208 Reflections on some Mineralogical Systems. 



confidence which the state of our knowledge should insure 

 it, we should apply it only to the physical molecules which 

 have been previously separated one by one to carry off 

 the foreign matter interposed. But, as the true physical 

 molecule is situated beyond our means, and the thought 

 only can reach it, a knowledge of the chemical element 

 would seem to be too remote for us ever to aspire to it. 

 Nevertheless nature and labour offer us some means. First, 

 it does not always happen that the physical molecules are 

 embarrassed by foreign matter: next, suppose several mi- 

 nerals whose physical division gives, for instance, an irre- 

 gular tetrahedron, but in all of uniform dimensions, and 

 that chemistry finds in one the elements a, b, c,d,e; in an- 

 other, a, b> c, J; in short, that J, e, and others if we please, 

 may be variable, but that a, b, c, may be sensibly invariable 

 in all the different pieces: now the species is unchangeable ; 

 therefore we have a right to conclude that a, b, c, are the 

 chemical elements of the species, and that the others are ac- 

 cidental. Jt is thus that chemistry itself furnishes a method 

 of correction which has been found sufficiently rigorous, 

 and the two molecules are still in out power. 



Taking the point in its most general sense; every time 

 that we can discover in any mineral whatever the relative 

 connexions of the simple component substances which 

 have been observed to be invariable, as well as the relative 

 dimensions of the solid which is produced by division in 

 all the divers directions of the cleavage, we have every thing 

 necessary to define a species. All minerals, however, do 

 not present the?e data; and this principle of specification, 

 however precise it may be, does not embrace the whole of 

 the mineral kingdom. 



Let us suppose a thousand individuals or mineral .mole- 

 cules of a single species suspended in the same solvent. 

 By a diminution of the dissolving power, these individuals 

 would tend to unite themselves in groups; it might then 

 happen either that the molecules should assume such an 

 arrangement as the aggregate would easily yield to me- 

 chanical division, whence we might extract the integral 

 molecule of the species or its representative J or that the 

 molecules might unite confusedly in an irregular mass, so 

 that the t vpe of the species could not be. recognized. Again ; 

 if wc suppose the molecules of several species in the same 

 solvent, we shall have two analogous cases ; the molecules 

 of each species might unite to form aggregates of sensible 

 dimensions, and aherwards concur in the formation of the 

 mass, in which each species would be perfectly discerni- 

 ble; 



