With the multifocal technique, as developed by Erich Sutter and colleagues, scores of focal electroretinogram (ERG) responses can be obtained in a matter of minutes. Although this technique is relatively new, it has already provided insights into the mechanisms of retinal disease. However, because it is new, there also remain questions about how it works and what it measures. This chapter considers some of these insights and some of these questions. The first part (Section 2) describes how the multifocal ERG (mERG) is recorded and considers its relationship to the full-field ERG. The mERG responses are shown to be from relatively local regions of the retina and are comprised of the same components as the full-field ERG. The diagnostic advantage of the mERG as compared to the full-field ERG is also illustrated. In Section 3, the effects of damage to different cell layers of the retina are shown to affect the mERG differently, and these changes are summarized within a conceptual framework. It is argued, for example, that when diseases of the receptor outer segment, like retinitis pigmentosa, result in small, depressed mERG responses, then the damage is, as expected, at the outer segment. However, when these diseases result in mERG responses that are reasonably large but very delayed, then the damage is beyond the outer segment, probably in the outer plexiform layer. The implicit time of the mERG, not amplitude, is the more sensitive measure of damage in degenerative diseases of the receptors. On the other hand, diseases, like glaucoma, which act on the ganglion axon, do not result in easily identified changes to the mERG unless inner retinal damage is involved as well. Inner retinal damage changes the waveform of the mERG and decreases the naso-temporal variation normally observed. Finally, diseases, like diabetes, that act on more than one layer of the retina can have a range of effects. In Section 4, recent work with the monkey mERG is reviewed, with emphasis on the relevance to human diseases. For example, blocking the sodium-based action potentials produced by ganglion and amacrine cells eliminates the naso-temporal variation in the monkey mERG and these altered mERG responses resemble those from some patients with diabetes or glaucoma. Finally, in Section 5 the second-order kernel is described. The presence of a second-order kernel has important implications for understanding the shape of the mERG response (first-order kernel). Full-field simulations of the mERG paradigm illustrate that the first-order kernel is comprised of responses with different waveforms. Further, it is argued that the nonlinear, adaptive mechanisms that produce the second-order kernel are involved in shaping the time course of the response. Patients with large, but abnormally delayed mERG responses (first-order kernel), do not have a detectable second-order kernel. It is speculated that a markedly diminished second-order kernel is diagnostic of outer plexiform layer damage, not inner plexiform layer damage as is commonly assumed.