421 As the birds are accustomed to visit different islands when the wind is favourable, the people send by them small presents (fish-hooks, Arc.) to their friends. Christian mis sionaries also occasionally use them as letter-carriers for communicating with one another. Elevated Coral Islands. There are comparatively few of the elevated coral islands in Polynesia, but they are so distinct from both the atolls and the volcanic islands that they need a separate description. They all lie within or near the lines marking off the volcanic ridge upon the map. South of the volcanic ridge there are many coral reefs forming shoals. The elevated coral islands doubtless were once such reefs. Lying within the area of volcanic action, they have participated in the upward movement, and have been raised from shoals to become islands. Some have evidently been lifted by successive stages and apparently by sudden movements. This is clearly seen in the Loyalty Islands. On approaching them one sees high coral cliffs, in appearance much like the chalk cliffs of England, except that they are often some distance inland and not close on the shore. The island of Mare may be taken as a good type of the class. Here, between the shore and the coral cliffs, there is a tract of level land varying from a few yards to perhaps one-fourth of a mile or more across. On this level tract the people mainly dwell. At the back of this there rises a perpendicular wall of coral, in some places as much as a hundred feet high. The cliff is water-worn, and has in it large caverns, showing that for a long period it was the coast-line. Still farther inland there are two similar though smaller cliffs, indicating that there were three distinct upheavals. These must have been at very long intervals. At present the island is fringed with a coral reef, and if it were now to be lifted from fifty to one hundred feet the present coast-line would form another cliff, while the present coral reef would form another low plot similar to that upon which the people now dwell. These islands are old enough to have a considerable depth of vegetable soil upon them. The low land between the coast and the first cliff is well stocked with cocoa-nut and other trees. None of the islands can be compared with the volcanic islands for fertility, all having a less rich soil and being much drier; still they are fairly fertile. They suffer sometimes from drought, but are much less seriously affected in this way than the atolls. The flora of the elevated coral islands is less rich than that of the volcanic islands, but much richer than that of the atolls. The island of Niue may be taken as a fair specimen of this class. Its flora probably contains between 4 00 and 500 species, nearly all being such as are found on adjacent volcanic islands. The fauna is also much richer than that of the atolls, but poorer than that of the volcanic islands. Birds are numerous. While most of the species are identical with those found in neighbouring volcanic islands, there are some interesting local variations well illustrating the modifications which take place from isola tion under changed surroundings. In some instances the differences are so great that local forms have been classed not only as varieties but as distinct species. Climate. The climate of the islands varies considerably, as may be naturally expected when the wide area covered is remembered, and the vast difference there is between the islands themselves. Some, especially the elevated coral islands, are very healthy for tropical regions. Speaking generally, the average reading of the thermometer over a large extent of Polynesia is about 80 Fahr. It very seldom sinks lower than 60, and, owing to the small size of most of the islands, and the prevalence of trade-winds during the greater portion of the year, the heat is always moderated, and rarely becomes intense. Yet, owing to the constant heat and to the humidity of the atmosphere, the climate in the mountainous islands is trying to the European constitution. But in this respect there is a great difference even between groups which, looked at superficially, appear to be similar, and which lie within almost the same parallels of latitude. All the islands eastward from and including Fiji are much more healthy than are those to the west. In the eastern section fever and ague are of rare occurrence ; in the western section European missionaries do not find it expedient to remain for long periods on the islands owing to the weakening effects of frequent attacks of these diseases. The most remarkable thing is that natives of the eastern section suffer even more than Europeans when they go to live in the western islands, the mortality among them being very great. Numerous attempts have been made to evangelize the New Hebrides through the agency of natives of the Samoan, Cook, and Society groups; but, owing to the great mortality among the agents, their efforts have failed. Yet these people have lived there under condi tions very similar to those they were accustomed to at home, the heat being about the same, and the food similar, as well as the general mode of life. The causes of the difference are as yet unknown. Possibly the explanation will be found in differences of natural drainage. It has often occurred to the present writer, though only as an unverified theory, that the bases of these western islands are, like that of New Caledonia, of older formation, and that the islands are only superficially volcanic. If so, this may account for their unhealthiness as compared with the purely volcanic islands within the same parallels of lati tude. In comparison with most tropical countries there is little dysentery in Polynesia ; but this also is more com mon in the west than in the east. The elevated coral islands are always much more healthy than are those of volcanic formation in their immediate neighbourhood. They are drier, being always well drained, have much less dense vegetation, and receive the benefit of the trade-winds which blow right across them. They, however, sometimes suffer from drought such as is unknown on the volcanic islands. The atolls may be called if the term can be applied to tiny islets scattered over the expanse of ocean the deserts of the Pacific. The soil being almost entirely sand, and the vegetation afford ing little shade, the heat and glare, especially of those lying close to the equator, are exceedingly trying to European visitors. Being so low only a few feet above the ocean there is nothing to attract the clouds, and the rainfall is small. The islands are therefore subject to frequent droughts, which are sometimes of month s dura tion ; and at such times even the fronds of the cocoa-nut palm get a shrivelled appearance, and the trees cease to bear fruit. Sometimes the people suffer greatly during these long-continued droughts, many being starved to death. At best their food supply is confined to cocoa- nuts, pandanus, fruit, and fish, but in times of drought they are forced to chew the roots of shrubs. Hurricanes. A great portion of southern Polynesia is subject to destructive cyclones. The tract over which they pass may be said to be, generally, that of the volcanic chain indicated by the lines on the map, although the northern edge of this region is not so subject to cyclones as the southern portion. A line drawn parallel to the lines of the ma]i, through the middle of the New Hebrides group, and extending south of Fiji, will well represent the centre of the cyclone tract. The hurricane season is from December to April. Some islands are visited by a more or less destructive cyclone nearly every year ; Samoa lies on the upper edge of the tract, and gets one, on an average,
about every seven or eight years. Although these cyclones