ATOLL RESEARCH BULLETIN No. 50 Geology of Kapingamarangi Atoll

ATOLL RESEARCH BULLETIN

- No. 50 Geology of Kapingamarangi Atoll, Caroline Islands

by Edriin D. PIcKee

THE PACIFIC SCIENCE BOARD National Academy of Sciences--Natianal.Research Council

Washiii&ton,. D. C. June 30, 1956 .Page Introduction ...... 1 Acknovledgriients ...... 1 Atoll fra~nework...... Geology of the islands ...... Character of islaxis ...... Sedimentary rocks ...... Lagoon beaches and bars ...... Sear.iard beaches ...... Ragart and rampart wash ...... Migration of islands ...... Problem of Xapingamararigi eachroclr...... Petrology ...... Sediments ...... Sedimentary rocks ...... Soils ...... General statement ...... Parent materials ...... Alteration processes ...... Eelation of soil to posi-ti03 on is:.and ...... DevelopneilL of p11osphor.i-ie ...... Ground water ...... Generai features ...... Objectives of investigatioi? ...... Methods of si;udy ...... Depth to ground water ...... Variations in water level ...... La!: in rise and fall of g~ound7.iater ...... Relative permeability of seaj.riier~tarymateri.als ...... Effects of island size on leiis ...... Quali-ty of the ground water ...... Near-shore currents ...... Sedimenta-tion in the lagoon ...... Geology of the patch reefs ...... References ci,ted ......

Figure 1. Geologic map of Torongahai Island ...... 2 . Geologic map of Ringwtoru and Rikwnanu Islands .. 3- Geolagic map of Nunalrita Island ...... li . Geologic map of Werua Island ...... 5 . Geologic map of Taringa Island ...... 6. Geologic Map, of' Matiro 1sl.arld ...... 7. Geologic map of Matulietuke Island ...... 8 . Geologic map of Hare Island ...... 9 Geologic map of Pmatahati Island ...... 1.0 . Geologic ma2 of Natukerekere Island ...... 11 . Erofiles of lagoon beaches and distribution of beach sediment ...... Page Figure 12, 'Iypical soil profi.les in test pl.ts on Ringui;oru end k'umatahat,i ...... lac 13. Typical soil profiles in test pits on Riirvmanu, Pwakahi and Tokmgo 1sl.aniis ...... 1.8d 1.4. Lithology and water leveis in wells on 'l'aringa Island ...... 22a 15. Lithclogy and water leyels in wells on Werua Island ...... 22b

1.6s. Tide and well. fluctuations, , Taringa Island, June 23-24, 1954...... 26a 16b. Tide and weli fluctutitions, , Taringa Island, July 30-31, 1954 ...... 26b 16c. Tide and well fluctuations, Taringa Island, Aug. 20-21, 19jk ...... 26~ 17a. Tide and .*r,ell ,fluc,tuations, hTerua Isl.ar.d, J-me 23-%!$, 1954 ...... 26d l.7b. Tide arid well flccixations, Werua Island, July 30-3i, 1954 ...... 26e 172. 'I'icle and well fluctuations, Werua Island, Aug. 20-21, 1954 ...... 26r :1.8a. Tide tiad :re].:. fluctutitions, Hukuniu arid. ParaIrahi IsleLs, July 30-31, 1954 ...... 26g 18b. Tide and well fluctuations, Xuiruniu and Parakahi Islets, Auk:. 20-21, 1951: ...... 26h 19 Near-shore currerits of two soutlie:rn, ixo eastern ana one nort;?ern islands of KapYngarnarar.gi .... 28a 20. Near-shore currents of two eastern and two norther;l islai~dsof Kapingmarangi ...... 28'0 21. Sections across lagoon, ver-tical exagzeration approximately X30 ...... 30a 22. Offshore profiies, eas.t and south sections of atoll arid distribution of bottom sedjments .... 3Ob 23,. Profiles of Ivianin patch reef ...... 3Oc 24. lso~oathiclines of Kapingamarangi Ato1.l comgiled from data on U, S. Navy Hydrographic Cha,r-t 6042 306 25. Relative proportions of princlpal ty~esof sediment at 5-root depth intervals as determined by sampling ...... Fieid sketches of patch reefs showing dist-ibution of sediments and corals ......

Table 1. Characteristics of lagoon beaches ...... 2. Distribution and characteristics of ramparts and rampart wash ...... 3. Mineral composition of Kapingamarangi sediments as determined by X-ray diffraction pa-tterns ..... 4. Elements in Kapingamarangi soils determined by semiquantitative spectrographic method ...... -.--Page Table 5. Runice analyses; elements determined by semi- qua.ntj.ta-Live syec'irogra;~hic method ...... 1% 6. Percentages of apatite as deterinined by X-yay diffraction analyses ...... 21 7. hal~pesof' %rater. samples fro^:! test wlls and comparative data for normal sea water ...... 27 8. Depth (in feet) of :.agoon floor surro~.urdir~gMmirl patch reef ...... 311. 9. Some characteristic coi-a1 assemblages on Kzplnga- lnarangiAtol1 ...... 35 10. Thickness (in inches) of sand accumulations on patch reef ...... 35 Kapingamaraiigi Atoll is located at .the southerrmost end of the Caroline .roup of islmds, slightly rimre than 1' north of the i$quatf>i.. It is about k10 miles s0xthcas.t of Tnk and 350 mlles novth of I?a'uaul im New Bribain. This isolated atoll surromds a lagoon al>out 6 miZes across fzom ex-l; to xest and slightly less ffora nwth to south. It is inhabited by 1125Po!.ynesians (19511), mos-t of vhom live on Timhou and Werua islands on the eastern side. DE.GR~~Sof geographic setting, cl.ima:te, and general e:ivix-o~:~e~~thilve been treatec?. H. J. %icw so will not be iacluried hwe.

This report on the gectlogy of Ihpingamar'angi Atoll covers prelimi.riary resul", ffr~ilfield investiga%ioi:s conciuc'ted during late Jwe, J,and AU- gust of 1954 and lzi~oratolyexamination of specimens made duri??~-the winter of 1951!-1955. Much attention has been devsted to stuiies of the islands, with special emphasis oil the classificaticn of sedimeuts an& sedimentary rocks, the occurrence %,id. behavior of grouiXl ::ater, and. the distribu-tion of soils. De- tailed stu6i.c~also were made of -t;ile processes of seiliinen-tation that curreiltly are qerating bot;h oa the reef &id in the lagoon. These s-Ludies incl.uded the making of deta5.3.ed aina,lyses of beach aad bar str~ctures,the recor::stru.ction of reef acd offshore profiles, and the sysi,ema-ti2 sampling of the lagoon floor.

Termiiiology med in this report folio>~s,in general, that reconumnded by Trace:? and othe;s (1995) in their receiit paier tl-iled "Conspicuous features of organic ~eefs." The terra "atoll" is useu for the peri;3lieral reef arid everythiiig ~rithii~it. The upper surface of the peripheral reef, except where covered by islands, is refened to as the reef flat. Small reefs within the lagoon, icdiaritis that heve been ca1l.ed patch reefs, small table reefs, reef knolls ma reef pinriacles, are discussed under the ger.ei-al desig- nation of patch reef. Dense growths of staghorn or branchinl: corals are referred to as t1iic;rets. Other terms for geomorphic and organic features are expl.ained in the text wherever thcir meaning is not apparent.

The geokogical iilvestigation described in this repa?t ms conducted as part of a project of Vne Pacific 'Science B0ar.d of the NzrLiona!. Hesearch Coun- cil and ?!as supported 'oy funds from the Office of' Naval Research through its contrac-t Wronr-29104 ('NR ,388 001) with the PJaational Acade~~yof Scimces. The U. S. Navy Depa.rtment aild the ~MiliiaryAir Transport Service furnished trans- portation to aiid from the atoll and the Navy Department assistea greatlf by supplying many items of equipment. The Civil Administi,ative staff of the Trust Territory of the Pacific Islands was helpful in milily vays . Many thanks are due Harold Coolidge, Lenore Smi:th,and Ercestine Akers of the Pacific Science Board for the help t,hat they renaered. the party. Ap- preciation is also e,upressed to iienneth P. &lory of the Bishop hiuseurn in Honolulu for his briefings on the atoll and its inhabitan-& and for personal introduction to its peop2.e. Preston Cloud and Joshla Tracey of the U. S. Geological Survey \,rere both extremely helnful with sugges-tioils for the work and with the loan of scj.enti.fic equi-gme;i,t.

My associates in the field. pazty were Cadet Iiand, Robert R. EIarry, Id. Jar1 Newhouse, William Niering, and H. J. Wierls, all of whom cortributed in various ways totrard the developmer-t of this report. The base maps of the islads prepared by Wiens were especialiy important in carrying out a gen- logical nrogram arrd the assistaxe of Niering in soil. studies arid tide re- cording is gratefully ackrrowledged. I

In the study of collections and samples of materials, I have been as- sisted by a nurilber of specialists, adit is with pl.easuz-e tha.t I ackno:~rledge Che following cooperation:

Coral identification -- John W. Wells, Cornell U!liversity. Foraminifera identificatic~n-- Ruth Todd, U. S. Geological Survey. Analyses of water from wells arid. lagoon -- john Hem, U. S. Geological Survey. X-ray 'diffraction ideiltificatisn of mirierals -- A. J. Gude 111, U. S. Geological Survey. Analyses of elements in soils -- P. R. Barnett, U. S. Geological Survey.

For critical reading of -he manuscript and many helpful su~;gestioi.ls, the writer is indebted 'GG G. D. Robinsoil of the U. S. Geological Survey, John Chronic of t'ne University of Colors~do, and Halka P, Chronic.

THE ATOLL ~.UEI.JORK

The peripheral reef of Kapiiigamararrgi Atoll (fig. 24) encircles ail oval lagoon, almost cormple-tely sep,rating its waters, especially at times of lo%! tide, from those of the surrounding sea. The only major breaks are relativsly narrow passes on the south side where channels nearly 20 feet deep permit strong currents to flow in and out continuously. This peripheral reef includes what has aptly been termed the "frame" of the reef complex or reef mass, arid is a lattice constructed -through the ermth of organisms, especially -; corals and coralline algae. As stated by Cloud ili)52a, p. 2128), these organisms "serve to holc! it bhe reef7 together, and the frame they build is a trap for clastic or chenicallg pyecipitated sediments."

The surface of the peripheral reef is widest -- nearly 4,000 feet across -- near the northernmost sector, although it is almost as wide in the extreme western part (fig. 24). On the sov.thern arc, west of the main passes, the surface is no wider than 1,000 feet. Tfie 33 islands distributed along the eastern half of the reef a.pgear as very low mounds that rise slightly above the general flat top along its lagoonmrd margin. They alone stand above the level of high-tide waters. El.sewhere along ?he atoll, espe- cia,l?.y on the southwectern, western, and nor'cnwestem seciors, the reef surface is d.iride4 into a seavard slope anu a lagoolivai-d. slope by a lovr but definite crest. Bokh s?.opes a,re extreneiy geii+~Lein most ~laces. The:. a?:e referred to in this paper as the outer reef fiat,and the imer reef flat.

The essential1.y flat top of ?:early all the outer park and sone of the inner part of the reef fc~rmsa pitted rock pavemen-t. Over wide areas this surface contains no organisms now contributing appreciably to grcwth of the reef. To the contrary, i-t appears to be solely the resu-lt of the wearing do~il,by waves and currents, of coral mmsses that once stood. higher than the present surface. Eviderlce from Okinawa obtained by MacNeiL (1950) indicates that similar surfaces there are the result of recent wave or solution planation. Other siniilar reef flats reported from many islands of tile :Pz.cific suggest, as sLated by various geologists, that this wiciespread feature is the effect of a recent lowering of sea level irith resultant death of all corals back from the reef eGge, fol.lmred by a be.veling of higher parts of the reef structure.

The present surface of the reef flat appears to result from an essentially static position of the sea for a time sufficieiit to develop a general evenness of planation and a condition of near sterility over wide areas. small living algae of the genus Boodlea cover extensive areas of the Today, ----. rock surface; mollusks and other marine anima2.s con@;regateal-ound and under limestone boula.ers and cora.1 heads that are strewn over. tiie rock surface -- debris-washed across from the seaward side. Brl-ctle stars., eels, and marine worms inhabit cracks and cavities that extend down into the rock mass. Liv-' ing corals and coralline algae, howver, are largely res-tricted to the seaward margins 01 the reef and, where isl.a!lds are absen-t, to wa.tei-s of' the inner reef area bordering the lagoon ed~e.

Clastic sediments are relatively.scaxe over much of the present reef- flat surface, but are p~ogressivelymilo:-&! abundaat toward the 'imer margins of the reef where they fokm a veneer tnat covers a framework of coral. The small amount of fragental rraterial on the outer side of the reef' fiat consists largely of co~alan& coralline algal de'bris, up to and including boulder size,, strewn over the surface, and of elastic particles from mollmk shells and echinoid spines, concentrated in pockets and cracks. Foraminifera of the genus Calcarina that are reported as common on the reef flats of m.@njj Pacific atolls are absent here; few Foraxinifera of any type occur on the' seawardparts of the atoll. The tests of these arlimals accumulatein quanti- ties sufficient to' forin 'deposits.df lime sand o:~lx bn the island 1-agoon beaches and on the inner parts of -the atoll. As poi.nted out by Sollas (1904, p. 6, 7)however, such line sands ultimately fill many of the interstices in the framework limestone mass, and they develop into extensive deposits on the inner side of the ''retaining... wall" of coral and algal structures. Attempts were made at ~a~in~amaran~ito determine the lithologic character and structura~'features of the framework limestone through studies of material exposed in cracks belo7.1 the pavement and of large aoulders washed up from seaward exposures. The investj.gatio13 was hamered b~ the extreme difficulty of digging into the reef aid by a genezal i.ack of na.tural breaks elcpocing sections, but from a series of samples some genera1.Lzations can be made. In ma,n:r spec?.rr.e?~s,the reef rock appears aphanitic, but in others the relict structures of corals are clearly preserved.

The framework limestone as a whole seems to be very cavernous, although small masses and hand specimens c~~mlonlyhave a relatively low apparent porosity. Cavernous structuxe, observed in near-surface e:xavations, probably extends to considerable depths. This is inaicated by the behavior of the fresh-water lens on various islands, in ~hichall parts of the lens rise simultaneously and with a simil.ar tidal lac (2escribed under "Ground water"). It also is suggested by evidence from the Crill holes at Funafvti (~ollas, 1904, p. 6) and at Eniwetok Atoll a add et al., 1953, p., 2259). Oil the o-ther hand, some of vhat originally were cavities appear to have been filled with clastic sediment. This feature has been noted by Neweli (1954, p. 10) in reference to reef blocks from the outer edge of the reef flat at Raroia.

