We have approached the problem of nerve excitability through three questions: (a) What is the diagram for a channel? That is, what conformational states can the protein assume, and what transitions between these conformations are permitted? (b) What is the channel conductance associated with each conformation the channel can assume? (c) How do the rates for conformational transition depend upon membrane potential? These three questions arise from a standard statistical mechanical treatment of a nerve membrane containing several classes of identical, independent channels. Gating of channels, in this view, is associated with conformational changes of the channel protein, and it is assumed these conformations are distinct. The precise formulation of these questions is presented in terms of the theoretical treatment, and the approaches we have taken to answer the questions are indicated. Our present results indicate: Transition rates should depend exponentially on membrane potential over a limited voltage range, but probably will show a more complex dependence for extremes of the range; channels probably can take on only two conductances, open and shut, but more complicated situations are not entirely excluded; the diagram for a channel cannot be determined from standard voltage clamp data alone, but by studying gating currents and conductance fluctuations, it should be possible to select between alternative plausible physical mechanisms.