THE MANAGEMENT OF NON-WOOD FOREST PRODUCTS IN PROTECTED AREAS: LESSONS FROM A CASE STUDY OF MULTIPLE-USE IN BWINDI IMPENETRABLE NATIONAL PARK, UGANDA

A.B. Cunningham

1. Introduction

A primary goal of this CARPE meeting is to assess the potential role of non-wood forest product (NWFP) use in contributing to forest conservation. In this presentation, the example of the multiple-use programme at Bwindi Impenetrable National Park (BINP), Uganda, is used to illustrate the potential and pitfalls of processes leading to NWFP use within protected areas. As we all hope to conclude this meeting with a series of pragmatic "take home messages" that will help organisations and individuals interested in this issue move forward in policy and practice, I have structured this paper by stating a series of points (the "take home messages"), illustrating each one through examples from the BINP case-study or from other cases. In this process, I hope that this paper may sharpen the focus of the CARPE programme's goals of identifying the most promising forest products, the most promising conditions and exploitation systems which lead to forest conservation through NWFP use, rather than a situation where a policy of increased use of NWFPs leads to species-selective over-exploitation of the "most valued, most vulnerable" sub-set of NWFPs.

By taking Bwindi Impenetrable National Park and its Afromontane forest as a case-study, I am able to highlight the potential and many pitfalls of NWFPs which may not yet be evident in massive forested areas such as the Zaire basin or Amazonia where human population densities are 1-10 people/km². Secondly, it is located near the border of Uganda, Rwanda and the Democratic Republic of Congo (former Zaire) in a region characterised by political instability as much as by endemic species.

2. Location and historical background to the case-study

Bwindi Impenetrable National Park, while large (330 km²) by comparison to most remaining Afromontane forests (generally <1-20 km² in size) is still relatively small compared to lowland tropical forests and is surrounded by 100-320 people/km². As a result, due to high demand and lower stocks of some species, it provides a very useful case-study from which valuable lessons can be drawn so that NWFP policy is cautioned by practice. At the same time, it is important to recognise that the extensive and more sparsely populated lowland forests in the CARPE region of interest offer the greater opportunity for harvest due to large stocks and lower local (but not necessarily international) demand for NWFPs.

In 1990, a public inquiry around Bwindi Impenetrable Forest recorded strong local opposition to the creation of a national park, primarily because local people believed they would be deprived of the use of forest resources (Hamilton et al., 1990) so that "white people can watch gorillas" (Wild and Mutebi, 1996). Nevertheless, in 1991, Bwindi Impenetrable forest changed its status from a forest reserve to a national park. With this proclamation came a rapid change from easy access to forest resources during a 15 year period of turmoil in Uganda (during the Amin/Obote period), to the situation where access to any forest resources was stopped. By 1990, following recommendations made by Butynski (1984), the destructive activities of agricultural clearing, pit-sawing and gold mining in Bwindi Forest Reserve were stopped and in 1991 Bwindi Impenetrable forest was declared a national park. Consistent with national park legislation in many parts of the world, this change of legal status also precluded natural resource use within the boundaries of the national park. The result was an upsurge of public opinion against the new national park from the surrounding community. This was stimulated on one hand by loss of access to forest resources and on the other by some foresters being embittered by loss of revenue from illegal timber loggers and gold-miners in Bwindi forest. A consequence of this was increased fire damage through little or no community control of accidental fires or arson in the dry season of late-1991 and early 1992. Fire thus became a significant threat to parts of the national park. It was under these circumstances that networking and resource management planning at a local community level first began.

To some extent conflict is inherent in any protected area management programme due to the difference between the long-term goals of conservation and peoples shorter-term needs. These conflicts increase with higher human population densities, higher arable potential soils and with resource scarcity. Richard Bell (1987) points out with his characteristic clarity:

"Any programme that emphasises long-term communal benefits at the expense of short-term individual benefits will meet with resistance. The problems and costs of conservation are proportional to the extent of the conflict between these two sets of interests. For a conservationist programme to develop and survive without external enforcement, the benefits conferred must be real and they must not be long delayed" (Bell, 1987).

Multiple use, not just of NWFPs, but through other consumptive and non-consumptive uses plays an important part in this process. In 1992, a survey around BINP recommended a "middle-road" where use of high-value, low-impact resources (bee-keeping, medicinal plants and basketry fibre use) should be permitted within multiple-use zones (Cunningham, 1992). In theory, two approaches could be taken. First, not to allow any use of wild plants by local people until potential sustainable yields had been determined. However, this was not practical given the diversity of species and short time needed before decisions had to be made on the people/protected area conflict. The second option is to take an adaptive management approach, allowing harvesting and monitoring the response of harvested populations as described by Walters (1986). In the Bwindi Impenetrable National Park case, we took an approach mid-way between these two options. Neither the money, the time nor the manpower were available for determining productivity or sustainable yields of all harvested species. Equally, it would have been counterproductive to repeat over-exploitation problems experienced elsewhere in Africa with building poles (Hall and Rodgers, 1986; Muir, 1990), craftwork resource or medicinal plants (Cunningham and Milton, 1987; Cunningham, 1991) or fuelwood (Leach and Mearns, 1989). If harvesting is not sustainable, then it is a false solution providing brief respite from land-use conflict by putting off the real solutions to the problem.

