**2. Maize participatory breeding in Portugal – VASO project**

#### **2.1. Historical prespective**

Maize domestication resulted from a single event involving its wild progenitor teosinte (*Z. mays* subspecies *parviglumis*), introgression from other teosinte types and the segregation in‐ to two germplasmpools between which much hybridisation occurred (reviewed by [5]. Por‐ tugal, by its privileged historical and geographical position as an enter point of new species into Europe, was among the first European nations to adopt maize (*Zea mays* L.) in its agri‐ cultural systems, more than five centuries ago [34]. The idea of hybridization among differ‐ ent maize introductions all over the country, rather than a slow northward dispersion accompanied by selection for earliness from one only germplasm introduction is supported in the case of Portugal (and Spain). The Iberian maize germplasm display no close relation‐ ship with any American types, but sharing alleles with both Caribbean and North American flints [35, 36].

After its introduction in the 16th century, maize spread rapidly throughout the country leading to an agricultural revolution, enhancing the rural population's standard of living. Numerous landraces (open pollinated varieties, OPV) have been developed during the centuries of cultivation, adapted to specific regional growing conditions as well as farm‐ er's needs.

However, after World War II, Portugal was one of the first European countries to test the American maize hybrids which initially were not well accepted by the Portuguese farmers due to several handicaps such as late maturity or kernel type, not fitted for food or poly‐ cropping systems. Nevertheless, several breeding stations were established within Portugal at that time, from North to South, in the cities of Braga (NUMI), Porto, Viseu, Elvas and Ta‐ vira, releasing adapted hybrid varieties based on inbreds developed from Portuguese and American germplasm. This was the case specially of NUMI and the breeding station at Porto with their famous white flint double-crosses being the preferred seed by the Portuguese farmers, during the 60's and 70's. In fact, during that period of time the commercial yellow dent hybrids from the international seed companies never reached the same level of prefer‐ ence by our farmers, who needed a type of plant with a cycle that could fit their poly-crop‐ ping system and with a type of quality kernel for human consumption and not for feed. Accordingly, an enormous decrease in the number of cultivated landraces occurred by re‐ placement with hybrid varieties. Due to this, a growing concern that numerous Portuguese maize landraces may have been lost forever started to be felt since the late 1970's. This awareness of genetic erosion led Silas Pego to initiate collection missions for maize in 1975. In the following years a more in-depth collection supported by FAO/IBPGR covered the en‐ tire country in successive missions. These materials gave rise to the first long-term cold stor‐ age facilities, the precursors of the present Portuguese Plant Germplasm Bank (BPGV), with one of the best European maize germplasm collections (ca 3000 landraces accessions) [37].

duction risks, since the farmers gain knowledge of the variety's performance as part of the

The essential advantages of PPB over CPB involve: better targeting of local environmental conditions, better definition of selection criteria important to the end-users, faster and high‐ er adoption of improved cultivars by the farmer and increase/maintenance of genetic variabil‐ ity. PPB also gives voice to farmers and elevates local knowledge to the role of science [19].

Participative approaches to agricultural research and development are now extensively used throughout the world to help define and address the practical research needs of farmers. They have proved useful in solving practical problems in complex and diverse farming sys‐ tems, characteristics typical of organic farming and low input systems. In the case of maize breeding, very effective PPB projects are reported all over the world. This is the case of the Andean region (Ecuador, Peru and Bolivia) [21], Brazil [22, 23], China [24], Ethiopia [25], Ghana [26], India [27, 28], Kenya [29], Mexico and Honduras [30], Nepal [31] and in Nigeria [32]. In Portugal, a very successful long running PPB project in maize (the VASO project, Vale do Sousa - Sousa Valley) is on going since 1984 [33]. This PPB project was developed to cover the needs of small maize farmers, with scarce land resources, in polycropping systems for human uses (bread production). This project has recently been enlarged and extended to other regions of the country and special attention is now given to quality traits such as nutri‐ tional and health beneficial quality aspects besides the already considered technological

