*Production of Globules from Films.*—When a film of oil spreads over a water surface it sometimes remains as such indefinitely. Certain other liquids, however, form films which after a short interval break up into globules, and the process of transition is at once striking and beautiful. In order to show it, I project a water surface on the screen, and pour on to it a very small quantity of _dimethyl-aniline_—an oily liquid related to but distinct from ordinary aniline. It spreads out into a film of irregular outline, which floats quietly for a short time. Soon, however, indentations are formed at the edges, which penetrate the film, and from the sides of the indentations branches spread which in turn become branched; and shortly the whole film becomes ramified, resembling a mass of coral, or, to use a more homely illustration, a jig-saw puzzle (Fig. 38). The various branches join in numerous places, cutting off small islands from the film; and immediately each island becomes circular in outline—and the resolution into globules is complete. We have witnessed one of the beauty-sights of Nature. The same method of globule formation is shown by nitro-benzol and _quinoline_, and as the action is more gradual in the case of the latter substance, I show it in order that we may study the process in greater detail. Notice the formation of the indentations and their subsequent branching; and also that holes form in the skin from which branchings also proceed. In this instance the film is broken up in sections, but the action continues until nothing but globules remain on the surface.⁴ It is not easy to see why the canals of water penetrate the film and split it up into small sections, nor why entry takes place at certain points on the edge in preference to others. Some orderly interplay of forces, not yet properly understood, gives rise to the action; and a satisfactory explanation has yet to be given. ⁴ The breaking-up of films on the surface of water was first noticed by Tomlinson about 50 years ago. He used essential oils, and called the patterns “cohesion figures.”