Zambezia (1989), XVI (i).A PARTICIPATORY MODEL OF AGRICULTURALRESEARCH AND EXTENSION: THE CASE OF VLEIS,TREES AND GRAZING SCHEMES IN THE DRYSOUTH OF ZIMBABWEI. SCOONESImperial College, University of LondonandB. COUSINSCentre for Applied Social Sciences, University of ZimbabweWHY HAVE PEASANT farmers in Africa not adopted modern agricultural techno-logy more readily? In the past the most common diagnosis was peasant ignorance orcultural conservatism. The answer then 'obviously' lay in programmes of educationand extension Š hence the major investments in improving extension services in the1950s and 1960s. The oft-repeated exhortation to 'educate the farmers' can still beheard today in some quarters.In the 1970s and 1980s a new trend of thought emerged amongstagricultural economists and development planners which proclaimed therationality of decision-making in rural farm-households. Obstacles to improvedproduction were seen to be mainly external constraints on decision-makersŠconstraints such as restricted access to resources and the need of such farm-households to minimize risk. Thus improved understanding of the nature ofthese constraints by agricultural research scientists and extension agents becameimportant, so that more appropriate technologies and more adaptable extensionrecommendations would be passed on to the farmers. Other suggestedinterventions were: better prices for agricultural produce, upgraded transportand other infrastructures, the wider provision of credit, improved inputs, supplyand marketing systems, and so on.THE NEED FOR A NEW MODEL OF AGRICULTURALRESEARCH AND EXTENSIONThe results of these kinds of diagnoses and prescriptions have, however, generallybeen disappointing. Significant improvements in 'practice-adoption' and in-creases in production have generally been achieved by only a minority ofresource-rich farmers. The bulk of the rural population of Africa is made up ofresource-poor farmers, defined as those 'whose resources of land, water, labourand capital do not currently permit a decent and secure family livelihood'(Chambers and Ghildyal, 1985,3), and here success has been much more elusive.4546 AGRICULTURAL RESEARCH AND EXTENSIONAs a result, attention is being shifted to the deficiencies of the technology itselfand, even more importantly, to the process by which the technology is generated.In the older models the roles of research scientist, extension agent and farmer werewell, defined and seen as unproblematic: researchers generated the technology,extension officers communicated it, and farmers adopted (or failed to adopt) thesolutions to their problems.In some versions of this model, such as the Training and Visit system ofextension, the problems that farmers faced were communicated to researchscientists by field extension staff, who also assisted in the adaptation of standardrecommendations (e.g. for fertilizer application) to suit particular conditions.Farmers, however, remained essentially passive recipients of the 'medicine'prescribed for them by development experts. It is these roles that are nowincreasingly being brought into question, and the active participation of farmersin the whole process of technology generation is at the heart of the search for anew paradigm for agricultural research and extension.The Transfer-of-Technology ModelOne of the most influential voices in this debate is that of Chambers. Chambers andhis co-authors have typified the dominant paradigm, as described briefly above, asthe Transfer-of-Technology model. In their analysis the main reason why thismodel has not been effective with regard to resource-poor farmers is because'technologies . . . bear the imprint of the conditions in which they are generated.They are then adoptable in similar conditions, but often not adoptable whereconditions differ' (Chambers and Ghildyal, 1985, 6). Because conditions onresearch stations, where most agricultural research has been carried out in the past,are often similar to those on resource-rich farms, and usually very different fromthose on resource-poor farms, it is hardly surprising that the technology generatedthrough this process does not meet the latter's needs. The contrasts in conditionswhich give rise to this phenomenon are summarized in Table I.In the'Zimbabwean context the model fits reality most closely if we equateresource-rich farmers with large-scale commercial farmers, and indeed this wouldhave been entirely apposite before Independence. But even within the majorreorientation towards communal area farmers by research and extension that hastaken place since 1980 the same kinds of contrasts may be found. If we see'resource-rich' as a relative term, then the bulk of current agricultural researchprogrammes are still servicing farmers in the high-potential regions of the countryand, within these regions, those farmers who are either users of purchased inputsor cash-crop growers.An Alternative Model: Farmer-First-and-LastChambers and his colleagues have suggested an alternative model of agriculturalresearch which entails fundamental reversals of learning and location. In theirTable ITYPICAL CONTRASTS IN PHYSICAL, SOCIAL AND ECONOMIC CONDITIONS*Physical conditionsSoilTopographyNutrient deficiencyPlot size and natureHazardsSize of management unitSocial and economic conditionsAccess to inputsAccess to creditPricesPriority for food productionResearchstationsDeep, fertileFlat or terracedRareLarge, squareNil or fewLarge, contiguousUnlimited, reliableUnlimitedIrrelevantNeutralResource-richfarmDeep, fertileFlat or terracedOccasionalLargeFew, usually controllableLarge or medium, contiguousHigh, reliableGood accessLower than resource-poorfarmers for inputs, higher thanresource-poor farmers foroutputsLowResource-poorfarmShallow, infertileOften undulating, slopingCommonSmall, irregularCommon: floods, droughts etc.Small, often fragmentedLow, unreliablePoor accessHigher than resource-rich farmersfor inputs, lower thanresource-rich farmers foroutputsHighcoOOomw9°FOOCOzCOŁNot all factors apply all the time, but most apply most of the time.Source: Adapted from Chambers and Ghildyal, 1985, and Chambers and Jiggins, 1986.48 AGRICULTURAL RESEARCH AND EXTENSIONFarmer-First-and-Last model one starts with the perceptions and priorities offamilies of resource-poor farmers rather than those of scientists. Research