Delayed primary closure or secondary closure of skin and subcutaneous fat in contaminated laparotomy incisions virtually eliminates the risk of wound abscess in clinical practice. Incisional hernias rarely develop in these wounds. This experimental study offers a possible explanation. Longitudinal incisions in the linea alba of female Wistar rats healed under skin incisions which were either sutured ('closed' subgroup) or left to close by secondary intention ('open' subgroup). Postoperative breaking strengths and collagen contents (measured as hydroxyproline) were studied at intervals of 3-120 days. Measurable strength developed by 5 days, at which time 'open' subgroup wounds were found to be weaker and to have less collagen. At all other times, however, 'open' subgroup wounds were stronger, a property which could be ascribed at earlier, but not at later, periods (42 or 120 days), to a higher content and concentration of collagen. At 3 days the rate of collagen production was significantly (P less than 0.025) lower in 'open' than in 'closed' subgroup wounds but the converse was true at 6 and 9 days (P less than 0.025), thus accounting for the changes in biomechanical properties. It is suggested that initially the stimulus for collagen synthesis is greater in the 'open' wounds and this leads to the development of a collagenous structure which is better adapted to resist tensile forces.