'' : .t-s*£>% \ -•4 v r ^■.^ ■*Ali .'W> ** ; 3f v -■ "v /. ,* ,. H *•» '* $»"?% -4. *-v " v *> **&*. ^'"•^- .■■-"'*, !*&&#' *< v • &K' ,4 .4 fc* : .. r?*N* HARVARD UNIVERSITY Library of the Museum of Comparative Zoology MEMOIRS OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. VOL. XXIX. CAMBRIDGE, U.S.A. PRINTED FOR THE MUSEUM. 1903. MEMOIRS OF THE |T MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. VOL. XXIX. CAMBRIDGE, U.S.A. PRINTED FOR THE MUSEUM. 1903. University Press : John Wilson and Son, Cambridge, U.S.A. CONTENTS. THE CORAL REEFS OF THE MALDIVES. By Alexander Agassiz. pp. i-xxv, 1-168. 82 Plates. December, 1903. One Volume Text. One Volume Plates. . Uttmoirs of % Utastum of Comgaratib* ^oblogg AT HARVARD COLLEGE. Vol. XXIX. THE CORAL REEFS OF THE MALDIVES. By ALEXANDER AGASSIZ. WITH EIGHTY-TWO PLATES. TEXT. CAMBRIDGE, U.S.A.: Printeo for tfje JHuseum. December, 1903. TABLE OF CONTENTS. Page Introduction ix Hydrography of the Maldives. Plates 1-8 c 1 Topography of the Indian Ocean West of Malabar. Plates 7, 8 5 Analysis of the Soundings taken by the "Amra." Plates 1-8 c 12 List of Soundings 29 Description of the Atolls : North Male. Plates 1, 3, 4 ; 8 a, fig. 6 ; 8 b, figs. 11, 19 ; 8 c, fig. 24; 9-18 ; 19, fig. 1 35 Gaha Faro. Plates 1, 3 ; 8 a, figs. 6, 8 ; 8 b, fig. 11 53 Goifurfehendu. Plates 1, 3 ; 8 a, fig. 8 ; 8 b, fig. 10 ; 19, fig. 2 ; 20 54 Karidu. Plates 1, 3; 8b, fig. 11; 21 55 Malosmadulu Plateau. Plates 1,3; 8 a, fig. 5 ; 8 b, figs. 9, 10 ; 8 c, fig. 28 ; 22-30 ; 31, fig. 1 56 South Malosmadulu. Plates 1, 3 ; 8 b, fig. 10: 8 c, fig. 28; 22-28; 29, fig. 2 . . 59 Middle Malosmadulu. Plates 1, 3 ; 8 b, figs. 9, 10 ; 29, fig. 1 . : 64 North Malosmadulu. Plates 1,3; 8a, fig. 5; 8b, fig. 9; 8c, fig. 26; 30; 31, fig. 1 65 Makunudu. Plates 1, 2 ; 8 a, fig. 7 ; 31, fig. 2 ; 32 ; 33, fig. 1 71 Fadiffolu. Plates 1, 3 ; 8 a, fig. 4 ; 8 b, fig. 11 ; 8 c, fig. 28 ; 33, fig. 2 ;• 34, fig. 1 . 73 Tiladummati-Miladummadulu Plateau. Plates 1, 2, 3; 8 a, figs. 2-5, 7; 34, fig. 2; 35-53 ; 79, fig. 3 78 Miladummadulu. Plates 1, 2, 3 ; 8 a, figs. 4, 5, 7 ; 34, fig. 2 ; 35-44 ; 79, fig. 3 . 83 Tiladummati. Plates 1,2; 8 a, figs. 2, 3; 45-53 93 Ihavandiffulu. Plates 1, 2 ; 8 a, figs. 1, 3 ; 54, 55, fig. 1 100 Toddu. Plates 1, 4 102 Rasdu. Plates 1, 4 102 Ari. Plates 1, 4 ;- 8 b, figs. 12, 14 ; 8 c, fig. 25 ; 55, fig. 2 ; 56, 57, 58, fig. 2 . . . 103 North Nilandu. Plates .1, 4, 5 ; 8 b, fig. 14 ; 59-61 ; 78, fig. 1 107 South Nilandu. Plates 1, 5; 8 b, figs. 13, 17 ; 62, 63, 64, fig. 2 ; 78, fig. 2 . . . 110 South Male. Plates 1, 4; 8b, figs. 12, 16; 8c, figs. 19, 27; 79, fig. 2 .... 113 Felidu. Plates 1, 4, 5; 8b, figs. 15, 16 114 Wataru Reef. Plates 1, 5 ; 8b, fig. 15 ; 58, fig. 1 118 Mulaku. Plates 1, 5; 8b, figs. 13, 15, 18; 64, fig. 1 ; 65 118 Kolumadulu. Plates 1, 5 ; 8 b, figs. 17, 18 ; 8 c, fig. 20 ; 66 122 Haddummati. Plates 1 , 6 ; 8 c, figs. 20, 22 ; 67-70, 77, fig. 1 127 Suvadiva. Plates 1, 6 ; 8 c, figs. 22, 23 ; 71-75 ; 79, fig. 1 134 Viii TABLE OF CONTENTS. Description of the Atolls ( Continued) : Page Fua Mulaku. Plates 1, 6 ; 8 c, fig. 23 145 Addu. Plates 1, 6; 8c, figs. 21, 23; 76, 77, fig. 2 145 Temperature of the Lagoons 149 Bottom Deposits in the Lagoons 152 The Pelagic Fauna of the Maldives 155 INDEX 161 INTRODUCTION. It had always been my intention, on the completion of my explorations of the atoll and coral regions of the Pacific, to make an expedition to the Maldives, the only great group of atolls I had not visited, and which promised interesting results ; the coral reefs of the Maldives differing radi- cally, according to the charts, from every region I had examined. 1 We spent part of December, 1901, and January, 1902, in exploring the Maldives. The steamer "Amra," chartered from the British India Steam Naviga- tion Company, proved a most serviceable vessel for examining the coral reefs of the Maldives. 3 She was commanded by Captain William Pigott, R. N. R,., who proved himself a most careful navigator among the maze of atolls through which we steamed for over sixteen hundred miles. Both he and the officers of the "Amra" showed the greatest interest in the objects of the Expedition. Captain Pigott took charge of the sounding machine, and himself superintended all the soundings we took (more than eighty in number) ; he became exceedingly skilful at this work. Several of the deeper soundings were taken successfully under most trying circum- stances. The " Amra " was equipped with a deep-sea Lucas sounding machine built for me by the Telegraph Construction and Maintenance Company of London. The machine differs radically from the American type of sounding machine developed by Captain Sigsbee with which I was familiar, and which I had in commission on all my former expeditions. 1 This is the last of the series of monographs I shall publish on Coral Reefs, and I hope, as soon as practicable, to give a resume and connected account of the results obtained during my expeditions to all the important coral regions of the Atlantic, Pacific, and Indian Oceans. 2 A sketch of the work of the Expedition was published at Colombo in the "Observer," January 29, 1902, and in the American Journal of Science for March of the same year. X INTRODUCTION. Excellent as is the Sigsbee machine, the Lucas sounder has some advantages in its compactness, it being self-contained and practically automatic. But its greatest advantage lies in the use of malleable wire for sounding in place of the hard-drawn wire in use in American machines. This greatly simplifies the making of splices and lessens immensely the danger of kinking while handling the wire. We had in addition a Sir William Thomson sounding machine for moderate depths in the lagoons or at our anchorages. I also placed on board a steam winch of the Bacon pattern with a drum large enough to hold six hundred to eight hundred fathoms of wire dredging rope. This winch was used for deep-sea towing down to one hundred and fifty fathoms and for the few hauls of the dredge we found time to take. Dr. W. McM. Woodworth, my son Maximilian, and Mr. H. B. Bigelow accompanied me as assistants. My son and Dr. Woodworth took a large number of photographs. Dr. Woodworth had general charge of the collec- tions. They were intentionally limited, as we could not hope in the short time at our command to add much to the material obtained by Mr. J. Stanley Gardiner during his prolonged stay at the Maldives and Laccadives. Mr. Bigelow collected thirty species of Medusae, interesting mainly for the geographical distribution of the genera represented. As might be expected, they were principally Hydroids, exclusive of the Sipho- nophores, Discophores, and Ctenophores. The pelagic Fauna was at times very rich, and some of our deeper hauls were most productive. The surface hauls inside the lagoons were also frequently very rich, far more than in the lagoons of any other coral-reef region I have visited. This may be accounted for from the open condition of the lagoons of the composite atolls of the Maldives. No attempt was made to collect any plants, the collections of Mr. Gardiner having supplied the material for an exhaustive list of the Flora of the Maldives. 1 Messrs. Willis and Gardiner have published a most interesting account of the Flora of the Maldives and of Minikoi based upon the extensive collections made by Mr. Gardiner. They have made a careful analysis of the flora, 1 The Botany of the Maldive Islands, by J. C. Willis and J. Stanley Gardiner, Ann. Royal Botanic Gardens, Peradeniya, Vol. I., Part II., December., 1901. INTRODUCTION. xi showing the sources from which it has been derived and the manner in which it has reached the islands. Messrs. Willis and Gardiner consider the ilora of the islands to be recent, and that if the Maldives " were now to be submerged and to rise again above the waves, they would probably acquire in time a flora almost identical with that which they now possess." 1 They look upon the effect of oceanic currents as the most important of all the agencies in stocking oceanic islands. To the considerable direct trade which exists between the Malabar Coast, Calcutta, and Male is probably due the introduction of many Indian weeds 2 and a large number of cultivated plants and trees. At Huludu (Addu) there are also " traditions of an Arabian trade in times past." 3 The natives of Tiladummati and Ihavandiffulu are most enterprising, and they have undoubtedly introduced a number of plants and trees in the northern groups of the Archipelago. For an archipelago with such active traders, the agency of man has undoubtedly been an effective factor in modifying the flora of the group and in introducing many plants and trees both from India and Arabia. It was interesting to watch the native ships taking advantage of the shelter of the islands to beat northward against the northeast monsoon in comparatively smooth water, until they reached Ihavandiffulu and could fetch the southern coast of India and Ceylon. In discussing the origin of the flora of other oceanic islands of the In- dian Ocean, Messrs. Willis and Gardiner come to the conclusion that " there is probably no need to assume complete land connections across the Indian Ocean to explain the flora of its islands, or perhaps even the affinities of the African and Indian floras." We might even go a step further and say that there is no need of assuming the former existence of larger islands like those sketched by Mr. Gardiner as occupying areas indicated by the two thousand fathom line. 4 We made a fair ethnological collection, the better part of which we owe to the kindness of His Highness Muhammadu Imadudin, the Sultan of the Maldives. I have also to thank the Agents of the British India Steam Navigation Company at Colombo, Messrs. Bois Brothers and Company, 1 Loc. cit., Willis and Gardiner, p. 162. 3 Loc. cil., p. 123. a Loc. cit., p. 140. * Loc. cit., Willis and Gardiner, PI. II. fig. 1. xii INTRODUCTION. and Captain Fenwick for having carried out my instructions regarding the equipment of the " Amra." We found the steamer ready for us on our arrival at Colombo. To the Right Honorable Joseph Chamberlain I am indebted for his kind- ness in giving me letters to the Government officials at Ceylon, and for writing to His Excellency Sir West Ridgeway in regard to our proposed visit. The Governor kindly wrote to the Sultan of the Maldives notifying him of the objects of our Expedition, and also gave us letters to the Sultan. His Highness took great interest in our work, gave us a circular letter to the chiefs of the various atolls, and in addition sent a representative and an inter- preter to accompany us on our trip. Thanks to this, we were everywhere received with the greatest cordiality. The Sultan's Barge We started on our explorations from Male after having, on making the atoll, obtained a glimpse of the east coast of North Male as we skirted the atoll, from Mirufuri to Male Island itself. This stretch of coast together with the islands near Male and the adjoining faros 1 of the lagoon to the northwest of Male contain all that is most characteristic of the atolls of the Maldives. A glance at the Chart gave us an approximate idea of the prob- lems to be solved in the study of the coral reefs of the group. 1 Faro is the name given by the natives to the small atolls which rise in the interior of the large lagoons or are found on their rims. It may be convenient to retain, as Mr. Gardiner has suggested, the Maldivian terms " faro " and " velu," * for the small atolls or ring-shaped reefs with their enclosed basin. Gardiner extends the term " velu " to the deep pools in the linear circumscribing reefs of many of the banks. This is likely to lead to confusion, as the mode of formation of the last is not necessarily the same as that of the velu proper. We are also confronted with the difficulty of distinguishing a faro from an atoll when the former constitutes a separate bank which may or may not be larger than a faro forming part of a land rim and larger than many an atoll in the Tropical Pacific. * The Maldives and Laccadives Archipelagoes, p. 155. INTRODUCTION. xiii After examining North Male 1 we passed to Ari, then to North and South Nilandu, crossing to Mulaku, making our way to Kolumadulu, to Haddum- mati, to Suvadiva, and to Addu, the southernmost atoll of the Maldives. On account of the heavy sea we were unable either on our way south or north to stop at Fua Mulaku, a small island between Suvadiva and Addu ; but, judging from the chart and such accounts as I could obtain, it probably does not differ from similar islands in the Maldives. Going north we modified our course so as to visit the faces of the atolls we had not seen on our way south, and to cross the lagoons from a different direction, taking thus a bird's eye-view of the atolls and islands. Our route was further daily modified according to the position of the sun to enable us to navigate the interior of an atoll in safety, or to take photographs as we passed, without loss of time. On our way north we examined the atolls of the eastern chain which we had not seen : Wataru, Felidu, and South Male. From Male Island we ex- amined the western parts of North Male which we had not visited before, passed on to Gaha Faro, to Kariclu, to Fadiffolu, to South Malosmadulu, to Goifurfehendu, to Middle and North Malosmadulu, to Miladummadulu, crossing to Makunudu and Tiladummati. We left the Maldives through one of the passages on the east face of Ihavandiffulu, the northernmost atoll of the group, after having steamed nearly sixteen hundred miles among the atolls of the Maldives. Although the waters within the groups of atolls of the Maldives have been most carefully sounded by Commander Moresby and Lieutenant Powell, 2 yet very little was known of the depths on the two sea faces of the great plateau upon which the atolls of the Maldives have developed, or of the depths in the channels separating them, until Mr. Gardiner's visit. He took a number of soundings across the central basin and a few of the chan- nels separating the Central Maldives. 3 The soundings of the Admiralty Charts give an excellent idea of the topography of the bottom of the lagoons of the composite atolls ; their greatest depth is not much more than forty 1 PI. 1 shows the track of the " Amra " through the Maldives. 2 B. A. Charts Nos. 66 a-c. 8 The Fauna and Geography of the Maldive and Laccadive Archipelagoes, edited by Mr. J. Stanley Gardiner, Vol. I., Part I., p. 19, and Introduction to the above, pp. 10, 11 ; Part II., p. 150, PI. X. x i v INTRODUCTION. fathoms. The depths indicate considerable variation over the bottom, and in some regions these changes in depth are very abrupt, from eight to over twenty fathoms in short distances. The character of the bottom varies greatly according to the locality and its vicinity to gaps, to passes, to islands or islets or sand-bars. In many cases the bottom is hard, swept clean by the currents, or covered with fragments coated with Nullipores, or it is cov- ered with Corallines or made up of fragments of broken corals, or of coarse or fine coral sand. On one occasion the claspers brought up a piece of Mil- lepore cut off from a living cluster from a depth of thirty-nine fathoms. This is an unusual depth for a reef builder, as in the Maldives the reef corals rarely extend below seventeen fathoms ; twelve fathoms is the usual depth I have observed, there usually begin the sand lanes and patches which sepa- rate them, and finally end in covering the bottom. We did not attempt to check any of the soundings on the Admiralty Charts. Within the atolls they exist in sufficient number for all theoretical and practical purposes, and for a party not sufficiently numerous and not properly equipped for the most delicate surveying operations it would have been hopeless to add any exact information to that already existing. Our explorations were immensely facilitated from the existence of the admirable charts published by the Admiralty of the Survey of the Maldives by Commander Moresby and Lieutenant Powell from 1834 to 1836. The accuracy of these charts is something wonderful when we remember the conditions under which the surveys were made nearly seventy years ago, with sailing vessels and rowboats. In our extensive and intricate navigation among the Maldives we were guided absolutely by these charts and never found them in error. Of course some minor changes have occurred on the reef flats, such as the disappearance of an islet or of a bank, or the addition of a sand-bar and the junction of adjoining islands or islets on the same reef flats. Otherwise the charts stand to-day as they did seventy years ago, a monument to the unsurpassed skill of the surveyors of those days. It was a simple task to pick out one's work in each atoll by an examination of the chart, and thus much time was saved. I have endeavored to adopt the spelling of the Admiralty Charts revised to May, 1903, but undoubtedly a few discrepancies will have occurred from our constant use of the older editions. INTRODUCTION. XV In exploring any coral-reef region with a steamer a good deal of planning is necessary to enable one to see the interesting points of an atoll. The lagoon can be explored only with the sun in certain positions ; one cannot steam east in the morning or west in the afternoon. To take photographs successfully the eastern face of an atoll must be explored in the morning and the western in the afternoon, that the sun may strike shallows and banks to define their limits and contrast their coloring with that of the deeper navigable waters of the atoll ; and finally the programme must be so arranged as to reach an anchorage before sunset, and an anchorage so selected as to entail but little loss of time in leaving in the morning with the sun in a favorable position either for further exploration or to pass out of the channels of the atolls when the flats and banks flanking them are well defined and rendered plainly visible from the angle at which the sun shines upon them. A mere glance at the Admiralty Charts of the Maldives 1 cannot fail to show how very different in structure are Makunudu, Gaha Faro, Karidu, Goifurfehendu, Rasdu, Toddu, Wataru, Fua Mulaku, and Addu from such groups of atolls as North and South Male, Ari, North and South Nilandu, Felidu, the Malosmadulu atolls, Miladummadulu and its northern extension Tiladummati which might be called the Maldivian group of atolls par excel- lence. Both these groups may be contrasted again to such atolls or groups of atolls as Fadiffolu, Felidu, and Mulaku, which have as it were combina- tions or modifications characteristic of the Maldivian atolls with features common to a number of Pacific atolls. And finally they may be compared to a third class of atolls like Kolumadulu, Haddummati, and Suvadiva, which remind us of the larger atolls of the Pacific in the Marshall, Ellice, Gilbert or Caroline Islands ; atolls noted for the absence of shoals and of islands in the lagoons. The groups of small typical Maldivian atolls along the forty to thirty fathom line of the great Maldivian plateau form an agglomeration of small atolls along that belt resembling the great land-rim reef flats of the Pacific atolls, but they have grown up as distinct parts and are separated by deep channels. These small atolls vary in size from a couple of hundred feet in diameter to atolls of seven miles in length. In 1 B. A. Charts Nos. 66 a, b, c. xvi INTRODUCTION. such clusters of atolls as those of North and South Male, of Ari, of Malos- madulu, and of other groups, their arrangement is such as to form well- defined rims, reminding us of the rims formed by great reef flats such as are common in the Pacific. In such groups of atolls as Miladummadulu and its Tiladummati extension we have the key to a rational explanation of the formation of the atolls and groups of atolls in the Maldives. The two atolls I have just named are not atolls in any sense of the word. They are so ill-defined that their division for political purposes is marked on the chart by a mere dotted line. In fact, they are composed of a great number of small atolls, often separated by con- siderable distances, as much as five to ten miles, which have gradually grown up on that part of the Maldivian plateau from depths of twenty-five to thirty fathoms, and where they can be seen in all possible stages of growth. The shape of the rings (faros) or bars or flats is not necessarily circular ; it varies greatly and is indirectly controlled by the topography of the bottom. Some of the faros are elliptical, pear-shaped, or crescent-shaped, differing greatly in outline and dimensions on the outer lines of the composite atolls. Their outer slopes are covered with corals growing with great luxuriance from the edge or even from the surface of the flats of the ring to a depth of from eight to twelve or fifteen fathoms. The superb growth of corals found in all the lagoons of the Maldives is in marked contrast with the scanty growth of corals in the lagoons of the atolls of the Pacific. We have a simple explanation of this in the fact that the rim of the atolls in the Maldives is full of wide and deep passages. In fact, the extent of the passes is generally much larger than the space occupied by the small atolls (the atollons or faros) and reef flats. As soon as the rim flats of the rings have reached the surface, either wholly or in part, sand-bars begin to form ; these develop rapidly into islets, and finally into large islands more or less covered with scrub vegetation and bushes. The rings or faros either retain a central lagoon or it becomes partly or wholly filled up. In the former case they appear as small atolls with islands or islets on the reef flats ; in the next stage there is either a smaller lagoon on the lee face of a larger island, or else the island has grown to occupy the whole flat of the faro with only very narrow flats on the INTRODUCTION. xvii lagoon faces of which corals grow. The larger islands are often covered with fine vegetation, large trees occupying the space bordered by the outer belt of bushes growing close to the beach. The greater number of the many islands which dot the large lagoons of the Maldives have been formed in the manner described, and it is comparatively easy to trace the progress of their development in all the stages intermediate between a mere ring, not yet rising to the surface, and a steep to island with its rich vegetation such as we find either in the lagoons or on the outer edge of a Maldivian composite atoll. The small atolls which form the outer rim of the composite atolls owe their existence to the same causes, and their development can be as easily traced from a mere ring which has risen to the surface having more or less extensive flats on which islands or islets or sand-bars have been thrown up. The sea reef flats of the outer lines of atolls are usually wider than those of the lee face, and naturally so. From my observations I am inclined to look upon Nullipores much as does Mr. Gardiner, 1 as especially efficient in consolidating the corals and in the protection of reef platforms and flats. Nullipores also flourish on the sea faces of coral reefs to the greatest depths at which corals grow, but they are far more important below that depth in lagoons. As far as I have examined the outer slopes of the reef flats or faros of the various banks, they are covered with a growth of corals down to a depth of twenty to twenty-one fathoms, taking their greatest development at from five to nine or ten fathoms and rapidly disappearing from seven- teen to twenty-five fathoms, where the coral slope passes into a region barren of reef corals. Mr. Gardiner comes to the conclusion that about thirty fathoms is the extreme limit of the growth of the effective reef-building corals. He also noticed, what is generally the case in the Pacific, that, modified by local conditions, corals grow more luxuriantly as the water becomes shallower to within three to six fathoms of the surface. From his observations 2 Gardiner concludes Darwin is correct in stating that the extreme depth of reef corals is not more than twenty-five fathoms. There are great diffi- 1 Loc. cit., p. 177. a Loc. cit., p 325. xviii INTRODUCTION. culties in ascertaining the lateral rate of increase on the outer slope of a coral reef where it depends so largely upon local conditions within the bathymetrical range at which corals grow. There is great difference in the method of growth of corals on the reef flats or slopes of lagoons and of the exterior faces of coral reefs. The method of solidifying is quite different from the amount of sand on the flats within a lagoon, the excess of encrusting Nullipores on the sea faces, and the mass of coral fragments in constant movement on that face. The rate of growth of branching corals is quite rapid, as is readily seen in the channels, cleared out every few years by the natives, leading through the outer reefs to the canals existing between them and the land. Mr. Gardiner has calculated from his observations that the coral reefs of Hulule might grow at the rate of a fathom in sixty years. 1 On a bank of twenty-five to thirty fathoms depth 2 patches of corals would grow over its surface wherever conditions were especially advantageous, and the outer patches might, according to the direction of the current, form a rim, such parts as were lower forming the passages through the encircling reef. This, according to him, would account for such atolls as Addu, Goifurfehendu, Gaha Faro, Makunudu, Wataru, Rasdu, as well as for atolls of greater size, like Felidu, Ihavandiffulu, Mulaku, and others of the central Maldives where the corals built up also small atolls on the rims; while on the great Tiladummati-Miladummadulu Bank the oceanic conditions prevailing have maintained the individuality of the separate faros. Mr. Gardiner says no faro has reefs or shoals. This cannot be said of Rasdu, of the many faros of the northern part of Tiladummati, from Hanimadu to Kelai and to Gafuri ; they are full of small shoals and banks and sand-bars. Gardiner is right in looking upon the condition of the shoals and reefs of the bank as greatly influenced by the extent of the circulation of the water. If the bank is represented by a small lagoon it is of course limited, but if by a great lagoon as Suvadiva, or a more or less open bank 1 Loc. cit., p. 332. 2 It seems to me that the presence of such genera as Lophohelia on deep banks does not necessarily imply that they play so important a part in building up the deep coral banks of the Atlantic as to raise them from four hundred and thirty-five to fifty fathoms, the depth Mr. Gardiner assumes as the upper limit of growth of reef corals. INTRODUCTION. xix as in Tiladuramati and Miladummadulu, the circulation becomes almost oceanic. The reef and islands of portions of each bank will show the effect of the greater or more limited circulation. This is admirably shown by the condition of the reefs and heads described by Mr. Gardiner 1 as existing in the "jungle of reefs" in North Malos. Gardiner calls attention to the contrast of the slight growth of corals off a reef in an enclosed atoll or clusters like Haddummati or Kolumadulu to the luxuriant growth off a reef in open banks like Miladummadulu. This, in such large, comparatively open atolls, is often due quite as much to many minor local causes as to the prevailing oceanic conditions. In the Maldives the increase in size of the islands on the outer land rim of atolls goes on much as we have observed it in the Gilbert, Ellice, Marshall, and Paumotu Islands. Small islets or islands on the same reef flats are gradually united by the formation of sand spits on the lee face of the islands, thus forming bays on the sea face of an atoll; the spits gradually approach, become connected, and the filling up of the bay from the sea face unites adjacent islands ; their former disconnected state is merely indicated by a difference in the growth of the vegetation, a distinction which gradually disappears with years. The bay may also be formed on the lee side by the throwing up on the sea face of a bar on the edge of the reef flat between separate islands, and the bay may then be filled up both from the lee and weather side and thus unite separate islands or sand-bars. The existence of lagoons completely shut off from the sea in some of the atolls of the northern part of the Maldives is readily explained by their mode of formation ; this can be traced in all its stages, from the time the atoll con- sists of a crescent-shaped island occupying only a portion of the reef flat of the ring, the remainder of the reef flat of the ring still enclosing a compara- tively deep lagoon sometimes with six to seven fathoms of water. The island throws out spits from the horns of the crescent until there is only a narrow pass left between them, and finally this gap is closed by a sand or shingle beach and we have the ideal atoll, — a closed ring of land enclosing a deep lagoon, which exists so rarely but is always the atoll to which one refers when discussing the coral-reef question. These phenomena are well 1 Loc. cit., p. 166. xx INTRODUCTION. illustrated in the long line of crescent-shaped atolls occurring on the east side of Miladummadulu from Nalandu south as far as Bomasdu. Such a change, from an open crescentic island flanked by a lagoon to a closed land rim surrounding a lagoon, may take place with considerable rapidity. Bodu Mandu is represented on the chart as an open crescent-shaped island ; we found it seventy years later a closed land ring completely sur- rounding a small lagoon with a depth of two fathoms. The formation of fresh-water or brackish sinks edged with mangroves 1 in some of the atolls can be traced to the same process which has formed enclosed lagoons. They occur on Kendikolu, Ekasdu, Nalandu, Madidu, and Filadu. The sinks differ from the lagoons only in being shallow, hav- ing been cut off by spits and bars extending across portions of the adjacent reef flats, covered only by water of a couple of feet or more in depth, while the enclosed lagoons were cut off from lagoons of atolls of considerable depth, six to seven fathoms or more. What has been written above seems to me to point to the uselessness of our present definition of atolls. There is every possible gradation between a curved, crescent-shaped open bank of greater or less size and an absolutely closed ring of land surrounding a lagoon without direct communication with the sea. The evidence of a great number of atolls scattered on an extensive bank or plateau like that of Tiladummati and Miladummadulu shows that reef corals will grow upon any foundation where they find the proper depth, and that local conditions will determine their existence as fringing reefs, barrier reefs, or atolls. In fact, in the Maldives, reefs that once formed an atoll may in time, when the atoll or faro is changed into an island, become fringing reefs, — a transformation which is quite common both on the outer lines of islands and on islands in the interior of the larger basins. The composite atolls of the Maldives have arisen upon minor elevations upon the greater Maldive plateau which have given to the reef-building corals a base at the proper depth from which they have risen to the surface. In such smaller plateaus as North Male, Ari, and others, there is found on the secondary plateaus in their turn a number of bases on which the atolls and faros have grown. In the central and most of the northern plateaus the con- 1 Messrs. Gardiner and Willis have called attention to the scarcity of mangroves on the Maldives. INTRODUCTION. xx i dition of exposure to oceanic currents is such that an immense body of water is constantly flowing across the plateau during both the northeast and the southwest monsoons. Where the plateaus are smaller or not so open to the flow of currents as in such atolls as Addu, Karidu, Goifurfehendu, Gaha Faro, Wataru, Makunudu, and others, we have only a single atoll developed ; again in such plateaus as those upon which Kolumadulu, Haddummati, and Suvadiva have developed, the conditions are oceanic, more similar to those we find in the widely separated atolls of the Ellice, Gilbert, or Marshall Islands. At the same time the lagoons of the southern atolls are far less open to oceanic circulation than those of the northern plateaus, and thus we find fewer banks or islands in the lagoons, and only here and there a trace of those remarkable rings which are so characteristic a feature in Maldivian coral-reef scenery. I have seen nothing so striking in my experience of coral reefs as these rings with a light-colored rim standing out from the deep blue water surrounding them, like ghosts of an atoll which had sunk, and enclosing a lighter blue or emerald-colored lagoon indicative of its depth below the surface. The conditions existing at the Maldives are repeated to a certain extent on the Yucatan Plateau, where the Alacran reef, a regular atoll, rises from it at a depth of about thirty fathoms. It is true that it is the only atoll on this extensive plateau ; but there are also other irregularly sbaped patches of coral reefs. The absence of atolls may be traced to the fact that the plateau is not within an area of such regular trades as are the northeast and southwest monsoons in the region of the Maldives. The central banks in the Tonga Archipelago, the Haapai and Nomuka groups, with their irregularly shaped lagoon reefs and land-rim flats, remind us also somewhat of Maldivian conditions. The moderate strength of the prevailing winds in the Maldives has greatly influenced many of the characteristic features of its atolls. The effect of the southwest and of the northeast monsoons cannot be compared to that of the trades in the Pacific. We have nothing in the Maldives corre- sponding to the incessant breakers of the huge rollers which pound upon the reef flats of the Paumotus and of the atolls and barrier reefs of the central and western Pacific. The boulders thrown upon the reef flats are mere pyg- xxii INTRODUCTION. mies compared to the gigantic masses moved on some of the reef flats of the Pacific Islands. The boulder belts of the Maldives seem like a newly macadamized road as compared to the quarry blocks which often line miles of the beaches of some Pacific atolls. But the same forces are at work in the Maldives only on a diminutive scale, even during the prevalence of the southwest monsoon. The beaches both sand and shingle are as a whole remarkably steep ; they rarely rise to more than five or six feet, though in some of the northern atolls they are fully twelve feet high. Mr. Gardiner informs me he has seen dunes rising to twenty-eight or thirty feet in height. The Maldives are remarkable for the small amount of coral breccia and conglomerate or beach rock which is rehandled. The material moved upon the reef flats is nearly all coral sand, — a great contrast to the Pacific atolls, where the wearing and crushing and rehandling of the reef rock ledges, both old and recent, are the most prominent features in the modification of the reef flats of the Gilbert, Ellice, Marshall, and Paumotu Islands. There exist in the Maldives, it is true, a number of shingle beaches and boulder beaches on some of the islands exposed to the full action of the southwest monsoon both in the southern and northern part of the group ; but these beaches are insignificant as compared with the wide reef flats and beaches of the Pacific atolls, which are strewn for miles with shingle and boulders often of a great size. This marked difference is due in the one case to the comparatively feeble sea and rollers of the southwest monsoon, in the other to the huge rollers of the southeast and northeast trades acting continuously with a force which bears no comparison to the forces at work in the Indian Ocean. Mr. J. Stanley Gardiner spent over three months at Minikoi ; he has given a detailed account of the atoll. 