**2. Maize participatory breeding in Portugal – VASO project**

Maize domestication resulted from a single event involving its wild progenitor teosinte (*Z. mays* subspecies *parviglumis*), introgression from other teosinte types and the segregation in‐ to two germplasmpools between which much hybridisation occurred (reviewed by [5]. Por‐ tugal, by its privileged historical and geographical position as an enter point of new species into Europe, was among the first European nations to adopt maize (*Zea mays* L.) in its agri‐ cultural systems, more than five centuries ago [34]. The idea of hybridization among differ‐ ent maize introductions all over the country, rather than a slow northward dispersion accompanied by selection for earliness from one only germplasm introduction is supported in the case of Portugal (and Spain). The Iberian maize germplasm display no close relation‐ ship with any American types, but sharing alleles with both Caribbean and North American

After its introduction in the 16th century, maize spread rapidly throughout the country leading to an agricultural revolution, enhancing the rural population's standard of living. Numerous landraces (open pollinated varieties, OPV) have been developed during the centuries of cultivation, adapted to specific regional growing conditions as well as farm‐

selection process [19].

258 Plant Breeding from Laboratories to Fields

ability for bread production.

**2.1. Historical prespective**

flints [35, 36].

er's needs.

After Portugal and Spain entered the European Community, in 1986, a new political reality took place with a consequent change in our agriculture policy, smashing down the tradition‐ al small farming characterized by a poly-cropping, quality oriented and sustainable agricul‐ ture. In two decades these small farmers were pushed to bankruptcy. Also later on, during the 1980's, the scientific community became aware of the importance of the genetic resour‐ ces co-evolution and the need for *in-situ* / on-farm conservation. The main question now was, how to restore this sustainable and quality oriented agricultural system in Portugal, and bring it back to business? Why not to apply some science to those genetic resources that had been selected by our traditional farmers during the last four centuries?

To provide an incentive for *in-situ* conservation of traditional maize landraces Silas Pego had the idea of engaging local farmers and their seeds in a participatory maize breeding program. By doing this, his goals were not only to conserve, but also to improve the social well-being of this rural community by increasing farmers' income through rising yields from some of their own seeds. To bring that idea to practice he led, in 1984, a detailed sur‐ vey on farmer's maize fields at «Vale do Sousa» Region (Sousa Valley Region) in the North‐ west of Portugal. The collected materials were the starting point of a PPB project, with simultaneous on-farm breeding and on-farm conservation objectives (VASO- ''Vale do Sou‐ sa''- project). This project aimed to answer the needs of small farmers (e.g., yield, bread making quality, ability for polycropping systems) with scarce land availability due to a high demographic density, where the American agriculture model did not fit and the multina‐ tionals had no adequate market to operate. From our previous knowledge of the small farm‐ ing reality of the mountainous North of our country we knew that a project oriented to one crop integrated within a system should be oriented under a general developmental frame‐ work, understood and accepted by the farmer. This would require a different philosophic approach that we coined as *integrant philosophy* in opposition to the *productivist philosophy*. While in this last model the plant breeder occupies the center of decision, in the *integrant philosophic approach* is the farmer who occupies that position. Furthermore, we needed to link our philosophy with the basic formula of any production system: energy + raw material+ science = final product. This commanded our decisions, preference for renewable energy in‐ stead of fossil, local genetic resources instead of exotic germplasm, and an adequate breed‐ ing methodology accessible to the farmer's understanding and participation. Besides, a set of parallel decisions came along, as:


If we could get at least some of these achievements, we needed to be sure that they had to be reached at the farmer's speed and under his own constraints. This model also embraced a quality oriented perspective (human food) where the quality factor must be financially va‐ lorized in order to sustain the agricultural system.

To implement this project several main decisions had to be made, such as the choice of the location to represent the region, the farmer to work with side-by-side, the germplasm source, and the breeding and management methodologies to apply.

#### **2.2. General procedures**

In this section, and taking the VASO project as a model, a generalised description of the most important decisions to be made on a maize PPB project implementation is presented.

As in any collaborative approach all the plans and decisions have to be made between all the involved partners and with that purpose breeders need to meet with local farmers and discuss selection strategies to understand farmer's objectives and constraints. Following this, four main decisions must be taken to implement correctly the participatory breeding approach:

#### *2.2.1. Farmers, breeders and environmental choices*

The selected farmer's fields must be located in a traditional production area where the crop to be breed is important and where support from local authorities/farmers associations is guaranteed. Farms should represent the different soil and climatic conditions under which the crop is grown, the different size of farms and farm types.

