Much of the literature of the Pacific Islands contains some mention of a native rat, described as living in the bush or infest-ing the houses and feeding upon vegetables and fruit, but for the page 166most part no scientific description of the animal is attempted, nor is reference made to previous records in other islands. Although the rat is frequently mentioned, it has not in all cases been thought of sufficient interest to be indexed, and therefore many possible records are not apparent. The geographical distribution of the Pacific rat is so wide, and therefore of such interest, that I have thought it wise to include all definite localities met with during casual reading, to form a basis on which to build.

Before doing so, however, some notice of its identity is necessary.

Identity.

Apart from the Maori rat, the only technical notice appears to be that by Peale,^18* who named (and figured) rats obtained from widely separated islands as Mus exulans, In this con-nection it may be mentioned that the Editor of the second edition of the work cited, remarks that he is not without suspicion that the animal is either Mus pencillatus, Gould,^† or Mus jacobice, Waterh.^‡ There is, however, small likelihood of the Pacific rat being identical with either of these species, and indeed Thomas,²² by adopting Peale's name, has. practically decided that it is distinct. His interesting note reads as follows:—

"The Rats from Sunday Island, Kermadec group, apparently belong to a species widely spread over the Pacific, the earliest name of which seems to be Mus exulans, Peale, based on Fijian examples. It is possible that examples from the different groups of islands may hereafter show certain differences from each other, but, so far as we can see at present, all should be united under one heading. Indeed the fine Maori Eat of New Zealand (Mus maorium, Hutton) seems to be very doubtfully separable from the same form, which has probably travelled from island to island in native canoes, or on floating logs &c., long before European ships began to bring over the ubiquitous Grey and Black Rats, which now threaten to exterminate the native species throughout the world."

It will be remarked that Fiji is not included in the localities enumerated by Peale at which Mus exulans was obtained: for rats from this group that writer proposed another name—Mus vitiensis; there can be little doubt, however, that notwith-standing the slight differences mentioned, the two forms are not specifically distinct.

page 167

All circumstances being taken into account, it appears probable that the Maori rat is also identical with this widely distributed Pacific species, and in one of his papers Hutton¹³ has pointedly remarked:—"It will be interesting to compare these skulls with specimens of the black rat^* from Polynesia, for they will probably be found to be identical." And again, writing on Mus novce-zealandice, Buller, he adds¹⁴ "There can, I think, be no doubt that these rats belong to the Polynesian variety." More recently Thomas has also expressed doubts as to the specific identity of the Maori rat, in the note previously quoted, and as mentioned by Buller,⁵ who further remarks that there are specimens of this form in the British Museum from the Fiji Islands, Norfolk Island, and New Caledonia. This view is supported by Maori tradition as related by Hochstetter,¹² to the effect that:—"the Kumara, or sweet potato (Convolvulus batata), the taro (Arum esculentum), the calabash-plant Hue (Lagenaria vulgaris), the Karaki tree (Gorynoearpus Icevigata), the rat Kiore, the Pukeko (Porphyrio), and the green parrot Kakariki, are said to have been imported from Hawaiki." This traditional ancestral home is considered by modern Ethnologists to be Savaii, one of the Samoan Islands.

The New Zealand rat has a literature to itself, which will be found mainly in Trans. and Proc. N.Z. Institute. This literature I have not attempted to epitomise, and have referred to it only for odd records of habits. There is apparently still room for research among the New Zealand rats. The Kiore rat is said to be extinct, the Mus maorium to swarm, fide Meeson,¹⁷ Rutland,¹⁹ etc.

Distribution.

If, as seems probable, the rat from all the Pacific Islands is referable to Mus exulans, the range of the species is very great indeed. Considering the native interchange which has taken place between islands hundreds of miles apart for ages past, this is not so remarkable as would at first sight appear.

For a long distance in the West Pacific there runs an enormous chain of islands, extending in a semi-circular sweep from the Marshall Archipelago, north of the equator to the Austral or Tubai Islands in the south-east. Our colleague has written of this as the Marshall-Austral chain, and dealt with it more particularly in his report. ^†

From each of the main links of this long chain of islands, we possess records of the occurrence of a native rat, as below enumerated.

page 168

Wake Island, an isolated atoll, which I would regard as an extension of the chain, is recorded by Peale,¹⁸ and is at the same time the most northerly and westerly (with New Caledonia) rat-inhabited island of which I have notice. Passing southward and westward the rat next appears to have been observed at Odia in the Marshall Group, and is represented by Kotzebue¹⁵ in an illus-tration as impudently trespassing in a Marshall Island house.

