The isolated arterially perfused eye preparation has become a valuable tool in the investigation of ocular function. Normal retinal function can be maintained for several hours with the measurement of the gross electroretinogram (ERG) serving as a useful monitor of the electrophysiological condition of the preparation. This paper deals with the manner in which the amplitude of the ERG is affected by the electrode recording locations and describes the characteristic distribution of potentials on the surface of the isolated arterially perfused dog eye. The ERG is recorded between a fixed corneal contact lens electrode and a movable electrode on the uppermost surface of the sclera. The scleral electrode is moved in small increments parallel to the optic axis and lowered onto the uppermost surface of the sclera on a line from the corneal limbus to the optic nerve. The resulting ERG profile is characterised by minimal b-wave amplitude at the corneal limbus and little growth until a point 5 or 6 mm from the corneal limbus is reached, followed by a region of rapidly increasing amplitude up to a maximum at the point where the optic nerve exits from the globe. There follows a region of large but stable signal amplitude along the optic nerve. The ERG profile is largely unaffected by a reorientation of the globe about the optic axis, demonstrating that the potential distribution on the surface of the globe may be described in terms of points equidistant from the limbus being isopotential. A similar distribution of potentials is shown to exist with two types of retinal illumination, that from a simple flash lamp and that from a more uniform retinal stimulation provided by the incorporation of a ganzfeld sphere. The implications of this finding are discussed with reference to clinical and experimental electroretinography.