The University of Reading, School of Plant Sciences, Department of Agricultural Botany, TOB2 Earley Gate, Whiteknights, Reading RG6 2AU, UK

Background and objectives
Many foliar pathogens infect poorly in the presence of large populations of phylloplane microorganisms. Artificially augmenting phylloplane communities for biological control has usually failed because augmented populations rapidly decline and to be effective often have to be introduced before the pathogen. Populations of saprophytes may be enhanced in such systems with high organic matter content at the soil surface, because of the proximity of splash-borne nutrient and inoculum sources. This may explain observations that straw incorporation on average suppresses Septoria tritici; however, large popuiations of saprophytes may themselves reduce yields. These studies aim to compare phylloplane microflora and disease severity on wheat grown in plots with different ground preparation: straw incorporated, manure incorporated or fallow. Hypotheses of potential mechanisms of disease suppression have also been tested.

Materials and methods A field trial was conducted at Shinfield near Reading on land which had previously been Pasture. Twelve 6X6 ; m plots were used, arranged as three randomized blocks. Three plots had 4 ;kg/m2 manure added, three had 1 ;kg/m2 straw added and six were untreated. The plots were sown in October with winter wheat cv. Mercia at 400 ;kg/ha. Each plot was separated by a 2-m barley guard. In March, a herbicide treatment was applied and soil surface nitrogen levels were equalized by adding ammonium nitrate. Between February and June, at regular intervals, phylloplane microflora were assessed by dilution plating of washings of the top two leaves of sampled plants. Disease severity was recorded on the top four leaves. Representative samples of the microflora were identified using Biolog or PCR. The ability of these microorganisms to produce antibiotics, degrade chitin and produce siderophores was tested.

Results and conclusions
Microorganism numbers fluctuated slightly, but increased overall during the season. Bacteria predominated but yeast increased later in the season. Pink colonies were mostly Sporobolomyces roseus, yellow were commonly Pantoea agglomerans, Pseudomonas fluorescens, P. ;viridiflava<,i>, Erwinia chrysanthemi and P. ;tolaasii, and white included P. ;corrugate, P. ;fluorescens, Cystofilobasidium spp. and Leucosporidium scottii. Some cultures, especially the yellow bacteria, produced antibiotics against S. ;tritici when stressed. Competition for iron is likely to be important in the phylloplane. Approximately 20% of the cultures sampled produced high levels of siderophores and hence can be assumed to be effective competitors for iron. These were mostly Gram-negative bacteria, including P. ;fluorescens. The main disease present was S. ;tritici. It was more severe in the straw plots in late February than in the other two treatments. This was correlated with greater microorganism populations. The straw could have acted as a local source of S. ;tritici and saprophyte inocula but it also interfered with sowing, resulting in a sparser crop and increasing splash dispersal. S. ;tritici severity rose sharply in March. By the end of April straw-treated plots still had slightly more disease but microorganism numbers were now similar in all treatments. During early May, disease severity and microorganism numbers were approximately equal between treatments. Later in the month there was more disease (P<0.1) on the manure-treated plots, which also had greater microorganism populations. The straw-treated plots now had the lowest disease and microbial numbers; perhaps the plants had been primed for resistance by the early exposure to disease and/or high levels of saprophytes. This trend continued throughout June with the manure-treated plots having the most disease and microbial flora while the other two treatments had similar, lower, levels. On average, straw-treated plots had slightly less disease and fewer microorganisms than the untreated plots.