The joint development of JM-12.7: A technographic description of the making of a bean variety

Abstract

Participatory Plant Breeding (PPB) is one of the areas of Participatory Technology Development (PTD) in which collaboration of researchers and farmers has been reported as quite successful although its institutionalization remains problematic. This paper aims to contribute to better understanding of PPB processes. It focuses on the practices of developing a common bean variety (Phaseolus vulgaris L.) by a group consisting of a breeder, farmers and NGO technician in northern Nicaragua. The description is an example of a technography and uses the concept of boundary object to analyse how actors come together around a shared goal and how their knowledge and practices are combined in the material making of five varieties, eventually leading to JM-12.7 as a formally released variety. The material making of five bean varieties is central in the first part of the process and shows how in practice different knowledges within the group interact. The second part of the process leads to distinguishing socio-political boundaries. The formal registration of JM-12.7 required crossing of these boundaries and prompted the reorganization of the group into a co-operative.

Keywords:

• PPB
• Variety development
• Common bean
• Boundary object
• Social making

1 Introduction

Over the last 20–30 years, Participatory Plant Breeding (PPB) has developed as a complementary strategy in crop improvement [1–3], in response to the recognition that formal plant breeding in the Green Revolution era has not provided small-scale farmers in complex, variable and marginal environments with suitable varieties [e.g.,4,6]. Lack of understanding of the conditions under which small-scale farmers grow their crops is explained as a main cause for the shortcomings of formal breeding in developing countries. Seeking farmers’ involvement in breeding was a logical idea: they know best what suits their socio-economic situation, know the agro-ecological environment in which they grow crops, and crop evolution is the evidence of their expert knowledge in seed selection [7]. The concept of PPB fitted well with the general idea in the area of agricultural research and development that farmer participation could make technology more relevant to the users and, in addition, could empower farmers and rural households [e.g.,5,8]. Since the discussions about the concept of PPB and its challenges in the 1990s [9–11] a wide range of initiatives has been implemented, with different crops and in different contexts [1,2,12,13]. The initiatives included farmer-led and breeder-led ones, and farmers’ participation varied between ‘consultative’ and ‘decisive’; farmers were involved in identifying parental materials and making crossings, in selection in early generations of segregating materials (PPB) and advanced materials (PVS), in farmers’ fields and on-station, and in a range of crops [e.g.,1–3,10]. Two issues are relevant in a reflection on PPB as an alternative approach in plant breeding: its success and institutionalization.

Although there is large variation in the way farmers have been involved, many if not all reports on PPB initiatives mention success: varieties that were developed outperformed in farmers’ fields those available from the conventional breeding programmes, and farmers who participated were strongly empowered [e.g.,1,13]. How can the successes of PPB be understood whereas on the other hand the interaction between farmers and researchers in Participatory Technology Development (PTD) is often reported as problematic [e.g.,11,14,15]. Because farmers and researchers are part of different social worlds, they do not easily meet and when they do, collaboration is often accompanied by tension, misunderstanding and different expectations because their ways of knowing and doing are different [14–18].

Despite the repeatedly reported successes of PPB, its institutionalization so far remains problematic, as one can learn from the breeders and other players in this field.¹ The problems of institutionalization (i.e., making the alternative way of doing normal practice) are mostly attributed to the inflexibility and culture in research institutes, no incentives for the researchers to truly engage with farmers, and a lack of an enabling political climate [13,19]. There is, however, little information on the concrete limitations for researchers and institutions to make farmers’ participation part of their normal practice.

This paper looks at knowledge interaction and institutionalization in a PPB case in northern Nicaragua using detailed analysis of the process and practices of making a common bean (Phaseolus vulgaris L.) variety by a group of actors. The analysis follows a technographic approach as described by Richards [20] and Jansen and Vellema [21; in this issue]. It presents PPB as task-oriented and collaborative work of a group of actors, in this case a number of farmers, a breeder and an NGO technician. I collected the data over a period of 12 years during visits, meetings and through interviews with the people involved.

To analyse the making of the bean variety I have made use of ‘boundary’ concepts. Different authors have used these concepts in somewhat variable ways to look at collective work of actors with different knowledges and to point out the difficulties involved [e.g.,17,22,24]. In this case I constructed the analysis around the boundary object as being the material or abstract object around which people coalesce and act [22,23]. The boundary object does not require consensus for successful collaboration and is sufficiently abstract or flexible to be part of different social worlds [22,23]. I looked at how the ‘improved bean variety’ functions as boundary object, and how it transforms from a shared goal into five candidate varieties and ultimately in the release of JM-12.7. First, I analyse how the actors come together around the boundary object, i.e., the improved bean variety, and how the different roles and knowledges of the farmers, breeder and technician become apparent in the material making of five bean varieties. In the following part of the process these five varieties and the group face a series of challenges of socio-political and institutional character. These challenges can be seen as the boundaries of the project space that the group and the varieties needed to cross in seeking acceptance by a wider social world.

