The effects of small mammals on invertebrate prey populations in grassland were investigated by means of an exclosure experiment over a period of 2 yr. Using direct searches and suction sampling, numbers of invertebrates were compared in four experimental plots fenced to exclude small mammals and four control plots where they had free access. There was no evidence of differences in guild structure or species composition in either large or small invertebrates as a result of mammal predation, but numbers of large invertebrates were consistently and significantly smaller (by 22.7-64.8%) in the presence of mammals, due principally to the predatory effects of shrews. Most affected were populations of Isopoda, Mollusca, Myriapoda, Araneae and Opiliones. Numbers of less commonly occurring invertebrate taxa, such as Coleoptera, showed more erratic differences between control and experimental plots. Numbers of small invertebrates collected by suction sampling were highly variable and, although often reduced in the presence of mammals, the treatment differences were not always significant. It is concluded that small mammal predators, with their capacity for high population densities and consumption rates, have a major effect on invertebrate communities in grassland.