In the case of the VASO project, the Sousa Valley was chosen after the results of the 1984 survey and taking in account several factors:

**a.** At that time only 15% of the Portuguese farmers were using hybrid seeds. However, in the fertile Sousa Valley this percent was higher - 25%. This meant that even in this re‐ gion 75% of the farmers were still using their own regional varieties of maize. This cre‐ ated a perfect situation for developing alternative production systems.

crop integrated within a system should be oriented under a general developmental frame‐ work, understood and accepted by the farmer. This would require a different philosophic approach that we coined as *integrant philosophy* in opposition to the *productivist philosophy*. While in this last model the plant breeder occupies the center of decision, in the *integrant philosophic approach* is the farmer who occupies that position. Furthermore, we needed to link our philosophy with the basic formula of any production system: energy + raw material+ science = final product. This commanded our decisions, preference for renewable energy in‐ stead of fossil, local genetic resources instead of exotic germplasm, and an adequate breed‐ ing methodology accessible to the farmer's understanding and participation. Besides, a set

If we could get at least some of these achievements, we needed to be sure that they had to be reached at the farmer's speed and under his own constraints. This model also embraced a quality oriented perspective (human food) where the quality factor must be financially va‐

To implement this project several main decisions had to be made, such as the choice of the location to represent the region, the farmer to work with side-by-side, the germplasm

In this section, and taking the VASO project as a model, a generalised description of the most important decisions to be made on a maize PPB project implementation is presented.

As in any collaborative approach all the plans and decisions have to be made between all the involved partners and with that purpose breeders need to meet with local farmers and discuss selection strategies to understand farmer's objectives and constraints. Following this, four main decisions must be taken to implement correctly the participatory breeding approach:

The selected farmer's fields must be located in a traditional production area where the crop to be breed is important and where support from local authorities/farmers associations is guaranteed. Farms should represent the different soil and climatic conditions under which

In the case of the VASO project, the Sousa Valley was chosen after the results of the 1984

of parallel decisions came along, as:

260 Plant Breeding from Laboratories to Fields

**1.** start with the farmer's genetic resources,

**2.** move to the farmer's place and develop the project at his own plots,

**3.** work side-by-side and let the farmer be the decision maker,

**5.** always be ready to share knowledge and enthusiasm.

lorized in order to sustain the agricultural system.

*2.2.1. Farmers, breeders and environmental choices*

survey and taking in account several factors:

the crop is grown, the different size of farms and farm types.

**2.2. General procedures**

**4.** respect the agricultural system and the farmer's preferences and

source, and the breeding and management methodologies to apply.


Choosing the right person to work with is also a major decision on participatory ap‐ proaches, where the work is carried out side-by-side and the power of decision shared. A high farmer initial acceptance of this type of approach and enthusiasm for joining this kind of projects are the best guarantee of success. Nevertheless, a careful respect of the breeder for the local traditional agriculture is also crucial.

The farmer' selection in the VASO project was made having into account both a previous information obtained in the mentioned sociological survey, and from direct contacts within the farmers' organization CGAVS. From this collected information, the most willing and al‐ so contradictory people were chosen. In its beginning only 3 farmers were directly involved.

In the process of choosing the right person, it is very important, as highlighted by [18], to clarify


These farmers will need to be trained/updated with practical examples of how selection could be improved. The plant breeder will have to take into account the selection objectives of the farmer and the farmer will learn simple population genetics methodologies that will help him to progress more rapidly and efficiently in their seed selection.

In a PPB program it is very important to maintain contacts with farmers beyond and besides specific scientific activities. These 'courtesy' visits are not only instrumental in building and maintain good human relationships between scientists and farmers by bridging gaps, but are an incredibly valuable reciprocal source of information [18].

In the VASO project the initial enthusiasm of some of the contacted farmers that still today collaborate with the project was fundamental. Also the support of a local elite farmers' asso‐ ciation (CGAVS) which agreed to be part of the project was very beneficial to the success of the project.

#### *2.2.2. Starting germplasm and variability generation*

According to the VASO philosophy, the project should start by using local landraces as its genetic resources, selected by the breeder as the most representative of the local farming. A survey was made by the breeder in 1984 during the summer time along the maize fields of that region in a close look for particular plant phenotypes and ear size. Further, at harvest time, several sets of store houses ("*sequeiros*") were visited and farmers contacted. From this first survey two regional varieties were selected: "Pigarro", a white flint type FAO 300 cycle with strong fasciation expression, used in the best soils for human consumption, and "Amarelo miúdo" ("Amiúdo"), a yellow flint type FAO 200, adapted to the poorest soils with low ph, water stress and aluminum toxicity, but also with quality for bread production. Afterward, the VASO project was also conserving additional landraces such as "Basto", "Al‐ jezur", "Aljezudo", "Castro Verde", "Verdial de Aperrela" and "Verdial de Cete". In paral‐ lel with the landraces approach, a synthetic population, Fandango, was also included [38].