1 According to him the sea is rapidly encroaching upon the east face of the island. He looks upon the Minikoi conglomerate, found at the base of the outer beach, as a reef rock similar to the conglomerate I have observed at so many of the Pacific atolls 2 ; show- ing that Minikoi, as well as the Pacific atolls, must at one time have been subject to a slight elevation (which he places at twenty-four feet). The 1 Loc. cit., p. 13. 2 A. Agassiz, Mem. M. C. Z., Vol. XXVHL, Pis. 8, 11, 12, 40, 63, 135, 136, 158, 163, 171. INTRODUCTION. xxiii boulder zone, as he calls the belt of masses of conglomerate, is what remains of the elevated coral rock which has not been washed or dissolved away. It is far more prominent at Minikoi than in the Maldives, but it does not attain the importance it has in many of the Pacific atolls. Judging from what I have observed in the Pacific, Mr. Gardiner is justified in assuming the ex- istence of a lagoon in Minikoi at the time of its elevation, and in considering the atoll to have been stationary for a considerable period of time on account of the broad reef flat round most of its circumference. 1 It is unfortunate that Mr. Gardiner should speak of a perfect atoll or of a perfect reef. He thus characterizes not only Minikoi, but also Kiltan, Chetlat, and Kavaratti among the Laccadives, and speaks of Cardamum as an atoll either in a late or an early stage. Judging from the charts and in analogy with the formation of Minikoi, he considers their present land con- figuration due to elevation and subsequent erosion, as has been shown to be the case for many Pacific atolls. We might indeed speak of Niau in the Paumotus as a perfect atoll where the rim is continuous, or of Pinaki or Maraki where we have only a small narrow boat passage opening through the otherwise continuous wooded rim of the atoll. These atolls represent a stage of growth which ultimately will end in the filling of the lagoon and leaving a central shallow sink or pool as the only trace of the former existence of a lagoon. While it is impracticable to subdivide atolls representing areas more or less circumscribed by belts of land and of reefs awash, or submerged, they vary from submerged banks through all the intermediate stages of banks with crescent-shaped rims with or without islands to the so-called perfect atolls with reef flats and passages and land rims. It seems to me mislead- ing to refer to Addu as differing from all the other banks of the Maldives and Laccadives in its more perfectly typical atoll form. Certainly Minikoi, Goifurfehendu, Rasdu, Gaha Faro, Wataru, Makunudu are as typical atolls .as Addu. Mr. Gardiner has given an account 2 of the mode of formation of atolls which calls upon solution as the essential factor for deepening the lagoon of I Mr. Gardiner lias given a number of sections of Minikoi on page 30, loc. cit. , showing the width of the reef flats at different points of the atoll. II Loc. cit., pp. 180, 181. • xxiv INTRODUCTION. an atoll after it has " become fairly perfect," ' though he recognizes " the outwash of mud as an important and direct subsidiary cause." 2 It has con- siderable bearing on the natui'e of the deposits on the outer face of an atoll. The amount of detritus washed out of a lagoon depends again in great measure on the position of an atoll. If sheltered or exposed to the sweep of the monsoons, the amount of silt will greatly vary with the depth of the lagoon, and will be more or less detrimental to the growth of corals wherever they happen to grow, on the lagoon or sea face. No one can cross one of the Fiji atolls of the Lau group 3 without being struck with the evidence of the action of solution on the islands, islets, and flats within the lagoon. Such atolls as Fulanga, Ongea, the Yangasa cluster, the Argo reef, and others of the Lau group show this effect most plainly. Islands like Kambara, Vatu Leile, Mango, are on a large scale, pitted, honeycombed, and under-cut by the combined wearing and solvent action of the sea. But with our present knowledge it seems difficult to assign to each factor its proportionate value. Wherever in the Maldives reef rock was examined I found it without exception of the most modern character, a few exposures as horses on the beaches and on the reef flats would seem to indicate a slight elevation of the Maldives. If the existing conditions at the Maldives have been brought about by subsidence, it is strange we should not find anywhere on this ex- tensive plateau, from the northernmost atolls as far as Kolumaduln, some trace, some outlier or some central rocks indicating the nature of the rocks composing the underlying plateau forming the base upon which the in- numerable atolls of the Maldives have been formed. The conditions are in many ways similar to those of the Lau Islands on the eastern plateau of Fiji. There the elevation has been considerable (to a height of one thousand feet), and everywhere indications are found of the character and age of the un- derlying strata. A similar condition exists in the Paumotus, where some of the tertiary elevated reefs attain an elevation of about three hundred feet. At the Maldives there is, however, only evidence of a very slight elevation. 1 This does not seem to hold good for Minikoi and other Laccadive atolls with shallow lagoons. 2 Loc. cit., p. 334. 8 A. Agassiz, Bull. M. C. Z., Vol. XXXIII., pp. 43, 88. INTRODUCTION. xxv In addition to the information to be derived from the charts and from Mr. Gardiner's Memoir on the Maldives, he was kind enough on two occa- sions, after his return from the Maldives, where he preceded me, to give me valuable details regarding many interesting localities. I can only close by expressing my admiration for the amount of work Mr. Gardiner was able to accomplish with the facilities at his command, and for the pluck and energy displayed by him and his companion, Mr. Cooper, in meeting, in native boats, the winds and currents which impeded their progress at every point, to say nothing of the constant discomforts to which they were subjected during their stay in the Maldives. I have especially to thank Admiral Sir J. W. Wharton for the interest he took in our expedition, and for allowing me access to those of the original charts of Moresby deposited in the London Hydrographic Office of the Admiralty. Newport, Rhode Island, July 15th, 1903. Street in Male. HYDROGRAPHY OF THE MALDIVES. Plates 1-8 c. It was greatly to our advantage to have as a guide the surveys pub- lished by the Admiralty. 1 We could obtain at a glance an accurate idea of the points of interest needing further examination and thus worked with great rapidity, losing no time in obtaining information by a tedious examination on shore of points of unknown value. This was specially important in a group of such simplicity and uniformity of structure as the Maldives. It seems to me as if the conditions of portions of the Maldive plateau were very similar to those existing upon the Yucatan Plateau inside the one-hundred-fathom line, where from a depth of about forty fathoms rise independent reefs such as the Triangles, Alacran Reef, the English, Opisbo, and Areas Banks. There is however a great difference in the number of atolls found in the two regions. On the Maldivian plateau in a distance of over four hundred and fifty miles there must be hundreds of atolls ; while on the Yucatan pla- teau, three or four only of the patches can be designated as atolls in about the same length. The number of atolls and islands existing on the Maldives is estimated to be as great as ten thousand. This is probably an exaggera- tion. I have counted roughly the number of atolls and rings and islands of Ari and other groups. There are three hundred and ninety-one in the Ari 1 The Maldives were surveyed during 1834-1836 by Commander Moresby and Lieutenants Powell and Young. These surveys were published by the Admiralty as sheets 66 a, 66 6, and 66 c (Pis. 1-6). The survey of Minikoi in 1891 by Commander Hoskyn is published as Admiralty Chart 2738; it also shows the relations of Minikoi to the Indian peninsula as well as to the Maldives and Laccadives. The Laccadives and the opposite coast of the Indian peninsula were surveyed in 1848 by officers of the East Indian Government, and by officers of the Marine Survey of India in 1891-92; see Admiralty Chart 2737 (PI. 8). H. O. Chart No. 1591 shows in a convenient form the connection of the Maldives and of Minikoi with the southern part of the Indian peninsula and with Ceylon (PI. 7). From Admiralty Chart 748 B (PI. 7) has been constructed. 2 THE CORAL REEFS OF THE MALDIVES. group, eighty-two in North Nilandu, one hundred and twenty-eight in South Nilandu ; but in this enumeration each lagoon reef as well as each reef hat was counted as a unit, and no account was taken of the many islands which sometimes occur on each flat. In groups like Ari, North Male, and North Nilandu, the increase in number could not be great. But in an atoll like Suvadiva and such groups as South Male, South Nilandu. Mulaku, Koluma- dulu, and Haddummati, where are often a great number of small islands forming a chain on the extensive reef flats, then the number of islands on the reef flats we have considered as units becomes very great. The coral reefs of the Haapai and of the Nomuka groups, the central part of the Tonga Islands, is another region which may be compared to the Maldives. Both the Tongan groups are independent banks separated by narrow and moderately deep channels from the banks to the north and south of them, and themselves separated by a channel of about five miles with depths varying from three hundred fathoms in the central part of the channel to three hundred and fifty and nearly five hundred at its eastern and western extremities. The Haapai and Nomuka groups hold to each other much the same relation which North and South Malosmadulu have. The plateaus of these groups are well defined by the steep falling off near the one-hund red-fathom line ; from their shallower parts rise secondary plateaus forming the coral reefs of Nomuka and Haapai. The circular or elongated lagoon reefs of the Haapai Plateau are flanked on the east and south by linear or circular reefs, while the Nomuka reefs are flanked by linear reefs, only on a part of the east face of the Nomuka Plateau. 1 The Maldive Islands (PI. 7) are about four hundred miles west of Ceylon. They extend north and south for a distance of about four hun- dred and seventy miles ; from 7 ° north latitude to fift}^ miles south of the Equator. The central part of the archipelago (PI. 1) consists of two chains of groups separated by a channel, varying in width from less than ten miles between Miladummadulu and North Malosmadulu to over twenty- five miles between Ari and South Male. 1 Admiralty Charts 2421, 3099, 3100, 474 ; A. Agassiz, Mem. M. C. Z. Vol. XXVIII., pp. 175, 192-197, Pis. 214-218. HYDROGRAPHY OF THE MALDIVES. 3 The northern extremity of the Maldives (PI. 1) consists of a single line of groups, Ihavandiffulu and Tiladummati flanked on the west by an isolated atoll, Makunudu. Then follows the eastern chain of the central Maldive basin, composed of the southern part of Miladuinmadulu, Fadiffolu, Gaha Faro, North and South Male, Felidu, Wataru Reef, and Mulaku, while the western chain of groups consists of North, Middle, and Southern Malosma- dulu, Goifurfehendu, Ari, and North and South Nilandu. Then comes the single southern row of groups (PI. 1), Kolumadulu, Haddummati, Suvadiva, and Addu. The depth of the channel between the eastern and western chains varies from five hundred and thirteen fathoms in the central part of the channel, between Miladummati and North Malosmadulu, to one hundred and ninety- four, the greatest depth between South Nilandu and Mulaku (PI. 1). The depth of the central basin increases gradually as we go north. It is two hundred and seventeen fathoms between Ari and South Male ; its northern part is separated by a shallow shelf west of North Male from the deeper basin to the east of Goifurfehendu, between it and Gaha Faro, where the greatest depth is two hundred and fifty-eight fathoms, increasing to two hundred and sixty-six, between South Malosmadulu and Fadiffolu and to five hundred and thirteen fathoms between North Malosmadulu and Miladummadulu. The depths of the channels separating the southern groups of the double chain north of Kolumadulu indicate that the southern part of the central plateau dips rapidly to the eastward (PI. 1). The greatest depth of Kuda- huvadu Channel between Mulaku and Kolumadulu is nearly six hundred and fifty fathoms, while it is only two hundred and fifty-one fathoms in the channel between South Nilandu and Kolumadulu. The eastward dip of the central basin can be traced from Mulaku to Miladummadulu from the depths of the western channels which are some- what shallower than those dividing the groups of the eastern face. On the western flank of the central basin the depth of the centre of the channel between South Nilandu and North Nilandu is 200 fathoms. On the eastern flank in the centre of South and North Wataru channels we find 253 and 283 fathoms. The depth of Ariyaddu Channel is 231 fathoms, that of 4 THE CORAL REEFS OF THE MALDIVES. Fulidu Channel, 374 fathoms. The greatest depth between Goifurfehendu and South Malosmadulu is 302 fathoms, that of North Kardiva Channel, 372 fathoms. From South Malosmadulu the central basin dips from 236 to 519 fathoms, and between Fadiffolu and Miladummadulu the greatest depth is 769 fathoms. The groups which constitute the western and eastern chains of the central Maldives (PI. 1) are elevations of considerable size ; they form the rim of an inner basin. The channels between the groups both on the east- ern and western faces of the basin correspond as far as we know very nearly in depth with those of the nearest part of the central basin. Our soundings (PI. 1) indicate that Kolumadulu is on the southern ex- tremity of that part of the plateau of the Maldives which extends unbroken from Ihavandiffulu south to Kolumadulu, or rather of a plateau of which the undulations are nowhere separated by depths greater than three hun- dred fathoms ; although between Fadiffolu and Miladummadulu a tongue of deeper water projects to the westward a considerable distance towards North Malosmadulu. South of Kolumadulu the conditions are different ; Haddummati, Suva- diva, and Addu occupy small plateaus separated by considerable depths and distinct from the larger one to the north. Haddummati, although less than twenty miles south of Kolumadulu, is divided from the northern plateau by a channel over eleven hundred fathoms in depth (PI. 1). Suvadiva rises from depths of over eleven hundred fathoms in the One and half degree Channel, fifty-five miles in width, which separates it from Haddummati ; and finally the small plateau of Addu, forty-five miles south of Suvadiva, rises from depths of nearly thirteen hundred fathoms, indi- cated in the central part of the Equatorial Channel. TOCOGRAPHY OF THE INDIAN OCEAN. TOPOGRAPHY OF THE INDIAN OCEAN WEST OF MALABAR. Plates 7, 8. The Admiralty Chart (PI. 7) contains a sufficient number of soundings to throw very important light on the depths of the Indian Ocean between the Malabar coast and the region south, and the Laccadives and Maldives. The one-hundred-fathom line runs nearly parallel to the coast at a distance vary- ing from twenty-five to sixty miles, and again parallel to this from twenty- five to thirty miles further west runs the one-thousand-fathom line. The Laccadives (PI. 8) rise from a plateau bounded by the one-thousand- fathom line, separated on the eastern side from the continental shelf by a channel of from ten to forty miles in width with a depth between ten and twelve hundred fathoms, the width of the channel and its depth gradually increasing as we go south ; so that between Kalpeni and Kiltan the distance between the eastern edge of the Laccadives and the one-thousand-fathom continental line is fully seventy-five miles, and the depth separating it from the Indian coast more than thirteen hundred fathoms. Within the one-thousand-fathom line surrounding the Laccadives plateau, rise quite abruptly the smaller banks and plateaus upon which are the coral reefs of the group. The five-hund red-fathom line of the plateaus and banks are within moderate distances of the outer edge of the reef Hats. This is indicated by the charts for Cora Divh, Sesostris Bank, and Bassas de Pedro, as well as for Cherbaniani Reef, Byramgore, Clietlat, Bitra, Peremul Par, Kiltan, Cardamum, Piti Bank, and Agatti. The same is the case for the more southern Laccadives, Elicalpeni Bank, Androth, Karawatti, Suheli Par, and Kalpeni, which rise from the deeper parts of the channel, over one thou- sand fathoms, separating the southern Laccadives from the Malabar conti- nental shelf. It is interesting to note in the Laccadives the existence of extensive banks similar in outline to some of the Maldive Banks or Plateaus, such as Cora Divh, Sesostris Bank, the Bassas de Pedro, and Piti Bank with 6 THE CORAL REEFS OF THE MALDIVES. its northern extension, Cardamom Bank, which equal in depth and nearly equal in size the largest plateaus of the Maldivian groups, for Bassas de Pedro is seventy miles in length, and the Piti Bank about forty. Three of the larger banks have no islands or lagoon reefs, the others culminate in coral reefs, having the structure of atolls ; some of the small banks are no larger than many lagoon reefs or faros in the Maldives. The soundings (PL 7) also indicate a narrow swale from fifteen to twenty miles wide, limited by the one-thousand-fathom line, as uniting Ihavandiffulu in the northern Maldives with Minikoi. A similar ridge has been sketched out by the Valdivia, 1 south of Addu, towards Diego Garcia, a distance of nearly three hundred miles; but Addu and Diego Garcia are separated by a much deeper channel, of twenty-five hundred fathoms in depth (PL 7), fully as great a depth as the average of the Indian Ocean to the east and west of the supposed Addu ridge. A long ridge bounded by the one-thousand-fathom line seems to unite Kalpeni with Suheli Par with a somewhat deeper channel between it and the Laccadives to the north, of a greatest depth of over twelve hundred fathoms; while a much broader and deeper channel separates this rise from Minikoi and the small banks immediately to the north with a greatest depth of nearly thirteen hundred fathoms. The fifteen-hundred-fathom line of the great tongue of the ocean (PL 7), between Ceylon and the Maldives, extends north as far as the ninth degree of north latitude, to the east of Minikoi and the Laccadives about halfway between them. On the west of the Laccadives and Maldives the fifteen- hundred-fathom line runs close to the western shore of the latter off Ari ; it is not more than eight miles from the one-hundred-fathom line to the west of the Laccadives ; it runs in a northerly direction thirty miles west of Agatti to west of Cora Divh. 1 Wissenshaftliehe Ergebnisse d. Deutschen Tiefsee Expedition auf d. Dampfer " Valdivia." Erster Band, Oceanographie ... v. Dr. G. Schott, 1902, p. 117., Taf. HE. The soundings on which the existence of this ridge are based are so widely separated that it is not possible to determine if the deep soundings of twenty-five hundred fathoms to the west of the line run by the " Valdivia " do not extend across it as a deep valley fully as deep as the oceanic depths, both to the east and west. It seems somewhat far-fetched to speak, as does Dr. Schott, of the Carolines, Marshall, and Gilbert Islands as connected by a plateau with a depth of four thousand metres ; and to speak of coral island districts as generally connected by such plateaus. TOPOGRAPHY OF THE INDIAN OCEAN. 7 The plateau upon which the Laccadives rise is nearly two hundred miles in latitude and about ninety in longitude. The two-thousand-fathom line runs nearly parallel with and at a short distance from the fifteen-hundred- fathom line, only it does not follow the tongue of the latter to the east of the Maldives. 1 There are other isolated banks to the north of the Laccadives off the con- tinental slope (PI. 7). A bank of about twenty miles in length, of a least depth of one hundred and seventy-eight fathoms, lies in latitude 14° 30' north, longitude 73° east, it is separated from the one-hundred-fathom con- tinental line by a depth of more than five hundred fathoms. Farther to the north in latitude 16 30' north, longitude 73° east, is Angria Bank, twenty- five miles in length and twelve in breadth ; it is separated from the one- hundred-fathom line by a channel, with depths varying from one hundred and seventy to over two hundred fathoms; see Admiralty Chart 2737. On the same chart are plans of Chetlat and of Kiltan Islands. They are small atolls, each with an island occupying the eastern reef flat, and a narrow reef flat on the western face of a shallow lagoon with a couple of boat passages leading through the western reef. These islands are steep to on the eastern face, but on the western face shallow soundings extend towards the one-hundred-fathom line about a quarter of a mile. Minikoi (A. C. 2738) 2 is usually considered as one of the Laccadives, yet it is only sixty-nine miles from the northernmost of the Maldives, while it is one hundred and eight from the southernmost of the Laccadives. This is quite an artificial subdivision, and it seems to me that Minikoi and the Laccadives in the north stand to the Maldives much in the same rela- tion which Addu and Suvadiva do in the south. The Laccadives are independent plateaus within the one-thousand-fathom line. Kalpeni, the southern bank, is separated from Minikoi by a channel with a depth of twelve hundred and forty-seven fathoms, while Minikoi is 1 The " Valdivia " took some soundings between Colombo and Suvadiva, the greatest depth obtained being 4754 metres. Schott, loc. cif., p. 96. 2 Mr. Gardiner * has taken Minikoi as a typical atoll to which he refers the structure of the Mal- dives ; it appears from the chart and from Mr. Gardiner's description as quite similar to some of the characteristic atolls of the Pacific. Not only Minikoi but the Laccadives as well as the southernmost Maldives do not possess the characteristics of the composite Maldive atolls ; on the contrary, they greatly resemble characteristic Pacific atolls. * Loc. cit., p. 27. 8 THE CORAL REEFS OF THE MALDIVES. separated from the northern Maldives by a channel with not over ten hun- dred and seventy-two fathoms, — depths not greater than those between Haddummati and Suvadiva, or Suvadiva and Addu. From what has been stated of the relation between Minikoi and the northern Maldives, it belongs geographically rather to them than to the Laccadives. The atoll of Minikoi is somewhat triangular, about five miles long, of a greatest width of about three miles. The eastern face of the atoll is flanked by a long narrow island ; there is no other land on the outer reef flat except a small island (Small-pox Island) on the southern part of its west face. At the northern apex of the atoll, a passage with from ten to twelve feet of water leads into a lagoon with a greatest depth of seven fathoms. The western and southern sides of the lagoon are flanked by a wide shallow reef flat full of coral heads in the southern part, with a few small shallow banks in the central part of the atoll. Minikoi rises quite abruptly from a depth of about one thousand fathoms. The northern Maldives are within the limits of the* northeast and south- west monsoons ; at the Equatorial Channel, however, the prevailing winds from May to December are westerly and southerly winds ; the northeast monsoon reaches the Equator and even beyond it during January, Febru- ary, and March. Addu is in the region of the variable winds beyond the influence of the monsoons. The southwest monsoon blows with greater strength than the northeast, and it is interesting to trace the modifications brought about by these op- posite winds on the faces of the islands exposed to their action. One can trace on the west coasts the action of the southwest monsoon with its eastern lee, while during the northeast monsoons the eastern face becomes the windward face. The position of an island, lagoon reef, faro, or ring in the interior of a group also greatly modifies the action of either of the mon- soons upon its shores, according to the range they may have, due to the trend of the island or of the group. Thus a great number of combinations are formed in position and in time, all having more or less influence upon a given area. They give us generally simple causes to account for the wast- ing of an island or for its increase, for the existence of coral sand or of shingle beaches, or for the presence of reaches or of belts of boulders, for TOPOGRAPHY OF THE INDIAN OCEAN. 9 the formation of bays by the shutting off of parts of the reef flats either on the windward or on the leeward side of an atoll, for the existence of horses of coral reef rock across certain beaches, and for the formation of coral reef rock patches, pitted and weather-worn, rising above the general level of the surrounding reef flats, for the formation of sinks, for the throwing up of dams, or even for the formation of a gap in the rim of an atoll. To climatic conditions, as affected by the position of islands and groups of islands, we also owe the degree of development of the flora of certain sections, or even its final destruction ; in others these forces begin to act only when the lagoon reefs have risen to sufficient height for the outer rim to come under the influence of the action of the sea upon its submerged reef flat rims. The soundings taken by the " Amra " do not bear out the statement of Mr. Gardiner that all the various banks of the Maldives (PI. 7) and Lacca- dives (Pis. 7, 8) appear to have been built on the same set of foundations. 1 Several of the Laccadive banks are separated by depths of more than one thousand fathoms, and the banks from Cora Divh to Cardamum are enclosed by the one-thousand-fathom line. Minikoi is similarly isolated, and the plateau of the fifteen-hundred-fathom line alone can be said to form the foundation of the Laccadives and Maldives (Pis. 7, 8). At any rate, the main chain of the Maldives does not lie alone upon a relatively shallow plateau at a depth of about two hundred fathoms.'- Goi- furfehendu is separated from South Malosmadulu by a channel of more than three hundred fathoms, South Malosmadulu from Fadiffolu by over five hundred fathoms, North Malosmadulu from Miladummadulu by about the same depth, Makunudu from Tiladummati by a channel of nearly eight hundred fathoms, Fadiffolu from Karidu by three hundred and seventy-two fathoms, and from Miladummadulu by seven hundred and sixty-nine fath- oms, — depths greatly in excess of those separating the banks of the central Maldives from North Male to South Nilandu and Mulaku (PI. 1). To the south, Mulaku is separated from the eastern face of Kolumadulu by a depth of nearly six hundred and fifty fathoms, and Kolumadulu from Haddum- mati by over eleven hundred fathoms, — depths nearly as great as those separating the banks of the southern Maldives (PI. 1). 1 Loc. cit., p. 172. ' Loc. at., pp. 150, 172. 10 THE CORAL REEFS OF THE MALDIVES. Gardiner has well shown that the topography of the central Maldivian basin almost precludes the idea that it was formed by the subsidence of a large central island. 1 But it seems to me he is hardly correct in stating that the shallower water lies in the centre of the plateau and that the depth gradually increases in the passages. 2 The deepest part of the central basin between South Male and Ari is in the centre ; but as I have stated, there is a more marked dip of the central plateau to the eastward than to the west- ward, the eastern passages being deeper than the western ones. That the precipitous slopes of the banks feeing the central basin indicate that they have been built up of corals or other organisms from a deeper plateau from a depth of over one hundred fathoms does not seem to me probable. 3 I should on the contrary consider each of these banks as an inde- pendent one as much as Makunudu, Goifurfehendu, Fadiffolu, Haddummati, and the southern banks. The existence of such banks as North Malos- inadulu and Kolumadulu, — the former rising from a depth of about two hundred fathoms at its southern, of nearly five hundred at its northern extremity, and from over twelve hundred fathoms on the western side ; the latter rising on the northwest from a depth of over two hundred and fifty fathoms, on the northeast from over six hundred, and on the southeast from over eleven hundred fathoms, — can only be explained on the supposition that they are independent banks. The existence of a deeper central valley between North Malosmadulu and Miladummadulu, off Fadiffolu and Koloma- dulu, developed by the soundings of the " Amra," shows a far greater irregu- larity in the depth of the plateau than is assumed by Mr. Gardiner, 4 and naturally modifies to a considerable extent the conclusions he draws from the existence of such a great level central plateau. 5 According to Darwin 6 there has been a disseverment of large atolls into smaller ones in the Maldives, but that " this process can be observed only in the northern half of the group, where the atolls have imperfect margins con- sisting of detached basin-formed reefs." He imagines the strong currents 1 Loc. cit., p. 172. 2 Ibid. 3 Darwin * calls attention to the steepness of the slope off the Maldives in spite of the sandy nature of the bottom of the slopes, slopes which exceed fifty-five degrees. 4 Loc. cit., p. 174. 6 Loc. cit., p. 175. 6 Coral Reefs, p. 143. * Coral Reefs, p. 31. TOPOGRAPHY OF THE INDIAN OCEAN. 11 which flow across these atolls to have broken "through the many wide breaches in their margin," and to have removed the finer sediment and deepened the channels. Such channels as divide Middle Malosmadulu from North and South Malosmadulu and the separate parts would then form distinct atolls like those of Ari and Ross, or the two Nilandus. " Further subsidence would render these channels unfathomable, and the dissevered portions would then resemble Felidu, Mulaku, and Horsburgh, which are related only in proximity." From what the soundings made by the " Amra " teach, the ingenious suggestions of Darwin regarding the formation of the Maldives are not borne out. Nor does it seem possible, as he suggests, that the Maldive Archipelago originally existed as a barrier reef of nearly the same dimensions as that of New Caledonia. The central basin which separates the eastern and western Maldives has no analogue in New Caledonia, and even if the whole island were to subside, the section across the centre would in no way resemble that of the great Maldivian basin with its fringe of independent atoll-like groups. 