Mr. W. Francis's Examination of a crystallized Nickel Ore, 337 



were adopted, the proportion 



7*74 : 19-35 

 would differ more from the result of the analysis. According 

 to the lirst supposition, the combination, in 100 parts, should 

 consist, admitting only nickel and arsenic as essential consti- 

 tuents, of 



Arsenic ...... 35*25 



Nickel 64-75 



and with this the analysis agrees exceedingly well. According 

 to the second formula (Ni 5 As 2 ), on the contrary, the compo- 

 sition should be 



Arsenic 33-69 



Nickel 66-31. 



Now, whichever formula may be considered as most cor- 

 rect, thus much is certain, that the combinations described in 

 the preceding Paper, and in this, are by no means composed 

 according to simple relations, as t : 2, or 2:3; nor will this 

 appear at all extraordinary, in a metal such as arsenic, which 

 forms an acid consisting of two atoms of radical and five atoms 

 of oxygen. It is, indeed, questionable whether the formula Ni 3 

 As 2 , laid claim to by Prof. Wohler for the nickel ore analysed 

 by him, does not likewise possess an atomic relation similar 

 to that of the combinations here treated of. 



Now the proportion 2:3, 3:5 approach one another very 

 nearly, as the following calculation shows : — 

 Ni 3 As* Ni 5 As 3 



Nickel 54-15 56'75 



Arsenic . . . 45*85 43-25. 



Professor Wohler, however, found the composition to be 



Nickel 55 



Arsenic 44, 



approaching, therefore, nearer to the formula Ni 5 As 3 than 



to Ni 3 As 9 . It is therefore very probable that the number 5 also 



occurs in the atomic relation of the constituents in this combi- 



tion ; and there would then have been found on the whole the 



following proportions between cobalt (or nickel) and arsenic : 



Third cobalt-combination Co 5 As (?) 



First and second do. as like- f Co 5 \* 2 S^° ? \a 3 



wise purified nickel ore ^Ni 5 j ' ' \Ni 7 J 



Common nickel ore Ni 5 As 3 . 



The middle combination will, according to theory, be re- 

 garded as far more probably Co 5 As 2 ; although the results of 

 the analyses bring it nearer to the formula Co 7 As 3 . 



Phil Mag. S. 3. Vol. 17. No. 111. Nov. 1840. Z 