So, as highlighted from the VASO project, one of the prerequisites for the implementation of a PPB project is the existence of local adapted germplasm. In this way the farmer's selection pursued over several centuries (quality preferences) will be respected and the environmen‐ tal adaptation already achieved either for the soil or climate will be assured [33].

The inclusion of high quality parents is of particular importance when considering the quali‐ ty objectives of the population. Quality is difficult for farmers to access if they grow a crop for the commercial market, and is not necessarily improved by natural selection [39].

When needed, genetic diversity can be generated by crossing genetically diverse and adapt‐ ed local cultivars as starting genetic base. Other foreign materials (exotic germplasm) can al‐ so be added after this point to overcome any limitation typical from the local cultivars (as disease susceptibility).

#### *2.2.3. Breeding methodologies*

Technically the process is similar to conventional breeding, with three main differences. Tri‐ als will be grown in farmers' fields rather than on-station, covering a range of target envi‐ ronments and using farmer's agronomic practices and levels of inputs. Selection will be conducted jointly by breeders and farmers (and other end-users when appropriate), so that farmers participate in all key decisions. This process can be independently implemented at a large number of locations [19].

of the farmer and the farmer will learn simple population genetics methodologies that will

In a PPB program it is very important to maintain contacts with farmers beyond and besides specific scientific activities. These 'courtesy' visits are not only instrumental in building and maintain good human relationships between scientists and farmers by bridging gaps, but

In the VASO project the initial enthusiasm of some of the contacted farmers that still today collaborate with the project was fundamental. Also the support of a local elite farmers' asso‐ ciation (CGAVS) which agreed to be part of the project was very beneficial to the success of

According to the VASO philosophy, the project should start by using local landraces as its genetic resources, selected by the breeder as the most representative of the local farming. A survey was made by the breeder in 1984 during the summer time along the maize fields of that region in a close look for particular plant phenotypes and ear size. Further, at harvest time, several sets of store houses ("*sequeiros*") were visited and farmers contacted. From this first survey two regional varieties were selected: "Pigarro", a white flint type FAO 300 cycle with strong fasciation expression, used in the best soils for human consumption, and "Amarelo miúdo" ("Amiúdo"), a yellow flint type FAO 200, adapted to the poorest soils with low ph, water stress and aluminum toxicity, but also with quality for bread production. Afterward, the VASO project was also conserving additional landraces such as "Basto", "Al‐ jezur", "Aljezudo", "Castro Verde", "Verdial de Aperrela" and "Verdial de Cete". In paral‐ lel with the landraces approach, a synthetic population, Fandango, was also included [38].

So, as highlighted from the VASO project, one of the prerequisites for the implementation of a PPB project is the existence of local adapted germplasm. In this way the farmer's selection pursued over several centuries (quality preferences) will be respected and the environmen‐

The inclusion of high quality parents is of particular importance when considering the quali‐ ty objectives of the population. Quality is difficult for farmers to access if they grow a crop

When needed, genetic diversity can be generated by crossing genetically diverse and adapt‐ ed local cultivars as starting genetic base. Other foreign materials (exotic germplasm) can al‐ so be added after this point to overcome any limitation typical from the local cultivars (as

Technically the process is similar to conventional breeding, with three main differences. Tri‐ als will be grown in farmers' fields rather than on-station, covering a range of target envi‐ ronments and using farmer's agronomic practices and levels of inputs. Selection will be

tal adaptation already achieved either for the soil or climate will be assured [33].

for the commercial market, and is not necessarily improved by natural selection [39].

help him to progress more rapidly and efficiently in their seed selection.

are an incredibly valuable reciprocal source of information [18].

*2.2.2. Starting germplasm and variability generation*

the project.

262 Plant Breeding from Laboratories to Fields

disease susceptibility).

*2.2.3. Breeding methodologies*

The goal of any plant breeding is to develop a plant population composed of phenotypes that meet farmers' criteria, and that farmers will, therefore, adopt and maintain. So selection criteria must consider farmers objectives like quality traits, such as flavour or nutritional value, pest and disease resistance and enhanced capacity to survive in highly changeable en‐ vironment typical of low input/organic farming systems (yield stability) through genetic di‐ versity maintenance or enhancement.