12 THE CORAL KEEFS OF THE MALDIVES. ANALYSIS OF THE SOUNDINGS TAKEN BY THE "AMRA." Plates 1-8 c. A sounding (No. 76) 1 taken in the centre of Gallandu Channel between Tiladummati and Ihavandiffuln (PI. 8 a, fig. 3) gave a depth of two hundred and fifty-one fathoms with a bottom of coarse coral sand. On the Admiralty Chart 2 are a number of soundings off IhavandifEulu, running into ten hundred and seventy-five fathoms, at a distance of eleven miles to the westward, normal to the trend of the west face (PI. 8 a, fig. 1). The outer edge of the slope must be quite flat, as nine hundred and ninety- two fathoms was obtained five miles from the face of the atoll. The sound- ings within a distance of five miles off the northern, the northeastern, and southeastern faces indicate a slope similar to that off the western face. The sounding of seven hundred and eighty-one fathoms, seven miles to the east of Kelai, even indicates a somewhat flatter slope (PI. 8 a, fig. 2), off the northeast horn of Tiladummati. From this a tongue of deep water runs westward towards the ridge connecting Tiladummati with Ihavandiffulu (PI. 2). The other lines we sounded to ascertain the slope of the oceanic faces of the Maldives are off North Male (PI. 8 c, fig. 24), off the west face of North Malosmadulu (PI. 8 c, fig. 26), off the west face of Ari (PI. 8 c, fig. 25), off the east face of South Male (PI. 8 c, fig. 27), and off Addu (PI. 8 c, fig. 21), in addition to the lines between the southern Maldives (PI. 8 c, figs. 22, 23). Line across the channel between Miladummadulu and the northern ex- tremity of Makunudu (PI. 8 a, fig. 7). Makunudu is separated from Miladummadulu by a channel with an oceanic bottom of a greatest depth of nearly eight hundred fathoms; it rises from a small secondary plateau on the western slope of the northern 1 The numbers refer to the List of Soundings, p. 2!). - Admiralty Chart, 06 a. ANALYSIS OF THE SOUNDINGS TAKEN BY THE "AMRA." 13 part of the Maldives. One mile west of the northern extremity of the faro, southwest of Goadu Island, we sounded in five hundred and four fathoms with a bottom of fine greenish sand with Globigerinae (No. 73). Four miles from the same starting-point in the centre of the channel, we obtained seven hundred and ninety-two fathoms (No. 74) with a bottom of broken shells. One mile east from Makunudu we sounded in five hundred and sixty-six fathoms with a bottom of coarse coral fragments (No. 75). Its western face must be steep and cannot be far from the fifteen-hundred-fathom line. The soundings between Miladummadulu and North Malosmadulu (PI. 8 a, fig. 5) seem to indicate that a tongue of deep water extends northward from the centre of the channel between Fadiffolu and Miladummadulu (PI. 3) towards the northern part of the central basin, and also southward towards the western point of Fadiffolu. We ran three lines of soundings across the northern part of the central basin of the Maldives. A northern line across its narrowest part between Mavila (Miladumma- dulu) and Furidu (North Malosmadulu) (PI. 8 a, fig. 5). One mile west of Mavila, we sounded in four hundred and twelve fathoms (No. 70) with a bottom consisting of green coral sand mixed with broken shells and Pteropods. Four and a half miles from the same point in the centre of the channel, we obtained five hundred and thirteen fathoms (No. 71) with a bottom of fine coral sand and Globigerina ooze. One mile east of Furidu, we obtained two hundred and eighty-six fathoms (No. 72) with fine coral sand and Globigerinae. The soundings we took off the west coast of North Malosmadulu (PL 8 c, fig. 26) indicate a less steep slope than further south, off the western part of the plateau. We obtained off Ari (PL 8 c, fig. 25), at a distance of seven miles, a depth of nearly fifteen hundred fathoms (No. 9) ; while off the coast of North Malosmadulu we sounded in twelve hundred and forty-seven fathoms (No. 65), at a dis- tance of eight miles. A second line across the northern basin ran from Kanifuri (Fadiffolu) to Mabaru (South Malosmadulu) (PL 8c, fig. 28). 14 THE CORAL REEFS OF THE MALDIVES. One and a half miles west of Kanifuri we found five hundred and nine- teen fathoms (No. 57), with a bottom of sticky yellow coral ooze, with but few Globigerinae and little sand and Pteropod shells. Five and a quarter miles from the starting-point in the centre of the channel, we sounded in three hundred and sixty-six fathoms (No. 58), with a bottom of green, fine coral ooze, with few Globigerinae. A mile and a half from Mabaru Island the depth had decreased to two hundred and thirty-six fathoms (No. 59), with a bottom of broken shells and coral sand. The plateau evidently dips rapidly to the eastward from the centre of the channel. Our third line across the northern central basin ran to the west of Gaha Faro towards Goifurfehendu; the soundings of the other extremity of the line from Goifnrfehendu east were taken while running to Goifurfehendu after leaving South Malosmadulu (PI. 8 a, fig. 8). One and a half miles to the west of the western entrance to Gaha Faro, we sounded in one hundred and forty-nine fathoms (No. 49), with a bottom consisting of coral sand and masses of large Globigerinae. 1 Nine miles from our starting-point, we obtained two hundred and fifty-eight fathoms (No. 50). In continuation of this line we obtained, seven miles east of Goifurfehendu, a depth of two hundred and thirty-seven fathoms (No. 60), with a bottom of fine green sand and masses of Globigerinae or Globigerinae sand. One and a half miles east of Goifurfehendu, we found two hundred and twelve fathoms (No. 61), with fine green sand and Globigerinae sand, showing a comparatively shallow plateau extending to the west of Gaha Faro towards Goifurfehendu, though it deepens somewhat to the west of Goifurfehendu, as is shown by the line from Goifurfehendu to South Malosmadulu, which has a greatest depth of three hundred and two fathoms in the centre of the channel separating them. The plateau again deepens to the northward, as is shown by the depth obtained between Gaha Faro and Fadiffolu, which varies between two hundred and forty-one and three hundred and seventy- two fathoms (Nos. 51, 56). 1 Iu the channels open to the sea, Globigerinae often occur in such quantities that we might with advantage call that kind of bottom Globigerina sand. ANALYSIS OF THE SOUNDINGS TAKEN BY THE "AJ1RA." 15 Line off the west coast of the southern part of North Malosmadulu (PL 8 c, fig. 26). Eight miles off the coast in a northwesterly direction from Pemfuri Faro we sounded in twelve hundred and forty-seven fathoms (No. 65). One mile off Maregiri, we obtained three hundred and seventy-eight fathoms (No. 66), and already found a bottom consisting of green coral sand mixed with Pteropod shells and Globigerinae. Pteropod shells are common in shallow soundings near the atolls; we collected them in number, as well as Globigerinae, in nearly all our surface tows made off the outer faces of the archipelago. We did not sound any of the Malosmadulu Channels (PL 8 b ; figs. 9, 10), the soundings on the Admiralty Charts being sufficiently numerous to define their depth. According to the Admiralty Chart (No. 66 a), South Malosmadulu is separated from Middle Malosmadulu by a channel from two to three miles wide, with soundings in the axis of the channel ranging from one hundred and ten to one hundred and forty fathoms. Middle Malosmadulu is sepa- rated from North Malosmadulu by Moresby Channel, somewhat wider than the channel to the south of it, about three miles in width, with soundings in the centre of the channel ranging from one hundred and twenty to one hundred and thirty-five fathoms. Line across the channel between Furudu (Goifurfehendu) and Madu- wari (South Malosmadulu) (PL Sb, fig. 10). One mile north of our Goifurfehendu base-line we obtained two hundred and forty-eight fathoms (No. 62), with a bottom of coarse coral sand and broken shells. In the centre of the channel, three and one-half miles from the starting-point, we sounded in three hundred and two fathoms (No. 63) ; bottom of coral sand and broken shells. One mile from Maduwari we sounded in one hundred and ninety-three fathoms (No. 64), bringing up in the claspers a few fragments of broken corals. The bottom of this channel is apparently covered with coarse coral sand, broken shells, and corals. 16 THE CORAL REEFS OP THE MALDIVES. Line across the channel between Fadiffolu and Miladnmmadulu (PI. 8 a, fig. 4). From Hurawali (Fadiffolu) to Karema (Miladummadulu). One mile north of Hurawali we found three hundred and forty-two fathoms (No. 67), hard bottom. Four miles north, in the centre of the channel, we obtained seven hundred and sixty-nine fathoms (No. 68). One mile south of Karema we sounded in three hundred and forty fathoms (No. 69). Unfortunately, no sample of the bottom of this channel was obtained. No. 67 indicated a hard current-swept bottom. Line across Kardiva Channel, from Gaha Faro to Aligau (Fadiffolu) (PI. 8b, fig. 11). One and a half miles north of Gaha Faro we sounded in two hundred and forty-one fathoms (No. 51), and brought up a fragment of Oculina. Six miles north of Gaha Faro, halfway to Karidu Island, we obtained two hundred and fifty-eight fathoms (No. 52), with a hard bottom; and a mile and a half sonth of Karidu we got two hundred and ninety-eight fathoms (No. 53), bringing up coral sand and small fragments of Pteropod shells. One mile north of Karidu we found three hundred and twelve fathoms (No. 54), and brought up small pieces of pumice from a hard bottom. In the centre of the channel between Karidu and Fadiffolu, seven and a half miles north of Karidu Island, there seems to be a distinct ridge, as we only obtained one hundred and ninety-seven fathoms (No. 55), with a bot- tom of clean coarse Globigerinae sand. One and one-half miles south of Aligau (Fadiffolu), we sounded in three hundred and seventy-two fathoms (No. 56), with a bottom consisting of coarse coral sand and Globigerinae sand. The edge of the Maldivian plateau drops gradually to the north of Gaha Faro from two hundred and forty-one to three hundred and seventy- two (Nos. 51, 56) near Fadiffolu. The northern part of the central basin is very open to the sea (PI. 1); between Gaha Faro and Fadiffolu it slopes to the east, only a few higher knolls, like Karidu, and the sounding of one hundred and ninety-seven fathoms (No. 55) to the north, indicating its boundary. ANALYSIS OF THE SOUNDINGS TAKEN BY THE "AMKA." 17 Line from North Male to Gaha Faro (PI. 8 a, fig. 6). In the centre of the channel, between Gaha Faro and North Male, we obtained a depth of one hundred fathoms (No. 48). The claspers only brought up a fragment of pumice, covered with Bryozoa, Algae, and Nullipores. Mr. Gardiner states that pumice was found on all beaches of the Mal- dives, and that it was not known before the eruption of Krakatoa in 1883. But pumice must have found its way on the beaches of the Maldives earlier, as it occurs inland on many of the islands. The bottom of all the channels separating the composite atolls of the central Maldives appears to be quite flat, the soundings off shore drop rapidly ; generally at a distance of a mile and a half from either face they attain a depth but little inferior to the greatest depths in the centre of the channels. Line of soundings west of North Male (PI. 8 c, fig. 24). One mile due west, off the normal to the pass west of Hembadu, in the centre of the western face of North Male, we obtained one hundred and twenty fathoms, with coral sand (PI. 1). At a distance of three and one-half miles the depth only increased to one hundred and fifty-nine fathoms with a bottom of sticky coral sand, and at seven and one-half miles we ran only into one hundred and eighty-six fathoms, with a bottom of fine green sand filled with Globigerinae. Judging from this line of soundings and from the sounding's existing ' between the northern part of Ari and Rasdu and Toddu, as well as to the westward of these atolls, these atolls are on the western edge of a shallow plateau extending west of North Male and north to Goifurfehendu. This plateau separates the northern part of the central basin from the southern ; it drops off very rapidly to the westward of Ari, if we may judge from the line of soundings we ran to the west of the pass south of Mandu (Ari) (PL 8 7 4 O ' 73 34 1460 Fine green ooze, glo- bigeriuae. 36 miles E. of Beramundu (Ihavan- diffulu). 78 (( ft 7 4 74 27 1547 Fine green globi- geriua ooze. 91 miles E. of Beramundu (Ihavan- diffulu). 79 tt tt 7 7 75 46 964 Fine green globi- gerina ooze. 181 miles E. of Beramundu (Ihavan- diffulu). DESCRIPTION OF THE ATOLLS. North Male. Plates 1, 3, 4; S a, Jiff. 6; 8 b, Jiffs. 11, 19; 8 c, fig. 24; 9-18 ; 19, Jig. 1. North Male is irregularly triangular in form (PL 4), with a western convex side ; the northwestern and northeastern faces are concave ; it is about thirty-two miles long with a greatest width of twenty-two miles. Over fifty islands, rings, and banks are enclosed within the basin of North Male ; it is flanked both on the east and west faces by large faros, only on a few of those of the west faces are found islets. The principal islands are on the southeast face from Mirufuri to Wilingili Island to the west of Male Island. There is no group in the Maldives in which there is such a large number of rings (faros) cropping up in all directions to the surface or within three or four fathoms from the surface. These submarine rings or embryo sub- merged atolls are specially numerous immediately towards the north of Male Island and in the northern part of the group to the southwest of Mirufuri (PI. 4). A great number of small banks are dotted over the central area of North Male, either reaching the surface or awash or at depths of three to four fathoms. These banks as well as the rings generally rise from a depth of from twenty to thirty fathoms. There are comparatively few islands scattered through the central area of North Male. Mr. Gardiner has noticed marked discrepancies between Moresby's Charts of North Male and the positions he assigned to several of the reefs. This may be due to the large areas of that atoll left unsurveyed within Moresby's lines of sounding and the difficulty of locating from a low level rings and low islands which become clearly outlined from a higher station like the bridge of a steamer. Wherever we followed lines of soundings we usually found the chart wonderfully accurate, but naturally in the unsurveyed portions we were frequently at sea. 36 THE CORAL REEFS OF THE MALDIVES. The rings of the inner parts of North Male and some of the islands on the outer rims are all built of sand, as has also been observed by Mr. Gardiner. But on many of the outer islands remnants of elevated reef rock are found forming the base behind which the sandy growths of the islands have been piled up since the slight elevation of the Maldives took place. Mr. Gardiner considers North and South Male to be in a very stable con- dition. 1 On North and South Male as well as Mulaku there is comparatively little land on the east side and still less on the west side of these atolls (Pis. 3, 4, 5). The character of the faros which follow in rapid succession on the south- east face of North Male varies but slightly. They differ in size, in the width of the reef flats, in the size and depth of the lagoons, in the number and dimensions of the islands on the faro rims, and in their position with reference to the sea or inner face (PI. 4). The east point of North Male is formed by the northern part of the wide outer flats of an elongated faro about seven miles in length and two at its greatest width. The inner rim of the faro is quite narrow ; at its southern extremity are found three small islands and a small secondary lagoon enclosed in a sharp spit forming the north side of the deep pass, with from twenty to twenty-seven fathoms, separating this faro from the one immediately to the south of it. Mirufuri and Difuri are two small islands on the northern parts of the east face of this eastern faro. The soundings in the lagoon of the faro vary from four to seven fathoms (PI. 4). Tulusdu, the island on the northeastern face of the next faro, is close to the pass ; the deepest part of the lagoon of the faro is near the northern extremity. The rim of the faro is nearly of uniform width on the two faces except at the northern and the southern points of the faro where the flat is flanked by a wooded island (PI. 4). The outer edge of the faros on the southeast face of North Male are all flanked by lines of coral boulders or small horse heads of modern reef rock. The lagoon reef to the south of Tulusdu faro is marked by the small size of its lagoon and the central position of the principal island of the faro flats. Imma, the larger faro to the south, is a wide reef flat covered with from one 1 hoc. cit., pp. 402, 403. NORTH MALE. 37 to six feet of water, with two large islands in the axis of the flat. The lagoon of this faro has practically been filled up by the washing into the central basin of coral sand from both the sea face and the lagoon side during the prevalence of the northeast or southwest monsoons. A small faro with an islet and a sand-bar at its eastern extremity occupies the central part of the deep pass separating it from the faro to the north. We find on the chart twenty-one fathoms in the channel on either side of these obstructions (PI. 4). The great faro at the horn of the southeast face of North Male is sepa- rated from Irama by a pass nearly three miles wide. It is well seen from off the northern extremity of Hulule, where one can take in at a glance the whole length of the lagoon and look over the flats to the east of Hulule and to the great reef flats forming the northern horn of the faro. The northern and southern parts of this southernmost faro (Pis. 10, 11) are occupied by wide reef flats; a small island, lying diagonally across the flats, occupies the central part of the northern horn of Hulule, a long island flanks the western face of the southern horn, and a small islet lies to the southeast near the southernmost extremity of the faro. Near the small southern islet a line of recent reef rock rises inside of the boulder belt, from a foot to eighteen inches above the reef flat. It is one of the outliers of the older reef flat conglomerate which has been denuded to its present level. This ridge is eroded, deeply pitted, undercut, and honeycombed, and is the first indication we saw of evidence of a slight elevation of the coral reef of the Maldives in most recent times, judging from the character of the reef rock as it appears there and at other points in the group. The outer eastern edge of Hulule faro is flanked by a boulder belt of fairly large blocks. The belt passes round the southeast horn and extends along the face of the little islet on the edge of the faro near its southeast horn. This faro is in reality an atoll, fully as large as many a Pacific island atoll with wide reef flats, flanked on the sea face by coral boulders enclosing a shallow lagoon and with islands on the rim flats. The southern part of Imma pass is occupied by a small faro and an islet to the north of it. The small faro is separated from Furena by a narrow pass with thirty fathoms of water. In the main channel the depths vary from 38 THE CORAL REEFS OF THE MALDIVES. twenty-four to thirty-five fathoms. Another faro in the pass to the north of Imma (PI. 15, fig. 2) is crescentic. One side is formed by an island extending into a sand-bar, while on the other the island extends into a submerged rim ; the horns of the crescent are widely separated, enclosing a deep lagoon. The southern part of Tulusdu faro as seen from the west appears like a narrow, shallow lagoon, enclosed between two strips of shallow flats. We steamed round a large faro with a wide reef flat rim nearly a mile and a half in diameter, about four miles to the west of Difuri, and with nine fathoms in the lagoon ; a wide flat makes off the eastern part of the rim of this faro. The great Difuri faro has even more than Tulusdu and Hulule the characteristics of an atoll ; as seen from the west, the lagoon studded with coral heads is seen to be separated from the inner waters of North Male by a narrow rim flat extending from the northern horn to the southern extremity. The eastern reef flat rim is in places nearly a mile wide. To the north of this lagoon reef in the centre of the pass (PI. 4) is situated a long comma-shaped faro nearly three miles in length and perhaps half a mile wide at the club. The faros of North Male are most irregular in shape, some are crescentic, others more or less rectangular, others somewhat polygonal, or dumb-bell- shaped, or circular with long spits, or they have assumed the shape of a long series of loops or cusps forming angular spits. An examination of the chart of North Male (PI. 4) will show their great diversity in shape and size. Yet it is a simple matter to reduce these protean shapes to a simple and common origin. The circular faros occur in the sheltered waters of North Male, where corals are left to grow with but little interference from the monsoons or currents which have tended to modify their outlines on the outer rim parts of the Maldive plateau, where they are exposed to the full force of the prevailing winds. To the west of the comma-shaped faro we came upon a maze of banks and small faros; that part of North Male not being as yet surveyed, we did not dare to push north, but shaped our course for the west face of North Male. We found the western slope of Difuri faro much less abrupt than that of other faros we had examined. About quarter of a mile from the NORTH MALE. 39 edge we were in nine to ten fathoms, the slope running very gradually up to the edge of the flat into from four to six feet of water at three quarters tide. There was a fine growth of corals on the western face in marked contrast to the condition of corals usually found on the flats and slopes within the lagoons of the Pacific atolls. Nothing shows, perhaps, as plainly as this the atoll character of these numerous faros, scattered over a great plateau. In the lagoon of the faro, coral heads occur occasionally in small patches as well as on the rim flat, where they are separated by wide lanes of sand. In the lagoon of the faros the corals are retarded in their development by the same cause, which acts so efficiently in the lagoons of atolls of other coral regions. On the outer rim flat there are fewer patches of living corals than on the western rim. The outer sea edge of the reef flat is edged with large coral blocks and boulders, more or less pitted and honeycombed ; these in some cases formed ridges of modern reef rock near the outer edo-e of the boulder belt, more or less undercut, rising a couple of feet at most above the level of the outer edge of the rim, and indicating perhaps an elevation of a few feet, these ridges are outliers of a former higher plane of the outer rim, which has been cut down to its present level. The depths of the rims of the rings and of the sand banks varies greatly. They are found at all depths from fifteen fathoms to the surface ; where they have grown up to within six or seven fathoms of the surface, the color of the water over the rims becomes an excellent guide of their depth below it. It is difficult to form an accurate opinion regarding the junction or coalescence of adjoining faros on the outer faces of such an enclosed basin as the great lagoon of North Male. It would seem natural to consider some of the longer and narrower faros which form the outer face of the eastern horn of Felidu, parts of the eastern face of Mulaku, of the outer rim of Kolumadulu and of Haddummati, as having originally been composed of a large number of small faros or banks separated by comparatively shallow passes. With the growth of the corals from the two sides those passes become gradually choked, and may finally have become united to form the long narrow reef flats just mentioned, which with the exception of the existence of small lagoons now show so little trace of their former composite condition. 40 THE COEAL REEFS OF THE MALDIVES. However, when we speculate upon the origin of the larger faros of some of the northern groups, as Ari, North Male, Malosmadulu, and Tiladuramati (Pis. 2, 3, 4), the conditions now existing do not warrant the conclusion that they have been formed by the junction or coalescence of adjoining faros. The great size and depth of the lagoons of some of the faros, the depth of the passages separating them, their great width, — all these factors indicate that the faros have grown from depths of eighteen to twenty fathoms on secondary elevations of the greater plateau, with much the same shape they now have. In their upward growth they have been modified so as to become faros differing in the width of their rims, their outline, the depth of the lagoons, the number of banks and islets or islands thrown up on the rim flats ; or the lagoons may even have been completely filled, and the faro changed to a great reef flat or even an island steep to with a fringing reef, where all trace of the original faro has been lost. Of course we do not deny that faros may join in time if they were originally separated by comparatively shallow passes. This we see taking place by the growth of corals in cusp-like spits from the slope of a rim flat, which enclose shallow parts of the adjoining waters — a process somewhat similar to that by which adjoining islands on wide reef flats become connected into a larger island from the throwing out of sand-spits or of shingle-spits, either on the lagoon or the sea face of the outer rim of an atoll, forming great bays which become eventually filled ; a process which takes its greatest development, perhaps, in the Ellice, Gilbert, Marshall Islands, and also occurs in other groups of atolls in the Pacific. An examination of the charts (Pis. 2-6) shows a few instances where adjoining faros in the interior of Ari, North Male, and the groups just men- tioned may have coalesced. The distance separating the rings which crop up within the enclosed area of Male and Ari furnishes no evidence that they have been formed by the coalescence of adjoining faros. The irregularly shaped banks and faros are additional evidence of their having grown up on secondary inequalities of the plateaus of these groups. The western rim of one of the rings on the northwest face is circular; the other is digitate with deep indentations cutting into the wide irregular eastern flat (PI. 4). NORTH MALE. 41 On the outer line of faros of North Male, Malosmadulu, and Ari (Pis. 3, 4), with the exception of the reef flat bounding the southern face of the latter, and in the maze of coral heads and banks of its southeastern face and the reef flat on the northwest face of North Male there is no evidence of the coalescence of adjoining faros ; they stand out separately and distinctly, much as they originally arose from the plateaus upon which they are scattered. Steaming south from Mirufuri along the eastern face of North Male, we could see across the outer line of faros a number of well-wooded islands in the interior of the great enclosed sheet of water to the westward from Asdu and south to the islands to the northwest of Male Island and near it. Such a land rim as we have described from Mirufuri to Male gave to the so-called lagoon of North Male a totally different character from that of any other atoll I had previously seen. The existence of well-wooded islands in the interior of the great enclosed sheets of water of the Maldives we subse- quently found to be a distinctive feature of the archipelago. The outer edge of the faros on the east face of North Male is clearly marked at low tide by a belt of boulders or by the surf breaking on the outer face of the wide reef fiat; the depth of the rim below the surface can readily be detected from the color of the water on the flat. The inner edge, at the time of the northeast monsoon, is only marked by the coloring of the inner flat, a yellowish-green belt of varying width separating the light blue of the lagoon of the faro from the deeper, darker-colored water of the Larger enclosed sheet. Extensive patches of corals spread over both the inner and outer reef flats of the faro, gradually becoming thinner and more distant and less luxuriant as they recede from the outer edges. Their presence can be detected at a considerable distance from the dark, more or less purple brown coloring of the patches. The southern corner of North Male is occupied by the island of Male, on which is built the capital of the Maldives. The greater part of the island is covered by an old fort erected by the Portuguese and Dutch during their occupancy of the islands. To the west and south extends a wide reef flat dished near the outer edge, while on the northern and eastern faces the island is steep to. Extensive works have at different times been built 42 THE COKAL REEFS OF THE MALDIVES. Bastion at Male. either as breakwaters or bastions, which conceal the structure of the island, but it is identical with that of many similar islands throughout the group. Advantage has been taken of a narrow reef flat on the northwest face of the island to build a breakwater enclosing a small harbor with suffi- cient depth to shelter native boats of considerable size. Tbe southern face of Male Island is also flanked on the outer edge of the somewhat dished reef flat by a belt of angular coral blocks more or less eroded and weath- ered, forming a low wall. From our anchorage off Male we could see to the northwest two of the remarkable light green rings which are so characteristic a feature of the coral physiognomy of the Maldives (PI. 4). These rings and their endless modifications, due to their age, their size, their position on the primary or secondary plateaus of the group, their exposure to the prevailing winds, their depth from the surface, give us the key to the structure of the coral reefs of the Maldives. 1 The only structures to which they can at all be compared are the small atolls known as " boilers " in the Bermudas, and which flank the south shore near Sinkey Bay. 2 But the diminutive atolls of the Bermudas owe their origin to other causes than those which have built the rings of the Maldives. The latter are atolls, and no matter what their size may be, they are built up by corals, while in the Bermudas these small atolls are really only large pot-holes, the outer rim of which is protected by the incessant growth of Serpulaa and Alga3 on the edge. All the faros of which the nucleus con- sists of elevated reef rock planed down to the level of the sea have 1 Darwin * considers the ringlike structure to be contingent on breaches into the lagoon being wide and numerous, thus placing the inner side of the rings in the same condition with the outside of an ordinary atoll exposed to the sea. Hence the margins have grown vigorously outwards, while they have grown upwards during the subsidence to which, according to Darwin, the whole archipelago has been subjected, subsidence and upward growth converting the central space of each little reef into a small lagoon. 