622 Mr. N. R. Campbell : Experiments on 



In the conclusions given below it is assumed that the 

 regular and consistent experiments give the normal result, 

 and that the irregularity o£ the others is due to some 

 unessential disturbance. The mutual agreement o£ the 

 former series affords a logical inductive proof of the correct- 

 ness of this assumption. 



Ko single sufficient cause for the inconsistencies can be 

 suggested, but some of them at least may be due to hydro- 

 carbons present as impurities, for the introduction of a mere 

 trace of coal-gas, ether, or benzoline was always sufficient 

 to produce them. All gases before passing into the apparatus 

 were bubbled through KMnO.^ ; this procedure minimised, 

 though it did not entirely prevent, the annoyance. There 

 seems, in these cases^ to be some deposit formed on the glass 

 of the apparatus, for, when the irregularities had once 

 appeared in a series of experiments, it was necessary to clean 

 all the glass with chromic acid before regular results could 

 be again attained. Even this was not sufficient after the 

 hydrocarbons had been introduced ; a new bulb had to be 

 blown from fresh glass. 



After these general observations, the results will be described 

 in greater detail. 



Air. — The air was drawn from outside the window of the 

 laboratory and purified by passing through a solution of 

 potassium permanganate and a plug of cotton-wool. 



Points of the following metals were employed : Pt, Au, 

 Ag, Cu, Ni, Al, Zn, Steel, German-Silver, Carbon, Mg, and 

 Copper coated with Sodium Amalgam. All these metals 

 except the last two gave the same result. Out of 34 experi- 

 ments, 19 were irregular ; the others gave : — 



Discharge from positive point — Fog with expansions for 



which — lay between 1*254 and 1"251. 



Discharge from negative point — Fog with expansions for 

 which — lay between 1*252 and 1*247. 



Twelve experiments made with the Mg and Sodium 

 Amalgam pomts were all quite irregular and inconsistent. 

 This will be referred to later. 



Ilydi'ogen. — Prepared from commercial zinc and HCl and 

 passed through a solution of KMnO^. At first great care 

 was taken to procure perfectly pure gas, but it was found 

 that if these precautions were omitted and the discharge 

 passed for five minutes before making observations, the same 

 results were obtained as with the pure gas. 