andlearning are located on the resource-poor farm rather than on the research stationand in the laboratory. Problems in need of research are identified by reference tothe needs and opportunities of the farmer, and the research station has a referraland consultancy role serving the resource-poor family. Success is judged not bythe rigour of research or by yields in resource-rich conditions, but by the spread oftechnology among the resource poor. These reversals are summarized in Table II.Thus the three major components of the Farmer-First-and-Last model are:(i) a distinctive diagnostic procedure, learning from farmers;(ii) generating technology on the farm and with the farmers; and(iii) evaluation of technology by adoption or non-adoption.Is this not the same as the now widespread practice of Farming SystemsResearch? It is clear that in a number of ways Farming Systems Research hasdeparted from a conventional Transfer-of-Technology approach. It seeks tounderstand the complexity of farm-household systems, including the needs andobjectives of farmers and their families, and does so by means of multi-disciplinary teams which encompass biological, social and economic dimensions.In the diagnosis phase Farming Systems Research involves consultation withfarmers on their problems. Relatively homogeneous groups of farmers who arelikely to encounter similar problems and opportunities are identified as'recommendation domains'. On-farm trials are often farmer-managed, andfarmers assist in the evaluation of research results.But does this go far enough? Chambers and Jiggins (1986) consider FarmingSystems Research an adaptation of the Transfer-of-Technology model ratherthan as a fundamental break with it. They claim that in Farming SystemsResearch the power of choice remains primarily with the scientists, who make theimportant decisions about what to try and what to do. The linear sequence ofclassical Transfer-of-Technology has been modified by building in feedback loopsand cycles of referral and evaluation; but, 'in the absence of farmers'determination of research agendas, the process remains Transfer-of-Technology,with the scientist first and the farmer last' (Chambers and Jiggins, 1986, 19).These views have been criticized recently by Farrington and Martin (1987),who see a need for participatory approaches which complement FarmingSystems Research rather than attempt to replace it. In their view Chambers andJiggins's misgivings relate more to the application of Farming Systems Researchthan to the basic concepts involved, and the extreme farmer-centric stance ofFarmer-First-and-Last is unjustified: scientists and the scientific method do havean important role to play. The critical component is the partnership betweenresearchers and farmers.Table IICONTRASTS IN LEARNING AND LOCATIONTransfer-of- TechnologyFarmer-First-and-LastResearch prioritiesdetermined by ...Crucial learning is that of . . .Role of farmer:Role of scientist:Location of research anddevelopment:Features of research anddevelopment determined by.Non-adoption explained by .Evaluation by ...Needs, problems, perceptionsand environment of scientistsFarmers from scientistsBeneficiaryGenerator of technologyResearch station, laboratory,glasshouseScientists' needs and preferencesResearch station resourcesFailure of farmer to learnFarm-level constraintsPublications,scientists' peersNeeds, problems, etc., of farmersScientists from fannersClient and professional colleagueConsultant and collaboratorFanners' fields and conditionsFarmers' needs and preferencesFarm-level resourcesFailure of scientist to learnResearch station constraintsAdoption,farmersCOOOOmCOw8COCOSource: Adapted from Chambers and Ghildyal, 1985.50 AGRICULTURAL RESEARCH AND EXTENSIONAGRICULTURAL RESEARCH AND FARMERS IN THESEMI-ARID AREAS OF ZIMBABWEIn Zimbabwe two-thirds of the communal area population live in NaturalRegions IV and V, both of which are characterized by low and erratic rainfall.They are recommended for extensive and semi-extensive livestock production,although drought-resistant cash crops can be grown in favourable localities. Yetcommunal area farmers do not own enough cattle or indeed have sufficient landto engage in commercial ranching, and hence all grow food crops for subsistenceand occasional sale as well as cash crops such as cotton and sunflowers. Thehazards of crop production in these semi-arid areas are emphasized when oneconsiders the probability of a 'normal' season (one in which rainfall is adequate tosustain plant growth without adverse dry spells) occurring in these regions is ofthe order of 40 per cent in Natural Region V and 35 per cent in Natural Region IV(Hussein, 1987).Since Zimbabwe attained Independence in 1980 communal area fanners,with the help of increased levels of credit and improved supplies of inputs, havebegun to purchase significantly larger amounts of fertilizers and chemicals, andmarketed surpluses from this sub-sector have risen dramatically. Zimbabwe hasbeen hailed as a rare success story in respect of peasant production for the market.However, these increases in production and sales have not occurred 'acrossthe board' but rather reflect an increasing differentiation of the communal areapopulation which is based both on agro-ecological differences and on inequalitiesin the means of production. Thus in 1983/4, 63,1 per cent of the maize marketedby communal area farmers came from Natural Regions I and II where only 15 percent of the communal area population lives (Moyo, 1986, 189). A surveyconducted in 1984 found that rural households in the drier regions are much morelikely than those in the higher rainfall zones to experience food scarcity in mostyears, and that the great majority of households in Regions IV and V receiveddrought-relief food in 1983 (see Table III).Class-based differentiation is indicated by the marked differences betweenrural farm-households in all communal areas in respect of land holdings,livestock, availability of draught power, access to off-farm income and access tocredit. It is suggested by many researchers currently investigating this issue thatthe increases in marketed surplus are benefiting only a minority of better-offfamilies (Moyo, 1986; Jackson etaL, 1987; Adams, 1987; Weiner, 1988).How has agricultural research and extension attempted to meet the needs ofthe majority