2 A. Agassiz, Bull. M. C. Z., Vol. XXVI., No. 2, Pis. 2-2-26. * Loc. cit., p. 141. NORTH MALE. 43 increased by sand being heaped up in the direction of the prevailing winds or currents, forming horns eventually uniting and enclosing a part of the greater inner lagoon. On the face of the banks, corals obtained a foothold and consolidated the rims of the faros. Their mode of formation can best be explained when tracing the changes undergone by the rings. In thirty fathoms at our anchorage off Male Island the bottom was covered with fragments of broken coral coated with Nullipores, or was quite hard, having been swept comparatively clean by the strong current rushing in and out of North Male through the southeast channel. At our anchorage off the west face of Difuri about three-quarters of a mile from the rim of the faro, we found coarse coral sand in twenty-one fathoms of water; a haul of the dredge made in about twenty-seven fathoms nearly in the centre of North Male brought up coral sand mixed with more or less sticky ooze and broken shells. In all cases we found quite a dif- ferent bottom from that which characterizes the bottom of a Pacific lagoon. The greatest depth of North Male is thh'ty-seven fathoms. The majority of the soundings indicated vary from twenty-five to thirt}'-one fathoms. As a type of an inner island, we examined a small well-wooded island (PI. 9, fig. 2) which rises in the middle of the southeast passage into North Male with nearly thirty fathoms on either side of it. The island is elliptical, and is placed on the northwestern horn of an elliptical flat which stretches out in a southeasterly direction. The flat slopes very gently to the sea, is edged on the outer rim by a sink forming a shallow ditch of varying width, flanked by coral boulders or masses of beach rock extending as an irregular wall along the greater part of the outer edge of the reef flat. The greatest width of the reef flat is from two hundred and fifty to three hundred feet. It diminishes gradually in width to the northwestern corner, where the island is steep to. Corals grow in great abundance at a depth of from five to three fathoms upon the steep slopes of the reef flat; they grow with less profusion to six or seven fathoms, where they are separated by wide lanes and patches of sand which eventually cover the whole bot- tom at a depth of from eight to nine fathoms. From the three-fathom line, they also diminish in number towards the surface and spread over the edge of the flat, which is partly bare at low water; they extend but 44 THE CORAL REEFS OF THE MALDIVES. a short way over it, the greater part of the flat being covered by dead corals overgrown with Nullipores. The sand beaches surrounding the island are steep, from six to seven feet in height. The central part of the island is lower than the top of the beaches which surround it, form- ing a shallow sink from twelve to eighteen inches or more in depth. Considerable moisture accumulates in this central sink, and in the rainy season a pool is probably formed of more or less brackish water. This structure is most characteristic of the islands of the Maldives, whether they occur in the interior of the great sheets of enclosed waters or on the outer rims of the plateaus. The sink has been formed by the washing up of the beaches round a central area, as we have seen it in the Paumotus and elsewhere in other Pacific atolls. Before the vegetation became too dense, beach sand was blown towards the interior and partly filled the central area, until this was prevented by the growth of bushes and shrubs, when the beaches merely increased in height and the sand of the upper ridge of the beach was driven sparingly towards the centre of the island, or its further passage stopped by the belt of denser vegetation which had come up on the higher parts of the coral sand beach. The island of Male has gradu- ally developed and been formed much in the same manner as this island. The corals growing on the slopes of this island are marked for their luxuriance ; they grow as abundantly as they do on the sea face of any atoll. This is in striking contrast to their scanty development in the interior of typical lagoons. It can readily be explained from the great depths of the passes and the great mass and purity of the water passing into the interior of the enclosed basin of North Male. Throughout the northern and central Maldives corals grow in great profusion on the slopes of the islands and faros within the enclosed basins of the group. The branching corals consist mainly of species of Madrepores, of Pocillopores, and Millepores. While the massive corals are usually Astreans, Porites, and the like, Maeandrinte are not common, The vegetation of this island is remarkably fine. In spite of its small size large trees are found upon it. It is a favorite resort for flying foxes ; its aspect is in striking contrast to the meagre flora found on similar islands in the Paumotus, Ellice, or Gilbert Islands. NORTH MALE. 45 The first ring we examined (PI. 13) was nearly circular, about half a mile in diameter. It rose quite abruptly from a depth of nearly twelve fathoms and sloped very gradually to the general level of the bottom of the greater lagoon (thirty fathoms); the outer edge of the rim being about one hundred and fifty feet from the vertical at the base of the rim slope. The width of the rim varied from sixty to one hundred and seventy-five feet, the general depth of the rim varying from two feet to two fathoms ; the slope of the rim towards the interior of the enclosed lagoon was even steeper than the outer slope. On the outer face corals grew in great profusion from the very base to the edge of the rim and in patches on its surface. The inner slope as well as the bottom of the lagoon consisted of coral sand driven across the rim and gradually widening the eastern or western rim according to the direction of the prevailing monsoon, the process at the same time filling the central lagoon, which in this case was about seven fathoms in depth. The corals were of the same genera as those growing on the island in the southeast pass into North Male, and fully as luxuriant. We did not observe any corals on the bottom of the small lagoon ; they must be killed by the inflowing sand. A few patches of corals extend from the outer base of the ring into somewhat deeper water, but they fast disappear and the slope consists again of fragments of coral and coral sand. Towards the northeastern face of the rim flat the sea breaks, the depth of a part of the rim not being more than a few inches below the surface. The sand of that part of the rim is kept greatly agitated, and we should expect there the early formation of a sand-bar or diminutive islet on the rim flat of the ring. The next faro we examined represented a stage in the formation of an island considerably more advanced than that of the simple ring rim enclos- ing a comparatively deep lagoon. It consisted of an extensive elliptical flat (PI. 15, fig. 1) with but three or four feet of water at low tide, its rim flanked on the east by a small narrow island covered with bushes rising from a steep coral sand beach ; towards the west a small shallow lagoon, judging from the color of the water, occupies a part of the reef flat. It seems a simple process to follow the transition of the first ring we 46 THE CORAL REEFS OF THE MALDIVES. described to such a larger reef flat. Its original large lagoon has been filled up and left a diminutive pool ; the incipient sand-bar has become a constantly growing islet and is covered with scant vegetation derived from the adjoining faros. It needs but an additional stage in the filling of the lagoon and increase in the size of the islet to reach the condition of the island we described lying in the southeast pass to the northeast of Male Island. With the growth of the land larger trees will obtain a foothold, and the original lagoon become transformed into a great reef flat similar to that of the former island. This island rose quite abruptly from a depth of twenty-five fathoms much as did the island to the eastward of Male. It is evident from the mass of broken and dead corals towards the base of the outer slope that the rim flat increases in width by the reaching out of the outer talus. Corals obtain a foothold on the upper slope of the flat and form extensive patches reaching over its surface from the outer rim. The reef flat gradu- ally becomes covered with fragments and small masses of coral coated with Nullipores which bind the whole together. These rings are apparently formed by the upward growth of circular or elliptical patches of corals occupying slight elevations above the general level of the surrounding plateau. These patches may grow up uniformly, forming elliptical banks, or they may grow as annular structures, the corals of the outer face only rising towards the surface, while those in the central part of the patch are killed either from want of food or of clean sea water, or are choked by the sand derived from the dead corals of the outer rim, which is washed into the interior and covers the corals. The outer ring growing faster than the inner dish is filled, a central lagoon is formed, this is filled in time after the ring rim has reached the surface ; the annular structure is then changed into a bank from which all traces of the original lagoon have disappeared and upon the outer rim of which sand-bars or islets or both have formed. There is thus eventually formed either an island surrounded by a more or less extensive reef flat or an island and reef flat with the remnants of a lagoon, or an island or islet on the rim of a ring with a well-marked and deep lagoon, all different stages of growth of an annular coral reef formed upon a base laid within the range of depths NORTH MALE. 47 at which reef-building corals begin to grow. The final stage, that of a well-wooded island steep to, is well seen in the three small islands we examined lying to the northeast of the first ring we visited in North Male waters. These are all steep to ; the original wide reef flats are reduced to a narrow insignificant strip flanking the islands. About three miles to the northwest of Male Island rises an interesting faro with a small islet on the eastern rim of its lagoon, that has been reduced to a minimum, nothing but a small hole being left of the original lagoon, which has gradually been filled by coral sand driven into it over the sum- mit of the reef flat rim until there is only an insignificant pool left to indi- cate the original lagoon. Fine patches of corals extended from the outer edge over the reef flat; a belt of low vegetation has found a foothold on the islet of the faro which is protected by coral sand beaches. Vehamafuri, another large faro to the northwest of Male Island, deserves mention for its narrow rim ; its greatest diameter is fully a mile ; the enclosed lagoon has a greatest depth of twelve fathoms. It rises from a depth of twenty-seven fathoms. To the westward rises another much smaller circular faro, less than a quarter of a mile in diameter, the rim of which is fully three fathoms below the surface. On our way to the island of Tulagiri from Vehamafuri we passed a number of faros with light green rims and darker bluish-colored lagoons of different shades varying according to the width of the rim, its depth below the surface, and the depth of the lagoon. The island of Tulagiri (PI. 16, fig. 2) is on the southern rim of a small faro, not quite half a mile m diameter; its southern rim is remarkably nar- row and covered with a great many patches of corals. To the eastward of this faro could be seen the outline of a circular faro, nearly a mile in diameter. Two others somewhat elliptical rose to the northward. Nearly halfway between Vehamafuri and Tulagiri rises Bundusi and a small island to the south of it; both are steep to; the latter is nearly a mile in diameter with extremely narrow reef flats. Between Tulagiri and Kudahitty we passed six faros. With the excep- tion of the northern one, which is triangular in shape, they are circular or elliptical, with fairly wide rims and moderately deep lagoons. Kudahitty 48 THE CORAL EEEFS OF THE MALDIVES. (PI. 17, fig. 1) is a small island on the eastern edge of a small lagoon, less than a quarter of a mile in length. Hitty (PI. 17, fig. 2) is a small island a couple of miles east of the great faro on the central part of the western edge of North Male ; it is on the southern extremity of an elliptical ring enclosing a deep lagoon. Hitty is flanked by a hook-like faro, with a broad eastern rim and open to the north. To the north we passed a similar double faro on our way to Hembadu. A small island covered with trees occupies the greater part of the rim, which separates the two loops of the adjoining faros. On our way across the Male group, we passed a number of coral patches or sand-bars of irregular or circular shape, at varying depths below the surface, the color of the patch indicating fairly its depth below the surface. Many of the flats were covered here and there with irregular patches of corals, but as a rule they were not flourishing and not to be compared with the vigorous growth on their slopes. One of the rings we passed enclosed a well-marked blue lagoon indicat- ing at least seven to eight fathoms ; its rim extended to the westward in a long spit awash, on which sand-bars were forming. The spit was separated by a deep channel from a narrow sand bank covered with but a few feet of water. To the north rose small faros and banks, on the rims of which corals were growing in great profusion. One large circular faro was specially noted for its diminutive lagoon, the only indication left of a former larger lagoon which has gradually been filled up by the sand driven from both sides across the rim, and deposited on the bottom of the lagoon during the prevailing monsoons. To the south of Hembadu we could see three interesting stages in the de- velopment of a faro. In the immediate foreground rose a large, irregularly circular reef flat of a light green color, its rim edged with closely packed great violet brown patches of corals, extending down along the outer slopes of the faro into twelve or fifteen fathoms ; on the northern face a small, bare sand bank had been thrown up on the reef rim flat. To the west a shallow lagoon of considerable size occupied the greater part of that side of the faro. Immediately to the north of this rose a small elliptical bank, its surface covered with a magnificent growth of corals in six to seven feet NORTH MALE. 49 of water, and a small sand bank thrown up on the outer edge of the rim where the corals were awash. Still further, a large bank of the same char- acter, but with a sand bank rising a couple of feet, and on which a few o-rasses and shrubs had obtained a foothold ; while in the distancea similar bank had been blown to a sufficient size and height to be covered with small trees and be surrounded by a considerable belt of low bushes. Passing out of North Male through the wide passage west of Hembadu, we could not fail to compare it with the narrow passes which characterize the entrances into so many of the Pacific atolls. Hembadu Pass is fully four miles in width, with from seventeen to twenty-two fathoms in depth ; and flanked by low faros, it gave one the impression of still greater width and marked more sharply the contrast to such narrow passes as those of the Paumotus, Ellice, Gilbert, and more northern atolls. The passes into the central area of North Male are only low portions of the rim separating the more elevated parts of the plateau, upon which the faros have grown either on the outer face or in the inner area of North Male. An examination of the soundings within North Male as well as those of any other group of the central and northern Maldives shows how great are the irregularities of the bottom of the secondary plateaus of the Maldives. To the westward of Hembadu the bottom is very uneven; on our way out of North Male through the wide pass to the west of Hem- badu, we found rapid changes in depth, and could distinctly see the exten- sion of the coral patches indicated on the charts scattered over a good part of our track. They extended all the way across the northern part of the pass, west of Hembadu. Off the eastern face of the faros of the western face of North Male one passes rapidly from fifteen to thirty fathoms ; in the vicinity of an agglomeration of small faros the depth varies from twenty to thirty fathoms. The lines of soundings as they exist throughout North Male indicate a series of elevations rising somewhat abruptly at from twenty-four to twenty fathoms; they constitute the base upon which the banks, faros, or islands and islets which stud North Male have been built up. This indicates a condition of the bottom of the central area of North Male different from the great level and unbroken flats which characterize the bottom of the lagoons of the Pacific atolls. There the great fall in 50 THE CORAL REEFS OF THE MALDIVES. depth occurs close to the lagoon face of the outer rim and it slopes suddenly to the general dead level of the great central area of the lagoon. 1 To the south of Hembadu Pass the faros on the west face of North Male are very similar in structure, though greatly differing in shape and size. The largest one, four miles in length by one and a half wide, immediately to the south of the pass, is edged on the western rim by a wide reef flat, on the eastern by a narrow rim. A small island covered with low bushes rises in the southern part of the western rim of the reef flat. The reef flats enclose a large lagoon nearly as long as the faro. Running south, parallel with the western face of the outer faros, all the way to Wilingili Island, on the southern face of North Male, we remained in sight of a number of faros of all sizes and shapes. Their western rims are generally wider than the eastern, with deep blue or lighter-colored lagoons, according to their depths, often edged on the east by extremely narrow rims, some of them a mere thread but a few feet in width, the wider reef flats being always the outer rims, as on the east face of North Male. The western edges of the outer rims were lined with a belt of great patches of corals extending sometimes across the sandy rim flats of the faro. The lagoon of the third faro south of Hembadu Pass is divided into two by a transverse bar like the one of the northeast horn of Mulaku, figured on the Admiralty Chart and mentioned by Mr. Gardiner. The three faros to the south are triangular with narrow rim flats. One of them is nearly two miles wide. The faro at the southwest angle of North Male is narrow, and on its eastern extremity a small island, fairly wooded, has been thrown up. On the eastern extremity of the faro to the west of Wilingili Island a couple of small islands have been thrown up which are not indicated on the charts. On the west face of North Male north of Hembadu Pass, heaps of coral boulders occur on the exposed points of the great faros which flank the northern part of the group to the south and to the north. 1 Mr. Gardiner's statement that in the Maldive Atolls the bottom is perfectly smooth and that no new banks are springing up does not seem in accordance with the soundings of such groups as North Male and North Malosmadulu. In both these groups as well as in others are many sand banks and rings with varying depths, of from three to four feet on the rims to five or six fathoms. Nor is it the case in the "jungle," a tract covered with hundreds of heads within the twenty-fathom line in the central part of North Malosmadulu. This as well as the general variation of the soundings shows a far greater range in the depth of the bottom than one would infer from the remarks of Mr. Gardiner. NORTH MALE. 51 The outline of the reef flat to the south of Akirifuri is very striking. The lagoon face of the flat is formed by a series of great bays separated by points jutting out eastward two to three miles, in marked contrast to the rectilinear course of the western edge of the flat. This peculiar mode of growth of the corals shows that, if the bays they formed were closed across the east face, what might appear as a series of faros which had coalesced or become united had a very different origin, due to the rapid extension of coral spits at right angles to the western edge of the faro. Similar reef flats occur in other groups of the Maldives in Felidu, in South Nilandu, in Kolumadulu, in Suvadiva and others to a less, extent. But nowhere are these cusps developed to such an extent as in this reef flat of the northern part of North Male. The vegetation on Akirifuri is scanty, and much of it is dying. The beaches of the island are steep, of very coarse yellow coral shingle with stretches of small coral boulders at the base of the beach. The vegeta- tion of the islands of the faro to the north of Hembadu Pass is not vig- orous ; many of the trees and bushes on the outer belt of vegetation are dying or dead, having been more or less buried by the advancing sands blown into the islands from the beach (PI. 19. fig. 1). Several of the sand-bars and banks marked on the chart as covered with low vegetation have disappeared, or the vegetation has been killed by the encroaching sands. To the westward of our track on the way to our anchorage off Kagi we passed a narrow crescent-shaped sand bank flanked with shingle. The vegetation on Kagi is most meagre, perhaps from its exposure to the full force of the northeast monsoon. This island is edged by coral sand beaches. On the southern face of Kagi corals are growing in abundance within the usual limits of depth. The northern extremity of Male is open to the northwest, where it is flanked by a narrow reef flat for a distance of nearly five miles, similarly the northern faces of South Male and of Felidu are not well protected, and the greater part of the northeast face of Ari is open for a stretch of more than twelve miles. The northern face of North Malosmadulu is also open, as well as the northeast face of Ihavandiffulu and of Tiladummati. 52 THE CORAL KEEFS OF THE MALDIVES. On passing ont of North Male through the eastern of the passes of the northern face of the group, we saw no houlder belt or heaps of shingle either on the sea faces or the pass faces of the two northern faros of North Male. The faro to the west of the pass is really a diminutive irregularly rec- tangular atoll. It has a narrow entrance to the north, the western face of which is flanked by a small sand-bar. Within the lagoon are a few heads and diminutive faros (a faro within a faro) ; the greatest depth of the lagoon indicated on the chart is eight fathoms. The absence of vegetation on the west face of North Male to the south of Hembadu Pass is a very marked feature, not only of that group, but also of most of the other groups of the Maldives. There are but few small islands on that face of the groups as compared to the number and size of the islands clothed with vegetation found on the eastern faces of the Maldive groups. One need only compare the many islands on the east face of Ari, well clothed with vegetation, with the great faros of the west face, on which only here and there rises a diminutive island. It is on the eastern face of Ari also that the greater part of population is gathered, sheltered to a certain extent from the strength of the southwest monsoon. For the east face though exposed to the northeast monsoon, is, as it were, the lee side, if we can call that face by such a name, when we have a southwest and a northeast monsoon season. It merely indicates that the lee face is exposed to a monsoon blowing with less strength than on the western face. The Sultan's Buogalow, at Anchor off Male. GAHA FAKO. 53 Gaha Faro. Plates 1, 3 ; 8 a, figs. 6, 8 ; 8 b, fir/. 11. Gaha Faro is eight miles in length by four in width, four to five times larger than many of the rings and faros of the Maldives. The greatest depth of its lagoon is twenty-two fathoms. Gafaru Island is the only land on the rim of this atoll, with the exception of small sand banks, one on the southern face to the west of Gafaru and the others on the eastern spit of the northern and the southern spit of the western entrances to Gaha Faro. Gafaru Island is edged by shingle beaches with a few patches of coral bould- ers on the outer edge of its southern reef flat. The reef flat enclosing the lagoon of Gaha Faro is very narrow and shallow, the sea breaking over its surface and along the continuous outer line of the flat broken only by a narrow pass to the north and a similar one opening to the west. The con- tinuous reef flat with its two passes gives Gaha Faro all the features of an atoll which it is difficult to distinguish from a gigantic faro. Had we seen it in the Pacific it would pass as an atoll in no way differing from the many of its kind in the Ellice, Gilbert, or Marshall Islands. Yet in the Maldives, Gaha Faro and similar atolls are regarded as faros. Gaha Faro is separated from its adjoining cluster of faros on the northern point of North Male by a channel of not more than a mile and half in width, but with a depth of one hundred fathoms. 1 While the faros of most groups are separated by narrower channels with a greatly varying depth, often not more than seven to eight fathoms, generally about twenty fathoms in the central part of the lagoon. The freest possible circulation exists between the lagoon and the sea over the shallow reef flats of the atoll. Coral patches are scattered in abundance over the rim flats. At Gaha Faro as at many of the openings of faros and off the mouth of the passes leading into the larger enclosed basins, the water is often seen to be turbid, evidently carrying in suspension a considerable amount of silt stirred up by the action of the sea on the exposed face of the inner part of the lagoons, or of the islands and banks within an atoll, or from its shallow bottom. 1 See the depths of the channels separating North, Middle, and South Malosmadulu which range from one hundred to one hundred and fifty fathoms (Pis. 3 ; 8 b. figs. 9, 10). 54 THE CORAL REEFS OF THE MALDIVES. Goifurfehen&u. Plat.es 1,8; 8 a, fig. 8 ; 8 b, fig. 10; 19, fig. 2; 20. We skirted the eastern extremity of Coifurfehendu 1 as well as the eastern half of its northern face ; this gave us a fair idea of the structure of the atoll. It is elliptical, about ten miles in length and nearly five across its widest part. The atoll is surrounded by a continuous reef flat somewhat wider on the north and east sides than on the southern face ; it has a single pass near the southwest horn. The deepest part of the lagoon is twenty- three fathoms; the average depth is from eighteen to twenty fathoms. The central part of the atoll is clear of dangers. On the eastern part of the northern reef flat of Goifurfehendu are three large islands of which the island of Goidu, at the very eastern point of the atoll and on the widest part of the reef flat, is the largest. The southern extremity of Goidu Island appears to be wasting away, a small adjoining island marked on the chart has disappeared. A narrow sand beach extends along the central part of Goidu parallel with the outer belt of vegetation close to the water's edge. At the northeast point of Goidu the upper part of a steep shingle beach has been driven far into the outer belt of vegetation. On the northern face of Goidu Island the vegetation is not flourishing, and it reaches the water's edge. The eastern face of Goidu Island is steep to close to the edge of the outer insignificant reef flat; in places the surf beats directly upon the beach. The projections of the northern face of Goidu form a large rectangular bay ; the island seems to be increasing in size towards the west. The northeastern part of Goifurfehendu atoll is exposed to the full force of the northeast monsoon. On the outer edge of the reef flat are numerous great patches of corals, showing that they grow abundantly on the northern slope of the atoll. On Fehendu, the next island to the west on the wide northern reef flat, the vegetation is fine with many clusters of large trees. Opposite Fehendu the outer reef flat is wide, the island thus appears to be placed well in the middle of it. Inafuri is the westernmost island of Goifurfehendu. Gardiner 2 notes the "more purely oceanic conditions of this atoll than any other of the more perfect atolls of small size except Addu " as well as 1 On Plates 1 and 3 Goifufehendu should be Goifurfehendu. 2 Loc. cit., p. 377. KARIDU. 55 its slight protection by the east horn of South Malosmadulu ; he recognizes the former recent connection of Fehendu and Furudu which were probably formed by the piling up of sand on the flats fringing the lagoon. Accord- ing to him, " the lagoon is increasing in size at the expense of its en- circling reef," the outer contour of his figure being the same as that on the Admiralty Chart, but the reef flat is charted much broader than in Mr. Gardiner's figures. 1 Furudu Island is wasting away. 2 Karidu. Plates 1, 8; 8 b, fig. 11; PA. Karidu is a small elliptical atoll about two miles in greatest diameter, nearly halfway between Gaha Faro and Fadiffolu. It rises abruptly in a chan- nel nearly thirty miles across, from a depth of about three hundred fathoms. It would be called a faro were it placed on the face or the central area of one of the larger of the Maldive groups. A wide curved island, follow- ing the line of the reef flat, occupies the greater part of its southern rim, to the west of which rises a small islet. The reef flats to the north and south are shallow and enclose a comparatively shallow lagoon. The point of the south face of the east end of Karidu is covered with coarse shingle ; that part of the island is steep to ; the reef flat gradually widens to the west ; along the centre of the south face of Karidu rises a steep shingle beach which extends to the base of the outer belt of low bushes. A few large forest trees grow in the centre of the island, much of the vegetation of Karidu is dead ; many of the bushes formerly above the beaches now stand in the water on the edge of the reef flat. At the western extremity of Karidu the central shingle beach has gradually passed into a coral sand beach which extends in an ill-defined sand spit along the narrow reef flats of the northern face of Karidu. On the reef flats corals abound ; huge patches extend from the outer rim of the atoll towards the lagoon. On the shallow northern reef flat the breakers due to the northeast monsoon form an almost continuous line. l hoc. cit., fig. 90, p. 377. 2 Loc. cit., fig. 91, p. 378. 56 THE CORAL REEFS OF THE MALDIVES. The lagoon has gradually filled ; it now consists only of a large bottle- shaped pool towards the eastern end of the atoll, probably with four or five fathoms of water, judging by its color. A narrow opening leads into the lagoon ; the pass is indicated by a heap of boulders on the western face and a small sand-bar on the eastern. The island of Karidu as it exists to-day does not, owing to the wasting away of the western extremity, occupy as large a proportion of the area of the reef flats as is indicated on Moresby's chart (1836). Malosmadulu Plateau. Plates 1, S; 8 a, fig. 5 ; 8 b, figs. 9, 10; 8 c, fig. 28; 22-80 ; 81, fig. 1. South, Middle, and North Malosmadulu resemble Male and Ari as far as the location of the faros and of the islands on the outer faces of the Malosmadulu plateaus is concerned (PI. 3). Their western face is flanked by large faros, some of them over five miles in length, separated by wide and deep passes. In fact, the faros are atolls of considerable size with as great a variety in size and shape as exists among widely separated atolls in other coral reef regions. Compare, for instance, such faros as Maduni Faro, Ekuru Faro, Ma Faro, Defili Faro, Mawa Faro, which are lagoons surrounded by wide reef flats, with faros having one or two islands on the rims, such as Kandu Gandu, Bodu Faro, Femfuri Faro and Turadu, 1 having all the features of small independent atolls with a minimum of land masses on the reef flats. The former really belong to the same category as the atolls, though no islands or islets have as yet been thrown up on the rim flats. The east face of North Malosmadulu and the southeast face of South Malosmadulu are flanked by a few faros and a number of islands running at right angles to the trend of the east face of Malosmadulu ; they are sepa- rated by wide and deep passes, some of them over three miles in width. North Malosmadulu bears a close resemblance to Ari; like it, it is flanked on the west by large faros and on the east by numerous islands, all steep 1 Turadu, according to Gardiner, has become modified into a linear reef flat ; loc. cit., p. 380, fig. 94. MALOSMADULU PLATEAU. 57 to, well wooded, and separated by deep passes. It has, like Ari, many faros and rings within the enclosed basin of the group ; the central basin, inside the twenty-fathom line, is crowded with a mass of faros, islands, banks, bars, rings, and heads, which constitute a perfect labyrinth and render the navigation of that part of North Malosmadulu impracticable. The islands near our truck, both in South, Middle, and North Malosmadulu, seem to have been built on one pattern. At first they are small banks without vegetation thrown up on the rim of a small flat or faro. The bank gradually increases in size, occupying a larger part of the rim. The lagoon then becomes silted up, and eventually the sand bank occupies the greater part of the area of the faro, becoming clothed with vegetation as it increases in size, and finally passing into a wooded island, steep to or with a narrow reef flat, as in Wadu, Medu, and other islands in the southern part of North Malosmadulu, or of Karidu, Mararrekellu, and Anghenufuri in Middle Malosmadulu and Kendu, Hurudu, and others on the northern and southern faces of South Malosmadulu. The deepest sounding indicated in the South Malosmadulu group is thirty- eight fathoms ; the greater number of the soundings are between twenty -five and thirty fathoms. In North Malosmadulu thirty-one fathoms is the deepest sounding indicated on the chart, and outside of the area enclosed by the twenty-fathom line the average of the soundings is between twenty and twenty-five fathoms. In Middle Malosmadulu the deepest sounding is twenty-seven fathoms, the soundings average about twenty fathoms. It is interesting to compare the position of the atoll of Goifurfehendu (Pis. 1, 3), separated from South Malosmadulu by a channel nearly six miles in width and a greatest depth of three hundred and two fathoms, with that of the typical Maldivian group of South Malosmadulu, separated from Middle Malosmadulu by a channel varying in width from one to two miles with a greatest depth of one hundred and forty fathoms, or with that of Middle Malosmadulu, divided from North Malosmadulu by a channel of nearly three miles in width, and a greatest depth of one hundred and thirty-five fathoms. Contrast - this with Femfuri Faro, which is north of the southwest horn of North Malosmadulu and separated from the faros to the north and to the south by channels, the one about a mile and 58 THE CORAL EEEFS OF THE MALDIVES. a quarter wide, with a greatest depth of twenty-seven fathoms, the other nearly two miles wide, with a depth varying from twenty-two to twenty-six fathoms. The relation which these faros or small atolls hold to Middle Malosmadulu are similar to those of Goifurfehendu to South Malosmadulu, or of South Malosmadulu to Middle and to North Malosmadulu, or of Gaha Faro to North Male. The distances and depths separating the atolls are merely questions of quantity. The faros, or atolls, or lagoon reefs, built up on the summits of the plateaus rising from various depths, are all atolls, though they may differ greatly in size and shape, they all represent variations similar in kind ; there is no fundamental difference between them. The lagoons of the largest Maldivian atolls, like Addu, Gaha Faro, Goifurfehendu, Makunudu, Wataru, are but little deeper than the lagoons of smaller faros like those I have just mentioned, or of the many others scattered over the various groups of the plateau of the Maldives. We cannot distinguish these faros from atolls, whether isolated or forming a part of an extensive chain, or occupying a portion of the face of the Maldivian plateau, or scattered within the basins enclosed by these chains. The faros and atolls have the same characteristic features, and nowhere is this more prominently brought out than in the group of islands forming the so-called Tiladummati and Miladummadulu atolls (Pis. 2, 3). Other cases of isolated faros and of atolls are the Powell Islands, about eight miles west of Miladummadulu; Makunudu, nearly ten miles off the west coast of the same group with a depth of nearly eight hundred fathoms in the centre of the dividing channel ; Ihavandiffulu, separated from the northern face of Tiladummati by a channel of nearly four miles in width, with a greatest depth of two hundred and fifty-one fathoms. The large faros at the northern extremity of Tiladummati, separated as they are by wide and deep channels, show the impossibility of distinguishing faros from atolls. An extreme case, perhaps, is that of Karidu (PI. 3), distant about six- teen miles from Fadiffolu and twelve from Gaha Faro, with greatest depths in the channels of three hundred and seventy-two and two hundred and SOUTH MALOSMADULU. 