78 JOURNAL OF THE 



nation of this ratio. Much excellent work lias been done u})()U 

 it, and especially in the last two or three years. Ostwald has 

 summed these np, and says there is an error of at least 0.3 per 

 cent, which has not been removed by recent workers. Branner 

 agrees with him that the variations are irreconcilable, and though 

 Meyer and Seubert think the ratio can vary but little from that 

 assumed by them as justifiable by the best determination, it is 

 manifestly a point on which the best authorities differ and hence 

 one of uncertainly. 



Why should we then make use of the number 15.96 if it is 

 not fixed by incontrovertible, unerring, universally accepted 

 experiments? It makes the matter no whit better for Meyer and 

 Seubert to profess their willingness to recalculate their table 

 should a change in the number 15.96 prove necessary. It is 

 merely a confession of the insecurity of their own position. We 

 do not wish any recahndation. We wish a standard bv which 

 the calculations can be made once and for all, one that will give 

 us the least possible error and is itself independent of other cal- 

 culations. The present use of the double standard 15.96 and 

 16 seems puerile and leads to all manner of inaccuracies. 



The second possibility is to use oxygen as the standard. 



The question reduces itself, then, really to this: Shall we use 

 O = 15.96 or O = 16? For oxygen must be used from very 

 necessity. If O = 15.96 is not the absolute ratio or is not gen- 

 erally accepted as such then the reason for its use ceases. 



It is not necessary to hriDu- forward aro^uments as to the rela- 

 tive convenience of the two, nor as to their effect upon the peri- 

 odic ln\v. Such arguments have little weight or significance 

 when it is seen that the question lies between 0= 15.96 and 

 O^ 16, and that hvdroiren can never be the actual standard or 

 fa(;tor from i)urely chemical reasons. Nor yet is there much in 

 any argument from analogy with other standards and units. 

 Such only lead us away from the one nil-important considei^a- 

 tion — the avoidance of unnecessary errors. 



I have said that the present proposition could be looked upon 

 in the light of a compromise. If oxygen takes the place of 

 hydrogen as standard, what falls to the share of the latter ele- 