of communal area farmers who live mostly in the dry south and westof Zimbabwe? For many years, of course, this was not even recognized as aproblem, let alone seriously addressed. From the point of view of large-scalecommercial farmers these regions have no dryland cropping potential, and henceI. SCOONES &B. COUSINS 51the attention of researchers was focused mostly on beef ranching or irrigated cropssuch as sugar cane. Since Independence this attitude has begun to change, and theFarming Systems Research Unit within the Department of Research andSpecialist Services in particular has attempted to test potential technologicalinterventions aimed at overcoming or alleviating the powerful constraints onmixed crop-livestock farming in the semi-arid areas.Table IIIFOOD SCARCITY AND DROUGHT VULNERABILITY BY NATURAL REGIONNatural Region 'Food is scarce Received droughtin most years' relief food in 1983(Percentage of households)II 13,9 1,0III 16,0 41,3IV 20,4 67,4V 57,1 87,5Thus trials involving the traditional drought-resistant small grains (sorghumand the millets) have been initiated on both research station and on-farm sites(Farming Systems Research Unit, 1985). The Lowveld Research Station inChiredzi has been working on techniques of water conservation and concentra-tion, appropriate fertilizer regimes, relay cropping, variety selection and sowingdates (Jones, Nyamudza and Nyati, 1987). Attention has also been turned toresearch aimed at improving the productivity of communal area livestock, whichhave been almost completely ignored and hence almost nothing is known aboutthem (Mombeshora, 1985).However, few reliable recommendations have yet resulted from this researchprogramme and it may be many years before they do. Some of the results of theCommunal Area Research Trials programme have thrown researchers intoconfusion, with a senior scientist revealing in 1986 that there was now a need toreduce the total number of trials in order to make the programme moremanageable, and also to concentrate on understanding basic soil-plant climateinteractions in the sandy soils which predominate in the Communal Lands.*ŁE. Whingwiri, remarks at an Agritex Senior Officers Conference held at the University ofZimbabwe, January 1986.52 AGRICULTURAL RESEARCH AND EXTENSIONIn the meantime, extension field staff in the semi-arid areas have continued topromote 'packages' of improved farming practices which include hybrid seed,fertilizers, pesticides and knapsack sprayers. Indeed, in the absence of any moreappropriate knowledge, what else could they be expected to do?Research in Chirumanzu (which is in Natural Region III) has shown that thefew farmers who do make a profit on crops such as maize do so by applying muchless than the recommended amounts of fertilizer (Drinkwater, 1987,21). A recentanalysis of the economics of fertilizer-use on maize in Natural Regions III and IVsuggests that the most economic level of application is around half of thatcontained in the present Agricultural Finance Corporation package on offer, andaround a quarter to a half of the recommendations based on soil analysis(Whingwiri etal, 1987).A few extension officials do listen to farmers and attempt to modify standardrecommendations on the basis of local experience. Dialogue at the extensionagent-farmer interface has even led to innovations being developed, as inMaranda, where farmers are advised to plough twice in every second or thirdfurrow to achieve the same effect as a ripper tine. These are the exceptions, foron the whole extension practice still reflects a 'top-down, message-orientedapproach', even though many Agritex staff recognize that 'the research that hasbacked them until now looks neither at the generation of sustainable yield levelsnor the viability of communal area production' (Drinkwater, 1987, 2). Thedominant mode of agricultural research and extension is still Transfer-of-Technology, and the result is a dearth of appropriate technical recommendationsfor peasant farmers in the semi-arid areas of Zimbabwe.EXPERIENCES OF FARMER PARTICIPATORY RESEARCHAND EXTENSION IN ZVISHAVANE DISTRICTParticipation is the central theme of two projects being developed in MazvihwaCommunal Land in Zvishavane District, which is located in Natural Regions IVand V. Both projects have resulted from local demands generated as aconsequence of research work being carried out in the area. The Oxfam-supported community water resources project is investigating the potential of vleiutilization, and the ENDA-Zimbabwe trees project is concerned with thedevelopment of community forestry initiatives. Both projects tackle individualfarm potentials (arable production on vleis; agroforestry in fields and aroundhomes), as well as communal resource issues (vleis in grazing land; indigenouswoodland development). Both involve the close interaction of research andextension activities. Neither aspect is currently addressed by government policyon dryland agriculture or forestry. It is the general contention of this article thatthis policy gap on crucial issues (at least as perceived by farmers) can be attributedI. SCOONES & B. COUSINS 53in part to the lack of appropriate research and extension techniques that caneffectively address these types of problems.The Vlei ProjectClose contact with farmers and a continuing dialogue resulting from researchbeing carried out in Mazvihwa has highlighted vlei areas as being 'key resources'crucial to the sustainability of the dryland agropastoral system (Wilson, 1986;Scoones, 1987). However, very little is known about vlei systems, especially inthe drylands. A more detailed appraisal of their functioning and potential forsustainable use is called for.The aims of this focused appraisal have been(i) the characterization of dryland vlei systems in terms of their potential foruse and constraints to sustainable development;(ii) the investigation of local innovations for testing; and(iii) the identification of key issues for policy, further research, institutionaldevelopment and extension.The appraisal techniques used have been simple and rapid, aimed at encouraginga close alliance between farmer, extension worker and researcher at all stages.Initial characterization: Historical surveysHistorical precedents are often central to highlighting development constraintsand potentials. Recollections of past