APRIL 1st. -- We arrived in view of the Keeling or CocosIslands, situated in the Indian Ocean, and about six hundredmiles distant from the coast of Sumatra. This is one of thelagoon-islands (or atolls) of coral formation, similar tothose in the Low Archipelago which we passed near. Whenthe ship was in the channel at the entrance, Mr. Liesk,an English resident, came off in his boat. The historyof the inhabitants of this place, in as few words aspossible, is as follows. About nine years ago, Mr. Hare,a worthless character, brought from the East Indianarchipelago a number of Malay slaves, which now includingchildren, amount to more than a hundred. Shortly afterwards,Captain Ross, who had before visited these islands in hismerchant-ship, arrived from England, bringingwith him his family and goods for settlement along withhim came Mr. Liesk, who had been a mate in his vessel.The Malay slaves soon ran away from the islet on whichMr. Hare was settled, and joined Captain Ross's party. Mr.Hare upon this was ultimately obliged to leave the place.

The Malays are now nominally in a state of freedom, andcertainly are so, as far as regards their personal treatment;but in most other points they are considered as slaves. Fromtheir discontented state, from the repeated removals fromislet to islet, and perhaps also from a little mismanagement,things are not very prosperous. The island has no domesticquadruped, excepting the pig, and the main vegetable productionis the cocoa-nut. The whole prosperity of the placedepends on this tree: the only exports being oil from the nut,and the nuts themselves, which are taken to Singapore andMauritius, where they are chiefly used, when grated, in makingcurries. On the cocoa-nut, also, the pigs, which areloaded with fat, almost entirely subsist, as do the ducks andpoultry. Even a huge land-crab is furnished by nature withthe means to open and feed on this most useful production.

The ring-formed reef of the lagoon-island is surmountedin the greater part of its length by linear islets. On thenorthern or leeward side, there is an opening through whichvessels can pass to the anchorage within. On entering, thescene was very curious and rather pretty; its beauty, however,entirely depended on the brilliancy of the surroundingcolours. The shallow, clear, and still water of the lagoon,resting in its greater part on white sand, is, when illuminedby a vertical sun, of the most vivid green. This brilliantexpanse, several miles in width, is on all sides divided, eitherby a line of snow-white breakers from the dark heavingwaters of the ocean, or from the blue vault of heaven bythe strips of land, crowned by the level tops of the cocoa-nuttrees. As a white cloud here and there affords a pleasingcontrast with the azure sky, so in the lagoon, bands ofliving coral darken the emerald green water.

The next morning after anchoring, I went on shore onDirection Island. The strip of dry land is only a few hundredyards in width; on the lagoon side there is a white calcareousbeach, the radiation from which under this sultryclimate was very oppressive; and on the outer coast, a solidbroad flat of coral-rock served to break the violence of theopen sea. Excepting near the lagoon, where there is somesand, the land is entirely composed of rounded fragments ofcoral. In such a loose, dry, stony soil, the climate of theintertropical regions alone could produce a vigorous vegetation.On some of the smaller islets, nothing could be moreelegant than the manner in which the young and full-growncocoa-nut trees, without destroying each other's symmetry,were mingled into one wood. A beach of glittering whitesand formed a border to these fairy spots.

I will now give a sketch of the natural history of theseislands, which, from its very paucity, possesses a peculiarinterest. The cocoa-nut tree, at first glance, seems tocompose the whole wood; there are however, five or sixother trees. One of these grows to a very large size, butfrom the extremes of softness of its wood, is useless; anothersort affords excellent timber for ship-building. Besides thetrees, the number of plants is exceedingly limited, and consistsof insignificant weeds. In my collection, which includes,I believe, nearly the perfect Flora, there are twentyspecies, without reckoning a moss, lichen, and fungus. Tothis number two trees must be added; one of which was notin flower, and the other I only heard of. The latter is asolitary tree of its kind, and grows near the beach, where,without doubt, the one seed was thrown up by the waves. AGuilandina also grows on only one of the islets. I do notinclude in the above list the sugar-cane, banana, some othervegetables, fruit-trees, and imported grasses. As the islandsconsist entirely of coral, and at one time must have existedas mere water-washed reefs, all their terrestrial productionsmust have been transported here by the waves of the sea.In accordance with this, the Florula has quite the characterof a refuge for the destitute: Professor Henslow informsme that of the twenty species nineteen belong to differentgenera, and these again to no less than sixteen families!

In Holman's

Chamisso,

It is also said that palms and bamboos from somewherein the torrid zone, and trunks of northern firs, arewashed on shore: these firs must have come from an immensedistance. These facts are highly interesting. It cannotbe doubted that if there were land-birds to pick up theseeds when first cast on shore, and a soil better adapted fortheir growth than the loose blocks of coral, that the mostisolated of the lagoon-islands would in time possess a farmore abundant Flora than they now have.

The list of land animals is even poorer than that of theplants. Some of the islets are inhabited by rats, which werebrought in a ship from the Mauritius, wrecked here. Theserats are considered by Mr. Waterhouse as identical with theEnglish kind, but they are smaller, and more brightly coloured.There are no true land-birds, for a snipe and a rail(Rallus Phillippensis), though living entirely in the dryherbage, belong to the order of Waders. Birds of this orderare said to occur on several of the small low islands in thePacific. At Ascension, where there is no land-bird, a rail(Porphyrio simplex) was shot near the summit of the mountain,and it was evidently a solitary straggler. At Tristand'Acunha, where, according to Carmichael, there are onlytwo land-birds, there is a coot. From these facts I believethat the waders, after the innumerable web-footed species,are generally the first colonists of small isolated islands. Imay add, that whenever I noticed birds, not of oceanicspecies, very far out at sea, they always belonged to thisorder; and hence they would naturally become the earliestcolonists of any remote point of land.

Of reptiles I saw only one small lizard. Of insects I tookpains to collect every kind. Exclusive of spiders, which werenumerous, there were thirteen species.

