1894.J On the Electrical Properties of Pure Substances. 33 



This method consists in observing the galvanometer indication of a 

 carrent passing through the substance of high resistance under a 

 known voltage, and subsequently causing the galvanometer to give 

 the same deflection, by supplying it with a known fraction of the 

 voltage of a Clark cell, and allowing this to act on the galvanometer 

 when the latter is in series with a wire megohm standard. The dis- 

 cussion of this method, depending as it does on the behaviour of 

 Clark cells from which currents are being taken, shows that it is 

 reliable ; but it is not intended here to go over the preliminary 

 ground covered by the papers referred to. A considerable portion of 

 the investigation of which the paper is an account, and which 

 extends from 1889 to the present time (October, 1893), was made 

 conjointly with Mr. Pollock. 



The first part of the paper deals with the purification of sulphur 

 as obtained from several sources, with the result that, in the end, 

 the following method was exclusively adopted. This method is 

 based on the use of sulphur recovered by the Chance process, which 

 comes into commerce as pure to at least one part in ten thousand, 

 and results from burning hydrogen sulphide from alkali waste with 

 insufficient air for complete combustion. The commercial product is 

 melted, and filtered through glass wool and platinum gauze. It is 

 then twice distilled, in such a manner as to be free from exposure to 

 dust : and sometimes it was subsequently freed from gas, by heating 

 in a vacuum to near the boiling point. The purity of the resulting 

 sulphur is tested by the following criteria. It must be free from 

 smell. It must leave no residue on evaporation from a platinum 

 dish. When cooled suddenly from a high temperature, it must 

 remain of a clear yellow colour : when perfectly crystalline it must 

 be absolutely soluble in carbon bisulphide. The absence of arsenic 

 and selenium from the sulphur employed, was proved to about one 

 part in a million by burning the sulphur to trioxide and applying the 

 appropriate tests. The reaction by sulphur dioxide test for selen- 

 ium, when properly carried out, is more delicate than the Marsh's 

 test for arsenic, even when the smell of the hydrogen is adopted as a 

 criterion. If a perceptible mirror of arsenic is to be accepted as a 

 criterion, the arsenic test is still less delicate. 



A number of experiments are described, tending to show that 

 neither arsenic nor selenium can possibly exist to any appreciable 

 extent in alkali waste produced in the Leblanc process, so that the 

 Chance sulphur is probably more free from these impurities than the 

 limit we can reach by analysis. All other known impurities are got 

 rid of by distillation and exhaustion in vacuo. 



Section 3 of the paper deals with a discussion of various methods 

 of measuring high resistances, and gives the detail of the method 

 adopted by us for rapidly effecting alternate measurements of 



VOL. LYI. D 



