Large amounts of antimicrobial agents are in the production of food animals used for therapy and prophylactics of bacterial infections and in feed to promote growth. The use of antimicrobial agents causes problems in the therapy of infections through the selection for resistance among bacteria pathogenic for animals or humans. Current knowledge regarding the occurrence of antimicrobial resistance in food animals, the quantitative impact of the use of different antimicrobial agents on selection for resistance and the most appropriate treatment regimes to limit the development of resistance is incomplete. Programmes monitoring the occurrence and development of resistance are essential to determine the most important areas for intervention and to monitor the effects of interventions. When designing a monitoring programme it is important to decide on the purpose of the programme. Thus, there are major differences between programmes designed to detect changes in a national population, individual herds or groups of animals. In addition, programmes have to be designed differently according to whether the aim is to determine changes in resistance for all antimicrobial agents or only the antimicrobial agents considered most important in relation to treatment of humans. In 1995 a continuous surveillance for antimicrobial resistance among bacteria isolated from food animals was established in Denmark. Three categories of bacteria, indicator bacteria, zoonotic bacteria and animal pathogens are continuously isolated from broilers, cattle and pigs and tested for susceptibility to antimicrobial agents used for therapy and growth promotion by disc diffusion or minimal inhibitory concentration determinations. This programme will only detect changes on a national level. However, isolating the bacteria and testing for several antimicrobial agents will enable us to determine the effect of linkage of resistance. Since 1995 major differences in the consumption pattern of different antimicrobial agents have occurred in Denmark. The Danish monitoring programme has enabled us to determine the effect of these changes on the occurrence of resistance. The Danish monitoring is, however, not suited to determine changes on a herd level or to detect emergence of new types of resistance only occurring at a low level.