Continuing the chain, the rat is recorded from the Gilbert Group by Woodford,²⁵ who remarks that the only wild mammal he met with was a small species of rat common to the islands "in this part of the world."

The next group is that of the Ellice, of which the island of Funafuti at least, is tenanted, and supplied the examples, to be more fully described, and which prompted the present essay.

Mention has previously been made of Savaii being the traditional ancestral home of the Maori rat, but further evidence of its occurrence in Samoa is indicated by the reference to it in ancient tradition detailed by Turner,²³ and direct evidence is afforded²⁴ by this writer in the following note:—"The only indigenous quadruped is a small rat, something between a mouse and the Norwegian rat, the latter of which was introduced some years ago."

The last in the direct chain to which I have reference to the rat is the Cook Group, its occurrence being mentioned at Raratonga and Mangaia by Gill.^{9 &10}

Of localities to the east of the main chain the following have been published. In the Phoenix Group Peale¹⁸ records it from Hull Island, and Arundel² from Sydney Island. Much further to the east it has been met with by Dixon⁷ at Maiden Island, and also further to the south by Lamont, as quoted by Smith²¹ at Penhryn Island, and Dixon⁸ at Caroline Island, all isolated atolls. In the Paumotu Group or Low Archipelago Peale again records it from Disappointment and Dog Islands, and also from the Society Islands, remarking that the species was seen on but one "high" island, Tahiti.

Its north-eastern limit is suggested by a statement by Brigham³ that "Rats and mice have always been a pest on the Hawaiian Islands; and the old Hawaiian, before the introduction of cats, used a bow and arrows to destroy them. It is curious that knowing the principle of the bow they never used it as a weapon of offence, nor developed it beyond a very feeble instrument only suited to the killing of 'rats and mice and such small deer.'"

To the westward of the main chain Allardyce¹ records it from Rotumah, and it is once more mentioned by Peale from Fiji as Mus vitiensis, and from Hoonga in the Tonga Islands by page 169Mariner.¹⁶ The occurrence of a rat in the Kermadec Islands was first recorded by Smith,²⁰ who wrote:—"The only animal native to the island is a small grey rat, which is very plentiful in summer, but is supposed to hybernate during the winter. We saw one that had been partly eaten—by a hawk probably,—it was about five inches long." Thomas also received it from Sunday Island in this group, as already quoted. Away to the west it appears in New Caledonia, and again at Norfolk Island on the authority of Buller,⁵ who states that there are specimens in the British Museum from these localities.

The list of localities is closed by the inclusion of New Zealand as the most southern limit, and to which previous mention has been made in notices by Hutton, Thomas, Hochstetter, and others.

Although a systematic search of the literature of the Pacific Islands would doubtless disclose many more references to the rat, the above are the only definite localities I have so far met with. There is little doubt that the rat exists, or rather did exist, at one time or another on all the islands of the Pacific. Gill⁹ writing in 1876, and mentioning the islands of the South Seas as being inhabited by dogs, hogs, and rats, says:—"The rat alone is universal."

Arundel,² who called at many of the atolls in the Central Pacific, states:—"I have never visited an island, however small or barren, without finding these animals living upon it."

Habits.

Unlike its European relative, the Pacific rat is usually said to feed only on vegetable substances. "Writing of Mangaia, in the Cook Group, Gill⁹ states that it feeds exclusively upon cocoanuts, bananas, arrowroot, candle nuts, and papao (pawpaw) apples, and that it was usual to defend growing cocoanuts from the depredations of the native rat by making a sort of screen cleverly secured all round the tree, close to the fronds at a great height from the ground. In Mariner's¹⁶ book the rats are described as living chiefly upon such vegetable substances as sugar cane, bread fruit, etc., and it is incidentally mentioned that roasted cocoanut was used as a bait.

Peale¹⁸ adds the Pandanus to this list, and states that the fruit of this plant forms the principal food of the rat, hazarding the suggestion that if its appetite was at all carnivorous it would be found to feed upon the land crabs and molluscs on the shore, such however not being the case. He describes it (Mus vitiensis) as attacking pockets and packs containing edibles.

