Document(s) 19 de 38

Abstract: Farmers are increasingly participating in agricultural research as scientists and development workers become more aware of the philosophy of 'farmer first and its effectiveness. Many farmer participatory approaches are possible in farmer participatory research for improved crop cultivars by farmers. They should be broadly categorized into farmer participatory varietal selection (PVS) and farmer participatory plant breeding (PPB) since they conveniently define two approaches that are very different, and are likely to have very different impacts. Methods are reviewed in PVS and PPB that employ differing levels of farmer participation and researcher inputs. Depending on the situation, either PVS or PPB can be the most appropriate method to use. PPB often follows from the successful participatory identification of cultivars.

Although both PVS and PPB can increase biodiversity found in farmers fields, PPB has the greatest effect. Indeed, if PPB is used with the maximum possible involvement of farmers using material generated from landrace x exotic crosses then it represents a dynamic form of in situ genetic conservation. This method is likely to have the greatest impact on conserving biodiversity.

Little attention has been paid to the impact of farmer participatory research on biodiversity. In the published examples on participatory plant breeding (Salazar, 1992 and Worede and Mekbib, 1993) the idea of preserving biodiversity has been in the mind of the researchers, and there has been emphasis on improving landraces or using them as parents. The issue of biodiversity has hardly been considered in the work on PVS.

In this paper, work and results are reported in PVS and, to a lesser extent, in PPB of the Crops Programme of the Krishak Bharati Cooperative Indo British Rainfed Farming Project (KRIBP).

Introduction

Participatory varietal selection

• a means of identifying farmers needs in a cultivar,
• a search for suitable material to test with farmers, and
• experimentation on farmers fields.

Identification of farmers needs

• participatory rural appraisal (PRA),
• the examination of the type of crops in farmers fields at or near to maturity, or
• the pre-selection by farmers of cultivars by the inspection of trials of many entries grown on a research station or in farmers fields.

Employment of such methods will help to reduce the possibility that farmers will be given obviously unacceptable varieties to test.

Search process

Amongst already released cultivars: One method, employed by the authors in India, is to include in the search cultivars that have already been released. A key assumption made in participatory varietal selection on released cultivars (Joshi and Witcombe, 1995) is that cultivar replacement rates are lower than optimal because farmers have not been exposed to a range of new cultivars. It is therefore assumed that amongst the released cultivars there are ones that will be preferred by farmers over those they are currently growing. All that is required is to expose the farmers to the suitable cultivars for the project area that already exist, but have not been released or are not available in that area. For many crops in India, cultivars can be introduced from other states for a participatory varietal selection program since many cultivars have only been released in single states.

There is a considerable body of evidence to support the assumption that farmers are not rapidly adopting new cultivars because most cultivars under cultivation are old. There is also good evidence that only a few of the released cultivars are widely grown. For example, in wheat in India the average age of cultivars under breeder seed indent is 9 years, and the average of cultivars in certified seed production is 13 years in the three states of the KRIBP project, Gujarat, Madhya Pradesh and Rajasthan (Witcombe et al., unpublished). The two most popular cultivars are Sonalika (released in 1967), and HD-2285 (released in 1982) and these account for a large proportion of the area. However, for wheat, there is a good choice of cultivars as there have been 44 releases in the period 1984 to 1992 inclusive. In most crops, cultivars are on average older than those in wheat. For example, the average age of cultivars under breeder seed indent is 11 years in rice, 13 years in chickpea, 15 years in groundnut, 16 years in sorghum, and 17 years in maize (Witcombe et al. unpublished).

Choosing from amongst released cultivars has the advantage that any non-governmental organization or governmental organization (NGO or GO, respectively) can, in principle, readily procure seeds in sufficient quantities for testing with farmers. If they are identified as being farmer-acceptable it should be much easier, than is the case for pre-release or breeders lines, to provide large quantities of seed to the farmers with little delay. Nonetheless, to increase the size of the basket of choices and exploit recent outputs from plant breeding research, pre-release cultivars might also be included in the search process and a number have identified as being suitable for testing with farmers in the KRIBP project (Table 1). Some of the pre-release cultivars would be defined by others as advanced material.

