ATOMIC WEIGHTS 17 



Third Series. — Hydrogen from Caustic Potash. 



.42205 3.7865' 15.943 



.4284 3.8436 15.944 



.4205 3.7776 15.967 



.43205 3.8748 15.937 



.4153 3.7281 15.954 



.4167 3.7435 15.967 



Mean, 15.952, ± .0035 

 Mean of all as one series, 15.953, ± .0020 



Shortly after the appearance of this paper by Cooke and Eichards 

 Lord Eayleigh pointed out the fact, already noted by Agamennone, that 

 a glass globe when exhausted is sensibly condensed by the pressure of 

 the surrounding atmosphere. This fact involves a correction to the fore- 

 going data, due to a change in the tare of the globe used, and this cor- 

 rection was promptly determined and applied by the authors.^ By a 

 careful series of measurements they found that the correction amounted 

 to an average increase of 1.98 milligrammes to the weight of hydrogen 

 taken in each experiment. Hence equals not 15.953, but 15.S69, the 

 probable error remaining unchanged. The final result of Cooke and 

 Eichards' investigation, therefore, is 



= 15.869, ±.0020 



Keiser's determinations of the ratio were published almost simul- 

 taneously with those of Cooke and Eichards. He burned hydrogen oc- 

 cluded by palladium, and weighed the water so formed. In a preliminary 

 paper " the following results are given : 



Mean, 15.864, ± .015 



IN^ot long after the publication of the foregoing data Keiser's full paper 

 appeared.^ Palladium foil, warmed to a temperature of 250°, was satu- 

 rated ^\ith hydrogen prepared from dilute sulphuric acid and zinc free 

 from arsenic. From 100 to 140 grammes of palladium were taken, and 

 it was first proved that the metal did not absorb other gases which might 

 contaminate the hydrogen. Before charging, the foil was heated to bright 

 redness in vacuo. After charging, the tube containing the palladium 



1 Proo. .\mer. Acad., 23, 182. Am. Chom. -lourn., 10, 191. 



= Berichte, 20, 2323. 1SS7. 



SAmer. Chem. Journ., 10, 249. 18SS. 



