Link to home

Number of Experiments Needed to Determine Wheat Disease Phenotypes for Four Wheat Diseases

January 2007 , Volume 91 , Number  1
Pages  103 - 108

William W. Bockus , Department of Plant Pathology, Throckmorton Hall, Kansas State University, Manhattan 66506 ; Zhaohui Su , Statistical and Data Analysis Center, Harvard School of Public Health, Boston, MA 02115 ; Karen A. Garrett , Bikram S. Gill , and James P. Stack , Department of Plant Pathology , and Allan K. Fritz and Kraig L. Roozeboom , Department of Agronomy, Kansas State University ; and T. Joe Martin , Kansas State University Agricultural Research Center-Hays, Hays 67601



Go to article:
Accepted for publication 08 September 2006.
ABSTRACT

Disease phenotypes for winter wheat cultivars were determined in numerous inoculated greenhouse and field experiments over many years. For four diseases, Fusarium head blight, tan spot, Septoria leaf blotch, and Stagonospora leaf blotch, at least 20 cultivars each had been evaluated in a minimum of five experiments. Reference cultivars of known disease reaction were included in each experiment, which allowed transformation of the percent disease severity data to a 1-to-9 scale for comparisons between experiments. Variations in scale values obtained for individual cultivars among the different experiments were used to calculate standard deviations for disease phenotype data. Standard deviations were used to calculate the number of experiment repetitions needed within each disease to achieve different levels of accuracy (margins of error). A margin of error of ±1.5 for the 1-to-9 scale was chosen as the best level of accuracy. Rounding values within this range would put the estimated disease phenotype within ±1 unit of the actual phenotype. To achieve a margin of error of ±1.5 for Fusarium head blight, tan spot, Septoria leaf blotch, and Stagonospora leaf blotch would require a mean that was calculated from a minimum of five, five, seven, and eight experiments, respectively. Personnel who report disease phenotype data to wheat producers or breeders should be aware of the number of experiments upon which they are basing their reports and adjust any disclaimers accordingly. Similarly, wheat breeders should be aware of the inherent variability in phenotyping these four wheat diseases and make appropriate adjustments to their selection protocols. With a minimum of five experimental repetitions, disease phenotype values obtained from inoculated greenhouse and field experiments had very high correlations (r = 0.81 to 0.92, P < 0.0001) with published Kansas State University Research and Extension ratings obtained from commercial fields.


Additional keywords: Fusarium graminearum, Septoria tritici

© 2007 The American Phytopathological Society