of stages in

Synarundinaria alpina succession in East African montane forest on the Aberdare Mountains,Kenya. The six stages do not exist for similar time periods. The course of change of acidity andpercentage cover of three floristic elements are shown as concentric plots. (b) Local Abayanda-

Twa plant expert Yakobo Bandusya at Mgahinga, Uganda, next to a patch of tall, long-internodeSynarundinaria alpina bamboo used for basket weaving

Gametrail

Pioneer bambooFlowered bamboo

Mature bambooBuilding bamboo

Old sambucus Young sambucus

Acidity(7-pH)

ClimbersRuderals

Woodland herbs

3 16

4250

18 20

0.62 1.93

1.94

1.841.76

1.3428

6

079

31 31391343

5

(a)

26

Arundinariaalpina

Sambucusafricana

Climbersvariousgenera

Ruderals

Cyperusdereilama

b


tracing paper’, which provides a quickway of doing overlays of importantvariables such as vegetation change overtime or of settlement patterns, roads andmarkets. Although more time consuming,these overlays can be done without high-tech equipment – for example, from aerialphotographs as described in Box 6.1.What is less recognized amongst urban-oriented planners are the participatoryapproaches to mapping that have beensuccessfully developed with many ruralcommunities in Asia (Poffenberger et al,1992), or the role that indigenous knowl-edge has played in soil mapping in Africa(Trapnell, 1953).

A key to the participatory mappingprocess is the production of maps whichillustrate the spatial distribution ofresources, of resource-flows or oflandscape features of significance (see Box6.3 and Chapter 2). As Poffenberger et al(1992) point out with specific reference toforest resources, in an approach which isjust as relevant to mapping geologicalfeatures, soil or vegetation types:

‘Through interactive exerciseswith the community and observa-tion, the research team can helpcreate a picture of spatial resourceuse patterns by developing sketchmaps, product flow maps andtransects of resource use patterns.The main purpose of diagnosticsketch-mapping is to create avisual representation of theresource system which can easilybe understood by both villagersand foresters.’

Resource mapping of landscapes familiarto local people can follow a series of stepsfor key resources and their dynamics inspace and time. The extent to which thissequence is followed depends upon howmuch time and funding are available and

what level of certainty and detail isrequired. If well done, participatorymapping is a good way of establishingcommon ground and for joint planning,which helps to identify regional landscapeareas with outstanding biological andcultural significance – not from theperspective of formally trained planners,but from the viewpoint of the localcommunities themselves.

Participatory mapping needs as muchcare in recording local terminology forland types, habitats and localities as anethnobotanist would take in recordingvernacular names of plants. Local termi-nology for land types and localities isusually at a very fine scale. Despite thisproblem of scale, it can be extremelyuseful to ‘translate’ the ‘mental maps’ andpatterns of resource use developed fromdiscussions and ground maps ontotopographic maps and, where possible,into more widely recognized terms forvegetation or land types.

A recent Australian example of partic-ipatory planning is the interaction betweenscientists and people of the Western Desert(Anangu) community in conservationplanning in Uluru (Ayers Rock–MountOlga) National Park, which has beenjointly managed by the Anangu people andthe Australian National Parks and WildlifeService (ANPWS) since the granting offreehold title to the Aboriginal traditionalowners in 1985 (Baker and MutitjuluCommunity, 1992). As part of the process,Anangu classification of major landscapetypes was used in addition to internation-ally recognized (but in some cases, lessprecise) geographic terms for landscapeunits.

At a finer level of resolution, fieldwork within habitats facilitates a betterunderstanding of ecosystem dynamicsamongst conservation biologists. This isparticularly relevant in forest ecosystems,which are extremely complex in terms of


217


Applied Ethnobotany

218

BOX 6.3 PARTICIPATORY MAPPING EXERCISES: RESOURCES AT

LANDSCAPE AND SPECIES POPULATION LEVELS

Local participation in mapping has been developed and popularized, and is increas-ingly used to enable local people to share their experience as part of participatory ruralappraisal (PRA) methods (Chambers, 1992). Participatory mapping exercises can beapplied widely to natural resources management; but they also have weaknesses whichshould be avoided, where possible, through complementary use of aerial photographsor ‘photo-maps’ (local people mapping onto a transparent sheet placed over large-scale aerial photographs). A common problem with PRA mapping is the difficulty ofestablishing scale (and therefore transferring data onto other maps). In Nepal, forexample, Richard Mather and his colleagues (1998) worked with villagers using large-scale (1:1250–1:2500) enlargements from 1:50,000 aerial photographs in conjunctionwith a GIS system to prepare management plans for community forests. They describehow the advantages of spatial accuracy, authenticity, consensus and trust through useof these ‘photo-maps’ were illustrated:

‘…when one woman, frustrated at trying to interpret a paper [PRA] map drawnby an earlier group, cast it aside and picked up an aerial photograph withwords to the effect that: “This is real, let me see it.” During the course of theevaluation it became more and more apparent that, in spite of all facilitationmeasures to the contrary, participatory maps largely represented the percep-tions of one or two dominant people.’

Combining conventional PRA mapping with the use of aerial photographs provides auseful context in which to interact with local resource users and develop an under-standing of natural resources from a local perspective. Mapping can apply to naturalresources, whether mammals, insects, soils, reptiles or plants, in space as well as time. Innorthern Canada, for example, a map drawn for a project field worker by local hunterson northern Baffin Island illustrated the interaction between narwhals and killerwhales, in their habitat at the margins of sea ice and open water. In this case, mappingenabled a greater understanding of how killer whale and narwhal behaviour related tothe location and timing of spring ice break-up and land ice composition than discus-sions with local hunters would have done (Brody, 1976). Participatory mapping caninclude topics as diverse as vegetation mapping and maps by local beekeepers showingtheir knowledge of bee-hive location in relation to vegetation and topography.

Photo-mapping also has its disadvantages, however, particularly the difficulty ofcovering large areas (for example, in arid environments) and the cost of gettingenlargements from 1:50,000 photographs. In these cases, thorough toponymic (localplace name) surveys mapped onto standard 1:50,000 scale or smaller can providecommon reference points that get round some of the problems of scale through provid-ing common reference points. Alternatively, an innovative system linking localknowledge with user-friendly GIS technology also has the potential for resourcemapping by local people over large areas. In this case, local trackers, who have no liter-acy but excellent field skills, have been using palm-pilot field computers designed withuser-friendly icons by expert tracker Louis Liebenberg and computer scientist LindsaySteventon to monitor wildlife such as rhino. In Zimbabwe, local basket makers aretesting a modification of this system for monitoring the impact of Berchemia discolorbark dye use (see Chapter 5).

Suggestions for participatory mapping are listed below:


life form and species diversity, wherehardwood tree generation times are longand understanding of patch dynamics islimited.

On shorter time scales, traditionalecological knowledge can provide valuableinsights into ecosystem functioning orhabitat/species associations. Goodexamples are the Kayapo terminology fordifferent forest types in Amazonia (Parkeret al, 1983), and information from

Anangu people on faunal habitat associa-tions of reptiles in Uluru National Park,central Australia (Baker and MutitjuluCommunity, 1992).

Local knowledge and spatial scale

The spatial scale of local knowledge variesaccording to social factors, such as mobil-ity, and with biological factors. Knowledgeof most resource users is usually focused


219

• Only undertake mapping exercises after permission is granted by the community towork in the area, and after you have established a good understanding of localvegetation and people–resource interactions. Mapping should ideally be precededby ‘walks in the woods’ and informal discussions with knowledgeable local people,and possibly by market surveys. Aerial photographs and soil maps can be very usefulat this stage.

• In any culture, and most of all in cultures with strong religious or symbolic attach-ment to the landscape, as in Aboriginal communities in Australia, mapping canonly take place on the basis of an established position of trust with regard to intel-lectual property and sites of cultural sensitivity. The religious and spiritualdimension of landscapes to many indigenous peoples is widely acknowledged. Sotoo are the varying perceptions of landscapes with gender and age. These factorsalone provide a strong argument for a community-based research and planningprocess to enable community control over sensitive information, and for separatemapping exercises with men or women or specific user groups.

• Once key resources have been identified through community discussions (resourcescarcity, importance), market surveys (high commercial demand) or community andinternational conservation concerns (overexploitation, endemic species), you needto consult key people within the community to select local experts on these issues.This will determine the type of map you will produce and what it will show(resource distribution, extraction routes, sale points).

• Choose a suitable place for ground mapping, where local people will feel relaxedto do the mapping exercise. Maps can be drawn in the sand or marked on hardersoil or grassy sites. Have suitable markers available (stones, sticks, tins). You mayneed to bring along large sheets of paper, despite the fact that this constrains mapscale and participation in the mapping process. In some places, for example,villagers have objected to drawing on the ground, affronted by the apparentassumption that they could not use pen and paper.

• Help people get started with the mapping exercise, but remember that it is theirmap, to be drawn on their initiative. Ground mapping often needs a lot of space.The first local person to put key points onto the ground map usually sets the scaleof the map. In some cases, when marking is easy (such as drawing map lines in thesand with a stick), the map may end up 15m or 20m long. This can be an advan-tage, enabling resource users to walk over a ‘conceptual landscape’, marking andnaming places and areas of resource richness.

• At the end of the mapping exercise, encourage someone to transcribe the maponto paper as a permanent record. Where culturally appropriate, add the names ofthe contributors.


within a ‘home range’ which will varyfrom large spatial scales (2000km2 ormore) among pastoralists and hunter-gatherer groups, to smaller scales amonggatherers in tropical forests or more seden-tary agricultural communities. Thepastoralists and hunter-gatherers havegreat mobility as they ‘track’ resourcepatches in space and time that are createdthrough natural events such as rainfall orby people through disturbances such asfire. In the semi-arid Kalahari sandveld insouthern Africa, for example, small groupsof part-time and full-time hunter-gatherergroups each ranged over an average areaof 1500km2 (Hitchcock, 1978). Withinthese ‘home ranges’, local ecologicalknowledge can be very detailed; know-ledge may decrease with distance awayfrom home ground.

In high rainfall areas, farmers’ know-ledge is much more localized, relating bestto mapping scales of 1:10,000 or less. It isequally important, however, to take intoaccount the more subtle influences onresource use at a much smaller scale,whether these relate to topography andsettlement patterns (see Chapter 2) orclumped distribution of resources. For thisreason, it is often useful to be aware thatone is working within a set of hierarchiesof scale and level of detail (see Figure 6.1).

Linking local knowledge to aerialphotographs or, in some cases, to satelliteimages enables the localized knowledge ofdifferent adjacent communities to belinked into a ‘patchwork’ covering a widearea and range of ecological zones. Innorthern Australia, for example, O’Neillet al (1993) have used Landsat imageryand a geographic information system(GIS) to compare the effects of burning oneucalyptus woodlands on different landtypes over a 3-year period in a 995km2-sized study area. This enabled acomparison of timing and size of burns byAboriginal people with those by pastoral-

ists, or ‘hazard reduction’ burns started bypark rangers with aerial incendiaries.Local people, including Aboriginal peoplefrom the Marralam community andpastoralists, assisted in the interpretationand ageing of burns. Field informationwas essential in order to identify theignition sources. Between 1988 and 1991,O’Neill et al concluded that, contrary tothe widespread belief that burning by theAboriginal community was destructive,impact was low and these perceptionsunfounded when considered at alandscape level. In addition, the studysuggested that lack of burning in eucalyp-tus woodland by Aboriginal people overthe past 100 years has increased thechances of high-intensity, destructive firesin the late dry season across northernAustralia. Indeed, efforts have been madein Kakadu National Park, northernAustralia to re-establish, as far as possible,traditional practices of burning in order tocreate a richer patchwork of habitats, tothe benefit of fire-dependent groups ofplants and animals (Lewis, 1989).

Cultural views of landscape

The cultural context of mapping stronglyinfluences how people interpret andunderstand landscape maps. RobertRundstrom (1990), who analysed Inuitmap-making, points out the importance ofunderstanding the ‘cultural filters’ throughwhich geographical information isorganized and retains meaning. On onehand, participatory maps help to reveallinks between people and the landscape.On another, interpretation of these mapsis improved by an ethnographic under-standing of these cultural links.

In the mapping process with localpeople, it is important to realize thatlandscapes and disturbance processes areviewed from very different perspectivesboth within and between cultures. In the

Applied Ethnobotany

220


same way that a ‘walk in the woods’approach gives the opportunity to improvecross-cultural communication forresearchers who are outsiders to thecommunity, so it is important to recordand cross-check local landscape classifica-tion systems.

Amongst formally trained researchersor planners, views of landscape are oftenclouded by orientation towards one acade-mic discipline or another. Appliedecologists are primarily concerned withbiological factors, while anthropologistsor social scientists focus on cultural andsocio-economic factors. Local resourceusers, on the other hand, may seelandscapes or the resources within themthrough a different set of ‘filters’ whichvary according to their age, gender, andspecialization (herbalist, honey-hunter,master fisherman, midwife). Within thesame human community, for example, aforest may appear to be a threateninghabitat, or a resource-filled series ofpatchy habitats for hunting, trapping or

harvesting timber. To African traditionalhealers, for example, as Victor Turner(1967) pointed out in his classic book oncognitive anthropology in Zambia, thesame forest would be a ‘forest of symbols’.Events such as rainfall, fire, cyclones orlightning, which create disturbancepatterns and resource patchiness, all havestrong cultural and symbolic meaning.Events which ecologists consider ‘natural’may be thought to be under supernaturalcontrol. In many societies, for example,rainfall, lightning and storms are widelyconsidered to be susceptible to the influ-ence of shamans and symbolic medicines.

Only with careful understanding andsensitive discussion can ethnobiologistsbegin to bridge the gap between differentinterpretations of the same disturbanceevent or landscape. This empathy is worththe effort, since cross-cultural understand-ing can give crucial insight into people’sattitudes as to how or why resourceswithin patches should be conserved. Thistopic is discussed in Chapter 7.


221


In many ways, the term ‘natural resourcemanagement’ is misleading since it hasmore to do with ‘people management’ thanwith managing natural resources. There isno doubt that we need to prioritize themost valued or vulnerable species, or findout how much can be sustainablyharvested (see Chapters 4–6). For goodscience to become good management,however, there needs to be wide socialacceptance of management plans or regula-tions. Achieving this, in turn, requires anunderstanding of the social, economic,ethical, religious and political factors thateither encourage resource conservation orlead to resource depletion. These are asimportant and as complex as the biologicalcomponent discussed in Chapters 4 to 6.Background information on these factorscan be gathered through local socialsurveys (see Chapter 2). There is also aclose link with commercial trade, whichbrings in opportunistic ‘outsiders’ as wellas income, as village economies shift fromsubsistence use and barter to links withexternal markets (see Chapter 3).

This chapter focuses on methodswhich can lead to a better understandingof tenure and boundaries, and resource-user characteristics which can form the

basis for common ground between localcommunities and modern conservationpractice. In particular, due to theircomplexity and significance in Africa, thischapter focuses on the interface betweencommunal lands and land set aside forconservation by state governments. Thisrelates directly to land and resource tenure(the rights and responsibilities underwhich land or resources are held) and theways in which individuals or groups ofpeople can use and control the area orresource, and the relations between stategovernment (such as national parksauthorities) and local people.

At first glance (particularly to mostbiologists), religion and social relationshipshave little to do with natural resourcemanagement or conservation. However, tolocal people and professionals in the socialsciences, the links between conservation,religion, social issues and land andresource tenure are clear (see, for example,Nhira and Fortmann, 1993; Richards,1996). Many national parks and forestreserves are recent. Seventy-six per cent ofCentral American protected areas, 65 percent of those in the Caribbean and 38 percent of those in South America weredeclared in the 1980s (McNeely, Harrison

Chapter 7

Conservation Behaviour, Boundaries and Beliefs

Introduction


and Dingwall, 1994). Although the firstAfrican national park was proclaimed acentury ago (1897), most have been estab-lished since the 1960s. Much of the landon which national parks have been createdhas a much longer human history, oftenwith complex cultural links to the present.Anthropologist Parker Shipton (1994)eloquently describes this link between landand culture:

‘…religion, ritual and cognition onone hand, and adaptation, suste-nance, and production on theother, cannot be kept pure of eachother. Land holding is at the centreof the confluence. Nothing evokesmore varied symbolic connota-tions or more intricate legalphilosophies. Nothing excitesdeeper passions or gives rise tomore bloodshed than do disagree-ments about territory, boundariesor access to land resources. Nor isanything more likely to preventmisunderstandings across cultures,harmful to both humans, and theirhabitat, than are thoughtful defin-itions of land holdings in the firstplace.’

This inevitably binds conservation (as oneform of land use) to the social world of

politics and religion – and, whether we likeit or not, this has to be better understoodif we are to achieve conservation andresource management objectives.

Earlier chapters of this manual weredevoted to outlining ecological and scien-tific approaches to developing a functionalclassification of vegetation, as well asdescribing species’ resilience or vulnerabil-ity to harvesting. This required usingmethods of analysis at a variety of spatialand time scales. The same applies to theanalysis of factors that influence localresource management and conservationbehaviour, with the additional element ofcommunication that is central to humansociety. This chapter deals, firstly, withconservation behaviour and the ‘ingredi-ents’ which are considered likely forsuccessful community-based conservation.It then deals in more detail with themethods used to define some of those‘ingredients’: in particular, methods usedto identify tenure types, boundaries andthe most significant stakeholders inresource management. In the process,since conservation organizations aregenerally staffed by people with a biologi-cal sciences background, the chapter willsuggest relevant literature, giving theoreti-cal background to these importantpractical issues in resource management.


223

Conservation and the ingredients for common property management

The protected area concept has broadenedfrom strictly protected sites to protectedareas within a bioregional or ecosystemframework. In most cases, this requirescomanagement of resources within desig-nated zones (buffer zones, multiple-useareas). This change in approach hasincreased the need for methods which leadto community-based conservation.

It is now widely recognized that legis-lation on its own is often ineffective.Conservation areas in private hands orunder state control often face similarproblems. Results from a postal survey ofmanagers of private reserves, for example,listed poaching (81.3 per cent), lack ofcooperation from government (71.95 percent), budget deficiencies (56.3 per cent),


political unrest (53.1 per cent), andcommunity opposition to loss of access tothe reserves’ resources (43.8 per cent) asmajor problems (Langholz, 1996).

In many countries with high biologicaldiversity but impoverished governments,centralized control of national parks, forestreserves or the natural resources withinthem by patrol rangers is often ineffectivein preventing hunting in ‘strictly’ protectedareas. Effective in-situ conservation forblack rhino, for example, would costUS$400 per km2 (Martin, 1994). Based onhis experience in East Africa, for example,John Hall suggests that patrolling of forestreserves generally requires 2 forest guardsper 500ha (or 4 guards per 10km2) (Hall,1983). In most cases, neither this level offunding nor staffing are available. Even themost important forest conservation areasfail to meet this criterion. Bwindi-Impenetrable National Park, Uganda, forexample, is a 330km2 rugged, forested areaworld-renowned as the home of half theworld’s mountain gorilla population andthe forest with the highest biodiversity inEast Africa. Instead of the 130 patrolrangers considered necessary on the basisof John Hall’s experience, Bwindi has 30patrol rangers. Implementing a sustainablelogging programme requires even morestaff. In Afromontane forest in SouthAfrica, which has a relatively low speciesdiversity compared to tropical forests, andwhere only a single product (timber) andfew species are involved, a marking teamof one forester and two staff who onlyselect trees greater than 30cm dbh is onlyable to cover 5ha per day (Seydack et al,1995).

The challenge is to find appropriateand lasting solutions to conflicts overconservation land. Local participationthrough community-based conservationprogrammes within protected areas or inbuffer zones around their margins hasbecome widespread as one potential

solution. Much of the success or failure oflocal participation in conservationprogrammes hinges on the social factorsof relations, rights and responsibilities (seeFigure 7.1). To avoid expensive failures,the design of Integrated Conservation andDevelopment Projects (ICDPs) has to takelocal institutions, tenure and resourcemanagement systems into account. A firststep towards this is to analyse the physicalfactors (bioclimatic, topographic) of thearea, as well as the social factors (politicalorganization and institutions, socio-economic, religious) covered in thischapter. Understanding people’s conserva-tion behaviour in terms of ‘what theyconserve, why, where, when and how’ isalso an important step towards consensusand reduced conflict with modern conser-vation objectives.

In theory, this seems straightforward;but in practice, defining relations, rightsand responsibilities is far from simple.There has been too much generalizationon a range of very diverse and dynamicsituations. For every claim that ‘ruralpeople have sophisticated systems ofnatural resources management which havemaintained biodiversity for thousands ofyears’ (IIED, 1994), there are cases wherelocal people have destroyed high-diversityhabitat, or where the people livingadjacent to protected areas are recentmigrants. All of this is confusing to fieldresearchers, rural development workersand protected area managers in trying towork out what should apply locally to thelong-term benefit of both conservationand local communities. For this reason, weneed to be able to look beyond the smokescreen of conservation politics and untan-gle the complex interplay of ecological,political, religious, economic and socialundercurrents behind successful or failedexamples of resource conservation.

Chapters 4 to 6 discussed biologicalfactors that lead to resilience or vulnerabil-

Applied Ethnobotany

224


ity of plant populations to overharvesting,methods for collecting data, and the role ofthis information in developing practicalmanagement plans. A parallel set ofquestions relates to human societies. Underwhat set of circumstances is conservationbehaviour likely to take place without beingimposed upon people by those from theoutside? What makes people comply withcustomary controls, and what leads to theirbreaking down? How (and why) should wedistinguish between intentional conserva-tion practices by local communities, andones that are merely the inadvertent endproduct of some other action?

Laws, whether an oral tradition ofcustomary law, or written into nationallegislation, are only effective when just asmall minority of the population are likelyto break those laws. When most peopledisagree with regulations and disregardthem, the costs of law enforcement becomeimpossibly high. For this reason, control isonly effective when the majority of people

choose to abide by resource conservationregulations. How long are these systems ofcustomary conservation or resource tenurelikely to last when conservation goals arebased on long (ecological) time scales, andwhen cultural change is often rapid?Looking at a range of different situationsfrom more remote to less remote areas,where population densities are higher,communities less homogeneous andcommercial trade more prevalent, can leadto useful insights into this question.

Customary (‘traditional’)conservation practices

Are people conserving habitat orresources, or doing this for otherpurposes? Based on her work on tradi-tional conservation practices in Oceania,geographer Margaret Chapman (1987)suggested that people traditionallyconserve habitat or a resource when threeconditions are fulfilled:


Source: adapted from Scott, 1995

Figure 7.1 Much of the success or failure of local participation in conservation programmeshinges on the social factors of relations, rights and responsibilities

225

OBJECTIVES

RESULTS

GOAL

RelationsImproved relations

between conservationarea management and

neighbouringcommunity

RightsOffer the communitysustainable benefits

from the conservationarea into the future

ResponsibilitiesEnlist people withinthe neighbouring

community as locallyresident caretakers

Reduced conflictbetween

neighbouringcommunity (or local

interest groups)

Conservation arearemains of value to thecommunity as intact

habitat

The community takesup an active role inmanagement and

monitoring

Enhanced management capacity

Conservation


• when the area or species is valued;• when there is widespread realization

that these values are threatened byhuman impacts; and

• when there are social or political insti-tutions in place to enforce controls.

I believe that these three requirements arewidely applicable to conservation behav-iour. Whether conservation practices areintentional or not is a topical issue.

One approach, taken by Gary Klee(1980), was to divide local managementstrategies into ‘inadvertent’ and ‘inten-tional’ conservation practices. Each ofthese can be further subdivided under thedifferent forms of control, such as limitedaccess, seasonal restrictions, limits placedon harvesting methods and so on. On theother hand, anthropologist RaymondHames (1987) suggests that: ‘Any persua-sive account of conservation as a humanadaptation requires a theory that showsthat conservation is by design, and not aside effect of some other process, specifiesthe conditions under which conservationwill evolve, and predicts how individualswill systematically regulate their behaviourto conserve resources.’ To do this,researchers and resource managers need totest how resource users behave underconditions of resource scarcity orabundance as well as during socio-economic change. They also need to avoida romantic view of customary (‘tradi-tional’) conservation practices that raiseexpectations but lead to resource deple-tion. What Thayer Scudder and ThomasConelly (1985) point out about manage-ment systems for traditional fisheries couldalso be said about some aspects of terres-trial community-based natural resourcemanagement (CBNRM) programmes:

‘As fisheries managers search forother strategies to “shore up” orcomplement ineffective govern-

ment regulatory mechanisms, thereis a real danger that too much willbe expected of traditional manage-ment practices. While somepractices may be directly relevant,others were designed to handlevery different circumstances.Without major adjustments andadaptations, it is unlikely that theycan be utilized today, especially ifthey have died out because thepeople involved no longer acceptthe type of leadership involvedand/or the underlying sanctions.Furthermore, even managementpractices which are still bothstrong and relevant must be fittedinto a more comprehensive strat-egy which includes strong externalsupport.’

Ingredients for community-basednatural resource management

A useful analytical framework for fieldworkers to use in sorting out the complexissues in predicting the likelihood ofsuccessful community-based conservationincludes the design principles for commonproperty resource management. The workof Elinor Ostrom (1990) on systems ofself-governance and Robert Wade’s (1987)study of village regulations on irrigationand grazing systems in South India arerecommended reading on this subject.These ‘design principles’ for viablemanagement of common propertyresources provide valuable guidelines indeciding where CBNRM programmesmight succeed or fail.

This is no ‘magic formula’ for success,particularly when dealing with the addedcomplexity of common property resourcemanagement: there are too many ecologi-cal, social, political and economicvariables for this. What we have a respon-sibility to do, however, is to highlight

Applied Ethnobotany

226



227

BOX 7.1 INGREDIENTS FOR SUCCESSFUL COMMUNITY-BASED

NATURAL RESOURCE MANAGEMENT (CBNRM) PROGRAMMES

Conservation and resource management through decentralized control to local commu-nities has been widely advocated in the past decade. The merits of local managementof communal resources have also been rightly questioned, as the capacity of local insti-tutions for resource management has been weakened by economic, political andreligious change. Before decentralization is advocated, it is important to have a predic-tive understanding of where community-based conservation is likely to succeed or fail.Land-use factors have a strong influence on this (see the following section on‘Ecological Factors, Land Use, Tenure and Territoriality’).

Boundaries

These include clear, accepted, controllable boundaries around the resource. They needto be defined and small enough to be controlled and monitored (see Figure 7.2e).

Tenure and Territoriality

Trust in long-term access increases the chances of sustainable management. Securetenure is one part of this. Long-term tenure is a major incentive for successful resourcemanagement and conservation, whether the land or the resources themselves areprivately or communally owned. Another part of this is that there should be local confi-dence of long-term access to resources, as agreed between the group which has legaltenure (such as national parks) and resource users within comanagement agreements(eg a memorandum of understanding). Based on his work in Nepal, Robert Fischer(1995) points out that formal tenure is less important than the confidence that agree-ments will be honoured.

Resource Predictability and Mobility

This includes predictability and low/no mobility. The greater the resource predictabilityin space or time, the greater the incentive for establishing property rights or manageduse. Examples are the strong rights attached to long-lived perennial resources thatprovide a predictable resource in unpredictable environments, such as wild tree speciesthat are sources of productive, favoured fruits or provide browse in arid/semi-aridenvironments (eg Boscia trees in East and Southern Africa), or the widespread privaterights to bee hives or trees with wild hives (see Figure 7.2f). With mobile resources suchas game animals or fish, private or common property rights apply to traps and trappingsites, rather than to the resource itself.

Relationship between Resources and the User Group

Resource value

The resource must be valuable to the group. As Marshall Murphree (1995) points out,‘effective management of natural resources is best achieved by giving the resource afocused value – to determine whether the benefit of managing a resource exceeds thecost, the resource must have value to the community.’


Applied Ethnobotany

228

Resource scarcity

The relationship between resource scarcity and vulnerability to human impact needs tobe recognized by resource users. If the resource users’ belief system does not link humanimpact (such as overhunting) with resource depletion (for example, as occurs amongstAlgonquin Indians in the Boreal forests of North America), then this poses a problemthat may even exacerbate overexploitation.

Resource commercialization

Customary controls break down rapidly with the shift from subsistence use to commer-cial harvest, particularly when commercialization is accompanied by an influx ofoutsiders.

Multiple-use lands, resources and multiples of users

CBNRM is favoured with fewer, rather than more, users and where resources have fewerrather than multiple uses. The more uses and users there are of a particular landscapeor resource, the more complex and potentially conflicting management becomes.

Composition of the User Group(s)

Group size

Smaller numbers of users are better than larger groups, but groups should not be sosmall that they have no social influence.

Group identity

The more clearly defined the user group, the greater the chance of success (eg localbeekeepers, herbalists, midwives, basket makers; see Figures 7.2b and c).

Location of resource users

Ideally, resource users need to live near the resource, or amongst mobile or semi-nomadic communities, and to frequent the resource area regularly. In either case, thissimplifies the monitoring of who is using the resource or resource area, and helpsrestrict access by outsiders.