Table 1: Cultivars identified by participatory varietal selection in the KRIBP project and performance of recommended cultivars
Crop	Performance of recommended cultivars	Cultivars tested¹		Cultivars identified	Release location	Year of release
		Rel.	Pre-rel.
Rice	All recommended cultivars tried failed in farmers' fields	9	7	Kalinga III Sathi-34-36	Orissa Gujarat	1983 1955
Maize (white)	Recommended cultivar not distinguished from local by farmers	3	3	Shweta	U.P.	1980
Chickpea	One of the recommended cultivars, Dahod yellow, is the local cultivar. Other failed.	3	22	ICCV 2 ICCV 10 ICCV 88202	A.P. SZ² Not rel.	1989 1992 -
Black gram	Recommended cultivar T-9 was less preferred. Others not tested as very old.	2	2	TPU-4 IU8-6	National Not rel.	1992 -
1. Rel. = released cultivar and Pre-rel. = pre-released cultivar 2. SZ= Southern Zone comprising Andhra Pradesh, southern Madhya Pradesh and Tamil Nadu (ICCV 10)

Amongst advanced material: Maurya et al. 1988, after a PRA on farmers' needs, searched among characterized advanced lines to find suitable material for testing with farmers. In other cases, breeders have searched amongst pre-released cultivars (entries in advanced stages of testing) having local adaptation and have deliberately chosen material that represents a wide range of phenotypes (Weltzien et al. 1995, this volume). Sperling et al. (1993) and Lechner (pers. commun.) have successfully used farmer visits to research stations trials to identify suitable cultivars amongst the trial entries of non-released material.

Experimentation for PVS

There are other methods that lie between the extremes of maximum and minimum inputs from outsiders. Scientists participation increases when farmers are asked to grow more than one introduced cultivar, since it necessitates an experimental design in which farmers, if unaided, can easily make planting errors. The scientists' contribution will also vary according to the quantity and quality of the data that is collected, and is greatest when quantitative estimates of yield using field-sampling techniques are employed. Therefore, a method that requires considerable input from scientists is when farmers are asked, with researcher help, to grow a set of cultivars, and quantitative data on their yield is taken by the researchers. However, most researchers when using participatory methods have asked farmers to grow only one introduction side by side with their local with no change in management, and have collected data that pertains to farmers perceptions of the cultivars.

The methods used by the authors are described in detail as an example of PVS. The varietal trials were carried out by farmers in Farmer Managed Participatory Research (FAMPAR) trials. The trials were divided into introductory and adaptive trials. The main difference between these two stages is that small quantities of seed were given to farmers in the introductory trials, but, to avoid overestimating the acceptability of cultivars to farmers, seed was sold at commercial rates in the adaptive trials (Joshi and Witcombe, 1995). The trials were made as simple as possible. Each participating farmer was randomly assigned a single variety, and asked to grow it alongside the local variety in the same field. Each farmer was given two bags of seed, so that the plot could be resown, if required, from the second bag of seed. Farmers were asked to mark the plots and to do this they sometimes grew a row of different crop between the two cultivars. They were also asked not to change the management of the crop in any way. Enough seed was provided for an average plot size 100 m² which is much larger than that used in advanced on-station trials.

In the introductory trials, data were collected by means of Focus Group Discussions (FGDs) before and after harvest, on all aspects of the crop including taste, market value, threshing characteristics and storability. Evaluation was facilitated by 'farm walks in which the participating farmers visited each other's plots. All the cultivars could then be compared in the discussions, and it permitted the assessment of the reactions to each cultivar of all of the farmers that participated in the farm walks. The focused group discussions were followed by questionnaires completed for individual households, called household level questionnaires (HLQ), in which the household members reactions to the variety were assessed by means of a detailed questionnaire that included questions on post harvest traits, such as cooking quality and market value of the grain.

Results obtained from participatory selection in KRIBP

PVS in rice: In 1993, introductory trials of rice were planned with 25 participating farmers in six villages, making a total of 150 farmers. Successful trials were conducted by 128 farmers, because some farmers failed to plant the seed. In each village, five cultivars were grown and every cultivar was replicated across three to five farmers. The cultivars were Kalinga III, Sathi-34-36, Jaldi Dhan-1, Jaldi Dhan-3, and GR-3.

