The corneal discs of 41 patients with scarring reminiscent of herpetic infection were organ cultured for HSV isolation. Of the 41 patients, 34 had a definite history of herpetic keratitis, from 10 of whom (29.4%) HSV was isolated. There were no clinical features which distinguished between these groups; there was however an indication that those from whom HSV was not isolated had been previously treated with substantial amounts of topical acycloguanosine. In three patients of 12 patients when the disc was separated into 7 parts using a punch technique, virus was isolated exclusively from those portions demonstrating clinical scarring. Electron microscopy (EM) demonstrated HSV particles in stromal cells in the cultured corneas of seven patients. In two of the patients no virus was detected prior to culture with EM. In one patient HSV antigen was not found using peroxidase-antiperoxidase (PAP) staining prior to subsequently positive organ culture.Studies were made to determine how HSV accedes to the corneal stroma using a murine model in which keratitis occurs by zosteriform spread of HSV following inoculation of the snout. Preliminary evidence using PAP staining indicates that the virus reaches the stroma at the same time as the epithelium, via the sensory nerves. Evidence of HSV persistence in anterior segments was obtained in the same model, in contrast to which no virus could be isolated following direct inoculation into the cornea. It is speculated that for virus to set up a longterm association with the stromal keratocyte, it must be introduced via the sensory nerve. It is proposed that the original source of stromal virus lies in the stroma beneath the dendrite, and that spread of virus from this site occurs by intercellular spread, rather than by extracellular spread from the ulcerated epithelium.