Wels shea dies ce er eS “ Nia y 5 He i (NR gi E 7 nade ct ei oy ates Ea a yaiae atari an ch Pee] Traine sinh suet on i Kaa Say AT AR it tanta i RRO EL IRE NN FR i igtacnte ies } cs } afedclst uni abea seal ‘ cates rr ah * aes Bis Se a eae ff. ies he i Hon sg eszaniahazaest) op )Sae DEM nendeyss aad eb alicia onat id a RD r cen Re eon ve ue i f Ogg Ranieri ale 5. Sin ea saneiay ant yeas ane A ny pei i aes Mer Miksa sta ahs een Ay Araneta an tasnuead mideca em ea gains aap uta aa UT GTR aL Wa sents ear teha bight ue ane SPA baad. Bun isheataena ridaaehs Heth pv ee tise mere ERE arp aioe Mics iudastas vbeSiehs Senate fie ee seater : Missi co iste toed Ed es HS) a Toren EMDR Hi 9 fie PORTS Pet! saFwd SAN deaths a ta han sist yviened tt a crimes Cte “wr Rare Any rane CT ee TAR aR, Meant ence: fiegmasuagie, . Pate ie te Cr Woar et ne Anis 4 satel eee : pa - i Asmara vatatatiee CHM 9 ive Cerrar MAcsiy nw Sy SeRGADD Fp a ky aa lites SH et} ae teeauierars cepa tise syne fushe. yt renee aie Paria, panan gs 15 Tee aay Taalatetesbasrnteen Sn pen na ied ors heres uesy eur * mies bid " ! Eien: Haat hie Mneusan haere Na iecpeutest bis 5S a Oly Ratt ina = ETE Tete Perma ee Mita sake onty Hela iCueriatysrigeen ches Teed titanate tee eat) ue Pagatanty Fil ey au taser gree a hatpties Pata ch ide hesaper hoe ee einen dy tee i tec Nratedi ety B eetceate ty a ieteaee eee ite plese pale we et dietabatis eli iS tts wa aaa cet Bed pptie kad oe i oy ci etgtatead sy tito aselty ee iy stun a ae faves 2 fea bein Ieaioh tater ata! ie faa cayaatiee ee ae Yatierss sh oh “ peeivgetoepsine RAY Aa tb rntes a mies Foy We ets " i iets a, RE eat a iene VA eae an te 3 Dh etamata lun isisy OA me iy + ro f BI ATOLL RESEARCH BULLETIN NOS. 399-414 RESEARCH BULLETIN ECOLOGY AND GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS Issued by 2 e) Fe iS se EY < Z, ey fe) = =) Fa N =) = , : EY < Z INSTITUTION D.C. U.S.A. SMITHSONIAN WASHINGTON, FEBRUARY 1994 ATOLL RESEARCH BULLETIN NOS. 399-414 ECOLOGY AND GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS EDITED BY COLIN D. WOODROFFE ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 ATOLL RESEARCH BULLETIN NOS. 399-414 NO. 399. SCIENTIFIC STUDIES IN THE COCOS (KEELING) ISLANDS: AN INTRODUCTION BY C.D. WOODROFFE AND P.F. BERRY NO. 400. CLIMATE, HYDROLOGY AND WATER RESOURCES OF THE COCOS (KEELING) ISLANDS BY A.C. FALKLAND NO. 401. LATE QUATERNARY MORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY D.E. SEARLE NO. 402. GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE, R.F. McLEAN AND E. WALLENSKY NO. 403. REEF ISLANDS OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE AND R.F McLEAN NO. 404. VEGETATION AND FLORA OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS ‘NO. 405. AN UPDATE OF BIRDS OF THE COCOS (KEELING) ISLANDS BY T. STOKES NO. 406. MARINE HABITATS OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS NO. 407. SEDIMENT FACIES OF THE COCOS (KEELING) ISLANDS LAGOON BY S.G. SMITHERS NO. 408. HYDRODYNAMIC OBSERVATIONS OF THE COCOS (KEELING) ISLANDS LAGOON BY P. KENCH NO. 409. HERMATYPIC CORALS OF THE COCOS (KEELING) ISLANDS: A SUMMARY BY J.E.N. VERON NO. NO. NO. NO. NO. 410. 411. 412. 413. 414. MARINE MOLLUSCS OF THE COCOS (KEELING) ISLANDS BY F.E. WELLS ECHINODERMS OF THE COCOS (KEELING) ISLANDS BY L.M. MARSH , FISHES OF THE COCOS (KEELING) ISLANDS BY G.R. ALLEN AND W.F. SMITH-VANIZ BARNACLES OF THE COCOS (KEELING) ISLANDS BY D.S. JONES DECAPOD CRUSTACEANS OF THE COCOS (KEELING) ISLANDS BY G.J. MORGAN (Manuscripts received 15 March 1993; Revised 8 December 1993) ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 ACKNOWLEDGMENT The Atoll Research Bulletin is issued by the Smithsonian Institution to provide an outlet for information on the biota of tropical islands and reefs and on the environment that supports the biota. The Bulletin is supported by the National Museum of Natural History and is produced by the Smithsonian Press. This special issue is financed and distributed with funds from the Department of the Environment, Sport and Territories, Canberra, Australia, the Australian Nature Conservation Agency, Directorate for Northern Australia, and Atoll Research Bulletin readers. The Bulletin was founded in 1951 and the first 117 numbers were issued by Pacific Science Board, National Academy of Sciences, with financial support from the Office of Naval Research. Its pages were devoted largely to reports resulting from the Pacific Science Board’s Coral Atoll Program. All statements made in papers published in the Atoll Research Bulletin are the sole responsibility of the authors and do not necessarily represent the views of the Smithsonian nor of the editors of the Bulletin. Articles submitted for publication in the Atoll Research Bulletin should be original papers in a format similar to that found in recent issues of the Bulletin. First drafts of manuscripts should be typewritten double spaced and can be sent to any of the editors. After the manuscript has been reviewed and accepted, the author will be provided with a page format with which to prepare a single-spaced camera-ready copy of the manuscript. COORDINATING EDITOR Ian G. Macintyre National Museum of Natural History MRC-125 Smithsonian Institution Washington, D.C. 20560 EDITORIAL BOARD Stephen D. Cairns (MRC-163) National Museum of Natural History Brian F. Kensley (MRC-163) (Insert appropriate MRC code) Mark M. Littler (MRC-166) Smithsonian Institution Wayne N. Mathis (MRC-169) Washington, D.C. 20560 Victor G. Springer (MRC-159) Joshua I. Tracey, Jr. (MRC-137) Warren L. Wagner (MRC-166) Roger B. Clapp (MRC-111) National Museum of Natural History National Biological Survey Smithsonian Institution Washington, D.C. 20560 David R. Stoddart Department of Geography 501 Earth Sciences Building University of California Berkeley, CA 94720 Bernard M. Salvat Laboratoire de Biologie & Ecologie Tropicale et Méditerranéenne Ecole Pratique des Hautes Etudes Labo. Biologie Marine et Malacologie Université de Perpignan 66025 Perpignan Cedex, France Bt fa al ba -Fypse = ; wed Sg Rene ma ete? Wercee Sp Mined at kw ” Teese tay shay sah aA x ms) . >} err ‘is 15 aca AACE? Navan bet Yes os Yui ein geet sfyreayet £8 ats ES Ge ety ty ae) a ny enn ep sea — * ‘mat, a a a. aoe, FORE BA * i nati hareatigfae \amemterescreaniaibe pet ar eeecamndatl WLuee- selhshoit Bare dvznonet. Lindh oft ath cotitmod tii Both Best: Rape ire wi Fo MOE hoe sk et del ool coe, Lembat poe Py mh, ek - ahaa Sines Fancesayee ae ures “gaps ot be rogy? ad BApowts sap ioneainime Pon fe erwra kwh obo PM ot at ed? 2a Large es wealth Sav 5 Re ge RY = Sia Bae * esl Pee meatal 05 aR i bank ae Week etiow i es a ty tect aah. ” bial + ea wate de - Cos i 2 Cr mee taaokaun t. by ae ‘ @Ls-ORM ; it opts bak woke hy , OD pe daa Oe OR 1 i yee a eee hata (fat =, . : Ghee pi IGM ws ener H : (éal ~ Ol) fie tao cee | E ‘4 . otro hha aX é a vt ; ‘ : toe - We! ee (SEL . Sete a 7 ie Se ee ee ee he Oe” a '. 6% uae 6 iy 4 8 (U3l.. pee Tecra | AE TER CTT ra 7 a ane tt ~ Bs ji Sa r PE Se > eer : va ra : % Ej fe at > oh v - a - “a | r is ATOLL RESEARCH BULLETIN NO. 399 CHAPTER 1 SCIENTIFIC STUDIES IN THE COCOS (KEELING) ISLANDS: AN INTRODUCTION BY C.D. WOODROFFE AND P.F. BERRY ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 96°49 96°50'E -. 96°55 12°05'S A, NORTH ee KEELING INDIAN OCEAN TENN NN Pulu Ampang Q O\ Pulu Kembang @_\ COCOS Pulu Cepelok \.. Pulu Wak Banka \ Pulu Pandan‘) co (KEELING) Pulu Siput x / Pulu Labu=>\ West Is. ISLANDS 12°10'S iy es YA 12°10'S : Pulu ‘\ Kambing 96° 50°E 96°55'E CHAPTER 1 SCIENTIFIC STUDIES IN THE COCOS (KEELING) ISLANDS: AN INTRODUCTION. BY C.D. WOODROFFE * AND P.F. BERRY ** INTRODUCTION The Cocos (Keeling) Islands are a particularly isolated group of islands in the eastern Indian Ocean (Fig. 1). They comprise a southern group, the South Keeling Islands (12° 12' S, 96° 54’ E), which form a coral atoll with a shallow lagoon fringed by a series of reef islands. A single horseshoe-shaped island (atollon), North Keeling (11° 50' S, 96° 49' E) is located 26 km to the north of the main group. They lie 900 km west of Christmas Island and 1000 km southwest of Java Head. DISCOVERY AND SETTLEMENT The group is named after the coconut (Cocos nucifera), which grew there in profusion, even before deliberate planting of all of the southern atoll as a part of the Clunies Ross estate, and Captain William Keeling. Keeling is believed to have been the first European to sight the islands in 1609 on his return from Bantam, on behalf of the East India Company, though there is no record of that sighting. The islands are not shown in the 1606 edition of Ortelius Theatrum Orbis Terrarum, but do appear in Blaeu's appendix to the third edition produced about 1631. They are recorded with the name Cocos Eylanden in a manuscript map drawn by Hessel Gerritsz in 1622, and on Dudley's Arcano del Mare (1646) on which it says that they were discovered by the English. On a Dutch chart produced in Amsterdam in 1659 they are called the Cocos Islands; though around this time they were also called the Triangular Islands. The English hydrographer Thornton used the name Kelling Island in his Oriental Navigation of 1703. Captain Ekeberg from Sweden visited North Keeling in 1740. There is an account of the islands in van Keulen's Zeefakkel, (6th edition, 1753), with a map attributed to the Dutch navigator, Jan de Marre (1729). In his sailing directory for this region of the Indian Ocean, compiled in 1805, the British hydrographer, James Horsburgh, called them the Cocos-Keeling Islands, and named one of the islands after himself. After settlement the early inhabitants called them the Borneo Coral Reefs after the supply vessel the Borneo, owned by John and Joseph Hare and Co, and captained by John Clunies Ross. They were also known as the Keeling- Cocos Islands, and after 1955 they became officially the Cocos (Keeling) Islands. Despite knowledge of the islands for 200 years or more, it was not until the early nineteenth century that they were settled and an interest was taken in them because they lay ‘ Department of Geography, University of Wollongong, Northfields Avenue, Wollongong, New South Wales, 2522. ee Western Australian Museum, Francis Street, Perth, Western Australia, 6000. on a trade route from Europe to the Far East. The first settlement was accidental, Captain Le Cour and the crew of the brig Mauritius lived on Direction Island for several weeks after their ship was wrecked on the reef. Captain Driscoll in the Lonach went ashore on 24 November 1825, shortly after the men were rescued, and noted the wreck of the Mauritius. Shortly thereafter, on 6 December 1825, Captain John Clunies Ross, a Scottisl trader sailing the Borneo for Alexander Hare's company, made a brief landing on the islands. During his visit he sounded the main channel, cleared an area on Horsburgh and Direction Islands, and planted cereals and vegetables. In the following year a settlement was established by Alexander Hare, a somewhat discredited trader and associate of Ross, who brought with him a crew of largely Sumatran and Javanese seamen and an entourage of women of various nationalities, on the Hippomenes. John Clunies Ross and his family returned on 16 February 1827 (this is the date given by Ackrill 1984; several other authors state that he returned on 27 November 1827, i.e. Hughes 1950; this error appears to have been perpetuated since an account in Gleanings in Science, Anon 1830; the correct dates are given by Gibson-Hill 1952), with the intention of commencing a settlement on the islands. Ross expressed surprise at finding Hare and a large group of people there, although it would seem that he should have known that Hare might already be there, since he had not found him at Cape Town on the outward journey. Relations between Ross and Hare deteriorated; Hare initially lived on Home Island, though with some occupation of West Island, Direction Island, and Horsburgh Island. Ross initially settled on Pulu Gangsa, just to the north of Home Island (presently joined onto Home Island, and the site of the cemetery); but he soon moved his house down to the central part of South Island. Hare, on the other hand, eventually moved from Home Island, to Prison Island, according to poorly substantiated reports because the women of his party were being molested. It appears that he maintained a party of womenfolk, and most of the children, under close supervision in the lower storey of his house. Rear Admiral Sir Edward Owen noted in a letter of 10 July 1830, that Hare engaged in unrestricted intercourse with such females of his establishment as he (might) deem worthy of his attentions’. The history of this time is particularly colourful, although it has also been subject to a lot of misrepresentation (Gibson-Hill 1947, Ackrill 1984). In part this results because the only records that survive reflect Ross' view of the circumstances, and his objectives and those of Hare differed. Ross endeavoured to establish a successful settlement, aspiring to a great trading 'entrepot'’. Hare, on the other hand, sought obscurity and his behaviour, according to Ross, became increasingly unbalanced and debauched. Initially both claimed ownership of the islands, but in 1831 Hare finally left the atoll, and he died in Batavia in the following year. The Malay workers, who joined with Ross, are the ancestors of many of the Cocos Malays there today, and the Clunies Ross family, who became known as 'Kings of the Cocos’, reigned over the islands for more than 150 years (Hughes 1950). John Clunies Ross died in 1854. He was succeeded by his son John George Clunies Ross until 1871. After his death in 1871 George Clunies-Ross took possession until 1910. John Sydney Clunies-Ross ruled from 1910 to 1944, and John Cecil Clunies- Ross thereafter. The papers of the Clunies-Ross family contain much useful information on the islands. The early papers of John Clunies Ross are particularly valuable for their description of the islands, although some were lost in a fire. In 1857 the islands were declared a part of the British Dominions by Captain Fremantle who arrived aboard H.M.S. Juno, having misread his directions which instructed that he annex Cocos in the Andaman Islands (Gibson-Hill 1947). Responsibility for supervision of the islands was transferred over the years to the Governments of Ceylon (1878), the Straits Settlements (1886), Singapore (1903) and Ceylon again (1939-1945). In 1886 Queen Victoria granted all of the islands, under certain provisions, to John George Clunies Ross, in perpetuity. They became a Territory of the Commonwealth of Australia in 1955, and in 1978 Australia purchased all of the lands, excepting the family home, from the Clunies Ross family for Aus$ 6.25 million. In 1984 the Malay population voted to become a part of Australia. Initially the Territory was administered by the Commonwealth government, but it is now being transferred to the responsibility of the state government of Western Australia. Today the islands are inhabited by the descendants of the original (Malay) settlers, though there have been several additional intakes of workers from various parts of Southeast Asia. There are about 400 Malays living in the kampong on Home Island. The Great House of the Clunies Ross family, Oceania House, is also on Home Island, set amongst 5 ha of grounds. Across the lagoon on West Island more than 200 people, associated with various departments of the Australian Government, live alongside the airfield, which was used as an important refueling stop on commercial air routes between Europe and Australia until the advent of larger aircraft in the 1950s allowed Cocos to be by-passed. NATURAL HISTORY AND SCIENTIFIC DESCRIPTION OF THE ISLANDS The Cocos (Keeling) Islands have held a special place in the literature on coral atolls because they represent the only atoll that Charles Darwin visited, and they played a central role in his discussion of his theory of coral reef development. The natural history of the islands was, in fact, uncharacteristically well-known by the turn of the century, because of the visits of a number of naturalists in addition to Darwin. It is interesting to note that rather than confirming Darwin's observations and interpretations of the atoll, many of the works of subsequent naturalists lead them into conflict with Darwin's views, and that Cocos was also subject to interpretations completely contrary to Darwin's. Thus, Guppy (Guppy 1889, 1890a, 1890b) subscribed, at least in part to the views of John Murray, who funded his visit. Wood-Jones put forward an entirely alternative view to Darwin's, in his book Coral and Atolls (Wood-Jones 1912). However, the strongest criticism of Darwin came from John Clunies Ross himself, who was absent at the time of Darwin's visit. His review of Darwin's book, published posthumously, was a bitter, vitriolic attack on Darwin's ideas (Ross 1855). Early accounts of the nature of the islands (Anonymous 1830) include a paper on the formation of the islands by Ross (1836), and descriptions from short visits. Owen's description (1831) appears to be based upon observations by Captain Sandilands who visited in February 1830 in the Comet (Owen, 1831), and the account in Holman (1840), is based largely on the accounts of Keating who left the atoll in November 1829 after less than a year there with Hare, and those of Captain Mangles. Van der Jagt (1831) visited and mapped the islands in 1829 in the Blora. H.M.S. Beagle called into the Cocos (Keeling) Islands on its way home, in its third year at sea, arriving off the islands on 1 April, and leaving on 12 April 1836 (Darwin 1842, 1845, Fitzroy, 1839). The visit lasted only ten days, and for much of that time Captain Fitzroy and the crew of the H.M.S. Beagle were engaged in survey work; this forming the basis of many of the subsequent charts. A more thorough hydrographic survey was undertaken in June and July 1983 by RAN Moresby. Darwin's visit has been analysed in detail by Armstrong (1991). The islands were subsequently visited by Henry Forbes, who arrived at around 17 January 1879, and left on 9 February 1879. Forbes speculated on the origin of the islands, and devoted two chapters (chapters ii and 111) to Cocos in his book on the Eastern Archipelago (Forbes 1879, 1885). Henry Brougham Guppy visited the Cocos (Keeling) Islands in 1888, his visit being funded by John Murray. Murray had examined a series of rock specimens collected from Christmas Island (Indian Ocean) by Captain Pelham Aldrich who had been there in H.M.S. Egeria in 1887, and had found some that were rich in phosphate. He sent Guppy to visit Christmas Island for a further examination. Guppy needed fine weather for a landing on Christmas Island, and took passage on the Clunies Ross vessels as the only way to get there. The weather was apparently not calm enough within the period that Guppy could stay to get to Christmas, and Guppy spent 10 weeks on Cocos waiting, before returning to Europe without getting to Christmas. Gibson-Hill (1947) suggests, however, that during the 5 months that Guppy endeavoured to get to Christmas, George Clunies-Ross' brother managed to visit the island and establish a settlement of a group of Malays in Flying Fish Cove. As a consequence of that settlement, the Clunies-Rosses began with about half of the shares (with John Murray) of the Christmas Island Phosphate Company when that commenced in the 1890s. Guppy meanwhile undertook an extremely detailed account of the Cocos (Keeling) Islands, with extensive observations on the nature and rate of operation of geological processes (Guppy 1889, 1890a, 1890b). There were other visits and accounts during this time; W.E. Birch, with Rev E.C. Spicer as naturalist, was on Cocos 20-28 August 1885 (Birch 1866). The lone round-the- world sailor Joshua Slocum arrived on 17 July 1897 in the Spray, and considered that ‘if there is a paradise on this earth it is the Keeling-Cocos' (Slocum 1899). Wood-Jones was the medical doctor at the Cable Station on Direction Island, and was on Cocos from June 1905 to September 1906. This afforded him ample opportunity to look around the atoll. He revisited Cocos briefly in 1907 as a guest of George Clunies- Ross, and subsequently married one of his daughters. He wrote a book, Coral and Atolls, incorporating his observations. However, the book suffered 'in parts from a considerable carelessness, and an over-optimistic acceptance of unconfirmed visual records’ (Gibson- Hill 1947 p.159). His view of the mode of formation of the reef differed both from that of Darwin, and that of Murray which received some modification from Guppy. The detailed observations of Wood-Jones are surpassed only by those of Gibson- Hill. Gibson-Hill was medical officer on Direction Island from 20 December 1940 until 10 November 1941. He made various collections of organisms, some material of which disappeared from the Raffles Museum during the Japanese occupation of Malaya in the World War II. Nevertheless his fascination with Cocos persisted, and Gibson-Hill published a series of notes on the islands, including both his own observations and collections (Gibson-Hill, 1947, 1948, 1949, 1950a, 1950b, 1950c, 1950d, 1950e, 1950f, 1950g), and his reprinting of earlier literature on the atoll (Gibson-Hill 1953). Since Gibson-Hill's reviews of Cocos natural history, there have been expeditions by the Academy of Natural Sciences of Philadelphia (1963 and 1974) and the Western Australian Museum (Berry 1989), as well as some visits from individual naturalists (e.g. Alfred 1961). Williams has produced an annotated bibliography of the natural history of the islands, which appears in a recent Atoll Research Bulletin (Williams 1990). A summary of major collections on Cocos is given in Table 1. BIOGEOGRAPHICAL RELATIONSHIPS OF THE COCOS (KEELING) ISLANDS BIOTA The Cocos (Keeling) Islands are not only extremely isolated, but they also lie at the western extension of the Western Pacific marine biogeographic province. For many species Cocos represents their western limit of distribution. The biota is derived, therefore, primarily from that of the tropical Indo-West Pacific; taxa from the western Indian Ocean are poorly represented. The plants contain a large component of drift- dispersed pantropical species, found throughout the Indian and Pacific Oceans, but the major source is western Java (Guppy 1890a). Similarly, in the case of the marine biota, the most likely source of larval recruitment is also the Indonesian and eastern Indian Ocean region. For all groups covered of marine biota described in subsequent chapters in this volume, significant additions have been made to existing knowledge of taxa occurring at Cocos (Table 1). Total numbers of taxa in the groups discussed in this volume are summarised in Table 2. The level of collecting now undertaken at Cocos means that the terrestrial ecology is relatively well-known, and in the marine groups is such that significant numbers of additional taxa are unlikely to be added. On this basis several generalised conclusions can be drawn regarding the biogeographical relationships of the Cocos (Keeling) Islands. There is almost no endemism in the Cocos biota. The Buff-banded Rail, Rallus philippensis andrewsi, is considered an endemic subspecies, restricted to North Keeling, and the rat on Direction Island, Rattus rattus keelingensis, has been accorded subspecies status, and was considered by Wallace (1902) to be an example of rapid divergence; it can be traced back to the Mauritius which was wrecked in 1825. The angelfish Centropyge joculator is recorded only from Christmas and the Cocos (Keeling) Islands. The Cocos subspecies of Pandanus tectorius, which is only localised in occurrence, is also considered endemic (Williams 1990, this volume). This lack of endemism may reflect the effect of rapidly oscillating sea levels during the late Quaternary, and the pattern of development of coral atolls, whereby the limestone plateau which was exposed at the last glacial maximum perhaps resembling the modern- day Christmas Island, was rapidly flooded during postglacial sea-level rise and all land was submerged in the early Holocene (10,000-8,000 years ago). The present reef islands appear to be no more than 4000 years old (Woodroffe et al. 1990a, 1990b). This implies that all the terrestrial biota must have recolonised the atoll in the last few thousand years. These sea-level fluctuations would also have had substantial implications for shallow-water ca biota, as the nature of the habitats must have altered drastically over that period also. Some taxa which might be expected are conspicuously absent from Cocos. Guppy (1890a) has drawn attention to the absence of mangroves and Nypa palm, despite the arrival of propagules on the shore (The stand of mangroves on the northern end of Horsburgh Island can be attributed to planting by John George Clunies Ross). Some shallow marine taxa usually common on coral reefs are conspicuously absent in apparently suitable habitat (i.e. there are no benthic skates or rays in the lagoon). Marine taxa must either be pelagic as adults or have long-lived larvae or juveniles to reach Cocos. Christmas Island is the nearest island to Cocos, and there is less similarity than might at first be expected between the biota found at Cocos and that at Christmas (Table 2). Undoubtedly this results from the contrasting physiography of the two islands. Christmas Island is an uplifted (and apparently still uplifting) limestone island with outcrops of volcanic rocks on it. It reaches a maximum elevation of 361 m, and has probably been above water since the Eocene. It is covered by dense forest, and has only poorly-developed reef fringing its cliffed coastline. There are not the extensive reef flats or shallow lagoonal sandy or muddy areas which are found on Cocos. _ In addition to the very different late Quaternary history, and the great contrast in the time available for establishment of terrestrial biota, the Cocos (Keeling) Islands are probably more subject to periodic catastrophic influences on the biota. The atoll has experienced several devastating tropical cyclones, which tend to have an impact all over the restricted land area of Cocos, as well as in shallow parts of the lagoon and reefs. There have been a series of coral and fish kills in the lagoon. Darwin (1842) noted an extensive area of dead coral in the southeastern corner of the lagoon. He speculated initially that this might have been due to slight emergence of the atoll, but then attributed it to the closure of a series of interisland passages through South Island and more resticted circulation in this part of the lagoon. Forbes recorded that in 1876, inky and foul smelling water had spread through the lagoon from the islands on the eastern rim (Forbes 1885). A similiar fish and coral kill has recurred, most noticeably in 1983, but also in intervening years. Its cause is still unclear. Forbes considered that the 1876 event may have been caused as a result of an earth tremor. The 1983 event was correlated with E] Nino in an incisive, but unpublished account of it by Blake and Blake, who attributed it to an ‘algal bloom’ (red tide). An alternative explanation, considered more likely by members of the Western Australian Museum expedition (Berry 1989), is that it was caused by mass coral spawning at a time of poor circulaton as has been described by Simpson et al. (in press) on the Western Australian coast. It is significant that the mortalities at Cocos and Western Australia both occurred in March and this hypothesis for the cause of mass mortality of corals and other organisms in the Cocos lagoon would be further supported if it could be established that coral spawning occurs there in March. A minor episode of fish kill was also observed in 1992 (J. Tranter, pers. comm.). Infestations of Acanthaster have been reported from the reefs (Colin 1977), and are reviewed in more detail by Marsh (this volume). Present diversity and abundance of reef organisms closely associated with living corals may have been reduced by the reduction of coral abundance and diversity on the reef slopes and in the lagoon as a result of these events. In view of the isolation of Cocos and the distance to be covered by propagules, if species are lost from the atoll as a result of such events, they are likely to be slow to recolonise. Human impact on the Cocos (Keeling) Islands has been most devastating on the South Keeling Islands (the southern atoll), where the vegetation has been almost totally altered to coconut plantation. The birds which once characterised the atoll have all but disappeared, and it is the absence of large numbers of seabirds, which strikes one as the most conspicuous difference between North Keeling and the southern atoll. There have also been impacts on marine organisms which are eaten, Tridacna gigas, Lambis lambis (gong gong), Birgus latro and the palinurid lobsters, though it is to be hoped that management measures prevent the total elimination of any more of these species. At the same time occupation of the atoll has resulted in an influx of new species to the Cocos (Keeling) Islands. These vary from ornamental and food plants, and deliberate animal introductions, such for instance as the Green Jungle Fowl on West Island. Sheep, cattle, alpacas and black rhinoceroses and other animals have been temporarily contained within the Quarantine Station on West Island. Accidental releases, include the rats and the lone King Parrot on West Island. In addition there are numerous insects which have been introduced, and a large number of insect pests (Gibson-Hill 1950f, Holloway 1982). STUDIES IN THIS VOLUME The frequent visits of naturalists, and the detailed studies that they undertook, ensured that the Cocos (Keeling) Islands were perhaps the best-known, and certainly the most hotly debated group of reefs in the world, by the end of the nineteenth century. This continued in the first half of the twentieth century. Since the overviews of Gibson-Hill, however, there has been relatively little extension of knowledge about the islands. The terrestrial ecology of the South Keeling Islands has changed markedly since settlement as large areas have had their natural vegetation replaced by coconut plantation (now sadly largely overgrown), and as birds and other food resources have been exploited. It is now possible to look in greater detail at the geological history of the atoll, and the marine ecology of its reefs and lagoon as the result of the development of several new techniques. Radiometric dating allows a new insight into the age of formation of coral limestones. Subsurface drilling, seismic profiling and dating allow geomorphological insight into those questions which intrigued the early naturalists to study Cocos. In the case of marine ecology, the development of more sophisticated underwater research methods, and particularly the use of SCUBA, has enabled the extension of collections into water depths that were previously only examinable by dredge, or not at all. It is thus in the areas of geological investigations, and in terms of the numbers of marine species on the atoll that the greatest advances in knowledge have been made in recent years. This volume brings together a series of recent studies on the Cocos (Keeling) Islands. A preliminary examination of the birds of Cocos was undertaken by T. Stokes of the Australian National Parks and Wildlife Service (ANPWS) and others in 1982. Since that time ANPWS has carried out a number of studies of birds, particularly at North Keeling. Vegetation and marine habitats of the atoll have been mapped by D. Williams, who was seconded as Environmental Resource adviser to the Territory of the Cocos (Keeling) Islands in 1986-7. Mapping was undertaken using SPOT satellite imagery, aerial photographs and ground survey on a habitat data sheet, used by Williams and ANPWS staff. A field survey of the inshore marine fauna and habitats of the Cocos (Keeling) Islands was undertaken 7-28 February 1989 by the Western Australian Museum under the leadership of P. Berry, sponsored by the Australian National Parks and Wildlife Service. During that fieldtrip a total of 37 stations was occupied (some more than once), in order to sample the major marine habitats (Fig. 2), which had been partly described and mapped by D. Williams. The survey of marine fauna involved SCUBA, snorkelling and reef walking, as well as the use of an ichthyocide and a small hand spear in the case of fishes and the photographing of habitats. Specimens have been lodged in the Western Australian Museum. Accounts of the corals, collected and identified by J.E.N. Veron, and of marine molluscs by F. Wells, echinoderms by L. Marsh, fishes by G. Allen and W. Smith-Vaniz, barnacles by D. Jones, and decapod crustaceans by G. Morgan appear in this volume. Water Resources represent an important aspect of the present settlement of Cocos, and have been studied over the last 6 years by A. Falkland. He describes the climate of Cocos, and the implications for Water Resources in this volume. Drilling undertaken for the water resources study has also provided data for a geomorphological study of the Cocos (Keeling) Islands by C. Woodroffe and R. McLean. Their geological reassessment of Cocos has involved detailed descriptions of reef islands and subsurface stratigraphy and chronology, as well as other studies. The results of seismic profiling are described by D.E. Searle, lagoon hydrodynamics by P. Kench and lagoonal sediments by S.G. Smithers. There are some aspects of the ecology and geomorphology of the Cocos (Keeling) Islands which have not been adequately treated in this volume. There is no account of the insect fauna, for instance, nor have the marine algae been studied in any great detail. It is hoped that the studies which are presented here provide a stimulus for continued scientific study and research on the Cocos (Keeling) Islands. ACKNOWLEDGEMENTS Research on the Cocos (Keeling) Islands which is presented in this volume has been made possible through the support of the Cocos (Keeling) Islands Administration, the Cocos Island Council, and the Australian National Parks and Wildlife Service. Collecting by the Western Australian Museum was funded by Australian National Parks and Wildlife Service, while geomorphological research was funded by the Australian Research Council and the National Geographic Society. The assistance of the Government Conservators, Andrew Grant, Paul Stevenson and Jeff Tranter has been very important. Logistical support from the Cocos Islands Co-operative and Australian Construction Services and John Clunies-Ross is greatly appreciated. REFERENCES Abbott, R. T. 1950. The molluscan fauna of the Cocos-Keeling Islands, Indian Ocean. Bull. Raffles Mus., 22: 68-98. Ackrill, M. 1984. The origins and nature of the first permanent settlement on the Cocos- Keeling Islands. Historical Studies, 21: 229-244. Alfred, A. E. 1961. Some birds of the Cocos-Keeling Islands. Malay. Nat. J. 15: 68-69. Anonymous 1830. Some account of the Cocos or Keeling Islands: and of their recent settlement. Gleanings in Science (Calcutta), 2, 293-301. Reprinted in: J. Malay. Br. Asiat. Soc. (1952) 25, 174-191. Armstrong, P. 1991. Under the blue vault of heaven: a study of Charles Darwin’s sojourn in the Cocos (Keeling) Islands. Indian Ocean Centre for Peace Studies, University of Western Australia. Berry, P.F. 1989. Survey of the Marine fauna of Cocos (Keeling) Islands, Indian Ocean. Unpublished report to the Australian National Parks and Wildlife service, 133pp. Birch, E. W. 1866. The Keeling Islands. Proc. Roy. Geog. Soc. N.S. 8: 263-265. Bleeker, P. 1855. Derde bijdrage tot de kennis der ichthyologische fauna van de Kokos- eilanden. Natuurk. Tijd. voor Ned. Ind. 7: 353-358. Chancellor, G., DiMauro, A., Ingle, R., and King, G. 1988. Charles Darwin's Beagle collections in the Oxford University Museum. Arch. Nat. Hist. 1988: 197-231. Clark, A. H. 1912. Crinoids of the Indian Ocean. Echinoderma of the Indian Museum, part VII, Crinoidea. Calcutta: Clark, A. H. 1950. Echinoderms from the Cocos-Keeling Islands. Bull. Raffles Mus: 22: 53-67. Colin, P. L. 1977. The reefs of Cocos-Keeling atoll, eastern Indian Ocean. Proc. 3rd Int. Coral Reef Symp. 1: 63-68. Darwin, C. 1842. The structure and distribution of coral reefs. London, Smith, Elder and Co. Darwin, C. 1845. Journal of researches into the natural history and geology of the countries visited during the voyage of H.M.S. Beagle round the world, under the command of Capt. Fitzroy R.N. London, John Murray. Fitzroy, R. 1839. Narrative of the surveying voyages of His Majesty's ships Adventure and Beagle, between the years 1826 and 1836, describing their examination of the southern shores of South America, and the Beagle's circumnavigation of the globe. London, H. Colburn. Forbes, H. O. 1879. Notes on the Cocos or Keeling Islands. Proc. Roy. Geog. Soc. 1: 777-784. Forbes, H. O. 1885. A Naturalist's wanderings in the Eastern Archipelago. A narrative of travel and exploration from 1878 to 1883. London, Sampson Row. Forest, J. 1956. La faune des Iles Cocos-Keelings, Paguridea. Bull. Nat. Mus. Singapore, 27: 45-55. Gibson-Hill, C. A. 1947. Notes on the Cocos-Keeling Islands. J. Malay. Br. Roy. Asiat. Soc. 20: 140-202. Gibson-Hill, C. A. 1948. The island of North Keeling. J. Malay. Br. Roy. Asiat. Soc. 21, 68-103. Gibson-Hill, C. A. 1949. The birds of the Cocos-Keeling Islands (Indian Ocean). Ibis 91: 221-243. Gibson-Hill, C. A. 1950a. Hemiptera collected on the Cocos-Keeling Islands, January- October 1941. Bull. Raffles Mus. 23: 206-211. Gibson-Hill, C. A. 1950b. The Myriapoda found on the Cocos-Keeling Islands January- October 1941. Bull. Raffles Mus. 22: 103-104. Gibson-Hill, C. A. 1950c. A note on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: Gibson-Hill, C. A. 1950d. A note on the reptiles occurring on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 206-211. Gibson-Hill, C. A. 1950e. Notes on the birds of the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 212-270. 10 Gibson-Hill, C. A. 1950f. Notes on the insects taken on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 149-165. Gibson-Hill, C. A. 1950g. Papers on the fauna of the Cocos-Keeling Islands. Based on material and data collected in the group by C.A. Gibson-Hill, M.A., between December 1940 and November 1941. Introduction. Bull. Raffles Mus. 22: 7-10. Gibson-Hill, C. A. (Editor). 1953. Documents relating to John Clunies Ross, Alexander Hare and the settlement on the Cocos-Keeling Islands. J. Malay. Br. Roy. Asiat. Soc. 25: 1-306 Guppy, H. B. 1889. The Cocos-Keeling Islands. Scott. Geog. Mag. 5: 281-297, 457- 474, 569-588. Guppy, H. B. 1890a. The dispersal of plants as illustrated by the flora of the Keeling or Cocos Islands. J. Trans. Vict. Inst. London 24: 267-306. Guppy, H. B. 1890b. Preliminary note on the Keeling Atoll. Proc. Vict. Inst. London 23: 72-78. Henslow, J. S. 1838. Florula Keelingensis. An account of the native plants of the Keeling Islands. Mag. Nat. Hist. 1: 337-347. Holloway, J. D. 1982. On the Lepidoptera of the Cocos-Keeling Islands in the Indian Ocean, with a review of the Nagia linteola complex (Noctuidae). Entomologia Gen. 8: 99-110. Holman, J. 1840. Travels in China, New Zealand. (2nd ed.). London. Hughes, J. S. 1950. Kings of Cocos: the story of the settlement on the atoll of Keeling- Cocos in the Indian Ocean. London: Methuen. Jenyns, L. 1842. The zoology of the voyage of H.M.S. Beagle, under the command of Captain Fitzroy R.M. during the years 1832 to 1836. Part IV. Fish. London: Smith, Elder. Maes, V. 1967. The littoral marine mollusks of Cocos-Keeling Islands (Indian Ocean). Proc: Acad: Nat? Sct., Phila 11993-2135. Marrat, F. P. 1879. Notes on shells from the Keeling or Cocos Islands, Indian Ocean. Proc. Lit. Phil. Soc. Liverpool, 33: ili-iv. Marshall, N. B. 1950. Fishes from the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 166-205. Owen, E. W. C. R. 1831. Account of the Cocos, or Keeling Islands. J. Roy. Geog. Soc. 1: 66-69. Randall, J. E. 1975. A revision of the Indo-Pacific angelfish genus Genicanthus, with descriptions of three new species. Bull. Mar. Sci. 25: 393-421. i} al Rees, W. J. 1950. The cephalopods of the Cocos-Keeling Islands (Indian Ocean). Bull. Raffles Mus. 22: 99-100. Ridley, S. O. 1884. On the classificatory value of growth and budding in the Madreporidae, and on a new genus illustrating this point. Ann. Mag. Nat. Hist. (Sth series). 13: 284-291. Ridley, S. O., and Quelch, J. J. 1885. List of corals collected in the Keeling Islands. In H. O. Forbes (Eds.), A naturalist's wanderings in the Eastern Archipelago (pp.44- 47.). London, Sampson Row. Ross, J. C. 1836. On the formation of the oceanic islands in general, and of the coralline in particular. Singapore Free Press, 2 June 1836. reprinted in J. Malay. Br. Roy. Asiat. Soc. (1952). 25: 251-260. Ross, J.C. 1855. Review of the theory of coral formations set forth by Ch. Darwin in his book entitled: Researches in Geology and Natural History. Natuur. Tijds. voor Neder. Ind. 8: 1-43. Simpson, C.J., Cary, J.L., and Masini, L.J. in press. Destruction of corals and other reef animals by coral spawn slicks on Ningaloo Reef, Western Australia. Coral Reefs. Slocum, J. 1899. Sailing alone around the World. London, Sampson Row, Marston. Smith-Vaniz, W. F., & Randall, J. E. 1974. Two new species of angelfishes (Centropyge) from the Cocos-Keeling Islands. Proc. Acad. Nat. Sci. Phil. 126: 105-113. Tweedie, M. W. F. 1950. The fauna of the Cocos-Keeling Islands, Brachyura and Stomatopoda. Bull. Raffles Mus. 22: 105-148. Van der Jagt, H. 1831. Beschrijving der Kokos - of Keeling-Eilanden. Verh. Batav. Gen. v. Kunsten en Wetenschappen (Batavia), 13, 293-322. translated in. J. Malay. Br. Roy. Asiat. Soc. (1952), 25, 148-159. Vaughan, T. W. 1918. Some shallow-water corals from Murray Island, Cocos-Keeling Islands, and Fanning Island. Carnegie Inst. Washington, Pub. 213: 49-234. Wallace, A. R. 1902. Island life, or, the phenomena and causes of insular faunas and floras: including a revision and attempted solution of the problem of geological climates (3rd ed.). London, Macmillan. Wells, J. W. 1950. Reef corals from the Cocos-Keeling atoll. Bull. Raffles Mus. 22: 29- Se Williams, D. G. 1990. An annotated bibliography of the natural history of the Cocos (Keeling) Islands, Indian Ocean. Atoll Res. Bull. 331: 1-17. Wood-Jones, F. 1909. The fauna of the Cocos-Keeling Atoll, collected by F. Wood- Jones. Proc. Zool. Soc. 1909: 132-160. Wood-Jones, F. 1912. Coral and Atolls: a history and description of the Keeling-Cocos Islands, with an account of their fauna and flora, and a discussion of the method of 12 development and transformation of coral structures in general. London, Lovell Reeve and Co. Woodroffe, C. D., McLean, R. F., Polach, H., & Wallensky, E. 1990a. Sea level and coral atolls: Late Holocene emergence in the Indian Ocean. Geology 18: 62-66. Woodroffe, C. D., McLean, R. F., and Wallensky, E. 1990b. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. Nat. Geog. Res. 6: 262-275. Date 1836 1854-5 1878 1905-6 1940-1 1963, 1974 1961 1985 1986/7 1989 Expedition/ Group Collectors Collected H.M.S. Beagle crustacea fishes plants shells A.J. Anderson and fishes G. Clunies-Ross H.O. Forbes corals plants F. Wood-Jones brachyurans corals echinoderms plants C.A. Gibson-Hill fishes brachyurans stomatopods anomura cephalopods echinoderms molluscs corals birds Academy of Natural Sciences of Philadelphia molluscs fishes A.E. Alfred birds I.R. Telford plants D.G. Williams plants marine algae Western Australian coral Museum echinoderms fishes molluscs barnacles 13 Table 1. Summary of collections of flora and fauna from the Cocos (Keeling) Islands Publication Darwin 1845, Chancellor et al. 1988 Jenyns 1842 Henslow 1838 Marrat 1879 Bleeker 1855 Guppy 1889, Ridley 1884, Ridley and Quelch 1985 Forbes 1885 Wood-Jones 1909 Vaughan 1918 Clark 1912 Wood-Jones 1912 Marshall 1950 Tweedie 1950 Tweedie 1950 Forest 1956 Rees 1950 Clark 1950 Abbott 1950 Wells 1950 Gibson-Hill 1949, 1950e Maes 1967 Randall 1975, Smith-Vaniz and Randall 1974 Alfred 1961 Flora of Australia This volume Berry 1989, This volume decapod crustaceans 14 Table 2. Number of species recorded at Cocos (Keeling) and Christmas Islands. GROUP Cocos Is. Christmas Is. Reef-building coral 99 85 Decapod crustaceans 198 204 Molluscs c 610 c 490 Echinoderms 88 90 Fishes C550 568 Native birds 38 88 Plants 130 386 15 . ORK KR HS SOK KOK SKK xO oe OOH 60.0.0 en Trench PKAOS © Me © oN e . ee ONa aoe PROROS, CHRISTMAS IS. *“cocos (KEELING) IS 2 SR ROOD E ~. OOD SOOO? . RRR S56 258 SOI KX "a, (s) > fo) co) (>) MALDIVES, .= S MAURITIUS yMADAGASCAR A> SELL .> ALDABRA ye K? xx LS <> SKS o ‘ Sx x We, OL OL Ox QQ 5 OPO LQ GS 2 SS % x sssteiestatorsssstecoetniestcton sorsenesioneestenmnnennee ene The Indian Ocean, showing the location of the Cocos (Keeling) Islands. Figure 1. NORTH KEELING ISLAND 0 kilometres 2 all cocos (KEELING) ! ISLANDS OCEAN LOCALITY DIAGRAM INDIAN i “i ° PRISON 30", ISLAND Turk Reet ja 15 HOME ISLAND COCOS PULU KEMBANG © PULU CEPELOK ¢ PULU WAK BANKA PULU PANDAN (KEELING) PULU SIPUT =~ iL ose Y\ WEST ISLAND PULU sel ll kilometres Figure 2. The Cocos (Keeling) Islands, showing stations from which collections were made during the Western Australian Museum expedition in 1989. These stations cover a series of different marine habitats (see Chapter 8, Fig. 1), which can be summarised as: Outer Reef Slope (9 sites: stations 4, 7, 13, 15, 19, 22, 25, 32, and 33), Reef Flat (13 sites: stations 1, 3, 6, 8, 10, 11, 12, 14, 20, 21, 24, 27, and 30) and lagoon (14 sites: stations 2, 9, 16, 17, 18, 23, 26, 28, 29, 31, 34, 35, 36 and 37). ATOLL RESEARCH BULLETIN NO. 400 CHAPTER 2 CLIMATE, HYDROLOGY AND WATER RESOURCES OF THE COCOS (KEELING) ISLANDS BY A.C. FALKLAND ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 2 CLMATE, HYDROLOGY AND WATER RESOURCES OF THE COCOS (KEELING) ISLANDS BY A.C. FALKLAND * CLIMATE AND HYDROLOGY GENERAL FEATURES The Cocos (Keeling) Islands are situated in the Humid Tropical zone. For most of the year they are under the influence of South East Trade Winds. Cyclonic conditions are sometimes experienced, particularly between November and March. Rainfall on the island is influenced to some extent by El Nifio events. The main climatic features are: - annual rainfall varying between about 850 and 3300 mm, - annual potential evaporation of about 2000 mm, - relatively uniform temperatures, ranging from about 18°C to 32°C, - relative humidity varying from about 65% to 84%, - daily atmospheric pressures ranging from 973 to 1018 hectopascals, and - mean daily wind speeds varying from 4.7 and 8.1 metres/second with a maximum gust during a cyclone recorded at 48.8 metres/second (176 kilometres/hour). METEOROLOGICAL DATA CURRENT NETWORK A meteorological station (No. 200284, Cocos Island A.M.O.) has been operated continuously on West Island on the eastern side of the airstrip by the Bureau of Meteorology (Australia) since February 1952. It is located at latitude 12°11’S, longitude 96°50’E and at an altitude of 3 metres. At the station, the following meteorological parameters, important to water resources assessment, are measured and recorded: - air temperature (wet and dry bulb, and dew point), - atmospheric pressure, - cloud cover, - wind speed and direction, - rainfall, and - pan evaporation. Temperature, atmospheric pressure, cloud cover and wind are measured every 3 hours. Daily averages can be derived from eight readings. Daily total wind run is also * Hydrology and Water Resources Branch, Australian Capital Territory Electricity and Water, P.O. Box 366, Canberra, Australian Capital Territory, 2601. 2 recorded. Rainfall and pan evaporation are measured daily at 9 a.m. Rainfall is also recorded on a pluviograph (continous recorder). In addition to the above surface level measurements, upper air data is collected via regular balloon releases from the station. Daily rainfall is measured and recorded at two other sites on the South Keeling atoll. The first site is located on a peninsular about 100 metres north of the jetty on Home Island. Data has been recorded at this site by the Cocos (Keeling) Islands Council (formerly Home Island Council) since 28 May 1986. This site is now described as station number 200733 by the Bureau of Meteorology. The second site is located on the eastern side of the administration building at the Quarantine Station on West Island. Data has been recorded at the second site by Quarantine Station personnel since 1 January 1989. Some of the data at both of these sites has not been recorded each day but rather recorded as a total for two or three days. Therefore, these records cannot be used for accurate daily rainfall analyses but they are suitable for monthly rainfall analyses. No meteorological data is recorded on the North Keeling atoll. PREVIOUS DATA COLLECTION SITES In 1904 a rainfall station was opened on Direction Island and operated by staff at the cable station. A climatological station was opened "probably soon after, although no files survive to prove this" (Bureau of Meteorology 1978). Other notes in the file entry indicate that this station operated to 1952 when the current station opened on West Island. There is some anecdotal evidence that the Clunies-Ross family recorded rainfall on Home Island. However, no records were sighted and it is not known whether any of these records were incorporated into the Bureau's records. As the rainfall records extend back to December 1901, it is probable that rainfall was recorded on Home Island from then until the rainfall station on Direction Island was opened in 1904. Rainfall records under the heading 'Cocos Island Composite’ are available from the Bureau of Meteorology for all but 17 months from December 1901 to the present. Months with missing data are November and December 1914, all months of 1915, January 1916, April 1946 and February 1952. An early entry in a Bureau of Meteorology file states that "the records for 1915 were incomplete on account of the instruments being destroyed in November 1914 by German warship 'Emden'. Records were recommenced in November 1915". The missing data in 1946 and 1952 appears to be due to staff and location changes. Overall, the length of record and the small number of missing months (and hence days) of record, have ensured a very good data set for the Cocos (Keeling) Islands. SELECTED METEOROLOGICAL DATA FROM WEST ISLAND Graphical and tabular summaries of important meteorological parameters recorded at the West Island meteorological station are presented in this and the next two sections. Summaries of temperature, relative humidity (a derived parameter), atmospheric pressure, cloud cover, wind speed and wind direction are presented in this section. More extensive summaries of rainfall and evaporation are provided below. The data used for these summaries, unless otherwise indicated, are for the period February 1952 to December 1991. TEMPERATURE Mean, maximum and minimum daily temperatures are shown in Figure 1 and in Table 1 for each month of the year. The maxima and minima are extreme values derived from all of the 3 hourly data. The mean values were estimated by averaging mean daily maxima and minima. The mean daily temperature is highest in March (27.5°C) and lowest in July and August (25.8°C). The extreme maximum temperature is 32.4°C recorded in February 1979 and the extreme minimum temperature is 18.3°C recorded in August 1979. RELATIVE HUMIDITY Mean, maximum and minimum daily relative humidities are shown in Figure 2 and in Table 2, for each month of the year. These values are derived from 9 a.m. and 3 p.m. readings of wet and dry bulb temperature and atmospheric pressure using standard meteorological methods. The mean daily relative humidity is highest from April to July (77%) and lowest from September to December (72%). The mean daily maximum relative humidity is 84% recorded in the months of June and July 1960, April and August 1973, and April 1974. The mean daily minimum relative humidity is 65% recorded in November 1956. Extreme maxima and minima were not computed. ATMOSPHERIC PRESSURE Mean, maximum and minimum daily atmospheric pressures are shown in Figure 3 and in Table 3, for each month of the year for the period February 1952 to July 1987. The shorter period was used as extreme maxima and minima had not been computed for the full data set. The maxima and minima are extreme values for each month and, as with the mean values, they are derived from the full 3 hourly data set. The mean daily atmospheric pressure is highest in September (1012.2 hectopascals) and lowest in February (1008.4 hectopascals). The extreme maximum atmospheric pressure is 1018.5 hectopascals recorded during the months of October 1952 and July 1984 and the extreme minimum atmospheric pressure is 970 hectopascals recorded during cyclone 'Doreen' on 21 January 1968. It is noted that the atmospheric pressures referred to in this section are at the level of the station. The atmospheric pressure at mean sea level is obtained from these readings by adding an amount of less than one hectopascal. Due to this small difference and the low altitude of the islands, the atmospheric pressure at the station can be used as an indicator of atmospheric pressure throughout the islands. CLOUD COVER Cloud cover (or cloudiness) is measured in oktas or the number of eigths of the sky filled with cloud. Mean, maximum and minimum daily values for each month of the year are shown in Figure 4 and Table 4. Data are derived from all observations for the period 1952 to 1987. The mean daily cloud cover varies between 5.0 and 5.3 oktas, from April to July and 4 oktas for the other 8 months. The mean daily maximum cloud cover is 6.8 oktas recorded in November 1973 and the mean daily minimum cloud cover is 3 oktas recorded in December 1987. Extreme maxima and minima were not computed. WIND SPEED Mean, maximum and minimum wind speed are shown in Figure 5 and in Table 5, for each month of the year. The mean values are taken from the 3 hourly data set from February 1952 to July 1987. The maxima and minima are extreme instantaneous values for each month for the full period of record (February 1952 to December 1991). The mean daily wind speed is highest in January (8.1 metres/second) and lowest in February (4.6 metres/second). The mean daily maximum wind speed is highest in August (14.2 metres/second) and the mean daily minimum wind speed is lowest in March (just above zero). The extreme maximum wind speed was recorded at 48.8 metres/second during cyclone 'Doreen' on 21 January 1968. This wind speed is equivalent to 95 knots or about 175 kilometres/hour. The extreme minimum wind speed is zero (calm) which has been recorded on many occasions during all months of the year. WIND DIRECTION Figure 6 shows the wind direction resolved as a percentage of time for each month (January to June in left hand graph and July to December in right hand graph). Eight points of the compass are used. The graphs are for average wind directions from 9 a.m. and 3 p.m. readings for the period February 1952 to December 1990. The legends show the percentage of time that calm periods were recorded for each month . The predominant wind direction is east to south east for all months, showing the influence of the South East Trade Winds on the islands. South easterly winds varied from 37% of the time in February to 60% of the time in November and December. Easterly winds varied from 17% in January to 44% in September. By comparison winds from the north, north east, south west, west and north west occurred for less than 6% of the time in all months and were often 2% or less. Southerly winds were experienced from a low of 2% of the time in September and October up to 17% of the time in January. OCCURRENCE OF CYCLONES A cyclone database maintained by the Bureau of Meteorology shows that a number of cyclones have affected the Cocos (Keeling) Islands. Table 6 presents data about cyclones since 1959 which have passed within approximately 100 kilometres of the island. One of the most damaging cyclone appears to have been ‘Doreen’ which passed directly over the South Keeling atoll. An interesting account of this cyclone is provided in Ryan (undated). RAINFALL DATA AND ANALYSES INTRODUCTION As rainfall is one of the most important determinants of the water resources of the islands, a more detailed analysis of rainfall is presented. This section describes characteristics of the annual, monthly and daily rainfall. In addition to analyses of the temporal distribution of rainfall, comments about the spatial distribution of rainfall on the South Keeling atoll are made. It is noted that from a water resource viewpoint, rainfall at time scales of days and months are of most significance. For recharge analysis, as part of groundwater studies on atolls such as South Keeling, daily rainfall data has been found to have sufficient time resolution. Monthly rainfall data can be used instead of daily rainfall data with a small loss in accuracy. For rainwater catchment studies, long sequences of daily rainfall data are ideal. SPATIAL VARIATION OF RAINFALL At the scale of the Indian Ocean, the variation of rainfall has been reported in Stoddart (1971) and Unesco (1977). Stoddart (1971) reviewed earlier reports and produced updated isohyetal maps for annual, seasonal and monthly rainfall based on coral island stations, primarily those with records longer than 10 years. The mean annual rainfall distribution is shown in slightly modified form in Figure 7. The isohyetal map of mean annual rainfall shown in Unesco (1977), not shown here, is quite different particularly in the centre of the ocean. In the region of the Cocos (Keeling) Islands, however, the two maps are similar with isohyets approximately horizontal. The reason for the differences is not clear. On the South Keeling atoll, the spatial variation rainfall has been analysed by Falkland (1992a). ANNUAL RAINFALL The mean annual rainfall recorded at the West Island meteorological station is 1954 mm for the 40 year period of record from 1953 to 1992. Using the available annual record from 1902 to 1992, the mean annual rainfall is 1982 mm. In the longer data set, the annual rainfalls in 1914, 1915, 1916, 1946 and 1952 are missing giving a total number of 86 years. For the period 1902-1952, during which time the rainfall was recorded primarily on Direction Island, the mean annual rainfall is 2006 mm. Figures 8 and 9 show, respectively, histograms of annual rainfall for the periods 1902 to 1952 and 1953 to 1992. Using the meteorological station records, the highest annual rainfall on record is 3291 mm which occurred in 1942 (Direction Island) while the lowest is 856 mm in 1991 (West Island). By comparison, the annual rainfall in 1991 on Home Island and at the Quarantine Station were, respectively, 837 and 820 mm. The difference in the mean annual rainfalls for the two periods 1902-1952 and 1953-1992 is 2.6% of the latter period mean annual rainfall. This is considered a minor difference, given that site differences can easily account for long term rainfall depth differences of 10% or more. This result shows that the total depth of rainfall in the first 6 half of the century (1902-1951), recorded primarily on Direction Island, is similar to that recorded in the second half of the century (1953-1991 ) on West Island. The standard deviations of the annual rainfalls for the periods 1902-1992 and 1953-1992 are, respectively, 519 and 594 mm, showing a higher variability in the second half of the century. The coefficient of variation (Cv) of annual rainfall (obtained by dividing the standard deviation by the mean) for the two periods are, respectively, 0.26 and 0.3, again indicating the higher variability of recorded rainfall in the second part of this century on the island. These Cv's of annual rainfall are moderate when compared with other islands, especially low lying coral atolls. Christmas Island (Australia), a raised limestone island, about 900 kilometres north east of the Cocos (Keeling) Islands, has a similar Cv of 0.29. The atolls of Tarawa and Kiritimati (Christmas Island), Republic of Kiribati in the Pacific Ocean have higher Cv's (0.42 and 0.64, respectively). By comparison, Kwajalein atoll in the Marshall Islands Pacific Ocean has a much lower Cv of 0.14 (Falkland et al. 1991). The variation in annual rainfall between the three raingauge sites can be seen in Table 7. From the data in Table 7, the rainfall for 1987 on Home Island is suspect as it is very low (only 65% of that at the meteorological station) and is inconsistent with the relative rainfall pattern between Home and West Island for the following years. Disregarding the suspect 1987 data for Home Island, there is slightly less rain occurring on Home Island than on West Island. The total rainfall recorded on Home Island from 1988 to 1992 is 5.3% less than at the meteorological station. The data also shows that the rainfall recorded on Home Island and at the Quarantine Station is, respectively, 3.4% less and 5.3% greater than at the meteorological station during the four year period 1989 to 1992. Overall, the variation of rainfall between the three sites is not greatly significant when it is considered that the recording accuracy of rainfall at any one site is generally not better than about 10%. Although the period of concurrent rainfall records is short and therefore not suitable for making long term predictions, it is reasonable to conclude that the annual rainfall on the South Keeling atoll can be adequately described by the rainfall record at the West Island meteorological station. MONTHLY RAINFALL Mean, maximum and minimum monthly rainfalls recorded at the West meteorological station are shown in Figure 10 and in Table 8, for the March 1952 to December 1991. The mean monthly rainfall is highest in April (234 mm) and lowest in October (70 mm). The maximum monthly rainfall is 649 mm recorded in June 1988 and the minimum monthly rainfall is 2.8 mm recorded in September 1986. A comparison of the monthly rainfalls recorded at the meteorological station and on Home Island for the period 1987 to 1991 is shown in Figure 11. A comparison of the meteorological station and the Quarantine Station monthly rainfalls for the period 1989 to 1991 is shown in Figure 12. A comparison of the cumulative monthly rainfall recorded at the three sites for the period of concurrent records (1989-1991) is shown in Figure 13. Figures 11 and 13 indicate that the rainfall recorded at the Home Island site is less than that at the meteorological station. The lower rainfall recorded on Home Island may be due to a rain shielding effect of nearby tall vegetation at the raingauge site. Similarly, Te Figures 12 and 13 show that the rainfall recorded at the Quarantine Station site is slightly greater than that recorded at the meteorological station. Double mass curves using cumulative monthly rainfall were plotted to check if any changes in the relative rainfall at the Home Island and the Quarantine Station sites had occurred during the periods of concurrent record. Figure 14 is the double mass curve using the rainfall data from the meteorological station and Home Island for the period 1987-1991. The plotted line shows some variation in slope, particularly from the early (1987) data to later data. Figure 15 is the double mass curve using the rainfall at West Island meteorological station and at the Quarantine Station for Home Island for the period 1989-1991. The corresponding Home Island data is also plotted in Figure 15. Very little variation in slope is shown for the Quarantine Station curve, indicating that there has been no major changes in this site or the method of recording since data collection commenced. However, the Home Island data indicates a greater variation. The variations in the Home Island record could be due to a number of reasons including progressive ‘shading’ of the raingauge from nearby trees, and errors or changes in the method of reading and recording rainfall data. Regression analyses of monthly rainfalls at Home Island and the Quarantine Station with the meteorological station are summarised in Figures 16 and 17. The analysis using 60 monthly rainfall pairs for Home Island and the meteorological station gave a correlation coefficient (r) of 0.92 which indicates a reasonably good correlation. The value of r for 36 pairs of data from the Quarantine Station and the meteorological station was 0.97 which indicates a very good correlation. The regression equations are shown in Figures 16 and 17, respectively. These could be used to estimate monthly rainfalls at the two sites from the meteorological station monthly rainfall. They should be updated with additional data if they are considered for future use. The Home Island equation should be treated with caution as some of the Home Island data is suspected of being in error. DAILY RAINFALL While daily rainfall records are available from the Bureau of Meteorology for the full period of record (1904-1992) from the stations on Direction Island and West Island, only the data from the latter station were obtained for analysis. Daily rainfall has been recorded at the West Island meteorological station from 15 February 1953 to the present. Daily rainfall has also been recorded on Home island (28 May 1986 to present) and at the Quarantine Station (1 January 1989 to present), as described previously. Daily data from all 3 stations was reviewed to the end of 1992. The maximum recorded daily rainfall at the West Island meteorological station is 287 mm on 28 August 1956. There have been 54 days when the rain exceeded 100 mm and 6 days when it exceeded 200 mm. The maximum daily rainfall on Home Island, 242 mm, was recorded on 11 November 1989. On the same day the rainfalls recorded at the meteorological station and the Quarantine Station were, respectively, 203 mm and and 220 mm. The maximum daily rainfall at the Quarantine Station was 248 mm, recorded on 4 July 1992. On the same day the rainfalls recorded at the meteorological station was 252 mm. No daily rainfall was recorded at Home island on 4 July 1992 but the three day total to 6 July was only 141 mm. On a daily basis, the rainfall records show considerable variation between the three rainfall recording sites, as some of the above results show. This is confirmed by general 8 observations that individual storms can affect small areas of the atoll while leaving other areas quite dry. Hence, in the short term the rainfall pattern on Home island or elsewhere on West Island cannot necessarily be deduced from the West Island meteorological station records. Daily variability can also be seen from a number of high rainfall days in 1990 and 1992 when all three rain gauges were operational. The list below shows, in order, the date and the rainfalls at the meteorological station, Home Island and the Quarantine Station: 12 January 1990: 162, 164 and 171 mm, 14 April 1990: 117, 83 and 116 mm, 17 July 1990: 158, 210 and 115 mm, 6 September 1990: 30, 23 and 116 mm, 28 February 1992: 26, 112 and 87 mm, 14 April 1992: 6, 102 and 8 mm, and 4 May 1992: 11, 122 and 29 mm. The longest period without any rainfall at the meteorological station is a period of 28 days in November 1985. The longest period when the total rainfall was less than 10 mm occurred between November 1985 and January 1986 when only 6.2 mm fell in 69 days. Long dry periods are of particular interest in the study of the island's water resources, as described below. PLUVIOGRAPH RECORDS A pluviograph (continuous rainfall recorder) is operated at the West Island meteorological station. These records enable rainfall patterns to be analysed (at time resolution in minutes). Such data are useful to analyse storm events and to construct rainfall intensity-frequency-duration (IFD) curves, for possible use in the design of stormwater facilities (e.g. roof gutters and downpipes). The Bureau of Meteorology has processed IFD information from pluviograph records between 1971 and 1991. This information is not presented here. EVAPORATION DATA AND ANALYSES INTRODUCTION Estimation of actual or catchment evaporation is essential for any water resources study. Evaporation from a catchment includes evaporation from soil, water and other open surfaces such as paved areas and from the leaves of grasses, plants and trees. Evaporation from the stomates of leaves is called transpiration and the combined effects of this process and other evaporation is often described as evapotranspiration. The two processes are basically variations of the one process, namely, the conversion of water from a liquid to a gaseous state and some authors use the term evaporation instead of evapotranspiration. The term evaporation will normally be used instead of evapotranspiration for present purposes. The estimation of actual evapotranspiration (ET4) is generally done as a two stage process. Firstly, ETp is estimated using a method based on meteorological data, such as the Penman (or Contbination) formula (Penman 1948, 1956), or from pan evaporation data multiplied by appropriate pan coefficient(s). The Penman equation has generally been found to be a good ETp estimation method in the humid tropics (Fleming 1987). Estimations using both Phe pan and Penman methods were made for the study of 9 groundwater resources on the South Keeling atoll (Falkland 1988). Secondly, ET, is determined using a water balance procedure taking into account the soil and vegetation conditions present on the island. The estimation of ETp, using both pan evaporation records and the Penman approach, is described below while the estimation of ETg is described in a later section on water balance. PAN EVAPORATION DATA Daily pan evaporation has been recorded at the West Island meteorological station using a U.S. Class A pan from December 1981 to the present. Mean, maximum and minimum monthly pan evaporation totals are shown in Figure 18 and Table 9, for the period January 1982 to December 1991. The mean monthly pan evaporation is highest in December (241 mm) and lowest in June (171 mm). The maximum monthly pan evaporation is 273 mm recorded in both December 1983 and December 1985. The minimum monthly pan evaporation is 146 mm recorded in May 1987. EVAPORATION ESTIMATION (Penman equation) The following meteorological parameters were available for use in the Penman equation: - dry bulb temperature, - wet bulb temperature, = dew point temperature, - cloud cover, and - wind speed. Using mean monthly values of the parameters above, estimates of monthly ET, were made using the Penman equation (Penman 1948, 1956) for the period January 198 to March 1986. This period was the longest period of available concurrent data at the time of investigations (Falkland 1988). Water balance simulations (Falkland 1988) showed that similar results in terms of groundwater recharge were obtained from monthly data sets using either actual or mean values of ETp. This shows the relatively constant nature of potential evaporation for a given month from year to year in a humid tropical environment such as the Cocos (Keeling) Islands. In the humid tropics, the net radiation energy term dominates the aerodynamic term in the Penman equation and it has been found that the simplified Priestly-Taylor method can also be used (Chang 1989). In the Priestley-Taylor method ETp is equated to 1.26 times the energy term from the Penman equation (Priestley and Taylor 1972). EVAPORATION ESTIMATION (pan method) Pan evaporation requires multiplication by an appropriate pan coefficient to obtain estimates of ET). An initial estimate of the pan coefficient of between 0.7 and 0.75 was 10 obtained by a procedure developed by Doorenbos and Pruitt (1977) using meteorological and specific site parameters. The pan coefficient was later adjusted to 0.8 after sensitivity analyses were conducted with trial data using water balance simulations. The water balance results in terms of recharge to groundwater using five years of rainfall data (1982 to 1986) were found to be very similar for Penman estimates of ET, and for pan data using a pan coefficient of 0.8 (Falkland 1988). Results were also similar for simulations using actual and mean monthly pan data. Figure 19 shows the comparison of mean monthly ET» estimates using both the Penman and pan methods for the five year period 1982 to 1986. The pan estimates are mean values for each month. The Penman estimates are based on mean monthly values of the relevant meteorological parameters. The mean annual ET, based on the pan method was 1983 mm compared with the annual ET, of 2048 mm based on the Penman method. This difference of about 3% is insignificant for practical purposes. As the results from the two methods are very similar, the pan method using mean monthly data was adopted as ETp estimates could be more easily computed with this method. Later studies (Falkland {991, 1992a) used mean monthly pan evaporation data for the period 1982 to 1987. Recently, additional pan data to December 1991 was obtained and the mean monthly estimates of ET, for the periods 1982-1987 and 1982-1991 were compared. The results are very similar. The mean annual values are in fact only 1 mm different (1986 and 1987 mm for the shorter and longer periods, respectively). TRANSPIRATION MEASUREMENTS At the commencement of detailed water resources investigations in 1987, it was realised that coconut trees (Cocos nucifera), prolific on most atolls including the Cocos (Keeling) Islands, are a major source of transpiration and, hence, loss from freshwater lenses. Direct measurements of coconut tree transpiration were, therefore, undertaken during the study. Due to time limitations, lysimeter or ventilated chamber methods could not be used. Instead, measurements were undertaken using a heat pulse velocity meter. The meter and its associated electronic data logger measures and records the velocity of an injected heat pulse in the sapwood of a tree by timing movement over a known distance. The technique had been used successfully on other types of trees but never, to the author's knowledge, on coconut trees. The results obtained from the one-week study suggested that transpiration rates per tree varied from about 70 to 130 litres/day (Bartle 1987). The range of values was considered to be the result of diurnal climatic variations. The values obtained must be considered preliminary owing to a number of simplifying assumptions and the short period of observations. Further study over a longer time period is warranted as part of general scientific research. Based on this limited data, the total transpiration rate due to coconut trees is about 400-750 mm per year per tree in areas with 100% tree cover, where typical tree spacings of about 8 metres prevail. This has implications for water resources management and it may be prudent to selectively clear coconut trees from some freshwater lens areas to maximise the supply of water. 11 INFLUENCE OF EL NINO ON THE CLIMATE Considerable research has been undertaken into the influence of the El Nifio phenomenon (also called the El Nifio Southern Oscillation or ENSO) on climatic patterns, particularly in the Pacific Ocean. Effects of strong El Nifio events in the Pacific Ocean include significant sea surface temperature changes, ocean current and wind direction reversals, extreme variations in rainfall patterns, higher tides, storm activity in some locations and severe droughts in others. The influence of the El Nifio phenomenon is felt more widely than just the Pacific Ocean. Some research has been conducted into the connections between El Nifio events and the weather patterns occurring in the north-eastern Indian Ocean area around Indonesia. Quinn et al. (1978) studied the connections between El Nifio events and droughts in Indonesia. Their general conclusion was that droughts in Indonesia, indicated by low rainfall periods on Java, occurred in years when El Nifio events were evident. A significant connection between low rainfall years and El Nifio events was found for Christmas Island in the Indian Ocean (Falkland 1986). The influence of El Nifio events on the rainfall of the Cocos (Keeling) Islands is outlined in Falkland (1988, 1992a). A graph showing the relationship between the Southern Oscillation Index (an index of the strength of ENSO activity) on an annual basis and annual rainfall (expressed as a percentage of mean annual rainfall) is shown in Figure 20 for the period 1953 to 1991. Negative values of SOI are associated with El Nifio activity with the more negative values indicating increased strength. Positive values indicate that El Nifio activity is absent. Figure 20 shows that there is a reasonable correlation between SOI and annual rainfall, with negative annual SOI values corresponding in general with less than average rainfall and vice versa. This trend is not always present, an example being the highly negative SOI during the 1982/83 El Nifio when the rainfall was near average. Using linear regression analysis between annual SOI and rainfall data, a correlation coefficient of only 0.58 was obtained, indicating that the correlation is not strong. In certain periods (for example, 1953 to 1960, 1967 to 1981) the correlation is much better as can be seen in Figure 20 (r=0.89 and 0.82, respectively). It can be concluded that there is a reasonable correlation between El Nifio activity in the Pacific Ocean and rainfall in the Cocos (Keeling) Islands. WATER RESOURCES TYPES The water resources of the Cocos (Keeling) Islands consist essentially of groundwater and rainwater. Where conditions are favourable, fresh groundwater occurs on coral islands in the form of shallow freshwater lenses. Such lenses are found in some of the larger islands within the Cocos (Keeling) Islands. The groundwater from these lenses has been and is currently used as the major source of freshwater for potable and other uses on Home and West Islands. Due to the generally porous nature of the soils and underlying geology, there is no significant surface runoff. Runoff only occurs in localised areas where the ground is compacted or paved and only for very short periods after heavy rain. Rainwater collected directly from roofs of buildings is a valuable supplementary source of water. 12 GROUNDWATER OCCURRENCE FRESHWATER LENS CHARACTERISTICS Freshwater lenses occur beneath the surface of some islands. The upper surface of a freshwater lens is the water table and the lower surface is a boundary between freshwater and saline water. The lower boundary is not a sharp interface but rather is in the form of a transition zone. Within the transition zone the water salinity increases from that of freshwater to that of seawater over a number of meters. A typical cross section through a small coral island showing the main features of a freshwater lens is presented in Figure 21. It must be noted that there is considerable vertical exaggeration in the diagram. In practice, the vertical scale is much smaller compared with the horizontal scale. The transition zone tends to be as thick as or thicker than the freshwater zone on many small coral islands. As shown in the diagram, there is often an asymmetric shape to the lens with the deepest portion displaced towards the lagoon side of the island. The salinity of the upper surface of a freshwater lens‘can be obtained by measurements at exposed water surfaces such as wells and pumping galleries. The lower surface can be determined accurately by establishing a recognisable salinity limit for freshwater and drilling through the lens and testing the water at different depths for salinity. It can also be estimated approximately by surface geophysical (electrical resistivity and electromagnetic) techniques. The salinity limit adopted for freshwater for the Cocos (Keeling) Islands is 600 mg/l chloride ion concentration. This limit is approximately equivalent to an electrical conductivity (specific conductance) reading of 2600 umhos/cm at the standard temperature of 25°C (Falkland 1988, 1992a). According to classical 'Ghyben-Herzberg' theory (Badon Ghyben 1889, Herzberg 1901), for every unit height of fresh water occurring above mean sea level there will be about 40 equal units of underlying fresh water below mean sea level. This theory assumes that the two fluids, freshwater and seawater, are immiscible (i.e. that they do not mix). In practice, the two fluids do mix due to mechanical and molecular diffusion and a transition zone forms with salinity gradually increasing from that of freshwater to that of seawater. In practical situations, the 1:40 ratio can be used as a guide to determine the mid-point of the transition zone from the water table elevation above mean sea level. It does not provide a means of determining the base of the freshwater zone and other methods described above are required. INFLUENCING FACTORS ON FRESHWATER LENSES The size and salinity distribution of freshwater lenses, particularly the thickness of freshwater and transition zones, are dependent on many factors but the most important are: - rainfall amount and distribution, - amount and nature of surface vegetation and the nature and distribution of soils (these factors influence the evapotranspiration), 13 - size of the island, particularly the width from sea to lagoon, - permeability and porosity of the coral sediments, and the presence of solution cavities, - tidal range, and - methods of extraction and quantity of water extracted by pumping. For small coral sand islands, an approximate relationship has been derived (Oberdorfer and Buddemeier 1988) between freshwater lens thickness, annual rainfall and island width as follows: H/P = 6.94 log a - 14.38 where an rl lens thickness (depth from water table to sharp interface or mid-point of transition zone in metres), P = annual rainfall (metres), and = island width (metres). This equation indicates that no permanent freshwater lens can occur regardless of rainfall where the island width is less than about 120 metres. Using the mean annual rainfall (1938 mm measured at the West Island meteorological station) for the Cocos (Keeling) Islands, the minimum island width for a small freshwater lens (say 5 metres thick) to occur is about 280 metres (say 300 metres). Thus, as an approximate guide, it is unlikely that a permanent freshwater lens suitable for groundwater extraction could be found on the South Keeling atoll where the width of the island is less than about 300 metres. It is noted, however, that other factors which are not accounted for in the above relationship, particularly the permeability of the coral sediments and the density of vegetation, have an effect on the occurrence of freshwater lenses. Further comments based on observed data on West Island are given later. The geological influences are considered in more detail below. GEOLOGICAL INFLUENCES ON FRESHWATER LENSES The geology of the South Keeling atoll consists of coral sediments, several hundreds of metres thick, overlying a volcanic seamount. From a hydrogeological viewpoint, the geology of most interest is that of the upper part of the atoll where freshwater lenses are found to occur. From a number of recent water investigations on the South Keeling atoll (Falkland 1988, 1991, 1992a, 1992b), freshwater lenses do not exceed 20 metres in thickness. Within this 20 metre zone, two major geological layers are found: a younger (Holocene), upper layer consisting of unconsolidated coral sediments and an older (Pleistocene), deeper layer of coral limestone. While no extensive investigations of surface geology have been undertaken on North Keeling atoll, it is expected that similar geological conditions would prevail there. Similar to findings on other atolls in the Pacific Ocean, an unconformity was found from drill cores between the relatively low permeability Holocene sediments and underlying higher permeability Pleistocene limestone at depths of less than 20 metres (Falkland 1988). Using the early results and data from additional boreholes on West Island, Home Island, South Island and Horsburgh Island, the unconformity was found at depths varying between about 8 and 17 metres below ground surface (Woodroffe et al. 1991). These depths correspond, respectively, to depths between 7 and 16 metres below mean sea level. 14 The presence of this unconformity is due to a period of emergence of the island with solution and erosion forming a karst surface. Uranium-series dating of the older limestone indicates that it was formed during the last inter-glacial period about 120,000 years ago (Woodroffe et al. 1991). The upper sediments have been laid down in the Holocene since about 10,000 years ago. Three phases of deposition have been identified in the Holocene (Woodroffe et al. 1990a, this volume). From the start of the Holocene to at least 5000 years ago, sediments accumulated rapidly as sea level rose. A conglomerate platform radio-carbon dated at 3000 to 4000 years ago was then formed during a period of relatively stable sea level. Since then unconsolidated sands and larger sediments have been deposited to form the present reef islands. Dating of in-situ corals has shown that the sea level was about 0.5 to 1.5 metres higher about 3000 years ago than today (Woodroffe et al. 1990a, 1990b). The unconformity described above is very significant to the formation of freshwater lenses. The limestone sediments below this unconformity have relatively high permeabilities and mixing of freshwater and seawater is readily facilitated. In the relatively less permeable upper sediments, mixing is less likely to occur. The unconformity, therefore, is one of the main controlling features to the depth of freshwater lenses. WATER BALANCE AND RECHARGE ESTIMATION RECHARGE The freshwater lenses in the Cocos (Keeling) Islands are recharged naturally from rainfall. Not all rainfall incident on the islands, percolates to groundwater, as much of it is evaporated or transpired. Essentially, natural recharge is the net input from rainfall to groundwater after all evaporative losses have been deducted and soil moisture requirements have been met. It is important that accurate estimates of recharge be obtained as it is one of the main determinants of the sustainable yield of freshwater lenses. Recharge can be estimated by a number of techniques. One of the most common and useful techniques is a water balance (or water budget) approach where water inputs to, and water outputs from, the surface of the island are quantified. This approach was used in water resources investigations of Home and West Islands (Falkland 1988, 1992a) and South Island (Falkland 1991). WATER BALANCE EQUATION Recharge can be described by a water balance equation using a specified reference zone and a specified time interval. The reference zone for a freshwater lens on a coral atoll is that zone extending from above the surface of the island down to the water table. In this zone, the flow of water is essentially vertical. The water balance equation for the upper zone on a coral island, such as those in the Cocos (Keeling) Islands, can be described as: R = P-ET,+dV where R = recharge, P = rainfall, AS ET, dV actual evaporation from all surfaces, and change in storage within the soil moisture zone (it can be a positive or negative change) As noted earlier, there is no term for surface runoff as this does not occur due to the very high infiltration capacity of the coral soils. The actual evaporation term (ET4q) includes evaporation from interception storage (for example, the leaves of trees, bushes and grass), from vegetation tapping water from the soil moisture zone and from trees with roots that penetrate to the water table and thus transpire water directly from the freshwater lens. Computations with this equation were conducted using a daily time interval, as recommended by Chapman (1985). It has been shown that computations using a monthly time step leads to an under-estimation of recharge for the Cocos (Keeling) Islands (Falkland 1988) and on other atolls (for example, Kwajalein: Hunt and Peterson 1980). Daily rainfall data and mean daily evaporation estimates were, therefore, used. DESCRIPTION OF THE RECHARGE MODEL A recharge model was developed, and a computer programme (WATBAL) written, to simulate the water balance in the upper zone and derive a monthly time series of recharge. The model is shown in Figure 22 and a brief description follows. The recharge model allows for interception storage by vegetation. A maximum value for the interception storage (ISMAX) can be defined and it is assumed that this store must be filled before water is made available to the soil moisture storage. Typical values of ISMAX are 1 mm for predominantly grassed catchments and 3 mm for catchments consisting predominantly of trees (particularly coconut trees). The airfield area on West Island is predominantly grassed while South Island and some of the northern parts of West Island consist predominantly of trees. Much of Home Island is intermediate between these two limits. Evaporation is assumed to occur from the interception storage at the potential rate. The recharge model incorporates a soil moisture zone from which the roots of shallow rooted vegetation (grasses, bushes) and the shallow roots of trees can obtain water. Water requirements of plants tapping water from this zone are assumed to be met before any excess drains to the water table. Maximum (field capacity) and minimum (wilting point) limits are set for the soil moisture in this zone. Above the field capacity, water is assumed to drain to the water table. Below the wilting point, no further evaporation is assumed to occur. The thickness of the soil moisture zone (SMZ) for the Cocos (Keeling) Islands was estimated as 500 mm based on observations of the soil profile and from studies on other atolls. Field capacity (FC) was assumed to be 0.15 based on observations of local soil type and typical values for this type of soil. Wilting point (WP) was assumed to be 0.05 based on typical values (for example, Linsley and Franzini 1973) for sand-type soils and from studies elsewhere. The operating range of soil moisture is thus assumed to be from 25 mm to 75 mm. In the model, the amount of evaporation from the SMZ is assumed to be related to the available soil moisture content. At WP, zero losses due to evaporation are assumed to 16 occur from this zone. Maximum or potential evaporation is assumed to occur when the soil moisture zone is at FC. A linear evaporative loss relationship is assumed to apply between the two soil moisture limits. Thus, at a soil moisture content midway between FC and WP, for instance, the evaporation rate is half that of the potential rate. Water entering the water table is 'gross recharge’ to the freshwater lens. A further loss, however, is experienced due to transpiration of trees whose roots penetrate to the water table. 'Net recharge’ is that water remaining after this additional loss is subtracted from ‘gross recharge’. Observations in dug pits and trenches on Home and West Islands reveal that a considerable number of roots penetrate to the capillary fringe just above the water table which typically occurs at depths of one to two metres below ground level. It is estimated that about 50% of the roots from mature coconut trees penetrate to the water table. Because the movement of the water table is relatively small, even during drought periods, these roots allow transpiration to occur even when the soil moisture store has been depleted. This is the reason that coconut trees are able to survive prolonged drought periods on coral atolls when other shallow rooted vegetation has reached wilting point and possibly died. Vegetation is assigned a 'crop factor' (Doorenbos and Pruitt 1977) according to its type. Each plant (or crop) type has its evaporative potential compared with that of a ‘reference crop’. The reference crop evaporation is equal to the potential evaporation, as derived from an appropriate method. The crop factor is a coefficient which is used to derive an adjusted potential evaporation of other crops from the potential evaporation (or the reference crop evaporation). The crop factor for most grasses and other shallow rooted vegetation is assumed to be 1.0. The crop factor for coconut trees was taken as 0.8 based on values for similar types of trees listed in Doorenbos and Pruitt (1977). Thus, the potential evaporation rate for coconut trees is taken to be 80% of that for grasses or other shallow rooted vegetation. The proportions of freshwater lens areas covered by deep rooted vegetation were estimated from coloured aerial photographs taken in April 1987 and from ground inspection. From recent investigations (Falkland 1991, 1992a), the proportions were estimated to be 0.15 for Home Island, 0 for the West Island Airfield and 0.8 for the northern part of West Island and South Island. RESULTS AND DISCUSSION Water balance analyses were conducted for freshwater lenses on West, Home and South Islands in a number of studies (Falkland 1988, 1991, 1992a). Series of monthly recharge estimates were obtained in each case, enabling drought sequences to be further analysed for estimation of sustainable yields. Graphical comparisons of annual recharge and annual rainfall (obtained by summation of monthly values) for the period 1953 to 1991 are provided in Figures 23 and 24 for, respectively, the West Island Airfield Lens and the West Island Northern Lens. A significant variation in recharge from year to year can be seen from Figures 23 and 24. In some years, recharge is actually negative (i.e. there is a net loss of water from the freshwater lens). Figure 24 shows that 'negative recharge’ occurred in the Northern Lens in 1953, 1962, 1977 and 1991 with the most negative value occurring in 1991 (corresponding to the lowest annual rainfall). In general, years of high annual rainfall 17 result in years of high annual recharge and vice versa. However, there is no simple relationship between the two parameters. This is because annual recharge is a function of the pattern of daily rainfall and not simply a function of the annual rainfall total. For the 39 year period of record (1953-1991), the following mean annual recharge estimates were obtained: - West Island Airfield Lens: 950 mm/year (49% of rainfall), - Home Island Lens: 855 mm/year (44% of rainfall), - West Island Northern Lens: 564 mm/year (29% of rainfall). The results for South Island are the same as for the Northern Lens as similar parameters were used in the recharge analysis. Figure 25 compares the annual recharge estimates from three lenses (West Island Airfield and Northern Lenses and the Home Island Lens). There are significant recharge differences between the three lenses, the main cause being differences in the density of the deep rooted vegetation, predominantly coconut trees, above the freshwater lens areas. Figure 26 shows the relationship between mean annual recharge (as a percentage of rainfall) and the percentage tree cover. This graph and the tabulated results above show that recharge can nearly be doubled by reducing the tree cover from 80% (as for the Northern Lens) to zero (as for the Airfield Lens). Due to the significant effect that coconut tree density has on groundwater recharge, one management option for increasing freshwater supplies is to selectively clear vegetation in areas where freshwater lenses occur (see also section on transpiration measurements). Cumulative annual recharge graphs for the West Island and Home Island lenses are shown in Figure 27. These graphs enable sequences of dry and wet years to be easily seen. For instance the lowest 5 year recharge period occurred from early 1976 to the end of 1981. Another low recharge period of 5 years occurred from early 1961 to the end of 1965. GROUNDWATER INVESTIGATIONS In the previous section, a number of freshwater lenses were named (e.g. West Island Airfield Lens, West Island Northern Lens and Home Island Lens). Groundwater investigations over a number of years were conducted to locate and quantify the depth and areal extent of these lenses. This section briefly describes these investigations and details of the freshwater lenses. PRELIMINARY INVESTIGATIONS The groundwater resources were first studied by Jacobson (1976a, 1976b). His investigation was limited to Home Island and involved observations of water table elevations and salinities of shallow water obtained from wells. Using this limited information he estimated the thickness of the freshwater lens at 10 to 15 metres and the sustainable yield to be 200 kilolitres per day. He recommended that more detailed investigations were warranted to confirm the preliminary results obtained. Later investigations showed that the actual thickness of the lens was not greater than 6 metres and that the estimated sustainable yield was approximately half of his estimate. 18 DETAILED INVESTIGATIONS Detailed investigations of the groundwater resources were undertaken from 1988 to 1992 (Falkland 1988, 1991, 1992a, 1992b). The aims of the groundwater investigations were to determine the location, lateral extent and depth of freshwater lenses and to determine hydrogeological properties necessary for an analysis of long-term sustainable yields from the lenses. A combined drilling and geophysical programme was used. This combined approach allowed for an accurate determination of the thickness of lenses at selected locations using the drilling programme and for reasonable estimates at intermediate sites using the electrical resistivity method. The drilling programme was relatively slow and costly but yielded accurate data whereas the resistivity programme was relatively quick and inexpensive but had a lower level of accuracy. The latter method, however, provided good estimates of lens thickness after correlation with salinity profiles obtained at borehole sites. A limited amount of seismic work was conducted at boreholes to gain a better understanding of the subsurface geological properties. Observations of topographic features, measurement of salinity levels at exposed water surfaces (wells, ponds, pumping galleries) and recording of water table movements relative to tidal movements were conducted to provide additional data. Details of all the investigations are beyond the scope of this report. Some details about the drilling programme are provided, however, as they were the most useful in terms of initial and continuing data about the freshwater lenses. A total of 29 boreholes were drilled from 1988 to 1992 on West Island (16 holes), Home Island (12 holes) and South Island (1 hole) and equipped with salinity permanent monitoring systems. Details of these holes including year of drilling, reduced level (RL) relative to mean sea level (MSL), depth to water table and depth to the unconformity between Holocene and Pleistocene sediments are shown in Table 10. The location of the boreholes are shown in Figure 28. Drilling logs with further details are contained in Murphy (1988), Falkland (1991), Murphy and Falkland (1992a) and Falkland (1992b). The permanent salinity monitoring system used in each borehole is shown diagrammatically in Figure 29. Water samples are pumped to the surface from each of the separate tubes by a portable electric pump and tested for electrical conductivity. Using the monitoring data, salinity profiles can be constructed for each borehole at intervals of typically one to three months. By obtaining a set of such salinity profiles, the salinity distribution over time can be viewed for each borehole. This data has yielded valuable information about the response of the freshwater lenses to variations in recharge. Figures 30 and 31 show the variation in the depth to the base of the freshwater zone in a number of West Island (Airfield Lens) and Home Island boreholes together with monthly recharge for the period 1988 to 1991. The antecedent recharge in 1987 is also shown. The permeability of the coral sediments was measured in-situ using falling head tests in some of the boreholes during drilling. The average permeability in the Holocene sediments was about 6 metres per day while the average permeability in the upper part of the Pleistocene sediments was about 30 metres per day. On occasions during drilling below the unconformity, karst zones such as solution channels were intersected where circulation (of water and drilling mud) was lost. In some of these zones, the permeability 19 was estimated to reach 1000 metres per day. The specific yield (or effective porosity) was estimated to be 0.3. FRESHWATER LENS DETAILS Using the results of the drilling, geophysical and other investigations, freshwater lenses were located on West, Home and South islands. On West Island, two permanent freshwater lenses have been identified underlying, respectively, the airfield and the northern part of the island. These have been named, respectively, the Airfield Lens and the Northern Lens. A permanent freshwater lens has been identified on Home Island underlying the inhabited area. In addition, one large and two smaller lenses have been identified on South Island (Falkland 1991). The locations of these lenses are shown in Figure 28. Approximate areas, maximum freshwater thicknesses, volumes and turnover times of these lenses are shown in Table 11. The areas and volumes vary with time according to antecedent recharge conditions. The areas shown in Table 3 are the maximum values and the volumes are the range of values estimated during the period of record. The turnover times are a measure of the average residence time of water within the freshwater zone and are calculated by dividing the average thickness of the freshwater zone by the mean annual recharge. A cross section through one of the lenses including details of a number of boreholes is shown in Figure 32. Some of the other islands in the Cocos (Keeling) Islands also have small freshwater lenses. Based on limited on-site tests (Jacobson 1976a, Falkland 1988), a freshwater lens is known to exist on Horsburgh Island but its sustainable yield cannot be assessed without further investigation. Preliminary investigations on North Keeling (Falkland 1988, 1992b) indicate the presence of a very thin freshwater lens at least on part of the island. It is not known whether the lenses on Horsburgh Island and North Keeling are permanent. A major influence on the thickness of the thicker lenses, particularly the Airfield Lens, is the geological unconformity between upper and lower sediments. As stated earlier, this unconformity is a very significant influence on the formation of freshwater lenses as the sediments below this unconformity have relatively high permeabilities and mixing of freshwater and seawater is readily facilitated. The depths to the unconformity are shown in Table 10. In all but one borehole in the Airfield Lens, the freshwater limit (2600 umhos/cm) occurs at all times within a zone about 2 to 3 metres below this unconformity. In general, it is evident that the unconformity is providing a limit to the formation of a deeper freshwater lens. When recharge is low, as occurred in 1991, the lens contracted to a position close to or above the unconformity. In dry periods, therefore, the lens becomes limited by recharge at this location while in wetter periods the lens is limited by the geology. It can be concluded that the underlying geology has a strong influence on the freshwater lens at the Airfield. Some of the boreholes in the Northern Lens exhibit similar behaviour while others within that lens and all of the Home Island boreholes show that the freshwater lens is contained wholly within the Holocene sediments. An interesting observation was made in the most recent investigations (Falkland 1992b).-At borehole W1 22 near the southern end of the Northern Lens a reasonably thick freshwater zone of 7 metres was found during drilling in August 1992. The thickness of 20 the lens to the mid-point of the transition zone (25,000 umhos/cm) was about 11 metres. This result was better than expected as the width of the island at this location is only about 270 metres. Based on the approximate relationship outlined earlier, the minimum width required to support a freshwater lens of this thickness is over 750 metres. This shows that the approximate relationship should be treated with some cautign as other factors not accounted for may have a significant bearing on lens thickness. At this borehole the unconformity occurs at almost precisely the same depth as the limit of the freshwater zone, indicating that it is a major influencing factor. Based on thickness and salinity of the freshwater zone at borehole WI 22, it is considered that the lens at this location will not disappear during drought periods. Future monitoring data will be used to establish the validity of this assumption. FRESHWATER LENS DYNAMICS AND MODELS Flow through freshwater lenses is complex and is influenced by hydrologic (variable recharge), geologic (variable permeabilities with depth and with distance from one side of island to the other), oceanic (tidal movements) and anthropogenic (water extraction) factors. Early conceptual models and solution techniques for freshwater lens flow assumed a sharp interface between freshwater and seawater. Observations have shown that this is not the case on atolls and wide transition zones are the norm. Sharp interface models can at best only provide an estimate of the depth to the mid-point of the transition zone, yielding no information about transition zone width. Such models also assumed horizontal flow within the lens with freshwater outflow occurring around the perimeter of the island and did not account for tidal movements. A more realistic conceptual freshwater lens flow model has evolved (Buddemeier and Holladay 1977, Wheatcraft and Buddemeier 1981, Oberdorfer et al. 1990, Peterson 1991, Underwood et al. 1992) based on detailed observations on atolls. The conceptual model accounts for vertical and horizontal tidal propagation through a dual aquifer system consisting of the upper (Holocene) and lower (Pleistocene) sediments. This conceptual model is supported by observations on a number of atolls in the Pacific (Buddemeier and Holladay 1977, Hunt and Peterson 1980, Wheatcraft and Buddemeier 1981, Anthony et al. 1989) and in the Cocos (Keeling) Islands (Falkland 1988) which have shown that tidal lags and efficiencies at water level monitoring locations within atolls are largely independent of horizontal distance from the shore. Tidal lag and efficiency (or the time difference between, and amplitude ratio of, water table movement to tidal movement) are in fact greatly influenced by the depth of the holes used for water level monitoring. Vertical propagation of tidal signals tends to be dominant in the middle of the island whereas both horizontal and vertical propagation are significant near the edges. Using the above conceptual model, the numerical solution of freshwater lens flow problems can more realistically be made with models which can account for a two layered hydrogeologic system, flow of variable density water and the mixing of fresh water and seawater. One such computer model, SUTRA, developed by the United States Geological Survey (Voss 1984) has been applied to the study of freshwater lenses and coastal aquifers on a variety of islands. Case studies of atolls and small carbonate islands include Enewetak atoll, Marshall Islands (Oberdorfer and Buddemeier 1988, Oberdorfer et al. 1990), Majuro atoll, Marshall Islands (Griggs and Peterson 1989) and Nauru, a raised atoll, (Ghassemi et al. 1990). 21 SUSTAINABLE YIELDS The sustainable (or safe) yield of an aquifer is the rate at which water can be extracted without causing adverse effects. For non-coastal mainland aquifers, the sustainable yield can be approximately equated to the long-term recharge. For freshwater lenses on small islands and some coastal mainland aquifers, such an approximation is not valid as only a small portion of the recharge is available as sustainable yield. Most of the recharge is required to counteract the effects of dispersion between the freshwater layer and underlying saline water. To avoid adverse effects from extraction (i.e. to avoid an increase in the salinity of extracted water), the overall extraction rate from the lens should not exceed the sustainable yield. An additional requirement is that pumping be distributed over the surface of the lens to avoid local upconing of saline water. Methods for estimating sustainable yield range from simple empirical approaches to complex numerical models (e.g. SUTRA). Due to time limitations, an empirical approach suggested by Mink (1976) was adopted for the Cocos (Keeling) Islands. Mink suggested that an extraction equal to 25% of the 'flux' or flow through the lens was a good first approximation to the sustainable yield. This is equivalent to 20% of the mean annual recharge based on the simple water balance equation for the freshwater lens outlined below. The water balance equation within the lens can be expressed simply as: R=Q+X+dVvV where R is the recharge into the lens after all evapotranspiration losses have been taken into account, including transpiration directly from the lens by deep-rooted vegetation, Q is the lens ‘flux’ (outflow at the edge of the lens and mixing with the transition zone at the base of the lens), Xx is the total amount of water pumped from the lens, dV is the change to the freshwater volume. In the long term, dV tends to be negligible and can be removed from the equation. Hence, the equation can be written as: R=Q+X This indicates that the maximum extraction (or sustainable yield) is 20% of mean annual recharge based on the condition that extraction should be less than 25% of flow through the lens. Given that mean recharge in the Cocos (Keeling) Islands is in the order of 25 to 50% of mean rainfall, the allowable extraction (or sustainable yield) is about 5 to 10% of mean rainfall. In relatively stable lenses, a proportion greater than 20% of the available recharge can be extracted without adverse effects on the lens. In a study of the 'Central Lens’ on 22 Bermuda, for instance, it has been suggested that about 75% of recharge could be extracted (Rowe 1984). This, however, is not considered appropriate for thin lenses, such as the Home Island Lens, at least until further monitoring results provide a more accurate insight into lens dynamics. In fact, because the Home Island Lens is a very thin and fragile lens, there is a strong case for lowering the sustainable yield estimate to slightly less than 20% of recharge. A value of 17% of recharge based on current pumping there (115 kilolitres/day) was adopted as the sustainable yield at least until more extensive salinity monitoring records are obtained and analysed. Under present vegetation conditions, the sustainable yields of the major lenses are estimated to be (Falkland 1991, 1992a, 1992b): - West Island Airfield Lens: 520 kilolitres/day, - West Island Northern Lens: 300 kilolitres/day, - Home Island Lens: 115 kilolitres/day, - South Island lenses: 220 kilolitres/day, If vegetation was substantially cleared from above some of these lens areas, the sustainable yield could be increased. In particular, it is estimated that the yields from the Northern Lens and from the lenses on South Island could be increased, respectively, to 400 and 330 kilolitres/day. It is noted that the sustainable yields for the Cocos (Keeling) Islands are based on an empirical approach. This approach, based on observations of the effects of pumping and on the results of extensive modelling on other atolls, has been shown to be at least a good approximation. It is noted that a similar 20% of mean annual recharge was used to estimate sustainable yield for the island of Laura on Majuro atoll in the Marshall Islands (Hamlin and Anthony 1987). The effects of pumping at different rates were investigated by Griggs and Peterson (1989) using the SUTRA model. They concluded that the lens was capable of extracting at least 20% and up to 30% of mean annual recharge. At extraction rates of 40% of mean annual recharge, the upconing of seawater below the gallery systems was found to be excessive. For the freshwater lenses in the Cocos (Keeling) Islands, it is intended that salinity monitoring at the network of monitoring boreholes will continue. Long term records obtained from these boreholes will enable the effects of recharge and extraction on the lenses to be evaluated. Adjustments to the present sustainable yield estimates may then be warranted. GROUNDWATER DEVELOPMENT On small coral islands, such as the Cocos (Keeling) Islands, small hand dug wells have been used for extraction of small quantities of water (e.g. at the household level). For larger centralised water supply systems, more extensive systems are required. There are three main alternative systems for larger scale pumping of water from freshwater lenses, as follows: - borehole systems, - wells, and - infiltration galleries. 23 Boreholes and wells, while possibly suitable in large freshwater lenses, are not considered suitable in the Cocos (Keeling) Islands because they extract from a localised area and can lead to excessive drawdowns. To avoid excessive drawdowns, many boreholes or wells would need to be drilled or dug. The cost of drilling or excavating, pumps and pipework would not be economical for the quantity of water extracted. Infiltration galleries (or ‘skimming wells’) are considered to be the best solution as they skim water from the surface of the lens thus minimising drawdown. These types of systems have recently been installed on Home Island and in the West Island Northern Lens. In many cases they have replaced earlier dug well systems with short radial pipes extending from their bases. Infiltration galleries consist of a horizontal, permeable conduit system laid at or close to mean sea level, enabling water to be easily drawn towards a central pump pit. Figure 33 shows the type of infiltration gallery used on Home Island (Falkland 1988). Salinity data collected before and after the galleries were installed have shown a general reduction in the salinity of the pumped water. Current water usage from each of the major lenses as a proportion of the estimated sustainable yields are as follows: - West Island Airfield Lens: 27% - West Island Northern Lens: 50% (present vegetation), 30% (cleared vegetation), - Home Island Lens: 100%, and - South Island lenses: 0% It can be seen there is ample capacity for expansion of current water usage at the Airfield Lens and no spare capacity on Home Island. The Northern Lens has sufficient spare capacity for some additional use, particularly if some clearing of the existing thick vegetation occurred. South Island remains at present an untapped resource. OTHER WATER RESOURCES Roof catchments and relatively small tanks (mainly 4.5 kilolitre capacity) provide supplementary rainwater to Home Island residents. Rainwater is also collected from a limited number of buildings on West Island. Desalination of seawater or brackish groundwater is a possibility but would be expensive. A detailed analysis of water resources (Falkland 1988) showed that groundwater is the cheapest resource to develop. Using Home Island for a pilot study, it was found that the unit costs (capital plus operating costs) of a desalination plant using the reverse osmosis principle would be about 7 times more expensive than the development of groundwater. Even if groundwater was piped from South Island to Home Island, an option which in the long term may be preferable if the Home Island lens becomes polluted, desalination would be more expensive by a factor of 3. By comparison, rainwater catchment as the sole source of water is the most expensive, being about 10 times more expensive than groundwater development on Home Island. The most appropriate option from an economic, quality and security viewpoint was the development of groundwater as the primary source of water with rainwater being used as a supplementary source. This option has been implemented. 24 REFERENCES Anthony, S.S., Peterson, F.L., MacKenzie, F.T. and Hamlin, S.N. 1989. Geohydrology of the Laura fresh-water lens, Majuro atoll: a hydrogeochemical approach. Geol. Soc. Am. Bull. 101: 1066-1075. Badon Ghijben (Ghyben), W. 1889. Nota in verband met de voorgenomen putboring nabij Amsterdam. (Notes on the probable results of the proposed well drilling near Amsterdam). Tijdschrift het Koninklijk Instituut voor Ingenieurs. The Hague, 8- 22. Bartle,G.A. 1987. Report on the evaluation of the Aokautere thermoelectric heat pulse for measuring transpiration in coconut palms. Division of Water Resources, CSIRO, Australian Government. Buddemeier, R.W. and Holladay, G.L. 1977. Atoll hydrology: island ground-water characteristics and their relationship to diagenesis. Proc. 3rd Int. Coral Reef Symp. 2: 167-174. Bureau of Meteorology 1978. Note on file 45/2311 entitled 'Cocos Island (A) M.O.' and dated, 22/5/78. Chang 1989. Hydrology in humid tropical Asia. Paper presented at the Internat. Colloquium on the Development of Hydrologic and Water Management Strategies in the Humid Tropics, James Cook Univ., Townsville, Australia, 15-22 July. Chapman, T.G. 1985. The use of water balances for water resource estimation with special reference to small islands. Bulletin No 4, prepared for the Pacific Regional Team, Australian Development Assistance Bureau. Doorenbos, J. and Pruitt, W.0. 1977. Crop water requirements. FAO Irrigation and Drainage Paper 24 (revised). Food and Agriculture Organisation, Rome, 144pp. Falkland, A.C. 1986. Christmas Island (Indian Ocean) water resources study, in relation to proposed development at Waterfall. Hydrology and Water Resources Unit, Rpt. No HWR 86/19, Department of Territories, Australian Government. Falkland, A.C. 1988. Cocos (Keeling) Islands water resources and management study. Vol. 1, Main report. Hydrology and Water Resources Unit, Rpt. No 88/12, ACT Electricity and Water. Falkland, A.C. 1991. South Island water resources study, Cocos (Keeling) Islands. Hydrology and Water Resources Branch, Rpt. No 91/02, ACT Electricity and Water. Falkland, A.C. 1992a. Review of groundwater resources on Home and West Islands, Cocos (Keeling) Islands. Vol. 1, Main report. Hydrology and Water Resources Branch, Rpt. No 92/01, ACT Electricity and Water. 25 Falkland, A.C. 1992b. Further investigations of the Northern Freshwater Lens, West Island, Cocos (Keeling) Islands. Hydrology and Water Resources Branch, Rpt. No 92/02, ACT Electricity and Water. Falkland, A.C. (editor), Custodio, E., Diaz Arenas, A. and Simler, L. 1991. Hydrology and water resources of small islands: a practical guide. Studies and Reports in Hydrology. No. 49. Unesco, Paris, France. Fleming, P.M. 1987. The role of radiation estimation in the areal water balance in tropical regions: a review. Arch. Hydrobiol. Beih. 28: 19-27. Ghassemi, F., Jakeman, A.J. and Jacobson, G.J. 1990. Mathematical modelling of sea water intrusion, Nauru island. Hydrological Processes. 4: 269-281. Griggs, J.E. and Peterson, F.L. 1989. Groundwater flow and development alternatives: a numeral simulation of Laura, Majuro atoll, Marshail Islands. Tech. Rpt. No.183, Water Resources Research Centre, University of Hawaii, Hawaii, USA. Hamlin, S.N. and Anthony, S.S. 1987. Ground-water resources of the Laura area, Majuro Atoll, Marshall Islands. U.S. Geol. Survey, Water Resources Investigations Report 87-4047, Hawaii, U.S.A., 69pp. Herzberg, A. 1901. Die Wasserversorgung einiger Nordseebdder (The water supply on parts of the North Sea coast). Jour. fur Gasbeleuchtung und Wasserversorgung, Miinchen, 44: 815-819; 45: 842-844. Hunt, C.D. and Peterson, F.L. 1980. Groundwater resources of Kwajalein Island, Marshall Island. Tech. Rpt. No. 126, Water Resources Research Center, Univ. of Hawaii, Hawaii, U.S.A., 91pp. Jacobson, G. 1976a. Preliminary investigation of groundwater resources, Cocos (Keeling) Islands, Indian Ocean, 1975. Record No. 1976/64, Bureau of Mineral Resources, Australian Government. Jacobson, G. 1976b. The freshwater lens on Home Island in the Cocos (Keeling) Islands. BMR J. Geol. Geophys. 1: 335-343. Linsley, R.K. and Franzini, J.B. 1972. Water resources engineering. 2nd Edition, McGraw-Hill Kogakusha Ltd, London. Mink, J.F. 1976. Groundwater resources of Guam: occurrences and development.Tech. Rpt. No. 1, Water Resources Research Center, University of Guam. Murphy, P. 1988. Drilling Report. Vol. 3. Cocos (Keeling) Islands Water Resources and Management Study. Rpt. No.88/G6, Scientific Services Branch, Australian Construction Services. Murphy, P. and Falkland, A.C. 1992. Review of groundwater resources on Home and West Islands. Vol. 2, Drilling Report and Appendices to Vol. 1.Hydrology and Water Resources Branch, Rpt. No 92/01, ACT Electricity and Water. Penman, H.L. 1948. Natural evaporation from open water, bare soil and grass. Proc. Royal Soc. London, series A, 193: 120-145. 26 Penman, H.L. 1956. Estimating evapotranspiration. Trans. Am. Geophys. Un. 37: 43- 46. Oberdorfer, J.A. and Buddemeier, R.W. 1988. Climate change: effects on reef island resources. Proc. 6th Int. Coral Reef Symp. 3: 523-527. Oberdorfer, J.A., Hogan, P.J. and Buddemeier, R.W. 1990. Atoll island hydrogeology: flow and fresh water occurrence in a tidally dominated system. J. Hydrol. 120: 327-340. Peterson, F.L. 1991. Recent advances in atoll hydrogeology. in Groundwater in the Pacific Rim Countries, Proc. Symp. of the Irrigation and Drainage Division, American Society of Civil Engineers, Hawaii, July 23-25, 1991 (ed. H.J. Peters), 2-8. Priestley, C.H.B. and Taylor, R.J. 1972. On the assessment of surface heat flux and evaporation using large-scale parameters. Month. Weather Rev. 100: 81-92 Quinn, W.H., Zopf, D.O., Short, K.S. and Kuo Yang, R.T. 1978. Historical trends and statistics of the Southern Oscillation, El] Nifio, and Indonesian Droughts. Fish. Bull. 76: 663-678. Rowe, M. 1984. The freshwater "Central Lens" of Bermuda. J. Hydrol. 73: 165-176. Ryan, P.L. (undated) . Cyclone over Cocos. File report, Government Conservator, Australian National Parks and Wildlife Service, Cocos (Keeling) Islands office, Opp. Stoddart, D.R. 1971. Rainfall on Indian Ocean coral islands. Atoll Res. Bull. 147: 21 pp. Underwood, M.R., Peterson, F.L. and Voss, C.I. 1992. Groundwater lens dynamics of atoll islands. Water Res. Res. 28: 2889-2902. UNESCO, 1977. Atlas of world water balance. (ed. V.1. Korzoun). USSR Committee for the IHP. Unesco, Paris, 34pp + 65 maps. Voss, C.I. 1984. SUTRA: A finite-element simulation model for saturated-unsaturated, fluid-density-dependent ground-water flow with energy transport or chemically- reactive single- species solute transport. U.S. Geol. Survey, Water- Resources Investigation Report, 84-4389, 409 pp. Wheatcraft, S.W. and Buddemeier, R.W. 1981. Atoll island hydrology. Groundwater. 19: 311-320. Woodroffe, C.D., McLean, R.F. and Wallensky, E. 1990a. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. Nat. Geog. Res. 6: 262-275 . Woodroffe, C.D., McLean, R.F., Polach, H. and Wallensky, E. 1990b. Sea level and coral atolls: late Holocene emergence in the Indian Ocean. Geology. 18: 1-96. Ou Woodroffe, C.D., Veeh, H.H., Falkland, A.C., McLean, R.F. and Wallensky. 1991. Last interglacial reef and subsidence of the Cocos (Keeling) Islands, Indian Ocean. Marine Geol. 96: 137-143. 28 Table 1. Temperature (°C): 1952-1991 data. Month Daily Mean Extreme Max. Extreme Min. January 27.0 32.1 20.1 February 27.4 32.4 20.1 March PU fe) 32.1 19.8 April 213 32:2 19.6 May 27.0 31.4 19.4 June 26.3 30.7 20.1 July 25.8 29.9 20.4 August 295.8 29.8 18.3 September 26.0 30.0 19.0 October 26.3 30.6 20.6 November 26.6 Se) 19.3 December 26.8 32.2 21.1 Table 2. Relative humidity (%): 1952-1991 data. Month Daily Mean Mean Daily Max. Mean Daily Min. January 74 82 66 February 74 81 66 March 16: 81 69 April al 84 69 May Ti 82 70 June 99) 84 69 July aA 84 va August 74 84 69 September 2 81 67 October 72 82 66 November 2 84 65 December 7d 82 65 29 Table 3. Atmospheric pressure (hectopascals): 1952-1987 data. Month Daily Mean Extreme Max. Extreme Min. January 1008.8 1015.8 970.0 February 1008.4 1018.1 992.1 March 1008.8 1015.7 987.0 April 1008.8 1015.8 998.2 May 1009.5 1017.7 1000.4 June 1010.3 1017.2 1002.5 July 1011.1 1018.5 1001.0 August 1011.6 1018.5 987.5 September 1012.2 1017.7 1004.8 October 1011.6 1018.5 1001.6 November 1010.8 1017.8 989.4 December 1010.0 1017.6 993.8 Table 4. Cloud cover (oktas): 1952-1987 data. Month Daily Mean Extreme Max. Extreme Min. January Spl 6.0 Ss) February 5.1 6.3 3.0 March 5.1 6.4 4.2 April 5.3 6.3 4.1 May 5.3 6.4 4.4 June 2 6.6 4.2 July 5.3 6.2 3.8 August 5.1 6.1 3.9 September D5 6.5 3.8 October Sal 6.4 3.8 November 5.0 6.8 3.8 December 5.1 6.4 3.5 30 Table 5. August September October November December Wind speed (metres/sec). Daily Mean (1952-1987) ANN WWINAAA RY ONNKH OOD OR WD W Extreme Max. (1952-1991) 48.8 34.0 28.3 29.3 27.8 30.9 24.7 36.0 24.2 24.2 40.1 27.3 Extreme Min. (1952-1991) SX AIO OOO IovoeeS} 3 1 Table 6. Cyclones passing over or close to the Cocos (Keeling) Islands, since 1960. Name and Date of Approx Min. central Max. wind Total number passing distance pressure when speed on rainfall closest away (km) passingclosest island (mm) to island toisland/ min. (km/hour) pressure on island(hecto- pascals) Unnamed (668) 13/2/61 60 991/992 122 71 Hazel (542) 9/3/64 80 988/991 102 121 Carol (548) 27/12/65 100 997/1002 95 48 Nancy (555) 14/3/66 100 997/1000 91 26 Doreen (566) 21/1/68 20 970/970 176 219 Dianne 6/1/70 80 996/1005 98 124 Paula (591) 27/3/73 20 999/1006 83 25 Annie (682) 25/11/73 40 995/1002 145 71 Deidre (684) PAV AV Ae) 30 995/994 85 72 Denise (455) 23/5/75 30 995/1002 100 28 Daphne (711) 15/1/82 60 995/998 81 66 Annette (696) 5/2/84 30 994/999 81 142 Daryl (699) 11/3/84 80 984/1001 92 253 Ophelia (742) 12/1/86 30 986/1002 93 252 Alison (752) 8/4/86 30 988/1002 106 Wp Frederic (784) 30/1/88 40 988/995 111 90 Herbie (783) 19/5/88 50 990/995 87 Si) John (767) 25/1/89 10 997/1000 61 106 Leon (769) 17/2/89 90 990/1006 63 27 Pedro (786) 10/11/89 100 982/1001 137 299 Graham (802) 5/12/91 100 925/1004 98 49 Harriet (803) 27/2/92 10 975/982 163 80 Ken 21/12/92 35) 990/1001 ath 162 Notes: - data are from the Bureau of Meteorology's cyclone database and other information - listed are cyclones which passed within 100 km of the islands and had a minimum central pressure less than 1000 hectopascals - rainfall totals are to the nearest 1 mm over a 2 day period. 32 Table 7. Annual rainfall (mm) at three raingauge sites on the South Keeling atoll. Year West Island Home West Island % variation (Met Station) Island (Quarantine) from West Island 1987 1871 1222 - 1988 2220 1976 - 1989 1963 1863 2210 1990 2271 2145 2410 1991 856 837 820 1992 2579 2568 2637 Total (1987-92) 11760 10611 - -10.8 Total (1988-92) 9889 9389 . -5.2 Total (1989-92) 7669 7413 8077 -3.4/+5.3 Table 8. Monthly rainfall (mm): 1952-1991 data. Month Monthly Mean Mean Monthly Max. __ Mean Monthly Min. January 199 561 i February 161 409 / March 232 630 39 April 234 551 21 May 194 646 17 June 201 649 7 July Diy? 643 25 August 201 468 18 September 2, 211. 3 October 70 S12 4 November 94 574 3 December 114 447 5 33 Table 9. Pan evaporation (mm): 1982-1991 data. Month Monthly Mean Mean Monthly Max. Mean Monthly Min. January 220 254 195 February 198 220 170 March 205 226 195 April 189 207 165 May 189 211 146 June 171 186 150 July 186 198 174 August 208 226 171 September 219 249 201 October 229 251 217 November 231 246 222 December 241 273 217 34 Table 10. Monitoring borehole details. Lens and Year of RL relative Depth of Depth of Borehole drilling to MSL (m) water table (m) _unconformity (m) Airfield Lens, West Island WI 1 1987 2.16 1.8 12.6 WI 2 1987 2.84 2.6 NONE 7/ WI16 1987 283 te7/ 11.8 WI17 1987 Saki 2.5 12.9 WI8 1988 1.33 0.9 10.3 WI19 1988 3{57 333 13.4 WI 11 1990 2.93 1.9 12.9 Northern Lens, West Island WI13 1987 1.74 i35) 12.0 WI14 1987 1.65 12 14.9 WI5 1987 1.85 1.9 14.8 WI 10 1988 1.70 12 132 WI 12 1990 1.43 1.4 10.5 WI 13 1990 1.18 13 15.2 WI 14 1990 ey 15 14.7 WI 15 1990 Lvl 1.4 14.6 WI 16 1990 n.a. 1.4 123 WI 17 1992 2.56 PED? 1297 WI 18 1992 2.09 1.6 10:5 WI 19 1992 2.19 1.8 15.4 WI 20 1992 1.81 1.6 122 WI 21 1992 AST AG 1.4 122 WI 22 1992 2.21 Zl 93 Home Island HI 1 1987 2.05 1.6 S55 HI 2 1987 1.26 1.0 10.6 HI 3 1987 1.89 1.6 11.9 HI 4 1987 1.70 1.1 9.6 HI 5 1987 1.65 1.1 >12.2 HI 6 1987 1.59 1:2 2 HI 7 1987 Bey 1.0 >15.6 HI 8 1987 127 10 >12.2 HI 9 1990 1.27 ial >15.8 HI 10 1990 1.26 1.0 >16.0 HI 11 1990 1.08 | Bes >16.1 HI 12 1990 1.14 7. >15.9 South Island SI 1 1990 n.a. 1.4 13.8 n.a. = not available 35 Note: Maxima and minima are the extreme values for each month 1 PLELLEELLLELSI SDSS ETISSESILMIIL ODPL DL ALIISSD ASSES TS, | | } | : | SS ea ee 1 CAAILLLIELEPLELLDADSALIOTIID AAA LE ALAA IIIT L LESS SDSS | | | | GAACCLEELOLEAALSSSAALA AAAS ASA AAAS SPINAL SEATS DS | | | | | GALOPLILESLLSSAASAOSAVSLIS SILAS ALAA SOI SA DEES SDSS | | | | \ PLEAILILLELSAAILISESELLLIILLISTISLLALDELS DLA L SASS To | { { | [Foes Ee | “4 WAALLAA AAA, nS La ISILLIML APPLES PSMALAPPAPPOISAAILA AAS AS TEATS | \ (age EEL LLL, VAOLPICPLLMMDAPLIALAESPIILISLIESLAAPAISS SALAS IDES AALS DLs | i | ! | u SEE RN Se aS SUAALLILELISLILLAESELAAADPLLAAA TIS SID DALAL LASS ALAS | jo) uw) oO w (2) w Oo ~ N N ro vo (9 seesBap) ainyesoduway | ERSTE SE PET a | SLLSLALELLLIOALIADALEOLAAAAD ALA LISS AAAI LSD API LILES Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month E =} = iS = MS Maximum 9 Mean daily temperature. inimum i dm , Maximum an Mean = S = &D = ae LTOPLLILLI MILLIS SEAMED OAL IA SAIS ASSDAI DS ESTTES SS TSST EST TY | | | | | CL ALIALELOLLELSELLLLALEDVLLLIEL LADO DADPLALESTDSSOLTTSS ASSES LEE VLLPLLLIIIVALISILLAISELELELEID ALIA LOPL DOPED DSAAD SLES [eee sel VLLILELIALISELSLALTTMLSA TITAS ALA ALA AALS L ASSISTS TTT ES DIALED OS I APOLAIL ISIS DOL IDLETI OLA SOALEAELAP ILI MALAA AILS LSS SL LAIATIETOLISSELILLEAOLLL ED AED LISI TAA EEL AALS | | | | | CLOOLEOPLLALLOIDLALESIISDISAS PST IALIS DADA OOS AS APOS ESA AAAS OLLLOLILLALSILADLSDLSILEILLLEELADDSS LEMMAS ETLLED SEES, | | | | eo © & oO o - N (%) APIWINH eAneley oO = PISLELILLLLIILIVAS LEAL DALEAELS DISSED ALE AADD TS TSTD LDS (o) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month (= =) a= & = SSQ§ Maximum ij Mean humidity. Ive lat InimumM re i dm , Maximum an Mean Figure 2. 36 Oo @ GLMLMLLLLLLLDLTLLLLLLLLLLLLLLILELLLTLILLLLETLLLTLLT LESTE > | Oo LLLLLLALILLDLILLLLLLLLLLLLITLELLLELLLELLLLLSLTLLLLLLLTLLLLL LL al e) ; | | | o o “ a fa} 3 fo) E (= | > ND = Q {e) ‘= oO (ok ‘= E WALALALAL AAA AAA AMAA AMMA 2 5 p= @® = 5 | | | ou e Q. SLLLILIISILELLELLELLLAAATIDLLSA AA OTTT ADA EETSTTTSSLLS TT OLSET SS 7. (dp) é c G Oo = oO l | = 5 o c te = VL IAALLLASIPPLIDAALEOIMPOPSAAAEESPIILASEA DAA STDS DIA LESS oO s of =, Oo Bs ° | >| A | . ar x CAA ALLELE LILLE IILLLLLLLLLLL LLL LLLLL LLL N iS $s A ALLILLLLLILLLLLLLILILLLLLLLLSLLLLLLSLLLLLLL LLL ILLS oO, | oO ig | | | e)) iS Src ao} S 5 ALD AILELLILLLLLTLLLELILLLOLLLLLLILLLLLTOLLLTTTTL TTT) Mice 5 i= , 4 SALAPPECEPALLEADSEASALAETAEAAEPPAD DEEP SADLPAEADEL SSE r<¢ iS | | 5 Ce fo) o | | | = oO = % | ee ry (= eS = = s oO . CLLIELLLLLEPLELLELLIPLLASEDLAOLLELEAEDEIL ELLE LLL LL MS c i“ = ae eee ULLMLLMMAELAA —~) & é c Oo = = Se |e = s a= E © E o PLLILLULLLL LLL LLLELLLLLLLL LILI LLL LLL LLLL LLL LLL es . DASA A LAMA MMM A AML MLIDLS AMM MS A A LN LE = =| f | | > & | | | = = {= © EATS PRR IETS HEY 3 % £ Ct phHLL esse LI ttt TEL ttt beh beh heh shh sets bbs st hl thts thts. = J & * (ok e =I Ee F j i 5 ne) GALA AVILELILILLSLISSIULELLLLDA LADLE LLAVLTD LTTE TL TS ae co 2 : a E =| | | | a é| 2 o PLILILLLLL ALLL ELL ILLLLLILLILELLLLLLLLLLL LLL LLL Lv x 3 co = 3 o | | | = © GALL AALEACMTLLLLTLLLLLSLELLELLLLLLLLLELLLLLELELLELL ELLE LLL Z : aera Pw ES = =| 7 Y = SOLS ILLOLILD ADA PEPELELLIDOLELEPLENAELED ELISE ELET = Y “C) Y4, 3 Y 7} 3 a Yy é € WEITZEL LLL LLL LLL LALLA, \\_ A = 3 Zz Le S | S z oo) 3 | c = SS Ss } o SUELO LLLEDALLA EDA ATLAS ATL AELLITTTTLTT LTS L ELIT LTT LELAS A VAAL hhh hhh n ; | wn “ =e) =| | | c H | 3 o | | | oO = | # oe = jo) jo) je) je) (>) (o) je) @ » © w + (ap) N - (S) BES = Bos 4 = 2 (Se}40) JAAOD pnojyD AjyjUOW Inssaig Dueydsow}y Figure 3. Figure 4. Figure 5. Percentage of Time Figure 6. Wind Speed (metres/second) ine) oO | | | 37 Note: Maxima are the extreme values for each month; minima are all zero PO SLLPLISPS PS AEDD A | CHULA SLLASLS, S : . x, % \ RS > N N S NY N : N N N S 207--_- §¥—§ —f— ~ N N N 3 Xs SS N N ey. (eee) ae) 15 S N — N N N ‘ S S ‘ N = - 10 N N N N N < x N x S N S N 54 Ns — 8 "ee S S N S BS BS S » S VSSTSTSESLILLS LDL OSSAAALIIA LEE EL. Ti 2. SSIIPITLELLA IED LALA STALEMATE ELA LEG ED AA, 7 | | | SLLEAALS TAL. i | SSODTASAAAPSODIDLSS ALLA I DALAL EDEL EEL LG VEAL LLL i SSISISIAAATIPDELISSASS LAID AMAA L EAL ASS ALAA EDF 0 VILLA LA Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month NSS Maximum fj Mean Mean, maximum and minimum wind speed. N NE E SE s Wind Direction —@— Jan (Calm 5%) —*— Feb (Calm 7%) —*— Mar (Calm 4%) —=- Apr (Calm 3%) —— May (Calm 1%) —&— Jun (Calm 2%) Wind direction. N NE E SE Ss SW WwW NW Wind Direction —®— Jul (Calm 2%) —+— Aug (Calm 1%) —*— Sep (Calm 0%) —>— Oct (Calm 0%) — Nov (Calm 1%) —#— Dec (Calm 2%) 38 0° Nv INDONESIA es) 90° Figure 7. Isohyetal map of mean annual rainfall for the Indian Ocean (modified from Stoddart 1971). 3500 Missing data in 1914, 1915, 1916, 1946 and 1952 NNRS -- SNES 5 SN a: SN Se —_ ‘ 8 NAB Nw ‘ N . BBS “aK . N see a. Ny RK N . : N . 1907 1912 1917 1922 3000 2500 2000 1500 Annual Rainfall (mm) 100048 § 500-N § 1952 1927 1942 1947 Year Figure 8. Annual rainfall, 1953-1991 (composite record with some missing data). 39 3500 LLL AE GLIA ELL. POLLILLIALLLLL AD GL. CLLIATAEIEALL OLED LAL VID ILELILLLE = LOPE PLOLILSSIA LL LT ES | C2 CELLSLELAS ES LAD LILA 7 CATOLTEOPPEDEDAETTSAL ELE LEELA ES SLELIIAL LED AEs LAALLLL POL LIL EIS bg 2 LILIA ALIELELLL LDA ILVLA EDS j=) S iS) N TAL ILLELEL EL ELLA f=) oO w) ad SSLSELILIPALLL ELA ES LEAS EPS LLAD IDEALS ALAMEDA ELIS AAAS LILA LALA DS LOCLES LES LIOTLILLLE SLISLISSA LLL ECLIS LEP AL EAL AL ELS. LAIPALEL IDLE 1000 (uw) JJeyureYy jenuuYy DALLELIL N aD io?) LLL EPPS S ALLS LS LEP STILLLELL DULIZA CALLED ] station record). 1ca 1991 (West Island meteorolog 1953- 9 fall In Annual ra Figure 9. 700 i | | | | | VLLLLILSSPLSDAESSA ALAA PDAAA EDD OA LATED SAAT ELL | | QZZZZZLLLLALLA. | GILIDISESSSALSELI SS DAELSL ASS SL VITAL ALA | GIIAIPLILLTLEPDLDSPPLSIS EDDA AASLLDAA DADA DAAASL EIDE ATT TDA TLDS LAID LSS VIAAPALILEPSLILIEPLEDADIEDPSD LADS ALLA AD LIA TEAS ELL AEAA ASSESS SAA ASSAD CZZALLA WTZZAA AA hdd VZTTAAALA AAA ULIMAMAA AMA | TAAL | | | SAASIALESIOPSSSAASS ASA ALS | | SIDLILAAETLATLLLDAIALADTT LD STTTSA IDA SASS Se SLPLLIPLESSIILIELATISVI DIALS DLA ESAS S SLD ALLE TALL LALA AAA AAA AAA he acmeamsad PILED IE LPELEOPSLESSASAAS SSSA L ETE 2 WLLITALLELALIDEADESADT DA AA ES EDEAADAIE ALLA TDL LIS ISS TITS AE? | t | | i ! | i T T [o) jo) fo) (2) [e) oO (2) oO (o) oO jo) ice) w vT isp) - (ww) jyeyurey AjyzUO fo) fo) or W Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Minimum i= 3 £ x oO = hly rainfall. ° inimum mont ° i dm , Maximum an Mean Figure 10. 40 West Is Met Stn {ij Home Is 700: =) =) 5 5 vt N (wuuu) jeyurey AjUJUOW, NDJFMAMJJASOND 1989 1990 1991 Month & Year 1988 1987 = On ON =— ' ™ 00 ron — oO 8 NM = o = fe) a E S Bo) — 8 (7) oa aS ot 9) fo) — is) =) o) 2 oO & oO 0) = — J 3 a & Ss bm = Ss r= S Figure 11. West Is Met Stn ij West Is Q Stn 450 400 COLLPLLLELELLEDAELLAALL ELL. ZALLE SLILLIPLAAA ASAD ALAA L AEA LOD SAAD ESAS LISI AAAI ALIA j=) =} oO o wo oO CILOLLLLLLILOLLLLELELLDLAOLL LATE LE LLOLLELLE ELLE PLEEPLIOLLI ESS ALLELES VIALE AAAAAREAEE PPOLLEAILA ELLA LL LIL. ' LOLLALLALLIOLLALLELLEL EEA AALAAADL ELLE LLL ELLL EL POLLELEPLEALLLALALOPLLEPELIALALLE LEAL ALLELES LLL COLLLILOALLLDLEL LLL IAAI LAA AIA ASL SOLAS ALAELLILALALLAALLOLLP LEIA OLELELELLELLII ELL S SOLIOLPAATSAAALAAS ALAA AAA ALAA ASA ALLL, AAAAAAAAA AAA AAA AMAA RADAR Le SOLEIL OLPLLLLADE WHAAAALAAAAA AAA So oO i=) o wn oOo © N N POLOLIAAAL AAMAS ALL AL ALIA ALLS TIAMAT AAAAAAD ARR VSLLLL ELLE ELELLL LLL Le) PPLE PLALLEEALLLEL Waa SLOCLLLELS POLE EL ELEPLELELLLALL aan az ae SOLLLLLS SLLELPLLLILALLILILLELLS SLLELELOELLLSSLLSAEE LLL CLLLELALLLLLSSLLLLLELEE COLLOLLL (AAAAAAAAAAEE SALILALLAILLILL ALLL OLLLLLLOLSALLLEL LS i ACCLELLELLELL SOLLLLLLLLIS ELLE PILLELSILL LLL ite) (uuu) yeyurey AjUyUOY —E: CTP LOLILLLLLELELELECLLLLLILLOLOLALAPLDLLL LL. JFMAMJJASONDJFMAMJJASONDJFMAMJJASO 1990 1991 Month and Year 1989 Monthly rainfall at the meteorological and Quarantine stations, 1989-1991. Figure 12. 4l — — West Is Q Stn —#- West Is Met Stn —+— Home Is (Wu) jeyurey AjUJUOWY SAITE|NWIND JFMAMJJASOND 1991 1990 Month and Year 1989 Cumulative monthly rainfall at the meteorological station, Home Island and the Quarantine Station, 1987-1991. Figure 13. 8000 9000 10000 7000 West Is Met Stn Monthly Rainfall (mm) 6000 5000 2000 BaD is seemless cd im in meee (=) oO a Be ee 3 vy oO N - (uw) jweyurey AjUuJLOW S| WOH y rainfall for the meteorological station and Home Double mass curve of monthl Island, 1987-1991. Figure 14. 42 Monthly Rainfall at other gauge (mm) Figure 15. Home Is Monthly Rainfall (mm) Figure 16. 0 1000 2000 3000 4000 5000 6000 West Is Met Stn Monthly Rainfall (mm) —#- Home Is —- Quarantine Stn Double mass curve of monthly rainfall for the meteorological station and Home Island and the Quarantine Station, 1989-1991. Regression Equation: H=0.865*W+2 where H = Home Is monthly rainfall (mm) and W = West Is Met Stn monthly rainfall (mm) 0 100 200 300 400 500 600 700 West Is Met Stn Monthly Rainfall (mm) Regression analysis of monthly rainfall at the meteorological station and Home Island, 1987-1991. 43 (mm) | Quarantine Stn monthly rainfall (mm) West Is Met Stn monthly rainfall 1.044*W+4 j HS é where Q (wuuu) jeyureY AjYJUOW LIS © S| ISOM ( and W 200 250 300 350 400 450 150 West Is Met Stn Monthly Rainfall (mm) 100 50 Regression analysis of monthly rainfall at the meteorological station and the Quarantine Station, 1987-1991. 300 Figure 17. VAALLILEELEDPSILDILIDL ALOE LAL AAEASLEELPDOEPDDELEPELTTADL ED A A PEEDILOSS LS LDA L ODL ALE. PILLILIAIIL ALIA ELDAEL ASSL LITLE SSS SSIS SS So LOLLPIL ISLIP ILL , PPLE PEPLISLILOPL ESE OSASLSIDOPLSSS SPOS LOIS PPLE EIA PLL LL OLLIPLESL PEPE DA , VOEUDIOADILLILLSSLSSLLD DIP LLDL As aa SIL IL DLS VLPILLAEOLIILIDAAESAAANAEEL PSA LLEL ESL ASIELLAL EDA Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month (oe) (eo) fo) (oe) [e) (eo) Ww (oe) w) [e) w N N _- - (ww) uolesodeng ued AjyjuoW Minimum Maximum {jj Mean Mean, maximum and minimum pan evaporation. Figure 18. 44 250 (ww) LLL hhihde ee | | | ada sEsssitihidititthitEsidttshhtidittihthis “assitthiititiissstittttttthti;¢ddddlithtt¢sséisss AILS ISASSDASSALSS IDS ADDS DDS Uusititittttiiittit SLITS LLLADISSS DS ASE bd hedhitithtit¢seeisssithttttishhsssda COL ALLISIS PIAL APLAL LAL SIDPLLSPAPSA SDE. S, MOANA MISA AM ptt tht) tht ltt ltl th YULIIDIILILILILLILILLILLIDDIDI DLL LLLLLLD SDSS Sb WLLL dldddldddddddddéiidddés WIMIMMMMIUMMM LILI LL LL LSA SEL, | oO Oo = 200 5 c onesodeng jeiuajog Ajujuopy ueayy Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec jo) oO Ww Month GB Penman method Pan method Mean monthly potential evaporation estimates using pan and Penman methods. Figure 19. (uw) JJesuleYy jenuulyy UBAaY JO % 1970 1975 1980 1985 1990 1995 1965 1960 1955 15 ° w ° rs) = -10 15 (IOS) xepuy UOHE)|!OSC WaUNOS Year ~s<- West Is Rainfall SOl Relation between Southern Oscillation Index and annual rainfall. Figure 20. 45 Evapotranspiration Rainfall MAW // mitre \ iar See ee eres 2-3m Mean sea level Lore ee — an e LAS q ? Z Selected gravel \ 6mm maximum. ork Base of bore. Figure 29. _ Borehole salinity monitoring system. 51 West Is Recharge —+— Bore WI6 >< Bore WI7 —- Bore WI8 Monthly Recharge (mm) Freshwater Depth (m) Month & Year Figure 30. Depth of freshwater zone and recharge for boreholes WI 16, WI 17 and WI 18 on West Island. West Is Recharge —+— Bore HI1 >< Bore HI3 Freshwater Depth (m Monthly Recharge (mm) ZB NNN NE NS N N N N ‘meh SN No NNN ; 0 AMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND 1987 1988 1989 1990 1991 Month & Year Figure 31. Depth of freshwater zone and recharge for boreholes HI 1 and HI 3 on West Island. (ALLA LALLIL ALA LLL LL ALLL CL OLTT OLED. \-ZZZZzAaaee 52 , (Electrical Conductivity for PS5 is offset 10m PS2 Seawater = 46000 umhas/cm) Lagoon ' approx. ground surface ——— Sees. Ny Gowan ar eS Se a7 Mean Sea Level x | [eugene Kd —s Z =» =i 5 _— ae - = i" ages 2 B50: \ Electrical Conductivity = 2600 umhos/cm (Base of Freshwater Zone) 7200 / / / aa ie Electrical Conductwity eat = 23000 umhos/cem HIS. (Midpoint of Transition Zone) a NOTE: Boreholes show depth in metres on the left and 18 + 37200 on the right Electrical Conductivity of sample Horizontal Scale: 41. :/5000 obtained during drilling in umhos/em Vertical Scale: 1 : 200 21 HI6 Figure 32. Cross-section through Home Island lens showing shape of lens at time of drilling in Oct/Nov 1987. Cover with good seal for pump well ‘Gatic’ Cover *Gatic’ Cover at or 50 dia. PVC pipe at surface Level just below surtace (for Salinity tests) 1500 dia. (approx.) To supply Back fill conc. pump well Mean Sea Level (MSL) ph ee Water Table ens: 150 | Concrete base 1 ir | 100, 150 of 225 3 . slotted PVC pipe impermeable membrane }—__—_——_—_—_—- Maximum 100 metres ——--—+ (Class 6 water (thick polytfiene sheet) supply pipe) placed across trench, about 100-200 above top of pipe — RL-0 30 To manhole (relative to MSL) NOTES: 1. All joints between pipes and pump wells/manholes to be well sealed 2. Float switch level to be as close above base of pump suction as possibje (without causing vortex) 3. Optional housing over pump well and pump/meter (not shown) 4. All figures in mm but drawing not to scale Figure 33. Cross-section through typical Home Island infiltration gallery. ATOLL RESEARCH BULLETIN NO. 401 CHAPTER 3 LATE QUATERNARY MORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY D.E. SEARLE ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 3 LATE QUATERNARY MORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY D.E. SEARLE * ABSTRACT Seismic profiles have been obtained from Cocos lagoon and correlated with radiometric-dated core data. They show that the Last Interglacial atoll had a very similar morphology to the present atoll, and lies 8 to 28 m below sea level. It is overlain by 8 to 18 m of Holocene reefal deposits. Two deeper surfaces, of similar shape to the overlying reef surfaces, are recorded beneath the lagoon. These are tentatively identified as solution unconformities formed during glacial lowstands at oxygen isotope stages 6 and 8. Blue holes in Cocos lagoon are interpreted as solution dolines modified by clastic-controlled growth during submergence and relief enhancement by facies-controlled solution weathering during emergence. The data support the antecedent model of reef development on a subsiding base. However, unlike Purdy's (1974) antecedent karst model, both constructional relief and differential erosion are emphasised. Preservation of thin reef caps indicates that the atoll surface may have been lowered by as much as 18 m during each glacial cycle by subaerial erosion and solution. This allows an improved estimate of 0.02 mm/yr for subsidence of the atoll during the late Quaternary. INTRODUCTION Charles Darwin visited the Cocos (Keeling) Islands in 1836 and collected field evidence in support of his subsidence theory of atoll formation. This theory recognises an evolutionary sequence by vertical reef growth from volcanic island fringing reefs, through barrier reefs, to coral atolls driven by gradual subsidence of the volcanic island core. Drilling through Pacific atolls has encountered thick, shallow-water, reef-associated limestone overlying basal basalt; these results have substantiated Darwin's theory. Isostasy and plate tectonics provide explanation for subsidence and co-existence of evolutionary reef phases along linear island chains. Darwin's field evidence from Cocos consisted of observation of erosion of the shoreline and collapse of coconut palms by undercutting at the shoreline. He inferred subsidence from this geomorphic evidence - an interpretation that has been strongly questioned, for example by Ross (1855). A cemented coral conglomerate platform underlies many of the Cocos Islands and is exposed along the ocean shoreline. Dating of material from this platform, which has been identified as a former reef flat, indicates that sea level was about 1 m higher than present some 3000 years ago (Woodroffe et al. 1990). Drilling on the Cocos Islands has intersected the "Thurber" discontinuity, a solution unconformity separating Holocene from Pleistocene reef that, at one site, has been dated at about 120, 000 years B.P. This corresponds to the Last interglacial oxygen * Department of Minerals and Energy, P.O. Box 194, Brisbane, Queensland, 4001. 2 isotope substage 5e (Woodroffe et al. 1991). The discontinuity lies at 6-16 m below sea level (revised depths, see Fig. 1 and Woodroffe et al. this volume). Woodroffe et al. (1991) interpreted the data to imply a subsidence rate of 0.1 mm/yr over the last 120, 000 years. Thus, the Cocos (Keeling) Islands have been undergoing gradual, long term subsidence. Superimposed on this are millennia scale, glacially-induced eustatic movements in relative sea level and consequent hydro-isostatic adjustments. The most recent relative sea-level change has been a slight fall over the last few thousand years. This paper is concerned with the structure of the Cocos (Keeling) Atoll during the Late Quaternary, and the influence of antecedent topography on the morphology of the modern atoll. REGIONAL SETTING The Cocos (Keeling) Islands, in the eastern Indian Ocean, consist of an atollon - North Keeling Island, and a horseshoe-shaped atoll - South Keeling Islands. The South Keeling atoll lies 40km south of North Keeling and consists of a shallow circular lagoon fringed by a series of reef islands. The islands vary in size from West Island, which is 11km long, to small islands less than a kilometre long (Fig. 1). The lagoon of South Keeling atoll is 10 km across east to west, and 12 km from north to south. To the south and east, the lagoon connects with the open ocean through inter-island channels that are about 1 m deep. To the north and north west, deep openings occur on either side of Horsburgh Island (see Kench, this volume). Cocos lagoon is shallow to the south, with much of it being exposed at low tide. To the north the lagoon deepens to over 15 m. Much of the central south-eastern part of the lagoon is occupied by steep-sided "blue-holes", some over 15 m deep (Fig. 5). The holes average 100 m across but are generally smaller and more isolated to the south east while towards the centre of the lagoon the holes commonly coalesce. Submerged patch reefs occur in the central lagoon (see Williams, this volume). METHODOLOGY A high resolution continuous seismic reflection survey was carried out in the lagoon of the South Keeling Islands. The continuous seismic profiling (CSP) was conducted using a "Uniboom" sound source, triggered every half-second at an energy level of 200 Joules. A single-channel, 8-element hydrophone was used to receive the seismic signal after reflection from the seabed and subsurface. The signal was filtered to remove noise below 500 Hz and displayed as a seismic cross section on a graphic recorder. Seismic profiling was carried out at a speed of 4 knots. Position fixes were obtained every 2 minutes (about every 250 m) using a Trimble GPS unit. In order to limit the presence of seabed multiples the sound source and hydrophone were deployed from opposite quarters of the vessel. This field technique imposes a geometrical depth scale on the records, and corrections have been applied to depths measured from the seismic records. RESULTS A total of 70 km of seismic profile was recorded over the northern and central part of the lagoon, Figure 1. This includes profiles across the blue holes, and profiles approaching Home and West Islands to allow correlation of seismic with drill data. Seismic profiling was not feasible in the very shallow southern part of the lagoon. Seismic data quality was only fair owing to noise from wind waves in the lagoon generated from the persistent Trade Winds. In the shallow seagrass-covered parts of the lagoon, particularly off Home Island, no seismic signal was received by the hydrophone array. This was due to attenuation of the signal by gas bubbles adhering to the seagrass. The seismic profiles from Cocos lagoon record subsurface data down to 45-50 milliseconds of reflection time, being limited by the system used and the nature of the sediments encountered. This converts to about 35 m in depth, assuming a conversion velocity of 1500 m/s. This value is close to the seismic velocity of sea water and reef limestone of Holocene age. Older sediments, which may be more compacted and cemented, can have higher velocities. Thickness of older units may be overestimated since a conversion velocity of 1500 m/s has to be used in the absence of measured values. INTERPRETATION The seismic profiles record the seabed and subsurface reflectors. The subsurface reflectors are derived from changes in physical properties at geological discontinuities. Experience in reef provinces elsewhere suggests that these discontinuities are commonly solution unconformities formed when reefs were exposed to subaerial processes during glacially-induced, sea-level lowstands (Orme et al. 1978, Searle and Harvey, 1982). Thus reflections from unconformities imply atoll emergence, while the seismic sequences bounded by these reflections are due to reef and lagoonal deposition during interglacials. . Lagoonal bathymetry is one of high relief except for the marginal areas where shallow subtidal flats commonly extend for 2 to 3 km into the lagoon. The deepest water encountered during the seismic survey was in the blue holes. One hole was 18 m deep with a rim barely deep enough to allow passage (0.5 m at high tide). Water to 16 m deep was traversed in the centre of the lagoon and close to the southern tip of Direction Island. Three subsurface reflectors (referred to as A, B, and C in descending order) are present on the seismic records (Figs. 2 and 3). These reflectors, together with the seabed reflection, form the sequence boundaries of 4 depositional units. The seismic sequence bounded by reflector A and the seabed varies in thickness from 8 to 18 m. It is thickest towards the southern part of the lagoon where blue holes appear to be infilled (Fig. 3). Unlike seismic data from the Great Barrier Reef Province (see Orme et al. 1978; Searle, 1983; Searle and Flood 1988) little internal structure is apparent in this sequence, and it is not possible to differentiate between reef rock and bioclastic facies. Reflector A varies in depth from 11.5 m where it rises towards West Island, to 28 m in the centre of the lagoon. The shape of the subsurface defined by reflector A closely matches that of the lagoonal seabed. Both have high relief, except where they rise gently towards the islands; are deepest in the centre of the lagoon; and steepest dips are apparent over the blue holes and their rims. The data have been tested for velocity distortion using 4 published refraction velocities for reefal limestone of Holocene age (Harvey and Hopley 1981). Less than 20% of the relief on reflector A can be accounted for by this effect. Similarities in the morphology of the lagoon and its subsurface as seen on the seismic records are, therefore, real and not attributable to distortion. A thin sequence separates reflectors A and B. This sequence averages only 1 - 2m in thickness except off the northern part of West Island where it increases to 5 m (Fig. 2). Reflector B lies at depths of 16 to 32 m, and generally conforms with the shape of reflector A. It is deepest beneath the centre of the lagoon and shallowest off the northern part of West Island. Here it forms a terrace at -17 m overlying a -20 m terrace defined by reflector C (Figure 2). Beneath the blue holes reflector B lies at a depth of 24 to 26 m, being deeper below the holes and somewhat shallower below their rims. Sequence B/C, where recognised, is only 1-2 m thick (Figs. 2 and 3). Reflector C, although commonly lost below the limit of penetration, lies at depths of 18 to 32 m. Reflector C is best developed between the northern part of West Island and the central lagoon where it forms the -20 m terrace that then dips towards the centre of the lagoon (Figure 2). Reflector C is also present at about -22 m off Direction Island, dipping towards the centre of the lagoon. DISCUSSION Drilling on the Cocos (Keeling) Islands (Woodroffe et al. 1991; Woodroffe et al. this volume; Falkland, this volume) intersected the "Thurber" discontinuity at depths of 6 to 16 m below mean sea level (Fig. 1). Projection of reflector A from the seismic profiles that come closest to the drillholes on West Island allow correlation with the top of the older limestone at the "Thurber" discontinuity (Figs. 2 and 4). On Cocos this older limestone has been dated as Last Interglacial in age (123 + 7 ka B.P., Woodroffe et al. 1991). Thus reflector A is interpreted as the weathered surface of the Last Interglacial atoll modified by subaerial exposure prior to the Postglacial transgression, and upon which Holocene reef has developed. Reflector A marks the Holocene/Pleistocene boundary. Reflectors B and C (Figs. 2 and 3) are interpreted as older solution unconformities formed by subaerial exposure on progressively older atoll surfaces. The sea-level curve of Chappell (1983) shows glacial lowstands reaching minina at 150, 000 and 260, 000 years B.P.; reflectors B and C, respectively, may represent the subaerial surfaces formed during these episodes (oxygen isotope stages 6 and 8) and upon which subsequent reef development took place. The sequence below reflector C probably consists of reefal deposits dating from the highstand of oxygen isotope stage 9. Seismic sequences B/C and A/B represent reefal limestone deposited during highstand stages 7 and Se, respectively. Although reflectors B and C are imperfectly recorded by the seismic system, the shape of the ancestral atoll lagoon (after subaerial weathering) appears to have been preserved during the Late Quaternary through to the present day. Even at the relatively fine scale of blue holes, shape has been preserved, and possibly enhanced (Fig. 3). The most noticeable difference in atoll shape since the formation of surface B has been the progradation of Last Interglacial age reefal sediments over the -17 m terrace off West Island (Fig. 2). The unconformity defined by reflector A is interpreted as the weathered surface of the Last Interglacial reef. This surface lies 6-16 m below mean sea level beneath the islands (Woodroffe et al. 1991, and this volume), and dips beneath the lagoon to a depth of 28m below mean sea level. Off West Island, where the best subsurface data is available, this Holocene/Pleistocene interface dips at 4 m/km into the lagoon (Fig. 4). Below the blue holes and the submerged patch reefs in the centre of the lagoon surface A is substantially mimiced by the modern reefal surface (Fig. 3). Thus the morphology of the modern atoll is inherited from its antecedent platform, which, in its turn appears to have been inherited from older atoll landforms. In this respect, the data from Cocos support the antecedent karst theory of Purdy (1974). It appears that atoll landforms maintain and even amplify their antecedent relief. This may be due to either differential accretion during submergence, or to differential erosion during subaerial exposure, or a combination of both processes. Accretion by vertical reef growth during marine transgressions would be more rapid in the shallow photic zone, aided by a process of clastic control whereby coral growth can be retarded by biodetritus deposited in low areas (Goreau and Land 1974). Subaerial erosion may also amplify relief by facies control of solution weathering (Bloom 1974). Under this process the less permeable lagoonal sand and mud facies would dissolve away faster than the more permeable reef framework and coarse rubble/conglomerate facies. On a smaller scale, facies control accounts for the development of blue holes. Once elevation differences were present in the proto-atoll lagoon the positive features would be exploited by coral colonisation and vertical growth. Then, as now, the lagoon would tend to silt up by progradation of bioclastic sand and mud behind the windward margin. The negative areas between reef patches would gradually infill as an intertidal reef flat developed in the lagoon. Upon emergence the surface would experience facies-controlled differential weathering through internal drainage and solution. This would result in the formation of residual prominences at the site of former patch reefs, and the development of solution dolines in the intervening lagoonal mud and sand facies. Subsequent glacial cycles would emphasise relief by constructional (highstand phases) and solution weathering (lowstand phases) processes forming relatively deep blue holes. While blue holes are relatively numerous in the Caribbean, few have been reported from the Indo-Pacific region (Purdy 1974). Deep blue holes are not common in the Great Barrier Reef; they occur singly, and partly for that reason are considered as collapsed doline features (Backshall et al. 1979). The blue holes on Cocos, by contrast, dominate the southern central area of the lagoon. They are unlikely to be juxtaposed collapse features, and are considered to be multi-generational solution dolines. The Holocene pattern of atoll development may be taken as a model for earlier cycles. The difference in thickness between the older Pleistocene reef and associated lagoonal sediments, and Holocene deposits is presumably due to subaerial erosion of the older sediments. It is interesting to speculate on the relationship between reef growth and subaerial erosion, and consider implications of sea level and subsidence for the preservation of the resulting reef cap. The data from Cocos implies a mean accretion rate of 2 mm/yr, based on an average thickness of Holocene of 16 m deposited since the sea flooded the atoll some 8000 years ago. Reefal accretion could only occur during submergence for, say 10% of each Late Quaternary glacial cycle. This would add 20 m to the reef cap during each interglacial cycle. As sequences A/B and B/C are only 1-2 m thick, 18 m of reef cap must have been 6 eroded during the remaining 90% of time when the atoll was emergent, hence an erosion rate of 0.2 mm/yr is implied. Data on erosion rates is sparse and values vary greatly. For instance, Trudgill (1976) provides an average value of 0.26 mm/yr from Aldabra Atoll from a range of values that vary with lithology, mineralogy and soil cover. Purdy (1974), quoting Land et al (1967), provides values of 0.01-0.04 mm/yr that vary with rainfall. Data for reefal accretion appears less extreme and better established. On Cocos, Holocene accretion estimated from the seismic and core data falls within the range of estimates for other atolls, see, for example, Marshall and Jacobsen (1985). For the reef cap to be repeatedly accommodated, highstand sea levels must rise, or the atoll must subside. Accepting the latter as the most feasible in the longer term, a subsidence rate of 0.02 mm/yr is implied from the data. This allows for a 2 m increment to the reef cap during each glacial cycle. Subsidence rates for Cocos (Woodroffe et al. 1991) can now be refined to the lower value of 0.02 mm/yr, which is closer to global average oceanic subsidence rates (Ladd et al. 1970, Menard 1986). CONCLUSIONS Continuous high resolution seismic data has been obtained from Cocos lagoon. Correlation with radiometric-dated core data from the Cocos islands allows reconstruction of the shape of the Last interglacial atoll. The Last Interglacial atoll surface, both island and lagoon, has a very similar morphology to the present atoll, and lies 8 to 28 m below sea level. It is overlain by 8 to 18 m of Holocene reefal deposits. Two deeper surfaces, of similar shape to the overlying reef surfaces, are recorded beneath the lagoon. These are tentatively identified as solution unconformities dating from subaerial exposure of earlier atolls during glacial lowstands at oxygen isotope stages 6 and 8. A field of partially-infilled blue holes seen in the southern central part of the lagoon is underlaid by antecedent depressions. These are interpreted as solution dolines modified by clastic-controlled growth during submergence and relief enhanced by facies-controlled solution weathering during emergence. The seismic data, correlated to core data, supports the antecedent model of reef development on a subsiding base. However, unlike Purdy's (1974) antecedent karst model, both constructional relief and differential erosion are envisaged as mechanisms for preserving atoll morphology through late Quaternary time. The preservation of only thin former reefs indicates that the atoll surface was significantly lowered by subaerial erosion and solution during each glacial cycle. The data allow an improved estimate of 0.02 mm/yr for subsidence of Cocos Atoll during the late Quaternary. ACKNOWLEDGMENTS The seismic work in the Cocos (Keeling) lagoon was undertaken at the kind invitation of Colin Woodroffe, Roger McLean and Eugene Wallensky. The survey was 7 conducted aboard John Clunies-Ross' vessel "Madaeke". Major John Mobbs (Australian Defence Force Academy, Canberra) provided full survey control for this work. Colin Woodroffe has provided useful comments on this paper. The research was funded by the Australian Research Council. This paper is published with the permission of the Chief Government Geologist, Department of Minerals and Energy, Brisbane. REFERENCES Backshall, D.G., Barnett, J., Davies, P.J., Duncan, D.C., Harvey, N., Hopley, D., Isdale, P., Jennings, J.N. and Moss, R., 1979. Drowned dolines-the blue holes of the Pompey Reefs, Great Barrier Reef. BMR J. Aust. Geol. Geophy. 