Community homogeneity

Although no community is completely homogeneous in outlook, many are far moredivided in terms of socio-economic status and diverse interests. Social control overresource use is more likely to occur in homogeneous than in heterogeneous communi-ties. Homogeneity breaks down with the influx of outsiders into the area.Religious/ritual belief systems are widely accepted; these systems maintain grouppressure for actions that encourage short-term individual sacrifice in favour of longer-term group benefit.


where there is a greater likelihood ofsuccess or failure. Community-basednatural resource management is often anexperimental process, and it is far betterto experiment where success, rather than

failure, is most likely. Both broad and fine-scale processes need to be taken intoaccount in this process. It is tempting tofocus on local circumstances, but this israrely sufficient to assess local systems of


229

Setting and Maintaining Limits

Users’ knowledge

Users’ knowledge is best built on existing local knowledge of sustainable yields andresource status.

Rules for resource use

These need to be simple, practical, enforceable and appropriate (see Figure 7.2f).

Maintaining obligations

Mutual agreements reached on resource use need to be kept and there need to bedisincentives against individuals exploiting resources at the expense of the group.

Free riders

People who try to abuse the system need to be easy to detect. This largely dependsupon having small, clearly defined boundaries around the resource, and a small andidentifiable group of resource users who live near the resource.

Punishments against rule breaking

Consensus needs to be reached on punishments for breaking agreed rules. There shouldbe a sliding scale of punishments, but punishments for serious offences require ‘bite’ insocial or material terms.

Conflict Resolution

Well-developed mechanisms for conflict resolution should be established. Mediationmay be brokered by outside institutions such as non-governmental organizations(NGOs), or there may be internal mechanisms, such as conflicts which are expressedthrough witchcraft accusations and resolved by cleansing rituals and therapy.

Resource Management Groups and the State

The state should support and encourage, and be careful not to undermine, decentral-ized control. Where resource groups are effective in preventing an open-accesssituation and are managing resource use on a sustainable basis, state control should beminimized. It can play a crucial role, however, when local institutions need support –whether in law enforcement or through technical input.

Source: adapted from Wade, 1987


Applied Ethnobotany

Figure 7.2 (a) Ecological impacts and land-use conflicts increase as savanna and grasslands usedby pastoralists are encroached upon by wheat farmers (Rift Valley, Kenya). (b) Stretcher-bearersocieties (ekibiina by’engozi) are strong local institutions which form the basis of forest societies

(ekibiina kya’beihamba) linked to smaller groups of resource users (herbalists, midwives,beekeepers, basket makers) licensed to harvest specific forest products within multiple-use zones

of Bwindi-Impenetrable National Park. (c) Group identity: user groups are first nominated by thecommunity and then each issued with cards for that user group, building on the memorandum of

understanding (MOU) between the national park and community, and avoiding conflict withpark patrol rangers. (d) Ritual boundaries, tenure and timing: Nuxia congesta fuel wood shiftedfrom virtual ‘open access’ before harvest in Mount Kilum Forest Reserve, Cameroon, to privaterights after harvest through placing charms (ju-jus) on sticks next to logs. (e) Mapping ‘invisible’

boundaries: John Makombo’s work with local elders enabled customary boundaries to be mappedwithin and around Rwenzori Mountains National Park. (f) Private rights to bee hives are strictlyenforced by magic and customary law. (g) Community game guards in northern Namibia are at

the forefront of local conservation efforts facilitated by the integrated rural development andnature conservation (IRDNC) programme

230


resource use and management. All areobviously influenced by ecological deter-minants (climate, soils) on land use, andvery few (if any) local communities areisolated from wider economic and politi-cal systems. These factors stronglyinfluence the potential to develop coman-agement systems or community-basedmanagement systems. This provides thebackground for developing site-specificmethods and approaches to community-based natural resource management.

Approaches that have been acclaimedin one part of the world, such asCAMPFIRE (Communal AreasManagement Programme for IndigenousResources) in Zimbabwe or Joint ForestManagement (JFM) programmes in WestBengal, India, cannot simply be trans-ferred to closed-canopy tropical forest, forexample. What can be done is to getinsights into what failed or succeeded inother places, since there are often commonprinciples that structure the way thatpeople in different parts of Africa, ordifferent parts of the world, see particularissues, events or places (see Box 7.1). Inaddition, a legacy of past research andthought by ecologists, economists andsocial scientists provides importanttheoretical background for understandingthe social processes behind establishingaccess rights to land or resources. Theseare summed up in Box 7.2 for those inter-ested in further reading.

In your study area, it is important toidentify where consensus and conflicts ofinterest occur in terms of the following.

Land use

What are the current land-use systems andland-use potential? This is a good measureof habitat change and the lost opportunitycosts if land is set aside for conservationpurposes (aerial photograph analysis, cost-benefit analysis of land-use options,

overlapping users such as hunter-gather-ers, pastoralists, farmers, traders). Whatare the conflicts (with wildlife, access towater points, grazing versus farming)?What is the direction and rate of change?What is the gap between conservation andother forms of land use, or betweenexploitation now or sustainable use in thelong term? Cost-benefit analysis is a usefultool for assessing who pays the real costsof conservation. Where the ‘lost opportu-nity costs’ of communities are low, there ismore chance that community-basedmanagement will succeed. Where they arelarge, then local-level management isunlikely to succeed unless ‘lost opportu-nity costs’ are met through internationaland national support.

Institutions and stakeholders

How many institutional ‘layers’ exist, andwhere do these conflict? In today’s world,resource use and conservation are influ-enced at multiple levels, from global tolocal, and it is important to record wherethese are supportive or in conflict. Is thelevel international or regional (eg trans-frontier parks or land uses such asmigratory pastoralists); national (egnational parks and forestry departments);district or subdistrict (eg parishes such asUganda, or taluk such as India); commu-nity or group level (eg stretcher-bearersocieties – abataka – in Uganda, groups ofbeekeepers, herbalists, midwives); house-hold; or individual? Who has access? Areresource users the local people, ‘outsiders’or both? Who authorizes access? Are theywomen, men or children? Is this changing,and why?

As discussed elsewhere in this chapter,political power is derived or soughtthrough control over key resources. If acomanagement system is desired, then thequestion of ‘who controls what?’ needs tobe carefully considered. In several cases in


231


Applied Ethnobotany

232

BOX 7.2 ACCESS RIGHTS, ENVIRONMENT AND CULTURAL PRACTICE

Although well known by anthropologists and social scientists, few biologists are awareof important studies on tenure and territoriality. These can be broadly divided intothree groups.

Group 1

The first group, starting with Julian Steward (1936, 1938), ‘the father of humanecology’, developed hypotheses on how human territorial behaviour would be influ-enced under different conditions of resource distribution. Key contributions are thepapers developing two important conceptual models. The first model is spatial bound-ary defence, proposed by Rada Dyson-Hudson and Eric Smith (1978), which makes across-cultural comparison of hunter-gatherer territoriality. Dyson-Hudson and Smithtermed this the ‘economic defensibility model’, which was elaborated on in later papers(Dyson-Hudson and Dyson-Hudson, 1980; Smith, 1983). The second model is socialboundary defence. This was hypothesized on the basis of optimal foraging theory andfield work with foraging societies by Elizabeth Cashdan (1983) from studies withbushmen (San) peoples in Southern Africa, and Nicholas Peterson’s (1975) work withAboriginal communities in Australia. The social boundary defence model suggestedthat as resources become more sparsely distributed and less predictable, it is not adefence on the basis of territorial space, but on the basis of social groups.

Group 2

The second group includes key studies that contribute to our understanding, and offer astructure for analyses, of how people conceptualize the world, how moral order ismaintained and how this is evident in religion, behaviour and natural symbols. Thisenables, for example, a cross-cultural understanding of land and resource tenure insocial terms – how, when and why boundaries are maintained in order to keep outsidersout – and leads to an incentive to manage those resources. Major contributors in thisgroup are Claude Levi-Strauss (1963), Mary Douglas (1973, 1984) and Reichel-Dolmatoff(1996) in Amazonia; Victor Turner (1967, 1969), J M Schoffelleers (1985) and W M J vanBinsbergen (1985) in Central Africa; and Monica Wilson (1957), Alex-Ivar Berglund (1976),Robert Thornton (1980) and David Hammond-Tooke (1981) in East and Southern Africa.

Group 3

The third major contribution is made by Anthony Giddens (1984) and Pierre Bourdieu(1973, 1977), who elaborate on how social practices and cultural concepts link topeople’s use of space – its creation, control and maintenance over time – and whoexplain the processes through which people in authority exercise power. Two contribu-tions are particularly useful in linking institutional controls to resource management.Firstly, there is the link between social action and spatial arrangements through differ-ent forms of control (see ‘Boundaries and Tenure, Meaning and Mapping’ below).Secondly, there is the recognition that social meaning is produced and maintainedthrough specific activities. The frequency of these practices, such as boundary-markingrituals (whether within or around households, or at larger spatial scales) is one measureof the continued social recognition of these ‘invisible’ boundaries (see ‘Ritual, Religionand Resource Control’ below).


Africa, for example, forestry or conserva-tion legislation has been formulated toconserve valuable resources (such asindigenous fruit-bearing trees) that werealready conserved under customary law.This overlap between national legislationand customary law, although intended asa conservation mechanism by the state (byforest guards, for example), can be seen bylocal institutions as a mechanism forundermining local control.

Economic factors

What scale of commerce is associated withlocal communities? Even the most remotevillages are usually linked in some way tonational, regional and international trade.One measure is the proportional contribu-tion of income from external sources(through labour migrancy, for example)compared to different local sources (suchas agriculture, livestock, gathering,hunting). Based on his work in Lesotho,Steven Lawry (1990) suggests that relianceon external sources of income contributedto local lack of interest in intensifyingresource management. Another measureof association with external markets iscommercial trade.

Culture

Mutual obligations between people,households and different groups are at thecore of cultural logic behind land andresource tenure. This is reflected in, andinfluenced by, local religions and the way

they are changing. This change stronglyaffects what is locally viewed as ethical.Concepts of tenure and boundary demar-cation are strongly linked to belief systemsand symbolism. Supernatural power, suchas through the use of traditionalmedicines, is drawn on to enforce sociallyacceptable behaviour, and social conflictsare played out in the idiom of witchcraft.Again, this occurs at different spatialscales.

Change

No researcher or resource manager,whether local or an outsider, is starting atthe ‘beginning’ of things. Inevitably, therewill be a dynamic situation of overlappinginterests and forms of control at differentspatial scales (global, national, district,ward, household, individual). Some ofthese situations are ‘traditional’, othersrecent, and all will probably change in thefuture. Whether you are in the field orexamining aerial photographs, askyourself: how have things changed? Whatfactors appear to determine change orresilience in tenure, or social controls overnatural resources? Just as formerly isolatedvillages have been opened to widereconomic markets, so they are changing inresponse to religious and political change.This, in turn, strongly influences inadver-tent conservation practices, such asprohibitions on entry to hills that are rain-making ritual sites, or seasonal restrictionson some medicinal species.


Ecological factors, land use, tenure and territoriality

233

The factors influencing success or failure ofcommunity-based natural resource manage-ment programmes have to be seen againstthe background of climate, land form andland use (see Box 7.1 and Table 7.1).

Climate, soil type and land form canbe valuable predictors of:

• the likely level of conflict betweenconservation and other forms of land


use (such as agriculture);• resource predictability; and• local people’s territoriality and tenure

systems.

On the basis of work in Tanzania andMalawi, for example, ecologist RichardBell (1984) points out that in general thereis a link between lifestyle and land form(see Chapter 6). In Malawi, subsistencecultivators in areas with poor soils tendedto be the heaviest poachers, while fishingcommunities (with alternative proteinsources) and cash-crop farmers were lessinvolved in poaching. Climate, soil typeand land form also influence people’s terri-torial behaviour and tenure. All of theseare elements in the success or failure ofbuffer zone establishment or community-based natural resource management(CBNRM).

They also highlight the need for verydifferent approaches in conservation and

resource management work in differenthabitats such as tropical forest, savannaor semi-desert. Based on their work withhunter-gatherers in Southern Africa andAustralia respectively, Elizabeth Cashdan(1983) and Nicholas Peterson (1975)suggest that as resources become moresparsely distributed and less predictable,territories are not maintained on the basisof space (‘perimeter defence’), but on thebasis of social groups (‘social boundarydefence’) (see Figure 7.3). A similar situa-tion is suggested by Michael Casimir(1992), based on his analysis of factorswhich determine the rights of ‘traditional’pastoralists to grazing areas. Since plantbiomass production is closely linked torainfall (Lieth, 1975), Casimir analysedanthropological studies of pastoralists todetermine whether territorial behaviourwas related to the amount and predictabil-ity of rainfall. He found that fewcommunities grazing their flocks in areas

Applied Ethnobotany

Table 7.1 Opportunities for community-based wildlife management (CWM) vary with rainfall,soils and land use, socio-economic factors and the composition of local communities

Band Ecological Distribution of Comparative Stakeholder Opportunitiescharacteristics income wildlife value structure for CWM

1 High amounts of Skewed towards the Low Diverse interests, Lowwell-spread rainfall, better-off, or differentiatedyear-round supplies outside interests communitiesof surface water, moderate slopes and fertile soils

2 High amounts of Skewed towards the High Diverse interests, Possiblewell-spread rainfall, better-off, or differentiatedyear-round supplies outside interests communitiesof surface water, moderate slopesand fertile soils

3 Uncertain and low Equitable High Strong linkages Highlevels of rainfall, between and within poor or seasonal communities,supplies of surface reciprocal accesswater, steep slopes rights, mobileand poor soils livelihood strategies

4 Uncertain and low Equitable Low Strong linkages Possiblelevels of rainfall, between and within poor or seasonal communities,supplies of surface reciprocal accesswater, steep slopes rights, mobileand poor soils livelihood strategies

Source: IIED, 1994

234


with rainfall of less than 600mm per yearattached fixed ownership rights to theirpasture, confirming the hypotheses ofRada Dyson-Hudson and Eric Smith(1978) and Elizabeth Cashdan (1983) with

respect to hunter-gatherer territoriality.Conservation is just one form of land

use, and it is important and useful toevaluate and compare benefits that couldbe expected from the same piece of land


Sources: (a) Cashdan, 1983; (b) Dyson-Hudson and Smith, 1978; (c) Casimir, 1992; (d) Smith, 1994

Figure 7.3 Conceptual models and variation in access rights amongst hunter-gatherers and‘traditional’ pastoralists. (a) The effects of competition and territory size on territoriality. (b) The

conceptual model of economic defensibility linking resource distribution and defensibility. (c) Long-term annual rainfall averages and different levels of access rights to grazing. (d) Whatboundaries? Beware of imposing Western concepts of tenure: a map of three !noresi (resourceareas on which !Kung groups depend, drawn by a !Kung man showing what has been termed

‘zero dimensional tenure’)

235

LOW(large territory)

(a)

Abundance of resources

Competition

HIGH(small territory)

Social boundarydefence

Perimeterdefence

No territoriality No territoriality

LOW

(b)

LOW

Resource predictability

Resource density

HIGH

HIGHa. High mobility,

informationsharing,

spatio-temporalterritories

c. Geographicallystable territorial

system

b. Increaseddispersion

and mobility

d. Home-rangesystem

I

(c)

Categories of pastoral access rights

Annual rainfall (mm)1200

1000

800

600

400

200

0

Generalizedrights of access

Ownershiprights

(d)

284.9

II

379.8

III

358.0

IV

1121.6

V

989.6

VI

917.9

LOW

HIGH


under different forms of land-use options.The inherent conflict in any conservationaction is between short-term individualbenefits and long-term communal ones.Analysis of different types of maps atdifferent scales (using aerial photographsor local community sketch maps),combined with field observation, can raiseimportant questions about the influencesof people, livestock and wildlife onlandscapes (see Chapters 2 and 6). Whereare boundaries established at differentspatial scales? Where are apparent‘anomalies’, such as trees or woodland andforest patches that have not been felled ina landscape where most of the vegetationhas been cleared for agriculture? In somecases, patches of conserved vegetation,such as sacred forests and burial sites, willbe more apparent from aerial photographsthan from ground surveys.

An evaluation of different land-usesgives a large-scale picture of the gapbetween benefits from conservation andother forms of land use such as farming orkeeping cattle. Macro-scale land-use assess-ments provide important background to themicro-scale assessments of resource use thatare the focus of this manual. Land-useassessments can be a useful tool in giving anestimate of the extent of local or nationalsacrifice (the ‘lost opportunity costs’) thatmay be involved through setting aside landfor conservation. In an ideal situation,CBNRM programme planning should bepreceded by a regional land-use planningprocess. In the 40,000km2 Sebungwe regionof Zimbabwe, for example, at an early stageof what was to become the CAMPFIREprogramme, ecologists Russell Taylor andRowan Martin (1983) carried out a land-use planning process prior to recommendingbuffer zones outside of national parks.These buffer zones were suggested for landof low agricultural potential. Where arableland occurred, it was set aside for drylandcropping, rather than for any other form of

land use. In this case, the process ofplanning to minimize conflicting land useswas possible due to the fact that the vicinitywas relatively undeveloped, was a tsetse flyarea which precluded domestic livestock,and was mostly marginal for agriculture.

High-conservation priority habitats onhighly arable potential soils, in denselypopulated landscapes, are a far more diffi-cult situation, requiring greater state andinternational support for conservation inthe long term. The opportunity for priorplanning often no longer exists. In the highsoil fertility, densely populated mountainregions of East Africa, or on the alluvialsoils around Lake Victoria, agriculturalexpansion has already taken place right upto the boundaries of forest reserves andnational parks. In these cases, cost-benefitanalysis is a useful measure of where thegreatest conflicts and conservationchallenges are likely to occur.

At a fine spatial scale, group social-survey methods such as PRA can be usefulin initial identification of conflicts overland and resources. Farmers in Senegal, forexample, identified a matrix of types ofconflict during a rapid rural appraisal ofresource conflicts (Freudenberger, 1994;see Figure 7.4).

At a broad geographical scale, the netand gross revenue from agriculture indifferent agro-ecological zones in Kenyagraphically illustrates the high ‘lost oppor-tunity costs’ of land set aside forconservation in higher rainfall sites(Norton-Griffiths and Southey, 1995).Norton-Griffiths and Southey (1995), forexample, estimated that the cost of wildlifeconservation in Kenya already exceeds thebenefits from it, and that by maintainingwildlife conservation areas, instead of usingthe land in some other way, Kenya may belosing up to US$161 million per year.

Non-monetary cultural values cannotbe ignored, however. While cost-benefitexercises are useful, it is important to bear

Applied Ethnobotany

236


in mind that although economic rational-ism may be the goal of governments,minimizing risk rather than maximizingprofit often characterizes pastoralist(Coughenour et al, 1985) and small-scalefarming communities.

Using participatory scoring methods(see Chapter 2) with separate groups ofmen, women and boys from Jinga villagein Zimbabwe, for example, researchersshowed that non-monetary benefits, suchas water retention and rain-making rituals,were rated far more important thanproducts such as building poles, fruits orfuelwood to which monetary values couldbe attached (Hot Springs Working Group,1995; see Figure 7.5). For this reason,modelling based on ‘economic rational-ism’ does not present the full picture.

Diversifying activities in a risk-proneenvironment is a rational approach. Toachieve this, many small-scale farmingfamilies and pastoralists prefer tominimize risk through involvement in arange of activities rather than aiming tomaximize profits by focusing on a singleactivity. Access to harvestable resourcesand religious links to ancestors or ritualcontrol of rainfall are all considered tominimize risk. It is also important torecognize that while the ‘lost opportunitycosts’ of conservation may be low on anational scale, they can be very highlocally. A basic principle behind multipleuse is to help off-set some of these lostopportunity costs, and to better justifyconservation as a form of land use.


Source: Freudenberger, 1994

Figure 7.4 A matrix of types of conflict over different natural resources at various institutionallevels developed during a rapid rural appraisal of resource conflicts

237


Secure tenure is considered an importantingredient for resource management andconservation. In some cases, the oppositecan also be true. Based on years of fieldwork, anthropologist Paul Richards(1996) is convinced that many forests inWest Africa survived not because theywere under the control of a single, centralauthority, but because they are old

contested domains (‘boundary wilder-nesses’) over which undisputed controlhad not been established (see Box 7.4).Many of these forests are now proclaimedas forest reserves or national parks.

Whether you are working in areas thatare privately owned, in a national parkbelonging to the state, or in a communalarea, it is crucial not to fall into the trap

Applied Ethnobotany

Property rights: land and resource tenure

238

Source: Hot Springs Working Group report, 1995

Figure 7.5 A diagram drawn by villagers showing the location of sites around Jinga village inZimbabwe and restrictions on land and resource use in those sites that lead to maintenance of

habitat and species in these localities


of simplistic assumptions or simple classi-fication of different types of tenure. Thisis particularly important where localpeople or indigenous communities haveestablished complex rules controllingaccess to communal land, which in turnhas been overlapped by state control (as inthe case of many national parks or forestreserves). It is also important to recognize,firstly, that land tenure can take manydifferent forms (anthropologist PaulBohannan’s 1963 paper is still recom-mended reading on this issue), andsecondly, that land tenure and resourcetenure may differ.

It is equally important to avoid misun-derstandings about the term ‘commonproperty’. These have raged on for 30years in the belief that they inevitably leadto a ‘free for all’ situation of resourceexploitation. A centre of much of theconfusion was the famous essay by humanecologist Garrett Hardin (1968), entitled‘The Tragedy of the Commons’, whichintended to show that freedom to have anynumber of children would lead tooverpopulation. To do this, he used theexample of a common grazing area,where:

‘The rational herdsman concludesthat the only sensible course forhim to pursue is to add anotheranimal to his herd. And another;and another…but this is theconclusion reached by each andevery rational herdsman sharing acommons. Therein lies the tragedy.Each man is locked into a systemthat compels him to increase hisherd without limit – in a worldthat is limited…Freedom in acommons brings ruin for all.’

This essay was as controversial as it wasinfluential; but it was unfortunate in thatcommon property resources were

confused with open-access resources. Toavoid similar misunderstandings, a usefulstep is to analyse types of tenure. Differentways of doing this are described in thesections that follow.

Basic types of tenure and tenurialniches

In order to clarify the differences betweenopen-access and common propertyresources, Daniel Bromley and MichaelCernea (1989) distinguished four basictypes of property rights (see Box 7.3).

These basic definitions are useful inmaking this distinction, but it is essentialto understand that the situation is far morecomplex. Tenure and property rightsclearly have to be seen in cultural context.It is important to look beyond the politi-cal, economic and legal aspects of tenureand examine the cultural logic and socialrules that support local tenure systems.

It is also useful to explore cases ofconflict over land or resources and identifythe systems of conflict resolution. Recordsof conflict from district courts, or conflictdealt with by traditional leadership undercustomary law, can also provide usefulinsights into rule-breaking and fines.Where the same issues are dealt withthrough an overlap between customary lawand constitutional law, it may also beuseful to compare the severity of thesedifferent processes. There can be a greatvariation between the two processes, withcustomary law being far more severe thanstate law, or vice-versa, reflecting differentperceptions of the interest groups in eachcase. Who are the different groups involvedin the conflict? Is the conflict withinfamilies (men/women, immediate orextended family)? Is it between long-estab-lished people and ‘newcomers’? What isthe severity of punishment for infringe-ments and what form does this take? Whoare the beneficiaries of these fines?


239


Specific court cases can provide a veryuseful illustration of the complexity oftenure. Melissa Leach (1994), for example,describes a local court case where palmwine was stolen from a Raphia palm treebeing tapped for palm wine in SierraLeone. Two levels of tenure were involvedin this case. Firstly, those held by the palm-wine tapper and secondly, those held by thefarmer who granted rights to the palm-wine tapper in exchange for periodic giftsor labour. Somewhat counter to whatmight have been expected, it was not thepalm-wine tapper who took the suspectedthief to court, but the farmer who grantedthe right to tap the trees in the first place.

Two useful approaches have been usedto get beyond the simple classificationshown in Box 7.3. The first identifies tenur-ial niches and the second (see the followingsection) analyses different types of tenure interms of specific characteristics.

The tenurial niche approach wasdeveloped by John Bruce and LouiseFortmann of the University of WisconsinLand Tenure Centre (Bruce and Fortmann,

1989). It was used to describe differentproperty relationships in terms ofcategories of land used by different groupsof people for a range of purposes. A tenur-ial niche is defined as ‘a space in whichaccess to and use of a resource is governedby a common set of rules’ (Bruce et al,1993). Identifying tenurial niches is auseful way of categorizing differentproperty regimes where systems of statecontrol are superimposed over ‘tradi-tional’ tenure systems. This is a commonsituation in Africa, for example, wherestate control, based on Western notions oftenure, and traditional tenure systemsoften occur side-by-side. The tenurialniche approach is also useful in definingthe different claims to plant resources –such as trees, tree products (fruits,medicines from roots or bark) or thatch-ing grass – by different groups of peopleon land under different forms of tenure(see Figure 7.7). On the basis of field workin communal areas in Zimbabwe, forexample, Calvin Nhira and LouiseFortmann (1996) identified six different

Applied Ethnobotany

240

BOX 7.3 FOUR TYPES OF PROPERTY RIGHTS

Property rights comprise the following:

1 Private property resources: an individual has a right to exclude all others from usingthat resource. Decisions are made by the single owner.

2 Common property resources: this may be thought of as private property for a socialunit or community, where outsiders are excluded. This is the main form of propertyright in sub-Saharan Africa. Decisions require consensus between members of thegroup before actions such as exclusion are taken; consensus must be established,generally through a community meeting.

3 State property resources: in order to prevent overuse and/or to gain revenue, thegovernment restricts the way that people may use a resource.

4 Open-access resources: there are no property rights and no rights of exclusion. Forthis reason, this is sometimes known as a no property regime. Without any rights ofexclusion, individual resource users ignore the costs that their resource use willhave for future users, often leading to the inevitable resource degradationdescribed by Garrett Hardin (1968).

Source: adapted from Bromley and Cernea, 1989



241

BOX 7.4 ELEVEN CHARACTERISTICS OF TENURE SYSTEMS

In order to understand complex forest tenure systems, David Haley and Martin Luckert(1990) chose the following 11 characteristics to describe and compare different foresttenure systems used in Canada because of the link they made between economic behav-iour and tenure. Most (94 per cent) forest land in Canada is public land controlled bythe state. In many cases, timber-harvesting rights and, in some instances, forest manage-ment responsibilities are allocated to the private sector. Just as with other plantresources and tenure systems in communal areas in Africa, the complex tenure systemscontrolling access to timber in Canada are an important influence on how tenureholders behave.

1 Exclusiveness

To what extent are tenure holders able to control access? Exclusiveness, referring to theright of tenure holders (individuals or a group) to exclude others through controlledaccess, is a crucial means for tenure holders to derive benefits from a valuable resource.In an open-access situation, there is no exclusivity. Exclusivity is also difficult to maintainwith mobile resources such as migratory birds, fish or mammals, and much easier forplants, which generally stay in one place! For this reason, controlled access may applyto traps or trapping sites for birds or fish, rather than to these mobile resourcesthemselves. Amongst many farming communities throughout Southern Africa, privaterights are also accorded to wild fruit trees conserved in cleared fields or near tohomesteads, whereas anyone can collect fruits in uncleared woodlands. Clear bound-aries around the resource or land area and a clearly identified group of users areimportant factors in controlling access (see Box 7.1).

2 Comprehensiveness

Over what range of resources does the tenure holder have control? In Canada, rightsare granted to harvest timber, but not to wildlife resources. Similarly, on theMozambique coastal plain in Southern Africa, palm-wine tappers have exclusive rightsto tap Hyphaene palms in a particular area, but this does not apply to grazing rights orrights to fruits from trees in that area.

3 Rights of the Tenure Holder to Benefits

In the Canadian case, the economic benefits flowing to the private companies who hadrights to harvest timber were limited by various taxes and fees (such as stumpage fees)levied by the state. In Southern Africa, 50 litres of palm wine had to be provided to thelocal headman at the start of tapping a newly granted area. In addition, income tomen tapping Hyphaene palm wine in palm savanna to which they had tapping rightswas split in a 2:1 ratio with women who transported the palm wine (see Chapter 3).The flow of benefits to the tenure holder may also be influenced by social rules whichaffect how the benefits are divided up. In Sierra Leone, for example, anthropologistMelissa Leach describes a common situation where Mende wives have rights to use thefruits and fuel wood from trees their husbands conserve in coffee and cocoa farms(Leach, 1994).


Applied Ethnobotany

242

4 Transferability

In addition to the right of the tenure holder to the benefits discussed above is theimportant issue of allocating use rights to other people. In the Canadian example,monetary benefits dominated the transfer of timber licences, subject to state (minister-ial) approval; but in other circumstances, a much wider range of benefits may beinvolved, such as political allegiance, barter goods or services. It is important to identifylocally or nationally established rules that place limits on how (or to whom) tenureholders are able to pass on rights to resources or land to someone else. This transfer ofproperty or products can take place in various ways, such as through sale, barter orinheritance.

