This study assessed the inter-ocular and inter-session variability of the transient pattern electroretinogram (PERG) in a group of non-human primates. The transient PERG was measured both eyes of 29 non-human primates, and again after three months in 23 eyes of 23 of these animals. Signals were elicited using a contrast (90%, 75 cdm(-2)) reversing (5 reversals sec(-1)) checkerboard pattern (0.56 cpd). PERGs were also measured for stimuli of varied spatial frequency (n=8, 0.07-2.22 cpd), contrast (n=4, 20-100%), mean luminance (n=4, 4.7-75 cdm(-2)) and defocus (n=5, +1, +2, +3 diopters). The inter-eye and inter-session limits-of-agreement (LOA; 95%) were determined for each PERG parameter. Variability was also compared with previous studies using the coefficient-of-variability (COV). Pharmacological blockade of the inner retinal contributions to the PERG measured under these conditions was conducted in one animal using intravitreal injection of tetrodotoxin (approximately 6 microM) and N-methyl-D-aspartic acid (approximately 6 microM). The N95 component of the primate transient PERG showed spatial tuning, with a peak between 0.14 and 0.28cpd. This spatial tuning was not as apparent for the P50 component. A linear relationship between P50 and N95 amplitude was found with contrast and mean luminance. Both components were attenuated with the introduction of +2 diopters or more of defocus. The inter-session COV for the P50 and N95 components were 23.8 and 19.2%, respectively, while the LOA were 58 and 46%, respectively. The N95:P50 ratio had smaller inter-session variability, was robust to changes in contrast, mean luminance and defocus, and was effective for characterization of inner-retinal dysfunction after pharmacologic block.