12 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 11, NO. 1 



were followed by a very gentle heating after dryness to drive 

 off the ammonium chloride. It would thus appear that the so- 

 dium carbonate fusion was the only factor that might lead to loss 

 in weight, whether through the temperature of ignition or through 

 the action of the carbonate. The same balance and the same 

 weights were used throughout the series. 



On June 12, 191 8, the weight of the palau crucible was 32.0712 

 g., and the weight diminished steadily and with great regularity 

 through the series of 47 carbonate fusions until the last weighing, 

 on November 8, 1920, when it was 32.0613 g. The difference is 

 0.0099 g., giving an average loss in weight for each fusion of 

 0.00021 g., that is, about 0.2 mg. 



For comparison with this there was taken a series of 2 1 weigh- 

 ings of an iridium-platinum crucible (percentage of iridium 

 unknown). The first weighing of the series was made just after 

 the crucible had been scrubbed with sea sand (Dec. 5, 19 14), 

 and the last when another set of weights was put into use (Sept. 9, 

 191 6). This crucible, likewise, was reserved for the sodium 

 carbonate fusions, and only occasionally for the evaporation of 

 the alkali chlorides. The initial weight of this crucible was 

 30.1493 g., and the final, after a series of steadily diminishing 

 weights, was 30.0656 g. This gives a total loss in weight of 

 0.0099 g. (by coincidence identical with the other), and an 

 average loss in weight of 0.00047 g. Thus the iridium-platinum 

 crucible showed an average loss, due to the fusions, rather more 

 than twice that of the palau. It is well known that crucibles 

 made of platinum containing iridium lose weight very noticeably 

 on ignition, which is commonly ascribed to volatization of 

 iridium, and this factor would seem to have had influence here. 



No estimate was made of the average loss in weight suffered 

 by pure platinum crucibles which had been in use before the 

 iridium-platinum one, as my notes showed that they had been 

 used occasionally for the fusion with potassium pyrosulfate, 

 which would cause a serious and rather irregular loss (which 

 may be as much as 5 mg.). For the same reason a long series of 

 data on my crucibles that had been used in my own laboratory 

 at Locust, New Jersey, from 1896 to 191 2 was rejected. 



