It is well-known that the collapsing motion of cavitation bubbles may cause various kinds of physical, chemical and mechanical actions, In our experiments we took up two actions, "sono-luminescence" and "cavitation erosion" and examined the behavior of cavitation bubbles through them. In our experiments on cavitation produced by a nickel transducer with a frequency of 5kHz we reached the same conclusion as A. S. Bebchuk, L. A. Chambers et al. (Figs. 3 and 6) that the intensity of sono-luminescence decreases monotonically with rise of temperature, while cavitation erosion has a maximum effect at a certain temperature of liquids. Thus we confirmed that these phenomena of cavitation actions are independent of experimental environment and so on. It has already been verified in our experiments that both sono-luminescence and cavitation erosion arise in the completely collapsing stage of cavitation bubbles irrespective of acoustic intensity. However, the reason why these two actions have such different inclinations from each other as to the temperature of liquids has not yet been clarified. We assume that the temperature dependence of cavitation actions is mainly attributed to a change in vapour pressure of liquids tested, because the vapours contained in collapsing bubbles have a "cushion effect" against their abrupt collapse. According to the experimental results the effect of cavitation erosion reaches its maximum at a certain temperature of liquids at a vapour pressure of about 0.1 bar, and with the rise of the temperature of liquids the intensity of sono-luminescence decreases. Using alchol as a specimen we confirmed the fact the higher the vapour pressure of liquids tested, the lower the intensity of sono-luminescence is (Fig. 5). With regard to the stability of collapsing spherical bubbles we examined theoretically the effect of pressure of gas contained in them. In the analysis of the stability we adopted a method of investigating whether slight perturbation given on the spherical surface of the bubbles would grow or damp as the bubbles contract. This method of analysis is conformed to that put forward by M. S. Plesset et al. Assuming that bubbles remain spherical till they contract to their minimum radius, W. Gtith already obtained the analytical values of maximum temperature and pressure reached in the bubbles (Figs. 7 and 8). When the initial pressure of gas contained in collapsing bubbles is low, the results obtained by him are not good. Because with an advance in contraction bubbles deviate from spherical shape greatly and his assumption does not hold good. In the final stage of collapse they may be divided into a lot of minute bubbles and so the maximum value of actual pressure in each bubble may be comparatively small. On the other hand, when the initial pressure of gas contained in collapsing bubbles is higher, their deviation from spherical shape is not so great and the value of actual gas pressure in the bubbles is nearly equal to the results obtained by Guth (Figs. 9 to 11). Thus we have made it plain from above-mentioned examination that, when in the initial stage of collapse cavitation bubbles have a suitable pressure of gas or vapour, the value of pressure in the collapsing bubbles reaches its maximum (Fig. 13) and so cavitation erosion has a maximum effect at a certain temperature of liquids, while with increase in the initial pressure in the collapsing bubbles, the temperature in it, that is, the intensity of sono-luminescence montonically decreases (Fig. 14).