The reef-buildiag organisms thai; contribute chiefly to the framework linestone, as judged by the forms currentiy growing almg the eastern seaward margin of the atoil, are very largely corals but include soce nasues of the coralline algz -.--Porolithon onimdes. The corals are represented by many species (table 9); the most cmimon belong to the genus Acropora, except along the inner margin, where microatoils of Porites lute% are dominant. Suggestion that much of the limestone beneath the presenb reef flat is composed of a comparable assemblage is found in -th- mineral content of selected samples examined by X-ray diffraction met!nods,~7 All specimens tested show -- -.I./ Analyses by A. J. Gude 111, U. S. Geological Survey. 90 percent or more aragonite, which percentage probably reflects the proportion of coral, thouzh a small atnoun-t of aragonice roc$ be due to irl-terstiti 1 deposits of clastic shells. Five to10 percent of high-magnesium calcite-27

+.. +.. 2/ Iiigh-magnesium calcite, as opposed to low-magnesium or norrial calcite, is discussed by Chave (1954, p. 267). Iie points out that calcite of alga!. structures contains more than 10 percent magnesium carbonate, whereas that in mollusks and certain ot'ner organisms contains less than 2 percent. in some specimens suggests the amount of algal contribution. A lack of normal calcite in all samples indicates Vnat the common reef-dwelling Foram- inifera of the genus Amphistegina were not included in the samples examined. Unfortunately, no data are yet available on the lithology or structure of the Kapingamarangi reef at appreciable depths below the pavement; of the reef flat. Judging from records of vells drilled on other atolls (F'airbridge, 1950, p, 384), however, one might expect to find zones of clastic ma-terials representative of vafious depths and einvironments, alternating vith zones of reef-forming corals and algae similar to those on the surface today. N.1 such changes at depth appear to be related to times of advance and retreat of the actively growing framework corals and coral~inealgae and probably were con- trolled by relative changes of sea, lewl and still-stands. Both the upward and the lateral deoe1oprnen.t of the lj-mestone framework are a record of tec- tonic and clima-tic events.

GEOLOGY OF TIa ISTAWS

-Character---....- of the islands.-. Thirty--three islands are distributed along tne arc -that forms the eastern, rrindviard peripheral reef of Kapingmarangi Atoll. The largest is IIare Island, vhich is more than a mile long and 600 feet vide, and the s1:iallest is Matukerekere Island, vknj.ch is about 130 feet long and supports but a single mature tree. Sou~eof -i;hese islands are composite, having attained their presen-t sizes by the combining of two or more small islands through processes of sediaentation. Other islands prob- ab1.y represent various stages of diminution through partial destruction or dissection resulting from cyclonic storms.

The origin of islands perched on oceanic a-tolls is a problem about which man has speculated for a long time. Some early views on this subJect are recorded in the log of Captain Cook (LI-ojd, 1949, p. 266) under the 6at.e of April 17, 1777:

"There are different op;nions amongst ingenious theorists, coi?cerning the formation of such low islands as ~almerston's. Some will have it, that, in remote times, t,hese li.ttle separ'ate heads' or islets were joined, arid formed one contiliued arid more elevated trac-t of land, which the sea, in the revolution of ages, has washed away, leaving only the hi;;her grounds; which, in time, also, will, according to this theory, share the same fate. Another conjecture is, that they havebeen thrown up by earthquakes, and are the effect of internal convulsions of the globe. A third opirioii, and which appears to me as the most probable dne, maiiltains that they are formed from shoals, or coral banks, and of consequence increasing."

I11 order systema-tically to accumulate data relative to the islands on Ka.pingamarani$. Atoll, geologic maps were prepared during the suumer of 1954 for all of the larger and some of the smaller ones (figs. 1 to 10). IsLand maps on a scale of ore inch ecpals 100 feet, compi!.ed and surveyed by H. J. Wiens, furnished bases to work on. The distribution of materials was plotted according to the classifica.i;ion discussed in the section of this report on "l?etroiogy." Dips in sedimentary strata were recorded on the maps.

Analysis of the maps indicates that Xapineamarangi islailds are forrned of three principal classes of material: (1) sedimentary rocks formed through the cemensation of clas'cic particles and organic remains, (2) unconsolidated sediments of beaches and bars, and (3) surficial deposits forming the ramparts, rampart wash, and soils (not mapped) that gartly cover and mask the other two. On the basis of the distribution and structure of these three classes of material, much of the island history may be interpreted.

Sedimentary rocks.- Beds of sedimentary rock rise above the reef flats along the seaward margins of all the large islands and many of the small ones, and they crop out locally within many islands. Because in most places the clastic partic2.e~of which they are formed are clearly d.iscernib1.e and because their stratification cownonly is prominent after weathe~ingthese rocks, for the most part, are readily distinguishable from the reef rock on vhich they re&. Isolated pedestals and undercut bl.ocks formed. of similar clastic rock stand on the reef flat considerable distances seaward fmm some islands -- reimants of earlier island. masses. Conspicuous illus-brations are on the flats east of Werua Island. Likewise, relict pla.tforms ofstratified clastic rock, mrn to a low level .through planation but stand-ing above the coral rock of the reef flat, extenir 11.00 feet northeastward from T'orongahai Island.

Evidence furnished by the distribution of erosional remimts of stratified rock suggests that the seaward shores of all Kapin~aroarangiislands have been retreating: for a considerable time. Similar evidence on Raroia, iwdicating reef-flat extension at the expense of islacds, has keen noted by Kevell (1954, p. 14). A second feaLure that tends to support this concept is the orientation of cross-s+,ra';ification along shore-line exposures of island rock. On many islands, of which Werua is an excellent illustration (fig. h), the s-brats in rocks that form the seaward coast dip exclusively lagoomard over lorig stretches in the same manner and degree as strata in tihe modern lagoon beaches. This fact suggests that strata in the two places were developed in a similar manner and., therefore, that the deposits forming these rocks accumulated at a time when the lagoon pargin was in the present position of the seaward shore. Still further evidence of island retreat is presence of the mineral apatite in coastal stratified rock on Nunakita and Ringutoru Islands in wave-washed areas beyond the present limits of trees, This location indicates that phosphorite must have developed at some time in the past vhen these localities were the interiors of islznds and vhen guano from birds was accumulating nearby.

The present distribution of stratified rock indicates not only a former, more sear,rard posi-tion of the islands but also a higher surTace level. Strat- ified rocks on various islaads, especialiy prominent on Rikumanu ald Ringutoru, stand 4 to 5 l'eet above the present high-tide level. These rocks are formed of clastic particles and. are cross-stratified; they tippear to be leaclled and partly phosphatiz,ed. Rocks having a similar high-level position in many atolls of the Pacific have been recorded avid and Sweet, 1904, p. 67-66; Ladd et al., 1950, p. 413) and are generally considered to represent deposits residual from a time twhen, owing to eustatic changes, sea level was higher than at present.

Lagoon .- beaches and bars.- Larse parts of most of the present islands,

inclu-~airtua1.1~' - all of-he lagoonward sides, are composed of unconsolidated. clastic materials and foraminifera: sands. Such sediments are accumulating today on the bars or horns projecting into the lagoon from both er:ds of each island and along the incurving beaches between these bars (figs. 1 to 10). Test pits an& .wells on various islands indicate that similar uncon- solickted materials extend dcr~rnwardin many places at least as far as low- water level, 3 or 4 feet below the surface. Sedimentary s'zuctures, especially stratification, further show that t'nese sediments were deposited largely as beaches, dipping lagoor?ward, and as bars. EXPLANATION 0 Beach or bar sand Aphanitic to granular limestone Puraka pit pzJ doO Beoch or bar rubble Rampart boulders Strike and dip of beds

a* * a Corol rubble con lomerate RampartFxa rubble House Limestone bouldera canglomerote Wash rubble FIGURE I. - GEOLOGIC MAP OF TORONGHAI ISLAND 0 500 Feet RIKUMANU

FIGURE 2.- GEOLOGIC MAP OF RINGUTORU AND RIKUMANU ISLANDS 0 500 Feet EXPLANATIONrn Ramparti':,.i,.q boulders . Romport rubble DJ Wash rubble

Beach or bar sand

Beach or bar rubble m Coral rubble conglomerate m Limestone boulder conglomerate m Aphanitic to granular limestone a Purako nit 0 House

FIGURE 3- GEOLOGIC MAP OF NUNAKITA ISLAND 0 500 Feet I I I , I EXPLANATION

Beach or bor sand Rampart boulders m rn 100.0 1 Beach or bar rubble Ram~ortrubble

Coral rubble conglomerate Wash rubble EsY a Limestone boulder conglomerate Puroka pit m 1~2" Aphanitic to gronular limestone Strike and dip of beds

C] House

FIGURE 4. -GEOLOGIC MAP OF WERUA ISLAND 0 500 Feet I I I I I EXPLANATION

Beachpq or bar sand Rampart boulders Beach or bar rubble Rampart rubble

Coral rubblem conglomerate Wash rubble Q Puroka pit 9O a 1 House Strike and dip of beds

FIGURE 5.-GEOLOGIC MAP OF TARINGA ISLAND 0 500 Feet EXPLANATION

RampartrJ rubble Beach or bar sand

Beach or bar rubble

Coral rubble conglomerate

FIGURE 6 --GEOLOGIC MAP OF MATIRO ISLAND

0 500 Feet I I I I I EXPLANATION

limestone

Limestone boulder conglomerate Strike ond dip of beds

FIGURE 7.-GEOLOGIC MAP OF MATUKETUKE ISLAND 0 500 Feet FlGU RE 8 .- GEOLOGIC MAP OF HARE ISLAND 0 500 Feet w 1954 EXPLANATION

BeachWl or bar sand Rampart boulders Beachm or bar rubble Rampartrn rubble Carol rubble conglomerate Wash rubble ,so m Strike and dip of beds Limestone boulder conglomemte 0 House 1 E 9.-GEOLOGIC MAP OF PUMATAHATI ISLAND FIGURE 10 .- GEOLOGIC MAP OF MATUKEREKERE ISLAND 0 50 100 Feet I.2 Beaches and bas on allof the major islands were n~~ped,and details of their struc-Lure were recorded from trenches dug at right angles to the strand on Taringa, Pemkahi, IuIatiro, l%matahati, and Singutom Islands. Re- sults of these stild.ies must at1ai.t d.escription and axtPysis at a later date, but some salient features are shown on three iypical bedl profil-es (fig. 11). The foreset beds of. the beaches dip from h0 to 11' lagoonward, except for short distances immediately below beach crests, whei.e many dip as much as 15' to 20' (table I). Backshore beaches commonly are horizo~~kal,but sGme dip inland 3' or 4'. No evidence that any of the lagoon beaches are being converted into beach i'ock covld be found, and beach deposits encountered in test holes on the lagoon sides of many islands 'were unconsolidated.

Lagoon beach d.eposits vary in composition from time to time and from place to place 8,ccording to effective wave conditions. Four principal com- ponei~kare recogrized, and these are normally sufficiently well sorted that they are seprated into parallel bands along the beaches.. The components are (1) lime sand canposed largely of an orange, ovoid-shaped foraminifer (&iphistegina nadagascariensis), vi-th lesser mounts of avhite, disc-shaped -,---- foia?linif~~rgi~.o~ora---.--.- -vertebralis); -. (2) white coquiina sand composed largely of cotminuted pieces of she11;73T-gray coral rub'ole, worn and washed, up to 1-l/2 or 2 inches long; (4) pieces of gray ol. black pumice averaging from 1/4 to 1/2 inch in diameter but some as much. as 5 inches, worn and rounded except where broken along fresh fractures. . . The sorting of beach ma,terials according to weight and specific gravity results inpumice being concentra-bed on the backshore and the &her three types of sedininen-t on foreshore beaches. The orange forminiferal s3nds today domirate the foreshore surfaces of most lagoon beaches,but on Fumata- hati and par.ts of Ha:z Island, coral rubble covers the su.rface because of special conditions responsibie for s'cro~?gwave action. Test trenches show that at vario,us times the beaches of all the islarlds have-heen covered by deposits of rubble, but most of these beaches were later buried by sand as they built forward under conditions of normal wave action. Island horns or bars that are ilOW building ou'c into the lagoon commonly consist in part of foraminifera1 sand and in part of coral rubble, the distributioa of each ma-berial depending oil conditions of locel current aad wwe strength. ------Table 1.--Characteristics of lagoon beaches -1 - - . . .-- .- --- Angle of dip toward lagoon (degrees) -. --. - - Island Width, iow tideT--- ores shore ~10'~e/subcrest sl~pe~Backshoreslopel 1

Parakahi - lo Matiro hatahati 20 4-9 Ringutoru 26 Torongahai Unconsolidated deposits of beach and bar that form major parts of some islands on Kapingamarangi furnish evidence that sedimeatation has cause& the layoonward sides of -these islands to build forward ad; an appreciable but i~ndetermiiiedrate. Beach .trenches show the recei~tacc.ai~~laiion of backshore deposits over earlier foreshvre sediments; test pits near -the shore ex2ose humus layers mixing with backshore deposits; wellsin 'die interior bring to view sequences of 7-agoon1,~ard-dippingforaminifera1 sands of foi.mer beaches. Most conspicuous of all features furnishing evidence or) beach migration is the presence of relict zones of pumice, located back from the margins of; some beaches and marking the baclishore accumulatio~~sof earlier . .. periods ,g --.-- 1/ Similar bands of punice on Addu Atoll have been reported by Sewell (19-93G P. 77).

Whether this agsradational process is as rapid as or more rapid than the ra-t,e of island destruction oil the seaward side is not knobm. Presumably the rate of wearing back of land has decreased in proportion to the distance from the reef firon'; during the current still.-stand of the sea. Although the present reef flat is relatively wide in most places, there is ample evidence that erosive forces of the sea are still vei-y effective on many of Yne islands. On the other hand, the rate of island building through sedinienta,tion majr be retarded as the shore moves forward into contimally greater depths of the lagoon, requiring greater amounts of sediment to build up the bottom. --Seaward beaches.- Beaches are ?elatively scarce on the seawo-d sides of Kapingamarangi islands; furthermore they are small and short-lived. Most of them are perched on the bevelled surfaces of stratified island rock at various distances above low-tide level. As i:idicated on the geologic maps (P.:..&s. pto lo), few of them are continuous for long distances along the shores. It is doubtful that any of -these beaches make permaneit co!itrr;.ou- tions to the growth of the islands.

