Perennial ryegrass was ensiled in laboratory silos after addition of formic acid (850 g kg-I) or sulphuric acid (906 g kg-I) at rates of 0, 2, 4 and 6 litres t-' fresh grass. Silos were opened after 6, 18 and 90 days and the silage subjected to chemical and microbiological analysis. The untreated control was poorly fermented with a final pH of 4.7, a butyric acid concentration of 19 g kg-I dry matter (DM) and an NH3-N content of 275 g kg-I total nitrogen (TN). For the formic acid treatments the 2 litre t-I and 6 litre t-I levels both produced well-preserved silages but they were of different types. The silage treated with 2 litre t-I had a p H of 4-0, a lactic acid concentration of 92 g kg-I DM and 161 g NH3-N kg-I TN, whereas with the 6 litre t-I treatment, fermentation had been severely restricted. The pH was 4.2, the lactic acid concentration was only 8 g kg-' DM and the NH3-N content was 80 g kg-' TN. However, formic acid at 4 litre t-' produced a badly fermented silage of final pH 5.0 with lactic acid and butyric acid concentrations of 16 and 15 g kg-1 DM, respectively, and an NH3-N content of 149 g kg-I TN. Sulphuric acid at 2 and 4 litres t-I produced silages of low lactic acid contents, 36 and 24 g kg-I DM, and they also contained butyric acid in concentrations of 13 and 11 g kg-I DM; respective NH3-N contents were 206 and 114 g kg-I DM. When sulphuric acid was added at 6 litres t-l, despite a reduction in the pH of the grass to 3.5, fermentation was not restricted as it was with the equivalent level of formic acid. Lactic acid was present at 27g kg-I DM and the ethanol concentration was very high at 66 g kg-I DM; the sulphuric acid-treated silages were characterised by high yeast counts. A t the higher rates of addition, formic acid reduced the 217 1. Sci. Food Agric.