2 The Nicaragua bean case: the task and the team

In 1998, farmers in Pueblo Nuevo and Condega, two neighbouring villages in northern Nicaragua, were having increasing difficulty growing common bean (P. vulgaris L). The pressure of Golden Mosaic Virus (GMV) had increased rapidly and none of the commonly grown bean varieties showed an acceptable level of resistance. Beans are, together with maize, the most important subsistence and cash crops for small-scale farmers in Central America. In Pueblo Nuevo and Condega, farmers traditionally produced maize and beans with few inputs because of the high risk of crop failures (droughts, hurricanes) and the unpredictability of market prices. The increase of GMV pressure was linked to the success of tobacco and tomato production in the area, which had led to an increased pressure of aphids, the vectors of the virus. When the aphids had grown resistant to chemical control the virus spread rapidly. This affected not only the tomato and tobacco crops, but also beans, which are susceptible to GMV as well. The fields in the lower and warmer parts of the villages were suffering most and growing a bean crop there had become impossible. Only the improved variety DOR 354 showed some resistance, but the red colour of its seeds was too dark to get a good price from the traders who were serving an urban market that was used to bright-red beans. In addition, farmers and their wives did not like the taste and texture of DOR 354, making it ill-suited for domestic consumption. When there was a Dutch initiative to pilot PPB approaches, the NGO CIPRES proposed a project to develop a GMV-resistant bean variety. A breeder from the national research programme (Instituto Nacional de Tecnología Agropecuaria, INTA) and 45 farmers in Pueblo Nuevo and Condega responded to the invitation to participate.

The project breeding team consisted of 45 farmers, a technician of the NGO and the INTA bean breeder. The first time they came together, in September 1999, in a start-up meeting organized by the NGO technician, none of them clearly understood what they were engaging in: PPB was new to all of them and there were no earlier experiences on which they could draw. But each of them was motivated to face the challenge. Farmers knew the type of bean they were looking for: drought resistant, adapted to low soil fertility, resistant to GMV and with a particular red colour seed for which middlemen would pay the full price. They had between 0.7 and 3.5 ha of land with an average of 2 ha, i.e., slightly higher than the average farm size in both villages. The proposition to develop their own varieties sounded quite unreal to them, but they trusted the NGO; many of them had been involved in earlier seed projects with the same NGO and knew it was serious. The CIPRES staff in Managua saw this donor-funded project as an opportunity to emphasize the importance of agricultural technologies, in particular seed, for the small-scale farmers and their contribution to national food security. In addition, it fitted their general rural development strategy to organize farmers into groups and co-operatives. Their office in Pueblo Nuevo had an office co-ordinator, and with project funds they hired a local technician who had received formal training in agriculture but till then had made his living from growing beans and tomato on rented land and from day-labouring for others. He was also well known in the village because he played in the local baseball team and had been a prominent Sandinista fighter during the contra-revolution. The breeder was sent to the meeting by his superior who had received an invitation from CIPRES for INTA to collaborate in the initiative. Later, the breeder said that he had heard something about PPB, but had never considered practising it. During the meeting, his interest and eagerness to develop varieties that farmers would adopt overcame his reservations. The limited budgets of INTA confined most of his work to the research station and he could only work outside the station when there were special programme resources from international research institutions like the Centro Internacional de Agricultura Tropical (CIAT) to pay for petrol and per diems. His engagement with the farmers and CIPRES was a personal decision. Only after insistence of the CIPRES staff with his superiors, a memorandum of understanding was signed with INTA that officially allowed the breeder to participate on the condition that he would not use his INTA time and that the NGO would pay the petrol for the INTA car he used.

3 The material making of five varieties

3.1 The experimental design

During the set-up meeting of September 1999 the group decided on the first step in the development of an adapted GMV-resistant bean variety. The breeder suggested crossing the variety that was most popular in the villages before the GMV became problematic, with GMV-resistant advanced lines he had received from the CIAT bean breeding programme in Cali. Because the crossing and production of sufficient seeds to plant trials would take more than a year, he proposed to provide the farmers with F3-seeds of 15 bean families, originating from 3 crosses that had various sources of GMV resistance [26]. With these, the breeder, the NGO technician and the farmers would start to experiment and learn about evaluation and selection until they had sufficient seeds originating from the new crosses.

In the same meeting it was agreed who would host the experimental plots. The breeder had suggested plots at different altitudes to capture the different micro-climates in the village. Five farmers agreed to host a plot on their land, each plot representing a different micro-climate (Table 2). Field plots were laid out in January 2000 on the basis of a plan that the breeder had explained and left with the NGO technician. The technician helped the farmers to mark the field by measuring the length of the rows and counting the rows to be planted, with a rope and chalk provided by the breeder. In each plot 520 seeds per bean family were planted and signs were put up with numbers from 1 to 15, corresponding with the 15 families. Farmers managed the plots in their own usual way but because they were looking for a bean variety adapted to their conditions, they agreed not to use any fertilizer and to irrigate only after consultation. Regular meetings were organized by the NGO technician to which all 45 farmers were invited and where the breeder explained the very basics of genetics and selection, using examples and a vocabulary that farmers could understand. In these meetings, they also jointly decided how to plant, evaluate and select. Field visits were organized to involve the other 40 farmers in the group in the evaluations of the materials.

Table 2 Yield (kg ha⁻¹) of the five best bean families selected by the five farmer–breeders in evaluation plots on their own farms (no repetitions).

Farm	Location and altitude (masl)	Origin of the line (farmer–breeder)
		Pedro Gómez	Juan García	Jose M. González	Santos L. Merlo	Jairo Videa	Test variety ^b	Average per farm
Pedro Gómez	La Lima, 1000	969	1948^a	839	1098	1164	1551	1514
Juan García	Santa Rosa, 850	2005	2717^a	1551	2069	2127	1875	2469
José M. González	Paso Hondo, 630	969	2134	2522^a	2134	2263	1616	2327
Santos Luis Merlo	El Rosario, 650	1035	1180	1016	1722^a	1275	1057	1457
Jairo Videa	Rio Abajo, 600	2328	1357	1616	1482	2522^a	2269	2415
Average per line		1461	1500	1834	1701	1870	1674	2016

a The best yielding material on the farm of each of the five farmer–breeders.

b On the farm of Santos Luis Merlo the variety DOR 364 was used as the local test variety, whereas on all other farms it was the variety INTA-Masatepe.

