Reactive synaptogenesis and terminal proliferation are known to occur in the dentate gyrus of the rat hippocampus following removal of specific afferents. In the present study we have examined the relation of synaptophysin immunoreactivity to the immunohistochemical staining pattern of GAP-431 1 Abbreviations used: COD, corrected optical density; CTA, crossed temporoammonic pathway; GAP-43, growth-associated protein 43; IML, inner molecular layer; MML, middle molecular layer; OML, outer molecular layer., a putative marker of neuritic growth. Within the molecular layer of the normal dentate gyrus, synaptophysin immunolabeling shows a trilaminar pattern, with the inner and outer layers having the greatest density of staining. Within the first week following denervation, there was a significant decrease in the staining density in the outer two-thirds of the molecular layer, followed by a moderate recovery at 14 days and 80% recovery by 30 days. This pattern is consistent with the time course of denervation and reinnervation in this system as determined previously by electron microscopy. By comparison, the staining pattern for GAP-43 in the intact dentate gyrus showed the middle and outer thirds of the molecular layer to be less densely stained than the inner third. Within a week following deafferentation, the outer twothirds of the molecular layer displayed decreased levels of GAP-43 immunoreactivity, followed by recovery to normal levels by 30 days. By 84 days postlesion, patterns of both synaptophysin and GAP-43 immunostaining reflected an increased width of the inner molecular layer. Laser confocal imaging of double-immunolabeled sections at 14 days postlesion showed a 370% increase in the number of GAP-43-positive terminals in the molecular layer as compared to unoperated controls. Many of these GAP 43-positive terminals were synaptophysin negative. We conclude that GAP-43 may play a role in the synaptic remodeling that occurs in the denervated rat hippocampus and that quantitative morphometry of synaptophysin immunolabeling accurately reflects the fate of presynaptic terminals in this model of degeneration and reinnervation.