The excitable membrane of nerve cells has a large number of ion channels, whose gating is usually voltage-dependent. The dynamic interaction between various types of ion channels plays a key role in the electrical activity of the cell. We present an approach to computing an approximation of the channel density of one ionic species along a dendritic segment from a single microelectrode stimulus, using measured voltage responses to a specified input current. We consider first a linear cable model, then the nonlinear Morris-Lecar model with spatially distributed p otassium channels. This is joint work with Jonathan Bell.