The five farmer–breeders (i.e., the farmers who hosted the plots) collected data on emergence, flowering, GMV resistance, and yield, according to the instructions of the breeder. Data collection was very time-consuming, but they were all convinced that it was necessary to keep good records if their new variety was to be taken seriously. One of the farmers who could not read or write made sure that his wife assisted him in taking notes, and the technician helped another illiterate farmer in keeping records. Based on the selection scheme of the breeder, the five farmers selected within or among the families. They evaluated the bean plants in their fields taking into account how the plants held up against the GMV, as well as final yield and seed colour. The screening for GMV was based on a scoring that the breeder had explained to them and which they carried out regularly. The bean yield was weighed after harvesting and threshing. Colour was a more subjective criterion. Although the criteria of the entire group were important for forming an opinion, each of the five farmers eventually decided by himself which materials to eliminate and which to keep. The elimination of selections proved difficult for the farmers: they felt that each of the selections could potentially be useful and were reluctant to discard anything. In the second season, each of the five farmers had 300 entries or more in the field. According to the breeder, they retained far too much material, despite his advice to keep only the best. Over the following seasons, the farmers reduced the number of entries through selection from 300 to around 80, 30, and 7, eventually ending up with their 5 best families (Table 1). To select their best performing lines, the farmer–breeders considered the opinion of other farmers but they mostly relied on their own criteria. They did not use the field data they had recorded following the breeder's instruction.

Table 1 Selection scheme of five farmer–breeders in Pueblo Nuevo and Condega, northern Nicaragua.

The five participating farmer–breeders followed, with some small variation, very similar selection procedures in the following primera^a and postrera^b planting seasons:
2000	Primera: each of the 5 participating farmers planted about 520 seeds (4 rows of 5 m) of 15 F3^c progenies/families. For each of the progenies the farmers selected up to 20 plants, resulting in 300 lines for the next planting
	Postrera: 300 progeny F4–F5 lines (rows of 30 m) → selection of about 80 lines
2001	Primera: 80 selected progeny F5–F6 lines were planted → 30 best lines were selected
	Postrera: 30 selected F6–F7 progeny lines planted → about 7 best progeny lines selected
	Irrigated trial after postrera: 7 F7–F8 selected progeny lines planted by 1 farmer → 5 lines selected for evaluation in the fields of 4 other farmers
2002	Primera: 5 F8–F9 selected progeny lines planted → the 2 best lines were selected. Hence at this point, each farmer contributed one single line derived from his individual selection from the common starting material of 15 families provided by INTA in 2000
	Postrera: 26 trials with the best selection of each of the 5 farmers being planted on farms in and around the 2 communities. In addition the best improved bean variety INTA-Masatepa was planted as the control variety
	The five farmer selections and the control were evaluated for yield and for consumption quality by the farmers and their wives
2003	Primera: 22 more trials were planted with the 5 selections, with INTA Masatepa as the control. This completed the evaluation data of a total of 48 evaluation trials. The results of the 48 comparative trials were statistically analysed and the 2 best selections identified
	Postrera: The two selections were named as varieties (JM-12.7 and Santa Elena) and multiplied for further distribution
2004	October: informal release of the varieties

a May–July planting season.

b September–November planting season.

c F3 stands for the third generation after the crossing of two parental lines. In a self-pollinating crop like common bean this implies that the material is still strongly segregating.

In the fifth growing season, the first planting of 2002, each of the five farmers provided seed of what he considered his best line and these five lines were compared with the best available commercial variety, INTA Mazatepa. Five evaluation plots were laid out on the farm of each of the participating farmers (Table 2). The breeder had suggested coded labelling of the varieties to avoid biased evaluation by the larger group of farmers; only he and the technician having the key to the codes. The five farmer–breeders considered the coding a funny formality as the five varieties were very distinct to them, so that they easily recognized ‘their’ variety. The differences between the varieties were a logical outcome of the farmers’ knowledge, selection criteria and environments. The following examples illustrate how these elements came into play in the practices of the farmers Santos Luis Merlo, José Manuel González and Juan García.

3.2 The farmers

The farm of Santos Luis Merlo consisted of several plots of land close to the river, with a total area of 7 ha, usually just enough to produce the maize and beans for the family, and selling some surplus in most years. The soils of Santos Luis Merlo's plots were poor and sandy, so that for him yield and drought stress were the principal selection criteria, next to GMV resistance and seed colour. He looked for plants with no more than 3 pods per branch because his experience told him that these would fill-up even if the rains stopped early and the growing season was cut short. With more pods per branch the seeds would be wrinkled and have no volume. Such a variety would threaten the food security of the family.

José Manuel, who had slightly more land (10.5 ha) with better soils was very keen on finding a bean variety that would fetch a better price in the negotiations with the middlemen since the commercialization of beans formed an important part of the family's cash income. When he saw his bean plot suffering from drought stress in one of the selection seasons he realized that it was often not the drought that made his crop fail, but the lodging of the plants in the first torrential rains after a drought period. Therefore, he decided to inundate his field as a way of mimicking a torrential rain. He then selected plants that had not fallen over because they apparently had a well-anchored root system. The NGO technician had told him not to irrigate the field unless the breeder had agreed. The technician was shocked when he visited and saw that the field had been irrigated. He decided to keep José Manuel's action out of the reports because he considered it a big mistake.

