A boat was lowered down, and I went on sh.o.r.e with the botanical gentlemen, to look about the island. It is little better than a bank of sand, upon a basis of coral rock; yet it was covered with shrubs and trees so thickly, that in many places they were impenetrable. The north-western part is entirely sand, but there grew upon it numbers of _panda.n.u.s_ trees, similar to those of the east coast of New South Wales; and around many of them was placed a circle of sh.e.l.ls of the _chama gigas_, or gigantic c.o.c.kle, the intention of which excited my curiosity.

It appeared that this little island was visited occasionally by the Indians, who obtained from it the fruit of the panda.n.u.s, and probably turtle, for the marks of them were seen; and the reef furnishes them with c.o.c.kles, which are of a superior size here to those we had found upon the reefs of the East Coast. There being no water upon the island, they seem to have hit upon the following expedient to obtain it: Long slips of bark are tied round the smooth stems of the panda.n.u.s, and the loose ends are led into the sh.e.l.ls of the c.o.c.kle, placed underneath. By these slips, the rain which runs down the branches and stem of the tree, is conducted into the sh.e.l.ls, and fills them at every considerable shower; and as each sh.e.l.l will contain two or three pints, forty or fifty thus placed under different trees will supply a good number of men. A pair of these c.o.c.kle sh.e.l.ls, bleached in the sun, weighed a hundred and one pounds; but still they were much inferior in size to some I have since seen.

The fruit of the panda.n.u.s, as it is used by these Indians and by the natives of Terra Australis, affords very little nourishment. They suck the bottom part of the drupes, or separated nuts, as we do the leaves of the artichoke; but the quant.i.ty of pulp thus obtained, is very small, and to my taste, too astringent to be agreeable. In the third volume of the Asiatic Researches, the fruit of the panda.n.u.s is described as furnishing, under the name of _Mellori_, an important article of food to the inhabitants of the Nicobar Islands; and in Mauritius, one of these species is planted for its long and fibrous leaves, of which sacks, mats, and bags for coffee and cotton are in a made.

This little island, or rather the surrounding reef, which is three or four miles long, affords shelter from the south-east winds; and being at a moderate day's run from Murray's Isles, it forms a convenient anchorage for the night to a ship pa.s.sing through Torres' Strait: I named it _Half-way Island_. It is scarcely more than a mile in circ.u.mference, but appears to be increasing both in elevation and extent. At no very distant period of time, it was one of those banks produced by the washing up of sand and broken coral, of which most reefs afford instances, and those of Torres' Strait a great many. These banks are in different stages of progress: some, like this, are become islands, but not yet habitable; some are above high-water mark, but dest.i.tute of vegetation; whilst others are overflowed with every returning tide.

It seems to me, that when the animalcules which form the corals at the bottom of the ocean, cease to live, their structures adhere to each other, by virtue either of the glutinous remains within, or of some property in salt water; and the interstices being gradually filled up with sand and broken pieces of coral washed by the sea, which also adhere, a ma.s.s of rock is at length formed. Future races of these animalcules erect their habitations upon the rising bank, and die in their turn to increase, but princ.i.p.ally to elevate, this monument of their wonderful labours. The care taken to work perpendicularly in the early stages, would mark a surprising instinct in these diminutive creatures. Their wall of coral, for the most part in situations where the winds are constant, being arrived at the surface, affords a shelter, to leeward of which their infant colonies may be safely sent forth; and to this their instinctive foresight it seems to be owing, that the windward side of a reef exposed to the open sea, is generally, if not always the highest part, and rises almost perpendicular, sometimes from the depth of 200, and perhaps many more fathoms. To be constantly covered with water, seems necessary to the existence of the animalcules, for they do not work, except in holes upon the reef, beyond low-water mark; but the coral sand and other broken remnants thrown up by the sea, adhere to the rock, and form a solid ma.s.s with it, as high as the common tides reach. That elevation surpa.s.sed, the future remnants, being rarely covered, lose their adhesive property; and remaining in a loose state, form what is usually called a _key_, upon the top of the reef. The new bank is not long in being visited by sea birds; salt plants take root upon it, and a soil begins to be formed; a cocoa nut, or the drupe of a panda.n.u.s is thrown on sh.o.r.e; land birds visit it and deposit the seeds of shrubs and trees; every high tide, and still more every gale, adds something to the bank; the form of an island is gradually a.s.sumed; and last of all comes man to take possession.