For this reason, the CARE Development Through Conservation Programme (DTC), which supports the Uganda Wildlife Authority (UWA) in implementing this integrated conservation and development project (ICDP), has placed an emphasis on providing viable alternatives to high volume uses such as fuelwood, building materials and bean-stakes outside the national park through agroforestry and rural development programmes (Wild and Mutebi, 1996). The UWA and the Uganda Forest Department recognised that the needs of neighbouring communities have to be considered and met, if protected areas are to have a long term future. Methods developed to meet these needs include, the sharing of tourism revenues, development activities in park adjacent areas, conservation education and resource utilisation.

3. "Take home messages"

Point 1: When NWFP use forms part of a multiple-use (and conflict resolution) strategy between local communities and protected areas, we must seek a "middle-road" between the "fences and fines" preservationist approach and "conservation" programmes driven almost solely by short-term benefits through "people's participation" which take no account of the local peoples resource needs in the long-term. A short-term approach may provide a temporary "bonanza" of resources now scarce outside the protected area, but it will undermine the primary goal of any protected area, i.e. the maintenance of habitat and species diversity. The common ground is that if a valued but vulnerable resource is overexploited, local people and conservation both lose out.

Over the past few decades and particularly since the late 1980s, there has been a strong move away from a "fences and fines" preservationist approach to protected areas to one stressing sustainable use and community development. This broader approach is evident in the different IUCN categories of Protected Areas which were developed in the mid-1980s and recently modified at the IV World Congress on National Parks and Protected Areas in 1992. In part, this change in approach to conservation came through the recognition that conservation attempts to maintain the integrity of protected areas by excluding all types of local community interaction and use has largely been unsuccessful (Wells and Brandon, 1992). In some cases, the pendulum has swung so far towards the local community/rural development side that Katrina Brandon (1997) recently wrote that "the majority of conservation programmes are in fact large and complicated social programmes" and that "the attention to sustainable development and poverty alleviation, while important, does not address environmental issues. If anything, these concerns broaden the agenda and dilute the message". Oates' (1995) cautionary tale from a forest conservation programme in Nigeria is a good example of this.

We need to be able to look beyond the smokescreen of "bio-politics" and untangle the complex interplay of ecological, political, religious, economic and social undercurrents behind successful or failed examples of resource conservation. There has been too much generalisation on a range of very diverse and dynamic situations. For every claim that "rural people have sophisticated systems of natural resources management which have maintained biodiversity for thousands of years" and that "people living adjacent to protected areas have found themselves deprived of resources which for thousands of years they had a right to utilise" (IIED, 1994), there are cases where local people have destroyed high diversity habitats, or where the people living adjacent to protected areas are recent migrants.

The recent pendulum swing back to "preservationism" called for by several prominent expatriate biologists who have worked in the tropics (Kramer, van Schaik and Johnson, 1997) is unworkable anywhere in tropical Africa. Reasons for this are that protected areas need to be maintained under circumstances of political turmoil, lack of funds, changes of government and a "brain drain" of senior park staff - with expatriates often those who leave first and fastest. As Jonathan Kingdon (1990) points out:

"...the realities of power are exactly the opposite to those perceived by most of the participants of this struggle to conserve key areas of high endemism and biodiversity because the long-term future of Africa's Centres of Endemism lies with local peasantries rather more than with transient governments or enthusiastic conservationists, yet locals seldom receive the respect that is generally accorded to those that wield power. Meanwhile, both populations and resentments grow.....The conservationists' answers should not lie in propaganda campaigns, which are generally seen for what they are, but in a shared growth of knowledge and debate. The minimal demands of local communities will include sustained, not ephemeral programmes of action in which their own people can find meaningful, decisive and dignified roles".

Several forest conservation areas in the borderlands of western Uganda/eastern DR Congo (formerly Zaire) and north-west Rwanda, including BINP are prime examples of this. Despite the fact that three of these forested national parks form the stronghold of one of the "ultimate" flag-ship species for international conservationists (the Mountain Gorilla (Gorilla gorilla berengei) and BINP is the first African protected area to receive funding through an international Trust Fund, it only has 30 park staff for an area of 330 km² for a national park surrounded by 100,000 people living in immediately adjacent parishes. Under these circumstances, total protection is not possible.

A basic principle behind multiple-use (and NWFP use) is to help off-set some of these lost opportunity costs and better justify conservation as a form of land-use. In principle, benefits need to be directed to those living closest to the protected area. In most cases, these are the people who are most affected by crop raiding animals and loss of access to plant resources inside protected areas. This is well illustrated by the household surveys by the CARE-DTC project recording the number of respondents from communities adjacent to the Bwindi Impenetrable National Park collecting forest products, pit-sawing or affected by crop-raiding animals prior to park closure, compared to those away from the forest (Wild and Mutebi, 1996).

Benefit sharing, including that from eco-tourism revenue should take place through local community institutions set up for this purpose, which should be representative of the communities and of resource users. This is often easier said than done. Protected areas are often located in more remote areas where access to literacy skills is limited. In many cases, resource users are from a sector of the local community with the least economic or political power. For these reasons, local resource users are generally not well represented, even at the lowest level of formal local government, although they may be highly influential members of their own communities. In addition, the administrative boundaries that form the basis for local government within the nation state rarely conform to the territorial boundaries of local communities. This may further skew the relationship of who "represents" communities surrounding protected areas.

Point 2: We need to better use the predictive ecological, anthropological and economic tools we have to avoid situations where resource over-exploitation and conflict develop despite (or because of) good intentions which then lead to worsened circumstances for local communities and for conservation.

