The main PERG component of the transient contrast reversal pattern electroretinogram (PERG) in cats was a negative wave (3.5 microV average, SD 1.7 microV) peaking at about 130 msec (N130) with a spatial resolution above 5.5 c/deg, close to behavioural estimates. The early positivity (P35) was more variable, smaller and had lower spatial resolution. Different anaesthetic protocols affected both the waveform and the amplitude by spatial frequency functions. Responses of urethane anaesthetised cats were like those reported previously for decerebrate cats or cats paralysed and ventilated with N2O/O2/CO2 (75%/24%/1%). P35 was evoked only by coarse stimuli and N130 amplitude decreased linearly as spatial frequency increased. When the luminance response amplitude, predicted from the optical transfer function of the eye, was subtracted, spatial tuning appeared. An anaesthetic mixture of ketamine hydrochloride and xylazine depressed both P35 and N130 at low spatial frequencies while enhancing them at high frequencies. In paralysed animals ventilated with N2O/O2 (67%/33%) P35 was larger and recordable to 1.6 c/deg. Peak times were reduced and the inter-peak time halved. Other anaesthetics depressed the ERGs. These effects suggest that cats are a good model for studying N130 in isolation or its interaction with P35 and that both PERG peaks include luminance and pattern components.