An experimental study was designed to test the efficacy of a silicon barrier around a nerve root in order to prevent post operative epidural root scar adhesions. In 32 Sprague-Dawley rats a lumbar nerve root was microsurgically exposed bilaterally. In 16 animals a silicon tube, prepared with a longitudinal cut along half of its wall, was placed around one nerve root, while the contralateral side served as control. In another group of 12 animals, an autologous subcutaneous fat graft was placed on the exposed root. A group of 4 animals served as the control group which had not been operated upon. No post operative neurological deficit was observed in any of the animals. All animals were sacrificed 60 days after the operation and a block of tissue including the nerve roots were resected bilaterally without removal of the silicon or fat graft. Longitudinal and transverse cuts of the roots were stained with Haematoxilin and Eosin and with Masson's trichrome collagen stain. The roots in the "unoperated" control group were clean of any scar tissue. In 13 out of 16 animals, silicon prevented scar formation around the root as opposed to scar adhesions around control root on the contralateral side and as compared to unoperated roots. In the silicon group, adhesions penetrated only through the longitudinal narrow cut edge of the tube. Fat did not prevent adhesions in 11 out of 12 animals. We conclude that a silicon barrier is an effective method preventing post operative epidural root scarring in rats.