RELATION TO MESOZOIC ROCKS IN THE LEPOXTINE ALPS. 221 



occasionally somewhat rounded, as if waterworn. This, indeed, is 

 SO marked in many cases that at the conclusion of my field-work I 

 was quite prepared to find that these minerals were derivative, and 

 represented the garnets and staurolites of the schist in a rolled 

 condition. But, as will be seen from the detailed account of the 

 microscopic structure in the Appendix, the spheroidal minerals are 

 totally different from the garnets, the prismatic minerals from the 

 staurolites or kyanites of the Piora schists, and all have been formed 

 in situ *. It was, indeed, already evident from the analyses quoted 

 by Dr. von Fritsch t, that the spheroidal mineral could not be a garnet 

 unless it included a large number of grains of quartz ; for it contains 

 over 53 per cent, of SiO.^, while the usual proportion for a garnet is 

 from 38 to 42 per cent.J The mineral also contains 6-06 of water, 

 and there is more than 3 per cent. " loss." Thus it is very different 

 in all respects from garnets in the Piora schists. Its analysis does 

 not correspond with that of any mineral known to me. This is not 

 surprising, because evidently so much impurity is present. One 

 statement only may be confidently made, that whatever the 

 cementing material may be to which the form is due, it is not 

 garnet. The prisms, on microscopic examination, are seen to be less 

 crowded with the detrital material, and the analysis, as Dr. von 

 Fritsch points out, suggests that of couseranite § ; and, as will be 

 seen from the description in the Appendix, its microscopic characters 

 fairly correspond with those of a mineral allied to dipyre. I have 

 examined the specimens of couseranite (all from either Poujac or 

 some other locality in the Department of the Ariege) in the British 

 Museum collection and at University College, and find that though 

 less perfect in external form and less pure, tliese Alpine specimens 

 present a general resemblance to them. I think it possible that in 

 the Lukmanier district a third hydrous silicate is also present. 



Thus the resemblances of these fossiliferous " spotted " rocks to 

 the Black-garnet schists are only superficial, while the differences 

 both of the larger included minerals and of the matrix, as will be 

 seen from the description in the Appendix, are marked and 

 essential. 



* It may be instructive to record iny exact impressions. The evidence at the 

 Lukmanier seemed to me on the whole to suggest a formation of the minerals 

 in situ; that at the Nufenen seemed to accord better with a derivative origin. 

 But in regard to the latter I was from the lirst conscious of a grave dif!icultv. 

 If the minerals were detrital — rolled garnets and staurolites — if the bands in 

 which they occurred were a kind of grit, how was it that I failed to find any 

 other fragments, such as vein-quartz or the harder parts of some of the schists ? 

 As, however, one or two bands (without the minerals) do contain detritus of 

 some crystalline " rock," I still thought it possible that something might be 

 revealed on microscopic examination. This, however, as may be seen, places it 

 beyond all doubt that the minerals liave been formed in sitn. 



t Lot: cif. p. 127. I have quoted these on p. 233. 



X Dana (Mineralogy, a.v. Garnet) quotes 84 analyses, in 82 of which the 

 SiO., is less than 43 per cent., and it is more ofteu below than above 40. (^ne 

 rises to 44, the other very exceptional analysis contains r)2 per cent. 



§ There is, however, too little both of silica aiul of alkalies. 