The long strips of land, forming the linear islets, havebeen raised only to that height to which the surf can throwfragments of coral, and the wind heap up calcareous sand.The solid flat of coral rock on the outside, by its breadth,breaks the first violence of the waves, which otherwise, in aday, would sweep away these islets and all their productions.The ocean and the land seem here struggling for mastery:although terra firma has obtained a footing, the denizens ofthe water think their claim at least equally good. In everypart one meets hermit crabs of more than one species,

Sunday, April 3rd. -- After service I accompanied CaptainFitz Roy to the settlement, situated at the distance of somemiles, on the point of an islet thickly covered with tallcocoa-nut trees. Captain Ross and Mr. Liesk live in a largebarn-like house open at both ends, and lined with mats made ofwoven bark. The houses of the Malays are arranged alongthe shore of the lagoon. The whole place had rather a desolateaspect, for there were no gardens to show the signs ofcare and cultivation. The natives belong to different islandsin the East Indian archipelago, but all speak the same language:we saw the inhabitants of Borneo, Celebes, Java, andSumatra. In colour they resemble the Tahitians, from whomthey do not widely differ in features. Some of the women,however, show a good deal of the Chinese character. I likedboth their general expressions and the sound of their voices.They appeared poor, and their houses were destitute offurniture; but it was evident, from the plumpness of the littlechildren, that cocoa-nuts and turtle afford no bad sustenance.

On this island the wells are situated, from which shipsobtain water. At first sight it appears not a little remarkablethat the fresh water should regularly ebb and flow with thetides; and it has even been imagined, that sand has the powerof filtering the salt from the sea-water. These ebbing wellsare common on some of the low islands in the West Indies.The compressed sand, or porous coral rock, is permeated likea sponge with the salt water, but the rain which falls on thesurface must sink to the level of the surrounding sea, andmust accumulate there, displacing an equal bulk of the saltwater. As the water in the lower part of the great sponge-like coral mass rises and falls with the tides, so will thewater near the surface; and this will keep fresh, if the massbe sufficiently compact to prevent much mechanical admixture;but where the land consists of great loose blocks ofcoral with open interstices, if a well be dug, the water, as Ihave seen, is brackish.

After dinner we stayed to see a curious half superstitiousscene acted by the Malay women. A large wooden spoondressed in garments, and which had been carried to the graveof a dead man, they pretend becomes inspired at the full ofthe moon, and will dance and jump about. After the properpreparations, the spoon, held by two women, became convulsed,and danced in good time to the song of the surroundingchildren and women. It was a most foolish spectacle;but Mr. Liesk maintained that many of the Malays believedin its spiritual movements. The dance did not commence tillthe moon had risen, and it was well worth remaining to beholdher bright orb so quietly shining through the long armsof the cocoa-nut trees as they waved in the evening breeze.These scenes of the tropics are in themselves so delicious,that they almost equal those dearer ones at home, to whichwe are bound by each best feeling of the mind.

The next day I employed myself in examining the veryinteresting, yet simple structure and origin of these islands.The water being unusually smooth, I waded over the outerflat of dead rock as far as the living mounds of coral, onwhich the swell of the open sea breaks. In some of thegullies and hollows there were beautiful green and othercoloured fishes, and the form and tints of many of the zoophyteswere admirable. It is excusable to grow enthusiastic overthe infinite numbers of organic beings with which the sea ofthe tropics, so prodigal of life, teems; yet I must confess Ithink those naturalists who have described, in well-knownwords, the submarine grottoes decked with a thousand beauties,have indulged in rather exuberant language.

April 6th. -- I accompanied Captain Fitz Roy to an islandat the head of the lagoon: the channel was exceedinglyintricate, winding through fields of delicately branched corals.We saw several turtle and two boats were then employed incatching them. The water was so clear and shallow, that althoughat first a turtle quickly dives out of sight, yet in acanoe or boat under sail, the pursuers after no very longchase come up to it. A man standing ready in the bow, atthis moment dashes through the water upon the turtle's back;then clinging with both hands by the shell of its neck, he iscarried away till the animal becomes exhausted and is secured.It was quite an interesting chase to see the two boatsthus doubling about, and the men dashing head foremostinto the water trying to seize their prey. Captain Moresbyinforms me that in the Chagos archipelago in this sameocean, the natives, by a horrible process, take the shell fromthe back of the living turtle. "It is covered with burningcharcoal, which causes the outer shell to curl upwards, it isthen forced off with a knife, and before it becomes coldflattened between boards. After this barbarous process theanimal is suffered to regain its native element, where, aftera certain time, a new shell is formed; it is, however, toothin to he of any service, and the animal always appearslanguishing and sickly."

When we arrived at the head of the lagoon, we crossed anarrow islet, and found a great surf breaking on the windwardcoast. I can hardly explain the reason, but there is tomy mind much grandeur in the view of the outer shores ofthese lagoon-islands. There is a simplicity in the barrier-likebeach, the margin of green bushes and tall cocoa-nuts,the solid flat of dead coral-rock, strewed here and therewith great loose fragments, and the line of furious breakers,all rounding away towards either hand. The oceanthrowing its waters over the broad reef appears an invincible,all-powerful enemy; yet we see it resisted, and evenconquered, by means which at first seem most weak andinefficient. It is not that the ocean spares the rock of coral;the great fragments scattered over the reef, and heaped onthe beach, whence the tall cocoa-nut springs, plainly bespeakthe unrelenting power of the waves. Nor are anyperiods of repose granted. The long swell caused by thegentle but steady action of the trade-wind, always blowingin one direction over a wide area, causes breakers, almostequalling in force those during a gale of wind in the temperateregions, and which never cease to rage. It is impossibleto behold these waves without feeling a conviction thatan island, though built of the hardest rock, let it be porphyry,granite, or quartz, would ultimately yield and be demolishedby such an irresistible power. Yet these low, insignificantcoral-islets stand and are victorious: for here another power,as an antagonist, takes part in the contest. The organic forcesseparate the atoms of carbonate of lime, one by one, fromthe foaming breakers, and unite them into a symmetricalstructure. Let the hurricane tear up its thousand hugefragments; yet what will that tell against the accumulatedlabour of myriads of architects at work night and day, monthafter month? Thus do we see the soft and gelatinous body of apolypus, through the agency of the vital laws, conqueringthe great mechanical power of the waves of an ocean whichneither the art of man nor the inanimate works of naturecould successfully resist.