4: 99- 109. Bloom, A.L., 1974. Geomorphology of reef complexes. In Laporth, L.F. (editor) Reefs in Time and Space. Soc. Econ. Palaeont. Mineral. Spec. Pub. 18: 1 - 8. Chappell, J,. 1983. A revised sea level record for the last 300 000 years from Papua New Guinea. Search 14: 99 - 101. Goreau, T.F., and Land, L.S. 1974. Fore-reef morphology and depositional processes, north Jamaica. In Laporth, L.F. (editor) Reefs in Time and Space. Soc. Econ. Palaeont. Mineral. Spec. Pub. 18: 77 - 89. Harvey, N. and Hopley, D. 1981. The relationship between modern reef morphology and a pre-Holocene substrate in the Great Barrier Reef Province. Proc. 4th Int. Coral Symp. 1: 549-554. Land, L.S., Mackenzie, F.T, and Gould, S.J. 1967. Pliestocene history of Bermuda. Geol. Soc. Amer. Bull. 78: 993-1006. Marshall, J.F. and Jacobsen, G. 1985. Holocene growth of a mid-Pacific atoll: Tarawa, Kiribati. Coral Reefs. 4: 11-17. Menard, H.W. 1986. Islands. Scientific American Library, 230pp. Orme, G.R., Flood, P.G. and Sargent, G.E.G. 1978. Sedimentation trends in the lee of outer (ribbon) reefs, Northern Region of the Great Barrier Reef Province. Phil. Trans. R. Soc. Lond. A291: 85 - 99. Purdy, E.G. 1974. Reef configurations: cause and effect. In Laporth, L.F. (editor) Reefs in Time and Space. Soc. Econ. Palaeont. Mineral. Spec. Pub. 18:9 - 76. Ross, J.C. 1855. Review of the theory of coral formations set forth by Ch. Darwin in his book entitled: Researches in Geology and Natural History. Natuurkd. Tijdschr. Ned. Indie. 8: 1-43. Searle, D.E. 1983. Late Quaternary controls on the development of the Great Barrier Reef: Geophysical Evidence. BMR J. Aust. Geol. Geophy. 8: 3, 267-276. Searle, D.E., and Harvey, N. 1982. Interpretation of inter-reefal seismic data: A case study from Michaelmas Reef, Australia. Marine Geol. 46: MR-MR16. 8 Searle, D.E., and Flood, P.G. 1988. Halimeda bioherms of the Swain Reefs - southern GBR. Proc. 6th Int. Coral Reef Symp. 3: 139-144. Trudgill, S.T. 1976. The subaerial and subsoil erosion of limestones on Aldabra Atoll, Indian Ocean. Z. Geomorph., Suppl. Bd. 26: 201 - 210. Woodroffe, C.D., McLean, R.F. Polloch, H.A. and Wallensky, E. 1990. Sea level and coral atolls: late Holocene emergence in the Indian Ocean. Geology. 18: 62-66. Woodroffe, C.D., Veeh, H.H., Falkland, A.C., McLean, R.F. and Wallensky, E. 1991. Last interglacial reef and subsidence of th Cocos (Keeling) Islands, Indian Ocean.Marine Geol. 96: 137-143. Se Se ae cg Island Q \ \ KN \ | | | \ 12°05’S » Direction Island 96° S0'E Figure 1. Cocos (Keeling) Atoll, showing the location of seismic tracks, coreholes and depths (below msl) to discontinuity. 10 APPROXIMATE DEPTH IN METRES ‘(V) SONS Joa [RLIL[SIIUT 3SC] Ot BUIATIOAO ‘Yor WY] Ajayeurtxoidde ‘aouanbas SUSSOIOH PON *S10I99]Ja1 soeyINsgns - D‘g‘y ‘paqeas - ¢ ‘aovjns Jo1eM - MA *Pl0deaI paxaidiajuy (q “pueysy isaA4 Jo wed WISYWOU JJO ploda1 GSD (ke q2@ AYuNDIA ‘7 aunsij ee AYWNDIA aesess REFLECTION TIME IN MILLISECONDS re 4 yt of We Ca TL | { 11 | “OWN YSNOI) Joljar pooueyUs pur ‘sadvjins JUsposaque PUL SdELMS [BUOCOSE] Wopou! UsaMm3eq odeys ul AjUIETIUMS feloUed SION “psosel payaiduajuy (q (SHY) $Jeer yored pure ‘sojoy onjq deep ‘(SH’T) Q[OY an]q paul SuIsJoAey ‘UOOSE] [eMUSD WOYINOS WIOIJ PIOIII JSD (ke "€ onBly qe auNndia ee 3YHNDISA NigMhes ye wm Pan (SABA Aes bang : AES APPROXIMATE DEPTH IN METRES REFLECTION TIME IN MILLISECONDS 12 1 metre contour interval ® Coreholes Seismic profile Datum — MSL } _ oN 10) i | Island \ eso a 12°10'S Figure 4. Contour plot of Last Interglacial atoll surface beneath western part of Cocos lagoon, based on seismic and drillhole data. 8) Figure 5. Oblique aerial photograph, looking east, of southen part of Cocos lagoon showing field of blue holes and atoll rim (between Home and South Islands). The holes are infilled to the south, coalesce in the midfield and pass into submerged patch reefs to the north. The seismic profile shown in Figure 3 passed close to the foreground in this photograph. Foreground width = 1500m. ATOLL RESEARCH BULLETIN NO. 402 CHAPTER 4 GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE, R.F. McLEAN AND E. WALLENSKY ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 4 GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE *, R.F. McLEAN ** AND E. WALLENSKY *** INTRODUCTION Charles Darwin's subsidence theory of coral reef development has gained wide acceptance. The initial idea had occurred to Darwin while he was in South America, and he refined it during his voyage across the Pacific, writing an early draft of a manuscript on Coral Islands, probably between Tahiti and New Zealand (see Stoddart 1962), much of which subsequently appeared in his book on the structure and distribution of coral reefs (Darwin 1842). The Cocos (Keeling) Islands, which Darwin visited in April 1836 during the voyage of H.M.S. Beagle, were the only coral atoll on which he ever landed. He sought evidence there in support of the theory of coral reef development, and he left convinced that he had found such support. He wrote enthusiastically to his sister Caroline on 29 April 1836, some days after leaving Cocos, saying “I am very glad we called there, as it has been our only opportunity of seeing one of these wonderful productions of the Coral polypi.- The subject of Coral formation has for the last half year been of particular interest to me. I hope to be able to put some of the facts in a more simple and connected point of view, than that in which they have hitherto been considered”. Lyell (1832) had earlier proposed that atolls, with their characteristic annular reef rims which encircle a central lagoon, consist of a thin veneer of coral growing over the rims of submerged volcanic craters. Darwin considered it improbable that so many volcanic rims would lie within the narrow depth range required for reef growth, ard proposed that there “is but one alternative; namely, the prolonged subsidence of the foundations on which atolls were primarily based, together with the upward growth of the reef-constructing corals. On this view every difficulty vanishes; fringing reefs are thus converted into barrier-reefs; and barrier-reefs, when encircling islands, are thus converted into atolls, the instant the last pinnacle of land sinks beneath the surface of the ocean.” (Darwin 1842 p109). Darwin used his observations during his brief visit to Cocos in support of his theory of coral reef development, and wrote a manuscript (termed the Cocos Coral Manuscript by Armstrong 1991) shortly after leaving the islands. Much of the debate for the next 100 years also centred around the Cocos (Keeling) Islands. Thus, although John Murray did not himself visit Cocos, during the voyage of H.M.S. Challenger, he funded the visit of Henry Brougham Guppy in 1888 (though his prime interest seems to have been to get Guppy to examine phosphate deposits on Christmas Island). Guppy td Department of Geography, University of Wollongong, Northfields Avenue, Wollongong, New South Wales, 2522. i Department of Geography and Oceanography, Australian Defence Force Academy, Canberra, Australian Capital Territory, 2601 *k* Department of Biogeography and Geomorphology, Australian National University, P.O. Box 4, Canberra, Australian Capital Territory, 2601 was already critical of Darwin's subsidence theory of reef development, having observed fossil coral reefs elevated above modern sea level in the Solomon Islands, and was sympathetic to the alternative theory put forward by Murray (1889). Murray proposed that atolls were the result of solutional processes concentrated in the lagoon. Guppy (1889) clearly demonstrated that the Cocos lagoon was infilling with sediment, and he described the reef islands in detail. He was on Cocos for 10 weeks and he propounded the view that the reef rim was building out episodically , as Murray had suggested. Wood-Jones was the doctor on the Cable Station on Direction Island, 1905-1906. He spent a considerable period examining the atoll, and based on his observations wrote a volume entitled Coral and Atolls, in which he put forward an alternative view, that the present morphology of the atoll had developed in response to the pattern of sediment production and deposition. Thus the three views, summarised by Wood-Jones as the Subsidence, Solution, and Sedimentation hypotheses each had a particular connection with Cocos, and had each been tried and accepted by one of its major proponents in these islands. The only other hypothesis deserving serious consideration according to Davis (1928), in a review of the coral reef problem, was the glacial control hypothesis proposed by Daly (1915, 1934). In this account we present results from a geomorphological reappraisal of the atoll based on a number of visits, and a series of surface and shallow subsurface observations. Our concern is primarily with the development of the surface morphology of the atoll rather than the atoll's structure, a distinction which needs to be made (Stoddart 1973), but which is not always clear in the preceding references. All of the surface features have formed in the mid-late Holocene, over a pre-Holocene surface; there are no surface outcrops of late Pleistocene or older materials anywhere on the atoll. Nevertheless, our interpretations of the surface morphology and Holocene evolution have implications for the structure and longer-term development of the atoll. REGIONAL SETTING The Cocos (Keeling) Islands comprise the main atoll of the South Keeling Islands (lat. 12°12'S, long. 96°54'E) and an isolated atollon, North Keeling (lat 11°50'S, long. 96°49'E), 27 km to the north. These are connected by a submerged ridge at a maximum depth of 1000 m. They comprise a single feature rising from an ocean floor depth of about 5000 m (Fig. 1). The age of the ocean floor at this location is not clear, but lies in the range 60-90 million years (Jongsma 1976). In this region the seafloor gets younger to the north, and appears to have been formed from a spreading centre that has been subducted into the Java Trench. The Cocos (Keeling) Islands lie on the Cocos Rise. To the south, the Umitaka Mary seamount reaches to within 16m of the sea surface. This chain of seamounts, the Vening Meinesz seamounts, can be traced northeastwards towards Christmas Island. They are not, however as regular as linear chains of islands and seamounts seen in the Pacific, and it is uncertain whether they have developed from a single hotspot in the same manner as may Pacific seamounts (Scott and Rotondo 1983). That Cocos represents a carbonate reefal capping on a volcanic seamount seems extremely likely although the depth to oceanic basalt is unknown. Magnetic surveys show an anomaly, reading 250nT in vertical intensity (Chamberlain 1960, Finlayson 1970). There is also a pronounced gravity anomaly over the island. In addition a basalt and tuff pebble has been dredged from the western end of the Cocos Rise (Bezrukov 1973), further supporting the idea that carbonate overlies a volcanic basement. The southern atoll consists of a reef rim that surrounds the atoll with two major passages, one to the northwest and one to the northeast. A series of reef islands (described in detail by Woodroffe and McLean, this volume) occur on the horseshoe- shaped reef rim which is continuous from Direction Island to the northern end of West Island. Horsburgh Island is isolated at the north of the lagoon (see Fig. 2). Marine habitats of the atoll are described in detail by Williams (this volume); the reef front, which is relatively barren of living hard corals rises to the reef crest which is algal- veneered with surge channels at intervals of 50-250 m. The reef flat is covered by 1-2 m of water at high tide and part of it dries at low tide. The northern part of the lagoon averages around 15 m deep and is covered with dead coral or sand (see Smithers, this volume). The southern part of the lagoon is shallow, but contains a network of ‘blue holes' (see Chapter 3, Fig. 5). Individual holes are 12-20 m deep, but their rims are emergent at low spring tides. Extensive sand flats and sand aprons occur around the margin of the lagoon (see Smithers, and Williams, this volume). METHODS Surface sediment characteristics were examined in exposures of lithified sediments, and along surveyed transects. Subsurface investigations were carried out by drilling and by seismic survey. The results of seismic reflection surveys within the Cocos lagoon are described by Searle (this volume). Seismic refraction on the islands was undertaken using a 12-channel Geometrics seismograph. Drilling was undertaken during several visits. In addition to drilling, aimed specifically at unravelling the geomorphological history of the islands, drillcore logs and in some cases cores drilled as part of a water resources survey (see Falkland, this volume) were also examined. An initial reconnaissance visit was made by Woodroffe in 1986. In 1988 Woodroffe, McLean and Wallensky undertook drilling with a portable, trailer mounted Jacro 105 rotary drill, to depths of up to 9 m on Home Island, West Island and Pulu Wak Banka. In 1990 deeper drilling was undertaken using a Jacro 500, operated by P. Murphy; holes were sunk on Horsburgh, South, West and Pulu Blan Madar Islands. In 1991 seismic reflection surveys were undertaken in the lagoon, and exploratory drillholes were put down to 2.3 m using a hand-held Mindrill on the reef flat to the east and south of West Island. Several vibrocores were taken from the southern and eastern part of the lagoon (Smithers et al., in press), using 75 mm diameter aluminium pipe, vibrated into the lagoon floor with a concrete vibrator. Three long cores of up to 4.2 m length were recovered, together with several shorter ones. Sediment compaction was around 30% in most vibrocores. Recovery in drillholes varied, but was rarely greater than 70%. Samples of coral and Tridacna recovered in the cores were submitted for radiocarbon dating. Samples of coral from the conglomerate platform, or from pits within islands were also submitted for dating. Radiocarbon dating was undertaken principally at the ANU Radiocarbon Dating Laboratory. Marine carbonate samples usually require an environmental correction for ocean reservoir effect because organisms at their time of death are already somewhat depleted in radiocarbon, as the oceanic reservoir. has a substantial circulation time. Marine carbonates therefore have an apparent age at the time of their death. For marine shells from the Australian coast this environmental correction varies, but is generally minus 450 + 35 years from the conventional radiocarbon age (Gillespie and Polach 1979). On coral atolls there has been some questioning of the estimate of the correction factor, and indeed whether any correction needs to be made. In particular, Pirazzoli has argued that no correction should be necessary for dating of coral samples from within the lagoons of Tuamotu atolls, where there may be limited exchange of waters with open ocean (Pirazzoli et al. 1987). We have examined this in Cocos, taking advantage of earlier coral collections, and by dating samples of corals collected by Wood-Jones in 1906 and Gibson-Hill in 1941 (pre-1950 samples of known age are necessary for such dating as post-1950 samples have elevated radiocarbon levels as a result of radiocarbon released by bomb tests). Results are shown in Table 1, indicating the average correction to be 460 years. The correction is indeed around 450 years, and that is the value that we have used throughout this paper to conform with similar studies elsewhere. PLEISTOCENE LIMESTONE Boreholes undertaken as a part of a Water Resources survey of the Cocos (Keeling) Islands (see Falkland, this volume), together with our own drillholes, indicate that a well-lithified, but porous, limestone underlies the poorly consolidated coral shingle and sand deposits of the Cocos reef islands, at depths of around 11-14 m (see Chapter 2, Table 10 and Chapter 3, Fig. 1). This older limestone contains corals, with some travertine deposits in voids, and in at least one drillcore, cemented oolites. Uranium-series disequilibrium dating of a sample of coral at the top of this facies, from 12.6 m depth (10.5 m below MSL) in drillhole WI1 gave an age of 118,000 + 7000 years B.P. on a bulk sample, but after preparation removing calcite under binocular microscope, the age determined was 123,000 + 7000 years B.P. (Woodroffe et al. 1991). The age of this limestone suggests it formed during the last interglacial, when the sea was around or slightly higher than present, about 125,000 years ago. The morphology of the Pleistocene atoll has been revealed in greater detail by seismic reflection profiling across the lagoon, results of which are discussed by Searle (this volume). Woodroffe et al. (1991) argued that the atoll had subsided at a rate of about 0.1 mm/yr based upon subsidence of this surface from above present sea level to 8-11 m below present sea level. The seismic results indicate that the surface actually has a considerable slope on it, becoming much deeper with distance into the lagoon, towards the centre of which it is more than 24 m below present sea level. This morphology seems likely to result from solution during subaerial exposure of the atoll when the sea was low, and Searle (this volume) suggests that the subsidence rate may be only 0.02 mm/yr. CONGLOMERATE PLATFORM An important feature of reef islands on Cocos is a platform of coral conglomerate which underlies most of the islands on the atoll rim. The platform was termed ‘brecciated coral-rock’ by Darwin (1842), 'reef conglomerate’ by Guppy (1889), and ‘breccia platform’ by Wood-Jones (1912). This near-horizontal conglomerate platform comprises cemented clasts of coral shingle or rubble, found especially along the oceanward shore of many of the islands, but also underlying part of the islands as shown in pits and wells. It occurs up to 0.5 m above MSL, and is thus inundated by the highest tides. Individual coral boulders of Porites of up to 2 m in diameter occasionally protrude from the platform, the highest points of which may reach up to 1 m above MSL. In some places, notably on North Keeling and Horsburgh, and at the southwestern end of West Island, the surface of the conglomerate platform is composed of arcuate, seaward-dipping beds of cemented coral cobbles. These appear to have been interpreted as former reef crests by Guppy (1889), but there is nothing in their composition to substantiate this. Instead, they resemble the foot of the modern beach where there is a rubble component, and we interpret them as beach conglomerate, marking the former position of rubble-strewn beaches. Guppy (1889) indicated that compositionally the conglomerate platform resembles the modern reef flat. The reef flat is characteristically 1.0-1.5 m lower than the surface of the platform, which is undergoing erosion on its seaward side. The reef flat often forms a hard, relatively smooth surface with encrustation by calcareous algae. This veneer may cover formerly truncated conglomerate platform. On the basis of constituent materials, gross fabric and surface morphology, we have interpreted the conglomerate platform as a fossil emergent reef flat (Woodroffe et al. 1990a, 1990b). Radiocarbon dates on corals from within the conglomerate platform indicate a spread of ages from 4010 + 85 to 3050 + 85 years B.P. (Table 2, Fig. 2). There are a number of sites at which apparently in situ fossil Porites microatolls, both massive and branching, have been found within the conglomerate, and which further serve to indicate that sea level was higher than the modern sea level when the conglomerate was formed. These provide a discontinuous record of the pattern of sea- level change over the late Holocene, and are discussed in greater detail below. Shallow drilling on Home Island and Pulu Wak Banka on the eastern rim of the atoll indicates that the platform is generally better cemented, and consists of coarser clasts nearer to the ocean. Shingle sticks of Acropora, often cemented by calcareous algae, form a major component of oceanward drillholes. Similar drilling on the modern reef flat has revealed that that it is also underlain by Acropora sticks cemented by calcareous algae. On lagoonward exposures of the platform drilling often encountered sand at 1-2 m depth. HOLOCENE REEF GROWTH The stratigraphy and chronology of the Holocene reef rim was examined along a series of transects around the atoll (Fig. 2), with drillholes through islands and the conglomerate platform which surrounds and underlies them, and through the reef flat. Figure 3 shows a cross-section (transect I) of the southern part of Home Island where observations were obtained from a trench and associated drillhole in the trench floor (CK7). The conglomerate platform was encountered in the floor of the trench at the same elevation that it outcrops on the oceanward side of the island. A radiocarbon date of 5760 + 95 years B.P. at 310 cm (-2.4 m below MSL) was obtained from the drillhole CK7. The island sediments, which are discussed in more detail in the next chapter, range in age from 1840 + 125 years B.P. to 1440 + 80 years B.P. The age of the conglomerate platform on Home Island is indicated by a radiocarbon date of 3680 + 105 years B.P. on a coral cemented into the top of the platform at the site of CK1, a drillhole in a sequence (transect II) on the platform at the southern end of this island. The cross-section at this point includes drillhole CK3 which is on the oceanward edge of this platform and is more then 6 m deep. This core contains coral shingle, generally well-cemented with calcareous algae, and at its base is dated 6160 + 95 years B.P. (see Table 3). The ages above are reversed, but their errors render them statistically indistinguishable. There is apparently a decrease in age as the lagoon is approached with a radiocarbon date of 3490 + 85 years B.P. at 2.4 m below MSL in the lagoonward core CK2. The platform is less well cemented in this core, and drilling was aborted in sand. A similar sequence of drillholes was drilled through the conglomerate platform on transect III at the southern end of Pulu Wak Banka. The channel south of this island contains numerous living microatolls. A fossil microatoll was identified within the conglomerate, and this has been radiocarbon dated at 1960 + 80 years B.P. (Fig. 5), indicating that some material has been added to the conglomerate platform on the margin of the channel in the last 2000 years. A coral within the conglomerate has been dated at 3220 + 85 years B.P. just near the transect (Fig. 2). The conglomerate platform becomes thinner closer to the lagoon, and drilling in CK6 was aborted in sand. A core into the sand spit extending from the southern end of Pulu Wak Banka into the lagoon, contains coral shingle at about 1 m depth, which has been dated 3170 + 85 years. It can also be inferred from this date that the lagoon has partially filled since the time of conglomerate platform formation. Figure 6 indicates the stratigraphy beneath South Island on transect IV, and Figure 7 indicates the stratigraphy beneath Pulu Blan Madar. Both are similar, intercepting the Pleistocene limestone at 13.8 m depth (11.6 m below MSL) and 12.6 m depth (11.2 m below MSL) respectively. The cores recovered shingle or shingle and sand. Coral fragments at 10 m in CK15 dated 6790 + 80 years B.P. and Tridacna at 6 m in CK14 dated at 6040 + 80 years B.P. Figure 8 shows shallow cores into the reef flat on transect VI, south of Pulu Maria. The reef flat appears about 1000 years older close to the oceanward edge, than beneath the island 750 m lagoonward. Radiocarbon dates of 5800 +70 years B.P. at 2.15 m depth, and 5630 + 205 years B.P. at 0.6 m depth in CK21 (not statistically significantly different), compare with a date of 4740 + 85 years B.P. at 2.1 m depth in CK23 (Fig. 8). The conglomerate platform on Pulu Maria contains only relatively fine clasts, and recovery in CK23 was poor, but generally also indicated weakly cemented sand. Figure 9 shows transect VI across the spits at the eastern end of West Island. Radiocarbon dates on coral shingle, recovered from a series of pits into the sandy spits, demonstrate progressive development of the spits from around 1400 years ago to present (Table 6), with the last spit giving a modern age. CV1 is a vibrocore sunk into the muddy sediments flanking the recent spit and penetrating sand and shingle, and from that vibrocore a date at around 0.9 m (about halfway down the core, allowing for compaction of the core) is 3240 + 85 years B.P.. This date appears to reflect lagoonal infilling, which must have occurred before the spits began to form. Though hard pan was encountered beneath some spits, this area is not underlain by typical conglomerate platform. Figure 10 at the southern end of West Island is a combination of several different drillholes undertaken at different times, and amalgamated schematically into a single transect (Transect VIII). Pleistocene limestone was encountered at 6.5 m below MSL. in CK13, but was not encountered in CK9 which went slightly deeper. CK13, drilled next to a telok (lagoonlet), but through a shingle substrate, encountered mud at 2-4 m depth, similar to that being deposited in the telok, implying that the island sand and shingle have been deposited over the surface of a formerly larger telok (Fig. 10). The sequence of 4 radiocarbon dates from drillholes CK13, CK9 and CK10A show the opposite trend to that generally observed on other transects, in that the older dates (6140 + 85 years B.P. at 4.4 m in core CK13) are to lagoonward, with younger dates beneath the reef flat (4770 + 85 years B.P. at 1.5 m in CK10A). Figure 11 shows transect IX which is at the southern boundary of the Quarantine Station on West Island and crosses the island where it is both especially low and particularly narrow. The most interesting feature of this transect is that there are a number of microatolls, up to 2 m in diameter, which are found along the shore, above the modern limit to coral growth. This together with elevated beachrock at this site provides convincing evidence that the sea has been relatively higher than it is at present. Radiocarbon ages of 2690 + 85 and 2730 + 85 years B.P. have been determined on two massive Porites microatolls at this site, where they are about 50-60 cm above their modern, living equivalents. In addition a sample of Porites, almost certainly a microatoll, was recovered from drillhole CK8 and dated 3190 + 85 years B.P. at a very similar elevation, indicating that similar, though in this case slightly older microatolls continue beneath the island sediments. The significance of these emergent, in situ corals will be examined below. Figure 12 shows the stratigraphy of two deep holes drilled on transect X on Horsburgh Island. The Pleistocene substrate was only encountered in the more oceanward drillhole, where it was found at a depth of 13.1 m (10.6 m below MSL). A coral sample from a massive coral colony recovered from almost directly above the Pleistocene/Holocene contact was radiocarbon dated 5540 + 80 years B.P., while a coral from 4.8 m has been dated 5260 + 80 years B.P.. These samples are more than 8 m apart and imply a rapid rate of reef accretion in the order of 25-30 mm/yr. CK11 at the lagoonward end of Horsburgh Island was presumably not drilled deep enough to encounter the Pleistocene surface, as subsequent seismic profiling in that part of the lagoon has indicated a reflector, believed to represent the last interglacial surface at depths of more than 20 m. Nevertheless the date of 4610 + 85 years B.P. on a sample of Tridacna at the base of that core, indicates that sedimentation here lagged about 1000 years behind that on the more oceanward side of Horsburgh. MICROATOLLS AND HOLOCENE SEA LEVEL The radiocarbon dates from Tables 3, comprising dated samples from transects I- X, have been plotted on an age-depth plot in Figure 13. These ages do not permit the accurate reconstruction of early to mid-Holocene sea-level history on Cocos, because it cannot be established that the corals in the cores are in their position of growth, and even if they were in situ they could have grown in water depths of up to several metres. Samples from the conglomerate platform (Table 2) on the other hand are manifestly not in growth position, and do not indicate the level of the sea at time of deposition. Some corals, however, do record former sea level. Microatolls are flat-topped colonies of coral which have been constrained in their upward growth by subaerial exposure at low spring tides, and have therefore continued to grow only laterally (Scoffin and Stoddart 1978). Their upper surface is related to sea level and the upper surface of fossil microatolls can be used to reconstruct late Holocene sea-level change (Chappell 1982). On the Cocos (Keeling) Islands there are modern, living microatolls, of massive and branching Porites, on the reef flat, in interisland passages, and within the lagoon (Woodroffe and McLean 1990). These corals were described by Wood-Jones (1912), though he attributed their form to sedimentation on their upper surface. Detailed survey of modern microatolls around the atoll indicates that they occur in a relatively narrow elevational range (around 0.3 m below MSL), and supports the idea that they are limited by water level (Smithers, unpublished results). Fossil microatolls, though by no means common, have been identified within the conglomerate platform at several sites on West Island, and on Pulu Pandan and South Island. At the southern end of West Island fossil microatolls of both massive and branching species of Porites are found together. Fossil microatolls can also be identified on the reef flat, and at one location in a telok on South Island, where the oldest microatolls so far dated on Cocos have been found (3560 + 85 years B.P., see Fig. 2). The exact elevation of these remains uncertain, but they appear to be lower than younger specimens on West Island (Table 4). In addition, Porites cored in CK8 (CK8.1B/2) almost certainly represents a microatoll dated 3190 + 85 years B.P. and at a similar elevation to those on the present shoreline (see transect IX, Fig. 11), and so too does that in CK5 (CK5.1B), dated 1960 + 80 years B.P. and found at around MSL (see transect Ill, Fig. 5). The upper surface of these microatolls gives an indication of the elevation of sea level. When compared with the modern elevation of microatolls (around -0.3 m MSL), these corals indicate a trend of gradually falling sea level (Fig. 13), from about 0.9 m above present around 3000 years ago, on the basis of branching microatolls on Pulu Pandan and slightly less if massive microatolls on West Island are considered. A discontinuous record of relative sea-level fall is preserved at the foot of the beach near the Quarantine station on West Island, where a series of microatolls is located (see plot in Fig. 13). LAGOONAL INFILL The nature of lagoonal sediments in the South Keeling Islands has been examined by Smithers (this volume), who demonstrates that sediments in the lagoon range from strongly fine skewed gravelly muds (as in the teloks and blue holes) to gravelly sands where sand aprons have encroached upon lagoonal corals. Compositionally they are entirely biogenic, dominated by three factors, coral sediments, molluscan sediments and sediments in which calcareous algal fragments and Halimeda plates are an important constituent (Smithers et al., in press). The lagoon is particularly shallow around much of the southeastern corner, and the surface, which is covered by seagrass, dries out at low tide (Williams, this volume). There is anecdotal evidence that it has filled in rapidly in historical times. Captain John Clunies Ross established his settlement on the middle portion of South Island, where access is now extremely difficult for a boat of any draft at almost any stage of the tide. Rapid infill has been inferred by Forbes (1879, p779). Guppy detected sediment in suspension being carried into the lagoon through the passages, by the predominantly unidirectional currents. He made a series of calculations of sediment transport and sedimentation into the lagoon (Guppy 1889). His estimates were based upon rates of coral growth and sediment production, distributed across the area of the reefs and lagoon that contained coral cover. He calculated that 5000 tons of sediment was carried into the lagoon each year. The majority (5/6) he considered to be deposited in the first 700 m of the lagoon, on the sand aprons. He estimated that these aprons were prograding at a rate of around 1 m/yr (1 yard per year). Vertical sedimentation averaged over the southern part of the lagoon, Guppy estimated to be 1ft/100 years (c. 3mm/yr). Extending these calculations to the northern lagoon, Guppy considered that the lagoon would require a further 4000 years to infill, and that the total time from initiation to complete infill for a lagoon would be about 15-20,000 years. Wood-Jones, not only realised the importance of this sediment accumulation in the lagoon, but he interpreted sedimentation as the prime control on the formation of the atoll. He compared the atoll as a whole with single colonies of Porites microatolls (he termed them an ‘atoll reef in miniature’, Wood-Jones 1912, p108-109), which he interpreted to be limited in their upward growth by sediment accumulation on their upper surface. Wood-Jones proposed his sedimentation theory of atoll development in Opposition to Darwin's subsidence theory, and the solution theory of Murray. We have examined lagoonal sedimentation in Cocos, based on a series of vibrocores taken in the southern and eastern lagoon. The stratigraphy, sediment grain- size and components, and radiocarbon dating from vibrocores indicate spatial and temporal variations in the nature and rate of sedimentation, controlled primarily by the pattern of sea-level change and the response of the atoll environments, particularly the formation of reef islands on the atoll rim (Smithers et al., in press). The main contrast is between sand apron sediments, on the one hand, which are composed of skeletal grains typical of a reef flat assemblage, being coarse, clean sands and shingle, with fragments of the algal rhodoliths, Spongites sp., and island lee sediments, on the other hand, which are higher in mud content, with occasional coral fragments. The base of vibrocores contains more shingle, and coral, algae and Halimeda are generally more common, perhaps reflecting lagoon reefs which have been covered by sand apron and island lee sediments. Sand aprons have encroached episodically into the lagoon and sand appears to have spilled into blue holes as the sedimentation front advanced. It seems highly probable that the sands of the southeastern section of the lagoon have already filled over a patchwork of blue holes, and this may explain the patchy penetration of vibrocores; the shorter ones reaching shingle at much shallower depths than those which penetrated into former sand-filled blue holes. The sands radiocarbon dated in vibrocores were all younger than 4000 years B.P. (Table 5); the base of CV11 dated 3850 + 80, while CV1 and CV12 had dates of 3240 + 85 and 3530 + 80 years B.P. respectively. These older dates are in those cores closest to islands, and consequently also close to the reef. Cores further into the lagoon had younger dates: CV15 dated 420 + 65, and CV10 910 + 80 at its base 2.2 m below the sediment surface, and 130 + 110 years B.P. in the centre, recording the present progradation of the sand sheets into the area of blue holes. Radiocarbon dates record the time of death of the coral shingle, and not the time of its deposition. Sediments flooring the lagoon are also likely to be subject to considerable bioturbation. Nevertheless, despite minor age reversals in vibrocores such as in CV2, the dates are generally stratigraphically consistent and indicate the general trend of sedimentation. Vertical accumulation rates are higher in sand aprons than in island lee sediments, being 0.5-1.0 mm/yr in the former, and in the latter, over the last 2000 years, ranging from 0.25-0.5 mm/yr (Smithers et al., in press). 10 HOLOCENE EVOLUTION OF THE ATOLL The Holocene atoll has developed over a Pleistocene limestone surface, which has been shown by seismic reflection profiling to be basin shaped probably as a result of solutional weathering during the glacial sea-level low (Searle, this volume). This surface has been flooded by the sea during the postglacial marine transgression. Seismic and drilling results indicate that there is a relatively continuous Pleistocene rim at about 9-10 m below MSL, around the western, southern and eastern sides of the atoll, with a deeper basin to the north, which opened out to the northeast, and perhaps also northwest. When the sea was 12-15 m below present, the Pleistocene rim was still emergent, and lagoonal exchange must have been predominantly through the northern passages. Since that surface has been inundated, as a result of the final stage of the postglacial marine transgression, there has been a phase of rapid vertical reef growth, following the rising sea level, recorded by radiocarbon dates from cores, and shown in Figure 13. During periods of rising sea level reef growth has adopted one of three strategies, keep-up (where the reef closely tracks the rising sea), catch-up (where reef growth lags behind sea level) and give-up in which there is negligible net reef growth (Neumann and Macintyre 1985). As described above, the coral dates do not indicate the position of the sea, except for during the last 3000 years where there are dated microatolls which have been constrained by water level. Regional sea-level curves indicate a sea-level history in which sea level rose rapidly up until around 6000 years ago when it reached a level close to its present level and has changed by only a metre or so since (Thom and Chappell 1975, McLean et al. 1978, Geyh et al. 1979, Thom and Roy 1985). Vertical reef accretion on Cocos appears to have lagged behind sea level, as also shown on atolls in the Pacific (Marshall and Jacobsen 1985). The three modes of response can be found at different points around an atoll. Reef growth on Cocos has varied from place to place; nowhere does it seem to have kept up with sea level (there are no 6000 year dates at present sea level), but it has lagged behind sea level by different amounts at differing points on the atoll rim. We identify this period of catch-up reef growth as the first of three phases in the Holocene development of the atoll. The second phase was a period of reef flat consolidation, represented by the conglomerate platform. On the basis of fabric and morphology we attribute the conglomerate platform to formation as a reef flat under conditions of sea level slightly higher than present. That the sea was higher than present is shown most convincingly by the presence of microatolls above the modern limit to coral growth (Table 4). Other data, such as the elevated beachrock (in Fig. 11), and the consistent difference between the conglomerate platform and the modern reef flat, also substantiate that the sea was relatively higher in the mid-Holocene. Elsewhere similar conglomerate platform has been interpreted as lithified storm- rubble ridges or ramparts similar to those deposited as a result of Tropical Cyclone Bebe on Funafuti in Tuvalu in the Pacific, in front of the elongate reef islands, and subsequently observed to migrate landwards and redistribute over the shoreface (Baines and McLean 1976). We discount this interpretation of the conglomerate platforms on Cocos because of the horizontal nature and width of the platform, and the relatively narrow range of radiocarbon ages from coral clasts. Storm rubble is evidently a component of the platform, with addition of material under non-storm conditions, bound by biological and chemical processes, in a similar way that material is supplied to and incorporated into the interisland reef flat areas on the modern atoll. Reef blocks, as seen in the Pacific storm belt, are relatively rare, though there are blocks of more than 1 m 11 diameter on the reef flat at the southern end of the atoll, one of which has been dated to 610 +75 years B.P. (Table 6). The third phase is a phase of reef island development. The modern reef islands lie primarily on an oceanward outcrop of conglomerate platform, and the appearance of islands in their modern location and form must therefore postdate the formation of the platform. Islands have formed during the last 3000 years when the sea level has been undergoing a relative fall. The age structure of reef islands is still poorly known; some dates are given in Table 6, indicating substantial deposition in the period 1800-1000 years B.P. The issue is examined in greater detail in the next chapter (Woodroffe and McLean, this volume). Progradation of the southern end of West Island in the last 1400 years is apparent from Figure 9. The three phases that are identified (Fig. 13), have not necessarily been discrete, but some overlap between them is likely. Reefs at the southern end of the atoll appear to have grown fastest, and although they did not keep up with sea level, they lagged only slightly (<1000 years) behind sea level, whereas reefs at Horsburgh appear to have undergone a greater lag before commencing to grow, but to have accreted vertically at a faster rate. Similarly the early stages of island formation appear to have occurred within the final stages of conglomerate platform development, as indicated by beachrock and beach conglomerate outcrops within the platform (i.e. Horsburgh, West Island, and North Keeling). Progression from one phase to another must have been accompanied by substantial changes of energy regime, particularly in the lagoon, which must have been a relatively high energy environment before the reef rim caught up with sea level, and subsequently underwent a further reduction in energy as reef islands were formed around the margin. DISCUSSION This three phase model of the Holocene evolution of the atoll incorporates components of each of the earlier theories on the development of atolls. We now examine some of these issues. The first issue at Cocos is whether the atoll has subsided, and whether it is continuing to subside. As discussed above, the depth of the last interglacial surface is taken by us as an indication of subsidence. The exact elevation of the sea at the peak of the last interglacial is contentious; in some parts of the world it is considered to have been around 5-6 m above present, elsewhere up to 10 m above present. Lambeck and Nakada (1992) indicate that flexural responses of the earth's crust and upper mantle need to be taken into account, and that it need not have been any higher than present. There are many parts of the world where the last interglacial reef is found above present sea level, the nearest being Christmas Island, where coral-bearing reef reaches 12 m above present sea level, but with associated deposits reaching 30m (Woodroffe 1988). We believe that the last interglacial reef on Cocos would have caught up with sea level during the oxygen-isotope Se sea-level high stand, and we attribute the fact that it is everywhere below present sea level to subsidence. Seismic results indicate that solution is likely to have occurred and deepened the lagoon, but we attribute the fact that contemporaneous limestones are around 12 m above sea level on Christmas Island and 12 m below sea level on Cocos to uplift of the former and subsidence of the latter. Indeed if the atoll were not subsiding, there would be no reason why interglacial limestones should occur in layer-cake fashion beneath the Holocene, as appears to be indicated by seismic profiling (Searle, this. volume). 12 Darwin himself realised that confirmation of his subsidence theory would come from deep drilling of coral atolls. He postulated that such drilling should reveal extensive thicknesses of shallow water limestones, in excess of the present depth range over which corals can grow. The final proof came after World War II with drilling of Enewetak, Mururoa and Midway atolls in the Pacific, in which basalts were encountered beneath the coral limestones at depths down to 1200 m (Ladd et al. 1953, Emery et al. 1954, Lalou et al. 1966). Deep-drilling has not been undertaken on Cocos, and so the ies of the volcanic basement, probably around 400-500 m (Jongsma 1976), is still unknown. Confirmation of the volcanic basement of open-ocean atolls, and more recently demonstration of the way in which plate tectonics can provide a mechanism (horizontal plate motion), by which subsidence can occur (Scott and Redondo 1983, Grigg and Epp 1989), give strong support to Darwin's theory of atoll origins. Relative and absolute dating of the basement volcanic rocks, and palaeontologic and diagenetic changes within the limestone, indicate that subsidence rates are “imperceptibly slow except in geological perspective” (Stoddart 1973). While we believe that gradual subsidence is continuing at Cocos, we are unable to accept the local evidence that Darwin invoked to prove his theory. Darwin was shown erosion of the shoreline on West Island, with undercutting of coconuts, which he believed was “tolerably conclusive evidence” of subsidence. His interpretation of this geomorphological evidence has been disputed by several other scientists who have visited the atoll. Most vehement of those disagreeing with Darwin was John Clunies Ross, who was resident on Cocos but, much to his subsequent regret, had been absent on the occasion of Darwin's visit. Evidently concerned at Darwin's suggestion that the islands of which he was in possession were about to disappear beneath the sea, he claimed that a “moderable attentive investigation of the Cocos islets affords ample reasons for believing that they have stood up to the present time above the level of the ocean during hundreds if not thousands of years” (Ross 1855). Guppy described each of the reef islands but decided that there was evidence for neither recent uplift nor recent subsidence of the atoll (Guppy 1889). On the other hand Forbes, and later Wood-Jones, believed that there was evidence that the sea had been higher with respect to the atoll. Wood-Jones wrote “the undermining of trees and the denudation of shore-lines do not necessarily indicate subsidence, for they are inconstant effects, and an area of land denudation is compensated for by an area of land construction at another part of the island ring” (Wood-Jones 1909 p674). Forbes had similarly argued that the erosion of islands was compensated by the debris being deposited further around the shore, and interpreted elevated fossil shells of clams and oysters on Workhouse Island (south of Direction Island, no longer existing as a distinct island), as an indication of former higher sea levels (Forbes 1879, 1885). The views of Forbes and Wood-Jones for a sea level higher than present revolved around their interpretations of the coral conglomerate that underlies much of the reef islands. Our data confirm the interpretation of recent emergence (that is a slight fall of sea level relative to the atoll) since the mid-Holocene (Woodroffe et al. 1990a, 1990b). However, although the geomorphological data on the surface morphology of the atoll were misinterpreted by Darwin, the overall atoll structure we do attribute to gradual subsidence as Darwin postulated. 13 Radiocarbon dates from western, eastern and southern reef flat or oceanward conglomerate platform, imply that the reef near the reef crest lagged only about 1000 years behind sea level. The reef caught up first near the reef margin, and there was more gradual infill to lagoonward; thus on each of these sides we see radiocarbon ages getting younger with distance into the atoll lagoon. This is contrary to most of the views of other geologists on the atoll. Thus Guppy, following the suggestion of Murray that reefs prograded out over their own talus, considered that there was proof that the reef flat had built out in a series of steps. His view appears to have been based upon description of the reef buttresses off the atoll given him by George Clunies-Ross, and his interpretation of the fossil reef rims as he viewed imbricated, arcuate ridges on North Keeling and Horsburgh Islands. The latter he called parallel lines of old reef margins that protrude above the reef flat, but we have interpreted these as beach deposits, marking instead the former foot of rubble-strewn beaches. Wood-Jones similarly supposed that the breccia (conglomerate platform) was oldest towards the lagoon and youngest toward the ocean (Wood-Jones 1909); whereas our results indicate the reverse trend. In Figure 14 we summarise the late Quaternary development of the Cocos (Keeling) Islands. Wood-Jones recognised three theories, Subsidence, Solution and Sedimentation (ie. the theories of Darwin, Murray and himself respectively), to which we may add Sea-level (the glacial control theory of Daly). The surface of the last interglacial at depths of 10-20 m below sea level indicates that the island was not planated as Daly has suggested at the time of sea level low, although our interpretation does emphasise sea-level fluctuations which have been the principal control over the late Quaternary of periods of reef establishment and their demise. We also interpret that it indicates that the atoll is continuing to subside as Darwin envisaged, although at a rate that is imperceptibly slow, even compared with rates of sea-level fluctuation. The form of the lagoon does not result from solution as Murray envisaged, but nevertheless solution does appear to have been important at times of low sea level in accentuating, through karst erosion, the basin-shape of the lagoon as inferred by Purdy (1974) in his antecedent karst hypothesis. Holocene reef morphology mimics this antecedent surface as suggested by Purdy. Finally, the lagoon has been one of the major areas of sediment accumulation, along with reef islands, over the last 3000 years, as envisaged by Wood-Jones, but we interpret this not as the cause of atoll morphology, but more as a response to the evolving morphology. ACKNOWLEDGMENTS This research has been funded by the Australian Research Council, National Geographic Society, and the Department of Arts, Sports, Environment and Territories. We are grateful to the Cocos (Keeling) Islands Administration, Cocos Island Council, the Cocos Co-operative, and Australian Construction Services for assistance on the island. We thank Peter Murphy, Scott Smithers and Paul Kench for their help in the field. 14 REFERENCES Armstrong, P. 1991. Under the blue vault of heaven: a study of Charles Darwin's sojourn in the Cocos (Keeling) Islands. Indian Ocean Centre for Peace Studies, University of Western Australia. Baines, G.B.K. and McLean, R.F. 1976. Sequential studies of hurricane bank evolution at Funafuti atoll. Marine Geol. 21: M1-M8&. Bezrukov, P. L. 1973. Principal scientific results of the 54th cruise of the R.V. Vitian in the Indian and Pacific Oceans (Feb-May 1973). Oceanology, 13: 761-766. Chamberlain, N. G. 1960. Cocos Island magnetic survey, 1946. Bureau of Mineral Resources, Record, 1960/124. Chappell, J. 1982. Evidence for smoothly falling sea levels relative to north Queensland, Australia, during the past 6000 years. Nature 302: 406-408. Chappell, J. and Shackleton, N.J. 1986. Oxygen isotopes and sea level. Nature 324: 137- 140. Daly, R.A. 1915. The glacial-control theory of coral reefs. Proc. Am. Acad. Arts Sci. 51: 155-251. Daly, R. A. 1934. The changing world of the Ice Age. New Haven: Yale University Press. Darwin, C. 1842. The structure and distribution of coral reefs. London: Smith, Elder and Co. Davis, W.M. 1928. The coral reef problem. Am. Geog. Soc. Spec. Pub. 9, 596pp. Emery, K.O., Tracey, J.I. and Ladd, H.S. 1954. Geology of Bikini and nearby atolls. U.S. Geol. Surv. Prof. Pap. 260-A: 1-265. Finlayson, D. M. 1970. First-order regional magnetic survey at Cocos Island, Southern Cross and Augusta. Bureau of Mineral Resources, Record, 1970/101. Forbes, H.O. 1879. Notes on the Cocos or Keeling Islands. Proc. Roy. Geog. Soc. 1: 777-784. Forbes, H.O. 1885. A Naturalist's wanderings in the Eastern Archipelago: A narrative of travel and exploration from 1878 to 1883. London, Sampson Row. Geyh, M.A., Kudrass, H-R. and Streif, H. 1979. Sea-level changes during the late Pleistocene and Holocene in the Strait of Malacca. Nature 278: 441-443. Gillespie, R. and Polach, H.A. 1979. The suitability of marine shells for radiocarbon dating of Australian prehistory. Proc. 9th Int. Conf. on Radiocarbon. Univ. California Press, 404-421. 15 Grigg, R.W. and Epp, D. 1989. Critical depth for the survival of coral islands: effects on Hawaiian Archipelago. Science 243: 638-641. Guppy, H. B. 1889. The Cocos-Keeling Islands. Scott. Geogr. Mag. 5: 281-297, 457- 474, 569-588. Jacobsen, G. 1976. The freshwater lens on Home Island in the Cocos (Keeling) Islands. BMR J. Aust. Geol. Geophys. 1: 335-343. Jongsma, D. 1976. A review of the geology and geophysics of the Cocos Islands and Cocos Rise. Bureau of Mineral Resources, Record, 1976/38. Ladd, H.S., Ingerson, E., Townsend, R.C., Russell, M. and Stephenson, H.K. 1953. Drilling on Eniwetok Atoll, Marshall Islands. Am. Ass. Petrol. Geol. Bull. 37: 2257-2280. Lalou, C., Labeyrie, J. and Delebrias, G. 1966. Datations des calcaires coralliens de l'atoll de Muroroa (archipel des Tuamotu) de l'époque actuelle jusqu’a -500,000 ans. C. R. Acad. Sci. Paris 263-D: 1946-1949. Lambeck, K. and Nakada, M. 1992. Constraints on the age and duration of the last interglacial period and on sea-level variations. Nature 357: 125-128. Lyell, C. 1832. Principles of Geology. London: Murray. Marshall, J.F. and Jacobsen, G. 1985. Holocene growth of a mid-Pacific atoll: Tarawa, Kiribati. Coral Reefs 4: 11-17. McLean, R.F., Stoddart, D.R., Hopley, D. and Polach, H.A. 1978. Sea level change in the Holocene on the northern Great Barrier Reef. Phil. Trans. Roy. Soc. Lond. A 291: 167-186. Murray, J. 1889. Structure, origin, and distribution of coral reefs and islands. Nature 39: 424-428. Neumann, A.C. and Macintyre, I.G. 1985. Reef response to sea level rise: keep-up, catch-up, or give-up. Proc. 5th Int. Coral Reef Congr. 3: 105-110. Pirazzoli, P. A., Delibrias, G., Montaggioni, L. F., Saliége, J. F., & Vergnaud-Grazzini, C. 1987. Vitesse de croissance latérale des platiers et évolution morphologique récente de l'atoll de Reao, tiles Tuamotu, Polnésie francaise. Ann. Inst. Océanogr. 63: 57-68. Purdy, E. G. 1974. Reef configurations, cause and effect. In L. F. Laporte (Eds.), Reefs in time and space (pp. 9-76.). Society of Economic Palaeontologists and Mineralogists Special Publication. Ross, J.C. 1855. Review of the theory of coral formations set forth by Ch. Darwin in his book entitled: Researches in Geology and Natural History. Natuur. Tijds. voor Neder. Ind. 8: 1-43. Scoffin, T.P. and Stoddart, D.R. 1978. The nature and significance of micro atolls. Phil. Trans. Roy. Soc. Lond. B 284: 99-122. 16 Scott, G. A. J., & Rotondo, G. M. 1983. A model to explain the differences between Pacific plate island atoll types. Coral Reefs 1: 139-150. Smithers, S.G., Woodroffe. C.D., McLean, R.F. and Wallensky, E. in press. Lagoonal sedimentation in the Cocos (Keeling) Islands. Proc. 7th Int. Coral Reef Congr. Stoddart, D.R. (editor) 1962. Coral Islands by Charles Darwin, with introduction, map and remarks. Atoll Res. Bull. 88. Stoddart, D.R. 1973. Coral reefs: the last two million years. Geography 58: 313-323. Thom, B.G. and Chappell, J. 1975. Holocene sea levels relative to Australia. Search 6: Thom, B.G. and Roy, P.S. 1985. Relative sea levels and coastal sedimentation in southeast Australia in the Holocene. J. Sedim. Petrol. 55: 257-264. Wood-Jones, F. 1909. The coral island question. Proc. Zool. Soc. 1909: 671-679. Wood-Jones, F. 1912. Coral and Atolls: a history and description of the Keeling-Cocos Islands, with an account of their fauna and flora, and a discussion of the method of development and transformation of coral structures in general. London: Lovell Reeve and Co. Woodroffe, C.D. 1988. Relict mangrove stand on Last Interglacial terrace, Christmas Island, Indian Ocean. J. Trop. Ecol. 4: 1-17. Woodroffe, C. D., McLean, R. F., Polach, H., and Wallensky, E. 1990a. Sea level and coral atolls: Late Holocene emergence in the Indian Ocean. Geology 18: 62-66. Woodroffe, C. D., McLean, R. F., and Wallensky, E. 1990b. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. National Geographic Research 6: 262-275. Woodroffe, C.D. and McLean, R.F. 1990. Microatolls and recent sea level change on coral atolls. Nature 344: 531-534. Woodroffe, C.D., Veeh, H.H., Falkland, A.C., McLean, R.F. and Wallensky, E. 1991. Last interglacial reef and subsidence of the Cocos (Keeling) Islands, Indian Ocean. Marine Geol. 96: 137-143. 17 Table 1. Radiocarbon dating results on museum specimens of coral, Cocos (Keeling) Islands. ANU Lab Coral species Date of Conventional Age Reservoir No. collection radiocarbon age _(pre-1950) correction 6151 Acropora scherzeriana 1906 [W-J] 370 + 60 40 330 + 60 6152 Montipora foliosa 1906 [W-J] 670 + 60 40 630 + 60 6153 Porites nigrescens 1906 [W-J] 410+ 60 40 370 + 60 (=P. cylindrica) 7638 § Montipora ramosa 1941 [G-H] 510+ 70 10 500 + 70 7639 __ Montipora lobulata 1941 [G-H] 480 + 60 10 470 + 60 Note: W-J = Wood-Jones, G-H = Gibson-Hill. 18 Table 2. ANU Lab No. 5411 5412 5414 5416 5417 5418 5419 5420 5421 6220 6221 6222 6224 7134 Pocillopora_merate transect VIII Radiocarbon dating results on conglomerate platform Cocos (Keeling) Islands. No. C4 C5 C16 C26 C43 C48 C59 C64 C65 CK1/5 C172 C158 C154 C202 Sample Material Coral Coral Coral, Porites Coral Coral Coral, Porites Coral Coral Coral Coral Coral Coral, Porites Coral Coral, Location from conglom- erate in base of well from conglom- erate platform beneath beach conglomerate from conglom- erate platform from conglom- erate platform cemented to conglomerate platform lower unit of conglomerate platform upper unit of conglomerate platform from conglom- erate platform from conglom- erate platform from conglom- erate platform from conglom- erate platform from conglom- erate on tran- sect IX In upper conglo- Island West Is. West Is. North Keeling Direction Is. Pulu Wak Banka North Keeling Pulu Labu Pulu Labu Home Is. Home Is. South Is. West Is. West Is. West Is. Conventional radiocarbon age 1770 + 70 3890 + 80 3480 + 80 3740 + 80 3670 + 80 4290 + 80 3950 + 80 3940 + 80 4000 + 80 4130 + 100 3500 + 80 4460 + 80 3550 + 80 3690 + 80 Environmentally corrected age 1320 + 80 3440 + 85 3030 + 85 3290 + 85 3220 + 85 3840 + 85 3500 + 85 3490 + 85 3550 + 85 3680 + 105 3050 + 85 4010+ 85 3100 + 85 3240 + 85 19 Table 3. Radiocarbon dating results from drillholes; Cocos (Keeling) Islands ANU Sample Material Location Island Elev- Conventional Environmentally Lab No. ation radiocarbon corrected No. (m)? age age 6223 C153 Coral frag from lagoonal Pulu Wak -1.0 3620 + 80 3170+ 85 ments infill Banka 6227 CK3-5B-4 Coral Depth 6.2m in Home Is. -6.1 6610 + 90 6160 + 95 core CK3 6641 CKI10A.1B Coral, 150cm in core West Is. -1.9 $220 + 80 4770+ 85 Porites on reef flat 6642 CK9.3B Coral, 230cm in core West Is. ~1.7 6000 + 80 5550 + 85 Faviid on conglomerate 6643 CK9.4B Tridacna 380cm incore West Is. -3.0 6370 + 90 $920 + 95 on conglomerate 6644 CK8.1B/2 Coral, 140cm in core West Is. +0.2 3640 + 80 3190 + 85 Porites through beachrock 6645 CK7.3B Coral 310cm in core Home ls. -2.4 6210+ 90 5760 + 95 6646 CKS5.1B Coral, Scm in core Pulu Wak 0.0 2410+ 70 1960 + 80 Porites on conglomerate Banka 6647 CK3.3B Coral 300cm in core Home ls. -3.0 5530 + 80 5080 + 85 on conglomerate 6648 CK3.2B Coral 140cm in core Home ls. -1.4 5610 + 80 5160 + 85 on conglomerate 6649 CK2.2B Coral In core Home ls. -2.4 3940 + 80 3490 + 85 on conglmerate 7546 CKI11/10B Tridacna 1305cmincore Horsburgh -11.1 5060 + 80 4610 + 85 Is. 7547 CK12/3B Coral 480cm in core Horsburgh -2.4 5710 + 70 5260 + 80 fragments Is. 7548 CK12/12B Coral 1300cmincore Horsburgh -10.6 5990 + 70 5540 + 80 Flaviid Is. 7549 CK13/5B_ Coral 440cm in core West Is -3.0 6590 + 80 6140+ 85 7550 CKI14/5B Tridacna 600cm incore South Is. = -2.5 6490 + 70 6040 + 80 7551 CKI15/6B Coral 1000cmincore PuluBlan -9.0 7240 + 70 6790 + 80 fragments Madar 8196 CK21215 Coral 215cm in core south of -2.4 6250 + 60 5800 + 70 on reef flat West Is. 8198 CK2160 Coral 60cm in core south of -0.8 6080 + 200 5630 + 205 on reef flat West Is. 8200 WI16/12 Coral 1220cmincore West Is. ? 41,100 + 890 8201 WII16/5 Coral 500cm in core West Is. ? 6170+ 70 5720 + 80 8404 HI12/15 Coral, 1550cmincore Homels. -14.0 7480+ 110 7030 + 115 Acropora 8197 CK17-235 Coral 235cm in core West Is. -2.5 6410+ 70 5960 + 80 8199 CK23-210 Coral 210cm in core Pulu Maria -1.7 $190 +70 4740+ 85 on reef flat Note: 2 metres relative to Mean Sea Level. 20 Table 4. Radiocarbon dating results on fossil microatolls; Cocos (Keeling) Islands ANU _ Sample Lab No. No. 5415 C18 6218 C156 6226 C174 6228 C155 7135 C 204 7136 C 206 7552 2 75534 75546 8408 4 8409 5 Note: Material Location Microatoll, branching Porites Microatoll, massive Porites Coral, Porites Microatoll, massive Porites Microatoll, massive Porites Coral, branching Porites Microatoll, Porites Microatoll, Porites Microatoll, Porites Microatoll, Porites Microatoll, Porites in situ within Pulu conglomerate Pandan platform in situ West Is. adjacent to C155 microatoll in South Is. lagoon sediments in situ beneath West Is. beachrock in situ West Is. in conglomerate in situ West Is. in conglomerate foot of beach West Is. foot of beach West Is. foot of beach West Is. foot of beach West Is. foot of beach West Is. 4 metres relative to Mean Sea Level. (m)? +0.6 +0.15 -0.15 +0.15 +0.35 +0.35 +0.1 +0.07 3400 + 80 3180 + 80 4010 + 80 3140 + 80 3690 + 80 3710 + 80 1500 + 60 2990 + 70 3470 + 80 3160 + 50 3430 + 60 Island Elev- Conventional Environmentally ation radiocarbon age _ corrected age 2950 + 85 2730 + 85 3560 + 85 2690 + 85 3240 + 85 3260 + 85 1050 + 70 2540 + 80 3020 + 85 2710 + 60 2980 + 70 Table 5. Radiocarbon dating results from vibrocores; Cocos (Keeling) Islands ANU Sample Material Location Island Conventional Environmentally Lab No. No. radiocarbon age corrected age 7531 CV2-240 Coral 240cm in vibrocore Lagoon 2780 + 100 2330 + 105 fragments 1532 CV2-300 Coral 300cm in vibrocore Lagoon 3050 + 90 2600 + 95 fragments 7533 CV2-414 Coral 414cm in vibrocore Lagoon 2660 + 130 2210+ 135 fragments 7534 CV3-80 Coral 80cm in vibrocore Lagoon 1670+ 110 1220+ 115 fragments 7535 CV3-240 Coral 240cm in vibrocore Lagoon 2980 + 70 2530 + 80 fragments 7536 CV10-110 Coral 110cm in vibrocore Lagoon 580 +100 130 + 105 fragments 7537 CV10-222 Coral 222cm in vibrocore Lagoon 1360 + 80 910+ 85 fragments 7538 CV12-50 Coral 50cm in vibrocore Lagoon 2520 + 110 2070 + 115 fragments 7539 CV12-158 Coral 158cm in vibrocore Lagoon 3980 + 80 3530 + 85 fragments 7540 CV1-90 Coral 90cm in vibrocore Lagoon 3690 + 80 3240 + 85 fragments 7541 CV5-66 Coral 66cm in vibrocore Lagoon 2490 + 80 2040 + 85 fragments 7542 CV6-48 = Coral 48cm in vibrocore Lagoon 1850 + 90 1400 + 95 fragments 7543 CV8-76 Coral 76cm in vibrocore Lagoon 2970 + 120 2520 + 125 fragments 7544 CV11-96 Coral 96cm in vibrocore Lagoon 4300 + 80 3850 + 85 fragments 7545 CV15-130 Coral 130cm in vibrocore Lagoon 870 + 60 420+ 70 fragments 8398 CV3 378 Coral 378cm in vibrocore Lagoon 3190 + 50 2740 + 60 fragments 8400 CV3 320 Coral 320cm in vibrocore Lagoon 3140 + 60 2690 + 70 fragments 8402 CV2 360 Coral 360cm in vibrocore Lagoon 3180+ 50 2730 + 60 21 fragments 22 Table 6. Radiocarbon dating results on reef island sediments; Cocos (Keeling) Islands ANU Sample Material Location Island Conventional Environmentally Lab No. No. radiocarbon age corrected age 5413 Gi, Coral boulder exposed North 2070 + 70 1620 + 80 in eroded ridge Keeling 6219 C171 Coral, reef block on southern 1060 + 70 610+ 80 Porites reef flat atoll rim 6225 C138 Coral, in bedded sand Home Is. 1890 + 70 1440+ 80 Pocillopora in wall of trench 7127 C72 Coral, In pit West Is. 1570 + 80 1120+ 85 Porites 7128 C 104 Bulk sand, _In pit, eastern West Is. 102.6 + 3.7 %M MODERN foraminifera ridge 7129 C106 ~=Coral, 80cm in pit West Is. 1550 + 90 1100 + 95 Pocillopora 7130 C118 ~~ Coral, 70cm in pit West Is. 830 + 100 380 + 105 Acropora qAtsy C133 Coralsand intrench Home Is. 2290 + 120 1840 + 125 and shingle F132 C135 Coralsand intrench Home Is. 2020 + 170 1570 + 175 and shingle 733 C136 ~—Coral shingle in trench Home Is. 1950+ 70 1500 + 70 Porites 7747 C108 Coral 20-30cm in pit _ West Is. 1890 + 60 1440 + 60 23 MALDIVES ,; SEYCHELLES CHRISTMAS Is. _ COCOS (KEELING) IS.” MADAGASCAR MAURITIUS , CHRISTMAS ISLAND ~ COCOS (KEELING) ISLANDS Figure 1. Location of the Cocos (Keeling) Islands and bathymetry (in metres) of the northeastern Indian Ocean. 96°50'E 96° 55'E 12°05'S /z=~Direction Is. NS Ate ——sZ 50'S ~ c NORTH \ 3290 KEELING INDIAN COCOS West Is. (KEELING) ISLANDS Coral Boulder s In Situ Microatoll a Vibrocore e 96° 50°E 96°55'E Figure 2. Cocos (Keeling) Islands, showing locations of stratigraphic transects I-X, vibrocores, and radiocarbon dates on coral from conglomerate platform (Table 2) and fossil microatolls (Table 4). Figure 3. + MSL 400 Sand metres Conglomerate platform Shingle (core) leat Large coral (core) 4 Cemented sand egee=25 Shingle cemented by calcareous algae Transect I, Home Island (see Fig. 2 for location): stratigraphy and radiocarbon dates. D5 26 CK2 3680+105 MSL 200 a Neal Floor of channel — —~ — My ky ‘ ky Hi ly lial | ! VI In = 1 | a Vy VI ce Hs i I! 5160-85 It Vl | = i md 3490=85 Mi i F uy : 11 = Lt =3)),22 ‘Ht nS E W | pang 4 5080:85 ~*“ Conglomerate platform Mi . x 4 ‘4s Shingle (core) De @® Large coral i Cemented sand gal -5 ar Shingle cemented by in calcareous algae ral 1 | 6160:95 mC Figure 4. Transect II, Home Island (see Fig. 2 for location): stratigraphy and radiocarbon dates. 27 W E ue CK6 Sand spit 2 MSL ae pl 100 & ale SSCE a eae Pa ot oe Al ) Sri —e? Floor of channel i i 3170=85 : | y i 4 tl | 1 lel \ ul mM } i fe G ee Li il L o a PI 3 fH ul E Cl ig i y lame A id «Sand NM < q ' ~~ Conglomerate platform 4 ul [ey I \~ Shingle (core) \ Hl @ Large coral ti im Cemented sand No recovery “tl p75 aa Shingle cemented by i calcareous algae al lll Reef flat i +-6 =e: Sand/mud/shingle ut Figure 5. Transect III, Pulu Wak Banka (see Fig. 2 for location): stratigraphy and radiocarbon dates. 1 2) (< © IA -3 5 E 6040-80—~Lay }-4 MA ry Kes I- Sand x ky lL xxx Conglomerate platform is Dy hee coro Shingle uf F L-8 4 Tridacna WJ 14 ee I|1I11 Reef flat cl ra “It Pleistocene limestone D L-10 K /) Figure 7. A S< Conglmerate platform 0 eee m1 #5 Sand and shingle it i Ill Reet flat Ms ay ais) | , Pleistocene limestone oe 6790+80 Six Sand and shingle (core) is I Py la Transect V, Pulu Blan Madar (see Fig. 2 for location): stratigraphy and radiocarbon dates. Reef crest metres 800 29 ss Sand e¢e Shingle s Coral lll Reef flat UV Shingle cemented by calcareous algae “~~~-/ Shingle (core) CK23 No recovery Figure 8. MSL Figure 9. — 5630-205 y H uN nN it 4740-85 i! iF | a 5800-70 Transect VI, Pulu Maria (see Fig. 2 for location): stratigraphy and radiocarbon dates. Bulldozed topography ss Conglomerate platform : @ Coral shingle -—= Sand and mud it -Il<~ Sand and shingle (core) Transect VIL, eastern end of West Island (see Fig. 2 for location): stratigraphy and radiocarbon dates. 30 3 S N CK9 2 ES g950:105— CiKig ? 23 ’ 1 ae | jo CK20 At 1 MSL 100 200 300 400 Ny 500 SS CK10B CK10A hy WI 600 K4 m1 ckig OK ! Ter A i 2 nl 1 A 3 ] iy iw a F i ¢ ; e ae 15550285 2 ( x Hey age =) ! = i a ares a Ly | ga Wl 4 kf g Sy hi ta S l-5920-95 EI e -3 (! i — 6140285 \! Pi Ss Conglomerate platform Gi a \ hy 4/> Shingle (core) al No recovery at | ra @ Coral ‘| y i I+ =f “2. Cemented sand 4 b ax Shingle cemented by in | calcareous algae eS re ry Al ae (II| Reef flat A 4 e.°. Shingle IS : a = Sand and shingle rm Sy U f === Mud ni “1 ~Pleistocene limestone re i : aT f : | i -10 i Figure 10. _—- Transect VIII, southern West Island (see Fig. 2 for location): stratigraphy and radiocarbon dates. Sit W ES 3 AorH n i CK8 ae Sand Microatolls CK8A <“ Conglomerate platform =~ Sand and mud \-7= Shingle (core) iil Reef flat 2 Road = Large coral Shingle (core) \ Wa le Figure 12. Transect X, Horsburgh Island (see Fig. 2 for location): stratigraphy and radiocarbon dates. 32 RADIOCARBON YEARS BP 8000 6000 4000 2000 0) MSL (W) Hld3ad / NOILVAA14 e Cocos, coral « Cocos, microatoll > 14 Figure 13. | Age-depth plot of radiocarbon dates from Cocos, showing dates from drillholes (Table 3), from conglomerate platform (Table 2) and from fossil microatolls (Table 4). Three phases can be recognised: 1) catch-up reef growth, 2) reef flat consolidation, and 3) reef island formation. See text for details. 33) INTERGLACIAL GLACIAL MODERN ATOLL FORM ne ees oe ATOLL Karst HOLOCENE REEF GROWTH AND LAGOONAL SEDIMENTATION PRESENT SEA LEVEL DEPTH (metres) 100 150 160 120 80 40 te) TIME (thousands of years before present) Figure 14. | A model of the late Quaternary development of the Cocos (Keeling) Islands. The sea-level curve is derived from Chappell and Shackleton (1986). The atoll is gradually subsiding. The interglacial atoll surface is subject to solutional weathering particularly when the sea is low. During the postglacial marine transgression reefs have re-established over the pre- Holocene surface and the three phases of Holocene atoll development identified in Figure 13 can be recognised. ATOLL RESEARCH BULLETIN NO. 403 CHAPTER 5 REEF ISLANDS OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE AND R.F. McLEAN ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 5 REEF ISLANDS OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE * AND R.F. McLEAN ** INTRODUCTION Reef islands have developed during the final phase of the Holocene development of the Cocos (Keeling) Islands. The islands are low-lying, and are generally composed of unconsolidated, biogenic sand. In places coral shingle is an important element of the sediments, and in a few localities, as for instance towards the southern end and on the oceanward shore of Pulu Wak Banka, coral rubble, composed of boulders of more than 50 cm diameter, covers the islands seaward margin. Reef islands often overlie a cemented coral breccia, referred to as conglomerate platform (see previous chapter). Guppy (1889) noted the compositional similarity between this and the material comprising the modern reef flat and reef crest, and as we have demonstrated this platform appears to be an emergent, fossil reef flat dating from the mid-Holocene (Woodroffe et al. 1990a, 1990b). The conglomerate platform underlies many of the islands, although is not found everywhere (Jacobson 1976); it appears to act as an anchor, determining island location. A clinker-like conglomerate, undercut by solution and ringing metallic to the hammer, is found along the lagoonward shore of many islands, and is particularly conspicuous around the perimeter of the lagoonlets or teloks. In addition to the conglomerate platform there are also cemented deposits of beachrock. Beachrock can be distinguished from conglomerate platform because it is bedded and exhibits a seaward dip (Russell and McIntire 1965, p35); beds are scarped at the landward face, and the deposits are undercut along bedding stratification (Stoddart 1971, p9). In places beachrock overlies conglomerate platform, as on the western shore of Horsburgh Island. In addition there are some isolated cemented deposits which resemble cay sandstone. These are generally less well lithified, and are horizontally bedded; an example occurs on the oceanward shore at the southern end of the airstrip on West Island. In this section the physical and major vegetation characteristics of individual reef islands are described and mapped. Mapping was undertaken from 1:10,000 colour vertical aerial photographs taken in 1987, supplemented by ground truthing. Elevational information was derived from a series of profiles surveyed across reef islands and related to mean sea level datum. Additional data come from existing surveys and benchmark records. * Department of Geography, University of Wollongong, Northfields Avenue, Wollongong, New South Wales, 2522. ** Department of Geography and Oceanography, Australian Defence Force Academy, Canberra, Australian Capital Territory, 2601. The general form of reef islands was summarised diagrammatically by Darwin, who recognised a ledge of conglomerate platform which protruded on the oceanward side of islands, and a ridge, generally sandy, but also of coarser material, which formed the oceanward beach. In common with reef islands on other atolls, Cocos reef islands often show a lower lagoonward beach ridge, in addition to the oceanward beach ridge. Individual reef islands were described by Guppy (1889). Details of perimeter and island area are given in Table 1; they are examined sequentially below. REEF ISLANDS DIRECTION ISLAND Direction Island (also known as Pulu Tikus, or Rat Island) is a crescent-shaped island. It appears to have been the first island to be inhabited, for it was on Direction Island that the crew, together with the rats, of the Mauritius were marooned after their ship was wrecked in 1825. The island was also inhabited for a short period by Alexander Hare and his followers in the same year. It is 1.6 km long and 300 m maximum width, with an area of 0.34 km2. It is dominated by coconut woodland, but with a band of Scaevola scrub along its eastern margin. This island was the site of the Cable Station, with undersea links to Australia, Singapore and South Africa, which came into operation in 1901 and ceased in 1966. In the late 1960s buildings on the island were either translocated to Home and West Island, or bulldozed into the sea along the oceanward margin, and considerable building rubble is conspicuous along the oceanward shore (Fig. 19a). The eastern portion of the oceanward shore comprises a prominent ledge of conglomerate platform, extending up to 35 m seaward from the beach (Fig. 14). This conglomerate platform is composed particularly of heads of branching Acropora. The oceanward beach ridge is composed of coral rubble and shingle along most of the island, and fragments of Pocillopora are especially frequent. This coarse substrate overlies sand, and the island can be seen to be composed of sand just over the ridge crest which reaches a height of around 3.0 m along much of the island, but 3.5 m towards the northern end (Fig. 11). The lagoonward shore is dominated by a broad sandy beach. At the northwestern end of Direction Island there are a series of shingle berms marking periodic accretion, in relation to which Guppy proposed that 'as the reef grows seaward the island also gains on the reef flat by a succession of ridges, thrown up during heavy gales, the remains of which are still to be seen in its interior' (Guppy 1889, p463). Small outcrops of beachrock are found at the northern and southern ends of the lagoonward beach, indicating minor recession of this shoreline at some stage in the past. There is a strong current running through the inter-island channel at the southern end of Direction Island, and called the Rip. Guppy (1889) proposed that such currents served to give the island its crescentic shape. WORKHOUSE ISLET There is presently no permanent land at the site of Workhouse Islet, or Pulu Pasir. However, there has been an island of variable size there in the past. A sand bank 150 ft by 100 ft was described by van der Jagt in 1829; Guppy (1889) records that in 1888 it was 6 ft high and dominated by one seaward leaning coconut. PRISON ISLAND The island north of Home Island is known as Prison Island; it has been known as Pulu Beras (Rice Island) or Pulu Tuan (Master's island), from the time of Alexander Hare. It is now considerably smaller (88x75 m) than it must have been when Alexander Hare moved his household there in 1827 from Home Island. van der Jagt (1831) recorded that it was 20 ft high in 1829, and Guppy (1889) also states that it was 20 ft high in 1888 and was composed of blown sand. It presently reaches a height of 6.7 m, and so has changed little in overall height in this time, though it is now eroding on all sides (Fig. 19b). It contains a mixture of coconut, Scaevola and Tournefortia. Bunce (1988) implies that much of this erosion has taken place in the 30 years since Pulu Gangsa has been connected to Home Island. BUTTON ISLETS Guppy (1889) records that there were a series of islands, termed the Button Islets, on the 1829 map of van der Jagt, between Prison Island and Pulu Gangsa. The sandy reef islands had already disappeared by 1888, with only conglomerate platform remaining (this platform can be seen in the foreground in Fig. 19b). HOME ISLAND Home Island has been a centre of habitation since Alexander Hare chose it for his first permanent settlement in 1826. The burial island, Pulu Gangsa, termed Clunie Island by Guppy (1889), was artificially joined to Home Island by placing coconut logs and concrete-filled drums across the channel in the late 1940s (Bunce 1988). In 1888 the channel between the two had been less than 2 ft deep at low tides (Guppy 1889, p464). The island is also known as Water Island, New Selima or Pulu Selma. It is covered by well-managed coconut woodland, with extensive groves of Calophyllum. Casuarina was reported as widespread on the island in 1888 by Guppy (1889). The combined islands have a length of 2.6 km, and reach a maximum width of 800 m (Fig. 2). Their area is 0.95 km?. Some part of this has been reclaimed from the sea; this is especially true of the landing area north and south of the present jetty, and the part of the village called kampong baru (new village), reclaimed by teams of women earlier this century (Bunce 1988). Oceania House was designed and built by George Clunies-Ross in 1893. There is considerable survey data available for Home Island. Most of the kampong is 1.20-1.60 m above mean sea level. The island rises generally to an oceanward beach crest that is around 3.30 m (Fig. 11). This beach is covered with coral shingle and coral boulders, but as excavations into the island have shown, these are underlain by sand which is dipping gradually oceanward at 2-6’. North of the village there is the remnant of a wind-blown dune (see Fig. 11); sand from similar dunes appears to have been removed and used to assist fill in reclaimed areas. The dune presently rises to 5.50 m above mean sea level and contains one of the few remaining stands of Pandanus. The sandy lagoonal shore has been extensively modified; sand has been bulldozed, and there is evidence of a series of seawalls along parts of the shore. The village extended along the southern shore, east of Oceania House earlier this century (Gibson-Hill 1950). On the oceanward shore of Home Island there is a narrow outcrop of conglomerate platform within which branching corals are especially prominent (Fig. 15). Individual Porites blocks within the conglomerate platform reach up to 1.0 m mean sea level. The conglomerate platform is overlain by boulder and shingle deposits. It widens into a broader platform at the southern end of the island. Conglomerate also underlies much of the island, as can be seen on the profiles in Figure 11, and from descriptions in Jacobson (1976). PULU AMPANG KECHIL Pulu Ampang Kechil is the small island south of Home Island, and forming the northern outlier of the Ampang Islands (Fig. 3). It was called Scaevola Islet by Guppy (1889). It is dominated by Scaevola, although with individual coconut and Tournefortia. The island is composed of coral shingle, with a sand spit extending to seaward and to lagoonward. It lies on an outcrop of conglomerate platform which contains some particularly large heads of Porites coral of over 1 m diameter. PULU AMPANG The Ampang Islands, termed Steward's group in van der Jagt's map of 1829, are a group of several small islands on one outcrop of conglomerate platform. The term Pulu Ampang is generally applied to the northernmost of the group, distinguished as Ampang Major by Guppy. This island is horseshoe shaped 625 m long and 275 m wide, with sandy spits extending into the lagoon around the island margins. Pemphis grows on these spits and also occurs along a minor bar cutting off the interior lagoonlet. This lagoonlet dries at low tide, and Guppy recounts that J.C. Ross remembers this feature silting up (Guppy 1899, p466). It appears to have changed little from the account given by Guppy over 100 years ago (Fig. 3) The conglomerate platform is extensive along the oceanward shore of each of the Ampang Islands (see Fig. 19c). It forms a much flatter surface than on Pulu Ampang Kechil, and rises up to heights on individual coral heads within the platform of 0.76 m mean sea level. Much of the conglomerate platform is inundated at high tide, particularly when there is a large swell. The oceanward beach on Pulu Ampang is composed of shingle overlying sand, with individual boulders at the foot of the beach of diameters up to 1 m. The crest of this ridge has a cover of Scaevola scrub, which is replaced 10-20 m inland by coconut woodland (Fig. 3). PULU WA-IDAS Pulu Wa-idas, called Ampang Minor by Guppy, is separated from Pulu Ampang to the north by a deep pool, which resembles other inter-island passages except that it does not continue through the conglomerate platform as a channel. The conglomerate platform is fissured, and is evidently eroding at this point, and given time it would appear that a channel will form between these two islands. The island is 75 m from north to south, and 320 m from oceanward to lagoonward; it is covered with coconut woodland, with a margin of Pemphis along its southern side (shown in Fig. 19c). PULU BLEKOK The southernmost of the Ampang islands is Pulu Blekok, called Pulu Bruko by Guppy. In form it is a mirror image of Pulu Wa-idas, with a fringe of Pemphis along its northern margin. It is 230 m from oceanward to lagoonward (Fig. 3). The conglomerate platform, although embayed on the oceanward side at this point, does not show the same indications that a channel will form as to the north of Pulu Wa-idas. Indeed Guppy suggested that the vegetation of the two islands was encroaching, and that over time the two islands would unite (Guppy 1889, p466); they have not done so in the 100 years since he observed them. There is a lagoonlet, largely cut off from the lagoon, on the lagoonward side. PULU KEMBANG Pulu Kembang, spelt Pulu Kumbang by Guppy, but not described in any detail, Sits on its own outcrop of conglomerate platform. The island is 150 m north to south and 390 m from oceanward to lagoonward; it is predominantly sandy, but it has extensive shingle along the margins flanking inter-island channels, and shingle is found at the beach top, where there is a narrow band of Scaevola. Much of the island is covered with coconut woodland, though Pemphis occurs on the lagoonward most parts of the sandy spits. The sand on the oceanward beach comprises an abundance of foraminifera; it appears to be actively accreting, particularly at the southeastern corner, where Pemphis is colonising this sand. PULU CHEPELOK/PULU WAK BANKA The island south of Pulu Kembang is a long island which has several names. It was called Armstrong Island on the 1829 map of van der Jagt; Guppy called it Gooseberry Island. On the recent maps its northern part is called Pulu Chepelok (also spelt Pulu Cepelok), while the southern half is named Pulu Wak Banka. The island is 1.15 km long, and up to 400 m wide. It has prominent spits at the northern and southern ends. There are also a couple of similar features extending into the lagoon in the middle of the island, giving the impression that this island may have comprised two or three islands in the past (Guppy 1899, p466). The island along its oceanward shore is underlain by an extensive conglomerate platform. This contains large boulders in places; there is a large boulder 1.5 m long and 1 m higher than the general conglomerate platform level, reaching an elevation of 1.50 m mean sea level, at the southern end of the island. The nature of the beach sediments on the oceanward shore changes markedly along the island. There are coarse coral rubble deposits, with boulders up to 1 m in diameter, along much of the southern half of the island, reaching up into the Scaevola scrub which is dense along the ridge crest (Fig. 20e). On the other hand, where the island is narrowest, the conglomerate platform is no longer present along the oceanward shore, and instead there is a broad sand beach. The ridge crest rises to 3.5 m mean seal level at the southern end of the island, but is only 2.1 m mean seal level where a profile has been surveyed across the island in the centre. On that profile (Fig. 11), it can be seen that the coral shingle overlies sand, and that the conglomerate platform also continues under much of the island. Conglomerate forms a thin crust along the margin of the channel along the southern end of the island. PULU PANDAN Pandan Island or Pulu Pandan (also called Misery Island) is the clearest example of a horseshoe shaped island (Fig. 3), called an atollon by Guppy (1889). Despite its name, Pandanus is no longer conspicuous element of the island's vegetation. It consists of two distinct sandy spits with a shingle veneer, both covered by coconut woodland with Pemphis on the lagoonward extremity (Fig. 16). Maximum width is about 800 m, and the island measures 1.8 km from the end of one spit to the end of the other spit. These spits serve to partially enclose a lagoonlet, with a soft muddy floor and cover of seagrass. The southern spit in particular has recently extended into the lagoon, and there is a further shoal of intertidal sand, with an outpost of Pemphis on it which represents a continuation of the island. On the oceanward shore there is a margin of conglomerate platform extending along the island. For much of its extent this conglomerate comprises an upper unit of shingle-sized clasts cemented into a near horizontal layer (Fig. 19d); this overlies some in situ fossil microatolls of branching Porites at the eastern end of the surveyed transect. The conglomerate platform surface rises to 1.20 m mean sea level, rather higher than on other islands on the eastern rim of the atoll, suggesting that this shingle conglomerate surface may overlie the more conventional conglomerate platform surface. The transect (Fig. 11) illustrates that the island is composed primarily of sand, rising up to a ridge crest of 4.50 m mean sea level, and does not have the shingle or rubble veneer characteristic of the oceanward shore of islands to the north. Shingle does form low elevation ridges along the lagoonward shore, and there are small outcrops of a clinker- like conglomerate around the margin of the lagoonlet. PULU SIPUT Pulu Siput (also called Goat Island) is dominated by coconut woodland, and is 660 m oceanward to lagoonward, and 240 m from north to south (Fig. 3). It has formed on an outcrop of conglomerate platform, and the island is predominantly sandy, with foraminiferal sand accumulating at the northeastern corner of the island. Spits, with a cover of Pemphis, but also with occasional Suriana, extend into the lagoon. PULU JAMBATAN Pulu Jambatan is the name given to the island formed largely of coconut woodland, 340 m long, at the lagoonward end of a long, narrow outcrop of conglomerate platform. There is a much smaller island, less than 50 m long, at the seaward end of this conglomerate platform outcrop, apparently without a name, with a vegetation cover of Scaevola with some coconut and Tournefortia (Fig. 3). The conglomerate platform between these two islands is smooth, and cover with a veneer of pink algal mat. Seaward of the more oceanward island, the platform is highly irregular, and contains much coarser coral boulders. PULU LABU Pulu Labu is the island directly north of South Island, separated from it by a narrow channel (Fig. 3). Most of the island, which is 430 m long, is composed of coconut woodland; there is a broad band of Scaevola scrub along the oceanward ridge crest, and on the southeastern corner where sand has recently accumulated, Pemphis is established. The form of the island is very similar to that of a series of lagoonward promontories on neighbouring South Island, and it is not unlikely that similar islands to Pulu Labu may have existed in the past, but have now been united with South Island. SOUTH ISLAND South Island, also called Pulu Atas (meaning top island in reference to it being upwind), Scott Island and Southeast Island, is the windward island of the atoll. It was chosen as the site for the first settlement by Captain John Clunies Ross in 1827, who dredged a boat channel through the southern lagoon to the centre of the island. The long lagoonal shore is the preferred site for a number of Home Islanders pondoks (weekender shacks), and was also home to a regiment of Kenyan soldiers, the Fifth African Rifles, who were stationed at the southwestern end near the highest point termed “Gunong’, as coastwatchers in World War II. The island is 9.5 km long, and reaches a maximum width of 1.1 km and is mapped in Figures 5 and 6. Much of the oceanward shore of South Island is formed of a dune (see Fig. 19f). Windblown sand reaches up to 6.3 m on profile I and profile J (Fig. 11). A dune reaches up to 11 m at the 'Gunong’ at the southwestern corner of the island. A coral rubble veneer reaches 4.7 m mean sea level on profile H. The vegetation of the dunes is primarily Scaevola, though with considerable Tournefortia, particularly as isolated shrubs within blowouts along the dune crest. Guppy (1889) recorded that Pandanus was found along this dune crest, but it is not a conspicuous element of the vegetation now. While dunes, which are rare on coral atolls, characterise much of the shoreline of South Island, there is also a substantial outcrop of conglomerate. This takes two forms; conglomerate platform occurs in irregular outcrops along much of the eastern part of the island, often rising up to 1.20 m mean seal level. There are also outcrops of conglomerate ramp, a highly worn form of conglomerate platform, which has been bevelled back to a steep ramp-like profile (Fig. 19e). The latter superficially resembles beachrock, which can also be found at sites along the oceanward shore of South Island, but is not imbricated, and on inspection can be seen to have been bevelled to form the dipping outcrop, rather than deposited in dipping stratification. Similar conglomerate ramps are described on Diego Garcia, an atoll in the Chagos group (Stoddart 1971, p18). The interior of the island is now covered by thick, overgrown coconut woodland which has degenerated from the organised and harvested coconut plantations of the heyday of the Clunies Ross estate. On the oceanward shore and over the narrow necks of the island, there is dense, impenetrable Scaevola scrub. Little remains, except isolated stumps of the Pisonia and Cordia stands which were once widespread on the island. There is a large stand of Calophyllum at the southwestern corner of the island (Fig. 5). The lagoonward shore of South Island is highly irregular. The lagoonal flats are composed of mud or sandy mud, and there are irregular linear shoals, covered by Pemphis and inundated at high tide, partially enclosing some of the larger lagoonlets, termed Teloks (Fig. 20c). It is said to have been silting rapidly, which may have lead to the abandonment of the first settlement there; though there can be little doubt that the southern flats of the lagoon must have been shallow even at that time, and access cannot have been easy. The western end of the island has a series of recurved spits; these are not as distinct as those of West Island; nevertheless they were Lae by Guppy (1889) to indicate that the island had been extending to the west. Upon first impression this elongate island appears to have been made up from several islands which have been joined together. There are two areas, traversed by profiles H and I respectively, which resemble infilled passages between these former islands. These are covered mainly by Scaevola scrub, with few coconuts; those coconuts which do grow there are stunted, and stressed. There is no freshwater lens developed beneath these narrow areas. Soil is absent or poorly developed, and the lagoonward portion of the island is composed of clinker coral shingle. Darwin interpreted these as former channels, and his interpretation was aided by a map that Leisk, the manager in charge of the islands at the time of his visit, told Darwin he had seen. Guppy was rather dismissive of the likelihood that the channels had been infilled as recently as Darwin implied, pointing out that they were closed, and the island one entity even in the map shown in van Keulen's Atlas of 1753 (Guppy 1889, p467). We examine this issue in more detail below. PULU KLAPA SATU Pulu Klapa Satu, the island directly west of South Island, is about 125 m long and 75 m wide (Fig. 5). It sits on a long linear exposure of conglomerate platform, which in common with the other islands of the southern passage, is relatively free of large coral clasts, and contains largely sand-sized grains cemented together. In petrography it resembles beachrock, but lacks the stratification which distinguishes the latter. PULU BLAN AND PULU BLAN MADAR Pulu Blan and Pulu Blan Madar, also known as Burial Island and East Cay, sit on the same outcrop of fine-grained conglomerate platform. They are composed of sand with some shingle, and carry a vegetation of coconut and Scaevola. The oceanward shore of Pulu Blan Madar rises up to a height of 1.20 m mean sea level. PULU MARIA Pulu Maraya or Pulu Maria lies on an outcrop of fine-grained conglomerate platform just west of the eastern end of West Island (Fig 8), and is named after one of two European children who disappeared without trace from the island shores in the 1860s. The island is predominantly sandy though with a series of shingle berms on the oceanward shore. It is dominated by coconut woodland, with a fringe of Scaevola, replaced with Pemphis along the lagoonward flanks. WEST ISLAND West Island, also known as Ross Island, or Pulu Panjang (Long Island), is the island upon which the airstrip was built, initially in 1944, but seeing little action in the war, and revamped for use by Qantas in 1951. It was first settled in 1826 by some of Alexander Hare's followers, probably in the vicinity of Rumah Baru, and has been inhabited discontinuously since. It was home to more than 7000 troops from Britain, Canada, Australia and India in 1944, and has been associated with the airstrip and contains an Australian expatriate population at present. The island is 12.6 km long and reaches up to just less than 1 km wide at its maximum width. It is mapped in Figures 6, 7 and 8. Most of the 6.2 km2 was covered by coconut plantation, but much is now covered by buildings, the airstrip, or radio transmitter and receiver aerials. The coconut woodland has ceased to be cleared regularly, and has become largely overgrown, and penetrable with difficulty. The island comprises three broad sections, connected by narrow sections which may have been former inter-island passages. These lead into the two large lagoonlet areas, Telok Jembu (Fig. 20a) and Telok Kambing (Fig. 20b). Much of the western shore is a sandy beach, with a dune, reaching more than 7 m high, at Beacon Heights, which has been excavated. Groynes have been constructed in front of the settlement to stop northwards movement of sediment, but accumulation within them indicates little net movement. There are extensive outcrops of beachrock, particularly at the southern end of the island, and adjacent to the Quarantine station, at those sites which appear to mark former passages. There is a large area of conglomerate platform at the southwestern end of the island, and isolated outcrops at the westernmost point and to the northwest. The outcrop to the southwest is one of the more elevated outcrops on the atoll rising up to 1.20 m mean sea level, with a further cemented shingle conglomerate up to 1.80 m mean sea level outcropping on the beach behind the conglomerate platform. There are a number of dipping arcuate ridges within this platform, especially at the southwestern corner, resembling the bassett edges recorded on the Great Barrier reef islands (Stoddart et al. 1978). The easternmost end of the island is characterised by a number of sand spits and ridges, suggesting gradual buildout of the island into the southern passage. Radiocarbon dating of coral shingle from shallow pits in those shown in Figure 9, in the previous chapter, indicates that these spits have been built progressively. The ages are shown on an aerial photograph of the spits in Figure 17. HORSBURGH ISLAND Named after James Horsburgh, the British hydrographer, who compiled detailed sailing directions of this part of the Indian Ocean in 1805, Horsburgh Island is also known as Pulu Luar (Outside island). It is 1.7 km long and 0.9 km wide, covering an area of over 1 km? (Figs. 9 and 18). It was almost continuously inhabited from 1826 until after World War II. Initially Alexander Hare put people on the island to grow vegetables and fruit for other islands. This tradition was maintained by the Clunies Ross proprietors, and George Clunies Ross kept deer on the island for hunting. In 1941, gun emplacements were installed on the southern point of the island and manned by Ceylonese troops. This island sits partly on an outcrop of conglomerate platform. The conglomerate differs from that on other islands; on the eastern shore of Horsburgh it is generally narrow, and often bevelled into a conglomerate ramp. On the western shore there is a broad platform which consists of a series of strata dipping seaward at up to 5°, which 10 resemble beachrock. The platform appears to combine conglomerate platform and beachrock, and suggests that islands here may have formed almost contemporaneously with the development of the emergent reef flat. Along the southern shore there is a broad sandy beach, in places with outcrops of beachrock which indicate that in the past the shoreline has had a slightly different orientation in this part of the island. The northern shore of the island is particularly exposed and consists of a bevelled conglomerate platform ledge, and boulder deposits over the top. A particularly noteworthy feature of this island is the small lagoonlet which occurs within the interior of the island to the northeast. This feature, blocked of from the sea by a boulder rampart, presently contains brackish water, and a stand of mangrove Rhizophora apiculata. Associated with the mangroves are Cordia stumps and Sesuvium. In his account of Horsburgh in 1888, Guppy (1889) describes the inland lakelet, but does not record mangroves growing there naturally; indeed he makes the point that mangrove propagules are regularly brought to the shores of Cocos, but have not colonised (Guppy, 1890). Guppy indicates that mangroves were planted there by John George Cluines Ross (Guppy 1890, p278). In a photograph of the lakelet, taken in 1941, the mangroves can be still seen (Gibson-Hill, 1950). Much of the northern part of Horsburgh is composed of shingle or rubble, while the southern part is predominantly sand. Coconut scrub is especially open over the southern part of the island with a sward of grass and the sedge Fimbristylis, but forms denser coconut woodland to the north. Scaevola scrub is widespread over the island; to the south it is relatively open, but to the north it is dense, and made almost impenetrable by a tangle of Turnera, Triumfetta, Wedelia, Premna and the parasitic Cassytha. NORTH KEELING ISLAND North Keeling Island is named after Captain William Keeling who is believed to have sighted the island in 1609. It was sketched, showing coconuts, by the Swedish captain Ekeberg in 1749 and appears on the chart reproduced by Dalrymple the English hydrographer in 1787. Fitzroy examined and mapped it from H.M.S. Beagle in 1836, but made no landing. Unlike the South Keeling Islands, North Keeling has been visited relatively infrequently by naturalists, and therefore does not have the same history of description. It was first described in detail by Guppy (1889) who was there for 6 days in 1888. Wood-Jones (1912) spent a few hours ashore in June 1906, and the most detailed account, especially of the fauna is that of Gibson-Hill (1948, 1950) who visited for 1 day in January and 2 days in early July in 1941. The island has not been inhabited for any continuous period, and is presently relatively little changed in comparison with the South Keeling Islands. It was visited from Cocos by the Clunies-Ross family, and Home Islanders (up to 40-60) stayed there for up to three months over the November-February period cutting firewood. The Emden beached on the southern shore of North Keeling after being routed by the Sydney in 1914; and the longest period of settlement was probably during the salvage of the Emden October 1915 to January 1916. The island is 2.0 km long and 1.3 km wide, with a reef crest around all of the island, except the northwestern corner (Fig. 10). Reef island is almost continuous around the perimeter of a shallow lagoon, reaching a maximum width of 320 m and a minimum width of 50 m. There is one major opening into the lagoon on the southeastern corner of the atollon. This is the windward side, and the opening has no channel through the reef, 11 but is a shallow conduit which drains almost totally at lowest tide. The lagoon is shallow, reported as nowhere deeper than 8 feet by Guppy. It's surface sediments are muddy sands, except for two sandy spits which trail in through the entrance. These did not appear on Fitzroy's chart of the island; Guppy added them in his sketch of the island, but shows them scrolled back on themselves. As can be seen in Figure 10 they are presently linear features which extend flanking the channel. Much of the lagoon is covered with sea grass. The island varies from sand to rubble. On the northern shore there is a broad sandy beach. This continues along the western shore but with varying amounts of shingle. On the profile (Fig. 10) the sandy beach rises up about 4 m above mean sea level. A pit shows some shingle fragments, but indicates that the majority of the substrate is sand. This becomes coarser to the south, where rubble outcrops on the beach, and there is an erosional cliff cut into this rubble. The southern shore of the island is composed of a spectacular steep shingle beach, with a series of berms identifiable. Much of the eastern shore is composed of a series of shingle berms, these are particularly well- developed just south of the channel into the lagoon, but continue to the north as well. Guppy (1889) recorded that pumice from the eruption of Krakatoa had advanced the shore into the lagoon; no evidence of this can be seen today. There are also outcrops of coral conglomerate. A broad platform of conglomerate extends out over the reef flat at the eastern part of the island, almost closing the channel into the lagoon completely. Along much of the southern and eastern shore the conglomerate outcrops at the foot of the beach but contains a series of parallel rubble ridges, dipping and stratified like beachrock. These appear to be the lines described by Guppy (1889) as old reef margins, and upon which he based his argument that the reef built out by a series of jumps rather than prograding gradually. Similar boulder conglomerates have been described from other reef settings; they closely resemble the adjacent beach in structure and composition and we call them beach conglomerate, believing that they mark the position of former beach lines rather than reef crests (Fig. 20f). At the site of the southern transect (Fig. 10) there are a number of algal terracettes veneering these old beach deposits. Beach conglomerate overlies conglomerate platform in some places (Fig. 10). The vegetation of the island was conveniently divided into four zones by Gibson- Hill (1950). Much of the island is dominated by Pisonia forest (see Williams, this volume, Chapter 6). Coconuts are a conspicuous element of all stands of Pisonia, and over much of the island we have chosen to map this as Pisonia and coconut woodland. Tournefortia is a conspicuous element of the vegetation of the eastern shore, dominating the crest of the shingle or rubble ridges. In some cases Tournefortia is monospecific, north of the channel into the lagoon it occurs with Scaevola also. Around the margins of the lagoon, Pemphis forms a thicket. Cordia is also important in this location, and it was to cut this latter species that the Clunies-Ross sent workers. It may have been less important when Gibson-Hill visited because of this history of cutting. Where Cordia forms a lagoonal fringe at present it is often fairly even-aged, and much may have grown back since cutting ceased. The final area that Gibson-Hill identified are cleared areas; the grassy and Sesuvium covered area to the northwest of the lagoon is the most extensive area of this type. There has been considerable speculation as to how North Keeling has developed. In particular it seems unusual because the remaining entrance to the lagoon occurs on the most windward side, rather in the shelter that might develop on the leeward. Indeed the island is quite the inverse of the horseshoe shape that Guppy considers the typical style of 12 development on the main atoll. This has lead a number of observers, starting with Fitzroy, to suggest that the island developed from a series of formerly unconnected islands. REEF ISLAND MORPHOLOGY The surveyed traverses (Fig. 11 and 12) show three basic cross-island profiles, the simplest of which was first described by Darwin (1842) and illustrated by a woodcut in the chapter on Keeling atoll (this illustration is in fact a section across Whitsunday atoll and not Cocos (Keeling)). Darwin notes that the highest part of the islets is close to the outer beach and that “from the outer beach the surface slopes gently to the shores of the lagoon”. Such simple asymmetric profiles are common on West and South Islands and across the centre of the small horseshoe shaped islands on the atoll’s eastern side. The second type of profile is basin shaped, again with a prominent seaward ridge which slopes inland to a central depression before rising to a lower lagoonward ridge. Such profiles are illustrated from Direction Island and the southern end of Home Island (Fig. 11). The third profile type is more complex being composed of a series of subdued ridges and swales between the ocean and lagoonward ridges. This form suggests a more complicated accretionary history. A characteristic feature of the islands on Cocos is their plan shape, which Guppy (1889) described as semi-crescentric or horseshoe shaped with their convexities to seaward. “The crescentric form is possessed in various degrees by different islands; some of the smaller ones are perfect horse-shoe atollons and enclose a shallow lagoonlet, others again exhibit only a semi-crescentic form, whilst the larger islands have been produced by the union of several islands of this shape.”” Examples of the first type would include Pulu Ampang, of the second type Direction Island and of the third type South Island. To Guppy the islands fitted into an evolutionary sequence all stages of which are represented on Cocos “from the islet recently thrown up on the reef to the perfect horse- shoe atollon”. Critical in Guppy’s interpretation are the lagoonward recurving extremities of the islands which he believed were formed from material brought in by uni-directional currents through the interisland passages and “heaped up in such a manner as to prolong the extremities of each island lagoonward in the form of two horns”. In the case of the larger islands a crescentric form results, while for the smaller islands a more perfect horseshoe shape is first attained. After the two horns are stabilised by vegetation, and providing there is an adequate supply of sand, the horns would tend to approach each other and ultimately they would be joined by a bar enclosing a lagoonlet on the island’s lagoon side, Guppy called this occluded island form an atollon, and noted that Horsburgh Island “represents the last condition of an atollon, the earlier stages being illustrated by Pandan Island and Pulu Ampang Major”. This view of island evolution differed from that of Darwin who envisaged a difference in formative processes between islands on the windward and leeward sides of the atoll (Darwin 1842). On the windward side, the islands “increase solely by the addition of fragments on their outer side”. Thus the gently sloping surface on the western side of the windward island predates the high ridge to seaward, and is lower because waves had further to go from the reef edge and “had less power to throw up fragments”. On the leeward islands, Darwin recognised a combination of two processes operating. First, waves from seaward formed the high ocean ridge, and second “‘little 13 waves of the lagoon, heap up sand and fragments of thinly branched corals on the inner side of the islets on the leeward side of the atoll”. As a result “these islets are broader than those to windward, some being even eight hundred yards in width, but the land thus added is very low”. Both Darwin and Guppy, as well as subsequent workers; recognised the association of islands with the conglomerate platform and the fact that the unconsolidated sands and gravels which go to make up the island commonly rest on a solid foundation of conglomerate platform. Indeed Guppy went so far as to suggest that where bare level patches of conglomerate are exposed on the windward side of the atoll these were “the foundation of the islets that have long since swept away” (Guppy 1889, p462). In our view, the evidence for such an assertion is generally lacking, except in those places where linear or arcuate bands of beachrock or beach conglomerate are firmly cemented onto the conglomerate platform. Examples of such exposures are found along the northwestern side of Horsburgh Island, adjacent to the Quarantine Station on West Island, around the southwestern corners of West and South Island and on North Keeling Island. We believe that these outcrops are residuals from the earliest phase of island building and represent shorelines developed concurrently with the formation of the conglomerate platform at the time of higher sea level. A radiocarbon date of 3030 + 85 years B.P. from a coral boulder in beach conglomerate to the southwest of North Keeling gives some support to this argument. Landward erosion of these shorelines has subsequently occurred. In some other locations high beachrock or beach conglomerate is found congruent with the present shoreline. In such cases the position of the initial shoreline has been maintained. While the association of islands and conglomerate platform is the norm, Guppy (1889) also recognised that conglomerate platform is not everywhere present beneath the islands being “absent in those situations where ancient passages have been filled up with sand and reef debris, and also in those places where recent additions have been made to the land surface” (p. 462). Our drilling and field data confirm the validity of this comment, particularly with respect to the “horns” and “bars” of the horseshoe islands and atollons, as well as the extensive area of lagoonward recurving spits at the western end of South Island and eastern end of West Island. Radiocarbon dates recording the development of the last area are shown in Figure 17. REEF ISLAND FORMATION Reef islands post-date the conglomerate platform, and it has been demonstrated that the conglomerate platform was deposited 4000-3000 years ago, as shown by the narrow range of radiocarbon ages from within it (Woodroffe et al. 1990a, 1990b, this volume; see chapter 4, Fig. 2). The platform has been interpreted as a former reef flat, deposited under a sea level around 1 m above present, and the islands have formed in the last 3000 years during the time that the sea has fallen to present level. Some indication of island age has already been given for Home Island (Woodroffe et al. this volume, see last chapter Fig. 3). Samples of coral shingle from a trench through island sediments (shown in Fig. 20d), indicate an age range of 1400-1800 years B.P. Nevertheless there remains a series of different possible models of island formation, both in terms of oceanward or lagoonward accretion, and in relation to the 14 gradual or episodic nature of deposition of sediment. In order to examine the chronology of island formation in greater detail, three transects of pits were examined on West Island (Fig. 13) and samples of coral shingle submitted for radiocarbon dating. The radiocarbon dates shown in Figure 13 confirm that the islands contain few sediments greater than 3000 years old. The date of 4280 + 70 on transect O (T2) came from a depth around mean sea level which would be within conglomerate platform elsewhere. There is no lithified platform at this site, but the date appears to indicate a similar chronology of deposition. Although this transect is across a narrow neck of island flanking a telok (see Fig. 12), termed a barachois in relation to the atoll of Diego Garcia, a date of 3030 + 70 elsewhere on the transect indicates island formation at an early stage at this site. Transect P (T1) has been date in some detail. The oldest date 2710 + 90 years B.P. comes from pit 7 to lagoonward. There is then a progressive decrease in age towards the ocean. Thus contrary to Darwin's expectation, the island appears to have built out towards the ocean even here on the leeward side of the atoll. Dates from pit 3 are stratigraphically consistent and indicate rapid vertical build up. The dominant mode of accretion is horizontal. A similar trend of older dates to lagoonward, and younger ages to oceanward is seen for transect L (T3), which also ranges from 3000 years B.P. to present. This is particularly significant because this eastern part of West Island has been extending further eastward over the last 1500 years (see Woodroffe et al., this volume, Fig. 9). Radiocarbon ages on individual spits are shown in Figure 17. The main part of this southern section is evidently 3000 years old, like the northern section of West Island. North Keeling is morphologically distinct from the South Keeling Islands and may have developed differently. It is not unusual in other Pacific and Indian Ocean atoll archipelagoes for the smaller reef platforms to be occupied by one island which is low in the middle, with a lagoon that may or may not be connected to the open ocean. The history of development of these is not known in detail, although there are some radiocarbon dates available from table reefs, or reef-top islands, of this type in Tuvalu (McLean and Hosking 1991). We have three further radiocarbon dates from North Keeling. A coral from conglomerate on the northeast of the island dated 3840 + 85 years B.P., similar to but at the older end of the range of dates for conglomerate platform from the South Keeling Islands. An age of 3060 + 60 years B.P. was derived for coral shingle in a pit in the centre of the island, suggesting little time difference between the formation of the beach at the margin of the reef platform, and the formation of the island. The final date was on a boulder exposed within an erosional scarp in the rubble beach on the southwest of the island, which gave an age of 1620 + 80 years B.P. Guppy (1889) suggested that the boulders on this beach indicated that it was prograded by coral blocks piled up during a cyclone; this age implies that cyclones may have occurred over the last 1500 years or more. We note that this equates with a phase of island building on other parts of the Cocos (Keeling) Islands. The radiocarbon ages suggest continual addition to islands over the last 3000 years, but we have insufficient dates to indicate whether this accretion was gradual or whether it occurred in a series of episodes. At this stage we have no dates which allow us to address the morphological issues raised by Guppy. Nevertheless reef islands are geologically young and morphologically dynamic; sediment is continuing to be produced 15 and supplied to islands and the islands are continuing to change through the addition of sediment at some points, but its erosion from elsewhere. REFERENCES Bunce, P. 1988. The Cocos (Keeling) Islands: Australian Atolls in the Indian Ocean. Milton: Jacaranda Press. Gibson-Hill, C. A. 1948. The island of North Keeling. J. Malay. Br. Roy. Asiat. Soc. 21: 68-103. Gibson-Hill, C. A. 1950. A note on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 11-28. Guppy, H. B. 1889. The Cocos-Keeling Islands. Scott. Geog. Mag. 5: 281-297, 457- 474, 569-588. Guppy, H. B. 1890. The dispersal of plants as illustrated by the flora of the Keeling or Cocos Islands. J. Trans. Vic. Inst. London, 24: 267-306. Jacobson, G. 1976. The freshwater lens on Home Island in the Cocos (Keeling) Islands. BMR, J. Aust. Geol. Geophys. 1/4: 335-343. McLean, R. F., and Hosking, P. L. 1991. Geomorphology of reef islands and atoll motu in Tuvalu. South Pacific J. Nat. Sci. 11: 167-189. Russell, R. J., and McIntire, W. G. 1965. Southern Hemisphere beach rock. Geog. Rev. 55: 17-45. Stoddart, D. R. 1971. Geomorphology of Diego Garcia Atoll. Atoll Res. Bull. 149: 7- 26. Stoddart, D. R., McLean, R. F., and Hopley, D. 1978. Geomorphology of reef islands, northern Great Barrier Reef. Phil. Trans. Roy. Soc. Lond. B 284: 39-61. Van der Jagt, H. 1831. Beschrijving der Kokos-of Keeling-Eilanden. Verh. Batav. Gen. v. Kunsten en Wetenschappen (Batavia), 13: 293-322. translated in J. Malay. Br. Roy. Asiat. Soc. (1952), 25: 148-159. Wood-Jones, F. 1912. Coral and Atolls: a history and description of the Keeling-Cocos Islands, with an account of their fauna and flora, and a discussion of the method of development and transformation of coral structures in general. London: Lovell Reeve and Co. Woodroffe, C. D., McLean, R. F., Polach, H., and Wallensky, E. 1990a. Sea level and coral atolls: Late Holocene emergence in the Indian Ocean. Geology 18: 62-66. Woodroffe, C. D., McLean, R. F., and Wallensky, E. 1990b. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. Nat. Geog. Res. 6: 262-275. 16 Table 1. Perimeter and area of the Cocos reef islands Island Perimeter (km) ‘Area (km2) Horsburgh Island 4.4 1.04 Direction Island 3.4 0.34 Prison Island 0.4 0.02 Home Island 6.7 0.95 Pulu Ampang 1.8 0.06 Pulu Wa-idas 0.7 0.02 Pulu Blekok 1.1 0.03 Pulu Kembang 1.6 0.04 Pulu Wak Banka 2.4 0.22 Pulu Pandan 3:9 0.24 Pulu Siput 2D 0.10 Pulu Labu 1.3 0.04 South Island 28.5 3.63 Pulu Klapa Satu 0.5 0.02 Pulu Blan Madar 0.7 0.03 Pulu Blan 0.8 0.03 Pulu Maria 0.7 0.01 West Island 38.5 6.23 Table 2. Radiocarbon dating results on reef island sediments, West Island and North Keeling Island. Beta Sample Island Depth of Material Conventional No. No. sample radiocarbon age (cm) 59845 NKI-60 North Keeling 60 Coral shingle 3060 + 60 59846 WI-T1-P295 West Island 95 Coral shingle 570 + 60 59847 WI-TI P3 85 West Island 85 Coral Shingle 1990 + 70 59848 WI-T1 P3120 West Island 120 Coral Shingle 2010 + 60 59849 WI-T1 P3160 West Island 160 Coral Shingle 2110+ 60 59850 WI-T1 P4140 West Island 140 Coral Shingle 2130 + 60 59851 WI-T1 P7 75 West Island WS Coral Shingle 2710 + 90 59852 WI-T1P260 West Island 60 Coral Shingle 3030 + 70 59853 WI-T1 P4140 West Island 140 Coral Shingle 4280 + 70 59854 WI-T1 P1200 West Island 200 Coral Shingle 420 + 50 59855. WI-T3P270 West Island 70 Coral Shingle 1970 + 70 59856 WI-T3P455 West Island 55 Coral Shingle 3100 + 70 Note: Radiocarbon ages determined by Beta Analytic have not been corrected for 0 C}3 or for environmental reservoir effect. These corrections are of similar magnitude (c 400 years), but cancel each other out. Thus these Beta dates are more-or-less comparable to the environmentally-corrected ages given in the previous chapter. 17 DIRECTION ISLAND WORKHOUSE | Conglomerate platform PRISON Beachrock = Coconut woodland Scaevola scrub 500 metres Figure 1. Direction Island, mapped from 1987 aerial photography. 18 Open coconut woodland Coconut woodland Calophyllum woodland Pandanus Sand Shingle Rubble Beachrock Artificial marine structures Conglomerate platform SOUTH KEELING ISLANDS 500 metres Figure 2. Home Island mapped from 1987 aerial photography. ee al PULU AMPANG KECHIL PULU CHEPELOK -«: ye ae 19 SOUTH KEELING ISLANDS Rubble Sand and shingle Intertidal sand Conglomerate platform Coconut woodland Scaevola scrub Scaevola and Tournefortia scrub Pemphis scrub PULU ‘SIPUT-. Cordia stumps Barringtonia 500 metres Figure 3. mapped from 1987 aerial photography. Islands of the eastern rim of the atoll from Pulu Ampang to Pulu Labu, 20 | ; : SOUTH ISLAND | Sand Intertidal sand :*:) Sand and shingle “SK | Beachrock ++ Conglomerate ramp Conglomerate platform Coconut woodland | Low coconut woodland hSsd ee ne Scaevola scrub Lagoon Pemphis scrub 0 500 metres a aaa refi] Figure 4. South Island, northern section, mapped from 1987 aerial photography. 21 -Kydes3ojoyd [esse 186 qnios ejOneeos ISS qnios siydwad LZ puejpoom wnj|Aydojedg pue jnuod0D nage pue}poom wnjijAydojeo ae PUR|POOM yNUudd0d MOF = pue|poom jnuos0D = aNv1sI HLNOS | wos poddew ‘uonoes wioyjnos “purys] YINoS ¢ oInsIy (ia sehen Geel Ge ee wuojjejd 9a}ye1aw0j;bu0D se a cue i yoouyoeag |Z ajbulys pue pues |.~. pues jepeyul |. | | ce | pues : NS SQNVISI ng = po e ONIN334 ; ei ee ee — vA \ ) \ sae Tae 4s S y, K xs F . Ss ‘i ‘i A sa}jaoeuel -\ «— jebiyv | sea ‘Fe a OF saujaw 00S TP SOUTH KEELING . metres as Sand metres D cS ley S 0) ———— [| Beachrock metres Conglomerate platform 4 F [ mm | Reef flat metres msl if eee je eee Le = Se = aye 3 Bs 200 400 600 800 1000 metres metres =e = = 600 800 metres Figure 12. | Surveyed sections across West and Horsburgh Islands. 29 Sand xs Conglomerate platform 30% Shingle | ; 44 280=70 200 300 400 500 E P3 metres ty "Ny i] 1990-704 ea) aod Il ty r 2010-608) te i t ie HI 2130-60 mad 2110-60 b! ih 4 + aaa + + 300 400 500 600 700 800 900 1000 Figure 13. Cross-section and pits from three transects on West Island (see Fig. 12 for locations), showing radiocarbon dates (see Table 2). Aerial photograph of Direction Island, 1987 (reproduced by permission of the General Manager, Australian Surveying and Land Information Group, Department of Administrative Services, Canberra). Figure 15. 3) Aerial photograph of Pulu Ampang and neighbouring islands, 1987 (reproduced by permission of the General Manager, Australian Surveying — and Land Information Group, Department of Administrative Services, Canberra). 32 Figure 16. = jim i 4 Aerial photograph of Pulu Pandan and neighbouring islands, 1987 (reproduced by permission of the General Manager, Australian Surveying and Land Information Group, Department of Administrative Services, Canberra). 33 Figure 17. _ Aerial photograph of eastern end of West Island, 1987. Radiocarbon dates on coral shingle indicate the progressive buildout of the spits (reproduced by permission of the General Manager, Australian Surveying and Land Information Group, Department of Administrative Services, Canberra). 34 Pe. 2 Figure 18. Aerial photograph of Horsburgh Island, 1987 (reproduced by permission of the General Manager, Australian Surveying and Land Information Group, Department of Administrative Services, Canberra). oo Figure 19. a: Oceanward shore of Direction Island; rubble is from ruins of Cable Station, b: View looking North from Home Island. Conglomerate platform in middle distance is where Button Islets were, Prison Island is in the middle of the photograph and Direction Island in the distance, c: Conglomerate platform on Ampang Island, d: conglomerate platform on Pulu Pandan; it appears to consist of a shingle conglomerate layer overlying typical conglomerate platform, e: Conglomerate ramp, oceanward shore of South Island, f: Sandy and beach dune on the southern side of South Island. Figure 20. a: Telok Jambu, West Island viewed from the north, b: Telok Kambing, West Island viewed from the west, c: Sheltered telok on South Island with stand of Pemphis on ridge at the mouth of lagoonlet, d: Ocean-dipping bedding revealed in trench on Home Island, e: Rubble-strewn shoreline on Pulu Wak Banka, f: Arcuate ridges, southern North Keeling; these appear to have been termed former reef margins by Guppy, but are reinterpreted as beach conglomerate marking foot of former rubble-strewn beaches. ATOLL RESEARCH BULLETIN NO. 404 CHAPTER 6 VEGETATION AND FLORA OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 6 VEGETATION AND FLORA OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS * ABSTRACT The vegetation and plant species occurring on the 22 vegetated Cocos (Keeling) islands have been classified numerically based on sample plots for the large islands as well as from checklists developed for each island from extensive reconnaissance. The vascular flora of about 130 species comprises approximately 50% native species, but there are no endemic species described. Most of the 69 introduced species are to be found on the larger settled islands and only one of these species has spread to the smaller islands. The relationship between island area and indigenous species richness shows a close fit to a power relationship. The more remote island of North Keeling has a distinct species composition compared to similar-sized islands of the main atoll. The pre-settlement vegetation has been extensively modified for coconut plantations, except for certain parts of North Keeling, where tall Pisonia grandis - Cocos nucifera forest occurs with small amounts of Cordia subcordata and other species. These forests are fringed on the lagoon shore by Pemphis acidula tall shrubland and on the exposed ocean shores by Argusia argentea shrubland. Each of these communities support breeding colonies of seabirds. On the main atoll, remnant vegetation occurs most commonly along the strand and in some places appears to be relatively recent. Many species on the Cocos (Keeling) atolls are restricted in their distribution there. In some cases these represent relict distributions, whilst a few could be considered to be pioneer populations. INTRODUCTION The Australian external Territory of the Cocos (Keeling) Islands is situated in the north-eastern Indian ocean at 12°S, 96°, is 2400 km north-west of North West Cape on the Australian mainland and 960 km south-west from Java. The Territory consists of two coral atolls about 25 km apart with a maximum ground elevation of 9 m above mean sea level. The smaller, northern atoll, is known locally and historically as Keeling Island and was inhabited intermittently and on a seasonal basis between ca. 1830 and 1929 (Gibson- Hill 1948). The main atoll, which continues to be known by various names and is here referred to as the Cocos atoll, consists of 21 vegetated islands, some of which have been Applied Ecology Research Group, University of Canberra, P.O. Box 1, Belconnen, Australian Capital Territory, 2616. 2 inhabited since 1827 and all of which have been extensively cleared for coconut plantations. Keeling Island (North Keeling), on the other hand, has retained more natural vegetation and is still a major seabird rookery for at least six species (Stokes et al. 1984). As the only atolls in the eastern Indian Ocean, and being relatively recently settled, these islands are of considerable scientific interest, but their isolation has prevented intensive scientific study. A series of naturalists have visited, the most notable being Charles Darwin for 11 days in 1836 (Darwin 1845); H.O. Forbes for 22 days in 1879 (Forbes 1879, 1885); H.B. Guppy for 5 months in 1888 (Guppy 1889); F. Wood-Jones for a year in 1905 (Wood-Jones 1912); and C.A. Gibson-Hill for 11 months in 1941 (Gibson-Hill 1950). Gibson-Hill (1948) presented the only systematic account of the vegetation in the form of a description of dominants and a sketch map of the plant communities of Keeling Island. Major plant collections have been made by C. Darwin (Henslow 1838), H.O. Forbes, F. Wood-Jones (1912), H.B Guppy (1889), I. Telford (1985) and the author in 1986-7. The Cocos (Keeling) Islands lie on an isolated spur of the submarine Ninety-East Ridge and are formed on a volcanic base rising from a depth of 5,000 m, with an unknown thickness of coral over the base (Jongsma 1976). Large solution and/or collapse dolines up to 20 m deep cover the south-eastern bed of the Cocos lagoon and possibly indicate a considerable depth of underlying limestone. Most of the islands are developed from coral sand, shingle and rubble deposits over a breccia platform that is just above mean sea level (Woodroffe et al. 1990) and beach rock commonly outcrops on the more exposed sandy shores. The highest elevations occur on the south and east ocean shores where sand and shingle deposits rise briefly to a maximum of 9 m forming a single elongate coastal dune best developed along the entire length of Pulu Atas as a sand dune and on North Keeling as a shingle ridge. Apart from these dunes, most of the islands are less than 3 m above sea level. The most leeward island of the Cocos atoll (Pulu Luar) displays a more complex geomorphology, being a mature moat island (Stoddart & Steers 1977) with a small saltwater lagoon. Jacobson (1976) studied the freshwater lens on Pulu Selma (Home Island) and concluded that the minimum width of island to sustain an exploitable Ghyben-Herzberg lens was 400 m. However during the present study in April and May 1986, fresh water was observed in wells on islands down to 100 m width (see Falkland 1988). The only naturally occurring surface freshwater is at the seasonal swamp known as Bechet Besar on the north-east shore of Pulu Panjang (West Island). On Pulu Luar (Horsburgh Island) there is a seasonal groundwater swamp identified by a ground layer of Mariscus javanicus. Meteorological records for various periods and locations on the Cocos atoll are available as a composite record from 1902, at least for rainfall. The annual average rainfall for the period 1902-1982 was 1994 mm with a range from 1099 mm to 3288 mm and a median of 1969 mm. Mean monthly rainfall varies from 81 mm in October to 256mm in April, with the dry season extending from September to December. Temperatures and humidities vary little throughout the year with an absolute annual temperature range from 21°C to 31°C. The wind régime is one of predominantly south- east trades for over 300 days per annum. Wind direction frequency analysis show south- easterly winds dominating from December to March whilst for the rest of the year there is a strong easterly component as well. Cyclone frequency within a five degree cell is about 0.25 in the region. Cyclones have passed near enough to the main atoll to cause damage in 1862, 1876, 1893, 1902, 1909, 1944, 1968, 1973 and 1988. 3 The purpose of the present survey was to establish the present composition of the flora on an island basis and to analyze the floristic and vegetation patterns of the entire Territory. The delineation of communities, relict stands and rare species distributions will serve as a basis for land use planning, the establishment of conservation priorities and the development of management aims. METHODS Circular sample plots of radius 10m were located on the six largest islands (Keeling, Luar, Tikus, Selma, Atas, Panjang) along transects selected to maximize the detection of vegetation and floristic change along environmental gradients (Gillison & Brewer 1986). The major environmental gradients considered in the layout of the transects were:- - ocean strand to lagoon strand; - potential for a freshwater lens; and - ocean coastline aspect. Transects were oriented at right angles to the ocean coastline (Fig. 2) and sample plots were positioned on both strandlines (ocean/reef and lagoon) and at 60 m intervals along the transect; unless there was a change in the height or composition of the top stratum, in which case additional plots were selected. In each plot, plant species present were recorded, as well as litter depth, soil surface texture and canopy height and dominants. All vegetated islands were surveyed (Pulu Pasir supports only occasional sprouting coconuts and was not included). For consistency, the Cocos-Malay names for the islands have been used throughout. Several islands have two Malay names apparently related to their origin from separate islands. In this report, Pulu Selma includes Pulu Gangsa (joined by human intervention) and Pulu Cepelok includes Pulu Wak-Banka, apparently joined by storm deposits before settlement in 1825. During the course of systematic collecting on the islands, the presence of each species on each island was recorded as well as an overall estimate of the species abundance on the island. A six-point ordinal abundance scale from very rare (less than 10 plants sighted) through rare, occasional, frequent, common to abundant was used. A complete set of voucher specimens is deposited at the Australian National Botanic Gardens (CBG) and nomenclature follows the Flora of Australia (1993). The floristic and quantitative data were primarily analyzed using the Pattern Analysis Package (PATN, Belbin 1992) at the C.S.I.R.O. Division of Water and Land Resources. The dissimilarity coefficient used for sites was the Bray-Curtis measure or the Kulczynski coefficient and the two-step procedure (Austin & Belbin 1981) was used for the between-species dissimilarity. Cluster analysis was hierarchical agglomeration using UPGMA fusion with B set to -0.1 to minimize space distortion (Belbin 1992). Vegetation patterns were also derived from panchromatic aerial photographs taken in 1976 ata scale of 1:44,400 (R.A.A.F. Film No. 8737) and 1987 colour photography at 1:10,000 (A.S.O. Film SOC760). Interpretation was done using a Zeiss Interpretoscope and transferred to a base map using a Zoom Transfer Scope. Island areas were measured off the R.A.S.C. Series R811 Cocos Island Sheet Special (1:25,000). RESULTS THE FLORA Exclusive of plants found only in cultivation the total vascular species count for these islands is 130 (Appendix 1). Given the variation in sampling intensity by past collectors (Table 1), it is difficult to be certain which species are introduced, except by examination of their biology, biogeography and present-day distribution as well as the historical record. Species found only in heavily disturbed areas and often on one or two large islands only, have usually been regarded as introduced in this analysis, and these account for about 50% of the flora (Table 2). Most of these species are pantropical herbs (Table 2) found on Pulu Panjang and many were probably introduced since the airfield was built on Pulu Panjang in 1944. The vast majority of the native species are Indo-Pacific strand plants that are predominantly sea-dispersed. There are no endemic species described at this stage, save for the variety cocosensis of Pandanus tectorius (Appendix 1). Of the 19 vascular species collected by Darwin (Henslow 1838) all but one have been recorded by recent collectors. Most are still common except for Cordia subcordata, Achryanthes aspera, Neisosperma oppositifolia and Laportea aestuans (Appendix 2). ISLAND FLORISTICS INDIGENOUS SPECIES The relationship between island area and indigenous species richness (Fig. 2) shows a closer fit to a power relationship (r? = 0.87) than a logarithmic one. When exotic species are included, the power relation is still a good fit (r? = 0.82), as the larger islands are also the most disturbed and colonized by exotic species. Cluster analysis for the 22 islands (i.e. including Keeling) based on the species abundance scores shows a clear grouping of islands by size, with Keeling being the most distinct floristically (Table 3, species groups A & E). The strand species form a distinct group (Table 3, species group D) well represented on all but the two smallest islands, Beras and Ampang Kecil, which have areas less than 0.5 ha and support only three of the six common strand species. This strand group comprises Argusia argentea, Pemphis acidula, Guettarda speciosa, Cocos nucifera, Scaevola taccada and Ipomoea macrantha, all of which have marine dispersal powers. Some species were found almost exclusively on islands larger than 20 ha, and most of this group were more abundant on Keeling (Table 3, group A). These included the trees Cordia subcordata, Hernandia nymphaeifolia and Pisonia grandis as well as Achryanthes aspera, Dicliptera ciliata, Portulaca oleracea, Boerhavia repens, Stenotaphrum micranthum, Lepturus repens and Sesuvium portulacustrum. Species group B (Table 3) represents those species common on the larger Cocos islands but absent or less abundant on Keeling. Of those species which do occur on Keeling, most are rare there, often recorded from one or two locations only. Species group C consists of three species each found at just a single site on Pulu Panjang. These are Lepturopetium sp., Ximenia americana and Enicostema axillare (Appendix 2). EXOTIC SPECIES Only one exotic species, Turnera ulmifolia, has spread to all the islands of the Cocos atoll, and it is usually abundant wherever it has established (Table 4). The large but relatively unsettled Pulu Atas has been colonized by six exotic species and five have reached Keeling. Most exotic species are confined to the four large islands that have had or still have intensive settlement. Thirty two of the 63 exotic species occur only on Pulu Panjang and/or Pulu Selma (Table 4, groups C & D, part of A) and nineteen occur on Pulu Panjang only (group C). At the other end of the size scale, the four islands without exotic species (Beras, Blan, Blekok, Jambatan) are all less than 2.5 ha. VEGETATION PATTERN ON KEELING Analyses of the transect plot data for Keeling were done with the total set of 26 species recorded in 65 plots along 11 transects. Another 10 species were recorded for North Keeling in reconnaissance. The floristic classification analysis does not exactly correspond with the dominance-based units able to be mapped from aerial photography and ground checking. Stands of Pemphis acidula tall shrubland (2-4 m) and Cordia subcordata tall shrubland (3-6 m) occur close to the lagoon shore and are commonly mono-specific (Table 5, site groups 1, 2 & 3), and, where there are finer sediments accumulated, a Sesuvium portulacustrum herbland is developed (Table 5, group 1), often lying between or within the two former types (Fig. 3). Site group 4 (Table 5) is characterised by exposed shore halophytes, such as Portulaca oleracea, Lepturus repens and Boerhavia repens, Cocos is absent. Site groups 5 to 8 highlighted floristic sub-units within the closed forest stands characteristically dominated by Pisonia grandis and/or Cocos nucifera (Table 5, Fig. 3). Group 5 contains the beach halophytes (species group A), group 6 has an understorey of forest mesophytes (species group C), group 7 are stands of pure Cocos and Pisonia, while group 8 are virtually pure Pisonia. Group 5 mainly represents relatively richer plots (s = 6.7) found within 20 m of the shore which have strand forest dominated by Cocos along with halophytic shrubs and herbs which typically occur only near the shoreline. Pisonia grandis and Stenotaphrum micranthum are constants and the former may be co-dominant on sheltered shores. Some plots in this class fall within areas which are mappable as Pisonia shrubland occurring on exposed shores usually behind a beach-fringing Argusia shrubland. Site group 6 (Table 5) corresponds with relatively species-poor areas of forest (mean richness of 3.9) dominated by Cocos nucifera and/or Pisonia grandis. The associated species include broad-leaved plants such as the climber Canavalia cathartica, Morinda citrifolia, Rivina humilis and Carica papaya. Sites in groups 9 & 10 contain species which are uncommon on North Keeling (species groups A & D). Many species on Keeling have a restricted distribution and most of these are found on the northern peninsula at the entrance to the lagoon or on the north-west lagoon shore and adjacent habitats. The same pattern is evident for the species recorded in transect plots. The richest floristic units, apart from the herblands, are the forest types found near the lagoon entrance and on the northwest side of the island. 6 VEGETATION CHANGE A comparison of Fig. 3 with the vegetation map of Keeling Island produced by Gibson-Hill in 1941 (Gibson-Hill 1948) shows geomorphic and a a changes evident over the intervening 45 years. The west-building peninsulas at the lagoon entrances have extended considerably into the lagoon. On the northern peninsula the Argusia shrubland mapped by Gibson-Hill is possibly the small area in a similar position mapped in 1986, since it is on a rocky substrate which may have conferred a degree of site stability. The southern arm has a similar clump of Argusia that has apparently extended and been flanked by Cocos and Pemphis. Observations on sand and shingle bars in the lagoon entrance channels show that Argusia is the first woody plant to colonize such places in exposed locations and Pemphis does the same on more sheltered shores. Both peninsulas show a definite sequence of changing dominance along their strands from Argusia, Pemphis and Cocos on the younger shores to Cordia and Pisonia on the older ones, reinforcing the interpretation of rapid development of these peninsulas. Being more sheltered, the southern peninsula has developed an area of Pisonia forest on its southern end, perhaps largely since 1941. Another change in the vegetation since 1941 is the loss of the Argusia zone along the southwest coastline and possibly on the southeast also. Both locations have 4-5 m high shingle ridges with their seaward faces lying at the repose angle, and show evidence of episodic deposition of shingle into the Pisonia zone suggesting storms have removed the Argusia. Gibson-Hill (1948) mapped an area of open grassland on the south-eastern lagoon shoreline which he said was a breeding habitat for shearwaters (Puffinus sp.). These birds have not been seen on the island for some years (Stokes et al. 1984) and this grassland has now become in parts an open shrubland colonized by Pemphis, Pisonia and Argusia to 3 m high. VEGETATION OF THE COCOS ATOLL Cluster analysis of the 106 sample plots on the larger islands (containing 52 species, native and introduced), produced at the ten-group level five major site groups and five further groups represented by a few sites each (Table 6). Floristic definitions of the major site groups emphasise variation in the coconut woodlands and forests in relation to ground layer composition and location relative to the ocean and lagoon. On the most exposed southern and eastern strands of Pulu Atas there are areas of Argusia - Scaevola shrubland and patches of Lepturus - Triumfetta herbland on the sand and shingle ridge topping the beach. The minor site groups reflect distinctive relict communities, mainly on Pulu Panjang and Pulu Luar, where strand trees have survived land clearing or colonized recent deposits. Species such as Calophyllum inophyllum, Guettarda speciosa, Hibiscus tiliaceus and Barringtonia asiatica characterize these sites and their distribution over the Cocos atoll is very restricted (Fig. 1, Appendix 2). These species occur scattered along the lagoon shores and on the sheltered west shore of Pulu Panjang, mostly as single trees or small clumps. The largest remaining stands of these species are to be found on Pulu Panjang along the northwest shore and adjacent to the swamp at Bechet Besar; along the 7 lagoon shore opposite the northern end of the runway; and on the southern lagoon shore of Pulu Luar (Fig. 1). Pemphis acidula shrubland forms 2-4 m high shrublands scattered all along the lagoon shore, particularly in areas where sand deposition is occurring (e.g. at Tanjong Klikil at the east end of Pulu Panjang) and also at the lagoonward edge of intertidal sandflats. No other species of plants grow in these offshore strands except for an occasional coconut seedling and epiphytic mosses and lichens. DISCUSSION FLORA Island floras originate from a variety of sources depending on their geographic location and suitability of their habitats for immigrant diaspores. The origins of the Cocos (Keeling) biota have long fascinated biologists, particularly those who have examined their plants and insects (see Guppy 1890, Holloway 1982). Renvoize (1979) suggested that island structure in terms of elevation and geological substrate are key factors in determining the richness of island floras. In this respect, the Cocos (Keeling) islands bear the greatest similarity to the central Indian Ocean islands (Laccadive, Maldive, Chagos) and to only some of the western Indian Ocean group (Cargados Carajos, Tromelin, Agalega, Amirante group, Alphonse, Gloriosa, Europa and Farquhar group). All these low islands have evolved in isolation from a continent, through the combined forces of vulcanism, subsidence and coral growth, and presently rise less than 10 m above sea level. The low habitat diversity of these islands leads to a flora characterized by very low endemicity with indigenous taxa of pantropical or Indo-Pacific distribution dominating (Renvoize 1979). Cocos (Keeling) is no exception to this general pattern; it has no endemic flora save for the variety cocosensis of Pandanus tectorius, and with 61 indigenous species (Table 2), is comparable with the Laccadives (40 indigenous species), the Addu atoll in the Maldives (52 species) and the Chagos group (ca. 100 species) (Renvoize 1979). It is also similar in species richness to western Pacific atolls such as Nui (44 species) and Kapingamarangi (50 species) (Woodroffe 1986). The mechanisms of natural dispersal to oceanic islands include wind, ocean currents, birds and bats. Undoubtedly all of these have contributed to the Cocos flora, (even bats have been occasionally sighted, Marlow 1970), but the only agent for which evidence is certain is that of oceanic drift. Most of the strand species are found as seeds on beaches and there is a further component of the flora that is found only on the drift line (Guppy 1890). The main currents around Cocos (Keeling) are westward and would be expected to derive propagules from northern Australia, Torres Strait and Java. These currents are reinforced for most of the year by the prevailing southeast trade winds. ISLAND RICHNESS AND COMPOSITION The power relationship established between indigenous species richness and island area (Fig. 2) is similar to that reported for Nui atoll in the Pacific (Woodroffe 1986) and 8 for the lagoonal islands of Aldabra (Hnatiuk 1979). None of these three species-area relationships support the notion of the small-island effect suggested for Kapingamarangi (Niering 1956, 1963), where very small islands tended to have area-independent species richness. The six commonest strand species are not always present on the smaller islands and show an increase in frequency of occurrence up to an island size of 10 ha, and are always present above this area (Table 3). If one sets aside these species, i.e. Argusia argentea, Pemphis acidula, Guettarda speciosa, Cocos nucifera, Scaevola taccada and Ipomoea macrantha, then the similarities between the smaller islands are very low. Among the 20 other species, there are only 41 occurrences on the 16 islands smaller than 25 ha. The presence of these species was often correlated with minor and possibly ephemeral habitats; e.g. Hibiscus tiliaceus on a small sheltered lagoon-facing shore of Pulu Jambatan where a channel had cut across the prograding western (lagoon) side of the island; Suriana on recently formed sandy spits extending lagoonwards on Pulu Siput; and Mariscus javanicus, Fimbristylis cymosa and Lepturus repens where lagoon shorelines distant from inter-island channels had relatively flat shingle embayments at upper tide levels. These observations support the idea that habitat diversity needs to be considered in modelling species richness on islands (Buckley 1982). Dispersal routes may also contribute to differences between the smaller islands, especially where islands are more likely to receive a high density of propagules. For example, Pulu Labu probably intercepts a higher number of propagules because of its position at the tip of Pulu Atas, where the equatorial current flows northward along Pulu Atas then some sweeps into the lagoon. An exceptionally high abundance of drift seeds was found to occur on the sand dunes at the northern tip of Pulu Atas and in a similar situation on the ocean beach of Pulu Gangsa. This may help to explain the occurrence of Calophyllum, Barringtonia and Neisosperma on the former island and on no other small islands, save for a single Calophyllum on Pulu Beras. The latter is also an island which is well situated to receive propagules concentrated by northward transport along the ocean shores of Pulu Selma. However these data are insufficient to suggest such islands have more species for their area, and there are presumably other factors operating, such as stability and age of an island. Amongst the six larger islands, Pulu Atas has a relatively low richness for its area, possibly related to its more uniform geomorphic structure, rugged ocean coastline and lack of currents flowing along its lagoon shores (although they would have done so in times past before channels closed off). North Keeling has seven species not found on the main atoll but is also different in composition from the other large islands. Some species (Table 3, group A) are relatively more abundant there, either because of greater areas of suitable habitat on Keeling (e.g. for Boerhavia repens and Portulaca oleracea in exposed herblands; Sesuvium portulacustrum in saltmarshes; Stenotaphrum micranthum, Achryanthes aspera, Dicliptera ciliata in the Pisonia grandis rainforest); or due to clearing over the last 160 years (e.g. Cordia subcordata, Hernandia nymphaeifolia and Pisonia grandis). Species group E (Table 3) found only on Keeling may represent in large part the extreme effects of vegetation clearance on the southern atoll. The restricted distribution of Thespesia populnea (in a clump of six individuals on Pulu Tikus and in a small mixed stand with Cordia on Pulu Luar), probably also represents a relict distribution resulting from extensive cutting in the past, as the bark fibres were once used for netting (Gibson-Hill 1947). ACKNOWLEDGEMENTS This study was initiated in 1986 whilst the author was on study leave from the University of Canberra (then Canberra College of Advanced Education). I am grateful to the Cocos (Keeling) Islands Council for its interest in the project and for allowing access to all parts of the Territory; to the Administration of the Territory for providing field logistic support; and to the Bureau of Flora and Fauna and the Australian National Botanic Gardens who supported the plant collecting. Ian Telford provided a plant species list and identified voucher collections. Amat Noor bin Anthoney, Peter Goh and Tony Stokes helped with survey work on the atolls. The C.S.I.R.O. Division of Water and Land Resources provided facilities for numerical analysis and I am grateful to Mike Austin, Lee Belbin, Dan Faith and Peter Minchin for advice on these analyses. REFERENCES Austin M.P. & Belbin L. 1982. A new approach to the species classification problem in floristic analysis. Aust. J. Ecol. 7: 75-89. Belbin L. 1992. PATN Analysis Package: Technical Reference Manual. Division of Wildlife and Ecology, CSIRO, Canberra. Birch, E.W. 1866. The Keeling Islands. Proc. Roy. Geog. Soc. N.S. 8: 263-265. Buckley, R. 1982. The habitat-unit model of island biogeography. J. Biogeog. 