NATURAL SCIENCES OF PHILADELPHIA. 93 



ranean solutions, should even acquire more than a relative purit}'. 

 So that when we glance over the anal^^sis of the same mineral 

 made by different cliemists, it is often hard to say which elements 

 are those characteristic of the species. (See Glaucon-ite, Conarite, 

 Beudantite, etc. etc.) 



In some cases, the admixture of the two minerals can be proved 

 to be mechanical (gold dust in magnetic sand) ; in others it is 

 merely strongly suspected, but the mineral cannot be separated 

 by mechanical means (gold in pi/rites, etc.), while in the great 

 majority of cases the formula maker contents himself with two 

 arrangements of the atoms present connected by the sign +5 ^iit^ 

 each representing a different mineral. 



It is this that forms one of the greatest difficulties to the stu- 

 dent, at the present state of science, in forming any probable 

 conception of the mutual chemical relations of the various ele- 

 ments represented. Such an hypothesis is certainly tenable in 

 two cases: 1st. When the resulting mass cannot be classed under 

 any of the crj'stal systems (limonite, etc.) and the mineral is set 

 down as amorphous or crystalline; and 2d. Where the percentage 

 of one or two of tlie supposed compounds is so small that the 

 morphological properties of that which greatly preponderates are 

 assumed by the whole mass. ( Quartz containing scales of specular 

 iron or needles of rutilite.) But when a mineral cr3^stallizes dis- 

 tinctly in one crystal sj^stem and is deliberately assumed to be 

 made up of two others, each crystallizing in a different one, the 

 case is exceedingly perplexing to the mind. Yet the greater 

 number of all the formulas determined by chemists exhibit this 

 anomaly. 



It was not surprising that this should be the case at a time 

 when the electro-polar theory of Berzelius was unquestioned, and 

 the bases and acids were placed opposite each other frequently 

 separated by the -f sign like the partners in a Virginia Reel, 

 and it was thought necessary to make them analogous b3' dividing 

 the oxygen between them. CaO, SO3 (anhydrite) or CaO -\- SO.j 

 which has a crystal form differing from that of S0 and CaO seems 

 to justify (Ag.,S),, + Sb.^S., (proustite), for in this latter case the 

 mineral is rhorabohedral (hexagonal) while one of its constituents 

 crystallizes isometric and the other rhombic. It is true that this 

 mineral belongs to the class of those of metallic habit, and is 

 therefore opaque, and thus its optical properties cannot be deter- 

 mined, but if it were transparent we should be under the necessity 

 of recognizing the power of a mineral which can only pi'oduce 

 the ordinarily refracted my + one that can produce two extraor- 

 dinary rays to form a mineral which can give rise to one extraor- 

 dinary ray, and so of pleochroism, etc. In this case the fancy is 

 pleased by the accidental coincidence of the crystal form of the 



0-4-2 

 resultant with the average of the other two ^!1-^ = 1 ; but we 



