ON ELECTROLYSIS IN ITS PHTSICAL AND CHEMICAL BEARINGS. 347 



' Treatise on Electricity and Magnetism,' Vol. I., p. 309, speaks of ' the theory of 

 Clausius ' simply. 



I trust, therefore, that you will kindly arrange so that terms conveying a 

 meaning not intended by the Committee may be avoided by them, lest they 

 excite a false impression in the minds of readers. — With assurances, &c., faithfully 

 yours, 



E. Clatjsius. 



On the Conduction of Alloys and Solid SuljjMdes. 

 JBij J. H. Gladstone, Ph.D., F.B.S., and Walter Hibbert, F.I.G. 



The following experiments were instituted in connection with the work of the 

 British Association Committee on Electrolysis. One of the questions originally 

 raised was whether there was any resemblance between the conduction of an alloy 

 and that of an electrolyte ; or, to put it in anotlier way, whether the passage of an 

 electric current through an alloy is associated with any separation of its con- 

 stituents. 



The method adopted by us depends on the assumption that any such separation 

 would produce changes of resistance, the changes being, in ail probability, different 

 at one electrode from what they would be at the other. 





The apparatus consisted of a U tube of glass of the following dimensions : — 



Internal diameter, 0'45 centimetres ; total length, 28 centimetres ; horizontal 

 portion, 11 centimetres; each vertical portion, 8'5 centimetres. The ends of the 

 tube were carefully fitted with corks, so boi'ed as to form large terminal cups, as 

 shown in section in the above diagram. 



In the middle of the horizontal part of the U a platinum wire was fused 

 through the glass for the purpose of making connection with the central part of 

 the alloy whilst the cun-ent was passing through it. 



The alloy used was the ' fusible alloy,' sold by Messrs. Hopkin & Williams, 

 containing bismuth, lead, tin, and a little cadmium. It was chosen because of its 

 low fusing-point, which enabled us to keep it perfectly liquid at the temperature 

 of boiling water. A suitable quantity was fused and poured into the U tube, 

 some care being needed to drive out all air-bubbles. The tube was then placed in 

 a water oven and kept at 100° Centigrade. Connections were made through the 

 lid of the oven by means of three thick insidated copper wires, two of them going 

 to the cups at the end of the U tube, and the third to a mercury cup at the bottom 

 of the oven, into which there dipped the platinum wire from the middle of the U. 

 By means of these connections the two halves of the tube were joined to the ends 

 of a metre bridge wire, and the ratio of the resistance of one half to that of the 

 other obtained in the usual way. 



In the earher experiments a slight modification of Carey Foster's arrangement 

 of the bridge was tried, but it was found a little inconvenient for these particular 

 experiments, and was therefore replaced by the ordinary simpler form. 



The currents employed varied in different experiments from 5 down to 0*2 

 amperes, running generally for four or five hours, with two or three tests during 

 that time. On one occasion a current of about 0'4 ampere was continued for 

 twenty-four hours. 



_ The results of the experiments were always negative, no change of resistance 

 being detected which did not fall within the limit of probable error. 



It ought to be pointed out that the sources of error were considerably greater 

 than usual. The resistances under comparison were so small, and such great 

 temperature diff'erences existed between the various junctions, that great pre- 