In the Namib desert of Namibia, for example, inherited rights by extended Topnaar(≠Aunin-Nama) family groups (!hao-!nas) are attached to !nara melon patches(Acanthosicyos horridus) in the Khuiseb delta (Budack, 1983; Dentlinger, 1977). Topnaarcouncillors are called ‘fathers rich in !naras’ (!naraaxa //gun) due to their role in settlingboundary disputes about !nara ‘fields’ (Budack, 1983). Another example of inheritablerights are those applying to fish trapping sites in Kosi lake system and Phongolo flood-plain, South Africa. In both cases, the barriers, which are built to accommodate the fishtraps, are erected only after discussion, leading to approval of the prospective trapperwith nearby trappers. This agreement is formalized by the headman and thereafter thesite is ‘owned’ by the trapper and is inheritable. Although there are still numerouspotential sites available to erect barriers in both wetlands, the trappers limit access,making it difficult for new trappers to enter the fisheries. Both systems have been inoperation for centuries and both are believed to be sustainable fisheries. The limitedentry into the fishery through site ownership is one of the reasons for this.

Complex social rules that affect who gets which benefits often exist and need to berecorded. When you investigate transferability, it is important to take extended familyrelationships and gender issues into account, and how these may be changing overtime. Inheritance may be passed on to men within a lineage, but not to men outsidethat lineage, or even women within it. Working in Mhondoro district, a communal areain Zimbabwe, for example, Fortmann and Nabane (1992) investigated how transferabil-ity was affected when women who had planted trees when married were laterwidowed or divorced. The method they used to investigate transferability was to inter-view widows and divorcees. They found that all 18 divorcees had lost the rights to thetrees they planted while they were married, even if they continued to live nearby. Theyfelt unable to do anything about this, and regretted planting the trees in the firstplace. The only exception applied to trees they had planted when they were part of awomen’s group, leading the researchers to identify this as a potential opportunity forstrengthening women’s rights to trees on land controlled by women’s groups. Nine of15 widows interviewed, however, had retained rights to family trees. The main reasonfor this was that they remained in the family home. Those that left the family homelost all rights to the trees.

5 Use Restrictions and Changes in Tenure Types

Does any form of zoning apply? For example, can garden plots controlled by women beconverted to agricultural fields controlled by men? Restrictions may also apply to howor when a resource is used. Closed seasons are a common way in which resource use islimited under customary law. Closing particular sites to any use is also widespread. Bothclosed seasons and closed areas are often linked to religious belief systems. Bans on



243

setting fires, collection of fuel wood and building poles, and cultivation or entry bymenstruating women commonly apply in Southern Africa to sites of religious signifi-cance such as burial sites or hills used for rain-making rituals. In South Africa, forexample, collection of the medicinal plants Siphonochilus aethiopicus (Zingiberaceae),Alepidea amatymbica and Peucedanum thodei (Apiaceae) only takes place duringwinter due to fear that collection during the growing season (spring/summer) willattract lightning. Enforcement of post-growing season gathering of Siphonochilus andAlepidea through the fear of causing storms and lightning is likely to be intentionaldevelopment of an avoidance practice by specialists (diviners and herbalists) in responseto depletion resulting from trade.

6 Duration

How long does right of tenure apply? This is a key issue in resource management andconservation when providing incentives to prevent short-term exploitation by individu-als at the expense of maintaining long-term benefits for the wider social group. Tenureover a longer time span is an important incentive for conservation and resourcemanagement. Based on his experience in developing joint forest management (JFM)systems in Nepal, Robert Fisher (1995) suggests that the longer the period during whichresource users are confident that they have use of a resource, the more likely they areto manage it sustainably. Duration of tenure is closely linked to security of tenure.

7 Security

Both duration and security are key factors affecting resource management behaviourand both depend upon the level of trust and extent of mutual obligations betweenthose granting tenure (whether individuals, a community group or the state) andresource user group(s). Secure tenure is also influenced by perception (how easy is it todetect rule-breaking ‘free-riders’), which in itself is influenced by how well the bound-aries of the resource and the resource user group are defined (see Box 7.1). A processleading to a written memorandum of understanding (MOU) between the state andresource-user groups within the local community is one way of increasing the level oftrust between resource users and the state through a set of mutually agreed rules andobligations. This can be useful where collaborative management or multiple-usearrangements are being developed for protected areas.

8 Operational Stipulations

To what extent can tenure holders make the more detailed rules and develop theirown management plans? In the Canadian forestry case, this covered three categories:management, harvesting and processing. Management requirements are there toensure sustainable use. In this case, reafforestation and protection were required.Harvesting requirements focus on efficient resource use and limits to ensure sustain-able yield. In Canada, a common requirement for processing was that tenure holdersneeded to construct or maintain a particular type of timber-processing plant. Limitsbased on the need for processing can also apply to tropical non-timber forest products,such as carving timber, rattan and medicinal bark.


tenurial niches where savanna woodlandmanagement occurred:

• indigenous woodland in communaland resettlement areas;

• trees controlled by district councils oncommunal land;

• woodland and forest controlled by the

state (forest reserves and nationalparks);

• trees planted by groups and institu-tions;

• trees planted and protected by individ-uals on individually controlled land;

• trees on commercial farms.

Applied Ethnobotany

244

9 Operational Control

Although the Canadian forestry example refers specifically to the ways and extent towhich the state ensured that the tenure holders followed operational requirements,size specifications and the allotment type agreed on, operational controls may also bedevolved to other levels. In protected areas (national parks, forest reserves), this will bebetween the state and resource users (individuals, groups). Outside protected areas,management responsibilities and operational controls are checked by locally basedauthorities, with minimal intervention by the state.

10 Size Specifications

How big is the area for which tenure is granted? Size of the area under tenure influ-ences management costs. David Haley and Martin Luckert (1990) point out that in afree-market system, if property rights are transferable and divisible, then the areaunder tenure tends to change to a point that suits a private economy of scale. This isnot possible with public land, such as Crown forest tenure in Canada; but a balance hasto be reached between size of the area for which tenure is granted and the creation ofmonopolies over very large areas.

11 Allotment Type

On what basis are tenure rights granted? This may incorporate one or several of thefollowing limitations on the basis of area, quantity of material harvested, defined usergroup and defined number of users. In some cases, limitations are granted for a specificarea (area-based). Alternatively, the limit is on the quantity of resource extracted(resource quantity-based). It may also be based on user groups and numbers of resourceusers. Depending upon the context and resource involved, quantities will be measuredin different ways: volume (m3) in the cases of commercial timber, kilograms or tonnes inthe case of formal medicinal bark extraction, or in local units such as bundles or handfuls.

Outside protected areas, and particularly within fields or the homestead area, rightsto resources are also affected by their location. This is an important factor to bear inmind during field work. In Sierra Leone, for example, Melissa Leach (1994) found thatthe small gardens behind kitchens were places where women acquired and retainedcontrol through individualized tenure. These spaces were intensively cultivated andcomposted, with medicinal plants and ‘wildings’ (self-sown seedlings transplanted fromnatural habitat) of indigenous trees such as Dialium guineense and oil palms plantedalongside other fruit trees.

Source: adapted from Haley and Luckert, 1990


Characteristics of tenure types

A second method provides a finer classifi-cation of different tenure types or niches.This is useful in avoiding simple classifica-tion of tenure into just four types (asshown in Box 7.3), and provides a farbetter basis for comparing and contrastingdifferent types of tenure. This method wasused in Canada by David Haley andMartin Luckert (1990) to analyse differentforest tenures which were employed asmechanisms to transfer timber-cuttingrights from state control to privateharvesters (see Box 7.4). Using thesecharacteristics enabled them to distinguish34 different types of forest tenure inCanada. The method has also been usedin Zimbabwe to develop a framework forclassifying different types of tenure incommunal areas. Each of the characteris-tics is explained in Box 7.4, with theaddition of African examples to emphasizethe wider applicability of this approach.

Secure tenure is an important ingredi-ent in natural resource management, but itis not a ‘cure all’. Neither weak tenure norhigh-value harvest, strong tenure nor lowresource values provide much incentive forlocally based resource management. Inaddition, when resource values are very

high, so is the temptation for resource‘mining’ rather than resource manage-ment, even if tenure is strong. Funds fromquick, high returns may be invested withinother areas, such as education or the casheconomy. Nevertheless, sustainableharvesting systems and resource manage-ment are both more likely where land andresource tenure is strong and the benefitsfrom sustainable harvest are high.

Recognition of boundaries that demar-cate particular resources or parts of thelandscape is a key aspect of tenure.Although obvious to local people, manyof these boundaries are ‘invisible’ to theoutsiders who are so often involved inland-use planning or conservation.Examples of such boundary markers arestone cairns, paths, river valleys, ridge topsor specific species planted at strategicpoints across the landscape, all formingboundaries which do not appear on anymap. To avoid misunderstanding andland-use conflicts, it is crucial to try to seethe world from a local perspective and tounderstand where such boundaries are andwhy they are widely recognized andrespected. This is covered in the followingsections.


Boundaries and tenure, meaning and mapping

245

Effective establishment and local accep-tance and recognition of boundariesaround a resource area are an importantbasis for avoiding an open-access, ‘free-for-all’ scramble for scarce resources. Thisapplies to private and state rights, as wellas to common property resources. Whenmany people think of boundaries, theythink of boundary demarcation usingfences, walls, lines of planted trees ornatural features. There is a second type of

‘boundary marking’ which it is equallyimportant to recognize. These boundariesare widely known within the localcommunity, yet are often ‘invisible’ tooutsiders, including urban-orientedresearchers in rural development orconservation. Physical structures used todemarcate boundaries, as well as treesconserved in fields, are therefore linked tosocial practice and belief systems that areassociated, in turn, with rules of behaviour


Applied Ethnobotany

Sources: (a) Fiona Walsh; (b) collection of J Kean, reproduced with permission

Figure 7.6 (a) Patch-burning of spinifex (Triodia) grasslands in Australia reduces the chances ofwildfires and creates vegetation varying in age, composition and resource richness. (b) A painting

of spinifex landscape, with black patches where the Nyananan men hunted for wallabies atTjikari, is rich in symbolic meaning and reflects the joy of the artist, Johnny Warangula

Tjupurrula, at returning to his country

246

a

b


at various spatial scales – from the individ-ual through to the district level.

The main reason why this is relevantto resource management is that it involvessome of the key requirements for resilientcommon property resource-managementsystems shown in Box 7.1: the need forclearly defined boundaries of the resourceor resource area, a defined user group,widespread acceptance of the rules govern-ing access to the resource, and effectiveways of detecting and punishing rulebreakers or of resolving conflicts.

For anyone interested in resourcemanagement, it is crucial to develop anunderstanding of what institutional andreligious factors stop people, eitherindividually or in groups, from depleting aresource in the short-term and favour theestablishment of managed long-termharvests. To understand people’s conser-vation behaviour, it is important tounderstand their ‘worldview’ or culturalperspective. This also is closely linked tothe relationship between people and thelandscape. The best examples of people’spowerful sense of place are those held bynomadic pastoralists and hunter-gatherersocieties (see Figure 7.6).

Insight into the social, symbolic andeconomic significance that land andresources have for people provides a betterunderstanding of what induces people tokeep to local rules about harvesting, wherethey apply, or what happens to rule break-ers as a result of unethical behaviour.

‘Invisible’ boundaries, unwritten rules

Developing an understanding of the socialprocesses that delimit space can be doneusing a range of social survey methodsdescribed in Chapters 2 and 6.

These include field observation ofpeople’s everyday activities, participatorymapping and participant observation of

special occasions such as boundary-marking rituals. The challenge is tounderstand what you see or are being told– not an easy task for outsider researchers.Ethnobotanists have many opportunitiesto observe activities directly linked tocultural control over land, resources andtheir harvesting, yet often neither ‘see’ noranalyse those activities. Amos Rapoport(1977) has suggested that any social activ-ity can be analysed into four components:

• the activity itself;• the specific way of doing it, and where

it is done;• additional, adjacent or associated

activities which become part of theactivity system;

• the symbolic aspects and meaning ofthe activity.

These are useful to bear in mind duringfield work aimed at unravelling the social,economic, ethical, religious and politicalfactors that either encourage resourceconservation or lead to resource depletion.Five important tools which can be used inthis process include the following:

1 Mapping

Local recognition and acceptance ofboundaries is a key determinant of who isallowed to do what, when and where. AsFigure 7.7 shows, ‘invisible’ boundariesare locally recognized at a range of spatialscales. While most anthropological workhas recorded how household space issubdivided (such as Adam Kuper’s 1980Southern African work, Pierre Bourdieu’s1977 study of Kabyle houses in NorthAfrica and Henrietta Moore’s 1996anthropological study in Kenya), maps canalso be produced at a broader spatial scale.Mapping these boundaries is a useful start-ing point, which can lead to a betterunderstanding of past land use. Ideally, use


247


recent aerial photographs as a startingpoint since they give a very useful perspec-tive on patterns of land use and tenure(Figure 7.8).

2 Identifying different socialgroups and working with keyassistants

The complex social groupings involved inestablishing boundaries at frontiers needto be recognized. So do the pitfalls ofworking through a few people (see Box7.5). However, local leaders and ritualspecialists will often have a more detailedknowledge of cultural codes and meaningsthat delimit space; the fact that someone isunable to explain these does not mean thatthey are unaware of them. Based on herfield work with Marakwet people inKenya, Henrietta Moore (1996) points outthat most people are rarely able to explaincultural codes, but can be very aware ofthe practical aspects that these codes implyfor social behaviour. For this reason,working in western Uganda, JohnMakombo (1998), a local researcher,mapped the boundaries controlled byridge-leaders using aerial photographs (seeFigure 7.2e). Field work with ridge-leaders(omuhulha wa balhombo) was a crucialpart of this process. Ridge-leaders alsoprovided information on the frequencywith which boundary-marking ritualstook place.

3 Recording how and when peopleuse space

Glen Mills (1986), an architect who hasstudied the social meaning of domesticspace, has suggested it is necessary toknow: what activities take place; who isinvolved; where this is done; when; inwhat order; and for how long. The sameapproach can be applied to ritual activitieswhich lead to habitat protection, such assacred forests, or to boundary marking.

Changes in the frequency of boundary-marking rituals is one indicator of how‘clear’ and controlled the edges of territo-ries are likely to be. The masay ritualsdescribed by Robert Thornton (1980)took place annually. By comparison, as aconsequence of political and religiouschange, John Makombo recorded thatboundary-marking rituals had not takenplace for 10 to 20 years in parts of theRwenzori Mountains, Uganda. With lowand declining frequency of renewal,control can be expected to be weak.

4 Plants and boundaries

Plants are commonly used as boundarymarkers, either planted at strategic pointssuch as entrances, boundaries of fields, orin fences of cattle pens, or pounded andsprinkled around the homestead. For thisreason, planting certain trees is synony-mous with claiming land, and the ‘simple’process of tree planting can result in bitterdisputes (see earlier section on‘Characteristics of Tenure Types’).

5 Toponymic surveys

Place names can be easy to record and saya lot about the past history of land andresource use (see the following section on‘Toponymic Surveys: Meanings of PlaceNames’). Geo-referenced place names canalso be a valuable guide in transferringresource-use information from participa-tory sketch maps or the field observationsof local people onto topographic mapsheets.

Toponymic surveys: meanings ofplace names

The study of local place names and theirmeanings (termed toponymy) as part ofthe mapping process is a useful way ofgetting insight into the past history of landand resource use. If this includes geo-refer-

Applied Ethnobotany

248



Figure 7.7 ‘Ritual topography’ at different spatial scales influences tenure and access to differentpeople. The symbolic division of the homestead after Kuper (1980) and of the hut after Davison(1991), where a = homestead head; b = other men of the homestead (umzi); c = primary womanof umzi; d = bride; e = other women of umzi; f = visiting men; g = visiting women; x = hearth; P

= strategic points (entrances, boundaries) where protected charms are placed

249

P

Lan

dsc

ape

scal

e(1

:20,

000–

1:50

,000

)

Co

mm

un

ity

scal

e(1

:120

0–1:

10,0

00)

Mic

ro-s

cale

(1:

100)

Ho

mes

tead

(fa

mily

)H

ut

(of

fam

ily h

ead

)(m

etre

s an

d le

ss)

The

ind

ivid

ual

Up

(W

est)

Do

wn

(Ea

st)

P

P

Cat

tle

pen

Bat

chel

ors

’si

de

Wiv

es’

sid

e

Fam

ily h

ead

’s h

ut

(in

dlu

nku

lu)

Wo

men

’ssi

de

Men

’ssi

de

P

P

a b

cd

e

gf

PPP P

P

PR

iver

P

P

Do

mes

tic

spac

e(p

osi

tive

: ho

mes

tead

s,cu

ltiv

ated

fie

lds)

Inte

rmed

iate

(am

big

uo

us

(neg

ativ

e/p

osi

tive

):g

rass

lan

d, s

avan

na,

sac

red

fore

sts,

lan

d-s

hri

nes

)

Fore

st(w

ild (

neg

ativ

e), p

lace

to

hid

e(p

osi

tive

)) (

acce

ss b

y m

en,

hu

nte

rs, t

rad

itio

nal

hea

lers

)

Inte

rmed

iate

spac

e

PP

Rig

ht

Left

P


Applied Ethnobotany

250

Source: (a) and (c)–(e) Mills, 1984, reproduced with permission

Figure 7.8 Aerial photograph analysis, combined with ‘ground-truthing’ with local people, is anexcellent method of gaining insight into the ‘nested layers’ of tenure, boundaries and local institu-tions and how these are changing. (a) A family homestead and farm in the Tsandi area, Namibia,demarcated with brushwood, where private (family) rights are attached to the Hyphaene palms.

Space within the family compound (eumbo) is further subdivided. (b) ‘Privatization’ with scarcityand local political change: former communal grazing on a drainage channel (oshana) fenced off

for exclusive use by a local person with economic (and local political) power. (c) Part of aneighbourhood (omikunda) in Ongwediva district showing farms as ‘islands’ divided by drainage

lines, each farm with a distinct boundary. Each family has strict rights to clay from termitaria,water from the hand-dug well and fruit from trees (eg Disopyros mespiliformis) within the farm

boundary, whereas fruit from trees within communal grazing areas is available to the widercommunity. (d) At a broader scale, neighbourhoods within part of a district (oshilonga). (e)

Diagrams showing the nested hierarchy of space, each governed by rules of access and rights



251

BOX 7.5 MAPPING METHODS: POTENTIAL AND PITFALLS

A recurrent problem facing the development of conservation or resource managementprogrammes is that, while boundaries are widely recognized locally as part of a culturallandscape, they are often not recognized by the state. Participatory mapping withdifferent groups to produce sketch maps (Chapters 2 and 6) is a useful starting point. InZimbabwe, for example, field workers worked with local villagers who mapped outhills to which access was restricted for ritual purposes such as rainmaking (see Figure7.5).

Another starting point is to map boundaries by walking with representatives of alllocal interest groups using aerial photographs and a global positioning system (GPS) tolocate and obtain coordinates for points on the landscape (geo-referencing). The use ofaerial photographs in developing community forestry-management plans in Nepal (seeBox 6.3) is a good example of this. Cheaper, hand-held GPS systems, costingUS$300–$400, have a limited accuracy. Differential GPS, which has a greater accuracy(to around 5m) requires a base unit with verified coordinates and a field unit, so is veryexpensive (around US$10,000). An advantage of GPS is that this technology enablesvery large areas to be covered. In the Amazon, for example, 470 Mengraknoti Kayapodemarcated their 4.4 million-hectare territory (Poole, 1995). Based on his recent surveyof the opportunity for applying this mapping technology to community-based conser-vation, Peter Poole (1995) concluded that mapping technology (GPS and GIS) enabledlocal communities to achieve five objectives:

1 Conserve and reinforce local/traditional knowledge.2 Enhance community capability to manage and protect lands.3 Raise and mobilize local awareness of environmental issues.4 Increase local capacities to deal with external agencies.5 Enable local and global groups to play reciprocal roles in global programmes for

biodiversity conservation.

Alternatively, if you do not have access to a GPS, then place-name surveys (see theabove section on ‘Toponymic Surveys’), combined with topographic maps (which showcontour lines), are useful in linking local knowledge to mapping scale. Both are effec-tive methods for mapping parts of the landscape under strong customary controls,working towards resolution of local boundary disputes or for understanding whetherlocally recognized boundaries can be used to demarcate multiple-use zones orrotational harvest areas.

It is crucial to be aware of three things, however. Firstly, boundary mapping can bea sensitive issue and it is important to avoid becoming a pawn in territorial disputesthat worsen the situation. Participatory mapping of the same area by different interestgroups can produce very different results. Secondly, some conservation areas are notplaced within customary boundaries, but in ‘separation zones’ between social groups.This is not an uncommon or unfamiliar situation in many parts of the world. The marchlands of medieval Europe, such as the disputed borderlands between Wales andEngland, are one example. African equivalents are the mopane woodlands betweendifferent pre-colonial kingdoms in what today is northern Namibia/southern Angolaand the ‘boundary wildernesses’ between pre-colonial African states (Erkkila andSiiskonen, 1992; Richards, 1996). Due to their no-man’s land status, these areas arefrontiers settled by people from many places and backgrounds. The net result, asanthropologist Paul Richards (1996) points out:


encing of place names, then this canprovide a network of geographic informa-tion points across the landscape which willenable the transfer of information onresource use from participatory sketchmaps, or the transfer of field observationsonto topographic map sheets. As part of asurvey of potential multiple-use zoneboundaries around the northern portionof Rwenzori Mountains National Park,Uganda, for example, John Makombo(1998) conducted a toponymic survey oflocal place names. With the enormousadvantage of being a Mukonjo researcherworking in his home area, he worked withfour to six people locally known to havethe deepest knowledge of place names,sometimes stopping every 200m during

‘transect walks’ when surveying eachmountain ridge. He also held groupmeetings where he cross-checked placenames to confirm their meanings andwhether any names had been left out.

In total, 38 focus group interviewswere followed later by site walks to placeswhich were identified during the meetings.Not all names had meanings, but 79 of therecorded names did, and their analysisprovides interesting results. Thirty-one (39per cent) of the names described the terrain(steep, rocky, boggy, infertile soils), 28(35.4 per cent) described places of resourceharvest, 14 (17.7 per cent) describedcultural/historical sites, and 6 (7.7 per cent)referred to habitat. Examples of placenames on the mountain ridge known as

Applied Ethnobotany

252

‘…is that the margins of most West African forests are extraordinarily complexfrom a sociological point of view. Typically, they will comprise a complex mix ofold refugee populations with new (colonial, post-colonial) migrant popula-tions. Each group in this mix will tend to seek to establish, to legitimize and tomaintain its own rights of occupancy and usage of forest resources reflectingthe circumstances of arrival of the group founders, and the history of the subse-quent incorporation of their descendants…to my mind, there can be no doubtsthat if conservationists are to devise effective schemes for the sustainablemanagement of forest reserve margins in West Africa, the careful study andanalysis of the cultural processes involved in creating these multi-layered, multi-valent identities is a pre-requisite.’

Probably the single biggest obstacle to understanding these complexcultural processes effectively is a prevailing belief among some conservationagencies that forests and forest margins have a single “true” owner with whoma once-and-for-all resource management deal might be struck, and that allother local groups are in some sense “impostors”.’

Under these circumstances, care and a healthy amount of scepticism need to accom-pany the interpretation of participatory maps. The outcome of the mapping processmay say more about the history and negotiating skills of the people who drew thatmap than about the geographic boundaries. Good examples are some of the mapsdrawn by local Solomon Islanders in the process of trying to resolve boundary disputes.These maps were studied by Peter Larmour (1979), who pointed out how two maps ofexactly the same land may be very different.

Thirdly, linked to the above point, boundaries and separation zones change. In1866 in northern Namibia, for example, missionary Hugo Hahn recorded a 60km widebelt of mopane woodland between the Uukwanyama and Ondonga kingdoms. Fiftyyears later, this woodland was 40km wide, and in the 1950s 10km wide. Today, this‘boundary woodland’ no longer exists (Erkkila and Siiskonen, 1992).


Kakuka, for example, are places referringto resource use, the first outside thenational park boundary, the rest inside thepark: Malindi (‘wild animal traps’), Bihya(‘pit-fall traps for wild animals’),Kamusonge (‘source of honey’), Masule(‘place of the liana Smilax anceps’ – usedfor basketry) and Bukungunia (‘place forresource harvest’ – bamboo). Forty-three(54.4 per cent) of the place names with

specific meanings are located inside thenational park. In addition, out of the 28names that describe places of resourceharvest, 24 (85.7 per cent) are locatedinside the park, with harvesting still takingplace in most of these today. Of the 43places located inside the park, 24 (55.8 percent) describe places whose names refer toresource harvest/utilization; 3 (7.0 percent) refer to cultural/ historical sites.


Ritual, religion and resource control

253

If we are to understand people’s conserva-tion behaviour, we have to understand the‘worldview’ that people have – that is, theway that people conceptualize particularevents, such as health or disease, good luckand misfortune, or ‘good’ or ‘bad’ behav-iour. If harvesting restrictions are going tobe employed, it also makes sense toidentify which restrictions will ‘resonate’with local people in terms of existingcultural controls on resource harvesting.

Just as many societies have ritual ritesof transition from one stage of life toanother, so there are ‘invisible’ boundarymarkers for individuals, men and womenwithin households, extended families, orneighbourhoods. These boundaries areclearest and most strongly enforced atfiner spatial scales (the house, the house-hold, a burial site, spring or a hill) andtend to weaken ‘outwards’ with increasein area from the neighbourhood to thecommunity, to the broader landscapeoccupied by other communities (seeFigures 7.8 and 7.9). People’s behaviour,sometimes with a conservation outcome,is linked to these boundaries. Somemountains, for example, are land-shrineswith a ban on access by menstruatingwomen. Single women are commonlydiscouraged from entering forests,although access by hunters (men) or divin-

ers gathering medicines is acceptable. At ahome level, coding of spatial divisions isstrong but still dynamic. The smallestspatial scale is the individual whose skin isa boundary crossed by incisions (symbolic‘doors’) where protective medicines(known as insizi in Zulu) have beenapplied at strategic points on the body, atstrategic times of his or her life. Rituals forboundary marking and protection arelinked by a common worldview at allspatial scales (see Figure 7.7). In manysocieties, the human body and the land areconceptually linked through ritual practiceand metaphor (van Binsbergen, 1988;Tamasari, 1995).

Entrances are the strategic pointswhere protective medicines are placed andreplaced in a process of social recognitionof boundaries and control over the spaceinside them. Each of these domains (wild,domesticated) ‘pushes’ against the bound-ary, requiring regular reinforcementthrough ritual process. Similar worldviewsare held by other agricultural societies,such as Aouan farmers in the Côte d’Ivoire(van den Beemer, 1994) and amongstBakiga farmers around Bwindi-Impenetrable National Park who havefelled forest to ‘domesticate’ this wild partof the landscape: it is no coincidence thatBwindi means ‘dark’.


Ritual control is strongest and itsrenewal most frequent at the homesteador individual level. In South Africa, forexample, Zulu traditionalists wouldundertake ritual cleansing (using Scillanatalensis bulbs and several other speciesmentioned below in the section on‘Metaphor and Meaning, Botany andBoundaries’) every one to two months(and even more frequently in cities). Ritualcleansing and strengthening of the house-hold would be less frequent (once or twicea year). The lower frequency of renewingboundaries, and weaker control at largerspatial scales, clearly has conservationimplications, where large areas are moreviable in terms of maintaining ecologicaland genetic processes than small areas. InKenya, for example, the average size ofsacred kaya forests is about 70ha. InZambia and Zimbabwe, the area set asideby wild places used as ‘land-shrines’, suchas large trees or spectacular hills, water-falls or caves, is also small and localized(van Binsbergen, 1985). In Nepal, forestsunder community management averageonly 23ha (Wily, 1995). A consequence ofthis is that while these small sites can bevery important for the conservation ofplant species, they have limited value interms of maintaining minimum viablepopulations of large mammals.

Reaffirmation of boundaries, ritual and control

Many landscapes in which conservationareas have been proclaimed are overlainby a ‘ritual topography’ dotted by farolder ritual features, such as hills andmountains of religious significance, or at asmaller scale, by trees, piles of stones orwatering points that demarcate bound-aries and tenure (see Figure 7.9). Thesehave direct links to land or resource tenureand customary restrictions motivated bylocal belief systems.

Establishing boundaries is important,but it is not by itself enough. Just as thepaint that marks lines on a tar road has tobe repainted, so the boundaries andentrances at all spatial scales, from theindividual level to that of the neighbour-hood, need to be periodically renewed (seeFigure 7.7). Without this, they graduallylose their significance and, if this occurs,weaken one of the mechanisms throughwhich local authorities exercise power(Giddens, 1984; Bourdieu, 1973, 1977).The frequency of these practices, such asboundary-marking rituals, is an importantmeasure of the continued social recogni-tion of these ‘invisible’ boundaries andculturally significant sites that are markersacross a cultural landscape.

Political control, whether throughtraditional leadership or, in many cases inAfrica, through political leadership ingovernment, is linked to symbolic orreligious power, where ritual specialists,such as diviners or shamans, mediate withancestor spirits to maintain social order.Table 7.2 shows a useful framework forthinking about ways in which controls areplaced on access to land or resources.