We did not return on board till late in the evening, for westayed a long time in the lagoon, examining the fields ofcoral and the gigantic shells of the chama, into which, if aman were to put his hand, he would not, as long as the animallived, be able to withdraw it. Near the head of thelagoon I was much surprised to find a wide area, considerablymore than a mile square, covered with a forest of delicatelybranching corals, which, though standing upright,were all dead and rotten. At first I was quite at a loss tounderstand the cause afterwards it occurred to me that itwas owing to the following rather curious combination ofcircumstances. It should, however, first be stated, that coralsare not able to survive even a short exposure in the air tothe sun's rays, so that their upward limit of growth isdetermined by that of lowest water at spring tides. It appears,from some old charts, that the long island to windward wasformerly separated by wide channels into several islets; thisfact is likewise indicated by the trees being younger on theseportions. Under the former condition of the reef, a strongbreeze, by throwing more water over the barrier, would tendto raise the level of the lagoon. Now it acts in a directlycontrary manner; for the water within the lagoon not onlyis not increased by currents from the outside, but is itselfblown outwards by the force of the wind. Hence it is observed,that the tide near the head of the lagoon does notrise so high during a strong breeze as it does when it iscalm. This difference of level, although no doubt very small,has, I believe, caused the death of those coral-groves, whichunder the former and more open condition of the outer reefhas attained the utmost possible limit of upward growth.

A few miles north of Keeling there is another small atoll,the lagoon of which is nearly filled up with coral-mud. CaptainRoss found embedded in the conglomerate on the outercoast, a well-rounded fragment of greenstone, rather largerthan a man's head: he and the men with him were so muchsurprised at this, that they brought it away and preserved itas a curiosity. The occurrence of this one stone, whereevery other particle of matter is calcareous, certainly is verypuzzling. The island has scarcely ever been visited, nor is itprobable that a ship had been wrecked there. From the absenceof any better explanation, I came to the conclusion thatit must have come entangled in the roots of some large tree:when, however, I considered the great distance from thenearest land, the combination of chances against a stone thusbeing entangled, the tree washed into the sea, floated so far,then landed safely, and the stone finally so embedded as toallow of its discovery, I was almost afraid of imagining ameans of transport apparently so improbable. It was thereforewith great interest that I found Chamisso, the justlydistinguished naturalist who accompanied Kotzebue, statingthat the inhabitants of the Radack archipelago, a group oflagoon-islands in the midst of the Pacific, obtained stonesfor sharpening their instruments by searching the roots oftrees which are cast upon the beach. It will be evident thatthis must have happened several times, since laws have beenestablished that such stones belong to the chief, and apunishment is inflicted on any one who attempts to steal them.When the isolated position of these small islands in themidst of a vast ocean -- their great distance from any landexcepting that of coral formation, attested by the valuewhich the inhabitants, who are such bold navigators, attachto a stone of any kind,

During another day I visited West Islet, on which thevegetation was perhaps more luxuriant than on any other.The cocoa-nut trees generally grow separate, but here theyoung ones flourished beneath their tall parents, and formedwith their long and curved fronds the most shady arbours.Those alone who have tried it, know how delicious it is tobe seated in such shade, and drink the cool pleasant fluidof the cocoa-nut. In this island there is a large bay-likespace, composed of the finest white sand: it is quite leveland is only covered by the tide at high water; from thislarge bay smaller creeks penetrate the surrounding woods.To see a field of glittering white sand, representing water,with the cocoa-nut trees extending their tall and wavingtrunks around the margin, formed a singular and very prettyview.

I have before alluded to a crab which lives on the cocoa-nuts;it is very common on all parts of the dry land, andgrows to a monstrous size: it is closely allied or identicalwith the Birgos latro. The front pair of legs terminate invery strong and heavy pincers, and the last pair are fittedwith others weaker and much narrower. It would at firstbe thought quite impossible for a crab to open a strongcocoa-nut covered with the husk; but Mr. Liesk assures methat he has repeatedly seen this effected. The crab beginsby tearing the husk, fibre by fibre, and always from thatend under which the three eye-holes are situated; when thisis completed, the crab commences hammering with its heavyclaws on one of the eye-holes till an opening is made. Thenturning round its body, by the aid of its posterior and narrowpair of pincers, it extracts the white albuminous substance.I think this is as curious a case of instinct as everI heard of, and likewise of adaptation in structure betweentwo objects apparently so remote from each other in thescheme of nature, as a crab and a cocoa-nut tree. TheBirgos is diurnal in its habits; but every night it is said topay a visit to the sea, no doubt for the purpose of moisteningits branchiae. The young are likewise hatched, and live forsome time, on the coast. These crabs inhabit deep burrows,which they hollow out beneath the roots of trees; and wherethey accumulate surprising quantities of the picked fibresof the cocoa-nut husk, on which they rest as on a bed. TheMalays sometimes take advantage of this, and collect thefibrous mass to use as junk. These crabs are very good toeat; moreover, under the tail of the larger ones there is amass of fat, which, when melted, sometimes yields as muchas a quart bottle full of limpid oil. It has been stated bysome authors that the Birgos crawls up the cocoa-nut treesfor the purpose of stealing the nuts: I very much doubt thepossibility of this; but with the Pandanus

Captain Moresby informs me that this crab inhabits theChagos and Seychelle groups, but not the neighbouring Maldivaarchipelago. It formerly abounded at Mauritius, butonly a few small ones are now found there. In the Pacific,this species, or one with closely allied habits, is said