We have the advantage of hindsight from successes and failures in many conservation and development case studies, not only in forests, but also in savannah woodlands and grasslands. We also know that sustainable use of NWFPs depends as much on a predictive understanding of the biological component as it does on the social and economic aspects of NWFP use. We have many of those predictive tools based on key issues drawn from ecological, social and economic studies of tenure, trade and NWFP use or abuse. It is essential that we use these tools to avoid cases where good intentions catalyse resource depletion. While sustainable harvest of any NWFPs is possible in theory, it is often more complex than people think. What is often glossed over is that high conservation priority habitats, with a high species diversity and slow-growing, habitat specific species require a level of management of an intensity that is not possible with the economic constraints that are a feature of many conservation departments.

For all interest groups, whether resource users, rural development workers or national park managers, it is far better to have pro-active management and to stop or phase out destructive harvesting in favour of suitable alternatives before over-exploitation occurs, than to have the "benefit" of hindsight in the midst of a devastated resource. Marilyn Hoskins (1990) puts this well in her paper on forestry and food security:

"All research and management by outsiders must remember that their activities come and go, but food security -- land and resources surety -- is a long-term, life and death issue for rural peoples".

Point 3: The complexities of implementation increase exponentially with increasing numbers of species coupled with a high number of resource users.

The culture/nature interface of Afromontane forests and surrounding farming communities is a very different and more complex situation to resolve. Firstly, resource sharing is focused on a wide range of non-wood forest products and on eco-tourism revenues from gorilla viewing rather than hunting. Secondly, because of their structural complexity and higher species diversity, forests provide a very different situation to harvesting of plant resources in several other African protected areas. Muir (1990), for example, working with local wood-cutters in Afromontane forest in southern Africa, has demonstrated that cultivating alternative sources of building material outside indigenous forest can be over ten times cheaper than the cost of an intensive monitoring programme for sustainable use of that resource.

In savannah parks in southern Africa, harvesting commonly focuses on plants from wetland or disturbed grassland situations. Examples are harvesting of Phragmites australis reeds and Cymbopogon validus thatch grass from wetlands or disturbed grasslands. Unlike forests, these vegetation types have a wide distribution, low species diversity and high biomass production of annual stems which are resilient to harvesting for hut-building purposes (Cunningham, 1985; Shackleton, 1990). Managed harvesting of reeds and thatch-grass is also facilitating late autumn and winter cutting when disturbance to nesting birds is minimised. Similarly, at a species level in African savannah, harvesting usually applies to common, fast-growing thatch or encroaching woody species such as Acacia karroo, Acacia nilotica, Dichrostachys cinerea (Fabaceae) and Euclea divinorum (Ebenaceae) which are harvested as an aid to management objectives for the savannah parks.

In contrast to reeds and thatch species, data on abundance, productivity and population biology of Afromontane forest trees, even those of major economic importance, are limited. This applies even more to the hundreds of species representing "minor forest products". Concern over loss of access to wild forest plant resources is an important local issue as wild plants provide craft and building materials, fuel, medicines, food supplements or are a source of honey to people around the forest. The question is, once resources have been identified, how does one decide whether uses are sustainable or not ? In the Bwindi case-study, based on ethnobotanical surveys of the forest, local markets and households, wild plant resources were divided into three categories:

First, a low impact, high value category, where impact is low due to harvesting of small volumes of plant material by specialist users, particularly where leaves, fruits or flowers are used. This includes non-commercial harvesting of medicinal plants by traditional healers and midwives or for veterinary medicines, occasional felling of the secondary forest tree Polyscias fulva by traditional blacksmiths, harvesting basketry materials or bee-keeping. All of these uses have low biological impact but high social value through this harvesting to a large sector of the community.

Meetings with bee-keepers have led to mutually accepted regulations printed on Bee-keeping Society Membership cards requested by this forest user group. To date, nearly 500 bee-keepers in four parish bee-keeping societies are registered to keep an estimated 3 000 hives within multiple-use zones. In addition, the DTC project is assisting bee-keepers with processing and marketing of surplus honey. Initial ethnobotanical surveys with traditional healers and basket-makers have been followed up with PRA work in three pilot-study parishes, leading to harvesting of selected species. Each of these activities falls within Joint Forest Management (JFM) agreements developed within each of these parishes and signed by the Forest Societies and Uganda National Parks (UNP), including access to certain footpaths through the forest and to a hot spring site considered to have spiritual healing qualities.

Second, a monitoring adaptive management category, where subsistence demand was high relative to supply or where species selective commercial harvesting was beginning to develop. This included seasonal and rotational management to specialist harvesters of the montane bamboo, Synarundinaria alpina, a slow growing forest climber Loesneriella apocynoides (Celastraceae), medicinal plant species such as Hallea rubrostipulata (Rubiaceae) whose bark is gathered on a small-scale commercial basis, and the secondary forest tree Rapanea melanophloeos (Myrsinaceae), which is used by woodcarvers.

Third, a substitution category, of continued closed access to resources where sustained use is not possible due to either complexity, high demand or slow growth rates where the emphasis needs to be placed on providing alternatives outside the national park.