The choice of the breeding methods cannot be made without considering whether and how farmers are handling genetic diversity. An issue related to the choice of the breeding meth‐ od is how much breeding material farmers can handle. The choice of the breeding method also depends on the desired genetic structure of the final product, i.e. pure lines, mixtures, hybrids or open pollinated varieties [18].

Common mass selection, usually performed by all farmers, consists on the identification of superior individuals in the form of plants from a population, and in the case of crops like maize, the bulking of seed to form the seed stock for the next generation [40]. It requires rel‐ atively little effort compared with other selection methods and, practiced season after season with the same seed stock, has the potential to maintain or even improve a crop population, depending upon the heritabilities of the selected traits, GxE for the trait, the proportion of the population selected, and gene flow in the form of pollen or seeds into the population [15]. PPB is a cyclic process where the best selections will be used in further cycles of recom‐ bination and selection or selection will just give rise to experimental cultivars, to be tested again on the farmers' fields.

Within the VASO project a tacit agreement was made between the breeder and the farmer involved. While the breeder would apply his breeding methodologies, the farmer would continue a parallel program with their own mass selection criteria, starting with the same initial populations. With this agreement the breeder had to accept low-input and intercrop‐ ping characteristics, as well as to accept and respect the local farmer as the decision maker. On the other hand the farmer was able to compare the effectiveness of the two breeding sys‐ tems allowing him to base his decisions on solid grounds. Due to the choice of locally adapt‐ ed germplasm, diversity and quality were considered as selection priorities.

In relation to the selection approach followed by the VASO project the breeder initially opt‐ ed by the *S2 lines recurrent selection*, due to its potential for favoring a good amount of addi‐ tive gene action. This methodology worked very promisingly with the "Pigarro" germplasm, but not so good with "Amiúdo". In fact, while we could observe a surprisingly low inbreeding depression when we tested the S2 lines of "Pigarro", a different situation came out from the S2 lines of "Amiúdo". Here, the inbreeding depression was so high that we had to move to the *S1 lines recurrent selection methodology*. On the other hand, farmers were advised to use improved mass selection approaches, such as with a two parental con‐ trol (*stratified mass selection*), where selection takes place not only after harvesting, at the stor‐ age facilities (ear traits), but also during crop development in the field such as in a cross pollinating species before pollen shedding, for the male parent selection by detasselling all undesirable plants, and before harvesting, for the female parent selection (ear size, root and stalk quality and indirectly pest tolerance; [33]). As the time was passing by, the Common Agriculture Policy (CAP) from the EU, which favored the big farming oriented for feed pro‐ duction, pressed the small farming to bankruptcy. As a consequence, our VASO project also suffered a sudden lack of governmental support leaving the breeder without any support to adequately pursue the mentioned methodologies, both requiring a big amount of hand pol‐ linations. Again, the breeding methodology had to be changed, now for *stratified mass selec‐ tion* to all the other landraces in the project.

The improvement program included yield, lodging performance, pest and disease tolerance, and indirectly, adaptation to climate changes.

During the VASO project, some pre-breeding methodologies were developed such as the HUNTERS (High, Uniformity, aNgle, Tassel, Ear, Root lodging and Stalk lodging) or the Overlapping Index [7, 38]. These are nowadays very useful on our PPB maize landrace se‐ lection on-farm.

In all cases in the VASO project, the traditional poly-cropping technology was followed: sur‐ rounded by vineyards, the plots were usually planted with three main crops: maize *(Zea mays L.),* beans *(Phaseolus vulgaris L.)* and a forage crop *(Lolium perene L.)*. The first two si‐ multaneously planted along the same row in May and the forage later on in July between the rows. This organic system has three main sustainable advantages:


#### *2.2.4. Seed management and dissemination*

Since the beginning of the VASO Project, phenotypic data were collected and seed of each selection cycle, either from phenotypic recurrent selection or from S2 recurrent selection, was kept at 4ºC in our national Plant Germplasm Bank (BPGV) cold storage facilities [38].

Several maize OPVs were selected within this project with the joint collaboration of the breeder Silas Pego and the farmers.