I was a good deal surprised by finding two species ofcoral of the genus Millepora (M. complanata and alcicornis),possessed of the power of stinging. The stony branches orplates, when taken fresh from the water, have a harsh feeland are not slimy, although possessing a strong and disagreeablesmell. The stinging property seems to vary indifferent specimens: when a piece was pressed or rubbed onthe tender skin of the face or arm, a pricking sensation wasusually caused, which came on after the interval of a second,and lasted only for a few minutes. One day, however, bymerely touching my face with one of the branches, pain wasinstantaneously caused; it increased as usual after a fewseconds, and remaining sharp for some minutes, was perceptiblefor half an hour afterwards. The sensation was asbad as that from a nettle, but more like that caused by thePhysalia or Portuguese man-of-war. Little red spots wereproduced on the tender skin of the arm, which appeared as ifthey would have formed watery pustules, but did not. M.Quoy mentions this case of the Millepora; and I have heardof stinging corals in the West Indies. Many marine animalsseem to have this power of stinging: besides the Portugueseman-of-war, many jelly-fish, and the Aplysia or sea-slugof the Cape de Verd Islands, it is stated in the voyageof the Astrolabe, that an Actinia or sea-anemone, as well asa flexible coralline allied to Sertularia, both possess thismeans of offence or defence. In the East Indian sea, astinging sea-weed is said to be found.

Two species of fish, of the genus Scarus, which are commonhere, exclusively feed on coral: both are coloured of asplendid bluish-green, one living invariably in the lagoon,and the other amongst the outer breakers. Mr. Liesk assuredus, that he had repeatedly seen whole shoals grazing withtheir strong bony jaws on the tops of the coral branches: Iopened the intestines of several, and found them distendedwith yellowish calcareous sandy mud. The slimy disgustingHoluthuriae (allied to our star-fish), which the Chinesegourmands are so fond of, also feed largely, as I am informed byDr. Allan, on corals; and the bony apparatus within theirbodies seems well adapted for this end. These Holuthuriae,the fish, the numerous burrowing shells, and nereidousworms, which perforate every block of dead coral, must bevery efficient agents in producing the fine white mud whichlies at the bottom and on the shores of the lagoon. A portion,however, of this mud, which when wet resembledpounded chalk, was found by Professor Ehrenberg to bepartly composed of siliceous-shielded infusoria.

April 12th. -- In the morning we stood out of the lagoonon our passage to the Isle of France. I am glad we havevisited these islands: such formations surely rank highamongst the wonderful objects of this world. Captain FitzRoy found no bottom with a line 7200 feet in length, at thedistance of only 2200 yards from the shore; hence this islandforms a lofty submarine mountain, with sides steeper eventhan those of the most abrupt volcanic cone. The saucer-shapedsummit is nearly ten miles across; and every singleatom,

I will now give a very brief account of the three greatclasses of coral-reefs; namely, Atolls, Barrier, and Fringing-reefs, and will explain my views

[picture]

une merveille de voir chacun de ces atollons, environne d'ungrand banc de pierre tout autour, n'y ayant point d'artificehumain." The accompanying sketch of Whitsunday Islandin the Pacific, copied from, Capt. Beechey's admirable Voyage,gives but a faint idea of the singular aspect of an atoll:it is one of the smallest size, and has its narrow islets unitedtogether in a ring. The immensity of the ocean, the fury ofthe breakers, contrasted with the lowness of the land and thesmoothness of the bright green water within the lagoon, canhardly be imagined without having been seen.

The earlier voyagers fancied that the coral-building animalsinstinctively built up their great circles to afford themselvesprotection in the inner parts; but so far is this fromthe truth, that those massive kinds, to whose growth on theexposed outer shores the very existence of the reef depends,cannot live within the lagoon, where other delicately-branchingkinds flourish. Moreover, on this view, many speciesof distinct genera and families are supposed to combine forone end; and of such a combination, not a single instancecan be found in the whole of nature. The theory that hasbeen most generally received is, that atolls are based onsubmarine craters; but when we consider the form and size ofsome, the number, proximity, and relative positions of others,this idea loses its plausible character: thus Suadiva atoll is44 geographical miles in diameter in one line, by 34 miles inanother line; Rimsky is 54 by 20 miles across, and it has astrangely sinuous margin; Bow atoll is 30 miles long, and onan average only 6 in width; Menchicoff atoll consists of threeatolls united or tied together. This theory, moreover, istotally inapplicable to the northern Maldiva atolls in theIndian Ocean (one of which is 88 miles in length, and between 10and 20 in breadth), for they are not bounded like ordinaryatolls by narrow reefs, but by a vast number of separatelittle atolls; other little atolls rising out of the greatcentral lagoon-like spaces. A third and better theory wasadvanced by Chamisso, who thought that from the corals growingmore vigorously where exposed to the open sea, as undoubtedly isthe case, the outer edges would grow up from the generalfoundation before any other part, and that this would accountfor the ring or cup-shaped structure. But we shallimmediately see, that in this, as well as in the crater-theory,a most important consideration has been overlooked, namely,on what have the reef-building corals, which cannot live ata great depth, based their massive structures?