59 fifty-eight fathoms. This small faro, for it would be considered small even if it formed a part of one of larger groups, differs in no way from a large number of the characteristic faros of the Maldives. But Karidu has all the characters of a small Pacific atoll ; it is isolated, separated by wide channels from the nearest groups, and instead of rising, as do other Maldivian faros, from a broad bank with an average depth of from twenty Jo thirty-five fathoms, it rises as the summit of a diminutive mound, surrounded by depths of nearly three hundred fathoms. Side by side iden- tical structures exist which have been built up in one case from the mini- mum depths of twenty to thirty-five fathoms ; and in the other, upon summits rising from depths of over three hundred fathoms ; and certainly the former do not owe their origin to subsidence. Otherwise Karidu has been formed by the subsidence of a plateau to a depth of nearly 300 fathoms, Gaha Faro to 100, Makunudu to nearly 800, Ihavandiffulu to 250, Middle Malosmadulu to 100 or 150, Rasdu to 120, Wataru to 200, the hosts of small atolls on the outer faces of the Maldivian groups and of the rings and faros within these groups to twenty or thirty-five fathoms (PI. 1). The different depths to which these atolls have sunk all occur within the same district, and postulate an irregularity in the rate of sub- sidence of different and adjoining tracts of this area which it is difficult to imagine, and for which a very slow and regular rate of subsidence has always been demanded. We can far more readily understand the existing conditions as due to the different levels to which the irregular bottom of the great Maldivian plateau has been elevated. In fact, the topography of this wide submarine ridge is neither more nor less varied than that of a mountain plateau. South Malosmadulu. Plates 1, 8 : 8 b, fig. 10; 8 c, fig. 28; 22-28 ; 29, fig. 2. South Malosmadulu (PI. 3) is irregularly triangular, its apex facing east; the greater part of the enclosed basin of the group, in a belt parallel to the south face is filled with large faros, many of them more than three miles 60 THE CORAL REEFS OF THE MALDIVES. in length. In addition a number of small banks, islands, and islets or faros are clustered in the northeastern area of the group. The slope of the eastern face of South Malosmadulu is, according to the soundings (PI. 1), very gradual, far more than that of many of the other groups. The one hundred-fathom line is at a greater distance from the edge of the outer row of islands, faros, and reef flats. This is not the case with the western face of either North or South Malosmadulu (PI. 1). Some of the islands in the northeastern area of South Malosmadulu are of considerable size, and many of them are well wooded. An island to the westward of Mabaru is specially noted for its fine trees. The eastern edge of the reef flat to the south of Mabaru is covered with patches of corals, and fringed with stretches of boulders and heaps of shingle. The extremities of Mabaru (PI. 22, fig. 2), exposed to the action of the northeast monsoon, are wasting away on the sea face, as are the islands to the south, many cocoanut palms lying prostrate at the base of the beaches. The vegetation on the islands of the east face of North and South Malosmadulu is far more flourishing than that of Facliffolu, probably be- cause it is not exposed to the force of the southwest monsoon. On the southeast face of South Malosmadulu both the monsoons blow along the trend of the face, so no heavy swell or sea pounds upon the beaches of that part of South Malosmadulu. The deep wide passes separating the islands of the southeast face of South Malosmadulu give the impression that the islands are scattered, irregularly over an indefinite sea space, and not that they are part of the land rim of an atoll (PI. 23). Aidu Island (PI. 22, fig. 1) is steep to, with deep wide passes on each side. It is surrounded on the sea face by steep, coarse shingle beaches, passing into coral sand beaches towards the west ; it is covered with fine vegetation. To the south of Mabaru, between it and Aidu, a large, irregu- larly shaped faro, Wandu Faro, fully three miles in length, extends west- ward. Two islands are found on its western rim. Another large faro, Huni Faro, lies to the south of Aidu, with islands at the eastern and western extremities; its lagoon has a greatest depth of six fathoms. The islands SOUTH MALOSMADULU. fil are flanked by steep coral sand beaches ; the outer edge of the rim flat is covered with large patches of thriving corals. On the sea face of Kum- finadu are outlying patches of boulders and angular reef rock, honey- combed, pitted, and undercut, similar to patches found elsewhere in the Maldives, indicative of a slight former elevation. Gardiner l says most of the reefs to south and east show traces of a broad line of elevated reef rock on the sea face, with extensions along the faces of the passages which are still going on. He also states that the bottom of South Malos is everywhere hard sand, and singularly barren of organic life, 2 and that dredging points to a general growth of corals on the outer edge, though it may only be in isolated patches. After skirting the east face of South Malosmadulu we entered it through the pass east of Hitadu ; on either side of the island are Olugeri and Maduwari, both steep to. The vegetation of Olugeri as seen from the pass appears to be quite luxuriant. The west face of Maduwari (PI. 25) is flanked with a steep beach of coarse coral shingle, so steep as to form an almost vertical wall along part of the island. The shingle has been driven in among the base of the trees and between the clumps of bushes well back from the beach towards the interior of the island. The coarse coral shingle surrounds the sea face of the island and the greater part of the eastern face along the pass. On the lagoon side of the island it is flanked by steep coral sand beaches. On the steep to sea face slope of Maduwari corals are very luxuriant down to a depth of twelve fathoms, where they become small, and are separated by belts and lanes and patches of sand until they are completely choked and disappear in depths of sixteen to twenty fathoms. 3 The southwestern horn of South Malosmadulu is formed, according to the chart, by Turadu, a large faro about four miles in length. The main part of the faro has at the northeast angle a regular entrance like that of any atoll of its size, is bounded on the southwestern and southeastern faces by a wide reef flat, on which are two islands, and which forms the southern rim of an elongated lagoon with from three to six fathoms in 1 Loc. cit., pp. 384, 385. - Loc. cil., p. 381. 8 See Gardiner's account of tins island, loc. cit., p. 164, fig. 30, and p. 384. 62 THE CORAL EEEFS OF THE MALDIVES. depth. The northern rim, according to the chart, is a very irregular narrow belt awash here and there with a secondary lagoon at the northern spit of this rim. The changes at Turadn from the conditions marked on the Admiralty Chart are, as stated by Gardiner, very marked 1 and were also noted by us at the time we passed by Dunikolu and anchored near Embudu, but I cannot agree with Mr. Gardiner that the extensive changes observed there are wholly due to solution. 2 The cyclone by which the island was struck would account for the movement of a great mass of sand such as perhaps formed the northern rim of Turadu in 1836. Boats from Turadu. To the northeast of Turadu rises Dunikolu, a small island with a low coral sand beach and covered with a scanty vegetation. Near our anchor- age stretched another, smaller faro (Velengeli), a little over a mile in diameter, with an island on its western rim having somewhat the character of an outer island. The wide rim of this faro enclosed a shallow lagoon the great part of which had been filled up. The little island of Embudu (Pis. 26, 27) to the north of our anchorage is covered by a scanty vegetation ; it is surrounded by a narrow reef flat and is flanked by steep almost vertical beaches of fine coral sand. At their 1 Loc. cit., fig. 94, p. 380. 2 Gardiner states (loc. cit., p. 386) that in the northeast part of South Malos there seemed to be little change within the lagoon. SOUTH MALOSMADULU. 63 base there are patches of beach rock. The deeper slopes are covered with a magnificent growth of corals extending down to a depth of from twelve to fourteen fathoms, where the corals become patchy and gradually disappear in from eighteen to twenty fathoms in a sandy bottom. The beaches are covered with windrows of small masses of pumice varying from the size of hazel-nuts to that of one's fist. Many cuttle-fish bones and Spirilla shells have also been blown up above high-water mark. On the outer edge of the rim of Mutalifoori, projecting here and there high above the outer rim, extends a belt of coral boulders. The little island on the western rim is bordered with low sand beaches ; the rim of the faro is comparatively wide, enclosing a small lagoon. The outer belt of the rim flat is thickly covered with great patches of corals which, as off Embudu, seem to be very flourishing on the slopes of the lagoon reef. To the east of our track on the way north about a mile and a half from Mutalifoori, we passed a large elliptical ring with from three to six fathoms in depth and more than two miles in length. On the western side of our track nearly opposite, another faro was passed, fringed on the outer edge of the east face of the rim with irregular patches of boulders. We passed close to the eastern extremity of Wakaru, a small island on the eastern point of a large ring with from three to four fathoms in the lagoon. On the east face of Wakaru a steep beach has been thrown high up in the outer belt of bushes. At the eastern point the base of the beach is flanked by stretches of beach rock. At both Embudu and Wakaru the outer beaches have been thrown up considerably higher than the general level of the central part of the island. This forms a rim to an inner sink or dish from twelve to eighteen inches or more lower than the summit of the beach. This apparent dishing is quite common among the small central islands of the Maldives, and when it occurs on the rim of an island which is steep to, the inner sink might readily be mistaken for the bottom of a dry lagoon, or for an atoll the lagoon of which had been slightly elevated above the general level of the surrounding reef flat. As we passed Megeli we noticed a small, low sand-bar on the southern face of the faro to the west of it, which is not indicated on the chart. 64 THE CORAL REEFS OF THE MALDIVES. Megeli Is] and is crescent-shaped. It occupies the eastern part of the rim flat of a faro of which the lagoon has been completely filled. The small islet on the east face of Muduwari (PI. 28, fig. 2) is covered with low bushes. From Muduwari north towards Kendu and the islands which form the northern boundary of South Malosmadulu the atoll is clear. Kendu (PI. 29, fig. 2), like the other islands to the eastward, is steep to, with steep and high sand beaches, much higher than any sand beaches we have seen in the central groups of the Maldives. The western point of Kendu, which is exposed to quite a reach of the southwest monsoon, is covered with coral boulders. The islands and faros of the north face of South Malos- madulu are divided by deep and wide passes. At Kiadu (PI. 28, fig. 1) the apparent depression due to the height of the enclosing beaches is very marked. Kiadufuri is a beautiful elliptical ring with wide reef flats enclosing a lagoon with from five to six fathoms in depth, and an extensive sand bank on the southern face ; this is not indicated on the charts of 1836. At Hanikandu Faro, which forms the northwestern angle of South Malosmadulu, the enclosed lagoon is said to be twenty fathoms in depth. Middle Malosmadulu. Plates 1, 3; 8 h, figs. 9, 10; 29, fig. 1. The large triangular bank wedged in between South and North Malos- madulu may be called Middle Malosmadulu; it is separated from them by deep channels of from two to three miles in width, and with depths varying from one hundred and ten to one hundred and forty fathoms in the axis of the channels. The enclosed area of Middle Malosmadulu is free from patches or faros. Its faces are bounded by comparatively few islands and large faros, separated by wide and deep passages. The greatest depth of this bank is twenty-six fathoms, with a general average of about twenty fathoms. We examined some of the islands and faros of the southern and north- western faces. Karidu Island is steep to, like Kendu and the islands on NORTH MALOSMADULU. 60 the northern face of South Malosmadulu, and flanked by high and steep sand beaches. Kari Faro (PI. 29, fig. 1), a large faro to the west of Karidu, has at the western extremity a broad crescent-shaped reef flat with steep sand beaches, and a narrow elliptical lagoon of a dark blue color, from three to four fathoms in depth ; the extremity of its eastern rim is occupied by a small sand-bar topped by a clump of vegetation. Many of the islands of both North, Middle, and South Malosmadulu appear as slightly dished on account of the height to which the summits of the encircling beaches have been tbrown up. Mararrekellu, an island on the northwestern face, is, like Karidu, steep to, with steep coral sand beaches, and covered by a scanty vegetation. Anghenufuri, marked as a small lagoon flat to the east of Mararrekellu, since 1836 has become en- tirely filled. It is now an islet on the eastern flank of a small reef flat. Many of the smaller islands of this part of Middle Malosmadulu seem to be old faros of which the lagoons have been silted up, and subsequently the islands have encroached on the reef flats so as to occupy the whole of the former area of the faro and the surrounding reef flats. So that former diminutive circular barrier reefs have gradually been changed to fringing reefs, as Karidu has in part. North Malosmadulu. Plates 1, 8 ; 8 a, fig. 5 ; 8 b, fig. 9 ; 8 r, fig. 26 ; SO : 31, fig. 1. The northern extremity of North Malosmadulu is open like that of Ari (PI. 3) ; on the northern part of the eastern and western faces the faros and islands are more distant and separated by wider and deeper passes than in the southern half of the group. Comparatively few faros, rings, and islands occur in the interior of the northern half of North Malosmadulu, and in a belt parallel to the southern part of the east face. But in the central part of North Malosmadulu, towards the western face, the twenty- fathom line encloses a large area with innumerable islands, islets, sand-bars, faros, and coral heads. The southern part of this area is flanked by a belt of a few more widely separated large islands and faros. The islands on the east face of North Malosmadulu are large (as Maldive Islands go), and 66 THE CORAL EEEFS OF THE MALDIVES. are separated by wide and deep passes, gradually becoming wider towards the northern part of North Malosmadulu. Some of the northern passes are over two miles in width. The southwestern horn of North Malosmadulu is formed by Mamanaga (PI. 30, fig. 1), a large triangular faro the base of which is nearly five miles. It has a wide rim on the western face, and a lagoon with from five to six fathoms in depth. A small sand bank has been thrown up on the northwest horn of the faro. On the southern face to the east of Mamanaga are two smaller faros (Velengeli and Furuwari), somewhat more than a mile in length, with very wide eastern rims, a small lagoon towards the western extremity, and each with a small bank on the eastern horn of the faros. Velengeli is flanked with coarse shingle. The southwestern part of North Malosmadulu is quite clear of faros, banks, rings, and islands, so that when entering the group through the pass to the east of Mamanaga, we opened a great stretch of clear water to the northeast as if we had passed out into the open sea. To the north of the southwestern horn and separated from it by a channel fully a mile in width lies Femfuri Faro, a long triangular faro fully three and a half miles in length, with the apex to the east and the eastern horn occupied by a small island. Its northern and southern rims are narrow, they pass gradually into a wide western reef flat. Tura is a small island facing the wide pass on the west face of North Malosmadulu between Femfuri Faro and Ma Faro ; it is steep to, bounded by a high steep sand beach ; near the southern point extends a stretch of coarse shingle beach, and on the western spit of the island a belt of small coral boulders. After passing Tura one sees to the northward rising upon the horizon the islands occupying the inner belt of North Malosmadulu to the south of the twenty-fathom curve. Maregiri 1 Island lies south of Bodu Faro, the large faro which occupies the western horn of North Malosmadulu. On the southwest it is flanked by steep sand beaches from ten to twelve feet high, with stretches of coarse shingle driven towards the interior among the bushes and between the base of the trees beyond that belt. A narrow boulder belt flanks the base of the beaches. At the western point there is a narrow reef flat 1 Gardiner, loc. cit., p. 387. NORTH MALOSMADULU. 67 covered with beach rock, and a patch of reef rock honeycombed and weath- ered, pitted and undercut ; elsewhere the island is steep to. Facing the pass to the south, the beach is composed of coarse shingle and reaches of sand with large boulders at the east point. Unfortunately we could not land and examine the high dunes on the island mentioned by Mr. Gardiner. 1 Mr. Gardiner has made a careful comparison of Moresby's Chart : of a stretch of the western face of North Mahlos, extending from Kukuludi Faro to Maduni Faro, with sketches of his own. 2 He finds well-marked velus in Dina Faro and Ma Faro ; none are on Moresby's Chart. 3 At Telin Faro and Bodu Faro, the broad eastern rims of the vein shown by Moresby have disappeared and they have become connected with the general lagoon. The faros to the north do not indicate any special change. Mr. Gardiner considers the enlargement of the velus as due to the solvent action of sea water. Similar differences on the chart between Turadu and Dunikolu he attributes to the same cause. When, as is often the case, the rims of the velus are merely sand, their demolition and the increase of size of the enclosed velus is not necessarily clue to solution. The filling up a vein would result in a reef flat. A study of the northeastern point of Tila- dummadulu indicates an increase of the land areas of the faros. I fully agree with Mr. Gardiner's 4 view that every large reef on the bank has grown up by itself. It does not follow, however, from the changes he has observed, that North Malos is approaching the condition of a perfect atoll. We can trace the passage of such rims composed of atolls, into long linear reefs only in very limited areas. We passed out of North Malosmadulu through the pass west of Tura, and entered again by Maregiri Island (PL 30, fig. 2; 31, fig. 1). Steaming in a southeasterly direction, we crossed the corner of the area bounded by the twenty-fathom curve, and obtained an idea of the character of the "jungle of reefs," 5 crowded as it is with banks, rings, coral heads, faros, flats, and islands and . islets. The first islet we came upon was a small sand islet with a clump of bushes in the centre, with steep coral sand beaches and 1 Loc. cit., p. 1G6. 2 Loc. cit., p. 169. 3 Though the original charts might have indicated such a velu from the position of the coral rocks on the outer edges on both these faros. 4 Loc. cit., p. 171. 5 Gardiner, loc. cit., p. 1C7. 68 THE CORAL REEFS OF THE MALDIVES. a narrow flat, the outer edge of which is indicated by a line of small coral boulders. The island of Fusmundu fills the eastern rim of a well-defined faro about a mile in length. It is flanked by high steep coral sand beaches. The outer northwestern edge of the rim of the faro is lined with small coral boulders. From the great height of the beaches enclosing the central part of the island it presents a dished appearance. A few fine large trees rise in the central part of the island, but many bushes and smaller trees of the outer belt have been killed by the spreading sand or are in a dying condition from the encroachment of the sea. To the eastward we came upon Muduwari, a small nearly circular island occupying the greater part of the flat upon which it has been thrown up. The coral sand beaches are high and steep, the summit of the beach rising far above the central part of the island. The island seems to be wasting away ; a small islet covered with a little vegetation once a part of the main islet now stands isolated upon the reef flat near the eastern point of the island. On the central part of the north side of Hoholundu Island sand has been blown from the summit of the high sand beaches well towards the interior, and has overwhelmed many of the bushes of the outer belt and buried the base of a number of the larger trees. Hoholundu Island, like the islands we have just described, occupies the eastern rim of the flat upon which it has developed. Hoholundu Faro is nearly a mile and a half in length. At no time during our cruise through the Maldives did we pass through a region containing so many banks, rings, faros, heads, islands, and islets as on our way from Maregiri to Medu, when we cut across the only part of the labyrinth of North Malosmadulu which seemed to promise a clear passage. The various islands, the different kinds of flats, bars, rings, and faros in the interior of North Malosmadulu, differ in no way from those on the outer rim of the so-called atoll of North Malosmadulu. They have all been formed in the same way both in the interior of the group and on the outer belt. The outer faros and islands being more exposed to the modifying influences due to currents and the action of the winds, have developed more freely and with greater rapidity. Yet some of the islands NORTH MALOSMADULU. 69 and faros of the interior compare favorably in size, in the extent of the boulder belt, the height and steepness of the sand beaches, the width of the rims and reef flats with similar structures on the outer belt of the group. This is specially the case when the islands in the interior basin are in such a position as to face a long reach either of the northeast or southwest monsoon, and are thus placed in conditions more similar to those of islands and faros on the exterior faces of the groups. Medu, one of the most important islands of North Malosmadulu, is steep to ; it is surrounded by a high steep coral sand beach with many large trees in the interior basin of the island formed by the high beaches, the summit of which rises from four to six feet above the inner flat. The upper part of the beach we found covered with windrows of masses of pumice. On the wide sea slope flat corals flourish in great abundance at from six to seven fathoms of depth, where the slope drops rapidly, and corals become less numerous ; at a depth of about twelve fathoms they disappear very rapidly as they become separated by lanes and patches of sand and over- whelmed by sand-bars. The vegetation of the central part of the island is made up of numerous large trees of Breadfruit, Pandanus, Banyans, and other forest trees surrounded by a thick outer belt of low bushes and trees ; many of the bushes belong to genera found on the beaches of Pacific atolls. The summit of parts of the outer coral sand beach of Medu is from fourteen to fifteen feet high ; the wide path leading across the island drops from three to four feet towards the interior after passing the flat summit of the beach. The islands we passed to the north of Medu — Digeli, Tahwahtah, Roon- gelly, as well as others we saw in the distance — all rise upon flats of con- siderable size ; like Medu, they are surrounded by high and steep coral sand beaches, but their vegetation is less flourishing, consisting mainly of low bushes and shrubs. We examined the islands on the east face of North Malosmadulu from Kotafuri north to Kuda Kura and near Anguretin ; those further south and north, which we did not visit, present, according to the chart, no different features from the former. The islands to the north of Kotafuri all have long reef spits extending to the westward. The beaches of Kotafuri are 70 THE CORAL REEFS OF THE MALDIVES. wasting, clumps of bushes and trees standing isolated on the surrounding reef flats. A coarse shingle beach faces the south side of Wandu ; towards the east and the west the coral sand beach is high and steep. There are no faros on the central part of the east face of North Malosmadulu and only one on the northern and one on the southern extremity of the face ; the other islands are generally steep to with narrow reef flats. The vegeta- tion of these islands is not prominently developed, though occasionally a fine clump of trees rises above the low outer belt of bushes and smaller trees. The vegetation of Wandu is better developed than that of either Rasmadu or Kuda Kura, the southern beach of which, like that of Kotafuri, is wasting away. The vegetation on Makara is perhaps finer than that of any other island on the east face of North Malosmadulu ; it consists mainly of extensive clumps of large trees. The island is surrounded by coral sand beaches with little or no shingle; while on the other islands on the east face of the group the eastern extremities are flanked with shingle. On Inamadu very coarse shingle extends from the eastern point half- way towards the western extremity on the southern face of the island, where it suddenly changes to fine coral sand. The beaches of the east face of North Malosmadulu are less steep than those of the islands in the interior of the group or on the west face of North Malosmadulu. The islands to the north of Inamadu have the same general character ; the vegetation is fairly developed ; they are all steep to on the southern and northern faces, with spits or reef flats extending to the westward, but steepest at the eastern points, where the reef flats are narrow and the islands are edged with steep shingle and boulder beaches. On the north faces the shingle and coral sand beaches are distributed much as on the southern faces of the islands. Kuda Kura is a long narrow island with scanty vegetation ; it is nearly in the centre of an extensive reef flat projecting to the westwai'd. A somewhat smaller reef flat, edged by a belt of small coral boulders, forms the eastern point of the island, which terminates in a steep shingle beach. An exceedingly coarse shingle beach has been thrown up on the east face of Rasmadu, as well as on the northeastern horn of Inamadu, where the shingle has been forced inland between the trees over the summit of MAKUNUDU. 71 the beach. The narrow reef flats of the eastern faces of these islands are all edged by a belt of small boulders. The eastern spits of the islands to the south of Wandu nearly all indicate a certain amount of wasting away. Anguretin and the adjoining islands of the northern extremity of the east face of North Malosmadulu are noted for their fine vegetation. Like the southern islands of the east face, they are on the whole steep to, espe- cially on the sea face, with comparatively small reef flats. The islands are surrounded by steep, coarse shingle beaches on the sea face, passing into fine coral sand beaches towards the west. Both the shiny-le and sand are driven far in between the base of the large trees across the outer belt of bushes and low trees. All the islands of the east face appear slightly dished from the heights to which the outer summits of the beaches rise above the enclosed interior area of the island. The deepest sounding in North Malosmadulu is thirty-one fathoms. The average of the soundings in the group to the north of the "wilder- ness," as well as to the south and west, is between twenty-four and thirty fathoms. In Middle Malosmadulu the depths vary between seventeen and twenty-four fathoms. In Southern Malosmadulu the greatest depth indi- cated on the chart is thirty-five fathoms. The interior of this group is somewhat deeper than either North or Middle Malosmadulu, a great many soundings indicating a depth greater than thirty fathoms. Off the northern extremity of North Malosmadulu, separated by a channel of three miles in width, are Powell's Islands, Etingili and Alifuri, at the two extremities of a narrow isolated bank about three miles in length. We did not visit them. They stand to North Malosmadulu in much the same relation as do other atolls or lagoon reefs and islands such as Wataru Reef, Rasdu and Toddu Atolls, to Ari and Mulaku. Makunudu. Plates 1, 2; S a, fig. 7; SI, fig. 2; 32; 83, fig. 1. We examined the northeastern part of Makunudu (PI. 2), an atoll about nine miles west of Miladummadulu. It is about fifteen miles long, run- ning nearly north and south, and quite narrow, — not more than three 72 THE CORAL REEFS OF THE MALDIVES. miles wide. The northern face of the atoll is angular; its northeastern horn is well defined by Makunudu Island (PI. 31, fig. 2) ; the rounded southern face of the atoll forms its southern horn. The lagoon of Makunudu is of considerable depth ; seventeen fathoms are indicated on the chart. As seen from the outer reef flat, it appears to be thickly studded with bars, flats, and heads. There are two shallow passes on the east face of the atoll. Immediately south of Makunudu Island lies the small island of Fengbu Hurah, in the centre of the east face the islet of Faro Doru (PI. 33, fig. 1), covered with tall bushes, and on the northwestern horn a cluster of diminutive islets covered with bushes. The reef flat which connects Fengbu Hurah with Makunudu is flanked along its outer edge by coarse shingle, and the islands themselves really constitute one island, connected as they are by long sand spits, forming a more or less continuous coral sand beach between them. On the chart the outer reef flat of Makunudu is characterized as a " dry reef." As far as we examined it, and from what we could see looking across the atoll, the outer edge of the narrow reef flat is edged by a low wall of angular coral boulders and heaps of coarse shingle (PI. 32) ; the boulder wall is undercut and greatly weathered. This belt and the shingle heaps form a dry reef, as it were ; it extended as far as we could trace it, south to Faro Doru, on the northern face of Makunudu, and on the western face, judging from what we could see looking across the atoll. We examined the wall of weathered coral boulder off the east face of Makunudu, where it is better developed than at any other point of the atoll. It seemed to be made up of coral re^ef rock and coral boulders cemented together, undercut and greatly weathered ; rising above the gen- eral level of the surrounding reef flats, they indicate a slight former eleva- tion at that point. We observed at many places in the Maldives patches of similar reef rock, all pointing to a slight elevation at many and widely scattered localities in the Maldives. The narrow reef flat on the east face of Makunudu Island widens out at its northern face where the low wall of coral boulders is broken through to form a shallow boat passage leading into a small secondary lagoon of moderate depth which connects with the FADIFFOLU. 73 larger lagoon round the northern point of the island. The coral sand beaches of the northern part of Makunudu Island are high and steep, the beaches of the southeastern face are flanked at the base with reaches of beach rock. Occasionally a larger coral boulder rises on the outer edge of the reef flat above the general outline of the low wall. On the outer edge of the reef flat of the northeastern horn of the atoll (PI. 32), thrown up along a part of the low boulder wall, lies a small island covered with a clump of bushes. The existence of this islet in its present position, identical with that marked on the chart, seems to indicate that no great changes have taken place in the topography of the northern part of Makunudu, in spite of the existence of the fringing coral boulder wall which has been in part elevated and in part probably thrown up on the east face by the northeast monsoon and on the west face by the southwest monsoon. Fadiffblu. Plates 1, s) S>u fig- 4 • 8b, fig. 11 ; 8 c, fig. 28 ; 38, fig. 2 ; 34, fig. 1. The southeastern face of Fadiffolu (PI. 3) is bounded by a reef flat, the greater part of which is nearly two miles wide. Aligau, a small island, occupies the southwestern horn of the group ; a belt of boulders edges its southern point ; the northern and western faces of the island are flanked by sand beaches. The vegetation of Aligau (PI. 33, fig. 2) is not flourishing; much of it is in a dying condition. A belt of coral boulders and heaps of shingle extend along the outer edge of the eastern reef flat. There are comparatively few islands on the eastern reef flat, though one of them, Difuri, 1 is over three miles long. 1 hoc. cit., fig. 103, p. 397, shows the changes observed by Mr. Gardiner in Moresby's Chart of Fadiffolu. It seems to me rather hazardous to base upon native records of the former existence of islets and shoals in the centre of the lagoon any conclusions as to the breaking up of reefs and shoals in the enclosed parts of Fadiffolu. With such deep and wide openings as exist on the southwest face of Fadiffolu and from the existence of the broad reef flats on the eastern faces over which a great amount of water incessantly passes, it can hardly be called " one of the more circumscribed atolls.'' It is difficult to reconcile Mr. Gardiner's statement that " the lagoon is increasing on all sides at the expense of the encircling reefs" (p. 396) with the impression " that everywhere the rim reefs were growing together to form a single enclosing band." The position of Inawari {loc. cit., p. 399) " too far west on its reef " may be due to its eastern migra- tion, as is the case of many other sandy islands which increase either to the east or west or migrate bodily. 74 THE CORAL REEFS OF THE MALDIVES. Were it not for the existence of velus, 1 the appearance of the reef flats of the southeast face of Fadiffolu is quite that of some of the reef flats of the Pacific atolls, with their double row of islands. Fadiffolu is irregularly rectangular in shape with rounded corners. The northeastern and southwestern faces are concave, the others convex. It is nineteen miles in length and about twelve miles in width. On the southwestern face a deep gap five miles in width separates Aligau and Lowalafuri (PI. 34, fig. 1), the next island to the north; between it and Kanifuri, the large triangular reef flat which forms the western horn of the atoll, there are only two small reef flats and a small island on the southwest face of Fadiffolu. Near the centre of the southwestern gap rises Maduwari, a small island, steep to, covered with low vegetation, much of which is dead or dying, as is the case with the vegetation of so many of the islands and islets in the northern part of Male and north of that group. Lowala- furi is steep to; on the western face a steep shingle beach extends into the base of the outer belt of vegetation ; much of this is dying ; the sea is evidently encroaching upon the south side of the island. *A lagoon extends off the eastern face of the main island, which is joined by a sand-bar to the islet on the eastern rim of the lagoon reef, which probably was once a part of the larger island. Lohi is also wasting at the southern extremity; both it and Kurangdu are steep to on the western face. Some large trees are growing on Kurangdu, though nearly all in a poor condition ; this may be due to the effect of the southwest monsoon. On the west and northwest faces of Fadiffolu we have a series of inde- pendent lagoon reefs, most of them with islands such as occur on the pla- teaus of Tiladummati and Miladuinmadulu. In the latter case the faros or atolls are often separated by gaps of from five to ten miles, while in the former case they are only from one to three miles apart. A long reef flat forms the western face of the northern pass into Fadiffolu. We entered Fadiffolu by the wide pass south of Maro. The difference in structure of the southwestern and northwestern faces 1 Two of the eastern reef flats of Fadiffolu have been figured by Mr. Gardiner and the differences with the Admiralty Chart carefully described by him, loc. cit., p. 400, fig. 104. FADIFFOLU. 