The seavard.beaches differ from the beaches on the lagoon sides of islands not only in being more patchy and far less extensive, but also in composition and color. They are foymed largely from comimted shells of mollusks and are white, in contrast to .the orange lagoon beaches, r,rhich are mostly composed of Foraminifera. A small beach at the seaward end of ;)ungu- pungu Island (fig. 6) forms a coquina composed almost entirely of unbroken shells of a small pelecmod of the genus Trigonocardia,

--Ranparts and rampart wash.- Accumulations of clastic debris consisting largely of coral rubble, coral heads, add limestone bloclcs are deposited on tne borders of most islands, abovk wave-cut benches and beach crests, by occasional violent s-torms. The deposits are classified as boulder rzmnparts and rubble ramparts,' according to their constitue~~ts,or as rampart wash where the debris has been spread out as a sheet below and beyond the iinland part of the rampart ridge.

Present distribution of these surficial deposits, shown on the geologic maps of the islands (figs. 1 to lo), gives some indication of the directions Main crest

Foreshore 7

Backshore I Crest Foreshore ---- -

Crest

~ &,ckshore ....-....7 -Foreshore ------? 0' Subcrest

.. .. x X 0 0 .. 20 .,. . ',. 0 -0.J ...). ,

EXPLANATION

Farominiferal sand (orange) Coral rubble

Clastic shell sand (white) Pumice gravel

- FIGURE I I. -PROFILES OF LAGOON BEACHES AND DISTRIBUTION OF BEACH SEDIMENT

0 5 fl iLllli of approach and the intensities of recent storms. Large islands along the northern arc of the atoll -- Torongahai, Ringutoa~,and hJunaltita -- all. have large ramparts on both eastern a;id western shores. The widest are on the western sides and in these the largest blocks (3b inches i: diameter on Torongahai) are concentrated near the ilol.thern ends (table 2). Islands of the eastern arc -- Werua, Matiro, and Hare -- have ramparts on their eastern seaward siLes, but these ramparts are forlred mostly of fine rubble with very small areas of boulders. Individual bloclis are on the average much Smal.Ler than on northern is1and.s (table 2). On -the southern arc, at Puniatahati Island, large ramparts :have been formed or1 both seaward and lagooiiward shoses. The rampart to seavard consis-ts largely of boulders; that on the lagoon side of coral rubble oidy. The unique development on this islaad of a lagoon-facing rampart mus;t be:attributed to a fetch sufficiently great to allow easterly storms in -the 18ggon.to reach high intensity-.

Rampart wash, as indicated on most of the:maps (figs. 1 -to lo), extends inland from the ranparts for distances as much as 250 feet, forming a surface covering of coi-a1 ~Xbble. Because ranpart wash ha.s developed largely ou the seaward sides of islands, it rests on bedrock in manjr places, on soil j.11 others. Apparently it is the result of vatem washing a put' of the gravels and other small constituents of the rampart beyond its crest. On the northern islands especially, rampart .wash covers extensive areas back from the seaward margins. Migration----- of is?.a,nd.s . - In preceding sections of %!]is paper, geological evideke has been cited showing that the seaward sides of islands are cur- rently being eroded back so that the cater reef flat is becoming wider at the expense of island stratified rock, and that the lagoon sides of lslands are constantly growing through accumulation and deposition of beach and. bar sediments. The total effect, therefore, is that of island migrakion across the reef, from seaward to lagoonward side. So long as sea level remains relatively stable with respect to the atoll, this process may be expected to con- tinue, though increasing distance from the, reef front causes the effective cutting power of iqaves to aecreaie, &d the advance, of the lagoon shore into progressively deeper ria-ters requires more and more sedimen'cary material to build up a comparable amount of beachjfront. J. I. Tracey (written comuni- cationj July 1355) suggests that solution of. rock is more effective than abrasion by waves in the retreat of islands and that wide reefs favor rapid solution. He also suggests that wide reefs may favor rapid sedimentation as they probably provide more sediment. Thus, it is not !mom uhe'chei- or not these islands migrate at a progressively slower rate. In addition to the evidence of isla~ldmigration already given, many biological features support-- the thesis. One of the chief among these is the undefiiiiing of trees -- especially coconut palms -- by waves 01; the seamrd shores. Mmv such trees are tilted or have fallen outward from zhe islands, The Kapingans are aware of the effectiveness of this process and on both Touhou and Werua Islands have constructed stone embankments along the southeast coasts to protect the land. Evidence of the recent growth of lagoon shores is found on several islands where successive lines of certain shore-fringing species of trees are now standing in the interiors, marking the sites of former shore lines. Detailed studies of the relations of island m&ration to glant life were niade by William Eiering duriiig the siun- mer of 1954, and results nave iiecn reported by him. On ICapingamara,ni;i Atoll., isl.and mizratior appears to have been, in general., grea-test for the n6rther.n islands ant leas's for tlie :;oYcherr?. This is indicated not ori?y by a greate;. width of outer reef flat in the emf-Lh than in the south, but al.so by the exte~tto vhich beielied relllranl-,s of s'mati- fied bedrock extend outvard from the seaward coasts of Toi-on:~&hai,Fingutoru, Nunakita, aiid lderua Islai?ds (figs. 1, 2, 3, and 4). In contrast, Hare Island, or. the soutlierii par% of the ai;oll, appeers to have been relatively stable; it shows little evidence of shore recessio?? on fhe seaward side and does not appear to be building out rapidly lagooiivard. 11 Its adjacent outer reef flat

L/ Its lagoon beach probably is the ty~eof t:?at at Ari~oAtolldescribed by We& (1.951, p. 5) as II degrading." Much of its surface is covered by rubble c~ncent~ates,the sand having been removed, and several coconut palrns growins near itare partly underniined. --. --.--- is coi~iparativelynarrov.

debris aa6 'oclr fragments, cemen-ted by calci~mcarbonate, It varies greatly in texture and in degree of cenientzbion, some being ex'creuiely friable, some well cemented and dense. MLIC~of it is sbratified or cross-stratifiedi and. it has all the charac,te-istics normally attribute& to beachrock. Details of its origin are not entirely clear, however, despite the numerous suggestions cancerning the oriein of beachrock that have been offered in the many publi- cations on this subject.

Because modern beachrock is limited to tropical seas, its relation to warm waters apgears tc be beyond question. Beca~~seit develops excl.usively in the i.ntertida1 zone, its relation to the rise and fall of tides seems equally d-efinite. Other fac.tors -- those responsible for its localization oi certain beac&es or parts of beaches -- are less clear. basic rerpi?:ements in the development of beachrock have i-ecently been summarized by Gizsburg (1953, p. 88) as (1) high temperature, (2) rapid rate of beach draiiiage between high tides, and (3) a permai?eixx of beach deposits sufficient to allow the cementation process to operate. The localization of beachr~ckis com.only explained as due to the precipitation oii calcium carboaate from sea irater as a resul-t of heating and evtiporation. The probable inportacce of blue-green aigae as agents that bind sand grains toge-bher until they can be cemented has been su.ggested by Cloud (1952b, p. 28) and others. As a result of observations on vszious Pacific atolls, a diiver&ence of conclusions has 'been reached concernin;; the areas in which beachrock is cur- rently developing. Ladd and others (1950, p. U) state that on Bikini Moll. "beachrock may be fornied bekween the reef fla-t and the beach." These &u-thors refer to the Fresh look of the rock a?d the original color of organ5.c inch- sions as evidence that lithification is going on today, bu-b also point out that exposed parts of the beachrock are beins eroded. Regardin:: Onotoa Atoll, Cloud (1952?o, p. 28) describes "bonded limesand," which he consi6ers to be incipient beachrock, forming on lago~nbeaches, in tide pools, and in spray pools, and states that it was not fot~qcl021 this atoll anywhere c;n the seaward Table 2.--Distribut-ion of ramparts and ra~par-twash

v-Boulc er ranpart Rubble rampart Rampart wash Island Location Width ' Size of bovJders Location 'id th Location Mid th I (feet) (inches) - feet) A - ______/I____' I I Torongahai / West shore 36 South part i East shore i 28-32 of east and I -- west shores i I North shore 115-164 24 Part of west Sout:l.west 1 30-200 Ringutoru I! Wss~shore 18-77 6-10 shore end borders 1 23-72 / East shore 21-54 6-12 esst raxpart t I ! / E .. - al I $o Rikumanu surface 100 18-2b z Nunakita West shore 22-1'20 12-36 South part of Nortk end East shore 31-68 6-8 east and west borders east j sii0:-es rampart - wema Soutk%east 1 57-120 24 South. s liore Borders east / I i silo rE East shore ranpar t I I

Matiro Southeast 6-24 East shore

East end I Northeast / 42-67 / 6-12 East shore Southeast- sh -- rampart I I I Puniatahati South shore 33-100 6-12 North shore West end borders 30-1 00 beach. Concerning Raroia Atoll, Newel1 (1954, p. 32) says that beachrock is quite rare along lagoon shores, but that it appears to be forming in "moats" in the island interiors -- behind lagoon shore ridges and, especially, back of seaward ramparts. This is probably the Cay Sandstone of British and Australian geologists. Most of the beachrock exposed at Kapingamarangi Atoll todtiy apgears to be relatively old; This is suggested by (1) the presence of typical beds preserved as relict deposits standing above present high-tide level on several islands, (2) the bevelled remnants of typical beachrock extending seaward a few hundreii feet, across the reef flat, from the present island shores, and (3) evidence of replacement of calcium carbonate by apatite in strata both in the interiors and on the shores of several islands.

Although considerable evidence indicates that active erosion dominates the seaward shores of most Kapingamaraiigi islands today, there is reason to believe that, locally at least, some lime precipitation is going on contem- poraneously. On the reef flat near the shore,of Parakahi Island a rectangle of boulders placed there by man at an unkaom date has been firmly welded in place through cementation. On the east shore of Tirakavme Island, a 3-foot block of stratified limestone from an ancien-k outcrop of beachrock is now standing on end, incorporated in the present middle beachrock layer. How recently this development took place is not known, but it clearly shows a second stage in beachrock development. Such illustrations of recent, though local, precipitation of calcium carbonate on the seaward shores of islands are supported by records from other atolls.. These include the record of a fragment of green glass from a Japanese fishnet float, embedded in beachrock deposits at Bikini a add et al., 1950, p. 416) and the report of a firmly bonded gravelly sand containing brass cartridge shells on the seaward shore at Tarawa (cloud, 1952b, p. a).

. , On Kapingamarangi Atoll no development of beachrock on the lagoon shores of the islands could be detected. Although a large number of test trenches were dug across the beaches of various islands and most of the beaches were examined in connection with mapping or the islands, without exception they were found to be entirely of normal unconsolidated deposits of sand and gravel. Likewise, no clear evidence of beachrocls development vas found in the island interiors, for "moa.is" that are periodically flooded by sea >rater as described by Nevell (1954) for Raroia do not occur at Fapingamarahgi. Thus, if beachrock is forming today in appreciable amounts on the islands of%- pingamarangi, i-t must be on the seaward sides. Most of the rocks ' on these sides, however, appear to be wearing away rapidly, so it is doubtful that' beachroclr development is extensive there at present. The theory that bea,chrock may develop by the worls of ground water that dissolves calcium carbonate from lime sand and precipitates it as it seeps through the beach at low tide was proposed many years ago (Field, 1920, p. 215). Considerable evidence opposed to this idea has since been presented by Daly (1924, p. 138) and others. On the islands at Kapingarnarangi, it clearly cannot apply because (1) intertidal sedimei~tson the lagoonward margins of islands are not lithified, (2) bedrock on the seaward side rises above high-tide level and extends inland a considerable distance, and (3) no correlation between the water levels and island bedrock can be detected in the series of wells across Taringa and CJerua Islands (figs. lb. and 15). Sedimentary material tkat foms the atoll of Kapingamerangi is, with few exceptions, comgosed of calcium carbonate. Part of it is uiiccmoli- da%ed accmulatioiis of clastic particles. The re~ainder,referred to as sedimentary rock, is lithified mate-is1 developed through cenient,ation of these clastic sediments and froin the reef-building processes of co-als, algae, and other orgauisms. For pwposes of studying and mapping, principal varieties of sedirneiit and sedimen.tary rock have been ciassified lnto easily reccgni.zable types, based primarily on texture wad secondarily on the major contributing ~rg~ismsor rock inp;redi?nts.

Sediments.---.-- Unconsolid.ated materials are separated into three groups according to particle size, folloving ';he Wentworth classification of detrital sediraeints. These groups are (1) lime gravel, in which par.ti:les are greater than 2 mm in diameter, (2) lime sand, in which they are between 1/16 and 2 mm, and (3) lime mud, in which they are less tinan 1/16 mm. De- posits of all three of these gr.oups form significan-t parts of tine atoll, 3;ld e8ch accumulates in characteristic situations,. as will be described later. 20th Lime grsvel and lime sand are concentrated on the reef flats, locally forming deposi-ts from which the islands are built. They also cover mnch of the lagom floor. Lime muds, on the other hand, are confined largely to the deepest parts of the lagoon.