Numerous soundings were carefully taken by Captain FitzRoy on the steep outside of Keeling atoll, and it was foundthat within ten fathoms, the prepared tallow at the bottomof the lead, invariably came up marked with the impressionof living corals, but as perfectly clean as if it had beendropped on a carpet of turf; as the depth increased, theimpressions became less numerous, but the adhering particlesof sand more and more numerous, until at last it was evidentthat the bottom consisted of a smooth sandy layer: to carryon the analogy of the turf, the blades of grass grew thinnerand thinner, till at last the soil was so sterile, that nothingsprang from it. From these observations, confirmed by manyothers, it may be safely inferred that the utmost depth atwhich corals can construct reefs is between 20 and 30 fathoms.Now there are enormous areas in the Pacific and IndianOcean, in which every single island is of coral formation,and is raised only to that height to which the waves canthrow up fragments, and the winds pile up sand. ThusRadack group of atolls is an irregular square, 520 miles longand 240 broad; the Low Archipelago is elliptic-formed, 840miles in its longer, and 420 in its shorter axis: there areother small groups and single low islands between these twoarchipelagoes, making a linear space of ocean actually morethan 4000 miles in length, in which not one single islandrises above the specified height. Again, in the Indian Oceanthere is a space of ocean 1500 miles in length, includingthree archipelagoes, in which every island is low and ofcoral formation. From the fact of the reef-building coralsnot living at great depths, it is absolutely certain thatthroughout these vast areas, wherever there is now an atoll,a foundation must have originally existed within a depth offrom 20 to 30 fathoms from the surface. It is improbable inthe highest degree that broad, lofty, isolated, steep-sidedbanks of sediment, arranged in groups and lines hundreds ofleagues in length, could have been deposited in the centraland profoundest parts of the Pacific and Indian Oceans, atan immense distance from any continent, and where thewater is perfectly limpid. It is equally improbable that theelevatory forces should have uplifted throughout the abovevast areas, innumerable great rocky banks within 20 to 30fathoms, or 120 to 180 feet, of the surface of the sea, andnot one single point above that level; for where on the wholesurface of the globe can we find a single chain of mountains,even a few hundred miles in length, with their many summitsrising within a few feet of a given level, and not onepinnacle above it? If then the foundations, whence the atoll-building corals sprang, were not formed of sediment, and ifthey were not lifted up to the required level, they must ofnecessity have subsided into it; and this at once solves thedifficulty. For as mountain after mountain, and island afterisland, slowly sank beneath the water, fresh bases would besuccessively afforded for the growth of the corals. It isimpossible here to enter into all the necessary details, but Iventure to defy

Before explaining how atoll-formed reefs acquire theirpeculiar structure, we must turn to the second great class,namely, Barrier-reefs. These either extend in straight linesin front of the shores of a continent or of a large island, orthey encircle smaller islands; in both cases, being separatedfrom the land by a broad and rather deep channel of water,analogous to the lagoon within an atoll. It is remarkablehow little attention has been paid to encircling barrier-reefs;yet they are truly wonderful structures. The following sketchrepresents part of the barrier encircling the island of Bolabolain the Pacific, as seen from one of the central peaks.In this instance the whole line of reef has been convertedinto land; but usually a snow-white line of great breakers,with only here and there a single low islet crowned withcocoa-nut trees, divides the dark heaving waters of the oceanfrom the light-green expanse of the lagoon-channel. Andthe quiet waters of this channel generally bathe a fringe oflow alluvial soil, loaded with the most beautiful productionsof the tropics, and lying at the foot of the wild, abrupt,central mountains.

Encircling barrier-reefs are of all sizes, from three milesto no less than forty-four miles in diameter; and that whichfronts one side, and encircles both ends, of New Caledonia,is 400 miles long. Each reef includes one, two, or severalrocky islands of various heights; and in one instance, evenas many as twelve separate islands. The reef runs at agreater or less distance from the included land; in theSociety archipelago generally from one to three or fourmiles; but at Hogoleu the reef is 20 miles on the southernside, and 14 miles on the opposite or northern side, from theincluded islands. The depth within the lagoon-channel alsovaries much; from 10 to 30 fathoms may be taken as anaverage; but at Vanikoro there are spaces no less than 56fathoms or 363 feet deep. Internally the reef either slopesgently into the lagoon-channel, or ends in a perpendicularwall sometimes between two and three hundred feet underwater in height: externally the reef rises, like an atoll, withextreme abruptness out of the profound depths of the ocean.

What can be more singular than these structures? We see

[picture]

an island, which may be compared to a castle situated on thesummit of a lofty submarine mountain, protected by a greatwall of coral-rock, always steep externally and sometimesinternally, with a broad level summit, here and there breachedby a narrow gateway, through which the largest ships canenter the wide and deep encircling moat.

As far as the actual reef of coral is concerned, there is notthe smallest difference, in general size, outline, grouping,and even in quite trifling details of structure, between abarrier and an atoll. The geographer Balbi has well remarked,that an encircled island is an atoll with high land rising outof its lagoon; remove the land from within, and a perfectatoll is left.

But what has caused these reefs to spring up at suchgreat distances from the shores of the included islands? Itcannot be that the corals will not grow close to the land;for the shores within the lagoon-channel, when not surroundedby alluvial soil, are often fringed by living reefs;and we shall presently see that there is a whole class, whichI have called Fringing Reefs from their close attachmentto the shores both of continents and of islands. Again, onwhat have the reef-building corals, which cannot live atgreat depths, based their encircling structures? This is agreat apparent difficulty, analogous to that in the case ofatolls, which has generally been overlooked. It will beperceived more clearly by inspecting the following sectionswhich are real ones, taken in north and south lines, throughthe islands with their barrier-reefs, of Vanikoro, Gambier,and Maurua; and they are laid down, both vertically andhorizontally, on the same scale of a quarter of an inch toa mile.

It should be observed that the sections might have beentaken in any direction through these islands, or through

[picture]

many other encircled islands, and the general features wouldhave been the same. Now, bearing in mind that reef-buildingcoral cannot live at a greater depth than from 20 to 30fathoms, and that the scale is so small that the plummets onthe right hand show a depth of 200 fathoms, on what arethese barrier-reefs based? Are we to suppose that eachisland is surrounded by a collar-like submarine ledge of rock,or by a great bank of sediment, ending abruptly where thereef ends?

If the sea had formerly eaten deeply into the islands,before they were protected by the reefs, thus havingleft a shallow ledge round them under water, the presentshores would have been invariably bounded by great precipices,but this is most rarely the case. Moreover, on thisnotion, it is not possible to explain why the corals shouldhave sprung up, like a wall, from the extreme outer marginof the ledge, often leaving a broad space of water within,too deep for the growth of corals. The accumulation of awide bank of sediment all round these islands, and generallywidest where the included islands are smallest, is highlyimprobable, considering their exposed positions in the centraland deepest parts of the ocean. In the case of the barrier-reefof New Caledonia, which extends for 150 miles beyondthe northern point of the islands, in the same straight linewith which it fronts the west coast, it is hardly possible tobelieve that a bank of sediment could thus have beenstraightly deposited in front of a lofty island, and so farbeyond its termination in the open sea. Finally, if we lookto other oceanic islands of about the same height and of similargeological constitution, but not encircled by coral-reefs,we may in vain search for so trifling a circumambientdepth as 30 fathoms, except quite near to their shores; forusually land that rises abruptly out of water, as do most ofthe encircled and non-encircled oceanic islands, plungesabruptly under it. On what then, I repeat, are these barrierreefs based? Why, with their wide and deep moat-like channels,do they stand so far from the included land? We shallsoon see how easily these difficulties disappear.

