VIEW ARTICLE

Equations for Predicting Wheat Stem Rust Development. M. G. Eversmeyer, Research Plant Pathologist, Plant Science Research Division, ARS, USDA, Manhattan, Kansas 66506; J. R. Burleigh(2), and A. P. Roelfs(3). (2)Former Research Plant Pathologist, Plant Science Research Division, ARS, USDA; (3)Research Plant Pathologist, Plant Protection Programs, Animal and Plant Health Inspection Service, USDA, St. Paul, Minnesota 55101, (2)Present address: Division of Agriculture, Chico State College, Chico, California. Phytopathology 63:348-351. Accepted for publication 20 October 1972. DOI: 10.1094/Phyto-63-348.

Stepwise multiple regression techniques were used to identify those meteorological and biological variables useful in explaining variation in stem rust (Puccinia graminis f. sp. tritici) severities 7, 14, 21, and 30 days after severity estimates were made. Variables which were most significant in the successful prediction of stem rust development were: disease severity estimates; weekly and cumulative number of urediospores deposited per cm²; cultivar; wheat growth stage; maximum temperature; minimum temperature; a fungal-temperature growth function; and a fungal infection function. Coefficients of determination (R²) for the most efficient combinations of these variables were .745, .664, .509, and .362 for the 7-, 14-, 21-, and 30-day forecast periods, respectively. Stem rust development was more accurately explained by use of disease severity estimates as the inoculum variable than by the use of urediospore numbers. We were unable to accurately predict stem rust development using only meteorological variables. Equations were in the form Y_i = K_i + b₁X_1i + ... +b_nX_ni.