Epidemics of Phymatotrichum root rot (Phymatotrichum omnivorum) in cotton: environmental correlates of final incidence and forecasting criteria

Abstract

SUMMARYEpidemics of Phymatotrichum root rot (PRR), caused by Phymatotrichum omnivorum, in cotton were monitored in field plots at the Blackland Research Center, Temple, Texas during the years 1969–1982. In most years disease incidence, assessed at periodic intervals during the growing season, increased smoothly and levelled off at values in the range 0–04‐0‐99 (proportion of plants killed). During two years there were marked increases in PRR late in the growing season. Increments in PRR were associated with preceding increments in precipitation but only when the latter were large. The final incidence of PRR late in August was directly related (in all but one of the 14 years) to cumulative precipitation in the range 36–100 cm, and inversely related (in all 14 years) to air temperatures greater than 34 oC. Regression was used to relate final PRR incidence to derived weather variates but the equations obtained did not provide a realistic basis for forecasting: the best predictive variates included cumulative precipitation up to mid‐August. A threshold criterion given by P/T > 1·45, where P was cumulative precipitation (cm) from 1 January and T was the mean maximum temperature (°C) during the preceding 10 days, was derived. This criterion satisfactorily grouped years with high (>0·;50) and low (<0–50) incidences of PRR. In those years where a forecast of high incidence was made, the criterion was satisfied prior to the first appearance of PRR (mid‐June); where a forecast of low incidence was made, the criterion was not satisfied until late August. The forecasting criterion was evaluated at three sites in the Blacklands region in 1983, at two in 1984, and at Temple only in 1985. In 1983 and 1984, final PRR incidences of less than 0–50 were forecast and observed; in 1985, cumulative annual precipitation exceeded 50 cm by early June, the threshold criterion was satisfied prior to the first observation of symptoms, and PRR incidence was 0–80 late in August.