Ritual boundary demarcation not onlydivides ‘wild’ from ‘domesticated’ space, itcan also be used to expand and establishcontrol over new territories, in addition tomaintaining old ones (see Figure 7.10). Agood example of this is anthropologistRobert Thornton’s (1980) field work withIraqw agropastoralists in Tanzania, whichillustrates the links between land, religionand tenure:

‘The ritual condition of the land isinfluenced most greatly by the actsof men and the events that involvemen. These acts and events endan-ger the condition of the land in allrespects, interfering not only withthe order imposed on it by theperformance of the masay ritual

Applied Ethnobotany

254



255

Source: (c) Y Aumeeruddy

Figure 7.9 Ritual topography, boundaries and tenure: the frequency of reaffirming key points inthe landscape through religious practice is an important ‘field measure’ of the social significance

of boundaries that cross the landscape. (a) A pile of stones in Colophospermum mopanewoodland says little to an outsider, but speaks volumes to local people. This type of cairn, addedto stone-by-stone by passers-by, is found widely in Southern Africa. Like many things of symbolicsignificance, their local names have multiple meanings. (b) Situated inside a high-diversity tropical

forest conservation area, Etinde Mountain (‘Small Mount Cameroon’) retains strong religioussignificance to the Bakweri people, with libations to the ancestors made by everyone climbing tothe summit. (c) Each of these figures (thorma), made of barley, yak butter and water, representslandscape features of the Gunasa high pasture of Dolpo region, Nepal. These are made during a

special ceremony called Yulsa Cholsa, during which lamas propitiate the deities and demonsinhabiting the landscape in order to bring peace and prosperity to the local people, agriculture

and natural resources

a b

c


itself, but also with fertility, thegermination of seeds, predationsof birds, and with the rain itself.The object of the ritual is the landitself.

Of course, the land is theprimary productive resource of theinhabitants of the aya, and theorder and control of the land musttherefore be political by its verynature. The masay ritual is there-fore political. Ritual is politics andpolitics is ritual: these two spheresthat are distinct in other societiesare scarcely separable in Iraqwlife. The performance is politicalin at least two arenas: within thecommunity of the aya itself, andwithin the larger arena of interac-tion that includes all other Iraqwand all other ethnic groups in theregion.

First of all, since the ritual isconducted by the elders and by thekahamuse (“the speakers of the

community”), who are ultimatelyresponsible in disputes over landrights, the ritual legitimates theiroffices by virtue of the fact thatthey “create the aya” in the firstplace. The ritual thus reaffirms thepolitical status quo. But since italso draws up boundaries betweenthe inhabitants of the aya andother groups, thereby laying claimto a portion of land, the ritual alsohas important bearing on thepolitical ecology of the wholeregion. Where the Iraqw haveexpanded onto lands formerlygrazed, cultivated or hunted on byother peoples, they have used themasay rituals as the chief politicalinstrument by which they gaincontrol over land, and to legiti-mate their claim to it. Once themasay ceremony has beenperformed for a piece of land,“creating” it, as it were, out of thebush, the settlers who inhabit it

Applied Ethnobotany

Table 7.2 A schematic way in which Nguni people structure the world: a ‘traditionalist’ viewwhich has resonance in several other agricultural societies, such as Aouan farmers in the Côte

d’Ivoire and Bakiga farmers who have felled (‘domesticated’) much of Bwindi (meaning ‘dark’)Forest in western Uganda

Wild space Intermediate Domesticated space(wild = bad) (ambiguous) (domestic = good)

negative negative/positive positive

Spatial forest savanna/grassland/river cultivated fields, homestead(sacred forests, land-shrinesand grave sites)

Animal wild animals river crocodiles (negative) domestic animals(carnivores, inedible) cattle (positive) (cattle, goats (edible))

Social individualism social freedom community

Ethical unmerited misfortune fortuity merited misfortune

Spirit witches river people ancestral spiritsfamiliars mythical snakes, spirits

Human witches diviners/shamans moral people

Sexual forbidden, evil sex foreplay between accepted marital relations(witches with familiars) unmarried adults

Sources: Hammond-Tooke, 1975; van den Beemer, 1994

256


cling to it in the face of repeatedraids and adversity. The masayceremony is in such cases theirdeed to land.’

Moral authority relates not only to terri-tories, but also to valued resources or siteswithin territories. Examples are custom-ary restrictions against the felling ofcertain tree species; the hunting of totemicanimals; seasonal restrictions of hunting,fishing or gathering valuable commoditiessuch as salt; bans on hunting or collectionof products such as fuelwood; or eventotal exclusion from places of ritual signif-icance. People who break these rules areanti-social, and lay themselves open topublic sentiment against them, played outin the form of misfortune through witch-craft. Ritual specialists mediate betweenthe two main dangerous realms ofwildlands (forest and bush, the source ofpredatory animals) and heavens (thesource of lightning). Fear of lightning anddestructive storms, wild predatory animalsor the direct accusation of witchcraft byother people in the community all

maintain moral order, including people‘doing the right thing’ by conforming torestrictions on resource use. Ritual special-ists also play a key role in resolvingconflicts that result from anti-social activ-ity. If these cannot be resolved, then exilefrom that community can follow, eithertemporarily, with people returning afterritual cleansing, or permanently.

These links may seem strange toconservation biologists or protected areamanagers who come from a biologicalsciences background. With our interest inresource management, however, it iscrucial that we develop an understandingof what institutional and religious factorsconstrain people, either individually or ingroups, from depleting a resource in favourof managed harvests for long-term gain.

Metaphor and meaning, botanyand boundaries

There are many ways of seeing the world,and seeing the world in different ways canavoid misunderstanding and conflict –including conflict over conservation areas.


Source: Porter et al, 1991

Figure 7.10 Diagrammatic representation of the mediating role of an African ritual specialist(diviner). This conceptual model is widely applicable in East and Southern Africa, in this case

relating to Giriama diviners whose ritual space, the kaya, has played a crucial role in conservingimportant small remnants of East African coastal forest

257

DIVINTYANCESTORS/SPIRITS

DIVINERS

PHYSICALENVIRONMENT

Earth/vegetation/climate

PEOPLE(individuals)

andSOCIETY


To a botanist, for example, a marula(Sclerocarya birrea) tree (a commonSouth-East African indigenous fruitspecies) in someone’s field may be a maleor female tree and a remnant of thewoodland that originally covered thelandscape. To the local farmer, the sametree is not only a source of fruits andshade, but also evidence of the authorityhe (or she) has to determine who may haveaccess to the tree’s fruits. It may also be a‘property peg’, marking out a boundarywithin the field. To the farmer and hisfamily, the tree may be a land-shrine,where offerings are made to ancestralspirits who influence their lives. To thedistrict, this and other trees are a sourceof ubuganu beer, which acts as a ‘socialglue’, linking the community together atthe annual first fruits ceremony.

It is important to realize that conserva-tion and resource management are directlylinked to social behaviour and, throughthis, to ritual practice and religion. In manysocieties around the world, ancestor spirits,through ritual specialists, are a source ofmoral authority over the living. Thiscreates strong social pressure to conform,and is a powerful motivation behindmaintaining control over land and overvalued, scarce resources within thoseboundaries. Following these rules plays acrucial role in maintaining social harmonyand community health. Rule breaking thennegatively influences the production ofcrops, cattle, and children, as well assuccess in hunting and fishing (Berglund,1976; Knight, 1991; Reichel-Dolmatoff,1996).

While an understanding of tenure orbelief systems is unfamiliar to biologists,they have an advantage: that of recogniz-ing ‘invisible’ boundaries through botany.In her work in Sumatra, for example,Yildiz Aumeeruddy (1994) records howbetel nut palms (Areca catechu), whosefruits are used in offerings and chewed as

a social practice, are planted to discretelymark the boundaries of gardens. In Africa,plant characteristics such as succulence,spinyness, longevity, vegetative reproduc-tion from cuttings, and leaf colour allinfluence the selection process. In manycases, several of these characteristics arecombined in genera such as Aloe,Commiphora, Crassocephalum, Dovyalis,Euphorbia, Erythrina, Ficus and Solanum,all commonly used in boundary marking.They are also found in species introducedto Africa for ornamental or economicpurposes, which are rapidly appropriatedand adopted as boundary markers withsymbolic significance (such as Agave,Pereskia and many Cactaceae). Selectionof these same species for demarcatingnational park boundaries, as is commonaround forests in western Uganda, has acultural ‘resonance’ which results in widerecognition of the boundary-markingprocess. Boundary-marking species alsolink symbolism and social practice.

Less obvious, but equally important, isthe selection of certain tree species forentrances or as protective charms. Thisprocess, the plant medicines and themeanings of the names of the plant ingredi-ents often all have highly significantmultiple meanings (polysemy). Sprinklingmedicines around the home boundary forprotection against lightning or misfortuneis termed ukubethelela in Zulu, but can alsomean to hammer, to put up a fight or tobind the affections of a girl through use ofa charm. The term for this type of medicineis intelezi, which means ‘to smooth over adifference (theleza)’ or ‘to counteract’(other charms). The names of plantscommonly used as protective charms alsoreveal this strong link with boundaries andbelief systems: Drimea elata (indonganaizibomvana) means ‘red walls’, where red isa symbolic colour which consolidatesaction; Rapanea melanophloeos(umaphipha) means ‘wipe away dirt’ and

Applied Ethnobotany

258


‘help through difficulty’ as a result of itsritual cleansing function; Scilla natalensis(inguduza) means ‘make a beaten track orpath’ or ‘grope about in the dark’.

This polysemy is also worth noting byfield researchers, for three reasons. Firstly,lack of recognition of polysemy can lead tomisinterpreting records from local people.Secondly, the meanings of symbolicobjects, such as plants used to demarcate‘invisible’ boundaries, say a lot about thebelief systems that contribute to social andmoral order. Thirdly, just as overdifferenti-ation in plant names reflects their highervalue within that society, so multiplemeanings of things reflect symbolic power.The ‘memorial heaps’ of stone (cairns)found in many parts of Southern Africa

and known in Zulu as isivivane are oneexample (see Figure 7.9). The word isivi-vane not only refers to the cairns that markpoints on the landscape (to which passers-by have to add a small stone or suffer theconsequences of misfortune); the rootword viva refers directly to mobilizationand to social groups of people (such as acompany of soldiers). Survey beacons orpegs marking a boundary are anotherexample. Known in Zulu as anisikhonkhwane, this word also refers to thefoundation of a house, a wooden stake fortethering livestock, a boundary-markingpeg, or a wooden peg (often made ofPtaeroxylon obliquum wood) treated withsymbolic medicines to ward off lightning.


Who are the stakeholders?

259

Identifying areas or resources to whichdifferent forms of tenure apply is animportant step. Equally important, and ascomplex in terms of resource manage-ment, is the process of identifying who thestakeholders are. The view that ‘peopleliving adjacent to protected areas havefound themselves deprived of resourceswhich for thousands of years they had aright to utilize’ (IIED, 1994) has caughtthe imagination of policy makers and‘biopoliticians’. However, this is moreoften the exception than the rule. Tounderstand changes in tenure and accessrights to land or resources, it is useful tofind out what proportion of the localcommunity really fit the image of long-time residents of land within or adjacentto protected areas. This is often fewer thanone might expect. Using household surveymethods, for example, a 1990 surveyaround Liwonde National Park, Malawi,showed that only 5 per cent of inhabitantsliving around the national park were there

in the late 1960s, 13 per cent had movedinto the area during the 1970s, and 70 percent of the 1990 population had moved inthe 1980s from other parts of Malawi orhad arrived as refugees from Mozambique(Hall-Martin, 1993). The social complex-ity of West African forest areas is anotherexample (see Box 7.5).

Important questions in this processare: who are the resource users; what localinstitutions would be most appropriateand representative? The answers to thesequestions will vary depending uponwhether we are dealing with people livingwithin a protected area or adjacent to it,and whether they are sedentary or semi-nomadic. Should the beneficiaries be thosepeople who are the most disadvantagedfrom proximity to the protected area? Orshould they be the ‘locals’, rather than‘outsider’ settlers? Discerning who is‘local’ and who is not, in terms of accessto land or resources, however, is obscuredby the many ways in which people access


land: for example, by settlement,birthright, ‘creating’ extended family linksto people already resident there, or bygetting land through allocation from localgovernment, traditional authority, loan,rental or direct sale.

In principle, benefits need to bedirected to those living closest to theprotected area. In most cases, these are thepeople who are most affected by crop-raiding animals and loss of access to plantresources inside protected areas. This iswell illustrated by the household surveyswhich Rob Wild and his colleagues withCARE-Uganda conducted. Wild andMutebi (1996) recorded the number ofrespondents from communities adjacent toBwindi-Impenetrable National Park whowere involved in collecting forest productsand pit-sawing, or who were affected bycrop-raiding animals prior to park closure,compared to those away from the forest(see Figure 7.11).

In theory, resource-sharing arrange-ments should take place through local

community institutions set up for thispurpose, which should be representativeof the communities and of resource users.This is often easier said than done.Protected areas are often located in moreremote areas, where literacy skills may belimited. In many cases, resource users arefrom a sector of the local community withthe least economic or political power. Forthese reasons, local resource users aregenerally not well represented – even at thelowest level of formal local government –although they may be highly influentialmembers of their own communities. Inaddition, the administrative boundariesthat form the basis for local governmentwithin the nation state rarely conform tothe territorial boundaries of local commu-nities. This may further skew therelationship of who ‘represents’ communi-ties surrounding protected areas.

Three participatory rural appraisal(PRA) methods which are useful in thisprocess during short-term surveys are theuse of household mapping, wealth

Applied Ethnobotany

Source: Wild and Mutebi, 1996

Figure 7.11 Mean percentage of people from communities adjacent to Bwindi-ImpenetrableNational Park (n = 978) who were involved in collecting forest products or pit-sawing, or who wereaffected by crop-raiding animals prior to park closure compared to those further away (n = 1405)

260

Kilometres

Percentage of respondents100

80

60

40

20

0

Forest products

Animal damage

Pit-sawing

0 1 2 3 4 5 6 7 8 9 10


ranking and Venn (or ‘chapatti’)diagrams (see Chapter 2). One exampleis the result of the Venn diagram exercisefacilitated by Moses Saranta atOlorosoito Emurua in Transmara district,Kenya, where local institutions andleaders were identified by the Maasaiparticipants. This example illustrateshow close (or far) several local institu-tions are from the local community.

Carol Pierce Colfer, an anthropologistat the Centre for International ForestryResearch (CIFOR), has proposed amethod for identifying and defining themost significant stakeholders in sustain-able forest management (Colfer, 1995).This has been field-tested in the Côted’Ivoire, West Africa, in North Americaand in Kalimantan (Borneo), Indonesia. Itis based on a matrix, with different stake-holders on one axis and six factorsconsidered most relevant to the relationsbetween forests and groups of people on

the other axis. The factors defining themost significant stakeholders are:

• proximity to the forest;• pre-existing rights of tenure (recogniz-

ing that this varies from place toplace);

• dependency on the forest (for productssuch as food, fibre and medicines);

• level of local/indigenous knowledgeabout the forest;

• culture/forest integration in terms ofsymbolic links with the forest; and

• power deficits (people with little powercompared to other stakeholders).

Using a rating system on a scale of 1–3 (1= high, 2 = medium, 3 = low, with anadditional category for ‘var’ = variable),based on field observation and experience,Colfer then calculated mean scores fordifferent stakeholders (see Table 7.3). Shefound this a quick and easy method, but


Table 7.3 Criteria for identifying the most significant stakeholders in sustainable forestmanagement: an example from East Kalimantan, Indonesia

Stakeholders DimensionsProximity Pre-existing Dependency Indigenous Culture/forest Power Value

rights knowledge integration deficit

Dayak 1 1 1 1 1 1 1.00

Kutai 1 1 1 1 1 1 1.00

Transmigrant 1 variable 1 variable variable 1 1.00

Forest workers 1 3 1 variable variable 1 1.50

Small scale entrepreneur 2 variable 2 2 2 2 2.00

Company officials 2 3 1 3 3 3 2.50

Forestry officials 3 3 1 3 3 3 2.67

Environ-mentalists 3 3 2 3 2 3 2.67

National citizens 3 3 2 3 3 variable 2.80

Consumers 3 3 3 3 3 variable 3.00

Source: Colfer, 1995

261


points out that problems may arise withidentifying and defining the six factorsthat describe stakeholders, the weightingof those dimensions and the level of detailof the scoring method and cut-off point.Although refinements are needed, this isan interesting approach to a truly knottyproblem.

Although Rob Wild and JacksonMutebi (1996) did not use the ratingsystem suggested by Carol Colfer, thecharacteristics she employed werecommon to the stakeholders and commu-nity groups with whom they worked inBwindi-Impenetrable National Park,Uganda. It also illustrates the point thatlocal institutions involved in the processcan be recent, rather than ‘traditional’ones. New organizations should not bediscounted: in some cases, religiousgroups, rural women’s savings clubs, evensoccer clubs may have a role to play. In thecase of Wild and Mutebi’s study (1996),three community organizations wereidentified by the local community as themost appropriate to deal with multipleuse, forming a ‘forest society’ which wouldcoordinate the multiple-use activities. Twoof these, the local (formerly resistance)council (LC) and the stretcher-bearersocieties (ebibiina by’engozi) are recentorganizations, while the third, theabataka, has a longer history. Due to theirdifferent history and social and politicallinks, each is briefly described here.

The LC system is a form of govern-ment introduced to Uganda by the currentNational Resistance Movement (NRM)government of President YoweriMuseveni, which came to power in 1986after a long period of civil war. The LCsystem allows significant local self-deter-mination and consists of five levels, fromthe village (LC1) level to the parliamen-tary level (LC5). At the LC1 level, everyadult member of the community (approxi-mately 150 households) is a member of the

LC1 council, which elects a committee of9 to manage the day-to-day running of thevillage. The ebibiina by’engozi arestretcher-bearer societies which wereformed in the 1980s, possibly followingthe idea of similar groups in Rwanda. Themotivation for stretcher societies camewhen individuals found it increasinglydifficult to mobilize their friends andrelatives to carry their sick to the clinic.Membership within each community andthe attendance at meetings is compulsory,and there is a small monthly fee. Tomaintain this level of support, discipline isvery tough. The penalty for being absentwithout a good excuse when a patientneeds to be carried can be a drum of localbanana beer (200 litres), a fine represent-ing the wage-labour equivalent of US$30.

The abataka is the traditional commu-nity organization amongst the Bakigapeople of western Uganda. Loosely trans-lated, abataka refers to responsible adultswithin a geographical area, such as on aridge or hill, who form citizens’ groups. Insome cases, therefore, the word abatakarefers to an extended family group relatedto one great-grandfather. Abataka leader-ship is drawn from elders in thecommunity, frequently with a hierarchicalstructure of chairman, secretary andtreasurer.

The forest society liaises at the inter-face between the local community and thenational park management and is the‘umbrella’ organization under whichspecialist user groups have formed. Priorto this, most categories of resource user,such as herbalists, midwives, basketmakers or beekeepers, worked on anindividual basis and did not belong to anyformal association.

The signed ‘memorandum of under-standing’ (MOU) reached between thecommunity and the national park authorityon multiple use in the national park was animportant step once agreement had been

Applied Ethnobotany

262


reached on rules, rights and responsibilities.The MOU includes the agreement that thecommunity controls activities incompatiblewith the goals of the conservation area,such as gold mining, commercial loggingand illegal hunting. Also important is therequest for identification cards by specialistusers (beekeeper, herbalist, midwife, etc),which have the mutually agreed rules andresponsibilities printed on them. To date,for example, nearly 500 beekeepers in fourparish beekeeping societies are registered tokeep an estimated 3000 hives within multi-ple-use zones. In addition, theCARE-Development through Conservation(DTC) project assists beekeepers withprocessing and marketing surplus honey.Initial ethnobotanical surveys with tradi-tional healers and basket makers have beenfollowed up with PRA surveys in threepilot-study parishes, leading to harvestingof selected species. Each of these activitiesfalls within joint forest management (JFM)agreements developed within the parishesand signed by the forest societies and theUganda Wildlife Authority, including accessto certain footpaths through the forest andto a hot spring site considered to have spiri-tual healing qualities.

Each of these factors – clear bound-aries, group size, group identity, agreedrules, conflict-resolution mechanisms, andso on – relates to the ‘ingredients’ forcommunity-based conservation given atthe start of this chapter (see Box 7.1).

The question is: how successful is thisexperiment? In answering this, as with anysimilar project, it is important to resist thetemptation to claim early success. I suggestthat a ten-year minimum period is neededbefore this can be done. Too many well-

publicized text-book cases fail after a fewyears. There are positive signs, however.The incidence of fires has been greatlyreduced, and so has the problem ofbeekeepers setting snares (Wild andMutebi, 1996). The Mpungu ForestSociety has also been remarkably tough onrule breakers. In one example in 1994, theforest resource users alerted park staff toillegal hunting in the forest and a rangerpatrol apprehended one person whileanother ran away. The person who wasapprehended was fined 4000 UgandaShillings (about US$4) in the local policecourt. By contrast, the man who ran awaywas identified to the community by hisown stretcher society, through the mediumof the LCs and the local forest society, andwas fined 60 litres of beer and a goat (theequivalent of about US$20), a five timeslarger penalty.

There have also been changes withinthe Uganda Wildlife Authority. Just as thecommunity has built on its existing systemby forming a forest society, the UgandaWildlife Authority has added to its ownnetwork, strengthening its own capacityto interact with the community through acommunity conservation unit. Patient andparticipatory conservation work aroundthese parks is bearing fruit. Park staff nolonger interact just with law-breakers, butalso with responsible members of thecommunity who are constructivelyinvolved in multiple-use areas. As a result,instead of a situation where a mere 30patrol rangers are faced with policing arugged 330km2 of forest surrounded by100,000 people, there are the beginningsof a more promising future emerging outof a complex and conflict-ridden past.


263


The beginning of Chapter 7 mentions thatconservation behaviour requires threebasic conditions before an area or aresource is conserved: appreciation ofvalue, the realization of scarcity, andappropriate social or political institutionsto implement conservation or resourcemanagement regulations. Recognizing thatconservation practice requires an under-standing of social and cultural factors,economic driving forces and ecologicalprinciples, the preceding chapters havedescribed a variety of tools for identifying,and in some cases for measuring, thesethree ingredients of conservation behav-iour in more detail. The ingredients are:value, monitoring scarcity of the mostvalued and vulnerable categories ofspecies, and the cultural factors underpin-ning control of access to land or resources.In particular, this manual has focused onsteps used in priority setting when makingdecisions on conservation and manage-ment of wild plant resources. These can becondensed into 15 basic steps, hopefully amid point between overcomplication (a‘horrendo-gram’) and oversimplification(see Figure 8.1).

Two key themes of Chapter 7, tenureand boundaries, have a message for all of

us: there are limits to how much demandcan be met from sustainably managedharvests. These limits can be extendedwith more intensive management or culti-vation (Tiffen et al, 1994; Netting, 1993;Boserup, 1965), but there are boundariesto this as well. Slow-growing, slowlyreproducing, habitat-specific species whichare destructively harvested – the verycategory most vulnerable to overexploita-tion – are the least likely to be cultivatedbecause of biological attributes: returnsper unit time are just too low, unless valuesare extraordinarily high.

The Brazilian economist AlfredoHomma’s (1992) analysis of forest-product harvesting in the Amazon isequally applicable in Africa and to othervegetation types (see Figure 8.2). Thisshows four phases in the harvesting offorest resources: an expansion phase at theresource-rich frontier (or when a monop-oly is held on the resource in the market);a stabilization phase where there is anequilibrium between supply and demand;a decline phase brought about by resourcedepletion; and for some species, a subse-quent domestication phase.

Prices begin to rise during the stabi-lization phase when there is no increase in

Chapter 8

Striving for Balance: Looking Outward and Inward

Introduction



Figure 8.1 Fifteen basic steps towards resource management. Setting priorities on the basis ofecological principles, social and cultural factors and economic driving forces is essential. So is the

recognition and review of past mistakes, successes and political context

265

9. Who controls what,where, when and how?

Access rights and rules of use:• mapping land tenure (sketch

maps, aerial photographs);• restrictions and rules:

harvesting methods and gear,season, gender, locals orspecialist users only;

• institutions and stakeholdergroups (Venn diagrams,wealth ranking, discussions,rating of stakeholder traits);

• resource tenure (micro-scalemapping, tenurial niches andtypes, discussions andinterviews of rules andpunishments;

• customary law and conflictcase-studies;

• national and internationallegislation and conventions

USES AND USERS QUANTITIES, YIELDSAND IMPACTS

ACCESS ANDCONTROLS

RECOMMENDATIONS

PRIORITIZING SPECIES

REVIEW ANDREFLECT

IMPLEMENT, EVALUATE(AND CHANGE)

2. Who are the users?• identify primary user

groups (eg basketmakers,herbalists);

• other users in themarketing chain;

• men? women? children?• ‘locals’ or ‘outsiders’• wealth and education

status• national or international

trade

8a. No/low levelmonitoring

• local community/patrolranger records of rulebreaking (type, time,locality);

• trends in ‘harvest per uniteffort’

8b. Detailed monitoringof ‘indicator’ species

• harvest impactassessments;

• periodic regenerationsurveys;

• population size-classdistributions;

• matrix populationmodels and extinctionprobabilities

8c. Core conservationareas: monitorecological impacts

• indicators species fordecline in ‘keystone’species due tooverharvesting;

• loss of habitat/niches (egspecies inhabiting hollowdeadwood)

5b. Adaptivemanagement andmonitoringcategory

• locallycommon/limiteddistribution, large tomedium-sizedpopulations, fairlyrapid growth;

• few uses, lowcompetition betweenuses;

• use of exudates,deadwood, phloemsap and above-groundstems of vigoroussprouters;

• low–medium speciesdiversity vegetation

5c. Substitution category,low/no opportunityfor sustainableharvest

• restricted distribution,habitat specific, smallpopulation size, slowgrowth rates;

• harvest of whole plants,bark, roots, apicalmeristems (eg non-sprouting palms);

• high diversity vegetationwith multiple uses andusers;

• species that need specialconservation efforts:keystone species,phylogenetically distinctspecies and endangered,vulnerable or rare species(IUCN categories)

13. Recommendations:conservation andresource management

• buffer and multiple-use zoneswith clear boundaries;

• rules of access (including noaccess);

• sustainable harvest levels;• development of alternatives

and resource substitution;• practical monitoring systems:

– illegal activity monitoring;– impact monitoring (few

species);– population and

regeneration monitoring(very few ‘indicator’species (high value andvulnerability));

• policy and legislative changes

15. MonitoringHow effective is the implementationof conservation resourcemanagement plans andrecommendations? Beware ofclaiming successes too soon: acommon donor-driven response.Decide on (1) what to monitor: whatindicators and what scale (species?vegetation cover?, what level ofprecision?); (2) what frequency?; (3)who will do the monitoring?; (4)who will process and store the data?;(5) who gets the results?; (6) whatchanges need to be made?

12. Learning from thepast

Oral histories from ‘old timers’,PRA methods (stick graphs,time lines), analysis of aerialphotograph series (vegetationchange, settlement patterns),historical records and pastresearch; palynology andarchaeology for longer timeseries:• human ecology: land-use

change, settlementpatterns, health andepidemiology,population change:census records, localhospital records, aerialphotographs;

• habitat change andspecies loss;

• conservation andresource managementfailures and successes(why? how?);

• social, cultural, politicaland economic change

5a. High resilience and good opportunity forsustainable harvest

• common, large populations, species with rapid growth;• single or non-competing uses;• use of leaves, flowers, fruits and above-ground biomass

of grasses, sedges and reeds;• low species diversity vegetation;• low input monitoring

3. Which species aremost valued?

• PRA methods;• household and trader

interviews;• quantitative surveys

– marketplaces;– prices along

marketing chains;– field plots

4. What is the potentialfor sustainableharvest?

• geographic distribution;• population size;• part of plant harvested;• level of demand; single

vs multiple use;• vegetation diversity

7. What harvest levelsare possible?

• biomass, growth rate andyield studies;

• vegetation mapping andsampling;

• harvester criteria:selective harvest;

• identify areas closed toharvesting;

• regeneration surveys;• predictive models and

identification ofvulnerable life-stages

6. Where do mostvalued or vulnerablespecies occur?

• sketch maps bycommunity;

• aerial photographs;• herbarium records;• habitat requirements and

understandingdisturbance

1. What is used?Ethnobotanical surveys:• field surveys;• marketplaces;• households

OR OR

OR

OR

11. What alternatives tooverharvesting?

Identify locally acceptablealternaives: what are peopledoing already? under whatcircumstances (scarcity, higherpopulation densities)?:• alternative

species/resources;• pricing and control

mechanisms;• appropriate technology

and developmentalternatives

10. Learning from commercialtrade: history anddynamics

Commercial trade in wild plantresources catalyses changes inharvest intensity, frequency, whoharvests or controls harvesting andproduction of alternatives.Understanding trade leads to abetter understanding of resourcemanagement problems andpotential solutions:• studies of marketplaces and

trade (local, regional, national,international);

• who sells, who buys and wheredo products go?;

• prices and production ofalternatives

14. ImplementationConservation behaviour andbringing about change:• communication of valid

research results:;– local participation in

monitoring;– involvement of local

motivators;– use of drama and role plays;– use of video and television;

• incentives (social, religious,cultural, economic);

• institutional support at local,national or international levelsfor:– development of alternatives

to overharvest and habitatdestruction;

– policy and behaviouralchange


production to meet growing demand.Control over access and resource manage-ment may then take place. Resourcedepletion, whether through habitatdestruction or overexploitation (or both)is represented by the decline phase. Goodexamples of this are the decline of ebony(‘mpingo’) (Dalbergia melanoxylon) inAfrica and rosewood (Aniba rosaeodora)stocks in Brazil.