experience can point to local sources ofknowledge of earlier technologies or management practices and institutions thatmay assist current development attempts. Historical surveys can also show up theshortcomings of previous policies and help to avoid the simple resurrection offormer (failed) approaches. The value of a historical survey has been particularlysignificant in the vlei project, since nineteenth- and early twentieth-centuryproduction in the dryland hilly areas of Zimbabwe is thought to have been largelyreliant on vlei land (Wilson, 1986). As a consequence a number of productiontechniques (e.g. ridging, intercropping) were developed and are still remembered.Historical surveys are best carried out on the basis of a selection of key informantinterviews to derive a locally specific view of major changes and recollections ofparticular practices. Older, long-term residents, both men and women, areobviously the best source of such information.Classification and typologiesInitial investigations require a preliminary categorization of system types. This isthe initial phase of defining relevant 'recommendation domains'. These need to beflexible and iteratively redefined. In the vlei research project a number of differentvlei types, differentiated according to environmental characteristics, have beenrecognized:(i) Wet vleis: central marshy area; waterlogged in wet years; highly concen-trated drainage; multiple catchments; sandy soils surrounding vlei.54 AGRICULTURAL RESEARCH AND EXTENSION(ii) Dry vleis: shallow sloping catchments; no year-round surface water;vegetation distinct from wet vleis; sandy soils surrounding vlei.(iii) Red soil vleis: heavy red soil catchments; rarely waterlogged; variablecatchment types,(iv) Eroded vleis: gully erosion often stopped; normally in grazing areas.These vlei types can be located either in grazing areas or in arable blocks. In arablearea vleis there is usually a seasonal grazing component of unfarmed portions. Aworking typology, that needs constant revision, can be derived only from anumber of visits to different areas; it is developed through direct observation anddiscussion with farmers. Farmers often have useful local classifications withassociated terminologies that can form the basis of further elaboration.Agro-ecosystem functioningA focused appraisal of system functioning requires the investigation of a numberof interrelated aspects which can be subsequently analysed. The context we havefound most useful for this appraisal is a group workshop. The workshop isfacilitated by the extension worker and researcher together, who encourage theopen discussion of selected issues. In the vlei research project these have been vleiproduction and use, environmental determinants, intervention potential andinstitutional issues.Vlei production and use: The parameters of economic sustainabUity: Discussion ofproduction and use is aimed at assessing the constraints and potentials of theexisting prodiction system. An attempt is made to get a picture of the range ofpractices rather than a description of the ideal. A useful technique for qualitativeand rapid description is the construction of a flow diagram of productionactivities. This can incorporate both agricultural and economic factors in the sameframework and can be constructed while in the process of discussion withfarmers. A flow diagram can be a useful tool in identifying problems andopportunities that may involve interactions of agronomic issues, managementpractices and economic factors.Seasonality has a great impact in dryland agro-ecosystems and an appropriateway of looking at this is through the construction of a seasonal calendar. Againinteractions and trade-offs (e.g. between arable and livestock use) are highlighted.Environmental determinants and ecologicalsustainabUity: A full understanding ofthe production system requires an investigation of the environmental factors thatunderlie it. In vlei workshops we discuss in detail soil structures, hydrologicalpatterns, vegetation ecology and the problems of environmental degradation.This is based on farmers' often very detailed local observations and impression ofcausality.I. SCOONES & B. COUSINS 55Institutional issues: Issues of conflict, for instance between grazing and arable useof the land or between technical development and local beliefs surrounding thesacredness of wetlands, sometimes arise (usually obliquely) in discussion. Theseneed to be carefully investigated, and we have used interviews with selectedinformants to gain a better understanding of the issues before raising the questionof community management of resources in open discussion.Development potential Innovator exchange workshops: A discussion of farmers'experiences in their own development efforts is a useful starting-point forprioritizing development options and testing/adapting farmer-designed inter-ventions. The exchange of ideas among local innovators is a good way ofencouraging the dynamics of farmer-controlled development. These discussionscan serve as an entry point for 'outsider' suggestions that can be incorporated intolocal attempts without imposing solutions. In the vlei project the Oxfamextension worker has attempted to form farmer groups with a common interest invlei development and it is hoped that these can be the focus for the design andtesting of vlei cultivation techniques, low-cost irrigation technologies, vlei wellsand ponds and fish farming systems in the future.Sustainability analysisA vast array of often disparate, apparently unconnected information and ideascan be generated through a single farmers' workshop. The aim, however, is togenerate useful questions for research and development, and flow diagrams andseasonal calendars help to distil out the ecological and economic problems andopportunities. One example from the vlei project illustrates the results of their'sustainability analysis'. The phenomenon of dry-season (August-September)water-rise, is central to the production success of vleis, allowing relay cropping,diversification of crops and early marketing. The key questions now become:What are the determinants of this phenomenon? What aspects of watershedmanagement are necessary to ensure its continuation? How can late dry-seasonwater be more fully exploited in the development of vlei production?Defining the action/research agendaThe process of appraisal involves taking an interdisciplinary view based onfarmers' perspectives and local knowledge. H opefully what is generated is a set ofuseful questions for further attention, some of which will need to be addressed bydifferent components of the administration, extension and research structure.Some examples are:(i) Policy questions: vlei cultivation is officially discouraged for conservationreasons,(ii) On-farm research questions: some farmer-derived technologies can betested immediately, but other trials may need to be more structured and beeither fanner- or researcher-managed.56 AGRICULTURAL RESEARCH AND EXTENSION(iii) Basic research questions: can be referred to appropriate institutions such asresearch stations (e.g. the causal factors in the rise of dry-season water tables).