ON THE ATOMIC WEIGHT OF CHLORINE. 177 



determine its change of volume, if any. When immersed in water to a mark on the 

 stem the bulb weighed 



Before heating 230-314 grammes. 



After heating for two hours to a dull red heat .... 230 '311 



After a second heating 230-307 



After a third heating 230-308 



The alteration in volume of the bulb, after heating to dull redness, was there- 

 fore so slight that the difference in its displacement of air was negligible. In the 

 actual experiments the bulb was never heated beyond 550 C. 



The palladium was used in the form of thin foil. We are indebted to Messrs. 

 JOHNSON and MATTHEY for kindly supplementing our stock for the purpose of this 

 investigation. The bulb contained sufficient palladium foil (360 grammes) to absorb 

 about li grammes of hydrogen. When the bulb A had been detached from the rest 

 of the hydrogen apparatus and had been cleaned, it was suspended by platinum wire 

 from one arm of the balance, from the other was suspended a counterpoise (of the 

 same Jena glass) which had nearly the same displacement as the palladium bulb. 

 It was then weighed by GAUSS' method of reversals. The balance case, after each 

 reversal, remained closed at least half-an-hour before a new weighing was started. 

 The air displaced by the small weights added to secure equilibrium was allowed for. 

 By equalising the volumes of the systems suspended from the arms of the balance, 

 errors due to variations of temperature in the balance case and to any deposition of 

 moisture on the bulb were avoided. 



3. Preparation of Chlorine. 



SHENSTONE,* in 1893, first proposed the electrolysis of fused silver chloride in a 

 vacuum as the best means of obtaining pure chlorine. He stated that his chief 

 difficulty was the rapid formation of silver trees, which eventually made contact 

 between the electrodes and thus prevented any further decomposition of the fused 

 chloride. In 1901, MELLOR and KussELLf substituted for SHENSTONE'S tube a V-tube 

 of the hardest Jena glass, so that the silver tree had to travel along the two limbs of 

 the V before making contact, and thus the decomposition of silver chloride could be 

 carried on to a greater extent than in SHENSTONE'S apparatus. They fastened their 

 carbons to glass tubes (ground into the necks of the V) by means of a plaster of Paris 

 joint. 



We have modified their apparatus by drawing out the upper portion of each arm of 

 the U, and melting it on to the carbon electrode for a length of about 2 - 5 centims. 

 We fused a little silver chloride round the top junction of glass and carbon. A 

 mercury cup completed the joint, and served for making electrical contact with the 

 carbon. Such a joint, even with a vacuum in the interior of the tubing, is quite 



impervious. 



* ' Journ. Chem. Soc.,' 71, 471 (1897). 



t 'Journ. Chem. Soc.,' 82, 1272 (1902). 

 VOL. CCV. A. 2 A 