Juan García farmed more than 15 ha of fertile hill-side land where rain and mist are more frequent. He selected a larger and more open plant type. He said that plants needed an open structure to allow ‘the air to flow’ and the leaves to dry quickly, thus reducing disease pressure.

As a result of the farmers’ individual selection environment, criteria and practices, the five bean lines, derived from the same crossings and selected with the same basic selection criteria (resistance, yield and seed colour) were nevertheless quite distinct in plant and grain type. Although genetic analysis of the five lines (using a PCR-based RAPD) did not show significant differences among the 5 lines [27], they nevertheless carried the visible fingerprints of their selectors and their respective environments.

3.3 The breeder

Throughout the process, the breeder regularly visited the fields and participated in meetings, providing explanation and advice on the planning, the evaluation and selection practices and the lay-out of plots. Over time his role changed. In the beginning, the breeder's suggestions weighed heavily. Although he did not try to impose his ideas on how things should be done, neither farmers nor the technician had other sources of knowledge to rely on. However, as the farmers became more knowledgeable, through the teachings of the breeder and their own experiences, they became more confident, started to formulate their own criteria and act accordingly. The breeder remained, however, important in the overall planning of the process. Later on, he took charge of collecting the yield data in the comparative evaluation trials (see below).

3.4 The technician

The NGO technician facilitated and co-ordinated the entire process. He organized the meetings where results were discussed and plans were made. He also arranged the field days on which other farmers evaluated the materials of the five farmer–breeders. He maintained the contact with the breeder, visited the farmers to notify them of an upcoming meeting and stimulated them to attend. He also monitored the selection plots of the five farmer–breeders. In practice this meant that he visited them regularly and had to come up with solutions for all kind of problems. When a farmer was losing confidence or when he was concerned about the drought affecting his plot, the technician sought to motivate the farmer, even when at times he himself was unclear about what should be done. He also needed to attend to a whole range of other logistical worries, like how to get the cash from the NGO project budget to buy fuel for the pump when irrigation was really critical; support the farmers when a hurricane hit the village and the river washed away their land; transport farmers’ family members to the hospital; and help out when there was a shortage of cash in the house or problems with alcohol in the family. These supporting actions had no direct relations with the farmers’ breeding and selection activities, but did threaten their capacities to undertake them. Farmers have frequently emphasized that without the technician, they would not have managed to carry on with the project. There were too many moments when they did not know what to do or how to solve a problem they had. They felt they could always rely on the technician because he understood their needs. The technician explained that he knew how to talk with the farmers and how to motivate them because he shared their life in many different ways: some farmers he knew well from the baseball games, others he knew from the time of the contra-revolution and when he worked on the land before CIPRES hired him as technician. Because, as he said, “I know how they talk and how they think” and because he had basic understanding about genetics from his agricultural training, he was the one who best understood the breeder and translated the breeder's knowledge for application in the field. The breeder was explaining the crossing and selection of plants in very basic language so that farmers could understand. But his explanations were hard to recall for the farmers when they were confronted with the real situation in the field. Making decisions without the breeder being present was therefore to a large extent depending on the technician's input, in particular in the early phase of the process.

4 Institutionalization

4.1 The selection

As mentioned above, the five selected lines, one from each of the five farmer–breeders, were first evaluated on the farms of the farmer–breeders. The results of these five trials did not clearly show one single best yielding variety (Table 2). It is interesting to note that on four of the five farms the best performing line was the one that had been selected on that specific farm. However, since the trials did not have replications (there was not enough seed to do so), it was not possible to test for statistical significance. These results could point to a narrow adaptation of the lines as a consequence of selection in distinct environments by a single person with a unique ideotype in mind. One possible implication of this was that none of the varieties would have a wide adaptation. Continuing with five varieties was not an option. The farmers wanted to produce and sell seed of the new bean varieties on the national market and the breeder helped them realize that they could not possibly manage with five different varieties. In order to decide which varieties to continue with, the breeder suggested testing on a wider scale. In addition, he said, the data from these evaluations would help them later in a possible formal registration of the variety. In 2002 and 2003, they laid out a total of 48 trials in different fields in and around Pueblo Nuevo and Condega. Most of the trials were on the farms of the 40 other farmers who had participated in the process since the beginning. Following these trials, the entire group was involved in the decision to select the lines developed by José Manuel García and Juan García. The final decision was based on the yield and greater stability of these lines over various environments, as shown by the breeders’ statistical analysis [26]. The variety of José Manuel was found to be the most stable, meaning that it had the widest adaptation and largest commercial potential. It was suitable for lower and warmer areas, the prevailing bean production conditions in Nicaragua, whereas the variety of Juan García was better for higher and more humid conditions. With the decision to continue with these two lines, they also had to name them. José Manuel named his variety Pueblo Nuevo JM-12.7, pointing to his village, his own initials and the row from which he had selected the plant early in the process. He said he wanted to use codes of letters and numbers just like he had learned that breeders do. Juan García gave his variety the name Santa Elena, after his daughter Elena.