(iv) Extension or implementation questions: practical issues of what furtheraction needs to be taken.The Tree Resources ProjectTree resources have been another problem area pointed to by farmers inZvishavane District. Farmers argue that trees are needed for a wide range of usesŠ as firewood, timber and browse in the communal woodlands; for fruit, shadeand windbreaks around homes; and for fertility inputs and fencing in arable lands.This calls for a multispecies approach to rural afforestation. In dry areas treesadapted to the local environment can help to meet this need. People in differentareas point to different problems. For some, firewood or poles may be in criticalshortage; others want to increase trees in fields or to establish fruit trees aroundhomes. The resource problems require a locally focused planning approach andan adaptive strategy of extension using a variety of tree species. Current policywhich essentially offers just the gum tree/woodlot package has failed to consultfarmers in the dryland areas about their priorities.The trees project is evolving a local planning approach to extension that isaimed at providing locally-appropriate community forestry options. Thetechniques used are basically the same as those used in the vlei project describedearlier: a combination of group and individual interactions between extensionworker, researcher and local communities that are aimed at encouragingparticipatory planning and involvement in natural resource development.An outline of the procedure being followed is given in Figure 1. It is managedby a locally resident extension worker and is closely supported by a small researcheffort. In contrast to the vlei project which has been focused on problem andopportunity identification, the trees project is implementation oriented. This leadsto different requirements. Discussion workshops, especially when held at wardlevel when the project is being introduced, are preferably large groups since theproject is aimed at whole community involvement at the outset rather than atencouraging participation of a representative sample in the research and designprocess.However, there is still a need for focused appraisal, concentrating onparticular 'user groups' (e.g. women and the fi rewood question) and a continuingrequirement for a research input to study the historical context, to investigate localsuggestions and monitor progress. The extension worker does much of thisinvestigation in the course of extension activities, but is supported by the projectco-ordinator.The project aims to assist tree planting in fields and around homes, but alsoaims to encourage planting and community management initiatives in theFigure 1: PROCEDURES IN THE TREE RESOURCES PROJECTProcedureKey informants: elders, political officials,chiefsOpen meeting for all in community;attendance 50-80 peopleInterviews with local informants; rapidassessments of environmental situation(maps, tree lists, etc.)VIDCO members, includingcommittee, village heads;30-40 peopleLocal nursery run by group, schoolor individualVIDCO-organized plantingDETAILED INTERVIEWS(1-2 hours each)ŁWARD WORKSHOP .(3+ hours)t9zccVIDCO APPRAISALS1VIDCO WORKSHOP(2-3 hours)WARD NURSERY LPLANTING ACTIVITIESi«<L_rcJJ/3AlmsHistorical survey: context forproblem in areaDefinition of key issues and localpriorities; discussion of problems;introduction of projectAppraisal of VIDCO-level problems,local management practices. Notingof key areas for reafforestationPlanning planting activities; lists oftrees needed; discussion of rules forwoodland managementProvision of trees to VIDCOs inline with requirementsCommunity planting activitiescooOOmCOs°CDOOwzCO58 AGRICULTURAL RESEARCH AND EXTENSIONcommunal indigenous woodlands. The Village Development Committee (VIDCO)structure is proving a more or less appropriate scale for community planning oftree resources. The pilot project has completed the local planning phase in MototiWard, Mazvihwa Communal Land.i"his has revealed an immense diversity of extension needs. This is not onlydetermined by differences in the local environment (e.g. hilly, sandy soil areas vs.heavy soil plains), but also by differences in management abilities and locally-specific requirements. For instance, among the six VIDCOs, two have plannedfor indigenous tree planting in the communal grazing area, one needs a gum treewoodlot for a local pole shortage, one was particularly concerned aboutwoodland management and needed to resolve a conflict between neighbouringcommunities over the use of trees within the grazing area, and all wanted a mix ofindigenous and exotic fruit trees for planting around homes.This diversity reflects a very real patchiness in natural resource situations andhighlights the necessity of a localized planning and management approach. Thisallows communities to become involved in their own resource issues andempowers them to act upon perceived problems.Methodological IssuesIndigenous technical knowledgeThe use of local knowledge can be an invaluable tool in designing focusedresearch agendas, suggesting appropriate technological interactions and adaptingextension messages. The important advantages of indigenous technicalknowledge, as experienced in the Zvishavane project, should be reiterated:(i) The identification of problems. Farmers necessarily have a holistic view of thefarming system, unconstrained by disciplinary training, which enables theidentification of key problems and opportunities often not apparent from acommodity-based or single-enterprise perspective. The identification of the vleiresources potential for arable and livestock production is an example.