Control of septoria and eyespot in winter wheat fungicide applications in response to growth stage and forecasts of disease risk


Cereals & Oilseeds
Project code:
01 May 1987 - 30 April 1991
AHDB Cereals & Oilseeds.
AHDB sector cost:
£163,092 From HGCA (Project No. 0008/4/87)
Project leader:
S R Parker and D J Royle IACR-Long Ashton M J Hims CSL-Harpenden



About this project


This co-ordinated HGCA project attempts to provide farmers with options for saving fungicide sprays and improving disease control pre-anthesis by integrating decision-making, particularly on septoria and eyespot control. Under a wider range of conditions than would be possible otherwise, it also seeks to define more precisely critical conditions for disease development and study the effects on yield of disease epidemics which develop at different times during the season.

In 1987 observations were made at Long Ashton, Rothamsted and Department of Agriculture for Northern Ireland DANI), Belfast, to develop hypotheses and experience for a series of field experiments of a common design to be established in the 1987-88 season. In an experiment at Long Ashton with 4 cultivars of winter wheat (Avalon, Slejpner, Brock and Moulin) grown at 4 sites which differed with respect to aspect, soil type and cropping sequence, Septoria tritici differed in severity according to site but not cultivar. Eyespot also varied between sites rather than cultivar, though Slejpner seemed least affected. In a separate experiment, the onset of severe S. tritici on the top leaves of cv. Longbow was delayed to a similar extent by different forecast-guided timings of prochloraz; this was unexpected because the first application was before the top leaves had emerged. Observations at Rothamsted confirmed earlier evidence that the R-pathotype of eyespot initially progresses more slowly from leaf sheaths to stems than the W-type. Further evidence was also obtained that fungicide sprays at GS 37 can be as effective at controlling eyespot as at GS 31. In N. Ireland yields and 1000-grain weights of cv. Norman increased as the number of sprays of several fungicide combinations were increased in a field experiment. The effects of treatment on disease were not, however, consistent between assessments made at GS 57 and GS 71.

In the harvest seasons between 1988-90 field experiments were done by Long Ashton, Rothamted and DANI to examine the effects of different growth stage based and forecast guided spray programmes on disease progress, control and grain yield. Greatest benefits were from sprays applied to the flag leaf, though differences in disease control and yield response between sprays applied at different times after flag leaf emergence varied and were inconsistent.

In terms of yield none of the spray programmes consistently out-performed the others. However, the Long Ashton Septoria forecast, and a programme comprising sprays at GS 37 & 59 performed consistently over cultivars, sites and seasons. A 3 spray programme always provided a significant yield increase above that of the untreated, but there was strong evidence that the 3rd spray was not economically justified in a high proportion of cases.

Satisfactory models explaining disease intensity and yield were not obtained for these experiments. In part this was because the experiments were designed for analyses by other techniques. However, recent studies at Long Ashton have shown that the interaction between winter wheat and S. tritici is far more complex than previously believed. As a consequence, accurate models that are reliable in the long-term will need to account for these interactions. A model explaining yields in terms of eyespot and the maximum rate of increase of S. tritici on the flag and 3rd leaves was fitted to the DANI data. This model explained 84% of yield variation, but it was only appropriate for cv. Brock. None of the models fitted to Long Ashton data were convincing.

Spray programmes starting at GS31 tended to control eyespot better than those starting at GS37. Light rainfall occurring after a spray generally enhanced eyespot control, presumably because this provided a mechanism for redistribution of the chemical to the base of the crop. In one case there was evidence that heavier rainfall may have reduced fungicide performance by washing-off, because both S. tritici and eyespot control were reduced.

Between 1988-89 ADAS conducted a complementary programme of 6 experiments designed to create a series of different disease epidemics. The aim was to investigate the properties of a range of protectant and eradicant fungicides, active against several foliar wheat diseases.

Moderate to severe disease caused by Septoria tritici occurred in five of the experiments. These experiments successfully identified some of the periods conducive for S. tritici infection and provided an insight into factors affecting the development of mildew and the rusts. The experiments also indicated the effect of fungicide applications, upon disease development and yield, when they were made either side of the infection periods.

Analysis of disease progress and weather records suggested that critical conditions for initial development of S. tritici occurred during late April to mid-June at all sites. Heavy rain, giving 5 mm on 1 day or a total of 10mm or more on 2 or 3 successive days, occurred at all sites prior to the appearance of symptoms on a particular leaf layer. At two of the sites rainfall during May was considerably lower than at the other sites. This was reflected by a delay in symptom development and disease did not become severe on the top two leaves. At all sites, the length of the incubation period on any of the top three leaves varied between 295 and 448 degree days, although precise identification of the primary splash-event responsible was not always possible.

Mildew developed at only two of the sites. Rapid increase of mildew occurred only once the crop canopy was complete, indicating that crop microclimate may have an influence by governing the rate of inoculum production from within the crop. Disease progress curves supported this possibility because mildew was developing on the lower leaves prior to its appearance on younger leaves. Brown rust and yellow rust occurred at only one site each. Their appearance and progress were associated with temperatures reported to favour infection, tissue colonisation and sporulation. However, it was impossible to estimate the times of primary or secondary infection because records of leaf wetness were not taken at either site. Eyespot developed to slight or moderate levels at three of the six sites. A moderate or high degree of control was achieved by one or more elements of the sequential spray programmes. This effect was associated with the occurrence of moderate to substantial amounts of rain within a few days of fungicide application.

Sequential spray programmes commencing later than growth stage (GS) 31 but immediately prior to the critical periods (splash-events) for leaf infection by S. tritici provided the best disease control and yield benefit. Similar effects were apparent with mildew and rusts. Regression models incorporating, as independent variables, percentage leaf disease on leaves 1, 2 and/or 3 satisfactorily explained yield loss at the six sites.