1SAC, Ferguson Building, Craibstone Estate, Bucksburn, Aberdeen AB21 9YA, UK; 2IACR-Rothamsted, Harpenden, Herts AL5 2JO, UK; 3ADAS Boxworth, Cambridge CB3 8NN, UK; 4Central Science Laboratory, MAFF, Sand Hutton, York Y04 1LZ, UK

Background and objectives
In the UK, an estimated 3-9 M are spent each year on fungicides to control winter oilseed rape diseases, but losses in excess of 30 M still occur [1]. Light leaf spot (Pyrenopeziza brassicae) is one of the major diseases of winter oilseed rape. The optimum time for control of severe light leaf spot is in the autumn [2], but symptoms often do not appear until later (December-February). Disease epidemics differ in intensity between seasons and between regions. However, many growers make spray decisions on the basis of local practice rather than on the actual disease levels present. Consequently, many crops with severe light leaf spot epidemics are left unsprayed whereas crops with little disease receive a full fungicide programme. Growers require a more accurate method to determine the need for fungicide sprays. Work on a forecasting system, being developed to predict severe light leaf spot epidemics and associated field experiments, is reported.

Materials and methods
Data on the incidence of light leaf spot in winter oilseed rape crops was collected from ADAS/CSL and SAC surveys (1977-95). Step-wise regression analyses relating the incidence of light leaf spot on leaves in March or on pods in July to environmental and agronomic factors were used to develop a forecasting scheme based on regional and crop risk indices. Replicated field trials were carried out at three sites to estimate yield loss from the disease. Oilseed rape stubble was applied to the two sites in eastern England (Rothamsted and Boxworth) to provide inoculum for artificial epidemics. Different disease epidemics were created by applying the fungicide tebuconazole at various timings during the autumn and/or spring. Disease progress was followed by sampling, incubating and assessing plants for light leaf spot.

Results and conclusions
Regression analyses indicated that seasons with severe light leaf spot epidemics could be predicted from disease incidence on pods in the previous July. Individual crops were at risk where cultivar resistance was <5, sowing date was in August and where there was a high rainfall. In October 1996, the forecasting system predicted severe risks of light leaf spot for Scotland, northern England and Wales and low risks for east and south-east England. These predictions were confirmed by survey results in the spring of 1997. Light leaf spot epidemics were severe in the field experiments at Aberdeen (natural inoculum) and Rothamsted (high inoculum density) but less damaging at Boxworth (low inoculum density). The disease was first observed in untreated plots of the susceptible cultivar Bristol (resistance rating 2) between November and January. Light leaf spot development was rapid at Aberdeen and Rothamsted, reaching 92-100% plants affected, with 12-24% leaf area affected, by March 1997. Disease development at Boxworth was slow and by April only 65% of untreated Bristol plants were affected, with 3% leaf area affected. At all three sites development of light leaf spot in the resistant cultivar Capitol started later in the season, but by March/April disease incidence was only slightly lower than that on Bristol. Monthly applications of a half dose of tebuconazole from October to March or a half dose in November/December + a half dose March gave the largest decrease in incidence and severity of light leaf spot. At Rothamsted, monthly applications of fungicide resulted in a 1.5 t/ha yield benefit but there were no yield responses to fungicide application at Aberdeen or Boxworth. Work is continuing to further improve the forecasting system.

1. Fitt BDL, Gladders P, Turner JA et al., 1997. Aspects of Applied Biology 48,135-142.
2. Sansford CE, Fitt BDL, Gladders P et al., 1996. Home-Grown Cereals Authority Project Report OS17.