4.2 Formalizing JM-12.7 and the formation of the co-operative COSENUP

In October 2004, the two bean varieties were informally launched in a big celebration organized in Pueblo Nuevo. The event was covered by the local radio station and newspaper. During the following months, handfuls of seeds were freely distributed at regional agricultural fairs to interested farmers. In order to get organized for larger-scale seed production, the farmers established a co-operative. An important reason for this was to create a legal entity that could formally register the variety in name of the farmers. Registration was necessary if they wanted to commercialize its seeds. An important channel for commercializing their seed was the government seed programme known as ‘Libra por Libra’ and this programme required formally certified seed. In addition, the registration of the variety through the co-operative had symbolic and political importance for the farmers and the NGO. For these reasons, the staff of CIPRES initiated the registration of the variety JM-12.7, on behalf of the co-operative [20]. They had jointly decided not yet to apply for registration of Santa Elena, because the costs of registration were considerable. Santa Elena only out-yielded JM-12.7 at higher elevations and this implied that it would probably have a smaller market, and thus a smaller chance to recover costs of registration and maintenance.

The CIPRES staff had, however, difficulties in finding out from the Ministry of Agriculture and Forestry (MAGFOR) and the seed office CONASEM the meaning of certain words in Law 280 and the steps to follow in the bureaucratic procedures [28,29]. Law 280 regulates seed issues in Nicaragua and indicates that both the variety and the co-operative had to be registered. The co-operative would have to register as a seed producer and demonstrate that it had expertise in the maintenance of varieties. For production of certified seed the variety had to be registered as well and the co-operative would have to apply production measures that are demanding and costly. Production of certified seed involves registration of seed production plots, control in the field by the authorities, germination tests, bagging and labelling. With the application for variety registration, the co-operative also needed to submit a description of the variety indicating its distinctness from other varieties, results from field evaluations and the proof that they had at least 800 kg of quality seed of the variety for commercialization. The breeder compiled the statistical data from the 48 evaluation trials in the required format as well as the requested description of the variety. In the description of the variety it was significant that the plants of variety JM-12.7 had a red marking on the stem and that there was also a red spot in the curvature of the dried pods. These red marks distinguished it from other varieties, an important condition in the variety registration procedures of UPOV, which is used in most countries in the world and which is known as the DUS condition (Distinctness, Uniformity and Stability).² The breeder had experienced earlier occasions in which a candidate variety with very good characteristics could not be registered because it could not be visibly distinguished from other already registered varieties.

In January 2007, COSENUP had more than 2000 kg seed of the new variety stored in the silos but no options for commercialization. However, just when the seed activities of the co-operative seemed about to fail, there was an important change. The 2006 national elections were won by the Sandinista party and as a consequence the agricultural policy environment became more favourable for small-scale farmers. This included government support for initiatives from CIPRES, which had close links with the new government. In the months after the new government took office, crucial positions in the Ministries were filled by officials who followed the new policy lines. As a consequence, the lower-ranking officials, like those in the Ministry of Agriculture, became more co-operative and willing to push the case of JM-12.7 [28]. In April 2007, four months after the new government had taken office, the official registration of Pueblo Nuevo JM-12.7 was announced by CONASEM, quite unexpectedly for the CIPRES staff and COSENUP members. The variety was registered as Pueblo Nuevo JM (Fig. 1). Later that year, COSENUP sold practically all their seed in stock to an NGO that used it for a project elsewhere in the country.

Fig. 1 The official registration of variety JM-12.7 as Pueblo Nuevo JM.

4.3 Recent developments

Currently, the co-operative is doing good business. Since 2007 they have increased their yearly production and sales of certified seed, principally of the variety JM-12.7, now with the official name of Pueblo Nuevo JM. Farmers have increased their bean yields because they have better seeds, but more importantly, the price they receive for seed is good. This means a significantly higher income for the farmers who invested in the co-operative. Most of them have been able to improve their houses, buy more tools and provide better for their families [30]. The co-operative now counts 82 members and has a full-time administrator to manage the business, together with a technician, an accountant and a committee that is in charge of the routine decisions. They have been able to access credit and invest in a ventilated building and silos to store the seed. To meet the demand for seed they complement their own production with contracted production from other farmers and co-operatives in the region. The government has also become an important client. It buys seed to distribute through the programme ‘Hambre Zero’ (Zero Hunger)’, which replaced the ‘Libra por Libra’ seed programme [28]. The technical committee of the co-operative feels they have good relations with the officers from the seed certification unit of the Ministry who inspect the bean fields at least once during the growing seasons. The technician now supports the co-operative in planning and running the business. In addition, he helps spreading the approach to other parts of the country, together with the breeder and staff from CIPRES offices in other locations. The official INTA policy is to collaborate with farmers ‘the way CIPRES has been doing it’. INTA staff works in the villages with promoters who are laying out evaluation trials that are meant to jointly evaluate with farmers which varieties do best in their environment. The approach is also used for sorghum, rice and maize.

From the crossings between the old popular local bean variety and sources of GMV resistance, made early after the start the programme (see Section 3.1), promising materials have been selected, like Luisito, Siete Panas, Marrojo and Rio Rojo. Only JM-12.7 is officially registered; the other varieties are informally released and included in the variety trials that CIPRES and INTA jointly implement in other communities. So far, none of the more recently selected varieties have surpassed the qualities of JM-12.7. Especially the GMV resistance of JM-12.7 is superior. The crosses of local varieties proved less stable in their GMV resistance than the varieties selected from the 15 families with which they started in 2000. The breeder explains this as a consequence of the fact that the local varieties from which they departed were only single-crossed with the sources of GMV resistant germplasm. The candidate variety of Santos Luis Merlo is still grown by neighbouring farmers and by relatives and farmers in an area with similar growing conditions and to whom he gave seeds. They appreciate the variety for its drought tolerance.

