370 REACTION OF THE INTERIOR OF THE EARTH 



The Island of Amsterdam (37 48' S.), or 50' of lati- 

 tude to the north of St. Paul's, consists, according to 

 Valentyn's drawing, of a single, wooded, rather rounded 

 mountain, on the highest part of which rises a small 

 cubical rock almost similar to that on the Cofre de 

 Perote in Mexico. In the expedition of D'Entrecasteaux, 

 in March 1792, the island was seen for two days wrapped 

 in fire and smoke. The smell of the smoke seemed to 

 indicate a forest and earth fire ; the voyagers did indeed 

 think they saw columns of smoke rising here and there 

 out of the ground near the shore ; but the naturalists 

 who accompanied the expedition came to the conclusion 

 that the phenomenon was at least not to be ascribed to 

 the eruption ( 503 ) of the high mountain as a volcano, 

 We might adduce as more secure evidences of a more 

 ancient and genuine volcanic activity in the Island of 

 Amsterdam, the beds of pumice (uitgebranden puim- 

 steen) mentioned already by Valentyn, according to 

 Vlaming's ship's journal in 1696, 



To the south-east of the extremity of Africa there 

 are Marion's or Prince Edward's island (47 2' S.) and 

 Possession island, belonging to the Crozet group (46 28' 

 S. lat., and 51 58' E. long.). Both show traces of 

 former volcanic activity ; they are small conical hills ( 504 ) 

 with eruption-orifices surrounded by columnar basalt. 



To the eastward, in almost the same latitude, follows 

 Kerguelen island (Cook's Island of Desolation), of which 

 we owe the first geological description to Sir James 

 Koss's antarctic expedition, so fruitful in successful 

 results. At the place called by Cook Christmas Har- 

 bour (48 41' S. lat., 69 04' E. long.), basaltic lavas se- 

 veral feet thick envelope fossil trunks of trees. The 



