VIEW ARTICLE

Ecology and Epidemiology

Quantitative Relationships Between Yield and Foliar Diseases of Alfalfa. S. C. Broscious, Former graduate research assistant, Department of Plant Pathology, University of Illinois at Urbana-Champaign, Urbana 61801; J. K. Pataky, and H. W. Kirby. Assistant professors, respectively, Department of Plant Pathology, University of Illinois at Urbana-Champaign, Urbana 61801. Phytopathology 77:887-892. Accepted for publication 20 November 1986. Copyright 1987 The American Phytopathological Society. DOI: 10.1094/Phyto-77-887.

Regression models relating various measures of disease level to yield were developed for the complex of fungi that cause foliar disease on alfalfa (Medicago sativa). Different levels of leaf spot epidemics were established in each of 21 trials using natural inoculum, fungicide applications (mancozeb alone or with benomyl), and inoculations with Phoma medicaginis and/or Stemphylium botryosum. Models to predict yield were fit by ordinary-least-squares regression for each trial using weekly disease severity assessments, defoliation index, or area under disease progress curve (AUDPC) as the independent variable. More than one model adequately described the relationship between disease and yield in each trial. Disease severity 1 wk before harvest (S₁), severity on the day of harvest (S₀), and AUDPC were the variables most consistently related to yield in individual trials. Linear models using each variable to predict the percentage of the maximum yield in the trial explained 52–60% of the variation in the yield data when pooled over 15 of the trials. Models estimated a yield reduction of 2.40, 1.83, and 0.14% for each unit of S₁, S₀, and AUDPC, respectively. The accuracy and precision of the models were validated using data other than those from which the models were developed. Model predictions were relatively accurate within the range of disease levels commonly observed in alfalfa fields. However, at higher levels of disease, all three models underestimated the effects of foliar disease on yield when compared with data from the six additional trials. Mean differences between the model prediction and observed values were +5.0, +2.3, and +2.7% of maximum yield for the S₁, S₀, and AUDPC models, respectively.