Fifty stapled gastric partitions were performed upon dogs using five techniques, three of which used strips of Marlex or Teflon to reinforce the staple line. At autopsy, five weeks after operation, 22 of 50 partitions were intact and 19 of 50 were completely disrupted. Nine of 50 dogs died of leak and peritonitis within three days of operation. The incidence of disruption was not significantly reduced by reinforcing the stapled partitions with Teflon or Marlex, but the incidence of leak and peritonitis which resulted in death was significantly higher with reinforced partitions, P less than or equal to 0.05. Disruption of the staple line appears to begin at the partition opening and proceeds laterally from that point. Distortion of staples was noted in most of the intact partitions, and in several, the staples remained closed but pulled through the full thickness of the anterior gastric wall. Dilation of the partition opening was noted with each technique. It is our impression from the experiments and the reported clinical experience with ever-changing operative techniques that mucosa-to-mucosa stapling of the undivided stomach is not a reliable technique for gastric partitioning and that additional measures to ensure the success of this simple technique may lead to other complications. At this point, we do not know of an operative technique for the treatment of morbid obesity which we consider to be both safe and effective enough for us to use.