The aim of the present study was to assess whether the presence of polymorphonuclear leukocytes in the first stage of fracture repair was of any importance. In anesthetized male Wistar rats, a transverse osteotomy was performed at midshaft in one femur and immediately stabilized by an intramedullary Kirschner nail. The animals were allowed unprotected weight bearing immediately. Nine animals were made neutropenic before operation by injection of a specific rat polymorphonuclear leukocyte antiserum (antineutrophil serum, ANS) raised in sheep. The number of circulating polymorphonuclear leukocytes were kept below 20% of their normal value for 72 h postoperatively by intraperitoneal injections of ANS every 12 h. Control animals were injected with the same amount of normal sheep serum (NSS). All animals received cefuroxime (100 mg/kg day) concomitantly with the ANS injections. Six weeks after operation the animals were killed, the amount of callus formation was measured, and the bones were radiologically examined. The nails were then removed and both the healing fractures and the nonoperated femurs mechanically tested for bending. There were no differences in the amount of callus measured or in radiological healing. However, there was a significantly higher bending moment in femurs from animals treated with ANS than in those given NSS (P less than 0.02). No differences were observed in rigidity or total energy absorption.