420 POLYNESIA its home on New Caledonia ; and in Samoa the Didunculus strijirustris has its habitat. This bird is remarkable as being the nearest relative of the extinct dodo. Some time ago it was rarely found, and was becoming extinct. It fed and nested on the ground, and was destroyed by cats and rats after they were introduced. Of late it has changed its habits : it now feeds, nests, and roosts upon trees, and is, in consequence, increasing in numbers. Certain non- venomous snakes are found in many of the islands. Insect life is abundant, and some of the butterflies are very beautiful. The lagoons formed by the coral reefs around the islands invariably abound in fish, many of them most gor geous in their colouring, vying in this respect with the parrots of Australia. Fish form a very important part of the food supply. One of the most wonderful creatures in the marine fauna of Polynesia is the palolo (Palola viridis), an annelid which appears upo n the surface of the ocean, near the edge of the coral reef, at certain seasons of the year. The palolo are from 9 to 18 inches long, and about th of an inch thick. They are eaten by the natives, and are esteemed a great delicacy. They live in the inter stices of the coral reef, and are confined to a few localities. About 3 o clock on the morning following the third quartering of the October moon they invariably appear upon the surface of the water ; generally they are in such quantities that they may be taken up by the handful. Soon after the sun rises they begin to break, and by 9 o clock A.M. they have broken to pieces and disappeared. The morning following the third quarter of the November moon they again appear in the same manner, but usually in smaller quantities. After that they are not again seen until October of the next year. They appear thus to deposit their ova, which is done by the break ing to pieces of the female worms ; the males also break in the same manner, the ova being fertilized while floating in the water. Thus the parents are destroyed in propagating their species. The eggs gradually sink down to the reef where they are hatched. The young papolo then live about the reef until the next year, when they repeat the process. Year by year these creatures appear according to lunar time. Yet, in the long run, they keep solar time. .This they do by keeping two cycles, one of three years and one of twenty- nine years. In the short cycle there are two inter vals of twelve lunations each, and one of thirteen lunations. These thirty-seven lunations bring lunar time somewhat near to solar time. But in the course of twenty-nine years there will be suffi cient difference to require the addition of another lunation ; the twenty-ninth year is therefore one of thirteen instead of twelve lunations. In this way they do not change their season during an entire century. So certain has been their appearance that in Samoa they have given their name to the spring season, which is vae- pilolo, or the time of palolo. 1 The Atolls. The atolls differ in almost every respect from the islands of volcanic origin. Little that is said of one class would be true of the other. These coral islands are all low, generally not more than 10 or 12 feet above high-water mark. They are simply sandbanks formed by the accumulation of debris washed on to the reefs during .strong winds. Hence they are usually in the shape of a narrow band, varying from a few yards to one-third of a mile across, near the outer edge of the reef, with a lagoon in the centre. In some of the smaller atolls the circle of land is almost or entirely complete, but in most of those of larger dimensions there are breaks to leeward, and the sea washes freely over the reef into the lagoon. Where the circle of land is complete the sea-water gains access to the central lagoon through the reef underneath the islands. In some it bubbles up at the rise of the tide in the midst of the lagoons, forming immense natural fountains. This has been observed producing a specially fine effect at Nui in the Ellice group. Some of these atolls are not more than 5 or 4 miles in their greatest length. Others are many miles long. They are not all circular, but are of all con ceivable shapes. 2 1 For fuller details, see article by the present writer in Proc. Z lftl. Soc. of Loud., 1875, p. 496. - On the formation of atolls and of coral reefs generally, see t oHALS, vol. vi. 377, and PACIFIC, vol. xviii. p. 128. Two of the atolls known to the present writer are remarkable. The lagoons in them are of fresh water. One of these is Lakena in the Ellice group, the other Olosenga, or Quiros Island, in 11 2 S. lat. and 171 W. long. Both are small circular islands, and in both the lagoon is shut off from the sea. Olosenga is less than ! miles in diameter, the lagoon occupying over 3 miles, leaving a ring of land around it less than half a mile across. In some places the lagoon is at least 6 fathoms deep. This bulk of fresh water cannot, therefore, be the result of drainage. There is much to favour the opinion that both this island and Lakena are situated over the craters of former volcanoes, and that there is submarine connexion between them and some of the larger islands situated on the volcanic ridge from which the body of fresh water must come. Olosenga is about 200 miles distant from Samoa. In that group mountain streams sometimes fall into chasms and totally disappear under ground. In this way subterranean lakes may be formed in some of the cavities which we may suppose volcanic eruptions to leave. It is not difficult to suppose that there would be subterranean connexion between these lakes and an isolated crater 200 miles distant. If so, as the crater participated in the subsidence of the region on the edge of which it is situated, the water would rise in it until, if the supply were sufficient, it there found an outlet. This appears to be what occurs at Olosenga. The lake has never been properly examined and sounded. It is, however, of considerable depth in the centre, where the water is said sometimes to bubble up as if from a great spring, and at low tide it is seen to percolate through the sand on the outer or sea side of the land. The vegetation of the atolls is extremely poor, not more than about fifty species of plants being found in the Tokelau, Ellice, and Gilbert groups, in all of which groups collec tions have been made. All the species consist of littoral plants found in the volcanic islands. Most of them have their seeds enveloped in thick husks, which specially fit them for being carried by currents. Doubtless it is in this way that the atolls have received their flora. The cocoa-nut is abundant on most of these islands. This most useful palm will grow on any sandbank in the tropics, and it is benefited by having its roots in soil saturated with sea water. Unlike the natives of volcanic islands, those dwelling on the atolls eat the raw kernel of the nut in large quantities. Indeed that, with fish and the fruit of a screw-pine (Pandanus), constitutes the main food supply on some atolls. The people make the palp of the pandanus into a kind of cake, in appearance much like a quantity of old dates. In some atolls a somewhat elaborate system of cultivation has been adopted, by means of which a coarse kind of taro, banana, the bread-fruit, &c., are grown. These low islands suffer much from drought, and the natural soil is nothing but sand. The people, therefore, form wide trenches by removing the sand until they get within about 2 to 3 feet of the sea -level. Into the trenches they put all the vegetable refuse and manure they can obtain, and, as there is more moisture at this level, those excavated gardens are comparatively fertile. Under the influence of a Christian civilization, which is growing, and by the introduction of new food-producing plants, the condition of the natives is improving; but they still suffer much at times from long-continued seasons of drought. The fauna of. the atolls consists mainly of a few birds, some lizards, and insects. Fish abound about the reefs, and most of the natives are deep-sea fishermen. In the Ellice Islands the people domesticate frigate-birds. Large
numbers of these pets may be seen about the villages.