< Two principal varieties of lime gravel are recognized: One is referred to as boulder gravel, for it contains angular blocks of reef rock, rounded coral heads, or masses of coralline a.lgae which are of boulder ( >lo-inch) size. The other is formed almost exclusively of coral dble, derived largely k.orn broken fragaen-cs of Acropora--- or elk-horn coral ranging from about 1/2 to 2-i/2 inches in length. Although many mixtures or intermediate stages between these varieties of grave.': exist, the importance of recogxizing them as separate types is that the boul3er gravel lndicaies the action of violent storms awl accompanying large waves, whereas the much smaller coral rubb1.e is tramportea also by normal waves and currents, and so is being deposited' almost coatinuously. Lime sands of Ihpinganarangi Atoll vary according to the oi-gailisms or combinations of orgar!isma Trom which they have been derived. Those of the beeches and shc.llow offshorre wxters are com~osedlargely of the small orange foraminifer Ar;;;?hj.sl-eginamsdagasca~iensis ..- ._-----w-_-_- 1 the white ',?heel-shaped foraminifer Mar.rginop~:r.: vxtebmXs or botll, Some 'Jeaches contaii; a high ) percentage of white i-'rafperr'cs of molluslr shells; elsewhere beach sands locally consist of tiny, un;7xolrerr pelecypod shells forming coquinas. Some shallov-water 1.irne sands, espec-ially on the tops of patch reefs, include larze amounts of tj.11~cc,rnl frag!iients; and in waters more than 80 feet deep,

sands are formed a.Lmst ex-l~isive;.. of framents of 8a!.iz1eda.~ ~~ (cal.cereous ., - ,~ ~- green algae) or of the hemis:>heri.cai pale-olive foraminifer Pmghistegina lessonii . Pmice is abw-&ant on -C?ebackshore beaches of mast islands. Muck of it is of granule or small-pebble size, but a few fragments are as much as 6 inches in di.m.ie'ter. Most cf -the pumice is light gray; some is black. Large specimetis conunonly are well roun2ed. This material is so concentrated on some backshores as to form essentiallj pure pumice layers. Lime muds include (lj some small depgsits of silt (1/16-1/256 ma) derived from shells or ot!~er organic stri~c-taresand (2) extensive deposits of very fine pale-olive oozes vl~ichare s-ticlry when wet, The silts develop largely at moderate depths in lagom channels or other plcces where there are curren-ts strong enough to gather and transport them. The calcareous oozes form only in the deepest parts of the lagoon bottom. They are composed mostly of clay-sized payticles of calcium carbonate but include perhaps 10 percent of Foraminifera, largely of very small species not fouiid in the shallower waters (discussed under "Sedimentation in the l.agoon" ) . This calcareous ooze has high plasticity and contaim a slimy residue of organic matter conspicuous for its fetid odor 1.7hen wet. The minerd composition of Kapingamarangi sediments as determined -11 -- -1/ Analyses by A. J. Gude 111, U. S. Gedogical Survey. with X-ray diffractorneter patterns shows wide variation depending upon the type of organism that dominates any particular samp1.e (table 3). It is significant that the two comon species of shallow-water Foraminifera are composed of essential.ly pure calcite, and the common foraminifer 'oe1.o~ depths of 100 feet, $nphis.tegina les~or&, is of nearly pure calciLe. In coxtrast, corals including ---Acropola and other cornon genera are virtually pure aragonite except for a 1ittI.e calcite in their 'dead" interior portions. Coral- line algae of the genus Porolithon, which grow in ab~nbanceon the algal --- - . .- .- ridge of the outer reef, appear to be formed cf calcite having a space- lattice diffepent frcm that of normal calcite and interpreted by Gude as resulting from high magnesium content. Green algae of the genus ll$ims%, rihich form important coritributions -GO Lagoon sediment behw 80 feet, contain about 98 percent aragonit? together with small. amounts of calcite. The bottom calcareous oozes are about three-fourths aragonite and orie-fourth calcite.

-----..-Sedinentary rocks.- Carbonate rocks of various types form the outer reef, parts of the islands on the outer reef, and patch reefs within the lagoon. Four mappable varjeties of these rocks are recognized, based on texture and s-tructure. These are (1) coral and coralline algal limestone, (2) aphanitic to stratified elastic limestone, (3) coral rubble limestone, and (4) boulder conglomeratic limestone. The first and second types canno-t everywhere be distinguished because of recrystallization and other modifi- cations resulting from secondary processes; the third and fourth, in many places, grade from one into the other. Nevertheless, an attempt to recognize and map these types must be mad.e 2.f the processes developing the atoll are to be understood.

Limestone from corals and coralline algae forms the framework of the outer reef and of the patch reefs and cm be seen developing today wherever contributing organisms are growing, especially along the outer margins of the reefs. The rock developed from these organisms forms the pavement on the present bevelled reef surface. It is difficult to examine in section, however, because of its extreme resistance to breakage and because of the scarcity of exposed natural cuts into it. Samples from immediately below the pavement surface commonly sho\~relict structures of organisms, as might be expected, but many appear aphanitic as a result of secondary processes acting on the calcium carbonate. Although hand specimens comiiiiollly appear to be low in porosity, the rock mass is cavernous and con~ainsmany cracks, some of which are filled with clastic debris.

Aphanitic to stratified &astic limestone forms an appreciable part Of the bedrock on all islands and occurs in many places on the reef flats as pedestals and raised ridges considere? to be relicts of former islmds. Where the rock is aphanitic, little can be determined concerning its origin. Where planes of stratification remain or are etched out on weathered surfacss and characteristics of grain are preserved, however, the history of the rock can readily be determined. Dips of cross-strata and other characteristics show close similarity to features of the unconsolidated deposi-ts of island beaches

Boulder conglomeratic limestone and coral. rubble limestone are major consrituents of most islands on Kapiiigamarangi Atoll and form conspicuous ledges and shelves along many of the seaward shores. The limestones are easy to recognize because the gravel within the lime matrix normally weathers into prominence and, in some places, is extremely conspicuous because of color contrasts. Boulder limestone includes angular blocks of reef limestone, rounded coral heads, and masses of coralline algae of many sizes and shapes similar to those in modern boulder ramparts. The coral rub'ole limestone is formed entirely of small rubble in a lime matrix. These clastic limestones are distinguished in mapping, as are their nonlithified equivalents, because of the different origins that they imply.

The varied mineralogy of different limestones on the atoll doubtless is in part due to secondary processes of recrystallization and replacement. Many of the differences, however, are directly attributable to dirfei-ences in the source materials. Foraminifera of which some rocks are composed are calcitic, corals in other rocks are aragonitic, and algal deposits in still others are high in magnesium (table 3). Therefore, fundamental genetic differences may account for the different compositions of many of these youthful limestones.

On a few islands phosphorite rather than limestone locally forms stratified rock. Phosphorite, presumably developed from the guano of sea birds through replacement of calcium carbonate, occurs on parts of Fumatahati, Nunakita, and possibly a few other islands. It is an earthy, light-colored rock resembling Lhe local limestone in texture and structure, but commonly lighter in weight o5ring to high porosity. Mineral studies by Gude 'with X-ray diffraction patterns show that in some specimens the entire rock is composed of a variety of apatite. Table 3.- Mineral composii;ion of Kapin@marangi sedimeiits as dibermined by X-ray diffraction patterns . .

*Calcite A is interpreted as normal calcite; calcite B as magnesian calcite.

-15- SOILS Genera.1 statement. - Soils of Kapinganiarangi Atoll are necessnrily j.our,g: the islands oil wlij.ch they occur are of recent origiri. Most of the soil profiles are cl.assed as A-C profiles and consist of materiais little altered from the parent rock or sediment, cove'ed by or mixed with varying aniour~tsof dark humus. The soils show few fea4tures of well-developed hori- zons such as norma.l.ly result from long periods of deccmposit,ion of varied source maberials . Kapinganarangi soils cliaracteris-tical1:r are well drained and alkaline, They are largely mechanizal mix:ures of organic material and gravel., sand, or fine calcium carbonate particles. They vary from dark bro;m to gray ad creamy white, accordiag to the proportions of various consti-tuents. Soil profiles all have dark layers at the top and light-colored bedded rock or sediment below, buc some layers axe separated by transition zones of in'Ler- mediate color aad composition, vhereas others are marked b;, abrupt change.

In order to obta.in quantitative Cata on the soils of [email protected] Atoll, soil profiles were measured and samples for mialjrsis were collected from eight wells, dug in connection iiith ground-vater st~dies,and f'rom eight test pits. These profiles were dis~ributedonseven islands as follows: Wema 3, Taringa 3, Parakahi 1, Pumatahati 3, Ringutoru 3, Tokongo 1, Rilru- manu 1. Conclusions resulting from studies of the soil profiles and of the anal.yses of samples constitute -the basis for most of the following discus- sion on soils.

Parent materials of soils.- The simplicity of soils on the islaads of Kapingamarangi stems, for the most part, from the fact that they are formed aln1os.t eatirkly from- only two basic ingredients -- (1) parent; rkk or sediment composed of calcium (and some magnesiunl) carbonate, and (2) humus from vegetable matter. The carbonate material varies considerably in physical form. Much of it is limestone which makes up all the bedrock; tke rest of it occurs as unconsolidated sediment inc'uding gravel, lime sand, and lime mud. Large clastic fragments include blocks broken from the reef, coral heads, and masses of corailine algae. Small gravel is almost en-tirely coral ru6hle. Lime sand consists of shell fragments and tests of Foraminifera, with local contributions of the alga &limeda and other organisms. The lime mud appears to 'be principally from the decomposition of stratified rock or of limestone blocks. All these clastic .materials mixed with 'numus are relatively little decomposed even though they are in an area of prevailing warm, hwid climate. This iudicates that the soil is very youthful.

Variations in the soils of Kapingamarangi are in part due to differences in the amount of original contamination and in downward filtration of vege- Labie carbon into the calcareous host matecials. High perraeabil.ity of much of the sediment, allowing rain water to enter readily, the penetration of roots as fourrd in most test pits, and the work of various animals all contribute in varying degrees to domiriard xixing. In some places a "transition" zone (A~horizon) in which small percentages of carbonaceous matter are mixed with coral rubble or lime sand has develoged below the normal zone of incorporated organic rua.tler (.A,, horizon) to a distance of' 1.2 to 3.5 inches. In other places no "transitioh" zone is present.

Appreciable variations iil the chemi.cal character of the island soils (table 4) were not detected in the present study. In general, the youthful- ness of the parent limestone vLrtually precludes the possibility that any a'gpreciable concentrations of rioucarbonate miiierals have &evelope6 throvgh leaching or decomposi.tion. On the other hand, local addi-Lions to the soil. may consist of puiiinice, bird guano, shells of crustaceans and echihoids, or skeletons o? fish; the overflovs of sea water possibly have affected the soil in some places. Bacisshore beaches composed largely or entirely of pumice (table 5) are present on the lagoon sides of sevcral islands, and some pwnice has been found in test pits, but its influence on vegetation in general is not kno+m. Bird gu.ano has contribu-ted to the local development of phospha;tic soil on some islazlds, especially Pumataha'ti, where accordkg to ::a.tive reports many frigate birds formerly roosted. Salts from sea water and sea spray that locally make soils high3.y saline have not ]..eft any conspicuous record of saic crusts, probably because evapora.tion is low as compared with coiitribu-Lions of fresh water; the salts d.o .lot seem to have affected l.nfge a.reas. Alteration processes .- Factors conspi.cuous in bringing about the a,ltera- tion both of soils aild7f the parert materials of soils on Kapingamarangi Atoll include' %hose that remove substances, those that add substances, and those that mix substances. In the first category, one of the most important factors appears to be rain. Showers on the iclasds normal1.y are short but violent, and permeable sxrfaces allow most of the later to enter readily with a fl.ushi.og effect that prob3bly accounts for the gelera1 lack of saline residues. Also signifizant is the abuidant evidence of solution work in limestone beds, probably largely the result of carbonic acid from piants, which enables the migrating waters to dissolve car6ona6,es. At the low lev-el,s in the centers of some islands, especially where pits for growing puraka 1 have been excavated by the natives, ground water reaches ---.- 1/ A plant with a tuberous, starchy root, related to the taro. -.-.-- -.-.-- - the soil level and, locally, appears at tie surface of the grounu. Where this occurs, normal oxidation of organic matter is retarded or stopped by the water &d black mucks develop. The fresh-water lens, moving up and down in response to tidal fluctuations, appareiitly has a considerable solvent effect on calcium carbonate in the soil wherever it comes within range, as shown by the calcium bicarbonatecoiltent of water samples from the wells. Extraneous but significant elemenk in the soil include nitrogen, added by legumes and some tmes of algae. Calcium carbonate introauced through evaporat,ion does not seem to be important, for no caliches such as occur in arid or semiarid regions were observed, and fresh vater above the water table aspears to be removing rather than conkributing calcium carbonate. A process that is especially important in the forming of soils through the mixing of..materials is the burrowilg of certain animals. Especially , . conspicuous are: holes resulting from the activities of la.nd crabs. System- atic studies of the di~trib?!~tioniand effects of these animals on several. isle.!ids were made by ~illiGnNlering of the 1954 Pacific Science Boaud party. Probably also in~portantthough less apparent are -the borings of earth worms; observations 05 the distribution of ihese were also made by Niering. The effects of roots in peiletrating qiid breaking U~Jsoil mate?ials vere see3 in the sides of all ve1.l~aid test pits. Roots are effective only in tile vpper 1 or 2 feet of soil and sand, however, fo;. their abundance and size diminish rapidly with depth. Oaly a few were observed as deep as 4 feet.

Relation of soil to position on island.- Because seaward and ls$;oo:ltiard --- -.-- --. -.--- - -parts of most islands on Kapinxamaran?i- - - P.toll differ corisiderab:L:ji in age,-. the stage of soil development oil each side liiceriise is varied. On larse and medium-sized islands the seaward sides are composed largely of 1il;;es.tone bed- roclc,.believed to be relict fromaeposits of a foraer and higher stage of sea level, overlain in most places by rubble. In con+,rast, the lagoon sides of these islands are formed of relatively recent beach and bar deposits such as are developing today along the lagoon shores.

