SUMMARYExperiments are described in which successive groups of healthy, susceptible, potted hop‐plants were exposed, each 24 h, to the weather and to sources of Pseudoperonospora kumuli inoculum in an unsprayed hop garden for 48 h periods during May and June in 1969, 1970 and 1971. The infection which arose after standardized incubation in isolation was measured and then related, both by inspection and by multiple regression analysis, to the conditions during exposure.Severe infection could be associated with certain sequences of events in which rain contributed substantially to plant wetness; relatively light or no infection developed when dew provided the wetness of a period. Infection was markedly correlated with the occurrence of infection periods, as defined from earlier growth‐room studies, especially when dew‐wetness was omitted.The regression analysis showed that, in all years, infection was highly correlated with variables reflecting wet conditions. It was inconsistently correlated with vapour pressure deficit (VPD) and with sunshine, not at all with temperature, and with airborne spore concentration only in the combined years. The spore catches of a funnel trap, used in 1971, which depended upon rain varied closely with infection. R2 values for multiple regression equations indicated that, in the years separately and together, well over 70% of the variation in infection could be explained by the environmental variables and, in 1971, almost 90%. Variables (temperature, VPD, airborne spores) which singly were poorly correlated with infection often significantly improved regression equations which were based on variables expressing wetness. When prediction of each year's infection levels was attempted using selected regression equations calculated from data of the other 2 years there was, in general, very good agreement between measured and estimated values. The best predictions were given by an equation utilizing rain‐wetness duration, rainfall amount and airborne spore concentration while an equation based solely on relative humidity and rainfall was only slightly inferior.The results are discussed in relation to the possible value of the regression models for short‐term forecasting of hop downy mildew.
Read full abstract