This was considered to be the case for some wood uses (beer boats, bean stakes, building poles, fuelwood) due to the combination of past impacts, high demand and limited staff for monitoring or management of multiple-use species in an uneven-aged, high diversity forest. To foresters, whose objective is hardwood timber production, saplings of "reserved species" represent regenerating timber trees. To people from local rural communities they also represent an important source of beer boats (>50cm diameter at breast height (dbh), building material (5-15cm dbh) or bean poles (1.5-5cm dbh) with high density wood favoured due to greater resistance to borer attack or fungal infection. These species are not only an immediately useful resource to local people. However, they also represent the future forest canopy of the next century which needs to be conserved if the forest is to survive.

Provision of alternatives to harvesting of beer boats, bean stakes and building poles was recommended on the basis of their widespread use, the high volumes used, the focus on hardwood species and problems that have arisen in other Afromontane and coastal forests (Cunningham, 1992). It was also recognised that the success of forest conservation and community relations strongly depends on how effectively tree cultivation is implemented. Shortages of fuelwood, building poles and bean stakes are being experienced in the DTC area, and a shortage of large trees for beer boats can be expected in the future. Reasons for wood scarcity, and solutions to the problem are recognised by local people. Cultivation of trees is widely practised in the DTC area already. Elephant grass (Pennisetum purpureum) and trees (particularly Eucalyptus) are also planted for bean stakes, while Ficus cuttings are planted for beer boats. In at least one area, even hardwood timber trees have been planted, with Entandrophragma (Meliaceae) reaching a diameter of 90 cm within 40 years. Such local initiatives need to be recognised and encouraged.

In a recent survey conducted in the DTC area, for example, Eucalyptus (88% of respondents) and Acacia mearnsii (49%) were the species most preferred for building and had respectively been planted by 77% (92) and 36% (43) of respondents (Kanongo, 1990). From field observation, it is clear that many homes in the DTC area are built from these cultivated tree species (particularly Eucalyptus), with the use of exotic species increasing with distance away from the forest. Over 125 farmer-tree nurseries have now been developed. Guided by local farmer preferences, the focus has been on Eucalyptus woodlots and the agroforestry species Sesbania sesban, a nitrogen fixing species useful for bean stakes. In addition, two indigenous tree nurseries have been maintained, with 5 000 trees distributed in the past year. Local Conservation Extension Agents (CEA's) employed by CARE are also assisting local farmers with climbing bean variety trials, banana plantation management, soil conservation and vegetable growing with the aim of sustainable agricultural production.

Point 4: It is crucial that any forest conservation/NWFP harvest programme takes the specific needs and roles of the pygmy peoples, who are the ancestral inhabitants of these forests (e.g. Baka, Mbuti, Batwa), into account.

There are several reasons for this. Firstly, it is crucial to work with them as an important forest user group. Many pygmy communities are not only heavily involved in the bush-meat trade, often in a "protein for starch exchange" with agriculturists, but they are also involved at the lower end of the commercial marketing chain for many timber loggers and for NWFPs such as commercial collection of Pausinystalia johimbe bark, Gnetum leaves and forest tubers (e.g. Dioscorea yams) and forest fruits (e.g. Irvingia, Ricinodendron heudelotii etc.). Secondly, throughout the region, they also place great cultural and religious value on some species which can exceed the barter or commercial value of those NWFP species (e.g. Dioscorea). Thirdly, this is an issue of cultural survival. In the Bwindi Impenetrable National Park base, the Batwa (or Abayanda people as they call themselves) have faced an historical situation where encroachment of Bantu agriculturists over many centuries (Taylor, 1993) has cleared the forest and built up to a situation today where there is strong social pressure against the Batwa having access to farmland. At the same time, the game resource that was a major barter source has been over-hunted, timber overexploited and access to large areas of the forest for harvesting purposes lost. Multiple-use has taken many of these needs into account such as use of stingless bees, yams and forest fruits. Although extensive research has been done in the lowland tropical forests of the CARPE region (e.g. by the University of Kyoto, Hart and Hart, 1986 and others), the conservation and NWFP harvesting programmes that consider the specific needs and roles of pygmy communities need to be taken into account to a greater extent.

Point 5: Most NWFPs with commercial value are already traded by local people who have access to markets - but some "wild cards", not recognised by local people as having commercial value, exist which have high potential for international trade with "green marketing" premium prices through more direct marketing. This can play an important role in raising living standards for communities in or adjacent to the forest, but it is crucial that strong tenure systems and simple, robust monitoring programmes are established prior to large-scale marketing.

Examples from the Bwindi Impenetrable National Park case which were identified in the initial 1992 survey we did were Carapa grandiflora and Allanblackia stuhlmannii, which are both sources of oils/vegetable fats. The same two genera occur in the CARPE region of interest. In the Amazon, Carapa seed oil is an important medicinal which is commercially harvested and in the Eastern Arc region of Tanzania, Allanblackia fruits have been commercially harvested for many years for soap making. However, both of these genera are classic cases of tropical trees which bear relatively few, yet very large fruits where recruitment into the "seedling bank" can be seriously disrupted by commercial harvesting (See, for example, Peters, 1994).

Point 6: Too many NWFP utilisation programmes assume that sustainable harvesting will take place and do not take the costs of monitoring into account. Monitoring the success of multiple-use programmes is essential but, for this to be sustainable, monitoring programmes must be robust, pragmatic and cheap. Before any monitoring starts, it is crucial to decide on what is to be monitored, at what scale (landscapes, populations, individuals), at what levels of detail and by whom. Once a few key species have been selected for population level monitoring, then it is important to set up permanent plots for long-term monitoring.