The Kiore Maori is described by Dieffenbach ⁶ and others as being a frugivorous rat.

page 170

Rutland¹⁹ writes of the New Zealand bush-rat:—"Considering the vast numbers of these rats that periodically congregate round the homes of settlers in the bush, the mischief done by them is extremely small. This is owing to their food during the time being green vegetables. In kitchen gardens they are certainly annoying, devouring peas, beans, cabbages, and even onions, as they appear above ground, climbing up poles to nip off the shoots of the vines, etc."

Of Sydney Island Arundel² ascribes a partial animal diet to them, writing:—"Before any settlements are formed they live in the ground and roots of trees, and subsist on young birds, birds' eggs, seeds, etc. As soon, however, as anyone comes to live on an island they gather round the settlement, particularly round the native quarters, natives being, as a rule, rather wasteful in their eating, and scattering round about them rice, bread, pieces of fish, etc."

If the native rat preceded the human inhabitants of the atolls, the pandanus, being indigenous, would probably be its staple food, and as the cocoanut, breadfruit, arrowroot, etc., were intro-duced, the rat would acquire a taste for these articles.

As to its nesting habits the accounts are somewhat varied, Peale describes it as constructing a nest in the tussocks of grass, and making shallow burrows like an Arvicola. He describes Mus vitiensis as being a great pest in most of the houses of the Fiji Islands, making its nest in the thatched roof. Being an excellent climber it sallies forth at night in such numbers as to be exceedingly troublesome. Gill⁹ relates an instance of a pair having made a nest within a mummy conserved in a cave.

Of Caroline Island Dixon⁸ writes:—"The brown rat has a foot-hold, but is not numerous. Their nests were made in the cocoanut trees, just at the base of the fronds." Our colleague understood that it nested in similar situations in Funafuti.

In New Zealand, too, Rutland ¹⁰ records how nests, evidently of rats, were found in the crowns of tree ferns and also under the roots of trees and among rushes. This writer describes the rats as being awkward on the ground but extremely active when climbing trees, ascending with the nimbleness of flies and running out to the very extremities of the branches. Hence, he adds, "when pursued they invariably make to trees if any are within reach." Peale mentions a similar habit in connection with the rats recorded by him.

In Tonga, Mariner¹⁶ describes it as being an inhabitant of the bush, writing:—"Every now and then the natives make a peculiar noise with the lips, like the squeaking of a rat, which frequently brings them out of the bushes."

In Mangaia, as mentioned by Gill,⁹ and as previously recorded, rats inhabited the mountain fern, whence they were occasionally page 171driven by fire, Arundel² describes the rats of Sydney Island as naturally living in the ground and roots of trees, but gathering round the dwellings as soon as a settlement is formed.

As elsewhere, the great enemy of the native rat is the common brown rat of Europe, introduced by ships throughout the world. Its depredations are such that Gill states that in many of the islands the indigenous breed has been exterminated by the imported rat. Some idea of the successful war waged by the introduced rat may be gathered from the following graphic account by the same writer⁹:—" In 1852 a solitary male Norway rat got ashore at Mangaia from the wreck of an American whaler. It made war upon the native rat, so that one of the bed-rooms of the mission-house became uninhabitable. On removing the flooring about thirty dead native rats were found. We caught the offender in a trap."

Writing of Raratonga, another island of the Cook Group, the same author¹¹ incidentally records how the native rat has been subjected to even more deadly onslaught, being almost exter-minated by the domestic cats which, originally introduced by missionaries and afterwards emigrating to the bush, took to hunting birds when rats became scarce.

On p. 59 of the present Memoir we read:—"Cats have long been introduced, they are known to the natives by the name of 'pussy,' and have proved of service in destroying the brown rat, formerly a great pest to the Islands." Dieffenbach,⁶ writing on New Zealand, states that the cat often runs wild and is another cause of the extermination of indigenous animals.

The natives themselves destroy the rats: first, as vermin; second, shooting them for sport; third, killing them for food.

When unchecked, rats became very numerous on some of the islands. Writing of Sydney Island, Arundel² mentions how on moonlight nights he has often seen hundreds of rats gathered together round the native quarters feeding upon waste rice, bread, pieces of fish, etc., thrown out. He adds that they frequently caught one hundred a night in tubs made into traps in the store.