75 of Fadiffolu is most marked when compared to similar faces of atolls like Addu, Minikoi, Gaha Faro, or Goifurfehendu, and others where an exten- sive lagoon flat slopes gradually from the inner face of the outer reef flat towards the deeper part of the lagoon. In such an open group as Fadiffolu where the lagoon reef flat is limited to the slope of the rims of the compo- nent faros the difference becomes most striking ; we seemed when at an- chor off Mafilefuri to be off the steep slope of an isolated atoll with other atolls looming up in the distance. The gap between Maro and Mafilefuri forms a deep bay, shut off from the sea face by a belt of boulders running across the gap. On the east face of Mafilefuri a lagoon extends northward, parallel to the island. The eastern beaches of Maro and Mafilefuri are alternate stretches of shingle and coral sand. The islands off our anchorage, like those to the south on the southwest face, are also wasting away. The western faces of Maro and Mafilefuri are flanked by a steep, coarse shingle beach; the western face is steep to ; a narrow reef flat, however, exists at the northern ex- tremity of the island. We skirted the southern face of the triangular reef flat forming the western horn of Fadiffolu. The extremities of the long narrow islands on the southern edge of that flat are generally covered with coarse shingle beaches, they run close to the southern edge. When not steep to, they are flanked with coral sand beaches or smaller shingle. The point of one of the islets was specially marked by the mass of small boulders at the base of the beach, forming a low wall of modern reef rock. These boulders were angular, undercut, pitted, honeycombed, and greatly weath- ered. The action of the southwest monsoon on the southwest face of Fadiffolu is evidently very considerable. It sweeps north, unbroken, through the open space between Ari and Goifurfehendu, so that the south- west face of Fadiffolu is exposed to its full strength. To the east of Kanifuri the outer edge of the reef flat is fringed by sand banks and is^ protected by a belt of large boulders ; on the south shore of Dehu a steep, coarse coral shingle beach encroaches upon the belt of vegetation. Dehu appears to be wasting at its western extremity, where patches of vegetation are standing on the reef flat, and close to the beach 76 THE COKAL REEFS OF THE MALDIVES. portions of the belt of bushes are dying. All along the southwest face of Fadiffolu the inclination of the belt of vegetation and its slant to the east- ward shows the effect of the southwest monsoon. On the south shore of Kanifuri, which is not quite so exposed to the full sweep of the southwest monsoon, the vegetation becomes quite flourishing- again. The southern beach of Kanifuri is a steep, coarse shingle beach ; the island is edged by a narrow reef flat, and is steep to at the western extremity, which forms the western horn of Fadiffolu. From five to seven feet of water are found on the reef flat of the western horn of Fadiffolu. Though there are no passes through the great southeastern reef flat of Fadiffolu, yet a great mass of water passes over it. On the northeast face, from Faidu the extremity of the reef flat on the southeast face, four narrow reef flats occupy the outer rim as far as Kuredu, the northern horn of the group. 1 Two of the flats are separated by narrow passes, but the northeastern pass is wide and of considerable depth. There are but few rings, banks, and faros within the Fadiffolu group, all in the western parts, the eastern parts of the atoll being quite clear. The greatest depth of Fadiffolu is thirty-two fathoms ; the greater number of the soundings are between twenty and twenty-five fathoms. On the west face of Kanifuri the corals were found to be most flourish- ing off the western horn, in the belt from eight to nine fathoms, gradually becoming less numerous towards the surface. They extend to seventeen fathoms, where the sand-bars and lanes and patches of sand gradually become wider, separating the coral masses into distinct clusters, until they become irregular, decrease gradually in size, and finally disappear in the belt of dead and dying fragments of corals at the base of the talus of the reef flat. The trend of the northwest face of Fadiffolu with reference to the northeast and southwest monsoon is such that it is raked by both, while the eastern and the western faces of most of the groups are raked by only one of the monsoons, and remain under the lee during the season of the other. This readily explains the distribution of the shingle, of the coral sand beaches, and of the boulder belts on the outer edge of the rims of faros 1 Gardiner, loc. cit., p. 400, fig. 104. FADIFFOLU. 77 or of reef flats, as well as the relative number of islands and islets scattered either on the windward or the lee face of the different groups. We might as a rule call the eastern the lee face, for the strength of the northeast monsoon is not to be compared with that of the southwest monsoon upon the western faces of the Maldive Islands. Of course the west face is the lee during the period of the northeast monsoon. These conditions again are somewhat modified to the south from Haddummati to Suvadiva and Addu, where we are outside of the region of the monsoons, in the area of calms and variable equatorial winds. Native Boat, Male. THE CORAL REEFS OF THE MALDIVES. TILADUMMATI-MILADUMMADULU PLATEAU. Plates 1, 2, 8 ; 8 a, figs. 2-4, 5, 7 ; 34, fig. 2 ; 85-58 ; 79, fig. 8. Miladummadulu and Tiladummati, the largest and nearly northern most groups of the Maldives, are in reality parts of the same plateau ; their boundary is an artificial one, a mere political division line running east and west south of Mavaida (Pis. 2, 3). This great northern plateau is fully eighty miles in length with a greatest breadth of twenty and a minimum width of ten miles. It is in striking contrast to the southern groups, of which the outer faces are more or less well defined by belts of faros or of islands, or by reef flats as in Ari, North and South Malosmadulu, North and South Male, North and South Nilandu (Pis. 1, 4, 5). The outer faces are still better defined in the more southern groups like Mulaku, Felidu, Kolumadulu, Haddummati, Suvadiva and in such smaller atolls as Addu, Goifurfehendu, Makunudu, Ihavandiffulu, Wataru, Gaha Faro, and Karidu. This northern part of the Maldivian plateau forms a great arc convex to the west with a slight indentation in the central part of the western face and one less marked to the south (Pis. 2, 3). On the east face, how- ever, are two great bights nearly fifteen miles across, forming re-entering curves. The southern part of the interior of the plateau is studded for twenty miles with distant islands and faros, some of them of considerable size. They are, however, few in number compared with the many islands, faros, and banks often crowded together as in parts of the interior of North and South Malosmadulu and in some of the central groups like Ari, North and South Male, and Nilandu. In the central part of the plateau the islands are less numerous than in the southern part; there are only fifteen in a distance of nearly forty miles, leaving that part of the plateau quite an open sea. In the northern extremity of the plateau the islands are again somewhat more numerous ; there are nineteen of them, all except five of considerable size, the largest of the interior islands of Tiladummati. TILADUMMATI-MILADUMMADULU PLATEAU. 79 The southern part of the east face of the Tiladuramati-Miladummadulu Plateau is bounded mainly by distant crescent-shaped islands or faros with or without sinks or lagoons, and by two large reef flats, on the eastern rim of one of which there are two islands and on the other Kendikolu Island, one of the largest in the Maldives (PI. 3). On the southeast face there are two large faros with islands occupying the eastern face of the reef flats, one of which, Ma Faro, is over two miles long and forms the southeastern rim of a small atoll about three and a half miles in greatest length. There is an opening with ten fathoms of water on the northeast face, leading into the lagoon of the atoll, which has a greatest depth of about eight. A couple of islets have been thrown up on the northeastern edge of the rim flat. The other, Edu Faro, to the south, is also a small atoll bounded on the east by a large crescentic island on the outer face of the wide eastern reef flat, which encloses a well-defined, elongated lagoon flanked on the west by a narrow rim flat. It resembles Karidu, but is larger. Atolls similar to Edu Faro, differing, however, in shape, occur on the northern part of the eastern face of Tiladummati. One of them, Hanimadu, is four miles in length. By far the most interesting feature of the northern horn of Tiladummati are the three atolls which form its northeastern point (PI. 2). Filadu, the southern atoll, with a long bow-shaped island flanking the eastern reef flat, with sinks in the two clnb-shaped extremities of the island, and a wide pass of nearly a mile in breadth leading into a circular lagoon of nine fathoms in depth, bounded on the south by a reef flat nearly two miles wide. Kelai, the northern atoll, forming the eastern horn of Tiladummati, is over four miles in length. Its eastern face is flanked by a large hook- shaped island on the outer edge of a wide reef flat, tapering to the west and towards the southwest to form the narrow rim of a large lagoon, with a greatest depth of eight fathoms. The western rim of the lagoon is narrow, much broken, with narrow passages leading into the lagoon through the open western rim. A small island occupies the western horn of the faro. The western atoll, Warifuri, is a pear-shaped faro with a narrow island on the southern edge of the narrow rim of the eastern reef flat ; the lagoon enclosed by it is circular and a wide triangular western rim 80 THE CORAL EEEFS OF THE MALDIVES. reef flat forms the northwestern horn of the faro. The lagoon has a greatest depth of more than seven fathoms. Nearly the whole western face of Tiladummati and of Miladummadulu is edged with large faros with or without islands, which may vary from a sand islet to islands of more than a mile in length. Filadu, Kelai, and Warifuri occupy a striking position. They are placed on the steep edge of the horn of a plateau, which on the eastern face slopes in a distance of about five miles to seven hundred and eighty-one fathoms, on the northern face to six hundred and one fathoms, and two hundred and fifty fathoms in a northwesterly direction, while the depth of the central part of the plateau within the area enclosed by these lagoon reefs is nowhere greater than twenty-nine fathoms. So that on the theory of sub- sidence we should have to account for the sinking of the eastern face of Kelai, Filadn, and Baura to a depth of over seven hundred fathoms, while the western faces were formed during a subsidence of less than thirty fathoms. On the same theory the southern faces of Kelai, Warifuri, and the other lagoon reefs of the northern face of Tiladummati were formed during a subsidence of from twenty to thirty fathoms, while the northern faces subsided from two hundred and fifty to six hundred fathoms! Again, in the same inner basin of Tiladummati and Miladummadulu are a number of smaller lagoon reefs, atolls in every sense of the word, which on the theory of subsidence could only have been formed during a subsidence of about thirty fathoms. Some of them in the central part of the group have become islands ; others are still lagoon reefs, as will be seen by a glance at the charts. To the former category belong Muradu, Mahafai, Kumberidu, and nearly all the islands in the central part of Tiladummati and Miladum- madulu, such as Faidu, Madidu, Kandute, Rebudu, and a number of small and nameless islands in the eastern and southern part of Miladummadulu. To the latter series belong Goadu, Maruri, Wagaru with a lagoon of eight fathoms in depth, Kabafaro, Dureadu with a lagoon of eighteen fathoms in depth, Maddedu with a lagoon of eight fathoms, and a number of unnamed rings and diminutive lagoon reefs in the southern part of Miladummadulu. The conditions which have shaped the formation of the faros of the reefs and islands of Tiladummati and Miladummadulu are similar to those which TILADUMMATI-MILADUMMADULU PLATEAU. 81 have led to the formation of the faros, the inner and outer lagoon reefs of North and South Malosmadulu, of North and South Male, of Ari, of Nilandu, and other Maldive groups. They show in great detail that subsidence cannot account for the formation of atolls upon banks of such great differences in depth as those of the Maldives, and all within short distances of one another. It would require an irregularity and a range in the vertical oscillations of different parts of the Maldivian plateau entirely without precedent. The deepest soundings indicated on the chart for the Tiladummati-Miladummadulu plateau is thirty fathoms, but the greatest number of soundings range between twenty and twenty-four fathoms. Entering, as we did, Miladiunmadulu from the west, to the south of Digu Faro (PI. 2), one cannot fail to observe, looking either east or south, how little land is visible, how open and unobstructed is this stretch of Miladiunmadulu, and how exposed are its opposite shores, according to the season, to the full action of the prevailing winds. One gets quite the impression of being in the open sea, with a few islands looming up here and there on the horizon. They are, however, so far apart and so few and insignificant in size, that they afford little or no shelter from the monsoons to islands even at a small distance. A glance at the chart cannot fail to bring this out. Great stretches of the interior of the central and northern part of Miladiunmadulu are bare of islands, with here and there a diminu- tive patch. They give the impression of a great open tract of sea, — fully sustained when one steams through the area delineated on the chart. The most characteristic features of this plateau are the open nature of the interior waters and the number and great width and depth of the passages between the islands and faros on the outer faces of the group. On the central part of the east face is a wide opening nearly ten miles across; five passages are over four miles in width, four are nearly three miles wide, twelve are over two miles, and five are two, with at least fifteen or sixteen passes of a mile between the smaller islands and faros. On the west face there are nine passes of more than one and a half miles. In fact, the length of deep open water on the faces of the Tiladummati-Mila- dummadulu group far exceeds in length that of the islands, faros, and flats 6 82 THE COEAL REEFS OF THE MALDIVES. (Pis. 2, 3), a condition which can hardly be called that of an atoll. Such a state of things seems to indicate that coral reefs and atolls may begin at any point or knoll or ridge or range or summit or flat of a plateau, having the depth at which corals may begin to form, provided its waters have the necessary temperature and the' proper exposure to currents or trades. The surface of the Maldivian plateau consists of a host of such knolls, summits, and ridges of greater or less dimension, upon which have risen the complicated system of islands, faros, reef flats, and atolls with which we have become familiar during our exploration of the Maldives. To understand the coral reefs of the Maldives,, we must look upon each ring or faro or islet or island on a reef flat as a thing by itself, just as much entitled to be called an atoll as any of the smaller atolls found in the Pacific, — atolls which only incidentally play a secondary part as reef flats of the outer faces of the groups, but which in spite of their position do not hold to the enclosed water the same relation which the reef flats or land rim of an atoll hold to the enclosed lagoon. Nowhere in the Maldives have we met with better examples of the modi- fying effect of the regular southwest and northeast monsoons as compared to the effect of winds in regions of southerly variable winds and squalls than in the northernmost groups of the Maldives. One can readily trace the direct action of the monsoons upon similar islands when situated either on the east or west face or on the inner waters of the group. Of course all that part of the Tiladummati and Miladummadulu groups which lies north of North Malosmadulu cannot fail to be greatly affected by the action of the south- west monsoon on the western face of the group. 1 Entering Tiladummati to the south of Muradu, we had to stem a current sweeping west with great violence, it being the season of the northeast monsoon. It is interesting to note the changes which have taken place in the outlines of many of the islands on the rim reef flats of both the western and eastern faros of the group since 1836. Some of the islands which were isolated at that time and are represented on the charts as single distant islands on opposite extremities of the rim of a faro have either both greatly 1 Gardiner, loc. cit , p. 315, has called attention to the small importance of the agency even of heavy gales on the edges of the reefs. MILADUMMADULU. 83 increased in size so as nearly to become united, or they form only a single crescent-shaped island on the outer edge of the rim reef flat. The further change these crescentic islands undergo from the action of the monsoons are often very striking. In fact, the crescent-shaped islands on the east face of the group almost seem to move in the direction of the prevailing monsoon, like great horseshoe-shaped dunes. Only their motion is limited to the extension of the horns of the crescent in the direction of the wind until the horns have locked, and the ring, once a crescent-shaped island partly enclosing a small lagoon, ends in completely surrounding the lagoon and forming the typical and mythical atoll so often described by writers on geography, but which really is only found very rarely in the great atoll regions of the Pacific, but finds its greatest development in the thousand isles of the Maldives. The islands of the interior of the plateau are wasting; the inner banks appear to be all of the same height ; and from the bridge we could see no indications, from the discoloration of the water, of banks growing up or having come near the surface. The crescentic islands on the eastern face of the northern part of Tiladummati show this process of growth admirabty, and there we can readily trace the development of a large single island from two smaller islands placed at each extremity of the atoll. We need only look at the charts and follow the gradual junction of the two islands of Hanimadu now connected by a low narrow sand spit, or those of Filadu or of Baura and Kelai or Noliwang Faro, and pass to such an island as Kuludu Faro, which occupies fully one half of the faro flat, to Komangdu and finally to Nuriwari, each occupying a gradually greater part of the reef flat rim till the island occupies the whole, and appears as a circular or elliptical island steep to, surrounded by an insignificant fringing reef flat ! Miladummadulu. Plates 1, 2,3; Sn, figs. 4, 5, 7 ; 34, fig. 2 ; 35-44 > 79, fig. 3. The southern part of Karema, the southernmost island on the south face of Miladummadulu, is wasting away; many of the trees and bushes are well 84 THE CORAL REEFS OF THE MALDIVES. out on the small reef flat of the southern end of the island. 1 Karema, like most of the islands on the outer faces of Miladummadulu, is steep to; it is bordered by coral sand beaches. On Wataru (PI. 34, fig. 2; 35, fig. 1), to the northeast of Karema, a large island occupies the southern rim of the faro. The vegetation of the island is poor; most of it is dead or dying, the sand and shingle of the high steep beaches having been blown far in between the trees through the outer belt of bushes. The southwest mon- soon must strike the south face of Miladummadulu with considerable force ; the tops of many of the cocoanut-trees are blown off, leaving nothing but the trunks. The lagoon of Wataru is circular and occupies the northern part of the faro ; it is surrounded by comparatively wide rims and has a greatest depth of three fathoms. On Wataru flats as well as on the southern flat rims of the east face of Edu Faro are extensive violet and dark green patches of corals. The south face of the southern island of Edu Faro is bordered by steep shingle beaches, and a belt of boulders and shingle extends along the outer edge of the reef flats. Edu Faro (PL 35, fig. 2) is steep to, with an extremely narrow reef flat on the sea face. The long lagoon to the west of the wide eastern rim is shallow ; its greatest depth is only three fathoms ; it is flanked on the west by a moderately wide reef flat rim. The two islands on the eastern face of Edu Faro (PL 35, fig. 2) must once have been separated by a gap of greater width than the one now existing. From the opposing extremities of the islands are forming extensive sand spits and sand flats broken up into a number of bars and islands and islets, covered with more or less vegetation. These islets, bars, and spits, as is more clearly seen in more northerly islands of this group, will eventually unite into one large island. Denduni and a nameless island to the eastward with a wide reef flat on the northern face present no peculiar features ; they both are steep to, and have steep coral sand beaches with an outer belt of low bushes and rather scanty vegetation. On Huludu, a small island nearer the western face of 1 Darwin, loc. cit,, p. 103, mentions that one of the Maldive reefs " which within a few years existed as an islet, bearing cocoanut-trees, was found by Lieutenant Prentice, entirely covered with live coral* and Madrepore." The natives believe that the islet was washed away. Darwin, however, attributes the change to subsidence. MILADUMMADULU. 85 Miladummadulu, the vegetation is again quite flourishing; some large forest trees occur on the western face. Huludu is also steep to and flanked by steep coral sand beaches. Similarly Lofara and Watu, small islands to the south of which we passed, have steep coral sand beaches, are steep to. and covered with scanty vegetation. All the islands of the southern part of Miladummadulu, with their high and steep coral sand beaches, enclosing an inner area somewhat lower than the summit of the beaches, give one the impression of being slightly dished. Mavila, on the west face of Miladummadulu, is a triangular island with a reef flat on the northwestern face. It is steep to, surrounded with steep coral sand beaches, and fairly well wooded ; some of the trees on the eastern point are of considerable size. About six miles to the north on the west face of Miladummadulu, and separated from it by a circular faro of fully two miles in diameter, with a lagoon of a greatest depth of eight fathoms, rises Digu Faro, an elliptical faro pointed to the south. The small islet with its clump of trees indicated on the chart as existing on the southern spit of the faro has disappeared ; nothing can be seen but a diminutive sand bank on which a few terns were collected. Delia Faro consists of two small islands on opposite faces of an irregular elliptical reef flat covered with low bushes ; they are bounded by coral sand beaches ; the eastern island is wasting away ; the outer belt of bushes runs close to the water's edge. Maddedu (PI. 36, fig. 1), in the very centre of the southern part of the group, is a triangular faro with a reef flat occupying the eastern part and an island at its eastern spit ; while the western half is a small lagoon, bounded by a sunken narrow rim flat, having a greatest depth of eight fathoms. The island is covered with low vegetation, much of it in a dying condition ; the eastern face is wasting ; the outer edge of the reef flat is bordered by a belt of boulders ; a beach of coarse coral shingle flanks the east face. To the southeastward of our track we passed by Tolandu, a small steep-to island on the east face of the group, with a scanty outer belt of low bushes surrounding the cocoanuts of the interior of the island. The western face is a long steep coral sand beach ; the base of the coarse 86 THE CORAL REEFS OF THE MALDIVES. coral shingle beach of the northeast face is edged by a belt of boulders ; as far as we could judge from the charts, and from what we saw of the small islands to the north and west of Ma Faro, they present no special features of interest. The deep pass with a depth of ten fathoms leading into the lagoon of Ma Faro, with its islets in the reef flats in addition to its larger island flanking the east face, gives to that faro all the appearance of a small atoll. 1 Kendikolu (PI. 37) is one of the largest reef flats of the southern part of the east face of Miladummadulu, the greater part of which is occupied by the island of Kendikolu, one of the largest in the Maldives. It is steep to on the east face ; the western face is occupied by a comparatively narrow reef flat somewhat wider at the northern extremity. The southern point of Kendikolu is bordered by steep shingle beaches, while all the way to the north extends a magnificent steep coral sand beach. Corals grow in great profusion on the slope off the western face of Kendikolu. The island is noted for the long sinks which occupy the greater part of the central portion of the island. We visited the northern one (PI. 38); it is some- what nearer the eastern face of the island, separated from it only by a narrow ridge of shingle and of sand. The depth of this brackish pool is only about eighteen inches ; its shores are overgrown with mangroves. There are many Pandanus, Hibiscus, Breadfruit, and other forest trees on Kendikolu; the vegetation is quite dense. The islands on the eastern face of Miladummadulu to the north of Ken- dikolu, as far as Nalandu (Pis. 2, 3), proved most interesting. In general they are small crescentic islands occupying a greater or less area of the edge of the rim reef flat of the faro encircling in part a small lagoon. From this stage we find all possible conditions of development from a closed ring surrounding a diminutive lagoon or only a sink in an island where all trace of its mode of formation has disappeared. Komandu (PI. 36, fig. 2) is a circular island, steep to, with large boulders on the south beach ; a coarse shingle beach extends round the eastern face, and a steep sand beach flanks the west side. A small reef flat covered with shingle extends off the north point. The vegetation of both Komandu and 1 Ma Faro is three and a half miles long, Gardiner, loc. cit., p. 395. MILADUMMADULU. 87 of Breyfasdu (PI. 39, fig. 1) is tall and dense; the latter has, like the former, a stretch of large boulders on the southeast face. 1 Both islands are wasting away. In Bomasdu (PI. 39, fig. 2) the lagoon has become entirely silted up; nothing is left of it except a very shallow sink almost filled with patches of sand, which cross it in every direction. Tbe sand beaches of the sea face are steep ; the outer edge of the reef rim flat is crowded with large patches of flourishing corals. Bodu Mandu (PI. 40, fig. 1), also on the east face of Miladummadulu, is steep to ; the greater part of the lagoon has been filled up to the general level of the rim reef flat ; the lagoon is shut off from the sea by a steep shingle beach which now joins the two horns of the crescent- shaped island and encloses a lagoon with a depth of five fathoms. In 1836 this lagoon was, according to the chart, connected with the sea across the narrow still slightly submerged rim reef flat, but it is now closed. In all the crescent- shaped or linear or other islands on the reef flats of the faros of the east face of Miladummadulu or of other groups which have increased in size and been formed by the coalescence of smaller islands, it is always possible to detect, by the difference in the character of the vegetation, those parts of the land rim which are of more recent origin and have grown up on the connecting spits or horns. No atoll we have seen shows in a simpler and more convincing manner than Bodu Mandu how an enclosed lagoon of considerable depth, completely shut off from the sea, has been formed by the throwing up of a shingle dam connecting the horns of the crescent- shaped island and enclosing a part of the old rim flat and central area of the faro. In Kuda Mandu (PI. 40, fig. 2), the next atoll to the north, the lagoon is not quite shut off, it is only partly closed by a bar of shingle thrown up on the shallow rim of the reef flat. There is to the eastward of the bar a boat passage giving free access to the sea at all stages of the tide. Another interesting crescent-shaped atoll is Dureadu (Pis. 41, 42, fig. 1), situated in the centre of the narrowest part of Miladummadulu. It is a faro 1 Hainbuddu I did not visit. According to Gardiner, loc. cit., p. 394, it is unique among the Maldives in having a definite reef flat on all sides except the west. 88 THE CORAL EEEFS OF THE MALDIVES. or atoll with a wide-rimmed reef flat that has only reached the surface, close to the horns of the island. The rest of the rim flat is submerged, leaving the lagoon open for the whole width of the faro, with as much as five fathoms of water on parts of the rim. The principal island of Dureadu is clothed with luxuriant vegetation ; at the extremity of the eastern horn rises a small island covered with large bushes and separated from the principal island by a line of large coral boulders. The east face of the islet and of the principal island is flanked by a shingle beach with stretches of small coral boulders which extend also on the northern face of the main island of the atoll. The lagoon of Dureadu is one of the deepest of the small independent atolls in the Maldives ; it is fully a mile across with a depth of eighteen fathoms. Dureadu resembles Nalandu, though the lagoon of the latter is much smaller and its greatest depth is not more than two fathoms. 1 All the islands we passed on the east face of Miladummadulu are steep to ; there are no reef flats or spits extending from the flats on the lee side, as in the islands of many of the other groups. The original flats must have been limited to the areas now occupied by the islands or faros. Ekasdu (Pis. 42, fig. 2 ; 43) is perhaps the most advanced stage of an enclosed central lagoon. It could be well seen from aloft. The outer beaches are all steep coarse shingle slopes. The lagoon of Ekasdu is of a dark blue color, indicating a depth fully as great as that of the lagoon of Bodu Mandu with a greatest depth of five fathoms. Ereadu, the next island to the north, though less than a mile in diameter, is the largest of the chain of small islands between Kuludu and Furnadu on the east face of Miladummadulu. The crescent-shaped island, open to the west, encloses within its horns an elliptical lagoon, shallow at its southern extremity with a wide shallow western rim over which the sea flows freely into the lagoon ; parts of the rim are barely awash. Ereadu represents one of the earliest stages in the closing off of a lagoon where no dam or shingle heap has as yet been thrown up on the reef flat between the horns. 1 Of the central islands it was only at Dureadu that Gardiner found on the northwest face the exist- ence of a definite but very limited reef flat with a well-formed fissure zone of gradual slope to the general level, loc. cit., p. 388. MILADUMMADULU. 89 The reef flat enclosing the islands of Furnadu (PL 44, fig. 1) and Faru- kolu 1 is nearly two miles wide and four in length ; the islands are separated by a wide reef flat with islets and sand banks thrown up on the outer edge of the eastern face. At the extremities of the central part of the reef fiat a small shallow lagoon has been formed. The western face of the reef flat is covered with patches of flourishing corals. The two islands are hook-shaped ; the narrow deep bays formed by the extension of the hooks or spits to the north and south enclose a shallow part of the reef flats, indicating clearly how the shallow sinks of such islands as Kuludu and Kendikolu are formed. Such sinks or lagoons or rather shallow enclosed bays of parts of the reef flats must not be confounded with the enclosed lagoons of faros of such islands as Nalanclu, Milandu, and the small crescent-shaped islands to the south or the larger crescent-shaped islands on the northeastern horn of Tiladummati. Sinks or enclosed bays like those of Kuludu exist on Makandudu and to the north in Kuludu Faro, and the two extremities of Filadu. In the northern faros, as in Kuludu, these sinks occur on faros or atolls where the lagoons are well developed, plainly showing the distinct differences in the mode of formation of the two. The northwestern point of Makandudu is flanked by a large pile of coral boulders and beach rock slabs on the east face, exposed to the northeast monsoon ; the island is bounded by steep coral shingle beaches. The island is steep to with a small shallow sink in the central part of the island. The southern face of Milandu is wasting away; off the beach great clumps of trees and bushes are standing on the reef flat. The coral shingle has been driven by the southwest monsoon far in between the clumps of trees and bushes. Milandu is a narrow crescent-shaped island, somewhat dumb-bell-shaped ; it lies on the outer edge of the eastern reef flat which forms the eastern rim of the shallow lagoon enclosed between it and the western rim flat. On the chart the greatest depth of the lagoon is marked as two fathoms ; its light green color would not indicate even that depth. The western rim flat connecting the horns of the island is quite shallow, the sea breaking over the greater part of it. The horns of the island have extended westward somewhat beyond the position indicated 1 Gardiner, he. cit., p. 391, fig. 99. 