When Bwindi Impenetrable National Park was a forest reserve, the 1964 Ugandan Forest Act allowed controlled harvest of forest plant resources, including "minor forest products". This was based on a permit system intended to generate government revenue and facilitate monitoring. This system broke down due to political and economic upheaval. During the period of political turmoil in Uganda, the buying power of forestry staff salaries declined dramatically (Howard, 1991). As a result, control systems for hardwood and "minor forest product" harvesting broke down. By 1983, an estimated 140-280 people were involved in pit-sawing, and an estimated 100-200 were people involved in panning gold from river-valleys inside Bwindi forest (Butysnki, 1984). As a result, an estimated 10% of the forest reserve remained intact, 61% had been heavily exploited by pit-sawyers and 29% "creamed" of the best hardwoods by selective pit-sawing (Howard, 1991). In addition, Butynski (1984) estimated that 10-20 people a day entered the forest daily for bee-keeping purposes or to hunt for wild bee-hives, whilst 25-50 people went to collect fuelwood, bamboo and building materials.

In many countries with high biological diversity, adequate control is too expensive for the State. Effective in situ conservation for black rhino, for example, would cost $400 per km² (Martin, 1993). Based on his experience in East Africa, for example, John Hall suggests that patrolling of forest reserves generally requires two forest guards per 500ha (or four guards per 10km²) (Hall, 1983). In most cases, neither this level of funding nor staffing are available. Situated in rugged, forested terrain, for example, Bwindi Impenetrable National Park, 330km² in extent, has 30 patrol rangers, rather than the 130 guards deemed necessary by Hall (1983). Implementation of a sustainable logging programme requires even more staff. In Afromontane forest in South Africa, which has a relatively low species diversity and where only a single product (timber) and few species are involved, a marking team of one forester and only two staff selecting trees >30cm dbh is only able to cover 5ha per day (Seydack et al, 1995). For these reasons alone, it is necessary to design programmes for protected areas or the land around them that takes local institutions, tenure and resource management systems into account.

Point 7: A large component of "resource management" is actually about "people management", often dealing more with human behaviour than with resources. For good science to become good management requires wide social acceptance of management plans or regulations. Low tenure and high NWFP prices, coupled with potentially destructive harvest (bark, root, tuber) use can spell disaster. Low prices with secure tenure mean little incentive to conserve for the long term. The challenge is to design and develop a situation where there is both secure tenure and high price coupled with a practical resource management plan.

Secure tenure is an important ingredient for resource management and conservation. However, whether you are working in areas that are privately owned, in a national park belonging to the state or in a communal area, it is crucial not to fall into the trap of simplistic assumptions or simple classification of different types of tenure. This is particularly important in much of Africa (and in many other sub-tropical and tropical regions) where local people or indigenous communities have established complex rules controlling access to communal land, which in turn has been overlapped by State control (as in the case of many national parks or forest reserves). It is also important to recognise that land tenure and resource tenure may be very different.

At BINP, key steps to establishing a mind-set of more secure tenure amongst local resource users were the demarcation of multiple-use zones based on agreed and clearly recognisable boundaries (Scott, 1992); the identification and nomination of individuals to small, locally based user groups (basket-makers, bee-keepers, herbalists etc.) in a process of community

BOX 1. Setting priorities for harvested species based on ethnobotanical surveys

participation, the provision of licenses to those individuals who were recognised by the national park authorities, and the signing of a written MoU between the communities and Uganda Wildlife Authority (UWA) which clearly spelt out the mutual obligations of the community and UWA (Wild and Mutebi, 1996). These principles are common ingredients for successful community based management as well (See Box 1).

Eight years ago, at a meeting of several hundred villagers, the chairman of Mpungu Parish presented a letter on behalf of the community expressing their opposition to the national park and asking why "innocent Ugandans should be made to suffer by creating a National Park" (Hamilton, et al., 1990). In April 1994, the same parish chairman in a speech to the Director of Uganda National Parks, made during the signing of the memorandum of understanding, said: "Mr Director, allow me to express our sincere and greatest appreciation to the new concept of UNP. Allowing us an area of the park to collect ensuli (Smilax kraussiana) and [medicinal] herbs . . . We were strongly and bitterly opposed to the idea of Bwindi becoming a national park, but as I say, this attitude has completely changed. We have made a complete turn from negative to positive . . . We used to refer to the national park as 'their' forest; now we call it 'our' national park".

Point 8: Inventories are a very important first step to priority setting. They are also part of a networking process which is ideally conducted by a multidisciplinary team, or more often in practice by researcher and local research participants who are aware of ecological, economic and local cultural/institutional issues.

Lowland and montane forests are a dynamic and diverse mix of species and life-forms with a multiplicity of users, so too are rural communities with their complex mix of political and socio-economic hierarchies, user groups, jealousy and hidden agendas. The higher the species diversity, the higher the number of wild plant uses and users would be expected. The first step to unravelling this complexity is fieldwork with resource users in forests and the surrounding community, enabling identification of both species and resource users groups. This is as much a method as a process towards credibility and communication at a local level.

Inventory surveys of plant or animal species are often the first step in identifying unique biological components within protected areas. Conservation managers do not have the luxury of time. Facing the pressures on protected areas they have to make decisions. Skilled biologists and taxonomists are a scarce resource in most tropical countries. Uganda is no exception. Neither is most of the CARPE region. Under these circumstances, folk taxonomic knowledge can be invaluable in inventory work, followed up through identification of voucher specimens. In this case, inventory of plants or edible insects in the forest with knowledgeable local people was combined with discussion on ecological or social issues relating to these resources or to sites of cultural importance. Separate discussions were held with specialist users groups (such as bee-keepers, midwives, bamboo basket-makers) or Batwa people on species used, species most favoured and whether resources were available or not outside the national parks (Cunningham, 1992; Cunningham et al., 1993; Scott, 1992).

