Sixty-five ankle arthrograms of the patients who underwent surgical reconstruction for their chronic ankle instability from 1987 to 1995 were studied retrospectively by comparing their operative findings. There were 15 cases (23.1%) with definite positive arthrographic findings. In the remaining 50 cases (76.9%), the arthrograms showed a "tear-drop" extravasation of dye beneath the fibular tip which we called "suspectable positive". Of these all but one, which had normal ligament, proved to correspond to the operative finding. As there was only one false positive finding among our surgical cases with both the "definite" and "suspectable" positive arthrograms, we concluded that ankle arthrography is relatively reliable and may be worth doing even in chronic cases if only for reference. Even when other diagnostic methods are negative, arthrography might help surgeons make surgical decisions in clinical symptomatic cases.