ORSTOM-IRRI Project on Rice Pest Characterisation, IRRI, Los Banos, The Philippines

Background and objectives Rice sheath blight, caused by Rhizoctonia solani is a disease with specific characteristics: in a first phase, the onset of epidemics is largely dependent on the mobilization of predominantly soil-borne inoculum, while in a second phase, the disease is leaf-borne and intensification depends on environmental conditions prevailing in the canopy. In this second phase, spread of the disease is ensured by progress of fungal hyphae from diseased to healthy tissues.

Results and conclusions
Spatio-temporal analysis of epidemics in farmers' fields indicates correlation of disease incidences at very short distances, and very short time lags (not exceeding 1 ;week). Semi-vadograms of such epidemics often do not suggest any structure beyond a distance of four hills and a range that never exceeds 13 hills. Epidemics therefore appear to have a very strongly aggregated spatial structure and to evolve very rapidly over time. Rapid increase in disease incidence at epidemic onset appears to be associated with a uniform distribution of soil-borne inoculum and lack of spatial structure of initial infections. The amount of primary inoculum and terminal disease incidence do not appear to be related. Another well known characteristic of the disease is that epidemics reach higher incidences in the rainy season than in the dry season. Experiments on focal expansion of sheath blight demonstrate that contact frequency between infected and healthy tissues and leaf wetness duration are key factors determining disease spread [1]. Detailed experiments document further the effects of these factors: increased leaf-to-leaf contacts, intermittent (12/12 ;h) leaf wetness periods - but not the nitrogen content of the host tissues -favour increased efficiency of the inoculum in the second phase of epidemics [2]. Simulation modelling shows the great importance of processes taking place in the second phase, relative to those to the first phase: sensitivity analysis shows that the intrinsic rate of secondary infections has a much stronger effect on simulated epidemics than the intrinsic rate of primary infection [3]. These elements allow us to to explain why a dry season epidemic may start faster but later taper off, compared with a rainy season epidemic in a farmer's field. They may also provide avenues for developing management techniques of this disease.

1. Savary S, Castilia N, Elazegui F, McLaren CG, Ynalvez MA, Teng PS, 1995. Phytopathology 85,959-965.
2. Castilia N, Leaho RM, Elazegui FA, Teng PS, Savary S, 1996. Journal of Phytopathology 144, 187-192.
3. Savary S, Willocquet L, Teng PS, 1997. Agricultural Systems 55, 359-384.