RICHARDS. — THE ATOMIC WEIGHT OF STRONTIUM. 605 



Two hundred grams of the purest crystals, after having been dried at 

 139°, were dissolved in about a liter of the purest water and filtered into 

 a large platinum disli, into which was passed at first pure ammonia gas 

 and then pure carbon dioxide through a platinum tube. The pure stron- 

 tic carbonate was washed by decaiitation eight or ten times, dried on the 

 steam bath, and ignited in a double platinum crucible over a Berzelius 

 lamp. Thus were residual traces of nitrate decomposed. 



The mixture of oxide and carbonate obtained in this manner was dis- 

 solved in very pure hydrochloric acid, which had been distilled in a plat- 

 inum still. The chloride was evaporated to dryness, fused in a platinum 

 dish to render silica insoluble, dissolved in purest water, allowed to stand, 

 filtered, acidified, crystallized, and dried. That used for Analysis 1 was 

 crystallized only once, that used for Analysis 4 was crystallized three 

 times, while that used for Analyses 2 and 3 came from the evaporated 

 mother liquors of the last crystallization. The essential agreement of 

 the results from these three fractions shows that if any barium had 

 escaped separation as sulphate, it was eliminated by the first recrystalli- 

 zation as chloride. This is not surprising, since baric chloride is heter- 

 omorphous with the strontic salt, these substances crystallizing with two 

 and six molecules of water respectively, at ordinary temperatures. The 

 previous double crystallization as nitrate could not have eliminated 

 barium, since strontic and baric nitrates cystallize in similar anhydrous 

 forms. 



The silver used was similar to that used in the work on magnesic 

 chloride.* It was finally prepared for weighing by fusing the elec- 

 trolytic crystals in a boat of lime in a vacuum. Subsequent work has 

 shown that this silver may have contained as much as 3-0000 ^^ '^^^ weight 

 of oxygen, derived from the decomposition of the included argentic 

 nitrate from which the crystals were deposited ; but no correction is made 

 for this trace of impurity, because its amount is uncertain, and because 

 the analytical work can hardly be accurate to within one part in thirty 

 thousand. 



Before analysis the strontic chloride was fused in a current of hydro- 

 chloric acid and nitrogen, cooled in pure dry nitrogen, and transferred in 

 its platinum boat to the stoppered weighing bottle in pure dry air. For 

 the first determination the apparatus used with the bromide was employed 

 for these operations, but in the last three analyses the improved form 

 devised for the work on magnesic chloride was employed. f 



* Rlclianls and Parker, These Proceedings, 32, 62 (1896). 

 t Ibid., 32, 58 (1896). 



