milligrm. This direct reading balance permits of very rapid weighing, a condition which is essential in the case of the anhydrous chlorides of radium and barium, which gradually absorb moisture, in spite of the presence of desiccating substances in the balance. The bodies to be weighed were placed in a platinum crucible; this crucible had been long in use, and its weight did not vary the tenth part of a milligrm. during the course of one operation.
The hydrated chloride obtained by crystallisation was placed in the crucible and heated till converted into the anhydrous chloride. When the chloride has been kept for several hours at 100° its weight becomes constant, and does not change even if the temperature is raised to 200°. The anhydrous chloride thus obtained constitutes, therefore, a perfectly definite body.
The following is a series of determinations on this point. The chloride (100 m.g.) is dried in the oven at 55°, and placed in a desiccator over anhydrous phosphoric acid; it then gradually loses weight, which proves that it still contains moisture; in the course of twelve hours the loss was 3 m.g. The chloride is replaced in the stove, and the temperature raised to 100°. During this process, the chloride lost 6·3 m.g. in weight. After being left three hours fifteen minutes in the oven, it lost 2·5 m.g. more. The temperature was maintained for forty-five minutes between 100° and 120°, which caused a loss of weight of 0·1 m.g. Then after being kept for thirty minutes at 125°, the chloride showed no diminution in weight. Then, however, after thirty minutes at 150°, it lost 0·1 m.g. Finally, after being heated for four hours at 200° it lost 0·15 m.g. During these operations the crucible varied from 0·05 m.g.
After each determination of the atomic weight, the radium was converted into the chloride in the following manner:—To the solution containing the weighed radium nitrate and excess of silver nitrate was added pure hydrochloric acid; the silver chloride was filtered off; the solution was evaporated to dryness several times with excess of pure hydrochloric acid. In this way the nitric acid is entirely removed.
The precipitated silver chloride was always radio-active and phosphorescent. In determining the amount of silver contained in it, I satisfied myself that no ponderable amount of radium had been carried down with it out of the solution. The method I pursued was to reduce the silver chloride precipitated in the crucible by hydrogen generated from dilute hydrochloric acid and zinc; after washing, the
