1903.] Heat and of Solvents on TJiin Films of Metal. 231 



In the first part of this paper it was shown that when the surface 

 of metals is subjected to various forms of mechanical treatment a 

 distinct covering film or layer is produced, which is in many respects 

 different from the mass of metal below. By the regulated action of 

 solvents this covering film can be wholly or partially removed. The 

 observations now to be described show that the behaviour of this 

 covering film under the action of heat annealing is in many respects 

 analogous to that of thin films supported on glass. 



For the study of these surface effects plates of pure gold were used. 

 These were planished by beating on a polished surface, and polished 

 by rubbing them lightly on rouged leather. The difficulty of obtaining 

 a scratchless, untorn surface in a soft metal like pure gold is very 

 great, but, fortunately for the purpose in view, a perfect surface was 

 not needed, indeed the presence of lines of flow was sometimes an 

 advantage. 



Fig. 19 shows the surface of the polished plate. This and the three fol- 

 lowing figures are direct photographs without enlargement, by normally 

 reflected light, with a 3 mm. oil immersion apochromat of 1-4 N.A. the 

 magnification is x 1500. 



Fig. 20 is the same plate after heating to between 700 and 800 

 for ] minutes. The surface film has become detached and broken 

 up, evidently by the escape of air which had been imprisoned during 

 the planishing and polishing operations. The surface has a 

 viscous appearance and the holes made by the escaping gas add to this 

 effect. 



Fig. 21 is the same plate after further heating for an hour. The 

 surface film appears to be aggregating and subsiding into the mass of 

 metal below. 



Fig. 22 is the same after further heating for an hour. In this case, 

 as exactly the same spot has been photographed as in the preceding 

 figure, it is possible to trace the further subsidence and disappearance 

 of the original surface film. 



Fig. 23 is another specimen of polished gold after heating for 

 1 hour. The ruffled surface again shows the effects of escaping gas. 



In Fig. 24, after a further hour's heating, the final subsidence and the 

 viscous flow of the surface are clearly seen. 



To complete the parallelism between the behaviour of thin films on 

 glass and surface films on massive metal, a plate of polished gold was 

 exposed for a short time to the action of mercury, which was then 

 driven off by heat. 



The exact nature of the change brought about by this treatment 

 was at first both puzzling and obscure. The greatly increased 

 absorption of light by the altered surface could not be accounted for 

 on the mere supposition that it was a scattering effect due to the 

 breaking-up of the surface, because there was evidently loss of light 



