A random sample of 200 males aged 50 years or more was selected from the National Register in order to investigate various aspects of spontaneous uroflowmetry and to estimate the frequency of prostatism and symptoms of lower urinary tract dysfunction. All in all the data from 112 persons were analysed. The prevalence of prostatism was found to be 17% (95% confidence limits 11-26%), while 88% (95% confidence limits 83-95%) experienced various degrees of symptoms. Only few associations between single symptoms and uroflowmetry variables were demonstrated. However, statistically significant correlations were revealed between obstructive and total symptom scores on one hand and maximum and average flow rates on the other hand. As the correlations were modest and a considerable overlap of uroflow variables in persons with and without prostatism were proved, the diagnostic specificity and sensitivity of maximum flow rate as well as other uroflow variables were low in the screening for prostatism. In conclusion uroflowmetry appears inefficient to confirm a clinical impression of prostatism.