154 Mi. II. Roliinson. A Redetermination of [June 19, 



nndecomposcd. When a low red heat was reached a slow current of 

 carbon dioxide was passed in along with the hydrochloric acid to 

 burn the little carbon left from a small quantity of the oxalic acid 

 nearly always charring. When the slightly grey appearance of the 

 chloride due to this small quantity of carbon disappeared the current 

 of carbon dioxide was stopped, that of the hydrochloric acid was 

 continued for an hour longer, while the chloride was kept at a full 

 red heat. With the hydrochloric acid still flowing over the chloride 

 the furnace was gradually cooled until it was safe to handle the glass 

 tube ; the exit end of the tube was then plugged, then the other end, 

 and the tube was removed from the furnace and wiped, clean from 

 the magnesia it had been lying in. The cork of the tube was then 

 withdrawn and the chloride slipped into the weighing flask which 

 was securely closed by a glass cap well ground on to its neck. It was 

 then quickly weighed while still quite warm. The object in 

 weighing the chloride at this stage in the process was to make quite 

 sure no water was absorbed before the final weighing was made, and 

 such was found to be the case, for although there was always a small 

 increase in weight, about 0'0025 grm. on 6'5 grras. of the chloride, it 

 was no more than was easily accounted for by the loss that would 

 arise from weighing the body warm. When weighing the chloride it 

 was placed in a small flask with a closely fitting cap and an exactly 

 similar flask was used as a tare ; the latter required the addition of 

 O'OOOS grm. to balance the weighing flask; whatever was done to the 

 one flask was always done to the other, and they were always kept 

 beside one another. To avoid repetition, I may say here that in all 

 the other weighings the body to be weighed was placed in one vessel 

 and another vessel as similar in material and weight as possible was 

 used as a tare. The weight having been taken, in order to remove 

 extraneous hydrochloric acid, the flask containing the chloride, and 

 the tare, were placed over sulphuric acid and surrounded by quick- 

 lime under the receiver of a large air-pump which could be exhausted 

 very rapidly to within 3 millims. of a vacuum. This was usually 

 done during an afternoon, they were left here the whole of the two 

 following days, and on the morning of the fourth day air, dried by 

 passing through sulphuric acid and phosphoric anhydride, was let 

 into the receiver. This treatment was found to completely free the 

 chloride from adhering hydrochloric acid, as when it was dissolved 

 its solution was perfectly neutral. The two flasks were now removed 

 from the receiver, securely closed, and the final weighing of the 

 chloride made ; it was then cautiously dissolved in water. 



In making the following determinations pure silver was used 

 prepared, according to Stas's directions, by precipitation from an 

 timmoniacal solution of the nitrate by ammonium sulphite. For 

 each experiment, after making the first weighing, 1 calculated how 



