THE COMPLETED ATOLL. 321 
one first meets with a stratum of sulphate of lime (sometimes 
compact and crystalline, sometimes soft and amorphous) fre- 
quently two feet thick, beneath which are successive strata 
of coral sand and shells, deposited one above the other in the 
gradual process by which the lagoon was filled up. These 
horizontal strata were penetrated to a depth of about twenty 
feet. They were composed chiefly of fine and coarse sand 
with an occasional stratum of coral fragments and shells. 
Of the origin of this sulphate of lime there can hardly be | 
any doubt. As the lagoon was nearly filled up, while, by the 
gradual elevation of the island, the communication between 
the outer ocean and the inner lake was constantly becoming 
less easy, large quantities of sea-water must have been evap- 
orated in the basm. By this means deposits would be formed 
containing common salt, gypsum, and other salts found in the 
waters of the ocean. From these the more soluble parts would 
gradually be washed out again by the occasional rains, leaving 
the less soluble sulphate of lime as we find it here. 
Some additional light is thrown on this matter by the dif- 
ferent parts of the surface, which, though nearly flat, shows 
some slight variety of level. The higher parts, particularly 
around the outer edges, are composed chiefly of coral sand, 
either mixed with or underlying guano. Nearer the centre © 
is a large tract, rather more depressed, forming a shallow 
basin, in which the bulk of the sea-water must have been 
evaporated, the surface of which (now partly covered with 
guano) is a bed of sulphate of lime, while, further, there 1s 
a still lower point, the least elevated of the whole, where the 
lagoon waters were, without doubt, most recently concen- 
trated. This latter locality is a crescent-shaped bed, about 
six hundred feet long by two hundred or three hundred feet 
wide, having a surface very slightly depressed from the outer 
21 