90 THE CORAL REEFS OF THE MALDIVES. on the chart, the opening into the lagoon over the western rim being much narrower than when it was surveyed in 1836. The beaches of Milandu are all steep, of coral shingle, even the inner beaches flanking the eastern face of the lagoon. The dumb-bell shape of the crescent-shaped Milandu Island indicates that at one time there were two islands on the extremities of the reef flats of the faro ; the spits making both ways and extending north and south from them, they gradually joined and finally formed a single crescentic island. The structure of Nalandu (PI. 44, fig. 2), the northernmost of the crescent-shaped islands of Miladummadulu, throws considerable light on the mode of formation of atolls which, like Kuludu, Filadu, and others, have both a lagoon and a sink, or enclosed bay, once forming a part of the outer rim reef flat. Nalandu is a small island, less than a mile in diameter, with a lagoon two fathoms deep and a pass open to the south ; it also has a closed sink or bay, one fathom in depth, communicating with the lagoon and con- siderably larger than the lagoon itself. At low tide a great part of the sink is bare, the bottom being fine coral ooze ; it is an exceedingly pretty sheet of water surrounded on all sides by clumps of large forest trees, with bays and bights reaching out from all faces. The sink is separated from the eastern part of the lagoon by a low sand point covered with a thick tangle of bushes and trees ; a wide rim covered by tall forest trees and man- groves divides it from the sea on the eastern and western sides ; it is flanked on the north by a narrow rim flat. On the outer face Nalandu is surrounded by steep coral shingle beaches ; off the western face of the island lies a small reef flat awash ; on its outer edge a belt of boulders and heaps of shingle have been thrown up; they pass into the boulder and shingle spits of the horns of the island. The southeastern face of the island, as well as the western point, is wasting away. To the west of Ereadu begins an inner line of small, steep to islands, with the exception of Kabafaro, which is crescentrshaped and flanked by a small reef flat. This inner line of islands runs northerly from Ereadu to Fivaku, nearly parallel to the eastern face of Miladummadulu, at a distance of from two to three miles. They have no distinguishing features from similar inner or outer islands of other groups in the Maldives. MILADUMMADULU. 91 We passed at a short distance off Kuraidu, the northernmost of this inner chain of islands ; its southern face is flanked by steep, coarse shingle beaches, with a mass of boulders along the southwestern point. The island is steep to; its vegetation is scanty. Faidu and Madidu, which are more in the central part of the group, are steep to, and both have, according to the chart, small central sinks. They are, like Kuraidu, surrounded by steep, coarse shingle beaches. The vegetation of both is poor ; it seems to be affected by the action of the southwest monsoon. Looking to the northeast, while off Madidu, we are facing the great central gap of Miladummadulu, between Fivaku and Mavaidu, nearly fifteen miles wide, with only Nu and Numara on the horizon across this great and deep passage. On the western face of Miladummadulu, to the south of Bilifuri, are a number of large faros and reef flats and islands, similar in all respects, according to the charts, to those we examined, as well as a few small islands scattered towards the centre of Miladummadulu to the north of Dureadu, similar to the inner line of small islands near the east face of the group. Kofenbe, to the south of our track, is a large faro over two miles in diameter, with a lagoon of a greatest depth of three fathoms, and a wide rim flat. A large triangular island occupies the greater part of the east- ern rim. The vegetation consists of low bushes, though on the eastern face there are a few clumps of large trees. The island is flanked by steep coarse coral shingle beaches. The island of Bilifuri occupies a good part of the eastern rim flat of the faro. The southern extremity of the island seems to be extending west- ward as a sand spit forming the northern face of a small bay surrounded with large forest trees. The southern spit itself is flanked by a coarse coral shingle beach. The island is covered by fine trees and a thick outer belt of bushes and smaller trees. Along the eastern face of the island the steep shingle beaches are of still coarser material, with an outer belt of coral boulders as large, if not larger, than any we have seen thus far in the Maldives. Close to the northern point of the island the narrow entrance to the lagoon is still plainly visible, and as we swept past it we obtained an 92 THE CORAL REEFS OF THE MALDIVES. excellent view of the western shore of the island and of the part of the lagoon enclosed by the wide rim of the reef flat on the western face of the faro. The lagoon of Bilif uri is quite shallow ; its depth is not more than one and a half to two fathoms ; its light-green color scarcely marking the outline of the lagoon, blended as it is with the greenish tints of the wide rim enclosing it, and covered with a few feet of water ; on the northern face of the rim flat the sea breaks along the outer edge and all the way across. Beyond the diminutive pass into the lagoon the outer edge of the rim reef flat is flanked by a belt of heaps of shingle and of small boulders. Magnificent coral patches dot the outer rim of Bilifuri as well as of the eastern rim flat of the large faro to the north of it on the western face of Miladummadulu. This large faro is over four miles in length, pear-shaped ; the larger of the enclosed lagoons is open at the northern extremity of the faro ; near the southern face a smaller circular lagoon less than half a mile in diameter is enclosed by the wide crescentic reef flat rim of the faro. The narrow sand bank on the eastern face of the faro is scantily covered with trees and an outer belt of small bushes. The beaches of the narrow islet of the faro to the north of Bilifuri are flanked with beach rock, and near the northern point with a few boulders at the base of the steep, coarse shingle beaches. Much of the beach rock and some of the boulder reaches are undercut, pitted, honeycombed, and greatly weathered. Stretches of this weathered coral boulder belt extend along the northeastern face of the faro ; they indicate a former slightly greater elevation. As we passed out of Miladummadulu west of Goadu, we steamed by the long narrow channel leading into the lagoon at the tip of the northwest horn of the faro to the west of Goadu, a mere strip of lighkblue water flanked on both sides by a light-greenish flat rim. Kandute and Goadu are the two northernmost islands of the west face of Miladummadulu. To the north of them a large elliptical faro nearly five miles in length, with a spur off its eastern face, marks the artificial northern boundary drawn east to west between Tiladummati and Miladummadulu. Kandute and Goadu are both flanked with sandy beaches along the eastern part of the faces of the islands, while steep shingle beaches surround the TILADUMMATI. 93 western extremities. At the western extremity of Kandute a small de- tached island rises upon the small reef flat; it is flanked with beach rock and coral boulder masses, greatly weathered. That part of the island is exposed to the full force of the southwest monsoon. Natives of Kuludu. Tiladummati. Plates 1,2 ; 8 a, jigs. 2, S; 45-53. Immediately north of the artificial boundary drawn between Miladum- madulu and Tiladummati, on the west face, lies a bank with irregular soundings varying from three to fourteen fathoms (PI. 2). Next, to the north, comes Muradu (Pis. 45 ; 46, fig. 1), a circular faro of about two miles in diameter with a wide rim reef flat. On the chart of 1836 a long narrow island is drawn upon the eastern face of the rim flat, and a smaller island on the inner edge of the western rim (PI. 2). We were surprised to find that these two islands are now connected, forming a single crescent-shaped island, a long sand spit having extended from each somewhat diagonally across the inner face of the northern rim. They are connected by a high coral sand ridge partly covered with low bushes and a scanty vege- tation near the old islands, and bare on the more recent central part. The 94 THE CORAL REEFS OF THE MALDIVES. original islands form knobs at the extremities of the single somewhat dumb- bell-shaped island. The outer edge of the southern reef rim flat is bordered by small boulders, and the eastern face of the island is flanked by shingle beaches and boulders. The outer belt of the rim of the large pear-shaped faro to the north of Muradu is, with the exception of a belt of large boulders on the northern face of the ring, covered with patches of flourishing corals. The northern rim reef flat of the faro slopes into deeper water than the shallower southern rim. There is no well-marked pass into the central lagoon. The next faro to the north is elliptical, with a wide opening leading from the east into the central lagoon ; this is bounded by a wide reef flat rim on the southern face. On the horn of the southern rim flat a few large boulders stand as sentinels at the entrance, with heaps of shingle on the northern rim ; the opening of the pass is flanked by isolated boulders, and to the westward, on the northern face, the faro is bordered by a long reach of large boulders. A small sand bank has been thrown up on the inner face of the northern rim near the pass ; this is not shown on the chart. Muadu is a faro to the northeast of the last, dividing the eastern entrance of the wide and deep pass between it and Naguri into a northern and southern channel. The eastern end of Muadu is occupied by a long triangular island, which has thrown out along the northern rim a sand spit covered with low bushes. A similar coral sand spit extends along the southern face of the faro enclosing in part a wide bay; these spits enclose the greater part of the lagoon of the faro. From the northwestern corner of the island a coral sand spit forms the western boundary of a wide bay open to the north. The southern face of the faro is bordered by a belt of boulders and the eastern face of the island is flanked by steep, coarse coral shingle beaches. Naguri (Pis. 46, fig. 2; 47) is a large faro nearly three miles in diameter with a large island extending towards the central part of the lagoon, club- shaped at the western extremity and T-shaped at the eastern. The south- ern rim flat is flanked by a low wall of shingle and large angular undercut boulders, all greatly weathered and separated from the beaches of the island by a shallow reef flat. The club-shaped and the T-shaped parts of the T1LADUMMATI. 95 island were evidently once distinct islands, the one on the outer edge of the eastern rim flat, the other extending across the wide southern rim of the faro. They have become united by a high sand bank, the extension of the shank of the T-shaped island. The island of Muheri, about two miles east of Naguri, represents the type of islands characteristic of the inner waters of Tiladummati. They are all steep to with high beaches, generally shingle beaches on the southwest faces and coral sand beaches on the northern ones ; according to the trend of the islands and their greater or less exposure to the prevailing monsoons, they have wider or narrower reef flats, with accumulations of small boulders at the terminal spits. Kurimbe Island in the central part of Tiladummati is about a mile in length, steep to, with high and steep shingle beaches on the northern as well as on the southern face, they are exposed to the southwest and the northeast monsoon, both of which have a sweep of considerable extent over that part of the central area of Tiladummati. The shingle beaches alternate with sandy reaches, according to the position of the shore with reference to the prevailing winds. The summit of the sand beach of the central part of the southern face of Kurimbe is fully fifteen feet high, and at other points it is from eleven to twelve feet high. A coarse, high, and steep shingle beach flanks the southeastern point of Kurimbe, and huge boulders are scattered along the edge of the narrow reef flats of the point. Kurimbe is surrounded by an outer belt of large bushes and smaller trees, enclosing several fine clumps of large forest trees. On the northern islands of Tiladummati and at Makunudu we find many Pandanus ; they are not as common in the central and south- ern parts of the Maldives. There are some large forest trees on the northern points of Kurimbe and Kumberidu (PI. 48, fig. 2). On both these islands as well as on the northern islands of Tiladummati, both the shingle and sand beaches are much higher and steeper than in any of the southern groups. Kurimbe, Kumberidu, as well as Mahafai, Muradu, two inner islands near the northern part of Tiladummati, and a few other islands in the central part of Tiladummati which we did not visit, belong to the same 96 THE CORAL REEFS OF THE MALDIVES. type as Muheri. We examined Mahafai and Muradu (PL 48, fig. 1), two of the inner islands near the northeastern extremity of Tiladummati. They are both steep to on the eastern face and flanked by steep shingle beaches on the east face with steep coral sand beaches along the western side (PL 49, fig. 2) ; on the western face of Mahafai there is a narrow reef flat. We examined the faros on the east and north faces of Tiladummati, north of Nuriwari. Those on the east face all have islands on the eastern rims of the faros occupying a great part of the area of the eastern reef flat. The southern point of Nuriwari is covered by low vegetation, swept by the prevailing winds ; the island is flanked by high, steep, coral sand beaches driven far into the outer belt of bushes. At the northern point and on the eastern part of the north face, which is evidently wasting away, large coral boulders are found along the outer edge of the rim of the diminutive secondary lagoon at that extremity of the island. At the southern extremity of the crescentic-shaped island of Noliwang Faro, a few isolated islands and islets flanked with coarse shingle and covered with tall bushes rise on the reef flat rim ; the edge of the reef flat is bordered by a few boulders. The small islands form the eastern face of the entrance to the shallow lagoon extending off the west face of the island; the western face of the pass is indicated by an accumulation of coral shingle. The western rim of the lagoon is narrow with but few coral patches on the rim of the slope which falls off very gradually towards the west. The northern part of the island of Noliwang Faro is club- shaped, with a hook-like extension to the south, occupying the northern part of the western rim reef flat, partly enclosing a deep, open shallow bay. This northern part was clearly at one time an island distinct from the south- ern extremity ; it is now connected with it by a long coral sand ridge ; the vegetation of the connecting ridge clearly indicates that its formation is of more recent date than that of the two extremities. The eastern rim of Hanimadu, the next atoll to the north, is flanked by a convex dumb-bell-shaped island, the clubs of which are united by a ver} 7 narrow coral sand ridge ; this gradually reached north and south from the disconnected extremities of the original islands at the northern and southern TILADUMMATI. 97 horns of the atoll until the spits joined in the central part of the rim of the east face. The vegetation of the connecting ridge was undoubtedly at one time as scanty as that of the more recently formed connecting ridge be- tween the two islands of Muradu on the western face of Tiladummati. The western face of the northern part of Hanimadu is flanked by a coarse coral shingle beach which must have been thrown up on that face of the island before the formation of the ridge connecting the north and south islands of the atoll. The connecting ridge is very narrow, as one can see the eastern face of the atoll through the low belt of bushes growing upon it ; the western rim of the shallow lagoon in the southern part of the atoll is narrow; the wide reef flat rim to the north of the lagoon is shallow. Hanimadu is a narrow atoll about four miles in length with a greatest width of a mile. At our anchorage off Hanimadu we seemed to be surrounded by land, and well shut in by islands rising up in all directions ; in marked contrast to the open character of the plateau south of the boundary between Miladum- madulu and Tiladummati as seen from the western face looking towards the wide open gap between Mavaidu and Fivaku. Seen by moonlight with the moon high above the horizon, the rims of the faros present a most striking appearance ; they appear like great flats painted a dull white with the thin glaring line of coral sand-beach in the background topped with the dark line of vegetation. The flats stand out so prominently that one might almost be tempted to navigate between the islands at night. Baura (PI. 49, fig. 1), the next atoll to the north of Hanimadu, is a circular lagoon reef over two miles in diameter, surrounding a small central lagoon with a greatest depth of three fathoms. The southern half of the reef flat is occupied by a broad dumb-bell-shaped crescentic island open to the north. The land of the southern rim of Baura undoubtedly consisted at one time of two distinct islands situated the one on the eastern, the other on the western face of the reef flat. These islands have become united by a broad ridge in the same manner as we have described the junction of other crescentic islands on the east face of Tiladummati. Seen from the north across the wide rim, the central lagoon forms a deep bay between the two horns 7 98 THE CORAL REEFS OF THE MALDIVES. of Baura. On the sea face of Baura the beaches are covered by coarse coral shingle, and the outer edge of the reef flat is flanked by a belt of large boulders. The larger crescent-shaped islands on the eastern face of Tiladummati, such as Hanimadu, Baura, Filadu, Kelai, have, like Muradu and Naguri on the west face and others elsewhere in the Maldives, been formed by the coa- lescence of separate islands, gradually united into one by the extension of sand spits and bars. This is admirably shown both in Hanimadu and Filadu (PI. 50, fig. 1); in the latter the connecting ridge is a mere sand-dam flanked by shingle on the sea face, and so low and so narrow that, steaming along the sea face, one can look across far over the reef flat rim of the western face of the lagoon reef. The lagoon reef of Filadu was at one time a lagoon open to the north with a wide eastern reef flat rim, on the northern and southern extremities of which were situated small islands (much as at Muradu in 1835); these must have gradually increased in size and become united by a narrow sand ridge reaching little by little north or south from the existing terminal islands, — a connection at first made by sand-bars or by piles of shingles. The vegetation of the connecting ridge of Filadu consists of low bushes and is in marked contrast with that of its club-shaped extremities. The lagoon reef of Filadu is nothing but a small atoll, with a well-defined land rim, a broad reef flat, and an opening nearly a mile wide with three fathoms of water leading into the lagoon, the greatest depth of which is nine fathoms. A small islet of coarse coral shingle, not indicated on the chart, marks the horn of the reef flat rim on the western side of the pass into the lagoon. There is a secondary lagoon in each of the extremities of Filadu ; the northern one we could look into as we passed ; its color indicated a depth of about one fathom, probably formed like that of Kuludu. Filadu, Kelai (PI. 50, fig. 2), and Warifuri (PI. 51, fig. 2), the three lagoon reefs forming the northeastern horn of Tiladummati, are in reality three atolls of considerable size. Kelai is over four miles in length and nearly two miles wide, Filadu is about three miles in length and two miles in greatest width, while Warifuri is somewhat smaller. On the western face of Kelai, there is an entrance to the lagoon, with a TILADUMMATI. 99 depth of three fathoms, somewhat blocked by coral heads and sand banks. The northern side of the entrance is inarked by a low sandy islet at the western extremity of the narrow northern rim of the lagoon, the depth of which varies from four to eight fathoms. The western rim of the lagoon is narrow, and both the northern and southern rims are extensions, tapering westward, of the wide reef flat rim on the eastern face of the lagoon reef, the greater part of which is occupied by the double hook-shaped island of Kelai. The widest reef flat rim of the lagoon reef of Warifuri forms the north- western horn of the atoll, its southwestern rim is quite narrow as well as the southern rim reef flat, which is edged by a long narrow coral sand island covered with low vegetation, and from each extremity long sand-bars are extending northward along the western and eastern rims of the lagoon reef. The lagoon of Warifuri has a greatest depth of over seven fathoms. Two small narrow islands covered with low vegetation extend along a part of the northern rim of Warifuri. In the course of time these islands may reach south on the western reef and unite with the southern island so as to form a single island on the western face of the lagoon reef, as we find it to be the case with Gafuri and Miledu, two lagoon reefs to the west- ward on the northern face of Tiladummati, in both of which the islands are on the western rim flat, as are also usually the widest rim flats on the faros of the western face of Tiladummati. Dedu (PI. 51, fig. 1) occupies the eastern face of a triangular lagoon reef, the western rim of which is narrow ; at one time it was composed of two islands now united by a central connecting ridge, as can be seen from the character of the vegetation covering the ridge. The western part of the southern rim of Gafuri is bordered by a wide belt of boulders extending northward from Dunacoori (PI. 52, fig. 1), a small islet at the southern horn of the rim. The southern part of the lagoon is very shallow, though the charts indicate a depth of seven fathoms at the northern end of the lagoon. Gafuri is separated from an adjoining ring to the west by a very narrow deep channel, a mere line of dark-blue water of a depth of ten fathoms. The outer edge of the wide rim of the ring is crowded with large colored patches of corals extending nearly to the edge of the small lagoon of the ring. 100 THE CORAL REEFS OP THE MALDIVES. The eastern face of Miledu (PL 52, fig. 2) is flanked by coarse coral shingle beaches; the northern point is buried in large coral bcmlders. Miledu, Tukandu, and Marandu are all steep to, and characterized by a belt of large boulders on the eastern face. The finest boulders are on Tukandu (PI. 53), towards the southeastern extremity and on the eastern point of Marandu. The reaches of boulders are separated by sand or coarse shingle beaches. On the western sides the islands are bordered by high and steep sand beaches. Ihavandiffulu. Plates 1,2; 8 a, figs. 1, S ; 54, 55, fig. 1. Ihavandiffulu (Pis. 1, 2), the northernmost atoll of the Maldives, is irreg- ularly rectangular in shape, about thirteen miles in length, and seven in width ; its trend is from southeast in a northwesterly direction. It is separated from Tiladummati by Gallandu Channel, about two and a half miles in width at its narrowest point and with a depth of two hundred and fifty-one fathoms in the centre. A sharp ridge must connect Ihavandift'ulu and Tiladummati, as both to the east and west are found soundings of over six and seven hundred fathoms (PI. 1). The southern horn of Ihavandiffulu is occupied by Digufuri, an elon- gated, pear-shaped reef, with wide flats at the two extremities enclosing a shallow lagoon, occupying the central area of the reef flat. The outer rim of this flat is edged by a low wall of small coral boulders like that which fringes Makunudu. A few ill-defined sand-bars rise on the outer belt of the northern face of the. reef flat towards lhavandu Island. North of Digufuri the concave western face of Ihavandiffulu is bordered by a reef flat from one to two miles in width ; its northwestern horn sweeps eastwardly to form the northern boundary of the atoll. On the sea face of the western reef flat extends a long chain of narrow islands and islets set back somewhat from the outer edge of the reef flat (PI. 54, fig. 2). The western faces of these islands and islets are bordered with sand or coarse shingle beaches with short reaches of beach rock at the base of the beaches. A belt or low wall of small, weathered, angular boulders similar to that of IHAVANDIFFULU. 101 Digufuri extends away to the northward on the outer rim of the western reef flat. The vegetation on the islands of the western reef flat of Ihavan- diffulu is meagre, but those in the interior as well as on the northeastern face of the atoll are covered with a flourishing outer belt of bushes sur- rounding inner clumps of taller forest trees. A belt of small boulders skirts the edge of the narrow reef flat of the northern point of Ihavandu Island ; a steep, high, coarse coral shingle beach mixed with small boulders has at times been driven through the outer belt of bushes between the base of the trees to a height of more than twelve or thirteen feet. We examined Manafur as a type of the islands in the central part of Ihavandiffulu : they are all steep to, with small reef flats. The eastern half of the northern face of Ihavandiffulu is open, with a single small lagoon reef island (Wagaru) near the centre of the northern side. On the northeastern face of the atoll are three small islands, steep to with small reef flats, and towards the northeastern horn a larger island, Uleguma, also steep to, and over two miles in length. The southern part of the northeast face ends in a long lagoon reef flat, at the extremities of which are two islands of considerable size, Muladu and Gumati ; the gap between them (PI. 54, fig. 1) is apparently closing, as the low vegetation of the inner extremities clearly indicates their recent origin. The long sand spit forming the northern extension of Muladu with its low vegetation and ending with small distinct tufts of bushes indicates clearly the extent of the recent addi- tion to the island. The corresponding spit of the southern part of Gumati is much shorter. The small reef flat forming the northern horn of Gumati is nearly covered by small boulders. The southern point of Beramundu is formed by a high, steep sand beach, while its northern point (PI. 55, fig. 1) and the southern point of Muladu are high, steep, coarse coral shingle beaches, with heaps of boulders at the base. The passes on the northeast face of Ihavandiffulu are wide and deep; those of the northern face are as wide, but of a less depth. The western pass through which we entered the atoll is more than sixteen fathoms in depth and fully a mile and a quarter Avide. On the southeast face, for a 102 THE CORAL REEFS OF THE MALDIVES. distance of over six miles, the small island of Galandu and a few coral heads alone bar the free circulation from the eastward. We passed out of Iha- vandiffulu atoll through the deep pass to the south of Murdu. The deepest part of the lagoon of Ihavandiffulu is thirty-one fathoms ; the majority of the soundings are between twenty and twenty-six fathoms, and in the northwestern part of the atoll they are considerably less, from twelve to sixteen fathoms. Toddu. Plates 1. 4- Toddu is a small steep-to island to the north of Rasdu Atoll, about a mile and a half in length by a mile in width, rising on the western edge of the plateau extending to the west of North Male, and which unites Toddu, Rasdu, and Ari. The greatest depth between Toddu and Rasdu is one hundred and fifty fathoms. There are no soundings to the north of Toddu. The island of Toddu is surrounded by a narrow reef flat somewhat wider on the western face ; it resembles in every respect the larger steep-to islands of the interior or outer faces of the larger Maldivian groups. We did not visit either Toddu or Rasdu. Rasdu. Plates 1, 4- Rasdu Atoll resembles in a general way Wataru Reef ; it is irregularly circular in outline, somewhat more than four miles in diameter. The western face is formed by a continuous reef flat. At each extremity of the eastern face of the atoll is found a faro and two small islands. The south- ern pass is divided by an islet; its western side is occupied by a small island. The greatest depth of the atoll is twenty fathoms. The lagoon is studded with coral heads, and a belt of rings and banks extends diag- onally across it in a northwesterly direction from the southern pass. Rasdu is separated from the northern part of Ari by a channel four and a half miles wide with a greatest depth of one hundred and forty fathoms. AE1. 103 Ari. Plates 1, 4 j 8 b, Jigs. 12, H : 8 c, fig. 25 ; 55, fig. 2 ; 56, 57, 58, fig. 2. No island group in the Maldives has the characteristic features of the archipelago so well marked as Ari. We may call it a great agglomeration of banks and faros over an elliptical area fifty miles in length and fifteen in width (PI. 4). The western face of the group is flanked by comparatively few faros, some of considerable size, over five miles in length. The eastern face, on the contrary, is bounded by a great number of small faros and banks, and towards the north the northeast face of Ari is quite open. Within this great area are dotted nearly two hundred lagoon reefs (faros) and banks, many of them, especially in the northern half, over two miles in length. A number of the lagoon reefs are connected with islands. Over twenty islands and islets on banks varying in size from a few yards in length to more than two miles are scattered irregularly through the central part of Ari. There are but few islands on the faros of the west face, while on the east face there are islands on nearly every lagoon reef or bank. Some of the islands are a mile in length. The greatest depth of Ari is forty-three fathoms. The majority of the soundings within the group averages thirty fathoms. Many of the faros on the east face of Ari are most irregular in outline ; they all tail westward, 1 and are separated by deep channels with from twenty to thirty fathoms in the centre of the passes. Shingle or large coral boulders flank the eastern spits of many of the faros, and the surface of the rims of the lagoon reefs is covered with patches of flourishing corals. To the north of Midu a pear-shaped faro extends westward for nearly two miles, with two islands on the eastern and southern part of the rim, which encloses a large lagoon with a depth of seven fathoms. Midu itself is a large island occupying nearly the whole of the reef flat upon which it has arisen. 1 The trend and shape of the islands on the east faces of the northern groups show in general the direction in which the sand is driven from the sea faces to the westward, both by currents and by the prevailing winds. The same effect can be traced on the faros and islands of the interior of the larger basins and on their western or other faces. 104 THE CORAL REEFS OF THE MALDIVES. As we steamed along the eastern face of Ari, we could see that the eastern part of Ari was filled with numerous diminutive banks and faros and coral heads which render its navigation dangerous (PI. 4). We entered Ari north of Digura by a pass (PI. 55, fig. 2) nearly two miles wide. To the north of the pass the faros are closely packed and irregular in shape, — dumb-bell, comma, circular, or crescentic in outline. Dugati, immediately north of the pass, is a crescent-shaped faro enclosing a lagoon of from five to seventeen fathoms and edged on the western face by a similar crescent made up of a line of small banks, the southern horn of which is flanked by a belt of corals awash. The southern face of Ari is flanked by two large reef flats separated by Ariadu, a large irregularly circular island with deep passes on either side. On the eastern part of the western reef flat are four islands, the largest of which is Mamigeli (PI. 56). Three lagoons are enclosed within the flat, the two largest on the western part of the reef flat. Digura Island is nearly two miles in length ; it flanks the eastern rim of a wide lagoon with from four to seven fathoms of water. The northern part of Digura Faro is more than two miles wide. On the western part of the same lagoon reef flat are Didu and two other islands. To the south of Digura Island a line of small islets extends towards Kurafuri ; the gaps between the islets are nearly filled by sand spits and bars which will eventually unite the two larger islands. On the outer face of Digura a long line of large black boulders crops up on the edge of the eastern rim flat; on its western slope a magnificent belt of corals extends from twelve fathoms to the surface ; they extend eastward along the edge of the rim flat far towards the lagoon slope. The edge of the western reef flat is quite irregular ; at a short distance from the outer edge huge coral knolls rise nearly to the surface from a depth of twelve to fifteen fathoms, form- ing a series of submarine buttresses more or less connected with the corals growing on the western face. At many points they have thus materially widened the area of the rim flat on the west of Digura lagoon. We found only a few patches of corals on the bottom or slopes of the rim of the lagoon; nowhere perhaps have we seen a finer example of the great development of corals on the so-called lagoon face of the land rim of an atoll, as at Digura. ARI. 105 Off the western face of Digura we found only fine and coarse coral sand, but none of the sticky ooze so characteristic of the bottom of the lagoons of the Pacific atolls. To the north of our anchorage off Didu we saw a large, most regularly elliptical faro fully a mile in length. Immediately to the west of Digura Pass are a number of large faros of various shapes, with very regular rim fiats and some with lagoons of considerable depth. Some of the banks we passed have no velus indicated on the chart; it is difficult to say that this is an error of the survey, or betokens a change, as is suggested by Mr. Gardiner for North and South Nilandu. 