303 Intelligence and Miscellaneous Articles, 



less had not been previously exposed to a white heat, such as those 

 mentioned in M. Dufrenoy's TraiU de Mineralogie, although the 

 oxygen of the protoxides always presents a slight deficiency, we 

 obtain results very similar to the preceding*. Certain groups of 

 minerals, although belonging to the class of those M'hich are formed 

 in the manner of lavas, are nevertheless capable of containing in a 

 state of mixture quantities, sometimes considerable, of minerals 

 formed in the manner of sulphur. It is thus that we may explain 

 the presence of fluorine, boron, titanium, &c. in pyroxene, and espe- 

 cially in hornblende. 



The following is an immediate consequence of this remark. The 

 efforts which have been made by mineralogists to reduce to a simple 

 formula the analyses of hornblende, in which the alumina may amount 

 to 20"0, or be entirely wanting, are well known. The opinion gene- 

 rally adopted for want of a better is that of M. BonsdorfF, who, sup- 

 posing that alumina is isomorphous with silica, represents the mineral 

 pretty well by a silicate of protoxides, in which the proportions of 

 the oxygen in the acid and in the base would be 9:4, instead of 

 8 : 4 as in pyroxene. But everything is explained and simplified if 

 we suppose that the alumina is foreign to the normal composition of 

 the mineral, and that it only exists there as an integral portion of a 

 substance formed essentially in the manner of sulphur ; a spinellide, 

 that is to say, a body of the form R^ O^ RO, like the spinels of iron 

 or magnesia, protoxide of iron, &c., which are found mechanically 

 mixed with hornblende, and consequently never alter its crystalhne 

 form. 



In applying this hypothesis to a great number of analyses taken 

 at random from amongst those contained in M. Rammelsberg's work, 

 we find, after previously deducting the intermixed aluminate, that 

 the oxygen of the silica is nearly equal to double that of the prot- 

 oxides. The analyses which present a noticeable discrepancy are 

 those of the hornblendes, which contain much magnesia, and this fact 

 bears a relation to the difficulty which is experienced in determining 

 this base exactly. 



I wished my idea to be checked by analyses made in the laboratory 

 of the Ecole Normale with the precision afforded by the methods 

 recently introduced there. The following are the results furnished : — 



* I do not mean to say that the boron is necessarily in the form of 

 boracic acid in tlie tourmaline, nor that fluorine may not be normally con- 

 tained in it. This is a point that can only be settled by analyses in which 

 these bodies shallbave been very exactly determined. My hypothesis consists 

 simply in considering fluorine as capable of substituting itself atomically 

 for oxygen. 



I have as yet arrived at nothing so clear as regards mica, but I have no 

 doubt that the apphcation of the same principle to complete analyses of 

 this singular body will explain its anomalous composition ; and it may be 

 that the optical anomalies which M. de Senarmont has simpUtied and de- 

 fined so remarkably in his memoir, are only due to the variable proportions 

 in which the mutual substitution of fluorine and oxygen takes place in it ; 

 it is a question which I propose, with my brother's assistance, to examine 

 in a separate memoir. 



