A prospective study of genital infection was conducted in four inner-city family–planning clinics. Fifteen per cent of routine attenders had symptoms and signs of vaginal infection and many more women attended primarily because of symptoms. Among the women with both signs and symptoms, 70% had positive laboratory findings, Trichomonas vaginalis, Candida albicans and bacterial vaginosis being equally prevalent. Measurement of vaginal pH in the clinic was the single most useful clinical finding for directing empirical therapy. Among patients with a discharge confirmed on examination and an abnormally high pH, 72% had either T. vaginalis or bacterial vaginosis. Neisseria gonorrhoeae was isolated from 4% of women with, and 1% of those without, symptoms. We believe that it is worthwhile to investigate patients presenting to family-planning clinics with vaginal symptoms. No single specimen was found ideal for all pathogens, a cervical swab is better for gonococci and also for T. vaginalis but a vaginal swab is needed for Candida and bacterial vaginosis.