ON THE ATOMIC WEIGHT OF CHLORINE. 191 



measured volume against the standard solution of iodine in potassium iodide. A 

 measured amount (about 6 to 7 cub. oentims.) of the sodium thiosulphate solution 

 was run into the wash-bottle and made alkaline by the addition of sodium hydrogen 

 carbonate. The taps D and K were now opened, and the combustion apparatus 

 evacuated (in a stream of water-vapour) as far as possible by the pump. A rapid 

 stream of water- vapour was produced by immersing a large condenser, fused to the 

 pump, in a freezing mixture of ice and salt, and by gently warming the lower part of 

 the combustion globe with warm water. This was done to facilitate the removal of 

 traces of air and nitrogen, and that this was accomplished we concluded from the 

 small amount of nitrogen discovered in the subsequent gas analysis. During the last 

 period of the exhaustion, the calcium chloride and ice freezing mixture, in which the 

 bulb containing liquid chlorine was immersed during the combustion, was prepared, 

 placed in a wide-necked, unsilvered Dewar tube, and packed well round the liquid 

 chlorine bulb. 



The evacuation completed, the taps D and H were closed, and the glass cul-de-sac M 

 broken by jerking the glass rod N against it. 



The heating of the palladium bulb, enclosed in a stout copper box covered witli 

 asbestos sheet, was next started, the temperature being noted by means of a mercury- 

 nitrogen thermometer. 



The temperature of the liquid chlorine was now between 25 C. and 30 C., and 

 the pressure on the special tap Q was therefore not greatly above atmospheric. 

 Q was slightly turned so as to admit chlorine slowly into the combustion globe. 

 When the pressure of gas in the globe had become nearly atmospheric, the tap Q was 

 closed. This point was determined by the change in the faint hissing noise which 

 attended the entry of chlorine into the vacuum. When the palladium bulb had 

 reached a suitable temperature, all lights were turned out. 



Next came the ignition of the jet of hydrogen. Whilst a rapid succession of sparks 

 was passed between the platinum-iridium tips, the tap N was very cautiously opened 

 so as to admit the hydrogen slowly into the combustion globe. The moment the jet 

 of hydrogen had ignited the sparks were discontinued, and all attention was centred 

 on the flame. To cool the globe during the combustion, ice was packed round the 

 lower portion, while that part which was immediately above the flame was cooled by 

 a stream of cold water. 



To avoid, as far as possible, any diffusion of hydrogen through the flame, the 

 combustion was carried out at a pressure only slightly below atmospheric. The 

 atmosphere of chlorine was constantly replenished through the tap Q, whilst the 

 tap N regulated the admission of hydrogen to the flame. 



The combustion of hydrogen in chlorine at a glass jet is an interesting phenomenon. 

 The flame can be divided into two zones an inner zone of a light apple-green colour, 

 with an outer zone of less pronounced hue. We learnt by experience that three 

 points in connection with the flame were important for our purpose. Firstly, the 



