Pacific Science (2000), vol. 54, no. 4:423-442 ? 2000 by University of Hawa?'i Press. All rights reserved Stratigraphy and Whole-Rock Amino Acid Geochronology of Key Holocene and Last Interglacial Carbonate Deposits in the Hawaiian Islands^ PAUL J. HEARTY,^ DARRELL S. KAUFMAN,^ STORKS L. OLSON,* AND HHXN F. JAMES* ABSTRACT: We evaluated the utility of whole-rock amino add racemization as a method for the stratigraphie correlation and dating of carbonate sediments in the Hawaiian Islands. D-alloisoleucine/L-isoleucine (A/I) ratios were deter- mined for carbonate sand and sandstone samples from 25 localities in the archi- pelago. The superposition of A/I ratios within stratigraphie sections and the re- gional concordance of ratios within geologieal formations support the integrity of the method. To correlate the A/I ratios with an absolute chronology, com- parisons were made with previously published uranium series dates on corals and with **C dates on carbonate sand and organic material, including several new dates reported herein. The A/I mean from four marine isotope stage (MIS) 5e U-series calibra?on sites was 0.505 ? 0.027 (w = 11), and 12 "test sites" of previously uncertain or speculative geochronological age yielded an A/I mean of 0.445 ? 0.058 (H = 17). Similarly, extensive Holoeene dunes on Moloka'i and Kaua'i were correlated by a mean A/I ratio of 0.266 + 0.022 (n = 8) and equated with a '*C bulk sediment mean age of 8600 yr B.P. Our results indicate that the eolian dunes currently exposed in various localities in the Islands origi- nated primarily during two major periods of dune formation, the last inter- glacial (MIS 5e) and the eariy Holocene (MIS 1). MIS 5e and MIS 1 A/I ratios from the Hawaiian Islands show close agreement with previous whole-rock studies in Bermuda and the Bahamas. We discuss these results in terms of their relevance to models of lithospheric flexure and to imposing constraints on the time frame for the extinction of fossil birds. GEOCHRONOLOGY AND SEA-LEVEL history have last interglacial (MIS 5e, ca. 125 ka been particular^ important in the scientific [thousand years] old) deposits from O'ahu history of the Hawaiian Islands. Thermal (Ku et al. 1974, Easton and K.u 1981, Muhs ionization mass spectrometric (TIMS) and and Szabo 1994, Szabo et al. 1994). ^*C dates alpha U-series dates augmented by '*C dates from terrestrial sites on O'ahu, Moloka'i, from the Holocene (marine isotope stage 1 Hawai'i, Maui, and Kaua'i (Olson and [MIS 1]) provide a temporal framework for James ?982?I,?>, James et al. 1987, Paxinos important events in geology and paleobiol- 1998; this study, Table 1) have placed con- ogy during the late Quaternary. More than straints on the recent extinctions of dozens of 70 U-series dates have been obtained from species of fossil birds. Nonetheless, the age of many important sites remains uncertain. U-series dating is ^ Manuscnpt accepted 1 February 2000. effective on coral samples, ideally those ^Schoo! of Earth Sciences, James Cook University, ? , ? .^u ??? A ^^^J^^^A ^r Townsvitle, Queersland 4811, Australia (paul.hearty? collected m growth pOSl?on and COmpOSed of jcu.edu.au). pure aragonite. However, like other reia?vely 'Departments of Geology and Environmental stable carbonate-producing areas (Bermuda Sciences, Northern Arizona University, Flagstaff, Ari- ^j^^ jj^g Bahamas), the majoritv of deposits in zona 86011-4099 (darreU.kaufman@iiau.cdu). , Ua^?:;?_ islands are not corals or coral "National Museum of Natural Historv, Smithsonian ^^ Hawaiian ISianas are not cor?is or corai institution, Washington, D.C. 20560. reefs, but carbonate grainstone deposited m 423 424 PACIFIC SCIENCE. Volume 54, October 2000 TABLE 1 "C DATES FROM FOSSU. SITES IN O'AHU, MOLOKA'I, AND KAUA'I LOCALITY ISLAND (AAR FOSSIL SAMPLE = A/I RATIO (REF NO.) SAMPLE NO.) '?hikilolo Sllb SAMPLE NO. MATERIAL YKB.P. ? Iff (TABLE 2) O^ahu (2) Beta-130725 Organic clay (OOHle) beneath -i-2 m beachrock 4,990 ? 60 >0.215? Kaua'i (2) Makaweh) dvines (KMWlc) Makawehi dunes K2 Beta-122589 Calcareous sand 8,900 ? 70 0.245 Kaua'i (1) K2 SI-3792 Land snail shells 6,740 -1- 80 K-aua'i (1) Makawehi dunes K2 SI-3793 Land crab claws 5,145 ? 60 Kaua-i (1) Makawehi dunes K2 AA-2976 (AMS) Bone, Brania sandvicensis 4,690 ? 100 Moloka'i (2) '?lio Point (MIP6x) 'l?o Point (MIP4X) '?lio Point Beta-12259 i Calcareous sand 12,740 ? 90 0.277 Moloka'i (2) Beta-122590 Calcareous sand 12,710 ? 90 0.308 Moloka'i (1) Site 20 SI-3791B Large land snail 5,510 ? 65 shells Moloka'i (1) '?lio Point Site 20 SI-3791A Small land snail shells 5,245 ? 65 Moloka'i (4) Kaunakakai Beia-60345 Calcareous sand 4,750 ? 70 0.270 West Beta-71518 Calcareous sand 5,730 + 80 Moloka'i (2) Mo'omomi (MMM4x) Mo'oraomi Beta-122592 Calcareous sand 8,360 ? 60 0.246 Moloka'i (1,3) Site 1 HIG-35 Land snail shells 25,150 ? lOOO .Vcre: Uncalibmed '"C ages are younger than their calendar ages. Because of the marine reservoir effect, bulk caleartous sedi- ment ages are several hundred years younger than the quoted age, References: (1) Olsor. and James (1982a,*); (2) neu "C ages from this study; (3) Stearns (1973); (4) Fletcher et al. (1999). ?.^/? ratio from beadjrock ttuncaung dated organic clay. subtidal, beach, or eolian environments. The need to date numerous sites lacicing in age control, as well as the availabihty of several previously well-dated sites, provides incen- tive for the application of new dating tech- niques such as whole-rock aminostratigraphy (Hearty et al. 1992, Hearty 1998). The Hawaiian Islands (Figure 1) are a succession of hot-spot volcanoes that, after formation, were conveyed WNW with the migration of the Pacific plate (Jackson et al. 1980). Those islands located farther west of the hot spot are progressively older; for ex- ample, the ages of formation of Moloka'i, O'ahu, and Kaua'i are about 1.6, 3.0, and 5.0 my (miUion years ago), respectively (Clague and Dalrymple 1989). A "rejuvena- tion" phase of volcanic activity occurred on these islands well into the middle and late Pleistocene. As they aged, the volcanoes ac- cumulated greater volumes of limestone de- posits (Darwin 1839) that were emplaced on shorelines, mainly during high stands of sea level (Bretz 1960). According to the Uthosphenc model of Grigg and Jones (1997) (Figure 2), the is- lands of Lana'i, Moloka'i, and O'ahu should experience net uplift, whereas Hawai'i and Maui, lying within the subsidence moat, and Kaua'i, lying beyond the forebulge ridge, should experience net subsidence during the late Quaternary'. This and other lithospheric models (Watts'and ten Brink 1989) can be tested, provided sufiBcient age and sea level information is available. Anthropogenic perturbations of Hawaiian ecosystems caused massive extinctions of birds and other elements of the biota in the approximately 1500 yr since the arrival of the first humans in the archipelago (Athens Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 425 ?az? I KAUA'I VRlWAU isr Mini??? Dmm T' O'AHU Kahutai 6T* ?Kaiila -2I* BO* ?I _J Z^j^OLOKA'l Kahuhri ) MAUl LANA'! ?IWl^ iw ?^ KAHOOLAWE ^^^*. ?Loman X- -?