(ii) Detailed baseline observation. Fanners are often skilled observers and cancontribute an immense amount to the basic understanding of agro-ecosystemfunctioning. Detailed observation, often within an extended time-frame, allowsa qualitative insight for further investigation that is often ignored.(iii) Classification. Local classifications or typologies are essential to make thesubsequent appraisal workable and locally applicable. Classificationsderived from indigenous perceptions are generally very effective (Howe andChambers, 1979).(iv; Historical/local context. The value of gaining a local and time-basedperspective to appraisals for development has been repeatedly stressed above.The only source of this contextual information is local knowledge.I. SCOONES & B. COUSINS 59(v) Rapid assessment. With limited resources available for exhaustive research onevery facet of the problem, techniques that are rapid and cost-efficient, as well ascost-effective, are necessary. The research for the vlei project appraisal has beencarried out at a number of 3-4-hour workshops and during a series of in-depthinterviews lasting 1-2 hours. The cost has been minimal as these workshopsand interviews have fitted into the ongoing programme of research andextension work in the area.(vi) Fanner involvement. Farmer participatory research should not simply extractinformation and co-opt farmers in the research process; they should be fullyinvolved. The revelation that their own knowledge and experience is valuableand that they can be part of the determination of local development is animportant aspect. This has been a central aim of the vlei projectThere are limits, of course, to the role of indigenous technical knowledge.Local knowledge is rarely as effective in defining causality as reductionistscientific investigation. For instance, the explanation for the dry-season water risehas been variously described as the result of 'boiling underground water like akettle', the power of benevolent ancestral spirits, and the release of water frommysterious underground caves.Group WorkshopsThe vlei project has used group workshops for much of its investigation. Thesehave a number of benefits, problems and requirements. Workshop discussions, ifwell facilitated, can generate dynamic and open discussions. An unstructuredformat is essential, as a comment from one person can spark off a train ofdiscussion from others.Workshops are a good place for eliciting insights on technical issues orexchanging views on appropriate technologies. For politically sensitive or secretmatters a less public encounter is desirable. Similarly, for issues where there is adiversity of situations Š especially socio-economic or gender based Š a selected'focus' group is preferable. Open group workshops are sometimes dominated by'better' (usually richer) farmers or political officials Š always men.This can be overcome to some extent by successful facilitation andarrangement of the workshop. It is essential to avoid letting an appraisalworkshop coincide with a political meeting or a standard extension worker'smeeting. The informal, participatory and unconstrained atmosphere of theresearch workshop is vital to its success.With the right atmosphere, workshops are a fruitful place for the successfulinteraction of researchers, extension workers and farmers. Researchers are forcedto address development issues and farmers' priorities, extension workers arerequired to be responsive and investigative, and farmers are allowed a voice.60 AGRICULTURAL RESEARCH AND EXTENSIONGrazing Schemes and Participatory ResearchThe Communal Lands are generally viewed as being both overstocked andovergrazed. The consequences are said to be low levels of productivity and aserious threat of irreversible degradation. The two most commonly suggestedsolutions are (a) a restriction of stock numbers to within the carrying capacity ofthe veld, and (b) improved grazing management by means of fenced paddocks(i.e. 'grazing schemes').These are not new ideas: in one form or another they have informed stateinterventions in peasant land-use systems since the 1930s, when Alvordintroduced the notion of'centralization*. An examination of the history of theseinterventions reveals that virtually all of them failed in their objectives, in somecases (e.g. the Native Land Husbandry Act of 1951) because they generatedwidespread political opposition. Coercive state action to enforce compulsorydestocking has always been resisted, either covertly, with farmers 'hiding' theiranimals in various ways, or in the form of support for guerrillas promising an endto state interference in peasant agriculture (Beinart, 1984; Ranger, 1985).In the early 1970s, however, the promotion of Short Duration Grazingschemes proved more successful, particularly in Masvingo Province. No mentionwas made by extension staff of destocking measures, and instead the benefits toveld and cattle were emphasized. A much greater effort was also made to involvecommunity leadership in the delineation of resource boundaries and in the activemanagement of these schemes (Froude, 1974; Danckwerts, n.d.)Since Independence grazing schemes have again been promoted and anumber of pre-Independence schemes have been resuscitated. A survey inDecember 1986 revealed 50 schemes which claimed to be operating and a further56 at the planning stage (Cousins, 1987). Voluntary adoption and communitymobilization have been strongly emphasized by Agritex. Recently the NationalConservation Strategy has reiterated the need for 'planning [which] will involvethe active participation and commitment of the local communities and not beimposed upon them' (Zimbabwe, 1987, 23).On the other hand, government policy has continued to pull strongly in thedirection of 'adjusting stocking rates to within carrying capacity' (Zimbabwe,1987, 25) and 'destocking where necessary' (Zimbabwe, 1986, 27). These twoobjectives Š popular participation in land-use planning on the one hand, andreduction in stock numbers on the other Š sit very uneasily together. Theoverwhelming evidence from research carried out by the Farming SystemResearch Unit and others is that 30-40 per cent of households in the communalareas own no cattle at all, and