PURPOSE: To demonstrate the clinical usefulness of ultrasonography for detecting esophageal varices. MATERIALS AND METHODS: In 20 cases of esophaged varix, the authors analysed the transabdominal ultrasono-graphicfindings of the esophagogastric junction and compared mural thickness, the anteroposterior diameter of theesophagus, and the echogenic nature of the esophageal mucosal layer with those of 78 normal patients. RESULTS:The anterior and posterior mural thickeness of normal esophagus was 2.2 +/-0.7 and 2.4 +/-0.8mm re-spectively, butfor variceal esophagus, the corresponding readings were 5.9 +/-1.3 and 5.2 +/-1.3mm respective-ly. Theanteroposterior diameter of normal esophagus was 7.9 +/-2.1mm and that of variceal esophagus was 1 4 . 0 +/-1.8mm.There was a stastically significant difference (p<0.01) in mural thickness and anteroposterior diameter of theesophagus between a normal and variceal patient with regard to change of echogenic nature at the esphagogastricjunction. Normal esophageal mucosa showed a thin and uniform echogenic line, but for variceal mucosa, theechogenic pattern was irregular, tortuous and thick. CONCLUSION: The athors believe that transabdominal US ishelpful for detecting esophageal varices in patients with liver cirrhosis and UGI bleeding. Important clinicallyuseful sonographic findings in diagnosing e-sophageal varix are as follows: 1) mural thickness more than 6mm; 2)anteroposterior diameter of the esopha-gus of more than 15mm; 3) irregular, tortuous and thickened echogenicmucosa.