Plain abdominal radiographs and enteroclysis studies were reviewed blindly in 117 consecutive patients undergoing enteroclysis for suspected small bowel obstruction. Plain radiographs were unreliably predictive of the presence of obstruction as determined by enteroclysis and surgery. Among patients with normal or abnormal nonspecific plain radiographs, varying degrees of small bowel obstruction were demonstrated by enteroclysis in 22%. Conversely, of patients with obstruction on plain radiographs, 42% had either normal enteroclysis studies or only minor adhesions. Enteroclysis correctly predicted the presence of obstruction in 100%, the absence of obstruction in 88%, the level (proximal vs distal) of obstruction in 89%, and the etiology of obstruction in 86% of operated patients. Enteroclysis is advocated as the definitive study in patients with clinical uncertainty about the diagnosis of small bowel obstruction.