5 Changing knowledges

The breeder reckons that he learned a lot from the farmers. He says that because they worked so closely together, he now understands local production conditions much better and looks at the performance of bean crops differently. He does not only consider GMV resistance and yield, but realizes how important it is to combine these characteristics with drought tolerance, good yields on poor soils and the right seed colour. He also recognizes that farmers are in a better position and have ‘a better feel’ for evaluating some of these characteristics than he, especially now they have learned from the PPB activities. However, he also realizes that his contribution to the making of varieties will not become redundant when farmers become involved in breeding: access to valuable germplasm and testing for wider adaptation remain essential steps in the crop improvement process. He continues working with the farmers and through him the farmers also collaborate with other breeders from the region and CIAT's international bean programme. Whereas he currently enjoys political support, he knows that many of his colleague breeders still do not fully believe in his approach.

Through basic technical training in breeding, the farmers are more confident in discussing performance of the materials, resistance to GMV, tolerance to drought and their preferences for either evaluating early segregating families or advanced lines. Also the exchange visits and meetings with PPB-farmers in other countries in the region have made strong impact on the farmers and changed their social worlds. The farmers continue managing evaluation trials for the breeders in their fields and this provides them with access to the newest breeding materials.

The farmers say their learning has been very important and they are enormously proud of their results. They now understand that varieties do not just ‘come to them’, but that characteristics like seed colour, plant shape and yield are actually defined by the parent plants and by the selection thereafter. Apart from variety improvement in beans and maize, the farmers are now also considering PPB in other crops. According to the technician, they even discuss the colour and productivity of their chickens and the appearance and behaviour of their dogs and family members. He has also noted that the farmers are now looking at their crops more analytically and make decisions with more criteria and logical reasoning. Furthermore, it is important to them that breeders, authorities and policy makers take them seriously and recognize the contribution of small-scale farmers to national food production in Nicaragua. For them, the formal registration of JM-12.7 to the name of COSENUP as evidenced by the official document (Fig. 1) represents that recognition. But they have also learned that the development of the varieties is much work with little financial compensation. They recoup some of their investments by commercializing seed production. The farmers differ in opinion on the importance of the learning as compared with the commercial benefits, and the dependence on the breeder for future new varieties.

6 Regional programme

To understand the wider context of this case it is necessary to realize that the group work did not take place in isolation but formed part of a larger regional programme with PPB pilot projects in other countries. The programme originated from an initiative in 1998 of the Dutch Genebank – where I worked at that time. The initiative was supported by the Ministry of International Co-operation, which wanted programmes that combined their objective of poverty alleviation in developing countries with the commitment made by the Netherlands government in signing the Convention of Biodiversity (CBD, 1992). PPB was such an activity: it potentially combined the improvement of farmers’ livelihood with the conservation of agrobiodiversity. As representative of the Dutch Genebank I was mandated to find organizations interested in a regional PPB programme and support them in the formulation of a proposal. This suited my personal interest to explore the technical feasibility of PPB. Eventually, the Dutch government did not fund any of these projects in the programme because after a national election a new government changed the priorities in international co-operation. In the end the programme started with funding by a Canadian and a Norwegian donor organization.

The regional programme counts with many more PPB activities, most of them with common bean and maize, but also with other crops. The farmers, technician and breeder in Pueblo Nuevo have also been developing better maize varieties. The activities started with a similar set-up around five farmer breeders but developed much more slowly because of the cross-pollinating character of maize. They achieved developing two varieties, Pueblo Nuevo SL and Condega JG, but they have only been informally launched. Other groups of farmers in northern Nicaragua work together with a French breeder in improving rice and sorghum varieties [31,32]. In Honduras the bean breeder from Escuela Agrícola Panamericana (EAP) Zamorano collaborates with farmer research groups, i.e., Comités de Investigación Agrícolas Locales (CIALes) in marginal hill-side conditions to develop adapted bean and maize varieties [33,34]. Also in Costa Rica [35], Guatemala, El Salvador and Cuba [36,37], farmers and breeders jointly develop better performing bean and maize varieties. Together these initiatives form the Regional Collaborative Program for Participatory Breeding in Meso America.³ The programme was initiated in 2001 and is governed by an Assembly in which breeders, NGO technicians and farmers of each of the participating countries are represented. Every two years the Assembly selects a new Executive Committee. Farmers and breeders also organize joint presentations of their experiences at national and regional meetings of the Central American research community and other forums. There are regular farmer exchange visits between the countries, which have shown to be important for sharing experiences and building networks. Lately, farmers in the programme have started talking about a regional commercialization network to help find the best market opportunities for the seed and grain. Production of black beans for the market in Venezuela (enabled through recent commercial agreements between the current governments of Venezuela and Nicaragua) and links to the market in Costa Rica via the partners in the programme are examples of how the regional links have served the farmers of COSENUP.

The network that emerged from the PPB pilot projects in the region also serves the breeders. Apart from the functional relations with skilled farmers, they make use of each others’ expertise and access to germplasm. For example, the bean breeder from Honduras is the only breeder in the region with a crossing programme. He crosses local varieties that farmers in the different countries have identified, with the most advanced breeding materials from his own programme and CIAT. He multiplies the early generation materials and after a first screening for resistances in nursery beds then sends the seeds to the farmers who requested the cross. Similarly, the maize breeder in Guatemala has a crossing programme and generates progenies that are valuable for the farmers in other countries. The other breeders and NGO technicians in the programme function as intermediaries. All together, the regional network forms an alternative structure that functions partly in parallel to and partly integrated with conventional R&D activities, but highly depends on external donor funding.