Half-way Island is well advanced in the above progressive state; having been many years, probably some ages, above the reach of the highest spring tides, or the wash of the surf in the heaviest gales. I distinguished, however, in the rock which forms its basis, the sand, coral, and sh.e.l.ls formerly thrown up, in a more or less perfect state of cohesion; small pieces of wood, pumice stone, and other extraneous bodies which chance had mixed with the calcareous substances when the cohesion began, were inclosed in the rock; and in some cases were still separable from it without much force. The upper part of the island is a mixture of the same substances in a loose state, with a little vegetable soil; and is covered with the _casuarina_ and a variety of other trees and shrubs, which give food to paroquets, pigeons, and some other birds; to whose ancestors it is probable, the island was originally indebted for this vegetation.

The lat.i.tude of Half-way Island, deduced from that of the preceding and following noons, is 10 8' south, and longitude by time keeper corrected, 143 18' east. From the time of anchoring, to nine at night, there was a set past the ship to the north-east, of half a knot; it ceased for three hours, then recommencing at a slower rate, ran to the same point. Thus far in the strait, the current had been found to run at the rate of fourteen miles a day to the westward; and the above set might have been an eddy under the lee of the reef, for it seemed too irregular to be a tide.

[SUNDAY 31 OCTOBER 1802]

At daylight in the morning the south-east trade blew fresh with squally weather. We steered south-westward, pa.s.sing at seven o'clock between two dry sands, three or four miles apart, with a depth of 15 fathoms; at eight, another dry bank was left two miles to the southward, and a small, low island set at N. by W., two or three leagues. From this time, and running at the rate of seven knots, nothing was seen until ten; a dry sand then bore N. 78 W., two miles and a half, and two more low isles were seen to the northward; the soundings had become regular, between 10 and 9 fathoms, and the bottom was of mixed sand and sh.e.l.ls, fit for anchorage. Our lat.i.tude at noon was 10 26' 45", and longitude 142 39'; and we had high land bearing S. 3 E. ten or twelve miles, which I supposed might be the easternmost of the York Isles, although captain Cook's longitude of it was 38' more westward. The weather being hazy, no other land was seen, nor any reefs; but at one o'clock, I set these bearings:

York Isle, high flat top, S. 35 E.

A more northern, double isle, S. 84 W.

A high peaked hill (Mt. Ernest of Bligh), N. 16 W.

[NORTH COAST. PRINCE OF WALES' ISLANDS.]

At two o'clock, when we pa.s.sed on the north side of the double isle, it was seen to be surrounded with a coral reef, and there were rocks on its west and south sides. We then hauled tip S. W. by S. for some rocky islets lying, as I supposed, off Cape York; but finding no shelter there, bore away round the north end of an island, of which Mr. Westall took a view (Atlas, Plate XVIII. View 12), and anch.o.r.ed in 7 fathoms, gravel and sh.e.l.ls, one mile and a half from the land, and two or three cables length from a shoal to the southward, which became dry at low water. Our lat.i.tude here was 10 30' from bearings, and longitude by time-keeper 142(? illegible in book) 18' east; but I was altogether at a loss to know what islands these were, under which we had anch.o.r.ed. Supposing the flat-topped island to have been the easternmost York Isle, the land we had in sight to the southward should have been Cape York; but no such isles as those around us were laid down by captain Cook, to the north of that cape. On consulting the sketch made by captain Bligh in the Bounty's launch (Voyage to the South Seas, p. 220), it appeared that the first land was not the easternmost isle, but one much nearer to Cape York; and that our anchorage was under the southern group of the Prince of Wales'

Islands, the longitude of which, by captain Cook, is 1 12' west of what I make it.* The north-eastern isle of this group, under which we more immediately lay, is that named Wednesday Island by captain Bligh; to the other isles he gave no name; but the one westward of the ship seems to have been the Hammond's Island of captain Edwards, when pa.s.sing here with the Pandora's boats. So soon as the weather cleared a little, the subjoined bearings were taken.