Excavations representative of 'the seaward and lagoonward sides and the centers of islacds demonstrate tire d:iffe?er;ces ip their soil profiles. Wells and pits. locate6 approximately 100 feet from the lagoon beaches of Ringutoru, Taringa and Werua Islands show soil lajrers of 24-, 2- and 5-inch thickness, respectively (figs. 12, 14 and 15). Soil layers measured near the centers of the same islands are 8, 11 and lj inches thick, and on the seaward sides 8, 14 aild 32 inches. Small islands like Rikunanu and Tokongo, composed largely of bedrock from an earlier stage of sea level, probably correspond in a.ge to the seaward portions of the larger islands. Commensurate with this age, soil layers of 29 ;inches on Rikumanu, and 18 and 7 inches on TO~OI~~Owere noted in test pits..( fig. 13). ,

Buried profiles occur 'on the seaward sides of at least three: islands and may be expected on others in corresaondirig positions. On Ringutor:u,-,Taringa, and Werua (~i~s.12, 14. and 15) the soil profiles include a relatively thin upper layer of dark-brown soil, separated from a lower, much thicicei, soil by 9 to 24 inches of light-gray sediment, iilcl.uding sand, coral rubble, and small amounts of humus. This light-colored layer is interpreted as repre- senting an interruption in the soil development process, during which t,ilne waters deposited elastic sediments. It is postulated that this deposition was a result of sheet wash at a time of storm activity. Much sand and son;e rubble, but no' very coarse inaterials, are included. Judging from the appreciable thickness of the overlying layer of soil (4-9 inches), which is comparable to that of the entire soil layer on the other side of the island, and from the fact that buried profiles have r1o.t been found on the lagoon sides of any islands, a period longer than that required for development of the present, lagoon side is considered probable-for this interruption in soil formation. Thus, although the buried profile may be the result of some recent storm, more likely it dates back .to the time of an earlier sea-stand. . . Extremely youthful soil occurs locally where two islands have, within historic times, become joined through sedimentation resulting from causeway Table 4.- Elements in Kapingamarangi soils determined by semiquantitative spectrographic method. Analyses by Paul Barnett, U. S. Geological Survey. Elements reported by percent ranges. Depth Well (inches) Si Al Fe Ti Mn P Ca Mg Na B Ba Cr Cu Pb Sr .01 .02 .1 Not .002 .1 .2 .001 .01 .0005 .OOl .a02 .5 ~erua#5 4 .02 .05 2 found .005 5 )lo .2 .5 .002 .02 .001 .002 .OW5 1.0 .002 .002 .005 Not .0002 e5 .5 2 -001 .0005 .a02 .0001 Not -5 Werua 6 20 .cq .oo5 .ole found .ow5 1.0 1.0 .5 .002 .001 .0005 .a02 found1.0 .O1 -01 0 Not .0002 Not .5 .2 ,001 .002 .om2 .coo2 -0002 .5 'nierua #5 31 .02 0 -05 found .OW5 found l.O .5 .002 .005 .0005 .COO5 .0005 1.0 .002 .1 .2 .QOl ,002 .W2 .0005 .2 < .Ol '02 .05 Taringa #2 5 -05 c.005 .oo5 10 >lo 2 .5 .002 -005 .0005 ,0010 0 5 .2 .2 -001 .a05 .om1 .0005 .5 c.01 1::; (.005 0 0 >10 Taringa #2 20 .005. - .5 .5 .OO2 .0010 .0002 .0010 O 1.0 .002 .001 .2 ,2 .001 .0005 .0001 .0002 .5 <.01 <.005 0 0 >10 Taringa #2 39 .a05 -002 .j -5 .002 .0010 .0002 .0005 O 1-0 -002 .005 <.oo5 .0002 .5 .2 -001 .0005 .0005 .0002 * 5 <,01 Taringa #2 55 .005 .OlO .0005 ' >lo 1.0 .5 .002 .0010 .0010 .0005 O 1.0- .002 .002 .0005 .2 2 .oo1 .ooo5 .ooo1 .0005 Taringa #3 2 "01 -005 .OO5 "Oo5 - .0010 O >lo -5 .5 .002 .0010 -0002 .OOlO O 1:; -002 -001 .5 .2 $002 .0005 .0002 .5 Taringa $3 8 '.O1 .005 ,002 <.0°5 0 O >lo 1.0 '5 .>O> .OOlO .0005 0 1.0 .0005 2 ,-remi COP .go05 05 < .01 '01 <.005 > 10 C . Tzringa #3 18 .02 .02 .0010 5 . .5 .002 -005 .0010 .0010 .002 .002- ‘ '5 .2 .00l .coo5 .0001 .0005 c.01 .005 .005 c.005 0 .5 Taringa #3 34 O >lo 1.0 .5 ,002 .0010 .0002 .0010 O I,O The folloving elements were also looked for but not found: Sensitivity limit Elements .00005 ng .0001 Be, Yb .OOO~ Co, Ge, Ni, Pd .001 Bi, Ga, In, Mo, Nb, Sn, V, Y, Zr .003 A,~t .005 Cd, Er, Gd, Ir, La, Os, Re, Rh, Xu .01 Li, Nd, Sb; Sm, T1, W -03 Te, Zn .05 As, Ce, D, Hf, a Th, U .5 K 1.0 K? Table 5.- Pumice analyses; elements determined by semiquantitative spectrographic method. Analyses by Paul Barnett, U. S. Geological Survey. Elements reported by percent ranges.

Gray pumice Black pumice Gray pumice Black pumice lement* Hare Island Element* Haze Island 1 Si .0001-.0002

A]. .001-.002 Fe .ooo5-.oolo Ti .oo5- .ole Mn .001-.002 Ca .005- .010

Mg .005- .OlO Na .0002--0005 K .OoO5- .0010 B -02-05 Ba .001-.002 Be .005-.010 Ce .ooo5-.o010 Co .02-.05

qhe following elements were looked for but not found: Ag, As, Au, Bi, Cd, Dy, Er, Gd, Ge, Hf, Hg, In, I i N 0,P Pd, Pt, Re, Rh, Ru, Sb, Sm, Sn, Ta, Te, Th, T1, U, W, Zn. RlNGUTORU ISLAND

(105 ft from lagoon) (132 ft. fra ~eo) W end N. Of center 1 end

EXPLANATION ...... m . Humus ...... _: ...... C0rOl rubble ...... /-J ...... _..., Lime sand ._...... Q3 ..... Limestone ...... -.... Scale

PUMATAHATI ISLAND

(16'2 ft from sea; 230 ft. from lagoon) W. center Center E. center

P. 0'

...... 0.. . ._ ...... 0 u...... FIGURE 12- TYPICAL SOIL PROFILES 1N TEST PITS ON RINGUTORU AND PUMATAHATI ISLANDS RIKUMANU PARAKAHI 40 ft from shore. 136 ft. from lagoon shore

EXPLANATION

Humus m Limestone f ragments

Coral rubble

pLJ Lime sand IT3 Limestone

TOKONGO 100 ft S.logoon. 60 ft. S. lagoon

FIGURE 13.- TYPICAL SOIL PROFILES IN TEST PITS ON RI KUMANU, PARAKAHI AND TOKONGO ISLANDS construction. Several illustrat~.onsof this process, in various stages of development, axe irnom; and the natives are fu3.ly awa!ce of the poss;'oil.S.tles of increasing the size of their islands by this method. Oile of the best examples of sedimen-ta,tion resulting from the process is near the south end. of Hare Island, where a small islet has been attached and completely incorporated into the main island (fig. 8). Eemaiits of a stone causeway about 15.0 feet long are still visible in the middle of the area though sediments have accurnul.ated to a considezable height on both sides of it. Seaward from the causeway an area about 300 feet Ion,? has been entirely filled in by natural processes, locally wi-th sand and elsewhere with mixed sand acd rabble Soil has not yet developed. On the lagoon side sand aad rubS2.e have been piled up on the shoreward part, leaving a iiepressed area inland toward the causeway; this low area is swampy and already has accumultited some black sticky mud.

According to soil ncmenclatwe used on okhe; atolls of the Pacj.fic (stone 1951, p. 19-37), most of the soil at Kapingamarar~giprobab1.y should be classed as belonging to the "Shioya soil series." This is especially true where the soil has developed in the very young seaiments on the lz,goon sides of islands. In some ?laces on the older, seaward parts of islands, development has gone beyonu the Shioya stage and greater, more concen4Lrated a.ccumulations of organic material have resulted in a darker soil which pe~htipshould be assigned to the "Amo soil series." In these two soil types there is great textural variation both vertically and laterally. Such variations, as shown in test pits (figs. 12 to 15), include orgarlic materials that are relatively pure and others that are mixed with lime sand, coral rubole, or limestone blocks. Because nearly all the soils are developed on surfaces aone the water ,table that allow rain water to pass thruugh rapidly and that are conducive to relative dryn~ss,mangrove swamps do not exist and muck or peat is scarce. Only puralra pl.'is end a few bomb craters where water acc~~ulatesgermanen-tiy have an enviroment favorable to these soils.

Many illustrations of correlations between plant indicators and types of soil and sediment are apparent on Kapingamarangi. The constant associa- tion of Scaevola with sandy beaches and'of --- Guettarda with coarse rubble ramparts or with smal.1 islets covered by rubble are examples. This subject has been studied in detail by William Niering and will be described by him.

DEVELOPMENT OF PHOSPHORITE

Phosphatic soils and rock phosphate or phosphorite occur at numerous places on various isianas of Kapingamarangi Atoll. The soils appear to be very similar to ather soils that are composed of lime sand and humus, but X-ray diffraction tests show that mineralogically they contain apatite in addition to calcite and aragonite and that a few samples consist of apatite only. The phosphorite or rock phosphate resembles in texture and structure the limestone from which it developed, but commonly it is very light in weight owing to high porosity. In many places it is various shades of brown, and in general has a powdery, earthy appearance. The ori.gin of the phosphorite from the &ino of birds is attributed to pho~phori~ationof salts in the guano,. inyolving reactions with forminiferal or coral. limestone. According to Aso '(1953;p.'19) the principal components of this tne,of phosphate normally are secolaary ca1:iurc p?.ios:,i~ate, CaXFO4; tertiary calcium phcs:~haiaSe, CajP2U~;,,and cakciuii carhona-te: present in varying ratios. ':il:us, the phcspha~epLcbabl$ represents a s.'.a7;e intemediat,e betveen fresh glans and tertiary: calciuii tke coaversion taking place under the ind..uence of orga:iic acid and carb5nic 6cid gac. The ease with which ca~.cium,carbonateisdissolved by phos~hatic'so:l.utionsis a significant factor in this process (.!so, 1953, p., 17'): . . . . The distribution of ph3sphati.c soils and .p$osph;>r:itc at Kapingamarangi 2nd the concen~;.ati:on.,of these .p-ociucts (tabr to have a sigrnifi- can-t relatioiiship co tk:.e history of island dev.elq!n,?:~;. i!.t F;iii.a-taha-Li Island, test pit 1 rear the island center shows cqj>l.e-te phos:>horiza:tion of 1.imestone in both :soil and stratified rock 50 a depth of 2 feet or more (table 6). Test pit 2, farther east on 'the s&e island, shows o11l.y a little apatite, pcesent, and tha:t is near the surface. Nevertheless, the fact that surface soil is ar'fected. inboth -places sug:;ests recent development. High concentration in tke central area is expectable; for. this. island. is reported by the na-Lives to hkve been the roosting place of i&rge iiumbei-s of frigate birds for man; year;. . . Varying concen.trationsof. phosphatxic soil are recognized on several

' :islands: other than Puiatahati. These irlcludeTari~ga(table 6) and ld'iJerua (table 4) and probably others for which analyses have not been made. he lccalization of apatitenear the soil surfaces on these islands sugxests that its development may be relatively recent. Phosphorite comprising the stratified rock onseveral other islands, however, clearly is ofconsiderable age, i .e., formed before the i.slands. had migrated to their pyesenc dositions . On Ringutoru and Rilrmanu, fqr instance, high-level remants' of stratified rock considered to antedate the latest fall in sea. level, and stand.ine; above areas of soil on the island margins, are highly phosphorized. On Hare, Iuunalrita, and Ri1:-umanu, stratified rock that today is on the seaward shores, beyond the ou.ter fringeof vege-Lation, and is periodically covered by high- tide waters, likewise is high in apatite (table 6): .These deposits arg interpreted as rel'icts from a time when they were in island interiors, such as the frigate birds inhabit today, and when birds in numbers frequented the areas...... GROrn kITEX .. ,. . General features.;- The presexe on most small islands of fresh or brackish gFound water floating on the relatively heavy salt riater within island rocks and the lens shape normally assumed by s-~chbodies of fresh water have long been recog~izedby. hydzologists. . The riame "Ghyben-Herzberg lens" com- monlyis used in referring to these lenses of. fresh water (stearns and Vaksvik, 1935, p. 237-:239.).

. .. ~ Nost of the islands on. Kapingamarangi. A.boll contain ground water of varying degrees of freshness. On the larger islaxis this water is potable .e 6.- Percentages of ayatite as de-termined by X-1-a y diffraction analyses. Data from .Janies Gude, U. 8. Geological Survey.

\patibe 3ercent; Com~ents Island Stra+igzapMc- -- Position .- - - - -.- fii!;torT- High-level rermant 50 Relict bedrock., Rikumanu High-level remnant 75 Relict bedxocls. IRilwanu Stratified shore deposits 20 Relict bedrock. ilrumanu Stratified shore deposits 40 RelicC bedrock. Iunakita Stratified shore deposits 30 Relict bedrock. %ratified shore deposits 20 Relict bedrock. Llt ahati Soil, top 100 Pit 1, near i i.sland center. IPuma-tahati Upper bedrock 100 Pit 1, near island center.

Fmatahati Lower bedrock LOO Pit 1, near island center.

Pumatahati Top soil 2 Pit 2, eastern interior.

IPumatahati Transi-tion soil 2 Pit 2, eastern interior.

Pumatahati Unconsolidated 0 Pit 2, ea.stern interior.

Taringa Top soil 50 Well 2, near island center.

Taringa Transition soil 0 Well 2, near island center.

Taringa Coral rubble 0 Well 2, near island center. and of sufficient quarltity for domes-tic use. The natives have dug wells pene- trating it in a number of places, bwt in general they prefer cistern water for drinking. This is fortunate, for the wells near the places of habitation are zpt to be polluted. In the village on Touhou, water from the community well is used largely for washing purlsoses; much of it is carried in buckets to bath houses but some is used in the immediate vicinity.

Because of drouths the supply of cistern rmter becomes low from time to time and locally is'exhausted. At such times the ground water of the islands constitutes an impor-tant supplementary supply of potable water.. Factors that control the location, quality, and amount of this water are similar to those that have been described in connection with the ground water of other atolls (COX, 1951; Arnon, 1954), but distinctive featires of climate, lithology, and te?rain are responsible for local variations and make desirable an analysis of the local water problems.

Objectives--. of investigation.--- Ground-'water studies on Kapingamarangi ktoll, made periodically during June, July adAugust of 1954, wei-e conducted on four islands -- Taringa, Werua, ~araliahi,and Hukuniu. These islands were selected partly because of accessibility but largely because they were considered representative of the principal island types. Taringa is an example of a moderately long but narrow (600 f.t.) island with stratified rock forming the seaward shore and sand beaches the lagoon shore. Werua is similar but considerably wider (1000 ft) and with more variation in topographic character resulting from larger size. Parakahi is a small island (400 ft x 300 ft) the surface of which is formed entirely of ulconsolidated sediments. Hukuniu is a very small island (250 ft x 150 ft) in which stratified rock forms all of the shores and also the surface, except for a thin mantle of humus in the interior.

