About six years ago, a comprehensive horizontal resistance breeding programme in bean was started for the Mixteca region in Mexico. The programme selected bean, since it is one of Mexico's major food crops and protein sources. However, due to diseases, the crop is no longer profitable and has tended to be abandoned by farmers in Mixteca. Mexico is considered one of the centres of origin of the common bean, and thus offers a wide genetic variability in both bean and its pathogens. The project is carried out by a multi­disciplinary team from the Colegio de Postgraduados, a Mexican agricultural research and education institution, in cooperation of the University of Guelph, Canada.
The programme has followed quantitative methods of plant breeding, which are based on the idea that many genes contribute to the expression of the desired resistance. These genes are present in the population, but need to be concentrated in the improved resistant varieties. Therefore, it is required to have a broad genetic base, genetic flexibility, and a strong selection pressure by the parasites to which the beans are to develop resistance.
The programme selected seven parent lines out of 70 landraces that had been collected from the highlands in Mixteca. They were selected specifically for their susceptibility to the two main local parasites: bacterial common blight and bean common mosaic virus. Susceptibility ensures that no vertical resistance genes are functioning in these landraces. With the parent lines and the subsequent generations, all possible direct and reciprocal crosses were undertaken. A strong selection was applied by artificially infecting the offspring of these crosses with the pathogens. Out of a the total population of 6000 plants, of which all showed disease symptoms of bacterial common blight and bean common mosaic virus, only the best looking 60 out of the first breeding cycle, and 600 individuals out of the subsequent three breeding cycles were selected for further crossing. Finally, an additional selection of the fourth generation (F4) to avoid potentially misleading effects of the hybrid vigour was made, even though this increased the period of the breeding cycle. The difference with this procedure and so­called 'recurrent mass selection' is the exceptionally strong selection pressures.

The programme has produced 14 lines with increased resistance. The average yields of these elite lines are nearly three times higher than the original landraces, without the use of crop protection chemicals. Some of the lines are now being tested by 15 farmers in their own fields. These farmers receive instructions for testing, such as to grow each line on well defined plots and to weight their yield. Although it is too early for precise evaluation, the elite lines look healthier than local landraces. The elite lines have higher resistance to soil­borne diseases and recover better from the bean common mosaic virus attacks.
The farmers, who have been neglected by formal breeding institutes, are enthusiastic about the preliminary results. They are eager to participate in the selection of improved lines according to their own preferences. Aside from resistance, they value taste and cooking qualities.
Farmers are not involved in the breeding itself. One of the most critical steps in the breeding process is the hand made crosses. These need to be carried out under greenhouse conditions and require special training. Because of the distance between the Colegio de Postgraduados and the Mixteca region, all crosses have been done outside the region. The hand pollination is done by undergraduate students or young farmers from these areas. The selection process under high infection pressure has to be done in the area of future cultivation and under usual agricultural practices, except for the use of pesticides. Farmers look after additional traits such as precocity, the growth habit, the pods length, the seeds size, colour and brightness and general health condition of the genotypes.
Because of the simplicity of the breeding process for horizontal resistance, there is huge potential for establishing farmer­participatory breeding programmes. In the future, the farmers organized into breeding clubs could conduct the breeding programme themselves. The guidance of scientists could be minimized to the statistical design of the experiments to ensure that the observations are not random errors, and to make yield analysis comparable among genotypes. Also, the participation of scientists is required to ensure that all plants are infected with the designated pathogens.
Breeding for horizontal resistance is a continuous process, which is a great advantage over the breeding for vertical resistance. Every breeding cycle (crosses, multiplication, selection) provides new and slightly better cultivars in all respects. Horizontal breeding in the hands of farmers, could provide them with the new genotypes capable of withstanding locally important parasites.
Horizontal resistant varieties are able to withstand the impact of locally important parasites if selection takes place on the site and under the agronomic practices of future cultivation. This means that even in one region such as mountainous Mixteca, advanced genotypes should be tested by as many farmers as possible under diverse agro­ecological conditions. In the programme, farmers are interested in cooperating since they can keep all the improved lines that they please. Thus, the farmers' participation ensures the widespread adoption of improved lines.
Roberto García­Espinosa

Colegio de Postgraduados, Instituto de Fitosanidad, Montecillos, Texcoco, Mexico 56230, Mexico.
E­mail rogar@colpos.colpos.mx