In Mangaia they were also numerous, for Gill⁹ states that, like most of the Pacific Islands, it was literally overrun with rats, and describes how a large bottle-shaped hole was dug in the earth and baited with candle-nuts, adding that when the hole was pretty well filled with rats, two men would go down with knobbed sticks to kill them. A hole which would contain two men would accommodate a goodly number of rats ! If the Mangaian rats were equally vicious with those mentioned by Peale, rat-killing under such conditions would not be unattended by danger, for he states that the animal resists pertinaciously and bites severely.

page 172

The reduction in the number of rats was a matter of such importance with the inhabitants, that we find a number of ingenious traps were in use for the purpose; these will be treated of in the Ethnological Report.

We have mentioned that sport may constitute a second way in which the rat is subject to persecution by the natives, Mariner¹⁶ has given an exhaustive account of the sport of "fanna gooma" or rat shooting, as practised on the island of Hoonga in the Tonga Group, from which it appears that it was an amusement in which only chiefs were permitted to participate, and was undertaken with much ceremony. The rats attracted by bait previously distributed, were shot with unfeathered arrows six feet long, projected from bows of similar length. The game was a party and not an individual affair, the party first killing ten rats was accounted the winner. If, Mariner adds, there be plenty of rats, they generally play three or four games. For a full account of the rules of the game the reader is referred to Mariner's book, which contains much of interest about the Tonga Islands. In Honolulu, as mentioned by Brigham,³ the bow was exclusively devoted to "killing rats and mice and such small deer."

The third reason for the native destruction of rats is of greater interest, and may be more fully mentioned.

In many of the islands of the Pacific the native rat formed an article of food with the inhabitants; feeding upon fruit or vegetables it would be less objectionable than the omnivorous European rat, and indeed Buller⁴ remarks that:—"Unlike the common rat, the rat of New Zealand is perfectly free from odour of any kind, probably due to the nature of its food, this consisting almost entirely of fruits and berries." The introduced rats were nowhere eaten: it may be that they were considered to be unpalatable, but it is equally possible that at the time they obtained a footing on the islands, pigs and other edible animals would also be introduced, and the necessity of eating rats removed. These native rats must have been considered good eating, for Grill,⁹ writing on the Cook Islands, states:—" The proverb 'sweet as a rat' survives in Mangaia to this day, although the adults of this generation have given up the disgusting practice of rat-eating."

This prolific and entertaining writer¹⁰ has given a valuable historical account of the capture and cooking of rats as practiced in Mangaia: it may be epitomised as follows:—

"In those days—ere the cat had been introduced—rats were very plentiful. Rat-hunting was the grave employment of bearded men, the flesh being regarded as most delicious. The rat, though but slightly larger than the English mouse, was the only quadruped on the island.

page 173

"Tamangoru, a solitary cannibal, on one occasion discovered two boys roasting a number of rats over a fire,—a joyful sight for a famishing Mangaian, he ambiguously remarked, 'cooked rats are capital eating.' The word 'rats' thus used might apply to the lads as well as to the little quadrupeds. A cooked boy would be indifferently called a 'fish' or a 'rat.'

"These two brothers subsisted chiefly by rat-catching, in which they were adepts.

"On the previous evening they dug a deep hole in the earth and covered the bottom of it with candlenuts, of which rats are excessively fond. A narrow pathway was made on either side for the rats to get down and eat. The lads lay in wait at a little distance, until they thought that the hole must be pretty full. Each lad carried a lighted torch in one hand, and a stout iron-wood stick in the other. They quickly killed a large number of rats.

"The boys now made a fire to roast the spoil. They then thrust long green reeds (previously prepared) through the rats, eight on each reed, and grilled them over the fire. There were four skewers or reeds of rats, that is, thirty-two in all. When the rats were done, the elder took two reeds of rats (sixteen) to Tamangoru; the famished man greedily devoured them and called for the remaining two reeds."

The same author⁹ informs us that in the neighbouring island of Raratonga, rats were not eaten, the inhabitants reviling the natives of Mangaia as the rat-eating Mangaians.

It would, however, appear that rats were not eaten when fish was procurable, for Gill relates how, when the sea was too rough for fishing, the boys set fire to the mountain fern, so that the rats rushing out of the fern, half blinded with fire and smoke, were easily killed with long sticks.