* 200 [79* SOO KLOMETERS (TO- PACIFIC OCEAN ^ Be* tss* FiGifl?E 1. Location map of study sites in the Hawaiian islands (modified from Olson and James [19S2aj), 1997). The extent of avian extinctions has been revealed in Pleistocene and mainly late Holocere fossil deposits on six o? the main Hawaiian islands. Carbonate dunes record an important chapter in the unfolding story of prehistoric human-induced extinctions, par?cularlv of birds, in the Hawaiian Archi- pelago (Olson and James \9Z2a,b, 1991, James and Olson 1991). -\s yet, the only fossil bird remains collected on the island of Moloka'i are from dtme deposits in the vicinity of Mo'omomi Beach and 'Dio Point. Until recently (Bumey et al. 2000), the oniy fossil record of birds from the island of Kaua'i was likewise obtained entirely from dune deposits at Makawehi on the south- eastern coast. Knowledge of the geology and age of these deposits is important for estab- hshing the chronolog>^ and probable causes of extinction, as well as interpreting the past environmental conditions under which now- extinct species flourished. We use the whole- rock amino acid racemization (.'V'VR) geo- chronology results to constrain the ages of terrestrial fossil deposits in Hawaiian dunes. AAR ANALYTICAL PROCEDLTRJES Whole-rock aminostratigraphy has been used to unravel Quaternary stratigraphie questions in Bermuda (Hearty et al. 1992) and the Bahamas (Hearty 1998, Hearty and Kaufman 2000) and has been effective for the estimation and correlation of ages of deposits on diverse islands. For example, in the Ba- hamas over 100 whole-rock D-alloisoleucine/ L-isoleucine (A/1) ratios from stratigraphi- cally defined last interglacial MIS 5e oolite yielded a consistent and unambiguous cor- relation across 700 km of the Bahamas Archipelago (Figure 3). The individual and collective last interglacial A/I means from 13 major island groups do not overlap with 426 PACIFIC SCIENCE, Volume 54, October 2000 Kauai Elovated Fsef deposits A Tide records ; Drowned rMis O _i_ -L ^0 100 200 300 4O0 500 600 Oistanc? along Hawaiian Hidge from hotspot, Hlauea volcano (km) FIGURE 2. Lithospheric ?exure model of Grigg and Jones (1997) showing subsidence of Hawai'i and Maui wiihin the moat of the Big Istand; uplift of L?na'i, Molo- ka'i, and O'ahu in the forebulge region; and subsidence of Kaua'i beyond the forebulge area. In general terms, our findings support this model. either younger or older modal classes of A/I ratios or "aminozones" (see Hearty et al. 1986). The underlying theory and various appli- cations of the AAR method are summarized in Rulter and Blackwell (1995). The AAR method is based on the racemization of amino acids preserved in biominerals (Hare and Mitterer 1967). Through time, L-amino acids racemize (or, more specifically in the case of the amino acid isoleucine, epimerize) to their D-isoraer form. The ratio of A/I amino acids measures the extent of epimeri- zation. In the epimerization reaction of iso- leucine, the A/I ratio is initially zero (0.011 with laboratory preparation) in truly modem organisms and increases to an equilibrium A/1 ratio of about 1.3 with time after death of an organism. Because the whole-rock method analyzes aggregates of comminuted skeletal and precipitated carbonate grains that form offshore over time, entirely "mod- em" material, even on active beaches, is not expected. Thus, each whole-rock sample will have some "inherited" age. The implicit assumption is that the inherited age is similar for ail samples of the same age. As seen in the results, the local and regional consistency of A/I ratios from equal-age units supports the validity of this assumption in most cases. Like many chemical reactions, the rate of racemization/epimerization depends on the ambient temperature of the reaction medium. Thus, sites at lower latitudes and warmer temperatures are expected to yield incremen- tally higher ratios. Single A/I ratios with- out stratigraphie context are obviously un- acceptable indicators of age or correlation. However, within a local or regional setting where several separate sites are determined to be stratigraphie equivalents by field geology, single ratios from each of several outcrops that yield similar A/1 ratios are considered to be an effective demonstration of the method. The Hawaiian Islands surveyed in this study (Hawai'i, Maui, O'ahu, Moloka'i, and Kaua'i) lie in the Tropics between 20.5? and 22? N. Relative to the Hawaiian Islands, most of the Bahamas are situated at higher, cooler latitudes (28? to 23? N), but also ex- tend southward to similar latitudes (Inagua, at 21? N). Bermuda lies well north of the Tropics and near the limit of reef growth at 32.3? N. Historical temperature records from the Hawaiian Islands generally yield higher (25?C) mean anntial temperatures (MATs) than Bermuda (20?C) or the Bahamas (22- 24? C), and notable intraisland MAT differ- ences are evident in Hawai'i. Because of the inferred difference in temperature histories between these distant localities, we cannot correlate A/I ratios directly. Instead, the ki- netics of racemization predict that deposits from Hawai'i should have higher ratios than deposits of similar age in the Bahamas, which have higher ratios than those in Bermuda (Hearty et al. 1992, Hearty 1998). Sample Preparation and Analysis The whoie-rock sample preparation pro- cedure follows that of Hearty et al. (1992) and Hearty (1998). However, in contrast to previous study sites in Bermuda and the Bahamas, some samples from the Hawaiian Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 427 eg tL cc I o o 0.7T 0.6 0.5- 0.4- 0.3 0.2- 0.1- 0.0 Hawaii a E si g? UIUJ2b(2) 0.523 ? 0.015 25nC 0KP2a(l) 0.541 ?0.013 251 ID 0KP2a (1) 0.532 ? 0.000 120 ka; w/o volcanics 120 ka; w/o volcanics 130 ka; w/o volcanics 130 ka; w/o voicanics Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. TABLE 2 (continued) 429 LOCALrrY A/1 RATIO ? 1(7 AGE LABORATORY NO. HEUS NO. (ANALYTICAL ERROR") ('"C, ?-SERIES OR STRATIGRAPHIC) Barbers Point (Shennan et al. 1993) 2509 OBAlc (2) 0.460 ? 0.041 117 ka mudstone 2508B OBAIb{2) 0.493 + 0.008 140-120? ka beachrock 2508C OBAIb (2) 0.493 + 0.018 140-120? ka beachrock 2508A OBAlb (1) 0.479 + 0.014 140 ka grainstone Makai Range Pier (Szabo et al. !994) 2503 OMIla 0.504 ? 0.021 134 ka Kalic Point Beach Park (Szabo et al. 1994) 2841 OHE2b 0.480 + 0.023 MIS 5e reef 2994 OHE4C 0.627 + 0.013" (~.120ka)MIS5econgI. 2840 OHElc 0.608 + 0.013" (~120ka)MIS5econgl. 2993 OHElc 0.543 + 0.018" (-.120ka)MIS5econgt. Kahviku Point (Ku ei : al. 1976, Szabo et ai. 1994) 2704A OKKJc 0.397 + 0.004 (< 121 ka) MIS 5e or 5a dune? 2704B OKKle 0.397 + 0.006 (<121 ka) MIS 5e or 5a dune? Laniloa Peninsula 2703A OLAlc 0.379 ? 0.003 MIS 5e or 5a? eoBanile 2502 OLAla 0.640 ? 0.019 Mid-Pleistocene (MIS 7/9) M?kua Valley 2992 OMUla 0.550 + 0.008 MIS 5e shoreiine at +9 m Maui Kahului 2832 AMCCla 0.378 - 0.017 MIS 5e eolianiie 2837 AMZBla 0.380 ? 0.021 MIS 5e eolianite Moloka'i '?lio Point 252 IB MIP3a 0.3S6? 0.019 MIS 5e eolianite 2521A MlPla 0.552 ? 0.025 MIS 5e eolianite Mo'omomi 2522C MMM2a 0.470 + 0.002 MIS 5e eolianite 2522D MMMla(l) 0,470 + 0.005 MIS 5e eolianite 2522E MMMla (2) 0.433 ? 0.008 MIS Se eolianite Kaua'i Makawehi 2525 B KMW2a 0.441 ?0.011 MIS 5e eolianite 2525C KMW3a 0,478 + 0.013 MIS 5e eolianite Aweoweonui Beach 2527 KAAla 0.420 ? 0,002 MIS 5e eolianite 2706 KAAla 0.444 ? 0.000 MIS 5e upper shore sands Barking Sands 2839 K-BSlz 0.465 ? 0.010 Late Pleistocene sands 2838 KBS2a 0.528 ? 0.050 MIS Se upper foreshore sands " Analytical error for multiple anaivses of same vial sample. A/I values for the Inter-Laboratory Companson Standards !LC-A. ILC-B. and ILC-C measured at Northern Arizona Univei?ity <1?8-1999) were 0.148 ? 