The farmers perceptions of Kalinga III, the most preferred variety, were compared in six villages. For yield, there was perfect agreement that Kalinga III was higher yielding in all villages in Madhya Pradesh and Gujarat (Fig. 1). From observations of farmers fields, Kalinga III was seen to be considerably higher yielding. However, perceptions that Kalinga III was higher yielding than the local were far less marked in Rajasthan, but were always considered to be so by 50% or more of the farmers (Fig. 1). The probable reason is that the land is less sloping and more fertile in these Rajasthan villages, so there is a reduction in the advantage of Kalinga III, a cultivar that is highly adapted to low input conditions.

Figure 1: The infinite possibilities of replacement of existing variability with new genetic material

PVS in chickpea: In rabi (dry season) 1992/1993, five chickpea cultivars were grown in six villages. In each village, each cultivar was grown by four farmers to give a total of 120 farmers. After the harvest in 1993, a HLQ was conducted and farmers were asked if they would grow their chickpea cultivar again. Three cultivars, ICCV 2, ICCV 88202 and ICCV 10, were selected by a good proportion of the farmers, but ICCC 4 the official released variety was not liked (Table 2). Differences between the cultivars were less clear when the cultivar choice was restricted to the cultivar the farmer had grown, than when choice was restricted to the remaining cultivars (Table 2). Probably the availability of seed of the cultivar the farmer has grown greatly influences the decision in favor of regrowing it.

Table 2: Number of farmers participating in the 1992-1993 participatory varietal selection trials who said they would grow one of the test cultivars in the following year
Cultivar choice	Sample size per cultivar¹	Number of farmers who said they would grow the cultivar next year²
		ICCC 4	ICCV 37	ICCV 10	ICCV 88202	ICCV 2
Cultivar farmer had grown in the trials	24	10	13	17	21	17
Remaining cultivar	96	1	5	13	25	35
Total	120	11	18	30	46	52
1. The maximum number of farmers that could choose an individual cultivar. 2. The farmers when asked this question were asked to assume that the seed would have to be purchased at commercial rates if they did not already have the seed.

The second-year adoption rates of the preferred cultivars were found for the cultivars by interviewing farmers in two villages, (those that had participated in trials in 1993/94), at the end of the 1994/95 season . The differences in adoption rates into the second year showed that ICCC 4 was even less liked than was indicated by the HLQ (Table 3). ICCV 37 was resown by two of the four farmers and no seed was given to others. Adoption rates were higher for the three preferred cultivars in the HLQ. ICCV 88202 and ICCV 2 were more preferred than ICCV 10 in terms of both adoption rate and the number of recipients of seed and this result agreed with those from the HLQ. Since the two surveys had only one village in common, this agreement was even more impressive. The three farmer-preferred cultivars will spread with farmer-to-farmer seed supply, and the area under these cultivars will increase quite rapidly, even if no further seed is supplied from outside. Multiplication rates are conservatively estimated at between two and three times and there is a high rate of spread of seed to new areas (Table 3).

Table 3: The results of a survey in 1995 in two villages on resowing rates of cultivars first grown in rabi 1993-1994
Cultivar	Sample size	Percentage resowing	Increase in area in project village (times)¹	Number of new farmers given seed of cultivar	Mean distance of new farmers from village (km)	Total amount given (kg)
ICCC 4	5	0	-	0	-	0
ICCV 37	4	50	2.4	0	-	0
ICCV 10	7	57	2.9	5	3.4	7
ICCV 88202	8	75	1.8	10	10.5	25
ICCV 2	8	75	2.8	4	5.6	3
1. Ratio of sown seed in 94/95 over seed sown by all farmers in 93/94. Area sown in 94/95 excludes new farmers given seed in 93/94 as they were not interviewed.

An FGD for 1993-1994 crop was conducted in three villages after the 1994 harvest. There was a total of 23 participants. It was found that the preference ranking changed somewhat, as ICCV 10 was preferred the most over the local. There was little difference in the preference over the local between ICCV 88202 and ICCV 2.

