540 PEOF. J. W. SPENCER ON THE GEOLOGICAL AND [NoV. I9OI, 



Mr. Watt, of Antigua. It resembles the White Limestones of the 

 other island. Among the collections sent by Nugent to London, 

 Duncan ^ found the coral Astrcea Antillarum closely allied to 

 A. endotheca, which is an old Miocene type occurring in Antigua 

 and Santo Domingo. Thus in Montserrat, there is a fragment of 

 the same geological foundation, of the recent topographical features, 

 as in other adjacent islands. 



YIII. Eeosion-Peattjees. 



The sunken tableland now forming the Saba Banks and the chain 

 of islands from Saba to Montserrat, rises above the floor of the 

 submarine Antillean plateau to a height of 2000 feet or more, in 

 the same manner as the other groups of the north-eastern Antilles, 

 which are regarded as the remains of a vast tableland dissected in 

 the earlier part of the Miocene-Pliocene Period. But the information 

 derived from erosion-phenomena upon the surface of these islands 

 is of the most fragmentary character, as the vast Saba plain 

 and the corresponding foundation of the volcanic islands have been 

 either submerged or covered by more recent ejectamenta, leaving 

 only the limited, central, base-level peneplains and the northern 

 hills of Statia, the south-eastern hilly part of St. Kitts, and frag- 

 meuts of Montserrat, comparable with the mountain-districts of 

 Antigua and St. Martin. Consequently, we are left to consider 

 only topographical outlines more recent than the Miocene-Pliocene 

 Period ; and so far as the surface is concerned, more recent than 

 even the period of great elevation of the early Pleistocene Period, 

 which was characterized by very deep valleys, dissecting the more 

 rounded outlines of the earlier topography. However, upon the 

 submerged margins of the area under consideration occur some 

 interesting erosion-features. Thus there is an embayment indent- 

 ing the sunken ridge between Saba and Statia, reaching to a 

 depth of more than 2562 feet, while the adjacent part of the sea- 

 floor is only about 1200 feet below the surface.^ The channel 

 over 2000 feet deep, between Saba and Saba Banks, is not pro- 

 bably due to erosion, but to the elevation of the volcanic cone at 

 the toot of the tableland now submerged. The channel, reaching 

 to a depth of 2592 feet, between Nevis and Eedonda, dissects the 

 sea-floor (which here generally reaches to about 1800 feet), and is a 

 repetition of the valley-feature just mentioned. The margins of 

 the Saba Banks are also indented by cirques. The shelf, norlh-east 

 of Statia, submerged to less than 150 feet, which is apparently 

 covered with a limestone-floor like that shown at Brimstone Hill 

 and in Statia, is indented by an amphitheatre having a depth of 

 more than 1476 feet. This phenomenon suggests that tlie epoch 



1 'On the Fopeil Corals of the West Indian Is.' Quart. Journ. Geol. Soc. 

 vol. xix (1863) p. 452, and following papers. 



2 See U.S. Hydrograpbic Chart ]No. 40, or the corresponding British- 

 A( niiraHy Charl, and also a chart on a still larger scale. 