The seed however has not yet been nationally distributed /commercialized due to the lack of appropriate legislation allowing the certification of OPV with a certain level of heterogenei‐ ty. These aspects will be further developed under the PPB success evaluation and the seed dissemination and ownership sections.

#### **2.3. PPB Vaso project success evaluation**

During more than a quarter of a century of continuous participatory on-farm maize breed‐ ing, some *ups* and *downs* came along, mainly influenced by political fluctuations that affect‐ ed both the governmental support and the market prices of its quality oriented output. In 2001, the mayor of the town of Lousada presented the situation of the VASO project govern‐ mental funds cutting to his general assembly that unanimously decided to substitute our Ministry of Agriculture institution (DRAEDM) as the sponsors of this long term PPB project.

In spite of these *ups* and *downs,* our main achievements can be centered in three areas:

**1.** breeding output,

age facilities (ear traits), but also during crop development in the field such as in a cross pollinating species before pollen shedding, for the male parent selection by detasselling all undesirable plants, and before harvesting, for the female parent selection (ear size, root and stalk quality and indirectly pest tolerance; [33]). As the time was passing by, the Common Agriculture Policy (CAP) from the EU, which favored the big farming oriented for feed pro‐ duction, pressed the small farming to bankruptcy. As a consequence, our VASO project also suffered a sudden lack of governmental support leaving the breeder without any support to adequately pursue the mentioned methodologies, both requiring a big amount of hand pol‐ linations. Again, the breeding methodology had to be changed, now for *stratified mass selec‐*

The improvement program included yield, lodging performance, pest and disease tolerance,

During the VASO project, some pre-breeding methodologies were developed such as the HUNTERS (High, Uniformity, aNgle, Tassel, Ear, Root lodging and Stalk lodging) or the Overlapping Index [7, 38]. These are nowadays very useful on our PPB maize landrace se‐

In all cases in the VASO project, the traditional poly-cropping technology was followed: sur‐ rounded by vineyards, the plots were usually planted with three main crops: maize *(Zea mays L.),* beans *(Phaseolus vulgaris L.)* and a forage crop *(Lolium perene L.)*. The first two si‐ multaneously planted along the same row in May and the forage later on in July between

**1.** a probable symbiosis between a *Leguminosae* (beans) and *Rhyzobium sp*, as a natural

**2.** any possible plant damage or failure along the row could be compensated in favor of

**3.** after harvesting the maize crop in September the soil was already covered with a forage layer that not only functions as a protection against erosion but also as a source of 3-4

Since the beginning of the VASO Project, phenotypic data were collected and seed of each selection cycle, either from phenotypic recurrent selection or from S2 recurrent selection, was kept at 4ºC in our national Plant Germplasm Bank (BPGV) cold storage facilities [38].

Several maize OPVs were selected within this project with the joint collaboration of the

The seed however has not yet been nationally distributed /commercialized due to the lack of appropriate legislation allowing the certification of OPV with a certain level of heterogenei‐ ty. These aspects will be further developed under the PPB success evaluation and the seed

the rows. This organic system has three main sustainable advantages:

the other crop to which was allowed more sunlight and

cuts of forage for animal feeding during the winter.

source of nitrogen for the *graminea* (maize),

*2.2.4. Seed management and dissemination*

breeder Silas Pego and the farmers.

dissemination and ownership sections.

*tion* to all the other landraces in the project.

264 Plant Breeding from Laboratories to Fields

and indirectly, adaptation to climate changes.

lection on-farm.


#### *2.3.1. Breeding output*

All the outputs from this project are improved OPVs. "Pigarro" is the one which received our main investment in breeding, with several cycles of *S2 lines recurrent selection* and repeat‐ ed cycles of *stratified mass selection*. Its yield level is now between 8 and 9 t/ha, and is the type of seed that better fits the high quality standards of our most famous maize bread – *Broa de Avintes*. "Amiúdo" went to some cycles of *S1 lines recurrent selection* and fits the conditions of tolerance to soil (low ph and Al) and water stress. As a yellow vitreous flint and small kernel size, this quality enables this variety to be used both for maize bread as well as for the specific market of feed for carrier-pigeons and pigeon breeding. "Fandan‐ go", an ear size champion (at the Sousa Valley best ear contest), had its original FAO700 cycle reduced to FAO600 and a competing yield over 10 t/ ha. A set of other improved varieties like "Aljezur" FAO 400 yellow flint, "Aljezudo" FAO 300 yellow flint, "Castro Verde" FAO 600 yellow flint, and some others, complete a minor output that can be im‐ proved in future attempts.