The results from the three different surveys all agree in finding ICCV 2, ICCV 88202 and ICCV 10 as the preferred varieties, but the order of preference changed somewhat. This is perhaps unsurprising as the three cultivars all have markedly different characteristics. ICCV 2 is very early kabuli type, the seeds of which fetch a higher market price. ICCV 88202 is a very early desi type, and ICCV 10 is later than ICCV 88202 but higher yielding. Different villages may have different preferences according to such factors as access to markets and soil types. By exposing the farmers to a diverse range of genotypes a number of cultivars have been adopted, and biodiversity has been increased since they are all being adopted and partially replacing the uniform single 'landrace cultivar, Dahod yellow, that was exclusively cultivated in the area before these introductions.

Participatory plant breeding

The methods used in participatory plant breeding are poorly documented since there are no reports in the literature of a completed participatory plant breeding program. Sthapit et al.(this volume) have used F5 bulk families as the starting point for their participatory breeding program. These were derived from seed harvested from F4 families that were grown in the farmers village. The breeding scheme is at the F5 stage in the monsoon season of 1995, and it is intended to monitor progress in the farmers fields until a finished product is produced. In contrast, Thakur (1995) has screened material in farmers' fields at the F2 stage, but subsequent generations have been grown by researchers. The authors, in collaboration with Dr. Goyal of Gujarat Agricultural University, are starting a participatory plant breeding in maize with the fourth random mating generation of a composite created from six farmer-acceptable open-pollinated cultivars. In Ethiopia, farmer enhancement of landraces by mass selection has been done in collaboration with scientists from the Plant Genetic Resources Centre (Worede and Mekbib, 1993). In participatory plant breeding in rice in the Philippines, farmers are involved in selecting from progeny of crosses between traditional and improved cultivars but, unfortunately, the methods used are not described in detail (Salazar, 1992).

A range of participatory plant breeding methods are possible with predominantly self-pollinating crops, and they have been ordered by degree of farmer participation in Table 4. The methods vary according to which generations are grown by farmers and by the extent of researcher participation. The method with the greatest farmer participation and the greatest number of generations requires little breeder input during the selection stages. However, an essential role is played by breeders and participatory plant breeding is not intended to make plant breeders redundant. In all of the methods, the plant breeder is the facilitator of the research. Only the plant breeder can make the crosses between the parents and have the essential understanding of the underlying genetics in the segregating generations. Moreover, only the plant breeder has the knowledge of the official release system, and cultivar release is still a very desirable end product to make the results of the participatory research more widely available.

Table 4: Methods of participatory plant breeding in predominantly self pollinating crops.
Methods in increasing order of farmer participation	Site specificity	Reference
1. Early generation (F2) in farmers fields. All other generations and procedures with plant breeder.	Single location.	Thakur et al., 1995
2. Best advanced lines at F7 or F8 given to farmers for testing. Closest method to participatory varietal selection since farmers given nearly-finished product.	Easy to use across locations.	Recommended by Galt, 1989 Maurya et al. 1988
3. From F3 onwards farmers and plant breeders work together to select and identify the best material. Farmers are the selectors. Plant breeders facilitate the process by giving advice on which characters are heritable, and on selection methods. Pre-release multiplication can take place in parallel to the participatory plant breeding. Release proposal prepared by plant breeder.	Possible to run selection procedures in more than one location.	Sthapit et al. 1995
4. Breeder gives F3 or F4 material to farmers. All selection and advancement of generations left to farmers. At F7 to F8, or later, stage breeders monitor diversity in farmers fields. They identify, by phenotypic appearance and farmers perceptions, best material to enter in conventional trials and pre-release multiplication.	Extremely easy to run selection schemes in many locations.	-

For predominantly open-pollinating crops, plant breeders can create composites in isolation and give the third or fourth random mating to farmers for mass selection. Large plots of composite have to be grown by farmers, or small plots need to be isolated by time of flowering or distance from other plots of the same crop. Because of these constraints, it is difficult to carry out the breeding scheme in many locations. Mass selection can be done with or without an off-season generation controlled by the plant breeder. However, to breed for wider adaptation, plant breeders can recombine selections from different farmers in the off-season. Although it is likely to reduce the progress made by selection over the best farmers selection, it avoids the risk of continuing a population from a poor selection by one farmer or from a population where outcrossing with other farmers crops happened to be higher than expected.