The principal objectives of the ground-water investigation were to obtain data on (1) the depth to water in various parts of the islands, 2 the amount of change in water level as compared with tidal changes in adjoining sea and lagoon, (3) the amount of lag between the extremes of tide and the corresponding changes represented in well levels, (4) tine effects of permeability in different types of rock and in unconsolidated sediment, (5) the variations in ground-water lenses resul-Ling from differences in size of islands, and (6) the quality of the water resulting from various situations. All these features have a close relationship to the availability and usefulness of ground-water on Kapiugamarangi.

Methods of study.- A series of eight wells -- three on each of the large islazds and one on each of the small islands -- were dug by native labor to a depth below .the lowest level of the water table for each area involved. Records were made of the types of sediment penetrated (figs. 14 and 15) and a staff gage was installed in each well for measuring water fluctuations. Three times during the summer hourly reaaings were made for 24-hour periods to obtain i'rom this series of wells relative measurements of the times and extent of water fluctuatioi~s. Some water samples were analyzed in the field for salinity and pH. Others, representative of all the wells, were brought back to the U. S. Geological Survey in Denver, where analyses were made by John D. Hem. EXPLANATION a 3.Seaward side Limestone rn - Coral rubble Calcareous mud

Sand: Forminifera and/or shell fragments I. Lagoon side

..... -' ' 'July 30-31 -_ - -Aug 20-21 June 23-24 -I- 0 0 -

0 0

FIGURE 14- LITHOLOGY AND WATER LEVELS IN WELLS ON TARINGA ISLAND Rfl 6 Seaward side yEzJ yEzJ Limestone Coral rubble Colcoreous mud a,. Sand: Forminifera and/or shell fragments

5 inches

Logoon side

5 lslond middle

-Aug. 20-21 "July 30-31

High woter levels

Low woter - level

-- FIGURE 15- LITHOLOGY AND WATER LEVELS IN WELLS ON WERUA ISLAND The vells on both Tari.nga and Werna Islarids were dug in e row across tile islands from west (lagoon shore) to east (seaward shtre) a.nd alao~ztu~idvay along the lenqkh (noi.th.south di;nension) of each island. Locations of weils are shom 011 maps (figs. 4 and 5). The spacing of the three wel&.relati~e to the coasts on these islarlds was as follom:

1 2 Lagoon to well-- 1 Well- to well 2 7Jel.L to- well. 3 We1.l 3 to sea ' ST---Tarlnga 68 st.. I-- 2; rt Ii 160 ft \ 125 ft

The single wells OQ the two small islands were dug near the centers. Lo- cations were as foll3ws: From lagoon From sea (w) side (z) side Frm north si&e From south side - -.. ---- . -- -. l~aralcahi / 136 ft 2 t r 145 ft 1

The riells on Tal3nga and hlerua Islands ranged in depth from b feet 2 inches $0 7 feet 1 inch, dependiug on the depLh .!LO grour?d-water at its lowest stage. The t~iodeepest wells were those adjacent -to the seaward coast. They weye not located on boulder ramparts but OIL lower ground inl.and from the ramparts. These wells were the most difficult to dig, for stratified rock vas encoun-te~edin each a little more tinan halfway devil.

--Depth to p,round ...- water.- As stated in the Ghjbec-Herz'uerg principle, the hi&est pcrt of the fresh-water. lens within the rocks of most small islands is sonev7hat higher than sea level adjacent to tne island. Because the ratio of fresh- to salt-water density that controls this lens is approximately 40/41, it follov~sthat the elevation of fresh water above sea level. is slight on islands the size of those on Kapingaxarangi. CO:~ (1351, p. 22) and Arnow (1.954, p. 3) show that an average height of at-oi~ta foot is -to be expected on the islands tinat they examined in different parts of the Marshall group.

Equipment was not adequate to measure the elevation of ground-water lenses with respect to sea level at Kapingamarangi, It seems safe to assume, however, that the elevation on any of those islands does not exceed 10 to 12 inches and in most places is far less. Thus, the water level on those islands may be regarded as roughly equal to sea level, and the depth to water in any well is approximately equal to the elevation above sea level of the surface at the well site. Because the lagoon margins of islands commonly are s1ie;ktly higher than the island centers and because the seaward margins are considerably higher than the lagoon margins, the depth to ground water normally is greatest near the seaward margins of the islands and least near the centers. Figures 14 and 15, using low-vater level as a datum plane (not allowing for the increase in elevation in the island center that results from the lens effect of the fresh water body), show the following depths to the h-ighest recorded water level: Lagaon side 1-- -Middle island-- Seaward side 43 inches 43 inches \~'-~K'T~CZZ---.~ I ~erua 1 @ inches / 29 inches I 66 inches 1

The figures in the preceding table probably are representative of de2-ths required to reach the upxr limits of water on all -the larger isl.a.nds at Kapingamxrangi, but to reach the level of permanent water (I&-water level) an additional 8 to 12 imhes is required. Further, if wells are to be used for domestic purposes, allowailces must be ~mdefor normal drawdown, for lowering due to peri0d.s of little recharge and for adeq~~.aLestorage, so ad- ditionai depth is necessary. The d.ep-th should riot be greater than necessary to achieve the e,bove o'bjectives, however, or selt-water contamination of the wells may occur.

On the veiy small islands at Kapingamarangi, the water table cmnonly is closer to the surface tnan on Taringa and Werua. This is due to loiier topography. Hovever, the amouut of fresh ilater (if present at all on these small islands) normally is much less and tine degree of salinity higher than on islands with large recharge surfaces,.as shown by the wells on Parairahi and Hukuniu Islands.

Variations in water level.- Fluctuations in water level in islands like ----ap-. - those of Kapinganarangi Atoll are principally of two types. In one the changes are gradualand arethe re'sult of.:increases or decreases in tine amount of recharge; in the other the movement is observable from hour to hour and results frorn rising and falling tides' that cause the salt water under the fresh water lens to go up or down. Fluctuations due to gain or loss in recharge were not detected duriflg the period spent OJ Kapingamarangi Atoll, though doubtless such changes were coiitinuous1.y affecting -the water levels of all islands, especially the very small ones. On the other hand, significant aata were obtained on the effects of tide on ground-vater levels

Figures lli and 15 show the total rise and fall of water levels in the wells on Taringa and Werua Islands, respectively, as shown by readbgs taken oa Jilrie 28, July 30, and August 20. Tile low-vater level for each well, as measured with respect to the well bottom, r;ia$ essentialiy the same for the 'three dates. The high-water-levels, however, show ma.riced differences, especially on Taringa, and these differences are directly related -Lo tidai fluctuations as indicated by tide records. The amount of rise in the two wells located in island centers is slightly to moderately less tha;n that of the r,~ellson the corresponding island margins, This difference v.ncloubtedly resuLts frorn incomglete readjustment of the fresh-water lens, due to in- creased friction from distance of travei, during the short intervals between tidal changes. It supports the principle, pointed out, for example, by Cox (1951, p. 1,that fluctuations in the water table of small islands are inversely proportional to distance from the coast. Lag in rise and fall of ground water.- The elapsed time between attain- -.--- -.-.- - ment of high- or low stage- by- %he tide in waters adiacent to an island and tine attainment of a corresponding high or low level by the ground waters in the island is variable according to the size of the island, the permeability of rock or sediment invol.ved., and other lesser factors. In any eveiit, this elapsed time is represented by a definite lag, vhich on Taringa Island was 3 to 4. hours and on Werua 4 to 5 hours (figs. 16a, 16b, 2.62, 17a, -17b, 17~). On the sma'll island of Hukwiiu, however, there was vir-Luaily no 3.ag between the time of high tide and high-water level (fig. 18a and 18b). This is interpreted as pyinlarily the result of high permeability, as wiil be discussed later, rather -than merely the small size of the island.

Analysis of the charts showing tide and well fluctuations (figs.' 6, 17, 1%) shows that the lagoon tidal variation is consistent1.y greater than the variation on the seaward side of the islands, but that in the areas studied, the high and lor? tides arrive inboth places at about the same time. Water levels in the wells on each island. 1i.kewise show differences in. amount of fluctuation, but on each island they reach their peaks at approximaLely the same time. Tnese data Suggest that differences in permeability as riel1 as in distances from the shore and in tidal fluctuation at the shore are responsible for differences in tne mount of rise and fall of the water 'cable in any particular part of an island. Thus, in these small islands the fresh- water lens acts as a unit insofar as the time of rise and fall is concerned..

Relative nerrneability of sedimentary n1ateriais.- An asymmetry in the "- ----.------.-- permeability of atoll islands has been' indicated in the work of wmerous geololjis.ts (COX,1951, p. 19). In most islands the lagoon shores are composed of beach sands, wheream the sea,%rardshores are comp~sedof stravified rocks, boulder ramparts, or both. The fine-grained sedinient of the lagoon side is far less permeable than the coarse detrital fragments of Lne rampart or the cavernous limestone that constitutes much of thestratified rock. Fluctuations in water table on atoll islands appear to be proportional to the perrneabili-ty of the sediment or rock through which the water moves.

On the islands of Taringa and Werua the seavard coasts are formed of stratified rock up to and above high-tide level and of narrow boulder ramparts above the rock. The rampal-'ts are not significant insofar as the movements of the fresh vater lenses are concerned, for, as illustrated in well proriles, all of the ground-water movement on this side of the islands is within the stratified rocks. Fbrthermore, water from wells on the seaward side is definitely brackish, mostly not potable, which suggests that it enters through open cavities or cracks with Pree circulation rather than by the slow percolation that constitutes intergranular movement in the sanas. In contrast, ground water on the lagoon sides of these islands is fresh and drinkable almost to the beaches. Parakahi and Hukuniu illustrate the effects of permeability on the ground water of very small islands. Well levels on Parakahi show approximately the same lag behind tide fluctuazions as represented in wells of the larger islands. The water, which is potable, murj't enter through sand deposits from any direction. In contrast, water level in the well on Hulruniu Island, which is formed al.most entirely df stratified rock, reaches its high and low points essentially at the same 'times that high arid lo~,rtides arrive. Furthernore, the salinity .of the water in this wel1.i~close -to -that of sea water -- again sugg,estingeasy access through large' channels, , .. . Effects of island size on lens.-. Observation weils in the centers of the .-.- -.- .- .. -. - ti10 very small islands -- Parakahi (1400 fi; r: 300 ft)and HuBuniu (250 f-t x 150 ft) -- suggest that, on these islands at lea&, permeabiliLy of the rock or sediment, rather than size of the island, controls the amount of wa-ter- level rise and fall. 'Figure 18 show thaton Parakahi .the rise is comparable to that of xells dug in similar sediments on the larger islands, whereas on IIukuniu, where. cavernous, stratified rock is involved, the rise is donsider- ably greater, even approaching in amount the corresponding total rise of the . . tide. . . . . The real significance of island size in regard to the fresh-watsr problem is whether or not a lens can develop and, if developed, can be kintained in a par-t5culs.r 'area. Because continued operation of a lens dep'enls in large measure on the amount of recharge, a sizable surface are8 for catching precipitationa;.id a sufficient amount of precipitation are basic. In con- sideririg how small An island may maintain a fresh-water lens, Parakahi Island, in a region of on1.y moderate rainfall, :is noteworthy for having a lens of fresh, potable water. Hukuniu., which is still smalier, has highly saline water but this may be due as much to contamination from free circu- latioi? through open channels in limestone as to small size of the island, Thus, only a rough qualitative measure of the minimumsize requirement, furnished by the exaniple 'of Parakahi, is available for .&pirigamarangi Atoll.. . . Quality of the--.--- ground water. - Water samples .collected on August 20, 195h,?rom each of the eight test wells and analyzed by the U. S. Geological Survey are s~wmarizedin table 7

Data presented in table 7 illustrate that:well waters froid Taringa and Wema Islands are progressively higher in chemical coinponeni;s, toEa1 haru- ness, and percent sodium from the lagoonsrard to the seaward side of each island. This trend almost certainly is related to the amount of mixing with .sea water in each place and probably results from differences in permeability of island sediments and rocks. The cavernous nature of rock beds on the seaward sides apparently allows relatively greater mixing than is possible in the intergranular spaces of sands on the lagoo!mard sides.

Cchparison of water from the tr.10 small islands -- ~a.raliahiand Hukuniu -- show differericessimiiar to those noted from one side to the other of the large islands, but contrasts are greater. These con-crasts apparently also result f~omvariations in permeability and in degree of mingling with sea water. . iJat,er. from the Parakahi well illustrates relatively poor mixinj; that in the Yukun&u well has nearly the coriiposition of sea watei-.

The hardness of the well waters of '&2ingamarangi Atoll undo&tedly reflects contributic~sof calciun and magnesium fromlimestone and lime sand. The hardness increases across Taringa and Werua from lagoon to sea, reflecting an increase in the degree of mixing of ground >rater with sea water. ------maa 33% P M

FIGURE 16b TIDE AND WELL FLUCTUATIONS, TARINGA ISLAND JULY 30-31,1954

-.-.- Tide on seaward shore Well #I, Lagoon side Tide on lagoon shore ---- Well #2, Center of island --Well #3, Seaward side FIGURE 16c TIDE AND WELL FLUCTUATIONS, TARINGA ISLAND AUG. 20-21, 1954

Tide on seaward shore Well # 1,Lagoon side Tide on lagoon shore ----- Well #2,Center of island . . ..-..Well #3, Seaword side iches

10 12 2 4 6 8 10 12 AM.

FIGURE 170-. TIDE AND WELL FLUCTUATIONS, WERUA ISLAND JUNE 23-24, 1954

Tide on seaword shore Well#4, Lagoon side -- --Tide on -lagoon shore ---- Well #5, Center of island iches

FIGURE 17b TIDE AND WELL FLUCTUATIONS, WERUA ISLAND JULY 30-31, 1954

Tide on seaward shore Well #4, Lagoon side Tide on lagoon shore ----- Well #5, Center of island ...... Well #6. Seoword side iches 30

FIGURE 17c TIDE AND WELL FLUCTUATIONS, WERUA ISLAND AUG. 20-21, 1954

-- -Tide on seaward shore -Well #4, Lagoon side -- - -Tide on lagoon shore Well #5, Center of island ...... Well #6, Seaward side nches

A.M.

FIGURE 180-. TIDE AND WELL FLUCTUATIONS, HUKUNIU AND PARAKAHI ISLETS JULY 30-31, 1954

---Tide on seaward shore Well #7, Po rokohi ----Tide on lagoon shore ...... Well #8, Hukuniu iches

i? M. A.M.

