Net blotch severity is best assessed at early grain filling with respect to its effect on grain weight of spring barley

Appropriate disease assessment methods and parameters reflecting whole-season disease severity levels in field plots remain important issues in studies related to plant disease epidemiology, disease resistance of crop cultivars, and disease-induced yield losses. Such methods and parameters should be yield-related to ensure relevance. Net blotch severity was determined over time in inoculated and non-inoculated field plots of three spring barley varieties by whole plot assessments and by assessments of individual leaves of single main tillers. Disease severity measures such as the area-under-disease-progress-curve, mean and maximum severity as well as severity levels at specific growth stages (GS) were derived from the data. Their relation to thousand grain weight (TGW) and their inter-correlations were examined by means of general linear model (GLM) and factor analyses (FA), respectively. All parameters of net blotch severity were significantly negatively correlated with TGW. Disease parameters derived from whole-plot assessments gave a slightly better explanation of TGW than parameters derived by assessing single main tillers. Net blotch severity at GS 70 (beginning of grain filling) of whole plot assessments yielded the highest adjusted R-squared (0.43) while the adjusted R-squared values resulting from using the same parameter of assessments of the upper three, four or all leaves of single tillers were between 0.34 and 0.35. Also, the residuals of TGW of GLM’s using disease covariates from whole-plot assessments and variety effects as independent variables exhibited less pattern related to other sources of variation than residuals of the corresponding models that used single-tiller-based disease covariates. FA revealed that all disease parameters were highly inter-correlated and co-varied along the 1 principal component axis. The results indicate that disease assessments at GS 70 are appropriate to reflect whole-season severity levels of net blotch. In this respect, the time consuming single-tiller method is in this respect not superior to the simpler whole-plot method. However, assessing individual leaf layers of single tillers allows to observe the epidemic development and thus to examine the dynamics of epidemics in much greater detail than assessing whole-plots. This showed, for example, how much each leaf layer contributed at any given time to the total disease and revealed that a substantial fraction of the total disease is being removed during the course of an epidemic by senescence of diseased lower leaves. This level of detail in examining the dynamics of epidemics cannot be achieved by the whole-plot method.