0.004, 0.498 ? 0.022. and 1.049 ? 0,025. These values are well within the range measured for the same samples by other laboratories (Wehmilkr 1984). 'Holocene deposits composed largely of reworked Pleistocene sands. 'Dates frotn Fletcher et al. (1999). 'Higher ratio probably the result of shallow burial (<1 m). o-phthai aldehyde (OPA) and fluorescence de- tection. Each sample solution was analyzed three to five times and the results averaged. The coefficient of similarity (?T/Z) of average peak-height A/I ratios was typically <3%, which represents the internal reproducibility (analytical precision). The analytical preci- sion accompanies all data presented in Table 2. Error resulting from analyses of several different samples from the same geological 430 PACIFIC SCIENCE, Volume 54, October 2000 A MOKAPU POir?T, O'AHU alluvium incontCMmity U-series: 127*4 ka A/I (0KP2b) =0^16*0.010(2) 134*4,13116.134*7 ka A/I (0KP2a) = 0.S3710.006(2) UTKonionnity =^^==.^=7 Ciater late sediments B MAKAI RANGE PIER. O'AHU TIMS13Z.6?3.3ka (OMI1a)sa.S04 A/1 (0MI1z) = 0.103 UiWV:115*10lia A/I iOBA1c) = 0.460 Unit IV: A/((0BA1b(2))* 0.493 xOOOOfZ) Unit III: 145t1S. 160*15 Ita A/I (0BA1b{1)) =0.479 uncontorrnty pateosol D KAHE BEACH STATE PARK, O'AHU IS-i ,11-series: 120*3,134j4,142*12 ka /VMS: 114,3*4.0 Ka (R^) 1 (OHEIC) = Q,593*0J)44 (3) (haaiod sani|]te) (OHEZb) 0.4S0 FiGUT?E 4. Stratigraphie sections of U-series calibration sites discussed in text. Mean U-series ages and whole rock A/! ratios are plotted in each section. A. M?kapu Point. O'ahti (in situ reefs at -i-5.5 m and +9 m) from Mubs and Szabo (?994). B. Makai Range Pier. O'ahu. Shoreline deposits at -i-2 m dated at 134 ka (Szabo et al. 1994). C. Barbers Point. O'ahu. Section indicating two high stand event at -5.5 m and >+6 m asi (Shennan et al. 1993). D. Kahe Point Beach Park with -i-5 and -10 m paleoshoreline levels (Hearty section). U-series ages are from (1) Muhs and Szabo (1994), (2) Easton and Ku (1981), and (3) Szabo et al. (1994). unit is reported in all other tables and figures in the format of the mean (X) ? 1 standard deviation (la), and the number of samples analyzed (in parentheses) (e.g., 0.464 ? 0,042 [n = 20]). To monitor analytical drift and to facilitate comparison with data from other laboratories, ?ie NAU laboratory routinely calibrates with the Interlaboratory Compar- ative Standards of Wehmiller (1984) (see footnote in Table 2). CALIBRATION SITES FOR THE LAST INTERGLACL^L (MIS 5e) Independently dated Pleistocene sites in O'ahu were selected as the control group for calibration of the AAR method. At most of the undated sites, it was possible to determine on the basis of field criteria whether deposits were Holocene or late or middle Pleistocene, Stratigraphie sections are illustrated in Fig- ures 4-6, which include previously published U-series dates and the A/I and radiometric data from Tables I and 2. O'ahu M?KAPU POINT. The geology of M?kapu Point and Ulupa'u Crater has been the subject of geological investigations for over a century (Dana 1890, Steams and Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 431 Vaksvik 1935, Wentworth and Hoffmeister 1939, Winchell 1947, Gr?mlich et al. 1971). Although undated, the eruption of Ulupa'u Crater probably took place during the early to early-middle Pleistocene, probably among the older "rejuvenation stage" volcanics (Ko'olau) of O'ahu (Clague and Dalrymple 1989). Sometime after the construction of the crater, a lake formed within its walls. Through much of the middle Pleistocene, both lacustrine and coUuvial sediments filled the crater to at least 20 m above current sea level. A considerable ntmiber of species of fossil birds have been found within the lake sediments (James 1987 and unpubl. data) and provide an important snapshot of avian evolution during the middle Pleistocene. Be- fore the Waimanalo transgression (~125 ka), marine erosion removed the eastern half of the crater, and reef fiats from that transgression directly abut the cliffs of crater fill. A com- posite stratigraphie section of the Waimanalo deposits at M?kapu Point (after Muhs and Szabo 1994) is shown in Figure 4A. Ku et al. (1974) determined alpha U-senes ages of the Waimanalo deposits at Mokapu to be 131 ? 8 ka at -7,8 m and 134 ? 7 ka at -rll m from coral cobbles in the upper conglomerate. Muhs and Szabo (1994) pro- vided alpha U-series ages of 134 + 4 ka and 127 ? 4 ka from coral heads in growth posi- tion at +8.5 m and marine conglomerate at +12.5 m. Fourteen TIMS ages from a study by Szabo et al. (1994) ranged from 123 to 141 ka, of which four samples from growth- position corals in the conglomerate at +7.3 m and +8.6 m yielded ages of 130 ? 2 ka and 124 + 3 ka, respectively. For dating comparison by AAR analysis, samples were collected at Mokapu Point from the reef matrix at +5 m and from shallow subtidal sands at +10 m. The two samples from reef matrix sedi- ments at +5 m (OKP2a) produced a mean A/I ratio of 0.537 ? 0.006 (volcanic grains excluded) (Table 2). Two similarly prepared samples from 4-10 m (0KP2b) generated a mean A/I ratio of 0.516 ? 0.010. Thus, A/I ratios from Mokapu are in stratigraphie order and yield ratios re?ecting a several- thousand-year interval, in agreement with previously published U-series ages. MAKAi RANGE PIER. This simple outcrop exposes beach conglomerate at +2 m (Figure 4B). The deposit was attributed to the L?'ahi shoreline of Steams (1978), which he viewed as a regression phase from the Waimanalo high stand. At face value, a single TIMS U-series age of 132.6 ? 3.3 ka (Szabo et al. 1994) appears to contradict Stearns' late MIS 5e interpretation. Both the Pleistocene con- glomerate and modem beach sands were sampled for AAR. A single A/I ratio of 0.504 suggests a correlation with Mokapu and Barbers Points and agreed with the older MIS 5e TIMS age of Szabo et al. (1994). BARBERS POINT. Evidence of two sea-level oscillations separated by a minor regression were interpreted from the sequence by Sher- man et al, (1993) (Figure 4C). Alpha U-series dating provided three wide-ranging ages of 115 ? 10 ka, 145 ? 15 ka, and 160 ? 15 ka. It is clear from stratigraphie relations and the younger MIS 5e U-series age (115 ? 10 ka), however, that the deposits are last interracial. Lower to upper units II (in situ bafBestone), IV (beachrock slabs of grain- stone), and V (in situ bafBestone) were col- lected for AAR analysis from Sherman et al.'s (1993) (Figure 4C) section at Barbers Point. Resulting A/I ratios are. respectively, 0.479 (n = 1), 0.493 + 0.000 (n = 2), and 0.460 in = 1). The lower two units II and IV were in- terpreted by Sherman et al. (1993) to belong to the older sea-level oscillation (128 ka?) and the upper unit V to the younger (120 ka?). Despite the variable composition of the limestone, A/I ratios confirm a similar older age of the lower two units, and the upper unit indicates a somewhat younger age. The mean A/I ratio of 0.481 ? 0.016 (? = 4) from the deposit closely corresponds to the Mokapu sec?on A/I ratio of 0.526 ? 