THE COMPLETED ATOLL. 317 
been observed at the Gilbert Group, and also at Enderbury’s 
Island, and many other coral islands in the Pacific. The 
stones at the Gilbert Islands, as far as could be learned, are 
generally basaltic or volcanic, and they are highly valued 
among the natives for whetstones, pestles, and hatchets. The 
logs are claimed by the chiefs for canoes. Some of the logs 
seen by the author, like those of Enderbury’s Island, were 
forty feet or more long. Several large masses of compact 
cellular lava occur on Rose Island, a few degrees east of the 
Navigator Group; they were lying two hundred yards inside 
of the line of breakers. The island is uninhabited, and the 
origin of the stones is doubtful; they may have been brought 
there by roots of trees, or perhaps by some canoe. 
Fragments of pumice and resin are transported by the 
waves to many of the islands in the Central Pacific. We 
were informed at the Gilbert Islands that the pumice was 
gathered from the shores by women and pounded up to fer- 
tilize the soil of their taro patches; and that it is common 
for a woman to pick up a peck a day. 
Where this pumice comes from is not ascertained. It is ‘ 
probably drifted from the westward, and perhaps from vol- 
canic islands of the Ladrones or Phillipines. In addition, 
volcanic ashes are sometimes distributed over these islands, 
through the atmosphere. In this manner the soil of the 
Tonga Islands has been improved, and it may thus have de- 
rived its reddish color. This group has its own active vol- 
cano to supply the ashes, and the volcanic group of the 
New Hebrides is not far distant to the southwest. 
The mineralogy of an atoll is usually confined to one 
species, — calcite, or carbonate of lime,— the material of 
the coral rock. On some of the smaller islands, in the drier 
equatorial part of the ocean, there are, in addition to this 