7 Discussion

The technographic description of this PPB process shows that the making of a bean variety is more than a process of genetic selection. The germplasm with GMV resistance was a vital ingredient for the success, but the combined knowledge and practices of farmers and breeder defined the material form of the bean varieties they developed. The process did not end with the material making: to become accepted beyond the group of involved actors, the varieties faced a process of social selection. In the following I analyse more closely the knowledge interaction in the making of the five varieties and the socio-political and institutional mechanisms that played a role in the formalization of variety JM-12.7. For the analysis of the experiences I used the concept of boundary object to follow how the development of an improved bean variety as a shared objective created a space for collaborative work of farmers, a technician and a breeder [17,22,24,25], and resulted in five varieties and ultimately ended in the formal registration of one of them. I first will analyse the knowledge interaction in the material making of the five varieties in the first part of the process. I then look more closely into the second part of the process that leads to the formalization of one of the varieties.

7.1 Knowledge interaction in the material making of an improved bean variety

‘The improved bean variety’ functioned as a boundary object around which the farmers, breeder and technician formed the task group in this PPB case. At first sight, the making of a red bean variety with GMV resistance was a concrete clear-cut task, but the outcomes show the contrary: five quite distinct varieties ended-up in a comparative trial (Table 1), with the breeder acknowledging he would have selected a different type of variety. For the farmers the bean variety was naturally situated in their poor fields, often facing drought, fetching low market prices, and an essential ingredient in their daily dishes. For them the bean variety-to-make was therefore a variety adapted to these field conditions, and with a good seed colour and taste. The breeder would have ended up with another bean variety. His variety would have had good GMV resistance and would have been selected under a range of field conditions for the necessary broad adaptation. The variety would have been less adapted to low soil fertility and with a darker seed colour because he had not been aware of the importance of the last two characteristics for the farmers. For the NGO technician the material form of the improved bean variety was a variety that would yield well and fetch a good price; it meant in the first place a possibility to improve the livelihood of farmers of whom many were also his friends. For the NGO and the donors the variety-to-improve involved scientific and socio-political interests that brought them to initiate and support the collaborative work. This shows that without the concrete material shape being explicit, or rather because the concrete shape was not explicit, the boundary object provided space for a range of actors to coalesce around a shared objective.

The fact that there were five distinct candidate varieties, selected from the same crossings, is not only genetically interesting, but also shows that the concrete shape of the bean variety differed for the five farmer–breeders. These differences, which are expressed in the selection practices, are related to the different agro-ecological conditions in which the farmers farm and to the socio-economic role of beans in their household. For example, Santos Luis Merlo produces primarily for family consumption. In his drought-stressed and poor field, his selection of fewer pods per plant became the explicit criterion in selecting for the food security of his family. Only through asking Juan García about his selection criteria it became clear, even to him, that he focused on ‘open plant types’ because of the humid conditions in his fields. José Manuel's idea to flood his bean plot to simulate torrential rains after a dry period is another example of knowledge becoming explicit in the task of making the bean variety. José Manuel is also a farmer who commercializes more of his production than Santos Luis Merlo. This explains the fact that the variety he selected is of a commercially attractive grain type. These examples illustrate how the knowledges of the actors are situated in their respective natural and social worlds. They also show how these knowledges influence the making of a technology and become visible and explicit in the making.

Next to understanding how the knowledges combined in the making of the five bean varieties, it is important to understand how the knowledges affected each other through interaction. Literature frequently points at the conflicting character of knowledges from different social worlds [14,17,22,24]. In this case, however, the working across the knowledge boundaries was demanding and required commitment but was not conflictive: farmers combined criteria and practices with the breeders’ advices, planning and procedures, all aspects and phases with the technician as intermediate. Instead, there was recognition of each other's expertise and a growing realization of mutual benefit. In this realm, farmers started mixing their practices with those of the social world of the breeder: they took data, together with the NGO technician they picked-up on methodologies of field experiments and eventually in the name of JM-12.7 the coding-culture of breeders was practiced. In the process, also the knowledge of farmers’ knowledge transformed. Farmers have become more analytical and make decisions with more criteria. Whereas it was not studied in detail how their bean production practices have changed, the PPB experiences in combination with other project activities, have changed their social world. The breeder in turn was impressed by the knowledge of farmers, enriched his understanding of local production conditions and modified his weighing of different selection criteria. His knowledge did not conflict with the knowledge of the farmer world, but rather experienced tensions within his own institution where many of his colleagues still do not consider his practices as professional.

It is obvious that the knowledge boundaries between farmers and the breeder were not very tight. There were multiple passages through which knowledge crossed over into the other field and became incorporated. The permeable character of the boundary is to a large extent explained by the technician's role in translating the breeder knowledge, and supporting its application in the field with farmers. This role of the technician as knowledge broker, together with the shared goal, mutual respect and recognition of the complementarity of the knowledge emerge as crucial elements in the effective interaction of knowledges in this PPB case. In this respect, crop breeding may be a special knowledge field. For farmers who normally save their seeds for next planting, breeding knowledge closely relates to their own world of practice. For that matter, the explanations of phenomena that they observe daily in their seed and plants turn out to be true eye-openers for the farmers and are easily and eagerly accommodated. This can explain the repeatedly reported enormous empowering effect of PPB on farmers. In other fields of PTD, like soil fertility and pest management, this may be very similar.

