AbstractCattle slurry (50 m3ha−1equivalent to 68 kg N ha−1) was applied to grassland plots 70 d (early application) and 34 d (late application) before ensiling and the retention and survival of slurry and epiphytic micro‐organisms on the growing herbage were examined and compared with those on herbage from corresponding fertilizer‐treated plots. The populations of lactic acid bacteria, enterococci and enterobacteria on herbage increased dramatically after slurry application. Thereafter, numbers of lactic acid bacteria declined, although they were always higher than on untreated herbage. Number of enterobacteria also declined but were higher on chopped grasses at ensiling [106colony‐forming units (CFU) g−1fresh matter (FM)] than they were on hand‐cut. unchopped herbages at all previous sampling times. Clostridia numbers were lowest on untreated and highest on slurry‐treated herbage, particularly after the late application (>103CFU g−1FM). Herbage was harvested 70 d and 34 d after slurry application, chopped and ensiled in laboratory silos. All herbages, irrespective of treatment, had low dry matter (DM) values (ranging from 149 to 170 g kg−1FM) and fairly low water‐soluble carbohydrate (WSC) concentrations (130 g kg−1DM or less). The initial rate of pH decline up to 4 d was most rapid in slurry‐treated herbages, with all pH values falling to4. 5. accompanied by a marked decline in lactic acid concentration. Lactic acid‐fermenting Clostridia increased in numbers, reaching peak values of 107CFU g−1FM by day 21, remaining high until opening, and were probably responsible for increases in butyric acid levels in all silages, with the highest concentrations occurring in those prepared from slurry‐treated herbages. The results suggest that, although some faecal lactic acid bacteria may have beneficial effects in the early stages of fermentation, Clostridia from slurry can survive on herbage for extended periods. The results indicate that the potential for growth of Clostridia in silage may be independent of source or size of the initial population even at tow pH, if