Greater plant breeder input is possible by using a method of progeny testing. Plant breeders can produce progeny before giving material to farmers, and can produce progeny between generations of farmer selection. In the most extremely breeder-oriented system, the farmers grow progeny trials of full-sib or S1 families, and the breeder recombines from remnant seed the farmer-preferred progeny.

Impact of farmer participatory research on biodiversity

Biodiversity in crops

Biodiversity can be over both space and time. When one cultivar totally replaces another, there is an increase in biodiversity over time. There is also a temporary increase in biodiversity over space until the replacement is complete because, while replacement is occurring, there are two cultivars in farmers fields instead of one. The pattern by which this replacement takes place, from many or only a few foci, will also be important. We can assume that there is greater biodiversity when the new cultivars spreads from many foci. An example of this type of spread is seen in the case for new chickpea cultivars in KRIBP (Table 3). The spread from many foci will give a more complex pattern between farmers fields, providing a useful increase in biodiversity. The vulnerability of a crop to a disease is reduced when there are many field-to-field differences than when there are few. This strategy of field-to-field variability has been recommended by Priestly and Bayles (1980, 1982). However, regional variation in cultivar diversity is also suggested as useful in disease control by Frey et al. (1977).

When participatory research increases replacement rates, and thus reduces the longevity of individual cultivars, biodiversity is increased over time. It is again a useful increase in biodiversity since pathogens and pests are exposed to a particular genotype for less time and have less chance to overcome host plant resistances.

Participatory varietal selection

Figure 2: Farmers' perceptions on a village to village basis of yield of Kalinga III in comparison to local in six villages in three states, kharif 1993

Table 5. Examples of the effect of participatory varietal selection (PVS) on biodiversity in a single agroecological zone
Environment	Biodiversity		Comments
	Decrease	Increase
Marginal	New cultivar or cultivars replace landraces.		Common situation in marginal areas where adoption rates of improved cultivars are so low that landraces remain in farmers fields.
	Single cultivar replaces a range of improved cultivars.		Under conventional varietal testing system the existence of a range of introduced cultivars in a marginal environment is unlikely. Moreover, PVS exposes farmers to a range of cultivars and often more than one will be adopted.
		New cultivar partially replaces existing cultivar or cultivars.	Increase in biodiversity over medium to long term.
High potential		New cultivar completely replaces existing cultivar.	Biodiversity over time increased
		Replacement rates increased.	Increase in biodiversity over time.
		Greater number of cultivars adopted than those existing.	Likely because farmers exposed to a greater choice of cultivars.

Often the most important variable will be the range of diversity that farmers can be offered in the 'basket of choices. The more variability in the basket for quality traits, and the better the adaptation of the cultivars to the local environments, the more likely that several cultivars will be adopted. The basket of choices is likely to be larger when there are local breeding programs with specific objectives producing a range of products, rather than networked breeding programs targeting the production of cultivars with wide adaptation.

The greater the physical diversity in the environment, the more likely it is that more than one cultivar will be adopted by farmers. Diversity of economic use will also make it more likely that several cultivars are found acceptable and are adopted. Often farmers will prefer different grain types for different purposes. High-yielding cultivars with poor quality grain may be grown as a cash crop or to reduce risk, whilst cultivars with high quality grain will be grown for home consumption, and for special social and religious occasions.

Nonetheless, when conditions are right for several or many cultivars to be adopted then it is likely that great diversity already exists in farmers fields. When PVS is employed in areas where highly variable landraces are grown and there is no or little adoption of improved cultivars then its success will reduce biodiversity. This dilemma is faced by NGOs that wish to conserve biodiversity and help resource-poor farmers. One example of the problem is given by Cromwell and Wiggins (1993). An NGO, the Save the Children Federation in The Gambia, were faced with "the dilemma of seeing just one of its introduced rice varieties almost completely replace the range of local varieties previously grown, which were no longer suitable because of declining rainfall."

