*Movements of Aniline Globules on a Water Surface.*—If we allow a small quantity of aniline to run on to the surface of water, it forms itself into a number of floating globules. I now project on the screen a water surface on which a little aniline has been poured, and we are thus enabled to watch the movements which occur. All the globules appear to be twitching or shuddering; and if you observe closely you will notice the surface of each globule stretching and recoiling alternately. The recoil is accompanied by the projection of tiny globules from the rim, which becomes scalloped when the globule is stretched. The small globules thrown off appear to be formed from the protuberances at the edge (Fig. 36), and after leaving the main globule they spread out over the surface, or dissolve. This process continues for a long time, gradually diminishing in vigour, until small stationary globules are left floating on the surface, which is now covered with a skin of aniline. This action is in striking contrast to the tranquil formation of floating globules of oil, and calls for some special comment. Let us recall again the three forces at work at the edge of a floating globule (Fig. 35). The surface tension of the water, acting horizontally, tends to stretch the globule, and is successful momentarily in overcoming the opposing tensions, each of which pulls at an angle to the surface. Enlargement of the upper surface of the globule, however, reduces the angles at which the tensions B and C act, and in consequence their effective strength is increased. The spreading of the aniline over the water surface diminishes the pull A, which B and C combined now overcome, and hence the surface of the globule shrinks again. For some unexplained reason both the stretching and recoil of the globule occur suddenly, there being an interval of repose between each, and these jerky movements result in small portions of the rim being detached, each of which forms a separate small globule. The aniline which spreads over the surface of the water dissolves, and the water tension A, which had been enfeebled by the presence of the aniline skin, recovers its former strength, and again stretches the globule; and so the whole process is repeated. When the surface of the water becomes permanently covered with a skin, which occurs when the top layer is saturated with aniline, the globule remains at rest, and has such a shape that the tensions B and C act at angles which enable them just to balance the weakened pull of A. Why the edge of the globule becomes indented during the movements, and why these movements are spasmodic instead of gradual, has not been clearly made out. It is interesting to recall that a spheroid of liquid on a hot plate also possesses a scalloped edge, and it may be that the two phenomena have something in common. [Illustration: __Fig._ 37.—Orthotoluidine globules on a water surface._]