Page:A Lady's Cruise in a French Man-of-War.djvu/290

258 that other group over which, also, I grieve as over a lost inheritance.

The Paumotus (i.e., cloud of islands)—or, as we used to call it, the Low or Dangerous Archipelago—is a cluster of eighty very flat coral-isles, most of which are of the nature of atolls, some shaped like a horse-shoe, others so nearly circular that only small canoes can enter the calm lagoon which occupies the centre; and some are perfect rings, having no visible connection whatever with the ocean, which, nevertheless, finds a subterranean passage through which the waters rush in a strong current as the tides rise and fall. Such lagoons as these are generally encircled by a belt of swamp, which can only be crossed by laying down pathways of long branches; these act in the same way as huge Canadian snow-shoes, and enable the light-footed natives to pass in safety across the treacherous green surface to the margin of the lake, where they keep small canoes in which to paddle about in search of eels and shrimps, and various kinds of fish. The water-supply is generally deficient, and only by sinking wells in the coral-sand can even brackish water be obtained. There are, however, exceptions to this rule, and some isles have good springs. But at best, the people depend greatly on their cocoa-nuts for drink even more than for food, and happily this supply rarely fails them. The coral-bed supplies neither soil nor water sufficient to raise any regular crops. Here and there a sort of caladium with edible root grows wild, but yams or taro are only known as imported luxuries. Forest-trees and scrub, however, contrive to find a living, and form a dense growth over most of the isles; and here and there clumps of carefully cultivated bananas, orange-trees, or bread-fruit, tell of a richer and deeper soil, probably accumulated with patient toil. Fig-trees and limes also flourish.

Some of the lagoon-reefs have a diameter of about forty miles, and only rise above the water in small isles, forming a dotted circle.

I believe it is generally supposed that such coral-rings as these have in many cases been the encircling reef formed round some volcanic isle, which has gradually subsided, whereas the