0.014 (? = 4), suggesting a temporal correlation between the deposits on opposite sides of O'ahu. The range of A/I ratios suggests a depositional interval of several thotisand years, rather than the wide range of U-series and electron spin resonance (ESR) ages (30 to 45 ka) re- ported in Sherman et al. (1993). KAHE POINT BEACH PARK. Reef dcpOSitS form a broad terrace at around -r5.5 m at 432 PACIFIC SCIENCE, Volume 54, October 2000 Kahe Point, and coral and volcanic boulder conglomerate rises to nearly ?12 m eastward (east side of Fanington Highway) of the point (Figure AD). Uranium dates have only been obtained for the +12 m upper conglom- erate. The lower reef deposit has not been previously dated. Hasten and Ku (1981) as- certained an age of 142 ? 12 ka for the upper conglomerate, Muhs and Szabo (1994) deter- mined two ages at 120 ? 3 ka and 134 + 4 ka, and Szabo et al. (1994) derived TIMS ages ranging from 110 ? 4 ka to 117 ? 2 ka. Considering the conglomeratic nature of the +12 m deposit, a range of ages is expected, and the youngest ages (120 to 115 ka?) likely approach the apparent age of the deposi- tional event late in MIS 5e. However, coral ages between 110 and 115 ka are problematic because they center on the MIS 5d low stand. AAR samples from the previously undated -i-5.5 m terrace yielded a MIS 5e ratio of 0.480 (1), and the conglomerate at +12 m produced three ratios averaging 0.593 + 0.044. In the +12 m deposit, it was necessary to collect the AAR sample from less than ideal conditions (<1 m shallow burial of sample); it is likely that the sample experi- enced some surface heating, resulting in a somewhat elevated A/I ratio. AAR CORRELATION OF UNDATED PLEISTOCENE AND HOLOCEN-E SITES O'ahu KAHUKU POINT. Growth position corals, capped by a coarse conglomerate and eolian- ite. are exposed near sea level at Kahuku Point on the northeastern end of O'ahu. Ku e? al. (1974) obtained an alpha U-series age of 137 + 11 ka from the reef unit and 115?6ka from the conglomerate. Subse- quendy, Szabo et al. (1994) determined three TIMS ages averaging 121 + 3 ka from the reef unit. Collections for AAR were only possible from the stratigraphica?y youngest Pleistocene eolianite unit in the section, which yielded .A/I ratios of 0.397 + 0.000 (? = 2). Constrained by stratigraphy and TIMS ages (<121 ka), these uppermost, eolianites may record the final MIS 5e regres- sion between 120 and 115 ka or perhaps a subsequent high stand later in MIS 5. LANiLOA PENINSULA. Steams (1978) classi- fied the eolian deposits at Laniloa Peninsula along the east coast of O'ahu with the middle Pleistocene L?'ie high stand of sea level. Our field investigation revealed that the peninsula is actually composed of at least three eolian units, separated by soils. (These units are conspicuous on nearby Kukuiho'olua Island.) We concur with Steams' middle Pleistocene interpretation of the landward part of the peninsula, but with the addition of a younger seaward eolianite of late Pleistocene age (MIS 5e or 5a?) forming the eastern part of the peninsula. Our collections from the most landward and the most seaward units indi- cated these eolianites to be middle Pleisto- cene (MIS 7-9) and late Pleistocene (late MIS 5e or 5a?), with A/I ratios of 0.640 (1) and 0.379 (1), respectively. .M?KUA VALLEY. A road cut along Far- rington Highway 0.5 km south of the mouth of Makua Stream and 1 km north of '?hiki- lolo exposes loose, carbonate-rich intertidal sediments and growth position Pontes corals up to +9 m. Steams (1974) described another exposure at lower elevation along the beach in the vicinity of our site. A whole-rock A/I ! itio from the site yielded a ratio of 0.550. /dthough somewhat high, the A/I ratio and the similarity of the site to the stratigraphy of Kahe Point Beach Park point to a correla- tion with MIS 5e. '?H1KILOLO. In situ Pleistocene reef forms the base of the section, which is succeeded by interbedded fluvial conglomerate and dense, brown organic (marsh?) peat and clays dated at 4990 + 60 yr B.P. (Table 1). The fluvial conglomerate and peat are truncated by in- lertidal beachrock grainstone, which is, in turn, capped by a complex series of brown to reddish brown, silty to sandy coUuvial de- posits. Spherical borings 3^ cm in diameter (Echinometra) up to +2 m high in the beach- rock indicate that sea level rose to this level at some point during the Holocene. The bored beachrock deposits and marsh peat most likely correspond with a mid-Holocene Carbonate Deposits in the Hawaiian Islands?HEARTY FT AL. 433 MO-OMOMI (KW_ANl PT.), MOLOKA'I (SITE 1) HOLOCENE DUNE I (MMM3X) = 0.274 ) Paleosol A/I (MMM1e) = 0531 ?0.002{2) MIS 5e eolia^ite B KAIEHU PT., MOOMOMI, MOLOKA'I ta i>- H0LOC?NE DUNE 25150?10ClOyrBP (Steams, 1973) A/I (MMMIc) = 0.34610.018(2) Organic, brown marsh clay l(approx. sea level) MIS 5e eoli^ite M (MMMIa) = 0.452*0.026(2) uncxjnformity paieosol MIS 5a eolianite A/I (MMM2d) = 0.343 Organic, brown marsh (appnox. sea level) MIS 5e eolianite 'M (MMM2d) = 0.470 C MAKAWEHI DUNES, KAUA'I (SITE K2) HOLOCENE DUNE Bird Bones 4690.100 yr BP Cnb Claws S14S. 69 yr 8P Land Snails 6740 ? eo yi BP Bulk Sediment B900 > 70 yr BP D KAHULUl, MAUl . red silt (paieosol) day lens blond silt lens A/I (KMW1c)= 0.245 uncontarmUy A/1 KMW2a/3a) = 0.460*0.026(2) HOLOCENE DUNE A/I(AMCC1C) = 0.372 {sediment reworked from MIS 5e dune) FIGURE 5 Stratigraphie sections of several "test siles" for the AAR method. Whole-rock A/I rauos (plotted m sections) are "used for correlation of undated sites with calibration sites (Figure 4). A. KalaniPomt, Mo'omomi dunes, Moloka'i: Late Pleistocene and Holocene dune sequence. B. Kaiehu Pomt. Moiaka'i. C Makau^lu Kaua . Pleist^ cene and Holocene sequence of eolianite. D. Community college site, Kahului, Maui. Whole-rock A/I rauos indicate that "Holocene" dunes were largely reworked from Pleistocene dunes- high Stand of sea level around +2 m, else- where dated at approximately 3500 yr B.P. (Jones 1992, Fletcher and Jones 1996, Gross- man and Fletcher 1998). An A/I ratio of 0.215 (1) reflects a younger age than the underlying organic clays at 4990 yr B.P. and an older age than the 3500 yr B.P. high stand. It appears that around 5000 yr B.P., a beach barrier was established, impounding local island drainage and forming a slackwater marsh. Moloka'i Pleistocene eolianites capped by thick cal- crete and orange-brown soils are exposed at the base of sections at Mo'omomi (Figure 5A.B) and llio Point on Moloka'i. AAR 434 PACIFIC SCIENCE, Votian? 54, October 2000 samples from these sites provided mean A/I ratios of 0.455 ? 0.020 (n = 3) and 0.469 ? 0.II7 (rt = 2), respectively, confirming a cor- relation with MIS 5e calibration sites. Steams (1973) described the dunes at Kalanj Point (Figure 5^) where a fossil flightless anserifonn was discovered (hoio- type of Thambetochen chauliodous Olson & Wetmore [1976], discovered by Joan Aidem of Moloka'i). Steams (1973) interpreted the dunes as glacial age, partly on the basis of a 25,150 ? 1000 yr B.P. ^*C date on land snails taken from the intercalated soil. However, the geological setting of the units beneath a thick cal?rete and pale brown (lOYR 6/3) soil clearly places the dunes in the Pleistocene, most likely associated with high stands of sea level during the last interglaciation. A/I ratios of 0.452 ? 0.026 from the base of the Kalani section are correlated with MIS 5e. This basal unit is capped by a red-orange soil and higher by a dense, clayey brown peat. A/I ratios from middle eolianite units above the brown peat and bracketing Aidera's fossil level yield a solid mean of 0.339 ? 