!92 PEOFESSOE H. B. DIXON AND ME. E. C. EDGAE 



gradual elongation of the outer zone, together with a lessening of the luminosity ol 

 the inner zone, indicated that the atmosphere of chlorine was riot being renewed 

 quickly enough. Secondly, when the flame became smaller and more luminous, we 

 knew that the pressure of chlorine was in excess, and that the gas was being admitted 

 into the globe too quickly. Lastly, a gradual shrinking in the size of the flame, 

 unattended by any change of luminosity, indicated that the supply of hydrogen was 

 failing. This was, of course, remedied by raising the temperature of the palladium bulb. 



When the combustion had been carried to such a point that only a drop of liquid 

 chlorine was left in the condensation bulb, the tap Q was finally closed and the flame 

 made very small. As the atmosphere became rarefied, the outer zone of the flame 

 became elongated and less luminous ; the inner zone changed also, but to a less extent. 

 In one experiment (IV.), the flow of hydrogen not being reduced as the chlorine- 

 atmosphere became rarefied, a flame passed through the whole globe. Just before 

 the point of extinction the tap N was closed and the combustion was ended. The 

 duration of the combustion was about three hours, during which constant watching 

 was necessary. The palladium bulb was now allowed to cool to the ordinary 

 temperature. 



The two small bulbs, containing concentrated solution of potassium iodide, were 

 then 'broken by dashing them against the interior of the combustion globe, when the 

 residual chlorine was absorbed with precipitation of iodine. The precipitated iodine, 

 however, soon dissolved in the excess of potassium iodide. The tap D was opened 

 and the residual gases were sucked out of the combustion globe in a current of water- 

 vapour through the alkaline solution of sodium thiosulphate contained in the wash- 

 bottle H, in which the vaporised iodine was absorbed. The residual gases were 

 collected in the gas analysis apparatus. 



During this exhaustion the long glass tube R connected with the three-way tap D, 

 and containing NaHCO :i , had been heated. The evacuation completed, D was turned 

 and CO a admitted until the combustion globe was full. This was indicated by the 

 escape of gas through the manometer. The tubulure G was now opened, cleaned 

 from adhering phosphoric acid, and the residual iodine titrated in the atmosphere of 

 carbonic acid by means of the sodium thiosulphate solution of known strength 

 contained in a calibrated burette.* As sufficient potassium iodide was originally 

 contained in the thin glass bulbs to dissolve easily the precipitated iodine, the titration 

 was quickly and accurately carried out, five drops of starch solution being added 

 towards the end of the titration. f One drop of the standard solution of iodine 

 restored the blue starch-iodide colour to the decolourised liquid in the combustion 



h The two burettes employed were carefully calibrated by means of an Ostwald calibrator of 

 2 cub. centims. volume. The mean results of two calibrations were tabulated and used in determining 

 the volumes. 



t Owing to the action of hydrochloric acid on a solution of sodium thiosulphate, we were unable to add 

 excess of the sodium thiosulphate solution and titrate back with the standard solution of iodine. 