We come now to our third class of Fringing-reefs, whichwill require a very short notice. Where the land slopes abruptlyunder water, these reefs are only a few yards in width,forming a mere ribbon or fringe round the shores: wherethe land slopes gently under the water the reef extendsfurther, sometimes even as much as a mile from the land;but in such cases the soundings outside the reef always showthat the submarine prolongation of the land is gently inclined.In fact, the reefs extend only to that distance from the shore,at which a foundation within the requisite depth from 20 to30 fathoms is found. As far as the actual reef is concerned,there is no essential difference between it and that forminga barrier or an atoll: it is, however, generally of less width,and consequently few islets have been formed on it. Fromthe corals growing more vigorously on the outside, and fromthe noxious effect of the sediment washed inwards, the outeredge of the reef is the highest part, and between it and theland there is generally a shallow sandy channel a few feet indepth. Where banks or sediments have accumulated near tothe surface, as in parts of the West Indies, they sometimesbecome fringed with corals, and hence in some degree resemblelagoon-islands or atolls, in the same manner as fringing-reefs,surrounding gently sloping islands, in some degree resemblebarrier-reefs.

No theory on the formation of coral-reefs can be consideredsatisfactory which does not include the three great

[picture]

classes. We have seen that we are driven to believe in thesubsidence of those vast areas, interspersed with low islands,of which not one rises above the height to which the wind andwaves can throw up matter, and yet are constructed by animalsrequiring a foundation, and that foundation to lie atno great depth. Let us then take an island surrounded byfringing-reefs, which offer no difficulty in their structure;and let this island with its reefs, represented by the unbrokenlines in the woodcut, slowly subside. Now, as the islandsinks down, either a few feet at a time or quite insensibly,we may safely infer, from what is known of the conditionsfavourable to the growth of coral, that the living masses,bathed by the surf on the margin of the reef, will soon regainthe surface. The water, however, will encroach little by littleon the shore, the island becoming lower and smaller, and thespace between the inner edge of the reef and the beachproportionately broader. A section of the reef and island inthis state, after a subsidence of several hundred feet, is givenby the dotted lines. Coral islets are supposed to have beenformed on the reef; and a ship is anchored in thelagoon-channel. This channel will be more or less deep,according to the rate of subsidence, to the amount of sedimentaccumulated in it, and to the growth of the delicately branchedcorals which can live there. The section in this state resemblesin every respect one drawn through an encircled island: in fact,it is a real section (on the scale of .517 of an inch to a mile)through Bolabola in the Pacific. We can now at once seewhy encircling barrier-reefs stand so far from the shoreswhich they front. We can also perceive, that a line drawnperpendicularly down from the outer edge of the new reef,to the foundation of solid rock beneath the old fringing-reef,will exceed by as many feet as there have been feet ofsubsidence, that small limit of depth at which the effectivecorals can live: -- the little architects having built up theirgreat wall-like mass, as the whole sank down, upon a basisformed of other corals and their consolidated fragments.Thus the difficulty on this head, which appeared so great,disappears.

If, instead of an island, we had taken the shore of a continentfringed with reefs, and had imagined it to have subsided,a great straight barrier, like that of Australia or NewCaledonia, separated from the land by a wide and deep channel,would evidently have been the result.

Let us take our new encircling barrier-reef, of which thesection is now represented by unbroken lines, and which, asI have said, is a real section through Bolabola, and let it goon subsiding. As the barrier-reef slowly sinks down, thecorals will go on vigorously growing upwards; but as theisland sinks, the water will gain inch by inch on the shore --the separate mountains first forming separate islands within

[picture]

one great reef -- and finally, the last and highest pinnacledisappearing. The instant this takes place, a perfect atollis formed: I have said, remove the high land from within anencircling barrier-reef, and an atoll is left, and the land hasbeen removed. We can now perceive how it comes thatatolls, having sprung from encircling barrier-reefs, resemblethem in general size, form, in the manner in which they aregrouped together, and in their arrangement in single ordouble lines; for they may be called rude outline charts ofthe sunken islands over which they stand. We can furthersee how it arises that the atolls in the Pacific and IndianOceans extend in lines parallel to the generally prevailingstrike of the high islands and great coast-lines of thoseoceans. I venture, therefore, to affirm, that on the theory ofthe upward growth of the corals during the sinking of theland,

It may be asked, whether I can offer any direct evidenceof the subsidence of barrier-reefs or atolls; but it must beborne in mind how difficult it must ever be to detect amovement, the tendency of which is to hide under water the partaffected. Nevertheless, at Keeling atoll I observed on allsides of the lagoon old cocoa-nut trees undermined and falling;and in one place the foundation-posts of a shed, whichthe inhabitants asserted had stood seven years before justabove high-water mark, but now was daily washed by everytide: on inquiry I found that three earthquakes, one of themsevere, had been felt here during the last ten years. AtVanikoro, the lagoon-channel is remarkably deep, scarcelyany alluvial soil has accumulated at the foot of the loftyincluded mountains, and remarkably few islets have beenformed by the heaping of fragments and sand on the wall-likebarrier reef; these facts, and some analogous ones, ledme to believe that this island must lately have subsided andthe reef grown upwards: here again earthquakes are frequentand very severe. In the Society archipelago, on theother hand, where the lagoon-channels are almost choked up,where much low alluvial land has accumulated, and where insome cases long islets have been formed on the barrier-reefs-- facts all showing that the islands have not very latelysubsided -- only feeble shocks are most rarely felt. In thesecoral formations, where the land and water seem strugglingfor mastery, it must be ever difficult to decide between theeffects of a change in the set of the tides and of a slightsubsidence: that many of these reefs and atolls are subject tochanges of some kind is certain; on some atolls the isletsappear to have increased greatly within a late period; onothers they have been partially or wholly washed away. Theinhabitants of parts of the Maldiva archipelago know thedate of the first formation of some islets; in other parts, thecorals are now flourishing on water-washed reefs, whereholes made for graves attest the former existence of inhabitedland. It is difficult to believe in frequent changes in thetidal currents of an open ocean; whereas, we have in theearthquakes recorded by the natives on some atolls, and inthe great fissures observed on other atolls, plain evidence ofchanges and disturbances in progress in the subterraneanregions.

