S« On the Dallonian Theory of [Acg. 



Weight of an 

 atom. 



Silver 13*714 ° 



Iron 7* I4.i p 



Nickel 7-305 i 



Zinc 4-095 r 



Bismuth 8"994 s 



Antimony . . , 1 1*249 ' 



Tellurium 4-027 u 



Arsenic 6-000 x 



° This is founded on the analysis of chloride of silver, which is, 

 I think-, susceptible of more accuracy than any of the other methods 

 of estimating the quantity of oxygen with which silver combines. 

 I conceive oxide of silver to be a compound of 100 metal -f 7 291 

 oxygen. 



p In my first table I stated the weight of an atom of iron C'666 s 

 supposing the black oxide a compound of 100 iron + 30 oxygen, 

 and the red of 100 iron + 45 oxygen. These numbers coincide 

 with the experiments of Hassenfratz, and do not deviate much 

 from the experiments of Berzelius, and from some of my own 

 published some years ago in Nicholson's Journal. But a set of 

 experiments which I made last winter to ascertain how much iron 

 was taken up by a definite quantity of sulphuric acid, which I 

 consider as more susceptible of precision, led to the result that 

 black oxide of iron is a compound of 100 iron + -B oxygen. As 

 this agrees with the experiments of Gay-Lussae and of Dalton, I 

 think it preferable to my preceding number, especially as it nearly 

 agrees with try own preceding experiments on the composition of 

 black oxide of iron. 



^ The number for nickel in mv original table is altogether erro- 

 neous. 1 do not know how it crept into the table. 



r This number is the result of my experiments on the composi- 

 tion of blende, published in the present number of the Annals of 

 Philosophy, to which therefore I refer the reader. Berzelius con- 

 siders zinc as capable of combining with two doses of oxygen; but 

 J have not been able to form more than one oxide of this metal, 

 which 1 therefore consider as a protoxide. 



s Berzelius informs us {Annuls of Philosophy, vol. iii. p. 356) 

 that he has discovered the existence of a purple protoxide of bis- 

 muth, besides the common yellow oxide. If this prove correct, 

 we must double the weight of an atom of bismuth. 



1 See my paper on the analy-is of sulphuret of antimony pub- 

 lished in the present number of the Annals of Philosophy. 



1 The number in the original table was erroneous in consequence 

 of an arithmetical error. 



x I have not room here to assign my reasons for refusing to admit 

 BerzeHus' new determination of the composition of the oxides of 