Yield trials for evaluation at the farmer's place, were the most difficult task to carry out in a way to fit the levels of statistical significance. Nevertheless for "Fandango" and "Pigarro" OPVs these trials were already established, and for some of the other VASO OPVs are now under way. In the already established evaluation trails we should take into account the orig‐ inal specific objectives defined for the farmers' and breeders' selections. For the "Pigarro" *stratified mass selection* the farmer aimed at obtaining bigger size ears while maintaining the flint and white type of kernel. The breeder with its alternative *S2 lines recurrent selection* method aimed at increasing favorable alleles both for yield, ear placement and stalk quality. For the "Fandango" *stratified mass selection* the farmer was looking for big ear size maximiza‐ tion and the breeder for yield maximization. Evaluation field trials for selection gain on these two OPVs were established in several locations in Portugal and in the case of "Fandan‐ go", also in the USA [38, 41]. The statistical analysis indicated that *stratified mass selection* in "Pigarro" lead to an increase in days to silk and anthesis, ear diameter, kernel row number and fasciation. On the contrary, ear length decreased significantly [38]. Molecular SSR mark‐ er data, from [42], on three selection cycles (C0-1984, C9-1993 and C20-2004), revealed that no effective loss of genetic diversity has occurred during the selective adaptation to the farmer's needs and the regional growing conditions. Variation among selection cycles repre‐ sented only 7% of the total molecular variation, indicating that a great proportion of the ge‐ netic diversity is maintained in each selection cycle. Genetic diversity has not been reduced from the "Pigarro" bred before 1984 to those examples improved after 2004, but the genetic diversity maintained is not exactly the same. Mass selection seems to be an effective way to conserve diversity on-farm, and interesting phenotypic improvements were achieved as the bigger ears farmers' objective [42].

On the other hand, the response to the "Pigarro" *recurrent selection by S2 lines* indicated that after three cycles of selection, days to silk, uniformity and the cob/ear ratio increased signifi‐ cantly, but also without a significant yield increase [38].

The "Fandango" *stratified farmers mass selection* evaluation performed in Portugal revealed that the ear length and the thousand kernel weight decreased significantly and simultane‐ ously plant and ear height increased significantly [41], accomplishing the bigger ears farm‐ ers' objective. These traits had no significant changes during the selection cycles performed by the breeder. Additionally, days to silk had a significant increase during selection. For yield no significant changes were observed during selection when all the evaluation loca‐ tions were considered. Nevertheless, when considering only the trials performed at the loca‐ tion where the PPB took place (Lousada), a significant yield increase was recorded especially during breeders selection cycles (3.09% gain per cycle per year), being less pro‐ nounce during farmers selection cycles (only 0.63% gain per cycle per year). The lack of sig‐ nificant progress in yield for both "Pigarro" and "Fandango" can be explained by low selection intensity due to the exclusion of stalk lodged plants in the basic units of selection [38, 41] what must be taken into consideration in future selections for yield increase.

Both selection methods used in "Pigarro" and "Fandango" different phases of selection, sug‐ gest that *stratified mass selection* is better than *S2 lines recurrent selection* due to the following reasons: *Stratified mass selection* is a cheaper methodology, technically more accessible to farmers, with one cycle of selection completed each summer, without reducing the con‐ served genetic diversity [41].

The evidence of genetic diversity maintenance by this PPB project with simultaneous pheno‐ typic improvement, fundaments the preservation of these on-farm selection programs where threatened landraces of great interest for future use in breeding programs and for de‐ veloping new farming systems are preserved.

#### *2.3.2. Tecnological improvement*

A modified planting system was developed in such a way that two rows of beans (20 cm apart) were planted between two consecutive rows of maize (130 cm apart). The difference to the original technology was that one row of maize was eliminated, but the seed density was maintained by doubling the plant density along each row. On one hand, this higher plant density was compensated by the larger space between rows that the plants could take advantage from. On the other hand, the now separated rows of beans, due to its fast grow‐ ing, rapidly covered the soil avoiding weeds. This new technology, besides being weed con‐ trol efficient, facilitates harvest of beans in July and the following planting of the forage crop over that empty space between rows of maize. Additionally, it had a final positive effect in September at the harvest of maize, because the soil was already protected against the ero‐ sion from the rainy season that was coming in. Moreover, during fall and winter, 3 or 4 cuts of forage could be made for animal feeding.