FIGURE 18b TIDE AND WELL FLUCTUATIONS, HUKUNIU AND PARAKAHI ISLETS AUG. 20-21, 1954

--- Tide on seaward shore -dell #7, Porakahi ---. Tide on logoon shore .--...Well # 8, Hukuniu Table 7.- Analyses of water samples from test wells a~d comparative data for normal sea water. (~rialysesby U. S. Geological.Survey)

Taringa Werua j Paraltahi ! Hukuniu Normal sea water - -. :hemica1 components (pPm .. Calciun (ca)

Magnesium (mg)

Sodium (?$a)

Potassium (K)

Sulfate (SO&)

chloride (~1)

?hysical characteristics

Hardness (ppm)

Percent sodium*

*Fercentage of sodium among the principal cations (sodium, potassium, calcium, and magnesium), all expressed as chemical equivalents. These rela.tionsbi:ps are illustrated by the fact that calcium is much hi&er than magnesium in the near-l.a,goon sainples, reflecting solution of calcim caybona-Le and rel.ative3.y libtle mixing with sea crate?, :.~t~ereascalcim ex- ceccis nia.gr;esium on1.j moderately in -the near-sen smpl.es, reflecting mixing til:Lh sea water, r.rhi.c:i has a proportlorlately h:i.yher mag~.iesi~mcon-tent.

The hardness of water from wells on the Lagoon sides and in the islarid ce!-kers ranges Se.b.;eeii 300 and 70s pgm, so t,he:i are inuch higher than for soft waters ( < 50-60 p:pm) as recognized in the United States. These i.iakeel.s would require "water soi'-ter?erst' for dc:ynest-ic use in &ueyice. They are similar i'ci qisli-ty to waters from the noi-thern ivlazshall 1sland.s recorded by l::rno>i (l95li, P 7).

Water samples from test wells on Tminga and Werua 1sla:lds were checked at various times during the sm!ein for PH arid ten@erature. The PH readings rziiged from about '1.0 to '7.5, alid those from the relatively brackish waters of wells on the seaward side were corisisteintly high, apparectly ref'lec-Ling a sl-ight alisaline increase from sea :.:ater mixing. Temperahire readings for alL veil vatfrs were 2-1.5' to 28O C (81.5O-ti3O F) in the ea.ily morning, but the) couunonly rose lo or 2O C during tile warmer part of the d.ay.

:,later from >rells on t!ie lagoon margins of islands is ::enerally more potable than that from other parts and is easler to reach by di~ging, pr~~b- ably all wells thus situated will furnish water of &oad quality which, if protected from pollution, can be used to advarrtage by the Kapingans. The U. S. Public Health Service recomuiends 250 ppm of sulfate (~0:~)and tine same amoun-t of chloride (~1)as upper iimits for water used in norm1 domestic cons~ur~ption,although water considerably higher in these componeilts may be used without apparent harm by people who have become adjusted to it. Water fro= the island centers is moderately good, but that on the seaward sides is too brackish to be acceptable.

NEAR-SHORE cams Currents moving along the shores of the Kapingamarangi islan(ls are, in part, normal longshore currents generated by waves and, in part, the results of tides. They contribute to the r?orlc of erosion and deposition; also they are significant in maintaining ventilation within the lagoon, introducing and circulating new sea water with each rise in tide. The tide rises 5 to 13 inches higher within the lagoon than it does on tine seaward sid.es of the islands (figs. 16, 17, and 18), probably because the passes between islands form constrictions that limit the movements of water entering and leaving the lagoon. In order to obtain specific data on the movemelits of near-shore currents, records were tabulated on the direction and relative rates of ri:ovemen-t wi-th respect to nine islands (figs. 19 and 20). TPxee of these islands -- ihina-. kita, Torongahai and Rin~xtoru-- are in the northern section of the atoll, four --. Werua, Natiro, Hare, and Taringa -- on the eastern side, facing the dominant wind direction, and two -- ;Pumatahaa-tiand Matukerekere -.- or1 the southern arc. To observe currents, fluorescein dye, which pro&uces an orange color readily observable even at a distance, was poured in the water. Move- ments recorded at times of both rising and falling tides were plotted for contrast. LAGOON

Lowerfng tide

LAGOON --.

LAGOON

O- O- 300 ft.

O- O- 300 ft. LAGOON &. 300 ft. '?.. LAGOON LAGOON . .

PUMATAHATI MATUKEREKERE Lowering fide Lowering tide 0- 400 fl. ,' FIGURE IS.- NEAR-SHORE CURRENTS OF 2 SOUTHERN, 2 EASTERN AND I NORTHERN ISLANDS OF KAPINGAMARANGI cuuseway

4. i

LAGOON :,

' WERUA ! Wngtide .1

, *---/, ... , \ ...... , / \-. rr . causeway

TORONGAHAI Rising fide

FIGURE 20.- NEAR -SHORE CURRENTS OF 2 EASTERN AND 2 NORTHERN ISLANDS OF KAPINGAMARANGI The results of observations on near-shore currevts as swwarized in figures 19 and 20shot~that on the seaward sides of most islands tidal currents normally move in opposite directions from a cen-tral point. During rising tides, however, longshore currents may develo:, wit? sufficient strength to direct all movement in one direction. The maps also sho>~that on the concave lagoon sides, near the sand beaches, vhich are protected by horns or bars at the island extremities, movement of currents is extremely small at most times and has no dominant direction.

The most significant current movements adjacent .'to isl.antls are those passing along the sides between islands. These are especially effective during rising tides. Such curreilts consistently travel to~iardthe lagoon, moving elastic sedhents and building rubble and sand bars out into the lagoon from the endsof each island. Furthermore, between the bars of adjacent islands these curren-ts pro6uce submerged deltas that protrude well out into the lagoon. The deltas are prominent in airplane photographs and are con- spicuaus even from a boat; normally they are compbsei? dominantly of coral rubble in their headward parts and of foimniuiferal sand in deeper i,~aters beyond. Because bare reef flats and little sandoccur seaward of the Kapingama- rangi islands,- relatively little sediment has accumulated on this side.. I/ This is i!? contrast to the conditions on Nu!moro and certain other atollz where Foraminifera (~alcarinaf---- live in abundarcce on Cne reef flat -----outside the islands arid form extensive sand deposits.- shores at and below sea level are largely of stratified rock, whic'n shows the effects of pla;lation as should be expected from knotm current 'trends. Tb.e few lime sands that wash into this area normally do not remain but are moved along by currents and waves through the gap between ish3ds and thence into the lagoon. Even coral rubble from the reef edge, which is ictroduced in considerable amounts, appears to be transitory with respect to most seaward shores. Only the ramparts of, boulders or rubble that stand a'oove high-tide level and are formed by: the waves of occasional large storm form signifi- . . cant deposits on seaward parts of the islands.

' In the near-shore waters on. the lagoon sides of islands, protected from the 13revalent easterly winds by the islands themselves and from tidal currents by the horns or bars at islar$d'ex~r6iities, the, normal, extremely weak and variable currents seem incapable either of depositing or of removing sediments of the beach. Most of these beaches are of foraminifera1 sand and the species represented appear to live in theshallow waters nearby,.but

their accumulation and that of coral r&oble as fou,nd at Hare and some other '' islands probably result fkom the, occasional reversals in wind direction,. which bring waves f~o&west to east' across the lagoon. These wind reversals result in a considerable piiing up of water andsediment against lagoon

shores of the islands. ' " The island of. watahati, on the southern rim of the atoll and near the vestern end of its islands, is unique in that its lagoon beach is formed almost entirely of coral rubble rather than sand and in that it has a well- developed rubble rampart on the lagoon side (fig. 9). Both of these features can be explained only as the results of the action of large waves caused by strong winds. Furthermore, the winds must have come from east to west across the la6oo11, for the source of rubble on the rampari; and the greatest conceii-txtion of rubble on the beach are at the eastern end. Such feal;ures developed on this island. and not on others probably because of Its far westerly position on the southern reef.

SEDIMElKT!ATION IN THE LAGOOB The lagoon a-t Kaping~a~angicovers an area of about 15.5 square miles and has maximum dimensiow of 5 by 6 nautical miles (~ugent,1946, p. 755). Although it is roughl.~circular in plan, its symme-try is far from perfect because of an almost straight southwestern side. Profiles of the lagcon bottom, disregarding the many irregularities caused by patch reefs that rise from i-t, are those of a shallow basin (fig. 21). Even. with the greatly ex- aggerated vertical scale used in the sections, the slope apnears gentle, and sections of near-shore areas (fig. 22) and of Manin knoll (fig. 23), made with the same vertical and horizontal scales, demonstra'ce the low- angled slopes on which sediraeri-i;s are accumulating.

The deepest par-ts of the lagoon, in the north-central and east-cen-tzal. areas (fig. 2)are recorded as 4.3 fathoms on U. S. Navy Hjdrogmphic Of- fice Chart 5042. The greatest depth measured by the writ,er in more %hail 200 soundings was :!O fathoms (2b0 feet); however, this depth was reacheii in at least five places, suggesting a relatively fiat bottom. Beeawe of the slightly aspmetrical distribution of the deepest areas, resul.tiiig in gent1.e slopes on the south and wes-t sides of the lagoon, Nugent (1946, p. 756) postulated that Kapingatnarangi "is apparently tilted to the nor+;heast." This feature of distribution can be equally well explained in other Trays, however, and the relative narrowness of the southwest arc of the atoll as compared to the northeastern arc argues against the postulate.

In order to obtain a detailed record of bottom sediments in the lagoon, samp1.e~were col.lected systematically along many lines forming a modified grid pattern. Both grab samplers and bottom drags were employed durinz the vork, and in relatively shal.10~waters samp3.e~were obtained by diving. Approximately 250 samples, representative of essentiallj all par-ts of the lagoon, were collected. Time has not yet permitted study of these sani$.es except in a general way, so details regarding their characteristics, dis- tribwtion, and significaxe must await publication at a later date. Only a generalized statement can be made at this time.

A relationship between type of sediment and depth of water (fig. 25) is apparent in the lagoon at Kapingamaran&i. Of seven principal tyjjes of sediment that are recognized, six are restricted to relatively narrow limits in depth, forming a series of bands encircling the deepest parts of the lagoon. The seventh type of sedimen-t is a white silt, ap;3arently formed of comminuted shells, corals, and other debris; it is distrib~~ted.al.ong %t~.e paths of stronz bottom currents at depths ranging from a few ree-t down to at least 200 feet. The other sediments and tiieir general ranges are as f ollorrs : 0 F 3 Z e LL I 0 2 4 5 3 I 2 5 3 I 4* $? a (L I ,-

NORTH-SOUTH SECTION ".W 2 t" = <,- Z 0 Y w W D. i: :: eW 9 "i I ::

NORTH-SOUTH SECTION

FIGURE 21.- SECTION ACROSS KAPINGAMARANGI LAGOON / VEM/CAL EXAGGERAVON APPROX/MATEL Y X30)

I Nautical mile (6080 Feet1 O- O- r 40 Fathoms (240 Feet; Low lid8 /eve/ ...... , ...... - - ...... - ~+mrr,,_ , - -

Y V ,, /, FROM MATiRO BEACH WESTWARD

-" ____- .. Low tide level .- 30OOOOOOOODOOOOoO~ll

FROM MATIRO-MATUKETUKE PASS (N) WESTWARD

......

FROM MATIRO-MATUKETUKE PASS (S) WESTWARD

Low lids leod ____..__~ -___...... --

FROM S.W. REEF (WEST OF TIUTUA KNOLL) NORTHWARD

EXPLANATION Ezg F14 Coroi rubble Microatolls

Foiominiferai sand Branch corals

FIGURE 22- OFF-SHORE PROFILES IN EAST AND SOUTH SECTIONS OF ATOLL AND DISTRIBUTION OF BOTTOM SEDIMENTS 0 50 Feet u Verf,coi ond horizonto1 scole WEST-EAST PROFILE

NORTHWEST-SOUTHEAST PROFILE -

NORTH-SOUTH PROFILE

NORTHEAST-SOUTHWEST PROFILE

FIGURE 23.- PROFILES OF MANlN PATCH REEF 0 50 100 Feet u Verhcol ond hormntol rcole EXPLANATION Depth Z in feef u

FIGURE 25.-RELATIVE PROPORTIONS OF PRINCIPAL TYPES OF SEDIMENT AT 5-FOOT DEPTH INTERVALS AS DETERMINED BY SAMPLING Sand dominantly of Foraminifera fhphis tegina madagascariensis 1. - --- - .. dlOrbigny, Marginopixa vei'tebraiis Slainville, and Elphidirun -- . ------..- craticulatw. (~=cTtelar:d ~:;ollr/ (50 feet; mostly 0-25 - -- -. -- feet. 2. Sand dominantly of clastic shell. frabments -- 10-50 feet; mostly 25-50 feet. 3. Coarse debris, largely of dead branch corals, acclmiulated between and below thickets of livir-g corals -- 30-105 feet. Accumlations of -Halimeda -macroloba fragzents -- 90-160 feet

------Spiroloculina comunis Cuslmian .. and Todd; ;.: 2...... 20

--.---.-Virgulina comp!.ana.ta Eg,ger 10 -Textulmia - --agglutinans dtOrbigny )...... 12 . I -T. -.--foliacea Heron-Allen and ~arlacd) Amphistegina .mada?;asoariensis d'orbigny ) 10 ------,- ...... ) -A. sp. (small, flat worm) 1 "The remaining 28% is.cotnposed of gbout 32 additional species among which one species eachof Tretompha+$-, Globigerina, C~&o~gporetAa,Lo=- st%= and OperculLna make up the-bull; (perhaps as much as15 cr 20% of the whole fauna). The Globiger*~ i.s a planktonic form arid the T~zouip:n&l>.s- is an,attached form with a plankt,oriic stage. Cc~mbalopoiettaand Cibe.ci.d.es may be attached, but are not necessarily al'rrays attached. (Ul the others are benthonic forms, so far as imor.rn."