Biodiversity can also be considered over a wider area such as at the national level. We can assume that the widespread adoption of participatory methods will increase the replacement rate of cultivars, so that the average age of cultivars grown by farmers will be reduced and biodiversity over time increased. It is also likely that adoption ceilings of improved cultivars will increase at the cost of reduced biodiversity. Overcoming inefficiencies that limit the adoption of improved cultivars to relatively small areas will be balanced by the adoption in any area of a greater number of cultivars. PVS reduces biodiversity when a cultivar is adopted over a wider area, and a good example is the rapid adoption of Kalinga III in western India following PVS by farmers when the cultivar was previously only released in eastern India (Joshi and Witcombe, 1995). However, in the longer term, PVS should have only beneficial effects on biodiversity. If many more farmers are exposed to many more cultivars, the number of cultivars adopted will increase and the patchwork of cultivars between fields, districts and regions will increase in complexity.

Participatory plant breeding

Table 6. Comparison of farmer participatory varietal selection and participatory plant breeding in relation to biodiversity
Participatory varietal selection	Participatory plant breeding
No effect on intra cultivar variability.	Increase in intra-cultivar variability.
Smaller increase in number of cultivars in cultivation.	Large increase in number of cultivars.
Fairly uniform impact in increase in cultivar number over large areas.	Variable impact. Maximum and considerable impact in participating villages. Number of new cultivars decreases with increasing distance from these villages.
Breeding strategies to produce finished cultivars unchanged.	Breeding strategies changed. All changes tend to favour increase in biodiversity.

In predominantly self-pollinating crops, the adoption of PPB brings a change of methodology (Table 7). Instead of conventional pedigree breeding, bulk population breeding is employed whereby farmers mass select within segregating populations, such as the F4 bulk families used by Sthapit et al. in Nepal. Hence PPB increases biodiversity at the intra-cultivar level overcoming the 'disease of cultivar uniformity (Lopez, 1994). Intra-cultivar variability in the form of multiline cultivars is recommended as a strategy for reducing disease (Browning and Frey, 1981), but in this method intra-cultivar diversity is deliberately minimized apart from variability for disease resistance genes. The intra-cultivar diversity generated from PPB is more akin to a varietal mixture and such mixtures have been effective in disease control (Wolfe, 1990).

Table 7: The influence of participatory plant breeding on cultivar uniformity in predominantly inbreeding crops and its implications
Participatory plant breeding	Conventional breeding
Bulk population breeding is used since it is suitable for participatory breeding. It requires less resources than pedigree breeding.	Pedigree breeding may be more effective, (but pedigree breeding not likely to be more cost effective).
Method does not produce a pure-line cultivar. Increases biodiversity within the cultivar but causes seed certification difficulties. It is possible to overcome this by reducing diversity; single pure-line cultivar, or a version selected for uniformity, could be produced using little extra resources.	Method produces uniform pure-line cultivar. Procedures for testing release, and certification of cultivar already in place.
Biodiversity promoted in participatory villages both within and between cultivars.	More difficult to monitor spread of any variety in participating villages because of biodiversity created. Concern on part of breeders that farmers may have produced many varieties so which one to identify and promote? However, any improved preferred variety is better than none.

PPB is also a logical second stage to PVS, and if the appropriate breeding methods are employed then it comes closest to the ideals of genetic conservation. After PVS has been successful, the farmer-preferred cultivars can be crossed to other materials for farmers to select in the progeny. Breeding strategies will involve crossing the cultivar identified by participatory varietal selection (termed the PVS cultivar) with landraces and with high-yielding released cultivars. In the first strategy, the landrace is chosen as a parent to give genes for adaptation, and, in the second, a released cultivar is chosen to give genes for high yield potential. When landrace x PVS cultivar crosses are used and there is maximal farmer input in the breeding (the last method in Table 4). then we have a breeding strategy that most closely resembles in situ conservation of landraces. Farmer experimentation on naturally existing genetic variation has produced landraces, and this method enhances such farmer experimentation. Genetic variation is increased by the hybridization between the landraces and the PVS cultivar, and selection procedures by farmers and farmer awareness are enhanced by interaction with scientists. Nonetheless, in this process there is a possibility that some useful genes present in the landraces will be lost so ex situ conservation will still be desirable. Certainly, if there is a desire to wholly preserve the existing landraces, then ex situ conservation is essential. We can say that in situ PPB conserves genetic resources in farmers fields whereas ex situ conservation preserves genetic resources.