Local communities often have good reason to be suspicious of conservation motives. In the past, conservation has often meant evictions, resource loss and harassment with few positive benefits to the community in general. Networking to develop credibility in the eyes of local leaders can be extremely important in this process, requiring time, transparency, patience and constancy. Community members are very good judges of human nature. In the minds of community members are questions such as : What attitudes do park managers have? How much do they understand of the resources themselves? In the Ugandan case-study, it took several sessions over nine months to establish that the communities of Mpungu Parish adjacent to Bwindi Impenetrable National Park had most confidence in the Stretcher Societies as key organisations forming Forest Societies (Wild and Mutebi, 1996).

Point 9: Although predictive models (Point 2) enable us to attain a "first approximation" level of what is likely to succeed or fail, multiple-use (and NWFP use) programmes need to be developed on a case-by-case basis. They CANNOT be developed on a "recipe book" basis from other socio-economic or ecological circumstances.

Managed use of resources within national parks or in buffer zones around them has become a widespread strategy to defuse land-use conflicts. As early as the 1940s, resource sharing arrangements were started in Africa to address some of these lost-opportunity costs (Bell, 1987). Forty years later, these are now termed Integrated Conservation and Development Programmes (ICDPs) (Wells and Brandon, 1992). Although the concept is well established, Wells and Brandon (1992) found almost no examples of resource use within buffer zones on the edges of the protected areas in their survey of the 23 most promising ICDPs. Nevertheless, a few ICDPs have been well documented in Zimbabwe and Zambia (Martin, 1986; Lewis et al., 1990). These examples are primarily from savannah regions with high game biomass, low densities of people, low arable potential and, in several cases, the occurrence of tsetse flies. In these savannah regions, the major focus has been on benefits derived from tourism, trophy hunting or venison from culling operations (Martin, 1986; Bell, 1987).

Innovative, decentralised approaches to conservation outside protected areas also have a way of catching on and spreading. Three examples are CAMPFIRE (Communal Areas Management Programme for Indigenous Resources) in Zimbabwe (Martin, 1984; Child, 1996), the Luangwa valley in Zambia (Lewis, Kaweche and Mwenya, 1990) and Joint Forest Management (JFM) Programme projects spread across India and Nepal (Poffenberger et al., 1992; Fischer, 1995). Although small, and started in isolation, these programmes have built up experience and common ground that has been more widely applied - sometimes in problematic socio-economic and ecological circumstances.

Figure 1. Sign protecting community agroforestry woodlot (Photo: T. Sunderland).

The Mount Kilum example in Cameroon is a local example of the blind application of the JFM recipe from Nepal to Afromontane Africa. Surrounded by high densities of local people and with very different ecological circumstances to the JFM areas of Nepal or India, the montane forests of Cameroon are home to 23 endemic bird species, making them a major African bird conservation priority site (Collar and Stuart, 1988). Several rare species of barbets, woodpeckers or hornbills are deadwood dependant species, with populations limited by nest-site availability. Claimed as a successful case of JFM (Fischer, 1995), the agreement reached at Mt. Kilum through participatory approaches with the local community for deadwood harvesting from this small and extremely important Afromontane forest reserve promises to be a conservation disaster. The JFM agreement has not only led to the collection of fallen deadwood, but also to the felling of very large trees such as Nuxia congesta, some in excess of 70cm diameter at breast height (dbh)). This represents a major removal of an ecological niche providing nest sites or shelter for rare birds, small mammals and reptiles in a forest reserve. (Predictably on the basis of existing information, it would have been far better to put effort into identifying and supporting the provision of alternative sources of fuelwood sources on farm or from fuel efficient stoves). Once the "Pandora" of the JFM agreement with the local community is out of the box, it is very difficult to retract such agreements and get them back into the box again.

Point 10: Predicting the sustainability of harvesting requires an assessment both of the biological factors influencing resilience or vulnerability to harvesting and of the socio-economic factors that drive demand. Ethnobotanical surveys of local markets are a very useful step in this process.

From a resource management perspective, there are several reasons why the marketing and sale of wild plants should be the focus of ethnobotanical surveys. First, commercial trade or barter reflect demand. If demand for a species or resource category (such as fuel, basketry fibre, herbal medicine) is high, then these species or resource categories will be sold in many marketplaces. Conversely, a species or category of plant in low demand would be less common in marketplaces.

The most useful species will be frequently sold by more sellers in many more markets than species for which there is little demand. Systematic market surveys therefore provide a useful way not only of classifying the species on sale, but also of arranging them into hierarchical levels which reflect their relative popularity and usefulness. However, some of the most useful and popular species no longer feature in markets, due to over-exploitation. Second, price reflects resource supply in relation to demand. Locally common species are rarely sold in local marketplaces unless it is bulk-sale for processing or retail elsewhere.

When a popular species is scarce, whether due to geographical distribution or to over-exploitation, then trade occurs from resource-rich areas to the places where there is demand but little or no supply. As scarcity increases, so does the price. When alternatives are not available, the higher the price, the greater the incentive to go further and further afield for a scarce species. Improved roads and cheaper transport reduce this cost. As a result, internal marketing systems change in two ways, each shortening the marketing chain. First, cheaper transport enables rural people to get to larger centres to sell their products. Second, better roads improve the access that outsiders have to more remote plant resources. Outsiders frequently have more buying power than local people in remote, resource-rich areas. If this takes place and resource tenure starts to break down, then this hastens the scramble for resources in high demand.