9: 339- 344. Darwin, C. 1845. Journal of researches into the natural history and geology of the countries visited during the voyage of H.M.S. Beagle round the world, under the command of Capt. Fitz-Roy, R.N. John Murray, London. Faith, D.P., Minchin, P.R. & Belbin, L. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69: 57-68. Falkland, A.C. 1988. General Report. Vol. 1. Cocos (Keeling) Islands Water Resources and Management Study. Hydrology and Water Resources Unit, A.C.T. Electricity and Water, Report No. 88/12. Flora of Australia 1993. Volume 50, Oceanic Islands 2. Australian Government Publishing Service, Canberra. Forbes, H.O. 1879. Notes on the Cocos or Keeling Islands. Proc. R. Geog. Soc. 1: 777- 784. Forbes, H.O. 1885. A Naturalist's Wanderings in the Eastern Archipelago. A narrative of travel and exploration from 1878 to 1883. Sampson Row, London. Fosberg, F.R. & Sachet, M.H. 1982. Micronesica 18: 73. Gibson-Hill, C.A. 1948. The island of North Keeling. J. Malay. Br. Roy. Asiatic Soc. 211: 68-103. 10 Gibson-Hill, C.A. 1950. A note on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 11- 28. Gillison, A.N. & Brewer, K.R.W. 1986. The use of gradient directed transects or gradsects in natural resource surveys. J. Env. Mgt. 20: 103-127. Guppy, H.B. 1889. The Cocos-Keeling Islands. Parts I-III. Scot. Geog. Mag. 5, Part I, 281-297, Part I, 457-474, Part II, 569-588. Guppy, H.B. 1890. The dispersal of plants as illustrated by the flora of the Keeling or Cocos Islands. J. Trans. Victoria Inst. Lond. 24: 267-303. Henslow, J.S. 1838. Florula Keelingensis. An account of the native plants of the Keeling Islands. Nat. Hist. Mag. 1: 337-347. Hnatiuk, S.H. 1979. Numbers of plant species on the islands of Aldabra Atoll. Phil. Trans. R. Soc. Lond. B 286: 247-254. Holloway, J.D. 1982. On the Lepidoptera of the Cocos-Keeling Islands in the Indian Ocean, with a review of the Nagia linteola complex (Noctuidae). Entomologia Generalis 81: 99-110. Jacobson, G 1976. The freshwater lens on Home Island in the Cocos (Keeling) Islands. Bureau of Mineral Resources Australian Journal of Geology and Geophysics. 1: 335-343. Jongsma, D. 1976. Review of geology and geophysics of the Cocos Islands and the Cocos Rise. Bureau of Mineral Resources, Australia, Record 1976/38 (unpublished). Marlow, B.J. 1970. A record of a Mastiff Bat, Tadarida plicata, from the Cocos (Keeling) Islands, Extrait de Mammalia, 34. Niering, W.A. 1956). Bioecology of Kapingamarangi Atoll, Caroline Islands: terrestrial aspects. Atoll Res. Bull. 49: 1-32. Niering, W.A. 1963. Terrestrial ecology of Kapingamarangi Atoll, Caroline Islands. Ecol. Monogr. 33: 131-160. Renvoize, S.A. 1979. The origins of Indian Ocean floras. in Plants and Islands. (ed. D. Bramwell). pp107-129. Academic Press, London. Stoddart, D.R. & Steers, J.A. 1977. The nature and origin of coral reef islands. In Biology and Geology of Coral Reefs. Vol. IV. Geology 2. (eds. O.A. Jones & R. Endean) pp59-105. Academic Press, New York. Stokes, T., Sheils, W. & Dunn, K. 1984. Birds of the Cocos (Keeling) Islands, Indian Ocean. Emu 84: 23-28. Williams, D.G. 1990. An annotated bibliography of the natural history of the Cocos (Keeling) Islands, Indian Ocean. Atoll Res. Bull. 331: 1-17. 11 Wood-Jones, F. 1912. Coral and Atolls: a history and description of the Keeling-Cocos Islands, with an account of their fauna and flora, and a discussion of the method of development and transformation of coral structures in general. Reeve & Co. Ltd., London. Woodroffe, C.D. 1986. Vascular plant species-area relationships on Nui Atoll, Tuvalu, Central Pacific: a reassessment of the small island effect. Aust. J. Ecol. 11: 21-31. Woodroffe, C.D., McLean, R. & Wallensky, E. 1990. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. Nat. Geog. Res. 6: 262-275. 12 Table 1. Sampling intensity and the number of native and naturalized (not horticultural) vascular plant species recorded by major collectors and naturalists on the Cocos (Keeling) Islands. Collector Year Period of visit Islands visited ’ Number of species C.R. Darwin 1836 10 days Cocos atoll 21 H.O. Forbes 1879 22 days Cocos atoll 38 W.E. Birch 1885 8 days Cocos atoll 11 H.B. Guppy 1888 10 weeks Both atolls 53 F. Wood-Jones 1909 15 months Cocos atoll 46 LR. Telford 1985 2 weeks Panjang, Tikus, Atas, 93 Selma, Keeling D.G. Williams 1986/7 9 months All, including Keeling 130 Table 2. Life forms of the native and naturalized flora. Origin Climber Forb Graminoid Seagrass Shrub Tree. otal Native 7 13 11 3 9 18 61 Naturalized 3 29 2 0 10 6 69 Table 3. Two-way classification for native species occurring on all islands. Classification based on abundance data standardised by species maximum. The first four letters of the generic name and specific epithet are read vertically. Numerical values represent the abundance scores standardised by species maximum. Island Species groups BOSS ASSSe> E, BOSSSS2309355 Be — s-=--> <-D--> <-E--> | ABPHDSASACPL | CDCCCPTZITEHTCMMMF PSCPVTR|ELX|AIPCSG|V|BN|Q|CEPACL| | COOETECTCOTE | AARAYHHOPEUIRAEAOIRULAIHH | NEI |RPEOCU|I|AE|U|ARALLA | | AERRCSHEARSP | EDILPYUYORPBISLRRMERENTEI | IPM| GOMCAE|G|RI|E|NYSLEP | | LRTNLURNLDOT | SONOELASMMHIUSAIIBMIRDESZ | CTE | UMPOET|N|RS|E|ATPOOO | | | | | HU vo | LRONCPAMISGR | BVAISAIMPCATPFBJCCSMITTPA | AMA | AMANTS |M|AO|H|CVVCGA| | AELYIOSINURE | O[ISNTMNAEATIRIIAIYEANEROP | XAM|RACUAP |A|SP|Y|AAAOYE | | NPEMLRPCDBAP | NSIOOAVTSTOLOLFVTMRRECIPI | IRE|GCICCE|R|IP|A|TRGBNS | | CERPITERICNE | DCAPLRORCATICILARORIRTFUC|LSR|ERDICC|I|AO|L|HIIBAT| Keeling 1166666666666|1..... oh ob Die, 0) ook A si Slope overlies so) 450025. - 1666666] Luar |.43.111..311]6626641646163655563....66]...|452662|. a clo so'o.c.c0 | Panjang |.1361..1.223111151166666645455634.1...|666|/466664|.|/1.].]......1] Selma le TR RES ree eR Ee cs, oalletiliis) «er 016 | Atas [hei AGrewsuoneneens 3|...4.1414..665545.41.....]...1646662|.].1].]......1 Tikus 16.3......3..11162664146...21666..3..1.]...1]122661|.|]..|/6]...... | Cepelok loScooogoocodabllooadtocoototooke lost ooo olloco IMOSOGGRI lls sifollons oor | Pandan lSCoo0c000 pL Vet SilioveHeNovolloMerielelloHellellsueyold Mr MioievenereyeronltereNo iO OO O.GlOilrellwoul veil teoiteweronononll Ampang loobocannnc co llondooonopoU OOOO Ooto OOOO ooo SOOO SIGI So lls ooacas | Kembang VosbSSb6o0bb00 owooogG 6 be cotex ooo do05 ollobo IOTSGGS Ic lS loons so | Wak-Idas oogoaoeodocolloococccsonodotloosdoodoosolloeolS50005lcllaclloiliccc oa | Blekok Voe60 60006004 losogoodooood obo 7456 00006 6 [oloo JEKSSOA I cio ollolleocéoc | Blan lo ooo ogo0006 codecs coco ada bO8 O0'd “|... 11566151. oollolloog eon | Kambing 7 erase RP Aaa Ue ac ede nny Meets nal eg © ofSCSG6 [ollos lolioos soo Kelapa Satu loo ooosoocoDR loodaodad eKelisfoneyellevole Sooo go olblooolloMoOoIDolloollollaooocs | Blan Madar lo Go5 6 0b0 S60 polloocoooodbodoo oO Kdss g oie SISSON No oo [loilleson5c || Maraya looodocoo ods loSoSdosedagccgoshoosoodcollooo LESBO lolloollollodsacs | Siput loGooSoo cos0ollogoan0b concn odbaon06DocodollsooSoA606| oho ollelloodoss | Jambatan lWemcttonoweitsHeiteere so llocccoodoooododgbncooooddb ooo IMoBOOS ooo lho lecoccs | Labu logoooodio6e collosotlotocoddcoagoo oboe loo ICMOER io lGSlle lec u0s | Beras Hio-o O:0'0-0'0.0 cooollooodboooudodooodocceoooooollooo Io st!Oallatlos lic llocoson'l| AMPANG PREC al ieoveiey si siehelienes siete evekonclelencdelciedersieker evens coo oo loool Ob oOo'llai}o%s' [ls lo7adi6 35 | Table 4. Two-way classification for exotic species occurring on all islands. Classification based on abundance data standardised by species maximum. The first four letters of the generic name and specific epithet are read vertically. Numerical values represent the abundance scores standardised by species maximum. Island Species groups groups BR OSSS0S555 FN Pe 6a Ba ie inn C=----=—=—— > <2) | BTVEEEMACDEEESDSHCSMRP | LCSPR| BBCCCHDCCEIPDAEBSGS |CS|RTT|IS|VS|CBLSSIAZ| | OREURUUUEALULPIPEYTAHH | EYCHI | OREHYIEHOMMAEPRRORT |AI|IRU|SE|EO|AREOENEE | | ETRPAPNSNCEPEOGEDNACOY | PNOYV | TENLPPSRNIPSSLIARAR|RD|CIR|CS|RL| SYUNNDRP | | RDNHGHTTCTUHURIRYOCRES | IOPLI | HYCOEPMYYLESMUOCGST | IA| IPN|HB|NA|UOCCNIVH | | | | tol lip ghea al | DPCHTPCIEAICRISACDJASM | VADNH | BDCBPLTABSCFBMMBBSA | PA| CTU |MG|CA|EPLOOHLR | | IRITERANCENYUNESOAATPTI | IRUOU | LITAOORCOOYOIUERIPS |AC|ORL|UR| IM|QIELCIAO | | FONRNOLUHGDADDTSRCMRAN | RCLDM| ASLRLNIINNLEPTYIC1TI | PU|MIM|TA|NE|UNUECRNS | | FCETESALIYITEIIUYTAOTI | GUCII | DTIBYGFCACITIIEZOOA|AT|MFI|IN|ER| INCRISAE | Keeling [PSP Men okel oxelotatel shettotcvetetetel of oll POO 1 O00. 6 OOO. 0 OORTEOS OGiP Lea serllReks. [evsiexenon owes | Atas [retenctie\thiee arse e cloveie ofete re) o sel Ol otel Ol lloveterelenoneteretetete e)/ctens sient besten Loe lisse healleurorcaos 6 | Luar GOD h eit er eleratareretere HY V7 Bosra SS omc. coop OOOO C [rere [1Gs.:53]'GGilhy clio ete eres crems | Tikus [rover eG ST ret rover d dds Giotel'|! reverent lietettovetevetetetctete eo ererolencrele: [voted [Ascii lKereM LOO ill otehenelerens i| Selma 11164666116666666661664|..... [Roterevetohetetetetel ciel eelete elers |..1164|..1..166666666| Panjang 16666666666666166611111 | 66111 | 6666666666666666666|..1]..6]..]..]....266. | Kambing [Keervetorstetchelerstate av oreh sVetel of oi of |Holel Olet onl hefetocete tel shictictetelel lel cheloters Gel beuex’j hoo lomlloonoa5 o.0|| AMDPANG RECA) "css stes sues el cl el oi eiene rere eheiereke [Rexel suet [Rokshetetclelotetscoteveleretsl sf clots [Foren lher cdl vere! Ineredl kenehenetenerete | Ampang |hahedshetelstenetete aictiotavetemetetahetele [foteretete [RokakeheteteNetotateteleletetoletelets eaeel eeteal lecacal Waxed OtC00-0.0.0 | Cepelok [Fopeteustevere’ clevetetereorereleyerenciors [Reterorete [Berctevertoncieicharstetsyereveterere’s [Pico [Ps oc! | ere l csc caltovenenctonetene | Pandan Wpasatotersteletayetene ale Mev hctotetoton|Vohetetelod iotetoletetet Reteletetete | cle leletons [icken Foret DilHenen[Neke: IWenoeenenete rel Siput [Po emeetc! cratelereleteleter everetshetetove [BototteRoteMlichetohelet hetolictetetslstetoiehene [retetl revel) [hetoa led cl we) orenestonetem Labu [Retetetel eles tote skated eteNeVoksletvolehet Retetotetel lNehetotstotclcleReletetotelohetet eters Pesecul boson heel [es | srenenonedl Blan Madar [sotonetokeketetolehetsfetevonere oh ofot ater’ lovee Noted Weketotetehoncteletote teks Better [eoten [retsi Di lioted [VoroMioretonaione vel Maraya |Hercxchetelstelsteloleretetehateteretetelete [Nokekctete [Pokerctchctehetststakehahet alate shotel [sree [beret Stfleron hele [ila itetetamerets | Wak-Idas 5-5-0. 00-5-3.0-6 cS ESO OO.Oe otol lberetoNetol oMekoreteekeNed sotto! olekehetoters Reset Wee] bitiecte eronl Heron b:5:0.010 | Kembang |;e¥evatotolatatatatehotetereteretoletetetote |Koweharete [HcFaltetetotelotetersteterotetelclctels Wezel booe tl Focdloallo a o6.06.0 | Kellapa: Satu « s[hcicctoe.rcyere cn etcret svete avototetetete [Wetemettete [NellebeteWelalaterstete ts ahekevetolote? [ster [fete Ail boron Weheull aterekettete cnet 15 Table 5. Two-way classification for all species recorded in transect plots on North Keeling Island. First letter of the site code represents the island name, second and third letters refer to the aspect of the ocean coastline. Site Species groups groups CRCPCCDMT AIOIOAIOR NVCSRRCRI AIOODILIP CHNGSPCCT AUURUAIIR TMCABPLTI HIINCAIRP uw a my ~ e a 2g ~ e eee te eee Ss) z — wo eo a ~ a 16 Table 6. Two-way classification for all species recorded in transect plots on the Cocos atoll. First letter of the site code represents the island name, second and third letters refer to the aspect of the ocean coastline. Asterisks indicate presence of a species. Species Site groups groups | LTLLLLLLSSLLPSSPPPPPP | LAAAATPPPSPPPPPPPPAAAPPP | LAAATAAAAAPAAPAPAAAP PAPPAAAAAPAAAAPAA | TTATTAPPPPPTTT | AAAAAAAAP |A| | NENNNNNNNNWNSNNWWWNWW | NSSSSNSSSNWWWHWNNREWWHNY | HSSWESSSNSSSSWSWHNSSSESSNNSHEWNNNNWS S | NNSNNSSSSSWNEE | NESNSSSES|S| |\////////88/ /8BB111W11| /BWWW/BEBB221W1iWWwW/ //Ww| /BW/ /EHEEERBH2W1E8HEE/ HERES / /1HBHB2BW| //B/ /BEBEE2///|B/BEBEB/B|B/| 100462538//792////////\14///S119/////////112///\16/0346316133/ //2222152221434/2111/2/ | 0144221111/312 | 202132222}3) | 32 /04347656/0/234 ///101132012 748) /0 SISIIILIESOSII11 S1T1/) OFF FF2/11 FI A11713 \/ ///17 SNA | 1 14 132 16 12240043 01548 32233 32013 | 5S 45467 14 05021 0/1) eeeeecerrerreererreerrere * te ef a i a ° wo ~~ ° BOER REPE | | | ' hs I*1 LEPI VIRG | I a | ! | iil ~ = * a = 4 z re] * . BNIC AXIL | | ' | | Plat PAND TECT | | | ' | uel CALO INOP | ** eS, CYPB STOL | *** * bd STAC JAMA |***** u IPOM PBSC |* oe DODO vIsc |* be * teteeeree er eee © © | * eervore ‘| * * ef ee ee] FPIMB CYMO | * **#tteeeene eee ZOYS MATR I* . eerereeerrerereee) ener tee esee * | TRID PROC | trees wee 1 COCO NUCI |**#teetereerrareerere ieee tee eeree Pee eee ee ereeeeKeHHHESe | HetEseeseHeeEe| * SCAB TACC |** *### tees & & [eereeres TURN ULMI | eee leeee eneee MBLA BIFL | we oe eeeerel|eeres eeeeeee e| ee ter 1 . Il teeere oo je eereecere « teeeres [ete ee | teeweeres cere tere fete eeeeerets serete|sereeeete see | * MORI CITR | * + tee eeeeeee] Ca {* tee * [teeeeereree 2; CRIN ASIA | i} * ] | * tee] oe tees jeeee tee! IU we eee . reece) ° eee] y 2 w q ~~ 5 BARR ASIA | | eal | | UW | CABS BOND | | Pica | | lint HIBI TILI | | oh AS | | av SS SS SES 65 17 64 62 60 59 55 x 5! 50 49 (Pulu Panjang) Horsburgh Island (Pulu Luar) Directi (ACHES INDIAN OCEAN Pulu Pasir a om t Pulu Beras = ith Pulu Gangsa | Home Island ie (Pulu Selma) WESTERN ENTRANCE 9 Pulu ameang Kecil | ; Pulu Ampan ee a Pulu Wak Idas | |—#2_. Pulu Blekok ———t— PuluKembang &. | Pulu Cepelok | = Pulu Wak Banka | Pulu Pandan { if See} Pulu Jambatan ) Pulu Sleu =} — ai Pulu Labu Pulu Pulu. Kelapa Pulu EIN SSN : i West Island Maraya —I South Island (Pulu Atas) 62 Figure 1. Location map showing island names, localities mentioned in the text and remnant vegetation patches on the Cocos atoll. Refer to Appendix 3 for a description of the remnant vegetation units. Gridlines represent 1000 m grid of the Australian Map Grid, Universal Transverse Mercator Projection. Map base derived from R.A.S.C. Series R811 Sheet Special Cocos Island 1979. 18 Figure 2. Number of native species 100 10 / Kelapa Satu / A 74 Beras ,/ 0.01 0.1 1 10 100 1000 Area (ha) Species-area relationship for indigenous plant species richness for the 21 vegetated islands of the Cocos atoll and the island of North Keeling, showing approximate 95% confidence limits. Non-linear regression fitted to obtain the equation s = 6.73a°-28. Cocos Forest Pisonia Forest Cordia Tall Shrubland Pemphis Tall Shrubland Argusia Shrubland Pisonia Shrubland Sesuvium Herbland Boerhavia Herbland Mixed Shrubland Premna Tall Shrubland 1000 m Figure 3. Vegetation map of North Keeling Island. 19 20 Figure 4. Pulu Beras, one of the smallest vegetated islands, with a cover of Cocos nucifera, Argusia argentea and Scaevola taccada. Figure 5. Boerhavia repens herbland grading into wind-sheared Argusia argentea shrubland on the south-east coast of North Keeling island. This is the breeding habitat for the Brown Booby. 21 Figure 6. Sesuvium herbland adjoining Pemphis shrubland with Cocos -Pisonia forest in the background, on North Keeling island. Figure 7. Rhizophora apiculata grove on Pulu Luar lagoon. Sesuvium herbland on coral shingle in the foreground. 22 Appendix 1. Vascular plant species list for the Cocos (Keeling) Islands. Asterisk indicates introduced species. Family Generic Specific Authority Var,/ Common __ Local Name Epithet Subsp. Name Name ACANTHACEAE Dicliptera ciliata Decne. AIZOACEAE Sesuvium portulacustrum fee) nL Sea Purslane AMARANTHACEAE Achryanthes aspera (es var villosior Chaff Flower (Henslow) D.Porter *__Aerva lanata L.) Schult. APOCYNACEAE Neisosperma oppositifolia (Lam.) Fosb. Kayu Laki & Sachet ASTERACEAE * Austroeupatorium inulifolium (Humb.,Bonpl. Stinkweed et Kunth) King et H.Robinson *__Eleutheranthera ruderalis Sw.) Sch. Bip. * Emilia sonchifolia L.) DC. Melanthera biflora (L.) Wild. Beach Sunflower *___Tridax procumbens -L. *__ Vernonia cinerea L.) Less. var. cinerea * Vernonia cinerea (L.) Less. var lanata J.T. Koster BORAGINACEAE Argusia argentea (L.f.) Heine Octopus Bush Kayu Sireh Cordia subcordata Lam. Sea Trumpet Geron- ggang BRASSICACEAE *_Lepidium virginicum L. CARICACEAE * Carica papaya LE. Pawpaw Katis CASUARINACEAE Casuarina equisetifolia Le: subsp. Coastal Cemara equisitifolia Sheoak CLUSIACEAE Calophyllum inophyllum L. Alexandrian Nyampl- Laurel ong COMBRETACEAE Terminalia catappa ke: Sea Almond Ketapang CONVOLVULACEAE |pomoea macrantha Roem. & Moon Flower Schult. CONVOLVULACEAE |pomoea pes-caprae (L.) R.Br. subsp. Goat's-foot Kangkong brasiliensis Convolvulus Meryap L.) Ooststr. Acalypha lanceolata Willd. * — Breynia disticha J.R.Forst. & G.Forst. Euphorbia atoto G. Forst. * Euphorbia cyathophora Murray Dwarf Poinsettia * Euphorbia hirta Le EUPHORBIACEAE FABACEAE FLACOURTIACEAE GENTIANACEAE GOODENIACEAE HERNANDIACEAE LAURACEAE LECYTHIDACEAE LYTHRACEAE MALVACEAE ‘ MIMOSACEAE MYRTACEAE NYCTAGINACEAE OLACACEAE PASSIFLORACEAE PORTULACACEAE RHIZOPHORACEAE * Euphorbia Phyllanthus * Ricinus * Sauropus * _ Alysicarpus Canavalia * Crotalaria * __ Desmodium Erythrina * _ Indigofera *_ Macroptilium * Sesbania *___ Sesbania Vigna * — Muntingia Enicostema Scaevola Hernandia Cassytha Barringtonia Pemphis Hibiscus *___ Sida Thespesia * Leucaena * Eugenia * — Psidium a Boerhavia Pisonia Ximenia * Passiflora Portulaca Rhizophora prostrata amarus communis androgynus vaginalis cathartica retusa triflorum variegata hirsuta atropurpureum cannabina grandiflora marina calabura axillare taccada nymphaeifolia filiformis asiatica acidula tiliaceus acuta populnea guajava albiflora grandis americana foetida oleracea apiculata Aiton Schumach & Thonn. L.) Merr. L.) DC. Thouars. in A.N. Desvaux (le DC.) Urb. (Retz.) Poir. L.) Poir. Burm.) Merr. (Gaertn.)Roxb. (C.Presl) Kubitzki fe (L.) Kurz. J.R.Forst. & G.Forst. ( Burm. f. (L.) Sol. ex Correa Fosberg var. retusa var. cannabina subsp. litiorale (Blume) A.Raynal subsp. tiliaceus var hispida (DC. ex Triana & Planch.) Killip Castor Oil Plant Sea Bean Coral Tree Siratro Sea Lettuce Sea Hearse Devil's Twine Box Fruit Cotton Tree Portia Tree Leucaena Pisonia Yellow Plum Stinking Passionflower Pigweed Spider Mangrove Pokok Jaru Jarak Kayu Dedap Kayu Kankong Kayu Jambu Hutan Kayu Besagi Kayu Keriting Pokok Waru Ampol Rukam 24 RUBIACEAE Guettarda speciosa i Kembang Melati Morinda citrifolia L. Cheesefruit Mengkud u *_Oldenlandia corymbosa LE: *___ Spermacoce assurgens Ruiz & Pav. RUTACEAE * — Triphasia trifolia (Burm.f.) ; Buah P.Wilson Kengkit SAPINDACEAE Allophylus cobbe L.) Blume Dodonaea viscosa Jacq. subsp. Hopbush viscosa SCROPHULARIACEAE *_ Scoparia dulcis Ee * Striga angustifolia (D.Don) Saldanha SOLANACEAE * ___Physalis minima E Chepelok * Solanum americanum Mill. Blackberry Nightshade SURIANACEAE Suriana maritima L. TILIACEAE Triumfetta repens (Blume) Merr. Bingil Burr & Rolfe URTICACEAE Laportea aestuans L.) Chew. VERBENACEAE *___ Clerodendrum indicum L.) Kuntze Clerodendrum inerme (L.) Gaertn. Sorcerer's Flower poe ehvia nodiflora L.) Greene Premna serratifolia EE * — Stachytarpheta jamaicensis (L.) J.Vahl Blue Snakeweed Vitex trifolia ARECACEAE Cocos nucifera Coconut Kelapa COMMELINACEAE *__Rhoeo spathacea Sw.) Stearn CYMODOCEACEAE Syringodium isoetifolium Asch.) Dand sea grass Thalassodendron ciliatum (Forssk.) sea grass Hartog Fimbristylis cymosa R.Br. Mariscus javanicus (Houtt.) Merr.& F.P.Metcalfe * — Pycreus polystachyos (Rottb.) P.Beauv. Queenslandiella hyalina (Vahl) F.Ballard HYDROCHARITACEAE Thalassia hemprichii (Ehrenb.) sea grass Asch. LILIACEAE Crinum asiaticum LE: Crinum Lil * — Zephyranthes rosea (Spreng.) Lindl. PANDANACEAE Pandanus tectorius Park. Var. Screw Palm Pandan cocosensis B.C.Stone | * — Bothriochloa bladhii (Retz.) S.T.Blake * Brachiaria brizantha (Hochst.ex A.Rich.)Stapf * Cenchrus ciliaris * Cenchrus echinatus Sand Burr * Chloris barbata Sw. D5 *_ Cynodon arcuatus J.Presl. & C.Presl. *__ Dactyloctenium aegyptium L.) Willd. * Eleusine indica (L.) Gaertn. Crowsfoot Grass * Eragrostis tenella (L.) P.Beauv. ex Roem. & Schult. * _Eriochloa meyeriana Nees) Pilg. * __Imperata cylindrica L.) P.Beauv. var. major Bladey Grass : Ischaemum muticum i Lepturopetium sp. aff. marshallense Lepturus repens (G.Forst.) Stalky Grass R.Br. * Panicum repens (ky Paspalum vaginatum Sw. * Sporobolus fertilis (Steud.) Sand Couch Clayton Stenotaphrum micranthum (Desv.) Beach Buffalo C.E.Hubb, Grass Thuarea involuta (G.Forst.) Bird's-beak R.Br.ex Roem. Grass & Schult. Zoysia matrella (L.) Merr. subsp. matrella *_ Unidentified sp. ‘ 26 Appendix 2. Notes on plant species of restricted distribution on the Cocos (Keeling) Islands, along with collection numbers held at CBG. Achryanthes aspera: Common on Keeling and found elsewhere only near some senescent Pisonia trees around the small lagoon on Pulu Luar. [D.G. Williams 45,52,211] Barringtonia asiatica: A solitary tree of great stature occurs on Pulu Panjang with numerous supressed-advance seedlings underneath the canopy. The only other Barringtonia seen were a few saplings in the recent strand forest along the lagoon shore 150 m south of the northeast point of Pulu Panjang. [D.G. Williams 110] Casuarina equisetifolia: Planted indiviuals are found in and near the settlements on Pulu Panjang and Pulu Selma and one large tree is on the lagoon shore north of the kampong. Guppy (1890) reported that the plant was introduced and spreading from island to island but no evidence was found for the latter. No seedlings were seen. [D.G. Williams 155] Cordia subcordata: Now occurs on the Cocos atoll only as large senescent individuals along the lagoon shore. No young plants were seen, although germination was common on North Keeling lagoon shore in April. [D.G. Williams 20,54] Enicostema axillare: Known only from two adjacent locations on southern Pulu Panjang, where it occurs amongst Zoysia matrella and Ipomoea pes-caprae in open coconut woodlands. It occurs from the ocean beach up to 100m inland. Although flowering freely, these populations appear to be extending largely by rhizome extension, to judge by their compact, circular distribution. [D.G. Williams 79] Erythrina variegata: A small but healthy grove of trees found at the north end of Keeling in Pisonia forest [D.G. Williams 53]. Likewise Allophylus cobbe [D.G. Williams 44,48] and Cleome gynandra [D.G. Williams 36] were found only in this area, the latter at the upper limit of the saltmarsh. Laportea aestuans: Previously collected here only by Darwin on the Cocos atoll in 1836. Since collected only on the beach top along the western shore of North Keeling island. [D.G. Williams 154] Lepturopetium sp.: A western Pacific genus of putative hybrid origin (Fosberg and Sachet 1982), found here only at the southern end of the runway on Pulu Panjang, growing on low-lying land occasionally inundated by rain or heavy seas. [D.G. Williams 267] Neisosperma oppositifolia: Found occuring as a stand only on Pulu Labu, where there are twenty or so mature trees forming abundant fruits. A solitary specimen without fruit was found on Pulu Atas and two apparently planted trees occur in the Pulu Panjang settlement. [D.G. Williams 25,145,175] Pandanus tectorius var. cocosensis: The only stands are on Pulu Selma where some of the clumps on high dunes have died out recently, possibly due to firing. A single clump on Pulu Panjang at the entrance to Telok Jambu appears to be all male, and therefore probably a single genet representing a solitary establishment event. [D.G. Williams 103] 27 Pisonia grandis: A few small clumps remain on the Cocos atoll of what must have been the dominant tree on the larger, higher islands before settlement. [D.G. Williams 21,43] Rhizophora apiculata: Occurs around the saline swamp on Pulu Luar and produces numerous seedlings there. One established seedling was found on the southern point of Pulu Selma but had disappeared a year later. Guppy (1890) stated that the populations were derived from beach drift planted on Pulu Luar by J. G. Clunies-Ross about 1850-60. [D.G. Williams 171] Suriana maritima: Occurs, in any abundance, only on recent sand deposits. Said by Guppy (1890) to have first colonized the atoll in about 1850, when it appeared on the ocean side of Pulu Cepelok, although it was not found there in this survey. [D.G. Williams 176] Ximenia americana: Found only as a few plants on the lagoon shore of Pulu Panjang. [D.G. Williams 183] 28 Appendix 3. Remnant native vegetation of the Cocos atoll referred to map units indicated by letters on Fig. 1. The map does not show the following types of native vegetation:- — Pemphis acidula and Suriana maritima shrublands on sheltered shores; — Scaevola taccada and Argusia argentea shrublands along exposed coastlines; — solitary individuals or small clumps of native species. Pulu Panjang (West Island) Major area of strand forest with single large Barringtonia asiatica, several Cordia subcordata, Calophyllum inophyllum, Hibiscus tiliaceus, Hernandia nymphaeifolia, and Morinda citrifolia. Strand vegetation of Pemphis acidula with the only stand of Pandanus tectorius on West Island (burnt in October 1987) and some Hibiscus tiliaceus and Hernandia nymphaeifolia. A small clump of mature Pisonia grandis. Scattered individuals of Hernandia nymphaeifolia occur in this area, most of which is cleared for the aerial field. Strand lined with patches of large Calophyllum inophyllum and freshwater swamp lined with Hibiscus tiliaceus. Suriana maritima, Guettarda speciosa and Pemphis acidula occur locally. Guettarda speciosa and Scaevola taccada scrub along the strand opposite the aerial field and merging northward with tall Guettarda speciosa and Calophyllum inophyllum strand forest which extends inland, indicating a former shoreline. Pulu Luar (Horsburgh Island) Well developed strand forest of Calophyllum inophyllum, Terminalia catappa, Dodonaea viscosa, Hibiscus tiliaceus, Premna serratifolia, Guettarda speciosa. Stand of Thespesia populnea and Cordia subcordata trees growing along a saltwater seep. Disturbed forest of Morinda citrifolia, Premna serratifolia, Guettarda speciosa, Terminalia catappa. Associated with a seasonally water-logged swamp dominated by Mariscus javanicus. Saltwater swamp with fringing Cordia subcordata, Rhizophora apiculata and a few Pisonia grandis. Achryanthes aspera and Sesuvium portulacustrum also occur. Pulu Tikus (Direction Island) A small clump of Thespesia populnea occurs here amongst Scaevola taccada at the top of a rubble beach. 29 Pulu Selma (Home Island) L Scattered clumps of Pandanus tectorius occur on the coastal dune; some burnt in 1987. Guettarda speciosa and Premna Serratifolia are also present. Pulu Labu M The interior of this island has several large Barringtonia asiatica and a number of small and large Neisosperma opposSitifolia. Pulu Atas (South Island) N Strand forest ranging from exposed to sheltered with Calophyllum inophyllum mainly, but also Hibiscus tiliaceus, Guettarda speciosa, and Premna serratifolia. ATOLL RESEARCH BULLETIN NO. 405 CHAPTER 7 AN UPDATE ON BIRDS OF THE COCOS (KEELING) ISLANDS BY T. STOKES ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 7 AN UPDATE ON BIRDS OF THE COCOS (KEELING) ISLANDS BY T. STOKES * INTRODUCTION The birds of the Cocos (Keeling) Islands were reviewed by Stokes et.al. (1984) and this paper summarises that paper and provides additional species records. It also provides an update of conservation comments provided in that paper. Forty-four species of bird (6 introduced) have been recorded from the Cocos (Keeling) Islands (Appendix). The following species notes are additional to those recorded by Stokes et. al. (1984). SPECIES Herald Petrel Pterodroma arminjoniana. A few were recorded on two occasions at North Keeling Island in April and June 1986 on the ground and in the air, suggesting nesting (Stokes and Goh 1987). Christmas Frigatebird Fregata andrewsi. An adult female was recorded on 21 March 1986 at North Keeling Island (Stokes and Goh 1987). Glossy Ibis Plegadis falcinellus. Four were recorded by a party of touring bird- watchers at the airstrip on 8-9 May 1990 (Richard Jordan and Peter Goh pers. comm.). Greater Flamingo Phoenicopterus ruber. A vagrant was recorded on North Keeling Island from April-June 1988 (Marchant and Higgins 1990). Australian Kestrel Falco cenchroides . A pair were recorded over several weeks at the West Island airport in May-June 1990 (P. Stevenson, pers. comm.). They are common on Christmas Island and may have originated from there. Buff-banded Rail Rallus philippensis andrewsi. An endemic endangered subspecies, formerly widespread on the main atoll and now virtually restricted to North Keeling Island where it is common (Stokes et. al. 1984). However occasional birds are still seen on the main atoll and the latest was a carcass, probably cat killed, found in 1991 at the West Island settlement (P. Stevenson, pers. comm.). Bridled Tern Sterna anaethetus. A specimen was collected on North Keeling Island by Gibson-Hill (1948) and there was a local report that it nested there. It has not been recorded since. * Great Barrier Reef Marine Park Authority, P.O. Box 1379, Townsville, QLD 4810 CONSERVATION When Charles Darwin visited the Cocos (Keeling) Islands in April 1836, the "immense number" of marine birds recorded in 1828-29 on the main atoll (cited Gibson- Hill 1949) were probably still present though somewhat diminished. He refers to trees on the island where he first landed on the main atoll, as being occupied by many nests of gannets (sic), frigatebirds and terns, and to a smell in the air which led him to call it a "sea rookery" (Darwin 1979). From his diary it would appear that the forest of the main atoll was by then well on the way to be being transformed into the monoculture coconut plantation that it was by 1885 when most main atoll birds had been eliminated (Forbes 1885). The reason for the decline in main atoll birds was almost certainly due to habitat change, intense hunting by people and predation by cats and rats . Today there are still very few birds on the main atoll. However birds remain in large numbers on North Keeling Island due to its isolation, the difficulty of landing, and access restrictions placed by the former Clunies- Ross clan rulers up to the mid 1970s. In the late 1970s and early 1980s, the lifting of access restrictions and the acquisition of more efficient boats and weapons by the Cocos- Malay people greatly increased the frequency and efficiency of bird-hunting on North Keeling Island. In the early 1980s the Australian Government became aware of concern on the islands and elsewhere that the number of seabirds being taken was not sustainable. I was despatched in January 1982 by the Australian National Parks and Wildlife Service (ANPWS) to investigate the situation and urgent control of hunting was recommended (Stokes et. al. 1984). In March 1986, the ANPWS Conservator on Christmas Island, about 900 km east of Cocos, was required also to provide conservation advice to the Cocos (Keeling) Islands Administrator. As incumbent at the time, I or my assistant (Peter Goh) flew to the Islands every 4-6 weeks to discuss conservation matters on the Islands and to survey North Keeling Island seabirds. By July 1986 agreement had been reached with the Cocos-Malay people : - to introduce a moratorium on seabird hunting on North Keeling Island pending seabird survey results, - to restrict any future seabird hunting to red-footed boobies, - to permit while the North Keeling moratorium was in place, certain numbers of red-footed boobies to be taken on Horsburgh Island on the main atoll (where small numbers come to roost in certain weather conditions at certain times of the year), - that hunters would try to avoid killing adult birds, and on - a series of administrative arrangements to regulate seabird hunting. The conclusion of the 1986 seabird hunting agreement brought considerable praise to the Cocos-Malay community from the Australian and international community. ANPWS created a permanent Conservator position on the Cocos (Keeling) Islands in December 1986. The moratorium on hunting at North Keeling Island eventually extended to December 1988 by which time surveys had revealed that about 34 000 pairs of red- footed boobies nested annually on the Island. Although hunting was prohibited on North Keeling between 1986-88 many illegal hunting trips probably occurred and one person was prosecuted. Hundreds and possibly thousands of red-footed boobies are believed to have been taken in the period. A cyclone in January 1989 caused considerable damage to the North Keeling Island vegetation and post-cyclone surveys suggested that more than 40% of red-footed booby chicks raised in the previous breeding season , and 1300 (or 3 1.9%) of breeding adults were killed (ANPWS 1989). In the subsequent breeding season there was a 60% reduction in the number of red-footed booby nests on North Keeling Island compared to the number at the peak of the best previous season in 1987 (ANPWS 1990). Although seabird poaching was reported to have declined in the 1989/90 with the purchase of a new patrol vessel, no seabird population surveys occurred in 1990 or 1991 due to lack of suitable transport (ANPWS 1990, J. Tranter pers. comm. ). Monthly surveys recommenced in the second half of 1992. A clear picture of the red-footed population status will not emerge until late 1993. However it appears that a cyclone in February 1992 caused sufficient damage to significantly reduce breeding success in the year (J. Tranter pers. comm.). The level of seabird poaching is reported to have been very high in 1990 and 1991 (J Tranter pers. comm. ). This was exarcerbated by a lack of adequate patrol vessels. However a new vessel was acquired in late 1992. An action plan to assist the recovery of the endangered rail population has been proposed by Garnett (1992). It includes enhanced access by management staff to North Keeling Island, a research program on the Island, declaration of the Island as a reserve, rat and cat control on main atoll islands and, if necessary, the re-introduction of rails to predator-free islands as conditions become suitable. Education to assist conservation is also proposed. Restrictions to ensure that North Keeling Island remains cat and rat free should also be considered. Since the mid-1980s ANPWS has developed a limited conservation education program in the Islands. This is being enhanced (J. Tranter pers. comm.) and should be accompanied by the declaration of North Keeling Island as a Nature Reserve, as recommended by Stokes et. al. in 1984. Seabird populations have low natural recruitment, and usually only inhabit and survive on islands free of predators. Although there may be a case for arguing that seabird hunting on the Cocos (Keeling) Islands is a legitimate part of the traditional Cocos-Malay culture, it should only continue if the take is sustainable and hunting is restricted to red-footed boobies on the main atoll under tight and enforcable controls. There will remain justifiable cause for conservation concern about seabird status in the Islands until this occurs. For this reason the International Council for Bird Preservation (ICBP) and other conservation agencies should continue to monitor the situation and seek rectification where necessary. ACKNOWLEDGEMENTS I thank Richard Jordan (New South Wales), Peter Goh (Christmas Island), Paul Stevenson (former Conservator, Cocos (Keeling) Islands), and Jeff Tranter (current Conservator, Cocos (Keeling) Islands) for assistance with this paper. REFERENCES Australian National Parks and Wildlife Service Annual Report 1988/89 Australian National Parks and Wildlife Service Annual Report 1989/90 Darwin, C. 1979. The journal of a voyage in HMS Beagle. Guildford (England). Genesis Publications. Forbes, H.O. 1885. A naturalists's wandering in the eastern archipelago. London: Sampson, Low, Marston, Searle and Rivington. Garnett, S. 1992. The Action Plan for Australian Birds. Canberra : Australian National Parks and Wildlife Service. Gibson-Hill, C.A. 1948. The Island of North Keeling. J. Malay Br. Royal Asiatic Soc. 21: 68-103. Gibson-Hill. C.A. 1949. The birds of the Cocos-Keeling Islands. Ibis 91 : 221-243. Marchant, S. and P. Higgins (Co-ordinators). 1990. Handbook of Australian, New Zealand and Antarctic Birds. Vol. 1. Melbourne: Oxford University Press. Stokes, T. and P. Goh. 1987. Records of Herald Petrels and the Christmas Frigatebird from North Keeling Island, Indian Ocean. Aust. Bird Watcher 132-133. Stokes, T., W. Shiels and K. Dunn. 1984. Birds of the Cocos (Keeling) Islands. The Emu 23-28. LIST OF BIRDS RECORDED FROM THE COCOS (KEELING) ISLANDS KEY TO SYMBOLS * = introduced, fenu= | resident. m = migratory, v = vagrant, e = nolonger occurs on the Islands , b = breeding, ? = unknown status. Herald Petrel Pterodroma arminjoniana (?) Wedge-tailed Shearwater Puffinus pacificus (mb) Red-footed Booby Sula sula (rb) Masked Booby Sula dactylatra (rb) Brown Booby Sula leucogaster (rb) Christmas Frigatebird Fregata andrewsi (v) Great Frigatebird Fregata minor (rb) Least Frigatebird Fregata ariel (rb) Red-tailed Tropicbird Phaethon rubricauda (rb) White-tailed Tropicbird Phaethon lepturus (rb) White-faced Heron Ardea novaehollandiae (r - possibly breeding) Cattle Egret Ardeola ibis (v) Eastern Reef Egret Egretta sacra (rb) Rufous Night Heron Nycticorax caledonicus (rb) Glossy Ibis Plegadis falcinellus (v) Greater Flamingo Phoenicopterus ruber (v) Unidentified hawk (v) Marsh Harrier Circus aeruginosus (v) Australian Kestrel Falco cenchroides (v) Feral Chicken Gallus gallus (rb) Guinea Fowl Numida meleagris (*rb) Buff-banded Rail Rallus philippensis andrewsi (rb-endemic subspecies) Lesser Golden Plover Pluvialis dominica (m) Ruddy Turnstone Arenaria interpres (m) Little Curlew Numenius minutus (m) Common Sanfdpiper Tringa hypoleucos (m) Greenshank Tringa nebularia (m) Pin-tailed Snipe Gallinago stenura (m) Sanderling Calidris alba (m) Oriental Pratincole Glareola maldivarum (m) White-winged Tern Chliudonias leucoptera (m) Sooty Tern Sterna fuscata (rb) Bridled Tern Sterna anaethetus (v b - specimen record of Gibson-Hill 1948, overlooked by Stokes et. al. 1984) Common Noddy Anous stolidus (rb) 6 White Tern Gygis alba Christmas Island Imperial-Pigeon Ducula whartoni Unidentified dove Unidentified nightjar Unidentified swift Barn Swallow Hirundo rustica Unidentified wagtail Christmas Island Thrush Turdus poliocephalus erythropleurus Christmas Island Whiteye Zosterops natalis Java Sparrow Padda oryzivora (*e) (*rb-restricted to Horsburgh and possibly West Islands) (*e) ATOLL RESEARCH BULLETIN NO. 406 CHAPTER 8 MARINE HABITATS OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 8 MARINE HABITATS OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS * ABSTRACT The marine environments of the main atoll of the Cocos (Keeling) Islands have been mapped at a field compilation scale of 1:25,000 scale using 1976 and 1987 aerial photography and field survey. Twenty two habitat units were recognized and mapped, with greater differentiation possible in the lagoon because of easier access and shallower depth. Qualitative descriptions of the habitats were made based on surface and SCUBA observations in the field. INTRODUCTION The Cocos (Keeling) Islands, located in the north-east Indian Ocean, comprise a main atoll (96° E; 12° S) called here the Cocos atoll and consisting of 25 vegetated coral cays, and a northern atoll 27 km away, called North Keeling Island, which is now a single island atoll. The nearest other land mass is Christmas Island, 900 km to the northeast and Java, 1200 km to the north. The islands were discovered (apparently not previously inhabited) in the early 17th century but were not settled until 1826 (Bunce 1988). Various naturalists have visited the islands, including Charles Darwin on H.M.S. Beagle in 1836 (see Armstrong 1991), just 10 years after settlement. Of the hundreds of atolls seen by Darwin during the voyage, this was the only atoll which he explored on foot. The experience was significant in the development of his coral reef theory (Woodroffe et al. 1990). Faunal studies of the corals, echinoderms, molluscs, crustacea and fishes were made by Gibson-Hill in the 1940's but, save for one survey by Maes (1967) on molluscs and Colin (1977) on fishes, there were no modern studies of the marine environments and organisms prior to those contained in the present volume (Williams 1990). Applied Ecology Research Group, University of Canberra, P.O. Box 1, Belconnen, ACT 2616 Despite the various studies mentioned, no consistent overview of the marine habitats was able to be achieved until aerial photographic coverage was available. In this context, and that of an increasing need to better manage the atoll's marine living resources, the objective of this study was to describe and delineate the distribution of the major marine habitat types for both the lagoon, reef flat and outer reef slopes of the Cocos atoll. Difficulty of access prevented sub-tidal survey on North Keeling. METHODS A preliminary photo-interpretation was done on panchromatic 1976 Royal Australian Air Force 1:44,400 complete coverage photography and this was used as a basis for field sampling. Location of sample sites attempted to include several examples of each of the photo-pattern units defined on the basis of tone, texture, bathymetry and location. The exact placement of field sites was able to be done with an accuracy of 100 m in most cases by reference to the photographs in the field. Over 50 sites were observed using SCUBA for depths over about 8 m and on snorkel or walking for shallower areas and observations recorded on a data sheet. Initially, fifteen different lagoon photo-patterns were described using the panchromatic photography. Further detail of the reef flats and of the outer reef was obtained from colour photographic coverage of the land areas only (1987 Australian Survey Office 1:10,000). These included the location of the "drop-off", (i.e. the top of a Pleistocene cliff at 18 m depth), the presence of sand chutes and deposits on the outer reef slope, spurs and grooves, reef crest surge channels, aligned coral flats, seagrass beds, conglomerate platform (Woodroffe et al. 1990) and beach rock deposits (Russell & McIntire 1965). Some further refinement of the map units was done from a classified SPOT image taken in May 1987. This imagery enabled more consistent delineation of some of the shallow water habitats but did not provide as much detail in the deeper lagoon and outer reef slope zones. Seagrass beds of two species in the deeper lagoon were not detected on the classified image, although no specific training was attempted. A description of each of the final mapping units is given below with cross referencing to map units (Table 1) and grid references on Fig. 1. The descriptions include habitat structure and dominant organisms as observed in the field, with other comments on structure or function within and between habitats. Reef classification and terminology varies considerably between authors and, where possible, the terms used here follow those of Hopley (1982). The marine areas of atolls and other reefs are commonly divided into three major zones. These are the seaward reef front which lies outside the line of breaking waves, the reef flat, which extends from the breakers to the shore or into the lagoon, and the lagoon itself. Each of these major units are further subdivided and described in the following scheme. RESULTS REEF FRONT The reef front (unit A) on Cocos is marked by a major slope change, usually at 15- 18 m depth, and below this the slope is greater than 45 degrees to depths over 50 m. Above the abyssal slope there is a gentle terrace of about 50 m to 2 km width, rising gradually to the reef crest at approximately the mean low water mark. On aspects more exposed to the south-east trades, the terrace has a well-developed spur and groove (buttress) morphology, which in several locations on the west side appears almost to form a secondary reef front due to very large buttresses located beyond the reef crest. The buttress systems are most pronounced on the southern reef. The reef terrace is the habitat of most abundant and diverse coral growth. In places the terrace has up to 60% cover, mainly of soft corals, whereas elsewhere there are sand deposits or very sparsely covered rock or sand slopes. Much of the wide terrace opposite the West Island settlement (grid reference 6251), for example, has a sparse coral cover on a hard basement, whereas the terrace further north (6256) is well covered in coral from the reef crest to the outer terrace. Because of the water depth over the terrace there were few mappable features. These were:- the break of slope at ca. 18 m, the presence and orientation of buttresses (spurs and grooves) and the presence of large sand deposits on the terrace. Terrace sand deposits are most abundant on the leeward side of the atoll. Opposite the northern half of West Island they form shore-parallel dunes at around 10 m depth (6257), usually lying between a buttress system and the coral-covered lower end of the terrace. Around Horsburgh Island sand is abundant and often covers much of the terrace and flows into deeper water via sand chutes (6561). THE REEF FLATS This component of the reef is also very varied and grades into the lagoon habitat in the channels between islands. The following units or features of the reef flat have been mapped:- — position of the coralgal crest and its surge channels; — sand and coral flats; - seagrass beds (Thalassia hemprichii); — aligned coral flats. Components of the flats which were not able to be mapped were boulder zones and algal flats. The former are especially well developed on parts of the southern inter-island reef flat where the combination of southeast wind and southwest swell produce the highest wave energy conditions. This is also reflected in the greater development of buttress systems and surge channels along the southern reef front. The structure of the coralgal crest also varies in relation to aspect. On the eastern and northern sides of the atoll it is poorly developed and very close (< 50 m) to the shore. Along the southern crest, calcareous red algae dominate and often form a honeycomb- like matrix that is exposed only during calm weather low tides. Along the western reef opposite West Island, the crest is dominated by seasonally varying stands of brown algae (mainly Turbinaria sp.), or a turf of red algae growing on smooth rock or algal-encrusted 4 pavements. The crest exposed at the lowest tides is usually less than 10 m wide and rises 30-50 cm above the adjacent reef flat. The reef crest has well-developed surge channels every 50-250 m, particularly along high-energy shores opposite islands. Associated with these channels are pockets of deeper water on the reef flat, these being possibly important in the movement of larger animals onto the reef flat, e.g. turtles, parrotfish and crayfish. The reef flats opposite islands (unit B) are generally zoned landward from the crest with a sparse coral zone, then an algal zone (usually brown algae), then a sand or rock zone and in places an inshore seagrass bed on trapped sediment. However, much variation exists and these zones are not always present nor clearly defined. Beds dominated by the seagrass Thalassia hemprichii (unit E) are developed on the inshore reef flat where sand has accumulated to about 5 cm or more depth and this usually occurs within 20 m of sandy beaches. In a few places sand has been trapped by exposed beach rock formations and this has enabled seagrass beds to develop. The coral zone of the reef flats (unit C) is well developed in only two localities — at the northern end of West Island (6257) and the south-western side of Horsburgh Island (6563). In both locations there are reef flats dominated by hard branching corals and this may be attributable to both sites being of moderately low wave energy as well as experiencing a constant flow of fresh oceanic water that is relatively less turbid. The aligned coral zone (unit D) is developed only on inter-island reef flats and usually extends from the crest into the lagoon to a point opposite the lagoon shores of the adjacent islands. The aligned corals are mostly hard massive species orientated parallel with the current. This is a faunally rich area dominated by clams, echinoderms and holothurians. At the lagoon end of the aligned coral zone and where the water is deeper than a metre, there is sometimes a well-developed area of outcrops composed of massive hard corals, surrounded by sand sheets. Beach rock deposits are common around the seaward beaches of Cocos (Keeling) and when exposed they provide a protective rampart to the coastline, which they often parallel. In some places they diverge from the shore and so indicate recent changes in shoreline position. THE LAGOON Moderately large by Pacific standards, the Cocos lagoon is distinctive for the high proportion of its area covered by what appear to be fields of collapse and/or solution dolines, commonly known as "blue holes" (units K, L & M). These depressions vary in their size, shape and depth of their surrounding patch reefs. Other major habitats mapped are the prograding sand sheets (unit F), opposite to and fed by the aligned coral zone; seagrass beds dominated mainly by Thalassia hemprichii (unit H) and with smaller areas of Syringodium isoetifolium (unit U) and Thalassodendron ciliatum (anit T); intertidal sand and mud flats (unit G); sand flats and shoals; coral flats and patch reefs. The continuous supply of sand from the outer and inner reef flat produces extensive sediment fans (unit F, Fig. 2), which are size proportional to the area of reef flat supplying them and the energy level of the coastline. The largest prograding sand sheets are the ones opposite the southern pass (between West and South Islands) and another fed by passes either side of Pulu Siput in the east. These deposits slope gently down into the lagoon and slowly engulf lagoon patch reefs and blue holes. Their margin with the aligned coral zone is not always distinct and it may be that the coral zone is extending slowly lagoonward across the sand sheet in places. In their most active places the sand sheets appear almost devoid of surface life other than algal crusts, but in more stable locations there is a sparse covering of Thalassia hemprichii, Halimeda (a sand-producing alga), Padina and Hydroclathrus (both brown algae). Areas of high bioturbation due to sea cucumbers and worms are also common. The spider shell, Lambis lambis, is locally abundant. Thalassia hemprichii dominates the seagrass beds (unit H, Fig. 3) developed from the Low Water Mark to depths of about one metre, close inshore to all the islands except Direction and Horsburgh. There is some variation in the structure of the beds with water depth and distance from the shore. Closer inshore, Thalassia is dominant but forms a discontinuous cover due to wave effects. Shallow intertidal areas are dominated by the seagrass and algae such as Gracilaria and Acanthophora, whilst subtidal areas have more coral and algal cover. In several places these seagrass beds (unit I, Fig. 3) show obvious lineation, and this represents overgrowth among the remnants of a prior aligned coral zone in places where a channel between islands has been closed off by storm activity. These sites are off South Island (7351, 7353) and West Island. The most conspicuous fauna of the inshore seagrass bed is the Portunid crab, Thalamita crenata, which makes shallow burrows, as does the much less common mud crab, Scylla serrata. The crabs appear to feed on burrowing bivalves and winkles, and numerous small fish shelter in the extensive crab burrow systems. Syringodium isoetifolium (units U, V) is a more restricted seagrass growing at depths from 1-6 m on sand in relatively clear water, mainly at the northern end of Home and West Island and in the bay of Direction Island (6962). It grows with Thalassia and various green algae including Caulerpa spp. in the former shallow sites, but is almost mono-specific dominant at Direction Island in 2-6 m of water. Thalassodendron ciliatum, a large, robust seagrass, grows attached to rubble and rock. There is one large bed (unit T) of this species in the central northern lagoon in 8 m of water on sandy rubble, and smaller patches of this species possibly occur on the outer reef south-east of Horsburgh Island. The species grows in the lagoon in large circular clumps, some of which appear to be showing central die-off and hence taking on an annular shape. Small hard coral outcrops occur within some of these beds. All three seagrasses appear to be grazed by green turtles, which are regularly seen along the northeast shore of West Island and in Direction Island bay. None were seen near the Thalassodendron bed but it did show signs of grazing by large animals — possibly turtles. Shoreward of the Thalassia beds along the southern and eastern sides of the lagoon there are extensive intertidal sand and mud flats (unit G, Fig. 3). These are often demarcated from the seagrass by a sandbank which is usually colonized by a line of the shrub Pemphis acidula, which thereby stabilizes the bank and forms a mangrove-like vegetation. It is rather surprising that no mangroves have colonized these flats since such habitats are generally occupied by them on other reefs around the world. The finer sediments tend to accumulate in the more sheltered inlets of the lagoon side of the islands and generally where current and wave action is minimal. These form the major habitat 6 for the fiddler crab Uca chlorophthalmus and Macrophthalmus verreauxi, which occur here in high numbers. Major subtidal sand flats (unit Q) which are not directly associated with passes between islands occur northeast of West Island and south of Horsburgh Island. These locations are where sand is accumulated on the leeward lagoon margin and where it may eventually be carried to the outer reef. At West Island the sand flat has a highly mobile offshore sand bank (unit R) and there is major sand accretion onshore. The sand flat south of Horsburgh Island lies in a westerly current flow and is connected to the outer reef slope by a major sand chute which continues across the reef terrace. Much of the central lagoon is a mosaic of "blue holes" of varying sizes and shapes and depth of surrounding patch reefs. Smaller, more discrete holes (unit K) occur towards the shallows and these ones are, in some cases, being filled by prograding sand sheets. Around the fringe of each depression there is a band of staghorn coral rubble which extends down the sides of the hole for about 2-5 m before passing into a sand slope going to as deep as 20 m. The water in the smaller holes is usually fairly turbid and the bottom is composed of sand and finer sediments. In a few places large clumps of the foliose hard coral, Echinopora, were dominant on the edges of holes. These seemed to be more common in the small patch reefs developed in the middle of large holes (units L, M, Fig. 2) where turbidity was low. Away from the fringes of holes the coral cover is usually sparse and consists of the branching Acropora and outcrops of massive corals such as Porites. Bivalve molluscs and holothurians are often abundant and there is, in places, a high degree of bioturbation, perhaps associated with areas of finer sediments. At least two species of mushroom corals also occur around these patch reefs. Although several different blue hole morphologies have been mapped, sampling has not been sufficiently intense to determine whether there are significant biological differences. The main difference evident on photography is between small, iso-diametric holes which cover 10-30% of the area, through to the holes in deeper water which are large, irregular and occupy more than 80% of the area. These sub-units tend to be separated by east-west lines of shallow water (of unit J), these perhaps representing former beachlines. The deeper central lagoon opposite Home Island appears to show a broad pattern of reticulate reefs (unit N) but without deep blue holes. Much of the bottom cover is very thick staghorn rubble, tightly interlaced and bonded by overgrowth of algae, particularly a Padina sp. There are occasional sandy patches, as well as large mostly dead outcrops of Porites spp. North of the previously described habitat is an extensive area (unit O) composed of mainly dead outcrops of massive corals with intervening sand patches. There are large areas of included sand flat on the eastern side and generally very small amounts of hard coral regrowth. Well-developed coral flats lie between West Island and Horsburgh Island and, since there is no reef crest here, these lagoon flats at a depth of around 8 m merge into the outer reef terrace. Between the seagrass beds on the east and west sides of the lagoon and the deeper habitats of the central lagoon, there is a large and variable shallow-water habitat (unit J) 7 with various mixtures of sand, algal and coral cover. The sandy substrate, which appears to be rather thin, is often strewn with fine coral and shell fragments and there may be a sparse cover of Thalassia and/or Caulerpa spp. Patches of very dense Halimeda cover occur and these are often overgrowing small coral boulders. A small black sea cucumber is very abundant here. Large skeins of the net-like alga Hydroclathrus drift along and accumulate on obstructions. Colonies of Porites occur, some forming microatolls of considerable size. Acknowledgements This project was completed in 1987 whilst the author was appointed as Environmental Resource Adviser to the Territory of Cocos (Keeling) Islands. I am grateful to the Cocos (Keeling) Islands Council for its interest in the project and for allowing access to all parts of the Territory; to the Administration of the Territory for providing field logistic support and to the members of the Cocos Dive Club and Amat Noor bin Anthoney for assistance in the field surveys. The Australian National Parks and Wildlife Service provided access to the classified SPOT satellite image. REFERENCES Armstrong, P. 1991. Under the Blue Vault of Heaven: A study of Charles Darwin's sojourn in the Cocos (Keeling) Islands. Indian Ocean Centre for Peace Studies, Nedlands, Western Australia. Bunce, P. 1988. The Cocos (Keeling) Islands. Australian Atolls in the Indian Ocean. Jacaranda, Milton, Queensland. Colin, P.L. 1977. The reefs of Cocos-Keeling atoll, eastern Indian Ocean. Proceedings, Third International Coral Reef Symposium. University of Miami, Florida. Hopley, D. 1982. The Geomorphology of the Great Barrier Reef: Quaternary Development of Coral Reefs. J. Wiley & Sons, New York. Maes, V. 1967. The littoral marine molluscs of Cocos-Keeling Islands (Indian Ocean). Proceedings of the Academy of Natural Science of Philadelphia 119: 93-217. Russell, R.J. and McIntire, W.G. 1965. Southern hemisphere beach rock. Geographical Review 55: 17-45. Williams, D.G. 1990. An annotated bibliography of the natural history of the Cocos (Keeling) Islands, Indian Ocean. Atoll Research Bulletin 331: 1-17. Woodroffe, C., McLean, R. & Wallensky, E. 1990. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. National Geographic Research 6: 262-275. Table 1. Marine Habitats of the Cocos (Keeling) Islands Main Atoll and their corresponding map units. A stipple areas indicate sand deposits hatching indicates the approximate length, width and direction of major buttresses outer solid line indicates lower depth limit of the outer terrace REEF FLAT | Coral and Algal Flat |e includes areas of platform rock, beach rock and beaches e reef crest (coralgal rim) shown as a dashed line e surge channels shown as an arrowhead Aligned Coral Flat - small massive corals dominant Seagrass Flat (Thalassia hemprichii) LAGOON | Prograding Sand Sheet Intertidal Sand and Silt Flat e lines of Pemphis acidula on offshore sandbanks shown stiooled Seagrass Sand and Silt Flat (Thalassia hemprichii) Seagrass Flat - on prior aligned coral flat (Thalassia hemprichil Coral and Algal Flat with sparse Thalassia hemprichii Blue Hole Mosaic 1 - small scattered isodiametric depressions Blue Hole Mosaic 2 - large, irregular depressions, prominent shallow coral rims Blue Hole Mosaic 3 - large, irregular depressions, deep rims Algal-covered Staghorn Rubble with occasional sandy patches Massive Coral Outcrops with sandy patches Emergent Reef (Turk's Reef) Sandy Lagoon Floor with occasional Coral Outcrops Sand Shoal Sandy Lagoon Floor with scattered Coral Outcrop and Seagrass Beds (Thalassodendron ciliatum) Seagrass Bed (dense Thalassodenaron ciliatum) and occasional coral outcrops Seagrass Bed (Syringodium isoetifolium) and occasional coral outcrops Mixed Seagrass & Algal Bed (Syringodium isoetifolium, Thalassia hemprichii, Halimeda spp., Caulerpa spp.) ali Sail a) - ale | crit sv Ber ers P| Mice | tai T aT 65 __\ AN ex |__HORSBURGH\ \A INDIAN |OCEAN ISLAND AY Ga (PULU LUAR) Ses 63 S B Mi a i = | ek a ~ AF ~|-\DIRECTION] ISLAND i of \\\,. ) (PULU TIKUS) 62 mils 2 (aa 2 5 : Bi: a i / FU, 1 al ay \ f ce) Vv \A 60 Bs came T - 7 i : << HOME| ISLAND 59 i s : 2 Md p/P ‘ Be Pall [ H \ DAR * Vv Q ie) N F \ L J NO 57 XS M H oS \ Z SS 56 = : a0 A D fe, ai ‘ DY L - A ( 55 a= p \ N -\ \E ‘\ pa G fl ¥ 54 2a +} S A 4 J H \e ey J 1 \ 53 fs sN 2 Gi K | | it \ B G re Z S ZAB fA K 52 — ws -# aS "ss : 4 as NS | Ue 51 » H SG is WEST ISLAND \°k \ D is (PULLU PANJANG) \\ *\ e < 50 2 ss . casey SOUTH = ha ISLAND “9 LS (PULU ATAS) | | 62 63 64 65 66 67 68 69 70 7\ 72 73 74 75 Fig. 1. Marine Habitats of the Cocos (Keeling) Main Atoll. Numbered grid lines are 1000 m intervals of the Universal Transverse Mercator Grid, Zone 47. Habitats mapped from field survey by D.G. Williams in 1986/7, using R.A.A.F. 1976 panchromatic (1:44,400) and A.S.O. 1987 colour (1:10,000) aerial photography. Map field compiled at 1:25,000 on base derived from R.A.S.C. Series R811 Sheet Special Cocos Island 1979. 10 Figure 2. Aerial oblique view over the southern lagoon, looking east towards the Figure 3. Aerial oblique view of the southern lagoon, looking east across Pulu Atas, showing seagrass beds (Unit H) and lines of Pemphis acidula at the edge of sand and silt flats. ATOLL RESEARCH BULLETIN NO. 407 CHAPTER 9 SEDIMENT FACIES OF THE COCOS (KEELING) ISLANDS LAGOON BY S.G. SMITHERS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 9 SEDIMENT FACIES OF THE COCOS (KEELING) ISLANDS LAGOON BY S.G. SMITHERS * ABSTRACT Surficial sediments from the Cocos (Keeling) Islands lagoon were classified according to texture and composition using factor analysis. Six main textural facies: 1) slightly gravelly coarse sands, 11) slightly gravelly medium sands, iii) gravelly sands, iv) sandy gravels, v) gravelly muds, vi) slightly gravelly fine sands; and three main compositional facies (i) coral-type sediments, ii) molluscan mud sediments, i11) coralline algae/Halimeda type sediments were identified, accounting for over 90% of sediment variation in the lagoon. These facies can be related to the provenance of constituent components and lagoonal hydrodynamics. INTRODUCTION The main atoll of the Cocos (Keeling) Islands (96948'-56'E; 12°04'S) consists of a horse-shoe shaped reef rim, on which 26 reef islands lie, surrounding a central lagoon of approximately 190 km2. The lagoon can be divided into two broad provinces, the deeper (8-15 m) northern basin and the shallower southern flats (0-3 m) (Fig.1). Blue holes exceeding 20m depth occur in several parts of the lagoon, but are most obvious across the shallower southern flats. A more detailed description of lagoonal marine habitats is provided by Williams (this volume), and they are mapped in Figure. 2. At the north of the atoll, deep and wide passages either side of Horsburgh Island connect the lagoon to the open ocean. Other exchange between the lagoon and ocean is restricted to 11 shallow reef flat passages situated on the eastern and southern atoll rim. Currents through these reef passages are predominantly unidirectional into the lagoon, probably driven by the persistent southeast trade winds which prevail for most of the year, and wave set-up generated by the swells which continually break over the windward reef crest. The hydrodynamics of this atoll have been examined recently by Kench (this volume). Lagoonal infilling by sediments produced on the reef rim is generally accepted as the dominant constructional process on atolls after the reef rim has reached a stable sea level (Marshall and Davies 1982, Frith 1983, Tudhope 1989). Upward growth of the reef rim has been limited by sea level for more than 2000 years on the Cocos (Keeling) Islands (Woodroffe et al. 1990a, 1990b, this volume), and historical accounts (Darwin 1842, Guppy 1889) indicate that much of the southern part of the lagoon has been rapidly infilled. During his visit in 1836 Darwin sailed to the south of the lagoon through channels dredged through living coral. Sand sheets or seagrass meadows which are often exposed at low tides now cover these areas. Vibrocore data were used to establish the nature and chronology of longer term (mid-late Holocene) accretion in several parts of the lagoon (Smithers et al. in press). id Department of Geography, University of Wollongong, Northfields Avenue, Wollongong, New South Wales, 25272. Sediments infilling atoll lagoons consist almost entirely of skeletal carbonate secreted by reef organisms, and facies development within atoll lagoons is governed by interaction between the supply and physical properties of the source material and the various processes which degrade, redistribute and stabilize sediments (Maxwell et al. 1964, Milliman 1974). Biogenic carbonates may consist of either rigid reef framework or unconsolidated detrital material, and may be produced and deposited in situ (autochthonous) or produced outside the lagoon and transported in before deposition (allochthonous). The relative contributions of allochthonous and autochthonous sediments usually varies around a lagoon, and can be determined from the texture and provenance of contributory components (Swinchatt 1965, Orme 1973). Reconciling the habitat zone of the source organism with the location of the depositional zone, and the determination of textural gradients between sediment sources and sinks, allows hydrodynamic, sediment transport, and facies development processes to be inferred. There have been few studies of the lagoonal sediments of Indian Ocean atolls and the lagoon of the Cocos (Keeling) Islands differs in several ways from other atolls where sedimentation has been examined. Firstly, the sediment producing biota of the Cocos (Keeling) Islands appear to differ from other atolls, possibly due to its extreme isolation. Secondly, the size, bathymetry and hydrodynamics of this lagoon differ from other atolls where carbonate sedimentation has been investigated. Early studies chiefly examined sediments from the relatively large and deep Pacific atolls with lagoons which deepen towards the centre (e.g. Kapingamarangi - McKee et al. 1959, Bikini, Rongelap, Enewetok - Emery et al. 1954). Smaller atoll lagoons with complex bathymetry have more recently received some attention (Mataiva, Takapoto - Adjas et al. 1990, Henderson Reef - Chevillon and Clavier 1990, Mataiva - Desalle et al. 1985), but once again are concentrated in the Pacific. This paper reports on an examination of the surficial lagoonal sediments of the Cocos (Keeling) Islands, a moderately sized Indian Ocean atoll with a complex lagoonal bathymetry. The primary aims were to: (1) determine the textural and compositional characteristics of lagoon surface sediments; (2) identify and map textural and compositional facies; and (3) relate facies distribution to specific biotic/physiographic environments. METHODS AND MATERIALS The lagoon floor was examined on a series of boat and snorkel transverses. A total of 167 sediment samples were collected from the lagoon bed (Fig. 3), using a sampling strategy based on environments determined from aerial photographs and SPOT satellite imagery. Not surprisingly, the lagoonal environments delineated in this way are very similar to the marine habitat units established by Williams (this volume). Samples were collected by scooping unconsolidated sediments into plastic bags except in depths that exceeded 8m when a weighted steel dredge was used. Seventy-six sediment samples were analysed granulometrically using the techniques of Folk (1974), making sure that several samples from each lagoonal environment were examined. Where necessary the mud fraction was first separated by wet sieving; these samples were washed with 200 ml of distilled water and approximately 1ml of 10% Calgon for each gram of estimated mud content. The mixture was left to stand overnight then mechanically stirred for 3 minutes and washed through a 4¢ sieve. Sediments larger than 49 were dried, weighed and between 50-70grams transferred to a nest of sieves ranging from -2¢ to 4g, with a 0.5¢g interval. The sieves were mechanically shaken for 15 minutes and the fraction retained on each sieve (and the pan) weighed. Mean grain size, sorting and skewness were determined using the graphic methods of Folk and Ward (1957). The skeletal compositions of 50 sediment samples were examined. Representative subsamples were taken from sieve fractions greater that 3.5¢ and grains were identified and point-counted using a binocular microscope. Approximately 100 grains were identified for each sieve fraction. Fifteen component categories were recognized: (i) coral shingle and grit; ii) Halimeda fragments; 111) coralline algae (principally Spongites rhodolith fragments); iv) Homotrema; v) gastropod fragments; vi) pelecypod fragments; vii) unknown molluscan fragments; viii) Marginopora; ix) Amphistegina; x) other foraminiferans; xi) echinoids; xii) annelids; xiii) alcyonarian spicules; xiv) crustacean fragments; and xv) indeterminate or unrecognisable grains. Component representation in the total sample was expressed as a weight percentage of the total sample. Granulometric and compositional data were analysed using Q-mode factor analysis (Klovan 1966, 1975, Gabrie and Montaggioni 1982, Montaggioni et al. 1986) in order to classify sediments according to their compositional and textural characteristics (Smithers 1990). RESULTS SEDIMENT TEXTURE The textural characteristics of seventy-six sediment samples from the Cocos (Keeling) Islands lagoon are presented in Table 1 and descriptive statistics for each of the lagoonal environments provided in Table 2. These results indicate that the Cocos (Keeling) Islands lagoon is dominated by poorly sorted, slightly gravelly (<10%) fine to coarse sands. Several general trends in sediment texture can be identified. Mean grain size is greatest in the interisland channels and is lowest in the seagrass meadows and intertidal sand and mud flat areas. Gravel abundance appears closely related to coral outcrop proximity, the highest mean values occuring in the interisland channels where autochthonous gravels are deposited with allochthonous gravels transported from the reef flats, and in the blue hole mosaic where gravels derived from patch reefs are common. Occasional high gravel values in samples collected elsewhere in the lagoon can largely be attributed to the deposition of autochthonous carbonates. Mud content peaks at around 45% in the seagrass meadows but generally comprises 0-2% of most sediment samples. Sorting is typically poor, but improves in the exposed sandy areas in the north of the lagoon. Skewness values range from strongly fine skewed to strongly coarse skewed, and in different areas may reflect either in situ sediment production or else the selective removal of certain grain sizes by incident currents. The significance of variation in the textural traits of lagoon sediments will be addressed in the discussion. Six factors were extracted from the data matrix of 14 variables (weight % of sediment in each sieve fraction) and 76 observations (sediment samples) using a Q-mode factor analysis which can account for 91.7% of the data variance. Communality values are high for all samples indicating that a good description of all samples is given by these factors. Sediment samples were classified according to the factor axis each was most heavily loaded upon; samples belonging to each class are listed in Table 3. The grain size distributions of samples with the highest loading on each factor axis are presented in Figure 4 and the average textural statistics of sediments assigned to each factor are given in Table 4. Descriptions of the textural sediment types classified on each factor axis are provided below and their distribution is shown in Figure 5: Factor One - Slightly Gravelly Coarse Sands. These sediments account for over one third of the samples and are chiefly composed of coarse sands with a minor gravel component (Fig. 4a and Tables 1 and 2). The grain size distribution is characterised by a primary mode in the 0¢-0.5¢ range and the mean grain size is around 0.5g. Muds usually form less than 1% of these sediments. Sorting and skewness are variable; ranging from moderately well to poorly sorted and from strongly fine to strongly coarse skewed. Slightly gravelly coarse sands occur throughout the lagoon, but appear to be most concentrated in the exposed areas of the deeper northern part of the lagoon and around the interisland channels. Factor Two - Slightly Gravelly Medium Sands. These sediments are very similar to those defined by factor one, however the sand fraction is finer with the principle mode being between 1.59g-29 (Fig. 4b). Sediments represented by this factor range from moderately well to poorly sorted and show a tendency to be coarse skewed. These sediments cover much of the lagoon floor, being patchily interspersed with the slightly gravelly coarse sands in the north of the lagoon and covering large areas north of the southern passage. Factor Three - Sandy Gravels. High gravel content is the definitive trait of these sediments, with the grain size distributions peaking in the >-2¢ interval (Fig. 4c). They are typically finely or very finely skewed and range from poorly to moderately sorted. A second, smaller modal peak may occur in the sand sized range. Sandy gravel patches are sporadically distributed within the lagoon, with three distinct patches located in the centre of the lagoon and another occuring south of Horsburgh Island. Smaller pockets of sandy gravel are located just north of both Pulu Maria and the seagrass meadows behind South Island. Factor Four - Gravelly Sands. These sediments are composed principally of sands, but also have a moderate gravel content (Fig. 4d). Grain size distributions are often bimodal, reflecting the poor sorting and variable skewness of most of these sediments. Gravelly sands are also patchily distributed over the lagoon, with a distinct band located lagoonward of the islands on the eastern rim. Several smaller patches occur towards the lagoon centre. Factor Five - Gravelly Muds. Abundant fine sands and muds characterize these sediments, although gravels are also moderately well represented (Fig. 4e). Sorting, therefore, is typically poor and most grain size distributions coarsely skewed. Gravelly muds occur in the lee of the windward islands and in the shallow embayments locally known as Teloks. Factor Six - Slightly Gravelly Fine Sands. Fine sands in the 2.5¢-3.0¢ range dominate these sediments. The fine sands may grade into muds in some samples and they are usually coarse skewed and poorly to moderately sorted (Fig. 4f). Patches of slightly gravelly fine sands are found throughout the lagoon, however they are more common in the north central areas. SEDIMENT COMPOSITION The skeletal compositions of 50 samples collected from the Cocos (Keeling) Islands lagoon are listed in Table 5 and the average composition of sediments deposited in each lagoonal environment presented in Table 6. It is evident from this table that the abundance of skeletal constituents may vary markedly between different lagoonal environments. Furthermore, relatively large standard deviation values suggest that sediment composition may also vary markedly within lagoonal environments. Nevertheless, several general statements can be made about the composition of sediments deposited within this lagoon. Coral debris clearly dominates most samples (range: 81.46% in sample 12 to 11.05% in sample 58), comprising the major identifiable component in all lagoon environments (see Table 6). Halimeda and coralline algae also contribute significantly to many samples (Halimeda >15% of samples 24, 29, 45, 48, 49, 58, 124, 171; coralline algae >15% of samples 6, 34, 58, 66, 125, 164, 165, and vibrocore cv15), particularly those collected where hard coral substrates exist, such as the blue hole mosaic and the interisland channels. Coralline algae may either encrust other constituents or consist of rhodolith debris, the later being spherical coralline algae colonies which are particularly abundant in the high energy interisland channels. The Acropora shingle which is widespread over the central lagoon floor is also heavily encrusted with coralline algae and represents a potential source of this material. Homotrema is a minor contributor to lagoonal sediments (range: 2.48% in sample 24 to 0% in many samples) but appears most abundant close to high energy, hard substrate environments. Gastropod detritus comprises around 5% of the sediment in most lagoonal environments, rising to an average of over 10% in the intertidal sand and mud flat areas, and accounting for more than 10% of some samples from the seagrass meadows (117, 122). Pelecypods comprise less than 5% of most samples, but contribute 9.8% and 9.35% of samples 108 and 38 respectively. Marginopora tests make up 0-4% of most samples with no clear pattern to their distribution being immediately apparent. Amphistegina is a widespread but locally significant component, being most prolific on the reefs south of Horsburgh Island and in the sandy lagoon floor region in the north of the lagoon. Annelida, alcyonarian spicules, crustacean debris and echiniod spines are generally present in small quantities. Crustacean detritus can, however, occasionally be quite high in areas where living crustaceans are plentiful (i.e. sample 39 from Telok Jambu - 7.85%). Alcyonarian spicules represent only a small proportion of most sediments (range: 6% in sample 50 to 0% in many) but appear most abundant in samples just lagoonward of the reef rim. Indeterminate sediments include sediments <3.5¢ and those not readily recognisable because of corrosion. As outlined in the textural results, the abundance of fine sediments is greatest in the seagrass meadows and intertidal sand and mud flat areas. The -0.5¢ fraction of a sheltered seagrass meadow, interisland channel, interisland channel/ sand apron and central lagoon sample is presented in Figure 10a-d. Three factors were extracted from the data matrix covering 15 component variables and 50 sediment samples. All samples except 153 have high communality values, suggesting that a good description of most samples is given by these factors. The lower value for sample 153 probably reflects the exceptionally high representation of Amphistegina in this sample, this being more than five times greater than in the sample with the next highest representation. Samples were classified according to the axis upon which they were most heavily loaded except where samples had similar loadings on more than one axis. Loadings were considered similar if the absolute difference between loadings on different axes was less than a third of the larger loading value, and where this occurred samples were deemed to be hybrids. Samples belonging to each class defined by the factor analysis are listed in Table 7 and pie charts showing the composition of the sample most heavily loaded on each factor axis are presented in Figure 6. These and the average compositional facies statistics are presented in Table 8. Descriptions of the compositional sediment types discriminated by the factor analysis are provided below and their distribution is presented in Figure 7. Factor One: Coral-Type Sediments. Coral-type sediments are chiefly characterised by the compositional dominance of the sample by coral debris. More than 60% of samples collected from the Cocos (Keeling) Islands lagoon are classified as coral-type sediments, conforming with the preponderance of coral evident in the raw compositional data. Skeletal material derived from organisms commonly associated with hard coral substrates (i.e. Homotrema, Amphistegina, annelids and alcyonarian spicules) also reach their highest representation in this facies. Most of the lagoon bed is covered by sediments most adequately described as coral-type, the main exceptions being the areas in the lee of the windward islands. Factor Two: Molluscan Mud Sediments. A large indeterminate component is characteristic of these sediments and they contain a noticeably smaller quantity of recognisable coral debris than the coral-type sediments depicted by factor one. Gastropod debris is also found in these sediments in moderate amounts, reaching its highest representation in this facies. Crustacean debris is also significantly more abundant in these sediments than in any of the other facies, and Marginopora is most prolific in these sediments. Molluscan mud sediments are predominately restricted to the shallow protected parts of the lagoon, however there are outlying patches in the north and central lagoon. Factor Three Coralline Algae/Halimeda Type Sediments. These sediments are essentially differentiated because they contain a relatively high proportion of coralline algae and Halimeda and a relatively low proportion of coral debris. Abundant rhodolith debris determines that sediments deposited in the lee of Pulu Maria and Pulu Siput are most heavily weighted on this factor, whilst Halimeda debris is responsible for sediments on the edge of the seagrass meadows in the lee of South Island being loaded on the third factor axis. DISCUSSION The nature and distribution of sedimentary facies in the Cocos (Keeling) Islands lagoon essentially reflects the interaction of wave and current energy on skeletal sediments derived from a range of organisms growing in different lagoon environments. The reef islands and a discontinous reef rim control the distribution of wave and current energy within the lagoon; directly controlling the entry and distribution of allochthonous sediment, indirectly controlling the distribution of autochthonous sediments by influencing biotic zonation, and controlling the redistribution of sediments within the lagoon. Three main features characterise the sedimentary facies of the Cocos (Keeling) Islands lagoon, these being: 1) the domination of the lagoon by coral derived sediments; 2) sediment sorting in areas of relatively high hydrodynamic energy and the deposition of predominantly poorly sorted sands and gravels in the centre of the lagoon; and 3) the concentration of mud deposits in the lee of the windward islands, almost exclusively in the seagrass and intertidal sand and mud flat environments. The predominance of coral derived sediments and subsequent coverage of most of the lagoon by the coral-type compositional facies is a striking feature of the Cocos (Keeling) Islands lagoon (Fig. 7), which is even more remarkable considering the dearth of living coral presently on this atoll. Compared to other carbonate lagoons coral components comprise an inordinate proportion of the sediments deposited in this lagoon (Fig. 8). Several possible reasons exist for the high representation of coral sediments in this lagoon, including its relatively small size and shallow nature. Milliman (1974) suggested that because the ratio of lagoonal area to reef rim becomes smaller as atoll size declines smaller lagoons are more likely to receive a higher proportion of reef flat sediments, including a substantial proportion of coral material. Alternatively, because much of the Cocos (Keeling) Islands lagoon is less than 10 m deep, a depth range dominated by corals in many reef environments (Emery et al. 1954 Stoddart 1969, Milliman 1974), it is perhaps not surprising that coral sediments are abundant here. Indeed, coral outcrops are common throughout most of the Cocos (Keeling) Islands lagoon, imparting a reefal character on most lagoonal sediments. Moreover, lack of net bathymetric relief has restricted the habitat potential of this lagoon and many components and facies derived from organisms normally found in deeper water are poorly represented here (e.g. the deep water Halimeda facies reported from deeper lagoons like Suwarrow (Tudhope et al. 1985), Kapingamarangi (McKee et al. 1959) and Enewetok (Emery et al. 1954)). The geomorphic history of the atoll may provide another explanation for the abundance of coral derived sediments in this lagoon. Woodroffe et al (1990a, 1990b, this volume) have established that approximately 3000 years ago sea level on this atoll was close to 1m higher than present, and that at this time a sea-level limited reef flat encircled much of the lagoon. Sea-level has subsequently fallen to its present level and most of this higher reef flat has been substantially eroded, remnants existing as the contemporary conglomerate platform. Clearly the erosion of this fossil reef flat comprises a potentially significant source of coral sediments which may have been transported around the atoll under different physiographic conditions as the atoll has developed. Prior to the consolidation of the larger islands (particularly South Island), for example, coral sediments were presumably transported into the lagoon through more numerous interisland SHAME S and could potentially achieve a more widespread coverage of the lagoon. Although coral-type sediments veneer most of the lagoon (Fig. 7), specific areas are covered by sediments which are more or less coral-type than others (i.e. are more or less heavily loaded on the first factor axis due to variations in the abundance of coral and other components), and textural parameters delineate two distinct source zones, the reef rim and the lagoon. Deposits formed by allochthonous material transported from the reef rim are typically most strongly defined as coral-type sediments and those composed of autochthonous material produced within the lagoon less so, reflecting a change from a strongly reefal component assemblage (i.e. coral, coralline algae, Homotrema, alcyonarian spicules, Amphistegina) to a moderately lagoonal one (reefal components less well represented, fine indeterminate sediments more abundant) (Table 6). The sandy lagoon floor region is exposed to high levels of wave and current action due to the discontinuous nature of the reef rim at the north of the atoll and the textural traits of sediments deposited there reflect this position. Extensively rippled coarse sands which are near symmetrically skewed and well sorted dominate this area, interupted sporadically by localised seagrass patches and small coral bommies. Sediments deposited here are texturally mature; reflecting the relatively high levels of wave and current energy affecting this area and the rarity of locally generated gravels. Speculation of a peripheral reef source for these sediments is supported by high Homotrema, Amphistegina and alcyonarian spicule content; these components normally originating from high energy reef zones and confering a strong coral-type classification on these sediments. Ripple orientation suggests that most of this material is transported from the northeast reef rim. Unlike in much of the southern part of the lagoon these sediments remain submerged at all tidal stages and are continually affected by waves and currents, enhancing their sorting potential. Similarly well sorted and rippled sands are described from the Alacran Reef Complex, Mexico (Kornicker and Boyd 1962) and the lagoon of Enewetok atoll (Wardlaw et al. 1991) and are thought to have developed under similar environmental conditions. Interisland channels link the high energy and highly productive outer reef flats to the lagoon along the eastern and southern atoll margin and act as a conduit for hydrodynamic energy and sediments entering the lagoon. Waves and currents forced over the windward reefs are concentrated through these channels developing relatively high levels of hydrodynamic energy which dissipates into the lagoon. The composition and texture of sediments deposited through these channels is distinctly reefal, consisting of sands and gravels derived from organisms typically located on high energy reefs such as coral, coralline algae, alcyonarian spicules, Homotrema and Amphistegina. Not suprisingly these sediments are unequivocally coral-type in composition. The relatively high levels of hydrodynamic energy which affect these channels is reflected by the mean grain size (0.02¢: the largest in the lagoon), and by the deficiency of fine sediments which are continually winnowed and transported into the lagoon. Despite the winnowing of fine sediments interisland channel deposits are generally poorly sorted and texturally immature, reflecting the heterogeneity of contributing organisms and the continual addition of variably degraded 'in-train' clasts. Three samples from the Southern Passage illustrate the coarse nature, in-train addition and textural immaturity of sediments deposited in the interisland channels, these samples (23, 24, 58) located in close proximity to each other and classified respectively as a slightly gravelly coarse sand, a sandy gravel and a gravelly sand. Coral-type sediments dominate the slightly gravelly medium sands which extend from the interisland channels over the sand aprons and through much of the lagoon centre (Figs. 4 and 6). Despite the continuity of these facies beyond the sand apron fringe (Fig. 2), however, textural gradients in samples collected from the sand aprons and changes in minor component abundance suggest that sediments deposited over sand aprons are allochthonous whilst those deposited beyond these features are autochthonous. The evolution of analogous textural attributes in skeletal carbonate deposits due to either hydrodynamic sorting or skeletal architecture is a principal shortcoming of carbonate texture as an environmental discriminator (Stoddart 1969, Montaggioni et al. 1986) and is well demonstrated here. The redeeming usefulness of textural gradients for environmental interpretation is, however, also confirmed. Extending into the lagoon over the sand aprons a marked decline in gravel content (24.89% to 7.94%) and an increase in the proportion of sands (74.92% to 90.97%) and muds (0.18% to 1.08%) occurs, conforming elegantly with models of lagoonal sedimentation which predict a systematic decline in mean grain size with distance from the reef rim (Frith 1983, Chevillon and Clavier 1988). Size-sorting is characteristic of backreef sand aprons on other reefs where hydrodynamic energy levels abate into the lagoon and are paralleled by a decline in mean grain size (Macintyre et al. 1987). Size- sorting generally becomes evident from around the mid-range of sand aprons extending into the Cocos (Keeling) Islands lagoon; sediments deposited at this distance from the interisland channels sufficiently removed from locally generated sediment sources to attain some degree of textural maturity. Textural gradients and composition indicate that the sand aprons predominantly comprise allochthonous sediments shed from the reef rim. Similar backreef sand deposits are described in the Pacific (Marshall and Jacobson 1985, Scoffin and Tudhope 1985, Tudhope 1989), where medium grade coral sands also dominate the lagoonward fringe. The penetration of allochthonous sand aprons in the Cocos (Keeling) Islands lagoon is similar to that reported from other reefs (Scoffin and Tudhope 1985), however at this atoll they are spatially restricted to where interisland channels link the outer reef flat to the lagoon and concentric backreef facies belts do not develop. The systematic decline of mean grain size ceases at the lagoonward margin of the sand aprons essentially marking the limit of allochthonous slightly gravelly medium sand penetration into the lagoon. Grain component data (Table 6) support the assertion that allochthonous sediments (greater than mud-sized) penetrate the lagoon only as far as the sand apron margins, sediments deposited over the lagoonward parts of the sand aprons being generally rounded whilst those deposited beyond sand apron fringes are predominantly angular and autochthonous. The range of the coralline algae/Halimeda facies which extend from the interisland channels immediately east of West Island and north of South Island further supports this speculation, and demonstrates the utility of skeletal carbonates derived from habitat specific organisms as biogenic tracers of sediment transport. These facies are chiefly comprised of rhodolith debris originating from these channels which can be traced, and is size-sorted, towards the lagoonward sand apron fringe. Kench (pers. comm) has suggested that the flood tidal wave entering the lagoon from the north opposes currents flowing through the Southern Passage around the lagoonward sand apron fringe, possibly impeding the transport of allochthonous sediments beyond this point. Immediately beyond the lagoonward sand apron margins the textural trends imposed by hydrodynamic sorting are corrupted by the addition of autochthonous gravels and sands and the skeletal architecture of contributing organisms becomes the principal determinant of facies texture. The lagoonal limit of allochthonous sediments may be obscured, however, when they prograde over gravel bearing reefs such as those fringing the blue holes behind the eastern reef islands. Here a band of gravelly sands has developed when transported and sorted allochthonous sands mix with and are texturally overwhelmed by gravels derived from the lagoonal patch reefs. The irregular mosaic of textural facies covering the central part of the lagoon suggests that sedimentation is chiefly governed by the locally abrupt bathymetric (and environmental) change imposed by the blue holes and the sporadic occurrence of patch reefs and lag gravel deposits. Formed as autochthonous material is deposited in situ, the textural characteristics of these facies are dependant on the grain sizes yielded as contributing organisms degrade, and the extent to which hydrodynamic conditions modify these deposits. Sediments through the centre of the lagoon are characteristically poorly sorted and coherent textural gradients are lacking, indicating the absence of significant hydrodynamic modification. Low mud values suggest, however, that fines may be winnowed from exposed deposits. The prevalence of coral debris through the centre of the lagoon is convincingly demonstrated by the distribution of the coral-type compositional facies, and the mosaic of textural facies which occurs through the same region can largely be ascribed to the variable representation of epilithic gravels derived from lagoonal patch reefs. The irregular bathymetry around the blue holes further ensures an erratic distribution for textural facies in this part of the lagoon via its control of patch reef distribution. Essentially these sediments are composed of medium to coarse coral sands supplemented with varying amounts of epilithic coral gravels to form various grades of gravelly sand and sandy gravel facies. The distribution of compositional facies other than coral-type is related to the occurrence of the definitive organisms, the presence of which may also impart distinctive textural properties. Isolated molluscan mud and coralline algae/Halimeda facies in the central part of the lagoon, for example, occur where the representation of their definitive components is high, and where largely intact and gravel sized mollusc shells and Halimeda segments respectively induce local coarsening of facies texture. Though coral detritus undoubtedly dominates most sediments through this area of the lagoon, the extent to which it does so and the representation of minor components varies considerably both within and between lagoonal environments (Tables 5 and 6), largely reflecting the diffuse and weakly zoned distribution of contributing organisms and the in situ deposition of derived sediments. Despite local variations in the representation of minor components, however, the overwhelming dominance of coral debris and the relative constancy of the component assemblage through the lagoon centre, which can be attributed to the lack of strong environmental and hydrodynamic gradients, has determined that except for at the extreme environments in this lagoon distinctive correlations between lagoonal environment and compositional/textural facies are difficult to define. Widespread facies-environment coincidence has been demonstrated in many carbonate emvironments (Ginsburg 1956, Swinchatt 1965, Boscence et al. 1985), however similarly poor 10 correlations between facies distribution and lagoonal environment are reported from other lagoons where environmental/hydrodynamic condtions remain constant over most of their area (Colby and Boardman 1989). The distribution of fine sediments within the Cocos (Keeling) Islands lagoon exhibits the strongest and most consistent textural/compositional facies and lagoonal environment correlation. In contrast to other lagoons where muds winnowed from the high energy peripheral zones accumulate in the lagoon centre (McKee et al. 1959, Roy and Smith 1971), significant mud deposits in the Cocos (Keeling) Islands lagoon are confined to the sheltered depositional environments in the lee of the windward islands. The concentration of mud facies behind windward reef islands is also described from the Tarawa atoll and Chesterfield Islands lagoons where reef islands effectively isolate the lagoon in their lee from erosional waves and currents. In the Cocos (Keeling) Islands lagoon muds are almost exclusively deposited in the seagrass meadow and intertidal sand and mud flat environments in the lee of South Island and in the West Island teloks (Figs. 2 and 4), with a marked concurrence of environment and facies boundaries. In addition to the sheltered position, the current reducing affects of benthic flora may enhance fine sediment deposition over the seagrass meadows (Ginsburg and Lowenstam 1958, Swinchatt 1965, Scoffin 1970), and intertidal periods of subaerial exposure may aid the accumulation of fine sediments in the intertidal sand and mud flat areas. Adjas et al. (1990) have demonstrated that most carbonate muds deposited in atoll lagoons are biogenic rather than chemogenic, and it is likely that the muds deposited in the Cocos (Keeling) Islands lagoon are produced by the attrition of larger skeletal carbonates (due to biological and physical action). Although some of these fine sediments are no doubt produced in situ it is likely that fines winnowed from elsewhere in the lagoon and transported to these sites comprise a significant proportion of these muds. In these low energy settings ‘currents of delivery’ rather than ‘currents of removal’ (Orme 1973) principally govern facies texture. The muds are deposited with autochthonous gravels and sands derived from indigenous molluscan, and to a lesser extent crustacean and coral gravels to form the gravelly mud/molluscan mud facies depicted in Figures. 3, 4, 5 and 6. Abundant molluscan and crustacean faunas presently inhabit the areas of the lagoon where muds are deposited and generate significant quantities of gravel sized sediment, however coral gravels in these deposits usually consist of lag material deposited under different physiographic conditions (i.e. prior to being isolated from the reef rim by the reef islands) or else brought to the surface by bioturbation. The skeletal remains of organisms indigenous to the seagrass meadows and intertidal sand and mud flats are particularly well represented in the recognisable fraction of these sediments (e.g. crustaceans, gastropods, Halimeda, Marginipora), and are normally deposited reasonably intact. The fragile tests of the epibiontic foraminiferan Marginopora, for example, remain relatively undamaged in these deposits but are usually fragmented in sediments deposited elsewhere. Furthermore, minor components derived from high energy reef areas (e.g. Homotrema, Amphistegina, alcyonarian spicules) are poorly represented. Muds are only nominally present outside of these areas, isolated deposits of finer sediment elsewhere in the lagoon essentially developing due to local modification of the hydrodynamic regime by seagrass beds, patch reefs and bathymetric change. Isolated patches of slightly gravelly fine sand amongst the generally coarse sediments of the high energy sandy lagoon environment can be directly attributed to patches of the seagrass Thalassodendron, the blades of which reduce current velocity and induce the deposition of finer sediment which is then stabilised by the root system (Scoffin 1970). The association of molluscs (and molluscan debris) and seagrass evident in the Thalassia seagrass meadows behind South Island is also apparent in the isolated Thalassodendron patches, and sediments over these patches are compositionally classified as molluscan mud 11 sediments. Muds also settle from suspension and accumulate at the base of many of the blue holes where low energy levels predominate, and pockets of muddier sediment are often deposited around patch reefs which impede current flow. These sediments are also often compositionally classified as molluscan muds, however it is the domination of fine indeterminate/mud sediments in these areas which confers this classification. The deposition of muds adjacent to patch reefs due to their modification of lagoonal currents has similarly been reported by Frith (1983) and Delasalle et al. (1985) and muds are reported to accumulate at the bottom of lagoonal 'pools' in Fanning Lagoon (Roy and Smith 1971). The concentration of fine sediments in sheltered areas behind the windward islands and their general absence elsewhere suggests that ambient lagoonal currents are sufficient to entrain and transport most fines out of the lagoon. The burrowing shrimps which inhabit areas of the lagoon bedded by sand may aid this process by resuspending sediments ejected from their burrows into the water column (Tudhope and Scoffin 1984, Scoffin and Tudhope 1985, Tudhope 1989). A sizable sediment shute extending seawards between Turk's reef and Horsburgh Island physically records the transport of sediment out of this lagoon, although the character of these sediments is not known. The purging of sands and muds outside of reef systems has, however, been well documented (Neumann and Land 1975, Roberts and Suhayda 1983, Frith 1983). It is interesting to note that the sediments deposited in this lagoon do not appear to conform with the Sorby principle (Folk and Robles 1964) which predicts the generation of size specific grain size populations controlled by the skeletal architecture of the contributing organisms (Fig. 9). Non-conformance with the Sorby principle is not uncommon however, with several authors reporting no apparent size specificity in sediments derived from different constituent organisms (Clack and Mountjoy 1977, Flood and Scoffin 1978, Gabrie and Montaggioni 1982). The ubiquity of coral sediments at all grain sizes is apparent in Figure 9, and may possibly distort the recognition of distinctive component- specific grain size populations simply by dominating grain counts. CONCLUSION The lagoonal sediments of the Cocos (Keeling) Islands are principally composed of gravels and sands derived from corals with minor components such as mollusc, Halimeda and rhodolith debris becoming locally important. Coral-type sediments overwhelmingly dominate the lagoon, reflecting the lack of significant populations of carbonate producing organisms other than coral on this atoll. Textural and compositional trends indicate that allochthonous sediments are deposited in this lagoon only as far as the sand aprons and sandy lagoon floor environments, beyond which sediments are almost entirely autochthonous. Allochthonous coral-type sediments can be identified by the inclusion of significant quantities minor components which are of distinctly high energy reef origin and by size-sorting along established hydrodynamic gradients. The irregular distribution patttern of textural facies in the centre of the lagoon reflects the deposition of epilithic gravels and sands produced as sporadically distributed patch reefs and lag material degrades in situ . The concurrent distribution of the gravelly mud textural facies, the molluscan mud compositional facies and the seagrass meadow and intertidal sand and mud flat environments is remarkable, and largely reflects the extent to which depositional conditions in these facies/environments are differentiated from the rest of the lagoon. Depositional conditions in these areas are characterised by low hydrodynamic energy levels, either as a function of position relative to the high energy interisland channels and/or as a function of the current reducing action of benthic flora. Fine sands and muds, which may be both 12 allochthonous and autochthonous are deposited in these zones with a coarse gravel component derived from the remains of indigenous organisms such as gastropods and crustaceans ACKNOWLEDGEMENTS This paper is largely based on an Honours thesis submitted to the Geography Department of the University of Wollongong. I am most grateful to the supervisor of this thesis, Dr. Colin Woodroffe, for the opportunity to undertake such a study and for encouragement, advice and constructive criticism at all stages. Thanks also go to Prof. Roger McLean and Eugene Wallensky for helpful advice and guidance in the field, and to Assoc. Prof. Ted Bryant for assistance with problems statistical. The logistical support of the Cocos (Keeling) Islands Administration and Cocos Islands Council whilst in the field was also much appreciated, with special thanks to Paul Stevenson, the Government Conservator whose field and administrative assistance proved invaluable. REFERENCES Adjas, A., Masse, J.P. and Montaggioni, L.F. 1990. Fine-grained carbonates in nearly closed reef environments: Mataiva and Takapoto atolls, Central Pacific Ocean. Sedim. Geol. 67: 115-132. Bosence, D.W.J., Rowlands, R.J. and Quine, M.L. 1985. Sedimentology and budget of a recent carbonate mound, Florida Keys. Sedimentology. 32: 317-343. Chevillon, C. and Clavier, J. 1988. Sedimentological structure of the Northern Lagoon of New Caledonia. Proc. 6th Int. Coral Reefs Symp. 3: 425-430. Chevillon, C. and Clavier, J. 1990. Preliminary sedimentological results on Chesterfield lagoon (New Caledonia). Proc. ISRS. Noumea. 173-178. Clack, W.J. and Mountjoy, E. 1977. Reef sediment transport and deposition off the East Coast of Carriacou, West Indies. Proc. 3rd Int. Coral Reef Symp. 2: 97-103. Colby, N.D. and Boardman, M.R. 1989. Depositional evolution of a windward, high energy lagoon, Graham's Harbour, San Salvador, Bahamas. J. Sedim. Petrol. 59: 819-834. Darwin, C. 1842. The structure and development of coral reefs. London, Smith, Elder and Co. Delesalle, B. and others. 1985. Environmental survey of Mataiva Atoll, Tuamotu Archipelago French Polynesia. Atoll Res. Bull. 282: 1-34. Emery, K.O. 1962. Marine Geology of Guam. U.S. Geol. Surv. Prof. Papers 403-B, 76pp. Emery, K.O., Tracey, J.I. Jr. and Ladd, H.S. 1954. Geology of Bikini and nearby atolls. U.S. Geol. Surv. Prof. Papers 260-A. Folk, R.L. 1974. Petrology of sedimentary rocks. Texas, Hemphill Publishing Co. 13 Folk, R.L. and Robles, R. 1964. Carbonate sands of Isla Perez, Alacran Reef Complex, Yucatan. J. Geol. 72: 255-272. Folk, R.L. and Ward, W.C. 1957. Brazos river bar : a study of significance of grain size parameters. J. Sedim. Petrol. 27: 3-26. Frith, C.A. 1983. Some aspects of lagoonal sedimentation and circulation, One Tree Reef, southern Great Barrier Reef. BMR J. Aust. Geol. Geophys. 8: 211-221. Gabrie, C. and Montaggioni, L. 1982. Sedimentary facies from the modern coral reefs, Jordan Gulf of Aqaba, Red Sea. Coral Reefs. 1: 115-124. Ginsburg, R.N. 1956. Environmental relationships of grainsize and constituent particles in some South Florida carbonate sediments. Am. Assoc. Petrol. Geol. Bull. 40: 2384-2427. Ginsburg, R.N. and Lowenstam, H.A. 1958. The influence of marine bottom communities on the depositional environment of sediments. J. Geol. 66: 310-318. Guppy, H.B. 1889. The Cocos-Keeling Islands. Scott. Geogr. Mag. 5: 281-297; 457- 474; 569-588. Hoskin, C.M. 1966. Coral pinnacle sedimentation, Alacran Reef, Mexico. J. Sedim. Petrol. 36: 1058-1074. Klovan, J.E. 1966. The use of factor analysis in determining depositional environments from grain-size distributions. J. Sedim. Petrol. 36: 115-125. Klovan, J.E. 1975. R- and Q-mode factor analysis. In McCammon, R. (ed.) Concepts in geostatistics. Berlin, Springer-Verlag. p21-69. Kornicker, L.S. and Boyd, D.W. 1962. Shallow-water geology and environments of Alacran Reef Complex, Campeche Bank, Mexico. Am. Assoc. Petrol. Geol. Bull. 46: 640-673. Ladd, H.S., Tracey, J.I., Wells, J.W. and Emery, K.O. 1950. Organic growth and sedimentation on an atoll. J. Geol. 58: 410-425. Macintyre, I.G., Graus, R.R., Reinthal, P.N., Littler, M.M. and Littler, D.S. 1987. The barrier reef sediment apron : Tobacco Reef, Belize. Coral Reefs. 6: 1-12. Maiklem, W.R. 1968. Some hydraulic properties of bioclastic grains. Sedimentology. 10: 101-109. Marshall, J.F. and Davies, P.J. 1982. Internal structure and Holocene evolution of One Tree Reef, southern Great Barrier Reef. Coral Reefs. 1: 21-28. Marshall, J.F. and Jacobson, G. 1985. Holocene growth of a mid-Pacific atoll: Tarawa, Kiribati. Coral Reefs. 4: 11-17. Maxwell, W.G.H., Jell, J.S. and McKellar, R.G. 1964. Differentiation of carbonate sediments in the Heron Island Reef. J. Sedim. Petrol. 34: 294-308. 14 McKee, E.D., Chronic, J. and Leopold, E.B. 1959. Sedimentary belts in the lagoon of Kapingamarangi Atoll. Am. Assoc. Petrol. Geol. Bull. 43: 501-562. Milliman, J.D. 1974. Recent sedimentary carbonates. Berlin, Springer-Verlag. Montaggioni, L.F., Behairy, KA., El-sayed, M.K. and Yusuf, N. 1986. The modern reef complex, Jeddah area, Red Sea: a facies model for carbonate sedimentation on embryonic passive margins. Coral Reefs. 5: 127-150. Neumann, A.C., and Land, L.S. 1975. Lime mud deposition and calcareous algae in the Bight of Abaco, Bahamas : A Budget. J. Sedim. Petrol. 45: 763-786. Orme, G.R. 1973. Aspects of sedimentation in the coral reef environment. In Jones, O.A. and Endean, R. (eds.) The geology and ecology of coral reefs. Geol 1: 129- 182. Roberts, H.H. and Suhayda, J.N. 1983. Wave-Current interactions on a shallow reef, Nicaragua, Central America. Coral Reefs. 1: 209-214. Roy, K.J. and Smith, S.V. 1971. Sedimentation and coral reef development in turbid water: Fanning Lagoon. Pacific Sci. 25: 234-248. Scoffin, T.P. 1970. The trapping and binding of subtidal carbonate sediments by marine vegetation in Bimini Lagoon, Bahamas. J. Sedim. Petrol. 40: 249-273. Scoffin, T.P. and Tudhope, A.W. 1985. Sedimentary environments of the central region of the Great Barrier Reef of Australia. Coral Reefs. 4: 81-93. Smithers, S.G. 1990. The surficial lagoonal sediments of the Cocos (Keeling) Islands, Indian Ocean. Unpublished Hons. Thesis. Uni. Wollongong. Australia. Smithers, S.G., Woodroffe, C.D., McLean, R.F. and Wallensky, E.P. (in press). Lagoonal Sedimentation in the Cocos (Keeling) Islands, Indian Ocean. Proc. 7th Int. Coral Reef Symp. Stoddart, D.R. 1969. Ecology and morphology of recent coral reefs. Biol. Rev. 44: 433- 4 Swinchatt, J.P. 1965. Significance of constituent composition, texture and skeletal breakdown in some recent carbonate sediments. J. Sedim. Petrol. 35: 71-90. Tudhope, A.W. 1989. Shallowing-upwards sedimentation in a coral reef lagoon, Great Barrier Reef of Australia. J. Sedim. Petrol. 59: 1036-1051. Tudhope, A.W. and Scoffin, T.P. 1984. The effects of Callianassa bioturbation on the preservation of carbonate grains in Davies Reef Lagoon, Great Barrier Reef, Australia. J. Sedim. Petrol. 54: 1091-1096. Tudhope, A.W., Scoffin, T.P., Stoddart, D.R. and Woodroffe, C.D. 1985. Sediments of Suwarrow atoll. Proc. 5th Int. Coral Reef Congress. 6: 611-616. 15 Wardlaw, B.R., Henry, T.W. and Martin, W.E. 1991. Sediment facies of Enewetak atoll lagoon. U.S Geol. Surv. Pap. 1513-B. Weber, J.N. and Woodhead, M.J. 1971. Carbonate lagoon and beach sediments of Tarawa Atoll, Gilbert Islands. Atoll Res.Bull. 157. Woodroffe, C.D., McLean, R.F. and Wallensky, E. 1990a. Darwin's coral atoll: geomorphology and recent development of the Cocos (Keeling) Islands, Indian Ocean. Nat. Geog. Res. 6: 262-275. Woodroffe, C.D., McLean, R.F., Polach, H. and Wallensky, E. 1990b. Sea level and coral atolls: Late Holocene emergence in the Indian Ocean. Geology 18: 62-66. JUIUIUONAUY PNA% puesy jaaeiyy 16 6L°L8 col (SNUIEOOM S(Sa)ISOIE OG Oils Tks SiS = SSiG Bike HOG, “98 86) IES His “Hic. Lz 691 ; I's TR'TR 19'S (Gos) GS0R (Ganga OWL Sel uAS WO GOW Sto PEI HOLL tS = ONS Share tas} SY) £91 78°68 IL6 (SD) OOF) SCOR eS Omea SiON ee OO eInCleaGle VAS = Cl Ole “916 SL- Gee Ril. '6iS 8F Te8 9c Fl (SN)6010" (Sd) PAIS WLIO) SiGe “9 t Cie Omer Lion GCs cismeacOle 1GiSearsl 6S Spy” CS 6¢ Ce 8h LOIS (SN)810 (Sd)O0c Il L60- 0 0 OR CORON SiGe “iS Sip cil PANO ES) 346'9:-+ 9972 Or Cr'8L Ie Ic CSS) aTO ee (Says eSiOm EOm 60) = 1HIOS S692 86 Rik OS OI. 6S hte 9 V9 LL 8 76°98 16Tl (SD ISO a(SAIEP = s9SOm C0}, SiO) “cele eir” Se LOl= 6c] LI _ 901 9S ESI vies “Ve 19 9 1c°L6 BLT (SH910 (SW)6L'0 SI'0 0 0 OP GO SV Cea VC (68 C96 Calle BLCeV LY Ele 9'0:= (6:0 6rT 81°66 10 (SD)IEGO SAIS OR REE EO) EG SLOG SIIES Salta 85S) ep Sit il F:0- S50 “L0l-=i0) 0 Ltl 60°18 10) (SOSIa0s Sd OGG VBI Srl” Sc ekin ys 98k 19 Sh 6G WS Cl 6 (iO) S10 0 aul 6lt cel (SDSISIIOn= (SAAS hee Selb SIO ee BiG 6S Sh iG. Lal 80 Lit “Site — S38 cel Slee OF IT (SOSVSGOm (SdINISEIC OLS Sivte (SO) tel SiG WES py Gr bP eLie- 81 91 — CO air {6i0l Oct COP 99°C | (SOSNG6H0ms (Sd) 9ul= AOLO SGi0F 20) “Si0RStic el Oe 9S Sle 9 8l IOI 89 GL - See abe Gr LIT 10°18 9L el (SOS) 09'0-, (Sd) 821 Oc'l OOS Olas Clo Me Siecle Si OS saSitee Sar yi6; Eey CG C6 68 81°96 Ei (SD) (SA) OOM (L8alae Sil ON ES SIL Gicc, le Ball £ 8h Pl 9:2 60. 20. 90 v8 68°6L (Tél (SAS)tri0® \SaHiGSI= (OVO Gils 9G See Sic -2- pL Ol €6 6 88 F8l 66 .9b OF o¢ St €9f (SHS) PEO (Sd) P61 ZO Sl 8t Fe ¢ iS ed: Oi oy eas AS AS A 7A s SPs €8'rl (SS)ISOr es Sd) ZO tS GiOe a ali EA Li eale8s “Orc le CO GIG Or ls8ie OG IS LG. OIG Ge £ TEER LOE! (SO)CGOms BSA COMS 00M Vice Ge OIG = LS isle ipl I. (916, G8 bo GL Si Sit 9% (é £°8L Pool (SNVIEO0# ASAIZEI (O90 (SiGe bel SiG aril [Ge SMT, 898 TL 9 Tp LI Om 1GiS= 158 sot 9¢°06 88'S (SN)600 (Sd)99'l OO1 8&€ SE 79 LE F6 F6 98 68 DOI 8S 61 92 61 EI POT 6£°96 6re (SH)SHON] (GaGa “oi TOM MO! (80) 2:8; “Gi GOl Sill SOl--21e TOl sis 6 Sil. fo Tvl bb v6 LS’ (SO) 070-. (Sd) 671 101 l AO Gale 89 ev GAl Owl Cl Oe Cr Cle 6c, | 90 eI Lael 90°78 1O'LI (SOS)SSiOz - (SaSIAl” 980k G0’ “Cel SiSE SEG SiSIe Girl. Sail C8! =6'9) Bic 18 Gc. “70> iE! 8¢ TE €6 (Aes) (SD) MAO (Sa) EG ROR 90 FiO) Pal BES “8 esl pr SLi cls Sip 8G) 91S Sil c se 9F'68 99°6 (SOSNGHO= ASAOGT OL0> G0S G0 S0- _:G 3c9e SIS! 1S ~“9i8I- SONS 89 Gk Sick TE 6F ve 65°96 STE (SO)810-, (Sd) 711 S01 Os 0 GOSS. Oe leiGiGl) Hole Vive Ri 9 Pe86e ret 802 10 ce 8°L6 TU (SD) Z10-. (Sd) 901 920 0 0 ORS V0 OO ACS Tle Bisse GOle Cal Sule: vO e0 I L9°86 cel (SN) 800 (SM)S0O O0F0 0 0 OSs OMe COR LVAS SESS (GPC GES aL 2 8:Se 0 FiO = c0 LOT 8P°L9 IS'Ze (SO)ETO-. (Sd) Pst LO OR VOM TSOP eS) Sie le Biol Oe sO S528 ORs OC oe 116; 619 = LS) = 02 LST S786 Pll (SN) 600 (SW) 690 260 0 KOM COS a Calee Site SL OC rE) Sealer tel I BC) 5 Oats €sT 6S°CE LGLS (SASDRSiOF (SAVER Oleh 160K iO" “E10! = 8'0 VEG GS: 8c9) IS. Ph LIS 9 BON 8:Oe a Les cst 79°6S 76'8E (SHENOe- (Sd) ei Oli0w PW=Hlsl- GE VES GEL- 6 GE 89-9 Bic 816 19 9ie=-€'6¢ 8s £L°S8 Iv vl (SN)TO0= (SaviGcal” SCO SOle KO) EOS SCG “VSS G8 cl PEL VOLE LIE Vicl S9 ~ y 936 #7 1e°08 89°61 (SN) €0'0 (Sd) 871 710 0 0 LOZ Sv0) SSOP eS esl Tl aeSIe SONS Oe “90 Or -1P € 66°76 v8'P (SN)I7Oi0s, (Sd) 60: = 0L0= COM SO Isle MiG “SO 8cl Ll Vel 8S = TI 6 Leese 010260 (a6 6S°S9 Ip ve (SO) P10-, (Sd) POI SLO 0 0 0 Om £0 ZO l VE VELL Slot 1v0l- 68 a 1:St rat 86°S9 lOve (SN) 8007, (Sd) 601 180 _0 0 TOE aOR Ci Ome 9i0) Se OS G1 896 8) esa eS Sl I OVS bee Steet; SiC Coe Sale tt SOR EO S02 ots a Sle: SSAUMIYS BUIIOS UBa\] % WS IZISUICI) ajdures ‘Z QIQv, 1oJ se suoMerAaIgqe 07 Ady “‘UCOSE] Spueys] (Sul[29y) $090z dy} WO soydwes JUSWIpas Q/ JO} SONSHeIS pue elep [RdIsOOIUSWUIPIS “] WBQeL 17 x! 2>>>r>r rr>rsss 68°66 8E't6 POOL 69°66 S8°S6 £666 6¢'06 S706 76°96 $8°86 S68 06 85°66 8L'€6 8°06 cS'08 [168 79°88 POOL 88°68 1S°86 ers 6r'66 96°S6 L16 1L'L6 vr'09 Lv'69 c9' 16 C178 16 9P'8S l6'TP CT SE 8t°16 v'68 97 b6 v0'S8 10 19°9 S6EC Lv0 vy 90'0 €o°¢ 688 £70 LOT 6c 01 766 vl0 1c9 £16 6t°6l c0'8 66 01 IT? [r'8 L0'0 LO'18 Il0 cel €c9 £1 sel eel LO9 y9'91 161 10°6£ STL 9P'¥9 ZA) 65°01 79S 9¢ ZI (SO) $7°0-, (SOS) OF'0-, (SO) 91°0-, (SH) S10 (SN) 80°0-, (SD) E1'0-. (SOS) ZE°0-. (SOS) Lr'0-, (SN) ¥0'0 (SD) 0Z'0-, (SH) O10 (SN) 90°0-, (SHS) O£°0 (SN) Z0'0-, (SD) 11'0-, (SN) 00'0 (SHS) $9°0 (SN) 10°0-. (SOS) 9¢'0-, (SD) I1'0-, (SN) 60°0 (SHS) 86°0 (SN) 80°0 (SOS) Lr'O-. (SHS) Z€°0 (SH) €7'0 (SN) SO'0-. (SO) Z1'0-, (SH) 110 (SN) 00°0 (SN) €0°0-, (Sd) LV'0 (SHS) ZL'0 (SHS) vr'0 (SOS) €€°0-, (SD) 17'0-. (SD) 17'0-. (SO) LO. (SW) ¢S°0 (Sd) ZO'I (Sd) 8£"1 (SW) 68°0 (SW) 760 SMW) 05°C (Sd) 6$'I (Sd) 09'1 (SW) 780 (SW) 780 (Sd) 171 (SW) S60 (Sd) 0771 (Sd) €7'1 (Sd) 911 (Sd) Iv'l (Sd) €€°1 (Sd) 67'I (SdA) €0°7 (Sd) 9¢'1 (Sd) SUI (SW) 96'0 (SW) 76'0 (Sd) €P'l (Sd) SS'I (SW) €8°0 (SdA) P17 (SdA) 617 (Sd) 9€'I (Sd) 79'I (Sd) b7'1 (Sd) 8L'I (Sd) $8'I (Sd) 911 (Sd) O€'I (Sd) I'l (Sd) 171 (Sd) vL'| LEC 850 IO Or'l 9r'0 €9'1 v6'l cr Ire scl bc 0 £0°0 rl £L0 SA) 610 89°0 09°0 $90 06'0 vel 09'I- gsl 817 660 69°0 LYy'l vr'l 99°0 0L0 61 L10 6r'0- £0'I- 99°C 10 sel Ol C7 cl £0 Ldl 0 a0) v 9°6 ee 3 60 V1z m0) Te ery) 6v 9°¢ 60 617 L0 Stl 88 69t Est a £07 9ST ei el oll 6LI cIl 80 [X35 Vol Ol €c Lil 8S LY 81 901 8°8C Vee SOc Lut V0 oF Gis 61 v8 £0 cc IE cl BC Ic Vel ss 6s CL TSI €9 el Tl BP ims c8 80 eC cs LY Ls 6LI SL VP LC cE LY Ve 90 eS 6€ Be eS Lsl ve VSI VL 8Tl TO 300 yt CP! Te lilsvae TOS io GE 6! 0 0 EN Shi iS OC co O00 vo 0 Siva iC Gy 18 TOO OG <3 Sioa Onl IS) OS VG 3 91S 1631 co 670 vr 9¢ 10 860 Gell 7! EO; = 150 Gl) GW Gils Sal SO 6¢$0 lalate Co eCAC Com al 98 8Y TO) WO) chal (Si 88 LI? CSC BSC TO T0 XS GS GE We WS) SAS ponuluos 0 ssl Teal Pst Ps Ist 10 orl L0 Prt 0 a (Gi POT 67 6P (0) OF 90 SP rs val 87 €l 0 9ST SI OST Le StI HL Srl 0 9ET re vel Lice cel rl 6L 0 OL 87S 99 0 s9 0 LS 97 OT Vl ST 8 191 SL 6¢ Ll 17 TE IZT el OCT Cale 871 VLZ Sel TEI 71 0 09 Call os 0 6 87 OLI “T A981 18 Table 2. Summary of textural statistics for each lagoonal environment: (n) denotes number of samples per environment. Lagoonal Environment I II Hl IV Vv VI VII Vill IX (n) (10) (9) (12) (8) (8) (3) (6) (8) (12) 0.02 0.87 1.34 0.56 0.72 1.46 0.83 0.74 1.05 Mean (@) (0.68) (0.15) (1.02) (0.70) (1.22) (0.02) (1.20) (0.24) (0.89) Coarse Coarse Medium Coarse Coarse Medium Coarse Coarse Medium Sand Sand Sand Sand Sand Sand Sand Sand Sand 1.28 les, 1.54 E52 1.41 PPh 9 1.23 1.39 1.00 Sorting (0.45) (0.27) (0.53) (0.24) (0.30) (0.04) (0.26) (0.30) (0.34) (PD) PS PS eS PS PS VPS PS PS MS 0.08 -0.15 -0.21 -0.04 0.05 -0.09 0.23 0.03 -0.10 Skewness (0.28) (0.20) (0.35) (0.17) (0.39) (0.05) (0.67) (0.35) (0.15) NS GS GS NS NS NS FS NS NS Gravel % 24.89 7.94 12.82 18.55 24.45 12.81 2.03 10.98 6.90 (20.71) (6.19) (10.46) (13.86) (25.59) (0.68) (2.68) (6.14) (7.93) Sand % 74.92 90.97 13°35 80.29 73.36 65 96.63 88.13 92.58 (20.83) (6.78) (23.10) (13.38) (24.47) (6.40) (3.26) (5.94) (7.79) Mud % 0.18 1.08 13.70 1.15 2.18 22.19 1:33 0.88 0.61] (0.45) (1.24) (19.16) (1.07) (2.84) (7.03) (0.94) (1.16) (1.35) Key Sorting Skewness PS Poorly Sorted NS Near Symmetrical VPS Very Poorly Sorted Cs Coarse Skewed MS Moderately Sorted SCS Strongly Coarse Skewed Ws Well Sorted FS Fine Skewed MWS Moderately Well Sorted SFS Strongly Fine Skewed Environment I Interisland Reef Flats VI Intertidal Sand and Mud Flats I Sand Aprons Vil Algal Covered Acropora Rubble II Seagrass Meadows Vill Massive Corals Interspersed with Sandy Patches IV Variable Coral and Algal Flat IX Sandy Lagoon Floor Vv Blue Hole Mosaic 19 Table 3. Sediment sample textural classification based on factor analysis. Bold numbers represent samples with the highest loading on each factor axis. Sediment Classification Samples 1, 6, 12, 13, 14, 15, 16, 21, 24, 34, 50, 70, 134, 136, 141, 144, 148, 149, 150, Factor One LS TISSS IS4e SON 67e . 2, 3, 9, 32, 35, 38, 45, 46, 48, 49, 84, 89, 114, 128, 130, 143, 155, 163. Factor Two Factor Three Factor Four Factor Five Factor Six 5; 10; 585665125, 12: 8, 23, 29, 30, 164, 165, 169, 170, 171. SOUT 20%1225 16: 57, 60, 104, 142, 147. Variable Mixtures 11, 65, 79, 124, 126, 132, 145, 146, 157. Table 4. Textural characteristics of sediment types discriminated by factor analysis. Abbreviations as per Table 2. Standard deviations in parentheses. Factor 1 2) 3 4 5 6 (n) 25 19 6 9 5 5 Var. % 3m 23.4 12.3 6.8 5.8 5.3 Cum. Var.% B75 61.1 73.4 80.2 86 91.3 “ 0.55 1.31 -0.74 0.63 1.82 2.49 Mean (9) (0.44) (0.55) (0.60) (0.35) (0.42) (0.41) 1.11 1.26 1.56 1.61 NOY] Dui? Sorting (9) (0.24) (0.45) (0.42) (0.15) (0.14) (0.32) Skewness 0.03 -0.14 0.53 0.07 -0.34 -0.33 (0.23) (0.23) (0.35) (0.15) (0.26) (0.54) ae > 8.02 8.32 55.48 15.45 15.16 1.55 Gravel % (7.89) (9.51) (17.20) (5.10) (3.74) (2.27) 91.51 90.29 43.84 82.83 49.32 91.60 Sand % 5 (7.82) (10.12) (16.62) (4.21) (14.97) (6.86) 0.46 1.25 0.67 1.82 35.52 6.83 Mud % (0.84) (1.46) (0.72) (1.19) (12.66) (7.05) AN AN A iil Ill I A XI Wh iW WA NA WA HA AI I TA wuswUOdAUA tll LS) 9919 O6t Ofcs EES SEOl OPO SLIT 6TR ITS ITIZ OOO OIbl Gor OOR LL6I ajqeuluisajapuy 790 760 WO LIO 000 Et 600 O00 £80 70 080 000 000 Lz70 +00 OOO 110 splouryog 0010 F710 66 Zl EST OBE PRI OOO SOE L90 OFE O00 OOO LIO £€FO0 L60 S60 sueaoeysns) 000 800 000 110 000 6II SZO 861 000 LOT 000 9F0 000 OF0 OOO L90 6rO sajnoidg uelseuodaly 9Lz 710 If0 POO OOO I70 LOO 000 €6I 610 000 000 000 900 000 000 0 epljauuy 000 100 Sot 70 ESI OOO ECT SII PLr COP ZI E97 000 OOE LZZ OOO 960 SuesaylUTWes0y UMOUyUQ IZ] 000 ZI0 910 O00 6cI 800 ISL 000 970 690 000 000 000 £00 190 000 puisajsiyduy OSE 000 610 +70 89F OL6 6I'l 860 OOF LIT 7% SLI OOO LET 990 B6c SEI viodourssv yy TelECl 8Sle fe Ole Int OLE SOS 180 OF ILE 49% “O00 izz~ E60 Les ict sdsnjjop UMoUyUA O81 Ir 69€ OF9 OTE I9Z 086 8LZ 000 060 OFF 000 OOO IIZ bir E80 8c spoddaajaq SLZ 9S ITrl 69Bl ESZI LOO €6S IGE E89 OPE 668 LSE OOO Led P90 FIL 089 spodo.yses) ErZ £00 000 000 000 70 000 160 000 000 ZI0 000 000 870 780 €S0 000 puta.4ouloyy Oss! LLL 000 OSI OOO 678 ILZI OOO SZI IFO 861 OOO E76! 861 I10 S6rr ESE aesty auryjes05 009 SOLI 6L'E FIZ] 69 69S SLI SOL EZ 859 S86 8Ib OOO 68'El 80S IP9I LREI ppaunoyy L69S 9O'LS SSE SOIS RPI LOS IOIE PELL BPZS 19°89 TIS ZED LLO8 89'SS SOE SOI Libr 18107) Stl 71 727i Oc CLIT bit sor Or 68 8 6L LL 99 $9 O09 85 LS ajauies DIA Al A NS STITAS STAs I] II 0 sal Se ei I NIG VAT ion []] JuowWuUonAUY I9bl ZIO 660€ Ofb 660E IPOr ISS Pel 971 SOSc P86 CIC OO O00 FIS O00 ETF 9] QeUIUIa}9pUT 000 «6Seél)6©000 «€000 «©6000 )3=6000 «36000 »«=86190 «6000 «6000 «©6000 «6000 :«€6LthO 6CSI10 «610 )=6LS0)=—000 Sproulyoy LOilameGOlOM Corea ClUOIn Coca eSS8se- TiO) = SSIOS EPO, SSal EST O0i0) 8610) 90NI (OSi0) (0010): £70 suvade}sniy 000 «6009 «©6690 «=6Le0 «6690 «60000 «2160 «CFI «6860 «68850 690) «6870 «=6S61 «6000 «=6EL0 6CUPS0 C601 sanoidg uemeuoday ooo «LOO )«€6h00 «6000 «600: «6000 «2000 «32000 «2326000 2386000 «86000 )=6LLt (0C0 «6tHO «6000 )=—000 SBE epljeuuy Gicce 6s COIs 6b COIS Shic. “SEIS SIS 8bic Sil IH IE 950) O00 FOROR 90> (68ic 1 -19;2 SUBJOJIUIWICIO J UAOUAUL 000 %0 900 L470 900 O00 eEr0 ELI £60 000 OOO 857 SOI O00 OOO B80 8Frl puisajsiyduy TOE we O00), blecemStalee tite SCE = 9ST VO0IES OF “Y9rc VcGI= Sle 0010 Tee0y LZo0” Tel = 070 psodouiss0y SERCO. (Glee elite sclee. - F590 VC0it= p60) 67/0! 88h LOIS ~~ C8ie 19IT BSS*CR SB 659%" 6670 SosnijoyA’ UMOUyUA Q0iG OHO) 6LiOm S9I0m 620) 000) aSS6"-8ile E8el L2I0) “8ki0 8k ekO!) fOPNe O0iC > Ec SiO spod(sa}ag SSiLs alibi 1S Sua teS em Lt Sem OSs SO OSC. | 16” “8022 (Svi6s =C8al= O53: FOLIOG 8Sic- “C89 = vole spodo.jsesy 000 «€©61L0.)6000 «©6000 «©6000 «=6000 «©2090 «36180 «6480 «6000 «6000 )«68rt «(O70 «6000 )2=—000 S 000-s« 0 pUajouor OOO 06. “L6G SEC Slit 0010, 68:9 -FSSl SSI Ge -Le6 LES 9G 90S. OC ST ee66c, LE TI aesTV IUul]e10+ SGP BCGTI 208/81 9G: R OSS wCCGe, lic Coal VES OOST CBIOl- Crice SLIM IL0%) LO6 Lib 3990 ppaunjoH L96S LOZ BOLT 9BVOS BOLT OSPZ 7TE9D HOSS OOSL 160F S66E COPH OFI8 CEI9 IS'8S ES hr EIBS Lee) 9S 80S: 6b Sk SP. 6& SE. Pe. ce OF 60 ve cl OF 6 9 € y9jawes 20 “UOOSE] SPURS] (SUITIOS) SODOD oY} WOIJ soTdures JUSUIPas EG JO UOIISOdUWIOD [eIBJaYS “¢ 9{qeL 74) ee nn A Al II I A AX Tan TL LLL Te Xe Is TASTES AT Sow uCTAU 6L7. 90S EL ILEl whl 9S8T B80 Sr6 970 HO'8E 000 soo 6LS CL7l 879 O67! d/QeUIWJa}aPUT ZLO €v0 9L0 870 OcI OOO SIO €10 ceO O10 OVE Lvl €tl 000 O00 650 Sploulyoy OG = OF cei SIO) =CCON SOS constituents having amplitudes of 29 and 14 cm in the north of the lagoon and 25 and 5.6 cm in the south of the lagoon. These amplitudes also display a marked attenuation from north to south within the lagoon. The Kj, constituent has an amplitude of 12.19 cm and 10.28 cm for the north and southern parts of the atoll respectively (Table 1). The southern lagoon tides lag those of the northern lagoon (Table 1). This lag is evident in the tidal records with observed high tides lagging those in the north by 15-55 minutes dependent on the tidal range and regime (spring or neap tides). Shallow water tides (M4) are much larger in the shallow southern region of the atoll than the deeper northern lagoon. The form of the tides as identified by the amplitude ratio [F = (K; + O1)/(S2 + M2)] is 0.44. This describes the tides as mixed mainly semi-diurnal. These tides characteristically have large inequalities in range and time between the highs and lows each day (Pond and Pickard, 1983). Mean spring tidal ranges using the equation S = 2(M2 + S2) are 87 cm and 61.2 cm for the north and south of the atoll. These amplitudes display a marked attenuation of the tide from north to south. The analysis also identifies a 6.02 cm (north) and 7.21 cm (south) increase in Mean Sea Level between neap and spring tide conditions (Msf constituent, Table 1). Current Circulation Progressive vector plots of current information are shown in Figures 2-5 for forty- eight hour periods during neap and spring tides. Distance between the 6.5 hour increments 4 allow comparison of the relative velocity and/or duration of tidal flow to be made between stations. Lagoon In all areas of the lagoon, except for the western shoreline, currents appear to be tidally modulated during both neap and spring tides (Figs. 2 and 3). The western shoreline of the lagoon exhibits a unidirectional northwestward flow throughout the rising and falling of the tide. In the eastern part of the lagoon the net flow westward, during neap tides (Fig. 2), is indicative of the longer duration or faster velocity of the westward flowing ebb currents. This pattern changes during spring tidal conditions to a net movement to the northwest (Fig. 3). The southeastward rising tide current is however more prominent during spring tides and penetrates the lagoon to within 1 km of the lagoon shoreline of the eastern islands. In the southeast of the lagoon net current movement is toward the southwest (neap tides) and west (spring tides). Westward drift is reduced during spring tidal conditions due to the greater oscillation of rising and falling tide currents in the southeast and northwestward directions (Fig. 3). Current patterns in the southwest lagoon are similar for both the neap and spring tidal conditions. There is a net movement northwestward throughout the 48 hour period displayed. It is evident that rising tide currents penetrate south to within 1 km of the southern passage. Tidal elevation plays an important role in the magnitude and time period of current reversal in the east and southeast sections of the lagoon. Comparisons of the neap and spring tidal currents identify a greater net movement of water during neap tidal conditions as shown by the distance between starting and finishing points of each 48 hour period. Net flow direction in the east and southeastern sectors of the lagoon rotate 45° toward north during spring tide conditions. This highlights the increased importance of the south/north tidal flow on current direction during spring tides. Current measurements taken at the deep northeastern passage (Fig. 2) convey a tidally modulated reversal in direction between the ocean and lagoon. This is accompanied by a net movement to the west. Measurements taken in the northwest passage, however, display a unidirectional flow to the southwest. This movement is hard to interpret. There appears to be no marked reversal of current direction with the oscillation of the tide and the direction seems to suggest water is leaving the lagoon. These currents are weak (5-10 cm/s)when compared to those of the northeastern passage. Measurements taken at mid- depth and 1m above the sea bed display similar current patterns to those shown by the surface current record in the deeper passages. Currents in the mid-lagoon are weak (0-7 cm/s) and appear to oscillate with the rising and falling tide (Fig. 2). Shallow Passages The shallow passages display a unidirectional flow from ocean-side reef to lagoon during neap tide conditions (Fig. 4). During spring tides the southern passages maintain the unidirectional flow (Fig. 5), while the eastern passages display a reversal in flow direction around low tide (Fig. 5). Due to the intertidal nature of the eastern passages at spring low tide the current record is broken causing distortion to the observed pattern. It is 5 however clear that currents do travel from lagoon to ocean for part of the spring tidal cycle in the eastern passages. Tidally Driven Currents As tidal currents vary in speed their direction rotates, usually with a semi-diurnal period dominating (Pond and Pickard 1983). The figure traced out by the tip of a vector representing the tidal current will be an ellipse. Tidal current constituents are presented in Table 2 as properties of tidal current ellipses for the M2 and Kj constituents. Figure 6 presents the physical appearance of the M2 tidal ellipses within the lagoon and shallow passages. Lagoon currents are dominated by the M2 tidal constituent (Table 2, Fig. 6) which is strongest in the southwest section of the lagoon at 16.95 cm/s. This constituent is also strong in the eastern side of the atoll at 11.36 cm/s. While the K1 currents are the second strongest, they are much weaker than the M2 currents ranging between 0.77cm/s and 4.84cm/s. M3 and Mg constituents (not listed) are the next strongest but are generally less than 1 cm/s. The narrow Mp) current ellipses in the east and southwestern sections of the lagoon (Fig. 6) reveal the strong oscillatory nature of tidal currents in these zones which are enhanced by the shallow passage flow. The small and wide ellipse of the southeastern lagoon portrays the weaker currents experienced in this zone. The shallow nature of this area and the curvature of South Island may contribute to the observed weak currents and direction of net flow to the west. The Mp currents in the southeast section of the lagoon have a phase lead over the eastern and southwestern areas of the lagoon. The K1 constituent, however, shows a phase lead of 5° in the east of the lagoon followed by the southwestern section of the lagoon, with the southeastern portion lagging the east by 13°. This lag may reflect the location of the current metre in the wide shallow western section of the lagoon causing shoaling of the tide. This is supported by the shallow water constituent (M4) having its largest magnitude in this zone (2.2 cm/s). As with the lagoon currents, the shallow passage tidal currents are dominated by the M2 tide which is strongest in the southern passages at 7.6 cm/s. The K] constituent is of secondary strength within the passages (Table 2). Residual currents are those components of the observed currents that cannot be explained through gravitational tidal forces. They are produced by wind stress, wind waves and or internal waves (i.e. temperature or pressure gradients). Analysis of lagoon currents identifies a mean of 15.6% of observed currents that are produced by these ‘other’ forces (Table 2 and Figs. 7a and 7b). Within the lagoon residual currents have their greatest magnitude in the southeast lagoon where 17.2% of the northing component and 27% of the easting component not driven by tidal forces (Table 2). Residual currents account for up to 52% of the observed shallow passage currents. This explains the small semi-major axes lengths of the M2 ellipses within the passages (Fig. 6), with residual currents being twice as strong as the M2 tidal current. The orthogonal with the greatest residual strength coincides with the orientation of the passage, i.e. in the southern passages the residual current is greatest for the north-south component of velocity (Table 2 and Fig. eye Figures 7a, b and c show the observed and residual current data for locations within the southeast and southwest regions of the lagoon and the southern passage. Apart from the first four days of observations within the southeast section of the lagoon the 6 residual easting component currents flow to the west (Fig. 7a). It is suggested that residual currents in this section of the lagoon are driven by the southeast trade winds. The strength of the residual current would, therefore, depend on wind strength. Residual currents in the southwest section of the lagoon (Fig. 7b) appear to fluctuate with tidal elevation. The magnitude of these currents is small. The large node in the southern passage residual current (Fig. 7c), coincides with tropical cyclone activity that influenced the island from December 5-8, 1991. The marked velocity increase, may have been the result of increased wave action and tidal elevation at the reef crest together with increased wind speeds. Correlation of wind strength and direction, and current direction, is required to identify the driving mechanisms of the residual currents. Shaliow Passage Water Flux, Tidal Prisms and Lagoon Exchange Lagoon tidal prisms calculated for a 1m (spring) and 0.4m (neap) tide are presented in Table 3. The spring tidal prisms for each shallow passage are shown in Figure 8 and cumulative spring and neap prisms for the shallow passages are presented in Table 3. A relationship is found between the cross-sectional area of each passage and the tidal prism, with larger passages transmitting greater volumes of water from ocean to lagoon (Fig. 8). On the rising spring tide the shallow passages contribute 10% to the tidal prism. The deeper passages to the north and northeast therefore must transmit 90% of the rising spring tide prism. During the falling tide shallow passages still transmit water from the ocean to the lagoon (approximately 50% of the flood tide contribution). Over a full tidal cycle, therefore, the shallow passages contribute 14% of the total spring tidal prism. Invoking a neap tidal range of 0.4 m the rising tide prism is much smaller than the spring tide prism (Table 3). The shallow passages contribute 16.9% to the neap rising tide prism. Over an entire tidal cycle the contribution of the shallow passages to the prism is proportionately much greater than it is for spring tides (22%). In calculating the flushing time of the lagoon several assumptions have been made. First, the volume of water that enters the lagoon during the rising tide equals that leaving the lagoon during the falling tide. Water entering the lagoon through the shallow passages during the falling tide, will however, be retained within the lagoon. The semi-diurnal nature of the tidal regime would also result in more water being retained in the lagoon if the low tide did not equal the original tide level. Second, the falling tide prism expels water that resided in the lagoon at low tide. With these assumptions it requires a minimum of 2.36 days (spring tide) and 5.54 days (neap tide) for the lagoon to exchange its volume with the ocean. DISCUSSION It is evident that the Cocos (Keeling) Islands lagoon experiences mixed semi- diurnal tides with a marked diurnal inequality (Table 1). The spring tidal range in the north of the lagoon of 0.82 m, although higher than most central Pacific atolls, falls within the lower range of Indian Ocean tidal ranges (Farrow and Brander 1971) and is considerably less than the 2.74m experienced at Aldabra atoll. A phase lag is identified within the lagoon; with the tide in the south of the atoll lagging the north. This phase lag is also manifested in the tidal current properties for the K; currents (Table 2). Southern passage M2 currents lag those of the eastern passages by 111°. This evidence supports the tidal lag relationship, suggesting that the tide sets from the east-northeast and travels south through the lagoon. 7 Attenuation of tide heights was observed from north to south within the lagoon but due to the short length of data in the south of the atoll (20 days) it is not appropriate to place great significance on these differences. The broad, shallow nature of the southeastern section of the lagoon is responsible for the significant shallow water effects (M4 constituent) in this region. Shallow water effects were identified by Pugh and Rayner (1981) in Salomon atoll, which they attributed to the more enclosed nature of the lagoon. Pugh and Rayner (1981) highlighted the importance of an atoll's tidal characteristics in contributing to the ecological behaviour of reef systems. Farrow and Brander (1971) established that the timing of extreme low water at Aldabra Atoll was synchronous with maximum solar radiation. This was thought to maximise stress on many organisms on the reef. Within the Cocos lagoon the maximum exposure of reef flats occurred at midnight with the second, higher low-tide of the day occurring around midday. Reef organisms were, therefore, not stressed by solar radiation during the most extreme low tide levels. Lagoon currents and circulation are driven by the tidal regime (Table 2, Figs 2 and 3), with rising tide currents penetrating south into the lagoon within 1km of the eastern and southern passages. The component of observed currents not attributable to tidal forces, may be driven by the influence of wind and lagoon generated waves, or internal salinity differences within the lagoon. These currents are small within the lagoon (mean 15%), during the observation period, and have not been further investigated. It is suggested however, that the southeast trade winds would play a major role in driving these currents. The unidirectional ocean to lagoon flow that occurs in the shallow passages, for all but spring low tides in the east of the atoll, can be explained through the interaction of tidal height and wave action, with the height of the reef crest. As the tide rises above the reef crest, waves incident at the reef break, reform (Gourlay 1990), travel over the reef and through the passage. These translatory currents comprise the bulk of the current accounted for by the residuals (Table 2). As the waves travel across the reef-flat, friction (induced by the reef flat morphology) slows the wave induced currents. This produces a build up of water at the reef crest which forms a hydraulic gradient from the reef crest to lagoon (Tait 1972). The movement of waves across the reef is still possible on the falling tide until the water height falls below that of the reef crest. The reversal in current direction in the eastern passages, around spring low tide, results from the interaction between tidal height at the reef crest and the height of the lagoonward sand bodies. As shown in Figure 9 the maximum height of the lagoonward sand body is greater than the reef crest. As the spring low tide falls below the level of the reef crest, water is ponded inside the sand apron with no connection to the deeper lagoon. A surface water gradient forms from lagoon to ocean producing a slow reefward flow. The magnitude of this reversing flow is small. That tidal currents were observed within 1km of shallow passages, indicate that shallow passage currents have negligible effect in driving circulation beyond 1km of the passage exit. They may, however, be important in retarding and deflecting lagoon currents. During the rising tide there is an opposition of currents entering the lagoon through shallow passages and the tidal currents penetrating the lagoon from the north-northwest toward the lee of the islands. The angle of opposition of these currents on the eastern side of the lagoon would indicate there is a deflection of rising tide water to the southeast. Net flow westward is the result of longer duration ebb flow which is reinforced by the continued flux of water through the shallow passages. Water behind South Island flows southward on the rising tide and west-southwest on the falling tide, as if of a semi-circular 8 nature, driven by the shallow bathymetry and closed lagoon shoreline in this region (Figs. 2 and 3). Flow in the southwest is predominantly in the north-south direction with net movement northward. This net flow direction is thought result from the increased volume of water entering via the south passage and that volume of the lagoon prism that flows from the eastern side of the lagoon. The build up of water in the west of the lagoon is evacuated by the unidirectional northward flow along the western shoreline. Water may also be built up along this shoreline due to the southeast trade winds forcing surface water northwest. As the trade winds drop there may be a small flow of water toward the east to equalise this pressure gradient. The northeastern passage experiences strong current reversals with tidal stage, whilst the northwestern passage exhibits a unidirectional and slow movement exiting the lagoon. The increased influx of water entering through the shallow passages for the duration of the tidal cycle must increase the length and or velocity of currents exiting the lagoon on the falling tide through these large passages. It is proposed that as the tidal wave sets from the east-northeast, the northeast passage is the major conduit for tidal inflow, while the northwest passage is dominated by net flow out of the lagoon (Fig. 2). A general circulation model of the lagoon, therefore, has water entering and flowing down the eastern and central sections of the lagoon, on the rising tide; being deflected west inside South Island and flowing north up the central and western sides of the lagoon, and exiting the northwest pass during the falling tide (Fig. 10). The northeast pass also evacuates water from the lagoon on the falling tide. Results from vector plots (Figs. 2-5) identify a net flow within the lagoon which mirrors that of the falling (Fig. 10b) except for the northeast passage experiencing a westward flow. Tidal oscillation is superimposed on this net flow. There are two possible mechanisms of this net northwestward flow. First, the prevailing southeast wind may produce a small surface current, as seen in the southeast lagoon residual current (Fig. 7a). If this mechanism is important, ocean-lagoon exchange time will decrease during stronger wind conditions. Second, due to the ebb prism being greater than the flood prism, because of the continuous shallow passage input, the falling tide is of longer duration, producing a net movement in the vector plots toward the deep northwest passage. The unidirectional ocean to lagoon flux of water through the passages also retards the rising currents penetrating the lagoon from the north. Continued influx of water at slack tide and the falling tide may create a pressure head from the south and east toward the northwest and may also accelerate flow toward the northwest deep passage. That circulation in the Cocos (Keeling) Islands lagoon is driven by tidal forces is consistent with other shallow lagoon studies in Pacific atolls (e.g. Stroup and Meyers 1974). While tidal forces dominate ocean-lagoon water exchange, the wave induced water flux through the shallow passages plays an important role in this exchange, especially during neap tides where shallow passage flux can represent 22% of the entire prism. Shallow passage water flux is also identified as a primary mechanism of lagoon-ocean water exchange by Atkinson et al. (1981) in Enewetak Atoll. The flushing time of the lagoon is estimated at a minimum of 2.36 days and a maximum of 5.4 days for spring and neap tides. This time period is comparable with lagoons of similar size and depth including Aldabra atoll (Pugh and Rayner 1981) and Britomart lagoon, Great Barrier Reef (Wolanski and Pickard 1983). Fanning and Canton Atoll lagoons which have similar dimensions to the Cocos lagoon (15 km wide, 6 m deep) have far greater flushing time scales (50-95 days, Canton and <300 days Fanning) due primarily to the enclosed nature of the lagoon with few connections to the ocean. CONCLUSION The Cocos (Keeling) Islands are influenced by mixed mainly semi-diurnal tides. Tides in the shallow south of the lagoon and currents in south passage, lag those of the deeper northern lagoon and eastern passages. The Cocos (Keeling) Islands lagoon circulation is tidally driven with strong tidal currents penetrating the lagoon from the northeastern passage to within 1km of the shallow passages. The tidal range has a significant impact on net flow direction in the east and southeastern sections of the lagoon. Shallow tidal constituents are important in the shallow southeastern section of the lagoon that is bordered by South island. A general circulation model has been derived in which water entering the northeastern passage travels down the eastern shoreline of the lagoon, is deflected westward, and flows northwestward exiting through the northwest deep passage. The role of wind was found to be small in driving lagoon circulation. Wind stress may contribute to the residual currents which produce a northwestward flow. However, these currents were weak within the lagoon. Shallow passages experience unidirectional ocean to lagoon flow throughout neap tides and during spring tides in the south of the atoll. This unidirectional flow contributes to the net movement of lagoon water toward the northwest throughout the tidal cycle. Eastern passages display a reversal in current direction around spring low tides due to the interaction of tidal height at the reef crest and height of the sand apron lagoonward of the passage. Exposure of the sand apron crest produces a temporary reversed hydraulic gradient and current flow. Shallow passage currents are dominated by translatory wave motion across the reef flat. Tidal currents contribute less than 50% to observed passage currents. The influence of shallow passage hydrodynamics on lagoon circulation is negligible. Shallow passage currents penetrate up to 1km lagoonward of the passage exit. These currents deflect water entering the deep lagoon to the southeast in the eastern side of the lagoon. Unidirectional passage flow increases the ebb prism and ebb tide current velocities. Shallow passages were found to be important mechanisms for the exchange of water between ocean and lagoon with up to 15% (springtides) and 22% (neap tides) of total water entering the lagoon over a tidal cycle being transmitted through the shallow passages. Flushing times of the Cocos lagoon were found to vary between 5.4 and 2.3 days for neap and spring tidal conditions respectively. These results were consistent with atoll lagoons of similar dimension and high degree of connection with the ocean. 10 REFERENCES Andrews, J.C., and Pickard, G.L. 1990. The Physical Oceanography of Coral-Reef Systems. Chapter 2 In Dubinsky, Z. ed. Coral Reefs. p 11-48. Atkinson, M., Smith, S.V. and Stroup, E.D. 1981. Circulation in Enewetak Atoll Lagoon. Limnology and Oceanography, 26:1074-1083. Farrow, G.E. and Brander, K.M. 1971. Tidal Studies on Aldabra. Phil. Trans. Roy. Soc. Lond. B. 260: 93-121. Foreman, M.G.G. 1979a. Manual for Tidal Heights Analysis and Prediction. Institute of Ocean Sciences, Patricia Bay Canada. Pacific Marine Science Report pp77-10. Foreman, M.G.G. 1979b. Manual for Tidal Currents Analysis and Prediction. Institute of Ocean Sciences, Patricia Bay Canada. Pacific Marine Science Report pp78-6. Gallagher, B.S., Shimada, K.M., Gonzales, F.I., Jr. and Stroup, E.D. 1971. Tides and Currents in Fanning Atoll Lagoon. Pacific Sci. 25: 191-205. Gilmour, A.J. and Colman, R. 1990. A Pilot Study of the Outer Island Development Program Republic of Kiribati. Report to AIDAB on environmental studies, Kiribati. Gourlay, M. 1990. Wave Set-up and Currents on Reefs. Cay Formation and Stability. Conference on Engineering in Coral Reef Regions, Magnetic Island Townsville Nov. 5-7. Neumann, G. 1968. Ocean Currents. Elsevier Oceanography Series Volume 4. Elsevier publishing company, 351pp. Platzman, G.W. 1984. Normal Modes of the World Ocean. Part IV: Synthesis of Diurnal and Semidiurnal Tides. J. Phys. Ocean. 14: 1532-1550. Pond, S. and Pickard, G.L. 1983. Introduction to Dynamical Oceanography, Second Edition. Pergamon Press, 329pp. Pugh, D.T. 1979. Sea Levels at Aldabra Atoll, Mombasa and Mahe, Western Equatorial Indian Ocean, Related to Tides, Meteorology and Ocean Circulation. Deep-Sea Res. 26: 237-258. Pugh, D.T. and Rayner, R.F. 1981. The tidal regimes of three Indian Ocean atolls and some ecological implications. Estu. Cstl. Shelf Sci. 13: 389-407. Smith, S.V. and Jokiel, P.L. 1978. Water Composition and Biogeochemical Gradients in the Canton Atoll Lagoon. Atoll Res. Bull. 221:15-54. Stroup, E.D. and Meyers, G.A. 1974. The flood-tide jet in Fanning Island Lagoon. Pacific Sci. 28: 211-223. 11 Tait, R.J. 1972. Wave Set-up on Coral Reefs. J. Geophys. Res. 77: 2207-2211. Tchernia, P. 1980. Descriptive Regional Oceanography. Pergamon Press. von Arx, W.S. 1948. The Circulation Systems of Bikini and Rongelap Lagoons. Trans. Am. Geophys. Uni. 29: 861-870. von Arx, W.S 1954. Circulation systems of Bikini and Rongelap Lagoons, Bikini and nearby atolls, Marshall Islands. U.S. Geol. Surv. Prof. Pap. 260-B: 265-273. Wiens, H.J. 1962. Atoll Environment and Ecology. Yale University Press, New Haven, 532 pp. Wolanski, E., and Pickard, G.L. 1983. Currents and Flushing of Britomart Reef Lagoon. Coral Reefs 2: 1-8. 12 Table 1: Harmonic analysis of lagoon tides, centimetres and degrees in relation to Green- which. Major tidal constituents at the permanent Home Island tide gauge and temporary location in South Passage. Symbols indicate tidal properties of amplitude (a) and phase (g). Due to the short length of record from the temporary tide gauge (co days) the No con- stituent was unable to be resolved. ‘Constituents Frequency Home Island + South Passage (hours) a (cm) g(deg) a(cm) g(deg) MSF 354.37 6.02 60.16 Wael 237.33 O; 25.82 7.82 236.50 8.23 256.37 Ky 23.93 12.19 259.44 10.29 282.27 No 12.65 10.92 118.03 — — M2 12.42 29.39 140.84 25.02 149.35 So 12.00 14.26 186.36 5.58 186.48 M4 6.21 0.21 205.31 3.92 244.04 Table 2: Tidal current constituents at M2 (12.42 hrs) and K1 (23.93 hrs) frequencies. The symbols indicate current ellipse properties of semi—major axis length (a), semi—minor axis length (b), phase (g) and orientation (6) measured anticlockwise from east. Residual, per- centages of the orthogonal components of velocity (northing — n and easting — e) indicate the percentage of observed currents not able to be accounted for by tidal forces. Mo Ky Residuals Mooring a b g (3) a b g 8 n e (cm/s) (cm/s) (deg) (deg) (cm/s) (cms) (deg) (deg) (%) (%) E Lagoon 11.36 0.56 219.36 146.98 2.61 0.17 325.44 142.01 13.8 5.0 SE Lagoon 7.03 3.68 216.32 117.84 2.49 0.99 338.25 140.17 17.2 27.0 SWLagoon 16.95 1.03 219.09 105.81 556 0.03 330.13 111.03 12.1 19.1 South Pass 7.60 2.11 337.79 263.88 4.13 0.80 235.32 82.21 46.4 29.9 East Pass 5.15 1.13 22633 2947 219 0.18 181.89 210.49 38.96 528 Table 3: Lagoon volumes and tidal prism calculations for the Cocos (Keeling) Islands Lagoon. Low-High = low to high tide shallow passage flux. Lagoon Volume Tidal Prism Shallow Passage Total (M3 x 108) (M3 x 10®) Flux (M2? x 10) (M®x 108) High Tide Low Tide Low-High High-Low SprinTides 6125 905:5 107 10.5 Bal 15:6 Neap Tides 571.9 524.5 47 8.0 3.6 116 Maldive Is. , INDIAN OCEAN 5 COCOS (KEELING) ~ Mauritius ISLANDS 60°E 90°E Deep Lagoon (10-12m) SOUTH KEELING ISLANDS Shallow Lagoon (<3m) West Is. South Is i. mea B Reef Crest Current metre deployment — 17 days . Current metre deployment — 1 day Tide gauge stations Sand aprons t SD 96°50'E 96°55'E Figure 1. Field and Instrument location 13 14 0001 10 Jan 92 “A 631 30 Dec 91 400 . F \ ‘ — Ss ». : [ 130 . ! \ 28 Dec 91 / 30 Dec 91 ‘ H Ki \ x H \ XQ we Yi ow x 1450 18 Jan 92 ee S 130 30 Dec 91 SS) 3.5 km Distance travelled / / t 400 SS 28 Dec 91 400 \ 28 Dec 91 30 Dec 91 130 Figure 2. Lagoon progressive vector plots, neap tide. The four shallow lagoon locations show a 48 hour period starting 28/12/91 at 4am (low tide). The northwest passage is a 12 hour record and the northeast and mid lagoon locations are 24 hour records of current speed and direction. Tian \ Uy / / yell EB \\ AN 7X s I — 19 Dec 91 21 Dec 91_ 3.5 km Distance travelled 830 1 1 1 600 9 Dec 9 21 Dec 91 Figure 3. Lagoon progressive vector plots, spring tide. Each location displays a 48 hour period beginning 19/12/91 at 830am (low tide). hd 16 510 - We Nov 91 \ 2. “15 Nov 91 N . ‘ \ 1700 3 Dec 91 14.5 km Distance travelled 1930 2 Dec 91 Figure 4. Shallow passage progressive vector plots, neap tide. 48 hour periods starting at low tide on the 15/11/91 at 510am in the eastern passages and 1/12/91 at 730pm in the southern passages. Rad 24 Nov 91 Distance (Km) Distance Cea i; a ey 14.5 km 920 Distance travelled 11 Dec 91 Figure 5. Shallow passage progressive vector plots, spring tide. 48 hour periods starting at low tide on the 22/11/91 at 1030am in the eastern passages and 9/12/91 at 1150am in the southern passages. WW 18 Figure 6. M2 tidal current ellipses for the shallow lagoon and passages. All display anti-clockwise rotation. Observed Currents Southeast Lagoon - Easting Velocity ntototaigditttoecoct. -20 mn oF T 21 23 DEC Residual Currents Southeast Lagoon - Easting Velocity 10] -10 23 ‘DEC Residual Currents Southwest Lagoon - Northing Velocity 104 \ ATA NA aaron -20] 30 u NOV 91 Residual Currents South Passage - Northing Velocity Cyclone Graham Figure 7. Observed and residual current information for selected sites. Residuals are derived by subtracting tidal current components from the observed current record. (a) Southeastern lagoon Showing the easting component of velocity. Negative values indicate flow to the west. (b) South west lagoon, showing the northing component of velocity. Negative values indicate flow to the south. (c) Southern passage, northing component of velocity. as 20 Tidal Prism (M3) 10 100 1000 Cross-sectional Area (M2) Tidal Prism - Eastern Passages @ Tidal Prism - Southern Passages a Figure 8. Spring tidal prism vs cross-sectional area (at MSL) relationship for the shallow passages. Zero velocity Opposing flow | from lagoon | Lagoonward flow <---> eee Deposition Sand apron Passage,\ Figure 9. Relative height of sand apron and reef crest, eastern side of the atoll. The opposition of currents entering the lagoon via the deep passes and shallow passages is also shown. (A) Rising tide (B) Falling tide Figure 10. General circulation of the Cocos (Keeling) Islands lagoon on the rising (A) and falling (B) tide. 2 1 ATOLL RESEARCH BULLETIN NO. 409 CHAPTER 11 HERMATYPIC CORALS OF THE COCOS (KEELING) ISLANDS: A SUMMARY BY J.E.N. VERON ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER I1 HERMATYPIC CORALS OF COCOS (KEELING) ISLAND: A SUMMARY BY J.E.N. VERON * ABSTRACT Ninety nine species of reef corals are recorded from Cocos (Keeling) Atoll. Of these, all but twelve are known from Western Australia. Nine species are not recorded elsewhere in the eastern Indian Ocean and two (one being taxonomically doubtful) are possibly endemic. This account is a summary only of Re-examination of the reef corals of Cocos (Keeling) Atoll (Veron 1990a) SYSTEMATIC ACCOUNT FAMILY Astrocoeniidae Koby Genus Stylocoeniella Yabe and Sugiyama Stylocoeniella guentheri (Bassett-Smith) Records: Wells (1950), Veron (1990a) Notes: Found on most reef slopes. Inconspicuous. Usually dark green, encrusting to submassive. Septa strongly alternate. Primary septa do not reach the boss-like columella. Stylocoeniella armata (Ehrenberg) Records: Veron (1990a) Notes: Rare, inconspicuous. Septa clearly alternate. Primary septa reach the columella which is thin, style-like. Australian Institute of Marine Science, P.M.B. No. 3, Townsville M.C. Queensland, 4810. 