Pathogenic variability of downy mildew (Sclerospora graminicola) on pearl millet. II. Statistical techniques for analysis of data

Abstract

SUMMARYDifferent cultivars of pearl millet were exposed to infection by asexual propagules of Sclerospora graminicola in Polythene tunnels. Subsequent incidence of disease was measured on one‐month old seedlings. The data consisted of binomial proportions based on differing total numbers of seedlings. It is shown how to utilise the logit link in the GLIM statistical package for analysis to compare variety performances. In addition the possibilities of interplot interference and intraplot variation were investigated using GLIM. Interplot interference occurred infrequently and did not affect the rank order of the cultivars in terms of their response to the fungus infection. Intraplot variation was also shown to be non significant in spite of proximity of some plants to the disease spreader plants. Once these potentially confounding phenomena were eliminated the cultivars could be compared directly with impunity. GLIM was also used to confirm that the type of symptom expression was a stable feature of each host genotype.