Ethnobotanical surveys of local markets provide a means through which we can filter out international or local priority species in the series of steps listed below:

BOX 2. Influences on successful community-based natural resources management: (CBNRM)

Point 11: When we look at ecological impacts, we must see beyond the individual plant level and understand impacts and conduct monitoring at the plant population and forest dynamics/forest system level. We also need to be very selective in choosing which plant populations are monitored.

In diverse and dynamic habitats, with many hundreds of species harvested, yet with limited funding and time, we have to carefully choose where to focus detailed studies at a plant population level. The first steps in this process are to short-list the species which are most valued and used in greatest quantity or enter commercial trade. In terms of resource management and monitoring, the plant resource categories of greatest concern are cases where destructive harvesting is taking place, particularly where species are scarce, slow growing, habitat specific and where roots, bark, stems or the whole plant are harvested. Conversely, this process also highlights species which are likely to be most tolerant of continued harvest and those which are unpopular, rarely used and would be less of a priority for quantitative work at a species population level.

Although the response of individual plants to harvesting impacts provides useful information, it is crucial to avoid getting side-tracked when we see destructive harvest at the individual plant level. Harvesting impacts need to be seen from the perspective of the population dynamics of that particular species. Harvested plant populations in turn need to be viewed in terms of how they are influenced by disturbance and succession.

Point 12: Multiple-use programmes in protected areas need to take ecological impacts into account.

This issue is most evident in Afromontane and coastal forest protected areas, which are small and are surrounded by high numbers of rural farmers. High volumes of hunting or deadwood removal both have complex ecological impacts. In the Bwindi Impenetrable National Park case, fuelwood use was not allowed within the multiple-use zones, but was a candidate for development of alternatives (on-farm agroforestry production, fuel-efficient stoves) outside the protected area. The ecological impact of dead tree felling at Mt. Kilum in Cameroon is a good example of a case where this has been ignored (See Point 10 above).

Point 13: Training and employment of people from communities around the park area are very important, providing a real benefit from the existence of the park and a resilience in terms of staffing in times of political turmoil.

This has not been sufficiently catered for in many African protected areas, including Uganda. This point is best made by the following quote :

"...that vehicles, buildings, and short-term consultants supported by large multi-nationals do not make a conservation project. Instead, conservation is achieved by people with commitment. Project personnel recruited from the local population who demonstrate qualities of leadership and commitment, who receive regular hands-on training that empowers them to take responsibility for the management of their natural resources, are the formula proven to sustain long-term conservation efforts under difficult conditions. The combination of a few dedicated individuals, together with the support of a non-governmental organisation (independent of political constraints) with a long-term commitment to conservation, is the best recipe for achieving lasting success in countries where political stability is in question, or perhaps anywhere" (Hart and Hart, 1997).

Two Central African examples highlight the need for training hand-picked local people in protected area management. One of the strongest tests of conservation strategies is how resilient they are to the chaos of civil conflicts. Recent tests of this stem from conservation areas in Rwanda and Zaire (now Democratic Republic of the Congo) engulfed by conflict (Hart and Hart, 1997; Fimbel and Fimbel, 1997). These Central African examples highlight the crucial need for appropriate training for hand-picked local people at various levels (rangers, technical staff, research professionals and managers) to take responsibility for conservation programmes. International non-governmental organisations have a key role in this process, and one of these is to support this training process. In both cases, international funding was disrupted and ex-patriate staff left or were evacuated due to conflicts in or around the Nyungwe Forest Conservation Project in Rwanda and four World Heritage Sites in Zaire. What maintained these conservation areas during these conflicts were local people connected to these projects. The important lesson from both cases is summed up from the Rwandan case, where Nyungwe forest, an Integrated Conservation and Development Project (ICDP) and a priority area for conservation was held together in the face of lawlessness and land-grabs. Four local people with exceptional leadership qualities continued to collect and safeguard project records and liaise with people neighbouring the park and local government representatives. Of 45 local staff, all from villages bordering the conservation area, 40 remained, continuing to undertake forest patrols without salaries or communications from former supervisors or senior staff who had fled.

4. Conclusion

The CARPE programme is working in a large region with many challenges. This region is also one which really requires practical support to a region of great importance for forest conservation. One of the advantages that a relatively new initiative like CARPE has is that it can learn from the research, conservation and development lessons learned from other African forest conservation programmes as well as from tropical forest research by ecologists, anthropologists and economists in other parts of the tropics. I sincerely hope that the programme will succeed and lead to long-term support for forest conservation in this region.

Acknowledgements

I would like to thank the CARPE programme and USDA for funding my travel to this meeting and in particular Terry Sunderland and Mark Buccowich for the invitation to attend. This overview of lessons learned from the BINP case-study has been possible due to support from the CARE-DTC programme, which in early 1992 commissioned me to carry out the survey and make recommendations for multiple-use around the national parks, and subsequent support from the WWF/UNESCO/Kew "People and Plants Initiative", which has funded Ugandan students at the Institute for Tropical Forest Conservation at BINP. In particular I would like to thank Jacob Bandusya, Z. R. Bukenya, Dominic Byarugaba, Onesimus Muhwezi, Maud Kamatenesi and Alan Hamilton for company and participation in stimulating discussions in the field over the past six years.

