72 JOURNAL Ol^ THE 



became clearer for the discovery of the inter-relation 

 of the elements and the dependence of their properties 

 upon their atomic weights. It must be remembered 

 that for several of the elements the atomic weigfhts are 

 unknown and for others are very poorly determined. 

 It is extremely important that these be correctly 

 determined, and work spent upon them is far from 

 wasted, but I must confess that I can not but feel that 

 further efforts to discover the ratio between hydrogen 

 and ox3'g-en are of little value to science and that chem- 

 ists generally would be more grateful were the same 

 labor devoted to such elements as thorium, cerium, or 

 nickel, and many others. 



The question of the variation of the elements in their 

 atomic weights is a very elusive one and scarcely 

 capable of being finally settled by even the most accu- 

 rate work of the chemist. The supporters of the 

 hypothesis have always a loop-hole of escape in the 

 limits within which these numbers may be supposed to 

 vary. This variation is now narrowed down to the 

 decimal places. As the determinations become more 

 accurate, it is easy for the limit to be moved from 

 one decimal to another and so defy pursuit. Nothing 

 but absolute accuracy, an accuracy shown in every 

 experiment, with all sorts of varied proportions, and 

 not an averaged result, could finally end the discussion. 

 But with a settled standard and the atomit weights 

 known to the first decimal place, the way would be 

 clear for laws dependent upon their inter-relation. 



APPROXIMATION TO WHOI^E NUMBERS. 



Speculations upon the numerical relations existing 

 between the atomic wei«"hts be^fan almost with the 



