Transected axons are often assumed to seal at their cut ends by the formation of continuous membrane barriers that allow for the restoration of function in the axonal stumps. We have used several electrophysiological measures (membrane potential, input resistance, injury current density) and several morphological measures (phase- contrast, video-enhanced differential interference contrast, light, and electron microscopies) of living and fixed material to assess the extent and mechanism of sealing within hours after transecting giant axons of squid (Loligo pealei and Sepioteuthis lessoniana) and earthworms (Lumbricus terrestris). Our electrophysiological data suggest that the proximal and distal ends of transected squid giant axons do not completely seal within 2.5 hr in physiological saline. In contrast, the same set of measures suggest that proximal and distal ends of transected earthworm giant axons seal within 1 hr in physiological saline. Our morphological data show that the cut ends of both squid and earthworm axons constrict, but that a 20–70-microns- diameter opening always remains at the cut end that is filled with vesicles. Axonal transection induces the formation of vesicles that are observed in the axoplasm within minutes in standard salines and that rapidly migrate to the cut ends. These injury-induced vesicles are loosely packed near the cut ends of squid giant axons, which do not functionally seal within 2.5 hr of transection. In contrast, vesicles formed a tightly packed plug at the cut ends of earthworm medial giant axons, which do functionally seal within 1 hr of transection in physiological saline. Since we detect no single continuous membrane that spans the cut end, sealing does not appear to occur by the fusion of constricted axolemmal membrane or the formation of a membranous partition at the cut end. Rather, our data are consistent with the hypothesis that a tightly packed vesicular plug is responsible for sealing of earthworm giant axons.