176 A. FE. Verrill-—-The Bermuda Islands; Geology. 
rotary motions, loosen the sand about themselves and thus cause the 
wind to excavate conical or cup-like cavities. Such cavities, if pre- 
served by clay ‘soil washing into them by rain, might be the starting 
points of deep cavities excavated by solution. 
Whatever the cause may have been, in particular cases, such shal- 
low cavities, if in a calcareous soil covered by clay, must have been 
filled by the red clay washed in by rain. The clay core on drying 
would shrink away from the surrounding materials, leaving a nar- 
row crevice about it into which rain water would percolate and 
slowly dissolve away the surrounding limestone, redepositing part 
of it a little farther away, as the moisture evaporated. The enlarged 
crevices would be filled, again and again, by additional clay material, 
and so the clay core would be increased in size and length as the 
solvent action went on. Thus there would be no definite limit to the 
depth or size of the cavities, provided the time were very long and 
no insoluble obstructions were encountered. The ordinary effects of 
gravitation account, in this theory, for the extension downward 
being most rapid. The presence of clay deposited on the sides of 
the cavities accounts for the water not spreading much laterally in 
the more porous layers. 
A similar effect may be seen when pebbles rest upon porous ice 
or snow in sunny weather. The ice melts away under and around 
the stone, but mostly beneath, so that the stone soon sinks into a 
hole but little larger than itself. 
That the solvent action referred to will result in forming circular 
pits may be demonstrated experimentally by resting balls or cylin- 
ders of clay on shallow indentations in pieces of limestone and allow- 
ing very dilute acids to trickle very slowly over the surface of the 
clay, so that the solution will evaporate almost as soon as formed. 
The best examples of these structures that I saw were near 
Hungry Bay on the south shore. See plates xix, xx. At this place 
there is a bench of hard limestone, believed to be of the Walsing- 
ham formation (see above, pp. 62, 72), just above high-tide, which 
has been quarried for building purposes. So that good sections of 
some of the cavities have been made, as in pl. xix.* The upper sur- 
face of this limestone is partially covered with indurated red clay, 
the softer parts of the clay stratum having been worn away by the 
sea. This surface is perforated by a large number of these cavities, 
most of them nearly round and a foot or more in diameter. If due 
* See also ‘‘ The Bermuda Islands,” plates lxxxiv-v, and these Trans., vol. Xi, 
same plates. 