[* Mr. Wales deduces from captain Cook's observations in the Endeavour, that the error of his chart here, is 35' west (_Astron. Observations_, p.

131).]

Wednesday I., distant 1 to 3 miles, S. 89 E. to 21 W.

Hammond's Isle, dist. 4 or 5 miles, S. 52 W. to 71 W.

Hawkesbury I. (of Edwards), highest part, N. 52 W.

Mount Augustus (of Bligh), N. 2 W.

A small isle, distant three leagues, N. 24 E.

Mount Ernest, peak, N. 36 E.

Double Isle, pa.s.sed at 2 p.m., N. 70 E.

Breakers on a reef, distant 3 miles, N. 64 to 30 W.

[MONDAY 1 NOVEMBER 1802]

This evening and all the next day, the wind blew so strong that it was impossible to land; nor did I think it prudent to quit the anchorage, though anxious to commence the survey of the Gulph of Carpentaria. Upon Hammond's Island some fires were seen; but Wednesday Island showed no signs of being inhabited, unless some whitish, conical figures like sentry boxes, were huts; there were bushes and small trees scattered over both islands, but their general appearance was rocky and barren.

The tide here ran nine hours to the westward, at the strongest two-and-half knots; and three hours north-eastward, but scarcely perceptible; which deviation from the regular order was probably caused by the current setting westward. So far as the soundings taken every hour could ascertain the rise, it was at least two fathoms, and high water took place _four or five hours after_ the moon's pa.s.sage over and under the meridian, and was completed by the three hours tide. According to this, it would be high water here, and low water at Murray's Islands at the same time, which would present a remarkable a.n.a.logy between this strait and that of Ba.s.s to the southward; this however is certain, that the tide set E. by S. one knot and a quarter, at Murray's Islands, at four in the morning; and that two days afterward, at Wednesday Island, it set from one-and-half to two-and-half knots W. by S., from one till seven in the morning. I will not venture to say that the latter part of the flood comes from southwest at the Prince of Wales' Islands, though appearances bespoke it; because captain Cook, who had better opportunity for observation, found it setting from the east, in Endeavour's Strait.

He also gives the time of high water at one or two hours after the moon, which comes nearer to what I observed at Murray's Islands.

From azimuths with the surveying compa.s.s when the head was S. E. by E., the variation was 3 32', or corrected to the meridian, 4 52' east.

TUESDAY 2 NOVEMBER 1802

In the morning of Nov. 2, the wind being more moderate and at E. S. E., we steered between Hammond's Island and the north-western reef, with soundings from 6 to 9 fathoms. Another island appeared beyond Hammond's, to the south-west, which, as it had no name, I called _Good's Island_, after Mr. Good, the botanical gardener; and we hauled up for it, pa.s.sing a rock and a small reef between the two. On seeing an extensive shoal ahead, which would have carried us off the land to go round it, we anch.o.r.ed in 7 fathoms, dead coral and sh.e.l.ls, with the north end of Hammond's Island bearing N. 64 E., four or five miles. The botanical gentlemen landed on Good's Island; and in the afternoon I took these bearings amongst others, from a hill near its south-west end.

The ship, distant 1 miles, N. 58 0' W.

Wallis' Isles, over the Shoal Cape of Bligh, S. 23 5 W.

b.o.o.by Isle, centre, S. 80 0 W.

Northern isles, the westernmost visible, N. 28 10' to 24 5 W.

Hawkesbury Island, N. 9 15 to 4 0 W.

North-west reef, its apparent termination, N. 38 50 W.

The shoal which stopped our progress did not run off from Shoal Cape, as captain Bligh had supposed, but from a smaller and nearer island, two miles from my station. Within the large island, of which Shoal Cape forms the north-western point, I saw water like an inclosed port, probably the Wolf's Bay of captain Edwards; and it seemed possible that the land may be there divided; but the best information I can give of the forms and extent of all these islands, will be seen in the particular chart.

