A prospective randomized study was conducted to measure the serial thickness of the lower uterine segment (LUS) by transvaginal ultrasonography in a control group of 80 women having no history of uterine surgery and in a study group of 43 women having a history of previous cesarean section (C/S). In the study group, more than 2 mm of thickness of the LUS was considered as good healing and less than 2 mm of thickness as poor healing. After serial sonographic examination, the women with good healing were given trial for labor unless an obstetrical indication for C/S existed. The appearance of the LUS during surgery was compared with antenatal ultrasonographic assessment by direct inspection. Twenty two (79%) of 28 women with a well healed scar had trial labor with the result that 46% had a successful vaginal birth without any uterine rupture or dehiscence. Eight women with poor healing all had elective C/S. Seven women with a 2 mm LUS thickness were individually categorized for delivery mode. Two of those women delivered vaginally. The LUS was found to be thin to translucent in these later two groups. Two mm or less as a criterion for poor healing had the sensitivity and specificity of 86.7% and 100% respectively. The positive predictive value was 100% and the negative predictive value was 86.7%. Ultrasonographic evaluation is effective in predicting the quality of a uterine scar and in differentiating the risk group of probable uterine rupture from the non risk group.