The role of electroosmosis was studied directly in Nitella. The cells were mounted in a water-tight barrier between two chambers containing reversible electrodes for the application of potentials, and fitted with calibrated capillaries to measure water movement. No water movement was found when small existing bioelectric potentials were short-circuited through an external connection, nor when external potentials up to 1 or 2 volts were applied (producing currents up to 5 microa). Higher potentials (up to 10 volts) caused small movements of water, toward the negative pole. Larger and often irreversible water movements were produced by potentials up to 20 volts-sometimes persisting after current flow. A variety of evidence suggests that the effects are caused by injury at the cathodal end of the cell, allowing water to be attracted osmotically at the intact end and forced out at the injured end (transosmosis). This injury is reversible under small applied potentials, irreversible after large ones (100 to 200 times the natural bioelectric values). Such water flows persist in low salt concentrations (up to 0.09 M NaCl) but almost completely vanish in isotonic (0.26 M) mannitol. This confirms the osmotic, rather than the electroosmotic nature of the water movement. It is estimated that electroosmosis cannot account for more than 1 per cent of the water movement (or turgor) in Nitella cells. The dead cellulose walls display a small electroosmotic water flow at very high current densities (under 20 volts applied potential).