FOR some time it has been suspected that the most relevant parameter when using electric shock as a stimulus is the applied wattage. Forbes and Bernstein1, Gilmer2, Hill et al.3, Campbell and Teghtsoonian4, Green5, all found evidence to support this contention. Although constant current and constant voltage outputs have long been in use, a device capable of supplying a constant power output compensating for the changes in impedance found in any biological circuit has not hitherto been available.