SUMMARY1. When light scattering was measured during hyperpolarizing and depolarizing voltage-clamp steps, relatively large scattering changes were found during the depolarizing steps. These large changes were found to depend on the time integral of the ionic current and not on the changes in conductance or potential.2. The current-dependent changes were examined at several scattering angles, and three distinct time courses were found. At 30-120, the main change occurred after the current when steps of 2-5 msec duration were used. This change was called I-90. At 15-30°, the change occurred with the same time course as the time integral of the current. This change was called I-25o. At 5-15o the scattering change occurred with a time course intermediate between that of I-90 and I-25o. This change was called I-10o.3. In all experiments, outward potassium and outward sodium currents led to similar light scattering changes indicating that specific effects of the cation carrying the current across the membrane were not involved.4. The size of I-900 was reduced by 29 % when an isethionate artificial sea water was substituted for the normal chloride artificial sea water. This reduction equalled the reduction predicted for a transport number effect at the membrane-solution interface. The time course of I-900 was similar to the predicted time course for a volume change in the periaxonal space, and such volume changes were tentatively identified as the origin of I-90.5. Because of difficulties in measuring the time course of I-25', it was not possible to distinguish between a water of hydration effect and a transport number effect as the cause of this change. Similarly, the origins 728 L. B. COHEN, R. D. KEYNES AND D. LANDOWNE of I-1O' were not identified. Only I-1IO was altered in size and time course when the external refractive index was increased with bovine albumin.6. When the scattering changes during the action potential were examined in light of the voltage-clamp experiments, we concluded that the forward-angle change was potential-dependent and that the long-lasting change at right angles probably represented a swelling of the periaxonal space resulting from the fact that chloride carried a significant fraction of the outward current during the action potential.