704 Professor Sir James Dewar [Jan. 21, 



If one film be broken, the remainder rearrange their positions, 

 those nearest the disturbance l)eino: most affected. 



Examination of tlie manner of the fall of the films in the tube 

 shows that the rate is approximately steady through the o-reater part 

 of the tube, when not too many films are present. As the films 

 become more crowded in the lower part, the rate of fall decreases. 

 If the tube is not strictly cylindrical, then, provided the narrower 

 part is above, the films may either remain almost steady, or if the 

 difference of diameter is sufficient, the films will steadily rise ; they 

 can thus indicate differences from uniformity of section, too small to 

 be perceptible to the eye. In a similar way small irregularities or 

 slight unevenness of surface are indicated l)y a proportionate distor- 

 tion in the film itself from a perfect plane, because the film sets 

 perpendicular to the supporting surface. Variations of temperature 

 are indicated by an interruption to the smooth progress of the films 

 through the tube. Thus, in a quiet room, where the temperature is 

 falling, the mass of liquid will cool more slowly than the thin glass 

 walls and supporting films ; and a slight evolution of w^ater vapour 

 rising from the liquid will drive the films up the tube. Such 

 variations are well shown in a graph of continuous records of such 

 a set of films. (Fig. 5.) 



The draining of the liquid from the tube w'alls is well exhibited 

 by the steady decrease of thickness of the Hquid rings of contact 

 around each film. The interactions, in these films, of the surface 

 tension and other forces, such as gravity, have been very 

 thoroughly investigated by the American mathematician, Willard 

 Gibbs, and this ring of contact is therefore appropriately termed 

 the Gibbs ring or layer. It is seen that the films remain all 

 black, i.e. of uniform thickness, w'hile the various Gibb layers are 

 variable according to the excess of liquid on the neighbouring glass 

 surface. The amount of liquid they contain will vary from two or 

 three times that present in the film up to a considerable amount. 

 This relative amount can also be varied by small temperature 

 variations imposed on the Gibb layer, which can cause a stream of 

 liquid to pass out therefrom into the film which thereby becomes 

 thickened. A similar result is obtained by differences of vapour 

 pressure on the two sides of a film caused by small alterations of 

 temperature, condensation on the film then taking place. 



A special form of exhausted vessel constructed for the purpose of 

 keeping the film stationary is shown in Fig. 6. It has two equal bulbs 

 A and B with a slight, but rather sharp constriction between them, 

 forming a neck on which the film is thrown. The two bulbs are 

 drawn out on opposite sides of the neck to a tubular bend to form 

 an enclosed space freely open to each side of the film, and the 

 equality of pressure thus secured prevents the film from being 

 disturbed by variations of vapour tension with temperature. A 

 small bulbed reservoir is sealed in the centre of the tubular bend 



