One eye was chosen randomly to receive timolol therapy twice daily; the fellow eye received placebo (timolol vehicle). The primary end point of the study was reproducible visual field loss detected on three consecutive tests. Automated static threshold visual fields were added to the protocol as the study proceeded, and criteria for reproducible defects for the automated fields were developed. The secondary end point was progressive optic disc cupping confirmed by examination of stereoscopic disc photographs. Intraocular pressure was not used as an end point (i.e., eyes were not withdrawn from the study because they reached a predetermined level of intraocular pressure).