The aim of the study was to explore whether penicillin was superior to placebo in altering the clinical course of proven streptococcal pharyngitis. A randomised, parallel, double blind placebo controlled trial of 10 days duration was undertaken in 42 general practices in the Gent region (Flemish part of Belgium). Phenoxymethylpenicillin (adults 250 mg t.i.d. and children 125 mg t.i.d.) or placebo were administrated to 173 patients, aged 5 to 50 y, with acute sore throat and a positive culture for Group A beta-haemolytic streptococci. Penicillin and placebo tablets were identical. Patient compliance was monitored by assay of penicillin in urine (Sarcina lutea method). The primary outcome variable was sore throat as recorded by the physician on Day 3. The experiences of the patients themselves over the 10 day period were also assessed. Secondary outcome variables were other local and general symptoms and signs of streptococcal throat infection. In the penicillin group on Day 3, 23.2% of the patients still complained of sore throat versus 65.9% in the placebo group: difference 42.7% (C.I. 29.4%, 56.1%). This finding was confirmed by survival analysis of the symptom 'sore throat', as recorded by the patients. The physicians recorded on Day 3 a significant positive effect on another symptom (malaise: P < 0.04) and certain clinical signs (abnormal throat: P < 0.07; and redness of throat: P < 0.003). Penicillin had more adverse effects than placebo (P < 0.007). It also inhibited the rise in ASLO (P < 0.001). In this study in general practice, penicillin had a slight but definitive positive effect on the clinical evolution of streptococcal pharyngitis.(ABSTRACT TRUNCATED AT 250 WORDS)