0.006 {n = 6). Obtained from several sites along the coastline, these ratios are too low for MIS 5e. We interpret these ratios to represent a high stand of sea level late in the last mterglaciation (sensu lato) at MIS 5a around 80 ka, which has been observed elsewhere along relatively stable coastlines (Vacher and Hearty 1989, Ludwig et al. 1996, Hearty and Kaufman 2000). The level of the brown peat approximates relative paleo- sea level of circa +2.0 m during the late in- terglacial high stand. A younger generation of Holocene dunes between Mo'omomi Beach and Kap?lauo'a Point were described in detail by Wentworth (1925) and Steams (1973) (Figure 5A.B). These dunes have been an important source of fossil bird bones (Olson and Wetmore 1976. Olson and James 1982a,?). The loca- tion of these fossils near and parallel to the coast, as well as the presence of well- deveioped soils in isolated remnant patches in the same zone indicate that the original dunes may have formed as a coastal dune cordon. We suggest that dune migration 8 km inland occurred recently as blowouts, initiated by enviroimiental degradation from deforestation, trampling, and overgrazing by ruminants and excessive agricultural use since the arrival of humans in Hawai'i around 1500 yr ago (Athens 1997). A/I ratios from two sites (MMM3x/4x) were 0.274 and 0.246, respectively. The MMM4x ratio of 0.246 + 0.016 was determined from two bulk sediment samples from the same collection that produced a ^*C age of 8360 ? 60 yr B.P. (Table 1). A modem beach sample from Mo'omomi returned an A/I ratio of 0.126, indicating that reworking of Pleistocene sediments along that shoreline is not signifi- cant. A whole-rock sample from a fossil beach ridge along the soutiiem coast west of Kaunakakai yielded a mid-Holocene ratio of 0.270. This ratio is directly associated with two whole-rock **C ages of 4750 ? 70 and 5730 ? 60 yr B.P. (Fletcher et al. 1999). Extensive Holocene dune deposits also cover most of 'Ilio Point. Land snail shells associated with the bird fossils have yielded ^*C ages of 5245 ? 60 and 5510 ? 65 yr B.P. (Olson and James 1982?i) (Table 1) for these dunes. Like Mo'omomi, destruction of the vegetative cover probably initiated reactiva- tion of the dunes in more recent times. A/I ratios from the 'Ilio Point dunes (MIP6x = 0.277; MIP4x = 0.308; MIP2x = 0.262) were concordant with those from Mo'omomi. Two "Holocene" sediment samples from 'Ilio Point, however, yielded older '*C ages of 12,740 ? 90 and 12,710 ? 90 yr B.P. These older ages may be attributed to incorporation of a substantial percentage of Pleistocene grains into the Holocene samples. Further- more, although providing a maximum (Holocene) age of dune sand formation, we cannot consider these dates to be accurate, because at that time sea level was positioned over 80 m lower than present (Fairbanks 1989). Because the -80 m contour lies several kilometers offshore of 'Ilio Point, it is im- probable that substantial marine sedimenta- tion could have occurred at that time on the current shoreline. Kaua'i Both Pleistocene and Holocene eolianites are present on the southeastem Makawehi- M?ha'ulep? coast of Kaua'i. Figure 5C Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 435 shows a composite section (KMWl/3) near the fossil bird site "K2" of Olson and James (1982?). Farther north, an upper backshore Pleistocene deposit outcrops along the south end of a small pocket beach in Ha'ula Bay. Much older limestone (early Pleistocene?) is exposed on the north margin of the same bay. Samples from the Pleistocene eolianite returned A/I ratios of 0.460 ? 0.026 {n = 2) at Makawehi, 0.432 + 0.017 {? = 2) at Ha'ula Beach, and 0.497 ? 0.045 (? = 2) at Barking Sands along the M?n? Plain. On the basis of these A/I ratios, these sites are cor- related with the O'ahu control localities of MIS 5e age. The Holocene dunes along the Makawehi and Ha'ula Bay coastline are notable for im- portant bird fossil discoveries from 1976 to the present (Olson and James l9S2a,b). A bulk sediment sample from the Makawehi dunes yielded an A/I ratio of 0.245 (1) and a '*C age of 8900 ? 70 yr B.P. These data reflect the maximum age of the deposit, identifying the interval when sediments were formed offshore. Land snail, crab claw, and bird bone samples found within the dunes yielded ^*C ages of 6740 ? 80, 5145 ? 60, and 4690 ? 100 yr B.P. (Table 1), respec- tively. These ages reflect the minimum, or "occupation age" of the dune environment. Together they document a 2000- to 3000-yr interval between time of formation of the sediments offshore (ca. 8900 yr B.P.), em- placement of the dunes, stabilization by veg- etation, and occupation of the dunes by or- ganisms (5500 yr B.P.). The Polihale dunes of western Kaua'i have produced no fossils and are considerably younger than Makawehi, reflected by a submodem A/1 ratio of 0.153 (1). Maui Extensive eolianite deposits in the vicinity of Kahului, Maui (Figure 5D), were exam- ined and collected for AAR analysis. A/1 ratios from two exposures of red-stained (2.5YR 3/6) eolianite near the commimity college (AMCC) and zoological gardens (AMZB) yielded concordant ratios of 0.378 (1) and 0.380 (1), equivalent to late MIS 5e. It is interesting that the straligraphically younger eolianite, determined to be "Holo- cene" on the basis of field criteria (loose to weakly cemented light brown sand without capping calcrete and red soil), superimposed on the Pleistocene unit at AMCC (Figure 5D) and Nehe Point, returned A/I ratios of 0.372 (1) and 0.349 (I). With A/I ratios nearly identical to those from Pleistocene de- posits, the most acceptable interpretation is that the bulk of "Holocene" eohanite along the north isthmus of Maui consists largely of reworked sediments of last interglacial age. Because the Pleistocene dunes are not firmly cemented, they are subject to reactivation through a number of human or natural pro- cesses including fire, deforestation by fire or overgrazing, prolonged aridity, or washover by tsimamis. Because of the absence of data- ble material, it is uncertain whether reactiva- tion of the Kahului dunes took place previ- ous to or since human arrival on the island. Unlike the mid-Holocene dunes of Moloka'i and Kaua'i, the Maui dunes have yielded no bird fossils, which possibly may be linked to diagenesis and/or the reworking of Pleisto- cene sands. SUMMARY OF AMINOSTRATIGRAPHIC REStJLTS The Last Interg?acialion, MIS 5e and 5a, Aminozones E and C Four TIMS and alpha U-series dated sites in O'ahu provided age calibration for Ami- nozone E, the last interglaciation. The mean of A/I ratios from the calibration sites was 0.505 + 0.027 (? = 11). Twelve "test sites" on Maui, O'ahu, Moloka'i, and Kaua'i, of previously uncertain geochronological age, produced a grand mean of 0.445 + 0.058 {n= 17). Although the variance of the test sites (0.387-0.503) is statistically equivalent to that of the calibration sites (0.478-0.532), the somewhat lower mean of the test sites may be explained by one or more of the fol- lowing: (I) Most test site collections were from late MIS 5e eolianite. Thjs eolianite marks the regression from the high stand (115-120 ka?) and thus marginally postdates the emergent marine deposits of the calibra- tion sites (135-120 ka). We interpret the 436 PACIFIC SCIENCE, Volume 54, October 2000 FIGURE 6 Comparison of U-series and '*? calibrated whole-rock epimerizauon data from Bermuda, the Baha- mas and the Hawaiian Islands for the past 125 ka (A) and for the Holocene {B}. Kinetic pathways show similar trends over the sample interval, with higher rates of epimerization occurring in the Hawaiian Islands as a result of assumed wanner thermal histories among the islands. Significantly higher rates dunng the Holocene of Hawai ! (upper curve in B) may be due to local effects, wanner temperature history, or a degree of mixing with oldw deposits. What- ever these effects, they are neg?gible in Pleistocene deposits. E - F represents age of fonnation ( F ) venus age of emplacement ("E") in B. Dashed line with arrow aiong the ver?cal axis in ? is an extrapolation of the curve and an approximation of "inherited age" of modem beach deposits. entire range of values encompassed in the means of 0.505 ? 