It is evident, on our theory, that coasts merely fringed byreefs cannot have subsided to any perceptible amount; andtherefore they must, since the growth of their corals, eitherhave remained stationary or have been upheaved. Now, itis remarkable how generally it can be shown, by the presenceof upraised organic remains, that the fringed islands havebeen elevated: and so far, this is indirect evidence in favourof our theory. I was particularly struck with this fact, whenI found, to my surprise, that the descriptions given by MM.Quoy and Gaimard were applicable, not to reefs in generalas implied by them, but only to those of the fringing class;my surprise, however, ceased when I afterwards found that,by a strange chance, all the several islands visited by theseeminent naturalists, could be shown by their own statementsto have been elevated within a recent geological era.

Not only the grand features in the structure of barrier-reefsand of atolls, and to their likeness to each other in form,size, and other characters, are explained on the theory ofsubsidence -- which theory we are independently forced toadmit in the very areas in question, from the necessity offinding bases for the corals within the requisite depth -- butmany details in structure and exceptional cases can thus alsobe simply explained. I will give only a few instances. Inbarrier-reefs it has long been remarked with surprise, thatthe passages through the reef exactly face valleys in theincluded land, even in cases where the reef is separatedfrom the land by a lagoon-channel so wide and so muchdeeper than the actual passage itself, that it seems hardlypossible that the very small quantity of water or sedimentbrought down could injure the corals on the reef. Now,every reef of the fringing class is breached by a narrowgateway in front of the smallest rivulet, even if dry duringthe greater part of the year, for the mud, sand, or gravel,occasionally washed down kills the corals on which it isdeposited. Consequently, when an island thus fringed subsides,though most of the narrow gateways will probablybecome closed by the outward and upward growth of thecorals, yet any that are not closed (and some must always bekept open by the sediment and impure water flowing out ofthe lagoon-channel) will still continue to front exactly theupper parts of those valleys, at the mouths of which theoriginal basal fringing-reef was breached.

We can easily see how an island fronted only on one side, or onone side with one end or both ends encircled by barrier-reefs,might after long-continued subsidence be convertedeither into a single wall-like reef, or into an atoll with agreat straight spur projecting from it, or into two or threeatolls tied together by straight reefs -- all of whichexceptional cases actually occur. As the reef-building coralsrequire food, are preyed upon by other animals, are killed bysediment, cannot adhere to a loose bottom, and may be easilycarried down to a depth whence they cannot spring up again,we need feel no surprise at the reefs both of atolls andbarriers becoming in parts imperfect. The great barrier ofNew Caledonia is thus imperfect and broken in many parts;hence, after long subsidence, this great reef would not produceone great atoll 400 miles in length, but a chain orarchipelago of atolls, of very nearly the same dimension withthose in the Maldiva archipelago. Moreover, in an atoll oncebreached on opposite sides, from the likelihood of the oceanicand tidal currents passing straight through the breaches, itis extremely improbable that the corals, especially duringcontinued subsidence, would ever be able again to unite therim; if they did not, as the whole sank downwards, one atollwould be divided into two or more. In the Maldiva archipelagothere are distinct atolls so related to each other inposition, and separated by channels either unfathomable orvery deep (the channel between Ross and Ari atolls is 150fathoms, and that between the north and south Nillandooatolls is 200 fathoms in depth), that it is impossible to lookat a map of them without believing that they were oncemore intimately related. And in this same archipelago,Mahlos-Mahdoo atoll is divided by a bifurcating channelfrom 100 to 132 fathoms in depth, in such a manner, thatit is scarcely possible to say whether it ought strictly tobe called three separate atolls, or one great atoll not yetfinally divided.

I will not enter on many more details; but I must remarkthat the curious structure of the northern Maldiva atollsreceives (taking into consideration the free entrance of thesea through their broken margins) a simple explanation inthe upward and outward growth of the corals, originallybased both on small detached reefs in their lagoons, such asoccur in common atolls, and on broken portions of the linearmarginal reef, such as bounds every atoll of the ordinaryform. I cannot refrain from once again remarking on thesingularity of these complex structures -- a great sandy andgenerally concave disk rises abruptly from the unfathomableocean, with its central expanse studded, and its edgesymmetrically bordered with oval basins of coral-rock justlipping the surface of the sea, sometimes clothed withvegetation, and each containing a lake of clear water!