Another technological improvement was discovered by the farmer. In a certain morning of June, when the maize was in its four to five leaves stage, the farmer was observing an in‐ tense flight of birds over the maize field. With a closer look he noticed that birds were catch‐ ing the larvae of the pest *Agrotis segetum* L. He then realized that the soil was still humid from his irrigation the afternoon before. Our conclusion became obvious: the larvae were su‐ perficial because the soil was still cold, but go underground when the sun heats. We realized that a sustainable insect control tool was available and a light superficial irrigation became usual in each afternoon.

#### *2.3.3.Seed difusion*

er data, from [42], on three selection cycles (C0-1984, C9-1993 and C20-2004), revealed that no effective loss of genetic diversity has occurred during the selective adaptation to the farmer's needs and the regional growing conditions. Variation among selection cycles repre‐ sented only 7% of the total molecular variation, indicating that a great proportion of the ge‐ netic diversity is maintained in each selection cycle. Genetic diversity has not been reduced from the "Pigarro" bred before 1984 to those examples improved after 2004, but the genetic diversity maintained is not exactly the same. Mass selection seems to be an effective way to conserve diversity on-farm, and interesting phenotypic improvements were achieved as the

On the other hand, the response to the "Pigarro" *recurrent selection by S2 lines* indicated that after three cycles of selection, days to silk, uniformity and the cob/ear ratio increased signifi‐

The "Fandango" *stratified farmers mass selection* evaluation performed in Portugal revealed that the ear length and the thousand kernel weight decreased significantly and simultane‐ ously plant and ear height increased significantly [41], accomplishing the bigger ears farm‐ ers' objective. These traits had no significant changes during the selection cycles performed by the breeder. Additionally, days to silk had a significant increase during selection. For yield no significant changes were observed during selection when all the evaluation loca‐ tions were considered. Nevertheless, when considering only the trials performed at the loca‐ tion where the PPB took place (Lousada), a significant yield increase was recorded especially during breeders selection cycles (3.09% gain per cycle per year), being less pro‐ nounce during farmers selection cycles (only 0.63% gain per cycle per year). The lack of sig‐ nificant progress in yield for both "Pigarro" and "Fandango" can be explained by low selection intensity due to the exclusion of stalk lodged plants in the basic units of selection

[38, 41] what must be taken into consideration in future selections for yield increase.

Both selection methods used in "Pigarro" and "Fandango" different phases of selection, sug‐ gest that *stratified mass selection* is better than *S2 lines recurrent selection* due to the following reasons: *Stratified mass selection* is a cheaper methodology, technically more accessible to farmers, with one cycle of selection completed each summer, without reducing the con‐

The evidence of genetic diversity maintenance by this PPB project with simultaneous pheno‐ typic improvement, fundaments the preservation of these on-farm selection programs where threatened landraces of great interest for future use in breeding programs and for de‐

A modified planting system was developed in such a way that two rows of beans (20 cm apart) were planted between two consecutive rows of maize (130 cm apart). The difference to the original technology was that one row of maize was eliminated, but the seed density was maintained by doubling the plant density along each row. On one hand, this higher plant density was compensated by the larger space between rows that the plants could take

bigger ears farmers' objective [42].

266 Plant Breeding from Laboratories to Fields

served genetic diversity [41].

*2.3.2. Tecnological improvement*

veloping new farming systems are preserved.

cantly, but also without a significant yield increase [38].

The diffusion of the improved seeds from this VASO project, while limited to the Sousa Val‐ ley area, has been very easy, as expected. Farmers are always ready to share seeds and it happens frequently. However, a private initiative from a local farmers' cooperative (Cooper‐ ativa Agricola de Paredes), in the early 90's, with a regional contest for maize - *The maize best ear of the Sousa Valley* turned to be the best diffuser of our seeds since they became quickly champions of ear size. Among our improved varieties the long cycle "Fandango" became a real champion beating, year after year, the best commercial hybrids in ear size. One of our collaborative farmers, *Francisco Meireles,* became the most rewarded Portuguese farmer, with more than 50 trophies. This has contributed to the recognition of the farmer by the commun‐ ity, but also attracted new farmers and new germplasm to this PPB project that in this way could be identified and preserved on-farm by the same approach [38].
