The ideal method of abdominal wound closure remains to be discovered. This should be technically so simple that the results are just as good in the hands of the trainee as in those of the surgical master; it should be free from the complications of burst abdomen, incisional hernia and persistent sinuses; it should be comfortable to the patient and it should result in a reasonably aesthetic scar. Since 1975, we have been engaged in a series of trials of different techniques of closure of abdominal incisions and, although we can report improvement in our results, these are still far from ideal.Our first investigation was into the importance of suture of the peritoneal layer. This is stressed in the standard surgical textbooks and yet every surgeon encounters the situation in the straining patient with poor tissues where the peritoneal stitches tear through. Moreover, some surgeons defy the rules and leave the peritoneum unsutured without apparent disaster and, in our own animal studies, spread over many years (Ellis 1962(Ellis , 1978, we know that raw peritoneal defects heal rapidly and smoothly.Over There were 4 burst abdomens (2.5%) in the two-layer group, and 5 (3%) in the one-layer group. To date, 8 patients (5%) in the two-layer group and 14 (8.5%) in the one-layer group have developed an incisional hernia. There is no statistical difference between these two groups.In searching for the causes of wound failure in this series, jaundice at the time of laparotomy was found to be of particular importance. There were 21 patients who were jaundiced at the time of surgery, among whom 3 developed a burst abdomen and 4 an incisional hernia. Male sex, closure performed by junior staff, obesity, postoperative chest complications, postoperative distension and wound infection all produced a relatively high incidence of wound failure, but these failed to reach statistical significance. We have previously shown, in rats rendered jaundiced by ligation of the common bile duct, that the healing of the abdominal incision, as well as of visceral wounds, is seriously impaired (Bayer & Ellis 1976).Dismayed by the high incidence of burst abdomen and incisional hernia in patients subjected to layered closure by the technique which is probably that most commonly employed in the United Kingdom, we were naturally attracted by the theoretical arguments supporting the value of non-absorbable sutures in a mass-closure technique which includes all layers of the abdominal wall apart from skin and which incorporates wide bites of tissue on either side of the line of incision, placed close together and under no-tension. Jenkins (1976), who is the principal proponent of the mass-closure technique in this country, records a remarkable series 1