One more point in detail: as in the two neighbouringarchipelagoes corals flourish in one and not in the other, andas so many conditions before enumerated must affect theirexistence, it would be an inexplicable fact if, during thechanges to which earth, air, and water are subjected, thereef-building corals were to keep alive for perpetuity on anyone spot or area. And as by our theory the areas includingatolls and barrier-reefs are subsiding, we ought occasionally tofind reefs both dead and submerged. In all reefs, owing to thesediment being washed out of the lagoon-channel to leeward,that side is least favourable to the long-continued vigorousgrowth of the corals; hence dead portions of reef notunfrequently occur on the leeward side; and these, though stillretaining their proper wall-like form, are now in severalinstances sunk several fathoms beneath the surface. TheChagos group appears from some cause, possibly from thesubsidence having been too rapid, at present to be much lessfavourably circumstanced for the growth of reefs than formerly:one atoll has a portion of its marginal reef, nine milesin length, dead and submerged; a second has only a fewquite small living points which rise to the surface, a thirdand fourth are entirely dead and submerged; a fifth is amere wreck, with its structure almost obliterated. It isremarkable that in all these cases, the dead reefs and portionsof reef lie at nearly the same depth, namely, from six toeight fathoms beneath the surface, as if they had been carrieddown by one uniform movement. One of these "half-drownedatolls," so called by Capt. Moresby (to whom Iam indebted for much invaluable information), is of vastsize, namely, ninety nautical miles across in one direction,and seventy miles in another line; and is in many respectseminently curious. As by our theory it follows that newatolls will generally be formed in each new area of subsidence,two weighty objections might have been raised,namely, that atolls must be increasing indefinitely in number;and secondly, that in old areas of subsidence each separateatoll must be increasing indefinitely in thickness, if proofsof their occasional destruction could not have been adduced.Thus have we traced the history of these great rings ofcoral-rock, from their first origin through their normalchanges, and through the occasional accidents of theirexistence, to their death and final obliteration.

In my volume on "Coral Formations" I have published amap, in which I have coloured all the atolls dark-blue, thebarrier-reefs pale-blue, and the fringing reefs red. Theselatter reefs have been formed whilst the land has beenstationary, or, as appears from the frequent presence ofupraised organic remains, whilst it has been slowly rising:atolls and barrier-reefs, on the other hand, have grown upduring the directly opposite movement of subsidence, whichmovement must have been very gradual, and in the case of atollsso vast in amount as to have buried every mountain-summit overwide ocean-spaces. Now in this map we see that the reefstinted pale and dark-blue, which have been produced by thesame order of movement, as a general rule manifestly standnear each other. Again we see, that the areas with the twoblue tints are of wide extent; and that they lie separate fromextensive lines of coast coloured red, both of whichcircumstances might naturally have been inferred, on the theoryof the nature of the reefs having been governed by the natureof the earth's movement. It deserves notice that in morethan one instance where single red and blue circles approachnear each other, I can show that there have been oscillationsof level; for in such cases the red or fringed circles consistof atolls, originally by our theory formed during subsidence,but subsequently upheaved; and on the other hand, some ofthe pale-blue or encircled islands are composed of coral-rock,which must have been uplifted to its present height before thatsubsidence took place, during which the existing barrier-reefsgrew upwards.

Authors have noticed with surprise, that although atollsare the commonest coral-structures throughout some enormousoceanic tracts, they are entirely absent in other seas,as in the West Indies: we can now at once perceive thecause, for where there has not been subsidence, atolls cannothave been formed; and in the case of the West Indies andparts of the East Indies, these tracts are known to have beenrising within the recent period. The larger areas, colouredred and blue, are all elongated; and between the two coloursthere is a degree of rude alternation, as if the rising of onehad balanced the sinking of the other. Taking into considerationthe proofs of recent elevation both on the fringedcoasts and on some others (for instance, in South America)where there are no reefs, we are led to conclude that thegreat continents are for the most part rising areas: and fromthe nature of the coral-reefs, that the central parts of thegreat oceans are sinking areas. The East Indian archipelago,the most broken land in the world, is in most parts an areaof elevation, but surrounded and penetrated, probably inmore lines than one, by narrow areas of subsidence.

I have marked with vermilion spots all the many knownactive volcanos within the limits of this same map. Theirentire absence from every one of the great subsiding areas,coloured either pale or dark blue, is most striking and notless so is the coincidence of the chief volcanic chains withthe parts coloured red, which we are led to conclude haveeither long remained stationary, or more generally have beenrecently upraised. Although a few of the vermilion spotsoccur within no great distance of single circles tinted blue,yet not one single active volcano is situated within severalhundred miles of an archipelago, or even small group ofatolls. It is, therefore, a striking fact that in the Friendlyarchipelago, which consists of a group of atolls upheavedand since partially worn down, two volcanos, and perhapsmore, are historically known to have been in action. On theother hand, although most of the islands in the Pacific whichare encircled by barrier-reefs, are of volcanic origin, oftenwith the remnants of craters still distinguishable, not one ofthem is known to have ever been in eruption. Hence in thesecases it would appear, that volcanos burst forth into actionand become extinguished on the same spots, accordingly aselevatory or subsiding movements prevail there. Numberlessfacts could be adduced to prove that upraised organic remainsare common wherever there are active volcanos; but until itcould be shown that in areas of subsidence, volcanos wereeither absent or inactive, the inference, however probable initself, that their distribution depended on the rising orfalling of the earth's surface, would have been hazardous. Butnow, I think, we may freely admit this important deduction.

Taking a final view of the map, and bearing in mind thestatements made with respect to the upraised organic remains,we must feel astonished at the vastness of the areas, whichhave suffered changes in level either downwards or upwards,within a period not geologically remote. It would appearalso, that the elevatory and subsiding movements follownearly the same laws. Throughout the spaces interspersedwith atolls, where not a single peak of high land has beenleft above the level of the sea, the sinking must have beenimmense in amount. The sinking, moreover, whether continuous,or recurrent with intervals sufficiently long for thecorals again to bring up their living edifices to the surface,must necessarily have been extremely slow. This conclusion isprobably the most important one which can be deduced from thestudy of coral formations; -- and it is one which it isdifficult to imagine how otherwise could ever have beenarrived at. Nor can I quite pass over the probability of theformer existence of large archipelagoes of lofty islands,where now only rings of coral-rock scarcely break the openexpanse of the sea, throwing some light on the distribution ofthe inhabitants of the other high islands, now left standingso immensely remote from each other in the midst of thegreat oceans. The reef-constructing corals have indeedreared and preserved wonderful memorials of the subterraneanoscillations of level; we see in each barrier-reef aproof that the land has there subsided, and in each atoll amonument over an island now lost. We may thus, like untoa geologist who had lived his ten thousand years and kept arecord of the passing changes, gain some insight into thegreat system by which the surface of this globe has beenbroken up, and land and water interchanged.