CHEMICAL SCIENCE. 193 



their atomic weights, we may find, in certain cases, that it is identical with 

 the ratio subsisting between the atomic weights of two other substances, and 

 so on through a considerable number of elements. The ratio between the 

 atomic -weights, for instance, of oxygen and nitrogen, is that of four to 

 seven, so likewise is that between those of zirconium and potassium, potas- 

 sium and barium, with absolute exactitude. What renders this the more 

 remarkable is, that all three of these last substances are striking exceptions 

 to Prout's law, that the equivalents of the elements are exact multiples of 

 that of hydrogen; they all have decimals, zirconium 22.4, potassium 39.2, 

 barium 68.6. Now, the ratio just mentioned gives these numbers with their 

 decimals with perfect exactness. The same species of relation also exists 

 between many other elements. 



" Again, the atomic weight of carbon stands to that of nitrogen in the ratio 

 of three to seven, a proportion which is found exactly or approximately to 

 extend to certain other elements. Apart, also, from these more general 

 ratios, many elements may be classed together in double or treble pairs, 

 such that the two elements in one pair stand to each other in the same 

 numerical ratio as the two elements of a second or third pair, the two ele- 

 ments constituting each pair being more or less closely allied to each other 

 iu properties, though the pairs are not necessarily analogous with those with 

 which they are compared. 



" For example, arsenic stands to antimony in the same numerical ratio as 

 selenium to tellurium, within an extremely small fraction, so that by multi- 

 plying and dividing we have : 



Arsenic 75 X Te]hmum 6 * = 120, Antimony = 120.3." 

 Selenium 40 



Our space does not allow further reference to the details of this paper than 

 to give its conclusion, which is as follows : 



" It is not easy to fix the exact amount of importance which attaches to 

 the numerical relations up to this time ascertained to exist between the 

 atomic weights of the elements. Some are, no doubt, mere casual coinci- 

 dences, and relations remarkably exact and symmetrical may exist between 

 the atomic weights of bodies which have no analogies in their properties: 

 for example, we may take calcium twenty, scier.ium forty, uranium sixty, 

 bromine eighty, mercury one hundred. Here the differences are not only 

 exact, but all the subsequent numbers are multiples of the first, and this 

 between bodies remarkably dissimilar in their properties, a striking proof 

 of the necessity of caution in inferring relations of properties as following 

 from relations of numbers. But, on the other hand, to reject the relations 

 of number when accompanied by analogy of properties as unmeaning and 

 unimportant, would be to err quite as much on the other side. When the 

 received equivalent of an clement, forming a term in a well marked series, 

 differs from that obtained by calculation, it naturally leads, as Professor 

 Mallet has remarked, to suspect an error and desire a redetermination. The 

 fact that a group of elements, allied in their chemical characters, may be 

 arranged in a scries having a common difference or a definite ratio between 

 its terms, confirms the propriety of grouping those elements together, and 

 such analogies may, in doubtful cases, assist us in arriving at a correct 

 classification." 



17 



