50 . The American Geologist. January, 1903. 
islands, there is not merely a combination of late and ancient 
eruptive deposits, but there are several formations secondarily 
derived from the remains of the older basement, and here is 
room for more study than has been attempted. 
The history of one is more or less the history of all these 
conical islands. For instance, in Dominique there is the old 
andesitic rock overlaid by volcanic breccia or conglomerate. 
At other points the age of the tuffs cannot at present be as- 
signed, but some of them have been denuded into relatively 
large valleys which have been partly refilled with still newer 
tuffs (like that of the Roseau valley) containing an abundance 
of water-worn pebbles, often arranged in lines among the more 
angular material. Such may correspond to the early Tertiary 
sub-aqueous tufaceous beds of Grand Terre (Guadeloupe). 
And these beds have been subsequently tilted outwards at 
considerable angles. As in St. Martin, Antigua and Grand 
Terre, there is nothing to show that there were any mid-Ter- 
tiary eruptions when the whole region was somewhat elevated 
and the denuding agents were moulding the surface into 
rounded outlines. From the corresponding topography in the 
more volcanic islands, where not surmounted by the modern 
cones, the impression is left that the volcanic activity of the 
region was quiescent during much of the Miocene-Pliocene 
period, before the building up of the cones and ridges ; which 
were constructed at a relatively late date ; for we find the sea- 
bed elevated along with these ridges. Thus we find in Statia 
and in St. Kitts volcanic cones raised by an upward thrust 
which carried along with it the sea-floor covered by about 30 
feet of marl, now forming broken mantles surrounding the 
cones to elevations of from 400 to 900 feet. Elsewhere, how- 
ever, we find fragments of a similar formation appearing with 
the volcanic rocks brought up by a general elevation of the is- 
land. These limestone marls contain practically a living fauna, 
thus showing the elevation to date no farther back than the end 
of the Pliocene period. Again there are two series of gravel 
formations, one of which is older than the coralline strata just 
mentioned as interbedded with the volcanic ejectamenta; but 
this gravel formation had its surface greatly denuded before 
the formation of the marl. Again, both the marl and the gravel 
have been further subjected to erosion so as to often be only 
