The increased content of soluble glycosaminoglycan-containing forms in sciatic nerves during recovery from crush injury [Shum & Chau (1996) J. Neurosci. Res., 46, 465] suggests that the glycosaminoglycans modulate the environment for post-traumatic tissue remodelling and axonal regrowth. To test this, defined amounts of soluble heparan sulphates from bovine kidney or guinea-pig nerve were introduced into the regenerating environment via silicone conduits that bridged 8-mm gaps of transected sciatic nerves of adult guinea-pigs. Controls were bridged using the phosphate-buffered saline (PBS) vehicle or a chondroition sulphate preparation from whale cartilage. After timed periods of recovery, the animals were assessed for electromyographic signals at the target gastrocnemius muscle to determine the conduction velocity across the bridged nerve. Sections of the bridge were also histologically examined for nerve fibres. Transected sciatic nerves bridged with heparan sulphates or chondroitin sulphate showed earlier stimulated myelination of axons (week 5–6) than PBS-bridged nerves (week 9). Initial electromyographic indication of reconnection with the target was at week 9 post-transection. In the course of 20 weeks, transected sections of the bridge indicated similar numbers of unmyelinated axons irrespective of bridge material, but distinctly higher numbers of myelinated axons in heparan sulphate-bridged nerves than either PBS- or chondroitin sulphate-bridged nerves. At the end of the same period, heparan sulphate-bridged nerves resumed normal conduction velocities, but both PBS- and chondroitin sulphate-bridged nerves remained at 50% of that of the intact contralateral nerves. These results are the first to demonstrate that supplementation of soluble heparan sulphate to the fluid regenerative neural environment can restore functional, axonal reconnection of the severed nerve with the target muscle.