28 



ON THE ELEMENTARY AND CONSTITUENT 



bodies are conducted in a definite and invariable ratio of relative weight 

 or measure as that of one part, to one part, one part to two parts, one 

 to three, or one to four ; and, consequently, that every change in the 

 compound thus produced, whether of addition or diminution, is a pre- 

 cise multiple or divisor of such ratio ; or, in other words, that the differ- 

 ent elementary bodies which enter into such compounds can never unite 

 or separate, never lay hold of or let go each other, in any other proportions. 



Let us exemplify this remark by a familiar instance or two. It is now 

 well known to every one, that the calxes, oxydes, or, as they are often 

 called, rusts, of metals, consist of a certain portion of oxygene with a 

 certain portion of the metal, which is thus converted into a calx or oxyde. 

 It is also known in the present day to most persons, that the greater 

 number of metals are possest of two or more kinds of oxydes, produced- 

 by an union of different proportions of the oxygene and the metal, and 

 often distinguishable even by their colour ; as minium or red lead, and 

 cerusse or white lead, which are equally oxydes of the metal whose name 

 they bear. Now, in whatever proportion the oxygene unites with the 

 metal to produce an oxyde of one kind, it invariably unites by a multiple 

 or divisor of the same proportion to produce every kind of oxyde belonging 

 to the same metal. Thus we have discovered not less than four different 

 oxydes of antimony in different parts of the world : the lowest or simplest 

 of them contains 4^ parts of oxygene to 100 parts of metal ; the next sim- 

 plest contains 18 parts of oxygene to 100 parts of metal, which is four 

 times 4^ ; the third oxyde consists of 27 parts of oxygene to 106 parts of 

 metal, which is six times 4^ ; and the fourth oxyde, 36 parts of oxygene 

 to 100 parts of metal, which is eight times 4i. So tin, which possesses 

 three discovered oxydes, has for its lowest the proportion of 7 parts of 

 oxygene to 100 parts of metal ; for its second oxyde, 14 parts of oxygene 

 to 100 parts of metal, which is twice 7 ; and for its highest, 21 parts of 

 oxygene to 100 parts of metal, which is three times 7. I have given the 

 proportions in round numbers ; but if I were to use the fractions that be- 

 long to them, the comparative results would be precisely the same. Nor 

 can we possibly combine these substances in any other proportions so as 

 to produce oxydes, for the corpuscles of which they consist will not lay 

 hold of or let go each other in any other ratios. It is possible that we 

 may hereafter detect an oxyde of antimony consisting of a less proportion 

 of oxygene than 41 ; but if we ever should, we are confident beforehand 

 that such proportion will be 2i. It is also possible that we may meet with 

 an oxyde containing more than 4^ and less than 18 parts of oxygene in 

 103 ; but if we should do so, we can nearly anticipate that such proportion 

 will be 9. And hence, as these proportions, though constantly iriie to their 

 respective series, are constantly diversified in different substances, their 

 radical figures or numbers may be employed, and now actually are em- 

 ployed, and that very generally, and in perfect coincidence with the system 

 of the Pythagorists, as synonyms of the simple forms or substances whose 

 progressive character they describe. This curious coincidence of ancient 

 and modern philosophy, for at present I will call it nothing more, 1 cannot 

 but regard as a very marvellous fact ; and am not a Httle surprised that it 

 should not hitherto have occurred, as it does not appear to have done, to 



* The only apparent exception I am aware of to this general principle is in the combina- 

 tion of the elements of M. Dulong's detonating substance, or azotane, as described by Sir 

 Humphrey Davy, Phil. Trans, for 1813, p. 260: and it is, hence, probable that we are not 

 yet put into possession of the proper results. 



