The appropriateness of empiric therapy, how closely prescribers adhered to hospital guidelines for use of antimicrobial drugs and how relevant these guidelines proved in providing cover for the infecting pathogen were assessed from a retrospective survey of culture and sensitivity reports issued during a one-month period in a district genera hospital. Blood, stool, pus, sputum and mid-stream urine specimens were reviewed. Urine specimens from a hdholding general practice unit were also included. Catheter specimens of urine and specimens from other body sites were excluded. Of 290 specimens sent for culture, 99 (34 per cent) were positive. One hundred patients, of whom 54 (19 per cent of the total)were subsequently demonstrated to have a positive culture, had been treated empirically. In 38 per cent of cases the pre-report empiric treatment was considered appropriate for the isolated pathogen, while 46 per cent of patients had negative cultures and 16 per cent had positive cultures but had been given unsuitable antibiotics. In hospitalised patients, 73 per cent of antibiotic treatments were prescribed according to hospital guidelines, although this was the case for only 45 per cent of community-based patients. Hospital guidelines were a relevant and useful aid to rational prescribing in that 75 per cent of organisms isolated were sensitive to guideline antibiotics. The guidelines were least appropriate for urine specimens from hospital. This reflects the range of organisms encountered and the emergence of resistant strains, reinforcing the need for regular updating and amending of prescribing recommendations.