We describe the first investigation into the effect on stimulation efficiency of varying the output of a commercial magnetic stimulator based on our original clinical design. Over the range of magnetic field waveforms considered, it is shown that the stored energy required to achieve stimulation, both cortically and in the periphery, varies by approximately 2:1. Greater efficiency is obtained by using shorter risetime magnetic fields. This results in more effective stimuli for the same stored energy, or, for the same stimulus, a decrease in energy storage, power dissipation and peak currents, thus simplifying hardware design. A novel method of processing the data obtained from different waveforms is presented which enables neural membrane time constant to be calculated. Data from normal subjects is presented showing both peripheral and neural time constants to be of order 150 microseconds. The cortical measurements represent the first non-invasive determination of cortical membrane time constant in man. Time constant measurements using magnetic stimulation may be clinically useful because they give information concerning the electrical properties of the nervous system not available from present techniques. Finally a method of quantifying the output of magnetic stimulators and coils is described which enables laboratory comparisons to be made, and takes into account magnetic field waveforms and coil geometry. The proposed symbol for this new measurement is Et150 with units volt seconds/meter.