474 Dr. W. F. G. Swann on the Electrical 



deposit during the 130 second period would account for the 

 departure of the corresponding point shown from the smooth 

 curve *, but such changes have no effect on conclusions 

 which we may draw with regard to the variation of the 

 resistance of any one film with temperature. For such 

 purposes the observations may be looked upon as accurate 

 to the extent of the agreement shown in columns 3 to 8 of 

 Table III. 



If the curves for -180° C. and 99°-8 C. are plotted, it 

 turns out that instead of the distance of the bends from the 

 origin of thickness being in the ratio of 8 "58 to unity as 

 would be required by the simple theory, the two curves are 

 practically coincident in this region, both with each other 

 and with the curve for 13 0, 9 C. In fact, such temperature 

 variations as the films show are, except in the case of very 

 thin films, extremely small. Further, the rate of increase of 

 the resistance at the critical thickness is enormously greater 

 than would be indicated by the simple theory, which in 

 virtue of the comparatively small variation of log X/t with t, 

 (equation 2), practically gives the curve relating to the 

 specific resistance and t as an hyperbola. The fact that 

 the experimental bend is sharper than the theoretical one 

 has been noted by Patterson, and it has been suggested 

 that it is due to a want of uniformity in the densities of 

 the free electrons at the surface and inside, but while 

 such a fact might be invoked to explain even a 100 per cent, 

 difference, it is difficult to see how it can be the expla- 

 nation of the enormous discrepancy found, a discrepancy 

 amounting to thousands of per cent, on the steep part of 

 the curve. 



We can obtain an explanation of the extremely sharp bend 

 if we assume that in the deposition of the film the molecules 

 come down in groups, and the phenomenon with which we 

 are concerned is not one of a film of a gradually increasing- 

 thickness, but in the early stages is somewhat more analogous 

 to the condition of affairs on a pavement when rain-drops 

 begin to fall, and before enough drops have fallen to cover 

 the pavement. Taking the results of Patterson, who measured 

 the thickness of some of his films, we find that the amount 



* In the experiments which were first performed with a different 

 apparatus, a film was deposited for, say, 50 seconds, and measured in situ. 

 Another film was then deposited on the top, and the measurements were 

 repeated, and so on. This method has the advantage of reducing con- 

 siderably the disadvantages arising from fluctuations in the coil, but 

 difficulties arose in carrying out the experiment in this way, which 

 rendered the method impracticable. 



