1819.] and on the Atomic Weight for Iron. 299 



these elements ; namely, the iron, as half an atom only ; and 

 this palpable absurdity can only be obviated by one of the two 

 following methods : that is, either by multiplying all the atoms 

 by 2, so as to get rid of the fraction, or by considering the 

 number hitherto admitted for the atom of iron as being double 

 what it ought to be ; but by the former of these methods, we 

 should obtain a number representing the weight of an atom of 

 ferro-chyazate of potash, which would not be equivalent to the 

 known quantities of the different chemical agents which decom- 

 pose it. Such a number, therefore, must be wrong, and the 

 method of producing it must be so likewise. The other method I 

 have, therefore, no hesitation in resorting to ; and on trying how 

 far the number thus deduced for the atom of iron (namely 17*5) 

 could be reconciled with the well-ascertained composition of its 

 oxides, sulphurets, and chlorides, I had the great satisfaction of 

 finding that it was easily reconcileable to them ; and, moreover, 

 that it was the only weight hitherto deduced for iron which was 

 reconcileable to them all; the others being so only partially. 

 Thus although the number 35 for iron with 1 for oxygen = 45 

 exactly expresses the composition of protoxide of iron, yet it does 

 not adapt itself to that of the peroxide, which to 35 iron contains 

 15 of oxygen, or lj atom of the latter : here the absurdity of 

 half an atom shows itself. The difficulty in this case was appa- 

 rently well got over by Dr. Thomson, who considered the perox- 

 ide as composed of two atoms of iron with three atoms of oxygen, 

 or as 70 + 30 ; and if there were only two oxides of iron, this 

 explanation would very well apply. But there is an oxide of 

 iron distinctly established both as existing in nature and as a 

 product of art (1 mean that which Berzelius has called the 

 oxydum ferroso ferricum, and which Gay-Lussac obtained artifi- 

 cially by passing the vapour of water over red-hot iron), which 

 contains 27-i- per cent, of oxygen — a proportion intervening 

 between that in the protoxide and that in the peroxide, the com- 

 position of which cannot be stated in whole atoms either by 

 taking the weight of iron as 35, or as 70 : with the former it 

 must be set down as with 13*3, or lJ- oxygen, and with the latter 

 as with 26-6, or 2$ oxygen. The attempt to consider this oxide 

 as a combination of one atom protoxide with two atoms peroxide, 

 instead of a distinct oxide of itself, can only be considered as 

 one of those ingenious expedients which have been resorted to 

 in order to make an experimental result harmonize with a theo- 

 retical one when the two appeared discordant ; it will doubtless 

 be abandoned now that it can be shown that this discordance 

 was only apparent, and resulted from taking the weight of an 

 atom of iron at double its real amount. 



The following table of the compounds of iron with sulphur, 

 oxygen, and chlorine, will show the manner in which the new 

 number now proposed for iron adapts itself to all those com- 

 pounds, several of which the whole number is inapplicable to. 