2 Stylocoeniella cocosensis Veron 1990 Record: Veron (1990b) Notes: Corallites are irregularly exsert. Septa are in two sub-equal cycles, fine. Columellae are very small. Coenosteum spinules very fine. Each corallite has a prominent style. FAMILY Pocilloporidae Gray Pocillopora is abundant in almost all coral communities, Seriatopora is usually uncommon. The other genera, notably Stylopora, have not been recorded. Genus Pocillopora Lamarck Pocillopora damicornis (Linnaeus) Records: Ridley and Quelch (1885) (as P. brevicornis), Vaughan (1918), Wells (1950), Veron (1990a) Notes: | Uncommon but found in a wide range of environments. Usually pink in colour. Indistinguishable from mainland Australian colonies. Pocillopora verrucosa (Ellis and Solander) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: Common on most upper reef slopes. Yellow or pinkish in colour. Indistinguishable from mainland Australian colonies. Pocillopora meandrina Dana Records: Vaughan (1918), Wells (1950), (both as P. elegans Dana), Veron (1990a) Notes: | Common on most upper reef slopes. Distinguished from P. verrucosa by having smaller verrucae and as described by Veron and Pichon (1982). Pocillopora woodjonesi Vaughan Records: Vaughan (1918) with the Cocos (Keeling) Islands as type locality, Wells (1950), Veron (1990a) Notes: | Uncommon. Difficult to distinguish from P. eydouxi. Colonies identified as P. woodjonesi in situ did not have the species specific skeletal characters described by Vaughan (1918) and Wells (1950) and used by the present author. The “taxonomic status of this species requires further study. Pocillopora eydouxi Edwards and Haime Records: Vaughan (1918), Veron (1990a) Notes: Common in most coral communities. Indistinguishable from mainland Australian colonies. Genus Seriatopora Lamarck Seriatopora hystrix Dana Records: Vaughan (1918), Wells (1950) (both as S. angulata Klunzinger), Veron (1990a) Notes: The few colonies observed during the present study were small and isolated. Indistinguishable from mainland Australian colonies. Family Acroporidae Verrill Genus Montipora de Blainville Montipora monasteriata (Forskal) Records: Veron (1990a) Notes: Common ina wide range of environments. Indistinguishable from mainland Australian colonies. Montipola tuberculosa (Lamarck) Records: Veron (1990a) Notes: | Common. Indistinguishable from mainland Australian colonies. Montipora lobulata Bernard Records: Wells (1950), Veron (1990a) Notes: | Has not been recorded elsewhere in Australia. Montipora mollis Bernard Records: Veron (1990a) Notes: Probably uncommon. Indistinguishable from mainland Australian colonies. 4 Montipora peltiformis Bernard Records: Veron (1990a) Notes: Common on some reef slopes. Montipora capricornis Veron Records: Veron (1990a) Notes: | Uncommon except in the atoll lagoon where this species is an early coloniser of denuded areas. Montipora spumosa (Lamarck) Records: Vaughan (1918) and Wells (1950) Notes: Possibly now extinct at Cocos (Keeling). Montipora danae (Edwards and Haime) Records: Veron (1990a) Notes: | Uncommon. Indistinguishable from mainland Australian colonies. Montipora angulata (Lamarck) Records: Vaughan (1918) (as M. cocosensis Vaughan, with Cocos (Keeling) Island as type locality), Veron (1990a) Notes: Uncommon. Found only on reef flat or sub-tidal sand flats with M. digitata. Thick branches becoming columnar, with conspicuous open corallites. Montipora digitata (Dana) Records: Ridley and Quelch (1886) (as M. laevis Quelch), Wells (1950) (as M. laevis Quelch M. ramosa Bemard and M. rubra Quoy and Gaimard), Veron (1990a) Notes: Forms extensive monospecific stands on intertidal sand flats. Intermixed with Montipora sp. Indistinguishable from mainland Australian colonies. Montipora sp. Records: Veron (1990a) Notes: A sub-arborescent species similar to M. digitata, primarily distinguished by high reticulum ridges between corallites and flattened branch tips with few corallites. The present specimens do not belong to any previously recorded or described species known to the author. Montipora efflorescens Bernard Records: Veron (1990a) Notes: _Indistinguishable from mainland Australian colonies. Montipora grisea Bernard Records: Veron (1990a) Notes: —Indistinguishable from mainland Australian colonies. Montipora informis Bernard Records: Vaughan (1918), Veron (1990a) Notes: Rare. Indistinguishable from mainland Australian colonies. Montipora foliosa (Pallas) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: | Uncommon.Indistinguishable from mainland Australian colonies. Montipora aequituberculata Bernard Records: Veron (1990a) Notes: _ Common on some outer slopes. Usually dark grey or brown. Indistinguishable from mainland Australian colonies. Genus Anacropora Ridley Anacropora forbesi Ridley, 1884 Records: Ridley (1884), with Cocos (Keeling) Atoll as type locality Notes: Many now be extinct at Cocos (Keeling). Genus Acropora Oken One of the most distinctive characters of Cocos (Keeling) Island corals is the low diversity and, usually, the low abundance of Acropora. The only extensive stands of living 6 Acropora are on reef flats. Very extensive stands of dead arborescent species, mainly pulchra and formosa, occur in the lagoon and extensive dead tabular colonies, no longer identifiable, occur at North Keeling Island. Acropora palifera (Lamarck) Records: Vaughan (1918), Veron (1990a) Notes: Seldom common. Both reef slope and lagoon colonies are similar in growth form and corallite structures and represent only a small part of the variation described by Veron and Wallace (1984). Acropom ocellata (Klunzinger) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: | Uncommon. This species belongs with the A. humilis group, with a growth form similar to A. humilis (Dana). Axial corallites are similar in size and shape to those of A.monticulosa (Bruggemann). Radial corallites are large, round, irregular, some immersed, others large and elongated, becoming incipient axials. Living colonies are pale brown with white branch tips. Acropora robusta (Dana) Records: Wells (1950) (as A. pinguis, described as a new species from Cocos (Keeling) Atoll), Veron (1990a) Notes: Very rare. Acropora danai (Edwards and Haime) Records: Wells (1950) (as A. irregularis, described as a new species from Cocos (Keeling) Atoll), Veron (1990a) Notes: Rare. Growth form is the same as mainland Australian colonies. Corallites near branch tips may become relatively elongate. Acropora sp. 1 Records: Vaughan (1918), Wells (1950) (as A. pharaonis Edwards and Haime), Veron (1990a) Notes: | Sometimes common in shallow water. Colonies are arborescent, forming thickets in shallow water where some branches may be fused. Branches are mostly straight and tapered. Radial corallites are of two sizes, the larger arranged in rows. They are similar in structure to those of A. valenciennesi. Acropora formosa (Dana) Records: Wells (1950), Veron (1990a) Notes: Uncommon except on some reef flats. Reef flat colonies have short branches with proliferous sub-branches. No colonies with long undivided branches were seen. Mostly yellowish in colour. Acropora microphthalma (Verrill) Records: Veron (1990a) Notes: Common on reef flats and some reef slopes. Indistinguishable from mainland Australian colonies. Acropora exquisita Nemenzo Records: Possibly Wells (1950) (as A. irregularis (Brook), Veron (1990a) Notes: | Uncommon. Indistinguishable from more robust colonies from mainland Australian North-west shelf reefs. Pale colours. Acropora aspera (Dana) Records: Vaughan (1918) (possibly as A. spicifera), Wells ( 1950) (as A. hebes), Veron (1990a) Notes: Mostly uncommon and only found on reef flats. Reddish-brown in colour. Indistinguishable from mainland Australian colonies. Acropora pulchra (Brook) Records: Vaughan (1918), Veron (1990a) Notes: Formerly very abundant throughout much of the southern lagoon, forming very extensive stands often over 20 m across. Now common on some reef flats and also found on some reef slopes. Indistinguishable from mainland Australian colonies. Acropora cytherea (Dana) Records: Veron (1990a) Notes: Uncommon. The largest colonies observed were < 1 m diameter. Indistinguishable from mainland Australian colonies. Acropora paniculata Vernill Records: Veron (1990a) Notes: Rare. It appears that this is a distinct geographic sub-species of A. paniculata, but as the latter is know in the Indian Ocean from only a single specimen (from Ashmore Reef, Veron and Marsh, 1988), no definite conclusion is possible. Acropora hyacinthus (Dana) Records: Veron (1990a) Notes: Rare. Only stunted reef flat colonies were found. Acropora latistella (Brook) Records: Veron (1990a) Notes: Rare, found only on reef flats. This species was not found as large tabular colonies. Branchlets are thinner than usual for shallow-water mainland Australian colonies. Acropora nana (Studer) Records: Wells (1950), Veron (1990a) Notes: Found only on outer reef flats and upper slopes. Colonies are relatively small, otherwise indistinguishable from mainland Australian colonies. Acropora subulata (Dana) Records: Veron (1990a) Notes: Rare. Nothing is known of environment-related growth form variation. Acropora valida (Dana) Records: Vaughan (1918) and Wells (1950) (as A. variabilis (Klunzinger)), Veron (1990a) Notes: Rare, Gibson-Hill records this species from several reef flat localities (Wells 1950). Corallites are smaller and have thinner walls than usual for the species, but nothing is known of environment-related variation. Coralla from Cocos (Keeling) illustrated by Vaughan (1918, pl. 80) have the characters of the species more clearly developed. Gibson-Hill (and this author) records the colour as "dirty-white, with faint lavender-blue tips" (Wells 1950). Acropora sp. 2 Records: Veron (1990a) Notes: Rare. Colonies are irregularly arborescent. Corallites are very irregular, some being valida-like and strongly oppressed. The species was not sufficiently abundant for detailed study and nothing is known of environment-related skeletal variation. Acropora schmitti Wells Records: Wells (1950), described as a new species from Cocos (Keeling) Atoll, Veron (1990a) Notes: Not found during the present study. Gibson-Hill notes, "This coral, which is rather similar to [A. valida] in both colour and form, occurs in shallow pools on the middle section of the barrier, and on part of its seaward edge. It is not very plentiful, but it seems to be most numerous at the back of Pulo Tikus, where [five] specimens were taken" (Wells 1950). Wells (1950) notes that "the distinctive character of this species is the extraordinary thickness of the outer lip of the radial corallites, which gives them the appearance of hemispherical bowls attached to the branch by one side or by a very short thick handle”. Genus Astreopora de Blainville Astreopora myriophthalma (Lamarck) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: | Common in a wide range of environments. Indistinguishable from mainland Australian colonies. Colours vary from dark purple to cream and pale pink. Astreopora gracilis (Bernard) Records: Veron (1990a) Notes: | Usually uncommon. Indistinguishable from mainland Australian colonies. Colours are cream and pale pinkish-purple. Family Poritidae Gray Goniopora and Alveopora have not been recorded from Cocos (Keeling) Atoll. Genus Porites Link Porites solida (Forskal) 10 Records: Vaughan (1918) and Wells (1950), Veron (1990a) Notes: Uncommon. Two specimens studied were indistinguishable from mainland Australian coralla. Porites lobata Dana Records: Veron (1990a) Notes: _Indistinguishable from mainland Australian colonies. Porites australiensis Vaughan Records: Veron (1990a) Notes: Corallites have a very distinct wall formed by lateral fusion of denticles. Porites somaliensis Gravier Records: ?Guppy (1889) (as P. clavaria), Vaughan (1918), Veron (1990a) Notes: |The most abundant massive Porites on some reet flats. Colonies from shallow water usually have a knobbly growth form. Corallites are closest to P. stephensoni but the present species appears to be distinct from any mainland Australian species. The triplet is sometimes fused and columellae are laterally compressed in the line of the directive septa forming a conspicuous line. Porites cf. evermanni Vaughan Records: Veron (1990a) Notes: Rare, but very distinctive. Indistinguishable from specimens of this species recorded from Australia, the Philippines (Veron and Hodgson 1989) and elsewhere. Porites cylindrica Dana Records: Guppy (1889) (as P. palmata), Ridley and Quelch (1885) (as P. levis Dana), Vanghan (1918) (as P. nigrescens), Wells (1950) (as P. nigrescens and P. gibsonhilli). Cocos (Keeling) atoll is the type locality of P. gibsonhilli Wells. Porites cocosensis Wells, described from two specimens from Cocos (Keeling) Atoll, may also be a synonym of P. cylindrica Notes: |The most common species of intertidal reef flats and forms extensive stands on some upper reef slopes. Indistinguishable from mainland Australian coralla. 11 Porites lichen Dana Records: Vaughan (1918) Notes: _Indistinguishable from mainland Australian coralla. Porites rus (Forskal) Records: Veron (1990a) Notes: Common. Forms extensive flat plates with short, irregular columns and branches. Usually fawn or brown. Porites sp. Records: Veron (1990a) Notes: Forms plates and irregular branches and columns. Corallites are essentially similar to those of P. rus and P. latistellata Quelch, but are smaller than both. The species appears to be undescribed. Usually brightly coloured: green, blue or yellow. Family Siderastreidae Vaughan and Wells Pseudosiderastrea and Coscinaraea have not heen recorded from Cocos (Keeling). Genus Psammocora Dana Psammocora digitata Edwards and Haime Records: Wells (1950) (as P. togianensis Umbgrove) Psammocora superficialis Gardiner Records: Vaughan (1918) (as Psammocora sp.), Veron (1990a) Notes: Uncommon. Indistinguishable from mainland Australian colonoies. Colonies are encrusting and may be over 1 m diameter. These large colonies have relatively coarse skeletal characters. Colour is very uniform within colonies, mostly battleship grey, rarely bright green. Psammocora profundacella Gardiner Records: Vaughan (1918) and Wells (1950) (as P. haimeana), Veron (1990a) Notes: | Very common in a wide range of environments. Indistinguishable from mainland Australian colonies. It may form coralliths. Usually pale pink or green, but may be dark green. Sometimes with blue centres. Gibson-Hill, le referring to reef-flat colonies, notes that “it is a pearl-grey colour” (Wells 1950) Family Agriciidae Gray Genus Pavona Lamarck Pavona cactus (Forskal) Records: Wells (1950), Veron (1990a) Notes: | Common only in small isolated patches. Indistinguishable from mainland Australian colonies Pavona frondifera Lamarck Records: Veron (1990a) Notes: | Common only in small isolated patches intermixed with P. cactus. Colonies are partly encrusting and have small, irregular, upright fronds. Dark greenish- brown with pale fronds. Pavona decussata (Dana) Records: Vaughan (as P. danai (Edwards and Haime), Wells (1950), Veron (1990a) Notes: | Known from two reef flat colonies only. Coralla are colllposed of highly anastomosed plates, a growth form common on reef flats. Skeletal detail is indistinguishable from mainland Australian coralla. Pavona explanulata (Lamarck) Records: Veron (1990a) Notes: | Usually uncommon but conspciuous. Colonies are massive or columnar. Pale or dark brown in colour. Plate-like colonies common in Australia, were seldom seen. Pavona minuta Wells Records: Veron (1990a) Notes: | Common on some exposed reef sites. Colonies are massive or columnar, rarely encrusting. All colonies observed were < 0.5 m. Grey in colour. 13 Pavona varians Verrill Records: Vaughan (1918 ), Veron (1990a) Notes: Very common in a wide range of reef slope environments. Forms large encrusting plates under overhangs. Very dark colours except in niches exposed to strong sunlight. Pavona venosa (Ehrenberg) Records: Veron (1990a) Notes: Septa are very coarse making the single specimen found very distinctive. Pavona maldivensis (Gardiner) Records: Vaughan (1918), Veron (1990a) Notes: Rare. Indistinguishable from mainland Australian colonies. Pavona sp. Records: Veron (1990a) Notes: Rare. Colonies are flat unifacial plates. Corallites are very small similar to those of P. bipartita Nemenzo, but with smaller calice centres and tendency to become subplocoid. Genus Leptoseris Edwards and Haime Leptoseris papyracea (Dana) Records: Veron (1990a) Notes: Forms an extensive carpet of some hundreds of square metres at one lagoonal site. Indistinguishable from fine, highly compact mainland Australian colonies. Pale pinkish-brown in colour. Leptoseris explanata Yabe and Sugiyama Records: Veron (1990a) Notes: Rare. The single specimen studied is indistinguishable from mainland Australian colonies. 14 Leptoseris mycetoseroides Wells Records: Veron (1990a) Notes: Rare. Indistinguishable from mainland Australian colonies. Genus Gardineroseris Scheer and Pillai Gardineroseris planulata (Dana) Records: Veron (1990a) Notes: | Uncommon although found in a wide variety of habitats. Colonies flat or dome- shaped, up to 1 m high, pale brown in colour. Indistinguishable from mainland Australian colonies. Genus Pachyseris Edwards and Haime Pachyseris speciosa (Dana) Records: Veron (1990a) Notes: Forms very extensive monospecific stands south of ‘Boat Passsage’. Indistinguishable from mainland Australian colonies. FAMILY Fungiidae Dana Genus Fungia Larnarck Fungia fungites (Linnaeus) Records: Wells (1950), Veron (1990a) Notes: | Uncommon. Indistinguishable from mainland Australian coralla. Fungia concinna Verrill Records: Veron (1990a) Notes: This is the only record of the species. The single specimen collected is indistinguishable from mainland Australian coralla. Fungia granulosa Klunzinger Records: Veron (1990a) Notes: This is the only record of the species. Indistinguishable from mainland Australian coralla. 15 Fungia scutaria Verrill Records: Vaughan (1918), Wells ( 1950), Veron (1990a) Notes: _ Common on reef slopes. Indistinguishable from mainland Australian coralla except for colour. Usually cream with blue or white tentacular lobes, occasionally pink. Genus Herpolitha Eschscholtz Herpolitha limax Houttuyn Records: Vaughan (1918) (as H. crassa Dana), Wells (1950), Veron (1990a) Notes: Seen, but not examined by the author. Genus Sandalolitha Quelch Sandalolitha robusta (Quelch) Records: Veron (1990a) Notes: | Usually rare. Colonies are up to 0.5 m diameter, flattened. Small colonies are oval, larger ones are contorted according to irregularities in the substrate. The flattened irregular appearance combined with wide corallum margins free of centres, suggests a different species from that found in Australia is involved. There are, however, no skeletal details which reliably distinguish Cocos (Keeling) coralla from those of Australia. Sandalolitha dentata Quelch may be a distinct species with the growth form of the present species, but this has yet to be established. FAMILY Pectiniidae Vaughan and Wells This family is represented only by Oxpora lacera Genus Oxypora Saville-Kent Oxypora lacera (Verrill) Records: Veron (1990a) Notes: Rare. Indistinguishable from mainland Australian colonies. FAMILY Mussidae Ortmann This family is represented only by Lobophyllia hemprichii 16 Genus Lobophyllia de Blainville Lobophyllia hemprichii (Ehrenberg) Records: Veron (1990a) Notes: | Usually uncommon but very conspicuous. Indistinguishable from mainland Australian colonies and shows the full range of the species except that very large colonies were not found. Often brick red in colour. FAMILY Merulinidae Verrill This family is represented only by Hydnophora microconos. Genus Hyndophora Fischer de Waldheim Wells (1950) lists H. exesa (Pallas) as recorded from Cocos (Keeling) by Vaughan (1918). This appears to be a mistake. Hydnophora microconos (Lamarck) Records: Vaughan (1918), Veron (1990a) Notes: | Uncommon but occurs in a wide range of habitats. Indistinguishable from mainland Australian colonies. FAMILY Faviidae Gregory Genus Favia Oken Favia Stelligera (Dana) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: | Common in most communities with moderate diversity. Indistinguishable from mainland Australian colonies. Favia pallida (Dana) Records: Vaughan (1918) (as F. speciosa), Veron (1990a) Notes: Colonies are small submassive to encrusting. They are mostly mottled dark colours. Favia matthaii Vaughan Records: Veron (1990a) Wi Notes: Uncommon. Corallites are smaller than those of eastern mainland Australian colonies but similar in size to those from equatorial localities. Skeletal detail is similar throughout this range. Genus Barabattoia Yabe and Sugiyama Barabattoia amicorum (Edwards and Haime) Records: Veron (1990a) Notes: Rare. Indistinguishable from mainland Australian colonies. All specimens observed were dark brown in colour. Genus Favites Link Favites abdita (Ellis and Solander) Records: Vaughan (1918), Veron (1990a) Notes: Usually uncommon. Colonies are small, usually encrusting. Corallites of colonies in high energy environments may have greatly thickened walls. Favites pentagona (Esper) Records: Vaughan (1918) (as F. melicerum Ehrenberg), Veron (1990a) Notes: Common. Coralla have most of the range of corallite characters described by Veron et al. (1977) except that all have exsert irregular septa and no ecomorphs associated with very strong wave action were found. The size of corallites overlaps with those of eastern mainland Australian colonies, but most are slightly smaller. Genus Leptoria Edwards and Haime Leptoria phrygia (Ellis and Solander) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: Usually uncommon. Always a uniform dark grey. Indistinguishable from mainland Australian colonies. Genus Montastrea de Blainville Montastrea curta (Dana) Records: Veron (1990a) 18 Notes: Usually uncommon. Colonies are small, encrusting, pale coloured. Corallites are small (most <6mm diameter with calices <3mm) and are uniform in size. This identification is tentative only because the species is very variable and lacks conservative character and also because no colonies were found on reef flats where it would be expected to be most abundant. Genus Plesiastrea Edwards and Haime Plesiastrea versipora Edwards and Haime Records: Vaughan (1918), Veron (1990a) Notes: Genus Leptastrea Edwards and Haime Leptastrea transversa Klunzinger Records: Veron (1990a) Notes: Rare. Colonies are pale cream, submassive to encrusting. Skeletal structure is indistinguishable from mainland Australian colonies. Uncommon. The characters of the species are better defined than in most mainland Australian coralla. Corallites are of relatively uniform size, with well- defined walls. Leptastrea pruinosa Crossland Records: Veron (1990a) Notes: Uncommon. Indistinguishable from mainland Australian colonies. Usually brightly coloured. Leptastrea bottae (Edwards and Haime) Records: Vaughan (1918), Wells (1950), Veron (1990a) Notes: Common over a wide range of environments. Colonies are submassive or encrusting. Corallites are relatively uniform in size, circular, with well defined walls. Septa are thin, with little ornamentation. Colonies from exposed environments are mostly creamy coloured with very dark calices. Genus Cyphastrea Edwards and Haime Cyphastrea serailia (Forskal ) Records: Wells’ (1950) record of C. chalcidicum (Forskal) appears to be this species. Veron (1990a) Ls Notes: Common ina wide range of environments. Indistinguishable from mainland Australian colonies. Cyphastrea microphthalma (Lamarck) Records: Vaughan (1918), Veron (1990a) Notes: Common. Indistinguishable from mainland Australian colonies. Cyphastrea agassizi (Vaughan) Records: Veron (1990a) Notes: Uncommon. Colonies are encrusting with widely spaced, exsert, corallites. Colonies are nearly uniform white in colour. May form coralliths. Genus Echinopora Lamarck Echinopora lamellosa (Esper) Records: Ridley and Quelch (1885), Vaughan (1918), Wells (1950), Veron (1990a) Notes: Only three small colonies were observed in situ. Indistinguishable from mainland Australian colonies. FAMILY Dendrophylliidae Gray Genus Turbinaria Oken Turbinaria reniformis Bernard Records: Veron (1990a) Notes: Usually rare but forms very extensive monospecific stands at 2-20 m depth north of “Boat Passage’. Indistinguishable from mainland Australian colonies and has the same yellow polyps as Great Barrier Reef colonies. Polyps were extended during the day. BIOGEOGRAPHIC AFFINITIES Many common and widespread Indo-Pacific taxa have not been recorded from Cocos (Keeling) and are almost certainly absent. There are no Oculinidae or Caryophylliidae. The Pectiniidae, Mussidae and hermatypic Dendrophylliidae are represented by only one species each. There are no recorded Stylophora, Goniopora, Alveopora, Coscinaraea, Cycloseris, Polyphyllia, Lithophyllon, Podabacia, Goniastrea, Platygyra and many minor east Indian Ocean genera. 20 Of the genera that are present, only Sandalolitha does not have a distribution range crossing the Indian Ocean (Veron 1986). At species level, the isolation of the atoll from Australia is reflected in: (a) the number of species which are known from western Australia but are absent from the atoll: (223 species or 70 % of the western mainland Australian total of 318 species). (b) the number of species which are present but have not been recorded from anywhere in Australia (12 species: Stylocoeniella cocosensis, Montipora lobulata, Montipora sp., Acropora ocellata, Acropora sp. 1, Acropora sp. 2, Acropora schmitti, Porites somaliensis, Porites sp., Pavona Frondifera, Pavona sp., Cyphastrea agassizi), and (c) the substantial proportion of species (perhaps 30 %) which are present but show points of difference from their western mainland Australian counterparts (e.g. differences in colour, habitat preferences as well as skeletal and growth form differences). It may be noted that of the 12 species not recorded from Australia (“b’ above), 3 have been recorded from the Philippines (Veron and Hodgson 1989). The remaining 9 have not been previously recorded from any eastern Indian Ocean locality, but only Stylocoeniella sp. (a doubtful species), Porites sp. and Pavona sp. have not been previously recorded anywhere. Although it is possible that the latter are endemic, the corals of Indonesia are poorly known and they, along with most or all Cocos (Keeling) species, may well occur in Indonesia. The principal difference between the corals of Cocos (Keeling) and Christmas Islands, is in the much greater number of species of Montipora at Cocos (Keeling) and the greater genetic richness of Christmas Island. The latter however, is a high island with a very restricted range of habitats, especially sheltered ones. The presence or absence of corals is therefore likely to be as much a function of habitat diversity as geographic isolation or relative dispersal ability. The only general observation of this data made here is that there is no clear evidence that Christmas Island has acted as a ‘stepping stone’ for the dispersal of corals to Cocos (Keeling). REFERENCES Bernard, H.M. 1897. The genus Montipora. The Genus Anacropora. Cat. Madreporarian Corals Br. Mus. (Nat. Hist.) 3: 1-192. Guppy, H.B. 1889. The Cocos-Keeling Islands. Scot. Geog. Mag. 5: 281-297, 457- 474, 569-588. Ridley, S.O. 1884. On the classificatory value of growth and budding in the Madreporidae, and on a new genus illustrating this point. Ann. Mag. Nat. Hist. (Sth series). 13: 284-291. Ridley, S.O. and Quelch. J.J. 1885. List of corals collected in the Keeling Islands. In: H.O. Forbes A Naturalist’s wanderings in the eastern Archipelago. pp. 44-47. 21 Vaughan, T. W. 1918. Some shoal-water coarls from Murray Islands, Cocos-Keeling Islands and Fanning Island. Pap. Dep. Mar. Biol. Carnegie Inst. Wash. 9: 51-234. Veron, J.E.N. 1986. Corals of Australia and the Indo-Pacific. Sydney, Angus and Robertson 644pp. Veron, J.E.N. 1990a. Re-examination of the reef corals of Cocos (Keeling) Atoll. Rec. West. Aust. Mus. 14: 553-581. Veron, J.E.N. 1990b. New Scleractinia from Japan and other Indo-West Pacific countries. Galaxea 9: 95-173. Veron, J.E.N. and Hodgson, G. 1989. Annotated checklist of the hermatypic corals of the Philippines. Pacific Sci. 43: 234-287. Veron, J.E.N. and Marsh, L.M. 1988. Hermatypic corals of Western Australia. Rec. Western Aust. Mus. Supplement 29: 1-136. Veron, J.E.N. and Pichon, M. 1982. Scleractinia of Eastern Australia. 1V. Family Poritidae. Aust. Inst. Mar. Sci. Monogr. Ser. 5: 1-159. Veron, J,E.N., Pichon, M. and Wijsman-Best, M. 1977. Scleractinia of Eastern Australia. II. Families Faviidae, Trachyphylliidae. Aust. Inst. Mar. Sci. Monogr. Ser. 3: 1-233. Veron, J.E.N, and Wallace, C. 1984. Scleractinia of Eastern Australia. V. Family Acroporidae. Aust. Inst. Mar. Sci. Monogr. Ser. 6: 1-485. Wells, J,W. 1950. Reef corals from the Cocos-Keeling Atoll. Bull. Raffles Mus. 22: 29- 52, pl. 9-14. ATOLL RESEARCH BULLETIN NO. 410 CHAPTER 12 MARINE MOLLUSCS OF THE COCOS (KEELING) ISLANDS BY F.E. WELLS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 12 MARINE MOLLUSCS OF THE COCOS (KEELING) ISLANDS BY F.E. WELLS* INTRODUCTION Compared to other localities in the eastern Indian Ocean, the molluscs of the Cocos (Keeling) Islands were relatively well known prior to the Western Australian Museum survey in February 1989. Two short papers on the molluscs of the atolls were presented by Marratt (1879) and Rees (1950). A much more extensive list was prepared by Abbott (1950). Mrs. R.E.M. Ostheimer and Mrs. V.O. Maes spent the first two months of 1963 on Cocos collecting for the Academy of Natural Sciences of Philadelphia, as part of the International Indian Ocean Expedition. Maes (1967) presented a complete list of the species collected, and included records of species recorded by Marratt (1879) or Abbott (1950) that she did not collect on the islands. A total of 504 species were recorded, 379 of which were identified to species. With their longer time on the atoll Maes and Ostheimer naturally collected more species than the Western Australian Musuem expedition, but their collections were primarily restricted to relatively shallow water as they did not scuba-dive. They did however do some dredging in the lagoon. The Museum team collected in many of the same localities as Maes and Ostheimer, but also dived in a number of areas. Because of this many of the species which live in deeper water that were recorded by only a few specimens by Maes (1967) were shown to in fact be common. _ The following list shows all of the mollusc species known from Cocos (Keeling). Station numbers are those of the Western Australian Museum expedition (see Chapter 2). Indications of abundance are given in four categories: abundant, common, uncommon and rare. These are subjective and not quantitative. The abundance categories are based partly on the number of stations at which a species was collected, but also on the number of specimens collected and whether or not the specimens were collected alive or as a broken dead shell. Thus it is possible for a bivalve species collected at two stations as single dead valves to be rare while another species collected at one station may be listed as common because several live animals were collected. Despite these limitations use of the four categories can provide an idea of the relative abundances of the different species. In a number of cases species recorded by previous papers were not collected by the Western Australian Museum team; these are included in the species list and are annotated. In some cases Maes (1967) recorded a species whose name has subsequently been changed or provided a photograph that we consider represents a different species. To avoid confusion by people not familiar with molluscs in comparing the two species lists I have annotated our identification of the species with the name used by Maes. Several pelagic molluscs recorded by Maes have been deleted as they occur in the water column and not on the bottom. Maes recorded species in a number of families that were identified only to generic level. To avoid the possibility of duplication, these have been excluded from the list where the Western Australian Museum team found species recorded in the family that were not Western Australian Museum, Francis Street, Perth, Western Australia, 6000. 2 found by Maes. In a survey such as this where molluscs of all groups have been collected species identifications should be regarded with caution. The list however does provide a good basic knowledge of the molluscs which live on the Cocos (Keeling) Islands. The following list shows that 610 species of molluscs are known from the Cocos (Keeling) Islands. There are 496 gastropods, 109 bivalves, 1 chiton, and 4 cephalopods. No monoplacophorans, aplacophorans or scaphopods are known from the islands. The fauna is diverse, and compares favourably with the total number of species known from nearby areas that have been studied: 543 from the Maldives (Robertson, unpublished list cited by Maes), 490 from Christmas Island (Wells and Slack-Smith, 1987), and 581 from atolls off the coast of northwestern Australia (Wells, 1987). Maes (1967) was intrigued by the zoogeographical relationships of the 379 mollusc species she was able to identify to species: 82% were widespread Indo-Pacific forms, 15% were Pacific species and only 3% had Indian Ocean affinities. She thus concluded that while Cocos (Keeling) is in the Indian Ocean, the islands have a greater faunal affinity with the western Pacific than with the remainder of the Indian Ocean. However since her paper was published the Western Australian Museum has had a number of expeditions to the coral atolls off the northwest coast of Western Australia and also to Christmas Island, some 900 km northeast of Cocos (Keeling), the molluscs of which are reported in a series of papers and reports (Wells 1986; 1987; Wells and Slack-Smith 1987). Comparison of these recent data with Cocos (Keeling) will provide a better idea of the zoogeographic relationships of the atoll. Many of the species at Cocos and in the other areas are either not identified to species or are identified only provisionally. For these reasons Wells (1986) selected 20 families of prosobranch gastropods for a detailed examination of the zoogeographic relationships of the molluscs of the atolls off northwestern Australia. The families were selected because the individual species are generally large and well known taxonomically, and they are well represented in museum collections. The same 20 families are examined here. For the Cocos material 248 species of the total of 584, or 42%, of all species collected belong to the 20 families. Thus they can be considered as representative of molluscs as a whole. An index of overlap (Krebs 1978) was calculated for all combinations of Cocos (Keeling), Christmas Islands and atolls off the northwestern Australian coast. The index varies from 0 where there is no overlap to 1 where the overlap is total. Values obtained were: Cocos-Christmas 0.52 Cocos-Northwestern Australian atolls OS7 Christmas-Northwestern Australia 0.52 All three overlaps are relatively low, probably due to our rather limited knowledge of the fauna of the three areas. However the overlaps are similar among the three areas. Most of the species recorded were found at two or three of the areas. Maes (1967) commented upon several species as not occurring in Western Australia; all were found on the surveys of offshore coral reefs. Based on the increased data now available it appears that the molluscs of Cocos (Keeling) have very close faunal affinities with those of Christmas Island and the offshore areas of Western Australia. Many of the species that Maes considered to be western Pacific are in fact found throughout the three areas of the eastern Indian Ocean and should be considered to be Indo-Pacific species. Perhaps if there is a specific Indian Ocean mollusc fauna it occurs primarily in the western Indian Ocean 3 and only a few species reach as far east as Cocos (Keeling), or in the case of species such as Drupa lobata even as far as the west coast of Western Australia. Neither Maes (1967) nor the Western Australian Museum survey recorded the largest of the giant clam species, Tridacna gigas, as living on Cocos (Keeling). However large numbers of long dead shells line the shoreline of Home Island. This suggests that T. gigas occurred on the atoll when it was first inhabited but became locally extinct as it was collected by Cocos Malays for food. The spider shell Lambis lambis occurs in large numbers in shallow water in the southern part of the lagoon at Cocos. It is easily collected and is regarded as a delicacy by the Cocos Malays. The same species is also collected for food in many other areas of the Indo-Pacific but a thorough literature search failed to find a single study of the fishery biology of any species of Lambis. Being a relatively large species that occurs in shallow water L. lambis could be easily fished out, and become locally extinct as did T. gigas. If there is to be effective management of the marine environment of the Cocos (Keeling) Islands a study of the population biology of Lambis lambis is urgently needed. ACKNOWLEDGEMENTS I would like to personally thank the people and organisations acknowledged at the beginning of this report for their help to the Museum team, and would also like to thank the other team members for help in the field and provision of specimens. In addition to his technical work for the entire group, C.W. Bryce collected numerous mollusc species and photographed most of the opisthobranchs alive. G.M. Hansen and G.W. Buick identified most of the specimens during my absence on other museum projects, and C.W. Bryce identified many of the opisthobranchs; their contribution is significant and very much appreciated. REFERENCES Abbott, R.T. 1950. Molluscan fauna of the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 68-98. Krebs, C.J. 1978. Ecology. The explanation of distribution and abundance. Second edition. Harper and Row, New York. Marratt, F.P. 1879. Note on some shells from the Keeling or Cocos Islands, Indian Ocean. Proc. Lit. Philos, Soc. Liverpool, 33: 53-54. Maes, V.O. 1967. The littoral marine mollusks of Cocos-Keeling Islands (Indian Ocean). Proc. Acad. Nat. Sci., Phila. 119: 93-217. Rees, W.J. 1950. The cephalopods of the Cocos-Keeling Islands collected by C.A. Gibson-Hill. Bull. Raffles Mus. 22: 99-100. Wells, F.E. 1986. Zoogeography of prosobranch gastropods on offshore coral reefs in northwestern Australia. Veliger 29: 191-198. 4 Wells, F.E. 1987. Molluscs. In: Berry, P F. (Ed.) Faunal surveys of Ashmore reef and Cartier Island. Unpubl report to Australian National Parks and Wildlife Service. Wells, F.E. and Slack-Smith, S.M. 1987. Molluscs. In: Berry, P F. (Ed). Faunal survey of Christmas Island (Indian Ocean.). Unpubl rept to Australian National Parks and Wildlife Service. LIST OF MOLLUSCS CLASS POLYPLACOPHORA ACANTHOCHITONIDAE Acanthochitona sp. CLASS GASTROPODA SCISSURELLIDAE Sp. FISSURELLIDAE Emarginula sp. TROCHIDAE Ethalia striolata (A. Adams, 1853) Euchelus foveolatus (A. Adams, 1851) Euchelus cf. instrictus (Gould, 1851) Monilea cf. nucleus (Philippi, 1849) Trochus maculatus Linnaeus,1758 STOMATIIDAE Broderipia rosea (Broderip, 1834) Stomatella impertussa (Burrow, 1815) Stomatia phymotis Helbling, 1779 Stomatia cf. rubra (Lamarck, 1822) Synaptocochlea sp. TURBINIDAE Astralium calcar (Linnaeus, 1758) Astraea helicina (Gmelin, 1791) Leptothyra solida Preston, 1908 Parviturbo parvissima (Hedley, 1899) Turbo lajonkairii Deshayes, 1839 Turbo petholatus Linnaeus, 1758 PHASIANELLIDAE Hiloa variabilis (Pease, 1860) NERITIDAE Nerita albicilla Linnaeus. 1758 Nerita costata Gmelin, 1791 Nerita maxima Gmelin, 1791 Nerita plicata Linnaeus, 1758 Maes Maes Maes Maes Maes Maes 9, 13, 23, Rare 1, 2, 6, 9, 10, 12, 14, 15, 1617 ,.23, 25; 268 27.4295 36, Abundant. Maes Maes Maes Maes Maes 32, Uncommon 6, 9, 10, 12, 15, Uncommon Maes Maes Sr LO ellen el 5) 23, 27, 30, 32, Abundant 2, 16, 17, 23, Uncommon Maes 2, 6, 10, 12, 27, 30, Common Maes Maes 12610312) 21230, Abundant 6 Nerita polita Linnaeus, 1758 Nerita undata Linnaeus, 1758 Nerita sp. Smaragdia rangiana (Recluz, 1841) Smaragdia souverbiana (Montrouier, 1865) PHENACOLEPIDAE Phenacolepas cf. senta Hedley, 1899 Phenacolepas sp. NERITOPSIDAE Neritopsis radula (Linnaeus, 1758) LITTORINIDAE Littoraria coccinea (Gmelin, 1791) Littoraria glabrata (Philippi, 1846) Littoraria scabra (Linnaeus, 1758) Littoraria undulata (Gray, 1839) Nodilittorina pyramidalis (Quoy and Gaimard, 1833) Tectarius granularis (Gray, 1839) VITRINELLIDAE Teinostoma sp. Vitrinella sp. TRUNCATELLIDAE Truncatella guerinii A. and J. Villa, 1841 RISSOIDAE Haurakia isolata Laseron, 1956 Pyramidelloides cf. miranda (A. Adams, 1861) Rissoina ambigua Gould, 1851 Rissoina balteata Pease, 1870 Rissoina ephamilla Watson, 1886 Rissoina exasperata Souverbie, 1866 Rissoina polytropa Hedley, 1899 Rissoina cf.tenuistriata Pease, 1867 Rissoina triticea Pease, 1862 Rissoina turricula Pease, 1860 Rissoina sp. Zebina semiplicata (Pease, 1862) Zebina tridentata Michaud, 1860 ASSIMINEIDAE Assiminea sp. OMALOGYRIDAE Omalogyra sp. RISSOELLIDAE 3 undetermined species 27, Common 2, Common 2, Uncommon Maes Maes. Maes Maes 13, 23, Rare Abbott Maes 2, Uncommon 2, 12, 21, Common Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes ARCHITECTONICIDAE Heliacus sp. Philippia radiata (Roding, 1798) VERMETIDAE Dendropoma maxima (Sowerby, 1825) CAECIDAE Caecum sp. PLANAXIDAE Planaxis lineatus (DaCosta, 1776) MODULIDAE Modulus tectum (Gmelin, 1791) CERITHIIDAE Bittum sp. Cerithium atromarginatum Dautzenberg and Bouge, 1933 Cerithium columna Sowerby, 1831 Cerithium echinatum Lamarck, 1822 Cerithium egenum Gould, 1849 Cerithium cf . ianthinum Gould, 1851 Cerithium nesioticum Pilsbry and Vanetta, 1905 Cerithium nodulosum (Bruguiére, 1792) Cerithium piperitum Sowerby, 1855 Cerithium purpurascens Sowerby, 1855 Cerithium rarimaculatum Sowerby, 1855 Cerithium rostratum Sowerby, 1855 Cerithium trailli (Sowerby, 1855) Diala albugo (Watson, 1886) Gourmya gourmyii (Crosse, 1861) Obtortio diplax (Watson, 1886) Rhinoclavis asper (Linnaeus, 1758) Rhinoclavis diadema Houbrick, 1978 Rhinoclavis fasciatus (Bruguiére, 1792) (referred to R. procera (Kiener, 1841) by Maes) Rhinoclavis sinensis (Gmelin, 1791) Rhinoclavis vertagus (Linnaeus, 1767) Clypeomorus bifasciata (Sowerby, 1855) CERITHIOPSIDAE Cerithiopsis four unidentified species 13, 23, Rare 22, Rare 5, 6, 7, 10, 12, 19, 36 Abundant Maes Maes Maes 9, 16, 23, 29, Common 10, 12, 13, 27, Uncommon POM 1B) 23427382) Abundant STAB 519022. 23: 25, 32, Abundant 13, 27, Rare 13, Rare 13, 19, 22, 32, Common 12) SHOW 2s lg 9: 35, Common Maes Maes 7, 13, 22, 25, 32, Common 9, 13, 23, Uncommon 16, 26, 29, Uncommon Maes 19, Rare Maes 12, 22, Uncommon On OZ IS 15019) 22, 23, Adundant SNA US 7922, 132: Abundant 1, 5, 6, 9, 10, 27, Common 2 OMIT 27929536 Common ZONA DT) SOs SO; Adundant Maes 8 TRIPHORIDAE Triphora alveolata Adams and Reeve, 1850 Triphora concors Hinds, 1843 Triphora rubra Hinds, 1843 Triphora ustulata Hervier, 1897 Triphora verrucosa Adams and Reeve, 1850 Triphora violacea Quoy and Gaimard, 1834 Triphora sp. Viriola cancellata (Hinds, 1843) Viriola intergranosa (Hervier, 1897) Viriola interfilata (Gould, 1861) EPITONIIDAE Epitonium alata (Sowerby, 1844) Epitonium martinii (Wood, 1828) Epitonium "muricatum” Risso (of Kiener, 1838-39) Epitonium cf. symmetrica (Pease, 1867) Nodiscala? attenuata (Pease, 1860) Epitonium 3 unidentified species EULIMIDAE Balcis cumingi (A. Adams, 1854) Balcis 4 additional species Sp. 1 STILIFERIDAE ?Sulifer dubia Sowerby, 1878 FOSSARIDAE Couthouyia stoliczkanus Nevill, 1871 VANIKORIDAE Vanikoro cancellata (Lamarck, 1822) Vanikoro distans (Recluz, 1843) HIPPONICIDAE Sabia concia (Schumacher, 1817) CALYPTRAEIDAE Cheilea equestris (Linnaeus, 1758) Cheilea hipponiciformis (Reeve, 1858) STROMBIDAE Lambis chiragra (Linnaeus, 1758) Lambis lambis (Linnaeus, 1758) Lambis truncata (Humphrey, 1786) Strombus aurisdianae Linnaeus, 1758 Strombus gibberulus Linnaeus, 1758 Strombus lentiginosus Linnaeus, 1758 Strombus luhuanus Linnaeus, 1758 Strombus microurceus (Kira, 1959) Maes Maes Maes Maes. Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes Maes, 13 Rare Maes 10, Rare Abbott Maes 25, Rare 25, Rare 12, 23, 27, Common 12, 23, Uncommon Maes 1, 7, 12, Uncommon 9, 12, 35, 36, Adundant 23, Rare Maes oF O12 > 15.27, 29: Common 5, 23, Rare 9A, 23, Uncommon 9A, Uncommon] Strombus mutabilis Swainson, 1821 NATICIDAE Eunaticina papilla (Gmelin, 1791) Natica fasciata (Roding, 1798) Natica gualtieriana (Recluz, 1844) Natica lemniscata Philippi, 1852 Natica lineozona Jousseaume, 1874 Natica marochiensis Gmelin, 1791 Natica orientalis Gmelin, 1791 Natica robillardi Sowerby, 1893 Polinices mammilla (Linnaeus, 1758) Polinices melanostomus (Gmelin, 1791) Polinices simiae (Deshayes, 1838) Polinices tumidus (Swainson, 1840) LAMELLARIIDAE Lamellaria sp. ERATOIDAE Proterato sulcifera (Gray, 1832) Trivia insecta (Mighels, 1845) Trivia oryza (Lamarck, 1810) Trivia pellucidula (Reeve, 1846) Trivia producta (Gaskoin, 1836) OVULIDAE Calpurneus lacteus (Lamarck, 1810) Calpurneus verrucosus (Linnaeus, 1758) Ovula ovum (Linnaeus, 1753) Pseudocypraea adamsoni (Gray, 1832) CYPRAEIDAE Cypraea annulus Linnaeus, 1758 Cypraea arabica Linnaeus, 1758 Cypraea boivinii Kiener, 1843 Cypraea caputserpentis Linnaeus, 1758 Cypraea carneola Linnaeus, 1758 Cypraea caurica Linnaeus, 1758 Cypraea chinensis Gmelin, 1791 Cypraea coloba Melvill, 1888 Cypraea depressa Gray, 1824 Cypraea erosa Linnaeus, 1758 Cypraea fimbriata Gmelin, 1791 Cypraea globulus Linnaeus, 1758 Cypraea helvola Linnaeus, 1758 Cypraea hirundo Linnaeus, 1758 Cypraea histrio Gmelin, 1791 Cypraea isabella Linnaeus, 1758 1, 5, 9, 12, 27, Common 19, Rare 17, Rare 2, 27, 35, Common Maes Maes Maes Maes 9, 29, Rare Maes 7, Rare Maes 5, 6, 95°10; 12; 13,15, 23, Adundant Maes Maes Maes Maes Maes Maes 13, Rare Maes Maes Maes 1, 6, 10, 27, Common 12, 27, Uncommon Maes 1, 11, 12, 27, Common 13, Uncommon Maes Maes 9A, Rare 1, 10, 12, 27, Common 5, 12, 27, Uncommon 13, 19, Rare 3, 13, Rare 1, 12, 13, 23, Uncommon 1, 19, Uncommon 1, 6, 7, 12, 21, 23, Common 9, 12, 13, 25, 32, Common 10 Cypraea labrolineata Gaskoin, 1849 Cypraea lynx Linnaeus, 1758 Cypraea mauritiana Linnaeus, 1758 Cypraea microdon Gray, 1828 Cypraea moneta Linnaeus, 1758 Cypraea nucleus Linnaeus, 1758 Cypraea poraria Linnaeus, 1758 Cypraea punctata Linnaeus, 1771 Cypraea stolida Linnaeus, 1758 Cypraea talpa (Linnaeus, 1758) Cypraea teres Gmelin, 1791 Cypraea testudinaria Linnaeus, 1758 Cypraea tigris Linnaeus, 1758 Cypraea vitellus Linnaeus, 1758 TONNIDAE Malea pomum (Linnaeus, 1758) Tonna canaliculata (Linnaeus, 1758) Tonna perdix (Linnaeus, 1758) CASSIDAE Casmaria erinaceus (Linnaeus, 1758) Cypraecassis rufa (Linnaeus, 1758) CYMATIIDAE Charonia tritonis (Linnaeus, 1758) Cymatium aquatile (Reeve, 1844) Cymatium lotorium (Linnaeus, 1758) Cymatium nicobaricum (Réding, 1798) Cymatium pileare (Linnaeus, 1758) Cymatium pyrum (Linnaeus, 1758) Cymatium rubeculum (Linnaeus, 1758) Cymatium vespaceum (Lamarck, 1822) Distorsio anus Linnaeus, 1758 Gutturnium muricinum (Gmelin, 1791) Septa gemmata (Reeve, 1844) Gelagna succincta (Linnaeus, 1771) Linatella clandestina (Lamarck, 1816) BURSIDAE Bursa bufonia (Gmelin, 1791) Bursa cruentata (Sowerby, 1841) Bursa granularis (R6ding, 1798) Bursa lamarcki (Deshayes, 1853) Bursa rhodostoma (Sowerby, 1835) Bursa tuberosissima (Reeve, 1844) Tutufa rubeta (Linnaeus, 1758) Merl 3syl5,. 19922..23; Common 3, 7, Uncommon 27, Rare Maes. 12. 5556.09, 102s Lome 26, 27, 29, 35, 36, Adundant Maes 27, 32, Uncommon 13, Uncommon 32, Uncommon 13, 15, Rare 1133255,;32- Rare Maes 12, 18, 22, 36, Rare 36, Uncommon 12, Rare 36, Rare 12, 27, 30, Rare Maes Maes 12, 21, Rare 16, Rare Maes 2, 6, 9, 10, 15, 16, 17, 27, 29, 35, 36 Maes 13, 23, Uncommon 9, 32, Uncommon Maes 6, Rare 1, 9, 29, 36, Uncommon L105 172 1827829: Common 13, Rare Abbott 1, 10, 12, 27, Common 13, 23, 32, Common 19:6) 9S 100120 27 Common 19, Rare 9, 23, Uncommon 25, Rare 23, Rare COLUBRARIIDAE Colubraria nitidula (Sowerby, 1833) Colubraria muricata (Lightfoot, 1786) MURICIDAE Aspella anceps (Lamarck, 1822) Chicoreus saulii (Sowerby, 1841) (listed as Murex torrefactus Sowerby, 1841 by Maes) Murex ramosus Linnaeus, 1758 Naquetia triquetra (Born, 1778) THAIDIDAE Cronia crassulnata (Hedley, 1915) Drupa clathrata (Lamarck, 1816) Drupa grossularia Roding, 1798 Drupa lobata (Blainville, 1832) Drupa morum (Roding, 1798) Drupa pophyrostoma (Reeve, 1846) Drupa ricinus (Linnaeus, 1758) Drupa rubusidaeus Roding, 1798 Drupella chaidea (Duclos, 1832) Drupella cornus (Roding, 1798) Drupella ochrostoma (Blainville, 1832) Maculotriton digitalis (Reeve, 1844) Maculotriton sculptilis (Reeve, 1846) Maculotriton serriale (Deshayes, 1834) Morula anaxeres (Kiener, 1845) Morula biconica (Blainville, 1832) Morula fiscella (Gmelin, 1791) Morula fusconigra (Dunker, 1871) Morula granulata (Duclos, 1832) Morula margariticola (Broderip, 1832) Morula marginaira (Blainville, 1832) Morula nodicostata (Pease, 1868) Morula spinosa (H. and A. Adams. 1835) Morula uva (Roding, 1798) Nassa serta (Bruguiére, 1789) Purpura persica (Linnaeus, 1758) Thais aculeata (Deshayes, 1844) Thais armigera (Link, 1807) Thais hippocastanum (Linnaeus, 1758) CORALLIOPHILIDAE Coralliophila deformis (Lamarck, 1822) 11 15, Rare 7, 9, Uncommon Maes 1.6, 9, 134954195 21) 22° 23,32, Abundant Abbott 23, Rare 27, Rare Maes Maes 23, Uncommon 195316, 11 Ola 12427, Common 6, Rare te6snl25 15527430; Common 13, 15, 23, 32, Common Maes 10, 13, 16, 29, 30, 36, Common Maes 13, 27, Uncommon Maes 6, 10, 13, 27, 30, Common 6, Rare Maes 27, Rare Maes 6, 10, 12, 27, 30, Common eG! 22-7) 29 3243.6, Common Maes 13, 30, Common be M3995 25.32), Common 155619) 1Of12; 13.19, 22, 23, 27, 30, 32, Abundant Maes Maes 5, 6, 10, 19, 27, Uncommon 1, 10, 12, 27, Uncommon Maes Maes 12 Coralliophila erosa (R6ding, 1798) Coralliophila robillardi (Lienard, 1870) Coralliophila violacea (Kiener, 1836) Quoyula madreporarum (Sowerby, 1832) Rapa rapa (Gmelin, 1791) COLUMBELLIDAE Aesopus cumingi (Reeve, 1859) Mitrella marquesa (Gaskoin, 1851) Pyrene obtusa (Sowerby, 1832) Pyrene turturina (Lamarck, 1822) Pyrene varians (Sowerby, 1832) Zafra sinensis (Sowerby, 1894) BUCCINIDAE Cantharus cf. fragaria (Reeve, 1846) Cantharus fumosus (Dilwyn, 1817) Cantharus iostomus (Gray, 1834) Cantharus pulcher (Reeve, 1846) Cantharus undosus (Linnaeus, 1758) Engina incarnata (Deshayes in Laborde and Linant, 1834) Engina lauta (Reeve, 1846) Engina lineata (Reeve, 1846) Engina melanozona Tomlin, 1928 Engina mendicaria (Linnaeus, 1758) Engina parva Pease, 1867 Engina zonalis (Lamarck, 1812) Nassaria pusilla (R6ding, 1798) Pisania fasciculata (Reeve, 1846) Pisania marmorata (Reeve, 1846) Pisania truncata (Hinds, 1844) NASSARIIDAE Nassarius gaudiosus (Hinds, 1844) Nassarius graniferus (Kiener, 1834) Nassarius margaritiferus (Dunker, 1847) Nassarius oneratus (Deshayes, 1863) Nassarius papillosus (Linnaeus, 17 58) FASCIOLARIIDAE Latirus nodatus (Gmelin, 1791) Latirus polygonus (Gmelin, 1791) Latirus turritus (Gmelin, 1790) Latirus sp. Peristernia fragaria (Wood, 1828) Peristernia nassatula (Lamarck, 1822) Peristernia ustulata (Reeve, 1847) 13, 36, Uncommon 13,30, Uncommon ST OMI2 AIS lS lO R22 23, 26, 32, Abundant 7, 10, 13, 23, 32, Common 7, 13, 25, 32, Common Maes Maes 5, 7, 13, 23, 32, Common 55°79) 10124133 19a 23, 29, Common 9, 13, 32, Uncommon Maes 7, 13, 19, 23, Common 9, 17, Uncommon 19, 22, 29, Uncommon 7, 19, 32, Uncommon 1, 5, 6, 10, 27, Common 7, Rare Maes 13, 27, Uncommon Maes 27, Uncommon 13, Rare 6, 27, Uncommon 32, Rare Maes Maes Maes 1, 7, Rare 5579 9,:135 lo AS wae? Abundant 6, 12, 27, Uncommon 23, Common 5, 13, Uncommon 10, Rare 5: 110,216;) 17 S19 26827 30, 32, Common 32, Rare 13, 29, 32, Common Maes 6: 7) 135 1519237275 Common 27, Rare OLIVIDAE Oliva annulata (Gmelin, 1791) Oliva caerulea (R6ding, 1798) Oliva episcopalis Lamarck, 1811 Oliva panniculata Duclos, 1835 MARGI NELLIDAE Marginella neville Jousseaume, 1875 MITRIDAE Cancilla carnicolor (Reeve, 1844) Cancilla filaris (Linnaeus, 1771) Imbricaria conovula (Quoy and Gaimard, 1833) Imbricaria olivaeformis (Swainson, 1821) Imbricaria punctata (Swainson, 1821) Imbricaria vanikorensis (Quoy and Gaimard, 1833) Imbricaria virgo (Broderip, 1836) Mitra acuminata Swainson, 1824 Mitra ambigua Swainson, 1832 Mitra columbelliformis Kiener, 1838 Mitra contracta Swainson, 1820 Mitra decurtata Reeve, 1844 Mitra eremitarum RGding, 1798 Mitra fraga Quoy and Gaimard, 1833 Mitra imperialis Réding, 1798 Mitra litterata Lamarck, 1811 Mitra mitra (Linnaeus, 1758) Mitra paupercula (Linnaeus, 1758) Mitra.oleacea (Reeve, 1844) Mitra rosacea Reeve, 1845 Mitra scutulata (Gmelin, 1791) Mitra Stictica (Link, 1807) Mitra ticaonica Reeve, 1844 Mitra turgida Reeve, 1845 Neocancilla papilio (Link, 1807) Pusia areolata (Reeve, 1844) Pusia cancellarioides (Anton, 1839) Scabricola fusca (Swainson, 1824) Scabricola fissurata (Lamarck, 1811) Scabricola granatina Lamarck, 1811 Scabricola scabricula (Linnaeus, 1758) Subcancilla annulata (Reeve, 1844) COSTELLARIIDAE Vexillum armigera (Reeve, 1845) Vexillum cadaverosum (Reeve, 1844) Vexillum cancellarioides (Anton, 1838) Vexillum cf. corallina (Reeve, 1845) 13 HRS e122 523.9325 Common 6, 13, 19, 25, Common Maes 13, Rare TO 2s A328 Common 15, 19, 22, Common 22, Rare Maes Maes 5, Rare Maes 12, Rare 13, Rare 6, 10, 27, Uncommon 9, Rare 13, 29, Rare Maes 6, 27, Common Maes 6, 10, 27, Uncommon 1, Rare 19, Rare 27, Rare 1, Rare 13, Rare Maes 19, Rare Maes Maes 23, Rare 5, 9, 15, 19, 22, 23, Common Maes Maes 19, Rare Maes 5, 9, 19, 22, Uncommon 32, Rare 7, Rare 14 Vexillum crocatum (Lamarck, 1811) Vexillum mutabile (Reeve, 1845) Vexillum pardalis (Kuster, 1841) Vexillum speciosum (Reeve, 1844) Vexillum tuberosa (Reeve, 1845) Vexillum turrigerum (Reeve, 1845) Vexillum unifascialis (Lamarck, 1811) Vexillum zelotypum (Reeve, 1845) VASIDAE Vasum turbinellum (Linnaeus, 1758) HARPIDAE Harpa amouretta Réding, 1798 TURRIDAE Carinapex sp. Clavus laeta (R6ding, 1798) Clavus lamberti (Montrouzier, 1860 ) Clavus sp. Crassispira sp. Daphnella atractoides Hervier, 1897 Daphnella cf. boholensis (Reeve, 1843) Daphnella delicata (Reeve, 1846) Daphnella sp. Etrema scalarina (Deshayes, 1863) Eucithara souverbii (Tryon, 1884) Eucithara stromboides (Reeve, 1846) Hemidaphne reeveana (Deshayes, 1863) Hemidaphne rissoides (Reeve, 1843) Tredalea pygmaea (Dunker, 1860) Tredalea sp. Lienardia sp. Macteola cf. thiasotes (Melvill and Standen, 1897) Mitromorpha atramentosa (Reeve, 1849) Mitromorpha lachryma (Reeve, 1845) Mitromorpha stepheni (Melvill and Standen, 1897) Philbertia barnardi (Brazier, 1876) Philbertia pustulosa (DeFolin, 1867) Philbertia granicostata (Reeve, 1846) Philbertia tincta (Reeve, 1846) Philbertia sp. Turridrupa cincta (Lamarck, 1822) Turridrupa sp. Turris spectabilis (Reeve, 1843) Xenoturris cingulifera (Lamarck, 1822) Xenoturris kingae Powell, 1964 CONIDAE Conus arenatus Hwass in Bruguiére, 1792 Conus aulicus Linnaeus, 1758 7, Rare Maes 13, Rare 15, 25, 32, Uncommon 13, 22, 23, Uncommon 9, 19, Uncommon 9, Rare 7, 22, 23, Uncommon 15°5, 75, 10) 115 PAS eho: 22, 23, 25, 27, Abundant Maes Maes 7, 13, 15, 22, 23, Common 7, 23, 32, Uncommon 13, Rare 13, Rare Maes Maes Maes 13, Rare Maes Maes Maes Maes Maes Maes 15, Rare Maes Maes 5, Rare Maes Maes Maes Maes Maes Maes 27, Rare Maes 32, Rare Maes 13, 19, 22, Uncommon 15, 22, 23, Uncommon 1, 9, 19, 22, 23, Common Maes Conus betulinus Linnaeus, 1758 Conus capitaneus Linnaeus, 1758 Conus catus Hwass in Bruguiére, 1792 Conus chaldeus (RG6ding, 1798) Conus coronatus Gmelin, 1791 Conus distans Hwass in Bruguiére, 1792 Conus eburneus Hwass in Bruguiére, 1792 Conus ebraeus Linnaeus, 1758 Conus flavidus Lamarck, 1810 Conus frigidus Reeve, 1848 Conus hevassi A. Adam, 1853 Conus imperialis Linnaeus, 1758 Conus leopardus Roding, 1798 Conus litoglyphus Hwass in Bruguiére, 1792 Conus litteratus Linnaeus, 1758 Conus lividus Hwass in Bruguiére, 1792 Conus marmoreus Linnaeus, 1758 Conus miles Linnaeus, 1758 Conus miliaris Hwass in Bruguiére, 1792 Conus moreleti Crosse, 1858 Conus musicus Hwass in Bruguiére, 1792 Conus obscurus Sowerby, 1833 Conus pertusus Hwass in Bruguiére, 1792 Conus pulicarius Hwass in Bruguiére, 1792 Conus quercinus Solander, 1786 Conus rattus Hwass in Bruguiére, 1792 Conus sponsalis Hwass in Bruguiére, 1792 Conus Straitellus Link, 1807 Conus striatus Linnaeus, 1758 Conus tenuistriatus Sowerby, 1857 Conus tessulatus Born, 1780 Conus textile Linnaeus, 1758 Conus tulipa Linnaeus, 1758 Conus vexillum Gmelin, 1791 Conus virgo Linnaeus, 1758 Conus vitulinus Hwass in Bruguiére, 1792 Conus zonatus Hwass in Bruguiére, 1792 TEREBRIDAE Hastula penicillata (Hinds, 1844) Terebra affinis Gray, 1834 Terebra areolata Link, 1807 Terebra argus Hinds, 1844 Terebra babylonia Lamarck, 1822 Terebra casta Hinds, 1844 WS) Maes 9, 22, 32, Common 1, 12, 27, Uncommon 1, 6, 10, Common 1, 6, 10, 27, Common 7, 10, 23, 27, Common 9, 15, Uncommon 1, 6, 10, 12, 27 Abundant 1, 7, 10, 23, Common Maes Maes 1, 7, 10,13,14, 19, Common 9, Uncommon 7, 22, Uncommon 7, 9, 23, 30, Uncommon PSs ON lO 2 1S O27: Adundant 23, Uncommon SOR SHG 722023: 25.027, 32, Abundant Le Om 12S BIS 23427532: Abundant 7, 13, Uncommon Sy He US le), I We os Ds). Uncommon 5, 22, 23, 32, Common 13, Rare 9 DARI2NES OS wi L9) 23, 27, Abundant 9, 12, Uncommon 5, 6, 7, 13, 27, Common LE6P 910 MI 2S IS; 19.322" 23, 30, Abundant 13, 22, Uncommon 7, Rare 13, Rare 19, 22, Uncommon Maes Maes 9, 32, Uncommon 9, Rare Maes Maes 9, 13, Uncommon Sh 5,9. SS g22 5:23); Common Maes 9, 13, Uncommon 13, 15, 32, Uncommon Maes 16 Terebra cerithina Lamarck, 1822 Terebra chlorata Lamarck, 1822 Terebra columellarisaris Hinds, 1844 Terebra cerithina Lamarck, 1822 Terebra crenulata (Linnaeus, 1758) Terebra dimidiata (Linnaeus, 1758) Terebra felina (Dillwyn, 1817) Terebra funiculata Hinds, 1844 Terebra guttata Burch, 1965 Terebra hectica (Linnaeus, 1758) Terebra laevigata Gray, 1834 Terebra lanceata (Linnaeus, 1767) Terebra maculata (Linnaeus, 1758) Terebra nebulosa (Sowerby, 1825) Terebra pertussa (Born, 1778) Terebra subulata (Linnaeus, 1767) Terebra solida (Gmelin, 1791) Terebra undulata Gray, 1834 Terenolla pygmaea (Hinds, 1844) SUBCLASS OPISTHOBRANCHIA PYRAMIDELLIDAE Odostomia peasei Dautzenberg and Bouge, 1933 Odostomia 6 species Otopleura mitralis (A. Adams, 1854) Pyramidella acus (Gmelin, 1791) Pyramidella dolabrata (Linnaeus, 1758) (Listed as P. terebellum (Miiller, 1774) by Maes Pyramidella sulcata (A. Adams, 1854) Turbonilla 2 species ACTEONIDAE Pupa sulcata (Gmelin, 1791) (Listed as Pupa glabra (Reeve, 1842) by Maes) Pupa nitidula (Lamarck, 1816) BULLINIDAE Bullina sp. HYDATINIDAE Hydatina amplustre (Linnaeus, 1758) Hydatina physis (Linnaeus, 1758) Micromelo guamensis (Quoy and Gaimard, 1825) BULLIDAE Bulla vernicosa Gould, 1859 Maes 5, 15, Uncommon 15, 19, Uncommon 5, Rare 5, 12, 13, 23, 32, Common 5, 6, 12, 13, 19, Common 7, 9, 12, 22, 29, Common 9, 15, 32, Uncommon 19, Rare Abbott Maes 9, 13, 15, 19, Uncommon 5; 9, A215, LON22s De Common 7, 15, 19, 22, Uncommon 15, Rare Oo 12 15> 19.23: Uncommon 19, Rare 15, 19, Uncommon 7, 13323, Rare Maes Maes Maes 12, 13, 17, 23, Common 9, Rare 9, Uncommon Maes 9, 17, 23, 29, 35 Uncommon Maes Maes Maes Maes Maes 9. 12517;59; 27.30; Uncommon ATYIDAE Atys cylindricus (Helbling, 1779) Haminoea cymbalum (Quoy and Gaimard, 1835) Phanerophthalmus cylindricus (Pease, 1861) GASTROPTERIDAE Gastropteron sp. SCAPHANDRIDAE Cylichna sp. AGLAJIDAE Philinopsis gardineri (Eliot, 1903) APLYSIIDAE Aplysia dactylomela Rang, 1828 Dolabella auricularia (Solander, 1786) Dolabrifera dolabrifera (Rang, 1828) PLEUROBRANCHIDAE Bertheliina citrina (Riippell and Leuchkart, 1828) Pleurobranchus cf. forskali Riippell and Leuckart, 1828 UMBRACULIDAE Umbraculum sinicum (Gmelin, 1791) ELYSIDAE Elysia sp. PLAKOBRANCHIDAE Placobranchus ocellatus van Hasselt, 1824 CYLINDROBULLIDAE Ascobulla sp. OXY NOEIDAE Lobiger sp. Oxynoe delicatula (G. and H. Nevill, 1869) VOLVATELLIDAE Volvatella cincta (G. and H. Nevill, 1869) Volvatella sp. JULIIDAE Julia borbonica (Deshayes, 1863) HEXABRANCHIDAE Hexabranchus sanguineus (Riippell and Leuckart, 1828) AEGIRIDAE Notodoris minor Eliot, 1904 9, 17, Uncommon 7, 10, Uncommon 12, Rare 13, Rare Maes 5, 23,Rare Maes 1, 12, Rare 1, 3, 12, 24, Rare 12, 27, Rare Maes 1, 18, Common 1, 5, 12, 27, 35, Common 18, Common Maes Maes Maes Maes Maes 5, 12, Rare 1, 29, Uncommon 17 18 Notodoris citrina Bergh, 1875 DORIDIDAE Halgerda tessellata (Bergh, 1880) Jorunna funebris (Kelaart, 1858) Platydoris cruenta (Quoy and Gaimard, 1832) Platydoris scabra (Cuvier, 1804) Dorid sp. DENDRODORIDIDAE Dendrodoris nigra (Stimpson, 1855) CHROMODORIDIDAE Chromodoris elisabethina Bergh, 1877 PHYLLIDIIDAE Phyllidia coelistis Bergh, 1905 Phyllidia elegans Bergh, 1869 Phyllidia cf. pustulosa Cuvier, 1804 Phyllidia sp. 1 Phyllidia sp. 2 Phyllidia sp. 3 TETHYDIDAE cf. Melibe sp. SUBCLASS PULMONATA SIPHONARIIDAE Siphonaria atra (Quoy and Gaimard, 1833) Siphonaria cf. normalis Gould, 1848 ELLOBIIDAE Auricula sp. Melampus castaneus (Muhlfeld, 1818) Melampus flavus (Gmelin, 1791) Melampus fasciatus (Deshayes, 1830) Pythia sp. CLASS BIVALVIA LIMOPSIDAE Cosa sp. ARCIDAE Arca plicata (Dillwyn, 1817) Arca ventricosa Lamarck, 1819 Barbatia decussata (Sowerby, 1833) Barbatia tenella Reeve, 1844 Barbatia velata (Sowerby, 1843) 1, Rare 13, Rare 12, Rare 1, Uncommon 13, Uncommon 24, Rare 1, 10, Uncommon 4, 15, Rare 23, Uncommon 13, Uncommon 12, Common 12, 15, 26, Common 4, 25, Uncommon 13, 15, Uncommon 1, Rare Maes 28, Common 2, Uncommon 1, 7, Common 2, 7, 21, Common Maes Maes Maes 7, 13, Rare 12, Uncommon Maes 22, Rare 7, 12, 22, 23, Uncommon MYTILIDAE Lithophaga nasuta (Philippi, 1846) Lithophaga teres (Philippi, 1846) Modiolus phillipinarum Hanley, 1843 (listed as Modiolus modulaides by Maes) Modiolus sp. PINNIDAE Atrina vexillum (Born, 1778) Pinna muricata Linnaeus, 1758 Streptopinna saccata ( Linnaeus, 1758) PTERIIDAE Electroma alacorvi (Dillwyn, 1817) Pinctada margaritifera (Linnaeus, 1758) Pteria penguin (Roding, 1798) Pteria sp. ISOGNOMONIDAE Isognomon ephippium (Linnaeus, 1758) Isognomon isognomum (Linnaeus, 1758) TIsognomon legumen (Gmelin, 1791) Isognomon perna (Linnaeus, 1767) PECTINIDAE Chlamys coruscans (Hinds, 1844) Chlamys irregularis (Sowerby, 1842) Chlamys lentiginosus (Reeve, 1853) Chlamys squamosus (Gmelin, 1791) Chlamys sp. Decatopecten radula (Linnaeus, 1758) Pecten pyxidatus Born, 1778 Semipallium tigris (Lamarck, 1819) PLICATULIDAE Plicatula chinensis Morch, 1853 SPONDYLIDAE Spondylus lamarckii Chenu, 1845 Spondylus nicobaricus Schreibers, 1793 Spondylus sanguineus Dunker, 1852 LIMIDAE Lima cf. annulata Lamarck, 1819 Lima fragilis (Gmelin, 1791) Limaria orientalis (Adams and Reeve, 1850) OSTREIDAE Ostrea sp. ike Maes 16, Common 1, 2, 22, 27, Uncommon 12, 25, 35, Uncommon Maes 12, 36, Uncommon 13, 19, 22, Uncommon Maes 23, Rare Maes 27, Rare Maes 9, Rare Maes 6, 7, 10, 13, Uncommon Maes 7, 15, 22, 23, 32, Common 23, 25, Common Maes 13, Rare 9, 29, 36, Uncommon Maes 26, Rare Maes 6, 9, 29, Uncommon 13, 16, 17, 19, Uncommon 13, Common 6, 13, 15, 26 29 Uncommon 9, 12, 36, Uncommon 7, 16, 19, 26, 30, 32, Common 16, 27, 28, 30, Common 20 GR YPHAEIDAE Hyotissa hyotis (Linnaeus, 1758) 9, Uncommon CHAMIDAE Chama aspersa Reeve, 1846 Maes Chama imbricata Broderip, 1834 Maes Chama cf. iostoma Conrad, 1837 6, 12, 15, 29, Uncommon Chama lazarus Linnaeus, 1758 6, 26, 28, 29, Uncommon Chama sp. 1, 9, 10, 16, Common LUCINIDAE Anodontia edentula (Linnaeus, 1758) 15, 31, Uncommon Anodontia pila (Reeve, 1850) 21, 23, 30, Uncommon Cavatidens sp. Maes Ctena sp. 36, Rare Codakia divergens (Philippi, 1850) Maes Codakia punctata (Linnaeus, 1758) 5, 9, 10; 12; 19, 32535586; Common Glycodonta sp. 15, 19, Rare Wallucina gordoni E. A. Smith 17, 30, Rare ERYCINIDAE Barrimysia incerta (Deshayes, 1863) Maes Erycinacea sp. Maes Fronsella cf. fugitaniana (Yokoyama, 1927) Maes Hitia ovalis Dall, Bartsch and Rehder, 1938 Maes Besobornia pacifica (Hedley, 1899) Maes GALEOMMATIDAE Scintillona sp. 5, Rare CARDITIDAE Cardita variegata (Bruguiére, 1792) 5, 7, 13, 15, 16, Common DIPLODONTIDAE Diplodonta sp. Maes SPORTELLIDAE cf. Anisodonta sp. Maes CARDIIDAE Acrosterigma alternatum (Sowerby, 1841) 29, 36, Uncommon Acrosterlgma orbita (Broderip and Sowerby, 1833) 6,°7; 13, 15, 22, 2359258 Common Corculum cardissa (Linnaeus, 1758) Maes Fragum fragum (Linnaeus, 1758) 2. Se 9512235436 Fragum unedo (Linnaeus, 1758) 17, Rare TRIDACNIDAE Tridacna derasa (R6ding, 1798) Maes Tridacna gigas (Linnaeus, 1758) 6 dead valves Tridacna maxima (R6ding, 1798) TELLINIDAE Arcopagia palatum Iredale, 1929 Cadella semitorta (Sowerby, 1867) Arcopagia scobinata (Linnaeus, 1758) Macoma obliquilineata (Conrad, 1837) Quadrans gargadia (Linnaeus, 1758) Tellina chariessa Salisbury, 1934 Tellina clathrata (Deshayes, 1835) Tellina crassiplicata (Sowerby, 1869) Tellina crucigera Lamarck, 1818 Tellina dispar (Conrad, 1837) Tellina linguafelis Linnaeus, 1758 Tellina obliquaria (Deshayes, 1854) Tellina palatum (Iredale, 1929) Tellina perna Splenger, 1798 Tellina pinguis (Hanley, 1845) Tellina pulcherrima (Sowerby, 1867) Tellina robusta (Hanley, 1844) Tellina tenuilirata (Sowerby, 1867) Tellina tongana (Quoy and Gaimard, 1835) Tellina sp. PSAMMOBIIDAE Asaphis violaceans (Forskal, 1775) Gari sp. SEMELIDAE Semele crenulata (Sowerby, 1853) Thyella cf. lamellosa H. Adams, 1873 TRAPEZIIDAE Trapezium oblongum (Linnaeus, 1758) VENERIDAE Katelysia cf. striata (Gmelin, 1791) Lioconcha castrensis (Linnaeus, 1758) Lioconcha hebraea (Lamarck, 1818) Periglypta chemnitzii (Hanley, 1844) Periglypta clathrata (Deshayes, 1853) Periglypta crispata (Deshayes, 1859) Periglypta puerpera (Linnaeus, 1771) Pitar cf. affinis (Gmelin, 1791) Pitar prora (Conrad, 1837) Protothaca marica (Linnaeus, 1758) Tapes cf. literatus (Linnaeus, 1758) Ventricolaria toreuma (Gould, 1846) 21 1, 5, 6, 7, 10, 12, 19, 36, Common Maes Maes OL 2S SON 22423) 29, 35, Common 29, Rare 15, Rare Maes 9, 30, 36, Common Maes 23, Rare 2, 9, 17, 36, Uncommon 26, Rare Maes 2, Common 12, 35, Rare Maes Maes 9, 13, 36, Uncommon Maes 30, 32, Rare 17, Rare 35, Rare 17, Rare 17, 29, Rare Maes 35, Rare Maes 29, Rare 6, 9, 17, 22, 23, 26, 36, Common 9, Rare Maes 29, Rare 9, Rare 22, Uncommon 9, 17, 29, 35, Uncommon Maes 9, 26, 29, Rare 25, Rare 22 CORBULIDAE Corbula ustulata Reeve, 1844 Maes Corbula sp. Maes GASTROCHAENIDAE Gastrochaena cuneiformis (Splengler, 1783) Maes CLASS CEPHALOPODA NAUTILIDAE Nautilus pompilius Linnaeus, 1758 5, Rare LOLIGINIDAE Sepioteuthis lessoniana Lesson, 1830 Rees OCTOPODIDAE Octopus cyanea Gray, 1849 1, 12, Common SEPIDAE Sepia latimanus Quoy and Gaimard, 1832 9, Rare ATOLL RESEARCH BULLETIN NO. 411 CHAPTER 13 ECHINODERMS OF THE COCOS (KEELING) ISLANDS BY L.M. MARSH ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 13 ECHINODERMS OF THE COCOS (KEELING) ISLANDS BY L.M. MARSH * INTRODUCTION The first extensive collection of echinoderms of the Cocos (Keeling) Islands was made by C.A. Gibson-Hill who was Resident Medical Officer at the Cable Station on Direction Island (Pulu Tikus) between December 1940 and November 1941. Prior to this Gibson-Hill was Resident Medical Officer at Christmas Island from September 1938 to December 1940, where he also made extensive natural history collections. His specimens and field notes were deposited in the Raffles Museum, Singapore in 1941. One of Gibson-Hill's aims was to be able to compare the fauna of the Cocos (Keeling) Islands with that of Christmas Island. His other aim was to document the terrestrial and reef fauna of the Cocos (Keeling) Islands as it stood at that time. Unfortunately some of his notes and collections were lost during the wartime occupation of Singapore. Among the marine invertebrates the specimens and field notes of soft corals and anemones and most holothurians were lost but a copy of the field catalogue of the holothurians remained (Gibson-Hill 1950a, b). A.H. Clark (1950) described Gibson-Hill's echinoderm collection and included his very useful field notes. Clark notes that specimens of the family Linckiidae (Ophidiasteridae) were sent to Dr H. Engel of Amsterdam who was preparing a monograph on this family for the Siboga expedition reports. Unfortunately neither the “Linckiidae’ of the Siboga Expedition nor those of Cocos (Keeling) were published. ' The present collection numbers 82 species collected from 13 reef flat sites, nine outer slope sites and 13 lagoon sites (Chapter 1, Fig. 2). It consists of 2 species of Crinoidea, 15 Asteroidea, 17 Ophiuroidea, 14 Echinoidea and 34 Holothurioidea listed at the end of this report. Most are widespread Indo- West Pacific species but there are several westward extensions of range from Indonesia or Christmas Island and one south-eastward extension from Sri Lanka. When added to the species recorded by Clark (1950) the total known echinoderm fauna is now 89 species (4 crinoids, 17 asteroids, 17 ophiuroids, 17 echinoids and 34 holothurians). Crinoidea Clark (1950) noted that crinoids were rare on the accessible portions (reef platforms) of the Cocos (Keeling) Islands. From the present survey I can confirm this and note that they are also rare on the outer slopes. Crinoids were only collected at four sites, three on the outer slopes and one in the northern part of the lagoon. Only two species were represented, one of which was also recorded by A.H. Clark (1950). Colin (1977) notes that in five weeks of collecting (fishes) in 1974 only a single small crinoid was found. Western Australian Museum, Francis Street, Perth, Western Australia, 6000. The crinoid fauna is even more depauperate than that of Christmas Island where eight species were recorded (Marsh 1988). This compares with 38 species found at Ashmore Reef (Marsh et al. in press) and 17 at the Rowley Shoals and Scott Reef (Marsh 1986) off north-western Australia and five from the isolated Western Indian Ocean atoll of Aldabra (Sloan et al. 1979). Clark and Taylor (1971) did not record any crinoids from Diego Garcia in the central Indian Ocean. Coral reef crinoids have a short larval life (Mortensen 1938) and few species are widely distributed in the Pacific and Indian Oceans. The species found at the Cocos (Keeling) Islands are among those whose distribution extends from the Red Sea or Western Indian Ocean to the western Pacific Ocean, apart from Stephanometra spinipinna (recorded by A.H. Clark), which is known only from Indonesia and northern Australia. Asteroidea Clark (1950) recorded only four species of starfishes from Gibson-Hill's collection but these did not include any members of the family Ophidiasteridae (Linckiidae), the family generally best represented on coral reefs. Gibson-Hill collected 141 specimens of this family, which were sent to Engel in Amsterdam who was currently working on a collection of the same family from Indonesia. Unfortunately Engel did not complete either project. I had been told that the Cocos specimens were still at the Natural History Museum in Amsterdam but on a recent visit there the collection could not be found. Neither is it in the University Museum collection in Singapore (formerly the Raffles Museum Collection). There is therefore no historical record of Ophidiasterids from the Cocos (Keeling) Islands. The present collection contains 15 species of Asteroidea and a further two were recorded by Clark (1950). This is the same number as that recorded from the Rowley Shoals (Marsh 1986), however only 11 species are in common. If the Rowley Shoals, Scott Reef and Ashmore Reef are taken together, 15 species are in common with Cocos, the same number as are in common between Cocos and Indonesia; however, Ashmore Reef has a much richer fauna (28 species) including four Oreasterids, generally regarded as ‘continental’ species. Cocos has more asteroid species (17) than Christmas Island (13) with only eight species in common, probably due in part to the more extensive reef flats at Cocos. When compared with the isolated atoll of Aldabra, in the Western Indian Ocean only seven of Aldabra's 19 species are in common with the Cocos Islands. A single species, Culcita schmideliana, is recorded from Diego Garcia, in the central Indian Ocean (Clark and Taylor 1971). As at Christmas Island and Aldabra the small Linckia multifora, which reproduces asexually by autotomy, is the most common asteroid and is found in all habitats at Cocos (Keeling), from the outer slopes (6 sites) and reef flats (6 sites), where it is usually small, to the lagoon where exceptionally large individuals (for the species) were found at site 35. Most other species were found at only two or three sites but Ophidiaster granifer was unexpectedly found on the outer slope, on reef flats and in the lagoon. This species is usually confined to reef flats. Nardoa tuberculata, usually found in the open on reef flats, was only found in lagoon holes on coral rubble. Several species were only found at one site and it is possible that other uncommon species may be found in areas not sampled in this survey. Acanthaster planci (crown-of-thorns starfish) was found on the outer slope, on a reef flat and in the lagoon but was only seen at three sites. In a lagoon hole, south of Direction Island about 20 individuals were observed but the greatest number (> 50) were seen on the outer slope of Turk Reef (stn 15) at 10-45 m, where there was very little coral. The following records indicate that large populations of A. planci have been present more than once in the past. Clark (1950) quoted from Gibson-Hill's field notes stating that "A. planci is very Conspicuous on the atoll but not very common. It occurs among coral rocks near the low tide level over the centre and outer portions of the barrier. It is most plentiful on the north and east coasts of the atoll". In 1971 a former resident of Cocos (Keeling) reported to the Western Australian Museum that the reef off Ujong Tanjong, at the north end of West Island, which had been a flourishing coral reef in 1963 was, by the end of 1969, considerably damaged by A. planci predation and there was a very large population of young specimens from about 100 to 320 mm in diameter. She did not find any near Pulu Beras where Gibson-Hill had collected them in 1941. There is no record of any observations on the outer slopes during this period. Ms Anne Waldron, who collected echinoderms on the reef flats at a number of localities around West Island in January 1972, did not find any A. planci, but noted that residents of West Island said they were present but not in large numbers. During the course of an ichthyological survey of the Cocos reefs in 1974 Colin (1977) found extensive areas of dead coral on the outer slopes to a depth of 45 m which he attributed to Acanthaster predation. Large A. planci were abundant at depths of 15-30 m on the outer slopes, at a density of about 1 per 200-400 m2 but few small individuals were seen. He also noted ‘islands’ of living coral on some of the buttresses of the outer slope while adjacent areas were dead. It seems likely therefore that recurring high levels of Acanthaster predation are responsible for the low level of coral cover on the outer reef slopes. Of great interest was the finding of several specimens of Tegulaster ceylanicus, on the outer slope of the Home Island reef. This species was previously known only from Sn Lanka and the Lakshadweep Islands, although a congener has been collected on the Great Barrier Reef, Queensland. The two species differ slightly but, as both are described from single specimens it is not possible to determine whether they are variations of the one species. The collections made on the Cocos (Keeling) reefs have extended the known distribution of several species: Celerina heffernani for which the only previous Indian Ocean locality was Christmas Island (Western Australian Museum coll.); it is also known from Indonesia (Guille and Jangoux 1978) and the Western Pacific. Similarly, the range of Neoferdina cumingi, is extended westward from Christmas Island; published records are from the Central Pacific to Christmas Island (Jangoux 1973). Ophidiaster granifer has not previously been recorded from the Indian Ocean although there is a specimen in the Western Australian Museum collection from Madagascar; previous records are from the western Pacific to Indonesia. Nardoa tuberculata is a common species on reef flats in Indonesia and northern Australia but the only previous record from the Indian Ocean is Andrews (1900) record from Christmas Island. However, this species was not found there by the Western Australian Museum in 1987 and its occurrence may be sporadic. Ophiuroidea Clark (1950) recorded only eight species of Ophiuroids, most of these being large species common on the reef flats. The present collection numbers 17 species (including all the species recorded by Clark) a low number compared with Christmas Island (33) (Marsh 1988), the Rowley Shoals (28), Scott Reef (38) (Marsh 1986), Ashmore Reef (42) (Marsh et al. in press), Aldabra (39) (Sloan et al. 1979) and Diego Garcia (10) (Clark and Taylor 1971). The low number of species may to some extent reflect less collecting effort on the outer slopes than at Christmas Island. However, extensive sampling of lagoon and reef flat habitats including sand sifting, examination of weed mats and breaking up rocky substrate yielded very few small species. Large ophiocomids were common and in some cases abundant under boulders on the reef flats. Ophiocoma scolopendrina and Ophiomastix annulosa were the most abundant, often with 4-5 of the latter under almost every boulder. O. scolopendrina was found under boulders but also occupied crevices in the reef from which it extended 3-4 arms which turn upside down to sweep the surface scum on the incoming tide. Ophiocoma brevipes was moderately common among seagrasses on sandy areas of the reef flats while O. erinaceus was found under boulders on the mid and outer reef flats. O. anaglyptica was found on the mid and outer reef flats, exposed to surf. The only new record for the Indian Ocean is Ophiarachnella similis whose range is extended westward from Indonesia. Fifteen of the 17 species are in common with Christmas Island and all occur off north-western Australia and Indonesia. Eleven are in common with Aldabra and six with Diego Garcia. Echinoidea A.H. Clark (1950) recorded 15 species of echinoid of which we failed to find three, but added another two making a total of 17 species now known from the Cocos (Keeling) Islands. The species are all widespread throughout the Indo-West Pacific including north-western Australia. Twelve species are in common with Christmas Island which apparently lacks all but one of the sand-dwelling Brissids and the Clypeasterid but has several species on surf-swept rocky shores, not found at Cocos, giving it a total of 18 species. In comparison with north-western Australian reefs there are more species than at the Rowley Shoals, where 14 are recorded although 22 have been found on the Rowley Shoals and Scott Reef combined and 23 on Ashmore Reef. Fourteen of the species are in common with Aldabra which has a total of 31 species (Sloan et al. 1979) and eight of Diego Garcia's nine species are in common with Cocos (Keeling). The brissids and clypeasterid were found only in the South Passage area and no live specimens were taken apart from one freshly predated specimen of Metalia spatagus. Extensive observation of the sand flats at the south end of the lagoon failed to find any others. Clark (1950) reports that Gibson-Hill found brissids near passages on the eastern side of the atoll but we were unable to examine this area. Despite the extensive die-off of lagoon fauna in 1983 (see Woodroffe and Berry, this volume) we found Parasalenia gratiosa to be abundant under dead coral slabs in lagoon ) holes in the same habitat as that described by Gibson-Hill (1950). This was the only habitat in which this species was found. Holothurioidea As noted in the introduction, Gibson-Hill's collection of holothurians from the Cocos (Keeling) Islands was destroyed during World War II in Singapore. The present collection is thus the only record of holothurians from the islands. The Cocos (Keeling) Islands have a fairly rich fauna of holothurians, including most of the species used for trepang (Béche-de-mer). Thirty four species were collected compared with 16 at Christmas Island, 28 at Scott Reef/Rowley Shoals (Marsh 1986), 47 at Ashmore Reef (Marsh et al. in press) and 35 species at Aldabra (Sloan et al. 1979). Although the extensive sand flats in the lagoon might be regarded as suitable habitat for holothurians, the majority (30 species) were found either on reef flats or in sandy areas adjacent to reef flats, as at South Passage. Only four species were found on the outer slopes. Ten species were found at lagoon sites but only one of these (Synaptula recta) was not found in other habitats. All but three of the holothurians are species widespread in the Indo-West Pacific, the exceptions are Holothuria (Acanthotrapeza) coluber, H. (Metriatyla) aculeata and Chiridota rigida the range of which is extended westward from Indonesia and north- western Australia to the Cocos (Keeling) Islands. The zonation of common reef flat species near the settlement on West Island is shown in Figure 1. Trepang (Béche-de-mer species) Species of a large size with a thick body wall are the only ones suitable for processing for food. At the Cocos (Keeling) Islands seven species of commercial value have been found. No estimates of population size could be made in the time available but indications are given of the sites where the commercial species were most common. The most valuable species are the teat fish, Holothuria (Microthele) nobilis, and other species of the subgenus. H. (Microthele) spp. are nowhere common but individuals were seen or collected at five reef flat and two lagoon sites (List of echinoderms). Thelenota ananas (prickly red fish) is another large, valuable species but this was only found at one outer slope site. Other commercial species, their value depending to some extent on size, are two species of Actinopyga (A. echinites and A. mauritiana), both common to abundant on reef flats particularly at West Island; another commercial species A. miliaris may occur at Cocos but was not found during the survey. Several less valuable commercial species were also found: Bohadschia marmorata (chalky fish), Holothuria (Metriatyla) scabra (sand fish) and H. (Halodeima) atra. B. marmorata and H. scabra were found on the lagoon side of South Passage near Pulu Maria where the former was moderately common. B. marmorata was also seen in the lagoon south of Direction Island. H. atra is the most widespread of any species at Cocos and was common in all habitats but it is of very little commercial value unless individuals are of a very large size; it is a highly toxic species. Bohadschia argus (leopard or tiger fish), although of fairly large size, and common on some of the sandy reef flats has a very low commercial value partly because of 6 the toxic cuvierian tubules ejected when the animal is touched. None of the other species listed is believed to have any commercial value. Little is known of growth rates of commercial species and any attempt at exploitation of the populations should be carefully monitored and certain areas reserved from exploitation. Quantitative population studies need to be made of the potentially commercial species before any fishing takes place and on-going studies of recruitment and growth should be initiated. It should be noted that all the commercial species have water soluble toxins in the body wall and can only be eaten after correct preparation. ACKNOWLEDGEMENTS I am very grateful to Ms Anne Waldron who collected echinoderms for the Western Australian Museum from various localities during a visit to West Island in 1972 and to Ms Diana Applehof for her observations of Acanthaster at the Cocos (Keeling) Islands. REFERENCES Andrews, C.W. 1900. I Introductory note. In: Andrews, C.W., Smith, E.A., Bernard, H.M., Kirkpatrick, R. and F.C. Chapman, On the Marine fauna of Chirstmas Island (Indian Ocean). Proc. Zool. Soc. (Lond.) 1900: 115-117. Clark, A.H. 1950. Echinoderms from the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 53-67. Clark, A.M. and Taylor, J.D. 1971. Echinoderms from Diego Garcia. Atoll Res. Bull. 149: 89-92. Colin, P.L. 1977. The Reefs of Cocos-Keeling Atoll, eastern Indian Ocean. Proc. 3rd Coral Reef Symp.,1: 63-68. Gibson-Hill, C.A. 1950a. Introduction to papers on the fauna of the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 7-10. Gibson-Hill, C.A. 1950b. A note on the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 11-28. Guille, A. and Jangoux, M. 1978. Astérides et Ophiurides littorales de la region d'Amboine (Indonesie). Ann. Inst. Oceanogr., Paris 54: 47-74. Jangoux, M. 1973. Le genre Neoferdina Livingstone. Revue zool. afr., 87: 775-794. Marsh, L.M. 1986. Echinoderms, Part VI Faunal Surveys of the Rowley Shoals, Scott Reef and Seringapatam Reef, north-western Australia, ed. P.F. Berry. Rec. West. Aust. Mus. Suppl. No. 25: 63-74. ii Marsh, L.M. 1988. Echinoderms, Part VII of Survey of the Marine Fauna of Christmas Island, Indian Ocean, ed. P.F. Berry. Unpubl. Report to Aust. National Parks and Wildlife Service: 59-67. Marsh, L.M. Vail, L.L., Hoggett, A.K. and Rowe, F.W.E. in press. Echinoderms, In Survey of the Marine Fauna of Ashmore Reef, Australia, ed. P.F. Berry, Rec. West. Aust. Mus. Mortensen, T. 1938. Contributions to the study of the development of larval forms of echinoderms IV. D. Kgl. Danske Vidensk. Selsk. Skrifter, Naturv. og Math. Afd., 9, 7(3): 1-59. Sloan, N.A., Clark, A.M. and Taylor, J.D. 1979. The echinoderms of Aldabra and their habitats. Bull. Br. Mus. Nat. Hist. (Zool.) 37: 81-128. LIST OF ECHINODERMS KEY TO SYMBOLS: A.H. Clark, 1950 (C) species by another name Cc = + = Spec. from various sources in Western Australian Museum eg6 = Western Australian Museum 1989 station numbers V = Visual records x - New records # = Extension of distribution Echinodermata Previous Collection Records Station Crinoidea COMASTERIDAE * Comaster multifidus (Miiller, 1841) - - 15,19 MARIAMETRIDAE Stephanometra indica (Smith, 1876) C - - S. spicata (Carpenter, 1881) C - 13°23 S. spinipinna (Hartlaub, 1890) C - - Asteroidea OREASTERIDAE Culcita schmideliana (Retzius, 1805) (C) - TAI OPHIDIASTERIDAE #* Celerina heffernani (Livingstone, 1931) - . W235 * Cistina columbiae Gray, 1840 - - 25 * Dactylosaster cylindricus (Lamarck, 1816) - - P1224 “3 Fromia milleporella (Lamarck, 1816) - - 13,32 Linckia guildingi Gray, 1840 - - 4,19 3 L. laevigata (Linnaeus, 1758) - - 12,30V * L. multifora (Lamarck, 1816) - . 33,4.6.7, 1012.13.14. 15319,30;32,35 #* Nardoa tuberculata Gray, 1840 - - 29,36 * N. galatheae (Liitken, 1864) - - 7,9,19,23 #* Neoferdina cumingi (Gray, 1840) - - 13922 #* Ophidiaster granifer (Liitken, 1872) - - O. cribrarius Liitken, 1872 4,6,26,27,29,30 8 MITHRODIIDAE Mithrodia clavigera (Lamarck, 1816) C - - ASTERINIDAE #* Tegulaster ceylanicus (Déderlein, 1889) - - 33 ACANTHASTERIDAE Acanthaster planci (Linnaeus, 1758) C - 8,9V,15V ECHINASTERIDAE Echinaster luzonicus (Gray, 1840) C - - Ophiuroidea AMPHIURIDAE 3 Amphipholis squamata (Delle Chiaje, 1829) - - 24,35,37 OPHIACTIDAE x Ophiactis savignyi (Miiller and Troschel, - - 9,12,20,28,32,35 1842) OPHIOTRICHIDAE Macrophiothrix longipeda (Lamarck, 1816) C - 1,12,13,14,24,32 OPHIOCOMIDAE Ophiarthrum elegans Peters, 1851 C - 13 * Ophiocoma anaglyptica Ely, 1944 = + 1,12,14,24,27,30 a O. brevipes Peters, 1851 - + 1,5,10,14,24,27,30 O. dentata Miiller and Troschel, 1842 C + 1,10,12,13,14,20, 24,27 O. erinaceus Miiller and Troschel, 1842 C - 1,6,8,9,12,13V,20, 24,27,30,32 O. pica Miiller and Troschel, 1842 Cc - PSA 2732 z O. pusilla (Brock, 1888) - - 32 O. scolopendrina (Lamarck, 1816) C + 1,6,10,12,20,24,27 id O. schoenleini Miiller and Troschel, 1842 _ - - 9 5 Ophiocomella sexradia (Duncan, 1887) - + 3,20,24,35 Ophiomastix annulosa (Lamarck, 1816) C - 1,3V,12,20,24,27,30 OPHIONEREIDIDAE co Ophionereis porrecta Lyman, 1860 - - 9 OPHIODERMATIDAE #* Ophiarachnella similis (Koehler, 1905) - - 32 Ophiopeza spinosa (Ljungman, 1867) C - 14 Echinoidea CIDARIDAE Eucidaris metularia (Lamarck, 1816) C + UND S32 10 DIADEMATIDAE Diadema savignyi Michelin, 1845 Echinothrix calamaris (Pallas, 1774) E. diadema (Linnaeus, 1758) TEMNOPLEURIDAE Mespilia globulus (Linnaeus, 1758) TOXOPNEUSTIDAE Toxopneustes pileolus (Lamarck, 1816) Tripneustes gratilla (Linnaeus, 1758) PARASALENIIDAE Parasalenia gratiosa A. Agassiz, 1863 ECHINOMETRIDAE Echinometra mathaei (de Blainville, 1825) * Echinostrephus molaris (de Blainville, 1825) Heterocentrotus mammillatus (Linnaeus, 1758) ECHINONEIDAE Echinoneus cyclostomus Leske, 1778 CLYPEASTERIDAE Clypeaster reticulatus (Linnaeus, 1758) BRISSIDAE * Brissus latecarinatus (Leske, 1778) Metalia dicrana H.L. Clark, 1917 M. spatagus (Linnaeus, 1758) M. sternalis (Lamarck, 1816) Holothurioidea HOLOTHURIIDAE 2 Actinopyga echinites (Jaeger, 1833) * A. mauritiana (Quoy and Gaimard, 1833) od Bohadschia argus Jaeger, 1833 ES B. graeffei (Semper, 1868) = B. marmorata Jaeger, 1833 i Labidodemas semperianum Selenka, 1867 #* Holothuria (Acanthotrapeza) coluber Semper 1868 - H. (Cystipus) rigida (Selenka, 1867) * H. (Halodeima) atra Jaeger, 1833 * H. (H.) edulis Lesson, 1830 * H. (Lessonothuria) lineata Ludwig, 1875 QOe@ ee) Q@ee 1,12,30V 30 1,12,24V,27V,30V 16,29 13V,9512V-50 16,17V,36 1,12,24 PH 4, 12, 19 153°V5 12:24 VE27NG 30V 1,3V,5V,10V,12V, 20V,24,27V,30V 5V,8V,12,30V OV; 12 1 1,2,3V,5V,6V, 9V, 12V, 16, 18V, 19, 20V, 24V, 27V, 30V, 34V,36V,37V 9,16,19,30 3312 eS H. (L.) pardalis Selenka, 1867 * H. (Mertensiothuria) leucospilota (Brandt, 1835) * H. (Metriatyla) scabra Jaeger, 1833 #* H. (M.) aculeata Semper, 1868 * H. (Microthele) nobilis (Selenka, 1867) H. (M.) sp. - H. (Platyperona) difficilis Semper, 1868 * H. (Semperothuria) cinerascens (Brandt, 1835) cS H. (Stauropora) pervicax Selenka, 1867 a H. (Thymiosycia) hilla Lesson, 1830 * H. (T.) impatiens (Forskal, 1775) STICHOPODIDAE * Stichopus chloronotus Brandt, 1835 - S. horrens Selenka, 1867 * S. variegatus Semper, 1868 8 Thelenota ananas (Jaeger, 1833) PHYLLOPHORIDAE = Afrocucumis africana (Semper, 1868) SYNAPTIDAE i Euapta godeffroyi (Semper, 1868) S Opheodesoma grisea (Semper, 1868) “2 Synapta maculata (Chamisso and Eysenhardt 1821) * Synaptula recta (Semper, 1868) CHIRIDOTIDAE #* Chiridota rigida Semper, 1868 2 Polycheira rufescens (Brandt, 1835) 11 1,30 10,12 12 12 1,12,24,27,36V 12,14,23 1,24,27 1, 20, 24, 27 3 8,12,23,30V 8,9,12 1,3V,5V,6V,9, 12, 27V, 30V,34V,36V 12V,16 5 19 1,3,6,14,24,27 12 H. (Halodeima) atra (abundant) Actinopyga mauritiana Actinopyga echinites Stichopus chloronotus (few) H. (Microthele) nobilis (few) H. cinerascens H. cinerascens (abundant) Zonation of holothurians on reef flat near settlement, West Island. Figure 1. ATOLL RESEARCH BULLETIN NO. 412 CHAPTER 14 FISHES OF THE COCOS (KEELING) ISLANDS BY G.R. ALLEN AND W.F. SMITH-VANIZ ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 14 FISHES OF THE COCOS (KEELING) ISLANDS BY G.R. ALLEN * AND W.F. SMITH-VANIZ ** ABSTRACT Extensive fish collections were obtained at the Cocos (Keeling) Islands by the Academy of Natural Sciences of Philadelphia in 1973 and the Western Australian Museum in 1989. The combined collections document the occurrence of 533 species. For many Indo-west Pacific fishes (about 85 species), Cocos represents their westernmost limit of distribution. The faunal composition is typical of Indo-West Pacific coral reefs. Only 5 percent of the ichthyofauna consists of exclusively Indian Ocean species. The largest families are as follows (number of species in parentheses): Labridae (54), Gobiidae (51), Pomacentridae (38), Apogonidae (30), Serranidae (30), Acanthuridae (25), Muraenidae (24), Chaetodontidae (23), Blenniidae (21), Scaridae (20), and Holocentridae (20). Collectively these 11 families comprise 63 percent of the fauna. Little or no endemism is present. INTRODUCTION The first fishes collected at the Cocos (Keeling) Islands were taken by the crew of the, Beagle and subsequently reported by Jenyns (1842), who listed 10 species. Numerous fish records were recorded in a series of papers by Pieter Bleeker (1854-1859) that were published in Natuurkundig Tijdschrift voor Nederlandsch Indie. Fish specimens were sent to Bleeker from the Cocos Islands by A. J. Anderson and G. Clunies-Ross. The Cocos fauna was summarised by Bleeker (1859) as consisting of approximately 104 species. Most of the fishes reported by Bleeker were eventually deposited in the Rijksmuseum in Leiden. The only other major ichthyological collection from Cocos was made by C. A. Gibson-Hill, who visited the islands from January to November 1941. The fishes from this expedition were reported by Marshall (1950) and included 119 new records, thus raising the total species known from the island group to approximately 220. The Gibson- Hill collections are presently housed at the British Museum (Natural History). In 1973 an expedition from the Academy of Natural Sciences of Philadelphia (ANSP) collected marine organisms at Cocos (Keeling) during February and March. Team members, including ichthyologists Patrick L. Colin and William Smith-Vaniz, collected fishes at 68 stations. A variety of collection methods were employed: rotenone oe U.S. Fish and Wildlife Service, National Fisheries Research Center, 7920 N.W. 71st Street, Gainesville, Florida 32606, U.S.A. 2 (35 stations), explosives (17), quinaldine (14), angling (1), and spearing (31), usually as a supplemental means of collection. This effort resulted in a large fish collection containing 6,780 specimens in 1,443 lots. Approximately 425 species were obtained. A team of biologists from the Western Australian Museum and Australian Institute of Marine Sciences conducted a faunal survey at Cocos (Keeling) during February 1989. Fishes were collected mainly with rotenone (24 stations), supplemented by spear and dipnets (1 station each). These efforts yielded a total of 1,814 specimens, including 465 lots and approximately 245 species. In addition, underwater observations were conducted in which the name of each species seen was written on a waterproof sheet. These "spot" identifications were based on the senior author's extensive experience in the Indo-Pacific region. Only fishes whose identity was absolutely certain were recorded. This method provided an additional 203 records, thus a total of 448 species was noted. The combined ANSP-Western Australian Museum collections (and observations) include a total of 533 species. Many of the species reported by Bleeker (1859) and Marshall (1950) are junior synonyms or were based on misidentifications. It is beyond the scope of the present study to determine the current status of most of the species they listed. However, we estimate that only about 20 of their species represent additions not seen during the 1973 and 1989 visits. Therefore, the total known fish fauna of Cocos (Keeling) is approximately 550 species. The following species listed by Marshall (1950) almost certainly represent additions to the Cocos ichthyofauna: Albulidae - Albula vulpes (Linnaeus); Exocoetidae - Cheilopogon atrisignis (Jenkins) and C. cyanopterus (Valenciennes); Syngnathidae - Hippocampus trimaculatus Leach; Ostraciidae - Lactoria cornuta (Linnaeus). Some authors have questioned the providence of material Bleeker reported as originating from Cocos (Keeling), including several new species he described from there. Dawson (1982) cited the type locality of Micrognathus andersonii (Bleeker) as "Novaselma, Kokos [Cocos] Is. [Indonesia]," and stated (p. 677) that Marshall's (1950) listing of the species (as M. brevirostris) from Cocos-Keeling is based on a "misidentification of the type-locality...". We do not agree with Dawson's assertion that Bleeker's material did not originate from Cocos-Keeling. In the introduction of his first paper Bleeker (1854a) gave the correct coordinates for "Kokos-eilanden" and referred to J. C. Ross. (These islands were originally settled in 1827 by a Scottish sea captain named John Clunies-Ross.) In a later paper Bleeker (1858) thanked both Dr A. J. Anderson "geneesheer" [= physician] and J.G.C. Ross "tegeneewoordigen beheerder" [= present- day administrator] of Cocos Island. Through the kindness of Dr. Tyson Roberts we received a copy of a letter signed by A. J. Anderson (original deposited in the RMNH archives), with the heading "Cocos - July -1860," and addressed to Dr. Bleeker, concerning specimens of "trepang fish" [= Carapidae] that had been sent to Batavia. In the same letter Anderson asked to have Bleeker's European address (Bleeker returned to the Netherlands in September 1860) "in the event of my attaining other interesting specimens." No additional correspondence in the RMNH Bleeker files apparently exists concerning Cocos Is. (T. Roberts, in lit.). Presumably much of Bleeker's correspondence while he was in Batavia did not survive. In the absence of any evidence to the contrary, we conclude that all Bleeker's material stated to have come from "Kokos-eilanden" refers to Cocos-Keeling. Although we did not duplicate Bleeker's Cocos record of Micrognathus andersonii, it is a broadly distributed Indo-Pacific species and we have no reason to doubt its occurrence there. SPECIES COMPOSITION AND ZOOGEOGRAPHY The ichthyofauna of Cocos (Keeling) consists primarily of species that are associated with coral reefs. The largest families are summarised in Table 1. The eleven most speciose groups (Labridae, Gobiidae, Pomacentridae, Apogonidae, Serranidae, Acanthuridae, Muraenidae, Chaetodontidae, Blenniidae, Scaridae and Holocentridae) account for 63 percent of the total fauna. These families are typically abundant throughout the tropical Indo-Pacific region. Most of these 11 families contain fishes that are diurnally active which either dwell on or near the reef surface, or forage on plankton a short distance above it. Exceptions are the nocturnal apogonids and holocentrids, and the crevice- dwelling morays (muraenidae). The fish fauna is similar to that of Christmas Island, the nearest land mass, lying about 850 km to the northeast. Allen and Steene (1987) recorded 575 species from Christmas, of which about 350 species also occur at the Cocos group. The approximately 175 species that are present at Cocos, but lacking at Christmas are primarily inhabitants of the lagoon, a habitat that does not occur at Christmas Island. It is more difficult to explain the occurrence of approximately 210 species of Christmas fishes that are apparently absent from Cocos (Keeling). Two of the most notable disparities are shown by the Serranidae and Blenniidae in which 25 and 14 species, respectively, and 7 genera in each family were found only at Christmas. Perhaps this anomaly can be at least partly be explained by the greater isolation of Cocos (Christmas Island is only about 290 km from Java). Compared to other oceanic atolls the Cocos Group appears to have a relatively impoverished fish fauna. For example Randall and Randall (1988) and Winterbottom et al. (1989) recorded approximately 800 and 700 species, respectively, for the Marshall Islands in the central Pacific and Chagos Archipelago (incorporating several atolls) in the western Indian Ocean. There are probably several reasons for the diminished Cocos fauna including (1) its small physical size; (2) relative isolation and lack of surrounding "island stepping-stones;" and (3) lesser collecting activity. In addition, the extensive coral die- back at Cocos (Colin, 1977) may be responsible for the exclusion of some species. Although we believe the ichthyofauna of Cocos has been reasonably well sampled (except for epipelagic fishes), we certainly did not collect all of the species of shorefishes that occur there. That the fauna may not be as well sampled as we would like to believe is suggested by the fact that a number of shallow-water, cryptic species were taken at only one of our combined 59 rotenone stations. Scorpaenids are cryptic bottom dwellers (except Pterois spp.) yet, inexplicably, only 2 of 16 species we report from Cocos were collected or observed by both museum expeditions. A zoogeographic analysis of the Cocos fauna is presented in Table 2. The majority of fishes have distributions that cover relatively wide areas in the Indo-Pacific region. There is a much greater affinity to the Western Pacific than to the Indian Ocean region. Indeed, only about 5.1 percent of the species are Indian Ocean forms. There is no endemism in the Cocos fish fauna, although one anglefish, Centropyge joculator, is known only from Cocos and Christmas Island and an undescribed goby of the genus Trimma may have the same distribution. The Indian Ocean coasts of Java and Sumatra are poorly sampled, however, and it is likely that one or both species will be found there eventually. Scorpaenoides keelingensis Marshall which, as the name suggests, Marshall (1950) believed to be endemic to the Cocos group is almost certainly a junior synonym of Scorpaenoides kelloggi (Jenkins) (W. N. Eschmeyer, pers. comm.). Nearly all of the reef 4 fishes found at Cocos are characterised by a pelagic larval stage of up to several weeks duration. Hence, the widespread nature of the individual species distributions. ACKNOWLEDGMENTS We thank the following individuals for asisting us in obtaining the correct scientific names for species in this checklist: Eugenia B. Béhlke, Bruce B. Collette, William N. Eschmeyer, Ross W. Feltes, Thomas H. Fraser, Douglass F. Hoese, W. Holleman, Theodore W. Pietsch, John E. Randall, Jeffrey T. Williams, Richard W. Winterbottom, Thosaporn Wongratana, and David J. Woodland. Curatorial assistance at the Academy of Natural Sciences was provided by Eugenia B. Bohlke and William G. Saul, and at the Western Australian Museum by Kevin Smith. REFERENCES Allen, G.R. and Steen, R.C. 1987. Fishes of Christmas Island, Indian Ocean. Christmas Island Natural History Association, 197 pp. Baldwin, C. and G.D. Johnson. 1993. Phylogeny of the Epinephelinae (Teleostei: Serranidae). Bull. Mar. Sci. 52 (1): 240-283. Bleeker, P. 1854a. Bijdrage tot de kennis der ichthyologischefauna van de Kokos- eilanden. Nat. Tijdsch. Ned. Ind. 7: 37-48. Bleeker, P. 1854b. Over eenige nieuwe van de Kokos-eilanden. Nat. Tijdsch. Ned. Ind. 7: 353-358. Bleeker, P. 1855a. Derde Bijdrage tot de kennis der ichyologische fauna van de Kokos- eilanden. Nat. Tijdsch. Ned. Ind. 8: 169-180. Bleeker, P. 1855b. Vierde Bijdrage tot de kennis der ichthyologische fauna van de Kokos- eilanden. Nat. Tijdsch. Ned. Ind. 8: 445-460. Bleeker, P. 1858. Vijfde bijdrage tot de kennis der ichthyologische fauna van de Kokos- eilanden. Nat. Tijdschr. Ned. Ind. 15: 457-468. Bleeker, P. 1859. Tiental vischsoorten van de Kokos-eilanden verzameled door Dr Anderson. Nat. Tijdschr. Ned. Ind. 20: 142-143. Burgess, W.E. 1973. Salts from the seven seas. Tropical Fish Hobbyist 21 [May]: 37-38, 40-41. Colin, P.L. 1977. The reefs of Cocos-Keeling Atoll, eastern Indian Ocean. Proc. 3rd International Coral Reef Symposium, Univ. of Miami, Miami, Florida: 63-68. Dawson, C.E. 1982. Review of the genus Micrognathus Duncker (Pisces: Syngnathidae), with description of M. natans, n. sp. Proc. Biol. Soc. Wash. 95 (4):657-687. Jenyns, L. 1842. Fish Jn The Zoology of the voyage of H.M.S. Beagle, under the command of Capt. Fitzroy R.M. during the years 1832 to 1836. Part IV. London. 172 pp. Marshall, N.B. 1950. Fishes from the Cocos-Keeling Islands. Bull. Raffles Mus. 22: 166-205. Randall, J.E. 1991. Revision of Indo-Pacific groupers (Perciforms: Serranidae: Epinephelinae), with descriptions of five new species. Indo-Pacific Fishes 20: 1- 332} Randall, J.E. and Hoese, D.F. 1985. Revision of the Indo-Pacific dartfishes, genus Ptereleotris (Perciformes: Gobioidei). Indo-Pacific Fishes, No. 7: 1-36. Randall, J.E. and Randall, H.A. 1987. Annotated checklist of the fishes of Enewetak Atoll and other Marshall Islands. Jn The Natural History of Enewetak Atoll. Vol. 2 (Devaney, D.M., Reese, E.S., Burch, B.L. and Helfrich, P., editors) U.S. Dept.of Eneregy Office of Scientific and Technical Information. Schultz, E.T. 1986. Pterois volitans and Pterois miles: Two valid species. Copeia 1986 (3): 686-690. Winterbottom, R.W., Emery, A.R. and Holm, E. 1989. An annotated checklist of the fishes of the Chagos Archipelago Central Indian Ocean. Life Sciences Contrib. 145 Royal Ontario Mus.: 1-226. CHECKLIST OF COCOS (KEELING) FISHES The following list includes fishes that were either collected or observed during the 1973 and 1989 surveys. Asterisk or numbers preceding species names indicate the following distributional data: * = also known from Christmas Island; 1 = widespread Indo- Pacific or Indo-west Pacific; 2 = West Pacific species that reach their western distributional limit at Cocos (Keeling); 3 = Indian Ocean species (may include western extremity of west Pacific); 4 = Circumtropical or cosmopolitan; 5 = uncertain extralimital distribution; Square brackets appearing after each species citation indicate that specimens are deposited at the Academy of Natural Sciences, Philadelphia [P], the Western Australian Museum [W], or were observed only [O]. Sphyrnidae - Hammerhead sharks 4 *Sphyrna lewini (Grffith and Smith, 1834) [O] Carcharhinidae - Requiem sharks 1 *Carcharhinus amblyrhynchos (Bleeker, 1856) [P] 1 *C. melanopterus (Quoy and Gaimard, 1824) [O] 4 *Galeocerdo cuvier (Peron and LeSueur, 1822) [O] 1 *Triaenodon obesus (Riippell, 1837) [O] Mobulidae - Manta rays 4 *Manta birostris (Donndorff, 1798) [O] Moringuidae - Worm eels 1 Moringua ferruginea Bliss, 1883 [W] 1 *M. javanica (Kaup, 1856) [W] 1 *M. microchir Bleeker, 1853 [P,W] Chlopsidae - False morays 3 *Kaupichthys n. sp. [K. Tighe, pers. comm., 1993] [P,W] Muraenidae - Moray eels 2 Anarchias cantonensis (Schultz, 1943) [P] 1 *A. seychellensis Smith, 1962 [P,W] 1 *Echidna nebulosa (Ahl, 1789) [P] 1 *E. polyzona (Richardson, 1844) [P] 2 *Enchelycore bayeri (Schultz, 1953) [W] 1 *E. pardalis (Temminck and Schlegel, 1842) [P,W] 2 *Enchelynassa canina (Quoy and Gaimard, 1824) [P] 1 *Gymnothorax enigmaticus McCosker and Randall, 1982 1 *G. buroensis (Bleeker, 1857) [P] 1 G. fimbriatus (Bennett, 1831) [P,W] 1 *G. flavimarginatus (Riippell, 1830) [P,W] 1 *G. javanicus (Bleeker, 1859) [O] 1 *G. margaritophorus Bleeker, 1865 [P,W] 1 *G. melatremus Schultz, 1953 [P] 1 G. monochrous Bleeker, 1864 [P] 1 *G. monoStigma (Regan, 1909) [P] 1 *G. rueppelliae (McClelland, 1845) [P,W] 1 G. undulatus (Lacepéde, 1803) [P,W] 1 *G. zonipectus Seale, 1906 [P,W] 1 *Siderea picta (Ahl, 1789) 1 *S. thrysoidea (Richardson, 1845) 1 *Uropterygius concolor (Riippell, 1838) 1 *U. marmoratus (Lacepéde, 1803) 1 *U. xanthopterus Bleeker, 1859 Ophichthidae - Snake eels 2 Callechelys catostomus (Bloch and Schneider, 1801) 1 *Leiuranus semicinctus (Lay and Bennett, 1839) 2 Muraenichthys brevis Ginther, 1876 1 *M. laticaudata (Ogilby, 1897) 1 M. macropterus Bleeker, 1857 2 Schultzidia johnstonensis (Schultz and Woods, 1949) Congidae - Conger eels 1 *Conger cinereus Riippell, 1830 3 Gorgasia maculata Klausewitz and Eibl-Eibesfeldt, 1959 1 *Heteroconger hassi (Klausewitz and Eibl-Eibesfeldt, 1959) Clupeidae - Herrings 1 Sardinella melanura (Cuvier, 1829) 1 Spratelloides delicatulus (Bennett, 1831) Synodontidae - Lizardfishes 1 *Saurida gracilis (Quoy and Gaimard, 1824) 1 *Synodus englemani Schultz, 1953 Chanidae - Milkfishes 1 Chanos chanos (Forsskal, 1775) Ophidiidae - Cusk-eels 1 *Brotula multibarbata Temminck and Schlegel, 1846 Bythitidae - Viviparous brotulas 1 *Brosmophyciops pautzkei Schultz, 1960 5 *Ogilbia sp. Antennariidae - Anglerfishes 1 *Antennarius coccineus (Lesson, 1831) 1 A. dorehensis Bleeker, 1859 Notocheiridae [=Isonidae] - Surf spites 3 Iso natalensis Regan, 1919 Hemirhamphidae - Halfbeaks 1 Hyporhamphus affinis (Giinther, 1866) 1 H. dussumieri (Valenciennes, 1847) 2 Zenarchopterus dispar (Valenciennes, 1847) Belonidae - Needlefishes 1 Platybelone argalus platyura (Bennett, 1831) 1 *Tylosurus crocodilus (Peron and LeSueur, 1821) 8 Holocentridae - Squirrelfishes 1 Myripristis adusta Bleeker, 1853 [P,W] 1 *M. berndti Jordan and Evermann, 1903 [W] 1 M. chryseres Jordan and Evermann, 1903 [P] 1 *M. kuntee Cuvier, 1831 [W] 1 *M. murdjan (ForsskAl, 1775) [P] 1 *M. parvidens Cuvier, 1829 [P] 1 M. pralinia Cuvier, 1829 [W] 1 *M. vittata Valenciennes, 1831 [Ww] 1 M. violaceus Bleeker, 1851 [O] 1 Neoniphon argenteus (Valenciennes, 1831) [P,W] 1 N. opercularis (Valenciennes, 1831) [O] 1 N.sammara (Forsskal, 1775) [P,W] 1 *Plectrypops lima (Valenciennes, 1831) [P,W] 1 *Sargocentron diadema (Lacepéde, 1801) [P,W] 3 *S. lepros (Allen and Cross, 1983) [W] 1 *S. microstoma (Giinther, 1859) [W] 1 *S. caudimaculatum (Riippell, 1838) [W] 1 *S. punctatissimum (Cuvier, 1829) [P] 1 *S. tiere (Cuvier, 1829) [P] 1_ S. spiniferum (Forsskal, 1775) [P] Aulostomidae - Trumpetfishes 1 *Aulostromus chinensis (Linnaeus, 1766) [P] Fistulariidae - Cornetfishes 1 *Fistularia commersonii Riippell, 1838 [O] Syngnathidae - Pipefishes 1 *Choeroichthys sculptus (Giinther, 1870) [P,W] 1 Corythoichthys flavofasciatus (Riippell, 1838) [P,W] 1 *Cosmocampus banneri (Herald and Randall, 1972) [W] 1 *Doryrhamphus excisus excisus Kaup, 1856 [P,W] 2 *Micrognathus brevirostris pygmaeus Fritzsche, 1981 [W] 1 Phoxocampus belcheri (Kaup, 1856) [W] Scorpaenidae - Scorpionfishes 1 Parascorpaena mossambica (Peters, 1855) [P] 1 *Pterois antennata (Bloch, 1787) [W] 1 *P. radiata Cuvier, 1829 [P,W] 2 *P. volitans (Linnaeus, 1758) [We follow Schultz (1986) in recognizing Pierois miles Bennett as an Indian Ocean species distinct from the Pacific P. volitans.] 1 *Scorpaenodes albaiensis (Evermann and Seale, 1907) [P] 1 *S. guamensis (Quoy and Gaimard, 1824) [P] 1 *S. hirsutus (Smith, 1957) [P,W] 1 S. kelloggi (Jenkins, 1903) [P] 1_ S. littoralis (Tanaka, 1917) [P] 1 S. minor (Smith, 1958) [P] 1 *§. parvipinnis (Garrett, 1863) [P] 1 *Scorpaenopsis diabolus (Cuvier, 1829) [O] 1 *Sebastapistes cyanostigma (Bleeker, 1856) [P] 1 *S. strongia (Cuvier, 1829) [P,W] 5 Sebastapistes sp. [P] 1 *Synanceia verrucosa Bloch and Schneider, 1801 [P] Platycephalidae - Flatheads 1 *Thysanophrys otaitensis (Cuvier, 1829) [P] Caracanthidae - Orbicular velvetfishes 2 *Caracanthus maculatus (Gray, 1831) [P] 1 *C. unipinna (Gray, 1831) [P] Serranidae - Sea basses [We follow Baldwin and Johnson (1993) in including the Grammistidae and Pseudogrammidae in this family.] 1 *Anyperodon leucogrammicus (Valenciennes, 1828) [P]} 1 *Cephalopholis argus Bloch and Schneider, 1801 [P,W] 1 *C. leopardus (Lacepéde, 1801) [P,W] 1 *C. polleni (Bleeker, 1868) [P] 1 *C. spiloparaea (Valenciennes, 1828) [P,W] 1 *C. urodeta (Valenciennes, 1828) [P,W] [Randall (1991 p.70) noted that this species consists of two allopatric color forms, the western Indian Ocean C. nigripinnis (Valenciennes) and the nominal Pacific form; he regarded them as conspecific because Christmas Is. specimens have somewhat intermediate color patterns. The color pattern of Cocos specimens agrees well with the Pacific form.] 3 Epinephelus faveatus (valenciennes, 1828) [W] 1 E. fuscoguttatus (Forsskal, 1775) [O] 1 *E. hexagonatus (Bloch and Schneider, 1801) [W] 2 E. maculatus Bloch, 1790 [O} 1 *E. merra Bloch, 1793 [P,W] 1 E. macrospilus (Bleeker, 1855) [P] [Randall (1991 p.187) noted that this species consists of two allopatric color forms, the western Indian Ocean E. cylindricus Giinther said to differ from the Pacific and eastern Indian Ocean E. macrospilos by larger and more closely spaced spots. Because only spot size appeared to distinguish the two forms they were considered to be conspecific. ] 1 E.microdon (Bleeker, 1856) [P] 1 *E. spilotoceps Schultz, 1953 [P,W] 1 *E. tauvina (Forsskal, 1775) [P,W] 1 *Gracila albomarginata (Fowler and Bean, 1930) [P] 1 *Grammistes sexlineatus (Thunberg, 1792) [P,W] 5 *Luzonichthys sp. [O] 2 *Plectranthias nanus Randall, 1980 [P,W] 1 Plectropomus areolatus Riippell, 1828 [P] 2 P. leopardus (Lacepéde, 1802) [P] 2 P. maculatus (Bloch, 1790) [P] 1 Pseudanthias cooperi (Regan, 1902) [P,W] 3 *P. evansi Smith, 1954 [P,W] 2 *P. smithvanizi (Randall and Lubbock, 1981) [P,W] 5 Pseudanthias sp. [P] [Winterbottom et al. (1989) give color photographs (plates IVE,F) of this species, which they report as Anthias sp. from the Chagos Archipelago. ] 2 Pseudogramma bilinearis (Schultz, 1943) [P] 1 *P. polyacantha (Bleeker, 1856) [P,W] 10 2 *Suttonia lineata Gosline, 1960 1 *Variola louti (Forsskal, 1775) Pseudochromidae - Dottybacks 1 Pseudoplesiops n. sp. 2 P.multisquamatus Allen, 1987 Plesiopidae - Longfins 1 *Plesiops coeruleolineatus Riippell, 1835 2 *P. corallicola Bleeker, 1853 Kuhliidae - Flagtails 1 *Kuhlia mugil (Bloch and Schneider, 1801) Priacanthidae - Bigeyes 1 *Heteropriacanthus cruentatus (Lacepéde, 1801) Apogonidae - Cardinalfishes 1 *Apogon angustatus (Smith and Radcliffe, 1911) 2 A. bandanensis Bleeker, 1854 . crassiceps Garman, 1903 . cyanosoma Bleeker, 1853 dispar Fraser and Randall, 1976 . evermanni Jordan and Snyder, 1904 . exostigma (Jordan and Starks, 1906) guamensis Valenciennes, 1832 . kallopterus Bleeker, 1856 . leptacanthus Bleeker, 1856 melas Bleeker, 1848 . novemfasciatus Cuvier, 1828 . taeniophorus Regan, 1908 . taeniopterus (Bennett, 1835) *Apogonichthys ocellatus (Weber, 1913) 1 A. perdix Bleeker, 1854 1 *Cercamia eremia (Allen, 1987) 1 Cheilodipterus lineatus Cuvier, 1828 2 *C. macrodon (Lacepéde, 1802) 1 *C. quinquelineatus Cuvier, 1828 1 *Fowleria aurita (Valenciennes, 1831) 1 F. isostigma (Jordan and Seale, 1906) 1 F. variegata (Valenciennes, 1832) 2 Gymnapogon urospilotus Lachner, 1953 1 Neamia octospina Smith and Radcliffe, 1912 1 Pseudamia gelatinosa Smith, 1955 2 *Pseudamiops gracilicauda (Lachner, 1953) 1 Rhabdamia gracilis (Bleeker, 1856) 2 Siphamia majimae Matsubara and Iwai, 1959 2 Sphaeramia nematoptera (Bleeker, 1856) * * fb be he Se oe 1 1 2 1 1 1 1 1 2 Pax is 1 1 Malacanthidae - Tilefishes 1 *Malacanthus brevirostris Guichenot, 1848 1 *M. latovittatus (Lacepéde, 1801) Carangidae - Trevallies 1 *Carangoides ferdau (Forssk§l, 1775) 1 *C. orthogrammus (Jordan and Gilbert, 1882) 1 *Caranx ignobilis (Forsskal, 1775) 4 *C. lugubris Poey, 1860 1 *C. melampygus Cuvier, 1833 1 *C. sexfasciatus Quoy and Gaimard, 1825 4 *Decapterus macarellus (Cuvier, 1833) 4 *Elagatis bipinnulatus (Quoy and Gaimard, 1825) 1 *Scomberoides lysan (Forsskal, 1775) 1 *Trachinotus bailloni (Lacepéde, 1801) 1 T. blochii (Lacepéde, 1801) Lutjanidae - Snappers 1 *Aphareus furca (Lacepéde, 1802) 1 *Aprion virescens Valenciennes, 1830 1 *Lutjanus bohar (Forsskal, 1775) 1 *L. fulvus (Bloch and Schneider, 1801) 1 *L. gibbus (Forsskal, 1775) 1 *L. kasmira (Forsskal, 1775) 1 L. monostigma (Cuvier, 1828) 1 *Macolor niger (Forsskal, 1775) Caesionidae - Fusiliers 1 *Caesio teres Seale, 1906 1 C. xanthonota Bleeker, 1853 1 *Pterocaesio lativattata Carpenter, 1987 1 *P. tile (Cuvier, 1830) Haemulidae - Sweetlips 2 Plectorhinchus chaetodontoides Lacepéde, 1800 Lethrinidae - Emperors 1 *Gnathodentex aurolineatus (Lacepéde, 1802) Gymnocranius grandoculis (Valenciennes, 1830) Lethrinus atkinsoni Seale, 1909 L. harak (Forsskal, 1775) L. hypselopterus Bleeker, 1873 L. lentjan (Lacepéde, 1802) L. microdon Valenciennes, 1830 L. obsoletus (Forsskal, 1775) L. xanthochilus Klunzinger, 1870 *Monotaxis grandoculis (Forsskal, 1775) pee ee LD Nemipteridae - Threadfin breams 2 Scolopsis lineatus (Quoy and Gaimard, 1824) Gerreidae - Mojarras 1 Gerres acinaces Bleeker, 1854 Mullidae - Goatfishes 1 *Mulloides flavolineatus (Lacepéde, 1801) 1 *M. vanicolensis (Valenciennes, 1831) ql 12 1 Parupeneus barberinus (Lacepéde, 1801) 1 *P. bifasciatus (Lacepéde, 1801) 1 *P. cyclostomus (Lacepéde, 1801) 1 *P. macronemus (Lacepéde, 1801) 2 *P. multifasciatus (Quoy and Gaimard, 1824) 1 *P. pleurostigma (Bennett, 1831) Kyphosidae - Rudderfishes 1 *Kyphosus cinerascens (Forssk§l, 1775) 1 K. vaigiensis (Quoy and Gaimard, 1825) Pempheridae - Sweepers 1 *Pempheris oualensis Cuvier, 1831 Ephippidae - Batfishes 1 *Platax orbicularis (Forsskal, 1775) 1 *P. teira (Forsskal, 1775) Chaetodontidae - Butterflyfishes 1 *Chaetodon auriga (Forsskal, 1775) 1 C. bennetti Cuvier, 1831 1 *C. citrinellus Cuvier, 1831 1 C. ephippium Cuvier, 1831 3 *C. guttatissimus Bennett, 1831 1 *C. kleinii Bloch, 1790 1 *C. lineolatus Cuvier, 1830 1 *C. lunula (Lacepéde, 1803) 3 *C. madagaskariensis Ahl, 1923 1 *C. melannotus Bloch and Schneider, 1801 1 *C. meyeri Bloch and Schneider, 1801 3 *C. mitratus Giinther, 1860 1 *C. ornatissimus Cuvier, 1831 1 *C. semeion Bleeker, 1855 1 *C. trifascialis Quoy and Gaimard, 1824 1 *C. trifasciatus Park, 1797 2 C. ulietensis Cuvier, 1831 1 *C. unimaculatus Bloch, 1787 1 *C. vagabundus Linnaeus, 1758 1 *Forcipiger flavissimus Jordan and McGregor, 1898 2 *Hemitaurichthys polylepis (Bleeker, 1857) 1 Heniochus chrysostomus Cuvier, 1831 1 *H. monoceros Cuvier, 1831 Pomacanthidae - Angelfishes 1 *Apolemichthys trimaculatus (Lacepéde, 1831) 2 Centropyge colini Smith-Vaniz and Randall, 1974 1 *C. flavissimus (Cuvier, 1831) 3 *C. joculator Smith-Vaniz and Randall, 1974 2 C. multifasciatus (Smith and Radcliffe, 1911) 2 Genicanthus bellus Randall, 1975 1 *Pomacanthus imperator (Bloch, 1787) Pomacentridae - Damselfishes 1 *Abudefduf notatus (Day, 1869) 1 *A. septemfasciatus (Cuvier, 1830) 1 *A. sordidus (Forsskal, 1775) 1 *A. vaigiensis (Quoy and Gaimard, 1825) 2 *Amblyglyphidodon aureus (Cuvier, 1830) 2 A. curacao (Bloch, 1787) 1 *Amphiprion clarkii (Bennett, 1830) 2 *A. perideraion Bleeker, 1855 2 *Chromis alpha Randall, 1988 2 *C. amboinensis (Bleeker, 1873) 2 *C. atripes Fowler and Bean, 1928 2 *C. caudalis Randall, 1988 2 *C. delta Randall, 1988 3 *C. dimidiata (Klunzinger, 1871) 1 *C. elerae Fowler and Bean, 1928 1 *C. lepidolepis Bleeker, 1877 2 *C. margaritifer Fowler, 1946 3 *C. nigrura Smith, 1960 3 *C. opercularis (Giinther, 1867) 1 *C. ternatensis (Bleeker, 1856) 1 C. viridis (Cuvier, 1830) 2 *C. xanthura (Bleeker, 1854) 1 Chrysiptera biocellata (Quoy and Gaimard, 1824) 1 *C. glauca (Cuvier, 1830) 1 Dascyllus aruanus (Linnaeus, 1758) 2 *D. reticulatus (Richardson, 1846) 1 *D. trimaculatus (Riippell, 1828) 1 *Plectroglyphidodon dickii (Liénard, 1839) 1 *P. imparipennis (Vallant and Sauvage, 1875) 1 *P. johnstonianus Fowler and Ball, 1924 1 *P. lacrymatus (Quoy and Gaimard, 1825) 1 *P. leucozonus (Bleeker, 1859) 1 *P. phoenixensis (Schultz, 1943) 1 Pomacentrus pavo (Bloch, 1787) 1 *Stegastes albifasciatus (Schlegel and Miiller, 1839) 1 *S. fasciolatus (Ogilby, 1889) 1_ S. lividus (Bloch and Schneider, 1801) 1 S. nigricans (Lacepéde, 1802) Cirrhitidae - Hawkfishes 1 *Amblycirrhitus bimacula (Jenkins, 1903) 2 Cirrhitichthys aprinus (Cuvier, 1829) 1 *C. oxycephalus (Bleeker, 1855) 1 *Cirrhitus pinnulatus (Schneider, 1801) 1 *Oxycirrhites typus Bleeker, 1857 1 *Paracirrhites arcatus (Cuvier, 1829) 1 *P. forsteri (Schneider, 1801) 2 *P. hemistictus (Giinther, 1874) Musgilidae - Mullets 1 *Crenimugil crenilabis (Forsskal, 1775) 1 Liza vaigiensis (Quoy and Gaimard, 1824) 13 14 Sphyraenidae - Barracudas 4 *Sphyraena barracuda (Walbaum, 1792) 1 *S. flavicauda Riippell, 1838 Polynemidae - Threadfins 1 Polydactylus sexfilis (Valenciennes, 1831) Labridae - Wrasses 1 *Anampses caeruleopunctatus Riippell, 1829 1 *A. meleagrides Valenciennes, 1840 1 *A. twistii Bleeker, 1856 1 *Bodianus anthioides (Bennett, 1830) 1 *B. axillaris (Bennett, 1831) 1 *B. diana (Lacepéde, 1801) 1 Cheilinus bimaculatus Valenciennes, 1840 1 C. chlorurus (Bloch, 1791) 1_ C. fasciatus (Bloch, 1791) 1 *C. trilobatus Lacepéde, 1801 1 *C. undulatus Riippell, 1835 2 *C. unifasciatus Streets, 1877 1 *Cheilio inermis (Forsskal, 1775) 1 *Cirrhilabrus exquisitus Smith, 1957 2 Cirrhilabrus rubrimarginatus Randall, 1992 1 *Coris aygula Lacepéde,1801 2 *C. dorsomacula Fowler, 1908 1 *C. gaimard (Quoy and Gaimard, 1824) 1 Cymolutes praetextatus (Quoy and Gaimard, 1834) 1 *Epibulus insidiator (Pallas, 1770) 2 *Gomphosus varius Lacepéde, 1801 2 Halichoeres chloropterus (Bloch, 1791) 1 H. hortulanus (Lacepéde, 1801) 1 *H. marginatus Riippell, 1835 2 *H. melasmapomus Randall, 1980 2 *H. ornatissimus (Garrett, 1863) 1 *H. scapularis (Bennett, 1831) 2 *H. trimaculatus (Quoy and Gaimard, 1834) 1 *Hemigymnus fasciatus (Bloch, 1792) 1 *H. melapterus (Bloch, 1791) 1 *Hologymnosus doliatus (Lacepéde, 1801) 1 *Labroides bicolor Fowler and Bean, 1928 1 *L. dimidiatus (Valenciennes, 1839) 2 *L. pectoralis Randall and Springer, 1975 1 *Labropsis xanthonota Randall, 1981 2 Macropharyngodon meleagris (Valenciennes, 1839) 1 Novaculichthys macrolepidotus (Bloch, 1791) 1 *N. taeniourus (Lacepéde, 1801) 1 *Pseudocheilinus hexataenia (Bleeker, 1857) 1 *P. octotaenia Jenkins, 1900 2 Pseudocoris aurantifasciatus Fourmanoir, 1971 1 *Pseudodax moluccanus (Valenciennes, 1839) 2 *Stethojulis bandanensis (Bleeker, 1851) 1 *S. strigiventer (Bennett, 1832) 1 *Thalassoma amblycephalum (Bleeker, 1856) 1 *T. hardwickei (Bennett, 1828) 1 *T. jansenii (Bleeker, 1856) 1 *T. lunare (Linnaeus, 1758) 1 *T. lutescens (Lay and Bennett, 1839) 1 *T. purpureum (Forsskal, 1775) 1 *T. quinquevittatum (Lay and Bennett, 1839) 1 *T. trilobatum (Lacepéde, 1801) 2 Xyrichtys aneitensis (Ginther, 1862) 1 *X. pavo Valenciennes, 1840 Scaridae - Parrotfishes 1 *Bolbometopon muricatum (Valenciennes, 1840) Calotomus carolinus (valenciennes, 1840) C. spinidens (Quoy and Gaimard, 1824) Hipposcarus longiceps (Valenciennes, 1840) Leptoscarus vaigiensis (Quoy and Gaimard, 1824) Scarus atropectoralis Schultz, 1958 S. enneacanthus Lacepéde, 1802 *§. forsteni (Bleeker, 1861) 1 *S. frenatus Lacepéde, 1802 1 *S. ghobban Forsskal, 1775 1 S. globiceps Valenciennes, 1840 1 *S. niger Forsskal, 1775 2 *S. oviceps Valenciennes, 1840 1 *S. prasiognathos Valenciennes, 1840 1 *S. psittacus Forsskal, 1775 1 *S. rubroviolaceus Bleeker, 1847 2 *S. schlegeli (Bleeker, 1861) 1 *S. sordidus Forsskal, 1775 3. S. strongylocephalus Bleeker, 1854 NWNRNR [This species, restricted to the Indian Ocean and Indonesia, has frequently been misidentified as S. gibbus Riippell, a closely related Red Sea endemic. ] 3 S. viridifuratus (Smith, 1956) Pinguipedidae - Sandperches 1 *Parapercis clathrata Ogilby, 1911 1 P. hexophthalma (Cuvier, 1829) 1 *P. schauinslandi (Steindachner, 1900) Creediidae - Sandburrowers 2 *Chalixodytes tauensis Schultz, 1943 3. Limnichthys nitidus Smith, 1958 Tripterygiidae - Triplefins 3 *Enneapterygius elegans (Peters, 1876) 5 *Enneapterygius tutuilae Jordan & Seale, 1906 5 *Enneapterygius sp. 1 2 Helcogramma capidata Rosenblatt, 1960 Blenniidae - Blennies 1 Aspidontus dussumieri (Valenciennes, 1836) 1 *A. taeniatus Quoy and Gaimard, 1834 1 *Cirripectes castaneus (Valenciennes, 1836) [P] 16 3 C. gilberti Williams, 1988 1 *C. polyzona (Bleeker, 1868) 1 C. quagga (Fowler and Ball, 1924) 1 *Escenius bicolor (Day, 1888) 1 *E. midas Starck, 1969 2 *Entomacrodus caudofasciatus (Regan, 1909) 1 *E. epalzeocheilus (Bleeker, 1859) 1 E. striatus (Quoy and Gaimard, 1836) 1 *Exallias brevis (Kner, 1868) 1 Glyptoparus delicatulus Smith, 1959 2 *Istiblennius chrysospilos (Bleeker, 1857) 1 *7. edentulus (Schneider, 1801) 1 */. lineatus (Valenciennes, 1836) 1 *J. periophthalmus (Valenciennes, 1836) 1 Petroscirtes xestus Jordan and Seale, 1906 1 *Plagiotremus rhinorhynchos (Bleeker, 1852) 1 *P. tapeinosoma (Bleeker, 1857) 1 Stanulus seychellensis Smith, 1959 Callionymidae - Dragonets 1 Diplogrammus goramensis (Bleeker, 1858) Gobiidae - Gobies 2 Amblygobius decussatus (Bleeker, 1855) 2 A. phalaena (Valenciennes, 1837) 3 A. semicinctus (Bennett, 1833) 3. A. tekomaji (Smith, 1959) 1 Asterropteryx semipunctatus Riippell, 1830 1 *Bathygobius cocosensis (Bleeker, 1854) 1 *B. cyclopterus (Valenciennes, 1837) 1 Bryaninops ridens Smith, 1959 1 Cabillus tongarevae (Fowler, 1927) 1 Callogobius maculipinnis (Fowler, 1918) 1 *C. sclateri (Steindachneri, 1880) 5 Callogobius sp. 1 Discordipinna griessingeri Hoese and Fourmanoir, 1978 2 Eviota lachdeberei ? Giltay, 1933 2 E. latifasciata ? Jewett and Lachner, 1983 1 E. melasma Lachner and Karanella, 1980 1 E. prasina (Klunzinger, 1871) 5 *Eviota sp. 1 5 *Eviota sp. 2 5 *Eviota sp. 3 5 Eviota sp. 4 1 Exyrias belissimus (Smith, 1959) 1 Fusigobius duospilus Hoese and Reader, 1985 1 F. neophytus (Giinther, 1877) 5 *Fusigobius sp. 1 Gnatholepis anjerensis (Bleeker, 1850) 3. G. caurensis (Bleeker, 1853) 5 Gnatholepis sp. 2 *Gobiodon okinawe Sawada, Arai, and Abe, 1973 1 *G. rivulatus (Riippell, 1830) 5 Oplopomops sp. 1 Oplopomus oplopomus (Valenciennes, 1837) 1 Palutrus pruinosus (Jordan and Seale, 1906) 1 Paragobiodon echinocephalus (Riippell, 1830) 1 *Priolepis cincta (Regan, 1908) 1 P.inhaca (Smith, 1949) 1 *P. semidoliatus (Valenciennes, 1837) 2 Psilogobius prolatus Watson and Lachner, 1985 1 Sueviota lachneri Winterbottom and Hoese, 1988 1 *Trimma emeryi Winterbottom, 1985 1 T. hoesei Winterbottom, 1984 1 T.macrophthalma (Tomiyama, 1936) 1 *T. taylori Lobel, 1979 1 T. undisquamis (Gosline, 1959) 3 T. winchi Winterbottom, 1984 3 *Trimma sp. 2 *Trimmaton sagma Winterbottom, 1989 1 *Valenciennea helsdingenii (Bleeker, 1858) 1 *V. sexguttata (Valenciennes, 1837) 1 *V. strigata (Broussonet, 1872) 1 Vanderhorstia ornatissima Smith, 1959 Xenisthmidae - Sandfishes 3 Xenisthmus africanus Smith, 1958 2 X. clara (Jordan and Seale, 1906) Microdesmidae - Hovergobies [We follow Randall and Hoese (1985) in including Nemateleotris and Ptereleotris in this family.] 1 *Gunnelichthys monostigma Smith, 1958 1 *Nemateleotris decora Randall and Allen, 1973 1 *N. magnifica Fowler, 1938 1 *Ptereleotris evides (Jordan and Hubbs, 1925) 1 *P. heteroptera (Bleeker, 1855) 1 *P. microlepis (Bleeker, 1856) 1 *P. zebra (Fowler, 1938) Kraemeridae - Sand darts 1 Kraemeria samoensis Steindachner, 1906 Acanthuridae - Surgeonfishes 1 *Acanthurus blochii Valenciennes, 1835 2 *A. guttatus Bloch and Schneider, 1801 2 *A. leucosternon Bennett, 1832 1 *A. lineatus (Linnaeus, 1758) 2 *A. maculiceps (Ahl, 1923) 1 *A. mata (Cuvier, 1829) 2 *A. nigricans (Linnaeus, 1758) 1 *A. nigricauda Duncker and Mohr, 1929 1 *A. nigrofuscus (Forsskal, 1775) 1 A. nigroris Valenciennes, 1835 2 *A. olivaceus Bloch and Schneider, 1801 1 *A. pyroferus Kittlitz, 1834 18 1 *A. thompsoni (Fowler, 1923) [P] 1 *A. triostegus (Linnaeus, 1758) [P] 1 *A. xanthopterus Valenciennes, 1835 [O] 1 *Ctenochaetus striatus (Quoy and Gaimard, 1825) [P,W] 1 *C. strigosus (Bennett, 1828) [P,W] 1 *Naso brevirostris (Valenciennes, 1835) [P] 1 *N. hexacanthus (Bleeker, 1855) [P] 1 *N. lituratus (Bloch and Schneider, 1801) [P,W] 1 *N. unicornis (Forsskal, 1775) [P,W] 1 *N. vlaminghi (Valenciennes, 1835) [P] 1 *Paracanthurus hepatus (Linnaeus, 1766) [P] 3 Zebrasoma desjardinii (Bennett, 1835) [P] [Most recent authors have recognized this Indian Ocean surgeonfish as a subspecies of the Pacific Z. veliferum. We follow Burgess (1973) in recognizing them both as distinct species, and note that in contrast to Cocos, Christmas Is. fish have the typical veliferum coloration. ] 1 *Z. scopas (Cuvier, 1829) [P,W] Zanclidae - Moorish Idols 1 *Zanclus cornutus (Linnaeus, 1758) [P] Siganidae - Rabbitfishes 1 Siganus argenteus (Quoy and Gaimard, 1825) [P] 2 S.puellus Schlegel, 1852 [P] 2 S. punctatus (Bloch and Schneider, 1801) [P] 3. S. stellatus Forsskal, 1775 [O] Scombridae - Tunas 4 *Acanthocybium solandri (Cuvier, 1831) [O] 1 *Gymnosarda unicolor (Riippell, 1836) [O] 4 *Thunnus albacares (Bonnaterre, 1788) [O] Bothidae - Flounders 1 *Bothus mancus (Bonnaterre, 1782) [P,W] 1 *B. pantherinus (Riippell, 1830) [W] Soleidae - Soles 5 *Aseraggodes sp. 1 [P] 5 Aseraggodes sp. 2 [P] Balistidae - Triggerfishes 1 *Balistapus undulatus (Park, 1797) [P] 1 *Balistoides viridescens (Bloch and Schneider, 1801) [O] 1 *Melichthys indicus Randall and Klausewitz, 1973 [P] 4 *M. niger (Bloch, 1786) [P] 1 *M. vidua (Solander, 1844) [P] 1 *Odonus niger (Riippell, 1837) [P] 1 Pseudobalistes flavimarginatus (Riippell, 1829) [P,W] 1 Rhinecanthus aculeatus (Linnaeus, 1758) [P,W] 1 *R. rectangulus (Bloch and Schneider, 1801) [P] 1 *Sufflamen bursa (Bloch and Schneider, 1801) [P] 1 *S. chrysopterus (Bloch and Schneider, 1801) [P] 1 S. fraenatus (Latreille, 1804) [P] 1 *Xanthichthys auromarginatus (Bennett, 1831) 1 *X. caeruleolineatus Randall, Matsuura and Zama, 1978 Monacanthidae - Leatherjackets 4 *Aluterus scriptus (Osbeck, 1765) 1 *Cantherines dumerilii (Hollard, 1854) 1 *C. pardalis (Riippell, 1837) 1 *Pervagor aspricaudus (Hollard, 1854) Ostraciontidae - Boxfishes 1 *Ostracion cubicus Linnaeus, 1758 Tetraodontidae - Puffers 1 *Arothron hispidus (Linnaeus, 1758) 1 *A. nigropunctatus (Bloch and Schneider, 1801) 1 *Canthigaster amboinensis (Bleeker, 1865) 1 *C. bennettii (Bleeker, 1854) 1 *C. janthinoptera (Bleeker, 1855) 1 *C. valentini (Bleeker, 1853) Diodontidae - Porcupinefishes 4 *Diodon hystrix Linnaeus, 1758 LS) 20 Table 1. Comparison of total ichthyofauna' and selected families of fishes occuring at Cocos-Keeling (CK) or Christmas Island (CI); numbers in parentheses are percent of total fauna; data for Christmas Island based on Slightly updated version of checklist given in Allen and Steene (1987). number of species shared CK cI Family CK (%) CI (%) spp. only only Labridae 54 (10.2) 61 (10.8) 43 as 16 Gobiidae 51 (9.6) 36 (6.4) 18 33 18 Pomacentridae 38 (7.2) 44 (7.8) 3 7 11 Apogonidae 30 (5.7) 22 (3.9) 12 20 10 Serranidae 30 (5.7) 44 (7.8) 19 11 25 Acanthuridae 25 (4.7) 26 (4.6) 24 al 2 Muraenidae 24 (4.5) 34 (6.0) 20 4 14 Chaetodontidae 23 (4.3) 27 (4.8) 19 4 8 Blenniidae 21°'(450) ° 28° (5.0) 14 7 14 Scaridae 20 (3.8) US 12/07.) 11 9 6 Holocentridae 20 (3.8) 1S) 627571) 12 8 3 Scorpaenidae 16 (3.0) 19 (3.4) 11 5 8 Balistidae 14 (2.6) 12 (2.1) 11 3 1 Carangidae V1 (2). 1) aSI(272:3)) 10 1 2 Lethrinidae 10 (1.9) 2 (0.0) 2 8 0 Lutjanidae $902.5) 15) (2:7) 7 a 8 Cirrihitidae 8 (1.5) 7 (1.2) 7 a 0 Mullidae 8. (PSS). 27 12) 7 1 0 Pomacanthidae > Keb) 121 (2;s1) 4 3 8 Microdesmidae qe) CL e3') Te \(lser2}) 7 ) 0 Tetradontidae Gren(dist)). ¢i9res( 136) 6 0 3 Ophichthidae 6 (1.1) mm (lie:2)) 2 4 5 Syngnathidae 6 (1.1) 7 (1.2) 4 2 3 Total fauna! 530 563 351 176 212 'The following families of epipelagic fishes were unsampled or under- sampled at Cocos (Keeling) Island, and to make the above faunal comparisons more meaningful, species of these families are not included in the total fauna counts (percentages were also calculated using the adjusted totals): Rhincodontidae, Exocoetidae, Coryphaenidae, Gempylidae, Scombridae and Istiophoridae. Table 2. Zoogeographic analysis of the Cocos (Keeling) fish fauna. Distribution Widespread Indo-Pacific or Indo-west Pacific West Pacific & Cocos Is. Indian Ocean Circumtropical Uncertain No. species 388 85 31 12 percent of total fauna ZA Also known from Christmas Is. ATOLL RESEARCH BULLETIN NO. 413 CHAPTER 15 BARNACLES (CIRRIPEDIA, THORACICA) OF THE COCOS (KEELING) ISLANDS BY D.S. JONES ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 15 BARNACLES (CIRRIPEDIA, THORACICA) OF THE COCOS (KEELING) ISLANDS BY D.S. JONES * INTRODUCTION The barnacle fauna of the Cocos (Keeling) Islands has not been documented prior to the present report. Previous reports on the crustaceans of these islands have listed Brachyura, Anomura, Caridea, Stomatopoda and Paguridae (Calman 1909, Wood-Jones 1909, Gibson-Hill 1947, 1948, Tweedie 1950, Forest 1956). The present collection of barnacles was made by sampling a wide variety of habitats throughout the atoll. Sampling stations included all reef flat zones and a range of lagoonal habitats and outer reef slopes (see Chapter 1, Fig. 2, List of barnacles). Specimens were collected by walking on shores and reef flats during low tide, snorkelling and SCUBA diving. A total of 13 species of barnacles in 11 genera are now recorded from the Cocos (Keeling) Islands. Nine of these species were collected at one locality only. Although the number of barnacles is small, this collection is of considerable interest since nothing is known of the barnacles of these islands, and knowledge of the barnacle fauna of coral atolls in the Indian Ocean in general is scanty. Eleven species were collected during the Western Australian Musum Cocos (Keeling) Island expedition of February, 1989. Two additional species from these islands (Capitulum mitella, Megabalanus ajax) are housed in the crustacean collection of the Western Australian Museum and are included in the species list. The species list given below must be considered provisional and further detailed collecting may well reveal additional species, particularly from sub-tidal areas. DISCUSSION The barnacle fauna of the Cocos (Keeling) Islands is composed of widespread Indo-West Pacific species (8) and species exhibiting cosmopolitan distributions (5). Of the cosmopolitan barnacles, L. anatifera and L. anserifera are pelagic in habitat and attach to floating objects. Freshly beached specimens were collected on reef platforms and beaches, attached to bamboo, wood, etc. However, large numbers of the abundant terrestrial hermit crab Coenobita perlatus Milne Edwards were observed actively predating these barnacles. Consequently strandings of pelagic barnacles may be more numerous, and more pelagic species may be represented than are presently recorded, but ensuing rapid predation (by terrestrial hermit crabs in particular) makes the collection of all but recently stranded specimens unlikely. Western Australian Museum, Francis Street, Perth, Western Australia, 6000. The cosmopolitan pedunculate barnacle O. lowei occurs on the gills of decapod crustaceans and was obtained from the gills of the rock lobster Panulirus penicillatus Oliver at the Cocos Islands. The cosmopolitan fouling species, M. tintinnabulum, occurred on mooring buoys in the main lagoon. This species may have been introduced via shipping since it was not found elsewhere in island waters and the species is a well-known fouler of ship hulls. T. divisa has a circumtropical insular and occasional mainland distribution. This species was first described from material collected on the west coast of Sumatra and the Java Sea. At Cocos (Keeling) the species was rare, only a few specimens being collected from deep, narrow crevices in beachrock. The Indo-West Pacific species T. wireni was the only sessile species collected at more than one locality. It was collected subtidally from a variety of hard substrata (mooring buoys, carapace of P. penicillatus, shell of Trochus maculatus Linnaeus) as well as from a sponge collected in beach drift. L. nicobarica bores into limestone substrates in the Indo-West Pacific region. The species was rare at the Cocos (Keeling) Islands, boring into the upper areas of coral and limestone boulders on intertidal and seaward reef flats. The Indo-West Pacific species C. mitella occurs in crevices in mid-tidal areas under conditions of semi to full wave exposure. T. fissum occurs on the mouthparts of decapod crustaceans in the Indo-West Pacific region, and was collected from the third maxillipeds of P. pencillatus. Two balanomorphs, E. hembeli and M. ajax, both very large, robust species, are known from the Indo-West Pacific region although both are extremely rare. E. hembeli occurs on high intertidal rocks and shores and at the Cocos (Keeling) Islands a solitary individual was collected from high up on the limestone boulder on an ocean reef flat. M. ajax occurs mainly in the subtidal on corals. The low live coral cover at Cocos is reflected in the very low number of coral barnacles collected. Although numerous samples of live and dead coral were examined, only one sample harboured coral barnacles - viz. S. dentatum on the coral Favia stelligera (Dana). The genus Acasta occurs in sponges and has many Indo-West Pacific representatives. Parietal plates of Acasta sp. were obtained from a sponge found in beach drift at the Cocos (Keeling) Islands. Barnacles are relatively rare and inconspicuous intertidal organisms at Cocos (Keeling) and their paucity in the overall marine invertebrate community is notable. The total of 13 species is small and may reflect inadequate sampling procedures. It may, however, be a true representation, since coral reefs are known to be unfavourable habitats for cirripedes (Darwin 1854, Borradaile 1903). The lack of development of barnacle populations on coral reefs has been documented in the tropical West Pacific (Newman 1960), the Tokara Islands, Japan (Utinomi 1954) and Heron Island, Australia (Endean et al. 1956). The scouring action of waves by rolling light coral limestone fragments and boulders is restrictive or adverse to barnacle settlement, especially in intertidal areas, where settlement would be restricted to crevices and underhangs. Newly settled and juvenile barnacles are indirectly predated on by herbivorous fish, which rasp limestone and coral reefs for micro-algae. Consequently barnacles may be restricted to higher intertidal areas or to boring into limestone substrata. Other known predators in the marine environment are molluscs (e.g. whelks) and sublittoral echinoderms (e.g. Diadema). These factors may all contribute to the general lack of intertidal barnacles at the Cocos (Keeling) Islands. In intertidal areas barnacles occur rarely and very sparsely, in interstices in or between and under rocks (e.g. C. mitella, T. divisa) or high on rocks which are only covered during high tides (E. hembeli). Burrowing forms occur in limestone and coral boulders (L. nicobarica) or live coral (S. dentatum). Some T. divisa individuals collected at Cocos (Keeling) exhibited gastropod bore holes in parietal plates. On fouling buoys in the main lagoon a small fouling community (e.g. sponges, ascidians, barnacles) is developing and many large specimens of the fouler M. tintinnabulum 3 were collected here as well as individuals of T. wireni. The presence and abundance of M. tintinnabulum and T. wireni at this site compared to the paucity of barnacle species elsewhere may be associated with a lack of coral. This, combined with a lack of shelter from piscivorous fish, may result in an absence of reef fish (e.g. Scaridae) and hence the predation pressure on newly settled barnacles and juveniles may be correspondingly reduced. The origin of the Cocos specimens of M. tintinnabulum is unknown. However, the presence of this species may be of some concern since it is a noted fouling species overseas. The thoracic cirripede fauna of the Indian Ocean is relatively well-known, with upwards of 280 species estimated to occur there (Stubbings 1936, Nilsson-Cantell 1938, Daniel 1972). The nearest localities to the Cocos (Keeling) Islands are Christmas Island, 900 km to the east and Java Head, 1000 km to the north-east. Only two cirripede species are recorded from Christmas Island (Nilsson-Cantell 1934 Daniel 1972), but the Western Australian Museum crustacean collection holds an additional five species making a total of seven. Java is part of Indo-Malay faunistic province, an area rich in both number of species and in the geographical distribution of cirripedes. At least 246 species are recorded from this area (Hoek 1907, 1913, Broch 1931, Nilsson-Cantell 1934). Compared to the nearest mainland shores (Sumatra, Java) which exhibit rich barnacle faunas in both the intertidal and the sub littoral (Nilsson-Cantell 1921), the fauna of the Cocos (Keeling) Islands must be considered depauperate. The number of cirripedes recorded from the Cocos (Keeling) Islands is larger than that at present recorded from other isolated Indian Ocean atolls (e.g. Diego Garcia, Chagos) but less than that presently known from larger atoll groups (e.g. Maldives and Laccadives). Table 1 compares the numbers of barnacle species recorded from islands and atolls in the Indian Ocean, and the species in common with the Cocos (Keeling) Islands. However, meaningful comparisons with other atolls and islands are difficult to make since the collecting effort at these localities is not known. ACKNOWLEDGEMENTS For identifications I thank Mrs. L.M. Marsh (coral) and Mr. P. Unsworth (gastropod). REFERENCES Annandale, N. 1906. Report on the Cirripedia collected by Professor Herdman, at Ceylon, in 1902. Rep. Pearl Fish. Manaar, 5: 137-150. Borradaile, L.A. 1903. Marine crustaceans. Part 7. The barnacles (Cirripedia) - In Gardiner, J.S. The fauna and geography of the Maldive and Laccadive Archipelagoes Part 1, 2: 440-443. Broch, H. 1931. Papers from Dr. Th. Mortensen’s Pacific Expedition 1914-1916. LVI. Indomalayan Cirripedia. Vidensk. Medd. dansk naturhist. Foren. Kbh. 91: 1-146. Calman, W.T. 1909. Crustacea of Cocos (Keeling). Proc. Zool. Soc. Lond. 1909: 159- 160. Daniel, A. 1972. Marine intertidal barnacles in the Indian Ocean. Proc. Indian Acad. Sci., Part B, 38: 179-189. Darwin, C. 1854. A monograph on the subclass Cirripedia with figures of all species. The Balanidae, the Verrucidae, etc. (Roy. Soc., London.) 684 pp. Endean, R., Stephenson, W. and Kenny, R. (1956). The ecology and distribution of intertidal organisms on certain islands off the Queensland coast. Aust. J. Mar. Freshw. Res. 7: 317-342. Forest, J. 1956. La Faune des iles Cocos (Keeling): Paguridae. Bull. Raffles Mus. 27: 45-55. Gibson-Hill, C.A. 1947. Notes on the Cocos Keeling Islands. J. Malay. Br. Asiat. Soc. 20: 140-202. Gibson-Hill, C.A. 1948. The Island of North Keeling. J. Malay. Br. Asiat. Soc. 21: 68- 103. Gruvel, J.A. 1909. Etudes des cirrhipédes de l’ocean Indien. Trans. Linn. Soc. Lond. (Zool.), ser. 2, 13: 23-27. Hoek, P.P.C. 1907. The Cirripedia of the Siboga Expedition. A. Cirripedia Pedunculata. Siboga Exped. Monogr. 31a: V-XXV, 1-127. Hoek, P.P.C. 1913. The Cirripedia of the Siboga Expedition. B. Cirripedia Sessilia. Siboga Exped. Monogr. 31b: I-XXV, 129-275. Newman, W.A. 1960. The paucity of intertidal barnacles in the tropical Western Pacific. Veliger 2: 89-94. Nilsson-Cantell, C.A. 1921. Cirripeden-studien. Zur Kenntnis der Biologie, Anatomi und Systematik Diesser Gruppe. Zool. Bidrag. Uppsala 7: 75-395. Nilsson-Cantell, C.A. 1934. Indo-Malayan cirripedes in the Raffles Museum, Singapore. Bull. Raffles Mus. 9: 42-73. Nilsson-Cantell, C.A. 1938. Cuirripedes from the Indian Ocean in the collection of the Indian Museum, Calcutta. Mem. Indian Mus. 13: 1-81. Smith, W.A. 1971. Crustacea: cirripedes from Diego Garcia. Atoll Res. Bull. 149: 103. Stubbings, H.G. 1936. Cirripedia. John Murray Expeditions, 1933-1934. Sci. Rep. Murray Exped. 4: 1-70. Taylor, J.D. 1968. Coral reef and associated invertebrate communities (mainly molluscan) around Mahe, Seychelles. Phil. Trans. Roy. Soc. Lond. (Biol.), B254: 129-206. Tweedie, M.W.F. 1950. The fauna of the Cocos-Keeling Islands; Brachyura and Stomatopoda. Bull. Raffles Mus. 22: 105-148. Utinomi, H. 1954. Invertebrate fauna of the intertidal zone of the Tokara Islands IX. Cirripedia. Publ. Seto Mar. Biol. Lab. 4: 17-26. 5 Wood-Jones, F. 1909. The fauna of the Cocos-Keeling Islands. Proc. Zool. Soc. Lond. 1909: 132-160. LIST OF BARNACLES x = specimens without precise locality data ORDER THORACICA Suborder Lepadomorpha Family Scalpellidae Pilsbry, 1916 Subfamily Lithotryinae Gruvel, 1905 Lithotrya nicobarica Reinhardt, 1850 Subfamily Pollicipinae Gruvel, 1905 Captitulum mitella (Linnaeus, 1767) Family Lepadidae Burmeister, 1834 (Fam. Lepadea") Lepas anatifera Linnaeus, 1767 Lepas anserifera Linnaeus, 1767 Family Poecilasmatidae Annandale, 1910 Temnaspsis fissum Darwin, 1851 Octolasmis lowei (Darwin, 1851) Suborder Balanomorpha Family Chthamalidae Darwin, 1854 Subfamily Euraphiinae Newman & Ross, 1976 Euraphia hembeli Conrad, 1837 Family Tetraclitidae Gruvel, 1903 Subfamily Tetraclitinae Gruvel, 1903 Tetraclita divisa Nilsson-Cantell, 1921 Tesseropora wireni Nilsson-Cantell, 1921 Family Archaeobalanidae Newman & Ross, 1976 Acasta sp. Family Pyrgomatidae Gray, 1825 Savignium dentatum (Darwin, 1854) Family Balanidae Leach, 1817 Megabalanus ajax (Darwin, 1854) Collection Station 132 24, 27, 30 Zils SOE 10 20 10 1, 2; 6, 20528 31 xX Megabalanus tintinnabulum (Linnaeus, 1758) 28 Table 1 : A comparison of the numbers of thoracic cirripede species recorded from Indian Ocean atolls (A) and islands, and the species in common with the Cocos (Keeling) Islands. Locality Total Spp. in common References Species with Cocos (Keeling) Christmas I. ih 4 Nilsson-Cantell 1934; Daniel 1972; WA Museum Collection Diego Garcia (A) 3 8 Smith 1971 Chagos (A) 6 3 Gruvel 1909 Andamans & Nicobars 17 D Gruvel 1909; Daniel 1972 Sri Lanka 31 3 Annandale 1906; Daniel 1972 Maldives & Laccadives (A) 25 Z Borradaile, 1903; Annandale 1906; Daniel 1972 Seychelles 9 2 Gruvel 1909; Taylor 1968 Providence I. My, 0 Gruvel 1909 ATOLL RESEARCH BULLETIN NO. 414 CHAPTER 16 DECAPOD CRUSTACEANS OF THE COCOS (KEELING) ISLANDS BY G.J. MORGAN ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. FEBRUARY 1994 CHAPTER 16 DECAPOD CRUSTACEANS OF THE COCOS (KEELING) ISLANDS BY G.J. MORGAN * INTRODUCTION Prior to the present study, the crustacean fauna of the Cocos (Keeling) Islands had been collected intensively only once, by C.A. Gibson-Hill in the years 1940-41. The Brachyura and Stomatopoda of his material were taxonomically reviewed by Tweedie (1950) and the hermit crabs by Forest (1956). Gibson- Hill's collection represents almost exclusively intertidal and terrestrial faunas. In the present survey, SCUBA was employed extensively to augment the poorly known subtidal faunas. Xanthoid crabs (families Xanthidae, Trapeziidae, Pilumnidae, Menippidae) and marine hermit crabs (families Diogenidae and Paguridae) are dominant decapod components of tropical rocky and coral reef ecosystems and were collected preferentially. Conversely, some taxa (e.g. caridean shrimps) are poorly represented in the present collection. Specimens were sampled by reef walking in the intertidal and by SCUBA or snorkelling in the subtidal habitats. A total of 198 species of Decapoda is recorded of which 78 are new records for the Cocos (Keeling) Islands (see list of species). Station localities are listed by number (see Chapter 1, Fig.2) and some additional sites sampled are recorded by name. Station 9 (Direction Island) has been divided for this list into 9(a): blue hole south of Direction I. and 9(b): sand shallows between island and blue hole. In order to compile as complete a record of decapod crustaceans as possible, the list includes species recorded by the previous workers noted above. The species names are those currently used in the scientific literature and not necessarily the names cited by historical workers. The historical collections have not been examined and hence the accuracy of early identifications cannot be ascertained. DISCUSSION The most diverse decapod taxa of the Cocos (Keeling) Islands are the xanthoid and paguroid crabs with 77 and 48 species recorded respectively. Both taxa are particularly evident in subtidal and intertidal reef habitats, although the abundance of xanthoids was found to be somewhat lower than expected on the 1989 sampling expedition. The islands can be divided broadly into several major habitat types with a convenient dichotomy between lagoon and outer oceanic environments. In the lagoon, only a few species of hermit crab notably Clibanarius longitarsus and Calcinus laevimanus are present in the sheltered shallow intertidal areas of fine mud sediment. Hermits appear to be absent from the very extensive soft grey sediment flats in North and South Lagoons of West Island, but these flats support high numbers of 2 Western Australian Museum, Francis Street, Perth, Western Australia, 6000. Z Macrophthalmus verreauxi and Uca chlorophthalmus. The latter produces a distinct pattern of hexagonal territories in some upshore areas. The portunid Thalamita crenata is common in the shallows of the lagoon. Tweedie (1950) identified T. spinimana as the species plentiful in the shallow sandy, slightly weedy water of the lagoon’ (Gibson-Hill, in Tweedie 1950) but in 1989 the common portunid in this habitat was T. crenata. It is possible that Tweedie's identification was spurious but he has recorded T. crenata from the ‘outer edge of the atoll’. The large edible mud or mangrove crab Scylla serrata occasionally is caught by locals in the very sheltered areas of the lagoon but is rarely seen and presumably occurs in very low numbers. Ghost crabs, Ocypode ceratophthalma and O. cordimana, forage across the lagoon flats from their upshore burrows and are also numerous on the oceanic beaches. O. ceratophthalma occurs in its two colour morphs, with the olive green form far outnumbering the cream and brown morph. The terrestrial crab Cardisoma carnifex also feeds on exposed flats of the lagoon at low tide. The continual natural process of sedimentary infilling of the lagoon, together with the occurrence of the above crabs in very high numbers, indicate that the populations of soft sediment crustaceans are relatively secure for the foreseeable future. In deeper areas of the lagoon and near the major channels allowing entry of oceanic water, hard and soft corals are present to a variable extent. The 1983 El Nino effect resulted in very extensive coral death and over much of the lagoon only small colonies of live hard corals have re-established. As many species of crustaceans are either symbiotic with live corals or prefer the live coral habitat, loss of corals is reflected in crustacean occurrences. Where live coral, especially Pocillopora spp., is present, the crustacean assemblage resembles that discussed for outside reef coral habitats. The dead coral rubble supports a lower diversity of hermit crabs with the diogenid Calcinus latens and several small species of pagurids dominant in numbers. Portunids and xanthoids are also present but often difficult to collect in the deep layer of coral fragments. Sandy areas and beds of the seagrass Thalassia and the alga Caulerpa support Calcinus latens and C. laevimanus in relatively low populations and several pagurids in high numbers, especially Micropagurus vexatus. Sandy areas are also habitat for several portunid species and Calappa hepatica. If further live coral dieback were to occur in the lagoon, the symbiotic faunal communities would be placed at considerable risk. Presumably there has been, and would continue to be, replacement of coral crustacean communities by rubble and sand-living species. The oceanic reefs of the Cocos (Keeling) Islands have also experienced major reduction of the live coral habitat. Percentage cover of live hard corals is low and many of the outside reefs are dominated by bare, wave scoured dead coral or coral rubble. The shallow subtidal and intertidal reef habitats are home to a variety of hermit crabs with Calcinus minutus, C. latens, C. sp. 1 (an undescribed species), Dardanus crassimanus and D. lagopodes common. The large Dardanus species, D. megistos, D. guttatus, D. gemmatus and D. deformis occur on shallow reef flats and adjacent sandy environments. A variety of xanthoids especially Pilodius areolatus, are fairly numerous under the coral and coral rubble. Where live branching coral is present, especially Pocillopora, symbiotic species of xanthoids (e.g. Trapezia and Cymo), and alpheids (e.g. Alpheus lottini), occur. Under coral slabs, particularly in the subtidal, several species of pagurid are common with the bright lilac Pylopaguropsis magnimanus especially obvious. As was recorded for Christmas Island (Morgan, unpublished), pagurids can be quite diverse and common in tropical reef habitats and their taxonomy requires considerable attention. Interestingly, the populations and diversity of porcellanids (porcelain crabs) were low at Cocos (Keeling). 