Disease Progress Curves in the Rice Blast Pathosystem Compared with the Logistic and Gompertz Models

Comparison of the linear coefficients from the regression analysis of 307 blast (Pyricularia grisea) disease progress curves of 42 rice genotypes, tested over a period of nine seasons, were made by fitting into the logistic and Gompertz models. The regression parameters Yintercept (a), regression coefficient (b), its standard error (SEb), coefficient of determination (R) and the time required for the disease to reach specific level of severity (T50) were estimated. The estimates of the linear coefficients following Gompertz model, revealed lower values of b, SEb and negative 'a'; higher values of T50 as well as R compared with those by logistic model. In general, the logistic apparent infection rates were higher than the Gompertz rates, the differences being wider for the slow-blasting genotypes compared with the fastblasting genotypes. The lower Gompertz infection rates accompanied with low negative ‘a' values, resulted in higher T50 values, compared with the respective lower estimates through logistic model. Both b and T50 values estimated by Gompertz model were more consistent among the slow-blasting genotypes over seasons. Out of the 307 disease progress curves, 91.2 % fitted well in to the logistic model and 90.8% in to the Gompertz model, thus suggesting that both the models fit well in to the rice blast pathosystem. The estimates of T50 were proportional to the corresponding apparent infection rates in both the models. The benefit from sanitation in terms of the epidemic to reach 0.5 level, was estimated to be 21 and 47 days following logistic and Gompertz models, respectively.