Every year between 2002 and 2005 approximately 100 samples of oats from fields of known agronomy were analysed by GC/MS for 10 trichothecenes: deoxynivalenol (DON), nivalenol, 3-acetylDON, 15-acetylDON, fusarenone X, T-2 toxin (T2), HT-2 toxin (HT2), diacetoxyscirpenol, neosolaniol and T-2 triol. Samples were also analysed for moniliformin and zearalenone by HPLC. Of the 10 trichothecenes analysed from 458 harvest samples of oat only three, 15-acetylDON, fusarenone X and diacetoxyscirpenol, were not detected. Moniliformin and zearalenone were absent or rarely detected, respectively. HT2 and T2 were the most frequently detected fusarium mycotoxins, present above the limit of quantification (10 microg kg(-1)) in 92 and 84% of samples, respectively, and were usually present at the highest concentrations. The combined mean and median for HT2 and T2 (HT2 + T2) was 570 and 213 microg kg(-1), respectively. There were good correlations between concentrations of HT2 and all other type A trichothecenes detected (T2, T2 triol and neosolaniol). Year and region had a significant effect on HT2 + T2 concentration. There was also a highly significant difference between HT2 + T2 content in organic and conventional samples, with the predicted mean for organic samples five times lower than that of conventional samples. This is the largest difference reported for any mycotoxin level in organic and conventional cereals. No samples exceeded the legal limits for DON or zearalenone in oats intended for human consumption. Legislative limits for HT2 and T2 are currently under consideration by the European Commission. Depending on the limits set for unprocessed oats intended for human consumption, the levels detected here could have serious consequences for the UK oat-processing industry.