References

Bell, R.H.V. 1987. Conservation with a human face: Conflict and reconciliation in African land-use planning. pp. 79 - 101. In D. Anderson & R. Grove (eds) Conservation in Africa: People, Politics and Practice. Cambridge University Press.

Butynski, T M. 1984. Ecological survey of the Impenetrable (Bwindi) forest, Uganda, and recommendations for its conservation and management. Unpublished report to the Ugandan Government.

Child, B. 1996. The practice and principles of community-based wildlife management in Zimbabwe: The CAMPFIRE programme. Biodiversity and Conservation 5 : 369-398.

Collar, N.J. & S.N. Stuart. 1988. Key forests for threatened birds in Africa. International Council for Bird preservation Monograph no. 3, Cambridge, IUCN/ICBP.

Cunningham, A.B. 1985. The resource value of indigenous plants to rural people in a low agricultural potential area. PhD thesis, Faculty of Science, University of Cape Town.

- 1991. Development of a conservation policy on commercially exploited medicinal plants: A case study from southern Africa. pp. 337 - 358. In Conservation of medicinal plants, edited by V. Heywood, H. Synge & O. Akerele. Cambridge University Press.

- 1992. Park, people and plant use: recommendations regarding multiple-use zones and development alternatives around Bwindi Impenetrable National Park, Western Uganda. Report for CARE-International, Kampala, Uganda.

FAO. 1990. The community's toolbox : the idea, methods and tools for participatory assessment, monitoring and evaluation in community forestry. Food and Agriculture Organisation (FAO), Rome.

Fimbel, C & R. Fimbel. 1997. Conservation and civil strife. Rwanda: the role of local participation. Conservation Biology 11 : 309-310.

Fischer, R. 1995. Co-management of forests. IUCN, Gland, Switzerland

Hall, J.B. & A. Rodgers. 1986. Pole cutting pressure on Tanzanian forest. Forest Ecology and Management. 7: 57-66.

Hamilton, A, J. Baranga & J. Tindigarukayo. 1990. Proposed Bwindi (Impenetrable) Forest National Park: Results of a public enquiry and recommendations for establishment. Unpublished report.

Hart, T.B. & J.A. Hart. 1986. The ecological basis of hunter-gatherer subsistence in African rainforests: The Mbuti of Eastern Zaire. Human Ecology. 14(1): 29-53.

- 1997. Conservation and civil strife. Zaire : new models for an emerging state. Conservation Biology 11 : 308-309.

Howard, P C. 1991. Nature Conservation in Uganda's Tropical Forest Reserves. IUCN, Gland, Switzerland.

IIED. 1994. Whose Eden? An overview of community approaches to wildlife management. International Institute for Environment and Development (IIED), London.

Kanongo, M. 1990. Draft report of baseline data, report for Development through Conservation project. Phase 1. Dec. 1989 - Feb. 1990. Development through Conservation project, Ikumba, Uganda.

Kramer, R, C. Van Schaik & J. Johnson (eds). 1997. Last stands: Protected areas and the defense of tropical biodiversity. Oxford University Press, New York.

Leach, G & R. Mearns. 1988. Beyond the fuelwood crisis: People, land and trees in Africa. Earthscan Publications, London.

Lewis, D., G.B. Kaweche, & A. Mwenya. 1990. Wildlife conservation areas outside protected areas - lessons from an experiment in Zambia. Conservation Biology 4 : 171 -180.

Martin, R.B. 1986. Communal areas management programme for indigenous resources (CAMPFIRE). CAMPFIRE Working Document No. 1/86, Department of National Parks and Wildlife Management, Zimbabwe.

Milton, S.J. 1987. Effects of the basket weaving industry on the mokola palm (Hyphaene petersiana) and on dye plants in NW Botswana. Economic Botany 42 : 386-402

Muir, D.P. 1990. Indigenous forest utilisation in KwaZulu: a case study of the Hlatikulu forest, Maputaland. MSc thesis, University of Natal, Pietermaritzburg.

Oates, J F. 1995. The dangers of conservation by rural development - a case study from the forests of Nigeria. Oryx 29 : 115-122.

Peters, C M. 1994. Sustainable harvest of non-timber plant resources in tropical moist forest: An ecological primer. Corporate press Inc., Landover MD.

Poffenberger, M., B. McGean, A. Khare, & J. Campbell. 1992. Field methods manuals in joint forest management, vols 1 and 2. Society for Promotion of Wastelands Development, New Dehli.

Scott, P. 1992. Fringe benefits : National Park boundaries as areas for local community utilisation : the case of the Impenetrable (Bwindi) forest. MSc thesis, Agricultural University of Norway.

Shackleton, S.E. 1990. Socio-economic importance of Cymbopogon validus in Mkambati Game Reserve, Transkei. South African Journal of Botany. 56: 675-332.

Taylor, D.M. 1993. Environmental change in montane southwest Uganda: A pollen record for the Holocene from Ahakagyesi swamp. The Holocene. 3: 324-332.

Walters, C. 1986. Adaptive management of renewable resources. Macmillan Publishers. London.

Wade, R. The management of common property resources: Collective action as an alternative to privatisation or state regulation. Cambridge Journal of Economics. 11: 95-106.

Wells, M. & K. Brandon. 1992. People and parks: Linking protected area management with local communities. World Bank, Washington.