In Tonga (Hoonga Island) the rats formed an article of food with the lower orders of people, but in the account above referred to, Mariner¹⁶ says they are not allowed to make a sport of shooting them, this privilege being reserved for "chiefs, mata-booles, and mooas."

Of the rat in New Zealand, Dieffenbach⁶ tells us that the frugivorous Kiore Maori was formerly largely eaten by the natives, but that it had in 1843 become so scarce, owing to the extermination carried on against it by the European rat, that he could never obtain one.

Buller⁴ describes how during certain seasons the New Zealand rat was captured by thousands and eaten, or potted down in their own fat for future use.

At Penhryn Island, Smith²¹ informs us that the only animal on the atoll was a small rat, which was not eaten.

page 174

In Funafuti the native rat is described to me as having been driven from the village, and indeed almost exterminated upon the main islet by the foreign rat. Upon the other islets it exists and in some cases swarms, but as these islets are not permanently tenanted the rat can scarcely be regarded as a pest.

It constructs its nest in the cocoanut trees, just at the base of the fronds, and Mr. Hedley tells me that he frequently noticed the rats peeping out of the matting that sheathes the butts of the cocoanut fronds, and scampering about the heads of palms, fifteen or twenty feet high. In pre-civilised times these rats were a great plague to the natives, who did not use them as food. By law each individual was at times obliged to catch and destroy a set number of these vermin, for which purpose an ingenious trap was used.

Native Rat.
Mus Exulans, Peale.

Mus exulans, Peale, U.S. Expl. Exp., Mamm., 1st Ed., 1848, p. 47.

(Plate viii., figs. la-f.

Fur fine, scanty, and of medium length; colour warm brown, reddish on the nape and back, basal half of the hair delicate grey, the tips yellowish or brown. On the back the fur is mixed with longer and comparatively thick hairs of bristly texture, these are white or very pale yellow throughout their length, the extreme tip only being dark brown. Muzzle and face warm brown; the hairs on the sides of the body are tipped with pale yellow with no longer or darker hairs intermixed, The whole under surface including the inside of the limbs white, fur pale grey at the base. Ears rounded and of considerable breadth, but on being laid forward they do not reach the eye. Outside of limbs coloured like the back; on the hind foot the colour extends scarcely further than the heel leaving nearly all the foot white. Foot and claw-pads very large. Tail longer than the head and body, quite rat-like. Hairs longer than the scales, but not so long as two scales, excepting towards the tip which is inclined to be pencilled. Scales 9½ to the centimeter; mammae 2.2= 8.

Skull of delicate proportions; the nasals project considerably beyond the line of the premaxillary; supraorbital ridge thin but very prominent, it becomes lower in the temporal region and is little more than discernable above the aural aperture: condition of occipital region unknown. The anterior palatina foramina are somewhat broad and reach the anterior margin of the molar alveoli. The anterior zygoma root is rounded above and the front edge scarcely emarginate.

page 175

Teeth.—The teeth do not call for special reference, the character of the molar pattern being sufficiently represented on the accompanying plate (fig. Id).

Dimensions.	Millim,
Head and body	125·0
Tail	148·0
Length of head	37·5
Muzzle to ear	30·0
Ear	17·5
Forearm and hand	35·0
Hind foot	28·0
Heel to front of last foot-pad	13·7
Last foot pad	5·0

Skull·
Greatest length	?35·0	?
Basal length	3O·0	?
Greatest breadth	17·6	?
Nasals, length	14·0	14·5
Nasals, greatest breadth	4·0	4·1
Interorbital breadth	5·5	?
Interparietal length	4·7	?
Interparietal breadth	1O·0	?
Brain-case, breadth	13·6	?
Anterior zygoma root	3·5	3·8
Diastema	9·0	9·8
Palate, length	18·4	?
Anterior palatina foramina	5·7	6·0
Upper molars, length	5·7	6·2
Lower molars, length	6·0	?
Condyle to incisor tip	? 23·0	?
Coronoid tip to angle	9·2	?

Peale¹⁸ states that in his examples "the females have two pectoral and four abdominal teats," whereas in mine the pectorals are four. This may be reconciled by supposing that Peale overlooked a pair of mammas, an error, as I have in a former article indicated, easily committed.

