ON THE ATOMIC WEIGHT OF CHLORINE. 



181 



After many failures we finally designed the vessel shown in fig. 4. The chlorine 

 weighed in the bulb A could only reach the tap when the sealed end of the inner 

 tube B was broken off by the rod of glass C falling on it. The tap D was an 

 inversion of the ordinary form of tap, that 

 is, its smallest diameter is at the top of 

 the tap ; so that instead of the key having 

 to be pushed into its socket, it has to be 

 pulled into it to fit. Internal pressure, 

 instead of tending to loosen the key, only 

 made it fit more tightly. Of course, if 

 the internal pressure became too great, the 

 key was so firmly driven into its socket 

 that it stuck, and then became useless. 

 However, the taps we used, when lubri- 

 cated with viscid phosphoric acid, with- 

 stood a pressure of four atmospheres 

 without sticking. Their chief disadvan- 

 tages lay in the difficulties of cleaning and 

 lubricating them, and in the fact that it 

 was necessary to affix to them weights, 

 suspended from a pulley, when carrying 

 out exhaustions of vessels to which they 

 were attached. We are not .aware that such taps have been used before in 

 scientific research work ; they were made for us by the University glass-blower. 



The small space E (less than - 5 cub. centim.) immediately below the key of the 

 special tap D, and the glass tubes connected with it, were first evacuated and then 

 filled with pure chlorine from the silver chloride through the tube F, which was 

 sealed off while the apparatus was cooled by immersion in a freezing mixture. On 

 the removal of the freezing mixture, the gas trapped between E and F (about 4 cub. 

 centims.) tended to expand, and thus held the tap D firmly in position. 



The chlorine condensation bulb, filled with approximately 37 grammes of liquid 

 chlorine, was weighed in a precisely similar manner to that detailed for the palladium 

 bulb. 



5. Preparation of Reagents. 



Iodine. Pure iodine was prepared by the first of the two methods proposed by 

 STAS. A strong solution of potassium iodide was saturated with resublimed 

 commercial iodine. To this, sufficient water was added to precipitate one half of the 

 dissolved iodine. The supernatant liquid was decanted and the precipitated iodine 

 repeatedly washed with small quantities of distilled water. It was then divided 

 into two portions. The iodine, in the first, was distilled in steam, the solid distillate 



Fig. 4. Chlorine bulb. 



