Twenty-one patients referred for evaluation with a diagnosis of questionable ascites were examined independently by three investigators who performed five different physical examination maneuvers. With ultrasonography as the reference standard, six patients had ascites. The sensitivity and specificity of the physical examination maneuvers ranged from 50% to 94% and 29% to 82%, respectively. The overall accuracy of the maneuvers was only 58%. The results of this study indicate that routine physical examination has definite limitations in the precise diagnosis of equivocal ascites. The only conclusion that can be made by a physical maneuver with over 90% accuracy is that ascites is absent if no flank dullness is elicited. Ultrasonography is recommended in questionable cases.