Aids to the rapid diagnosis of urinary tract infection were assessed by the examination of 325 consecutive urine samples taken in the normal course of work in a general practice. Of these samples 103 produced a pure growth of at least 10(5) organisms per ml. The appearance and smell of each sample was noted and it was then tested by simple low-power microscopy of a drop of urine and by a dipstick which measured leucocyte esterase and nitrite, together with protein, blood and pH. In addition, pus cell counts per mm3 were performed on 272 of the samples using a cytometer chamber. This method is too time-consuming for routine use in the surgery. Neither a cloudy appearance nor haematuria were sufficiently specific to be of much use in the diagnosis of urinary tract infection. In the prediction of a 'positive' culture the sensitivity and specificity of the other tests were as follows: drop method microscopy 95% and 76%, respectively; cytometer count 95% and 81%; leucocyte-esterase estimation 89% and 68%; and nitrite 57% and 96%. These figures may underestimate the true values of the tests in the diagnosis of urinary tract infection because infection may be present in some cases producing growths of less than 10(5) organisms per ml. It is concluded that the most useful aid to the diagnosis of urinary tract infection is low-power microscopy of a drop of urine.