The behaviour of water under hydrostatic tension: II

A study is made of the behaviour of water under hydrostatic tension. A review of the literature reveals very considerable differences in the values of critical tension obtained by different methods, the factor of disagreement being sometimes as high as 50. Two methods are studied in detail: the Berthelot method, which makes use of the difference in expansions of glass and water and the Reynolds centrifugal method. Evidence is brought forward that it requires high pressures to persuade water to stick to glass under subsequent tension. A modification of the Berthelot method gives results in much better agreement with other methods. Examination of the Reynolds method shows that account must be taken of the fact that water does not move as a rigid body. Although this has no appreciable effect on the pressure distribution, it does imply a considerable stirring of the water, and any small bubbles would tend to be brought near the region of greatest tension in a time that might be comparable with that of the experiment. The flow of water in constricted tubes is also studied. The critical tension appears to be low, but this is to be expected in view of the fact that the motion is probably turbulent before cavitation occurs. It is concluded that if tension is applied statically, ordinary water can stand tensions of the order of 40 atmospheres, even if it is not perfectly air-free. Water nearly saturated with air has been shown to stand tension up to 6 atmospheres.