304 THE CANADIAN NATURALIST. [Sept. 



water can be found. At early morning the redbreasted mergan- 

 sers fly out to sea in large flock?, but return to fresh water in the 

 evening : its provincial name is " shell bird." 



Hooded Merganser, Lophodytes cucullatus (i/i7i?i.)Apparently 

 rare on the north-west coast, and generally obtained in the imma- 

 ture plumage. 



(^To he Continued.) 



ON THE ORIGIN AND CLASSIFICATION OF 

 ORIGINAL OR CRYSTALLINE ROCKS. 



By Thomas Macfarlane. 

 ^Continued from June Nuiiiber.) 



IV. — CHEMICAL COMPOSITION. 



Crystalline or original rocks have been hitherto regarded and 

 described as aggregates of minerals. No doubt the larger number 

 of them may be correctly enough thus characterised, but it is 

 doubtful whether the description applies to all the original 

 rocks. For instance, obsidian has always been classed among 

 these, and, on all hands, it is admitted that no minerals are 

 discernable in it, that it is perfectly vitreous, as much so as bottle 

 or window glass. A similar vitreous substance, unresolvable by 

 the microscope, forms, according to Vogelgesang, part of the 

 matrix of all true porphyries. Then we have many instances 

 of rocks, almost impalpable in texture, belonging to various 

 families, in which the microscope certainly reveals the presence 

 of separate minerals, but, frequently, leave their nature and^ 

 always, their composition undetermined. Besides the uncertainty 

 which thus very frequently surrounds our knowledge of the 

 mineralogical constitution of fine-grained rocks, there are other 

 considerations which tend to shew that the composition of a rock 

 is not ascertained even after its constituent minerals have been 

 determined. In the first place, the relative quantities of these 

 present cannot be ascertained, and, secondly, even when this is 

 done approximatively, the uncertain composition of the mineral 

 species renders the chemical composition of the rock almost as 

 doubtful as before. It would therefore appear simpler and tend 

 to a juster view of the nature of original rocks, to regard them 



