NATURAL SCIENCES OP PHILADELPHIA. Ill 



these minerals, and while the mass could hav^e no crystal form of 

 its own, the mutual ratios of the resulting compounds would be 

 more constant as the process of decomposition was more perfect. 



It often results that in calculating a formula for a so-called 

 species the results of analyses of specimens from widely distant 

 localities, and made by different persons, agree remarkably well 

 together, while the atomic ratio is such as to resist all efforts to 

 bring these atoms into one homogeneous compound Often, too, 

 the student sees clearly that he is dealing with a partiallj' decora- 

 posed mass, and would, perhaps, be justified in writing "a, per 

 cent, of the mineral A with &, per cent, of the mineral B dissemi- 

 nated through it," but it is obvious that he must assign wide 

 limits to a and 6; and if the species possess that patent of genuine- 

 ness, crystal form, unless he define those limits his formula loses 

 its value. 



Take the case of smaltite. This mineral has veiy well-marked 

 physical properties and unmistakable crj'^stal form, and is an arse- 

 nate of cobalt, iron, and nickel, but its per cent, of As varies from 

 59 percent. (Salvetal & Wertheim) to about 75 per cent. (Karsten); 

 the Co from per cent. (Rammelsberg) to 20 per cent. (Stro- 

 meyer) ; the Ni from per cent. (Yarrentrapp & Stromeyer) to 

 29.50 (Rammelsberg); the Fe from trace (Rammelsberg) to IS 

 (von Kobell) ; Cu from per cent. (Lange, Booth, Karsten, &c.) 

 to 2 per cent. (Jackel). Besides these very large varieties of com- 

 position there are frequentlj^ found other elements with it, such 

 as Bi and S. Howls a formula to be constructed for such a 

 mineral? 



The only recourse is to the R's, and we have no less than four 

 groups of formulas proposed b}' Dana, under one of which every 

 smaltite yet analyzed can be brought. The first two of these are 

 really identical, and differ only in the different proportions in 

 which the analogous elements Is^'i, Co, and Fe, replace each other, 

 and may be written RAs^. But the next group (C) has the for- 

 mula RAs -f RAs,, and the third (D) RAs. + RASg. 



Independently' of the presumptive evidence against such a mix- 

 ture producing a beautiful octahedron of smaltite, where is to be the 

 limit to such formulas ? Why not RAs -|- RAs, -|- RASg -j- etc. ? 

 And would it not be well to adopt some more definite rule for 

 assigning formulas to minerals of such variable composition ? 



Without naming these laws certain facts can be assumed on 

 which to base them: 1st. If there is no single chemical formula 

 which expresses the constitution of a crj-stallized mineral, then 

 that mineral is a mixture. 2d. Its form is determined by the 

 preponderance in quantit}^ or in crystallizing power, or both, of 

 one of its constituents. 3d. The mineral can only present its 

 characteristics when the foreign ingredients are present under a 

 given per cent. 



The plan would seem to be to deduce from the known charac- 



