VIEW ARTICLE

A Statistical Method of Predicting Outbreaks of Rice Panicle Blast. Tsugio Sasaki, Tohoku National Agricultural Experiment Station, Omagari, Akita-ken; Hajime Kato, National Institute of Agricultural Sciences, Nishigahara, Kita-ku, Tokyo, Japan. Phytopathology 62:1126-1132. Accepted for publication 7 April 1972. DOI: 10.1094/Phyto-62-1126.

A statistical method of forecasting panicle blast after heading has been developed. Cumulative numbers of diseased spikelets, plotted against time, formed a sigmoid curve for all cultivars grown under different conditions for the years 1962 through 1967. Based on 112 sets of readings, each linear equation has been developed, relating the logit of the percentage of diseased spikelets 12 days after the middle stage of heading (MSH) and the rate of increase during the following 6 days. With these equations, numbers of spikelets blasted in the next 6 days can be forecast by extrapolation. Predicted values and values calculated from each regression formula based on real readings had a coefficient of determination R² = .714 (P < .01). Between logits of diseased spikelets 24 days after MSH and of diseased neck nodes 30 days after MSH, the R² was .591 (P < .01). The amount of neck node blast corresponding to the amount of spikelet blast predicted can be estimated by means of this regression formula. The amount of diseased portions in panicle branches can be predicted by the same procedure.