1IRRI, PO Box 933, 1099 Manila, The Philippines; 2Jiangsu Academy of Agricultural Sciences (JAAS), Nanjing, China; 3Department of Agriculture, Bangkok, Thailand

Background and objectives
The challenge of meeting the continually increasing demand for food in third world countries, resulting from population increases and reduced arable land area, remains serious even with improved crop genotypes. In rice, with the increasing intensity of production, diseases are one of the major constraints to the stability of production. Disease management in rice has been largely dependent on host plant resistance or, in some countries, on chemical control which may not be consistent with sustainable production. It is urgent that we search for other options for disease management for tropical conditions. One promising option emerging in recent years is biological control technology. Following a standard isolation procedure, many biological control agents (BCA), mostly antagonistic bacteria, have been found to be associated with the rice ecosystems. Many of these bacteria have the ability to suppress the development of more than one diseases.

Field performance evaluation
Through the collaborative project on biological control of rice diseases, the functional relationship of BCA to sheath blight (Rhizoctonia solaniAG1) in the course of the epidemic process was assessed in terms of reduction of disease foci, lesion and focal point expansion. The effective density of BCA was critical at time of lesion initiation and expansion. When BCA was introduced to the canopy its population declined rapidly (often to an undetectable level within 24 h). Timing BCA application at initial sheath blight lesion development stage was key to the success of BCA deployment. To maintain that 'effective density' of BCA on the rice plant surface, BCA should be applied when sheath blight lesion was visible in the field. With this initial result, we applied BCA for sheath blight control in direct seeded rice in farmers' fields for the past 3 years in central Thailand. A downward trend of sheath blight foci from 77% in the first crop to 20% in the fifth crop cycle was noted, while sheath blight in the neighbouring fields remained very high. In Jiangsu, China, BCA strain 916 (Bacillus subtilis) was introduced with a high inoculum density at maximum tillering or early booting stage when sheath blight lesions were visible in the field. The disease control value ranged from 70 to 80%. Based on a 3-year experiment with BCA applied two times: one at early booting stage and the other 10 days afterward, and under moderate disease pressure, sheath blight could be suppressed to a level at which minimal crop damage occurred. However, under conditions of high disease pressure, BCA alone was not adequate to suppress disease development and spread, and a one-time fungicide application was needed. When BCA was mixed with a fungicide (at half or one-third of the usual dosage), the control value was improved. Fungicide appears to provide a 'window' of protection before BCA was active in the canopy. Strain B-916 has been applied for the past three years over thousands of hectares and initial results indicated that sheath blight was decreased in some fields at particular sites. Full-scale experiments are now in progress to assess sustainable disease management with BCA.

A system to scale-up biological control technology
For biological control technology to show impact there is a need to identify extension mechanisms to link to farmers in BCA development, formulation and production. Based on large-scale testing, sufficient B-916 inoculum was prepared to meet the demand from the farmers. The Plant Protection Institution of JAAS in collaboration with the Unit of Microbial Fermentation Plant, Nanjing Agricultural University produce the amount needed for a year. In 1997, the total volume of B-916 fermented product (in liquid form) was 10 tons whereas the demand in 1998 will be 16 tons for one site alone. We believe that this 'local production with local strains for local use' may be a model to scale up the biological control technology for rice disease management. This model offers a system in which quality control of the product is met and production responds to demand so that issues of shelf-life and storage of BCA are minimized.