After settling down from the excitement of seeing single channels open and close in the membrane of a living cell for the first time (Neher & Sakmann, 1976), the question that arose was what could these extraordinary recordings tell us? At the time we had no idea how many different applications we would soon find for the patch clamp technique. Most of these applications followed from the subsequent improvement of the gigaseal, which dramatically increased the power of the patch clamp. But as a ‘first application’ for single channel recording, in an article in The Journal of Physiology, Neher & Steinbach (1978) turned to a question about the mechanism of action of local anaesthetics. It was known that local anaesthetics reduce the current through the membranes of excitable cells. The endplate current at the neuromuscular synapse becomes smaller in the presence of a local anaesthetic, and its decay follows a double-exponential time course. These results could be accounted for by a model in which a channel first opens, and a local anaesthetic molecule then binds to a site within the open channel to block it (Ruff, 1977). These drugs could thus be pictured as acting like a plug in the drain of a bathtub. This model fitted the data well, and most researchers regarded it as very appealing. Neher and Steinbach noted a number of studies that had invoked a sequential mechanism of channel blockade, but this mechanism was not generally accepted as rigorously proven. The key result of two exponentials in the decay of the synaptic current indicated that the drug must induce the appearance of a new kinetic state of the channel. From an analysis of macroscopic currents alone, it was very difficult to decide whether the new state was an open or closed channel, or whether entry into the new state occurs from an open or closed channel. Single channel recordings from a frog skeletal muscle fibre Patch electrode solution containing 0.1 μm suberyldicholine alone in A, together with 5 μm (B), 10 μm (C), and 50 μm (D) QX-222. (Reproduced from Fig. 2 of Neher & Steinbach, 1978.) QX-222 and QX314 are both charged, and if their binding site resides within the channel, then the transmembrane potential should influence binding. The single channel data permitted a direct determination of the blocking and unblocking rates (F and G), and these rates did indeed depend on voltage. The magnitude of this voltage dependence suggested that the drug traverses 78% of the membrane potential upon binding. This study established open channel block as an important mechanism of drug action. But the study had much greater significance in providing the first illustration of the remarkable power of single channel recording in the elucidation of kinetic mechanism. By observing the closed state, the open state and the blocked state directly, some very strong statements could be made about how these states interconvert. For any ion channel mechanism, the single-molecule transitions contain information not readily extracted from the macroscopic currents that report the activity of populations of channels. This work thus initiated a new era of mechanistic studies of ion channels. A rigorous mathematical formulation of single-channel kinetics provided the essential theoretical tools for interpreting complex lifetime distributions (Colquhoun & Hawkes, 1982). Single channel recordings revealed correlations in the lifetimes of adjacent gating intervals that added to the information content uniquely available in this form of data (Jackson et al. 1983), and again a theoretical framework served to guide these studies (Fredkin et al. 1985). The enduring contribution of this paper was its illustration of the extraordinary power of single-channel recording in model discrimination and the elucidation of kinetic mechanism. The many applications of single-channel kinetics that followed this work brought our understanding of ion channels to a new level of detail. These studies offer valuable lessons that should prove useful in the study of other forms of single-molecule kinetics, most notably single-molecule fluorescence.