The early detection of optic nerve damage is a prerequisite for effective therapy in primary open-angle glaucoma (Kass et al., 1990). An alarmingly large proportion of retinal ganglion cells may be damaged, however, before visual field defects are observed with conventional manual or automated perimetry (Quigley et al., 1982; Quigley et al., 1989). This implies that either there is widespread redundancy within the retina such that substantial cell death can occur without loss of visual performance or that our current clinical procedures do not test the function of the cells that are damaged in early glaucoma. Recent studies lead us to believe that the latter of these two hypotheses may be correct.