A quantitative re-investigation of the time course of the initial rise of the potassium current in voltage-clamped squid giant axons is described. The n4 law of the Hodgkin-Huxley equations was found to be well obeyed only for the smallest test pulses, and for larger ones a good fit of the inflected rise required use of the expression (1-exp[-t/tau n1])X-1(1-exp[-t/tau n2]), where both of the time constants and the power X varied with the size of the test pulse. Application of a negative prepulse produced a delay in the rise resulting mainly from an increase of X from a value of about 3 at -70 mV to 8 at -250 mV, while tau n1 remained constant and tau n2 was nearly doubled. The process responsible for generating this delay was switched on with a time constant of 8 ms at 4 degrees C, which fell to about 1 ms at 15 degrees C. Analysis of the inward tail currents at the end of a voltage-clamp pulse showed that there was a substantial external accumulation of potassium owing to the restriction of its diffusion out of the Schwann cell space, which, when duly allowed for, roughly doubled the calculated value of the potassium conductance. Computations suggested that the principal effect of such a build-up of [K]o would be to reduce the fitted values of tau n1 and tau n2 to two-thirds or even half their true sizes, while the power X would generally be little changed; but it would not affect the necessity to introduce a second time constant, nor would it invalidate our findings on the effect of negative prepulses.