It was now ascertained, that the figures resembling sentry boxes were ant hills, of eight or more feet high; Pelsert found similar hills on the West Coast, and says they might have been taken for the houses of Indians, as in fact we did take them at a distant view. They were also seen by Dampier on the North-west Coast, who mistook them in the same way; but says he found them to be so many rocks, probably from not making the examination with his usual care. The insects which inhabit, and I suppose erect these structures, are small, reddish, with black heads, and seemed to be a sluggish and feeble race. We found the common black flies excessively numerous here; and almost as troublesome as Dampier describes them to be on the North-west Coast.

Good's Island is between one and two miles long, and resembles the rest of the cl.u.s.ter in being hilly, woody, and rocky, with small beaches on the leeward side. The stone is granitic and brittle; but there is also porphyry, and in one place I found streaks of verdegrease, as if the cliffs above had contained copper ore. A log of wood, resembling the cedar of Port Jackson, was thrown up on the beach, but none of the trees were seen; those scattered over the island, though of various kinds, were small and fit for little else than the fire. A species of silk-cotton plant was plentiful; the fibres in the pod are strong, and have a fine gloss, and might perhaps be advantageously employed in manufacture.

From two supplements of the sun's meridian alt.i.tude to the north, the _lat.i.tude_ of our anchorage would be 10 34' 12"; but the supplements observed on the 31st having given 1' 14" too far south, the correct lat.i.tude is taken to be 10 32' 58". The _longitude_ from nine sets of distances of the sun west of the moon, was 142 23'; but by the corrected time keeper, which I prefer, it was 142 10' east. To compare this longitude with that of captain Cook, it must be reduced to some point distinctly laid down by him, and I take b.o.o.by Island, which was in sight.

According to that navigator, b.o.o.by Isle is in 140 38' east (Hawkesworth, III, 214); whereas I made it to lie in 141 57', or 1 19' further east, a difference which certainly appears very extraordinary; but it is still more so, that the island should be laid down 63' of longitude to the west of the high, flat-topped York Isle, instead Of 43' or 44'. To show that the longitude by my time keeper was not much, if any thing too great, I have to observe, that in captain Bligh's ma.n.u.script chart of 1792, Mount Augustus is laid down from his time keepers in 142 14'; and the mean of his lunar observations, taken eight days before and six days afterward, was 16' _more east_. My time keeper now placed Mount Augustus in 142 18', or only 4' more east than captain Bligh's chart, consequently in 12'

less than by his lunar observations; by which quant.i.ty it was also less than the nine sets of distances now taken by lieutenant Flinders.

No run of tide was perceptible at the anchorage, from eight in the morning to two p.m.; but it then set westward, and continued so to do until four next morning, and was then running one knot and a half. The time of high water appeared by the soundings, to be nearly as they gave it at Wednesday Island.

WEDNESDAY 3 NOVEMBER 1802

In the morning of the 3rd, the wind was moderate at E. S. E., and we made sail to get in with the main land to the south of the Prince of Wales'

Islands. In hauling round the dry part of the shoal, we fell into 3 fathoms, and were obliged to steer round off; nor was it until after many attempts, and running four or five miles further to the south-westward, that the shoal would allow us to steer a southern course. At 8h 45', being then in 5 fathoms,

b.o.o.by Isle bore, N. 56 W.

Cape Cornwall, S. 58 E.

Station on Good's Island, dist. 11 miles N. 54 E.

From hence we carried 6 to 7 fathoms until past ten, and afterwards irregular soundings between 3 and 9 fathoms, to noon; the lat.i.tude from a supplement to the north, with the same correction as applied on the 2nd, was then 10 50' 44", and the bearings of the land were these;

Station on Good's Island, N. 29 E.

Cape Cornwall, N. 68 E.

Wallis' Isles, the highest, distant 2 miles, N. 84 E.

Wallis' Isles, a lower and broader, dist. 3 or 4 miles, S. 71 to 64 E.

Main land, low sandy point, dist. 8 miles, S. 43 E.

Main land, furthest extreme near a smoke, S. 77 E.