Alfredo Homma points out that thedomestication phase occurs during thedecline phase but requires three condi-tions. Firstly, the technology forcultivation must be available. Secondly,substitutes (natural or synthetic) for theproduct do not yet exist, and thirdly,prices must remain high. What also needsto be stressed is that the biological attrib-utes of the plant (such as growth rates andwhat part of the plant is harvested) play acrucial role. These are summarized inBoxes 5.1 and 5.2 (see Chapter 5).

Only when prices become extremelyhigh for destructively harvested products –often in response to high demand andgreat scarcity – does cultivation start.Medicinal bark from Warburgia salutarisin Southern Africa (and Zimbabwe, inparticular), is a good example. In theZimbabwe case, local extinction due tooverexploitation resulted in loss of allgrowing stock. Even with local prices forlocal medicinal use reaching US$30–$50per kilogram of bark (dry mass), there wasno chance of cultivation, even if thetechnology had been there. It took speciesreintroduction and work with SAFIRE(Southern Alliance for IndigenousResources), a local NGO, for this speciesto enter the cultivation phase.

While this manual focuses on local-level harvesting and management of wildplant resources, Chapter 3, which dealtwith marketplaces and trade, was areminder that even the remotest areas and

Applied Ethnobotany

Source: modified from Homma, 1992

Figure 8.2 The historical cycle of forest product extraction, with examples from Amazonia (*)and Africa (–) showing part of the plant harvested (l = leaves, fr = fruits)

266

EXPANSION PHASE

Time

Production or extraction

* Timber– Palm basketry– Woodcarving

* Brazil nuts* Rubber trees– Yohimbe bark

* Rosewood* Guarana* Timbo– Ebony (Dalbergia)– Eru (Gnetum)– Pepperbark– Chewing sticks– Prunus africana

* Coca (l)* Guarana (fr)* Quinine (fr)– Marula (fr)– Dacroydes (fr)– Voacanga (fr)

STABILIZATION PHASE DECLINE PHASE CULTIVATION PHASE


plant species can be linked with interna-tional trade networks. For this reason,although the methods for analysingmacro-economic issues are beyond thescope of this manual, it is important,

briefly, to look outward at the widercontext. Unless these wider issues areaddressed, conservation programmes endup winning at the local level but losing ata global one.


Looking outward

267

Harvesting of wild plant resources, oftencarried out by rural people with littleaccess to political or economic power, isdirectly linked to the wider issues: macro-economic factors, per capita consumptionrates, population growth, technologicaldevelopment and living standards. Therecent problems in the Asian economies,where events on New York StockExchange computer screens can affectharvesting of forest products in South-EastAsia, are a good illustration of one of theseglobal linkages.

The point that wild plant resources arethe ‘green social security’, providing abuffer against drought or economic decline,was made at the start of this manual. Onemeasure of the extent to which these plantresources are under increasing pressure inAfrica is the proportion of the total popula-tion in formal employment. In Zimbabwe,this has fallen steadily from 15.6 per centin 1977 to 11.6 per cent in 1993, and thereal average, including wages in the farmingsector, dropped by 35 per cent between1985 and 1993 (Bond, 1996). Many otherAfrican countries are in a similar situation,as are parts of South-East Asia which werehit by the financial crisis in 1998. Thisboosts the number of people relying on wildplant resources, including their sale in local,regional and international markets. It canalso result in weaker tenure, with the influxof ‘outsiders’ who arrive to harvestcommercially valuable resources or to clearfrontier land bordering on conservationareas.

Other measures of environmentalimpact are human population density,population growth rates and consumptionlevels. These are emotive, highly politi-cized issues, but they cannot be ignored(Meffe et al, 1993; Baltz, 1999). Neithercan each of these factors be taken in isola-tion: they have to be considered togetheras important indicators of loss of habitator increase in demand for resources. Over30 years ago, Paul Erlich and otherssummarized this into the simple equation:I = PAT, indicating that environmentalimpact (I) is a function of population size(P), affluence (A) and technology (T)(Erlich and Holdren, 1971). Energyconsumption rates are a good example ofenvironmental impact: a North Americanfamily of 4 people has the same environ-mental impact as 80 Costa Ricans and 280Bangladeshis (Erlich et al, 1995). In orderto present a realistic picture of humanconsumption on the environment, theWWF Living Plant Report (1998)combines population numbers andconsumption rates as an index of thepressures people put on the environment(see Figure 8.3).

Projected world-population growthrates under different fertility rates suggestthat if global fertility rates stabilize at justover 2 children per woman, the world’spopulation will expand from 5.7 billionpeople in 1995 to 9.4 billion in 2050, andwill reach a maximum of almost 11 billionby about 2200. However, even smalldifferences in fertility rate can make a huge


difference to these projections (see Figure8.4a). In Africa, for example, the popula-tion would treble over the next 50 yearsunder the ‘medium fertility scenario’. In

1995, 700 million people lived in Africa;by 2050 there will be just over 2 billion(see Figure 8.4b).

Applied Ethnobotany

268

Source: Loh et al, 1998

Figure 8.3 Consumption pressure: a measure of the burden placed on the environment by people, 1995

1040

1 = worldaverage

consumer in 1995

China

United States

India

Japan

Russian Federation

Indonesia

Brazil

Germany

Pakistan

Republic of Korea

Italy

France

Mexico

Thailand

United Kingdom

Spain

Taiwan

Canada

Philippines

Iran

0.85

7232.74

4430.47

2942.35

2251.53

1570.79

1420.88

1241.52

1100.78

1062.35

971.70

921.58

920.98

891.52

831.43

761.92

733.42

692.35

670.99

650.96

Consumption pressureper person

(pressure units)

Consumption pressureof whole country

(million pressure units)


Ethics are at the core of conservationbehaviour, and ethical change, in this caseethics that constrain individual action forlong-term benefits to a wider group, oftenarises through crisis. One of the require-ments, however, is that claims about crisisare validated and popularized. In societieswith high literacy levels, this can be donethrough posters or books, such as AldoLeopold’s classic, A Sand County Almanac(1949). In other cases, verbal and visualcommunication, through the use of songs,theatre, drama, film or video, is a far morepowerful tool for communicating researchresults.

Religious practice can play a powerfulrole in instilling (or re-instilling) a landethic. In Zimbabwe, for example, projectsthrough ZIRRCON (Zimbabwean

Institute of Religious Research andEcological Conservation), which was initi-ated by anthropologist Enus Daneel, andAZTREC (Association of ZimbabweanTraditional Ecologists) – both workingwith the African Independent Churchesand traditional healers – have developed atheology of the environment with links totree planting. Using practices which havelinks to traditional practices of blessingseeds and the ecological religion of territo-rial cults studied by Schoffeleers (1978)and van Binsbergen (1985) (see Chapter7), congregations are held where peoplecontribute tree seeds, not money, into acollection. These are then blessed andplanted. Ceremonies conducted by tradi-tionalists, separately from the IndependentChurch members, also offer sorghum beer


Looking inward; examining innovative local approaches

269

Source: UNESCO, 1998

Figure 8.4 (a) World population, projected to 2150. (b) Population by region, 1995 andprojection for 2050 under a medium-fertility scenario

1950

(a) World population

Forecasts (billion)

High

Medium

Low

30

25

20

15

10

5

01995 2050 2100 2150

Fertility scenarios

0

(b) Regional population

Billion

1995

2050 (medium-fertility forecast)

Other Asia

Africa

India

China

Latin America*

Europe

North America

0.5 1.0 1.5 2.0

* including Caribbean

2.5


to their ancestors, sprinkling snuff andmixing beer into the water for the seeds(Holt-Biddle, 1994).

Drama has also been used verysuccessfully in Kenya. Actors fromwoodcarving families (or the woodcarversthemselves) tour woodcarving centres,acting out a drama in Kikamba written bylocal Kamba playrights Fidelma Kyalo andVinette Mbaluto (see Figure 8.5). The useof low-cost, but high-quality, videofootage – taken by programme staff butprofessionally edited to scripts they havewritten – has also been a successfulmethod used by the People and PlantsInitiative to transfer knowledge of

resource management methods, problemsand solutions.

In the South Pacific, the Wan SmolbagTheatre (translated as ‘one small bag’),based in Vanuatu, has produced plays,videos and songs related to marine conser-vation issues. In Southern Africa, NicholasEllenbogen’s ‘Theatre for Africa’ has sentan equally powerful message about localcommunities and conservation issues,including a play on the controversial issueof the Southern African position onelephant culling and the ivory trade, actedat the recent CITES (Convention onInternational Trade in Endangered Speciesof Wild Fauna and Flora) congress in

Applied Ethnobotany

270

Figure 8.5 Encouraging conservation through drama, rather than through the barrel of a gun:woodcarvers and their relatives from Wamunyu, Kenya, toured carving centres in Kenya, actingout a drama which communicated research results about the depletion of favoured wild species

and the cultivation of viable alternatives


Harare, Zimbabwe.Creative projects give signs of hope,

even under the bleakest of circumstances(see Figure 8.5). One of the strongest testsof conservation strategies is how resilientthey are in the face of civil conflicts.Recent tests of this stem from conserva-tion areas in Rwanda and Zaire (now theDemocratic Republic of Congo), engulfedby conflict (Hart and Hart, 1997; Fimbeland Fimbel, 1997). These Central Africanexamples highlight the crucial need forappropriately training hand-picked localpeople at various levels (rangers, technicalstaff, research professionals andmanagers) to take responsibility forconservation programmes. Internationalnon-governmental organizations have akey role in this process, and one of these isto support this training process. In bothcases, international funding was disruptedand expatriate staff left or were evacuateddue to conflicts in or around the NyungweForest Conservation Project in Rwandaand four World Heritage Sites in theDemocratic Republic of Congo. Whatpreserved these conservation areas duringthe conflicts were local people connectedto the projects.

Innovative, decentralized approachesto buffer zones and landscapes outsideprotected areas also have a way of catch-ing on and spreading. Two examples areCAMPFIRE (Communal AreasManagement Programme for IndigenousResources) in Zimbabwe (Child, 1996)and Joint Forest Management programmeprojects spread across India and Nepal(Poffenberger et al, 1992a, b; Fischer,1995). In both cases, however, the contextis crucial.

CAMPFIRE has succeeded mainly inareas such as the Zambezi Valley, where

there are big populations of large mammals(from which revenue is derived throughtrophy hunting), low arable potential andlow densities of people. In India, JointForest Management has worked best in‘tribal’ areas such as West Bengal, whichalso have poor lateritic soils and where thedominant tree species, Shorea robusta (sal),has vigorous resprouting ability.

The multiple effect of innovativeapproaches also applies to the marineenvironment. In Vanuatu (Johannes, 1998)and in Fiji (Biodiversity SupportProgramme, 1998), experiments in village-based marine conservation, includingmonitoring of key resources (such astrochus shells and edible mussels), incollaboration with local government andNGOs, have led to the development ofcontrolled harvests of other valued speciesand the closure of some areas to harvest-ing. Although small, and begun inisolation, these programmes have built upexperience and common ground that havebeen more widely applied.

The practical methods covered in thismanual can similarly be used by trainedlocal people. What is essential is to reviewand reflect on these successes and failuresand to work out what made them succeedor fail – whether social (see Chapter 7),ecological (see Chapters 4 to 6) oreconomic factors (see Chapter 3).

The final important ingredient forsuccessful projects is the one mentioned inthe preface to this manual: long-term insti-tutional support, independent of politicalconstraints. Applied ethnobotanicalstudies can be a catalyst in this process – itis clear that useful plants are a key toopening people’s minds to seeing the worldin different ways so that a land ethic maygrow.


271


AI artificial intelligenceANPWS Australian National Parks and Wildlife ServiceAVHRR Advanced Very High Resolution RadiometerAZTREC Association of Zimbabwean Traditional Ecologistsba basal areabd basal diameterBP before presentBSP Biodiversity Support ProgrammeCAMP Conservation Assessment and Management PlanCAMPFIRE Communal Areas Management Programme for Indigenous ResourcesCBNRM community-based natural resource management CFA Communauté Financière AfricaineCIFOR Centre for International Forestry ResearchCITES Convention on International Trade in Endangered Species of Wild

Fauna and FloraCLC Community Land Companies, Indiacm centimetreCNES Centre National d’Études Spatiales, Francedbh diameter at breast heightDFID Department for International Development, UKDTC Development through Conservation projectFAA formalin-acetic acid alcoholFAO Food and Agriculture Organizationgbh girth at breast heightGIS geographic information systemGPS global positioning systemha hectareIAWA International Association of Wood AnatomistsICDP Integrated Conservation and Development ProjectICS index of cultural significanceIGBP International Geosphere-Biosphere ProgrammeIIED International Institute for Environment and Development, UKILDIS International Legume Database and Information ServiceIRDNC integrated rural development and nature conservationISD initial size-class distributionIUCN The World Conservation UnionJCM Joint Forest Management programmekg kilogram

Acronyms and Abbreviations


km kilometreLC local councilLEAP List of East African PlantsMAB Man and the Biosphere programmeMVP minimum viable populationMOU memorandum of understandingNASA National Aeronautics and Space Administration, USNGO non-governmental organizationNLCB National Lottery Charities Board, UKNRM National Resistance Movement, UgandaODI Overseas Development InstitutePAME participatory assessment, monitoring and evaluationPAR participatory action researchPC personal computerPCQ point-centred quarter methodPRA participatory rural appraisalPSP permanent sample plotRBG Royal Botanic Gardens, KewRRA rapid rural appraisalSAFIRE Southern Alliance for Indigenous ResourcesSARARES Southern African RARES databaseSE standard errorSLA specific leaf areaSPOT Système pour l’Observation de la TerreSSD stable-stage distributionUNEP United Nations Environment ProgrammeUNESCO United Nations Educational, Scientific and Cultural OrganizationWCMC World Conservation Monitoring Centre, CambridgeWCU World Conservation UnionWRI World Resources InstituteWWF formerly the World Wide Fund For NatureZIRRCON Zimbabwean Institute of Religious Research and Ecological

Conservation

Acronyms and Abbreviations

273


Further Reading

The following key references are recommended for pursuing the subjects covered inChapters 2–8.

Chapter 2 Local inventories, values and quantities of harvested resourcesBaker and Mutitjulu Community (1992) ‘Comparing two views of the landscape: aboriginal

ecological knowledge and modern scientific knowledge’, Rangeland Journal 14(2): 174–189Bernhard, H R, P Killworth, D Kronenfeld and L Sailer (1984) ‘The problem of informant

accuracy: the validity of retrospective data’, Annual Review of Anthropology 13: 495–517Carlquist, S (1991) ‘Anatomy of vine and liana stems: a review and synthesis’ in F E Putz and H

A Mooney (eds) The Biology of Vines. Cambridge University Press, Cambridge, pp53–71Chambers, R (1992) Rural appraisal: rapid, relaxed and participatory. Discussion Paper 311,

Institute of Development Studies, University of SussexDufour, D L and N I Teufel (1995) ‘Minimum data sets for the description of diet and

measurement of food intake and nutritional status’ in E F Moran (ed) The ComparativeAnalysis of Human Societies. Lynne Reiner Publishers, London, pp98–128

Hoskins, M (1990) ‘The Community Toolbox: the idea, methods and tools for participatoryassessment, monitoring and evaluation in Community Forestry’, FAO Community ForestryField Manual 2. FAO, Rome

IAWA (1957) ‘International glossary of terms used in wood anatomy’, Tropical Woods 107: 1–36IAWA (1981) ‘Standard list of characters suitable for computerized hardwood identification’,

International Association of Wood Anatomists (IAWA) Bulletin 2(2–3): 99–110IAWA (1989) ‘IAWA list of microscopic features for hardwood identification’, International

Association of Wood Anatomists Bulletin 10: 219–332Junikka, L (1994) ‘Survey of English macroscopic bark terminology’, International Association of

Wood Anatomists (IAWA) Bulletin 15(1):1–45Miller, R B (1981) ‘Explanation of coding procedure’, International Association of Wood

Anatomists (IAWA) Bulletin 2(2–3):111–145Mori, S A and G T Prance (1990) ‘Taxonomy, ecology and economic botany of the Brazil nut

(Bertholletia excelsa Humb & Bompl: Lecythidaceae)’, Advances in Economic Botany 8:130–150

Nichols, P (1991) Social Survey Methods: a Field-Guide for Development Workers. DevelopmentGuidelines No 6, Oxfam, Oxford

Phillips, O L (1996) ‘Some quantitative methods for analysing ethnobotanical knowledge’ in M NAlexiades (ed) Selected Guidelines for Ethnobotanical Research: a field manual. New YorkBotanical Garden, New York, pp171–197

Phillips, O and A H Gentry (1993a) ‘The useful plants of Tambopata, Peru: I. Statistical hypothe-ses tests with a new quantitative technique’, Economic Botany 47: 15–32

Phillips, O and A H Gentry (1993b) ‘The useful plants of Tambopata, Peru: II. Additionalhypothesis testing in quantitative ethnobotany’, Economic Botany 47: 33–43

Pratt, B and P Loizo. (1992) Choosing Research Methods: Data Collection for DevelopmentWorkers. Development Guidelines No 7, Oxfam, Oxford

Pretty, J N, I Guijt, I Scoones and J Thompson (1995) Participatory Learning and Action: atrainer’s guide. IIED Participatory Methodology Series, International Institute forEnvironment and Development, London


Toledo, V M, A I Batis, R Bacerra, E Martinez and C H Ramos (1992) ‘Products from thetropical rain forests of Mexico: an ethnoecological approach’ in M Plotkin and L Famolare(eds) Non-Wood Products from Tropical Rainforests. Conservation International,Washington, DC

Turner, N J (1988) ‘The importance of a rose: evaluating the cultural significance of plants inThompson and Lilooet Interior Salish’, American Anthropologist 90: 272–290

Weller, S C and A K Romney (1988) Systematic Data Collection. Qualitative Research Methods,vol 10. Sage Publications, Newbury Park, California

Chapter 3 Settlement, commercialization and changeBye, R A and E Linares (1985) ‘The role of plants found in Mexican markets and their

importance in ethnobotanical studies’, Journal of Ethnobiology 3: 1–13Smith, C A (1985) ‘How to count onions: methods for a regional analysis of marketing’ in S

Plattner (ed) Markets and Marketing. Society for Economic Anthropology, University Press ofAmerica, Lanham, MD, pp49–77

Trager, L (1995) ‘Minimum data sets in the study of exchange and distribution’ in E F Moran(ed) The Comparative Analysis of Human Societies. Lynne Reinner Publishers, London,pp75–96

Chapter 4 Measuring individual plants and assessing harvesting impactsEnright, N J and A D Watson (1992) ‘Population dynamics of the nikau palm, Rhopalostylis

sapida (Wendl et Drude), in a temperate forest remnant near Auckland, New Zealand’, NewZealand Journal of Botany 30: 29–43

Green, D F and E A Johnson (1994) ‘Estimating the mean annual seed production of trees’,Ecology 75: 642–647

Lamont, B B, D C le Maitre, R M Cowling and N J Enright (1991) ‘Canopy seed storage inwoody plants’, The Botanical Review 57: 277–317

Peters, C M (1996) ‘Beyond nomenclature and use: a review of ecological methods forethnobotanists’ in M N Alexiades (ed) Selected Guidelines for Ethnobotanical Research: afield manual. New York Botanical Garden, New York, pp 241–276

Ruiters, C, B McKenzie and L M Raitt (1993) ‘Life-history studies of the perennial geophyteHaemanthus pubescens (Amaryllidaceae) in lowland coastal fynbos, South Africa’,International Journal of Plant Science 154: 441–449

Rutherford, M C (1979) ‘Plant-based techniques for determining available browse and browseutilisation: a review’, The Botanical Review 45: 203–228

Chapter 5 Opportunities and constraints on sustainable harvest: plantpopulations

Alder, D and T J Synnott (1992) Permanent Sample Plot Techniques for Mixed Tropical Forest.Oxford Forestry Institute, University of Oxford, Oxford

Alvarez-Buylla, E R (1994) ‘Density dependence and patch dynamics in tropical rain forests:matrix models and applications to a tree species’, The American Naturalist 143(1):155–191

Bell, A D (1998) Plant Form: an illustrated guide to flowering plant morphology. OxfordUniversity Press, Oxford

Bernal, R (1998) ‘Demography of the vegetable ivory palm Phytelephus seemannii in Colombia,and the impact of seed harvesting’, Journal of Applied Ecology 35: 64–74

Caswell, H (1989) Matrix Population Models. Sinauer Associates, Sunderland, MassachusettsDesmet, P G, C M Shackleton and E R Robinson (1996) ‘The population dynamics and life-

history attributes of a Pterocarpus angolensis DC. Population in the northern province,South Africa’, South African Journal of Botany 62(3): 160–166

Further Reading

275


Enright, N J and Watson, A D (1992) ‘Population dynamics of the nikau palm, Rhopalostylissapida (Wendl et Drude) in a temperate forest remnant near Auckland, New Zealand’, NewZealand Journal of Botany 30: 29–43

Midgley, J J, M C Cameron and W J Bond (1995) ‘Gap characteristics and replacement patternsin the Knysna forest, South Africa’, Journal of Vegetation Science 6: 29–35

Nantel, P, D Gagnon and A Nault (1996) ‘Population viability analysis of American ginseng andwild leek harvested in stochastic environments’, Conservation Biology 10(2): 608–621

Peters, C M (1994) Sustainable Harvest of Non-Timber Forest Plant Resources in Tropical MoistForest: an ecological primer. Biodiversity Support Program, Washington, DC. Available fromBiodiversity Support Program, 1250 24th Street, NW, Washington, DC, 20037, US

Peters, C M (1996a) ‘Beyond nomenclature and use: a review of ecological methods forethnobotanists’ in M N Alexiades (ed) Selected Guidelines for Ethnobotanical Research: afield manual. New York Botanical Garden, New York, pp241–276

Sheil, D (1995) ‘A critique of permanent plot methods and analysis with examples from Budongoforest, Uganda’, Forest Ecology and Management 77: 11–34

Chapter 6 Landscapes and ecosystems: patterns, processes and plant useBond, W J and B W van Wilgen (1996) Fire and Plants. Chapman & Hall, London Clark, D B (1996) ‘Abolishing virginity’, Journal of Tropical Ecology 12: 735–739Fairhead, J and M Leach (1996) Misreading the African Landscape: society and ecology in a

forest-savanna mosaic. Cambridge University Press, CambridgeHubbell, S P and R B Foster (1986) ‘Commonness and rarity in a neotropical forest: implications

for tropical tree conservation’ in M E Soulé (ed) Conservation biology: the science of scarcityand diversity. Sinauer Associates, Sunderland, Massachusetts, pp205–231

IUCN (1994) IUCN Red List Categories. IUCN Species Survival Commission, Gland,Switzerland

Lewis, H T (1989) ‘Ecological and technological knowledge of fire: Aborigines versus parkrangers in Northern Australia’, American Anthropologist 91: 940–961

Mather, R, M de Boer, M Gurung and N Roche (1998) Aerial Photographs and ‘Photo-Maps’ forCommunity Forestry. Overseas Development Institute (ODI), London. Rural DevelopmentForestry Network 2e: 13–22

Miller, K R (1996) Balancing the Scales: guidelines for increasing biodiversity’s chances throughbioregional management. World Resources Institute, Washington, DC

Noss, R F (1990) ‘Indicators for monitoring biodiversity: a hierarchical approach’, ConservationBiology 4: 355–364

Pickett, S T A and P S White (1985) The Ecology of Natural Disturbance and Patch Dynamics.Academic Press, New York

Poole, P (1995) Indigenous Peoples, Mapping and Biodiversity Conservation: an analysis ofcurrent activities and opportunities for applying geomatics technologies. Biodiversity SupportProgram (BSP), Peoples and Forest Program Discussion Paper series. BSP/WWF/The NatureConservancy, Washington, DC

Rabinowitz, D, S Cairns and T Dillon (1986) ‘Seven forms of rarity and their frequency in theflora of the British Isles’ in M E Soulé (ed) Conservation Biology: the science of scarcity anddiversity. Sinauer Associates, Sunderland, Massachusetts, pp182–204

Rundstrom, R A (1990) ‘A cultural interpretation of Inuit map accuracy’, The GeographicalReview 80: 155–168

Sharpe, B (1998) ‘Forest people and conservation initiatives: the cultural context of rainforestconservation in West Africa’ in B Goldsmith (ed) Tropical rain forest: a wider perspective.Chapman & Hall, London

Wagener, W W (1961) ‘Past fire incidence in Sierra Nevada forests’, Journal of Forestry 59:739–748

Whitmore, T C (1988) ‘The influence of tree population dynamics on forest species composition’in A J Davy, M J Hutchings and A R Watkinson (eds) Plant Population Biology. Blackwell,Oxford, pp271–291

Applied Ethnobotany

276


Chapter 7 Conservation Behaviour, Boundaries and Beliefs Bohannan, P (1963) ‘“Land”, “tenure” and land-tenure’ in D Biebuyck (ed) African Agrarian

Systems. Oxford University Press, Oxford, pp101–111 Bromley, DW and M M Cernea (1989) The Management of Common Property Natural

Resources: some conceptual and operational fallacies. World Bank Discussion Paper 57.World Bank, Washington, DC

Bruce, J and L Fortmann (1989) Agroforestry: tenure and incentives. Land Tenure Centre Report135, Land Tenure Centre, Madison, Wisconsin

Casimir, M J (1992) ‘The determinants of rights to pasture: territorial organisation and ecologicalconstraints’ in M J Casimir and A Rai (eds) Mobility and Territoriality: social and spatialboundaries among foragers, fishers, pastoralists and peripatetics. Berg, New York,pp153–203

Colfer, C J Pierce (1995) Who Counts Most in Sustainable Forest Mmanagement? Working PaperNo 7, Centre for International Forestry Research (CIFOR), Bogor, Indonesia

Fischer, R (1995) Collaborative Management of Forests for Conservation and Development.Issues in Forest Conservation. IUCN, Gland, Switzerland

Ostrom, E (1990) Governing the Commons: the evolution of institutions for collective action.Cambridge University Press, Cambridge

Poole, P (1995) Indigenous Peoples, Mapping and Biodiversity Conservation: an analysis ofcurrent activities and opportunities for applying geomatics technologies. Biodiversity SupportProgram, Washington, DC

Richards, P (1996) ‘Forest indigenous peoples: concept, critique and cases’, Proceedings of theRoyal Society of Edinburgh 104B: 349–365

Shipton, P (1994) ‘Land and culture in tropical Africa: soils, symbols and the metaphysics of themundane’, Annual Review of Anthropology 23: 347–377

Wade, R (1987) ‘The management of common property resources: collective action as an alterna-tive to privatisation or state regulation’, Cambridge Journal of Economics 11: 95–106

Chapter 8 Striving for balance: looking outward and inwardBaltz, M E (1999) ‘Overconsumption of resources in industrial countries: the other missing

agenda’, Conservation Biology 13: 213–215Homma, A K G (1992) ‘The dynamics of extraction in Amazonia: a historical perspective’,

Advances in Economic Botany 9: 23–31Meffe, G K, A H Erlich and D Ehrenfeld (1993) ‘Human population control: the missing agenda’,

Conservation Biology 7:1–3

Further Reading

277


References

Acevedo-Rodriguez, P (1990) ‘The occurrence of piscicides and stupefactants in the plantkingdom’, Advances in Economic Botany 8: 1–23

Adu-Tutu, M, Y Afful, K Asante-Appiah, D Liebermann, J B Hall and M Elvin-Lewis (1979)‘Chewing stick usage in Ghana’, Economic Botany 33(3): 320–328

Agnew, A D Q (1985) ‘Cyclic sequences of vegetation in the plant communities of the Aberdaremountains, Kenya’, Journal of the East Africa Natural History Society and National Museum75: 1–12

Ake-Assi, L (1988) ‘Plantes medicinales: quelques Legumineuses utilisées dans la medecine detradition Africaine en Côte d’Ivoire’, Monogr. Syst. Bot Gard. 25: 309–313

Alder, D and T J Synnott (1992) Permanent Sample Plot Techniques for Mixed Tropical Forest.Tropical Forestry Papers No 25. Oxford Forestry Institute, University of Oxford, Oxford

Alvarez-Buylla, E R (1994) ‘Density dependence and patch dynamics in tropical rain forests:matrix models and applications to a tree species’, The American Naturalist 143(1): 155–191

Anderson, J C (1984) ‘The Political and Economic Basis of Kuku-Yalanji Social History’.Unpublished PhD thesis, Department of Archaeology and Sociology. University ofQueensland, Queensland

Anderson, A B and D A Posey (1989) ‘Management of a tropical shrub savanna by the GorotireKayapo of Brazil’, Advances in Economic Botany 7: 159–173