1 In one of the faros we find from six to seventeen fathoms, in another from four to ten ; another has a greatest depth of nine, and several have depths of five to six fathoms. "We did not check these soundings, yet we could see that these faros had undergone only slight changes since they were surveyed in 1834. In areas protected from the action of the monsoons a faro with a deep lagoon and well-submerged rim must change slowly; its rim increases very gradually in width and height, and its lagoon fills at a very insignificant rate. . The horn of the eastern face of the south pass into Ari, as well as of the western face of the pass, is edged with large boulders, and a mass of coral rubble is thrown up on the flats of Ariadu, on the southern face of the island. To the west of Mamigeli a number of sand-bars have been thrown up on the south side of the eastern, irregularly shaped lagoon of the faro. The northern rim of the lagoon is a mere narrow flat ; its slope and summit are covered with a magnificent growth of corals. The western islands of Mami- geli Faro are separated from the terminal lagoons by a wide green flat covered with from two to six feet of water. The northern horn of this faro is sharp, and is separated by a pass with not more than six fathoms from the faro to the north of it, the first link of the chain of large faros flanking the western face of Ari ; six of these are at least four miles in length ; they are quite irregular in outline. At the northwest angle of Ari is Matiwari, an irregularly shaped faro resembling Dugati, at the southeast angle of Ari, only the reef flat area is more extensive in the northern faro ; like the southern faro it seems 1 Loc. cit., p. 405. 106 THE CORAL REEFS OF THE MALDIVES. to have been formed by the coalescence of smaller faros or patches, once separated by comparatively shallow water. The depth of the passes be- tween the faros of the west face of Ari is not as great as on the eastern, where the average depth of the narrow passes is fully twenty-four fathoms. On the western face are several passes with not more than from six to fifteen fathoms. The lagoons of the faros of the west face of Ari are many of them quite deep ; ten, fourteen, fifteen, sixteen fathoms are the depths indicated on the charts. These faros are also characterized by the great width of the western rim flats of the lagoons. Throughout the southern part of Ari are a number of small and incipient faros in all stages of development. On the little island which divides the second southern pass on the west face of Ari, the small reef flat is covered with great slabs of beach rock. 1 About five miles to the southeast of Mandu Faro and about three miles and a half south of Furadu, is a large elliptical ring over two miles in length (PI. 57, fig. 1), with a fairly wooded island on the southern rim flat and a small sand bank on the western face of the rim, which is nearly all awash. The faros in this part of Ari have been somewhat modified since they were surveyed by Captain Moresby. On the rims of several we find islets and islands, or sand-bars and banks, which did not exist at the time of the survey. The changes in this part of the group may be due to the position of the faros exposed to the action of the northeast and specially the south- west monsoon in that area of Ari. The existence of the lines of boulders on the outer rims of the faros of the southern and western faces of Ari shows the extent of the action of the southwest monsoon on the faces of the faros exposed to them. About a mile and a half south of Furadu there is a circular faro (PI. 57, fig. 2) fully a mile in diameter with a lagoon evidently deepest at the eastern extremity, where it is of a dark blue color, and shallower, of a greenish-blue tint, towards the western end. The lagoon is surrounded by a wide greenish rim of uniform width, flanked with fine patches of corals on 1 Gardiner (loc. cit., p. 341) has given an excellent account of the formation of terraces of beach sandstone, off a sand beach ; sandstone beach rock is not necessarily, as is stated by him (loc. cit., p. 342), formed on a beach, to a large extent protected from the heavy ocean rollers, or the waves within a lagoon. Among other cases in the Pacific coral reefs (Mem. Mus. Comp. Zool., Vol. XXVIII.) where this is not the case, I might mention especially the beach rock on the east face of Nanuku Levu in Fiji (PI. 106, A. Agassiz, Bull. M. C. Z., XXXIII.). NORTH XILAXDU. 107 the outer edge ; the rim is awash on the northern face where sand is accumulating on the rim flat. Furadu (PI. 58, fig. 2), a small island covered entirely with low atoll vegetation, is flanked with steep sand beaches and a narrow reef flat. To the east and north of Furadu, a number of large rings are found, one of which is nearly four miles in length. On the south face of Ari a heavy swell coming from the southwest was breaking hard upon the outer edge of the reef flats. The outer face of the reef flats of the southern face of Ari was everywhere lined with fine patches of corals. The northern part of Ari is not well defined; for a stretch of nearly twelve miles it is open to the northeast ; this long stretch is only broken by one large faro, two much smaller ones, and a few diminutive banks, cropping up at about the thirty-fathom line. The northern parts of Ari and of North Malosmadulu have the open character of Tiladummati and Mila- clummadulu ; they illustrate the passage between the very open groups to those that are more closed, like North Male and the central groups of the Maldives. North Nilandu. Plates 1, 4. 5; S b, fi'g. 14 ; 59-61; 78, fi At ten in the morning we made a surface haul about halfway between Miladummadulu and Makunudu. The sea was smooth with a bright sun shining, and the water was swarming with life. We collected a great num- ber of Porpita, among them a few very diminutive ones, a colossal Cestum with bright yellow pigment at the extremity of the flappers, Globigerinae, Halobates, young Flying Fish, Diodon, Balistes, and Mollusks; Sergestes, masses of a large blue Copepod, large blue Shrimps, young of Physalia, many different species of floating fish eggs, Sagittae, Sargassum, and Thrichodesmium. We found a species of Amelia in great abundance in the northern part of Tiladummati, specially at our anchorage off Hanimadu, and in the northern Maldives. It is found as far south as Addu. Off Timarafuri Pass (Kolumadulu) we made three hauls with the tow net in fifty, one hundred, and one hundred and fifty fathoms. We obtained free swimming Amphioxus larvae, perhaps the young of some pelagic species. It seems to be the same larva as that collected by Gardiner in the south of Miladummadulu Atoll and in Fadiffolu Atoll. 1 The net also contained hosts of Hyalea, Atlanta, Criseis, Cuvieria, Sty- liola, Phyllosoma, Hyperia, three species of Siphonophores, two species of Medusae, huge Sagittae, Sapphirina, Sergestes, brilliantly colored Ma- crurans, Oxyrhynchus, large blue, red, and small light-brown Copepods, and numerous large Appendiculariae. About two miles west off Kandudu Pass (Kolumadulu), we made a pelagic haul in about seventy fathoms. We obtained masses of Copepods, Sagittae, Pyrosoma, Macruran and Brachiuran larvae, Sergestes, Collozoum, Aglaura, Rhopalonema; a remarkably large transparent Zoea with a huge carmine patch, Oxyrhynchus, transparent, Syllis-like Annelids, Diphyes, and two other genera of Siphonophores, Sapphirina, Atlanta, and Spinalis. 1 C. Foster Cooper Cephalochorda, p. 354, PI. XVIII., fig. 3, Part I. Vol. II. J. S. Gardiner, Fauna and Geog. of the Maldive and Laccadive Archipelagoes, June 15, 1903. THE PELAGIC FAUNA OF THE MALDIVES. 157 At anchor in Kolumadulu off Kandudu we obtained from the surface a number of large-sized Appendiculariae in their houses, Mnemiopsis and Aglaura, small Isopods, Sagittae, Collozoum, and many Crustacean larvae. At anchor in Felidu Atoll off Buri Hura, we obtained from the surface Sagittae, Copepods, a large species of Appendicularia, Crustacean larvae, both Macrurans and Brachiurans, a young Turris-like Medusa and Thricho- desmium. While crossing the Veimandu Channel (between Kolumadulu and Had- dummati) in a dead calm, we passed a great mimber of flying fish : they could, as a rule, rise but little ; the tail, or the pectorals and tail, both touched the water as they skimmed along. In the former case they left a line of a single series of rings, lapping more or less, formed by striking of the tail at regular intervals. In the second case the contact of the tail and fins formed three lines of rings which interfered in a great many ways, and in both cases left the delicate rings thus formed to disappear gradu- ally long after the flying fish had sunk below the surface. Track of Flying Fish (tail and fins). Tail Track. In the Veimandu Channel we also passed a great number of Eutimas and of Aurelias, the same species we have met with before. We made two pelagic hauls off the southern entrance to Haddummati in one hundred fathoms, and obtained many Sagittae, a few large red Copepods, Collozoum, Rhopalonema, and parts of Medusae, fragments of Diphyes, Sergestes ; numerous larvae of Crustaceans, both Macrurans and Brachiurans, Thrichodesmium, Tomopteris, and Alciopa. 158 THE CORAL REEFS OF THE MALDIVES. At our anchorage off Funadu (South Haddummati), just before sunset we found the water swarming with Sapphirinae ; when gathered in a jar they produced the most beautiful effects from the variety and intensity of the metallic colors reflected by them while swimming about. A large species of Eutima and a species of Berenice were also abundant at this locality. We found the Sapphirina and Eutima again swarming round the ship early in the morning. At our anchorage off Funadu (south part of Haddummati), we made a most successful surface haul. It was specially marked for the large number of Medusae it contained ; there were Stomobrachium, Ephyrae, Hybocodon, a Medusa allied to Gonionemus, Geryonia, Glossocodon, many Styliolas, a number of Halobates, young of Ianthina, fragments of Ctenophores, Diphyes, Veligers, Sea-urchin larvae, Sagittae and Copepods in great number. In a haul made later in the evening we collected only a few Copepods, Sagittae, Crustacean larvae, and Appendiculariae. In a surface haul off the south pass of Haddummati, we obtained a great number of Crustacean embryos, both Macrurans and Brachiurans, Sagittae, Salpae, Sapphirinae. Two species of Diphyes, Styliola, Spinalis, Appendi- cularia, and a few Copepods. Salpae were very abundant in the lagoon as well as a species of Aurelia, which we found everywhere in the Mal- dives and in the Indian Ocean, all the way from Ceylon to Aden. In another surface haul, at our anchorage, in the southern part of Haddum- mati we obtained Zygodactyla, numerous Halobates, very many Aureliae, and small Copepods shining like Sapphirina. We made a haul off Dandu (Suvadiva) and collected fine large trans- parent Macrurans, Sergestes, larvae of Crustaceans, among them a huge Megalops, masses of Copepods, large transparent Syllis-like Annelids, To- mopteris, with colossal lateral appendages ; Sagittae, Salpae, Beroe, Ocyroe, Cydippe, Mertensia, Bolina, Geryonia in all stages, two species of Diphyes, Tima, Dipurena, Cunina, Aglaura, Zygodactyla, Stomobrachium, Oceania, Circe, Ephyra, and Rhopalonema. We made a pelagic haul off the Gan Channel (Suvadiva), in twenty-five fathoms, with poor results. THE PELAGIC FAUNA OF THE MALDIVES. 159 We made a surface haul at our anchorage off Gadu (south part Suvadiva), and obtained Sagittae, Diphyes, young Fishes, Collozoum, Crustacean larvae, Salpae, and Thrichodesmium. In a surface haul at our anchorage off Gang (Addu), we obtained Sa- gittae, Sergestes, Copepods, Crustacean larvae, Collozoum. B. I. S. N. Co. S. "Amra. INDEX. Adago, 107, 109 Addu, 145, Pis. 1, 6; 8 c, figs. 21, 23; 76, 77, fig. 2 Absence of islands in lagoon of, 147 Bottom deposits of, 154. Connection with Chagos Archipelago, 6 Corals on lagoon slope of, 148 Gardiner on Addu, 147, 148 Growth of corals in, 148 High shingle beach at, 146 In Equatorial region, 145 Isolation of, 145 Lagoon encroaching on land of, 147 Position of islands of, 147 Resemblance of, to Pacific reefs, 145 Shape of, 146 Temperature of lagoon of, 150 Vegetation of, 146 Addu Ridge, 6 Agatti, 5 Aidu, 60, PI. 22, fig. 1 Akirifuri, 51 Bays of, 51 Vegetation of, 51 Alifuri, 71 Aligau, 73, PI. 33, fig. 2 Vegetation of, 73 Alimata, 117 Androth, 5 Anghenufuri, 57, 65 Lagoon of, now filled, 65 Anguretin, 71 Vegetation of, 71 Ari, 103, Pis. 1, 4; 86, figs. 12, 14; 8 c, fig. 25; 55, fig. 2; 56-58, fig. 2 Bottom deposits of, 153 Bottom of lagoon of, 104 Changes in islands and rings of, 104, 105, 107 Depth of lagoon of, 103 Maldivian characters of, 103 Northern part of. open, 107 Temperature of lagoon of, 149 Trend of faros of east face of, 103 Ariadu, 105, PI. 55, fig. 2 Atolls, number of, in MaldiveB, 1 Description of the, 35 Bassas de Pedro, 5 Baura, 80, 97, PI. 49, fig. 1 Distinct islands of, 97 Beramundu, 101, PI. 55, fig. 1 Bilifuri, 91, Coral patches on rim of, 92 Great size of faro of, 92 Small pass into lagoon of, 92 Bitra, 5 Bodu Faro, 66 Bodu Mandu, 87, PI. 40, fig. 1 Junction of islands of, 87 Lagoon of, nearly filled, 87 Boilers and faros, 42 Bomasdu, 87, PI. 39, fig. 2 Shallow sink of, 87 Steep beaches of, 87 Bottom Deposits of, 152 Addu, 154 Ari, 153 Fadiffolu, 152 Haddummati, 153 Kolumadulu, 153 Miladummadulu, 153 Mulaku, 153 North Male, 152 North Malosmadulu, 152 South Malosmadulu, 152 Suvadiva, 153 Tiladummati, 153 Breyfasdu, 87, PI. 39, fig. 1 Wasting of, 87 Bundusi, 47 Buri Hura, 116 Byramgore, 5 Cardamum, 5 Cherbaniani, 5 Chetlat, 5 Cooper, C. Foster, on Cephaloehorda, 156 Cora Divb, 6 Corals in lagoon of North Male, 39 Corals in lagoon of faros, 39 Dambidu, 132 11 162 INDEX. Dandu, 136, PI. 75 Darwin, Charles, on disseverment of the large Maldive atolls, 10 On South Maldives, 135 On subsidence of Maldives, 84 On the Maldives as a barrier reef of great dimensions, 11 Dedu, 99, PI. 51, fig. 1 Defili Faro, 56 Deha Faro, 86 Dehu, 75 Denduni, 84 Diaddu, 137 Didu, 105 Difuri, 36, 38 Digeli, 69 Digeri, 117 Digu Faro, 81, 85 Reef flat of, 86 Digura (Ari), 104 Digura (Suvadiva), 141 Dina Faro, 67 Gardiner on velu of, 67 Diyageli, 123, PI. 66 Consolidation of islands at, 123 Dugati, 105 Dunacoori, 99, PI. 52, fig. 1 Dunikolu, 62 Dureadu, 87, Pis. 41, 42, fig, 1 Great depth of lagoon of, 88 Vegetation of, 88 Edu Faro, 79, 84, PI. 35, fig. 2 Corals on flats of, 84 Gap separating islands of, 84 Ekasdu, 88, PI. 42, fig. 2 Central lagoon of, 88 Ekuru Faro, 56 Elicalpeni Bank, 5 Embudu, 62, Pis. 26, 27 Cuttle-fish bones and Spirula on beach of, 63 Height of beaches of, 63 Pumice on beaches of, 63 Ereadn, 88, PI. 43 Dimension of, 88 Early stage of closure of lagoon in, 88 Etingili, 71 Fadiffolu, 73, Pis. 1, 3 ; 8 a, fig. 4 ; $b, fig. 10 ; 8c, fig. 28 ; 33, fig. 2 ; 34, fig. 1 Absence of rings in, 76 Compared to other Maldive atolls, 75 Corals on western horn of, 76 Dimensions of, 74 Exposure of, to monsoons, 77 Lagoon reefs of, 74 Fadiffolu — Continued Resemblance of, to Pacific atolls, 74 Southeastern reef flat, 76 Southwest face of, 75 Trend of coasts, 76 Velus of, 74 Western horn of, 75 Fahala, 122 Coalescence of islands at, 122 Faidu, 76, 91 Farna, 109 ; Pis. 59 ; 78, fig. 1 Faro Doru, 72 Faros of North Male, 36, 38 Bottom deposits of, 152 Coalescence of adjoining, 39 Depth of lagoon of, 40 Origin of large, 40 Rims of, 39 Temperature of lagoon of, 150 Farukolu, 89 Feartu, 101, PI. 61, fig. 1 Fehendu, 54, PI. 19, fig. 2 Felidu Atoll, 114, Pis. 1, 4, 5; 86, figs. 15, 16 Bank of angular boulders in lagoon of, 116 Closed periphery of, 117 Corals on reef flats of, 117 Faros on outer faces of, 115 Irregular outline of, 114 Islands of, 117 Lagoon cusps of, 116 Narrow passes into, 115 Passes of east face of, 117 Temperature of lagoon of, 150 Femfuri Faro, 57, 66 Fengbu Hurah, 72 Filadu, 80, 98, PI. 50, fig. 1 Connecting ridge of, 98 Lagoon reef of, 98 Secondary lagoon of, 98 Fiori, 138 Fivaku, 91 Flying fish, track of, 157 Fua Mulaku, 145 Funadu (Haddummati), 129 Funadu (Suvadiva), 141 Furadu, 107, PI. 58, fig. 1 Furena, 37 Furnadu, 89, PI. 44, fig. 1 Small lagoon of, 89 Furudu, 64 Wasting of, 55 Furuwari, 66 Fusmimdu, 68 Gadu (Haddummati), 129 Gadu (Suvadiva), 139 INDEX. 163 Gafuri, 99 Gaha Faro, 53, Pis. 1, 3; 8a, figs. 6, 8; 86, fig. 11 An atoll like Pacific atolls, 53 Free circulation of sea over reef flats, 53 Galandu, 102 Gallandu Channel, 34, 100 Gan (Suvadiva), 139 Gan (Addu), 147, Pis. 76 ; 77, fig. 1 Gang, 130 Gardiner, J. Stanley, On beach sandstone, 106 On bottom in Maldive atolls, 50 On changes at Fadiffolu, 73 On changes at Haddummati, 127 On changes at northeast horn of Addu, 147 On changes at Suvadiva, 144 On changes at Turadu, 62 On changes in Kolumadulu, 122 On changes in North and South Nilandu, 108 On changes in South Nilandu, 110 On changes of faros in Ari, 105 On comparison of Moresby's Chart of North Malosmadulu with his sketches, 67 On depth of Suvadiva lagoon, 143 On elevated reef rock of South Malosmadulu, 61 On enlargement of velus, 67 On estimate of land and water area of Addu, 148 On Farukolu and Furnadu, 89 On fissure zone in Maldives, 88 On growth of corals at Addu, 148 On growth of individual banks, 67 On Hainbuddu, 87 On islands off Gan, Addu, 147 On Kolumadulu as a typical atoll, 134 On Moresby's Charts, 35 On oceanic condition of Goifurfehendu, 54 On small importance of monsoons on edge of reef, 82 On solution in lagoon of Addu, 148 On soundings in Addu, 148 On soundings across Ariyaddu Channel, 27 On soundings across Wadu Channel, 27 On soundings in central basin, 27 On soundings in channels of Maldives, 26 On soundings in North Male, 36 On soundings in Suvadiva, 143 On the filling of Addu, 148 On topography of central plateau, 10 On varying depths of banks of Maldives, 134 On western face of Suvadiva, 143 Globigerina sand, 14 Goadu, 92 Goidu, 54, PI. 20 Goifurfehendu, 54, Pis. 1, 3 ; 8 a, fig. 8 ; 8 b, fig. 10 ; 19, fig. 2 ; 20 Goifurfehendu — Continued Depth of lagoon of, 54 Exposed to northeast monsoon, 54 Former connection of Fehendu, 54 Furudu on rim flat of, 55 Islands on, 54 Gumati, 101, PI. 64, fig. 1 Closing of gap of, 101 Guradu (South Nilandu), 114 Guradu (Kolumadulu), 125 Guru, 114 Deep faro north of, 114 Haapai group reefs, 2 Haddummati, 127, Pis. 1,6; 8 c, figs. 20, 22 ; 67-70 ; 77, fig. 1. Bays enclosing part of lagoon of, 131 Belongs to Pacific atoll type, 128 Bottom deposits of, 153 Clear of islands, 128 Corals on lagoon slopes of, 129 Depth of, 128 Effect of monsoons on outer faces of, 130 Formation of lagoons of, 133 Islands on west face of, 128 Large triangular northeast horn of, 132 Narrow boulder belt of, 130 Narrow gap at Gang in, 131 Reef flats of, 133 Temperature of lagoon of, 160 Vegetation of, 129 Hainbuddu, 87 Hanikandu Faro, 64 Hanimadu, 96 Connecting ridge of islands of, 97 Narrow atoll of, 97 Hanlus, 141 Vegetation of, 141 Haru Hura, 114 Islands on flat of, 114 Hatedu, 136 Hauwandu, PI. 54, fig. 1. Havaru, 138 Hekara, 109, PI. 60, fig. 1 Boulders at, 109 Hembadu, 48 Hibadu, PI. 74, fig. 2 Hirilandu, 126 Hitadu (Addu), 146 Hitadu (South Malos), 61 Hitadu (Haddummati), 128, 129 Hitty, 48, PI. 17, fig. 2 Hoholundu, 68 Hondedu, 138 Hoskyn, Commander, 1 Huludali, 111 164 INDEX. Huludu (Addu), 146 Huludu (Miladummadulu), 84 Hulule, 37, Pis. 10, 11 Huluwa, PI. 62, fig. 1 Huluwarolu, 137 Huni Faro, 60 Huradu, 136 Hura Mula, 136 Hurudu, 57 Hute Kolu, 114 Hydrography of the Maldives, 1, Pis. 1-8 c Ihavandiffulu Atoll, 100, Pis. 1,2; 8 a, figs. 54, 55, fig. 1 Irregular shape of, 100 Ihavandu, 101 Imma, 37 Inamadu, 70 Inawari, 73 Isdu, 132, PI. 69, fig. 2 ; 70 Jeweller's Islands, 111, PI. 78, fig. 2 Kabafaro, 90 Kadu, 130 Kagi, 61 Meagre vegetation of, 51 Kakum, PI. 68 Kalpeni, 6 Kandimas, 112, Pis. 62, fig. 2; 63, 64, fig. 2 Kandudu (Kolumadulu), 126 Kandudu (Suvadiva), 138 Kandufuri, 126 Kandu Huludu, 137, PI. 74, fig. 1 Kandute, 92 Exposed to southwest monsoon, 93 Kanifuri, 75 Corals off, 76 Exposure of, to southwest monsoon, 76 Karaidu, 132, Pis. 69, fig. 1 ; 77, fig. 1 Karawatti, 5 Karema, 83 Wasting of, 84 Corals on rim flats of, 84 Kari Faro, 65, PI. 29, fig. 1 Karidu (Middle Malos), 64 Karidu, 55, PI. 21 Compared to Maldivian atolls, 59 Compared to Pacific atolls, 59 Faro is an atoll, 58 Forest trees in centre of island of, 55 Lagoon of, partly filled, 56 Opening into lagoon of, 56 Wasting of island of, 56 Keadu, 116 Kelai, 79, 80, PI. 50, fig. 2 Entranae to lagoon of, 98 Kendikolu, 79, 86 Mangroves at, 86 Sink of, 86 Rendu, 67, 64, PI. 29, fig. 2 Kiadu, 64 Kiadufuri, 64 Kiltan, 5 Kimbudu, 126 Kofenbe, 91 Kolufuri, 121 Kolumadulu, 122, Pis. 1,5; 8 b, figs. 17, 18 ; 8 c, fig. 20; 66 Bottom deposits of, 153 Depth of, 122 Consolidation of islands at, 123 Interior basin of, shut off from sea, 124 Line of angular boulders on east face of, 123 Position of islands in lagoon of, 122 Structure of outer faces of, 125 Temperature of lagoon of, 150 Vegetation on sea face of, 126 Komandu, 86, PI. 36, fig. 2 Wasting of, 87 Konipafuri, 112 Kotafuri, 69 Wasting of, 70 Kuda Kura, 70 Kuda Mandu, 87, PI. 40, fig. 2 Kudahitty, 47, PI. 17, fig. 1. Kudu, 142 Kukuludi Faro, 67 Gardiner on changes at, 67 Kuludu Faro, 89, Pis. 37, 38 ; 79, fig. 3 Kumberidu, 95, PI. 48, fig. 2 Kunahandu, 120 Kumfinadu, PI. 3 Kurada, 144 Kurafuri, 104 Kuraidu, 91 Kurangdu, 74 Kuredu, 76 Kureli, 120, PI. 65 Boulders on west face of, 120 Kurimbe, 96 High beaches of, 95 Labadu, 141 Laccadive Archipelago, 6, PI. 8 Laccadives, Plateau of the, 7 Soundings among the, 6 Survey of the, 1 Lofara, 85 Lohi, 74 Wasting of, 74 INDEX. 165 Lowalafuri, 74, PI. 34, fig. 1 Encroachment of shingle on beaches of, 74 Mabadu, 132 Mabaru, 60, PL 22, fig. 2 Madali, 111 Maddedu, 86, PI. 36, fig. 1 Madidu, 91 Madu, PL 3 Maduni Faro, 56, 67 Maduwari, 61, PL 25 Ma Faro (Miladummadulu), 79, 86 Ma Faro (Malosmadulu), 56, 66 Gardiner on velu of, 67 Mafilefuri, 75 Mafuri (South Male), 114, PL 79, fig. 2 Mafuri (Suvadiva), 135, Pis. 71, 72; 79, fig. 1 Magudu, 107 Mahafai, 95, Pis. 48, fig. 1 ; 49, fig. 2 Maimbudu, 112 Makandudu, 89 Sink of, 89 Makara, 70 Makunudu, 71, Pis. 1, 2; 8a, fig. 7; 31, fig. 2; 32; 33, fig. 1 Boulder wall of, 73 Depth of lagoon of, 72 Dry reef of, 72 Elevation of, 72 Makunudu Island, 72 Maldive Islands, Absence of elevation in, 124 Bottom deposits of lagoons in the, 152 Character of atolls of the, 120, 124 Changes in the, going south, 120 Contrast of outer rims of the, to Pacific atoll rims, 142 Description of atolls in the, 35 Depth of channels between eastern and western chain of the, 3 Dip of plateau of the, 3 Distance from Ceylon, 2 Double chain of the, 3 Limits of monsoons in the, 8 Pelagic fauna of the, 153 Plateau of the, 9 Single line of the, 3 Male Faro (South Nilandu), 112 Male Island, 41 Malosmadulu Plateau, £6, Pis. 1, 3; 8a, fig. 5; 81,, figs. 9, 10 ; 8 c, fig. 28 ; 22-30 ; 31, fig. 1 Comparison of, with Goifurfehendu, 57 Deepest sounding in lagoon of, 57 Deep passes on east face of, 56 Faros in basin of, 67 Faros of, compared to atolls, 58 Malosmadulu Plateau — Continued Faros with wide reef flats on, 66 Formation of islands on, 57 Position of faros and islands on, 56 Resemblance of, to Ari, 56 Mamanaga Faro, 66, PL 30, fig. 1 Mameta, 143 Mamigeli, 104, 105, Pis. 55, fig. 2 ; 56 Manafur, 101 Mandu (Ari), 106 Mandu (Haddummati), 132 Manganese nodules, 24 Marandu, 100 Mararrekellu, 57, 65 Maregiri, 66, Pis. 30, fig. 2; 31, fig. 1 Dunes of, 67 Maro, 74 Matiwari, 106 Matoda, 138 Mavaidu, 91 Mavaru, 128, PL 67 Mavila, 85 Mawa Faro, 56 Mawafuri, 109, PL 61, fig. 2 Medu, 67, 69 Vegetation of, 69 Height of beaches of, 69 Megeli, 63 Island not on chart of, 63 Middle Malosmadulu, 64, Pis. 1, 3 ; 8 b, figs. 9, 10; 29, fig. 1 Changes in islands of, 65 Islands of, dished, 65 Midu, 146 Miladummadulu, 83, Pis. 1,2,3; 8a, figs. 4,6,7; 34, fig. 2 ; 35-44 ; 79, fig. 3 Bottom deposits of, 153 Central gap of, 91 Crescentic islands on east face of, 86 Exposed to southwest monsoon, 84 Formation of atolls from crescentic islands in, 87 Inner line of island near east face of, 90 Islands on east face of, steep to, 88 Islands with high beaches in, 85 Milandu, 89 Wasting of, 89 Dumb-bell shape of, 90 Miledu, 100, PL 52, fig. 2 Minikoi, 7 Mirufuri, 35 Moresby, Commander, Survey of Maldives,l Muadu, 94 Spits off east end of, 94 Muduwari (South Malos), 64, PL 28, fig. 2 Muduwari (North Malos), 68 Muheri, 95 166 INDEX. Muladu, 101, PI. 54, fig. 1 Mulaku, 118, Pis. 1, 5; 86, figs. 13, 15, 18; 64, fig. 1; 66 Absence of islands on west face of, 121 Bottom deposits of, 153 Corals on inner flats of, 121 Depth of lagoon of, 118 Faros and rings in lagoon of, 119, 120 Faros in outer faces of, 119 Reef flats on east face of, 119 Resemblance of, to Pacific atolls, 119 Similar to atolls of single line, 119 Transverse dikes of boulders at, 120 Munafuri, 128 Muradu (West Tiladummati), 82, 93, Pk. 45 ; 46, fig. 1 Change in islands of, 93 No pass into lagoon of, 94 Muradu (Northeast Tiladummati), 95, PL 48, fig. 1 Murdu, 102 Murray, Sir John, Bottom deposits of the Maldives, 152 Mutalifoori, 63, PI. 24, fig. 1 Nadale, 138 Nalandu, 90, PL 44, fig. 2 Sink of, bare at low water, 90 Lagoon and sink of, collected, 90 Naguri, 94, Pis. 46, fig. 2 ; 47 Peculiar shape of, 94 Naraka, 126, PL 66 Nipafura, PL 60, fig. 2 Noliwang Faro, 83, 96 Nomuka group reefs, 2 North Male, 35, Pis. 1, 3, 4 ; 8a, fig. 6 ; 8 b, figs. 11, 19; 8 c, fig. 24; 9-18; 19, fig. 1 Absence of vegetation on west face of, 52 Bottom deposits of, 152 Bottom of lagoon of, 43, 49 Central sink of islands in lagoon of, 44 Corals on faros of, 46 Corals on slopes of island in lagoon of, 44 Depth of lagoon of, 43 Open northern extremity of, 51 Passage of rings to islands in, 47 Passes into, 49 Purity of water of lagoon of, 44 Rings and faros of, 42, 45, 46 Soundings in, 49 Temperature of lagoon of, 150 Vegetation of islands of lagoon of, 44 Wooded islands in, 41 North Malosmadulu, 65, Pis. 1, 3 ; 8 a, fig. 5; 8 6, fig. 9; 8 c, fig. 26; 30; 31, fig. 1 Bottom deposits of, 162 Depth of, 71 Dunes at, 67 North Malosmadulu — Continued East face of, 69 Enlargement of velus of, 67 "Jungle of reefs " in, 67, 68 Large islands on east face of, 66 Open at northern extremity, 65 Sand encroaching on vegetation of, 67 Southwest part of, clear, 66 Temperature of lagoon of, 151 Vegetation of east face of, 71 Vegetation of islands of, 68 Wide passes into, 66 North Nilandu, 107, Pis. 1, 4, 6; 8 6, fig. 14 ; 59-61 ; 78, fig. 1 Banks with velus in, 108 Channel between South and, 108 Cusps on west face of, 108 East face of, 108 Elevation of parts of outer rim of, 109 Islands in interior of, 107 Rim of, 107 Well-defined reef platforms of, 110 Nu, 91 Numara, 91 Nuriwari, 96 Olugeri, 61 Pelagic Fauna of the Maldives, 155 Peremul Par, 5 Phares, 129 Piti Bank, 5 Powell Islands, 58, 71 Isolated atolls of, 58 Powell, Lieutenant, survey of Maldives, 1 Prentice, Lieutenant, on disappearance of an island in the Maldives, 84 Pteropod ooze, 24 Rakidu, 116 Narrow lagoon along, 116 Rasdu, 102, Pis. 1, 4 Depths of atoll of, 102 Rasmadu, 70 Roongelly, 69 Sesostris Bank, 5, PL 8 Soundings taken by the "Amra," 12, 29, Pis. 1-8 c Across Ariyaddu Channel, 21, 30, PL 86, fig. 14 Across Central Basin of the Maldives, 13, 19, 30, 33, Pis. 8 a, fig. 5; 86, figs. 12,13; 8 c, 28 Across Equatorial Channel, 25, 31, PL 8 c, fig. 23 Across Gallandu Channel, 12, 34, PL 8 a, fig. 3 Across Fadiffolu to Miladummadulu, 16, 34, PL 8 a, fig. 4 Across Fulidu Channel, 20, 32, PL 8 6, fig. 16 INDEX. 167 Soundings taken by the "Amra" — Continued Across Goifurfehendu to South Malosmadulu, 15, 33, PI. 8/>, fig. 10 Across Kardiva Channel, 16, 33, PI. 8 6, fig. 11 Across Malosmadulu Channels, 15, PI. 8 6, figs. 9, 10 Across North Male to Gaha Faro, 17, 32, PI. 8 a, fig. 6 Across One and half degree Channel, 21, 31, PL 8 c, fig. 22 Across Veimandu Channel, 23, 31, PI. 8 c, fig. 20 Across Wadu Channel, 18, 29, PI. 8 6, fig. 19 Between Kolumadulu and South Nilandu, 22, 32, PI. 8 6, fig. 17 Between Miladummadulu and Makunudu, 12, 34, PI. 8 a, fig. 7 Between Miladummadulu and North Malosma- dulu, 13,34, PI. 8 a, fig. 5 Between Mulaku and Kolumadulu, 22, 30, PI. 8 6, fig. 18 Between Rasdu and Ari, 102, PI. 4 Between South Male and Ari, 29 Between Toddu and Rasdu, 102, PI. 4 East of Kelai, 12, PI. 8 a, fig. 2 East of South Male, 18, PI. 8 c, fig. 27 From Gaha Faro to Goifurfehendu, 14, 33, PI. 8 a, fig. 8 From Mulaku to Wataru and Felidu, 21, 32, PI. 8 b, fig. 15 From Murdu to Colombo, 26, 34, PI. 1 From North to South Nilandu, 22 Off oceanic faces of the Maldives, 12, 20, 29, 30, 32, 34, PI. 8 c, figs. 21-27 Offlhavandiffulu, 12, PI. 8 a, fig. 1 South of Addu, 25, 31, PI. 8 c, fig. 21 West of North Male, 17, PI. 8 c, fig. 24 South Male, 113, Pis. 1, 4; 8 b, figs. 12, 16; 8 c, figs. 19, 27 ; 79, fig. 2 Deep lagoon on east face of, 114 Dimensions of, 113 Similarity to North Male, 113 South Malosmadulu, 59, Pis. 1, 3; 86, fig. 10; 8 c, fig. 28 ; 22-28 ; 29, fig. 2 Bottom deposits of, 152 Changes in southeastern horn of, 61 Corals on rim flats and lagoon slopes at, 63 Exposure of, to northeast monsoon, 60 Sink in islands of, 63, 64 Size of islands of, 60 Slope of east face of, 60 Temperature of lagoon of, 151 Vegetation of islands of, 60 Wide passes of southeast face of, 60 South Nilandu, 110, Pis. 1, 5; 86, figs. 13,. 17 ; 62- 64, fig. 2; 78, fig. 2 Distribution of banks and faros in, 110 South Nilandu — Continued Formation of islets in, 113 Formation of lagoon on southeast face of, 111 Jewellers Islands in, 111 Nullipores at, 113 Outliers of coral reef rock in, 112 Passes separating outer faros of, 112 Suheli Par, 5, PI. 8 Surveys of the Maldives, 1 Suvadiva, 134, Pis. 1, 6; 8 c, figs. 22, 23; 71-75; 79, fig. 1 An open oceanic atoll, 143 Bottom deposits of, 153 Boulders on flats of, 139 Climate of, 136 Comparison of, with Pacific atolls, 135 Corals in passes of, 141 Corals on lagoon slopes of, 135 Depth of, 138 Extent of land rim of, 140 Formation of bays in, 137 Faros on parts of outer faces of, 134 Gaps between islands of, 139 Great size of, 134 Islands and banks in lagoon of, 135 Narrow reef fiats of, 135 Reef fiat of northeast horn of, 142 Resemblance of, to Marshall atolls, 135 Shallow passes into, 138 Temperature of lagoon of, 151 Vegetation of, 141 Width of passes of, 140 Tahwahtah, 69 Telin Faro, 67 Temperature of the lagoons, 149 In Addu, 150 In Ari, 149 In Fadiffolu, 150 In Felidu, 150 In Haddummati, 150 In Kolumadulu, 150 In Mulaku, 149 In North Male, 150 In North Malosmadulu, 151 In South Malosmadulu, 151 In Suvadiva, 150 In Tiladummati, 151 Tiladummati, 93, Pis. 1, 2 ; 8 a, figs. 2, 3; 45-53 Artificial boundary of, 93 Atolls of northeast horn of, 98 Bottom deposits of, 153 Changes in outline of islands of, 93 Coalescence of islands of, 98 Crescent-shaped islands of, 98 Effect of moonlight on reef flats of, 97 168 INDEX. Tiladummati — Continued Faros of east and west face of, 96 Islands characteristic of inner waters of, 95 Islands of northern part of, closely packed, 97 Pandanus on northern islands of, 95 Temperature of lagoon of, 151 Tiladummati-Miladummadulu Plateau, 78, Pis. 1, 2,3; 8«, figs. 2-5,7; 34, fig. 2 ; 35-53; 79, fig. 3 Atolls and faros of, based on separate knolls, 82 Atolls and faros of, not formed by subsidence, 80 Atolls of, become islands, 80 Atolls of northeast horn of, 79 Boundary line between, 78 Change in outlines of islands of, 82 Compared to southern groups, 78 Condition of, not that of an atoll, 82 Crescent-shaped islands of east face of, 79 Current sweeping into, 82 Faros passing to islands in, 80 Formation of faros of, 80 Formation of separate islands of, 83 Great width of passages of, 81 Islands of interior of, 80 Junction of separate islands of, 83 Motion of islands like that of dunes in, 83 Modifying effect of monsoons on, 82 Open interior waters of, 81 Position of islands on, 78 Shape of, 78 Small land area of, 81 Subsidence does not account for formation of atolls of, 81 West face of, 80 Timarafuri, 124 Tinadu (Felidu), 117 Tinadu (Suvadiva), 144 Toddu, 102, Pis. 1, 4 Tolandu, 85 Topography of Indian Ocean, 5, Pis. 7, 8 Tukandu, 100, PI. 53 Boulders of, 100 Tulagiri, 47, PI. 16, fig. 2 Tulusdu, 36 Tura, 66 Turadu, 61 Gardiner on changes at, 62 Tuvaru, 121 Ulegnma, 101 Vaimandu, 126 " Valdivia " Soundings south of Addu, 6 Vehamafuri, 47 Veimandu Channel, 23, 31, PI. 8 c, fig. 20 Velengeli, 62, 66 Wadu (North Malos), 57 Wadu (Suvadiva), 139 Waduni, 132 Wagaru, 101 Wakaru, 63 Height of beaches of, 63 Wandu, 70 Wandu Faro, 60 Wani, 112 Warifuri, 79, 80, 99, PI. 51, fig. 2 Depth of lagoon of, 99 Islands of, 99 Wataru, 84, PI. 35, fig. 1 Wataru Reef, 118, Pis. 1, 5; 8 6, fig 15; 58, fig. Islets on rim flat of, 118 Narrow entrance into, 118 Reef flats of, awash, 118 Wataru Channel, 21, 32, 118, PI. 8 b, fig. 15 Watu, 85 Wegeli, 141 Welingandu, 126 Wilingili (North Male), 35, 50 "Wiringili (Addu), 146 Wiringili (Suvadiva), 142, PI. 77, fig. 2 Elevated reef rock at, 162 Yucatan plateau, reefs of, 1 PUBLICATIONS OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. There have been published of the Bulletin Vols. I. to XXXVIII., and also Vol. XL. ; of the Memoirs, Vols. I. to XXIV., and also Vols. XXVIII. and XXIX. Vols. XXXIX., XLI. to XLV., of the Bulletin, and Vols. XXV. to XXVII., and XXX. of the Memoirs, are now in course of publication. A price list of the publications of the Museum will be sent on application to the Librarian of the Museum of Comparative Zoology, Cambridge, 3Iass. 3 2044 066 30 s-^r — - .'rv '*.. fc^* i£* ' ./* tfo£ « f > •J,. *-/■ ^.' >v .5 . «K* £."<#. - -j*^Sf >• ,y « . .. S>*3 .-•#*■, * VK i 3 *