92 Mi:ssi;s \. sMITHELLS, II. M. DAWSON, AND H. A. WILSON: ELECTRICAL 



of electrodes was placed between the cones, and the electric currents between them 

 measured under various E.M.F.8. 



It was found that any unsteadiness of the flame produced considerable and 

 irregular deflections of the galvanometer which prevented the current from being 

 measured in a satisfactory manner. The steadiness of the flame could be judged 

 readily by observing the lower cone, which, in consequence of the large amount of air 

 mixed with the gas (about 5 vols. air to 1 of coal-gas) in order to produce separation 

 of the cones, was very sharply defined. This cone in our earlier experiments could 1 e 

 seen to oscillate up and down and from side to side. 



With a view to reducing this oscillation to a minimum, so as to obtain a steady 

 galvanometer deflection, care was taken to regulate the quantity and pressure of the 

 gas and air supplies with as much nicety as possible. This greatly diminished the 

 oscillations, and a further great improvement was effected by admitting the gas close 

 to the nozzle of the sprayer, so as to produce a more perfect admixture. 



With the arrangements finally adopted, it was found that the values obtained for 

 any particular solution remained approximately constant during the whole period 

 covered by our experiments. 



The air was supplied under high pressure by means of a Westiughouse air-pump, 

 the amount of air used being but a small fraction of what the pump could supply. 

 The air was filtered through a long cotton-wool plug, P (fig. 1), and a first adjustment 

 of the supply was afforded by a tap close to a small mercury manometer, M. The 

 excess of air thus diverted blew off through a tube dipping 160 centims. below the 

 surface of water contained in the cylinder, C. The air to be used was next passed 

 into an iron drum, D, of 500 litres capacity, in order to damp down any pulsations of 

 pressure. 



The coal-gas was regulated by admission into a large gas holder, G, whence it 

 passed through a micrometer screw tap, S, to the flame tube. 



The adjustment of the sprayer is a matter of great importance, as the constancy of 

 its action determines mainly the accuracy of the experiments. To obtain constant 

 action it is necessary to work the sprayer with a much greater air supply than is just 

 sufficient to actuate it, otherwise small variations in the pressure produce considerable 

 changes in the amount of spray produced. Besides this, a strong air supply, producing 

 a large amount of spray, permits the use of more dilute solutions than would other- 

 wise be necessary. The difficulty of obtaining reliable results increases rapidly as the 

 concentration of the solutions is increased. 



The arrangement of the sprayer finally adopted is shown in fig. 2. The outer tube, 

 0, was blown like a test-tube, with a hole 2 millims. in diameter. The inner tube, I, 

 which was made narrow, so as to leave as much free space as possible, was joined to a 

 wider tube, and, after exact adjustment, this was cemented to the outer tube. The 

 sprayer was fitted into a large paraffined cork, C, which closed the end of the tubulated 