0.027 and 0.445 ? 0.058 as equivalent to the duration of MIS 5e, docu- mented by uranium ages between 135 and 115 ka; (2) A significantly lower A/1 ratio of 0.339 ? 0.006 {n = 6) from the "middle eolianite" unit Mo'omomi, Moloka'i, situated above MIS 5e marine deposits and capping soil probably represents a near-present high stand late in the interglacial (sensu lato) or MIS 5a. The possibility remains that eolian- ite units at Kahuku and Laniloa, O'ahu, could also correspond with MIS 5a. Addi- tional tests are warranted on these sites; (3) The possibility of local and regional temper- ature varibility of the test sites (e.g., wind- ward versus leeward); and (4) Some sites may be more susceptible to reworking (receiving old sediment) than others and may account for some of the variability in A/I ratios. The Holocene, MIS 1, and Aminozone A Three subgroups of A/1 ratios are rec- ognized among Holocene deposits of the islands. Samples from modem beach and re- cent dune deposits from several islands return A/I ratios averaging 0.112 ?0.025 (n = 7) (Table 2). Extrapolation of the A/I versus ^*C age (Figure 6B, dashed hne) estimates an "inherited age" of 1000 to 2500 yr for the modem beach and dunes. A second unit is represented by T-2 m beachrock at '?hikilolo in O'ahu, which yielded a ratio of 0.215. In- dependent studies (Jones 1992, Fletcher and Jones 1996, Grossman and Fletcher 1998) Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 437 suggest a correlation of the beachrock with a mid-Holocene high stand. Because of the apparent "inherited age" of the beachrock sand, the A/1 ratio indicates a time of for- mation of the sand some 1000-2500 yr earlier. The oldest Holocene subgroup from dune sites on Moloka'i and Kaua'i yielded consistent ratios averaging 0.266 ? 0.022 (rt = 8), which equate with bulk sediment ages of 8600 ? 70 yr B.P. ''*C ages on organ- isms inhabiting the dunes center on 5500 yr B.p,, bracketing the time of emplacement of the dunes within this interval. Comparison of Whole-Rock Ratios with Those of Bermuda and the Bahamas As predicted by epimerization kinetic models (Miller and Brigham-Grette 1989), the increasing temperature histories (MAT by proxy) from Bermuda, Bahamas, and Hawai'i yield incrementallv higher overall mean ratios of 0.29 + 0.03, 0.38 ? 0.02, and 0.47 ? 0.05, respectively, for MIS 5e (Figure 6 and Table 3). In finer detail. MIS 5e sites from similar latitudes (~2rN) in the Bahamas and Hawai'i produce similar ratios. For example, mean island A/I ratios from Inagua (0.477 - 0.014 [n = 3]) in the Bahamas (Hearty and Kaufman 2000) and those from Kaua'i (0.450 ? 0.022 [? = 5]), O'ahu (0.485 ?0.055 [n = 15]), and Moloka'i (0.462 ? 0.061 [n = 5]) show close corre- spondence (Figure 3). Because of the sub- stantially higher epimerization rate during the Holocene (Hearty and Aharon 1988, Hearty and Dai Pra 1992, Miller et al. 1999), the separation of A/I ratios between the Baha- mas and Hawai'i are greater: the whole- rock ratio at 5000 yr B.P. from the Bahamas is 0.11 ?0.03, whereas bulk sediment ages of 8600 yr B.P. from Hawai'i average 0.27 ?0.02. This large degree of separation of ratios appears to be primarily the result of the warmer temperature history of Hawai'i during the Holocene, but may also be affected by "fast" epimerization grain constituents and some degree of mixing with older sediments. Overall, whole-rock A/I ratios mirror morphostratigraphic relations, increase ap- propriately with greater stratigraphie age, and produce concordant numbers from both independentiy dated and stratigraphically equivalent age deposits from widespread localities. IMPLICATIONS OF A HAWAIIAN AMINOSTRATIGRAPHY Sea Leve! History and Lithospkeric Flexure The AAR results correlate 16 last inter- glacial sites on four Hawaiian islands. Cali- bration is provided from four O'ahu sites for the remaining 12 sites of previously uncertain geochronological age. At each of the calibra- tion sites, sedhnentary structures and in situ coral growth indicated higher than current sea levels ranging from +5.5 to +9 m above sea level, supporting the sustained uplift of O'ahu. In comparison with Bermuda and Bahamas MIS 5e sites, the difference between both eariy and late high stand levels of 3 m would yield an O'ahu uplift rate of 0.024 m/ ka. If this rate were applied to the interpreted MIS 5a site at Kalani Point, Moloka'i, it would predict an 80 ka sea level 1.9 m above present, given previous doctimeniation of a near-present MIS 5a sea level on stable coasthnes (Vacher and Heart>' 1989. Hearty 1998). The presence of a marsh peat at -2 ra associated with MIS 5a eolian deposits at Kaiehu and Kalani supports a similar rate of uplift between O'ahu and Moloka'i. Older emergent carbonate deposits have been de- scribed on Moloka'i (Grigg and Jones 1997), but their possible origin by tsunamis (Moore et ai. 1994) as well as their ages remain equivocal. On Hawai'i, Maui, and Kaua'i, only eolian and uppermost backshore de- posits have been observed above sea level, suggesting subsidence at rates of greater than 0.048 m/ka (6 m/125 ka) of these islands. In general, these findings lend support to the Grigg and Jones (1997) lithospheric flexure model. The ^*C and AAR data from Holocene deposits indicate that a major depositional event occurred between 8600 and about 5500 yr B.p. Extensive dunes were emplaced along windward coastiines as postglacial sea level < 2 +i? S IS+I ? M ? oes? o ? s? _^ -H-H -H3 r^ ^ ?n vo do s? o -H 5-5 O < ?<<.--0 ' : ? ? .? ?^^ 2 ? ?- W ^- I = = S o - 5 T3 .- ^ ce H ) *^ es < 2 < p 3 ! 1 1^ 'S ?O 5 5? 1 ? c 1 TJ .?-? 5 " ?>?' '^ :ci u i s o ^ ? ?2 ? ? "Sj s rt? := w 's ? ^ r? . s i^ ^ 11 o c O iZ < f^-s ? c ??? .-^n s. < r^ m -? -^ O z 1 >^ i fi? O = ri n ?? rsi f^ i ZI = e e S o ? 1 : 3: < = e o C! ? o , I 'rc ^ I 2 e: -^? -H-H -H +! -H S < r^ fn o\ a\ \o r^ V ?.:^ S? ? O c o (5 < < _? i? < ?= _^- 2 II S ?I < i i i re i? 1 II S ? 3 5 z E "H = 5 ca 5 >. c HS- V ^ u B; es *> c ^ t- ^ 4J ^ ^i u u O i t- c 2C = y E a. > -^ < ^^ s ^? ? Si X .ir c ? < i r^ fN ? <<< oS 'J: i?. OS ^ ! ^ < ^ ?f? ;r j^ s Js 5:i ^=2 ; 'S 5 in .CI Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 439 approached the current datum, rising from -15 m to near present during this 300O-yr interval of the Holocene (Grossman and Fletcher 1998). Evidence of a higher than present sea level around 5000 yr B.P. is in- ferred from beachrock with Echinometra borings and marsh peat above +2 m at '?hikilolo, O'ahu. The sediments composing the modem beaches and recent dunes of the Hawaiian Islands are marked by A/I ratios averaging 0.112 ? 