638 Dr. W. Wilson on the Discharge of 



by simply dipping the platinum wire in distilled \*ater, the 

 discharge in the case of potassium sulphate was due mainly, 

 if not entirely, to water. The similar effect observed by 

 A. E. Garrett, I. c, in his experiments on hot aluminium 

 phosphate is possibly an increased activity of the platinum 

 on which he heated his salt, and occasioned by the com- 

 paratively large quantity of water which the salt contains. 



To obtain this increase in the positive discharge from 

 platinum, it is not necessary that there should be water 

 vapour present in the atmosphere surrounding the wire 

 during the measurement of the discharge. This is shown by 

 the following experiment : — The platinum wire was placed 

 in an atmosphere saturated with water vapour and heated for 

 some minutes by a current, the temperature being raised as 

 far as w r as possible without risking fusion of the wire. The 

 latter, after being allowed to cool in this atmosphere, was 

 then placed in the measuring apparatus (fig. 2), the air in 

 which was perfectly dry, and with the usual potential- 

 difference and temperature the leak was found to be greatly 

 increased. It would appear therefore that water causes or 

 accelerates the production in the platinum of something 

 which can emerge at sufficiently high temperatures in the 

 form of positive ions. In all cases the activity induced by 

 water decreased with great rapidity, which indicates that it 

 is confined to the outer portion of the wire. It should be 

 mentioned that a much greater increase in activity was pro- 

 duced by actually dipping the wire into water and introducing 

 the wet wire directly into the measuring apparatus. 



Another experiment, which showed the effect of water in a 

 very striking way, consisted in dipping the wire into a solution 

 of calcium nitrate. It was then heated in the measuring 

 apparatus (fig. 2), till the salt was reduced to calcium oxide 

 and the discharge was reduced to small dimensions. So long 

 as the ware w T ith its coating of oxide remained in the dry 

 atmosphere of the apparatus, the leak continued to be almost 

 inappreciable. A 5 minutes' exposure of the coated wire to 

 the air of the room, however, was sufficient to induce a leak 

 of 50 to 60 scale-divisions. Again the leak decayed with 

 great rapidity. 



Nature of the Ions in the Induced Discharge. — The carriers 

 of the positive electricity in the case of the normal leak from 

 hot platinum consist probably to a large extent of carbon 



monoxide. The value of — obtained by Richardson for the 



m J 



positive ions from platinum, and Horton's spectroscopic work 



in the case of aluminium phosphate heated on platinum 



