REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 139 



but it is much more probable that denudation has taken a leading part in the formation 

 of these terraces, the limits of the erosion being determined by the alternations of 

 massive and vesicular structure. We shall see that the surfaces of many of the super- 

 imposed layers have been directly exposed to atmospheric agencies, the influence of 

 which has been most powerful on their scoriaceous parts. 



The sheets of basalt contain masses of trachyte and phonolite, which are often 

 associated, and form the escarpments crowning the heights of the island. These crests 

 of trachyte or phonolite are shown in Table Mountain, in the region of Betsy Cove, 

 at Koyal Sound, and, above all, at Greenland Harbour. The stratigraphic relations of 

 the basalts and trachytes, on which we have insisted in describing Mr. Buchanan's 

 observations as confirmed by Dr. Studer, undoubtedly go to show that the phonolitie 

 and trachytic masses were erupted before the outflow of the basalt sheets. In this con- 

 nection we may recall an observation of Professor Both which establishes this order of 

 succession. He found that a trachytic rock from the neighbourhood of Mount Peeper 

 had been exposed to the caustic action of basalt. On the other hand, we have stated 

 that at Greenland Harbour, where basalt and phonolite are found in contact, it is the 

 latter rock that has undergone the mechanical effects of the intrusion, which has formed 

 a true friction-breccia. This necessarily implies the pre-existence of the phonolite. 



Taking account, then, of all these observations, it is necessary to admit that in 

 Kerguelen trachyte and phonolite have preceded the basaltic eruptions. There is also 

 sufficient reason for the statement, based on the structure and composition of the 

 trachyte and basaltic series as shown in the island, that their eruption is comprised 

 within the recent volcanic period. 



We may also recall the fact that all these rocks, generally altered, are filled with 

 minerals of secondary formation, such as delessite, zeolites, quartz, chalcedony, agate, 

 &c. This greatly complicates the question which must now be put, viz., Are there 

 erupted rocks in Kerguelen which belong to more remote geological periods ? Professor 

 Roth and Dr. Studer were inclined to think so. The reasons which led the former 

 to suppose that paleo-volcanic rocks were found there are as follows : — Amongst the 

 specimens from near Mount Crozier he found a micaceous diorite and a fragment of red 

 porphyry, from Lake Margot a labradorite porphyry, and at Winterhafen a rock was 

 picked up which resembled certain dolomites of the crystalline schists. The existence 

 of ancient crystalline rocks in oceanic islands appears incontestable, and we have shown 

 their presence in many of them. Still, in the present state of our knowledge, we 

 think it premature to state positively that outcrops of these ancient rocks exist in 

 Kerguelen. 1 While freely admitting the correctness of Professor Roth's determinations, 

 one may reasonably inquire whether the specimens he examined have not been con- 



1 Mr. Eton says that limestone has been found near Foundry Branch ; he adds that Mr. Stone of H.M.S. " Supply " 

 showed hiui the cast of a fossil shell which a sailor picked up near Thumb Peak ; Phil. Trans., vol. clxviii. p. 2. 