0.025, which reflects the interval of formation and aggregation or "inherited age" of modem coastal sediments. Dime Bird Fossils Previous '*C dates obtained from land snails and crab claws associated with bird fossils at '?lio Point, Moioka'i, and Maka- wehi, Kaua'i, as well as from bird bones themselves (Olson and James 1982?!), al- though once considered equivocal, now ap- pear perfectly in Une with the maximal ^*C ages for the dime sand itself (Jable 1), sup- ported by concordant AAR ratios (Table 2). A Holocene age is inferred from the loosely consolidated to unconsolidated carbonate sands composing the dunes, with hollow root casts typical of very young deposits in other locations (White and Curran 1988), their youngest stratigraphie position, and the weak development of soils. The ?lio Point and Makawehi sites are of further interest because the sands here are perched above erosiona! cliffs of basalt that would have severed the supply of sand. Well before 5500 yr B.P., organisms inhabited the dunes, which were presumably stabilized by vegetation. The fossils from which the '*C dates were obtained are unlikely to have been buried after the sand source was cut off, sug- gesting in turn that the erosional features along these coasts are younger than indicated by the '^C dates. Radiocarbon dates from dune sites and other fossil locahties have shown that most (maybe all) of Hawai'i's extinct fossil birds were still aiive in the mid- to late Holocene, well after any climatic changes of the Pleis- tocene that might be invoked as possible causes of extinction (Olson and James 1982a, 1991, James et al. 1987, Paxinos 1998, Bumey et al. 1999). Indeed, it may be that human activity had a profound impact on the land- scape itself, where dunes that were perhaps stable and vegetated for several thousand years were defoliated and deflated. At Mo'omomi, parabolic and "blowout" dunes migrated in some cases several kilometers inland, perhaps facilitated by sustained trampling and grazing by ruminants. It was this renewed activity and transport of the dunes that exposed the avian fossils that once lay buried within the dunes. At Makawehi, Kaua'i, hawk bones re- ported by Olson and James (1997) came from deposits that, on the basis of field relation- ships, are also interpreted to be Pleistocene. Whole-rock A/I ratios from this deposit, sit- uated beneath a thick calcrete and paleosol, yield ratios that correlate directly with sites on O'ahu. Mean A/I ratios are 0.460 ? 0.026 (n = 2), equivalent to MIS 5e. Exposure of previously buried fossils in the dunes appears to be slow, especially on Kaua'i. All of the major outcrops of bird fossils in the Makawehi dunes were found when these dunes were first explored paleon- tologically in 1976. Since then, little else of signjficance has been found eroding naturally in the dunes despite two powerful hurricanes that passed over this area in 1982 ('Iwa) and 1992 ('Iniki). The process is being slowed further by spread of vegetation, especially introduced plants such as Casuarina and Prosopis. Two dune sites that have produced fossil birds are Pleistocene rather than Holocene in age. At Site 1 (Figure 5^), at Mo'omomi, Moioka'i, the complete articulated skeleton of an extinct flightless waterfowl (holotype of Thambetochen chauiiodous Olson & Wetmore, 1976) was recovered by Joan Aidem from a weak soil interbedded with eolianite (Steams 1973). Whole-rock samples from Kalani Point yielded MIS 5e (125 ka) A/I ratios from the base of the section, probable MIS 5a (80 ka) ratios from the middle units containing the fossils, &nd Holo- cene (8-5 ka) ratios from the upper part (Figure 5B, Table 2). 440 PACIFIC SCIENCE, Volume 54, October 2000 CONCLUSIONS U-series and ^*C calibration of whole- rock A/I ratios has enabled the correlation and dating of numerous sites of previously unknown or uncertain geochronological age. Three aminozones are recognized: (1) Ami- nozone A, represented by three subgroups of the Holocene with correlated ages of 8500- 5500 yr B.P., 5000 to 3000 yr B.P., and mod- em beaches and dunes. The "inherited" ages of whole-rock samples from these subgroups average about 1000-2500 yr; (2) Aminozone C (MIS 5a, 80 ka), tied to eoUanite deposits on the north shore of Moloka'i associated with Thambetochen chauliodous, a flightless anserifonn that was once abundant in the Hawaiian Islands; and (3) Aminozone E (MIS 5e, 125 ka), composed of numerous in- dependently dated last intergiacial sites from O'ahu, from which it is jjossible to make correlations with noncoraliferous deposits on several islands. Older deposits and amino- zones have also been defined and will be presented in forthcoming papers. On the basis of the height of emergent shoreline deposits, these data confirm uplift rates on the order of 0.020 ? 0,005 m/ka on O'ahu and Moloka'i. Because emergent MIS 5e subtidal deposits are not observed above sea level on Hawai'i, Maui, and Kaua'i, these data also suggest subsidence rates in excess of 0.048 m/ka for those islands. These findings generally support the lithospheric flexure model of Grigg and Jones (1997). Finally, our findings place time constraints on the extinction of a variety of avian forms that occupied the Hawaiian Islands through the late Quaternary up to historical times. ACKNOWLEDGMENTS We are indebted to A. Zeigler (Bishop Museum), D. Bumey (Fordham University), and J. Aidem (Moloka'i) for sharing their knowledge of the natural history of the Hawaiian Islands. S. Anthony assiduously edited the manuscript. Work on Kaua'i and several '*? dates were determined under NSF grant DEB-9707260 to D. Bumey. Whole-rock AAR samples were analyzed under a collaborative agreement with the Amino Acid Laboratory of Northern Ari- zona University (D. Kaufman, director). LITERATURE CITED ATHENS. J. S. 1997. Hawaiian native lowland vegetation in prehistory. Pages 248-270 in P. V. Kirch and T. L. Hunt, eds. Histori- cal ecology in the Pacific Islands. Yale University Press, New Haven, Connecti- cut. BRETZ, J. H. 1960. Bermuda: A partially drowned late mature Pleistocene karst. Geol. Soc. Am. Bull. 71:1729-1754. Bt;RNEY, D. A., L. P. BURNEY, M. BURNEY, D. MCCLOSKEY, S. L. OLSON, H. F. JAMES, F. V. GRADY, W. L. WAGNER, W. KiKUCHi, D. KiKUCm, R. GAGE II, and R. NiSHEK. 2000. Holocene lake sediments in the Maha'ulepu Caves of Kaua'i: Evi- dence for a diverse biotic assemblage from the Hawaiian lowlands and its transfor- mation since human arrival. Ecol. Mono- gr. (in press). CLAGUE, D. A., and G. B. DALRYMPLE. 1989. Tectonics, geochronology and origin of the Hawaiian-Emperor Chain. Pages I SS- II 7 in E. L. Winterer, D. M. Hussong, and R. W. Decker, eds. Geology of North America: The eastern Pacific Ocean and Hawaii. The Geological Society of Amer- ica, Boulder, Colorado. DANA, J. D. 1890. Characteristics of volca- noes, with contributions of facts and prin- ciples from the Hawaiian Islands. Dodd, Mead, and Company, New York. DARWIN, C. 1839. On the structure and dis- tribution of coral reefs. Reprint. Ward, Lock, Bowden and Company, London. EASTON, W. H., and T. L. Ku. 1981. 2?Q'^? 234" dates of Pleistocene deposits on Oahu. Bull. Mar. Sei. 31:552-557. FAIRBANKS, R. G. 1989. A 17,000-year glacio-eustatic sea-level record: Influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature (Lond.) 342:637-642. FLETCHER, C. H., Ill, and A. T. JONES. 