Since its introduction for general use, the role of rigid ureteroscopy in the diagnosis and therapy of urological disease has been in evolution. We evaluated retrospectively the experience at our institution with rigid ureteroscopy from January 1983 to July 1986 in an attempt to identify clinical situations or techniques that tended to increase the incidence of either success or complications. We determined that ureteroscopic complications were rare in procedures performed for diagnosis compared to those performed for calculi (2 of 33 or 6 per cent versus 27 of 99 or 27 per cent, p less than 0.025). The likelihood of failure or complication was greater for stones above than for those below the pelvic brim (15 of 25 or 60 per cent versus 26 of 75 or 35 per cent, p less than 0.05). Major complications were more common early in our experience (9 of 63 or 14 per cent versus 2 of 69 or 3 per cent for the combined years 1983 and 1984 compared to 1985 and 1986, p less than 0.05). Our success rates in the treatment of calculous disease were similar to those reported previously, and they were somewhat better for stones located below (62 of 75 or 83 per cent) than for those above (17 of 25 or 68 per cent) the pelvic brim. Based on our findings we conclude that carefully performed diagnostic ureteroscopy has little potential for major complications, ureteroscopy for stones above the pelvic brim should be avoided when possible, and an increased rate of complications and failures is expected early in any series owing at least partly to the learning curve effect.