3 There seemed no obvious explanation for this apart from the possible effects of a paucity of live coral. Intertidal rock and rock-sand platforms support quite high populations of hermit crabs with Calcinus laevimanus most widespread and C. latens and Clibanarius humilis common in areas. High on the platforms, in very warm pools flushed only by spring tides, the only hermit crab is Calcinus seurati. Xanthoids are also characteristic of rocky platforms, especially species of Leptodius. Grapsids are usually the most conspicuous crabs on intertidal hard substrata, with Grapsus tenuicrustatus and G. intermedius the largest species on Cocos (Keeling). Three species of Pachygrapsus occur in a range of intertidal habitats, usually on or under rock or coral slabs. Prior to this study, two species of rock lobster, Panulirus penicillatus and P. versicolor were recorded from the Cocos (Keeling) Islands (George 1968). The presence of a third species, P. ornatus, commonly referred to on the islands as the “leopard cray', was confirmed during the 1989 study. All three species are very widespread in the Indo- West Pacific area. Panulirus species have extended larval stages of several months with the planktonic phyllosoma capable of drifting great distances on ocean currents before settlement as the puerulus. It is probable that rock lobster stocks at Cocos (Keeling) are dependent upon larvae originating considerable distances from the islands and hence fishing overexploitation of the local breeding population is unlikely to severely effect settlement. It is certainly possible, however, that heavy fishing might deplete the population of table-size lobsters. Some form of monitoring of fishing effort would provide information on distribution of the species and their present abundance. The Cocos (Keeling) Islands do not support the numbers or diversity of true terrestrial crabs so obvious on Christmas Island (Indian Ocean). The gecarcinid Cardisoma carnifex is very common on West Island, with apparently lower populations elsewhere. No specimens of C. rotundum recorded by Tweedie (1950) (as C. frontalis), were collected on the 1989 expedition. The grapsids Geograpsus crinipes and G. grayi co- occur on the islands, in lower numbers than Cardisoma. The presence of the Christmas Island ‘red crab', Gecarcoidea natalis, was confirmed on North Keeling Island, but the species occurs in only small numbers. Tweedie (1950) stated that its occurrence was due to introduction with soil from Christmas Island to Direction Island and thence by larvae to other islands in the Cocos. This argument is convincing, given the small population of G. natalis on Cocos (Keeling) and the great distance (900 km) to Christmas Island, the only other habitat of the species. Only one species of the terrestrial hermit crab family Coenobitidae had been recorded previously from Cocos (Keeling), namely Coenobita perlatus (Forest 1956). It is odd that the two additional species collected in this study, C. rugosus and C. brevimanus, were not represented in Gibson-Hill's collection as the former in particular is not uncommon. The coconut or robber crab, Birgus latro, was not found during the 1989 expedition, despite searches for it, but is reported by local people to occur on at least West and North Keeling Islands, the latter occurrence being confirmed by the ANPWS ranger, Mr Paul Stephenson (pers. comm.). In addition, a specimen of B. latro is on display in the local museum on Home Island, reportedly collected on Cocos (Keeling). Charles Darwin (1845) noted that B. latro was common at the time of his visit in 1836 and it might be suggested that the presently very low population of the species is due to overexploitation by the islands’ local inhabitants. Protection of the existing specimens will be necessary to ensure their continued survival on Cocos (Keeling). There are no naturally occurring bodies of freshwater on the islands, although a brackish lake (Bechet Besar) is present towards the north end of West Island. Freshwater must be accessed by sinking wells into subterranean reserves. It is scarcely surprising therefore that no freshwater crustacea were collected, unlike Christmas Island. Cardisoma carnifex was observed to concentrate at temporary rainwater pools and several specimens were seen in a shallow well, totally immersed in freshwater. REFERENCES Darwin, C. 1845. Journal of researches into the natural history and geology of the countries visisted during the voyage of H.M.S. Beagle round the world, under the command of Capt. Fitroy R.N. London. John Murray Forest, J. 1956. La Faune des iles Cocos-Keelings Paguridea. Bull. Raffles Mus. 27: 45- 59! George, R.W. 1968. Tropical spiny lobsters Panulirus spp., of Western Australia (and the Indo-West Pacific). J. Roy. Soc. W.A. 51(2): 33-38. Tweedie, M.W.F. 1950. The fauna of the Cocos-Keeling Islands, Brachyura and Stomatopoda. Bull. Raffles Mus. 22: 105-148. LIST OF DECAPOD CRUSTACEA KEY TO SYMBOLS + New record for Cocos (Keeling) Islands = Not collected during this survey $ Not collected but occurrence confirmed Numbers = sampling station (see Chapter 1, Fig.2) STENOPODIDEA STENOPODIDAE + Stenopus hispidus (Olivier, 1811) 1 CARIDEA ALPHEIDAE + Alpheus lottini Guérin, 1829 WW + Alpheus macrodactylus Ortmann, 1890 1 +) Alpheus strenuus strenuus Dana, 1852 PD + Alpheus sp. 1 + Synalpheus stimpsoni (De Man, 1888) 19 PALINURA PALINURIDAE Panulirus penicillatus (Olivier, 1791) 1,10,20 $ Panulirus versicolor (Latreille, 1804) +$ Panulirus ornatus (Fabricius, 1798) ANOMURA DIOGENIDAE + Aniculus retipes Lewinsohn, 1982 4,19,32 Aniculus ursus (Olivier, 1811) 1,11 + Aniculus sp. 1 + Calcinus argus Wooster, 1984 4 Calcinus elegans (H. Milne Edwards, 1836) 1,6,11,12,18,30 Calcinus gaimardii (H. Milne Edwards, 1848) 1,6,9(a),11,25 + Calcinus guamensis Wooster, 1984 1 Calcinus laevimanus (Randall, 1839) 2,6,10,12,30,34 +++ ++4+4+ + *¥¥+ ++ * +++ Calcinus latens (Randall, 1839) Calcinus minutus Buitendijk, 1937 Calcinus pulcher Forest, 1958 Calcinus seurati Forest, 1951 Calcinus sp. 1 Calcinus sp. 3 Calcinus sp. 4 Calcinus sp. 5 Calcinus sp. 6 Clibanarius corallinus (H. Milne Edwards, 1848) Clibanarius eurysternus Hilgendorf, 1878 Clibanarius humilis Dana, 1852 Clibanarius laevimanus Buitendijk, 1937 Clibanarius longitarsus (De Haan, 1849) ?Clibanarius merguiensis De Man, 1888 Clibanarius striolatus Dana, 1852 Dardanus crassimanus (H. Milne Edwards, 1848) Dardanus deformis (H. Milne Edwards, 1836) Dardanus gemmatus (H. Milne Edwards, 1848) Dardanus guttatus (Olivier, 1811) Dardanus lagopodes (ForskAl, 1775) Dardanus megistos (Herbst, 1804) Dardanus scutellatus (H. Milne Edwards, 1848) Diogenes sp. Paguristes sp. Trizopagurus strigatus (Herbst, 1804) PAGURIDAE + ++ttet+ +4 +FAQ + Micropagurus vexatus Haig and Ball, 1988 4,7,9(a),13515,19522, 23325326332 Nematopagurus cf. muricatus (Henderson, 1896) Pagurixus anceps (Forest, 1954) Pagurixus tweediei (Forest, 1956) Pagurixus sp. Pylopaguropsis magnimanus (Henderson, 1896) Pagurid sp. 1 Pagurid sp. 2 Pagurid sp. 3 Pagurid sp. 6 OENOBITIDAE Birgus latro (Linnaeus, 1767) Coenobita brevimanus Dana, 1852 Coenobita perlatus H. Milne Edwards, 1837 Coenobita rugosus H. Milne Edwards, 1837 PORCELLANIDAE + Petrolisthes asiaticus (Leach, 1820) 1,6,8,9(a),9(b),12, 17, 18,22,23,30,34, 36 4°6,7,13,15319325332 4,9(a),15,32 10,30 4A TASS 19225255 32 4 A6, 15925 5 25 6,10,12,30 2,30 2 2 1,4,8,15,19,23,32 12 1,24 1,6,8,11,30 7,8,12,13,15,19,32 6,9(b),11,12,17,18, 19,34 6,9(b),12,17,34 9(b),19,22 4,6,25,32 722 22 1,6;12-30 1 4 1,6 9(a),15,16,26,36 1,16,17,18226 2: North Keeling 21 1,2,6,10,21 256;10313,21 + Petrolisthes carinipes (Heller, 1861) 1 GALATHEIDAE + Galathea sp. 1,19,23 BRACHYURA Dromiacea DYNOMENIDAE * Dynomene hispida Desmarest, 1825 Dynomene cf. pilumnoides Alcock, 1899 4,32 Dynomene praedator A. Milne Edwards, 1879 Dynomene sp. 4,25,32 + *+4+ Oxystomata CALAPPIDAE Calappa hepatica (Linnaeus, 1758) 9(b),19 Oxyrhyncha MAJIDAE EPIALTINAE + Huenia grandidierii A. Milne Edwards, 1865 20 Menaethius monoceros (Latreille, 1825 1,20,24 Perinia tumida Dana, 1852 1,24 + Simocarcinus obtusirostris (Miers, 1879) 18 MAJINAE Cyclax suborbicularis (Stimpson, 1907) 1 + Schizophrys aspera (H. Milne Edwards, 1834) 1,32 MITHRACINAE + Micippa thalia (Herbst, 1803) 11 PARTHENOPIDAE AETHRINAE Actaeomorpha erosa Miers, 1878 20 EUMEDONINAE = Eumedonus pentagonus (A. Milne Edwards, 1879) PARTHENOPINAE Daldorfia horrida (Linnaeus, 1758) Cancridea ATELECYCLIDAE * Kraussia integra (De Haan, 1835) + Kraussia cf. nitida Stimpson, 1858 23 = Kraussia rugulosa (Krauss, 1843) Brachyrhyncha PORTUNIDAE CATOPTRINAE af Carupa tenuipes Dana, 1851 PORTUNINAE + Charybdis erythrodactyla (Lamarck, 1818) i Charybdis obtusifrons Leene, 1936 Portunus granulatus (H. Milne Edwards, 1834) +$ Scylla serrata (Forskal, 1775) Thalamita admete (Herbst, 1803) + Thalamita chaptali (Audouin and Savigny, 1825) Thalamita crenata H. Milne Edwards, 1834 Thalamita dakini Montgomery, 1931 Thalamita demani Nobili, 1905 Thalamita integra Dana, 1852 Thalamita picta Stimpson, 1858 Thalamita spinimana Dana, 1852 Thalamitoides quadridens A. Milne Edwards, 1869 Thalamitoides tridens A. Milne Edwards, 1869 + *¥ eX +4 XANTHIDAE POLYDECTINAE Lybia tessellata (Latreille, 1812) i Polydectus cupulifer (Latreille, 1812) CYMOINAE Cymo andreossyi (Audouin, 1826) * Cymo quadrilobatus Miers, 1884 LIOMERINAE Liomera bella (Dana, 1852) Liomera caelata (Odhner, 1925) Liomera laevis (A. Milne Edwards, 1873) Liomera monticulosa (A. Milne Edwards, 1873) Liomera pallida (Borradaile, 1900) Liomera rugata (H. Milne Edwards, 1834) Liomera stimpsoni (A. Milne Edwards, 1865) Liomera tristis (Dana, 1852) + Liomera venosa (H. Milne Edwards, 1834) + Liomera sp. + *¥+ % & * EUXANTHINAE x Euxanthus exsculptus (Herbst, 1790) + Paramedaeus simplex (A. Milne Edwards, 1873) ACTAEINAE Actaeodes consobrinus (A. Milne Edwards, 1873) Actaeodes tomentosus (H. Milne Edwards, 1834) Gaillardiellus orientalis (Odhner, 1925) Gaillardiellus superciliaris (Odhner, 1925) Paractaea rufopunctata (H. Milne Edwards, 1834) Psaumis cavipes (Dana, 1852) Pseudoliomera granosimana (A. Milne Edwards, 1865) Pseudoliomera speciosa (Dana, 1852) Se 36 tee *% 12 el 1,6,9(a),12,17,18,36 36 2 18,24 4 20,25,32 9(a), 16,36 PM | APA] 13819 6,27 12,20,27 6,12,18 23 ZOSIMINAE = Atergatopsis signatus (Adams and White, 1848) Lophozozymus dodone (Herbst, 1801) Lophozozymus pulchellus A. Milne Edwards, 1867 Platypodia cristata (A. Milne Edwards, 1865) Platypodia granulosa (Ruppell, 1830) Platypodia cf. pseudogranulosa Serene, 1984 Zozymodes pumilus (Jacquinot and Lucas, 1852) Zosimus aeneus (Linnaeus, 1758) + * * XANTHINAE Lachnopodus gibsonhilli (Tweedie, 1950) Lachnopodus subacutus (Stimpson, 1858) Lachnopodus tahitensis De Man, 1889 Leptodius exaratus (H. Milne Edwards, 1834) Leptodius gracilis (Dana, 1852) Leptodius nudipes (Dana, 1852) Leptodius sanguineus (H. Milne Edwards, 1834) Lioxanthodes alcocki Calman, 1909 Macromedaeus nudipes (A. Milne Edwards, 1867) Neoxanthias impressus (Lamarck, 1818) + % & *X ETISINAE + Etisus bifrontalis (Edmondson, 1935) + Etisus demani Odhner, 1925 Etisus dentatus (Herbst, 1785) + Etisus frontalis Dana, 1852 x Etisus laevimanus Randall, 1840 Paraetisus sp. CHLORODIINAE Chlorodiella barbata (Borradaile, 1900) Chlorodiella cytherea (Dana, 1852) Chlorodiella laevissima (Dana, 1852) Phymodius granulosus (De Man, 1888) Phymodius monticulosus (Dana, 1852) Phymodius ungulatus (H. Milne Edwards, 1834) Pilodius areolatus (H. Milne Edwards, 1834) * Pilodius pubescens Dana, 1852 Pilodius scabriculus Dana, 1852 + Tweedieia odhneri (Gordon, 1934) + ¥+ TRAPEZIIDAE TRAPEZIINAE Tetralia glaberrima (Herbst, 1790) Be Trapezia areolata Dana, 1852 Trapezia cymodoce (Herbst, 1799) = Trapezia digitalis Latreille, 1825 Trapezia ferruginea Latreille, 1825 Trapezia guttata Ruppell, 1830 Trapezia rufopunctata (Herbst, 1799) 6,30 30 6,10,12,27 24 18 fel 18 1 1 17,18,20 13 9(a),17,36 1527 1 17,36 6 156, 125118220) 23.27 1,6,12 On 225 32 16,27 1,4;7,13524;25, DD) ey M32 25,36 TDD 10 + Trapezia septata Dana, 1852 DOMECIINAE Domecia hispida Eydoux and Souleyet, 1842 CARPILIIDAE * Carpilius convexus (Forskal, 1775) s Carpilius maculatus (Linnaeus, 1758) MENIPPIDAE OZIINAE Lydia annulipes (H. Milne Edwards, 1834) * Ozius tuberculosus H. Milne Edwards, 1834 ERIPHITINAE * Eriphia scabricula Dana, 1852 Eriphia sebana (Shaw and Nodder, 1803) DACRYOPILUMNINAE Dacryopilumnus rathbunae Balss, 1932 PILUMNIDAE + Pilumnus minutus (De Haan, 1835) INCERTAE SEDIS Daira perlata (Herbst, 1790) Pseudozius caystrus (Adams and White, 1848) PALICIDAE + Crossotonotus brevimanus (Ward, 1933) OCYPODIDAE OCYPODINAE Ocypode ceratophthalma (Pallas, 1772) Ocypode cordimana Desmarest, 1825 Uca chlorophthalmus (H. Milne Edwards, 1837) MACROPHTHALMINAE Macrophthalmus verreauxi H. Milne Edwards, 1848 GRAPSIDAE GRAPSINAE Geograpsus crinipes (Dana, 1851) Geograpsus grayi (H. Milne Edwards, 1853) Grapsus intermedius De Man, 1887 Grapsus tenuicrustatus (Herbst, 1783) Metopograpsus thukuhar (Owen, 1839) + Pachygrapsus minutus A. Milne Edwards, 1873 Pachygrapsus cf. planifrons De Man, 1888 Pachygrapsus plicatus (H. Milne Edwards, 1837) SESARMINAE + Cyclograpsus integer H. Milne Edwards, 1837 1 AD,22027 9(a),27 Loc. unrecorded 27 ISS 242732 6,30 19 1,2,6,10, West Island 1, West Island 2; » Nnww ) ,buoys(lagoon) RAPENNNNS — Sa LY ON Sesarma (Parasesarma) sigillata Tweedie, 1950 Sesarma (Parasesarma) lenzii De Man, 1895 NNW + Sesarma (Chiromantes) sp. PLAGUSIINAE Percnon abbreviatum (Dana, 1851) 12 a Percnon affine (H. Milne Edwards, 1853) + Percnon guinotae Crosnier, 1965 eg) Percnon planissimum (Herbst, 1804) WHI) Plagusia depressa tuberculata Lamarck, 1818 Buoys (lagoon) VARUNINAE + Pseudograpsus albus Stimpson, 1858 6 Thalassograpsus harpax (Hilgendorf, 1892) 10 GECARCINIDAE Cardisoma carnifex (Herbst, 1794) 2, West Island ce Cardisoma rotundum (Quoy and Gaimard, 1824) Gecarcoidea natalis (Pocock, 1888) Di CRYPTOCHIRIDAE + Hapalocarcinus marsupialis Stimpson, 1859 4 > i ATOLL RESEARCH BULLETIN NOS. 399-414 NO. 399. SCIENTIFIC STUDIES IN THE COCOS (KEELING) ISLANDS: AN INTRODUCTION BY C.D. WOODROFFE AND P.F. BERRY NO. 400. CLIMATE, HYDROLOGY AND WATER RESOURCES OF THE COCOS (KEELING) ISLANDS BY A.C. FALKLAND NO. 401. LATE QUATERNARY MORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY D.E. SEARLE NO. 402. _GEOMORPHOLOGY OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE, R.F. MCLEAN AND E. WALLENSKY NO. 403. . REEF ISLANDS OF THE COCOS (KEELING) ISLANDS BY C.D. WOODROFFE AND R.F. MCLEAN NO. 404. . VEGETATION AND FLORA OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS NO. 405. AN UPDATE OF BIRDS OF THE COCOS (KEELING) ISLANDS BY T. STOKES NO. 406. MARINE HABITATS OF THE COCOS (KEELING) ISLANDS BY D.G. WILLIAMS NO. 407. SEDIMENT FACIES OF THE COCOS (KEELING) ISLANDS LAGOON BY S.G. SMITHERS NO. 408. HYDRODYNAMIC OBSERVATIONS OF THE COCOS (KEELING) ISLANDS LAGOON BY P. KENCH NO. 409. _HERMATYPIC CORALS OF THE COCOS (KEELING) ISLANDS: A SUMMARY BY J.E.N. VERON NO. 410. _MARINE MOLLUSCS OF THE COCOS (KEELING) ISLANDS BY F.E. WELLS NO. 411. . ECHINODERMS OF THE COCOS (KEELING) ISLANDS BY L.M. MARSH NO. 412. FISHES OF THE COCOS (KEELING) ISLANDS BY G.R. ALLEN AND W.F. SMITH-VANIZ NO. 413. .BARNACLES OF THE COCOS (KEELING) ISLANDS BY D.S. JONES NO. 414. _DECAPOD CRUSTACEANS OF THE COCOS (KEELING) ISLANDS BY G.J. MORGAN ATOLL RESEARCH BULLETIN NOS. 415-425 RESEARCH BULLETIN Issued by NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C. U.S.A. AUGUST 1994 ATOLL RESEARCH BULLETIN NOS. 415-425 NO. 415. NO. 416. NO. 417. NO. 418. NO. 419. NO. 420. TIKEHAU AN ATOLL OF THE TUAMOTU ARCHIPELAGO (FRENCH POLYNESIA) PARTI. ENVIRONMENT AND BIOTA OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY A. INTES AND B. CAILLART PART Il. NUTRIENTS, PARTICULATE ORGANIC MATTER, AND PLANKTONIC AND BENTHIC PRODUCTION OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY C.J. CHARPY-ROUBAUD AND L. CHARPY PART Ill. REEF FISH COMMUNITIES AND FISHERY YIELDS OF TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY B. CAILLART, M.L. HARMELIN-VIVIEN, R. GALZIN, AND E. MORIZE COLONIZATION OF FISH LARVAE IN LAGOONS OF RANGIROA (TUAMOTU ARCHIPELAGO) AND MOOREA (SOCIETY ARCHIPELAGO) BY V. DUFOUR CAVES AND SPELEOGENESIS OF MANGAIA, COOK ISLANDS BY JOANNA C. ELLISON SHALLOW-WATER SCLERACTINIAN CORALS FROM KERMADEC ISLANDS BY VLADIMIR N. KOSMYNIN DESCRIPTION OF REEFS AND CORALS FOR THE 1988 PROTECTED AREA SURVEY OF THE NORTHERN MARSHALL ISLANDS BY JAMES E. MARAGOS QUATERNARY OOLITES IN THE INDIAN OCEAN BY C.J.R. BRAITHWAITE NO. NO. NO. NO. NO. 421. 422. 423. 424. 425. LARGE-SCALE, LONG-TERM MONITORING OF CARIBBEAN CORAL REEFS: SIMPLE, QUICK, INEXPENSIVE TECHNIQUES BY RICHARD B. ARONSON, PETER J. EDMUNDS, WILLIAM F. PRECHT, DIONE W. SWANSON, AND DON R. LEVITAN CHANGES IN THE COASTAL FISH COMMUNITIES FOLLOWING HURRICANE HUGO IN GUADELOPE ISLAND (FRENCH WEST INDIES) BY CLAUDE BOUCHON, YOLANDE BOUCHON-NAVARO, AND MAX LOUIS THE SIAN KA'AN BIOSPHERE RESERVE CORAL REEF SYSTEM, YUCATAN PENINSULA, MEXICO BY ERIC JORDAN-DAHLGREN, EDUARDO MARTIN-CHAVEZ, MARTIN SANCHEZ-SEGURA, AND ALEJANDRO GONZALEZ DE LA PARRA A PRELIMINARY EVALUATION OF THE COMMERCIAL SPONGE RESOURCES OF BELIZE WITH REFERENCE TO THE LOCATION OF THE TURNEFFE ISLANDS SPONGE FARM BY J.M. STEVELY AND D.E. SWEAT SPATIAL AND TEMPORAL VARIATIONS IN GRAZING PRESSURE BY HERBIVOROUS FISHES: TOBACCO REEF, BELIZE BY PETER N. REINTHAL AND IAN G. MACINTYRE NEWS AND COMMENTS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 ACKNOWLEDGMENT The Atoll Research Bulletin is issued by the Smithsonian Institution to provide an outlet for information on the biota of tropical islands and reefs and on the environment that supports the biota. The Bulletin is supported by the National Museum of Natural History and is produced by the Smithsonian Press. This issue is partly financed and distributed with funds from readers. The Bulletin was founded in 1951 and the first 117 numbers were issued by the Pacific Science Board, National Academy of Sciences, with financial support from the Office of Naval Research. Its pages were devoted largely to reports resulting from the Pacific Science Board's Coral Atoll Program. All statements made in papers published in the Atoll Research Bulletin are the sole responsibility of the authors and do not necessarily represent the views of the Smithsonian nor of the editors of the Bulletin. Articles submitted for publication in the Atoll Research Bulletin should be original papers in a format similar to that found in recent issues of the Bulletin. First drafts of manuscripts should be typewritten double spaced and can be sent to any of the editors. After the manuscript has been reviewed and accepted, the author will be provided with a page format with which to prepare a single-spaced camera-ready copy of the manuscript. COORDINATING EDITOR Ian G. Macintyre National Museum of Natural History MRC-125 Smithsonian Institution Washington, D.C. 20560 EDITORIAL BOARD Stephen D. Cairns (MRC-163) National Museum of Natural History Brian F. Kensley (MRC-163) (Insert appropriate MRC code) Mark M. Littler (MRC-166) Smithsonian Institution Wayne N. Mathis (MRC-169) Washington, D.C 20560 Victor G. Springer (MRC-159) Joshua I. Tracey, Jr. (MRC-137) Warren L. Wagner (MRC-166) Roger B. Clapp (MRC-111) National Biological Survey National Museum of Natural History Smithsonian Institution Washington, D.C. 20560 David R. Stoddart Department of Geography 501 Earth Sciences Building University of California Berkeley, CA 94720 Bernard M. Salvat Ecole Pratique des Hautes Etudes Labo. Biologie Marine et Malacologie Université de Perpignan 66025 Perpignan Cedex, France PUBLICATIONS MANAGER A. Alan Burchell Smithsonian Institution Press ey eae ae Bevan Cul oper: Noveuber 1996. ANDRE GUILCHER 1913-1993 For over 40 years, André Guilcher was an eminent coastal geomorphologist and marine geographer. He was born in Brest, France, received his PhD from the Sorbonne in 1948, taught in several French universities and retired from the Universitiy of Brest, with which he was affililiated, in 1981. He served on many national and international editorial boards and scientific committees, received numerous awards and honors and had been proposed for the last Darwin Award. He started work on coral reefs in the early 1950s, taking part in the first Calypso expedition in the Red Sea. The discovery of reefs was one of immense excitement for Guilcher, and permanently marked his future research work. Subsequent field expeditions took him to Madagascar, Mayotte, New Caledonia, French Polynesia, the Solomon Islands, Kiribati, Vanuatu, Lord Howe Island, Florida, Sinai, the West Indies, Brazil, Kenya and elseware. The results of his work were published in over fifty books and articles. One of his later works was Coral Reef Geomorphology (Wiley, 1988), the first global synthesis of the morphology and typology of reefs, and a thorough review of the evolution of modern regional variations in reef structure and development. This book was also significant in introducing into the international literature important examples of reef geomorphology found in areas that are seldom visited by English-speaking workers. Guilcher assisted and influenced many people in the course of his long academic career and was a model for many of his students. His scientific and leadership skills and his integrity earned the respect of all his colleagues. These and his many friends will greatly miss him. Paolo A. Pirazzoli ATOLL RESEARCH BULLETIN NO. 415 TIKEHAU AN ATOLL OF THE TUAMOTU ARCHIPELAGO (FRENCH POLYNESIA) PART I. ENVIRONMENT AND BIOTA OF THE TIKEHAU (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY A. INTES AND B. CAILLART PART If. NUTRIENTS, PARTICULATE ORGANIC MATTER, AND PLANKTONIC AND BENTHIC PRODUCTION OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY C.J. CHARPY ROUBAUD AND L. CHARPY PART Il. REEF FISH COMMUNITIES AND FISHERY YIELDS OF TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY B. CAILLART, M.L. HARMELIN-VIVIEN, R. GALZIN, AND E. MORIZE ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 FOREWORD In 1982, ORSTOM, a french institution for the development of cooperation in scientific research (Institut Francais de Recherche Scientifique pour le Développement en Coopération) launched the program "ATOLL" in French Polynesia under the dynamic leadership of André INTES, who built up the program while going on with his own research on pearl oyster. The launch of this program took place in a historical context of scientific research and of institutional specialisms. Research on coral ecosystem in French Polynesia had started to develop at the end of the nineteen-sixties and the beginning of the seventies (Research on the atolls of the south-east of Tuamotu was carried out in liaison with the Direction des Centres d'Expérimentations Nucléaires and the implementation on Moorea in the Society Islands, of the antenne Museum EPHE Research Center). In 1974, research started under the auspices of the MAB program (Man and Biosphere) of UNESCO, an interdisciplinary program for ecology and the rational use of insular ecosystems. The objectives of that program were, even at that time (20 years ago), to specify the exact, natural and social sciences, which were the necessary basis for the rational use and conservation of island ecosystems. The two islands of Moorea and Takapoto were selected. A large number of natural, social and medical research organisations from both mainland France and the French Territories participated in the work. A great many reports (1, 2, 3, 4) were produced on these two ecosystems whose coral reefs are among the most studied and documented in the world. With the program "ATOLL", ORSTOM started in 1982 its study of the Polynesian coral reef ecosystem. The choice of Takapoto Atoll was made by the legislature of the Territory in 1973 when the pearl industry was developing, and at the request of scientists. For scientific reasons the latter had asked for the selection of a closed atoll which had no pass : there is less complexity in studying such a lagoon thus giving the best approach to the scientific work. The choice of Tikehau by ORSTOM for the siting of a field station and the development of programs was due to the fact that it was necessary to select an open atoll with a pass to go further in the study of (1) a small scale fishery and (2) the assessment of the exchanges between the lagoon and the ocean. It is in this context that research was developed on Tikehau between 1982 and 1987, research by scientists from ORSTOM and by other organisations working in collaboration with them, and research which has proved to be very beneficial. This station is still in existence with an on-going program, "CYEL" (Energy and Matter Cycles in Atoll Lagoons) having followed on from "ATOLL". In a document edited by ORSTOM, André INTES and his collaborators (5) have already described in minute detail their activities on Tikehau during this five years period. It has now become necessary to condense this information into synopses which are to be published in English. Such is the objective of the three aticles which follow. The first (by A. INTES and B. CAILLART) describes the nature and human environment of the atoll, detailing its characteristics. This contribution is well suitable for publication in Atoll Research Bulletin, now the most important reference for useful descriptions of atolls all over the world. The second article by C.J. CHARPY-ROUBAUD and L. CHARPY deals with matter and energy budgets for Tikehau's coral reefs. This subject is very topical: what about nutrient enrichment for coral reef lagoons and atolls? what about autotrophy and eutrophication due to human activity? The third one by B. CAILLART, M.L. HARMELIN-VIVIEN, R. GALZIN and E. MORIZE presents all the available information on the fish fauna of the atoll lagoon and outer slope, including the communities and yields. Such an account is necessary when looking at the managment of fishing, recruitment, and the movement, or not, of populations between islands. Bernard SALVAT Moorea and Perpignan, march 1993. References (1) SALVAT B. and J. FAGES, 1977 - Takapoto, Tuamotu, Polynésie Francaise. Programme MAB- UNESCO - Compte rendu de recherches 1974-1976. Document roneo, Papeete, march 1977, 150 pp. (2) SALVAT B., 1988 - Bibliographie de l'atoll de Takapoto, archipel des Tuamotu. Bull. Soc. Et. Océaniennes, Papeete, Tahiti, 243 : 55,60 (3) OFAI, 1989 - Bulletin de liaison Centre de l'Environnement Antenne Muséum EPHE de Moorea: 1, 188 (4) INTES A. (ed.), 1991 - Ecosystémes de lagons de la Polynésie Francaise. Rapports et Etudes du PNUE sur les mers régionales n°137, PNUE, 1991 et SPREP Rapports et Etudes n°51, Programme régional océanien de l'environnement, 1991, 298 pp. (5) INTES A., CHARPY-ROUBAUD C,J., CHARPY L., LEMASSON L. and MORIZE E., 1990 - Les lagons d'atolls en Polynésie Francaise : Bilan des travaux du programme "ATOLL". ORSTOM-Tahiti, Notes et Doc. Océanogr., 43 : 1,136 Manuscript received 8 April 1992; revised 30 April 1993 ACKNOWLEDGMENTS All the work done on the field would not have been possible without the kind hospitality and active cooperation of the population of TIKEHAU. It is a great pleasure for the editors to dedicate this volume to the people of this atoll. Many scientists from different institutions contributed to the program "ATOLL" in a friendly collaboration. They came mainly from Centre National de la Recherche Scientifique (CNRS), Ecole Pratique des Hautes Etudes (EPHE), Museum National d'Histoire Naturelle de Paris (MNHN), Institut Francais de Recherche pour I'Exploitation de la Mer (IFREMER), Service Mixte de Contréle Biologique (SMCB), but also from the ORSTOM center of Nouméa (New Caledonia). Most of them became our friends. The preparation of the report involved few people who gave unrestricted thought and time to bring this collective work into publication. We wish to extend our thanks to all of them, listed below, regardless of the nature of their contribution. Some of them may have been overlooked, and we apologize for these unfortunate oversights. Arnaudin H. Blanchet G. Blanchot J. Bonnet S. Bourrouilh-Le jan F. Cremoux J.L. Faure G. Florence J. Galzin R. Harmelin-Vivien M. Jamet R. Laboute P. Le Borgne R. Lenhardt X. Moll P. Monnet C. Monniot C. Monniot F. Orempuller J. Paoaafaite J. Peyrot-Clausade M. Poulsen M. Poupet P. Saliot A. Sandstr6m M. Sarazin G. Sodter F. Teuri J. Trichet J. Vigneron E. Villiers L. edited by André INTES'~ - Senior Editor Benoit CAILLART Loic CHARPY Claude J. CHARPY-ROUBAUD Lionel LEMASSON Eric MORIZE PART I. ENVIRONMENT AND BIOTA OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY A. INTES AND B. CAILLART THE REGIONAL BACKGROUND The islands of French Polynesia are scattered throughout a considerable oceanic area located on the eastern boundary of the Indo-Pacific Province. This area stretches from 134°28' W (Temoe Island) to 154°40' W longitude (Scilly Island), and from 7°50' S (Motu one Island) to 27°36' S latitude (Rapa Island). Out of the 118 islands constituting French Polynesia, 35 are high volcanic islands and 83 are low-relief islands or atolls. Altogether, the territory of French Polynesia represents an area of 4000 km2 of dry land, 12,000 km? of lagoonal water and a huge Exclusive Economic Zone (EEZ) covering 5,500,000 km2 of oceanic water (Gabrie and Salvat, 1985). French Polynesia is divided into five archipelagos all oriented parallel to a northwest- southeast axis (Fig. 1). These are the Society archipelago, the Tuamotu archipelago, the Austral archipelago, the Marquesas archipelago and the Gambier archipelago. The Tuamotu archipelago stretches over a distance of 1800 km. Its 76 atolls cover a total area of 13,500 km2 of which 600 km2 are dry land. GEOLOGY OF THE TUAMOTU ARCHIPELAGO As figured by Montaggioni (1985), the Tuamotu atolls cap the top of cone-like volcanoes which rise steeply from the floor of a huge ridge forming wide shelves ranging in depth from 1,500 to 3,000 m. Geomorphological and geochronological evidences support the fact that the formation of the Tuamotu chain is much older than that of other neighboring islands of French Polynesia. The foundations of extinct volcanoes appear to have been simultaneously, and not sequentially, active for at least the Northwestern Tuamotu chain. The existence of a massive submerged ' ridge and the lack of high volcanic islands are in accordance with average ages found out by the Deep Sea Drilling Project (summarized by Clague, 1981 and Schlanger, 1981) : reef debris of the early to late Eocene (50-51 mybp) have been sampled in two holes drilled on the northeastern flank of the Tuamotu archipelago and on the ridge itself. The occurrence of these fossils leads to the conclusion that vulcanism would have stopped between late Cretaceous and early Eocene for at least the northwestern part of the Tuamotu chain. The large number and the close-spacing pattern of the Tuamotu atolls are indicative of their origin in shallow waters close to the East Pacific ridge. Centre ORSTOM de Brest, B.P. 70, 29280 Plouzané, France ue: “Motu One Ratu hi Noku Hive oe =a Hoke Fatu Huku ° . Md ew hive Os Tehuals # + Mohotani 4 “Motu Nao ”) S TIKEHAU ii Tele epuha ARCHIPE; vane Wr oo d LES SOUS d Beene Joa, Kavehi Takume, OFangaiau EQ sphaata —Taengs PRaroia € oo Fokahne any Siw Ma vf ORES AS TepaTOISR Rebate Anas Motutungs Sete ol _Tevere alatakoto ‘16° Nau Fake Anuanurunga~ - Nukutepipi Vanavana, Turew Tenararo Tenor Vahanga sein Marutea(S) > Mowuroa tureivavad “Mana ILES GAMBIER lematang: eFangateuta © JCJ ORSTOM TAHITI. 025/03-85 Fig. 1 : Map of French Polynesia and location of the Tikehau atoll, Tuamotu - 3 In the northwestern Tuamotu, several atolls (including Tikehau), located in the vicinity of recently active volcanoes (Tahiti, Moorea, Mehetia), have been lifted up. Lambeck (1981) pointed out that the tectonic uplift of these atolls was a result of the loading effects of the nearby Tahiti volcanic complex. The magnitude of uplift was a few ten meters with respect to present sea level, without considering the unusual case of Makatea which, located near the center of the load, has its highest point 133 m above present sea-level. Since age dating of the oldest volcano (Moorea) is around 1.5 mybp, the tectonic uplift of atolls is thought to have been initiated in early Pleistocene. At Tikehau, the magnitude of uplift can be related to the present day elevation of the numerous old reef remnants (locally termed Feo) that stand on the atoll rim (Plate 1). Feo are highly recrystallized and dolomitized old reef remnants, that witness a long period of subaerial weathering. Although there is insufficient evidence to accurately date the Feo, a comparison with analogous structures on Makatea confirms their emergence in the early Pleistocene (Pirazzoli and Montaggioni, 1985). On the southern shores of Tikehau, Feo can be as high as 12m above present sea-level. Many other lower Feo are located on the eastern and western reef flats. Studies from Holocene fluctuations in sea level can be useful in interpreting individual reef histories. In Tikehau as well as in most of the islands of French Polynesia, cemented coral reef conglomerates, often found around islands and alongshore shallow channels, and notches, undercutting Feo near their bases are evidences to support a higher-than-present relative sea level between 5200 and 1200 ybp. This higher sea level may have been 0.9 m greater than the present sea level. In this region, the lowering of the sea level near to its present datum is thought to be a very recent phenomenon which has occurred not earlier than around 1200 ybp. GENERAL OCEANOGRAPHY OF THE TUAMOTU ARCHIPELAGO Current and oceanic water characteristics Tikehau, as all the northwestern parts of the Tuamotu archipelago, resides in the South Equatorial Current. The current has a general westward drift between 40 to 50 cm s-! down to 200 m depth, steady throughout the year. The southern boundary of the current moves northward during the southern hemisphere summer and shifts back toward the south in winter. Currents near the atolls may vary in both speed and direction due to the dynamic topography and rather permanent eddies probably exist, though their occurrence has never been demonstrated in French Polynesia. In the south of the northwestern part of the Tuamotu archipelago, variations in the current directions are considerable and may influence to some extent oceanographic patterns around Tikehau. Rougerie and Chabanne (1983) pointed out that during the summer, the current in the vicinity of Tahiti may be an extension of the South Equatorial Countercurrent which originates in the Solomon Sea and has a general eastward-southeastward drift. The salinity is low (34.8 %o) reflecting the annual net rainfall in the South Pacific Convergence Zone (SPCZ) which occurs along a Solomon-Samoa-Tahiti line at this time of year. During winter, the SPCZ shifts northward and the trade winds strengthen somewhat over the Tahiti region. The current flows westward carrying high salinity water (36.4 %o) drawn from the Central Pacific barren zone where evaporation exceeds precipitation by 50 cm per year (Rougerie, 1981). Surface water temperature varies seasonally in a spread of 25.5°C to 29°C and reaches an average value of 28°C. The main thermocline is weak and is located between 400 and 600 m in depth with a temperature of 10°C at 400 m. By 1000 m, the temperature drops to 3°C. 4 Chemical data for ocean water (summarized in Table 1) show that the ocean surrounding the Tuamotu archipelago is nutrient-poor and is actually one of the poorest oceans in the world. For example, the average copepod surface population living in that oligotrophic environment is of 20 individuals m-3 near the surface while it is between 50 and 80 between New-Caledonia and Tonga, greater than 100 in the south of the Coral Sea, reaches 300 at the equator and exceeds 1000 to the south of Galapagos Islands. Table 1 : Average chemical data for ocean water (0 to 100 m in depth) in the Tahiti zone. (DOM : Dissolved Organic Matter, chl a : Chlorophylle a, PProd : Primary Production, * : data measured during winter of southern hemisphere, ** : data measured in summer). Adapted from Rougerie and Wauthy (1985). Temp Sal PO4P NO3N _ SiO3 DOM chl a PProd (°C) (%o) (mmoles m-3) mg m-3 gCm2g-1 25.5% 36.4* Ocean 0.25 0.10 1.0 0.5 0.1 30 29.0% 34.8** Waves and Tides Waves in the Tuamotu archipelago are mostly from the east, a consequence of persistent trade winds. Waves are generally between 1 and 3 m high, breaking at a 6 to 9 s period. Swells from distant storms can reach the Tuamotu and create a different situation in which shores exposed to the swell (which can be either windward or leeward) are heavily pummeled, whereas the waves produced by the wind in the immediate area may be small. Northern hemisphere storms that occur in the Alaskan gulf during summer of the southern hemisphere generate waves from north-northwest which are generally about 4 m high and break at a period between 10 and 18 s on Polynesian shores. In winter, southern hemisphere storms generated in the lower southern latitudes may send associated waves to the Tuamotu zone. These 7 to 10 s period waves can exceed 3 m high and reach the southeast shores of the islands. Several cyclones and near- cyclones have passed by or over the Northwestern Tuamotu in 1982 and 1983. These storms have produced waves greater than 10 m high from various directions related to the storm track. The tides at Tikehau are usually in good agreement with the French Navy (SHOM) tide table established for Tahiti. However, ocean tide records at Tikehau show differences in timing. The time-lag between the tide at Tikehau and the tide at Tahiti is approximately 72 hours. The amplitude of spring tides is only 15 cm in the vicinity of Tikehau whilst neap tide amplitude is almost zero. Spring tides occur three days before the new moon and the full moon. WEATHER AND CLIMATE The Tuamotu archipelago lies in the tropical oceanic climate area. It has a distinct wet-dry annual cycle. The wet and hot season occurs from November through April and the dry and cold season from May to October. Since all of the islands are low and of a small area, they do not alter weather conditions by their presence. Data presented hereafter were provided by the National Meteorological Station of Rangiroa, except when otherwise mentioned as no weather records are available at Tikehau. Weather conditions have been recorded continuously from 1972 onward by this station. Air Temperature Average air temperature on a monthly basis for the 1972-1985 period ranges from 25.5 °C in August to 27.5 °C in March. There is little variation in these quantities through the year. Extreme temperatures have been recorded but provide little additional information since they rarely occur. The absolute minimum value recorded was 18 °C and the maximum 32 °C. Wind A summary of surface wind data is shown in Fig. 2. Tikehau is within the trade wind belt with a nearly consistent easterly wind. During much of the year, the wind blows from northeast to southeast 70 % or more of the time (i.e. 250 days/year). From June to September average wind speed increases slightly but rarely exceeds 6 of the Beaufort scale. The occurrence of wind from west around to north is very low. The maximum frequency is in November, December and January when the South Pacific Convergence Zone is closest on the average and subsequently, disturbances most common. Tropical storms and cyclones strike Tikehau infrequently, mostly during the wet season. An average of four cyclones per century is likely to occur in this area. Precipitation, evaporation The rainfall distribution throughout the year is shown in Fig. 3. The annual average rainfall of 1780 mm is not distributed uniformly throughout the year as about 65 % comes during the wet . season. The maximum monthly average value is 229 mm in January and the minimum 75 mm in August. The variability of rainfall is high from year to year and data presented should be considered only as a general trend. Evaporation reaches an average of 1800 mm a year and balances precipitation, as measured by the National Meteorological Station of Mururoa. Maximum evaporation takes place in December, January and February (179 mm, 193 mm and 188 mm), minimum values in June, July and August (109 mm, 109 mm and 124 mm). Frequency of calms : 37% wee Wind speed greater than 5 m s-1| Wind speed below 5 m s-1 Frequency scale ty ) SO */ee Fig. 2 : The yearly mean of surface wind data recorded at Rangiroa (Rangiroa is located 30 km to the east of Tikehau). Data provided by Meteorologie Nationale, Tahiti-Faaa. Rainfall (mm) 250 200 150 100 10 J F M A M J J A S O N D Vea Fig. 3 : Average rainfall amounts for each month recorded at Rangiroa (Rangiroa is located 30 km to the east of Tikehau). Data provided by National Meteorological station of Tahiti-Faaa. LOCATION AND SIZE OF TIKEHAU ATOLL Atolls can be described as more or less continuous coral reefs (corals or other calcium carbonate producing organisms) which surround a deeper lagoon and drop steeply to oceanic depth on the seaward margin. All islands are typically low with soil derived primarily from reef rubble and sand. Although almost identical in their general shape, atolls of the Tuamotu archipelago are different by the characteristics of their lagoon and the number of passes on which depends the amount of water circulation. As shown in Table 2, a few atolls have one or two passes and are termed open atolls, most have no pass and are termed closed atolls, four atolls have a dry lagoon, filled up by reef detritic and are termed filled atolls, one island, Makatea, has been uplifted by tectonic movements probably linked with the formation of Tahiti and is termed a raised atoll. Lastly, Portland is a submerged atoll . Table 2 : Classification of the atolls of the Tuamotu archipelago based on the geomorphological characteristics of their lagoon. Atoll group Example Open atolls with more than one pass Rangiroa Fakarava Toau Open atolls with one pass Tikehau Mataiva Arutua Closed atolls Takapoto Hikueru Reao Filled atolls Akiaki Nukutavake Tikei Raised atoll Makatea Submerged atoll Portland Tikehau has a large somewhat elliptically shaped lagoon, numerous shallow channels cutting the reef flat especially on the windward side, one pass between the lagoon and ocean, and narrow shelves dropping steeply into deep sea on all sides. A succession of small islands, locally termed Motu, constitute the dry land. 8 Tikehau is located in the northwestern Tuamotu archipelago with its center at 15°00'S and 148°10'W (Fig. 1). It is approximately 300 km from Tahiti to the south, 30 km from Rangiroa to the east and 20 km from Mataiva to the west. By French Polynesia standards, Tikehau is a relatively large elliptical atoll, about 20 km by 28 km in size, covering a total area of about 420 km2 . Among the 76 islands of the Tuamotu archipelago, Tikehau is the 11th largest. It is exceeded, among others, by Rangiroa (1640 km2 ), Fakarava (1220 km2 ) or Makemo (910 km2 ), the three largest islands. Small atolls are generally a pattern of the Tuamotu since 45 out of the 77 islands cover less than 100 km2 (Mataiva :50 km2, Taiaro : 14 km2, Tikei : 4 km2 ). THE TERRESTRIAL ENVIRONMENT Since few scientists conducted research in the terrestrial environment with respect to the marine environment of the Tikehau atoll, little is known about this part of the ecosystem. This section contains a review of what is known about atoll soils and associated vegetation, reports on the terrestrial fauna (birds and other vertebrates), and finally provides an overview of Tikehau human population evolution over the last century. Information is drawn from Jamet (1985) for soils, Florence (1985) for vegetation, Poulsen et al. (1985) for avifauna, and Sodter (1985) for human demography. SOILS AND VEGETATION Tikehau soils exclusively originate from the alteration of a mother rock made up of reef forming or reef living organisms. The micro-splitting of particles mostly through chemical processes leads to the formation of clay or carbonated silt in which the percentage can reach 40 % in sandy soils. Carbonates accumulate above the upper zone of the groundwater lens forming a calcareous crust. Organic matter is usually mixed with fine materials in a topsoil horizon fairly thick but it can also accumulate superficially in marshy depressions. Following these alteration processes, Tikehau atoll soils fall into four types : - Rough mineral soils formed of accumulations of unaltered recent sediment. - Weakly developed soils made up of coarse materials with low content of organic matter, found mostly on the oceanic shoreline of the island. - Magnesium-calcite soils with a dark, developed A horizon more or less thick and with variable organic matter content. This kind of soils covers the majority of the islands at Tikehau. - Marsh soils very rich in organic matter, located in island floor depressions (Plate 2). Unlike high island soils, atoll soils almost lack silica, aluminium and iron. The mineralogy is almost exclusively calcium carbonate ( 80 to 95% of calcite and aragonite) which fine soluble particles represent 3 to 5 % in rough soils, more than 20 % in humus horizon, and exceed 30 % in marshes. Magnesium carbonate represents less than 1 %. Potassium and phosphorus content are generally less than 0.05 % in subsoils but fecal matter of birds and vegetation remains locally contribute to a ten fold increase in K and P concentrations of topsoils. Sodium concentration is lower (0.5 %) than expected in this kind of ecosystem. The fertility of atoll soils is almost entirely dependant on the content of organic matter. Accounting for less than 2 or 3 % in rough soils, organic matter can reach 15 % in humus horizons and much more in marshy areas. Organic matter not only carries out the normal role of soil organic matter in storing and recycling nutrients, but it is also the major moisture storage component in the soils, since coral sands and rocks have an extremely limited moisture storage capacity. This is all important in atolls where evaporation exceeds precipitation eight months 9 of the year. Organic matter lowers pH which ranges between 8 and 9 in subsoils. The pH is almost neutral in topsoils, leading to a better nutrient assimilation. Atoll soils have considerable influence on the composition of vegetation. There is a marked gradient from the beach toward the center of the island. - Unaltered rough sediment soils are constantly rearranged and do not enable settlement of durable vegetation. - Weakly developed soils constituted by coarse materials oceanward and fine sands lagoonward have a low organic matter content in the first 10 or 20 cm, forming patches or stretches. Two types of vegetation settled there : Vegetation of the oceanic side of the island is a low and open assemblage of Guettarda speciosa, Scaevola sericea and Tournefortia argentea. Behind the beach, the assemblage gets richer with Euphorbia atoto, Timonius polygamus and Pandanus tectorius. On cemented coral substrata, Pemphis acidula forms bushes and on sand patches, Suriana maritima and Lepturus repens develop. Vegetation of the lagoon side of the island is still well represented eastward the atoll whereas it was cleared by coconut plantations in other places. Bush assemblage is dominated by Suriana maritima and Scaevola sericea, but Guettarda speciosa and Tournefortia argentea occur sporadically. The herbaceous stratum is varied with Triumfetta procumbens and Lepturus repens dominating. Magnesium calcite soils (found in the center of the island) cover the area of the forest. Although coconut plantation cleared much of the original vegetation, two facies can be distinguished : - On weakly developed soils, Pandanus tectorius dominates the tree stratum along with Tournefortia and Guettarda. Among the bush, Scaevola sericea, Pipturus argenteus and Timonius polygamus were recorded. Herbaceous vegetation is rare : Psilotum nudum, Cassytha filiformis and Nesogenes euphrasioides were recorded nonetheless. - On sandy soils, the original forest of Pisonia grandis has almost disappeared whereas Guettardia speciosa still occurs. Bushes are made up of Pipturus argenteus, Morindia citrifolia or Euphorbia atoto. The herbaceous stratum is varied with Achyranthes velutina, Laportea ruderalis, Digitaria stenotaphrodes, Boerhavia tetrandra. Coconut plantations Sand and gravel soils which are the most favourable for coconut agriculture have been planted mostly during the last century. The western coast has coarse substratum only allowing coconut trees to be planted on the lagoon side of the islands. From place to place, components of the original vegetation are encountered such as Guettarda speciosa, Pisonia grandis or Pandanus tectorius for trees, Euphorbia atoto or Morinda citrifolia for the bush, Lepturus repens, Boerhavia tetrandra and Triumfetta procumbens on the ground. Feo vegetation Feo are located north of the island supporting the main village and surrounded by coconut plantations. Being as high as 7 meters, they present a compact substratum with a low moisture storage capacity, and tiny soils in small caves. Tree stratum is composed of Pandanus tectorius and Thespesia populnea. Bush is made up of Pipturus argenteus, Euphorbia atoto and rarely Capparis cordifolia. On the ground, Lepturus repens and Triumfetta procumbens were recorded but also ferns such as Asplenium nidus, Nephrolepis biserrata or Phymatosorus grossus. In a hydromorphic depression, vegetation is a Cyperaceous assemblage. Cladium jamaicense is so overwhelmingly dominating that the assemblage is almost monospecific, reaching 3 m high. On the edges of the marsh, Mariscus pennatus and Eleocharis geniculata are found. 10 THE TERRESTRIAL FAUNA Avifauna Sedentary terrestrial species widespread in all atolls are Tuamotu warbler (Acrocephalus atypha) of which 150 pairs were counted at Tikehau, at least 10 pairs of green pigeon (Ptilinotus coralensis), between 150 and 200 individuals of pacific reef heron (Egretta sacra). Tuamotu loriket (Vini peruviana) resides in a few atolls of western Tuamotu and breeds on Tikehau on western islands. About 20 pairs were counted. Sooty crake (Porzana tabuensis) was observed in the marshy area of the island supporting the village. New Zealand cuckoo (Eudynamis taitensis) is the only migrant species exclusively terrestrial sighted in autumn (May). Shorebirds : especially present in winter of southern hemisphere, the most readily observable species are lesser golden-plover (Pluvialis dominica), bristle-thighed curlew (Numenius tahitiensis), wandering tattler (Heteroscelus incanus). Less abundant though regular visitors of Tikehau, ruddy turnstone (Arenaria interpres), sanderling (Calidris alba) and pectoral sandpiper Calidris melanotos account. Seabirds : Red-footed booby (Sula sula) forms some small nesting colonies in trees (Pisonia grandis). Brown booby (Sula leucogaster) and frigatebirds (Fregata ariel and great frigatebird Fregata minor) regularly occur but no breeding evidences were recorded. Terns are well represented with gray-backed tern (Sterna lunata) of which a 20 pairs nesting colony was sighted, sooty tern (Sterna fuscata) which seems only vagrant, great crested tern (Sterna bergii) of which about 50 pairs nest on the atoll, blue tern (Procelsterna coerulea) of which a few pairs are supposed to nest at Tikehau, the very common brown noddy (Anous stolidus) of which 1500 nests were censused in small trees (Pemphis, Tournefortia), 800 nests of black noddy (Anous tenuirostris ) in trees (Pisonia, Guettarda), and lastly a nesting population of 3000 individuals of white tern (Gygis alba). Other seabirds as petrel (Pterodroma rostrata) and skua (Stercorarius pomarinus) sometimes approach Tikehau. Other vertebrates Except marine green turtle (Chelonia mydas), reptiles reported to occur at Tikehau are lizards azure-tailed skink (Lygosoma cyanurum) which occurs in sunny forest floor, and house gecko (Hemidactylus frenatus). Among mammals, rodents are best represented with polynesian rat (Rattus exulans) and Norway rat (Rattus norvegicus). Some domestic cats returned to the wild were seen wandering in coconut plantations and marshes. 11 PEOPLE OF TIKEHAU, PAST AND PRESENT Tikehau could have been one of the atolls discovered by Turnbull in February 1803 when he was sailing from Tahiti to Hawaii, but neither positions nor name were given. In 1816, Otto von Kotzebue, Russian master of the "Rurick" first identified the atoll of Tikehau and gave it the name of one of his shipmates : "Krusenstern". In May 1848, the atoll was once again described by a trader, Lucett. Given the poor knowledge of the danger of the island, none of those first discoverers attempted to land and as a result, nothing is known about the population during this period. However, archeological remains of the past (marae) attest that Tikehau was inhabited at the time of the Christ. In all probability, age dating of the sites would push the date for settlement of Tikehau further back in time. The first information on the total number of inhabitants was given in 1862. Ten persons were counted but this total is probably unrealistic because the census method did not take into account the frequent seasonal movements of the population around the island. In 1902, an official census made by French Authorities gives a total of 156 inhabitants for both Tikehau and Mataiva. The following census in 1911 gives a total of 95 inhabitants on Tikehau and 58 on Mataiva. Until the end of World War II, no information on Tikehau and Mataiva population is available. In 1946, 376 persons had been registered on both Mataiva and Tikehau. From 1950 to 1983 From 1950 on, census at regular intervals allow to follow population variations. Census reliability depends however on how the important mobility of the population is taken into account. Census methods sometime tend to overestimate the population by counting a single person twice (in 1956 for instance). As illustrated by information displayed in Table 3, the level of the population of Tikehau in 1983 is slightly the same as in 1951 though the total population of French Polynesia doubled meanwhile. Trends in population variation are difficult to analyse without precise individual information on place of birth and location of main home. Table 3 : Variations of Tikehau, Mataiva, both Tikehau and Mataiva, and French Polynesia populations between 1946 and 1983. All census were made by French Authorities. Results of 1956 census are to be interpreted very cautiously since census method tends to over-estimate the actual population by counting some single person twice. 1946 1951 1956 1962 1967 1971 1977 1983 Tikehau -- 259 349 275 287 246 266 279 Mataiva -- 126 241 162 138 147 178 183 Both 376 385 590 437 425 393 444 462 French Polynesia 55,424 62678 76,327 84551 98,378 119,168 137,382 166,753 12 The Population in 1983 A typical feature of the population of Tikehau , like of the population of French Polynesia, is the high proportion of young people : 28.7 % of the population is less than 10 years old and 52.7 % less than 20. The worker age class is very low, especially between 30 and 39 which represents only 5 % of the population. The high migratory rate of the population, the main source of difficulties to accurately measure the population, is confirmed by the proportion of Tikehau inhabitants (21 %) having spent more than six months in a row in another district of French Polynesia. For 91 % of the people having moved, the island of Tahiti is the main destination, and 70 % of the 91 % go to Papeete. The working population is composed by 90 persons, 76 men and 14 women. Among the 76 men, 50 put up with copra production, 14 are fishermen, 9 are artisans, 2 are employed by the government and 1 is administrative officer. Among women, 8 are administrative officers, 3 are shopkeepers and 3 are artisans ; 3 men and 33 women have declared to be looking for a job. Conclusion The main characteristics of Tikehau population are: - High proportions of young people under 20 years old and, as a consequence, low proportions of people in the worker age class. - High migratory rate. - Men professional activities primarily oriented toward agriculture, especially copra production. Women professional activity is low but high employment request exists. 13 THE MARINE ENVIRONMENT PHYSIOGRAPHY OF THE MARINE AREA Definition of reef units In order to classify reef units of the atoll, Faure and Laboute (1984) described three main types of reef units as follows : - Compartments defined on a physiographic basis - Zones assessed on a morphological basis - Zones divided into biota according to bionomic field data The three compartments are the outer slope, the reef flat, and the lagoon formations. A fourth virtual compartment termed morphological discontinuities grouping pass and shallow channels is also studied. Fig. 4 lists the different units classically found around an atoll. BIOTA ZONES COMPARTMENTS Lower part >60-70 m | ) Middle part 35-60 m_ | Deep slope 25-75 m | Upper part 25-35 m i) | Lower part 15-25 m_ ) outer terrace 10-25m | OUTER SLOPE Upper part 0-15 m ) Fore reef platform 4-10m_) Fore reef area 0-10 m Spur and groove zone 0-4 m Algal ridge Submerged reef flat Inner flat Inner slope outer reef flat inner reef flat inner slope Coral patches Upper zone 0-2 m Middle zone 2-6m Lower zone 6-15 m Pass Gutter (or Hoa) Low depth coral patches Pinnacles Pass and Hoa MORPHOLOGICAL DISCONTINUITIES We eS i Se FS ess SS SS ) ) ) ) Down slope reef patches ) | | ) ) ) Fig. 4 : Inventory and repartition of the different atoll units (from Faure and Laboute, 1984), 14 The outer slope The outer slope is the seaward part of an atoll which drops more or less steeply to oceanic depth. In Tikehau, the outer slope is divided into three zones : - The fore reef area located between 0 and 10 m depth. This zone itself is subdivided into the spur and groove zone (0-4 m) and the fore reef platform (4-10 m). The spur and groove system is a succession of reef fingers projecting seaward where the waves break, oriented perpendicular to the reef front. Spurs are relatively flat on top. Their width ranges from 8 to 12 m with a low slope gradient of 2 to 4°. Grooves are shallow (1 to 3 m) and relatively narrow (2 to 3 m). The walls of the grooves are sub-vertical and their bases are floored with cobbles and dead coral boulders, precluding the development of any significant sessile benthic life (Plate 4). Just seaward of the spur and groove, the bottom flattens somewhat with the fore reef platform. - The outer terrace begins at a depth of 10 m with a distinct change from a gentle slope of a few degrees to an angle of approximatively 45 °. It presents an irregular surface with small periodic shallow grooves oriented parallel to the slope direction, well distinguished from the spur and groove formations described above. This structure could be considered as an old spur and groove system (Chevalier, 1973). - The deep slope (or drop off) begins below 25 m at an angle often greater than 45°. According to coral community distribution, the slope can be separated into three biota : the upper part (25-35 m), the middle (35-65 m) and the lower part from 70 m downwards. The reef flat The outer reef flat begins seaward by the algal ridge and ends lagoonward by the emerged conglomerate or the island. On leeward reef, the algal ridge is low (20 cm above low tide level) and has a width of 10 m while on windward shelves, the algal ridge is larger (30-40 m) and higher (40 cm above low tide level). Numerous deep grooves (1-3 m) extend across the windward algal ridge. Just inshore of this formation, a slight depression of the reef flat can occur, especially on the windward sides of the atoll. To this follows an emerged hardened conglomerate which can be considered as the remains of a past algal ridge. As shown in Fig. 5, the outer reef flat morphology varies considerably in different area, particularly between the windward and the leeward sides. Its width ranges from 150 to 180 m as measured on West- Southwest transects to 20 to 40 m on eastern shores. It consists of an area of rock pavement derived from an old conglomerate submerged under 10 to 50 cm of water with a rough bottom, pitted by small erosional pools. The floor is covered by a thin sedimental detritic layer that get thicker toward the beach. The inner reef flat begins just lagoonward of the motu or the emerged conglomerate, and ends where the bottom starts to slope into the lagoon. The width of the inner reef flat varies considerably around the atoll rim. On windward shores, the inner lagoon margin flat is protected from trade winds by the outer reef flat, the motu and its associated vegetation. The inner reef flat is therefore somewhat dead and resembles a slight sandy slope. A different situation exists on the lagoon border in the lee of the atoll. Because of exposure to trade wind across the fetch of the lagoon, these areas are well-formed reef flats composed of an old conglomerate covered by a thin sedimental layer that gets thicker lagoonward. Fig. 5 : Cross atoll rim sections at various locations of the Tikehau atoll (OS : outer slope; AR : algal ridge; OF : outer reef flat; MO : island; CO : conglomerate IF : inner reef flat; IS : inner slope; LA : lagoon; PR : pinnacle reef) - from Harmelin-Vivien (1985). Lagoon structures Immediately after the inner reef flat, the lagoon inner slope is a relatively steep sediment and rubble slope which extends to depth of 2-6 m. When the slope begins to flatten out, numerous coral patches protrude from the sediment bottom. Where the water flow across the reef is ‘ unimpeded by islands, these coral colonies can be numerous and healthy. The lagoon floor is essentially flat with a low slope gradient. The bottom is primarily a fine sandy bottom. This is studied in more detail in a subsequent chapter of this volume. Many coral pinnacles are scattered all over the lagoon. Up to 300 pinnacles visible from the surface (locally called Karena) have been counted. They are quite unevenly spaced throughout the lagoon since half of them can be found in the southwest part. Pinnacle reefs vary greatly in size, ranging from a few tens of meters to over 200 m in diameter for the largest. The shape of the pinnacles can be roughly related to their size : the smaller the pinnacle reef's diameter, the steeper its slope. On small pinnacles, much of the slope is nearly vertical. The largest pinnacles are somewhat flat on top and an emergent one can support bushes or small trees. However, they still slope to the lagoon floor at an angle of at least 20°. Pinnacle reef shapes also vary considerably between the windward and the leeward sides. The windward side usually presents a steep subvertical slope whereas the leeward side of the pinnacle gently slopes to a detrital zone. Coral community zonation is rather regular among pinnacle reefs, and three biota may be recognized : the upper (0-2 m), middle (2-6 m) and lower zones (6 m downward). 16 Morphological discontinuities Passes can be defined as major deep channels between the ocean and the lagoon. In Tikehau, the only pass, called Tuheiava, is located at the western part of the atoll. Localisation of the pass on the leeward side of atolls appears to be a general trend of the Tuamotu archipelago. Minimal depth of the pass is about 4 m and is sufficient for a small boat to traverse. The bottom, of bare eroded flagstone, slopes steeply seaward and gradually deepens lagoonward until it merges with the lagoon floor. Current direction in its vicinity reverses, depending on the tide and on the height of water in the lagoon. When tidal currents run against the trade winds, steep standing waves can occur in the pass, hampering sailing and fishing activities. When sea- level is high in the lagoon, the current is mostly unidirectionnal out of the lagoon. This occurs frequently when storm associated waves break on windward reef and drive large quantities of water into the lagoon. On an annual basis, the net flow through the pass is an outflow. Main features of water circulation at Tikehau are studied out in details in a subsequent chapter of this volume. Hoa, also termed rips or gutters, are shallow channels which cut the reef flat superficially. They draw their flow from the shallow reef flat and channelize the flow of water into the lagoon between the motu (Plate 5). In Tikehau, their width ranges from a few tens of meters to 500 m. When currents can freely flow through hoa across the atoll rim, hoa are termed open or functional hoa. On the contrary, when currents do not flow or flow only when a storm occurs because the channel is obstructed by boulders on the outer flat and/or is closed by littoral sand shoals or rubble accumulations lagoonward, hoa are termed closed or non functional hoa. The shape of these channels is subject to major changes owing to sedimentation and high erosion. Hoa are shallow on the outer reef flat (10 to 20 cm) but deepen towards the lagoon (1-3 m). On their lagoonward end where current flow slows down, hoa usually have shallow sand shoals. Hoa have sandy bottoms sculptured by current fluxes with more or less patch reefs. The amount of patch reefs are directly related to the intensity of water circulation through them. Up to 150 hoa have been counted around the atoll rim. More than 100 are concentrated on the southeast coast and are mostly open channels. The current flow through them is almost unidirectional, from ocean into the lagoon, depending on water level in the lagoon. The 50 other hoa are principally located on the northwest coast. A few of them are functional. Hoa, as well as the pass, are the major source of water movements between the ocean and lagoon. BATHYMETRY OF THE LAGOON Knowledge of the bathymetry of the lagoon is basic data for all marine research carried out on the lagoon. There is no detailed marine chart of the lagoon available and an attempt to map the Tikehau lagoon bottom from the satellite LANDSAT proved to be unsuccessful. The main cause was the high and uneven turbidity of the water in the lagoon which prevented LANDSAT from mapping the bottom efficiently. The bathymetry of the lagoon of Tikehau was mapped using field measurements (Lenhardt, 1987). A SIMRAD EY-M echo-sounder (frequency 70 khz, range 0-60m) was used on board a motor boat and depth continually recorded on eight transects across the lagoon. The boat was steaming at a regular speed of 3.2 knots on the magnetic North-South axis (declination : 13 °). Results obtained have been cross-checked with results from five other transects. To discretize the continuous series of data, one depth measure was taken every 80 mand raw data smoothed to eliminate numerous local minor unevenness of the bottom. Pass morphology was studied by recording depth on three transects, one through and two across the pass. 17 Bathymetric map Smoothed depths recorded on transects were contoured to map the bathymetry of Tikehau (Fig. 6). The greatest depth recorded was 38 m in the central northeastern part of the lagoon. The main lagoon basin appears to be a relatively flat area with gentle slopes. Depth histogram The histogram presented on Fig. 7 shows that the general shape of the lagoon is that of a basin with steep walls. Depths between 0 and 15 m represent only 7 % of the total surface of the lagoon S (S = 420 106 square meters). Other geometric data The mean depth P as well as the total volume V of the lagoon is of great interest for other studies. Mean depth is calculated by computing the mean depth of each transect with mean length and width of each transect. The result is an average depth of P = 25 m. The total volume of the lagoon is calculated by multiplying the total surface of the lagoon by its mean depth. The result is V = 1019 cubic meters. The confidence interval of those data is about 5% which meets the requirement for precision of other research. “Mott Deoe « 35 15°05’ S \ 148°05'W \. Tuherahera (village) afi Fig. 6 : Contoured depths (in m) in the lagoon of Tikehau. 18 % 30 20 10 Fig. 7 : Average Tikehau lagoon surface per 5 m depth intervals expressed as a percentage of total lagoon surface . Characteristic of the pass Fig. 8 shows the bathymetric map of the pass, transects to record the depth and the smoothed shape of the bottom. The minimal depth recorded in the pass is 4 m. For further modelisation of the flow through this channel and water circulation in the lagoon, the mean section of the pass is estimated at 1000 square meters and average length at 600 m. on ESWC ay es Lary I SLRS GEREN Re A AAAI Kt wt PLSS TUMHELIAVA LAGOON Fig. 8 : Bottom profile of Tikehau atoll pass. 19 NATURE AND DISTRIBUTION OF LAGOON SEDIMENTS The bottom sediments of the Tikehau lagoon were characterized by Intes and Amaudin (1987). Fifty four samples of bottom sediment were taken in various location of the lagoon and subsequently sorted and analyzed in the laboratory. Findings were that the sediments are all calcareous organic sands. They consist of the following chief components : Halimeda segment sand, Foraminifera (family Miliolidae) test sand, mollusc shell sand and miscellaneous debris. Halimeda segment sand is the first common material. It is present in almost all samples and predominates in most of them, especially in samples taken in the western part of the lagoon. As shown on Fig. 9, Halimeda segment distribution follows an horizontal west-east gradient and a vertical gradient since its abundance decreases steadily below a 20 m depth. Foraminifera test is the second most common material. Large stretches of foram sand occur on shallow bottoms less than 10 m deep along the southern and southeastern lagoon margin. Moreover, a large foram sand patch extends in the northern central part of the lagoon between a depth of 20 and 35 m. Mollusc shell sand and gravel are never abundant and never dominate the sediment composition. NS SS // Y S Mi SS. SS yi , [LT LE S = Ta M ~~ i) EMM Ci? Spy i) dy) Fig. 9 : Contoured bottom sediment data of Tikehau lagoon characterized by their chief components (H : Halimeda sands, F : Foraminifera sands). 20 The mean size of the sand taken in every sample was estimated by sieving in order to classify the sediment into three size classes : very fine sand (STF - less than 0.25 mm), fine sand (SF - between 0.25 and 0.50 mm) and medium sand (SM - greater than 0.50 mm). Sediment smaller than 0.04 mm was not studied. Contoured data (Fig 10) show that lagoon sediments are mostly fine sand, found in a wide range of depths. However, a large stretch of very fine sand was found in the central northern part of the lagoon between 15 and 30 m, and on an irregular discontinuous strip located 20 to 30 m leeward the eastern and southern reef at depth between 9 and 30 m. Some small medium sand patches are scattered along the lagoon margin down to a 10-15 m depth and a large stretch occurs across the deep central basin of the lagoon. Most of the materials of the sand sample taken proved to be quite heterogeneous with a large size range and a symmetric distribution, except for foram medium sands that are homogeneous. Sands easily driven by currents are logically found in area of important water transport (i.e. lagoonward hoa, in the vicinity of the pass) and unexpectedly at a 34 m depth in the center of the lagoon which is an area thought to be calm. 14, wie Fig. 10 : Contoured bottom sediment data of Tikehau lagoon characterized by mean sand size (STF : very fine sand, SF ; fine sand; SM medium sand; see text for definition of size-classes). ———————o 21 BIOLOGICAL COMMUNITIES OF THE OUTER SLOPE This chapter, as well as the following ones, does not attempt to provide a comprehensive description of the fauna in Tikehau but descriptive information about marine habitat. Determination of many coral and sponge species is still underway and furthermore, fauna description is limited to the first 90 m in depth. Depths below are unpractical for sustained SCUBA diving operations and should be sampled remotly. The information presented hereafter has been drawn from publications of Faure and Laboute (1984) for coral species census and distribution, Peyrot-Clausade (1984) for cryptofauna distribution and unpublished data of Intes for zoobenthos of sediments. Fish fauna distribution will be studied in a subsequent chapter of this volume. The fore reef area In the spur and groove zone (0-4 m), corals are largely dominated by calcareous algae which become increasingly dominant as exposure to trade wind increases. Total coral coverage rate ranges from 5 to 25 %. Spur and groove have on them coral adapted to withstand this high energy environment. The top of the spur and the parts of the wall of the grooves are colonized by small branching Pocillopora (P. verrucosa, P. meandrina, P. damicornis) ; small massive Favia rotumana, F. stelligera, Montastrea curta, Pavona clavus ; encrusting forms of Montipora caliculata, Acropora robusta, Millepora platyphylla and Acropora abrotanoides. Algae are the major component of this substrate. The main species are green algae Halimeda opuntia, H. discoidea, Caulerpa pickerengii, C. seuratii, Neomeris van bosse, Microdyction and the red algae Dasya to some extent. Sessile cryptofauna is abundant, sheltering among coral branches or small grooves in the rock. The highest richness is reached on overhangs and in small caves as noticed on the south southeast coast (Sponges, Hydroids Solanderia, Bryozoans, Stylasterids, Dendrophiliids, Didemnid and Polyclinid Ascidians). An important motile cryptofauna occurs in this zone. Polychaetes and Crustaceans dominate and borers (mainly Sipunculids) are rare. On the fore reef platform (4-10 m), the amount of coral coverage increases (60-80%) as well as its diversity. The most conspicuous coral species are short bunches of Acropora humilis, ‘A. digitifera, A. variabilis, Astreopora myriophtalma, the first noticed Fungid (Fungia fungites and F. scutaria), and all the coral species cited above. On the northwest shelves where the platform is wide, the fore reef platform is also highly colonized by algae of the genus Microdyction, Halimeda and Caulerpa. Cryptic community biomass decreases slightly, still dominated by Polychaetes and Crustaceans. The outer terrace Between 10 and 15 m depth, grooves floored with rubbles alternate with coral ridges. Communities are made up of Favia stelligera, Pocillipora eydouxi, Astreopora myriophtalma, Acropora abrotanoides, Platygyra daedalea, Porites lobata, Favia rotumana, Millepora platyphylla. Coral coverage rate is about 60 %. From 15 m to 25 m, Porites lobata progressively dominate followed in, order of abundance, by Pocillipora eydouxi, Favia stelligera, F. rotumana, Astreopora myriophtalma, Acanthastrea echinata, Pavona varians, Herpolitha limax and Acropora sp.. Calcareous algae Porolithon and Peyssonnelia along with soft algae colonize the outer terrace bottom to depth of 25 m. On the windward coast, algae coverage can be important to some extent. 22 Sessile fauna, except for some sponges as Astroclera sp., shelter among coral patches and is rather scarce. Echinoderms, Molluscs and Crustaceans are abundant but remain hidden during daytime. The only occurrence of the invertebrate coral predator, the crown-of-thorn starfish Acanthaster planci, was recorded in this zone. Its population level seems to be very low and no evidence of an extensive coral predation was found. Motile cryptofauna is dominated by Polychaetes and Crustaceans. Its biomass is low and is reduced by half between 10 and 20 m in depth. The deep slope The living coral coverage rate of the deep slope is high ranging from 50 to 100 %. At these depths, the amount of light energy reaching the bottom decreases steadily, inducing a colonization of the bottom by coral species fitted to dim recess conditions. The plate-like Pachyseris speciosa is thus the major component of the coral fauna, so much that this zone is termed the "Pachyseris speciosa area". In the upper zone (25-35 m), most of the dominant species of the previous zone (Porites lobata, Pocillipora eydouxi, Favia stelligera and Acropora sp.) are progressively replaced by species more shade-tolerant as Gardinoseris planulata, Lobophyllia sp., Coscinarea sp. and small colonies of Pachyseris speciosa are located in the shade of large vasiform coral species. The live coral coverage rate is above 50 %. In the middle zone (35 to 60-70 m), Pachyseris speciosa sharply predominates colonies of Porites lobata, Pavona varians, Leptoseris incrustans, Gardineroseris planulata and Echinophyllia aspera. In the lower zone (from 60-70 m downwards), numerous colonies of Leptoseris (Leptoseris hawaiensis, L. scabra, Leptoseris sp.) and Echinophyllia (E. aspera, E. echinata) appear among well-developed colonies of Pachyseris speciosa with a nearly flat upper surface well adapted for best capturing sunlight. Stylaster and especially the red coral Stylaster sanguineus are common species below 60 m but they are already found from 40 m downward in places well protected from light (i.e. : on overhangs, in small caves or in the shade of larger species). No SCUBA diving investigations were made below 85 m but given the transparency of the water, it has been observed that the Pachyseris-Leptoseris population is still developing with the same energy below a depth of 90 m. Algae zonation is not restricted to the upper strata of the slope. When a dead coral leaves a living site, the vacated space can be reoccupied by Caulerpa urvilliana (to depths of 65 m), Caulerpa seuratii (15-70 m), Caulerpa bikinensis (to depths of 75 m and probably below), Microdyction sp., (1-65 m) or an unidentified species of Halimeda. The black sponge Astroclera sp. remains the most conspicuous species among the 25 observed sponge species. It reaches its maximum density at a depth of 40 m but is still abundant in small colonies at 70 m. The paucity of other sessile invertebrates like Gorgonians, Anthipatharians or large Hydroids is noteworthy. Unlike many Indo-Pacific coral reefs, this is a common feature of outer slope of atoll in the Tuamotu archipelago. BIOLOGICAL COMMUNITIES OF THE REEF FLAT In all aspects, the reef flat is quite variable. Various authors have described this zone at Tikehau, usually in combination with a description of cross-reef flat transects on several parts of the atoll rim. The algal ridge The algal ridge is mostly built by calcareous algae of genus Porolithon and Chevaliericrusta along with Pocockiella variegata and algal turf made of Caulerpa, Halimeda, Microdyction and Liagora. 23 Coral coverage rate is low, less than 2 % on windward reef and less than 10 % on leeward shelf. Coral colonies generally flourish on the wall of the grooves which are emergent only when low spring tide occurs. Coral species are short, small and blunt ecomorphs fit to withstand the high- energy environment generated by wave action. The main components of this community are : Pocillipora damicornis, Porites lobata, Pocillipora verrucosa, Montipora caliculata, Acropora humilis, A. digitifera and Millepora platyphylla. Three echinoderms, Heterocentus mammillatus, Colobocentrotus pedifer, Actinopyga mauritania and lesser number of juvenile sea-urchins sheltered in holes with Echinometra sp., are often abundant. Among molluscs, the most seaward species is Platella flexuosa followed lagoonward by Drupa ricinus, D. morum, Turbo setosus, Morula uva and two species of Vermetid. The algal ridge structure provides numerous small cavities which afford protection from the wave surge (and predation) to motile cryptofauna. Its biomass is high and mostly consists of Polychaetes and Crustaceans. Sessile cryptofauna has a high species richness index but since its living space is restricted by the availability of cavities always submerged, the total biomass is low. The outer reef flat The coral community of the outer reef flat is extremely poor with only two species censused, Pocillipora damicornis and Porites lobata, covering less than 1 % of the area. Algae are scarce. Among the most conspicuous molluscs, Drupa grossularia, Conus sponsalis, C. ebraeus, Erosaria moneta, Cerithium alveolus, young Tridacna maxima and Chama imbricata are frequently encountered. The conglomerate has evidence of extensive rock boring by numerous Lithophagids (up to 50 individuals m-2). Various sessile invertebrates shelter under blocks and in cavities but their abundance is quite low. Motile cryptofauna is scarce and of low biomass, dominated by Polychaetes and Molluscs. The inner reef flat Algae are the main components of the innermost part of the reef flat. The primarily fine sandy bottom is colonized by Halimeda opuntia, Caulerpa serrulata and C. urvillana. Many sand mounds of the mud shrimp Callichirus armatus are scattered all over this area, attesting an important burrowing activity. Lagoonward the inner reef flat and down to a 2 m depth, the coral community consists of Pocillipora damicornis, Acropora digitifera, A. abrotanoides, A. corymbosa, A. humilis, Favia stelligera, Montastrea curta, Platygyra daedalea. Live coral fauna covers about 25 % of the substrate. Algae coverage is high over blocks and sites left by dead coral colonies. Species identified are Dictyotales, Halimeda sp., Mycrodyction sp., Pocockiella sp. and a thick algal turf. A great abundance of Cerithium alveolus and Erosaria moneta is remarkable. Shapes and extensions of dead coral colonies provide a lot of space for a motile cryptofauna settlement. As a consequence, its biomass is high, dominated by motile and boring Molluscs and Sipunculids which account for more than 30 % of the total cryptofauna biomass. Main species of Ascidians are from the families of Didemnidae and Polycitoridae. 24 BIOLOGICAL COMMUNITIES OF THE LAGOON The lagoon slope The important sedimental layer of the inner slope hampers somewhat the development of coral communities. Live coral coverage rate is less than 10 % and consists of Pocillipora sp. mostly settled in the shallowest part of the slope, followed downward by massive and encrusting forms of Porites lobata, Leptastrea purpurea, Pavona varians, Platygyra daedalea, Montipora verilli and Fungia sp.. Algae are generally scarce. Microdyction have colonized hard substrate on the windward side of the atoll. On the west coast, not far from the pass, an algal flat community dominated by Caulerpa sp. extends on the slope with some Halimeda sp. found where the slope flattens out and merges with the lagoon bottom. Invertebrates occur sporadically in this barren area. Between 6 and 12 m depth, the lagoon margin has areas of abundant patch reefs. Particularly in places exposed to an abundant water circulation but protected from sediment overwash, lagoon margin patch reefs have well developed coral communities which can be either multispecific coral heads of Pseudocolumastrea pollicata, Platygyra daedalea, Leptastrea purpurea, Pavona varians, P. minuta, Stylocoeniella sp., Astreopora sp., Fungia ssp., Porites lutea, Stylophora pistillata, Montipora verrucosa, M. verrilli or paucispecific bunches of Acropora formosa and Acropora vaughani. A few sessile bivalves grow on the sides of the patch reefs. The most frequently encountered species are Arca ventricosa, Pinctada maculata and rarely the pearl oyster Pinctada margaritifera. In addition to some echinoids hidden in shelters provided by coral patches, beche-de-mer Halodeima atra, Thelenota ananas and several species of synaptid commonly feed in this area. The lagoon bottom The nature of lagoon bottom substrate has not been studied in detail but obviously, soft sediment substrata overwhelmingly dominate hard coral substratum. The lagoon floor can be characterized as large stretches of sand with occasional patch reefs. Wide areas of soft bottom are covered with a thin algal mat of brown Cyanophycae. Other Cyanophycae are visible on the sediment as large red balls. Species of Halimeda (principally H. opuntia) and Caulerpa (C. serrulata, C. urvilliana) can build up large algal flat areas. Caulerpa sp. grows via rhyzomes, spreading out over the bottom in easily distinguishable patterns. Sea grass beds of the marine phanerogame Halophila ovalis reach high density in some places. Little is known on the fauna buried in, or on, sediments. Epifauna is composed mostly of sponges (Echinodictyon, Axinella) scattered over sediment surface. From place to place, some bivalves of the genus Pinna raise from the sediment. Holothurian Halodeima atra concentrates in great numbers, up to 10 individuals per square meter in the shallowest sandy areas. Polychaetes and Molluscs alternatively dominate the endofauna. Sedentary Polychaetes are well represented mostly by the families of Spionidae (Prionospio sp., Aonides oxycephala), Maldanidae (Axiothella sp.), Capitellidae (Dasybranchus and Notomastus) and Terebellidae. A few errant species live in the sediments ; they belong to the families of Glyceridae, Eunicidae and Nephtyidae. Molluscs are essentially little bivalves like Tellinidae but some carnivorous gastropods exist (Naticidae, Strombidae). Among Echinoderms, only a few ophiuroids may be encountered. Very motile crustaceans as Portunid crabs may bury in sediments. Lancelets are locally abundant, especially in coarse sands. Small conical mounds disrupting the sediment surface are evidences of the presence of mud-shrimps. Their density is quite high, approximately 15 mounds per ten square meters. Mounds can be as large as 50 cm diameter on bottom of deep zones. 25 Stretches of hard bottom are found down to a 20 m depth. They are often restricted to areas of a few square meters but around the largest pinnacle reefs, they can cover areas of hundreds of square meters. Coral communities are made up of either multispecific patches of Porites lobata, Psammocora sp., Montipora sp. Astreopora sp., or paucispecific bunches of Acropora spp.. Macroalgae, particularly species of Halimeda, are often found among the corals. Pinnacle reefs The pinnacle reefs of Tikehau cover only a few percent of the lagoon bottom area but concentrate a great biological diversity, harbouring a wealth of various organisms. The upper zone (0-2 m) can be emergent at low tide. The center of the largest pinnacle reefs is rugged with sediment and rubble, colonized by algae Halimeda, Pocockiella, Caulerpa, Padina and Lobophora. On the windward side, very large colonies (6 to 8 meters) of Porites lutea and Millepora platyphylla grow along with a few Pocillipora meandrina and Acropora abrotanoides. The leeward side supports branched colonies dominated by Acropora variabilis, A. hyacinthus, A. hemprichi and Montipora spp.. The motile cryptofauna of the upper zone is rich and dominated by Molluscs. Borers are rare. The bottom of the middle zone (2-6 m) is largely a rocky substrate with shelves on which considerable quantities of sediment and rubble are retained, promoting the development of Halimeda and Caulerpa algal flat. Some coral heads of Montipora verrucosa, Astreopora sp., Psammocora sp., Porites lobata, Platygyra daedalea and Pavona varians intersperse among sediment stretches. Cryptofauna abundance decreases somewhat, being still dominated by Molluscs. Boring invertebrates, mostly Moliuscs and Sipunculids, represent almost 40 % of the total biomass. The important sediment deposition in this zone precludes the development of any significant sessile fauna. The lower zone (6-15 m) of the pinnacle reef is covered by coral patches of Montipora verrucosa, Stylocoenilla sp., Platygyra daedalea and branched forms of Acropora formosa, Stylophora pistillata and Favia favus. Below 15 m, coral colonies vanish under the sediment. . BIOLOGICAL COMMUNITIES OF MORPHOLOGICAL DISCONTINUITIES The pass Numerous live coral ridges oriented parallel to the pass axis are scattered over the bottom of the pass. These colonies are characterized by an outstanding development of Pocillipora (P. meandrina, P. verrucosa, P. eydouxi, P. damicornis) measuring up to 3 m. Live coral coverage rate is about 80 % and reaches 100 % alongshore. Small colonies of Leptastrea purpurea, Montipora sp. Fungia fungites, F. scutaria and Millepora platyphylla are found protected from current on the flagstone between ridges. Algae and invertebrates are rare in the pass owing to extremely rough current conditions. Some sponges (Aurora sp.), beche-de-mer (Thelenota ananas) or Asterids settled there nonetheless. Hoa The seaward bare bottom of hoa is colonized by a great number of Cerithium alveolus and often bored by numerous lithophagid. The top of many dead coral blocks is covered with the black Cyanophyceae Hassalia byssoides. Pink sands rich in Cyanophyceae and bacteria have accumulated alongside the edge of the channel. Coral patches of Porites lobata, Leptastrea purpurea, Porites cf andrewsi and Platygyra daedalea appear lagoonward, downcurrent of the reef where sediment overwash is weaker. Echinoid species (Echinometra, 26 Echinothrix, Diadema) are concentrated in some numbers around these coral heads. Some bivalves Tridacna grow on top of them. The motile cryptofauna dominated by Crustacean and Polychaetes is poor due to the paucity of suitable substrates. OCEANOGRAPHY OF THE TIKEHAU LAGOON The understanding of lagoonal water circulation is of primary interest for all biological and chemical studies carried out in the lagoon. Current data were recorded by four current meters set in the hoa of the windward reef, of the southwestern reef, of the northwestern reef, and in the pass. A tide gauge continually recorded water height in the lagoon. Modeling of the lagoon circulation was furthermore attempted. Pass currents The pass current is reversing. Its speed and timing are in phase with the tide. Data from Lenhardt (1991) show that the current speed usually ranges from 0 to 120 cm s-1, increasing from 0 to maximum speed in about 3 hours and then decreasing to slack water in another 3 hours. The period of slack water in the pass is only a few minutes. When water level is high in the lagoon, the pass has a nearly continuous outflow. The current speed may exceed 3 ms! under these frequent conditions and can last from 3 to 10 days. The volume of transport of the pass current varies between neap and spring tide and more importantly with water height figures in the lagoon. Data available did not enable us to estimate the outflow through the pass but the average inflowing water transport was estimated to be 400 m3 s-! over a tidal cycle. It was furthermore estimated that the net transport volume of the pass is approximatively zero over a tidal cycle under normal lagoon water level conditions. Hoa currents Hoa currents or cross-reef currents involve a shallow flow over the reef margins. Hoa draw their flow from the outer reef flat and channelize the water off the reef into the lagoon between islands. Hoa currents are a result of breaking waves over the algal ridge on the windward reef. They vary in response to surf height and therefore to regional oceanographic patterns. Lenhardt (1991) pointed out that Hoa currents do not reverse direction at Tikehau, they always flow from ocean to lagoon even when water level is high in the lagoon. Cross-reef currents do not flow in any well developed pattern as summarized in Table 4. The average current speed is low, generally less than 30 cm sl. It ranges from 0 to 120 cm s-! when high surf occurs. Current speed also varies seasonally, it is highest in hoa of the northwestern coast in summer of southern hemisphere while it is maximum in hoa of southeastern and southwestern reefs in winter. Table 4 shows that volume transport across the windward reef (southeast shore) accounts for more than 60 % of cross-reef water input whereas southwestern water transport accounts for 31 % and northwestern reef for 9 %. Average inflow per unit length of open coast (hereafter expressed lineic flow rate in m3 s-1/m or m2 s-!) is an useful data. Its variations over the year are summarized in Fig. 11. Bases for calculation are : length of South-East coast 23 km; length of South-West coast 11 km; length of North-West coast 11 km. 27, Table 4 : Estimates of monthly inflowing current speed and associated water transport in hoa and pass of the Tikehau atoll. Sse : current speed in southeastern channels, Ssw : current speed in southwestern channels, Snw : current speed in northwestern channels, Tse, Tsw and Tnw : relevant water volume transport. Rt : residence time. Bases for calculation : equivalent section of southeastern reef front : 5000 m2 ,southwestern reef : 2600 m2 , northwestern reef : 80 m2 . Total inflow is the sum of inflow through pass and hoa. Month Cross-reef current and transport Pass Total Rt inflow inflow Sse Tse Ssw Tsw Snw Tnw cm s-1 m3 s-1 cm s-1 m3 s-1 cm s-1 m3 s-1 m3 s-1 m3 s-l days January 55 270 25 65 18 15 260 600 February 3.0 140 3 75 21 18 180 400 230 March 4 100 17 14 150 April 75 200 28 18 250 May 15 800 105 280 9 i 100 1200 June 20 1000 6 160 13 11 70 1200 105 July 11 550 7.5 190 3 2 100 900 Fig 11 : Variations of lineic flow rate through hoa of Tikehau over the year. 28 Lagoon currents Suspended acoustic drifter releases at a 4 m depth, in various points of the lagoon, show that the surface current in the lagoon is primarily wind-driven. The general surface drift is downwind (or roughly westerly) at a speed of approximately 1 % of the wind speed. Lagoon currents in the vicinity of pass and hoa is strongly altered by local current conditions. Current patterns were found to vary with depth. Current in the deep layers of the lagoon is upwind and flows at about one half of the surface current speed. Water budget and residence times Because the cross-reef currents never reverse, the volume transport over the reef through hoa represent a net input of water into the lagoon. Water furthermore flows into the lagoon from the pass. The water can flow out of the lagoon only from the pass. As the quantity of water entering the lagoon and the quantity of water exiting out of the lagoon must be balanced, the net inflow must exit as outflow out of the pass. Thus, water transport in the pass is indicative of water budget figures in the lagoon. The volume transport Q (m3 s-1) flushing in the pass was found to be a simple function of the difference in m between height of lagoon (h) and ocean (Zz) according to: Q=e3000 \ |h-z| € being -1 if water flows out of the lagoon, +1 if water flows in. The average residence time of water in the lagoon can be estimated by dividing the lagoon volume (i.e. 10. 10? m3) by the net rate of water input presented in Table 4. Under these very simple assumptions, the calculation yields a residence time of 230 days in summer and 105 days in winter, the yearly average value being 170 days. Because the water entering the southeastern lagoon must transit the entire lagoon before exiting through the pass and because it probably undergoes mixing by the wind-driven circulation, during that transit, the residence time of this part of the inflow will be longer. Conversely, water inflowing through northwestern channels will be expected to have a residence time shorter than the average. The residence times estimated at Tikehau can be compared to those estimated in the lagoon of a high island (Moorea, 6.5 hours) and in the lagoon of a closed atoll (Takapoto, between 4 and 5 years). Circulation model Lagoon circulation can be explained as a response to three sources of energy : the tide, the surf on the ocean reef and the wind. Lenhardt (unpublished data) proposed a bi-dimensional finite difference model using vertically integred Navier-Stockes equations to model the response of lagoon circulation to these three sources of energy taken separately then altogether. Fig. 12a shows the residual tide-induced circulation in the lagoon. Tidal currents influence the flow of water only in the immediate area of the pass. Current speed is greater than 1 cm s-! only within 4 km of the pass. Elsewhere in the lagoon, tidal current speed is very low (a few mm s-1) and can hardly be measured. The tide generates a conspicuous clockwise local eddy south to the pass and a weak counter clockwise eddy north of the pass. Fig. 12b shows the circulation induced by the surf on the windward reef. The oceanic water spreads into the lagoon, moving downwind toward the pass. Currents are significant only in the southern part of the lagoon with speed of about 1 cm s-!. The northern part of the lagoon appears to be poorly affected by this circulation with modeled current speed less than 0.5 cms-1. 29 0.001 m/s tere erm n eee Fig 12a : Lagoon residual circulation generated Fig 12b : Lagoon circulation generated by by the tide at Tikehau surf on the windward reef of Tikehau (amplitude 10 cm ; arrows : current vectors). (lineic flowing current : 0,3 m3 s-l/m). Fig. 12c shows the circulation generated by a constant, unidirectionnal easterly windstress of 10 ms-1 (about 20 knots). Closed and impermeable boundaries were set to the atoll rim in order to avoid effects of water inflow for modelisation purposes. As measured by drogue releases, current speeds are strongly related to wind speeds (e.g. current speed is 1 to 2 % of the wind speed). In shallow part of the lagoon, the wind creates a downwind drift and an upwind drift in the deepest area. The wind-driven circulation induces two large conspicuous counter- rotating bodies of water, a counter clockwise northern eddy and another clockwise southern eddy. Fig.12d shows the connection the effects of the three sources of energy, the tide, the surf and the wind. Given the low current speed of the tide-induced circulation, the connection of the three sources is actually the sum of the wind and surf-driven circulation. Fig. 12d shows a general downwind drift toward the pass in the southern lagoon and a weak water circulation in the northern lagoon. The relevance of the model is somehow limited by the fact that it is bi-dimensionnal. The model gives only an average of the current speed and direction over the whole water layer and cannot take into account changes in current with depth. In all probability, and as confirmed by acoustic drogue releases and current speed records at various depths, the pattern of circulation in the lagoon must be a general downwind surface drift oriented toward the pass, balanced by an upwind low-depth drift. Moreover, the lagoon rim is closed neither to the pass nor windward. Large quantities of water are introduced along the windward edge and flushed out of the pass, and considerably modify wind-driven current patterns in the lagoon. Further circulation models should therefore be developed in three dimensions though information supplied by the bi-dimensional model meets biological research requirements. Fig 12c : Lagoon circulation generated by Fig. 12d : Lagoon circulation generated a constant 10 ms"! easterly wind at Tikehau. by the connection of the three sources of energy (tide : 10 cm, surf : 0,3 m3 s-1/m and wind :4ms-1) at Tikehau. AN EXAMPLE OF MAJOR DISTURBANCE : CYCLONES About four cyclones per century are likely to occur in the Tuamotu archipelago but during the hot season 1982-1983, no less than four cyclones (Orama, Reva, Veena and William) and a tropical storm (Lisa) ravaged French Polynesia, causing severe damages to human installations, vegetation and coral reef ecosystems. This meteorological phenomenon of an outstanding frequency and intensity induced waves measuring more than 10 meters, a sea- level rising of 3 meters above predicted tidal levels, and a wind blowing in excess of 160 km per hour with gusting to over 200 km per hour. In particular, the cyclone Veena occurred in the vicinity of Tikehau between the 6th and 13th of April 1983. SCUBA Dive surveys carried out on the same transects before and after that disaster allow to assess the magnitude of damage caused to coral communities, and to explain destruction mechanisms. Two years later, SCUBA Dive surveys carried out on the same transects showed that time of recovery is very long. It will take at least five decades to have the coral communities in the state observed prior 1983. Damage assessment Damaged area (SD) is assessed by multiplying the length of destroyed reef (L) by 200 m (this value is the average length / of the outer slope between the depth of 0 and 90 m where hermatypic corals live). L is an estimation based on field observations made during 22 SCUBA Dive surveys around the atoll rim. In Tikehau, an area of 13 million square meters is estimated to have been damaged. This represents 80 % of the outer slope (Laboute, 1985). The area located windward of the atoll (from the north around to the south-west through the east) has been destroyed to a magnitude exceeding 90 %. Areas located north-north-west and west have been damaged to a magnitude varying between 30 and 80 %. The western side of the atoll remained intact. 31 Destruction mechanisms Based on the outer slope morphology, three types of destruction mechanism were figured by Laboute (1985) and Harmelin-Vivien and Laboute (1986) following field observations. These theoretical processes are subject to variations with local details of the slope. The reef flat is narrow, the fore reef area above a depth of 15 m and the slope very steep (> 45°) as on the western coast of the atoll. Between 0 and 15 m, plate-like and branching madreporic species are abundant. They are of the genus of Pocillipora, Acropora, Montipora, Astreopora, Favia and Pavona. All those species are fragile and have been uprooted then reduced to rubble by the strength of the swell. The remains have contributed to destroy more resistant species by recurrent impacts and abrasion. Most of the remains have been thrown up on the reef flat. At 12-15 m on the fore reef area, massive and heavy madreporic species like Porites lobata and Montipora prevail. As above, those species have been uprooted by the swell but since they were growing near or on high-angle substrates, remains rolled down the steep slopes weeping deeper colonies often more fragile, such as the plate-like Pachyseris speciosa, the dominant species below 40 m. Destructive effects of this underwater avalanche increased with depth as shown in Fig. 13. It is likely that coral blocks accumulated at some level between 300 and 500 m where the slope begins to flatten out, contributing to the formation of a detrital cone surrounding the atoll (Harmelin-Vivien and Laboute, 1986). So far, this avalanche phenomenon has never been described elsewhere. All previous hurricane effect descriptions were done for islands without this typical steep outer slope and as a result, cyclone damages were thought to be limited to the upper level of the ecosystem. OUTER REEF LAGOON PSB boulder rampart sea level direct destruction by storm induced SS 2RP Tey PSS eee coral destruction Soooces 50% to 80% : destruction by avalanche MOVEMENT OF BROKEN CORALS SURVIVING COLONIES 100% coral SAND AND RUBBLE destruction Fig. 13 : Cross section of the west coast of Tikehau with explanatory hypothesis of deep coral destruction : direct coral destruction by storm-induced waves occured between the surface and 20 to 22 m depth. Most remains of broken coral rolled down the slope breaking fragile deeper colonies (after Harmelin-Vivien and Laboute, 1986), 32 The reef flat is large with a low slope gradient, the fore reef area deeper (25 m) and the outer slope less steep (< 45 °) as on the northern coast of the atoll. Coral colonies have been broken down to 15 m. Part of their remains were thrown up on the reef flat, the other part remaining on the same area. Many coral skeleton were covered by algae (genus Microdyction and calcareous algae). Massive species between 15 and 25 m have been uprooted too but given the distance to the steep slope, no avalanche occurred and therefore deeper colonies remained undamaged. The reef flat is equally large or narrow but there is a depression or a very low-angled zone just before the fore reef area. Madreporic species located above a depth of 15 m have been destroyed following the previous process, but part of their remains have been trapped in that depression, preventing an avalanche. Thus, Porites lobata and Pachyseris speciosa colonies of deeper zones have been preserved. Whatever the shape of the deep slope is a shallow fore reef coral community (composed of small colonies well adapted to withstand high energy level environment) suffered less than deeper reef communities. Coral destruction, estimated at 50 %, resulted primarily from abrasions by dislodged material, rolling remains and scouring sand. Destruction mechanism in the pass and in the vicinity is of a particular nature. The Tikehau pass, located at the west of the atoll, was colonized by numerous species of the genus Pocillopora (P. meandrina, P. verrucosa, P. eydouxi and P. damicornis). The live coral covering rate was high, reaching 80 % to 100 % alongside the edge of the pass, between 2 and 8 meters. Those colonies remained intact as observed during SCUBA Dive surveys carried out six months after the cyclone Veena. More than one year later, all species were dead and skeleton still in place. This can be explained by the following observations : after the cyclone, the sea-level in the lagoon was high and currents constantly flowed out of the lagoon. Water was loaded with a considerable amount of suspended particulate matter and was, as a result, very turbid. Sedimentation was very important in the vicinity of the pass and therefore the amount of light energy reaching the bottom decreased dramatically and subsequently, corals died. In all cases, most of the coral associated fauna (sessile fauna and fish) disappeared owing to a high mortality rate and to a lack of food and shelter. Recovery processes Coral resettlement became visible only one year after the cyclone in an area restricted to the upper 15 m as pointed out by Laboute (1985). In May 1984 Pocillipora (measuring between 2 and 4 cm) were dominating followed by encrusting forms of Favia stelligera, Acropora (size range : 2-7 cm), Millepora platyphylla and to some extent, small colonies of Pavona minuta and Favia rotumana. In the ravaged areas, no new madreporic colonies were seen below 15-20 m. In February-March of 1985, an important madreporic and algae resettlement could be seen on the West coast of the atoll between 3 and 15 m. Species inventoried were : Pocillipora - obviously the most numerous - (size range : 1-20 cm), Montipora caliculata, Montipora verrili, Astreopora myriophtalma (size range : 10-30 cm), Favia stelligera (size range : 2-12 cm), algae Halimeda taenicola, Mycrodyction and Caulerpa urvilliana. In June of 1985, algae and madreporic species were still resettling on the outer slope especially between 3 and 15 m. Species were : Pocillipora (size range : 4-23 cm), two or three species of Acropora (4-20 cm), Favia rotumana (4-15 cm), Porites lobata (2-11 cm), Pavona minuta (5-9 cm), Millepora platyphylla (15-50 cm). Algae Halimeda taenicola, Microdyction and Caulerpa urvilliana were as abundant as before the cyclone. In the area partly damaged and below a depth of 15 m, madreporic resettlement seemed to be faster owing to the presence of some sparse colonies which survived the cyclone. For instance : At 20m Acropora robusta was 60 cm large but still very encrusting except on its edge, Favia stelligera numerous in some place (size range : 10-40 cm), sparse Porites lobata (6-17 cm). 33 At 30 m, numerous Fungia fungites (size range : 8-10 cm), Millepora platyphylla as large as 1 m but still thin and encrusting. At 40 m, Astreopora myriophtalma measuring between 8 and 35 cm, sparse 17 cm Porites lobata very thin which had settled on previous base of same species. At 50 m, rare Fungia sp. of 3 cm and Pachyseris speciosa (size range : 4-7 cm). The impact of cyclones on fish fauna has been studied in detail and the results are presented in a later number of this issue. Literature cited CHEVALIER (J.P.) - 1973 - Geomorphology and geology of coral reefs in French Polynesia. In : Geology and Biology of Coral Reefs, I. Academic Press, New York : 113-141 CLAGUE (D.A.) - 1981 - Linear island and seamount chains, aseismic ridges and intraplate volcanism : results from DSDP. Soc. Econ. Paleontolog. Mineralog., Spec. Publ., 32 : 7-22 FAURE (G.), LABOUTE (P.) - 1984 - Formations récifales : I - Définition des unités récifales et distribution des principaux peuplements de Scleractiniaires. In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 108-136 FLORENCE JJ.) - 1985 - Introduction 4 la flore et a la végétation. In : Contribution a I’étude de J’atoll de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 24 : 98-113 GABRIE (C.), SALVAT (B.) - Généralités sur les iles de la Polynésie Frangaise et leur récifs coralliens. Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 1 : 1-16 HARMELIN-VIVIEN (M.L.) - 1985 - Présentation générale de l'atoll. In : Contribution a l’étude de I’atoll de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 24 : 2-27 HARMELIN-VIVIEN (M.L.), LABOUTE (P.) - 1986 - Catastrophic impact of hurricanes on atoll outer reef slopes in the Tuamotu (French Polynesia). Coral Reefs, 5 : 55-62 INTES (A.) - 1984 - Présentation générale de l'atoll. In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22: 4-12 INTES (A.), ARNAUDIN (H.) - 1987 - Esquisse sédimentologique du lagon. In : Contribution a l'étude de l'atoll de Tikehau : IV. ORSTOM Tahiti, Notes et Doc. Océanogr., 35 : 71-100 JAMET (R.) - 1985 - Les sols de I'atoll. in : Contribution a l’étude de I’atoll de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 24 : 74-97 LABOUTE (P.) - 1985 - Evaluation des dégats causés par les passages des cyclones de 1982-83 en Polynésie Francaise sur les pentes externes des atolls de Tikehau et de Takapoto (Archipel des Tuamotu, Polynésie Francaise). Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 3 : 323-329 LAMBECK (K.) - 1981 - Flexure of the ocean lithosphere from island uplift, bathymetry and geoid height observations : the Society islands. Geophys. J. R. Astr. Soc., 67 : 91-114 LENHARDT (X.) - 1987 - Etude bathymétrique du lagon de Tikehau. In : Contribution 4 l'étude de l'atoll de Tikehau : IV. ORSTOM Tahiti, Notes et Doc. Océanogr., 35 : 53-70 LENHARDT (X.) - 1991 - Hydrodynamique des lagons d’atoll et d’ile haute en Polynésie Frangaise. ORSTOM Paris ed., Etudes et Theses : 132 p 34 MONTAGGIONI (L.F.) - 1985 - Makatea island, Tuamotu archipelago. Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 1 : 103-158 PEYROT-CLASADE (M.) - 1984 - Cryptofaune mobile des formations récifales. In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 137-144 PIRAZZOLI (P.A.), MONTAGGIONI (L.F.) - 1985 - Lithospheric deformation in French Polynesia (Pacific Ocean) as deduced from quaternery shorelines. Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 3 : 195-200 POULSEN (M.K.), INTES (A.), MONNET (C.) - 1985 - Observations sur l'avifaune en octobre 1984. In: Contribution a l’étude de I’atoll de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 24: 114-124 ROUGERIE (F.) - 1981 - Evaporation and salinity survey in French Polynesia. Tropical Ocean- Atmosphere Newsletter, 7 : 8 ROUGERIE (F.), CHABANNE JJ.) - 1983 - Relationship between Tuna and Salinity in Tahitian coastal waters. Tropical Ocean-Atmosphere Newsletter, 17 : 12-13 ROUGERIE (F.), WAUTHY (B.) - 1986 - Le concept d'endo-upwelling dans le fonctionnement des atolls-oasis. Oceanol. Acta, 9 : 133-148 SCHLANGER (S.O.) - 1981 - Shallow-water limestones in oceanic basins as tectonic and paleoceanographic indicators. Soc. Econ. Paleontol. Mineralog., Spec. Publ., 32 : 209-226 SODTER (F.) - 1985 - Eléments d’une histoire démographique. In : Contribution a l’étude de I’atoll de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 24 : 125-134 Plate 1 : Uplifted reefs, locally known as "FEO", are distributed along the northern coast of the atoll. Some of them, basaly notched, may be seen on the outer reef flat. (Photo Intes) Plate 2 : The marshy depression is covered with the cyperaceous assemblage dominated by Cladium jamaicense. (Photo Intes) Plate 4 : The spur and groove zone (eastern coast, 4 m). (Photo Laboute) Plate 5 : The ocean water flows into the lagoon through shallow channels called "Hoa". (Photo Intes) Plate 6 : A large Acropora colony overturned by the wave action induced by the hurricane "Veena" in April 1983 (eastern outer slope, 8 m). (Photo Laboute) Plate 7 : The avalanche effect may destroy 100% of the corals of the outer slope (eastern outer slope, 22 m). (Photo Laboute) Plate 8 : Recolonisation of the fore reef platform two years after hurricane "Veena" (eastern coast, 8 m ). (Photo Laboute) PART II. NUTRIENTS, PARTICULATE ORGANIC MATTER, AND PLANKTONIC AND BENTHIC PRODUCTION OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY C.J. CHARPY-ROUBAUD AND L. CHARPY NUTRIENTS IN OCEANIC AND LAGOONAL WATERS Matter and energy budgets for coral reefs, their components, and the world around them can, do, and must balance in a theorical context (Smith and Kinsey, 1988). In this paper, we will try to establish nitrogen, phosphorus and silica budgets between ocean and lagoon waters of Tikehau atoll in the purpose to learn more about the functioning of coral reef lagoons. Nutrient concentrations (dissolved components and particulate organic matter) were measured in the lagoon and in the surrounding oceanic surface waters between 1983 and 1987. DISSOLVED COMPONENTS Ocean waters Five oceanic stations shown on Fig. 1 have been sampled to a 500 m depth. At the most remote ocean station, OS7 (considered as representative of oceanic conditions), the upper 50 m surface layer displayed the characteristic temperature (29.5 °C) and the salinity (35.5 %o) of oceanic waters. Below this superficial layer (120 to 150 m) a temperature of 25 °C and a maximum salinity of 36.2 %o were recorded. Nutrient profiles were typical of offshore oceanic waters, with a very low nutrient . concentration down to 200 m depth. As summarized in Fig. 1, nutrient concentrations increased to 15 mmol NO3 m-3 and 2.5 mmol PO4 m-3 at 500 m. The nitrite concentration displayed a maximum at 175 m (0.1 mmol m-3). Nutrient concentrations in the upper 200 m proved to be higher at the stations located in the immediate vicinity of the atoll. The atoll therefore seems to disturb the standard vertical profiles of nutrient observed at station OS7, which in turn, results in an enrichment of nitrogen and phosphorus in the euphotic layer. The enrichment is probably due to a turbulent vertical mixing caused by the atoll mass effect, i.e. by internal waves (so, Andrews and Gentien , 1982, consider upwelling to be a source of nutrients for the Great Reef ecosystem ) and also, perhaps, by an up surge of deep waters from the atoll coral base as hypothesized by Rougerie and Wauthy (1986). ORSTOM-Tahiti, BP 529, Papeete, French Polynesia mmol m7? O01 nt 10 100 0.01 0.1 1 10 100 TIKEHAU 100 0.01 01 1 10 100 1 (= WO iS (= jou as 5-10 10-15 15-20 20-25 25-30 30-35 35-40 Depth (m) Fig. 9 : Benthic, planktonic and total primary productions in the Tikehau lagoon by a 5 m depth interval (from Charpy-Roubaud, 1988). RELATIONSHIP BETWEEN LIGHT AND PRIMARY PRODUCTION IN THE LAGOON. Light energy Percentages of light energy measured at surface decreases with depth d (m) according to the equation : % of incident energy = e(4.45 - 0.06d) It was measured that 17% of surface light energy reached 25 m ( average depth of the lagoon). Phytobenthos and phytoplankton production According to the foregoing sections, phytoplankton production (PP) can be related to incidental light energy (Eh) with a standard error of 0.6 mg C m-3 hr! by the equation : PP = 1.29 Eh0.39 (1) 18 Phytoplankton production per unit volume can then be converted into production per unit area according to the equation : d PPd = | PPz dz (2) 0 where PPd (mg C m-2 h-1) = production at depth d per unit area, PPz (mg C m-3 h-!) = production at depth z per unit volume. Using equation (1) and (2), phytoplankton production is related to light by the equation : d PPd = | @ 0.254 [e(4.45-0.062)E},, /100 ]0.39 0 where Ehs (E m2 h-1) is light energy at the surface. Similary, the relationship between phytobenthos production at depth d (BPd in mg C m2 hr1) and light energy Ehs at surface level can be written : BPd = 28.78 (e(4.45-0.06d)\ F(Ehs,100)) 9-454 (3) Predicted plankton and benthic production were calculated at depths between 0 and 36m (ie. : maximum depth of the lagoon) for different Ehs values observed in natural conditions (1-8Em-72h-}). Results presented in Fig. 10 indicate that phytobenthos production exceeds phytoplankton production in the upper 18 m. The total primary production (PT) is relatively constant with depth and depend primarily on light energy reaching the surface. Therefore, an average PT can be obtained for each Ehs value and subsequently, a linear relation linking PT and Ehs, i.e. : PT = 6.5 Ehs + 31.5 Therefore, daily total production PTd can be calculated assuming a sun time of 10 hours a day by the equation : PTd (mg C m2 d-1) = 6.5 Eds (E m2 d-1) +315 (4) The PTd value for Eds = 0 is 315 mg C m-2 d-1; may be interpreted as the respiration in the water column and sediments. Daily light energy data obtained in 1986 can be converted into daily primary production using equation (4). Monthly averages of PTd are plotted in Fig. 11. Over the entire year, the surface lagoon received a total of 15,550 E m-2 of which water column and sediments produced a total of 216,709 mg C m-2, equivalent to a daily average of 0.59 mg C m-2 d-1, not very different of the value of 0.69 mg C m-2 d-! estimated through field experiments. 19 mg C m-2 hv! 0 18 36 0 18 36 depth (m) Fig. 10 : Predicted phytoplankton production (dashed line), phytobenthic production (dotted line) and total production for different surface light energy levels (Ehs : E m2 hv!) in Tikehau lagoon from Charpy and Charpy-Roubaud (1990a). The photosynthetic efficiency of lagoonal communities can be summarized as follows : - One mg of Chi a allows an hourly growth production of 4.2 mg carbon - One Einstein reaching lagoon surface allows the growth production of 14 mg of organic carbon in _ the water column and sediments. mg C m2 g-1 month Fig. 11 : Monthly averages and confidence intervals (p=95%) of predicted total primary production (plankton and benthos) in 1986 in Tikehau lagoon (from Charpy and Charpy-Roubaud, 1990a). 20 mg dry wt m7? mean: 33.5 mg m-3 1») A 1965 1966 [J bioom [ Xp Fig. 12 : Variations in mesozooplankton dry weight from April 1985 to April 1986 in the Tikehau lagoon. Blooms of Thalia democratica (Salps), appendicularians (Appendic.), the pteropod Creseis chierchiae and the copepod Undinula vulgaris are shown. D: large amount of detritus (from Le Borgne et al., 1989). Table 11 : Relative contributions (%) of the main taxa to total zooplankton biomass in Tikehau lagoon. Percent contributions of non-living organisms (detritus) to total dry weight is also shown (from Blanchot and Moll, 1986 and Blanchot et al., 1989). a: Three samples of mesozooplankton were measured in April 1986 (from Le Borgne et al., 1989). Date Size class Taxon % total biomass % detritus April 1985 = Microzooplankton nauplii 39.2 35-200 pm copepods 30.6 70 bivalve larvae 19.8 polychaete larvae 10.4 Mesozooplankton copepods 68.4 500-2000 um chaetognaths 19.3 14 salpids 8.8 Macrozooplankton salpids 63.2 >2000 zm copepods 22.8 chaetognaths 1.8 April 1986 | Mesozooplankton copepods 73.8 200-2000 um larvaceans 5.0 3,10,494 brachyuran larvae 8.7 chaetognaths 5.6 21 ZOOPLANKTON BIOMASS AND METABOLISM The fate of the abundant particulate organic matter in the Tikehau lagoon and the level of efficiency of utilization by consumers are two aspects which are of primary importance to an understanding of the functioning of the lagoon pelagic food-web. Once sunk to the bottom, is seston consumed mostly by pelagic animals in the water column or by benthic ones ? Do they give rise to a significant predator biomass or is the efficiency of energy transfer between food and consumers low ? In order to provide answers to these questions, zooplankton have been studied at Tikehau during two 10 d periods in April 1985 and April 1986, and additionally by weekly samples taken in between. ZOOPLANKTON BIOMASS Fraction 3 to 35 zm (nanozooplankton) An average number of 71 + 14 heteroflagellates ml-! and 7 ciliates ml-! were found in samples taken in April 1986 in the lagoon. The mean individual volumes were of 542 + 96 tm? for heteroflagellates and of 14 246 + 7 427 um for cilates. In order to assess the relative biomass of these taxa, mean individual volumes were converted into carbon using a conversion factor chosen equal to 0.08 pg C m3 according to Sherr et al. (1984). The average biomass of heteroflagellates is approximately 3.1 + 0.7 mg C m-3 while ciliate biomass is 7.6 + mg C m-3. These values are of the same order of magnitude as those found by Hirota and Szyper (1976) in Hawaii though the method used does not make any difference between dead and live plankters. Fraction 35 - 200 j1m (microzooplankton) Microzooplankton consists of 43% of organisms smaller than 100 um of which 73% are protozoans (tintinnids:Rhabdonella sp., Codonellopsis sp. and Epiplocyclis sp. accounting for 64% ; naked ciliates for 5%). Foraminifers and radiolarians are poorly represented (less than 1%). Metazoans account for only 27%. The second most abundant taxon smaller than 100 um consists of naupliar copepods (18%) and the third of meroplankton bivalve larvae (8%). Out of organisms larger than 100 um, protozoans account for 33% of total numbers of which 23% are tintinnids. Metazoans are noticeably dominant, accounting for 67% of the total. Copepod nauplii are the most abundant organisms (41%). Bivalve ' larvae do not exceed 7% of total numbers. By using a C/ATP ratio of 125, live microzooplankton biomass can be estimated at 1.52 mgC m-3- Together with nanozooplankton, biomass of heteroprophs smaller than 200 jum reaches a value of 3.3 mg C m:3- Fraction 200 - 2000 jzm (mesozooplankton) Variations in mesozooplankton dry weight were monitored as regularly as possible in the lagoon and displayed a measurement of two maxima in October 1985 (65 and 93 mg dry wt m3). Zooplankton at Tikehau are characterized by periodic blooms of copepods, larvae, pteropods and salpe. As shown in Fig. 12, the annual mean value is 33.5 mg m-3, which is six times greater than the oceanic plankton concentrations measured in the vicinity of the pass (i.e. : 5.4 mg m-3). Relative contributions of various size classes and taxa to zooplancton biomass are displayed in table 11. 22 Table 12 : Live biomass and detritus in terms of C, N, P estimated from the percentage of detritus in seston and from particulate organic carbon values in the Tikehau lagoon. - a: ATP x 125 b: POC - (ATP x 125) Sizeclass year Live Dead (um) Cc N P Cc N P 35-200 1985 0.36 0.07 0.03 0.84 0.16 0.06 1986 1.02 0.19 0.08 2.37 0.45 0.18 1986 1.5a 2.30b 200-2000 1985 5.22 0.98 0.39 0.85 0.16 0.06 1986 2.80 0.53 0.21 0.74 0.14 0.06 1986 2.5a 4.4b > 2000 1985 3.23 0.61 0.24 0 0 0 1986 0 0 Table 13 : Metabolic atomic ratios and contribution of inorganic excretion to total excretion (%) in the Tikehau lagoon. O:O2 respired; NH4, Nt, PO4 Pt : ammonia, total nitrogen, phosphate and total phosphorus excreted. nd : no data (from Le Borgne et al., 1989). Date Size class/species O:NHqg O:Nt O:POq O:Pt NH4;POq Nt:Pt NHg4:Nt POg:Pt April 1985 Microzooplankton thee 6.9 144 104 12.1 11.1 85.4 74.4 Mesozooplankton 10.8 10.8 119 92 11.4 8.6 100.0 76.5 Undinula vulgaris 18.4 13.2 132 100 7.9 79 48.2 43.1 Thalia democratica 15.5 8.1 103 56 74 6.9 58.3 54.3 April 1986 Microzooplankton nd ed, 124 96 nd 12.5 nd 77.5 Mesozooplankton nd 7.1 144 110 nd 8.4 nd 76.3 Table 14 : Zooplankton. Net growth efficiencies (K2) in terms of nitrogen and phosphorus for total populations and sorted species or taxa, calculated from N: P ratios of particles (a1), zooplankton excretion (a2), and body constituents (a3). Number of replicates in (brackets) (from Le Borgne et al., 1989). Date Size class/species a a a K K 1 2 3 2,N 2,P 1985 Mesozooplankton (9-10 April) 13.9 (25) 8.6 (4) 15.0 (2) 0.894 0.828 Undinula vulgaris 13.9 (25) 7.9 (2) 25.7 (2) 0.623 0.337 Thalia democratica 13.9 (25) 6.9 (12) 20.4 (2) 0.761 0.519 1986 § Mesozooplankton (7-9 Apr.) 18.2 (4) 10.8 (2) 26.0 (1) 0.695 0.487 (10-12 Apr.) 16.3 (4) 10.8 (2) 26.0 (1) 0.695 0.487 (13-16 Apr.) 13.9 (4) 9.2 (2) 20.5 (1) 0613 0.416 mean 16.1 8.4 23.3 0.748 0.517 Mixed copepods 16.1 (12) 8.4 37.6 (1) 0.616 0.264 Microzooplankton 16.1 (12) 12.5 (6) 16.7 (5) 0.889 0.857 23 Zooplankton biomass in terms of C, N, P By removing detritus of samples and by calculating their relative contribution to seston dry weight, live zooplankton biomass can be estimated. Detritus dry weight represents 70% of 5-200 um particles, between 1% and 21% of 200-2000 jm particles, and 0% of particles larger than 2000 pm. Contributions of carbon, nitrogen and phosphorus to dry weight of detritus and zooplankton taken separately were estimated by Le Borgne et al. (1989) enabling them to calculate biomass as shown in Table 12. METABOLISM Zooplankton respiration and excretion were measured in three organism size-classes and in species, Undinula vulgaris and Thalia democratica, which are abundant in the lagoon. Results of metabolic atomic ratios and contribution of inorganic excretion to total excretion are summarized in Table 13. Growth efficiency for the total population and sorted taxa are displayed in Table 14 and assimilation efficiencies of copepods are listed in Table 15. As a general pattern, efficiencies estimated for microzooplankton are greater than for mesozooplankton. Excreted nitrogen and phosphorus are mostly inorganic. They meet 32% and 18% of phytoplankton nitrogen and phosphorus requirements. Production was then estimated. All P/B ratios presented in Table 16 are greater than 34% which is equivalent to a three day turnover time of the biomass. The turnover rate is shorter for mixed zooplankton, close to one day and even shorter for the salp Thalia democratica due to, in all probability, asexual reproduction, high water temperature (29.5 °C) and abundances of food as emphasized by Le Borgne and Moll (1986). On the average, P/B ratios for zooplankton are 5.7 times lower than P/B ratios estimated for phytoplankton. The nitrogen and phosphorus assimilation rates in Table 17 are the sum of production and total excretion rates. Ingestion is calculated from assimilation and its relevant effficiency. Table 15 : Assimilation efficiencies (D) of C, N, P and their ratios (a4=DN:Dp a'4=Dc:DN). Calculations made by the method of Conover (1966) using organic carbon, nitrogen and phosphorus percentages of dry weight in feces and food. nd : no data (from Le Borgne et al., 1989). Date Species Faeces Food D (%) a a’ C N P C N P CoN P 1985 Undinula vulgaris 0.40 0.10 0.048 2.36 0.25 0.081 83.6 60.1 38.3 1.57 1.39 1986 Small copepods 0.59 0.17 0.025 15.4 1.86 0.258 96.8 91.0 90.2 1.01 1.06 0.62 0.09 325/79 0:85m OSIS9 890695 ae ndiaend 0:99 24 Table 16 : Production rates in terms of C, N, P (ug mg7! dry wt d-1) and daily P:B ratios at Station 6. Body C, N and P, as percentages of dry weight, allows the conversion of rates into P:B (from Le Borgne et al., 1989). Date Sizeclass /species Production rates Body constituents P:B C N P C N etre? (%) April 1985 Mesozooplankton 164 44.65 3.565 20.7. 4.12 0.37 102.0 Undinula vulgaris 108 29.97 2.604 314 873 0.76 343 Thalia democratica 190 4288 5.859 1.91 0.43 0.0725 816.0 April 1986 Mesozooplankton 331 73.64 7.061 384 850 0.81 864 Mixed copepods __ 147 39.80 2.346 = 27.17.31 ~—0.43 54.4 Table 17 : Zooplankton. Assimilation and ingestion rates (ug C, N or P mg“! dry wt d-) in the Tikehau lagoon (from Le Borgne et al., 1989). Date Size class/ species Assimilation Ingestion Cc N P Cc N P April 1985 Mesozooplankton 656 TACT 10.57 785 102 15.1 Undinula vulgaris 515 48.3 Ups) 616 80 20.2 Thalia democratica 56.4 11.28 617 81 16.1 April 1986 Mesozooplankton 724 98.4 13.54 804 109 15.0 mixed copepods 473 64.6 4.67 526 7? 2 TROPHIC STRUCTURE AND PRODUCTIVITY OF THE ECOSYSTEM With all data estimated in the foregoing sections of this chapter, biomass and fluxes of matter in plankton and in benthos can be assessed. In order to enable comparisons between benthic and planktonic ecosystems, standing stocks are in mg C m2 and fluxes in mg C m2 d-!. The trophic web is illustrated in Fig. 13. We did not measure directly the biomass of bacteria, but we think that we can obtain an order of magnitude by the difference between liv C estimated from ATP and the other biomasses measured or estimated. Therefore, free bacteria biomass (BB) may be calculated with the equation: BB = liv C(<5 ym) - phy C(< 5 pm) with: liv C(< 5 um) = percentage of ATP(< 5 ym) x ATP x 250 phy Cc< 5 ym) = percentage of chl ac< 5 ym) x chl a x (C/chl-a ratio). The average value C/chl a = 50 (Charpy and Charpy-Roubaud, 1990 b) lies within the range reported by Takahashi et al. (1985) for picoplankton and is very close to the ratio of 46 found by Laws et al. (1987) for oligotrophic Pacific waters. Therefore : BB = (0.46 x O.11 x 250) - (0.5 x 0.18 x 50) = 8.2 mg C m3 25 The liv CS - 35 ym) is made up of heterotrophs and phytoplankton (4.5 mg C m-3). It can be calculated using ATP(5 - 35 ym) data : 0.54 x 0.11 x 250 = 14.9 mg C m°3. The carbon content of heterotrophs in the size range from 5 to 35 lm was therefore equal to : 14.9 - 4.5 = 10.4 mg C m3. The biomass of ciliates and heteroflagellates was equal to 1.8 mg C m-3 (Table 3), and the difference (= 8.6 mg C m°3) was certainly due to bacteria adsorbed onto the detritus (Charpy, 1985). The total biomass of bacteria was therefore equal to : free bacteria + adsorbed bacteria = 16.8 mg C m-3. Such a biomass is commonly observed in waters over reefs ; Sorokin (1974) summarizes data for biomasses of bacteria which range from 11 to 170 mg C m°3. More recently, Moriarty et al. (1985) reviewed the productivity and trophic role of bacteria on coral reefs. They give biomass values ranging from 19 to 150 mg C m3. Linley and Koop (1986) observed in the coral reef lagoon of One Tree Island (Great Barrier Reef) a biomass of heterotrophic bacteria ranging from 1.2 to 16.2 mg C m-3, and Hopkinson et al. (1987) observed a bacterial biomass of 2 mg C m°3 in the water column of Davies Reef (Australia). In Tikehau, in April 1986, the biomass of bacteria was estimated at 17.1 mg C m-3 by Blanchot et al. (1989). The observed ratio free bacteria / adsorbed bacteria = 2 is consistent with the ratios given by Moriarty (1979) and Moriarty et al. (1985) in coral reef areas. The estimated bacterial biomass was 2 times higher than the phytoplanktonic C in the Tikehau lagoon. Dominance of bacterial biomass was also observed in the oligotrophic waters of the Sargasso Sea by Fuhrman ef al. (1989); the interpretation of these authors was that bacteria consume sigificant amounts of carbon probably released from phytoplankton directly or via herbivores. DIC A = 440 ZOOPLANKTON |_P >35 ym PHYTOPLANKTON 101 36-200 pm : 38 > 200 pm : 63 P=? E=3 225 <5 pm : 113 >5 pm : 112 DETRITUS 4268 5 pm : 2050 5-35 pm : 2050 35-200 pm: 58 > 200 pm: 110 | 420 <35 pm 46 ates : 33 cll fi BOO OOOO OO Dae OOOO OOOO Fig. 13 : Trophic structure and productivity of the Tikehau lagoon communities. Standing stocks (mg C m2) are in boxes , and fluxes (mg C m2 d-1) are represented by arrows. A = assimilation, E = excretion, Ex = export, I = ingestion, P = production, S = sedimentation, DIC = dissolved inorganic carbon (from Charpy and Charpy-Roubaud, 1990b). Detritus, smaller than 35 ym, represent the most important particulate organic carbon pool in the lagoon. They originate from lagoonal primary production (Charpy and Charpy-Roubaud, 1990b) and their sedimentation onto the bottom exceeds benthic primary production. Plankton bacteria biomass is of same order of magnitude as microphytobenthos biomass and is equal to twice the phytoplankton biomass. Pelagic bacteria dominance can be interpreted by a microbial loop returning energy released as dissolved organic matter by phytoplankton and zooplankton, but also energy released as mucus from lagoon coral communities. 26 ZOOBENTHOS BIOMASS IN SEDIMENTS : FIRST RESULTS A survey of 20 random stations was realised in the lagoon by A. Intes to provide a first assessment of the macro-zoobenthos of the soft bottoms. These unpublished data have to be considered as preliminary results. Taxonomic structure - results (Fig. 14) In terms of abundance, the Molluscs and the Polychaetes dominate the endofauna with a mean density of 11.3 and 10.4 ind m- respectively. The crustaceans count for less than 2 ind m-2 and the Echinoderms as well as the lancelets less than 1 ind m-2. The abundance of the Crustaceans is probably under estimated as only large apparent Invertebrates were counted and no digging operations were carried out. Large Invertebrates burrowing in the sediment are not taken into account and their biomass remains unknown. However, the average density of the big burrowing species can be estimated to stand around 0.3 ind m~ as revealed by hole density. Regarding the epifauna, the Sponges clearly dominate with around 0.5 colony m-2. The other organisms are generally scarce except in some very localised areas harbouring high densities of holothuroid (Halodeima atra). TAXONOMIC STRUCTURE Epifauna ME Endofauna Fig. 14 : Taxonomic structure of zoobenthos biomass in sediment (SW : Sea weeds ; ML: Molluscs ; SP : Sponges ; SC : Sipunculids ; PL: Polychaetes ; CR : Crustaceans ; ED : Echinoderms) expressed in Ash Free Dry Weight (AFDW). In terms of biomass, the global structure is not very different : the Molluscs remain at first rank with 36% of the animal organic matter (AOM), followed by the Sponges with 27% of AOM. The third class includes the Sipunculids, but also all the fauna living in the dead shells and in all probability some micromolluscs. This represents about 25% of AOM. The other groups stand far behind in importance : the Polychaetes are mainly small animals and contribute for only 9% to AOM. The crustaceans are essentially little species such as Tanaidacae or Mysidacae. Some crabs (mostly Portunids) are rarely collected in the endofauna and the contribution of the group to the biomass is poor : 2% of AOM. Once again, this is under estimated by the lack of information on the big burrowing forms. The Echinoderms are very few represented on the soft bottoms (Ophiuroids) with only 1% of AOM. 27 Considering the total living biomass, the primary producers largely dominate with about 60% of the total organic matter. They are sea weeds such as Halimeda or Caulerpa belonging to several species, but they may also be Phanerogams Halophila ovalis. (the sole species encountered in this lagoon). Trophic structure - results (Fig. 15) The deposit feeders are considered as a unique group. No distinction between the surface deposit feeders and the burrowing deposit feeders is made. They are slightly the best represented with 46% of AOM, but it must be emphasized that they belong to the endofauna (except the Holothuroid Halodeima atra). TROPHIC STRUCTURE Epifauna MB Endofauna Fig. 15 : Trophic structure of zoobenthos in sediments (SW : Sea weeds ; DF : Deposit-feeders ; SF : Suspension -feeders ; CR : Carnivorous), The suspension feeders stand in the same order of magnitude with 42% of AOM. Most of them belong to the same epifauna as Sponges do, but also some molluscs such as the bivalves Pinna are found. The carnivores are the least represented with only 12% of AOM. They are collected in the epifauna (Molluscs as Strombus spp. or Nassarius spp.) as well as in the endofauna (some Crustaceans and Polychaeta). The deposit feeders are under estimated because the mud shrimps, or the Cardiidae for example, are not taken in account by the sampling method. Discussion: A first biocenotic survey carried out by Faure and Laboute (1984) concluded from the study of the coral species distribution that the whole lagoon belongs to a sole community inside which the species distribution depends on geomorphological or physical factors. The study of the fish community by Morize et al. (1990) leads to the same finding : there is only one fish community in the lagoon, but its observed structures (which varies in space and biomass) are greatly heterogenous without any evident explicative factor. The spatial heterogeneity of the benthic biomasses cannot be explained by the classic factors such as sediment characteristics, bathymetry, and distance to the rim. 28 Two hypothesis can be advanced nevertheless : 1 -The mapping of the biomasses matches the mapping of the bidimensional modelisation of the water circulation under the influence of the trade winds (see fig.12 of the first chapter). If the model can be considered as relevant, the vortex generated by the wind will be a facilitating factor for the organic matter sedimentation and consequently, would allow the development of the greatest amount of benthic biomass of bottom invertebrates. 2 - The sample distance from a lagoon reef construction may be the best explicative factor. These reefs, considered as a source of organic matter via the detritus released, may govern the bottom invertebrates distribution. This hypothesis will be tested in the forthcoming research program "Cyel”. As only the soft bottoms were investigated in this quantitative survey, no cnidarians were collected : the free living scleractinia as Heteropsammia, Cycloseris, or Trachyphyllia do not exist in French Polynesia. Apart from this, the general structure is close to the structure observed in the New Caledonia lagoon by Chardy and Clavier (1989). The three main groups in terms of biomass are the same, respectively the macrophytes, the sponges and the molluscs. The crustaceans and especially the echinoderms are relatively less represented in the Tikehau lagoon than in New Caledonia. Few species of crabs and quite no pagurids, no urchins, no asterids were collected in or on the sediment. The biomass of the macrophytes is higher than the total animal biomass, but its trophic role in the bottom network cannot be appraised as few invertebrates seem to feed on them. Among the animals, the suspension feeders stand to the same order of magnitude of biomass than the deposit feeders in the Tikehau lagoon. Chardy and Clavier (1989), in New Caledonia, observed a suspensivores biomass about twice the one of the deposit feeders. The lack of soft bottom free living scleractinians may partly explain this fact. The general trends of the trophic network described on the soft bottoms of the Tikehau lagoon may be summarised as : A highly simplified fauna in which most of the zoological groups are represented, but with a little number of species ( about 80 taxa). The trophic network is basically highly dominated by the macrophytes production, but their consumers do not belong to the invertebrates fauna. Among the animals, the suspension feeding guild is mainly represented on the sediment surface by the sponges, but also by some molluscs or polychaetes living in the sediment. The deposit feeding guild is better represented among the zoological groups, where the molluscs and shell living organisms dominate in weight. The two main guilds are quite equally represented at least on the soft bottoms. A first attempt to explain this may be the weakness of the currents in the lagoon, allowing a near vertical sedimentation on which only filter feeding organisms (i.e. actively pumping the sea water) may feed in sufficient quantity before bottom accumulation. This hypothesis mark a basic difference with the bottom network of New Caledonia, where the water circulation allows a higher contribution of the suspension guild to the benthic network. 29 Literature cited ANDREWS .C.), GENTIEN (P.) - 1982 - Upwelling as a source of nutrients for the Great Barrier Reef eco- systems : a solution to Darwin’s question ? Mar. Ecol. Prog. Ser., 8 : 259 - 269 BLANCHOT JJ.), CHARPY (L.), LE BORGNE (R.) - 1989 - Size composition of particulate organic matter in the lagoon of Tikehau atoll (Tuamotu Archipelago). Mar. Biol., 102 : 329-339 CHARDY (P.), CLAVIER JJ.) - 1989 - An attempt to estimate the carbon budget for the south west lagoon of New Caledonia. Proceeding of the Sixth International Coral Reef Congress, Townsville, 1988, 2 : 541-546 CHARPY (L.) - 1985 - Distribution and composition of particulate organic matter in the lagoon of Tikehau (Tuamotu archipelago, French Polynesia). Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 2: 353-357 CHARPY (L.), CHARPY-ROUBAUD (C].) - 1990a - A model of the relationship between light and primary production in an atoll lagoon. J. mar. Biol. Ass. U.K., 70 : 357-369 CHARPY (L.), CHARPY-ROUBAUD (C,J.) - 1990b - Trophic structure and productivity of the lagoonal communities of Tikehau atoll (Tuamotu archipelago, French Polynesia). Hydrobiologia, 207 : 43-52 CHARPY (L.), CHARPY-ROUBAUD (C]J.) - 1991 - Particulate organic matter fluxes in a coral reef lagoon (Tikehau atoll, French Polynesia). Mar. Ecol. Prog. Ser., 71 : 53-63 CHARPY-ROUBAUD (C]J.) - 1988 - Production primaire des fonds meubles du lagon de Tikehau (archipel des Tuamotu, Polynésie Francaise). Oceanol. acta, 11 (3) : 241-248 CHARPY-ROUBAUD (C]J.), CHARPY (L.), LEMASSON (L.) - 1989 - Benthic and planktonic primary production of an open atoll lagoon (Tikehau, Tuamotu archipelago, French Polynesia). Proceeding of the Sixth International Coral Reef Congress, Townsville, 1988, 2 : 551-556 CHARPY-ROUBAUD (C]J.), CHARPY (L.), CREMOUX (J.L.) - 1990 - Nutrient budget of the lagoonal waters in an open central South Pacific atoll (Tikehau, Tuamotu, French Polynesia). Mar. Biol., 107 : 67-73 CONOVER (R,J.) - 1966 - Assimilation of organic matter by zooplankton. Limnol. Oceanogr., 11 : 338-345 CROSSLAND (C].) - 1983 - Dissolved nutrients in coral reef waters. In : Perspectives on coral reefs. Barnes DJ. (eds) Australian Institute for Marine Science, Townsville, : 49-56 FAURE (G.), LABOUTE (P.) - 1984 - Formations récifales : I Définition des unités récifales et distribution des principaux peuplements de Scléractinaires. In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Frangaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 108-136 FUHRMAN J.A.), SLEETER (T.D.), CARLSON (C.A.), PROCTOR (L.M.) - 1989 - Dominance of bacterial biomass in the Sargasso Sea and its ecological implications. Mar. Ecol. Prog. Ser., 57 : 207-217 HESSLEIN (R.H.) - 1976 - An in situ sampler for close internal pore water studies. Limnol. Oceanogr., 21: 912 -924 HALL (P.O J.) - 1984 - Chemical fluxes at the sediment sea-water interface ; in situ investigations with benthic chambers. Ph. D. Thesis, Chalmers university of technology, Goteborg, Sweden. HIROTA (.), SZYPER J.) - 1976 - Standing stocks of zooplankton size classes and trophic levels in Kaneohe bay, Oahu, Hawaiian islands. Pacif. Sci., 30 : 341-361 HOPKINSON (C:S.), SHERR (B.F.), DUCKLOW (H.W.) - 1987 - Microbial regeneration in the water column of Davies Reef Australia. Mar. ecol. Prog. Ser., 41 : 147-153 KOOP (K.), LARKUM (A.W.D.) - 1987 - Deposition of organic material in a coral reef lagoon, One Tree Island, Great Barrier Reef. Estuar. coast. Shelf Sci., 25 : 1-9 30 KROM (M.D.), BERNER (R.A.) - 1980 - Adsorption of phosphate in anoxic marine sediments. Limnol. Oceanogr., 25 : 797-806 LAWS (E.A.), Di TULLIO (G.R.), REDALJE (D.G.) - 1987 - High phytoplancton growth and production rates in the North Pacific subtropical gyre. Limnol. Oceanogr., 32 : 905-918 LE BORGNE (R.), BLANCHOT (J.), CHARPY (L.) - 1989 - Zooplankton of Tikehau atoll (Tuamotu archipelago) and its relationship to particulate matter. Mar. Biol., 102 : 341-353 LE BORGNE (R.), MOLL (P.) - 1986 - Growth rates of the salp in Tikehau atoll (Tuamotu is.) Oceanogr. trop., 21 : 23-29 LENHARDT (X.) - 1991 - Hydrodynamique des lagons d’atoll et d’ile haute en Polynésie Francaise. ORSTOM Paris ed., Etudes et Theses : 132 p LINLEY (E.A.S.), KOOP (K.) - 1986 - Significance of pelagic bacteria in a coral reef lagoon, One Tree Island, Great Barrier Reef. Mar. Biol., 92: 457-464 MORIARTY (D,J.W. ) - 1979 - Biomass of suspended bacteria over coral reefs. Mar. Biol., 53 : 193-200 MORIARTY (D.J.W.), POLLARD (P.C.), ALONGI (D.M.), WILKINSON (C.R.), GRAY (.S.) - 1985 - Bacterial productivity and trophic relationships with consumers on a coral reef (MECOR I). Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 3 : 457-462 MORIZE (E.), GALZIN (R.), HARMELIN-VIVIEN (M.), ARNAUDIN (H.) - 1990 - Organisation spatiale du peuplement ichthyologique dans le lagon de l'atoll de Tikehau, Polynésie francaise. ORSTOM- Tahiti, Notes et Doc Océanogr., 40 : 1-44 QUASIM (S.Z.), SANKARANARAYANAN (V.N.) - 1970 - Production of particulate organic matter by the reef of Karavatti atoll (Laccadives). Limnol. Oceanogr., 15 : 574-578 ROUGERIE (F.) , WAUTHY (B) - 1986 - Le concept d’endo-upwelling dans le fonctionnement des atolls-oasis. Oceanol. Acta, 9 : 133-148 SHERR (B.F.), SHERR (E.B.), NEWELL (S.Y.) - 1984 - Abundance and productivity of heterotrophic nanozooplankton in Georgian coastal waters. J. Plankton Res., 6 : 195-202 SMITH (S.V.) - 1984 - Phosphorus versus nitrogen limitation in the marine environment. Limnol.Oceanogr., 29 (6) : 1149-1160 SMITH (S.V.), CHANDRA (S.), KWITKO (L.), SCHNEIDER (R.C.), SCHOONMAKER J.), SEETO (.), TEBANO (T.), TRIBBLE (G.W.) - 1984 - Chemical stoechiometry of lagoonal metabolism. U. Hawaii/U. South Pacificic Internat. Sea Grant Prog. Tech. Rept. UNIHI-SEAGRANT-CR-84-02, 30 p SMITH (S.V), JOKIEL (P.L.) - 1975 - Water composition and biogeochemical gradients in the Canton atoll lagoon : 2 Budgets of phosphorus, nitrogen, carbon dioxide, and particulate materials. Marine Science communications, 1 (2) : 165-207 SMITH (S.V), KINSEY (D.W.) - 1988 - Why don't budgets of energy, nutrients, and carbonates always balance at the level of organisms, reefs, and tropical oceans ? an overview. Proceeding of the Sixth International Coral Reef Congress, Townsville, 1988, 2 : 551-556 SOROKIN (Y.I.) -1974 - Bacteria as a component of the coral reef community. Proceeding of the Second International Coral Reef Symposium, Manilla, 1 : 3-10 TAGUCHI (S.) - 1982 - Sedimentation of newly produced particulate organic matter in a subtropical inlet, Kaneohe Bay, Hawaii. Estuar. coast. Shelf Sci., 14 : 533-544 TAKAHASHI (M.), KIKUSHI (K.), HARA (Y.) - 1985 - Importance of picocyanobacteria biomass (unicellular, blue-green algae) in the phytoplankton population of the coastal waters off Japan. Mar. Biol., 89 : 63-69 PART II. REEF FISH COMMUNITIES AND FISHERY YIELDS OF TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY B. CAILLART!, M.L. HARMELIN-VIVIEN’, R. GALZIN*®, AND E. MORIZE’ INTRODUCTION Fish communities in the lagoon of the Tikehau atoll were studied by only a few researchers. Harmelin-Vivien (1984) studied the distribution of the main herbivorous families (Scaridae and Acanthuridae) in the lagoon and on the outer slope to 30 m in depth. The total fish community of the outer slope was studied by Galzin (1985, 1987) at 12 m in depth. These studies were carried out in the southwestern part of the atoll. Spatial organization of coral associated fish community was studied throughout the lagoon by Morize et al. (1990). Most of the other studies undertaken at Tikehau involved the artisanal fishery (Morize, 1984, 1985 ; Caillart and Morize, 1986, 1988) and the biology of some target species to the exploited stock (Caillart et al., 1986; Caillart, 1988; Morize et Caillart, 1987). It seems worthwhile to present in this special issue of ARB, all the available information on the fish fauna of Tikehau. Furthermore, this overview allows us to compare our results with others in the Indopacific region. FISH COMMUNITIES OF TIKEHAU ATOLL METHODS To study the fish communities of Tikehau, two complementary methods were used : visual census and rotenone poisoning. Many synthesis (GBRMPA, 1978; Barans and Bortone, 1983; Harmelin-Vivien et al., 1985) describe the method for estimating in situ fish communities and populations using visual censuses. These methods, widely used on coral reefs, enable scientists to study fish communities without perturbation. In the Tikehau atoll, visual censuses were carried out by SCUBA diving on 50 m length and 5 m width transects. The transects on the outer slope and the inner reef flat were parallel to the reef. Around the pinnacles, the sampling line was curved around them. In each transect, abundance inside three caregories of size (small - medium - large) were recordered for all species encountered (St. John et al., 1990). 1: Centre ORSTOM de Brest, BP 70, 29280 Plouzané, France 2: Centre d'Océanologie de Marseille, CNRS-UA 41, Station Marine d'Endoume, 13007 Marseille, France 3: Ecole Pratique des Hautes Etudes, CNRS-URA 1453, Laboratoire d'Ichtyoecologie Tropicale et Mediterraneenne, 66860 Perpignan Cedex and Centre de l'Environnement, BP 1013 Moorea, French Polynesia. 2 Fish hidden in reef shelters and in sediment were caught with ichtyotoxic rotenone. Individual fish were measured to the nearest millimeter (standard and total length), weighted to the nearest gram and preserved in a 10% neutral formalin. Length-weight relationships were subsequently computed for all species caught, providing that the sample size was large enough. The two methods were used to study fish communities in the lagoon, put only visual censuses were used on the outer slope. FISH COMMUNITIES OF THE OUTER SLOPE Fish communities of the outer slope are strongly influenced by environmental factors : primary substratum types, slope gradient, level of wind exposure, and magnitude of the 1983 cyclonic damages on coral assemblages. Several surveys of the outer slope fish fauna were carried out on the southwestern outer slope of the atoll (Fig. 1). For the damages induced by the cyclones and the description of the outer slope, see the previous chapter on the environment by Intes and Caillart. 148°15' 14°55° ~ Motu Oeote Motu Puarua a vi y ,Paaie Mamaa ‘Motu Mamaa RS Tohuaao ae” ‘a in x ov a yy \ « Mauu eos 29 > YZ 15°05’ S XN %, so . are ue “Sy os 148°05°W uherahera ( . fulage) oe Fig. 1 : Location (Mamaa-arrow) of fish community sampling station on the outer slope of the Tikehau atoll . The fore reef area (0-10 m) The spur and groove zone is an area of very high fish abundance. In particular surgeonfish Acanthurus achilles, A. nigroris, A. guttatus, A. lineatus and parrotfish Scarus sordidus , Scarus sp. are typical features of this zone. Small coral dependant fish found are Cirrhitidae, small Serranidae (Cephalopholis urodelus), Chaetodontidae (Chaetodon quadrimaculatus) Pomacanthidae (Centropyge sp.), Labridae (Thalassoma fuscum), numerous Balistidae (B. viridescens, B. undulatus and Melichthys niger and M. vidua in mid-water). The shark Carcharhinus melanopterus and a great variety of Carangidae are also frequentely encountered in this productive and well oxygenated area. 3 On the fore reef platform (4-10 m), benthic fish fauna (e.g. : Gobiidae, Chaetodontidae, Acanthuridae, Serranidae, Labridae) can be distinguished from zooplankton-feeding mid- water fish fauna (e.g. : Anthias spp., Pomacentridae, nocturnal Holocentridae and Naso spp.), and upper-water fish fauna (Balistidae, sharks, tunas and Sphyraenidae). The outer terrace (10-25 m) The fish fauna of this zone present a great diversity (more than 100 species), and an abundance of fishes. The most conspicuous families are Holocentridae (genus Holocentrus, Sargocentron, Myripristis) numerous around coral patches, Lutjanidae (Lutjanus bohar, L. gibbus, L. kasmira) forming schools of several hundred individuals, Acanthuridae (Ctenochaetus striatus, C. strigosus, Zebrasoma scopas, Acanthurus glaucopareius, A. nubilus and schools of Naso spp.), Serranidae (genus Variola, Gracila and the common grouper Epinephelus microdon), Chaetodontidae and some Scaridae (Scarus gibbus, S. niger, Cetoscarus bicolor). The deep outer slope (from 25 m) Abundance and diversity of fish fauna decrease somewhat but a new, more characterized, species assemblage occurs with depth. Holocentridae and Scaridae are less important while the abundance of large Serranidae, some Labridae (genus Bodianus, Cirrhilabrus), Zanclidae and Heniochus noticeably increase. Among Chaetodontidae still present, species of the genus Hemithaurichtys appear. Among the Acanthurid censused are, Acanthurus bleekeri, A. pyroferus, A. xanthopterus and large schools of Naso hexacanthus and Naso vlamingii. Lutjanidae, with large Lutjanus bohar, are numerous as well. The abundance of parrotfish decreases rapidly below 30 m. Fish assemblage was not studied below 40 m on the outer slope of the Tikehau atoll. Temporal variations of fish communities Numerous authors working on coral reef ecosytems, and Bell and Galzin (1984) and Galzin et al. (1990) in French Polynesia, emphasized that a strong relationship exists between the live coral coverage rate and fish repartition. As shown in Table 1, dramatic changes occured in live coral coverage rate on transect under investigations in five years, inducing a renewal of fish assemblages. Most of these dramatic changes were induced by six cyclones which ravaged french Polynesia during the hot season 1982-83 (Harmelin-Vivien and Laboute, 1986). Table 1 : Live coral coverage rate of the southwestern outer slope of the Tikehau atoll before, immediatly after and five years after cyclones. 1982 1983 1987 Depth Before the cyclones After the cyclones (Galzin et Harmelin- (Faure et Laboute, 1984) (Harmelin-Vivien et Vivien, unp data) Laboute, 1986) 3m 5 to 25 % <5% - 5m 40 to 60 % 20 to 25 % 56 to 62 % 10 m 40 to 60 % 20 to 25 % 42% 20 m 40 to 60 % 15% 22 to 24 % 30 m 40 to 60 % 15% 16 to 24 % 4 Data displayed in Table 2 permit the assessment of fish fauna temporal variations. Between 1983 and 1987, total number of species on the fore reef area increased from 46 to 56 due to a conspicuous resettlement of Serranidae, Pomacentridae and Labridae. On all other biota of the outer slope, the total number of species decreased between 1983 and 1987. Most of the Scaridae, Acanthuridae and Balistidae left the 10 m depth area whereas most of Holocentridae, Lutjanidae and Mullidae usually encountered around 20 m in depth, moved away. Fish densities at 20 m depth decreased dramatically between 1983 and 1987 (i.e. : from 3.4 ind m-2 in 1983 to 2.6 ind m-2 in 1987 on the average). Table 2 : Main characteristics of the ichtyological fauna on the outer slope of Tikehau at different depths before, just after, and five years after cyclones of late 1982 - early 1983. Nhs : Number of herbivorous species, Dih : Number of individuals of herbivorous species . 100 m-2,,Nst : Total number of species, Dsi : Number of all individuals . 100 m-2. (— : no data). Depth 1982 1983 1987 (m) Before cyclones After cyclones Nhs Dih Nst_ Dsi Nhs Dih Nst_ Dsi Nhs Dih Nst_ Dsi 5 155% G59 -- - We 213 46 - 197) 143 56 -- 10 20 «188 -- - 21 78 40 337 12 55 69 260 20 21 199 -- - 19 9174 78 - 25 152. 67. - 30 19 =140 - - 7 epel Ot 58 - 220152 - - Herbivorous species were studied in more detail. Data listed in Table 2 and 3 show that for herbivorous fishes the mean number of individuals is relatively constant at 5, 20 and 30 m depths between 1982 and 1987. As previously noticed, the only anomality is found at a 10 m depth where the number of herbivorous species on the outer slope undergo a veritable decrease : 1.7 ind. m-2 in 1982, 1.4 ind. m-2 in 1983 and 1.2 ind m2 in 1987. After the cyclones, fish fauna decreased considerably. A great number of cryptic species died with associated corals, another part remained unsheltered and suffered subsequently from higher predation by piscivorous species like Epinephelus microdon that became more abundant after the cyclones. Another part of fish fauna escaped toward undamaged reef areas. A re- arrangement of fish fauna was noticed on the outer slope ; a greater number of species were counted in shallow areas. COMPARISON WITH FISH COMMUNITIES OF OTHER OUTER SLOPES Galzin (1985) has compared fish communities in the outer slopes of 2 high islands (Moorea, Mehetia) and 3 atolls (Tikehau, Takapoto and Mataiva) of French Polynesia. Qualitative and quantitative studies show that fish communities found at a 12 m depth on atoll outer slopes are different than those found on high island outer slopes (Moorea, Fig. 2). Out of the 189 species censused in ten sampling stations, 8 (4%) are found exclusively at Tikehau. These are : Elagatis bipinnulata, Lethrinus elongatus, Lethrinus xanthochilus, Chromis margaritifer, Bodianus loxozonus, Cetoscarus bicolor and Scarus niger. Pomacentrid Chromis xanthura is unexpectedly absent from Mataiva and Tikehau outer slopes whereas it is present at the 8 other sampling stations. Differences in coral coverage can also be a major factor since outer slope sampled at Tikehau and Mataiva have been damaged by cyclones to a greater extent than the southwestern outer slope of Takapoto (Galzin, 1987 ; Harmelin-Vivien and Laboute, 1986). The current state of knowledge does not enable to isolate the major factors influencing fish repartition on atoll outer slopes in French Polynesia. ep ee ee ee © © © © © © © © © co He © © © © © eo ew © oo 16 17 18 19 2015 13 14 11 12 Tikehau are a A = N é 19 as? ar Aa Zia 5hm Mataiva Jaccard’s similarity indices Stations Fig. 2 : Location of sampling stations on each of the five islands and dendrogram derived from similarity matrices. Numbers refer to the 10 sampling sites distributed among the five islands (from Galzin, 1987). Table 3 : Temporal variability for two families of herbivorous fish (Scaridae and Acanthuridae) on the outer slope of the Tikehau atoll. (number of individuals . 1000 m-2)), 1982 1983 1987 =5 7-10-20)" -30...-5 =10 5-20) =30) -5) =10 .=20)-30 SCARIDAE Cetoscarus bicolor 7 3 4 2 6 2 2 Hipposcarus longiceps 1 4 2 11 2 Scarus altipinnis 5 12 6 7 3 5 4 1 4 1 Scarus forsteri Scarus frenatus 5 9 11 14 2 5 4 1 5 2 Scarus frontalis 8 2 5 1 4 Scarus ghobban 5 10 2 1 4 2 3 4 3 Scarus gibbus So eg 304 16 1 10 pe a2 3 1 4 Scarus globiceps 1 4 1 Scarus niger 1 8 3 16 3 2 Scarus oviceps 1 Scarus psittacus 2 6 1 Scarus rubroviolaceus 1 2 1 Scarus schlegeli 1 9 72 6 Scarus sordidus 27 SNS ge 21 Vea 77. 7 SOM Se 14; 9 132 748 Scarus juv. iis) Poee74s) 6 10 Number of species 6 1GW | pa) 5 8 10 8 7 8 4 10 7 Number of individuals 657 103°) ¥88i5 (524 | 107.) 41 78s 100) 7) 4250) 1539) eos ACANTHURIDAE Acanthurus achilles 24 «8 51 10 Acanthurus bleekeri Acanthurus glaucopareius 219 26, yl 5), 35 2 360 4 6 Wis 3) Acanthurus guttatus 13 13 4 Acanthurus nigricauda 1 2 2: 4 24 «13 Acanthurus nigroris 108 85 145 = 74 170 75 Acanthurus nubilus 9 4 14 4 2 Acanthurus olivaceus 11 9 7 8 Acanthurus pyroferus 2 24 46 40 1937, Acanthurus thompsoni BA 227 377730 Acanthurus triostegus 2 6 3 86 Acanthurus xanthopterus 14 ~—-60 Ctenochaetus striatus 90149 ena 16) 9-153 o15.S 72a 2 5 14 50 48 Ctenochaetus strigosus 60 6 164 = 104 2 148 74 74 2 75\ 150 Naso brevirostris 10 7 7 NE: 12 5 Naso hexacanthus 8 Naso lituratus 7 13) 52), 15s 12 34. 24 6 20 8 23 «= 46 Naso vlamingtii 1 6 3 6 Zebrasoma rostratum 8 4 1 6 8 7 1 10 3 13 2 Zebrasoma scopas 17 6 21 3 15) 18 1 7 Vey Zebrasoma veliferum 2 2 1 3 1 8 1 Acanthurus juv. 2 5 Number of species 9 9 abt 14 9 11 11 10 11 8 15) 5 Number of individuals 333. 367 410 299 425 154 358 192 315 123 341 318 FISH COMMUNITIES ASSOCIATED WITH CORAL FORMATIONS In the lagoon of Tikehau, three main types of biotopes can be distinguished : coral reef formations, sediments and mid-water. Coral reef formations are composed by the inner reef flat, pinnacles and coral patches. They are scattered all over the lagoon but are more numerous in the front of channels (Harmelin-Vivien, 1985). In the southern and western part of the lagoon, the inner reef flat that edges the atoll rim lagoonward does not extend deeper than 5-6 m. Live corals extend down to 15 m depth on pinnacle slopes. Pinnacles are more abundant in the western part of the lagoon, especially between the village and the pass. Fish abundance on Takapoto’s outer slope (4 to 5 ind . m-2) appears to be greater than that on the outer slopes of Tikehau and Mataiva (3 to 4 ind . m2) (Table 4). However, the difference is not statistically significant. This difference can be explained either by geomorphological considerations (presence/absence of a pass) or by variations in longitudinal position. Table 4: Comparative quantitative data for the coral reef fish community at 12 m depth of the outer-slope of 5 islands in French Polynesia. NI : Number of individuals . 100 m-2, NS : Number of species . 100 m-2, NI NS stations NI Mean NI Mean (s.d.) (s.d.) 11 575 44 12 428 32 MOOREA 13 220 400 27 35 14 378 (146) 37 (7) 16 487 46 TAKAPOTO 17 435 442 41 45 18 404 (42) 47 (3) MEHETIA 15 516 46 TIKEHAU 19 337 418 43 45 MATAIVA 20 400 (91) 47 (2) Only the fish community associated with coral formations was studied in detail (Harmelin- Vivien, 1984; Morize et al., 1990). The total fish fauna of the lagoon is obviously richer because soft-bottom and mid-water fish communities were under-sampled (St. John ef al., 1990). Lagoon fish communities are divided as follows : 1) fish species remaining in the lagoon during their entire life span after recruitment to the reef, 2) fish species that, at least at one time of their life, live on the outer slope or in ocean water, 3) and species living on the outer slope but that migrate toward the lagoon for reproduction. Species with different life cycles gather especially near the pass. Structure of the coral associated fish fauna in Tikehau lagoon. A total of 164 fish species, belonging to 34 families were censused around the coral formations in the Tikehau lagoon : 99 species were observed by visual census and 108 species were caught by rotenone poisoning (Appendix 1). The most diversified families are Labridae (21 spp.), Acanthuridae (20 spp.), Scaridae (14 spp.), Serranidae and Chaetodontidae (7 spp.). All species recorded only by visual census live in mid-water. These species belong to families of Carcharinidae, Fistularidae, Echeneidae, Carangidae, Lutjanidae, Lethrinidae and Zanclidae. gS ff, ples, as ot oa 15° suo D aA PASSE TuUHElava © 2 fe) : QQ hee - abe y {- TIKEHAU ° c\) oc ry we oO p : TUHERAHERA Lh a need gy 1489 10° ovest Km VILLAGE Fig. 3 : Mean fish densities at different sampling stations in the Tikehau lagoon (2 fish : 100 ind . 