Three examples of the Funafuti native rat were included in the collection: two of these I had not the opportunity of examining. The third had the skull a little but not very seriously damaged, and fragments of a fourth specimen enabled me to add the few figures in the second column of skull dimensions.

The stomach of the rat examined contained a white vegetable substance, possibly cocoanut or pandanus.

On the Funafuti Atoll this rat is known to the inhabitants by the name of "Tikimoa."

page 176

The Pacific Ocean being bounded by the land masses of Asia, Australia, South and North America, and the genus Mus being exclusively confined to the Old World, it necessarily follows that this rat has entered the islands of the Pacific from an Asiatic source. This agrees with the origin of the flora of the region as sketched by Guppy,^* and also with the distribution of the Lepidoptera independently remarked by Woodford.^*It is thus opposed to the theory of a migration westward from America across a Mesozoic Pacific continent as advocated by Hutton.^*

Postscript.

As previously indicated, it only needs more extensive reading to add materially to the known distribution of the Pacific rat, and already several localities may be added to those enumerated. A perusal of Brenchley, "Cruise of the Curac.oa,"² shows that it has been observed at Niue or Savage Island, situated between the Samoan and Cook Groups, and again at Aneitium, Tanna, and Efate, in the New Hebrides. It is also said to be indigenous to Upolu and Tutuila in the Samoan Islands, being at the latter place described as "the mouse." At Tongatabu the rat is said to be imported.

It may be mentioned that the only group in the South Pacific from which I have not quoted references is the Marquesas; an hiatus which would doubtless be filled did time permit to search its literature.

page 177

Native Rat.
Works Referred To.

1.	Allardyce, G. W. L.—Proc. Geogr. Soe. Aust. (Queensland Branch), i., 1886, p. 134.
2.	Arundel, John T.—The Phcenix Group and other Islands of the Pacific. 2a. Brenchley, Julius L.—Cruise of "Curacoa," 1873, pp. 25, 60, 86, 134, 199, 213, 231.
3.	Brigharn, William T.—Cat. Bernice Pauahi Bishop Museum, Honolulu,1892, p. 31.
4.	Buller, Walter—Trans. N.Z. Inst., iii., 1870, p. 1.
5.	Buller, W. L.—Trans. N.Z. Inst, xxv., 1892, p. 49.
6.	Dieffenbach, Ernest—Travels in New Zealand, ii., 1843, p. 185.
7.	Dixon, W. A.—Trans. Roy. Soc.N.S.W., xi., 1877, p. 172.—Foot-note.
8.	Dixon, W. S.—Mem. Nat. Acad. Sci., ii., 1883, p. 91.
9.	Gill, William Wyatt—Life in the Southern Isles, 1876, pp. 316-7.
10.	Gill, William Wyatt—Savage Life in Polynesia, 1880, p. 85.
11.	Gill, William Wyatt—Jottings from the Pacific, 1885, p. 127.
12.	Hochstetter, F. von.—New Zealand, 1867, p. 205.—Foot-note.
13.	Hutton, F. W.—Trans. N.Z. Inst., ix., 1876, p. 348.
14.	Hutton, F. W.—Trans. N.Z. Inst., xi., 1878, p. 344, and other papers.
15.	Kotzebue, Otto von—Voy. Discovery, ii., 1821, plate facing p. 63.
16.	Mariner, Wm. (Martin's)—Natives of the Tonga Islands, i., 1817, pp. 279-83.
17.	Meeson, John—Trans. N.Z. Inst., xvii., 1884, p. 199.
18.	Peale, Titian R.—Mamm., U.S. Explor. Exped., 2nd Ed., 1858, p. 38, pl. 4, fig. 1.
19.	Rutland, J.—Trans. N.Z. Inst., xxii., 1889, pp. 302-3.
20.	Smith, S. Percy—The Kermadec Islands, 1887, p. 24.
21.	Smith, S. Percy—Trans. N.Z. Inst., xxii., 1889, p. 99.
22.	Thomas, Oldfield—Pacific Rat, Proc. Zool. Soc, 1895, p. 338.
23.	Turner, George—Samoa, 1884, pp. 25, 187, 216, 218.
24.	Turner, George—Scot. Geogr. Mag., v., 1889, p. 246.
25.	Woodford, C. M.—The Gilbert Islands, Geogr. Journ., vi., 1895, pp. 347-9.