Populations of Fusarium spp. from plots growing continuous winter wheat were estimated by dilution plating of soil samples taken in February, April and July 1993. The plots, from two sections of Broadbalk field, Rothamsted, either had the straw removed, or chopped and incorporated into the soil after every harvest since 1986; the plots also bad different fertilizer treatments. Most species were more abundant in July than in February or April. Estimated propagule numbers of F. culmorum, the only cereal pathogen found frequently, were less than in a survey of the same plots in 1964. It occurred in greatest numbers in plots treated with farmyard manure and, in July, with the largest inorganic nitrogen applications. F. oxysporum and F. solani, non-pathogenic on cereals, were mast frequent in plots treated with phosphorus, potassium and magnesium but without added nitrogen. Other non-pathogenic species, of which the most abundant was F. merismoides, showed no consistent response to any of the fertilizer treatments. The method of straw disposal had no consistent effects. The results are discussed in relation to perceived changes in fusarium disease in the U.K. and changing cereal husbandry practices.