feet on the cliff, but grows narrower westward and appears to 
die out entirely between 200 and 250 feet from the cliff. It is 
probable, however, that the tuff really persists until it passes 
the edge of the overlying melaphyr, and is merged with the next 
bed of tuff, as shown on the map. 
The boundaries of this tapering edge of the fragmental bed 
are quite irregular. This is especially true of its lower surface, 
where it rests upon the second melaphyr. The form of this 
contact for 130 feet west from Valley Beach, or as far as it is 
clearly exposed, is shown in Fig. 4; for the sake of convenience, 
however, the curves are represented as following each other 
more nearly in the same direct line than is actually the case. 
The upper surface of the melaphyr presents smoothly rounded 
hemispherical protuberances or swellings one to three or four 
feet in diameter. The actual boundaries of adjacent protu- 
berances may usually be traced below the surface of the mela- 
phyr ; and some of these remarkably regular and graceful curves 
thus describe two-thirds or three-fourths of a complete circle. 
Probably the majority of the curves are really sections of rounded 
ridges or rolls, the surface flow-structure of the lava. But it 18 
very noticeable that these prominences exhibit semicircular pro- 
files to some extent on both the transverse and longitudinal &ec- 
tions of the bed, suggesting hemispherical tumefactions or 
superficial bubbles formed on the liquid lava. 
It seems impossible to regard these features as the product of 
erosion ; they must be entirely original ; the actual surface of 
the fresh submarine lava-flow, which was covered almost in- 
tact either by the ashes accompanying a second flow or by the 
débris worn from adjacent masses. The fragments are mainly 
angular, ranging in size from dust to three inches in diameter; 
and this is clearly a thin, local deposit on the surface of the 
sheet of lava. 
The well-marked north-south depression separating Atlantic 
and Centre Hills, and terminating at the lower end in Valley 
Beach, does not seem to be oceupied by adike, for the melaphyr 
shows half way across the bottom of it, and there is not a ves- 