Ankei, Y (1985) ‘A comparative study of the barter markets of the upper Zaire river’, AfricaStudy Monographs Supplementary Issue 4: 89–101

Ankei, Y (1981) ‘Agricultural livelihood and economic activities of the Songola–slash and burnagriculturalists in the tropical rainforest, central Africa’, Kikan Jinruigku 12: 96–178

Antos, J A and G A Allen (1988) ‘Relative reproductive effort in males and females of thedioecious shrub Oemleria cerasiformis (Rosaceae)’, Oecologia 76: 111–118

Arnold, J E M and W C Stewart (1991) Common property resource management in India.Tropical Forestry Papers No 24. Oxford Forestry Institute, University of Oxford, Oxford

Ash, J (1986) ‘Demography and production of Leptopteris wilkesiana (Osmundaceae), a tropicaltree-fern from Fiji’, Australian Journal of Botany 34: 207–215

Ash, J (1987) ‘Demography of Cyathea hornei (Cyatheaceae), a tropical tree-fern in Fiji’,Australian Journal of Botany 35: 331–342

Austin, D D and P J Urness (1980) ‘Response of Curlleaf Mountain Mahogany to pruningtreatments in northern Utah’, Journal of Range Management 33: 275–277

Baker and Mutitjulu Community (1992) ‘Comparing two views of the landscape: aboriginalecological knowledge and modern scientific knowledge’, Rangeland Journal 14 (2): 174–189

Balee, W and D C Daly (1990) ‘Resin classification by the Ka’apor Indians’, New Directions inthe Study of Plants and People. Advances in Economic Botany 8: 24–34 (New York BotanicalGarden)

Balee, W (1989) ‘The culture of Amazonian forests’, Advances in Economic Botany 7: 1–21Baltz, M E (1999) ‘Overconsumption of resources in industrial countries: the other missing

agenda’, Conservation Biology 13: 213–215Barnes, R F W (1990). ‘Deforestation trends in tropical Africa’, African Journal of Ecology 28:

161–173Beentje, H (1994) Kenya Shrubs, Trees and Lianas. National Museum of Kenya, NairobiBell, A D (1998) Plant Form: an illustrated guide to flowering plant morphology. Oxford

University Press, OxfordBell, R H V and T O McShane (1984) ‘Landscape classification’ in R H V Bell, T O McShane

and E Caluzi (eds) Conservation and Wildlife Management in Africa, US Peace Corps,Washington, DC, pp95–106


Bennett, B C (1992) ‘Plants and people of the Amazonian rainforests’, BioScience 42: 599–607Berglund, A-I (1976) Zulu Thought Patterns and Symbolism. David Phillip, Cape TownBerlin, B (1992) Ethnobiological Classification: principles of categorization of plants and animals

in traditional societies. Princeton University Press, Princeton, New JerseyBernal, R (1998) ‘Demography of the vegetable ivory palm Phytelephas seemannii in Colombia,

and the impact of seed harvesting’, Journal of Applied Ecology 35: 64–74Bernhard, H R, P Killworth, D Kronenfeld and L Sailer (1984) ‘The problem of informant

accuracy: the validity of retrospective data’, Annual Review of Anthropology 13: 495–517Berry, M I (1996) ‘The use of chewing sticks, khat and Indha kuul amongst the Somali

population of Liverpool’, poster abstract. Society for Economic Botany Meeting, London, 2–3July

Biodiversity Support Program (1998) ‘Fiji Case Story: accounting for success in community-basedmonitoring’, Lessons from the Field 1: 6–7

Bird, N M and G Shepherd (1989) Charcoal in Somalia: a woodfuel inventory in the bay regionof Somalia. Somali Forestry Series No 2. NRA/ODA, Somalia

Bohannan, P (1963) ‘“Land”, “tenure” and land-tenure’, in D Biebuyck (ed) African AgrarianSystems. Oxford University Press, Oxford, pp101–111

Bond, I (1996) ‘Employment, population growth and the demand for land in communal areas’ inB M Campbell (ed) The Miombo in Transition: woodlands and welfare in Africa. Centre forInternational Forestry Research (CIFOR), Bogor, p88

Bond, W J and B W van Wilgen (1996) Fire and Plants. Chapman & Hall, LondonBoom, B M (1989) ‘Use of plant resources by the Chacabo’, Advances in Economic Botany 7:

78–96Boom, B M (1990) ‘Giving native people a share of the profits’, Garden 14: 28–31Borrini-Fyerabend, G (1996) Collaborative Management of Protected Areas: Tailoring the

Approach to the Context. Issues in Social Policy. IUCN, Gland, SwitzerlandBourdieu, P (1973) ‘The Berber house’ in M Douglas (ed) Rules and Meanings. Penguin,

Harmondsworth, pp98–110Bourdieu, P (1977) Outline of a Theory of Practice. Cambridge University Press, CambridgeBridson, D and L Forman (1992) The Herbarium Handbook. Royal Botanic Gardens, KewBromley, D W and Cernea, M M (1989) The Management of Common Property Natural

Resources: some conceptual and operational fallacies. World Bank Discussion Paper 57.World Bank, Washington, DC

Bromley, R J (1971) ‘Markets in the developing countries: a review’, Geography 56: 124–132Bromley, R J (1974) ‘The organization of Quito’s urban markets: towards a reinterpretation of

periodic central places’, Institute of British Geographers Transactions 62: 45–70Bruce, J and L Fortmann (1989) Agroforestry: tenure and incentives. Land Tenure Centre Report

135. Land Tenure Centre, Madison, WisconsinBruce, J, L Fortmann and C Nhira (1993) ‘Tenures in transition, tenures in conflict: examples

from the Zimbabwe social forest’, Rural Sociology 8:626–642Budack, K F R (1983) ‘A harvesting people along the south Atlantic coast’, South African Journal

of Ethnology 6: 1–7Bulatao, R A, E Bos, P W Stephens, and M T Vu (1990) World Population Projections

1989–1990 edition; short- and long-term estimates. Johns Hopkins University Press, LondonBuwai, M and M J Trlica (1977) ‘Multiple defoliation effects on herbage yield, vigour and total

non-structural carbohydrates of five range species’, Journal of Range Management 30:164–171

Bye, R A and E Linares (1985) ‘The role of plants found in Mexican markets and theirimportance in ethnobotanical studies’, Journal of Ethnobiology 3: 1–13

Campbell, DG (1989) ‘Quantitative inventory of tropical forests’ in D G Campbell and H DHammond (eds) Floristic Inventory of Tropical Countries. New York Botanical Garden, NewYork, pp524–533

Carlquist, S (1991) ‘Anatomy of vine and liana stems: a review and synthesis’ in F E Putz and HA Mooney (eds) The Biology of Vines. Cambridge University Press, Cambridge, pp53–71

Carpenter, S G and G Cottam (1982) ‘Growth and reproduction of American ginseng (Panaxquinquefolius) in Wisconsin, USA’, Canadian Journal of Botany 60: 2692–2696

References

279


Casas, A and J Caballero (1996) ‘Traditional management and morphological variation inLeucaena esculenta (Fabaceae: Mimosoideae) in the Mixtec region of Guerrero, Mexico’,Economic Botany 50: 167–181

Cashdan, E (1983) ‘Territoriality among human foragers: ecological models and an application tofour Bushman groups’, Current Anthropology 24: 47–66

Casimir, M J (1992) ‘The determinants of rights to pasture: territorial organisation and ecologicalconstraints’ in M J Casimir and A Rai (eds) Mobility and Territoriality: social and spatialboundaries among foragers, fishers, pastoralists and peripatetics. Berg, New York

Caswell, H (1989) Matrix Population Models. Sinauer Associates, Sunderland, MassachusettsChambers, R (1992) Rural Appraisal: rapid, relaxed and participatory. Discussion Paper 311.

Institute of Development Studies, University of Sussex, SussexChapman, M (1987) ‘Traditional political structure and conservation in Oceania’, Ambio 16:

201–205Chazdon, R L (1992) ‘Patterns of growth and reproduction of Geonoma congesta, a clustered

understorey palm’, Biotropica 24: 43–51Child, B (1996) ‘The practice and principles of community-based wildlife management in

Zimbabwe: the CAMPFIRE programme’, Biodiversity and Conservation 5: 369–398Christaller, W (1966) Central Places in Southern Germany. Prentice-Hall, Englewood CliffsClark, D B (1996) ‘Abolishing virginity’, Journal of Tropical Ecology 12: 735–739Clark, R L (1990) ‘Ecological history for environmental management’, Proceedings of the

Ecological Society of Australia 16:1–21Cole, M M and R C Brown (1976) ‘The vegetation of the Ghanzi area of western Botswana’,

Journal of Biogeography 3: 169–196Coley, P D, J P Bryant, and F S Chapin III (1985) ‘Resource availability and anti-herbivore

defence’, Science 230: 859–899Colfer, C J Pierce (1995) Who Counts Most in Sustainable Forest Management? Working Paper

No 7. Centre for International Forestry Research (CIFOR), Bogor, IndonesiaConelly, W T (1985) ‘Copal and rattan collecting in the Philippines’, Economic Botany 39: 39–46Connah, G and A Jones (1983) ‘Photographing Australian prehistoric sites from the air’ in G

Connah (ed) Australian Field Archaeology: a guide to techniques. Australian Institute ofAboriginal Studies, Canberra, pp73–81

Cooke, H B S (1964) ‘The Pleistocene environment in southern Africa’ in D H S Davis, B deMeillon and J S Harington (eds) Ecological Studies in Southern Africa. Junk, The Hague,pp1–23

Coppen, J J W (1995) Flavours and Fragrances of Plant Origin. Non-Wood Forest Products 1.Food and Agriculture Organization (FAO), Rome

Coughenour, M B, J E Ellis, D M Swift, D L Coppock, K Galvin, J T McCabe and T C Hart(1985) ‘Energy extraction and use in a nomadic pastoral ecosystem’, Science 230: 619–625

Crouse, D T, L B Crowder and H Caswell (1987) ‘A stage-based population model forloggerhead sea turtles and implications for conservation’, Ecology 68: 1412–1423

Crumley, C L and W H Marquardt (eds) (1987) Regional Dynamics: Burgundian landscapes inhistorical perspective. Academic Press, San Diego

Cunningham, A B (1985) ‘The Resource Value of Indigenous Plants to Rural People in a lowAgricultural Potential Area’. Unpublished PhD thesis. University of Cape Town, Cape Town

Cunningham, A B (1988a) ‘Collection of wild foods in Tembe-Thonga society: a guide to IronAge gathering activities?’ Annals of the Natal Museum 29: 433–466

Cunningham, A B (1988b) Botswana Basketry Resources: research and management of plantresources supporting the Ngamiland basketry trade, 1982–1988. Botswanacraft MarketingCompany (Pty) Ltd, Gaborone

Cunningham A B (1988c) ‘Leaf production and utilization in Hyphaene coriacea: managementguidelines for commercial harvesting’, South African Journal of Botany 54: 189–195

Cunningham, A B (1988d) An Investigation of the Herbal Medicine Trade in Natal/KwaZulu.Investigational Report No 29. Institute of Natural Resources, University of Natal,Pietermaritzburg, South Africa

Cunningham, A B (1989) ‘Indigenous plant use: balancing human needs and resources’ in B JHuntley (ed) Biotic Diversity in Southern Africa: concepts and conservation. OxfordUniversity Press, Cape Town, pp93–106

Applied Ethnobotany

280


Cunningham, A B (1990a) ‘Income, sap yield and the effects of tapping on palms in south-easternAfrica’, South African Journal of Botany 56: 137–144

Cunningham, A B (1990b) ‘The regional distribution, marketing and economic value of the palmwine trade in the Ingwavuma district, Natal, South Africa’, South African Journal of Botany56: 191–198

Cunningham, A B (1991) ‘Development of a conservation policy on commercially exploitedmedicinal plants: a case study from southern Africa’ in O Akerele, V Heywood, H Synge (eds)Conservation of Medicinal Plants. Cambridge University Press, Cambridge, pp337–358

Cunningham, A B (1993) Ethics, Ethnobiological Research and Biodiversity. WWF-International,Gland, Switzerland

Cunningham, A B (1996a) ‘Professional ethics and ethnobotanical research’ in M N Alexiadesand J W Sheldon (eds) Selected Guidelines for Ethnobotanical Research: a field manual.Advances in Economic Botany, New York Botanical Garden, New York, pp19–51

Cunningham, A B (1996b) People, Park and Plant Use: recommendations for multiple-use zonesand development alternatives around Bwindi-Impenetrable National Park, Uganda. Peopleand Plants Working Paper 4. UNESCO, Paris

Cunningham, A B, E Ayuk, S Franzel, B Duguma and C Asanga (in press) An economic valuationof medicinal tree cultivation: Prunus africana in Cameroon. Working Paper,WWF/UNESCO/Kew People and Plants Initiative. UNESCO, Paris

Cunningham, A B and B R Gwala (1986) ‘Plant species and building methods used in Tembe-Thonga hut construction’, Annals of the Natal Museum 27: 491–511

Cunningham, A B and Milton, S J (1987) ‘Effects of the basket weaving industry on the mokolapalm (Hyphaene petersiana) and on dye plants in NW Botswana’, Economic Botany 42:386–402

Dallmeier, F (1992) Long-Term Monitoring of Biological Diversity in Tropical Forest Areas:methods for establishment and inventory of permanent plots. MAB Digest 11. UNESCO,Paris

Davis, S D, S J M Droop, P Gregerson, L Henson, C J Leon, J L Villa-Lobos, H Synge and JZantovska (1986) Plants in Danger: what do we know? IUCN, Gland, Switzerland

Davison, P (1991) ‘Material, Context and Meaning: a critical investigation of museum practice,with particular reference to the South African Museum’. Unpublished PhD thesis, Universityof Cape Town, Cape Town

Dawkins, H C (1958) The Management of Natural Tropical High Forest with Special Referenceto Uganda. Institute Paper, Imperial Forestry Institute, University of Oxford, No 34.University of Oxford, Oxford

Dentlinger, U (1977) The !Nara plant in the Topnaar Hottentot culture of Namibia. MungerAfricana Library Notes, No 38. California Institute of Technology, Pasadena

Desmet, P G, C M Shackleton and E R Robinson (1996) ‘The population dynamics and life-history attributes of a Pterocarpus angolensis DC population in the northern province, SouthAfrica’, South African Journal of Botany 62(3): 160–166

Digernes, T H (1979) ‘Fuelwood crisis causing unfortunate land-use – and the other way around’,Norsk. geogr. Tidsskft 33: 23–32

Dinerstein, E, D M Olson, J Graham, A L Webster, S A Primm, M P Bookbinder and G Ledec(1995) A Conservation Assessment of the Terrestrial Ecoregions of Latin America and theCaribbean. The World Bank, Washington, DC

Donaldson, J (in press) Changes in Abundance of South African Cycads during the 20th Century:preliminary data from the study of matched photographs. Proceedings of the 4th InternationalConference on Cycad Biology, Panzihua, People’s Republic of China

Douglas, M (1966) Purity and Danger: an analysis of concepts of pollution and taboo. Routledgeand Kegan Paul, London

Douglas, M (1970) Natural Symbols. Pantheon Books, New YorkDouglass, A E (1914) ‘A method of estimating rainfall by the growth of trees’ in E Huntington

(ed) The Climatic Factor as Illustrated in Arid America. Carnegie Institute of Washington,Publ, 192, Washington, DC, pp100–122

Douglass, A E (1936) Climatic Cycles and Tree Growth. Carnegie Institute of Washington, VolIII, Washington, DC

References

281


Dransfield, J (1981) ‘The biology of Asiatic rattans in relation to the rattan trade andconservation’ in H Synge (ed) The Biological Aspects of Rare Plant Conservation. John Wiley& Sons, Chichester, pp 179–186

Dufour, D L and N I Teufel (1995) ‘Minimum data sets for the description of diet andmeasurement of food intake and nutritional status’ in E F Moran (ed) The ComparativeAnalysis of Human Societies. Lynne Reiner Publishers, London, pp98–128

Durr, P C, L Richmond and C Eagar (1988) Site Classification and Field Measurements: MethodsManual. Science Division, Great Smoky Mountains National Park, Gatlinburg, Tennessee

Dyson-Hudson, R and Dyson-Hudson, N (1980) ‘Nomadic pastoralism’, Annual Review ofAnthropology 9: 15–61

Dyson-Hudson, R and E A Smith (1978) ‘Human territoriality: an ecological assessment’,American Anthropologist 80: 21–41

Enright, N J and Watson, A D (1992) ‘Population dynamics of the nikau palm, Rhopalostylissapida (Wendl et Drude) in a temperate forest remnant near Auckland, New Zealand’, NewZealand Journal of Botany 30: 29–43

Erkkila, A and H Siiskonen (1992) Forestry in Namibia 1850–1990. Silva Carelica 20. Faculty ofForestry, University of Joensuu

Erlich, P R and P Holdren (1971) ‘Impact of population growth’, Science 171: 1212–1217Erlich, P R, A H Erlich and G C Daily (1995) The Stork and the Plow: the equity answer to the

human dilemma. G P Putnam and Sons, New YorkEverard, D A (1992) ‘On the feasibility of developing a functional classification as a decision

support system for the management of indigenous forests in Natal’. Unpublished report.FOR-DEA 458, CSIR, Pretoria

Fairhead, J and M Leach (1996) Misreading the African Landscape: society and ecology in aforest-savanna mosaic. Cambridge University Press, Cambridge

Fimbel, C and R Fimbel (1997) ‘Conservation and civil strife. Rwanda: the role of local participa-tion’, Conservation Biology 11: 309–310

Fischer, J B (1981) ‘Wound healing by exposed secondary xylem in Adansonia (Bombacaceae)’,International Association of Wood Anatomists Bulletin 2(4): 193–199

Fischer, R J (1995) Collaborative Management of Forests for Conservation and Development.Issues in Forest Conservation. IUCN, Gland, Switzerland

Fjeldså, J, D Erlich, E Lambin and E Prins (1997) ‘Are biodiversity “hot-spots” correlated withcurrent ecocolimatic stability? A pilot study using the NOAA-AVHRR remote sensing data’,Biodiversity and Conservation 6: 401–422

Flannery, T F (1994) The Future Eaters: an ecological history of Australasian lands and people.Reed Books, Chatswood

Fleuret, A (1980) ‘Non-food uses of plants in Usambara’, Economic Botany 34: 320–333Fortmann, L and J Bruce (1988) Whose Trees? Proprietary dimensions of Forestry. Westview

Press, Boulder, ColoradoFortmann, L and N Nabane (1992) The fruits of their labour: gender, property and trees in

Mhondoro District. CASS Occasional Paper Series 6/1992. Centre for Applied Social Sciences,University of Zimbabwe, Harare

Fuegelsang, A (1982) About Understanding – ideas and observations on cross-culturalcommunication. Media Books, Inc, New York

Freudenberger, K S (1994) Tree and Land Tenure: Rapid Appraisal Tools. Community ForestryField Manual 4. Food and Agriculture Organization (FAO), Rome

Geldenhuys, C J (1992) The use of diameter distributions in sustained use management of forests:examples from southern Africa. Paper presented at the SAREC/Zimbabwe ForestryCommission Symposium: Ecology and Management of Indigenous Forests in Southern Africa,Victoria Falls, July 1992

Geldenhuys, C J (1994) ‘Bergwind fires and the location of forest patches in the southern Capelandscape, South Africa’, Journal of Biogeography 21: 49–62

Gentry, A H (1993) A Field Guide to the Families and Genera of Woody Plants of NorthwestSouth America (Colombia, Ecuador, Peru) with Supplementary Notes on Herbaceous Taxa.Conservation International, Washington, DC

Gerstner, J (1938) ‘A preliminary checklist of Zulu names of plants’, Bantu Studies 13: 49–64;131–149; 307–326

Applied Ethnobotany

282


Ghimire, K B and M P Pimbert (eds) (1997) Social Change and Conservation: environmentalpolitics and impacts of national parks and protected areas. Earthscan Ltd, London

Giddens, A (1984) The Constitution of Society: outline of a theory of structuration. Polity Press,Cambridge

Glowka, L, F Burhenne-Guilmin and H Synge (1994) A Guide to the Convention on BiologicalDiversity. Environmental Policy Paper 30. World Conservation Union, Cambridge

Gluckman, M (1965) The Ideas in Barotse Jurisprudence. Yale University Press, New HavenGodoy, R and K S Bawa (1993) ‘The economic value and sustainable harvest of plants and

animals from the tropical forest: assumptions, hypotheses and methods’, Economic Botany47(3): 215–219

Gomez-Pompa, A and S del Amo (eds) (1985) Investigaciones sobre la Regeneracion de SelvasAtlas en Veracruz. Mexico. II. Editorial Alhambra Mexicana, SA de CV, Mexico

Good, C M (1975) ‘Periodic markets and travelling traders in Uganda’, Geographical Review 65:49–72

Gourlay, I D (1995) ‘Growth ring characteristics of some African Acacia species’, Journal ofTropical Ecology 11: 121–140

Gourlay, I D and R D Barnes (1994) ‘Seasonal growth zones in the wood of Acacia karrooHayne: their definition and implications’, Commonwealth Forestry Review 73: 121–127

Green, D F and E A Johnson (1994) ‘Estimating the mean annual seed production of trees’,Ecology 75: 642–647

Gregory, M (1980) ‘Wood identification: an annotated bibliography’, International Association ofWood Anatomists Bulletin 1(1–2): 3–41

Grosenbaugh, L R (1991) ‘Tilted tree hypsometry’, Forest Science 37(6): 1581–1590Grossman, D, A A Ferrar and P C du Plessis (1992) ‘Socio-economic factors influencing conserva-

tion in southern Africa’, TRAFFIC Bulletin 13: 29–31Grundy, I M (1995) ‘Regeneration and Management of Brachystegia spiciformis Benth and

Julbernardia globiflora (Benth) Troupin in Miombo Woodland, Zimbabwe’. Unpublished PhDthesis. University of Oxford, Oxford

Haley, D and M K Luckert (1990) Forest Tenures in Canada: a framework for policy analysis.Information Report E-X–43. Forestry Canada, Ottawa

Hall, M (1987). The Changing Past: farmers, kings and traders in southern Africa 200–1860.David Philip, Cape Town

Hall, J B (1983) ‘Positive management for strict nature reserves: reviewing effectiveness’, ForestEcology and Management 7: 57–66

Hall, R C (1944). ‘A vernier tree-growth band’, Journal of Forestry 42: 742–743Hallé, F, R A A Oldeman and P B Tomlinson (1978) Tropical Trees and Forests: an architectural

analysis. Springer Verlag, BerlinHallé, F (1985) ‘Modular growth in seed plants’, Phil. Trans. R. Soc. London B(313): 77–87Hall-Martin, A (1993) ‘Liwonde: Malawi’s National Park of the future’, Africa Environment and

Wildlife 1: 71–79Hames, R (1987) ‘Game conservation or efficient hunting?’ in B McCay and J M Acheson (eds)

The question of the commons: the culture and ecology of communal resources. The Universityof Arizona Press, Tucson, Arizona

Hamilton, A C and R Bensted-Smith (1989) Forest Conservation in the East UsambaraMountains, Tanzania. IUCN Tropical Forest Programme, Gland, Switzerland

Hamilton, A C (1981) ‘The quaternary history of African forests: its relevance to conservation’,African Journal of Ecology 19: 1–6

Hammond, J E (1933) Winjan’s People. Imperial Printing, PerthHammond-Tooke, W D (1975) ‘The symbolic structure of Cape Nguni cosmology’ in M G

Whisson and M West (eds) Religion and Social Change in southern Africa. David Philip, CapeTown

Hammond-Tooke, W D (1981) Boundaries and Belief: the structure of a Sotho worldview.Witwatersrand University Press, Johannesburg

Hansen, A, J S Pate and A P Hansen (1991) ‘Growth and reproductive performance of a seederand resprouter species of Bossiaea as a function of plant age after fire’, Annals of Botany 67:497–509

References

283


Hardin, G (1968) ‘The tragedy of the commons’, Science 162: 1244Harper, J L (1977) Population Biology of Plants. Academic Press, New YorkHarris, D R (1989) ‘An evolutionary continuum of people-plant interaction’ in D R Harris and G

C Hillman (eds) Foraging and farming: the evolution of plant exploitation. One WorldArchaeology Series. Unwin Hyman, London, pp11–26

Hart, T and J Hart (1997) ‘Conservation and civil strife. Zaire: new models for an emergingstate’, Conservation Biology 11: 308–309

Hart, T B, J A Hart, R Dechamps, M Fournier and M Ataholo (1994) ‘Changes in forestcomposition over the last 4000 years in the Ituri basin, Zaire’ in L J G van der Maesen, X Mvan der Burgt and J M van Medenbach de Rooy (eds) The Biodiversity of African Plants.Kluwer Academic Publishers, Dordrecht, pp545–559

Hartshorn, G S (1989) ‘Application of gap theory to tropical forest management: natural regener-ation on strip clear-cuts in the Peruvian Amazon’, Ecology 70: 567–569

Hecht, S B and D A Posey (1989) ‘Preliminary results in soil management techniques of theKayapo Indians’, Advances in Economic Botany 7: 174–188

Hersh-Martinez, P (1995) ‘Commercialisation of wild medicinal plants from south-west Puebla,Mexico’, Economic Botany 49: 197–206

Hill, P and R H T Smith (1972) ‘The spatial and temporal synchronisation of periodic markets:evidence from four emirates in northern Nigeria’, Economic Geography 47: 345–355

Hilton-Taylor, C (1996) Red Data List of Southern African Plants. Strelitzia 4, NationalBotanical Institute, Pretoria

Hitchcock, R K (1978) Kalahari Cattle Posts: a regional study of hunter-gatherers, pastoralistsand agriculturalists in the western sandveld region, Central District, Botswana. Ministry ofLocal Government and Lands, Gaborone

Hobbs, R J and H A Mooney (1991) ‘Effects of rainfall variability and gopher disturbance ongrassland dynamics in northern California’, Ecology 72: 59–68

Hodder, B W and R Lee (1974) Economic Geography. Methuen and Co Ltd, LondonHolmgren, P K et al (1990) Index Herbariorum: Part 1. The Herbaria of the World. 8th edition.

Regnum Vegetabile vol 120. New York Botanical Garden, New YorkHolt-Biddle, D (1994) ‘Hondo yemiti: Zimbabwe’s “war of the trees”’, Africa Environment and

Wildlife 2: 59–63Homma, A K G (1992) ‘The dynamics of extraction in Amazonia: a historical perspective’,

Advances in Economic Botany 9: 23–31Hoskins, M (1990) The Community Toolbox: the idea, methods and tools for participatory

assessment, monitoring and evaluation in Community Forestry. FAO Community ForestryField Manual 2, Food and Agriculture Organization (FAO), Rome

Hot Springs Working Group (1995) Local-level economic valuation of savanna woodlandresources; village cases from Zimbabwe. IIED Research Series 3(2). Sustainable AgricultureProgramme, International Institute for Environment and Development, London

Howard, P C (1991) Nature Conservation in Uganda’s Tropical Forest Reserves. IUCN, Gland,Switzerland

Hubbell, S P and R B Foster (1986) ‘Commonness and rarity in a neotropical forest: implicationsfor tropical tree conservation’ in M E Soulé (ed) Conservation Biology: the science of scarcityand diversity. Sinauer Associates, Sunderland, Massachusetts

Huntley, B J (1977). Savanna Ecosystem Project – progress report 1975/76. South AfricanNational Scientific Programme Report 12, CSIR, Pretoria

IAWA (1957) ‘International glossary of terms used in wood anatomy’, Tropical Woods 107: 1–36IAWA (1981) ‘Standard list of characters suitable for computerized hardwood identification’,

International Association of Wood Anatomists (IAWA) Bulletin 2(2–3): 99–110IAWA (1989) ‘IAWA list of microscopic features for hardwood identification’, International

Association of Wood Anatomists Journal 10: 219–332IIED (1994) Whose Eden? An overview of community approaches to wildlife management.