that the majority of households have insufficientaccess to draught power (Zimbabwe, 1984; Farming System Research Unit,1985). Lack of adequate draught power constitutes a major bottleneck in cropI. SCOONES & B. COUSINS 61production, and so most people have a strong need to increase rather thandecrease their herds (Shumba, 1984; Cliffe, 1986).This is a complex issue and clearly not amenable to simple solutions. Part ofthe answer undoubtedly lies in policy shifts which would allow a much moresubstantial resettlement programme, more 'Model D' resettlement schemes, andthe granting of permission to communal area farmers to purchase adjoining farmsor ranches. All of these would address the problem by increasing the grazing areaavailable to hard-pressed communities, and are currently the subjects ofdiscussion within rural communities themselves.Institutional development which allows for the effective management ofcommon property resources also needs to be encouraged, and those communitieswhich are already operating grazing schemes are an important source of learningabout what is likely to work. In the Tagarika scheme in Mwenezi District, forexample, the community has begun to discuss the principle of allocating equalshares of grazing rights, with redistribution effected by means of traditionalmechanisms of lending out cattle (kuronzera).But there are also technical issues to be addressed, in particular, the thornyquestion of stocking rates and carrying capacity, and again a participatoryapproach to research and extension may assist in the development of sustainablesolutions to this pressing problem.It has become clear over the last few years that terms such as 'carryingcapacity' are not as unproblematic as they were once thought to be. An intensedebate as to the precise definition and application of such concepts is nowunder way, with at least three different viewpoints being expressed: theconventional wisdom of mainstream veld and pasture science, the dissentingvoice of Sandford and others, and the provocative stance adopted by Savory andadvocates of Holistic Resource Management (see Cousins, 1987, for an extendedsummary of these positions).These debates have been prompted by contradictory sets of evidence: on theone hand, the vegetation of the communal areas is undoubtedly very heavilyutilized by livestock, and in some areas the soil is very poorly covered. In thesesituations serious soil erosion is probably occurring, although we have little in theway of direct measurements of the extent of permanent soil loss (Van den WallBake, 1986).On the other land, although concern over stocking rates and environmentaldegradation in communal areas was first expressed in the 1920s, livestocknumbers have continued to increase and are today probably well over the threemillion mark. Most Communal Lands are carrying numbers of animals which areat least twice those recommended using conventional .estimates of carryingcapacity. But many cattle, particularly in the higher rainfall regions, can be62 AGRICULTURAL RESEARCH AND EXTENSIONobserved to be in excellent condition right at the end of the dry season, and after adrought year. As Sandford has commented:When actual stocking rates... are far in excess of carrying capacities for long periods oftiire then one should ask whether this apparent prolonged defiance of the Laws of Naturemay not have been due to an initial underestimate of the true carrying capacity (Sandford,1982, 57).Sandford recommended basic research on the relationship between stockingrates, primary productivity and soil erosion, and applied research on optimumstocking rates related to the needs of peasant farmers. Unfortunately thisrecommendation was not acted upon, and valuable time has been lost. Tensionsbetween the central state and local communities on the issue of land-use andsettlement patterns are emerging again as the result of the pilot villagizationprogramme, and the National Conservation Strategy announces that Agritex willundertake land-use planning in all VIDCOs within seven years. To avoid adamaging confrontation on stocking rates we urgently need to have a much betteridea of the true ecological carrying capacity of communal area grazing land.Grazing schemes, as currently designed, are a technological interventionbased on an ecological model which is increasingly in question. The assumptionhas been that if Short Duration Grazing works on large-scale commercial ranchesthen it must surely be appropriate for communal areas as well Š in short, a primeexample of the Transfer-of-Technology paradigm.The community-based research being carried out in Mazvihwa has begun toaddress both the underlying issue of ecological carrying capacity and the questionof appropriate management strategies for improved and sustainable livestockproduction. Again, the methods have been participatory in character and similarto those used in the vlei and tree projects. Since this aspect of the research is still inan embryonic phase, it will be only briefly summarized.From interviews with individual farmers, group discussions of the kinddescribed above, and also from direct observations of farmers' practices andlivestock behaviour, some of the properties of the dryland farming system havebeen identified. Two types of savannah ecosystem may be distinguished: adystrophic type, characterized by granitic sands, in which soil nutrients constitutethe major constraint, and a eutrophic type, in which soils are either clays or loamsand the major constraint is soil water. These correspond broadly to the localclassifications of soil types and vegetation known as mucheche and chiwomvoAlso important are riverine areas, vleis, drainage lines, and all alluvial soils.These form only a small proportion of the total land area, but together withbrowse are crucial to the survival of livestock at the end of the dry season. Theyare also important for crop production, as discussed above. They thus constitute'key resources' in an essentially heterogeneous agro-ecosystem.I. SCOONES & B. COUSINS 63In this environment a common method of coping with drought is to movelivestock from the eutrophic areas to the dystrophic, which often involves lendinganimals for a period of time to individuals in neighbouring communities. Mobility isa strategy that reflects spatial as well as temporal variability of