References

Cromwell, E. and S. Wiggins, 1993. Sowing beyond the state: NGOs and seed supply in developing countries. London:Overseas Development Institute.

Chambers, R., 1989. Institutions and practical change: reversals, institutions and change. pp. 181-195. In, R. Chambers, A. Pacey, L. A. Thrupp, eds. Farmer first. London:Intermediate Technology Publications.

Frey K.J, J.A. Browning, and M.D. Simons, 1977. Management systems for host genes to control disease loss. Annals of New York Academy of Sciences, 287:255-274.

Galt, D.L., 1989. Joining FSR to commodity programme breeding efforts earlier: increasing plant breeding efficiency in Nepal. Network Paper 8. Overseas Development Institute, London:Overseas Development Institute.

Joshi, A and J.R.Witcombe, 1995. Farmer participatory research for the selection of rainfed rice cultivars. Proceedings of the IRRI Conference 1995. Fragile lives in fragile ecosystems. Los Banos: International Rice Research Institute (in press).

Lopez, P.B., 1994. A new plant disease: uniformity. Ceres, 26:41-47.

Maurya D.M., A.Bottrall, and J. Farrington, 1988. Improved livelihoods, genetic diversity and farmers participation: a strategy for rice-breeding in rainfed areas of India. Experimental Agriculture, 24:311-320.

Priestly R.H. and R.A. Bayles, 1980. Varietal diversification as a means of reducing the spread of cereal diseases in the United Kingdom. Journal of the National Institute of Agricultural Botany, 15:205-214.

Priestly R.H. and R.A. Bayles, 1982. Evidence that varietal diversification can reduce the spread of cereal diseases. Journal of the National Institute of Agricultural Botany, 16:31-38.

Salazar, R., 1992. MASIPAG: alternative community rice breeding in the Philippines, Appropriate Technology, 18:20:21.

Sperling L, M.E.Loevinsohn, B. Ntabomvura, 1993. Rethinking the farmers' role in plant breeding: local bean experts and on-station selection in Rwanda. Experimental Agriculture, 29:509-519.

Sthapit, B.R., K.D. Joshi and J.R. Witcombe, 1995. Farmers' participatory high altitude rice breeding in Nepal: providing choice and utilizing farmers' expertise. Proceedings of workshop "Using Diversity": enhancing and maintaining genetic resources on-farm. New Delhi, (this volume).

Thakur, R., 1995. Prioritization and development of breeding strategies for rainfed lowlands: a critical appraisal. Proceedings of the IRRI Conference 1995. Fragile lives in fragile ecosystems. Los Banos: International Rice Research Institute (in press)

Weltzien R., E., M.L. Whitaker and M. Dhamotharan, 1995. Breeding pearl millet with farmers in Rajasthan. Proceedings of workshop "Using Diversity": enhancing and maintaining genetic resources on-farm. New Delhi, (this volume).

Witcombe, J.R., A.G. Bhasker Raj, A.J. Packwood, and D.S. Virk, 1995. Small farmer seed supply: reforming regulatory frameworks for testing, release and dissemination. Final technical report of the ODA funded project. CAZS, University of Wales, UK.

Wolfe, M.S., 1990. Intra-crop diversification: Disease yield and quality. In Crop protection in organic and low input agriculture-options for reducing chemical usage. British Crop Protection Council monograph 45:105-114.

Worede, M and H. Mekbib, 1993. Linking genetic resource conservation to farmers in Ethiopia. In W.de Boef, K. Amanor, and K. Wellard, eds., Cultivating knowledge: genetic diversity, farmer experimentation and crop research. London:Intermediate Technology Publications.

Footnote:

Document(s) 19 de 38