International Institute for Environment and Development, LondonIUCN (1980) World Conservation Strategy: living resource conservation for sustainable

development. IUCN/UNEP/WWF, Gland, SwitzerlandIUCN (1994) IUCN Red List Categories. IUCN Species Survival Commission, Gland,

Switzerland

Applied Ethnobotany

284


Jacoby, G C (1989) ‘Overview of tree-ring analysis in tropical regions’, International Associationof Wood Anatomists Bulletin 13(4): 359–379

James, S (1984) ‘Ligno-tubers and burls – their structure, function and ecological significance inMediterranean ecosystems’, The Botanical Review 50: 225–266

Jenkins, M and S Oldfield (1992) Wild plants in trade. TRAFFIC International, CambridgeJohannes, R E (1981) Words of the Lagoon: fishing and marine lore in the Palau district of

Micronesia. University of California Press, BerkeleyJohannes, R E (1998) ‘The case for data-less marine resource management: examples from

tropical nearshore fisheries’, TREE 13: 243–246Johns, T, J O Kokwaro and E K Kimanani (1990) ‘Herbal remedies of the Luo of Siaya district,

Kenya: establishing quantitative criteria for consensus’, Economic Botany 44: 369–381Johnson, D (1996) Palms: their conservation and sustained utilisation. Status Survey and

Conservation Action Plan. IUCN, Gland and CambridgeJoyal, E (1996) ‘The palm has its time: an ethnoecology of Sabal uresana in Sonora, Mexico’,

Economic Botany 50: 446–462Junikka, L (1994) ‘Survey of English macroscopic bark terminology’, International Association of

Wood Anatomists (IAWA) Journal 15(1): 1–45Kamatenesi, M (1997) ‘Utilization of the medicinal plant “nyakibazi” (Rytigynia spp) in the

multiple-use zones of Bwindi-Impenetrable National Park, Uganda’. Unpublished MSc thesis.Makerere University, Kampala

Kawano, S, A Hiratsuka and K Hayashi (1982) ‘The productive and reproductive biology offlowering plants. V. The life characteristics and survivorship of Erythronium japonicum’,Oikos 38: 129–149

Kingdon, J (1990) Island Africa: the evolution of Africa’s rare animals and plants. Collins,London

Kingston, B (1967) Working plan for Mgahinga Central Forest Reserve, Kigezi District, Uganda(1967–1977). Uganda Forest Department, Kampala

Klee, G A (1980) World systems of traditional resource management. V H Winston & Sons withEdward Arnold

Kloos, H (1976/1977) ‘Preliminary study of medicinal plants and plant products in markets ofCentral Ethiopia’, Ethnomedizin IV(1/2): 63–102

Knight, C (1991) Blood Relations: menstruation and the origins of culture. Yale University Press,New Haven

Kremen, C, A M Merenlender and D D Murphy (1994) ‘Ecological monitoring: a vital need forintegrated conservation and development programs in the tropics’, Conservation Biology 8:388–397

Kruger, L M, J J Midgley and R M Cowling (1997) ‘Resprouters vs reseeders in South Africanforest trees; a model based on forest canopy height’, Functional Ecology 11: 101–105

Kuper, A (1980) ‘Symbolic dimensions of the southern Bantu homestead’, Africa 50: 8–23Kurita, K (1982) ‘A market on boundary: the economic activities of the Pokot and the Marakwet

in Kenya’, Africa Study Monographs, Supplementary Issue 1: 71–103Lamont, B B and S Downes (1979) ‘The longevity, flowering and fire history of the grasstrees

Xanthorrhoea preissii and Kingia australis’, Journal of Applied Ecology 16: 893–899Lamont, B B, D C le Maitre, R M Cowling and N J Enright (1991) ‘Canopy seed storage in

woody plants’, The Botanical Review 57: 277–317Lange, D (1997) ‘Trade in plant material for medicinal and other purposes: a German case

study’, TRAFFIC Bulletin 17(1): 21–32Lange, D and U Schippmann (1997) Trade survey of medicinal plants in Germany: a contribution

to international species conservation. Bundesamt für Naturschütz. LV Druck, MünsterLangholz, G (1996) ‘Economics, objectives and success of private nature reserves in sub-Saharan

Africa and Latin America’, Conservation Biology 10: 271–280Larmour, P (1979) ‘Customary maps’ in Institute of South Pacific Studies Land in Solomon

Islands. Fiji Times and Herald, Suva, pp28–40Latz, P (1995) Bush Fires and Bushtucker: Aboriginal plant use in Central Australia. IAD Press,

Alice SpringsLawry, S (1990) ‘Tenure policy toward common property natural resources in sub-Saharan

Africa’, Natural Resources Journal 30: 403–422

References

285


Lay, D W (1965) ‘Effects of periodic clipping on yield of some common browse species’, Journalof Range Management 18: 181–184

Leach, M (1994) Rainforest Relations: gender and resource use among the Mende of Gola, SierraLeone. Edinburgh University Press, EdinburghLeakey, R and D O Ladipo (1996) ‘Trading on genetic variation – fruits of Dacroydes edulis’,

Agroforestry Today. April–June: 16–17Lefkovitch, L P (1965) ‘The study of population growth in organisms grouped in stages’,

Biometrics 21: 1–18Leloup, S (1984) The grapple plant project: an ecophysiological approach of the influence of

harvest on the population dynamics of the Grapple plant Harpagophytum procumbens DC.National Institute for Development Research and Documentation, Gaborone, Botswana

Leopold, A (1949) A Sand County Almanac and Sketches of Here and There. Oxford UniversityPress, New York

Leslie, P H (1945) ‘On the use of matrices in certain population dynamics’, Biometrika 33:183–212

Levin, D A and H W Kerster (1978) ‘Rings and age in Liatris’, American Naturalist 112:1121–1122

Levi-Strauss, C (1963) Structural Anthropology. Basic Books, New YorkLewis, H T (1989) ‘Ecological and technological knowledge of fire: Aborigines versus park

rangers in Northern Australia’, American Anthropologist 91: 940–961Liebenberg, L and L Steveton (1989) Cybertracker Software: Field computer system for trackers.

Noordhoek, Cape Town (http://www.icon.co.za/~ctracker)Lieth, H (1975) ‘Primary production of the major vegetation units of the world’ in H Lieth and H

Whittaker (eds) Primary Productivity of the Biosphere. Springer Verlag, Berlin, pp203–215Liming, F G (1957) ‘Homemade dendrometers’, Journal of Forestry 55: 575–577Lock, J M (1989) Legumes of Africa: a check-list. Royal Botanic Gardens, KewLoh, J, J Randers, A Macgillivray, V Kapos, M Jenkins, B Groombridge and N Cox (1998)

Overconsumption is Driving the Rapid Decline of the World’s Natural Environments. WWF-International/New Economics Foundation/WCMC

Lösch, A (1954) The Economics of Location. Yale University, New HavenMakombo, J (1998) ‘Mapping traditional resource use zones around Rwenzori Mountains

National Park, Uganda’. Unpublished MSc thesis. Makerere University, KampalaMcConnell, B R and J G Smith (1977) ‘Influence of grazing on age-yield interactions in

bitterbush’, Journal of Range Management 30: 91–93McKim, W (1972) ‘The periodic market system in north-eastern Ghana’, Economic Geography

48: 333–344MacNally, R and G P Quinn (1998) ‘Symposium introduction: the importance of scale in

ecology’, Australian Journal of Ecology 23: 1–7McNeely, J A, J Harrison and P Dingwall (eds) (1994) Protecting Nature: regional reviews of

protected areas. IUCN, MorgesMcPherson, K and K Williams (1996) ‘Establishment growth of cabbage palm, Sabal palmetto

(Areaceae)’, American Journal of Botany 83: 1566–1570Mace, G, N Collar, J Cooke, K Gaston, J Ginsberg, N Leader-Williams, M Maunder and E J

Milner-Gulland (1993) ‘The development of new criteria for listing species on the IUCN RedList’, Species 19: 16–22

Mann, C C and M L Plumme (1993) ‘The high cost of biodiversity’, Science 260: 1868–1871Marsh, A C and M K Seely (eds) (1992) Oshanas: sustaining people, environment and

development in Central Owambo, Namibia. Typoprint, Windhoek, NamibiaMarshall, L (1976) The !Kung of Nyae-nyae. Harvard University Press, Cambridge,

MassachusettsMartin, G J (1995) Ethnobotany: a methods manual. Chapman and Hall, LondonMartin, R B (1994) Alternative Approaches to Sustainable Use; what does and doesn’t work.

Paper presented at the conference on Conservation through Sustainable Use of Wildlife.University of Queensland, Brisbane

Martinez-Ramos, M (1985) ‘Claros, ciclos vitales de los arboles tropicales y regeneracion naturalde las selvas atlas perennifolias’ in A Gomez-Pompa and S del Amo (eds) Investigaciones

Applied Ethnobotany

286


sobre la Regeneracion de Selvas Atlas en Veracruz, Mexico. II. Editorial Alhambra Mexicana,S A de CV, Mexico, pp191–239

Martinez-Ramos, M, J Sarukhan, and D Pinero (1988) ‘Treefall age determination and gapdynamics: the case of Astrocaryum mexicanum at Los Tuxlas tropical rain forest’ in A J Davy,M J Hutchings and A R Watkinson (eds) Plant population Ecology. Blackwell ScientificPublications, Oxford, pp293–313

Mason, L R and G S Hutchings (1967) ‘Estimating foliage yields on Utah juniper frommeasurements of crown diameter’, Journal of Range Management 20: 161–166

Mason, P M (1995) ‘Wildlife conservation in the long term – Uganda as a case-study’.Unpublished MSc thesis in environmental economics. University of Oxford, Oxford

Mather, R, M de Boer, M Gurung and N Roche (1998) ‘Aerial photographs and “photo-maps”for community forestry’, Rural Development Forestry Network 2e: 13–22. OverseasDevelopment Institute (ODI), London

Mattos, M, D C Nepstad and I C M Vieira (1992) Cartilha sobre mapeamento de area, cubagemde maderia e inventario florestal. Woods Hole Research Centre, Belem, Brazil

Maxwell, S E and H D Delaney (1989) Designing Experiments and Analysing Data: a modelcomparison perspective. Wadsworth Publishing Company, Belmont, California

Meffe, G K, A H Erlich and D Ehrenfeld (1993) ‘Human population control: the missing agenda’,Conservation Biology 7: 1–3

Menke, J W and M J Trilica (1981) ‘Carbohydrate reserves, phenology and growth cycles of nineColorado range species’, Journal of Range Management 34: 269–277

Mentis, M T (1980) ‘Towards a scientific management of terrestrial ecosystems’, South AfricanJournal of Science 76: 536–540

Mialoundama, F (1993) ‘Nutritional and socio-economic value of Gnetum leaves in CentralAfrican Forest’ in C M Hladik, A Hladik, O F Linares, H Pagezy, A Semple and M Hadley(eds) Tropical Forests, People and Food: biocultural interactions and applications todevelopment. Man and the Biosphere Series, vol 13, UNESCO, Paris, pp177–182

Midgley, J J, M C Cameron and W J Bond (1995) ‘Gap characteristics and replacement patternsin the Knysna forest, South Africa’, Journal of Vegetation Science 6: 29–35

Midgley, J J (1996) ‘Why is the world’s vegetation not totally dominated by resprouters: an archi-tectural hypothesis’, Ecography 19: 92–95

Midgley, J J and R M Cowling (1993) ‘Regeneration patterns in a subtropical transition thicket:where are all the seedlings?’ South African Journal of Botany 59: 496–499

Midgley, J J, A Seydack, D C Reynell and D McKelly (1990) ‘Fine grain patterns in southernCape plateau forests’, Journal of Vegetation Science 1: 29–35

Miller, K R (1996) Balancing the Scales: guidelines for increasing biodiversity’s chances throughbioregional management. World Resources Institute, Washington, DC

Miller, R B (1981) ‘Explanation of coding procedure’, International Association of WoodAnatomists (IAWA) Bulletin 2(2–3): 111–145

Mills, G T (1984) ‘An inquiry into the structure and function of space in indigenous settlement inOvamboland’. Unpublished MArch thesis. University of Cape Town, Cape Town

Milton, S J (1987) ‘Effects of harvesting on four species of forest ferns in South Africa’,Biological Conservation 41: 133–146

Milton, S J (1991) ‘Slow recovery of defoliated seven-weeks fern Rumohra adiantiformis inHarkerville forest’, South African Journal of Forestry 158: 23–27

Mkanda, F X and S M Munthali (1994) ‘Public attitudes and needs around Kasungu NationalPark, Malawi’, Biodiversity and Conservation 3: 29–43

Moll, E J (1981) Trees of Natal. Eco-Lab Trust Fund, University of Cape Town, South AfricaMolur, S and S Walker (1996) Conservation Assessment and Management Plan (CAMP II) for

Selected Species of Medicinal Plants of Southern India. 12–14 February 1996. Zoo outreachorganization/CBSG, India, Coimbatore

Moore, H L (1996) Space, Text and Gender: an anthropological study of the Marakwet ofKenya. The Guildford Press, New York

Mori, S A and G T Prance (1990) ‘Taxonomy, ecology and economic botany of the Brazil nut(Bertholletia excelsa Humb. & Bompl: Lecythidaceae)’, Advances in Economic Botany 8:130–150

References

287


Muir, D P (1990) ‘Indigenous forest utilization in KwaZulu: a case study of the Hlatikulu forest,Maputaland’. Unpublished MSc thesis. University of Natal, Pietermaritzburg

Murphree, M (1995) Optimal principles and pragmatic strategies: creating an enabling politico-legal environment for community-based natural resource management – CBNRM. KeynoteAddress, Conference of the National Resources Management Programme, SADCC TechnicalCoordination Unit, USAID-NRMP. Chobe, Botswana

Mustart, P J and R M Cowling (1992) ‘Impact of flower and cone harvesting on seed banks andseed set of serotinous Agulhas Proteaceae’, South African Journal of Botany 58: 337–342

Nantel, P, D Gagnon and A Nault (1996) ‘Population viability analysis of American ginseng andwild leek harvested in stochastic environments’, Conservation Biology 10(2): 608–621

Nhira, C and L Fortmann (1993) ‘Local woodland management: realities at the grass roots’ in PN Bradley and K McNamara (eds) Living with Trees: policies for forestry management inZimbabwe. World Bank Technical Paper 210, World Bank, Washington, DC, pp139–156

Nichols, P (1991) Social Survey Methods: a Field-Guide for Development Workers. DevelopmentGuidelines No 6. Oxfam, Oxford

Norton-Griffiths, M and C Southey (1995) ‘The opportunity costs of biodiversity in Kenya’,Ecological Economics 12: 125–139

Noss, R F (1990) ‘Indicators for monitoring biodiversity: a hierarchical approach’, ConservationBiology 4: 355–364

Obermeyer, A A (1978) ‘Ornithogalum: a revision of the southern African species’, Bothalia 12:323–376

Obeso, J R (1997) ‘Costs of reproduction in Ilex aquifolium: effects at tree, branch and leaflevels’, Journal of Ecology 85: 159–166

Ogle, B M and L E Grivetti (1985) ‘Legacy of the chameleon: edible wild plants in the kingdomof Swaziland, southern Africa. A cultural, ecological, nutritional study. Part II –demographics, species availability and dietary use, analysis by ecological zone’, Ecology ofFood and Nutrition 17: 1–30

Oldfield, S (1997) Cactus and Succulent Plants: status survey and Conservation Action Plan.IUCN, Gland and Cambridge

Olmsted, I and E R Alvarez-Buylla (1995) ‘Sustainable harvesting of tropical trees: demographyand matrix models of two palm species in Mexico’, Ecological Applications 5(2): 484–500

Olson, D M and E Dinerstein (1998) ‘The Global 200: a representation approach to conservingthe earth’s most biologically valuable ecoregions’, Conservation Biology 12: 502–515

O’Neill, A L, L M Head and J K Marthick (1993) ‘Integrating remote sensing and spatial analysistechniques to compare Aboriginal and pastoral fire patterns in the East Kimberly, Australia’,Applied Geography 13: 67–85

Osborne, R, A Grove, P Oh, T Mabry, J Ng and A Seawright (1994) ‘The magical and medicinalusage of Stangeria eriopus in South Africa’, Ethnopharmacology 43: 67–72

Ostrom, E (1990) Governing the Commons: the evolution of institutions for collective action.Cambridge University Press, Cambridge

Oyama, K (1993) ‘Are age and height correlated in Chamaedorea tepejilote (Palmae)?’ Journal ofTropical Ecology 9: 381–385

Padoch, C (1988) ‘Aguaje (Mauritia flexuosa), L.f., in the economy of Iquitos, Peru’, Advances inEconomic Botany 6: 214–224

Park, S (1981) ‘Rural development in Korea’, Economic Geography 57: 113–126Parker, E, D Posey, J Frechione and L F Da Silva (1983) ‘Resource exploitation in Amazonia:

ethnoecological examples from four populations’, Annals of Carnegie Museum 52: 163–203Pate, J and K Dixon (1982) Tuberous, Cormous and Bulbous Plants: biology of an adaptive strat-

egy in western Australia. University of Western Australia Press, NedlandsPate, J S, R H Froend, B J Bowen, A Hansen, and J Kuo (1990) ‘Seedling growth and storage and

characteristics of seeder and resprouter species of Mediterranean-type ecosystems of south-western Australia’, Annals of Botany 65: 585–601

Pate, J S, K A Meney, and K W Dixon (1991) ‘Contrasting growth and morphologicalcharacteristics of fire-sensitive (obligate seeder) and fire-resistant (resprouter) species ofRestionaceae (S. Hemisphere Restiads) from south-western Australia’, Australian Journal ofBotany 39: 505–525

Applied Ethnobotany

288


Pellew, R A (1980) ‘The production and consumption of Acacia browse and its potential foranimal protein production’ in H N le Houerou (ed) Browse in Africa. International LivestockCentre for Africa, Addis Ababa, Ethiopia, pp223–231

Peterken, G F (1993) Woodland Conservation and Management. Chapman & Hall, LondonPeters, C M (1994) Sustainable Harvest of Non-Timber Forest Plant Resources in Tropical Moist

Forest: an ecological primer. Biodiversity Support Program, Washington, DCPeters, C M (1996a) ‘Beyond nomenclature and use: a review of ecological methods for

ethnobotanists’ in M N Alexiades (ed) Selected Guidelines for Ethnobotanical Research: afield manual. New York Botanical Garden, New York, pp241–276

Peters, C M (1996b) Pedoman inventarisasi dan Penghasilan Sumber Daya Alam Non-Kayu[Inventory and yield manual for non-wood forest products]. Social Forestry Developmentproject, GTZ and departemen Kehutanan, Sanggau, West Kalimantan, Indonesia

Peters, C M and E Hammond (1990) ‘Fruits from the flooded forests of Peruvian Amazonia: yieldestimates for natural populations of three promising species’, Advances in Economic Botany8: 159–176

Philip, M S (1994) Measuring Trees and Forests. Second edition. CAB International, WallingfordPhillips, O (1993) ‘The potential for harvesting fruits in tropical rainforests: new data from the

Peruvian Amazon’, Biodiversity and Conservation 2: 18–38Phillips, O L (1996) ‘Some quantitative methods for analysing ethnobotanical knowledge’ in M N

Alexiades (ed) Selected Guidelines for Ethnobotanical Research: a field manual. New YorkBotanical Garden, New York, pp171–197

Phillips, O and A H Gentry (1993a) ‘The useful plants of Tambopata, Peru: I. Statistical hypothe-ses tests with a new quantitative technique’, Economic Botany 47: 15–32

Phillips, O and A H Gentry (1993b) ‘The useful plants of Tambopata, Peru: II. Additionalhypothesis testing in quantitative ethnobotany’, Economic Botany 47: 33–43

Pickett, S T A and P S White (1985) The Ecology of Natural Disturbance and Patch Dynamics.Academic Press, New York

Picton, J and J Mack (1989) African Textiles. British Museum Publications, LondonPinard, M (1993) ‘Impacts of stem harvesting on populations of Iriartea deltoidea (Palmae) in an

Extractive Reserve in Acre, Brazil’, Biotropica 25(1): 2–14Pinard, M A and F E Putz (1992) ‘Population matrix models and palm resource management’,

Bull. Inst. Fr. Études Andines 21(2): 637–649Pitman, N C A, J Terborgh, M R Silman and P Nuñez (1999) ‘Tree species distributions in an

upper Amazonian forest’, Ecology 80(8): 2651–2661Poffenberger, M, B McGean, A Khare and J Campbell (1992) Field Methods Manual, volume 2:

community forest economy and use patterns. Joint Forest Management Support Programme,Society for Promotion of Wastelands Development, New Delhi, India

Poffenberger, M, B McGean, N H Ravindranath and M Gadgil (1992) Field Methods Manual,volume 1: diagnostic tools for supporting joint forest management systems. Joint ForestManagement Support Programme, Society for Promotion of Wastelands Development, NewDelhi, India

Polak, A M (1992) Major Timber Trees of Guyana: a field guide. Tropenbos Series 2. TropenbosFoundation, Wageningen

Poole, P (1995) Indigenous Peoples, Mapping and Biodiversity Conservation: an analysis ofcurrent activities and opportunities for applying geomatics technologies. Biodiversity SupportProgram (BSP), Peoples and Forest Programme Discussion Paper Series. BSP/WWF/TheNature Conservancy, Washington, DC

Porter, D, B Allen and G Thompson (1991) Development in Practice: paved with good intentions.Routledge, London

Posey, D A (1984) ‘A preliminary report on diversified management of tropical forest by theKayapo Indians of the Brazilian Amazon’, Advances in Economic Botany 1: 112–126

Prance, G T, W Balee, B M Boom and R L Carneiro (1987) ‘Quantitative ethnobotany and thecase for conservation in Amazonia’, Conservation Biology 1(4): 296–310

Pratt, B and P Loizos (1992) Choosing Research Methods: Data Collection for DevelopmentWorkers. Development Guidelines No 7. Oxfam, Oxford

References

289


Pretty, J N, I Guijt, I Scoones and J Thompson (1995) Participatory Learning and Action: atrainer’s guide. IIED Participatory Methodology Series. International Institute forEnvironment and Development, London

Prinsloo, J (1982) ‘Construction of a questionnaire on poverty; lessons from an urban survey’ inSALDRU (ed) Questionnaires Are No Short Cut. SALDRU Working Paper 43. South AfricanLabour and Development Research Unit, University of Cape Town, Cape Town, pp29–41

Prior, J A B and P E Gasson (1990) ‘Comparative wood anatomy of Afromontane and bushveldspecies from Swaziland’, IAWA Bulletin 11: 319–336

Rabinowitz, D, S Cairns and T Dillon (1986) ‘Seven forms of rarity and their frequency in theflora of the British Isles’ in M E Soulé (ed) Conservation Biology: the science of scarcity anddiversity. Sinauer Associates, Sunderland, Massachusetts, pp182–204

Rapoport, A (1977) Human Aspects of Urban Form: towards a man-environment approach tothe urban form and design. Pergamon, New York

Raunkiaer, C (1934) Life Forms of Plants and Statistical Plant Geography. Clarendon Press,Oxford

Rebelo, T (1995) Proteas: a field guide to the Proteas of southern Africa. Fernwood Press,Vlaeberg

Rebelo, A G and P M Holmes (1988) ‘Commercial exploitation of Brunia albiflora (Bruniaceae)in South Africa’, Biological Conservation 45: 195–207

Reichel-Dolmatoff, G (1996) The Forest Within: the world-view of the Tukano AmazonianIndians. Themis Books, Dartington

Reynolds, N (1981) The design of rural development: proposals for the evolution of a socialcontract suited to conditions in southern Africa. Parts 1 and 2. SALDRU Papers Nos 41 and42. University of Cape Town, South Africa

Richards, P (1996) ‘Forest indigenous peoples: concept, critique and cases’, Proceedings of theRoyal Society of Edinburgh 104B: 349–365

Ruiters, C, B McKenzie and L M Raitt (1993) ‘Life-history studies of the perennial geophyteHaemanthus pubescens (Amaryllidaceae) in lowland coastal fynbos, South Africa’,International Journal of Plant Science 154: 441–449

Rundstrom, R A (1990) ‘A cultural interpretation of Inuit map accuracy’, The GeographicalReview 80: 155–168

Rutherford, M C (1979) ‘Plant-based techniques for determining available browse and browseutilisation: a review’, The Botanical Review 45: 203–228

Rutherford, M C (1982) ‘Woody plant biomass distribution in Burkea africana savannas’ in B JHuntley and B H Walker (eds) Ecology of Tropical Savannas. Ecological Studies 42. SpringerVerlag, Berlin, pp121–141

Rykiel, E J, R N Coulson, P J H Sharpe, T F H Allen and R O Flamm (1988) ‘Disturbancepropagation by bark beetles as an episodic landscape phenomenon’, Landscape Ecology 1:129–139

Sada, P O, M L McNulty and I A Adelemo (1978) ‘Periodic markets in a metropolitanenvironment: an example of Lagos, Nigeria’ in R H T Smith (ed) Marketplace Trade.University of British Columbia Centre for Transportation Studies, Vancouver

Sakai, A, S Sakai and F Akiyama (1997) ‘Do sprouting tree species on erosion-prone sites carrylarge reserves of resources?’ Annals of Botany 79: 625–630

Salick, J (1992) ‘Amuesha forest use and management: an integration of indigenous forest useand natural forest management’ in K Redford and C Padoch (eds) Conservation ofNeotropical Forests: working from traditional resource use. Columbia University Press, NewYork, pp305–332

Salick, J, A Meija and T Anderson (1995) ‘Non-timber forest products integrated with naturalforest management, Rio San Juan, Nicaragua’, Ecological Applications 5: 878–895

Scharfetter, H (1987) ‘Timber resources and needs in southern Africa’, South African Journal ofScience 83: 256–259

Schoffeleers, J M (1978) Guardians of the Land. Mambo Press, GweloSchupp, J R (1992) ‘Effect of root pruning and summer pruning on growth, yield, quality, and

fruit maturity of McIntosh apple trees’, Acta Hort. 322: 173–183Schuster, J L and J George (1976) ‘Redberry juniper response to top removal’, Journal of Range

Management 29: 258–259

Applied Ethnobotany

290


Schutte, A L, J H J Vlok and B-E van Wyk (1995) ‘Fire-survival strategy –a character oftaxonomic, ecological and evolutionary importance in fynbos legumes’, Plant Systematics andEvolution 195: 243–259

Schweingruber, F H (1992) ‘Annual growth rings and growth zones in woody plants in southernAustralia’, International Association of Wood Anatomists Bulletin 13(4): 359–379

Scifres, C J, Kothmann, M M and Mathis, C W (1974) ‘Range site and grazing system influenceon regrowth after spraying honey mesquite’, Journal of Range Management 27: 97–100

Scott, P J (1993) ‘Fringe benefits: national park boundaries as areas for local communityutilization: the case of the Impenetrable (Bwindi) forest’. Unpublished MSc thesis. AgriculturalUniversity of Norway, Bergen

Scott, P J (1995) ‘Progress report on multiple-use management to Uganda National Parks’.Unpublished report. WWF Rwenzori Mountains Conservation and Development Project, FortPortal, Uganda

Scudder, T and T Conelly (1985) Management Systems for Riverine Fisheries. InternationalDevelopment Association Report. International Development Association, Washington, DC

Seydack, A H W (1995) ‘An unconventional approach to timber yield regulation for multi-aged,multi-species forests. I. Fundamental considerations’, Forest Ecology and Management 77:139–153

Seydack, A H W, W J Vermeulen, H E Heyns, G P Durrheim, C Vermeulen, D Willems, M AFerguson, J Huisamen and J Roth (1995) ‘An unconventional approach to timber yieldregulation for multi-aged, multi-species forests. II. Application to a South African forest’,Forest Ecology and Management 77: 155–168

Shackleton, C M (1993) ‘Demography and dynamics of the dominant woody species in a commu-nal and protected area of the eastern transvaal lowveld’, South African Journal of Botany 59:569–574

Shackleton, C M (1993) ‘Fuelwood harvesting and sustainable utilization in a communal grazingland and protected area of the eastern Transvaal lowveld’, Biological Conservation 63:247–254

Shackleton, C (1998) ‘Annual production of harvestable deadwood in semi-arid savannas, SouthAfrica’, Forest Ecology and Management 112: 139–144

Shackleton, S (1990) ‘Socio-economic importance of Cymbopogon validus in Mkambati GameReserve, Transkei’, South African Journal of Botany 56: 675–682

Shafer, C L (1990) Nature Reserves – island theory and conservation practice. SmithsonianInstitution Press, Washington, DC

Shankar, U, K S Murali, R U Shaanker, K N Ganeshaiah and K S Bawa (1996) ‘Extraction ofnon-timber forest products in the forests of Biligiri Rangan hills, India. 3. Productivity, extrac-tion and prospects of sustainable harvest of Amla (Phyllanthus emblica), Euphorbiaceae’,Economic Botany 50: 270–279

Shanley, P and L Luz (in press) ‘Eastern Amazonian medicinals: marketing, utilisation andimplications for conservation’, BioScience

Shanley, P, J Galvao and L Luz (1997) ‘Limits and strengths of local participation: a case-study inEastern Amazonia’, PLA Notes (IIED, London) 14: 64–67

Shanley, P, M Cymerys and J Galvao (1998) Fruitiferas da mata na Vida Amazonica. EditoraSupercores, Chaco, Belem-Para

Sharpe, B (1998) ‘Forest people and conservation initiatives: the cultural context of rainforestconservation in West Africa’ in B Goldsmith (ed) Tropical Rain forest: a wider perspective.Chapman & Hall, London, pp75–97

Sheil, D (1995) ‘A critique of permanent plot methods and analysis with examples from Budongoforest, Uganda’, Forest Ecology and Management 77: 11–34

Shiokura, T (1989) ‘A method to measure radial increment in tropical trees’, IAWA Bulletin 10:147–154

Shipton, P (1994) ‘Land and culture in tropical Africa: soils, symbols and the metaphysics of themundane’, Annual Review of Anthropology 23: 347–377

Schonau, A P G (1982) ‘Models for Stand Development in a Short Rotation crop’. Paperdelivered at the Symposium on Growth Modelling, University of Stellenbosch, South Africa

Siegfried, W R, G A Benn and C M Gelderblom (1998) ‘Regional assessment and conservationimplications of landscape characteristics of African national parks’, Biological Conservation84: 131–140