forage production.Taken together, these insights into how the existing farming system functionssuggest an alternative strategy for the management of forage resources. Instead of thecostly fencing of large areas of poor quality land into paddocks, for the purposes ofoperating a conventional rotational grazing system, it may be more useful toconcentrate on community regulation of access to the 'key resources'.For example, portions of the grazed vleis could perhaps be fenced off andreserved for selected draught or milking animals, particularly towards the end of thedry season. In general the heterogeneity of the environment needs to be recognized,and flexible, adaptive management strategies need to be devised to make optimumuse of it.Most importantly, however, to develop workable strategies the farmersthemselves would have to be involved in the planning process, contributing theirown knowledge of the local environment and suggesting possible avenues for furtherinvestigation and trials. The Transfer-of-Technology model, from which present-day grazing schemes seem largely to derive, needs replacing with an approach whichinvolves research, extension and resource planning in a dynamic adaptive andlearning process.CONCLUSIONIn effecting a marriage of extension and research in a participatory framework theanalytical techniques and adaptive extension procedures used by the vlei and treeprojects may prove useful. A participatory approach necessarily acts in acomplementary way to existing strategies. It is not intended to replace mainstreamresearch and extension efforts, but to assist in their effective operation throughexisting institutional structures. The failures of the Transfer-of-Technology andpackage model, especially in the dryland areas of Zimbabwe, suggest that the time isright to try an alternative approach.ReferencesADAMS, J. 1987 'Wage Labour in Mutirikwa Communal Land, Masvingo'(Harare, Univ. of Zimbabwe, Centre for Applied Social Sciences, SeminarPaper).BEINART, W. 1984 'Soil erosion, conservationism and ideas about develop-ment: A Southern African exploration, 1900-1960', Journal of SouthernAfrican Studies, XI, 52-83.64 AGRICULTURAL RESEARCH AND EXTENSIONCHAMBERS, R. AND GHILDYAL, B. P. 1985 'Agricultural research forresource-poor farmers: The Farmer-First-and-Last model', AgriculturalAdministration, XX, 1-30.CHAMBERS, R. AND JIGGINS, J. 1986 'Agricultural Research for Resource-Poor Farmers: A Parsimonious Paradigm' (Brighton, Univ. of Sussex,Institute of Development Studies, Discussion Paper 220).CLIFFE, L. 1986 Policy Options for Agrarian Reform: A TechnicalAppraisal (Harare, FAO).COUSINS, B. 1987 A Survey of Current Grazing Schemes in the CommunalLands of Zimbabwe (Harare, Univ. of Zimbabwe, Centre for Applied SocialSciences).DANCKWERTS, J. n.d. A Socio-economic Study of Veld Management inTribal Areas of the Victoria Province (Salisbury, Tribal Areas ResearchFoundation of Rhodesia).DRINKWATER, M. 1987 'Exhausted Messages, Training and Groups: AnEvaluation of Zimbabwe's Training and Visit System' (Harare, Univ. ofZimbabwe, Dept. of Agricultural Economics and Extension, WorkingPaper).FARMING SYSTEMS RESEARCH UNIT 1985 Annual Report (Harare, Dept.of Research and Specialist Services).FARRINGTON, J. AND MARTIN, A. 1987 Farmer Participatory Research:A Review of Concepts and Practices (London, Overseas DevelopmentAgricultural Administration (Research and Extension) Network, DiscussionPaper 19).FROUDE, M. 1974 'Veld Management in the Victoria Province TribalAreas', Rhodesian Agricultural Journal, LXXI, 29-37.HOWES, M. AND CHAMBERS, R. 1979 'Indigenous technical knowledge:Analysis, implications and issues', in Institute of Development StudiesBulletin, X, ii: Rural Development: Whose Knowledge Counts?, 5-11.HUSSEIN, J. 1987 'Agro-climatological Analysis of Cro p Growing Seasonsin Natural Regions HI, IV, and V of Zimbabwe' (Harare, Agritex Workshopon 'Cropping in the Semi-arid Areas of Zimbabwe').JACKSON, J. C. 1987 Rural Development Policies and Food Security inZimbabwe: Part II (Geneva, International Labour Office).JONES, E., NYAMUDZA, P. AND NYATI, C. 1987 'Rainfed Farming inNatural Region V (Harare, Agritex Workshop on 'Cropping in the Semi-arid Areas of Zimbabwe').MOMBESHORA, B. 1985 'Livestock production research', in M. Avila (ed.),Crop and Livestock Production Research for Communal Areas (Harare,Dept. of Research and Specialist Services, Zimbabwe Agricultural JournalSpecial Report 1), 84-96.I. SCOONES & B. COUSINS 65MOYO, S. 1986 'The land question', in I. Mandaza (ed.), Zimbabwe: ThePolitical Economy of Transition: 1980-1986 (Dakar, Codesria), 165-201.RANGER, T. 1985 Peasant Consciousness and Guerrilla War in Zimbabwe(Harare, Zimbabwe Publishing House).SANDFORD, S. 1982 'Livestock in the Communal Areas of Zimbabwe'(Harare, Ministry of Lands, Resettlement and Rural Development).SCOONES, I. 1987 'Key Properties of Dryland Farming Systems inZimbabwe's Communal Areas: Issues for Research and Development'(Harare, Univ. of Zimbabwe, Dept. of Agricultural Economics andExtension, Seminar Paper).SHUMBA, E. 1984 'Animals and the cropping system in the communal areasof Zimbabwe', Zimbabwe Science News, XVIII, 99-102.VAN DEN WALL BAKE, G. W. 1986 'The siltation and soil erosion survey inZimbabwe', in R. F. Hadley (ed.), Drainage Basin Sediment Delivery(Wallingford, International Association of Hydrological Sciences, Publication159), 69-80.WEINER, D. 1988 'Land and agricultural development', in C. Stoneman(ed.), Zimbabwe's Prospects (London, Macmillan), 63-89.WHINGWIRI, E. et al., 1987 'The Package of Higher Maize Yields in theSemi-arid Zones (Natural Regions III and IV): Is It Economical?' (Harare,Agritex Workshop on 'Cropping in the Semi-arid Areas of Zimbabwe').WILSON, K. 1986 'Aspects of the History of Vlei Cultivation in SouthernZimbabwe's Communal Areas' (Harare, Univ. of Zimbabwe, Dept. of CivilEngineering, Seminar Paper).ZIMBABWE 1984 Report on the Demographic Socio-economic Survey of theCommunal Lands of Masvingo Province 1983/84 (Harare, Central Stati-stical Office).ZIMBABWE 1986 First Five-Year National Development Plan 1986-1990:Volume I (Harare, Ministry of Finance, Economic Planning and Develop-ment).ZIMBABWE 1987 The National Conservation Strategy: Zimbabwe's Road toSurvival (Harare, Ministry of Information, Posts and Telecommunications).