348 KEPORT— 1888. 



cautions had to be taken against thermo-electric disturbance. Errors arising from 

 this aource were generally eliminated by repeating the observations with a reversed 

 current. In some cases the battery circuit was completed after the index of the 

 galvanometer had come to rest xinder the action of the thermo-current ; but this 

 was not considered so decisive as the other method. 



A few experiments were made by a method depending on the fall of potential 

 in one half of the tube compared with that in the other half, but they yielded the 

 same negative result. 



This conclusion is true for nearly aU our experiments to 1 part in 500 and for 

 many of them to 1 pvrt in 1,000. The conclusion is strengthened by the fact that 

 when small changes did occur they were shown to be due to some influence not 

 dependent on the direction of the curcent. Thus in one experiment already 

 alluded to, in which the current continued for twenty-four hours, a change of 

 zero amounting to 07 of a millimetre was observed. As this was a fairly large 

 change compared with the usual result, tlie current was reversed for twelve hours. 

 At the end of that time the zero had moved a little further in the smne direction, 

 showing that the alteration was due to some other cause than the one we were 

 seeking to detect. 



One or two data obtained incidentallj' may he of interest. 



The aUoy had a specific resistance at 100° C of about 107,000 c.g.s. units — i.e., 

 a cubic centimetre of it had between opposite faces a resistance of 'OOOIO? ohm. 



To assure ourselves that changes in composition would affect the resistance 

 two experiments were made. To the alloy in one half of the U tube 0-5 per cent, 

 of lead was added. This was done witli some care, but it coidd hardly be ex- 

 pected that the distribution of the lead would be uniform through the one half and 

 confined to it. However, a fall in the ratio of the resistances of the two limbs of 

 nearly 3 per cent, was noticed. In the second experiment the amount of lead 

 added was 1 per cent., and the fall in resistance of that half of the tube was about 

 6 per cent. 



This negative result is in accordance with that arrived at by Professor Roberts- 

 Austen with other alloj's and by a different method. 



Solid Sulphides. — It is well known that certain mineral sulphides conduct 

 electricity, and it is evident that if such a compound were placed between two 

 sheets of silver, and an electric current passed through, the condition of the silver 

 would at once reveal whether any sulphur had combined with it on the one side^ 

 or any metal bad been deposited on the other. Experiments made some years ago 

 gave a negative result, but they have been repeated more carefully recently. The 

 plates of silver were connected with an astatic galvanometer and a battery, and the 

 current varied in different experiments from a very small value up to about 1 '5 ampere. 

 The condition of the sulphide was also varied. Sometimes we used small 

 blocks of native galena, cinnabar, &c., ground smooth on two opposite faces ; in 

 other experiments we employed finely divided powders. 



The sulphides generally produced a slight tarnishing of the plates, but it was 

 found that something similar occurred without the current. In the case of three 

 sulphides — lead, mercury, and copper — however, the tarnishing was extremely 

 small compared with what would have occurred if the whole current had passed 

 electrolytically. We concluded, therefore, that there was no electrolytic conduc- 

 tion, or, if any, very little in comparison with the non-electrolytic. 



During the course of these experiments it came to our knowledge that Hittorf, 

 in 1851, had experimented on silver sulphide fused into solid cylinders, and showed 

 that the sulphide is electrolysed with the formation of threads of metallic silver. 

 This we confirmed, working with tightly compressed powder ; and we observed 

 that the current was at first small, but rose gradually, and then jumped to a mucb 

 greater value, no doubt at the moment when the space between the poles was 

 bridged over by the metallic threads. Under the microscope the reduced silver 

 seemed to be fused. Hittorf observed a similar decomposition of cuprous sulphide. 

 The following are the additional results we have arrived at : — 

 Lead Sulphide. — This compound, whether in the form of powdered galena or the 

 precipitated sulphide, conducts pretty easily- It always gave a slight indication of 