100 m-2) (Morize et al., 1990). 15° suo VILLAGE TUHERAHERA 1485 10° ovest Fig. 4 : Mean fish biomass at different sampling stations (numbers) in the Tikehau Lagoon (modified from Morize et al., 1990). 9 On the other hand, species recorded only by rotenone poisoning, are cryptic species or live buried in sediments (Congridae, Ophichthidae, Ophidiidae, Scorpaenidae, Blenniidae and Bothidae). Only a part of the fish fauna of coral formations can be recorded by each method (60% by visual census and 67% by rotenone poisoning). Only 30 % of all species are recorded by both methods. The composition of fish species is relatively homogeneous throughout the whole lagoon. Down to a depth of 15 m, the distribution of species does not show any gradient over the whole lagoon (Morize et al., 1990). The same fish community is found around coral reef patches of the lagoon of Tikehau. On a biomass basis, this community is made up of about 70% of carnivorous species, 14% of omnivorous species and 17% of herbivorous species (Table 5). However the trophic structure of the community observed is different according to the method of sampling. Samples obtained by rotenone poisoning allow to have a better estimation of the abundance of nocturnal plankton feeders, nocturnal carnivores and omnivores. On the other hand, diurnal plankton feeders, sessile invertebrates feeders and herbivorous species are better sampled with visual censuses. Table 5 : Comparison of trophic structure of fish community in the Tikehau lagoon related to the two assessment methods (expressed as percentage of total number of species ). Total Visual rotenone community counts poisoning Total number of species 161 97 108 % piscivorous 9.4 8.3 10.2 % other carnivorous nocturnal 18.7 115 20.4 diurnal 20,6 20,8 22,2 % planktivorous nocturnal 8.8 4.2 12.0 diurnal Sul S72 19 % sessile invertebrate browsers 9.4 146 9.3 % omnivorous 13.7 8.3 18.5 % herbivorous 16.3 27.1 55 Spatial distribution of fishes in the Tikehau lagoon Geographical distribution The small-scale spatial heterogeneity of fish community in the lagoon is considerable. However the distribution of this community follows a steady pattern all around the pinnacles. On the windward area of the pinnacles, species richness, density, and biomass of fish are always higher (between 1.5 to 4 times) than on leeward ones (Morize et al., 1990). In spite of a relatively homogeneous distribution of fish in the whole lagoon, densities, biomass and length frequencies of fish of this community present a heterogeneous spatial distribution. Densities : Depending on the sites, average fish density around pinnacles of the Tikehau lagoon vary from 102 to 1274 fishes per 100 m2: The highest mean densities are located windward of the atoll (in the northeastern part of the lagoon, Fig. 3). Biomass : The biomass of the 31 most abundant species varies from 0.8 to 34.4 kg . 100 m-2 and display a considerable spatial heterogeneity. The spatial variations of biomass seem to depend in part upon the localisation of studied sites from the reef flat, the village and the pass (Morize et al., 1990). The most important average biomass is recorded near the center of the lagoon and at the pass of the atoll (Fig. 4). 10 Table 6 : Mean demographic structure of fish populations around coral pinnacles in the Tikehau lagoon (D : mean density of individuals 100 m2;% : percentage of each total population size class). Station location 1 SW SW 2 4 7 3 E 5 6 8 NNW pass Tuheiava Size of juveniles © 80-120 mm () 170 mm Fig. 5: Size-class repartition of Naso brevirostris in the Tikehau Lagoon. (number Juveniles D % 12.4 8.0 6.0 12.9 20.7 18.4 7.9 Sy) 26.1 10.7 31.6 6.7 20.5 13.0 19.8 14.0 148°15' juveniles recorded per transect), Adults Olds % D 53.6 59.3 58.5 13.3 43.6 42.8 68.9 39.0 65.3 58.3 62.3 146.5 59)2. 43.8 54.6 41.4 as 148°05'W 24.0 31.0 27.8 29.4 Density of adults (nb. indiv./transect) A 10-90 A 100 - 300 : number of 11 Age structure : Generally, middle sized fish are the most numerous and represent from 44 up to 70% of the total number of fish (Table 6). The number of the largest fish varies between one fourth and one third of the total population, while juveniles are less numerous (5 to 8%). The low abundance of juveniles may be due to the fact that they are not easily seen by divers or that they recruit somewhere else to other biotopes. Furthermore, the length frequency distribution throughout the lagoon is not homogeneous. Juveniles are more numerous in northern and eastern parts of the lagoon. These areas receive oceanic water passing over the reef flat through hoa which are particularly numerous. The distribution of length frequencies of Naso brevirostris is a good example that shows differences in juvenile and in adult fish distribution (Fig. 5). All small juveniles (80-120 mm) were observed in the eastern part of the lagoon while larger juveniles (170 mm) were seen mostly in the western part (Caillart, 1988). Conversely, the density of adult fishes in the western part, and particularly near the pass, is four time higher than in the eastern part. Distribution with depth Specific_composition : The species richness of the fish community in the lagoon is greater between 3 and 5 meters depth : 87 species were recorded at these depths by visual censuses. From 10 to 15 m, the community is poorer (only 65 species censused) but is not qualitatively different. Only one species, Gobiidae Amblygobius phalaena, appears to be a characteristic species of this deeper zone. Inversely, some Mullidae (Mulloides spp., Parupaeneus porphyreus), Pomacanthidae, some Labridae (Gomphosus varius, Thalassoma amblycephalum), Scaridae (Scarus globiceps) and Acanthuridae (Acanthurus nigroris, Zebrasoma veliferum) were not inventoried deeper than 5 m. Density and biomass : For the whole community, there is no significant difference in mean fish density and biomass between 5 m and 15 m in the Tikehau lagoon (Table 7). However, most species or families are not uniformly distributed with depth : Scaridae and Acanthuridae densities are greater on the inner reef flat and on the top of pinnacles, and decrease with depth (Harmelin-Vivien, 1984). Similarly, Labridae are more numerous near the surface than at 15 m. Conversely, the density of Lutjanidae, Gobiidae and some Pomacentridae, like Pomacentrus pavo, are higher at 15 m (Morize et al., 1990). Table 7 : Mean density and mean biomass of reef fishes estimated from visual census according to depth (number of replicates n=8). Density Biomass Nd indiv. 100 m-2 g 100 m-2 -5m -15m -5m -15m mean 413.5 318.0 11465.6 10236.0 SD 365.5 195.0 10236.0 10109.6 Age structures : The average density of larger fishes on the whole fish community is more important between a depth of 3 to 5 m (Fig. 6). It decreases with depth and on reef flats (Harmelin-vivien, 1984 ; Morize et al., 1990). The average density of juvenile fish is in turn more important at 15 m deep than at 5 m. Meanwhile, distribution of length classes with depth differs among families. The highest densities of juveniles of Scaridae and Acanthuridae were observed in shallow waters (0-2 m) (Table 8). Conversely, juveniles of Lutjanidae, Labridae and Pomacentridae are more numerous between 10 and 15 m depth. Nb individuals . 100 m-2 [Tar] , m 100 YY I5m 80 Y y L é U7 7 7; = YY i YY SS LL Juveniles Adults old Adults size class Fig. 6 : Mean demographic structure of fish community at two depths (5 and 15 m) around pinnacle reefs of Tikehau lagoon. Table 8 : Mean density of juvenile parrotfishes (Scaridae) and juvenile surgeonfishes (Acanthuridae) with depth in Tikehau lagoon (number of individuals . 100 m-). 0-2 m 3-5m 10-15 m Scaridae mean 12.3 96 7.2 SD 5.4 8.3 5.6 Acanthuridae mean 2.9 0.4 0.2 SD 15 0.9 0.2 13 Comparison with other Tuamotu atolls The fish communities of atoll lagoons were studied by different authors with a different sampling design in five other Tuamotu atolls, Takapoto, Scilly, Mataiva, Fangataufa and Mururoa (Table 9). Each of these lagoonal communities differs somehow from the other, either by its specific composition or by its average density and biomass, whereas the outer reef slopes look much alike (Galzin, 1987). Table 9 : Comparison of lagoon fish communities associated with coral formations of six Tuamotu atolls : total number of species and mean density of individuals. Atoll Number of Depth Density reference fish species (m) (nb indiv 100 m-2) mean range Takapoto 170 0-20 - - 1 Scilly 180 0-30 - - 2 Tikehau 161 3-5 414 102-1274 3 10-15 318 104-612 3 Mataiva 157 0-3 50 3-125 4 Fangataufa 128 0.3 164 54-275 5 Mururoa 230 12 188 56-531 6 References : 1. Bagnis, Galzin and Bennett, 1979 (28 sites in lagoon, 16 in hoa) 2. Galzin, Bagnis and Bennett, 1983 (2 transects in lagoon, 2 transects in hoa, 4 transects on outer reef flat). 3. Morize, Galzin, Harmelin-Vivien and Arnaudin, 1990 (8 sites in lagoon, 4 transects on inner reef flat). 4. Galzin, Bell and Lefévre, 1990 (8 sites in lagoon surveyed 4 times in 8 years). 5 . Galzin, unpublished data (7 sites in lagoon). 6. Galzin, unpublished data (10 sites in lagoon, 6 sites on inner reef slope) The observed species richness is low in the Fangataufa lagoon (128 spp.), an atoll without a natural pass. It is in turn very high in the Mururoa lagoon (230 spp.), an atoll widely opened to Oceanic waters. The number of species recorded in the four other atolls are closely related in spite of differences in morphological structures : Tikehau and Mataiva have a pass whereas Takapoto and Scilly do not. The mean fish density is very low in the lagoon of Mataiva (Table 9) ; this phenomenon can be explained by a distrophic crisis that seems to affect this atoll (Galzin et al., 1990). On the other hand the mean density of fish is higher in the Tikehau lagoon, in spite of a considerable exploitation of fish stock. At a 12 m depth, the density of fish is lower at Mururoa than at Tikehau. However, the average length of fish is much larger at Mururoa where there is no fishery. The average biomass of fish is probably the same in these two lagoons. Conclusion Only one fish community is observed around coral formations (pinnacles) in the lagoon of Tikehau. The mean fish density and biomass do not vary with depth, although the species richness is lower at 15 m deep than between 3 and 5 m. The highest fish densities, characterized by a great proportion of juveniles, are generally found in the northern and eastern parts of the lagoon. Mean biomass per unit area is generally the highest in the southern and particularly the western part of the lagoon, near the pass, characterized by a great proportion of large-sized fishes. The depth vs age structure of population relationship varies according to families or species. Juvenile densities are higher in shallow water for some families (Scaridae, Acanthuridae) or in deeper water for other families (Lutjanidae, Labridae, Pomacentridae). 14 Lagoon - outer slope comparison Fish community of the outer slope of the Tikehau atoll is more diversified than the lagoonal fish community (Appendix 1). Indeed, twice as many species of fish were recorded on the outer slope as in the lagoon by visual censuses. Around Moorea island, Galzin (1987) recorded also a greater fish species richness on the outer slope than in the lagoon and on reef flats. Among families, some fish species are more numerous on the outer slope than in the lagoon of Tikehau and vice versa. Serranidae, Cirrhitidae, Carangidae, Lutjanidae, Chaetodontidae and Balistidae species are more numerous on the outer slope (31 spp.) than in the lagoon (24 spp.) (Harmelin-Vivien, 1984). However, the mean Acanthuridae density is higher on the outer slope, whereas Scaridae density is higher in the lagoon (Table 3). Other families, like Lethrinidae and Mullidae are more diversified and their populations are much more abundant in the lagoon as compared to the outer slope. The distribution of length class, and sex ratio may be also different into or out of the lagoon for a same species or a same family. The most juvenile Scaridae were observed in the lagoon. Immature males and females are more abundant in the lagoon, whereas ripe males are much more numerous on the outer slope (Harmelin-Vivien, 1984). Fish communities of the outer slope and of the lagoon of the Tikehau atoll are different not only by their species richness and population density, but differ also by their age and trophic structures. THE EXPLOITED LAGOON RESOURCE : THE FISHERY OF TIKEHAU The fishery of Tikehau is of artisanal nature, in which fish are sought for commercial and subsistence purpose. It is based principally upon the use of a relatively simple gear : bottom- fixed fish traps. Additionally, an important proportion of fish is occasionally taken with hook and line or spear gun. The fishery of Tikehau was thoroughly studied for four years. Numerous data on the fishery yields and on the biology and behavior of the target species have been gathered in order to assess the reef fish stock for management purposes. THE FISHERY OF TIKEHAU The fishing gear Traditionally, fish traps were built in shallow waters using rocks or coral boulders. Blanchet et al. (1985) pointed out that yields were low but satisfactory sufficient to meet the needs of the low-level human population. In the middle of the century, intensive phosphate mining on the neighboring island of Makatea created and kept a high sustained demand of fish to feed the population of workers (about 3,000 in 1962). As a result, the subsistence fishery of Tikehau developed into a commercial fishery by setting traps in more productive areas (in the vicinity of the pass), using modern building materials (wire net, iron stakes) as well as traditionnal wooden stakes. After the close-down of the mining site in 1966, fish trading logically reoriented toward the Tahiti fish market. CATCHING ROOM RETAINING Ve Sketch-view of a trap with two ROOM retaining room COEFECTHING EXTENSIONS Fig 7 : Diagram of a typical Tikehau fish trap (actually fish-trap n°2, see text for more details). As shown in the framed sketch-view, there can be two retaining rooms. Dotted line : wire-net, underlined number : depth at which the part of the trap is set. The general shape of a Tikehau fish trap is presented on Fig. 7. A fish - or a school of fish - coming across the large collecting extensions of wire net (locally termed Rauroa) are naturally driven toward the catching room (locally called Aua) in which they enter through a narrow opening. At least every day, trapped fish are herded off the trap by fishermen banging on the water surface and driven into a first retaining room (Tipua) where they can be held alive for a couple of weeks until they are sold. The fish are landed when the small trading vessel, able to load between 12 and 15 metric tons of catches, arrives at Tikehau (usually once a week), and subsequently shipped to Tahiti. 16 TRAP N°4 TRAP N*1 LAGOON Fig. 8 : Location of fish traps in the pass of Tikehau. Trap n°1 and n°2 located close to the middle of the pass are far more efficient than the two others located lagoonward on the reef flat. Although the origin of a great part of the catches has not been accurately determined, data available indicate that trap n°1 and n°2 yield at least 78 % of the total catches. The main fishery of Tikehau uses four fish traps, all located in or around the vicinity of the pass. Two traps (trap #1 and #2) are set quite in the middle of the pass by up to a 5 m depth (Fig. 8), and the two others traps (#3 and #4) are located lagoonward on the northern shore of the pass in shallower water (1 to 2 m). When fish is thought to be abundant in the pass and if current speed allows underwater work, a net is set across the pass between trap #1 and trap #2 and a "scare line" driving technique is used to increase catches. Handlines and spear-guns are used mostly for a subsistence purpose. The use of these kind of gears can however significantly contribute to commercial catches when there are huge concentrations of groupers or emperors in the pass making these species readily available in large quantities. Fishermen retain a portion of their catch for their own use and sell the remainder to the trader. 17 Fishery yield In Tikehau, statistical sampling of the catches was done by a local agent who noted down ona log sheet the species composition of the catch, weights sold on a species basis and the number of fish traps that provided the catch. Various information relevant to fishing such as current strength in the pass and weather were also recorded (Morize, 1984). Data were recorded from 1983 to 1987. As fishing activities are maximal by the end of the year and lower by July - August, a fishing year was defined to run from Ist of July to 30th June. Thus, the study of the fishery of Tikehau was carried out upon four fishing years : 83-84, 84-85, 85-86 and 86-87. Morize (1984) pointed out that fishing effort is somewhat difficult to appraise but since the shape, the number and the location of the traps have not been modified during the study, the fishing effort can be assessed as the number of days with a fully efficient presence of the traps on the fishing grounds. As the level of fishing effort can be estimated to have been constant, variations of catch per unit effort (c.p.u.e.) correspond with variations of catch. Table 10 gives an inventory list of species caught in Tikehau fish traps (comprehensive studies available in Morize, 1985 ; Caillart and Morize, 1986). Almost fifty species are likely to be trapped, covering a complete trophic spectrum of species ranging from piscivorous to herbivorous species. Although the selectivity of the gear appears to be poor, no more than fourteen species significantly contribute to the catch by accounting for about 85 % of the total landing. These fish include lutjanid Lutjanus gibbus and Lutjanus fulvus, lethrinid Lethrinus miniatus, carangid Caranx melampygus, Decapterus macarellus and Selar crumenophthalmus, serranid Epinephelus microdon, acanthurid Naso brevirostris and Acanthurus xanthopterus, mullid Upeneus vittatus and Mulloides spp., albulid Albula vulpes, sphyraenid Sphyraena forsteri, and lastly holocentrid Myripristis spp. Table 11 shows that total harvests obtained through trap fishing range from 144 metric tons to 207 metric tons a year with an average value of 165 metric tons. Lethrinus miniatus is the principal component of the catches with an average landing of 32 metric tons per year. It is followed by Lutjanus gibbus, Caranx melampygus and Selar crumenophthalmus representing a yearly average catch of respectively 17, 16 and 14 metric tons. These species can be dubbed target species though fishing activity is not specifically oriented toward them. Landings of the other species are less abundant ranging from 2 to 10 tons a year on the average. Local consumption of fish is difficult to appraise since every inhabitant of the atoll meets his needs himself. Morize (1984) had estimated that about 150 kg of fish per year and per person are likely to be consumed. Given the total population of Tikehau, an additional 40 metric tons of fish would be landed every year for subsistence. Species readily available to various simple gear (handline, spear) such as groupers, surgeonfish or parrotfish are probably the principal components of this secondary fishery. Temporal variations of the catch Although total landings are somewhat homogeneous from year to year (average value of 165 metric tons), with a slight upward trend (Table 11), the relative species abundance in the catches varies considerably. In 1985-86, about 14 tons of Lethrinus miniatus have been fished whereas more than 50 tons were caught the next year with the same fishing effort applied to the stock. At the same time, Epinephelus microdon yield changed from 5 to about 50 tons and that of Naso brevirostris dropped from 19 to 2 tons. These variations are extremes but in general, only a handful of minor species are equally harvested from year to year. For most of the target species, yield can double or conversely, be reduced by half from year to year without any predictive signs. However, the great number of species available to the traps tend to buffer large fluctuations in total catches by changes in recruited population levels of individual species. 18 Table 10: Check-list of species (italic) caught in Tikehau fish traps with indicative figures of their diet (P : piscivorous, I : invertebrate feeders, H : herbivorous) and indications on their relative abundance in the catches (* : low, erratic catches generally less than 1 % of the total catches; ** : medium abundance, species often fished but representing less than 5 % of the annual total catches ; *** : high abundance, species regularly catched accounting for more than 5 % of the total). Family Species Diet Harvest Holocentrid Sphyraenid Siganid Serranid Priacanthid Carangid Lutjanid Mullid Mugillid Chanid Lethrinid Chaetodontid Albulid Kyphosid Scarid Acanthurid Balistid Sargocentron spiniferum Myripristis sp. Sphyraena forsteri Siganus argenteus Epinephelus merra Epinephelus microdon Priacanthus cruentatus Alectis indicus Carangoides orthogrammus Caranx ignobilis Caranx lugubris Caranx melampygus Caranx sp. Decapterus macarellus Elagatis bipinnulata Scomberoides lysan Selar crumenophthalmus Lutjanus fulvus Lutjanus gibbus Mulloides flavolineatus Mulloides vanicolensis Parupeneus barberinus Upeneus vittatus Mugil cephalus Liza vaigiensis Chanos chanos Lethrinus mahsena Lethrinus miniatus Monotaxis grandoculis Chaetodon auriga Albula vulpes Kyphosus cinerascens Scarus gibbus Scarus sp. Acanthurus xanthopterus Ctenochaetus striatus Naso _ brevirostris Naso lituratus Naso unicornis Naso vlamingit Balistoides viridescens I ~ ~ — — ~ PO REO ee te commas ne oe OW a NOLES PEC ed ae — — ~ jooeces ecco ae LH 19 Table 11: Yearly total weight landed (kg) of the fourteen main species caught by Tikehau fish traps and yearly total (kg) including all species. Mean year calculated by averaging data of the four year. 83-84 84-85 85-86 86-87 Mean Lethrinus miniatus 34,812 29,923 13,961 50,983 32,419 Lutjanus gibbus 8,152 11,371 24,374 24,354 17,062 Caranx melampygus 24,357 21,332 10,213 11,214 16779 Selar 8,337 14,201 17,133 16,063 13,933 crumenophthalmus Epinephelus microdon 180 810 5,183 48,902 13,786 Lutjanus fulvus 11,226 15,962 13,050 7,694 11,983 Naso brevirostris 3,036 15,299 19,374 2,293 10,000 Mulloides sp. 9,593 8,506 11,066 5,359 8,631 Albula vulpes 12,292 7,889 6,391 5,099 7,918 Upeneus vittatus 9,454 882 6,206 1,085 4,406 Sphyraena forsteri 2,835 2,835 5,085 2,954 3,427 Acanthurus 2,085 6,229 307 1,661 2,270 xanthopterus Myripristis sp. 2,475 1,559 2,931 1,851 2,204 Decapterus pinnulatus m.d. 1,424 3,580 1,582 2,195 Other species 15,484 15,974 16,152 26,348 18,489 total 144,318 154,236 155,006 207,442 165,250 1000 900 800 700 600 C.P.U.E. 500 (Kg day-1) 400 300 200 100 (0) Jul. Aug. Sep. Oct.Nov.Dec. Jan. Feb.Mar.Apr.May Jun. YEAR Fig.9 : Temporal variations of Catch per Unit Effort (C.P.U.E.) over an average year in the fishery of Tikehau. 20 Fig. 9 shows fairly wide fluctuations in the overall monthly catch per unit effort through an average year. Yield ranges from 68 kg per day in July to 898 kg per day in November. The highest productivity of the fishery occurs from October through January and the lowest from April through August. Individual yields of the overwhelming majority of the target species follow these variations but maximal c.p.u.e. of a few species are reached at a different time of the year. Noteworthy is the example of Epinephelus microdon in which the presence on the fishing ground peaks in April. Behind these strong seasonal fluctuations, c.p.u.e.s have a clear relationship depending on the time of the lunar month. Yields of the target species noticeably increase the week prior to the new moon and drop around the full moon. Obviously, yields of the target species are strongly related to seasonal spawning aggregations in the vicinity of the pass. Biological sampling of landed fish carried out every month of the study confirmed that all fish trapped are adult fish, most of them having ripe gonads. Such spawning movements in other tropical areas are also well documented in numerous published observations reviewed by Johannes (1978). Thus temporal variations of c.p.u.e.s of the fishery of Tikehau would have a strong relationship with the time of the breeding period of the major components of the catches as emphasized by Caillart and Morize (1988). BIOLOGY AND ECOLOGY OF TARGET SPECIES The biology of the main species caught by fish traps in Tikehau has been studied. The overwhelming majority of the fish sampled was collected in the fishery landing. Additionally, some fish were collected by experimental fishing in the lagoon or on the outer slope using a handline or spear gun. The biological study presented hereafter is restricted to the seven major species : lethrinid Lethrinus miniatus, carangid Caranx melampygus, serranid Epinephelus microdon, lutjanids Lutjanus gibbus and Lutjanus fulvus and acanthurids Acanthurus xanthopterus and Naso brevirostris. Reproduction Reproductive patterns of the target species were followed throughout the year on a monthly basis. For all samples taken, gonosomatic indices (GSI) were calculated for individual males and/or females as GSI = 100xgonad wt/fish wt. Fig. 10 summarizes the observations gathered on the time of spawning of the target species. At Tikehau, fish typically have extended breeding seasons with more or less conspicuous seasonal peaks in breeding activity. For Lethrinus miniatus, spawning is virtually confined from September to December with most spawning through September. The snappers Lutjanus fulvus and Lutjanus gibbus appear to spawn between October and June with two seasonal peaks that occur in November and in March. The average GSI remain however at significant levels all year round suggesting that some individual spawning may occur at an odd time. The data for Caranx melampygus indicates that spawning occurs throughout the year with slight peaks in July, October and February. Lastly, spawning of Epinephelus microdon and Naso brevirostris is virtually confined to a short period of three months. The records for Epinephelus microdon show a maximum in the period between March and May with the greatest proportion of ripe fish found in April. The surgeonfish Naso brevirostris spawns between December and February with most spawning in December. For this last species, the time of spawning was confirmed by two additional methods : maturity stages assigned to female fish using a five stage scale and a study of frequency distributions of egg size within ovaries over the year (Caillart, 1988). Patterns in fecundity of Naso brevirostris were drawn from this last meaningful method. A female would spawn about 160,000 eggs, on the average, within a breeding season. Batch fecundity averaged over the complete breeding season, about 221 eggs g-! body weight, indicated that each female N. brevirostris must release its eggs in about three times, providing that discrete spawning occurs (Caillart, 1988). L. gibbus C. melampygus A. xanthopterus Fig. 10 : Summary of information on time of spawning of the target species of the fishery of Tikehau drawn from GSI variation study, and relationship with time of maximum catch per unit effort (cpue) over an average year (dark bars : breeding season, dotted bars : cpue). No data available to determine the breeding season of A. xanthopterus . For most species, occasional individual spawning is likely to happen in all months, but maximum activity takes place in the earlier months of the year. However, the sole study of GSI variations only gives general trends and is probably insufficient to accurately provide estimates of the occurrence of breeding seasons in the tropics. Table 12 : Fork-length at first reproduction (mm) of the target species of the fishery of Tikehau obtained from length-frequency data of the catches in fish traps. (* : length-frequency data inadequate to calculate length for both sexes ; *“ : relevant data available only for females). Species Male Female Lethrinus miniatus 410 370 Lutjanus gibbus 220 210 Caranx melampygus 270 250 Epinephelus microdon m.d. 310 ** Lutjanus fulvus 200 200 Acanthurus xanthopterus 320 * 320 * Naso _ brevirostris 260 220 Length at first reproduction was determined under the assumption that the relationship between fishery yields and spawning activity does exist. The first group in the length- frequency distributions of the catch is assumed to actually represent the earlier migrating spawner group (i.e. : fish newly recruited to the fishery). Therefore length at first reproduction was calculated as the length in which summed length-frequency reaches 50 % of the total number of fish in the first cohort (Table 12). 22 Growth Information on the age and growth of fishes is a central element in fishery management analysis. Common biological characteristics of fishes of Tikehau such as a missing seasonal growth and an extended breeding season throughout the year, have made growth rate determination difficult. Basically, three approaches to the determination of age and growth of the target species were attempted. These were 1) modal progression analysis in a time series of length-frequency histograms ; 2) tag-recapture study and 3) the aging of individuals on the basis of regular periodic (daily) markers in otoliths. The growth rate of fishes has been described by the Von Bertalanffy Growth formula (hereafter expressed VBGF) because it fits most of the data obtained on fish growth and it can be readily incorporated into models of stock assessment. The VBGF expression is : L(t) = Leo (1 - exp( -k(t-to)) where L(t) is the length at time t, Leo is the asymptotic length, k the rate at which the fish approaches the asymptotic length and to the origin of the growth curve. All length measurements presented herein are fork lengths in mm unless stated otherwise. Table 13 : VBGF growth parameter estimations for the main species caught in the fishery of Tikehau. Leo is given in mm, k and tg on a year basis. (6 : standard deviation of relevant parameter when available, Meth : method used ; 1 : modal progression analysis of length- frequency histograms, 2 : tag-recapture study and 3 : otolith microstructure examination), Leo 2) k OL, to Si Meth Lethrinus miniatus 560 110 0.42 032 049 1.09 1 Lutjanus gibbus 360 70 0.60 0.26 059 0.83 1 Lutjanus fulvus 280 _ 0.89 -- -0.05 — 3 Caranx melampygus 610 367 0.20 030 -180 150 1 Epinephelus microdon 610 _ 0.35 _ — — 2 Epinephelus microdon 690 301 0.31 0.03 022 0.08 1 Acanthurus xanthopterus 490 _— 0.30 _- -0.00 -— 1 Naso brevirostris (male) 380 _ 0.33 _ 0.39 — ut Naso brevirostris (female) 350 _— 0.26 _ -0.80 — 1 Length-frequency histograms were examined. A random length sample of the main target species was taken serially, whenever possible. For species in which spawning season is confined to a short period (Lethrinus miniatus, Lutjanus gibbus, Epinephelus microdon and Naso brevirostris), analysis was carried out under the assumption that cohorts are separated by a time interval of one year. The VBGF parameter estimations presented in Table 13 probably lack robustness but figures generated correspond to some extent to growth parameters reviewed by Munro and Williams (1985) and can be considered as reliable. Several limitations arise on the results presented on Caranx melampygus and Acanthurus xanthopterus because breeding seasons tend to be prolonged over several months and as a result, age classes are not readily separable from one another. In that case, mode discrimination involves a large part of subjectivity. 23 A tag-recapture study was undertaken on grouper Epinephelus microdon (Morize and Caillart, 1987). Between 1984 and 1987, over one thousand tags were released all over the lagoon. Most recoveries occurred within one month of tagging and very close to the point of release but there is a tendency for at least a part of the population to seasonally migrate toward the pass since a few fish tagged in various locations of the lagoon were recaptured in the vicinity of the pass during the breeding season (i.e. : April). For growth rate estimation purposes, all tagged fish were measured upon release and fishermen were asked to provide information on the length of fish recaptured. Out of the thousand tags released, only 47 tags recovered met this basic requirement. Data were fitted to the VBGF using the method of Fabens. The VBGF parameter estimations are presented in Table 13. Otoliths are structures that are commonly used to age tropical fishes (Panella, 1971). The relatively new finding that many fish deposit otolith growth increments with a daily periodicity appeared to offer a method of assessing age and growth with greater accuracy than was previously possible through other classical methods. Otolith microstructures of the target species of the fishery of Tikehau were examined (Caillart et al., 1986 ; Caillart, 1988) for Lethrinus miniatus, Lutjanus gibbus, Caranx melampygus and Naso brevirostris. Ages determinated through increment counts appeared to be obviously underestimated although the actual age-increment discrepancy has not been measured. Tetracycline injected into adult Epinephelus microdon reared for more than one year was used to verify the periodicity of increment deposition (Caillart and Morize, 1989). For this species held in captivity, one ring was laid down every two days on an average. If this result applies to Epinephelus microdon in their natural environment, aging fishes under the assumption that otolith increments are daily, would have lead to underestimate the actual age by a factor of two. In spite of all the limitations raised by the foregoing discussion, growth parameters of Lutjanus fulous were calculated by fitting the VBGF to the results of otolith increment counts because either the length-frequency histograms method or the tag-recapture operation failed to give results (Table 13). Table 14: Length (in mm) at age (in year) of the target species of the fishery of Tikehau during the exploited phase (data backcalculated with VBGF growth parameters presented in table 13). (*) : Data backcalculated with the tag-recapture VBGF, (**) : Data backcalculated with the modal progression analysis VBGF. Age L. Ib LE: € E. A. N. miniatus gibbus fulvus melampygus microdon xanthopterus brevirostris CDi Ge) male female 1 221 262 1d) 257 210 295 249 2 363 284 235 325 30799293 ZiE9) 400 304 251 352 S50r OU), 234 3 431 376 397 399 256 220 3.5 455 431 440 275 236 4 475 460 476 342 291 250 4.5 491 484 363 304 262 5 504 504 381 316 273 55 515 396 326 282 6 409 6.5 420 7 430 24 Lengths at age back calculated from the VBGF growth parameters are presented in Table 14. Only the portion of the growth curve covering the range of data used to establish the predictive equation was taken into account. Since this range of data corresponds with the exploited phase of the fishes, Table 14 gives insight into the duration of the phase. Certain patterns emerge pertaining to the main species and can be summarized as follows : the duration of the exploited phase is generally short ranging from three years (Naso brevirostris, Caranx melampygus) to four years (Lethrinus miniatus, Epinephelus microdon and Acanthurus xanthopterus). In the case of lutjanids, the vulnerability to fishing gear appears to last two years. Data furthermore suggest that fishes are fully recruited to the fishery at an average age of three years for acanthurids, and two years for the others. It is most likely that fishes disappear from the fishing ground due to a dramatic mortality rate since experimental fishing carried out in various locations of the lagoon and off the reef yielded a very few fish beyond the maximal size recorded in the catches. Caranx melampygus is however an exception. The adult population of this species shifts later in its life-cycle toward the pelagic environment of the outer slope, out of the reach of fishing gears. Length-weight relationships The relation of weight (W in g) to the fork length (Lf in mm) was calculated for the seven target species. The parameters a and b of the formula : W =aLb are listed in Table 15 (Morize, unpublished data). For all species under investigation, samples of a few hundred fish taken in the catches were used to derive the regression equations. Correlation coefficients r obtained ranged from 0.95 to 0.99. ASSESSMENT OF THE FISHERY OF TIKEHAU The problem of stock assessment in the fishery of Tikehau mostly relates to the fact that it is based upon at least fourteen species in which none of them is overwhelmingly dominant. Given the set of data obtained on the fishery (catch statistics, common biological parameters of individual species), two techniques are available for appraising potential harvests. Firstly, assessment can be based upon a comparison with known harvests per unit area taken by fisheries of a similar environment. Secondly, analytical models requiring reliable estimates of either biological or fishery parameters can be used in order to model the response of the stock to exploitation. Table 15 : Length-weight relationship for the main species caught in Tikehau fish- traps (a and b, parameters of the equation W=aL> where W = weight in g, L = fork length in mm). a (.10-) b Lethrinus miniatus 3.4 2.8 Lutjanus gibbus 21 3.0 Lutjanus fulvus 11.0 2.8 Caranx melampygus 6.4 2.8 Epinephelus microdon 0.5 3.2 Acanthurus xanthopterus 9.3 2.8 Naso brevirostris 3.8 2.8 25 Yield per unit area On the average, 200 metric tons of finfishes per year are caught in the main fishery of Tikehau. Additionally, 40 metric tons are taken for subsistence and another 40 tons are fished by occasional fishermen for commercial purposes (Morize, 1984 ; Morize, 1985). That is, the fishery of Tikehau produces an average of 280 tons per year (table 16). The area covered by the lagoon of the Tikehau atoll is about 420 km2 and the annual harvest per unit area of 0.7 tons . km-2. Marshall (1980) pointed out that a finfish harvest of 3 to 5 tons . km-2 may be upheld as a generalization for the potential fishery yields of coral reefs and adjacent shallow water environments. Although records presented in Table 16 fall far below the suggested potential, data are somewhat homogeneous, ranging from 0.6 tons . km-2 in the fishery of Ontong Java to 1.3 tons . km-2 in the fishery of Mataiva with the noticeable exception of Rangiroa where fishery harvests reach only 0.2 tons . km-2. However a limitation arises to permit the comparison of the different harvests per unit area recorded. Table 16: Harvests per unit area for a selection of exploited coral atolls (for the Tuamotu coral atolls, groups included in catch statistics are only finfishes. For Kapingamarangi and Ontong Java, composition of the catches is unknown), Total catch Lagoon area Harvest per (metric tons) (square kilometers) unit area Ref. (Tons/km2) Kapingamarangi 280 400 0.7 i (Caroline islands) Ontong Java 122 79 0.6 2 (Solomon islands) Rangiroa (Tuamotu) 350 1600 0.2 3 Kaukura (Tuamotu) 500 500 1.0 3 Mataiva (Tuamotu) 63 50 1.3 3 Tikehau (Tuamotu) 280 420 0.7 4 Reference 1- Stevenson and Marshall (1974) ; 2 - Munro and Williams (1985) 3 - Galzin et al. (1989) ; 4 - Caillart (1988) As a reef fishery is generally a patchwork of coral reef patches (which are highly productive) and sandy bottoms (which is not that productive) ; the yield per unit area can very much depend upon the area and the percentage of area that is actually covered by hard coral substrate. Some fishery records like these of Rangiroa cover a large area, only part of which is actually covered by coral, whereas other records of fish yield apply to very small areas like Mataiva or Ontong Java where a hard substrate coverage is much greater. Moreover the potential fish yield from a given area cannot be inferred from sole catch records without even a rough reference to the fishing effort. In Rangiroa and Mataiva, the level of exploitation applied to the stock is low with regard to fishing effort in Tikehau or Kaukura. Information on yield assessment and management in the fishery of Tikehau can be drawn from the comparison with the neighboring atoll of Kaukura. These two atolls have a comparable surface and morphology. In Tikehau, the fishery is based on bottom fixed fish traps all located in the vicinity of the pass. Yield relies on the behavior of species most prone to migrate for spawning. These fish are primarily carnivorous species as indicated by the specific composition of the catches. In Kaukura, bottom fixed fish traps are set not only in the vicinity of the pass but also all around the atoll rim, on the shallow inner reef flat. 26 Species caught are for a great percentage non territorial herbivorous species which wander to seek for food (Stein, in Galzin et al., 1989). So, higher yields in Tikehau could probably be achieved by setting traps in various locations of the lagoon which in turn would probably exploit the food chain more efficiently. The total harvest of the Tikehau fishery could also be increased by diversifying the fishing gears, and setting classical bottom free fish traps around the numerous coral knolls scattered in the lagoon. Although we believe that it would be quite impossible to reach the potential yield suggested by Marshall (1980) (i.e. : 3 to 5 tons . km-2), it would be at least possible to attain a harvest of 1 ton . km-2 recorded at Kaukura. This would result in a substantial increase of the catch of about 140 tons. If this simple but nevertheless useful approach can be used to set a likely estimate of the potential fish yield of Tikehau, it is obvious that more thorough evaluations must be undertaken in order to focus management issues not only on optimum yield but also on preferred species. Analytical assessment models Analytical assessment models have been widely used in temperate water fisheries but they have been applied to coral reef fisheries in a limited number of cases. If these models cannot take into account the numerous and intricate relationships between all the components of the multi-species fishery, they are nevertheless of great value in giving an insight into the state of the fishery. There were two means used to provide estimates of the status of the fishery of Tikehau. One mean was a length converted catch-curve analysis (in Ricker, 1980). The other mean was to use yield per recruit estimates in a length structured model in which fishing mortality vector (F) is obtained from a length cohort analysis (Jones, 1974). No adequate data sets on Tikehau fish stocks exist for an accurate determination of natural mortality (M). This parameter was estimated by two empirical formulas (Hoenig, 1984 and Pauly, 1980) that provided rough estimates of the value of M (Table 17). The real value of M is expected to lie in between these two estimates. The specific exploitation rate E is given by : : F ~ F+M E where F is the fishing mortality. E estimated through length-converted catch curve analysis is found greater than 0.5 for Lutjanus gibbus, L. fuluus, Caranx melampygus and Epinephelus microdon, and less than 0.5 for Lethrinus miniatus, Acanthurus xanthopterus and Naso brevirostris. Gulland (1973) pointed out that a value of 0.5 of the exploitation rate can be roughly set as a limit below which a fish stock is lightly exploited and over which over- fishing may occur. Table 17 : A range of values of natural mortality M (yr!) chosen for Tikehau target species. Mmin is given by Hoenig (1984) empirical formula, Mmax by Pauly (1980) equation. Species M min M max Lethrinus miniatus 0.43 0.66 Lutjanus gibbus 0.57 0.96 Lutjanus fulvus 0.46 0.88 Caranx melampygus 0.43 0.72 Epinephelus microdon 0.61 0.88 Acanthurus xanthopterus 0.43 0.72 Naso _ brevirostris 0.60 0.80 27 Yield per recruit model results listed in Table 18 are strongly related to the estimate of M chosen and have considerable different responses to F variations with respect to the species under investigation. For Lethrinus miniatus, Acanthurus xanthopterus and Naso brevirostris, a substantial increase of yield per recruit (more than 10% on the average) can be achieved if the fishing mortality vector is 50% or 100% higher. The snapper Lutjanus gibbus and L. fulvus yield per recruit estimates appears to be poorly increased (5% on the average) when fishing mortality vector increases. Lastly, yield per recruit estimates of Epinephelus microdon and Caranx melampygus do not significantly increase and can even decrease if an attempt to increase F is made. Table 18 : Range of yield per recruit variations of the target species of Tikehau fishery (in % of present yield per recruit) in response to variations of fishing effort (uF : Fishing mortality coefficient, lowest value of yield per recruit correspond to the highest natural mortality figure). Species p= 0:5 el WEES uF = 2 Lethrinus miniatus -30 / -20 0 +20 / +8 +27 / +10 Lutjanus gibbus -30 / -15 0 +12 / +3 +20 / +3 Lutjanus fulvus -25 / -12 0) +10 / +2 +15 / +2 Caranx melampygus -25 / -8 0 +10 /0 +20 / -1 Epinephelus microdon -15 / 6 0 +6 /-1 +8 / -2 Acanthurus xanthopterus -40 / -30 0 +20 / +10 +40 / +20 Naso brevirostris 0 +19 / +26 +32 / +20 According to the foregoing results, the Tikehau fishery appears to be well fitted to carnivorous fish stock exploitation. The evidence from these analytical models suggests that Tikehau fish stocks are being fished at or near the Maximum Sustainable Yield (MSY). No major change in the direction of the present trap fishing strategy (increases or decreases in effort) is justifiable, although yield per recruit of certain species (emperors, surgeonfishes and snappers to a lesser extent) could be improved by a moderate increase of fishing effort. And it is unlikely that the grouper and jack fisheries could tolerate a heavy effort increase. Higher harvests of carnivorous species could probably be achieved by using more selective fishing gears. For instance, the abundant stock of Lethrinus miniatus could provide substantial additional catches if handlines were more heavily used when the fish are abundant in the pass and hence, readily available. It has been mentioned that the herbivorous fish stock at Tikehau is very lightly exploited. The principal management issue would probably be to orient fishing pressure toward this part of the resource by setting traps on shallow areas all around the atoll rim where availability of herbivorous species is greater. CONCLUSION A total of 276 species belonging to 47 families have been recorded on the Tikehau atoll (Appendix 1). The real number of species is obviously under-estimated since rotenone poisonning was not used in all sites, and only one transect was regularly studied on the outer slope. The number of species censused in the lagoon was 167, 39 in the pass and 180 on the outer slope. 28 Only 17 species (6.2% of the total richness species) were encountered in the three environments : Sargocentron spiniferum, Epinephelus merra, Epinephelus microdon, Caranx melampygus, Lutjanus gibbus, Lutjanus fulvus, Lethrinus miniatus, Monotaxis grandoculis, Mulloides vanicolensis, Chaetodon auriga, Scarus gibbus, Acanthurus xanthopterus, Ctenochaetus striatus, Naso lituratus, Naso unicornis, Naso vlamingti and Balistoides viridescens. An unusual paucity of Carcharhinidae, Synodontidae, Apogonidae, Mugilidae, Sphyraenidae, Caesionidae and Tetraodontidae was noted while fish of the families of Holocentridae, Serranidae, Carangidae, Lutjanidae, Lethrinidae, Mullidae, Chaetodontidae, Pomacanthidae, Pomacentridae, Labridae, Scaridae, Acanthuridae and Balistidae were abundant. A key question in fishery management is the correspondence between adult stock size and the number of each new cohort reaching the mean size of capture by the fishing gear. Recruitment to the fishery is preceded by a pre-recruit phase from birth to recruitment to the ecosystem and followed by a post recruit phase consisting of a pre-exploited phase. No study of larval recruitment was carried out at Tikehau though the knowledge of this part of the life-cycle is critical for understanding the dynamics of reef fish populations. Recruitment processes in coral reef fishes are however well documented (reviews in Munro and Williams, 1985 ; Richards and Lindeman, 1987) and much of the findings can apply to Tikehau. Most reef fishes spawn externally in the water column above hard bottom structures. Off-shore larval dispersal is thought to be an evolutionary response to intense predation pressure in the adult habitat (Johannes, 1978). Fish community studies at Tikehau suggest that, adult fishes of various species gather off or in the pass to release their offspring in oceanic water. Larvae or fertilized eggs subsequently undergo oceanic advection and diffusion and juveniles enter the lagoon through shallow channels of the eastern coast. Most coral reef fishes characteristically present a two part life-cycle ; a pelagic larval phase during which extensive dispersal is possible and a relatively site-attached phase during in which movements are somewhat restricted. According to relevant data presented by Brothers et al. (1983), the duration of the pelagic stage of the main families exploited in Tikehau is estimated to range from about one month (Lethrinidae, Lutjanidae) to over three months in the case of Naso sp. (Acanthuridae). Absolute survivalship during planktonic life stages is a function of highly complex interactions among predation, oceanographic processes, growth and food availability. Mortality rates through this phase are subject to tremendous variations which considerably affect the availability of recruits to the atoll fish community. Although of a lesser order of magnitude, additional losses in subsequent post-settlement life due to innapropriate habitat and predation can in turn impact the number of recruit to the fishery. Variations in recruitment can also contribute to significant shifts in species composition within the exploited stock as it does occur in Tikehau. Knowledge on the extent of fish population exchange between islands through the pelagic phase is of particular importance to effectively manage a fishery. The management strategy will vary greatly depending on the extent to which recruitment to the atoll is derived from within the fished population or is spawned outside the system. Due to the close-spacing pattern of the atoll of the Tuamotu archipelago, it might be expected that the stocks of species having a long pelagic larval stage occurring in a given atoll may be recruited from parent stocks living in areas further upstream. If the exploited stock of Tikehau is recruited largely from atolls located upstream like Rangiroa and Arutua, regulations for the conservation of the spawning stock will be ineffective and will be of benefit only to islands lying downstream (Mataiva). We have yet insufficient information to determine any general patterns, but there is an urgent need for further studies aiming to determine the potential limits of stock exchanges between atolls and the unit stock of a given species. 29 ACKNOWLEDGEMENTS Authors are deeply indebted to the support of the Tikehaun people. We are very grateful to fishermen, for without their help, nothing would have been possible. Literature cited BAGNIS (R.), GALZIN (R.), BENNETT JJ.) - 1979 - Poissons de Takapoto. J. Soc. Océanistes, 62 (35) : 69-74 BARANS (C.A.), BORTONE (S.A.), Eds. - 1983 - The visual assessment of fish populations in the Southeastern United States, 1982 Workshop, South Carolina Sea Grant Consortium, Techn. Rep., 1:52 p BELL JJ.), GALZIN (R.) - 1984 - Influence of live coral cover on coral reef fish communities. Mar. Ecol. Prog. Ser., 15 : 265-274 BLANCHET (G.), CAILLAUD (L.), PAOAAFAITE (J.) - 1985 - Un aspect de la péche artisanale en Polynésie Frangaise, les piéges 4 poissons de Tikehau. ORSTOM Tahiti, Notes et Doc. Océanogr., 25 : 116 p BROTHERS (E.B.), WILLIAMS (D.McB.), SALE (P.F.) - 1983 - Length of larval life in twelve families of fishes at "One tree lagoon", Great Barrier Reef, Australia. Mar. Biol., 76 : 319-324 CAILLART (B.) - 1988 - Etude d'une pécherie artisanale de l'archipel des Tuamotu (Polynésie Frangaise). Biologie, ethologie et dynamique des populations d'une espéce caractéristique : Naso brevirostris (Poissons-Acanthuridae). ORSTOM Tahiti, Notes et Doc. Océanogr., 38 : 235 p CAILLART (B.), FRANC DE FERRIERE (M.), MORIZE (E.) - 1986 - Croissance de deux espéces du lagon, Lethrinus miniatus (Schneider) et Lutjanus fuluus (Schneider), évaluée par la lecture des otolithes. in : Contribution a l'étude de l'atoll de Tikehau, II]. ORSTOM Tahiti, Notes et doc. Océanogr., 30 : 1-44 CAILLART (B.), MORIZE (E.) - 1986 - La production de la pécherie de l’atoll en 1985. In : Contribution a l'étude de l'atoll de Tikehau, III. ORSTOM Tahiti, Notes et Doc. Océanogr., 30 : 45-71 CAILLART (B.), MORIZE (E.) - 1988 - Biology and reproduction strategy of Naso brevirostris and relationships with fishery yields. Workshop on Pacific Inshore Fishery Resource, Noumea, New Caledonia, March 1988 : 10 p CAILLART (B.), MORIZE (E.) - 1989 - Etude du rythme de dépdét des microstries sur les otolithes d'un serranidae tropical, Epinephelus microdon (Bleeker) a l'aide d'un marqueur fluorescent : l'oxytetracycline. Aquat. Living Resour., 2 : 255-261 FABENS (A J.) - 1965 - Properties and fitting of the Von Bertalanffy growth curve. Growth, 29 : 265-289 FAURE (G.), LABOUTE (P.) - 1984 - Formations récifales de Tikehau : I - Définition des unités récifales et distribution des principaux peuplements de Scléractiniaires.In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 108-136 GALZIN (R.) - 1985 - Ecologie des poissons récifaux de Polynésie francaise. Thése Doc. és-Sciences, Univ. Montpellier, 195 p GALZIN (R.) - 1987 - Structure of fish communities of French Polynesian coral reefs. 1- Spatial scales. Mar. Ecol. Prog. Ser., 41 : 129-136 30 GALZIN (R.), BAGNIS (R.), BENNETT VJ.) - 1983 - Premier inventaire et distribution de la faune ichtyologique de I'atoll de Scilly (Polynésie frangaise). J. Soc. Océanistes, 77 (34) : 17-28 GALZIN (R.), BELL (J.D)., LEFEVRE (A). - 1990 - Etude interannuelle du peuplement ichtyologique du lagon de I'atoll de Mataiva en Polynésie francaise. Cybium, 14 (4) : 313-322 GALZIN (R.), MORIZE (E.), STEIN (A.), CONTE (E.) - 1989 - Dégradations naturelles et anthropiques en zones cotiéres intertropicales et répercussions possibles sur l'économie des péches : le cas des récifs coralliens. Action CORDET, Rapport antenne EPHE-Museum, 1989, RA 32: 135 p GBRMPA - 1978 - Great Barrier Reef Marine Park Authority workshop on reef fish assessment and monitoring. Workshop Series 1, Heron Island, Australia : 64 p GULLAND JJ.A.) - 1973 - Fish stock assessment. A manual of basic methods. J. Wiley & Sons, Chichester : 223 p HARMELIN-VIVIEN (M.L.) - 1984 - Distribution quantitative des poissons herbivores dans les formations coralliennes de Tikehau (Polynésie frangaise). In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Frangaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 81-107 HARMELIN-VIVIEN (M.L.) - 1985 - Tikehau atoll, Tuamotu archipelago. Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 1 : 211-266 HARMELIN-VIVIEN (M.L.), LABOUTE (P.) - 1986 - Catastrophic impact of hurricanes on atoll outer reef slopes in the Tuamotu (French Polynesia). Coral Reefs, 5 : 55-62 HARMELIN-VIVIEN (M.L.) and al. (11 authors) - 1985 - Evaluation visuelle des peuplements et populations de poissons. Méthodes et problemes. Rev. Ecol. (Terre Vie), 40 : 467-539 HOENIG (J.M.) - 1984 - Empirical use of longevity data to estimate mortality rates. Fish. Bull., 82 (1) : 898-902 JOHANNES (R.E.) - 1978 - Reproductive strategies of coastal marine fishes in the tropics. Env. Biol. Fish., 3 (1) : 65-84 JONES (R.) - 1974 - Assessing the long term effects of changes in fishing effort and mesh size from length composition data. CIEM, CM 1974/F : 33 p MARSHALL (N.) - 1980 - Fishery yields of coral reefs and adjacent shallow-water environments. In : Stock assessment for tropical fisheries, ROEDEL (P.) & SAILA (S.B.) eds, Proc. Int. Work. held in Kingston (P.M.) : 103-109 MORIZE (E.) - 1984 - Contribution a l'étude d'une pécherie artisanale et de la dynamique des populations des principales espéces exploitées. In : L’atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise), premiers résultats. ORSTOM Tahiti, Notes et Doc. Océanogr., 22 : 35-80 MORIZE (E.) - 1985 - Etude d'une pécherie artisanale de l'atoll de Tikehau (Archipel des Tuamotu, Polynésie Francaise). Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 5 : 501-506 MORIZE (E.), CAILLART (B.) - 1987 - Modélisation de la croissance de Epinephelus microdon (Bleeker) obtenue a partir de marquage. In : Contribution a l'étude de I'atoll de Tikehau : IV. ORSTOM Tahiti, Notes et Doc. Océanogr., 35 : 101-114 MORIZE (E.), GALZIN (R.), HARMELIN-VIVIEN (M.L.), ARNAUDIN (H.) - 1990 - Organisation spatiale du peuplement ichtyologique dans le lagon de l’atoll de Tikehau (Polynésie frangaise). ORSTOM Tahiti, Notes et Doc. Océanogr., 40 : 1-44 31 MUNRO JJ.L.), WILLIAMS (D.McB.) - 1985 - Assessment and management of coral reef fisheries. Proceeding of the Fifth International Coral Reef Congress, Tahiti, 1985, 4 : 543-581 PANNELLA (G.) - 1971 - Fish otoliths : daily growth layers and periodical patterns. Science, 173 : 1124-1127 PAULY (D.) - 1980 - On the interrelationships between natural mortality, growth parameters and mean environmental temperature in 175 fish stocks. J. cons. Cons. Int. Explor. Mer, 39 : 175-192 RICHARDS (W,J.), LINDEMAN (K.C.) - 1987 - Recruitment dynamics of reef fishes : planktonic processes, settlement and demersal ecologies, and fishery analysis. Bull. Mar. Sci., 41 (2) : 392-410 RICKER (W.E.) - 1980 - Calcul et interprétation des statistiques biologiques des populations de poissons. Bull. Res. Bd. Can., 191 F : 409 p ST. JOHN J.), RUSS (G.R.), GLADSTONE (W.) - 1990 - Accuracy and bias of visual estimates of numbers, size structure and biomass of a coral reef fish. Mar. Ecol. Prog. Ser., 64 : 253 - 262 STEVENSON (D.K.), MARSHALL (N.) - 1974 - Generalisation on the fisheries potential of coral reefs and adjacent shallow water environments. Proceeding of the Second International Coral Reef Symposium, Manilla, 1 : 147-156 32 Appendix 1 : Check-list of the fishes of Tikehau atoll (L : lagoon, P : pass, O : ocean) CARCHARHINIDAE Carcharhinus melanopterus (Quoy et Gaimard, 1824) JE, ALBULIDAE Albula oulpes (Linné, 1758) P CHANIDAE Chanos chanos (Forsskal, 1775) 12 MURAENIDAE Echidna polyzona (Richardson, 1844) Gymnothorax buroensis (Bleeker, 1857) Gymnothorax fimbriatus (Bennett, 1831) Gymnothorax javanicus (Bleeker, 1859) Gymnothorax margaritophorus Bleeker, 1864 Gymnothorax zonipectis Seale, 1906 Gymnothorax sp.3 Gymnothorax sp. 16 Gymnothorax sp. 18 Uropterygius xanthopterus Bleeker, 1859 CONGRIDAE Conger cinereus Riippell, 1828 OPHICHTHIDAE Leiuranus semicinctus (Lay and Bennet, 1839) Muraenichthys macropterus Bleeker, 1857 ATHERINIDAE Atherinidae sp. SYNODONTIDAE Saurida gracilis (Quoy et Gaimard, 1824) Synodus variegatus (Lacepéde, 1803) ANTENNARIIDAE Antennarius sp. (juv.) OPHIDIIDAE Brotula multibarbata Temminck and Schlegel, 1846 HEMIRAMPHIDAE Hyporhamphus acutus (Gtnther, 1871) HOLOCENTRIDAE Myripristis kuntee Valenciennes, 1831 Myripristis murdjan Forsskal, 1775 Myripristis pralinia Cuvier, 1829 Myripristis violacea Bleeker, 1851 Myripristis sp. Neoniphon argenteus (Valenciennes, 1831) Neoniphon opercularis (Valenciennes, 1831) Neoniphon sammara (Forsskal, 1775) Sargocentron caudimaculatum (Ruppell, 1838) Sargocentron diadema (Lacepéde, 1802) Sargocentron microstoma (Gtinther, 1859) Sargocentron spiniferum (Forsskal, 1775) AULOSTOMIDAE Aulostomus chinensis (Linné, 1766) FISTULARIIDAE Fistularia commersonii (Ruppell, 1838) SYNGNATHIDAE Corythoichthys flavofasciatus Ruppel, 1838 L SCORPAENIDAE Scorpaenodes parvipinnis (Garrett, 1863) L SERRANIDAE Anthias lori Randall and Lubbock, 1976 Anthias olivaceus Randall and Mc Cosker, 1892 Anthias pascalus (Jordan and Tanaka, 1927) Anthias squamipinnis Peters, 1855 Cephalopholis argus (Bloch and Schneider, 1801) L Cephalopholis urodelus (Bloch and Schneider, 1801) Epinephelus fasciatus (Forsskal, 1775) Epinephelus hexagonatus (Bloch and Schneider, 1801) Epinephelus merra Bloch, 1793 Epinephelus microdon (Bleeker, 1856) Epinephelus socialis (Giinther, 1873) Epinephelus sp. Ee GS. Ro Ree Se oe ae Ce eae Ss ee ted eit ait) ©) ey SV e)(C) CK) 6) COE! © cay SiGe a) OOO000000000 Appendix 1 (cont'd) Gracila albomarginata (Fowler and Bean, 1930) Grammistes sexlineatus (Thunberg, 1792) Pseudogramma bilinearis (Schultz, 1943) Pseudogramma polyacantha (Bleeker, 1856) Variola louti (Forsskal, 1775) KUHLIIDAE Kuhlia marginata (Cuvier, 1829) PRIACANTHIDAE Priacanthus cruentatus (Lacepéde, 1801) CIRRHITIDAE Paracirrhites arcatus (Cuvier, 1829) Paracirrhites forsteri (Bloch and Schneider, 1801) Paracirrhites hemistictus (Gunther, 1874) APOGONIDAE Apogon angustatus (Smith and Radcliffe, 1911) Apogon coccineus Ruppell, 1838 Apogon fraenatus Valenciennes, 1832 Apogonichthys ocellatus (Weber, 1913) Cheilodipterus quinquelineatus Cuvier, 1828 Fowleria aurita (Valenciennes, 1831) Fowleria marmorata (Alleyne and Macleay, 1876) Ostorhynchus savayensis (Gtinther, 1871) Pristiapogon snyderi Smith, 1961 Pseudamia gelatinosa Smith, 1955 MUGILIDAE Liza vaigiensis (Quoy et Gaimard, 1825) Mugil cephalus (Linné, 1758) SPHYRAENIDAE Sphyraena forsteri Cuvier, 1829 ECHENEIDIDAE Echeneis naucrates Linné, 1758 CARANGIDAE Alectis indicus (Ruippel, 1830) Carangoides orthogrammus (Jordan and Gilbert, 1881) Caranx ignobilis (Forsskal, 1775) Caranx lugubris Poey, 1860 Caranx melampygus (Cuvier, 1833) Caranx sp. Decapterus macarellus (Valenciennes, 1833) Elagatis bipinnulata (Quoy et Gaimard, 1825) Scomberoides lysan (Forsskal, 1775) Selar crumenophthalmus (Bloch, 1793) LUTJANIDAE Aphareus furca (Lacepéde, 1801) Aprion virescens Valenciennes, 1830 Lutjanus bohar (Forsskal, 1775) Lutjanus fulous (Bloch and Schneider, 1801) Lutjanus gibbus (Forsskal, 1775) Lutjanus kasmira (Forsskal, 1775) Lutjanus monostigmus (Cuvier, 1828) LETHRINIDAE Gnathodentex aureolineatus (Lacepéde, 1802) Lethrinus elongatus Valenciennes, 1830 Lethrinus mahsena (Forsskal, 1775) Lethrinus miniatus Smith, 1959 Lethrinus variegatus Ehrenberg, 1830 Lethrinus xanthochilus Klunzinger, 1870 Monotaxis grandoculis (Forsskal, 1775) MULLIDAE Mulloides flavolineatus (Lacepéde, 1801) Mulloides vanicolensis (Valenciennes, 1831) Parupeneus barberinus (Lacepéde, 1801) Parupeneus bifasciatus (Lacepéde, 1801) Parupeneus ciliatus (Lacepéde, 1801) Parupeneus multifasciatus (Quoy et Gaimard, 1825) Parupeneus porphyreus (Jenkins, 1900) Upeneus vittatus (Forskall, 1775) Ge Ge ee Siesta ile "eS lS sa fa tS i Beet i Ges FG te tt OSvy iVUNTT Ee} ige) eo} tae) steele ire} 33 (e)(e)(e) Ee) loeye) O 00 0 00 90000000 OO000 O 34 Appendix 1 (cont'd) PEMPHERIDAE Pempheris oualensis Cuvier, 1831 KYPHOSIDAE Kyphosus cinerascens (Forsskal, 1775) CHAETODONTIDAE Chaetodon auriga Forsskal, 1775 Chaetodon bennetti Cuvier, 1831 Chaetodon citrinellus Cuvier, 1831 Chaetodon ephippium Cuvier, 1831 Chaetodon lunula (Lacepéde, 1802) Chaetodon ornatissimus Cuvier, 1831 Chaetodon pelewensis Kner, 1868 Chaetodon quadrimaculatus Gray, 1831 Chaetodon reticulatus Cuvier, 1831 Chaetodon trifasciatus Mungo Park, 1797 Chaetodon ulietensis Cuvier, 1831 Chaetodon unimaculatus Bloch, 1787 Forcipiger flavissimus Jordan and Mc Gregor, 1898 Forcipiger longirostris (Broussonet, 1782) Hemitaurichthys polylepis (Bleeker, 1857) Hemitaurichthys zoster (Bennett, 1831) Heniochus acuminatus (Linné, 1758) Heniochus chrysostomus Cuvier, 1831 Heniochus monoceros Cuvier, 1831 POMACANTHIDAE Centropyge flavissimus (Cuvier, 1831) Centropyge loriculus (Giinther, 1874) Pomacanthus imperator (Bloch, 1787) Pygoplites diacanthus (Boddaert, 1772) POMACENTRIDAE Abudefduf sexfasciatus (Lacepéde, 1801) Abudefduf sordidus (Forsskal, 1775) Amphiprion chrysopterus Cuvier, 1830 Chromis iomelas Jordan and Seale, 1906 Chromis margaritifer Fowler, 1946 Chromis vanderbilti (Fowler, 1941) Chromis viridis (Cuvier, 1830) Chromis xanthura (Bleeker, 1854) Chrysiptera glauca (Cuvier, 1830) Chrysiptera leucopoma (Lesson, 1830) Dascyllus aruanus (Linné, 1758) Dascyllus flavicaudus Randall et Allen, 1977 Dascyllus trimaculatus (Riippell, 1828) Plectroglyphidodon dickii (Liénard, 1839) Plectroglyphidodon johnstonianus Fowler and Ball, 1924 Pomacentrus fuscidorsalis Allen and Randall, 1974 Pomacentrus pavo (Bloch, 1787) Stegastes albofasciatus (Schlegel and Miiller, 1839-44) Stegastes aureus (Fowler, 1927) Stegastes nigricans (Lacepéde, 1803) LABRIDAE Anampses caeruleopunctatus Rtippel, 1828 Bodianus axillaris (Bennett, 1831) Bodianus loxozonus (Snyder, 1908) Cheilinus chlorourus (Bloch, 1791) Cheilinus trilobatus (Lacepéde, 1801) Cheilinus undulatus Ruppel, 1835 Cirrhilabrus exquisitus Smith, 1957 Cirrhilabrus scottorum Randall and Pyle, 1856 Coris aygula Lacepéde, 1801 Coris gaimard (Quoy et Gaimard, 1824) Cymolutes praetextatus (Quoy et Gaimard, 1824) Epibulus insidiator (Pallas, 1770) Gomphosus varius Lacepéde, 1801 Halichoeres hortulanus (Lacepéde, 1801) abisotisle Bile Eur Bute ict Rite Cpe Gea ee se eee C000 ODODODGCNDOCNOOCOOCOCC0CNO OO oe) Oo 0000 OO0000 ‘e) O00 O00000 love) Appendix 1 (cont'd) Halichoeres melasmapomus Randall, 1980 Halichoeres trimaculatus (Quoy et Gaimard, 1834) Hemigymnus fasciatus (Bloch, 1792) Labridae sp. (juv.) Labridae sp. 8 (juv.) Labroides bicolor Fowler and Bean, 1928 Labroides dimidiatus (Valenciennes, 1839) Novaculichthys taeniourus (Lacepéde, 1801) Pseudocheilinus hexataenia (Bleeker, 1857) Pseudocheilinus octotaenia Jenkins, 1900 Stethojulis bandanensis (Bleeker, 1851) Stethojulis strigiventer Bennett, 1832 Thalassoma amblycephalum (Bleeker, 1856) Thalassoma hardwicke (Bennett, 1830). Thalassoma purpureum (Forsskal, 1775) Thalassoma quinquevittatum (Lay and Bennett, 1839) Thalassoma trilobatum (Lacepéde, 1801) Wetmorella ocellata Schultz and Marshall, 1954 SCARIDAE Calotomus carolinus (Valenciennes, 1839) Cetoscarus bicolor (Ruppell, 1829) Hipposcarus harid (Forsskal, 1775) Hipposcarus longiceps (Valenciennes, 1839) Leptoscarus vaigiensis (Quoy et Gaimard, 1824) Scarus altipinnis Steindachner, 1879 Scarus brevifilis (Giinther, 1909) Scarus festious , Valenciennes, 1840 Scarus forsteri (Bleeker, 1861) Scarus frenatus Lacepéde, 1802 Scarus frontalis Valenciennes, 1839 Scarus ghobban Forsskal, 1775 Scarus gibbus Rtippell, 1828 Scarus globiceps Valenciennes, 1840 Scarus niger Forsskal, 1775 Scarus oviceps Valenciennes, 1839 Scarus psittacus Forsskal, 1775 Scarus rubroviolaceus Bleeker, 1849 Scarus schlegeli Bleeker, 1861 Scarus sordidus Forsskal, 1775 Scarus sp. rayé (juv.) Scarus sp. gris (juv.) Scarus sp. marron (juv.) Scarus sp. parc Scarus sp. vert (juv.) BLENNIIDAE Enchelyurus ater (Gtinther, 1877) Istiblennius periophthalmus (Valenciennes, 1836) Plagiotremus tapeinosoma (Bleeker, 1857) CALLIONYMIDAE Callionymus simplicicornis Valenciennes, 1837 GOBIIDAE Amblygobius phalaena (Valenciennes, 1837) Asterropteryx ensiferus (Bleeker, 1874) Asterropteryx semipunctatus (Ruppell, 1830) Callogobius sclateri (Steindachner, 1880) Eviota afelei Jordan and Seale, 1906 Eviota sp. Fusigobius neophytus (Gunther, 1877) Gnatholepis cauerensis (Bleeker, 1853) Gobiidae sp. 5 Nemateleotris magnifica Fowler, 1938 Ptereleotris evides (Jordan and Hubbs, 1925) Quisquilius eugenius (Valenciennes, 1836) ISTIOPHORIDAE Istiophorus platypterus (Shaw and Nodder, 1792) ZANCLIDAE Zanclus cornutus (Linné, 1758) Cee Be ts ed es eee BE ea eg Den te sc cm = BG Ge te ee ite: 9 tit feign fall ete SiS Haiti 35 okeye) C000 0 000000 O ODODDODDODOOOCOCCOCOCC0C0C00O OO 36 Appendix 1 (cont'd) ACANTHURIDAE Acanthurus achilles Shaw, 1803 Acanthurus bleekeri Giinther, 1861 Acanthurus glaucopareius Cuvier, 1829 Acanthurus guttatus Bloch and Schneider, 1801 Acanthurus leucopareius (Jenkins, 1903) Acanthurus lineatus (Linné, 1758) Acanthurus mata (Cuvier, 1829) Acanthurus nigricauda Duncker and Mohr, 1929 Acanthurus nigrofuscus (Forsskal, 1775) Acanthurus nigroris (Valenciennes, 1835) Acanthurus nubilus (Fowler and Bean, 1929) Acanthurus olivaceus Bloch and Schneider, 1801 Acanthurus pyroferus Kittlitz, 1834 Acanthurus thompsoni (Fowler, 1923) Acanthurus triostegus (Linné, 1758) Acanthurus xanthopterus (Valenciennes, 1835) Acanthurus sp. Juv.) jaune Ctenochaetus striatus (Quoy et Gaimard, 1825) Ctenochaetus strigosus (Bennett, 1828) Naso annulatus (Quoy et Gaimard, 1825) Naso brachycentron (Quoy et Gaimard, 1825) Naso brevirostris (Valenciennes, 1835) Naso hexacanthus (Bleeker, 1855) Naso lituratus (Bloch and Schneider, 1801) Naso unicornis (Forsskal, 1775) Naso vlamingii (Valenciennes, 1835) Zebrasoma rostratum (Giinther, 1873) Zebrasoma scopas (Cuvier, 1829) Zebrasoma veliferum (Bloch, 1795) SIGANIDAE Siganus argenteus (Quoy et Gaimard, 1825) BOTHIDAE Bothus mancus (Broussonet, 1782) BALISTIDAE Balistapus undulatus (Mungo Park, 1797) Balistoides viridescens (Bloch and Schneider, 1801) Melichthys niger (Bloch, 1786) Melichthys vidua (Solander, 1844) Odonus niger (Ruppell, 1837) Rhinecanthus aculeatus (Linné, 1758) Rhinecanthus rectangulus (Bloch and Schneider, 1801) Sufflamen bursa (Bloch and Schneider, 1801) Sufflamen fraenatus (Latreille, 1804) Xanthichthys caeruleolineatus (Randall, Matsuura, Zama, 1978) MONACANTHIDAE Aluterus scriptus (Osbeck, 1765) Amanses scopas (Cuvier, 1829) Cantherhines dumerilii (Hollard, 1854) OSTRACIIDAE Ostracion cubicus Linné, 1758 Ostracion meleagris Shaw, 1796 TETRAODONTIDAE Arothron hispidus (Linné, 1758) Canthigaster bennetti (Bleeker, 1854) Canthigaster solandri (Richardson, 1844) Canthigaster valentini (Bleeker, 1853) Paulas aa Goaeres Fea Sie Be Sr Se ee ei lav }iae| ne) |= tae] C00 OCO0000000N0 om Oleleleleloleloleyeyeoleleyeleleleoleleleleleyleyeleyeleyeye) eo) Riis Sirs — hoto Intes) ce a week. (P Id in the traps are collected about on : fishes held in Plate 2 : The Plate 4 : A "goelette" weekly brings the fishes caught in Tikehau to the Papeete market. (Photo Morize) ATOLL RESEARCH BULLETIN NO. 416 COLONIZATION OF FISH LARVAE IN LAGOONS OF RANGIROA (TUAMOTU ARCHIPELAGO) AND MOOREA (SOCIETY ARCHIPELAGO) BY V. DUFOUR ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 COLONIZATION OF FISH LARVAE IN LAGOONS OF RANGIROA (TUAMOTU ARCHIPELAGO) AND MOOREA (SOCIETY ARCHIPELAGO) BY V. DUFOUR ABSTRACT The colonization of the lagoon by coral reef fish larvae was compared between two islands of French Polynesia, the atoll of Rangiroa and the high volcanic island of Moorea. In both cases the larval flux coming into the lagoon followed a daily cycle. Larvae were mainly caught at dusk and during the night, and on both islands the colonization was higher during moonless than moonlit periods. The larval flux did not appear to be dependent on the waterflow in the lagoons. A comparison of larval abundance and taxonomic lists indicates that Scarids and Labrids were dominant in Rangiroa while Gobiidae was the major family on Moorea. This difference could be in part related to the different sampling periods, but other environmental and biological factors could also be important. INTRODUCTION Most reef fishes have a pelagic larval phase, ending with the colonization of the reef (Leis, 1991). This recruitment of fish larvae on coral reefs is now studied in detail “since it has been assumed that events occuring during this period determine the characteristics of reef-fish stocks (Sale, 1980; Richards and Lindeman, 1987; Doherty and Williams, 1988). Although some studies emphasized the importance of the processes during the settlement of fish larvae among coral reefs (Sweatman, 1985, 1988; Victor, 1986), this phenomenon is not clearly understood. For fifteen years, scientists have studied mechanisms of this return to the parental habitat. These studies have been limited mainly to continental reefs (e. g.: reefs of Central America), or patch reefs along continental platforms (e.g.: Great Barrier Reef of Australia) with very little data available on recruitment of reef fish species in oceanic islands and in atolls. This is a first attempt to compare some features of fish colonization of the lagoons of two geomorphologically different islands located in French Polynesia. Laboratoire d'‘Ichtyoécologie Tropicale et Méditerranéenne, Ecole Pratique des Hautes Etudes, URA CNRS 1453, Université de Perpignan 66860 Perpignan Cedex France et Antenne EPHE-Muséum Centre de l'Environnement BP 1013 Moorea Polynésie Frangaise Manuscript received 14 October 1993; revised 2 June 1994 Although the data were not obtained simultaneously in both islands, it is still useful to compare these two sets of data. It is also worth considering whether or not the observed differences are due to the location, the geomorphological features of the islands, or the time lag between sample collecting on the two islands. MATERIAL AND METHODS STUDY AREA Rangiroa Atoll (figure 1) is one of the largest atolls in the world and the most important of the Tuamotu Archipelago (Ricard, 1985). It is 70 km long, 30 km wide and the peripheral rim is 225 km long. One third of the rim is above the sea surface and consists of small cays separated by channels. The rim flat is generally wider in the northern than in the southern part (800 m vs 500 m). The lagoon is biologically very rich compared to the other atolls of Tuamotu and is one of the most important reef fisheries centers of this Archipelago. The maximum estimated depth is 35 m and a lot of pinnacles are evenly distributed on its surface. Two passes, 450 to 550 m wide and 14 to 35 m deep are located in the North coast and lagoon waters are flushed out through these passes during ebb tides (35 cm to 60 cm tide range). Oceanic waters flow into the lagoon through channels over the atoll rim and the two passes during flood tides and also when trade winds blow. The fish larvae were collected in a channel, midway between the two passes. Moorea Island (figure 1) is located 25 km north-west of Tahiti (Galzin and Pointier, 1985). This volcanic island has a triangular shape with a 61 km coastline and a relief of 1200 m. The island is surrounded by a barrier reef, which encloses a lagoon, 800 to 1600 m wide. The reef is intersected by several passes. Two bays are located on the northern part of the island. The lagoon is generally shallow (1 to 5 m), but deeper near the passes. The oceanic water enters the lagoon by waves breaking over the outer reef crest, and return to the ocean through the passes. The very weak tides on Moorea (average range 15 cm) do not reverse the current in the passes. Sampling was carried out on the outer reef crest, 600 m away from the pass. METHODS Samples were collected off the northern coasts of both islands. Fish larvae were collected with an anchored net that filtered the waterflow coming into the lagoon. The net with rectangular mouth (1 x 0.25 m) was of mesh size 0.5 mm. A General Oceanics flowmeter was fixed in the mouth of the net. On Rangiroa Atoll, the net filtered the water coming from the seaward reef flat to the lagoon. It was located 500 m from the outer reef front. The channel was made of gravel in a shallow area (0.5 m). On Moorea Island, the fish larvae were collected on the outer reef crest. The net was fixed on the reef substrate and filtered the water coming over the crest with the 3 breaking waves. Thus, the water flow was not constant but it was estimated over the time period of each sample (10 minutes). The time lag between two samples was lh or 2h. Two diel cycles were made in February 1989 on Rangiroa. At Moorea, two diel cycles were made in April 1989 and a third one was made in October 1989. At Rangiroa, the first cycle was 3 days before the new moon and the second was around the first quarter. The diel cycles were made during same lunar phases on Moorea. The collected larvae were then fixed with 5% formalin seawater and identified under a dissecting microscope at the lowest taxonomic level following the recommendations of Leis and Rennis (1983) and Leis and Trnski (1989). This correspond to a family-level identification for all the larvae but the Gobioids. In several cases, genus-level identification was accomplished. Most of the larvae caught were in the postflexion stage or in metamorphosis and were identified. Abundances of unidentifiable preflexion larvae and juveniles were very low and were simply pooled into preflexion and juveniles types. Results are expressed in larval abundance per sample which represents the larval flux, i. e. the number of larvae for 1m of reef section and for 10 minutes (Dufour, 1993). RESULTS DIEL CYCLES The diel cycles from Rangiroa (figure 2) represent changes in larval composition over a two night period. Although the samples do not cover 24 hours, they take into account the two consecutive daily changes at dusk and dawn. Our data clearly indicate that fish larvae were present only at night and during twilight. Because the waterflow ' may influence the larval flux, the volume filtered by the net was also presented. The Kendall coefficient correlation rank calculated with Statview software (Abacus Concepts, Inc, Berkeley, 1992) was significant for the comparison between the two larval fluxes of the two cycles and also between the two water flows (Table I). But it was positive between the larval fluxes (+0.524) while it was negative between the water flows (-0.486). This result indicates that the change of the larval flux was somewhat similar during the two sampling periods. But the water flow was negatively correlated “between the two cycles. The Kendall coefficient correlation rank calculated between the water flow and the larval flux for the two diel cycles made on Rangiroa was very small and not significant (table II). However, the second cycle shows a strong decrease of the water flow in the middle of the night and this decrease was also observed for the larval abundance. These results indicate that there was no clear link between water flow and larval flux, except for low water flow, which could hinden larval colonization in the lagoon. The diel cycles made on Moorea also show that most fish larvae were taken at night and dusk (figure 2). The Kendall 's tau calculated between the larval flux and the water flow (Table II) was not significant. The absence of significant correlation in the 4 four studied cycles confirms that larval flux did not seem to be quantitatively dependent of the water flow. The study of larval flux on the two islands reveals that the larval flux on Rangiroa reached 3 times the value of 500 larvae per sample, which was obtained only once on Moorea, despite a larger sampling effort. On both islands these larval peaks occured in the early evening. A second peak was found just before dawn on the second cycle on Rangiroa. The water flow during these larval peaks on Rangiroa was not very high and similar to that found during larval peaks of Moorea. As a result, these high peaks of larval colonization on Rangiroa and Moorea do not appear to be created by variation in water flow over the reef of these islands. The comparison of the average larval flux recorded on the two islands at different times indicates that this flux appears to be more significant on Rangiroa than on Moorea (Table III). It was obvious that a high larval flux from these islands was never recorded during full moon. However, during moonlight periods of the first lunar quarter, the larval abundance on Rangiroa was higher than the abundance on Moorea. TAXONOMIC ANALYSIS OF THE SAMPLES The number of larvae and the number of larval types were different between the two islands (Table IV). The total number of larvae from Rangiroa was almost half the number of those collected from Moorea during eleven months, although the number of samples was higher. Based on the two studied periods, the average larval flux on Rangiroa reached three times the average larval flux on Moorea. The number of larval types on Moorea was 56 for the three cycles. The number of larval types on Rangiroa during only two nights was 43. Several larval types from Moorea were not found on Rangiroa, while only one larval type from Rangiroa was missing from Moorea. Some of these types were represented by more than 50 larvae. The comparison between Rangiroa and all the samples of Moorea indicates that the number of types was twice as less as that found in all the samples of Moorea despite the fact that the number of samples collected was eight times higher and the sample period was much longer in Moorea. Therefore, the number of larval types caught in two nights on Rangiroa was significantly higher that those caught off Moorea. The list of the larval types and their abundance is presented for both Rangiroa and Moorea (Fig. 3). The pie diagrams show the percentages of the main larval types for each island. The abundance of the larvae from Moorea is presented for all the 358 samples made between March 1989 and November 1989 (grey bars) and for the three diel cycles previously studied (black bars). The most abundant larval type on Rangiroa was the Scaridae forming 52% of the total catch. The two most abundant larval types on Moorea were Gobiidae (Gobiidae type 1 and Gobiidae type 56). The abundance of Gobioid types on Moorea represents 63% of the total catch. Scaridae were the second most abundant family on Moorea but they represented half the number of Scaridae collected from Rangiroa. On Rangiroa Gobioid types were the second most important group but their number were far below those of the Scaridae. The other significant larval types were found in similar numbers on both islands although periods of sampling were 5 different. This was the case for the Labridae, the Callionymidae and the Schindleriidae. It is apparent that the number of larvae of these families would have been much higher on Rangiroa if the extent of sampling was similar to that carried out off Moorea. The Apogonidae type 2 were more abundant on Rangiroa but the total number of Apogonidae from both areas was not very different. Juvenile fishes were caught in both islands in relatively high number. Different families were gathered in this type (Mullidae, Holocentridae...). It is interesting to note that these juveniles were collected at dusk despite the fact that daylight was supposed to assist in a higher avoidance of the net. The Gobiidae type 8 was only collected at Rangiroa. DISCUSSION The daily patterns of the reef colonization by reef fish larvae have been demonstrated only recently on coral reefs (Dufour, 1991, 1993). The fish larvae that enter the lagoon were caught only at night and dusk. Their abundance was also found to be higher during moonless periods. This pattern has been confirmed by samples over a two years period. The data from Rangiroa in this study confirm this finding. Each cycle made at Rangiroa demonstrated that fish larvae were abundant during the moonless nights in the channel of the atoll. The larval abundance could reflect higher larval activity above the reef at night (Hobson et Chess, 1978). However, the fixed nets could not catch larvae that do not move into the lagoon. Hobson and Chess (1978, 1986) have demonstrated that planktonic organisms drifted at night over the reef of Enewetak atoll to enter the lagoon. Their appearence over the reef was related to a vertical migration at night, followed by a passive drift in a current flow induced by breaking waves. However, colonization by fish larvae at Rangiroa and at Moorea was only accomplished by individuals ready to settle. The larval flux observations do not include preflexion larvae because these larvae were scarce in samples, although they could have drifted . more easily than postflexion larvae. It is known that postflexion larvae are able to swim (Blaxter, 1986; Webb and Weihs, 1986). Moreover, reef fish larvae can avoid the reef area until they are competent for metamorphosis (Kingsford and Choat, 1989). These phenomena imply other mechanisms of colonization in addition to passive drift. The larval flux in the lagoon could thus be viewed as an active process made nightly by competent fish larvae. Night activity correllated to the darker phases of the moon cycle has also been demonstrated for other planktonic organisms over reefs (Aldredge and King, 1980, Tranter and al., 1981). These authors found that this moonless activity was an adaptative advantage against predation. In a similar way, the colonization of fish larvae occurs at night when predation is lower (Hobson, 1973, 1975). Therefore, larval colonization of the lagoons at night could be viewed as an adaptative process against predation, as predation plays a major role during the recruitment of reef fishes (Shulman and Ogden, 1987, Victor, 1986, Hixton, 1991). Both the geomorphology of the reef and hydrodynamic characteristics of the waters flushing into the lagoons appear to have no significant control on larval colonization. The difference of the abundance of fish larvae between the two islands can be explained by the difference of the sampling periods. Although it has not been established that fish larvae were more abundant in French Polynesia during February than during 6 April, the summer season was considered to be the recruitment season in other coral reef areas (Williams, 1983, Victor, 1987). Thus, the lower abundance in samples from Moorea could be explained by variations related to seasonal recruitment. The difference in abundance and diversity of fishes during colonization between these two islands could also be related to the size of the lagoon. The quotient of reef periphery to surface of the lagoon is also much lower for Rangiroa than for Moorea. This is because the lagoon of Moorea encloses the volcanic island and does not cover all the surface delimited by the outer reef like an atoll. On Moorea, the quotient of the lagoon surface to the reef length is around 0.86 km"! (60 km/70 km2), on Rangiroa it is 0.11 km-! (230 km/2100 km2), but the sand cays over one third of the reef lower this coefficient to 0.074. This last value is more than 10 times smaller than on Moorea. If we could assume that the density of the larval flux per unit of lagoon surface over the crest was related to this coefficient, the number of fish colonizing the lagoon should be proportionally higher. This assumption could explain the higher rate of colonization for Rangiroa. This hypothesis cannot be verified, however, because the larval flux over all the reef rim has not been determined. The difference between the major larval types from the two islands could also be explained by other hypothesis. The composition and diversity of adult fishes in both lagoons was probably not the same. It is possible that the number of fish species in the lagoons of atoll is related to the surface area of these atolls (Galzin et al., 1994). Scaridae and Labridae are among the most abundant fishes in atoll lagoons (Bouchon- Navarro, 1983, Morize et.al., 1990), while Pomacentridae and Acanthuridae are more abundant in Moorea lagoon (Galzin, 1987). Although we have no information about their density in Rangiroa atoll, the higher abundance of Scaridae larvae on Rangiroa was not surprising. But this higher abundance could be related to the low number of samples collected in Rangiroa, and the period when they were collected. It is possible, however, that the pattern of settlement of fish larvae on reefs could be relatively unpredictable and chaotic and peaks of larvae have be described as randomly distributed at different time scales (Doherty and Williams, 1988). Another explanation could be the reproduction period of Scaridae, which could occur earlier. Larvae of Scaridae, however, were caught on Moorea until the end of June and Scaridae and Labridae were also the most abundant families in samples made in May and June 1988 on Moorea. CONCLUSIONS The study of the larval flux over the reef on Rangiroa and Moorea was useful to the understanding of some aspects of the settlement processes of fish larvae in lagoons. This study has confirmed some trends in the diel and lunar cycles of reef colonization by fish larvae. The difference of larval abundance between samples on both islands can be related to the time lag between the sampling periods of each island. The sizes of the two lagoons could also play a role in this difference. It was more difficult to understand the taxonomic difference. It could be explained by the difference in size of the two lagoons, or by the period of fish reproduction or even by the density of the different families, but few data were available to confirm these hypotheses. REFERENCES Alldredge A. L., King J. M., 1980. Effects of moonlight on the vertical migration patterns of demersal zooplankton. J. exp. mar. Biol. Ecol., 44: 133-156. Bouchon-Navarro Y., 1983. Distribution quantitative des principaux poissons herbivores (Acanthuridae et Scaridae) de l'atoll de Takapoto (Polynésie frangaise). J. Soc. Océan. 39(77): 43-54. Blaxter J. H. S., 1986. Development of sense organs and behaviour of Teleost larvae with special reference to feeding and predator avoidance. Trans. Am. Fish. Soc. 115(86): 98-114. Doherty P.J., Williams D.McB., 1988.-The replenishment of coral reef fish populations. Oceanogr. mar. Biol. Ann. Rev., 26: 487-551. Dufour V., 1991. Variation d'abondance des larves de poissons en milieu récifal: effet de la lumiére sur la colonisation larvaire. C.R. Acad. Sci., Paris, t.313, série III: 187-194. Dufour V., Galzin R., 1993. Colonization patterns of reef fish larvae to the lagoon at Moorea Island, French Polynesia. Marine Ecology Progress Serie, 102: 143-152. Galzin R., 1987. Structure of fish community of French Polynesia coral reefs. 1/ Spatial scales. Mar. Ecol. Prog. Ser., 41: 137-145. Galzin R., Pointier J-P., 1985. Moorea Island, Society Archipelago. In: Proc. 5th int. Coral Reef Congress, B. Delesalle, R. Galzin and B. Salvat (eds.), 1: 73-102. . Galzin R., Planes S, Dufour V, Salvat B., in press. Variation in diversity of Coral reef fishes among French Polynesian atolls. Coral Reefs. Hixton , 1991. Predation as a process structuring coral reef fish communities. In : The ecology of fish on coral reef, P. F. Sale (ed.), Academic Press : 475-508. Hobson E. S., 1973. Diel feeding migrations in tropical reef fishes. Helgol. Meer. 24 : 361-370. Hobson E. S., 1975. Feeding patterns among tropical reef fishes. Am. Scient., 63 : 382-392. Hobson E. S., Chess J. R., 1978. Trophic relationships among fishes and plankton in the lagoon at Enewetak atoll, Marshall island. Fish. Bull, vol 76: 133-153. Hobson E.S., Chess J.R., 1986. Diel movements of resident and transient zooplankters above lagoon reefs at Enewetak atoll, Marshall islands. Pac. Sci., (40): 7-25. Kingsford M. J., Choat J. M., 1989. Horizontal distribution patterns of presettlement reef fish: were they influenced by the proximity of reefs? Mar. Biol., 91: 161-171. 8 Leis J. M., 1991 The pelagic stage of reef fishes: the larval biology of coral reef fishes. In: The ecology of fishes on coral reefs. Sale P. F., ed., Academic Press, San Diego, p. 183- 230. Leis J. M., Rennis D. S., 1983. The larvae of Indo-Pacific coral reef fishes. J. M. Leis (ed.) NSW Univ. press & Univ. of Hawaii press, 269pp. Leis J. M., Trnski T., 1989. The larvae of Indo-OPacific shorefishes. NSW Univ. Press & Univ. of Hawaii press, 371pp. Morize E., Galzin R., Harmelin-Vivien M., Arnaudin H., 1990. Organisation spatiale du peuplement ichtyologique dans le lagon de l'atoll de Tikehau (Polynésie francaise). Notes et Doc. ORSTOM, N°40, 44pp. Ricard M., 1985. Rangiroa Atoll , Tuamotu Archipelago. In: Proc. 5th int. Coral Reef Congress, B. Delesalle, R. Galzin and B. Salvat (eds.), 1: 159-210. Richards W. J., Lindeman K. C., 1987. Recruitment dynamics of reef fishes: planctonic processes, settlement and demersal ecologies, and fishery analysis. Bull. Mar. Sci., 41 (2): 392-410. Shulman M. J., Ogden J. C., 1987. What control tropical reef fish populations: recruitment or benthic mortality? An example in the Caribbean reef fish: Haemulon flavolineatum. Mar. Ecol. Prog. Ser., 39: 233-242. Sweatman H. P. A., 1985. The influence of adults of some coral reef fishes on larval recruitment. Ecol. Monogr. 55: 469-485. Sweatman H. P. A., 1988. Field evidence that settling coral reef fish larvae detect resident fish using dissolved chemical cues. J. Exp. Mar. Biol. Ecol., 124(3): 163-174. Tranter D. J., Bulleid N. C., Campbell R., Higgins H. W., Rowe F., Tranter H. A., Smith D. F., 1981. Nocturnal movements of phototactic zooplancton in shallow waters. Mar. Biol. 61 : 317-326. Victor B. C., 1986. Larval settlement and juvenile mortality in a recruitment-limited coral reef fishes population. Ecol. Monogr., 56(2): 145-160. Victor B. C., 1987. Growth, dispersal, and identification of planktonic labrid and pomacentrid reef fish larvae in the eastern Pacific Ocean. Mar. Biol., 95: 145-152. Webb P. W., Weihs D., 1986. Functional locomotor morphology of early life history stages of fishes. Trans. Am. Fish. Soc., 115: 115-127. Williams D.McB., 1983.-Daily, monthly and yearly variability in recruitment of a guild of coral reef fishes. Mar. Ecol. Prog. Ser., 10: 231-237. Table I: Values of the Kendall coefficient correlation rank for the larval flux and the water flow between the two diel cycles from Rangiroa (n.s: not significant at 5%, s: significant at 5%). Comparison of the larval fluxes 0.524 s Comparison of the Water flows -0.486 s Table II: Values of the Kendall coefficient correlation rank between the water flow and the larval flux (n.s: not significant at 5%). [____JRangira 1 [Rangioa2 [Moowal [Moorea [Moorea Kendall coefficient | 0.206 n.s -0.176 n.s_ |-0.109 n.s | -0.036 n.s| 0.345 n.s Tableau III : Average values of the water flow and the larval flux for the cycles from Rangiroa (R) and Moorea (M), standard deviation are in brackets. water flow : -m3-sample-! | abundance : larves.sample-! 122.5 (156.4) M 23.10 S50 (GID 38.1 (25.9) Tableau IV : Abundance of larvae and larval types from Rangiroa and Moorea | = SRangiroa | Moorea (3 cycles) | Moorea (all samples) 44 es 10° Marquesas *Archipelago Pacific Ocean Rangiroa,. Tuamotu Archipelago ° ow uC) ° SS -. ¢¢. Moorea Society Archipelago, Lec Tahiti 500 km @ high Island O Atoll 147°40'W Study area 17°30S Reef crest Pacific Ocean 149°S0W Figure 1. French Polynesia (above) with the atoll of Rangiroa, Tuamotu archipelago (middle), and the high Island of Moorea, Society archipelago (below). Larval flux water flow Rangiroa Larval flux water flow New moon Apr. 89 Water flow % 3 Lamy oS fe 3 | First quarter Apr. 89 Moorea Larval flux Water flow Larval flux Water flow Figure 2. Evolution of the larval flux expressed in number of larvae. sample7! (bars) and the water flow in m*?. sample"! (line) during nycthemeral cycles made on Rangiroa and on Moorea. The black thickness on the categories axis represents the night hours, the white frame on the same axis represents moonlit hours. Rangiroa Moorea Scaridae 1 Gobiidae 1 Labridae 1 OC 4 juveniles Callionymidae Labridae 6 a i SS Schindleriidae leptocephalii ¥ El ES Apogonidae 2 others Gobiidae 2 OM oO lopidae | Blenniidae | Pin BED wide Rangiroa abridae 3 Synodontidae Apogonidae 1 bepocepbali ridae 6 Apogonidae 2 Schindleriidae (ad.)_ = Callionymidae Labridae | 2229 Scaridae 1 0 100 200 300 400 500 0 500 1000 1500 2000 Figure 3. Percentage of the main larval types collected on Moorea and Rangiroa (above) and diagram of larval abundance (below) for all the samples from Rangiroa and for the three cycles of Moorea (black) and all the samples from Moorea (grey). n.i.: not identified to lower taxonomic level; juv.: juvenile; ad.: adulte fishes are also included in this neotenic family. ATOLL RESEARCH BULLETIN NO. 417 CAVES AND SPELEOGENESIS OF MANGAIA, COOK ISLANDS BY JOANNA C. ELLISON ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 oie edmaae . e ; ‘ i 7 ress at 8 ees on nit at ie L002 LAvGRVEDT meEIPAOOLINNG ane: CAVES AND SPELEOGENESIS OF MANGAIA, COOK ISLANDS BY JOANNA C. ELLISON ABSTRACT Ten caves in the makatea limestone of Mangaia, Cook Islands were explored and mapped, totalling over 3.7 km of passage. Of these, there was an apparent grouping by elevation that corresponds with previously described sea-level terraces in the makatea. Four caves have major level sections 10-20 m above sea-level, corresponding with a 14.5 m Pleistocene terrace. The high dimensions of these caves indicate downcutting during slow uplift, or multiple reoccupations by highstands of Pleistocene sea- levels. One major cave has level passage 20-30 m above sea-level, corresponding with a 26-27.5 m terrace. Three caves have level passage 40 m above sea-level, corresponding with a 34-39 m terrace. Active conduit caves are developed at the present sea-level, but are closed with heavy clay deposits from recent soil erosion. INTRODUCTION Mangaia is the second largest and most southerly of the Cook Islands (21°54'S, 157°58'W), with a land area of 52 kmé@ (Fig. 1). The island is divided into two concentric geological zones. The inner zone is a subdued ‘basaltic volcanic cone rising to 168 m, flattened at the summit possibly by marine erosion prior to its uplift (Wood, 1967), dating between 17-19 m yr BP (before present) (Dalrymple et al., 1975). The outer zone is a complete raised limestone rim or makatea, 0.7 to 2 km wide, up to 70 m in height, with erosional topography of steep terraces on the outer edge, and cliffs on the inner edge (Stoddart et al., 1985). Yonekura et al. (1988) showed from the identification of planktonic foraminifera in makatea limestone that these are up to 17 m years old, indicating that coral reefs developed shortly after the volcanic island formed. While the major part of the limestone is Tertiary, Pleistocene deposits occur on the seaward margins to an elevation of 14.5 m (Stoddart et al., 1985). Emergence occurred in late Tertiary to Quaternary times, during which there were two periods of stability in relative sea-level to cut marine terraces at 26- 27.5 m and 18-20 m (Stoddart et al., 1985). Department of Biogeography and Geomorphology, Research School of Pacific Studies, Australian National University, Canberra ACT 2601, Australia. Manuscript received 25 January 1993; revised 14 September 1993 mance eeet TTT een, pL Bart ee of =e’, pies aa To'uri F Taenoterau Teruarere | . Avarua < ae Fuss; Touroporu Piriteuemeume @ oO Ls BASALTIC CONE oe a e°e oe 0 oe ene eee ee ee MAKATEA LIMESTONE Tautua pe etal ae coe =o yee - meme FRINGING REEF Swamps 0 1 2 ® Entrances of cave mapped hieginstn, Nols in this study cale: km 3 Mangaia receives a mean annual rainfall of 1967 mm, with range in the period 1914-1984 of 1024 to 2983 mm (Thompson, 1986). There is a pronounced wet season from November to April, and dry season from May to October. Drainage is radial, with deeply incised first and second order streams off the central cone feeding lowland taro swamps collected against the inner makatea cliff. Water from the swamps sinks beneath the makatea limestone in radially draining cave systems. Stoddart et al. (1985) showed that stream water entering the makatea limestone is aggressive, supporting the interpretation that the cliffs of the inner makatea are erosional remnants from a former complete cover of limestone to equivalent elevations on the volcanic slopes (Stoddart and Spencer, 1987, Figure 4). The purpose of this study is to investigate how these events are expressed in the speleogenesis of Mangaia. Many cave entrances can be seen in the makatea cliffs, and the topographic map of Mangaia shows 65 sink holes on the top of the makatea, which considering that the map was made from air photographs of a heavily forested area must under-estimate total numbers. The sinkholes are strongly clustered to indicate traces of cave systems. A few are used for burials or settlement and hence are of archaeological interest. As commented by Gill (1894), "the numerous and extensive caves that honeycomb the makatea were formerly used as habitations, cemeteries, places of refuge, and stores. Scores of them are filled with dessicated human bodies". While very little work has been done on the cave systems of makatea islands, but the following principles on cave evolution are suggested from the surface geomorphological work. 1. As cave development is most active at the water table, those that are ' presently under active development can be found at the vadose conduits where streams enter the makatea. These decline slightly along their course, similar to a stream. 2. Caves above these levels are fossil conduits, with a positive relationship between elevation and age, resultant from uplift of Mangaia and general decline in sea-levels from the Tertiary to the Quaternary (Haq et al., 1987). 3. Fossil conduit caves should decline slightly from the influent entrance at the inner edge of the makatea to the coast, at an elevation that is slightly above the sea-level position at the time of development. On low limestone islands, such caves can be used as an indicator of former sea-leve! (Mylroie and Carew, 1988). 4. Higher elevation caves should therefore show features of older caves, with collapse, flowstone infill and large speleothem formations. Lower elevation caves should show features of younger caves, with more even walls and floor, and smaller formations. 5. A conduit cave could cut into its floor during uplift of the makatea, developing a deeply rifted cave. Such caves could indicate a period of slow uplift, while rapid uplift would result in abandonment of the former conduit and development of a new cave at a lower level. METHODS In July 1991 ten caves were explored and surveyed, to the British Cave Research Association Grade 5b standard, using a Suunto compass and clinometer, and 50 m fibreglass tape. This standard requires a station- to- station survey, with passage details recorded at the time (Ellis, 1976). Cave maps produced are shown in plan view, so it must be remembered that passages are not shown at their actual length unless they are horizontal. A profile view is also shown of the cave passage. No vertical techniques were possible in this study, and climbing risks not taken owing to lack of back-up support. Where possible, surface survey was continued to a known elevation to give the altitude of the cave as indicated on the maps, otherwise it was estimated from contours on the topographic map. DESCRIPTIONS OF THE CAVES Tuitini cave, Veitatei (Figure 2) Tuitini is the largest cave explored in inis study, with two entrances about 100 m to the east of Lake Tiriara. It would have been the conduit cave for the Veitatei drainage basin, which is the largest on Mangaia, when relative sea- level was 20 m higher. Survey was continued to the water level of the lake to establish the altitude. The main passage is large, with few formations, while the upper passages to the south and east are well decorated with formations. There are four burials in the cave, permission to explore should be sought from the chief in Kaumata village (Oneroa). The main passage indicates downcutting, with heights of 20 to 25 m, while the upper passages have low ceilings of 10 m or less. There are sections in the main passage with collapse, it is necessary to climb around or over large boulders. At the end of the surveyed section the main passage continues at the base of a 7 m pitch, at which point the cave is around 30 m high. KEY TUITINI CAVE Mangaia, Cook Islands + Skeleton GR 21°S7'S; 15 7°S6'w » Column bent ese m ni aE 8 > Stalactite epth-13m =" : : Alt. (main entrance) 23 m +, e Boulder Grade 5b ki = Pool CAYE PLAN cave enlarges fossil and continues : ossil poo main a ridges on floor at lower level entrance pitch ++ ee down makatea upper passage Af wall with many > hole stalactites int i calcite O entrance arch adjacent Pie <= O20 5 40 poe Scale: m to Lake Tiriara KEY H main passage + L. Tiriara entrance CAVE PROFILE | south upper passage 10 . % main east passage entrance Depth (m) 0 100 200 300 400 Distance (m) Figure 2. 6 To'uri cave, Tava'enga (Figure 3) The entrance to To'uri is where the depression to the west of Tava'enga swamp meets the makatea wall. A stream leads to the cave entrance, which is approached by a steep climb down over boulders and ferns into a large overhang. The cave is partially an active streamway, with heavy wet red clay deposits throughout. A bank of red clay at the entrance has been incised by headward erosion of the stream to create a 2 to 3 m profile. The small stream follows the left hand side of the cave for 60 m, then sinks into a hole. In the stream water are small black fish. To'uri cave indicates recent stream flow that has downcut through older clay deposits, leaving exposed mud sections along sections of the passage wall. The cave is large in passage dimensions relative to others in Mangaia, the roof a wide vertical fissure visible to 20-30 m, the passage 3-10 m wide. There are occasional rocks fallen down, and occasional large stalagmite or flowstone formations. Stalactites are the more common. After 490 m of passage there is a clay bank 2.5 m down to flowing water, which heads out through a tight passage to the north. The water tasted salty and waves could be heard. Opposite, up a clay climb, the cave continues to the west. This was not explored. Elevations for Tou'ri and Teruarere caves are based on the salt water being at sea-level. Teruarere cave, Tava'enga (Figure 4) This is the second largest cave explored in this study, located 40 m above the entrance to To'uri Cave, along the makatea wall to the SW. The cave is accessed from the top of the makatea. The entrance rift trends east- west, and there is a 10 m climb down to the cave entrance 50 m back from the makatea edge, assisted by roots. A large Hernandia moerehroutiana tree grows out of entrance. Where the rift reaches the makatea edge one can look down on entrance to To'uri cave, some 100 m to the left. Strong winds come through the cleft. The cave entrance leads away from makatea edge, under the entrance climb. Cave is simple fossil streamway, 10-30 m high, 1-10 m wide, dry. It contains at least 7 skeletons lying on the floor near the cave wall, surrounded by stones. This is a burial cave, in the charge of Tuara George of Oneroa village, from whom permission to explore the cave should be obtained. Cave has been studied by the ornothilogist D. Steadman, showing from fossil bird bones that early Polynesian settlement caused extinction of many species (Steadman, 1985, 1986). The rift opens above to four daylight avens in the first 200 m of the cave. Formations are mainly calcite cascades, curtains and flowstones, with vandalism of whatever possible. The floor is hard red mud, with occasional patches of rubble. Lower in the cave, calcite flows cover the floor, and there are places where ridges on the floor indicate fossil pools. Towards the end of the main passage of the cave, there are a couple of climbs down, then a fi,tnb ut buwaids . WRa}S oO SHULS O weat}s jes ysu tebe yueg hela IS ph Les psu tebe doo| of yueg hes pasojdxa you abessed pasojdxa you abessed N¥1d JA¥o "¢ anol (Ww) aue}siq a9ue1}Ua W :a|BoS 00d ———_—= Japnog Bea 8h <0 9} 1W68|2}S 93.1398 (8}S uwn|o7 AJA ile ysu tebe yueg hes Ww ¢ yydaq W 69P y}bue7 MS9So2S1 ‘S.S'PSolZ ad spue|s| yoo3 “erebuey) JAVI 1aNn.OL pasoldxa you uo hew Ayinb ut pasoldxa bu tads you uo her, WES1}S p anol A00[) UL [oy Wo} auoysol4 RY W: a[F9S joog =» OF Of oa uadAe Buf sap|nog y4DLihep a} ube [235 uaae a}.1,98[8}S yy Bui fiep uwno7 UO}AL9AS AqA UdAG 1ybiifep uaAe an; asde| [09 tes a Aapun [Me49 Wlu uada do} eayeyxeu wo UAOP qQULIDWO'OL - Ba}eAeUW JO aseq N¥1d JAVD 0}W O61 - abpa eayexeuw (Ww) aque}siq OOP H PIUBIJUS JACI doy eayey ew WON) UMOP GUILD 31140add JAD 400|) UO jood [isso uaae yybijfep UaAG ole yybiihep yybiifhep is qS apeig W oP VU W PZ yydaq W 16¢ 4ybua7 M.G'9So2S1 ‘SS PSolZ ad spue|s| 4007 “e.ebuey, JAVI JddAdV NAIL 9 pitch down that was not explored. Some daylight comes through high above the pitch. The upper passage continues, but is a tricky climb past the pitch, this was not explored. Touropuru cave, Ivirua (Figure 5) This is the third largest cave explored in this study. The entrance is 10 m up the makatea wall adjacent to the small Kirikiri swamp, between the larger Karanga and Ivirua taro swamps. Adjacent to the entrance to the south was a large cave shelter used for habitation, which has unfortunately recently collapsed. This cave is the burial cave of the Totongaiti tribe, with 22 skeletons in the main cave and 9 in a small cave above the lower entrance. Permission to explore this cave must be obtained from a member of the tribe in Ivirua village, such as Ma'ara Ora, Director of Forestry. Just inside the entrance is a side passage to the right, 39 m long to a window 10 m up the makatea wall. The passage is narrow and lined by 19 skeletons in open coffins, mostly of planks (post-contact), but some canoes (pre-contact). The main passage is a fossil vadose stream passage, generally 1-2 m wide and 3 m high, but sections which widen to some 6 m, and sections where the rift above is visible to some 15 m. After 350 m the passage forks, the left passage closing up with flowstone after 150 m, the right fork leading to the base of a doline entrance in Ivirua village, that has a 22.5 m pitch that can be climbed (GR 21° 54' 45" S, 157° 54' 15" W). The main passage has been closed in by flowstone deposits and some collapse, causing the route to go up and down, but from the entrance to the fork over a distance of 350 m there is no elevation change. There are three more skeletons at the start of the main passage, then no more. Formations are mainly curtains and .flowstones, though there are some large stalagmites and columns and more smaller stalactites. There is less vandalism of formations in the final left branch. After the fork, both passages climb some 4 m. There are some small passages off the main route, some of which may have unexplored sections. Erua Cave, Karanga (Figure 6) The cave entrance is located some 200 m to the south of the Karanga swamp conduit entrance, just north of where the path from Karanga village descends the makatea to the Karanga swamp. The cave entrance is conspicuous, 14 m above the base of the makatea, a rift 20 m tall, anda strong through breeze can be felt. The passage of this cave is wide and high, with few formations and a mud floor with frequent cobble-sized angular rocks. It was used as a refuge during prehistoric wars, and several human structures and midden deposits can be seen, but no burials. Like Tautua cave, which was also used as a war refuge, Erua has light inside the cave, from an aven. There is a debris slope from material which has fallen through the aven, including some animal skeletons. The few formations are large auoyserol4 [004d = Aap[nog o 9} LWBe[e}S S}LPIE LEYS uwn [079 U0} [34S Ad aUO}SMO} uaayaq yYybL} abessed ule, a youeig MS o Youe1g JS « Ada °G a4nbl Le }LWWNS Ba}exeLJ UO a[oyAULs 0} eslt auo}SAAOL) AaA0 pasojdxa jou cz A00[) UL afoy (wi) aaue}sig 31140dd JA¥I aseq Ba}exHeW 0} SUO}S[IAS YL Paul, WOOl- abessed apis Aowseu AODU LAS pasojdxa you Ba}EAeL AOL} UL ajoy qS apeag WO] C3dUBI}US }SaMI} “YL W 2+ Cabessed ulew) yjdaq W 2s yybue7 MG PSodGI SA PGolzZ ad spuels| 4007 ‘erebuey JA¥I NnOADdOANOGOL “9 a4AnBly (WwW) aaueysiq OO| S9UBI}US 0} Gulls WP] + 311d08d JA¥S Suap|nog jo yUawabuesie 4b (nbueysay su jojje\d pa}on4ysu07 we [Lea Ba}eyeLJ qS apeig Ybiheq N¥1d JA¥I WW: SLEDS oe $3|qq07 *. ai WOPr 9 4LY 02 oO 0 Ww 9 ujdag Wee} yybue7 a}IWBe[e}S ¢ ME OSoLS1 ‘SS PSolzZ 49 A}L}IB[EYS Se spue|s} 4009 “erebuey, AI JAVI ¥nas Japnog o 12 stalactites and stalagmites, and there are two avens which could extend the cave. Taenoterau Cave, Karanga (Figure 7) The cave entrance is located to the north of the Karanga swamp conduit entrance, 20 m above the level of the swamp. Behind the large rock at the entrance are trenches where a Japanese archaeological group dug in 1990. The lighted entrance is some 25 m tall, but falls to 5-10 m as the cave begins. Flowstone formations have blocked the passage twice in the first 20 m, so one has to climb over. The formations are all large, stalactites, stalagmites and dripstones, and crystalline flowstone walls. Tautua cave, Tamarua (Figure 8) At the makatea wall where the East Tamarua swamp collects there is a large cavern, over 30 m high and 100 m across (Plate 1). The floor is covered by large boulders from collapse, and there are old, vertical stalactites on the ceiling, indicating that the cavern has been exposed by makatea retreat. The stream from the swamp enters the cavern and through a conduit cave that is 15 to 20 m wide, mostly 1 to 2 m high, with extensive wet clay deposits. Above the conduit is the entrance to Tautua cave, up an 11 m climb. At the entrance the cave is 15 to 20 m high, and this height continues along the right branch just inside the entrance. This does not extend far, and has two connections with the lower streamway. On the left branch are habitation sites clustered around a window out of the makatea wall. The cave is well known from oral traditions as the primary refuge of the Tonga’iti tribe in times of war (Gill 1894, Buck 1934). Permission to explore the cave should be obtained from a member of that tribe, such as Noka Tumarama of Tamarua viilage. There are stone faced platforms, a marae (worship place), a tupe disk pitch (for playing a bowling- type game), and midden deposits. The site was analysed by P. Kirch and other archaeologists from University of California at Berkeley in 1989, and mapped by theodolite in 1991. The site was a retreat in times of war as it is easily defended. The lower entrance is a vertical climb, and defenders could bombard invaders with rocks. Past the archaeological site is a T junction, where the left branch leads up to a high makatea entrance, used as an escape route or otherwise blocked with stones. The right branch leads down to the main part of the cave, which was mapped here until a pitch down was reached. TEANOTERAU CAVE CAVE EROFILE Mangaia, Cook Islands GR 21°54.5'S; 157°54.5'w Length 5? m Depth 5 m Alt. 20 m Depth (m}) 20 40 Distance ¢m) Grade 5b archaeological excavations (Japanese, 1990) KEY oO Column ? Stalactite & Stalagmite : o Boulder +18 mclimb a Ce ee to entrance ; : Scale: m Flowstone Makatea wall Figure ?. 13 14 “9 aanbl (Ww) aaueysiq ool + O aguesjua abessed yjsou a Re suoja1ayxs oy pasoldxa abessed jsam + a Ol YpL Jaqueys you uo fie, abessed ulew o oe daddn {jews 02 & uAKOp fg youd uop ydjid AIA 3114d0dd JAYS Obl + Suu 40))e/d vo payonajsuos oy, 7 AQODU LA hevweays 6 AaAo| Ub 0} a9ue4}Ua Ww40j,e\d uo}a[ayS «+ payonaysu03 : pase) -auo}s auoysrol 4 F7 ujgaARa 190d goue4j}Ua Jo BueysaAo uwn[oj o ayelu xoidde $3(qqoj *. ese s aie 9S pe4g OZ al 0 w6 FY dapjnog o Ww o+ yjdag aplubejeys + = Ww See yybue7 A} 1}9B [BIS 4 = M.SSo2S 1 ‘SeSolZ a9 [[8. Ba}exeW UO spuels| yoo. ‘erebue, AA Ny N¥1d FAV yBiy aauesjua JAY] FNLNGL VS wam r m Cc re V. 1. Entran 16 Other Active Conduit Caves Ivirua Conduit Cave GR 21° 55' 30" S, 157° 54' 00" W. Where the Ivirua swamp drains through the makatea is a cave entrance similar to Tautua cave in the East Tamarua swamp. Above the stream is a large overhang some 25 m high, with old stalactites. The stream has red clay banks, and in pools there are many freshwater prawns (Macrobranchium lar). The conduit cave is wide and low, with few and small formations, and heavy red clay deposits, similar to East Tamarua. Neither of these active streamway caves were explored far because of problems of sinking in the wet clay. A small raft is recommended, as this could be dangerous. To the north of the streamway, there is a cave entrance up a 10 m climb, resembling the position of Tautua cave in Tamarua. The cave was not explored. Lake Tiriara Conduit GR 21° 57' 00" S, 157° 56' 15" W. Lake Tiriara in Veitatei, adjacent to Tuitini cave, drains through a large cave entrance (Plate 2). These dimensions continue for some 50 m, with 2 m depth of water, then the cave narrows and the roof meets the water level. Figure 2 shows that the main passage of Tuitini cave trends towards the Lake Tiriara conduit, and it is possible that they connect in the lower unexplored section of Tuitini cave. On the makatea summit directly above where the Lake Tiriara conduit enters the cliff, c. 10 m back from the cliff face, there is an arch some 15 m high, and a pitch cave entrance in the south wall of the arch. The topographic map shows a number of clustered cave entrances on the makatea summit above the Tuitini and Lake Tiriara conduit cave systems. Kauvava cave, Temakatea (Figure 9) Kauvava and Piriteumeume caves both have doline entrances, rather than makatea wall former streamway entrances. As they are of the highest elevation of caves explored, they could have been formed before erosion developed the makatea wall. The southern entrance to this cave is the large doline to the N of Temakatea village. Permission to explore the cave should be requested from Papa Tua, who lives to the east of the cave entrance. At the base of the doline is a 6 m climb down at the entrance, which is filled with boulders and other debris. The entrance passage has rounded walls, indicating meandering stream activity, but is presently dry. The cave is mostly a tall rift, visible to 20 m. Some 60 m into the cave the passage descends and narrows, indicating a sink which is now infilled by deposits of clay. This is the lowest part of the cave between the two avens. Towards the northern aven flowstone has closed the passage, but there is a tight climb up the rift over this. The aven is 23 m high, the opening is behind the Mormon church in Temakatea village. Below the aven the cave 17 tei. Lake Tiriara, Vei In n nr P 16 Aapinog 9} Le [e}S BPLPIS LEYS uwnjo4 Uo}a|axS + AJA au LLop W :a[(bIS SS Se Ue Oho —> NW "6 aanbly pasoidxa you Uo fier Ra}8>6W jo do} wou uaae yYybifieg (7464) Ls dn quiljyw get Mela YyyLer Pally 4uls W} aue\siq qs apeug oat aoue4jua W Og 9 (a9ue4jUa UaAe) “RLY Ww g¢ yydaq woz yybuaq ME eLSod GI S.G°SSolzZ a9 spue|s| y007 ‘erebue, JAVI VAVANVA 311408d JAI uaae yybifeg pal [ul “uaae AapUuNn Spuayxsa aoy NY1d JA¥o oO oO (W) aouesjua Woy Yyidag | ™ oO NM RS) continues to the north, but this is down a pitch so could not be explored in this study. Pirituemeume cave, Temakatea (Figure 10) The makatea summit is characterised by deep (5-10 m) dissection, with "karst streets” that follow the joints (Stoddart, et al., 1985). The entrance to Piriteumeume cave is a small doline in a karst corridor to the south of the inland road from Temakatea village, opposite the quarry. There are old (dry) flowstone walls at the entrance, and a strong draft. The cave is relatively old, indicated by its elevation, collapse features and large formations. It links three dolines, and has a low daylight aven towards the northern end. The height varies from 4 to 15 m, and the floor is irregular and filled with collapsed boulders. The flowstone walls are dry and rough, indicating that they are no longer under formation. The southern section of the cave has a cavern with large stalactites and dripstones, while the northern section has several large columns. Man 88 cave, Keia (Figure 11) No name could be found for this cave, so it is described by the archaeological site number of P. Kirch, who excavated both in the cave and below its entrance in 1991. The entrance is 8 m from the makatea base, a large lighted cavern adjacent to a pool of water further inside the cave, with signs of occupation. The cave does not extend far, and seems to have been Closed by calcite deposits. SPELEOGENESIS The ten caves described include the major caves of Mangaia, but many more remain to be explored. However, there is an apparent grouping of these caves according to elevation of major sections, as shown in Table 1. The lowest caves are the present conduits, such as To'uri cave, and the East Tamarua conduit below Tautua cave, as well as the Lake Tiriara and Ivirua conduit caves described. These contain an active streamway, few formations, and heavy clay deposits resulting from soil erosion off the volcanic cone. The elevations are close to sea-level, and they resurge at the coast as seen at Avarua landing, or on the reef flat as at Tamarua. Tuitini, Touropouru, Tautua and Taenoterau caves all have major level sections between 10 and 20 m above sea-level. This suggests that they were conduit drainage caves corresponding with the 14.5 m elevation of Pleistocene 20 “Ol aanbiy (Ww) aue}siq WW: a[E9S == el of oO 0 (Ww) yydaq NW — 41140dd JAYS UaAe 1y4b.ifep aouesjua au Lop auLjop uada aouesjUua a | [Op qs apesg — ides W 6 - Yyj}daq cee — wW 9s | yj6uaq Notes ot = Af MG'eSodSl (S.2'SGolzZ a9 AJA N¥1d JA¥I spue|s| Joo] “elebue, JAVI ANNAWANLIAld 21 MAN 36 CAVE Mangaia, Cook Islands GR 21°55.7'S; 157°57.4'wW Depth +3 m Alt.c. 40 m archaeological excavations (P. Kirch, 1991) +7? 8mclimb to entrance Grade 5b makatea edge CAYE PROFILE entrance o KEY M.N. oO Column ] ° Boulder & Pool Depth (m)} Oo 120 100 80 60 40 10 Ny Distance (m} Scale: m Figure 11. 22 Table 1. Dimensions and elevations of Mangaia caves (m) Cave Length Altitude Depth Elevation of Height explored _of entrance level sections __of rift To'uri 489 3 -3 1 20-30 Tuitini 830 23 -13 10 & 20 20-25 Touropouru) 587 16 +4 10& 18 15 Tautua 335 fe) +6 15&19 10-20 Taenoterau 57 20 -5 Ue 5-10 Teruarere 791 43 -24 20 & 30 10-30 Erua 177 c. 40 -8 c. 40 20 Kauvava 358 c. 60 -38 c. 40 20+ Man 88 94 c. 40 +3 c. 40 10-15 Piriteumeume 136 c. 60 -9 ©2595 4-15 23 limestones, probably formed at the same time as a relative sea-level 20 m higher than present (Woodroffe et al., 1991). Reef groove-and-spurs formed at this time are now 2 to 11 m above sea-level (Stoddart et. al, 1985). Thus it is indicated from three corresponding features that sea-level was in this position for considerable time. Teruarere cave is a major feature of Mangaia, a minimum of 791 m of continuous level passage ranks with the spectacular find of Ana Maui by the Tonga '87 caving expedition on 'Eua (the cave named after the Tongan demi-god) (Lowe, 1988). It declines from 30 to 20 m above sea-level, a simple fossil streamway, corresponding with the 26-27.5 terrace described by Stoddart et al. (1985). The terrace is deeply dissected, and was believed to represent a very old sea-level feature. Similarly, Teruarere cave has a number of high daylight avens, indicating long-term dissolution and collapse. The pitch at the end of the surveyed section indicates that one of these cut down to a deeper level after uplift of the main passage. Erua, Kauvava and Man 88 caves all occur at about 40 m above sea- level, and could correspond with the 34-39 m terrace identified by Schofield (1967), Wood and Hay (1970), and Ward et al., (1971), and shown by several profiles of Stoddart et al., (1985). However, this feature is so old that they must have eroded dowm from their formative elevation. While Erua and Man 88 caves are closed by calcite deposits, the pitch beneath the 23 m aven at the northern end of Kauvava cave indicates that flow through here allowed the cave to cut down to a deeper level after uplift. The highest elevation cave in this study is Piriteumeume, which has all the features of an old cave, with large formations, collapse and irregularity, and dryness. During development of the cave the dolines would have drained a higher surface, and the karst corridors leading to the entrance would also have been caves. The makatea surface has many such small relic caves to be found in the karst corridors, though the caver will probably - get lost finding them! CONCLUSIONS Cave systems are a relatively neglected aspect of study of raised limestone islands, possibly owing to the specialised techniques of survey, and potential dangers. This study shows that they can make a contribution to the knowledge of limestone geomorphology and changes in relative sea- level . Features of the caves explored indicate the validity of the principles of makatea cave development outlined in the introduction. It is apparent that caves of lower elevation are of larger dimensions than those higher. This could be because parts of the higher caves have been lost with time, but as well as length the lower caves have higher rifts or vadose slots. This could result from slower rates of island uplift that allowed the streams to downcut their caves, or it could indicate the sea-level changes of the Pleistocene that have caused multiple reoccupations of these caves. 24 ACKNOWLEDGEMENTS This research was carried out as part of the project "Anthropogenic Environmental Change, Agricultural Intensification, and Socio-Political Evolution in Polynesia", funded by the NSF Grant BNS-9020750, to Patrick V. Kirch, Principal Investigator. | should like to thank P. V. Kirch and D. W. Steadman for the opportunity to carry out this work, and their encouragement and comments. Thanks are given to the Government of the Cook Islands for their assistance in this project, particularly Tony Utanga, and also to Ma'ara Ngu and Tuara George for field assistance in Mangaia. REFERENCES Buck, P.H., 1934. Mangaian Society. B.P. Bishop Museum Bulletin, 122. Dalrymple, G.B., Jarrard, R.D. and Clague, D.A., 1975. K-Ar ages of some volcanic rocks from the Cook and Austral Islands. Geol. Soc. Am. Bull., 86: 1463-1467. Ellis, B.M., 1976. Cave surveys. In T.D. Ford and C.H.D. Cullingford (Editors), The Science of Speleology. London, Academic Press, 1-10. Gill, W.W., 1894. From Darkness to Light in Polynesia. London, The Religious Tract Society, 383 p. Haq, B.U., Hardenbol, J., and Vail, P.R., 1987. Chronology of fluctuating sea levels since the Triassic. Science, 235, 1156-1167. Lowe, D.J., 1988. ‘Eua Island Tonga ‘87 Expedition Report. Unpublished, 25); Mylroie, J.E. and Carew, J.L., 1988. Solution conduits as indicators of Late Quaternary sea level position. Quat. Sci. Rev., 7, 55-64. Schofield, J.C., 1967. Pleistocene sea-level evidence from the Cook Islands. J. Geosci. Osaka City Univ., 10, 118-120. Steadman, D.W., 1985. Fossil birds from Mangaia, southern Cook Islands. Bull. Br. Orn. Cl., 105, 58-66. Steadman, D.W., 1986. Two new species of Rails (Aves: Rallidae) from Mangaia, Southern Cook Islands. Pacific Science, 40, 27-43. Stoddart, D.R., Spencer, T. and Scoffin, T.P., 1985. Reef growth and karst erosion on Mangaia, Cook Islands: A reinterpretation. Z. Geomorph., N.F. , 57, 121-140. 25 Thompson, C.S., 1986. The climate and weather of the southern Cook Islands. N.Z. Met. Service Misc. Publ., 188 (2), 69 p. Ward, W.T., Ross, P.J. and Colquhoun, D.J., 1971. Interglacial high sea- levels, an absolute chronology derived from shoreline elevations. Palaeogeogr., Palaeoclimatol., Palaeoecol., 9, 77-99. Wood. B.L., 1967. Geology of the Cook Islands. N. Z. J. Geol. Geophys., 10, 1429-1445. Wood, B.L. and Hay, R.F., 1970. Geology of the Cook Islands. N. Z. Geol. Surv. Bull., n.s. 82, 103 p. Woodroffe, C.D., Short, S.A., Stoddart, D.R., Spencer, T. and Harmon, R.S., 1991. Stratigraphy and chronology of Late Pleistocene reefs in the southern Cook Islands, South Pacific. Quat. Res., 35, 246-263. Yonekura, N., Ishii, T., Saito, Y., Maeda, Y., Matsushima, Y., Matsumoto, E., and Kayanne, H., 1988. Holocene fringing reefs and sea-level change in Mangaia Island, Southern Cook Islands. Palaeogeogr., Palaeoclimatol. Palaeoecol., 68, 177-188. 3 100s inte ‘ad? 26 penis eimitoeseT 3061 Be AMER RIN’ atheiy sai aba Reaieripion Aig me wo “4 Bn ; re : } : ay - { 5) te i ’ iat \ i Rit } ae rf yA Ni \, 5 Rey ‘Ay eee” oy i a abs aw, ear : : 7 oa ; j im wey sn P ue i ~ 1 Vigne T te 9) Bai, NBER Se e062 el e . yibeurey WEe-AOrles, Oat Re patio theo or + se iReB arts, aioe Ws) ASCHUDNOpHS eRe, oatnil igrala f hitibraiou Ho “ve bhai 2387 fe Se an for =e rate. "ta y ot ht aan ae eg Sri siecle hose Aone i Mernics teeta “a hi Gye Me pod oc iy: Tor ie $y 165 hen 4, Cacrgy ton Kee Sepaienenins: tie eyed ae eon ihe So + +, My Ne ies we I Wg Sng) 4 >. x ATOLL RESEARCH BULLETIN NO. 418 SHALLOW-WATER SCLERACTINIAN CORALS FROM KERMADEC ISLANDS BY VLADIMIR N. KOSMYNIN ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 v WW MeOH VF SINIOATY - jay Yad AbesUA MORE BIAROD KA SHALLOW-WATER SCLERACTINIAN CORALS FROM KERMADEC ISLANDS BY VLADIMIR N. KOSMYNIN INTRODUCTION Shallow-water scleractinian corals were collected in the Kermadecs (South-West Pacific) in April- May of 1990 during the 18th cruise of the USSR Academy of Sciences’ R/V "Akademik Alexandr Nesmeyanov", which sailed from Vladivostok to the Western Pacific. The objective of this expedition was to study the influence of volcanic gas and hydrothermal activity on sublittoral communities. Dredging operations were performed on the slopes of Kermadecs: underwater observations were made near Curtis Island and corals were collected at several points near Raoul Island and nearby islets using SCUBA (fig. 1). Despite numerous previous visits to the Kermadec Islands by research vessels, very little information has been reported on the scleractinian fauna. The presence of corals in bottom communities in this area was mentioned by Nelson and Adams (1984) and coral communities were noted when Shiel et al. (1986) described the sublittoral zonation, but neither publication provided a list of species. The last publication dealt directly with corals from the Kermadec Islands was that of Vaughan (1917). This paper was based on the material collected early this century by R.W.B.Oliver, former director of the New Zealand National Museum of Natural History, who described the Kermadecs in several publications. Material from this collection are deposited in the Museum of New Zealand (MNZ) in Wellington and at National Museum of Natural History (NMNH) in Washington. Some additional Kermadec corals collected and kept in the New Zealand National Museum were identified by D.F.Squires. Soviet expeditions also worked there, but their results are not yet published. A SHORT DESCRIPTION OF SITES OF CORAL COLLECTING Near Raoul Island two sublittoral areas were studied: one north of Blue Lake and another in the southern part of Denham Bay. Meyers and Napier Islets were also studied and corals were collected (Figure 1). North of Blue Lake at the depth of 10-15 m the foreshore is characterized by a gently sloping floor. The floor shows an alternation of areas covered by ripple sands with separate boulders and patches of boulder pavement, among which some boulders that are 1-2 m high. Unlike the barren ripple sands, bottom communities on large boulders are rich and include fleshy algae as well as calcareous geniculated Corallina on upper surfaces and incrusting Lithothamnium under overhangings, and many sea urchins that include at least two species. Corals are abundant there and are mostly encrusting forms of Hydnophora, Goniastrea and Montastrea 50-60 mm in size on top surfaces of boulders, and Dendrophyllia under overhangings (Tabl. 1). Life is less diverse on the Moscow State University, Geographical Faculty, 119899 Moscow, Russia Manuscript received 10 November 1993; revised 5 December 1993 2 surface of boulder pavement in comparison to that found on large boulders probably because of sand abrasion. The foreshore in the southern part of Denham Bay is considerably different from that described above. It is steeper and at a distance of 50-70 m from the shoreline the depth is 15-20 m. The slope is composed of large basalt blocks, each several cubic meters. They are arranged on the slope either in dense groups or scattered individuals. Coarse-grained sand and gravel are accumulated between boulders. These boulders support rich flora and fauna on their tops, whereas at the base, where light is weaker and sand-gravel erosion is stronger, the boulders are poorly inhabited. Along with macrophytes, these rich communities contain a considerable number of corals, primarily colonies of Turbinaria bifrons, which form large flat coralla up to 1.5 m in diameter. There are also abundant encrusting colonies of Goniastrea australiensis with a corallum size up to 20 cm, and smaller coralla of Montipora, Leptoseris and Cyphastrea. These corals cover no more than 10-20% of the surface even at the top of boulders. The maximum number of corals appears to occur at depths of 13-15 m and decrease rapidly up the slope, probably because of the increasing abrasive effect of the resuspended sediments. The richest coral communities were discovered on the western side of the Meyers islets opposite a channel separating them. Here at depth of 22-24m there is a terrace that looks like a wave cut platform. It may be in fact a submerged coast-line. On the surface of the terrace there are thin lenses of coarse-grained sand and gravel, as well as individual colonies of Turbinaria radicalis with coralla up to 0.5m in diameter. Rocks of different height occur on the surface of the terrace, some of them reach a depth of 10-12 m, while others extend up to almost the surface. The rocks contain very rich communities in which corals play an important role sometimes covering up to 25% of the surface. Life forms of corals are mainly encrusting such as Goniastrea, Leptoseris and Montipora, or massive or columnar such as Hydnophora exesa. Branching forms are represented by sparse small colonies of Pocillopora damicornis, which inhabit the tops of boulders. Turbinaria radicalis and Turbinaria bifrons form the largest colonies and are more noticeable than other corals. Hydnophora exesa form the second largest colonies, with coralla up to 40 cm in diameter. On vertical cliffs there are frequent Dendrophyllia sp. together with gorgonian corals. The latter grow abundantly at a depth of more than 25 m, and up the slope on more gently sloping areas there are abundant alcyonarian corals. In the communities at the depths of less than 25 m other sedentary animals are numerous along with the corals, including sponges, tunicates, hydroids, sea lilies, sea urchins and the notorious sea-star Acanthaster planci. Similar communities were noted also near Napier island. The Kermadec Islands are also characterised by abundant deep water scleractinian corals at the depth of 375-1000 m. In particular, a large quantity of fragments of Goniocorella dumosa (Alcock,1902) were dredged from the depth of 1000 m by trawl at 30°28’0”S 178°37°2”’W, as well as corals such as Madrepora vitiae Squires & Keyes,1967 and Flabellum gracile (Studer,1877). DISCUSSION The deep water scleractinian corals mentioned above are related to the fauna from New Zealand, whereas the shallow water Kermadec corals are linked with the Tonga and Fiji Islands to the north and the Norfolk Islands on the west. There is a certain affinity of the latter with the Great Barrier Reef of Australia through the Norfolk and Lord Howe Islands but the diversity is considerably lower. Further studies will undoubtedly increase the list of corals for the Kermadec Islands. Some fossil corals were also noted by Vaughan (1917): Leptoria phrygia (Ellis & Solander,1786; USNM 93896), Alveopora sp. (USNM 93898), Acropora sp.(sample I failed to find in USNM) and Cynarina 3 lacrymalis (Milne Edwards & Haime,1848) (USNM 93899), which may quite possibly be alive on Kermadecs. The low diversity of reef-building coral species off the Kermadecs is probably due to several reasons. I tend to agree with Schiel et al.(1986) that it is not a result of competition for space with other organisms, in particular algae. Among the main factors are the relative isolation of the islands, their young age and recent activity of volcanic processes as well as wave erosion (abrasion). The development of coral reefs near Raoul Island is limited not only by low water temperatures but more by relief and coastal geomorphic processes. Quick wave erosion of volcanic shores and mobility of floor substrates limit the possibilities for establishment of reef building corals. The coral fauna adjacent to Raoul Island small islets is richer than the coral fauna of Raoul Island itself. The precipitous rocky foreshores of small islets almost lack sediments. For corals it is more suitable to settle and grow there than on the relativly gentle slopes of Raoul Island where the mobile coarse bottom sediments limit coral growth. The water temperatures near the Raoul Island are not optimal for reef corals but does not fall to the lethal winter limit of 15°C for the majority of hermatypic corals. The summer temperatures, which are about 24°C, is close to normal for coral reproduction. The establishment/recruitment of new species of corals on Kermadec Islands is quite possible. It can happen not only by the distribution of planula by but also by rafting, for example, by fragments of pumice-stone carried by currents from northern volcanic islands. The discovery of fossil colony fragments of such species as Acropora and Alveopora (Vaughan,1917) indicate that some species could have established in this area but the conditions are not favorable for survival of new populations. In particular, their survival is considerably limited by abrasion and mobility of sediment material, and by volcanic activity. ACKNOWLEDGMENTS The author is grateful for the support provided to me: by Dr. Vitaly Tarasov, the head of the expedition on the "Academik Alexandr Nesmeyanov"; Dr. Alan Baker, the director of MNZ, Wellington; other collegues of the MNZ for the opportunity to work with their collection; Dr. Paul Gillespie (Cawthron Institute, Nelson, NZ) for computer search of data on the corals from the Kermadec Islands; Dr. Ian Macintyre and Mr.Chad Walter (NMNH) for reviewing of manuscript. I would like to thank especially Dr. Stephen Cairns for reviewing of the article and for advice on work with the NMNH coral collection. LITERATURE Nelson W.A.; Adams N.M. 1984: Marine algae of the Kermadec Islands. National Museum of New Zealand occasional series 10: 1-29. Schiel D.R., Kingsford M.J., Choat J.H. 1986: Depth distribution and abundance of benthic organisms and fishes at the subtropical Kermadec Islands. New Zealand Journal of Marine and Freshwater research. Vol. 20: 521-535. Vaughan T.W. 1917: Some corals from Kermadec Islands. Trans. N.Z. Inst., vol.49: 275- 279. Table 1. The list of shallow water scleractinian corals from Kermadec Islands. Species name Raoul Hydnophora exesa (Pallas, 1766) Goniastrea favulus (Dana,1848)* Goniastrea australiensis (Milne Edwards & Haime,1857) —18,19,23 Leptoseris mycetoseroides Wells,1954 Leptoseris hawatiensis Vaughan, 1907 Cyphastrea serailia (Forskal,1775) Montastrea curta (Dana,1846) 1-10-3,8 Montipora cf. millepora Crossland,1952 Montipora cf. spumosa (Lamarck, 1816)** Pocillopora damicornis (Linnaeus, 1758) Plesiastrea versipora (Lamarck, 1816)** Turbinaria radicalis Bernard,1896 1-10-22 Turbinaria bifrons Bruggemann, 1877 Dendrophyllia sp. 1-10-24 * according to Shiel et al.,1986. North coast 1-10-4,5,6,7,9 1-10-4,7,17, Denham Bay 3-14-4,8,11, 13,14,17,18 3-14-7,15,16,24 3-14-4 3-14-5,6,9,12 3-14-22 Napier Islet 2-22-1,2,6,11,13 2-22-3,15 2-22-10,12 2-22-3 2-22-4,9 2-22-8 2-22-5 2-22-14 ** according to Vaughan,1917. Comments to some of these species see in text. Meyers Islets 4-5(15)-7,8 4-5(15)-1,2,15,19 4-5(15)-10 4-5(15)-13 4-5(15)-3 4-5(15)-12 4-5-30,31,32 4-5(15)-4,5,6,17 4-5(15)-14 4-5(15)-16 1-10-3 - Paleontological Museum of Moscow recent scleractinian corals collection number. First number is the site of collecting (the same as on the chart, fig.1); second number is depth of collecting (if 5(15) - it means range 5-15 meters); third is number of sample from this site. Dublicates of these spesimens with the same indexation were transmitted to MNZ in Wellington. New IN Caledonia Napier Isl Hutchinson Blue Lake @ (\ Raoul Island 29°18'S Wok W Figure 1. A. The location of the Kermadec Islands, north of New Zealand. B. Enlarged map of Raoul Island in the Kermadecs showing smaller ISMeeS with, Sibesioer coral collecting (1) = 4). . é Ta : _ ai of “ 3 % e 5 v - y 2 ‘i ih % ; ' it { . J hal a " - 1 ), a , + Lay ‘" i Vary a” f ae iq “ m) oe} ‘ ’ i U ea Ve ; 5 > ? = > 5 V po ‘ nec A it i en H - \ cw! j ; 3 hy é be y at rienibarieeiad dee alae? “a aoe ri fabs 4 & i 3 - ’ -~ . ” ‘ " ees H - # hast 4; i ca “* , \ Rite we init 7 2 e f ; : ‘ — 7 o Win rg + } A is ce ‘s bi 1 { tas ¢ , , + { s an , ‘ fe f Ys ¥ aNtae. ) ; = Tt ’ n 5 . +4 ry a £ Wed “i 4 ~ ae r i ) ' H | | iS ; F i we = wt t 4 iv | : i j ie be iP A gr ayesn — Deal ae [Sat * i ‘ 7 y 4 : . v -— 4 4 > i ‘ ‘ Gi “oy 2 P< ' 4 ; { ; 7 1 bee a 2m oy ror wi? . ‘, ‘7 . ey 7 ssh ms \ 7 . E : : fF cides aio on =m ave PON em RE, dimen —r Deyn a ae eee ea Le YU 2 cya ela ee meal @ Ce SY, ety : hi ; cS nad y * f i a i. } . * ee “i , ait ee * i % mt! i : ‘ S108 badd ISD : y sr a ae , i’ : 7 sett ~ oe ys i ad * t ; P ws ’ Let i i i . 4 ’ a ie a 4 \ ’ ¢ Ca \ ' fen) year ¢ { 4 4 - - i eit 4 F dy th 4 5 i ny 1 ; Hehe] , . ' 3 :- { j ' ' eae roa: i a wre f } j % | er ‘= > i oa ‘ “y 5 roe ety 2 ar NS s i t . oes % ; is Se 5b Ais F bi i | my i t P ke - “ : ran 4 2 = Yr tt * 2% Mi. Uy my death ? Ki iy pets i x vy 2 t > . y & ‘ nh “ ~ * he Pie he i 4 y ci ann D SND ay a ATOLL RESEARCH BULLETIN NO. 419 DESCRIPTION OF REEFS AND CORALS FOR THE 1988 PROTECTED AREA SURVEY OF THE NORTHERN MARSHALL ISLANDS BY JAMES E. MARAGOS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 AATDSTOAT SRY TAT HOVE * CGV AST AIAN eR AM VA) PO AR AT U4 ne A902 OVS @RATAIO-Vi FATHOM BHT AO VaVAUE AMAL ~~ : ae a 7 4 DESCRIPTION OF REEFS AND CORALS FOR THE 1988 PROTECTED AREA SURVEY OF THE NORTHERN MARSHALL ISLANDS BY JAMES E. MARAGOS Abstract The Republic of the Marshall Islands requested a natural and cultural biodiversity survey of 6 northern atolls (Bok-ak, Pikaar, Toke, Wotto, Rondik, Adkup) and one reef island (Jemg) which was accomplished over 17 days in September 1988. This report covers the results of the survey of the reefs and corals during the expedition. Ninety-five marine sites were snorkeled and the shorelines of all island were surveyed during the expedition. A total of 168 species and 55 genera and subgenera of stony corals were reported including several new species and one new genus recorded (Polyphyllia) for the Marshalls. Bok-ak Atoll, the northernmost atoll, supports large giant clam populations, a completely native flora, and the largest seabird populations in the Marshalls. Pikaar Atoll also supports large giant clam populations and the largest sea turtle nesting populations in the Marshalls. Both Bok-ak and Pikaar are isolated from other atolls and have shallow lagoons elevated slightly above sea level due to their geomorphological configuration. Toke Atoll is located about 10 km from inhabited Utrok Atoll, and supports healthy coral reef habitats and giant clams. Jemg Island supports large seabird populations and is the second most important sea turtle nesting site in the Marshalls. Boat access to Jemg, Pikaar and Bok-ak is hazardous due to wave exposure or strong currents. Rondik Atoll supports healthy coral reefs, blue coral habitats, pink foraminiferan sand beaches, and large coconut crab populations. Adkup supports abundant seabird populations, sea turtle nesting populations, and healthy coral reefs. Inhabited WOtto also supports healthy coral, coconut crab, sea turtle and giant clam habitats and has beautiful beach, reef and lagoon habitats. On the basis of the surveys, Bok-ak, Pikaar and Jemg are recommended for designation as limited entry ecological preserves. Toke is recommended as a national park accessible to both tourists and residents. Limited sport diving and beach-going is also suitable for Rofdik and Adkup. Assistance should be provided to the people of Wotto Atoll to fulfill their desire for small scale adventure tourism. Program on Environment, East-West Center 1777 East-West Road Honolulu, Hawaii 96848 U.S.A. Manuscript received 8 March 1990; revised 30 August 1994 II. IV. TABLE OF CONTENTS Introduction Materials and Methods Results Jemo Reef Atoll Geomorphology and Oceanography Bok-ak Atoll Pikaar Atoll Toke Atoll Wotto Atoll Rondik Atoll Adkup Atoll Corals: Combined Species List Discussion Marine Reserves Marine Parks and Recreational Areas Subsistence Activities Radiological Contamination Mariculture Agriculture Commercial Fishing Small Scale Tourism Variations in the Depths of Living Reef Flats Rapid Marine Field Assessment Procedures Summary Acknowledgements References Appendix A Appendix B Appendix C Appendix D Table i. Table 2. Table 3. Table 4. Table 5. Table 6. Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. LIST OF TABLES Physiolographic data on the atolls and table reef visited in 1988 Check list of corals observed in the Northern Marshalls, 1988 Relative diversity and distribution of stony coral genera and species observed in 1988 Previous coral records from the RMI atolls studied in 1988 Criteria for evaluating candidate reef areas as marine protected areas and parks in the RMI Relationships between various types of resource uses for Northern Marshalls reefs and atolls and the required criteria for their feasibility LIST OF FIGURES Northern Marshall Islands Map of Jemg Island and Table Reef Reef features at Jemg Reef Close up of Jemg Island and Nearby Reefs Map of Bok-ak Atoll showing sites of Sept. 1988 marine surveys Southern tip of Bok-ak Atoll Bok-ak Atoll in the vicinity of the cluster of large eastern islands Bok-ak Atoll in the vicinity of the western passage Bok-ak Atoll in the vicinity of the NW reef tip Pikaar Atoll showing locations of marine survey sites Detailed map of the single western passage at Pikaar Atoll Pikaar Atoll in the vicinity of the southern tip and island of Pikaar Pikaar Atoll at the eastern end Pikaar Atoll at the northwest tip Shoreline changes between 1944 and 1978 at selected locations at Toke and Pikaar atolls Toke Atoll showing locations of marine survey sites Toke Atoll, NW corner in the vicinity of a shipwreck NE tip of Toke Atoll SW side of Toke Atoll in the vicinity of 2 marine survey sites and the deep passage Detailed map of deep passage off western Toke Atoll SW corner of Toke Atoll in the vicinity of the largest islands 19 21 22 23 28 29 30 31 32 34 35 35 36 37 39 41 42 43 44, 44 45 Figure 272. Figure 23 Figure 24. Figures. Figure 26. Figure 27. Figure 28. Figure 29. Figure 30. Figure 31. Figure 32. Figure 33. Figure 34. Figure 35. Figure 36. Figure 37. Figure 38. Figure 39. Figure A-1. Figure A-2. Figure A-3. Figure A-4. Figure A-5. Figure A-6. Figure A-7. Figure A-8. Figure A-9. Figure A-10. Figure A-11. Figure A-12. Figure A-13. Figure A-14. LIST OF FIGURES - cont’d Map of Wotto Atoll Defense Mapping Agency Chart of Wotto Atoll Map of NE Wotto Atoll sketched from 1978 color aerial photos Map of southern Wotto Atoll sketched from 1978 color aerial photos Rondik Atoll showing locations of marine survey sites SW corner of Rondik Atoll showing the main south passage and the location of 2 marine survey sites Western end of Rondik Atoll showing Bock Island Main island cluster of NE Rofdik Atoll Shoreline changes between 1944 and 1978 at selected locations along NE Rondik Atoll Map of Adkup Atoll NE peripheral reef of Adkup Atoll NE rim of Adkup Atoll near a windward passage NW rim of Adkup Atoll in the vicinity of a pass and one marine survey site NW rim of Adkup Atoll near Aneko (Enego) Island SE end of Adkup Atoll near the main islands and pass Southeastern end of Adkup Atoll showing sites of some marine surveys Possible adverse effects of cutting channels through semi-enclosed atoll lagoons Progressive development of reef flat features on atolls Bok-ak Atoll reef profiles: lagoon and pass reefs Bok-ak Atoll reef profiles: eastern perimeter reefs Bok-ak Atoll reef profiles: western perimeter reefs Pikaar reef profiles: perimeter reefs Pikaar reef profiles: lagoon reefs and perimeter reef Toke Atoll reef profiles: perimeter reefs Toke Atoll reef profiles: western perimeter and lagoon reefs Jemg reef profiles: ocean slopes Wotto Atoll reef profiles: north and eastern perimeter reefs Wotto Atoll reef profiles: eastern and southern perimeter reefs Rondik Atoll reef profiles: southeast perimeter reefs Rofdik Atoll reef profiles: northeast perimeter and lagoon reefs Adkup Atoll reef profiles: southern perimeter reefs Adkup Atoll reef profiles: northern perimeter reefs I. INTRODUCTION Oceania represents one of the last frontiers for the assessment of biological diversity, especially for shallow water marine ecosystems. Coral reefs are among the most widely distributed ecosystems on the face of the earth, and within Micronesia they dominate in terms of area. They also provide critical physical and ecological support to most other ecosystems including those of low coral islands. Because of remote access, geographic isolation, and the physical limitations of underwater surveying techniques, most marine areas in Micronesia remain unstudied. Yet assessment of the ecological and biological importance of reef areas for conservation, subsistence, recreation, visitor and commercial uses must require on-site surveys to some extent. With many thousands of reefs and islands yet to be evaluated, new techniques must be employed to allow a rapid but technically adequate evaluation of reef sites. Existing regional evaluations (IUCN, 1989; 1988; Dahl, 1980) provided valuable information on many areas with respect to park and reserve potential, but the emphasis has been placed on terrestrial (island) ecosystems which are easier to visit and survey. The Marshall Islands study of September 1988 offered a unique opportunity to accomplish a marine oriented regional survey of reefs using a non-conventional rapid field assessment technique relying on a combination of field observations, teamwork, aerial photographs, underwater photographs, available maps, and interviews with knowledgeable islanders. The study is primarily based on the results of a three week expedition to six atolls and one table reef in the northern Marshalls during September 1988. The areas surveyed were the atolls of Bok-ak, (Taongi, Bokak, Pokak), Pikaar (Bikar), Toke (Taka), Adkup (Erikub), Roidik (Rongerik), and Wotto (Wotho) and the table reef of Jemg (Table 1, Fig. 1). The name spellings used above reflect the most current official RMI linguistical determinations. Those in parentheses above reflect spellings commonly used in the past. The field expedition concentrated on evaluating the following categories of resources with respect to potential justification and interest in protected area designation; names in parentheses refer to the expedition team members responsible for collecting information about the resources: island vegetation (Derral Herbst), seabird nesting, resting and feeding (James Juvik), other terrestrial animals (James Juvik, Peter Thomas), turtle nesting and feeding habits (John Naughton, James Maragos, Peter Thomas), nearshore reef fishes (John Naughton, R. Virgil Alfred and Paul Maddison), giant clams and other edible shellfish (John Naughton, R. Virgil Alfred, James Maragos), * coral and reef features (James Maragos), * pelagic fisheries and marine mammals (Paul Maddison, R. Virgil Alfred, and John Naughton), * * * = See * cultural historical and archaeological resources (Charles Streck Jr.), and * tourism, park and reserve feasibility (Peter Thomas) The present report describes the results of the surveys of reefs and corals with emphasis on ecology and related oceanographic and geological characteristics. It is meant to serve as a technical supplement to the report prepared by Thomas et al (1989), a summary of all survey results. cOl Crt X 8I 6S oS9T 90 01 ve $6 tl so OnOM 0 Cri X StI OV -691 LO IT cS 6 ¢ LS0 oxOL 0 8I X 8I IToLIL AColl 6S rl (E)| IT? yIpuoy 0 LX CI .LOoOLT SloTl LS'Le 9 60 Teeyld 0 tX8 OF 5691 90,01 0 I 810 Oulog 0 8 X 81 LS 9891 Votl BE 8L IT LLt ye 0 ve X LC =, TIOly dV 139NOU 4 20! Cia: ee TNNOLVY INIMIG edl bl 20 Ill. RESULTS General descriptions of the reef and island systems of the seven areas are found in Fosberg (1988) and Fosberg et al (1956). Maps and additional geological information are reported in MacNeil (1969). The six atolls (Bok-ak, Pikaar, Wotto, Toke, Rondik, and Adkup) are generally small compared to the average size of atolls in the rest of the Marshalls, in terms of island land area and lagoon area (Table 1). The island of Jemo iS also smaller than the average size of islands on the other four table reefs in the Marshalls (Mejit, Kili, Lib, Jabat). All areas visited are not permanently inhabited except WoOtto. Jemg, Toke, and Adkup are occupied for brief periods during seasonal harvesting of fish, turtles and their eggs, coconut crabs, or other resources. The other areas (Rofdik, Pikaar, and Bok-ak) appear to be visited less frequently based upon our 1988 field observations. A combination of factors discourages permanent occupation of the uninhabited areas, including limited fresh water supplies and rainfall, poor soil conditions for cultivation, remoteness from nearby population centers, difficult or hazardous boat access to main islands, and perhaps greater vulnerability of the small islands to exposure from storm waves and surges. General physiographic data on the seven areas are summarized in Table 1. Reef profiles of most stations are found in Figures A-1 to 14 in Appendix A, and maps of many reefs and islands constitute Figures 2 through 37 in the text. Climate and Oceanography The ocean in the region of the northern Marshalls is between 4,500 and 5,400m deep (Fosberg et al, 1956). The northern Marshalls are semi arid and experience less than average annual rainfall compared to atolls and islands more to the south. The two northernmost atolls, Bok-ak and Pikaar are the driest Marshall atolls, (excluding Enen Kio (Wake) which is under U.S. jurisdiction, drier still, and further to the north). Of the six atolls and one island surveyed in September 1988, Adkup, which is situated in the central Marshalls, is the wettest of the group. The dryness limits groundwater and vegetation development, and Fosberg et al (1956) divides the Marshalls into several vegetative zones (see Figure 1). The northern atolls are exposed to stronger tradewinds and associated wave action. Although tropical storms and typhoons tend to spawn in lower latitudes further to the west, sometimes the storms gain intensity, move into the Marshalls, usually from the south, and cause extensive damage to shorelines, islands and some reefs. Even infrequent storms can modify the distribution of islands on atoll reefs with long lasting effects, as reported for Arno Atoll (Wells, 1951). The typhoon frequency in the Northern Marshalls is of the order of 50 to 100 yrs, and the visible results of typhoons, especially on atoll islands are the record of at least a thousand years or more (MacNeil, 1969). The major tropical current system in the northern Marshalls is a large westward flowing current between latitudes 10 and 20 degrees, north termed the North Pacific Equatorial Current (NPEC). This current mostly affects deep ocean circulation patterns off shore. Nearshore effects of the NPEC are masked by much stronger but localized currents caused by the tides, winds, and wave action. 21 JEMO REEF (Figures 2-4, and A-8) Jemo is the only reef of the seven visited that is classified as a table reef and is one of only 5 table reefs (compared to 29 atolls) located in the Marshall Islands. A table reef consists of an isolated flat topped coral reef which reaches the sea surface but which lacks a lagoon (MacNeil, 1969). These reefs tend to be small, sometimes linear, and are exposed to wave action due to the lack of sheltered lagoon reefs. The table reef supporting Jemg Island is 8 km long, slightly arcuate, and is situated along a SW to NE axis (Figure 2). Jemg Island is egg shaped and about one-third mile long. Fig 2. Jemg Island and table reef showing sites of marine surveys, September 1988 Stars indicate reef sites of special 40.7 s interest. Scale: 1" = .5 mi. age sate -“ Submerged reef... aoe @stidllow reef flajy’ Exposure of the reefs and island to heavy waves and storms from virtually any direction has controlled and shaped reef development at Jemg. Underwater observations, published charts, and color aerial photographs all document that the flanks of the table reef drop off near precipitously to great depths within a kilometer of the reef crest. Shallow reef flats emerge at low tide only at the southwest end upon which rests the single island of Jemo (Figures 2-4). Elsewhere, an extensive system of sand covered surge channels (see Figures 3, 4) traverse the reef crest in a north-south axis and at depths of 2-4m. At the NE end, which receives the most exposure from trade-wind waves, the reef resembles a rounded knob in which the surge channels give way to well developed spur-and-groove formations. Elsewhere the outer margin of the reef crest consists of flat pavement-like and heavily scoured ramps descending at a moderate angle from a depth of 2-3m to a drop off at a depth of about 6-8m. Below the drop off, the reefs are steep vertical walls sometimes overhanging (Figure A-8, sites 4B, 4L, 4E, 4J, and 4D). Perhaps due to the long NW facing axis of the reef oriented away from the prevailing NE tradewinds, typical spur and groove formations are lacking along the reef margin except at the NE end. These features, along with the prominent series of sand bottomed surge channels across the reef crest, are clearly displayed in the 1978 color aerial photographs of the island and reef at Jemg taken at scales of 1:30,000 and 1:8,000. 