The permeation of Na + through gramicidin A channels shows a simple saturation with increasing Na + concentration that can be described by two different models. The first model assumes that one Na + binds to the channel with high affinity ( ~-30 M -1 ) and that conduction occurs by a 'knock-on" mechanism requiring double occupancy of the channel; the other model assumes that Na + binding is of low a|finity (< I M-1)) and that double occupancy of the channel is rare. NMR measurements have shown tight Na + binding, favoring the first model, but measurements of flux ratios and water transport support the second model. We present here a relatively model-independent measurement of the dwell time of Na + inside the channel, in which we characterize the fluctuations in H + current through the channel induced by 'block' from the more slowly permeating Na + ions. The mean Na + dwell time inside the channel is estimRted to be ~-10 ns at a membrane potential of 200 inV. This result is inconsistent with tight Na + binding, thus favoring the second model.