1Institute for Plant Sciences, Swiss Federal Institute of Technology, CH-8092 Zurich, Switzerland; 2Department of Plant Pathology, 2082 Cordley Hall, Oregon State University, Corvallis, OR 97331-2902, USA; 3Wakelyns Agroforestry, Fressingfield, Suffolk 1P19 5SD, UK

Background and objectives
Major problems in large-scale monoculture are high potential for disease and selection of pathogen races able to overcome host resistances. One solution is to increase host diversity both within and among fields and at the regional level. Management of resistance at the field level is largely the responsibility of the farmer and farmer groups; at the regional level, it depends largely on government organizations and private sector companies. A major potential advantage from exploitation of useful biodiversity is reduced pesticide use.

Monoculture can be defined at the levels of species, variety and resistance genotype [1]. Here we equate an increase in biodiversity with polyculture. This is usually considered to involve intercropping (mixed, row, strip, plot or relay intercropping), but can involve sole cropping provided there is rotation in space and time or sequential cropping of more than one crop per season in the same space. Different levels of diversity should be used simultaneously; for example, intercropping should also be part of a rotation. Diversification schemes and recommendations have been developed at different levels, from simple species diversification, to diversification among resistance genotypes based on pathogen phenotype distribution.

Results and conclusions
Large-scale diversification schemes are difficult to implement because farmer priorities are to grow for yield and/or quality. Farm-scale diversification can be much more successful, particularly where there is an economic advantage. Positive experiences have resulted where intercropping systems have been practised and monitored. However, continued over-use at one level only can be detrimental, for example with beans in East Africa, where diversity is common at the variety level but not at the species level, leading to problems with soilborne diseases.

Where one variety mixture is common, there could be selection of complex pathogen phenotypes able to grow simultaneously on several or all host components. However, theoretical considerations and practical experience indicate that such selection is likely to be only slow. This should allow time for introduction of new mixtures without the 'boom and bust' associated with monoculture.

Intimate intercropping also raises a question of product quality - but there are examples that show how this potential problem can be turned to advantage, for example with coffee in Colombia, barley in the former GDR, and wheat in the USA. Other advantages for intercropping include (i) yield stability; (ii) reduction in pesticide use; (iii) weed control; and (iv) at least in some cases, pest control. To maximize the advantages, there is a need for breeding programmes that involve selection for performance in heterogeneous crops. There is also a need for more promotion of intercropping systems.

1. Finckh MR, Wolfe MS, 1998. In Jones DG, ed, The Epidemiology of Plant Diseases. Chapman and Hall, London.