1996. Sea-level highstand recorded in Holocene Carbonate Deposits in the Hawaiian Islands?HEARTY ET AL. 441 shoreline deposits on Oahu, Hawaii. J. Sediment. Res. 66:632-641. FLETCHER, C. H., Ill (ED.), P. J. HEARTY, C. MURRAY-WALLACE, C. R. GLENN, C. E. SHERMAN, and B. POPP. 1999. Molokai field trip guidebook for participants. The non-steady state of the inner shelf and shoreline: Coastal change on the time scale of decades to millenia. University of Hawai'i, Honolulu. GR?MLICH, J. W., V. A. LEWIS, and J. J. NAUOHTON. 1971. Potassium-argon dating of Holocene basalts of the Honolulu Volcanic Series. Geol. Soc. Am. Bull. 82:1399-1404. GRIGG, R. W., and A. T. JONES. 1997. Uplift caused by lithospheric flexure in the Ha- waiian Archipelago as revealed by ele- vated coral deposits. Mar. Geol. 141 : II- 25. GROSSMAN, E. E., and C. H. FLETCHER. 1998. Sea level higher than present 3500 years ago on the northern main Hawaiian Islands. Geology 26:363-366. HARE, P. E., and R. M. MITTERER. 1967. Non-protein amino acids in fossil shells, Carnegie Inst. Washington Year Book 65:236-364. HEARTY, P. J. 1998. The geology of Eleuthera Island, Bahamas: A Rosetta Stone of Qua:emary stratigraphy and sea-level his- tory. Qual. Sei. Rev. n': 333-355. HEARTY, P. J., and P. AHARON. 1988. Amino acid chronostratigraphy of late Pleistocene coral reef sites: Huon Peninsula, New Guinea and the Great Barrier Reef, Aus- tralia. Geology 16:579-583. HEARTY, P. J., and G. DAI PRA. 1992. The age and stratigraphy of Quatemarj' coastal deposits along xhs Gulf of Taranto (south Itaiy). J. Coastal Res. 8:882-905. HEARTY, P. J., and D. S. KAUFMAN. 2000. Whole-rock amino stratigraphy and Qua- ternary sea-level history of the Bahamas. Quat. Res. 54. In press. HEARTY, P. J., G. H. MILLER, C. E. STEARNS, and B. J. SZABO. 1986. Aminostraiigraphy of Quaternary shorelines around the Mediterranean basin. Geol. Soc. Am. Bull. 97:850-858. HEARTY, P. J., H. L. VACHER, and R. M, MITTERER. 1992. Aminostratigraphy and ages of Pleistocene limestones of Ber- muda. Geol. Soc. Am. Bull. 104:471-480. JACKSON, E., I. KOISUMI, G. DALRYMPLE, D. CLAQUE, R. KIRKPATRICK, and H. GREENE. 1980. Introduction and sum- mary of results from DSDP leg 55, the Hawaiian-Emperor hot-spot experiment. Pages 5-41 in E. Jackson and I. Koisumi, eds. Initial reports of the deep sea drilling project. U.S. Government Printing Office, Washington, D.C. JAMES, H. F. 1987. A late Pleistocene avi- fauna from the island of Oahu, Hawaiian Islands. Doc. Lab. Geol. Fac. Sei. Lyon 99:221-230. JAMES, H. F., and S. L. OLSON. 1991. De- scriptions of thirty-two new species of birds from the Hawaiian Islands: Passeri- formes. Omithol. Monogr. 46:1-88. JAMES, H. F., T. W. STAFFORD JR., S. L. OLSON, D. W. STEADMAN, P. S. MARTIN, and P. McCoY. 1987. Radiocarbon dates on bones of extinct birds from Hawaii. Proc. Nail. Acad. Sei. U.S.A. 84:2350- 2354. JONES, A. T. 1992. Holocene coral reef on Kauai, Hawaii. Evidence for a sea-level highstand in the central Pacific. Pages 267-271 in C. H. Fletcher and J. F. Weh- miller, eds. Quaternary coasts of the United Stales: Marine and lacustrine sys- tems. SEPM Spec. Publ. 48:267-271, Ku, T. L., M. A. KIMMEL, W. H, EASTON, and T, J. O'NEILL. 1974. Eustatic sea level 120,000 years ago on Oahu, Hawaii. Science (Washington, D.C) 183:959- 962. LLT5WIG. K. R., D. R. MUHS, K. R. SIM- MONS, R. B. HALLEY, and E. A, SHINN, 1996. Sea-level records at ~80 kys: from tectonically stable platforms: Florida and Bermuda. Geology 24-.lU-214. MILLER, G, H,, and J. BRIGHAM-GRETTE. 1989. Amino acid geochronolog>': Reso- lution and precision in carbonate fossils. Quat. Int. 1:111-128. MILLER, G. H., J. W, MAGEE, B. J, JOHNSON, M, L. FoGEL, N, A, SPOONER, M, T. MCCULLOCH, and L. K. AYLIFFE. 1999. Pleistocene extinction of Genyornis new- 442 PACIFIC SCIENCE, Volume 54, October 2000 ?oni: Human impact on Australian mega- fauna. Science (Washington, D.C.) 283: 205-208. MOORE, J. G., W. B. BRYAN, and K. R. LUDWIG. 1994. Chaotic deposition by a giant wave, Moiokai, Hawaii. Geol. Soc. Am. Bull. 106:962-967. MuHS, D. R., and B, J. SZABO. 1994. New uranium-series ages of the Waimanalo Limestone, Oahu, Hawaii: Implications for sea level during the last interglacial period. Mar. Geol. 118:315-326. OLSON, S. L., and H. F. JAMES. 1982?. Fossil birds from the Hawaiian Islands: Evidence for wholesale extinction by man before Western contact. Science (Washington, D.C.) 217:633-635. . 1982b. Prodromus of the fossil avi- fauna of the Hawaiian Islands. Smithson. Contrib.Zool. 365:1-59. 1991. Descriptions of thirty-two isotope substage 5e. Geology 21:1079- 1082. STEARNS, H. T. 1973. Geologic setting of the fossil goose bones found on Moiokai Is- land, Hawaii. Occas. Pap. Bemice Pauahi Bishop Mus. 24:156-163. . 1974. Submerged shorelines and shelves in the Hawaiian Islands and a revi- sion of some of the eustatic emerged shore- lines. Geol. Soc. Am. Bull. 85:795-804. . 1978. Quaternary shorelines in the new species of birds from the Hawaiian Islands: Non-passerifonnes. Omithol. Monogr. 45:1-88. 1997. Prehistoric status and distri- bution of the Hawaiian Hawk {Buteo solharius), with the first fossil record from Kaua'i. Bishop Mus. Occas. Pap. 49:65- 69. OLSON. S. L., and A. WETMORE. 1976. Pre- liminary diagnoses of two extraordinary new genera of birds from Pleistocene de- posits in the Hawaiian Islands. Proc. Biol. Soc. Wash. 89:247-258. PAXINOS, E. E. 1998. Prehistoric anseriform diversity in the Hawaiian Islands: A mo- lecular perspective from the analysis of subfossil DNA. Ph.D diss., Brown Uni- versity, Providence, Rhode Island. RLTTER. N. W., and B. BLACKWELL. 1995. Amino acid racemization dating. Pages 125-167 in N. W. Rutter and N. R. Catto, eds. Dating methods for Quaternary de- posits: Newfoundland. Geological Associ- ation of Canada. SHERMAN, C. E., C. R. GLENN, A. T. JONES, W. C. BURNETT, and H. P. SCHWARCZ. 1993. New evidence for two highstands of the sea during the last interglacial, oxygen Hawaiian Islands. Bemice P. Bishop Mus. Bull. 237:1-57. STEARNS, H. T., and K. N. VAKSVIK. 1935. Geology and ground-water resotirces of the island of Oahu, Hawaii. O'ahu Divi- sion of Hydrography, Honolulu. SZABO, B. J., K. R. LUDWIG, D. R. MUHS, and K. R. SIMMONS. 1994. Thorium-230 ages of corals and duration of the last intergiacial sea-level high stand on Oahu, Hawaii. Science (Washington, D.C.) 266:93-96. VACHER, H. L., and P. J. HEARTY. 1989. History of stage-5 sea level in Bermuda: With new evidence of a rise to present sea level dtiring substage 5a. Quat. Sei. Rev. 8:159-168. WATTS, A. B., and U. S. TEN BRINK. 1989. Crustal structure, flexure, and subsidence history of the Hawaiian Islands. J. Geo- phys.Rei. 94:10.473-10,500. WEHMILLFR, J. F. 1984. Interlaboratory comparison of amino acid enantiomeric ratios in fossil Pleistocene mollusks. Quat. Res. 22:109-120. WENTWORTH, C. K. 1925. The desert strip of West Moiokai. Univ. Iowa Stud. Nat. Hist. 11:41-56. WENTWORTH, C. K.., and J. E. HOFFMEISTER. 1939, Geology of Ulupau Head, Oahu. Geol. Soc. Am. Bull, 50:1553-1572. WHITE, B., and H. A. CURRAN. 1988. Meso- scale physical and sedimentary structures and trace fossils in Holocene carbonate eolianites from San Salvador, Bahamas. Sediment. Geol. 55:163-1S4. WINCHELL, H. 1947. Honolulu Series, Oahu, Hawaii. Geol. Soc. Am. Bull. 58:1-48.