390 



NATURE 



[August 23, 1883 



specific resistance of the liquid, when drawn out into so 

 thin a film, is the same as that determined under ordinary 

 conditions. 



We have, by direct experiment, proved that this 

 assumption is true for films the thickness of which 

 exceeds 374 X io" 8 mm. (Philosophical Transactions, 

 1881, p. 447). The investigation was considerably pro- 

 tracted by the great difficulty experienced in maintaining 

 the constitution of the films even approximately constant. 

 Every change in temperature, every slight alteration in 

 the hygrometric state of the air in the glass chamber in 

 which the bubbles were formed, involved a loss or gain 

 of water which affected the specific resistance so largely 

 as to make any certain conclusion impossible. It is only 

 in our latest apparatus that we have secured the requisite 

 constancy in the conditions. In it the films are formed 

 in a chamber surrounded by water to keep the tempera- 

 ture constant. The base of the inclosed space is covered 

 by the solution used, and the complete saturation of the 

 air is further secured by an endless band of linen passing 

 over rollers which can be worked from the outside. The 

 lower roller is immersed in the solution employed, and 

 thus every part of the linen can in turn be dipped into the 

 liquid and kept completely saturated without opening the 

 case. The films are blown as spherical bubbles with air 

 which has been caused to pass over some of the liquid in 

 order to insure saturation ; they are converted from 

 spheres into cylinders adhering to two rings, and are 

 further put in communication at any desired point with 

 the electrical apparatus without opening the case, and 

 thus without affecting the temperature or saturation of 

 the air with which they are in contact. A thermometer 

 and a hair-hygrometer, placed in the closed chamber, 

 serve to detect any change of conditions which these pre- 

 cautions fail to obviate. 



The earlier form of apparatus described in our paper 

 "On the Electrical Resistance of Thin Liquid Films" 

 (Phil. Trans., local.), was in some respects less perfect. 

 By it, however, we were able to show that the specific 

 resistance of a film differed less and less from that of the 

 liquid from which it was formed, as the temperature and 

 hygrometric state of the air become more and more 

 constant, and that in the case of the six films in which 

 the desired constancy had been most successfully attained, 

 the difference amounted only to 1 '8 per cent. 



It was also shown that there was no indication of any 

 change in the specific resistance between thicknesses 

 corresponding to the middle of the red of the sixth and 

 of the yellow of the second order respectively. As the 

 smaller of these thicknesses is nearly the same as the 

 wave-length of the rays which bound the spectrum at the 

 blue end, this result may be roughly stated as proving 

 that the thickness of a film may be reduced to the length 

 of the shortest visible light wave without any change in 

 the specific electrical resistance of the liquid of which it is 

 composed. 



In the course of some of our earlier experiments (Proc. 

 Roy. Sac, 1877, p. 334) we had been fortunate enough to 

 make a soap solution, giving very persistent films, which 

 frequently thinned to the black of the first order. The 

 resistance of the black portion was measured on several 

 occasions, and it was found that the thickness was in all 

 cases nearly the same (the variations amounted to about 

 5 per cent.), and differed but little (if the specific resist- 

 ance was assumed equal to that of the liquid in mass) 

 from 12 millionths of a millimetre (12 X icr" mm.). 



We were anxious to try this experiment again with our 

 improved apparatus and methods of measurement, but 

 great difficulty was experienced in obtaining a liquid 

 which would both thin and last sufficiently for our pur- 

 pose. We have not, in fact, succeeded in again making 

 a solution, containing the proportion of glycerine recom- 

 mended by M. Plateau, which would behave in the desired 

 way, but we find that a liquid of similar constitution, in 



which the glycerine is replaced by water, will allow a 

 measurement of the resistance of the black to be made 

 in the case of about one film out of every three or 

 four. 



Films which do not contain glycerine generally 

 exhibit greater irregularities of behaviour than those 

 which do, and thus our later experiments are not in 

 as close agreement as the earlier ones. They indicate 

 that, whereas the thickness of the black portion of a film 

 remains constant however much its area may alter, it is 

 different in different films. All the values obtained lay 

 between 145 Xicr 6 , and 7'2X io~ 6 mm., and the mean 

 value 1 1 7 X 1 o- 6 differed only by two ten millionths of a 

 millimetre (2X io~ 7 mm.) from our previous result. 



In spite of this close agreement these results were open 

 to criticism. It is a long way, in terms of molecular 

 magnitudes, from the yellow of the second to the black of 

 the first order. We had no right to argue from results 

 on the specific resistance at the greater thickness to its 

 constancy at the less. It was, therefore, very important 

 to attempt to check our observations by some independent 

 method. 



We had often observed that plane circular films formed 

 in a glass tube thinned very readily to the black. This 

 was perhaps due to the fact that the small aggregation of 

 liquid all round the film affords a channel by means of 

 which the liquid can readily escape. However this may 

 be, it occurred to us that, though it was probably impos- 

 sible to measure the thickness of a single black film by 

 any optical method, it might nevertheless be possible to 

 determine the total thickness of a number of parallel films 

 in a tube. This we have succeeded in doing. The tube 

 and its contents were placed on an apparatus for pro- 

 ducing interference by thick plates. One of the inter- 

 fering rays passed through the tube. A few steel sewing 

 needles were included within it. When the films became 

 black, a number of them were broken by moving the 

 needles with a magnet, and the thickness could be calcu- 

 lated by observing the positions of the interference fringes 

 before and after the rupture. By this method the mean 

 thickness of the films was measured, on the assumption 

 that the refractive index of the films was the same as 

 that of the liquid in mass. Various considerations led to 

 the conclusion that this was probably correct, but in any 

 case the complete independence of the electrical and 

 optical methods made each a valuable check on the other, 

 though — if the fundamental assumption was correct — the 

 former was by far the more accurate. 



The result showed the two methods in most satisfactory 

 accord. The mean of all the electrical observations gave 

 a thickness of H'Sxio" 6 mm., that of all the optical 

 ii"4Xio -6 mm., an agreement which places it beyond 

 doubt that the mean value for all the films observed was 

 really about 1 1 6x IO" 6 mm. 



The methods employed then afford a definite measure of 

 thicknesses much smaller than the smallest that Newton's 

 scale of colours allows us to estimate. That scale 

 is very uncertain when colours of the first order are 

 employed. The difficulty or impossibility of obtaining 

 perfect contact between the lenses in the fundamental 

 experiment, and the possible distortion of their form in 

 the neighbourhood of the points of closest contact, make 

 colour estimates of thickness in the first order very 

 doubtful. The few observations we have made, on films 

 exhibiting the red and orange of the first order, show a dis- 

 cordance with Newton's results in striking contrast to the 

 agreement obtained in the case of most greater thick- 

 nesses. 



Our estimate of the thickness of the middle ot the red 

 of the first order (284X io" 6 mm.) differs from Newton's 

 by 20 per cent. In the blue of the second order our own 

 observations on Newton's rings differ from those on the 

 soap films by 6 per cent., and we were obliged, when aim- 

 ing at an accuracy of 1 per cent., to discard all observa- 



