12 PROCEEDINGS OF THE AMERICAN ACADEMY 



Below the temperature of 70° in somewhat moist air, or at the 

 ordinary temperature in perfectly dry air, it loses one of these 

 molecules.* The other one is retained until a temperature of from 

 100° to 130° is reached, according to the hygroscoj^ic condition 

 of the surrounding air. 



The accuracy of the final result for the atomic weight of course 

 depends uj)on the complete absence of water from the dried salt; 

 hence an especial series of experiments was made to determine the 

 conditions under which the water was completely expelled. Upon 

 heating to redness, the salt is very slightly decomposed; f hence 

 in all cases where a high heat was used the amount of baric hydrox- 

 ide and baric carbonate formed were determined by means of very 

 dilute standard hydrobromic acid, using phenol phthalein and 

 methyl orange respectively as the indicators. The accuracy which 

 it is possible to attain in this process was a great surprise. If a 

 very small amount of pure boiled water is used for the solution of 

 the baric bromide, a deficiency of less than a tenth of a milligramme 

 of bromine in five grammes of the salt is detected with the greatest 

 ease. The correction applied to the weight of the baric bromide 

 was of course always the calculated difference between the weights 

 of the bromine and the hydroxyl, or the carbonic acid, which had 

 taken its place. For example, a deficiency of 0.81 milligramme 

 of hydrobromic acid found by alkalimetry involved a correction of 

 0.63 milligramme if the alkaline earth had been found in the form 

 of hydroxide, or 0.50 milligramme if it had been found in the 

 form of carbonate. Since baric carbonate is very faintly alkaline 

 to phenol phthalein, this correction is not absolutely exact ; but its 

 error is an infinitesimal one so far as this work is concerned. It 

 is probable that, if any traces of oxide were formed, they were con- 

 verted into hydroxide or carbonate before the crucible cooled. 



One experiment showed that 1.6 grammes of baric bromide dried 

 at 136° lost 0.4 milligramme on being heated to dull redness. On 

 another occasion, three grammes of baric bromide which had been 

 dried at 200° to constant weight lost 0.50 milligramme upon heat- 



* 2.8688 grammes of crystallized baric bromide lost 0.1547 gramme on 

 heating to constant weight at 70-80°; the residue lost 0.1533 gramme more 

 upon heating to constant weight at 160°. Compare Graham Otto (Michaelis), 

 III. 662. 



2.0506 grammes of baric bromide which had been powdered and dried over 

 sulphuric acid to constant weight lost 0.1181 gramme upon drying at 200°. 



t Schultze, Jour. f. prakt. Chem., [2.] XXI. 407. 



