Quality of phytopathometric variables generated from a ranking scale for the CABMV-passionfruit pathosystem

Abstract

The Cowpea aphid-borne mosaic virus (CABMV) is one of the most important pathogenic agents in passionfruit culture, causing extensive loss throughout the national territory. Efficient quantification of disease symptoms is highly dependent on the methodology used, and is directly related to the quality of data
 generated for later manipulation and analysis. Thus, our objective was to evaluate different methods of using the data collected using a scale based on quality of the generated variables, using statistical parameters. Assumptions of additivity, homoscedasticity and normality of the errors in parametric analysis were tested.
 Experimental quality, for each phytopatometric variable (PV) was tested for calculated F (Fc), coefficient of determination (R²) and coefficient of variation (CV%). Four different PVs were generated through a ranking scale: AUDPC-III, AUDPC-GS, III and GS. All variables met the assumptions for analysis of variance, with
 AUDPC-III and III PVs having slightly higher values in terms of adherence to normality, and AUDPC-GS and GS PVs having slightly higher values in terms of significance for additivity and homoscedasticity. AUDPC-III and III had the highest calculated R² and F values, and the highest coefficients of variation. We recorded the inverse for AUDPC-GS and GS, with lower coefficients of variation and higher R² and F values. A lower correlation,
 though still significant, was observed between AUDPC-GS and AUDPC-III, while a higher correlation was recorded between III and GS. Overall the PVs III and AUDPC-III systems were the best for use in the analyzes of the studied pathosystem.