A review of the evidence for believing that gas concentration gradients can induce water movement in vivo shows that this could be a significant physiological parameter in many tissues, depending upon the magnitude of the gas-osmotic pressures. Estimation of these has suggested two experiments whose results show (1) that reflexion coefficients for gases may be appreciably higher than determined previously, (2) that water flow rates are high, and (3) that the list of semipermeable tissues should include articular cartilage. The clinical implications of gas-induced osmosis are then discussed with particular reference to gaseous anesthesia, hyperbaric arthralgia, gouty arthritis, aseptic bone necrosis, hyperbaric urticaria, pulmonary edema, and the edematous separation of tissue boundaries in general.