The surface area of 99 leg ulcers was measured from a scaled photograph using a computerized ultrasonic digitizer. This was compared with the area obtained by multiplying the two maximal perpendicular diameters of each ulcer. There was an extremely good correlation between these two methods (r = 0.951). Seventy-four patients were followed up as part of a placebo-controlled double-blind study. In this study, treatment was assessed by the time taken for the ulcerated limb to heal completely. The initial ulcer size was found to be a weak predictor of subsequent ulcer healing (r = 0.49). The healing rates of individual ulcers calculated over 1 month intervals from presentation proved to be a poor predictor of the time required for complete ulcer healing (Spearman rank correlation coefficients ranged from 0.15 to 0.61). The healing curves of individual ulcers showed considerable fluctuations during the process of healing. The product of the maximal dimensions of an ulcer provides an easy and accurate method of monitoring treatment. The reduction in ulcer size within a set time interval, used in many ulcer studies, is a poor predictor of eventual ulcer healing. The percentage of ulcers completely healed within a pre-determined time interval is a better method of assessing new treatments.