The origin of the calcareous mud is not known, and detailed studies of it have not yet been made. Evidence at hand, however, suggests &at the mud is not derived primarily from the residue of sediments occurring at higher levels. Although its composition is principally aragoni,Le, it is largely surrouiided by a bel-t of Amphistegina ---lessonii with a composiLion of normal calcite. If the aragonite were eerived from either coral or shell debris of shallower depths, the difficult;. of explaining how it bypassed the surround- irig calcite belt is encountered. On the other hand, the presence of some aragonite needles suggests the possibility that it formed as a chemical pre- cipitate. . . The overall pa-ttern formed by belts of sed.imentation in the lagoon at Ihpingamarangi is, of course, much complicated by the many pa.tch reefs that rise above the floor. Each ofthese has sediments on its top and sides of varieties in keeping with the general depth ranges, except that where slopes are especially steep the lower limits of various sedimentary types exlend deeper than otherwise. Further complications in the pattern of se2imen-kary belts are caused by bottom currerlts in certain areas. Tile white silts characteristic of these current zonesare especially well developed in and near the ma.in pass on the south, in a broad area bordering the sou-th side of the atoll and within a mile of it, and .in narrow zones bordering the ves-i;erri and northern arcs. The east-trending deposits of current-'cramported silts are prominerit in airplane photographs of the southern part of the la~oon. The problem of 37by belts of sedimen-t comparable to those in ICapingm!a- rangi lagoon are nut recognized in other atolls merits consideration. A likely answer is found in comparative depths: mosi; other lagoons that have . . . . been sampled are shal!;.oyer thah the one ai'Ibpingmarengi. Available de- scriptions indicate tnaf belts co~-res_noi~c?ii~gto those of the rel.alive3.y shallow waters do occur, i., the belts of near-shore foramiuiferal sand, of clastic sand, and of coral debris. At Raroia, in the deeper parts (20- 25 fa-thorns), Ner.~ell(1954, p. 25) reports, a lack of bo-mom mud but records . . some accumulatioris of Hal.i.meda. This 'dht:ibution paral3.els closely that - --.-- - for correspoi.ding depths at Kapingamarmgi . . .

GECLOGY OF TI% PATCH REEFS The evenness of the bowl-shaped basin chat contains the lagoon at Kaping- amarailgi is disrupted in many places by patch reefs that rise as moulds from :its sides and floor.. Many of these reefs are subcircular in plan, but others are linea?, and still others very irre&ar. They range in size from low :knolls or pinnacles a few dozen feet c.cross and 10 or 20 feet high to wide platforms that are 1.00 feet high or more. Among the largest is Sokoro, vhose 'tab is 2,900 feet long and 700 fee-t vide, aiid Tokope;, which has an upper surface approximately 1,1100 feet by 900 feet (fig. 20). . . The total number of pa-tch reefs is no-ci"imown, but 20 of those i.isted are more than 500 feet long.. Available mapsshoy aoproximately '75 of all sizes, yet even -this number probably is far from'Lhe tokal, for :nany very : sma1.l ones and others that 60 not reach th&'sul.fa&eare noi: included. Cou~lt- igg oT patch reefs is fVther co,pfus&d &caVs& sane reefs tha-t appear separate at the surface are coiected at 'diallor<'dep$hs. Of the az?roximately - -I 5 inciicated on maps, 35 are 'in wates deeper ,than 60'feet, including some of the very large oms that. rise from t?ne deepest pakis of the lagoon. The other 40 pa-tch reefs shown on. the niaps ,r~'mo~,tlysmall imd occur in shoal waters bordering the ikier garGih of The 'a;toll; ,especfally along its northern . . and western sides. . . . . , .. ~. .. . Nearly all the patch reefs. a:t :~b,billgmfl&ran~i'have flat tops at or slightly above ].ow-tide.. level, &.viilg to each ix3ul.a the i@pea;.ance of. a mesa rising up through. . the lagoop. Only on & few of the s.dbmeimged'.reefs do top surfaces appea? to be. soniqhat Poun~ledand irregular. Flat tops :are characteristically developed and mahtained inmost patch reefs 'because the ulx~ardgrowth of orgariisms is controiled by tide levels and the surface.is continually being bevelled by wave'action, The sides are steep, in general, an4 commonly slope off at angles of 30" Lo Cboo'.' hey have the appearance of being far steeper than this' and locally seem' tolbe nearly vertical, butmea- surements show that m~s-tvisual estimates are high. . , . . .. The general shape of Manin knoll in the northeast.em'~artof the lagoon wa:; determined from soundings (table 8). This p&h reef probably is typical of most at Kapingamarangi. Profiles (fig. 23) silo>: that its southwest side, in its steepest part, slo$es dor.m at about 50' within a vertical distance of h0 to 50 fee-L, and that the i~o~thwestside has a slope of about 40°. These are extremes, and other sides of the reef have more gentle slopes, including the southeastern sector, which extends as a slightly sub- merge3 promontory or ridge far out into the lagooil. Table 8.- Depth (in feet) of lagoon floor ,surrounding Manin patch reef

Distance (in ft) Directior outward from r-- reef margin 145~~'~80'~ - -- .- - -- -. - .. ..- - - --. 2 2

30 18 60 , 54. -- 1 80

96 PO2 .s4 i nb I - I *Along submarine ridge, 128 feet outward from 2-foot margin.

The distribution and orientation of patch reefs in the lagoon seem to reflect in large measure the trends of currents and waves. A cluster of patch reefs, some of them large, near the ship's pass on the south side of the atoll probably is directly related to the strong currents that maintain good circulation in that area. The many small patch reefs immediately in- side the northwestern and western arcs of the atoll apparently developed in response to waves of maximum fetch before the dominant winds. Likewise a general northwestward lineation of many reefs in the lagoon center probably reflects this dominant wave direction. The fact that many of the patch reefs are subcircular rather than elongate is believed to be due to the relative quietness of lagoon waters., allowing nearly equal growth in all directions, as suggested by Cloud (1952a, p. 2140).

All of the patch reefs in deep water, and many of those in shallow, rise from parts of the iagoon bottom that are covered with extensive deposits of foraminifera1 sand and calcareous mud. The patch reefs themselves, however, are largely mantled with growing corals. Dense thickets or forests of the yellow, branching Porites andrewsi cover most of the sides, and both microatolls and smaller coral head~~representingnumerous species, cover extensive areas on the flat tops, especially along the margins. Typical coral assemblages from three Kapingamarangi patch reefs, identified by J. W.. ldells, are listed in table 9. The list served to indicate the principal forms, though others contribute to the reefs also. A comparison of these lists with those of Wells (1954, p. 390-393) for the Marshall Islends shows that all of the genera and all but two of the species of Kapingamarangi occur also in the Marshall Islands. Table 9.- Some characteristic coral asscliiblages on Kapil~~amarangiAtoll, determiried by Z. W. Weils

1.5 i i6 b. .bilk i r( k aJ s m.0I""$ i zoo461 0 ikd ~i o i k ,-I k u c-i 1 $2 .,I , la:..;d o rn / 8~~jb/ rr\ lu ! [in ; 4 I-. I-. .- -- .- - - -. 1 --+ .!-- -.- Psammocnra- nlerstraszi v.d. Horst I , i j . .L..-.; ~ . t Seriatopora hystrlx Dana 1 X / -.. I ' 1 Poclllopora damcornis' (~innaeus) ' x i i ! daGe3FFFil-1 I

Wells - --'-- -i i prolifera ~rueggemai' ! -- "---.- . i i verrilli Vaurrhan !

i f

Lime sand ihat accumulates on the upper.sides ind tops of most patch reefs is similar to that accumula-tin6 at corresponding depths off the lagoon beaches and lagoon reef margins of the atoll. This sand consists largely of sliells arid corals. Some sand coriiains moderate amounts of locally derived Halimeda opuntia fragments. From field sketches of 10 representative pa-tch reef-ig. 2b), the cilaracteristic distribution of sand is apparer~t. In general, sand is concentrated in the centers and on the norbheastern margins, belaw which it cormonly covers slopes down to depths of 20 fee-t or more. The thickness of sand accumulations on most patch reefs does not appear to be great (table lo), and on some, where coral growth is especially hxur- iant, sand is confined Lo small pockets. In contrast, Sokoro, Tokopel, and a few other knolls contai:l sufficient sand to form bars of appreciable size that rise well above low-tide level.

.... hib1.e 10.- Thickness (in inches) of sand accumulations on patch reefs I ...... ~...... -... . i / Patch Reef - IE Margin Center SW Area I ! ...... , ...... -...... ! .. 1 I j Tisu 6 1 5 I 5 1 i I i I ! %tamtong 11 7 - i 1i I I I i 1 Sokoro - lo i lj ! .... i I .i...... I Why sand that extends over the rims of the lmolls is confined almost entirely to the northeastern slopes is not known. On other sides, where thickets of branching corals are absent, the slopes normally are covered wl-th tine debris of dead coral branches instead of sand.

As stated by Ladd and others (1950, p. b2l.), "the age, origin and internal constltution of the knolls is not known, as no structure of this EXPLANATION rn Dead coraln pavement Microatolls Sand surface Branch corals

Spherical coral heads

HAKAROPURAP Diameter 500 ft.

TIWAWE I200 ft x 800f1.

TOKOPEL 1400 ft. w 90011.

MATAMATONG 1000 ft. x 800 ft TlSU Diameter 500 ft

TOKOLARA Length 500 ft

THOKOMOTU Diameter 500 f1.

SOKORO 2900 ff x 700fl TOKOHUI 300 ft.

FIGURE 26.-FIELD SKETCHES OF PATCH REEFS SHOWING DISTRIBUTION OF SEDIMENTS AND CORALS toe has ever been drilled." Those writers suggest that -the ptch reefs filay- have developed through sporadic cora:L growth that was iriitiated during a part of the Pleis-Locene e~och,when sea >.eve1 was much lover than now. Such an hypothesis iiould fit very well tine mcager kno:.led;:e avail.db1e concer~ing the shape of the lagoon basin, the extensive bot"com saiids arid r;uds that are helping to fill it, and the patch reefs t.hat rise to 10.7-tide level within Che lagooii. Fu~thermore, the bevelled. surfaces of dead coi,al, ba.ck from tine actively y-owing rims on ;nos t of the high knolls, stronglj suggest that like the ou.kr reef of' the ato1.l they were higher than at prberii not far back in history and ha,ve been 'truncated because of a receut drop in sea level. From ?>hatmay be observed of their simple shaje and structure, it seeins probable that these patch reefs are siailar in all essential res2ects to the bioherins that have been described fr-om many ancient deposits, especial1.y those of Silurian and Devoxian age (~hroclr,1939).

RE>TRFfi?CES CITED

Arnox, Ted, 195it, Die hg-krology of the northern Karstall Islands: Atoll Research Bull. 30, 7 p. ; .. . .

Aso, Yavata, 1-953, A yrel.iminaly study of the propertizs and forr(iatioi1 of phosphate deposits on Tolwbei Island; report, Sou-;;'? Sea Developmmt. Company, Pe

Chave, K. E., 19:'+, Aspects ol' the biochemistry of magnes%um. 1. Cal- careous rnarine organisms: Jour. Geology, v. 62, no. 3, p. 266-233.

Cloud, ?. E., Jr., 1952a, F'acies relationships of' organic reefs: Am. kssoc. Petroleum Geologists Bull., v. 36, no. lk, p. 2125-2149. -.------, l952b, Preliminary report on geology and marine environ- ments of' Onotos Atoll, Gilbert Islands: Atoll Reseaxh Bull. 12, 73 p.

Cox, D. C., 1951, T'ne hydrology of Arno Atoll, Marshall Islands: Atoll Research Bull. 8, 29 p.

Dal~,R. A., 1924, The geology of American Samoa: Carne@ie Inst. ilashing- ton kb. 340, p. 93-143.

Davici, T. W., anG Sweet, G., lgOh-, The geology of li'unafu2;i: Royal Soc. London, Nept. of Coral Keef Corm., Atoll of I'unafwii, sec. 5, p. 61-24.

Fairbridge, R. W., 1953, Recent and Pleistocene coral reefs of Australia: Jour. Geology, v. 58, .no. 4, p. 330-401,

Fiela, R. M., 1920, Origin of the "Beach Xock" (~o~uiila)at Loggerhead. Keyi Tortugas: Geol. Soc. America Bull., v. 31, no. 1, _u. 215.

Ginsburg, R, N., 1953, Beachrock in south Florida: Jour. Sedimentary Petrology, v. 23, no. 2, p. 85-92. Ladd, H. S., Tracey, J. I., Wells, J, F., and Ekery, I:. O., 1950, Organic grow-th and sedimentatAon on iin atoll: Sour. Ge(Goyj, v. 58, no- '.I, p, 410-495.

Ladd, H. S., Ingerson, Earl, Tomsend, R. C., Russell, I.la.rtj.n, aild St,ephen- son, H. K., 1953, Drill-ing on Eniwntok PLu11, Miirsxiail Is!.anr;s: Am. Assoc. Petroleum Geolo

Llogd, Christcpiier, lg&J3 The voyaces of Captain James Cook rm3 t:ie world, selected from his journals: London, Cresset Press, 384 9.

MacNeil., F. S., i950, Planation oC recent reef fla-ts on OXinawa: Geol. SOC. Ari.erica Bull., v. 61, no. 11., p. 1307-1308.

Newell, N. D., 1954, Reef's and sedlneztary processes of Ilal-oia: Atoll Ne- search Bull. 36, 35 p.

Nugent, L. E., Jr., 1945, Coral reefs in t!le Gilbert, Marshall and. Caroline Islands: Geol, Soc. America Bull,, v. 57, :lo. 8, p. 735-790.

Shrock, R. R., 1939, Wisccnsin Silurian bioilerms (organic reefs): Geol. Soc. America Bull., v, 50, no. 4, p. $29-562.

Sewe1.1, R. B. S., 3.936, .b accoun-t of Adau ktoll: Eri-tish Kus. Nst. History, The Johr, Murray Ex@. 1931-34, Sci. Rep-ts., v. 1, no. 3 and 55, p. 63-95.

Sollas, W. J., 1904, Xarrative of the expedition in j.896: Royal Sac. London, Re3.l: of Coral Reef Comfi., The Atoll of Funafuti, sec. 1 p 1-28.

Stearns, H. T., an6 Vaksvik, K. N., 1935, Geology and ground-.rater' resources of' -the is]-and cf Oahu, Hawaii: Hawaii Dept. of Public Lsnds, Div. Iiyckograpily, Bull. I., 479 p.

Stone, E. L., Jr., 1951, The soils of Axo Atoll, Marshall Islands: Atoll Research Eull. 5, p. 1-55.

Tracey, J. I,, Emery, K. O., and Cloud, P. E., Jr., 1955, Conspicuous features of organic reefs: Atoll Research Eull. 46, 3 p.

Wells, J. W., 1951, The coral reefs of Arno Atoll, Marshall. Islands: Atoll Research Bull. 9, 14 p. --, 1954, Recent corals of the Marshall Islands: U, S. Geol. Survey Prof. Paper 269-1, p . 385-486.