330 The Paragenetic Relations of Minerals. 



widely in chemical composition from all the other members, 

 inasmuch as alumina is an essential constituent. 



It is in the study of these more ancient rocks, as in that of 

 the early history of Man, that the greatest mystery, the great- 

 est difficulties are encountered, and the progress of observation 

 would appear to prove not only that the conditions under 

 which they were formed were widely different from those 

 which now prevail, but also that these rocks have experienced 

 many successive alterations. 



With regard to porphyritic rocks, Werner entertained the 

 opinion that the imbedded crystals were of earlier formation 

 than the matrix ; however the contrary appears, with very 

 trifling exceptions, more probable, and for the following rea- 

 sons : — 



1. The occurrence of crystals which contain nuclei of the 

 mass in which they are imbedded. — This is the case with the 

 large rhombohedrons of magnesite occurring in the talcose 

 slate of the Tyrol. Crystals of leucite, from the old Vesu- 

 vian lavas, not unfrequently contain the same lava in their 

 interior. The twin-crystals of pegmatolite from Elbogen, 

 Bohemia, contain not only scales of phengite and granules of 

 quartz, but nuclei of granite, presenting exactly the same 

 appearance as the surrounding rock. Very finely developed 

 crystals of iron pyrites from Osterode, Harz, with smooth sur- 

 faces, contain nuclei of the gypsum in which they lie, and large 

 groups of iron pyrites crystals frequently contain in their centre 

 some of the clay by which they are surrounded. In the horn- 

 stones of Schneeberg (Saxony) hexahedrons of tin white cobalt 

 or smaltine occur with nuclei of the same hornstone. Some 

 pseudomorphous minerals constitute an exception to this gene- 

 ral rule ; thus when an envelope of red haematite has been 

 formed over calcite, assuming its scalenohedron form, and the 

 latter mineral has been subsequently removed, and red hsema- 

 tite deposited in its place. But in this instance the pseudo- 

 morph was at a certain period hollow. Again, at Bothen- 

 berg (Saxony), the fluorspar formerly existing, and of which 

 not a trace is now found, was first covered with fibrous red 

 haematite, and then by some means removed, leaving a hol- 

 low cast, which was afterwards filled with quartz. Some 





