210 SCIENCE PROGRESS. 



Several atomic weights have been redetermined with 

 great care, and of these determinations perhaps those 

 of T. W. Richards of barium and of strontium are the most 

 accurate and most interesting. By an exhaustive research 

 on barium bromide he deduces the value Ba = 137*434 

 (O = 16) (9). From a similar study of barium chloride the 

 value Ba = 137*440 is deduced (10). 



This value is notably higher than that usually accepted 

 and is no doubt due to the careful elimination of small 

 quantities of strontium and calcium which have contaminated 

 the preparations of earlier experimenters. From a study of 

 strontium bromide Richards found Sr = 87*659 (O = 16) 



Still more recently the atomic weight of zinc has been 

 determined by Richards and Rogers again by means of the 

 bromide and precipitation with silver, and as a mean they 

 find the value (Zn = 65*404) (O = 16) (12). 



In all the above determinations Richards estimated the 

 percentage of silver in his haloid silver salt and showed it 

 to be identical with that found by Stas, thus placing his 

 work on the same footing and guaranteeing in this way its 

 very high accuracy. 



In 1888 two other American experimenters, Burton and 

 Morse (13), published the results of their work on the same 

 atomic weight which they arrived at by means of the con- 

 version of the metal into the oxide by treatment with nitric 

 acid and ignition of the nitrate. Although their work 

 agrees throughout very well the value found is lower than 

 that of Richards, due no doubt to the retention by the 

 oxide of oxides of nitrogen as Marignac pointed out. In 

 defending their work against this objection they expose 

 their want of knowledge of the commonest reactions in such 

 a way as to make one distrust all their work. The perusal 

 of their paper provides much food for reflection of a serious 

 nature although it does give a certain amount of instruction 

 as well as amusement. They carry out their weighings to 

 *ooooi of a gram and pretend to detect differences of this 

 minute amount in porcelain crucibles which have been 

 heated up to the melting point of steel. In their account 