7.2 Moving across social boundaries

When the five candidate varieties and the group sought acceptance in the world beyond the project space they faced a series of challenges. These challenges can be seen as social boundaries of the project space that needed to be crossed. These boundaries were related to the rules and regulations of the socio-political and institutional environment of which the selected varieties were to become part, and with which they thus had to comply. How these rules and regulations functioned as a selection mechanism became visible first in the testing of the five candidate varieties in the fields of other farmers in Pueblo Nuevo and Condega. Whereas the NGO and the technician saw the wider testing as a way to more closely involve the other 40 farmers who participated in the project since the start, for the breeder this was part of the normal practice to multiply seed of the candidate varieties and collect the field data for a final selection and a future variety registration. In the functioning as a social selection mechanism, the variety and seed regulations prompted the creation of the co-operative COSENUP in order to create a juridical identity for the formal variety registration.

The official document of the registration of JM-12.7 as Pueblo Nuevo JM (Fig. 1) is the proof that the product of the alternative, participatory way of making passed the last social selection criterion and was accepted by all actors involved, including the wider socio-political and institutional world. JM-12.7 is the form of the boundary object that complied with requirements of farmers, the technician and the breeder and their respective organizations. Four candidate varieties were not formally registered. This does not mean that they failed to meet the objectives of the farmers who selected them: three of them where the highest yielders in the experiment on the land of the farmer who selected them (Table 2). Their wider use and acceptance were restricted only by socio-economical and institutional mechanisms. One reason that JM-12.7 could be registered was because the plants and the pods had a red mark that enabled it to comply with UPOV regulations. The other varieties may not have had such distinguishable characteristics and therefore may have failed to formally register if it had been tried. Also, their adaptation may have been too narrow to pass the requirements of good yields in multiple environments as described in the regulations of the variety registration, which is necessary for the varieties to be commercially viable. This implies that there may be environments where these varieties are the best performing ones, but the lack of an institutional infrastructure impedes their availability to farmers in those environments. An illustrative example from this case is the variety selected by Santos Luis Merlo that is still grown by farmers to whom he personally gave the seed. Further distribution of the variety is only supported by farmer-to-farmer seed exchange.

The experiences also show how institutional regulations are highly dependent on the people who implement them and on their political context. The formalization of the variety JM-12.7 needed a political change in the country for public-sector officials to become supportive to the registration process. The political–institutional context as an aspect that influences the behaviour of individual points at the contingent character of these processes as well. The importance of political context was also visible in the direct relation of the agenda of the donor and NGO to the creation of the space in which the group collaborated. Whereas this demonstrates that spaces for successful farmer–researcher collaboration can be created, it also emphasizes their dependent and temporal character. On the other hand, it is important to note that new forms of organization were established: the farmer-co-operative COSENUP and the regional programme as organizational forms transgressed the boundaries of the project space and are alternative organizational forms that have become partly integrated in the wider socio-political and institutional world.

8 Conclusions

This case shows that farmer and researcher knowledges are different but not necessarily conflicting, and that the boundaries between knowledges are not necessarily as tight and difficult to pass as is sometimes suggested. First of all, the farmer and scientist knowledges may be less different than suggested, and open-mindedness and a shared interest are likely to be essential for working across knowledge boundaries. In this respect it is important to understand the role of a knowledge-broker and task-organizer. The technician in this case did not only make the boundary work possible in terms of organization, but initially also translated the breeders’ knowledge and decided jointly with the farmers on the implementation in the field. In addition, he took care of all other matters that seemed not directly related but were significant for continued and effective interaction.

More than boundaries between the actors within the task group, there were boundaries between the group work and the world beyond that presented limitations. Socio-institutional and political mechanisms in combination with personal characters and coincidences were directing the destiny of the candidate varieties and final registration of one of them. Not only the social boundaries as selection mechanisms are relevant. Also the emerging of alternative organizational structures in response to these mechanisms are interesting phenomena in this case that indicate that temporary, contingent spaces potentially contribute to more structural changes in the organization of technology development.

Finally, I end this paper with a short reflection on the methodological angle of this analysis. The description of the collaborative making of bean varieties was inspired by technography as described by Richards [20] and Jansen and Vellema [21; in this issue]. This technography or the description of the process of making focuses on a group of actors and their practices and organization around a group-task. The combination of the technographic approach with boundary concepts brings out how objectives and practices varied within the task group and how these practices were informed by different realities of the social worlds of the actors. In addition, it showed how the material making of the improved variety crossed the boundaries of these worlds. After the making of the five candidate varieties, another phase followed. That phase highlighted that crossing inner group boundaries was not sufficient. For ‘institutionalization’ or acceptance in the wider social world, the varieties and the group had to cross a series of socio-political and institutional boundaries. This meant compliance to regulations and reorganization of the group. It also required a degree of contingency to accomplish the goal. In this case, the material making of the varieties and the wider acceptance of the varieties were remarkably distinct phases in the process. In contrast to the first phase of material making, I like to consider the later phase as a process of social making. This emphasizes that institutionalization or social making equally involves practices and knowledges of actors that are influenced by the social world of which they are part. The technographic description with its emphasis on practices of the people in the field of material and social making captured many essential elements that helped understanding this part of the process. This case shows that looking at institutionalization as part of the making can contribute to understanding the successes and failures of PPB and participatory technology development in general.

Acknowledgements

The author wishes to thank Todd Crane, Kees Jansen, Sietze Vellema and Gerard Verschoor for their contributions to the analysis of the case. The reviewers are acknowledged for their valuable comments on an earlier version of this paper. Rolando Herrera's and Julio Molina's unconditional sharing of information and their support to data collection by the author are specially acknowledged.

Notes

1 From my personal communication with breeders and other researchers in NARs and CGIAR institutes.

2 a http://www.upov.int/.

3 a http://www.programa-fpma.org.ni/.