REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 141 



belonged to the same outflow, we should not see an alternation of compact and 

 amygdaloidal rocks. The intercalation of beds of lignite and fossil wood also 

 proves and gives precision to this interpretation. The beds prove that the upper 

 layers of the sheets have been exposed to meteoric agencies, and that, thanks to their 

 scoriaceous structure, they were readily disintegrated and transformed into argillaceous 

 matter, on which vegetation could take root and develop. The growth of large trees 

 proves that there were long intervals of rest between the eruption of the two basalt 

 sheets which enclose these vegetable remains. 



Accepting this view of the original arrangement of the basalt sheets, we must 

 consider that Kerguelen formerly presented the aspect of wide basaltic plateaux broken 

 only by escarpments of trachyte and phonolite. It is principally to meteoric agencies 

 that the island owes its present shape. We have said that all the heights of one 

 region come to about the same elevation, and that on both sides of the valleys the 

 various strata occur at the same level. These topographical features show that the 

 hills belonged at one time to a plateau extending over the whole region, and that these 

 hills have been left when the valleys, which cut up and furrow the island, were carved 

 out of the original plateau by running streams, glaciers, and atmospheric agencies. 

 These agents, joining their powers with that of the sea, have formed the fjords and 

 bays which everywhere run into the central mass. These ragged coasts, these cliffs 

 and perpendicular crags and terraced mountains, in a word, the deeply trenched form 

 of Kerguelen, are all explained by the extreme abundance of the atmospheric pre- 

 cipitation which beats on those barren rocks, almost destitute of vegetation. On the 

 other hand, we have seen that glacial phenomena have left their mark everywhere, 

 and added their action to that of running water and of the sea. The oscillations of the 

 land, frequent elevation and subsidence, have also contributed to modify the shape 

 of the island. Everything indicates that these great topographical movements and the 

 epoch of the extension of glaciers have been subsequent to the last outflow of basalt. 

 Finally, we must admit that the causes which have produced the vertical relief and 

 outline of Kerguelen have extended their action beyond the present limits of the 

 island and encircling rocks, and that the central mass is but the remains of a great 

 denuded land. The present configuration shows this, and so does the development 

 of vegetation in earlier periods. As Dr. Studer observes, even if we admit a higher 

 mean temperature in order to explain biological facts, it does not suffice to explain 

 the existence of a flora, for which a much larger land is required, in order to afford 

 protection against the storms that now carry devastation to every part of the island. 

 "We are thus led to admit that in times anterior to our epoch Kerguelen was a vast 

 mass of land. The topographical features that we mentioned at the beginning, and the 

 results of soundings made by Ross, and on the "Gazelle" and Challenger, confirm 

 this view, and point to a probable extension towards the south-west. 