References

291


Singh, L and J S Singh (1991) ‘Species structure, dry matter dynamics and carbon flux of a drytropical forest in India’, Annals of Botany 68: 263–273

Skinner, G W (1964) ‘Marketing and social structure in rural China. Part 1’, Journal of AsianStudies 24: 3–43

Skinner, G W (1965) ‘Marketing and social structure in rural China. Part 2’, Journal of AsianStudies 24: 195–228

Smith, A B (1994) ‘Metaphors of space: rock art and territoriality in southern Africa’ in T ADowson and D Lewis-Williams (eds) Contested Images: diversity in southern African rock artresearch. Witwatersrand University Press, Johannesburg, pp373–384

Smith, C A (1985) ‘How to count onions: methods for a regional analysis of marketing’ in SPlattner (ed) Markets and Marketing. Society for Economic Anthropology, University Press ofAmerica, pp 49–77

Smith, R L (1980) Ecology and Field Biology. Harper and Row, New YorkSoetprayitno, T, T L Tobing and E A Widjaja (1988) ‘Why the Sundanese of West Java prefer

slope inhabiting Gigantochloa pseudoarundinacea to those growing in the valley’,BAMBOOS Current research. Proceedings of the International Workshop, 14–18 November1988, pp215–217

Sokal, R R and F J Rohlf (1987) Introduction to Biostatistics. Freeman, New YorkSouthwood, T R E, U K Brown and P M Reader (1986) ‘Leaf palatability, life expectancy and

herbivore damage’, Oecologia 70: 544–548Steiner, K E (1993) ‘Has Ixianthes (Scrophulariaceae) lost its special bee?’ Plant Systematics and

Evolution 185: 7–16Stern, W L (1978) ‘Index Xylariorum. Institutional wood collections of the world. 2’, Taxon 27

(2/3): 233–269Stockdale, M C and J D Power (1994) Forest Ecology and Management 64: 47–57Sussman, R W, G M Green and L K Sussman (1994) ‘Satellite imagery, human ecology,

anthropology and deforestation in Madagascar’, Human Ecology 22: 333–354Swart, J P J (1980) ‘Non-destructive wood sampling methods from living trees: a literature

survey’, International Association of Wood Anatomists Bulletin 1(1–2): 42Tailfer, Y (1989) La forêt dense d’Afrique tropicale. Identification practique des principaux

arbres. Tome 1, ACCT, ParisTait, N G and A B Cunningham (1988) ‘An identification guide to commonly sold medicinal

plants’. Unpublished report. Institute of Natural Resources, University of Natal,Pietermaritzburg, South Africa

Tamasari, F (1995) Body, names and movement: images of identity among the Yolngu of north-east Arnhem Land. Unpublished PhD thesis. London School of Economics, UniversityCollege, London

Taylor, D M (1990) ‘Late quaternary pollen records from two Ugandan mires: evidence forenvironmental change in the Rukiga highlands of south-west Uganda’, Paleogeography,Paleoclimatology, Paleoecology 80: 283–300

Thorsen W, W R J Dean and S J Milton (1997) ‘Simulated plant population responses to small-scale disturbances in semi-arid shrublands’, Journal of Vegetation Science 8: 163–176

Tietema, T (1993) ‘Biomass determination of fuelwood trees and bushes of Botswana, SouthernAfrica’, Forest Ecology and Management 60: 257–269

Tinley, K L (1987) ‘Achieving a balance between long and short term research’ in D A Saunders,G W Arnold, A A Burbidge and A J M Hopkins (eds) Nature Conservation: the role ofremnants of native vegetation. Surrey Beatty and Sons Pty Ltd, Chipping-Norton, NSW,Australia, pp347–350

Toledo, V M, A Batis, R Bacerra, E Martinez and C H Ramos (1992) ‘Products from the tropicalrain forests of Mexico: an ethnoecological approach’ in M Plotkin and L Famolare (eds) Non-Wood Products from Tropical Rainforests. Conservation International, Washington, DC,pp99–109

Trager, L (1995) ‘Minimum data sets in the study of exchange and distribution’ in E F Moran(ed) The Comparative Analysis of Human Societies. Lynne Reiner Publishers, London,pp75–96

Trapnell, C G (1953) The Soils, Vegetation and Agricultural Systems of North-Eastern Rhodesia.Government Printer, Lusaka, Zambia

Applied Ethnobotany

292


Trlica, M J, M Buwai, and J W Menke (1977) ‘Effects of rest following defoliations on the recov-ery of several range species’, Journal of Range Management 30: 21–27

Trollip, A-M (1995) ‘Meaning of blankets, towels and T-shirts in the context of acculturation’,South African Journal of Ethnology 18: 150–154

Turner, M G (1989) ‘Landscape ecology: the effect of pattern on process’, Ann. Rev. Ecol. Sys.20: 171–197

Turner, N J (1988) ‘The importance of a rose: evaluating the cultural significance of plants’,American Anthropologist 90: 272–290

Turner, R M 1990 ‘Long-term vegetation change at a fully protected Sonoran desert site’, Ecology71: 464–477

Turner, V (1967) The Forest of Symbols. Cornell University Press, Ithaca, New YorkTurner, V (1969) The Ritual Process. Aldine, ChicagoUkwu, U I (1969) ‘Markets in Iboland’ in B W Hodder and U I Ukwu (eds) Markets in West

Africa. Ibadan University Press, IbadanUNESCO (1998) The Economist, 14 Marchvan Binsbergen, W J (1985) Theoretical Explorations in African Religion. Kegan Paul

International, Londonvan Binsbergen, W J (1988) ‘The land as body: an essay on the interpretation of ritual among the

Manjaks of Guinea-Bissau’, Medical Anthropological Quarterly 2: 386–401van den Beemer, J P M (1992) ‘Ideas and usage. Environment in Aouan society, Ivory Coast’ in E

Croll and D Parkin (eds) Bush Base: Forest Farm. Culture, Environment and Development.Routledge, London, pp97–109

van Laar, A and C J Geldenhuys (1975) ‘Tariff tables for indigenous tree species in the southernCape Province’, Forestry in South Africa 17: 29–36

van Wyk, G F, D A Everard, J J Midgley and I G Gordon (1996) ‘Classification and dynamics ofa southern African subtropical coastal lowland forest’, South African Journal of Botany62(3): 133–142

Vance, J E (1970) The Merchant’s World: the geography of wholesaling. Foundations ofEconomic Geography Series. Prentice-Hall, Englewood Cliffs, New Jersey

Vasquez, R and A H Gentry (1989) ‘Use and misuse of forested harvested fruits in the Iquitosarea’, Conservation Biology 3: 350–361

Viljoen, A J (1988) ‘Long-term changes in the tree component of the vegetation in the KrugerNational Park’ in I A MacDonald and R J M Crawford (eds) Long-term data series relatingto southern Africa’s renewable natural resources. South African National ScientificProgrammes Report 157. CSIR, Pretoria, pp310–315

Wade, R (1987) ‘The management of common property resources: collective action as an alterna-tive to privatisation or state regulation’, Cambridge Journal of Economics 11: 95–106

Waddy, J A (1982) ‘Biological classification from a Groote eylandt Aborigine’s point of view’,Journal of Ethnobiology 2: 63–77

Wagener, W W (1961) ‘Past fire incidence in Sierra Nevada forests’, Journal of Forestry 59:739–748

Walsh, F J (1990) ‘An ecological study of traditional use of “country”: Martu in the Great andLittle Sandy deserts, Western Australia’, Proceedings of the Ecological Society of Australia 16:23–37

Walsh, F J (1993) ‘The relevance of some aspects of Aboriginal subsistence to the management ofnational parks; with reference to Martu people of the Western Desert’, in J Birkhead, T deLacy and L Smith (eds) Aboriginal involvement in parks and protected areas. AustralianInstitute of Aboriginal and Torres Strait Islander Studies, Canberra, pp75–97

Walter, S and J-C R Rakotonirina (1995) L’exploitation de Prunus africana à Madagascar. PCDIZahamena et la Direction des Eaux et Forêts. Antananarivo, Madagascar

Watkinson, A R and J White (1985) ‘Some life-history consequences of modular construction inplants’, Phil. Trans. Roy. Soc. Lond. B313: 31–51

Weller, S C and A K Romney (1988) Systematic Data Collection. Qualitative Research Methods,vol 10. Sage Publications, Newbury Park, California

Werner, P A (1978) ‘On the determination of age in Liatris aspera using cross-sections of corms:implications for past demographic studies’, American Naturalist 112: 1113–1144

References

293


White, F (1976) ‘The underground forests of Africa: a preliminary review’, Gardens Bull.(Singapore) 29: 55–71

Whitmore, T C (1988) ‘The influence of tree population dynamics on forest species composition’in A J Davy, M J Hutchings and A R Watkinson (eds) Plant Population Biology. Blackwell,Oxford, pp271–291

Wild, R G, A B Cunningham and J Mutebi (1995) ‘People, parks and plant use: networks toenhance the conservation of montane forests in Uganda, East Africa’ in D A Saunders, J LCraig and E M Mattiske (eds) Nature Conservation 4: the role of networks. Surrey Beatty &Sons, Sydney, pp112–121

Wild, R G and J Mutebi (1996) Conservation through Community Use of Plant Resources.People and Plants Working Paper 5. UNESCO, Paris

Williams, V L (1996) ‘The Witwatersrand muti trade’, Veld & Flora 82(1): 12–14Wilmsen, E N (1978) ‘Seasonal effects of dietary intake on Kalahari San’, Federation Proceedings

37: 65–72Wilson, M (1957) Rituals of Kinship Among the Nyakusa. Oxford University Press, LondonWily, E (1995) ‘What future for community forestry in Nepal?’ Social Development Newsletter 3,

Overseas Development Administration, LondonWitkowski, E T F, B B Lamont and F J Obbens (1994) ‘Commercial picking of Banksia

hookeriana in the wild reduces subsequent shoot, flower and seed production’, Journal ofApplied Ecology 31: 508–520

Witkowski, E T F and B B Lamont (1996) ‘Nutrient losses from commercial picking andcockatoo removal of Banksia hookeriana blooms at the organ, plant and site levels’, Journalof Applied Ecology 33: 131–140

World Conservation Union, United Nations Environment Programme, World Wide Fund ForNature (1991) Caring for the Earth: a strategy for sustainable living. WCU, UNEP, WWF,Gland, Switzerland

World Resources Institute (1992) Global Biodiversity Strategy: guidelines for action to save,study and use Earth’s biotic wealth sustainably and equitably. WRI/IUCN/UNEP, Washington,DC

Wyatt-Smith, J and A J Vincent (1962) ‘The swing from qualitative to quantitative assessment ofindividual tree crown parameters in the Malayan Forest Service’, Malayan Forester 25:276–291

Yeaton, R I (1988) ‘Porcupines, fire and the dynamics of the tree layer of the Burkea africanasavanna’, Journal of Ecology 76: 1017–1029

Zar, J H (1998) Biostatistical Analysis. Fourth edition. Prentice-Hall Biological Sciences,Englewood Cliffs, New Jersey

Applied Ethnobotany

294


Acacia karoo (sweet-thorn) 117, 144, 145,146

Acacia mearnsii (Australian black wattle) 108,109–10, 137

access height, fruits 129–30Adansonia digitata (African baobab) 137–9,

138adaptive management approach 158, 159aerial photographs 196, 199, 200–1, 202

for participatory mapping 218, 219, 220,247–8, 250, 251

Africa, soils 1, 2African baobab (Adansonia digitata) 137–9,

138ageing plants 98, 115

bulbs 123, 125corms 123–4, 125dendrochronology 115, 116, 116–17grass trees 118, 119, 120leaves 125–6palms 118–21, 119, 122scar counting 117–18, 119–20, 122tree ferns 118

Amaryllidaceae 100, 123ash 37Australian black wattle (Acacia mearnsii) 108,

109–10, 137

background information, sources 14bamboo (Synarundinaria alpina) 214–15, 216Banksia hookeriana 106–7bark

characters 34, 35, 39–40, 56–7damage 212damage ratings 135–7, 136harvesting impacts 135–9, 138, 144, 145,

174mass 109–11measuring 108–11, 110thickness 109, 110

bark gauges 109, 110basal area (ba) 98, 100, 102basketry 51, 54–9, 54, 58, 59, 133, 174belt transects 160–1, 161–2, 163biomass, measuring 102–4, 114, 114biosphere reserves 3boundaries 223, 227, 230, 241, 245–6, 258–9

‘invisible’ 245, 247–8, 253, 258–9mapping 247–53ritual 253–7

Brachystegia 117, 137, 138see also miombo woodland

Brunia albiflora 129–30Bruniaceae 118, 129, 153buffer zones 3–5, 236, 271building materials 19–20, 21, 111–12, 112,

133, 158, 181–2, 215bulbs 122

ageing 123, 125characters 34, 40–1, 42harvesting impacts 169–71measuring diameter 100sales 81

burning 8, 220bush plum (Dacryodes edulis) 92, 93, 94Bwindi-Impenetrable National Park (Uganda)

51, 199, 201–2, 212, 224stakeholders 260, 260, 262–3

CAMP (Conservation Assessment andManagement Plan) workshops 206

CAMPFIRE (Communal Areas ManagementProgramme for Indigenous Resources,Zimbabwe) xix, 231

canopy, measuring 105–7Catha edulis (khat) 94–5CBNRM (community-based natural resource

management) 176, 226–33, 234, 247central place theory 63, 68–70, 73change 233‘chapatti’ (Venn) diagrams 25, 260–1charcoal 37Christaller-Lösch land-use model 68–70, 73cinnamon tree (Cinnamomum zeylanicum)

133, 137classification 44–8climate changes 7–8clinometers 97, 98, 101–2, 102, 103, 104Colophospermum mopane see mopane

woodlandcolour 34, 35, 37, 43commercial forest inventories 164–5commercial harvesting 61–2common property 239, 240, 245

Index

Page numbers in italics refer to figures and tables


Communal Areas Management Programmefor Indigenous Resources, Zimbabwe(CAMPFIRE) xix, 231

community-based conservation 223–33, 225,263

community-based natural resourcemanagement (CBNRM) 176, 226–33,234, 247

community monitoring 176–80community support 15computer analysis, questionnaires 30computers, field 177–80, 178, 218conflicts 223–4, 230, 236, 237conservation 3

conflicts 223–4, 230, 236, 237ecological influences 233–7, 234innovative approaches 271non-monetary benefits 237, 238problems xix–xxresilience to civil conflict xix, 271social and cultural links 258species requiring special effort 62, 89–90‘traditional’ practices 1, 5–6, 225–6, 233see also resource management

conservation areas 3–5, 194–5, 199–202,206–7, 209

Conservation Assessment and ManagementPlan (CAMP) workshops 206

conservation behaviour 222–6, 225construction materials 19–20, 21, 133, 158,

181–2, 215traditional Owambo housing 111–12, 112

consumption pressure 267, 268coppice rotation 182–3cork oak (Quercus suber) 39, 108, 137corms 42, 100, 122, 123–4, 125, 169–71, 170cost-benefit analyses 236–7costs of monitoring systems 180cross-checking 19–22, 20, 21, 22, 48, 60–1

when ageing plants 115official data 14, 19participatory studies 25, 27, 28species names 18–19

crowns, trees 105–7, 130, 141, 142crystals 39cultivation 266cultural issues 1, 219, 221, 233, 247customary conservation practices 1, 5–6,

225–6, 233Cybertracker system 177–80, 178cycads 171, 172Cymbopogon validus (thatch grass) 157, 182

Dacryodes edulis (bush plum) 92, 93, 94Dalbergia melanoxylon (ebony) 43, 54, 266deadwood 20, 22, 97, 113, 158, 181defoliation, simulated 134–5

demand 10–14, 61, 88dendrochronology 115, 116, 116–17destructive harvesting 2, 132, 169, 202, 264,

266dioecious species 128, 129roots 140–1, 142

destructive sampling 100, 103diameter at breast height (dbh), measuring

98–100, 99dietary surveys 21–2, 26–7dioecious species 128, 129distribution of species 202–5, 204

and threat 205–8disturbance events 158, 161, 171, 190, 191,

192, 194cultural meaning 221fires 8, 161, 190, 191, 208, 209, 210in forests 209–12harvesting 208–9

documentation of specimens 31domesticated plants 9, 62, 92, 93domestication xviii, 9, 62, 266Drosera (sundews) 124–5dyes 43, 174

ebony (Dalbergia melanoxylon) 43, 54, 266ecological ‘filters’, classification systems 150–6edible wild greens 22, 27, 132elasticity analyses 189–91ensuli (Smilax anceps) 58, 59equatorial forests 8–9, 33equipment xv, 97, 176

for ageing plants 116, 117bark gauges 109, 110field computers 177–80, 178, 218fruit traps 107–8GPS systems 97, 179, 251for measuring plants 98, 99–102, 102,

103–4for plant population impact studies 144–5

ethical change 269–71ethical guidelines, specimen collection 32ethnobotanical inventories 30–2, 44–59ethnobotanical market surveys 87–95ethnotaxonomy 45extrapolation 16exudates 34, 37, 38–9, 130–2, 131

Faurea macnaughtonii 137, 138Ficus natalensis 139field characters, for taxonomy 32–44, 35field computers 177–80, 178, 218‘filters’ for choosing priority species 147–50fires 8, 161, 190, 191, 208, 209, 210

human-induced (burning) 8, 220fish poisons 43, 44flowers 107, 129–30

Applied Ethnobotany

296


fluorescence 44folk taxonomy 45–8

see also vernacular namesfood lists 26–7food records 26forest inventories, commercial 164–5frothiness 43, 44fruit traps 107–8fruits 107–8, 128–30fuelwood 16, 20, 22, 70, 89, 97, 113, 158,

181overexploitation model 70

geographic information systems (GIS) 215–17,218, 220

geophytes, harvesting impact 169–70girth at breast height (gbh), measuring 98,

99–100, 99GIS (geographic information systems) 215–17,

218, 220global linkages 267GPS (global positioning systems) 97, 179, 251grass trees, ageing 118, 119, 120‘green social security’ 1, 267greens, edible wild 22, 27, 132growth rates 154–6, 157

harvester assessments 50–1, 54harvesters 60, 62harvesting 96, 195–6, 264–6, 266

bark 108–9sustainability 1, 1–2, 3, 5–6, 147–50see also harvesting impacts; harvesting

patternsharvesting gradients 165–7, 168harvesting impacts 95, 96, 126–8, 127

bark 135–9, 136, 138, 144, 145, 174exudates 130–2, 131leaves 132–5, 134measuring using size-class distributions

173–6, 173, 174, 175reproductive structures 128–30roots 140–2, 140, 141stems 143whole plants 167–71, 170see also harvesting

harvesting patterns 212–21, 214harvesting rules 180–1height, measuring 100–2, 102, 103–4, 105herbal medicines 25, 44, 46, 48, 81, 167–9herbarium specimens 30–2Hevea brasiliensis (rubber) 131, 132household mapping 260Hyphaene coriacea (ilala palm) 59, 183Hyphaene petersiana 51, 54, 196hypsometers see clinometers

IAWA (International Association of WoodAnatomists), list of wood characters 43

ICDPs (Integrated Conservation andDevelopment Projects) 194, 224

ilala palm (Hyphaene coriacea) 59, 183individual interview surveys 26–8, 29–30, 46,

50–1, 54, 92‘informant’ accuracy 19informant consensus 48, 49, 51innovative approaches to conservation 271institutional support 271institutions 231–3, 259, 260, 262Integrated Conservation and Development

Projects see ICDPsInternational Association of Wood Anatomists

see IAWAinterviews 26–8, 29–30, 46, 50–1, 54, 92

cross-checking 20, 22inventories 30–2, 44–59

marketplace surveys 87–92‘invisible’ boundaries 245, 247–8, 253, 258–9Iridaceae 122, 123IUCN (World Conservation Union)

categories of threat 90protected area categories xvi, 3, 4recognized conserved areas 194, 194Red List categories 205–6, 207–8

Joint Forest Management Programme (Indiaand Nepal) xix, 231

Julbernardia see miombo woodland

khat (Catha edulis) 94–5

land-use, alternatives to conservation 236–7land-use assessments 236–7landscape classification 192–6language issues 18–19, 29, 45latex see exudatesleaves

ageing 125–6harvesting impacts 132–5, 134measuring 104–5

lianas 97–8life form classification system 150–1light requirements 154–6, 155, 156Liliaceae 100, 122, 123, 169–70Linnean classification 44–5, 47, 48local knowledge 212–19, 214, 220logging 211–12long-term surveys 16, 28–30Lösch land-use model 68–70, 73

mapping 23boundaries 247–53, 250participatory 160, 213, 215–21

Index

297


marketing chains 82–7, 85, 91–2see also markets

marketplaces see marketsmarkets 60–2, 84

classification 64, 69, 70–1, 73–6, 74, 75, 76location 63, 66, 68–9location models 68–70mapping 64–7micro-geography 63, 67, 76–8, 77seasonality 80–2, 80, 81size 73–6, 91species inventories 87–92, 87survey design 63–4, 66–7timing 63–4, 66, 78–80volumes sold 92–5see also marketing chains

Martin, G J, Ethnobotany: A MethodsManual (1995) xvi, 32, 46

marula (Sclerocarya birrea) 16, 17, 64, 82,196, 258

matrix population models 184–91matrix ranking exercises 25measuring individual plants 98

bark 108–11, 110biomass 102–4canopy 105–7diameter 98–100, 99flowers 107fruits 107–8height 100–2, 102, 103–4, 105leaves 104–5stem length 100–2, 102, 103–4, 105stem mass 111–13stem volume 111–13whole plant biomass 114, 114

medicines, traditional 25, 46, 48, 167–9‘mental maps’ 192, 217–21methods of survey 12–14miombo woodland (Brachystegia and

Julbernardia) 109, 139, 184, 197, 200monitoring systems 156–8mopane woodland (Colophospermum

mopane) 166–7, 168, 184, 215mukwa (Pterocarpus angolensis) 113, 185,

189–90, 189

Namibia 196, 197naming, plants 18–19, 44–8

see also taxonomy; vernacular namesnational parks xvi, xvii, 4, 194, 194, 222–3‘nodes’, settlement development 71

official data 14, 19open-access resources 239, 240, 245overexploitation 1, 2, 6, 8, 61, 72, 266

by destructive harvesting 132ebony 54

fuelwood 70slow-growing species 115, 264

palm-winesales 64, 65sap yields 183seasonality 16, 17, 64, 81–2tappers’ rights 241trade 94

palms 98ageing 118–21, 119, 122, 126leaf harvesting 133–4, 134, 134–5

PAME (participatory assessment, monitoringand evaluation) 12, 18

PAR (participatory action research) 23participant observation 28–30, 59participatory action research (PAR) 23participatory assessment, monitoring and

evaluation (PAME) 12, 18participatory mapping 213, 215–21participatory methods 10–12, 18, 23–8participatory rural appraisal see PRAPCQ (point-centred quarter) sampling 162People and Plants Initiative xvi, 7perishability, plant products 94permanent sample plots (PSPs) 146, 158, 160,

167, 169permit data 19–20, 21personnel xix–xx

see also resource usersphotographs 171–3, 172Phragmites reeds 157–8, 181–2place names 248–53plant architectural models 118, 151plant dyes 43, 174plant material, quantities used 53–9, 54plant populations, harvesting impacts 144–5,

147plot shape 161point-centred quarter (PCQ) sampling 162population growth, human 267–8population matrix models 184–91PRA (participatory rural appraisal) 12, 14, 18,

20–1, 48, 92, 263identifying stakeholders 260providing background information 146, 176studying resource conflicts 236

preservation of specimens 31prices 85–6, 91, 264, 266priorities for conservation 89–90priority species, choosing 145–6

‘filters’ 147–50Proteaceae 106, 118, 129, 130, 142, 153protected areas 5–6, 206–8, 209Prunus africana 11, 19, 20, 21, 36, 86PSPs (permanent sample plots) 146, 158, 160,

167, 169

Applied Ethnobotany

298


Pterocarpus angolensis (mukwa) 113, 185,189–90, 189

quantitative methods, species use assessment48–51

Quercus suber (cork oak) 39, 108, 137questionnaires 27–8, 29–30

rapid rural appraisal see RRArecovery rates 53–9, 55recycling harvested material 53–4Red List categories (IUCN) 205–6, 207–8religious practices 269–70reseeders 151–2, 153resin see exudatesresource management xv, 10–12, 15, 61–2

basic steps 265ecological influences 233–8priorities 89–90social and cultural influences 238–45see also conservation; tenure

resource users xiv, 10–12, 15–16, 228, 230assessing resources 50–1help with sampling 160–1as monitors 176–80, 178

resprouters 151–3, 153–4, 166–7resprouting 113, 143rights 234, 235, 242, 250

see also tenureritual 253–9, 255roots 34, 40–1, 140–2, 140, 141RRA (rapid rural appraisal) 12, 18, 48, 236rubber (Hevea brasiliensis) 131, 132

sal (Shorea robusta) 4–5, 8, 85, 184sampling

fruit fall 108markets 70–1plant populations 156–8, 159–65, 163, 164,

166questionnaires 29

sap see exudatessatellite images 196–202, 197, 220scales of study 192, 193scar counting 117–18, 119–21, 122scarcity 11, 67, 89, 91, 92scent 34, 35, 36, 43Sclerocarya birrea (marula) 16, 17, 64, 82,

196, 258seasonal calendars 23seasonal variations 16, 18, 213

markets 80–2, 80, 81palm-wine trade 16, 17, 64, 81–2

seeds, harvesting impacts 128–30sellers 67, 72, 76–8, 77, 82–4, 83, 86–7sense of place 246, 247sensitivity analyses 189–91

Shackleton, Charlie, fuelwood studies 20, 22,113, 181

shade-tolerance 154–6Shorea robusta (sal) 84–5, 85, 184short-term ‘snapshots’ 16, 30simulated defoliation 134–5size-class distributions 173–6, 173

for measuring harvesting impacts 174, 175,181

Skinner, hierarchical classification of markets73–4

SLA (specific leaf area) 154–6, 157Smilax anceps (ensuli) 58, 59social survey methods 12–14soils, sub-equatorial Africa 1, 2sound 36–7sources, background information 14species names 18–19, 44–8

see also taxonomy; vernacular namesspecies use assessment 48–51specific gravity 43specific leaf area (SLA) 154–6, 157specimens, preservation 31‘spinaches’ 22, 27, 132spreadsheet software 188stakeholders 231–3, 259–63, 260, 261stems

harvesting impacts 143measuring length 100–2, 102, 103–4, 105measuring mass 111–13measuring volume 111–13

sterile material 31, 32–3‘stick graphs’ 23, 24stochastic events see disturbance eventsStrychnos madagascariensis 55subjective allocation 49–50sundews (Drosera) 124–5survivorship curves 173–6, 173, 174, 175,

181sustainable harvesting 1, 1–2, 3, 5–6, 147–50

alternatives 158sweet-thorn (Acacia karoo) 117, 144, 145,

146Synarundinaria alpina 214–15, 216

tapping, exudates 130–2, 131taxonomy 44–8

using field characters 32–44, 35tenure 213, 222–3, 227, 230, 232, 238–40,

250characteristics 241–4, 245

tenurial niches 240–4territoriality 213, 227, 232, 234–5, 235territories 235, 257texture 34, 35, 36thatch grass (Cymbopogon validus) 157, 182threat, degree of 202, 205–8

Index

299


time lines 23toponymic surveys 248–53trade data 14trade networks 60, 61, 266–7

see also marketstraditional medicines 25, 44, 46, 48, 81,

167–9training, interviewers 30transect walks 23transects, belt 160–1, 161–2, 163transition matrices 184–8transport 62, 62, 91–2, 94–5tree-crowns 105–7, 130, 141, 142tree ferns 98, 118–21, 120tree ring counting 115, 116, 116–17tropical ecosystems xivtropical forests 8–9, 33tubers 122–3, 124–5, 169–71

Uganda 7, 8see also Bwindi-Impenetrable National Park

underground plant parts 40–1see also bulbs; corms; roots; tubers

units of measurement, plant products 51–3,52, 53, 91, 93–4

urbanization 71–2, 87user groups 11–12, 25–6

uses totalled method 50

vegetation change 7–9vegetation patterns 192, 195–6, 197–202vegetative characteristics 33vendors see sellersVenn (‘chapatti’) diagrams 25, 260–1vernacular names 14, 18–19, 31, 45–8, 217village mapping 25volume, measuring 102–4, 111–13volumes, plant material sold 92–5von Thünen land-use model 68–70voucher specimens 18–19, 30–2, 89

‘walks in the woods’ 16, 46, 213, 219, 221wealth ranking 25, 260–1wild plants xviii, 9, 267

uses xvii, 1sustainable use 3, 5–6, 147–50

woodcharacters 33–4, 36, 39, 41–4measuring volume 111–13

woodcarving 54, 113World Conservation Union see IUCN‘worldviews’ 247, 253

yields 180–4

Applied Ethnobotany

300


Date post:	22-Dec-2016
Category:	Documents
Upload:	anthony-cunningham
View:	219 times
Download:	2 times

Applied Ethnobotany: People, Wild Plant Use and Conservation

Documents