It has been suggested that the therapeutic use of oral chloramphenicol in animals is liable to select resistance to antibiotics and that the resistance may jeopardise the treatment of infections in man. At present this risk appears minimal; resistance to chloramphenicol in animal bacteria may well be selected by the increasing use of semi-synthetic penicillins because of linkage between genes coding for production of beta-lactamase and resistance to chloramphenicol. Among salmonellae, the strains causing enteric fever have no animal reservoir and the few food poisoning incidents in man that require therapy can be treated with antibacterial agents such as trimethoprim. Chloramphenicol is not now the antibiotic of choice for any human infection except perhaps a few caused by Haemophilus influenzae. Resistance to antibiotics in 'human' cultures has largely been selected by the use of antibiotics in human medicine. Control of salmonellosis is essentially a public health, not a therapeutic problem.