22 “O00SOE:T P1295 ‘“sydeubozoud [eltuae wolod g/6T Wous uMeup dey $994 94} UO SjauUeYD abuns puke SjJaa4 MOLL PYS JO UOLJNGLuisLp HuLMOYS jJaa4 a1qez pue puelsy bwar S[auueyd obuns —--. amie aie _ PaaAOD pues jO UOLzeYUaLUO . va ¢ ~ s JUL UOLZeZaH9A pue[st auL[asoys ydeaq pues (MIW) SUL, Jaau JnojUod wore} OT oe te a, -- ~ --- ct | Mh | “S949 ya ays “e ounbly 000°8:T 2Le9S -soqoud [eLuae 4OLOd B/G wos umeup dew “Ssjaau Aqueau pue pueis; Swar yo dn asol) Sjauueyd abuns BUL| UOL}ZeZAHAA puelst 0 02e¢ we © 0 og aUL[a4OYsS Yeaq pues 4nojuo0d woyurxes OT eee x ee lees 24 There are no major accumulations of loose boulders strewn on the reef crest. Small boulder beaches occur along the south and southeast facing shorelines of Jemo Island. The reef margins off the elongate SE facing reef axis appear to be depositional in character, with sand, gravel and other sedimentary materials apparently carried through the surge channels over the crest of the reef from the other (NW) side. Thick accumulations of sediments were observed at the base of canyons between steeply sloping massive rounded reef buttresses along the SE reef margin. In contrast, the NW reef margin lacked any loose sediment, and coral formations consisted of robust, low profile colonies. The bare reef pavement ramps appeared heavily scoured. Ailuk Atoll affords Jemo reef some protection from northeast swells, Likiep, provides protection from the southwest, and Wotje and Adkup provide protection from the southeast, leaving the northwest face of Jemg reef the most exposed to heavy seas. Movement of reef materials appears to be from a northwest to southeast direction, with the northwest face of the reef potentially more exposed to damage from wave energy. Another unusual feature of Jemg reef is the truncated sheer face of the southwest end of the reef adjacent to the island. This wall drops vertically from depths of 6m to 50m or more with some major overhangs (Figure A-8). The margin of the reef is nearly perfectly straight along a northwest to southeast axis (see Figs. 3 & 4) as if a giant knife had sliced away the reef mass further to the southwest. I can think of no explanation for this unusual geomorphic feature, except that a previously existing extension of this reef to the SE may have fractured and slumped down towards greater depths, leaving a vertical face on the remaining reef. Since none of the other table reefs in the Marshalls have been described in much detail, it is impossible to compare Jemg’s features to them. It is possible that Jemg’s table reef might be the peripheral remnant of a once larger atoll, the rest of which has been displaced by faulting below the depths required for active upward reef growth. Clearly Jemg’s reefs are unusual with features not having been previously reported in the scientific literature. Despite the abundance of hard reef surfaces and transparent well illuminated waters, coral abundance and diversity were low. Coverage on the pavement terraces was low and nearly absent from the walls of the sand channels. Highest coral development occurred in a NW facing reef indentation near the NE end of Jemg reef. This indentation apparently affords the reef slopes some protection from heavy wave action, allowing luxuriant, three-dimensional coral development. Typical vertical profiles of all reef sites sketched in the field are presented in Appendix A (Figure A-8 covers Jemo). The absence of a more extensive and shallow reef flat and the presence of the surge channels on the reef crest is curious. Perhaps the reef is too narrow or wave action too severe to facilitate shallower reef flat development. Wave action was observed to be approaching Jemg from all directions, although heaviest from the north during our visit. The bulbous NE terminus of the reef was the zone of maximum wave action. 25 Landing on Jemg Island is extremely hazardous as noted in Fosberg et al (1956), Fosberg (1988), and as experienced first-hand in 1988. The transition from deep to shallow water is abrupt off the SW end (the only safe "access" point to the island) and the spur and groove formations and unpredictable wave action renders boat navigation dangerous. During our landing, the shaft of the outboard motor struck the reef and broke off, causing our skiff to swamp. The lack of safe anchorage and access must have contributed historically to the lack of permanent habitation on Jemg. During heavy surf it would be impossible to land at the island from any direction. Rare Marine Species at Jemo Evidence of sea turtle nesting activity was high along Jemg Island’s sandy beaches and the level of evidence (53 pairs of turtle tracks) was second only to Pikaar’s. Signs of recent harvesting of green turtle was evident, and one nest with fresh eggs was discovered. Although uninhabited, J emg lies close to inhabited Wotje, Likiep, and Ailuk atolls. According to Fosberg (1988), Jemg was in pre-European times a turtle sanctuary, and only infrequent visits were permitted; turtles and eggs were taken in limited numbers under close supervision by priests (Jack Tobin, pers comm. to Ray Fosberg). Although not rare species, numerous sharks also inhabit the reef waters off Jemg. Jemg’s Corals Jemg’s coral fauna is noticeably depauperate with only 33 species and 16 genera reported after surveying 12 separate sites. These numbers compare to 74-93 species and 29-35 genera reported from atolls subject to the same sampling intensity (Table 2) during the 1988 survey. The most common Jemg species were ramose stony corals (Acropora (1) palifera, Pocillopora verrucosa, P. meandrina), robust firecorals (Millepora platyphylla), encrusting colonies (Montipora spp) massive brain corals (Platygyra daedalea, Favia spp, Hydnophora microconos) other reef corals (Symphyllia radians, Astreopora spp, Turbinaria sp, Porites spp, the soft coral Sinularia, and the blue coral (Heliopora coerulea). Maximum coral abundance and diversity occurred at depths between 7-10m and especially where reef indentations afforded corals some protection from heavy wave exposure. Free living forms such as the mushroom corals (Fungia and related genera) were not reported. Coral coverage appeared lower along the southeast facing slopes probably due to substrate instability of disturbance from moving sediment. ATOLL GEOMORPHOLOGY AND OCEANOGRAPHY Atolls are annular (perimeter) reefs enclosing lagoons which usually contain passes through the reefs, islets on the reef, and lagoon reef formations (MacNeil, 1969). The six remaining areas surveyed in 1988 are atolls with perimeter reefs affording protection and surrounding semi-enclosed lagoons. At least five or more small to moderate sized islets are situated on the shallow reef flats along the perimeter reefs of each atoll. All six atolls include at least one natural deep passage through the perimeter reefs, allowing sub-tidal exchange between ocean and lagoon waters. There are more islands and fewer passes on 26 the windward side than on the leeward side. The perimeter reefs of atolls in the Marshalls are usually between 1,000 and 2,000 feet wide, with windward reefs usually slightly higher in elevation. The six atolls can be divided into two major geomorphological groups with gradients between one another along a north-south axis: 1) Small Northern Atolls. These have shallow lagoons with maximum depths of 13mm or less, a very narrow single passage along the west side, and elevated (perched) water levels in the lagoon during low tide. This category includes Bok-ak and Pikaar. Toke Atoll is intermediate between the two groups. Wake (traditional Marshallese name Enen Kio) is located further to the north of the Republic of the Marshalls, is under U.S. military jurisdiction and is more closely allied to this group. The climate is dry and prevailing trade winds are heavy. In addition, the semi-enclosed nature of the lagoons of Ebon and Namorik (Namdik) Atolls and the two semi-enclosed sublagoons of Arno Atoll, all in the southern Marshalls, show some functional resemblances to the first group. 2) Central Atolls. These have deeper, more open lagoons with maximum water depths of 49m or more, larger, deeper or more numerous passages, and lagoon tidal fluctuations more closely corresponding to those of the adjacent ocean areas. Included in this category are Rondik, Wotto, and Adkup. These three atolls more closely resemble most of the other atolls in the Marshalls. Due to Wake’s proximity to Bok-ak and Pikaar, all three may have undergone a similar geological evolution. Wake has no natural passage through its reef, and the maximum depth of its lagoon is only 4m. Wake represents one extreme in the gradient between the two groups of atolls in terms of small pass development, shallowness of lagoon, and small lagoon area. At the other extreme would be the more typical atolls of the Marshalls charactered by larger and more numerous passages and lagoons, well developed perimeter reefs and islands, and passages generally concentrated along the south and west rims (Wiens, 1963). Toke Atoll, intermediate between the two extremes, most closely resembles nearby Utrok Atoll, with comparable angular shape, size, moderate depth of the lagoon, and small size and position of the single deep western passage. The Perched Lagoons of Bok-ak and Pikaar: The raised perimeter reefs and the single narrow western passage off both Bok-ak and Pikaar atolls restrict tidal exchange between the lagoon and ocean (Fosberg et al, 1956). More water is pumped into the lagoon by wave action along the eastern (windward) reefs than exits from the lagoon through the pass and over the reefs at low tide along the western (leeward) side. This factor causes average lagoon water levels to be higher compared to those on the ocean side. Since water levels in the lagoon never get as low as the low tide levels outside the lagoon, perimeter reefs may have continued to grow upward in response to the constant washing from higher lagoon water levels. 27 The raised nature of the peripheral reefs, especially along the leeward rim prevents tidal exchange over the reefs except at moderate to high tide, further increasing the accumulation of water pumped in the lagoon from the windward side wave action relative to that which exits the lagoon. Coupled with the limited drainage of lagoon waters out of the narrow pass of each atoll, the "low" tide water levels in the lagoon were observed to be two - three feet higher than corresponding levels on the ocean side. Lagoon tidal amplitude is very small and is nearly completely out of phase with ocean side tidal fluctuations. During low tides on the ocean side, lagoon waters were observed to stream out of the pass, dropping up to three feet and resembling white water "rapids" over a distance of about 100m. As a consequence, the leeward perimeter reefs of both Bok-ak and Pikaar atolls serve as natural dams and spillways, ponding lagoon waters and dampening outside tidal fluctuations. Only during the few hours of high tide do all perimeter and lagoon reef flats completely submerge, allowing free exchange of lagoon and ocean waters over the reef. At the time of highest tide (on the ocean tide) current flow out the channel reverses direction, running into the lagoon for an hour or so. Perhaps in response to less water level fluctuations in the lagoon, living corals and coralline algae grow to higher elevations, displaying prominent overhanging reef wall formations. In contrast, the central group of atolls (Adkup, Wotto, and Rofdik) display localized oceanographic conditions more typical of the rest of the Marshalls. Wave action along windward reefs pumps water over the perimeter reefs into the lagoon during virtually all stages of the tide. Water also enters the lagoon during flooding tides through all passes and over most shallow perimeter reefs. During ebb tide, water flow out of the passes, and ebb flow over leeward reefs is likely to be strong (but not specifically observed during the 1988 study). Lagoon and ocean tidal fluctuations appear more closely synchronized and show similar amplitudes. Wave action inside the lagoons is moderately high due to the more open configuration and larger size of the lagoons for the central atolls compared to those of the small northern atolls. BOK-AK ATOLL (Figures 5-9; A-1 through A-3) Aelon-in Bok-ak (also called Taongi, Bokaak, or Pokak Atoll) is the Republic’s most isolated atoll with the nearest reefs and islands located 150 NM to the southeast at Pikaar Atoll and 300 NM to the north at Wake Atoll (Enen kio). Also an unnamed bank at a depth of seven fathoms lies 100 NM south of Bok-ak (MacNeil, 1969). Bok-ak is unusual from several respects, not the least of which is the elevated configuration of its living lagoon reefs. The atoll is crescent shaped (Figure 5), curving to the west with reef horns extending off the northern and southern tips of the atoll reef, and is about 11 miles long from reef tip to reef tip. The 1988 team was able to spend three field days at Bok- ak allowing 20 marine sites, including five leeward ocean reef sites to be surveyed (Figure 5). 28 Fig 5. Bok-ak (Taongi) Atoll showing sites of marine surveys September 1988. Stars indicate reef sites of special interest. Scale: 1" = 1.970 mi. : : ne | Ane Jalto Bokon-Ak Bok-Dik 1R 7 Taongi Passag¢ 29 Fig 6. Southern tip of Bok-ak (Taongi) Atoll showing locations of some marine survey sites. Stars indicate reef sites of special interest. Scale: 1" = .6504 mi. A near continuous string of islands (Figures 6, 7) extends along the SE windward reefs. The NE windward reef and the entire western (leeward) reef lacks islands. Island land area totals 1.45 square miles (3.8 km’) for Bok-ak, the second most of any of the seven areas surveyed in 1988 after Wotto. There is no fresh water at Bok-ak and even wells dug in the center of large islands are quite salty (Fosberg et al, 1956). Bok-ak lies far enough north for tropical storms originating in the central Marshalls to have gained full typhoon intensity, and the atoll’s islands and reefs display extensive evidence of typhoon effects (MacNeil, 1969). Boulder ramparts, beaches, and concentrations of strewn boulders are thought to have been formed during intense storms and are most concentrated on the east and SE ocean facing sides of islands at Bok-ak Atoll (MacNeil, 1969; Fosberg, 1988). Many patch reefs throughout the lagoon are elongated into ribbon reefs with vertical or overhanging walls. The lagoon averages seven fathoms (13m) in depth with the greatest recorded depths being eight fathoms (15m), mostly in the western lagoon. The tops of many lagoon patch and ribbon reefs are awash at low tide with overhangs of profusely growing corals just below the surface. The tops of the shallowest reefs are smooth pavements of living crustose coralline algae. Over a full tidal cycle, lagoon water 30 levels were observed to fluctuate less than one foot. At high tide all reefs and corals are flooded to depths of a few inches or more. Water levels in the lagoon never dropped below mean tide level. Sand deposits covered the floor of the lagoon while most elevated surfaces were covered with live coral. Lagoon coral communities were very healthy with only a few dead corals observed. Giant clam populations in the lagoon were huge, including the species Tridacna maxima, T. squamosa, and Hippopus hippopus. Despite an intensive search neither live or dead remains of the largest species Tridacna gigas were reported. Neither were sea turtles observed at Bok-ak. Sharks were numerous, especially black tips inside the lagoon and grey and white tip reef sharks outside the lagoon. Fig 7. Bok-ak (Taongi) Atoll in the vicinity of the cluster of large éastern islands. Scale: 1"=.8953 mi. Kiden-Kan Ane Bokan Bokon-Ak Bok-Dik Eastern (windward) perimeter reefs at Bok-ak are different in shape compared to those of larger more open atolls to the south in the Marshalls. Observations at an elevation of about 8m, from the deck of a recently wrecked Japanese longliner fishing vessel on the windward reef (site 1-C), revealed that spur and groove formations are well developed and typical. However, the coralline algal ridge was a wide irregular feature, rather than a more typical elevated ridge measuring only a few meters in width (see also Fosberg, 1988). The reef crest was generally flat but elevated 2 or more feet above mean 31 low water. The back lagoon edge of the reef abruptly drops as a pronounced step 1-2m in depth. Elsewhere in the Marshalls back reef slopes towards the lagoon are generally more gentle. Lagoonward water flow over the reef was also not as swift as reported for many windward reef flats in the Marshalls. The higher observed lagoon water levels may prevent more rapid "downhill" movement of waters from the ocean side. Leeward perimeter reefs were unusually narrow, averaging less than 100m in width (Figures 8, 9). Except for a small coralline algal ridge-like feature at the crest of the Fig 8. Bok-ak (Taongi reef near its lagoon margin (see Atoll in the vicinity of Fosberg, 1956; et al, 1988), the upper the western passage. reef surface is smooth and covered Scale: 1"=.7075 mi.@ © with living crustose coralline algae 1R and slopes down two to three feet from the lagoon side to the ocean side. This tiny ridge, up to 10-15 cm in height, is also reported on windward facing edges of patch reefs in the lagoon and along the lagoon shores of some islets. The living reef flat serves as a coral-algal dam and spillway, holding back higher lagoon water levels except for excess water trickling downslope to the ocean margin. Even at low tide (outside), water was seen constantly spilling over the dam and down the spillway to the ocean, with flow presumably maintained by the constant wave action pumping water into the lagoon from the windward side. The ocean face of the leeward reefs resembled the steep slopes, reentrant/canyons, high coral cover, and diversity typically reported for such environments elsewhere in the Marshalls. Water currents were strong and turbulent off the leeward side of the southern reef extension or horn (site 1-O). Loose sediment and sand were absent from the shallow ocean reef slopes. Seven of the 20 sites surveyed at the atoll displayed exceptional or unique coral reef features (Figs. 5, 6, 9, A-1, A-2, and A-3). One unusual feature was an ocean patch reef (site 1-S) separated from the main ocean reef slope by a deep chasm. Other exceptional sites included a windward reef flat (site 1C), lagoon patch and ribbon reefs (sites 1K, IN, 11, 1G and 1H). Corals of Bok-ak Atoll Bok-ak was one of three atolls where the most species of corals were observed during the expedition. Ninety-three species belonging to 38 genera and subgenera were reported at Bok-ak based upon surveys at the 20 field sites, including five ocean sites. 32 In contrast, 93 species were reported at both Toke and Pikaar after surveys at only 13 sites which included no ocean sites at Toke and only two ocean sites at Pikaar. Thus, despite the high number of coral species reported at Bok-ak, Toke and Pikaar would appear to support high numbers of species, based upon equivalent sampling intensity. The condition of the coral communities of Bok-ak was healthy and flourishing at all observed lagoon and ocean reef sites. Thus a lower number of coral species does not appear to be related to environmental stress. Fig 9. Bok-ak (Taongi) Atoll in the vicinity of the north western reef tip. Marine survey sites shown with stars represent reef sites of special interest. Scale: 1" = .7149 mi. Because of the elevated nature of the lagoon coral communities, they may be more isolated from ocean reefs due to restricted tidal exchange. Furthermore, the remote position of Bok-ak from its nearest reef neighbors may reduce the number of coral species which can successfully migrate and establish at Bok-ak. Over prolonged periods this might be reflected in fewer total species of coral that are established at Bok-ak. Several reef genera which are common elsewhere in the Marshalls were absent from Bok-ak: Porites (Synaraea), Coscinaraea and Distichopora. Some genera were conspicuously more abundant at Bok-ak including Platygyra and to a lesser extent Anacropora. These observations lend further support to the hypothesis of geographic isolation of Bok-ak from nearby atolls. However, coral communities at Bok-ak achieve an unprecedented level of abundance and development. Lagoon habitats were complex three dimensional coral dominated environments, with many overhangs, mounds, walls and elevated ledges. The 33 protected shallow lagoon environment appears to promote optimal coral growth due to abundant light, transparent waters, lack of suspended sediment and only minor wave action. Along windward perimeter reefs, the stepped back reef margins included many abundant corals: Acropora (J) palifera, other Acropora spp, Porites lobata, Cyphastrea microphthalma, Goniastrea spp, Pavona minuta, Seriatopora aculeata, Heliopora coerulea, Stylophora pistillata, encrusting Montipora spp, Pocillopora spp, Favia spp, Leptastrea purpurea, Platygyra spp, Millepora spp, and Astreopora spp. On the slopes of lagoon pinnacles, the following corals were common: Porites cylindrica (fingercoral), Astreopora gracilis, Goniastrea pectinata, Favia pallida, Stylophora pistillata, Porites spp, Fungia fungites, Lobophyllia hemprichii, Montipora spp, Pocillopora spp, and Acropora spp, especially staghorn corals, and others. On leeward ocean reef slopes and margins, the following corals achieved prominence: Millepora spp, Acropora digitifera, A. palifera, Porites superfusa, Montipora tuberculosa, Stylophora pistillata, Ecinopora lamellosa, Goniastrea retiformis, Favia stelligera, Turbinaria stellulata, Symphyllia spp, Favia spp, other Acropora spp (tables), Porites spp, and Cyphastrea microphthalma. Many other species were common, and the leeward ocean reef slopes displayed the highest reef coral abundance and diversity observed of any habitat at Bok-ak. Rare Species at Bok-ak Atoll The smaller giant clam species which were abundant in Bok-ak lagoon are considered rare species. However, there was no evidence of the rarest and largest giant clam species nor of sea turtles or coconut crabs. Bok-ak, along with Pikaar and Jemg was regarded as a bird and turtle reserve by the Marshallese prior to the era of European influence (Jack Tobin pers. comm. to Ray Fosberg, in Fosberg 1988), and in the early 1960’s Bok-ak was designated as a reserve by the then District Administrator of the Marshalls. PIKAAR ATOLL (Figures 10-15, A-4 to A-5) Aelon-in Pikaar (also called Bikar Atoll) is the Republic’s second most isolated atoll with the nearest reefs and islands being Utrok and Toke Atolls some 80 NM (146 km) to the south and Bok-ak Atoll some 150 NM (247 km) to the north. An unnamed bank with a depth of seven fathoms lies about 50 NM (91 km) north of Pikaar Atoll. Pikaar most closely resembled Bok-ak in geomorphology but has much less land area. In fact, with only 0.19 square miles (0.49 km’) of land, Pikaar has the least amount of land of any atoll in the Marshall Islands, and only the table reef at Jemo has less island area. Storm generated boulder ramparts and concentrations of strewn boulders occur only along the northwest shoreline of Jeliklik Island and along the lagoon face of northwestern 34 Fig 10. Pikaar (Bikar) Atoll showing location of marine survey sites. Stars indicate reef sites of special interest. Scale: 1" = 1.042 mi. DS ae pe Almani ¢ Jeliklik OB a4 Jobwero Pikaar 35 perimeter reef flats (MacNeil 1969). Although no wells have been dug, the small size of the largest islands and dryer climate argue against potable groundwater at Pikaar Atoll (Figure 10). Bikar Passage lagoon Jobwero Fig 12. Pikaar (Bikar) Atoll in the vicinity of the southern tip and the island of Pikaar (Bikar). Stars indicate reef sites of special interest. Scale: 1" = 1 mi. Fig. 11. Detail of the single western passage at Pikaar (Bikar) Atoll. Scale: 1"=.25 mi The top of Pikaar’s reef formations are living and elevated some two to three feet above mean low water. Its one single pass (Fig. 11) is narrow and forked to the lagoon side along the western rim of the atoll. Small boat navigation through the pass at low tide is extremely hazardous, and many reef sharks patrol its waters. Pikaar Atoll’s largest island is Pikaar Island, resting on the widest section of _ the perimeter reef at the southern tip of the atoll (Figure 12). Remaining islands are very small or are only sand cays (Figures 13-14). Pikaar’s lagoon is deeper than Bok- ak’s, varying in depth 36 between five and 13 fathoms (9-24m). The lagoon near the southern tip and western pass region is shallower, averaging eight fathoms (15m), but most of the lagoon floor is situated at depths between 10-11 fathoms (18-20m). As with the reefs of Bok-ak, the tops of the shallowest reefs are smooth pavements of living crustose coralline algae. Ribbon reefs fill the lagoon with the walls dominated by live coral and with pronounced overhangs near the tops of the reefs. Sand deposits cover the floor of the lagoon. Fig.13. Pikaar (Bikar) Atoll at the eastern end showing locations of some of the marine survey sites. Scale: 1" = .8093 mi. Tidal characteristics at Pikaar appear to be very similar to those of Bok-ak although there was less time in the field (1-1/2 days) to observe them. At low tide, the tops of lagoon reefs are awash but rest some two to three feet higher than the margin along the ocean side of the reef at mean low water. Over a full tidal cycle lagoon water levels were observed to fluctuate less than one foot. At high tide all living reefs and corals are flooded to depths of one foot or more. Water levels in the lagoon never dropped below mean tide level. Huge populations of giant clams (especially Tridacna maxima, T. squamosa, and Hippopus hippopus) were found throughout the lagoon, resembling those of Bok-ak. Swimming green sea turtles were observed both inside and outside the lagoon, and evidence of recent sea turtle nesting activity was evident along the sand beaches of most islands, especially Pikaar. Si] Fig 14. Pikaar (Bikar) Atoll at the NW tip showing locations of some of the marine Survey sites. Scale: 1"= .5579 mi. Saye e eel e 2F numerous coral heads i os ee G lagoon ocean e 2 o 2M 7) Oo Se) 38 The windward, eastern facing perimeter reefs of Pikaar resembled those of Bok-ak in terms of form and coral species composition. The back lagoon edge of the reef flat drops down as a pronounced step one or more meters in depth. This feature shows up well in the color aerial photographs, and the stepped or double reef feature is also marked on available maps and charts. The fact that the stepped reefs were reported only from the reefs of Bok-ak and Pikaar suggests that higher average lagoon water levels may have something to do with the formation of the steps. Stepped reefs occur also along the southwest and northwest lagoon faces at perimeter reefs at Pikaar atoll. Compared to those of Bok-ak Atoll, leeward (western) perimeter reefs at Pikaar Atoll are generally wider. Only the NW section to the north of the pass region shows narrow perimeter reefs. The perched lagoon water levels are maintained by wave action pumping water into the lagoon at a rate faster than can drain out the deep western pass at low tide. The water circulation dynamics in Pikaar’s lagoon appear very similar to those of Bok-ak lagoon. Several coral and reef habitats (sites 2A, 2B, 2C) at Pikaar displayed exceptional or unique characteristics (Figs. 10, 11, A-4, and A-5). Site 2A is a deep reef flat moat to the west of the main island (Pikaar). Site 2B is a back reef environment on the windward side, and site 2L is an ocean reef slope along the leeward side of the atoll. Comparison between 1944-based maps/aerial photographs and 1978 color aerial photographs indicate that sandy beach habitat around Pikaar Island has increased substantially during the 34 year period (Figure 15). Hence, suitable nesting habitat for sea turtles may have increased during the interval. Fosberg (1988) reports that Pikaar showed signs of extensive change from a typhoon between 1945 and 1952 including possible loss of some of the small islets. During our visit in 1988 there was extensive damage to the Pisonia forest on Pikaar and near total destruction of it on Jobwero and Almani islands from high winds, possibly during a recent tropical storm or typhoon. Corals of Pikaar Ninety-three species of corals belonging to 35 genera and subgenera were reported from the surveys of Pikaar which encompassed 13 marine sites. Of interest was the presence of the purple fan coral Distichopora on ocean reef environments at Pikaar. This species was absent from Bok-ak although a related coral, Stylaster, was common. Stylaster was absent from Pikaar, suggesting that these two different species are filling the same niches in their respective atolls, occupying similar habitats. The presence of Pectinia in Pikaar’s lagoon is only the second record of this coral from the Marshall Islands. Other corals present at Pikaar but absent from nearby Bok-ak include Coscinaraea and Acanthastrea. Curiously the common corals Oulophyllia and Goniopora were absent at Pikaar although present at Bok-ak. Other "missing" corals from Pikaar 39 Ae eee PR RUC ye kas gates 10 fathom See contour a sandy ca =~ N mw reef line (MLLW) eal eect feogsart sandy shoreline (1978) —“—— island vegetation line \ (1978) 9000 ce00e island vegetation line (1944) ~_— FIGURE 15. Shoreline changes between 1944 and 1978 at selected locations in the Northern Marshall Islands. A. sand cay at northeast corner of Toke Atoll. B. Pikaar Island, southern Pikaar Atoll. Comparisons based upon U.S. Army 1944 black and white aerial photographs and E. G. & G. 1978 color aerial photographs. Scale: 1:30,000. cuprsinetoemt 40 which are normally common reef components include Psammocora, Porites (Synaraea), Cycloseris, and Halomitra. It seems plausible that some of these and other widespread genera would have been reported from surveys conducted at more sites and in deeper water, especially along ocean facing reef slopes. Common and abundant corals on reef flats and moat environments included the fire corals Millepora spp, the blue coral Heliopora coerulea, the corals Favia stelligera, Pocillopora spp, Pavona spp, Porites spp, Goniastrea retiformis, other Favia spp, Montipora spp, Cyphastrea spp, Astreopora spp, Leptastrea spp, Acropora spp, Seriatopora angulata, Stylophora pistillata, the soft corals Sinularia sp and Xenia sp, the free living corals Herpolitha limax and Fungia spp, the brain corals Montastrea curta and Platygyra sp, and the explanate corals Turbinaria spp and Echinopora sp. Common and abundant corals on pinnacles and ribbon reefs included: the fingercoral Porites cylindrica, the corals Stylophora pistillata, Favia stelligera, Cyphastrea microphthalma, Acropora spp, Herpolitha limax, Millepora platyphylla, Scapophyllia cylindrica, Favites halicora, other Favia spp, Fungia spp, and Seriatopora angulata. Common and abundant corals along ocean facing reef slopes along the leeward side of Pikaar include Pocillopora spp, Millepora spp, Acropora spp, Porites spp, Goniastrea spp, Favia spp, Montipora spp, Platygyra sp, Lobophyllia spp, Symphyllia spp, and Favites spp. Rare Species at Pikaar Pikaar is the most important sea turtle nesting area in the Marshall Islands. Over 264 sets of turtle nesting tracks were observed at the atoll around the perimeter of Pikaar (176), Jobwero (74), and Almani (14) islands. One set of fresh tracks was probably those of a hawksbill sea turtle while remaining tracks were of green sea turtles. One pair of green sea turtles were observed to be mating in waters offshore from Pikaar Atoll (see Thomas, 1989). Since pre-European times, the Marshallese have considered Pikaar to be a turtle and bird sanctuary (Fosberg 1988). TOKE ATOLL (Figures 16-21, A-6, and A-7) At its closest point Aelon-in Toke (also called Taka Atoll) lies only 7.3 km southwest of Utrdk (Utirik) Atoll. However, due to the position of Toke’s single deep pass along the western atoll rim, it takes about 46 km by boat to travel from the largest island of Utrdk Atoll to the largest island of Toke Atoll (Toke). Both atolls are roughly triangular in shape. Although Toke has a larger lagoon area than Utrok (94 km’ vs. 57 km’), the land area of Toke is comparatively quite limited (0.57 km’ vs 2.4 km’). In fact, Toke ranks only ahead of Pikaar with respect to land area for atolls in the Marshall Islands. Toke’s land area consists of five islands of which only Toke and Allook are large enough to support permanent vegetation. One centrally located well dug at Toke Island yielded non-potable groundwater (chlorides 440-840 ppm). Two other peripherally Watuwe-rok\ » Jejakiki-kan milook Fig 16. Toke (Taka) Atoll showing locations of marine Survey sites. Stars represent reef sites of special interest. Scale: 1" = 1.970 mi. 42 located wells yielded very salty non-potable water (Fosberg et al 1956). Although Toke’s pass is deep and narrow, boat passage is not hazardous. Toke’s windward reefs are also afforded some protection from the upwind position of Utrok Atoll. Storm generated boulder beaches are not well developed along Toke Atoll’s island shorelines. They are best developed along the eastern face of the southernmost island Watuwe-rok and the western (lagoon) faces of Toke and Lojiron islands along the eastern perimeter (MacNeil, 1969). Fosberg (1988) also describes the effects of a typhoon which passed over Toke inel9Ssile Unlike the lagoons of Bok-ak and Pikaar, Toke’s lagoon is deeper (maximum reported depth of 28 fathoms or 51m) and with many soundings between 18-22 fathoms or 33-40m (Figures 16-18). Jeig 17. Toke (Taka) Atoll, NW corner in vicinity of a recent shipwreck and one marine survey station. Dotted line represents , submerged reef. “. Scale: 1" = .8156 mi. Toke’s lagoon has fewer pinnacles and patch reefs (less than 50 total) all of circular shape and generally concentrated in the southern half of the lagoon and near the deep western passage (Figures 16-18). Despite Utrok’s smaller size, its lagoon contains over twice as many patch and pinnacle reefs. These factors help to explain why Toke is not permanently inhabited. Land, water, and lagoon reef resources are larger and more conveniently located on Utrok. Fishermen from inhabited Utrok Atoll occasionally visit Toke to fish and to harvest shellfish and sea turtles. The land owners and managers of Toke reside at Utrok, but the expedition was not able to visit Utrok because the limited field time was cut short due to an unscheduled stop at Aelon-in Kuwajleen (Kwajalein Atoll) for provisions. 43 Fig 18. NE tip of Toke (Taka) Atoll i in the vicinity of Several September #3E 1988 marine survery stations. : Dotted line represents submerged reef. Stars represent reef e 3D sites of special interest. Scale: 1" = 1 mi lagoon The lagoon margins of Toke’s perimeter reefs are not elevated above mean low water as is the case at Pikaar and Bok-ak atolls. Furthermore, lagoon tidal fluctuations are more closely in phase and amplitude with those on the ocean side. At the most, water levels in the lagoon at low tide were only a few inches higher than outside low tide levels. Thus, Toke’s lagoon does not have the perched or elevated lagoon reefs and water levels characterizing the other two atolls, and exchange of waters between the lagoon and ocean is more pronounced. Besides the deep passage near Toke island with a depth of 12 fathoms (22m) and a width of over 100m (Figures 19-20), Toke Atoll also has several other shallower passages through the western reef, three of which were visited during the 1988 surveys (sites 31, 3J, and 3L). Four of the 13 marine sites surveyed at Toke displayed unique or exceptional reef characteristics. All were situated along the lagoon margins of perimeter reefs, two near the passes (sites 3K and 3L) and two in the northeast corner of the lagoon (sites 3G and 3E) where reef pinnacle and patch reef formations are slowly being buried under accumulating sand deposits washing over the reef flats from windward directions (N & E). The team was unable to visit any ocean reef slope sites at Toke due to safety and time limitations. Observations from the ship indicate that live corals dominate the slopes of ocean facing reefs and that sharks were numerous. One of the largest sand cays visited during the expedition occurred at the northeast corner (Figure 18) of Toke Atoll (site 3F). Comparison between Army maps based upon 1944 aerial photographs and 1978 color aerial photographs reveal the shape and position of the sand cay has changed drastically. During the 34-year interval, the deposit elongated and shifted 200m to the north. Our sea level observations in 1988 could not determine whether additional changes had occurred since 1978. The instability of the deposit and low elevation may explain the lack of vegetation on the sand cay. 44 Corals of Toke Atoll Ninety-three species of corals belonging to 35 genera Fig. 19. SW side of Toke (Taka) and subgenera were reported Atoll in the vicinity of 2 marine survey sites and deep passage. collectively from the 13 Dotted line represents submerged marine sites at Toke Atoll. reef. Stars represent reef sites of Of interest was the presence special interest. . Scale! Am 8456 mi: of Porites (Synaraea) and Sandalolitha which were absent from both Pikaar and Bok-ak. The reported lagoon absence of several common ou reef genera from Toke may Taka at be attributed to the lack of Passage v observations along ocean facing reef slopes where different and more diverse coral assemblages are expected. As a result, the species diversity of corals at Toke might be higher than observed at Pikaar and Bok- ak. The number of species reported per site was also relatively high at Toke, and coral communities were well developed and diverse. In back reef environments along windward (eastern) reefs, the following corals were oy lagoon abundant or common: Acropora Bi Fig 20. Passage off spp, Cyphastrea microphthalma, western Toke (Taka) Pavona varians, Montastrea Atoll. Dotted lines curta, Tubipora musica, Favia Li submerged “ae ieee reef. spp, Platygyra daedalea, Fungia Seale: 1seadsornik spp, Astreopora spp, Montipora spp, Favites abdita, Pocillopora verrucosa, Sinularia sp, Porites : Taka spp, Millepora spp, Echinopora Passage lamellosa, Turbinaria stellulata, Goniastrea spp, Heliopora coerulea, Stylophora pistillata, Psammocora profundacella, and Leptastrea purpurea. 45 Further offshore on shallow pinnacles and on the slopes of larger patch reefs the following corals were common or abundant: Porites spp (both fingercoral and massive forms), Pocillopora spp, Acropora spp, Montipora spp, Fungia spp, Goniopora lobata, Astreopora spp, Pavona spp, Tubipora musica, Coscinaraea columna, Oulophyllia crispa, Leptastrea purpurea, Seriatopora spp, Lobophyllia hemprichii, Millepora platyphylla, Stylophora pistillata, Stylocoeniella armata, Goniastrea spp, Cyphastrea microphthalma, Favia spp, Porites (Synaraea) rus, Montastrea curta, Psammocora profundacella, Echinopora lamellosa, Herpolitha limax, Leptoseris mycetoseroides, Heliopora coerulea, Scapophyllia cylindrica, Turbinaria stellulata and Favites russelli. Rare Marine Species Evidence of green sea turtle nesting activity was observed along the shorelines of Toke Island (16 sets of tracks), Lojiron Island (4 sets of tracks), and Allook Island (4 sets of tracks) (Figure 21). Of the seven areas visited Toke Atoll ranks fourth behind Pikaar, Jemo, and Adkup with respect to the level of sea turtle nesting evidence. The only sighting of a Hawksbill sea turtle during the expedition was off the NE sand cay at Toke (site 3F). It was observed to be feeding and swimming at a depth of 2-3m off the bottom. lagoon / Toke Atoll was the first in which live specimens of the rare giant clam, Lojiron Tridacna gigas were observed, primarily in Toke shallow lagoon environments near islands or back reefs. However, there were many more dead shells of the species observed on the reefs (only 5 of 24 were alive). Live individuals of the smaller species were present Fig 21. SE corner of Toke (Taka) Atoll in the vicinity of the largest islands. Dotted but in smaller numbers than line represents submerged reported for Pikaar and Bok- reef. ak. It was reported by Allook Scale: 1" = 1.348 mi. islanders from Majro (Majuro) and WoOtto that overseas fishermen illegally poach live individuals of T. gigas to obtain the abductor muscles which fetch high prices in Asian markets. Evidence obtained during our surveys suggest that uninhabited Toke Atoll may be an inviting target for illegal poaching of the rare giant clam Tridacna gigas. Interviews with the residents of Utrok might shed additional light on the extent of traditional harvesting and illegal poaching of giant clams at Toke Atoll. 46 WOTTO ATOLL (Figures 22-25, A-9, and A-10) Originally, uninhabited Ailinginae Atoll was to be visited during the expedition. However, the failure to obtain approval to visit the atoll, and a subsequent invitation extended by the leaders of Wotto Atoll, led the expedition to visit WOtto instead of Ailinginae on 18-19 September 1988. Aelon-in WOtto (also referred to as Wotho Atoll) is the only inhabited atoll surveyed during the 1988 expedition, and is only one of two atolls visited within the Ralik (western or "sunset") chain of the Marshalls. W6tto is relatively small in terms of lagoon area and is located within the dryer belt of the RMI. However, land area is greater (4.2 km’), the most of any of the seven areas visited during the 1988 expedition (Figures 22-23). Of the 22 inhabited atolls in the Marshalls though, Wotto ranks only ahead of Utrok (Utirik) and Namdik (Namorik) in terms of land area. Its population is about 100, the smallest of any inhabited atoll, and ranks ahead of only Jabat and Lib which are inhabited table reefs. Of the 18 islands at Wotto, only the largest (WOtto) is occupied (Figure 24). The atoll, including its islands, reefs, and village setting, is very scenic, relatively undisturbed, and harbors considerable natural and cultural diversity. Apparently, large land areas were never cleared and planted to coconuts (Fosberg et al, 1956). The nearest atolls to WOtto are Kuwajleen (Kwajalein) to the southeast; Ujae to the south; Bikini to the north; and Ailinginae, Ronlap (Rongelap), and Rofdik (Rongerik) to the northeast. Due to the clustering of several large atolls in the vicinity of WOtto, only the atoll’s southwest quadrant is considered vulnerable or exposed to heavy open ocean seas. Wotto Atoll is roughly triangular in shape with the widest reefs and the largest islands situated at the apexes. WOtto Island, the atoll’s largest, is at the NE tip of the atoll. The longest axis is the southwest facing side of the atoll, some 20 km between Majur-wor Island at the NW tip and Kapen Island (Figure 25) at the south. Most of the SW facing axis consists of several wide but shallow passages between smaller clusters of shallow reefs. The main navigational passage occurs just north of Pik-en Island near the northwest end of the axis. The three islets along the SW axis, Pik-en, Anbwil-en, and Ane-aidik are very small. The rest of Wotto Atolls perimeter reefs along the N and E facing axes are shallow and contain numerous islands and cays. The largest sand dune reported by MacNeil (1969) in the northern Marshalls occurred along the lagoon side of Ane-aidik Island, Wotto Atoll. The three largest islands also display large boulder beaches and concentrations of strewn boulders most likely tossed up on the reefs during storms. The boulder beaches face seaward to the north on the two large northern islands (Majur-wor and Wotto) and face seaward to the south off the southern islands of Kapen and Ane-jaito. A small boulder beach also occurs on the north side of Kapen Island. Ane-barbar Ane-obnak Majur-won Anbwil-en Ane-aidik Fig 22. Wotto (Wotho) Atoll. Numbers indicate marine survey sites. Stars show reef sites of special interest. Scale: 1" = 2.058 mi. 48 Figure 23 . Defense Mapping Agency Chart of Wotto (Wotho) Atoll. Scale 1:316,120. Soundings in fathoms. WOTHO ATOLL Medyeron I. @ , Lat. 10°10'36"N.—Long. 165°55/147E SCALE 1:316,120 VAR 8°27°E (1980) ANNUAL CHANGE 4°E Lif Anchorese : BeginI\ Aree & Begin Chan **, Nautical Miles 5 Meters 49 ~ ee _-— O)5c BH PON o> ” 5 ie Figure 24. Wap of northeastern ie WOtto (Wotho) Atoll sketched ne from 1978 color aerial { photographs. Scale 1:30,000. \ Numbers refer to Septembér 1988 i marine survey sites. Underlined ( numbers represent reef sites of special interest. c y; Island vegetation am = line NT OG uv y oy A Sand beach shore- 4 es } line Reef line (MLW) = arm 10 fathom contour — — — —— 50 Figure 2:5, Map of Southern Wotto (Wotho) Atoll sketched from 1978 color aerial photographs. Scale 1:30,000. Numbers refer to September 1988 marine survey sites. Underlined numbers represent reef sites of special interest. XN ~—_ Island vegetations line > “se Cand beach shoreline aT VTi, Reef line (MLW) 10 fathom contour 51 Patch reefs are not numerous in WoOtto’s lagoon and consist of three small clusters. One cluster off Kapen island in the south lagoon includes five reefs. A second cluster of seven reefs occurs in the south central lagoon east of Anbwil-en Island. The third and largest cluster of 10 reefs in the north central lagoon occurs east of Pik-en Island. Our limited time at WOtto did not permit field visits to patch reefs in the deeper lagoon. There is very little easily available information about the depths and bathymetry of Wotto’s lagoon. The only published soundings (Figure 22-23) show depths of at least 20 fathoms (37m) in the lagoon anchorage east of the main navigation channel which shows a depth of 6 to 6-1/2 fathoms (11-12m). The channel south of Anbwil-en Island shows a depth of 3-1/2 fathoms (6.4m). Some lagoon areas closest to Wotto Island may be shallow, with published depths of 1-1/2 to 2-1/2 fathoms(3-4.6m). Inspection of 1978 color aerial photographs of WoOtto indicate that most of the lagoon appears deep and in excess of 10 fathoms (18m) and probably close to 20 fathoms (37m) or more. Thirteen field sites were visited at WOtto, all along perimeter reefs (Figure 22). Only site 5L was on the ocean side of the reef. Limited time did not permit more than cursory observations of water current and circulation patterns at Wotto Atoll. During late afternoon fieldwork on 19 September 1988 on the shallow pass area between Ane-aidik and Anbwil-en islands (site 5L), a strong 2 kt current coinciding with rising tide was flowing north through the channel and into the lagoon. The depth of the top of reef varied from 2-3m. WoOtto’s lagoon appears to be well flushed, based upon an analysis of the configuration of Wotto’s reefs and the presence of wave action along the ocean sides of the north and eastern reefs. The wave action continually pumps fresh ocean waters into the lagoon from the windward (NE) sides during essentially all stages of the tide. This cooler water probably sinks to the bottom of the lagoon displacing less dense water which exits the lagoon over the western reef during ebbing tides. As noted during our limited field observations flood tide currents are strong and ebb tide currents are expected to be strong if not stronger. Although the western passes are shallow, their great width enhances the exchange of lagoon and ocean waters during tidal fluctuations. Corals of Wotto A total of 88 species belonging to 36 genera and subgenera were reported from WoOtto based upon surveys at 13 sites. However, time was very limited at several stations where observations were hampered by low light conditions during a cloudy late afternoon. In general, coral communities were diverse and in good health. Genera reported at Wotto which were not observed elsewhere during the expedition were Cycloseris, Halomitra, and Polyphyllia, all being free living mushroom corals. The last (Polyphyllia) is a new generic record from the Marshall Islands. A few common coral genera should have been reported but were not and include: Psammocora, Lobophyllia, Hydnophora, and Echinopora. More extensive surveys on ocean facing reefs might have yielded some of these genera as well as others. a2 Four of the 13 stations surveyed displayed unique or exceptional characteristics worthy of some mention. Three of these were lagoon reef slope habitats (sites 5A, 5G, and 5K) and all supported live individuals of the rare giant clam Tridacna gigas. Three of the sites (except 5G) had high live coral coverage of 60% or more. The single ocean facing reef slope site surveyed showed spectacular relief and well developed coral communities (site 5C), but numerous sharks and strong currents hampered the collection of additional information. The slope was characterized by a series of large coral canyons with flat scoured floors. Abundant and common corals on the ocean reef slope site were Millepora platyphylla, corymbose and table coral species of Acropora, Pocillopora spp, Turbinaria stellulata, Porites lobata, Montipora spp, Stylohora pistillata, Platygyra pini, Favia spp, Favites spp, Pavona minuta, Acropora palifera and Pavona spp. Corals on a deep pinnacle along the western perimeter reef (site 5M) included Acropora spp, Millepora spp, Stylophora pistillata, Pavona minuta, Pocillopora meandrina and Montipora aequituberculata. Abundant and common corals along the sheltered lagoon slopes (sites SA-5K) of perimeter reefs included: Seriatopora hystrix, Acropora (many species), Astreopora spp, Stylophora pistillata, Miullepora exaesa, Fungia spp, Porites spp, Montipora spp, Oulophyllia crispa, Goniastrea retiformis, Pavona spp, Pocillopora spp, Scapophyllia cylindrica, Acropora palifera, the blue coral Heliopora coerulea, Favia spp, Platygyra spp, Porites (Synaraea) rus, Coscinaraea columna, Herpolitha limax, and Cyphastrea sp. Rare Marine Species Giant clams of several species were reported from WoOtto including the most living specimens (15) of the largest and rarest species, Tridacna gigas. Some individuals were very large, and WOtto was the only area where living specimens approached the numbers of dead shells of the largest species (15 vs. 16). Interviews with the islanders revealed the Taiwanese fishermen visited WoOtto "about six or eight years ago" to seek permission to harvest T. gigas. After it was granted, the fishermen proceeded to harvest many clams but taking only the abductor muscle and leaving the dead shells and remaining tissue "to rot in the sun." This experience seemed to have shocked the islanders and made them more conscious of the need to protect remaining giant clams. No doubt the presence of the islanders discourages further harvesting. Despite recording the highest number of living Tridacna gigas at Wotto Atoll, it is important to note that the ratio of dead to live clams is still relatively high despite a hiatus on harvesting the species over a period of six to eight years. This fact points to the vulnerability of such a population to overexploitation, even from occasional or one-time harvests (Thomas 1989). The smaller giant clam species are preferred by the islanders for consumption; for one, they are easier to collect and shuck. Although all common species of giant clams 5/9 were observed on WoOtto’s reefs, the smaller species seemed less numerous than reported at Pikaar, Bok-ak, and perhaps Toke. Green turtles nest at WOtto but only in low numbers. The most pairs of tracks (4) were spotted on the beaches of Pik-en, (Figure 22), with two pairs of tracks observed at Long Island (Bokon-aetok) and two pairs at Kapen Island (Figure 25). During the night of 18 September 1988, the ship’s crew captured a female Green sea turtle after it had laid its eggs on Long Island, and gave it to the villagers. The WoOtto islanders harvest the turtles only infrequently for special or ceremonial occasions, usually during the summer months off the beaches of uninhabited islands. The villagers seem very conscious of the vulnerability of the nesting turtle population and limit their harvesting practices accordingly (Thomas, 1989). Coconut crabs are heavily harvested by the islanders, primarily at Kapen Island, and approximately 500 crabs per year are captured and consumed locally. The average size of the crabs during our visit was about 0.5 kg, somewhat smaller than those observed at Rondik. Harvesting pressure on coconut crabs at Wotto could be higher except that a fuel shortage during most of 1988 prevented small boat travel and access to outer islands including Kapen. RONDIK ATOLL (Figures 26-30, A-11 and A-12) Aelon-in Rondik (previously referred to as Rongerik Atoll) is moderately sized and is located just east of Ronlap (Rongelap) and Ailinginae, northeast of Wotto, north of Kuwajleen (Kwajalein), west of Toke and Utrok, and northwest of Jemg and Likiep. It is generally afforded some protection from storms and large waves by the positions of these atolls. However, Rofidik is exposed to open sea conditions from the north. Rofdik is roughly circular in outline, has the second largest lagoon area (145 km’) and the third largest land area (0.81 sq. mil) of the six atolls and one table reef visited in September 1988. All but one (Bok) of its 17 islands are located along the eastern perimeter of the atoll (Figure 26). Rondik’s lagoon is very open with major gaps in the perimeter reefs along the northwest and west sectors. The navigation chart of Rofdik shows several major navigable passes: Jeteptep and an unnamed passage to the north, Bok Passage to the west, and Enewetak Pass (Ane-wetak) to the south (Figures 27-28). Bok Passage is over three miles wide, and together with the lack of islands and shallow reefs along the west rim, renders the lagoon and lagoon shorelines of islands facing to the southwest exposed to heavy wave action. 54 6B 6C a pa oe Fig 26. Rondik (Rongerik Atoll showing locations of marine Survey sites. Stars represent reef sites of special interest. Scale: 1" = 2.335 mi. Fig 27. Southwest corner of Rondik (Rongerik) Atoll showing the main south passage and the location of 2 marine survey sites. Scale: 1" = .3071 mi. 55 It is interesting to note that the shorelines of several islands and sand cays along the northeast rim of the atoll have undergone considerable change between 1944 and 1978 comparing earlier and later sets of aerial photographs. The island of lagoon Rondik shows minor accretion of the beach and vegetation along the SW facing lagoon Fig 28. Western end shoreline. However, the W tip of Rondik, of Rondik (Rongerik) the lagoon shorelines of two large islands Atoll showing Bok Island.| (Patpat and Pikonaden), and several sand mine Henn vegion\was cays between the islands have undergone not visited during the ; : 1988 survey, Bok is a beach erosion and some receding of the suspected turtle vegetation line, especially along W and SW nesting site. facing shorelines. Perhaps large waves or Seale: 1" = 9625 mi. storms traversing the lagoon from the exposed W and SW directions were responsible for these shoreline modifications. Elsewhere on the atoll, beach and vegetation lines along the shorelines of islands have remained unchanged. The lagoon side of Ane-wetak Island has an unusually high sand dune (Fosberg, et al, 1956). According to Fosberg (1988) the island was also much disturbed by construction of a radio station involving bulldozing a strip across the center and a road along the length of the seaward coast prior to 1956 (Figure 28). Rondik’s lagoon is fairly deep (maximum reported sounding of 28 fathoms or 51m) with many pinnacles and patch reefs which breach the sea surface. Most shallow lagoon reefs are patch reefs, circular or eliptical in shape and located in the eastern and central lagoon. Primarily deeper lagoon pinnacles are found in the western lagoon and are not as numerous or spatially dense. No soundings or bathymetric data are presented for the north and southwest extremities of the lagoon. The survey of Rofdik was cut short due to a sudden medical emergency and the need to transport a sick seaman to the nearest hospital on Epja (Ebeye) island at Kuwajleen (Kwajalein) Atoll. As a consequence, the western half of the atoll was not surveyed including Bok (Bock) island and the extensive reef and lagoon areas to the NW and SW of it. Small boat travel in the open lagoon was more turbulent the further away from the upwind (NE) reefs, and most of the 12 marine survey sites were within the lagoon shelter of the NE perimeter reef between Ane-wetak (Eniwetak) and Jeteptep islands. Six of the 12 marine survey sites demonstrated unique or exceptional natural characteristics worthy of mention (sites 6F, 6G, 61, 6J, 6K, and 6A) and of possible conservation importance (Figure 30). Site 6A was a shallow downwind lagoon reef 56 (» & Mot-lap Fig 29. Main island cluster of NE Rondik (Rongerik) Atoll. Numbers represent marine survey sites, stars being reef sites of special interest. Scale: 1" = .68 mi. complex with table corals and gigantic spectacular colonies of the yellow foliaceous coral Turbinaria. Site 6F was a shallow lagoon reef adjacent to Rondik island with exceptional coral diversity. Also one live but six dead Tridacna gigas giant clams were reported there. Site 6G was a spectacular deep reef flat moat environment on the windward side of Rondik dominated by extensive platforms and microatolls of the blue coral Heliopora coerulea. The high development of blue coral was unique and is probably maintained by wave-generated water currents which constantly flush clean ocean water through the moat system. Site 61 was the only deeper lagoon pinnacle reef surveyed and contained very high coral abundance (over 90% live coral coverage) and diversity of reef fishes. Sharks, however, were numerous and aggressive, preventing more detailed listing of corals. Sites 6J and 6K were shallow lagoon pinnacle habitats near the NE perimeter reef with fairly high coral coverage and very high coral diversity. The blue coral back reef zone of site 6K was exceptional, and underwater visibility and relief were also excellent. The expansive pink sand beaches of the NE islands of Rondik atoll are also worthy of mention. The beaches are formed primarily of the remains of pink foraminiferal tests. Although foraminiferal sand beaches are commonly observed in the Marshalls, their extensive development and coloration at Rondik added substantially to the natural beauty and aesthetics of the atoll’s island ecosystems. "000°0€:1 eTeos ‘sydei30j0yd TeTaee azojTos *) 9°95 *G pue syderaZ0joyud Tefiee ajtTym pue yIeTq HwHE, Awiy °S°p uodn peseq suosfieduoj <‘T[[OIV (AFiesuoy) YF_puoy jo wea Yseoayjiou BuoTte JaISN[O pue[S} upeW sSpueTs] TTeusieW ul IYyYIION 9y UF SUOTIeDOT Pe eTOS 3e 8/6 puke HHEI Usemjeq SaeBueYD sUF[eACUS “QE eANsTY 8L61T JO FsSoYyA wor; 12JJEP YOFYM SQueuZes (761) FUFTe1OYsS Apues xew KM 8261 JO eSoy Wo1Z 19zZTP YOFYM squseUses (HHE[) SUT[ UOTIEISBZOA PURTST ececcrcccce (8461) 2UFT uoz3zeqeden pueTsST —~~__L SUOF IIIS (8261) ouFfT9a0ys Apues ste eee c eens ees e., (MTT) auf, joo1r VAAL AA “SI MIGNOd Seas Anojuod woyyejJ OT —-— —— 58 Rare Marine Species Evidence of substantial recent sea turtle nesting was observed along the lagoon beaches of Ane-wetak (Eniwetak) island. A total of 33 pairs of tracks were observed along Ane-wetak and one additional pair was observed on Kaarooka island. A school of large bottlenose dolphins was seen swimming outside the reef on the ocean side. Although bottlenose dolphins are common throughout the tropical Pacific, marine mammals were observed only on this occasion during the 1988 expedition. There is no obvious explanation for this curious lack of sightings of marine mammals elsewhere during the northern Marshalls expedition. The several smaller species and the rare larger species of giant clams were present at Rondik but not common. Only four live individuals compared to 16 dead shells of Tridacna gigas were observed on the reefs, suggesting heavy collection during the recent past. Except during a brief period in 1946-1948, when Rondik was inhabited by the displaced Bikinians, there are no other known periods of occupation of Rofidik during the recent historical past. Hence the high mortality of giant clams may be best explained as the consequence of unauthorized (and unobserved) poaching. The largest concentration of coconut crabs observed during the expedition was reported from Rondik islet. Other large islands at Rondik Atoll may also harbor large coconut crab populations, but only Rofdik and Ane-wetak islands are said to be planted in coconuts (Fosberg et al, 1956). Not only were the crabs very numerous, but many individuals were large, exceeding two to three kg in weight. In a few hours time, several of the ships crew were able to collect over 100 crabs without much effort, all of which were 0.5 kg or more in weight. The large population of coconut crabs at Rondik is best explained by infrequent harvesting pressure by islanders (since Rondik is uninhabited), and the abundance of coconut trees. Coconuts are the preferred food of the crabs. Although coconut crab is a favorite delicacy of the Marshallese and other Pacific islanders, crab populations at Rondik atoll may be contaminated with radionuclides. In 1954, the U.S. accomplished BRAVO, the atmospheric testing of a large thermonuclear device (H-Bomb) at Bikini Atoll, some 230 km west of Rondik. Due to unanticipated adverse (westerly) wind conditions and higher than expected energy yields from the detonation, radioactive fallout from the BRAVO blast penetrated the upper atmosphere and drifted east. Fallout from BRAVO was observed to contaminate Bikini, Ronlap (Rongelap) and Utrok (Utirik) Atolls and may have contaminated other nearby atolls, some of which are uninhabited. Rofdik lies on a direct line between the contaminated atolls of Utrdk and Rojlap, and thus it is likely to have been contaminated by fallout from BRAVO. Radiological studies at Ane-wetak (Eniwetak) and Bikini Atolls reveal that the radionuclides cesium-137 and strontium-90 are taken up and concentrated in the tissues of coconut trees and nuts. The consumption of contaminated coconuts was 59 probably the most likely pathway explaining how returning Bikinians received excessive dosage of these radionuclides during their aborted resettlement during the 1970’s (see BARC, 1985; 1986; USACE, 1986). Coconut crabs can also become contaminated by eating contaminated nuts, and in turn humans can become exposed to the radionuclides by consuming effected coconut crabs or nuts. Although there has not been radiological surveys of the trees and crabs of Rondik Atoll, the consumption of these foods poses a potentially serious health risk, especially since the coconut crab population is very large and since the crabs are a favored delicacy, in great demand. Comparison of 1944 and 1978 shoreline configurations for NE Rondik Atoll did not reveal significant changes (Figure 30). Corals of Rondik Only 74 species belonging to 29 genera and subgenera of corals were reported from Rondik Atoll based upon the results of visits to 12 marine survey sites. These numbers reflect lower levels and diversity of sampling at Rondik compared to the other atolls. Only one lagoon pinnacle and no ocean reef sites were surveyed, and survey time at the most northerly sites were shortened due to the need to leave Rondik earlier than planned due to a medical emergency. The entire western lagoon of the atoll was also left unsurveyed. Many common coral species were not observed at Rondik, including Stylocoeniella, Porites (Synaraea), Psammocora, Hydnophora and Leptoria. Most of these would be expected on ocean reef slopes if they could have been examined. Underrepresented during the survey were several genera and species of common free living mushroom corals. Of interest was the recording of Plesiastrea versipora at Rondik, the only atoll where this coral was reported during the 1988 study. Abundant and common corals on lagoon pinnacle reef environments included: Acropora spp (table and staghorn coral), Pavona spp, Stylophora pistillata, Montastrea curta, Pocillopora spp, Cyphastrea spp, Astreopora spp, Montipora spp, the blue coral Heliopora coerulea, Porites spp, the leafy yellow coral Turbinaria, the fire corals Millepora spp, Fungia spp, and Seriatopora hystrix. Abundant and common corals along the windward (NE) lagoon perimeter reef slopes included: Heliopora coerulea, Acropora spp, Goniastrea retiformis, Fungia scutaria, Astreopora spp, Pavona minuta, Stylophora pistillata, the conspicuous table coral Acropora cytherea, Porites spp, Montipora spp, Leptastrea spp, Millepora spp, Montastrea curta, Favia spp, Pavona minuta, Turbinaria stellulata, Platygyra daedalea, Pocillopora spp, Goniastrea retiformis, Favites spp, the fingercoral Porites cylindrica, Seriatopora spp, the soft coral Sinularia sp, Cyphastrea spp, and the organ pipe coral Tubipora musica. Abundant and common corals within blue coral dominated reef moat and back reef flat environments included: Heliopora coerulea, Stylophora pistillata, Seriatopora hystrix, 60 Jabonwod Areto-jaion ° Aseto-jairok 7| Fig 31. Adkup (Erikub) Atoll. Bokan-aik “SS Numbers indicate marine survey Boke-lomjan ; Du Bokan-kowak sites with stars showing reef sites of special interest. Map created by CORIAL using MARIS. Scale: 1"= 3.146 mi. 61 Pocillopora damicornis, Acropora palifera, Porites spp, and Leptastrea purpurea. Giant clams were abundant in these environments and along the slopes of perimeter reefs, but most were dead and some stacked in piles - clear evidence of unauthorized poaching. Nurse sharks (a harmless species) were also numerous. ADKUP ATOLL (Figures 31-37; A-13, and A-14) Aelon-in Adkup (also referred to as Erikub Atoll) has the form of an ellipse with its long axis (about 27 km long) facing NE and SW (Figure 31). It is the largest and most southerly of the atolls visited during the 1988 expedition. In comparison to the rest of the Marshall Islands, Adkup is centrally located and of intermediate size in terms of lagoon surface area (232 km’), ranking 15th of 28 atolls. In terms of land area (1.53 km’) however, Adkup ranks much lower, 25th, and even the table reef of Mejit has more island area. Due to its lower latitude, Adkup probably experiences a greater average annual rainfall rate than the other six areas visited in 1988, and its islets are densely forested. Adkup is clustered among several other atolls including Wotje just five NM to the north, Maloelap and Aur to the southeast, and Likiep to the northwest. Perhaps due to limited land and water area, Adkup is not permanently inhabited. However, residents of nearly Wotje regularly visit Adkup to gather copra, fish, and other food. At the time of our visit on 22-23 September 1988, there was evidence of a very recent visit to the main island (Adkup) of the atoll probably to harvest sea turtles, fish, and crabs. Due to its central location, Adkup is sheltered by other nearby atolls from heavy exposure to storms, surges, and waves, except those approaching from the south. There are no _ recent available navigation charts of the atoll, and the 1978 aerial photography of the northern Marshalls did not include Adkup. Eventually, after considerable searching, a complete set of the U.S. Army Map Service topographic series maps of Adkup was obtained and analyzed (Figures 32-37). Fig 32. NE peripheral Adkup’s lagoon and reefs show some reef of Adkup (Erikub) unusual features. Lagoon patch reefs and Atoll showing marine pinnacles are rare, given the large size of the SUNG SIO CVE SeE lagoon. Three clusters of patch reefs, each interest. Pisce than DOMEEE a a atealhe Scale: 1" = 8611 mi. of less than reeIs occur opposite three o the atoll’s six passes (for example Figures 31, 36). Elsewhere in the lagoon there are only a few isolated patch reefs (only seven could be counted from the AMS maps) and reef pinnacles were only slightly more abundant (for example see Figures 31-34, 36). Unfortunately, time did not permit the team to visit patch or pinnacle reefs in the deeper lagoon. A limited number of soundings have been taken in Adkup’s lagoon, especially 62 near the passes, and reveal that the lagoon is deep. At least one lagoon area in the vicinity of the west central lagoon shows depths of 30 fathoms (55m) or more. ‘ Six deep passes cut through Adkup’s perimeter reefs (see Figures Fig 33. Northeast rim of 31, 33, 34, and 36) and all are Adkup (Erikub) Atoll near considered navigable. One pass to pass and in vicinity of the north of Areto-jairok Island is on several September 1988 the windward (eastern) side of the marine survey sites, two atoll (site 7J; Figure 33) and is the of which (stars) were of : special interest only windward pass present at any of Scale: 1"= 1.252 mi. the six atolls surveyed during the 1988 expedition. The AMS map lists the depth of this pass at 3-3/4 fathoms (7m) and a width of about 60 m, but based upon my snorkeling Areto-jairok )2 observations, the minimum depth of this pass may be less (about 4m). The other passes are deeper. For example, the northwestern pass near Looj Island is 12 to 23 fathoms (22- 42m) deep, and the southwest pass closest to Adkup Island is 18-20 fathoms (33-37m) deep. Two additional deep passes are located just north of this latter pass. The rest of the passes are generally spaced out and are probably effective in keeping all portions of the lagoon well flushed from tidal fluctuations and currents. The northeast Rig St Notes sim : : of Adkup (Erikub) Atoll orientation of the atoll’s long in the vicinity of the pass. reef axis maximizes the constant eS Scale: 1"= .6498 mi. pumping of fresh ocean water . into the lagoon from wave action along the windward ocean reef lagoon slopes. The lagoon gave the impression of a well mixed open system. Fourteen islands occur at Adkup Atoll, but most are concentrated at the southern end where most of the land area is also situated (Figures 36, 37). To the north of this 63 island cluster are found only three islands along the windward reef, and two islands along the leeward reef throughout the rest of the atoll. Rare Marine Species Fig 35. Northwest rim of There was extensive evidence Adkup (Erikub) Atoll in the of recent sea turtle nesting activity vicinity of Aneko (Enego) on the beaches of several islands. Island, an important sea Twenty-three pairs of sea turtle turtle nesting site. tracks were reported from Aneko Scale: 1" = 1.557 mi. Island (Figure 35), a large island near the northwest end of the atoll. Fewer pairs of tracks were recorded on the other islets: Adkup (13 pairs), Ijo-kan (6), Areto-jairok (3), and Looj (4). The collective totals for Adkup atoll rank it third behind Pikaar and Jemo with respect to the level of sea turtle nesting activity recorded during the 1988 expedition. | However, the nesting turtles and their eggs appear to be subject to heavy harvesting pressure. Recent human footprints were found along all beaches where turtle tracks were reported. Numerous nest marker sticks, temporary camps, and the remains of sea turtles and their eggs were also conspicuous. Most likely the turtle harvesting is accomplished by residents of Wotje, but we were not able to visit Wotje and query its islanders. Aneko Island Fig 36. Southeast end of Adkup (Erikub) Atoll near main islands and pass. Scale: 1" = .95 mi. lagoon Bokan-aik — : Boke-lomjan Bokan-kowak i 64 Fig 37. Southeastern end of Adkup_ og, pe (Erikub) Atoll showing sites of marine surveys, with those of special interest indicated by stars. Scale: 1” = .9232 mi. Jaltoneej 7F Bokan-kowak Interviews with the crew and Marshallese from Majuro indicate that Adkup Island is famous for its large coconut crab population. Indeed crabs were reported during the survey, but most all were small. Only one crab approached the large (2 kg) size of the many large crabs observed at Rondik Island (Rondik Atoll). The crabs were similar in size as those observed at Kapen Island (WoOtto Atoll), but were not nearly as numerous. We conclude that the Adkup crabs are also subject to intense harvesting pressure. Giant clams of all four species were observed on the lagoon reefs of Adkup but only Tridacna maxima and Hippopus hippopus were common. There were many more dead shells than live clams, and only one live individual of Tridacna gigas was reported. The lagoon slopes of the eastern perimeter reefs were unusually steep. These habitats may be suboptimal for the giant clam T. gigas at Adkup due to substrate instability. It is also possible that the giant clam populations at Adkup are subjected to heavy harvesting pressure from islanders at nearby atolls. Of the 13 marine sites surveyed at Adkup Atoll, four (7D, 7G, 71, and 7K) were considered unique or exceptional with respect to natural characteristics. All of the sites 65 were lagoon facing sides of perimeter back reefs and reef slopes along the windward side. All displayed complex three-dimensional coral communities on steep reef slopes characterized by moderate to high fish abundance and high coral cover and diversity. Live giant clams were present at all four sites and live coral coverage was 40% or more. At site 7D underwater visibility on the lagoon slope (low tide) was 45 m. The slope at site 7I consisted of a patch reef half buried by a sand talus where wave action and strong currents constantly transport sand lagoonward and down the slope. Site 7K included nurse sharks and several rarer corals among the 30+ species recorded. The windward pass area (site 7J) was also of interest. At the time of my surveys (morning of 23 September 1988) a strong flood tide of two to three kts was entering the channel. The floor of the channel consisted of a hard reef pavement at a depth of 4m which gradually deepened to 15m and transitioned to a sand bottom in a lagoonward direction. The sides of the channel supported exceptional live coral development and abundant reef fish populations. Corals of Adkup A total of 75 species belonging to 35 genera and subgenera were reported from Adkup Atoll based upon observations at the 13 marine sites. The lower totals compared to some of the other atolls reflect the lack of observations on ocean reef slopes, lagoon pinnacles and other reef habitats expected to harbor additional species. One species reported from Adkup at site 7E, Euphyllia glabrescens was not observed elsewhere during the 1988 expedition and is a rare coral elsewhere in the Marshalls. Several common genera or subgenera were not reported at Adkup, including Stylocoeniella, Porites (Synaraea), several mushroom corals, Lobophyllia, Leptoria and others (Table 2). Some of these would be expected to be seen after more intensive surveys. Abundant and common species along Adkup’s back reef flats and shallow lagoon reef slopes along windward perimeter reefs include: fingercoral (Porites cylindrica) Acropora spp (tables), Stylophora pistillata, Favia spp, Montipora spp, Pavona spp, other Porites spp, Acropora palifera, the soft corals Sinularia and Sarcophyton spp, Millepora spp, Pocillopora spp, Astreopora myriophthalma, Echinopora lamellosa, Cyphastrea microphthalma, Montastrea curta, Heliopora coerulea, Platygyra daedalea, Fungia spp, Turbinaria stellulata, Seriatopora spp, Favites halicora, Scapophyllia cylindrica, Goniastrea spp, and Leptastrea purpurea. Abundant and common coral species along the windward facing lagoon reefs on the western perimeter of the atoll include: staghorn and table coral species of Acropora, Pavona minuta, Millepora platyphylla, Favia stelligera, Pocillopora spp, and Montipora Spp. 66 CORALS: COMBINED SPECIES LIST Prior to the 1988 surveys, coral records for the seven areas were extremely limited, a combined 35 species from Bok-ak and Rondik (Wells, 1954) (Table 3). All but four of these species were subsequently reported in 1988, and a total of 168 species belonging to 55 genera and subgenera have now been recorded for the seven areas. Several of the species and one genus (Polyphyllia) are new records for the Marshall Islands. Despite limited deep water and ocean reef sampling, the 168 species is a sizable total comparable to the faunas of Bikini, Ane-wetak (Enewetak), and Arno Atolls (Wells 1951, 1954; Maragos, 1989; Devaney and Lang, 1986), where much more extensive surveys and sampling for corals was accomplished. The species totals for the individual atolls surveyed in 1988 are lower than reported from each of Bikini, Arno, and Ane-wetak Atolls due to the 1988 sampling limitations. 67 IV. DISCUSSION An excellent overview of the feasibility, justification and procedures to establish a system of protected areas in the Marshall Islands is found in Thomas (1989) and covers terrestrial, cultural, and marine factors. The present report concentrates on marine resources, and assesses the consequences of establishing marine parks and reserves (Table 5). The feasibility of other resource uses at the seven studied areas is also assessed since atolls and islands that the RMI does not establish as preserves or parks may be earmarked for other forms of development (Table 6). Most of the Republic’s natural resources are marine resources, and the seven studied areas represent a major proportion of the undisturbed reef systems in the country. Although none of the study atolls is large, the RMI is home to the world’s largest atolls. More atolls are found in the RMI (28) compared to any other country except the Federated States of Micronesia (42) and French Polynesia. But unlike the FSM and French Polynesia, the RMI is comprised of only atolls, table reefs, and low coral islands. As such, the land resources are small and lack rich and abundant soil and groundwater resources. Hence the RMI must look to marine and coastal environments for future economic development. The RMI is also faced with rapid population growth and the need to reduce balance of trade deficits and unemployment. Self reliance is the central theme for the Republic’s future development goals. Fully a quarter of the RMI’s atolls are uninhabited (including Ailinginae which was not studied), and at least one or two others have been temporarily evacuated (Ronlap and Bikini due to concerns over contamination from nuclear testing). Thus the RMI perceives most of these uninhabited areas as major resource development opportunities. The RMI government also recognizes and supports the traditional use of several of these atolls as wildlife reserves or "pantry" reserves, and specifically requested the study team to evaluate the uninhabited areas (and inhabited WoOtto) as possible parks and reserves. Those areas which are not established as protected areas are theoretically open for subsistence activities, resort development, small scale (nature based) tourism, agriculture, mariculture, urbanization or settlement, and industry. However, the inaccessibility, geographic isolation, small land areas, limited fresh water, and vulnerability to typhoons and other natural hazards render most of the areas unfavorable for intensive development (see Table 6). Marine Reserves At least portions of all seven areas visited harbor marine resources and sites worthy of marine reserve and preserve status. The entire reef ecosystems (along with the islands) of Bok-ak Atoll, Rikaar Atoll, and Jemo Island warrant reserve status, a designation which would be entirely consistent with the traditional and cultural uses of these reefs as practiced by the Marshallese for many centuries. All three areas are acknowledged reserves for either nesting seabirds, nesting sea turtles, or both. All three have unique coral reef features and habitats which have been little studied scientifically. 68 Ayiqisee} sey) 40) eeyUo pesjnbes ey} pue sjjoye pue syees sijeysseyy WIEYUON 40 sesn edJnosel jo sedA} snoyeA UeeMjeq diysuoejey “9 e\qeL jueyoduwiun - « juevodw; Ajeyesepow - O jueyodw AleA - juawdojevep jeujsnpu| sJueWO}JEeS JUBUREWIOYg swosey WSIINO} eyeoS |[eWS $J801 JOYJO Je SIOJEM |eOjIE) PUB OOM A\qissod onom Aiqissod rele Bulys|} jelosewwog einyjnoby dnypy pue yIpuoy ‘p01 Ajqissod pue ‘oom elNnyjnoysew/esnjjnoenby (Ajuo Seounoses eUeW) yIPUOY “ONOM B 6x01 ‘dnypy Hulusi} 9» Buleyye6 eouejsisqns seeie uoljeeioes @ Swed dnypy pue yIpuoy Jo suc|yOd ‘OOM ‘exOL SOAJOSOJ B SOAIBSOld sjjoye JOuJO JO SUOILOd pue 6wer ‘Weeryid ‘ye-40g SSS Ess MS S333Y GNV STIOLV ALVGIGNVS SoS Sy < GS GS SS ASS) “5 Sy 4} cS ky cS KS Ss 6S S3SN SSOuNOSSY : 2Q */ 2B YS SLT Y Ae) eA LS SV SS L/S SYS PSPS dS 2 9 > o (07,2 YE <> VASA Ky SS > SS Oo: Oy 2 cS Ss & SOY SESE, > x wo OWA YAN ARO) Ha GOS 2? » o K2 Cg & oe S > Oy Ae) Ro GS eo S > —WAYAS ‘AS > oOKS ES &/ © a S R VIHALIS 69 The lagoons of Bok-ak and Pikaar also support huge populations of several of the smaller species of giant clams which is perhaps the most important reason to establish them as marine reserves. The lagoons also support exceptional populations of reef corals and reef fishes and the whole ecosystems are in pristine condition. The designation of any area for reserve or protected status will require the cooperation of persons with traditional rights to these areas. At a smaller scale there are unique or exceptional coral reef habitats at the other atolls (Wotto, Adkup, Rondik, and Toke) that also merit marine reserve designation due to the abundance, diversity or unique features of coral reef populations and reef features. For example, the pink sand beaches and high dune systems, blue coral moats and flourishing lagoon pinnacles at Rondik deserve special recognition as do other lagoon reef formations at the other atolls. Marine Parks and Recreational Areas Several of the atolls offer major advantages with few disadvantages to support marine park designation. The variety, accessibility, safety, and pristine condition of lagoon reefs (and perhaps ocean reefs as well) at WOtto and Toke support the entire atolls being designated as National Marine Parks. Portions of Adkup and Rofdik also contain diverse and accessible reef habitats to support at least Regional Marine Park status. These parks would serve both the residents of the RMI and visitors and would be oriented to provide recreation and educational opportunities. Visiting tourists to these parks could provide the fees to support park management, jobs, travel, and educational opportunities for Marshallese residents, including school students. Of the four areas, Wotto has the most potential to serve as a national park, because of an onsite residential population interested in pursuing marine park and nature based tourism. WOtto also contains food and water supplies, and its airstrip allows the atoll to be serviced by weekly commuter flights from Majro or Kuwajleen (Kwajalein) Atolls. The large population of American defense workers at Kwajalein may find the opportunity to visit WOtto an attractive prospect. Many other people from the urban settings of Majro and Kuwajleen may also be interested in experiencing the natural and cultural resources of WOtto. The atolls of Toke, Adkup, and Rondik are somewhat less accessible to serve park visitors. The nearest airstrip and inhabited population from Toke is at Utrok atoll, and the Utrok islanders need to be queried as to their interest in park designation for Toke. Likewise, Adkup and Rojidik are close by other atolls (Wojte and Rofllap) where airstrips are present. Wotje is inhabited while Ronlap has been temporarily evacuated. On a long range basis, park designation and development is theoretically possible for these areas, as well as for Ailinginae (which was not visited) which is also near Ronlap (see Figure 1). Park development at Pikaar, Bok-ak, and Jemg is not feasible due to hazardous access, remoteness, vulnerability to large waves and typhoons, and lack of potable water. Heavy visitation to these areas could also disturb wildlife resources (especially sea turtles, 70 sea birds, and possibly clam populations). Any form of physical development at these atolls would compromise the value of a reserve, and could disrupt reef and island ecosystems. As with any other proposed use, park designation of these areas will require the cooperation of persons with traditional rights to these areas. Subsistence Activities Very limited harvest of sea turtles and seabirds for ceremonial purposes has been traditionally practiced at Bok-ak, Pikaar, and Jemg. Such visits, if continued to be limited to one or two per year, do not pose a danger to the vulnerable wildlife at these sites and would be in keeping with long established cultural practices. However, harvesting of larger numbers of wildlife for purely subsistence purposes would be disruptive to bird, turtle and clam populations, and may endanger their status as the most important populations in the RMI. The remaining four areas are suitable for traditional level subsistence activities and would be compatible with park reserve designations if planned properly. For example, marine resources and sites at WOtto used for subsistence activities could be identified and sustained for such uses, and visitor or recreation sites would be best located at separate sites. There should be controls established over the taking of rare marine species (sea turtles, their eggs, and giant clams) from Rofdik and Adkup Atolls to ensure that important breeding populations are not depleted or threatened. One way to accomplish this is first to establish critical habitat areas as reserves. Radiological Contamination Consumption of coconuts or coconut crabs from Rondik Atoll may pose as health hazards to islanders. In addition, subsistence use of terrestrial resources from Ailinginae, if any, should likewise be discouraged until radiological surveys document that consumption of these resources will not pose a hazard to public health. It is possible but highly unlikely that consumption of marine resources from these two atolls as well as from Toke would pose a health problem. Resident coconuts, breadfruit, pandanus, and coconut crabs from Rondik Atoll may be contaminated with the radionuclides cesium-137 and strontium-90. These radionuclides were generated during the BRAVO hydrogen bomb test at Bikini in 1954 and carried with the nuclear fallout from the blast. Although Bikini is located over 125 NM west of Rofidik (and 100 NM west of Ailinginae and Romlap), fallout from BRAVO was reported to have been carried into upper atmospheric winds and to the east, where some of it eventually rained down on Rofilap and Utrok Atolls, which were inhabited at the time. The fallout also probably rained down on other nearby atolls, but since they were uninhabited, evidence of fallout must be derived from analysis of plants and soils. Rojidik falls within a straight axis between Rofilap (25 NM to the west) and Utrok (140 yal NM to the east) (see Figure 1). Thus it is highly likely that Rondik, and perhaps Ailinginae were contaminated with the fallout. Studies at Bikini and Ane-wetak Atolls after the nuclear testing period reveal that coconut trees (especially the living nuts) and other crops take up and bioaccumulate cesium-137 and strontium-90 in their tissues. The concentrated radiation levels in the coconuts posed a much greater health risk than radiation in the soils, because resettled Bikinians subsisted regularly off locally grown but contaminated coconuts between 1969- 1978. Excessive whole body dose counts of the Bikinians in 1978 measured by Brookhaven National Laboratory prompted the evacuation of the Bikinians from their home atoll on short notice in 1978. Coconut crabs as well as humans subsist on coconuts, and the crabs can also bioaccumulate Ce-137 and Sr-90 in their tissues by foraging off contaminated nuts, as studies by Lawrence Livermore National Laboratory at Ane-wetak Atoll have demonstrated (BARC, 1984; 1985; 1986). The radionuclides could then be passed up the food chain to man if he eats contaminated coconut crabs. Unlike the small crab populations at Bikini and Ane-wetak, Rofidik supports huge coconut crab populations. This raises the possibility of a greater public health hazard since there may be many more potentially contaminated crabs at Rondik. Regular consumption of coconuts, breadfruit and other crops at Rondik is more likely and could also pose a risk. Toke and Ailinginae Atolls were also within the fallout zone. Although Toke’s edible vegetation and coconut crab populations are very small, the extent of Ailinginae’s is not known since the team was unable to visit the latter atoll. Thus radiological surveys may be warranted for Rondik as well as Toke and Ailinginae to document the extent of radiation hazard from ingesting food crops and coconut crabs. Romnlap, and Utrok are presently being monitored for radiation by Lawrence Livermore National Laboratory. Mariculture The shallow protected lagoons and broad reef flats and shelves within the Northern Marshalls offer ideal locations for certain forms of mariculture development. Several marine species with mariculture potential were observed during the expedition including: giant clams (Tridacna, Hippopus), topshell (Trochus), black-lipped pearl oysters (Pinctada), milkfish (Chanos), mullet (Mugil) and reef groupers (Epinephelus). Mariculture would be more feasible within atoll reefs and lagoons accessible by air and sea transportation and near population centers. Mariculture thus would be feasible at WoOtto, with its airstrip, protected anchorage, resident population and proximity to urban Kuwajleen. Toke is somewhat less feasible since the nearest airstrip and residential population is at Utrdk. Similarly, Adkup is removed from Wotje, the nearest population and airstrip. Mariculture would more likely be developed at the populated atolls (Utrok and Wotje) rather than at their uninhabited neighbors (Toke and Adkup). Rojfidik is less feasible for mariculture due to the lack of inhabited atolls nearby, although Ronlap, which 72 is serviced by an airstrip and protected anchorage was only recently evacuated and may be eventually resettled (Rongelap Reassessment Project, 1989). The remaining reef and atolls (Bok-ak, Pikaar, and Jemg) are not feasible due to remoteness, lack of safe access, lack of safe anchorages, and relatively inhospitable living conditions. Extensive mariculture development at Pikaar and Bok-ak might also conflict with other values and uses such as protected reserves or preserves. The large giant clam populations in the lagoons of Pikaar and Bok-ak may eventually serve as important brood stock and a source of giant clam seed, if the smaller giant clam species become severely depleted in the RMI. For this reason, the giant clam populations at the two atolls should be maintained as reserves indefinitely. Agriculture Although the focus of this report is on marine resources, a few comments on the feasibility of agriculture at the seven studied areas can be offered. Due to lack of water and good soil, agriculture at Pikaar, Toke and Bok-ak is not feasible. Although Jemo Island has thicker soil and more abundant water, agriculture is also difficult due to hazardous access and limited land and settlement options. Rondik Atoll should not be considered feasible for agriculture unless radiological surveys and analyses project it to be safe. Subsistence level agriculture is already practiced on inhabited Wotto, including some copra harvest. Likewise, Adkup is subject to limited copra harvest, but is small land areas limit greater agricultural development. Commercial Fishing Commercial fishing activity by resident Marshallese would be feasible at inhabited WoOtto, which has accessible ocean fishing grounds. Fishing vessels from Wotje, Utrok, and Rofilap could also fish the coastal waters of Adkup, Toke, or Rofidik, especially for game fish and tunas. Controlled commercial fishing in the lagoons of uninhabited atolls is also possible unless it competes with the subsistence needs of the nearby inhabited atolls which own or traditionally control lagoon fishing grounds. Commercial fishing is less feasible at Pikaar and Bok-ak atolls due to remoteness. Fishing in the lagoons would be further discouraged by numerous reefs and hazardous access through the single narrow meandering passes. Permanent occupation of the atolls to promote commercial fishing would be extremely disruptive to rare marine species, reef life and nesting seabirds. Controlled commercial fishing for tunas and other migratory species along the ocean sides (territorial waters) of all atolls is feasible but should be limited to Marshallese and monitored to avoid poaching of turtles, giant clams, and other rare species within or near designated marine reserves. Small Scale Tourism Small scale tourism in this report means small lodges or beach cabanas, limited to about 20 rooms which take advantage of the natural features, scenic beauty, and 713 cultural resources within the vicinity of the tourism facilities (e.g. nature based tourism). This style and level of tourism at Wotto is very feasible given the interests of the islanders to pursue it and the host of natural amenities and attributes. Reliable sources of water and food appear feasible to obtain and the major infrastructure requirements would include the accommodations, power generation (for lights and refrigeration, ceiling fans, etc.), catchment water storage, and waste disposal. WoOtto is already serviced by weekly commuter air service and has a protected anchorage. Thomas et al (1989) provides a detailed account of the feasibility of small scale tourism at Wotto Atoll. Many existing marine resources and features could be incorporated into a tourism operation. Visitor destination attractions include flourishing coral communities at safe and accessible lagoon areas. Corals, fish, giant clams and other reef life could be observed via diving and snorkeling. Mariculture and fishing activity at the atoll could provide fresh fish and shellfish to feed visitors. Beaches and other coastal sites at islands away from the existing village could also be visited. Interpretive and educational displays regarding marine life could be established. Swimming, gamefishing, diving, some boating, and sailing activities could also be included. The focus of small scale tourism at WoOtto should be nature based, with considerable emphasis on marine resources. Toke Atoll also has many attributes to support tourism visitation, but accommodations for tourists would be best placed on Utrdk, subject to the views, approvals and guidance of the Utrok people. The survey team could not find the time to visit the Utrok islanders and obtain their views on tourism. The water and land resources at Toke are too limited to support development of self-contained permanent tourism accommodations. Toke would best serve as a day visitor area or for limited overnight "rustic" camping. Access to the atoll would be gained by boat from Utrok. Land resources at Toke are too limited to justify an airstrip, and dredging and filling to construct a reef runway could have major adverse impacts on marine resources. Similar limitations on tourism development apply at Rofidik and Adkup atolls. Although land resources are more abundant, the lack of permanent residents would require infrastructure development for accommodations, power, water supply and transportation. Airfields would most likely be required to attract tourists to the atolls, and reef runways would be needed. Tourism development would be expensive and would need to be well planned to avoid serious socioeconomic and environmental impacts. Most importantly, tourism development, if any, would require the approval and support of the traditional land managers, owners, and users of both atolls. Tourism development is not feasible at Jemo, Pikaar and Bok-ak due to remoteness, hazardous accessibility, lack of reliable fresh water, lack of a permanent work force, and the significant expected economic costs and environmental impacts. Of the three areas, only Bok-ak has sufficient land for an airstrip, but substantial bird nesting on the islands poses serious constraints. Disruption of nesting activity and collisions between birds and airplanes are highly likely in nesting areas and contrary to the designation of the atoll as a wildlife refuge. Safe boat access to either Bok-ak or Pikaar would require major dredging and other coastal construction activity in the vicinity of the existing 74 passes. Additional lagoon reefs may also need to be dredged or knocked down to provide safe access across the lagoons to the destination islands. The clearing of safe navigation channels could have major adverse effects on coral reefs and giant clam populations, and may also disturb some sea turtle nesting. Major widening or enlarging of the passes could also lower lagoon water levels, exposing and killing many shallow reef flats, disrupting lagoon circulation, and possibly degrading coral reef and giant clam habitat (see Figure 38). For these reasons permanent occupation or settlement of Pikaar and Bok- ak for tourism and other purposes (e.g. resettlement, industrial, development) would pose serious threats to the natural resources of most value at the atolls and should be strongly discouraged. Variations in the depths of living reef flats: revisiting some widely held assumptions It is generally thought that living reef flats on coral atolls and barrier reefs can grow no higher than mean low water due to the requirement of the reef building organisms (e.g. corals, coralline algae), to be regularly immersed in sea water for survival. As growing coral reefs reach the sea surface, further upward growth is inhibited while lateral growth lagoonward and seaward can continue (see Figure 39). Over time and with stable sea level, cessation of upward growth and continued lateral reef expansion would result in the formation of wide reef flats. Their widespread occurrence is indisputable evidence on the limitation of marine organisms to grow above a level of regular exposure to sea water. One widely observed exception to the "mean low tide rule" is the presence of elevated living coralline algal ridges along the margins of many windward reef flats of many atolls, especially those of the RMI (Tracey et al, 1948. Wells, 1954; 1957a, b). The fact that these ridges occur only along windward reef flats suggests a relationship between the ridges and the wave action generated by the tradewinds. The most widely supported hypothesis is that the constant wave action generates wave wash or splash that constantly bathes the ridges, and keeping the reef building marine organisms (primarily crustose coralline algae and some reef corals) alive. This wave wash can regularly immerse surfaces up to two or three feet above mean low water during low tide conditions on the ocean side of the reefs which explains why ridges at this elevation remain alive. Wave wash passing over the ridge then moves downhill toward the lagoon and mixes with lagoon waters, eventually exiting through the passes or over the tops of reefs along leeward sides. Another less widely known exception to the "mean low tide rule" became evident after the visits to Bok-ak and Pikaar Atolls. Here, not only were windward algal ridges found to be alive and growing above mean low water, but so were many other lagoon reefs. The elevated living nature of some lagoon reefs were reported by Fosberg et al (1956) and Fosberg (1988) at Pikaar and Bok-ak. These authors and the 1988 field team also observed the elevated lagoon water levels at "low tide" at these atolls. Similar elevated lagoon and leeward reef flats were also reported by Wells (1951) at Arno Atoll’s northern sublagoon (Namdik), later corroborated by Maragos and Lamberts (1989). I also "sjao1 JuobioWa Buyjjiy pue jaro] sayem uoobe| Buuemo, ‘Ajyoinb ayow uleJp 0} sIa}eEM BSN PjNOM 4j9a1 ||OVe ue yons ybnoiy} jauueyo deep e Bung = ||eao] Jayem UCOCbe| JoyBiy 0} asuodses ul pue MO} J8JEM JULJSUOD 4O BSNedeq joao] ~«=CsBS_)~——sC eNO -~—FEWOU eaoge moh sjaoJ au, ‘JOJeM SS80xX8 9} UIBIP O} SjouUeYyo doap ‘abie| yo yor] 9u} puke Sj9e1 pyempuim J8Ao sayem Buidwnd uoloe aAeM JO a9snedeq Sjaag] Joyem ucobe; payeagjs eaey (ueeyiq pue ‘paunjoid SI YoIyM “ye-40g se yons) sjjoye aWOS ‘sucoCbR| jjoye + pasojoue-iwas YyOnoiy} sjouueyd BHuljnd jo sae OQSIBAPE 9IQISSOd ‘ge bi4 $jaa1 pasodxa Surry pue paray sayem Suisamoy] ‘uoode; ayy wos “TL PHAgis 1 4vo 1ovy 1uoey 7 Ppinb azow suicip 131eM sjuagin> 139u01S Jouueyp ny NV IDO JMO J2as ayy ysnory) sassed daap ou Ajjenqta ym [fay xy , jeai 22] 4300 UMOP j2a3 paempuim S2]4I12) JEM 43A0 pue dn 13;em sduind uonse aaey [aaa] 19eM payeaary LF “Desai ee o 4 BYP 4 uote N a (Troy F8uce], yesyog uT-uoTey NV IDO IOV 1oyy 1uoey, 76 A. Reef Platform at lower sea level stand (during previous glaciation) leeward side ¢—— windward side sea level (s.I) 20,000 yrs before present (b.p Me B. Submerged Bank (during post-glaciation period of rising Re? level) ABWAApRAIIeaerrr OOO Om s.1. 8,000 - 10,000 yrs. A b.p. s.l. 20,000 growing coral reef yrs. b.p. mean high tide (MHO) wave action present s.I- mean low tide (MLW) reef flat and lagoon coralline algal submerged ridge at MLW or leeward reef slightly higher oN pinnacle s.l. 20,000 reef yrs. b.p. (W6tto, Rofidik, Kuwajleen Atolls) wreamym reef line (MLW) I /1 1) raised living reef | Gage islands KAAS j .. submerged reef contour ROO raised dead reef submerged bank or 100m contour Figure 39. Progressive development of reef flat features on atolls. D. Complete Atoll E. Open Raised Atoll S aeN~ MHW MLW present s.I. leeward reef flat reaching MLW s.1. 20,000 —* yrs. b.p. coral-algal dam & spill- way above MLW (living) s.1. 20,000 YiSeD!p).—$==» F. Nearly Closed Raised Atoll 2AM raised (emergent) reefs (dead) windward reef flat projecting patch above MLW reef (Toke, Utrok, Adkup Atolls) stepped reef massive coral-algal ridge (living) ribbon reef (Bok-ak, Pikaar Atolls) land blocking lagoonward movement of water pumped by wave action (Kanton Atoll, Phoenix Islands, Enen Kio [Wake] Atoll) 78 observed elevated lagoon and leeward reefs at Namdik (Namorik) Atoll in 1971, and former residents of nearby Ebon Atoll also noted the elevated nature of perimeter reefs (pers. comm. to J. Maragos by N. Neimon). All five of these atoll lagoons have two things in common: 1) exposure of windward reef flats to prevailing wave action generated by the trade winds, and 2) restricted passages or no passages through the reef. These two factors are interrelated in explaining the presence of elevated living lagoon and leeward reefs (Figure 39). Progressive development of reef flat features on atolls explain how contemporary reefs in the northern Marshalls can grow above low tide level without the need for a higher "Holocene" sea stand. Stage A. Antecedent reef platform at the end of the previous ice age. Stage B. Subsequent melting of the glaciers causes sea level to rise, drowning the reef platform and renewing upward coral reef growth. Stage C. Upward reef growth eventually reaches sea level (mean low tide) on the windward side. Coralline algal ridges projecting above mean low tide may develop in response to constant wave action. Stage D. All perimeter (annular) reefs grow upward to mean low tide level except where passes cut through the reef. Coralline algal ridges and windward reef flats continue to broaden. Stage E. Eventually the passes close off to the extent that water pumped into the lagoon by constant windward wave action is greater than can exit the passes at low tide, causing average lagoon water levels to rise. Remaining (leeward) perimeter reefs, now constantly submerged, begin to grow upward, forming coral-algal dams, spillways, and perched lagoon reefs. Stage F. Storms naturally cast sand and rubble on top of windward reefs, forming cays, ramparts or islands, or man builds causeways along windward reefs to connect islands. In either case they block the pumping of seawater into the lagoon by wave action. As a result, water levels in the lagoon drop, permanently exposing raised reefs which dry out and die, leaving intact "fossil" raised reefs. The cutting or enlarging of passes through perimeter reefs can have the same effect by draining lagoon waters more quickly and lowering average water levels (see Figure 38). The restricted passages result in more water entering the lagoon over windward reefs than can exit through passes at low tide. In response to the restricted discharges, average lagoon water levels increase with the excess water spilling over leeward reef flats as well as through the passes. If the passes begin to close off, restrictions increase, causing lagoon water levels to rise further. Higher water levels in the lagoon, especially during low tide, result in more and more water spilling over the leeward reef flats until the latter are constantly immersed even at low tide. Prolonged immersion may in turn ultimately cause leeward reef flats to grow upward, since the reef organisms are no longer limited by exposure at low tide (see Figure 39). Eventually, perimeter reefs along leeward sides of the atoll grow upward. Supplementing the coralline algal ridges along windward reef margins are smaller coralline algal ridges and coral-algal dams and spillways along leeward reefs. Lagoon reefs also grow upward in response to the progressively higher lagoon water levels. Maximum upward reef growth depends upon the magnitude of prevailing wave action and the extent of open reef flats along the windward side of the atoll. Some of the kinetic energy of wave action is converted into potential energy by pumping water up on TS, higher reef flats. Wave action can constantly pump ocean waters into the lagoon over the ridge and reef flats. If lagoonward water movement is blocked by the presence of islands or rubble ramparts created during tropical storms, lagoon water levels could drop. Man’s intervention, either by building causeways along windward reefs (which blocks wave pumping of water into the lagoon), or by enlarging passes through leeward reefs (which drains water more quickly from the lagoon), can also lower average lagoon water levels. The lowered water levels could then result in the emergence and death of exposed reefs, which may have occurred at Kanton Atoll, where a near continuous causeway was built around the perimeter reefs of the atoll, and where recently exposed reefs were observed (Jokiel and Maragos, 1978; Smith and Jokiel, 1978). The hypothetical evolution of atoll reef flats based upon the above scenario is depicted in Figure 39. Geologists often rely the elevation of previously intact fossil reef flats to estimate the extent and age of relative sea level stands in various parts of the world. Two implicit assumptions in many of these studies is that all or most modern living reef flats grow no higher than mean low water elevation, and that intact previously living reef flats found emerged on present day reefs must have formed when relative sea level was higher. Based upon the 1988 observations at Bok-ak and Pikaar, supplemented by the observations at other atolls (Arno and Namdik), the first, and perhaps both of these assumptions may be incorrect. First, in the case of Bok-ak, Pikaar, Arno, Namdik, and perhaps other atolls, many present day living reef flats occur above low tide level due to factors other than a higher sea level stand. More importantly, some of these same reefs may become reexposed due to natural factors, such as islands, cays, and rubble ramparts forming on the windward sides of atolls with elevated leeward and lagoon reef flats, thereby blocking lagoonward movement of water pumped by wave action. As a consequence, the hypothesis of a higher Holocene sea level stand some 4,000-6,000 years ago that is based upon the evidence of higher stands of recent reefs less than one meter above present sea level may need to be reexamined. The complex interaction of prevailing wave action, restricted passages through reefs, open windward reef flats, the frequency of storms, and other factors can alternatively explain the upward growth of living reefs above normal low tide levels and their subsequent reemergence. Reliance on evidence from prehistoric reef stands in support of hypotheses on previous sea level stands must involve an examination of the geomorphology, oceanography, and geological history of the reefs in question. Rapid Marine Field Assessment Procedures The results of the 17 day visit to 95 marine sites and other numerous shoreline sites at seven atolls and reefs in the northern Marshall Islands demonstrate that qualitative data gathering procedures can be very useful in describing marine areas. Preliminary assessment of biological and ecological diversity can be accomplished without the need for transect and quadrant surveys if the purposes and goals of such studies are clearly identified in advance. With the primary emphasis of the 1988 expedition on evaluation of natural diversity and feasibility for park and protected area development, it was 80 possible to collect valuable information on species, habitats, bathymetry, geomorphology, and oceanography, relying primarily on shallow water snorkeling observations. Coupled with the availability of aerial photographs, and previous map sources, field work was designed to sample a greater variety of habitats than would have been otherwise possible. Although the literature was scant and the opportunity to interview knowledgeable informants limited (since all but one of the sites were uninhabited), good maps and aerial photographs can be consulted to improve the efficiency of field work. Modern satellite imagery from the French Satellite SPOT now has resolution (10m) which can supplement photo interpretation of maps, especially where conventional aerial photographs are not available. Collectively these data acquisition strategies may become increasingly important in evaluating the multitude of marine resources and habitats in the South Pacific. With many thousands of reefs and islands, and hundreds of atolls, many of which are remote, innovation will be required to allow rapid evaluation of particularly valuable areas. As population levels and development pressures increase, more and more natural marine areas will become vulnerable to exploitation and degradation. A systematic inventory and evaluation of candidate marine protected and park areas throughout the tropical Pacific will become an even more important goal of proponents of both conservation and development. 81 V. SUMMARY Six atolls: Bok-ak (Taongi), Pikaar (Bikar), Toke (Taka), Wotto (Wotho), Rofdik (Rongerik), and Adkup (Erikub) and one table reef (Jemg) were surveyed during a 17-day expedition in September 1988 to the Northern Marshall Islands to describe coral communities and reef formations as part of a larger natural diversity survey. Only observations using snorkeling gear, underwater writing slates and underwater cameras were possible during the approximately 2-day visit to each atoll. A total of 95 sites were surveyed, ranging from 12 to 20 sites per atoll. Additional observations were made during boat travel and walks along shorelines of islands. Over 160 species of reef corals belonging to 55 genera and sub genera were reported from the seven areas, including several species and one genus as new records from the Marshall Islands. The abundance and distribution of corals varied from one atoll to the next and may reflect geographic isolation from adjacent reefs, limitations on habitat diversity (in the case of Jemo), and limitations on larval recruitment (in the case of Bok-ak and Pikaar). Several of the coral communities and habitats were unique or have not been previously described. Many sites displayed exceptional coral development, and sites of special interest were identified on maps and are highlighted in the report. The reef geomorphology of the seven areas is also described and each belongs to one of three distinct physiographic categories: 1) small semi-enclosed atolls (Bok-ak, Pikaar, Toke) ii) larger open atolls (Wotto, Rondik, Adkup) iii) exposed table reef (Jemg) Lagoon and adjacent perimeter reef formations at Bok-ak and Pikaar are elevated two or more feet above mean low tide level. These elevated reefs are living and perhaps growing, and are maintained by a combination of water being pumped into the lagoon from wave action on the windward sides, and the inability of the narrow passes to drain water from the lagoon at an equivalent rate during low tides. Unique features at both atolls associated with the elevated reefs include overhanging ribbon reefs, coral-algal dams, spillways and steep water level gradients in each atoll pass during low tide. Navigation through the passes during low tide is treacherous due to the narrow and meandering configuration of the passes, and the turbulent water flow caused by a two to three foot drop from the higher lagoon water levels over a short distance. Huge undisturbed giant clam populations (Hippopus sp. and Tridacna spp, but not the largest species, T. gigas) occur extensively in the lagoons of both Bok-ak and Pikaar. The author has never observed such high giant clam densities elsewhere in the central west Pacific. Furthermore, extensive sea turtle nesting and swimming activity was reported at Pikaar. Toke Atoll is more properly intermediate in form between the semi enclosed atoll and open lagoon atoll groups. Like Bok-ak and Pikaar, Toke atoll has a single narrow pass on the western side. Unlike the other two atolls, Toke’s lagoon is deep and lacks the ribbon reef formations. Lagoon patch and pinnacle reefs are more circular in form, and lagoon reefs are not elevated above mean low tide levels as noted for Bok-ak and Pikaar. Giant clam populations at Toke are smaller but include live specimens of the rare 82 largest species, Tridacna gigas. The only sighting of a hawksbill sea turtle during the expedition occurred in northeast Toke lagoon. Although uninhabited, Toke is near Utrdk (Utirik) Atoll. The owners of Toke Atoll reside at Utrok, and Utrdk fishermen occasionally visit Toke to harvest fish and shellfish. Small boat navigation through the channel and landing small boats at islands along the lagoon shorelines are relatively safe. Several important reef areas of special interest due to good coral development and diversity were observed at Toke Atoll. Jemo is one of only five table reefs in the Marshalis and the only one which was visited. Due to the lack of a lagoon, and heavy exposure to ocean waves and swells from virtually any direction, Jemo’s reefs have unusual geomorphology and limited coral development. Only the southwestern end of the reef is shallow enough to form a reef flat exposed at low tide, upon which Jemo Island is situated. Elsewhere, Jemo’s reef crest does not emerge at mean low tide and is dominated by a curious but extensive network of sand covered surge channels oriented in a north-south axis The outer reef margins consist of scoured sloping pavements with limited coral growth. Coral species diversity is low, less than half of that of the other areas surveyed, and is probably controlled by exposure to heavy waves and limitations in habitat diversity and abundance. The very steep, deeper reef slopes showed higher coral diversity. The best coral development occurred within a semi-protected reef indentation on the north side. Jemo’s beaches support the second largest sea turtle nesting population observed in the Marshalls (only Pikaar’s population is reported to be larger). Jemo island itself is relatively inaccessible due to hazardous reef conditions, lack of protection from waves, and the lack of a safe approach to the island except during calm seas. The numerous sharks and large swells would also discourage snorkeling and diving interest at Jemo. Wotto, Rondik, and Adkup are the largest of the atolls visited during the September 1988 expedition, but are relatively modest in size compared to many other atolls in the Marshalls. All three have large passes, deep open lagoons, and a diverse set of lagoon and ocean reef habitats. Wotto and Rofdik in particular have unique and aesthetically interesting coral and beach habitats, including pink sand beaches. A blue coral reef moat occurs at Rofidik, and diverse and flourishing coral and clam habitats occur at Wotto. Boat passage through channels and lagoon access to islands are safe. The inhabitants of Wotto have expressed strong interest in promoting tourism at their atoll and prior to our expedition, requested financial and technical assistance to develop a tourism facility. From the standpoint of uniqueness of reef forms, Bok-ak, Pikaar, and Jemo all warrant special recognition and research interest. The huge giant clam populations at Bok-ak and Pikaar, and the large sea turtle nesting populations at Pikaar and Jemo also argue for marine and island reserve designations. When coupled with the extraordinary seabird nesting populations at Bok-ak and lesser but important bird populations at Pikaar and Jemo, all three areas should be established as part of a system of national ecological reserves (Thomas et al 1989). Such designation will require the cooperation of persons 83 with traditional rights to these areas. All three of these areas are unsafe with respect to boat access and landing, which reinforces their preferred status as limited entry reserves. In particular, the clam and turtle populations would be vulnerable to over exploitation, and access to the three areas should be strictly controlled in any case. Proposals to enlarge or widen the reef passes at Bok-ak and Pikaar to promote safe boat access would result in major and perhaps catastrophic impact to lagoon reefs. Aside from direct destruction of reefs, low tide water levels in the lagoons would probably drop, exposing and killing the living tops of elevated reefs throughout the atolls’ lagoons. Circulation in the lagoons would also change and possibly harm resident giant clam populations. Toke and Wotto Atolls seem well suited as possible national marine parks open to both tourism and resident recreational use. Residents could manage and monitor the areas as marine parks, perhaps as part of small scale tourism development. Access and landing at both atolls is relatively safe, accessible snorkeling areas exist, and diversity and development of coral reef environments is high. Although residents expressed strong interest in tourism and marine park development at Wotto, it was not possible to query the Utrok islanders on their views for similar development at Toke. Nature based tourism and park use would benefit the natural and cultural resources in both areas if properly planned and managed. Adkup and Rofidik atolls likewise have many attributes supporting marine park designation. Adkup, although uninhabited, is heavily utilized as a traditional harvest ("pantry") area by visiting fishermen and islanders from nearby Wotje Atoll. Any future designation of portions of Adkup for marine park and sanctuary status should reflect the coordination with and the views of the traditional resource users and owners. Although specific areas of both Adkup and Rofidik are suitable candidates for marine parks or reserves, there is less justification to designate the entire atolls as reserves or parks. Western Rondik Atoll could not be surveyed, but Bok (Bock) island is suspected as an important sea turtle nesting area due to its extensive white sand beaches. Follow-up observations could confirm the importance of the island for turtle nesting. Rondik Atoll’s pink sand beaches, blue coral moat, and luxuriant lagoon coral formations would be of great recreational interest to both visitors and residents. The large populations of coconut crabs at Rondik may eventually be heavily harvested since coconut crab is a favorite islander delicacy. However, Rondik was exposed to fallout from the "BRAVO" atmospheric thermonuclear detonation at nearby Bikini Atoll in 1954. Fallout from the blast contaminated Bikini and the atolls of Ronlap (Rongelap) and Utrok. Given the close proximity of Rondik to these other atolls, a radiological survey of the atoll is warranted to determine possible health hazard from ingestion of crabmeat and coconuts. Coconuts are known to concentrate the radionuclides cesium-137 and strontium-90 in their tissues, based upon sampling of coconut trees at both Bikini and Ane-wetak (Enewetak) Atolls conducted by Lawrence Livermore 84 Laboratories. Since coconut is the preferred food of the crabs, radiological contamination of coconut crabs is a definite possibility. Consumption of Rofdik coconut crabs should therefore be discouraged until radiological tests have determined the crabs are safe to eat. 85 VI. ACKNOWLEDGEMENTS This study of reefs and corals was part of a joint project of the South Pacific Regional Environment Programme (SPREP), the East-West Center Environment and Policy Institute, the U.S. government, and the Republic of the Marshall Islands (RMI) to assess the ecological and cultural conditions of selected reefs and islands in the Northern Marshalls to determine their suitability as candidates for a system of the RMI protected areas. The study was requested by the government of the RMI. Substantial financial support for the project came from a grant from the John D. and Catherine T. MacArthur Foundation, and matching support was provided in the form of scientific participation by the U.S. Army Corps of Engineers, the East-West Center, the National Marine Fisheries Service, and the U.S. Fish and Wildlife Service. This project also contributes to the RMI Coastal Inventory and Atlas program supported by U.S. Congressional appropriations to the U.S. Army Corps of Engineers and as described in the RMI Plan of Action for Water Resources (OEA, 1989). A number of officials in the Republic of the Marshall Islands provided critical support for the project: we thank Peter Oliver, Special Assistant to the Chief Secretary; Oscar De Brum, Chief Secretary; the late Stephen Muller, Director of the RMI Marine Resources Authority; Larry Muller, Captain of the patrol vessel Ionmeto I, and the vessel crew; Gerald Knight, Director of the Alele Museum; Alfred Capelle, RMI Resource Protection Officer; Abacca Anjain, Acting Secretary of Interior and Outer Island Affairs; and the atoll owners and land managers (alab) for granting access to their island and reefs. We thank the project leader, Dr. Lawrence Hamilton, Research Associate, EAPI for administrative support, and the late Dr. F. Raymond Fosberg of the Smithsonian Institution for sharing his extensive experience and advice on the Marshall Islands. Finally, I extend my appreciation to the other members of the expedition who assisted on the marine surveys, especially Peter Thomas, Virgil Alfred, Paul Maddison, and John Naughton; and to Linda Mizuguchi, Karen Tomoyasu, Chris Cabacungan, Jan Eber, Jody Oyama, Jennifer Gorospe, Mary Hayano, Gidget Tsui, and Karin Z. Meier for word processing and/or editing support. I thank Jim Laurel of Aspect Software Engineering Inc. and Karin Z. Meier of CORIAL for digitizing the maps, and preparing most of the map figures for this report. Finally, I thank Mr. Alfred Capelle for providing the correct spellings of place names used in the text and added to the maps. 86 VII. REFERENCES Bikini Atoll Rehabilitation Committee, 1984. Report No. 1. Resettlement of Bikini Atoll: Feasibility and Estimated Cost of Meeting the Federal Radiation Protection Standards. Submitted to the U.S. Congress, House and Senate Committees on Interior Appropriations, Wash. D.C. and Berkeley, CA. Bikini Atoll Rehabilitation Committee, 1985. Report No. 2 FY 1985. Submitted to the U.S. Congress, House and Senate Committees on Interior Appropriations. Washington D.C. and Berkeley, CA. Bikini Atoll Rehabilitation Committee, 1986. Report No. 4 Status March 31, 1986. Submitted to the U.S. Congress, House and Senate Committees on Interior Appropriations. Wash. D.C. and Berkeley, CA. Bikini Atoll Rehabilitation Committee, 1988. Summary Report (Report No. 6) July 22, 1988. Submitted to the U.S. Congress, House and Senate Committee on Interior Appropriations, Wash. D.C. and Berkeley, CA. Dahl, A.L., 1980. Regional Ecosystem Survey of the South Pacific Area. Technical Paper No. 179. South Pacific Commission, IUCN, Noumea, New Caledonia. Devaney, D.M., and J.C. Lang, 1984. Scleractinia (Stony Corals) of Enewetak Atoll (pp. 67-76) in: U.S. Dept. of Energy: The Natural History of Enewetak Atoll (Vol I) The Ecosystem: Environments, Biotas, and Processes. DOE/EV/007-03-TI-Vol 1: 228 pp. Emery, K.O., 1948. Submarine geology of Bikini Atoll. Bull. geol. Soc. Am. 59:855- 860. Emery, K.O., J.I. Tracey and H.S. Ladd, 1954. Geology of Bikini and Nearby Atolls. U.S. Geol. Survey Prof. Pap. 260-A. Fosberg, F.R., 1988. A review of the Natural History of the Marshall Islands. Prepared for East-West Center/MacArthur Foundation Project, June 1988, unpubl. 115 pp. Fosberg, F.R., T. Arnow, and F.S. MacNeil, 1956. Military Geography of the Northern Marshalls. Prep. for the Chief of Engineers, U.S. Army by Intelligence Division, Office of the Engineers, Headquarters, U.S. Army Forces Far East, Eighth United States Army, and U.S. Geological Survey. Tokyo, 320 pp. Ladd, H.S., J.I. Tracey, J.W. Wells, and K.O. Emery, 1950. Organic growth and sedimentation on an atoll. J. Geol. 58(4):410-425. 87 MacNeil, F.S., 1969. Physical and Biological aspects of atolls in the Northern Marshalls. Proc. Symp. Corals and Coral Reefs. Mar. biol. Assoc. India, pp 507-567. Lamberts, A.E. and J.E. Maragos, 1989. Observations on the reefs and stony corals of Majro Atoll, Republic of the Marshall Islands. In: Majro Coastal Resource Inventory, prep by Univ. of Hawaii Sea Grant Extension Service and the U.S. Army Corps of Engineers Pacific Ocean Division, Environmental Resources Section, in press. Maragos, J.E., 1990. Observations on the corals and reefs of Bikini Atoll 30 years after termination of the U.S. nuclear weapons testing program, in preparation. Maragos, J.E. and A.E. Lamberts, 1989. Observations on the reefs and stony corals of Arno Atoll, Republic of the Marshall Islands. In: Arno Atoll Coastal Resource Inventory. prep by the Univ. of Hawaii Sea Grant Extension Service and the U.S. Army Corps of Engineers, Pacific Ocean Division, Environmental Resources Section, in press. Odum, E.P. and H.T. Odum, 1956. Zonation of corals on Japtan Reef, Eniwetok Atoll, Atoll Res. Bull 52:1-3. Rongelap Assessment Project, 1989. Report, Corrected Edition March 1, 1989: prep pursuant to the Compact of Free Association Act of 1985 and Administered by the Republic of the Marshall Islands. 1203 Shattuck Ave., Berkeley, CA 94709 (this is also the address for Bikini Atoll Rehab. Comm. reports). Thomas, P.E.J. et. al., 1989. Report of the Northern Marshall Islands Natural Diversity and Protected Areas Survey, 7-24 September 1988. South Pacific Regional Environment Programme, Noumea, New Caledonia, and East-West Center, Honolulu, Hawai, 133 pp. Tracey, J.E. Jr., H.S. Ladd, and J.E. Hoffmeister, 1948. Reefs of Bikini, Marshall Islands Bull. geol. Soc. Am. 59:861-878. Veron, J.E.N., 1986. Corals of Australia and the Indo-Pacific. Australian Institute of Marine Sciences, Townsville. Veron, J.E.N. and M. Pichon, 1976. Scleractinia of Eastern Australia, Part 1: Families Thamnasteriidae, Astrocoeniidae, Pocilloporidae, Austral. Inst. Mar. Sci. Monogr. 1:1-86. Veron, J.E.N., and M. Pichon, 1980. Scleractinia of Eastern Australia. Part II]: Families Agariciidae, Siderastreidae, Fungiidae, Oculinidae, Merulinidae, Mussidae, Pectiniidae, Caryophylliidae, Dendrophylliidae, Austral. Inst. Mar. Sci. Monogr. 4:1-422. 88 Veron, J.E.N., and M. Pichon, 1982. Scleractinia of Eastern Australia, Part IV: Family Poritidae, Austral. Inst. Mar. Sci. Monogr. 5:1-159. Veron, J.E.N., M. Pichon, and M. Wijsman-Best, 1977. Scleractinia of Eastern Australia. Part II: Families Faviidae, Trachyphyllidae. Austral. Inst. Mar. Sci. Monogr. 3:1-233. Veron, J.E.N., and C.C. Wallace, 1984. Scleractinia of Eastern Australia, Part V: Family Acroporidae, Austral. Inst. Mar. Sci. Monogr. 6:1-485. Wells, J.W., 1954. Recent Corals of the Marshall Islands, Bikini and Nearby Atolls, Part 2, Oceanography (Biologic). Geological Survey Professional Paper 260-I, U.S. Government Printing Office, Washington D.C., pp. 385- 486 + pls 94-187. Wells, J.W., 1951. The coral reefs of Arno Atoll, Marshall Islands. Atoll Res. Bull. 5:56 pp. Wells, J.W., Coral Reefs, In: Treatise on Marine Ecology and Paleoecology. J. Hedgepeth (ed.). Geological Soc. of Am. Memoir 67. Wiens, H.J., 1962. Atoll Environment and Ecology. Yale University Press, New Haven, Connecticut, 532 pp. U.S. Dept. of Defense, Office of Economic Adjustment, 1989. Republic of the Marshall Islands, Water Resources Plan of Action prepared by U.S. Army Corps of Engineers Pacific Ocean Division, Honolulu, 64 pp + 21 appendices. Appendix A. Fig. A-1l BOK - AK Atoll REEF PROFILES: LAGOON AND PASS REEFS +0.5m Spondylis Es hard abundant Tridacna spp SITE 1B 4m coral reef & Hippopus sand sand SITE 1D ee large Tridacna spp hard reef & Pa live coral populations 3m +0.5m sand hard coral 6m live coral sand SITE 1N 15m TOP VIEW OF BOK-AK PASS lagoon entrance SITE 1R1 sire ik gene (TOP VIEW) scoured hard reef lagoon Ligne ss interior from lagoon heads 1R2 Coral scoured bottom mid point NORTH wae 0 REEF om ) 1R3 reef rubble : . slump blocks interior from ocean ak ard hard reef rock reef rock Beean rubble trench ocean entrance 1R5 ars scour trenches of hard reef rock & some rubble CROSS SECTIONS OF BOK-AK PASS FIG A-2 REEF PROFILES: EASTERN PERIMETER REEFS BOK - AK Atoll celagoon +1m ocean NE SITE 1A sic sated Pa iat coral head sand pavement pavement encrusting SITE 1c ; corals Holothuria at rubble tract ztdge SITE eard live coral 1L overhang ocean ~» SITE 1M1 live coral rubble XN coraliine SITE 1M3 ocean & ocean» live sand sand coral overhang dead pane coral ocean (southeast) sand SE 3-4m Pinctada margarjti mounds sand ; ocean > Many Tridacna spp microatolls SITE 1F coral sand heads pabbie sand head SITE 1G SITE 1H reef wall live coral Halimeda Bana sand sand an FIG. A-3 BOK - AK Atoll SITE 1T ocean 30m 25m 30m SITE 19 REEF PROFILES: WESTERN PERIMETER REEFS Coralline algal dam spillway coralline alage live coral SITE 1K ribbon reef SITE 1J Coralline Canyon wall. ~ SITE 1S YE algal — canyon floor gamis Y spillway SITE 10 coralline algae +0.5m spillway sacs 90% live coral cover on slopes 1m SITE 1T eras 1P coralline algae 90% live coral on slopes A-4 FIG. A-4 PIKAAR ATOLL REEF PROFILES: PERIMETER REEFS NORTH SS 1 = SITE 2F1 im 2F2 2F3 coral coral overhangs 4m live reef pavement coral pavement WEST sand SITE 2H spillway sand chute shingle spillwa SITE 2L coral-algal dam ‘ shallaw coralline algae spurs & grooyes~ Halimeda live coral SITE 2M live cora ocean (SE) —= EAST 1m reef pavement SITE 2B corals hard reef 5m sand ocean (E)—> 1m SITE 2D 1570 micro-atolls reef pavement hard reef pavement sand 0.5 as SITE 2E pavement F F corals Dictyosphaeria 2.5m reef hard pavement FIG. A-5 PIKAAR ATOLL REEF PROFILES: LAGOON REEFS & PERIMETER REEF LAGOON 2C i STE coralline algae corals ; Qe Dictyospheria live finger coral SITE 2G live coral SITE 21 coralline 15m small Tridacna abundant ee sand SITE 23 dead cora & live coral <— lagoon SITE 2K talus sand slope 5m SOUTH (PERIMETER) SITE 2A sw ~~ coral ocean reef flat island corals coral sand chute FIG. A-6 2 REEF PROFILES: PERIMETER REEFS TOKE ATOLL a z Simoes EAST Dictyosphaeria (ocean) SITE 3A pavement and rubble coral Culcita _ 0.5m 7m sand reef pavement 2m SITE 3C ile taza rubble blue coral coral OrSne es 2 XX SITE 3D m . rubble mixed corals pavement staghorn coral < EAN aa ee SITE 3F SE sand cay mixed coral & reef pavement hard coral beach sand chute lm sand N—> (OCEAN) NORTH coralline algae & corals SITE 3G live live coral SITE 3H sand dune WEST live coral Dictyosphaeria SITE 31 coral sand FIG. A-7 - REEF PROFILES: WESTERN PERIMETER AND LAGOON REEFS TOKE ATOLL WEST im live coral staghorn 4m staghorn coral coral ITE 3 Ses sand and rubble eee table coral deep reef live coral pool SITE 3K sand 10h 3m coralline , algae & live SITE 3L Tavencoral coral sand sand Bel ~ sand LAGOON live coral SITE 3B Grane! im sand live coral SITE 3E (Porites) 20m 20m sand A-8 FIG. A-8 JEMO REEF PROFILES: OCEAN SLOPES 3 hard reef pavement & Sens. < NW scattered corals live corals a 50% cover wy 3m SITE 4C Zi scoured reef pavement 45m 7 shallow channels = W SITE 4A 10-25% live coral cover =o j[=e = isha =~ (SETE 4D sand bottom surge channels scoured reef pavement sharks live coral pavement 6 N 11m high live trough pavement coral cover SITE 4E SITE 4F =< SE coral terrace 4 pavement and live corals <«s my ee SITE 43 1 SITE 4H pavement canyon / “ 4 5m “ SITE 41 wall & some live coral overhang aon pavement 20m “ shallow SITE 4L-2 channels ELS Cs corals & hard reef 40m SITE 4L-1 FIG. A-9 REEF PROFILES: NORTH AND EASTERN PERIMETER WOTTO ATOLL REEFS NORTH Ta RA IR Eee ELAS oS ST Ocean (N) -> table and staghorn Caulerpa Tridacna gigas corals en Halimeda 50-60% coverage ee sand Bielad ate — Holothoria ere edulis Bal — — sand oPe —— ew 10m —a SITE 5B small Hippopus Se 3m 2m coral reef flat (N) Ocean <&— Holothuria Dictyosphaeria Stichopus Ba 10m cig SITE 5C live reef rock coral ba sand coral aS nd staghorn rubble ———Yeéérf fiat rock coralli & coral — N (ocean) corals Dictyosphaeria 2m s SITE 5E and rubble ey SITE 5D om sparse corals SITENSF 3m 2m hare reef pavement A-10 FIG. A-10 WOTTO ATOLL REEF PROFILES: EASTERN AND SOUTHERN PERIMETER REEFS EAST 2m lagoon (W) 8m eef rock SITE 5G ovexys and rubble turn XXX 10m Coral xe staghorn coral sand 3m ne 8m SITE 5H coral mound coral sand — lagoon (W) am coral und coral 2m In te mound over turned ——_— rubble back re slope eS SOUTH 3m 2m coral SITE 5J mound coral fine sand 5m coralline aw 10m SITE 5K ——— 4m Jp Se ee eee coral head corals, sand & rubble 15m A A-11 FIG. A-11 RONDIK ATOLL REEF PROFILES: SOUTHEAST PERIMETER REEFS SOUTH SITE 6A <<— Ocean (S) ck reef slope Ss hard coral x e staghorn pavement head & coral take coral sand pinnacle cpa ae See cp ~ (S) Ocean-= SITE 6B (plex an blue cora hard 6m __ coral heads = pavement fine sand Acropora palifera (S) Island —7 Re hie beach SReEUce coralline algae beach rock 4m undercut lip en coral heads rubble & tables sand EIST trough avement EAST g P SITE 6D ee rilled reef 2m ere pavement Am VA om sand SEO atoll : y sand corals eee coral heads sand Ocean (SE)=> ~~ coral head nie SITE 6E rilled SG lei Pavement Ga gravels 4m \————_" staghorn sand coral sand ripples table corals , . im canis SITE 6F Tridacna & Hippopus beach beack rock sand sand live corals coral sand A-12 FIG. A-12 RONDIK ATOLL REEF PROFILES: NORTHEAST PERIMETER AND LAGOON REEFS : SITE 6G1 lagoon (W) > blue cora sand & mt rubble SITE 6G2 bank sand aS sand channel 0.5m SITE 6H island sand channel island SITE 6L coralline scoured algae reef corals 6 oH coral table sand ee ee Sand ‘ mound corals ripples SITE 6I crest of reef pinnacle coralline algae on 90% live coral 15m SITE’ 6J cover 15m live coral mound sand slope back reef SITE 6K pavement 2m microatolls rubble corals live corals SITE 6K 6m coral rubble staghorn table coral mounds sand A-13 FIG. A-13 ADKUP ATOLL REEF PROFILES: SOUTHERN PERIMETER REEFS coralline algae SITE 7Al1 and corals head SITE 7A2 coralline algae and corals sand staghorn sand sts coral slope BD Legions coral heads rubble beach rock & mounds island Halimeda SITE 7B lagoon (N) —_——_____ i eastep ai == cin ne reef pavement SITE 7C microatolls table rubble coral <— lagoon (N) Te eo Tae coe en ee — 3m microatolls pavement SITE 7D blue corals steep sand Ocean (E) >» slope -_-—— 2m SITE 7E steep mostly dead corals and rubble table coral sand dead) slope (dea Ocean (E)—>— reef pavement coral heads SITE 7F A-14 FIG. A-14 ADKUP ATOLL REEF PROFILES: NORTHERN PERIMETER REEFS are oe a 0 | na ocean (NE) SITE 7G1 = coralline algal ocean (NE)> a staghorn pavement coral Halimeda ee coralline algae SITE 7G2 corals (40% cover rubble on slope) coral coverage finger coral 10m 20% dead coral sane py go ie 0.3 ae ee OES 15m sand talus microatolls SITE 7H 1.5m corals rubble coral coverage 15% reef flat corall —> ocean (E) F ~ algae Pass 4m f 7 ingens aS SITE 7J3 = sand lagoon hard pavement with scour, hard corals SITE 71 and soft corals talus side to side cross section in pass ocean end corals hard scoured pavement with some corals d igh high lagoon en hig corals corals SITE 7K ——— sand staghorn coralline coz coral algae & Porites lutea Hippopus SITE 7L dead coral coralline algae high 50% coverage of the live corals live coral sand talus sandwedge coral hora coral dead 1 Zupps avement aie reef lip # SITE 7M 10% coverace of live corals sand talus Appendix B. Definitions for relative abundance terms used in the field for corals SYMBOL D TERM dominant abundant common occasional rare DEFINITION (WITHIN A_ ZONE OR HABITAT TYPE ON THE REEF) the coral constitutes a majority in abundance or coverage (50% or more of total) the coral contributes substantial abundance or coverage, or 1S very numerous coral present as several or more individuals or as a few larger colonies uncommon, present only as a few individuals, or present as a single conspicuous individual reported only once as a single individual DEFINITION (FOR THE REEF SITE AS A WHOLE) the coral contributes substantial abundance or coverage (25% or more of total) or is conspicuous in all zones coral is conspicuous in most zones or is dominant within a single zone coral conspicuous in only one or a few zones or locally substantial in a single zone present more than once but not substantially within a single zone reported only once from the reef C-1 Appendix C. Map sources for selected northern Marshall atolls. AMS = Army Map Service Series W861 (all at scale 1:25,000, except Wotto which has a scale of 1:50,000). DMA = Defense Mapping Agency charts (various scales). ADKUP (ERIK UB) AMS - 8249 I SW, II NW, SW, III NE; IV NE, SE DMA - none BOK-AK (TAONGI) AMS - 8066 III NE, SE; IV SE DMA - 81626A (1:50, 190) JEMO AMS - 8152 III SW DMA - none PIKAAR (BIKAR) AMS - 8258 IV NW, SW DMA - 81626B, C (1:10,110; 1:50,200) RONDIK (RONGERIK) AMS - 7755 IV NE, NW, SE, SW DMA - 81557A (1:50,000) TOKE (Taka) AMS - 8155 III NE, NW, SE, SW DMA - 81616A (1:10,000; 1:50,310) WOTTO (WOTHO) AMS - 7352 II (1:50,000) DMA - 81030C (1:316,120) ALINGINAE AMS - 7455 Il NW, SE, SW; III SE DMA - 81557B (1:72,500) Appendix D. Index of aerial photographs consulted during the study from 1978 color (EG&G) and 1944-1945 black-and-white U.S. War Dept. (VD3) sources, the latter at the Bernice P. Bishop Museum Map and aerial photo collections, Honolulu. RONDIK (RONGERIK) ATOLL EG&G Roll 8, perf 2234: frames 59-60, 62, 70-79, 82-90, 93-95, 101-110 (35 negatives, scale 1:30,000; 16 Aug 1978) VDB - oblique aerial photos only TOKE (TAKA) ATOLL EG&G Roll 7, perf 2205: frames 22-23, 102-106, 110-123, perf 2185: frames 208-216 (38 negatives, scale 1:30,000; 11 Aug 1978) VDB - p. 40316-36 frames 1-14; VD3-AP47A figures 1-41; and VD3 - AP47B photomosaic (55 negatives, scale unknown; 5 March 1944) PIKAAR (BIKAR) ATOLL EG&G Roll 12, perf 2305: frames 9-94, 98-103, 121, 123, 125, 137-139 (16 negatives, scale 1:30,000; 8 Aug 1978) VD3 - oblique aerial photos only WOTTO (WOTHO) ATOLL EG&G Roll 10, perf 2259: frames 4, 6, 8, 10, 54, 56-59, 108-109, 116-118, 134-138, 153- 157, (24 negatives, scale 1:30,000; 18 Aug 1978) VD3 - AP41B frames 1-39 (39 negatives, unknown scale; 29 Feb 1944) JEMO EG&G Roll 1, perf 2096: frames 33, 35, 36 (3 negatives, scale 1:30,000; 29 July 1978), Roll 2, perf 2122: frame 150 (1 negative, scale 1:8,000; 5 Aug 1978) VD3 - oblique aerial photos only ADKUP (ERIKUB) ATOLL VD3 - AP16A frames 1-8 and VD3-AP16B frames 1-2 (10 negatives, unknown scale; 4 Feb 1944) BOK-AK (TAONGI, POKAK) ATOLL VD3-12, frames 1-10 (10 negatives, unknown scale; 28 Mar 1945) Sey He PCTS: ii ats cig ee Nae at a eoes erent 12 pe “t i ee ae AaeeS BA Ul a FN Pte ea ‘6 i aaiiiet : ae ie 2 aa aRR OT zi yy mE) Fae | ’ A A isa | se) ie > : ‘ : ¢ Fie - Tinne 2 we rr - ved SESE SET A y af ‘ cf i eel YVxten Ny ' A oe $E-G Snrieket es eae a) i soe hae pee te ni aD ert d ep ee : ciiry ARAL . asi ESE RS STAN, 7 f = i ; A Maik a4 D ‘a 7 e. : ee ; ae a : ey ; - aes ra f c ‘URC aK ny ty ATIZR “ij wy t - a a ve PoP "4 4 th ‘ a4 - sf ; ee i 4 “pi is ‘ if — = LIOTTA KOU PEA BET-RED Bil -OEt MO aU) VEae pe wt 80 Sm ClsS ag, EOE S UN Sai acl BL on Mok giaiyitia f 4 é: ¥ ¢ - ' { 4 : 2 as ity rs | i . 4 Ji ) e) ' bv ER oD 5 tf 4 ; % ; ° uf ‘ ; % aris OEE toe (eerie ade EF assinat (a0Us + ’ sir ‘g 8 ~ ae LIOTTA te a ar So F, AACS ay (Pay vesrenyity Virose OEY eT | ATOLL RESEARCH BULLETIN NO. 420 QUATERNARY OOLITES IN THE INDIAN OCEAN BY C.J.R. BRAITHWAITE ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 ~- 20cotra = - Sri Lanka ee Swaine Loe” : ¢ Seychelles # Chagos Cele Aldabra * | Madagascar WESTERN INDIAN OCEAN > Mauritius. 2? ‘Rodrigues Reunion Fig. 1 General map of the Indian Ocean indicating positions of the principal locations referred to in the text. QUATERNARY OOLITES IN THE INDIAN OCEAN BY C. J. R. BRAITHWAITE ABSTRACT Aeolian calcarenites from Rodrigues in the Western Indian Ocean include oolites. No recent oolites are recorded in the Indian Ocean and there is only one other Pleistocene record. The calcarenites are compared with deposits in the Seychelles, East Africa, India, and South Africa which consist of marine-derived bioclasts. In general they indicate critical water depths over generating shelves and thus particular stages in sea-level change. They are characteristic of both early phases of cooling and later stages of warming events. There seems to be no explanation for the restricted distribution of oolites in the Indian Ocean. INTRODUCTION Aeolianites are widespread in the Pleistocene deposits of the Western Indian Ocean. Those on Rodrigues are unusual in that those on the east coast consist of oolites with a relatively small bioclastic component. With the exception of deposits in Kathianar in India, described by Chapman (1900) and Evans (1900), these are the only oolitic deposits known from the Pleistocene or Holocene of the Western Indian Ocean. FIELD DESCRIPTION Rodrigues lies about 650 km east of Mauritius, Lat. 19° 42' S, Long. 63° 25'E (Fig. 1). It is approximately 18.3 km long and 6.3 km wide, consisting predominantly of young undersaturated basalts erupted between 1.5 and 1.5 million years ago (McDougall et al. 1965). The south-eastern and south-western margins of the island are blanketed by a discontinuous cover of cross-bedded calcarenites (Fig. 2). These were known a hundred years ago when they were described by Balfour (1879) and Slater (1879) who visited the island during the Transit of Venus expedition to make observations on the general geology. Their reports describe caves in these limestones in which the bones of the Solitaire and other flightless birds were found. Although Rodrigues has a surface area of only about 120 km? the platform on which it rests is at least 1650 km2. The island surface slopes gently outwards to about 100 m depth before plunging into deeper water. The island is bordered on its south-western coast by a fringing reef 4-8 km from the shore and Admiralty Charts show that the lagoon area within this is generally very shallow (McDougall et al. 1965). However, the shallow platform beneath extends a considerable distance both to the west and the east and, as will be shown, probably played a significant part in the formation of the calcarenites. Most of the calcarenites are bioclastic and were originally marine, but those in Department of Geology and Applied Geology, The University, Glasgow G12 8QQ, Scotland, U.K Manuscript received 23 April 1993; revised 28 June 1994 2 south-eastern localities (Pointe Coton, Trou d'Argent) are clean-washed oolites. In both south-eastern and south-western areas grains are well rounded and well sorted within individual laminae. Cliff sections show well-defined high-angle cross bedding in sets of up to 10 m, forming cosets locally of more than 20 m thickness (Figs. 3 and 4). Cross- laminae dip at angles up to 30-35° and set-bounding surfaces generally dip a few degrees in a seawards, (generally southerly) direction. In some areas the upper margins of dunes are full of casts of branching root systems which imply colonization by substantial trees. There are, however, no distinctive terra rossa or palaeosol deposits and plants must have grown in unconsolidated sand before and perhaps during cementation. — The thickness of the calcarenites is quite variable. In western localities they blanket a gently undulating ramp, while in the east they occupy an erosional bench cut into the underlying basalts. Significantly, in both areas basal surfaces appear to extend below present sea-level. McDougall et al. (1965) recorded these limestones as occurring at more than 60 m above present sea-level, while Snell and Tams (1920) claimed that they extend up to 500 ft (165 m). Present observations suggest that the lower figure is more realistic. The thickest sequences are found in the south-west of the island in the area of Caverne Patate where Cave systems penetrate at least 30 m of limestone. Montaggioni (1973) suggested that typical thicknesses are 15-20 m). McDougall et al. (1965) thought that the depositional morphology of the calcarenites had been obscured by recent erosion. In fact, low dune ridges can be seen in south-western areas and air photographs reveal a spectacular series of parallel dune crests facing north and north-east and having a wave length of from tens to hundreds of metres. These extend for several kilometres from the present shoreline (Fig. 5). Groups of steep foreset laminae are visible on air photographs and can be traced across outcrops. Interdune areas are less conspicuous and are characterized in limited outcrops by low angle lamination and shallow troughs. PETROGRAPHY The oolites have a mean diameter of about 300m (medium sand), many nuclei are dark micritic peloids while others are bioclasts. Cortical layers have the tangential structure typical of recent aragonitic ooids and may be more than 100um thick (Fig. 6). Bioclast calcarenites consist predominantly of foraminifera, including Marginopora and Heterostegina, together with echinoderm plates, calcareous algal fragments (cf. Goniolithon and Halimeda), mollusc shell and coral fragments. In the samples examined grains have a mean size of 400-800uUm. The marked contrast in grain typebetween oolitic and bioclastic rocks are paralleled by differences in mineralogy and diagenetic history. The bioclastic limestones include both aragonite and calcite (identified by X-ray diffraction) but have only a sparse fine-grained blocky calcite cement. This is restricted to patches where it forms meniscus bridges between grains and occasional pendant drops, suggesting deposition in the vadose zone, although it lacks the fibrous or prismatic textures typical of such environments. By contrast, in oolites, the ooliths themselves consist of aragonite and they are commonly embedded in a coarse blocky calcite cement which may completely fill pores and locally extends inwards from grain surfaces as a neomorphic replacement of the original tangential structure. INTERPRETATION The distinctive high-angle lamination which characterizes the calcarenite deposits on Rodrigues, and the occurrence of calcarenites over such a range of altitude, seems to 3 confirm an aeolian origin. Submarine sand-waves might be of similar size but would generally be expected to have lower angle cross-bedding and are unlikely to have accumulated over such a slope. They would, moreover, have had to be related to a sea- level more than 60 m higher than at present. Deposition of these rocks must have occurred at a time when sea-level stood lower than at present, perhaps by about 10 m. However, the sediments were of marine origin and reflect generation in shelf environments. Thus, sea-level must have been high enough to maintain a permanent water cover over a sufficiently large area. The large area of gently-sloping surface beneath present reefs would have ensured that a suitable shelf was present over a range of sealevels, although not during the glacial maximum. Grains generated within the shelf were swept onshore by storm waves and accumulated as beach and, ultimately, dune deposits, the latter probably migrating several kilometres from the shoreline. Correlation with a lower sea-level implies a general correlation with a cooler climate. THE AGE OF THE DEPOSITS Two different calcarenite types are present on Rodrigues and since lithological variations are paralleled by diagenetic differences the rocks are likely to be of at least two different ages. Montaggioni (1970) referred to three separate dune assemblies in the eastern area but these have not been identified here. The calcarenites have not been precisely dated but it is clear that they are not Recent. The active phase of accretion was followed by a passive phase of colonization by trees, and the whole assemblage has been dissected by a mature karst system which, judging from the included fauna, is at least late Pleistocene (see again Slater, 1879). However, the ages inferred for these deposits have been determined by reference to associated limestones. Montaggioni (1970) regarded them as of the same age as calcarenites on Mauritius. In this interpretation they should equate with high sealevel stand dates of 230Th/234U ages of 120 15-20 ka obtained by Battistini (1976) for Pleistocene limestones on Mauritius, which were compared in turn with dates of 160 £40 ka and 110 +40 ka given by Veeh (1966) and of 114 £6-7 ka and 104 +4-6 ka given by Elbez (1976). However, none of these ages was obtained from calcarenites and, for reasons of environment, it is unlikely that the aeolian calcarenites on Rodrigues were contemporary with the marine deposits from which the dates were obtained. There are no observations of rocks overlying the Rodrigues calcarenites and this contrasts with the relative positions of aeolianites in Kenya (Braithwaite, 1984) and on Aldabra (Braithwaite et al., 1973) which are probably substantially older (see below). OTHER INDO-PACIFIC DEPOSITS Within the western Indian Ocean aeolianites (calcarenites with characteristics similar to those described) have been identified in a number of areas. In northern Madagascar Battistini (1976) recorded dune-bedded carbonate rocks low in the Pleistocene. These overlie his Recif I but appear to be older than Recif II, dated at 160 £10-15 ka. This dune assemblage is said to be more than 200 m thick but only about 15 m are shown on the published synthetic section extending for 25 km in the south. It is considerably thinner on the north and east coasts. Like the deposits on Rodrigues the top is marked by non- deposition, in this case resulting in a dissected erosion surface and what are described as "decalcified pockets”. In East Africa, calcarenites with high angle cross-bedding and consisting of marine- derived grains are exposed on the northern coast of Kenya, south of Malindi, and in the extreme south, near Shimoni and Wasini Island (Braithwaite, 1984). In both of these areas the deposits are believed to extend below present sea-level but are stratigraphically above 4 coral-bearing marine limestones which were apparently the source of material dated by Battistini (1976) as 240 40-70 ka. They are overlain by younger limestones which have been described by Crame (1980, 1981) and which appear similar to the youngest coral- bearing limestones on Aldabra. Aldabra has a small area of calcarenites with high angle cross-bedding exposed on the south coast (Braithwaite et al., 1973). These are overlain by two marine limestones, the youngest of which has been dated by Thomson and Walton (1972) as forming between 118-136 £9 ka. It is this which is equivalent to the younger coral limestone in Kenya. These calcarenites, in which grains are entirely of marine bioclasts, formed when Aldabra was a shallow marine platform with only small sand cays accumulating along what is now the southern coast and in a relatively large area to the north-east. The aeolian calcarenites of southern India have already been referred to (Chapman, 1900, and Evans, 1900). Knox (1977) described other calcareous aeolian dune deposits of supposed middle to late Pleistocene age from Saldanha Bay in South Africa. These average 30 m in thickness but are reported by Visser and Schoch (1973) to be as much as 88 m locally. Once more they consist predominantly of marine bioclasts. DISCUSSION The descriptions given reinforce the view that aeolian deposits are a common, even characteristic, feature of Pleistocene limestone successions in the Indian Ocean. They are of a variety of ages but have commonly (although not exclusively) formed in areas which do not have present day dunes. It might be argued that winds were stronger in the past and that the absence of present-day dunes is a reflection of the inadequacy of present day winds. However, this is unlikely, Bagnold (1941) and Wolman and Miller (1960) have indicated that in dune formation the net transport of sands is controlled more by prevailing winds above threshold velocity than by extreme wind events. In almost all of the Indian Ocean examples the sands were of medium to fine grade and therefore do not reflect excessive wind speeds. Elsewhere, in Bermuda (Mackenzie 1964), aeolianites show palaeowind vectors which are essentially the same as present directions while in the Turks and Caicos islands (Lloyd et al. 1987) palaeocurrents in both Pleistocene and recent aeolianites (which include oolites) can be related to south-easterly Trade Winds. The presence of oolites remains enigmatic. Their formation, and that of the bioclasts, was probably critically dependent on water depths over the generating shelves. If sealevel is too high shelves are deeply flooded and unlikely to reach the necessary saturation conditions. If it is too low shelves are restricted (or dry) and again fail to generate sediment. Thus it seems that the oolites and the aeolianites generally are characteristic of early stages of cooling and later stages of warming events when sealevel was only marginally below its present position. Why they were not more widespread in the Indian Ocean, where there are several extensive shallow banks, is not known. Marine-deposited oolites may be present but if they are it is likely that all lie beneath present-day shelves. ACKNOWLEDGEMENTS I wish to thank Prof. Gulam T. G. Mohammedbhai and Dr. André Chan Chim Yuk of the University of Mauritius for guidance during visits to the Rodrigues deposits as part of a programme financed by British Council. Work in Kenya was supported by the 5) Carnegie Trust for the Universities of Scotland and that on Aldabra by the Royal Society, and the Natural Environment Research Council, with logistical support and facilities provided by the Seychelle Islands Foundation. REFERENCES Bagnold, R. A. 1941. The Physics of blown sand and desert dunes. Methuen, London, 265pp. Balfour, I. B. 1897. The Physical features of Rodrigues. An account of the petrological, botanical and zoological collections made in Kerguelen's Land and Rodrigues during the Transit of Venus Expeditions 1874-5. Philosophical Transactions of the Royal Society of London 168:289-292. Battistini, R. 1976. Application des methodes Th239-Ur234 a la datation des depots marins anciens de Madagascar et des iles voisines. Association Sénégal Etudes Quaternaire Africain, Bulletin Liason Sénégal 49:79-95. Braithwaite, C. J. R. 1984. Depositional history of the Late Pleistocene limestones of the Kenya Coast. Journal of the Geological Society of London 141:685-699. Braithwaite, C. J. R., Taylor, J. D. and Kennedy, W. J. 1973. The evolution of an Atoll: the depositional and erosional history of Aldabra. Philosophical Transactions of the Royal Society of London, Series B 266:307-340. Chapman, F. 1900. Notes on the consolidated sands of Kathiawar. Quarterly Journal of the Geological Society of London 56:584-588. Crame, A. 1980. Succession and diversity in the Pleistocene coral reefs of the Kenya Coast. Palaeontology 23:1-37. Crame, A. 1981. Ecological stratification in the Pleistocene coral reefs of the Kenya Coast. Palaeontology 24:609-646. Elbez, G 1976. Application de la methode de datation 239Th/234U a la determination des niveaux marins de Madagascar et des isles environmentes. Memoires Maitre (Geophysique) soutenu a l'Universite de Paris-Sud, Centre Orsay. Evans, J. W. 1900. Mechanically formed limestones from Junagarh (Kathiawar) and other localities. Quarterly Journal of the Geological Society of London 56:559-583 and 588-589. Knox, G. J. 1977. Caliche profile formation, Saldanha Bay (South Africa). Sedimentology 24:657-674. Lloyd, R. M., Perkins, R. D. and Kerr, S. D. 1987. Beach and shoreface ooid deposition on shallow interior banks, Turks and Caicos islands, British West Indies. Journal of Sedimentary Petrology 57:976-982. Mackenzie, F. T. 1964. Bermuda Pleistocene eolianites and palaeowinds. Sedimentology 3352-64. 