1. Two methods have been used to determine whether differences in gas concentrations between adjacent regions of tissue can induce osmosis.
2. In a steady-state experiment water has been shown to move in the direction of increasing gas concentration across gross sections of excised tissues such as bladder and peritoneum.
3. In a second experiment it was found that more water was retained in subcutaneous pockets of saline saturated with a soluble gas (nitrous oxide or ethylene) than was retained in a control pocket saturated with nitrogen and simultaneously monitored in the same rabbit.
4. A value of the reflexion coefficient for nitrous oxide has been estimated from the results of the experiments with the gross tissue sections and shown to be compatible with those known for non-gaseous solutes by extrapolation with respect to molecular size.
5. The significance of possible osmotic effects due to transient gas concentration gradients are discussed in connection with dry joints, aseptic bone necrosis in caisson workers and inert gas narcosis, the approach offering a quantitative correlation of narcotic potency by a simple physical mechanism.
6. The steady-state gas concentration gradients which arise in tissue as a result of metabolism are suggested as a third driving force in homeostasis but, unless larger reflexion coefficients can be demonstrated, this effect would only appear to be significant for inspired oxygen partial pressures appreciably greater than normal.