6 McDougall, I., Upton, B. G. J. and Wadsworth, W. J. 1965. A geological reconnaissance of Rodrigues Island, Indian Ocean. Nature 206:26-27. Montaggioni, L. 1970. Essai de reconstruction palaeogeographique de I'ile Rodrigue (Archipel des Mascareignes, Ocean Indien). Compte Rendu de la Academie Sciences de France, Paris 271:1741-1744. Slater, H. S. 1879. Observations on the bone caves of Rodrigues. An account of the petrological, botanical and zoological collections made in Kerguelen's Land and Rodrigues during the Transit of Venus Expedition 1874-5. Philosophical Transactions of the Royal Society of London 168:420-422. Snell, H. J. and Tams, W. H. T. 1920. The Natural History of the island of Rodrigues. Proceedings of the Cambridge Philosophical Society 1916-1919 19:420-422. Thomson, J. and Walton, A. 1972. Redetermination of Chronology of Aldabra by 230Th/243U dating. Nature 240:145-146. Veeh, H. H. 1966 Th239/U238 and U234/U238 ages of Pleistocene high sea-level stand. Journal of Geophysical Research 71:3379-3386. Visser, N. H. and Schoch, A. E. 1973. The geology and mineral resources of the Saldanha Bay area. Republic of South Africa, Department of Mines, Geological Survey Mememoir 63:150pp. Wolman, M. G. and Miller, J. P. 1960. Magnitude and frequency of forces in geomorphic processes. Journal of Geology 68:54-73. 63°30’ E PORT aT ORIN PAS SA COTON \TROU d’ ¥ ARGENT S GRAVIERS DreSsT -© we Tees cs a1 St 0, WM ce POINTE CRABE CORAIL ©w” F> BASALT PRESENT REEF [| MARGIN CALCARENITES 01 3 #5Km Fig. 2 General map of Rodrigues showing distribution of main areas of calcarenites. Based on air photographs and Montaggioni (1973). Fig. 3. High angle dips in cross-laminae of calcarenites. Sante Francois, Rodrigues. Metre rule gives scale. Fig. 4 Two sets of high angle cross-laminae in calcarenites. Trou d'Argent, Rodrigues. Sets about 10 m high. Note on the right hand side the abutment against the sloping surface of the underlying volcanics. ‘sounp [Issoj Jofeul Jo UONRIO] Oy) pue ‘svUTWUR] SUIPpaq-ssoId ATTedt1ouud are YoTYM soul] [emnjonys ‘s}yeseq SutApJopun ou} uM Arepunog say} ‘sdo19jno ayTUSIeITeD JO 3Ud1X9 dy) ZUIMOYS SONSLIPOY JO 1SBOd UII}Sed-YINOS dy] JO UONLIsIdI|\UI YdeisojOYd IW ¢ “SIY \ ) Dw Nh 1 \ tN “A _ SS SS $01J9W OOOL enbijebuaeg [J21GD OYUIOd ie Wh aqiou,| ayulod 033NG 138d Je@A epuesy esuy Fadl \ Fig. 6 Thin section showing tangential structure of aragonitic ooliths. Note that ooliths are about 300u.m in diameter. ATOLL RESEARCH BULLETIN NO. 421 LARGE-SCALE, LONG-TERM MONITORING OF CARIBBEAN CORAL REEFS: SIMPLE, QUICK, INEXPENSIVE TECHNIQUES BY RICHARD B. ARONSON, PETER J. EDMUNDS, WILLIAM F. PRECHT, DIONE W. SWANSON, AND DON R. LEVITAN ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 D . s =: > eae ree A we j ih iP i) te a ai : j : : . i Sy SES ASTON TS Alar thc ad aka na Aa Hamed " 7 7 ms us i" 2 : e = & om tie | é€ , rb ; } is wey, rae na LARGE-SCALE, LONG-TERM MONITORING OF CARIBBEAN CORAL REEFS: SIMPLE, QUICK, INEXPENSIVE TECHNIQUES BY RICHARD B. ARONSON!, PETER J. EDMUNDS’, WILLIAM F. PRECHT?, DIONE W. SWANSON? AND DON R. LEVITAN* ABSTRACT With coral cover and diversity declining on many coral reefs, a clearer understanding of large-scale reef dynamics is imperative. This paper presents a sampling program designed to quantify the sessile biotas of Caribbean reefs on large spatiotemporal scales. For each reef sampled, data are gathered along replicate, 25-m transects located within the habitat of interest. Herbivore impact is estimated by fish and echinoid censuses along the transects. High-resolution videotapes are used to estimate the percent cover of corals, algae, and other substratum occupants, and to estimate coral diversity. Finally, topographic complexity is measured along the transects. In at least some reef habitats, this index of three-dimensional structure provides a measure of the total disturbance regime, with flatter areas having been subjected to more intense, more frequent, and/or more recent sources of coral mortality. The techniques and statistical analyses described in this paper are simple, quick and inexpensive. Repeated sampling on multiple reefs will enable the investigator to detect changes in community structure and to test hypotheses of the causes of those changes. INTRODUCTION Coral reefs are complex, diverse, productive tropical ecosystems in which multiple physical and biological processes covary in space and time (Huston 1985). Discerning the contributions of those processes to the community structure of reefs has been and will continue to be extremely difficult. The question of the appropriate scales at which to search for pattern and process is fundamental to unraveling these multiple causal connections (Jackson 1991, 1992; Karlson and Hurd 1993). Are ecological parameters such as coral cover and diversity determined primarily by small-scale processes, such as the local level of herbivore activity (Sammarco 1980; Lewis 1986), or are larger-scale, regional disturbances more important? Such questions are becoming increasingly ‘Marine Environmental Sciences Consortium, Dauphin Island Sea Lab, Dauphin Island, AL 36528; and Department of Invertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560. *Department of Biology, California State University, Northridge, CA 91330. ’Reef Resources & Associates, 7310 Poinciana Court, Miami Lakes, FL 33014. “Department of Biological Sciences, Florida State University, Tallahassee, FL 32306. Manuscript received 14 October 1993; revised 22 July 1994 2 germane as we confront the possibility that human interference is altering coral reef ecosystems (e.g., Brown 1987). Over the past two decades, the "health" of many coral reefs worldwide has deteriorated, as measured by indicators such as the diversity and abundance of reef- building corals (Rogers 1985; Dustan and Halas 1987; Hatcher et al. 1989; Grigg and Dollar 1990; Porter and Meier 1992; Ginsburg 1994). Hurricanes (Woodley et al. 1981; Rogers et al. 1982, 1991; Edmunds and Witman 1991; Hubbard et al. 1991; Bythell et al. 1993), coral bleaching (Oliver 1985; Brown and Suharsono 1990; Williams and Bunkley-Williams 1990; Glynn 1993), coral diseases (Gladfelter 1982; Riitzler et al. 1983; Edmunds 1991), mortality of the sea urchin Diadema antillarum (Lessios et al. 1984; Levitan 1988; Lessios 1988; Carpenter 1990) and outbreaks of the seastar Acanthaster planci (Moran 1986; Endean and Cameron 1991) may all be involved. At present, it is unknown whether any or all of these problems are due to recent human activities, or whether they are part of natural, long-term trends or cycles (Brown 1987; Richards and Bohnsack 1990). In addition, reefs are directly affected by human activities, particularly fishing, sedimentation, eutrophication and pollution (Brown 1987, 1988; Hatcher et al. 1989; Rogers 1990; Richmond 1993; Sebens 1994). These stresses reduce coral survivorship and growth and may promote macroalgal growth. Furthermore, marine pollution may increase the susceptibility of corals to disease (Brown 1988; Peters 1993). Reef dynamics are governed by multiple causes operating on multiple scales (Grigg and Dollar 1990). For example, the recent decline in coral cover and increase in macroalgal cover on the well-studied reef at Discovery Bay, Jamaica appears to have resulted from a variety of interacting processes: Hurricane Allen (1980), which opened substratum for colonization; feeding by corallivorous snails, which increased the mortality of hurricane- fragmented corals; the Diadema dieoff, which reduced herbivory drastically; and a history of overfishing by humans, which had previously removed herbivorous reef fishes (Hughes et al. 1987; Knowlton et al. 1990; Hughes 1994). Likewise, the catastrophic effects of the 1982-83 El Nino event on eastern Pacific coral reefs were due to the interaction of physical changes and a variety of biological processes (Glynn 1990). One of the greatest challenges to ecology is determining the relative importance of numerous causes to ecosystem structure and function (Quinn and Dunham 1983). Yet, even choosing the correct range of scales on which to understand coral reef dynamics seems a forbidding task. It is becoming increasingly clear that quadrat-scale observations (square meters or smaller) have less explanatory power than larger-scale observations (Jackson 1991). Those larger scales range from the landscape within a reef (hundreds of m’ to hectares), to the area encompassing multiple reefs within a locality (hundreds of km’), to an entire region such as the Caribbean. The need for large-scale and long-term (>5 yr) monitoring programs for coral reefs 3 has been emphasized over the past few years (Rogers 1988, 1990; Ogden and Wicklund 1988; D’Elia et al. 1991; Jackson 1991; Ray and Grassle 1991; Smith and Buddemeier 1992; Bythell et al. 1993; Glynn 1993; Hughes 1994), although less attention has been given to the hypotheses that should be driving the research (Hughes 1992). Few such studies have been undertaken, in part because hypotheses are so difficult to formulate and test (Hughes 1992). Such hypotheses might, for example, include an inverse relationship between the degree of disturbance and coral cover; the intermediate disturbance hypothesis (Connell 1978; Rogers 1993), which postulates that coral diversity should be highest at intermediate frequencies and intensities of disturbance; and complex relationships between herbivory, nutrients, coral and algal cover, and diversity (Littler et al. 1991; Knowlton 1992; Hughes 1994). However, with so many possible causes of community change, and with the ecological implications of each putative cause unknown, highly variable and/or controversial, it is hard to know which variables to monitor in order to test hypotheses adequately. This paper outlines procedures for comparing coral reef community structure and disturbance regimes across space and through time, with a view toward eventually testing hypotheses such as those listed above. The methods were developed during continuing studies of four western Atlantic coral reefs: Carrie Bow Cay, on the Belizean Barrier Reef; Discovery Bay, Jamaica; and Conch and Carysfort Reefs, off Key Largo, Florida. Since funding for monitoring programs is difficult to obtain and since field time is usually limited, our goal was to create an accurate, relatively inexpensive, “rapid assessment" sampling program for Caribbean coral reefs, with sufficient statistical power to detect biologically meaningful differences. Video transects are used to describe the sessile biota, and fish and echinoid censuses provide estimates of the intensity of herbivory. Since the importance of each type of reef disturbance has not been conclusively demonstrated, we advocate combining them in a single disturbance index, topographic complexity. Topographic complexity reflects the total disturbance regime in at least some reef habitats, integrated on a time scale of years. METHODS 1. Sampling Design Any study that compares reefs must be standardized with respect to habitat. The design described here is for a common Caribbean fore-reef habitat: the spur-and-groove habitat of windward-facing reefs at 12-15 m depth. Ten 25-m long transects are sampled at each site. The transects are placed along the central axes of replicate, haphazardly chosen spurs. Sandy areas and the edges of the spurs are avoided. In the present study, each transect was placed on a different spur, except at Discovery Bay. A number of spurs at Discovery Bay were >25 m wide; 2 transects, spaced >10 m apart, were surveyed on some of those wide spurs. It is important to choose sites that will accommodate at least 25-30 transects, so that the transects sampled during different visits to a site will not be identical. 4 The use of band transects is dictated in this particular application by the elongate shape of the spurs, and in general by the desire to encompass as great an area as possible in each sample. The 25-m transect length was chosen to be as great as possible while still restricted to a narrow depth and habitat range on the spurs. A 25-m transect, when extended over a 3-m depth range, requires spurs that slope at angles of no more than 7°. For each transect, scuba divers unreel a 25-m fiberglass surveyor’s tape, laying it taught along the center of the spur. The tape is left undisturbed for 3-5 min, at which point the divers commence surveys of the mobile fauna. 2. Fish and Echinoid Censuses Coral reef researchers have yet to agree on a single, reliable method for quantifying the activity of herbivores (Steneck 1983). The simplest measure is the abundance of herbivores, which correlates with their impact among habitats within reefs (Hay and Goertemiller 1983; Lewis and Wainwright 1985). Therefore, before the fish are disturbed further, a visual census is conducted along each transect. A diver swims along the tape at a standard slow speed, recording the number of parrotfish (Labridae, formerly Scaridae) and surgeonfish (Acanthuridae) within a visually-estimated 2 m on either side. The fishes are classified as small (<10 cm Standard Length), medium (10-25 cm), or large (>25 cm). The small transect width minimizes the underestimate of true fish population density inherent in the transect method (Sale and Sharp 1983). Divers then carefully explore the 100 m? area, recording the number and species identities of damselfish (Pomacentridae), as well as censusing Diadema and other regular echinoid species. The echinoids can be extracted from their shelters and measured (test diameter) with calipers. Size distributions of echinoid species can then be used to estimate their impacts on algal assemblages (Levitan 1988). It should be noted that daytime censuses underestimate echinoid densities; more accurate estimates can be obtained by censusing at night (e.g., Carpenter 1981, 1986). An alternative approach is to examine the process of herbivory by counting the number of bites that parrotfish and surgeonfish take per unit time from small areas of algal turf (Steneck 1983). This technique is more time-consuming than counting fish. A more serious concern is that some habitats on some reefs are currently so overgrown with fleshy macroalgae that finding even a square meter of algal turf would be problematic. For example, at Discovery Bay in 1992 macroalgal cover was >90 %, coral cover was <3 %, and the cover of algal turfs, crustose coralline algae, and bare space combined was <6 % at 15 m depth (as assessed by the video technique described below; Table 1, p. 11). Large differences in the availability of algal turfs could lead to differences in fish foraging behavior, independent of fish abundance. It is important to recognize that both the abundance and bite frequency methods yield short-term "snapshots" of herbivory, which may not adequately reflect longer-term variability. 3. Percent Cover and Scleractinian Diversity Photography provides the only practical means of sampling large areas underwater, 5 given the time constraints of scuba diving. Littler (1980) discussed the advantages of photography over recording data in situ. High-resolution video technology makes the approach all the more attractive because videotaping is easier and less time-consuming than still photography. While still photographs provide better resolution than videotapes, the resolution of videotapes is adequate for work of the type described here. Videotapes enable the investigator to cover a far greater area per unit sampling effort. In addition, video exposures are automatic, continuous and do not require developing. In this method, a diver swims slowly along the transect, videotaping a 40-cm wide swath of reef from a height of approximately 40 cm, using a high-resolution (Hi-8) video camera in an underwater housing, fitted with a wide-angle lens. A 15-cm gray plastic bar is attached to a rod that projects forward from the video housing. The bar, which is held at the level of the substratum during taping, provides scale in the videotaped images and also ensures that the camera is held a constant distance from the bottom. Individual video frames are displayed on a high-resolution monitor in the laboratory. A clear plastic sheet with 10 random dots is laid over the monitor screen, and the sessile organisms underlying the dots are recorded (Sebens and Johnson 1991; see Sample Sizes, p. 8, for number of dots per frame). The videotape is then advanced to a new, non- overlapping position. Each 25-m transect yields 50 video "quadrats", for a total of 500 points per transect. The point count data are used to calculate percent cover and the Shannon-Wiener diversity index, H’, for each transect. Since this and similar sampling methods tend to be biased against the inclusion of rare species, presence-absence data are also gathered for coral species by viewing the entire videotape of each transect (Chiappone and Sullivan 1991). By holding the camera perpendicular to the substratum, swimming slowly along the transect, and using a pair of video lights (50 or 100 watts each), it is possible to produce clear stop-action images. Corals, sponges, and some gorgonians and macroalgae can be identified to species, down to a diameter of approximately 5 cm. A drawback of this method is the difficulty of distinguishing fine algal turfs, crustose coralline algae and bare space from the tapes; these are lumped into a single category, which can be resolved by closeup, still photography if desired. Ecologists have devoted a great deal of effort to developing and comparing methods for quantifying coral reef community structure, with variable results (e.g., Loya 1978; Dodge et al. 1982; Ohlhorst et al. 1988; Chiappone and Sullivan 1991; Porter and Meier 1992). Porter and Meier (1992) examined some of the biases introduced to surface area estimates by photographic and video techniques. Such errors include non-orthographic projection, in which coral heads that stick up above the surrounding substratum are closer to the camera lens and therefore artificially enlarged, and parallax. These problems are difficult to correct (Porter and Meier 1992). Whorff and Griffing (1992) found that point counts from video frames overestimated 6 the percent cover of intertidal barnacles and bivalves, compared to computer image- processing of the video frames. More dots per frame yielded better percent cover estimates. On the other hand, Foster et al. (1991) concluded that point counts from photographs underestimated cover in multilayered, temperate subtidal assemblages, compared to point counts done in the field. The point count method is not as accurate as planimetry of the individual colonies in each frame or fully-automated image processing. However, planimetry is so time-consuming as to be impractical. Image processing is also difficult at present because subtle color and pattern differences must be detected; most corals and algae are quite similar in color, and reasonably-priced, hand-held video lights provide limited color saturation. The point count method is capable of detecting significant among-site differences in percent coral cover and diversity (see Sample Sizes, p. 8). Video has its problems and biases like other techniques, but it remains a simple, cost-effective comparative method. Another concern is the seasonality of macroalgal growth (Carpenter 1981; Hughes et al. 1987). Seasonal changes within a site could change estimates of coral cover, as more or less living coral is obscured by the algae (J. C. Lang, personal communication). The constraints of time, logistics and funding do not always permit the investigator to standardize sampling by season, and the error in coral cover estimates caused by seasonal variations in algal growth are unknown. In the present study, Carrie Bow Cay was sampled in the late spring, Discovery Bay in the winter, and Conch and Carysfort Reefs in the fall of 1992. If algal growth is maximal in the summer and algal destruction by storms is maximal in the winter, then, all other things being equal, macroalgal cover, and the error in coral cover estimates due to macroalgal cover, should have been greatest in Florida (after the summer’s algal growth), intermediate at Carrie Bow Cay (before the summer’s growth), and least at Discovery Bay (during the stormy season). In fact, macroalgal cover was highest at Discovery Bay, intermediate at Conch Reef and Carrie Bow Cay, and lowest at Carysfort Reef (Table 1, p. 11). In this study, differences in macroalgal cover among sites apparently outweighed any error in coral cover estimates associated with seasonal differences. The species diversity of scleractinian corals is evaluated as species richness, S, and as the Shannon-Wiener index, H’. S is measured for each site as the asymptote of the rarefaction curve (cumulative species numbers plotted against number of transects videotaped). Species richness is not calculated for each transect since reefs with lower coral cover are expected to have lower S per transect simply because fewer colonies are sampled (Magurran 1988). The Shannon-Wiener index is calculated for each transect as H’=-Y(p,[Inp,J), where p; is the proportion of the ith species in the sample. Vast size differences among coral colonies and colony fragmentation and fusion obscure the meaning of H’ calculated from numbers of "individuals". Unless the investigator is prepared to establish the genetic identity of all ramets, we recommend H’ indices based on areal coverage for a general characterization of reef communities. 4. An Integrated Measure of Disturbance Historical records of disturbance do not exist for most coral reefs. Even where such records are available, there is no obvious way to sum the different disturbances to reflect the total disturbance regime; one cannot simply score three disturbance points for a hurricane and two for a bleaching event. We present topographic complexity as a technique for measuring disturbance, along with its rationale and sources of error, so that the individual investigator can decide whether or not it will be useful. Topographic complexity is measured by carefully conforming a 5-m length of fine brass chain to the substratum adjacent to the central part of each transect tape. The chain is conformed to the finest topographic features that the 17-mm links permit; it is carefully inserted into small cavities and into the spaces within thickets of foliose and branching corals. The 5-m chain length was chosen so that the procedure could be completed in a reasonable length of time (10-15 min); the chain must be carefully straightened before it is conformed to the substratum. A complexity index, C, is calculated as C=1-d/l, where d is the horizontal distance covered by the conformed chain (measured against the transect tape) and / is its length when fully extended (e.g., Risk 1972; Rogers et al. 1982; Aronson and Harms 1985; Hubbard et al. 1990; Connell and Jones 1991). Disturbances that lead to the partial or complete mortality of coral colonies decrease this topographic complexity. Hurricanes decrease complexity directly by toppling branching and head corals (e.g., Rogers et al. 1982; Kaufman 1983), although this is not true in all reef habitats (Rogers et al. 1991; see below). In addition, once dead coral skeletons are exposed by a disturbance of any sort, they are colonized by bioeroders, including bivalves, sponges, sipunculans, polychaetes and echinoids, which break down the reef framework (Hutchings 1986). Disturbances that cause partial to complete mortality of coral colonies, provide fresh substratum for bioeroders, and in fact lead to increased bioerosion rates include hurricanes (Moran and Reaka-Kudla 1991), El Nifio- induced coral bleaching (Glynn 1990) and damselfish territoriality (Kaufman 1977), although damselfish can have a negative effect on bioerosion by excluding echinoids from their territories (Eakin 1988; Glynn 1990). Naturally and artificially high levels of nutrients on reefs also increase bioerosion rates (Highsmith 1980; Tomascik and Sander 1987; Hallock 1988). Overfishing off the coast of Kenya increased bioerosion and decreased topographic complexity as well, by releasing burrowing echinoids from predation (McClanahan and Shafir 1990). Furthermore, coralline algae, which are important in cementing the reef framework (and thus important in maintaining topographic complexity), are suppressed by macroalgae, which are promoted by nutrient input and the removal of herbivores (Littler and Littler 1985; Lewis 1986; Carpenter 1990). Coral growth, by contrast, generally increases topographic complexity at the scale under consideration (Dahl 1976). To a first approximation in certain situations, the topographic complexity index should be inversely related to total disturbance, with lower values indicating flatter terrain and suggesting more frequent, more recent and/or more intense disturbance. 8 No measure of disturbance is free of bias, including topographic complexity. One source of error is that coral mortality does not lead to the immediate loss of structure because bioerosion takes time. In addition to this time lag, the relationship between coral mortality and physical complexity is not always direct. Some coral species can survive breakage in storms and even reproduce asexually as a result (Acropora spp.: Highsmith et al. 1980; Tunnicliffe 1981); low complexity accompanied by high coral cover is a possible consequence. Conversely, bioerosion following partial or complete mortality of massive coral heads could initially increase complexity rather than decrease it. While these problems introduce error to estimates of disturbance, that error should be minimal in the spur-and-groove down to 15 m depth. Throughout Florida and the Caribbean, those habitats are now or were formerly (before disturbance) dominated by branching or other delicate coral species, including Acropora cervicornis (Belize, Jamaica, the Florida Keys and many other localities), branching Porites spp. (some reefs in St. Croix, U. S. Virgin Islands), and Agaricia tenuifolia (Belize). For all of these corals, complexity in their habitats should decline rapidly following mortality. Topographic complexity would not be as useful an indicator of disturbance in certain other reef habitats, such as shallow-water hardground areas, which are characterized by isolated head corals on limestone pavements. Similarly, deep-reef areas dominated by corals with a plating morphology could have high coral cover but low topographic complexity. Another potential complication in separating disturbance and herbivory effects is an observed positive correlation of fish abundance and topographic complexity. Herbivorous (and other) fish avoid low-relief areas, including those which have been disturbed (Hay and Goertemiller 1983; Kaufman 1983). Areas with fewer herbivores may experience increased algal cover, decreased coral cover, and increased bioerosion, leading to even flatter topography (reviewed in Hutchings 1986). This feedback loop does not appear to be a problem in the reefs studied (see Sample Sizes). SAMPLE SIZES The techniques outlined above are intended for testing hypotheses on large scales, ranging from a landscape scale (among spurs in the spur-and-groove habitat within reefs), to a subregional scale (among reefs within an area such as the Florida Keys), to a regional scale (among reefs throughout the Caribbean). In order to determine the appropriate sample sizes for statistical comparisons among reefs, preliminary surveys were conducted during 1992-93 in the spur-and-groove habitats of the four sites mentioned in the Introduction. Ten transects were completed at each site, and an additional 10 transects were videotaped only (see comments below on species richness). Two investigators can sample a site in 3-4 days, assuming 2-3 "full" transects per dive and 2 dives per day. Where funding and equipment are available, the use of nitrox diving techniques increases bottom time substantially, increasing the number of transects that can be completed per dive. 9 Percent cover and H’ are estimated by point counts from each of 10 video transects per site. Fifty non-overlapping, video frame "quadrats" cover most of the length of the transect videotape. The question is how many random dots to use per video frame. For each of the 50 frames in each of the 10 Carrie Bow Cay and 10 Discovery Bay transects that were analyzed, the substratum occupants under 25 random dots were recorded, in groups of 5 dots. The means and standard deviations of percent cover of hard corals and H’ for scleractinians remained virtually unchanged when the number of dots used was =10 (Figs. 1, 2). Therefore, an appropriate sampling protocol is 50 video frame "quadrats" per transect and 10 random dots per video frame. Each 500-dot data set constitutes a single sample, requiring 2-3 hours for trained personnel to extract from the videotaped transect. Because each video transect is a single sample, increasing the number of random dots per frame yields no advantage in terms of statistical power. The only parameter for which more than 10 video transects per reef are required is species richness, S. Plots of cumulative § for Conch Reef and Discovery Bay reach their asymptotes after 9 and 15 transects, respectively, and the curves for Carysfort Reef and Carrie Bow Cay both asymptote after 13 transects (Fig. 3). The difference in sample sizes required to estimate H’ and S is due to rare species, which add little to the coral cover from which H’ is calculated. We recommend 20 transects to estimate species richness for a site. The percent cover and topographic complexity data were arcsine-transformed and the fish census data logarithmically transformed, so that they conformed to the assumptions of parametric statistics (Sokal and Rohlf 1981). The square root transformation, often recommended for data in the form of counts (Sokal and Rohlf 1981), was not used on the census data. This transformation assumes that the data in question are Poisson distributed, but count data are not necessarily so distributed. Transformation of the H’ ‘data was unnecessary because H’ is normally distributed (Magurran 1988) and the data collected in this study were homoscedastic. (See Clarke and Green [1988] for a detailed discussion of data transformations.) A one-way analysis of variance (ANOVA) on the arcsine-transformed percent coral cover data showed significant among-site differences (p< 0.0005; Table 1). Tukey HSD multiple comparisons revealed the following differences between sites, listed in order of increasing cover (<, significantly less than; =, not significantly different from at a=0.05): Discovery Bay < Conch Reef < Carrie Bow Cay = Carysfort Reef. The ANOVA results were used to estimate the minimum detectable difference, 6, in transformed percent cover at a significance level of ~=0.05 with a power of 1-8 =0.90 (Zar 1984; Clarke and Green 1988). For an ANOVA comparing 4 sites, 5=0.13. This calculated 6 was then used to estimate the range of the minimum detectable difference in actual (untransformed) percent cover by the following procedure: 1. adding 6 to the (transformed) lowest mean, back-transforming that value to actual percent cover, and taking the difference between the back-transformed value and 10 —®— Discovery Bay 25 ]|—o— Carrie Bow Cay 20 15 10 Percent Cover 0 5 10 15 20 25 Number of Dots Fig. 1. Percent cover of hard corals as a function of the number of random dots used per video frame in transects from Discovery Bay, Jamaica and Carrie Bow Cay, Belize. Fifty frames were analyzed in each of 10 transects for each site. Error bars represent standard deviations. —®8—_ Discovery Bay 2.57|—o— Carrie Bow Cay 2.0 Ww 1S 1.0 0.5 0.0 0 5 10 15 20 25 Number of Dots Figure 2, Shannon-Wiener diversity of Scleractinia, H’, as a function of the number of random dots per video frame in transects from Discovery Bay and Carrie Bow Cay. Sample sizes as in Fig. 1. Error bars represent standard deviations. 11 Discovery Bay —oO— Carrie Bow Cay Conch Reef Carysfort Reef 0 5 10 15 20 Number of Transects Figure 3. Cumulative species richness curves for scleractinian corals in video transects from the four sites surveyed. Table 1. Means + standard deviations of parameters measured in 1992 at four coral reef sites. Field Site Parameter § Discovery Bay Carrie Bow Cay Conch Reef Carysfort Reef A. Percent Cover Hard corals 2.8+1.5 16.74+3.8 6.4+4.0 DNAse i) Macroalgae 91.2+4.3 63.2+6.9 65.7+8.8 40.9+12.0 B. Scleractinian Diversity 1’ 1.22+0.28 1.81+0.29 1.04+0.55 1.21+40.58 C. Integrated Disturbance Index C 0.184+0.078 0.403+0.089 0.194+0.077 0.335+0.106 D. Fish Abundance (number/transect) Parrotfish 31.5+10.0 24.6+7.4 17.0+8.3 34.24+15.8 Damselfish 7.5+4.5 38.5+7.9 58.5+14.2 18.5+7.8 12 the observed (untransformed) lowest mean, and 2. subtracting 6 from the (transformed) highest mean, back-transforming that value to actual percent cover, and taking the difference between the observed (untransformed) highest mean and the back-transformed value. The minimum detectable difference ranged from 5.2 % where coral cover is minimal (a few percent cover) to 9.8 % at high values of coral cover (approximately 20 % cover; Table 1). In fact, a posteriori comparisons after ANOVA on the arcsine-transformed percent cover data detected a significant difference between the sites with the two lowest means, Discovery Bay and Conch Reef, which differed by only 3.6 % cover (Table 1). If the study were expanded to include surveys of 10 sites instead of just 4, the minimum detectable difference would range from 6.3 to 11.1 % cover. A one-way ANOVA on the H’ data also revealed significant differences among sites (p<0.005), which are listed in the same format as for percent coral cover: Conch Reef = Carysfort Reef = Discovery Bay < Carrie Bow Cay. A power analysis gave 6=0.81 for an ANOVA comparing 4 sites and 6=0.93 for an ANOVA comparing 10 sites, at a=0.05 and 1-8=0.90. A posteriori comparisons detected a significant difference between the sites with the two highest means, Carrie Bow Cay and Discovery Bay, which actually differed in mean H’ by only 0.57. The video method thus provides a statistically powerful tool for detecting differences in coral cover and diversity among sites. The estimates of percent cover and H’ for Discovery Bay were compared to estimates obtained independently at the same time (during the winter of 1992) by the linear point- intercept (LPI) method (Ohlhorst et al. 1988). A 10-m surveyor’s tape was positioned randomly at 15 m depth, and the sessile organisms underlying the tape were recorded every 10 cm, for a total of 100 points per transect. The LPI method gave a higher estimate of coral cover (mean 4.4+0.9 SD, based on 5 transects) than the video transects (Table 1), but the two estimates were not significantly different by Student’s f-test (t,=2.207, df=13, 0.05 s POEWNOLAGE Coa TA Ae f (PTA Per (HVIIOQE AIA AIOY VOR ae ua iu) - iy . ote a a — : e ft, ‘ a 1 3 iP i Pa - oy y 4 CHANGES IN THE COASTAL FISH COMMUNITIES FOLLOWING HURRICANE HUGO IN GUADELOUPE ISLAND (FRENCH WEST INDIES) BY CLAUDE BOUCHON, YOLANDE BOUCHON-NAVARO AND MAX LOUIS ABSTRACT Hurricane Hugo swept the island of Guadeloupe (French West Indies) on 16 and 17 September 1989. Sustained winds were of 140 knots and gusts exceeded 160 knots. This hurricane was one of the most devastating of the century for the Lesser Antilles. The mangroves were completely defoliated and anoxic conditions of the water induced considerable fish mortality. Consequently, the fish community was modified in terms of species composition, structure and biomass. Four months later, the fish assemblages of the mangroves returned to conditions previous to the hurricane in species composition and community structure. The impact on the marine phanerogams was more destructive on the Syringodium filiforme seagrass beds than on those of Thalassia testudinum. In this ecosystem, the effect of the hurricane was minor on the fish community. Changes in the fish community occurred four months later in the seagrass beds and were apparently induced by a delayed mortality of the Thalassia testudinum. In the coral reef environment, the impact of the hurricane surge on the coral _ community mainly affected the branched coral species located between the surface and 15 m deep. The fish assemblages were not modified concerning their species composition. However, the proportion of juveniles in the community drastically dropped after the hurricane. Four months later, the proportion of juvenile fishes was still reduced. The overall effects of hurricane Hugo on the coastal fish communities of the island of Guadeloupe were minor considering the magnitude of the hurricane. I. INTRODUCTION In the Lesser Antilles, hurricanes are considered one of the major factors controlling the coastal marine ecosystems. In the island of Guadeloupe, these are represented by mangrove, seagrass beds and coral reefs. Hurricane Hugo reached Guadeloupe in the night of 16 September 1989, travelled the length of the island until the following morning, with the 37 Km-diameter eye passing over Université des Antilles et de la Guyane, Laboratoire de Biologie et Physiologie Animales, B.P. 592, 97167 Pointe-a-Pitre, Guadeloupe (F. W. I.) Manuscript received 20 June 1991; revised 4 August 1994 2 the Grande Terre (Fig. 1). The atmospheric pressure dropped to 941,5 millibars and the wind was recorded at 140 knots with squalls exceeding 160 knots. In some areas, rainfall reached 300 mm in one day. Such a rainfall rate has a probability of occurence lower than 1 per 50 years (Anon., 1990). Figure 1 : Guadeloupe Island. Track of the hurricane. Location of the study areas. Only few observations were available concerning the effects of the hurricane on the sea conditions.The theoretical calculation of the storm tide predicted a 3 m rise of the mean sea level (Anon., 1990). Our examination of the high-water marks after the storm indicated that the tide did not exceed 1,5 m. Offshore, the predicted amplitude of the swell was 5 m (Anon., 1990). On the shore, the structure of the waves is normally variable and depends on the morphology of the sea bottom and the incidence of surge along the coast. Unfortunately, no observations were made during the hurricane. However, the amplitude of the waves observed for similar hurricanes in the Caribbean area varied between 10 and 12 m (Stoddart, 1974 ; Woodley et al., 1981 ; Kjerfve et Dinnel, 1983). 3 The general impact of Hurricane Hugo on the different coastal communities of the island of Guadeloupe was previously reported by Bouchon et al. (1991). The present work summarizes the observations made on the changes in the fish communities during the months preceding and following the hurricane. Il. STUDY AREAS AND METHODS Observations were made in the Grand Cul-de-Sac Marin bay, for the fish in the mangrove and the seagrass beds. The coral reef fish community was studied near Pigeon Island, on the west coast of Guadeloupe (Fig. 1). These areas were chosen because previous data were available for them and provided a basis for comparison. After the hurricane, the first observations were made on 24 September, 1989 at Pigeon Island and on 25 September in Grand Cul-de-Sac Marin. The fish communities were studied with different sampling techniques because of the varied habitat. In the mangrove, where the water was turbid, fishes were sampled with a special fishing net called “capéchade”. This device consisted of a fence net (45 mlong and 2 m high), placed perpendicular to the mangrove front and three hoop-nets that trap the fishes. From the mouth to the extremity of the hoop-nets, the mesh-size decreased from 13.8 mm to 6mm (Fig. 2). The sampling station was located in the “Manche a Eau", a mangrove lagoon (Fig. 3) and important nursery zone for the fishes in Guadeloupe (Louis et Guyard, 1982). In the seagrass beds, fishes were collected with a seine net, 50 m long and 2 m high, used to encircle the sampling area. Two stations were chosen in Grand Cul-de-Sac Marin : one at Lambis Point and the other at Christophe Islet (Fig. 3). mangrove hoop net U CEO 04 Figure 2 : A “capéchade” : the fishing device used in the mangrove. The sampling area for the coral reef fish community was located near Pigeon Island at 15 m deep (Fig. 1). The fishes were counted, by SCUBA diving, inside a quadrat of 300 m? (150 m long, 2 m wide) defined by transect lines on the bottom. The water column investigated was about 3m high. Each fish censused was assigned one of three size-classes (juvenile, medium-size, big-size) based on the size range of each species (Bouchon-Navaro and Harmelin-Vivien, 1981). eons oenee " ° Paced nates Colas “Channel GRAND CUL-DE-SAC MARIN 1km Mangrove SS SS S % {\ BellePlaine | Coral reef j Co) BS Christophe Cy ess Sampling station \ . Belle-Plaine Figure 3 : Location of the sampling stations in Grand Cul-de-Sac Marin Bay. 5 From the data, several biological indices were computed such as species richness, species diversity (H’) calculated according to Shannon and Weaver (1948), and evenness index (E) of Pielou (1969) that gives an indication on the community structure. E fluctuates between 0 (only one species in the community) and 1 (all the species of the community have the same importance). These indices were calculated using biomass values for the mangrove and seagrass fishes and the number of individuals for the reef fishes. Data did not fit a gaussian distribution, even when using current transformation techniques (Log, square root, hyperbolic...). Three non parametric statistical tests were used to analyse the data : the Wilcoxon signed-ranks test, the Spearman rank correlation and the Friedman two-way analysis of variance by ranks (Siegel, 1956). Results are given with their exact probability of occurence. We considered that the results of the tests were statistically significant when probabilities were < 0.05. lil. RESULTS A. Mangrove areas The hurricane was accompanied by a storm tide which was followed by arise in sea level of at least one meter. The mud from the bottom in shallow waters was stirred up by the waves. Considerable amounts of freshwater runoff flushed the mangroves. These phenomena induced a drop in salinity (to 7%o) and quite anoxic conditions (0.2 mg.02.1") that lasted several days (Bouchon et al., 1991). After the hurricane, numerous dead fishes were floating at the surface of the water in the mangrove. Some fishes were observed dead on the substrate between the mangrove roots and up to 20 m inshore. The dead fish species were the following : Gerres cinereus, Eucinostomus gula, Eugerres brasilianus, Bairdiellaronchus, Lutjanus apodus, Haemulon ' bonariense, Mugil curema, Sphyraena barracuda, Chaetodipterus faber, Archosargus rhomboidalis, Diodon holacanthus and Sphoeroides testudineus. Fish surveys were conducted from 24th September 1989 (one week after the hurricane) and the data could be compared to data acquired previously at the same station. In the Manche-a-Eau lagoon, the results were compared to data from June 1989 (3 months before the hurricane) and additional samples made in January 1990 (4 months after the hurricane) and in March 1990 (6 months later) (Annex I). The Friedman two-way analysis of variance by ranks was used to compare the fish biomass among the four samples. A global statistical significant difference was found between the samples (X?= 11.709 ; p = 0.0084). The Wilcoxon signed-ranks test was used to compare the samples pairwise (Tab. 1). The results show that the fish community observed before the hurricane (June 1989) was different from the one observed after the hurricane (September 1989). Surveys conducted in January and March 1990 were also different from the September 1989 sample. But in January and March 1990, 4 and 6 months after the hurricane, fish biomass returned to the previous situation of June 1989. A drop in fish biomass was observed just after the hurricane. In January 1990, fish biomass had returned to the pre-hurricane values. Decreases in the number of species and 6 number of individuals, as well as the diversity indices, were also noticed one week after the hurricane. The fish community observed in the mangrove lagoon during the study period comprised 32 species. A Spearman rank- correlation coefficient calculated with the pre and post-hurricane data showed a significant inverse correlation between the quantitative fish dominances (Z = -2.817, p = 0.048). Before the hurricane, the dominant species in biomass were : Sphoeroides testudineus, Bairdiella ronchus, Archosargus rhomboidalis, Eucinostomus argenteus and Eucinostomus gula. These species usually correspond to the fishes permanently residing in the mangrove (Louis and Guyard, 1982). After the hurricane, these species were no longer present in the surveys, except for A. rhomboidalis (three individuals collected). Moreover, gobiid fishes which were not commonly sampled in the mangrove (chiefly Gobionellus oceanicus) were dominant in the community. Thus, significant changes in the fish community were observed just after the hurricane in the mangrove : 4 and 6 months later, the community had returned to its previous situation. Table 1 : Results of the Wilcoxon tests concerning the fishes of Manche-a-Eau lagoon ( Z = values of the Wilcoxon test ; p = probability of realization of Ho ; * = significant values). ae ee ee [wwe | sree [_psoom "| pore [a B. Seagrass beds In seagrass areas, a total of 50 fish species were collected in October 1988 (one year before the hurricane), in October 1989 (10 days after), in January 1990 (4 months after) and in March 1990 (6 months after the hurricane) (Annex II). At Christophe Islet, the Friedman analysis of variance revealed a significant difference between the fish biomass in the four samples ( X?= 17.891 ; p = 0.0013). The Wilcoxon test was used to test the difference between the samples pairwise (Tab. 2). Only samples collected in January 1990 appeared significantly different from those of October 1989 and March 1990. No significant difference was found in biomass between the samples collected in October 1988 and the 3 samples collected after the hurricane. Thus, there was no change in fish biomass immediately after the hurricane. Table 2 : Results of the Wilcoxon tests on the fish community of Christophe Islet (Z = values of Wilcoxon test ; p = probability of realization of Ho ; * = significant values). jae Werranaruies October 1989 January 1990 March 1990 | October 1988 aa Z =- 1.589 Z =- 1.663 Z = 0.368 January 1990 p=0.0964 | p=0.0001 * per aeeeae bn 2 =-3.346 March 1990 p=0.7132 | p=0.1919 p = 0.0008 * eae At Lambis Point, the Friedman test also revealed a significant difference between the samples (X?= 13.05 ; p= 0.011). The Wilcoxon test showed a significant difference only between the samples of October 1988 and January 1990, and between those of January 1990 and March 1990 (Tab. 3). As for the previous station, there was no change in fish biomass just after the hurricane. Conversely, a comparison of the Spearman rank correlation coefficients (rs)indicated that the fish community structure differed significantly before and after the hurricane in both stations (1s = -1.086, p = 0.277 at Christophe Islet and rs = 0.311, p = 0.756 at Lambis Point). These differences are partly due to the appearance in the samples of schooling transient fishes (Anchoa lyolepis, Diapterus rhombeus). Their suppression from the analysis increased the values of the correlation coefficients. October 1988 Table 3 : Results of the Wilcoxon tests on the fish community of Lambis Point. (Z = values of Wilcoxon test ; p = probability of realization of Ho ; * = significant values). [eso [omen [amine | C. The coral reef areas Pigeon island, a volcanic formation, is devoid of true coral reefs, but its steep slopes support the most flourishing hermatypic coral community of Guadeloupe. Concerning the fish communities, the results presented hereafter cover a 9 month period from April 1989 to January 1990. During this period, 12 censuses were made respectively before and after the hurricane. These censuses were separated by a 12-day interval. A total of 89 fish species were observed (Annex III). The Wilcoxon signed-rank test was used to compare the biological parameters obtained from the data collected before and after the hurricane, i. e., species richness, the total density of fishes; the number of juveniles; the number of medium-size fishes; the number of big-size fishes; the number of species possessing juveniles; the Shannon- Weaver diversity and the evenness index (Tab. 4). A significant difference was found for the total density of fishes, the number of juveniles, H' and the Pielou evenness. The other parameters such as the species richness, the number of big and medium-size fishes were not significantly different before and after the hurricane. Since there were no significant changes in the amount of medium and large fishes, only the juveniles were responsible for the observed changes in total abundance. Table 4 : Results of the Wilcoxon test concerning the fish community of Pigeon Island ( Z = values of the Wilcoxon test ; p = probability of realization of Ho ; * = significant values). a ets aera end a ee ee a ee Pielou evenness Moreover, a Spearman ranks correlation coefficient was computed between the profiles of fish abundances before and after the hurricane. The correlation was highly significant showing that there were no noticeable changes in the species composition or their dominance ranks within the community. Figure 4 shows the change in numbers of juveniles for the 24 samples distributed before and after the hurricane. An important drop in the abundance of juveniles can be observed just after the hurricane. The density observed remained low even four months after the hurricane and these conditions would probably persist until the next period of recruitment that occurs in summer. Number of juveniles We 20 3 4.56 7 8 29) JO) MN) 12,13) 14) 15) N67, 18) 19) 120) (21) 22) 23) 24. Samples Figure 4 : Change in number of juvenile fishes before and after the hurricane. IV. DISCUSSION AND CONCLUSION The effects of severe storms or hurricanes on the fish communities have been documented from many parts of the world. For the Atlantic region, reports can be found for Florida (Robins, 1957 ; Breder, 1962 ; Springer and McErlean, 1962 ; Tabb and Jones, 1962 ; Beecher, 1973 ; Bortone, 1976 ), Jamaica (Woodley et al., 1981 ; Kaufman, 1983 ; Williams, 1984 ), Puerto Rico (Glynn et al., 1964 ) and Texas (Hubbs, 1962. For the Indo- _ Pacific region, observations have been reported for Hawaii (Walsh, 1983), the Great Barrier Reef of Australia (Lassig, 1983), the Fiji Islands (Cooper, 1966), Japan (Araga and Tanase, 1966 ; Tribble et al., 1982 ) and Reunion Island (Letourneur, 1991). However, as pointed out by Walsh (1983), the effects of catastrophic storms on fish communities is still unclear. Some authors reported a high fish mortality after a hurricane, while others observed noticeable changes in the fish communities. Some did not observe any significant alterations in the community due to the storm. Among the authors who did not find noticeable changes in the fish communities after a hurricane are Springer and McErlean (1962) and Bortone (1976) in Florida. Springer and McErlean (1962) noticed that reef fish populations were not much disturbed after a hurricane although reef formations were destroyed. However, their observations occurred one month after the hurricane. Bortone (1976) concluded that no major changes occurred in the fish community as a result of Hurricane Eloise. He related this to the location of the study area (well oxygenated waters and not directly affected by the surge) and to the possible presence of protective shelters for the fishes. Robins (1957) was the first to report on the effects of a severe storm on fishes. He observed numerous dead specimens washed onshore after a severe storm in Florida. In the same region, Hurricane Donna also caused a high fish mortality (Tabb and Jones, 1962). After Hurricane Edith at Puerto Rico, Glynn et al. (1964) reported dead fishes floating near 10 the coast. Cooper (1966) presented a dismal picture of the reefs of Fiji Islands after the hurricane of February 1965 ; dead fishes were floating on the water and thousands were washed up on the beach. High fish mortality was also recorded in Japan after typhoons (Araga and Tanase, 1966; Tribble et al., 1982). Araga and Tanase (1966) made quantitative observations on the stranded fishes and noticed that about 84 % of the species and 98 % of the individuals were inshore inhabitants. In general, the fish communities from the shallow coastal waters are mostly affected. In the mangrove areas of Grand-Cul-de-Sac Marin, the trees were completely defiolated after the hurricane. However, the loss of wood biomass was variable according to the area. In the part of the mangrove areas dominated by the red mangrove, the estimation of the loss of biomass fluctuated between 25 and75 % (Bouchon etal., 1991). Fish mortality mainly occured in the mangrove areas where the fishes were exposed to low salinity, high levels of suspended sediments and oxygen depletion. The post-hurricane fish community was significantly different to the pre-hurricane community. The impact of Hurricane Hugo on the seagrass beds was varied. The Thalassia testudinum beds, even those situated in shallow waters, were only slightly affected by the direct impact of the cyclonic surge. On the contrary, the Syringodium filiforme beds were much more affected. A large amount of S. filiforme leaves and roots were washed onshore. In the months following the hurricane, a delayed mortality of the T. testudinum meadows was observed in the Grand Cul-de-Sac Marin. In some places, T. testudinum was progressively replaced by S. filiforme (Bouchon et al., 1991). In the seagrass beds, the observed changes in the fish community were more complex. They only appeared a few months after the hurricane. This may be related to the delayed mortality of Thalassia testudinum. In the coral reef environment the observed changes were less important than would be expected from the strength of the hurricane. For the benthic community, the damage due to the cyclonic surge mostly affected branching species of corals, such as Millepora alcicornis (especially in shallow waters), Madracis mirabilis, Acropora cervicornis, Porites porites and Eusmilia fastigiata. These colonies, broken and tossed by the waves, smashed the other benthic organisms. Massive corals withstood the hurricane better than branching corals. The soft benthic organisms, such as sponges and gorgonians were greatly damaged especially in shallow waters (Bouchoneral., 1991). During the weeks following the hurricane, a “bleaching” phenomenon affected many coral colonies. This bleaching consisted in the loss of their symbiotic unicellular algae (zooxanthellae). This is generally linked to a state of stress of the animals. Most of these corals finally died. Three months after the hurricane, the bleaching phenomenon progressively disappeared. Before the hurricane a dense algal community, dominated by species belonging to the genus Dictyota, were present at Pigeon Island. These algae were washed ashore by the storm waves. A few weeks after, an outbreak of a red algae belonging to the genus Liagora occurred. Three months after, the Liagora population disappeared and the Dictyota resettled (Bouchon et al., 1991). In the study area, Hurricane Hugo mainly affected the juvenile fishes. Their density on the study reef drastically decreased the week following the hurricane. The same observations were made by Lassig (1983) on the Great Barrier Reef of Australia who noted that “the cyclone had little effects on adults but caused high juvenile mortality and re- distribution of sub-adult individuals”. Beecher (1973) also reported a high mortality of 11 juveniles of a Pomacentrid fish, Pomacentrus (=Stegastes) variabilis, after Hurricane Agnes in Florida. In Guadeloupe, no specific changes in reef fish behavior were noticed after the hurricane. This is contrary to what had been described in Jamaica after Hurricane Allen (Woodley et al., 1981 ; Kaufman, 1983) where cryptic species were observed in the open waters and planktivorous species swam near the bottom. The territorial fishes such as Stegastes planifrons became more aggressive and schools of parrotfish were reduced in size. In Hawaii, Walsh (1983) reported that fishes from the reef flats moved down to the deeper zones. During the weeks following the hurricane in Guadeloupe, some acanthurid species (Acanthurus bahianus and A. coeruleus) were observed browsing the algae belonging to the genus Liagora that abnormally proliferated in the coral community. Nevertheless, examination of the survey results showed that the density of herbivorous fishes in the study areas did not increase significantly after the hurricane. This is contrary to what had been noticed in Martinique following the proliferation of Sargassum (Bouchon et al., 1988). In Jamaica, Williams (1984) and Kaufman (1983) had reported an increase in the number of Stegastes planifrons, an herbivorous species, after Hurricane Allen. The consequences of a hurricane on fish communities depend on various factors: the violence of the phenomenon ; the geographical location of the study areas ; the reef topography ; the depth location of the observations ; and above all, the magnitude of the damage on the reef associated benthic communities. In the island of Guadeloupe, the immediate impact of Hurricane Hugo was important for the fish communities situated in the mangrove. However, in this habitat, the fish community is well adapted to variations in environmental factors and apparently recovered within a few months. The changes which occurred in the seagrass beds reflect a long term decay of this habitat. As for the reef fishes, ' the drastic drop of juveniles may have an influence in the structuring of the fish community in the long term. ACKNOWLEDGEMENTS This research was funded by the Commission d’ Organisation de la Recherche dans les Départements et Territoires d’ Outre-Mer (C.O.R.D.E.T.) of the French Government. BIBLIOGRAPHY Anon., 1990. L’ouragan Hugo. Pub. Service Météorologique Interrégional Antilles Guyane. Météo France, 32 pp. Araga A., Tanase H., 1966. Fish stranding caused by a typhoon in the vicinity of Seto. Publ. Seto Mar. Biol. Lab., 14 : 155-160. Beecher H.A., 1973. Effects of a hurricane on a shallow-water population of damselfish, Pomacentrus variabilis. Copeia, 3 : 613-615. 12 Bortone S.A., 1976. Effects of a hurricane on the fish fauna at Destin, Florida. Fla. Sci., 39 : 245-248. Bouchon C., Bouchon-Navaro Y., Louis M., 1988. A firstrecord of a Sargassum (Pheophyta, Algae) outbreak in a Caribbean coral reef ecosystem. Com. oe Gulf Carib. Fish. Inst. Congress, St Thomas : 452-468. Bouchon C., Bouchon-Navaro Y., Imbert D., Louis M., 1991. Effets de l’ouragan Hugo sur les communautés cotiéres de Guadeloupe (Antilles frangaises). Ann. Inst. océanogr., Paris, 67 (1) : 5-33. Bouchon-Navaro Y., Harmelin- Vivien M., 1981. Quantitative distribution of herbivorous reef fishes in the gulf of Aqaba (Red Sea). Mar. Biol., 63 : 79-86. Breder C.M., 1962. Effects of a hurricane on the small fishes of a shallow bay. Copeia, 2 : 459-248. Cooper M.J., 1966. Destruction of marine flora and fauna in Fiji caused by the hurricane of February 1965. Pac. Sci., 20 : 137-141. Glynn P.W., AldomovarL.R., Gonzales J.G., 1964. Effects of hurricane Edith on the marine life in La Parguera, Puerto Rico. Carib. J. Sci., 4 : 335-345. Hubbs C., 1962. Effects of a hurricane on the fish fauna of a coastal pool and drainage ditch. Tex. J. Sci. , 14 : 289-296. Kaufman L.S., 1983. Effects of Hurricane Allen on reef fish assemblages near Discovery Bay, Jamaica. Coral reefs, 2 : 43-47. Kjerfve B., Dinnel S.P., 1983. Hindcast hurricane characteristics on the Belize barrier reef. Coral reefs, 1(4) : 203-207. Lassig B.R., 1983. The effects of a cyclonic storm on coral reef fish assemblages. Envir. Biol. fish, 9 (1) : 55-63. Letourneur Y., 1991. Modifications du peuplement de poissons du platier récifal de Saint- Pierre (ile de La réunion, Océan Indien) consécutives au passage du cyclone Firinga. Cybium, 15 (2) : 159-170. Louis M., Guyard A., 1982. Contribution al’ étude des peuplements ichtyologiques dans les mangroves de Guadeloupe (Antilles frangaises). Bull. Ecol., 13 (1) : 9-24. Pielou E.G., 1969. An introduction to mathematical ecology. Wiley, Interscience, New- York : 286 pp. Robins C.R., 1957. Effects of storms on the shallow-water fish fauna of Southern Florida with new records of fishes from Florida. Bull. Mar. Sci. Gulf and Caribbean, 7 (3) : 266-275. Shannon C.E., Weaver W., 1948. The mathematical theory of Communication. Urbana Univ. Press, Illinois : 117 pp. Siegel S., 1956. Nonparametric statistics for the behavioral sciences. Mc Graw-Hill, New York : 312 pp. Springer V.G., McErlean A.J., 1962. A study of the behavior of some tagged South Florida coral reef fishes. Am. Mid. Nat., 67 : 386-397. Stoddart D.R., 1974. Post-hurricane changes on the British Honduras reefs : resurvey of 1972. Proc. 2nd Inter. Coral Reef Symp., Brisbane, 2 : 473-483. Tabb D.C., Jones A.C., 1962. Effects of Hurricane Donna on the aquatic fauna of North Florida Bay. Trans. Am. Fish. Soc., 91 : 375-378. 13 Tribble G.W., Bell J.L., Moyer J.T., 1982. Subtidal effects of a large typhoon on Miyake- Jima, Japan. Publ. Seto Mar. Biol. Lab., 27 : 1-10. Walsh W.J., 1983. Stability of a coral reef fish community following a catastrophic storm. Coral reefs, 2 : 49-63. Williams A.H., 1984. The effects of Hurricane Allen on back reef populations of Discovery bay, Jamaica. J. Exp. Mar. Biol. Ecol., 75 : 233-243. Woodley J.D., Chornesky E.A., Clifford P.A., Jackson J.B.C., Kaufman L.S., Knowlton N., Lang J.C., Pearson M.P., Porter J.W., Rooney M.C., Rylaarsdam K.W., Tunnicliffe V.J., Whale C.M., Wulff J.L., Curtis A.S.G., Dallmeyer M.D., Jupp B.P., Koehl M.A.R., Neigel J., Sides E.M., 1981. Hurricane Allen’s impact on Jamaican coral reefs. Science, 214 : 749-755. 8s0 690 870 €9'0 XOpu] NOpIIg LS@ 167 68°0 8r'7Z XOpuy UouUeYS 02 61 6 sl soweds 7 ssouyou sotadg vIessy LOBE 6°8£667 0788 9'VEL6 09T 7 186L9 SL8 PIOL 8089 SI Te6l [t, €LIOVT She snauipnisaj sapiodaoyds AVALINOGOWLAL 197 9 ds snjjauoiqgoy v7 I 7L9 611 snJ1uvaI0 snjjauoiqoy VL@ €l 4oywsodos sniqoskywg avalgoo 67 I snaaidoyids skyyoiavyny AVCIH.LOG C871 I 4aqn{ snaajidipojavy aVdldIHda TOLIL cl TOST 144 Tol 9 SNIAIUID SALIBH cer v Lol I V'89€1 ZI snupiisvsg Sassaing vveez ber €rsrl ZIE 8°LLOZ 611 Dns snwojsourmng 6LSIZ tS Eve 91 € I Snajuasav snuojsouimngy P'SS6L Waa! Terr 196 T0ZIv ?9oT snaquoys snaajdviq avai 1679S OLT 68S8P we €LEO9T €€7 SnyIuod DJ aAIpsivg aAVCINAVIS €'CCOET vst S9OP9E 4 $989 € TO0S€6 60 SIDploquioys sn8avsoyo4y avdlvds 16 CG 6LS I asualivuog uojnuwavy] + AVAIASVWAWWOd L16 I snpodp snuv{inT AVGINVELAT vor I snans(1yo Snaquiorso10]yD 909 9 Usp 9 snanvs saj1ydo8Q SLICE 6 7ST (4 ces 9 SNID] XUDAD) AVAIONVUVO 0€ rd sndafisua snuodosjuaa vl £ £017 S SIfouJapun snuodosjuay FVAINOdOWLNAD secs L ESI v 9€9 v Dpn3v0440q vuavakyd§S AVGINAVYAHdS 60 (4 Duan [18nW avarnonw € Ori I DIIQUDSSOU UOpOsaY OLDS aAVaGTIHDIO ¢ SOE 8€ snyoiospfiun snydumysod{y] AVGIHdWVaINaH L€el I snzo snansojé [ AVGINOTAd 0vL6 v Sstaqaun{ xpsoyjouuky AVCINAVANWN 6'SOI € L'8€ I wnu1j3o0 puauoyisid¢ oT € (74! z LLI Zz sypsaumy vinsuasvp] 8781 60€ vL719 10€Z pjoadn]2 vjnsuasvj] avaladn 1 Il @ Saploadnjo viaoyouy oLPB81 799 10069 = 6SLP T081 I€ sidajok] voyouy = AVTINVUDNA 8618 8 poyuvio sdojpsay = AVAIdO1VDAW (M N (3) N (@M N @m N 06 YoreW 06 ‘uer 68 ‘Ides Boren S3ld3dS s3itiwv4 “‘yau-dooy oytoads 943 YIM payoayjoo saydures ysyy ay) 10J saorput AyIsI9.ATp pue (AA) SseWoIg ‘(N) SIoquINN, : | xouUY Annex II (continued) : Numbers (N), biomass (W) and diversity indices for the fish samples collected with a seine net in the FAMILY LUTJANIDAE POMADASYIDAE SPARIDAE SCIAENIDAE GERREIDAE BOTHIDAE SOLEIDAE SCORPAENIDAE CHAETODONTIDAE LABRIDAE SCARIDAE GOBIDAE ACANTHURIDAE MONACANTHIDAE TETRODONTIDAE DIODONTIDAE Total Species richness Shannon Index Pielou Index SPECIES Harengula clupeola Anchoa cf lyolepis Cosmocampus elucens Syngnathus sp Holocentrus rufus Sphyraena barracuda Hypoplectrus puella Serranus flaviventris Caranx latus Oligoplites saurus Selene vomer Lutjanus analis Lutjanus apodus Lutjanus griseus Lutjanus synagris Ocyurus chrysurus Haemulon aurolineatum Haemulon bonariense Haemulon chrysargyreum Haemulon flavolineatum Haemutlon plumieri Haemulon sciurus Archosar gus rhomboidalis Calamus sp Bairdiella ronchus Diapterus rhombeus Eucinostomus argenteus Eucinostomus gula Gerres cinereus Citharichthys spilopterus Achirus lineatus Pseudupeneus maculatus Scorpaena grandicornis Chaetodon capistratus Lachnolaimus maximus Sparisoma chrysopterum Sparisoma radians Gobidae sp.1 Gobidae sp. 2 Gobidae sp.3 Gobionellus oceanicus Acanthurus bahianus Acanthurus chirurgus Monacanthus ciliatus Sphoeroides nephelus Sphoeroides greeleyi Sphoeroides spengleri Sphoeroides testudineus Diodon holacanthus 50 species 138 215 W (g) N 81 97 5 32.3 2 31.3 2 he) 4 633.7 90 12.3 3 2 2 10.5 i 9 1.5 85.5 34 42 88.2 2 0.4 4 3.3 1.4 1 43.2 1 85 2 1045.7 2A5 16 2.14 0.54 922.2 28.1 6.3 287.5 151.7 15.1 LAMBIS POINT Oct. 89 W (g) Albula vulpes 98 Jan. 90 W(g) 309.1 85.3 193.3 60.4 157.3 152.2 3.1 0.8 seagrass beds of Grand Cul-de-Sac Marin Bay. March 90 N W (g) 6 27.1 3 6.1 10 37.3 1 15 153 396 Wi 14.2 7 42.4 9 57.3 12 83.9 1 2.6 2 20.4 3 12.8 2 0.3 1 23.8 4 3.1 2 1.6 1 5.2 1 293.8 225 1042.9 18 2.7 Annex II : Numbers (N), biomass (W) and diversity indices for the fish samples collected with a seine net in the seagrass beds of Grand Cul-de-Sac Marin Bay. CHRISTOPHE ISLET FAMILY SPECIES Oct. 88 Oct. 89 Jan. 90 March 90 N W (g) N W (g) N Wg), NN W(g) ALBULIDAE Albula vulpes 1 5.8 1 20.3 CLUPEIDAE Harengula clupeola 27 81 ENGRAULIDAE Anchoa cf lyolepis 8 13.9 347 995.3 360 550.5 SYNGNATHIDAE Cosmocampus elucens Syngnathus sp HOLOCENTRIDAE Holocentrus rufus 1 43.6 SPHYRAENIDAE Sphyraena barracuda 2 143.5 6 8.9 4 134.5 SERRANIDAE Hypoplectrus puella 1 2.8 8 77 2 14.7 Serranus flaviventris 10 42.7 3 3.8 17 51.6 4 13.1 CARANGIDAE Caranx latus 1 2 1 6 Oligoplites saurus Selene vomer 5 11.9 1 5 LUTJANIDAE Lutjanus analis 1 21.9 Lutjanus apodus 3 13 1 74.1 Lutjanus griseus Lutjanus synagris 75 55 43 278.2 12 88.5 Ocyurus chrysurus 152 595.2 9 69.3 115 583.8 5 18.1 POMADASYIDAE Haemulon aurolineatum 4 4.5 Haemulon bonariense 1 8.8 1 5.1 2 10.1 Haemulon chrysargyreum 2 17.6 Haemulon flavolineatum 13 18.2 2 5.2 Haemulon plumieri 14 114.1 Haemulon sciurus SPARIDAE Archosar gus rhomboidalis 2 81.7 1 31.8 2 45.5 Calamus sp SCIAENIDAE Bairdiella ronchus 3 16.9 GERREIDAE Diapterus rhombeus 128 945.3 398 1173.6 99 344.9 Eucinostomus argenteus 12 23.5 31 135.4 Eucinostomus gula 70 293.7 258 358.6 200 609.9 103 472 Gerres cinereus 2 8.6 8 63.4 BOTHIDAE Citharichthys spilopterus 1 15.7 9 23.7 4 19.1 SOLEIDAE Achirus lineatus 20 355.7 11 4 Pseudupeneus maculatus 1 14.3 SCORPAENIDAE Scorpaena grandicornis CHAETODONTIDAE Chaetodon capistratus 7 14.6 1 6.6 LABRIDAE Lachnolaimus maximus 1 49 SCARIDAE Sparisoma chrysopterum 1 58.3 Sparisoma radians 5 51.6 GOBIDAE Gobidae sp.1 17 57 Gobidae sp. 2 2 0.4 Gobidae sp.3 4 0.9 1 0.1 Gobionellus oceanicus 5 1:5 ACANTHURIDAE Acanthurus bahianus 1 17.2 1 2.9 Acanthurus chirurgus 1 8.4 1 1.2 MONACANTHIDAE Monacanthus ciliatus TETRODONTIDAE Sphoeroides nephelus 1 0.2 Sphoeroides greeleyi 1 4.3 2 0.9 Sphoeroides spengleri 1 12 1 0.6 1 0.5 Sphoeroides testudineus 1 0.5 2 9 1 1 DIODONTIDAE Diodon holacanthus 2 139 4 987.9 Species richness 50 species 19 17 29 20 Shannon Index 2.98 1.83 3.29 2.66 Pielou Index 0.7 0.45 0.68 0.61 land. igeon is Number of indivuals per species observed before (1 to 12) and after (13 to 24) the hurricane at P Annex III 23-24 16 17 15 12 10 Lycodontis moringa Muraena miliaris Synodus intermedius 10 Neoniphon marianus Holocentrus rufus Myripristis Jacobus 14 «(14 14 22 13 36 22 2 BS G3 e214 2222013 17 17 16 Aulostomus maculatus Fistularia tabacaria Scorpaena plumieri 14 Cephalopholis cruentatus Cephalopholis fulva Epinephelus adscensionis Hypoplectrus chlorurus Hypoplectrus cf guttavarius Hypoplectrus nigricans 11 Hypoplectrus puella Hypoplectrus sp. 1 Mycteroperca interstitialis Paranthias furcifer Serranus tabacarius Serranus tigrinus 1 Heteropriacanthus cruentatus Carangoides ruber Caranx latus 0 0 275 4 0 Selar crumenophthalmus Inermia vittata 0 115 40 0 0 0 20 20 Lutjanus apodus Lutjanus jocu Lutjanus mahogoni Sao a =O - OMNWO HBO KH OO KH NH OA SC SC - OM SO = GN 19) 16) Fest Fen eS OOO OO Oe OS. (Ole So N OnmOoOnornn et COR ONR COA ea i=) = ONION tN) IO) IO SCO OO ORE REO OT OTENG = 6LST SO (OF OMAN ia QE ORO Or Qi SO ieien o Ol Olen = zr Oo = CA) et OO) ot ONO) et (OO! OC) (OS 00 NN) et ie oooroo0o9oo so i=) _ SN ON XN own oonmoom oon =m fon) N _ N _ WMO R= OR Nw 1) (TO On Ouinig~=|| (Olen ~ an o mm ~ mooorcr gogo OoowON oF DOO oO oO mM 001 \o é~ SS OO 4OF SO) eH 1S LOMO COIN Oe tHs) OF OF OF 91 06S cooonorocewrocdeo0ocmcU8O Fr Kr OT NO So CS & 8 N CA Dh ASS on Hi) bas i Yin Cig PLN See FCN 3 3 — on a3 _ won oe Nooo on nan Oe an On Oo om an ~-m Wn aa eS SO) ONS eS 1 et OO Oren OO AN) tO) STON _ _ =a Orr omooqonj#oo aqaoonmnnm © © = (OMAN OP aan! ©) ix = foe) N ee) eye ey eo te) fe) Te) t=) poe oO OL O OO et (OO Oa Oe eat Oe Ol! ie =aOnmOoONwres OO OO MR NWO A CO SF mooocq*croeoanrtieocerOmvn coer neoseosd & _ @ ~ oocaonr rer CO CO KF MNO = 18€ if + Oo71C Cran oO = OO OM OO we ON © i= Yousws Sasv0yID}] avisod snaydaja snfn4 snuvipog ppnzv.10q vuavskydy souid snjiysnokjquy suosfiuvjd sajsv8aj¢ snjyaod saispsajg snans€ay uopoyswdsosn1py ELT LULL OU ETD) vauvko snuo1y7) syuoxos fnpfepnqy nipd snyjuvIDUOdg 40] 0914] SNYIUDIDIO}] snjo14js uopojavy) snyjo4jsidvo uopojavy7) snypajngv uopojavy) saaojoas snsoydky snjojnz0u Snauadnpnas gq SNIIUNAOU Sapio]jNp sniojound snyanbq SNUIDIDI snunjIDD opouolog snunjo07D Snani9s uojnuav}] 1aatumnyd uojnuavyy unjoauijoani { uojnuan fy umasksavs(iys uojnuan fy WN1ADU0GLDI UOINUIAD FY SnIMuig4ia Snuasosiuy snanskays snank3Q wv nN a aQ nN = nN —) AN fa) = ) = L oa = ‘eo = Le) = vt = 4) = N = = = Ss = i) t=) é .) 2) vr ise) nN = “purjst UoosTg 1e SUBOLLNY oY (PZ 01 ET) Jo1Je pue (ZT 01 [) B10Jaq paatosgo satoads sod spenarput Jo Joaquin] : (penuTIUO}) [J] XxouUY oO © O © 0 0) —O 0° O © OD 0 OO OO © 4 4 (4 Chew [ds sajmugane On 0c w0e Oi 0) 0 a0 OM, LO BODO ie sO 20 08 0 20; 0, 0= 0-0 80 40 20 snyjuvovjoy uopoiq 0-0 - O20 02-0 © 0-0) © 0 -0° 10-0 0) 0 Oo 0 0 Om © snyouuaqun skyiyay KZ [aa Ot paren Celene ye lce at 9 wen ee eee Ce oO, Cee ae nO ee ee Ce Ge 50 DIDAISOA 4a]SDSIYIUD Oe Omen Ob: Cl Cente ce Cs mle: 6 = 0) 0 Oe OO Oe al Ne’ te Oe aganbiay sKsydojov7] O =O 0.0 0 30 “02 0-0 OO. 0 OO OO FO 20 OS syopnong sKaydojav] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 t 0 0 0 O = snMo8Kjod uo1DD4s0YysUDIy oO 0 0 O20 - 6 tO Oi 0" 0 0 0 60 =0- Oo 020-0 M0 oO 4astu skyyoynay ee 0S Ole Oe One 220 fe Oe ee Cpe ee Ca Cale OSs Ieee lat — 0 el ee 9D snjnd sauiysayjuvg OS a= s0eOe Je Oe le <0) 10> 0) 002 0) 0-0 0? 2 0 0 02101 00 0 <0) 0 SNAIFJOATOU SIUIYIYJUDD Oo O° OO OO OOO 0 070-0 0 OOO 0-9 0 © © T-0 © snjoun] snyjog Orn Oe tO REROR Oe Oneal. <.0l- ol 20) 0 OF 20s le KOE SAI Sle 0 <0 Ge 0) C6 0e. 0 $1]0834 SNAOWOLIQUOIS ClearCase bint BRL 9% CS CMC Gi anGS se eel 0 SP eke aoe SE 01 0 gsnajn42oo snanyjuvdy QO 0 © O 0-8 0 GY .O 0 © © G2 O00 0-20-00 -D © 0 0 snsanaiy) snanyjuvoy (leer Ce lence OV SC COP SCO lem Cok IO Ce al el 0 Oe eee hI snupiypog snanyjuvoy 0 0 OO 8% OF 9 8h 2% 09 OPI 08 09 SOP SIE SOS SEZ S68 SLE COI 08 7 0 O O snyouossad snaadoydkioD © © - 0 -O © O02 © © 0 @ OD tf O60 0-0 0 0 O° OD GO © @ -O wnianwojn puosi1s0ds Pee LP ABUT etfs NTSC eG me aT cb Ro SVs Abe eet ee ee C= FC C= XG apiaia Duosisvds Omer 20% 52 Oj sh ale le ice Cae ON OS SIS abe Berne ce 0 OenC G eats oc 1G 20 auuidiaqns puosisvds 0 1k 0 OO -O- hh -O 0-0 -O O- OO O20 0 OO O- oO T O20 wnsajdoskys vuosisvds GeO 6 OB Se Ue ire OE Bh Ge ONL GE OB 4 ee OR. winjouasfoinv Duosisvds On Ones On eOR e oel ee SCo) oie meet Oe BOR eee Ie Ole Ce OO FG Oa AU DINJaA SNADIS [Pole Sie 1%) SIl CO. OG 6 CC mole oCCm, SE 8 Ole eT Bl 8G Gh LG 8G. OF th 0G. IT snaajdoiav] snvr¢ [eles One Ouran OPW KORO: -201—2 10-9 g0r ORT. ON C08 ae OO 0-25-08 070 0 tr 0 114981 SNADIS O09 O O02 OO Oe 1 OO hy a OS OO 0) OO 1) DIDUIDIVNS SNAvIS 8h 6€ OF €8 Srp L9 OL 8S SS 6€ 7B Bh SEL SII 8ZI BSI LOI b7Z 88% Ipl ESI €8 8h 89 winjniosvfiq DUOSSDIDY J OFF 0h 0-08 20" 20! a0 0) 0) 0) 50) <0. Io 0e 0ne0' 0 = 05-0 0- 0. 0 Or 0 SNUICOU SNUIDJOUYIVT OR 00) Re ONe Onss0e Me (0 AN. 0 Wee 1? 0 0 0 02 Om 8 0. 20 50 SNIDIPDA SALIOYINOH oe t¢ @ JW oOo 60 8 LE OO ST ot €f @ IW or 6 8 L 9 Smee, € 4 I ‘PUPS! UOSSIg 1B DUROLLMY oI (PZ 01 €[) JoIye pue (Z] 01 [) a10Jaq Poarasqo sardeds Jod sTenarput Jo Joqumy : (penuNUOS) FT xouuy i P Eki if ie } i 1 le ( | ee we’ [hig F m Me : ey eeaecee fs Teens - i : : , iy | o- “bos ts Wie ; 2 1 . _ a Se Rea at a BP ee : ; 5 ae ae adnan Oa i BS ie er Be ete His a \ 1 4 . 7 # | i F x ] ge | @ { ; 13 as ane came = in 5 3 Ps Uy le : ; ‘ q a ah a | a a | i a } aie ‘ il ge - Som tig? E> 4! 3 ) : 1 ae es Pos ' if h \ i 5 2c) c (fhe ? «4 : sh y i forks, ae Liane cys (ae +4 j vel a ae | ee. f < Pa | j A : -_ oo «2 Ke i t ] \ | AS , { es § fra \ 7 ‘ ; pf -~ io oy s i tee ; i . / & ! : dy i i H \ vee * a4 i ge | oe ve a ; , i : ~ f 5 .. j ‘ ee 1 i i 4 ae "SS z i = j be it ‘ = ; \ : ae | ahead ah ’ 2 > * a ; i. , ant = I J ae ' i: ‘ , a 4 . We x ; ’ Z ' | : f i ‘ 4 . 7 4 t of ; a; "J 4% i | 7 ot ee {i att i} a ay a z \ z + 7 2 7 : uy t1> he . ¥ ' ’ 7 ATOLL RESEARCH BULLETIN NO. 423 THE SIAN KA'AN BIOSPHERE RESERVE CORAL REEF SYSTEM, YUCATAN PENINSULA, MEXICO BY ERIC JORDAN-DAHLGREN, EDUARDO MARTIN-CHA VEZ, MARTIN SANCHEZ-SEGURA, AND ALEJANDRO GONZALEZ DE LA PARRA ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 ea ; | : j { ; | ‘ i ; ae {) < i ‘3 | th i | ie 1] i ei i i 7 H . if HI | i l LS) fo} (o} fo) 09 Ov 0 ooo! ooot Al 0002 0002 IA OOOE ooo€e 1SW sw Wt 1SW SuY3aLaw SY¥313~N Figure 2. Bathymetric profiles for each section. The horizontal line at the observations. -40m level marks the deeper limit of underwater *“squsueinsesu jo sqes Tereaes Jo sebereae jueseidei pue ajzemtxoidde erie suoTsueuTd -paze Apnys euR ut punoj sedAQ joer uTew ey FO saudqexs °€ eanbty GuvMv3as -— 30-70m —H 3YNOHS ite SNOOT1E-HYNdS SYOHS wooe-002 ——————> Wa SALAS ——— wosi-0sS ——}{ 3YOHS 1 —_——— wos- ofr CLL Y/ WS \\ ee Ole INO) ive (RICE FC Yank fond 2) On = NO Se i), (™) H1d3G (w) H1id30 (“) Hid30 TUK ie SUMP 2 & S4335Y VYOdOYNOV EEO FOiry a: fh 50- 100m — S4535H WHOALV Id 3YOHS -— wos:-00!1 ———4 ornmw ww 3y¥OHS a wosi-o001 —— 4 rs Soy es Z =u 0 (YW) Hid3ad (w) H1d 30 (Ww) H1d30 DEPTH Slope Sections Platform Spur block Reefs Reefs Acropora Reefs Hard Sand grounds Figure 4. Main reef and substrata distribution on the shelf of Sian Ka’an. Wi-a Vi-a IV— @ REAR REEF- Vil-—@ crest y-a HIGH COVER Vil-a Wi-c i —b PLATFORM iV - b REEFS l-a W-c ll-d NON-CORAL ll-a Vil-c¢ Vi-b IV—'c MIDDLE V-c SLOPE Vi-c ll!-—b = 2 LOW COVER ec Ii-d VII|—d DEEP SLOPE Vill-—c Vi-d V-b V-d Vil-d Vi-d (Sec ar ee OA SO Nee an SL 2.0 EUCLIDEAN OISTANCE 0.0 Figure 5. Dendrogram of cover attributes (excluding hard ground and sand). Cluster analysis of log- transformed cover data with average linkage. “fo 100 80 | 60 40 20 Oo <= - RR-C P-R {MIDD DEEP N-C | High Cover Low Cover LZABiotic Gi Dead coral C=JSand/rubble (J Hard ground Figure 6. Percentage of cover for biota, dead coral, hard ground and sand, for each of the clusters suggested by the dendrogram in Fig. 5. cm Lineal 800 600 400+ Et ul i \ 200 5 NN % 3 \\ ‘NN TNT] EEL \ NN NY 4 | NA eT AA \ NN N SCtNN “| NN N NN NN = \\ PENN IN 8 NN WN LEE NS aLLN : rN, ERE RR-C P-R N-C MIDD DEEP High Cover Low Cover ESS Coral EJ Sponge MSSGorgo [2] Algae Figure 7. Mean lineal cover of the main biotic attributes, for each of the clusters suggested by the dendrogram in Fig. 5. The vertical line indicates one standard deviation. cm Lineal LOCLSS CORES CAS OSS x 2x Cx xX Cx xX ras TORO Oe 3050 g-9-6:0% Pas oi yas 500 <2 yas x Bess O RR-C P-R MIDD DEEP High Cover Low Cover (XX A.palmata M8 M. annularis EZ]] Massive Branching L_lEncrusting CA eatty Figure 8. Mean lineal cover of the coral attributes for each of the clusters suggested by the dendrogram in Fig. 5. The vertical line indicates one standard deviation. i a 4" ; } T5 eee & : e a Fee)! ‘ " * “ * Le! ‘ ee ry > ‘ ’ ‘ h 24 i ; a * Pe - 3 pee Ne mtyer'é rg sy Han + et : iv +h) fees 16% fi = ctr 7 i i fe ‘s » t J " wy Sao bas. Ay Leo) SB TP tty f helena , ‘ yt CY eae Var eee ; Aol teivet Syrah aes: er _ / ee ATOLL RESEARCH BULLETIN NO. 424 | A PRELIMINARY EVALUATION OF THE COMERCIAL SPONGE RESOURCES OF BELIZE WITH REFERENCE TO THE LOCATION OF THE TURNEFFE ISLANDS SPONGE FARM BY J.M. STEVELY AND D.E. SWEAT ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 ete aan i a —— iE LT SNR A TG a | , err Peewee Shuma Vs ; 4 i - ” - ‘ 3 Dey 1 = - - ° re vi y 7 Meet ATR ty nH 4 Vi J AMGTT AV ; eS ak ; H y y it : WAURE I -- Sa eee x = # ate Te iye i Fs ; + ba ‘4 P a mi Pe ~ {~ i j aj A PRELIMINARY EVALUATION OF THE COMMERCIAL SPONGE RESOURCES OF BELIZE WITH REFERENCE TO THE LOCATION OF THE TURNEFFE ISLANDS SPONGE FARM BY J.M. STEVELY AND D.E. SWEAT SUMMARY OF FINDINGS Our discovery of concrete disks used for planting sponge cuttings confirms the location of the Turneffe Island sponge farm activities as described by Smith (1941). It is hoped that this information will be useful to the Government of Belize in identifying potentially historic or unique marine resources. We here also report the occurrence of the velvet sponge, Hippospongia gossypina, at Turneffe Islands. The occurrence of the velvet sponge is particularity worthy of note since a devastating commercial sponge mortality in 1938-39 drastically reduced velvet sponge abundance throughout a portion of its geographic distribution. We observed that the attachment substrate of the Turneffe Islands velvet sponge, mangrove peat, was different than that reported for sheepswool sponge, Hippospongia lachne, and different than the attachment substrate previously reported for velvet sponge. The quality of the Turneffe Islands velvet sponge is such that it is commercially marketable, but would be less valuable than the sheepswool sponge. The velvet sponge is sufficiently abundant at one location to support commercial fishing activity. However, our survey work was not adequate to establish whether the abundance of velvet sponges was sufficiently extensive to support a sustainable sponge fishery. The lack of more extensive data on abundance and the historical accounts indicating the effects of past sponge disease and fishing pressure on velvet sponge distribution warrant a conservative approach to managing velvet sponges as a commercial! fishery resource. Additional survey work will be required to more fully understand the commercial sponge resources of Belize. Florida Sea Grant College Program University of Florida P.O. Box 110400 Gainesville, Florida 32611-0341 Manuscript received 5 January 1993; revised 5 June 1994 INTRODUCTION The usefulness of commercial sponges is based on their soft, compressible nature and their ability to absorb and hold water. Sponges from the genera Spongia and Hippospongia have been used for personal bathing and household cleaning for thousands of years. The taking of sponges for commercial purposes was first practiced in the Mediterranean Sea and the sponge fishery was often noted in early Greek literature (Moore, 1951). More recently, commercial sponges have been used in manufacturing pottery and ceramics, and in a variety of applications in surgery, painting, polishing, printing, horse grooming, and professional cleaning services. Synthetic sponges have replaced natural sponges for many of these uses because they are cheaper and more readily available. Today, natural sponges are principally used for bathing (Josupeit, 1991) and the application of cosmetics by people in westernized communities because these people prefer a natural product (Wilkinson, unpub. mans.). Although synthetic sponges are less expensive, they cannot equal the softness and absorbency of natural sponges, and, importantly, natural sponges can be more easily and thoroughly cleaned due to the truly porous nature of the sponge skeleton. Until the 1840’s the world’s sponge supply was derived solely from the waters of the Mediterranean. However, the discovery of quality commercial sponges in the Bahamas and Florida Keys led to the rapid development of sponge fisheries in Caribbean and Gulf of Mexico waters. During the early part of the 20th century (1900 - 1940's) the commercial sponge fishery was the most economically important fishery in Florida, U.S.A. (Stevely, et al., 1978). Cuba produced 440,000 Ib (198,000 kg) of sponges, the Bahamas 670,000 (302,000 kg), while U.S. production totaled 610,000 Ib (275,000 kg) (Moore, 1951). The large-scale attempts to culture sponges in the Bahamas (Storr, 1964), Belize (Smith, 1939) and Florida (Moore, 1910a; Shubow, I969) further attest to the importance of sponges as a fishery resource during this period. A sponge disease swept through the Caribbean and Gulf of Mexico in 1938-39 and dramatically reduced commercial sponge abundance. The decline in supply caused by the disease and the outbreak of World War II, which curtailed production in the Mediterranean, resulted in dramatically higher prices. Although production in Florida declined precipitously in the 1940’s, rapidly escalating prices were sufficient to increase fishing effort and actually increase the total value of the fishery. The result was increased fishing effort at a time when fishing effort probably should have been curtailed to allow the commercial sponge populations to recover (Storr, 1964). The Florida commercial sponge grounds were depleted to the point of causing the virtual economic extinction of the fishery and many years were required for commercial sponges to increase to abundances that approached those found before the sponge disease epidemic. The effects of the sponge blight were similar throughout the Caribbean. The cause of the disease has been confused somewhat by the presence of bacteria including symbionts, that live in close association with sponge tissue (Lauckner, 1980). Although the Caribbean sponge disease and introduction of synthetic sponges in the post World War II era has resulted in reducing the world sponge trade to a fraction of its former importance, a significant sponge trade still exists. Prior to WW Il (1927-1936), world sponge production annually averaged 1,346.1 MT, and in more 3 recent times (1977-I986) it has annually averaged 222.1 MT (Josupeit, 1991). The world’s supply of bath sponges comes from the lesser-developed countries of the Mediterranean and Caribbean: Tunisia (48%), Greece (17%) and Cuba (26%) are the principal suppliers. Countries importing the largest volume of natural sponges are France (37%), USA (26%), Japan (10%), and Italy (9%) (Josupeit, 1991). In terms of the quantity landed by weight, most fishery managers would consider the world sponge fishery to be insignificant. However, it must be noted that the highest grade of commercial sponge can command a price of over U.S. $50.00/lb ($110.00/kg) in the export market. Currently, market demand for natural sponges is such that the opportunity exists for expanded production from the Caribbean and Gulf of Mexico. A decline in commercial sponge abundance caused by disease (Gaino and Pronzato), 1989), and pollution and overfishing (Verdenal and Verdenal, 1986) has significantly reduced the supply of Mediterranean sponges. Reduced supply has resulted in higher commercial sponge prices and focused attention on increasing sponge production on other areas of the world. For example, Josupeit (1991) reported that, in France, the reduced sponge supply, as well as a general trend for increased use of natural products, resulted in retail sponge prices doubling and even trebling. During |988, Florida sponge prices more than doubled and fishing effort and production significantly increased (Stevely, pers. obs.). Although Florida sponge prices have stabilized below the peak prices of 1988, they are still substantially higher than pre-I988 prices. Increased market demand and the consequent higher prices for Caribbean sponges has resulted in a need to carefully assess sponge fishery potential and fishery management needs throughout the region. The objectives of this project were to (1) evaluate the fishery potential for harvesting commercial sponges in the marine waters of Belize, (2) provide the Belizean Fisheries Unit with information pertinent to management of a commercial sponge fishery, and, (3) establish the location of the historic Turneffe Islands commercial sponge farm before its location was lost to posterity in general, and to the fisheries heritage of Belize in particular. Project funding was provided by the Smithsonian Institution’s Caribbean Coral Reef Ecosystem Program (CCRE Contribution No. 402). METHODS AND MATERIALS Field surveys to determine the distributions and abundance of commerical sponges were conducted from !l4 May to 30 May, 1989. These surveys were performed using dive mask, snorkel and fins. To cover large areas, a diver was towed by boat for 30 to 90 minutes. Three locations were surveyed (Figure 1): Carrie Bow Cay, Ambergris Cay, and Turneffe Islands. The maximum and minimum diameters of commercial sponges were measured with a pair of large calipers. Field notes on duration of tow and habitat type were recorded. Plans to collect commercial sponge abundance within quantifiable transect lines had to be aborted due to inclement weather and contractual problems with a local fishing guide. Surveys at Carrie Bow Cay, capitalizing on the Smithsonian Institute facilities there, were conducted to field-test survey procedures, and collect information on the distribution of commercial sponges in the vicinity of the barrier reef and associated Belize 30 mi 40 km ,7~ MEXICO ¢ /. TURNEFFE B3/ ISLANDS yah ) / =< — <= = Ld — < => [aie] CARIBBEAN SEA Figure 1. Location of commercial sponge survey sites (Carrie Bow Cay, Ambergris Cay, and Turneffe Islands). 5 habitats. Ambergris Cay was chosen as a survey site because anecdotal observations suggested that commercial sponges could be found there and because a fishery cooperative was located at San Pedro. Turneffe Island was selected as a survey area because it was the reputed site of a large scale sponge farm in the 1930’s and interviews with fishermen indicated that Turneffe Islands was the most likely area where commercial sponges could be found. To assist with the Turneffe Islands field work, the services of two fishermen were contracted, one of which was old enough to have personal knowledge of the sponge farm location. Fisheries Unit personnel, commercial fishermen and fishing guides were interviewed to obtain information prior to conducting the field work in each area. RESULTS A total of 19 locations were surveyed by either towing a diver or by having the boat follow 2 divers. These surveys represent a total of 24 hours of underwater observations. Distribution and Abundance of Commercial Sponges Carrie Bow Cay No commercial sponges of the genera Spongia and Hippospongia were found. Habitats surveyed included: seagrass beds surrounding Twin Cayes (occasional loggerhead sponges, Spheciospongia vesparium, were seen), seagrass beds west of South Water Cay, Barrier Reef sand/rubble zone, and Barrier Reef habitat accessible by snorkeling from South Water Cay through South Water Cut (Figure 2). Strong winds precluded surveying exposed areas, and, by necessity, surveys had to be conducted in nearby protected locations. The inability to travel appreciable distances limited the thoroughness of the survey. In general, the Carrie Bow Cay vicinity did not appear to be a productive area for commercial sponges. However, one of the authors (Stevely) has observed the presence of the reef sponge (Spongia obliqua) in patch reef areas located between the mainland and barrier reef during field work on a different project. The reef sponge is generally not considered to be of sufficient quality to support commercial harvest. Ambergris Cay No commercial sponges were found. Based on conversations with local fishing guides, five potential sponge habitat areas in the vicinity of Ambergris Cay were surveyed. These included: north of San Pedro, south of San Pedro, offshore of Laguna de Boca Ciega, seagrass beds near Congrejo Cay, and Cayo Romero (Figure 3). Survey sites north of San Pedro, south of San Pedro, and offshore of Laguna de Boca Ciega were "hard bottom" habitats with numerous "loggerhead" sponges (Spheciospongia vesparium), and represented habitat in which commercial sponges are sometimes found in Florida. . Tobacco E 2" Tobacco Sr daGay, Coco Plum: Cay Tobacco Reef Blue Grounds): eae i Range: o¢ Sp fae x . 8 — >. South Water Cay Ee OS Se 8 ut . ; : CARRIE BOW CAY Figure 2. Carrie Bow Cay commercial sponge survey sites. Key: G= survey site Laguna de Boca Ciegat y Boca Chica Congrejo Cay / Barrier } Reef J CARIBBEAN SEA Figure 3. Ambergris Cay survey sites. A - north of San Pedro, B - south of San Pedro, C - offshore of Laguna de Boca Ciega, D - seagrass beds near Congrejo Cay, E - Cayo Romero. Key: f = survey site 8 Turneffe Island A total of 9 areas within the Turneffe Islands lagoon were surveyed, but velvet sponges, Hippospongia gossypina, were found only in one location, the "Crooked Creek" area (Figure 4). Velvet sponges were abundant in water 3-6 ft (1-2 m) deep, 400-500 ft (400-500 m) north of the Crooked Creek entrance into the Turneffe Islands lagoon (a tidal cut between mangrove islands). Ircinia sp. and Spheciospongia vesparium, were also present at this location. Additional field observations made on October 19, 1991, as part of a different survey project, documented the presence of velvet sponges in other areas within the Turneffe Island lagoon. These areas included: the eastern shoreline of Soldier Bight and shallow waters near mangrove islands and mangrove shoreline in the vicinity of the western opening of Grand Bogue Creek into the main lagoonal area. The October I9th, 1991 observations did not include measuring the sponges or recording quantifiable data on whether the sponges were attached to the substrate. Data on maximum and minimum sponge diameter and type of attachment to the substrate was recorded for a total of |5 velvet sponges (Table 1). The velvet sponges were growing on a mangrove/seagrass peat substrate. A considerable percentage of these sponges had broken free from their attachment to the substrate, with only 53% found growing attached to the substrate. A species of Spongia, probably a variety of sponge that would commonly be referred to as "yellow sponge" in the commercial trade (Spongia barbara, sensu de Laubenfels and Storr, 1958), was found in patch reef habitat to the west of Douglas Cay (outside the lagoon environment). This sponge did not appear to have significant value for commercial trade. Unfortunately, the collected specimens were lost in the process of having the specimens shipped. Evaluation of Commercial Sponge Fishery Potential A commercially valuable grade of velvet sponge (Hippospongia gossypina) was found at Turneffe Islands (Figure 5). After evaluating a sample of I3 Turneffe Island sponges, tarpon Springs, Florida sponge buyers indicated that they would pay US $2.00-3.00 per sponge. The velvet sponge quality was such that it would be considered to be somewhat inferior to the sheepswool sponge (Hippospongia lachne), but would have a market value greater than that of other commercial sponge varieties. Velvet sponges tear more easily than sheepswool sponges, partially because of the characteristic presence of large pseudoscula or vents on the upper surface (Moore, 1910b). Based on our experience with sponge fisheries in Florida and the Bahamas, we estimate that the value and quantity of sponges at Turneffe Islands were sufficient to support sponge fishing activities. Field observations indicated that it would not be unreasonable for a 2-man fishing team to produce in excess of 100 sponges per day. For example, we collected 15 sponges in | hour, and this included time taken to measure each sponge and record data. However, there are several important factors and limitations which must be considered before the fishery potential can be properly evaluated. These are considered in detail in the following discussion section. Norierit Ho Ne ; Turneffe Flats Resort Crickozeen Creek & Cc ay £ Z RESEARCH = BASE CAMP244 fCrooked “= Creek =x}. Ambergris £ Joe’s Hole : ¢ & RIVA’S KEY \- SPONGE . FARM RELICT Blue Creek © RESEARCH BASE CAMP Grand Point “2 “ vo Turneffe Island Lodge Figure 4. Turneffe Islands survey sites and approximate location of the sponge farming area (stipled area). 10 Table 1. Maximum and minimum sponge diameters and category of attachment to the substrate for 15 velvet sponges (Hippospongia gossypina) collected at Turneffe Islands, Belize, May 24, 1989. Maximum Diameter Minimum Diameter Attachment to in (cm) in (cm) Substrate 10.8 (27) 9.6 (24) 11.2 (28) 10.8 (27) 10.8 (271 8.8 (22) ps 48.6000) 5]. n0.otze1 7/7 Tineteechoa aa ee eee ee Mean Maximum Minimum Percent Diameter 10.6 (27) 9.6 (24) Attached 53% Location of the commercial sponge farm Information provided by local fishermen (Mr. Joseph Garbutt, Mr. Carl Carbal) indicated the approximate location of sponge planting areas in the Turneffe Islands lagoon (Figure 4). The approximate location of the sponge planting area in the southern portion of the lagoon was verified by finding the concrete disks used to "plant" sponges (Figure 6) on the western side of Riva’s Cay, approximately 200 m south of the northern tip of the island (Figure 4). The disks found along the mangrove shorline and in the water immediately adjacent to the shoreline. This site marked the location where sponge farm workers docked a live-aboard boat. No navigation charts that located Rivas Cay could be found. However, the approximate location, as determined by triangulation on the open water is shown in Figure 4. Smith (1941) did not describe the precise location of the sponge farm but his figure showing the 11 Figure 5. Live velvet sponge (Hippospongia gossypina) removed from water. Note pseudoscula described by Moore, I9IOb. Figure 6. Concrete disks used to "plant" sponge cuttings at the Turneffe Islands sponge farm. 12 progression of sponge disease throughout the large planting areas is consistent with our location of the sponge planting areas. The area where sponges were planted was extensively surveyed, but no concrete disks were found. Wood stakes were found which the local guide (Joseph Garbutt) claimed were marking the sponge planting area. These wood stakes were obviously much older than stakes used to mark lobster traps (they were covered with an extensive growth of fire coral). Although all of Mr. Garbutt’s comments during the expedition proved to be accurate, the verification of these stakes as artifacts from the sponge farm was impossible. DISCUSSION Inclement weather (sustained 25-30 mph winds) and failure of a local guide to provide contracted services (a new Turneffe Islands expedition team had to be organized) reduced opportunities to conduct more extensive surveys. These factors prevented the collection of quantitative data on sponge density and population structure. However, two significant goals were realized: useful ecological and fishery information was collected on the velvet sponge, Hippospongia gossypina, in the marine waters of Belize and the location of the Turneffe Islands sponge farm was documented. These observations increase the information available for understanding and managing the marine biological and fishery resources of Belize. Distribution and Abundance of Commercial Sponges Within the Turneffe Island lagoon (the only location where commercially valuable sponges were found), velvet sponge were found growing on a mangrove peat substrate. Although the mangrove peat provides a substrate for sponge attachment (Figure 7), it crumbles easily and is sufficiently soft to allow the establishment of the seagrass Thallassia testudinum. Our observations indicate that, although this substrate is adequate for the attachment of sponges, many of the sponges eventually break loose (Table 1). Sponges that break loose from the substrate but continue to survive and grow are commonly called "rollers" in the commercial trade (Figure 8). The percentage of rollers was much higher than the percentage observed in the Florida commercial sponge harvest (Stevely, pers. obs). The occurrence of commercial sponge growing on a non-rock substrate was a new observation for us. In our extensive field observations in both the Gulf of Mexico and the Bahamas, we have found commercial sponge species (Hippospongia and Spongia growing either attached to rock outcroppings or to any suitable hard surface (coral/rock fragments, gorgonians, etc). De Laubenfels (1948) reported that sheepswool and velvet sponge were common on these substrates, often in precisely the same localities. Although the previously noted additional field observations taken on October 19, 1991 were extremely cursory, they did confirm the observation that the velvet sponge was found growing on mangrove peat/seagrass substrate. Also, there was a general impression that the percentage of "roller" sponges varied considerably from site to site. In some areas it seemed that most sponges were found growing attached to the substrate. In the Soldier Bight area it appeared as if almost all the velvet sponges were rollers and that they had accumulated along the eastern shoreline as a result of Gere ss tere ee ee Figure 7. Figure 8. 13 Live velvet sponge (Hippospongia gossypina) removed from water to show growth habit of attachment to mangrove peat substrate. Live velvet sponge (Hippospongia gossypina) removed from water to show growth habit of "roller sponge". Former point of attachment to substrate is now covered by ectosome. 14 tidal and/or wind driven currents. However, these observations need to be verified by more detailed field work. Smith (1941), in describing the Turneffe Islands lagoon sediments, stated that "the floor of the lagoon consists of calcarious mud, with admixed shell and coral sand predominate near the eastern entrances, and with organic matter formed from the detritus of eel-grass and mangrove roots present in varying degree throughout the lagoon". Mangrove/seagrass peat substrate was observed in several areas in the Crooked Creek and Chickozeen vicinity. The depth of the mangrove peat appeared to be considerable. On one occasion crevices were seen in the mangrove peat which extended 6-8 ft (2-2.5 m) in depth. Observations made while snorkeling Crooked Creek (a channel between mangrove islands carved boy tidal currents) indicate the depth of the mangrove peat deposits extended to at least 30 ft (10 m). These observations suggest that the Turneffe Islands represent an atoll formation consisting of mangrove peat and suggest that mangrove peat formation has kept pace with seamount subsidence/and or sea level rise. Prior to the 1938-39 sponge mortality, the velvet sponge was known for its commercial value and was considered to be the most valuable commercial sponge after the sheepswool sponge (Moore, 1910b). Although it was considered to be less compressible, absorbent, and durable compared to the sheepswool sponge, velvet sponges from some area (e.g., the Bahamas) were regarded as almost equivalent in quality to the sheepswool sponge. Moore (1910b) reported the velvet sponges were found in the straits of Florida, the Caribbean Sea, and the Bahamas. De Laubenfels and Storr (1958) stated that velvet sponges had been common around Florida and the west Indies. In Florida, velvet sponge was harvested from the fishing grounds between Key West and Cape Florida (Moore, !9I0b) in living coral areas at depths of 3-25 ft (1-8 m) (Storr, 1964). Florida sponge fishermen produced 8,000 Ib (3,600 m) of velvet sponge in 1899 (Moore, I9I0b). Although velvet sponges were sufficiently abundant to sometimes be reported in Florida commercial sponge landings, they were the least abundant of the commercial sponges (Smith, 1898). The best quality of velvet sponge was regarded to be from the Bahamas (Moore, 19IOb), and at one time it was the principal commercial sponge of the Bahamas (de Laubenfels and Storr, 1958). Moore (I9I0b) noted some commercial sponge production from the British Honduras, including velvet sponge. Moore also reported that sheepswool, velvet and grass sponges (Spongia sp.) were found along the entire coast of British Honduras, in the shallow waters about the numerous islands, rocks, and banks, and that many commercial sponge varieties grow "attached to staghorn corals and gorogonians". No mention was made of either sponges at Turneffe Islands or velvet sponges found growing attached to mangrove peat substrate. Cresswell (1935) mentioned an effort by commercial sponge fishermen in 1895 to explore the waters of British Honduras and reported that velvet sponge was the most common sponge harvested. After the devastating sponge mortality, the velvet sponge was thought to be essentially extinct in areas that had been known to produce commercial quantities, although many years later a few were reported from Cuba (de Laubenfels and Storr, 1958; Storr 1964). An extensive survey of the Florida sponge grounds in 1947 and 1948 15 was conducted to evaluate the condition of the sponge grounds following the effects of the sponge disease and overfishing during the early 1940’s. The resulting fishery report did not report the occurrence of any velvet sponge (Dawson and Smith, 1953). However, later taxonomic study of the sponges collected during the survey reported a velvet sponge specimen collected from a station in the northern Gulf of Mexico (de Laubenfels, 1953). Storr (1964) stated that the velvet sponge had not been reported in the Bahamas since the disease. In 1975 a report on the Bahamian sponge fishery indicated that the fishery was based on the harvest of sheepswool and grass sponge and contained no reference to velvet sponge (Thompson, unpubls. mans.). Wiedenmayer (1977) surveyed the shallow-water sponges of the western Bahamas and reported eighty-two sponge species, including 3 commercial species of the genus Spongia but no velvet sponge (his study was not intended as a survey of commercial varieties). Repeated communication with Florida sponge fishermen and sponge buyers has failed to indicate even the rare occurrence of the velvet sponge. The effects of disease and overfishing were apparently sufficiently severe to drastically reduce velvet sponge abundance throughout a major portion of its geographic range. In view of the long-term change in distribution and abundance of the velvet sponge and a lack of knowledge of its current distribution, the apparently healthy population at Turneffe Islands is worthy of note to fishery managers and scientists. The Turneffe Island’s velvet sponge population may represent a relatively small genetically isolated population. Storr (1964) indicated that the sheepswool sponge (Hippospongia lachne) larval state is short-lived (I-2 days) and does not have strong swimming capabilities. The Turneffe Islands atoll is separated from other shallow water habitats by waters at least 250 fathoms in depth. Prevailing surface current patterns (Hartshorn et al., 1984: Figure 9, this report) support the idea that larval recruitment (i.e., genetic exchange) may be limited between Turneffe Islands and other shallow water Caribbean sponge populations. If the prevailing surface currents depicted in Figure 9 accurately indicate a likely transport mechanism for velvet sponge larvae, then the larvae would have to traverse open Caribbean Sea waters in a relatively short time. The ability of the velvet sponge to grow attached to a mangrove peat substrate, and possibly the ability to utilize a food source enriched with detrital particles may help to explain to future investigators subtle differences and commonalities in the ecological niches occupied by commercial sponge species. Although speculative, field observations may suggest that seagrass/mangrove derived detrital particles may contribute to velvet sponge nutrition. On windy days, waves and currents agitated the shallow water where the velvet sponges were found, and, in effect, produced a suspension of detrital material consisting of decaying seagrass leaves and eroding mangrove peat. The detrital suspension was sufficient to noticeably reduce water clarity in mangrove peat areas exposed to windy conditions. Sponges are efficient filter feeders, evidently capable of filtering bacteria size food particles. It is possible that fine detrital particles in the water column could either directly or indirectly, by promoting bacterial growth or elevating the level of dissolved organic, contribute to nutritional intake. Lauckner (1980) reviewed data that suggest the presence of symbiotic bacteria associated with several sponge genera, including Spongia and Hippospongia, which possibly could assist the sponge in utilizing dissolved organic substances. 16 Belize 30 mi AO sake d anette } "y Pe Sab rl (teathe “ky / ighthouse Figure 9. Prevailing surface currents for waters of Belize (adapted from Hartshorn et al., 1984). 17 Evaluation of Commercial Sponge Fishery Potential Although a commercially valuable grade of velvet sponge was found to occur at a density capable of supporting fishery activities, several important factors must be considered before the fishery potential can be properly evaluated. Most importantly, the extent of sponge producing areas within the Turneffe Islands lagoon must be assessed more thoroughly. Although the additional field work conducted on October 19,1991 found more areas that supported velvet sponge growth, the areas were limited in size and might be easily depleted by continuous intensive fishing effort. The comments by Moore (I9lOb) indicating the presence of several varieties of commercial sponges from the coast of British Honduras suggest that additional surveys are required to more fully define the commercial sponge resources of Belize. The unique substrate to which the Turneffe Islands sponges are attached may not be conductive to allowing long-term harvest. Sponge fisheries in Florida and the Bahamas are supported by harvesting sponges that grow attached principally to carbonate rock outcroppings. Field observations in the Bahamas (Stevely, pers. obs.) and experimental work in the Florida Keys (Stevely and Sweat, 1985) indicate that when sponges are either cut or torn from such substrate, a significant quantity of sponge tissue sometimes remains attached to the substrate. This tissue is capable of regenerating to produce another viable sponge. In Florida, survival of harvested sponges ranged from 30% for sponges torn free using a sponge hook to 70% for sponges cut free with a knife. In Turneffe Islands, where sponges are anchored only to mangrove peat, it is likely that all the sponge tissue would be taken when the sponges were harvested by either hooking or cutting and that no attached sponge tissue would remain for regeneration. The same would be true for harvested roller sponges. The large size of the Turneffe Islands velvet sponges (Table 1) may suggest that the legal size for sheepswool sponge required by either Florida (5 in, 12.5 cm, minimum diameter) or Bahamian law (7 in, 17.5 cm, minimum diameter) would provide little protection for the resource if sponge harvesting was economically feasible. Only one of the |5 sponges measured (Table 1) would have been protected by a law requiring a minimum harvest size of 5 in (12.5 cm). However, may more sponges from several areas should be measured before fishery management regulations are suggested. Also, consideration must be given to managing a regularly harvested resource compared to harvesting a virgin stock. A conservative approach that insures adequate protection of the resource should be taken in managing Turneffe Islands sponge fishery development until more complete information is available. Establishing an enforceable minimum legal size of 8-9 in (20-22.5 cm) exemplifies such a conservative approach. Conservative management of the Turneffe Islands sponge fishery requires protection of reproductive stocks. Historically, research on reproduction in commercial sponges has focused on the more valuable sheepswool sponge, and essentially no information on velvet sponge biology is available. The sheepswool sponge attains reproductive maturity at a size of from 3 in (7.5 cm) in the Florida Keys to 5.5 in (14 cm) in the northernmost Florida west coast sponge grounds; reproductive maturity is attained at a smaller size in the warmer portions of its geographic range (Storr, 1964). Assuming a similar trend in the size of maturation in the velvet sponge it is reasonable 18 to assume that a minimum size of 8 or 9 inches (20-22.5 cm) in Belizean waters would protect sponges capable of a significant contribution to larval production. The physical remoteness of Turneffe Islands presents transportation problems for sponge fishermen. Most likely, it would be necessary to store sponges accumulated for several weeks before transport to the mainland. Fortunately, cured and dried sponges do not require refrigeration and could be collected over weeks or months while intermittently pursuing other fishing activities. Storage of cleaned sponges would require some shelter to prevent rotting and long-term exposure to the sun. Cleaned sponges are also lightweight and can be easily transported by small boats. Proper cleaning of sponges is hard work, but it is critical for receiving top price; improperly or incompletely cleaned sponge are either worthless or are worth only a fraction of their true value. In addition to time allotted for sponge harvesting, the sponge fishermen must commit an equal proportion of effort to cleaning, storing, and transporting the catch. Conversations with fishery cooperative manager indicated some lack of interest in exporting sponges for two reasons: they were unfamiliar with the current value of commercial sponge in the export market and they did not know whether sponges were sufficiently abundant to support fishery development. The fishery cooperatives are the obvious focal points for collecting and exporting sponges. Belizean fishery cooperatives routinely ship seafood to Florida. Sponges are a highly valuable commodity (e.g., US $20.00-50.00/lb; $44.00-110.00/kg), and reasonable shipping costs (e.g. US $l.00-2.00/lb; $2.00-4.00/kg) would not be prohibitively expensive. Thus, sponge exports may significantly increase the cash flow and profitability of fishery cooperatives. For example, a 5,000 Ib (2250 kg) annual shipment could easily result in an annual cash flow of $100,000 based on an estimated minimum price of US $20.00/lb ($44.00/kg) for velvet sponge. If a reliable supply of quality velvet sponges was established, it is reasonable to expect that the price paid for sponge to further increase. Location of the Commercial Sponge Farm A significant amount of historical information exists for sponge culture work in the Florida Keys, Bahamas, and Pacific Ocean (Stevely et. al., 1978). However, only sparse notes in the literature, referring principally to the occurrence of sponge disease and briefly describing sponge farm operations and location are available for the Turneffe Islands sponge farm (Smith, 1941). Smith (1941) stated, "Cultivation consists of cutting the sponge into small pieces, attaching these to stone or cement disks and allowing them to grow to market size on areas of the lagoon bottom most favorable to fast and healthy development". In the Bahamas, a length of palmetto string made from splitting a palmetto palm leaf was used to tie the cut sponge pieces to the concrete disk (Storr, 1964). One of the sponge disks found during our investigation still had a piece of aluminum wire attached through the small hole in the disk, suggesting that aluminum wire was at least sometimes used to attach sponge cuttings at Turneffe Islands. Aluminum wire also was used in sponge farming attempts in the Florida Keys (Stevely, et. al., 1978). The Turneffe Islands concrete sponge planting disks (Figure 6) 19 were similar in appearance to those used in both the Bahamas (Storr, 1964; Figure 9) and Florida Keys (Stevely et al., 1978, Figure I5). Smith (1941) stated that the farm was run by concessionaires (Messers. R.E. Foote and H.T. Grant) licensed by the British Honduras Government. At the time of the sponge disease mortality, approximately 700,000 sponges were under cultivation (225,000sheepswool sponges, and 475,000velvet sponges). Mortality of the densely planted sponge cuttings (in some places one per square meter) was estimated to be 95%. Some of the geographical advantages of attempting sponge farming in the Turneffe Islands lagoon are readily apparent. The lagoon contains extensive shallow areas 3-I2 ft (1-4 m) deep that are reasonably protected by mangrove islands. Sheltered and relatively clear waters would permit sponge farm operations to proceed in all but the most severe weather conditions. The remoteness of the Turneffe Islands would probably assist in enforcing security of the farm. Difficulty with protecting sponge plantings from theft has been a major problem for sponge farming attempts in many areas (Stevely et al, 1978). However, Turneffe Islands sponge farm workers living in the immediate vicinity of the farm could serve as security guards in the area. Potential thieves interested in stealing sponges from the farm would have to establish a camp for harvesting, cleaning, and storing a sufficient number of sponges to justify transport back to mainland. In general, it would be difficult for a potential sponge poacher to escape notice in these remote surroundings populated principally with sponge farm workers. Finally, the abundance of "wild" sponge stock in the lagoon may have played a key role in the decision to attempt sponge farming. The natural sponge populations may have been insufficient to support a fishery harvesting hundreds-of-thousands of sponges, but capable of producing tens-of-thousands of "seed" sponges for ‘ propagation. LITERATURE CITED Dawson, C.E., and F.G. Walton Smith 1953. The Gulf of Mexico sponge investigation. State of Florida Board of Conservation Technical Series No. 1. Marine Laboratory, Univ. of Miami, FL 28p. Cresswell, E.J, 1935. Sponges: their nature, history, modes of fishing, varieties, cultivation, etc. Pitmen’s common Commodities and Industries Series. Sir Isaac Pitman and Sons, Ltd. London. 126p. de Laubenfels, M.W. 1948. The order Keratosa of the phylum Porifera - a monographic study. Occ. Pap. Allan Hancock Fdn. (3):1-217. 20 de Laubenfels, and J.F. Storr. 1958. The taxonomy of American Commercial sponges. Bull. Mar. Sci. Gulf and Carib. 8(2):99-117. Gaino, E. and R. Pronzato. 1989. Ultrastructural evidence of bacterial damage to Spongia officinalis fibers (Porifera, Demospongiae). Dis. Aquat. Org. 6:67-74. Hartshorn, G. 1984. Belize country environmental profile. San Jose: Trejos. Hnos. Sucs. S.A. p. XVI-152. Josupeit, H. I99I. Sponges: world production and markets. INFOFISH International, Kuala Lumpura, Malaysia. I99I, No. 2, p. 21-27. Lauckner, G. 1991. . Diseases of Porifera. In: Kinne, O. (ed) Diseases of Marine animals, Vol. I, General aspects, Protozoa to Gastropoda. Wiley and Sons, Chichester, p. 139- 165. Moore, H.F. and P.S. Galtsoff. 1951. Commercial Sponges. In: Marine Products of Commerce, R.K. Tressler and J.M. Lemon eds. Rheinhold Publishing Corp. N.Y., N.Y. p. 733-751. Moore, H.F. 19l0a. A practical method of sponge culture. Bull. U.S. Bur. Fish., Vol. 28, 1908 Part |, p. 545-585. Moore, H.F. I9IOb. The commercial sponges and the sponge fisheries. Bull. U.S. Bur. Fish. Vol. 28, 1908. Part |, p. 399-511. Shubow, D. 1969. The Florida sponge industry. Master’s Thesis, Univ. Miami, Coral Gables, FL. 123 p. Smith, H.M. 1898. The Florida commercial sponge. In: Proceedings and papers of the National Fishery Congress held in Tampa, FL. Jan. 19-4, 1898. p. 225-250. Smith, F.G.W. 1941. Sponge disease in British Honduras and its transmission by water currents. Ecology 22(4): 415-421. Stevely, J.M., Thompson, and R.E. Warner. 1978. The biology and utilization of Florida’s commercial sponges. Florida Sea Grant College Program Tech. Rept. No. 8. 45p. 21 Stevely, J.M., and D. Sweat. 1985. Survival and growth of cut vs. hooked commercial sponges in the Florida Keys. Florida Sea Grant College Program Tech. Rept. No. 38. 12p. Storr, J.F. 1964. Ecology of the Gulf of Mexico commercial sponges and its relation to the fishery. U.S. Dept, Interior, Fish. Wildl. Serv., Spec. Sci. Rept. No. 466. 73p. Thompson, R.W. Unpubl. mans. Report on the sponge industry. Bahamas Ministry of Agriculture Fisheries and Local Government. 4p. Verdenal, B. and M. Verdenal. 1987. Evaluation de I’Interet Economique de la Culture d’ Esponges Commerciales sur les Cotes Meditereenes Francaises. Aquaculture, 64:9-29. Wiedenmayer, F. 1977. Shallow-water sponges of the western Bahamas. Birkhauser Verlog, Basel and Stuttgart. p. 1-287, pls. 1-43. Wilkinson, C.R. Unpubls. mans. Consultants report on the potential for commercial sponge farming in Pohnpei, Federated States of Micronesia. Prepared for Marine Resources Division, Federated States of Micronesia. FAO South Pacific Aquaculture Development Project, Suva, Fiji, 31p. ‘i, wa Mh, ho ae 1 ; ta 5 Ai { a Pet ‘ hh ; t i ane i ee ; ; aap Gk ee A, tie et JS8. fe a f ae rr AP Bets 7 fe st ni aa PaKGGe beater" wiguiage eh 30S f: or 1988, soot ‘mengoy ea. Ty f 0 Jorn maa: ely: 3 7, _ es 2 vat UE a7 pact dois, daa 5 ti Leo Pongne Mei: Syed a aN “if abeke 4 ahr sae. wie We ee : B45 a J ut WA aA Tees seit | at ava eR, ; rk Roper A * big Bak cm a on m4, Iie FM . RT Baty LET, WED The ict mUITBND “a alt WORE en Tie i 1 nih of fF 4 | , ot BS ? 4 t : fy 1 1 bel ty ; ; " c : '¢ = : el; Pa ’ : grAaghe i Vang a i MN eer R Le! Dif? —_- WOO8 ee j ’ Is oy ,d ry ae ; bse a e { Try ee i & Bint , > t 4 £ , F ‘ Eat Tei Bien S being an rn asap VEY it a i "y a, i ' LDA 7 ph Age pet IFPI TS PaO BL ee rie trig & a Sid. tu ~~ ATOLL RESEARCH BULLETIN NO. 425 SPATIAL AND TEMPORAL VARIATIONS IN GRAZING PRESSURE BY HERBIVOROUS FISHES: TOBACCO REEF, BELIZE BY PETER N. REINTHAL AND IAN G. MACINTYRE ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 5 o Y c io} o —lo) 20 = 1.0] U ) if South Water : Cay South Water Cut ve 2 Carrie Bow a “Lagoon ao S Cay lsu Reefs :/ cf Curlew Bank Curlew Cut FIGURE 1 - Index map showing location of study transect on Tobacco Reef. SPATIAL AND TEMPORAL VARIATIONS IN GRAZING PRESSURE BY HERBIVOROUS FISHES: TOBACCO REEF, BELIZE BY PETER N. REINTHAL! and IAN G. MACINTYREZ ABSTRACT Fish herbivory appears to play an important role in the pattern of macrophyte zones running parallel to the crest of the Belizean barrier reef. The experimental relocation's of reef macrophytes seaward of the reef in 1984 and 1989 revealed that algal species with herbivore-resistant strategies are predominant in the zones of strongest herbivory near the reef crest, whereas those highly susceptible to fish grazing occur well lagoonward of the reef crest. The foraging ranges of herbivorous fish are thought to depend in large part on their proximity to suitable shelter. These trends in herbivore activity were observed in both 1984 and 1989, although the grazing pressure was uniformly less for all algal species studied in 1989. These findings may be related to seasonal (May 1984 versus March 1989) or annual variations in grazing pressure, or to a general decrease in grazing pressure over the five-year study period. INTRODUCTION The patterns of abundance and distribution of coral reef macrophytes are well known to be influenced by grazing activities of herbivorous fishes (Stephenson and Searles 1960, Randall 1961, 1965, John and Pope 1973, Wanders 1977, Hay 1981a, b, Hay et al. 1983, Hay 1984, Lewis 1986, Horn 1989, Choat 1991, Hay 1991) and urchins (Ogden et al. 1973, Sammarco et al. 1974, Sammarco 1980, Lawerence and Sammarco 1982, Hay 1981a, b, 1984). However, only recently have studies been conducted on spatial variation in herbivore activity patterns and the effects of spatial patterns of herbivory on macrophyte distributions (Hay et al. 1983, Lewis 1986, Macintyre et al.. 1987, Morrison 1988; Hay 1991). Spatial heterogeneity in grazing intensity has been found to contribute to regional diversity among and within tropical reef habitats (e.g. Lewis 1986) and fishes appear to play a major role in structuring shallow water macrophyte communities (e.g. Morrison 1988). Department of Biology, Eastern Michigan University, Ypsilanti, Michigan 48197, and 2Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington D.C. 20560 Manuscript received 11 May 1994; revised 30 August 1994 FIGURE 2 — Aerial view of Tobacco Reef looking south towards South Water and Carrie Bow Cays. Note location of study transect and biogeological zones on sediment apron. (1. Coralline-Coral-Dictyota pavement; 2. Turbinaria-Sargassum rubble; 3. Laurencia-Acanthophora sand and gravel; 4. Bare sand; 5. Thalassia sand.) Emergent reef crest and deep fore-reef on left. 3 In this study we used transplanted samples of various reef macrophytes as a bioassay for an assessment of fish grazing in the backreef habitat. The grazing activity was then compared to observed patterns of macrophyte zonation and evaluated to determine relative grazing pressures on various macrophytes as a function of distance from the reef crest. Varying selective pressures from fish herbivory and substrate requirements appear to be important factors in determining the distribution and zonation patterns of reef macrophytes (Macintyre et al. 1987). Preliminary results found in 1984 (Macintyre et al. 1987) prompted us to repeat similar experiments in 1989 when we were able to replicate the treatments at three separate locations (versus one location in 1984) and provide a caged control treatment at one of the locations. This approach allowed us to look at spatial variation in grazing as both a function of distance from the reef crest and variation along the reef crest. The results raise a number of interesting questions concerning spatial and temporal variation in fish grazing activity in the back reef habitat. MATERIALS AND METHODS This study was conducted during May 1984 and March 1989 in the backreef habitat at Tobacco Reef, north of the Smithsonian Institution's field station at Carrie Bow Cay, Belize, Central America (16948' N, 88°05' W; Fig. 1). The topography, geology and zonation of the region and the floristic and fish distribution patterns are described in detail elsewhere (Rutzler and Macintyre 1982, Norris and Bucher 1982, Lewis and Wainwright 1985, Macintyre et al. 1987). A primary study site representative of the reef was identified on the sediment apron of Tobacco Reef approximately 1 km north of South Water Cay (Fig. 1). Three sites that were separated approximately 100 m (north, middle and south) were used as replicate localities. We transplanted samples of eight reef macrophytes to 0 m., 40 m., 90 m. and 150 m. from the reef crest to determine fish grazing activity as a function of distance from the reef crest. The distances were related to the observed patterns of macrophyte _ zonation and five distinct biogeological zones (Fig. 2): (1) 0 m: coralline-coral-Dictyota pavement, (2) 40 m: Turbinaria-Sargassum rubble, (3) 90 m: Laurencia-Acanthophora sand and gravel, (4) 150 m: bare sand and (5) Thalassia sand. These zonation patterns and various abiotic factors affecting macrophyte distribution are discussed elsewhere (Macintyre et al. 1987). The eight macrophytes used in the grazing assay were Turbinaria turbinata, Sargassum polyceratium, Thalassia testudinum, Padina jamaicensis, Acanthophora spicifera, Laurencia papillosa, Laurencia intricata and Dictyota sp. The T. testudinum blades and other species were free of epiphytes as were all macrophytes. These macrophytes were chosen because they are common and abundant members of the backreef community and show variation in distribution patterns. A mixed assay allowed us to determine grazing intensity with respect to the variety of herbivore groups observed to be active in this area. For example, previous studies (Lewis 1985) have shown that T. turbinata, S. polyceratium, T. testudinum and P. jamaicensis are preferred and primarily consumed by parrotfish (Sparisoma and Scarus spp.). In contrast the red algal species, A. spicifera, L. papillosa and L. intricata, are preferred and primarily consumed by acanthurid species (Acanthurus bahianus, A. coeruleus and A. chirugus) (Lewis 1985). Another herbivore, Diadema antillarum, was not considered in this study because it had undergone a massive die-off throughout the Caribbean (Lessios et al. 1983) and was never common on this back-reef sediment apron of Tobacco Reef. Moreover, fish are considered to be the grazers of prominent FIGURE 3 - Before and after photographs of test set of four macrophytes (left to right: Turbinaria turbinata, Sargassum polyceratium, Padina jamaicensis, and Thalassia testudinum) at reef crest site. 5 importance on many tropical reefs (Hay 1984, Lewis 1986, Morrison 1988). No other organisms were found grazing on any of the assays. The pattern of herbivore pressure as a function of distance from the reef crest was measured by placing complete spin dried and weighed sets of the assay in the back reef at distances of 0, 40, 90 and 150 m behind the reef crest. Samples of each of the macrophyte species were held in wooden clothespins attached to metal rods (Fig. 3). All trials were run for 8 h from 0830 to 1630 h at each experimental grazing site. The samples were then spin dried and reweighed. In 1984, four trials were conducted at each distance at one location each on different days. The amount of macrophyte used in the assays was consistent with the size of macrophytes observed in the study area. In all trials a small piece of the macrophyte remained in the clothespins where it was inaccessible to herbivores. The remaining macrophytes, grazing scars and field observations indicated that macrophytes were not lost by wave action. In 1989 the trials were replicated on different days three times at each of three locations approximately 100 m apart for all four distances from the reef crest. The middle site was approximately at the same location as that used in 1984. At the northern most locality, 1 m3 cages constructed from 1/4" wire mesh were placed at each distance from the reef crest. Grazing assays with the eight macrophytes were placed in the cages as controls for the amount of macrophyte lost due to wave action when fish grazers are excluded. Special care was taken to avoid placing macrophyte assays near the guarded territories of damselfish. Fish counts were made in 1984 along a series of four replicated visual line transect censuses (50 m x 2 m) between 0800 and 1600 h at each experimental grazing site. All parrotfish (Scarus and Sparisoma spp.) and surgeonfish (Acanthuridae) were counted to determine distribution patterns and densities. The methodology applied here is similar to that used by Lewis and Wainwright (1985) in their study of distributions of herbivores among reef habitats. RESULTS 1984 Experiments In 1984, a distinct fish-grazing pattern emerges from the data on percent weight loss for the macrophyte species used in this study. Average percent weight loss generally decreases with distance from the reef crest for seven of the macrophytes studied (the exception being Dictyota sp.) and shows highly significant weight-loss differences with respect to distance from the reef crest (ANOVA, P < 0.005 in all cases) (Table 1). For all macrophytes except Dictyota divaricata, grazing pressure was homogeneously heavy in 1984 at 0 and 40 m from the reef crest as indicated by Duncan multiple range tests (Table 1): the average percent weight loss at 0 and 40 m are included in the same homogeneous subset whose highest and lowest means do not differ by more than the shortest significant (.05 level) range for a subset of that size. For two macrophyte species, Padina jamaicensis and Thalassia testudinum, the homogeneity of heavy grazing pressure extends to 90 m. These two macrophyte species are highly preferred by parrotfish in field feeding trials (Lewis 1985, Lewis and Reinthal in prep.). Turbinaria turbinata and Sargassum polyceratium, also preferred by parrotfish, appear to be subject to the same heavy grazing pressure at 0 and 40 m, but little or moderate grazing pressure at 90 and 150 m. The situation was similar for the macrophytes preferred by surgeonfish (Acanthuridae) (Lewis 1985, Lewis and Reinthal, in prep.). The grazing pressure on Acanthophora spicifera and Laurencia papillosa is homogeneously heavy at 0 and 40 m, moderate at 90 m, and light at 150 m as indicated by the homogeneous subsets in the Duncan's Multiple range test (Table 1). Heavy grazing pressure is also observed at 0 and 40 m for L. intricata but grazing is homogeneously moderate to weak at 90-150 m. In the case of Dictyota divaricata sp., grazing pressure was homogeneously light and no significant differences were found between average weight loss at various distances (ANOVA, P = 0.68) (Table 1). The Duncan's Multiple range test included all distances in the same homogeneous subset. The weight loss in this case is thought to be due primarily to experimental error caused by the removal of small fragments through wave action and this was the only macrophyte that proved difficult to retain as one piece in a clothespin. Distribution patterns identified for herbivorous fish (Table 2) coincide with grazing patterns identified here. Fish densities were highest at 0 and 40 m, the areas with the greatest grazing pressure. Even though most of the fish observed at 90 m were juvenile Acanthurus bahianus, the sighting of a large heterospecific school of adult herbivorous fish outside the transect area (approximately 60 A. bahianus, 10 A. chirugus, 4A. coeruleus and 13 Sparisoma chrysopterum) indicates these fish do extend their foraging to 90 m. Many heterospecific schools were also observed at 0 and 40 m. No schools were sighted at 150 m and only 1 individual was counted in the transects. 1989 Experiments There were no significant differences between the three replicate locations for all eight macrophytes nor were there any significant distance*location interaction effects (ANOVA p > .05 in all cases). Thus, the three locations were grouped together for purposes of analyses. The caged treatment showed no significant distance effects for macrophyte loss due to wave action (ANOVA p > .05 in all cases) and all distances were included in the same homogeneous subset (Table 1). For the experimental assays, a similar pattern of spatial variation in grazing intensity emerges with respect to distance from the reef crest as that found in 1984 (Table 1) but grazing pressure was uniformly weaker. At all distances, for all macrophyte species, the percent weight loss was less in 1989 than that found in 1984 (ANOVA; p<.01). For six of the eight macrophytes there was a significant difference between the distances (ANOVA; p < .0005; Table 1). Only Turbinaria turbinata and Dictyota sp. showed no significant differences between distances. The grazing intensity observed in 1989 appears to be uniformly less than that found for all macrophytes in 1984. The Duncan's multiple range tests for 1989 show the same grazing pattern but a fairly consistent shift in the grazing pressure toward the reef crest and in no cases was grazing pressure heavier nor further extended than that found in 1984 (Table 1). DISCUSSION Spatial Patterns of Herbivory The data presented here show distinct spatial grazing patterns on the macrophyte species under study. For all macrophyte species, except Dictyota divaricata, the grazing pressure in 1984 is heavy at 0 and 40 m from the reef crest, moderate to light at 90 m and absent at 150 m. The same general results were found in 1989 but the grazing pressure was uniformly less. These results are associated with zonation patterns and biogeologic zones observed for this backreef habitat and the distribution of _ biological assemblages appears to be controlled mainly by a combination of grazing 7 pressure of herbivorous fish, which is a function of distance from the reef crest, and both physical factors and hydrodynamic conditions which are discussed elsewhere (Macintyre et al. 1987). Thus, while spatial heterogeneity in herbivore grazing represents a biotically generated mechanism contributing to high regional diversity among reef habitats (Lewis 1986, Hay 1981a, 1985; Morrison 1988), the patterns of zonation and habitat diversity within the Tobacco Reef backreef appear to be under a similar mechanistic control. Dictyota spp. in general, probably because they contain noxious chemical compounds (Gerwick 1981, Norris and Fenical 1982), are notably avoided by herbivorous fish (Montgomery 1980, Hay 1981a, Littler et al. 1983). Surgeonfish were occasionally observed taking single bites of D. divaricata but would not continue to feed or graze with the same intensity as they did on Laurencia papillosa or Acanthophora spicifera. The grazing pressure may be determining the zonation pattern through either direct effects or an indirect effect by inhibiting an efficient competitor. Lewis (1986) directly compared grazed backreef versus ungrazed backreef and found that Sargassium, Padina and Turbinaria did increase significantly in the absence of fish herbivory and outcompeted corals and other slow-growing benthic species. Hay (1981a, b) and Hay et al. (1983) showed that algal species characteristic of the deep sand plains and intertidal reef flats may be restricted from the reef slope by herbivory, and suggested that these species would represent potentially dominant competitors on the reef slope in the absence of herbivory. The same may be true for the Tobacco Reef backreef habitat. Macroalgae species found at the greatest distances from the reef crest and found in low-herbivory habitats were found to be the macrophytes most highly susceptible to grazing by herbivorous fishes. Other studies have indicated that many algal species characteristic of habitats with high grazing intensities are resistant to herbivorous fish grazing and, conversely, algae in habitats with low grazing intensity are susceptible to herbivory (Hay 1981b, 1984, Littler et al. 1983, Lewis 1985, 1986). For example, the coralline-coral-Dictyota zone (Macintyre et al. 1987) is the zone in which we measured the highest levels of herbivory. The dominant biota have well-documented herbivore . defense strategies (Norris and Fennical 1982, Paul and Hay 1986). Fish grazing intensity has been found to decrease with depth on forereefs and exhibits an inverse relationship to algal abundance (Morrison 1988). The decreasing herbivory is thought to be the result of diminishing trophic carrying capacity and increased risk of predation (Hay and Goertemiller 1983, Steneck 1983, Lewis 1986). The latter may well apply to the habitat studied here but the argument concerning trophic carrying capacity is not applicable in the shallow back-reef. Proximity to suitable shelter has long been recognized as a critical factor in determining herbivore foraging ranges on patch reefs in tropical seagrass beds (Randall 1965), on deep sand plains adjoining reef slopes (Earle 1972, Hay 1981a, b) and in the backreef areas remote from the reef crest (Lewis 1986, Lewis and Wainwright 1985). Although we have no experimental evidence to explain why herbivorous fish do not graze further from the reef crest, they may be constrained by predatory piscivorous fishes (Ogden et al. 1973) and birds. Barracuda (Sphyraenidae), jacks (Carangidae) and snapper (Lutjanidae) were often seen swimming in the study area. This might also explain why only heterospecific schools of fishes, not solitary adult individuals, were seen at the 90 m site. Heterospecific schooling is considered to provide predator avoidance advantages to participants (see Morse 1977 for review). All three study areas for 1989 showed the same pattern of herbivory and no significant differences were found between the different locations for all macrophytes. Thus, while the data clearly indicates that herbivore pressure varies perpendicular to the reef crest, it is also consistent along the reef at any one distance from the crest. Temporal Patterns of Herbivory In 1989, the grazing pressure for all macrophytes was homogeneously less than the grazing pressure found in 1984. These differences could represent seasonal fluctuations in herbivory pressure or resource availability (May versus March) or longer term patterns of variation (1984 versus 1989). The seasonal patterns could be the result of differences in productivity or resource availability to the herbivorous fish. Unfortunately, seasonal variation in grazing pressure or herbivore diets has been largely ignored and warrants further research. There may also be seasonal differences in fish movement patterns or community composition that may influence herbivory patterns. Alternatively the long term effects could be the result of the Diadema die-off (Lessios et al. 1983) such that there is less competition for resources among herbivores. Morrison (1988) demonstrated that fish grazing intensity tripled after the Diadema die-off. Thus the increased herbivore pressure in 1984 could be the fish responding to the lack of a competition and a new equilibrium is established by 1989. An alternative explanation could be that in 1984, the macrophytes had not yet responded to the overall decrease in grazing and the fish were showing the same grazing pattern had Diadema been present. By 1989 the fish were no longer required to venture as far into the backreef to graze on macrophytes since they were no longer competing with Diadema. Fish grazing intensity has also been found to vary between reefs because of heavy fishing pressure (Hay 1984). Thus the decrease in grazing pressure from 1984 to 1989 could also be the result of a decrease in the fish populations through an increase in fishing pressure. CONCLUSION From these data we may conclude that herbivorous fish have a strong impact on the distribution patterns of various reef macrophytes and this impact is a function of the distance from the reef crest. The data also indicate that there is temporal variation in herbivory patterns. While grazing pressure did not vary at three locations along the reef crest, it appears effective in preventing the establishment of many macrophytes within 90 m of the reef crest in the backreef habitat. The herbivory patterns found here are consistent with the macrophyte distribution patterns and suggests that fish grazers are of primary importance in controlling these distributional patterns. These patterns are, in turn, directly correlated with the distance from the reef crest. ACKNOWLEDGEMENTS Special thanks go to Cindy Stackhouse, Anthony Macintyre and Allan Macintyre for their assistance in the field. Sara Lewis made helpful comments in her review of an earlier version of this manuscript. Part of this work was done while PR was a postdoctoral Research Fellow at the American Museum of Natural History. This work was supported by grants from the Caribbean Coral Reef Ecosystem Program, National Museum of Natural History, American Museum of Natural History and Exon Corporation (CCRE Contribution No. 351). REFERENCES Choat, J.H. (1991). The biology of herbivorous fishes on coral reefs. Pages 120-155 in P.F. Sale, editor. The ecology of fishes on coral reefs. Academic Press, Inc. Harcourt Brace Jovanovich, San Diego. Earle, S.A. (1972). The influence of herbivores on the marine plants of Great Lameshur Bay, with an annotated list of plants. Results of the Tektite Program: ecology of coral reef fishes. Sci. Bull. 14:17-44. Gerwick, W.H. (1981). The natural products chemistry of the Dictyotacae. Ph.D. dissertation, University of California, San Diego, Calif. Hay, M.E. (1981a). Herbivory, algal distribution, and the maintenance of between- habitat diversity on a tropical fringing reef. American Naturalist 118:520-540. Hay, M.E. (1981b). Spatial patterns and grazing intensity on a Caribbean barrier reef: herbivory and algal distribution. Aquatic Botany 11:97-109. Hay, M.E. (1984). Patterns of fish and urchin grazing on Caribbean coral reefs: are previous results typical? Ecology 65(2):446-454. Hay, M.E. (1985). Spatial patterns of herbivore impact and their importance in maintaining algal species richness. Pages 29-34 in Proceedings of the Fifth International Coral Reef Congress, Tahiti 1985. Volume 4. Antenne Museum-Ephe, Moorea, French Polynesia. Hay, M.E. (1991). Fish-seaweed interactions on coral reefs: Effects of herbivorous fishes and adaptations of their prey. Pages 96-119 in P.F. Sale, editor. The ecology of fishes on coral reefs. Academic Press, Inc. Harcourt Brace Jovanovich, San Diego. Hay, M.E., Colburn, T., Downing, D. (1983). Spatial and temporal patterns in herbivory on a Caribbean fringing reef: the effects on plant distribution. Oecologia (Berlin) 58:299-308. Hay, M.E., Goertemiller, T. (1983). Between-habitat differences in herbivore impact on Caribbean coral reefs. Pages 97-102 in M.L. Reaka, editor. The ecology of deep and shallow coral reefs. Symposia series for undersea research. Volume 1. Office of Undersea Research, National Oceanic and Atmospheric Administration, Rockville Maryland, USA. Horn, M. (1989). Biology of marine herbivorous fishes. Oceanography and Marine Biology Annual Review, 27:167-272. John, D.M., Pope, W. (1973). The fish grazing of rocky shore algae in the Gulf of Guinea. Journal of Experimental Marine Biology and Ecology 11:81-90. Lawrence, J.M., Sammarco, P.W. (1982). Effect of feeding on the environment: Echinoidea. Pages 499-519 in M. Jangoux and J.M. Lawrence (eds.), Echinoderm nutrition. A.A. Balkema Press, Rotterdam, The Netherlands. 10 Lewis, S.M. (1985). Herbivory on coral reefs: algal susceptibility to herbivorous fishes. Oecologia (Berlin) 65:370-375. Lewis, S.M. (1986). The role of herbivorous fishes in the organization of a Caribbean reef community. Ecological Monographs 56(3):183-200. Lewis, S.M., Wainwright, P.C. (1985). Herbivore abundance and grazing intensity on a Caribbean coral reef. Journal Experimental Marine Biology and Ecology 87:215-228. Lessios, H.A., Glynn, P.W., Robertson, D.R. (1983). Mass mortality of coral reef organisms. Science 222:715. Littler, M.M., Littler, D.S., Taylor, P. (1983). Algal resistance to herbivory on a Caribbean barrier reef. Coral Reefs 2:111-118. Macintyre, I.G., Graus, R.R., Reinthal, P.N., Littler, M.M., Littler, D.S. (1987). The Barrier Reef sediment apron: Tobacco Reef, Belize. Coral Reefs 6:1-12. Montgomery, W.L. (1980). The impact of non-selective grazing by the giant blue damselfish Microspathodon dorsalis on algal communities in the Gulf of California, Mexico. Bulletin Marine Science 30:290-303. Morrison, D. (1988). Comparing fish and urchin grazing in shallow and deeper reef algal communities. Ecology 69(5):1367-1382. Morse, D.H. (1977). Feeding behavior and predator avoidance in heterospecific groups. BioScience 27(5):332-338. Norris, J.N., Bucher, K.E. (1982). Marine algae and sea grasses from Carrie Bow Cay, Belize. In: Rutzler, K. & I.G. Macintyre (eds.) The Atlantic Barrier Reef ecosystem at Carrie Bow Cay, Belize. Smithsonian Institution Press, Washington, DC (Smithson. Contrib. Mar. Sci. 12:167-238). Norris, J.N., Fenical, W. (1982). Chemical defenses in tropical marine algae. In: Rutzler, K. & I.G. Macintyre (eds.) The Atlantic Barrier Reef ecosystem at Carrie Bow Cay, Belize. Smithsonian Institution Press, Washington, DC (Smithson Contrib Mar Sci 12:417-431). Ogden, J.C., Brown, R.A., Salesky, N. (1973). Grazing by the echinoid Diadema antillarum Philippi: formation of halos around West Indian patch reefs. Science 182:715-717. Paul, V.J., Hay, M.E. (1986). Seaweed susceptibility to herbivory: chemical and morphological correlates. Marine Ecology Progress Series, 33:255-264. Randall, J.E. (1961). Overgrazing of algae by herbivorous marine fishes. Ecology 42:812. Randall, J.E. (1965). Grazing effect on seagrass by herbivorous reef fish in the West Indies. Ecology 46:255-260. Rutzler, K., Macintyre, I.G. (1982). The habitat distribution and community structure of the barrier reef complex at Carrie Bow Cay, Belize. In: Rutzler, K. & 1.G. 11 Macintyre (eds.) The Atlantic Barrier Reef ecosystem at Carrie Bow Cay, Belize. 1. Structure and communities. Smithsonian Institution Press, Washington, DC (Smithson. Contrib. Mar. Sci. 12:9-45). Sammarco, P.W. (1980). Diadema and its relationship to coral spat mortality: grazing, competition, and biological disturbance. Journal of Experimental Marine Biology and Ecology 45:245-272. Sammarco, P.W., Levinton, J.S., Ogden, J.C. (1974). Grazing and control of coral reef community structure by Diadema antillarum (Philippi) (Echinodermata: Echinoidea): a preliminary study. Journal of Marine Research 32:47-53. Steneck, R.S. (1983). Quantifying herbivory on coral reefs: just scratching the surface and still biting off more than we can chew. Pages 103-111 in M.L. Reaka, editor. The ecology of deep and shallow coral reefs. Symposia series for undersea research. Volume 1. Office of Undersea Research, National Oceanic and Atmospheric Administration, Rockville Maryland, USA. Stephenson, W., Searles, R.B. (1960). Experimental studies on the ecology of intertidal environments at Heron Island. Australian Journal of Marine and Freshwater Research 11:241-267. Wanders, J.B.W. (1977). The role of benthic algae in the shallow reef of Curacao (Netherlands Antilles) II: The significance of grazing. Aquatic Botany 3:357-390. Table 1. Results of macroalgal transplant experiments: mean percentage macrophyte weight loss (+ standard deviation) after exposure to or caging from herbivores at varying distances from the reef crest in the Tobacco Reef backreef habitat for 8-h trials. Four replicate trials were conducted at one location at each distance in 1984 (N=4) and three replicate trials were conducted at each of three locations in 1989 with no differences found between locations (N=9). 1989 Cage are the results from three replicate trials placed in fish exclusion cages at one location (N=3). Groupings between distances are based on Duncan's multiple range test. SPECIES MEAN PERCENT WEIGHT LOSS (4 S.D.) ANOVA Om 40 m 90 m 150 m F P Thalassia testudinum 1984 83.4 (6.2) 76.4 (10.7) 78.7 (6. > ae) (23: ») 24.98 <.0001 | 1989" 27 7-35(67) 2 9 (40. ) 0.0 (0.0) 0.0 (0. ») 30.39 <.0001 | 1989 0.0 (0.0) 2:31:33) 0.0 (0.0) 0.0 (0. ») 1.00 <.4411 Cage _ |------------------------------------------------------------------ Padina jamaicensis 1984 an 8 (3.0) 91.5, G31) 87.8 (9. =) 11.8 (6. 8) 142.2 <.0001 1989 81.3 (29. ? 42.4 (33. ey 6.4 (6.4) 11.0 (6. i 21.17 <.0001 | 1989 10.3 (5.6) 15.0 (6.4) 14.0 (4.5) 11.3 (8. “ 0.23 <.8707 Oe Turbinaria turbinata 1984 86.4 (25.3) 61.1 (44. ? 8.9 (3.5) TUG) » 9.37 <.0018 | 1989) 21.8 (28.5) 6.9 (4.5) 5.4 (5. 2 3.8 (2.0) 2.26 <.1031 | 1989 4.3 (0.9) Sle) 4.7 (2.9) 4.0 (0. oh 0.35 <.7934 CAC lonnana ans n rama nnn ar nn a nnn Sargassum polyceratium 1984 97.3 (2.0) 61.2 (42. “ 31.2 (36.1) 10.0 (4. _ 7.40 <.0046 | 1989 sie 730; HY i 4 (5.4) 5.8 (3.6) 7.4 (4. Z 9.32 <.0002 1989 = 4.7 (3.3) 10.7 (4.6) 8.0 (6.2) 6.3 (3. i 0.65 <.6023 Cage _ |----------------------------------------------------------------- Table 1 (cont.). SPECIES MEAN PERCENT WEIGHT LOSS (z S.D.) Om Acanthophora spicifera 1984 93.3 (2.8) 1989 1989 Cage 74.8 (26.7) | ee Laurencia intricata 1984 73.6 (19.7) 1989 Cage 19.7 (7.0) Laurencia papillosa 1984 1989 Cage 93.4 (3.1) 13.0 (8.5) Dictyota sp. 1984 1989 1989 Cage 28.5 (11.7) | 40 m 89.1 (11. » 90 m 712) Gi. ) | 150 m 4.0 (6. y | ANOVA F P 87.14 <.0001 31.22 <.0001 2.79 <.1090 12.35 <.0006 8.13 <.0005 0.25 <.8621 30.20 <.0001 85.65 <.0001 0.98 <.4477 0.52 <.6778 0.28 <.8371 0.00 <.9998 Table 2. Mean fish counts (densities per 100 m2, N=4) along line transects (50 x 2 m) at the four distances (m) from the reef crest (herbivory test sites) (from Macintyre et al. 1987). The Acanthuridae juveniles (Juv) and adults (Ad) were both counted and the species represented are Bah. = A. bahianus, Coe. = A. coeruleus and Chi. = A. chirugus. All Scaridae individuals were included in counts and species represented are Ise. = S. iserti, Chr. = S. chrysopterum, Rub. = S. rubripinne, Vir. = S. viride and Rad. = S. radians ACANTHURIDAE SCARIDAE Dist. Bah. Coe? Chi. ise. Chr Rubs) Vir | Rad? (m) Juv Ad Juv Ad Juv Ad 150 0.0 0:.0---40220 OF25- 7.0 0 0 0 90 6) 3054 40,0 108.0 0.50 0O 0 0 Pt /S) 40 23 45 On 2it i313 5:00 “0250 0 0 0.25 0 14.15 0:5.22.5-1010:8 LOS) (12541-0325 0 * U.S. GOVERNMENT PRINTING OFFICE: 1994-383-509 os NEWS AND COMMENTS F. RAYMOND FOSBERG (1908-1993) A CELEBRATION On May 19 1994, just a day before what would have been Ray Fosberg's 86th birthday we held a celebration in his honor in the Learning Center of the National Museum of Natural History. Approximately 60 people participated in this commemoration of Ray's life achievements, including several members of his family. The activities began with Warren Wagner, Chairman of the Department of Botany, who welcomed everybody to this celebration of Francis Raymond Fosberg's life and introduced the Guest Speaker David R. Stoddart from the University of California at Berkeley. Along with a general biographical review of both Ray Fosberg and David Stoddart, Warren, in a lighter vein, mentioned Ray's comment in a letter to Dr. William Stearn of the British Museum, that despite the poor quality of David's initial plant collection, David, unlike "most geographers" was capable of learning how to collect botanical specimens. Warren Wagner introduces the Guest Speaker. (photo by Bill Boykins) In reviewing Ray Fosberg's career, David Stoddart made particular mention of the early stage of development of the Coral Atoll Program and the initiation of the Atoll Research Bulletin. He also talked about Ray's wide range of interests in natural history, which extended well beyond his specialty in floristic and taxonomic studies. In addition, he described Ray's profound influence on his own career, starting with those first lessons in plant collecting. Finally he discussed Ray's strong commitment to conservation, highlighted with the successful preservation of Aldabra, in the western Indian Ocean -- an island that was scheduled to be converted into a military base in the mid-1960s. David worked very closely with Ray on this crusade. David Stoddart gives the Guest Speaker's address. (photo by Bill Boykins) Ian Macintyre followed with a report on the current status of one of Ray's major legacies -- the Atoll Research Bulletin. He indicated that there has been a considerable increase in the editorial board in the hope of attracting manuscripts from a wider variety of disciplines. The new members include Steve Cairns (Invertebrate Zoology), Brian Kensley (Invertebrate Zoology), Wayne Mathis (Entomology), Victor Springer (Vertebrate Zoology), Warren Wagner (Botany), and Roger Clapp (National Biological Survey). Ian also mentioned that three Special Issues of the Bulletin are in press, including a Fosberg Commemorative Issue along with issues on Caroline Atoll and the Cocos (Keeling) Islands. With another regular issue in preparation, he suggested that Ray would be rather pleased with our progress. Ian Macintyre reports on the status of the Atoll Research Bulletin. (photo by Bill Boykins) The last speaker was Dan Nicolson, who read a few letters that he had recently received from some of Ray's former colleagues. He then went on to cite Statistics on Ray's work that he had assembled for the Fosberg Commemorative Issue. At least 51 plants were named for Ray and each year for about 60 years Ne \ —< wa Dan Nicolson cites statistics on Ray Fosberg's work. (photo by Bill Boykins) Ray named about 20 new taxa, with a total of about 1,000. His publication record was even more impressive, with an average of about 10 papers a year, resulting in 625 papers to date with still more to come. Most spectacular of his achievements was his average yearly collection of 1,000 plants, with a total of 66,369. | After thanking the many people who had worked on Ray's backlog and assisted with the organization of this event, Dan asked the audience if anybody wanted to make a statement. At least 6 people responded, including Frank Whitmore, who hired Ray to work in the US Geological Survey to work on the Pacific Geologic Mapping Program; Elbert Little, who was one of the Cinchona Mission collectors who Ray lead during the Second World War; and Lee Talbot, who talked about Ray's contribution to conservation. A member of the audience, Frank Whitmore comments on his past association with Ray Fosberg. (photo by Bill Boykins) After about an hour of talking, all of which was recorded by the Smithsonian Institution Archives staff, the participants socialized over a light serving of food and drinks. It was at this time that Ray's family mentioned that they had just spent the moming driving out to the Blue Ridge Mountains where they scattered Ray's ashes in places where he used to love visiting with them. ATOLL RESEARCH BULLETIN NOS. 415-425 NO. 415. NO. NO. NO. NO. NO. NO. NO. NO. NO. NO. 417. 418. 419. 420. 421. 422. 423. 424. 425. TIKEHAU AN ATOLL OF THE TUAMOTU ARCHIPELAGO (FRENCH POLYNESIA) PART I. ENVIRONMENT AND BIOTA OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY A. INTES AND B. CAILLART PART II. NUTRIENTS, PARTICULATE ORGANIC MATTER, AND PLANKTONIC AND BENTHIC PRODUCTION OF THE TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY C.J. CHARPY-ROUBAUD AND L. CHARPY PART III. © REEF FISH COMMUNITIES AND FISHERY YIELDS OF TIKEHAU ATOLL (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA) BY B. CAILLART, M.L. HARMELIN-VIVIEN, R. GALZIN, AND E. MORIZE COLONIZATION OF FISH LARVAE IN LAGOONS OF RANGIROA (TUAMOTU ARCHIPELAGO) AND MOOREA (SOCIETY ARCHIPELAGO) BY V. DUFOUR CAVES AND SPELEOGENESIS OF MANGAIA, COOK ISLANDS BY JOANNA C. ELLISON SHALLOW-WATER SCLERACTINIAN CORALS FROM KERMADEC ISLANDS BY VLADIMIR N. KOSMYNIN DESCRIPTION OF REEFS AND CORALS FOR THE 1988 PROTECTED AREA SURVEY OF THE NORTHERN MARSHALL ISLANDS BY JAMES E. MARAGOS QUATERNARY OOLITES IN THE INDIAN OCEAN BY C.J.R. BRAITHWAITE LARGE-SCALE, LONG-TERM MONITORING OF CARIBBEAN CORAL REEFS: SIMPLE, QUICK, INEXPENSIVE TECHNIQUES BY RICHARD B. ARONSON, PETER J. EDMUNDS, WILLIAM F. PRECHT, DIONE W. SWANSON, AND DON R. LEVITAN CHANGES IN THE COASTAL FISH COMMUNITIES FOLLOWING HURRICANE HUGO IN GUADELOPE ISLAND (FRENCH WEST INDIES) BY CLAUDE BOUCHON, YOLANDE BOUCHON-NAVARO, AND MAX LOUIS THE SIAN KA’AN BIOSPHERE RESERVE CORAL REEF SYSTEM, YUCATAN PENINSULA, MEXICO BY ERIC JORDAN-DAHLGREN, EDUARDO MARTIN-CHAVEZ, MARTIN SANCHEZ- SEGURA, AND ALEJANDRO GONZALEZ DE LA PARRA A PRELIMINARY EVALUATION OF THE COMMERCIAL SPONGE RESOURCES OF BELIZE WITH REFERENCE TO THE LOCATION OF THE TURNEFFE ISLANDS SPONGE FARM BY J.M. STEVELY AND D.E. SWEAT SPATIAL AND TEMPORAL VARIATIONS IN GRAZING PRESSURE BY HERBIVOROUS FISHES: TOBACCO REEF, BELIZE BY PETER N. REINTHAL AND IAN G. MACINTYRE NEWS AND COMMENTS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 : Ay itd f y ' < 4 SMITHSONIAN INSTITUTION LIBRARIES ae peat ht 7 ¥5 i oh K A aa aaa Sf) rity ‘i : gee i i a aah é ” ri se ehh adh ty “ t ¢ eit . POI A EAR TED ‘ i : ' ; ae if Bh Ha aa eras : i i J fipte ple meena Js of fist athe 5 ~ H : ’ 6) ey Pas whe bad} r ‘ M i a ie Gt Seeds ‘a \ an 4 } ¥ : ia Pot i, ¢ : , Rea Ls 4 3 5 ¢ ve Se MN ry 3 9088 01375 3900 ay Wfimstetls eek ie vats re eae hie 1a hala rai) rile tech lesen EO SHES cahoots g tebe Apnant botedtfnn era “abe sy! ein ws eleieseen Wuenary he ot pre feos ahh feta hi set oP } Riedie Saute 4 tehebe cthye eter en tt TG oe SS aay ee sae i Trane : oan ea ue eh tant ure sehr ane f pire ia i mT Vet URNA Et vous 3 aS Qu yetsat ay Lue EM ? ArGhyeAs heats HiGy Nias A ER per ayn aes wt AIS wi peur! PERS eS ANY ‘i neareinne ebinegat as Ate Wanner diye umes nr ha MESO RT Ba seni Weeds Wyn ara Nes bi iNaihalng 4 Say ante ie Bani Rei AO Be t sahatoaite! SAO WE prog phat mea Piling ay, igen ake: nis it peer ah Wate: ef ROntay une Han sou ee aaa, Wi Hi eB 1A ae ie i fied ne ra a fe Veswik _ R att ats na